Notes The Reaction of 1-Alkynes with Organometallic Compounds. XII.' The Rate of Reaction Ethyl Grignards with 1-Hexyne in Certain Ethers JOHN H. WOTIZAND GORDON L. PROFFITT~ Department of Chemistry, Marshall University, Huntington, West Virginia
compared to a 10 to 1 molar ratio concentration in ethyl ether. The rate insensitivity of the presently used Grignard reagents to changes in reactant concentrations was recently reported.' In the present study attempts were also made to adjust the bromine to magnesium ratios in the Grignard close to 1.08 by the addition of appropriate quantities of magnesium bromide prepared in specified ethers. TABLE I THEREACTION OF ETHYLGRIGNARD WITH ~ - H E X Y N INE CERTAIN ETHERS C4H&=CH C2H6MgBr+C2He C4H9C=CMgBr
Received June 16, 1964
The rate of reaction of ethylmagnesium bromide with 1-hexyne has been extensively studied and was found to be a function of the structure of the acetylene3 and the Grignard reagenk4 The rate of reaction was also found to depend on the nature and structure of certain cosolvents. Thus, the addition of tertiary amines to ethyl ether solutions of ethylmagnesium bromide produced significant changes in the rate^.^,^ Similarly, the addition of a certain quantity of dioxane to the ether solutions of ethylmagnesium bromide also produced significant changes in rates which were dependent not only on the amount of added dioxane, but also on the time lapsed between the addition of dioxane and the reaction of the product with 1-hexyne.' On the other hand, the addition of tetrahydrofuran was found to be void of detectable changes in the rate of reaction. In order to ascertain the influence of the structure of an ether used as a solvent for ethyl Grignards in its reaction with 1-hexyne, the Grignard reagents were prepared directly in isopropyl and butyl ether, as well as in tetrahydrofuran and 2-methyltetrahydrofuran. In this way we avoided the chance of having traces of ethyl ether present, which is frequently the case in attempts to displace ethyl ether with some other solvent. The rates of reaction were established by measuring the evolution of ethane, using the experimental conditions previously described. In Table I we are listing the observed rates in terms of relative reactivities. The standard of comparison was again an equimolar solution of ethylmagnesium bromide and 1-hexyne in 1 N concentration with a bromine to magnesium ratio of 1.08. Such a reaction had an arbitrary assigned relative reactivity of 100. However, in the present study we have used a 10 to 1 molar ratio of hexyne to Grignard. This, in turn, was (1) P a r t XI: J. H. Wotiz and G. L. Proffitt. Proc. West. V a . Acad. S c i . , 88, 107 (1964). (2) Abstracted from a portion of the Master of Science Thesis of G . L. P., Marshall University, 1964. (3) J. H. Wotiz. C. A. Hollingsworth, and R . E. Dessy, J . Ow. Chem., 40, 1545 (1955). (4) J. H. Wotiz, C. A. Hollingsworth, and Soc., 77, 103 (1955).
+
Molar ratioa
a
+
Solvent
Br : Mg
1:1 Et20 1.08 1O:l i-Pr2O 1.04 10: 1 BUZO 1.06 10: 1 Et20 1.05 10:1 THF 1.05 10:1 2-MeTHF 1.07 Hexyne to Grignard. Standard of comparison.
Relative reactivity
1OOb 155 130 110 105
37 See ref. 3.
From the relative reactivities listed in Table I we see that the structure of the ether produced changes in the rate. The fastest reaction was in isopropyl ether. The difference in the reactivity between tetrahydrofuran and ethyl ether is probably within the limit of experimental error. The slowest reaction was observed in 2-methyltetrahydrofuran. The observed relative reactivities are generally in line with the determined basicities of ethers toward magnesium and other Lewis acids of comparable sizes8 Thus our relative reactivities were a measure of relative base strength'of studied ethers toward magnesium, which was a combination of electronic and steric effects. The found order of basicity was i-PrzO < BuzO < Et20 < T H F < 2-MeTHF. Since we have previously demonstratedgthat halogenfree diethylmagnesium in ether reacts about three times as fast with hexyne as a Grignard with a bromine to magnesium ratio of 1.08, it became of interest to ascertain the effect of the bromine to magnesium ratio on the rate of reaction in tetrahydrofuran. In Table I1 we are listing ouI findings. A comparison of previous data in ethyl ethere shows that there are only small differences between the rates in ether and tetrahydrofuran in spite of the fact that the rates in ether were ascertained a t equimolar concentration, whereas the rates in tetrahydrofuran were a t 10 to 1 molar concentration, hexyne to diethylmagnesium. We have thus demonstrated again' that with the presently used magnesium and/or Grignard solutions there
R. E. Dessy, J . A m . Cham.
(5) J. H. Wotiz, C. A. Hollingsworth, R. E. Dessy, and L. C. Lin, J . Ow. Chem., 48, 228 (1958). (6) J. H. Wotiz, C. A. Hollingsworth, and A. W. Simon, ibid., 44, 1202 (1959). (7) J. H. Wotiz. C. A. Hollingsworth, and R. E. Dessy. J. A m . Chem. Soc., 7 8 , 1221 (1956).
(8) (a) E. M. Arnett and C. Y. Wu, ibid., 84, 4999 (1960); (b) 0. P. Brand, M. Tamres. and S. Searles, ibid., 84, 2129 (1960); (0) H. H. Sisler and P. E. Perkins, ibid., 78, 1135 (1956); (d) P. A. D . de Moine, J . Chem. Phys., 46, 1199 (1957); (e) A. Kirrman. R. Hamlin, and S. Hayes, Bull. soc. chim. France. 1395 (1963). (9) J. H. Wotiz, C. A. Hollingsworth, and R . E. Dessy, J . 070. Chem..
41, 1063 (1956).
1240
NOTES
APRIL 1965
1241
carbonyl oxygen even in highly substituted ketones. TABLE I1 They were recovered unchanged aftel. refluxing with THEREACTION OF ETHYLMAQNESIUM BROMIDEAND DIETHYLMAQNESIUM WITH I-HEXYNE IN TETRAHYDROFURANphosphorus pentachloride, either in the presence or C~HBC-CH : organometal, molar ratio
1:l 10:1 10: 1 1O:l 1O:l
Br : Mg
Relative reactivity
0 0
1.oo 1.05 1.17
350 330 134 105 80
is little dependence of the rate of reaction on the molar concentration of reactants. Experimental The gas-collecting system and the determination of rates was the same as previously d e ~ c r i b e d . ~Ethers were purified by distilling from ethylmagnesium bromides prepared in them. 1Hexyne was purchased from Farchan Research Laboratories, redistilled, and sealed in vials until ready for use. Magnesium turnings were from an unknown lot purchased from Eastman Organic Chemicals. Analysis of the metal was not known. Diethylmagnesium was prepared by the dioxane precipitation method.l0 The solvents were removed in vacuo. The light green solid remaining wm redissolved in dry tetrahydrofuran forming a light yellow solution. It contained less than 0.3 mole yoof bromine. A tetrahydrofuran solution of magnesium bromide was prepared by treating bromine with magnesium in dry tetrahydrofuran. The solution was analyzed for total magnesium and bromine content. The preparation of magnesium bromide in ether was previously described." The Grignard reagents were prepared from ethyl bromide in the appropriate ethers. With the exception of ether and tetrahydrofuran, the bromine to magnesium ratios of the formed solution were sufficiently close to 1.08 so that they did not require adjustments with solutions of magnesium bromide in the appropriate ethers. (10) C. R. Noller and W. R. White, J . A m . Chem. SOC.,6S, 1354 (1937). (11) W. E . Doering and C. R. Noller, ibid., 61, 3436 (1939).
Perhalo Ketones. IV.' The Reaction of Perhaloacetones w i t h Phosphorus Pentachloride BASILS. FARAH AND EVERETT E. GILBERT^ General Chemical Research Laboratory, Allied Chemical Corporation, Morris Toumship, New Jersey 07960 Received September 33,1964
absence of inert solvents. Although pentachloroacetone is converted to heptachloropropane with PC16 at 180" in 6-8 hr.,6 the perhaloacetones gave extremely low yields a t 200" for 24 hr. with excess reagent in a rocking autoclave., Not until temperatures of 250" or higher were employed did substantial reaction occur, except in the case of hexachloroacetone which reacted at 230". A t 275-300' fair yields of the corresponding perhalopropanes were obtained (Table I) in 5 hr. Since the unreacted ketones are either readily soluble in aqueous base: or react with it, the perhalopropanes were obtained in high purity. These compounds have previously been prepared only by mu1t istep syntheses. TABLE I PREPARATION OF PERHALOPROPANES -yo Perhalopropane formed
(1) Paper I: B. Sukornick, Or#. S u n . , 40, 103 (1960); Paper 11: P. Eaton, E. J. Carbon, P. Lombardo. and P. Yates, J. Ora. Chem., 41, 1225 (1960); and Paper 111: B. Farah and S. Horensky. ibid.. 48, 2494 (1963). (2) To whom inquiries should be sent. (3) H. E. Simmonsand D. W. Wiley, J. A m . Chem. Soc., 84, 2288 (1960). (4) See, for example. ( a ) S. Andreades, U. 9. Patent 3,030,409 (1962); (b) S. Andreades, U. 9.Patent 3,040,085 (1962); (0) I. L. Knunyants, et aE., I r v . Akad. Nauk S S S R , Otd. K h i m . Nauk, 927 (1962). (5) See, for example, (a) D. C. England, J . A m . Chem. Soc., 89, 2205 (1961); (b) J. F. Harris and D. D. Coffman. ibid., 84, 1553 (1962).
fluorinCalcd. Found
FaCCClzCFs 33-34b 31 51.7 51.6 FaCCClzCFzCl 72c 50 40.2 40.0 ClFzCCC12CFzCl 112-113d 61 29.8 29.9 ClF&CClzCFCl2 152d 48 21.2 21.1 ClzFCCClzCFClz 194* 62 12.9 13.2 ClaCCC1zCCl3' 270-2710 85 a Based on the amount of ketone used. J. T. Maynard, J . Org. Chem., 28, 113 (1963). A. L. Henne, A. M. Whaley, and J. K. Stevenson, J . Am. Chem. Soc., 63,3478 (1941). A. L. Henne and M. W. Renoll, ibid., 61, 2489 (1939). e A. L. Henne and E. C. Ladd, ibid., 60, 2491 (1938). f Run a t 230". 0 F. Kraft and V. Merz, Ber., 8, 1296 (1875). Product gives an infrared spectrum identical with that of sample purchased from Aldrich Chemical Co.
This resistance of the perhaloacetones to phosphorus pentachloride is attributed to the negative inductive effect of the halogen clusters surrounding the carbonyl group,* which reduces electron availability at the carbonyl oxygen and inhibits attack by the chlorophosphonium ion, PC14+.9 I t would also make the formation of the carbonium ion I difficult, since departure of the oxygen atom with its pair of bonding electrons would be electronically unfavorable.
PCl,
Xs~COCX3
The perhalofluorochloroacetones are not readily susceptible to attack by reagents that normally lead to replacement of the carbonyl oxygen.3 This inertness is in marked contrast with their considerable reactivity toward nucleophiles4 and other polar function~.~ This study has indicated that these perhaloacetones are resistant to the action of phosphorus pentachloride under conditions which lead to replacement of the
% yielda
B.P., OC.
+
+
x'A
Cl-
XICCCX0 + XJC CXa +
I 0 I
PCl,
I
c1
PClr CI -
x*c~cx3 +x 3 c c c l ~ c x a I (6) P. Fintsch, A n n . , 497, 314 (1897). (7) (a) W. K. Pearlson. "Fluorine Chemistry," Vol. I, J . H. Simons. Ed., Xcademic Press Inc., New York. N. Y., 1950, p. 482; (b) C. Woolf. Ahstracts, 132nd National Meeting of the American Chemical Society, New York, N . Y.. Sept. 1957, p. 23M. (8) E. T. McBee, Y. S. Kim, and H. P. Braendlin, J . A m . Chem. SOC.,84, 3154 (1962); H. P. Braendlin and E. T. McBee, Advan. Fluorine Chem., 3, 1 (1963). (9) For an elegant discussion of the mechanism of the reaction of phos-
phorus pentachloride with ketones. see (a) M. S. Newman and L. L. Wood, J. A m . Chem. Sac., 81, 4300 (1959); (b) M. 9. Newman, G. Fraenkel, and W. N . Kirn, J . Ow. Chem., 98, 1851 (1963).
NOTES
1242 Experimental
Typical Procedure.-The ketone (0.5 mole) and phosphorus pentachloride (1.O mole) were sealed in a stainless steel autoclave, and the mixture was heated a t 275-300' for 5 hr. with shaking. The vessel was then cooled to room temperature and the contents were transferred to a separatory funnel containing 1 kg. of crushed ice. The mixture was shaken and the lower layer was withdrawn. It was washed three times with 50-ml. portions of 5y0 sodium hydroxide solution, dried over calcium chloride, and distilled.
The Effects of Hydrogen Bonding on the Absorption Spectra of Some Substituted Benzaldehyde Tosylhydraeone Anions HELENW. DAVIESAND MEYERSCHWARZ Harry Diamond Laboratories, Washington,D . C . 80488 Received August 19, 1964
The decomposition reaction of the anions of aryland alkylsulfonylhydrazones, referred to as the Bamfordatevens reaction, has attracted considerable attention in the past few The reaction affords a convenient way of generating carbenes when it is carried out in aprotic media (eq. 1-3), while the de-
VOL. 30
and co-workers617in their studies of the effects of hydrogen bond formation on the absorption spectra of many aromatic substances having hydrogen bond donor and acceptor sites. They showed that where the effect of hydrogen bonding with the solvent is to reduce the polarity of the solute molecule a blue shift of the T -+ a* absorption band (K-band) is observed. A relevant example is the blue shift which occurs in the absorption spectrum of the p-nitrophenol anion on changing the solvent from pyridine to water.' In this case hydrogen bonding localizes the negative charge on the oxygen, thereby decreasing the dipole moment of the anion and stabilizing its ground state and causing a shift of the primary band to shorter wave lengths. Brealey and Kashas have also discussed the effect of hydrogen bonding on electronic absorption spectra. They showed that the blue shifts observed in the n -+ a* absorption bands (R-bands) of pyridazine and benzophenone upon changing from hydrocarbon to hydroxylic solvents are mainly due to hydrogen bonding, which stabilizes the ground state more than the excited state of the molecule. This paper reports our findings of a hypsochromic shift of the primary absorption bands of several anions of substituted benzaldehyde tosylhydrazones upon changing from aprotic to protic solvents. It was also found that the rate at which the anions decompose is retarded by the same solvent changes. Experimental
Spectra.-Spectra were measured on a recording spectrophotometer, using 1-cm. quartz cells. Solvents.-Solvents used were all Spectroquality, with the exception of N,N-dimethylformamide (DMF), which was reagent grade. Distilled water was used. Toluene-p-sulfonylhydrazones.-The toluene-p-sulfonylhydrazones were prepared in ethanol solutions and were recrystallized from methanol-water mixtures. The p-diethylamino compound was prepared in an ethanol-acetic acid solution. R R The aldehyde tosylhydrazones which are shown in Table I were \ + - nonprotic prepared. C==N=NC ': (3) Excess diethylamine was added to lo-' M solutions of the solvents / / tosylhydrazones to obtain the anion solutions, except in solutions R R of the p-diethylamino compound, in which alcoholic NaOH was used. R R R The spectral absorption data are given in Tables I1 and 111 \ + - protic \ + \+ where Amax refers to the wave length of the primary absorption C=N=N + CH-NGN + (4) solvents / band. The p-nitrobenzaldehyde tosylhydrazone was studied / R R R most thoroughly, since it displayed the greatest shift. p-Nitropheny1diazomethane.-p-Nitrobenzaldehyde tosylhydrazone (1 g.) was dissolved in 10 ml. of DMF. Diethylcomposition in protic solvents leads to products formed amine (1 ml.) was added and the solution was allowed to stand via an ionic mechanism (eq. 1, 2, and 4). The intera t room temperature for about 1 hr. On the addition of water, mediates in both cases are aryl- or alkyldiazomethanes. p-nitrophenyldiazomethane, m.p. 80' dec., se orated and waa 370 mp (log e recrystallized from acetone; yield 59%,' : :A In the case of the tosylhydrazones of some aromatic 4.29). In the infrared spectrum a strong band a t 4.84 p (>CNzO) aldehydes and ketones, the alkaline decomposition rewas observed. The identical compound was also prepared by action proceeds at sufficiently low temperatures that the oxidation of p-02NC6H4CH=NNH2with active manganese the diazo compound may be conveniently i ~ o l a t e d . ~ dioxide. During the course of some synthetic work using this Anal. Calcd. for C7H6Na02:C, 51.5; H , 3.1. Found: c, 51.6; H,3.1. reaction we observed large differences between the abp-Nitrobenzaldehyde N-Methyltoluene-p-sulfonylhydrazone. sorption spectra of the anion of the tosylhydrazone of -p-Nitrobenzaldehyde N-methyltosylhydrazone was prepared pnitrobenzaldehyde in protic and nonprotic solvents. in 80% yield according to Dornow and Bartsch,l by alkylation Similar observations had been made earlier by Burawoy of the tosylhydrazone with diazomethane. It melted a t 165' (1) W. R. Bamford and T. S. Stevens, J. Chem. Soc.. 4735 (1952). (2) A. Dornow and W. Bartsch. Ann., 609, 23 (1957). (3) J . W. Powell and M. C. Whiting. Tetrahedron, 7 , 305 (1959). (4) L . Friedman and H. Shechter, J . Am. Chem. Soc., 81, 5512 (1959); 88, 1002 (1960); 88, 3159 (1961). (5) (a) D. G. Farnum, J . O w . Chem., 98, 870 (1963); (b) G. L. Closs and R. A . Moss, J . A m . Ckem. Soc.. 86, 4042 (1964).
(6) A. Burawoy, "Hydrogen Bonding," D. Hadzi, Ed., Pergamon Press, Ltd., London, 1959, p. 259.' (7) W. A. Lees and A. Burawoy, Tetrahedron, 19, 419 (19831, and references cited therein. (8) G. J. Brealey and M. Kasha, J. A m . Ckem. SOC.,7 7 , 4462 (1955). (9) P. Yates, et ai.. ibid.. 79, 5756 (1957).
1243
NOTES
APRIL 1965
TABLEI ALDEHYDETOSYLHYDRAZONES hmsx
-Calod.% C
Aldehyde
M.p., OC.
p-OzNCBH4CHO VZ-O~NC~H~CHO 2,PClzCsHsCHO CI-HOCSH~CHO P-(C~HS)ZNC~H,CHO CsHsCHO See ref. 1.
158 dec. 160-161 dec. 184 dec. 155 dec. 172-173 dec. 127 dec. (lit." m.p. 128)
TABLE I1 ABSORPTION MAXIMA OF THE p-NITROBENZALDEHYDE TOSYLHYDRAZONE ANION Solvent
A, mr
475 475 473 450 427 41 1 405 403 403 403 400 ccl4 395 CzHrOH-HzO (7 :3) 395 CHsOH 387 CHjOH-HzO (9 : 1 ) 386 Hz0 373 Estimated by extrapolation to zero time. HCON( CH3)z CsHsN CH3COCHs CHaCN CH&OCHa-H20 (9 : 1 ) (CH3)zCHOH n-C4H90H C2HSOH HCONH2 C~HB CHCls
a
log
ea
4.15 4.14 4.18 4.02 4.25 4.20 4.18 4.23 4.25 4.02 3.97 3.97 4.22 4.24 4.25 4.22
with decomposition after recrystallization from ethanol; A%": 323 mp (log c 4.24). A n d . Calcd. for C ~ S H I ~ N ~ OC,~ S54.0; : H , 4.5. Found: C, 54.1; H , 4.6.
Results and Discussion As reference to Tables I1 and 111 shows, all of the anions studied exhibited qualitatively the same behavior. As observed with the p-nitrophenol anion,I in each case replacement of a nonhydrogen-bonding by a hydrogen-bonding solvent resulted in a blue shift of the primary absorption band. That these shifts occur with the anions only was shown by the fact that the parent compounds showed only very minor shifts with changes of solvents. Neither was the spectrum of p nitrobenzaldehyde N-methyltosylhydrazone, in which anion formation is precluded, affected by the addition of base or change of solvent. The effect of protic solvents is to localize the negative charge at its original position on the nitrogen atom through hydrogen bonding, thereby inhibiting the redistribution of electrons required in the excitation process. The transition energy is thus increased in hydrogen-bonding solvents, as manifested by a shift of the absorption maximum to shorter wave length. It would seem reasonable then that in any solvent where hydrogen bonding is possible the amount of shift of the absorption maximum should be related to the hydrogen-bonding ability of the solvent. Indeed we find that for each solvent the hypso(10) G. C. Pimentel, J . A m . Chem. Soc., 79, 3323 (1957). (11) E . D. Becker, Spectrochim. Acto, 17, 436 (1961).
52.7 52.7 49.0 57.9 62.6
% H
4.1 4.1 3.5 4.8 6.7
-Found%C
52.7 52.4 49.2 57.7 62.5
(EtOH), % H
mr
log
4.2 4.3 3.7 4.9 6.7
317 265 285 273 341 273
4.20 4.34 4.29 4.15 4.37 4.24
t
chromic shift parallels the hydrogen-bonding abilitylOvll in the order HzO > CH3OH > C2H60H > n-CdHsOH. The absorption maximum of the pnitro compound in formamide, which has approximately the same protondonating power for hydrogen-bond formation as ethanol12was found to be a t the same wave length as in the latter solvent. Thus it would appear that for the solvents studied here the shifts in the absorption maxima of the tosylhydrazone anions can also be viewed as a measure of their relative abilities to function as hydrogen-bond donors. In benzene and carbon tetrachloride which cannot act as proton donors but which a t the same time are nonionizing solvents we find that absorption takes place a t the shorter wave lengths. This could be due to the formation of ion pairs, in which the proximity of the cation prevents the distribution of charge throughout the anion and localizes it on the nitrogen. In the case of chloroform, which has some proton-donating ability for hydrogen-bond formation,13and at the same time is a poor ionizing solvent, both effects may be operative. Another possible explanation of the observed blue shift suggested itself, namely that this shift might simply be the consequence of the different polarities of the solvents in question. If one assumes that the dipole moment of the excited state is less than that of the ground state, a blue shift would be expected on changing from nonpolar to polar solvents. Kosower14 has been successful in relating many solvent-dependent shifts to the solvent polarity using his 2-values. A plot of our transition energies us. 2-values gives us a fair correlation for the hydrogen-bonding solvents only; however, in the aprotic solvents (DMF, CeHe, CH3COCH3, CH3CN, CsHaN) and also in chloroform this is not the case. It has also been found that with these compounds the total energy shift in changing to a nonhydrogen-bonding solvent from the best hydrogen-bonding solvent (HZO) is related to the electron-donating or electronwithdrawing ability of the substituent in the benzaldehyde moiety of the molecule. Thus it can be seen in Tables I1 and I11 that the relative spectral shifts are the largest for the pnitro compound. We believe that the reason for this is that while hydrogen bonding stabilizes the resonance form with the negative charge on the nitrogen adjacent to the tosyl group (OzNCeH4-CH=N-N--tosyl) , in the absence of hydrogen-bonding resonance forms involving the nitro (12) 9. Mizuahima, et ol., i b i d . , 7 , 100 (1955). (13) A. Allerhand and P. von R . Schleyer, J . A m . Chem. Soe., 86, 1715 (1963). (14) E. M . Kosower. ibid., 80, 3253, 3261 (1958).
1244
NOTES
VOL.30
TABLEI11 ABSORPTION MAXIMAOF TOSYLHYDRAZONE ANIONS RCH=N-N-tosyl Solvent
2,4-ClzCsHc 7
a
HCON(CHa)z CHaCN CzHSOH CHaOH HzO Estimated by extrapolation
368.5 (4.13) 355 (4.28) 334.5 ( 4 . 2 5 ) 329 ( 4 . 2 9 ) 313 (4.27) to zero time.
--
m-OoNCaHh X , m y (log fa)--------
351 (4.18) 342 (4.30) 321 (4.25) 315 (4.25) 297 (4.29)
group (e.g., -OzN=C6H4=CH-N=N-tosyl) contribute greatly to the excited state. The increased resonance stabilization in these latter forms is manifested by the large energy shift in changing from a protic to a nonprotic solvent. On the other hand, the effect of an electron-donating substituent such as the diethylamino group would be to favor resonance forms with high electron density a t the nitrogen atom (e.g., Et2NC6H&H=N-K--tosyl). The contribution of resonance forms in which the negative charge is delocalized (Et2N--C6H3=CH-N=N-tosyl) will be small even in nonhydrogen-bonding solvents, hence leading to only small transition energy differences between protic and nonprotic solvents. The total energy changes observed range from 16.6 kcal./mole for the pnitro compound to 3.6 kcal./mole for the pdiethylamino compound. These energies are for the most part much larger than those observed for the energy of the hydrogen bond, which is on the order of 2 to 10 kcal./ m01e.l~ Therefore, the total energy shifts should be related to the energy gained through stabilization of the ground state plus the energy gained through the destabilization of the excited state in the hydrogen-bonding solvents: l 6 The effect of solvents manifests itself also in the rate a t which the tosylhydrazones decompose. At room temperature the pnitrobenzaldehyde derivative anion decomposed in aprotic dissociating solvents approximately 15 times as fast as in the hydrogen-bonding solvents, confirming the spectroscopically observed stabilization of the ground state by the latter solvents. The half-life of a 10-4 M solution was about 2 min. in DMF us. about 30 min. in ethanol. In the cases of the 2,4-dichlorobenzaldehyde, m-nitrobenzaldehyde, and benzaldehyde derivatives, the rates of decomposition in nonprotic solvents were also faster than in protic solvents, The rates were, however, considerably slower than for the pnitro compound in the expected order : 2,4-dichloro- > m-nitro- > benzaldehyde. These decomposition reactions were followed spectroscopically by observing the rate of disappearance of the anion bands. The o-hydroxy and pdiethylamino tosylhydrazone anions appeared to be stable a t room temperature indefinitely. In the case of the pnitro compound, pnitrophenyldiazomethane could be isolated as a stable solid. (15) L. Pauling, "The Nature of the Chemical Bond," 3rd Ed., Cornell University Press, Ithaca, N . Y . , 1960, p. 449. (16) This situation appears to be similar to t h e n -r r* carbonyl transition discussed by Kosower.1' We wish to thank the referee for pointing this out to us.
-
R CsHa
-
o-HOCsH, --A,
345 (4.23) 334 (4.15) 313 ( 4 . 1 2 ) 310.5 (4.26) 303 (4.24)
p(CnHa)zNCsH4 my (log e)-
342(4.34) 337(4.37) 325 ( 4 . 2 3 ) 322(4.23) 322(4.19)
337 ( 4 . 4 0 ) 332 ( 4 . 4 2 ) 325(4.53) 325 (4.41) 325 (4.36)
Conversion of Methyl 17~-Acetoxy-5-oxo-3,5-seco-4-norest ran-3-oate to
Ag(")-Testosterone'
JULIUSA. VIDAAND MARCEL GUT Worcester Foundation for Experimental Biology, Shrausbury, Massachusetts Received October 19, 1964
A review of recent total syntheses shows that most of them, especially the most efficient ones, led to 19norsteroids. A recently published total synthesis2 commanded our attention because an intermediate seemed to be a good starting material to produce analogs possessing a 19-methyl group. We had a further interest in this problem because of our desire to introduce a radioactive label at C-19.a The key intermediate in our synthesis of adrenosterone, 17~-hydroxy-5-oxo-3,5-seco-4-norestr-9-en-3-oic acid (I),4 had already been obtained by total synthesis. As an alternative, a partial synthesis makes use of the known methyl 17p-acetoxy-5-oxo-3,5seco-4-norestran-3-oate~ which was brominated selectively a t C-10 with N-bromosuccinimide and the resulting crude bromide was dehydrobrominated with lithium chloride in N,N-dimethylformamide to give 178hydroxy-5-0~0-3~5-seco-4-norestr-9-en-3-oic acid (I). There remained the introduction of the angular methyl group a t (3-10,already mentioned without disclosure of details, in a review-article by Velluz, et a1.,2a in a closely related case. To the solution of I in N,Ndimethylformamide sodium hydride was added, followed by the addition of a catalytic amount of methanol and then by a large excess of methyl iodide. The absence of any a,/?-unsaturated ketone in the isolated crude product demonstrates that the methylation had taken place a t (3-10, thus making it unnecessary to block C-6.* Simultaneous methylation of the (1) Supported, in part, by National Institutes of Health Grant H-5266. Presented, in part, before the Division of Organic Chemistry a t the 148th National Meeting of the American Chemical Society, Chicago, Ill., Sept. 1964, Abstracts of Papers, p. 405. (2) (a) L. Vellus, G. Nomine, and J . Mathieu, Angew. Chem., 79, 725 (1960); (b) L. Velluz, G . Nomine, G. Amiard, V. Torelli. and J. Cerede, Compt. rend., 967, 3086 (1963); (0) British Patent 914,738 (Jan. 2 , 1963). (3) S. Rakhit and M. Gut, J . Am. Chem. Soc., 86, 1432 (1964). (4) L. J. Chinn and H. L. Dryden, Jr., J . Org. "hem., 96, 3904 (1961). (5) Free acid described by J. A. Hartman, A. J. Tomasewski, and A. S. Dreiding, J . Am. Chem. Soc., 78, 5662 (1956). (6) Cf. R . B . Woodward, F. Sondheimer. D . Taub, K. Heusler, and W. M. McLamore, ibid., 74, 4223 (1952); L. B. Barkley. W. S. Knowles, H. Raffelson, and Q . E. Thompson, ibid., 78, 4111 (1956).
NOTES
APRIL1965
U
- 1
OH
OAc
IV
C-3 carboxyl gave rise to the methyl ester. This keto ester IIa was hydrolyzed to its acid IIb and the latter was converted to the enol lactone 111. A Grignard reaction on the lactone, followed by hydrolysis and cyclization in the usual manner' produced (IV). Its conversion the desired A9(11)-testoster~ne to adrenosterone has already been described.2as8 Experimental9
17p-Hydrory-S-oxo-3,5-seco-4-norestr-9-en-3-oic Acid (I).Methyl 17~-acetoxy-5-oxo-3,5-seco-4-norestran-3-oate (7.250 g., 20.7 mmoles) was dissolved in a mixture of 350 ml. of carbon tetrachloride and 100 ml. of n-pentane. To this solution 3.91 g. (22 mmoles) of N-bromosuccinimide was added and the mixture was refluxed and illuminated by a 500-w. photospot lamp for 30 min. The mixture was cooled, the precipitated succinimide (1.942 9.) was filtered off, and the solution was evaporated a t room temperature to dryness under reduced pressure. The residue was dissolved in 70 ml. of N,N-dimethylformamide, 3.4 g. of lithium chloride was added, and the solution was kept a t 100' for 4 hr. The solution was diluted with ether, and ice was added. The aqueous phase was separated and the ether layer was washed with water, 2 N hydrochloric acid, water, saturated sodium bicarbonate solution, and water. The ether solution was dried and evaporated to afford an oil (6.2 g.). This was chromatographed on 300 g. of silica gel (Davison 923), whereby the eluates with benzene-ethyl acetate (9.5:0.5) gave 1.9 g. of methyl 178acetoxy-5-oxo-3,5-seco-4-norestran-3-oate (starting material). Further elution with the same solvent mixture gave 504 mg. of a compound which appeared to be a mixture. Elution with a 9 : l mixture of benzene-ethyl acetate afforded 3.46 g. (73.5% conversion, 65% yield) of methyl 17~-acetoxy-5-oxo-3,5-seco4-norestr-g-en-3-oate which appeared uniform on a thin layer chromatogram. The methyl ester was dissolved in 50 ml. of methyl alcohol, and 5 ml. of 40y0 sodium hydroxide solution was added. The solution was refluxed under nitrogen for 2 hr. and then concentrated under reduced pressure a t room temperature. Acidifi(7) R . B . Turner, J . A m . Chcm. Soc., 71, 579 (1950); G. I. Fujimoto, i b i d . , 78, 1856 (1951). (8) J. Fried and E. F. Sabo, ibid., 76, 2273 (1953); R. H. Lenhard and S. Bernstein. ibid., 77, 6665 (1955). (9) Melting points were determined on a Fisher-Johns apparatus and are corrected. The ultraviolet absorption spectra were determined in methanol on a Cary Model 14 recording spectrophotometer and the infrared spectra in a potassium bromide pellet (Infracord), The elemental analyses were carried out by Schwarakopf Microanalytical Laboratory, Woodside, 77, N . Y. The n.m.r. spectra were obtained in deuteriochloroform solution with a Varian Model V-4300B spectrometer. using tetramethylsilane as an internal standard.
1245
cation with 2 N hydrochloric acid and crystallization of the resulting precipitate from aqueous ethanol afforded 2.60 g. (58.5%) acid: m.p. of 17~-hydroxy-5-oxo-3,5-seco-4-norestr-9-en-3-oic 195-198'; kff" 248 mp (e 14,600); vmax 3472, 1745, 1712, 1626, and 1600 cm.-']. 17~-Hydroxy-S-oxo-3,5-seco-4-norandrost-9 (11)-en-3-oic Acid (IIb) .-To the solution of 1 g. of 17~-hydroxy-5-oxo-3,5-seco-4norestr-9-en-3-oic acid in 100 ml. of distilled N,N-dimethylformamide 1 g. of sodium hydride (51y0 suspension in mineral oil) was added. The reaction flask was heated under nitrogen a t 100' for 1 hr., then cooled to O', and 2 drops of methyl alcohol was added. After stirring the reaction mixture for 15 min., 5 ml. of methyl iodide was added and the stirring was continued at room temperature for 2 hr. Ether was added and then ice and saturated sodium chloride solution. The product was extracted with ether; the ether solution was dried and evaporated. The crude reaction mixture was chromatographed on 100 g. of silica gel (Davison 923), whereby the eluates with benzene-ethyl acetate (7:3) gave 940 mg. (85%) of compound I I a which appeared uniform on a thin layer chromatogram. A 6:4 mixture of benzene-ethyl acetate eluted an additional 95 mg. of a compound which was not further identified. The methyl ester 11 was dissolved in 25 ml. of methyl alcohol and 2 ml. of 15% sodium hydroxide solution was added. The solution was refluxed under nitrogen for 1 hr., 2 ml. of a 15% sodium hydroxide solution was added, and heating was continued for another hour. The solution was concentrated under reduced pressure a t room temperature to a small volume, water was added, and, after acidification to pH 4 with 2 N hydrochloric acid, the compound was extracted with ether. The ethereal solution wm dried and evaporated to dryness. The residual oil was crystallized from ether-hexane, and the analytical sample was prepared by crystallization from aqueous ethyl alcohol to give colorless [aIz1~ +48' ( c 1.4, methanol); plates: m.p. 167.5-168'; infrared absorption maxima, vmax 3440, 3200-3100, 2900-2800, 1703, and 1633 cm.-l; no ultraviolet absorption between 220280 mp; n.m.r. spectrum, T 3.47 (COOH), 4.38 (11-H), 6.12 (17a-H), 8.75 (19-CHa), 9.18 (18-CH3); T (in CsD6N) 4.32 (11-H), 6.05 (17a-H), 8.73 (Ig-CH,), 9.00 (19-CH3). Anal. Calcd. for C L ~ H Z ~ C, O ~70.56; : H , 8.55. Found: C, 70.25; H , 8.55. 5-Hydroxy-17~-acetoxy-3,5-seco-4-norandrosta-S,9 (11) -dien3-oic Acid 3,s-Lactone (III).-To the solution of 1.0 g. of 178hydroxy-5-oxo-3,5-seco-4-norestr-9(ll)-en-3-oic acid in 20 ml. of acetic anhydride was added 200 mg. of fused sodium acetate, and the mixture was heated a t reflux under nitrogen for 2 hr. Then the solvent was evaporated to dryness under reduced pressure and ice was added. After acidification with 2 N hydrochloric acid, the compound was extracted with ether, and the solution was dried over sodium sulfate and evaporated to dryness. The oily residue was chromatographed on 100 g. of silica gel (Davison, grade 9231, whereby a mixture of benzene-ethyl acetate 9 : l eluted 850 mg. (78%) of enol lactone acetate. Recrystallization from ethanol. gave colorless plates: m.p. 118-120'; [ a l Z 1 ~ -86" (e 1.7, methanol); infrared absorption m.txima, vmax 1750, 1727, 1683, 1643, 1262, 1248 cm.-1; no selective ultraviolet absorption above 220 mp; n.m.r. spectrum, I 4.46 and 4.59 (6-H and 11-H), 5.29, 7.29, 7.41, 7.95 (17a-H), 8.73 (19CHs), 9.21 (18-CHa). Anal. Calcd. for C~oHZeOc:C, 72.70; H , 7.93. Found: C, 72.86; H , 7.77. A9(l1)-Testosterone (IV) .-A 10% excess of a standardized methylmagnesium iodide solution was added dropwise over a period of 1 hr. to a stirred solution of 265 mg. of enol lactone acetate 111 in 50 ml. of benzeneether (1 : 1) which was kept a t ca. -15' with exclusion of air for an additional hour. Then cold dilute hydrochloric acid was added to the reaction mixture, and the compound was extracted with ether. The organic layer was washed with water, dried, and evaporated to dryness under reduced pressure. A solution of 660 mg. of potassium hydroxide in 1 ml. of water and 20 ml. of methyl alcohol was added to the residue, and the mixture was heated a t reflux under nitrogen for 2 hr. The solvent evaporated under reduced pressure a t room temperature to a small volume, saturated sodium bicarbonate solution was added, and the compound was extracted with ethyl acetate. The organic layer was washed with water and chromatographed on 50 g. of silica gel. A mixture of benzene-ethyl acetate ( 9 : l ) eluted 8 mg. of a compound showing an infrared absorption a t 1700 cm. -1. Further elution with benzene-ethyl acetate (8:2) gave 13 mg. of a mixture, exhibiting in the infrared
NOTES
1246
spectrum a saturated ketone and an enone absorption, and benzeneethyl acetate (7:3) eluted 114 mg. (50%) of AQ(l1)testosterone. Crystallization from aqueous ethyl alcohol and methylene chloride-hexane afforded colorless crystals, m.p. 153-154", identical in all respects with an authentic sample.lO (10) The authors are indebted to the Upjohn Co., Kalamazoo, Mich., for kindly supplying an authentic sample of AP(ll)-testosterone.
Studies Concerning the Infrared Spectra of Some Substituted Benzofuran Derivatives
VOL. 30
The C-H in-plane bending bands (1270-1015 cm.-l) of furans are much stronger and more easily distinguished than those for the more normal aromatic compounds presumably due to the polar nature of the heteroatom. The compounds studied were divided into two groups. One group (Table I) included only substances with two
TABLE I BENZOFURAN DERIVATIVES WITH Two FURAN HYDROGENS
W. W. EPSTEIN,W. J. HORTON, AND C. T. LIN Department of Chemistry, University of Utah, Salt Lake City, Utah 84112 Recezved September 11, 1964
I
r
RI
Although a number of studies of the infrared spectra of furan-type compounds have been reported, l-* these investigations with one exception' have been limited to monosubstituted or simple polysubstituted derivatives. Three groups of these workers have made some generalizations concerning characterizing bands. Katritzky and Lagowski2 found nine bands to be common and identified with the furan nucleus in 24 2-monosubstituted furans. Kubota3 after considering the spectra of 43 furano compounds has suggested three bands as being characteristic of the furan group. On the basis of 20 compounds most of which were 2,3-disubstituted furan derivatives of complex furanoquinoline alkaloids and furanocoumarins, Briggs and Colebrook' proposed that seven bands characterized the furano group. However, there has been little effort directed toward empirical correlations between given bands and specific furan hydrogens. As a result of synthetic work directed toward furanocoumarins and other benzofuran derivatives, we have been able to gather and study the infrared spectra of a number of simple and complex substituted benzofurans. The findings of this study seem to indicate a relationship between the presence or absence of furan hydrogens and medium to strong bands in the 1180-1020-~m.-~ region. It is not surprising that a relationship exists since Katritzky and Lagowskig have shown that in the six-membered ring heteroaroniatic series, and presumably in the five-membered series as well, the number and relative orientations of the hydrogen atoms determine the positions of the in- and out-of-plane bending modes. The strong absorption of the out-of-plane bending modes of aromatic hydrogen have long been used to establish substitution patterns while the inplane bending modes are not of much diagnostic use because they usually give only weak bands and there are other types of absorption in the same region. (1) L. H. Briggs and L. D. Colebrook, J . Chem. Soc., 2458 (1960).
(2) A. R. Katritrky and J. M. Lagowski, ibid., 657 (1959). (3) T . Kubota, Tetrahedron. 4, 68 (19.58). (4) A. H. J. Cross, S. G . E. Stevens, and T. H. E. Watts, J . A p p l . Chem. (London), 7 , 562 (1957). (5) L. \\-, Daasch, Chem. I n d . (London), 1113 (1958). (6) A. H. J. Gross and T . H. E. Watts, ibid., 1161 (1958). (7) A . Quilico, F. Piozri, and M . Pavan, Tetrahedron, 1, 177 (1957). ( 8 ) R. Royer, E. Bisagui. C. Hudry, A. Cheutin, and M. L. Desvaye, Bull. 8 0 c . chim. France, 1003 (1963). (9) A. R. Katritzky and J. M. Lagowski, J . Chem. Soc.. 4155 (1958).
-
R4 --
Compd.--
Rz
Rs
--Bands
R4
Ha
(cm. -I)-HO
H H H H 1029 (s) 1125 (8) H H OH COCH3a 1047 ( m ) 1138 (m) H H. OCHD Et" 1048 (ms) 1142 ( 8 ) H H OH COCH," 1073 ( m ) 1160,(ms H OCHa OCHa Hb 1045 (m) 1139 (s) H OH COCHa HC 1042 ( m ) 1139 (8) H OH COOH HC 1041 (8) 1140 (s) OH OCHa OCHa Hd 1087 (m) 1134 ( 8 ) OOCCHa OCHa OCHa 1070 ( m ) 1136 (8) Hd OCHa OH COCH, OCHs",' 1064 (8) 1143 (ma) P. K. Ramachandran, A. T . Tefteller, G. 0. Paulson, T. Cheng, C. T. Lin, and W. J. Horton, J . Org. Chem., 28, 398 (1963). Unpublished work by P. K. Ramachandran and W. J. Horton. P. K. Ramachandran, T. Cheng, and W. J. Horton, J . Org. Chem., 28, 2744 (1963). Unpublished work of C. T. I i n and W. J. Horton. e Unpublished work of E . Paul and W. J. Horton. f Spectrum run in carbon tetrachloride.
furan hydrogens. This group had two bands, one in the range 1160-1125 cm.-' and one in the 1087-1029cm.-l region which we suggest are associated with the in-plane bending modes of the 0-and a-hydrogens on the furan nucleus and tend to be characteristic of furans with adjacent a- and P-hydrogens. The second group (Table 11) has compounds with only a single furan hydrogen and in the 0-position. These compounds all showed a band in the range 1172-1121 cm.-l and no medium or strong, sharp bands in the 1087-1029-cm.-' region which we feel is characteristic of the 0-hydrogen. It is not known if this generalization can be applied to simple furan derivatives but Quilico, Piozzi, and Pavan7 report that dendrolosin and tetrahydrodendrolosin, both monosubstituted furan derivatives, do have medium to strong bands at 1156 and 1075 cm.-l. Since the region of the spectrum under consideration is one where many types of absorptions occur, the generalization discussed above should be used cautiously and perhaps in a negative sense rather than a positive one, ie., the absence of a 1160-1125- or 1087-1020cm.-l band indicating the lack of an a- or 0-H on a benzofuran nucleus, rather than the occurrence of the proper absorption indicating the definite presence of the a- or 0-H. Experimental All the infrared spectra were taken on a Beckman I R 5 with sodium chloride optics. Chloroform was used as solvent unless
1247
NOTES
APRIL1965
TABLEII.-BENZOFURANDERIVATIVES WITH ONEFURAN HYDROQEN
RZ
-
Compd.
R4
Ri
Rs
Band, cm. -1, HD
H" 1135 (s) H COCHa 1143 (ms) H H COOCHa 1152 (ms) Hb OH COOCHa 1152 (ms) Hb OCHa COOCHI He 1157 (s) NHz COOCHa H" 1152 (s) NHCOCHa COOCHa He 1126 (m) COCHa COOCHa He 1163 (s) C( CHa)=NOH COOCHa HC 1122 (w) c1 COOCHa HC 1142 (ms) CzHs COOH HC 1152 (ms) COCHa CH(CHa)z He 1152 (ms) Cyclic ketal of COCHa COH(CH3)z GH6b 1152 (m) or 1131 ( m ) OH COOCHa COCH? 1129 ( m ) OH COOCHa H" 1142 (s) CONH C=CH&Ha H" 1136 (9) COCHa COOCHI Hc,d 1156 ( m ) or 1129 (m) Cyclic ketal of COCHa COOCHa Hd.e 1165 (m) or 1135 (m) H COOCHa OCH? 1143 (m) H COOCHa OCH? 1172 (ms) H COOCHa COCH?.' H 1135 (9) COOH HC 1136 (ms) CzHs COOCHs HE 1125 (m) COOCH, NOz OCHad3' H 1149 (ms) COOCHa NHzg H 1157 (s) COOCHa NHCOCHS' H 1149 (s) COOCHa COCHa' H 1125 (s) COOCH, C( CHa)=NOH' H 1121 ( 9 ) COOCH3 Hh OCHa 1130 (s) COCHa Hh OCHs 1134 (s) COCHa Hh OCHa 1130 (s) COCHs OH" COCHa 1135 (s) COOCHa OHo COCHa 1149 (s) COOCHa COCHa 00C CHa' 1138 (s) COOCHa OCH; COCHa 1152 (s) COOH Br OCH; 1136 (s) COOH OCHaE 1121 (s) CHa COOH OCHad" 1150 (s) COOCHa CHa OOCCHa COCHah 1127 (m) COCHa OOCCHa COCH; 1130 ( m ) COCHzCOCHa Footnote c, Table I. J. I. Degraw and W. A. Bonner, Tetrahedron, 18, 1295 (1962). * Footnote a, Table I. Spectrum run in Spectrum run as Nujol mull. Footnote b, Table I. Ir Footnote d, Table I. carbon tetrachloride. e Footnote e, Table I. otherwise indicated. The preparation of the compounds listed in Tables I and I1 aa previously unreported will be the subject of a later communication.
5H-Pyrimido[4,5-b][l,4]thiazin-6(7H)-one. The Product of a Novel Cyclization Reaction' JAMES R. PIPERA N D THOMAS P. JOHNSTON Kettering-Meyer Laboratory, Southern Research Institute, Birmingham, Alabama 55905 Received November 24, 1964
Heteroaromatic-substituted alkanethiols, such as 2-benzimidazolemethanethiol (I), became of interest in a search for an effective heterocyclic modification of
2-aminoethanethiol as an anbiradiation drug.2 The reported3 synthesis of I by the acid-catalyzed condensation of o-phenylenediamine and mercaptoacetic acid provided a good method for the synthesis of 2-benzimidazoleethanethiol (11) by a similar condensation in which 3-mercaptopropionic acid was used. The insolubility of 4,5-diaminopyrimidine (111) in ethyl mercaptoacetate, however, precluded the condensation expected to give the analogous purine-g-methanethiol (IV) under conditions suggested by Albert's synthesis of purine-8-methanol from I11 and ethyl g l y ~ o l a t e . ~ (1) This investigation was supported by the U. S. Army Medical Research and Development Command under Contract No. DA-49-193-MD-2028. (2) (a) T. P. Johnston and A. Gallagher. J. Ore. Chem., 17, 2452 (1962): (b) J. R. Piper and T. P. Johnston, ibid.,1 8 , 981 (1963): (c) T. P. Johnston and A. Gallagher, ibid.,1 8 , 1305 (1963). (3) G . K. Hughes and F. Lions, J . Proc. Roy. SOC.N . 8. Wales, 71, 209 (1938): Chem. Abstr., 83,5830 (1938). (4) A. Albert, J . Chem. Soc., 2690 (1955).
NOTES
1248
H I, n - 1
VOL. 30
concentrated to 200 ml. Slow cooling produced 0.78 g. of colorless crystals, and a second crop of 0.23 g. was obtained from the filtrate concentrated to about 50 mi., yield 33%. For analysis, the combined crops were twice recrystallized from methanol, the final sample being dried in vacuo over phosphorus pentoxide at 78": m.p. 295-300" dec. (capillary in aluminum block, from in m r (e X lo+), 217 (11.4),242 (11.1),302 (5.0), 260"); A, and 333 (sh) at pH 1, 213 (11.8), 242 (13.6), and 300 (6.3) at pH 7, 254 (13.8) and 297 (8.3) at pH 13; v::: 3000-2500 (acidic NH) and 1680 cm.? (amide CO, strong). The compound was homogeneous on a thin layer chromatogram [silica gel H (Merck), 9: 1 CHC13-CH30H] viewed in ultraviolet light after spraying with aqueous Ultraphor solution. Anal. Calcd. for CBHSNSOS:C, 43.10; H , 3.01; N, 25.13; S, 19.18. Found: C, 42.90; H , 3.16; N, 25.12; S, 19.28.
H IV
11, 1L = 2
When mercaptoacetic acid was used as reactant and solvent instead of the ester, t,he isolated product was shown to be not a purine but 5H-pyriniido[4,5-b] [1,4]thiazin-6(7H)-one (VII). Assignment of structure was made on the basis of analysis, strong amide carbonyl absorption in the infrared, and ultraviolet absorption dissimilar to that of purine-8-methanol. The formaAcknowledgment.-The authors are indebted to Dr. tion of VI1 can be rationalized as a cyclization of .the W. J. Barrett and his associates in the Analytical initially formed amide V by addition of -SH to Chemistry Section of Southern Research Institute for -C(NH2)=Ywith loss of ammonia from the unthe spectra and microanalyses reported. stable intermediate VI. Cyclization resulted in effect from a novel displacement of the C-4 amino group by a mercaptoalkyl group, the more nucleophilic C-5 amino group being favored for amide formation and disThe Addition of t-Butyl Chloride to Butadiene favored for displacement. The 2-amino-4-methyl derivative of this ring system had previously been obGUGLIELMO KABASAND RUDOLF GABLER tained by acid-catalyzed cyclization of (2,5-diamino-6Dewey and Almy A. G . , Zurich, Switzerland methyl-4-pyrimidiny1thio)acetic acid.5 7
?-
I
0
I
Received October 83, 1964
The cationic telomerization of butadiene with ali~ has phatic halides is recorded in the l i t e r a t ~ r e l -and been studied extensively by Petrov and co-workers. 4-7 This reaction may be represented by the following generalized equations. With R = (CH3)3C-the lower R-C1
+ n(CHQ=CH-CH=CH*) R-( CH&H=CH-CH?),Cl
VI Experimental
2-Benzimidazoleethanethiol (11).-A solution of freshly recrystallized o-phenylenediamine (2.00 g., 18.5 mmoles) and 3mercaptopropionic acid (2.00 g., 19.2 mmoles) in 4 N hydrochloric acid (20 ml.) was refluxed for 1 hr., allowed to cool somewhat, and treated with decolorizing carbon. The colorless filtrate, diluted with methanol (10 ml.) and chilled to about 5", was neutralized by the dropwise addition of cold concentrated ammonium hydroxide (8 ml.) . The resulting white precipitate was collected and dried in vacuo over concentrated sulfuric acid and sodium hydroxide pellets. The crude product (1.88 g.) was extracted with cold 5y0 sodium hydroxide solution (38 ml.) with minimal exposure to air; addition of glacial acetic acid to the filtrate in the cold precipitated pure I1 as white crystals, which were collected under nitrogen and dried zn vacuo over phosphorus pentoxide at 78": yield 1.58 g. (48y0), m.p. 168169.5", v",: 2560 cm.-' (SH). Anal. Calcd. for CsHlON2S: C, 60.65; H , 5.66; N, 15.72; S, 17.99. Found: C, 60.54; H , 5.74; N, 15.41; S, 18.17. 5H-Pyrimido[4,5-b][ 1,4]thiazin-6(7H)-one (VII).-A solution of 4,5-diaminopyrimidine (2.00 g., 18.2 mmoles) in freshly distilled mercaptoacetic acid (20 ml.) was heated at 148-152' for 2 hr. The resulting red sirup was thoroughly triturated in two 20-ml. portions of benzene, which was removed by decantation. A solution of the residue in 1,2-dimethoxyethane (5 ml.), when stirred, deposited a yellow-tan powder, which was collected by filtration of the cooled mixture with the aid of a little cold methanol. A solution of the crude product in boiling methanol (300 ml.) was treated with decolorizing carbon (1.0 g.), filtered, and (8) F . L. Rose,
J. Chem. Soc., 3452 (1952)
telomers (n = 1-3) were isolated and identified as straight-chain l14-addition compounds. 4,5 We were particularly interested in getting high yields of pure l-chloro-5,5-dimethyl-2-hexene, but none of the reported procedures appeared to be satisfactory. With the object of avoiding the formation of higher telomers, the influence of all reaction parameters, ie., catalyst, cocatalyst, temperature, time, and reactants ratio, on the conversion of starting materials and on the yields of chlorooctene and chlorododecadiene was investigated. As may be seen from the results summarized in Table I, the telomerization of t-butyl chloride with butadiene was catalyzed by a number of FriedelCrafts-type halides. Only mercuric chloride, antimony pentachloride, and antimony trichloride failed to show any efficiency. The highest proportion of l-chlor0-5~5-dimethy12-hexene was obtained when zinc chloride and bismuth chloride were used as catalysts. Furthermore, it was found in two more experimental series not reported here in detail, that in order to obtain a good conversion, an amount of 0.9% (based on total weight of (1) I. G.Farben Ind., French Patent 824,909 (1938). (2) W. H. Peterson and K. D. Detling, U. S.Patent 2,419,500(1947). (3) W. D . Niederhauser. U. S. Patent 2,689,873(1954). (4) A. A. Petrov and K. V. Leets, Dokl. Akad. Nauk S S S R , 96, 285 (1954). (5) A. A. Petrov and K. V . Leets, Zh. Obahch. Khim., 26, 1113 (1956). (6) A. A. Petrov and T. V. Yakovleva, Isv. Akad. Nauk S S S R Ser. Fiz., as, 1217 (1959). (7) A . A . Petrov and Z. N. Kolyaskina, Zh. Obahch. Khim., SO, 1450 (1960)
TABLE I TELOMERIZATION OF &BUTYL CHLORIDE WITH BUTADIENE' Cat.
--Yield n - 1
of products, % ' -n = 2 n > 2
Reactant8 conversion, %
HClO4 .. .. 100 21 1 38 61 30 Tic14 9 6 85 60 A1C13 13 6 81 59 SnC14 58 19 23 37 ZnClz 60 22 18 64 BiCll a In these experimente, 0.45 mole of t-butyl chloride and 0.3 mole of butadiene in the presence of 0.5 g. of the halide catalyst were shaken for 5 days in pressure bottles a t 20-25".
reactants) zinc chloride or bismuth chloride and a reaction time of 120-130 hr. were necessary. A reaction time of more than 130 hr. at 20-25", apart from being impractical, did not raise the conversion. It has already been stated in the literature,*g5 that concentrated hydrochloric acid suppresses the formation of high telomers. This is shown for the two most effective catalysts (ZnCl:! and BiC13) in Table 11. TABLE I1 INFLUENCE OF CONCENTRATED AQUEOUS HYDROCHLORIC ACIDON THE TELOMERIZATION O F &BUTYL CHLORIDE AND BY ZINC CHLORIDE AND BUTADIENE CATALYZED BISMUTHT R ~ C H L O R I D E ~ Cat.
Concd. HC1, ml.
--Yield n = 1
of products, yo---n = 2 n > 1 n > 2
Reactants conversion, %
ZnClz ... 58 ,.. 42 ... 37 ZnClz 0 . 1 86 ... 14 ... 11 ZnClz 0 . 2 85 ... 15 ... 8 ZnCl2 0 . 3 86 ... 14 ... 2 ... ... ... ... .. ZnCL 0 . 4 BiC13 ... 60 22 ... 18 64 BiCla 0 . 0 5 61 18 ... 21 64 BiCl, 0.1 62 19 .,. 19 60 BiC13 0 . 2 82 14 ... 4 46 BiC13 0 . 3 83 13 ... 4 38 2.9 51 BiC13 1.2* 8 0 . 5 12.9 . . , a In these experiments, 0.45 mole of t-butyl chloride and 0.3 mole of butadiene were shaken in the presence of 0.5 g. of the See large-scale preparahalide catalyst at 20-25" for 120 hr. tion in the Experimental part.
In both reactions, l-chloro-5,5-dimethyI-2-hexene can be prepared in 80-85% yield. However, in the zinc chloride catalyzed reaction, this is only possible a t a conversion of 11%, compared with a 40-50% conversion in the case of BiC13. It is evident from the results of Table I1 that, for the preparation of the lower telomers, BiC13 is the catalyst of choice. The influence of different t-butyl chloride-butadiene molar ratios on the telomer distribution in telomerizations catalyzed by BiC13 and aqueous HC1 (in amounts of 0.9 and 0.2%, TABLE I11 CONDENSATION O F &BUTYL CHLORIDE A N D BUTADIENE Molar ratioa
1249
NOTES
APRIL1965
-Yield
of product?, %-
n = l n = 2 n > 2 1:2 34 23 43 39 19 42 1:1.5 59 21 20 1:l 81 15 4 1 : 0 63b 2 1:0.4 93 5 Molar ratios of t-butyl chloride to butadiene.
Reactants conversion, Yo
89 84 78 63 40 See ref. 4.
respectively, based on the total weight of reactants) is shown in Table 111. From these results it may be seen that favorable conditions for the preparation of l-chlor0-5,5-dimethyl2-hexene exist when the molar ratio of t-butyl chloride to butadiene lies in the range of 1.6 to 2.2. The last reaction parameter investigated was the effect of temperature on the t-butyl chloride-butadiene conversions and telomer distribution. It was found that at 40" the results shown in Table I1 could be obtained after only 15-17 hr. The telomer distribution remained almost unaltered, as can be seen from an optimized up-scaled run described in the Experimental part. The chlorooctene and chlorododecadiene obtained in this telonierization process are all-trans compounds as evidenced by their infrared spectra. Experimental
A . Small-Scale Runs.-A 250-ml. glass pressure bottle was cooled to - lo", filled with the required amounts of reactants and catalysts, and shaken at room temperature for a certain number of hours. Excess butadiene was vented off and the brown liquid was decanted into water; the mixture was well shaken, the organic layer was washed until neutral with water, separated, dried with calcium chloride, and distilled. l-Chloro-5,5-dimethyl-2-hexene distilled at 65-68' (25 mm.), n% 1.4447 (lit.4 b.p. 46-47' at 10 mm., 72% 1.4456). The remaining higher boiling material consisted of chlorododecadiene distilling at 55-57' (0.1 mm.), ?&'OD 1.4658 (lit.6 68-73' at 2.5 mm., n% 1.4643) and higher telomers. B. Large-Scale Preparation.-A mixture of 440 g. (4.75 moles) of redistilled t-butyl chloride, 162 g. (3 moles) of butadiene, 5.4 g. of bismuth chloride, and 1.2 ml. of roncentrated aqueous HC1 was charged in a glass autoclave. The reaction mixture was heated to 40" and stirred at this temperatiire for 15 hr. At this point the manometer showed that the pressure in the reaction vessel had fallen to 1 atm. or less. The crude reaction mixture was neutralized with diethylamine and distilled over a 50-cm. Vigreux column. After a forerun of unreacted t-butyl chloride and dissolved butadiene, l-chloro-5,5-dimethyl2-hexene distilled at 65-68' (25 mm.) to give 306 g. (80.5Oj, yield at 51% conversion) of a pure product; chlorododecadiene distilled at 55-57' (0.1 mm.) to give 49 g. (12.9T0 yield), followed by a small amount of higher boiling material (b.p. 60-120" at 0.1 mm.), 11 g., corresponding to a 2.9% yield.
Pyridines from Cyanogen-Like Compounds and Unsymmetrical Dienes' GEORQE. J. JANZ AND ALANR. M O N A H A N ~ Rensselaer Polytechnic Institute, Department of Chemistry, Troy, New York Received October 19, 1,964
The synthesis of pyridines at 350-450' by the thermally initiated reaction of cyanogen-like compounds with lJ3-butadiene was reported in the preceding paper.3 With unsymmetrical dienes, a route to a series of new pyridines is apparent; the relative amounts of the isomeric pyridines from each unsymmetrical (1) Abstracted in part from a thesis submitted by A . R . Monahan t o Rensselaer Polytechnic Institute in partial fulfillment of the requirements for the P h . D . degree, Aug. 1964. (2) Union Carbide Fellow in Chemistry, 1963-1964. (3) G . J. Janz and A . R. Monahan. J . Org. Chem., 19, 569 (1964).
1250
NOTES
VOL. 30
TABLE I SYNTHESIS O F %CYANO- AND 2-PERFLUOROALKYLPICOLINES RCN (moles)
Diene (moles)
Contact time, sec.
Total time, min.
Temp.,
floc.
Material recovery, wt. yo
Conversions per single pass, mole % Diene Nitrile
CF3CN (0.062) Isoprene (0.058) 32.9 30 418.5 100 34.5 12.6 CF3CN (0,200) Pentadiene (0,385) 2770 300 400.0 99.0 62.0 60.0 CF~CFZCFZCN (0.101) Isoprene (0.056) 49.9 60 420.5 100 46.8 11.4 (CN)2(0.132) Isoprene (0.042) 36.6 48 412.9 93 1 83.3 12.9 a Yield of pyridinic products based on converted diene. Yield of pyridinic products based on converted nitrile.
TABLE I1 H1 K.M.R.OF METHYL-GROUP PROTONS ON SOMESUBSTITUTED PICOLINES Pyridines
4Methyl-2-heptafluoroprop ylpyridine 5-Methyl-2-heptafluoroprop ylpyridine 6-Methyl-2-cyano pyridine 4-Methyl-2-cyanopyridine 5-Methyl-2-cyano pyridine 3-Methyl-2-trifluoromethylpyridine 4-Methyl-2-trifluoromethylpyridine 5-Methyl-2-trifluoromethylpyridine 6-hlethyl-2-trifluoromethylpyridine a Isomeric mixture.
6 (p.p.m. 1 0 . 0 1 , downfield from TMS external standard)
2.43" 2.37" 3.19 2.86" 2.96" 2.51 2.18" 2.12" 2.42
diene, furthermore, offers the prospect of a more penetrating insight in the reaction mechanism. The present communication reports on a series of experiments with isoprene and pentadiene in reaction with (CK)Z,CF3CX, and CF3(CFZ),CN. Experimental Cyanogen (99.5% minimum purity, b.p. -21.7", American Cyanamid Co.) was vacuum degassed (-195') in the conventional manner three times prior to use. The cyanogen-like compounds, CF3CS and CF3(CF2)2CN (95% minimum purity, b.p. -64'; and 95y0 minimum purity, b.p. -5', respectively; Peninsular Chemresearch, Inc.), were similarly vacuum degassed. The unsymmetrical dienes, CH2=C( CH3)-CH=CH2 and CH3CH=CH-CH=CHz (99% minimum purity, b.p. 34'; and 90% minimum cis isomer, b.p. 40", respectively; Phillips Petroleum Co.), were distilled just prior to use. The purity of all samples was checked by gas chromatographic analysis. The continuous flow reactor, recovery of materials, and monitoring of gaseous and liquid reactants have been described in the preceding paper3 and were used without change. A summary of the experimental conditions, recoveries, and yields is in Table I. Quantitative separation of the pyridines was achieved by gas chromatography (Beckmann GC-2; 12-ft. di-n-decyl phthalate on firebrick column; 13@190°), and structural confirmation wm by vibrational spectroscopy, mass spectra, and H' magnetic resonance data. The latter results are in Table 11. The physical properties and microelemental analyses found for the new pyridines are given below. 3-Methyl-2-trifluoromethylpyridine had b.p. 174O, d 1.2402 g. ml.-*, andnzOo1.466. Anal. Calcd.: N , 8.70. Found: N, 8.51. 4-Methyl-2-trifluoromethylpyridinehad b.p. 170°, d 1.2199, and n Z o D 1.4298. Anal. Calcd.: N, 8.70. Found: N, 8.18. 5-Methyl-2-trifluoromethylpyridine had b.p. 171 d 1.2202, and n z 0 1.4298. ~ Anal. Calcd.: N, 8.70. Found: N , 8.48. 6-Methyl-2-trifluoromethylpyridine had b.p. 154", d 1.23259 and n z 01.4278. ~ Anal. Calcd.: N , 8.70. Found: N, 8.90. O,
Yields of pyridinic product, mole % ' a b
37.7 50.2 44.2 11.1
100 100 100 22
4-Methyl-2-heptafluoropropylpyridine had b.p. 176", d 1.3789, n% 1.40%. Anal. Calcd.: N,5.36. Found: N, 5.46. 5-Methyl-2-heptafluoropropylpyridine had b.p. 181', d 1.3792, and n 2 6 ~1.3934. Anal. Calcd.: N , 5.36. Found: N, 5.90. Isomer Ratios.-The relative amounts of the two isomeric pyridines in the products obtained by the techniques of gas chromatography and confirmed by H1 magnetic resonance spectroscopy were as follows (for the experiments summarized in Table I). RCN
CFaCN CF3(CF2)zCN (CN)z
Diene
Isoprene Isoprene Isoprene
4-CHs: 5-CHa isomer ratio
2.90: 1 . 0 2.52: 1 . 0 1.34:1 . 0
With pentadiene and CF3CN,similarly, the ratio of the isomers (6-CH3:3-CH3) in the pyridinic product was thus found to be 4.92: 1.O.
Discussion Inspection shows clearly that there is a correlation with the change in the electrophilic properties of the carbon in the C=N group; as the electrophilicity of this center increases, there is an increase in the 4-CH3 isomer in the pyridinic mixtures [e.g., with CFSCN, with (CN),, 74% 4-methyl-2-trifluorornethylpyridine; 57% 4-methyl-2-cyanopyridine]. That this is more probably due to the inductive effect rather than a volume factor4receives support from the near constancy of the isomer ratios when CF&N and CF,(CFz)ZCN, respectively, cyclize with isoprene. The isomer ratios further are found temperature invariant (400-420') and independent of reaction time (30-300 min.), Le., that these are thermally isomerized equilibrium ratios appears improbable. The markedly smaller amounts of the 3-CH3 isomer in the pentadiene-CF3CN reaction undoubtedly relates to the influence of the terminal CH, of the diene; this may be attributed, in part, to a steric factor with cis-pentadiene. Further discussion of the secondary forces must await more detailed studies of the reaction mechanism. The present results underline the need to consider the role of polar structures in any complete account of this high temture gas phase Diels-Alder process. Acknowledgment.-This work has been made possible, in large part, by grants-in-aid from the American Chemical Society, Petroleum Research Fund, and the Union Carbide Chemicals Company. (4) I. N . Nazarov, Yu. A. Titov, and A . I. Kuznetsora, Akod. Nauk. S S S R , Otd. Khim. Nauk., Engl. Tranal., 1412 (1959).
NOTES
APRIL1965
1251 Experimental
The Monohalogenation of 8-Quinolinols RAMESHWAR PRASAD, HOBERTL. D. COFFER, AND HENRY FREISER QUINTUSFERNANDO, Department of Chemistry, University of Arizona, Tucson, Arizona Received October 28, 1964
I n the course of a study of the kinetics of bromination of the chromium(II1) chelate of 8-quinolinol by a coulometric method, it was observed that the metal chelate brominated much faster than the free ligand under identical conditions. The same result was obtained with the cobalt(III), aluminum(III), and copper(I1) chelates of 8-quinolinol. Although the reaction rates of the metal chelates were too fast to measure by our method,' the stoichiometry of the reaction was readily obtained. One mole'of bromine reacted rapidly with every mole of 8-quinolinol present in the metal chelate. Further reaction of the metal chelate with bromine did occur, but a t a much slower rate. This behavior of the chelated 8-quinolinol is in direct contrast with that of the free ligand, since each mole of 8-quinolinol is known to react quite rapidly with 2 moles of bromine to form the 5,7-dibromo-8-quinolinol. Attempts to monohalogenate 8-quinolinol directly result in low yields and extensive contamination by the dihalo product. On a preparative scale, halogenation of the metal chelate, rather than of the ligand itself, with 1 mole of halogen/mole of the 8-quinolinol in the metal chelate and subsequent rupture of the metal chelate with acid should give the monohalogenated derivative in good yields and free of dihalogenated products. N-halosuccinimides were used as halogenating agents since they did not release hydrogen ions that can cause rupture of the metal chelate during the course of the reaction. A further means of avoiding chelate dissociation is the use of an organic solvent as reaction medium. If these conditions are met, then it would not be necessary to use inert chelates such as those of Co(II1) and Cr(II1) which may by their resistance to dissociation present problems in metal ion removal. The monochloro, monobromo, and monoiodo derivatives of 8-quinolinol, 2-methyl-8-quinolino1, and 4methyl-8-quitlolinol were synthesized by halogenating the copper(I1) chelate with N-halosucciniinde (1 mole/mole of the 8-quinolinol in the chelate). Halogenation of the lithium salt under identical conditions always gave the 5,7-dihalo derivative. An investigation of the effect of the metal ion on the yield of monobromo-8-quinolinol obtained by this method showed that the maximuin yield was obtained with the cobalt(I1) chelate (69%) followed by the aluminum(II1) chelate (66%) , the cobalt(II1) chelate (520j0), and the copper(I1) chelate (50%). The similarity of these results indicate that specific metal ion effects are minimal. It was deduced that halogenation occurred in 5-position in all cases by a comparison of the infrared and proton magnetic resonance spectra of the halogenated compounds that were synthesized with available authentic samples of 5-halo-8-quinolinols. (1) G. S Kozak. Q . Fernando, and H. Freiser. Anal. Chem.. 86, 296 (1964).
5-Bromod-Quinolinol .-The copper( 11)chelate of 8-quinolinolz (0.43 g.) was dissolved in a minimum volume of chloroform and a chloroform solution of N-bromosuccinimide (0.43 9.) was added. After the reaction mixture was allowed to stand for several hours a precipitate of the copper(I1) chelate of 5-bromo8-quinolinol was formed and separated by filtration. About 80% of the copper(I1) chelate was recovered in this step. The chelate was dissolved in 50 ml. of 6 M HCI and the copper(I1) wm precipitated by the addition of excess thioacetamide. The solution containing the copper(I1) sulfide was filtered and the filtrate waa neutralized with NaHC08 solution. The 5-bromo-8-quinolinol which precipitated was washed with water, recrystallized from ethanol, and finally vacuum sublimed: yield -0.31 g., m.p. 124" (lit.2 m.D. 124"). An alternative method that was used for the separation of copper(I1) from the 5-bromo-8-quinolinol was aa follows. The copper(I1) chelate of 5-bromo-8-quinolinol(O.85 9 . ) was dissolved in 1 : 1 HCl and a solution containing 1.13 g. of EDTA was added. When the pH of the solution was raised t o 4 with ammonium acetate, a yellowish precipitate was formed. The mixture was extracted with 100 ml. of chloroform, the solid obtained bv evaporation of the chloroform layer was redissolved in 1 : 1 HCl, and a solution containing 1.13 g. of EDTA was added. On raising the pH of the solution to 5 with ammonium acetate, a pale yellow precipitate was obtained. The mixture was extracted with 75 ml. of chloroform and on evaporation 0.63 g. of 5-bromo-8quinolinol waa obtained. The copper(I1) chelates of 8-quinolinol, %methyl- and 4methyl-8-quinolino18 were halogenated in a similar manner to give the corresponding 5-halo derivatives.
Acknowledgment.-The authors are grateful to the National Institutes of Health for financial assistance. (2) R. G. W. Hollingshead, "Oxine and its Derivatives," Vol. I , 11. and 111,Butterworth and Co., Ltd., London, 1956. (3) J. P. Phillips, R. L. Elbinger, and L. Merritt, J A m . Chem. Soc., T i , 3986 (1949).
The Structure of the Product from Reaction of the Dimedon Formaldehyde Derivative with Base and Iodine. An Example of Coupling through Five Bonds FREDH. GREENBERG Laboratory of Metabolism, National Heart Institute, Bethesda, Maryland Received July 15, 1964
In 1925 and 1927 Radulescu and Georgescu1t2proposed the cyclopropane structure I for the product, m.p. 211-212', obtained by treating iodine with the disodium salt of the diniedon derivative of formaldehyde. Repeating the above authors' procedure we obtained the compound described earlier. However the spectral properties are inconsistent with I and lead to assignment of the isomeric structure 11.
I
I1
( 1 ) D. Radulescu and V. Georgescu, Bull. aoc. chim. France, [4] S T , 187 (1925). (2) D. Radulescu and V. Georgescu, Bul. soc. stiinte CZuj, 3, 129 (1927);
Chem. Abstr., 81,3203 (1927).
NOTES
1252
t
8.0
7.0
6.0
5.0
4.0
3.0
2.0
1.0
0
1.0
0
P.p.m., 6.
Figure 1.-N.m.r. spectrum of 11.
8.0
7.0
6.0
5.0
4.0
3.0
2.0
VOL.30
1,3-~yclohexanedione (1722 and 1752 cm. -l). Model compounds for comparison are 3-ethoxy-5,s-dimethylcyclohexenone6 (CHCl,, 1642 cm. -I) ; 2,2,5,5-tet,ramethylcyclohexane-l,3-dione6(CHCl,, 1688 and 1717 cm. -l) ; 1,3,3-trimethylbicyclo [2.2.2loctane-2,6-dione' (IV, CS,, 1712 and 1749 cm.-'); and bicyclo[2.2.2]octane-2,6,7-trienes (V, KBr, 1725 and 1770 cm.-'). Surprisingly, the n.m.r. spectrum of I1 displays two , methylene groups a t 6 3.17 and 2.48 as symmetrical triplets (J = 2.0 =t0.1 c.P.s.) rather than sharp peaks. The origin of the long-range coupling is demonstrated by the spectrum of I1 prepared from deuterioformaldehyde. This spect,rum is altered from the original (Figure 1) by the disappearance of the 3.17 peaks and the conversion of the 2.48 triplet to a singlet. Therefore, coupling exists between the formaldehyde methylene and one of the methylenes of the cyclohexenone ring. Presumably the methylene involved in this coupling is adjacent to the ring-juncture double bond since protons in the corresponding positions of benzofuran have been foundQt'o be coupled. Radulescu and Georgescu2 subjected I1 to basic cleavage and obtained a compound, m.p. 102-103O, which they formulated as VI. We find that treatment of I1 with sodium hydroxide in aqueous dioxane, followed by acidificat'ion, yields an acid with nearly the same melting point, whose spectral properties are those of VII.
P.p.m., 6.
Figure 2.-N.ia.r. spectrum of VII.
@
COCH&(CH3)zCH&OzH
The n.m.r. spectrum of I would be expected to display four peaks in the area ratio 6 :6 :8 :2. Instead the observed spectrum (Figure 1) in deuteriochloroform exhibits three methyl peaks and four methylene peaks in the ratio 6 :3 :3 :4 :2 :2 :2. This is clearly compatible with I1 which has nonequivalent methyls attached to the cyclohexanedione ring, equivalent methyls attached to the cyclohexenone ring, and four kinds of methylene groups. Further evidence for I1 is gained from the ultraviolet spectrum 267 mp, E 11,800) which is characteristic e.g., 3-ethoxy-2,5,5of 3-alkoxy-2-alkylcyclohexenones, trimethylcycl~hexenone~ 268 mp, E 19,100) and the dimedon derivative obtained from both chloroacetaldehyde and hydr~xyacetaldehyde~ (111,*: : :A 268 mp, E 24,000). The ultraviolet spectrumof I would be expected to resemble that of 2,2,5,5-tetramethylcyclohexane-1,3-dione5 A('",: 274 mp, E 240).
0
VI
&~OCHzC(CH*(g-IDIH VI1
Confirmation of I1 is given by the infrared spectrum in chloroform which exhibits carbonyl bands due to a 3-alkoxycyclohexenone (1644 cm. -l) and a nonenolic
The infrared spectrum discloses carbonyl absorptions due to a 3-alkoxycyclohexenone (1637 cm.-'), a carboxyl group (1725 cm.-'), and an aliphatic ket'one " : :A( 269 (1708 cm.-I). The ultraviolet spectrum mp, E 13,900) is similar to that of 11. The n.m.r. spectrum in deut'eriochloroform (Figure 2) reveals two overlapping met'hyl peaks; four methylene peaks belonging to the side chain and cyclohexenone ring; two poorly resolved multiplets due to the dihydrofuran methylene which is coupled with the adjacent proton and a methylene of the cyclohexenone ring; and a quartet centered a t 5.14 (J = 10 c.P.s.) due to the remaining dihydrofuran proton coupled with the adjacent methylene. The low field a t which this single proton resonance occurs is consistent with a hydrogen that is bonded to a carbon attached to oxygen. The carboxyl proton occurs a t 6 10.08. The formation of I1 appears to be the first example of the formation of a dihydrofuran by O-alkylat,ion when the possibility exists for a cyclopropane via C-alkyla-
(3) E. G. Meek, J. H. Turnbull. and W. Wilaon, J . Chem. Soc., 511 (1953). (4) The preparation of I11 from both aldehydes was originally reported by D. Varlander, Z . Anal. Chem., 7 7 , 241 (1929). (5) E. G. Meek, J. H. Turnbull, a n d W. Wilson. J . Chem. Soc., 2591 (1953).
(6) K. Nakanishi, "Infrared Absorption Spectroscopy," Holden-Day, Inc.. 7an Franciaoo. Calif., and Nankoda Co., Ltd., Tokyo, J a p a n , 1962, p. 69. (7) D. Y . Curtin and R. R. Fraser, J . A m . Chem. SOL, 81, 662 (1959). (5) W. Theilacker and E. Wegner, A n n . , 664, 125 (1963). (9) J. A. EIvidge and R. G . Foster. J . Chem. Soc.. 590 (1960).
I11
IV
V
NOTES
APRIL1965 tion. The dominant factor leading to the production of I1 rather than I may be the greater stability of dimedon as the enol rather than the keto form.l0 Thus I1 with the preferred cyclohexenone ring is the observed product. Experimental l1
The Disodium Salt of the Dimedon Formaldehyde Derivative. -This and subsequent procedures were essentially those of Radulescu and Georgescu.192 A mixture of sodium ethoxide ( 1 .OO g of sodium, 0.0435 g.-atom) and the dimedon formaldehyde12 derivative (5.80 g., 0.0199 mole) in 100 ml. of absolute ethanol waa stirred a t room temperature for 1.5 hr. affording a finely divided white crystalline precipitate. After filtration and washing with ether and ethanol, the solid waa dried under vacuum for 12 hr. and weighed 6.35 g. (957,). The n.m.r. spectrum in D,O showed three peaks a t 6 0.90, 2.05, and 2.97 in the ratio 12:8:2. Spiro [4-keto-6,6-dimethyltetrahydrobenzofuran-Z(5,7,H)-l'4',4'-dimethylcyclohexane-2,6-dione] (11).-To a suspension of the disodium salt of the dimedon-formaldehyde adduct (4.81 g., 0.0143 mole) in 80 ml. of anhydrous ether was added dropwise with stirring a t room temperature during 2 hr. a solution of iodine (3.64 g., 0.0143 mole) in 80 ml. of ether. The deep brown mixture was stirred for 0.5 hr. and filtered; the precipitate was washed with sodium thiosulfate solution and water and dried. The brown solid was then dissolved in 40 ml. of chloroform and successively washed with aqueous sodium thiosulfate and water. After drying (hlgS0,) the solvent mas evaporated in vacuo leaving a yellow-white solid: 3.72 g. (90%); m.p. 207-209'; CHCl8 vmsi 1644, 1722, and 1752 em.-'; v::: 1633, 1713, and 1748 em. - l . Recrystallization from ethanol afforded 2.91 g. (70'%), m.p. 211-212", "X:," 267 m r ( C 11800). The infrared spectra of the crude and recrystallized solids were identical. Anal. Calcd. for C17H&4: C, 70.32; H , 7.64. Found: C, 70.47; H , 7.62. The preparation of 11-Q, m.p. 210-211", was carried out aa above employing formaldehyde-& which was obtained by pyrolysis of Ep-polyoxymethylene-dz (Tracerlab, Waltham, Mass. ). Basic Cleavage of 11.-To a solution of I1 (0.321 g., 1.12 mmoles) in 15 ml. of dioxane was added dropwise with stirring a t room temperature 35.0 ml. of 0.0314 N sodium hydroxide (1.10 mmoles). After stirring 4 hr., the mixture waa then acidified with 0.55 ml. of 2 A' hydrochloric acid and extracted with methylene chloride; the extracts were washed with water and dried (Na&O4); and the solvent was removed a t reduced pressure leaving a yellowish white oil (0.316 g.). Crystallization waa effected by dissolution in ethyl ether followed by addition of petroleuni ether. Several recrystallizations from ethyl ether gave a solid with melting point 98-100"; ",",A:: 269 m r ( c 13,900); 1726 and 1637, 1708 and 1725 cm.-l. 1611 cm.-l; ~2::~' Dimedon Derivative of Chloroaceta1dehyde.-A mixture of dimedon (3.02 g., 0.0215 mole), chloroacetaldehyde diethylacetal (1.62 g., 0.0106 mole), 15 ml. of ethanol, and 25 of ml. water was heated a t reflux for 5 hr. After standing overnight the solution deposited a white-yellow solid which \+asfiltered and washed with water. The dried material weighed 0.915 g., m.p. 211-221'. The filtrate deposited an additional 0.277 g. Crystallization of the combined solids (ethanol-acetone) yielded 0.490 g. (870): m.p. 237.5");: : A 268 m p ( C :$:OOO)la; m.p. 221-224" ( k 4 CHCla Amax 1613, 1703 (weak), and 1735 (weak) em.-'; vmsl 1592, 1629, and 1637 em.-'. The n.m.r. spectrum (dimethyl-& sulfoxide) revealed two methyl peaks a t 6 0.98 and 1.07, a dimedon methylene peak a t 2.22, and cyclohexenone methylene peaks a t 2.29 and 2.05 with the latter a doublet (J = 2 c.p.6.) presumably due to coupling with the lone proton of the dihydrofuran ring;
VI::
(10) Dimedon in aqueous solution has been found t o be 95% enolic by G. Schwarzenbach and E. Felder, Helu. C h i n . Acta, 47, 1044 (1944). (11) Melting points were obtained on a Kofler hot stage and are corrected. Carbon-hydrogen analyses were performed by Micro-Tech Laboratories. Skokie, Ill. Infrared spectra were determined with a Beckman IR-7 spectrophotometer. ultraviolet spectra with a Cary Model 11 spectrophotometer, and n.m.r. spectra with a Varian A-60 spectrophotometer. (12) Prepared by the method of R. L. Shriner, R . C. Fuson, and D. Y. Curtin, "The Syatematic Identification of Organic Compounds," 4th Ed., John Wiley and Sons, Inc., Kew York. N. Y., 1959, p. 220. (13) For 2,5,5-trimethylcyclohexane-l,3-dione, 264 m p ( e 13,600); ref. 5 .
1253
the dihydrofuran methylene is a complex multiplet centered a t 4.50 and the single dihydrofuran proton is a multiplet centered a t 4.20. A peak due to the hydroxyl proton of the dimedon moiety waa not seen.I4 Anal. Calcd. for ClsH2dOa: C, 71.02; H, 7.95. Found: C, 71.34; H , 8.06.
Acknowledgment.-The author wishes to thank Miss Susan Sklar who initiated this study and Drs. H. M. Fales, R. J. Highet, and H. A. Lloyd for many helpful suggestions. (14) T h e n.m.1. spectrum of formaldehyde bismethone does not display a n enolic hydrogen peak: R. F. Brown, et al., J . Org. Chem., 98, 146 (1964).
A Convenient Synthesis of Some Haloferrocenes R. W. FISHAND M. ROSENBLUM Department of Chemistry, Brandeis University, Waltham, Massachusetts Received October 30, 1964
Haloferrocenes constitute intermediates of considerable value for the preparation of hydroxy-,la-e amino-,la*c azido-,* thio-,3 and cyanoferrocenes, l o as well as biferrocenyl~4,~ and ferrocenyl Grignard reagent^.^ Of the halo derivatives, bromo- and iodoferrocenes were first prepared by Nesmeyanov, Perevalova, and Nesmeyanova in 1955 by the action of the free halogens on mono- and 1,l'-bischloromercuriferrocene.6 Although relatively simple and direct, the procedure has not found wide application, since, with the possible exception of iodoferrocene, the yields of haloferrocenes obtained in these reactions were poor. In the intervening years, the chloro- and bromoferrocenes have become considerably more accessible through the reaction of cupric halides with ferrocenylboronic and ferrocenyl-1 ,1'-diboronic acids, a procedure first introduced and largely exploited by the Russian group. la,? Our interest in the halofcrrocenes, in connection with studies related to the possible existence of ferrocynes, prompted us to re-examine the use of the readily accessible chloroinercuriferrocenes as potential precursors of these substances. The purpose of the present report is to set forth a general procedure which accomplishes this aim, and which constitutes an attractive alternative method for the preparation of the mono- and disubstituted bromo- and iodoferrocenes. (1) (a) A. N. Nesmeyanov, V. A. Sazonova, and V. N. Drozd, Ber., 98, 2717 (1960); (b) A. N. Nesmeyanov and V. A. Sazonova, Dokl. Akad. Nauk SSSR, 148, 1060 (1959); (c) ibid., 180, 1030 (1980); (d) ibid., 188, 126 (1960); (e) A. Nesmeyanov, V. A. Sazonova, and V. N. Drozd, I P U .Akad. Nauk S S S R , O t d . K h i m . Nauk, 45 (1962). (2) A. N . Nesmeyanov, V. N. Drozd. and V. A . Sazonova, Dokl. A k a d . Nauk S S S R , 160, 321 (1963). (3) M. D. Rausch, J . Org. Chem., 26, 3579 (1961). (4) E. G. Perevalova and 0. A. Nesmeyanova. Dokl. Akad. Nauk S S S R , 189, 1093 (1960); M. D . Rausch, J . A m . Chem. Soc., 89, 2080 (1960); M. D . Rausch, J . Org. Chem., 46, 1802 (1962); 9. J. Goldberg and R. L. Matteson. ihid., 49, 323 (1964). ( 5 ) H . Shechter a n d J. F. Hellina, ibid., 46, 1034 (1961). (6) A. N. Nesmeyanov, E . G. Perevalova, and 0. A. Nesmeyanova, Dokl. Akad. Nauk S S S R , 100, 1099 (1955). (7) A. N. Nesmeyanov. V. A. Sazonova, and V. N. Drozd, ibtd., 146, 1004 (1Y59); 181, 1088 (1960).
1254
NOTES
We have found the chloromercuriferroceness are readily transformed to haloferrocenes by treatment with positive halogen reagents in polar media. Thus, bromoferrocene is formed in 57y0 yield when chloromercuriferrocene is treated with N-bromosuccinimide in dimethylforniamide solution, and 1 , l '-bischloromercuriferrocene is converted to 1 , l '-dibromoferrocene in 47% yield under similar reaction conditions. The absence of other ferrocene derivatives in the crude reaction mixture makes isolation of the haloferrocenes especially convenient. Although comparable yields of bromoferrocene may be obtained from the reaction of chloromercuriferrocene with N-bromosuccinimide in methylene chloride, with pyridinium bromide perbromide, or with N-bromoacetamide in dimethylformamide, the conconiitant formation of ferrocene as well as bromomercuriferrocene and diferrocenylmercury makes these procedures somewhat less advantageous. Iodoferrocene may similarly be prepared from chloromercuriferrocene, in 85% yield, by treatment with Niodosuccinimide in methylene chloride solution, while the more insoluble 1 , l '-bischloromercuriferrocene is converted in dimethylformaniide solution to 1 , l '-diiodoferrocene in 42y0 yield. These procedures appear to be confined in application to the preparation of bromo- and iodoferrocenes, since several attempts to employ them for the synthesis of chloroferrocenes were unsuccessful. In one experiment, the direct conversion of ferrocene to 1 , l '-dibromoferrocene, without isolation of the intermediate chloromercuri derivative, was shown to be feasible, but the practicability of this procedures was not further investigated. Experimental
Ha1omercuriferrocenes.-A solution of 78.7 g. (0.25 mole) of mercuric acetate in 750 ml. of absolute methanol was added dropwise to a stirred solution of 93 g. (0.50 mole) of ferrocene in 500 ml. of dry benzene. The reaction was continued in a nitrogen atmosphere a t room temperature for 10 hr., and then 22 g. (0.52 mole) of lithium chloride in 200 ml. of a 1: 1 ethanol-water mixture was added dropwise. The resulting orange suspension was stirred a t room temperature for 2 hr., then heated a t reflux for 1 hr., and was finally collected, placed in a Soxhlet, and extracted with methylene chloride. The residue remaining was recrystallized from dimethylformamide to give 22.7 g. of 1,l'-bischloromercuriferrocene (180jo)as a fine yellow powder, m.p. >300°. The methylene chloride extract was washed thoroughly with water and dried over magnesium sulfate. After removal of solvent, the solid residue was sublimed in vacuo to remove unchanged ferrocene. In this manner, 56.9 g. of ferrocene were recovered. The unsublimed portion gave, on recrystallization from methylene chloride-petroleum ether, 59.8 g. (73Yo) of chloromercuriferrocene as golden platelets, m.p. 196-198" dec. (lit.8 m.p. 194196"). Bromoferrocem-A solution of N-bromosuccinimide (1.15 g., 6.4 mmoles) in 100 ml. of dry, nitrogen-flushed dimethylformamide was added dropwise to a cold, stirred solution of chloromercuriferrocene (2.10 g., 5 mmoles) in 50 ml. of the same solvent. Reaction was continued a t 0' in a nitrogen atmosphere for 3 hr., after which time 200 ml. of a 10% sodium thiosulfate solution was added, and the resulting dark solution was poured into 2 1. of cold water. The aqueous solution was extracted four times with 100-ml. portions of petroleum ether, and the combined organic extract was washed with water and dried over magnesium sulfate. Evaporation of solvent left an orange oil
( 8 ) A . N . Nesmeyanov, E. G . Perevalova, R . V. Golovnya, and 0. A. Nesmeyanova, Dokl. A k a d . Nauk S S S R , 9'7, 459 (1954); M. D. Rausch. M. Vogel. and H. Rosenberg. J . Org. Chem., aa, 960 (1957).
VOL. 30
which was taken up in Skellysolve B and chromatographed on an alumina column (Fisher, activity 1j . Evaporation of solvent from the single yellow band, which developed on elution with Skellysolve B, gave 0.81 g. of an orange oil, which solidified on cooling. The material was recrystallized from cold ethanol to give 0.75 g. (577,) of bromoferrocene as yellow plates, m.p. 31-32' (lit.4m.p.3&31"). Treatment of chloromercuriferrocene with an equivalent of Nbromosuccinimide in methylene chloride solution gave bromoferrocene in 50% yield together with ferrocene (5%), diferrocenylmercury ( 18Y0j, and bromomercuriferrocene (4y0), m.p. 201.5203'. This latter substance was identical by mixture melting point determination with the product obtained on treatment of acetoxymercuriferrocene with potassium bromide. Anal. Calcd. for CloHgBrFeHg:C, 25.79; H , 1.95. Found: C, 25.97; H , 1.87. With chloromercuriferrocene and pyridinium bromide perbromide, in dimethylformamide solution under similar reaction conditions, ferrocene and bromoferrocene were formed in 42 and rjS% yield, respectively. N-bromoacetamide in dimethylformamide solution gave 48y0 of bromoferrocene and 5% of ferrocene. 1 , l '-Dibromoferrocene.-Following the experimental procedure given for the preparation of bromoferrocene, bischloromercuriferrocene, on treatment with N-bromosuccinimide in dimethylformamide, gave 1,l'-dibromoferrocene, m.p. 51.552" ( k 3 m . p .50-51°j, in47%yield. 1odoferrocene.-A solution of N-iodosuccinimide (1.2 g., 5.1 mmoles) in 100 ml. of dry methylene chloride, previously flushed with nitrogen, was added dropwise to a stirred suspension of chloromercuriferrocene (2.10 g., 5 mmoles) in 75 ml. of methylene chloride cooled in an ice bath, and maintained in a nitrogen atmosphere. After allowing the reaction to continue for 12 hr., 50 ml. of an aqueous 10% sodium bisulfite solution was added, followed by an equal volume of a 10% sodium carbonate solution. The organic layer was separated and the aqueous solution was extracted thrice with 50-ml. portions of methylene chloride. The combined organic extract was washed successively with 10% sodium carbonate solution and water, and dried over magnesium sulfate. Removal of solvent left an oil which was taken up in a small volume of Skellysolve B and chromatographed on alumina. On elution with this solvent two bands developed. The first gave 1.32 g. (85%) of iodoferrocene, m.p. 45.546' (lit.4m.p. 44-45'). The second band, eluted with methylene chloride-Skellysolve B, gave 0.34 g. of golden brown needles, m.p. 174-175.5", after one recrystallization from this solvent pair. This compound was identical by mixture melting point with iodomercuriferrocene, prepared by treatment of acetoxymercuriferrocene with potassium iodide. Anal. Calcd. for CloHsFeHgI: C, 23.43; H, 1.77. Found: C, 23.61; H , 1.97. 1,1 '-Diiodoferrocene.-This substance was isolated as a redbrown oil, in 42% yield, on treatment of l,l'-bischloromercuriferrocene with N-iodosuccinimide in dimethylformamide solution following the general procedure employed for the preparation of iodoferrocene. Direct Conversion of Ferrocene to 1,1 '-Dibromoferrocene .-A solution of 14 g. of mercuric acetate (0.044 mole) in 150 ml. Of methanol was added to 3.72 g. of ferrocene (0.02 mole) dissolved in 100 ml. of dry benzene. The reaction was allowed to continue a t room temperature under nitrogen for a period of 12 hr. A t the end of this period 6.75 g. of sodium iodide (0.045 mole) in 100 ml. of methanol was added and the solution was stirred a t room temperature for an additional 3 hr., and then finally heated a t reflux for 1 hr. The mixture was cooled to O', 8.95 g. of Nbromosuccinimide (0.05 mole) in 150 ml. of dry methylene chloride was added slowly, and the resulting green solution was stirred a t 0" for 6 hr. Aqueous sodium bisulfite solution ( l o % , 200 ml.) was added, the organic layer was separated, and the aqueous phase was extracted several times with methylene chloride. The combined organic extract was washed twice with 100ml. portions of 10% sodium carbonate solution, then with water to neutrality, and finally dried over magnesium sulfate. Evaporation of the solvent left a brown oil, which on vapor phase frartometric analysis, employing a 20% QF column on Chromosorb W , indicated that it was a mixture of ferrocene and 1,l'-dibromoferrocene in a ratio of 1:9.6. N o monobromoferrocene was evident in the mixture.
NOTES
APRIL1965 The Structure of Ethyl Dehydroacetate E. EARLROYALS~ AND JOHNC. LEFFINGWELL~ Department of Chemistry, Emoky Unizersity, Atlanta 22, Georgia Received September 1 , 1964
During the course of an attempted diene synthesis employing cyclopentadiene and the enol acet,ate of ethyl acetoacetate (I), a solid, m.p. l l O o , identified as dehydroacetic acid (11), was isolated; no Diels-Alder adduct was formed. Under the vigorous experimental conditions, the enol acetate apparently underwent selfcondensation to yield ethyl acetate and dehydroacetic acid (11). 0
0
II
I
0
I1
Dehydroacetic acid has been assigned the Feist structure 113on the basis of the observed isomerization to 2,6-dimethylpyrone-4-carboxylic acid-3. A reexamination of structure I1 by Berson5has confirmed the structure of the carbon skeleton as formulated. The striking similarity of the ultraviolet absorption spectra of the acid I1 and its ethyl ester (enol ether) provided compelling evidence that dehydroacetic acid (11) was completely enolized in solution. Furthermore, OR 0
IIa
0
0
IIb
0
OR
1255
merits.* However, the case of dehydroacetic acid (and its esters) is complicated by the presence of the 3acetyl group. With these considerations in mind, a systematic study of the ultraviolet, infrared, and n.ni.r. spectra of dehydroacetic acid and its methyl and ethyl esters was undertaken, employing chemical methods to supplement the spectroscopic information. Homogeneous esters were obtained in the usual manner by the action of alkyl iodides on the dry silver salt of dehydroacetic acid.5 Initial comparison of the spectral data gave little more inforrnation than had been obtained by Berson5 and by Forsen and Kilsson6 as to the enol structure of the acid and its esters. Subjecting the acid to the action of deuterium oxide in the presence of sodium deuterioxide caused rapid deuteration of the 3-acetyl group (observed via n.m.r. by the disappearance of the singlet at T 7.43). This observation was not sufficient to rule out structure IIc inasmuch as the acid may be subject to tautomeric eyuilibria. The ethyl ester, in which the enol form is fixed, also exhibited rapid deuterium exchange on the 3acetyl group (disappearance of the singlet a t T 7.55). Hydrogenation of the ethyl ester, m.p. 9&9l0, in ethanol using 5% palladium on charcoal as catalyst afforded an oil, Amax 264.5 mp, which gave a strong positive ferric chloride test. This oil was readily identified as 3-acetyl-6-methyl-2-pyrone (IIIa) by its n.m.r. and ultraviolet spectra. The absence of the characteristic O-C& group in the n.m.r. spectra indicated that the enol ether had cleaved during hydrogenation. The 6-methyl group was now present as a doublet (3H, center a t T 8.63), while the &acetyl group showed as a singlet (3H, T 8.07) which readily underwent deuterium exchange. A single proton showed as a multiplet (center a t T 5.87) corresponding to the 6-position hydrogen. The remaining hydrogens showed as unresolved multiplets. The absence of
IIC
on the basis of the absorption data, Berson concluded that the conversion of the acid to its ethyl ester occurred with a minimal structural change, ie., with simple replacement of the hydrogen by an ethyl group. However,'no conclusion was drawn as to which of the three possible enol variants (IIa, IIb, or IIc) the acid (and the ethyl ester) possessed. Recently, Forsen and Nilsson'j have concluded on the basis of the infrared6*'and n.m.r. spectra that dehydroacet,ic acid and its methyl ester posses the a-pyrone structure IIa, although structure IIc was not ruled out. Investigations of similar substituted 2,4-pyrandione systems have shown that, depending on the method of preparation, the enol ethers which are formed are not necessarily homogeneous. Comparison of the ultraviolet and infrared spectral propert'ies of the enol ethers often allow a simple method of positional assign(1) To whom inquiries should be made a t : Newport Division. Heyden Newport Chemical Corp., Pensacola, Fla. (2) National Defense Education Act Fellow, 1960-1963. (3) F. Feist, A n n . , 86'7, 253 (1890). (4) C. F . Rassweiler and R.Adams, J . A m . Chem. S o c . , 46, 2758 (1924). (5) J. A . Berson, ibid.. 74, 5172 (1952). (6) S. Forsen and M. Nilsson, Arkiu K e m i , 17, 523 (1961). (7) H. M. Randall, R . G. Fowler, N . Fuson, and J. R . Dangl, "Infrared Determinations of Organic Structures," Van Nostrand Co., Inc., New York. N. Y . , 1949, p. 231.
IIIa
IIIb
IIIC
two ext'ra protons in the region of the n.m.r. spectra occupied by the single proton on the >CH-0- carbon ruled out IIIb, while the deuteration experiments (as well as the absence of a -CzH5 group in the n.m.r.)lO precluded IIIc. This combination of spectral data constitut'es a simple proof that the enol structure of ethyl dehydroacetate is IIa, and, if the direct relationship of the enol ether to the acid is valid, of the enol structure of dehydroacetic acid. (8) (a) S. Janiszewska-Drabarek, Roczniki Chem., 87, 456 (1953); Chem. Abstr., 49,3176 (1955); (b) I. Chmielewska, J. Cieslak, and T . Kraczkieaicz, Rocrniki Chem.. 80, 1009 (1956); Chem. Abslr., 61, 8733 (1957); ( e ) D. Herbst, W. B. Mers, 0 . R . Gottlieb. and C . Djerassi. J . A m . Chem. S O C . , 81, 2427 (1959); (d) W. Jachvmczyk and I . Chmielewska, Bull. Acad. Polon. Sei., Ser. sci. chim., 8 , 155 (1960); Chem. Abstr., 66, 11401 (1961): (e) J . D. Bu'Lock and H. G. Smith, J . Chem. Soc., 502 (1960); ( f ) E. .4. Chandross and P. Yates. Chem. I n d . (London), 149 (1960). (9) J. C. Leffingwell, Ph.D. Dissertation, Emory University. Aug. 1963. (10) The -&Ha group, for structure 1110, might have been expected to show up in the -CHz- region not as a quartet but as a multiplet due to the transient present of the acidic proton on the ring: this situation would have been resolved during the deuteration experiments. Such a situation, however, did not arise. The deuteration experiments were controlled so that no appreciable cleavage of the @-keto esters occurred in any of the ca8es reported herein.
1256
NOTES Experimental
Ultraviolet spectra were obtained on a Beckman Model DB spectrophotometer in 9570 ethanol (mean deviation f l .O mp). Infrared spectra were taken on a Perkin-Elmer Model 21 spectrophotometer equipped with sodium chloride optics. The n.m.r. spectra were obtained on a Varian Model A-60 spectrometer using tetramethylsilane as an internal reference arid are reported in T (p.p.m.) units. Melting points were taken in capillaries on a Mel-Temp melting point apparatus. Dehydroacetic Acid (II).-The enol acetate of ethyl aceto~ 61.2 g.], freshly distilled acetate [b.p. 83" ( 5 mm.), n Z 51.4423, cyclopentadiene (19.8 g.), and hydroquinone (0.2 g.) were mixed in a glass liner and placed in a high pressure bomb at 190" for 20.5 hr. under 2000 p.s.i. of an inert gas (hydrogen). On removal from the bomb, 4 g. of polymer was filtered off and the liquid was distilled, yielding three fractions. The low-boiling fraction consisted of mainly cyclopentadiene and was stripped off and discarded. The higher boiling material was fractionated to give 35 g. of unreacted enol acetate and 6.3 g. of a white crystalline material, b.p. 126" (7 mm.), m.p. 99-loo", which melted at 110" following recrystallization from carbon tetrachloride. The solid was identified as dehydroacetic acid, giving no melting point depression on admixture with an authentic sample, m.p. 110", prepared by the method of Arndt." The acid gave the following 225.5 mp (log e 3.99) and spectral properties: ultraviolet, A,, 5.70 (s), 5.75 ( m ) , 6.06 (s), 311.0 mp (log e 4.05); infrared,: : :A 6.16 (m), and 6.40 (s) p ; X:".:"' 5.82 (s), 6.06 (s), 6.16 (s), and 5.78 (s), 6.05 (s),6.17(m), and 6.40 (s) p ; n.m.r. 6.42 (8) p ; ::A: (CDCl,), 7 7.84 (3H, doublet, J = 0.8 c.P.s.), 7.43 (3H, singlet), and 4.22 ( l H , quartet). A sample of dehydroacetic acid was dissolved in DZ0 containing sodium deuterioxide and examination of the n.m.r. spectrum showed rapid deuteration of the 3-acetyl group as indicated by the disappearance of the singlet at 7 7.43. No exchange with the 6-methyl was observed. Sodium Dehydroacetate .12-Sodium dehydroacetate was obtained as a white salt from ethanol-water: A,, 230 mp (log e 4.22) and 295.5 mp (log e 3.92); ::A: 5.93 (s), 6.02 (s), 6.25 (sh), 6.34 (s), and 6.51 (ms) p ; Ai:?' 5.93 (s), 6.02 (s), 6.31 (m), 6.34 (s), and 6.51 (ms) g . Silver Dehydroacetate .12-Silver dehydroacetate was obtained as white crystals from hot water which turned gray on drying: Amax 229 mp (log B 4.21) and 297 mp (log e 3.92); Ai:: 5.93 (s), 6.03 (s), 6.35 (s), and 6.52 ( m ) p ; Af;:' 5.93 (s), 6.08 (s), and 6.41 (8) p . Methyl Dehydroacetate.lZ-Methyl dehydroacetate was obtained as faintly yellow needles on recrystallization from carbon 224.5 mp tetrachloride: m.p. 91-92' (lit.12 m.p. 93-95'); A, (s), 5.83 (8, sh), (log e 3.98) and 314.5 mp (log €3.94); A:%5'7.8 5.89 (s, sh), 6.08 (m), and 6.56 ( m ) p ; Ai:;' 5.78 (ms, sh), 5.86 (s), 5.98 (s), 6.09 (m), and 6.66 (s) p ; A:".: 5.76 (ms, sh), 5.83 (s), 5.95 (s), 6.06 (m), and 6.64 (s) p ; n.m.r. (CDCl,), 7 7.68 (3H, doublet, J = 0.8 c.P.s.), 7.55 (3H, singlet,) 6.04 (3H, singlet), and 3.77 ( l H , quartet). Ethyl Dehydroacetate.12-Ethyl dehydroacetate was obtained as white needles on recrystallization from carbon tetrachloride: 225.5 mp (log e 3.96) and m.p. 90-91" (lit.12 m.p. 93-94'); , , ,A 313.5 mp (log e 3.91); 5.79 (a), 5.87 (8, sh), 6.07 (m), and 6.53 (ms) p ; AI::"' 5.83 (8, sh), 5.89 (s), 6.09 (m), and 6.54 (8) p ; KRr Amax 5.69 ( m ) , 5.83 (s), 5.89 (s, sh), 6.07 (m), and 6.54 (8) p ; n.m.r. (CDCl,), T 8.57 (3H, center of triplet), 7.40 (3H, doublet, J = 0.8 c.P.s.), 7.55 (3H, singlet), 5.78 (2H, center of quartet), and 3.93 ( l H , quartet). A small sample of ethyl dehydroacetate was shaken in DzO in the presence of sodium deuterioxide for 20 min. and extracted with chloroform-d. The diminished singlet at T 7.55 in the n.m.r. indicated partial deuteration of the 3-acetyl methyl group while no exchange of the 6-methyl was observed. Ethyl dehydroacetate (2.5 9.) was hydrogenated on a Parr hydrogenation apparatus (60 p.s.i.) using 57, Pd-C (0.5 9.) as catalyst in 35 ml. of absolute ethanol. The catalyst was removed by filtration and the ethanol solution exhibited an ultraviolet A,;, at 264.5 m p . The solvent was removed under vacuum to give 1.5 g. of a dark oil which gave a strong, positive ferric chloride test in both water and alcohol. This material was assigned the structure of 3-acetyl-6-methyl-2-pprone on the basis of the
Az:
( 1 1 ) F. Arndt, O r g . Syn.,20, 26 (1940). ( 1 2 ) .J. Y Collie and H. R. LeSuerir, J . Chem. Soc.. 66, 254 (1894)
VOL. 30
spectral evidence: A,:t 5.81 (w-m), 6.04 (s), 6.19 (s), and 8.00 (8) p ; A?: 5.75 ( 8 1 , 6.04 (s), 6.19 (s), and 8.00 (s) p ; n.m.r. (CCla), T 8.63 (3H, center of doublet, J = 7 c.P.s.), 8.07 (3H, singlet), 5.87 ( l H , multiplet), and 7.84-8.50 (5H, unresolved multiplets). This oil rapidly exchanged deuterium with the 3acetyl hydrogens as was observed by the disappearance of the singlet a t T 8.07. The deuterium exchange reactions were carried out in DzO at ca. 10-15" with only a trace of sodium deuterioxide present so as to minimize any chance of @-ketoester cleavage. A quantitative hydrogenation of a 1.O-g. sample using 570 Pd-C in absolute ethanol gave absorption of 3.2 mole equiv. of hydrogen. The ethanol solution was examined by ultraviolet and showed an absence of absorptions at 225 and 313 but gave an absorption at 264.5 mlr indicating hydrogenation was complete. Hydrogena~ which tion using PtOz in ethanol led to a yellow oil, n Z 41.4924, had A, 294 mp and: : :A 5.74 (m, sh,) 5.85 (s), 6.05 (ms), 6.39 (s), 6.90 (ms), 7.21 (ms), 7.41 (m), 7.52 (wm), and 7.81 (8) p . This material, which was obviously different from the product using Pd-C as catalyst, was not identified.
Acknowledgment.-The authors wish to thank Mr. Ralph Mayo for the n.m.r. spectra in this work, and Drs. M. R. Willcott, 111, and C. E. Boozer for their valuable comments.
The Peroxide-Induced Decarbonylation of Phenoxyacetaldehyde JAMESW. WILTA N D MARTINP. STUMPF' Department of Chemistry, Lo yola University, Chicago, Illinois 60686 Received October 19, 1964
Our interest in the neophyl-radical rearrangement prompted an investigation of a possibly analogous oxygen to carbon migration by phenyl. This report concerns our efforts to convert the phenoxymethyl radical (1) into the benayloxy radical (2). Several ?
Ph-O-CHs.
+ .O-CHz-Ph 2
1
items in the literature made such an investigation seem worthwhile. (1) The rearrangement appeared possible even though no steric compression exists in 1, since Slaug'ls found some rearrangement of phenyl in CSH5CH2CH2.(using labeled compounds) and s5owed thereby that compressional strain in the radical is not necessary for rearrangement to occur. ( 2 ) Oxygen radicals of the type sought do not usually undergo the reverse of the above rearrangement unless three aryl groups are present. Thus, 3 does not rearrange to 4,4 while 5 does rearrange to 6 (the Wieland resrrangePh-CH( CH,)-O. 3 Ph3C-0. 5
CHaCH-OPh 4
--+ PhiC-OPh 6
(1) From the M.S. Thesis of M. P. S.. Loyola University, June 1962. (2) J . W. Wilt and C. A . Schneider. J . O r g . Chem., 96, 4196 (1961),
and earlier papers. (3) 1,. H. Slaueh. J . A m . C h e m Soc., 81, 2262 (1959). (4) N . Kornblum and H. E. De La Mare, ibid., 74, 3079 (1952). This work is incorrectly ascribed t o N . Kornblum and C. Teitelbaum in both Gould's text5 and in C. Walling, "Free Radicals in Solution," John Wiley and Sons. I n c . , New York, N. Y . , 1957, p . 473.
NOTES
APRIL 1965 ment).6 It has been noted,6 however, that the study of 3 was perhaps not done under conditions favorable for rearrangement (thiophenol was present) ; thus this example is possibly not apt. Another instance where a rearrangement of this type might have occurred was in the thermal decomposition of bis(l,3,5-tri-t-butyl2,5-cyclohexadien-4-one) peroxide (7) .' Some 2,6di-t-butyl-4-t-butoxyphenol (8) was isolated, possibly formed as the result of the process shown as path a.
Some quinone ( 9 ) , isobutane, and isobutylene were isolated also, however, and the formation of 8 could be accommodated just as well by a ,8-cleavage-readdition pathway (path b). (3) The general ease of oxygenoxygen homolysis compared with carbon-carbon homolysis indicates a somewhat greater stability for the oxy radicals, and the enthalpic change associated with the rearrangement of 1 to 2 might therefore be favorable. This last is further borne out by the heats of formation of alkyl vs. alkoxy radicals; the former are in the range of 4 to 30 kcal./mole (excluding the exceptionally stable trifluoroniethyl radical, whose heat of formation is -117 f 2 kcal./mole), while the latter are in the range of -0.5 to -30 kcal./mole.* While not necessarily so by any means, such literature items indicate tthat the process 1 -F 2 might proceed. The di-t-butyl peroxide (DTBP) induced decarbonylation of phenoxyacetaldehyde was therefore performed. It proceeded only partially to completion, however, and yielded only anisole as a monomeric product. No evidence was found for the rearrangement of 1 to 2. DTBP
PhOCH,CHO
-+ PhOCH3 + polymer A
+ CO
The results of the several runs made are given in Table I. I t is seen that the yield of anisole was usually low and that generally only about' one-third of the phenoxymethyl radicals (1) terminated as anisole. The exception was run 5 where the benzyl mercaptan present completely trapped out 1. In all runs made, careful searching failed to find any products likely from 2 such as benzyl alcohol or benzaldehyde. Of course, neither of these two products would probably be inert under the reaction condit'ions. Benzaldehyde is converted to meso-dihydrobemoin dibenzoate by DTBP,8 (5) E. S. Gould, "Mechanism and Structure in Organic Chemistry,"
H. Holt and Co.. New York. N. Y., 1959, pp. 758-759. (6) H. Wieland, Ber.. 44, 2550 (1911). (7) C. D. Cook, R. C. Woodworth, a n d P. Fianu, J. Am. Chem. S o c . , 7 8 , 4159 (1956). We thank a referee for this reference. We do not, however, agree with him t h a t it necessarily represents a rearrangement (see text). (8) N. N . Semenov, "Some Problems in Chemical Kinetics and Reactivity," Vol. 1, M. Boudart, Transl., Princeton University Press, Princeton, N . J.. 1958, p. 23. (9) F. F. R u s t , F. H. Seubold, and W. E. Vaughn, J. A m . Chem. Soc., TO, 32.58 (1948).
1257 TABLEI PHENOXYACETALDEHYDE DECARBONYLATIONS
% COO % anisolea 4 75 21 6 6.25 63 21 ac 6 25 53 21 4d 9.0 52 21 5' 8 5 81 81 a Baaed on the aldehyde consumed. Usually 30-50% of the Excess peroxide waa used. Noraldehyde waa recoverable. mally 30 mole yo of IITBP waa employed (e.g., run 1) a t a temperature of 172'. Here, however, 225 mole % of DTBP was employed a t this temperature. E 290 mole yoof DTBP a t 175". The aldehyde waa 1 M in chlorobenzene a t reflux temperature. e The aldehyde waa 1 M in chlorobenzene a t reflux with 5 mole yo of benzyl mercaptan present. Run
Time, hr.
1 2b
and benzyl alcohol would be expected to give other products via the a-hydroxybenzyl radical. l o The absence of either substance to the limit of detection (about 3 mole %), particularly under more dilute conditions, is considered significant. The reported attack of phenoxymethyl radicals upon anisole" to yield (inter alia) isomeric phenoxymethylanisoles, as shown, gave a clue to the nature of the DTBP
PhOCHa -+ o,m,p-CeH,OCHrCeH,OCHa A
(disappointingly intractable) residues that made up the bulk of the product in several of the runs made. These residues showed intense ether absorption (8.1 k ) in tJhe infrared. On the other hand, only weak carbonyl absorption was notJed. This argues against much meso-dihydrobenzoin dibenzoates in the residue. Small amounts nevertheless could have been there. These facts, t'ogether with the observed molecular weight of over 400 for these residues, suggest that 1 attacked phenoxyacetaldehyde in much the same manner as 1 attacks anisole. The ensuing radicals appear to be poor chain carriers, however, because the 30 mole % of peroxide needed in this decarbonylation, rat'her than the 10 mole % normally effective,2 indicates a shorter chain length than normal. Further complicat'ing work on these residues, however, was the observat'ion that phenoxyacetaldehyde resinified upon heating to 150° and undoubtedly thereby made up a portion of the residue itself. I t is conceivable that 2 was formed t'o some extent from 1 but that such rearrangement was then obscured through consumption of 2 by paths as yet' unknown. The available evidence, however, suggests that the rearrangement) did not occur. Dilution did not afford any change in the monomeric product, contrary to the results obtained in ot'her systems.'* Also, the ability of benzyl mercaptan to trap all of 1, with no sign of rearranged product's, indicat'ed that if 1 rearranged, it did so a t a slower rate than it underwent chain transfer. While this in itself is not uncommon, neither is it, uncommon to observe st'ill some rearrangement in the presence of the hydrogen donor.12 Sonet>heless, we feel that the present condit'ions, though worth trying, did not afford the best opportunity to observe t.he (10) C. Walling, ref. 4, p. 285 ff. (11) H. B. Henbest, J. A. Reid, and C. J. M . Stirling, J. Chem. S o c . , 5239 (1061).
(12) C. Walling, "Molecular Rearrangements," Vol. 1, P. de Mayo, E d . , Interscience Publishers, Inc.. New York. N. Y., 1963, p. 409 8.
1258
NOTES
desired rearrangement. Our eff0rt.s to find this and other “hetero” analogs of the neophyl rearrangement are therefore continuing. Experimental Phenoxyacetaldehyde was prepared by the oxidation of glycerol a-phenyl ether ( a gift from the Miner Laboratories, Chicago, Ill.) with lead tetraacetate according to Speer and MahlerL3(48%, b.p. 82” a t 4 mm., semicarbazone m.p. 144.5-145.5”, 2,4-dinitrophenylhydrazone m.p. 137.5-138’, in agreement with literature valuesL3). The decarbonylations were performed as described in other work.Z The reaction material was processed via gas chromatography (g.c.) on a Perkin-Elmer Model 154’2 instrument equipped with a column of Union Carbide X-525 silicone oil (15% on firebrick, 6 f t . X 0.25 in., 128”, helium carried gas pressure 20 p.8.i.g.) Assignment of structure to the products isolated by g.c. was via infrared and g.c. comparison with knowns. Mixing experiments indicated that about 3 mole % of benzyl alcohol and benzaldehyde each could readily be detected in the residues from decarbonylation . Carbon monoxide was collected over water and determined by absorption in cuprous sulfate-p-naphtho1 in a Fisher-Orsat apparatus. The molecular weights of the nonvolatile reaction residues from several runs were determined cryoscopically by the freezing point depression of benzene. Further details of the experiments performed may be found in Table I. (13)
R. J. Speer and H. R. Mahler, J. A m . Chem. SOC.,71, 1133 (1949).
VOL. 30
in no case does the cis: trans ratio exceed unity, and in most cases does not deviate appreciably from the “equilibrium value” which favors the trans isomer by a factor of 3 : l a t 25°.4 We are unable a t this time to describe the exact mechanism of the elimination of lithium hydride from alkyllithium compounds. It is clear, however, that the preference for the cis isomer is indicative of unusual steric requirements in the activated complex. This may be due to the well-known association of alkyllithium compounds in both the pure state and in organic solvents. Thus the crystal structure of ethyllithium consists of tetramer^,^ while hexamers and tetramers are indicated in the vapor phase6 and in benzene.’ t-Butyllithium is tetrameric in benzene.* It is conceivable, for example, that the orientation of a sec-butyllithium tetramer into the proper conformation for cis @-elimination results in the preference of cis-2-butene over the trans isomer. The increase in the percentage of the latter in octane may be accounted for by a rapid equilibrium between tetramers (or hexamers) and dimers in which the steric requirements are less rigid. Experimental
The Thermal Decomposition of sec-Butyllithium WILLIAMH. GLAZE,JACOB LIN, A N D E. G. FELT ON^ Department of Chemistry, Xorlh Texas State University, Denton, Texas Received October 6, 1964
Although the thermal decomposition of ethyllithium and n-butyllithium have been reported by Ziegler2 and Bry~e-Smith,~ no systematic investigation of the reaction has been made. In the course of such a study we have examined the products of the decomposition of sec-butyllithium in the pure, liquid state and in hydrocarbon solutions. The solid product has been identified as virtually pure lithium hydride. With the exception of a small amount (300". Less reactive isocyanates required heating a t 80". Trimers obtained in this manner were: o-methoxyphenyl, 1 hr., %Yo (lit.1 m.p. 95"); p-tolyl, 6 hr., 22%, m.p. 268" (1it.I m.p. 264'). Trimers obtained a t room temperature in heptane or DMSO over a period of 24 hr. were: m-chlorophenyl, 97%; phenyl, 96%; m-nitrophenyl, 71%, m.p. 245" dec.; toluene diisocyanate, loo'%, m.p. >300°; hexamethylene diisocyanate (5 mole % oxide used), 10070, m.p. >300"; the latter two were run in DMSO. Reaction of Tri-n-butylantimony Oxide and Phenyl Isocyanate. -When 7.14 g. of phenyl isocyanate (0.06 mole) was added to 15.4 g. of tri-n-butylantimony oxide (0.05 mole) in a drybox, an exothermic reaction was noted. After 4 days a t room temperature, the reaction mixture waa triturated with heptane to yield some triphenyl isocyanurate (m.p. 278-281 "); the heptanesoluble fraction, upon distillation of solvent, yielded 17.1 g. of a nondistillable oil ( n Z 4 ~1.5275) whose elemental analysis indicated it to be a stoichiometric complex of tri-n-butylantimony oxide and phenyl isocyanate. Anal. Calcd. for C10H27NOzSb: N, 3.3. Found: N, 3.4. The structure of this complex was confirmed by infrared analysis which indicated the absence of NH absorption, unreacted isocyanate (yc-0 2240 cm.-l), or triphenyl isocyanurate ( y c - 0 , 1705 cm.-1); a strong carbonyl absorption a t 1725 cm.-' was noted. Attempts to prepare this complex by heating triphenyl isocyanurate and the oxide were unsuccessful. The complex was an active trimerizing agent.
Synthesis of Isocyanatoorganosulfonyl Isocyanates and Organodisulfonyl Isocyanates1 J. SMITH,JR., T. K. BROTHERTON, AND J. W. L Y N N
VOL. 30
A few aliphatic and aromatic monosulfonyl isocyanates are described in the literature. These materials were prepared either by the reaction of sulfony1,chlorides with silver cyanate,2 by the reaction of sulfonic anhydrides with silver cyanate,S or by the direct phosgenation of monosulfonamides. By the silver cyanate methods, the.yields were generally low (5 to 38%); phosgenation gave yields of about 80%.
+ AgCNO A+ -3 CHsSOzNCO (CHsS0z)zO + AgCNO CeHsSOzNHz + COClz +CsH6SOzNCO CHsSOzCl
Sulfonyl diisocyanates I and 11, respectively, were prepared by the reactions of chlorosulfonyl isocyanate with silver ~ y a n a t e and , ~ potassium cyanate with sulfur trioxide.6 CNCl
AgCNO + so3 +ClSOZNCO +OCNSOzNCO
I 4SOa
+ 2KCNO +&SzO7 + Sz06(NCO)s I1
p-Isocyanatobenzenesulfonyl isocyanate (111) and the organodisulfonyl isocyanates IV-VI, representing new classes of diisocyanates, have now been prepared by the direct phosgenation of p-aminobenzenesulfonamide (VII) and organodisulfonamides, respectively. SOzNCO
OCNOzS- R- SOzNCO IV, R = -(CHz);r
V, R = -(CHzk NCO
I11
Synthesis and Derivatives of p-Isocyanatobenzenesulfonyl Isocyanate (III).-Two-stage phosgenation of VI1 in nitrobenzene solvent afforded 111 in 87% yield. This new diisocyanate is nonlachrymatory at room temperature, is obtainable in a high degree of purity by a simple distillation, and is extremely reactive with active hydrogen-containing materials, ie., water, alcohols, etc. The conversion of VI1 to I11 involves: (1) addition of a slurry of VI1 in nitrobenzene to a solution of phosgene in nitrobenzene at 0" to form presumably a mixture of p-aminobenzenesulfonamide hydrochloride and
0
SOzNCO
Research and Development Department, Union Carbide Corporation, Chemicals Division, South Charleston, West Virginia
SOzNHa
SOiNHz
i3HCl
Received November 4, 1964
Phosgenation of p-aminobenzenesulfonamide (VII) in an inert medium at 150" gave p-isocyanatobenzenesulfonyl isocyanate (111), while a t 90" the product was p-isocyanatobenzenesulfonamide (IX) . Several derivatives of I11 were prepared illustrating the difference in the reactivity of the sulfonyl and phenyl isocyanate groups contained in this molecule. Organodisulfonyl isocyanates IV-VI were prepared by the direct phosgenation of organodisulforiainides. ( 1 ) This paper was presented at the Southeastern Regional Meeting of the American Chemical Society, Charlotte, N . C., Nov. 1963.
Nco IX (2) 0. C. Billeter, Ber., SO, 2013 (1905). (3) L. Field, J . Am. Chem. SOC.,74, 394 (1952). (4) H. Krzekalla (to Badische Anilin- and Soda-Fabrik Aktiengesellsohaft), U. 9. Patent 2,666,787 (1950). (6) R. Appel and H. Gerber, Ber.. 91, 1200 (1958). (6) R. Appel and H. Gerber, Angew. Chem.. TO, 271 (1958).
NOTES
APRIL1965 p-(chloroformamido) benzenesulfonamide (VIII) ; (2) elevating the temperature of the reaction mixture to 150" and maintaining this temperature while passing phosgene through the reaction mixture for 3 hr.; (3) removal of the solvent by flash distillation; and (4)distillation of the product under reduced pressure. Low-temperature (90") phosgenation of the reaction product of VI1 and phosgene in the cold (0') gave I X in 710/, yield. I11 was allowed to react with amines and alcohols to give the derivatives shown in Table I. TABLE I
O H
H O
R X - b - P ! J ~ S O 2 - N - - C - XI R R
X
>NH CHFCHCHT >NH CdHr -0CH3-0CH-CHCHi0 Temperatures are uncorrected.
It
M.p., OC.a
Yield, %
203-204 192-193.5 203-205 158-1 59
60 85 38 96
Sulfonyl isocyanates have been reported to be much more reactive with common protic reagents than are aryl and alkyl isocyanates. This difference is shown by the products obtained from the reaction of water with 111, which contains both the aryl and sulfonyl isocyanate group. The reaction of equimoles of I11 and water gave a white, crystalline solid which was identical with IX. Further reaction of IX with water gave 4,4'-ureylenebis(benzenesu1fonamide) (X) , identical with the coinpound prepared by the reaction of 2 moles of I11 with 3 moles of water.
Q
+1 HzO
Q
NCO
I11
NCO
X
I11
Synthesis of Organodisulfonyl Isocyanates.-Organodisulfonyl isocyanates were prepared by the direct phosgenation of organodisulfonamides. I V and V were prepared in 57 and 62% yields, respectively, by the phosgenation of the corresponding disulfonamides in nitrobenzene at 170-180". At temperatures less than l50", no appreciable reaction was noted after several hours of phosgenating. To prepare VI it was necessary to utilize a temperature of 250" and a chlorinated phenyl ether (Arochlor 1260) as solvent for the phosgenation. This was presumably due to the insolubility of the aromatic disulfonamide a t the lower temperatures. VI was precipitated from the solvent in 54% yield and subsequently refined by sublimation.
1261 Experimental7
p-Isocyanatobenzenesulfonyl Isocyanate (III).-A slurry of 172 g. (1.0 mole) of p-aminobenzenesulfonamide in 1000 g. of nitrobenzene waa added to a solution of 396 g. (4.0 moles) of phosgene in 548 g. of nitrobenzene while the kettle temperature was maintained a t -10 to 0". The temperature of the resulting mixture was raised to 157' and the mixture was sparged with phosgene for 3 hr. The resulting clear solution was treated with phosgene for an additional 2 hr., and subsequently with nitrogen to remove phosgene and by-product hydrogen chloride. The solvent waa removed by flash distillation and the product WM distilled to give 194 g. (86.7%) of I11 with a boiling point of 108-109' (0.15 mm.), a n d a freezingpointof42.5". The product had infrared absorption a t 4.45 (NCO), 6.25 and 6.55 (phenyl C=C), 7.40 and 8.60 (SO,), and 11.95 p (para disubstitution). Anal. Calcd. for CsH4N2OB: C, 42.80; H , 1.78; N, 12.48; S, 14.30. Found: C, 42.62; H, 1.96; N , 12.25; S,14.25. p-Isocyanatobenzenesulfonamide (IX) .-The same general procedure was used aa described for the preparation of 111, except the temperature waa maintained a t 90". I11 (172 g., 1 .O mole) in nitrobenzene treated with liquid phosgene a t 0" and then with gaseous phosgene a t 90" for 4 hr. wm purged free of phosgene with nitrogen, cooled, and filtered. The solid product was washed with ethyl ether and dried to yield 141 g. (71.2%) of I X , which on recrystallization from toluene had m.p. 156-157'. The product had infrared absorption a t 2.99 (iYHZ), 4.45 (NCO), and 7.50 and 8.70 p (SO2). Anal. Calcd. for C,HBN,03S: C, 42.4; H , 3.03; N , 14.13. Found: C, 42.22; H , 3.30; N , 14.19. I X in methyl isobutyl ketone reacted with aniline to give the anilide, m.p. 229-230O: Infrared maxima a t 5.95 (substituted ~ urea C=O), 7.6 and 8.7 (SOz),and 13.35 and 1 4 . 4 (monosubstituted aromatic) are consistent with the proposed structure. Anal. Calcd. for CI3H13N3OaS: N, 14.43. Found: N , 14.41. p-Isocyanatobenzenesulfonamide (IX) by Reaction of Equimoles of I11 and Water.-Water (1.8 g., 0.1 mole) was added to 22.4 g. (0.1 mole) of I11 dissolved in 250 ml. of benzene. The reaction temperature rose from 28 to 38" during the addition. When the reaction had subsided, the mixture was filtered to yield 15 g. (75.8y0) of I X , which on recrystallization from toluene had m.p. 155-157'. Anal. Calcd. for C,Hs?J203S: C, 42.4; H , 3.03; N, 14.13. Found: C, 42.3; H, 3.42; N, 13.98. A mixture melting point of this material with that prepared by phosgenation was 156-157'. 4,4'-Ureylenebis( beazenesulfonamide) (X)by Reaction of I11 with an Excess of Water.-To an agitated solution of 22.4 g. (0.1 mole) of I11 in 100 ml. of acetone there waa added 9.0 g. (0.5 mole) of water in 25 ml. of acetone. Stirring was continued for 30 min. a t ambient temperature, after which the temperature was elevated t o reflux (55") and maintained for 2 hr. Cooling and filtration afforded X (10.0 g., 28% of theory) which on recrystallization from water had m.p. 285-286". Infrared absorption a t 3.02 and 3.1 (NH, NH,), 5.91 (substituted urea C=O), and 7.55 and 8.7 p (SO2) are consistent with the proposed structure. A n d . Calcd. for C I ~ H , , N ~ O ~C,S ~42.15; : H , 3.81; N, 15.13. Found: C,42.26; H,3.97: N, 15.21. 4,4'-Ureylenebis( benzenesulfonamide) (X)Prepared from IX and an Excess of Water.-IX (1.98 g., 0.01 mole) prepared by the reaction of I11 with equimoles of water, was stirred for 4 hr. with an excess of water in acetone. Subsequent filtration and drying under reduced pressure afforded X ( 1 . 4 g., 827, of theory) which on recrystallization had m.m.p. 284286" with X , prepared from I11 and an excess of water. N-(Methoxycarbonyl)4-methoxyformamidobenzenesulfonamide.-To a solution of 8.25 g. (0.04 mole) of I11 in 80 ml. of dry acetone was added slowly 5.9 g. (0.18 mole) of methanol in 20 ml. of dry acetone. The mixture was allowed t o stir a t ambient temperature for 30 min. after the addition, the temperature was elevated to 40", and heating was continued for 2 hr. Subsequent filtration and recrystallization from water gave the product (4.1 g., 38% of theory), m.p. 203-205". Infrared absorption a t 2.95 ( N H ) , 3.3 (aromatic CH), 3.35 (CH,), 5.78 (carbamate C=O), 6.47 (carbamate NH), 6.65 (aromatic C=C), 7.37 and 8.55 (SO2), and 8.1 p (carbamate C-0) are consistent with the structure. (7) All temperatures are uncorrected
NOTES
1262
Anal. Calcd. for ClOHl2N2O6S:C, 41.7; H , 4.17; N, 9.73. Found: C,41.99; H,4.22; N,9.48. N-( Allylcarbamoyl)4allylureylenebenzenesulfonamide .-In a manner similar to that employed in the synthesis of the allyl alcohol derivative, 97 g. (1.7 moles) of allylamine in benzene was treated with 168 g. (0.75 mole) of I11 in benzene. The product obtained in 60% yield was recrystallized from ethanol, m.p. 203-204'. Anal. Calcd. for C14H18N404S:C, 49.75; H , 5.36; N, 16.56. Found: C, 49.79; H , 5.67; N , 16.15. N-( n-Butylcarbamoyl)-4-(n-butylureylene)benzenesulfonamide.-In a manner similar to that employed in the synthesis of the allyl alcohol derivative, 73.1 g . (1.0 mole) of n-butylamine in benzene was treated with 101 g. (0.45 mole) of 111. The product, obtained in 8577, yield, melted a t 192-193.5'. Anal. Calcd. for Cl6HZ6N4O4S:C, 51.90; H , 7.70; N , 15.10. Found: C, 52.0; H, 7.50; N, 14.80. 1,4-Butanedisulfonyl Isocyanate (IV).-Phosgene was sparged through a nitrobenzene (279 g.) solution of 31 g. (0.15 mole) of 1,4-butanedisulfonamide a t a rate of 1.0 mole/hr. for 2 hr. a t loo0, and then for 5 hr. a t 160". The resulting solution was sparged with nitrogen for 1 hr., filtered, and the nitrobenzene was removed under reduced pressure. The residual liquid solidified on cooling. The solid was washed with anhydrous ether and dried. The product (22.5 g., 57%) was isolated with a melting point of 60-63" and infrared maxima a t 4.45 (NCO) and 7.5 and 8.60 p (SOZ). Anal. Calcd. for C6H8N2O6S2:C, 26.9; €3, 2.99; N, 10.45. Found: C, 26.66; H , 3.86; N , 10.78. 1,s-Pentanedisulfonyl Isocyanate (V).-The same general procedure was used as described for the synthesis of IV. Phosgene was sparged through a 10% nitrobenzene solution of 1,5pentanedisulfonamide (23 g., 0.1 mole) for 5 hr. a t 165". The crude product (17 9.) was isolated in 60.2% yield which was subsequently flash distilled a t 200" (0.1 mm.). Infrared absorption has the expected maxima a t 4.45 (NCO) and 7.4 and 8.6 p (SOZ). Anal. Calcd. for C7H10N206S2: C, 29.8; H , 3.55. Found: C, 29.45; H , 3.90. 1,s-Naphthalenedisulfonyl Isocyanate (VI).-A slurry of 143 g. (0.5 mole) of 1,5-naphthalenedisulfonamidein 2000 g. of Arochlor 1260 (chlorinated phenyl ether) was maintained at 250" as gaseous phosgene was added a t a rate of 1.0 mole/hr. for 10 hr. The resulting solution was sparged with nitrogen for 1 hr. and then filtered. Anhydrous ether (500 ml.) was added to the filtrate and 91 g. (54%) of product was isolated which melted a t 185-190'. The analytical sample was purified by sublimation a t 240-260' and 0.1 mm. Anal. Calcd. for Cl&?i&Sz: c , 42.79; H, 1.77; N, 8.31; S, 18.36. Found: C, 42.84; H , 1.96; N , 7.92; S, 18.56.
Syntheses and Reactions of Some Hindered Organophosphorus Compounds' A. GILBERT COOK^ Argonne National Laboratory, Argonne, Illinois Received J u l y 8, 1964
There has been little discussion in the literature of the effect of steric hindrance on the reactions of organophosphorus compounds. Triphenylmethylphosphony1 dichloride (1) can be hydrolyzed to the corresponding phosphoriic acid 2 only with difficult^.^-^ This is probably due to the steric shielding of the phos(1) Based on work performed under the auspices of the U. S. Atomic Energy Commission. (2) Correspondence should be addressed t o Department of Chemistry, Valparaiso Univeraity, Valparaiso, Ind. (3) D. R. Boyd and G. C h i g n e k J . Chem. Soc., 123, 813 (1923). (4) H. H. H a t t . ibid.. 2412 (1929). (5) G . hZ. Kosolapoff, O w . Reactions. 6 , 273 (1951). (6) M. Halmann, L. Kugel. and S. Pinchas, J . Chem. Soc., 3542 (1961).
VOL. 30
(e
.
0
(@cpQ
1. KOH, EtOH 2.HI
1
n
c,(ofi
2
phorus atom by the three phenyl groups. Application of the synthetic procedure by which phenylphosphorus dichloride was obtained from benzene' (namely by the use of phosphorus trichloride and aluminum chloride with the aromatic hydrocarbon) to mesitylene, durene, and pentamethylbenzene primarily produced the corresponding diarylphosphinic chlorides.8 Hydrolysis of these hindered phosphinic chlorides resulted in the formation of some surprisingly stable secondary diarylphosphine oxides which could not be oxidized to phosphinic acids by the normal procedure using alkaline hydrogen peroxide.8 Resistance to oxidation by alkaline ferricyanide seemed to increase with increasing methyl substitution, on the ring. These observed phenomena would appear to be due largely to steric effects. Resistance to both hydrolysis and oxidation by organophosphorus chlorides which possess a substantial amount of steric hindrance, as illustrated by the examples mentioned above, has been found to an even greater extent in a more radically hindered organophosphorus chloride, namely 2,4,6-tri-t-butylphenylphosphinic chloride (4). This compound was readily synthesized in a 71% yield by treating 1,3,5tri-t-butylbenzeneg (3) with phosphorus trichloride and anhydrous aluminum chloride followed by hydrolysis. 4 0
x
3
4
5
The attempts to oxidize 2,4,6-tri-t-butylphenylphosphinic chloride with alkaline hydrogen peroxide or chlorine were not successful as shown by complete recovery of unchanged starting material. A substitution product, 2,4,6-tri-t-butylphenylphosphinicanhydride ( 5 ) , was obtained in a small yield (15%) froin the attempted oxidation of compound 4 with alkaline potassiuni ferricyanide. Use of potassium permanganate in refluxing alkaline solution as an oxidizing agent for compound 4 strikingly illustrated the difficulty of oxidizing the phosphorus-hydrogen bond in this compound. It did oxidize one of the t-butyl groups to a carboxylic acid and hydrolyze the phosphorus chloride, thereby proacid ducing 2,6-di-t-butyl-4-carboxylphenylphosphinic (6) in a very small yield (7y0) ; but the phosphorushydrogen bond in this product remained intact. It was shown to be a dibasic acid by the two breaks in its titration curve. The infrared spectrum of 6, when run (7) B. Buchner and L. B. Lockhart, Jr., "Organic Synthesis," Coll. Vol. I V , John Wiley and Sons, Inc., N e w York, N. Y., 1963, p. 764. (8) A. W. Frank. J . O r g . Chem., 24, 966 (1959). (9) L. R . C. Barclay and E. E. Betts. Can. J . Chem., 88, 672 (1955).
NOTES
APRIL1965
4
6
in a potassium bromide pellet, showed no bands in the 3500-3600-cm.-l region. It did exhibit several bands in the 2500-3000-cm.-1 region and a band a t 1705 cm.-l. The infrared spectrum of 6 in dioxane solvent (in which carboxylic acids are usually nionomericlO) showed a strong doublet appearing a t 3550 and 3625 cm. - l along with a band a t 1735 cni.-'. This indicates that) there is intermolecular rather than intramolecular hydrogen bonding present. Therefore the carboxylic acid group probably is para to the phosphinic acid group, assuming no rearrangement has taken place. There have been very few reports of a t-butyl group on an aromatic ring being oxidized to an aromatic carboxylic acid. l 1 The treatment of 1,3,5-tri-t-butylbenzenewith phenylphosphonous dichloride and aluminum chloride resulted in the formation of p-t-butylphenylphenylphosphine oxide (7) in about a 4Oj, yield as the only product isolated. The excess of phenylphosphonous di-
0
7
chloride in conjunction with the aluminum chloride caused the dealkylation of the 1,3,5-tri-t-butylbenzene.'2 The position of the t-butyl group was assigned as para on the basis of the infrared spectrum of 7 which exhibited strong maxima a t 820 (due to para disubstitution), 755, and 725 cm.-', the latter two bands being due to a monosubstituted benzene ring. Experimental
2,4,6-Tri-t-butylphenylphosphinicChloride .-A stirred mixture 60 ml. (0.68 of 42.23 g. (0.17 mole) of 1,3,5-tri-t-but~lbenzene,~ mole) of phosphorus trichloride, and 33.3 g. (0.25 mole) of aluminum chloride was refluxed for 4 hr. After the reaction mixture was cooled to room temperature, 250 ml. of methylene chloride was added (when chlorine gas was added a t this point no oxidation was found to take place) and the reaction mixture was poured into a water-ice mixture. The methylene chloride phase was separated, and the aqueous phase was extracted several times with fresh methylene chloride. The combined methylene chloride extracts were dried and filtered and the solvent was removed to yield a crystalline solid product. After one recrystallization from petroleum ether (b.p. 30-60") a total of 38.0 g. (0.12 mole) of product was obtained representing a yield of 71%. I t crystallized from petroleum ether as colorless needles, m.p. 133-134", XI: 2340 cm.-l (P-H). Anal. Calcd. for C18H30ClOP: C, 65.74; H , 9.20; P , 9.42. Found: C, 65.62; H,9.24; P,9.76. The product could not be oxidized with alkaline hydrogen peroxide. Attempted Oxidation of 2,4,6-Tri-t-butylphenylphosphinic Chloride with Alkaline Potassium Ferricyanide.-A stirred mixture of 45.8 g. (0.14 mole) of 2,4,6-tri-t-butylphenylphosphinic chloride, 54.0 g. (0.16 mole) of potassium ferricyanide, 32.8 g. of sodium hydroxide, and 2 1. of water was heated a t 80-'100" for 22 hr. The solution was acidified with hydrochloric acid and extracted twice with benzene. The combined benzene extracts (10) M. St. C. Flett. J . Chem. Soc., 962 (1951). (11) I. V . Butina and V. G. Plyusnin, T r . Vsee. Soueehch. PO K h i m . Pererabotke N e / t . Uoleuodorodou v Poluprod. dlua Sinteza Volokon i Plast. Maele, Baku. 131 (1957). (12) V. S . Ipatieff and B. B. Corson, J . A m . Chem. Soc., 69, 1417 (1837).
1263
were dried and filtered and the solvent was removed. Some petroleum ether (b.p. 20-40") was added to the residual oil to remove starting material and all but some colorless solid dissolved. Upon filtration, a total of 6.2 g. (0.01 mole) of 2,4,6tri-t-butylphenylphosphinic anhydride was obtained representing a yield of 15%. It crystallized from ethyl acetate as colorless needles, m.p. 259.5-260.5', Azfl2340 cm.? (P-H). Anal. Calcd. for CaGHeoOaPz: C, 71.72; H , 10.03; P, 10.28. Found: C,71.74; H , 10.40; P , 10.21. Oxidation of 2,4,6-Tri-t-butylphenylphosphinicChloride with Alkaline Potassium Permanganate.-A stirred mixture of 18.6 g. (0.06 mole) of 2,4,6-tri-t-butylphenylphosphinic chloride, 39.5 g. (0.25 mole) of potassium permanganate, and 250 ml. of 2 ATaqueous sodium hydroxide was refluxed for 16 hr. All of the potassium permanganate had reacted by this time and the manganese dioxide precipitate was removed by filtration. The basic aqueous solution was washed with benzene, acidified with hydrochloric acid ( a t which time a gas was evolved from the solution), and extracted with benzene and 3-hexanone. The solvents were removed from the combined solutions and a thick oil remained. Addition of petroleum ether (b.p. 30-60") to the oil caused a white insoluble precipitate to separate. A total of 1.2 g. (0.004 mole) of the white crystalline product, 2,6-di-t-butyl-4-carboxylphenylphosphinic acid, was isolated by filtration representing a yield of 7%. I t was dissolved in aqueous sodium hydroxide, washed with benzene, and reprecipitated with hydrochloric acid. I t crystallized from n-heptane as colorless plates, m.p. 226-227'. Anal. Calcd. for C16H2304P:C, 60.39; H, 7.77; P , 10.38; neut. equiv., 149. Found: C,59.97; H,7.59; P , 10.77; neut. equiv., 144. The infrared spectrum (KBr pellet) exhibited a band a t 1705 cm.-l, and the spectrum using dioxane solvent showed bands a t 3625, 3550, and 1735 cm.?. Likewise bands were found (KBr pellet) a t 2360 (P-H) and 2650 cm.? (P-OH). Reaction of 1,3,5-Tri-t-butylbenzenewith Phenylphosphonous Dichloride.-A stirred mixture of 45.4 g. (0.18 mole) of 1,3,5-trit-butylbenzene, 128.9 g. (0.72 mole) of phenylphosphonous dichloride, and 40 g. (0.3 mole) of anhydrous aluminum chloride was refluxed for 4 hr. Then, after adding 28 ml. (0.3 mole) of phosphorus oxychloride, refluxing for 0.5 hr., and washing with petroleum ether (b.p. 30-60"), the residue was added to a waterice mixture. The aqueous solution was extracted with benzene, the combined benzene extracts were washed with aqueous sodium hydroxide and water, dried, and filtered, and the solvent was removed. Acetone was added to the residual oil and a colorless solid remained insoluble. A total of 2.02 g. (0.008 mole) of p-tbutylphenylphenylphosphine oxide was obtained which represents a yield of 4%. I t separated from n-heptane as colorless 2370 (P-H), 820, 755, and 725 needles: m.p. 130-131"; A;: cm.-l. I t can also be purified by sublimation. Anal. Calcd. for Cl6HlgOP: C, 74.40; H , 7.41; P , 11.99; mol. wt., 258. Found: C, 74.31; H , 7.39; P , 12.19; mol. n t . (ebulliscopic determination in benzene), 248.
Problems of Orientation in Arylphosphonic Acids. I. 3-Chloro-4-tolylphosphonic Acid' LEOND. FREEDMAN AND G. 0. DOAK Department of Chemistry, h'orth Carolina State of the University of Korth Carolina at Raleigh, Sorth Carolina Received August 4, 1964
3-Chloro-4-tolylphosphonic acid was first reported by RIelchiker2 and is listed in Kosolapoff's monograph3; it is also mentioned in the patent 1iteratu1-e.~The synthetic procedure used by Alelchiker involved a Friedel(1) The work was supported in part b y Research Grant GM-09479 from the National Institutes of Health, Public Health Service. (2) P. Melchiker, Ber., 81, 2915 (1898). (3) G . M. Kosolapoff, "Organophosphorus Compounds." John Wiley and Sons,Inc.. New York, N . Y . , 1950, p. 166. (4) K. C. Whitehouse and H. Z. Lecher, U. S . Patent 2,894.024 (July 7, 1959).
NOTES
1264
Crafts reaction between o-chlorotoluene and phosphorus trichloride to yield a phosphonous dichloride which was subsequently oxidized to a phosphonic dichloride; hydrolysis of the latter compound yielded a phosphonic acid which was reported to melt a t 190'. In an attempt to establish the structure of this acid, Melchiker allowed it to react with bromine to yield a bromochlorotoluene which was then oxidized to a bromochlorobenzoic acid, m.p. 15.5-156'. He believed that this melting point agreed with the melting point of 2chloro-4-bromobenzoic acid and that therefore the structure of his phosphonic acid must have been 3chloro-4-tolylphosphonic acid.
CH,
CH3
CO, H
There are several reasons for questioning the correctness of Melchiker's conclusions. In the reaction of electrophilic reagents with o-chlorotoluene, the entering group tends to go mainly para to the chloro group. This is true both in the Friedel-Crafts reaction5 and in nitration.6 Thus one might expect Rlelchiker's phosphonic acid to be 4-chloro-3-tolylphosphonic acid, although the possibility of the Friedel-Crafts reaction yielding mixtures of organophosphorus compounds should not be o ~ e r l o o k e d . ~ A possible flaw in Melchiker's proof of structure concerns the melting point of 2-chloro-4-bromobenzoic acid, According to Heilbron,8 the melting point of this acid is 166-167' and not 156' as stated by Xlelchiker. The melting point of 2-chloro-5-bromobenzoic acid, on the other hand, is 155-156O.R This evidence also suggested that Melchiker's phosphonic acid was 4chloro-3-tolylphosphonic acid.
I
c1
I
c1
In connection with another p r ~ b l e m we , ~ had the occasion to prepare 3-chloro-4-tolylphosphonic acid from the corresponding diazonium fluoroborate. The melting point of this acid was 160-162'. Kitration yielded a mononitro derivative which melted with decomposition a t 223-227O. Melrhiker had reported that nitration of his phosphonic acid yielded a compound, 1ii.p. 200'. A t this point in our study it seemed certain to us that he had not prepared 3-chloro-4-tolylphosphonic acid, and we assumed a t first that his mate(5) C. A. Thomas, "Anhydrous Aluminum Chloride in Organic Chemistry," Reinhold Publishing Corp., New York, N. Y . . 1941, pp. 223-225. (6) J. J3. Wibaut. Rec. trau. c h i n . , 93,243 (1913). (7) G . M. Kosolapoff, J . Am. Chem. Soc., '74, 4119 (1952); R . A. Baldwin, K. A. Smitheman, and R . M. Washburn, J. Org. Chem., 26, 3547 (1961); R. Schmuteler, J . Inorg. Nucl. Chem.. 3 6 , 3 3 5 (1963). (8) I. Heilbron and H. M. Bunbury. "Dictionary of Organic Compounds," Val. 1 . Oxford University Press, New York, N . Y.. 1953, p. 477. (9) Cf. L. D . Freedman, G . 0. Doak, and J. W. Clark, Jr., Clin. Med.. 71, 351 (1964).
VOL. 30
rial must have been 4-chloro-3-tolylphosphonic acid. When we prepared the latter compound by an unequivocal route, however, we found that it melted a t 158-161'; and thus the nature of Melchiker's phosphonic acid had still not been elucidated, In order to obtain further information concerning this material, the Friedel-Crafts reaction between ochlorotoluene and phosphorus trichloride was reinvestigated. We employed two different Friedel-Crafts procedures, viz., the one used by Melchiker and the more convenient technique recently described by Gefter.lo The results obtained by both procedures were very similar, except that the yield of phosphonous dichloride was 57% by Gefter's method and only 6% by Melchiker's. Hydrolysis of either phosphonous dichloride yielded the same phosphinic acid, m.p. 77.580'," which was oxidized to a phosphonic acid. The melting point of this phosphonic acid could not be raised above 149-151 ' even after several recrystallizations from 6 N hydrochloric acid and was not depressed by admixture with either 3-chloro-4-tolylphosphonic acid or 4-chloro-3-tolylphosphonic acid. The infrared spectra of the latter two acids were, of course, quite similar, but each spectrum contained a number of distinct maxima which were absent in the spectrum of the other compound. The spectrum of the phosphonic acid obtained from either Friedel-Crafts reaction exhibited every peak present in the spectra of the two known acids. When a mixture of equal weights of the two known acids was recrystallized from 6 N hydrochloric acid, a material was obtained which melted a t 147-150' and whose infrared spectrum was virtually identical with that of the phosphonic acid obtained from the Friedel-Crafts reaction. This evidence strongly suggests that Melchiker's recorded melting point of 190' is in error and that the Friedel-Crafts reaction between o-chlorotoluene and phosphorus trichloride yields a mixture of the 3-chloro-4-tolyl and the 4-chloro-3-tolyl isomers. Since o-chlorotoluene can be isomerized by heating with aluminum chloride and hydrogen chloride,l 2 there was a possibility that the methyl and chloro groups in the phosphonic acid obtained by the Friedel-Crafts reaction niight no longer be ortho to one another. This possibility was effectively eliminated when we found that heating this phosphonic acid to 300' caused it to decompose to give pure o-chlorotoluene (identified by its infrared spectrum) and a residue of inorganic phosphate. The usefulness of thermal dephosphonation as an aid in elucidating the structures of arylphosphonic acids has been previously described.'* Experimentall' 3-Chloro4-tolylphosphonic Acid and Bis(3-chloro4-tolyl)phosphinic Acid.-3-Chloro-4-methylaniline (Eastman P6877) (10) E. L. Gefter, Zh. Obshch. R h i m . , 39, 1338 (1958); Chem. Abstr., 63, 19999 (1958). (11) Melchiker' reported m.p. 70° for his phosphinic acid. The melting point of authentic 3-chloro-4-tolylphosphinic acid is 97.5-98.5" : L. D. Quin and J. S. Humphrey, Jr., J. Am. Chem. Soc., 98, 4124 (1961). (12) J. F. Norris and H. S. Turner, ibid., 61, 2128 (1939). (13) L. D. Freedman, G . 0. Doak, and E. L. Petit, J. 070. Chem., 36, 140 (1960); C. E. Griffin and J. T. Brown, ibid., 36, 853 (1961). (14) Melting points were determined as previously described: cf. ref. 15. The carbon and hydrogen analyses were performed by Galbraith Laboratories. Ino., Knoxville, Tenn.
APRIL1965 was purified by vacuum distillation and then converted to the corresponding diazonium fluoroborate by Roe's procedure IIA.lB The dried diazonium salt was suspended in ethyl acetate and treated with phosphorus trichloride and cuprous bromide in the usual manner When the reaction mixture was steam distilled, an oil, consisting largely of phosphinic acid, separated in the distilling flask. The hot aqueous layer was decanted through a filter, and the oil remaining in the beaker was washed several times with boiling water. The crude phosphinic acid was then purified by reprecipitation from alkaline solution and subsequent recrystallization from 9570 ethanol. The yield of pure bis(3chloro-4-toly1)phosphinic acid was 3%, m.p. 189-192". Anal. Calcd. for C14H13C1202P: C1, 22.50; P , 9.83; neut. equiv., 315.1. Found: C1,22.26; P , 9.86; neut. equiv., 310.5. The phosphonic acid was isolated as its hemi-sodium salt16from the aqueous layer mentioned above, and the salt was converted to the free acid by recrystallization from a mixture of 1 vol. of 95% ethanol to 5 vol. of 6 N hydrochloric acid. The yield was 50%, m.p. 16C-162'. Anal. Calcd. for C7H&103P: C1, 17.16; P , 15.00; neut. equiv., 103.3. Found: C1,16.97; P , 14.50; neut.equiv., 104.4. 2-Nitro-5-chloro-l-tolylphosphonic Acid.17-3-Chloro-4-tolyl phosphonic acid (40.6 9.) was added in small portions to 185 ml. of stirred fuming nitric acid ( d 1.5) maintained a t 15-20'. Stirring was continued for 1 hr. after the addition was complete, and the reaction mixture was then poured onto 400 g. of cracked ice. The mixture was allowed to stand in the ice box overnight after which the solid was removed by filtration. The yield of crude mononitrated material was 36.5 g. (73y0),m.p. 169-184". After several recrystallizations from 3 N hydrochloric acid, the material melted with decomposition a t 223-227'. Anal. Calcd. for C7H7ClNOsP: C1, 14.09; N, 5.57; P , 12.31; neut. equiv., 125.8. Found: C1, 14.12; N, 5.54; P, 12.06; neut. equiv., 126.0. The structure of this substance was not established unequivocally, but it is probably 2-nitro-5-chloro-4-tolylphosphonic acid for the following reasons. (1) The compound does not form a water-insoluble magnesium salt either at room temperature or when heated. This behavior is characteristic of arylphosphonic acids containing bulky ortho substituents such as the nitro group.I8 (2) The nitration of 3-chloro-4-toluenesulfonic acid yields almost exclusively 2-nitro-5-chloro-4-toluenesulfonic acid.'@ Since the sulfo and the phosphono groups have similar electronic structures, it seems reasonable to assume that the nitration of 3-chloro-4-tolylphosphonic acid gives 2-nitro-5chloro-4-tolylphosphonic acid. 4-Chloro-3-tolylphosphonicAcid and Bis(4-chloro-3-tolyl)phosphinic Acid .4-Chloro-3-methylaniline ,Zo m.p . 85.5-86.5 O , was diazotized in fluoroboric acid in the usual manner.16 The cold reaction mixture was filtered by suction on a sintered-glass filter, and the solid diazonium fluoroborate was washed with a small amount of cold fluoroboric acid (40 ml./mole of amine) and then several times with copious quantities of ether. After being dried in a desiccator, the salt was converted to the corresponding phosphonic and phosphinic acids, which were separated by the procedure used for the 3-chloro-4-tolyl isomers. After several recrystallizations from aqueous alcohol, the yield of pure bis(4-chloro-3toly1)phosphinic acid was 3y0, m.p. 178-181". Anal. Calcd. for ClaH13C1202P: C, 53.36; H, 4.16. Found: C, 53.48; H, 4.30. The phosphonic acid was recrystallized from 6 AT hydrochloric acid; the yield was 1570, m.p. 158-161'. Anal. Calcd. for C?HsC1O3P:C, 40.70; H , 3.90. Found: C, 40.78; H, 4.01.
Acknowledgment.-The authors wish to acknowledge the technical assistance of Mrs. Joyce E. Carevic, Mr. Bobby R. Ezzell, and Mr Dick E. Hoskins. (15) G . 0. Doak and L. D. Freedmsn, 3 . A m . Chem. Soc., 73,5658 (1951). (16) A. Roe, O w . Reactions, 6 , 204 (1949). (17) This compound is mentioned in ref. 4 , b u t no information concerning ita structure is presented. (18) L. D. Freedman and G . 0. Doak. J. A m . Chem. Soc., 7 7 , 6221 (1955). (19) E. G. Turner and W. P. Wynee, J . Chem. Soc., 707 (1936). ( 2 0 ) J. P. Lambooy and E . E. Haley, J. A m . Chem. Soc., 74, 1087 (1952).
1265
NOTES Dialkyl Esters of Acylphosphonic Acids' K. DARRELL BERLIN,D. M. HELLWEGE, A N D RI. NAGABHUSHANAM~ Department of Chemistry, Oklahoma Slate University, Stillwater, Oklahoma Received October 23, 1964
During the course of studies involving the reaction of compounds containing adjacent carbonyl-phosphoryl groups3 we have synthesized several new alkyl acylphosphonates by a very convenient method. Early work by Arbuzov4and more recent data by Kabachnik5 demonstrate that acetyl chloride, ethyl chloroformate, phosgene, and benzoyl chloride react with phosphites to give phosphonates in a Michaelis-Arbuzov rearrangement.6 Acetic anhydride behaves similarly7 while ahaloacyl halides* react in a pat tern similar to t'he Perkow react'ion of a-halo ketones with trisubstit'uted phosphorus esters. Addition of acyl halide to phosphite required cooling as I t was found that high yields of dialkyl acylphosphonates could be realized if the reaction mixtures were permitted to stand at room temperature overnight before distillation was att'empted. In the case of the cyclohexyl derivatives, an additional 1-2 hr. at reflux was necessary to complete the process after the mixture had stood for a day. In all examples, the vacuum distillation proceeded very sinoothly to give colorless oils when this procedure was followed, Physical properties are recorded in Table I for all esters, and the 2,4-dinitrophenylhydrazonederivatives were analyzed for nitrogen and phosphorus. Infrared (Table 11) and n.m.r. data (Table 111) support the structures. As in the examples with dialkyl aroylphosphonates, the dialkyl acylphosphonates are hygroscopic and presumably undergo carbon-phosphorus bond cleavage. The surprisingly low frequency of absorption for the carbonyl function in these novel esters is siniilar to the situation reported with the aroylphosphonates. I t was suggested that one of the nonbonding orbitals on the oxygen atom of the phosphoryl group could orient properly for overlap with the p-orbital on the carbonyl carbon at,oni which could result in alteration of the force const,antof the carbonoxygen bond.3 Such a phenomenon could be operative in the compounds described herein. (1) We gratefully acknowledge support of t h e National Institutes of Health, GM-10367-03. (2) Postdoctorate, 1963-1965. (3) K. D. Berlin and H. A. Taylor, J. A m . Chem. Soc., 86, 3862 (1964). (4) A. E. Arbuzov and A. A. Dunin, J. Russ. Phys. Chem. Soc., 46, 295 (1914); Chem. Abstr., 8 , 2551 (1914); see also Ber., 60B,291 (1927); Chem. Abstr.. 21, 1627 (1927). (5) M. I. Kabachnik and P. A. Rossiiskaya, Bull. acad. sci. U R S S , Classe sei. chim., 597 (1945); Chem. Abstr., 41, 88 (1947); M. I. Kabachnik, P. A. Rossiiskaya, and E. S. Shepeleva, Bull. acad. sci. C R S S , Classe sci. c h i n . . 163 (1947); Chem. Abstr.. 42,4132 (1948); XI.I. Kabachnik and P. A. Rossiiskayo, Irv. Akad. Nauk S S S R , Old. K h i m . Nauk, 48 (1957); Chem. Abstr., 61, 10366 (1957); Bull. Acad. Sci. C S S R , Diu. Chem. S a . , 1398 (1958); Chem. Abstr., 68, 6988 (1959). (6) This reaction has been reviewed recently: R . G. Harvey and E . R. DeSombre, "Topics in Phosphorus Chemistry," M. Grayson and E. T. Griffith, Ed., Interscience Publishers, Inc.. New York. N. Y., 1964. (7) H. W. Coover, Jr., and J. B. Dickey, U. S. Patent 2,784,209 (1957); Chem. Abstr., 61, 13903 (1957); G . Kamai and V. A. Kukhtin, Zh. Obshch. K h i m . , 1 7 , 949 (1957); Chem. Abstr.. 62, 36658 (1958); V. A. Kukhtin, Dokl. Akad. Nauk S S S R , 121, 466 (1958); Chem. Abstr., 6 3 , 1105 (1959). ( 8 ) A. N. Pudovik and L. G. Biktimirova. Zh. Obsch. K h i m . , 27, 2104 (1957); Chem. Abstr., 62, 6156 (1958).
1266
NOTES
VOL. 30
TABLE I ALKYLACYLPHOSPHONATES AND DERIVATIVES 00
/I t
RC-P(0R')z Compd.
R
R'
2,4-DNP B.P., 'C. (mm.)
Yield,
%
M.p., OC.
%-
-Calcd., N
--Found, N
P
CzHa CH3 181-186 (3-4) 73.6 CZHK 105(7) 85.7 2 CzHs 93-94 14.97 8.28 15.37 3" (CHa)zCH C2Ha 99-100 (4) 85.5 126-127 4 (CHa)aC CZHK 97-100 (5) 65 140-141 13.89 7.68 13.44 CZHK 5 Cyclopropyl 119-120(5) 69 114-115 14.50 8.02 14.30 CzHa 120-124 (5) 6 Cyclobutyl 64.5 136-136.5 14.00 7.74 14.20 7 Cyclopentyl CZHK 121-126 (3) 81.5 123-125 13.52 7.48 13.32 CHa 98-99 ( 0 . 5 ) 8 Cyclohexyl 82 125-126 13.94 7.75 14.00 CZHK 9 Cyclohexyl 89-90 ( 0 . 2 ) 88 113.5-115 13.07 7.22 13.08 10 Cyclohexyl (CHa)zCH 100-102 (0.4) 83 84-86 12.43 6.83 12.28 a Both 1 and 3 have been reported but the yields were much lower and thus the esters are included here for completeness. McConnellandH. W.Coover, Jr., J . A m . Chem.Soc., 78,4450(1956): 1, b.p.62-65" (1.5mm.), and3, b.p. 92-97' (3.9 mm.).
%P
1"
TABLE I1 INFRARED SPECTRA OF DIALKYL ACYLPHOSPHONATES~ Compd.
c=o
P-0
1695 1257 2 1265 1695 3 1691 1265 4 1666 1250 1670 1260 5 1257 1691 6 1253 1690 7 1266 1695 8 1263 1695 9 1260 1695 10 All spectra were made of films of the esters on ride plates; values are in cm.-'. 1
As part of a study to evaluate P-H coupling constants in a variety of organophosphorus compounds, the n.m.r. spectra of the acylphosphonates are particularly interesting. I n the example of diethyl acetylphosphonate (1) the P-H coupling constant for the methyl group adjacent to the carbonyl function is 5 C.P.S. This is a marked drop in value compared with that of methyl esters of phosphorus acids such as in trimethyl phosphate ( J = 11.2 c.P.s.), dimethyl methylphosphonate (CH,, J = 17.5 c.P.s.; OCHa, J = 11 c.P.s.), trimethyl phosphite ( J = 10 c.P.s.), and methyl ~ ~ adiphenylph~sphinate~( J = 11 c . p . ~ . ) . 1 0 ,The hydrogens, as shown below, in several of the esters were
complex from coupling with vicinal protons, but the multiplets displayed shoulders in all cases suggesting a P-H coupling.l2 A variety of solvents did not provide for better separation suggesting that the coupling may (9) K. D. Berlin, T. H. Austin, and M . Nagabhushanam, t o be published. (10) All spectra were measured in carbon tetrachloride for this comparison. (11) Recent results b y Martin and co-worker6 demonstrate the variation in J when sharp differsncea exist in the electronegativity of atoms attached t o phosphorus; see G. Martin and A. Besnard, Compt. rend.. 96'7, 898 (1963). T w o interesting examples reported which might be compared to our work are (CHaN)aP-0 ( J = 11 c.p.8.) a n d (CHaS)aP+O ( J = 9.25 c.P.s.), both measured in cyclohexane, however. (12) Double resonance experiments with triethyl phosphite have shown must be nearly equal and the resulting quarteta t h a t J C H s - p a I and JCHZ-CHI overlap to give a five-line system; see J. D. Baldeschwieler and E. W. Randall. Chem. Rev., 68, 81 (19631, ref. 95.
7.41 7.95 7.82 7.26 7.75 7.24 6.80 See R. L.
TABLE I11 N.M.R.CHEMICAL SHIFTSA N D COUPLING CONSTANTS OF DIALKYL ACYLPHOSPHONATES
P-0-c
1021 1020 1022 1022 1022 1022 1019 1040 1026 990 sodium chlo-
8.10
00
I1 t
RC-P( OR'), Compd.
1 2
3 4
5c
6d 7a 8'
-RCHa
Parameters of structure unitsa r R' CHz CH CHs CHz
2.43d (5)b 1.05 t 2 . 8 0 q (7.5) (7) 1.14d (7)
-CH
1.37 t 4 . 2 2 q t (7.5) (7) 1.32 t 4 . 1 6 q t (7.5) (7) 3 . 1 0 m 1.35 t 4 . 1 6 q t (7) (7.5) (7) 1.33 t 4 . 1 4 q t (7.2) (7) 2 . 5 8 m 1.22 t 4.10 qt (6) (7.5) (7) 3.70m 1 . 3 3 t 4.19qt (7.2) (7) 1 . 3 0 t 4.05 qt (7.3) (7) 3.81 d (ll)b
1.35t 4.18qt (7.5) (7) 10 1.43 d 4.85m (6.5) (7.5) In CClr solution; &values downfield from TMS (J,c.P.s.); multiplicity of signals: d, doublet; t, triplet; m, multiplet; q, A quartet; qt, quintet; s, singlet. Jp-H coupling constants. Complex multiplet for ring protons was centered near 6 1.15. multiplets for ring protons visible at 6 2.2 and 4.2. e Complex multiplets for ring protons visible at 6 1.65 and 4.05. JAll cyclohexyl compounds had complex multiplets for ring protons centered at approximately 6 1.65. 9
be small. Paramagnetic shielding by carbonyl groups on a-protons is well known,I3 but the resultant influence by an adjacent phosphoryl group is not easily defined. Proton decoupling (b-protons from a-protons) should be instructive in this matter and will be reported at a later date.'* (13) L. M. Jackman "Applications of Nuclear Magnetic Resonance Spectroscopy in Organic Chemistry," Pergamon Press Inc., New York N. Y., 1959, Chapter 4. (14) A recent paper summarizes considerable d a t a on Pal-H spin-spin coupling constants in several organophosphorus compounds; see J. B. Hendrickson, M. L. Maddox. J. J. Sims. and H. D. Kaesz. Tetrahedron, 20, 449 (1964).
NOTES
APRIL1965 ExperimentaP Preparation of Diethyl Propiony1phosphonate.-The following general procedure was followed in the preparation of all esters and the corresponding 2,4-DNP derivatives. To triethyl phosphite (35.9 g., 0.216 mole), stirred in a 100-ml. round-bottom flask under anhydrous conditions and under nitrogen, was added dropwise 20.0 g. (0.216 mole) of distilled propionyl chloride.16 An exothermic reaction (temperature maintained below 5 0 " ) resulted during which bubbles of a gas were evolved. The mixture was allowed to stand for 1 day under nitrogen and was then vacuum distilled to give 36.0 g. (85.7%)of a colorless liquid. The 2,4-DNP was prepared by adding a few drops of the dialkyl acylphosphonate to about 10 ml. of a stock solution of 2,4-dinitrophenylhydrazine and filtering the crystals which formed in the solution after standing a few minutes. Recrystallization from methanol gave a yellow, fibrous mass of crystals, m.p. 93-94".
Acknowledgment.-The senior author acknowledges partial support by the Research Foundation, Oklahoma State University. (15) All melting points are corrected; a l l boiling points are uncorrected. T h e infrared spectra were obtained on a Beckman IR-5 as films on sodium chloride ceus. T h e n.m.r. spectra were obtained on avarian A-60 instrument with carbon tetrachloride as t h e solvent. Tetramethylsilane was used a s a n internal standard. We gratefully acknowledge t h e gift from Stauffer Chemical Co., Victor Chemical Division, of samples of t h e various phosphites needed in this work. Analyses were performed by Galbraith Laboratories. Copies of spectra may be obtained from the senior author. (16) T h e acid halides were purchased or prepared b y standard procedures and were distilled prior t o use.
A Convenient Synthesis of Esters of Diphenylphosphinic Acid. 111',2 K. DARRELL BERLIN,T. HOWARD AUSTIN, AND hf. NAGABHUSHANAM' Department of Chemistry, Oklahoma State University, Stillwater, Oklahoma Received October 23, 1964
Phosphinates are a somewhat rare class of organophosphorus esters which have been prepared by a variety of methods none of which has been systematic in a p p r ~ a c h . Recently, ~ it was demonstrated that diazo alkanes can be employed with a phosphorus acid to give good yields of esters.ss6 Phosphonites are not useful precursors since aryl halides are not effective in promoting a Michaelis-Arbuzov rearrangement .' We now report several new phosphinates of the type (CsH&P(O)OR which were synthesized by a general procedure involving a classic reaction between diphenylphosphinic chloride and the corresponding alcohol in the presence of t'riethylaniine in ether. After a short time a t reflux, the mixture was filtered and the resulting organic solvent was evaporated to give a solid. A slight excess of the alcohol removed (1) We gratefully acknowledge support by the Air Force Officeof Scientific Research, Grant AF AFOSR-132-65. (2) Preceding papers in t h e series: (a) K. D. Berlin, T. H. Austin, and K. L. Stone, J . A m . Chem. Soc., 86, 1787 (1964); and (b) K. D. Berlin and M. Nagabhushanam, Chem. Ind. (London), 974 (1964). (3) Postdoctorate, 1963-1965. (4) T h e meager d a t a on t h e subject has been covered in a review; see K. Sasee, "Methoden der organischen Chemie," Vol. 12, P a r t 1, E. Muller, Ed., Georg Thieme Verlag, S t u t t g a r t , Germany, 1963. (5) K. Issleib and H. M. Mobius, Ber., 94, 102 (1961). (6) B. E. Smith and A. Burger, J . A m . Chem. Soc., 7 5 , 5891 (1953). (7) A review is available: E. R. De Sombre and R. G. Harvey, "Topicsin Phosphorus Chemistry," M. Grayson and E. J. Griffith, Ed., Interscience Publishers, Inc., New York. N. Y., 1964.
1267
all of the acid chloride which, via hydrolysis, could lead to diphenylphosphinic acid-a difficult contaminant to remove. The method is superior both with respect to simplicity of isolation of relatively pure products and yields of esters. An exception, t-butyl diphenylphosph,inate, was obtained from condensation of potassium t-butoxide with diphenylphosphinic chloride since the regular procedure gave only complex mixtures. A list of the esters is found in Table I. As expected infrared analyses revealed absorption by the phosphoryl group near 8.2 p while the P-0-C groups displayed peaks from 9.6 to 10.85p.s,g In addition the P-phenyl band was observed a t 6.93 k . Examination of the esters by n.m.r. showed complex multiplets, particularly for the a-protons, due to H1-Psl coupling. With isopropyl diphenylphosphinate, a symmetrical eight-line pattern was visible (each line was split again) for the tertiary hydrogen. This possibly, CH3
\ /
0
t
CH-OP( CaH6)z
CHa
may occur from overlap of two heptets, one appearing as a result of CH-CH3 coupling and the other from and J C H - ~ S I H-P3I coupling assuming the JCH-CH, constants were close in value. I t has been demonstrated that J C H ~and -PS JCH,-CH, I are equal or nearly so in triethyl phosphite and give rise to a symmetrical five-line grouping without further splitting.lO*llAdditional n.m.r. data is provided in Table 11. Experimental1* The following procedure is typical of that used for the preparation of the title compounds. Cyclohexyl Dipheny1phosphinate.-To a mixture of 40.04 g. (0.40 mole) of cyclohexanol and 50.57 g. (0.50 mole) of triethylamine in 350 ml. of anhydrous ether under nitrogen was added 78.80 g. (0.33 mole) of diphenylphosphinic chloride1ain 100 ml. of anhydrous ether with rapid stirring over a period of 0.5 hr. The reaction mixture was heated a t reflux for 1 hr. After filtering the hydrochloride salt,14 the mixture was concentrated to give a solid which, upon recrystallization from benzene-heptane, gave 78.47 g. (78 57,)of cyclohexyl diphenylphosphinate. &Butyl Dipheny1phosphinate.-To a rapidly stirred solution of potassium t-butoxide, prepared from 11.73 g. (0.3 g.-atom) of potassium and 300 ml. of anhydrous t-butyl alcohol, was added, under nitrogen, 43.13 g. (0.18 mole) of diphenylphosphinic chloride in 100 ml. of benzene over a period of 0.5 hr. The mixture (8) K. L. Paciorek,lnorg. Chem., 3 , 96 (1964). (9) L. C. Thomas and R . A. Chittenden, Chem. Ind. (London), 1913 (1961). (10) J. B. Hendrickson, M. L. Maddox, J. J. Sims, and H. D. Kaesz, Tetrahedron, 20, 449 (1964). (11) See J. D. Baldeschwieler and E. W. Randall, Chem. Rev., 63, 81 (19631, ref. 95. (12) All melting points are corrected and all boiling points are uncorrected. T h e microanalyses were performed b y Galbraith Laboratories, Inc.. Knoxville, Tenn. T h e infrared spectra were determined on a Beckman I R 5 spectrophotometer. A Varian Model A-60 high-resolution spectrometer fitted with a field-eensing stabilizer "Super Stabilizer" was used t o determine the n.m.r. spectra. (13) This compound was prepared from oxidation of diphenylphosphinous chloride which was obtained from Stauffer Chemical Co., Victor Chemical Division. We express our thanks t o W. D. E b e r t and E . L. Kubiak for generous samples of this material. (14) An alternative procedure of using a slight excess of diphenylphosphinie chloride was attempted. This necessitated a wash with sodium bicarbonate a t the end of the heating period t o remove diphenylphosphinic acid and t o destroy t h e amine salt. However, yields of t h e esters approximated those obtained by the method given. I n t h e cases of 7 and 8 , phenol and a-naphthol, respectively, were contaminants and could be removed by sodium hydroxide without saponifying 7 or 8 .
1268
NOTES
VOL. 30
TABLE I 0
t (CsHs)zPOR Yield,
%
R
Compd.
M.p. or b.p., '(2. (mm.)
-----Amax,
P-0-c
c
P+O
--
%--
----Calcd.,
p-
H
Found, %---
C
P
H
P
CHI 92.5 178(2.4)5 9.6 8.1 (CHshCHb 60.2 97-99 10.1 8.2 3 (CH3)aC 80.1 111-112 10.2 8.2 70.06 6.98 11.29 70.28 6.94 11.34 4 CeHii 78.5 120-121 10.1 8.2 71.98 7 . 0 5 10.31 71.86 6.80 10.36 5 trans-472.0 132-133 9.9 8.2 74.15 8.15 8 . 7 0 73.94 8.18 8.71 (CH3)aCCeKoC 6 CHFCHCHZ 75.0 162-165(5) 9.8, 10.05d 8 . 1 5 69.77 5.81 12.02 69.65 5.72 12.03 97.9 135-136 7 CsH5" 10.85' 8.2 8 a-CIOH7 92.7 122-124 10.95' 8.1 76.74 4.98 9.00 76.45 4.91 9.10 a This crystalline solid liquifies upon standing in air with no essential change in its infrared spectrum. A. E. Arbuzov, J . Russ. Phys. Chem. SOC.,42, 395 (1910); Chem. Zenlr., 81 (111, 453 (1910). The trans isomer is tentatively assigned for this compound in analogy with the corresponding acetates; see R. U. Lemieux, R. K. Kullnig, H. J. Bernstein, and W. G. Schneider, J . A m . Chew SOC., 80, 6098 (1958). I t was observed that the broad peaks for H-1 in the trans alcohol and acetate had half-widths of 22 C.P.S. while the cis isomer and its acetat8ehad a sharp signal with a half-width of 7 c.p.s. In 5 , H-1 appeared as a broad signal centered a t 6 4.3 with a half-width of 22 c.p.s. a t the following settings on a near-saturated solution in CDC13: filter band width, 4; radio frequency field, 0.2; sweep time, 250sec.; sweep width, 500; spectrum amplitude, 1.6. Two bands of equal intensity. e A. Michaelis and A. Link, Ann., 207, 193 (1881). 'See ref. 8 and 9. 12
TABLE I1 CHEMICAL SHIFT P A R l M E T E R S FOR DIPHENYL ALKYLPHOSPHINATES~ Compd.
-CHI
>CHz
H-C f
Aryl-H
7.25-8.20 m 4.640 7.28-8.25 m 7.25-8.05 m 1.57m 4.45m 7,28-8.25 m 0.75s 1.53m 4.38m 7.20-8.12 m 6d 4.40m' 7.20-8.05 m 7 6.97-8.28 m 8 7.08-8.45 m a Given in &values downfield from internal TMS. Except for 6 in which deuteriochloroform was the solvent, all other determinat,ions were made in methylene chloride as solvent. Multiplicity of signals: d, doublet; s, singlet; m, multiplet; 0 , octet. JP-= = 11.0 C.P.S. J = 6.0 c.p.5. Multiplets as expected for the alkyl group are observed centered a t 6 5.12, 5.39, and 6.0 which is similar to the pattern given by alkyl alcohol (see NMR Spectra Catalog I, Varian Associates, Palo Alto, Calif., 1962, spectrum 34). e This is the CH2 group attached to the oxygen atom. 1 2 3 4 5
3.70db 1.28d" 1.45s
was heated a t reflux for 1.5 hr. and, after washing with 15% aqueous ammonium chloride solution, was dried over magnesium sulfate and concentrated. A solid was isolated which was recrystallized from benzene-heptane; yield 40.05 g. (80.1%).
Acknowledgment.-The senior author gratefully acknowledges partial support by the Research Foundation, Oklahoma State University.
compounds inight be made by the interaction of sodium diethyl N-alkoxyphosphoramidates and carbon dioxide. Wadsworth and Emmons have reported that sodium diethyl N-alkylphosphoramidates and carbon dioxide produce alkyl isocyanates.2 Wadsworth and E ninons2reported the preparation of diethyl N-methoxyphosphoramide. By utilization of a similar synthesis diethyl N-ethoxyphosphoraniide has now been made. However, reaction of carbon dioxide with sodiuin diethyl N-alkoxyphosphorainidates did not yield N-alkoxyisocyanates but their trimers, 1,3,5trialkoxyisocyanuric acids, (ROXCO),. 1,3,5-Tribenzyloxyisocyanuric acid and several substituted benzyl derivatives have been made by JIcKay, et u1.,1c by the treatinent of the corresponding benzyloxyamine hydrochlorides with phosgene followed by treatment of the resulting intermediate with triethylamine. It seems rather probable that the S-alkoxyisocyanic acids were formed first in the reaction between carbon dioxide and sodium diethyl N-alkoxyphosphorainidates and that the monomer then polymerized in the presence of the anions present. Trimerization of isocyanates is well known.a Both 1,3,5-trimethoxy- and 1,3,5-triethoxyisocyariuric acid were relatively stable, crystalline compounds. 0
0
t -
3(CzHsO)zP-N-
Reactions of Diethyl N-Alkoxyphosphoramidate Anions with Carbon Dioxide and Carbon Disulfide
OR
Cobb Chemical Laboratory, Uniuersity of Virginia, Charlottesville, Virginia Received November 18, 1964
For over 60 years chemical investigators have tried, unsuccessfully, to synthesize N-alkoxyisocyanic acids, RONCO.' It seemed possible to us that these novel
-t
[3(C&LO)d'-YORI t
--c
c=o -0 0
t
-
3(CzHsO)zP-O RANDOLPH T. MAJORA N D ROBERT J. HEDRICK
+ 3 COz
OR
/
+ 3[RONCO]
o=c ",c=o I
L RON,
I
,NOR
! 0
(1) (a) L. W. Jones, A m . Chem. J . . 30, 1 (1898); (b) L. W. Jones and L. Neuffer. J . A m . Chem. Soc., 89, 652 (1917); ( c ) A. F. RlcKay, D. L. Garmaise, G. Y. Paris, and S. Gelblum, Can. J. Chem., 81, 343 (1960). (2) W. €3. Wadsworth, Jr., and \V, D. Emmons, J . A m . Chem. Soc., 84, 1316 (1962); J . O w . Chem.. 19, 2816 (1964). (3) (a) E . H. Rodd, "Chemistry of Carbon Compounds," Vol. I B , Elsevier Publiahing Co., New York. N . Y., 1952, p . 939; (b) V. E . Mhsshoua. W. Sweeney, and R. F. Tieta, J . A m . Chem. Soe., 81, 866 (1960).
NOTES
APRIL 1965 Their infrared and n.m.r. spectra were in harmony with the structure shown abovee4 Fusion of 1,3,5-triethoxyisocyanuric acid with potassium hydroxide gave N-ethoxyamine. Wadsworth and Emmons have also found that interaction of a sodium diethyl N-alkylphosphoramidate and carbon disulfide gave alkyl isothiocyanates.2 However, in our laboratory, when sodium diethyl N-ethoxyphosphoramidate was treated with carbon disulfide no evidence of the presence of N-ethoxyisothiocyanic acid in the reaction product was obtained. However, the distillate from the reaction product contained ethanol. The addition of water to the reaction product gave free sulfur and thiocyanate ions. A likely pathway of this reaction is given below. 0
0
t -
(CzHb0)ZP-N-OCzHs
t + [( CzHs0)zP-NOCzHs + I CS2
-&=s
0
t
[(CzHbO)zP-N=C=S]
+ S +-OCzHs
J. H * 0 0
t
(CzH~O)~P-OH
+ HSCN
When sodium diethyl N-ethoxyphosphoramidate was treated with 3-pentanone, 0-ethyldiethylketoxime was obtained. This finding is similar to Wadsworth and 0
t -
(CzH5O)z P-NOCzHs
+ (C2HS)zCO +(CzHs)zC=NOCzHs
Emmons' observation that the interaction of benzaldehyde and sodium diethyl N-methoxyphosphoramidate gave O-methylbenzaldoxime.2 Experimental Diethyl N-A1koxyphosphoramides.-To 0.06 mole of the Nalkoxyamine in 100 ml. of dry ether was added 5.2 g. (0.03 mole) of diethyl phosphorochloridate, prepared by the method of McCombie, Saunders, and Stacey.6 After the mixture had been stirred for 1 hr. at room temperature, the precipitated N-alkoxtimine hydrochloride was removed by filtration. The ether was wmoved by distillation and then the residual oil was fractionated zn vacuo. Diethyl N-methoxyphosphoramide2 was obtained in a yield of SO%, b.p. 107" (0.5 mm.). The diethyl N-ethoxyphosphoramide, b.p. 110-112' (0.5 mm.), which was obtained was a colorless liquid. The neat liquid showed strong infrared absorption bands a t 3.2, 3.45, 7.2, and 8.5 and broad, intense bands a t 9.2-10.5 p . This infrared spectrum was very similar to that of diethyl N-methoxyphosphoramide. The n.m.r. spectrum in carbon tetrachloride showed overlapping triplets, the one a t 7 8.6 having twice the area of the one at 8.8. Overlapping quartets were seen at'+ 5.9 and 6.2. The former had twice the area of the latter.6 The compound exploded when an elemental analysis was attempted. Sodium Diethyl N-A1koxyphosphoramidates.-A solution of 0.05 mole of freshly distilled diethyl N-alkoxyphosphoramide in 20 ml. of dimethoxyethane was added to 2.4 g. of sodium hydride (50% in mineral oil) suspended in 80 ml. of ( 4 ) L. Kakovic and K. W. F. Kohlrausch, (1944). (5) H. McCombie,
Z. P h y s i k . Chem., 199, 188
B. C. Saunders, and G . J. Stacey. J . Chem. SOC.,380 (1945). (6) A11 infrared spectra were taken on a Perkin-Elmer Infracord and all n.m.r. spectra were taken on n Varian A-60.
1269
dimethoxyethane. The mixture was then stirred a t room temperature for 1 hr. 1,3,5-TriethoxyisocyanuricAcid .-The above suspension of sodium diethyl N-ethoxyphosphoramidate from 0.5 mole of diethyl N-ethoxyphosphoramide was cooled to 3' and stirred. Carbon dioxide was bubbled through the suspension; a clear solution WBS obtained. The temperature of the reaction mixture was raised to that of the room and the mixture was stirred for an additional hour. When the temperature was raised to 65' a gummy precipitate formed. The mixture was cooled and filtered. The precipitate was a hygroscopic, water-soluble gum which did not melt. It contained no nitrogen and generally behaved like the inorganic phosphorus-containing salt obtained by Wadsworth and Emmons.2 Dimethoxyethane was removed from the filtrate by distillation. The residue was an oil which solidified when it was cooled. Recrystallization from benzene-petroleum ether gave 1.2 g. (28% of t h e x y ) of colorless crystals, m.p. 145'. Infrared KBr showed important bands at 3.4 (m), 5.7 (s), 7.1 (s), 7.3 (s), 9.7 (s), 9.9 (s), and 14.25 (s) p . The n.m.r. spectrum (in CDCla) showed a triplet a t 7 8.6 and a quartet a t 5.7 with relative are39 of 3:2. The solid was unaffected by boiling 6 N hydrochloric acid. Anal. Calcd. for CpHl6N&: C, 41.37; H , 5.80; N, 16.09; mol. wt., 261. Found: C, 41.69; H , 5.98; N, 16.58; mol. wt. (micro-Rast,? average of two determinations), 251. A small amount of the compound was mixed with solid potassium hydroxide in a flask fitted with an outlet tube leading to a solution of dilute hydrochloric acid. The mixture waa heated with a small flame. When no further reaction was apparent, the hydrochloric acid solution was evaporated under reduced pressure. A white solid remained as a residue which was shown to be N-ethoxyamine hydrochloride by comparison of its infrared spectrum with that of authentic N-ethoxyamine hydrochloride.8 1,3,5-Trimethoxyisocyanuric Acid.-Carbon dioxide was introduced a t room temperature into the above suspension of sodium diethyl N-methoxyphosphoramidate from 0.05 mole of diethyl N-methoxyphosphoramide. The reaction mixture was heated to 80" and stirred for 1 hr. Then the mixture was cooled and the gummy precipitate was filtered. The filtrate was distilled. The solid residue was recrystallized from carbon tetrachloridepetroleum ether, m.p. 197-198'. The infrared spectrum was similar to that of 1,3,5-triethoxyisocyanuric acid, showing important bands (KBr) a t 3.4 (m), 5.7 (s), 7.2 (s), 9.7 (s), and 14.2 (s) p . The n.m.r. spectrum (in deuterated dimethyl sulfoxide) showed only a singlet a t 7 6.2. Anal. Calcd. for Ce&N3O6: C, 33.44; H , 4.20; N, 20.37. Found: C, 32.88; H , 4.14; N,19.88. Reaction of Sodium Diethyl N-Ethoxyphosphoramidate with Carbon Disulfide.-A solution of 9.8 g. (0.05 mole) of diethyl N-ethoxyphosphoramide in 20 ml. of dimethoxyethane was slowly added to a suspension of 2.4 g. of 50% sodium hydride in dimethoxyethane. After the mixture had been stirred for 90 min. a t room temperature, 3.7 g. (0.047 mole) of carbon disulfide in 10 ml. of dimethoxyethane was added while the mixture was cooled in an ice bath. The clear solution which was obtained was warmed on a steam bath. No volatile material was collected in a Dry Ice-acetone cooled receiving flask which was connected with the flask containing the reaction mixture. Then, the dimethoxyethane was removed by distillation a t reduced pressure. The distillate gave a positive test for ethanol by Feigl's test,g whereas dimethoxyethane itself gave a negative test. The residue from the distillation was a thick, light red suspension. Water was added to this residue and then the aqueous mixture was filtered. The filtrate gave a positive test for thiocyanate ion.10 The precipitate was a yellow solid which was readily soluble in carbon disulfide. Its infrared spectrum was identical with that of elemental sulfur. 0-Ethyldiethy1ketoxime.-A solution of 5.8 g. (0.03 mole) of diethyl N-ethoxyphosphoramide in 50 ml. of dry dimethoxyethane was slowly added, with stirring, to a slurry of 1.7 g. (7) E. J. Cowles and M . T. Pike, J . Chem. Educ., 40, 422 (1963). (8) L. W. Jones, Am. Chem. J . , 3 0 , 46 (1898). (9) F. Feigl, "Spot Tests in Organic Analysis," Elsevier Publishing Go.. New York, N . Y . , 1960, p. 358. (10) F. Feigl, "Spot Teats." Vol. 1, R. E. Oesper, Transl., Elsevier Publishing Co., New York, N. Y . , 1954, p. 340
1270
VOL. 30
NOTES
(0.3 mole) of sodium hydride in 50 ml. of dimethoxyethane. To the suspension which was obtained was added 3.0 g. (0.036 mole) of 3-pentanone. The mixture was then stirred at room temperature for 30 min. and cooled, and the sticky precipitate was filtered and washed with dimethoxyethane. The filtrate was then fractionated. In addition to dimethoxyethane, a small amount of 0-ethyldiethylketoxime was obtained, b.p. 131-1340.11 The infrared spectrum of this liquid was identical with that of authentic 0-ethyldiethylketoxime.
Acknowledgment.-We are indebted to Merck and Company, Inc., for a grant which supported part of this work. (11) A. I. Vogel, W. T. Cresswell, G. H. Jeffery. and J. Leicester, J . Chem. SOC.,514 (1952).
for the reduction products was confirmed by elemental analysis, infrared spectra, and n.m.r. spectra. Reduction of N-ethoxyphthalimide with the weaker reducing agent, sodium borohydride, did not yield Nethoxyisoindoline but rather 2-ethoxy-3-hydroxyphthalimidine (11). This structure for the reduction product was demonstrated by elemental analyses, infrared spectra, and n.m.r. spectra. Horii, Iwata, and Tamura have shown that the analogous compounds, N-alkylphthaliniides, were reduced by sodium borohydride to 2-alkyl-3-hydro~yphthalimidine.~ However, reduction of IT-methoxyphthalimide with sodium borohydride did not yield a phthalimidine derivative. The only organic product which was isolated was phthalide (111). The same compound was obtained by Horii, Iwata, and Tamura in varying
Reduc tion of N-Alkoxyphthalimides
0
RANDOLPH T. MAJORA N D ROBERTJ. HEDRICK Department of Chemistry, University of Virginia, Charlottesville, Virginia Received November 17, 1964
Orndorff and Pratt have reported that the interaction of phthalic anhydride and hydroxylamine Igave two compounds, one white and the other yellow.' These authors were unable to assign definite structures to these substances. Pratt and Gibbs later made the methyl ethers of the white compound and of the yellow compound but also were uncertain as to the structures of these compounds. More recently, Roderick and Brown found that there was no detectable difference between the infrared spectra of the white and yellow forms of Orndorff and Pratt's products.' Also, they found that the yellow color of the "yellow form" was due to a slight amount of impurity that could be removed by several methods. They concluded that both socalled forms were N-hydroxyphthalimide. The findings of Kuhler and Wegler4 confirmed the conclusions of Roderick and Brown. I t occurred to the present authors that a new class of isoindoline derivatives, N-alkoxyisoindolines (I), might be made by the reduction of N-alkoxyphthalimides, provided that the alkoxy group was not removed during the reduction. The methyl and ethyl ethers of N-hydroxyphthalimide were prepared by the interaction of an aqueous solution of sodium bicarbonate and the appropriate alkyl sulfate with N-hydroxyphthalimide. Hydrolysis of the resulting N-alkoxyphthalimide with boiling dilute hydrochloric acid gave alkoxyamine hydrochloride~.~ N-Alkoxyisoindolines (I) were indeed obtained on the reduction of the methyl and ethyl ethers of Nhydroxyphthaliniide with lithium aluminum hydride. The correctness of the N-alkoxyisoindoline structure (1) W . R. Orndorff and D. S. P r a t t , A m . Chem. J., 47, 89 (1912). (2) D. S . P r a t t and H. D. Gibbs, Phillzppzne J . Scz.. SA, 165 (1913). (3) W. R. Roderick and W . G . Brown, J . A m . Chem. Soc., 79, 5196 (1957). ( 4 ) E . Kiihler and R. Wegler, Ann., 616, 183 (1958). (5) J . Petracaek, Ber., 16, 827 (1883).
0
I
TI
m
amounts by the reduction of N-alkylphthalimides.6 The formation of phthalide on reduction of N-methoxyphthalimide may be accounted for in a way analogous to that suggested by Horii, Iwata, and Tamura.6 It is suggested that the N-methoxyisoindoline which was formed, first was reduced further to a borohydride complex of o-hydroxymethyl-X-methoxybenzaniide which was later hydrolyzed to phthalide. It would appear that in the case of N-ethoxyphthalimide that the more bulky N-ethoxy group interfered with a similar reduction to o-hydroxymethyl-N-ethoxybenzamide. Experimental'
N-Alkoxyphtha1imides.-N-Hydroxyphthalimide (0.64 mole) was dissolved in a solution of 120 g. of sodium bicarbonate in 300 ml. of water. To this solution was slowly added 0.7 mole of dialkyl sulfate with stirring. In the case of diethyl sulfate the addition took place a t room temperature for 8 hr. while in the case of dimethyl sulfate the addition took place a t approximately 15' for 90 min. The precipitate of N-alkoxyphthalimide was filtered and recrystallized from 95% ethanol and then from 1butanol; yield, 60-64%. N-Methoxyphthalimide .-The infrared spectrum (KBr) showed important bands a t 3.4 (w), 5.8 (a), 6.8 (m), 7.2 ( m ) , 8.4 (a), 8.8 (a), 9.2 (m), 9.8 ( s ) , 10.0 ( a ) , 11.0 ( M ) , 11.3 (a), 12.6 ( m ) , and 14.2 ( s ) p ; m.p. 133°.2 Anal. Calcd. for C9H,NOa: C, 61.0; H, 3.98. Found: C, 61.37; H,4.11. N-Ethoxyphthalimide .-The infrared spectrum (KBr) showed bands a t 3.4 (w), 5.8 (a), 6.8 ( m ) , 7.1 (m), 7.3 (m), 8.4 (a), 8.9 (a), 9.2 (m), 9.8 (a), 10.2 ( s ) , 11.3 (s), 12.6 ( m ) , and 14.2 ( s ) p ; m.p.97-98".' Alkoxyammonium Chlorides.-The N-alkoxyphthalimides were refluxed with an excess of 6 N hydrochloric acid for 2 hr. On cooling, phthalic acid precipitated. The filtrate from this mixture was evaporated under reduced pressure. The residue was recrystallized from ethanol-ether; yield, 56-600/0. Meth( 6 ) Z. Horii, C. Iwata, and Y. Tamura, J . O w . Chem., 2 6 , 2273 (1961). (7) All infrared spectra reported were determined on a Perkin-Elmer Infracord. N.m.r. spectra were obtained on a Varian A-60 instrument.
oxyamine hydrochloride had m.p. 150-151°6; ethoxyamine hydrochloride had m.p. 129-131".* N-Alkoxyisoindo1ines.-Over a period of 90 min. a solution of 0.05 mole of N-alkoxyphthalimide in dry ether was slowly added with stirring to a slurry of 0.13 mole of lithium aluminum hydride in ether. After the mixture had been stirred a t room temperature for 8 hr., the reaction mixture was cooled and the complex present was decomposed by successive additions dropwise of 6.5 ml. of water, 4.5 ml. of 20% potassium hydroxide, and 10 ml. of water. The inorganic salts were removed by filtration. The filtrate was dried with magnesium sulfate and then fractionated. N-Ethoxyisoindoline Picrate .-A colorless liquid, b.p. 145152' ( 5 mm.), was obtained which became dark red on standing. The infrared spectrum of the freshly distilled, colorless liquid (neat) showed important absorption bands a t 3.5 (m), 6.8 ( m ) , 7.3 ( m ) , 9.5 (s), 11.2 (m), and 13.4 ( 8 ) p and no carbonyl absorption. A few drops of the freshly distilled liquid was added to a saturated solution of picric acid in 95% ethanol. The mixture was warmed over steam for a few minutes and then cooled. The yellow crystalline precipitate was filtered and recrystallized from ethanol, m.p. l i 0 " dec. The n.m.r. spectrum of this picrate (in deuterated dimethyl sulfoxide) showed a singlet a t 7 1.7 with a relative area of 2, a singlet a t 2.7 with an area of 4, a singlet a t 5.35 with an area of 4, a quartet a t 6.0 with an area of 2, and a triplet a t 9.0 with an area of 3. Anal. Calcd. for CI6Hl6N4O8:C, 48.98; H , 4.11; N, 14.27. Found: '2,4835; H,4.38; N, 14.78. N-Methoxyisoindoline Picrate.-Fractionation gave a light yellow liquid, b.p. 13.5-140" (15 mm.), which became colorless on repeated distillations; yield, 72% of theory. On standing in the cold, the color quickly changed to dark red. The infrared spectrum (neat) had bands a t 3.4 (s), 3.55 (m), 9.5 (a), 6.8 ( m ) , and 13.4 (8)p but no carbonyl bands. The pirrate was prepared by the addition of a few drops of the freshly distilled liquid to a saturated solution of picric acid in 95% ethanol. After the solution had been warmed, it was cooled and a yellow crystalline precipitate appeared. This was filtered and recrystallized from ethanol, m.p. 104-105'. Anal. Calcd. for C1bH14N408: C, 47.63; H, 3.73; N, 14.80. Found: C,47.66; H,3.97; N, 14.61. 2-Ethoxy-3-hydroxyphthalimidine.-To 10.0 g. (0.05 mole) of N-ethoxyphthalimide suspended in 90% methanol was added a solution of 3.7 g. of sodium borohydride in methanol over a period of 1 hr. After the mixture had been stirred a t room temperature for 18 hr., 5 ml. of glacial acetic acid was added. The methanol was partially removed by evaporation under reduced pressure. Water was added; the white solid which formed was filtered and then recrystallized from water: m.p. 120-121"; yield, 62y0 of theory. Characteristic infrared absorption bands (KBr) were found a t 3.2 (a), 3.5 (m), 6.0 (a), 6.8 ( m ) , 9.35 (a), and 13.3 (8) p . The n.m.r. spectrum (in CDCl,) showed a singlet a t T 2.0 (area 4), a doublet a t 3.7 (area l ) , a doublet a t 6.45 (area I), a quartet a t 5.5 (area 2), and a triplet a t 8.6 (area 3). Anal. Calcd. for CloHllNOa: C, 62.15; H, 5.74; N, 7.25. Found: C, 62.12; H , 5.71; N, 7.21. Phtha1ide.-To 3.9 g. (0.022 mole) of N-methoxyphthalimide suspended in 90% ethanol was added a solution of 1.9 g. (0.05 mole) of sodium borohydride in 100 ml. of ethanol over a period of 1 hr. The mixture was then stirred a t room temperature for 6 hr. The bases were neutralized with glacial acetic acid. The ethanol was partially removed by evaporation under reduced pressure. Water precipitated a white solid which was recrystallized from water: m.p. 73-74'; lit.g m.p. 73'; yield, 67% of theory. The infrared absorption spectrum (KBr) of this solid showed characteristic bands a t 3.5 (w), 5.7 (a), 6.8 (m), 6.9 (m), 7.2 (m), 7.5 (m), 7.7 (m), 8.1 (m), 9.4 (s), 9.85 (a), and 13.3 +
(9)
1271
NOTES
APRIL1965
cc.
Anal. Calcd. for C8HB02:C , 71.61; H, 4.52. Found: C, 71.15; H , 4.44.
Acknowledgment.-We are indebted to Merck and Company, Inc., for a grant in support of this research. (8) L. W. Jones. A m . Chem. J . , 30, 1 (1898). (9) J. Wialicenus, Ber., 17, 2178 (1884).
An Improved Method for the Preparation of Volatile Epoxides DANIELJ. PASTOAND CHARLES C. C U M B O ~ Department of Chemistry, University of Notre Dame, Notre Dame, Indiana Received November 9, 1964
The synthesis of the isomeric 2-butene oxides has been reported previously2,3and' involves the conversion of cis- and trans-2-butene to the corresponding epoxides via the halohydrins. The halohydrins are isolated and purified before treatment with strong aqueous base to give the epoxides. The over-all yields average 4448%. The first one-step preparation of the isomeric 2butene oxides was reported by Eliel and Delmonte.' The trans epoxide was prepared by treating the monotosylate (not isolated) of meso-2,3-butanediol with strong aqueous base. The reported yield of epoxide in this procedure is 28%. Various investigators who have used the 2-butene oxides in research have prepared the compounds by the methods outlined above. It is of interest to note that no direct oxidations of the olefins with peracids seem to have been reported. We have observed that direct epoxidation of cis- or trans-2-butene with m-chloroperbenzoic acid in dioxane gives the corresponding cisand trans-2-butene oxides in a high state of stereochemical purity and in good yield (5240%). The experimental procedure is much simpler in that the epoxide is distilled directly from the reaction mixture without prior removal of the m-chlorobenzoic acid. Under these conditions the epoxides are stable in the presence of the carboxylic acid. The general method also appears to be applicable to the preparation of higher boiling epoxides. Treatment of 1-hexene with m-chloroperbenzoic acid in diglyme (b.p. 162") followed by direct distillation of the l-hexene oxide at 116-119" gives the epoxide in 60% yield. The greater simplicity in the experimental procedure used and the saving in time and effort makes this procedure very attractive for the preparation of volatile epoxides. Experimental Preparation of cis- and trans-&Butene Oxides.-In a threenecked 1000-ml. round-bottom flask equipped with a delivery tube, magnetic stirrer, and Dry Iceacetone reflux condenser was placed 500 ml. of anhydrous dioxane and 65.3 g. (0.322 mole) of m-chloroperbenzoic acid (FMC Corp., 85% minimum purity). After dissolution of the acid, the contents of the flask was cooled to 0"; 18.1 g. (0.322 mole) of cis- or trans-2-butene (Matheson Coleman and Bell, 99.0% minimum purity) waa added via the delivery tube. The contents of the flask was stirred for 10 hr. under the Dry Ice-acetone condenser. A t the end of this period of time the butene stopped refluxing. The flask was stoppered and placed in a refrigerator overnight. The mixture was subjected to distillation, collecting the fraction up to 100'. This fraction was fractionated through a 2-ft. helices-packed column giving a 52-60% yield of cis-2-butene oxide, b.p. 58.0-59.0' (748 mm.), and trans-2-butene oxide, b.p. 52.0-53.0' (748 mm.). (1) National Institutes of Health Predoctoral Fellow. ( 2 ) C . E. Wilson and H. J . Lucas, J . Am. Chem. Soc., 58, 2397 (1936). (3) S. Winstein and H. J. Lucas, ibrd., 81, 1576 (1939). (4) E. L. Eliel and D. W. Delmonte, J . Org. Chem., Pi, 596 (1956).
VOL. 30
NOTES
1272
The stereochemical purity of the two epoxides was shown to be greater than 99.5% by gas-liquid chromatography on a 30-ft. 20% Carbowax 20M on Chromosorb P column a t 150'. The retention of the cis epoxide was 10.0 min. and that of the trans epoxide, 8.7 min. Preparation of 1-Hexene Oxide.--A mixture of 24.4 g. (0.119 mole) of m-chloroperbenxoic acid (85%) and 10.0 g. (0.119 mole) of I-hexene in 300 ml. of anhydrous diglyme was allowed to stand 24 hr. in a refrigerator. The mixture was subjected to distillation collecting the fraction up to 162". This fraction was redistilled through a 2-ft. helices-packed column giving 7.05 g. (60%) of 1-hexene oxide, b.p. 116-119', n Z o1.4051 ~ (lit.5 b.p. 117-119", nZoD1.4060).
Ia, A4-3-keto 1
1
Ira, I1 A4-3-keto I
OH-
Acknowledgment.-The authors wish to thank the National Institutes of Health for a predoctoral fellowship to C. C. C. (No. 1-F1-GM-21, 333-01). (5) W. D. Emmons and A. S. Pagaro, J . A m . Chem. SOC., 7 7 , 89 (1955).
The Decarboxylation of 3-Carboxy-2-isoxazolines G. W. MOERSCH, E. L. WITTLE,AND W. A. NEUELIS
VI1 VIIa, A 4 -3-keto
I11 IIIa. A 4 -3-keto /OH-
Parke, Davis and Company, Research Laboratorks, Ann Arbor, Michigan Received June 4, 1964
I n the course of preparing a series of steroidal isoxazolines by the method of nit'rile oxide addition to an olefin, we found that pregna-5,16-dien-3p-ol-20-one acetate reacted smoothly in ether solution with carbethoxyfornionitrile oxide2 to yield (80%) the ethyl ]ester (I) of 16a117a-[3-carboxy-3,l-(2-isoxazolino) pregn-5-en-3p-ol-20-one 3p-acetate ester. Hydrolysis of this ester in methanol solution with aqueous sodium or potassium hydroxide a t room temperat'ure yielded the salt of the corresponding acid from which the free acid (11) was obtained by acidification with hydrochloric acid. The 3-acetate ester was removed simultaneously. When this hydrolysis was carried out by refluxing the ester in aqueous methanol with potassium carbonate, the acidic product (in 67y0 yield) was accompanied by a neutral product in'7% yield. This latter material showed infrared absorption a t 2220 (cyano) and a t 1680 em.-' (conjugated carbonyl). The ultraviolet absorption, X 247 mk ( e 10,700), also indicated the presence of a conjugated ketone. This same product (VII) was also obtained by a second process. When the free acid, 3p-hydroxy-16aJ1-pregn-5-en-20-one 17a- [3-carboxy-3,1-(2-isoxazolino) (11) was heated on a hot plate a t 250-280' to give a clear melt (with loss of COz and initial foaming) and then cooled and crystaIlized, the product again showed cyano group absorption in the infrared but without conjugate absorption. Loss of the carboxyl group and isoxazoline ring opening produced the 16a-cyano-17ahydroxy steroid (111) as a stable product. Treatment of this product (111) with base caused a dehydration and gave rise to VII. The structure of I11 was confirmed by analyt'ical data, the presence of hydroxyl (3440 and 3240 cm.-'), 20-carbonyl (1717 em.-'), and cyano (2267 cm.-') group absorption in the infrared, ( 1 ) A . Quilico, in "Heterocyclic Compounds," R . Wiley. E d . , Interscience Publishers, Inc., New York, N. Y . . 1960. p. 99. ( 2 ) C . S. Skinner, J . A m . Chem. Soc., 46, 731 (1924).
VI
V
IV
and by the n.m.r. spectrum of the A4-3-keto analog (IIIa). The A4-3-ketoseries was obtained by the selective addition (55% yield) of carbethoxyformonitrile oxide to the AI6 double bond of pregna-4,16-diene-3,20dione to give Ia, followed by 'hydrolysis and pyrolysis to give IIa and IIIa, respectively. The A4-3-keto'analog (IIIa) was sufficiently soluble in deuteriochloroform to obtain the n.m.r. spectrum, whereas the A6-3hydroxy steroid (111) was not. The n.m.r. spectrum of IIIa shows singlets in deuteriochloroform for the C-18 methyl (6 0.68) and for the C-21 methyl (6 2.28) , both of which are characteristic of the normal steroid structure and which are not in accord with a D-homo ~ t r u c t u r e . ~The &value of 0.68 is also in better accord with a 17a-hydroxy-17p-acetyl configuration than with the ('iso" structure, since the 17p-hydroxyl group tends to shift the resonance band of the C-18 methyl protons downfield in this type of s t r ~ c t u r e . The ~ effects of the 16-cyano group cannot, however, be completely assessed. This interpretation of the n.m.r. data, together with the probable attack of the nitrile oxide from the a-face of the molecule, leads to the designation of I11 as 16a-cyano-3P117~dihydroxypregn-5-en-20-0ne.~~ (3) N . R . Trenner, B. H. Arison. D . Taub, and N. L. Wendler, Proc. Chem. S o c . , 215 (1961). (4) Unpublished d a t a from these laboratories. T h e C-18 methyl proton resonance in epi-testosterone has a value 6 0.70 compared with testosterone, 8 0.80. The (2-18 methyl proton resonance in 38-acetoxy-17a-hydroxypregn-5-en-2O-one has a value d 0.70 compared with. 38-acetoxy-178hydroxy-1 7-isopregn-5-en-ZO-one, 6 0.94. (4a) NOTEADOED I N PRooF.-The mass spectrum of compound I I I a gave additional evidence for the unrearranged steroid structure, showing a molecular ion peak a t m/e 355 and fragments corresponding t o the loss of acetyl a t M - 312 and to acetyl a t M - 43. The mass spectrum was determined with a n Atlas C H 4 mas8 spectrometer with a n ionizing potential of 70 e.v. and a n ionizing current of 18 aa. We wish t o thank Dr. C. D . DeJongh of Wayne State University for determining and interpreting these results.
NOTES
APRIL1965 Treatment of I11 or IIIa with strong base readily yielded an a,@-unsaturatedketone VI1 or a bis a,@-unsaturated ketone VIIa, respectively. These were shown by n.m.r. to have a D-homo structure.6 The C-18 methyl proton resonance is shifted downfield to 6 1.02 in VI1 and to 6 1.05 in VIIa, characteristic of a 17a-keto-D-homo ~ t e r o i d . ~Instead of a singlet for the C-21 methyl protons, a split resonance (6 2.08) is present which would be expected for a C-17 methyl in a Dhomo steroid, split unequally through the 16,17 double bond by the C-15 protons.' In the base-catalyzed D-homo rearrangement of a steroid with a 17a-hydroxy-l7@-acetylfunctionality, as in 111, the predominant product in past experience has been the 17a-methyl-17a-hydroxy-17-one D-homo system corresponding to migration of the 13,17 bond.8 Elimination of the elements of water from such a system in the present case would lead to a 17a-methylene group, a structure not allowed by the n.m.r. spectrum which shows no vinyl proton resonances and does show the D-homo 17-methyl proton resonance. The alternative possibility of C-18 methyl migration is precluded by the above considerations and the spectral requirement of an a,@-unsaturatedcarbonyl. Eli4nation thus removes the product VI1 from the equilibration mechanisms of the D-homo acyloin rearrangement in what corresponds to 16,17 bond migration. The structures described above also fix the direction of addition of the nitrile oxide to the 16,17 double bond (as shown in I) analogous to the addition of diazomethane to this bond.g The two 16a-cyano-17a-hydroxy steroids melt considerably higher (111, 275"; IIIa, 245") than the corresponding 3-carboxy-2-isoxazolines (11, 213' ; IIa, 210") and this may explain the erratic melting points of these acids, which decarboxylate a t their melting points and then exhibit a higher melting range. Drefahl and Horholde have commented in a preliminary statement on the thermal instability of the 3carboxy-2-isoxamlines during saponification of the esters, indicating the formation of an aldehyde (equivalent to loss of a carbon atom from the initial olefin), of carbon dioxide, and of an additional unidentified product. Whether this is a base catalyzed or only a thermal reaction is not clear from the description given. Experimental Melting points are corrected and were determined in a ThomasHoover ( T H ) apparatus (capillary tube) or on a Fisher-Johns (FJ) hot plate and are so labeled. The following measurements were made by Dr. J. M. Vandenbelt and his staff in our laboratories: rotations were measured a t room temperature in chloroform solution in a l-dm. tube; infrared spectra were determined in potassium bromide pressed disks in a Reckman IR-7 recording spectrophotometer; ultraviolet absorption spectra were determined in methanol in a Model 14 Cary recording spectrophotom(5) We are indebted to a referee of the original manuscript for valuable suggestions regarding a D-homo rearrangement reconciling our results with those of Drefahl and Horhold.8 (6) G . Drefahl and H. H. Horhold, Chem. BeT., 97, 159 (1964). (7) R e wish to thank Dr. Glenn Berchtold for discussions of the n.m.r. data. (8) N . L. Wendler in "Molecular Rearrangement," Vol. 11, P. DeMayo, Ed., Interscience Publishers, Inc., New York, N . Y . , 1964, p. 1116; N . L. Rendler, D. Taub, and R . Firestone, Ezperientia, 16,237 (1959). (9) J. A. Moore, W. F. Holton, and E . L. Wittle, J . A m . Chem. Soc., 84, 390 (1962); D . Taub, R . D . Hofsommer, H. L. Slates, C. H. Kuo. and N . t. Wendler, ibid.. 89, 4012 (1960).
1273
eter. Microanalyses were by Mr. C. E. Childs and his staff in our laboratories. 160,17a- [ 3-Carboxy-3,1-(2-isoxazolino)]pregn-5-en-30-01-20one 30-Acetate, Ethyl Ester (I).-A solution of 3.56 g. (0.01 mole) of pregna-5,16-dien-3p-ol-2O-one acetate and 4.58 g. (0.03 mole) of ethyl chloroximinoacetate* in 500 ml. of dry ether waa treated with a solution of 4 ml. of triethylamine in 125 ml. of dry ether a t 25" during 3 hr. with stirring. The reaction waa then stirred for an additional 3 hr. and let stand a t 25' overnight. The mixture was filtered and the filtrate waa evaporated 'to a yellow oil which solidified on standing. Crystallization from methanol gave 3.8 g. of pearly white plates (80% yield): m.p. 168-170' ( T H ) ; Amax 249 mp ( e 5000); [a]"D -14' (C 0.92, CHCl,); vmax 1728 (ester), 1717 (sh) (keto), 1588 (C=N) cm.-l. Anal. Calcd. for Cz7H3,NOe(471.6): C, 68.76; H, 7.89; N , 2.97. Found: C, 68.94; H , 7.99; N (Kjeldahl), 3.08. 1&,170-[3-Carboxy-3,1-(2-isoxazolino)pregn-5-en-3~-01-20one (II).-A mixture of 0.5 g. of I with 30 ml. of methanol, 0.6 ml. of 50% aqueous sodium hydroxide, and 3 ml. of water waa swirled to dissolve the steroid and was then let stand for 2 hr. a t room temperature. A thick white precipitate formed. The mixture was cooled in ice and filtered. The solid was dissolved in 50 ml. of water and 1 ml. of methanol, diluted to 100 ml. with water, cooled, and acidified to pH 1 with 3 N hydrochloric acid. The mixture was chilled in ice and filtered. The product weighed 0.27 g., m.p. 224-230" ( T H ) . Crystallization from 50% aqueous dioxane gave 0.21 g. of colorless needles: m.p. 213' (TH); 50y0 yield; ywx 3405 (OH), 1730 (carboxyl), 1715 (keto), 1593 (C=N) cm.-l. A n d . Calcd. for CZ3HalNO5(401.5): C, 68.80; H, 7.78; N, 3.48. Found: C, 68.84; H , 7.80; N (Kjeldahl), 3.56.
3~-Hydroxy-l7-methyl-l7a-oxo-D-homo-pregna-5,16-diene16-carbonitrile (VII).-A solution of 2.78 g. (5.9mmoles) of I in 100 ml. of methanol with 2 g. of potassium carbonate and 25 ml. of water waa refluxed for 2 hr. The methanol waa evaporated under reduced pressure and the neutral solid waa separated by filtration. The filtrate was acidified with acetic acid to yield 1.6 g. of acidic product (11). The neutral product, 0.29 g., was crystallized from aqueous acetone to give VII: 0.14 g.; m.p. 260-262" ( T H ) ; Amnx 247 mp (e 10,700);vmax 3470 (hydroxyl), 2220 (cyano), 1680 (keto) cm.-'. Anal. Calcd. for C Z ~ H Z ~ N (339.5): O~: C, 77.80; H , 8.61; N,4.12. Found: C, 77.76; H, 8.61; N, 4.21. Acetate.-VI1 with pyridine and acetic anhydride yielded white needles (from ethanol): m.p. 260"; kmax 246 mN ( C 10,600); vmr. 2210, 1728,1678,1640, 1623, 1259 cm.-l. Anal. Calcd. for CzrHslNOa(381.5): C, 75.55; H , 8.19; N, 3.68. Found: C, 75.54; H, 8.27; N, 3.77. 3~,17~-Dihydroxy-20-0xopregn-5-ene-1&-carbonitrile (III).-A sample (1.7g., 4.2 mmoles) of I1 was heated on a hot plate in an erlenmeyer flask a t 280" to give a clear yellow melt. On cooling, the product was taken into methanol and crystallized to yield 700 mg. (44%), m.p. 225265'. Recrystallization from methanol and ether gave 350 mg.: m.p. 275-277' ( F J ) ; no significant ~ ( c 0.59, MeOH); vmax ultraviolet absorption; [ a I z 3 -98.3" 3440 and 3240 (hydroxyl), 1717 (keto), 2267 (cyano) cm.-l. Anal. Calcd. for CnH31N03: (357.5); C, 73.91; H , 8.74; N , 3.92. Found: C, 73.74; H , 9.04; N, 4.31. 3B-Hydroxy-17-methyl-17a-oxo-D-homo-pregna-5,16-diene16-carbonitrile (VII) from 111.-A solution of 50 mg. of I11 and 500 mg. of potassium hydroxide in 15 ml. of methanol was heated on the steam bath for 10 min., cooled, and precipitated with water. The solid was filtered off, washed with water, and dried in air to yield 40 mg., m.p. 255" ( F J ) of VII, infrared absorption curve identical with that of VI1 obtained by the carbonate hydrolysis of I (see above). 160,17~~-[3-Carboxy-3,1-( 2-isoxazolino)]pregn4-ene-3,20dione, Ethyl Ester (Ia).-A solution of 1.6 g. of 16-dehydroprogesterone in diethyl ether, when treated with 2.3 g. of ethyl chloroximinoacetate by the same method as for I, yielded 1.2g., 55% yield, m.p. 215-217" (FJ). Two crystallizations from methanol gave 1.0 g.; m.p. 218-220" ( F J ) ; A,, 240 mp ( C 21,000); [ 0 l z a D f89.3" (C 0.56, CHCls); vmax 1727, 1676, 1625, 1597 cm.-'. A n d . Calcd. for C Z ~ H ~ ~(427.5): N O ~ C, 70.23; H, 7.78. Found: C, 70.22; H , 7.65. 3,17a-Dioxo-17-methyl-D-homo-pregna-4,16-diene-l6-~arbonitrile (VIIa).-A solution of 1.5 g. of I a in 100 ml. of methanol with 3 g. of potassium bicarbonate in 25 ml. of water was heated
NOTES
1274
to boiling and then allowed to stand overnight a t 25'. The mixture was filtered into excess dilute hydrochloric acid, diluted with water, and filtered; the product was washed with water. The wet acid IIa was melted in a 50-ml. erlenmeyer flask on a hot plate, cooled, taken into ether and filtered. The filtrate on evaporation gave an orange oil IIIa which was treated with 250 mg. of potassium hydroxide in 25 ml. of methanol and boiled for 15 min. The reaction was diluted with water to yield a crystalline solid, 400 mg., m.p. 170-174O. The product was purified by passing it over 10 g. of alumina in dichloromethaneether. The first eluates gave white crystalline material which was crystallized from ether to give 180 mg. of VIIa: 15y0yield; m.p. 190-193' ( F J ) ; A,, 240.5 mp ( e 26,800); umax 2215 (cyano), D (cO.54, CHC13). 1688, 1 6 2 4 c m . ~ ' ;[ ( Y ] ~ ~-18.5' Anal. Calcd. for C Z Z H ~ ~ N(337.4): OZ C, 78.30; H , 8.08. Found: C, 78.33; H, 8.22. 16~y 170r-[3-Carboxy-3,1-(2-isoxazolino)]pregn-4-ene-3 ,tO-dione (IIa).--A solution of 4.0 g. (0.01 mole) of I1 in 1 1. of acetone (distilled over K M n 0 4 )was treated a t ice-bath temperature and under nitrogen with 5 ml. of 8 N chromic acid (Jones reagent)'(' during 7 min., stirred far an additional 5 min., and quenched in 1.5 1. of ice-water. Most of the acetone was evaporated under reduced pressure, and the precipitated solid was separated by filtration, washed with water, and dried, 2.9 g., 727, crude. The crude product was refluxed for 25 min. in 200 ml. of 95% ethanol with 900 mg. of oxalic acid dihydrate. The mixture was evaporated under reduced pressure a t 50-60" and crystallized from dioxane-water to give 2.37 g. (60% yield): m.p. 210-212' ( T H ) ; A,, 240 my ( e 18,700); [aIz4r,+102" (c 0.51, CHCl,); vmsx 1741, 1719, 1675, 1633, 1587 em.-'. Anal. Calcd. for C23HzsN05 (399.5): C, 69.15; H, 7.32; N , 3.51. Found: C, 69.05; H , 7.13; Tu', 3.60. 17a-Hydroxy-3,2O-dioxopregn-4-ene-lba-carbonitrile (IIIa).-A sample (1.0 9.)of IIa was melted a t 250' until all foaming ceased. The melt was cooled and stirred with methanol to yield 280 mg. of a solid product, m.p. 245" (FJ). Two crystallizations from methanol and ether gave 80 mg.: m.p. 245-247" ( F J ) ; Am,, 240 mp ( e 16,000); [ Y ] ~ I D +92.2' (c 0.51, CHC13); vmsx 3440 (hydroxyl), 2250 (cyano), 1710 (keto), 1655 (keto), 1616 cm.-'. Anal. , Calcd. for C22H29NO3(355.5): C, 74.33; H , 8.22. Found: C, 74.14; H, 8.22. (10) K. Bowden, I. M. Heilbron, E. R . H. Jones, and B. C. L. Weedon J . Chem. Soc., 39 (1946).
VOL. 30
I
11
a t 3.0 p in its infrared spectrum. This material is instead assigned structure I1 on the basis of chemical and physical data cited below. The elemental analysis of 11, m.p. 225-226', indicates that it is an isomer of I. Dehydration with thionyl chloride-pyridine in benzene afforded in excellent yield l-p-toluenesulfonyl-3,5-diphenylpyrazole (111). Treatment of I11 with lithium aluminum hydride in tetrahydrofuran readily gave rise to 3,5-diphenylpyrazole (IV). Structure I11 was further confirmed by its unequivocal synthesis from 3,5diphenylpyrazole and p-toluenesulfonyl chloride. The various transformations leading to the products described above are outlined in Chart I. CHARTI
n
\
502CsH4CHs
SOCISI; pyridine
Me0 b
A Novel 2.2.1-Bicyclic Elimination of a N-Tosylpyrazoline ALBERTPADWA Department of Chemistry, The Ohio State University, Columbus, Ohio Received November 23, 1964
The pyrolytic decomposition of the alkali salts of tosylhydrazones of aldehydes and ketones in aprotic solvents has been reported to give products expected to arise from intermediate carbenes.'S2 I n view of our interest in heterocyclic small-ring compounds, we sought to define further the reactivity of a carbenoid center adjacent to a three-membered oxirane ring. I n an attempt to prepare the tosylhydrazone of trans-chalcone oxide (I), two different approaches were tried. Treatment of trans-chalcone oxide with tosylhydrazide in acidic ethanol for 5 min. a t 50' gave a product identical with that obtained by treating benzalacetophenone tosylhydrazone with 10% sodium hydroxide and 30% hydrogen peroxide in methyl alcohol. That the product from both reactions was not the desired tosylhydrazone I was evidenced by a strong peak M.C. Whlting. Tetrahedron, I , 305 (1959). (2) L. Friedman and H. Sheohter. J . A m . Chem. Soc.. 81,5512 (1959). (1) J. W. Powell and
Even with the mildest conditions, 4-hydroxypyrazoline (11) resulted from the reaction of tosylhydrazide with the trans oxide of benzalacetophenone. The mechanism of its formation probably involves the intermediate formation of tosylhydrazone (I) which then undergoes an intramolecular ring opening and closure to form 11. The configuration of I1 is such that the hydroxyl group on C-4 and the hydrogen on C-5 are on the same side of the pyrazoline ring. A related study of the reaction of hydrazine with an epoxy ketone3 has shown that the epoxide ring is opened to give a similar intermediate, which, on heating, loses water to yield the corresponding pyrazole. It has also been reported4 that the related trans-ethylenimine ketones and phenylhydrazine produce the analogous 4-aminotrialkylpyrazolines. Treatment of I1 with sodium hydride in diglyme or tetrahydrofuran led to an unexpected result. Under the basic conditions employed, I1 was converted in almost quantitative yield to 3,E~diphenylpyrazole (IV) (3) Jorlander, Ber , 49,2782 (1916) (4) N H Cromwell and H Hoeksema J A m Chem Soc 11, 716 (1949)
NOTES
APRIL1965 SCHEMEI
I1 NaH
I11
1.
b1 0 .
-
I
Id
H
P h l
v
Ph/t'$N
H
u
and p-toluenesulfonic acid. Scheme I is a summary of a possible sequence capable of describing pyrazole formation. The basic dehydration (path a) of an analogous 4hydroxypyrazoline to the corresponding pyrazole5 demanded that particular attention be paid to this possibility (ie., paths a and b). This sequence, however, is eliminated by the finding that I11 was stable under conditions a t which pyrazole formation from I1 proceeded smoothly. This result caused 111 to be rejected as a possible precursor of the pyrazole. The formation of IV can therefore be regarded as occurring directly from compound I1 and may be rationalized by sequence c through g. The scheme consists of the initial formation of an alkoxide anion followed by intramolecular migration of the p-toluenesulfonyl group from nitrogen to oxygen. This step involves a transition state which is geometrically comparable to a bicyclo [2.2.1]heptane and suggests that a minimum of strain energy would be required for the migration. Once the p-toluenesulfonyl group has migrated, the resulting species eliminates a tosylate anion and is followed by subsequent tautomerization to the stable pyrazole. Treatment of I1 with hot alcoholic base resulted in a 20% conversion to IV only after refluxing for 24 hr. Presumably the alkoxide anion of I1 in a protic solvent is well solvated, and achieving the 2.2.1-bicyclic transition state becomes increasingly difficult. Experimental
1275
30 min. On cooling, the product crystallized in small white plates, m.p. 225-226", 0.89 g., 47%. Anal. Calcd. for C22H20N203S: C, 67.32; H, 5.14; N, 7.14; S, 8.17. Found: C, 67.15; H,4.94; N, 7.26; S.8.40. The infrared spectrum of the 4-hydroxypyrazoline (11) has bands at 3.00, 6.27, 7.41, and 8.60 p. To 1.O g. of benzalacetophenone tosylhydrazide dissolved in 10 mi. of methanol was added 1.20 mi. of 2 N NaOH and I .60 ml. of 30y0 hydrogen peroxide. The mixture was allowed to stand at room temperature for 12 hr. It was then diluted with an equal volume of water, saturated with sodium chloride, and thoroughly extracted with ether. The ethereal extracts were washed over sodium thiosulfate solution and were dried over sodium sulfate. Evaporation of the solvent left 0.42 g. (4oyO,)of a colorless solid, m.p. 225-226". The infrared spectrum of this solid was identical in every detail with the solid described above. Preparation of l-p-Toluenesulfonyl-3,5-diphenylpyrazole (111). -A solution of 1.0 g. of 4-hydroxypyrazoline (11) and 3 ml. of pyridine in 50 ml. of benzene was cooled to 0" and 2.0 g. of thionyl chloride in 25 mi. was added dropwise over 10 min. The solution was stirred for an additional 30 min. during which time the temperature was allowed to rise to 20". The mixture was then heated under reflux for 1 hr. and then filtered to remove the pyridine hydrochloride. The solution was washed twice with water and dried over sodium sulfate. Evaporat,ion of the solvent left 0.89 g. of a pale yellow oil.. This material was taken up in hexane-benzene and a cryst'alline solid soon precipitated. The solid amoiuited to 0.78 g. (7474, m.p. 108-114", and was recrystallized from hexane-benzene to give a solid of m.p. 118-1 19O . Anal. Calcd. for C22Hl*N2O2S: C, 70.58; H , 4.85. Found: C! 70.76; H, 4.72. The infrared spectrum is characterized by bands a t 7.30, 8.41, and 8.5 p . The n.m.r. spectrum (CDC1,) shows a miiltiplet centered a t T 2.59, a singlet at 3.41 and a singlet a t 7.68. The peak areas are in the ratio of 14: 1 : 3 . , The infrared spectrum of the solid was superimposable on that of an authentic sample of l-p-toluenesulfonyl-3,5-diphenylpyrazole (prepared by heating equimolar quantities of 3,5-diphenylpyrazole and p-toluenesulfonyl chloride in pyridine for 4 hr. 011 a steam bath). This material can be readily reduced with lithium aluminum hydride to give 3,5-diphenylpyrazole and p-toluenesulfonic acid. Lithium Aluminum Hydride Reduction of the 4-Hydroxypyrazoline Derivative.-To 75 mg. of fresh lithium aluminum hydride in 10 mi. of ether, stirred magnetically, was added a solution of 500 mg. of 4- ydroxypyrazoline in 10 ml. of ether. There was slight warming on addition. The mixture was stirred for 1 hr. a t room temperature and was then hydrolyzed with saturated ammonium chloride. The aqueous layer was extracted with ether and the organic extracts were dried over sodium sulfate. Evaporation of the solvent left a colorless solid, m.p. 183-190", 0.82 g. Successive recrystallizations from methanol gave white needles, m.p. 201-202". Anal. Calcd. for Cl5HI2N2:C, 81.79; H , 5.49; N, 12.72. Found: C, 81.87; H, 5.64; N , 12.48. The infrared spectrum of this material was identical in every detail with an authtntic sample of 3,5-diphenylpyrazole.6 Reaction of l-p-Toluenesulfonyl-3,5-diphenyl-4-hydroxypyrazoline with Sodium Hydride.-h 5 ml. of tetrahydrofuran was dissolved 100 mg. of the 4-hydroxypyrazoline (11). To the mixture was added 15 mg. of sodium hydride. The solution was heated under reflux for 10 min. and was then poured onto ice. The mixture was extracted with ether and the extracts were dried over sodium sulfate. Evaporation of the ether afforded 70 mg. of 3,5-diphenylpyrazole. Treatment of 1-p-toluenesulfonyl-3,5-diphenylpyrazole (111) under identical conditions wit,h those described above gave, as the only isolated product, 957' yield of recovered starting material.
Attempted Preparation of the Tosylhydrazone of trans-Chalcone Oxide.-A mixture of trans-chalcone oxide (4.47 mmoles) and tosylhydrazide (4.50 mmoles) in 15 ml. of acidic ethanol (ly0 H&Od-95yO ethanol) was refluxed in a nitrogen atmosphere for
Acknowledgment.-The author wishes to acknowledge the excellent technical assistance of l\liss Linda Norling. This investigation was supported in part by The Petroleum Research Fund, administered by the American Society, through a Type G grant.
( 5 ) T. L. Jacobs, "Heterocyclic Compounds," Vol. 5, R . C. Elderfield, E d . , John Wiley and Sons. New York, N . Y.. 1957, p. 68.
(6) An authentic sample of 3.5-diphenylpyraaole was prepared by the procedure of A . Dornow and W. Rartsch, A n n . , 602, 23 (1987).
NOTES
1276 Ozonolysis. V. Reaction of an Ozonide with the Isopropyl Grignard Reagent FREDL. GREENWOOD AND BERNARD J. HASKE~ Contribution N o . 38'7, Chemistry Department, Tufts University, Medford, Massachusetts Received November 24, 1964
It has been reported2 that treatment of the reaction mixture resulting from the ozonation of trans-alkenes in ether solution at -112" with the isopropyl Grignard reagent gave rise to an a-diol and carbinols. Supposedly, the a-diol arose from reaction of the Grignard reagent with the molozonide, and the carbinols, which were obtained in much smaller quantity than the a-diol, were formed by reaction of the Grignard reagent with a small amount of ozonide in the reaction mixture. There are in the literature several instances of the treatment of ozonation reaction mixtures with Grignard reagents. Criegee and Schroder3 described the reaction of methyl and isopropyl Grignard reagents with the reaction mixture resulting from the ozonation of dit-butylethylene, and Sparks and Knobloch* reported the reaction of ethyl and phenyl Grignard reagents with reaction mixtures obtained by the ozonation of several 1-alkenes. In each case the products were carbinols which would result from the reaction of the Grignard reagent with the cleavage products of the ozonized alkene. Ozonation of an alkene in solvents devoid of active hydrogen atoms gives rise to polymeric peroxidic material, carbonyl compounds, and ozonide, all of which will react with the Grignard reagent. Indeed, the ozonide may be formed in least amount. There are no reports of the reaction of a pure ozonide with the Grignard reagent, and it was thought desirable to ascertain the products resulting from t,hese reactants. The ozonide was obtained in excellent yield by the ozonation of cis-3-hexene in n-pentane solution a t -70". The ozonide distilled with no temperature range; vapor phase chromatography (v.P.c.) indicated a single component, and the infrared spectrum confirmed the absence of carbonyl- or hydroxy-containing impurities. In order to make comparisons with earlier experimentsZthe reaction of the ozonide (26.6 mmoles) with excess isopropyl Grignard reagent was carried out as were the earlier reactions; i.e., the reactants were mixed at -112" and the reaction mixture was allowed to warm slowly. The temperature behavior of the reaction mixture in the early and later stages of the addition of the Grignard reagent indicated that some reaction did occur at -112'. The reaction mixture was eventually refluxed, and during the reflux period propane (17.1 mmoles) and propene ' (10.4 mmoles) were evolved. The ozonide and Grignard reagent reacted in the molar rat'io 1.00:3.89. The other identifiable products of the react,ion were isopropyl alcohol (12.8 mmoles) and ethylisopropylcarbinol (36.7). (1) Participant, summer 1964, National Science Foundation Research Participation for College Teachers Program. (2) F . L. Greenwood, J . O w . Chem., 19, 1321 (1964). (3) R . Criegee and G.Schroder. Chem. Ber., 98. 689 (1960). (4) J . W. Sparks and J. 0 . Knobloch, U. S. Patent 2,871,812 (1954); Chem. Abatr.. 49, 5508 (1955).
VOL. 30
The products resulting from the treatment of the solution which was obtained by the ozonation of cis3-hexene in ether solution at - 112" with isopropyl Grignard reagent have been reported.2 To compare the results of this experiment with those described above, one must correct for the larger amount of ozonide in the earlier experiment. After this correction the products were propane (17.7 mmoles), propene (11.8 mmoles), isopropyl alcohol (18.9 mmoles), and ethylisopropylcarbinol (36.0 mmoles). Other than the discrepancy in the amounts of isopropyl alcohol, there is good agreement between the two experiments. It is clear that ozonides react with the Crignard reagent to give carbinols which would result from reaction of the Grignard reagent with the cleavage products of the ozonide. Also, it is clear that not all of the carbon atoms of the ozonide react in this fashion, for the yield of ethylisopropylcarbinol is but 69y0 of what one would expect if all of the ozonide reacted to give this carbinol. Experimental Chemicals.-The n-pentane and isopropyl bromide were purified as described earlier.* The cis-3-hexene was an API standard sample (Carnegie Institute of Technology). Ether was an anhydrous grade which was stored over sodium ribbon. Ozonide Preparation.-A solution of 3.26 g. (38.7 mmoles) of cis-3-hexene in 250 ml. of n-pentane was ozonized with 36.8 mmoles of ozone a t - 70". The ozone-oxygen flow rate was 13.0 l./hr., and the ozone was added a t the rate of 583 mg./hr. The ozonation flask was placed in a -50" bath, and the pentane was removed through a rotary evaporator under a pressure of 0.1 mm. The pentane (243 ml.) was condensed ixi a liquid nitrogen cooled trap. The -50" bath was replaced by one at 25", and 4.22 g. of ozonide (87% yield) was collected in a clean trap. A small residue (0.40 g.) remained in the reaction flask. The ozonide was distilled through a Vigreux column to give 3.98 g. (30.2 mmoles, D mol. wt. 82y0 yield) of ozonide, b.p. 45" (22 mm.), T L ~ ~1.4012, 132 (cryoscopic, benzene). The nuclear magnetic resonance spectrum of the ozonide (35 wt. % in carbon tetrachloride) had two overlapping triplets centered a t 4.92, a multiplet centered a t 1.55, and two, unresolved, overlapping triplets centered a t 0.82 p.p.m. The integrated areas of these three peaks were in the ratio 1.00: 2.04: 3.12 and, accordingly, were assigned to the methine, metliylene, and methyl protons, respectively. For the ozonide of trans3-hexene, Criegee6reported b.p. 46.5" (26 mm.), ~ ) O D 1.4050. V.P.C. of a pentane solution of the ozonide on a silicon GE SF 96 column showed only one component. An infrared spectrum (Perkin-Elmer Model 421) of the ozonide (neat) indicated the absence of carbonyl- or hydroxy-containing impurities. Comparison of the spectrum with that of di-n-propyl peroxide showed the similarities to be very strong absorption a t 950, 1380, and 1460 for the ozonide and 955, 1375, and 1453 cm.-l, peroxide. Reaction of the Ozonide with Isopropyl Grignard Reagent.In a flask fitted with a stirrer, thermometer well, and dropping funnel was placed a solution of 3.52 g. (26.6 mmoles) of ozonide in 350 ml. of ether. To this solution was added 100 ml. (pipet) of isopropylmagnesium bromide (1.72 M ) a t such a rate that the temperature of the reaction mixture did not rise above -110'. After standing 3 days, the reaction mixture had warmed to 20'. The dropping funnel was replaced by a condenser which was attached to a gas-collecting bottle. Stirring and refluxing (1 hr.) the reaction mixture until no more gas was evolved gave 1183 ml. (STP)of gas, which was found by mass spectroscopy* to contain 17.1 mmoles of propane and 10.4 mmoles of propene. After cooling to - 10" the reaction mixture was hydrolyzed by the addition of a solution of 9.31 g. of concentrated sulfuric acid in 100 ml. of water. Refluxing the mixture liberated the propane from the excess Grignard reagent. Determination of this pr3pane showed that 104 mmoles of Grignard reagent had reacted with the ozonide; ozonide-Grignard reagent molar ratio, 1.OO :3 3 9 . (5) R . Criegee, A . Kerchow, and H. Zinke. Chem. Ber., 88, 1878 (1955).
(6) We are indebted to the Texas Research Laboratories, Beacon, N . Y., for the mass spectrometric analysis.
1277
NOTES
APRIL1965 The aqueous layer from the hydrolysis mixture waa extracted in a continuous ether extractor. The ether extract waa combined with the ether layer from the hydrolysis mixture, and the ether was removed from the dried (sodium sulfate) ether soution through a Fenske column. The residue wm distilled i n vacuo (5 mm.) through a Vigreux column. The distillate (4.69 9 . ) was condensed in a receiver which was in a -70" bath. A residue of 0.43 g. remained in the distilling flask. The constituents of this distillate were identified by comparison with authentic samples by V.P.C. (cyanosilicone column). The quantitative V.P.C. analysis of the distillate wm carried out by the method of internal normaliati ion.^ The distillate contained isopropyl alcohol (12.8 mmoles), ethylisopropylcarbinol (36.7 mmoles, 69% yield on the basis of 1 mole of ozonide giving 2 moles of carbinol), and unidentified material (0.14 g.). (7) R. L. Pecsok, "Principles and Practice of Gas Chromatography," John Wiley and Sons, Inc., New York, N . Y . , 1959, p. 143.
A New Synthesis of p -Methylaminobenzoyl- L-glutamic Acid' S.-C. J. F u
AND
MARIONREINER
The Children's Cancer Research Foundation, The Children's Hospital Medical Center, Harvard Medical School, Boston, Massachusetts AND
TI LI
LOO
Laboratory of Chemical Pharmacology, National Cancer Institute, National Institutes of Health, Bethesda, Maryland Received December 8, 1964
The preparation of p-methylaminobenzoyl-L-glutamic acid, an intermediate in the synthesis of N'O-methylfolic acid and analogs, by the aminolysis of sodium p-iodobenzoyl-L-glutamate with methylamine a t pH 8-9 has been reported.2 The desired acid failed to crystallize, and the crude sodium salt was isolated in unspecified yield. We were unsuccessful with this procedure until the use of sodium hydroxide during the condensation was eliminated. The p-methylaminobenzoyl-L-g1 u t a m ic acid was then isolated in low yield as the dihydrated barium salt from which the free acid was obtained as an oil. There was also a possibility of racemization of the optically active compound in the alkaline solution a t elevated temperature. In a procedure here described, p-methylaminobenzoyl-L-glutamic acid hydrobromide has been synthesized in .high purity and yield suitable for the synthesis of N'O-methylfolic acid and derivatives. Starting with p-methylaminobenzoic acid, the methylamino group was protected by carbobenzyloxylation before conversion of the acid into its chloride. The Np - [ (carbobenzyloxy)methylamino]benzoylchloride was allowed to react with diethyl L-glutamate as in the synthesis of derivatives of aminobenzoylglutamic acid.3 The resulting ester was hydrolyzed by alkali to afford N- { p - [(carbobenzyloxy)methylamin olben zoyl )-L-glutamic acid. The carbobenzyloxy group was removed by 40% hydrogen bromide i n glacial acetic and the p-methylaminoberizoyl-L-glutamic acid was isolated (1) Supported in part by grants from the National Institutes of Health, U. S. Public Health Service, No. CY-3335 and C-6516. (2) D. D. Cosulich and J. M. Smith, J r . , J . Am. Chem. Sac., T O , 1922 (1948). (3) S.-C. J. Fu, J . Med. Pharm. Chem.. 6 , 33 (1962). (4) D. Den-Ishai and A. Berger, J . Org. Chem., 11, 1564 (1952).
as the hydrobromide. Attempts to remove the hydrogen bromide from the dipeptide with pyridine, triethylamine, or sodium hydrogen carbonate yielded an oily product which could not be reconverted to the crystalline hydrobromide by treatment with hydrogen bromide in et,her. The diethyl ester was prepared by the decarbobenzyloxylation of diethyl N-{ p- [(carbobenzyloxy)methylamino]benzoylf +glutamate and it showed optical rotation identical with that reported.2 This indicates that no racemization occurred in the earlier aminolysis procedure. Experimental6 Barium p-Methylaminobenzoyl-L-glutamate from p-Iodobenzoyl-L-glutamic Acid.-p-Iodobenzoyl-L-glutamic acid,* 252 m p , 2 g. (5.3 mmoles), 5 ml. of 3Oy0 aqueous solution of methylamine, and 10 mg. of fine copper powder was heated in a sealed tube a t 150" for 10 hr. The reaction mixture waa diluted with 10 ml. of methanol and filtered. Upon concentration under nitrogen, a golden oil resulted, which defied attempts a t crystallization and which gave unreproducible refractive indices and erroneous elementary analyses. The oil was redissolved in 5 ml. of methanol and treated with 5 ml. of 0.5 M barium chloride in 50% methanol. The precipitated barium p-methylaminobenzoyl-1,-glutamate waa collected and washed with methanol; yield 0.6 g. (24%), 291 mp. Anal. Calcd. for C 1 3 H ~ ~ B a N ~ 0 5 . 2 HC, ~ 0 :34.6; H, 4.0; N,6.2. Found: C,34.1; H,4.0; N,6.0. Treatment of the barium salt with 1 N sulfuric acid to pH 6 afforded the free dipeptide as an oil which again gave unreproducible refractive indices and erroneous elementary analyses. N-p-[( Carbobenzyloxy)methylamino]benzoic Acid.-pMethylaminobenzoic acid (2 g., 13.3 mmoles) purified by the nitrosation procedure6 waa dissolved in 10 ml. of 2 N sodium hydroxide. To this solution was added carbobenzyloxy chloride (2.3 ml., 16 mmoles) alternately with 17 ml. of 2 N sodium hydroxide a t O D . Upon acidification to pH 3, the crude compound precipitated; it was collected and washed with cold water. The pure N-p-[( carbobenzyloxy)methylamino]benzoic acid was obtained after two recrystallizations from glacial acetic acid; yield 66%, m.p. 149.5-151.5". Anal. Calcd. for ClBHIdNO4: C, 67.4; H , 5.3; N , 4.9. Found: C,67.3; H , 5.4; N,4.9. Diethyl N-(p-[(Carbobenzyloxy)methylamino]benzoyl I-L-glutamate .-The N-p- [ (carbobenzyloxy)methylamino]benzoyl chloride was prepared by treating N-p-[(carbobenzy1oxy)methylaminolbenzoic acid (1.4 g., 5 mmoles) with phosphorus pentachloride (1.2 g., 5.5 mmoles) in 20 ml. of anhydrous ether. The acid chloride was not isolated; the ether solution waa washed with 10 ml. of ice-cold water and immediately poured into a mixture of diethyl L-glutamate hydrochloride7 (1.2 g., 5 mmoles) and sodium hydrogen carbonate (3.3 g., 40 mmoles) in 20 ml. of ethyl acetate and 20 ml. of water. This was stirred a t 0' for 15 min. and then a t 25-30' for 1 hr. The organic phase was separated and washed successively with 10 ml. each of water, 2 N hydrochloric acid, and water. The crude product obtained after evaporation of the solvent was recrystallized from ethyl acetate-n-hexane (1 :5 ) and acetone-n-hexane (1:5); yield 79%, m.p. 85", sintered a t 64-66". Anal. Calcd. for C*5H30N207:C, 63.8; H , 6.4; N, 6.0. Found: C,63.7; H,6.2; N,6.0. N-( p- [( Carbobenzyloxy)methylamino]benzoyl ) -L-glutamic Acid Hydrate .-Diethyl N-p- [(carbobenzyloxy)methylamino] benzoyl-L-glutamate (2.4 g., 5 mmoles) was hydrolyzed in 1 N sodium hydroxide ( 11 ml., 11 mmoles) and 11 In1 . of methanol, first a t 0" for 10 min. and then a t 25-30" for 1 hr. The resulting (5) All melting points are corrected. Optical rotation was measured with a Bellingham and Stanley polarimeter. Ultraviolet absorption spectra were recorded with a Cary Model 15 spectrophotometer. The elementary analyses were performed by Dr. C. K . Fitz, Needham Heights, Mass. (6) (a) A . R . Surrey and H. F. Hammer, J . Am. Chem. Sac., 66, 2127 (1944); (b) F. Klaus and 0 . Baudisch, Ber., 61, 1036 (1918). (7) [ a l * ' D +23.0° (absolute ethanol). [ a ] D +22.4O (water) is given in J. P. Greenstein and M. Winitz, "Chemistry of Amino Acids," Vol. 11, John Wiley and Sons, Inc.. New York, N . Y . , 1861, p. 932.
NOTES
1278
clear solution was acidified to pH 3 with 6 N hydrochloric acid. The crude waxy crystals were recrystallized from acetone-nhexane (1:5). The yield was quantitative; m.p. 74-75'. Anal. Calcd. for C ~ I H ~ Z N ~ O ~ . H C,Z O 58.3; : H, 5.6; N, 6.5. Found: C, 58.6; H,6.0; N, 6.5. p-Methylaminobenzoyl-L-glutamic Acid Hydrobromide.- To N-( p-[( carbobenzyloxy)methylamino]benzoyl )-L-glutamic acid (1 g., 2.4 mmoles), 3 ml. of 40y0 hydrobromic acid in glacial acetic acid4 was added with vigorous stirring a t 25" for 1 hr. The evolution of carbon dioxide had almost subsided in 15 min. and the clear solution was stirred for another 30 min. Anhydrous ether (30 ml.) was added and the mixture was stirred continuously a t 4" for 20 hr. After decantation of the ether, the semisolid was dried in vacuo over phosphorus pentoxide, concentrated sulfuric acid, and sodium hydroxide pellets. A white hygroscopic crystalline powder was obtained and no further purification was necessary. The yield was 787,; m.p. 72-74" dec. (sealed tube), 291 m p ( e 15,050), [ a I z 5-7.45" ~ (H20). Anal. Calcd. for C I ~ H I O N Z O ~ . HC, B ~43.2; : H , 4.8; Br, 22.1; N , 7 . 8 . Found: C,43.2; H,5.2; Br,21.8; N,7.7. Diethyl p-Methylaminobenzoyl-L-glutamate.-Diethyl N{ p - [( carbobenzyloxy )methylamino] benzoyl 1 -L-glutamate ( 1 g., 2.4 mmoles) and 3.5 ml. of 40% hydrobromic acid in glacial acetic acid were stirred a t 25" for 45 min., by the end of which time evolution of carbon dioxide had subsided. Anhydrous ether (30 ml.) was added and the reaction mixture was continuously stirred a t 4" for 20 hr. The ether was decanted and the residual semisolid was dried in vacuo over phosphorus pentoxide, concentrated sulfuric acid, and sodium hydroxide pellets. The crude diethyl p-methylaminobenzoyl-L-glutamate hydrobromide was obtained as a white hygroscopic powder. The crude product (0.65 9.) was dissolved in 8 ml. of absolute ethanol and filtered through Darco.8 To the clear solution were added 0.5 ml. of pyridine, a few crystals of sodium hydrogen sulfite, and 40 ml. of water. The crystallization of diethyl p-methylaminobenzoyl-L-glutamatewas complete a t the end of 72 hr. a t 0". The yield was 75%; white needles from either ethyl acetate-n-hexane (1 :5) or dilute ethanol, m.p. 89-91', [cY]"D -20.50" (1 N HC1); lit.' m.p. 89.8-91.0°, [cY]*'D -21' (1 N HC1). Anal. Calcd. for C ~ ~ H Z ~ N ZC, O ~60.7; : H , 7.2; N, 8.3. Found: C,61.0; H,7.3; N,8.5. This ester was also formed when a solution of 1 g. (2.8 mmoles) of p-methylaminobenzoyl-L-glutamic acid hydrobromide in 10 ml. of absolute ethanol saturated with hydrogen chloride was allowed to stand in a closed vessel for 36 hr. a t room temperature. After removal of the solvent under reduced pressure, the residue was treated with pyridine and then twice recrystallized from ethyl acetate-n-hexane (1 :5). The pure diethyl p-methylaminobenzoyl-L-glutamate was obtained in 60% yield; the physical properties were the same as those described above. Anal. Found: C, 60.9; H , 7.1; N,8.3.
Acknowledgment.-The authors wish to express their appreciation to Dr. Leonard T. Capell, Chemical Abstracts Service, for discussion of the nomenclature. (8) Darco, grade G-60, Atlas Chemical Industries, Wilmington, Del.
The Synthesis of Cyanovinylguanidines WILLIAM J. FANSHAWE, VICTORJ. BAUER,h N D
s. R. S.4FIR
Organic Chemical Research Section, Lederle Laboratories, A Division of American Cyanamid Company, Pearl River, New York
VOL.30 HzN
HzN
\
/
C=N-CN
\
/
C=N-CH=CH-CN
HzN
HzN
I
I1
idines VI1 and VIII, substituted examples of structure 11, and our ,unsuccessful efforts to isolate I1 itself. The first synthetic route employed was predicated upon conjugate addition of tetramethylguanidine (111) to 3-chloroacrylonitrile. Single different products were obtained when trans- and cis-3-chloroacrylonitrile (IV and V) were used. The spin-spin coupling constants (14 and 7 c.P.s.) of the olefinic protons in the n.ni.r. spectra of the respective :isomeric products indicated2 that they were trans- and cis-2-(2-cyanovinyl)-l,l,3,3tetramethylguanidine (VI1 and VIII).
These stereochemical results imply that a likely course of reaction is addition of the guanidine to the acrylonitrile, followed by rotation about the central C-C bond of the anion to achieve the trans coplanarity necessary for elimination of chloride. The sequence is illustrated for the trans and cis isomers by eq. 1 and 2. These results are in complete accord with the. recently published observations of Scotti and Frasza4on the addition of other nucleophiles to 3-chloroacrylonitrile. A second synthetic route consisted of the addition of tetramethylguanidine to propiolonitrile (VI). As expected, the product formed by trans addition to the acetylenic bond5 was the cis-cyanovinylguanidine VIII. trans-Cyanovinylguanidine VI1 formed a crystalline hydrochloride. Evidence for protonation on the central nitrogen (IX) was found in the ultraviolet spectrum, which showed a marked hypsochromic shift relative to the base. The trans isomer VI1 appears to be more stable than cis isomer VIII, as shown by the conversion of VI11 to VI1 with basic alumina.
Received December 1 , 1964
Cyanoguanidine (I) has enjoyed considerable utility as a reagent in organic synthesis.' The vinylogous cyanoguanidine system I1 is, however, unknown. In this communication, we describe the preparation and properties of tmns- and cis-tetramet,hylcyanovinylguan(1) V. Migrdichian, "The Chemistry of Organic Cyanogen Compounds," Reinhold Publishing Corp., New York. N. Y.,1947.
(2) I t has been shown8 t h a t t h e magnitude of the spin-spin splitting cons t a n t falls between 11 and 18 c.p.s. for trans olefinic protons and between 6 and 14 C.P.S. for cis olefinic protons. When values for a pair of isomers fall outside the range of overlap. structural assignments may be made with reasonable certainty. (3) L. M . Jackman. "Applications of Nuclear Magnetic Resonance Spectroscopy in Organic Chemistry," Pergamon Press Ltd., London, 1959, p. 85. (4) F. Scotti and E. J. Frazza, J . Org, Chem., 1 9 , 1800 (1964). (5) E. L. Eliel, "Stereochemistry of Carbon Compounds." McGraw-Hill Book Co., Inc., New York, N. Y., 1962,p. 349.
NOTES
APRIL 1965
NC
G
cTH -
I
An attempt to synthesize the parent cyanovinylguanidine I1 from IV and guanidine was unsuccessful; the only product isolated was 2,4-diaminopyrimidine (X). Apparently addition-elimination occurred and was followed by cyclization to the aromatic system. HzN \ HzN
C=NH
/
+
/H
c1\
/c=c\
H
CN
-
IV
H2N
I1
X Experimental6
trans-2-(2-Cyanoviny1)-1,1,3,3-tetramethylguanidine (VII).To a solution of 0.59 g. (5.0mmoles) of 1,1,3,3-tetramethylgiianidine in 5 ml. of benzene was slowly added 0.22 g. (2.5 mmoles) of trans-3-chl0roacrylonitrile.~After 16 hr. a t room temperature, the crystals, m.p. 202-205', of tetramethylguanidine hydrochloride which separated were removed by filtration. The filtrate was concentrated under reduced pressure to a solid, which upon recrystallization from ether yielded 0.15 g. (37%) of an off-white solid, m.p. 89-90.5'. Two recrystallizations from ether afforded the analytical sample as long colorless prisms, m.p. 9&91", Ai:: 4.55 p , and A$': 298 mp (e 26,600). The n.m.r. spectrum (CDCl,) showed doublets at r 2.48 and 5.27 ( J = 14 c.P.s., one proton each) and a singlet a t 7.07 (12 protons). Anal. Calcd. for C8H14N4: C, 57.80; H, 8.49; N, 33.71. Found: C, 57.44; H , 8.31; N, 33.44. trans-2-(2-Cyanoviny1)-1,1,3,3-tetramethylguanidine Hydrochloride (IX).-An ethereal solution of 1.0 g. (6.0mmoles) of trans-2- (2-cyanovinyl)-lt1,3,3-tetramethylguanidine was acidified with ethereal hydrogen chloride. The precipitate was collected and recrystallized from acetone to yield 0.53 g. (43%) of cream-colored crystals, m.p. 185-187'. An additional recrystallization afforded the analytical sample as long regular prisms, 4.50 p , and At:;"" 258 mp (e 27,000). The m.p. 187-188", (6) Melting points w'ere determined in a Hershberg apparatus and are uncorrected. Microanalyses were performed by Mr. L. M. Brancone and staff. Ultraviolet and n.m.r. spectra were determined by Mr. W. Fulmor and staff.
1279
n.m.r. spectrum (DMSO-&) showed doublets a t r 2.33 and 4.25 ( J = 14 c.P.s., one proton each) and asinglet a t 6.88 (12protons). Anal. Calcd. for CSHISCIN~:C, 47.41; H , 7.41; C1, 17.53; N, 27.65. Found: C, 47.20; H, 7.41; C1, 17.31; N, 27.15. eis-2-(2-Cyanovinyl)-lI1,3,3-tetramethylguanidhe(VIII). A . -To a cold, stirred solution of 14.7 g. (0.13 mole) of 1,1,3,3tetramethylguanidine in 100 ml. of benzene was added dropwise under nitrogen a solution of 5.5 g. (0.065 mole) of cis-3-chloroacryl~nitrile.~The mixture was stirred at room temperature for 16 hr., the tetramethylguanidine hydrochloride which separated was removed, and the solution was concentrated under reduced pressure to a brown, oily solid. Three recrystallizations from ether yielded 3.1 g. (29%) of regular prisms, m.p. 53-56", Amax KBr 4.55 p , and 299 mp (e 23,400). The n.m.r. spectrum (CDCI,) showed doublets a t r 2.88 and 5.78 ( J = 7 c.P.s., one proton each) and a singlet a t 7.10 (12 protons). The melting point was depressed to 48-53' upon admixture with the trans isomer VII. Anal. Calcd. for C ~ H I ~ V ,C, : 57.80; H , 8.49; N, 33.71. Found: C, 57.48; H , 8.57; N,33.34. B.-To a cold, stirred solution of 1.7 g. (0.033mole) of propiolonitrileT in 40 ml. of benzene was added dropw-ise under nitrogen a solution of 3.8 g. (0.033 mole) of 1,1,3,3-tetramethylguanidine in 30 ml. of benzene. The mixture was stirred a t room temperature for 16 hr. and filtered. The filtrate was concentrated under reduced pressure to a brown tar. Two recrystallizations from ether yielded 0.25 g. (4.57') of long colorless prisms, m.p. 5559'. Anal. Calcd. for C ~ H I ~ NC, ~ :57.80; H , 8.49; N, 33.71. Found: C, 57.80; H, 8.57; N, 33.93. The n.m.r. spectrum of the compound was identical with that of the product of method A. Isomerization of cis- to trans-2-(2-Cyanovinyl)-l,1,3,3-tetramethy1guanidine.-A mixture of 0.12 g. of cis-2-(2-cyanovinyl)1,1,3,3-tetramethylguanidine,0.2 g. of activated alumina, and 50 ml. of ether was heated under reflux for 5 hr. and filtered. The filtrate was concentrated under reduced pressure to 0.12 g. of liquid which crystallized on standing to a pale yellow solid, m.p. 84-85'. The melting point was not depressed upon admixture with trans-2-(2-cyanovinyl)-l,l,3,3-tetramethylguanidine. The n.m.r. spectrum of the product was identical with that of VI1 prepared above. 2,4-Diaminopyrimidine (X).-To a cold, stirred suspension of 3.0 g. (0.05 mole) of guanidine8 in 150 ml. of acetonitrile was added dropwise a solution of 2.2 g. (0.025mole) of trans-3-chloroacrylonitrile in 60 ml. of acetonitrile. The brown mixture was stirred overnight a t room temperature and filtered. The filtrate was concentrated under reduced pressure to 3.5 g. of a brown solid, which was extracted with chloroform. Concentration of the chloroform solution left 0.45 g. of a tan solid, m.p. 122-140'. Three recrystallizations from isopropyl alcohol-hexane afforded 149-150"). colorless microcrystals, m.p. 147-149' (lit.8 m. The ultraviolet spectrum of the compound, A$: 284 mp ( e 6400), was identical with that of authentic 2,4-diaminopyrimidine. (7) C. Moureu and J. C. Bongrand, A n n . cham. (Paris), [9]14, 47 (1920). (8) W. Marckwald and F. Struxe, Chem. Eer., 65, 457 (1922). (9) J. P. Engllsh and J. W. Clapp, U. S. Patent 2,416,617 (1947).
The Preparation of 2-Methyl-1-phenylbenzimidazole 3-Oxide JOHNW. SCHULENBERG AND S. ARCHER Sterling-Winthrop Research Institute, Rensselaer, New York Received October 26, 1964
Several years ago we prepared 2-chloro-2'-nitro-Xphenylacetanilide (I) and hydrogenated it with platinum in ethanol. A hydrochloride was obtained in good yield. On the basis of analytical results, mechanistic considerations, and the absence of a carbonyl
1280
NOTES
I11
-
I
VJI, X = H VIII, x=c1
1
hydrochloride of VI was obtained. On the other hand, when 1 mole or more of acid was present during the reduction, the hydrochloride of the N-oxide I1 resulted in good yield. Therefore, acid can catalyze cyclization of the intermediate hydroxylamine (VII) to the point where ring closure is faster than reduction to the primary amine.4 In the hydrogenation of the chloroacetyl derivative I, the hydrogen chloride generated by reductive dehalogenation must catalyze ring closure of VI1 and VIII. This catalytic dechlorination probably can occur a t several stages, since we have found that 2-chloro-N1N-diphenylacetamide and 2-(chloromethyl)benzimidazole can be hydrogenolyzed to N,N-diphenylacetamide and 2-msthylbenzimidazole hydrochloride, respectively. Experimental6
Ha-Pt EtOH.HC1
0 I1
NH*
V
VOL.30
VI
band in the infrared spectrum, the structure 2-methyl1-phenylbenzimidazole 3-oxide (11) was tentatively assigned. This could have resulted from reduction to the hydroxylamine stage, cyclization, dehydration to the Noxide, and reductive dehalogenation (not necessarily in this order). One alternative structure, l-phenylbenzimidazole-2-methanol (111), was eliminated by synthesis. Recently, a series of benzimidazole N-oxides was synthesized by Takahashi and Kano by reduction of onitroanilides with hydrogen sulfide and ammonia.' Included in their report was the preparation of I1 from 2'-nitro-N-phenylacetanilide (IV). We repeated this reaction and obtained material identical with our hydrogenation product, confirming the postulated structure. It was of interest to see if IV could be converted directly to the N-oxide I1 by catalytic hydrogenation. Under neutral conditions, IV has been reported to furnish the expected reduction product V.lg2 It was claimed in one paper that hydrogenation of IV with platinum in 50% ethanol gave V but the same reaction in ethanol afforded 2-methyl-1-phenylbenzimidazole (VI).s However, we have carried out this hydrogenation in Soy0 ethanol, 95% ethanol, absolute ethanol, and ethyl acetate and obtained V in good yield in all cases. The amide V can be converted to VI either by refluxing in xylene or by treatment with hydrochloric acid at room temperature. When the hydrogenation of IV was carried out in ethanol in the presence of 0.2 mole of hydrochloric acid and the product was then treated with excess acid, the ( 1 ) 5 . Takahashi and H. Ksno, Chem. Pharm. E d . (Tokyo), 11, 1375 (1963). (2) E.J. Forbes and R . T . Wragg. Tefrahedron,8 , 79 (1960). (3) P . A. S. Smith, B. B. Brown, R. K. Putney. and R. F . Reinisch, J . A m , Chem. Soc., 1 6 , 6335 (1953).
2-Chloro-2'4tro-N-phenylacetanilide (I).-A mixture of 107 g. (0.5 mole) of o-nitrodiphenylamine, 75 ml. (113 g., 1 mole) of chloroacetyl chloride, and 150 ml. of toluene was refluxed 4.5 hr. The solvent was removed in vacuo and the dark red oil was crystallized from absolute ethanol to give an orange solid. Recrystallization from absolute ethanol afforded 100 g. (69y0) of 232 mp ( E 16,300) and 300 yellow prisms: m.p. 122-124'; , , ,A m p ( ~ 1 6 0 0 ) a; n d 5 . 9 1 ~ . Anal. Calcd. for Cl4Hl1C1N2O3:C, 57.84; H , 3.81; C1, 12.20. Found: C, 57.72; H , 3.94; C1, 12.21. 2'-Nitro-N-phenylacetanilide (IV).@-A mixture of 21.4 g. (0.1 mole) of o-nitrodiphenylamine, 15 ml. (16.5 g., 0.21 mole) of acetyl chloride, 0.5 g. of zinc chloride, and 20 ml. of ben ene was stirred and refluxed 3 hr. After careful addition of 75 ml. of absolute ethanol, the product crystallized as a yellow solid, m.p. 133-136.5", yield 21.3 g. (83%). Recrystallization from absolute ethanol gave 18.6 g., m.p. 134.5137" (lit.6 m.p. 134135'). 2 '-Amino-N-phenylacetanilide (V).-A mixture of 5.12 g. (0.02 mole) of recrystallized 2'-nitro-N-phenylacetanilide (IV), 100 ml. of ethyl acetate, and 200 mg. of platinum oxide was shaken under hydrogen on a Parr hydrogenator for 30 min. Removal of catalyst and solvent left a yellow oil which was crystallized from benzene-hexane to give 3.70 g. (82Oj,) of white solid, m.p. 112-1 15' (lit.2,3m.p. 115-1 16'). Recrystallization from benzene-hexane furnished the analytical sample: m.p. 112.5-116.5'; Amax 296 mp ( e 3300); 2.88, 2.96 and 6.00 p ; 6 = 7.17-7.87 (aromatic), 4.45 (NH,), and 2.55 p.p.m. (CH3). And. Calcd for ClrHlaNzO: C, 74.31; H , 6.24; N, 12.38. Found: C, 74.73; H , 6.40; N , 12.24. Similar hydrogenations in absolute ethanol and 9.570 ethanol gave yields of 80-83y0 while a 71% yield was obtained when 50y0 ethanol was used. Melting points and spectra showed that V, rather than 2-methyl-1-phenylbenzimidazole (VI), wm the product in each case. In some runs, a lower-melting form of V resulted. The melting point, 105-108O, was unchanged on recrystallization from benzene-hexane. When recrystallized again from the same solvent pair, but seeding with the higher-melting form, colorless rods, m.p. 112.5-116', resulted. The infrared spectra of the two forms differed when run as solids (KBr), but chloroform solutions gave identical curves. The n.m.r. spectra (deuteriochloroform) of the two solids were identical.
(4) Catalytic hydrogenation of o-nitrophenylalanine has recently been reported to give similar results. The free base yields o-aminophenylalanine while the hydrochloride gives 3-amino-3,4-dihydro-l-hydroxycarbostyril: A. L. Davis, 0. H. P . Choun, D . E. Cook, and T. J. MoCord, J . Med. Chem., 1 , 632 (1964). (5) All melting points were determined in capillaries and are corrected. Ultraviolet spectra were run in 95% ethanol (Cary spectrophotometer), infrared spectra in potassium bromide disks (Perkin-Elmer 21), and n.m.r. spectra in deuteriochloroform using tetramethylsilane as the external standard (Varian A-60). We thank Dr. F . C. Nachod, Dr. R . K. Kullnig. Miss C. Martini, and Mr. M. Priznar for spectral determinations and Mr. K. D . Fleischer and staff for analytical results. (6) F . Kehrmann and E. Baumgartner, Helu. Chim. Acta, 9, 673 (1926). The literature procedure uses acetic anhydride.
APRIL1965
NOTES
2-Methyl-1-phenylbenzimidazole (VI).7-A solution of 2.26 g. (0.01 mole) of 2'-amino-N-phenylacetanilide(V) in 10 ml. of dry xylene was refluxed for 16 hr.'l*lg Solvent removal loft a yellow oil which was crystallized from benzene-hexane to give 1.36 g. (65%) of pale tan crystals: m.p. 70.5-72.5', lit.8 m.p. 72-73'; 6 = 7.50-8.33 (aromatic) and 2.87 p.p.m. (CH,). Hydrochloride. A.-A portion of the above base VI was dissolved in 2-propanol and treated with excess ethanolic HCI t o give a white solid, m.p. 220-223'. Recrystallization from 2propanol gave the analytical sample, m.p. 222-225". Anal. Calcd. for C1dH13ClNz: C, 68.71; H , 5.35; N , 11.45. Found: C, 68.77; H , 5.13; N , 11.28. B.-A solution of 0.9 g. (0.004 mole) of a'-amino-N-phenylacetanilide in acetoneether waa treated with 0.008 mole of ethanolic HC1 and kept overnight a t room t e m p e r a t ~ r e . ' ~ White needles began to crystallize about 1 hr. after the HCl was added. The product, 0.6 g. (60%), m.p. 220.5-223.5', waa shown by infrared spectrum and mixture melting point to be identical with material obtained by procedure A.l* The reaction could be carried out, with comparable results, on crude V obtained directly from hydrogenation of IV. 2-Methyl-1-phenylbenzimidazole 3-Oxide Hydrochloride (II.HC1). A.-In a typical run, a mixture of 14.55 g. (0.05 mole) of 2-chloro-2'-nitro-N-phenylacetanilide( I ) , 200 mg. of platinum oxide, and 200 ml. of absolute ethanol was shaken under hydrogen on a Parr hydrogenator for 30-60 min. The hydrogen uptake waa about 0.16-0.17 mole. The catalyst and solvent were removed and the resulting reddish brown gum or solid was crystallized from acetone-ether to give, in one case, 9.0 g. (69%) of off-white solid, m.p. 204-209' dec. Recrystallization from acetone gave white prisms, m.p. 205-210' dec. Anal. Calcd. for CldHlaCINZO: C, 64.49; H , 5.02; N, 10.75. Found: C,64.52; H,5.12; N, 10.96. Authentic hydrochloride waa prepared from I1 (made by the literature procedure)' and found to melt a t 202-207' dec. Infrared comparison and mixture melting point showed that the samples were identical. In other runs, yields of 51-75% were obtained. In some runs, lower melting points (around 192-204') resulted and the product waa more difficult to purify. Carrying out the hydrogenation with added ethanolic HCl had little effect except that the yields were slightly lower and the products were somewhat darker. The product could also be obtained by carrying out the hydrogenation with platinum in ethyl acetate or palladium on carbon in ethanol. The product deteriorates slowly on long standing (several years) a t room temperature. B.-A mixture of 7.68 g. (0.03 mole) of 2'-nitro-N-phenylacetanilide (IV), 200 mg. of platinum oxide, 30 ml. of ethanolic HCl (0.065 mole), and absolute ethanol (total volume 200 ml.) was hydrogenated on the Parr apparatus for 30 min. About 10% more than the theoretical amount of hydrogen waa absorbed. Removal of catalyst and solvent and crystallization of the residue from acetoneether gave 4.45 g. (57yo) of pale pink solid, m.p. 204-210". The material waa shown to be identical with the product obtained from I (procedure A, above) by infrared comparison and mixture melting point. When the hydrogenation was carried out using equimolar amounts of IV and HC1, a 62% yield of white solid, m.p. 2W205', resulted. When the hydrogenation was carried out with 0.02 mole of IV and 0.004 mole of HCl, the crude product then being treated a t (7) This compound has previously been prepared from 2'-anilinoacetanilide by treatment with acid,'-*O from N-phenyl-o-phenylenediamine, acetic anhydride, and aqueous hydrochloric acid,fi and from compound V and aqueous acid.' (8) L. Wol5, Ann., 894, 59 (1912). (9) M.A. Phillips, J . Chem. ~ o c . 2820 , (1929). (10) F . Hunziker, F. KILnrle, 0. Schindler, and J . Schmute, Helu. Chim. Acfo,47,1163 (1964). (11) The uncatalyred cyclization does not take place readily. Refluxing for 2 hr. in xylene gave a mixture which still contained starting material while heating V without solvent a t 150-15S0 for 4 hr. gave back over 50% of unrescted amide. On the other hand, a sample of V which had been kept in the dark at room temperature for 5.5 years had completely cyclized t o VI. (12) The isomeric amide, 2'-anilinoacetanilide,avfigave chiefly recovered starting material under these conditions. (13) We were unable to induce the hydrochloride of V to crystallize. It is possible that the reported hydrochloride of V, m.p. 210-21lo. ia actually that of the benzimidazole VI; no analytical values were given.' (14) 2'-Anilinoacetanilide%ggave the hydrochloride of VI in comparable yield under theae conditions.
1281
room temperature with excess ethanolic HCl in acetone, 2-methyl1-phenylbenzimidazole hydrochloride (VI.HCl), m.p. 218.5224.5', was obtained in 48% yield. This product depressed the melting point of the N-oxide hydrochloride (I1.HCl), but not that of authentic VI.HC1 (above). Infrared comparison confirmed the assigned structure. 2-Methyl-1-phenylbenzimidazole 3-Oxide GI).-The hydrochloride of I1 (obtained by procedure A) was treated with aqueous potassium carbonate and the freed base was extracted with chloroform. A yellow solid was obtained and recrystallized from ethyl acetate to furnish almost white prisms (79'%), m.p. 161167" dec. Another recrystallization from ethyl acetate, followed by drying in vacuo a t 100' (darkened) gave the analytical sample: m.p. 165-170' dec., lit.' m.p. 164-165'; A, 283 mp ( t 8000); 8 = 7.75-8.75 (aromatic) and 3.17 p.p.m. (CHa). The product was compared with authentic material prepared by the literature procedure.' The two samples had identical infrared and ultraviolet spectra and a mixture of the two gave no depression in melting point. Anal. Calcd. for C11H12N20: C, 74.99; H , 5.38; N , 12.49. Found: C, 75.42; H , 5.14; N , 12.39. In one run, the product, after drying a t only 75", melted a t 127-130' dec. The analyses suggested the formula C l ~ H l z N z O ~ 0.75H20, but this hydrate has been reported to melt a t 97-100'.1 2'-Anilino-2-hydroxyacetanilide.-To a solution of 1.5 g. (0.02 mole) of glycolic acid in 10 ml. of methanol was added 3.7 g. (0.02 mole) of N-phenyl-o-phenylenediamine in 10 ml. of chloroform. Next, 4.5 g. (0.022 mole) of dicyclohexylcarbodiimide in 5 ml. of chloroform was added. After standing overnight at room temperature, the precipitated dicyclohexylurea was filtered off and the solvent was removed from the reddish purple filtrate. The residual gum was taken up in 2-propanol, additional dicycloand the filtrate was diluted hexylurea was filtered off (total 8OyO), with hexane to give 1.6 g. (33%) of light brown product, m.p. 144-149.5'. This was recrystallized from ethyl acetate to give an almost white solid: m.p. 148-151'; A,, 282 mp ( e 13,300); and 3.02 and 6.03 p . Anal. Calcd. for C1~HlrNnOz:C , 69.40; H , 5.83; N, 11.56. Found: C, 69.55; H , 5.96; N, 11.89. 1-Phenylbenzimidazole-2-methanol (III) .-A mixture of 5.9 g. (0.024 mole) of 2'-anilino-2-hydroxyacetanilide (m.p. 143150'), 5 g. (0.026 mole) of p-toluenesulfonic acid hydrate, and 250 ml. of toluene waa slowly distilled over a 3-hr. period. After cooling, aqueous sodium hydroxide was added and the freed base was extracted with ether. Solvent removal left a brown solid which was recrystallized from benzene to give 4.3 g. (87%) of white product, m.p. 131-133'. Anal. Calcd. for C1~HI2NzO:N, 12.49; NAP, 6.25.lS Found: N, 12.38; NAP,6.17. The hydrochloride was prepared by adding ethanolic HCl to a solution of the base I11 in acetone. Upon recrystallizing from 2propanol-acetone, small white needles, m.p. 192-197', were obtained. A mixture melting point with the N-oxide hydrochloride (II.HC1) showed a marked depression. Anal. Calcd. for ClrHISCINIO: N, 10.75. Found: N, 10.53. N,N-Diphenylacetamide.-A mixture of 7.4 g. (0.03 mole) of 2 4 hloro-N,Ndiphenylacetamide,16 2 g. of 10% palladium on charcoal, and 140 ml. of absolute ethanol waa shaken on a Parr hydrogenator until the theoretical amount of hydrogen waa absorbed (5-6 hr.)." After removing the catalyst and solvent, the residue was crystallized from a small volume of 2-propanol to give 5.0 g. (79%) of white solid, m.p. 96-101". Identity with authentic diphenylacetamide waa proven by mixture melting point and infrared comparison. 2-Methylbenzimidazole Hydrochloride.-Hydrogenation of 5 g. (0.03 mole) of 2-(chloromethyl)benzimidazolewith 200 mg. of platinum oxide in 95 ml. of absolute ethanol waa complete in 2.25 hr. The product was crystallized from ethanol-ether to give 2 g. (40%) of tan solid, m.p. 293-298" dec. Comparison with authentic material by mixture melting point and infrared spectra proved the structure. (15) NAP refers to determination of basic nitrogen by titration with perchloric acid in acetic acid. (16) H. Frerichs, Arch. Pharm.. 441, 220 (1903). (17) When platinum we8 used, results were erratic. When the reaction we8 stopped after 1 mole of hydrogen had been t,aken up, over 60% of recovered starting material we8 obtained. Presumably, reduction of the aromatic rings occurred.
VOL. 30
NOTES
1282 Synthesis of a Thionobenzoate in the PolyhydroxytetrahydropyranSeries E. J. H E D G L E Y AND ~ ~HEWITTG. FLETCHER, JR.’~
Chemistry Department, Birkbeck College (University of London), London, and National Institute of Arthritis and Metabolic Diseases, National Institutes oj Health, Public Health Service, U.S. Department of Health, Education, and Welfare, Bethesda, Maryland
of 1,5-anhydro-3,4-0-isopropylidene-~-arabinitol afforded the 2-0-benzyl ether from which the isopropylidene group was removed by acid hydrolysis. The dibenzoate of the product, 1,5-anhydro-3,4-di-O-benzoyl2-O-benzyl-~-arabinitol, was hydrogenated to give 1,5anhydro-3,4-di-O-benzoyl-~-arabinitol;thiobenzoylation of this diester afforded the desired lJ5-anhydro3,4-di-O-benzoyl-2-O-thiobenzoyl-~-arabinitol. Possessing the C=S chromophore, thionobenzoates , ~ which markare characteristically yellow in C O I O a~ fact edly facilitates the chromatography of these substances.
Received December I , 1964 Experimental”
Although little investigated, thiono esters [RC(=S)Thiobenzomorpho1ide.-The substance was made from benzOR’] are normally prepared either through treatment aldehyde (42.5 g., freshly distilled), morpholine (52.0 g.), and of imino ethers with hydrogen sulfide2,3or by reaction powdered sulfur (19.2 g.) as described in “Method A ” of Peak of a chlorothionoformate with a Grignard reagent.2 and Stansfield,Eand the powdered, crude product was extracted Owing the the labilities of many masking groups which in a Soxhlet apparatus with ca. 750 ml. of ethanol, the contaminating sulfur being left behind. On cooling the ethanolic soluare normally used in the carbohydrate field, both of tion, the product separated as large prismatic needles: 78.2 g. these procedures are subject to limitations in the syn(94%), m.p. 137-138’. thesis of thiono esters of the sugar series. In the course Methyl Dithiobenzoate.-Thiobenzomorpholide (91.O 9.) was of studies of the rearrangement of esters4 it became dedissolved in 900 ml. of hot, anhydrous acetone, and the solution was cooled rapidly under the tap. Methyl iodide (40 ml.) was sirable to synthesize certain thionobenzoates of polyadded through an efficient reflux condenser, and the solution was hydroxytetrahydropyrans (1,5-anhydroglycitols) and rewarmed a t gentle reflux with vigorous stirring. After about 5 we therefore sought a method for the preparation of min., a yellow solid rapidly separated from the solution; the thionobenzoates which would be suitable for use in the suspension was heated for a further 30 min., cooled, diluted with carbohydrate field. Thiobenzoylation with thio150 ml. of dry pyridine, and stirred while a gentle stream of hydrogen sulfide was passed in. The reaction mixture turned benzoyl chloride in pyridine solution proved to be satisorange and then to the red color characteristic of methyl dithiofactory. However, while the reagent is readily prebenzoate, the yellow solid gradually being replaced by a colorless parable from dithiobenzoic acid as described in the literone.l* A slow stream of hydrogen sulfide into the reaction a t ~ r e ,the ~ , standard ~ preparation of dithiobenzoic acid mixture (at room temperature) was maintained overnight and the solution was then filtered, the solid being washed with from phenylmagnesium bromide and carbon disulfide5-’ absolute ethanol until colorless. The combined filtrate and proved troublesome in our hands. Methyl dithiobenwashings were concentrated in vacuo (50’ bath) to remove the zoate, a distillable red oil which is stable on storage, may acetone, ethanol, and the major part of the pyridine, solid separeadily be prepared from thiobenzoinorpholide methrating toward the end of the concentration. After cooling in ice, the magma was acidified with an excess of ice-cold 5 N hydroiodides and serves as a reliable source of dithiobenzoic chloric acid and the aqueous suspension was extracted with ether acid as the latter is required. Simple treatment of the (a total of 500 ml.) until no more color was removed. The commethyl ester with sodium hydrogen sulfide in dry methbined ethereal extracts were washed thoroughly with water and, anol, followed by acidification gives dithiobenzoic acid finally, aqueous sodium bicarbonate solution. Moisture was which need not, however, be isolated and can be conremoved with magnesium sulfate and the solution was concentrated in vacuo, the residue being distilled a t 90’ and 0.6 verted directly to thiobenzoyl chloride. The desired ester, 1,5-anhydr0-3,4-di-O-benzoyl-2- to give 67.3 g. (91yo) of methyl dithiobenzoate as a mobile red oil. Until required, the substance was stored in stoppered, 0-thiobenzoyl-D-arabinitol was synthesized in the foldark-colored bottles a t room temperature. lowing fashion. lJ5-Anhydro-~-arabinitola was conDithiobenzoic Acid.-Methyl dithiobenzoate (67 9.) was added to a solution of anhydrous, freshly prepared sodium hydrogen densed with acetone,lO the sirupy product, lJ5-anhydrosulfide (25 9.) in absolute methanol (150 ml.). The mixture 3,4-0-isopropylidene-~-arabinitol, being characterized was stirred at room temperature until homogeneous (2 hr.) as its crystalline 2-0-benzoyl derivative. Benzylation and then left overnight a t room temperature. Concentration (1) (a) Birkbeck College; (b) National Institutes of Health. (2) Houhen-WeyL “Methoden der organischen Chemie,” Vol. I X , E. Muller, Ed., Georg Thieme Verlag, Stuttgart, 1955,p. 759; E. E. Reid, “Organic Chemistry of Bivalent Sulfur,” Vol. IV, Chemical Publishing Co., Inc., New York, N. Y., 1962. (3) U.Schmidt, E . Heymann, and K . Kabitske, Ber., 96,1478 (1963). (4) E. J. Hedgley and H. G. Fletcher, Jr., J. A m . Chem. Soc., 86, 1615 (1963). (5) H . Staudinger and J. Siegwart, Helu. Chim. Acta, 3, 824 (1920); T . Bacchetti and A. Alemagna, Rend, I s t . Lombardo Sei. Pt. I , Classe Sei. Mat. e. Not., 91,617 (1957); Chem. Abstr.. 68, 6217 (1959). (6) F. Bloch. Compt. rend., 904,1342 (1937). (7) J. Houben, Ber., 39,3219 (1906). (8) D.A. Peak and F. Stansfield, J . Chem. Soc., 4067 (1952). (9) H . G. Fletcher, J r . , and C. S. Hudson [ J . A m . Chem. Soc.. 69, 1672 (1947)] prepared this substance through the desulfurization of various aryl 1-thio-D-arabinopyranoside triacetates with Raney nickel, We have, however, found t h a t the substance is more readily accessible through the catabromide, a lytic reduction of 2,3.4-tri-0-beneoyl-~-~-ar&hinopyranosyl synthetic procedure first introduced by L. Zervas and C. Zioudrou [J. Chem. Soc., 214 (195611. (10) p-Toluenesulfonic acid was used as a catalyst, and Molecular Sieve, in a Soxhlet extractor, was employed as a desiccant. I t is believed t h a t this technique may have general applicability for such condensations.
in vacuo (45” bath) afforded a very dark red residue which waa dissolved in 200 ml. of water, and the resulting solution was extracted with dichloromethane until the extracts were virtually colorless. Ether (250 ml.) and then ice-cold 5 N hydrochloric acid (150 ml.) were now cautiously added to the aqueous solution, and the system was shaken until the aqueous layer was colorless. The red-violet ethereal solution was washed twice with water and then dried with sodium ~u1fate.l~Dithiobenzoic acid (41 g., 67y0) was obtained as a red-brown oil on concentrating the solution under nitrogen at 0’; it was characterized through the formation of its lead salt, obtained as purple-red needles, m.p. 204-205’, from toluene. The lead salt may also be used for further purification of the dithiobenzoic acid.’J6 (11) Melting points are corrected. (12) Care must be taken to avoid clogging of the hydrogen sulfide delivery tube by t h e precipitate. (13) Also noted were h.p. 80-85’ (0.3mm.) and h.p. 76O (0.2 mm.). (14) I t is advantageous t o use this solution directly for the preparation of thiobeneoyl chloride, without isolating the dithioheneoic acid. (15) T . G. L,evi, Gatz. chim. ital., 61, 665 (1931); Chem. Absti., 96,1250 (1932).
APRIL1965
NOTES
1283
Anal. Calcd. for Cz6H2406 (432.45): C, 72.21; H , 5.59. Thiobenzoyl Chloride.-Dithiobenzoic acid (37.5g.) was disFound: C, 72.28; H , 5.78. solved in ether (50ml.) and the solution was treated with thionyl 1,5-Anhydro-3,4-di-O-benzoyl-~-arabinitol .-Palladium chlochloride (ca. 40 ml.) under conditions defined by earlier workers's6 ride (1 9.) was suspended in ethanol (15 ml.) and reduced with to give thiobenzoyl chloride (23.1 g., 61%) as a violet-red, hydrogen a t room pressure. 1,5-Anhydro-3,4-di-O-benzoyl-2-0lachrymatory liquid, b.p. 88" (3.75mm.). benzyl-D-arabinitol (2.16 g.), dissolved in 100 ml. of ethanol, 1,5-Anhydro-2,3,4-tri-O-benzoyl-~-arabinitol.-a-~-~rab~nowas added to the suspended catalyst and the mixture was agipyranose tetrabemoatel6 (22.64 9.) was warmed gently with 40 tated with hydrogen until no more gas was absorbed. The ml. of a 30% solution of hydrogen bromide in glacial acetic acid catalyst w&s removed by filtration and the solution was conuntil 2,3,4-tri-'0-benzoyl-8-~-arabinopyranosylbromide began centrated i n vacuo, finally a t 80" (bath) to give 1.68 g. (98%) to crystallize from the clear solution. The mixture was then of crude crystalline product, m.p. 134-136'. Recrystallized stirred a t room temperature for 2.5 hr., and the solution was from ethanol-hexane and sublimed a t 135" and 0.01 mm., decanted from the crystalline halide which was washed three the ester melted a t 134-135" and showed [ C Y ] ~ O D -212' (c 0.85, times (by decantation) with petroleum ether b.p. (60-70'). ethanol). Dissolved in 250 ml. of ethyl acetate, the halide was treated Anal. Calcd. for C18H1806 (342.33): C , 66.66; H, 5.30. with 2.0 g. of 10% palladium on charcoal (presaturated with Found: C, 66.39;H , 5.52. hydrogen) and 10.0 g. of triethylamine; on shaking with hydro1,5-Anhydro-3,4-di-O-benzoyl-2-O-thiobenzoyl-~-arabinitol.gen, the solution absorbed the theoretical amount of hydrogen To a solution of 3.4 g. of 1,5-anhydro-3,4-di-O-benzoyl-~-arabiin 5.8 hr.; after removal of the catalyst by filtration, the solunitol in 15 ml. of dry pyridine (5.4 g., 3.5 molar equiv.) of thiotion was washed with aqueous sodium bicarbonate and concentrated to give 16.29 g. (86YG)of 1,5-anhydro-2,3,4-tri-O-benzoyl- benzoyl chloride was added in one portion without cooling. A viscous sirup separated and the medium assumed a red-brown D-arabinitol, m.p. 119". Recrystallization from methanol coloration distinct from the violet-red color of the reagent. afforded pure material: m.p. 119-120", [a]% -219' ( c 0.11, After 12 hr., water (1 ml.) was added, then, after a further 30 CHCls). Fletcher and Hudson0 reported m.p. 120-121" and [a]zoD -220' (CHCl,) for 1,5-anhydro-2,3,4-tri-O-benzoyl-~- min., the pyridine was removed by evaporation a t 80". The residue was dissolved in dichloromethane and the solution arabinitol. 1,5-Anhydro-2-O-benzoyl-3,4-O-isopropylidene-~-arabinitol.- washed successively with dilute hydrochloric acid, water, and 1,5-Anhydro-2,3,4-tri-O-benzoyl-~-arabinitol was debenzoylated saturated aqueous sodium bicarbonate. After removal of in conventional fashion using ammonia in methanol to give 1,5moisture with sodium sulfate, the solution was concentrated anhydro-D-arabinitol as a sirup which crystallized on seeding. to a red, viscous oil (7.1 g.). With benzene as a solvent, thinWithout further purification, the anhydride (15.6g.) was suslayer chromatography on silica gel resolved the product into four pended in 200 ml. of anhydrous acetone containing 0.2 g. of p components, a fast moving red fraction (Rf -0.9)) two closely toluenesulfonic acid. The solution was boiled for 16 hr. under a associated orange and yellow fractions (Rf -0.5), and one Soxhlet extractor holding Molecular Sieve, type 5A, and then fraction which failed to migrate. Chromatography of 1 .O neutralized with ammonia. After filtration, the solution w&s g. of the mixture on a short column of silica gel (10 g., packed concentrated and the residue was distilled i n vacuo to give 1,5as a slurry in benzene) effected an excellent resolution of the anhydro-3,4-0-isopropylidene-~-arabinitolas a yellow oil: orange (0.039 . ) and yellow (0.40 g.) components. The charb.p. 78-80", (0.1-0.05mm.), 19.05 g. (94oj,) A sample of the acteristic color, optical activity, and infrared spectrum of the product was benzoylated with benzoyl chloride in pyridine to latter identified it as the required thionobenzoate; yield was give 1,5-anhydro-2-O-benzoy1-3,4-O-isopropylidene-~-arabin~tol equivalent to 62%, based on the 1,5-anhydro-3,4-di-O-benzoyl&s fine needles from ethanol: m.p. 102-103", [ a I m ~ -111' ( c D-arabinitol. Thin layer chromatography of the product thus 0.50,CHzCL). isolated revealed a trace of a contaminant which fluoresced Anal. Calcd. for ClbH1805 (278.31): C, 64.76; H , 6.52. under ultraviolet light. Final purihation by rapid distillation Found: C, 64.62;H , 6.76. a t 0.03 mm. and 220" (bath) afforded chromatographically 1,5-Anhydro-2-O-benzyl-3,4-C)-isopropylidene-~-arabinitol.- pure material which rotated [a]zo~ - 140.2 f 3.0' in ethanol (c 1,5-Anhydro-3,4-0-isopropylidene-~-arabinitol(3.36 9.) was 0.5): ultraviolet absorption data," : :A 282 mp (e 9680), mixed with 50 ml. of anhydrous tetrahydrofuran, 2.2 g. of pow292 mp (E 9400). dered potassium hydroxide,'? and 3.65 g. of benzyl chloride, and Anal. Calcd. for Cz6&?OeS (462.53): C, 67.52; H , 4.79. the suspension was boiled under reflux for 12 hr. Solid carbon Found: C, 67.81; H , 4.81. dioxide was added, the solution was concentrated to dryness, and Acknowledgment.-We are indebted to the staff of the residue was extracted with acetone. The sirup obtained on concentrating the extract was distilled to give a mobile yellow oil: the Section on Microanalytical Services and Instrub.p. 110-120' (0.1mm.), 4.5 g. (88%). A benzene solution of mentation (National Institute of Arthritis and Metathe crude product was passed through a short column of neutral bolic Diseases) for analyses and specific rotations. alumina and then concentrated, the sirupy residue being distilled ~ ( c 1.19, ethanol), nZoD1.5125. as before: 3.91 g., [ a ] %-57' Attempts to obtain the product in crystalline form were unsuccessful. An Improved Method for the Preparation of Anal. Calcd. for C1.,HzoOa (264.33): C, 68.16; H , 7.63. Found: C, 68.28;H , 7.50. Methyl 6-Chloro-6-deoxy-ru-~-glucopyranoside
1,5-Anhydro-2-O-benzyl-~-arabinitol.-l,5-Anhydro-2-O-benzyl-3,4-0-isopropylidene-~-arabinitol (22.4 9.) was dissolved H. B. SINCLAIR in 300 ml. of 50Oj, aqueous ethanol. Amberlite IR-12O(H) (5.09.) was added, and the solution was refluxed and stirred for Northern Regional Research Laboratory, 1 Peoria, Illinois 3 hr. After filtration and removal of the solvent, the product wm dried by azeotroping with benzene. Crystallization was spontaneous: 18.8 g. (99%), m.p. 99-100'. After two reReceived October 22, 1964 crystallizations from benzene, the product was obtained as fine ~ ( c 1.19, ethanol). needles: m.p. 101-102', [ a l Z o-35' The only reference found in the literature to the prepAnal. Calcd. for C1zH,eOa(224.25): C, 64.27; H , 7.19. aration of methyl 6-chloro-6-deoxy-~u-~-glucopyranoside Found: C, 64.44;H, 7.21. 1,5-Anhydro-3,4-di-O-benzoyl-2-O-benzyl-~-ara binitol .-1,5is by Helferich, Klein, arid Schaefer,2 who report an Anhydro-2-O-benzyl-~-arabinitol(18.80 9.) was benzoylated over-all 8% yield3 of methyl 6-chloro-6-deoxy-a-~-gluwith benzoyl chloride in pyridine to give a sirup which solidified ( 1 ) This is a laboratory of the Northern Utilization Research and Deto a crystalline mass: 34.83 g. (96%). Recrystallized twice velopment Division, Agricultural Research Service, U . S. Department of from petroleum ether (60-80"),the ester was obtained as needles: Agriculture. m.p. 62-63', [a]"D -155" (c 0.95,ethanol). (16) H. G . Fletcher, Jr., and C. S . Hudson, J . Am. Chem. SOC.,69, 1145 (1947). (17) Hooker Chemical Corp.. Niagara Falls, N . Y .
(2) B . Helferich, W. Klein, and W.Schaefer, Be?., 69B,79 (1926). (3) The yield reported in ref. 2 is much lower than 8 % , hut subsequent work by B . Helferich and H. Bredereck [Ber.. 60B,2002 11927)l on methyl 8-D-ghcopyranoside improved the chlorination step, and this improvement was used in the calculation.
NOTES
1284
copyranoside from methyl a-D-glucopyranoside. Their procedure probably discouraged further investigation of chlorinated hexoses for it not only resulted in a low yield, but also involved successively, tritylation with trityl chloride, acetylation with acetic anhydride, chlorination with phosphorus pentachloride, and hydrolysis with barium hydroxide. In contrast, a convenient procedure has been devised, namely, reaction of methyl a-D-glucopyranoside with sulfur monochloride (SzC12)and separation of the reaction products on a Darco G-60-Celite 5354 column. This improved procedure results in a 30-35% yield of methyl 6-chloro6-deoxy-a-~-glucopyranoside. HOCHi
ClCB
VOL.30
To eliminate the majority of contaminating Celite, the crystalline solid waa dissolved in boiling absolute ethanol (ca. 20 ml.) and filtered while hot. After concentrating the filtrate to dryness a t room temperature, a crystalline solid resulted containing a trace of Celite; this trace was easily removed by dissolving the crystals in boiling absolute ethanol (ca. 10 ml.), adding slowly ethyl acetate (ca. 50 ml.) to the boiling solution until a light tan floc separated, and then filtering the hot solution. From this filtrate, concentrated a t room temperature, separated crystalline’ 3.50 g. (32?&), methyl 6-chloro-6-deoxy-a-~-glucopyranoside, m.p. 110-112”. An analytical sample was prepared by recrystallization twice from ethanol-ethyl acetate, m.p. 111.2-112.0°, [ a ] ” D +141” (C 0.658, HzO)(lit.l m.p. 110-112’, [ a ] ” D f139’). Anal. Calcd. for C7H&105: C, 39.54; H, 6.16; CI, 16.67. Found: C, 39.75; H,6.10; Cl, 16.31. Linear horizontal chromatograph9 revealed only one spot with an Rf of 0.61 (methyl a-D-ghcopyranoside Rr 0.29) when the AgN08-NaOH dipping reagent of Smith waa used.@ (7) Occasionally an oil separates that readily crystallizes. ( 8 ) Paper was S and S 2043b; solvent was n - B u O H i - P r O H - H , O , 3: 1: 1
1
1
OH
I
(v./v.). For horizontal technique refer to R . G. Strobe1 and J. Holme, Cereal Chem., 40, 361 (1963). (9) I. Smith, “Chromatographic Techniques,” Interscienoe Publishers, Inc., New York, N . Y . , 1958, p. 169.
OH
When sulfur monochloride is mixed with methyl a-D-glucopyranoside in N,N-dimethylformamide, a slightly exothermic reaction occurs accompanied by the separation of sulfur. After the destruction of unreacted sulfur monochloride with water and removal of the sulfur by filtration, the filtrate is rapidly adjusted with sodium carbonate to pH 8. Following its concentration to a convenient volume, the solution is applied to a Darco G-60-Celite 535 c ~ l u m n . ~ Surprisingly, the chlorinated glucoside is strongly held on the Darco G-60; after elution with water to remove inorganic salts and with 5% ethanol to remove methyl a-D-glucopyranoside,6a t least 12% ethanol is required to elute the chloroglucoside. The yield of purified methyl 6-chloro-6-deoxy-a-~-glucopyranoside,m.p. 110-112”, is 30-35%. Experimental Methyl 6-Chloro-6-deoxy-~-~-glucopyranoside.-To a solution of methyl a-D-glucopyranoside (10 g., m.p. 164-166’) in N,N-dimethylformamide (250 ml., reagent grade used directly from the bottle) was added in one portion sulfur monochloride (15 ml., Eastman Kodak practical grade). Immediately, a canary yellow milkiness developed and the solution warmed slightly. The stoppered flask wa8 cooled in flowing tap water for 0.5 hr. and allowed to stand a t room temperature overnight (ca. 17 hr.) during which time the milkiness was replaced by crystalline sulfur. Water was then added with cooling in tap water as follows: Water, ml. Time, min.
2 0
2 35
2 45
44 60
50 105
50 115
After standing a t room temperature for 0.5 hr., the solution was filtered through a small pad of Celite 535 to remove sulfur; the filtrate was diluted with 750 ml. of water and adjusted to pH 8 (paper) with solid sodium carbonate. This solution was concentrated under vacuum (bath temperature ruma t 6.03 (method B) the intermediate acetate ester was not isolated but waa hydrolyzed directly. The reaction mixtures, after completion p ; A4-3-ethylene ketals are reported to absorb a t 6.00 of the acidolysis and subsequent hydrolysis (method B), were pZarb; and (c) when a solution of I1 in chloroform conmade basic with sodium hydroxide and extracted with petroleum t'aining a trace of water was left overnight a t room ether (b.p. 40-60"). The extract from each reaction mixture temperature, A-nortestosterone was obtained, The waa reextracted with 0.1 N hydrochloric acid. The acidic aquesusceptibilit'y of allylic ketals to mild acid hydrolysis ous extract was heated (50-60") in a water bath for 24 hr., made basic, and extracted with petroleum ether. After two additional has previously been noted.2asc cycles, the alcohols were recrystallized from the petroleum ether (1) (a) E . F. Fernhola and H. E. Stavely, Abstracts, 102nd National Meetsolution which had been dried over sodium sulfate, treated with ingof the American Chemical Society, Atlantic City, N . J..Sept. 1941,p.39M: charcoal, filtered through sintered glass, and concentrated. From l l z
-
(9) All melting points were obtained in a Hershberg [E. B. Hershberg. Ind. Ene. Chem., A n a l . E d . , 8 , 312 (1936)l silicone (550-Dow) filled melting point apparatus equipped with Anschutz full-immersion thermometers. The samples were placed in the circulating silicone bath 10' below the reported melting points and heated a t the rate of l-ZO/min. Elemental analyses were performed by Weiler and Strauss, Oxford, England. Isotopic Spe( 0 1 8 ) analyses were performed by Analytica Corp., New York, N . Y. cific rotations were determined with a Zeiss O.0lo polarimeter in a modified [C.Hite and J. Lyons, Chemist-Analyst, 68, 84 (1964)l 2-dm. (2-ml.) ayringefilling tube. The criterion for racemic products obtained from optically active starting materials was a level base line in the range 700-320 mp as determined with a Rudolph manual spectropolarimeter. T h e 0 1 8 enriched water was obtained from Isomet, Inc., Palisades Park, N . J.
(b) R. hntonucci, S. Bernstein, R . Littel, K. J. Sax, and J . H. \Yilliams, J . OW. Chem., 17, 1341 (1952); (c) G. I . Poos, G . E. .irth, R . E. Beyler, and L. H. Sarett. J . Am. Chem. Soc.. 76, 422 (1953). (2) (a) J . J . Brown, R . H. Lenhard, and S. Bernstein. Erperientin, 18, 310 (1962); J . A m . Chem. S o c . , 86, 2183 (1964); (b) Q. R . Petersen and E. E. Sowers, J . Org. Chem., 29, 1627 (1964); (0) J. \Y. Dean and R . G . Christiansen, ibid., 88, 2110 (1963). (3) J. Fried and E. Sabo [ J . A m . Chem. Sac., 84, 4356 (1962)l have reported the ketalization of a As-3-keto-A-norsteroid with p-toluenesulfonic acid gives a mixture consisting mainly of the Ae-58-3-ethylene ketal a n d a Ah-3-ethylene ketal. (4) F. L. \Yeisenborn and H. E. Applegate, ibid., 81, 1960 (1959). (5) R . Rull and G. Ourisson, Bull. aoc. chim. France, 1573 (1958). (6) L. F . Fieser and M. Fieser, "Steroids," Reinhold Publishing Corp., New York, N . Y., 1959, p. 309.
1326
NOTES
RJ$ OH R
III,R=O-CHz
I, R = O
I
II, R=O-CH2
0- CH2
1
IV,R = O
O-CH2 OR1
OR
V, R = R 1 = H VI, R=R1=Ac W, R=CHs; R'=H
These data indicated that the double bond in I1 was in ring A. Confirmation for this conclusion was obtained in the following manner. Treatment of I1 with osmium tetroxide and decomposition of the resulting osmate ester with an alkaline mannitol solution' afforded a single diol (n.m.r., t.1.c.) I11 which on acid hydrolysis gave a ketotriol IV, whose n.m.r. spectrum showed signals for the protons on the carbons 5.55 and 6.35. bearing the secondary alcohols a t The multiplet a t r 6.35 could be assigned to the C-17 hydrogen and since the peak at 5.55 was a singlet it was assigned to the C-3 hydrogen. If the hydroxyl were at C-6, the n.m.r. signal should be split by the hydrogens a t C-7. Further, dehydration of the ketotriol in ethanolic potassium hydroxide solution gave a product (V) which exhibited ultraviolet absorption characteristic NaOH of a diosphenol": :A( 265 mp, Amax 303 mp with a lower extinction coefficient). This evidence placed the cis-glycol at C-3, C-5 and left the configuration at these centers to be established. Direct evidence for the 3p15p-diol (IV) configuration was obtained from the optical rotatory dispersion (O.R.D.) curve. The O.R.D. curve of IV should be that of a 5@-(H)-2-keto-A-norsteroid since usually neither angular substitution (hydroxyl for hydrogen) at a position nonadjacent to a ketone, nor introduction of a hydroxyl function adjacent to the ketone affects the sign of the O.R.D. curve.8 Indeed, the O.R.D. curve of IV exhibited a negative Cotton effect ( [ a 1 3 2 1 - 1893O, minimum) similar in both sign and amplitude to methyl 2-keto-A-norcholanate9 ([01]312.5 - 1900°, minimum), The osmium tetroxide hydroxylation of cholestenone yields a mixture of the 4a,5a-diol and the 4@,5pdiol. lo The reaction of A-nortestosterone with osmium tetroxide was investigated with the hope of being able to isolate an isomeric diol. Decomposition of ( 7 ) R . Criegee, B. Marchand, and H. Wannowius, A n n . , 660, 99 (1942). ( 8 ) C . Djerassi. "Optical Rotatory Dispersion," McGraw-Hill Book Co., Inc.. New York. N . Y . , 1960. (9) C . Djerassi. R.Riniker, and B . Riniker, J . A m . Chem. Soc., '78, 6362 (1956). (10) J. Eastham. G . B. Miles. and C.A. Krauth, ibid., 81, 3114 (1950).
VOL.30
the intermediate osmate ester with an alkaline mannitol solution or biphasic treatment" of the osmate ester in benzene-methanol with sodium sulfite and potassium carbonate resulted in dehydration to the diosphenol V,12 and decomposition of the osmate ester with hydrogen sulfidela led to a single 3,5-diol in 60% yield identical with IV. From an inspection of models of I and I1 attack by the bulky osmic acid would seem to be favored from the p-side. Attack from the a-side is somewhat more hindered by the interference of the axial hydrogens at C-7 and C-9 and ring A which is folded toward ring B on the a-side. Acetylation of V afforded the diacetate VI,14 while treatment of V with methanol in the presence of boron trifl~oride'~ gave 3-methoxy-A-nortestosterone(VII). Since testosterone has been demonstrated to yield either the As-3-ketal or a mixture of the A4- and As-% ketals under similar reaction conditions, lbvZc the exclusive formation of the A3-2-ketal in the case of Anortestosterone merits discussion. In the normal steroid series a A3v5-dienolether which undergoes 1,2addition of the primary hydroxyl to the 3,4 double bond has been proposed as the intermediate in the formation of the A5-3-keta1,16whereas 1,Zaddition to a A2!4-dienol ether or displacement of water from its precursor would lead to the A4-3-ketal.2a*c The isolation of I1 as the sole product requires that either VI11 or IX be the intermediate in its formation. The failure to obtain any of the A5-2-ketal may be attributed to the fact that the A215-dienolether X does not form easily in A-norsteroids. In fact, the dienol acetate of A-norcholestenone has not been isolated even under forcing condition^.^^^ A235-
VI11
IX
X
(11) W. 9. Allen and 9 . Bernstein, ibid., '78, 1909 (1056). (12) E . Caapi, W . Schmid, and B . T. Khan [ J . Org. Chem., 46, 3898 (196l)l reported that reaction of adrenosterone with osmium tetroxide afforded the osmate ester in good yield, but that cleavage of the ester with an alkaline mannitol solution proved difficult and the yield of glycol was not satisfactory. A likely possibility, however, is that cleavage proceeded readily to a mixture of stereoisomeric cis-glycols, the major portion of which underwent dehydration t o 4-hydroxyandrenosterone. (13) D. H. R . Barton and D. Elad, J . Chem. Soc., 2085 (1956). (14) The alternate structure XI is excluded by the absence of a vinyl proton peak in the n.m.r. spectrum (see Experimental).
AcO 0
XI (15) R. Stevenson and L. F. Fieser, J . A m . Chem. Soc., '78, 1409 (1956). (16) C. Djerassi and M . Gorman, ibid., '76, 3704 (1954). (17) T. L. Jacobs and N . Takahashi, {bid., 80, 4865 (1958). (18) W. G. Dauben and G. A. Boswell [ibid.. 88, 5003 (1961)l obtained an oily product whose ultraviolet spectrum showed the presence of the dienol acetate contaminated with starting material.
NOTES
APRIL1965
1327
203-205"). Recrystallization of the residue obtained by evaporation of the mother liquor from ethyl acetate-hexane gave addiMelting points were taken on a Fisher-Johns melting point tional IV (581 mg., m.p. 197-199'; 323 mg., m.p. 198-200'). apparatus and are uncorrected. Values of [ a ] Dhave been ap3,17P-Dihydroxy-A-norandrost-3-en-2-one(V). A.-A mixproximated to the nearest degree and were taken in 9570 ethanol. ture of I (1.18 9.) and osmium tetroxide (1.1 g.) in pyridine Ultraviolet spectra were determined in 95% ethanol, infrared (1 ml.) and benzene (20 ml.) was stirred a t room temperature spectra in pressed potassium bromide pellets, and n.m.r. spectra for 65.5 hr. The reaction mixture was evaporated and the resiin deuteriochloroform with tetramethylsilane as internal standdue was dissolved in chloroform (100 ml.) and stirred with a 17G ard. All evaporations were carried out in vacuo. potassium hydroxide solution (200 ml.) containing mannitol 2-Ethylenedioxy-l7~-hydroxy-A-norandrost-3-en-2-one (11). (20 g.) for 2.5 hr. The layers were separated and the aqueous A.-A mixture of A-nortestosterone (I, 4 g.), p-toluenesulfonic phase was acidified with hydrochloric acid and extracted five acid monohydrate (165 mg.), benzene (250 ml.), and ethylene times with chloroform. The chloroform extracts were dried glycol (40 ml.) was stirred and refluxed for 7 days. The water over sodium sulfate and evaporated. Crystallization of the formed during the reaction was removed by a Dean-Stark residue from acetone-hexane gave V (292 mg., m.p. ?54.5moisture trap fitted with a calcium carbide thimble. The 255.5'). The analytical sample was prepared by recrystalreaction mixture was treated with pyridine (1 ml.), the benzene lization from acetone-hexane and had m.p. 259-260': [ a ] 2 9 ~ layer was separated, and the ethylene glycol layer was diluted +65" ( c 0.48); X 2.93 (17-OH), 3.18 (3-OH), 5.87 (2-one), with water and extracted with additional benzene. The comand 6.02 p (C-C); 7 9.19 (8, 18-Me), 8.83 ( 8 , 19-Me), 6.33 bined benzene extracts were washed with 8% salt solution, (m, 17-H), and 4.63 ( 8 , 3-OH); X 265 mp ( e 12,700); XNaOH dried over sodium sulfate, and evaporated. The residue was 303 mp ( e 9800). crystallized from isopropyl ether to yield I1 (2.45 g.): m.p. Anal. Calcd. for Cl8HZeO3(290.39): C, 74.44; H , 9.03. 149.5-150.5'; +34" (C 1.13); X 2.88 (OH) and 6.03 p Found: C, 74.31; H, 9.01. (C=C); I 9.22 (8, 18-Me), 8.96 (8, 19-Me), 6.32 (m, 17-H), B.-A solution of 11' (24 mg.) in chloroform ( 5 ml.) was stirred 6.05 (s, ketal methylenes), and 4.73 (d, J = 1-2 c.P.s., 3-H). with a 1% potassium hydroxide solution (10 ml.) containing Anal. Calcd. for C2oHSoOa (318.44): C, 75.43; H , 9.50. mannitol (19.) for 3 hr. and worked up as described above to give Found: C, 75.52; H , 9.47. T(5mg.,m.p.259.5-260.5"). Recrystallization of the residue obtained by evaporation of the C.-A mixture of I (920 mg.) and osmium tetroxide (853 mg.) mother liquor from isopropyl ether gave additional I1 (616 mg., in pyridine (1 ml.) and benzene (25 ml.) was stirred a t room temm.p. 146-147'). Plate chromatography of the residue using perature for 3 days. The reaction mixture was diluted with Woelm neutral alumina (activity V) as adsorbent and chloroform as the developing solvent gave a major band a t Rr 0.5, benzene (25 ml.) and treated with potassium bicarbonate (3.6 which was detectable by iodine vapor and which on elution with 9.) and sodium sulfite (3.6 g.) in water (45 ml.). Methanol ethyl acetate gave a residue which upon crystallization from (25 ml.) was added and the reaction mixture was stirred a t room isopropyl ether afforded additional I1 (113 mg., m.p. 145.5temperature for 1 day. The reaction mixture was filtered, and 146.5'; 152 mg., m.p. 143-145'). the layers were separated. The benzene layer was dried over B.-A-Nortestosterone (1 g.) was ketalized as described above sodium sulfate and evaporated. The residue was crystallized in benzene (150 ml.) with ethylene glycol (30 ml.) and p-toluenefrom ethyl acetate to give V (192 mg., m.p. 259-260'). Resulfonic acid monohydrate (15 mg.) to give a 5 5 7 , yield of 11. crystallization of the residue obtained by evaporation of the 2-Ethylenedioxy-3P,5~,17p-trihydroxy-A-norandrostan-2-one mother liquor gave additional V (96 mg., m.p. 257-258'; 27 (III).-A mixture of I1 (50 mg.) and osmium tetroxide (47 mg.) mg., m.p. 256-257"). in pyridine (0.4 ml.) and benzene (10 ml.) was stirred a t room 3,17j3-Diacetoxy-A-norandrost-3-en-2-one (VI).-A mixture temperature for 89 hr. The reaction mixture was evaporated of V (275 mg.), pyridine (0.4 ml.), and acetic anhydride (2 ml.) and the residue was dissolved in chloroform (8 ml.) and stirred was allowed to stand a t room temperature for 65 hr. The rewith a 1% potassium hydroxide solution (16 ml.) containing action mixture was poured onto ice and extracted three times with mannitol (1.6 g.) for 3 hr. The chloroform layer was separated ether. The ether extracts were washed with a saturated sodium and dried over sodium sulfate. Evaporation gave I11 (54 mg., bicarbonate solution and 87, salt solution, dried over sodium m.p. 204-207'). Recrystallization from benzene gave the anasulfate, and evaporated. The residue was crystallized from lytical sample: m.p. 208-209'; [ c Y ] ~ ~ +46" D (c 0.19); X 2.90 isopropyl ether to give VI (139 mg., m.p. 182-184'). Recrys(OH) and 3.00 p (OH); 7 9.25 (8, 18-Me), 9.01 (8, 19-Me), tallization from isopropyl ether gave the analytical sample: 6.37 ( m , 17-H), and 6.04 (s, ketal methylenes). 5.66 (3-acetate), m.p. 188.5-189.5'; [CX]~'D -3" ( c 0.22); Anal. Calcd. for C2oH32OS (352.46): C, 68.15; H , 9.15. 5.83 (17-acetate and 2-one), and 6.01 p (C-C); T 8.98 (8, Found: C, 68.19; H , 9.13. 18-Me), 8.78 (8, 19-Me), 7.96 (s, 17-acetate), 7.75 (8,3-acetate), 3~,5p,l7~-Trihydroxy-A-norandrostan-2-one(IV). A.-A and 5.40 (m, 17-H); X 240 mp ( e 15,000). mixture of 111 (225 mg.) and p-toluenesulfonic acid monohyAnal. Calcd. for C22H3005 (374.46): C, 70.56; H , 8.08. drate (17 mg.) in water (2 ml.) and acetone (8 ml.) was refluxed Found: C, 70.38; H, 7.91. for 5 hr. The reaction mixture was evaporated, diluted with 3-Methoxy-17p-hydroxy-A-norandrost-3-en-2-0ne (VII).-A water, and extracted three times with ethyl acetate. The ethyl solution of V (96 mg.) in methanol (20 ml.) containing boron acetate extracts were washed with a saturated sodium bicarbonate fluoride etherate (0.1 ml., freshly distilled) was refluxed for 18 hr. solution and 8% salt solution, dried over sodium sulfate, and The reaction mixture was evaporated, diluted with water, and evaporated. The residue was crystallized from ethyl acetate extracted three times with ethyl acetate. The ethyl acetate to afford IV (56 mg., m.p. 202-204'). Recrystallization from extracts were washed with 8% salt solution, dried over sodium ethyl acetate gave the analytical sample: m.p. 206-207'; sulfate, and evaporated. The residue was crystallized from [ a ] 2 2 D -40' (C 0.62); X 2.86 (OH), 2.93 (OH), and 5.75 p (2ether to afford VI1 (41 mg., m.p. 129-130"; 22 mg., m.p. 127.5one); T 9.20 (8, 18-Me), 8.85 (8, 19-Me), 6.35 (m, 17-H), and 128.5'). Recrystallization from isopropyl ether gave the 5.55 (8, 3-H); O.R.D.I9 ( c 0.06, methanol), [a1650-40", [a1321 analytical sample: m.p. 129-130"; \alzSD+55" (c 0.71); x -1893", and [ ~ ] z ~$1666". o 2.88 (OH), 5.88 (a-one), and 6.06 p (C=C); T 9.20 (8, 18-Me), Anal. Calcd. for ClsH2sO4 (308.42): C, 70.10; H , 9.15. 8.85 (s, 19-Me), 6.34 (m, 17-H), and 6.15 (8, 3-OCH3); X Found: C, 70.20; H , 9.11. 253 mp ( e 11,300) and 306 mp ( E 336). B.-A mixture of I (3.25 g.) and osmium tetroxide (3.0 9.) in pyridine ( 3 ml.) and benzene (60 ml.) was stirred a t room Anal. Calcd. for ClsH2803 (304.41): C, 74.96; H, 9.27. Found: C, 75.04; H, 9.28. temperature for 2 days. The reaction mixture was diluted with dioxane (100 ml.) and hydrogen sulfide was bubbled through Acknowledgment.-The authors wish to thank the the solution for 5 min. The precipitate was removed by filtration through Celite,20the filtrate was evaporated, and the residue following members of the Squibb Institute for their was crystallized from ethyl acetate to give IV (1.17 g., m.p. contributions to this work: Dr. A. Cohen for the n.m.r. Experimental
(19) The O.R.D. curve was determined by Professor A. K. Bose, Stevens
Institute of Technology. (20) Celite is Johns-Manville's trade-mark for diatomaceous silica products.
spectra, N r . J. Alicino and hlr. C. Sabo for the microanalyses, Miss B. Keeler and hIiss R. Karitzky for the infrared spectra, and Mr. W. Bullock for the ultraviolet spectra.
NOTES
1328
Kinetics of the Acid-Catalyzed Hydration of Isobutyraldehyde, Studied by Nuclear Magnetic Resonance Line-Broadening Techniques]
VOL. 30
terms of the reaction rate constant and various other parameters. By making approximations that should not lead to major errors in the present case the following simplified relation (eq. 1, CVK's eq. 5) is obtained
_1 -_
JACK HINEAND JAMES G. HOUSTON~
78
School of Chemistru, Georgia Institute of Technology, Atlanta, Georgia Received December 3, 1964
In a study of catalysis of the deuterium exchange of isobutyraldehyde, determination of the rate of hydration of the aldehyde became desirable in order to learn whether the dehydration or hydration reaction would ever be slow enough to become the rate-determining step. Experimentally it was noted that the two separate sets of n.m.r. absorption lines due to the hydrate and the free aldehyde became broadened and then fused as increasing concentrations of strong acid were added to the aqueous solutions of the aldehyde studied. It was therefore decided to study the reaction rate by n.m.r. line-broadening techniques. After the completion of our experiments an investigation by Gruen and McTigue was published in which the kinetics of the hydration of isobutyraldehyde were studied by the maximumtemperature-rise method.3 Initially it was also intended to study the kinetics of the hydration of acetaldehyde in order to provide a test of the kinetic method used since the hydration of acetaldehyde had been studied by Bell and co-workers using dilatometric4 and maximum-temperature-rise methods.5.6 After carrying out preliminary experiments on acetaldehyde, however, our study was discontinued when it was learned that the reaction had already been studied by n.m.r. techniques by Evans, Kreevoy, and Miller.'
Results
-
- VB VBTZB
vBm
(1)
where VB is the height of the absorption peak under the given set of conditions, V B is~ the height of the peak under conditions such that the rate of chemical reaction is negligible, T ~ isB the transverse relaxation time ( l / T z ~is equal to one-half the width of the peak13 at half-height under nonreacting conditions), and 1 / ~ is the first-order rate constant for the transformation of the species B to species A in the equilibrium being studied. In our case RCHO
ke + Hz0 & - RCH(0H)z ka
where both k B and k~ are treated as first-order rate constants. We were not able to determine V B m and T ~ reliably B since we were not able to obtain reaction conditions under which the hydration rate is really negligible. Aldehyde hydration is catalyzed by acids and our purification procedures did not succeed in removing all of the acidic impurities in our aldehyde samples. Since hydration is also catalyzed by bases, including salts of weak acids, it was not enough to "neutralize" our acidic impurities. Furthermore even uncatalyzed hydration proceeds a t an appreciable rate. However, by crude measurements of the hydration rate of aldehyde samples to which no acid had been added, we were able to show that the first-order rate constant for hydration under such conditions is no more than 0.2 see.-'. This proves that V B ~the , height, and T 2 ~ athe , reciprocal of the half-width a t half-height of the n.m.r. peak in solutions containing purified aldehyde and no added acid, are not too far from V B m and T ~ B . If the product of the peak height and half-width remains constant (as it would be expected to do and as it can be seen to do over a considerable range of reaction rates) over the range of rates from zero to the rate at which VB&and T z ~ ~ w e r e measured, then V B ~ / T ~will B be equal to VB&/T~B'. Equation 1 may then be rewritten as eq. 2. The second-
The strongest lines in the n.m.r. spectra of aqueous solutions of isobutyraldehyde8 are the doublets due to the methyl groups of the free aldehyde and its hydrate. When increasing concentrations of strong acid or strong base are added to such aqueous solutions, these doublets are broadened increasingly until they become unrecognizable; then a single merged doublet appears. The height of an n.m.r. peak that is being broadened (and hence shortened) by chemical reaction is expressed order rate constant for the hydrogen ion catalyzed - ~ l reaction is equal to the rate of change of the first-order by eq. 2 and 3 of Charman, Vinard, and K r e e v ~ y ~ in rate constant kg with respect to the hydrogen ion con(1) (a) This investigation was supported in part b y Public Health Service Research Grant AM 00829-01 MCB, from t h e National Institute of centration (eq. 3). Therefore no error in the deterArthritis and Metabolic Diseases. (b) Abstracted in part from t h e Ph.D. Thesis of J. G. Houston, 1965. (2) National Defense Education Act Fellow, 1960-1963. (3) L. C. Gruen and P. T. McTigue, J. Chem. Soc., 5224 (1963). (4) R . P. Bell and B. d. B. Darwent, Trans. Faraday Soc., 46,34 (1950). ( 5 ) R . P. Bell and J. C. Clunie. Proc. Roy. SOC.(London), PlPA, 33 (1952). (6) R . P. Bell, M . H . Rand, and K. M . A. Wynne-Jones, Trans. Faraday Soc., 62, 1093 (1956). (7) Private communication from R. P. Bell. (8) J. Hine. J. G. Houston, and J. H . Jensen, J . Org. Chem., 80, 1184 (1965). (9) H . B. Charman, D. R. Vinard, and M . M . Kreevoy, J . Am. Chem. S o c . , 84, 347 (1962). (IO) I n order t o conserve space these equations will not be repeated here. However, certain typographical errors should be pointed out. I n the last two terms in the numerator in eq. 2. 2A a n d 2B should be T ~ Aand T ~ B respectively , I n the denominator of eq. 3 a plus sign is missing just before A d .
mination of k H + will result from the use of V a a n d T ~ B & instead of V B and ~ T ~ Balthough , the k B values that would result will be in error by a const'ant amount and (11) According t o Alexander, the classical McConnell equations, upon which Charman, Vinard, and Kreevoy's derivation is based, are valid as long as t h e rate constant for exchange is small compared t o the difference in chemical shifts between t h e two environments oi the exchanging species (13 0.p.s. in our case). The Solomon-Bloembergen correction factor, which would never introduce a correction of as much as 4% into any of our d a t a , is said t o be generally valid.'' (12) S. Alexander, J . Chem. P h y s . , 88, 1787 (1963). (13) I n radians per second, not cycles per second.
NOTES
APRIL1965
1329
hydration of acetaldehyde a t 25" reported by Bell, Rand, and Wynne-Jones6 is 31% smaller than that reported by Gruen and McTigue. Inasmuch as isobutyraldehyde is 30% hydrated at equilibrium in aqueous solution a t 35",*a value of 1.0 X loa sec.-' may be calculated for the secondorder rate constant for the acid-catalyzed dehydration of isobutyraldehyde hydrate. Experimental
0.002
Figure 1.-Plot
of
kB'
[H 'I.
0.004
calculated from the data of Table I using eq. 4 us. [H+].
will hence be denoted k~'. Since V Bvalues were not so reproducible as desired they were measured by reference to a standard, t-butyl alcohol, added in known concentrations (cf. CVKg) so that we actually used eq. 4, (4)
were the Vr terms are the heights of the reference peaks. Plots of the values of kB' thus obtained us. the hydrogen ion concentration gave a straight line from as low a hydrogen ion concentration as could be obtained (without added base) to about 0.006 M , after which the slope decreased. The data obtained in solutions containing more than 0.006 M acid were neglected since the decrease in slope is presumably due to the overlap of one peak with another, a factor not allowed for in the equation used. To assess the validity of the assumptions that led to eq. 1, the rate constants were calculated analogously using CVM's complete eq. 3. The values thus obtained differed from those obtained using the simpler equation by less than 1%. A plot of a set of data obtained a t 35" is shown in Figure 1. From the slope of the line shown a secondorder rate constant of 438 M-' sec.-' may be calculated for the acid-catalyzed hydration of isobutyraldehyde in aqueous solution. Combination of this value with those obtained in several other such plots led to an average k H + value of 440 f 40 M-l sec.-l. Discussion The rate constant we have obtained for the acid-catalysed hydration of isobutyraldehyde in aqueous solution a t 35" (440 M-l sec.-l) is larger than that obtained a t 25" by Gruen and McTigue (370 M-' sec.-l)s using a different method. However our value is about 50% smaller than the value a t 35" that would be calculated from the rate constant of Gruen and McTigue a t 25" and an activation energy (16.2 kcal./mole) equal to that which may be calculated for the hydration of acetaldeh ~ d e . ~We ! ~have no explanation for this discrepancy but note that the rate constant for the acid-catalyzed
The n.m.r. spectra of isobutyraldehyde and its aqueous solutions were described earlier.8J4 In typical kinetic experiments 0.50-ml. samples of a 3% aqueous isobutyraldehyde solution were added to each of a number of n.m.r. tubes containing 0.25 ml. of aqueous perchloric acid of known concentration under nitrogen. The n.m.r. spect,rum of each sample was determined in the vicinity of the absorption due to the methyl groups a t a sweep width of 50 cycles and a sweep time of 500 sec. The spectra thus obtained were not changed significantly when the sweep time was changed to 250 sec., showing that the slow sweep assumption is valid. The results obtained in a typical run are shown in Table I. There was no significant difference in the rate constants obtained a t a different radiofrequency field strength; therefore, apparently the results are not complicated by saturation effects.
TABLE I N.M.R.MEASUREMENTS ON AQUEOUS ISOBUTYRALDEHYDE~ [H+l X 10' VB Vr kB ' 4.74 9.33 17.17 25.8 28.5 34.4 45.8 51.2 57.3
20,35 19.80 17.00 16.50 14.90 14.45 13.02 12.00 12.30
a The values of V B ~ Vra, , and respectively.
T2Ba
15.81 0.35 14,85 0.26 13.80 0.47 16.75 1.13 15.80 1.28 15,90 1.41 16.00 1.83 15.70 2.10 16.60 2.23 are 21.20, 14.20, and 0.455,
Crude measurements on the rate of hydration of isobutyraldehyde were made by injecting small amounts of the aldehyde into water and very quickly measuring the change in absorbance a t 2700 &. using a Cary spectrophotometer, Model 14.
Acknowledgment.-We are indebted to the National Science Foundation for grants that aided in the purchase of the n.m.r. spectrometer and the ultravioletvisible spectrophotometer, whose purchase was also made possible by a generous grant from the Charles F. Kettering Foundation. We also wish to thank Eastman Chemical Products, Inc., for the gift' of the isobutyraldehyde used in this study. (14) All n.m.r. spectra were determined using a Varian-A40 inatrument.
Acetylation of Triptycene' CHARLES J. PAGET A N D ALFREDBURGER Department of Chemistry, University of Virginia, Charlottesville, Virginia 88903 Received December B3, 1964
Although many derivatives of triptycene (9,lOdihydro - 9,lO - o - benzenoanthracene) have become (1) Supported b y a grant from Smith Kline and French Laboratories, Philadelphia, Pa. We are grateful to this company for their support of this work.
1330
NOTES
known, they have been prepared by one of two routes or their combinations : the condensation of anthracene or substituted anthracenes with benzyne(s) or quinone(s),6 and the further substitution or react,ions of derivatives of triptycene which had been formed by synthesis of the entire ring system.26 The direct electrophilic substitution of triptycene has not been recorded. We have now studied the Friedel-Crafts substitution of the hydrocarbon and have found that it orients the acyl group into the 2-position. Acetylation of triptycene in tetrachloroethane yielded 96.5Oj, of 2-acetyltriptycene; in carbon disulfide the yield of the same ketone was much lower, and an oily ketonic substance contaminated the main reaction product. Proof for the position of the acetyl group was obtained by treating 2-lithium triptycene, from syn-
VOL. 30
thetic 2-bromotriptycenela with acetonitrile and hydrolyzing the adduct; the resulting ketone was identical with that from the Friedel-Crafts reaction. In measuring the n.m.r. spectrum of 2-acetyltriptycene, a mirror image quintet centered a t 6 7.25 and characteristic of triptycene2Qwas found. There are two doublets (6 8.07 and 7.60) in the region of ortho protons and split with J a b a t 2 c.P.s., apparently due to meta splitting. A singlet a t 6 5.55 is due to the bridgehead protons, while one a t 6 2.45 represents the methyl protons. The ratio of these protons is 1: 1: 5 : 4:2:3. This spectrum is strikingly similar to that of 2-nitrotriptycene (see Experimental) which was synthesized from anthracene and 4-nitroanthranilic acid for spectral comparison. By contrast, the n.m.r. spectrum of synthetic 1-nitrotriptycene contains the aryl protons centered at 6 7.25 as two quintets which lack the symmetry found in unsubstituted triptycene. There are two singlets of equal intensity a t 6 6.6 and 5.55; the one a t 6 6.6 is evidently the bridgehead proton nearest the nitro group, and the other one the 10bridgehead proton. The ratio of the protons in the spectrum of 1-nitrotriptycene is 6 : 5 : 1: 1.
(2) T h e following derivatives of triptycene have been prepared by this method: 2-methyl, 9-formyl, 9-bromo. 9 - ~ a r b o m e t h o x y ,and ~ other 9substituted derivatives.4 (3) L. Friedman a n d F. Logullo, J . A m . Chem. Soc., 86, 1549 (1963). (4) E. W. Berndt, University Microfilms No. Mic 60-919; Chem. Abs.fr., 64, 14161e (1960). (5) T h e following have been prepared by these methods: g-bromo,"' 9-chloro,9 g-iodo,'O 9-nitr0,'tll 9-carboxy,a-lo g-formyl,lO 9-methoxy,' 9methyl,' g-phenoxy,g 9-cyan0,g 1,4-diamino,l2 1,4-diamino-9-carboxy,10 9-bromo-l,I-diamino,' 1.4-dibenaamid0,l~ 1-benzenesulfonamido-4-hydroxy,1',1S 9-bromo-10-methyl,' 9-chloro-10-methyl.' 9,10-dibromo,g 9,lOExperimental dimethyl'; the following derivatives of 1,4-dihydroxytriptycena~~~~e-l8: 2-methy1,u 9-formyl,' S-carboxy,'o g-carbomethoxy,IO 9-bromo,? 9-cyano,' All melting points were determined in a stirred bath and are 9-bromo-10-methyl,' Q,lO-dimethyl,' 9-carboxy-lO-methyl,1' S-carboxy,lQ corrected. Infrared spectra were measured with a Perkin-Elmer 9-(a-hydroxyethyl).' 9-chloro,' 9-cbloro-lO-methyl,~ 9-chloromethy1,' 9,lONo. 337 spectrophotometer, and proton magnetic resonance dibromo,' 9-methyl,' 9-nitro,' 9-phenoxy,' 2-t-buty1,20 2-t-amy1,m 9-bromo10-carbomethoxy," g-carbomethoxy,lQ 9-carboxy-lO-methy1,I' 5-carboxy,zl spectra with a Varian A-60 spectrometer. Triptycene was pre5-carboxy 1,4-diacetate,zl B-carboxy,21 6-carboxy 1,4-diacetatel; the folpared by the n-csthod of Friedman and Logullo.3 1,4-dioxime,lZ lowing l,4-dihydrotriptycenederivatives: 1,4-dioxo,'t~Z~~'~~z 2-Acetyltriptycene. A.-To a stirred mixture of 23 g. (0.0905 1,4-dioxo-9-formyl,' 9-carboxy-1,4-dioxo,~oQ-bromo-1,4-dioxo,' 9-cyano-1,4mole) of triptycene, 250 ml. of tetrachloroethane, and 24 g. of dihydroxy,' 9-cyano-1,4-dioxo,Q 9-acetyl-1,4-dioxo,Q 9-bromo-1,4-dioxo,g anhydrous aluminum chloride a t -20" was added 6.82 ml. 9-bromo-lO-metbyl-l,4-dioxo,~ 9-chloromethyl-1.4-dioxo,~ 9-chloro-1,4-di(0.0905 mole) of acetyl chloride in one portion. The mixture oxo, 9 9-chloro-10-methyl-1 ,4-dioxo,g 9,lO-dibromo-1 ,I-dioxo,9 9,lO-dimethylwas stirred a t -20 to -30" for 30 min. and then poured into a l,l-dioxo,' 9-methoxy-l,4-dioxo,9 9-methyl-1,4-dioxo,9 9-nitro-1,4-dioxo,g mixture of ice and 50 ml. of 37% HCl. This mixture was ex9-bromo-l0-carboxy-l,4-dioxo,~~ 9-carboxy-1,4-di9-pbenoxy-1,4-dioxo,' 9-carboxy-lO-methyl-1,4-dioxo,~~ 9ox0,10,19 9-carbomethoxy-1,4-dioxo,~~ tracted twice with 100 ml. of chloroform and twice with ether; carbomethoxy-l0-methyl-1,4-dioxo,~'9-amino-1,4-dioxo,2a 9-[(2-hydroxythe extracts were combined and dried (NQSOP). After removal l-napbthyl)azo]-1,4-dioxo2'; the following 1,4,4&,9a-tetrahydrotriptyof the solvents under vacuum, the remaining solid was decolorized cenes: g-brom0-1,4-dihydroxy,' 9-bromo-lO-methyl-l,4-dihydroxy,~9with charcoal during recrystallization from dilute ethanol; 9,10-dibromo-l,4chloro-1,4-dihydroxy, 9 9-chloro-lO-methyl-l,4-dihydroxy,~ yield 25.83 g. (96.5%), m.p. 20C&2Ol0. An analytical sample dihydroxy,' 9-ethyl-l,4-dihydroxy and 1,4-diacetate,' 1,4-dihydroxy,g 9had m.p. 200.5-201°; infrared bands a t 2935 (w) ( + C H ) , 1675 methoxy-1,4-dihydroxy.g 9-methyl-1,4-dihydroxy,Q9-nitro-1,4-dihydroxy,' ( 6 ) (C-0), and 740 and 625 cm.? ( s )(aromatic H ) . g-pbenoxy-l,4-dihydroxy,~ 1,4-dioxo,QJe 1-hydroxy-4-oxo,' 9-bromo-10Anal. Calcd. for CzzHlsO: C, 89.16; H , 5.44. Found: C, 9-oarboxy-10carboxy-1,4-dioxo,~~9-bromo-lO-carbomethoxy-1,4-dioxo,~~ 5-carboxy-1,4methyl-1,4-dioxo,l' 9-carbomethoxy-l0-methyl-1,4-dioxo,~Q 88.95; H, 5.52. 6-carboxy-1,4-dioxo.~~B-carbethoxy&oxo,?' 5-carbometboxy-1,4-dioxo,~~ 1,I-dioxo,21 9,lO-dimethyl-1 ,Pdihydroxy.g 9-bromo-l,4-dioxo.? 9-formyl-1,4(25) T h e following compounds have been prepared via other triptycene dioxo,' 9-carboxy-1,4-dioxo,~ 1,4-bia(beneoylimino),1a 1,4-bis(benrenederivatives: l-chloro,l2 g-chloro,ll 9-hydroxy,lO~ll*z( 9-acetoxy," 9-nitroso," sulfonimino) ,I4 1-benzenesulfonimino-4-oxo." g-phenylazo,Il 9-hydroxylamine," 9-(N-methylhydroxylamino),I1 9-benaal(6) P. D. Bartlett and E. 9. Lewis, J . A m . Chem. Soc., 71, 1005 (1950). amino and N-oxide," g-aaoditriptycene,Il 9-cyano-x-methyl.11 9-carbox(7) P. D. Bartlett, S. G. Cohen, J. D. Cotman, Jr., N. Kornblum, J. R. amide,'O 9-carbonyl chloride,lo 9-carboxazide.10 9-isocyanato,lo 9-carboxy Landry, and E. S. Lewis, ibid.. 71, 1003 (1950). anhydride,lO S-hydroxymethyl,lo 9-carboxy.8 g-carbomethoxy,a 9-carboxy(8) G. Wittig a n d U. Schollkopf, Tetrahedron, 3, 91 (1958). 1-hydroxy (and methyl ester and Iactone),l0 9-carboxy- and 9-carbomethoxy(9) W. Theilacker, U. Berger-Brose, a n d K. H. Beyer. Chem. Be?., 9S, 1658 1,4-diacetamido,Io 1,4-diacetamido,*z1,4-dibromo,l? 1,4-dihydroxy-1,2,3,4,5, (1960). 6-hexahydro and diacetate,'? 1,4-dihydroxytetradecahydro,l2 1,4-bis(ben(lo) P. D. Bartlett a n d F. D. Greene, J . A m . Chem. Soc., 76, 1088 (1954). 1,4-diol dibenzenesulfozenesulfonamido).l6 1,4-bis(methaneaulfonamido),~~ (11) W. Theilacker a n d K. H. Beyer, Chem. Ber., 94,2968 (1961). and 1,4-dibensoate,Qg-(l-hydroxynate,g 1,4-diol-1,2.3,4,4a,9a-hexahydro (12) P. D. Bartlett, M. J. R y a n , and 9. G. Cohen, J . A m . Chem. Soc., 64, ethyl)-1,4-dihydroxy.O 1,4-dibydro-l-hydroxy-4-oxo,~ 2-anilino-3-chloro2649 (1942). 1,4-dihydro-1,4-dioxo,*' 2-chIoro-1,4-dihydro-i,4-dioxo-3-(~-hydroxyethyl(13) R. Adams and D. S. Acker, ibid.. 74, 5872 (1952). amino),?' 1,4-dihydro-1,4-dioxo-2-oxy-3-pyridinium betaine.?' 2-chloro-1,4(14) R. Adams and J. D. Edwards, Jr., ibid.. 7 4 , 2603 (1952). dihydro-1,4-dioxo-3-(@-naphthylamino)," 9-(chlorodipbenylmethyl),*~ 9(15) R. Adams a n d C. R. Walter, ibid., 7 8 , 1152 (1951). (hydroxydiphenylmethyl),ze 9-triphenylboro,z' bis(S-triptycenyl)mercury,z' (16) E. Clar, Ber., 64,1676 (1931). 9-phenylazo,?e 2-anilino-3-chloro-l.4-dihydro-l,4-dioxo,~~ 2,a-dibromo- and (17) T. W. Chiu, K'o Hsueh T'ung Pao. 574 (1960); Chem. Abstr., 66, 2,3-dichloro-1,4-dihydro-l,4-dioxo,~~ 9-carbomethoxy-l,4-diamino,~~ 9-car13690d (1962). 9-carbobomethoxy,lo 9-methylamino," 2-chloro-1,4-dioxo-3-p-toluidyl,~' (18) S. Dershowitz, U. S. P a t e n t 3,065,075 (Nov. 20, 1962). methoxy-1,4-dioxo-l,4.4a.9a-tetrahydro,~~~~~ 9-triptycyl-9-triptoate,'0 1H(19) C. F. Wilcox. Jr., and A. C. Craig, J . 0x7. Chem., 16, 2491 (1961). 6,lOb-o-beneen0-6H-anthra[l,9-bc]furan-5-01, 2a.5,5a,lOc-tetrahydro,' 1H(20) E. Perotti and G. Castelfranchi, Chim. ind. (Milan), 42, 1333 B,lOb-o-benseno-6H-anthra[l,9-bc]furan,lo bis(9-triptyceny1)selenide and ( 1960). diselenide. 28 (21) A. Sonada, F. Ogura. and M. Nakagawa, Bull. Chem. Soc. Japan, (26) G. Wittig and W. Tochtermann, Ann., 660, 23 (1962). 86, 835 (1962). (27) B. D. Tilak a n d S. S. Rao, Chem. Ind. (London), 1320 (1957). (22) A. A. Balandin and E. I. Klabunovskii. Dokl. Akad. Nauk U S S R , (28) E. H . Daruwalla, S. S. Rao, and B. D. Tilak, J . SOC.Dyers Colouriste, 129, 102 (1959). 76, 418 (1960). (23) D. Y. Curtin. B. H. Klanderman, and D. F. Tavares, J . 0 r g . Chem., (29) T h e n.m.r. spectrum of triptycene displays a n AzBz pattern a t 1 7 , 2709 (1960). 6 7.15 resulting from t h e aromatic protons; a singlet a t 6 5.3 is attributable (24) R . Adams and J. D. Edwards, J r . , J . A m . Chem. Soe., 74, 2593 t o t h e two bridgehead protons. The ratio of the protons is 6:6:2. (1952)
APRIL1965 The 2,4-dinitrophenylhydrazone was prepared in the usual manner, washed with methanol and water, and dried over PZOSa t 55" for 6 hr. The crude material melted a t 271-272', and recrystallization from 2-propanol raised the melting point to 297". Anal. Calcd. for CZBHzON404: C, 70.58; H , 4.23. Found: C, 70.34; H , 4.03. Mixture melting points with the 2,4-DNP derivatives of the solid ketone prepared by method B, and from 2-lithium triptycene and acetonitrile (method C), respectively, showed no depression. The infrared spectra of the respective 2,4-DNP derivatives (s), [3290 (NH), 2930 (s) (triptycene SCH), 1620 (s) (-C=N-) 1595 and 1345 (s) (NO,), 832 (s) (1,2,4-trisubstituted benzene), and 740 cm. -1 (s) (ortho-disubstituted benzene)] were superimposable. Likewise, the 4-nitrophenylhydrazone of the ketone was prepared and recrystallized from dilute ethanol; m.p. 269-270". A mixture melting point with the 4-nitrophenylhydrazone from C was undepressed, and the infrared spectra of the two derivatives I3325 (w) (NH), 2930 (s) (triptycene SCH), 1610 (-C=N-), 1595 and 1335 (s) (NO%),and 745 cm.+ (s) (ortho-disubstituted benzene)] were superimposable. Anal. Calcd. for CzaHzlN,Oz: C, 77.94; H, 4.91. Found: C, 78.11; H , 5.20. B.-To 6.35 g. (0.025 mole) of triptycene in 60 ml. of carbon disulfide was added 6.67 g. (0.05 mole) of anhydrous AlCla. The mixture turned orange, then brown. With stirring, 2.55 ml. (0.025 mole) of acetic anhydride was added dropwise a t a rate so as to maintain reflux (25 min.). After refluxing for 20 min. the mixture was worked up with icehydrochloric acid as under A and extracted into methylene chloride. The pale yellow solid was recrystallized from cyclohexane to yield 4 g. of solid, m.p. 172-183". Thin layer chromatography (silica gel G, 8 : 2 methylcyclohexane-ethyl acetate, iodine vapor visualization) gave two spots, R f0.15 and 0.31 besides a faint spot a t Rr 0.58 (triptycene). Column chromatography of 350 mg. over F l o r i d using methylcyclohexane-ethyl acetate (8:2) furnished two ketonic products. Both had infrared bands a t 2935 (w) ( S C H ) , 1675 (aryl conjugated ketone), and 745 cm.-' (ortho-disubstituted benzene). On standing a t 4' for 1 week the material with Rr 0.31 crystallized, m.p. 201-202"; it showed no melting point depression with the ketone obtained by method A. The oily substance (Rr 0.15) yielded an orange 2,4-DNP derivative, R f 0.13, m.p. 298" (from isopropyl alcohol). A mixture melting point with the 2,4-DNP derivative (m.p. 297') of the ketone from method A (Rt 0.31 in the same solvent systems) was depressed (m.p. 275-285'). The infrared spectra of the two 2,4D N P derivatives were very similar, but the small amount of this oily ketone was not investigated further. C.-To 0.35 g. (0.05 g.-atom) of lithium in 20 ml. of ether under a nitrogen atmosphere a t -10' was added 2.45 ml. (0.025 mole) of n-butyl bromide in 75 ml. of ether. After stirring a t -10" for 2 hr., 3.33 g. (0.01 mole) of 2-bromotriptycene2was added in one portion. The mixture was stirred further a t 0' for 3 hr., 24 ml. of acetonitrile was added, and stirring was continued a t 26" for 10 hr. Decomposition of the mixture with a saturated solution of NHdC1, extraction with ether, drying (Na2S04), and removal of the ether furnished a solid which was recrystallized from dilute ethanol. A portion (1 g.) of 2-bromotriptycene crystallized first. The mother liquors on concentration gave 800 mg. (27%) of crude ketone; infrared bands a t 1675 (C==O), 2930 (triptycene), and 740 and 625 cm.-' (aromatic). This material still contained some 2-bromotriptycene as shown by thin layer chromatography on silica gel G in methylcyclohexane-ethyl acetate (8:2). The bromo compound on visualization with iodine vapor showed up as a pink spot while the methyl ketone appeared as a yellow spot. Separation of the ketone was effected best as the 2,4-DNP derivative which crystallized from isopropyl alcohol and after drying over PzOSfor 6 hr. melted a t 297". A sample of this derivative did not depress the melting point of the corresponding derivative prepared by method A. The infrared spectra of the two derivatives were superimposable. Likewise, the p nitrophenylhydrazone melted a t 270-271 O (from dilute ethanol); the mixture melting point with the corresponding derivative from method A was 270-271'. The two infrared spectra were superimposable. 1-Nitrotriptycene .-To a stirred refluxing mixture of 17.8 g. (0.075 mole) of 98% anthracene and 150 ml. of dry acetonitrile were added simultaneously over a 4-hr. period from separate dropping funnels 13.65 g. (0.075 mole) of 6-nitroanthranilic acid in 200 ml. of acetonitrile and 21 ml. of n-amyl nitrite in 75 ml. of
NOTES
1331
acetonitrile. The mixture was refluxed for 1 hr. after completion of addition and then the solvent wae distilled and replaced by 200 ml. of xylene. Maleic anhydride (12 9.) was added, and the mixture was refluxed for 45 min. It was cooled and diluted with 300 ml. of benzene. This mixture was extracted three times with 100 ml. of 10% NaOH, the benzene extract was dried (NaS04), and the solvent was removed. The remaining solid was chromatographed on alumina moistened with petroleum ether (b.p. 3060") and then diluted with 3000 ml. of petroleum ether followed by 2000 ml. of benzene. A white solid which crystallized from the petroleum ether eluate on standing was filtered off, washed with methanol, and dried; m.p. 221-221.5', yield 1.27 g. (5.6y0). Anal. Calcd. for C Z ~ H ~ S N O C,~ :80.25; H, 4.38. Found: C, 80.05; H, 4.31. The infrared spectrum contained bands a t 2930 (w) (.C-H, characteristic of triptycene), 1530 and 1355 (s) (NOz), and 735 cm. -1 (ortho-disubstituted benzene). 2-Nitrotriptycene was prepared essentially as was the 1-nitro isomer, starting with 71.4 g. (0.4 mole) of anthracene, 35.1 g. (0.3 mole) of n-amyl nitrite, 36.4 g: (0.2 mole) of 4-nitroanthranilic acid, and 49 g. (0.5 mole) of maleic anhydride. Work-up by continuous extraction with methylene chloride, and purification by chromatography over alumina with benzene, yielded 8.4 g. A small amount of (14%) of a yellow solid, m.p. 269-270'. contaminating anthraquinone was removed by sublimation a t 140' (0.075 mm.). Recrystallization from methanol gave the pure material, m.p. 270-271", yield 3.9 g. (6.5y0). Anal. Calcd. for CZoHlsN02: C, 80.25; H, 4.38. Found: C, 80.27; H , 4.41. The infrared spectrum showed bands a t 2930 (w) (SCH), 1520 and 1340 (s) (NOz),and 745 cm.-l (ortho-disubstituted benzene, strong); n.m.r. spectrum showed a mirror image quintet centered a t 6 7.10 (unsubstituted ring protons), doublets a t 8.12, and one centered a t 7.88 with J a b of 2 C.P.S. (in the region of ortho protons, apparently meta splitting), singlet a t 7.52 (probably the proton meta to NO%),doublet a t 5.7 (split of 3 c.P.s., probably due to a chemical shift), ratio of protons 1:1:1:4:4:2.
Synthesis of 3-Quinuclidinol by the Cyclodehydration of (4-Piperidyl)-l,Z-ethanediol HERBERTS. A A R O N OMER , ~ ~ 0. OW ENS,^' PAUL D. ROSENSTOCK,~~ STANLEY LEONARD,~~~O ELK IN,^^ AND JACOB I. MILLER^^ SAMUEL Chemical Research Division, Chemical Research and Development Laboratories, Edgewood Arsenal, Maryland, and Department of Chemistry, School of Pharmacy, Temple University, Philadelphia, Pennsylvania Recewed December 2.2,1964
The fact that 4-p-hydroxyethylpiperidinewas cyclized to quinuclidinez on activated alumina prompted us to carry out this type of reaction on the model (4piperidyl)-1,2-ethanediol in hope of obtaining a new synthetic route to an azabicyclic alcohol. The reaction scheme that was followed is depicted on p. 1332. Thus, 4-vinylpyridine (I) was oxidized by cold aqueous permanganate to (4-pyridyl)-l,Bethanediol (II).3 The glycol was best isolated as its hydrochloride, which was hydrogenated to the corresponding piperidyl glycol (111) hydrochloride. About 20y0 of 4-p-hydroxyethylpiperidinewas also formed as a hydrogenolysis by-product. The glycol hydrochloride (1) (a) Edgewood Arsenal; (b) Temple University; (c) Deceased. (2) S. Leonard and S. Elkin, J. 070. Chem., 17, 4635 (1962). (3) This glycol has not been described previously. although i t has been claimed IF. E. Cislak, U. S. Patent 2,743,277 (195S)l in unspecified yield from the reaction of the N-oxide of 4-(@-hydroxyethyl)pyridinewith acetic anhydride. However. this reaction could not be performed successfully in the Chemical Research and Development Laboratories, nor in those of one of our contractors.
NOTES
1332 CHZ - OH
CHZ- OH
CH - OH
CH-OH
I
CH
$-$ I
II
I
VOL. 30
obtained. The mixture was filtered and the potassium chloride filter cake was washed several times with acetone. The acetone washings were combined with the methanol filtrate and the solvents were removed under reduced pressure. The dark, viscous, oily residue waa dissolved in warm acetone (600 ml.) and filtered from the residual potassium chloride. After removing the acetone on a steam bath, the pink residue was recrystallized several times by dissolving the product in hot acetonebenzene (1: I ) , decolorizing with charcoal, filtering, and concentrating the filtrate to approximately one-half of its original volume. The solution was seeded (see below) and allowed to cool slowly to ambient temperature, with occasional swirling. In this way, the pure product (II), 21.2 g. (35%, based on recovered starting material), m.p. 71-73' dec., was obtained. Anal. Calcd. for C7HoN02: C, 60.42; H , 6.52; 0, 23.00; neut. equiv., 139.2. Found: C, 60.3; H , 6.4; 0, 22.8; neut. equiv., 140 (pK. 4.90 a t 0.015 ionic strength). Its infrared spectrum (dilute solution) showed a single, unresolved, broad (3550-3675 cm.? a t its base), 0-H stretching band with a maximum a t 3625 cm.-l. I n earlier work, attempts to distil the glycol led to extensive decomposition. However, a small quantity of the product distilled as an extremely viscous yellow oil, b.p. 142-145" ( < I p). The distillate, which gave a correct analysis, crystallized on standing a t ambient temperature to form a waxy solid. Recrystallization from ether-acetone (1 : 1)gave a product, m.p. 7C-72", which was used as the seed crystals mentioned above. (4-Pyridyl)-l,Z-ethanediolHydrochloride .-The acidified (pH 1) aqueous solution obtained from the oxidation (as above) of 52 g. of 4-vinylpyridine was concentrated, under reduced pressure, to about 100 ml. The potassium chloride which precipitated was filtered and the filtrate was allowed to evaporate spontaneously to dryness. The solid residue (50 9.) that contained some potassium chloride was recrystallized from ethanol to give 47 g. (0.27 mole, 54%) of the hydrochloride of 11, m.p. 126130'. Recrystallization from isopropyl alcohol gave a product, m.p. 131-I33', neut. equiv. 179 (calcd. 175.6). (4-Piperidyl)-l,2-ethanediol(III).-The pyridyl glycol (11) hydrochloride (18.3 g., 0.104 mole) in 15 ml. of 1 N hydrochloric acid and 30 ml. of ethanol was reduced in a Parr hydrogenator a t 66 p.s.i.g. over platinum dioxide (0.5 9.) a t ambient temperature. After 4 hr., hydrogen uptake had ceased. The fact that a 2OY0 excess of hydrogen had been absorbed indicated that some hydrogenolysis (see below) had probably occurred. The mixture was filtered to remove the catalyst, then concentrated, under reduced pressure, on the steam bath. The viscous residue was dissolved in 100 ml. of boiling ethanol. On standing a t room temperature there was obtained 12 g. (64%) of (4-piperidyl)-1,2-ethanediol hydrochloride (111HCl), m.p. 150-152'. An analytical sample, m.p. 155-156", was obtained by recrystallization from isopropyl alcohol. Anal. Calcd. for C7H&lNO,: C, 46.28; H, 8.88; 0, 17.61; neut. equiv., 181.7. Found: C, 46.2; H, 8.8; 0, 17.6; neut. equiv., 178 (pK. 10.85 a t 0.20 ionic strength). The hydrogenolysis product was isolated from another run. Thus, the total crude product (I1 HC1) obtained as a solid residue from the oxidation of 52.5 g. of 4-vinylpyridine (as above) was hydrogenated in 90 ml. of ethanol over 1 g. of platinum dioxide. The catalyst was filtered; the filtrate was concentrated under reduced pressure and then treated with an excess of sodium methoxide in methanol. After filtering the sodium chloride, the solvent was removed and the residue was distilled a t 0.050 mm. to give fraction 1, 4.7 g., b.p. 108'; fraction 2, 1.4 g., b.p. 108125'; and fraction 3, 20.0 g., b.p. 125-130', a viscous, colorless oil, n% 1.5155. Fraction 3 was shown (g.l.c., 245") to be essentially pure piperidyl glycol I11 (22.8 min.), contaminated by a few percent of 4-8-hydroxyethylpiperidine (4.4 min. ). The infrared spectrum of the piperidyl glycol (111) in dilute solution shows a doublet for the 0-H stretching band: 3635 (free OH) and 3590 cm.-l (intramolecularly bonded -OH, . .OH). G.1.c. analysis indicated fractions 1 and 2 were essentially pure 4+hydroxyethylpiperidine, the infrared spectrum (film) of which was identical to that of an authentic sample.? The 4-8hydroxyethylpiperidine could not be efficiently removed by redistillation of the glycol. Indeed, redistillation resulted in the formation of a few per cent of a new impurity, identified as 3quinuclidinol (4.1 min., 245") by g.1.c. analysis. The latter is apparently formed by thermal cyclodehydration of the glycol.
- b -cyoH H
--
m
Tv
was converted to the free base 111, which gave 3quinuclidinol (IV) on treatment with activated alumina a t 300". Depending upon the conditions and catalyst used, varying amounts of an unknown alcohol, possibly 1-azahicyclo[2.2.1Iheptyl-7-carbino1, plus several other low-boiling (presumably olefinic, not further investigated) products were also formed, as detected by gas-liquid chromatographic analysis. Cyclodehydrations of amino alcohols, although apparently only infrequently reported,' have been carried out under a variety of conditions. Of the various catalysts (aluminum, barium, calcium, strontium, and magnesium oxides) examined in our study, alumina proved to be the most effective. Interestingly enough, however, of five commercial aluminas screened in this study, only the Woelm neutral or (better) basic grade produced a satisfactory yield of 3-quinuclidinol. The simplicity and the results of our cyclodehydration study suggest that the method may be of synthetic utility as a general route to other azabicyclic alcohols. A related reductive cyclization of amino alcohols on Raney nickel has been used for the synthesis of az& bicyclic fused ring systems.6 This reaction, however, although formally a cyclodehydration, apparently proceeds by a mechanism that precludes the formation of bridged ring systems.6s6 Experimental Infrared spectra were recorded in dilute carbon tetrachloride in 1- or 2-cm. quartz cells on a Perkin-Elmer Model 237B grating spectrophotometer. Neutralization equivalents were measured electrometrically, using a Beckman Model H-2 pH meter. Gasliquid chromatographic (g.1.c.) retention times are given for a 10 ft. x 0.25 in. column of Carbowax 20 M (15%) on Gas-Chrom P (60/80) a t 120 ml./min. (He) a t the indicated column temperatures. (4-Pyridyl)-l,Z-ethanediol (11).-Potassium permanganate (52.6 g., 0.33 mole) in 1.5 1. of water waa added dropwise, over a 2-hr. period, to 4-vinylpyridine (52.5 g., 0.5 mole) in 450 ml. of water. During the addition, the reaction mixture was stirred vigorously and cooled in an external bath to maintain a reaction temperature of 2-4". After the addition, the mixture was allowed to stand overnight a t ambient temperature. The manganese dioxide was filtered (glass-sintered funnel, F) and washed several times with hot water (50 ml. per wash). The washings and filtrate were combined and extracted with three 500ml. portions of ether. On evaporation of the combined ether washings, 7 g. of vinylpyridine was recovered. The aqueous solution was acidified (pH 1)with concentrated hydrochloric acid and the resultant solution was evaporated to dryness under reduced pressure. Methanol (200 ml.) was added to the light yellow solid residue and a potassium hydroxide-methanol solution (13%) was added to the stirred suspension until pH 9 was (4) (a) Yu. K. Yur'ev, G. P. Mikhailovskii, and S. 2. Shapiro, Zh. Obahch Khim., 19, 2217 (1949); (b) L. J. Kitchen and C. B. Pollard, J . Am. Chem. S o c . , 69, 854 (1947); (c) C. Glacet and T. M. Deram, Compt. rend., 389, 889 (1954); (d) T. Ishiguro, E. Kitamura, M. Matsumura, and H. Ogawa, J . Pharm. SOC.Japan,71, 1370 (1055); (e) A. S. Sadykov, M. Xarimov. and Kh. A. Aslanov, Z h . Obahch. K h i m . , 8 8 , 3414 (1963): (5) M. G. Reinecke and L. R. Kray, J . O w . Chem., 39, 1736 (1964). ( 6 ) L. T. Plante, W. G. Lloyd, C. E. Schilling, and L. B. Clapp, ibid., Pi, 82 (1956).
(7) Kindly furnished by Reilly Tar and Chemical Corp.
APRIL1965
NOTES
The hydrogenolysis products is best removed from the glycol, therefore, by recrystallization of the hydrochloride salt before liberation of the free base. Fractions 1 and 2 were combined, treated (0.17 g.) with 2 ml. of methyl iodide in 4 ml. of isopropyl alcohol, warmed on the steam bath, then allowed to cool to room temperature. The solid thus obtained was recrystallized from isopropyl alcohol to give 1,l-dimethyl-4-(2-hydroxyethy1)piperidinium iodide, m.p. 169.5170.5'. Anal. Calcd. for C,H,JNO: C, 37.90, H , 7.07. Found: C, 37.7; H, 6.9. 3-Quinuclidinol (IV).-A 13-mm.-diameter glass reactor tube was packed with 10 g. of Woelm basic alumina (Alupharm Chemicals, New Orleans, La.), then heated to 300" in a muffle furnace, and maintained a t that temperature while a stream of nitrogen was passed through a t a rate of about 30 ml./min. Then, the piperidyl glycol III,4.8 g. (0.033 mole), in an open-end glass carrier tube (80 X 10 mm.), was inserted into the reactor tube just to the point where the latter entered the muffle furnace. The apparatus was inclined about 10" downward from the horizontal such that the glycol, when warmed to about 100" with a heating tape, slowly flowed down onto the alumina surface. Product formation was observed a few minutes later, and the reaction appeared to have been completed in less than 1 hr. The product, which had condensed as a viscous liquid in the cold portion of the exit tube, was dissolved in 23 ml. of methanol. Titration of a 1-ml. aliquot of this solution indicated a 43% yield (total base) was obtained. G.1.c. analysis (240") indicated a composition of about 80% quinuclidinol (4.1 min.), about 15% of an unidentified alcohol (from infrared, possibly 1-azabicyclo [2.2.l]heptyl-7-carbinol)(3.7 min.), and a few per cent of a second unknown product (2.9 min.). ( I n other runs, minor amounts of several other lower boiling products were also observed.) The remaining 22 ml. of the methanol solution was mixed with about 30 ml. of benzene (to azeotrope the water) and the solvents were removed under reduced pressure. The residue was recycled through another hot column, which now contained 3 g. of Woelm neutral alumina a t 190-200". The first condensate (liquid) was collected in a short collector tube (10-mm. diameter), which had been inserted into the column a t its point of exit from the furnace. This forerun was removed, and the product then condensed as a solid and was washed out with methanol. The methanol was evaporated, and heptane was added and evaporated to remove residual methanol. The solid that had precipitated was collected to give 1.34 g. (0.0105 mole, 32%) of 3-quinuclidinol, m.p. 205-210". The product thus obtained from another run melted a t 218-220" (lit.8 m.p. 221-223') after trituration with hot heptane, and gave an infrared spectrum (in potassium bromide) which was identical with that of authentics 3-quinuclidinol. (8) Although not investigated, this side reaction possibly could be prevented by decreasing the proportion of the reduction catalyst and the acidity of the medium, or by changing to an acetic acid solution. (9) L. H. Sternbach and S. Kaiser, J . Am. Chem. Soc., 74, 2215 (1952).
Organometallic Chemistry. IX. The Metalation of Benzocyclobutene with Sodium and Potassium alkyl^^^^ R . A. FINNEGAN Department of Medicinal Chemistry, School of Pharmacy, State University of New York at Buflalo, Buffalo 1.4, New York Received December 91, 1964
About five years ago, a reaction between butylpotassium and benzocyclobutene (1) was carried out as part of a projected study of the effect of strain on hydrocar(1) Grateful acknowledgment is made to the donors of the Petroleum Research Fund, administered by the American Chemical Society, for generous financial support of thia reaearch. (2) Paper V I I I : R. A. Finnegan and R . 9. McNeea. J . Orp. Cham., ID, 3241 (1964).
1333
bon acidity.8 Although this study, which was also to include indane and tetralin, did not materialize, the initial observations with benzocyclobutene, incomplete as they are, seem to be of sufficient interest for publication in this form. When a freshly prepared suspension of butylpotassium in pentane was stirred for 3.5 hr. with 1 equiv. of 1, followed by treatment with an excess of Dry Ice, only a single acid was obtained in purified form, and this in low yield. An identical acid was later obtained in somewhat higher yield as the sole product (other than caproic acid) after carbonation of a mixture of amylsodium with 1 equiv. of 1 which had been allowed to react for 1 week. This acid had m.p. 188-191' (sublimes) and therefore was not the known a-carboxylic acid which was reported4 to melt a t 76.5'. That this acid was isomeric was evident from the microanalytical results as well as the neutral equivalent, and the conclusion was reached that this material was either the 3carboxylic acid (2) or the 4-isomer (3). Consistent
2
3
with this conclusion was the carbonyl absorption in the infrared a t 1675 cm.-' (KBr) typical for an aromatic acid. Diazomethylation provided the methyl ester, m.p. 50.5-51' ( v f Z 1727 cm.-'), whose ultraviolet spectrum ruled out the possibility that it was a styrene derivative. In view of the fact that metalation of alkyl aromatic substrates usually occurs meta or para to the alkyl ~ u b s t i t u e n t ,structure ~ 3 was considered more likely to be correct. This tentative assignment was made in spite of the presence in the infrared spectra of the acid and ibs methyl ester (2a) of strong bands (779 and 749 cm.-' in the acid, 776 and 729 cm.-' in the ester) attributed to out-of-plane bending vibrations of aromatic CH groups and associated with a 1,2,3substitution p a t t e m 6 Reluctance to weigh this evidence heavily was conditioned by the lack of available infrared data on substances of this sort and by the observation that the spectra of methyl 2,3-dimethylbenzoate and 3,4-dimethylbenzoate were not so neatly classified as might have been anticipated. Thus, while the 2,3-isomer displayed a pair of strong bands (763 and 747 cm.-l) consistent with the normal correlation,6 the 3,4-isomer showed only a single strong band in this region (760 cm.-l) and even this one occurred a t lower frequency than expected.e, An unambiguous structural assignment was eventually made after comparison of the n.m.r. spectrum of 2a with those of the two isomeric diniethylbenzoate esters mentioned above. In the accompanying figure are shown tracings of the aromatic proton patterns observed in the spectra of these compounds, and it becomes clear on inspection that the acid resulting from (3) The potential synthetic utility of the reaction was. of course, not to be overlooked. (4) M. P. Cava, R . L. Litle, and D. R. Napier, J . Am. Chem. SOC.,80 2257 (1958). (5) R. A. Benkeser, D. J . Foster, D. M. Sauve. and J . F. Nobis, Chem. Rev., 57, 867 (1957). (6) K. Nakaniahi, "Infrared Absorption Spectroscopy." Holden-Day: Inc., Ban Francisco, Calif., 1962, p . 27.
1334
NOTES
I
I
8
7
A
8
I
a
C
B
7
7
3.25
D Figure 1.-Values are given in p.p.m.: A, methyl benzocyclobutene-3-carboxylate (2a); B, methyl 2,3-dimethylbenzoate; C, methyl 3,4dimethylbenzoate; D, benzylic protons iri 2a; center lines are 4 C.P.S. apart.
the metalation of 1 must be represented by structure 2, i.e., benzocyclobutene-3-carboxylic acid. In addition to the aromatic region, the features of the C-methyl region are also pertinent to the structural assignment. While the methyls of the 3,4-isomer give rise to a single sharp peak a t 7 7.71, the corresponding methyls in the 2,3-isomer are separated by about 9 c.P.s., with one a t 7 7.56 and the other a t 7.71. Similarly, the methylene groups of the four-membered ring are chemically shifted from one another in 2a, as indicated by the appearance of these protons as an AzBz multiplet centered a t 7 6.75. This multiplet is also shown (expanded five times) in Figure 1, the separation between the two center lines being 4 C.P.S. A final verification is inherent in a very recent report of the preparation of benzocyclobutene-4-carboxylic acid (3), whose melting point (139-140')' and spectral properties clearly distinguish it from 2. Thus, all three of the isomeric monocarboxylic acid derivatives of 1 have now been characterized. Since the first reaction described above was carried out, Benkeser and his studentss have shown that nuclear metalation of alkyl aromatic hydrocarbons is kinetically favored by both sodium and potassium reagents ; however, the final product distribution depends on the relative ease of equilibration of the metal to a benzylic position, as well as on subsequent transformations of the a-metalated species.8 The increased s character of the a-CH bond orbitals in l a might lead to the expectation (7) J. B. F. Lloyd and P. A. Ongley, Tetrahedron, 90, 2185 (1964). (8) R. A . Benkeser, J . Hooz, T. V. Liston. and A. E. Trevillyan, J . Am. Chem. Soc., 86, 3984 (1963), and previous papers. (9) G. Fraenkel. Y . Ashai, M. J . Mitchell, and M . P. Cava, Tetrahedron, 90, 1179 (1964).
VOL. 30
of increased acidity a t these positions. It is clear, however, that the conjugate base which would result from metalation a t a benzylic position would not enjoy efficient resonance stabilization without the introduction of an increased amount of strain. A similar consideration obtains in the case of cyclopentene where vinyl hydrogen rather than allylic hydrogen abstraction is the observed result of metalation.'O Unfortunately, the present findings do not rule out the possibility that the a-metalated species is indeed formed, and either exists in a prohibitively small amount in equilibrium with the ring-metalated isomer, or undergoes decomposition by, for example, an elimination reaction to yield an o-metallostyrene derivative whose fate was undetermined. l 1 In conclusion, there remains only to comment on the unique orientation observed in the metalation of 1. The avoidance of metalation ortho to the side chain previously noted in the reactions of simple alkyl aromatic hydrocarbons has generally been ascribed to the steric influence of the side chain, perhaps aggravated by the presumed heterogeneity of the system, as well as to the inductive effect which may serve to lessen the acidity of the ortho hydrogens. In the case of 1, not only is the steric repulsion to ortho attack markedly diminished, but also the hybridization state at C-7 and C-8 (ringjuncture atoms) is such that electron withdrawal from the remaining positions on the benzene ring would be expected. On this basis, then, metalation at C-3 would be predicted to predominate over reaction at C-4. The obtention solely of acid 2 is in qualitative accord with this prediction. Experimental The infrared spectra were recorded on a Baird Model B or a Perkin-Elmer Model 137 spectrophotometer. The ultraviolet spectra were recorded on a Perkin-Elmer Model 202 instrument. The n.m.r. spectra were determined using a Varian A60 spectrometer with solutions in carbon tetrachloride containing tetramethylsilane as internal standard. Melting points were observed o n a Fisher-Johns block, except where noted, and are uncorrected. Microanalyses were performed by Dr. A . Bernhardt, Mulheim, Germany. Benzocyclobutene-3-carboxylicAcid ( 2 ).-To a suspension of potassium sand (3.52 g . , 0.09 g.-atom) in pentane (250 ml.) was added during high-speed stirring and over a period of 45 min. 4.6 ml. (0.045 mole) of n-butyl chloride. The first small quantity of the halide was added at -10 to - 5 " , and then the mixture was allowed to warm up to 10" to ensure the initiation of (10) R. A. Finnegan and R. S. McNees, J . Oru. Chem.. 99, 3234 (1964). (11) I t seems pertinent to speculate that had the a-metalated species experienced a sufficient lifetime, a second metalation step might have proved to be energetically beneficial in view of the potential formation of the
-M+ 4
-M 5
theoretically very interesting disalt 4, which follows the Hnckel rule for aromaticity and would be isoelectronic with naphthalene. In order to facilitate the formation of the monosalt. and thus increase the possibility of forming a disalt, i t is proposed that 1-phenyl- or 1,2-diphenylbenzocyclobutene might prove to be an effective substrate for the production of species such 88 6 in a metalation reaction with sodium or potassium alkyls. W. Adam [Tetrahedron Letters, 1387 (1963) I has attempted, unfortunately without success, t o prepare a tetramethylcyclobutadienyl dianion by halogen-metal and hydrogen-metal exchange reactions of an appropriate substrate with lithium reagents. He has suggested that a tetraphenylcyclobutene derivative may prove efficacious, but cautions that the adverse effect of electron repulsion may negate the anticipated resonance stabilization.
NOTES
APRIL1965 the reaction. The mixture was cooled to - 10" for the remainder of the addition. Stirring was continued for 50 min. a t 0-5' before the addition of 4.27 g. (0.041 mole) of 1.12 The stirring was stopped 3.5 hr. later (0-5"), and the mixture was forced onto a large excess of powdered Dry Ice. Until this point, the reaction mixture had been maintained in an tmosphere of nitrogen. The carboxylate salts were dissolve8 in water and extracted with ether. The aqueous layer was acidified with concentrated hydrochloric acid and again extracted with ether. Titration of an aliquot of this ether solution indicated that acidic groups were obtained in 49% yield, based on the amount of butyl chloride used. Evaporation of the ether solution gave 1.80 g. of a residue which was digested in petroleum ether (b.p. 3G60"). The soluble portion was judged to be mainIy aliphatic acids by virtue of its odor and infrared spectrum. The insoluble residue provided crystals of 2, m.p. 187-190", from an aqueous ethanolic solution which had been cooled in an ice box. This material weighed 0.11 g. and represents approximately a 3y0yield based on the available organometallic reagent .I3 A sample of 2 was purified for analysis by sublimation (120°, 0.1 mm.): m.p. 188-191' (sealed capillary); Y",: 3300-2300, 1675,926, 779, and 749 cm.-I. Anal. Calcd. for C9Hg02: C, 72.95; H, 5.44; 0, 21.60; neut. equiv., 148. Found: C, 72.80; H, 5.37; 0, 22.53; neut. equiv., 148. In another experiment, 5.51 g. (0.053 mole) of 112 was allowed to react in a sealed bottle with 0.06 mole of amylsodium suspended in 150 ml. of pentane for 1 week a t room temperature. The amylsodium had been prepared from sodium sand and amyl chloride in a manner analogous to that described above for butylpotassium. Carbonation and similar work-up provided, after purification by sublimation, 0.45 g. of 2, m.p. 188-191" (sealed capillary) (6y0 yield, based on the amount of 1 used).I6 Methyl Benzocyclobutene-3-carboxylate(2a).-A sample of 2 was treated with ethereal diazomethane to give the ester 2a, which, after two recrystallizations from Skellysolve B, had m.p. 50.5-51°16; "2:; 1727, 776, and 729 cm.-*; 241 and 290 mp ( e 8130 and 2560). Anal. Calcd. for C10H1002: C, 74.05; H , 6.22; 0, 19.73. Found: C, 74.13; H, 5.97; 0, 19.90. The isomeric methyl dimethylbenzoates were similarly prepared by diazomethylation of the commercially available acids. (12) This sample was kindly supplied by Drs. M. P. Cava and M. J. Mitchell. (13) This low yield is no doubt owing in p a r t t o the fact t h a t insu5cient substrate was available t o be used as the suspending medium, BB in the recommended procedure for the metalation of alkyl aromatic hydrocarbons. Furthermore, subsequent observationslda have revealed inadequacies in the method used here for the preparation of b u t y l p o t a d u m as well as the surprisingly rapid decomposition of the alkyl potassium reagent, once formed.l'b80 These resultslds-c underscore t h e recommendation for having the substrate present in excess during the preparation of the potassium reagent, provided, of course, t h a t i t is inert t o the metal itself. (14) (a) R. A. Finnegan, Tetrahedron Letters, 1303 (1962); (b) ibid., 429 (1963); (c) ibid., 851 (1963). (15) The assistance of Mr. A. W . Hagen in carrying o u t this experiment is acknowledged.
A Study of the Mechanism of the Reaction of the Silver Salt of Phenylnitroacetonitrile with Triphenylmethyl Chloride N. E. ALEXANDROU~ Noyes Chemical Laboratory, University of Illinois Urbctna, Illinois Received November 11, 1964
The main product of the reaction of the silver salt of phenylnitroacetonitrile (I) with triphenylmethyl chloride has been found to be one of the stereoisomeric C Y , a'-bis(tripheny1methaneazo)stilbenes (11). (1) Laboratory of Organic Chemistry, University of Tbessaloniki, Thessaloniki, Greece.
1335
Ag+ -k (CeH6)aCl+
-]1:(-6...[
I CsHs-c=C-csHs
I
I
N N (C6H6)ac-
AA
-c(CsHr)s
I1
I n this reaction benzonitrile oxide (111) was also formed as shown by its interception with phenylacetylene to give 3,5-diphenylisoxazole. The benzonitrile oxide and the triphenylmethyl isocyanate (IV), which had previously been isolated by Wieland and HOchtlenl3 were proposed as possible intermediates + C~HI-C=N-O I11
(CsH6 W N C O IV
leading to the formation of the product, 11. The reaction of compounds I11 and IV was expected to give carbon dioxide and the nitrilimine V, which might then dimerize to the bisazoethylene derivative 11.2~4
CeHb-C=h=N-c( V
CBH~)~
However, the benzonitrile oxide has been found not to react with the triphenylmethyl isocyanate to give compound 11. The isocyanate IV was prepared according to Jones and H ~ r d . Its ~ infrared spectrum in carbon tetrachloride showed peaks a t 2260, 1590, and 700 cm.-l and the n.m.r. spectrum of a 12% solution in carbon disulfide had a singlet at r 2.80. The infrared spectrum of benzonit'rile oxide, immediately after its preparation, in carbon tetrachloride was similar to that which has previously been reported,6 with peaks a t 2290,1710,1365,1095, and 1025 cni.-1. The n.m.r. spectrum of a sample after the reaction of the triphenylmethyl isocyanate with the benzonitrile oxide, in carbon disulfide, had peaks a t 2.80 and 2.60. It was a mixture of isocyanate IV and 3,4-diphenylfuroxan (peak at r 2.60) produced from dimerization of the benzonitrile oxide. In order to establish whether the benzonitrile oxide takes any part in the formation of the bisazoethylene 11, experiments with labeled benzonitrile oxide (C6H6C1*NO)were carried out. This radioactive compound was prepared according to the method of Quilico and Speroni,' by using labeled benzaldehyde (C6H5C14HO). When to a mixture of the silver salt I and triphenylmethyl chloride a t -20" a solution of labeled benzonitrile oxide was added, 5.6% of bisazoethylene 11, in the form of red crystals, was obtained with m.p. 143" dec. The specific activity of the benzonitrile oxide was 0.693 pc./mmole, measured in the form of 3,4diphenylfuroxan. Before the addition of the beneonitrile oxide to the reaction mixture its infrared spectrum was obtained; it did not show the presence of (2) D.
PI, 4300
Y. Curtin, R.J. Crawford, and D. K. Wedegaertner, J . Org. Chem.. (1962).
(3) H. Wieland and A. Hochtlen. Ann., 505, 237 (1933). (4) D. K. Wedegsertner, P1i.D. Thesis, University of Illinois. 1962. (5) C. Jones and C. Hurd, J . A m . Chem. Soc., 43, 2422 (1921). (6) R. H. Wiley and B. J. Wakefield. J. 078. Chem., 25, 546 (1960). (7) A. Quilico and G. Speroni, Qazz. chim. ital., 7 6 , 148 (1946).
NOTES
1336
3,4-diphenylfuroxan. The specific activity of the product I1 was 0.0148 pc./mmole by combustion and measurement of the radioactivity of the produced carbon dioxide. In this reaction the total amount of the benzonitrile oxide was 17.32 mmoles (added 2.32 mmoles, produced from the reaction 15 mmoles). By assuming that 2 moles of benzonitrile oxide participate in the formation of the bisazoethylene 11, its radioactivity should be 0.1856 pc./nimole. Therefore, 8% of the used labeled benzonitrile oxide seems to be incorporated in this reaction. (In another experiment the incorporated amount was 5%.) However, when the isolated radioactive product (11) was d e c o m p o ~ e dby ~ ~heating ~ in benzene, the specific activity of the formed diphenylacetylene was found to be only 3.58 X pc./mmoles. The ultraviolet spectrum of this sample in ethanol was identical with that of an authentic sample of diphenylacetylene. Therefore, the almost inactive diphenylacetylene obtained from this reaction is evidence that the benzonitrile oxide does not participate in the formation of the compound 11, but is probably an intermediate of a side reaction. In any other case, the obtained diphenylacetylene should have the same activity as the compound 11. The radioactivity of compound I1 is probably due to radioactive impurities. The bisazoethylene I1 is insoiuble in a11 common s ~ i v e n t sand ~ ~for ~ this reason its purification by crystallization was impossible. The above conclusion, that the benzonitrile oxide is not an intermediate in the formation of bisazoethylene 11, is in agreement with the fact that the reaction of silver salt I with triphenylmethyl chloride in presence of p-chlorobenaonitrile oxide gave a product which was identical with the bisazoethylene 11. Another reasonable hypothesis for the mechanism of the reaction of I with triphenylmethyl chloride is that these two compounds form the nitroketimine VI, which further gives the bisazoethylene 11. Attempts to prepare compound VI from the sodium salt of phenylchloronitromethane (VII) and triphenylmethyl isocyanide (VIII) were unsuccessful.
1
NO2 CeHs-C=C=N-C(CeHs)a II
NOz VI
[GHs-C(
L
Na
ci VI1
VOL.30
a water bath for 1 hr., did not show the presence of diphenylacetylene. The same results were also obtained by changing the solvent, the amount of benzonitrile oxide, the time, and the temperature. Reaction of the Silver Salt I with Triphenylmethyl Chloride in the Presence of Labeled Benzonitrile Oxide .-A typical experiment is described. To 4 g. (15 mmoles) of I in 15 ml. of dry toluene, 4.2 g. of triphenylmethyl chloride in 30 ml. of toluene (kept over potassium carbonate for 10 min.) was added at -20' with stirring, under a nitrogen atmosphere, in a period of 30 min. After 1 hr., 0.276 g. (2.32 mmoles) of labeled benzonitrile oxide7 (freshly prepared) in 18 ml. of ether was added. The ratio of C6HsC"NO:Ag salt I was equal to 0.155. When this ratio was greater than 0.2, no product I1 was formed. The mixture was stirred for 1 hr. a t -20' and then filtered cold. The filtrate was allowed to stand overnight a t 0" to give 0.3 g. (5.6%) of red crystals (II), m.p. 143' dec. The specific activity of the benzonitrile oxide, measured in form of 3,4-diphenylfuroxan by using a liquid scintillation counter, found was 0.693 rc./mmole. (The specific activity of the benzaldoxime used was 0.698 pc./mmole.) If the entire amount of labeled benzonitrile oxide were incorporated in the reaction, the specific activity of the bisazoethylene I1 should be (2 X 2.32 X 0.693)/17.32 = 0.1856 pc./mmole. The specific activity of the isolated bisazoethylene I1 was found to be 0.0148 rc./mmole; it was determined by combustion and measurement of the radioactivity of the carbon dioxide, by using an ion chamber in conjunction with a vibrating-reed electrometer (Drift method). The sample before measurement was washed with benzene, ether, and then dried under vacuum. Anal. Calcd. for C52HaoNa:C, 86.6. Found: C, 87.1. When compound I1 was decomposed2j3by heating in benzene on a water bath for 30 min., it gave an oily product which was submitted to vapor-phase chromatographic analysis through a 4-ft. SE-30 column on Chromosorb W a t 180'. The product which had the same retention time (4.4 min.) as an authentic sample of diphenylacetylene was collected. Its specific activity was measured by using a liquid scintillation counter and found to be 3.58 x Mc./mmoles. The ultraviolet spectrum of this product in ethanol was identical with that of diphenylacetylene. In a similar manner the reaction was carried out in presence of p-chlorobenzonitrile oxide. The isolated red crystals had m.p. 143' dec. and the Beilstein test for chlorine was negative. Preparation of Phenylchloronitromethane .-This compound was prepared by treatment of phenylchloronitroacetamide with a solution of potassium hydroxide.g The oily product had peaks in the infrared spectrum (in carbon disulfide) a t 1560, 1350, 715, and 690 cm.?. The n.m.r. spectrum in carbon tetrachloride showed peaks a t r 3.30 (singlet, area 0.95) and 2.50 (multiplet, area 5.05). Preparation of Triphenylmethyl Isocyanide (VIII).-Compound VI11 was prepared by a method analogous to that of Hertler and Coreylo by treatment of 20 g. of N-triphenylmethylformamide," m.p. 199-201 with 17.5 g. of p-toluenesulfonyl chloride in 80 ml. of pyridine. The resulting solution was allowed to stand a t room temperature for 8 hr .; then cold water was added, the mixture was filtered, and the filtrate was extracted with ether. The ether layer was washed with water and then dried over anhydrous sodium sulfate. After chromat,ography on alumina (eluent carbon tetrachloride), 5.7 g. of VI11 were obtained, m.p. 130-133', after crystallization from benzene. The infrared spectrum in chloroform showed peaks a t 2130 and 1590 cm.-I and the n.m.r. spectrum of a lOyosolution in deuteriochloroform gave an unresolved singlet a t r 2.70. The infrared spectrum of triphenylmethyl cyanide,I2 m.p. 127-128", showed a strong absorption a t 2230 cm.-'. Anal. Calcd. for CZOHI~N: C, 89.2; H, 5.6; N , 5.2. Found: C, 89.0; H , 5.7; N , 5 . 2 . Reaction of the Sodium Salt of Phenylchloronitromethane(VII) with Triphenylmethyl Isocyanide (VIII).-To 0.5 g. of phenylchloronitromethane in 4 ml. of ether a solution 0.17 g. of sodium hydroxide in 10 ml. of absolute ethanol was added in portions with stirring. The resulting precipitate was filtered in the absence of air and washed with ether. I t was placed in a flask containing 10 ml. of dry benzene and t o this suspension 1 g. of triphenylmethyl isocyanide in 10 ml. of benzene was added. The O ,
+
(CeH5)aCNC
J
VI11
It is apparent that further work is required, in order to elucidate the mechanism of this unusual reaction. Experimental6 Reaction of Triphenylmethyl Isocyanate (IV) with Benzonitrile Oxide (III).-To 0.2 g. of IVs (m.p. 91-93') in 5 ml. of ether, a solution of 0.18 g. of 1117in 18 ml. of ether, dried over calcium chloride for 10 min. a t O', was added and the solution was allowed to stand a t 0' for 2 days. After evaporation of the solvent a yellowish solid was obtained, m.p. 8&1 IO", the infrared spectrum (in chloroform) of which showed that it is a mixture of IV and 3,4diphenylfuroxan. The n.m.r. spectrum of a 12% solution in carbon disulfide gave two singlets a t r 2.80 and 2.60 and was identical with that of a mixture of IV and 3,4-diphenylfuroxan. The ultraviolet spectrum in ethanol of the solid, after heating on (8) All melting points are corrected. N.m.r. spectra were obtained with a Varian Associates A-60 spectrometer at 60 Mc. The author is indebted to Mr. 0. Norton, Mr. D. Johnson, and their associates for these spectra. Infrared were measured with a Perkin-Elmer Model 21 spectrophotometer. Microanalyses were obtained in the Microanalytical Laboratory at the University of Illinois under the direction of Mr. J. Nemeth.
(9) A . van Peski. Rec. traa. chim.. 41, 687 (1922).
(IO) W. Hertler and E. Corey, J . Org. Chem., 98, 1221 (1958). (11) H. Bredereck, R. Gompper, and G . Theilig, Chem. Ber., 87, 537 (1954). (12) C. Sohimelpfenig, J . Org. Chem., 96, 4156 (1961).
NOTES
APRIL 1965 mixture was heated a t 45-50' with stirring in a nitrogen atmosphere for 70 hr. After filtration, the precipitate was dissolved in water and acidified with hydrochloric acid to give 0.1 g. of benzoic acid. The filtrate was submitted to chromatographic analysis on alumina to give 0.7 g. of triphenylmethyl isocyanide, 25 mg. of triphenyl carbinol, m.p. 160-163', and finally an unidentified oily product, which, however, waa not diphenylacetylene. When tetrahydrofuran was used as solvent instead of benzene, a part of the isocyanide waa isomerized to cyanide. The mixture had m.p. 128-132' and the infrared spectrum had peaks a t 2130 and 2250 cm.-'.
Acknowledgment.-The author is very much indebted to Dr. D. Y . Curtin for his advice and encouragement and expresses his appreciation to Dr. R. F. Nystrom for the radioactivity measurements. This work was supported by a grant (G-14,480) from the National Science Foundation.
The Application of Bredt's Rule to
1337
and recognized that this was probably a consequence of the strain associated with the hypothetical product. In a subsequent study Rabe' found that 3,5-dimethylbicyclo [3.3.1Inonan-1-01 (5) could not be dehydrated even under drastic conditions.
In contrast to these results, Meerwein5 later found that 6 was decarboxylated rapidly in water at relatively mild conditions (ca. lSOo) for a compound of this structure. Since this reaction is believed to proceed through formation of the enol, it would appear that a t least a transient double bond can exist a t a bridgehead of the bicyclo [3.3.1Inonane system. To investigate this pos-
Bicyclo[3.3. llnonanes JOHN P. SCHAEFER AND JOHN C. LARK Department of Chemistry, University of Arizona, T U C W Arizona ~, Received November $6, 1964
Bredt's rule' is an empirical formulation which, when applied to bridged ring systems, defines the minimum geometrical requirements necessary to accommodate a double bond a t a bridgehead position. In an attempt to determine the lower limit of n in the bicyclo[n.3.1] series compatible with unsaturation a t the bridgehead, Prelog2 studied the aldol condensation of 1 and found that the ratio of products 2 and 3 was a sensitive function of n. When n was 6, 5 , or 4,the yields of 2 and 3 were 76 and 0, 14 and 36, and 0 and 65%, respectively.
sibility further we have studied the base-catalyzed deuterium exchange of bicyclo [3.3.l]nonan-2-one (8) in detail. When 8 was heated at 95" in deuterium oxide (ca. 0.1 M NaOD) up to three atoms of deuterium per molecule were incorporated. Table I summarizes the TABLE I BASE-CATALYZED DEUTERIUM EXCHANQE OF B I C Y C L O [ ~NONAN-%ONE .~.~] EXchange time, Sample days
1 2 3
COzCHs
6 26 40
Do, %
DI, %
Da, %
Ds,%
Average number of D per molecule
1.5 0.8 1.1
8.1 3.2 2.45
32.0 35.1 19.3
58.4 60.9 77.2
2.46 2.56 2.72
1
2
3
Although these data clearly show that when n = 5 there is sufficient flexibility in 2 to tolerate an olefinic linkage a t the bridgehead, they cannot be used to place a lower limit on the value of n since the reaction studied is subject to thermodynamic control and the ratio of products observed reflects only the relative stabilities of these compounds. During a study of the bicyclo [3.3.l]nonanes, Rabe3 observed that 4 was extremely resistant to dehydration (1) F. S. Fawcett, Chem. Rev., 47, 219 (1950. (2) V. Prelog, J . Chem. SOC.,420 (1950); V. Prelog, P. Barman, and M. Zimmerman, Helu. C h i n . Acta, 99, 1284 (1949); V. Prelog, L. Ruzicka, P. Barman, and L. Frenkiel, $ b i d . , S1, 92 (1948). (3) P. Rabe, R. Ehrenstein, and M. Jahr, A n n . , 860, 265 (1908).
results of three separate, but similar, experiments. Under no conditions were more than three deuterium atoms per mole incorporated. In 8 the three hydrogens adjacent to the carbonyl group appear a t r 7.6 units in the n.m.r. spectrum. Since this absorption disappears after deuterium exchange and the remainder of the spectrum is unchanged, it is apparent that only the hydrogens adjacent to the carbonyl group are acidic enough to undergo exchange. Furthermore, since Bartlett and Woods6 have shown that the cumulative inductive effect of two adjacent carbonyl groups is insufficient to increase the acidity of the bridgehead hydrogen in bicyclo [2.2.2]octane-2,6-dione significantly, it must be concluded that there is sufficient orbital overlap in the enolate ion 9 to confer substantial acidity upon the bridgehead hydrogen, or, in terms of valence bond (4) P. Rabe and K. Appuhn, Ber., 76B,082 (1943). (5) H. Meerwein and W. Schtirmann, A n n . , SSS, 196 (1913). (6) P. D. Bartlett and G. F. Woods, J . Am. Chem. Soc., 61,2933 (1940).
1338
NOTES
cqoQ-Q 0
8
0-
9b
9a
pictures, structure 9b must be an important resonance form. A difficulty which arises when enolization data is used to evaluate the limits of Bredt's rule is that the degree of resonance stabilization of the enolate ion (and therefore, the acidity of the bridgehead hydrogen) is a function of the cosine of the projected angles between the interacting orbitals in the ion. Since it has not been ascertained low much orbital overlap is necessary to allow enolate formation, these data only require the conclusion that significant double-bond character can exist between C-1 and C-2 in the bicyclo [3.3.1Inonane series and not that a stable double bond can be formed between these carbon atoms. Although there is ample evidence that the preferred conformation of bicyclo[3.3.l]nonanes is the chair chair ni~dification,~ an examination of Dreiding models of 8 immediately shows that formation of a true enolate is not possible in this conformation since the C-1-H bond is orthogonal to the p-orbitals of the carbonyl group. However, if the ring containing the carbonyl group is in the boat form, the projected angle between the interacting orbitals decreases to 30" (cos 30" = 0.866) and resonance interactions in the transition state for removal of the proton and in the subsequent enolate ion can result in significant delocalization energies. The importance of the conformational factor on enol formation is dramatically illustrated by the stability of 10 toward decarboxylation8 since all rings are locked in the chair conformation, in contrast to 6 where the rings are mobile.
10
Experimental I n a test tube was placed 200 mg. (1.45 mmoles) of bicyclo[3.3.l]nonan-2-0ne,~5 ml. of freshly distilled dioxane, 5 ml. of deuterium oxide, and ca. 20 mg. of sodium. The tube was sealed, placed in a constant-temperature bath a t 95', and after several days (see Table I ) was removed and cooled and the contents were poured into 100 ml. of pentane. The pentane solution was washed with five 50-ml. portions of ice-water to remove the dioxane and dried over magnesium sulfate. After removal of the pentane the ketone was sublimed and analyzed on the mass spectrometer.
Acknowledgment.-The authors wish to express their gratitude to the Petroleum Research Foundation (Grant PRF-789) for their generous support of the initial phases of this work. ( 7 ) W. A. C. Brown, G. Eglinton, J . Martin, W . Parker, and G. A. Sim, Proc. Chem. Soc., 57 (1964). M.Dobler and J. D. Dunitz, H e h . C h i n . Acta, 47, 695 (1964). (8) 0 . Bottger, Ber., 70B,314 (1937). (9) H. Meerwein, F. Kiel, G. Klogsen, and E. Schoch, J . prakt. Chem., 104, 161 (1922). A more convenient procedure for t h e preparation of this compound will be published shortly.
VOL. 30 Reduction of Sulfoxides by Triphenylphosphine and Carbon Tetrachloride JOSd
P. A.
CASTRILL6N AND
H.
H.4RRY SZMANT'
Puerto Rico Nuclear Center and Department of Chemistry, University of Puerto Rico, Rio Piedras, Puerto Rico Received October 23, 1964
While the reduction of sulfoxides by triphenylphosphine is acid catalyzed2 and thus the reduction does not take place in an inert solvent such as benzene, it was found3 that di-p-bromophenyl sulfoxide was reduced to the corresponding sulfide when heated with triphenylphosphine in the presence of carbon tetrachloride. The purpose of this Note is to report this new reaction in view of its potential usefulness since, generally speaking, it produces good yields of sulfides even in the presence of nitro groups. Experimental The reductions of most diary1 sulfoxides were performed by refluxing, during 2 hr., a mixture of 0.01 and 0.02 moles of the sulfoxide and triphenylphosphine, respectively, in 100 ml. of carbon tetrachloride. The reaction mixture was then evaporated on a steam bath and the residue was worked up depending upon the solubility of the sulfide. Sulfides which crystallize readily from ethanol (di-p-bromophenyl and di-p-nitrophenyl sulfides) were obtained by dissolving the residue in ethanol in which triphenylphosphine oxide is readily soluble, while the lower melting sulfides were obtained by extraction of the residue with petroleum ether ( b .p. 35-47 ") and by evaporation of this extract. In the case of di-p-hydroxyphenyl sulfoxide which is highly insoluble in carbon tetrachloride, the dry, finely ground powder was suspended in a mixture of 100 ml. each of carbon tetrachloride and benzene and the reflux period was extended to 18 hr. The reaction mixture was concentrated as described above and the oily residue was then extracted with aqueous sodium hydroxide. This left behind most of the triphenylphosphine oxide, and the sulfide was isolated by acidification of the alkaline extract, extraction with ether, crystallization from water, and recrystallization from benzene. The reaction of dimethyl sulfoxide is highly exothermic and thus a solution of 0.01 mole of this sulfoxide in 50 ml. of carbon tetrachloride was added slowly to the solution of the phosphine in 100 ml. of carbon tetrachloride. The volatile dimethyl sulfide was swept by means of a stream of nitrogen into a 2.5% solution of mercuric chloride in water in order to precipitate the sulfide in the form of the metallic ~ o m p l e x . ~ The experimental results are summarized in Table I . The yields and melting points of the sulfides are those of the crude products. In all cases these were further purified and found to give satisfactory mixture melting points with samples of the authentic sulfides. The purity of the crude products was also examined by thin layer chromatography. Owing to the high solubility of triphenylphosphine oxide in ethyl alcohol the crudes obtained by crystallization from this solvent (p-bromo and pnitro) were free from the oxide; conversely all the crudes obtained by petroleum ether extraction contained a trace of it. The crude of di-p-hydroxyphenyl sulfide contained a larger amount of the oxide but crystallization from water brought about a satisfactory purification of the sulfide. The reaction between di-p-bromophenyl sulfoxide and triphenylphosphine in carbon tetrachloride was also carried out using equimolar quantities of the first-mentioned reagents, and the expected sulfide was isolated in 91 yoyield. (1) T o whom all inquiries should be addressed. (2) H. H. Szmant and 0. Cox, manuscript in preparation. (3) 0 . Cox, M.S. Thesis, University of Puerto Rico, 1964. (4) W. F. Faragher, J. C. Morell, and S . Comay, J . A m . Chem. Soc., 61, 2781 (1929).
APRIL1965
NOTES
1339
TABLE I REDUCTION OF SULFOXIDES BY TRIPHENYLPHOSPHINE AND CARBON TETRACHLORIDE Sulfide-----------, M.p.,O C .
7
Sulfoxide
M.p. "C.
Yield, % '
Lit. m.p., O C .
Lit. m . p . , O C .
Dimethyl" 82 157.5-163.5' 150-1 5 1' 94-95 92' 100 54-57.5 57.3d Di-p-tolyl Di-p-hy droxyphenyl 195-196 195" 27 135.5-147.5 151-151.5' Di-p-methoxyphenyl 95-96 93-94' 90 38.5-42.5 46' 154-155 153h 94 114.5-115.5 112' Di-p-bromophenyl Di-p-nitrophenyl 182-1 87 178-180' 83 150-153 158-160k Phenoxat hiin- 10-oxide 154-155 15&159l 91 51.5-55 57. 5-5gm Mercuric chloride complex reported in ref. 4. a Matheson Coleman and Bell product redistilled in vacuo. H. C. Parker, Ber., 23, E. Fischer, ibid., 48, 96 (1915). e S. Smiles and A. W.Bain, J . Chem. SOC.,91, 1119 (1907). I G. Tassinari, Gam. 1845 (1890). J. Boesecken, Rec. trav. chim., 29, 315 chim. ital., 17, 83 (1887). 9 S.Smiles and A. Le Rossignol, J . Chem. SOC.,93, 755 (1908). K. W. Rosenmund and H. Harns, Ber., 53, 2234 (1920). H. H. Szmant and J. J. McIntosh, J. Am. Chem. SOC.,73, 4356 (1910). H. D. K. Drew, J . Chem. SOC.,511 (1928). C. M. Suter and (1951). k C. C. Price and G. W. Stacy, Org. Syn., 28, 82 (1948). Ch. E. Maxwell, "Organic Syntheses," Coll. Vol. 11, John Wiley and Sons, Inc., New York, N. Y . , 1943, p. 485.
'
While the mechanism of this reaction is currently under study, it is very likely that the reduction of the sulfoxides is brought about by the dichloromethylidene derivative of the triphenylphosphine produced6 according to this equation. 2Ph3P
+ CCl, --+ PhaP=CClz + Ph3PClz
Furthermore, one is tempted to suggest that the ylide reacts subsequently with the sulfoxide to give an intermediate sulfur ylide, similar to the known reaction of carbonyl compounds.6 Ph,P=CClz
+ R2SO +PhaPO + (RzS=CClz)
Although the fate of the =CC12 moiety is not clear, the following possibility is suggested a t this time. (RzS=CClz)
+ PhaPClz +R2S + CCl, + PhaP
In favor of this suggestion is the observation that a nearly theoretical yield of di-p-bromophenyl sulfide is obtained when equimolar quantities of the sulfoxide and triphenylphosphine are allowed to react. We wish to add that a large-scale experiment employing dimethyl sulfoxide failed to provide evidence for the formation of tetrachloroethylene. This would indicate that the dichlorocarbene, a possible decomposition product of the sulfur ylide, is either not produced or efficiently scavenged6c7by triphenylphosphine . (5) R . Rabinowitn and R . Marcus, J . A m . Chem. Soc., 84, 1312 (1962). (6) G. Wittig, et at., Angew. Chem., 79, 324, 4 - 7 (1960). (7) D. Seyferth, S. 0. Grim, and T. 0. Read, J . A m . Chem. SOC.,89, 1510 (1960).
Solutions of Organic Compounds in Fused Alkali Thiocyanates THOMAS I. CROWELL A N D PAUL HILLERY Cobb Chemical Laboratory, University of Virginia, Charbttesville, Virginia Received November 1.9,1964
The number of reports of reactions between organic compounds and fused salts is increasing.l These reactions, understandably, involve a gas in contact with the high-temperature melt. Homogeneous reactions are rare, though some interesting current research makes use of fused tetra-n-alkylaninionium nitrates.2 The melting points of potassium thiocyanate (177') and of a 3 : 1 mixture by weight of KSCN and NaSCN (1) D.C. Coldiron, L. F. Albright, and L. G. Alexander, Ind. Eng. Chem., SO, 991 (1958); M. Fild. W. Sundermeyer, and 0. Glemser, Ber., 87, 620 (1964); and earlier references. (2) J. E. Gordon, J . A m . Chem. Soc., 86, 4402 (1964).
140
0
20
40 60 80 MOLE % HYDROQUINONE
100
Figure 1 .-Freezing point-composition plot for potassium thiocyanate-hydroquinone system.
(about 130') are low enough to permit the study of organic compounds in the fused salts. While investigating organic reactions in these media, we observed the high solubility of a number of polyhydroxy compounds. At 150' in fused ICWN-NaSCN, pentaerythritol, hydroquinone, ethylene glycol, methanol, and water are very soluble and sucrose, glucose, and p-nitroaniline are fairly soluble, glucose being recoverable as the osazone. The carbohydrate solutions begin to turn yellow after 10 min. at 150'. Triphenylniethyl chloride, triphenylcarbinol, hexyl bromide, phenol, borneol, p-toluidine, and pyridine are relatively insoluble. Benzoic acid is soluble and can be sublimed unchanged from the melt, but some decomposition of the thiocyanate occurs, probably due to the formation of HSCN. The hydroxy compounds are apparently un-ionized in solution. Alizarin dissolves to form an orange solution and only when a trace of sodium hydroxide is added, does the deep blue color of the anion appear. The addition of solid benzoic acid changes the color back to yellow. The n.m.r. spectrum of pentaerythritol in KSCN-NaSCN at 150' shows two singlets, separated by 0.4 p.p.m. a t 60 Me. The fused-salt solubility of the limited number of compounds observed shows some similarity to their water solubility. There are differences, for example ethanol is not miscible at 150', though there is con-
NOTES
1340
siderable mutual solubility in the two-phase system. Pyridine is insoluble, perhaps because of the impossibility of any hydrogen bonding, and sodium benzoate is not wet by fused potassium thiocyanate. A rough freezing point-composition diagram (Figure l ) , constructed from cooling curves for the system KSCN-hydroquinone, shows a minimum at about 140' and 60 mole % of hydroquinone. The organic compound (m.p. 171') and the fused salt are completely miscible above 177'. Experimental Reagent grade potassium thiocyanate and sodium thiocyanate were oven dried a t 150' for 6 hr. In the solubility tests, 0.1 g. of organic compound was stirred in a test tube with 10 g. of the molten solvent, 7593 KSCN-25% NaSCN. Temperature was maintained a t 150' by an aluminum-block thermostat containing recesses filled with silicone oil. The solubilities of methanol and ethanol in the fused thiocyanate were observed in sealed tubes behind a safety shield. The compounds classified above as insoluble are probably far below 0.1 g./lO g. in solubility. Pentaerythritoi (2.3 g.) easily dissolved in 10 g. of molten potassium thiocyanate, lowering the melting point to 156". Certainly the solubility would be much higher a t 176'. The hydroquinone-KSCN mixtures were sealed into Pyrex tubes 12 X 120 mm., each with a 20-mm. well pushed into the lower end to accommodate a chromel-alumel thermocouple. The tubes were heated in a furnace to 200" and the cooling curves were obtained on a Sargent recorder. The preliminary n.m.r. spectrum was obtained on a Varian A60 spectrometer with variable-temperature probe.3 The TMS standard is of course not feasible a t 150" and no other was used.
Acknowledgment.-The
authors wish to thank the
U. S. Army Research Office (Durham) for support, and Mr. Murray Margolis and Mrs. Edna RiI. Luck4for assistance in the laboratory. (3) We are indebted t o M r . Robert A. Pages for this determination. (4) National Science Foundation summer research participant, 1959.
A Convenient Preparation of S-Benzhydryland S-Trityl-L-cysteine' RICHARD G. HISKEYAND
JOHN
B. ADAMS,J R . ~
Venable Chemical Laboratory, The University of North Carolina at Chapel Hill, Chapel Hill, North Carolina Received December 10, 1964
In connection with our work on the synthesis of unsymmetrical cystine peptides, a simple, rapid method for the preparation of S-benshydryl-(DPM-) (I) and Strityl-L-cysteine (11) was desired. The method should give good yields and be readily adaptable to large-scale laboratory operations. The best published procedures3 for the preparation of I and 11, from L-cysteine hydrochloride or tosylate, suffer from the disadvantage of the rather low crude yields of I (53%) and I1 (75%). In view of these limitations the present procedure was devised. The method involves the direct Salkylation of L-cysteine hydrochloride using the ap(1) Supported in part by Grant A-3416 from the Institute of Arthritis and Metabolic Diseases of the National Institutes of Health, U. S. Public Health Service. (2) Shell Chemical Carp. Fellow, 1963-1964. (3) L. Zervas and I. Photaki, J. A m . Chem. Soc., 84,3887 (1962).
VOL.30
propriate alcohol and boron trifluoride etherate in acetic acid. The scheme provides high yields of pure products and, as illustrated in the synthesis of I, is applicable to large-scale synthesis. In this regard it should be noted that the present procedure for the preparation of I is superior both in yield and simplicity to the scheme classically used for the synthesis of Sbenzyl-~-cysteine~ (111). Experimental5 S-Benzhydryl-L-cysteine (I).--Into a 2-1. erlenmeyer flask were placed 157.6 g. (1 mole) of L-cysteine hydrochloride6 and 1 1. of glacial acetic acid. The mixture was heated on a steam bath with occasional swirling until the temperature reached 60', whereupon 184.2 g. (1 mole) of benzhydrol was added and the temperature was again brought to 60'. Then was added in one portion 140 ml. ( a 10% excess) of boron trifluoride etherate, and the mixture was heated and swirled for another 15 min. while the temperature rises to 80'. The thick mixture was transferred to a 4-1. beaker with the aid of 1500 ml. of ethanol, 500 ml. of water was added, and the mixture was stirred until homogeneous. The solution was treated with 300 g. of anhydrous, powdered sodium acetate, added in one portion with rapid stirring. The mixture was cooled to 10' and filtered, and the product was washed successively (and thoroughly) with water, absolute ethanol, and ether. The product was dried in vacuo over phosphorus pentoxide and sodium hydroxide and appeared as 259 g. (90yo) of white, odorless solid: m.p. 206-207" dec., unchanged on one recrystallizationa (86% recovery); [a]2 2 +15.2" f 0.3" (c 1.7, 0.1 ili ethanolic HCl); reported3 m.p. 202-203", [ Q ] * ~ D4-16.9" ( c 2.9, 0.1 N ethanolic HC1). Thin layer chromatography of I ("crude") on silica gel G shows (ninhydrin or iodine vapor) one spot. With silica gel GFZuone spot is revealed under ultraviolet light. Ascending paper chromatography (Whatman No. I ) shows (ninhydrin) one spot, Rr 0.92, and a faint trare, Rr 0.12. After the one recrystallization the compound was chromatographically homogeneous. Anal. Calcd. for C ~ ~ H I ~ N O ZC, S :66.86; H , 5.96; N , 4.87; S, 11.16. Found (for "crude"): C, 66.48; H , 5.84; N, 4.87; S, 11.28. Found (for recrystallized): C, 66.78; H , 5.63; N, 4.57; S, 11.42. S-Trityl-L-cysteine (II).-By the procedure previously described for the preparation of I , 1.58 g. (0.01 mole) of L-cysteine hydrochloride and 2.60 g. (0.01 mole) of trityl alcohol in 10 ml. of glacial acetic acid were treated with 1.40 ml. ( a 10% excess) of boron trifluoride etherate. The mixture was warmed 30 min. on a steam bath, kept a t room temperature for 45 min., and transferred to a beaker with 15 ml. of ethanol. The solution was treated with 5 ml. of water and 3 g. of powdered, anhydrous sodium acetate. The addition of 40 ml. of water provided a gum which solidified when triturated with cold water. After successive washings with water, acetone, and ether, the product w&s dried in vacuo over phosphorus pentoxide and sodium hydroxide and appeared as 3.08 g. (85%) of 11, m.p. 181-182' dec. One recrystallization from N,N-dimethylformamide-water raised the melting point to 183.5' dec., [aIz4o+114 f 2' ( c 0.832, 0.04 N ethanolic HCl); reporteda m.p. 181-182', [CY]~'D $108' ( c 1.45, 0.04 N ethanolic HCl). Anal. Calcd. for C ~ ~ H ~ I N O ZC, S :72.69; H , 5.82; N, 3.85; 5 , 8.82. Found: C, 72.20; H, 5.97; N, 3.99; S, 8.96. Thin layer chromatography of the "crude" material (m.p. ISl-18Z0) revealed one spot, as did paper chromatography of the recrystallized material, Rr 0.92. (4) Conditions for the removal of the S-benzhydryl ( D P M ) group from I using either sodium in liquid ammonia or refluxing trifluoroacetic acid are described in ref. 3. ( 5 ) Melting points are uncorrected a n d were taken in capillary tubea. Elemental analyses were performed by the Triangle Chemical Laboratories, Chapel Hill, N. C . Optical rotations were taken with a Rudolph polarimeter, Model 80, equipped with a Model 200 photoelectric attachment. All chromatographic procedures were carried out in the 1-butanol-acetic acidwater ( 4 : 1: 5) system. The L-cysteine hydrochloride wae obtained aa the C.P. monohydrate from the Mann Research Laboratories, New York, N. Y. (6) M. Bergmann and G . Michalis [Ber., 68, 987 (1930)l describe the conversion of the monohydrate to the anhydrous form.
~
