Dec. 5, 1961
2,2-AZOBISISOBUTYFUMIDINEHYDROCHLORIDE IN SOLUTION
4849
[CONTRIBUTION FROM YERKESRESEARCH LABORATORY, FILM DEPARTYENT, E. 1. DU P O N T DE NEMOURS AND CO., INC., BUFFALO,N. Y.]
Chemistry of 2,2'-Azobisisobutyramidine Hydrochloride in Aqueous Solution : A
Water-soluble Azo Initiator1 BY THOMAS J. DOUGHERTY RECEIVED APRIL8, 1961 Thermal decomposition of 2,2'-azobisisobutyramidine hydrochloride (ABAHs++2Cl-) in water has been shown to yield initially the product of free radical coupling, tetramethylsuccinamidine hydrochloride (I), which undergoes cyclization to tetramethylsuccininiidine hydrochloride (11). Hydrolysis of the imidine to 3,2,3-trimethyl-3-guanylbutyramide hydro(IV), and tetramethylsuccinimide (V)has been observed. The possichloride (111). tetramethyl-fj-imino-2-pyrrolidone bility of tetrarnethylsuceinimidine hydrochloride in the conjugated tautomeric form (IIa) as an intermediate is discussed.
Introduction There are several advantages in the use of watersoluble, charged azo initiators for polymerization reactions. While the salts of 2,2'-azobisisobutyramidine2 offer the desired combination of labile azo function and water solubility, the amidine moieties introduce the possibility of several side products which may be undesirable in the polymerization process and/or in the resulting polymers. A study was thus undertaken to elucidate the chemistry of 2,2' - azobisisobutyramidine hydrochloride (ABL4H2++2C1-) under conditions similar to aqueous phase polymerization (excluding monomer).
species absorbing a t 292.5 mp. Addition of hydrochloric acid completely retarded the formation of this adsorption without affecting the rate of azo loss (Table I). The rate of amidine consumption was followed iodometrically by application of a procedure developed by Stevens for quantitative analysis of certain aromatic amidiness (see Experimental). In the absence of added acid the change in absorption a t 292.5 mp with time could be followed (Fig. 1). The initial slope is concentration dependent, but the position of the maximum (approximately 300 min.) is concentration independent at 60". The short induction period was shown not to be a result of radical scavenging by oxygen HJJ+ CHI CH, +NH2 by carrying out the reactions in a carefully deoxy\ I I / / genated system. Deliberate addition of air tended C-C-N=N-C-C 2 c1/ I I \ to lengthen the induction period somewhat but H2N CHI CH: "2 not appreciably affect the remainder of the curve. ABAHi' +2ClProduct Identification.-Product isolation waq Results carried out under various conditions. Refluxing Kinetic Data.-To determine if the amidine func- aqueous ABAH:,++2C1- for 3.5 days resulted in tions were undergoing reaction a t a rate comparable a 60% yield of tetramethylsuccinimide V, (identito decomposition of azo, rate constants for both fied by elemental analysis, m.p. and infrared). Heating a solution of ABA4H~++2C1-for one reactions were measured at several temperatures in demineralized, deoxygenated water. The re- hour on a steam-bath gave a 45y0 yield of 2,2,3actions followed simple first-order kinetics. Table trimethyl-3-guanylbutyramide hydrochloride 111, (identified by elemental analysis, infrared and forI summarizes the results. mation of tetramethylsuccinimide upon hydrolysis TABLE I under neutral or basic conditions), KINETICDATAFOR LOSS OF AZO AND LOSS OF AMIDINE Partial decomposition by heating an aqueous Azo loiP Amidine l o d solution of ABAH2++2CI- in a constant temperaT,O C . k X 104, min.-lo i X 104. min.-lo ture bath at 60" for 300 minutes (approximate 40.15 .. 1.5 half-life for azo loss) permitted isolation of several 50.05 5.5 5.9 products. Continuous ether extraction yielded 60.20 24.4 24.3 trace amounts of tetramethylsuccinirnide (V). 70.10 92.7 91.3 Removal of the water under vacuum followed by 100.10 3990 ... continuous chloroform extraction of the resulting AH*, kcal./mole 29.1 29.3 solids (4 days) yielded three fractions4: a chloroAS*, e x . +5.5 S9.2 form non-extractable fraction (80% of total) conIn all runs equimolar or 100% mole excess HCl was sisting of unreacted ABAH2++2C1-, the amideadded. Rate constants were independent of amount of amidine salt 111, and ammonium chloride; a added acid. * T h e actual rate of amidine loss is twice that of azo since there are two amidine functions per mole- slightly chloroform-soluble mixture (17%) which Rate constants are averages of two or more runs. separated as an oil, consisting of I11 which slowly cule. crystallized out upon standing5 and the interrneDisappearance of the azo absorption in the ultraviolet at 365 mM afforded a convenient means for &ate as shown by absorption a t 292.5 mpO; and (3) F. H.Stevens, A n d . Chrm., 24, IS0 (1952). following the rate of azo decomposition but was (4) Thiv procedure was first carried out by W. R . Hendrix of this complicated by the appearance of an interfering department. (1) Presented before the 139th Meeting of the American Chemical Society, St. Louis, Mo., March, 1961, Abstracts, p. 17-0. (2) For a discussion of the syntheses of these cornpounda and their use in polymerization, see R. W. Upson, U. S. Potent 2,690,299 (1952).
(6) Since 111 is very slowly extracted by chloroform, the actual amount in the oil depends upon the length of time of the extraction. (6) An approximation of the rate of disappearance of the intermediate in water was determined spectrophotometrical!y using this oil. The reaction is first order with b o o 3 . 0 4 . 0 X 10-*min.-I.
-
THOMAS J. DOUGHERTY
4850
tlmin. j.
of change of optical density a t 292.5 mp. T = 60.2', [ABAH2++2Cl-]a = 0.0103 M .
Fig. 1.-Rate
the third fraction, a chloroform-soluble material (trace) identified as tetramethyl-5-imino-2-pyrrolidone (IV) by elemental analysis, infrared and hydrolysis to tetramethylsuccinimide under neutral or basic conditions. Tetramethylsuccinamidine hydrochloride (I) is probably a constituent of the chloroform-insoluble material (andlor the slightly chloroform-soluble oil) although i t was not specifically identified. No products resulting from disproportionation were found, although minor amounts may be formed. His+
CH3 CHI +K"?
I / / I \
\ \ I
c-c-c-c
/
I
HzN
CH3 CH,
2
c1-
NHz
1
0
\
l
CH3 CHI +NHz
l
c-c---c-c
/
c1-
Vol. 83
and infrared identical to that of picrate from authentic 111) indicating that hydrolysis may proceed by ring opening. A sample of tetramethylsuccinimidine nitrate7b examined in this Laboratory showed a maximum in water a t 292.5 mp with E 101. The magnitude of absorption in this region, however, was found to be pH sensitive as expected on the basis of its absence during kinetic runs in acid solution. When M solution of the imidine nitrate in a 2.03 X 302.5 mp, E 258)s was dimethyl sulfoxide"::A:( acidified with aqueous hydrochloric acid to an N , the optical density slowly extent of 2 X decreased from an initial value of 0.525 to 0.358 in 18 hours. Further acidification of this solution to a concentration of 0.2 N resulted in a drop of optical density to 0.02 in 8 minutes. When a solution of approximately the same concentration in anhydrous dimethyl sulfoxide was strongly acidified (-0.1 N in acid) with anhydrous hydrogen chloride, the optical density decreased to a constant low value (-0.05) in 10 minutes. A solution of the imidine nitrate in ethanol (A":, 298 mp) which was acidified to 0.2 N with 70% nitric acid resulted in almost complete loss of absorption after 10 minutes. The material isolated after 30 minutes (>go%) had an infrared spectrum essentially unchanged from that of the starting iinidine nitrate with the exception of the presence of a weak band a t 9.2 p. When a similarly acidified solution remained overnight, the isolated material had an infrared spectrum very different from the imidine nitrate (see Experimental) . After 20 hours the optical density of a solution of the imidine nitrate in absolute ethanol had dropped to approximately 70y0of its original value.
Discussion
AHN
O J \'
1'
Tetramethyl-5-imino-2-pyrrolidone (IV) was obtained in greater amounts from a saturated aqueous solution of ABAH2Cf2C1- which had been exposed to sunlight for approximately six months (see Experimental). (HdC"2W(CHj)L
€I.\
If
I1
Tetramethylsuccinimidine nitrate (11, X = NO3-) has been shown to be a product from thermal decomposition of 4 B h H ~ f T 2 N O a -in anhydrous dimethyl sulfoxide by Hammond and Neuman.'" When a of this compound was allowed to stand several days in water a t room temperature, 2,2,3-trimethyl-3-guanylbutyramidehydrochloride (111) was isolated as the picrate (m.p. (7) (a) G. S.Hammond and R . C. Neuman, Jr., private communication. (b) The author is indebted to these wnrkers for this sample of t h e imidine nitrate
The kinetic data (Table I ) indicate that both loss of azo and loss of amidine are occurring with the same rate-controlling process ; rate constants and thermodynamic parameters are identical within experimental error for both reactions. The relatively large enthalpy of activation (-29 kcal./ mole) and favorable entropy of activation (+8-9 e.u.) are indicative of a chain-scission process such as decomposition of azo compounds. These values may be compared with an enthalpy of activation of 31 kcal./mole and an entropy of activation of 11.7 e.u. for decomposition of 2,2'-azobisisobutyrnnitrile in b e n ~ e n e . ~It thus appears that the amidine functions are consumed in a fast step subsequent to rate-determining azo decomposition. The type of behavior noted for the material with 292.5 mp (Fig. 1) is characteristic of an intermediate (C) whose precursor is itself an intermediate (B).
+
ka
ABAH2++2Cl- + [intermediate B]
kb
+ ko
[intermediate C] +D (8) This sample was prepared in this Laboratory by a procedure similar t o Hammond and Neuman (see Experimental). (9) Calculated from data of P. Smith and S . Carhone, J. .4m. Cham. Sac. 81, 6174 (1959).
Dec. 3, 1961
2.2-L\ZOBISISOBcTYRAMIDINEHYDROCHLORIDE Ih’ SOLVTIOh’
485 1
The kinetic expression for C is given by eq. 1 which integrates to eq. 2.1°
[Cl =
(2)
where [C] is the concentration of intermediate C a t time t and [AIo is the initial concentration of XBAHzi +2C1-. The short induction period indicates that while k b > k a , it is not sufficiently different to allow the concentration of B to be treated by the steady-state approximation. A curve calculated according to eq. 2, which has the same form as the experimental curve (Fig. l), is obtained using values obtained a t 60”for ka (Table I) and kc6 of 2.44 X rnin.-’ min.-l respectively, and setting and 4.0 X k b = 2.4 x 10-2min.-1. The concentration of C a t the maximum may be estimated from eq. 3 (obtained from eq. 1by setting
Hac
\ dc/dt = 0). Setting [Ala = 0.0103 (Fig. 1) and using the above values for rate constants, [C],ax = 3.34 X M . Assuming adherence to Beer’s law and using the value of 4.211for the optical density corresponding to [ C l m a x (Fig. I), B = 1260. Since e a t 292.5 mp for the sample of tetramethylsuccinimidine nitrate obtained from Hammond and N e ~ m a nwas ~ ~101, i t may be estimated that the material exhibiting this absorption constitutes a maximum of 8% of the sample and that the sample obtained in this Laboratorys a maximumof 20% (e = 258).12 Two structures for the intermediate may be considered based upon its ultraviolet spectrum. Dimethyl-N-(2-guanyl-2-propyl)-ketenimine hydrochloride (VIII) could be formed as a result of carbon to nitrogen coupling of radicals13 according to Chart I . Since product isolation was not quantitative, failure to observe the hydrolysis product of VII, dimethyl-N-(2-guanyl-2-propyl)-isobutyramide hydrochloride (VIII)14 or cyano - containing compounds resulting from homolytic cleavage,13 does not rule out the mode of decomposition as outlined in Chart I. (10) These equations are derived similarly to eq. 15; A. A. Frost and R . G. Pearson, “Kinetics and Mechanism,” John Wiley and Sons, Inc., New York, N. Y.,1953, p. 154,with the exception that C is not a stable product. (11) This value was obtained by multiplication of the optical density of diluted samples by the dilution factor. (12) Equations 1-3 assume that the reaction proceeds entirely through the intermediate leading to a maximum value of [CImax and a minimum value for e. The concentrations of the intermediate in these samples may thus be much less than estimated. (13) Dimethyl-N-(2-cyano-2-propyl)-ketenimine, [ k k r 291 mp, c 150.5; P. Smith and M. Rosenberg, J. A m . Chem. So;., 81, 2037 (1959)I is an intermediate in decomposition of 2,2’-azobisisobutyronitrile in non-aqueous solution. This ketenimine may break down t o re-form 2-cyano-2-proppl radicals [M. Talht-Erben and S. Rywater, ibid., I ? , 3710, 3712 (1955)l. (14) Certain phenyl-substituted ketenimines are relatively stable in neutral solution but rapidly hydrolyzed t o the corresponding amides b y acid: H. Staudinger and 1. Meyer, Ber., I S , 62 (1920),and e. bb Stevenn and J. C . French, J . A m . Chem. Soc., 7 6 , 057 (1953):
0
CHs
11
CH-C-XH-C-C
/
H3C
I / / I \
CH3 VI11
XH?+
c1NH:!
The ketenimine structure however may be ruled unlikely on the basis of the estimated extinction coefficient a t 292.5 mp for the intermediate. The value of 1260 is an order of magnitude greater than the extinction coefficient of 150 a t 291 mp measured for essentially pure dimethyl-N-(2cyano-2-propyl) -ketenimine. l 3 l5 A second structure to be considered is the conjugated tautomer of tetramethylsuccinimidine hy-
IIa
drochloride (IIa, X = Cl-). Elvidge and Linstead16 have pointed out that structures of type I I a would be expected to absorb in the 300 mp region by analogy with compounds of type I X and X which show high intensity absorption between 280-300 mM. 17,l8
\ I C=C-CH=CH-N / IX
Et2
I
kC-CH=C-K X
I
Et2
Although tautomeric forms such as I I a have not to date been shown to exist in equilibrium with the diimino forms such as 11, Elvidge and Linstead observed that the fine structure of N,N’-diphenylsuccinimidine hydrochloride hemihydrate in the ultraviolet shows a maximum a t 312 mp in methanol ( E 23,000) indicating that there may be some contribution from the conjugated form corresponding to IIa. (15) A similar extinction coefficient would be expected for ketenimine VII. (16) J. A. Elvidge and R. P. Linstead, J . Chem. SOL.,442 (1954). (17) E. A. Braude, Ann. Refits., 42, 105 (1945).and J . Chem. Soc., 45 (1946). (18) Substitution of nitrogen for carbon has little effect on the position of maximum absorption in 1,3-diene systems (H. C. Barang, E.A, Braude and M. Pranha, ibid., 1898 (1949)).
48.52
THOMAS
J. DOUGHERTY
If ahsorption in the 300 mp region is due to the conjugated tautomer XIa, i t is interesting to speculate on the acid-catalyzed reactions. I n aqueous solutions this may simply involve acidcatalyzed hydrolysis. lo In non-aqueous solution, however, the disappearance of absorption near 300 mp may result from acid-catalyzed isomerization of form TIa to form 11.
Vol. 83 CHARTI1
II
IIa
This postulate implies however that rapid proton exchange (as expected a t least in aqueous solutions) should lead to rapid isomerization. Therefore, it would not be expected that the intermediate would be sufficiently long lived to be observed in kinetic measurements. Recent work has shown that rates of proton exchange of salts of ammonia, methylamine and trimethylamine may become quite slow in very acidic solution.20 The imidine salts may be representative of systems showing relatively slow rates of proton exchange even in neutral or slightly acidic s o h tion. -4n over-all. scheme consistent with product identification and kinetic data is shown in Chart I1 (dotted lines represent uncertain reaction paths). Although the diarnidine salt I, may or may not be a sufficiently short-lived intermediate to a'low neglect of its rate in the rate expression for loss of amidine (which occurs upon cyclization to imidine 1121,22), i t is likely that i t would appear to be so as a result of the particular method employed for determining amidine (see Experimental). Thus, the amide-amidine 111, does not analyze for amidine according to the method used in kinetic determinations perhaps ns a result of excessive stesic crowding around the amidine function. Arguing by analogy i t would be expected that the diamidine I: in which the amidine functions are similarly crowded, would also fail to give the amidiiie test by this method. It would appear therefore that amidine had been consumed upon radical coupling to I (a very rapid step), leading to a rate constant for amidine loss identical to that for azo decomposition.23
Experimentale4 2,2'-Azobisisobutyramidine hydrochloride
(ABAHs
+-
2C1-), prepared by addition of ammonia to the imino ether of
2,2'-azobisisobutyronitrile
according to reference 2, was
(10) An uncatalgzed hydrolysis may account lor consumption of intermediate observed in kinetic measurements (Fig. 1). (20) M . T. Emerson, E . Grunwald, M. L. Kaplan and R . A. Kromhout, J . Am. C k d m . SOL, 82, 6307 (1960). Mean lives of the N-€I covalent bond in N€L+ranged from 0.16 sec. in dilute, slightly acidic solution to 15 hr. in 69.38% HrSOd. (21) Cyclization to succinimidine hydrochloride occurs upou dissolution of siiccinamidinr hydrochloride in water. A. Pinner, "Die lmidoather und ihre Derivate," Oppenheim, Berlin, 1892. (23) If t t i s cyclization is not very rapid, its rate should appear in ihe rate expression for loss of amidine but not of course in that for azo decomposition. (23) T h e kinetics for amidine loss show no evidence of different rates for the two functions up to 75'30 reaction. While this is not definitiv: it is consistent with the mechanism shown. (24) All melting points are imcorrccted. Elemental Rnalvses bv Schwarzkopf Laboratories, Woodside, N. Y.
iecrystallized by preparing a saturated solution in water a t room temperature and storing several hours at 3". Kinetic solvent, demineralized, deoxygenated water was prepared by passing tap water through a commercial ion exchange resin and bubbling in nitrogen. Kinetic Methods.-Azo loss was determined by following the change in optical density at 365 mp (Beckman model DU spectrophotometer) of ABAH2++2C1- in solutions of the kinetic solvent (0.02 to 0.10 M ) in the presence of equimolar of 100% mole excess of hydrochloric acid. The rate was independent of the amount of added acid. Individual samples of 2 to 4 ml. in capped polypropylene test-tubes were used. Linearity of optical density with concentration of ABAH?+%?CI- was demonstrated, Since in the presence of acid none of the products absorb in the 250400 m s region, the final optical density was taken as zero. Plots of log optical density vs. time gave straight lines t o 90% reaction. Substitution of distilled water for demineralized. deoxygenated water had Iittle effect on the rates of azz loss. Rate constants at 50.05, 60.2, 70.1 and 100.1 were calculated from the slopes of the best straight lines drawn visually through the experimental points (Table I). Thermodynamic parameters of activation, AH and AS*, were calculated from the Eyring equation.2J .Amidhe disappearance was followed iodometricdlly.g A typical run consisted of withdrawing individual %mi. (room temperature volume) samples of ABAH.2""2C1solution (0.05 to 0.11 M in kinetic solvent) in polypropylene test-tubes, cooling rapidly in a stream of cold water, adding to excess standard iodine solution followed immediately by excess 2 N sodium hydroxide solution. After stixring 5 to 10 minutes26the fine yellow precipitate of iodoamidiiie was filtered off, washed several times with distilled water, the combined iiltrate and washings acidified with 4 N hydrochloric acid and excess iodine titrated with standard sodium thiosulfate solution. The analytical method was checked by determining known amounts of ABAHtt+2C1-.
*
(2.5) A. A. Frost and R . G.Pearson, ref. 10, p. 08, eq. 40. (26) IK stirring was continued much longer than 10 minutes, the precipitate turned dark and erratic results were obtained.
BENZYLDIMETHYLANILINIUM AND THIOCYANATE IONS
Dec. 5, 1961
4853
Plots of log [ABAH2++2Cl-] wersus time gave straight with chloroform for 4 days. An infrared spectrum of the lines to 75% reaction. Rate constants were determined a t non-dissolved solid (6.6 g., 80%) indicated ABAH2++2Cland 111. Ammonium chloride is also found in the chloro40.0°, [8.05', 60.2", 70.1 " and are the average of at least form-insoluble fraction. A slightly soluble oil (1.27 g., two runs (Table I). The rate of change of optical density of the intermediate 17%) which separates from the chloroform solution slowly 0.2:; 292.6 mg) was followed spectrophotometrically a t deposited crystals of ILI upon standing. The ultraviolet 00.2" a s described for azo loss with the exception that no spectrum of this oil in water showed the presence also OF the intermediate absorbing a t 292.5 mp. Evaporation of the acid was added and initial ABAH2++3Cl- concentrations were in the range O.OO&C).OlO iM. In addition to using chloroform solution yielded 0.015 g. of a white solid whose deoxygenated demineralized water, the test-tubes were infrared spectrum and m.p. are identical to tetramethyl-8imino-Spyrrolidone (IV) which was ohtained in greater flushed with nitrogen before capping. Plots of optical amounts lrom the photochemical degradation of ABAHz++density wersus time g o e curves with initial slopes dependent upon [ABA1&++2Cl-]o and maxima a t approximately 300 2C1- (see below). Photochemical decomposition of ABAH2++2Cl- was minutes (Fig. 1). carried out by storing in the sunlight of the laboratory a The constant temperature bath used in all runs regulated The spreads of rate constants are approximately solution of ABAHzf+2CI- (270 g., 0.995 mole) in 2 1. of a t f0.05'. deoxygenated, demineralized water contained in a poly3 ~ 5 from 7 ~ the means. ethylene bottle. After nearly 6 months thcre had pieDecomposition of ABAHzf+2CI- to form 2,2,3-trimethyl3-guanylbutyraniide hydrochloride (111) was carried out by cipitated approximately 20 g. (13% yield) of a white crystalline material, m.p. 285-288" dec., with NH band in the dissolving ABAH2++2Cl- (5.28 g., 1.9Fj X lo-' mole) in 100 ml. of distilled water and heating 1 hour on a steam- irifrared a t 3 . 0 8 ~and bands a t 5.90(s) and G.l(m) I.(. Hydrolysis of this compound under neutral (refluxing bath. The solution was cooled, the water partially removed in uucuo, and stored overnight a t 3". The white crystalline water, 24 hr.) or basic conditions (o\wnight, room temprecipitate amounted to 0.74 g. An additional 1.08 g. was perature) produced tetramethylsuccinimide. This conipound is assigned t h e structure te:rarcethyl-5obtained by further concentration in uucuo and cooling for a total yield of 467c. This yield is a minimum value since imino-2-pyrrolidone (IV). evaporation of the remaining water gave a mixture of amAnal. Calcd. for CsHlJTaO: C, G2.3; H, 9.1; K, 1Y.2. monium chloride and 111. The compound was recrystal- Found: C,62.3; H,9.2; N, 18.1. lized from cthanol-ether or water, m.p. >300°. Ionic T e t r a m e t h y l s u c c i d i n e nitrate was obtained by a prochloride was indicated with silver nitrate solution. The cedure similar to that of Hammond and Nei:mnnia: ABinfrared spectrum showed NH at 2.89 p and bands a t 5.90- AHz++2NO8- (0.3 g., 9.26 X IO-' mole, obtained by addi(m) and 6.10(s)I.(. tion of aqueous AgNO, to a solution of ABAH2++2CI-) was Hydrolysis of this salt in refiuxing water (2 days) gave mixed with 2 ml. of anhydrous dimethyl sulfoxide and heated tetramethylsuccinimide (by ether extraction) in 40% yield to 95'. Dissolution of material and loss of nitrogen oc(m.p. 1%3-J89°, characteristic carbonyl bands curred. After approximately 20 minutes, the solution was a t 5.60 and 5.85-5.9011). On this basis. the comDound is cooled and mixed with approximately 6 ml. of benzene. Just assigned the structure 2,2,3-trimethyl-3~guanylbu~yramide sufficient ethanol was added to give a homogeneous solution. hydrocliloride (111). After standing overnight, the white precipitate was filtered A~zul.Calcd. for C&&LlO: C, 45.4; IS, 8.8; N, 20.3; off. washed with ether and dried to yield 0.OOG g. of tetra0, 7 . 7 ; Cl, 17.1. Found: C, 46.3; 13, 8.6; N, 20.3; methylsuccinimidine nitrate (3% yield) whose infrared 0, i.8; C1, lG.9. The picrate has m.p. 262-265'. spectrum was identical to that of the sample supplied by Decompositior. of ABAHZ ff2Cl- -to form tetramethylHammond and Neuman.7b In dimethyl sulfoxide the succinirnide was accomplished by dissolving ABAHz++- sample showed XmSx 302.5 mp, c 258. To a sample of tetramethylsuccinimidine nitrate'b (0.0040 mole) in 100 ml. of distilled 2C1- (10.0 g., 3.69 X water and refluxing 3.5 days. The solution was cooled, g . , 1.88 X mole) in 5 cc. of ethanol was added 3 drops partially concentrated in wucuo and stored several hours a t of 70% nitric acid. After 30 minutes the solvent was evap3" to yield 2.17 g. of tetramethylsuccinimide. An addi- orated and the solids dried overnight to yield 0.0037 g. tional 1.2.5 g. was obtained by extracting with ether to give (92.5p/o) of material with an infrared spectrum essentially a total yield of 60%. identical to starting imidine (N-H band at -3.0(s) p , Decomposition of ABAH3++2CI- to SOc% reaclion was other bands a t -5.95(w) and -G.l(m) p ) with the excepeffected by dissolving AUAHZf+2C1- (10.0 g., 3.69 X 10-z tion of an additional weak band at -9.2 / I . mole) in distilled water and maintaining a t 60' for 300 A similar sample left overnight in the acid solution yielded minutes (approximate half-life for azo loss). The solution a material in which tlie 5.95 and G.1 p peaks had reversed mas cooled and extracted continuously with ether for 6 days in intensity. t o yield 0.1 g. of tetramcthylsuccinimide. The aqueous Acknowledgments.-The author wishes to thank solution was cvapoiated to dryness (in vacuo, low temperature) and the resulting solids (T.0 g.) continuously extracted Drs. J. I,. Hechi, J. J. Sparapany, D. F. Barringer (27) R. Auwers and J. A. Gardne, Der., 23, 3622 (ISeO), give m.p. 187'.
[CONTRIBUTION PROM
THE
Jr., and G. S. Hammond for suggestions and helpiul discussions concerning this work.
ICESEARCH LABORATORIES OF THE SFRAGUE ELRCTXIC Co., IL'oar~iADAJIS,Ll.4ss.J
Changes in Reaction Order Due to Ion Association in the Reaction of Benzyldimethylaniliniurn Ion and Thiocyanate Ion BY SIDNEYD. Ross, MANUEL PINKELSTBIN AND RAYMOND C. PETERSEN RECEIVED MAY29, 1961 The rates of reaction of benzyldimethylaniliniurn ion and thiocyanate ion to form benzyl thiocyanate and dimetliylaniline have been studied in four solvents, N-methylpropionamide, benzyl alcohol, s-tetrachloroethane and chloroform. There are significant changes in reaction order in this series of solvents. These variations can be explained by considering the appropriate ion association cquilibria and the specific rates for the reactions of ions and ion pairs.
In a previous report from this Laboratory,' it was shown that tlie rate of reaction of benzyldi( I ) S. D. ICoss, M. Finkelstein and R. C. Pctrscn, J . Am. Chrm. Soc.. 81, G335 (1960).
methylanilinium ion with ethoxide ion in ethanol varies with both changing initial concentrations and neutrd in a manner which is not satisfactorily accommodated by the Bronsted
