4932
OXAR 51:. STEWAKI) .IXD OGDENR. PIERCE
glycol column resolved carbon tetrachloride, t-butyl alcohol and t-butyl acrylate from a mixture of acrolein and acetone. The acrolein-acetone pair was resolved on P,B'-oxydipropionitrile. The blip corresponding to t-butyl acrylate was trapped ten times from the triethylene glycol column. Its infrared spectrum was identical with the spectrum of authentic gaseous t-butyl acrylate.Zd
[CONTRIBUTION FROM
THE D O W
A&nowledgment.-The
Vol. 83 authors are indebted to
>/lr,F.L. ~~d~~~~for patient assistance with the
experimentation. 128) This substance was prepared in poor yield from acrylyl chloride and l-butyl alcohol in t h e presence of P\TazCOj; b.p. 56' (80 m m . ) , l t D ? O . & 1.1102.
CORNING CORPORATION,
?dIDLAND, LfICH
]
The Effect of Polar Substituents on the Acid-catalyzed Hydrolysis of Organosilicon Hydrides BY
OhIAR
\v. STEWL4RDA S D OGDEN R.PIERCE RECEIVED JULY 19, 1961
The first-order rate constaiits for the acid-catalyzed hydrolysis of sixteen fluoroalkyl-, w-cyanoalkyl- and n-alkylsilicon hydrides have been determined using aqueous ethanolic hydrochloric acid (1.43 N , 95 v o l . - ~ethanol) , a t 34.8". The firstorder rate constants are correlated using the Taft equation. The correlations are discussed in terms of polar effects, steric effects and dative r-bonding.
Introduction The polar effect of substituted alkyl groups on the alkali-catalyzed hydrolysis of triorganosilicon hydrides has been reported previously.' An excellent correlation of the second-order rate constants mas obtained using the Taft equation* for triorganosilicon hydrides containing fluoroalkyl, w cyanoalkyl and n-alkyl groups.la Other linear relationships involving the inductive effect of substituent groups in silicon hydrides have been reported: (a) Hammett's u-values and modified c-values zts. log k for the base-catalyzed (b) v SiH (cm.-') hydrolysis of triorgan~silanes,~ L I S . empirically determined E-values and Taft's u * - d u e s , 3 (c) v SiH (cm.-l) DS. 7 (p.p.ni.) for a proton attached to silicon,6 (d) v SiH (cm.-l) L I S . log AT (2.5') for the hydrolysis of triorganochlorosilanes.6 The acid-catalyzed solvolysis of organosilicon hydrides has been studied by Taketa, et al.,i and Baines and Eaborn.8 In this paper a study is reported on the effect of polar substituents on the acid-catalyzed hydrolysis of mono-, di- and triorganosilicon hydrides. The first-order rate constants are correlated using the Taft equation. The correlations are discussed in terms of polar effects, steric effects and dative 7-bonding Experimental Starting Materials.-Trichlorosilane, 3,3,3-trifluoroprop~-ltrichlorosiiane,9~~-tricliloroiil~-lpropionitrile.lo l1 0-inethyl(1) ( a ) 0 . \T, Steward and 0. R. Pierce, J . A m . Chewi. S s r , 8 3 , 1916 (1961); (b) 8 1 , 1983 (19.59). ( 2 ) R . XI-, T a f t , Jr , "Steric Effects in Organic Chemistry." R l . S . S e w m a n . Ed., John \Viler arid Sons, Inc , S e w 'iork, S V . , 1936, pp. 55G67.5. ( 3 ) ( a ) H Gilman and G . E. Dunn. .I. A m C h e m . .Soc., 73, 3404 (1931); lh) G. Schott and C . Hnrzdorf. %. d x o i g . I I a l l f r i i i . Cheiir., 306, 180 (1QbO) (4) (a) X I>.Smith and N. C Angelotti, Spe(li.oihiii? At.lo, 15, 412 (1959); (b) H. \I-. Thompson, i b i d , 16, 235 (19G0> 1);. D. F: Welister, J Chen7. Sor , 21:32 C1!160). (6) J F. Hyde. P. I.. Brown anti A. I, Smith, J . .1m Chrm L o . ,8 2 , 5x54 (1000) ( 7 ) A Taketa, RI. K u m a d a a n d I(. T a r a m a , Bull. 1 7 z ( ? Chrnr. R P s r a r i h , Kyalo C-7175 , 31, 260 1193:3). (8) J. E Baines and C . I k b o r n . J . Chriri. Src., 1436 (lQ.56). (9) P. Tarrant, G. W. Dyckes, R . Dunmire and G. B Butler, J . d VI. Chtvq C o r , 79, 6536 (1937).
dichlorosilylpropionitrile~10~11 chloroinethyldimethylchlorosi lane12 and 3-bromo-1,111-triflu~ropropanel~ were available in research quantities. They were fractionally distilled before being used. Preparation of 3,3,3-Trifluoropropyl-n-butyldiethoxysilane
and
Bis-(3,3,3-trifluoropropyl)-diethoxysilane.--Absolute
ethanol (191 g., 4.2 moles) was added to 3,3.3-trifluoropropyltrichlorosilane (463 g., 2.0 moles) over a period of 0.5 hour. Throughout the additiorio the reaction temperature was maintained betvi-een 50-60 . Sfter addition v a s complete, the mixture was heated to 120' and allowed to cool to room temperature. The reaction mixture was then stirred for 16 hours with a slow nitrogen sweep to remove any remaining hydrogen chloride. The material in the flask, 3,3,3-trifluoropropylethoxychlorosilanes(493 g.), analyzed for one chlorine per silicon, a 9870yield. A?znl. Calcd. for C7Hl4C1F30?Si:C1, 14.15. Found: C1, 14.4. n-Buty11nagneiium bromide, dissolved in ether (550 i d . ) , prepared from n-butyl bromide (116 g., 0.85 mole) and magnesium (20.i g., 0.85 mole), TTas added to the above prepared 3,3,3-trifluoropropylethosychlorosilanes (213 g., 0.85 mole), dissolved in ether (200 ml.), over a period of 0.5 hour. Throughout tlie addition the reaction mixture was cooled in an ice-bath. ilfter refluxing for 16 hours, the reaction mixture was poured onto a mixture of cracked ice and sodium bicarbonate (212 g., 2.0 moles). The ether layer was w.aslied and dried over anhydrous calcium sulfate. .After removing the ether, tractional distillation gave: 3,3,3trifluoropropyltriethoxysilane14(33.5 g., 0.13 mole), b.p. 82" (30 mm.), n z j ~1.3650, a 1570 yield; 3,3,3-trifluoropropyl-n-butyldiethoxysilane (146 g., 0.54 mole), b.p. 105' (30mm.), X Z ~ D 1.3858, d"., 0.991, a 6.77,yield. A1ml. Calcd. for C1,HZ3F302Si: C , 48.5; €1, 8.31; F, 20.9. Found: C,49.0; H,8.65; F.21.1. 3,3,3-TrifluoropropyIdi-n-butyletl~oxysilane (5.3 g . , 0.019 mole), b . p . 126.5' (30 m m . ) , n% 1.4073. dZs40.944, L: 27; yield. .,lnaZ. Calcd. for C13H2;F30Si: C, 51.9; H, 9.57; F, 20.0. Found: C,55.8; H, 10.1; F,19.9. Using the same procedure described above, 3,3,3-trifluoropropylmagnesium bromide, dissolved in ether (500 ml.), prepared from 3-bromo-l,l,l-trifluoropropane(150 g., 0.85 mole) and magnesium (20.i g., 0.85 mole), was added to the 3,3,3-trifluoropropylethoxychlorosilanes (213 g., 0.85 mole). Fractional distillation gave: 3,3,3-triilaneI4 (15.4 g., 0.06 mole), b.p. ~~
110) J. C. Saam and J. I, Speier, J . O Y ~Chenz., . 2 4 , 427 (1959). (11) G D. Cooper and 11.Prober, ibid., 2 6 , 240 (1960) (12) R. H Kriehle a n d J R . Elliutt, J , A n t . Chtrn. So( , 67, 1810 (lQ45) cl3) P.Tarrant, A . Ll, 1,ovelace and 11. R. Lilyquist, zbid., 77, 2783 (1925) (11) 0. W. Steward and 0 . R. Pierce. J . O v p . C h o n , 26, 2943 ( 1r l i i 1 )
Dec. 20, 1961
HYDROLYSIS OF ORGANOSILICON HYDRIDES
4933
V,-values determined experimentally were within 2% of 82" (30 mm.), n*'D 1.3651, a 7yGyield; bis-(3,3,3-trifluorothe calculated values. Straight lines were obtained from propyl)-diethoxysilane (120 g., 0.385 mole), b.p. 98" 40 to >SO?@ reaction. The less reactive silanes showed the (30 mm.), 712% 1.3578, d2541.153, a 4570 yield. A n d . Calcd. for C I O H I ~ F ~ O C ~S , ~38.5; : H , 6.13; F, largest deviations, probably resulting from small losses of hydrogen due to diffusion. The data obtained for n36.5. Found: C,38.4; H, 5.86; F,36.9. propyldimethylsilane tended to give concave-upward Preparation of the Organosilicon Hydrides.-The orgmo- curves, probably a result of the slow reaction rate and the silicon hydrides were prepared by the methods given below. volatility of the silane. Their physical properties and analyses are given in Table 11. a . Tris-( 3,3,3-trifluoropropyl)-silane was prepared by Results the addition of three equivalents of 3,3,3-trifluoropropylThe kinetic data for the acid-catalyzed hydrolymagnesium hromide t o trichlorosilane in ether solvent using the procedure of Price.I5 The reaction mixture was re- sis of the organosilicon hydrides are reported in fluxed for 16 hours. Table I. The physical properties, analytical data b . Chloromethyldimethylsilane16 \vas prepared by the and yields of the organosilicon hydrides not rereduction of chloromethyldimethy!chlorosilane M ith lithium aluminum hydride in ether using the procedure of West.17 ported in the previous papersia,bare given in Table c. P-Methplsilylpropionitrile was prepared by a reverse 11. lithium aluminum hydride reduction of P-methyldichloroThe first-order rate constants were determined silylpropionitrile using the procedure of Steward, et aZ.'"J* using 1.43 N aqueous ethanolic hydrochloric acid d . 3,3,3-Trifluoropropyl-n-butylsilane was prepared from (95 vol.-yo ethanol) a t 34.8' so they could be com3,3,3-trifluoropropyl-n-butyldiethoxysilaneas follows : In a 1-liter, %necked flask equipped with a stirrer, condenser pared with the data of Baines and Eaborn.8 The and dropping funnel, were placed lithium aluminum hyrate constant obtained for triethylsilane is within dride (9.5 g., 0.25 mole) and ether (400 ml.). The system one standard deviation of the value reported by was vented to the atmosphere aia a Dry Ice-cooled trap and Baines and Eaborn.s drying tube. The system was purged with nitrogen, and 3,3,3-trifluoropropyl-n-butyldiethoxysilane (109 g., 0.40 The rate constants for the reaction of the first mole), dissolved in ether (100 ml.), was added over a period and second hydrogens of the diorganosilicon hyof 1 hour. The reaction mixture was refluxed for 16 hours drides can be determined from a single run as reunder a nitrogen atmosphere, and then poured onto a misported by Baines and Eaborn.* In each of the ture of cracked ice and concentrated hydrochloric acid (200 1111.). The ether layer was dried over anhydrous compounds studied, the rate constant for the recalcium sulfate, the ether removed by distillation, and the action of the first hydrogen is from 10 to 15 times higher-boiling material fractionally distilled. 3,3,3-Trigreater than the rate constant for the reaction of fluoropropyl-n-butylsilane (39.1 g., 0.214 mole), was the second hydrogen. The nature of the species obtained in a 537@yield. e. Bis-(3,3,3-trifluoropropyl)-silane was prepared by from which the second hydrogen is removed is not the reduction of bis-(3,3,3-trifluoropropyl)-diethoxysilane known. with lithium aluminum hydride using the above procedure. The rate constants for the reaction of the first Apparatus and Kinetic Procedure.-For the rate studies, the following apparatus and procedure were used. A 250- and second hydrogens of 6-silylpropionitrile are ml., three-necked flask x a s equipped with a Lex magnetic approximately equivalent. A first-order plot which stirrer,lg a self-sealing rubber stopper, and a glass outlet assumes equal rates for the reaction of the first tube leading t o a mercury-filled, 100-cc. gas buret and maand second hydrogens (7253, reaction) is slightly nometer. The bottom of the gas buret was connected to a balance tube. A stopcock located a t the bottom of the concave-upward, indicating the rate constant for buret allowed the system to be adjusted to atmospheric the reaction of the second hydrogen might be pressure by removing mercury. The flask was maintained slightly greater. The rate of reaction of the first a t 34.8' by a constant temperature bath. Throughout: single run, the bath temperature was constant within 0.1 ; and second hydrogens is approximately 72 times faster than the initial rate of reaction for the third the temperature varied 0.3" from run to run. Aqueous ethanolic hydrochloric acid (1.43 N,95 v 0 1 . - ~ ~ hydrogen. The reaction rate of the third hydrogen ethanol)1g*was freshly prepared from a stock solution for decreases rapidly with time, probably a result of each run since the acid solution was not stable on standing over a period of days. The standardized acid solution (50 formation of a three-dimensional polymer. With cyclohexylsilane, Baines and EabornB ml.) was placed in the 250-ml. flask, and the system was allowed to reach thermal equilibrium. The silane sample observed a constant rate of hydrogen evolution was injected into the reaction flask from a weighed hypoduring the first 20% of the reaction and suggested dermic syringe. Readings were taken a t various times by that all of the hydrogens have similar reactivities. adjusting the system to atmospheric pressure. The error 'They observed the initial rate was proportional in the buret reading was approximately 0.1 cc., and the error in timing was about two seconds. During a run, the to the concentration of the cyclohexylsilane. room temperature was constant within one degree. However, from the rate data for p-silylpropioniThe volume of hydrogen evolved was taken as a measure trile, it would seem likely that the rate constant of the amount of silane which had reacted. The rate confor removal of the third hydrogen of cyclohexylstants mere determined from the slope of the line obtained by plotting log [ Vm - V , / V , X lo*] vs. t , where V, = silane would also be much smaller than the rate volume of gas evolved a t infinite time and Vt = volume of constant for removal of the first and second gas evolved a t time t . In most cases, the rates of reaction hydrogens. w-ere too slow to determine Vm-values experimentally. Vm-values were calculated from the weight of the silane The rate constants were correlated using the and were corrected for the vapor pressure of the solvent. Taft equation2 U-ith the two most reactive diorganosilicon hydrides, the log kl = ( 2 u * ) p * c
+
(15) F. P. Price, J. A m . Chem. Soc., 69, 2600 (1947). (16) L. H. Sommer, M. P. Barie, Jr., and D. R. Weyenberg, i b i d , 81,251 (1959). (17) R. West, i b i d . , 76, 6012 (1954). (18) 0. W. Steward, Ph.D. Thesis, T h e Pennsylvania State Univezsity, 1957, pp. 132-134; Dissertation Abstr., 17,2827 (1957). (19) Saentific Glaso Apparatus Co.,Bloomfield, S. J. (19a) The 95 vol. - % ethanol (sp. gr. g z = 0.806) was prepared by mixing 50 ml. of water with 950 ml. of absolute ethanol.
Three linear correlations were obtained (Fig. 1): the triorganosilicon hydrides (13 points), the diorganosilicon hydrides (rate constants for the first hydrogen, 4 points), and the monoorganosilicon hydrides (ra,te constants for the first hydrogen, 2 points). The correlation data are given in Table 111.
OMARVir. STEWARD AND OGDENR. PIERCE
4934
RATEDATAFOR No
1
--
THE
R
Vol. 83
TABLE I ACID-CATALYZED €IYDROLYSIS OF ORGANOSILICON HYDRIDES AT 34.8' (1.43 N IX HYDROGEN CHLORIDE) RR'R"SiHa No. of ki X 103, R'
CHsCHz
-
R"
CHpCHz
runs
1.16 zt 0.04 1.19d 1.99 i 0.08 1.45 f 0 . 0 2 0.74Y
4 3
2 3 4
.O i d
5 6
.7od
"
05 V O L . -ETHANOL ~~ Relative rate
mrn.-'b
3
CHiCHz
IN
ze*c
1.0
-0.30
1.7 1.2d 0.63 .06 .59
-0.115 .23 - .345
-
- .57 - .39 - ,375
.20d .17 ,32 3.14 f 0.07 2.7 4.4 9 5.01 i: .26 .01 10 7.8 .R6 9 04 f .39 11 2.6 .la 2 98 f .04 12 3 . 1 1 f .16 2.7 .2 I 13 2.8 .36 3.29 f .12 11 .466 4.96 f .10 4.3 3 ,l7 2.8 15 3 3.26 f .09 1 05 1.8 16 3 2.08 A .09 1.8 0.60 17 2 . 13d 0.3 1.80 18 0 . 40d 22 0.20 26. 2d 19 (1.43)d" (1.2) 47 0.68 20 54.5 zt 0 . 6 CFjCH2CHz CHaCH2CHzCHz II ( 3 . 6 =k 0 . 2 ) f (3.1) 1.13 111 f2 96 21 CF3CHzCIlz CFaCHzCHz H (11.2 f 0 . 3 ) f (9.7) 0.95" 106 zt 5 91 22 KECCHpCHz CHo H ( 9 . 5 f 0.4)f (8.2) 143d 120 0.83 Cyclo-CcH11 H H 23 1.44" 340 394 f 128 SECCHzCHz H H 24 (5.5 f0.3p (4.7) Rate constants reported Ref. 2, p. 619. Concentration of silane, 0.0250-0.0915 mole 1.-1. b Standard deviation. by J. E. Baines and C. Eaborn using aqueous ethanolic hydrochloric acid (1.43 N , 95 vel.-% ethanol) a t 34.9". ref. 8. e The o*-values for the w-cyanoalkyl groups were calculated by dividing the o*-value for the cyanomethyl group by the factor Rate of hyof 2.8 for each intervening methylene group, ref. 2, p. 592. f Rate of hydrolysis of the second Si-H bond. Initial rate of hydrolysis of the third Si-H bond. drolysis of the first and second Si-H bonds. I
8
3 3 3 3 3 3
For the Hammett equation, Jaffd20 considers correlations which have a correlation coefficient of >0.95 as satisfactory. Both of the correlation coefficients in Table 111 fall in this range.
-060
-020
020
060
IO0
I40
Discussion Polar Effects.-The effect of polar substituents on the rate of the acid-catalyzed hydrolysis of organosilicon hydrides is quite small ( p * = 0.77), the reaction being facilitated slightly by electron withdrawal from the silicon atom. Baines and Eaborn* noted the same effect qualitatively from the acid-catalyzed hydrolysis of p-substituted phenyldimethylsilanes. This effect is in contrast to the very large increase in rate by electron withdrawal from silicon in the alkali-catalyzed hydrolysis ( p * = 4.27).la The reaction mechanism proposed by Eaborn'' involving electrophilic attack by a hydronium i o n and nucleophilic attack by a water molecule in the rate-determining step seems reasonable based on the above polar effects since the silicon atom must be more negative in the transition state than the ground state.
I80
R3SiH
I3 99 a The organosilicon hydrides not reported in this table by vapor phase chromatography. d Ref. 16. Ref. 14.
Se
89"
p*
(p*)
ab
RsSiH 0.77 0.06 -2.72 R2SiH2 .77 0.09 -1.77 RSiHa .72 -1.66 Standard error. Intercept.
..
(I
*
8e
(a)
0.08 0.03
,.
(log ki)
0.09 0.06 * .
Hydrogen, % -Fluorine. %Calcd. Found Calcd. Found
4.1 34.0 34.0 0.315 0.32 33.5 8.35 33.2 0.93 0.91 8.2 45.7 1.09 1 . 0 8 46.6 0.89 0.89 32.1 32.1 2.03 1.99 48.4 48.2 9.1 3.55 3.55 have been reported previously, ref. 1.
TABLE I11 CORRELATION DATA Correlation
-Carbon, %Calcd. Pound
;e
0.961 0.972
...
Correlation coefficient.
Steric Effects.-The points representing the mono-, di- and triorganosilicon hydrides fall on separate lines with the same slope (within one standard error), and the distance between these lines is smaller as the number of substituent groups on the silicon atom is decreased. All of the organosilicon hydrides included in the correlations have substituent groups which are not branched, except cyclohexylsilane. The above observations suggest that steric effects are important and the linear relationships represent organosilicon hydrides in which the steric effects of the groups around the silicon atom are very similar. The successful correlation of the second-order rate constants for the alkali-catalyzed hydrolysis of triorganosilicon hydrides with the Taft equation was ascribed t o constant steric effects.Ia I n the correlation of the data for the triorganosilicon hydrides, the points representing the compounds with two methyl groups tend t o lie above the regression line, and the points representing the compounds without methyl groups, below. The point representing the diorganosilicon hydride with one methyl group also lies above the regression line. These observations indicate that the methyl group has slightly smaller steric requirements than the larger groups. These differences in steric requirements are probably partly responsible for the observed scattering of the data. The low reactivity of triisopropylsilane and triisobutylsilane toward the acid-catalyzed hydrolysis of the Si-H bond is undoubtedly the result of increased steric effects due t o branching in the attached groups.* As expected, the point for triisopropylsilane shows the largest deviation from the regression line. Cyclohexylsilane does not appear t o show any greater steric effects than P-silylpropionitrile for removal of the first hydrogen. This would indicate that the rate of reaction of the first hydrogen in the monoorganosilicon hydrides is not very sensitive to branching in the organic group. The points representing the compounds phenyldimethylsilane, triphenylsilane and chloromethyldimethylsilane all fall below the regression line
4.3 8.4 5.3
53.4
53.4
30.9 50.9
31.2 50.8
8.7 Table I.
Determined
for the triorganosilicon hydrides. As expected, if steric effects were involved, the point for triphenylsilane shows the largest deviation. However, unusual steric effects would not be expected for chloromethyldimethylsilane. The deviation for chloromethyldimethylsilane is probably due to dative n-bonding which may also be responsible for the deviations of the phenyl-substituted silicon hydrides. Dative n-Bonding.-There is substantial evidence for dative ?r-bonding between a silicon atom and an attached atom or group of atoms in which filled p-orbitals are available for bonding.22 Dipole momentz3and nuclear magnetic resonance5 studies have indicated that contributions from structures such as I are important in the ground state of phenyl-substituted silicon compounds. If the
.ee -
Si=
I
electron release from a phenyl group t o silicon by dative a-bonding is important as evidence indicates, it would be in opposition t o the electronwithdrawing effect, ;.e., --I effect. The net result would be a decrease in the inductive effect of a phenyl group attached to silicon relative to carbon. Since the a*-values were determined for substituent groups on carbon, deviations from the Taft equation would be expected for phenylsilicon compounds. From the above discussion, i t would seem that the observed deviations of the phenyl-substituted silicon hydrides from the regression line for the triorganosilicon hydrides are probably partly a result of dative n-bonding, steric effects also being partly responsible. Recently, Brook, et aLjZ4have proposed a direct interaction between the filled p-orbitals of an oxygen atom and the vacant d-orbitals of a silicon atom when the oxygen is separated from the silicon by one carbon atom, e.g., a-silyl ketones and a-silylcarbinols. Frye, et a1.,26have reported results which indicate that a transannular interaction between the filled p-orbitals of a nitrogen atom and the vacant d-orbitals of a silicon atom (22) For general reviews on t h e subject see ref. 21, pp. 94-103, and F. G. A. Stone and D. Seyferth, J. Inorg. Nucl. Chem., 1, 112 (1955). (23) H.Freiser, M. V. Eagle and J. Speier, J. A m . Chem. Soc., 75, 2821 (1953). (24) A. G. Brook, M. A. Quigley, G. J. D. Peddle, N. V.1Schwartz a n d C. M. Warner. ibid., 82, 5102 (1960). (25) C. L. Frye, G. E. Vogel a n d J. A. Hall, ibid., 83, 996 (1961).
ROBERT EISENTHXL
VOl. 83
occurs when the spatial arrangement of the atoms are conducive to bonding. The point representing chloromethyldimethylsilane falls a substantial distance below the regression line for the triorganosilicon hydrides. In this case, it seems unlikely that steric effects would be responsible. Direct interaction between the filled p-orbitals of the chlorine atom and the vacant d-orbitals of the silicon atom, analogous t o the interaction proposed by Brook, et al.,??seems likely for chloromethylsilicon compounds. Resonance structures of the type I1 cannot be ruled out. Effects of this type would tend to reduce the in-
BaneyZ8 has correlated the hydrogen-bonding acidity of various triorganosilanols with the Taft a*-values and observed that chloromethyldimethylsilanol and the phenyl-substituted silanols were less acidic than predicted by the u*-values for the chloromethyl and phenyl groups. In this study, steric effects of the organic groups bonded to silicon should be of minor importance. Other atoms in addition t o oxygen and chlorine which have filled p-orbitals available for bonding may also interact directly with the vacant dorbitals of silicon when they are separated from the silicon by one carbon atom. The base strengths of a-trimethylsilylalkylamines do not decrease e a regularly as the number of methylene groups C1 C&=Si= between the silicon and nitrogen atoms is inI1 creased from one to three; trimethylsilylmethylductive effect of the chlorornethyl group on a amine is a slightly weaker base than p-trimethylsilicon atom relative to a chloromethyl group on a silylethylamine.29 These observations are inconcarbon atom and cause serious deviations from the sistent with inductive effects. The small K B value for trimethylsilylmethylamine could be Taft equation. The above postulate for a chloromethpl group explained by partial bonding between the nitrogen on silicon is supported by dipole moment studies. and silicon atoms. The proposed intramolecular rearrangement of Freiser, et U Z . , ~ ~reported 0.23 D. for the (CH3)XSiCaliph bond moment in chloroniethyldimethyl- a fluorine atom from carbon t o silicon in the thermal silane. Coleman and FreiserZ7reported 0.26 D. decomposition of a-fluoroalkylsilicon compounds is for the (CIIB) aSiCaT0, bond moment in p-chloro- in agreement with the above postulate.30 Acknowledgments.-We wish to thank the Dow phenyltrimethylsilane. Considerable dative T bonding between the benzene ring and the silicon Computation Research Laboratory for the least atom in p-chlorophenylsilicon compounds has square analyses, G. Wayne Holbrook for the and been proposed to explain dipole moment data.27 preparation of bis- (3,3,3-trifluoropropyl)-silane, Since approximately the same bond moment for Ronald H. Baney and Cecil L. Frye for helpful disthe (CH3)3SiC group is observed in both com- cussions. pounds, it seems reasonable t o assume that elec(28) R. H. Baney, private communication. (29) L. H. Sommer and J. Rockett, J . Am. Chem. Soc., 73, 5130 tron release from the chloromethyl group to the (1931). T h e following values were reported for K g X 10' a t 25': silicon atom is also occurring. (20) H. Freiser, R. Charles, J. Speier and X I . Eagle, J . A m Cheitz. Soc., 73, 5 2 2 5 (1551). ( 2 7 ) A. Coleman and H. Freiser, { h i d . , 83, 1127 ( 1 9 6 1 ) .
[COSTRIRUTION FROM THE T'EX-4BLE CHEMICAL
trimethylsilylmethylamine, 9.1; 6-trimethylsilylethylamine, 9.7; ytrimethylsilylpropylamine, 5.6. (30) R. S. Haszeldine and J. C. Young, Pmc. C h e w Soc., 344 (1939).
LABORATORIES, UNIVERSITY
O F NORTH CAROLISA, CHAPEL
HILL,N.
c.]
Interannular Electronic Effects in Ferrocene : Kinetics of Reaction of Substituted Ferrocenoic Acids with Diphenyldiazomethane BY WILLIAMF. LITTLEAND ROBERTEISENTHAL' RECEIVED APRIL 22, 1961 The kinetics of the reaction of heteroannular substituted ferrocenoic acids with diphenyldiazomethane in toluene were ~ t u d i e dunder second-order conditions a t 30'. For the substituents H, CHBCO-, C~HS-,CSHSCO-and C6H5CH2-, the logarithms of the second-order rate constants were correlated with the corresponding acid constants and with Hammett pare acid was much too fast for correlation with sigma constants The rate of esterification of 1-a-hydroxybenzyl-1'-ferrocenoic its acid constant. This can be ascribed to iriteranriular hvdrogen bonding between the or-hydroxybenzyl group and the carboxyl group in the acid.
Interannular transmission of electronic effects through the ferrocene nucleus has been demonstrated in the acid constants of heteroannular ferrocenoic acids2--" and in the shifts of the infrared (1) This work was taken from t h e Ph.1) dissertation of Robert Eisenthal, submitted t o t h e University of S o r t h Carolina, January-, 1961. (2) 4.N. Sesmeyanov and 0. A . Reutov, Doklndy A k i l d . S a u k , S . S . S . R . , 115, 518 (1957). (3) A. N. Sesmeyanov and 0. A. Reutov, Izvesl. .4knd. .YaiLk, S.S.S.R., 926 (1959); C. A , , 54, 191 (1953).
stretching frequencies of substituted ferrocenyl acids and esters.& The acid constants and spectral shifts vary in the direction expected from the electronic effects of the substituent groups. The reaction of carboxylic acids with diaryldiazomethanes has been thoroughly studied kineti(4) W. F. Little a n d R. Eisenthal, J. Am. C h e m . Soc., 82, 1577 (1960); and J . Ovg. Chem., 26, 3609 (1961). ( 5 ) L. A Kazitsyna. B. V. Lokshin and A. X, Nesmeyanov, Doklady A k a d S n i r k , S.S.S.R., 127, 333 (1939).
