M. 11. SPRUNG AND F.0. GUENTHER solid (0.45 g.) upon standing in a freeze chest. After crystallization from acetone, there was thus obtained 0.08 g. more of purified solid, m.p. 209-210”. Several small lots of this solid, obtaiiied as described above, were combined, resublimed, and again crystallized from acetone, m.p. 210-211’ iuiicor.) in a sealed tube. Tlic solubility and superficial appearance w ~ t the r ~ same :is were described above. .-lncil. Calcd. for CiiHlsSihO:,:C , 17.9; 1 1 , -4.5; mol. Ivt., 402.M. Found: C, 18.1; H , 4 . 5 ; mol. ~ v t . 374. , Iiifrared spectrograms of “hexametiiyl Tii” :itid “octamethyl T8” also mere obtained, both a s Sujol mulls. Both curves confirmed the complete absence of silanol, and shorved no features inconsistent with the assumed structures. Evidence was sought for the presence of six-membered (cyclotrisiloxane) rings in “liexamethyl but an unequivocal aiisIver to this question \vas not obtained. Analyses.---For cieterrnination of alkosJ-I, samples of suitable weight to give 0.03 to 0.09 g. OK were refluxed iti aqueous 50% potassium hydroxide. Solution usually was complete in 0.5 hour. T‘rre alcohol \\-;I; distilled directly ~~
118) C . \V. Young,
P.C .
Ssriais, C . T H I S J U T ‘ R B A L , 70, 3778 i l g 4 8 ) .
C C,ii-rie aud \ I . J . ltiinter,
from the alkaline solution and wascollectedas awater-alcohol mixture. The alcohol was determined by measuring the percentage absorption a t 4 i 5 mp in 5-mm. cells of the c01ored complex formed with ammonium hexanitrato cerate.IY Silanol usually was determined by reaction with methylniagiiesium iodide in dibutyl ether solution. In some cases, lithium aluminum hydride was used instead of the Grignartl reagent. I t is believed that the Grignard method, as employed for this work, is accurate to within 5 or, a t most, 10%. Silanol values usually were checked qualitatively by iiifrarecl spectrogrami. The molecukir weight determinations were carried otlt cryoscopicalll- in benzene.
Acknowledgment.-U?e are indebted to E. 11. Hadsell for precision distillations, to various members of the Analytical Chemistry Unit of this Laboratory for elementary analyses and molecular weight determinations, and to C. -1.Hirt for t.he infrared absorption curves. ( I O ) 1’. i < , Ijukr, Iud. i s i i q . C/~!:I!.,A d . Ii,l,,17, 572 (194.5), SCHti:NEC1..41)Y,
[ C O S T R I B U T I O N FROM THE
VOl. 77
s. T.
GESERALELECTRIC RESE.4RCH
LABORATORY]
The Partial Hydrolysis of Ethyltriethoxysilane BY 11.Izr. SPRUNG AND F.0. GCEXTHER RECEIVEDJASUA4RY 5 , 1955 Appreciable yields of distillablc liquids result from the acid-catalyzed rcaction of three molar equivaleiits of water with ethyltriethoxysilane in benzene solution. The major components are polycyclic polysiloxanes that contain free silanol groups and residual ethoxyl groups. Loxver yields are obtained in ethanol solution with an acid catalyst, and only gels are obtained with an alkaline catalyst. If smaller amounts of water are used, the major products are first less highly cyclic, and finally predominantly straight-chain polysiloxanes. Several ethoxy-substituted linear and monocyclic polysiloxanes were isolated. ilane and ethyltrichlorosilane gives products siniilar in physical Co-hydrolysis of equimolar mixture e, but with much less residual ethoxyl and silanol. A sublimable properties to those derived from the fied :is octa-(ethylsilsesquioxaiiej,( CtHeSiOl.jh. -4 lower lllelting high-melting solid obtained in small solid frequently present in even smaller ainounti: \ r a s identified a s hesa-iethylsilsesq~~iosane), ( C2H:,SiOl.:,)6.
Introduction The distillable liquids isolated in low yields when methyltriethoxysilane was treated with three molar equivalents of water in benzene solution are principally complex polycycli~polysiloxanes with ethoxyl and hydroxyl end-groups.‘ It was of considerable interest that complex silanols survived the rigorous catalytic and thermal treatment involved, including distillation a t high still-pot temperatures. Ethyltriethoxysilane is considerably more stable toward hydrolysis. Under the standard conditions previously described (;.e., 1.5moles of water allowed to react with 0.5 mole of the alkoxy silane, dissolved in 3.00 ml. of benzene and stirred rapidly a t the boiling point‘), about 2.5 to 3 hours is required to reach the equilibrium conditions that are established. h n acid or alkaline catalyst must be employed. Two high-boiling fractions were obtained: (1) b.p. 149-13S” a t 1 mm., f i 2 “ ~1.4263; (2) b.p. 167-176’ a t 1 mm., n Z o 1.4308. ~ The residue was a heavy, balsam-like liquid, in contrast to the gelled resin obtained from methyltriethoxysilane. The results of ultimate analyses, ethoxyl, hydroxyl and molecular weight determinations are shown in Table I. Infrared spectra of the distillable fractions confirmed the presence (1) \I t IUii)
AI
S p r u n g and P 0 (vrienther. THIS Tor
RUAI,
77, ?VIn
of Si-OH, Si-OC2116and Si-C2H5 bonds (peaks a t 2.93, 7.73 and 7.99 p ) . The two infrared curves were nearly indistinguishable. TABLE 1 FIYDRULYSIS
PRODUCTS Ol3l”AINED
IVIlli
EQUIVALESTS OF WATER 1101 w t . (in benzene) Carbon, % Hydrogen, % Silicon, % --OCzHa, %
--OH,70
Fraction 1
Fraction
540 33.9 7.4 29.2
,5i4 33.1
15.6 1.0
7.3
1167, 1 3 3 0 31 . A 0 9
29.9 1:3.5
31.2 92
4 . 3 (a\..
01
,
6)
..
The close siniilarity of the analytical constaIzts, and the trends from lower to higher molecular weight components, suggest that the individual components are very closely akin. The data for the lower boiling fraction conform closely to those required for a compound of the empirical composi: ;.e., (CzH5)&06(OC2H5)2(0H) tion C,4H36Si50~ (calcd.: mol. wt., 4SS.7; C, 34.4; H, 7.4; Si, 28.7; -OCzHb, 18.4; -OH, 3.V). The analytical data for the higher boiling fraction do not lead ( 2 ) Reasonable sttuctural formulas would h e j,7-diethosq.-(X,9’~pou~-l-hydrox~-l,3,;7,7.9-pentaeth~Icyciopentasiloxane nr 1 .7 rii ~ t h o s v - ~ R . F I ~ - ~ ~ n x y - 6 - h y n r o x y - l , 3p,r5n.t a7e.t0h ~ l r y c 1 n p r n t x i ; l o u l n r
Aug. 5, 1955
PARTIAL HYDROLYSIS OF ETHYLTRIETHOXYSILANE
3997
to a single empirical formula. However, a mixture particular case, an increase in the relative proporof I (or Ia) and 11, differing only by a molecule of tions of 11compared to I might be expected. water, would yield these analytical data. (Calcd. The Effect of Ethanol on the Initial Reaction Products.for C16H&i6ol1 (I or I a ) : mol. wt., 578.9; c, Higher yields of cyclic ethorylated products might be real33.2; H, 7.3; Si, 29.1; -OC2H5, 15.6; -OH, 5.9. ized if the sensitive silanol groups were converted to alkoxy1 to distillation. Accordingly, a half-mole of ethylCalcd. for C16H40Si6010 (11): mol. wt., 560.8; c, previous triethoxysilane was hydrolyzed and, after the solvents had 34.2; H, 7.2; Si, 30.0; -OCnH5, 16.0; -OH, 0.0.) been removed, an equal volume of absolute ethanol conCzH,
C,H,
CzH,
HOSt-O-Si-O-SIOC2H, I 0 0
I
I
C?H,OSi-O--$i-O-SiOH I
I
C.H,
I
CZHO CzH5
I CIH, OH
HO
o/
t;i b '
Ia C>H, C?H, I 0-Si-0-Si-0
C,H\
/ \
>Si
C?Hs
0
I
0
'0
si^'""
taining the usual catalytic concentration of hydrochloric acid was added. This mixture was heated again as before. Vacuum distillation gave a materially increased yield of liquid products (28.1 g., b.p. 132-223" a t 0.3 mm., n 2 0 ~ 1.4233, 26.5% -OC2Hj). The balsam-like residue (21.1 g.) had 19.7% -OC2H6. On redistillation of the volatile portion in a spinning-band column, the fractions shown in Table I1 were taken. Azeotropic separation of water after hydrolysis then was omitted. Solvents were removed immediately under vacuum. Absolute ethanol and hydrochloric acid were added again, the mixture heated under reflux as before, and distilled. The yield of crude distillate was 20.0 g . (b.p 138-196" a t 1 mm., T I ~ O D 1.4208, -0C2H5 28.7%) and of residue, 29.0 g. (20.6% -0CzHj). The volatile liquid was redistilled again through a spinning-band column (Table 111).
T o determine if siloxane bonds were split by reaction with ethanol, the non-volatile residue from a standard half-mole hydrolysis (average -OC2Ha content, 12.0%) was heated under reflux for 21 hours with absolutr ethanol and hydrochloric arid. Distillation gave a small fraction (2.2 g., 20.5% -OCzHj), but the boiling range was above that to which the original distillation had been carried. Since lower boiling fractions were not obtained, siloxane-bond cleavage is apparently not important.
Characterization of Hydrolysis Products through Physical Properties and End-group Analysis.I B multiplicity of complex polysiloxanes, polyC?Ha C2H, siloxanols, alkoxy polysiloxanes and alkoxy polyI1 siloxanols are encountered in the hydrolysis prodThese structures are analogous to (in fact, homol- ucts of trifunctional organosilanes. A systematic ogous with) those assigned the major distillable method of classification based on structure is hydrolysis products obtained from methyltri- needed. However, for the present purposes, a ethoxysilane under similar conditions. It should scheme that considers only the relative extensivebe emphasized again that closely related substances ness of the hydrolytic attack is useful. With having more or less than this number of -OH and increasing hydrolysis, the major products are first -OR groups are probably present in small amounts, linear, then monocyclic, then polycyclic and and that less probable isomeric species also must be finally "cage-like." The concomitant changes in taken into account. These data again demonstrate composition can be expressed in terms of the -OR'how readily structures containing a plurality of to -R2Si2O3 ratio. Polymer growth by polyconrelatively small rings are formed from trifunctional densation usually occurs randomly. However, in polysiloxane systems, the relative ease of ring organosilanes. The non-volatile residue can be represented as a closure and the high probability of formation of mixture of polymer-homologs of the distillable eight-membered rings lend a certain statistical components. The average composition corre- preference to some types of ring structures. These sponds to (C,HE,)IZS~~~O~,(OC,H,), (related to 11) considerations furnish a key to the nature of the (calcd. for CBH7oSi12019: mol. wt., 1047.5; C, 32.1; substances present. For distillation cuts of narrow boiling limits, the H, 6.8; Si, 32.2; -0CzH5, 8.6). A two-mole hydrolysis gave only a slightly greater nature of the predominant component can be yield of distillable products, owing probably to determined from the ethoxyl content and boiling longer heating times during concentration and dis- range only (knowing other physical constants, such tillation. This product was redistilled in a spin- as refractive index and density, or having an infraning-band column, and cuts were taken a t the mid- red spectrogram of the sample, is naturally helpful). points of the fractions described above (fraction 2, The ethoxyl content, in turn, need not always be b.p. 155-156' a t 4 mm.; fraction 6, b.p. 166-167' determined directly (although this was generally a t 1 mm.). Except for low silanol contents, they done as a matter of course) but can be estimated were analytically similar to original fractions 1 from the specific gravity. The correlation beand 2. (Anal. Found (fract. 2): Si, 29.0; -OH, tween these two quantities in a series of closely 2.1, 2.2. (Fract. 6): mol. wt., 540; C, 33.6; related cyclic alkoxy siloxanes is shown in Fig. 1. H, 7.0; Si, 30.8; -0CzH5, 13.5; -OH, 0.7.) The The predominant components of distillation decrease in hydroxyl content was confirmed quali- cuts given in Tables I1 and I11 were deduced tatively by infrared analysis. The complex silanols in this way. Treatment with ethanol in the obviously are destroyed by long heating. In this presence of acid did, indeed, convert sotrie of thc C?H,O'
0-s1-0-s1-0
/
'OC2HJ
11.M. SPRUNG AND F. 0. GUENTHER
390s
EDIS IS TILL AT ION Y,>i , ml.
Fraction
OF A
Vol. 77
TABLE I1 REACTION PRODUCT P R E V I O U S L Y TREATED IVITII EXCESS ETHANOL
n.p., OC. a t 1 mm.
Probable predominant conipunrnt b Theor. c$ -0WEtkhOa -0CsHr
-oc?II,,o
n-t. ,,f 1 ml.
ll?OD
'7a
--OISt/BtrSizOa
105-121 1.4137 1 ,023 41.1 4:2d ... 38.1 1.3 ... ... ... 0.9 124-120 1.4158 ... 120-127 1.4193 1 , OB6 31.8 3 : 2" 1: 2* no.4 1. A 1.5 127-131 1 .4220 1.095 2.5. G 3:2.5 ... 26.1 131-136 1 ,4232 ... ... , . . 1.1 1.3 136-141 1,4223 23.1 3:3 1:3 22.3 " I 1.6 144-16 1 1.4221 1.070 31.0 1:3 ... 28.4 8 1.6 151-155 1.42lG 1.OS0 28.8 1:s ... 28.4 ... , . . 9 1.0 15-16SC 1.421; ... ... ... in 1.5 168-158 1.4237 1,102 26.1 4:4 ... 22.6 11 1.1 ... ... 158- 1G 7 1 ,5252 ... Adduced from the residual ethoxyl a;id the physical constaiits. Estimated from the specific gravity. The pressure became erratic a t this point in the distillation. Tetraethyitetraethoxycyclotetrasiloxane. EIydro~ytetraethyltriethoxycyclotetrasiloxane. TABLE III REDISTILLATIOS O F A REACTION I'RODUCT PREVIOVSLY TRE.4TED IVITH ETHANOL'
1 2 3 4 5 G
I
n.p.,o c .
mi.
Fraction
at
Wt. < I f 1 nil.
1 mm.
Prolinhle predonlinant componentC T h e m '& -OEt/€:tzSizOJ -oc21ra
-OC'IH~,b m
silanol-rich components to alkoxy compounds, but did not reduce the complexity of the reaction product. The distillable products in both cases were predominantly polycyclic, having OEt/ EtZSi203ratios between 4 : 2 and 4 : 4.
r
113 1.10
3
3
1.05
5
.-.a
I
M
5
1.00
0 95 0.90
1-1
8
22
10
_ 34 ::-OCyHi.
8
_ 8
46
_
2s
~ 70
TS. L%-C2H50relatioilship of distilled hydrolysis products of ethyltriethoxysilane.
Fig. 1.-Density
Infrared absorption spectra were obtained for the products described in Table I1 and these data added considerable support t o the assumed structures. For example, in the crude distillate obtained after treatment of the hydrolyzate with ('xcess alcohol, the peak heights a t 7.7.5 aticl 2.9:; 11 indicated a relatively large atriotint of ethoxy a t i d
...
...
0.9 94-122 1,1108 . . , ... 1 .s 120-121 1.41s 1.041 37.2 1.5 121-130 1.4192 1 .OX 30.1 1.2 132-133 1.1220 1.102 24.2 1 1.3 131-146 1,4204 I ,096 25.4 G 1.6 146-151 1.4194 1. ( 6 1 32.2 7 1.0 143-155 1,4200 , . . ... 8 0.6 153-155 1,4216 ... ... ... 9 1.2 155-1 62 1.4237 ..* " Dean-Stark trap separation of water omitted. 0 Estimated from the specific gravity. ethoxyl and physical constants. Tetraethyltetraethoxycyclotetrasiloxane. 1 2 3 4
-E
.
. I .
Vd.,
1.20
.
4 : 3"
38. 1
...
...
4:2.5 4:3.5 4:s
26.1 26.1 28.4
I
.
.
.
... ...
.
I
...
... c
Adduced froin the residual
a silanol content of around 1 to 276. I n fraction 1 of the redistilled product the silanol peak was virtually absent, but reappeared in somewhat reduced intensity in fraction 3. In fraction 7 the silanol peak again was virtually absent. The SiOC2Hj peak a t 7.73 p was, however, nearly as prominent as in fractions 1 or 2. These observations accord well with the data in Table 11. Mechanism of Siloxane Bond Formation.-.\ sample of crude distillate from a 1-mole hydrolysis run under normal conditions (19.3 g., b.p. 12S163" a t 0.S Inni., T Z ~ " D1.4286, 15.1% -0CzH5, 2.3% OH) was heated under a water-cooled reflux condenser and under a pressure of 20 t o 35 mm. for 105 minutes a t 190-23O0. A volatile product weighing 0.613g., caught in a Dry Ice trap, was determined, by physical properties, to be mostly ethyl alcohol. The oil then was distilled, giving: (1) 6.2 g., h.p. 12S-190° a t 0.5 inrn.,n% 1.4270, 1-2.87, -OC2H6; (2) 12.2 g. of nun-volatile residue having 11.57, -OC2H5; and ( 3 ) 0.07 g. inore of the The average alkoxy content ~ ~ o l a t i lproduct. e dropped from 13.17, (initial) to 12.6% (final), and the average silanol content from 2.3 to 0.57G. The stoichiometry of the reaction insofar as it affects functional groups can be represented as
I
1.4--siOEt 1
I
+ 2.0 -4iOEf
--f
I
1 .I I ' t O l l
1 r i ::
Il?O j
I i
I
/
SiOSi
I 1
Aug. 5 , 1959
PARTIAL HYDROLYSIS OF ETHYLTRIETHOXYSILANE
3999
TABLE IV ACID-CATALYZED HYDROLYSIS IN BENZENE WITH VARYING AMOUNTS OF WATER Fraction
Wt , ,
B.p., O C . at 1 mm.
-OCzHs,
R.
(A) 1 2 3 4 6 Residue
...
(B) 1 3 3 4 5 Residue
EtSi(OEt)s:H20 = 1 : 1 . 5 1.4165 36.9, 34.1 1.4222 27.6 1.4252 31.9 1.4275 ... 1.4292 17.9 ... 15.1
4.2 4.7 7.2 6.4 7.3 18.5
103-115 122- 134 134-1 44 117-168 173-193
6.7 5.2 4.7 5.1 2.0 28.0
44-69' 74-94 130-150 159-181 188-196
...
%
?izOD
EtSi(OEt)p:H20 = 1:1.125 1 ,3923 ... 1.4024 60.1 1.4152 38.9, 42.9 1.4217 26.7 ... 1.4237 ... 20.2
Probable predominant component
Monocyclica Polycyclic' Polycyclicc
... Polycyclicd Polycyclic siloxanols Monomer Linear dimer Cyclic tetramer Polycyclic"
... Polycyclic
Theor. %
-OCZH~
38.1 26.1 22.3
... 17.6
...
... 58.1 38.1 28.4
... ...
(C) EtSi(OEt)8:H20 = 1:0.75 70.5 Monomer ... 69.8 30.3 1 50-66' 58.1 Linear dimer 2 67-79 4.9 1,4022 56.1 38.1 Cyclic tetramer 2.5 1.4138 40.1 3 120-1 45 ... ... ... 4 148-18-1 2.4 1.4212 ... Polycyclic 21.5 ... 20.3 Residue ... Average of 3 - 0 E t and Average of 4 - 0 E t groups for 2 EtzSi?Oaunits. * Average of 3 - 0 E t groups for 2.5 Et&inOs. 1 -OH for 3 Et2Si203. Average of 3-OEt and 1 -OH for 4 Et2Si20J, e Average of 4 - 0 E t for 3 Et?Si*Oa. f A t 20 mm.
On this basis, the 0.53 g. of volatile product consisted of 0.675 g. of ethanol and 0.055 g. of water. The reactions giving alcohol and mater can be formulated as
(B)
2(Et0)2(Et&O&
S(
