SPECTRA A N D STRUCTURE OF
April 5, 1956
ever, the curve produced did not break sharply, although extensions of the straight portions of the curve showed a break a t the molar ratio of 1: 1 for the niobium and pyrogallol. The absence of a sharp break here was further evidence t h a t the niobium pyrogallol complex was partially dissociated. The above observations, which seem to show t h a t one pyrogallol molecule complexes with one niobium, do not give the nature of the complex formed. I t is uncertain whether the pyrogallol simply adds itself to the niobium oxytrioxalato ion already
[CONTRIBUTION FROM
THE
DISILOXANE AND
DISILOX.\NE-&
1327
present or whether it displaces one oxalate ion. If the addition compound is formed, the coiirdination number of niobium would be eight, which is possible according to Sidgwick." Work in progress in this Laboratory, it is hoped, will give more information regarding the relative dissociation constants of the pyrogallol and oxalato complexes of niobium. (11) N. V. Sidgwick, "Chemical Elements and Their Compounds," Vol. I, Oxford University Press, London, 1950, pp. 840-841. AUBURN,
ALABAMA
DEPARTMENT OF CHEMISTRY A N D SPECTROSCOPY LABORATORY,MASSACHUSETTS INST~TUTE OF TECHNOLOGY]
Vibrational Spectra and Structure of Disiloxane and Disiloxane-d,' BY R.c. LORD, D. w.ROBIN SON^ AND w. c. SCHUhfB RECEIVED OCTOBER11, 1965 Disiloxane and disiloxane-d, have been prepared by the hydrolysis of chloro-ikdne and chlorosilnne-ds. Infrared and Raman spectra have been obtained of the gaseous and liquid samples, respectively, down to about 250 cm.-'. I n the infrared, the fundamentals have been located at 2169, 1107, 957 cm.-' (parallel) and 2183, 957, 764 cm.-' (perpendicular) in H,SiOSiH3, and at 1575, 1094, 713 cm.-' (parallel) and 1575, 720, 593 cm.-' (perpendicular) in D3SiOSiDa. I n the Raman effect fundamentals are located at 2174, 1009, 606 cm.-' (sharp) and 2174, 947, 716 cm.-' (diffuse) in H3SiOSiH1, and a t 1575, (diffuse) in D3SiOSiDs. One perpendicular infrared-active fundamental, T71, (555) cm.-' (sharp) and 1546, 699, 532 believed to lie well below 250 cm. - l , was not observed. High-resolution spectra were obtained but insufricient rotational structure was found t o permit interpretation. Vibrational analysis indicates that the point-group syrnmetry of the molecule is D3d, that is, the S i U S i bond angle equals or approaches 180'. The results are equally compatible with a structure in axis. which the silyl groups undergo free internal rotation about a linear Si+Si
110.8 f 1 O . S The present investigation confirnis Introduction Although considerable attention has been di- the unexpectedly wide Si-0-Si bond angle in unrected toward spectroscopic and other structural substituted disiloxane. investigations of silicon compounds, i t is rather Experimental surprising t h a t disiloxane, the parent of silicones, Preparation of Disiloxane and Disiloxaned,.-Although has never been the object of such studiesS Spec- many attempts were made to reduce hexachlorodisiloxane tra of substituted disiloxanes have been reported4 t o disiloxane using lithium aluminum hydride and other and electron d i f f r a ~ t i o n , ~ dipole moment,6 and similar reducing agents, in no case could a reaction be found nuclear magnetic resonance' studies have appeared yielding the product sought. The Si-@Si linkage was alcleaved giving SiH, as the only volatile product.9 with some structural interpretation. No precise ways Disiloxane was prepared by the method of Stock1o in which measurements of the Si-0-Si bond angle have been monochlorosilane is hydrolyzed. Silicon tetrachloride11was made on these compounds, but the fact has been es- reduced t o silane, SiH,, with lithium aluminum hydride b y the method described by Finholt, Bond, IVilzbach and tablished t h a t the angles are much greater-probably between 130 and 160°-than those found Schlesinger.'* The silane was converted t o monochlorosilane by heating when first period elements replace silicon. Further- for about 30 hours at 100° with hydrogen chloride in the more, the wide angle is not general for the second presence of aluminum chloride. The monochlorosilane was period since, for example, the angle in ClzO is only separated from unreacted silane and hydrogen chloride by (1) Based on the Ph.D. thesis of Dean W. Robinson, submitted to the Graduate Department of Chemistry, Massachusetts Institute of Technology. May, 1955. ( 2 ) Du Pout Instructor in Chemistry, 1954-1955. (3) After the present work was completed, the authors learned of the investigation of EmelCus, MacDiarmid and Maddock ( J . Inorg. N u c l . Chcm.. 1, 194 (1955)). Their structural conclusion that disiloxane is an asymmetric top based on infrared studies only is somewhat at variance with those drawn in this paper. We believe that the additional evidence of the infrared spectrum of disiloxane-ds and of the Raman spectra of both compounds enables firmer conclusions to he drawn. (4) C. C. Certo. J. L. Lauer and H. C. Beachell, J . Chcm. P h y s . . 21, 1 (1954); H. Murato and M. Kumada, ibid., 21, 945 (1953); L. Savidon, Birll. SOL. chim. France, 411 (1953); I. Simon and H. 0. Mahon, J . Chcm. Phys., 20, 905 (1952); N. Wright and M. J . Hunter. TEIS JOURNAL, 69, 803 (1947). (5) K. Yamasaki, A. Kotera, AT. Yokoi and Y. Ueda, J . Chsm. P h y s . . 18. 1414 (1950). ( 6 ) R. S . Holland and C. P. Smyth, THISJOURNAL, 77. 268 (1955). (7) E. G. Rochow and H. G . LeClair, J . Inorg. S u c l . Chem., 1, 92 (1955).
fractionation of these more volatile compounds through a trap immersed in allyl chloride slush. It was separated from dichlorosilane by repeated fractionation through a trap immersed in carbon disulfide slush. The monochlorosilane was hydrolyzed with twice the theoretical amount of water at 30"; the resulting disiloxane was then washed with about fifty times as much water at 0". A total of 0.78 g. of disiloxane was prepared in this manner. Disiloxaned, was produced by similar reactions with some modifications. Silicon tetrachloride was reduced with lithium aluminum deuteride purchased from Metal Hydrides, Inc. The silaned, formed was then chlorinated with deuterium chloride prepared by hydrolysis of benzoyl chloride with deuterium oxide. However, because long heating ( 8 ) J . D. Dunitz and K . Hedberg, THISJOURNAL, 72. 3108 (1950). (9) W. C. Schumb and D. W. Robinson. a b i d . . 77, 5391 (1955). (10) A. Stock and C. Somieski, Bel.. 62, G95 (1919). (11) Obtained from the City Chemical Company, Brooklyn. New York. The lithium aluminum hydride was kindly donated by Dr. M D . Banus of Metal Hydrides, Inc.. Beverly, Alassachusetts. (1'2) A. E. Finholt. A. C. Bund, Jr.. K. E. Wilzbach and H. I . Schlesinger, THISJOURNAL. 69, 2G9? (1947).
132s
K. C. LORD,D. W. ROBINSOK AND W. C. SCHUMB
WAVE NUMUERS IN C Y
.
WAVE NUMBERS IN CM.
VOl. 78
'.
Fig. 1.-Infrared spectra of disiloxane (above curve) and disiloxane-ds (bottom curve) vapor. The isolated polystyrene band a t 1582-1603 cm.-' in both spectra was recorded as a check on the wave number calibration. Note change of scale a t 2000 cm.-l. was found to favor strongly the formation of dichlorosilane, presumably by disproportionation of monochlorosilane, the chlorination was done in batches, each batch being heated to 100' with aluminum chloride for 1-3 hours. This reduced the ratio of dichlorosilane-dz to monochlorosilane-da. One attempt to chlorinate s i l a n e 4 with hydrogen chloride showed considerable hydrogen-deuterium exchange on the silicon atom. The monochlorosilane-d, was hydrolyzed with deuterium oxide and washed in the same manner as described for disiloxane. Yields for the hydrolysis step were as high as 94% and the silicon tetrachloride reduction is practically quantitative. The poorest yields were in the chlorination step, where the proportion of dichlorosilane could not be reduced below about 20%. Disiloxane-ds amounting to 1.3 g. was prepared. The isotopic purity, as shown by the very weak Si-H infrared band a t 2180 cm.-', was about 98 atom-% deuterium. Disiloxane was found to be reasonably stable a t room temperature. Confined as a liquid of vapor pressure about 4.75 atm. and stored in a small Pyrex tube, it decomposed 3 % in 4 months. The volatile product contained silane. Determination of Spectra.-The infrared spectra were obtained with a Baird Associates' Double-Beam Recording Infrared Spectrophotometer with a rock-salt prism and a 10 cm. path length through the gaseous samples and with a Perkin-Elmer Model 12B Single-Beam Spectrometer modifiedl3 to permit removal of atmospheric carbon dioxide and water from the radiation beam. With the latter instrument gas cells of 5 cm. path length and prisms of CaF2, NaCl, KBr and TlBr-TI1 were used to cover the range 250-5000 cm.-'. The band centers were estimated from the singlebeam records, each band being measured with the prism of optimum dispersion for the particular frequency. As no unambiguous rotational structure was resolved, the estimates of centers, particularly for oddly shaped bands, are somewhat approximate. Probably the greatest error in the location of any band is 3 c m . 3 . All spectra were calibrated against known bands in water, carbon dioxide, methanol, ammonia and carbon monoxide, each used in the appropriate sDectral refion. (13) R.C. Lord, R.S. McDonald and F. A. Miller, J . Ogl. Soc. A m . , 42, 149 (1952).
High-resolution infrared spectra were obtained of the two perpendicular bands a t 2183 and 957 cm.-' in disiloxane. The instrument used for the first band has been described recently,14 and for the latter a similar instrument was employed which used a Golay radiometer as a detector rather than a refrigerated lead telluride cell. Spectral resolution of 0.3-0.4 cm.-l was achieved. For the determination of the Raman spectra of the light and heavy compounds in the liquid phase, samples were sealed in tubes constructed of Pyrex with a n inside diameter of 3.5 mm. and a wall thickness of 1.4 mm. Flat circular windows 3.0 mm. thick were sealed to the tubes with a tiny oxy-hydrogen flame and then annealed. These easily withstood the vapor pressure of the samples at 40°,the temperature during excitation. Extremely slow photodecomposition was noticed. The Raman spectra were obtained with a Zeiss threeprism spectrograph and Hg-4358 excitation. Filters of a saturated aqueous sodium nitrite solution and a 0.15 g./l. aqueous rhodamine solution were used. In the case of the more weakly scattering deuterium compound, a saturated solution of praseodymium chloride also was used. Plates were calibrated against the spectrum of an iron arc imposed on either side of the Raman spectra. Spectra were measured by enlargement of the plates and comparison with a millimeter scale. Strong lines are probably accurate to +2 cm.?, while weak diffuse lines only to f 4 ern.-'. Because of the small sample size and poor Raman scattering power of the compounds, all spectra were weak. The small tube diameter and the necessarily long exposure times caused halation about the exciting line so that no Raman lines could be observed below about 250 cm.-I. No depolarization measurements were made. The infrared spectra above 650 cm.-' obtained with the Baird instrument are shown in Fig. 1, and the only band observed below 650 cm.-', that of disiloxaned6a t 593 cm.-', is given in Fig. 2. The numerical values of band transmission minima are listed in Tables I and I1 and the Raman data in Tables I11 and I V .
Interpretation of Spectra.-The spectra, if complete, are too simple t o be interpreted on the basis (14) R. C. Lord and T. K. McCubbin, Jr., i b i d . , 46, 441 (1955).
SPECTRA AND STRUCTURE OF DISILOXANE AND DISILOXANE-d6
April 5 , 1956
100
TABLE I
r-
IXFRARED BAXDSOF DISILOXANE VAPOR Observed between 280 and 5000 cm.?. Wave no.
Wave no. (cm.-l vac.) a n d estd. intensity
lcni. -1 vac.)
and estd. intensity
Assignment
764 s 957 V V S 1107 VS 1220 ms 1465 v w 1556 \V
VI?; v 7 Y6
? VI0
v3
YE
+
VI3
vz
-I-
Ya
1761 vw
Y7
+
+ vi; iva + + + vs; vs + ve;
VI
3320 V V W 4450 w
Y6;
+
v9
YII
v5
vi2
v9;
VI2
V5
VI2
+ ui; + +
~3
1700m
+
\
Assignment
1910 w 2169 vs 2183 vs 3135 w
VI3
1329
VI
VI
+ + + + + +
v12;
v9;
vi
v12;
Y9
V6;
Y6
VI
PATH 5 CM PRESSURE 92 MM
VI
0; VI1
Y8
VI];
Y8
VI1
vi3
TABLE I1 ISFRARED BANDS O F DISILOXANE-d6 VAPOR
Observed between 240 and 5000 cm.-l. b \ ' a v e no. (crn - 1 vac.) and estd intensity
593 s 713 vvs 720 s 882 w 109.2 vs 1196 ms 1260 VlV 1385 m
W a v e no. (ern.-1 vac.)
Assignment
a n d estd. intensity
-
1575 vs 1710m 1862 m 2290 m
Y13
~7
VI? VI]
YQ
Assignment Y S ; VII
? Y2
vI Y5
Y6
? Y3
vi
+
Yi;
V3
V S ; YY
+
+
Y8 Vi2 YIZ
2660 V W 3150 w
VI
VI v5
+ + + + + 3+
Y6
+ + va; + + + vu +
V I ; VI
us;
~ 1 2 ;
v7
Y'l
v12; Y4 V6;
Y6
Va
v 5 ; VI Ye;
VII;
VI1
TABLE 111 RAMAN SPECTRUM OF LIQUIDDISILOXANE Wave no. (cm. - 1 vac.)
Character
2174 (2174) 1009 947 716 606
Sharp Under Y I Sharp Diffuse Diffuse Sharp
RAMAN Wave no. (crn. - 1 vac.)
Estd. int.
Assign ment
VS
VI
-
W
VU
vi
s
v9
S
VI 0
S
Y3
TABLE IV SPECTRUM O F LIQUIDDISILOXAXE-&
1575 1546 771 699 ( 555) 532 468 Approximate.
Character
Sharp Diffuse Sharp Diffuse Under Y I O Diffuse
Estd. int.
Assignment
vs
VI
W
Y8
W
vz
W
YQ
s vw
Y10
-
v3
Spurious
of Czv or .Cs symmetry. T h e lack of coincidences between infrared and Raman frequencies, apart from those which can be regarded as accidental, implies a symmetry center in the molecule. T h e only plausible structure with a symmetry center is that of a point group D3d. However, it has been pointed out by BaumanI5 t h a t the vibrational selection rules for a free internal rotator of the ethane type are indistinguishable from those of point group D 3 d . Thus the interpretation of the spectra in terms of a D3d structure is ambiguous since vibrational spec(1.7)
R . P Bnurnnn. J C h ~ i i fP h y ~ . 2. 4 , I 3 ( I < l > t i l
550 600 650 700 Frequency in cm.-l. Fig.2.-Infrared band of disiloxane-d6 a t 593 c m . - I
450
500
tra cannot serve to distinguish a rigid D s d structure from a freely rotating one. I n Table V is listed an assignment of the infrared and Raman frequencies based on the DSd model. Since the evidence against this model comes mainly from the rotational structure of the infrared bands, t h a t evidence will be discussed below in connection with the results of the high-resolution studies. The assignment of observed frequencies to the various vibrational species was made as follows: Species AI,.-Since no polarization studies were carried out, totally symmetrical Kaman lines were identified by their sharpness and intensity. There is little doubt about the assignment of the Si-H and Si-D stretching frequencies a t 2174 and 1573 cm.-l, nor about the deformation frequencies at 1009 and 771 cm.-l. Comparable but slightly lower values have been observed in the Raman spectrum of disilane vapor by Stitt and Yost.I6 The third AI, vibration, Si-OSi symmetrical stretching, is observed a t 606 cm.-', considerably above the value of 435 found for the analogous vibration in disilane." This difference is not surprising in view of the greater stability of the Si-0 bond than that of the Si-Si bond. Unfortunately the corresponding vibration in disiloxane-de cannot be clearly observed. It is expected in the range 550-573 cm.-l, which would place it quite close to the broad line centered a t 532 cxn.-l. The estimated value of 555 ern.-' is in satisfactory agreement with t h e product rule. l7 (16) F S t i t t a n d D hI Yost, rhrd, 6, 90 (1937) (17) T a h l e V last column Diwu.;sion of t h e product rule is g i b e n by F H a l \ c r s < , n f(c18 lfod Phis 19, 87 (IS471
1330
R.C. LORD,D. iV. ROBINSON .IND b-. C.S c ~ i u m
1-01. 7s
T.4BLE \-
SYMMETRY SPECIES. SELECTION RCLESANI) FREQUESCY ASSIGSMESTS O F FUXDAMESTALS OF DISILOXAXE A N D DISILOXANEds [Dad STRUCTURE) Species
L'ibn. no.
A4,g
1 2 3
A,, 41,
4 5
6 7
E,
E,
X $1
10 11 12 13
11
Approximate iorm of vibration
Si-H stretching SiHj deformation Si-0-Si stretching Torsion Si-H stretchirig Si-0-Si stretching SiHj deformation Si-H stretching SiH; tlefortiiatioii SiIL rocking Si-H sttetching SiH:; deforinotion SiHl rocking Si-O--Si bending
Raman
Selection rules
Infrared
Active sharp (Pol.)
Silent
Silent Silent
Silent Active parallel
Assignmcnt (Si€13)?0 (SiDaliO
2174 1009 606 . .
21ti9 1lOi
Active diffuse (Dcpol.)
Silent
Silent
Active perpciitliculitr
93i 2174
947 7 lii
Species E,.--This is the only other species active in the Ramari effect. The Si-H stretching frequency of this species lies very close to the totally symmetrical one. It was not resolved in the spectrum of disiloxane and thus is given the same frequency, 21T4 cni.-l. In the deuterium compound the two Raman lines are not superimposed and the E, stretching frequency is observed a t 13413 cm. -'. The two remaining vibrations are the SiH3 deforriiation and rocking. In the hydrogen compound these are found a t 947 and T l t i tin.-', respectively, and in the deuterium compound a t 699 and 532 ern.-'. A weak line at 46s cm. in disiloxane-ds cannot be assigned. It occurs only after long exposure (>SO hoursj and falls below the lowest observed line in disiloxane by more t h a n a factor 1/45, It is believed not to belong to disiloxane-dc. Species A2,,.-The three infrared-active vibrations of this species, Si-H stretching, Si-0-Si antisymmetric stretching and SiH3 deformation. can be identified by the parallel structure of the infrared bands. In other silyl compounds such as SiH3Br18r19 and in GeH3C1and C,eD3C1,*Othe parallel vibrations occur a t very similar b u t slightly lower frequencies than the corresponding perpendicular vibrations. In disiloxane the parallel Si-H stretching vibration is shown by high-resolution spectra to lie a t about 2169 cm.-l, just below the perpendicular vibration. The band center was estimated from the band's transmission minimum. In the heavy compound, both parallel and perpendicular Si-D stretching vibrations appear to coincide a t 1575 cm.-'. Although in disilane the parallel and perpendicular SiH3 deformation vibrations are well separated2' because of close mechanical coupling between the silyl groups, i n disiloxane the two vibrations are superimposed a t 957 tin.-'." This can be seen in ' 181 I) 1%. ; ~ I a \ o € I I< Oilit%A t i d J 5 . Peake, J Cli,.n, P l i ~ r 23, , I J i l fIU.jiT.) (19) See also K . S D i x o n and S Sheppard. i b r d . , 23, 215 (IrJSS). i ? O ) R . C. Lord and C 51 Steese. i b i d . . 2 9 , 542 (1954) (21) H S Cutowsky and E 0 Stejskal, & b i d ,22, 939 (19341. f2?1 In the iiilrired spectrum of monr~chlorr~silane the appear.tnce nf t h e !ifiO rrn ! r e p < , ni s quite similar t o t h a t in di~iloxane.indicating that the parallrl and prrpendicular SiHs deformation vibrations :grv also superimposed in t h a t molecule. l h e band in SiHsCI a t 1011i.i cm - 1 attributed by Munlile irbrd., 19, 138 (1951); Coinpt ~ P I I ~ 236, . ,
211s:3
