K. J. PALMER
2360 !CONTRIBUTION FROM THE GATES AND
1'01. 60
CRELLIN LABORATORIES OF CHEMISTRY, CALIFORNIA INSTITUTE No. 6551
OF
TECHNOLOGY,
, Sulfur Dichloride, Oxytrichloride, and
Chromyl Chloride BY K. J. PALMER
Despite the rapid development of the electron diffraction method of studying the structure of molecules, comparatively few inorganic compounds have been investigated by this means. The stress which has been laid on organic molecules is due in part to the large number of organic compounds which have appreciable vapor pressures at, or near, room temperature, as compared with the number of inorganic compounds in this category, and in part to the fact that the organic chemist has in general been more interested in stereochemistry than the inorganic chemist, and has built up a large amount of purely chemical evidence supporting definite configurations. From a purely structural point of view, inorganic molecules are nevertheless of great interest. Especially is it desirable to determine whether or not the empirical relations connecting bond distance with bond type, which have proved so useful in the discussion of interatomic distances in organic compounds, are applicable to inorganic molecules as well. The investigations reported in this paper were carried out with these considerations in mind.
Preparation of Materials Sulfur Dichloride.-The sulfur dichloride used was prepared by passing dry chlorine into sulfur monochloride until the color of the liquid became deep red. This liquid was then distilled carefully in an all-glass apparatus, and only a small fraction boiling a t 50' at atmospheric pressure was retained. This product was redistilled twice, the middle fraction only being retained. There was no evidence of decomposition during the distillation. Sulfur Monoeh1oride.-Sulfur monochloride was prepared by passing chlorine over hot sulfur. The product was purified by distillation over activated bone charcoal and sulfur. The final product was light yellow and boiled a t 135436". Thionyl Chloride.-Eastman thionyl chloride was purified by fractional distillation under reduced pressure. Sulfur Trioxide.--Kahlbaum c. P. sulfur trioxide was used without further purification. Sulfuryl Chloride.--Sulfuryl chloride was prepared by passing a mixture of sulfur dioxide and chlorine over camphor. The product was purified by distillation. Chromyl Chloride.-Chromyl chloride was prepared by heating potassium dichromate, potassium chloride, and
concentrated sulfuric acid. The blood-red liquid was purified by distillation. Vanadium Oxytrichlorlde.-Vanadium oxytrichloride was prepared by passing hydrogen over heated Vn06until it was completely reduced to VzO*. The water which was generated by the reaction was carefully driven off, and chlorine was then introduced. The generated vanadium oxytrichloride was condensed in a trap cooled by a mixture of ice and salt. This product was purified by repeated distillation
Experimental Method The method of obtaining and interpreting the photographs already has been described in the literature.' All radial distribution curves were calculated using C (= sze-asi) in place of I as recently suggested.2 Because the vapor pressure of some of the compounds investigated is very low at room temperature it was necessary to use a high temperature n ~ z z l e . ~This design of nozzle was found to be particularly advantageous in these cases because of the hygroscopic nature of the compounds. Sulfur Dichloride.-The photographs show six well-defined rings. The second and fourth maxima appear to have shelves on the outer edge, the shelf on the fourth maximum being more pronounced than that on the second. The third and fifth minima (second and fourth on the reproduced curves) appear to be broad and less deep than the fourth. Values of so [= 47r sin /2/X], I (the visually estimated intensities) and C (= s2e-asH) are given in Table I. The radial distribution curve is reproduced as curve A in Fig. 1. The two well-defined peaks at 1.98 and 3.06 A. correspond to the S-C1 and CI-Cl distances, respectively. The Cl-S-CI angle is 1Ol01O'. Intensity c w e s calculated for S-C1 = 1.98 A. and CIS-GI angle equal to 1Ol01O', 109O28', 125", and 180' are reproduced as curves A, B, C, and D, respectively, in Fig. 2. Curve E is the intensity curve for chlorine, which has been included because of t h e (1) L. 0.Brockway, Rae. Modern Phys., 8,231 (1938). (2) V. Schomaker and C. Degard, to be published in Tars JOUK NU. JOURNAL, 69, 2181 (3) L. 0. Brockway and K. J Palmer, THIS (1987)
STRUCTURES OF SOME SULFUR, VANADIUM AND CHROMIUM COMPOUNDS
Oct., 1938
possibility of the sulfur dichloride decomposing to give sulfur monochloride and chlorine. This curve can be eliminated because the photographs do not appear to have the regular structure reTABLEI SULFUR DICHLORIDE Max. Min.
1
I
C
10
3
8
5
6
5
3
2
2
1
1
1
2 2 3 3 4
4 5 5 6 6
so
Sa
4.545 5.724 6.910 8.934 10.424 11.790 13.178 14.999 16.397 17.851 19.227
4.18 5.55 6.92 8.65 10.48 11.87 13.24 15.30 16.76 18.12 19.52
s/so
(0.913) ( ,970) 1.001 0.968 1.005 1.007 1.005 1.020 1.022 1.015 1.015
Average 1.006 Average deviation 0.011 S-CI = (1.98)(1.ooci) = 1.99 * 0.03 A. C1-C1 = (3.06)(1.006) = 3.08 * 0.04 A. Calculated for the model with S-CI = -1.98 A. and CI-CI = 3.06 A. (I
quired. Curve D is also in disagreement with the photographs, in that all of the peaks have shelves on the outer edge, and the fourth minimum is not deeper than the third and fifth as is required by the photographs. In curve C the intensity relations of the minima are again wrong, and in addifion the third maximum appears to be broad and flat, whereas the photographs show it to be quite sharp. Curve B can be eliminated because the shelf on the fourth maximum is on the inner edge instead of the outer edge, as required by the photographs, and also the third, fourth, and fifth minima have the wrong intensity relationship. On the other hand, curve A (1Ol01O’ model) agrees very well with the photographs. The outside shelves on the second and fourth maxima are present, and the fourth minimum is deeper and narrower than either the third or fifth. Since this model agrees also with the results of the radial distribution curve, it is accepted as correct. The quantitative comparison given in the last column of Table I leads to an S-C1 distance
I
I
n
1
2361
I
Fig. 1-A. Fig. 1-B. Fig. 1.-Radial distribution curves for (A) sulfur dichloride, (B) sulfur monochloride, (C) thionyl chloride, (D) sulfur trioxide, (E) sulfuryl chloride, (F) chromyl chloride, and (G) vanadium oxytrichloride.
.
Vol. 60
K. J. PALMEK
eludes the structure in which there are two chlorine atoms attached to one sulfur atom, because under these circumstances the radial distribution curve would exhibit only two peaks of about equal height (the outer peak being due to the long S-C1 aiid Cl-C1 distances), or perhaps three peaks, the outer two lying very close together.i Of these two outer peaks that representing the long S-C1 distance should be approximately twice as intense as that representing the Cl-C1 distance. Both the heights and the positioiis of the peaks are, however, compatible with the model in which one chlorine atom is attached to each sulfur atom. For this reason intensity curves were calculated only fox models compatible with this latter W I I figuration. ----- ---- - - 1 -1ssuming the value 1.99 .A\ for 1 10 13 90 S- C1 distance (this value the Fig 2 - 1' 1ier)ieticaIintensity ciirve; for sillfur drchloridc. heiiig found in both SCh and of L . W T 8. The final values art" therefore taken SQ2Clz),and considering the peak a t 2.01 A. to be as S-Cl = I W =tC) O:i .'1 a i d the aliglc Ci S C1 = the weighted rnean of the two S-C1 and the one S S diqtatice, it is then possible to calculate an 101 d= 4". Sulfur Monochloride.--The sample of sulfur TABLE11 monochloride was transferred to the high temperSVLFLTR h f O V O C l I 1 ORIDE ature nozzle inside a moisture-proof box. 'SVheii \ l d \ \ ~ l t l f c S= S/'" the nozzle was opened after the exposures were I IO 3 I tii2 4 00 (0 864) made there was 1 1 0 trace of sulfur and the rviiiaiii2 ,T 00 i 3 48 ( 978) S 1 ,Ti8 6 72 I 026 ing sample was stili light yellow in c d )r , The photographs I )f sulfur moiiochloridc s l l ~ n ~ , 2 \ 2.w 8 10 0 9 s eight maxima, s e c:i ~ of which were measured 4 '4 If); 9 00 982 accurately. 'The Srst maximum appears to brb 1 7 tl I O 21ri 10 18 996 symmetrical, the second has a verq. decided outer 1 1I 450 11 ;i2 1 005 I .i i 1:3 237 1.7 10 0 990 shelf which is called the third niaximtiiil in the i; t;, Of10 15 20 1 009 tables, the fourth is synimetrical, the fifth very (I 5 I If1201 lfi:E 1007 broad, the sixth sharp, arid the seventh is ti01 very prominent. The values of sfl, f, and C are given i 3 2 1u :I83 19 $0 1 .Ol(i in Table 11. The radial distribution curve, c-al riverage 1 002 culated using the values o f C,is reproduced as *\verage deviation 0 012 curve B in Fig. 1. The peaks at 2 01, ;{.lS m i t i ins^ values. s-cl = I w * 0 03*A. 3.96 A. are interpreted as the sum of the S -C1 S-S = 2.05 * 0 O3A c1-s-SL = 103 * 2' and S-S, the long S-C1, arid the C1-Cl distances, * Calcwlated for the madel with S-S = 2 05 8, , S-GI respectively I'he small peak at 2 47 ,I1 5 iiot I 99 A. end CI-CI = :MY,A k+eti any siKiuticaiicc.. 'Thr radial d i \ t r i b u t l i ~ i lc i i n e i i n i ~ i ~ ~ c l i ~ t l ~ l v (4) Refer t o c u r v e C 1 1 fcri t h i o i i ) I Lhloridc
1
t j
I
-
('\-
1
1):
Oct., 1938
STRUCTURES OF SOME SULFUR, VANADIUMAND CHROMIUM COMPOUNDS
2363
SS distance. The value found is 2.05 A.6 Using all but the static “right angled” structure. The ‘
the values 1.99 8.for the S-C1 distance, 2.05 8. quantitative comparison of s to so for curve A is for S-S, 3.18 8.for the long S C 1 and assuming given in Table 11. the value 3.96 8.for the C1-Cl distance, the S-SY. Morino and S. Mizushimas recently have C1 angle is calculated to be 103’35’ and the angle measured the dipole moment of sulfur monochloride, reporting the value of 1 Debye unit. They between the C11-S-S and C l & S planes 97’7’. In calculating intensity curves the S-Cl and S S distances and the C1-SS angle were given the values 1.99 k.,2.05 k.,and 103’35’, respectively. This is justifiable because the well resolved inner peaks on the radial distribution curve are probably reliable to within 1%. Curves A, B, and C for Fig. 3 were calculated for models in which the angle between the two Cl-S-S planes is 97’7’, 180’ (trans), and 0’ (cis), respectively. Model C can be eliminated because there is no shelf on the outside of the second maximum, this shelf being a very distinctive feature of the photographs. Curves A and B agree qualitatively with the photographs. The similarity of these two intensity curves, both as regards in5 10 15 20 tensity and position of the maxima and s -. minima, makes the reliability of the radial Fig. 3.-Theoretical intensity curves for sulfur monochloride. distribution peak at 3.96 k. very doubtful, since the distribution curve is calculated with t h e have also measured the Raman spectrum in diaid of just the SOvalues and the visually estimated lute hexane solution and have concluded from their results that the “right angled” static strucintensities. No intensity curves were calculated for models ture is the correct one. It can be said in support in which there was either free or restricted rotation of this view that the electron diffraction results for the reason that there is no feature of the curve are compatible with this configuration. Ackermann and Mayer’ have investigated the which appears to be sensitive to the C1-Cl separation, and it is consequently impossible at the molecular structure of sulfur monochloride by the present time to determine accurately the nature diffraction of comparatively slow (about 6 kv.) or height of the potential hump restricting free electrons. They concluded that there is one chlorine atom attached to each sulfur atom and rotation. The static trans model can be eliminated, how- claimed to obtain best agreement when they asever, from other considerations. Five Raman sumed S-Cl = 1.98 A., S-S = 2.04 A., and frequencies have been reported16 whereas only C1-SS angle = 105’. They also state that the three frequencies of the trans model are Raman intensity curve calculated for free rotation gave active. It is evident therefore that the molecule best agreement with their photographs. An incan have only one static configuration, namely, tensity curve calculated assuming free rotation is the so-called “right angled” structure, or the chlo- almost identical with curve A, Fig. 3, and conserine atoms can be oscillating or rotating with re- quently their results are in good agreement with spect to one another. It is unfortunate that those obtained in this investigation. Thionyl Chloride.-The photographs of thionyl more reliability cannot be placed upon the radial distribution peak at 3.96 8.because its position chloride show seven measurable rings, having the and sharpness would then immediately eliminate visually estimated intensities shown under I, Table 111. The third minimum (second on (5) This value is in good agreement with that found in HtSa by Beach and Stevenson (private communication). curves) appears to be very deep, the fourth com(6) Y .Morino and S. Miaushima, Sci. Papers 1. P . C.R., 32, 220
(1 937).
(7) Ackermann and Msyer, J . Chem. Phys., 4, 377 (1936).
2.764.
C1-C1 distances, respectively. The value 2.07 A. for the S-C1 distance seemed to be a little large. Xfax. Min. i C 10 sa S/SO 3b ',s o and to make sure that an impurity was not the 1 10 4 4.78 3.98 (0.833) 4.08 (0.884) 2 5.79 5.18 ( .894) 5.29 ( .914) cause of the apparent anomaly, electron diffrac2 8 fi 6.w 7 05 G.80 .g8* tion photographs were taken of a new sample of 3 8.30 8.32 1.002 8.38 1 010 3 9 II 9.78 9 79 1.001 0.89 o 091 thionyl chloride. When these were measured, it 4 11.67 I1.Gll 0.994 was found that the so values from the two sets of 4 4 0 13.07 I ? 12 1 004 13 32 I 010 5 14 ,9 14 hti I) t , ~ 4 14 58 093 photographs were in very good agreement, indi5 2 3 16.07 Iri.OR Ool 7; cating that the large S-Cl distance is probably 6 17 43 1 7 'J? 0 9'94 IT 35 Yl+i 6 3 4 lti.(it, 18.5s 'j*>tp 18 77 I ooo real. 7 Intensity curves were calculated for the follow7 1 1 z?.IO 1?li 1.003 2243 l0lh ing models: planar, with S-0 = 1.44 8.,s C 1 = .\verdyt. 1) 99'1 ",), 2.07 A,, and a CIS-Cl angle of 110, 113, 116, hvcrag~debiationn OW (I 012 i'inal values: SO = 1 43 * 02 A c ) - ~ - c ~=L * 1" arid 120"; pyramidal, with S-0 = 1.44 A., S-Cl S-Cl = 2 07 * 0 03 A CI-S-CIL 114 * 2" = 8.07 C1-0 = 2.84 -4. and CI-S-C1 angle CI-0 = 2.R4 = I' 03 A CI-CI = 3 47 * 0.03 x equal to 106, 110, 115, and 120'. These last Calculated for the model with S-0 = I .4aA., S-Cl = four curves are shown as curves A, B, C, and D, 2.07 A., c1-0 = 2.84 A., and CkCl = 3.50 A Cab- respectively, in Fig. 4. None of the intensity lated for the model with S-0 = 1.44 .&., S-CI = 2.07 .&, curves for planar models has been reproduced bec1-0 = 2.84 A., and Cl-Ci = 3.40 A. cause they are all in very definite disagreement paratively shallow and broad, the fifth fairly with the photographs. Curve D cat] be elimisharp but of even less depth than the fourth, the nated because the intensity relationship between sixth deep and well defined, and the seventh the first three maxima is wrong; also, the intenbroad and not as deep as the sixth. The fifth sity of the fifth maximum should be less than that maximum appears to be less ititerise than either of either the fourth or sixth. Curve C is in good the fourth or sixth. The values of sc, I , and C are qualitative agreement with the appearance of the given in Table 111. The radial distribution curve photographs. Curve B is also in good agreement calculated from these values is shown as curve C except for the appearance of the shelf on the inin Fig. 1. The peaks at 1.46, 2.07, 2.84, and 3.42 side of the fourth maximum, and the equality in .i. are interpreted as the S-0, S-C1, C1-0, and depth of the fifth and sixth minima. The model finally selected is a weighted average of these two. Curve A is less satisfactory than B because the shelf on the fourth maximum has become too prominent and the intensity 0 relationships of the fourth, fifth, and sixth maxima, which are about right in curves B and C, are unsatisfactory. The quantitative comparison for models I3 and C is given in the last column of Table 111. The final values selected, taking into account both the results of the radial distribution curve and the quantitative comparisons, are S-0 = 1.45 * 0.02 w., s-CI = 2.07 =I= 0.03 A., (21-0 = 2.84 0.03 K., a l ~ d Cl-S-Cl angle = 114 * 2". This leads to an 0-S-C1 angle of 106'17'. The same value for the 0-S-C1 angle has been observed in sulfuryl chloride. Fig.4.-Theoretical intensity curves for thionyl chloride. V
,
~
w.,
'
\
f
I
Oct., 1938
STRUCTURES OF SOME SULFUR, VANADIUM AND CHROMIUM COMPOUNDS
2365
values of C are given in Table IV. The radial distribution curve is so, reproduced in Fig. 1,curve D. The two well-defined peaks at 1.43 and A m , / - A 2.45 A. are interpreted as the 2-0 and 0-0 distances, respectively. These values lead to an O S - 0 angle of 117'50'. Theoretical intensity curves calculated for models having \ ii an $0 distance equal to 1.43 A. \ and an 0-S-0 angle of 120°, 117'50'. 116'. and 110' are shown , I 1 , I 1 as curves A, B, C, and D, respec15 20 25 30 5 10 tively, in Fig. 5. Curve D is not S. acceptable because the shelf on the Fig. 5.-Theoretical intensity curves for sulfur trioxide. second maximum is on the outside, TABLE IV rather than the inside, and also because the third SULFUR TRIOXIDE maximum does not possess a distinct enough outer sa S/SQ sb S/SD Max. Min. I C so shelf. 1 10 6 5.692 5 . 3 8 (0.945) 5.64 (0.973) The rather flat top of the second peak with the 2 7.422 7.13 ( ,961) 7.35 .990 n
n
pl,
5-
'
subsequent steep slope on the inside exhibited by curve C, and the inner shelf on the fifth maximum are also in disagreement with the photographs. Curves A and B are very similar, except for the slight change in the nature of the inside shelf of the second maximum, and are both in good qualitative agreement with the photographs. The intensity of this peak reaches its maximum value a t about the same point in both curves A and B, but falls off on the inside at approximately a constant rate in A, while for B the rate is at first less than for A and then becomes greater. This change in the rate of falling off of the intensity in curve B would, if exhibited by the photographs, probably give the impression of a definite edge to the shelf, whereas no such edge would be expected if the intensity followed curve A. As no edge can be discerned on the photographs, model A is to be favored over model B. The ratios s/so for curves A and B are given in
2
7
10
5
8
4
6
2
2
10.757 12.251 13.576 17.038 18.461 20.860 23.087
1
1
27.10
3 3 4 5 5
,921 ,957 .994 .994 .980 .969 ,981
10.20 12.10 13.90 . 16.83 18.60 20.50 23.32
.948 .988 1.024 0.988 1.007 0.983 1.011
,992 26.90 Average 0.974 Average deviation 0.019
27.20
1.003 0.994 0.017
....
6 6
9.91 11.72 13.50 16.30 18.10 20.20 22.63
...
...
...
,..
(0.974)(1.47) = 1.43 A. (0.994)(1.48) = 1.42 A. Final values: S O 1.43 * 0 . 0 2 A. 0-s-OL = 1200 6 20
-
Calculated for the model with S-0 = 1.47 A,, 0-S-0 Calculated for the model with S-0 = anglea = 120'. 1.43 A.,0-S-OL = 117O50'.
Sulfuryl Chloride.-The photographs of sulfuryl chloride have eight measurable maxima. The values of so, I , and C are given in Table V. The radial distribution curve (curve E of Fig. 1) shows principal peaks at 1.43, 1.99, 2.76, and 3.85 (8)
A. Smits, N. F. Moerman and J. C. Pathuis, 2.ghrsik. Chcm..
BIB, 80 (1987).
R. J. PALMEX
2366
1
j
I
r'
-
f
I
-
'-\
.
1
VOl. 60
unsatisfactory by the shelf on the inside of the fifth maximum. The quantitative comparison for curve C is shown in Table V as
STRUCTURES OF SOME SULFUR,VANADIUM AND CHROMIUM COMPOUNDS
Oct., 1938
TABLE VI FINALVALUES O F INTERATOMIC DISTANCES FOR SULFURYL CHLORIDE, OXYTRICHLORIDE SOnClZ
1.43 * 0.02A. s-c1 = 1.99 * 0.02 C1-0 E 2.76 * 0.03 CI-CI 3.28 * 0.10 0-0 = 2.48 * 0.10 0-S-OL = 119'48' * 5' CI-S-ClL = 111'12' * 2' CIS-OL = 106'28' * 2'
S-0
CrOPCli
Cr-0 = 1.57 * 0.03A. Cr-CI = 2.12 * 0.02 C1-0 3.03 * 0.03 CI-CI = 3.54 * 0.05 0-0 = 2.49 * 0.10 0-Cr-OL = 105'6' * 4" CI-Cr-ClL = 113'16' * 3" CI-Cr-OL = 109'34' * 3'
2367
CHROMYL CHLORIDE, AND VANADYL
VOClS
V-0 = 1.56 * 0.04A. V-Cl = 2.12 * 0.03 C1-0 = 3.00 * 0.04 CI-Cl = 3.50 * 0.03
...
... CI-V-ClL = 111'17' Cl-V-OL = 108'12'
* 2" * 2'
The radial distribution curve (curve F of Fig. 1) values of the ratio of s to SO shown in Table VIII. shows principal peaks at 1.57, 2.12, 3.03, and 3.54 The final values selected for the parameters are A. These are interpreted as the Cr-0, Cr-C1, given in Table VI. C1-0, and C1-C1 distances, respectively. The Discussion intensity curve calculated for this model agrees The configurations found for the seven molevery well, both qualitatively and quantitatively, with the photographs. This curve is reproduced cules studied in this investigation are all in agreeas curve C, Fig. 7. Curves A and B were calcu- ment with expectation. The C1-S-C1 angle of lated for models in which Cr-0 = 1.57, Cr-C1 = 2.12, C1-0 = 3.03, and the 0-Cr-0 angle was given the values 109'28' and 107', respectively. The value of the 0-Cr-0 angle in model C is 105'6'. Both curves A and B are less satisfactory than curve C because of the disappearance of the shelf on the outside of the third maximum and the appearance of a small maximum at approximately s = 17. Intensity curves were also calculated for a planar and a tetrahedral model, but the disagreement with the photographs was so marked that the curves have not been reproduced. The quantitative comparison for curve C is given in Table VII, and the finally selected values of the parameters are listed in Table VI. Vanadium 0xytrichloride.-The photogtaphs of vanadium oxytrichloride have seven measurable maxima. The values of SO, I , and C are given in Table VIII. The radial distribution curve (curve G of Fig. 1) shows I I I principal peaks at 1.56,2.12, 3.01, and 3.50 A. Fig. 7.-Theoretical intensity curves for chromyl chloride and These values are interpreted as the V-0, V-Cl, vanadium oxytrichloride. C1-0, and C1-C1 distances, respectively. The intensity curve calculated for-the model 101' found in sulfur dichloride lies between the with the above distances is given as curve D, limits 90' for pure p bonds and 109'28' for sp3 Fig. 7. This curve gives excellent qualitative hybridization. agreement with the photographs. All other The generally accepted structure for sulfur curves which were calculated for models varying monochloride, that in which there is one chlorine slightly from the above were less satisfactory and atom attached to each Sulfur atom, has been conhave not been reproduced. The quantitative firmed. The C1-SS angle of 103' is in agreeagreement is also very good as is evident from the ment with the angle to be expected from a con-
I(.1. PALMER
2365 TABLE VI1 CHROMYL CHLORIDE Max
Min.
1
I 10
ti
4 435
7
8
8.783
c'
s/so
2 2
5.629
h
1.5
x
19
3
7
122 9 304 11.310 12.785 14 076 15 157
1
2
18.210
2
3
3
3
%
4 4
6 5
6 6
(0 933) ( .946)
969 978 991 I ni2 1002 1 001 1 011
..
..
18.45
....
..
1 013
21.216
21 35
1 004
f
c
*
4.13 5.33 6 57 7.90 9.22 11 48 12.82 14.09 1 5 30
.
.
.
Average 0 997 Average deviation 0 O l g Calculated for the model with Cr-0 = 1.57 b., CrC1 = 2.12 A,, 0-Cr-0 angle equal to 105", and Cl-Cr-C1 angle equal to 113". TABLE
VI11
VANADIUMOXYTRICHLORIDE M a x . Min
1
I
c
so
S"
S/SQ
10
4
3.95 5.11
(0.943) ( ,964)
7
6
4.193 5.310 6.700 8.261 9.427 11.150 12.730 14.170 15.36 16.81 18.23 19.91 21 10
8.22 9.42
0.995 ,999
12.88 14.07 15.20
1.012 0.993 .990
18.48 19.90 21.66
1.014 0.999 1.026 1.004 0.011
2
2 3 3
7
10
5
9
4
7
2
3
1
1
4 4 5
*5 6
6 I
"
...
...
...
*..
... ...
Average Average deviation * Calculated for the model with V-0 = 1.56 A., V-Cl = 2.12 A,, 0-V-Cl angle equal to 108O, and Cl-V-C1 angle equal to 111".
sideration of the electronic formula for sulfur monochloride, since, as in sulfur dichloride, the sulfur atom is surrounded by an octet of electrons and both the S-S and S-C1 bonds are single bonds. Thionyl chloride is pyramidal. The molecule would be planar if the S-0 bond were a double bond, and the electronic structure were similar , The unshared pair of elecr atom, however, causes the The sulfur tri
molecule is isoelectronic
The confirmation of the planarity of sulfur tri-
VOl. 60
oxide is also in agreement with the dielectric constant measurements which recently have been carried out.* The electronic formula which usually is written for sulfuryl chloride shows the sulfur atom to be surrounded by eight electrons. This makes each S-0 bond a single covalent bond (semipolar double bond), and leads one to anticipate that sulfuryl chloride will have a tetrahedral configuration. The actual configuration is considerably distorted from a regular tetrahedral arrangement, the O-S0 angle being increased from 109'28' to about 120'. The CIS-C1 angle is only slightly larger than that expected, having the value 1 1 1 O . It is interesting to note that the C1-0 distances in both thionyl and sulfuryl chlorides are nearly equal. The increase in the 0-$0 angle in sulfuryl chloride is probably due to a considerable amount of multiple bond character which is brought about by the swinging in of one or even two pairs of electrons from the oxygen atom. The excited structures no doubt disobey the octet rule which is, however, of little significance in a discussion of second row elements because of the comparative proximity in energy of orbitals such as the 3d or 4s orbitals in sulfur which can be used for bonding purposes. In chromyl chloride the 0-Cr-0 angle has decreased to 105' as compared to the large increase in sulfuryl chloride. No explanation of this phenomenon can be given a t present. The Cl-Cr-C1 angle shows an increase of a few degrees over the tetrahedral value. The close analogy in both ,distances and angler: in chromyl chloride and vanadium oxytrichloride is very striking. A similar phenomenon is observed in the case of the sequence of molecules Sic&, PC13, SC12, and CL, where the observed interatomic distances are 2.00, 2.00, 1.99, and 1.98 A., respe~tively,~ as compared to the sum of the single bond radii 2.16, 2.09, 2.03, and 1.98 8., respectively. The constancy of these observed interatomic distances is due to the fact that the double-bond character of each of the above bonds, which is a function of the difference in electronegativity of the bonded atoms, just compensates for the change in value of the sum of the single-bond radii. This same effect accounts for the observed similarities in distance in chromyl chloride and vanadium oxytrichloride. This phenomenon is not observed for first row (9) L. 0.Brockway, Ran. Modern P h y r . , 8, 231 (1936).
Oct., 1938
STRUCTURES OF SOMESULFUR, VANADIUMAND CHROMIUM COMPOUNDS
2369
TABLE IX FINALVALUES OF INTERATOYIC DIST~CE AND S ANGLES A-Cl, A. A-0, A.
c1-0, A. c1-c1,A.
0-0, A. C1-A-C1 L 0-A-0 L C1-A-0 L
s-s A.
CI-s-s L
SClr
StClr
1.99 * 0 . 0 3
1.99 * 0 . 0 3
......
......
......
3.08*0.04
......
101 *4O
......
...... ......
......
...... ...... ......
...... ......
...... 2.05*0.03 103 * 2'
SOCll
SOaCla
2.07 * 0 . 0 3 1.99 * 0 . 0 2 1 . 4 5 * 0 . 0 2 1.43 * 0 . 0 2 2.84*0.03 2 . 7 6 t 0 . 0 3 3.47*0.03 3.28*0.10 ...... 2.48t0.10 114 * 2' 111'12' * 2' ...... 119'48' * 5" 106 * 1' 106'28' * 2"
...... ......
chlorides because in these cases double-bond formation can occur only when a t least one other bond becomes ionic. This does occur in the case of some fluorides'O where a very large difference in electronegativity between the bonded atoms exists. The S O distances observed in the four molecules sulfur dioxide,l 1 sulfur trioxide, thionyl chloride, and sulfuryl chloride are from 0.06 to 0.09 A. shorter than the sum of the normal doublebond radii which is 1.52 A. This is particularly surprising when one considers the electronic formulas of these molecules. These formulas indicate that the $0 bond in sulfur dioxide possesses one-half, in sulfur trioxide one-third, and in thionyl and sulfuryl chlorides no double-bond character. Making a correction for the formal charges on the atoms, the values for the $0 distance predicted on the basis of structures of the octet type are 1.52 8. for sulfur dioxide and sulfur trioxide, and 1.69 A. for thionyl and sulfuryl chlorides. The fact that the observed distances are all nearly equal and much less than 1.52 A. indicates that excited electronic structures in which double and triple $0 bonds are present must make a considerable contribution to the normal state of these molecules, being, however, more important for thionyl and sulfuryl chlorides than for sulfur dioxide and sulfur trioxide. It is difficult to give a quantitative discussion of the observed interatomic distances in chromyl chloride and vanadium oxytrichloride because there is some uncertainty as to the values of the (10) I;.0. Brockway, J . Phys. Chem., 41, 185 (1037). (11) P. C. Cross and L. 0. Brockway, J . Chem. Phrs., 8, 821 (1935).
...... ......
SOa
...... 1.43 * 0 . 0 2
......
......
VOCh
CIoIC12
2 . 1 2 * 0.03 1.56 * 0 . 0 4 3.00 0 . 0 4 3 . 5 0 * 0.03
2.12 " 0 . 0 2 1.57 * 0 . 0 3 3 . 0 3 * 0.03 3 . 5 4 * O 06 2 . 4 9 *.0.10 113 3' 105 * 4' 109 * 3"
f
2.48*0.03
......
......
111 =+=2'
120
f
2'
......
......
......
...... 108
* 2'
......
......
f
...... ......
covalent radii of chromium and vanadium. If one assumes reasonable values of about 1.15 for chromium and 1.20 for vanadium, however, it is again evident that the Cr-0 and V-O distances are considerably shorter than the s u m of the single bond radii, indicating that also in these two molecules structures involving multiple Cr-0 and V-0 bonds make important contributions to the normal state. I wish to express my sincere thanks to Mr. Ray Clinton, who prepared the samples of sulfur monochloride, sulfur dichloride, sulfuryl chloride, and chromyl chloride, and especially to Professor Linus Pauling for his many helpful suggestions and for his encouragement during the course of this investigation.
summary. The molecular structures of sulfur monochloride, sulfur dichloride, thionyl chloride, sulfuryl chloride, sulfur trioxide, vanadium oxytrichloride, and chromyl chloride have been investigated by the electron diffraction method. The final values of the interatomic distances and angles are given in Table IX. It has been shown that sulfuryl chloride, vanadium oxytrichloride, and chromyl chloride have tetrahedral configurations which are, however, considerably distorted. Thionyl chloride is pyramidal, and sulfur trioxide planar. Sulfur monochloride has been shown to have one chlorine atom attached to each sulfur atom. The positions of the chlorine atoms cannot be determined with certainty. PASADENA, CALIF.
RECEIVED JULY 1 1 , 1938
