484 The Constitution of Sulphur Vapour .
By Sir James J. Dobbie , F.R.S. , and J. J. Fox , D.Sc .
( Received January 31 , 1919 .
) [ Plates 13 and 14 .
] The constitution of sulphur vapour has been studied by many investigators , the method usually employed being based upon the determination of the density .
In 1835 Dumas and Mitscherlich found the vapour density at temperatures near the boiling point to be 6-56 , which corresponds closely with the molecular formula S6 .
Deville and Troost carried out determinations at temperatures ranging from 860 ' C. to 1040 ' C. and obtained the value 2*23 which is that required by the formula S2 .
More recently Biltz* has shown that below 800 ' C. the density is greater than is required by the formula S2 , and at 468 ' C. becomes 7'8 which corresponds approximately to the formula S7 , without any constant value being observed between these temperatures .
Bleier and Kohnf found that when determinations were made under reduced pressure between 192 ' C. and 310 ' C. the density of the vapour gradually rose with increase of pressure and slowly but asymptotically approached the value Sg .
Evidence of the existence of molecules containing eight atoms has also been obtained from an examination of solutions of sulphur .
Biltz holds that the value obtained by Dumas and Mitscherlich is only of significance for the conditions of temperature and pressure under which it was determined , and affords no evidence of the presence of hexatomic molecules in the vapour .
His view is that only oetatomic and diatomic molecules have any existence , the former gradually dissociating into the latter as the temperature is raised until about 900 ' C. , the dissociation of the heavier molecules is complete and the vapour is composed entirely of diatomic molecules .
Above this temperature no further change appears to occur .
Premier , on the other hand , from a study of the curve representing the change of density with change of pressure^ considers that it is not unlikely that hexatomic and tetratomic molecules are formed as intermediate products of the dissociation of the octatomic molecules .
Of this , however , the investigation of the vapour density does not afford any conclusive evidence .
* ' Zeit .
Phys. Chem. , ' vol. 2 , p. 920 ( 1888 ) .
t 'Monatsh .
Chem. , ' vol. 21 , p. 575 ( 1900 ) .
f ' Zeit , Phys. Chem. , ' vol. 44 , p. 733 ( 1903 ) .
I The Constitution of Sulphur Vapour .
The colour of sulphur vapour passes through a remarkable series of changes as the temperature is raised .
According to Howe and Hammer* it is orange yellow just above the boiling point , deep red at 500 ' C. , and straw yellow at 634 ' C. The method followed in their experiments , however , did not admit of any great degree of accuracy in the measurement of the temperature at which the changes of colour occur , and no attempt apparently has hitherto been made to correlate these changes with the changes of molecular complexity which occur as the temperature of sulphur vapour is raised .
In a paper contributed to the Society in 1910 by Mr. J. I. Graham , B.Sc. , f the author described the narrow absorption bands with which the spectra of sulphur vapour abound , and from an examination of their distribution under various conditions of temperature and pressure concluded that they form two systems , one belonging to molecules of the formula S8 , the other to molecules of the formula S* At atmospheric pressure he found no evidence of the presence of any other molecules .
At reduced pressures , on the other hand , the spectra obtained below 520 ' C. , gave evidence , he considered , of the existence of molecules intermediate in complexity between the octatomic and the diatomic molecules .
The paper makes no reference to the phenomena of the colour changes which appear to us to afford the only satisfactory evidence of the existence of such molecules , and it is with these phenomena that the present paper principally deals .
In the method of experimenting adopted , the sulphur under examination was placed in a silica tube 100 mm. long and 12 mm. in diameter , closed at each end by a flat ground silica plate fused on to the tube .
A weighed quantity of sulphur having been introduced through a narrow side tube , the air was displaced by nitrogen , the tube evacuated to 8 mm. and sealed .
It was then placed in the silica or porcelain tube of a small electric resistance furnace , the space between the inner and outer tubes being packed with asbestos to prevent any light reaching the spectroscope except through the inner tube .
This tube was arranged so that its axis was in line with that of the collimator of the spectrograph used for photographing the spectra , and the light to be examined after passing through the tube was directed by means of a lens upon the slit of the collimator .
The arrangement will be readily understood from the accompanying diagram ( fig. 1 ) .
Various sources of light were employed , that of the Nernst filament being found most generally convenient for the visible spectrum and the ultraviolet as far as \ 3100 .
For observations beyond the range of the Nernst filament , a cadmium arc proved most suitable , not only because the spectrum * ' J. Amer .
Chem. Soc. , ' vol. 20 , p. 757 ( 1898 ) .
t ' Roy .
Soc. Proc. , ' A , vol. 84 , p. 311 ( 1910 ) .
VOL. XCV.\#151 ; A. Sir J. J. Dobbie and Dr. J. J. Fox .
of cadmium reaches far down into the ultra-violet , but because it possesses a continuous background upon which the most delicate bands can be readily observed .
It is possible that the strong lines of the cadmium spectrum may conceal bands which happen to coincide nearly with them in position , but we know of no other convenient source of light which is free from this objection and extends far enough into the ultra-violet for our purpose .
* A G Fig. 1.\#151 ; A , source of light ; B , silica tube containing sulphur ; C , side tube for introducing sulphur ; D , tube of electrical resistance furnace packed with asbestos ; E , case of furnace filled with magnesia bricks ; F , quartz lens to slit of spectograph .
The instrument employed in the experiments was a large Hilger quartz spectrograph .
The temperature of the sulphur vapour was ascertained by means of a platinum\#151 ; platinum rhodium thermo-couple placed in intimate contact with the tube containing the sulphur .
Photographs of the spectrum of the light from the Nernst filament or cadmium arc were taken , after its passage through the sulphur vapour , at approximately equal intervals of temperature between 400 ' C. and 1200 ' C. The weights of sulphur employed in the experiments were 0*5 , P0 , 2'0 , 4'0 , and 8-0 mgrm .
Fig. 2 shows the results obtained with 1 mgrm .
of sulphur .
The successive bands of the figure represent the spectra at the temperatures marked opposite to them .
From an inspection of these it will be seen that the effect of raising the temperature is very remarkable .
At first there is an increase in the absorption , in other words , the spectrum transmitted becomes shorter as the temperature rises .
At the same time the colour gradually changes from orange yellow to red .
The shortening of the spectrum continues and the depth of the colour increases until a certain temperature is reached when the effect is reversed , the vapour then becomes more and more transparent , the length of the spectrum increases , and the colour changes again to light yellow as the temperature is raised .
The temperature at which the reversal of the phenomena occurs i The Constitution of Sulphur Vapour .
has been found , as the result of many experiments , to be somewhere near 650 ' C. The effect of increasing the quantity of sulphur in the tube to 8-0 mgrm .
is , as will be seen from inspection of fig. 3 , to increase the general absorption and , as it were , push back the curve towards the red end of the spectrum .
At the same time the shortening of the spectrum in the red becomes pronounced .
But the general result is the same ; the absorption first increases to a maximum at or near 650 ' C. and then falls off .
Numerous experiments made with other quantities of sulphur gave results similar to those shown in figs. 2 and 3 ( Plate 13 ) .
In each of the series of experiments to which figs. 2 and 3 refer the weight and volume of the sulphur vapour was the same throughout .
It is to be observed that while each spectrum in figs. 2 and 3 is the spectrum of the light after its passage through a layer of the same depth containing the same weight of sulphur , the pressure was different in each case , and this may have influenced to some extent the course of events .
The influence cannot , however , have been great within the limits of pressure obtaining in our experiments , and we found in every case , whether operating with 05 mgrm .
or 8 mgrm .
of sulphur , that the point of maximum absorption lay near 650 ' C. Whatever alteration of the maximum point the alteration of pressure produces is therefore clearly too small to be measurable by the method employed by us .
One series of experiments was carried out at the atmospheric pressure .
For this purpose , the usual apparatus was modified , as shown in fig. 4 .
The tube containing the sulphur vapour was connected Fig. 4.\#151 ; A , silica tube for by means of a silica tube of narrow diameter with connecting with Cm a reservoir 0f nitrogen , which was so arranged that fig. 1 ; B , silica tube .
, ' filled with nitrogen ; equilibrium could readily be established by means of C , inlet for nitrogen ; a stopcock between the pressure of the gases inside D , rubbei tube closed the tube and that of the atmosphere .
As the by pinch-cock .
r temperature of the furnace was raised , some of the sulphur vapour distilled into the side tube connecting with the nitrogen reservoir , where it was allowed to condense , but kept liquid by means of a small gas jet .
The plug of liquid sulphur thus formed prevented any nitrogen getting into the observation tube while allowing of the expansion of the jC C( k B Sir J. J. Dobbie and Dr. J. J. Fox .
sulphur vapour .
In this experiment the effect of varying pressure was eliminated , but the fact that the quantity of sulphur in the observation tube altered with the temperature introduced another source of complication .
While the different spectra in this case are not those of equal quantities of sulphur , the variation in the amount of sulphur is not great enough to overpower the effect of temperature at 650 ' C. , at which the maximum absorption is again seen to occur ( fig. 5 ) .
The remarkable series of changes which the spectra undergo with rise of temperature can only be explained by assuming changes in the molecular constitution of the vapour .
If , however , Biltz 's view that only two kinds of molecule S8 and S2 exist be correct , we should expect the spectrum to lengthen or shorten continuously as the one form passes into the other under the influence of change of temperature .
The fact that it first becomes shorter with rise in temperature , in other words , that the absorption increases up to a certain point and that then it lengthens , the absorption diminishing when that point is passed , can hardly admit of any other explanation than that molecules of some degree of complexity intermediate between those of S8 and S2 and more highly absorptive than either , are formed as intermediate products of dissociation .
The fact that Biltz did not find any constant value for the density of the vapour between 468 ' C. and 800 ' C. does not preclude the possibility of the existence of molecules other than S8 and S2 , stable only over a comparatively limited range of temperature .
He made no determinations apparently between 606 ' C. and 1400 ' C. , but it is a remarkable circumstance that the density value for the temperature 650 ' C. found by extrapolation of his curve is 3'4 , which is very nearly that required by the formula S3 .
By itself this is perhaps of little import , but it acquires significance when taken in conjunction with the fact that precisely at the temperature at which the vapour possesses this density it also possesses optical properties distinct from those of the vapour composed of molecules of diatomic or octatomic sulphur .
The evidence for the existence of triatomic sulphur , which is thus indicated , is strengthened if we take account of the analogies between sulphur and oxygen .
As is well known , ozone possesses much greater absorptive power for light than ordinary diatomic oxygen.* How , if we assume the existence of triatomic molecules of sulphur constituted in the same way as ozone , we should expect them to possess greater absorptive power than those of diatomic sulphur , and this condition , as we have seen , is fulfilled by sulphur vapour at 650 ' C. As has already been mentioned , the spectra of sulphur vapour exhibit * Hartley , ' J. Chem. Soc. , ' vol. 39 , pp. 57 , 111 ( 1881 ) .
/ The Constitution of Sulphur Vapour .
489 numerous bands , both in the visible and in the ultra-violet regions .
The following is a list of the bands given by the spectra of 2 mgrm .
of sulphur at a few of the temperatures between 485 ' C. and 1150 ' C. at which observations were made:\#151 ; 1030 ' C. 830 ' C. 650 ' 0 .
485 ' C. 4849/ 4970 4989/ 4755 4784/ 4910 4960/ 4685 w 4714/ 4840 4885 4555 4679/ 4763 4820w 4520 4655/ 4705 4750 w 4496 4640/ 4665 4714 4480 4620/ 4640 4685 4455 4595/ 4630 4650 4445 4586/ 4603 4635 4410 4549 f 4580 4611 4379 4530/ 4555 4595 4360 4505 4540 4560 w 4340 4485 4520 4535 4311 4458 4500 4520 ; 4307 4435 4480 4500 4276 4420 4450w 4480 4255 4410 4420 General absorption after 4240 4394 4410 X 4430 General absorption 4370 4360 4340 4320 4311 4276 4261 4242 4230 4215 4189 4150 4127 4085 4045 4039 4020 3960 3890 3790/ 3720/ 3681/ General absorption after X 3630 4379 4365 4355 4340 General absorption after \ 4220 One very wide band with centre about 5190 after X 4220 / = faint band .
w = wide band .
From an examination of this list it will be observed that some of the bands occur in all the spectra .
Seeing that no further change takes place in the spectra above 900 ' C. , and that the density of sulphur vapour at that temperature corresponds with the molecule Sa , it may be assumed that the bands which occur at 900 ' C. are the bands of Sa , and inasmuch as these same Sir J. J. Dobbie and Dr. J. J. Fox .
bands also occur at the lowest temperatures at which the vapour was examined , we have here direct evidence that the dissociation into molecules S2 begins at or immediately above the temperature at which the sulphur vapourises .
With very small quantities of sulphur ( 0*5 mgrm .
) interesting phenomena are observed in the ultra-violet region , which of course cannot be studied when larger quantities are employed .
At temperatures below 650 ' C. no absorptive effect whatever can be detected in any part of the cadmium spectrum ; at 650 ' C. there is a slight weakening of the spectrum at X 2824 ; and 100 ' higher a series of clearly defined narrow absorption bands make their appearance at X3027 , X 2834 , X2886 , X2762 , X 2730 , X2702 , and X 2628 .
These bands become beautifully distinct at 855 ' C. They can still be traced at 1010 ' C. , but not so clearly , owing to the whole spectrum beyond X 2750 being only faintly transmitted .
At higher temperatures there is general absorption beyond X 2940 , and only the bands nearest the red end can be seen .
Using the Nernst filament as the source of light , similar phenomena are observed with a somewhat different arrangement of the bands .
At 650 ' C. there is one band at X 3422 .
At 700 ' C. bands occur at X 3422 , X 3398 , X3369 , X 3348 , X3314 , and X3298 .
The number of bands increases as the temperature is raised , the maximum occurring at 900 ' C. Bands observed with 0*5 mgrm .
sulphur , using a Nernst filament .
1015 ' C. 900 ' C. 800 ' C. 700 ' C. 650 ' C. 3875 3775 3422 3422 3422 only and 3775 3675 3398 3398 very faint 3675 3605 3380 3369 3605 3570 3369 3348 3570 3510 3348 3314 3510 3480 3329 3298 3480 3456 3314 i 3456 3440 3298 3440 3422 3290 3422 3398 3272 j 3398 3380 3256 3380 3369 3230 3369 3348 3217 * 3348 3329 3183 3329 3314 3156 3314 3298 3136 1 3298 3290 3290 3272 3272 3256 3256 3230 General absorption 3217 after A 3050 * 3197 3183 The Constitution of Sulphur Vapour .
491 Above 900 ' C. , as already stated , no further change is to be observed in the spectra , and we conclude that at that temperature all the molecules are in the diatomic state as indicated by the vapour density determinations .
The spectroscopic examination affords no evidence of the resolution of the diatomic into monatomic molecules at the highest temperatures at which we worked .
Bands observed with 0-5 mgrm .
, using the Cadmium Arc. 1010 ' c. 855 ' C. 3550 2990 3298 1 Series of 3416 2987 3216 bands from 3392 2982 3170 2742 3374 2976 3070 to 3358 2973 3063 2724 3343 2970 3027 and from 3325 2966 3000 2724 3314 2962 2990 to 3298 2958 2967 2690 3284 2954 2962 2650 3246 2952 2954 2630 to 3233 2944 2924 2510 3215 2935 2864 2437 3203 2929 2860 2435 3195 2924 2848 2433 3185 2916 2846 2410 3176 2913 2843 2404 3170 2911 2835 2399 3164 2826 2390 3152 2824 3142 2822 3125 2806 3116 2777 3104 2750 3097 3074 3070 2064 3055 3047 .
3044 3035 3027 3023 3015 3010 3005 3000 2995 to = wide .
At 855 ' G. marked series of bands occur : four bands between X 2742 and X 2733 , sharp narrow bands at X2717 , X2710 , X2696 , X 2693 , and X 2690 .
At 660 ' C. one faint band at X 2806 was observed .
No bands occur in any spectrum below 660 ' C. At 1175 ' C. the spectrum was extended at the red end for about 70 units farther than at 1010 ' C. vol. xcv.\#151 ; a. 2 R i 492 Prince B. Galitzin .
An Apparatus for the In measuring the wide bands recorded in the above lists the middle position of the bands has been taken .
There is evidence that some of these bands could be resolved into series by using a different type of spectrograph .
DESCRIPTION OF PLATES .
Plate 13 .
Fig. 2.\#151 ; Spectrum of 1 mgrm .
of sulphur .
Fig. 3.\#151 ; Spectrum of 8 mgrm .
of sulphur .
Plate 14 .
Fig. 5.\#151 ; Spectrum of sulphur vapour at constant pressure of 768 mm. Tube 100 mm. long .
An Apparatus for the Direct Determination of Accelerations .
By the late Prince B. Galitzin , For .
Mem. K.S.* The question of the determination of the acceleration of the true motion of the ground in various seismic phenomena , or of the motion in different parts of buildings , bridges , and all kinds of artificial structures , due to explosions , shocks , or oscillations of the ground , has a considerable theoretical and practical interest , since the investigation of these accelerations serves as a guide in the study of the mechanical forces by which these movements are caused .
The knowledge of these forces is particularly important in the design of buildings of all kinds in seismological areas , and also for the calculation of various elements ( ties or reinforcements ) of structures which are often subjected to vibrations caused by the action of powerful engines by shifting of large ' masses ( in the case of bridges or buildings on yielding foundations ) , by sudden shocks ( in the case of ships armed with heavy guns ) , and so on .
The problem of the determination of the true magnitude of the accelera* The two papers , by Prince Galitzin , which follow , are translated from the originals , which appeared in the 'Bulletin of the Academy of Science of Petrograd3 for 1915 .
Their publication in the ' Proceedings3 has , for special reasons , been authorised by the Council of the Royal Society .
Prince Galitzin was elected a Foreign Member of the Royal Society on March 23 , 1916 , and died May 17 of the same year .
The papers contain , therefore , the last work of a distinguished man of science .
It was represented to the Council that they were of high scientific value , but inaccessible to most men of science , owing to their only having appeared in the Russian language .
Under these circumstances the Council felt justified in departing from the usual practice of not publishing communications that have already appeared elsewhere.\#151 ; A.S. i Bobbie and Fox .
Roy .
Soc. Proc. , A , vol. 95 , Plate 13 .
Fig. 2 .
Fig. 3 .
Dobbie an Fox .
Roy .
Soc. Proc. , A , 95 , Plate 14 .
Fig. 5 .
I

