R. FANELLI
1832
[CONTRIBUTION FROM THE T E n S
GULPSULPHUR COMPANY,
Vol. tii N E W YORR.
N. Y . ]
Supercooled Sulfur and its Viscosity R Y R. FANELLI I t is not generally known that liquid sulfur, even though not highly purified, may be supercooled far below its nielting point. Maximum supercooling may be brought about by allowing liquid sulfur, purified as directed in an earlier article,' to cool slowly, i. e., at the rate of about 0.5O per minute, at the same time taking care to reduce to a niiiiimum all of the usual factors that tend to promote crystallization. The supercooled liquid may be kept indefinitely at 100' but its longevity decreases as the temperature is lowered. In the neighborhood of 50' it is possible only occasionally to maintain the liquid state in bulk (5 cc. or less) for about twenty minutes. When the temperature falls helow about 50". solidification always results. The rate of cooling and the purity of the sulfur are not critical factors in obtaining extensive
supercooling. For exaniple, partially purified sulfur' may be supcrwoleil reaclily to as low as 70' by thoronglily heating 25 g. a t a temperature of 140° and above in a tube (20 X 3 cm.) completely immersed in an air-bath and cooling a t room temperature by removing and suspending the tube at an anele of about 20° to the horizontal. Liauid sulfur in the form of ciropiets (3 mm. or less in diameter) may be supercooled to room temperature. On standing at this temperature the larger drops crystallize within one or two weeks, while the smaller ones may persist in the supercooled state for years. These drops are indistinguishable from drops of sulfur in the true liquid state except that they appear to be of such high viscosity as t o be solid. On contact the clear transparent droplets become opaque with solidification spreading rapidly from the ooint of indentation. In Fig. l'are shown drops of \\ 1 supercooled sulfur at room temperature and a drop (A) which has crystallized spontaneously. When heated t o above 120' the clear drops show no visible transition chances t o the true liquid state. :1 would appear Fig,' &iscorn. that this form of sulfur may be eter used to declassed with such undercooled termine liquids of high viscosity as of -elass.. certain allovs. , . waxes. etc. ,UT_ Supercooling of sulfur is comoarable to that shown hv its nearest homoloebelenium. I
The Viscosity of Supercooled Sulfur
PIE.l.-supercmled drops of sulfur on the inner surldcc of a glass tube at room temperature. ''A" drop which crystallized spontaneously. (I) R. F. Bacon nod R. Faaelli. Ind. E*& Chm.. 14, 1043 (1942).
Farr and Macleod? determined the "time of flow" of sulfur at various temperatures down tu SOo but did not convert these values to C. G. S. units. I n n e w of the anomalous viscosityJ of sulfur above lGO", and its possible importance for fundamental study in chain formation, it was believed worth while t o determine viscosity of sulfur in the supercooled region and thus complete the viscosity-temperature curve for the known liquid range, namely, about 50 to 444'. (21 C. C. Farr nod D; E. Mnrleod. Pror. Roy. Sor. ILoodon). l l S A , 584 (1928). (31 R. F. Earno aod R. Fs+nelli,THISJOURNAL. 66, GR!l (111:l1.
SUPERCOOLED SULFUR AND
Oct., 1945
ITS
VISCOSITY
1833
io
50
Fig. 3.-Viscosity
90 110 130 150 160 Temperature, "C. of supercooled sulfur: 0 , extrapolated; 0 calculated from F a n and Macleod d a t a ; X , determined.
Experimental Methods A special viscometer shown in Fig. 2 was used. After charging with purified sulfur' the viscometer was sealed a t its open end and mounted in a cradle set in an air-bath.a The cradle could be rotated through 360'. The temperature was given by an iron-constantan couple fastened to the narrow end of the viscometer. The viscometer was rotated from the filled position through 180' and the time of emptying the larger bulb noted. The viscomcter was calibrated with a water-glycerol mixture of known viscosity. The densities of sulfur in the supercooled regioil are not known but the necessary values may be obtained by extrapolation of the known values' between 115 and 160°, all of which fall on a straight line. The data obtained are shown in Table I and are plotted X as in Fig. 3. The continuous curve above 115' has been determined by two different methods and independent inv e s t i g a t o r ~ . ~The * ~ determined values above 115' fall sharply on this curve, thus proving the validity of the method employed. Farr and Macleod's "times of flow" were converted into C. G. S. units using sulfur as the reference liquid. These data are shown as on the curve and are based on the known viscosity and density of sulfur a t 120.4' : namely 11.0 centipoises and 1.805, respectively,
at which temperature Farr and Macled's "time of flow" is given as 47.6 seconds. Attempts t o determine the viscosity below 80' were unsuccessful, but by plotting all the supercooled values on log/log paper (Fig. 4), extrapolated values for temperatures between 50' and 80' may be obtained. These are shown as 0 in Fig. 3.
70
50
40
f
'g a 30
.-
U
TABLE I VISCOSITY OF SULFUR BETWEEN 50 Temp.. "C.
50 60 70 80.8 pA.7 87 88.8 93.8
Visc., c. p.
73 50 36 26.2 23.6 23.3 21.4 18.4
E
u AND
POISES Temp., Method "C.
Extrap. Extrap. Extrap. Calcd. Calcd. Detd. Calcd. Calcd.
97.9 100 104.5 106 110.3 130 145 155
155' IN CENTI20
Visc., c. p.
Method
17.1 16.3 14.7 14.4 13.2 8.9 7.4 6.7
Calcd. Detd. Calcd. Detd. Calcd. Detd. Detd. Detd.
. -
(4) A . M. Kellas, .I. C h e w . S O ~ 113, . , 903 (1918). )I(! C.C. Parr and D. B. Macleod, Proc. R o y . Soc. (1,nodon).97A,
80
(inzo).
10 50
GO 70 80 90 100 120 Temperature, "C. I:ig. 4.-Viscosity of sulfur between 50 and 120'.
1534
K. FANELLI
Vol. 67
Discussion The viscosity of a liquid depends not only upon its molecular complexity ,but also upon the molecular structure. The uniform decrease in the viscosity and the density of sulfur with increase in temperature to about lG0' indicates that the liquid is made up of relatively simple molecules. LJTarren and Burwel16 by X-ray analysis show that the sulfur molecule in the orthorhombic and nionoclinic forins is an eightniembered puckered ring and that just above the melting point S g molecules predominate. Ewe11 and Eyring7 by calculation from viscosity data confirm the presence of Sa molecules in the liquid up to about 100'. According to Batschinski* the einpirical relationship v = w c / q where v is the specific voluine, 9 the viscosity and w,c are constants, holds for a number of non-associated liquids. The expression indicates a linear relationship between the fluidity 1/q and the specific volume. The relationship between the fluidity aiid the specific volume of sulfur is shown in Fig. 5 . The determined data for teiriperatures above and below 115' are showii by and X, respectively, the extrapolated data below SO' by 0. A straight line passing through nearly all of these values tends to confirm the presence of simple nonassociated molecules in the entire liquid range shown in Fig. 3 . For a discussion of the properties of liquid sulfur above ICiO', see Powell and IZyring.3 , B
-+
Summary Liquid sulfur may be supercooled in bulk to ab(J1lt .io' and i n droplets to rooin temperature. Viscosity values of supercooled sulfur are giver) (low11to No. The nature of the supercooled liquid is esseiitially the same as that between 115 to about 160'. 75 EAST 45TH ST. NEW YORK,N.Y . ~._._.___
Fig. 5. - --Kdationship hetween fluidity atid specific volulne of ~ i i l f t i rfrom 50-120"
RECEIVED JUNE 28, 1945
(6) B. 13. Warren and J . 'Y. Burwell, J . Clrem. Phys., 3,6 (1035). (7) R . 13. Ewell and €I. IZyrinK, ibid., 6, 726 (1937). J. ~ ~ ~ ~ ~ z, l , ,,,,,,,;k, i ~ ~ ~ kj ~ i ~ ,, 84, , ~ . 6.13 , (1913). 18) ' ! I 1 I< E. l'nwell n n d H livrina, ' i "J~ O I T R N A I , , 66. (543 (1943)
