56 A VERIFICATION OF ANDRADE'S FORMULA OF VISCOSITY N.S. CHENG (86 75 #8) Department of Physies, University of Cheiang (Received December 23, 1940) ‘The well known formula of Andrade? for the viscosity 7 of a liquid at its freezing point is T=4 volo. a where m is the molecular mass, ¢ the average intermolecular distance (center- to-center), and ¥ the freqiency of mole: cular vibration of the liquid. Denoting ‘molar volume by 7, Avogadro number by N, freezing point by 7, gas constant by R, molecular weight by Mf, and molecular radius by a, we may put =winyh and, as given by Lindemann’ theory? of fusion, veer hy "hy", where co¥Z glhwth, e p being'a fraction defined by dase -o) ° Equation (4) means that ag is the short. est face-to-face distance between the assumed spheres representing two neigh- boring molecules when they are at their mean positions. In support to Andrade’s theory, we have found a simple, though not very rigorous, way of verifying formula (1 The motion of each molecule may be treated as one in a discontinuous mediuun composed of all the rest molecules, The resistance offered by this medium to each molecule moving through it with velocity R=4nehlf, «) where pats Loteeso.2m “M4, (5 T being the mean free path of the mole- Mf we set £ equal to ep, i.e., twico the amplitude of vibration in Lind: mann’s theory of fusion, equation (9) gives cules, . 6) «ie Adopting Groneisen’s value! 2.8 x 10! for C, equation (2) yields p=0.124. By (6) and (5), we have f=1.27 for any Higuid freezing point. The equation of motion for any chosen at its direction x isA VERIFICATION OF ANDRADE’S'FURMULA OF VISCOSITY mos ax ae ae where X is the force causing motion and +X, a=4nalf, ” As in Langevin's theory of Brownian movement, multiplying through by x changes this equation into 2 9 at y am ae a 4 aa) Toa 7% Taking the statistical average of this ‘equation over all the molecules, we have FA preg ae ce di since Xz=0 and, by the principle of ‘energy equipartition, $m GA 4aT, 4% being the Boltzmann constant. The solution to this equation is ast) “at PAT (Ie @ If ¢ is greater than 10 second, the exponential term in the bracket is negligible compared to unity for all molecules except hydrogen. In this ‘case, neglect of this term gives fen 22 fia, os (Ee BE, 7 UL Cis the time between two successive collisions, i.e., equal to 1/2v, then s/t ‘is the average component speed, 3, of ‘the molecules along the x-direction Le. Substituting this expression of a into (2), we have ay which, by (3), becomes Putting in the munerical values of f and p previously given, we have (8) which is only 1.09 times that of An- Grade’s expression (1). For diatomic liquids, the molecular shape is not spherical and therefore58 not expected to give It however yields the equation (8) accurate values. correct tendency of change for different * ‘molecules as shown in the following table: Freezing Molecule FECES calculated observed Ch WK 7.7ep. 10.3e.p.% ire 2650) 016) = lise le, 286 14S B.S O° 4 4 ar Hy 14 0.037 0.0138 We note that the discreparcy is largest Lindelaann, Phys. Zits. 11, (1912) 608, orzog, R: and Kudat, HL, Zelts. . Phys, N. Sc CHENG in the case of hydrogen’ This is in the present thoory due to the fact that for hydrogen schich is certainly not riegligible come pared to unity so that the simple solution here used isnot applicable, The writer is indebted to professor HOFF LU for his suggestion’ of the pfoblem and his colstant advice and encouragement. ‘Andrade. E. N. 8, C., Phil. Mag, 17, (104) 497, 615, (169) 217, Davies C.N., Proc. Roy. Soe, ST, (045) 259, Gerunsisen,-Awn. der Phys., 99, (WI2) 257 Steacie and Johnson, J: Am. Chem. Soe. 97. (1925) 154. Rulako and Schubaikow, PAYS. Z. Sowjetunion, 6, (W&M) 470,” Verschatfelt ,}.E,, Comm Phys. Lab. Leiden, 180, b, (197) 16 eueeteowe