No. 3686, JUNE 22, 1940 NATURE 971 conditions have been rather neglected by physicists, Formation of Aggregates and Structures particularly in recent years. In my experiments, a form of apparatus has been in Dilute Solutions of Hydrogen devised in which a metallic behows, closed by a Bentonites needle valve, contains the gas-free liquid under test, and the possibility of inward leakage of atmospheric A YIELD value is noticeable in hydrog,m bentonites air is avoided by completely surrounding the bellows at a concentration lower than that at which thixo­ and valve by the same liquid. The needle valve tropy first appears. This seems to happen generally being closed, tension is gradually applied to the with bentonites. Some natural calcium bentonites liquid by slowly filling a suspended can by means show coacervation1 • Coacervation of fine sub­ of a small jet of water. The water flow is turned off fractions of a natural bentonite has been reported• the moment a sudden downward movement of the at concentrations above the thixotropic range. can indicates that the liquid has broken. The can Hydrogen bentonites, however, show coacervation is then weighed complete with the contained water, over a wide range and even below the thixotropic and the tension which existed in the liquid imme­ range of concentrations. Near about the concentra­ diately prior to breaking is calculated from the tion at which yield value becomes noticeable a sharp known effective cross-section of the bellows. change in several properties has been observed. The liquids so far examined by this method are The cataphoretic velocity rapidly increases and tends alcohol, ether and lubricating oil. It would be to a maximum. The conductivity, A, per gram. of premature to quote results in detail at the present colloid can be calculated from the relation A = 1000 stage. However, it appears that the maximum ten­ ().,c-Au)/0, where Ac is the specific conductivity at sion which the above liquids will withstand for long a concentration of C grams per litre and Au that of periods under the conditions described is simply its ultra-filtrate; it passes through a minimum if related to their surface tensions and vapour pressures. plotted against a or yO. The maximum tension increases as the temperature Aggregation during coagulation is often associated is lowered, and this increase can be largely explained with an increase in the cataphoretic velocity". Also by the changes in surface tension and vapour pressure. the mobility of 'monomeric ions' of 'colloidal R. s. VINCENT. electrolytes' has been found to increase when they Research and Technical Department, associate to form 'micelle ions' and pass through Callender's Cable and Construction Co., Ltd., a maximum, whereas the equivalent conductivity of Erith Works, the electrolyte as a whole passes through a minimum 4. Kent. The hydrogen ion activity of hydrogen bentonite May 17. changes almost proportionally with concentration. The total acidity calculated from the inflexion point5 in the sodium hydroxide titration curve does not generally vary with concentration, though a Crystal Structure of Phosphorus variation has been observed for one subfraction of a Pentabromide hydrogen bentonite•. IT has been shown 1 that phosphorus pentachloride The conductivities per gram of colloid calculated has an ionic type of crystal structure which is tetra­ from the relation A'=aH (lH + Zcon)/0, where lH Zcoll gonal and contains [PC14]+ and [PCl6]- groups. and are the mobilities of hydrogen and 'colloid' Phosphorus pentabromide is entirely different in ion respectively, are appreciably greater than A. crystalline form. It· is orthorhombic, and X-ray The free charge per particle calculated from aH, investigations give for the unit cell a= 5·6, b = 16·9, assuming the activity coefficient to be unity, is of c = 8 ·3 A., with four molecules per unit cell. The the order of 500 or 1,000 electron units. Onsager's 7 space group is Pbcm and the four phosphorus atoms conductivity equation . for a similar electrolyte in are therefore necessarily equivalent crystallographic­ true solution is inadequate to explain the form of ally and chemically. The possibility of a structure the A, vO curve. containing two different complexes similar to those The special features of these titration curves5 and in phosphorus pentachloride does not arise in this the discrepancy between observed and calculated case. It is impossible to fit in four molecules of PBr5 conductivities per gram colloid are, however, met of the trigonal bipyramidal form, and the appearance with in hydrogen clays and other similar colloidal of the diffraction patterns is entirely unlike that of a systems8 but are not of immediate interest. The molecular compound. Patterson and Fourier methods sols have strong absorption. Interpretation of the were used to elucidate the structure. It was found measurements of extinction coefficient is difficult. that the structure contains tetrahedral [PBr,]+ The latter decreases with the concentration and groups located so that a plane of symmetry passes passes through a minimum. Somewhat similar through each, and the fifth bromine is present as a changes have been observed with colloidal dyes• and bromide ion removed from the phosphorus at a ascribed to aggregate formation. The 'true viscosity' distance about twice that of the four covalently increases with O at a rate much greater than that linked atoms. It is natural to expect the stability of corresponding to a linear relationship. the octahedral complex to diminish with increasing Neither Einstein's equation nor Smoluchowski's radius of the halogen atoms so that the change-over modification of it can account for the high viscosity from the constitution [PC14]+ [PCl6]- to [PBr4 ]+ Br or its variation with concentration. Also the high is not surprising. viscosity cannot be fully accounted for by the H. M. POWELL. anisometry of the particles. Hauser and LeBeau10 Department of Mineralogy, D. CLARK. observed that the apparent specific gravity, pA, of University Museum, subfractions of a natural bentonite increases with Oxford. concentration and approaches a limiting value. April 30. However, pA observed by them in the most dilute 1 rowell, Clark and Wells, NATURE, 146, 149 (1940). solutions is definitely less than the true specific © 1940 Nature Publishing Group.