Document g0ameOn1LaYdLnxJJvr31YQ

(Reprinted from the Journal ol PbymcaJ Chemistry, 0, 381 f 19JM>.| Copyright 1956 by tha Awricm Chamical Society and reprinted by permieewo of the copyright owner. THE PROPERTIES OF ASBESTOS. II. THE DENSITY AND STRUCTURE OF CHRYSOTILE1 By Fred L. Pondsack Contribution from the Johna-ManviUe Research Center, Manthlle, New Jersey Rtctwd Auffutl W, 195S It is shown that the tubular structure hypothesized for chrysotile asbestos is not compatible with experimentally deter mined density values for sealed, solid block* of asbestos fiber. The data indicate that the fundamental fillers are packed together efficiently with very little void space. The only void space observed appears to lie associated with the volume oc cupied bjr sorbed water. Sorption and aesorotion of water probably causes reversible swelling of the fiber bundles. It is hypothesized that the fundamental fibers may exist ss sheet- or ribbon-like structures possessing a certain degree of distortion due to limited curvature about the fiber axis. The unusual nature of serpentine minerals (3MgO-2SiOj-2HiO) which occur in massive, pseudo-fibrous and fibrous (chrysotile) forms has led numerous investigators to attempt a structural (1) Preeedinc paper ia (hie aeries, F. L. Puodaaek, This Jocsnal, If, 892 (1955). analysis of the various species. Chrysotile was classified first as an amphibole structure with a re peating SigOu-* unit,' but this was revised later to a layer- or sheet-type structure with a repeating (2) B. E. Warns and W. L. Bragg. Z. Krut., T*. 201 (1030). produced JM " M MTC 000378 302 Fiieo I,. PtiNimr'K Voi. no SiA,-< unit.1 Although there appears to be general agreement that ehrysotile lias a sheet- or layertype structure, many workers have pointed out that the fundamental sheet structure must be dis torted in some way to account for the X-ray dif fraction patterns which are observed.*-* In recent years interest in the structural characteristics of ehrysotile has been stimulated by the almost simul taneous publication by Bates, Sand and Mink7 in this country and Noll and Kircher* in Germany of electron photomicrographs of ehrysotile which ap pear to show the fibers in the form of hollow tubes. Hillier and Turkevich* had previously noted this phenomenon in passing. Recently a number of at tempts have been made to correlate X-ray diffrac tion data from ehrysotile with the hypothesized tubular structure.1,-11 It is apparent that a tubular structure of the type hypothesized for ehrysotile will contain consider able void volume (f.e., space not occupied by chrysotile) when the fibers are packed into solid bundles of the type found in nature. Thus, if the density of naturally occurring blocks of fiber is determined in a liquid which does not penetrate the fiber blocks, the magnitude of the void volume can be ascertained, and the existence of a tubular structure can be shown to be either possible or highly improbable. Experimental Materials.--The ehrysotile fiber used for the density de terminations in water was selected from a supply of No. 1 Danville Crude from Canada. The analysis of this mate rial has been reported previously.1 Portions of the selected solid fiber blocks were broken off into unopened bundles about 2.4 cm. long and 0.4-O.S cm. in diameter. The samples were examined carefully for flaws and imperfections. Density determinations on these specimens are referred to as density values for ' 'blocks of fiber.'' After the experimental runs had been completed on these blocks of fiber, they were opened carefully by hand and found to be free of any for eign inclusions which might otherwise invalidate the meas urements. The density determination in air was carried out with an essentially flawless, translucent specimen of ehrysotile from Arizona. This sample was in the form of a slightly irregular block about 5 cm. X 3 cm. X 1.6 cm. with a mass of ap proximately 58 g. Distilled water with a specific resistance >5 X 10* ohms was used for density determinations in an aqueous medium. Density Measurements.--Two 25-ml. capacity pycnome ters fitted with thermometers and side-arm capillary tubes were calibrated to 0.002 cc. with the liquid being used in the density measurements. Sampler were weired out into the pycnometer flasks, and then they were outgassed in accordance with a method described by Tschapek1* and later used successfully by Culbertson and Weber.1* The procedure consists of adding enough liquid to the flask to cover the sample and then evacuating the system in a vacuum desiccator to just below the vapor pressure of the liquid. (3) B. E. Warren and K. W. Herisc, Phgt. Rn,, It. 92ft (1941). (4) V. A. Fock and V. A. Kolpinsky, J. PKy, U.3.3.8.. ft. 125 (1940). (ft) B. E. Warns. Am. Mineralogiit, ftT, 235 (1942). (ft) E. Anija, MinaraUg. .Va#.. St. 6ft (1944). (7) T. F. Bataa, L. B. Sand aod J. F. Mink. Scunca, 111. 512 (1950). (8) W. Nell and H. Kircher. Yatarm., ST. 540 (1950). (9) J. Hillier and J. Turkevteh, Anal. Cbam., II, 475 (1949). (10) E. J. W. Whittaker. Acta Crfti.. 7, 127 (1954) (11) E. J. W. Whittaker, ibid., ft. 201. 265 (1955). (12) H. Jacedaineki and G. Kunae, Yenee Jahrb. Mineral. ifanatah., 9ft, 113. 137 (1954). (13) M. W. Teebapak. RoUcvLZ.. tt, 343 (1933). (14) J. L. Culbertaon and M. K. Weber, J. Am. Cham. .tee.. 60, 2695 (1938). This method effectively removes adsorbed air from the sys tem. After the system hod lieen mitgnased a sufficient length of time (determined by repeated runs until a constant density was attained) the vacuum was broken and the flask filled with liquid and allowed to come to room temperature, 25 1. The pycnometer unit was assembled and the density determined. The actual temperature of the system was read to within 0 1" from the thermometer in the pyc nometer flask. After the density values had been obtained for the unsealed specimens the samples were collected, dried and the density calculated on both an ''as-ia" (.., original sample weight which includes sorbed water) and a dry basis. For the ``sealed block" density determinations unopened blocks of ehrysotile were weighed and then dipped in molten paraffin at about 78. This procedure formed a relatively smooth, even coating of solidified paraffin over the entire block of fiber. Subsequent gas adsorption measurements showed that the coating was impervious. The density of the ``sealed blocks" was determined in water, and then the value was corrected for the density of the paraffin present. Sorbed water content of the blocks was determined by de hydration runs on untreated fiber blocks from the same source. The untreated blocks were weighed under the same conditions of relative humidity os the blocks used in the density determinations. The density of the paraffin, 0.901 g./ce., was determined independently by coating glass rods of known volume with the molten paraffin. The density determination in air on the specimen of fiber from Arizona was carried out simply by mapping the con tours and profile of the block on graph paper and calculating the volume on the basis of the mapped dimensions. The block was weighed and then the sorbed water content was determined on a small portion of the block by the procedure described below. Water Determinations.--Experiments showed that re versibly sorbed water could be removed from ehrysotile by (hying the material at temperatures between 175 and 200. At temperatures much above 200 a slight, non-reversible water loss occurs in ehrysotile. Therefore, for purposes of calculating the density of the fiber samples on a dry sample basis the water loss up to 175 was considered sorbed water. As an approximation the sorbed water was assumed to oc cupy the same volume as an equal quantity of free water, and this correction was used to obtain the density of the solid fiber. Although the sorbed water may actually oc cupy a volume slightly leas than that of free water, this assumption is not critical to the conclusions reached in this work. Theoretical When hollow cylindrical tubes with an outer ra dius rt and an inner radius r, are placed together in hexagonal close-packing the ratio of the gross vol ume of the bundle of tubes, To, to the volume of the solid, Kg, is Va 2y/3 ( r, \ V~ r W - r,V (1) Equation 1 may be rearranged and reciprocal den sity values D substituted for corresponding volumes to give Ds (2) where Da * absolute density of the solid Do observed gross density of the fiber bundle Equation 2 reflects the relationship of the density of the solid to the observed bundle density when the liquid fails to penetrate the void volume. If the fibers are solid (t.e., r, = 0), or the liquid penetrates the intrafibril pores but not the interfibril pores, equation 2 reduces to D, ?^D0 (3) When the liquid penetrates all pores PRODUCE" Jill-83 MTC 000379 Mar., 1956 Tub Dunaitv and Pritucrt'itE nr ( 'iiuvsotii.e 363 Da, which is essential to the utilization of the pre ceding equations, can be calculated readily from X-ray diffraction unit cell data available in the lit erature. The data on the dimensions of a unit cell of the composition Mg*(OH)sSiOio are summa rized in Table I. Using theaverage values a = 5.33 A., 6 = 9.24 A., and e = 7.33 A. the absolute den sity of chiysotile is 2.56 g./cc. Since the samples used in this work contained about 2% FeO + FeiOj isomorphously substituted for magnesium,1 the theoretical density of 2.56 can be corrected to 2.58 g./cc. to take this into account. This correc tion is incidental since it is within the range of ex perimental error. Tabu I Unit Cbll (Mg^OH^UOi.) Data fob Chbtsotilb Axis Warns aasd Bragg* VUTM Henox4 Anije' Whitteker* Padurow* a 5.33 5.33 5.32 5.33 5 33 6 9.25 9.24 9.2 9.2 9 28 c 7.33 7.33 7.31 7.33 7.36 The magnitude of Do to be expected for hollowtubes of the dimensions which have been sug gested," rt = 175 A. and r* = 75 A., can be calcu lated from equation 2 by using Ds = 2.58. In this case Do = 1.91 g./cc. assuming a sealed block in which no penetration of liquid into the void volume occurs. Results and Discussion The data for a series of density measurements on both sealed and unsealed blocks of fiber are sum marised in Table II. It is clearly evident from the data on the paraffin-sealed blocks that they do not contain the void space required by a hollow tube structure. In fact, the results strongly suggest that the only appreciable void space in the solid blocks of fiber is that occupied by sorbed water. Further more, the void space is not constant, but increases as the sorbed water content increases. That is, the blocks of fiber must swell as water is sorbed. The only exception among the sealed blocks is sample Tabu II Dbnritt or Ssaud and Unsealed Solid Blocks or Fibbb Semple Sorhed HiO. Graee semple,* Chrysotile. * % *./. g./ee. Sealed 1 0 2.57 2.57 Sealed 2 0 2.55 2.55 Sealed 3 0 2 56 2.56 Sealed 4 0 2.56 2.56 Sealed 5 2.2 2.48 2 57 Sealed 6 1.3 2.48 2 53 Sealed 7 0.8 2.53 2 56 Unsealed S 1.0 2.51 2 55 Unsealed 9 1.0 2.53 2.57 Unsealed 10 0.7 2 56 2.58 Arizona* 2 0 2.45 2 53 * Density of gross sample including sorbed water but ex cluding paraffin. * Calculated on the assumption that sorbed water occupies a volume of 1 cc./g. The density of this sample was determined in air. no. 6 which has a density slightly lower than one would expect. In view of the consistent behavior of the other values it seems likely that the value for no. 6 may have been low because of a slight imper fection in the otherwise solid block of fiber. The fact that the sealed samples which contain essentially no sorbed water have density values closely approximating the theoretical absolute den sity of chrysotile indicates that the fibers in a de sorbed condition are packed together with very lit tle void space. This behavior is not compatible with a hollow tube structure. In order to account for the close-packing of the fibers it seems more plausible to view them as strip or ribbon-like struc tures which may be distorted by limited curvature about the fi ber axis. ' This type of structure awaits confirmation by X-ray diffraction studies. The density values obtained for the unsealed blocks of fiber in water indicate the consistent na ture of the measurements. In the unsealed blocks the only appreciable void space measured is also that occupied by sorbed water. Taken alone the values for the unsealed blocks could not conclusively establish the improbability of a tubular structure. However, the results on sealed blocks indicate that when unsealed blocks of fiber are placed in water, the blocks probably sorb water and swell although initially the only appreciable void space present is that occupied by water sorbed prior to immersion. Measurement of the density of the block of Ari zona chrysotile in air furnishes additional confirma tion of the improbability of the existence of a tubu lar structure in chrysotile. Considering the crude ness of the method, the results arc in good agree ment with the premise that the only significant void space in bundles of fiber is that occupied by sorbed water. This void space is much less than would be required by a tubular structure. A series of studies of the amount of water sorbed by chrysotile as a function of relative humidity indi cates that the maximum amount of sorbed water retained by blocks of fiber at 100 per cent, relative humidity and 25 is approximately 2.5 per cent. If, as seems probable, sorbed water causes the fiber bundles to swell, the maximum degree of swelling can be calculated using the sorbed water value at 100 per cent, relative humidity. One gram of de sorbed fiber occupies ^- 0.388 cc./g. When 1 g. of chrysotile sorbs the maximum amount of water the gross sample will weigh 1.026 g. and occupy a volume of 0.388 + 0.026 -- 0.414 cc./g. chrysolite The apparent change in volume of the block of fiber is 0.026 cc./g. chrysotile. The percentage in crease in volume over the range of complete desorp tion to complete sorption would be Experimental work is now in progress to observe di rectly the swelling of fiber bundles and to establish (IS) E. 1. W. WUiteker. Jtta Crttt., S, M3 (1032). (IS) N. N. Pedurow. Aid.. S, 304 (1930). (17) G. J. Ym|ud F. H. Hratoy, Tate Journal, M, Ml (1954). (18) Such alimited curvature bee been suggested for satigorite. a so* eaiied massive variety of serpeatiae, by J. Ztisemaa, JfinsreJeg. ,W09., SO. 498 (1954). produced JW-M MTC 000380 364 Vo!. GO whether the swelling is nf the predicted order of magnitude. No satisfactory explanation of the hollow-tube appearance of rhrvsotile libers when viewed in the electron microscope7-* " can be offered other than to suggest the obvious: the sample viewed in the electron microscope no longer bears a one to one re lationship with the native fiber. Whether this is the result of the treatment the fibers have received (19) R. K. tier. "The Colloid Chemistry of SUlce and Silicatee," Cornel] University Press. Itheee. N. Y.. 1955, p. 208. during the preparation or of the exposure to the electron beam in a high vacuum remains to be de termined. Acknowledgment--The successful execution of this work is due in large part to Mr. George Reimschussel who made many of the density measure ments reported here. Mr. Marion Badollet and Mr. William Streib made available certain excellent specimens of chrysotile which were measured in the course of this investigation. PRODUCED MTC 000381