Document Ner8eOwDOVwjZG9R2LgK7qLvb

wd Clay Minerals i the limits examined is not a ecompiished by an oscillatory iNT Hxe Baroid Division of The -ship to the senior author. neral alterations: Ph.D. DissertoCounty, Wisconsin: Amer. Mill. (1965) Investigation of selective Natl. Clay Conf., Berkeley, ver eilicatos in soils by chemical :>nf., Pergamon Press, Now York, PLAINTIFF'S EXHIBIT SA-510 SURFACE PROPERTIES AND TEXTURE OF CHRYSOTILES by J. J. Feepeat* and M. della Faelle'I' Laboratoire do Physico-Chimie Minerals, Institut Agronnmique, Universito do Louvain, Heveriee-Louvain (Belgium) ABSTRACT Nitbogen surface areas and pore-size distribution curves of various ehrysotiles have been measured using a continuous flow method. A model founded on a hexagonal close packing of fibers has been adjusted to fit the frequency distribution curve of the fiber outside diameters obtained from electron micrographs. From this model, theoretical distribution functions of the surface area versus the pore diameter were computed and compared to the experimental data. For one fiber only (i.e. Coalinga cbrysotiie), the good agreement between the computed and experimental data allows one to conclude that the external pores (between the fibers) and the internal pores (within the fibers) are free from any amorphous material. For the other studied ehrysotiles, the degree of filling of the pore system by amorphous materials was always higher than 50%. Under these conditions, hydration water cannot be removed unless the samples are pretreated in the 300s--400"C temperature range. On the oontrary, water is driven off from the "clean" Coalinga fibers at temperatures lower than 100C. Surface area measurements derived from water-adsorption isotherms correspond to those obtained with nitrogen after the hydration water has been removed. INTRODUCTION The early application of electron microscopy to the study of ehrysotiles had suggested fibers having an internal porous structure (Turkevioh and Hillier, 1949). The relationships between structure and morphology were emphasized in the following years (Noll and Kircher, 1.950,1951,1952) and consequently problems concerning the degree of filling of the void spaces were invoked. Young and Healey (1954), Healey and Young (1954) had shown that two kinds of surfaces might be distinguished. Internal pores seemed available to small polar molecules such as H2O but not to N2, unless dehydrated at rather high temperature. External pores between the fibers were considered as partially available to nitrogen. These assumptions were founded on the observation that removing water at about 400C increases the surface area determined with N2 to a value equivalent to that obtained with H2O. * The University of Louvain and SI.R.A.C.,- Tervuren (Belgium), f Etomit S.A., Kapelle-op-den-Boa (Belgium). 305 <4 306 Fifteenth Conference on Clays and Clay Minerals Pundsacfc (1956, 1961) suggested that the accessibility of external surfaces was not as complete as claimed by Young and Healey since mechanical treatment noticeably affects the B.E.T. measurements. Moreover, the apparent density deduced from water adsorption isotherms was close to the theoretical value. It was therefore concluded that the externa! pores (between fibers) as well as the internal pores (within individual fibers) are filled with materials of chemical composition similar to that of the bulk. In. the pore-size distribu tion curves computed from the HaO adsorption isotherms, two maxima were observed at 16 A and 13 A. They were assigned to the external pores and to the internal cylindrical pores respectively. Bates (1969), Bates and Comer (1959) tried to explain the divergent observations of Pundsack and of Young and Healey, assuming that the material filling the internal pores is a good water adsorbent. Maser, Rice, and Klug (1960) were able to prepare sections of chrysotile fibers perpendicular to their axis. The electron micrographs showed clearlv the existence of internal cylindrical pores hut did not permit any conclusion upon the degree of filling. In order to explain the fiber habits usually observed on electron micrographs, Whittaker (1957) calculated the screening effect of eventual amorphous materials on. the scattering of electrons. The scattering distribution curve across the fiber was not appreciably modified whether or not amorphous substances are present in the internal pore. Pacing these divergent interpretations, the present contribution aims to study systematically the surface properties of chrysotiles and, especially, to analyse carefully numerous pore-size distribution curves computed from Ns adsorption isotherms after various thermal pretreafcraents. Eeeent develop ment of a rapid continuous flow method (Cahen ef al., 1965) for measuring pore-size distributions has made this work possible. Moreover, the results obtained with Na have been compared with those obtained with H2O. The conclusions were checked with measurements carried out on electron micro graphs. PROCEDURES Chrysotile Samples The raw material was washed with water on a 400-mesh sieve. Pibers not smaller than 2 mm were separated by elutriation on a 16-mesh sieve. The origin and the structure formulae of the samples studied in this work are shown in Table 1. The X-ray diffraction diagrams as well as the D.T.A. cxirves fit very well the usual patterns described for chrysotiles. Typical T.G.A. curves are shown in Pig. 1. The weight loss obtained on calcining the samples ocours in two steps, below and above 450C respectively. Above 450C, the weight loss corresponds to the removal of 6.5 to 7 x 10~3 HaO mole per gram, in approxi mate agreement with the theoretical water content, i.e. 7.6 x 10~ H2O mole per gram. The weight loss of a few per cent occurring below 450C is assumed * Olay Minerals bility of external surfaces Healey since mechanical ts. Moreover, the apparent. ,vas close to the theoretical nal pores (between fibers) rs) arc filled with materials .. In the pore-size distribui isotherms, two maxima ,acd to the external pores . Bates (1959), Bates and rvations of Pundsaek and -) filling the internal pores pare sections of ch.rysoriie licrographs showed clearly not permit any conclusion iier habits usually observed dated the screening effect of electrons. The scattering liably modified whether or ual pore. sent contribution aims to .hrysotiles and, especially, ion curves computed from reatments. Recentdevelopet al., 196o) for measuring bio. Moreover, the results e obtained with H2O. The ied out on electron micro- 400-mesh sieve. Fibers not 1 on a 16-mesh sieve. The 3 studied in this work are D.T.A. curves fit very well al T.G.A. curves are shown the samples occurs in two >ve 450C, the weight loss mole per gram, in approxiit, i.e. 7.5 x 10-3 H2O mole Ing below 450C is assumed Surface Properties ash Texture of Chrysotix.es Table 1. --Omanis and Stbuctttkal Fobjiciae of Cfnnrsoin.ES 307 Fiber Origin Structure formulae 4M*v myi O10 (OH)b .--------------------------------Tetrahedral layer (MIV) Octahedral layer (MVI) ` Si>+ Al3-1- ' ` Mg3* Fe3- Al3* J.M. Jeffrey mines, 3.913 0.0747 5.775 0.1915 -- Russian Corsica Canada Oural, S.S.S.R. Canari mines, 3.9413 3.6712 0-0585 0.2076 5.5626 0.2824 0.155 5.757 0.243 ~ ** Corsica 11:: Cassiar Bell Asbestos 3.887 0.101 5.88 0.123 iff Mint;, Vancouver, !:. Canada Coalinga California 3.818 0.182 5.621 0.146 0.407 |ij Arizona Arizona 3.975 0.028 5.929 0.052 -- jjv |;.; |i; to represent the removal of hydration water either from internal or external if? pores. In order to investigate the effect of hydration water on surface proper if- ties, thermal pretreatments were carried out below 450C. The samples were :;S heated overnight in an oven at 100, 200, 300 and 400G and carefully protected against rehydration until use. In a few cases, the temperature was increased to 650C to produce a complete dehydroxylation. The changes in morphology and pore structure of the "thermally damaged materials" have |:'- been reported elsewhere (della Faille, De Kimpc, and Fripiat, 1966). & If Pore-size Distribution Curves In order to obtain a statistically significant number of experimental data, a rapid and continuous flow method was sot up for determining nitrogen desorption isotherms (Cahen et al., 1965). This technique involves passing a mixture of 15-20% nitrogen in helium through the sample cooled in liquid nitrogen at a total pressure of 4000 to 5000 mm of Hg, causing the N2 partial pressure to approach its liquefaction pressure and nitrogen to be adsorbed. By lowering then the total pressure in a continuous and steady way, the partial pressure decreases, the adsorbed gas is progressively desorbed and the amount ofN2 evolved is recorded as a function ofthe pressure. The continuous recording of a desorption isotherm requires a time of 1 to 3 hours, according to the nature of the sample, instead of days, with a conventional volumetric apparatus. The desorption diagram which gives the volume of nitrogen desorbed with respect to the nitrogen partial pressure is transferred to the computer. The calculation program proposed by Cranston and Inkley (1957) was used. Several presentations of the results are possible: the most adequate for this work was that showing the fractional surface area (AS, m2 per gram) i I s 1 308 Fifteenth Conference on Clays and Clay Minerals developed by pores having diameters between <I> and <3E>+5 A, being progres sively increased from 20 to 250A. Moreover, the computer provides the surface area calculated according to the B.E.T. method. The instrument developed for pore-size distribution measurements is derived from a conventional gas 5 chromatograph equipped with thermistor catharometers and with manom 1 eters measuring pressures up to 5 atm per cm2. The oven-dried {or calcined) I samples were degassed at the pretreatment temperature by passing a slow flow of helium (30 ml per min) overnight. . Fig. 1. Weight loss against temperature of several ehrysotile samples. Solid line: J.M.; broken line: Russian; dashed line: Corsica. Water Adsorption Isotherms Water adsorption isotherms were obtained using a technique described by Pundsack (1961). Samples, oven-dried at 300C and carofully protected against atmospheric moisture, were introduced in long glass tubes, immersed in a constant temperature bath maintained at 23C 0.1. Each tube contained H2O--H2SO4 mixtures of known water relative pressures, covering the range I 0.145<P/Po<0.895. The adsorption isotherm is obtained by weighing each sample when the equilibrium is reached, i.e. after 10-15 days. In a few cases the results were compared with those given by a vacuum electrobalance, operated at 23C. The agreement was found satisfactory. The pretreatment temperature of 3G0C was chosen in order to insure the removal of hydration water, as shown further. The surface area available to H2O was determined by the B.E.T. method, assuming the packing of physically adsorbed water molecules to be 11 A2. V | 1 *-.S^-nWM1T_------ J VM'Amv. jW.WWf.WlsjwiU'BAwXtfllWMMW.W ! D Clay Minerals -ud $+5 A, 3> being progres>mputer provides the surface . The instrument developed ed from a conventional gas rometers and with manom'he oven-dried (or calcined) perature by passing a slow Surface Properties and Texture of Chrysotiles THE THEORETICAL PORE-SIZE DISTRIBUTION A model has been set up for the theoretical calculation of the surface area distribution as a function of the pore diameter. The following assumptions were made: (1) the fibers are arranged in hexagonal close packing to form bundles, as shown in Fig. 2, in agreement with Fankuchen and Schneider (1944); there are consequently twice as many internal as external pores; (2) the length (L) of the fiber is considered as infinite with respect to its external diamoter (21?); (3) nitrogen cannot develop a monolayer in the regions (black in Fig. 2) where the distance between two adjacent walls is smaller than twice the molecular diameter (approximately 8 A). 700 *00 T(c) yeotile samples. Solid line: j: Corsica. =tg a technique described 1 and carefully protected mg glass tubes, immersed L 0.1. Each tube contained sures, covering the range .named by weighing each &-16 days. In a few cases vacuum electrobalance, rThe pretreatment I^W teiaoval of hydration IfjjQ was determined *?_... . - water not avatUbie to Nt available to Nj Fig. 2. Schematio representation of cylindrical hollow fibers in hexagonal close packing. The surface area 8 of the external pore is then =^1-^ ttRL -where sin J~, R being expressed in A. (2) The apparent radius of a cylindrical pore of equivalent surface area would be : The surface area of the internal pore is . s--'ltrrL. (4) In order to obtain the specific surface area (per gram), it is necessary to Tt 310 Fifteenth Conference on Clays axd Clay Minerals take into account the apparent density Ps given by (5) where 8 = 2.6. Consequently the specific surface area So is equal to the sum of two contributions, S* and a*, corresponding to the external and internal pores respectively, or: where +a,: (8) . 1--SG*cc--jiro! ~ R(\ -r2/7?2)8 and _ 2r a -V2/2)S' Let / and F be the frequency distribution functions of the internal pore diameter (2r) and the outside fiber diameter (2R) respectively. Let be the partial surface limited between the curve F(R) and increments of 20 A of R. Assume that r is proportional to R. It follows where and 5=2i 501^1=2i .^*^=22 1 --6aj/ir At l i Ri(i~r<2!Ri*)8 Eet*A, 2r ______A_t_.__ < JStHl-ri*/'Bt*)a (?) The assumption that r is proportional to R was introduced because of the uncertainty involved in the determination of the internal pore diameter from electron micrographs. The mathematical treatment of relationship (7) was performed for various distribution curves and various Rjr ratios. Figure 3 shows the F function obtained for Coalinga from measurements carried out on numerous electron micrographs, and the corresponding adjusted normal distribution. The most probable value obtained for 2R was 218 A. The most probable internal diameter was estimated to 45 A by measuring the widths of the void spaces running parallel to the axis of each individual fiber, as shown in Plate 1. The proportionality factor njRi was thus equal to 45/218 for this peculiar sample. Figure 4 shows the good agreement between the computed distribution of the specific surface area with respect to the apparent diameter 2rapp and the experimental distribution averaging data obtained after pretreatments between 100 and 400G. The apparent diameter (2rapp) has been chosen for this representation since the calculation of the pore-size distribution from the adsorption or desorption isotherms is founded on the assumption of cylindrical pores of infinite lengths. In Table 2 the \w/* !AV> mvnwwiko T * Clay Meverals >y (5) Tea- So is equal to the sum > the external and internal (6) Surface Properties and Texture of Chrysotiles 313. tions of the internal pore respectively. Let At he the d increments of 20A of R. Si*)Ai (?) At Af. introduced because of the ic internal pore diameter atment of relationship (7) various R[r ratios. Figure am measurements carried e corresponding adjusted flamed for 212 was 218 A. tsd to 45 A by measuring he axis of each individual i^tt rtjRt was thus equal to >good agreement between with respect to the 'Sribation averaging data .iX/|35iaapparent diameter fl5e:ithe.calculation of the -*w- * dherms is founded |> :Sn Table 2 the Fxo. 3. Frequency distribution curves obtained by measuring the fibers outside diameters on electron micrographs (Coalinga chrysotile). Brokenline: experimental distribution. Solid line: normalized distribution. contributions of the surface areas of the external (S*) and internal (a*) pores are tabulated, assuming either r< proportional to Rt or 2r=4oA. The total surface areas computed in both cases are in good agreement with the B.E.T. values. Table 2.--Computed and Experimental Surface Areas of the Coalinga Chrysotile (m2 per gram) r proportional to R ' 8* 15.9 S* 35.4 S 51.3 2r=4oA s* 's* B.E.T. after pretreatment at the indicated temperatures S ' 100C 200=0 300-500=0 15.7 35.4 51.1 54.6 52.0 53.4 From the comparisons in Table 2 it may be concluded: (1) that the Coalinga sample is composed of bundles of hollow fibers arranged in close hexagonal packing, (2) and that, from a statistical viewpoint, the frequency distribution curve of 212 approximates a single gaussian function. II < I ) s* ; /. Plate I. Electron micrograph of Coalinga chiysotile. An interesting consequence of the good fit betweon the computed and experimental distribution functions, as shown in Fig. 4, is the possibility of deriving the most probable outside diameter of individual fibers from the moBt probable apparent diameter 2rapp, whatever the degree of filling of internal pores. In order to demonstrate this point, consider the theoretical distribution functions of Fig. 5 obtained for variable r values but with the same frequency distribution function F(2R), shown in Fig. 3. The amplitude of the maximum changes greatly but it occurs always at the same apparent diameter. SYSTEMATIC STUDY OF THE PORE-SIZE DISTRIBUTION FUNCTIONS As shown in Figs. 6 and 7, the experimental pore-size distributions obtained for various chrysotiles have single maxima, the position and amplitude of which appear to be characteristic for each sample. This is still more apparent in Table 3, which contains the most probable apparent diameters, the outside SUEFACE PkOPEBTIES AND TeXTDBE OP CHEYSOTtt.ES 313 diameters derived from the former through equations (2) and (3), and the 11 E.E.T. surface areas. Some results are also given for two other fibers, i.e. Arizona (J.M. mines) and Advocate (Newfoundland). Table 3 shows the very peculiar habit of Coalinga chrysotile: the outside diameter is the highest of the series and the experimental specific surface area is almost identical to tmmitod and ssibility of from the Jitif' filling of ypS^-tfeorotical the Fig. 4. Distribution functions obtained for the Coalinga chrysotile. Broken line: computed function (equation 7) assuming r( being proportional to Stt. Solid line: experimental function averaging the results obtained after pretreatment in the lOO-oOO'C range. the theoretical value. This agreement is in favor of "free" porous spaces, not occluded with amorphous materials. For the other fibers, according to the previous calculations, the minimum surface areas expected from the most probable 2R value would be at least higher than 40 m2/g, whatever the value of the internal pore radii. Even after pretreatment at 400C, such surface areas have never been measured. Therefore it may be concluded that the internal or externalpores contain appreciable amounts ofamorphous materials. The degree of filling could be calculated if r could be determined accurately, but measuring this parameter from usual electron micrographs seems highly hazardous. Upon increasing the pretreatment temperature, the surface areas increase, sometimes slightly, sometimes more strongly. As proposed earlier, this may ^^slL At\'OAX.'AVwJ>w 314 FlFTSKNTH CONFERENCE ON CLAYS AND ClAY MbiERAiS result from removing water molecules filling partially the void spaces. How ever, the thermal pretreatment of Coalinga does not affect- tho surface area. Since the internal and external pore diameters are quite high for this fiber and the degree of filling with amorphous materials is very low, it seems 150 170 190 210 2R (A) 230 250 270 290 310 } i Fio. 5. Theoretical distribution functions obtained assuming the same Frequency distribution function as in Pig. 3, but for vsiiable internal pore radii (r). appropriate to suggest that water molecules diffuse freely in the channels and that they may be removed at temperatures as low as 100C. On the contrary, for a fiber such as Cassiar, the constancy of low surface areas at any pretreatment temperature should be explained assuming that tho degree of filling by amorphous materials is quite high and that eventual hydration water molecules cannot diffuse appreciably unless high temperatures are reached. Dehydroxylation and dehydration occur then simultaneously, deeply damaging the original structure (della Faille, De Kimpe, and Fripiat, 1966). : :- j | ; ! i ' i t ) r;: ------------------------------------ }; j. Stjbface Properties and Texture ok Chbysotiles 315 Table 3. Maik Textural axd Surface Characteristics i Fiber Pretroatmont Host probable S Most probable Most probable temperature 2rapp (B.E.T.) outside outside i diameter (2R) diameter (electron micrographs) (C) (A) m2/g (A) (A) Russian 100 27.3 13.5 200 32 5 14 0 300 27.5 21.0 156 l 400 32.5 14.2 500 32.5 19.7 Avorage: 30.0 (below 500CC) -- Corsica 100 200 300 400 500 37.5 37.6 37.5 32.5 32.5 6 7.3 11.5 17.0 12.2 176 -- Average: 37.5 (below 400:'C) 1 I Coalinga 100 49.0 64.6 218 200 49.0 52.0 210 220 (P) 300-300 49a 0 83.4 { Average: 49.0 : Arizona 100-300 22.5 12.9 136 400-500 22.5 17.G Average; 22.5 ! Advocate 100-200 32.5 17.2 300-400 32.5 21.2 162 500 32.5 21.5 I Average: 32.5 r J.M. t 100 27.5 lfi.6 200 27.5 18.4 Average: 27.5 150 i Cassiar 100-200 22.5 12.0 300 27.3 14.3 143 400 27.5 12.3 500 27.5 10.0 f; Average: 25.0 (below 400:>C) $ 159 (P) 160 (P) j Jf.B. The most probable outside diameters computed from electron micrographs and indicated with subscripts P are due to the courtesy of Dr. F. Pundsaek, Director of the Johns-Manville Basic Bescarch Department-. 1! 1 . <**i*v .0.`.v ,Ar..v-t.-. K i XkY Minerals CORSE Surface Properties and Texture of Chrysotiles 31" too* A 200* 17 300.400.50Cf 640* COAUNGA S w6 l| CO <1 i i i i J.M. 4h \ `1 3H \\\ni ill 2r \\ % 1r \ 70 80 90,, 100 20 30 40 SO 60 70 80 90, .100 0 (A) ______ 0 (&) Fla. 7. Distribution functions obtained for Coolinga and J.M. chrysotilea. WATER ADSORPTION ISOTHERMS It was particularly interesting to check whether the increase of the specific surface areas measured with nitrogen was due to removal of hydration water molecules upon heating. This idea, already proposed by Young and Healey (1954), needs to be adapted to take account of the presence of amorphous materials, evidenced above. Specific surface areas measured with water and obtained after pretreatment at 300C are compared in Table 4 with the maximum specific surface areas measured with nitrogen after pretreatment in the range of 300 to 400C. I 1 318 Fifteenth Conference on Clays and Clay Minerals Table 4.--Surface Areas (S, msg-i) Accessible to Nj and HaO Molecules and Corresponding Percentage Void Volumes {? %) Fiber S(HaO) s(N2) s(HaO) Arizona 17.6 18.4 1.82 ____ J.M. 13.4 15.0 4.10 3.7 Russian 21.0 13.8 4.61 3.45 Corsica 17 -- 3.27 5.58 Cassiar 14.3 12.3 2.83 3.12 Advocato 31.2 23.6 4.90 5.97 | Coalinga 53.4 94.4 19.30 22.0 The general agreement is rather good but the most amazing discrepancy is that observed for Coalinga. It may be easily explained by taking into account that polar molecules have access to void spaces not available to non-polar molecules (Escard, 1950). Consider, in Fig. 2, the spaces (in black) not accessible to N2 for energetic reasons, the cohesion energy of two close surfaces being higher than the van der Waals energy involved in the nitrogen adsorption process. In equation (1), the contribution of this region to the surface area of the external pore is given by 6oJtL. For polar molecules, this negative term cancels and S* becomes 2/J5(l -- r2jBz)8. The computation of the specific surface area from equation (7) with this new S* value has been performed using the characteristic textural parameters determined for Coalinga. As a result, the theoretical specific surface area increases from 51.3 m2 to 89.5 m2/g, the latter value being in good agreement with the area available to H0 (Table 4). It may be therefore concluded: (1) that outgassing chrysotile fibers at 300C removes hydration water with the consequence that the "nitrogen" surface area increases; (2) that this increase is a function of the presence of amorphous materials; and (3) that there are regions which are not accessible to nitrogen for energetic reasons. In Table 4, the percentage void volumes obtained from nitrogen and water adsorption isotherms are compared. Except for Coalinga where tho quoted values are rather high, the other ones are very low, in good agreement with data given by Pundsack (1961). Void volume percentages in the range of a few per cent strongly suggest the presence of amorphous materials filling the porous system to variable extent while the values obtained for Coalinga are close to that estimated by Whittaker (1957) from electron micrographs, assuming emptied porous spaces. CONCLUSIONS The systematic study of nitrogen and water desorption isotherms applied to various ehrysotiles leads to a model that fits reasonably well the observa tions made with other techniques such as electron microscopy and X-ray diffraction. I 1 al > Clay Minerals BLE TO Ma AND HrO Void Volumes (e, %) (N*) 1.82 1.10 1.61 ..27 -.83 .90 '.30 (HsO) ____ 3.7 3.45 5.58 3.12 5.97 22.0 sb amazing discrepancy is i.ed by taking into account >ot available to non-polar he spaces (in black) not fion energy of two close v involved in the nitrogen ion of this region to the For polar molecules, this "2)S. The computation of is new 8* value has been ametors determined for face area increases from agreement with the area mug chrysotile fibers at mee that the "nitrogen" action of the presence of which, are not accessible from nitrogen and water ilajga where the quoted i hi good agreement with l.ii^ges in the range of a hSissim&terials filling the Signed for Coalinga are .^electron micrographs. applied /0ie observa- X-ray Surface Properties and Texture of Cheysoth.es 319 The ideal arrangement- of individual fibers in hexagonal close packing, with emptied internal and external pores, allows one to compute specific surface area, pore-size distribution function, and percentage void volume in good agreement with those observed in one case only, i.e. for Coalinga. It is interesting to point out that the deposit in which tlfis fiber is found is quite remarkable since the chrysotile mineral is not interstratificd with serpentine, as usual. Of course, this suggests a peculiar formation process that might be responsible for the different habit of Coalinga fibers. .For the other samples, the internal and probably also the external pores are filled to an appreciable extent by materials with chemical compositions similar to the bulk. From the percentage void volumes, t-ho degree of filling is estimated to be of the order of magnitude of 50%, or higher. Tho pore-size distribution function does not permit one to compute this figure. The presence of one single maximum in this function suggests that the fiber distribution with respect to the outside diameter is nearly gaussian and that the radius of the internal cylindrical pore is either roughly proportional to the outside radius, or almost constant. The most probable outside diameter can be derived from the most probable apparent pore diameter corresponding to the maximum observed in the poresize distribution function. The outside diameters obtained in that way are in good agreement with those derived directly from electron micrographs. The degree of filling by amorphous materials and the outside diameter appear to he characteristic of each chrysotile deposit. For samples with a high degree of filling, the hydration water cannot be driven off, unless they are pretreated between 300 and 400C. ACKNOWLEDGMENTS One of us (M. della Faille) is indebted to Etemit S.A. for the financial f. support which has made this study possible. Discussions with Dr. Pundsack from Johns-Mauville Research Center have been very helpful. We want to acknowledge also the part taken by Drs. Marechal and Cahen from Labofina S.A. in the realization of the continuousflow method used for recording the pore-size distribution functions. Many % thanks are due also to Professor Meinguet-, Head of the Computer Center k of the University, for the calculations performed on theoretical models. REFERENCES Bates, T. F. (1959) Morphology and crystal chemistry of 1:1 layer lattice silicates: Anier. Min. 44, 78-114. Bates, T. F., and Comer, J. J. (1959) Further observations on the morphology of chrysotile and haUoysite: Clays and Olay Minerals, Proc. 6th Conf., Pergamon v Press, New York, 237--tS. i' Bates. T. F., Sand, T.. B., and Mink, J. F. (1950) Tubular crystal of chrysotile asbestos: fr Science 7, 512-13. VA' /.VVAIWAK*.VV | TtTv 3-20 Fifteenth Conference on Clays and Clay Minerals C.una, B., Marec-hal. J., della Faille, M,, and Frieiat, J. ,T. (1965) Pore-size distri bution by a rapid, continuous-[low method: Anal. Chem. 37, 133--7. Cranston, B*. IV., and Inkley, F. A. (1057) The determination of pore structure from nitrogen adsorption isotherms: Ado. in Catalysis 9, 143-6, Academic Press, New York. della Faille, 31.. De Kimpe, C., and Fripiat, J. J. (1966) Influence de la deshydroxyla- tion sur la morphologic et la texture des chrysotiles: Silicates Inttusimels 31, 460-73. Escard, J. (1950) Influence de la d&hydratation progressive sur i'oire de surface d.' -( montmori!Ionites: Jour. Chimie Physique 47, 113--17. j) Fankwchen, I., and Schneider, M. (1944) Low angle X-ray scattering from ehrysotiie: Jour. A-mer. Chem. Sac. 66. 500-1. Healey, F. H., and Young, G. L. (1954) Tho surface properties of chrysotile asbestos: J:]- Jour. Phys. Chem. 58, 885-6. Maser, 3L, Bice, B. V., and Ki.va, H. P. (1960) Chrysotile morphology: A filer. Min. 45,680-8. '^ ? Noll, \V., and. Kircher, H. (1950) Zur Morphologie des Chrysotil-Asbesten: 3faturwiss. 37, 640-1. ' Noll, W., and Kxroher, H. (1951) L'ber die Morphologie veu Asbesten und ihren Zusammenhang mit dor Kristal-Struktur: Neues Jahrh. Min. Mh. 10, 219--10. Noll, W,, and Kircher, H. (1952) Veranderung von Cbrysotil-Asbest im Eleotronen- mikroskop: Xaturwiss. 39, 188-91. Pundsack, F. L. (1956) Tho density and structure of chrysotile asbestos: Jour. Phys. Chem. 60, 361-4. PtrarosACK, F. L. (1961) The pore structure of chrysotile asbestos: Jour. Phys. Chem. 65, 30-3. Turksvich. J., and Hellieb, J. (1948) Electron microsoopy on colloidal systems: Anal. Cham. 21, 475-85. Whittaker, E. J. W. (1957) The structure of chrysotile. V. Diffuse reflections and fibers texture: Acta Cryst. 10, 149. Young, G. J,, and Healey, F. H. (1954) The physical structure of asbestos: Jour. Phys. Chem. 58, 881-4. ' Small Bentoni ranging| spacing! XaPO;!,f particlebecame) between the eqi| tration.' tration intorpa concern The scatfa betwi