THE AMERICAN
MI\TERALOGIST, VOL 50, JULY AUGUST, 1965
DENSITY OF BULK CHRYSOTILE AND MASSIVE
SERPE}iTINE
CuenlBs W. HuccrNs aNn H. R. Snnr-r, Bureau oJ Mines,
Iorris, Tennessee
Aesrnacr
Density measurements r,veremade on trventy-three bulk specimens of chrysotile and
seven massive serpentine samples, both by using mercury and by coating the samples l'ith
paraffin and immersing in water Arizona chrysotile had densities of 219 to 2.25 gmicc,
whereas canadian chrysotile gave slightly higher values that varied from 2.34 to 2.39
mgfcc. Arizona massive serpentine blocks had less porosity; the density ranged from 2.40
to 2.44 gm/cc. From published data obtained by"electron microscopy, a single fiber of
chrysotile has an average outside diameter of 340 A and an inside diameter of 80 A. The
theoreticai density of chrysotiie, composed of parallel bundles of hollou,cylinders of these
d i m e n s i o n s ,w a s c a l c u l a t e dt o h e 2 . 1 9 g m / c c
INtnonuctloN
Turkevich and Hillier (1919)first pubiished electronmicrographsindicating that chrysotilesinglefibersmight be hollow tubes.Their work was
verified by electronmicroscopicstudiesof chrysotileb1' Bates (1948)and
Bateset al. (t95O).The hollow tube theory gainedsupport by the work of
reNolI el ol. (195S)when he synthesizedseveralsubslancesthat closel-vwhen
was
evident
structure
tubular
sembledchr,vsotileand showedthat
the materiai was viewed with an electronmicroscope.Examination of Dr.
Noll's cobalt chrysotiie by Huggins (1962) also indicated tubular morphologyZussman,etal. (1957),Huggins (1959),and othershave supported
tubular morphology as electron micrographsof dispersedchrysotile appeared to show hollow cylinders for the ultimate structure. Maset et al.
(1960) succeededin cutting crosssectionsof chrysotile and showedseveral crosssectionsthat were predominately doughnut shaped.Whittaker
(1954, 1955a,b, c) contributed to the tubular morphology idea by comparing the :r-ray diffraction efiects expected from tubular or c1'lindrical
Iatticeswith thoseobtained experimentally.
ConverselyPundsack (1956) and Kalousek and Nluttart (1957) concluded that the density of compact bundles of chrysotile fibers was incompatible with either tubular or solid cylindrical shapes.Kalousek and
Muttart did not rule out tubular structure entirell' since they did find a
void volume in Globe, Lrizona, chrysotile ol 12.5 per cent. Ilowever,
their measurementsgenerallyrevealeda porosity far too low compared
to that required for massive specimens consisting of tubular structure.
Pundsack (1961) also presentedexperimentaldata which indicated that
if any pores existed in chrysotile they were less than 6O A in diameter.
More recently, Whittaker (1963), and Bates (1958),have hypothesized
that chrysotilefibers are filled with amorphousor partially oriented ma-
DENSITY OF SDRPENTI N T:,
terial and that this accountsfor the lack of voids in the fibers.Very surprising is the fact that nearly all of Pundsack's experimental bulk density
determination gave values very closeto the theoretical 2.56 gm/cc, the
Iatter being derived from r-ray difiraction.
Severalother literature sourceshave indicated a much lower value for
the density. Bowles (1955) gavea specificgravity oI 2.22for pure chrysotile and said that higher values were always obtained because of the
presenceof impurities. Berger (1963) gave the density of cleanedchrysotileas2.22, and Dana (1892)gave a specificgravity of 2.219.
The present investigation was initiated in an attempt to help resolve
some of the conflicting data and also to see if the density determination
would, or would not, support the tubular structure theory. It was reasonedthat a valid bulk density value should be obtained if large blocks of
choicespecimenswere used and correctionsmade for any contamination.
Merenrer,s
Many samplesof chrysotile were obtained from Canada; the best of
these were used for density measurements. Chrysotile from Africa and
another thought to be from New York were receivedfrom the Smithsonian Institution. Arizona chrysotile blocks of the highest quality, obtained from D. W. Jacquays Company, Phoenix, Arizona, were nearly
translucent and free of admixed minerals. All the chrysotile sampleswere
nearly free of cracks and admixed mineral impurities, except for 0 to 20
per cent magnetite in the blocks of Canadian chrysotile. All of the fiber
blocks except one exceededf; of an inch in cross-sectionand one inch in
length. The massiveserpentineblocks were from Arizona and were the
highest purity we could obtain.
DBNsrry MnasunpunNrs
The blocks of chrysotile were dried at 130oC. for 36 hours to remove
the sorbedwater and then weighed.Most were then immersedin mercury
(Figs. 1, 2) and the density determined. The crank on the apparatus,
(Fig. 1) was used to raise the tank until the block of chrysotile was completely immersed in the mercury and the tip of the pointer just touched
the top of the mercury (Fig. 2, center). The following equation was then
usedto calculatethe density of the block from the experimentaldata:
(w)(d)
.
os:1a7ap
where
\[':weight of sample in grams
d:density of mercury at its temperature during measurement
1060
C. W. HUGGINSAND H. R. SI]ELL
F:force to immerse, which equals weight without sample, ninus
weight with sampie.
After determination of density in mercury, the blocks were dipped in
molten paraffin until a smooth, impervious coating resulted.The density
of the paraffin-coatedblocks was determined in water, and a correction
was made for the paraffin. The measured densitl' of the paraffin was
0.904 gm/cc.
Theoretical.Figure 3 is a drawing of a cross-sectionof a bundie of fibers
predicted on perfect cylinders; the white areas are the void spaces.The
absolutedensity of the solid portion may be calculatedif one knows the
diameter (inner and outer) of the fibers:
D.:2j:
("{;)
"
Frc. 1. Apparatus for the determination of density by immersion in mercury.
Scale: 1 inch in photograph:4 inches on actual apparatirs'
DENSITY OF SERPENTINE
lrc.
1061
2. Close up of mercury container (Fig. 1) showing relationship of pointer, sample
h o l d e r . a n d r n e r c u r y . Dimensions of mercury container are 41"X5"X4f;".
F-ro. 3. Cross sectional vierv of chrysotile fiber bundle in hexagonal close packing.
C. W. IIUGGINS AND H. R. SHELL
where:
Du: absolutedensity of solid
D*:observed density of fiber block
h:outer radius
rz: inner radius
The model in Fig. 3 is idealizedand in nature irregularitieswould be
expectedin the diametersand packing arrangement.The electronmicrographshave shown the fibers to vary in outsidediameterfrom 150 400 ^i.
The percentageof volume taken up by the voids between rigid solid
cylindersof equal diameter is a constantindependentof the number and
sizeof the cyiinders.If the ratio betweenouter and inner diametersof the
tubes were constant, the bulk density would be constant. Nonuniform
outsidediameterswould increasethe bulk density, if the "hole" diameter
remainedconstant or decreased.
If the averagediameters (inner and outer) and the theoreticaldensity
of the fibers is known, thesevalues can be substituted into the previous
equation and the bulk density can be calculated ver)/ closely. By r-ray
diffraction,the unit cell dimensionsof chrysotilehave beenaccuratelydetermined; the theoreticaldensity (i.e. of spaceoccupiedby solid) based
on these measurementsis 2.56 gmf cc. Kalousek and Muttart (1957)
and Maser et al. (1960) gave, as an average, the outside diameter of
chrysotile as 340 A and the inside diameter as 80 A. Using these two
values (340 A and 80 A) ln ttre above equation, the bulk density value
obtainedis 2.19gm/cc. The averagetaken for the insidediameter mav be
a little high, Pundsack (1961),and might possibly-be as low as 50 A. Using as the outsidediameter 340 A and the inside value of 5OA in solving
the equation, a bulk density oI 2.27 gm/cc was obtained. Therefore,one
would expect the density of the bulk fi.berblock to be between2.l9 and
2.27 provided that no impurities filled the void areas. Also the use of an
averageoutside diameter does not causetoo iarge a deviation from the
theoreticalmodel becausethe fibers measuredwere at or near 340 A.
C ouposrrroN
Crrn'lrrc.q.r.
The bulk specimenswere checked for impurities by chemical analy-sis,
optical methods, and r-ray diffraction. Except for magnetite in the
Canadian chrysotile, all of the fiber blocks were nearly free of foreign
minerals. The only other appreciablemineral detected was calcite, but
this was present in the Arizona massiveserpentineonly' Tabie 1 gives the
analysis of three samples.The first is typical high-purity Arizona chrvsotile (note the low iron content). The secondis Arizona massiveserpentine
blocks, with the Ca2+and COu being largely or completely present as
1063
DENSITI/ OF SERPENTINE
'I'esr,r
Constituent
1. Cnnlrrcar Annrysns
Arizona
chrysotiler
wt. per cent
Massive
serpentine2
wt. per cent
43.29
.23
.40
.06
.25
41.54
0
.04
0
.24
SiOz
ALOr
Fe:Os
FeO
CaO
40 23
.37
44
12.91
.02
.02
.02
3.00
39.09
0
.02
0
2.36
119
1 3. 3 3
0
0
0
100.09
100.03
Meo
KzO
NazO
TiOs
COr
HzOHzO+
MnO
Nio
CrzOr
r07
Total
Canadian
chrysotile3
(Thetford)
wt. per cent
41.41
.55
103
.72
0
41.52
0
0
.34
|.47
t2.90
.02
.03
0
99.99
I Chrysotile from D. W.
Jacquays Company, Phoenix, Arizona.
2 Massive serpentine from D. W.
Jacquays Company, Phoenix, Arizona.
3 Chrysotile from
Johns-Manville Company, Ltd , Box 1500, Asbestos, Quebec.
Analysis of chrysotiles by R. L. Craig, and of massive serpentine by E. E. Sutton, both
of Norris Metallurgy Research Laboratory, Bureau of Mines, Norris, Tenn.
T,ters 2. Dtlsny
Reference material
Aluminum, purel
Qtartz, clear
Large test ball
Small test ball
Steel block, machined
o-r'Re-prneNc-EMar-enrl.ls
Densityusing
Mercury
gmfcc
2.696
2.644
7 .665
,/.o.)/
7 .841
Density
in rvater
Probable true
gm/cc
2.708
2.653
7. 8 3 1
2 702
2.648
7 .6662
/ . ooo"
7.830
1 Alcoa spectrographic standard.
2 Based on values for volume of test balls furnished with air uvrconometer. by Beckman
Instruments. Pasadena. California.
C, W. HIGGINS ANI) IT. R. SHD,LL
calcite.The third is hand-pickedCanadian chrysotileand, at 100 magnification, was free of magnetite.
Rosur,rs ANDDrscussroN
Severalreferencernaterialswere selectedto check the newly designed
and built apparatusfor density measurementsin mercury. The densities
of three of these were also determinedin water. The results are given in
Table 2. Aluminum and quartz were chosenbecausethey have a density
Tlllr
Sample
No.
13
2
3
4
5
6
7
8
o
104
115
126
13
1A
l.)
16
1a
187
t9
20
2l
22
ZJ
3. DnNsrtv or Sner-rn artn Unseernn
Sample
\rt. gms
118.2870
7 8. 3 7r 5
40.6490
166.5995
162 8144
64.3568
77.3860
185 6378
3 2. 1 7 4 3
13.7156
41.6885
205.6065
167.6960
111.2180
9 4 . 1 757
41.9352
33 0130
84 5540
104.0106
60 8961
39.2374
70.1196
67.5001
Br-ocKs ol Crlnnsotrr-n
Bulk density,
D e n s i t y , u s i n g Density, using after all correcparaffin
tions (magnemercury, as
coating as
tite structural
measured,l
\\,t Der
cent
gm/cc
measured2
Fe, air buoyancy, gm/cc)
,
50r be(1
.
arra tar
1.18
1. 0 9
1.08
1. 0 5
1.08
1.09
1.10
1.04
1. 0 8
93
1.10
84
89
.80
89
.89
.80
.89
.83
.88
76
.91
69
2.20
2.22
2.24
222
2.24
2.25
2 -23
2.40
2.41
2.5r
2.40
2.40
2.49
2.43
2.42
2.66
2.19
2.22
2.25
2.25
2.23
2.25
2.25
2.23
2.23
2.39
2.25
239
2.40
2.51
2.4r
2.38
2.53
I
t1
2.50
2.43
2.54
2.M
267
2.r9
2.22
a
at
2.22
2.25
2.25
2.22
222
2.36
z.,tl
2.34
2.38
2.38
2.38
2.38
2.36
2.39
1 Density of sample excluding sorbed rvater
2 Density of sample excluding sorbed water and paraffin.
3 Samples 1 through 9 from D. W.
Jacquays Co., Phoenix, Arizona
a Sample No. 10, Aboutr.ille (?), New York, Smithsonian Cat. No. 91261.
5 Sample No. 11 from Kaapsche Hoop near Barberton E. Transval, Africa, Smithsonian
No. 91197.
6 Sample No. 12-17,
Johns-Manville Co , Ltd., Bor 1500, Asbestos, Quebec.
7 Sampies No 18-23, Bell Asbestos Mines, Ltd., Thetford Mines,
Quebec.
DENSITV OF SERPENTINIi
1065
Tenr,r 4. DnNsrry ot AnrzoN.q.
MassrvB SnnppNtrNr lvrrn Coxcclrnar,FRACTURES
Sample
No.
1
2
3
a
6
7
Samplel
wt., gms
117.4810
63.4768
7 8 .1 7 8 0
59.4592
55.5614
80.5733
101 0990
Density after
soaking in
waterg
gmf cc
Density b1'
mercuryB
lmmersl0n
Density by
paralnn
coating
gm/cc
2.56
2.56
2.43
2.41
2.43
2.40
2.41
2.44
2.M
2.42
2.41
2.43
2 -40
2.40
2.42
2.44
z.Jt
2.55
2.56
2.56
2.56
I Samples dried 1250 C. for 36 hours.
2 Samples soaked in warm water for 48 hours, and
density was determined.
3 Samples again dried and density determined in Hg, and
then by coating with paraliin
and immersing in water.
near the theoreticalvalue for chrysotile.The apparatus for determining
density by weighingin mercury was very easy to manipulate, and based
on the values found for referencematerialsiisted in Table 2, had an acc u r a c yo f * . 0 1g m / c c .
The results of the density measurementson the block samples are
given in Table 3. Considerablemagnetite was present in manv of the
blocksof Canadianchrysotile.The magnetitewas separatedmagnetically
after heating to destrol. the chrysotile structure and determined chemicaily. The density valuesobtained were correctedboth for the magnetite
and for the Fe2+and Fe3+in the structure. The correcteddensity for the
fiber blocksis given in the right hand column of Table 3.
The densitiesof Arizona and African blocks of chrysotile show that
void areas exist, and the density measurementsagreevery closely with
theoretical data calculated from previouslv published electron microscopic measurementsof the single fibers (Kalousek and Muttart, 1957;
Maser et al., t960). Canadianchrysotileis a little too high in densitv to be
completely tubular, but the data in Table 3 certainlt' indicate that over
half of the chrysotile must exist as hollow tubes.
The authors have examined massiveserpentinein the past and have
noted tubular structure. However, the present specimensof Arizona
massive(rock-like) serpentinewere largely filled with unknown material
(Fig. a).
AIso, the determined density (Table 4) confirms considerablefilling of
the tubes. Immersion in water gave density values approximately equal
to the theoreticalas determinedby r-ray diffraction. This meansthat the
void.?aces were filled with water during the 48 hour immersion.Chryso,
1066
C. W. HUGGINS AND IT ]1. SI]ELL
Frc.4. Examples of filled and unfilled single fibers: serpentine on left, chrysotile on
right. Both are from Arizona Dark line up the middle of the serpentine fiber indicates filiing; the light line up the micldle of the chrysotile fiber indicates a hollow tube. Magnification 54,000X.
tile from some locaiitiesmight exist with the void areasbeing all or partially filied. At least one-third of the possible void volume of Canadian
chrysotile,that we examined,must be filled with solid material; or alternately, most or even all of the possiblevoid volume could be filled with a
material or materials with a density considerablyless than 2.56 gm/cc.
CoNcr,usroxs
1. The density of bulk chrysotile is lower than that postulated from
r-ray diffraction data-in somechrl'sotilesbv an amount approximately
equal to the calculated void volume based on hollow tubes. The actual
valuesare in the range2.2 to 2.4 gmf cc.
2. The density values presentedin this report essentiailyagree with
those given by earlf investigators (Berger, 1963; Bowles, 19551Dana,
1892),but disagreeand are irreconcilablewith some more recently pub-
D]'',NSIT Y OF SI,:.RPF,NTI N ]'
r067
lished data (Pundsack, 1956). The latter allowed only a solid structure
without holes or spacesbetweenfibers,and has been the subject of considerablecontroversv.
3. The density of both bulk chrysotile and massive serpentineindicatesa hollow-tube or partialiv filled holiow-tube structure.
A"**o-wLEDGMENT
The authors are grateful to the following companiesand institution for
bulk specimensof chrysotile: Bell AsbestosMines, Ltd., Thetford, Quebec; Johns-Manville Asbestos,Ltd., Asbestos,QuebeclD. W. Jacquays
Company, Phoenix, Lrizona; and to the Smithsonian Institution for samples numbered91197 and9l26t.
RrlrnnNcns
'I'.
Brrns,
F-. (1948) Electron microscopy of kaolin minerals. Geol.Soc. Am. Bull.59, 1310.
--L. B. Sartns eNo J. F. Mwr (1950) Tubular crystals of chrysotile asbestos.Science
lll,512.
--(1958) Selected electron micrographs of clays and other fine grained minerals. Penn.
State Univ. ColI Mineral Ind. Circ. Sl
Boncen, HeNs (Trans. from German by Ralph E. Oesper) (1963) Asbestos Funilamentals.
Chem. Pub. Co., Inc., N. Y.
Bowr.rs, Or.ivrn (1955) The asbestosindustry. U. S Bur. Mines Bull.552,
DeNr, Eowenr S. (1892) Q System of Mineralogy. John Wiley & Sons, Inc., N. Y
HuccrNs, Cnanlrs W. (1959) Electron micrographs of asbestiform minerals. U. S. Bur.
M i n e s R ep l . I n r c s l .5 5 5 1 .
--(1962) Electron micrographs of some unusual inorganic fibers [/. S. Bur. Mines
Rept. Int:est.6O2O
Ka.r-ousrx, G. L. eNo L. E Murr,mr (1957) Studies on the chrysotile and antigorite components of serpentine. Am ,ilIineral.42r 1.
Mason, M., R V Rrcn eNn H P Kr.uc (1960) Chrysotiie morphology. Am. Mineral.45,
680.
Nor.r., W., H. Krncsrn eNl W. Svr:r:nrz (1958) Ein u'eiteres Solenosilikat: KobaltchrysoLil. N,t I u ru'i ssensha.[tcn ..
PuNose,cr, F. L. (1956) The properties of asbestosII. The density and structure of chrysotlle. Jour Pkys. Chem.60, 361.
--(1961) The pore structure of chrysotile asbestos.f our. Phys. Chem.65,30.
Tunxnvrcn, J. aNn J. Hrr.r.rrn (1949) Electron microscopl' of colloidal systems. Anal.
Chem.2l, 475.
lVrrrrrarrn, E.J. W. (1954) The difiraction of X rays by a cylindrical latticeL Acta Cryst.
7,827.
-(tO55a) The difiraction of r-rays by a cylindrical lattice II. Acta Cryst ,8, 261
--(1955b) 'Ihe difiraction of r-rays by a cylindrical lattice III. Acta Cryst.8,265.
-(1955c) The diffraction of r-rays by a cylindrical lattice IY. Acta Cryst.8,726.
--(1963) Chrysotile fibers-fiiled or hollow tubes? Chern €r Eng. News, Sept.30,34.
ZussuaN, J. aur G. W BnrNlr.ry (1957) Electron difiraction studies of serpentine minerals
Am. Mi.neral.42, 133.
ilIanuscript recehed, December8, 1964; accepted.
lor publication, March 21, 1965.