American Minerulogist, Volume 59, pages 1123-1126, 1974
Determination
by
of Boronin Ghondrodite
lon MicroprobeMassAnalysis
Jeurs R. HrNrHonNe,
Hasler Research Center Applied Research Laborqtories,
Goletq.. Calif ornia 9 3 0 I 7
aNo Peur H. RrnsB
D ep art ment ol G eological Sciences,
Virginia Polytechnic Institute and State Uniuersity,
Blacksburg, V irginia 2406 I
Abstract
Boron concentrationsup to 0.71 atomic percent (the equivalent of 1.20 wt percent BzO")
have been recorded by ion microprobe mass analysis in chondrodites previously found to be
silicon-deficientby Jones,Ribbe, and Gibbs (1969'). The formulas [normalized to 2(Si + B)]
for the two most boron-rich chondrodites are:
gsrTiooreFrtt(OH)ottOooru
2Mgr saoFeo
saoBo
oooO,]'Mgo
66eMn6
662Ca6
oor[Sio
2Mgr grrFeooosMno
ort
srzBoorrOr]'Mgors6TioorrFtos(OH)o.zeOo
ooaCao
oor[Sio
The charge-balancing substitutions are thought to occur in the hexagonal closest-packedanion
array, probably as hydroxyl for oxygen. Only trace levels of monovalent cations (Li,Na,K)
were found in the mass spectra of these orthosilicates, and these are effectively charge-compensatedby trace levelsof trivalent octahedrally-coordinated
cations (Al,Cr).
National Museum #36081) had values of St as low
The humite minerals are hexagonal closest-packed as 0.928 and 0.953, respectively. Single crystal and
orthosilicates with the general formula nM2SiOa. X-ray powder patterns confirmed that both of these
Mr,Ti,(OH,F)r
r*Oro, where n : 1 for norbergite, were indeed chondrodites, and the cell volume of
n = 2 for chondrodite, n = 3 for humite, and n = Q #9 was observed to be significantly smaller (by
for clinohumite. and where x ( 1 and the octahe- 1.8 A3) than Jones' chondrodite #2 which contains
drally-coordinated M atoms are Mg, Fe, Mn, Ca, only one-third as many larger-cation substituents for
amdZn in decreasingorder of abundance.The crystal magnesium (see Fig. 1;cl Tables 1 and 3 in Jones
et at, 1.969). From these observations Jones el cl
chemistry of the humites was studied in detail by
R i b b e , G i b b s ,a n d J o n e s( 1 9 6 8 ) , J o n e s( 1 9 6 8 ) , a n d (1969,p.396) implied ". . . the presenceof another
tetrahedrally coordinated cation which is smaller than
Jones et aI (1969). Of 55 specimens analyzed by
conventional electron microprobe techniques,several Si." Careful microprobe analysisindicated that phosof the chondrodites were found to be significantly phorus is not present in detectableamounts in humite
minerals; thus the only alternative was boron, since
deficient in silicon.
beryllium is as large as silicon and aluminum is conFor stoichiometric humites, the ratio St : 2Si/
siderably larger and is, in any case, present in only
(Si
I)]M.n
1.000
should
be
and
M1i
ate,
+
[2n/(2n
respectively, the atomic proportions of tetrahedrally trace amounts.
This ion microprobe study determined that boron
coordinated Si and octahedrally coordinated cations,
Mri = Mg + Fe * Mn + Ca + Zn * Ti). Jones can indeed substitute for at least 6 percent of the
(1968) found that two of his chondrodites(#9 from
silicon atoms in chondrodites, although in the other
Franklin, New Jersey-U.S. National Museum humite minerals it has not yet been found in excess
of 0.7 percent.
#R14660; and # 10 from Amity, New York-U.S.
fntroduction
r123
tt24
73.O
MINERALOGICAL
o
alo
o
fa,
72.O
Y.
UJ
24o
.06
2g
.ol
(+.o3
=
f
J
o 7t.o
o'
o
70.o
o
o.2
o.4
RAT|O,(F+OH)/O
o.6
Frc. 1. The relation between the normalized cell volume
a x b x @*t/n) and (F + OH)/O in norbergite (solid
circles), cho'ndrodite (open circles), humite (solid tri
angles), clinohumite (open triangles), and forsterite
(squares). Numbers next to humite mineral data points
correspond to those in Table I of Jones et al (1969). The
solid line conects synthetic Mg,SiO, (data from Yoder and
Sahama, 1957), 3Mg"SiO,.MgF,, and Mg,SiOn.MgF, (data
from Van Valkenburg, 1961). Fo* data from Birle et al
(1968). Ratios indicate percent substitution of octahedral
cationsfor Mg.
Experimental Techniques
NOTES
tively high abundanceof boron (>100 parts per
million atomic [ppma]) in these chondrodites,an
integrationtime of only two secondswas adequateto
measurethe 1'B* isotope peak. The peak-to-background ratio for chondrodite#6, rhe samplewith
the lowest boron concentration,was greater than
100, and usingcountingstatisticsthe estimatederror
2o is 7.5 percentof the amountpresent,or 5 ppma.
For the most boron-rich specimens,2,ois 2 percentof
the amountpresent,or < 150 ppma.The boron data
were quantitatively reducedusing sensitivity factors
derived from a theoreticalionization model (Andersen and Hinthorne, 1973), thus eliminatingthe need
for boron calibrationstandards.
Analyses were also performed for other cations
which may occur in tetrahedralcoordination.Aluminum was the most abundantof theseelements,averagingabout 100 ppma with a rangein the 15 samples
from 75 to 200 ppma. Phosphorusconstitutedless
than 100 ppma and beryllium lessthan 20 ppma in
were alsomadefor
all samples.Quantitativeanalyses
the following elements:lithium (averagevalue, 13
ppma; range, 1.9-31), sodium (120; 16-530),
potassium( 100; 5-8 80), titanium ( 1150; 12-3700),
vanadium(16; 9-26), manganese(72O; 67-3400),
strontium(1.8; <0.1-7), yttrium (0.8; <O.l-2.7),
zirconium(2.7;0.a-8.7) and chromium(<130).
In addition,chlorinewas detectedin two of the samples, at the 40 ppma level in #24 and 8O ppma in
#25 (detectionlimit -20 ppma).
Discussion
The boron values reported in Table 1 were obTable 1 and Figure 2 summarizethe data for the
tained using the Ion Microprobe Mass Analyzer, 15 chondroditesanalyzedfor boron. The numberof
Iltue, at Applied ResearchLaboratories,Hasler Re- atomsper formula unit, normalizedto 2(Si + B),
searchCenter,Goleta,California (seeAndersenand was calculatedfor each sample using the Iltvr.q reHinthorne, 1972). A primary beam of 'uO- ions sults for boron and the electronmicroprobe analyses
accelerated
to 17 keV was usedto sputterthe sam- of Joneset aI (1969, Table 1). The *4 chargeon
ples. The negativelychargedbeam effectivelyelimi
titanium is assumedto be compensatedby oxygen
nated any electrostaticchargebuild-up on the sample substituting for fluorine and hydroxyl according
surface (Andersen,Roden, and Robinson, 1969). to the formula:
The beam diameter was approximately 20 pm and
Ti + 2(O)z: (Me,Fe) + 2(F,OH)
the beamcurrent6.0 nanoamps.
The samplesanalyzedin this study were grains Hydroxyl contents were calculatedin the manner
from the same electron microprobe mounts previ- suggested
by Joneset aI (1969, p. 394), assuming
ously preparedand analyzedby Jones (1968). The that the tetrahedral sites are filled with silicon and
grain mounts were carbon-coatedto facilitate sec- that eachspecimenis charge-balanced:
ondary ion collection.Using the light microscopeof
OH:0.4Mri-2Ti-F.
the Ivrul, it was possibleto avoid mineral and fluid
inclusions and microfracturesduring the analyses.
ffowever, since B3* replaces Sia*, there must be
Becauseof favorabledetectionlimits and the rela- either monovalent cations substituting for divalent
MINERALOGICAL
tl25
NOTES
TeeLB 1. Boron Content and Revised Cell Contents for 15 Chondrodites"
Number of atons
Specimen
number
I
1
6
9
10
11
I4
15
ao
22
23
z4
25
30
a
b
Boron
(ppna)
-t
"
per fornula
noroalized
uni!,
Si
Mg
to Z(li
t-D-
oHb
ll-
Stoic .c
1.002
1.009
0.975
1.012
0.928
1865
425
1530
80
7110
4,89
4.89
4.98
4.85
4.90
.026
.033
.064
.063
.138
.002
.008
.003
.008
.005
.000
.001
.002
.002
.002
.000
.002
.004
.003
.001
r,969
1.993
L . 9 74
L.999
1.880
.031
.007
.026
.001
.l-20
.000
.001
.009
.010
.016
r.. 03
L.25
1.51
L,23
1.88
0.93
0.72
0.49
0.73
0.11
l". 018
1.013
0.988
1.012
0.987
0.953
0.996
0.985
0. 986
0.994
6 2 75
310
1380
1180
990
4.81
4.84
4.79
4. 8r
4.76
.137
.141
.153
. 156
.159
.006
.0L2
.007
.010
.005
.002
.001
.001
.001
.002
.000
.000
.001
.000
.003
L.894
r,995
1.977
1.980
1.983
.106
.005
.o23
.020
.017
.014
.0r0
.071
,o44
.058
r.69
L.44
t.27
L.45
1.48
0.26
0.54
0.59
0.47
0.40
1.007
0.998
0,996
0,996
1,.002
0.995
0.991
0.998
0.965
0.985
640
100
830
t790
135
4.72
4,69
4.66
4.73
4.6L
.237
.25L
,275
.3r7
.353
.0r-5
.081
.015
.014
.089
.000
.004
.000
.003
.005
.001
.008
.001
.003
.0r.0
r,.989
1.998
1.986
L.969
1.998
.011
.oo2
.014
.031
.002
,023
.009
.o20
.O27
.001
0.98
1.31
1.06
r.47
1.38
0.98
0.68
0.89
0.51
0.65
1.001
0.992
1.005
0.981
0.989
specinen nu[Ders and data fot s. = 2.5 sillt--+ anil tot
Jones et af (7969, TabTe 7 -- rvlps pocunent'*00345).
js (8 + 2x) where x is the numbet of Ti atoms.
Hgdto*g7
content
" stoi"hio^"ttic
calculated
ratio
= 2.5(st
using
the nethod of
Mg, Fe, Mn, ca, zn, si' Ti and E fron
The nudbet of oxggens pet lotmuTa unit
Jones et al
(7969, p.
3g4):
0H = O.4MT. - 2Ti
+ B)/MTi.
metalsin the M sitesor further hydroxyl substitution
for oxygen in the hexagonal closest-packedanion
array. The anion substitutionis presumedlythe correct one, becausethe analysesof thesechondrodites
referred to above showed only trace concentrations
of monovalentcations(Li, Na, K). In any casethe
effect of thesewould be offset by the trace amounts
of trivalent cations (Al, Cr) also recorded in the
spectra.
To our knowledgethis is the only reported occurrence of boron substitution for silicon in orthosilicates.The levelsof boron observedin one norbergte
(Jones'#4) and,five humites(#3,4,5, 7, 9) were
less than 400 ppma, althoughtwo (#3, 4) of the
five clinohumites we analyzed contained 800 and
950 ppma,respectively.
This study also serves to indicate the absolute
analytical accuracy achievable with the electron
microprobeusing the proper standardsand analytical
techniques.Assuming that the tetrahedral sites in
theseminerals are completelyfilled with silicon plus
FIc. 2. Atomic percent boron determined by hvru.L plotted
against atomic percent silicon determined by electron microprobe (Jones et al, 1969) for 15 chondro'dites. The theoretical one-for-one substitution line is shown with parallel
lines marking -+0.05 atomic percentsilicon.
=
E O:5OO
z,
o
G
o
.D
s
o o.3oo
=
o
)o
il.40
ATOMIC% SILICON( EMX )
- F.
MINERALOGICAL
t126
boron (i.e., Si + B - 2.000), the errors in the
silicon determinations average approximately 0.5
percentof the silicon abundance.
Acknowledgments
The authors are grateful to Dr. Norris Jones for use of
his samples, many of which were originally obtained from
Professor Th. G. Sahama and the U.S. National Museum.
The National Science Foundation provided financial support for this project under grant 30864X to P. H. Ribbe and
G. V. Gibbs.
References
ANornsEN, C. A., eNo J. R. HrNrnonNe (1972) Ion microprobe mass analyzer.Science,1.75,853-860.
(1973) Thermodynamic approach to
AND the quantitative interpretation of sputtered ion mass
spectra.Anal. Chem. 45, 142l-1438.
H. J. RooeN, eNp C. F. RoerNsoN (19'69) Negative
NOTES
ion bombardment of insulators to alleviate surface
charge-up.I . AppI. Phys. 40, 3419-3420.
Brnre, J. D., G. V. Grsns, P. B. Moonr, .rNo J. V. StvrrfiI
(1968) Crystal structures of natural olivines. Am. Minerol. 53,807-824.
JoNps, N. W. (1968) Crystal Chemistry ol the Humite
Minerals. Ph.D. Thesis, Virginia Polytechnic Institute,
Blacksburg, Virginia.
P, H. Rrnun, lNo G. V. Gress (1969) Crystal
chemistry of the humite minerals. Am. Mineral. 54r 3914lt.
Rrnrr, P. H., G. V. Gnns, eNo N. W. JoNls (1968) Cation
and anion substitutions in the humite minerals. Mineral.
Mag. 36, 966-97 5.
Ver Vexr,Ngunc, A. (1961) Synthesis of the humites
nMg,SiOn'Mg(F,OH)". l. Res.Nat. Bur. Stand., A- Phys.
Chem. 65A.,415-428.
Yoorn, H. S., eNo Tn. G. S*rer"re (1957) Olivine X-ray
determinativecnrve.Am. Mineral. 42' 47 5-491.
Manuscript receioed,March 7, 1974; accepted
for publication,May 1, 1974.