Mineralogy Exercise 2: Trace Constituents in Minerals
The elements listed in the chemical formulas of minerals aren’t the only ones found in them, as you know from Homework
1
and our class discussions: any cation of similar Ionic Radius and
Valence
may substitute for a major mineral-forming cation, as it will coordinate similarly with oxygens, and both “fit” into available structural sites in minerals, and provide charge balance for the mineral. Such substitutions are called Camouflage substitutions, meaning that during the crystallization of the mineral no discrimination occurs between the major elements and these minor species.
However, camouflage substitutions are rare: most of the time cations are either larger or smaller than is ideal for mineral sites, and often they are of the wrong valence. Yet, we find significant, if small abundances of such “poorly fitting” elements in any mineral that we chemically analyze. Where are these species found?
A rather labor-intensive way to assess where in a mineral a given element might be was attempted by Onuma and coworkers in Japan in the 1960’s and 1970’s. The specific interest of Onuma was in how low abundance Trace
Element
sdistributed between precipitating minerals and melt in natural igneous systems. So, he (actually, his students!!) would collect 10+ kg samples of volcanic rocks from Japanese volcanoes that contained large and abundant crystals (a texture called Porphyritic), crush the samples, and separate out several grams of each of the Phenocryst minerals and of crystal-free rock matrix. They would then chemically analyze each of the mineral and matrix splits for a large variety of major and trace elements, and assess the affinities of each element based on a parameter called a Mineral/
Matrix
Partition Coefficient , which was just the ratio of the measured concentration of each element in the minerals versus its concentration in the matrix.
A)What we’re going to do is interpretive part of Onuma’s work. Below I have given you concentration data for several minerals and their associated matrix. I want you to calculate mineral/matrix partition coefficients for each element in each mineral, and tabulate this data for yourself.
ii) Compare the same elements in different minerals: are the coefficients similar? Why or why not?
Ionic radius |
Rock 1 |
Rock 2 |
|||||||||||||||||||||||||||
angstroms |
Olivine |
CPX |
Plag |
||||||||||||||||||||||||||
Li, ppm |
+1 |
0.7 |
1.1 |
3.7 |
1.4 |
5 |
5.6 |
||||||||||||||||||||||
Na, % wt. |
1.02 |
0.00298 |
1.7 |
6.32 |
2.82 |
2.33 |
|||||||||||||||||||||||
K, % wt |
1.38 |
n.a. |
0.00075 |
0. |
50 |
1.67 |
|||||||||||||||||||||||
Rb, ppm |
1. |
49 |
0.79 |
49.6 |
|||||||||||||||||||||||||
Mg, % wt |
+2 |
0.72 |
23.6 |
2.95 |
8.63 |
4.08 |
|||||||||||||||||||||||
Co, ppm |
0.74 5 |
202 |
39 |
61.2 |
54.7 |
||||||||||||||||||||||||
Ni, ppm |
2020 |
1 |
25 |
||||||||||||||||||||||||||
Zn, ppm |
0.75 |
73 |
29.8 |
72.3 |
|||||||||||||||||||||||||
Fe2+, % wt |
0.78 |
6.01 |
5.65 |
4.42 |
4.62 |
||||||||||||||||||||||||
Mn, %wt. |
0.8 |
0.1 |
0.166 |
0.14 |
0.0058 |
0.11 |
|||||||||||||||||||||||
Ca, % wt |
0.06 |
6.54 |
13.5 |
6.19 |
5.55 |
||||||||||||||||||||||||
Sr, ppm |
1.13 |
59.3 |
1600 |
543 |
|||||||||||||||||||||||||
Ba, ppm |
1.36 |
145 |
484 |
||||||||||||||||||||||||||
B, ppm |
+3 |
0.25 |
0.112 |
10.1 |
0.23 |
0.24 |
|||||||||||||||||||||||
Al, % wt |
0.53 |
0.105 |
9.06 |
4.65 |
13.8 |
8.46 |
|||||||||||||||||||||||
Cr, ppm |
0.615 |
565 |
134 |
||||||||||||||||||||||||||
V, ppm |
0.64 |
11.4 |
381 |
126 |
100 |
||||||||||||||||||||||||
Sc, ppm |
6.95 |
17.1 |
|||||||||||||||||||||||||||
In, ppm |
0.052 |
0.086 |
|||||||||||||||||||||||||||
Lu, ppm |
0.861 |
0.00694 |
0.4 |
0.316 |
0.332 |
||||||||||||||||||||||||
Yb, ppm |
0.868 |
0.0266 |
3.05 |
2.34 |
0.037 |
||||||||||||||||||||||||
Tb, ppm |
0.923 |
0.00376 |
0.727 |
0.83 2 |
0.022 |
0.856 |
|||||||||||||||||||||||
Eu, ppm |
+3, or +2 |
.947* |
0.0061 |
1.34 |
1.62 |
0.68 |
2.15 |
||||||||||||||||||||||
Sm, ppm |
0.958 |
0.0089 |
3.29 |
6.11 |
|||||||||||||||||||||||||
La, ppm |
1.045 |
0.009 |
3.42 |
2.85 |
4.57 |
33.8 |
|||||||||||||||||||||||
Si, %wt |
+4 |
22.8 |
24.4 |
21.8 |
26.7 |
||||||||||||||||||||||||
Ti, %wt |
0.605 |
0.001 |
0.749 |
0.953 |
n.a.: unable to analyze.
b) I have also provided a list of the “octahedral” ionic radii for each of the elements analyzed. (admittedly, these elements all can’t coordinate octahedrally with oxygen, but using radii for a specific coordination number facilitates comparison, and anyway, it’s what Onuma did!). An “Onuma Diagram is a plot of the natural log of the partition coefficient for elements versus the ionic radii of the elements, and that’s what we’re going to do. Put Ionic Radii on the “x” axis, and plot Partition Coefficients on the “y” axis – you can either calculate the natural log (ln x) of each value, or you can simply do this in Excel and set it to do the “y” axis on a log scale. Plot together only the different elemental partition coefficients for a specific mineral (in other words, plot all the olivine/matrix partition coefficients on one graph, then put all the pyroxene/matrix coefficients on a second graph, etc. You’ll end up with a graph for every mineral). Label the points with the symbols for the elements they represent.
c) The table above also lists the common valence states of the elements we’re working on. Label the elements with their valences.
d) What I want you to do now is a modified “connect the dots”:
i) On each graph identify the data for all plotted elements of similar valence
ii) connect elements of similar valence on each graph via a smooth curve. What you should end up with is a graph with a series of curves on it for +1, +2, +3, and +4 valence elements.
Bring your plots to class with you for the next session, and we’ll go over this (if you don’t feel confident to “connect the dots”, you can wait and do that after I show it to y’all! After we talk about it, we’re going to answer the questions below:
1) Where do trace elements “fit” in mineral structures?
2) is ionic radius or valence more important in determining the “fit”?
3) Do some minerals more readily accept trace constituents than others? If so, why do think this is the case?
Mineralogy Exercise 2: Trace Constituents in Minerals
The elements listed in the chemical formulas of minerals aren’t the only ones
found in them, as you know from Homework 1 and our class discussions: any cation of
similar Ionic Radius and Valence may substitute for a major mineral-forming cation, as
it will coordinate similarly with oxygens, and both “fit” into available structural sites in
minerals, and provide charge balance for the mineral. Such substitutions are called
Camouflage substitutions, meaning that during the crystallization of the mineral no
discrimination occurs between the major elements and these minor species.
However, camouflage substitutions are rare: most of the time cations are either
larger or smaller than is ideal for mineral sites, and often they are of the wrong valence.
Yet, we find significant, if small abundances of such “poorly fitting” elements in any
mineral that we chemically analyze. Where are these species found?
A rather labor-intensive way to assess where in a mineral a given element might
be was attempted by Onuma and coworkers in Japan in the 1960’s and 1970’s. The
specific interest of Onuma was in how low abundance Trace Elements distributed
between precipitating minerals and melt in natural igneous systems. So, he (actually,
his students!!) would collect 10+ kg samples of volcanic rocks from Japanese volcanoes
that contained large and abundant crystals (a texture called Porphyritic), crush the
samples, and separate out several grams of each of the Phenocryst minerals and of
crystal-free rock matrix. They would then chemically analyze each of the mineral and
matrix splits for a large variety of major and trace elements, and assess the affinities of
each element based on a parameter called a Mineral/Matrix Partition Coefficient ,
which was just the ratio of the measured concentration of each element in the minerals
versus its concentration in the matrix.
A) What we’re going to do is interpretive part of Onuma’s work. Below I have given you
concentration data for several minerals and their associated matrix. I want you to
calculate mineral/matrix partition coefficients for each element in each mineral, and
tabulate this data for yourself.
ii) Compare the same elements in different minerals: are the coefficients similar? Why
or why not?
Element
Li, ppm
Valence
+1
Ionic
radius
Rock 1
Rock 2
angstroms
Olivine
Matrix
CPX
Plag
Matrix
0.74
1.1
3.7
1.45
1.1
5.6
Na, % wt.
+1
1.02
0.00298
1.78
6.32
2.82
2.33
K, % wt
+1
1.38
n.a.
n.a.
0.00075
0.505
1.67
Rb, ppm
+1
1.49
n.a.
n.a.
n.a.
0.79
49.6
Mg, % wt
+2
0.72
23.6
2.95
8.63
n.a.
4.08
Co, ppm
+2
0.745
202
39.1
61.2
n.a.
54.7
Ni, ppm
+2
0.7
2020
125
n.a.
n.a.
n.a.
Zn, ppm
+2
0.75
49
73
29.8
n.a.
72.3
Fe2+, % wt
+2
0.78
6.01
5.65
4.42
n.a.
4.62
Mn, %wt.
+2
0.83
0.177
0.166
0.14
0.0058
0.11
Ca, % wt
+2
1
0.0615
6.54
13.5
6.19
5.55
Sr, ppm
+2
1.13
n.a.
n.a.
59.3
1600
543
Ba, ppm
+2
1.36
n.a.
n.a.
1.7
145
484
B, ppm
+3
0.25
0.112
10.1
0.23
0.24
25
Al, % wt
+3
0.53
0.105
9.06
4.65
13.8
8.46
Cr, ppm
+3
0.615
n.a.
n.a.
565
n.a.
134
V, ppm
+3
0.64
11.4
381
126
n.a.
100
Sc, ppm
+3
0.745
6.95
39
50
n.a.
17.1
In, ppm
+3
0.8
0.052
0.086
0.1
n.a.
0.06
Lu, ppm
+3
0.861
0.00694
0.499
0.316
n.a.
0.332
Yb, ppm
+3
0.868
0.0266
3.05
2.34
0.037
2.33
Tb, ppm
+3
0.923
0.00376
0.727
0.832
0.022
0.856
Eu, ppm
+3, or +2
.947*
0.0061
1.34
1.62
0.686
2.15
Sm, ppm
+3
0.958
0.0089
3.29
5
0.24
6.11
La, ppm
+3
1.045
0.009
3.42
2.85
4.57
33.8
Si, %wt
+4
0.4
22.8
24.4
21.8
26.7
23.6
Ti, %wt
+4
0.605
0.001
0.68
0.749
n.a.
0.953
n.a.: unable to analyze.
b) I have also provided a list of the “octahedral” ionic radii for each of the elements
analyzed. (admittedly, these elements all can’t coordinate octahedrally with oxygen, but
using radii for a specific coordination number facilitates comparison, and anyway, it’s
what Onuma did!). An “Onuma Diagram is a plot of the natural log of the partition
coefficient for elements versus the ionic radii of the elements, and that’s what we’re
going to do. Put Ionic Radii on the “x” axis, and plot Partition Coefficients on the “y” axis
– you can either calculate the natural log (ln x) of each value, or you can simply do this
in Excel and set it to do the “y” axis on a log scale. Plot together only the different
elemental partition coefficients for a specific mineral (in other words, plot all the
olivine/matrix partition coefficients on one graph, then put all the pyroxene/matrix
coefficients on a second graph, etc. You’ll end up with a graph for every mineral). Label
the points with the symbols for the elements they represent.
c) The table above also lists the common valence states of the elements we’re working
on. Label the elements with their valences.
d) What I want you to do now is a modified “connect the dots”:
i) On each graph identify the data for all plotted elements of similar valence
ii) connect elements of similar valence on each graph via a smooth curve. What
you should end up with is a graph with a series of curves on it for +1, +2, +3, and +4
valence elements.
Bring your plots to class with you for the next session, and we’ll go over this (if
you don’t feel confident to “connect the dots”, you can wait and do that after I
show it to y’all! After we talk about it, we’re going to answer the questions below:
1) Where do trace elements “fit” in mineral structures?
2) is ionic radius or valence more important in determining the “fit”?
3) Do some minerals more readily accept trace constituents than others? If so,
why do think this is the case?
EARTH AND PLANETARY SCIENCE LETTERS 5 (1968) 47-51. NORTH-tlOLLAND PUBLISHINGCOMP., AMSTERDAM
TRACE ELEMENT PARTITION BETWEEN TWO PYROXENES
AND THE HOST LAVA
Naoki ONUMA
Department o1″Chemistry. The Unfi~ersity of Torero. Itongo. Bunkyo.ku. Tokyo. Japan
Hideo HIGUCHI
The Institute fi~r Atomie Energy, Rikk~.~ Uniw.rsity. Yokostlka, Japan
ttiroshi WAKITA
.R~!dioisotope Sekool, Japan A tom& Energy Research Institute,
Honkomagome, Bunkyo.ku. Tokyo. Japan
and
Hiroshi NAGASAWA
Fae lily o/Science, Gakushu#l Unive’rsiry, Mejiro, 7bkyo. Japan
Received 15 August 1968
Partition coefficients for +l, +2, +3 and +4 ,talent trace ion~ between alkaline olivine basalt lava and the coexistintgortbo- and clinopyroxene phenoerysts from Takashima, Ne,rth Kyushu, ~apan, have been determined by neutron
activation analysis. Substitution of trace elements in crystal lattice sites is proposed for the mechanism of trace element partition by examining the relation between measured paetition coefficients and ionic radius and ionic ehaxge.
1. INTRODUCTION
The distribution of trace elements, as a measure of
magma.tic differentiation, provides useful informatior,
for the evolution of igneous rocks. There is, however,
very litlle quantitative info~rmation about the effect of
ionic radius and charge on the trace element partition
coefficients between s,ilieate melt and crystallizing
minerals.
Masulda and Matsui [1 ] first estimated the partition
coefficients of rare earth elements between silicate
melt and crystalilizing rock from the rare earth abundances in the earth’s crust and in chondrites, assuming
that the earth’s crust had been differentiated from a
wholly molten chondritic earth by fractional crystallization, Recently, Schnetzler and Philpotts [2] determined the partition coefficients of barium and rare
earth elements be(ween rock-forming minerals and
their matrix by the mass spectrometric isotope dilution method.
One of us [31 has used lattice defect theory to
interpret the eftect o f ionic radius on the partition
coefficients of some trace univalent ic,ns between
ionic crystals and their melts.
We determined partition coefficients of trace elements between two kinds of pyroxene (augite and
bronzite) and their host lava, alkaline olivine basalt
or hawaiite from Takashima. North Kyushu, Japan.
2. SAMPLE
Tile basalt and purified augite samples were provided by Prof. H. Kuno; the bronzite sample was pre-
‘~q.ONUMAet at.
,-8
Table t
Chemic,~.tcompositions of alkali olivine basalt, au#tn and
broazite from Takashima
a
b
c
5iO2
AI203
Fe203
FeO
MgO
CaO
Na20
K20
H20 +
[120TiO2
P205
MnO
Cr203
50.27
15.99
3.59
5.94
6.77
7.77
3,43
1.97
1.52
0.35
]1.59
0.47
0.17
n.d.
46.56
8.79
3.08
5.68
14.31
18.91
0.89
0.03
0.62
0.02
1:25
0.04
0.19
n.d.
51.62
4.35
2.67
10.83
2%as; t: +2 valent inn:s; -: +3
v~t~Bt ions; (~) : +4 valent ions. Errors larger than 5% are
given by vertical bars. Ionic radfi ~,f Goldschmldt ~e used.
TRACE ELEMENTPART(T~ONBI’~TWEENTWO PYROXENESAND THE HOST LAVA
49
Table 2
Re,~tdtsof neutron activationanaly!;tsand t~Lei,’aleulaledpaJrtitioncoefficientsof the elements
Element
tonic
Radius
(A)
Na+
K+
Mn2+
Co2+
Sr2+
Ba2+
Concentration (pp:l~)
Partition coefficiem
Basalt
Augite
Bronlite
Angite
BTonzite
0.98
1.33
0.91
(I.82
1.27
1.43
23300 ± 500
16700 — 100
1100 ± 60
54.7 4.-2.9
543 4.-16
484 ± 14
6320 ± 70
7.5 —0.3
1400 ± 50
61.2 ± 2.0
59.3 4. 5.0
1.7 4. 0.5
1740 + 40
0.271 ± 0.007
(4.5 +0.2) X IO- 4
(.27 ± 0.08
1.12 +-0.07
0.109 +-0,010
(3.5 ± 1.01X 10″3
0,075 ± 0.002
Se3+
In3+
La3+
Ce3+
Ns3+
Sin3+
I’u3+
Gd3+
Tb3+
Tm3÷
Yb3+
Lu3+
0.83
0.92
1.22
1.18
1.15
(.13
i.12
1.11
1.09
1.01
1.00
0.99
17.1 4.-0.5
0.060 ± 0.003
33.8 ±4.4
66.1 ~ 1,7
30.4 4.2.9
6.40+-0.15
2.15±0.11
6.11 ±0.60
0.856 +-0.047
0.364 -~0.029
2.33 +-0.39
0.332±0.030
50.0 +-0.8
O.(00 +-0.004
2.85 +0.42
( 1,0 4-_0.9
t (.6 4. 1.6
5.00+-0.31
(.62-+0.08
5.00 ±0.31
0.832 ± 0.040
0.398 ± 0.036
2.3¢ 4’0,018
0.316 +-0.021
21.0 ± 1.5
2.92 4. 0.12
1.67 4″0.11
0.084 ± 0,017
0.(66 ± 0.016
0.382 4. 0,063
0.736-+0,021
0,753+-0.053
0.82 4″0,(0
0.97 +-0,07
1.09 ± 0.13
1.01 ±0.19
0.95+-0.11
1.23 ±0.10
U4+ (10)
Th4+ (101
1.05
1.10
0.707±0.006
4.57 +-0.10
0.0118± O.00f~4
0.0555 ~0.0015
1590 +–50
1(4 + 3
0.17 ± 0.02
0,090± 0.(115
0.049+-0.004
0,039± 0.004
0.25 ±0.05
O.O37 – 11.05
1.44 ± 0.09
2.08 +0.13
0,0026 ± 0.0003
0.014-+-0.02
0.0234.0.002
(/.046 4.-0.05
0.(07 4.0.028
0.1 | | ±0.018
0.017±0.0007
0.013 +-0.0005
(Calculated error is the stanctard deviation from countingstatistics.)
These character:istie features could be explained by
the crystal structure of pyro~¢ene: ([Jl in augite,
MIM((Si206, there are two erystal)ographically dif-,
ferenll sites. M1, the larger site occupied mainly b y
Ca2+ in eightfold coordinaticm and MII, the smalil~r
site o~::cupiedmainly by Mg2+ and Fe2″~”in sixfold coordina~tion [) 1]. (I)) in bronzite, M2 Si206, there
are also two sites, but both sites are almost equivalent and occupied exclusively by ~,lg-~+ and Fe2+ in
sixfc,kl coordination. In the augite structure, the
larger ions such as Sr 2+, Ba2*, rare earths3″~, U4+ and
Th 4÷ would occupy the MI :sitesby preference, and
the sn~aller ions such as Co2÷ and Se 3÷ would occupy
MII sites. In the bronzite s’trueture, ol’t the other hand,
these ~ons could only substitute in the small sites
available for Mg2+ and Fe2+, so that the partition coefficients of larger ions, rare ear ths3+., Ca2÷ and Na +
are about an order of magnitud~ smaller in the bronzite-basa)t system than those in the augite-basalt systern.
It is also seen in figs. 1 and 2 thac the parallel lines
indicating the partition coefficients of +1 to +4 valent
ions suggest( that the free-energy change required |k~r
balancing the difference of charge between the host
and the substituting trace ions is ap~)roximately constant for the substitution of equall~ charged trace ions
and is independent q3f the ionic radi us However, in
the augite.bal~alt system, the inversi,)n of the lines of
+1 and +3 va)ent ions for ionic radii smaller than about
I A is nol consistent wilh the abovl~ discussion. This
complicated partition pattern woul,:l be caused by tl~e
difference of chemical nature of ea,~h ,~lement, i,e. the
covalency, the crystal field effect, etc., and/or comple.~:
crystal strucl:ure of augite. This inw;rsiJon does not (,ccur in the case of bronzite basalt system.
The lattice-site substitution of tlraee elements could
also be checked by examining the elastic-body apprt~x.
i~nation postulated by Nagasawa [3], If the elasticbody approximation can be applied to the pyro×encbasalt system, there should be the ~relation
50
N. ONUMA et aL
10
,
,
..-,
g
\\\
“E
0ol
g
OOt
e
0001
0.02
J
&O01
0.2
04
0.6
0.8
1.0
0!:~6
008
010
012
( rrare e~’t h – rt4 s )2
12
~onie radius ( ~ l
Fig. 2. Partition coefficients versus ionic radiu!; for the bronz e-ba~ sysem Symbosaret esan-eas n f g 1 onic
radii of Coldsehmidt are u~ed.
llt DIz = A ( r u — ro)2 + B ,
004
Fig. 3. The partition cocfficientzr’of rare earth elements in the
bronzite-basalt system are plotted le~arithmieall~l against the
square of tile difference of ionic radilts between rare earths
and Mg. The Illor¢ precise ionic iadii of rare cart~s ~:,yTempleton and Dauben [ 121 and rMg = 0.71 .~ (after Zachariasen)
used instead of tl-tgsehy Goldsehmidt.
(1)
where D u is the partition coefficient of an element between pyroxene arld lhe host basalt; ru and r0 are the
ionic radii of the trace clement and the most favorable
ion to a given site, re*;pectively; A is a constar~t given by
elastic constants of the crystal, and B is the constant
introduced by the charge balance. Since the relation
holds only for trace element partition between a pure,
isotropic crystal and its melt in the strict sense, it is
difficult to apply eq. (1) to augite, which is a ,complex
mixed crystal of several components and complicated
crystal structure. Because bronzite has less conrplex
s~.ructure, it is feasi,ble to test eq. (1) for the bronzitebasalt ~ystem. Using ~lte values of Templeton and
Duuben [ 12] and 0.71.8, for ionic radii of rare earth
elements and Mg, respectively, we obtained an approximately straight line shown in fig. 3. This relation also
supports the lattice-si te substitution of rare earth elements in bronzite, However, because of variability of
the value~; of ic, nic radius and of the partition coefficients based on the low rare earth contents in bron-
zite, we cannot draw a defin:ite conclusion. The decrease o f partition coeffi¢,ient at larger ionic r;~dius
values is so rapid, that the linear relations between
In D and I/r or In D and atomic mmtber postti]ated
by Masuda and Matsui [ I ] does not hold in this ease.
The partition patterns of rare earth elements in
figs. I and 2 are closely similar to those shown by
Schnetzler and Philpotts [2], but differ for the lighter
rare earths in augite-basalt system, and for all ‘the absolute values. These discrepancies may be caused by
the difference in the temperature of crystallization,
chemical composition of the system, and possiibly by
minute inclusions of other minerals in pyroxene crystals.
The regularities iz) the partition patterns discussed
above indicate the following co.nelusion:
(I) Trace elements possibly occupy lattice sites of
pyroxene crystals, rather than occurring in separate
grains on or within the crystals.
(2) For trace ions of the same charge, the difference between ionic radius of the trace ion and that of
TRACE ELEMENT PARTITION BETWEEN TWO PYROXENES AND TIlE I IO”T LAVA
the host iion is the most important factor determining,
the partir.ion coefficients.
(3) Tllte effect o f ionic charge on (he 19~rtition coefficients seems to be independent o f the effect of
the ionic radius.
ACKNOWLEDGEI~IENTS
We th;mk Dr. K, T o m u l a for laboratory facilities
at Rikkyo University: Mr. H. Takahashi for technical
assistance: Prof. H, Kuuo for provision o f samples;
Profs. K. Kigoshi, Y. Ma~sui, M. Honda, S. Banno and
H. Hamaguciti for helpful discussions.
REFERENCES
[ II A. Ma!;uda and Y. Matsui. The difference in lanthanide
abundance pattern between the crust and tile chondrite
and its. possible meaning to the genesis of crust and mantle, Geoehim. Cosmochim. Acra 30 11966) 239.
12 ] C. C. Sehnt:tzler and J. A. Philpotts, partition coefficients ,of tile rare-earths and barium betweeu igneous
matrix material and rock-forming-mineral phenocrysts.
in the Proceedings of tire Symposium of the International
Associ:atinn of Geochemistry and Cosmodlemistry, May
1967.
131 H. Nagasawa, Trace element partition coeft~cient in ionic
crystals, Science 152 (1966) 767.
[41 H. Kuno. Aiuminkm au~dte and bronzite in alkali olivine ba,~’.dt from Taka.shima, bi6rth E~.vushu,Jap;in.
AdvaneilLg Frontiers Geol. Gcophys, 1964 ( 1965 ) 2115.
[5[ A. O. Brunfelt and E, Steinness, Instrumental neutr~mactivation analysis of the ‘standard n)cks’, Geochien,
Cosmochilrl. Acta 30 1t966) 921.
161 H. Hamagtkchi, K. Tomara. N. Onuma. I1 Iliguehl and
K. Suda, Determination of indium in rock samples by
neutron activation analysis, lapan AL~aly~t 16 ( 196″/i
1233.
[71 H, Nagasax~a and II. Wakita, Neutron activation znalysis of potassium in ultrabasic roeks~ Geochem. J. 1
11967) 149,
[81 1t. Higuchi, K. Tomara, H. Takahash[, N. Onuma and
FI. l-lanlaguchi, Simultaneous determ ination ~t’ strontium and barium in rock ,s:lmples hy neutzon activ;Jtion
analysis using a GelLi) detector, in prcparati~n,
191 g. Tomum, Ft. Higuchi, N. ~-liyajl, N. Onutma and H.
H~maguclti, Determination of mre-e=lrth elements in
rock samples by neutron activation znaiy~;is with lithiumdrifted germanium detector alter chemical group-separation, Anal. Chim. Aeta 41 (1968) 217.
IlOl II. Nagasawa and 1t. Wakita. Partition of uranium and
thorium between augite and host lavas. Geochim. Cosmot:him. Acta, in press.
[1]1 L. Brag,L,’,G. F. Claringbull and W. I-I. Tayl,or, Crystai
structure of minerals (G. Bell and SorLs. t,ondon, 1965)
p. 232.
[121 D. H. Templeton and C. H. Dauben, Lattice parameters
of some raze earth compounds and a set uf crystal radii,
J, Am, Chum, Soc. 76 11954) 5237,