WILLIAM H.
DENNEN
IMPURITIES IN QUARTZ
Abstract: The impurity content of quartzsubsti-
tutional elements plus foreign inclusionsis differ-
ent for different rocks and may be sufficiently
distinctive to allow correlation within the same
Introduction
Quartz is known to contain small but varia-
ble amounts of structurally substituted or
interstitial ions, but since most quartz is also
crowded with microscopic inclusions, pub-
lished analyses such as those found in Frondel
(1962, p. 142-153) are largely restricted to
coarsely crystalline material wherein foreign
matter is minimal. Quartz formed under
particular environmental conditions might be
anticipated to contain a reasonably constant
and characteristic level of foreign matter. Such
"dirty" quartz is the subject of the following
study, and all chemical concentrations re-
ported are understood to be the sum of the
substitutional elements plus inclusions below
a few microns in size.
The elements reported are those which can
be detected by emission spectrography of small
samples with reasonable precision. These are
magnesium, iron, aluminum, and titanium.
Calcium and lithium plus, in some cases,
manganese, sodium, and potassium are de-
tectable but with poor analytical precision.
Previous work on the chemistry of quartz
has shown that distinct chemical differences
exist between the quartz of different granites
(I. M. Bray, 1937, PhD thesis, Mass. Inst.
Technology; Stavrov, 1962), between the
quartz of a granite and its associated pegmatite
(Bray, 1937 [thesis]; Stavrov, 1962; Tatekawa,
1954), and within the same metamorphic rock
(E. C. Perry, 1963, PhD thesis, Mass. Inst.
Technology). In general, however, there is a
paucity of chemical data for contaminants in
rock-forming quartz.
A reconnaissance study of the rate and
amount of the blackening of quartz under
X-ray irradiation (C. J. Ritter, 1962, SM
thesis, Mass. Inst. Technology) suggests that
quartz from different environments may have
Geological Society of America Bulletin, v. 75, p. 241-246, 3 figs., March 1964
241
igneous or metamorphic rock mass and distinction
between different quartz-bearing rocks. An appli-
cation to provenance is suggested.
marked physical-chemical differences and that
different rocks of the same environment, e.g.
granites, may also be distinguished.
Acknowledgments
This work is part of a study of contaminants
in quartz supported by National Science
Foundation Grant NSF-G19743. The writer
has received valuable assistance from his stu-
dents in the Cabot Spectrographic Laboratory,
notably J. Mielke and E. Perry, and has been
freely given samples by many people.
Sample Preparation and Analysis
Samples for study were collected in the
field or obtained from rock collections in the
Department of Geology and Geophysics,
Massachusetts Institute of Technology. A
suite of Nova Scotian granites and samples
from the Idaho batholith were made available
by H. W. Fairbairn. A. Ramo contributed the
Dedham Granodiorite-Hoppin Quartzite con-
tact material.
Samples for analysis were crushed with a
hammer and anvil and sized to 40 +100
mesh, and enough quartz for a spectrochemical
analysis was selected under a 10 X binocular.
A preliminary picking against a black back-
ground facilitated the distinction between
quartz and feldspars. Final picking was done
under acetone against a white background.
Acetone has an index of refraction close to
that of quartz, so that inclusions larger than a
few microns may be readily seen in the interior
of grains. Acetone also washes the grain surface
and evaporates without leaving any con-
taminating residue. About 10 grains aggre-
gating 2-5 mg were selected for each determi-
nation and stored in gelatine capsules until
needed. No grinding or other potentially con-
taminating operations were performed.
Spectrochemical determinations of Mg, Fe,
242 W. H. DENNENIMPURITIES IN QUARTZ

Al, and Ti were made on all samples. Ca was of several different samples and standards
recorded as a rather erratic contaminant, and scattered throughout the analytical program.
Mn was occasionally found. The precision is not high. Coefficients of
Transfer of samples from gelatine capsule variation of +15 per cent for titanium and
to spectrographic electrode was made via a about 30 per cent for magnesium, iron, and
nylon funnel. The sample electrode was a pre- aluminum were found as representative of
arced one eighth of an inch "specpure"
graphite rod provided with a 1.5-mm-diameter TABLE 1. SPECTROCHEMICAL DATA: ELEMENT SOUGHT,
crater 3 mm deep. The counter-electrode was WAVE-LENGTH OF SPECTRAL LINE USED, AND
a three sixteenths of an inch "specpure" MINIMUM SENSITIVITY ATTAINED
pointed carbon rod. A Hilger large quartz
prism spectrograph was used for all determina- Element Line used Detection limit
tions. Samples were arced as the anode at 6
amps to completion (about 45 sec) and the Manganese 2794.8 A 10 ppm
spectral region 2800 to 5000 A was recorded Magnesium 2852.1 0.2
Iron 3020.6* 5
on Kodak Spectrum Analysis-1 plates. Plate Silicon 2970.3 Standard
calibration was accomplished by a "two-line" Aluminum 3092.7 10
procedure, and all optical densities were Titanium 3234.5 5
measured by use of a Hilger nonrecording Calcium 3968.5 5
microphotometer.
Reference standards were prepared by ad- * A weak unlisted line of silicon lies at about 3020.0 A
mixture of appropriate metal oxides into a base
of rather pure powdered quartz. An addition single determinations of these elements. The
plot provided the necessary corrections to program, however, entailed replication of de-
standard content. terminations which reduced the analytical
The ratio of the spectral line intensities of variation of the mean to at least one third of
analysis element to silicon was found to be these values.
reproducible under the conditions used, and The expectable variations in the chemical
working curves of this ratio versus concentra- content of quartz from a granite were in-
tion were constructed and used for all determi- vestigated by means of a nested sampling
nations. The spectral lines used and the ap- sequence involving analysis of 10 samples
proximate detection limits for the various taken from the same hand specimen, 24 samples
elements studied are given in Table 1. obtained at intersections of an orthogonal grid
The precision of the analytical method em- of 2-foot spacing (containing the hand speci-
ployed was assessed by replicate determinations men), 15 samples from intersections of a 100-

STOCK
100'GRID
2'GPlD
R- HAND SPECIMEN

Q.
*

25.5124 8 16 32 64 128 16 32 64 128 512 & 16 32 64 256

PPM Mg PPM Ti PPM Al POM Fe
Figure 1. Abundance and distribution of some chemical impurities in quartz from
a granite. Quincy stock, Quincy Granite, Massachusetts
 SHORT NOTES 243

foot grid (containing the 2-foot grid), and 20 but less marked, variability as indicated by
grab samples from a small granite stock (con- Table 2. The contrast in chemical content of
taining the 100-foot grid). quartz from different environments is strik-
The results of this study are shown in Figure ingly shown by a comparison of modal values
1. The chemical values for the impurity con- of the different groups when plotted in rose
tent of quartz appear to have a lognormal diagrams (Fig. 2).
distribution and are so plotted. It is immedi-
ately apparent from an inspection of Figure 1 TABLE 2. PRESENCE OF SPECTROGRAI>HICALLY DE-
that single samples or samples covering only TECTABLE AMOUNTS OF VARIOUS ELEMENTS IN QUARTZ
a small area cannot be taken as representative FROM DIFFERENT ROCK TYPES
of the rock mass, i.e. the impurity content of
the quartz of a given rock apparently has Rock type Per cent of samples in which found
some random variability in space.
Mn Mg Fe Al Ti Ca
Differences in Impurity Content of Quartz Granitic rocks 8 98 94 100 96 94
Between Major Rocl^ Types Rhyolites 0 89 56 100 78 100
Pegmatites 3 64 48 94 39 48
Quartz from 66 different granites, 9 rhyolites, High-grade meta-
30 pegmatites, 19 veins, 37 high-grade meta- morphic rocks 0 100 57 100 16 54
Veins
morphic rocks, and 20 clear and colored crystal- Crystal 3 70 35 10 30 55
quartz 0 85 46 92 8 15
line samples was examined. The most distinc-
tive single element is titanium which is almost
universally present at levels of approximately
45 ppm in the quartz of plutonic and ex- Differences in Impurity Content of Quartz
trusive igneous rocks, but is not found spectro- from Different Granitic Rocl(s
graphically in the quartz of most metamorphic The quartz of a number of different granitic
rocks, veins, pegmatites, and coarsely crystal- intrusive masses has been studied to determine
line material. Other elements have a similar, whether distinction between different granites

20 CRYSTALS
66 GRANITES

19 VEINS

37
METAMORPHIC 30 PEGMATITES
ROCKS

Figure 2. Abundance of chemical impurities in quartz from different environ-

ments
244 W. H. DENNENIMPURITIES IN QUARTZ

is possible. There is a fairly large scatter
among individual samples from the same rock
mass, but the modal values for populations
of about 10 samples tend toward rather dis-
tinctive values. Figure 3 presents the modal
impurity contents for the quartz of eight
Island. The general parallelism of the plots is
consistent with these relationships.
A point of some interest is the comparative
chemistry of the quartz in the three granite
stocks in eastern Massachusetts, all of which
have been assigned to the Quincy Granite by

Sample
Ftopo lotion
Dedham Granodionfe
Basal Hoppin Formation
R.I.
(xlO)

Cape Ann Stock. 28

Peabody Stock 25
Quincy 5tocK zo
QUINCY GRANITE
MASS.

Quarl-j Monjonite 10
Granodiont-e IO
IDAHO BATHOLITH
(xo-l)
Granite , W. Side Chilkoot FbSS 10
ALA5KA- V.T.
Devonian Granite 9
NOVA SCOTIA
(xO-Ol)

Mg Fe. T| Al
Figure 3. Comparison of chemical impurities in quartz from different
granite rocks

different rocks. Some of these rocks are
genetically related, i.e. the quartz monzonite
and granodiorite of the Idaho batholith,
Peabody and Quincy stocks of the Quincy
Granite, Massachusetts, and the Dedham
Granodiorite and the derived metasedimentary
rocks of the basal Hoppin Formation, Rhode
various authors, e.g. Emerson (1917, p. 186jf),
and Clapp (1921, p. 25). In this work, distinct
similarities are noted between the Quincy
and Peabody stocks and significant differences
between these two and the Cape Ann stock.
It thus appears that the rock of the Cape Ann
stock is not Quincy Granite. This conclusion
 SHORT NOTES 245

receives firm support in the work of M. L.
Bottino (1963, PhD thesis, Mass. Inst. Tech-
nology), who has determined the ages of the
rocks in the Quincy and Cape Ann stocks to
be 325 + 15 m.y. and 415 10 m.y. re-
spectively.
Another point of interest in the age relation-
ships of the rocks of eastern Massachusetts and
Rhode Island is the similarity of the trace-
element content of the quartz in the contact
zone of the Dedham Granodiorite and basal
rocks of the Hoppin Formation. This similarity
is consistent with sediments derived by weath-
ering from the igneous rock as established by
A. O. Ramo (1963, SB thesis, Mass. Inst.
Technology) on the basis of field studies and
age measurement. The decrease in impurity
content, especially for titanium, in the Hoppin
Quartzite may be ascribed to partial recrystalli-
zation and expulsion of impurities from the
quartz in this rock.
Conclusions
The impurity content of quartz separated
from the same rock has a fairly large scatter,
but familial characteristics may be readily
recognized when the modal values obtained
from about 10 samples representing a given
rock are used.
It is believed that careful sampling and
analysis of quartz can lead to correlation of
genetically related rocks and that the general
procedure may be applied to problems of
provenance.

References Cited
Clapp, C. H., 1921, Geology of the igneous rocks of Essex County, Massachusetts: U. S. Geological Survev
Bull. 704, 132 p.
Emerson, B. K., 1917, Geology of Massachusetts and Rhode Island: U. S. Geological Survey Bull. 597,
289 p.
Frondel, C., 1962, Dana's system of mineralogy, Volume III, Silica minerals, 7th Ed.: New York, lohn
Wiley and Sons, 334 p.
Stavrov, O. D., 1961, On the content of rare elements in quartz: Geochemistry, no. 6, p. 542-549
Tatekawa, M., 1954, Spectrographic distribution of minor elements in the quartz and feldspars contained
in granites and pegmatites of the Oku-Tango district, Kyoto Prefecture: Mem. Coll. Sci. Univ.
Kyoto, v. 21, ser. B, p. 183-192

DEPT. GEOLOGY AND GEOPHYSICS, MASSACHUSETTS INSTITUTE OF TECHNOLOGY, CAMBRIDGE, MASS.

MANUSCRIPT RECEIVED BY THE SOCIETY, OCTOBER 23, 1963