Determination of Lithium Isotopes at Natural Abundance Levels by Atomic Absorption Spectrometry

Allen L. Meler US. Geological Survey, MS 973, Box 25046, Federal Center, Denver, Colorado 80225
The relationships of the absorption of 'Ll and 7Li hollow cathode lamp emlsslons are used to
determine ilthlum Isotopic composition In the natural abundance range of geologic materlais.
Absorption was found to have a nonlinear dependence upon total ilthlum concentratlon and Isotopic
composltlon. A method using nonlinear equatlons to describe the relatlonshlp of the absorption of
'Li and 7Li lamp radlation Is proposed as a means of caicuiatlng Isotopic composition that Is
Independent of total llthium concentration. The determination of the occurrence and relative concentration of
stable lithium isotopes in natural materials is of interest in geochemical exploration as a guide to hydrothermal
alteration. It may also have useful application in the exploration for ore deposits and geothermal reservoirs (I).
Lithium-6 is apparently enriched in rocks associated with hydrothermal alteration (I). Isakov et al. (2) reported an
increase in lithium-6 as the degree of rock alteration increased. The average abundance of lithium in the earth's crust
is about 20 ppm, ranging from 5 to 200 ppm in soils and averaging 10 ppm in basalt and 60 ppm in shale (3). Some
lithium minerals may contain as much as 4.7% lithium. With the exception of these minerals and highly anomalous
samples, the expected lithium concentration range for most geologic samples is about 10-200 ppm. The average
natural isotopic abundance of lithium is 7.42% Li and 92.58% 7Li (4). Some controversy exists over how much
natural variation in isotopic composition exists. Svec and Anderson (5) summarized published values, determined by
mass spectrometry, which showed isotopic abundance variation in lithium reagents and lithium separated from various
minerals to be from 6.90% to 7.98% 6Li. They attributed this variation to instrumental factors as well as natural
variation. In their study of lithium-bearing minerals, variations of only 7.34% to 7.61% 6Li were found. Isakov et
al(2) reported a range of 6.77% to 9.28% 6Li in mica samples analyzed by mass spectrometry. Natural water samples
analyzed by atomic absorption were reported to have as much as 13.38% 6Li (6). Divis and Clark (I) reported values,
determined by atomic absorption, of 7.30% to 27.4% "i in hydrothermally altered rocks and an average of 9.38% 6Li
in 24 unaltered volcanic rocks, with a standard deviation of 1.94. The values for the hydrothermally altered rocks have
the greatest reported variation in natural lithium isotopic abundance that is readily found in the literature. The
possibility of determining isotopes by atomic absorption was suggested by Walsh (7). The basis for the atomic
absorption determination of lithium isotopes is that each isotope emits a doublet at the 670.8-nm resonance line of
lithium. The two peaks of each doublet are separated by 0.015 nm, and the doublet of one isotope is shifted by 0.015
nm with respect to the other. The upper wavelength of the 7Li doublet overlaps the lower wavelength of the 6Li
doublet as a result of the isotopic shift and doublet separation being equal. The separate wavelength components
cannot be resolved by conventional atomic absorption spectrophotometers; therefore, the isotopes cannot be
determined independently. However, the difference in absorption of emissions of monoisotopic lamps by the two
isotopes provides a means of estimating lithium isotopic composition (I, 6-12). The most commonly reported way of
relating the absorption of 6Li and 7Li lamp radiation to isotopic composition is to plot the ratio of the absorbance
measured using a 6Li lamp to absorbance measured using a 7Li lamp vs. percentage 6Li (1, 6, 10, 11, 13). The
resultant curve is nearly linear over the entire isotopic composition range. Some investigators (1,13) have reported the
ratio to be insensitive to total Li concentration. However, the absorbance ratio was found by others (6, 10, 11) to be
dependent upon the total Li concentration; therefore, standards and samples must be matched to make a valid
estimate of isotope composition. In this study, the total Li concentration was found to have a significant affect on t,he
absorbance ratio. Atomic absorption has been applied to the determination of lithium isotopes in material used in the
nuclear industry (10, ll), lithium reagents (IO), natural water (6), and rocks (1). The use of atomic absorption for the
determination of lithium isotopes in geologic materials is attractive when compared to the conventional method of
mass spectrometry in regard to cost, speed, ease of sample preparation, and availability to most laboratories involved
with geochemical analysis. Techniques for the dissolution of geologic materials for atomic absorption determination
of lithium are applicable to isotopic determinations. These techniques usually result in solutions that contain 1% or
less of sample by weight per volume of solution. The expected lithium concentration range of these solutions is about
0.1-2.0 pg mL-l with isotopic composition of about 5% to 20% 6Li by vveight. In this study the relationship of the
absorption of 6Li and 7Li lamp radiation to isotopic composition is examined, and a method is proposed for
estimating isotopic composition within the expected range for geologic materials.
EXPERIMENTAL SECTION

Apparatus. An Instrumentation Laboratories Model 951 atomic absorption spectrophotometer equipped with a Model
254 autosampler manufactured by the same company was used in this study. This instrument has two double beam
channels allowing simultaneous measurements of a single solution with individual 6Li and ‘Li hollow cathode
lamps. A single-dot 10-cm burner was used with an air-acetylene flame. Reagents. Lithium-6 and lithium-7 stock
solutions of 1000 pg mL-l were prepared firom monoisotopic metal obtained from Oak Ridge National Laboratory by
dissolving 1.000 g of freshly cut metal in approximately 50 mL of distilled-deionized water. Twenty milliliters of
concentrated HC1 was added and diluted to a volume of 1 L, with distilled-deionized water. Mixed isotopic
solutions of 100 hhg mL-l total Li and 0, 5, 10, 15, and 20% %i by weight were prepared by mixing and diluting
appropriate aliquots of the 1000 yg 1nL-l 6Li and 7Li stock solutions. Mixed isotopic working solutions were prepared
by diluting aliquots of each mixed isotopic solution to give solutions with total Li concentrations of 0.1-2.0 pg mL-’
in 0.1 yg mL-l increments. For samples with unknown isotopic composition and various total Li concentrations,
solutions of reagent grade LiC1, LiNOB, Li2S04, and Li2C03 were prepared to contain approximately 1.5, 1.0, 0.5,
and 0.2 pg mL-l total Li. Procedure. The atomic absorption spectrophotometer was set up with a 6Li lamp in one
channel and a ‘‘Li lamp in the other. Each lamp was operated at approximately 4 mA. The wavelength
of each monochromator was peaked at 670.8 nm, with a bandwidth of 0.5 nm. An oxidizing air-acetylene flame was
used, adjusted to give maximum absorbance when a lithium solution was aspirated. The working solutions and reagent
solutions were placed in the autosampler, and a simultaneous 10-s integration of the absorbances expanded 10 times
was taken for each solution. The instrument was automatically zeroed after each four solutions. Five consecutive
readings were made on the set of solutions over a 6-h period. The flame conditions were then readjusted and another
five readings were taken in the same manner.
RESULTS AND DISCUSSION
The individual isotopic lines could not be resolved by this instrument even when a bandwidth of 0.04 nm was used.
The use of bandwidths of less than 0.5 nm did little to improve the signal but did increase the noise level. A nearby
neon line at 671.7 nm did not interfere with the 670.8-nm lithium line at a bandwidth of 0.5 nm. The dual-channel
capability of the instrument made it ideally suited to this work because simultaneous measurements of each solution
could be made by using individual hollow cathode lamps. This capability not only increases the speed at which data
are collected but helps to match the measurement conditions for each lamp. However, flame conditions have a
significant affect on the sensitivity of the isotopes. In this study the flame conditions gradually changed
over a 6-h period, resulting in a decrease in sensitivity of about 18% for the 7Li lamp and 13% for the 6Li lamp. This
unequal change in sensitivity invalidates comparisons of measurements
taken within the first few hours of operation. After the first 6 h of operation the flame was readjusted, and the
sensitivity of both lamps then remained relatively constant for the next 6 h of operation. Simultaneous absorbance
values using the 6Li lamp (A6) and absorbance values using the 7Li lamp (A7)
for each of the mixed isotopic working solutions and reagent solutions were taken in five consecutive runs after the
flame was readjusted. The mean and relative standard deviations of the two absorbance values for each solution were
calculated. Tables I and I1 list the means and relative standard deviations of A7 and A6 for five different total Li
concentrations of the mixed isotopic working solutions. These five concentrations are representative of the
concentration range of interest. However, all 20 concentrations for each isotopic composition were used in the
comparisons and calculations made in this study.
Table I11 shows the ratio of the A, to A7 of five total Li
concentrations for each isotopic composition studied. The table shows that the absorbance ratio (A6/A7) increases
with decreasing total Li concentration. Thus, use of the absorbance ratio alone, without matching the sample and
standard concentration, would lead to serious errors in the estimate of isotopic composition. The relationship between
As and A, for solutions of the same isotopic composition is defined by eq 1. The coefficients
A6 = u + bA7 + cA~ (1)
a, b, and c are determined by solving the normal equations
for the least-squares parabola of A, vs. A, for each set of solutions with the same isotopic composition. The
coefficients were determined for each set of mixed isotopic working solutions used in this study and are given in the
following equations:

0% %i A6 = 0.0011 0.42644, - 0.1051A7'

5% 'Xi A6 = 0.0016 + 0.45934, - 0.1044A7'
10% 6Li A6 = 0.0020 -!- 0.4921A7 - 0.1007A7'
15% 6Li A, = 0.0006 -!- 0.5381A7 - 0.1138A7'
20% 'Li A6 = 0.0014 + 0.57604, - 0.1156A7' (2)
Figure 1 shows the best-fit parabolas for these equations.
Each pair of absorbance values has a corresponding unique
isotopic composition, thus providing a means of determining the isotopic composition of unknowns by using A6 and
A7. The standard formula for calculating elemental concentration
in a sample is (sample abs - blank abs)(std concn) sample concn = (std abs - blank abs) (3) This formula can be used
to calculate isotopic composition by substituting the appropriate equations into the formula Blank absorbance and
standard absorbance have a nonlinear dependence on A,, as defined by eq 1. Therefore, the formula
becomes
(4)
(A6 - ao + + ~OA~~)(Z)
% 6Li =
(az + ~ Z A ~ C ~- A(a,o~ + )b oA7 + ~oA7')
where AB = absorbance of the sample measured with a
lamp, A7 = absorbance of the sample measured with a 7Li
lamp, ao, bo, and co = coefficients for 0% 6Li, and aZ, bZ, and
cz = coefficients for 2% 6Li, where 2 = percent 6Li in
standards.
The coefficients for each set of mixed isotopic working
solutions were substituted into the formula and reduced to
derive the following equations:
5% 6Li =
As - 0.0011 + 0.426447 - 0.1051A72
for 0 and 5% 6Li
(5)
0.0001 + 0.0066A7 + 0.0001A72
A6 - 0.0011 + 0.4264A7 - 0.1051A72
0.0001 + 0.0066A7 + 0.0004A72
% 6Li = for 0
and 10% 6Li (6)
A6 - 0.0011 +
0.426447 - 0.1051A,'
0.0074~47- 0.0006A7'
% 6Li = for 0
and 15% 6Li (7)
A6 - 0.0011 +
0.4264A7 - 0.1051A7'
0.0075A7 - 0.0005A72
5% 6Li = for 0
and 20% 6Li (8)
The differences in the coefficients for each of the isotopic compositions are due to experimental error and a difference
in curvature as the percent 6Li increases. By averaging the coefficients, one can derive a compromise equation which
can be used to calculate an estimate of percent 6Li over the entire range.
A6 - 0.0011 0.426447 - 0.1051A72
% 6Li = (9)
o.oo70A7 - 0.0002A~~
The A6 and A7 values for the mixed isotopic working solution were used in eq 9 to calculate the estimated percent
6Li of each solution. Table IV, which lists the mean and standard deviation of the estimated percent 6Li, shows that
the equation fits the data and the estimated percent 6Li is in close agreement with the actual percent 6Li of the
solutions. The estimated 6Li percentages for the Eiolutions of lithium reagents were calculated by using eq 9. Talble
V lists the mean and standard deviation of five analyses of reagent for different total Li concentrations. The mean and
13tandard deviation for all analyses of each reagent are also shown. The data show that the precision of tho
estimate decreases with decreasing total Li concentration. The estimates of isotopic composition are in close
agreement at all concentrations, indicating that the effects of total Li concentration are adequately accounted for in
the calculation. unfortunately, no statement can be made about the accuracy because no previously analyzed reference
material was available. For greater precision and accuracy, coefficients for standards that bracket the concentration
and isotopic composition Of interest may be used in the formula. ]For example, if the concentration range of interest
is 0.5-1.0 yg mL-l Li with isotopic composition of 5% to 10% "i, then standards of 0.5-1.0 yg mL-l total Li
containing 5% and 10% 6Li are used to determine the coefficients. The coefficients for 5% 6 Li are used in place of
the 0% 6Li coefficients and the coefficients for 10% 6Li are used in the denominator. Because 5% 6Li is the base
line, 5% is used as the standard concentration and 5% is added to the result to give the estimated 6Li% in the sample.
The findings of this study show that atomic absorption spectrophotometry is applicable to the determination of lithium
isotopes in the natural abundance range of geologic materials. The use of nonlinear equations to describe the
relationship of the absorption of "i and 7Li lamp radiation provides a means of calculating isotopic composition that is
independent of total lithium concentration. The precision and ability to resolve small differences in isotopic
composition decrease with decreasing total lithium concentration. Precision and accuracy are improved by keeping
the total lithium concentration as high as possible and by using standards that bracket the sample concentration
and isotopic composition.

ACKNOWLEDGMENT

The author thanks J. G. Viets, R. W. Leinz, and T. T. Chao

of the U.S. Geological Survey and R. K. Glanzman of Chevron.