Automatic Measurements and Computations for...
1 downloads
155 Views
3MB Size
Automatic Measurements and Computations for Radiochemical Analyses JOHN N. ROSHOLT, Jr.,l and J. R. DOOLEY, Jr. U. S. Geological Survey, Denver, Colo.
b In natural radioactive sources the most important radioactive daughter products useful for geochemical studies are protactinium-231 , the alpha-emitting thorium isotopes, and the radium isotopes. To resolve the abundances of these thorium and radium isotopes by their characteristic decay and growth patterns, a large number of repeated alpha activity measurements on the two chemically separated elements were made over extended periods of time. Alpha scintillation counting with automatic measurements and sample changing is used to obtain the basic count data. Generation of the required theoretical decay and growth functions, varying with time, and the least squares solution of the overdetermined simultaneous count rate equations are done with a digital computer. Examples of the complex count rate equations which may be solved and results of a natural sample containing four a-emitting isotopes of thorium are illustrated. These methods facilitate the determination of the radioactive sources on the large scale required for many geochemical investigations.
T
of the sources of natural radioactivity has provided clues to geochemical history in the study of geologic and geochemical processes. The basis for many of these clues is the degree and kind of radioactive disequilibrium present between the parents and daughter products of the uranium-238, uranium-235, and thorium-232 series (11, 12). For a maximum period of approximately '/4 million years, uranium, protactiniuni231, and all the naturally occurring thorium and radium isotopes can be useful as natural tracers in these studies. Chemical methods of analysis are used for the uranium determination (4, 15). Thorium-227 and radium-223 in most natural samples are assumed to be in radioactive equilibrium with protactinium-231 and their abundances are used for the determination of protactiniumHE DETERMINATION
Present address, U. S. Geological Survey The Marine Laboratory, University of hiarni, Miami 49, Fla.
231. With this assumption, the measurement of the thorium and radium isotopes will comprise the method used for the determination of all the longlived daughter products. The purpose of this paper is t o describe methods with several different types of computation which can be used to resolve the isotopic abundances of the thorium and radium. Multiple radioactive daughter products have been determined by other investigators. Baranov and Kuzmina (1) have described the radiochemical determination of thorium isotopes and radium-226 in deep-sea sediments. Muclear photographic emulsions have been used for the measurement of the four a-emitting thorium isotopes in sea water (6). Some of the thorium and radium isotopes along with other a-emitting decay products and uranium-238 in radioactive minerals have been resolved by a-spectrum measured with a grid ionization chamber (2). None of the published accounts describes the determination of all four a-emitting thorium isotopes and the four radium isotopes in solid samples as well as water samples. The method described here requires the measurements of the total a-particle radioactivity of individual thorium and radium separates by scintillation counting with zinc sulfide. The zinc sulfide scintillation phosphor is contained in the radioactive separate. Detailed standardization and efficiency calculations for the radiochemical separations and measurements of the thorium and radium isotopes are described elsewhere (IO). The determination of the abundances of each of the thorium and radium isotopes is a fairly simple problem for the restricted cases where only the two uranium series or only the thorium series contribute essentially all the a-radioactivity. There are only two isotopes each of thorium and radium to resolve. The problem becomes much more complicated when all three radioactive series are present in a natural sample with the occurrence of four isotopes of each of the two elements (Th232, Th23'3, Th228, ThZz7, and R a 2 2 8 , Ra226, R a z z 4 , and F t a z 2 3 ) , the determination of which may be desired. Complex decay curves are obtained when the combined radioactivities of four iso-
topes are counted many times. The theoretical growth and decay properties of each isotope and its immediate daughter products as defined by Bateman equations are used to resolve the abundances of the various thorium and radium isotopes. Growth and decay functions (3, 6, IO), derived from Bateman equations, are used in the calculation of the abundances of all the isotopes under consideration. Perkel (8) has also used Bateman equations to resolve complex decay curves for three radioactive isotopes. APPARATUS
A medium-sized digital computer is used for reduction of the count data, generation of the theoretical growth and decay functions, and the numerical analysis necessary to resolve the isotopic components which were counted. Automatic measurements and machine computations used in the analysis of artificial radioisotopes have been published (7, 9, 14). The construction of an automatic sample changer similar to the noncommercial sample changers used in this laboratory has been described in detail (19). S e v e r a1 s m a1 I-c a p a c i t y s a m p 1e changers u hich can accommodate from one t o eight radioactive preoipitate samples are used to complement the commercial sample changers which have a maximum capacity of 50 samples. The special-purpose sample changers are operated from 100-volt alternating current that is automatically switched by the scalers. Figure 1 shows a photograph of a disassembled machine. A 1-r.p.m. brake-stopped induction motor operates the turntable, which holds the samples. Each sample holder, G, has a cam, H , which is positioned when a sample is placed in that particular location. This cam actuates a microswitch, I , that stops the turntable motor when the sample is in the counting position. After the sample is counted for a predetermined number of counts, an actuation signal from the scaler starts the turntable motor and the next sample is brought into counting position. The sample-changing operations are repeated until a cycle of the turntable is completed; then a second microswitch, J , is activated which rezeros the index number on the printing timer and the samples are recounted. The sample holders without sample rings will be bypassed because the cam, VOL. 32, NO. 9, AUGUST 1960
1093
H, will not he held in position to actuate the microswitch. Both the commercial and the specialpurpose sample changers contain the same scintillation detector, scaler, and timer components. The multiplier phototube, C, used is a 10- or 12-stage 5-inch diameter end window Dumont Model K1391 or K1209. A specially designed dome, B, houses the bare multiplier phototube in each sample changer. Highly polished aluminum surfaces on the interior of the sample ring and on a reflector cone, D,placed
between the sample and the multiplier phototube, increase the reflected light from scintillations reaching the photocathode. A light seal attached to the base of the inverted adapter cone prevents leakage a t the junction of the turntable and phototube housing. The sample changer unit, enclosed in a fiber package drum, provides additional light shielding. Samples are loaded and removed throughlighhtight accessdoors. Binary-type scaling units, vhich inelude automatic resetting registers with a wide range of preset count selection,
are used. The time required to accumulate the predetermined number of counts is recorded on commercial models of automatic printing timers. The range of the timer is from 0 to 999.99 minutes. The automatic counh ing system is completed by connecting one pen of a multiple pen production recorder in parallel with the motor of each sample changer to record the time of day for each sample change. A counting laboratory with two commercial and three special-purpose .sample changers and associated recording e q u i p mentis shown in Figure 2. Trouble-free sample changer operations require that the size and shape of the sample rings he held within close tolerances. The mounted filter membrane must be ruaeed to withstand sample changing an&torage for several months. A cutaway diagram of a mounted sample is shown in Figure 3. OPERATION AND CALIBRATION
enibled special. orpose changer G. ! nple holder H. I itioning cam itioning microwitch
K.
1.
Turntable shaft Turntable
iic counting system
A set of samples is counted in the order in which they are placed in the sample changer. The same order is retained while a large number of repeated count measurements are made, automatically, until the sequence is manually interrupted. The time of day a t which each sample is counted is obtained from the production recorder and included with each piece of count data, When three or four a-emitting isotopes of thorium or radium are present in a sample, counting measurements are made which extend over periods of several months. This integrated time, during which the samples should he counted a large number of times, varies considerably with the isotopes to be resolved. Approximate integrated times for several specific cases are included in Table I. For the short integrated times and during the initial part of the longer integrated times, count measurements are made at frequent intervals. During the later stages of the long integrated time, measurements are made at less frequent intervals. Calibration of the counting efficiency of the instruments is necessary in making integrated measurements over periods greater than a. few weeks. For this purpose, several thorium-230containing precipitates older than 4 months and several radium-226containing precipitates older than 1 month are counted periodically and corrected for the growth of long-lived daughters. The a-radioactivity of these calibration standards is approximately 100 counts per minute and each standard is counted automatically, three or four times for each calibration. Correction for any deviation in the counting efficiency, based on the average calibration value, is applied to the count data hefore the data are used in the machine computations. Experience with several scintillation counters of this type shows
that very stable operation can be maintained when the machines are allowed to run continually. The counting units tyere not shut down for more than a few hours during a 2-year interval. Comparison of any of the periodic calibrations shows that deviations of the counting efficiency did not exceed 2% over the Zyear interval. Deviations in efficiency between adjacent 3-month intervals rarely exceeded 1%. For the smaller capacity special-purpose sample changers, each individual sample and each individual calibration standard are always counted in the same sample holder (Figure 1) to facilitate correction of the counting efficiency over extended time intervals. One Hollerith (IBNI) card is prepared for each count measurement. The follon-ing quantities are punched on the card for subsequent machine computations when sufficient counting data have been accumulated : the sample identification number, the time (year, month, day, hour, minute) a t TI hich the sample n a s counted, the predetermined count, the time required for this count, the background count rate, and a number representing the chronological order of each count measurement. With each set of data cards for each sample, a lead card is included to provide for the selection requirements for the machine computation. This lead card contains a numerical code
Table I.
PROTECTIVE CELLOPHANE CCIVER I10015") TRIMMED TO RING SIZE PRECIPITATE 0 CARRIER COVE:RS AREA OF 35 MM DIA 4'' CONTOURED RADIUS OF CURVATURE
SAMPLE RINGS MACHINED FROM 14 BW WALL GAUGE, TYPE 615 T 4L
. 2"O.D. 'UBING
CELLOPHANE 8 FILTER MEMBRANE CEMENTED TO UNDERSIDE OF RING OUTSIDE DlA 1985 f 002
Figure 3.
Cutaway drawing of a sample mount
for the proper Bateman constants for the isotopes selected. a code to select the type and size of calculation, and a scaling code to facilitate fixed point computer calculations. MACHINE COMPUTATIONS
A general-purpose program consisting of 1636 instruction steps and constants has been prepared for the Datatron 205 computer using fixed point arithmetic. IBM card input is used with Cardatron editing control and line printer output. The scaling conditions imposed by fixed point arithmetic limit the measured gross count rate of the sample to bztween 1.00 and 9999.99 counts per minute with a maximum of 50 separate count measurements requiring one I B M
data card per measurement. The number of thorium or radium isotopic components which the program can handle are from one to four. Each thorium or radium isotope which contributes t o the count rate must be specified for the computation along n-ith the background count rate. One of five basically different types of calculation can be selected for the sample, depending on the number and kind of isotopes involved. The compound count rate can be represented as a linear equation of Bateman functions. For the isotopic determinations described in this paper, the general count rate equation will have a dependent variable and a maximum of four coefficients and four variable functions.
Count Rate Equations Listed by Type of Calculation Used and Combinations of Isotopic Components Computed
Type of Calculation 1, two isotopes
Count Rate Equations for Thorium Isotopes"
Equation To.
Approximate Integrated Time of Count 15 days 10 days
Count Rate Equations for Radium Isotopes
Approximate Equation Integrated So. Time of Count 3 days 10 days 30 days
2, three isotopes
100-180 days
20 days
100-180 days
2 0 days
3, three isotopes
15 days 15 days 15 days
4, four isotopes
24-36 months
18-24 months
24-36 months
18-24 months
5, four isotopes
fThz" = 1, and except Equations 21 and 23 fThZ3' = 1, because of their constant decay rate and negligible growth properties of their immediate daughter products during protracted time of count measurements.
VOL. 32, NO. 9, AUGUST 1960
1095
CT =
Clfl
+ + + Czfz
c3f3
c4f4
(1)
where CT =
c
=
f =
measured total net count rate of mixture of radioisotopes count rate coefficient of each particular isotope which is the quantity sought variable function of decay of a particular isotope and subsequent growth and decay of its daughter products with time, as defined by the theoretical growth and decay relationship
A count rate equation of this form would represent each counting measurement, and with the accumulation of a relatively large number of measurements, n, the overdetermined simultaneous equations are solved for their coefficients by the method of least squares (16).
where i is the numbered count measurement (calculation card number) and i = 1 to n. To allow the sum of the squares of
measurement
the errors to be minimized, the first partial derivatives are taken with respect to cj, the parameters sought. 4 n
czfii
-~af3~ c4faal
(3)
where j = 1, 2, 3, and 4 for each i
i
and
=
1t o n
These first partial derivatives are allowed to become zero to determine thc most probable value of c,
dsi dCi
=
0 for j
4 n
0 =
.fix
j=1 i = l
[-
1, 2, 3, and 4
=
cr,
+
CIS,
cd..
+ +
Cd2,
crj4,l
+ (4)
where j = 1, 2, 3, and 4 for each i, and i = 1 to n, which can be stated in the generalized form used for solution of the simultaneous equations by matrix manipulation.
Evaluate decay functions
calcrilation and select isotopes for computation
A scaling code, preselected and punched on the lead card for each sample, is necessary to accommodate a broad range in the number of count measurements and the gross count rate using fixed point arithmetic in the computations. Three different kinds of quantities are involved in the coefficients of the normal equations: C r , and f j l , calculated for each count measurement, and n. The total number of count measurements, 71, will be used in a straight-line equation form for two isotopes discussed later. The growth-decay function, f, is the Bateman function. Z/ce-x‘, where t is the time elapsed, in hours, between the chemical separation and the individual count measurement and is evaluated by the use of prestored constants in the computer program. The Bateman coefficient, k , 2nd the decay constant, A , are se-
Calculate coefficients
Store and print out count rate equation coefficients
of count rate equa-
1,2,4 b
tions
1,2,3,4,5 P
J
( e . subroutine)
3 5
I
IL
I
equations
equations
Print out individual isotope count (Sample calculation completed) 1,3,5
2,4
Pre are two isotope mdified count rate equations with results of three- and four-isotope component data
Figure. 4
Block diagram of computations
2.4 refer to initial computotion of total of three or four isotopes 21,41 refer to additional calculations solving for two of three or four isotopes 216,41, refer to completed additional calculations
1096
ANALYTICAL CHEMISTRY
Count rate and 2,,4 hours elapsed after? isolation ~
I
i.l,*l
Determine percentage concentration of isotope equations
1,224
print out auxiliary equations used for two isotom solutions of thee isotope componenta
1
lected from the prestored constants by the use of the isotope codes specified on the lead card. The evaluation of e-xt is made by a power series expansion. The particular growth-decay function, jj,, is the summation of each heAxt term of the Bateman equation. Each individual value of j l t is used for the least squares solution as shown in the generalized equation (4a). After all of the data cards for a sample have been processed with completion of the calculations of the net count rates and growth decay functions, the coefficients of the least squares equations are evaluated by summation of the individual cross products for each measurement yielding an equation for each unknown. The solutions of the overdetermined simultaneous equations are made by matrix multiplication to determine the activity of each isotope, c,. For most types of samples some additional calculations are made using the results of the least squares determination. The descriptions of these additional calculations, referred to as postcomputation calculations, are included in the discussion of each separate type of calculation. -4code, punched in the lead card for each sample, is used t o select the size of the matrix for least squares determination along with the selection of the type of calculation desired. Table I shows the common count rate equations which may be processed with each type of calculation, The approximate length of time during which several count measurements are made is indicated for each equation. A straight-line equation of the form, y = a bx, is used for the count rate equations when only two isotopes are computed, so that the printout for each count measurement can be manually plotted for postcomputation checking and inspection purposes. The general count rate equation is then:
putation of the count rate Equations 13 to 15, 21, and 22. Further explanation in conjunction with Table I is required for some of the types of calculations. Type 3 utilizes the second generation radium isotope ratio to solve for two isotopes in a three-isotopic component mixture. The short-lived second generation radium isotopes, radium-223 and radium-224, are used to determine the ratio of their parent isotopes, thorium-227 and thorium-228, respectively, which was originally present in the sample. The original thorium isotope ratio is defined by Equations 17 and 18. I n the precipitate containing the isolated thorium isotopes, thorium-228 and thorium-227 will generate radium-224 and radium223, respectively, in a known proportion dependent upon their individual decay constants. The radium-224 and radium-223 isotopes are separated from the precipitate of the isolated thorium isotopes after the count measurements have been completed for the thorium isotopes (IO). The elapsed time for the growth of these radium isotopes is the time interval between the thorium isotope isolation and the subsequent radium isotope isolation. This elapsed
~
Table II.
Radium Isotope Growth, Hr.
30
314.5
Chemical Sepn. to Count, 35
and the least squares equations will be :
5 crj
= CI(71)
+
c2
2
i=l
f. f li
(6)
35
38
where i = 1 to n and n is the total number of count measurements. The general Equations 5, 6, and 7 for the two isotope forms are used for the computation of the specific count rate Equations 8 to 12, 16, 19, 20, and 23 to 28, shown in Table I. The general Equations 1 and 4 are used for the com-
Partial Calculations for Sample 2701 86 with Four Isotopes of Thorium
Calcn. Code
+
i=1
time is used to calculate the two growth constants, Equation 18. The results of Equations 12 and 18 are used to determine the ratio, K , Equation 17, which is used in Equations 19 and 20. I n Type 5 calculations, measurements for the long-lived isotopes, Equations 25 and 26, are made after the shortlived isotopes no longer reflect an individual level of radioactivity. The results of Equations 25 and 26 are subtracted from the measurements shown by Equations 27 and 28 taken during the initial period of the decay of the isotopes. After completion of the solution of the initial count rate equations, additional calculations are made for most of the samples. The per cent equivalent concentration (IO) is calculated for samples con-puted as Type 1 by comparison to standards. One of two kinds of additional calculations is made for samples computed as Types 2 and 4. For each count measurement, the values for each term in Equations 13 to 15, 21, and 22 are determined and printed out; or, transformations of the net count rates are made t o modified net count rates, using Equations 16, 23, and 24 for a maximum of any three
a
Hr. 26.2 86.9 171.3 238.6 265.2 296 7 314.4 334.7 366 5 393.8
f(v)R s Z z 4
f ( p ) RaZa3,
Eq. 18 0.4254
Eq. 18 0,9093
Net C T ~ -0.35
Background
fRaZz3
fRaZZ4
23.673 14.997 7.443 4.767 4.043 3 461 2.776 2 571 1.961 1.545
3.7379 3.1967 2.5715 2.1618 2.0183 i.86ii 1.7779 1.6872 1.5546 1.4489
3.2471 2,0068 1.0272 0.6022 0.4875 0.3798 0.3300 0.2809 0.2184 0.1758
6.333 4.619 2.894 2.205 2.003
1.860 1.561 1.523 1.261 1.066
3 9 15 20 22 24 25 26 27 28
0.8687 0.6278 0.3995 0.2786 0.2415 0.2041 0.1856 0.1665 0.1406 0.1213
(Table I, Calcn. 3)
CRa2235
CThZ2"
&224.
CTh
Eq. 17 10,226
Eq. 12 0.351
Eq. 17 0.740
Eq. 12 6,888
Eq. 17 7 ,575
Chemical Sepn. to Count, Hr.
Background
11.1 22.2 43.7 i2.6 145.6 221.8 307.5
45.576 48.958 53.340 58.901 70.864 76.149 79.894
Net C p -0.70
fTh"'
2.0797 2.1684 2.3282 2.5182 2.8864 3.1319 3.2797
1.2724 1.5533 2.0710 2.6627 3.6774 4.2575 4.5942
15.091 18.052 23,506 29.747 40.491 46.669 50.261
2 4 7 10 15 20 26
Counts per minute.
VOL. 32, NO. 9. AUGUST 1960
e
1097
preselected combinations of isotopes. Equation 24 is an example of one such combination. A block diagram of the computations for all of the types of calculations described is shown in Figure 4. RESULTS
To illustrate an actual numerical example, partial results appear in Table I1 as they are tabulated as output from the computer. The sample is a 65,000-year-old core segment of a globigerian ooze deep-sea sediment. Data for the three required count rate equations are shown with the sample number and calculation code. This abbreviated example illustrates the use of the second generation radium isotopes to determine the ratio of the two shorter-lived thorium isotopes, Th227and ThZZ6. In the follon-ing order the tabulation shows: the fraction of the total possible Ra223 and Ra224 isotopes allowed to grow in the thorium precipitate; 10 of the 28 count measurements made on the subsequent decay of the isolated Ra223and Ra224isotopes; the results of the least squares solution of the 28 count measurements (Equation 12); 7 of the 26 count measure ments previously made on the growth of the thorium isotopes; and the results of the least squares solution of the 26 count measurements, Equation 19. DISCUSSION
The use of a digital computer has its greatest advantage in the analysis of samples which contain all three natural radioactive series. The experimental
data accumulated over the past 4 years show that counting measurements extending over long periods of time are required for the simultaneous resolution of three and four isotopes. At least 2 years of measurements are required for four isotopes as shown in Table I. Not only are counting measurements which extend over several years impractical timewise, but a considerable amount of effort is also required to calibrate the instruments to constant detection efficiency over long periods. The best solution to this problem found here is to restrict analyses to those types of naturally occurring material where the thorium-232 series, above radon-220, is assured to be in close equilibrium. Using the assumption that thorium-228 is in equilibrium with thorium-232, the Type 3 calculation utilizing measurements of second generation radium isotopes can be used to determine all the isotopic components described in a period of approximately 45 days. ACKNOWLEDGMENT
This work as part of a program undertaken by the U. S. Geological Survey on behalf of the Division of Rarr Materials of the Atomic Energy Commission. The authors are indebted to F. E. Senftle and F. J. Flanagan of the U. S. Geological Survey for valuable suggestions and assistance in the course of this work. lye also acknowledge the help and assistance from E. E. Wilson of the U. S. Geological Survey for his part in the construction of the special-purpose sample changers.
LITERATURE CITED
(l)$Baranov, V. I., Kuzmina, L. A, in Radioisotopes in Scientific Research,’’ R. C. Extermann, ed., Vol. 11, Proc. of 1st UNESCO International Conf. 1957, pp. 601-18, Pergamon Press, London, 1958. (2) FaccFni, U., Forte, M., Malvicini, A., Rossini, T.. AVucleonics 14, S o . 9, 126 (1956). ‘ (3) Flanagan, F. J., Senftle, F. E., h x . 4 ~ . CHEiV. 26, 1595-601 (1954). (4) Grimaldi, F. S., May, Irving, Fletcher, AI. H., U. S. Geol. Survey Cirr. 199, 20 (1952). ( 5 ) Kirby, H. W., A s . 4 ~ . CHEX 26, 1063-72 (1954). (6) Koczy, F. F., Picciotto, E., Poulaert, G., Wilgain, S., Geochim. et Cosmochim. Acta 11, 103-29 (195’7). (7) Nucleonics 15,S o . 5, 53 (1957). (8) Perkel, D. H., Ibid., 15, S o . 6, 103-6 (1957). (9) Relf, K. E., Ibid., 15, S o . 4, 86-9 (1957). (101 Rusholt. J. S . . Jr.. h r a ~ CHEM. . 1398-407 (1957). (11) Rosholt, J. X., Jr., in Proc. 2nd International Conf. on Peaceful Csee of .;2tomic Energy, Vol. 2, United Sations, Geneva, pp. 231-6, 1958. (12) Rosholt, J. N., Jr., U. S. Geol. Survey Bull. 1084 A, 30 (1959). (13) Rydberg, J., J . Sei. Insir. 32, 343-5 (1955). (14) Stone, R. S., in “Advances in Nuclear Engineering,” Vol. 2, J. R. Dunning and B. R. Prentice, ede., pp. 352-6, Pergamon Press, London, 1957. (15) Wahlberg, J. S.,Skinner, D. L., Rader. L. F.. Jr.. -4x.4~. CHEM. 29, 954-7 ’( 1957). (16) Worthing, A. G., Geffner, J., “Treatment of Experimental Data,’’ pp. 23840, Wiley, New York, 1943. I
,
~
I
,
RECEIVED for review January 21, 1960. Accepted May 9, 1960. Publication authorized by the Director, U. S. Geological Survey.
Removal and Identification of Organic Compounds by Chemical Reaction, in Chromatographic Analysis RICHARD BASSETTE and C. H. WHITNAH Departments o f Dairy Husbandry and Chemistry, Kansas State University, Manhattan, Kan.
b A method is presented for yielding evidence for identification of some organic chemical compounds with gas chromatography. In addition to retention times, techniques based on selective chemical separation of some of the components from a mixture with sodium bisulfite, and/or mercuric chloride with an accompanying alteration in the gas chromatogram were employed. This technique has application in eliminating impurities that may b e inseparable by chromatography prior to mass spectrographic analysis. 1098
0
ANALYTICAL CHEMISTRY
G
AS-LIQUID
CHROMATOGRAPHY
(GLC) has provided a method of fractionating mixtures of chemical compounds heretofore inseparable. Recent improvements in chromatographic instrument design, detector mechanisms, and column materials have increased the value of this analytical technique. The feasibility of identifying components in a mixture by adding reagents that would selectively remove certain classes of chemical compounds-e.g., carbonyl compounds, sulfides, acids, amines-from the solution to be chro-
matographed was studied. Several preliminary physical separations of mixtures entering a GLC column have recently been described ( 5 , 8 ) . Identical GLC pictures of a mixture of methyl esters were used by James and Martin (3) to prove the completeness of a lead salt separation of unsaturated from saturated fatty acids. I n another test, many GLC peaks disappeared after bromination and n-ere identified as unsaturated acids. Landowne and Lipsky (6) also used bromination, not t o remove peaks but to separate ester peaks otherwise unresolvable by GLC.
