ANALYTICAL CHEMISTRY, VOL. 50, NO. 6, MAY 1978
827
ACKNOWLEDGMENT
sodium hydroxide trap. T h e lead acetate trap (8 mm dia. X 6 cm long glass tube packed with lead acetate) is connected between the water vapor trap and the arsine trap. T h e additional C o p trap (8 mm X 8 cm glass tube packed with NaOH beads) is connected between the water vapor trap and the H2S t r a p (Figure 1). Experimentally, it was discovered that after the water, C02, and H2S traps were dried and repacked, more consistent results were obtained if the system was “sensitized”. This was accomplished by processing a high concentration of standard (50 pg As3+ and 100 pg DMAA) through the normal procedure for As3+but not using the detector. After this was completed, the reaction chamber was carefully cleaned, and several reagent blanks were run before the system was standardized. This “sensitizing“ procedure apparently saturates arsine absorption sites in the system and decreases the number of analyses required to obtain reproducible calibration curves. Samples should be analyzed periodically by the method of standard additions. This is most important for samples of a new or complex matrix, since organic and inorganic compounds may either affect the arsine generation, cause positive emission peaks, or quench the dc arc.
The author thanks R. W. Sanders for his assistance in this work.
LITERATURE CITED (1) R. S. Braman and C. C. Foreback, Science, 182, 1247 (1973). (2) . . R. S. Braman, L. L. Justen, and C. C. Foreback. Ana/. Chem. 44, 2 195 (1972). (3) R. S. Braman, D. L. Johnson, C. C. Foreback, J. M. Arnrnons, and J. L. Bricker. Anal. Chem.. 49. 621 11977). (4) E. A.-Crecelius, D. E.’ Robertson, J. Fruchter. and J. 0 . Ludwick, “Chemical Forms of Hg and As Emitted by a Geothermal Power Plant’’ in “Proceedings of the Tenth Annual Conference on Trace Substances in Environmental Health, Columbia, Mo.. June 8-10, 1976”, University of Missouri Press. (5) E. A. Crecelius, Bull Environ. Contam. Toxicoi., 18, 227 (1977). (6) E. A. Crecelius, Environ. Health Perspec., 19. 147 (1977).
8.
RECEIVED for review September 12, 1977. Accepted January 16,1978. This paper is based on work performed under the
U.S.Energy Research and Development Administration (Contract EY-76-C-06-1830), now functioning as the Department of Energy.
Column Switching Technique for Gas Chromatographic Analysis Donald E. Willis Monsanto Chemical Intermediates Company, St. Louis, Missouri
63 166
A variety of gas chromatographic methods has been developed for the analysis of mixtures of fixed gases and organic compounds, particularly light hydrocarbon gases (1-12). These methods generally involve the use of a molecular sieve column and a second column, typically a porous polymer packing, in (a) a parallel column arrangement, (b) a series/ bypass configuration, or (c) an arrangement with the columns in series across the detector (12). Detector polarity reversal may be required in (a) or (c). These arrangements also result in accumulation of materials not eluted from the molecular sieve column, necessitating eventual regeneration of the column. A novel column switching technique has been developed which avoids many of the problems associated with the above approaches to the analysis of fixed gases and organic compounds. The method is not specifically limited to the analysis of light hydrocarbon gases, but, with proper choice of columns, can be used for the analysis of a wide variety of organic mixtures. Only those components which can be eluted from a molecular sieve column are allowed to enter the column, all components are detected on one side of a thermal conductivity detector (no polarity reversal required), and the need for a matching restrictor [as in (b)] is eliminated. This is achieved by using all columns in the flow system a t all times, but reversing the order of flow through the columns, while maintaining the flow through the columns in the same direction. T h e proposed method is reasonably easy to set up and can result in reduced analysis time. The method, as will be illustrated, can be used with two, three, or four columns in series. T h e term “foreflushing” has been applied to the technique, as all components are eluted in the normal or forward flow through the columns, as contrasted to backflushing techniques for the analysis of heavy ends in gas samples. 0003-2700/78/0350-0827$01 OO/O
EXPERIMENTAL All chromatograms shown were obtained on an F and M Model 810 gas chromatograph equipped with dual thermal conductivity and flame ionization detectors and a 1-mV Honeywell Model 16 recorder. The valves used were from Carle Instruments. Inc.. Fullerton, Calif. Column packings were all 6 6 8 0 mesh, obtained from Supelco, Inc., Bellefonte, Pa. All columns were made with premium grade stainless steel tubing, specifically manufactured for chromatographic analysis, from Supelco. Inc. The column., were prepared by conventional methods and conditioned at Rn appropriate temperature before use.
RESULTS AND DISCUSSION
A flow diagram of one version of the proposed method is shown in Figure 1. In valve position I (Figure 14) carrier flow goes first through column 1, then column 2, and to the detector. Switching the valve to position I1 (Figure 1B) reverses the order of flow through the columns, but with the flow through both columns in the same direction. In position 11. components eluting from column 1 elute directly to the detector; components eluting from column 2 flow through column 1 again and then to the detector. Figure 2 is the chromatogram of a mixture of components as indicated, analyzed at room temperature with the columns shown. T h e sample was injected in position I and switched O p ,and N P . Elution of the to position I1 after elution of HP, remaining components then occurs simultaneously, with CH, and CO being eluted from the sieve column and C 0 2 and the C2 hydrocarbons being eluted from the series Porapak columns. The baseline disturbance observed on switching has been damped by the use of a short column between the valve and the detector. After this disturbance, the baseline is reestablished and remains stable for the duration of the analysis. ,%, 1978 American Chemical Society
828
ANALYTICAL CHEMISTRY, VOL. 50, NO. 6, MAY 1978 He
+
SAMPLE
1 ,
&PORT VALVE
8
IAJ VALVE POSITION I
3
3
7
1 $-{ z
3
t
1
I 6
TO DETECTOR
b
E:
He
3
8
IBI VALVE POSITION II
8
5
11
1 2
4
TO DETECTOR
6
8
10
12
14
18
Retention Time. minutes
Figure 1. Foreflushingtechnique with two columns
Figure 3. Chromatogram of fixed gas-hydrocarbon blend. Column 1: 2m X inch Porapak N. Column 2: l m X inch Molecular Sieve 13X. l m X inch Porapak T between valve and detector. Other conditions and compound identification as in Figure 2 7
I
I
I
I
,
,
,
1
2
3
1
5
6
1
Retention Time, minutes
Figure 2. Chromatogram of fixed gas-hydrocarbon blend. Column 1: 2m X ' / e inch Porapak R + 2 ft X inch Porapak S. Column 2: I m X inch Molecular Sieve 13X, room temperature, 45 mL Helmin, thermal conductivitv detector 150 mA X 8. ComDound
Figure 3 is a chromatogram of the same mixture using a different packing for column 1. T o act as a pneumatic damper and to slow the elution of COz for the switching time shown, a l m X 1/8 inch Porapak T column was used between the valve and the detector. In this case, no baseline disturbance was observed and the order of component elution is different from Figure 1. Figures 4 and 5 are chromatograms of the blend to illustrate the elution order which can be achieved using the more retentive molecular sieve 5A in place of 13X. Switching times are shown on the figures. Depending on the packing in column 1,C 0 2 may be eluted before or after NZ. In both figures, the valve was returned to position I for elution of CO from the sieve column. T h e analysis for hydrogen using helium carrier gas is a well-known problem ( 2 3 , 2 4 ) . The present method should be well suited for use of the palladium-alloy hydrogen transfer system (15). Figure 6 is a flow diagram of the method modified
I
I
,
I
I
1
1
2
4
6
8
10
12
,
1
14
1
16
1
1
18
20
Retention Time, Minutes
Figure 4. Chromatogram of fixed gas-hydrocarbon blend. Column 1: 2m X ' I 8 inch Porapak R 2m X inch Porapak N. Column 2: I m X inch Molecular Sieve 5A. l m X inch 5 % SE-30 between valve and detector. Compound identification and conditions as Figure 2, except attenuation X16
+
for inclusion of the hydrogen transfer system and thermal conductivity detector as developed by Carle Instruments for N, and He carrier gases. Depending on the concentration of hydrogen and its resolution from O2 on column 2, a delay column between the exit side of the HTS transfer tube and the He side of the detector might be required. Figure 7 is a chromatogram of some of the possible products obtained in the catalytic oxidation of propylene. In this case, column 2 has been replaced by a catalytic methanator for the conversion of CO and C 0 2 to methane, permitting the total analysis to be performed using a flame ionization detector. Hydrogen for the methanation was added just ahead of the methanator and serves additionally as a carrier gas after the valve switch as indicated. Although the flow through the methanator could be by-passed using a four-port valve after the elution of C 0 2 (detected as CHI), the arrangement proposed continuously purges the methanator with hydrogen and permits maintenance of good catalytic activity. In other
ANALYTICAL CHEMISTRY, VOL. 50, NO. 6, MAY 1978
,
1
2
1
.
5
6
I
7
8
l
9
829
l
1
0
1
1
1
2
1
,
Retention Time, minutes
Figure 5. Chromatogram of blend illustrating unusual elution order. Column 1: same as Figure 2. Column 2: same as Figure 4. Compound identification and conditions as in Figure 2 He
+ SAMPLE
1
HTS TRANSFER TUBE
,W O R T VALVE
I
,
1
2
4
'
; 1 1
1
Retention Time, Minutes
Figure 8. Effect of valve position of Molecular Sieve column resolution columns, conditions, and compound identification as in Figure 2 He
+ SAMPLE
1.41 VALVE POSITION I
I
8-PORT VALVE
/q31
-
TO DETECTOR
DETECTOR
Nz
He
Figure 6. Foreflushing technique with hydrogen transfer system for H, analysis
ne
is) VALVE POSlTiON 1 I
COLUMN 1
TO DETECTOR
COLUMN 2
Figure 9. Foreflushing technique with three columns
2
4
6
I
1
I
10
12
1.
L-
18 I 8
-
20
22
Retention Time, minutes
Figure 7. Chromatogram of C,H, oxidation products. Column 1: 2m X '/, Porapak Q 0.5m X '/, inch Porapak R. Column 2: replaced by 6 inch X I / , inch Ni/Chromosorb P methanator at 400 OC flame ionization detector 100 X 8. 40 O C for 3 min, then 20 O/rnin to 200 OC. Flows: 20 mL He/min, 15 mL H,/min added ahead of methanator. 15 mL H,/min directly to detector. Compound identification: (1) CO, , , (6) C,H, (7) C&, (8) acetaldehyde, (2) CH,. (3) CO,, (4) C2H4,(5) CH (9) acrolein, (10) aiiyi alcohol, (1 1) acetic acid, (12) bialiyl, (13) benzene. (14) acrylic acid
+
analyses in which CO, CHI, and C 0 2 were not adequately resolved on a different column 1, the methanator could be preceded by a column to give resolution of these components. Figure 8 illustrates the effect of valve position on resolution inch molecular sieve 13X column with column for a l m X 1 as given in Figure 2. At t h e same column flow, the linear gas velocity through the sieve column is less in position 11, resulting in greater resolution of CH, and CO. The observed resolution of the two components is dependent on the time after injection when the valve is switched. Figures 9 and 10 are flow diagrams of the method using three or four columns. T h e flow diagram of Figure 9 might be used with column 1 in the column oven and columns 2 and 3 external to the oven a t room temperature. With the proper choice of column 1 and temperature, the separations shown in Figures 2-4 could be expanded to include, for example, water and higher boiling components. Other options are possible with the proposed method. Columns 1 and 2 may contain the same packing material, but be of different lengths. This could result in shortened analysis time for separations in which a longer column (column 1 +
830
ANALYTICAL CHEMISTRY, VOL. 50, NO. 6, MAY 1978
Figure 7; the hydrogen used for methanation also passes through column 1 in addition to the primary helium flow. Foreflushing can be combined with backflushing by placing column 1 on a four-port valve. Components can be eluted from column 1 either in the foreflush or backflush modes. If the separation performed by column 2 was not required in some instances, the valve could be switched to position I1 immediately after injection, bypassing column 2 completely. Similarly, a sample could be injected on column 2 (valve in position 11) and switched to position I to bypass column 1.
LITERATURE CITED (1) (2) (3) (4)
(5) (6) (7) (8) (9)
t TO DETECTOR
(10) (11) (12)
Figure 10. Foreflushing technique with four columns
column 2) would be required for resolution of early eluting components, but a shorter column (column 1 only) would be adequate for other components in the mixture. Foreflushing a t an accelerated rate can be achieved by the use of additional carrier gas after column 2 (Figure 1). In position I, only t h e primary carrier gas would flow through both columns; in position 11, column 2 would operate only with the primary flow, but column 1 would be eluted a t a higher carrier flow. This situation in modified form is present in
(13) (14) (15)
V. P. Chizhkov and G. A. Yushina, Zh. Anal. Khim., 31, 16-22 (1976). G. T. Noles and M. L. Lieberman, J . Chromatogr., 114, 211-214 (1975). S. Nand and M. K. Sarkar, J . Chromatogr., 89, 73-75 (1974). D. R. Deans, M. T. Huckle, and R. M. Peterson, Chromatographia, 4, 279-285 (1971). E. Malan and E. Brink. Chromatographia, 4, 178-180 (1971). S. Lukac, Chromatographia, 3 , 359-362 (1970). R. R. Forsey, J , Gas Chromatogr., 6, 555-556 (1968). R. I. Jerrnan snd L. R. Carpenter, J. Gas Chromatogr., 6, 298-301 (1968). E. L. Obermiller and G. 0. Charlier, J . Gas Chromatogr., 6, 446-447 (1968). D. Bennett, J . Chromatogr., 26, 482-484 (1967). T. Doran and J. P. Cross, J . Gas Chromatogr., 4, 260-262 (1966). B. Thompson, "Fundamentals of Gas Analysis by Gas Chromatography", Varian Associates, Inc., 1977. J. E. Purcell and L. S. Ettre, J . Gas Chromatogr., 3, 69-71 (1965). R. Villalobos and G. R. Nuss. ISA Trans., 4, 281-266 (1965). Carle Current Peaks, 7. No. 2 (1974), Carle Instruments, Inc.. Fullerton, Calif.
RECEIVED for review November 21,1977. Accepted February 13, 1978.
Variations in Mass Spectral Fragmentation Produced by Active Sites in a Mass Spectrometer Source Annemarie Wegmann Department of Chemistry, Stanford University, Stanford, California
94305
The fact that active sites in a mass spectrometer source can cause changed fragmentation patterns has been largely overlooked in the past. These sites can have several origins (I). In the case described here, marked effects were observed after the source of a MAT-711 double focusing mass spectrometer was electropolished (2) by a commercial company. After this treatment, t h e overall sensitivity increased, but changed fragmentation patterns were observed with various compounds such as free sterols and indole alkaloids. T o prove t h a t t h e pattern change was due to active sites in the source, test substances were introduced via gold crucibles and different sample concentrations applied by gradually increasing the temperature of the solid probe rod. Drastic fragmentation changes were obtained with low sample concentrations (low temperature), but all spectra improved with increased sample concentration until they reached the originally known pattern (see Table I and bar-plots, Figures 1, 2, 3, and 4). These results demonstrate t h a t a larger number of molecules have to be present to cover the active sites in the source so as to yield a reproducibly satisfactory spectrum. To minimize the effects of active sites when only very small amounts of sample material are available, perfluorokerosene (PFK) was tried as a "source filler" to increase the sample concentration in the gas phase. This added material ("source filler") seems to act by coating the active sites in the source, allowing more of the sensitive molecules to get ionized without 0003-2700/78/0350-0830$01 .OO/O
l""1 8O-
DESMOSTEROL AT 80°C +
366
271
25a
5 6OP P
a a
2
m e
DESMOSTEROL AT 14OoC
lo0l
'271
80P
60-
Figure 1. Mass spectra of desmosterol at 80 OC and 140 high sample concentration) 1978 American Chemical Society
O C
(low and