Deactivation of Polar Chemisorption in a Fritted-Glass Molecular Separator Interfacing a Gas Chromatograph with a Mass Spectrometer William D. MacLeod, Jr., and Bartholomew Nagy University of California at San Diego, L a Jolla, Catif. 92037
AMONGthe molecular separators currently described (1-5) for interfacing a gas chromatograph (GC) to a mass spectrometer (MS), the Watson-Biemann separator (2) is perhaps the most versatile and most widely employed (6). This must be owing to not only its simplicity of construction, but also to its effectiveness over a broad range of G C carrier gas flow rates (7), permitting it to interface both packed and open-tubular G C columns to mass spectrometers. As shown schematically (Figure l), the separator essentially consists of a heated, fritted-glass tube which connects the G C effluent (1 atm.) to the MS ionization chamber under vacuum (lOP mm Hg). The exterior of the porous glass tube is exposed to the intermediate vacuum (1 mm Hg) of a mechanical pump to preferentially remove the light carrier gas (He) which effuses through the frit pores much faster than the heavier organic G C iractions traveling down the interior of the tube to the MS ion source. In the process, the G C fractions are relatively enriched while the pressure is reduced for safe admission to the MS. Although the separator is constructed of glass, presumably for chemical inertness at elevated temperatures, the long history of adsorption of organic compounds on siliceous surfaces, sometimes accompanied by decomposition, cautions against undue confidence in the inertness of the separator surfaces at normal operating temperatures (150-200" C). Our experience with two commercial Watson-Biemann type separators supplied with the mass spectrometer (Hitachi RMU6E) indicates that there can be a significant loss of certain organic compounds during passage through the separator, even after acid washing. When GC-MS chromatograms of selected terpenoids (Table I) recorded from the MS total ion monitor were compared with their G C hydrogen-flame detector counterparts, complete absence of alcohols and aldehydes was observed in the ion monitor chromatograms up to approximately 10-6 gram of GC-MS sample. Ketones and esters were less affected, while no losses of ethers or olefins were observed down to the minimum GC-MS sampling level gram). The great advantage of tandem GC-MS analysis is that it offers the possibility of separating and identifying the components of complex organic mixtures more quickly and comprehensively with less sample than any other technique. For this reason, loss of polar organic samples well above the average minimum level of sensitivity is serious. The similarity of this behavior to the chemical adsorptivity observed with diatomite-
(1) R. Ryhage, ANAL.CHEM., 31,759 (1964). (2) J. T,Watson and K. Biemann, Zbid., p. 1135. (3) S. R. Lipsky, C. G. Horvath, and W. J. McMurray, Zbid., 38, 1585 (1966). (4) P. M. Llewellyn and D. P. Littlejohn, Pittsburgh Conference on Analytical Chemistry and Applied Spectroscopy,Pittsburgh, Pa., February 1966. (5) R. F. Cree, Zbid., March 1967. (6) W. H. McFadden, Separation Science, 1,723 (1966). (7) M. A. Grayson and C. J. Wolf, ANAL.CHEM., 39, 1438 (1967).
Table I. Approximate Minimum GC-MSa Sample Sensitivity with a Watson-Biemann Molecular Separator Compounds Untreated, gram BSA treated, gram Alcohols Linalool 10-6 10-9 Menthol 10-6 10-9 a-Terpineol 10-6 10-9 Aldehydes Geranial 1010-9 Neral 10-6 10-9 Esters Geranyl acetate IO-' 10-9 Linalyl acetate 10-7 10-9 Ketones Carvone 10-7 10-9 Menthone 10-7 10-9 Ether 10-9 10-9 1,8-Cineole Olefins a-Pinene 10-0 10-9 P-Pinene 10-9 I 0-9 a
Hitachi RMU-BE.
type G C column packings (8-11) suggested treatment with silanizing reagents (11, 12) to deactivate the porous glass surface. Bistrimethylsilylacetamide (12) proved to be satisfactory. EXPERIMENTAL
Gas Chromatograph y-Mass Spectrometry. A PerkinElmer 226 gas chromatograph connected to the built-in Watson-Biemann molecular separator (2) of a Hitachi RMU6E mass spectrometer was employed in GC-MS analyses. Samples were injected into the GC inlet/splitter with a Hamilton 7101 microsyringe. The sample was split approximately 100 to 1 and the smaller portion was chromatographed on a 100-foot X 0.010-inch i.d. open-tubular column coated with Apiezon L. Helium was used as the carrier gas at a constant G C inlet pressure of 15 psig. The terpenoids listed in Table I were obtained from Fritzsche Brothers, Inc., and the Givaudan Corp. Silanization. A clean, acid-washed Watson-Biemann separator was silanized in situ with all MS electronics turned off including filaments, heaters, and ion gages. The 1/16-inch 0.d. stainless steel transfer line (4-foot X 0.010-inch i d . ) leading from the GC to the separator was detached at the
(8) E. C. Homing, E. A. Moscatelli, and C. C. Sweely, Chem. Znd., 751 (1959). (9) W. L. Holmes and E. Stack, Biochim. Biophys. Acta, 56, 163 (1962). (10) D. M. Ottenstein, J. Gas Chromatog., 1, 11 (1963). (11) W. R. Supina, R. S. Henly, and R. F. Kruppa, J. Am. Oil Chemists' SOC.,43, 202A (1966). (12) J. F. Klebe, H. Finkbeiner, and D. M. White, J. Am. Chem. SOC.,88, 3390 (1966). VOL 40, NO. 4, APRIL 1968
841
aA
i
\
C
h Figure 1. Schematic diagram of Watson-Biemann separator with silanization attachment (a) septum, (6) capping nuts, (c) union, (4transfer line, (e) glass to metal seal, (f)oven, (g) fritted-glass tube, (h) mechanical pump inlet, (i) MS source inlet
GC end and fitted with a standard l/le-inch stainless steel SwageIok (No. 100-6-316) union. The open side of the union was sealed with a small GC silicone rubber septum (bored with a No. 2 cork borer) placed inside the capping nut. All MS and separator vacuum pumps remained on so that normal instrument vacuum was maintained. With the separator heated to 150' C and the capped transfer line at 25 O C, 0.10 ml of bistrimethylsilylacetamide (BSA), obtained from Applied Science Laboratories, Inc., was slowly injected by microsyringe through the septum into the transfer line. The BSA was drawn by vacuum into the porous glass separator for reaction with surface hydroxyl groups. Excess reagent and by-products were pumped out through both the frit pores and the MS inlet. After 30 minutes, the MS analyzer tube and traps were baked out for 2 hours and then pumped out for 2 days. RESULTS AND DISCUSSION
So much effort has been invested in recent years to deactivate the adsorptive properties of diatomite supports used in gas chromatography (8-11) that it seemed logical to attempt at least one standard procedure on the fritted-glass separator. Although followed carefully, a dimethyldichlorosilane procedure (11)failed to reduce significantly the loss of polar organic compounds in a new, acid-washed, commercial Watson-Biemann separator, when subsequently installed in the MS inlet. Rather than remove the separator from the glass inlet system in situ silanization with' bistrimethylsilylacetamide (12) was attempted, As Table I shows, the results proved satisfactory, even with rather labile terpenoids such as linalool and neral. Presumably, BSA deactivates the surface hydroxyl groups of the glass frit according to the reaction:
842
ANALYTICAL CHEMISTRY
O-Si(CH&
I CH&N-Si(CH&
(xs)
+ HO-Si-
11 (Surface)
-
0
I/
CH3C-N-Si(CHJ3
I
H
+ (CH&Si--O-Si-- I I
(Surface)
The by-product, trimethylsilylacetamide, may react analogously to form the by-product, acetamide. While it is not preferred that the MS be contaminated with excess reagent and by-products, they are innocuous compounds easily pumped out after a few hours' baking. Because BSA deactivated the separator chemisorption effectively on the first attempt, probably excess reagent was employed. Conceivably, a 10-fold or more reduction in dosage could still be effective. Replacement of BSA by the newer, more reactive bistrimethylsilyltrifluoroacetamide (13) may also be desirable.
RECEIVED for review December 18, 1967. Accepted January 2, 1968. Research supported by the National Aeronautics and Space Administration under Grants NgR-05-009-043 and NsG-541. (13) R. W. Zumwalt, D. L. Stalling, and C. W. Gehrke, Abs. C159, 154th Meeting, ACS, Chicago, Ill., September 1967.
