mass spectrometry

---

Anal. Chem. 1984, 56, 1533-1537

oxybutane, 106-88-7;2,3-epoxy-2-methylbutane, 5076-19-7;2,3epoxy-2,3-dimethylbutane,5076-20-0; 1,2-epoxypentane,1003-14-1; cis-2,3-epoxypentane,3203-99-4;cis-2,3-epoxyhexane,6124-90-9; 3,4-epoxy-4-methyl-L-pentanone, 4478-63-1; ethanol, 64-17-5; 2-propanol, 67-63-0;l,%propanediol,57-55-6; 2-hydroxypropanoic acid, 50-21-5; ethyl 2-oxopropanoate, 617-35-6; 2-hydroxy-1,4butanedioic acid, 6915-15-7; citric acid, 77-92-9; ethyl 3-oxobutanoate, 141-97-9; citronellal, 106-23-0;farnesol, 4602-84-0; limonene, 138-86-3;citral, 5392-40-5;pulegone, 89-82-7;cis-8,9epoxyheptadecane, 85267-93-2; trans-8,9-epoxyheptadecane, 85267-94-3; cis-9,10-epoxynonadecane,85267-95-4; trans-9,lOepoxynonadecane, 85267-96-5; methyl cis-9,lO-epoxyoctadecanoate, 2566-91-8; 1,2-(epoxyethyl)benzene, 96-09-3; trans-2,3-epoxy-3-phenylpropanal, 71403-94-6; trans-2,3-epoxy3-phenylpropanoic acid, 1566-68-3;cis-1,2-epoxy-1,2-diphenylethane, 1689-71-0; 1,2-epoxy-2-phenylpropane,2085-88-3; 1,2epoxy-1,1,2-triphenylethane,4479-98-5; 2,3-epoxy-3-phenylpropanoic acid, 5694-02-0;gheptadecene, 2579-04-6;9-nonadecene, 31035-07-1; 3-octanol, 589-98-0.

LITERATURE CITED (1) B i d , 6. A,; Beroza, M.; Collier, C. W. Science 1970, 170, 87-89. (2) HIII, A. S.;Roelofs, W. L. J. Chem. Ecol. 1981, 1 , 655-668.

1533

Uebel, E. C.; Schwarz, M.; Lusby, W. R.; Mlller, R. W.; Sonnet, P. E. Lloydia 1978, 4 1 , 63-67. Aplln, R. T.; Coles, L. Chem. Commun. 1967, 658-859. Tumllnson, J. H.; Heath, R. R.; Doolittle, R. E. Anal. Chem. 1974, 46,

1309-1312. Bierl, B. A.; Beroza, M.; Aldridge, M. A. Anal. Chem. 1971, 43,

636-641. Schwartz, D. P.; Welhrauch, J. L.; Burgwald, L. H. Anal. Chem. 1969, 4 1 , 984-966. Mizuno, G. R.; Elllson, E. C.; Chlapauit, J. R. Microchem. J . 1969, 14,

227-234. Attygalle, A. B.; Morgan E. D. Anal. Chem. 1983, 55, 1379-1384. Schwartz, N. N.; Blumberges, J. H. J . Org. Chem. 1984, 29,

1976-1979. Morgan, E. D.; Wadhams, L. J. J. Chromafogr. Sci. 1972, 10,

528-529. Hoffman, N. E.; Barborlak, J. J.; Hardrnan, H. F. Anal. Biochem. 1964,

9 , 175-179. Honda, S.;Fukuhara, Y.; Kakehi, K. Anal. Chem. 1978, 50, 55-59.

RECEIVED for review December 29,1983. Accepted March 6, and Nuffield Foundation provided lgg4* The British some financial assistance to A.B.A.

Open Split Interface for Capillary Gas Chromatography/Mass Spectrometry R. F. Arrendale,* R. F. Severson, and 0. T. Chortyk

Tobacco Safety Research Unit, Agricultural Research Service, United States Department of Agriculture, P.O. Box 5677, Athens, Georgia 30613 The successful application of capillary gas chromatography/mass spectrometry (GC/MS) relies heavily upon the interface system. Two alternatives exist a t this time. First, the capillary GC column can be directly connected to the mass spectrometer source (1,2). This method has many advantages; however, it also has several disadvantages, including: (1) all of the sample, including solvent, elutes directly into the mass spectrometer source, and (2) changing columns is laborious and time consuming, as the mass spectrometer vacuum system is vented with each column change. Direct connection has its greatest applicability for routine analysis, when capillary columns are changed infrequently. The second alternative for connecting GC columns to the mass spectrometer source is an open split interface (3). Some advantages of an open split interface include the following: (1)Capillary columns can be changed without an isolation valve or venting the mass spectrometer vacuum system. (2) The capillary column chromatography, in some modes of operation, is unaffected by the vacuum system. (3) The dynamic range of sample component concentrations can be greatly increased by purging portions of the highly concentrated sample, such as solvent components. (4) Chemical inertness of all surfaces, contacted by the sample prior to transfer to the mass spectrometer, is equivalent to that of the capillary GC column. Our laboratory analyzes a diverse array of complex samples, requiring a wide variety of capillary columns. Because capillary columns must be changed frequently, an open split interface was the best choice for our capillary GC/MS use. Although our interface design is functionally similar to commercially available or previously reported units (3-5), it has some unique characteristics. Some of these are as follows: (1) The interface body is completely inside the GC oven, improving accessibility, compared to interfaces housed in the heated interface oven. (2) The use of Pyrex glass allows precise

visual adjustment of fused silica (FS) capillary columns and the interface tubing [FS or vitreous silica (VS), Superox-4 deactivated] (6, 7). (3) The interface tubing (0.15 mm i.d.) can easily be inserted inside the FS capillary column (0.3 mm i.d.) for maximum chromatographic efficiency, without distortion by the vacuum system (‘‘the ideal mode”, discussed later). (4) Component parts are relatively inexpensive and assembly is simple and rapid. The platinum/iridium capillary tubing SGE isolation valve interface that came with our Hewlett-Packard 5985B GC/MS system was unsatisfactory, due to a high dead volume and to catalytic and adsorptive activities of the platinum/iridium and the glass-lined stainless steel (SS) tubing. Commercially available interfaces and those described in the literature are more complicated in design, are more expensive, and are composed of metal and/or glass-linked stainless steel (SS) tubing, making visual adjustment of the capillary column and interface tubing impossible. The design and construction of our interface and the modification of the mass spectrometer for capillary GC/MS applications are discussed. The application of the interface is illustrated by analyses of a variety of complex mixtures.

EXPERIMENTAL SECTION Materials. Vitreous silica (VS) capillary tubing (0.15 mm i.d.)

was obtained from Scientific Glass Engineering (Austin, TX), Superox-4from Alltech (Deerfield, IL),0-60 psi pressure regulator from Supelco (Bellefonte, PA), Swagelok lI8 in. X in. SS reducing unions, Swagelok ‘Il6 in. X ‘I4 in. SS reducing unions, Nupro “J”series miniature forged body shuboff valves, and Nupro “MG” series fine metering valve from Georgia Valve and Fitting Co. (Atlanta, GA), graphite and graphite-vespel ferrules from Scientific Glass Engineering (Austin, TX), and lI8 in. SS tubing (lI8in. 0.d. X 0.085 in. i.d.) and ‘Il6in. SS tubing (lis in. 0.d. X 0.03 in. id.) from Analabs (North Haven, CT). The Pyrex glass

This article not subject to US. Copyright. Published 1984 by the American Chemical Soclety

1534

ANALYTICAL CHEMISTRY, VOL. 56, NO. -

8,

JULY

1984

k

A

fl

HEWLETT-PACKARD 5 9 8 5 8

A

-/ -w ;,

;;

I/B'COPF€R

TUBING

TRANSFER LINE PROBE ASSEMBLY

~ M G SERIES ' FINE MASS SPECTROMFTER

METERING VALVE

INTERFACE OVEN

-118'

S S TUBING

8 VODIFIEC INTERFACE FOR CAPILLARY GClMS

, %

J SERIES

- R A N S F E R LIkE PROBE ASSEMBLI

U A S S SOECTPONETER

1Ab SSTUBING-

8

STAINLESS STEEL TUBING 1

15O*lm d FUSEO SILICA TUBING [SUPEROX - 4 DEACTIVA7EDI

l / l b S S TUBING

REDUCING UNION

Figure 1. (A) Hewlett-Packard 59858 isolation valve assembly for capillary G U M S applications. (B) Modified Interface for capillary G U M S applications after removal of the isolation valve assembly.

tubes for the interface body were manufactured by the University of Georgia Glass Shop. Modification of Capillary GC/MS Interface. The original Hewlett-Packard 5985B isolation valve assembly for the capillary GC/MS system is shown in Figure 1A. Two other inlets to the transfer line probe assembly, which are associated with the packed GC column GC/MS interface, were also present but not shown. The first step in modification of the capillary GC/MS interface was the removal of the isolation valve assembly. The isolation valve assembly was replaced with approximately 15 in. (38 cm) of in. SS tubing, which had been carefully rinsed with 500 mL each of benzene, methanol, water, and acetone and dried under vacuum. The SS tubing was bent as illustrated to allow it to protrude into the GC oven through the hole, where the isolation valve assembly had been. Connection of the SS tubing to the transfer line probe assembly was accomplished with a 1/8 in. SS Swagelok nut and a graphite ferrule. A 1/8 in. X '/le in. SS reducing union (with graphite ferrule) was connected to the other end, which protruded into the GC oven. Approximately 50 cm of 0.15 mm i.d. FS or VS tubing (Superox-4 deactivated) (6, 7) was slowly inserted through the reducing union and SS tubing until it reached the end of the transfer line probe assembly (Figure 1B). About 15 cm of the tubing remained in the GC oven for connection to the interface. Modification of the vacuum system was completed by sealing the 0.15-mm tubing at the reducing union with a 1/16 in. SS nut and a graphite or graphite-vespel ferrule. Thus, the 50 cm of 0.15-mm i.d. FS or VS tubing extends from the GC oven into the mass spectrometer source and restricts flow into the mass spectrometer. With the mass spectrometer system under vacuum and the end of the interface tubing exposed to atmospheric pressure, the mass spectrometer source pressure should read approximately 5 X 10" torr. Construction and Installation of the Open Split Interface. The basic design of our interface is shown in Figure 2. The interface body consists of a 6 cm piece of 6.35 mm X 2 mm i.d. Pyrex tubing which has been drawn down by a glassblower to a 0.5-1.00 mm i.d. constriction at the center. This constriction assisted in the alignment of the interface tubing (0.15 mm i.d. FS or VS) and the FS capillary GC columns. The Pyrex glass body was mounted between two modified Swagelok (1/4 in. X '/le

FUSEO SILICA G C COLUMN-

T

n

Flgure 2. Laboratory-constructed open split interface for Capillary GC/MS applications.

in. SS) reducing unions, using graphite vespel ferrules. The reducing unions were previously modified by drilling a hole in one of the flat faces of the union body and attaching (by silver-soldering) a l-m length of '/I6 in. SS tubing. The interface was positioned in the GC oven and the '/I6 in. SS tubing ends were extended through the top of the GC oven (through one of several holes). The interface may be supported by a clip connection to the GC oven wall or it may be held in position by the '/la in. SS tubing. Purge gas (helium) to the interface was supplied through the bottom reducing union by connecting the '/I6 in. SS tubing to an exterior shut-off valve, connected, in series, to a fine metering valve, a 0-60 psi pressure regulator, and a helium tank (Figure 2). The purge gas effluent exited the interface through the upper reducing union and the attached '/I6 in. SS tubing to an exterior shut-off valve. The capillary GC column and the fused silica interface tubing were inserted into the bottom and top of the interface, respectively, and were sealed with in. SS nuts and graphite or graphite-vespel ferrules. Leak Testing the Interface System. The portion of the interface system shown in Figure 1B should be leak-tested first. This is easily accomplished by sealing the end of the 0.15 mm i.d. FS or VS interface tubing which extended into the GC oven (Figure 1B) by inserting the end into a split septum. The source pressure and air counts should be approximately as obtained previously with the Hewlett-Packard 5985B isolation valve assembly, with the valve closed (Figure 1A). If leaks occur, they may be located by applying methanol or some other volatile solvent to specific connection points and checking for characteristic ions. Next, the interface itself can be checked by sealing the bottom (capillary GC column entrance) with a Swagelok female '/le in. SS or brass plug. The 0.15 mm i.d. FS or VS interface tubing from the mass spectrometer was connected to the interface and the "J" series miniature forged body shut-off valves were closed to provide a sealed vacuum system. The vacuum system

ANALYTICAL CHEMISTRY, VOL. 56, NO. 8, JULY 1984

was allowed to pump down and then the source pressure and air counts were observed and leaks, if any, were found and sealed. The number of air counts and source pressure should meet the specifications of the GC/MS system before modification. The source pressure for our modified system with 1/16 in. SS plug sealing off the interface where the column is normally connected was (1-2) x IO-’ torr.

RESULTS AND DISCUSSION The operation of our interface in the “ideal mode” has been the most useful, because it results in chromatographic integrity, low background, and high sensitivity mass spectrometry. In this ideal mode, the end of the 0.15 mm i.d. interface tubing was placed just within the 0.3 mm i.d. GC capillary column, inside the interface assembly. If a smaller diameter GC capillary column (0.20-0.25) was used, then a few inches of 0.3 mm i.d. FS tubing was connected to the interface and the FS capillary column was attached to the 0.3 mm i.d. FS tubing via a butt connection, using a commercial fitting for this purpose. The interface was continuously purged with helium (10-20 mL/min) during analyses. No solvent purging or removal of other highly concentrated sample components occurs in this mode of operation. However, capillary column flow may be more or less than the mass spectrometer intake. If the GC column flow was greater, then the excess was purged with the helium purge gas and any memory effect from excess GC column effluent was eliminated. If the GC column effluent was less than the mass spectrometer intake, then the mass spectrometer will make up the difference with the helium purge gas. In the ”ideal mode” of operation, the integrity of the chromatography was preserved and the mass spectrometer background remained low (yielding high sensitivity). The combination of capillary GC efficiency and mass spectrometer sensitivity is equivalent to that obtainable with direct connection of the capillary GC column to the mass spectrometer source, when the capillary GC column flow is equal to or less than the interface tube flow. However, the interface allows capillary GC columns to be changed without venting the vacuum system. The interface may also be sealed with the two externally located shut-off valves to provide a capillary GC/MS arrangement similar to direct connection of the GC column to the mass spectrometer source. However, leaks in the interface system could increase background and distortion of the capillary chromatography by the vacuum system would also increase. In the other modes of operation, which we refer to as the “solvent-purged modes”, the GC column and the 0.15 mm i.d. interface tubing ends are aligned inside the interface body, so that they almost touch. The amount of solvent or other highly concentrated components entering the mass spectrometer can be reduced by a high flow (15-20 mL/min) of purge gas (helium) through the interface. After the undesirable components have been purged away, interface flow options can then be varied, depending on the flow characteristics of the GC column. When the capillary GC column flow is less than or equal to the intake of the mass spectrometer, a low flow (1-2 mL/min) of helium purge gas could be applied through the interface or both shut-off valves could be shut to provide a closed vacuum system. However, a low flow of helium through the interface during the analysis prevents peak broadening and/or destortion of the capillary gas chromatography. The closed vacuum system provides maximum sensitivity for trace components. If the capillary GC column flow is more than the intake of the mass spectrometer, then the shut-off valve on the inlet side of the interface can be closed, after the solvent purge and the excess column effluent can exist through the vent. The maximum amount of sample reaches the mass spectrometer without distortion of the gas chromatographic conditions. A low flow of purge gas may also be maintained through the interface

1535

WCYLETI-PICKOID IPOLITION W L V F

INlERTPCF

A

OPEN SPLIl

INTERFPCF

C

1

I

Il

1 P

Kx

0

io

io

L

io

TIME (MINI

Flgure 3. Comparison of the separation of a sample containing fatty acid methyl esters (FAME) on a fused silica Silar 1OC Capillary column: (A), TI chromatogram obtained with the Hewlett-Packard isolation valve interface; (B) T I chromatogram obtained with our laboratory-constructed interface; (C) FID capillary GC chromatogram.

during the analysis; however, sensitivity will again be reduced. The separation of a sample containing fatty acid methyl esters (FAME) on a FS Silar 1OC capillary column illustrates the improvement in the total ion current GC/MS chromatogram (TI chromatogram) achieved with the laboratoryconstructed open split interface (Figure 3). Capillary GC conditions for these analyses were as follows: 140-220 “C at 4 “C/min; 28 cm/s He flow; split injection mode; 25 m X 0.25 mm i.d. Figure 3A shows the TI chromatogram obtained with the original Hewlett-Packard isolation valve interface while Figure 3C shows a chromatogram of this FAME sample obtained with a flame ionization detector (FID) (49).Figure 3B shows a TI chromatogram of the same sample after modification of the GC/MS system with our laboratory-constructed interface. Peak shapes were well defined and the appearance of the TI chromatogram (Figure 3B) was similar

1536

ANALYTICAL CHEMISTRY, VOL.

1'819

56, NO. 8, JULY 1984 27

15

m

m

t26

5

4

7

6

~-

In

9

8

I1

I?

1F

I*

1:

:5

E

C18

c20

m

2

3

4

4

1

7

8

9

I0

I1

I2

13

14

15

TIMEIMINI

Figure 4. T I chromatogram of a mixture of underivatized, volatile fatty acids on a fused silica SP-1000 bolumn: Peaks: 1, acetic acid: 2, pripionic acid: 3, isobutyric acid: 4, n-butyric acid; 5, isovalerlc and 2-methylbutyric acids: 6, valeric acid: 7, methylvaleric acid; 8, caproic acid: 9, heptanoic acid. (B) T I chromatogram of an actlvity standards mixture separated on a fused silica Superox-4 WCOT column.

to the GC/FID chromatogram (Figure 3C). Comparison of Figure 3A and Figure 3C revealed major distortions in the original TI chromatogram (Figure 3A). Peak broadening, peak tailing, adsorption of components, and discrimination between low and high boiling components were the result of high surface activity and void volume of the isolation valve interface. These data are not surprising when one considers that the void volume of the isolation valve assembly (Figure 1A) is several hundred times that of the interface tubing. The low surface activity of the interface is illustrated by the data in Figure 4. Figure 4A shows a T I chromatogram of a standard mixture of underivatized fatty acids separated on a FS SP1000 capillary column (6). Conditions for this analysis were as follows: 34 m X 0.2 mm i.d.; temperature program, 120-220 "C at 6 "C/min; 25 cm/s He flow; split injection mode. The separation of a standard activity mixture (Figure 4B) also illustrates the low surface activity of the interface (6). Capillary GC conditions for this analysis were as follows: 25 m X 0.3 mm i.d.; temperature program, 80-220 OC at 4 OC/min; 35 cm/s He flow; split injection mode. When analyzed on the original GC/MS system, the free acids and the polar components of the activity mixture were almost completely adsorbed by the isolation valve GC/MS interface. The chromatograms in Figure 5 illustrate the excellent chromatographic characteristics of our interface. Figure 5A shows the T I chromatogram of the capillary GC/MS analysis of an aliphatic hydrocarbon wax fraction from tobacco separated on a FS SE-54 capillary column (7, 10). Note the excellent peak shapes and the separation of the iso-, anteiso-, and straight-chain hydrocarbon isomers. Capillary GC conditions for this analysis were as follows: 34 m X 0.2 mm i.d.; temperature program, 120-220 'C at 4 OC/min; 28 cm/s He flow; split injection mode. Figure 5B depicts the capillary GC/MS analysis of a standard mixture of FAME. The ex-

i

,

TIME(MIN1

m

B

TIMEIPlINI

l'z 15 l k l b 218 2'2 2'4 2 k 2'8 38 3 2 3k 36 38 40 42 44 46 48 5 8 52 5 4 5 6 68

chromatogram of a sucrose ester (Me,Si) fraction from tobacco separated on a fused silica SE-54 WCOT column. (8) T I chromatogram of a wax ester fraction from tobacco separated on a fused silica SE-54 WCOT column. Figure 6. (A) T I

cellent chromatographic characteristics of the interface allowed separation of closely eluting saturated (CIS)and unsaturated

Anal. Chem. 1984,56, 1537-1539

(Cis:.) isomers, similar to GC/FID analyses (6,II).Capillary GC conditions for this analysis were as follows: 25 m X 0.3 mm i.d.; temperature program, 150-230 "C at 4 "C/min; 30 cm/s He flow; split injection mode. The ability of our interface to transfer molecules of higher molecular weight to the mass spectrometer source is illustrated in Figure 6. The first example (Figure 6A) is a T I chromatogram of the trimethylsilyl ethers (Me,Si) of a sucrose ester fraction from tobacco separated on a FS SE-54 capillary column (7,9). Capillary GC conditions for this analysis were as follows: 25 m X 0.3 mm i.d.; temperature program, 225-280 "C at 2 "C/min; 35 cm/s He flow; split injection mode. The sucrose esters are composed of sucrose molecules, containing an acetic acid moiety and three other volatile acids esterified to the glucose portion (7,9). These compounds vary in molecular weight from 900t~ 1OOO. Another group of compounds that was successfully characterized was the wax esters group of tobacco. These compounds are composed of fatty acids esterified with fatty alcohols. A T I chromatogram, showing the separation of a wax ester fraction, is shown in Figure 6B (7,9). Mass spectral data were obtained on these wax esters, up to the CS2ester. Capillary GC conditions for this analysis were as follows: 3.75 m X 0.3 mm i.d.; temperature program, 200-300 "C at 2 "C/min; 100 cm/s He flow; split injection mode. The temperatures of the capillary GC column and the interface body are identical during temperature programming since both are located in the GC oven. The interface oven and the transfer line probe assembly are independently temperature controlled and were maintained at 280 "C. Transfer of components to the ion source is facilitated by the transfer line probe assembly (Figure 1B) which fits directly into the ion source body and provides temperature control for the

1537

interface tubing all the way to the ion source. Studies have shown that the temperature of the interface tubing, especially the portion that protrudes into the ion source, has a great influence on the transfer of components to the ion source (12). Investigations using our open split interface showed that relatively high molecular weight, low volatility compounds are transferred to the ion source with little or no losses in GC resolution (Figures 5A and 6).

LITERATURE CITED Ten Noever de Braun, M. C.; Brunnee, C. Z . Anal. Chem. 1967, 229, 321. Henneberg, D.; Schomburg, G. Z . Anal. Chem. 1964, 221, 5 5 . Henneberg, D.; Henricks, V.; Schomburg, G. Chromatographla 1975, 8, 449. Wetgel, E.; Kuster, Th.; Curtius, H.-Ch. J . Chromafogr. 1962, 239, 107. Stan, H.; Abraham, B. Anal. Chem. 1976, 5 0 , 2161. Arrendale, R. F.; Severson, R. F.; Chortyk, 0.T. J . Chromafogr. 1983, 254, 63. Arrendale, R. F.; Severson, R. F.; Chortyk, 0.T. HRC CC, J . Hlgh Resolut Chromatogr Chromatogr. Commun. 1983, 6 , 436. Arrendale, R. F.; Chapman, G. W.; Chortyk, 0.T. J . Agric. food Chem. 1983, 3 1 , 1334. Severson, R. F.; Arrendaie, R. F.; Chortyk, 0. T.; Johnson, A. W.; Jackson, D. M.; Gwynn, G. R.; Chaplin, J. F.; Stephenson, M. G. J . Agric. food Chem. in press. Severson, R. F.; McDuffle, K. L.; Arrendale, R. F.; Chortyk, 0.T. Beitr. Tabakforsch. 1981, 1 1 , 27. Arrendale, R. F.; Severson, R. F.; Chortyk, 0.T. J . Chromatogr. 1981, 208. 209. Henneberg, D.; Henricks, U.; Husmann, H.; Schomburg, G. J . Chromatogr. 1978, 167, 139.

.

.

RECEIVED for review December 8, 1983. Accepted February 17,1984. Reference to a company or product name does not imply endorsement or recommendation by the U.S. Department of Agriculture.

Device for Studying Heat-Induced Surface Reactions by Infrared Spectrometry P. J. Zanzucchi* and W. R. Frenchu

RCA Laboratories, Princeton, New Jersey 08540 By polarized infrared reflectance at oblique angles of incidence, infrared spectra of thin films on metals can be obtained. In 1959, following from the work of Fry (I),Francis and Ellison (2)reported that absorption-like spectra of films, which were as thin as a monolayer, could be obtained by the oblique multiple reflection of polarized infrared radiation from metal surfaces. Since that time, the characteristics of the technique, which is referred to in the literature as differential-reflection or absorption-reflectionspectrometry, have been studied in detail. The reflectivity equations describing the method have been analyzed by McIntyre and Aspnes (3)in their broad analysis of the effects of thin film6 on the reflectance properties of (metal) surfaces. Greenler (4) has reported on experimental considerations and Allara et al. (5) have evaluated the degree of band distortion that occurs. Applications of the oblique reflectance technique are cited in the work of Boerio and Chen (6) and recently Yoshida and Ishida (7)used the technique to study the orientation of imidazoles on copper surfaces. Here we report on the design and application of a temperature variable device for studying heat-induced surface reactions by infrared oblique reflectance spectrometry. In selecting applications, the emittance of the sample must be considered. This was noted by Poling (8)in his study of metal oxidation. For transparent materials, McMahon (9) has shown that sample emission is not easily defined and, in part, depends on the transmittance of the material. For these ma-

terials, typically organics, the effects of sample emission are less severe and oblique reflectance measurements can be made, directly, at elevated temperatures. The applications for a temperature-variable device are directly related to the characteristics of oblique reflectance spectrometry. These are (1)detection, on metal surfaces, of films as thin as a monolayer, (2) detection of molecular orientation which has recently been discussed by Rabolt et al. (IO),and (3) wide application to organic materials provided a metal substrate can be utilized. To demonstrate use of the method and device, a study of the temperature-dependent changes in organic flux (water white rosin), on polished metals such as copper or aluminum, is reported here. An apparent temperature-dependent dimer-monomer equilibrium is found to exist. The thermally induced properties of flux are of technical interest in the electronics industry where water white rosin, and related fluxes, are widely used on copper, and other metal surfaces, prior to soldering (11).

EXPERIMENTAL SECTION The temperature-variable device is shown in Figure 1. The drawing is not to dimension. The device consists of an iron block (21/4 in. X 1 in. X 1/2 in.) (Figure la) with a press-fit cartridge heater (Tempco H80, HD06-0200N, 150 W/120 V, 0.373 in. 0.d. X 2 in) (Figure lb). The metal piece for study, e.g., copper plate (23/sin. X 1 in. X 0.025 in. nominal, polished) (Figure lg), is clamped t o the iron block with a piece of shaped iron, (Figure

0003-2700/84/0356-1537$01.50/00 1984 American Chemical Society