mass

---

Anal. Chem. 1989, 6 1 , 2126-2128

2126

glassy carbon. Line widths as small as 60 wm were obtained at a fairly high feed rate. It should be possible to make even narrower lines by using lower feed rates. Also worth pointing out is the versatility of the dicing saw for making cuts of different sizes. Changing the spacing between blade cuts merely requires the setting of a switch. In addition, gaps of different widths can be made by making multiple cuts with the blade or by using blades of different widths. The success achieved with the dicing saw in making smooth lines of glassy carbon in micrometer dimensions in such an easy fashion makes the technique noteworthy. The range of application can probably be extended to other electrode materials such as gold or platinum and to other sealing materials such as other epoxies or Pyrex (15, 16).

ACKNOWLEDGMENT We thank Mike Jackson and the Rochester Institute of Technology for the use of the dicing saw. Registry No. Fe(CN),4-, 13408-63-4;carbon, 7440-44-0.

LITERATURE CITED (1) Weisshaar. 5 3 , 1809

D.E.: Tallman, D.E.;Anderson, J. L. Anal. Chem. 1981,

(2) Falat, L: Cheng, H. Y. Anal. Chem. 1982, 5 4 , 2109. (3) Wang, J.: Frelha, B. A. J. Chromatogr. 1984, 298, 79. (4) Caudill, W. L.; Howell, J. 0.; Wightman, R. M. Anal. Chem. 1982, 5 4 , 2531. (5) Bond, A. M.; Henderson, T. L. E.;Thorman, W. J. Phys. Chem. 1986, 9 0 , 2911. (6) Thorman, W.; van den Bosch, P.; Bond, A. M. Anal. Chem. 1985, 5 7 , 2764. (7) Fosdick, L. E.;Anderson, J. L. Anal. Chem. 1986, 5 8 , 2481. (8)Fosdick, L. E.; Anderson, J. L.: Baginski, T. A,; Jaeger, R. C. Anal. Chem. 1986, 5 8 , 2750. (9) DeAbreu, M.; Purdy, W. C. Anal. Chem. 1987, 5 9 , 204. (IO) Moldoveanu. S.: Anderson, J. L. J. Electroanal. Chem. Interfacial Electrochem. 1985, 185, 239. (11) Anderson, J. L.: Ou. T. S.:Moldoveanu, S. J. Electroanal. Chem. Interfacial Electrochem. 1985, 196, 213. (12) Cope, D. S.;Tallman, D. E. J. Nectroanal. Chem. Interfaclal Electrochem. 1986, 205, 101. (13) Sparrow, E. M. National Advisory Committee for Aeronautics TN 3331, 1955. (14) Weber, S.G.; Purdy, W. C. Anal. Chim. Acta 1978, 100, 531. (15) Sambell. R. A. J.; Bowen, D. H.; Phillips, D. C. J. Mater. Sci. 1972, 7 , 663. (16) Sambell, R. A. J.: Briggs, A.; Phillips, D. C.; Bowen. D. H. J. Mater. S o . 1972. 7, 676.

RECEIVED for review March 8,1989. Accepted May 12,1989. This work was supported in part by the National Science Foundation under Grant CHE 8521200.

Tuning and Calibration in Thermospray Liquid Chromatography/Mass Spectrometry Using Trifluoroacetlc Acid Cluster Ions Steven J. Stout* and Adrian R. daCunha American Cyanamid Company, Agricultural Research Division, P.O. Box 400, Princeton, New Jersey 08540

INTRODUCTION Liquid chromatography/mass spectrometry (LC/MS) is a rapidly developing technique for the analysis of complex mixtures not amenable to gas chromatography/mass spectrometry (GC/MS) techniques ( I ) . Of the several approaches for interfacing liquid chromatography with mass spectrometry, thermospray (TSP) LC/MS appears to be the one best suited for the analysis of polar and labile organic compounds (2-6). One of the major drawbacks of T S P LC/MS is the need to install a separate ion volume which then must be tuned and calibrated. As discussed by Heeremans et al. (7), current methods of tuning TSP LC/MS suffer from serious shortcomings. Tuning on a solution of a particular analyte may not be applicable for the analysis of unknown compounds or when limited quantities of sample are available. Tuning solutions of poly(propy1ene glycol) (PPG), poly(ethy1ene glycol) (PEG) (8, 9),and sodium acetate ( I O ) offer a more universal method of tuning but result in rapid contamination of the ion source and memory effects. To overcome these shortcomings, Heeremans et al. (7) reported adding volatile acetic acid to an ammonium acetate eluent. Tuning on acetic acid-ammonia cluster ions from m/z 100 to lo00 was achieved with no ion source contamination. In this paper, we report on the use of trifluoroacetic acid to generate cluster ions for TSP LC/MS tuning and Calibration to m / z 4000 (the upper mass limit of our instrumentation) without ion source contamination. Additionally, the same tuning solution can be used for tuning in the negative ion mode of operation.

EXPERIMENTAL SECTION The experiments were performed with a TSQ-70 triple-stage quadrupole system equipped with a thermospray interface

* Author to whom correspondence should be addressed. 0003-2700/89/0361-2126$01.50/0

(Finnigan-MAT Corp., San Jose, CA). Operational parameters specific to the thermospray interface included the following: vaporizer temperature, 90 "C; aerosol temperature, 230 "C; repeller Torr. voltage, 70 V; mass spectrometer high vacuum, 2.7 X Solvent delivery was performed with an AB1 Kratos Spectroflow Model 400 LC pump. The mobile phase was CH,OH/H,O/trifluoroacetic acid (15/84.5/0.5,0.1 M ammonium acetate), flowed at 1.5 mL/min, and gave a back pressure of approximately 30 bar. ["NIAmmonium acetate (99% 15Nenriched) was obtained from ICON Services (Summit, NJ). Analytical reagent grade [14N]ammonium acetate was obtained from Mallinckrodt, Inc. (Paris, KY).

RESULTS AND DISCUSSION Since Heeremans (7)demonstrated tuning and calibration of a TSP LC/MS ion source to m / z 1000 using acetic acid (H0Ac)-ammonia cluster ions, we anticipated the greater mass of trifluoroacetic acid (TFA) would approximately double this mass range if TFA (molecular weight 114) exhibited the same level of cluster ion formation as acetic acid (molecular weight 60). As shown in Figure 1, this objective was achieved with cluster ions covering the mass range of m / z 100-2000. The predominant series of TFA-ammonia cluster ions corresponds to (TFA),(NH3),(NH4)+. This series starts with the (TFA)(NH,)(NH,)+ ion at m / z 149 and repeats in increments of 131 u (equivalent to TFA + NH,). This pattern is especially evident above m / z 400. Several ions below m / z 400 correspond to additions of NH3 to this series. Ions at m / z 192 (base peak) and m / z 175 correspond to (TFA)(HOAc)(NH4)+and m / z 192 - NHJ, respectively. The origins of the m / z 119 and 110 ions are uncertain at the present time. With switching to the high mass range of the instrument, the (TFA),(NH3),(NH4)+series is found to extend to m / z 4000 (Figure 2), the upper mass limit of the instrument. Unfortunately, extending the mass range is also accompanied by a degradation in mass resolution. 1989 American Chemical Society

ANALYTICAL CHEMISTRY, VOL. 61, NO. 18,

SEPTEMBER

15, 1989

2127

1 2

101 149

41 1

"1

Jxao

4

lor6

Y

t:

673

'OOll

SO4

W I-

tc -I

-i_L t

W

a

v)

2

w

,

T7

ai00

t

935

I

23'00

w

v)

z

2 1'00

c

7

-z

1001

2600

2800

3000

W

1 I-

"Is

4

3293

i 34r

1459

I

14'00

1600

3656 1852

3666

3817

394s

1983 1

1

The extent of NH3 incorporation in the cluster ions was verified by using a mobile phase containing 0.1 M 15NH40Ac in place of the 0.1 M I4NH4OAc. For every NH3 (or NH4+) contained in a TFA-ammonia cluster ion, the cluster ion is shifted 1 u higher on going from 14NH40Acto 15NH40Acin the mobile phase. The results of this experiment show the cluster ion series now starting with (TFA)(15NH3)(15NH4)+ at m / z 151 and repeating every 132 u (equivalent to TFA 15NH3)and support the proposed compositions of the TFAammonia cluster ions. Because of the electrophilic nature of the fluorine-containing TFA, the mobile phase was also evaluated for its potential for tuning TSP LC/MS in the negative ion mode. While being useful for negative ion tuning, the mobile phase did not work nearly as well as was found in the positive ion mode. As shown in Figure 3, approximately 90-95% of the ion current is carried by the (CF3COO-)(TFA)ion at m / z 227. This situation is dramatically different from the positive ion spectrum where no one ion carries the vast majority of the ion current. The cluster ions formed in the negative ion mode also do not follow as well-defined of a simple, repetitive pattern as found in the positive ion mode. As listed in Table I, after the (CF3COO-) (TFA)2ion at m / z 341, higher mass cluster ions incorporate varying amounts of NH3 with the TFA in the cluster. Finally, while negative ion clusters have been detected to mlz 2000, their responses diminish markedly above m / z 1000. While no ion source contamination has been detected or T S P LC/MS performance impaired after pumping as much as liter volumes of the tuning solution through the instrument, a solid precipitate did deposit in the vacuum tubing external to the heated mass spectrometer manifold. The chemical ionization (CH4)mass spectrum of this precipitate from the solid probe generated ions at m / z 18, NH4+,and m / z 115, the

+

227 341 455 472 489 603 734 865 996 1127 1144 1161 1275 1292 1406 1423 1554 1685 1816 1947

227 341 455 473 491 605 737 869 1001 1133 -0

1169 1283 1301 1415 1433 1565 1697

-

-

'Corresponding ion not detected in I6NHaOAcexDeriment. (M + H) ion of TFA, indicative of ammonium trifluoroacetate. The salt began volatilizing immediately upon entry into the ion source, volatilized predominantly a t 60 O C , and was completely vaporized by a solid probe temperature of 90 "C. The solid probe volatilization characteristics of ammonium trifluoroacetate explain why the material is volatile under vacuum in the heated ion source and mass spectrometer manifold but precipitates on reaching ambient temperatures in the

2128

ANALYTICAL CHEMISTRY, VOL. 61, NO. 18, SEPTEMBER 15, 1989

loOl

2271 I

m0

IX500

472 I

I

I

I

1

455,

341

1489

I

i

UI

I

b-

I

? b-

a

144

1:

I

n

I

5

I

I

14,06

1554

I

I I

I1007

12'00

I LL W 1400

18,l6 l ; T 5 It

i i

1700

1600

1947

1 1

1 1

I

1900

1

2100

m/z

Flgure 3. Negative ion mass spectrum of the thermospray LC/MS tuning solution.

vacuum tubing outside of the instrument. While giving an unsightly appearance in the vacuum tubing, the precipitate has never deposited in the ion source or metal pumpout lines connecting the ion source to the vacuum tubing and has never adversely affected subsequent TSP LC/MS performance. The precipitate can be readily removed by disconnecting the tubing and flushing with water when convenient. Although not tested, an alternative approach for eliminating the precipitate from the tubing would be to heat the tubing from the manifold to the rough pump isolation trap during TSP LC/MS operation. Reducing the concentrations of TFA and NH,OAc in the mobile phase to 0.1% and 0.05 M, respectively, has been found to reduce the rate of precipitate deposition without adversely affecting the intensities of the ion clusters. A further reduction of the TFA concentration to 0.01% decreased the intensities of the cluster ions by more than an order of magnitude. To test for potential memory effects caused by the tuning solution, the ion current was monitored after switching from

the tuning solution to a mobile phase of 18% H 2 0 in CH30H (0.1 M ammonium acetate). The tuning solution was monitored for 2 min after which the LC pump was switched to the new mobile phase. Following the drop accompanying the switchover, the response increased momentarily as residual tuning solution was pumped out of the connecting tubing between the thermospray interface and the LC pump. As the fresh mobile phase eluted, response to the tuning solution was eliminated over the next 2 min. The same pumpout characteristics were observed in both positive ion and negative ion modes. In conclusion, a tuning and calibration solution for T S P LC/MS has been developed that generates TFA-NH3 cluster ions to m / z 4000 in the positive ion mode. Not only is this mass range markedly higher than that provided by HOAcNH3 cluster ions (7), the predominant repetitive pattern of (TFA),(NH,),NH,+ is simpler than that for HOAc-NH, cluster ions where a fixed ratio of HOAc to NH3 is not maintained. The same mobile phase can be used for tuning and calibration in the negative ion mode but, unfortunately, does not work as well as in the positive ion mode. While temperature effects are expected to have some influence on cluster ion formation, they are expected to be similar to those previously reported for sodium acetate (10). Although a precipitate, determined to be ammonium trifluoroacetate, deposits in the vacuum tubing external to the instrument, no contamination of the ion source has been observed, and no degradation of TSP LC/MS performance in subsequent analyses has been detected. Better instrumental response to the high mass cluster ions is anticipated with a recently introduced 20-kV conversion dynode multiplier kit (11). Finally, the use of fluorine to increase the mass of the cluster ions without increasing their positive mass affect may prove useful for obtaining exact mass measurements when coupling highresolution mass spectrometry with TSP LC/MS.

LITERATURE CITED (1) Covey, T. R.: Lee, E. D.; Bruins. A. P.; Henion, J. D. Anal. Cbem. 1988,58, 1451A-1461A. (2) Blakley, C. R.: Vestal, M. L. Anal. Cbem. 1983, 55, 750-754. (3) Liberato, D. J.; Fenselau, C. C.; Vestal, M. L.; Yergey, A. L. Anal. Chem. 1983,55, 1741-1744. (4) Pilosof, D.; Kim, H. Y.; Dyckes, D. F.; Vestal, M. L. Anal. Cbem. 1984, 56, 1236-1240. (5) Vestal, M. L.; Fergusson, G. J. Anal. Cbem. 1985, 5 7 , 2373-2376. (6) Covey, T. R.; Crowther, J. B.; Dewey, E. A,; Henion, J. D. Anal. Chem. 1985,57, 474-481. (7) Heeremans, C. E. M.; van der Hoeven, R. A. M.; Niessen, W. M. A,: Tjaden, U. R.; van der Greef, J. Org. Mass Spectrom. 1989, 2 4 , 109-112. (8) Yang. L.; Fergusson, G. J. Annu. Conf. Mass Specfrom. Allied Top., 33rd 1985, 775-776. (9) Thermospray Accessory Operation and Service Manual, Finnigan-MAT Corp.: San Jose, CA, June 1987. ( 1 0 ) Robins, R . H.; Crow, F. W. RapidCommun. Mass Specfrom. 1988,2, 30-34 (11) Schoen, A,; Syka, J. E. P. Annu. Conf. Mass Specfrom. Allied Top., 36th 1988,843-2344,

__

RECEIVED for review March 31, 1989. Accepted June 22, 1989.