(19) T. Su and M. 1.Bowers, J. Chem. Phys., 58, 3027 (1973). (20) G. Gioumousis and D. P. Stevenson, J. Chem. Phys., 29,294 (1958). (21) F. H. Field, J. Am. Chem. Soc., 90, 5649 (1968).
RECEIVEDfor review December 12, 1975. Accepted February 11, 1976. R.V.H. is a National Science Foundation predoctoral Fellow and J.L.B. is a Camille and Henry Dreyfus
Teacher-Scholar (1971-1976). This research has been supported in part by the United States Energy Research and Development Administration under Grant No. AT(043)767-8 and the Ford Motor Company Fund for Energy Research administered by the California Institute of Technology.
Chemical Ionization Mass Spectrometry of Chromium Tris1,1,l-trifluoro-2,4-pentanedionate and Other Transition Metal p-Di ketonates S. R. Prescott, J. E. Campana, P. C. Jurs, and T. H. Risby' Department of Chemistry, The Pennsylvania State University, University Park, Pa. 16802
A. L. Yergey Scientific Research lnstruments Corporation, Baltimore, Md. 2 1207
The chemical ionization mass spectra for the following chelates are reported: manganese blsacetylacetonate (Mn( a ~ a c ) ~Ni( ) , a c a ~ ) Cu( ~ , acac)p, Z n ( a ~ a c ) ~ Cr(acac)3, , Fe(acac)3, C o ( a ~ a c ) ~Ru(acac)3, , Rh(acacI3, manganese bis-l,l,l-trlfluoro-2,4-pentanedionate ( M ~ ~ ( t f a ) ~Nl(tfa)2, ), C ~ ( t f a ) ~Z, ~ ~ ( t f a )VO(tfa)2, ~, Cr(tfa)3, Fe(tfa)3, Co(tfa)3r Ru(tfa)3, Rh(tfa)3, nickel bis-l,l,7,7-tetramethyl-4,6-heptanedionate ( Ni(thd)2), Cu(thd)2, Z ~ ~ ( t h d )VO(lhd)2, ~, Also a system Cr(thd)3, Fe(thd)3, M f ~ ( t h d ) ~and , C~(thd)~ . is described whereby solutions can be injected directly into the ionization source of a chemical ionization mass spectrometer (CIMS). Preliminary detection limits of 1 part per billion (ppb) for Cr(tfa)3 in toluene were obtaiqed with the direct injection system.
The natural abundance of chromium in igneous rocks in the form of chromite is of the order of 0.02%,and soils have been found to contain from traces to 2.4%of chromium. As a direct result of this natural level of chromium, most plants contain chromium in the range of 100-500 mg/kg of dry matter. Chromium is used industrially in the following products or processes: alloys, plating, catalysts, tanning, paints, mordants, and fungicides. The average total daily intake of 30-40 mg of chromium by man is derived from the following sources: food (30-100 mg), water (0-40 mg), and air (0.3 mg). This quantity of chromium more than satisfies this essential trace element requirement, since chromium I11 is known to be involved in carbohydrate and fat metabolism. As a result, excess dietary chromium is excreted in the feces and some of the absorbed chromium is returned to the intestinal tract in bile or excreted in the urine. Traces of chromium have been thought to be stored in an insoluble form in the lungs ( I ) . Various studies have been made in order to ascertain the toxic effects of exposure to chromium. It has been observed that the skin, nasal membrane, lungs and kidneys can be more prone to carcinomas with high levels of chromium (2-8). As a result of this interest in low-level concentrations of chromium, extensive use of electron impact (EI) mass spectrometry has been made via complexation with
6-diketones (9-20): All these analyses used the Cr(P-diketonate)Z+ ion which is the base peak in the E1 mass spectrum. Recent studies have shown that chemical ionization mass spectrometry (CIMS) for trace metal analysis is attractive because the resulting mass spectra are simple (21,22).This present paper is intended to show the analytical application of CIMS for the analysis of chromium 6-diketonates by direct injection of solutions of chromium tris-l,l,l-trifluoro-2,4-pentanedionate(Cr(tfa)s) in toluene into the source of the CI mass spectrometer. Also the mass spectra of a series of other metal chelates are reported.
EXPERIMENTAL Chelate Preparation. The 2,4-pentanedionates (acac), the l,l,l-trifluoro-2,4-pentanedionates (tfa), and the 2,2,6,6-tetramethyl-3,5-heptanedionates (thd) of VO(II), Cr(II), Mn(II), Mn(III), Fe(III), Co(III), Ni(II), Cu(II), Zn(II), Ru(II1) and Rh(II1) were prepared by methods which have been described previously (23, 2 4 ) . Apparatus. The CI mass spectra were obtained using a scientific Research Instruments Corporation BIOSPECT system with either isobutane or methane as the reagent gas, which has been described previously (21, 2 2 ) . Aliquots of known solutions of the metal chelates in toluene were evaporated onto the direct insertion probe, and the probe was inserted directly into the ionization source. The following are the source temperatures used to sublime the chelates into the mass spectrometer: 130-180 O C (acac), 130 OC (tfa), and 130 O C (thd). All other conditions were identical with those previously reported (21). The mass-to-charge ratios of the various peaks in the mass spectra were determined by the mass marker which has been calibrated with methyl stearate. The sensitivity studies for Cr(tfa)3 were made in the following manner: the quadrupole mass filter was tuned to unit resolution in the region of the protonated molecular ion [Cr(tfa)3H+](512 amu) and then it was set to repeatedly scan the mass region 507 to 517 amu in 200 ms.
RESULTS AND DISCUSSIONS CI Mass Spectra. In order that the CI mass spectra could be interpreted, a similar computation of relative intensities as has been reported previously (21) was performed, and the results are summarized in Tables 1-111. These results are based on the contributions due to the natural isotopic abundances of the metal isotopes, H2, C13, ANALYTICAL CHEMISTRY, VOL. 48, NO. 6, MAY 1976
829
512
I
I
I
450
500
a. m. u.
Figure 1. The methane CI mass spectrum of Cr(tfa)s ion current vs. amu. Source pressure 1.0 Torr, source housing 1 X Torr, analyzer section 5 X Torr, source temperature 130 OC
and Ols a t a particular mass-to-charge ratio. These tables illustrate the ease with which mixtures of the chelates may be analyzed. The CI mass spectra were obtained for each of the metal chelates contained in Tables 1-111 using either isobutane or methane as the reagent gas, and in every case the calculated intensities agreed with the observed intensities for the protonated molecular ions. An example of a CI mass spectrum for Cr(tfa)s using methane as the reagent gas is shown in Figure 1. I t can be seen from this spectrum that in the
region of 520 to 400 amu, apart from the parent ion, there is a very minor fragment ion which corresponds to the loss of 20 amu (HF). This fragment ion was observed for all the chelates based on H(tfa) with methane. Isobutane showed this fragmentation with Fe(III), VO(II), and Zn(I1). The chelates based on H(acac) and H(thd) showed no fragmentation within the region corresponding to the loss of a ligand and the protonated molecular ion. The loss of a 20 amu neutral fragment can be more readily explained than the earlier results (22) which showed losses of 18 amu for the lanthanide tris-1,1,1,2,2,3,3-heptafluoro-7,7-dimethyl4,6-octanedionates (Ln(fod)3). We therefore conclude that our earlier results were probably caused by the presence of an impurity in the ligand (H(fod)) which did not contain the perfluoropropyl group. Sensitivity Studies of Cr(tfa)a. A preliminary sensitivit y study was made using solutions of Cr(tfa)3 in toluene (10 ppb-10 ppm by weight in terms of Cr). This study showed that it was extremely difficult to use the direct insertion probe as a method of sample introduction since the chelate was lost irreproducibly when the solvent was evaporated. This problem was found to be present for any compound which is either volatile or easily sublimed. A new method of sample introduction based on a heated gas chromatographic injection port was developed. This
Table I. Calculated Protonated Molecular Ion (M(acac),H+) Intensities for Various 2,CPentanedionates m le 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270
Mn(I1)
88.6 10.1 1.2 0.1
Ion intensities Ni(I1)
60.3 6.9 24.1 3.8 3.4 0.4 1.0 0.1
Cu(I1)
61.2 7.0 28.2 3.2 0.4
Zn(I1)
Fe(II1)
Co(II1)
Ru(II1)
Rh(II1)
4.6 0.8 1.7 10.9 12.4 16.3 29.1 4.9 16.2 2.7 0.4
83.4 14.2 2.2 0.2
43.3 4.9 25.2 6.5 17.2 2.0 0.8 0.1
348 349 350 351 352 353 354 35 5 356 357 358 359 360 394 395 396 397 398 399 400 401 402 403 404 830
Cr(II1)
ANALYTICAL CHEMISTRY, VOL. 48, NO. 6, MAY 1976
3.6 0.6 70.0 20.0 5.2 0.7 0.1
4.9 0.8 76.5 14.9 2.6 0.3
83.4 14.3 2.2 0.2
system consists of a heated injection port (Hamilton, Model 86810) mounted directly on the vacuum housing, perpendicular to the CI source. This port is connected to the ionization source by silanized glass lined stainless steel tubing (Supelco Incorporated, /B' in. 0.d.). This tubing was continuous from the injection port septum to the source of the mass spectrometer in order to reduce any possibility of decomposition of the sample. The inlet, between the CI source and injection block, is wrapped with insulated heating wire to provide uniform temperature throughout. Electrical connections between the inlet and a power supply are made through a high vacuum feedthrough. Temperatures are controlled and monitored independently for the injection port, inlet, and CI source. A reagent gas inlet and appropriate metering valves are attached to the injection port. When solutions are injected into the mass spectrometer, the signal observed is inversely proportional to the gas flow through the injection port system. Thus, for signal optimization, reagent gas is leaked through the direct injection system. The source is brought to 1 Torr by adding reagent gas through the conventional source gas inlet. Few instrumental modifications are necessary, since the manufacturer already provides the vacuum housing and CI source with the appropriate ports and hardware for coupling the instrument to a gas chromatograph.
1. TIME
Figure 2. Two successive plots of the total ion current repetitively scanned from 507 to 517 amu (0.2 s/scan) as a function of time 100 ppb solutions of Cr(tfa)s in toluene were injected via the direct injection port. Total ion current vs. time. Source pressure 1.0 Torr, source temperature 130 OC. injection block and inlet 140 OC
Aliquots (0.15 111) of solutions of Cr(tfa)s in toluene (10 ppb-10 ppm in terms of Cr) were injected through this system, and the ion current was measured over the range of 507 to 517 amu as a function of time with an oscillographic strip chart recorder. The resulting area under the ion current envelope curve was measured with a planimeter, and this area reflected the number of ions in this 10 amu mass region. Two typical replicate injections of a 100 ppb solution of Cr(tfa)s in toluene are shown in Figure 2. The direct injection port suffers from problems when either solvents other than alkyl or aryl hydrocarbons are in-
Table 11. Calculated Protonated Molecular Ion M(tfa),H+ Intensities for Various l,l,l-Trifluoro-2,4-pentanedionates I o n intensities
nile
Mn(I1)
362 363 364 365 366 367 368 369 370 371 372 373 374 37 5 376 377 378
88.6
Ni(I1)
Cu(I1)
Zn(I1)
VWI)
43.3 4.9 25.3 0.2 6.5 17.2 1.9 0.8
0.2 88.3 10.0 1.4
Cr(W
Fe(II1)
RU(III)
CO(II)
Rh(II1)
10.0
1.2 0.1
60.3 6.8 24.2 3.8 3.4 0.4 1.0 0.1
61.2 7.0 28.3 3.2 0.4
0.1
0.1
510 511 512 51 3 514 515 516 517 518 519 520 521 522 556 557 558 559 560 561 562 563 564 565 566
3.6 0.6 70.0 19.8 5.1
0.7 0.1
4.9 0.8 76.6 14.8 2.6 0.3
83.5 14.2 2.1
0.2 4.6 0.8
1.7 10.9 12.4 16.3 29.1 4.9 16.2 2.7 0.4
83.5 14.2
ANALYTICAL CHEMISTRY, VOL. 48, NO. 6, M A Y 1976
2.1
0.2 831
Table 111. Calculated Protonated Molecular Ion (M(thd),Hf) Intensities for Various
2,2,6,6.Tetramethyl-3,5-heptanedionates Ion intensities m/e
Ni(I1)
425 426 427 428 429 430 431 432 433 434 435 436 437 438
52.6 13.2 22.3 6.3 3.6
600 601 602 603 604 605 606 607 608 609 610 611 61 2
0.8
0.9 0.2
Cu(I1) Zn(I1) VO(I1) Cr(II1) Fe(II1) Mn(II1) Co(II1)
from 10 ppb to 10 ppm with a correlation coefficient of 0.996, and the limit of detection was of the order of 1 ppb (8 X g of Cr). Currently, the CIMS is in the process of being computer controlled. This coupling will enable signal averaging to be improve this detection limit and done which fast analysis of mixtures of metal chelates. LITERATURE CITED
53.4 13.4 25.9 37.8 6.2 9.5 0.9 22.9 0.2 0.1 8.7 77.0 16.0 19.4 3.8 3.1 1.0 0.4 0.2 2.9 1.1
57.2 28.0 8.7 1.9
0.2
4.0 1.5 68.0 62.6 25.6 25.0 5.5 5.9 0.9 1.0
68.0 25.6 5.5 0.9
jected or when injection sizes greater than 0.2 111 are used. If these limitations are exceeded, then signals are observed which are independent of mass in addition to the expected peaks which are due to protonated solvent ion and solvent fragment ions. This could be due to neutrals or long-lived excited species passing through the quadrupole mass filter to be ionized by collision immediately prior to the ion multiplier. This phenomenon is not unexpected since 0.5 111 of toluene will produce a partial pressure in the source of about 0.25 Torr which is comparable in pressure to the reagent gas. The relationship between the area under the ion current envelope and the concentration of the Cr(tfa)s was linear
(1) L. Fishbein, "Chromatography of Environmental Hazards", Voi. 11, Elsevier, 1973, Chapter 2. (2) G. E. Morris, AMA Arch. Dermatol., 78, 612 (1958). (3) A. M. Baetter, C. Damron, and V. Budacz, Arch. lnd. Health, 20, 136 (1949). (4) T. F. Manguso, lnd. Med. Surg., 20, 393 (1951). (5) H. P. Brinton. E. S. Frasier, and A. L. Koven, U S . Public Health Serv. Publ., No. 87, 385 (1952). (6) A. M. Baetter, Arch. lnd. Health, 2, 487 (1950). (7) W. Machle and F. Gregorius, U.S., Public Health Serv. Publ., No. 63, 1114(1948). (8) P. L. Bidstrup, Br. J. lnd. Med., 8, 302 (1951). (9) C. G. MacDonaid and J. S.Shannon, Aust. J. Chem., 19, 1545 (1966). (10) A. L. Clobes, M. L. Morris, and R. D. Koob, J. Am. Chem. Soc., 91, 3087 (1969). (11) J. L. Booker, T. L. Isenhour, and R. E. Sievers, Anal. Chem., 41, 1705 (1969) (12) ?.H.-Risby and L. G. Sanchez, unpublished results, 1970. (13) R. Belcher, J. R. Majer, R. Perry, and W. I. Stephen, Anal. Chim. Acta, 45, 451 (1968). (14) B. R. Kowalski, T. L. Isenhour, and R. E. Sievers. Anal. Chem., 41, 998 (1969). (15) R. J. Majer, Sci. Tools, 15, 11 (1968). (16) L. C. Hansen, W. G. Scribner, T. W. Gilbert, and R. E. Sievers. Anal. Chem., 43, 349 (1971). (17) W. R. Wolf, M. L. Taylor, B. M. Hughes, T. 0. Tiernan, and R. E. Sievers, Anal. Chem., 44, 616 (1972). (18) N. M. Frew, J. J. Leary. and T. L. Isenhour, Anal. Chem.. 44, 665 (1972). (19) K. J. Eisentraut, M. S. Black, F. D. Hiieman, and R. E. Sievers, Geochim. Cosmochim. Acta, 2, 1327 (1972). (20) T. L. Isenhour, E. R. Kowalski, and R. E. Sievers, Dev. Appl. Spectrosc., 8, 193 (1970). (21) T. H. Risby. P. C. Jurs, F. W. Lampe, and A. L. Yergey, Anal. Chem., 46, 161 (1974). (22) T. H. Risby, P. C. Jurs, F. W. Lampe, and A. L. Yergey, AnalChem., 46, 726 (1974). (23) R. W. Moshier and R. E. Sievers. "Gas Chromatography of Metal Chelates'', Pergamon Press, Elmsford, N.Y., 1965, references cited therein. (24) G. Guiochon and C. Pommier, "Gas Chromatography in lnorganics and Organometallics", Ann Arbor Science, 1973, references cited therein.
RECEIVEDfor review November 20, 1975. Accepted January 14, 1976. Part of this work was presented a t the 22nd Annual Conference on Mass Spectrometry and Allied Topics in Philadelphia, Pa., May 1974. Part of this work was supported by the U S . Environmental Protection Agency, Grant No. R803651-10.
Quantitative Analysis of Low Alloy Steels with the Ion Microprobe Mass Analyzer Kouzou Tsunoyama,* Yoshiharu Ohashi, and Toshiko Suzuki Research Laboratories, Kawasaki Steel Corp., Chiba, Japan
The effect of carbides on the relative Intensity of sputtered Ions In low alloy steels was investigated with the Ion microprobe mass analyzer. The relative Intensity varied with the formation of carblde In the matrix. The varlation of the relative intensity imposed a severe limitation upon the analysis of alloy elements by means of the callbration curve procedure. A method to correct for the effect of carblde formation was proposed and applied successfully to the analysis of low alloy steels. 832
ANALYTICAL CHEMISTRY, VOL. 48, NO. 6, MAY 1976
The ion microprobe mass analyzer has recently been applied to various problems in steels. The spatial distribution of alloy elements and in-depth analysis of sheet surfaces were studied by several investigators (1-3). However, few reports have been published on quantitative analysis of alloy elements in steels. Andersen and Hinthorne ( 4 ) proposed a theoretical correction procedure for the quantitative analysis of sputtered ions. They demonstrated that the local thermal equilibrium (LTE) model worked well for
