Anal. Chem. 1980, 52, 398-400
398
LITERATURE CITED
has been shown to provide large gains in efficiency over existing systems. The very short processing times that are possible with this new algorithm are important for processing high resolution mass spectra in real time applications.
( 1 ) A. L. Robertson and M. C. Hamming, Biorned. Mass Spectrorn., 4,203
ACKNOWLEDGMENT The authors thank M. Bruce and J. K. MacLeod of the Australian National University for bringing to their attention Robertson and Hamming's algorithm. C. MacDonald and C.
(1977). (2) A. L. Burlingame, Adv. Mass Spectrom., 4, 15 (1968). (3) S. R. Shrader, "Introductory Mass Spectrometry",Allyn and Bacon, New York, 1971. (4) J. Lederberg, "Computation of Molecular Formulas for Mass Spectrometry", Holden-Day, San Francisco, Calif., 1964. (5) J. R. Chapman, "Computers in Mass spectrometry", Academic Press, London, 1978. (6) E. M. Reingold. J. Nievergelt, and N. Deo, "Combinatorial Algorithms", Prentice-Hall, Englewood Cliffs, N.J., 1977.
Whittle of CSIRO Canberra are also thanked for helpful discussions.
RECEIVED for review July 5,1979. Accepted November 8,1979.
Analysis of Sulfuric Acid Aerosol by Negative Ion Chemical Ionization Mass Spectrometry J. E. Campana' and T. H. Risby2* Department of Chemistty, The Pennsylvania State University, University Park, Pennsylvania 16802
SO3 + H 2 0
Positive and negative chemical ionization mass spectra have been obtained for sulfuric acid using methane as either the reactant gas or as the moderating gas. The only positive ions which could be attributed to sulfuric acid were due to the protonated molecular Ions and these ions had very low intenstties. In contrast there were many ions which could be attributed to sulfuric acid and its fragments and some of these ions were Intense. A preliminary investigation is reported on the use of negative ion chemical Ionization mass spectrometry for quantitative analysis. Various methods of sample introduction and source temperatures were investigated and the optimum conditions provided a limH of detection of 100 ng of sulfuric acid.
+ hv
-+
SOz* + 0,
so4 + so,
-
-
SO4
2s03
(1)
(3)
1Present address: Department of Pharmacology and Experimental Therapeutics, School of Medicine, The Johns Hopkiiis University, Baltimore, Md. 21205. 2 Author to whom all correspondenceshould be addressed; present address: Division of Environmental Chemistry, Department of Environmental Health Sciences, School of Hygiene and Public Health, The Johns Hopkins University, Baltimore, Md. 21205. 0003-2700/80/0352-0398$01 .OO/O
H,SO,
(4)
where SO,* represents an electronically excited state of sulfur dioxide and the SO4intermediate species is only speculative. Castleman ( 2 ) has shown that water vapor reacts readily with sulfur trioxide to produce sulfuric acid with a rate cm3/s and a equilibrium constant coefficient of 9.1 x of 7.8 X lo9 atm-' at 29 "C (Equation 4). Examination of thermochemical data reveals that the reaction for the condensation of sulfuric acid gas is favored thermodynamically. Heicklen concludes from these data that sulfur trioxide is rapidly converted to sulfuric acid under all atmospheric conditions ( I ) . Currently, there is no direct analytical method for the measurement of sulfuric acid aerosols. West and co-workers, authoritative workers in this field, have recently reviewed the problems, sampling techniques, analytical methodologies, preparation of standards, and new developments in this important research area ( 3 ) . Recently a gas chromatographic methodology for the analysis of sulfates in real atmospheric samples has been reported ( 4 )and the detection limit using a flame photometric detector was 0.5 ng. This method was reported to be sensitive to sulfuric acid, sulfurous acid, methyl and dimethyl sulfonic acids; all of which were found to be present in real samples without any interference from ammonium sulfate. This chromatographic methodology coupled with the specificity of mass spectrometry would appear to be a possible methodology for the determination of sulfur-containing acids and sulfates in the atmosphere. This research investigates the possibility of using chemical ionization mass spectrometry as a sensitive and specific detector for ambient air monitoring of sulfuric acid aerosol. The electron impact (EI) mass spectra ( 5 ) and the chemical ionization (CI) mass spectra (6, 7)of sulfuric acid and sulfates have been reported by other workers. The E1 spectra showed the prominent ions to be [SO,]', [SO]+,and [SI'; but attempts to quantify by separating the acid from its salts using a temperature-programmed solids probe did not provide adequate separation. The CI mass spectra study of these compounds gave impressive results using methane as a reactant gas. The major ions observed for the acid and its salts were [HS02]+,
T h e environmental and physicological significance of sulfur-containing compounds as pollutants in the atmosphere has been realized in the last decade. Morbidity in man from long-term exposure to toxic levels of sulfur dioxide is recognized as cardiorespiratory diseases and increased illness among patients suffering from bronchitis ( I ) . Sulfuric acid is produced by the photooxidation of sulfur dioxide to sulfur trioxide in the presence of oxygen and water under atmospheric conditions. The postulated reaction mechanism 1-4 for the formation of sulfuric acid in the atmosphere is as follows,
SO,
-
G
1980 American Chemical Society
ANALYTICAL CHEMISTRY, VOL. 52, NO. 3, MARCH 1980 ..c
Table I. Negative Ions of Sulfuric Acid at 200 " C relative intensity,
ion
399
7
%
14.0 100.0 6.0
7.9
112
160 176 177
194 195 275
[S0;SO4]-
1.2
9.2 12.7 5.4
[SO;SO,]-
[SO,.HSO,][H,SO,.HSO,][ H,SO;HSO,][H,SO;SO;SO,]
?
1.2 3.5
[HS03]+,and [H3S04]+; the acid and its salts were separated using a temperature-programmed solids probe. In that study water was used as the reactant gas to protonate H2S04selectively over SO2 (proton affinities: SO3 < SO2 < H 2 0 < H2S04). The only problem with this methodology was due to a hydrocarbon interference at m / z 99 that required the use of high-mass resolution (>1000). The CI studies showed a detection limit for sulfuric acid to be 20 ng for standard samples. T h e chemical composition and structure of sulfuric acid would appear to make it amenable to analysis by "negativeion" chemical ionization (NICI) mass spectrometry and this study discusses the NICI mass spectra of sulfuric acid.
EXPERIMENTAL Reagents. Sulfuric acid (Ultrex, J. T. Baker Chemical Co., Phillipsburg, N.J.), ethanol (loo%, Publicker Industries Co., Linfield, Pa.), benzaldehyde (J. T. Baker), and salicylaldoxime (Eastman Organic Chemicals, Rochester, N.Y.) were used directly. Apparatus. A chemical ionization mass spectrometer (Scientific Research Instruments Corporation Biospect System) which has been described previously (8) was used for this study. Methane was used as the reactant gas unless otherwise specified and all data were acquired using a computer data acquisition system (9). The reactant gas inlet pressure was 1 Torr and the emission current was regulated to 0.4 mA. The mass spectrometer was mass calibrated to the computer data acquisition system using salicylaldoxime and reactant gas ions as mass calibrants. Several abundant ions are available in the mass range of interest for mass calibration. A positive ion at m / z 138 corresponds t o the protonated-molecular ion of salicylaldoxime. The negative ion mass spectra revealed decreased background, there were no background negative ions in the source, a distinct advantage of NICI mass spectrometry. The major negative ions of salicylaldoximeare m / z 136 and m / z 120. The first is a result of the loss of a proton (on either the hydroxyl or oxime group) or the capture of an electron with subsequent loss of H.. The ion at m / z 120 is attributed to the loss of an OH. group from salicylaldoxime. The capture of an electron and subsequent loss of H- or OH. is analogous to mechanisms observed in the negative ion E1 spectra of organic compounds containing the hydroxyl group (IO).
RESULTS AND DISCUSSION Aliquots (10 1L) of concentrated sulfuric acid were evaporated from glass capillaries that were inserted into the solids probe and during the evaporation time-averaged spectra (10 scans) were collected over the range of m / z 10 to 200. Spectra were collected in both the positive and negative chemical ionization modes at nine ion source temperatures from 100 to 300 "C in increments of 25 "C. No appreciable ion currents from sulfuric acid were observed until temperatures in excess of 125 "C were reached. The positive ion spectra revealed only one ion that could be attributed to sulfuric acid and this was the protonated molecular ion. The intensity of this ion was less than 1%of the total ion current. At source temperatures greater than 150 "C, no positive ions attributable to sulfuric acid were observed.
\
4
Figure 1. Temperature study of the relative intensities of sulfuric acid
The negative ion spectra displayed several intense ions over the temperature range studied. The major ions observed a t 200 "C are listed in Table I. Several of the observed ions appear to be a result of either association reactions or fragmentation of the dimeric sulfuric acid species. For example, the ions observed a t mlz 176 and mlz 177 could be the result of the loss of a proton from molecular pyrosulfuric acid [H2S,0,] followed by loss of a hydrogen atom respectively. Alternatively, these ions could be a result of the association as expressed in Table I. The results of the temperature study are summarized in Figure 1. The relative concentrations of [HS04]- and its association or dimeric complex, [ HzS04.HS04]-,decrease pseudo-exponentially with temperature increase. This suggests that the thermal decomposition of sulfuric acid increases with temperature, but it should be noted that higher source temperatures produce a higher rate of vaporization of the sample from the probe. Under this condition, the high concentrations of the sulfuric acid species in the ion source could act as the negative chemical ionization reactant gas. The positive ion mass spectra showed high concentrations of the protonated monomer, dimer, and trimer of water which was contained in the concentrated sulfuric acid solution. This high concentration of water could result in the conversion of [SOJ to [HS04J-with subsequent association to form the [S03-HS04]-negative ion at higher temperatures. The disappearance of [SO3]-with temperature increase supports the former supposition. The species SO3could react with [OH](a product of the dissociative ionization of H 2 0 ) to form [ S 0 3 0 H ] -(or [HS04]-)which could subsequently dissociate to [SO,]-. I t is interesting to note that no ion currents for [O,]- and [SO,]- were obtained. Sulfur dioxide and oxygen both have high collisional stabilization efficiencies relative to helium and the following [OJ attachment reaction can be postulated 02-
+ so2
-
[S02.0,1-
(5)
from the observation that [02]promotes negative-ion formation of the sample adduct by an associative-ionization process ( 8 , I I ) . Fehsenfeld and Ferguson have discovered that SO2 rapidly replaces H 2 0 clustered to all negative ions (12) while our laboratory has reported the [ H20.0H]- association negative ion in chemical ionization mass spectrometry (13,16). Therefore, the following reaction can then be postulated
400
ANALYTICAL CHEMISTRY, VOL. 52, NO. 3, MARCH 1980
[H,O*OH]- SO2
-
[SOz.OH]- H20
(6)
and may define another reactive intermediate species. These are only several of the many possible reactions that may explain the results of this complex system. The highest absolute ion current which was due to the ion [SO4]- occurred at 275 "C and this temperature was used for subsequent trace analysis. Solutions of 1OOO,100,and 10 ppm of sulfuric acid by weight in ethanol were prepared and introduced into the mass spectrometer via the solids probe. Solutions of known volume (rL) were injected onto glass capillaries or onto alumina tubes (151, and inserted into the solids probe. Data were acquired in the region m / z 94 to 100 using the continuous repetitive measurement of the spectra mode of computer data acquisition. The resulting integrated peak areas were reproducible for the 1000 and 100 ppm solutions (ca. 5.0 and 0.5 pg of H2S04). Results for the 10-ppm solutions were irreproducible and suggested a problem with the sample introduction methodology or the need for an internal standard. The sulfuric acid solutions in ethanol also appeared to be unstable and benzaldehyde was used as the solvent ( 3 ) . The benzaldehyde-acid solution was also found to be unstable; the solution turning yellowish within a few hours. The benzaldehyde negative ion spectrum showed the molecular ion a t m / z 108 which is in the mass region of the analyte and could be used as an internal standard. A new method of sample introduction was investigated whose principle was similar to the method of direct sample injection which has been reported previously (16). The simplified injector was constructed from glass capillary and used a stainless steel injection port. Solutions were injected through a septum and the sample plug was swept into the mass spectrometer source by the reactant gas. The injector was kept a t 250 "C for the analysis and the glass capillary was treated with acetic acid to minimize reactive sites. Methane saturated with water was used as the reactant gas. Water-saturated methane reactant gas was used since it was thought that the water may react with the thermally decomposing sulfuric acid to force the reaction in the direction of the undissociated form. Higher intensities of [HS04]-and [SO4]-were observed with this reactant gas, but the results for solutions of about 10 ppm and loss of sulfuric acid were again irreproducible. The limit of detection for sulfuric acid in ideal samples was about 100 ng.
CONCLUSION T h e analysis of sulfuric acid represents a formidable problem by mass spectrometry. The data suggest a need for the investigation of some simple systems such as sulfur trioxide and sulfur dioxide with water reactant gas. The detection limit
reported here is for an ideal sample. This is within the sensitivity range of several other methods, but may be considerably poorer with real samples. The presence of other sulfates in real samples could cause a problem with interferences. The direct injection technique is simple and merits further investigation. If the injector is operated below 300 "C, no sulfate interferences should occur with real samples. Pyrolysis data of sulfates (17) indicate that ammonium sulfate will pyrolyze above 300 "C and the metal sulfates that could be present in real samples will pyrolyze above 400 "C. Other methods of ionization which can be performed at low source temperatures should be investigated in the hope that they may ensure thermal stability of the species formed.
ACKNOWLEDGEMENT The authors thank W. D. Smith for his help in this study.
LITERATURE CITED (1) J. Heicklen, "Atmospheric Chemistry", Academic Press, New York, 1976. (2) A. W. Castleman, Jr., Brookhaven National Laboratw, . Lona-Ishnd, N.Y., unpublished work. (3) V . Dharmarajan, R. L. Thomas, R. F. hkddalone, and P. W. West, Sci. Total Environ . 4. 279- 119751. (4) R - D . Penzhorn and W. G. Fllby, Staub-Reinhalt. Luft, 36,205 (1976). (5) D. Scheutzle, L. M. Isabelle, and J. G. Watson, presented at the 23rd Annual Conference on Mass Soectrometrv and Allied Tooics. - . Houston. Texas, May 1975. (6) T. M. HaNey, T. J. Prater, and D. Scheutzle, presented at the 24th Annual Conference on Mass Spectrometry and Allied Topics, San Diego. Calif., May 1976. (7) T. M. Harvey, T. J. Prater, and D. Scheutzle, SA€, Pub/. No. 770063, Detroit, Mich., February 1977. (8) S. R. Prescott, J. E. Campana, and T. H. Risby, Anal. Chem., 49, 1501 (1977). (9) J. E. Campana, T. H. Risby, and P. C. Jurs, Anal. Chim. Acta. Comput. Techno/. Optim., in press. (10) J. H. Bowie and 8. D. Williams, "MTP International Review of Science", Physical Chemistry, Series 2, A. Maccoll, Ed., Vol. 5, Butterworth and Co., London, 1975. (11) D. F. Hunt and F. W. Crow, Anal. Chem., 50, 1781 (1978). (12) F. C. FehsenfeM and E. E. Ferguson. J . Chem. Phys., 61, 3181 (1974). (13) J. E. Campana, Ph.D. Thesis, The Pennsylvania State University, University Park, Pa., 1979. (14) J. E. Campana, S. R. Prescott, P. C. Jurs, and T. H. Risby, the 172nd National Meeting of the American Chemical Society, San Francisco, Calif., September 1976. (15) J. E. Campana, P. C. Jurs. and T. H. Risby, the 25th Annaul Conference on Mass Spectrometry and Allied Topics, Washington, D.C., May 1977. (16) S. R. Prescott, J. E. Campam. P. C. Jurs. T. H. Risby, and A. L. Yergey. Anal. Chem., 48, 829 (1976). (17) D. F. Adams, Health Lab. Sci., 12, 150 (1975).
. .-
I
- I
RECEIVED for review July 12, 1979. Accepted December 27, 1979. The support of the U S . Environmental Protection Agency through Grant R 803651 and of E. I. Du Pont de Nemours and Company through unrestricted funds to the Center for Air Environment Studies is gratefully acknowledged.
