1180
Anal. Chsm. 1991, 63,1186-1192
Fixed-Wavelength Laser Ionization/Tandem Mass Spectrometry for Mixture Analysis in the Quadrupole Ion Trap Douglas E.Goeringer,* Gary L. Glish, and Scott A. McLuckey Analytical Chemistry Division, Oak Ridge National Laboratory, Oak Ridge, Tennessee 37831 -6365
The analytlcal potentla1 of flxed-wavelength laser photolonk zatbdtackm mau epoctrometry In the qwdnpde ion trap Is Investigated. Tho capablllty of the Ion trap for mass-08Iectlve kdatlon and dkroclatlon of laser+enorated Ions Is exploned for organic mlxture analysis. Deaplte the fact that c o l l ~ n d u c e ddkroclatlon (CID) ylddr product Ions no dMorenl from the fragment Ions produced via laser lonizatlon/fragmentation, tandem mass spectrometry Is shown to Improve the speclflclty of fixed-wavelength (266-nm), resonance-onhanced, two-photon ionization (R2PI). The technlque Is illustrated by udng mass-selectlve CID of lasergenerated molecular Ions from a mlxture of ethylbenzene, propylbenzene, and the Isomers N,Ndhnethylanlllne and N-ethylanlline.
INTRODUCTION Laeer multiphoton ionization (MPI) has proven to be a very useful method for detection of polyatomic species. MPI of a molecule can be achieved if the total energy of the photons absorbed during irradiation exceeds the ionization potential of the molecule. Resonance-enhanced multiphoton ionization (REMPI) occurs when the laser is tuned to a wavelength matching an allowed n-photon transition for the molecule, thus increasing ita MPI cross section by several orders of magnitude. Many aromatic molecules have strong absorption bands in the UV that are rather broad at room temperature due to the high population of excited rovibronic states. The ionization potentials for many such polyatomics are also often less than 9 eV (I). Therefore, they can be ionized with high efficiency by resonance-enhanced two-photon ionization (RPPI) by using the frequency-quadrupled output of the NdYAG laser at 266 nm (4.7 eV) (2,3).Ionization selectivity for polyatomics is necessarily limited when fixed-wavelength R2PI is used, however. Although improvementa in ionization selectivity can be achieved through the use of tunable UV laser radiation (449,limitations exist on the optical selectivity of tunable R2PI attainable a t room temperature due to the thermal population of excited rovibronic states. Improved optical ionization selectivity can be realized by using tunable R2PI in conjunction with supersonic jet techniques, which serve to depopulate excited rotational and vibrational levels (7, 8). Two-laser R2PI in a supersonic jet has also been demonstrated for isomer discrimination (9). However, the implementation of these techniques involves additional complexity and increases analysis time. Moreover, in some situations such as rapid screening of complex mixtures for compound class, it may be advantageous to use a relatively nonselective ionization method in conjunction with another technique that enhances the analytical specificity. In this paper, we examine tandem mass spectrometry as an alternative method for enhancing the specificity of fixedwavelength R2PI. Tandem mass spectrometry, or MS/MS, is a technique that can enhance the specificity of analysis over that provided by a single stage of mass spectrometry. The usual MS/MS experiment involves sequential mass-selective 0003-2700/91/0363-1186502.5010
isolation of a parent ionic species, dissociation of the parent ion, and mass analysis of the resulting product ions. Because the product ions are generally characteristicof the parent ion structure, their mass spectrum furnishes information for identifying that structure. The specificity of an analysis can sometimes be improved further by employing multiple stages of MS/MS (MS").Because the degree of fragmentation accompanying R2PI is a function of laser power, single-stage mass spectrometric analysis of such fragments also can often improve the specificity of the overall analysis (10,11). However, in the case of organic mixture analysis, the complexity of R2PI/fragmentation mass spectra may compromise compound identification. Specificity is distinctly differentiated from selectivity in this context. Selectivity refers to the ionization discrimination of compounds by judicious choice of some parameter, such as wavelength. Consequently, isomeric compounds may possibly be differentiated on the basis of R2PI wavelength. Specificity refers to identification based on characteristic information in the spectrum of a compound, such as relative intensitiea of fragment ions versus mass-to-charge ratio. An ionization method with relatively limited selectivity, such as fixed-wavelength RPPI, can thus be specific for a particular compound based on the unique nature of its mass spectrum. Either the selectivity or specificity of an analytical technique can be improved without necessarily enhancing the other. The time-of-flight mass spectrometer (TOFMS) has typically been used for mass analysis of laser-generated ions because it is inherently suited for use with pulsed ionization techniques. However, the TOFMS is not readily adaptable to perform tandem mass spectrometry. The three-dimensional quadrupole ion trap, which is also intrinsically suited for pulsed ionization methods, is capable of ion storage and subsequent ion manipulation, enabling a variety of experimenta to be performed. The combination of radio-frequency and dc voltages (12)or the use of resonance ejection (13) enables trapped ions to be mass-selectively stored or ejected. In addition, collision-induceddissociation (CID) of mass-selected parent ions is readily performed by mass-selective resonant excitation in the ion trap; product ions from CID also can be trapped and mass analyzed efficiently in the ion trap. When such methods for ion isolation, dissociation, and mass analysis are combined in the proper manner, MS" can be performed in the quadrupole ion trap (14,15). The utility of the quadrupole ion trap for tandem mass spectrometry has been demonstrated in many applications (1618).In addition, REMPI (191,laser desorption mass spectrometry (201,and tandem mass spectrometry coupled with laser desorption (21) in the ion trap have recently been reported. Such capabilities suggest that the combination of fixed-wavelength R2PI and tandem mass spectrometry in the quadrupole ion trap can provide a viable alternative to supersonic expansion/RSPI techniques coupled with time-of-flight mass spectrometry. In the work presented here, an ion trap mass spectrometer was adapted for RBPI/tandem mass spectrometry. Examples are presented that illustrate the use of tandem mass spectrometry for specificityenhancement in fixed-wavelengthRPPI of organic mixtures. 0 1991 American Chemical Society
ANALYTICAL CHEMISTRY, VOL. 63, NO. 13, JULY 1, lB9l
EXPERIMENTAL SECTION Apparatus. Various details of the standard ion trap mass spectrometer (ITMS) (Finnigan MAT, San Jose, CA) have been described previously (22). Briefly, ionic motion in a three-dimensional quadrupole field can be described by the Mathieu equation, which is characterized by the parameters a, and q,. These parameters are related to the radio-frequency(rf)amplitude (V),dc amplitude (U,angular radio frequency (!I), mass (4, and ring electrode radius (ro)by the equations a, = -2a, = -8eU/(mr&z) qI
= -2q, = -4eV/(mr:!I2)
The subscript u refers to either the radial (r) or axial (z) direction. Normal operation of the ITMS uses a 1.1-MHz rf voltage, known as the fundamental rf, applied to the ring electrode, and the dc amplitude is zero (a, = 0). Ions with q, values below 0.908 have stable trajectories. Thus, ions above a specific m/z, which is proportional to the fundamental rf amplitude, are stored in the ion trap. Maas spectra are normally obtained by the mass-selective instability technique (23) in which the fundamental rf amplitude is linearly increased, thus causing trapped ions to sequentially exit, from low to high m/z, through holes in the center of the exit endcap. At a given fundamental rf amplitude, each m / z has a set of characteristic frequencies of oscillation known as secular frequencies. Mass-selective kinetic excitation of stored ions can be effected when the frequency of a supplemental rf signal applied to the endcap electrodes is in resonance with a secular frequency for a specific m/z value. Such resonant kinetic excitation can lead to energetic collisions with the helium bath gas. Subsequently, CID can occur with the product ions remaining in the ion trap. MS/MS spectra can then be acquired by mass analysis of the trapped product ions. Because the CID and trapping efficiencies for the ion trap can often be very high (>95%), multiple stages of tandem mass spectrometryare easily performed with the ITMS provided certain prerequisites are met (24). Simultaneous application of rf and dc potentials to the electrodes of the ion trap, such that the rf/dc amplitude ratio is maintained at a constant value, permits mass-selective ion isolation. The specific mass/charge for trapped ions is determined by the exact values for the rf and dc amplitudes. Experiments described in this report were performed on a modified ITMS. Two 3-mm-diameterholes spaced 180 deg apart, with their centerlines passing through the center of the trapping volume, were drilled through the ring electrode. The ion trap aasembly was then mounted in the standard ITMS configuration with the ring electrode in a vertical plane. The collinear holes were horizontally aligned with an optical port on the side of the ITMS,enabling a laser beam to pass radially through the ion trap cavity. The electron beam gating signal from the ITMS control electronics was used to trigger the laser by activating a pulse generator whose output in turn triggered the oscillator control input of the NdYAG laser. The f i i e n t assembly was left intact, thus permitting E1 to be performed when desired. However, the filament current was always off during R2PI experimentsto ensure that an electron pulse did not occur. ITMS software for experiment control and data acquisition, running on an IBM PC/AT, permits flexible control of the fundamental rf, supplemental rf, electron gating, trigger, and other control parameters. The rf amplitude during photoionization was fixed at a level corresponding to a high-pass mass cutoff of m/z 25. The ion packet resulting from the laser pulse was stored for 10me before b e i i subjected to further manipulation. R2PI mass spectra were acquired via the standard mass-selective instability technique. Individual MS/MS and MS/MS/MS spectra for R2PI-generated molecular ions were a180 obtained. After being photoionized via unfocused laser radiation at 266 nm, the molecular ion was isolated in the ion trap via the rf/dc technique. In addition to application of a dc potential, this procedure also required elevating the rf amplitude to a correspondinglevel above that used during photoionization. Following the molecular ion isolation step, the dc component was then removed and the rf amplitude reduced. T h e MS/MS spectrum of the molecular ion was subsequently obtained via CID resulting from application
1187
of the supplementaryrf signal at an appropriate BBcular frequency. The average parent ion kinetic energy during resonant excitation is affected by its qEvalue, higher qEvalues resulting in more energetic collisions (14). However, because q, for the parent ion increases with the rf trapping level, the minimum m/z value at which product ions can be trapped also increases with collision energy. Therefore,the fundamentalrf amplitudeduring collisional activation was adjusted to an amplitude that would result in sufficient kinetic excitation for dissociationwhile allowing product ions above m/z 50 to be trapped. MS/MS/MS spectra were similarlyobtained by subsequent ion isolation and CID of selected product ions generated in the previous MS/MS stage. Laser and Optics. R2PI was performed by using the frequency-quadrupled output (266 nm) from a pulsed NdYAG laser (Quanta-RayDCR-2A, Spectra Physica, Mountain View, CA). The laser radiation was directed through a 2-mm-diameter aperture and variable attenuator (Model 935, Newport Corp., Fountain Valley, CA) before entering the ITMS vacuum housing via an optical port. The attenuator was typically adjusted to give approximately0.5 mJ/pulse, unless otherwisenoted, when measured in front of the ITMS optical port. For experiments involving focused laser radiation, a Wcm focal length, fused silica lens was positioned in front of the optical port such that the laser focus was located near the center of the ion trap cavity. Sample Handling. Ethylbenzene, propylbenzene, NJV-dimethylaniline, and N-ethylaniline samples (Chem Services, Westchester, PA) were introduced into the ITMS by sampling the saturated headspace vapor with a variable leak valve. A mixture of the sample vapors was generated by using separate leak valves, connected in parallel, for each compound. The samples were degassed before such experiments via several freeze-pumpthaw cycles. RESULTS AND DISCUSSION One of the attractive characteristics of resonance enhanced laser ionization is the variable degree of internal energy deposition and resultant fragmentation associated with ionization. By increasingthe laser power density, additional photons can be absorbed by the molecular ion subsequent to its formation; this results in the formation of excited ionic states that dissociate into ionic fragments. Further increases in laser power can lead to subsequent absorption of additional photons by the initial fragment ions, causing more extensive fragmentation. Another means of increasing the internal energy of the molecular ion, and thereby its extent of fragmentation, is by collision with a neutral target gas. The process, known as collisional activation (CA), is initiated via conversion of the ion's translational energy into internal energy. The lowelectronvolt collision energy region is accessed during CA in the ITMS. The energy deposited during CA in the ion trap can be varied by changing the amplitude or duration of the supplementary rf excitation signal. The m / z values and relative abundances of resultant fragment ions from either excitation process are determined by the parent ion internal energy distribution, rate constant versus internal energy (k versus E) for a series of consecutive and competing unimolecular dissociation reactions, and time before the fragments are analyzed. Although RBPIlfragmentation and CIDMS/MS can yield complementary data, situations may often occur in which the spectra reveal no unique structural information. Nevertheless, the combination of R2PI and CIDMS/MS can still provide an advantage over the use of RPPI alone as seen below. R2PI and Tandem Mass Spectra. RBPI/fragmentation and R2PI/tandem mass spectra obtained in the ITMS for ethylbenzene, propylbenzene, Nfl-dimethylaniline, and Nethylaniline are shown in Figures 1-4. The R2PI/tandem mass spectra were acquired by using unfocused radiation. The R2PI/fragmentation spectra were obtained by increasing the 266-nm laser energy to about 1 mJ/pulse and focusing the output in order to achieve laser-induced fragmentation. In addition to the molecular ion, major fragment or CID product
1188
ANALYTICAL CHEMISTRY, VOL. 63,NO. 13, JULY 1, 1991
I
25
a
a 120
35 45
55 65
75 85 95 105115125 b
120
25
35 45
55 6 5 7 5 8 5 9 5 1 0 5 1 1 5 1 2 5
I
II
55
65
75
85
95
105
115
125
55
65
mfz
75
85
95
105
115
125
m/z
Flguro 1. Quadrupole Ion trap mass spectra of N,Ndlmethylaniiine. (a) RPPI/fragmentatlon spectrum. (b) R2PI-MS/MS spectrum of m lecular ion 121 flrstgeneratlon products. (c) RPPI-MSIMSIMS spectrum of 121 -. 120 -.secondgeneratlon products.
Flgurr 2. Quadrupole ion trap mass spectra of N-ethylaniilne. (a) R2PI/fragmentaticm sped". (b) RPPIWSIMS spectrum of molecular ion 121 firstgeneration products. (c)R2PI-MS/MS/MS spectrum of 121 106 second-generation products.
ions can be identified in each of the spectra. The R2PI/ fragmentation (Figure la) and RSPI-MS/MS (Figure 1b) spectra for N,N-dimethylaniline indicate that its molecular ion ( m / z 121) fragments very efficiently to (M - He)+at m/z 120. The R2PI-MS/MS/MS spectrum from CID of (M - W)+ (Figure IC)has major second-generation product ions at m / z 77,91,103, and 118; the formulas for these ions have previously been assigned from MS/MS/MS spectra acquired in the ITMS from CID of EI-generated molecular ions (25). For N-ethylaniline and ethylbenzene, methyl loss to give (M CH3')+ product ions a t m / z 106 and 91, respectively, is the major process in the R2PI/fragmentation (Figures 2a and 3a) and RSPI-MS/MS (Figures 2b and 3b) spectra. Subsequent CID of (M - CH3')+ from N-ethylaniline produces R2PIMS/MS/MS second-generation product ions at mlz 77 and 79 (Figure 2 4 . These product ions correspond to those identified in the N,N-dimethylanilineMS/MS/MS spectrum. The propylbenzene RPPI/fragmentation (Figure 4a) and RZPI-MS/MS (Figure 4b) spectra indicate that its molecular ion (m/z 120) fragments to (M - C2Hs')+. Although the data are not shown here, RZPI/fragmentation mass spectra containing the molecular ion accompanied by a single major fragment ion were also experimentally observed for each compound by gradually increasing the energy/pulse with unfocused R2PI. In each case, the major RPPIlfragmentation ion corresponded to the major RBPIICID product ion in the MS/MS spectrum, e.g., (M - CH3')+ at m / z 106 from N ethylaniline.
The ionization energies for the sample compounds and the appearance energies for their corresponding major fragment ions observed are (1)N,N-dimethylaniline,7.12 eV, (M - IT)+, 10.8 eV; N-ethylaniline, 7.67 eV, (M - CH3')+, 10.6 eV; ethylbenzene, 8.77 eV, (M - CH3')+, 10.9 eV; propylbenzene, 8.72 eV, (M - C2H5*)+,11.6 eV. Thus, two 266-nm (4.7-eV) photons are required for ionization of each compound, and an additional photon is necessary to exceed the appearance energy for each corresponding major R2PI/fragmentation ion. It is, therefore, reasonable for the RPPI mass spectra obtained with low-power, unfocused laser radiation to be dominated by the molecular ions and for the relative abundance of the major fragment ions to gradually increase with the laser energy. The fragment ions seen at mass-to-chargeratios 39,51, and 65 in the high-power RZPI spectra (Figures la-4a) have considerably higher appearance energies, indicating a larger amount of internal energy was deposited during focused R2PI. The fact that product ions at these m/zvalues are absent in the R2PI-MSIMS spectra is a consequence of several factors. First, the MS" experiments were performed at an rf level corresponding to a high-pass m / z limit of 50. Therefore, product ions below that value are not seen in the RSPIICID spectra even if they are formed. As indicated above, there is a trade-off between the low mass limit for trapping product ions and the maximum kinetic energy attained during excitation. Thus, the observation of product ions from higher activation energies may be precluded by the fundamental rf level required to attain the requisite kinetic energy. Second,
-.
-- -
ANALYTICAL CHEMISTRY, VOL. 63,NO. 13, JULY 1, 1991
1189
a 121
120 106
25
35
45
55
65
75
85
b
9 5 105 115
25 55
65
75
85
95
105
115
-
Flgwe 3. Quadrupole ion trap mass spectra of ethylbenzene. (a) R2PI/fragmentetkn spectnm. @) R2PIMSIMS specbvn of molealar
fkst-generatbn products.
c
25
55
35 45 5 5 65
65
75
75 8 5 95 1 0 5 1 1 5 1 2 5
85
7 5 85
120
95 1 0 5 1 1 5 1 2 5
mlz
Flgurr 5. Quadrupole ion trap mass spectra from mixture of ethyl-
benzene, propylbenzene, Nethylaniline,and N,Ndlmethylanlline vapors following (a) unfocused 266-nm R2PI and (b) focused 266nm
m/z
ion 106
35 4 5 5 5 6 5
I
95
105
115
125
mlz
-
Flgurr 4. Quadrupole ion trap mass spectra of propylbenzene. (a) R2PI/fragmentetkn specbum. @) RPPIMS/MS spectnm of molecular ion 120 flrst-generatlon products.
there is an upper limit to the kinetic energy that an excited ion can possess before it is ejected from the ion trap (26). Because the internal energy of a collisionally activated ion depends on its kinetic energy, among other factors, this establishes a maximum value of excess internal energy that can be deposited via CA in the trap. Third, product ions formed in the first stage of MS/MS subsequently fall out of resonance with the supplementary excitation signal. Therefore,such ions are no longer kinetically excited and would not be expected to undergo further dissociation. Finally, internal energy deposition occurs in a stepwise fashion on the millisecond time
R2PI.
scale in the trap (27). Thus, the maximum value of internal energy achieved via CA in the trap is determined by the k versw E curves for the competing dissociation pathways (28). Mixture Analysis. On the basis of the above results, combining tandem mass spectrometry with R2PI appears to provide little advantage over the use of R2PI alone for the analysis of pure compounds. However, a single-stage mass spectrum for a mixture of compounds, whether obtained via RBPIIITMS or RPPIITOFMS, is a combination of the individual mass spectra. As such, the complexity of the overall spectrum generally compromises the identification of the individual components in the mixture. The well-known strengths of tandem mass spectrometry in mixture analysis are therefore expected to benefit R2PI in such situations. The effectiveness of RZPIItandem mass spectrometry compared with RBPIlfragmentation for mixture analysis was demonstrated using a mixture of the N,iV-dimethylaniline, N-ethylaniline, propylbenzene, and ethylbenzene. The R2PI mass spectrum in Figure 5a was obtained by using low-power, unfocused, 266-nm laser radiation, which minimized fragmentation; the R2PI/fragmentation spectrum in Figure 5b was acquired by increasing the laser energy and focusing the output to induce fragmentation. Assuming that the sample is comprised of a variable (unknown) number of components from the designated suite of four compounds, inspection of these spectra reveals the limitations of RPPI when employed for mixture analysis. The molecular ion signals at m/z values 106 and 120 in Figure 5a indicate the presence of ethylbenzene and propylbenzene, respectively. However, the mlz 121 ion signal could be due to either N&-dimethylaniline, N-ethylaniline, or both compounds. The fragment ion signals in Figure 5b at relatively low m / z values, e.g., 39,51,65, and 77, provide no distinguishing structural information. Fragment ion signals at higher mlz values contribute little additional characteristic data. Although the increase in the mlz 120 signal and the appearance of the mlz 104 ions suggest that N,N-dimethylaniline is also present, such effects could be attributed to propylbenzene. In any case, the presence of N-ethylaniline cannot be confirmed because its major fragment ion a t mlz 106 coincides with the molecular ion of ethylbenzene. Therefore, the origin of the ion signal at mlz
1180
ANALYTICAL CHEMISTRY, VOL.
63,NO. 13, JULY 1, 1991
I
55
55
65
75
85
95
105
116
a
0 120
75
65
85
95
105
115
125
125
m/z 0 106
C
-
propybenzene, N4JvknHb, and N,N-nOhe vapors rokwina focused 2 6 h m R P I and (a)MS/MS 120 flrst-generalion products and (b) MS/MS 106 firstgeneration products.
I
4 0
55
01
65
75
85
95
105
115
mlz
-
Flgure 6. Quadrupole ion trap mass spectra from mixture of ethylbenzene,propylbenzene, N-ethylanllne, and N ,Ndimethylanlline vapors following unfocused 266nm R2PI and (a) MS/MS 121 firstgeneration products, (b) MS/MS 120 firstgeneration products, and (c) MSlMS 106 flrst-generatlon products.
-
-
Figure 7. Quadrupole ion trap spectra from mixture of ethylbenzene,
-
121 cannot be confirmed. Thus, despite the fact that individual RBPI/fragmentation mass spectra for the components comprising a sample are unique, mixture analysis via R2PI mass spectrometry alone can be problematic when fragment ions are indistinguishablefrom molecular ions. As seen above, the RZPI-MS/MS spectra for each of the sample compounds contain major product ions corresponding to the major fragment ions in the RZPI/fragmentation spectra. Nevertheless, the use of tandem mass spectrometry in conjunction with R2PI can enhance specificity in such circumstances. An MS/MS spectrum (Figure 6) of the m / z 121 molecular ion from the same mixture was obtained by using unfocused laser radiation and the same CID parameters previously used for pure component spectra. As seen in the spectrum, the major product ions are a t m/z 106 and 120. Comparison with the pure component MS/MS spectra in Figures 1b-4b reveals that the m / z 106 and 120 product ions correspond to the Nethylanidine (M - CH,')+ and NJV-dimethylaniline (M - H)+ ions, respectively. Given this MS/MS structural information and the molecular ion data obtained from the low-power R2PI mass spectrum (Figure 5a), it can then be deduced that the sample comprises all four components. Of course, given any combination of possible mixtures generated from the same suite of assigned compounds, similar reasoning could be used to deduce the composition. Although MS/MS spectra of the ions a t m / z 106 and 120 were not needed to confirm the mixture composition, it is
interesting to compare such RBPIICID spectra taken a t low laser power with corresponding R2PI/CID spectra obtained at high laser power. As expected, the MS/MS spectra of m/z 120 and 106 ions from low-power R2PI (Figure 6b,c) were similar to the MS/MS spectra for the molecular ions from pure propylbenzene (Figure 4b) and ethylbenzene(Figure 3b), respectively, since little fragmentation was induced under such ionization conditions. Parts a and b of Figure 7 are MS/MS spectra of ions at m / z 120 and 106, respectively, following high-power R2PI of the four-component mixture. Under high-power R2PI conditions, part of the ion signal at m/z 120 and 106 should also be due to (M- H Y and (M - C H 3 Y fragment ions from N,N-dimethylaniline and N-ethylaniline, respectively. Thus, the MS/MS spectra in Figure 7 should differ from those taken under low-power RPPI conditions, viz. Figure 6b,c. As predicted, Figure 7a resembles a combination of the MS/MS spectrum for the propylbenzenemolecular ion (120 first-generation products, Figure 4b) and the MS/ MS/MS spectrum for the N,N-dimethylaniline (M - H ')' ion (121 120 second-generationproducts, Figure IC);Le., in addition to the m/z 91 signal, ion signals at m/z 77,103, and 118 are also present. Consequently, Figure 7a suggests that the signal at m/z 120 actually results from two ions of different structure. The MS/MS spectrum of the ethylbenzene molecular ion (106 first-generation products, Figure 3b) and the MS/MS/MS spectrum of the N-ethylaniline (M - CH,')+ ion (120 106 second-generation products, Figure 2c) contain unique product ions at m/z 91 and 77 and 79, respectively. Therefore, the presence of signals at these three m / z values in Figure 7b also suggests that the ion signal a t m / z 106 is due to the ethylbenzene M'+ and N-ethylaniline (M - CH,')+ ions. However, in order to definitively state that the m/z 120 and 106 signals following high-power R2PI each resulted from two ions of different structure, it was also necessary to ensure that the MS/MS spectra for the pure compounds were not subject to precursor ion internal energy effects. This was confirmed by additional high-power R2PI experiments that yielded pure-component MS/MS spectra independent of laser power density. These results for mixture analysis illustrate a typical situation in which the union of R2PI and tandem mass spec-
-
- -
-
- -
ANALYTICAL CHEMISTRY, VOL. 63,NO. 13, JULY 1, 1991
trometry can be beneficial. The use of MS/MS improves the specificity of fixed-wavelength R2PI/fragmentation despite the fact that no product ions unique to MS/MS may originate from CID. It is evident that the key element in this regard is the ion separation step, based on mass-to-charge ratio, which occurs between ionization and CID. Subsequent product ion spectra are free from signals due to unrelated molecular and fragment ions. Consequently, successive CID spectra of discrete ions can be compared with such spectra obtained from pure compounds to identify individual components of a mixture. Although tunable UV laser radiation was not available for use in this study, it is nevertheless interesting to envision the combination of wavelength-tunable or two-laser RPPI and tandem mass spectrometry. As noted earlier, discrimination in wavelength-tunable and two-laser R2PI is based on ionization selectivity, whereas fixed-wavelength R2PI/tandem mass spectrometry distinguishescompounds by the specificity of dissociation patterns. Although tandem mass spectrometry enabled the NJV-dimethylaniline and N-ethylaniline isomers to be differentiated in these experiments, the extent of discrimination for substitutional isomers may often be more limited due to similar fragmentation patterns. However, such isomers may have unique spectral characteristics or ionization thresholds that are sensitive to subtle structural differences. In such cases, wavelength-tunable or two-laser R2PI might be exploited to provide ionization selectivity. Although in some situations such optical selectivity can be achieved without cooling techniques ( 4 , 5 ) ,maximum ionization discrimination would be attained through the use of supersonic jet expansion methods. Such an instrumental combination would certainly be more complicated than the apparatus used in this work, yet ita complexity might be offset by increased analytical utility. Comparison of R2P1,EI,and CI. The exclusive formation of the molecular ion is not a prerequisite for an ionization method to be adaptable to tandem mass spectrometry. Such a requirement is obviated by the capability for discrete isolation of ions based on mass-to-charge ratio. However, the overall time required for MS/MS analysis of a mixture is generally related to the total number of ionic species generated during the ionization process. Therefore, a reduction in the number of different ionic species produced from a mixture is intrinsically advantageous for MS/MS analysis. Because the degree of molecular ion fragmentation associated with R2PI can readily be controlled, the ionization method is well-suited for application in tandem mass spectrometry. Chemical ionization (CI) is also used extensively in MS/MS analysis. It is often possible to restrict ionization to a single ion with CI, albeit not always the molecular ion. However, judicious selection of the reagant gas is required. Also note that electron attachment CI, which can have a sensitivity up to 2 orders of magnitude greater than that due to ion/molecule CI reactions, is problematic in the ion trap due to ita inability to store thermal electrons. In contrast, ionization via electron impact generally results in substantial internal energy deposition and subsequent fragmentation of the excited ion. Although the fragmentation can be controlled by varying the electron energy, this parameter is not easily regulated in the ITMS. In addition, the E1 cross section drops with electron energy as well, so that the resulting ionization efficiency also is reduced. In addition, low abundance of the molecular ion is seen in about one-fifth of the E1 spectra obtained for small organic molecules, thus complicating determination of the molecular mass for such compounds. RPPI has been demonstrated to have high sensitivity for aromatic molecules (29,30). In our experiments under R2PI conditions producing little fragmentation from N,N-di-
1191
methylaniline, the M + signal was typically = 15-20 times larger than the M + and (M - W)+ signals resulting from EI. The relative sensitivity for E1 and RBPI in the ITMS can be estimated. The number of analyte ions, nm, generated by E1 in the ITMS can be approximated by nEI = [%][I - exp(-aEINALEI)l where NAis the number density of analyte molecules, uEI is the E1 cross section, LEI is the E1 path length, and ne is the total number of electrons with sufficient energy to ionize A. Because u~NALEI
