TOF-MS: A Study of “Twin

While investigating the in-source CID fragmentation of nonsteroidal antiinflammatory drugs (NSAIDs), it was noticed that the same fragment ion (nomina...
0 downloads 0 Views 128KB Size
Anal. Chem. 2005, 77, 3394-3400

Technical Notes

Measuring the Mass of an Electron by LC/TOF-MS: A Study of “Twin Ions” Imma Ferrer* and E. Michael Thurman

Department of Analytical Chemistry, University of Almerı´a, 04120 Almerı´a, Spain

While investigating the in-source CID fragmentation of nonsteroidal antiinflammatory drugs (NSAIDs), it was noticed that the same fragment ion (nominal mass) formed in either positive or negative ion electrospray for a suite of NSAIDs. For example, ibuprofen with a molecular mass of 206, fragments to the m/z 161 ion in negative ion from its deprotonated molecule (m/z 205, [M - H]-) and fragments to the m/z 161 ion in positive ion from its protonated molecule (m/z 207, [M + H]+). This fragment ion was euphemistically called a “twin ion” because of the same nominal mass despite opposite charge. The CID fragmentation of the twin ions was confirmed also by LC/MS/MS ion trap. Accurate mass measurements in negative ion show that the loss was due to CO2 (measured loss of 43.9897 atomic mass units (u) versus calculated loss of 43.9898 u for N ) 10) and in positive ion the loss is due to HCOOH (measured loss of 46.0048 u versus calculated loss of 46.0055 u, N ) 10). It was realized that, in fact, the ions were not “identical mass twins of opposite charge” but separated in accurate mass by two electrons. The accurate mass measurement by liquid chromatography/time-of-flight-mass spectrometry (LC/TOF-MS) can distinguish between the two fragment ions of ibuprofen (161.13362 ( 0.00019 and 161.13243 ( 0.00014 for N ) 20). This experiment was repeated for two other NSAIDs, and the mass of an electron was measured as the difference between the twin ions, which was 0.00062 u ( 14.8% relative standard deviation (N ) 20 analyses). Thus, the use of continuous calibration makes it possible to measure the mass of an electron within one significant figure using the NSAID solution. This result shows the importance of including electron mass in accurate mass measurements and the value of a benchmark test for LC/TOF-MS mass accuracy. The analysis of pharmaceuticals in the environment has been a hot topic for environmental chemistry over the past 5 years,1-7 and in particular, the nonsteroidal antiinflammatory drugs have * Corresponding author. E-mail: [email protected]. (1) Ternes, T. A. Water Res. 1998, 32, 3245-3260. (2) Buser, H. R.; Poiger, T.; Muller, M. D. Environ. Sci. Technol. 1999, 33, 2529-2535. (3) Hirsch, R.; Ternes, T.; Haberer, K.; Kratz, K. Sci. Total Environ. 1999, 225, 109-118.

3394 Analytical Chemistry, Vol. 77, No. 10, May 15, 2005

received considerable attention because of their widespread use in over-the-counter medications.1-2 As a result of this environmental research interest, a number of papers have appeared recently on the LC/MS analysis of the nonsteroidal antiinflammatory drugs (NSAIDs).3-4,8-10 The methods used for the analysis of these compounds include all of the common LC/MS/MS analysis techniques, including triple quadrupole analysis,3,8 LC/ MS ion trap,5 and Q/TOF MS/MS.8 Furthermore, single quadrupole analysis has been used,4 and most recently, a comparison paper of triple quadrupole and Q/TOF MS/MS has appeared on the analysis of the NSAIDs.8 There is also a substantial, albeit, voluminous literature on the NSAIDs in pharmaceutical research, and the identification of metabolites of the NSAIDs has received attention in a recent abstract applying the use of LC/MS/MS ion trap.11 Thus, this family of compounds is important for analytical mass spectrometry. The widespread use of the NSAID attests to the importance in pharmaceutical chemistry, and indeed, ibuprofen itself is estimated as 1 of the top 10 drugs used worldwide. Coupled with this interest in the NSAIDs are the recent advances in LC/TOF-MS accurate mass analysis, which permits accurate mass analysis of NSAIDs and their metabolites for identification.5,10 We also are involved in the analysis of NSAIDs for surveys of pharmaceuticals,4,6-7 especially wastewater, using both LC/MS ion trap and LC/TOF-MS.5 Thus, we were interested (4) Kolpin, D. W.; Buxton, H.; Furlong, E.; Meyer, M.; Thurman, E. M.; Zaugg, S.; Barber, L., Jr. Environ. Sci. Technol. 2002, 36, 1202-1211. (5) Ferrer, I.; Thurman, E. M. Liquid Chromatography/Mass Spectrometry/Mass Spectrometry and Time-of-Flight Mass Spectrometry for the analysis of emerging contaminants, American Chemical Society Symposium Series 850, Oxford University Press: New York, 2003. (6) Ferrer, I.; Heine, C. E.; Thurman, E. M. Anal. Chem. 2004, 76, 14371444. (7) Thurman, E. M.; Heine, C. E.; Ferrer, I. Anal. Chem. 2004, 76, 14121417. (8) Marchese, S.; Gentili, A.; Perret, D.; D’Ascenzo, G.; Pastori, F. Rapid Commun. Mass Spectrom. 2003, 17, 879-886. (9) Benotti, M. J.; Ferguson, P. L.; Rieger, R. A.; Iden, C. R.; Heine, C. E.; Brownawell, B. J. HPLC TOF-MS: An alternative to LC/MS/MS for sensitive and selective determination of polar organic contaminants in the aquatic environment. In Liquid Chromatography/Mass Spectrometry/Mass Spectrometry and Time-of-Flight Mass Spectrometry for the Analysis of Emerging Contaminants; American Chemical Society Symposium Series 850; Oxford University Press: New York, 2003; Chapter 7, pp 109-127. (10) Ferrer, I.; Thurman, E. M. Trends Anal. Chem. 2003, 22, 750-756. (11) Ferrer, I.; Thurman, E. M. Abstract Book: Identification of emerging contaminants in wastewater and sediment samples by ion trap LC/MS/ MS, Pittcon 2003, Orlando, FL, 2003. 10.1021/ac0485942 CCC: $30.25

© 2005 American Chemical Society Published on Web 04/15/2005

in the fragmentation and analysis of the NSAIDs and the application of accurate mass. Typically accurate mass by LC/TOFMS has been at approximately the 5 ppm level for unknown compounds.5 Early instruments using LC/TOF-MS have a 2-m flight tube with an external corrected accuracy of ∼10 ppm. The accuracy was a function of changes in temperature that affect the length of the flight tube as well as problems with electronic drift. The introduction of the double sprayer with reference solutions corrected this problem of instrument drift by continuous calibration of the mass axis. This immediately improved mass accuracy to less than 5 ppm. Meanwhile improvements in counting electronics have further improved the signal-to-noise ratio so that the theoretical limits of LC/TOF-MS mass accuracy may be reached. Lee and Marshall12 have shown that the theoretical limit of LC/TOF-MS is a function of the mass resolving power and the signal-to-noise ratio. It is important to realize that LC/TOF-MS has lower resolution than FT-ICR mass spectrometry by a factor of 100 times or more, i.e., ∼10 000-20 000 resolution (at ∼1000 mass units) for LC/TOF-MS and 500 000 to the world record of 3 300 000 resolution.13 In fact, FT ICR has measured the mass resolution of two peptides at a difference that is less than the mass of an electron or ∼0.0005 u.13 Furthermore, simultaneous measurements of C60+ and C60- have been made by FT-ICR MS to show the mass resolution within the mass of 2 electrons.14 However, recent improvements in the measurement of mass accuracy by LC/TOF-MS has made it a useful tool for analytical analysis of small molecules (i.e., less than 500 mass units) that was in the past only the realm of FT-ICR MS. This paper will examine three pharmaceuticals in order to show that in clean samples, i.e., no background interfering ions present, mass accuracy may be able to reach the mass of an electron at one significant figure with two mass measurements, one in positive ion and one in negative ion electrospray. The mass of an electron is then measured by difference. Thus, a solution of the NSAIDs may be used as a benchmark solution for the measurement of accurate mass in both positive and negative ions for both molecular and fragment ions. In general, the LC/MS analysis of the NSAIDs has been carried out in negative ion electrospray as the method of choice for these compounds.3,4,8-10 In fact, we were not able to find any published examples of the use of electrospray positive for the analysis of the NSAIDs. The classic fragmentation in negative ion is the loss of carbon dioxide (with a nominal mass of 44), as pointed out by a recent paper on negative ion fragmentation of acidic pharmaceuticals.15 Furthermore, we have been interested in the electrospray process of compounds that ionize in both positive and negative ion electrospray ionization (ESI), especially pesticides, at unfavor(12) Lee, H.-N.; Marshall, A. G. Anal. Chem. 2000, 72, 2256-2260. (13) He, F.; Hendrickson, C. L.; Marshall, A. G. Anal. Chem. 2001, 73, 647650. (14) Schweikhard, L.; Drader, J. J.; Shi, S. D.-H.; Hendrickson, C. L.; Marshall, A. G. Quadrature detection for the separation of the signals of positive and negative ions in Fourier transform ion cyclotron resonance mass spectrometry. In Non-Neutral Plasm Physics V; Anderegg, F., Schweikhard, L.; Driscoll, C. F., Eds. American Institute of Physics: College Park, MD, 2002; pp 547-652. (15) Bandu, M. L.; Watkins, K. R.; Bretthauer, M. L.; Moore, C. A.; Desaire, H. Anal. Chem. 2004, 76, 1746-1753.

able regions of pH (called wrong way around electrospray16). Thus, this paper on NSAIDs extends our work to pharmaceuticals, especially the acidic compounds and their fragmentation in both positive and negative ion electrospray. This paper will discuss first the use of LC/MS ion trap to show the formation of fragment ions of the same nominal mass in positive and negative ion and their pathway. Second, the paper addresses the use of LC/TOFMS to measure the accurate mass of fragment ions and the importance of including the mass of the electron in ion fragmentation studies. Third, the paper shows that recent advances in continuous mass calibration allow the measurement of the electron to one significant figure using LC/TOF-MS, when combined with the analysis of the NSAID solution. EXPERIMENTAL SECTION Chemicals and Reagents. High-performance liquid chromatography (HPLC)-grade acetonitrile and methanol were purchased from Merck (Darmstadt, Germany). A Milli-Q-Plus ultrapure water system from Millipore was used to obtain the HPLC-grade water used during the analyses. Formic acid was obtained from Fluka. The NSAID standards (ibuprofen, naproxen, diclofenac) were obtained from Chem Service (Philadelphia. PA), dissolved in methanol, and analyzed at concentrations of 10 ppm by LC/TOFMS for electron mass measurements. Subsequently, ketorolac and ketoprofen were also obtained from Chem Service, dissolved in methanol, and analyzed at 10 ppm by LC/MS ion trap for further testing of twin ion formation. LC/MS Ion Trap Method. The analytes were separated using an HPLC (series 1100, Agilent Technologies, Palo Alto, CA) equipped with a reversed-phase C8 analytical column (Zorbax Eclipse XDB 4.6 × 150 mm, 5 µm). Column temperature was maintained at 25 °C. Mobile phase A was acetonitrile, and mobile phase B consisted of 0.1% formic acid. A linear gradient progressed from 50 (initial conditions) to 100% A in 10 min, after which the mobile-phase composition was maintained at 100% A for 2 min. The flow rate was 0.6 mL/min, and 50 µL of the standard solutions was injected. This HPLC system was connected to an ion trap mass spectrometer LC/MS/MS (Esquire LC, Bruker Daltonics, Bellerica, MA) system equipped with an ESI probe. The maximum ion accumulation time was set at 200 ms, capillary voltage was 3200 V, nebulizer 45 psig, drying gas 12 L/min, gas temperature 300 °C, and fragmentor 70 V. Manual MS/MS was carried out for the [M + H]+ or [M - H]- ion. The analysis by LC/MS ion trap was carried out by running each sample in positive ion and in negative ion full scan and in MS/MS mode. LC/TOF-MS Method. The same chromatographic gradient and analytical column as above was used for the LC/TOF-MS method. The HPLC system was connected to a time-of-flight mass spectrometer (MSD-TOF, Agilent Technologies, Santa Clara, CA) equipped with an electrospray interface under the following operating parameters in positive ion mode: capillary 4000 V, nebulizer 40 psig, drying gas 9 L/min, gas temperature 300 °C, fragmentor 215 V, and skimmer 60 V, Oct DC1 35V, OCT RF V 250 V. The mass axis was calibrated using the mixture provided by the manufacturer over the m/z 50-3200 range. A second orthogonal sprayer with a reference solution was used as a continuous calibration in positive ion using the following reference (16) Thurman, E. M.; Ferrer, I.; Barcelo´, D. Anal. Chem. 2001, 73, 5441-5449.

Analytical Chemistry, Vol. 77, No. 10, May 15, 2005

3395

Figure 1. Full scan in-source CID spectra of NSAIDs.

masses: 121.0509 and 922.0098 m/z (resolution: 9700 ( 500 at 922.0098 m/z). Reference A sprayer 2 flows at a constant rate during the run. Mass peaks were determined at the center of the mass peak using the width of the peak at half-peak height as the standard method by computer-generated algorithms so that human bias was removed. A handpicking method was also used as a comparison with the computer-generated algorithm using the same criteria. With the electrospray source in negative (ESI-), reference masses were 112.9856 and 1033.9881 m/z; resolution, 10 200 ( 500 at 1033.9881 m/z. The analysis for electron mass by LC/TOF-MS was completed with 20 chromatographic runs in positive ion over an 18-h period. No calibration of the instrument was done between samples. The instrument was then converted from electrospray positive to electrospray negative the following day and recalibrated, and the sample was analyzed 20 times again over a 18-h period. RESULTS AND DISCUSSION Twin Ion Formation in LC/MS Ion Trap. The mass spectra of the NSAIDs and the formation of twin ion spectra for ibuprofen, naproxen, and diclofenac are shown in Figure 1 for both positive and negative ion electrospray using LC/MS ion trap. First, the in-source collision-induced dissociation (CID) of ibuprofen gave the m/z 207 in positive ion [M + H]+ and the m/z 205 in negative ion [M - H]-. The major fragment ion for both positive and negative ions was the m/z 161 (Figure 1). Because the nominal mass of the fragment ion is the same, it is given the name of twin ion. The fragmentation of ibuprofen in positive ion electrospray results in the loss of a neutral fragment (HCOOH, formic acid) 3396

Analytical Chemistry, Vol. 77, No. 10, May 15, 2005

Figure 2. Inductive (i) cleavage of ibuprofen [M + H]+ or [M - H]to form the twin ions in either positive or negative ion electrospray.

and the transfer of charge from the protonated carboxyl group to the secondary benzylic carbon atom (Figure 2). The same process occurs in negative ion electrospray except that the neutral fragment that is lost is CO2 and again the negative charge migrates to the secondary benzylic carbon atom. In both cases for ibuprofen, the mechanism of ion formation is an inductive cleavage, which is initiated by the charged site and involves the migration of charge and cleavage of a single bond.17 Because both the precursor ion and the twin ion are even electron ions, a neutral loss is involved in order to maintain the even electron ion.17 The same mechanism of fragmentation is valid for both naproxen and diclofenac. Note that the formation of the twin ion (17) McLafferty, F. W. Interpretation of Mass Spectra; University Science Books: Mill Valley, CA, 1980; pp 56-57.

Table 1. Accuracy of CO2 and HCOOH Neutral Losses compound ibuprofen naproxen diclofenac average error from true value

compound ibuprofen naproxen diclofenac average error from true value

Figure 3. Chemical structures of NSAIDs and twin ion examples by MS/MS.

is also a major ion in the naproxen spectra, similar to ibuprofen (Figure 1). However, the twin ion is of less intensity with diclofenac in either positive or negative ion (m/z 250), which is probably due to the structural difference of diclofenac, which does not have the tertiary carbon R to the carboxyl group that contains the charge. The fragmentation experiments were repeated using the LC/ MS ion trap in MS/MS mode to confirm the formation of the twin ions from the protonated or deprotonated molecule. In all three compounds, the twin ion was the major product ion in either positive or negative ion electrospray, which was further evidence of the importance of this fragmentation, whether it is in-source CID or MS/MS fragmentation in the ion trap. Thus, the aromatic ring plays an important role in the stabilization of the charged fragment ion, either positive or negative ion. It was realized that this feature of ion formation may be useful in the identification of unknown compounds because of implied structural information from the losses of both HCOOH and CO2. Because of the common structural features of an aromatic ring and carboxyl group adjacent to a tertiary carbon, we quickly realized that the family of NSAIDs probably formed twin ions. This idea was tested by further MS/MS analysis in both positive and negative ions for two more common NSAIDs, ketoprofen and ketorolac (Figure 3). Ketoprofen is similar to ibuprofen and naproxen, which have similar structures in that they have a carboxyl group that is adjacent to an aromatic ring with a tertiary carbon to accept the charge transfer. Ketorolac is different in that it has a tertiary carbon from a five-membered ring and that the aromatic ring is heterocyclic. This structure also was able to stabilize the charged positive and negative ions. Diclofenac is different too in that it has a chlorinated structure and no tertiary carbon. Apparently the nitrogen bridge atom between the two aromatic rings plays a similar role in stabilizing the twin ion. All five compounds have the carboxyl group one carbon removed from an aromatic ring and all compounds contain at least one

CO2 mass calcd, u

CO2 mass obsd, u, av N ) 10

CO2 mass obsd, std dev, N ) 10

43.98983

43.98971 43.98984 43.98997 43.98984 0.00001

0.00012 0.00006 0.00006 0.00008

HCOOH mass calcd, u

HCOOH mass obsd, u, av N ) 10

HCOOH mass obsd, std dev, N ) 10

46.00548

46.00478 46.00524 46.00483 46.00495 -0.00053

0.00065 0.00016 0.00027 0.00036

aromatic ring. These structural features are important to the twin ion formation and the stability of the retained charge. The idea of fragment ions of the same nominal mass has also been reported in GC/MS using electron impact ionization18 and is, in fact, an old idea. The classic example is the case of nitrobenzene, which gives both molecular ions that are the same nominal mass and fragment ions that are the same nominal mass, (i.e., phenoxyl cation and anion fragments; see Acknowledgment). Accurate Mass of Neutral Losses. The accurate mass differences between the precursor ion (protonated or deprotonated molecule) and the twin ion can be used to determine the neutral loss by difference, using accurate mass. Table 1 summarizes the results of accurate mass analysis of ibuprofen, naproxen, and diclofenac for the neutral loss of CO2 and HCOOH and gives the mean and standard deviation for 10 injections. The negative ion results for the loss of CO2 was 43.989 71 u for ibuprofen, 43.989 84 u for naproxen, and 43.989 97 u for diclofenac, with the average value for CO2 loss of 43.989 84 u, which is within 0.01 mmu of the correct value. The values for HCOOH neutral losses were less accurate with the average value of 46.004 95 u versus the calculated value of 46.005 48 u for an error of 0.53 mmu. The greater error in positive ion was probably due to the lower signal for the protonated molecules of ibuprofen (nominal mass of m/z 207 in Figure 4) and protonated diclofenac (data not shown but intensity of counts was ∼16 000 counts compared to 60 000 counts in negative ion). These results indicated that the LC/TOF-MS system was well calibrated and stable, since these measurements were taken over a 18-h period without external calibration. There was external calibration applied between positive and negative ion according to manufacturer’s specifications, and the instrument was autotuned to within 1-5 ppm residual errors before injection of negative ion samples. Positive ion results were acquired on day 1 and negative ion results on day 2. The best accuracy was obtained with signals above 50 000 counts of peak height intensity, which were used for electron mass calculations as follows. Measuring Electron Mass. The absolute value between the twin ions for each of the three compounds (ibuprofen, naproxen, (18) Bowie, J. H. Mass Spectrom. Rev. 1984, 3, 1-37.

Analytical Chemistry, Vol. 77, No. 10, May 15, 2005

3397

Figure 4. Accurate mass spectra of ibuprofen in both positive and negative ion electrospray.

diclofenac) actually differs in accurate mass by two electrons. This is due to the fact that a protonated molecule of ibuprofen, naproxen, or diclofenac ionizes in electrospray as the addition of a proton, which of course is the mass of the hydrogen atom minus an electron (1.00780-0.00055 u). Conversely, the negative ion of the compound forms in electrospray from the loss of a proton leaving an electron behind. Although we write the protonated molecule as [M + H]+ and the deprotonated molecule as [M H]-,19 this notation is actually slightly incorrect for accurate mass calculations because the addition or the loss is a proton and not a hydrogen atom. Because the twin ion forms from its precursor ion by a neutral loss of CO2 or HCOOH, it is an even electron ion also.17 Thus, the charge remains with the fragment ion, and the absolute mass of the twin ions differ by 2 electrons. The negative ion contains two more electrons than the positive ion and should give an absolute, theoretical mass difference of 0.0011 u greater. Thus, it may be possible to measure this difference in mass if the accuracy and stability of the LC/TOF-MS is sufficient. Figure 4 shows an example of the mass of a single analysis of ibuprofen in positive and negative ion and the exact mass of the twin ion (nominal mass m/z 161). Note that the accurate masses for the twin ions differ by ∼0.0012 u (161.1336-161.1324). This is an example of a single measurement from the following set of data (Table 2). (19) Sparkman, D. L. Mass Spectrometry Desk Reference; Global Publishing: Pittsburgh, 2002; p 47.

3398

Analytical Chemistry, Vol. 77, No. 10, May 15, 2005

Table 2. Mean Value, Error of Mean Value, Standard Deviation, and Calculations of Electron Mass from the Difference of the Absolute Mass of the Twin Ionsa masses 185+ 185161+ 161250+ 250mean electron mass (N ) 60) std dev % rel std dev electron mass20 accepted value

mean value, m/z

av error, ppm

std dev, u

electron mass, u

185.09597 185.09721 161.13243 161.13362 250.01825 250.01955

-0.62 0.13 -0.31 0.42 -0.92 -0.13

0.00020 0.00017 0.00014 0.00019 0.00017 0.00008

0.00062 0.00059 0.00065

0.00062 0.00009 14.8 0.00055

a Computer algorithm was used for all calculations from the raw data.

Table 2 shows the summary table for the accurate mass measurements of the twin ion in positive and negative ion for 20 replicate measurements of the three NSAIDs. The calculation of the electron mass is carried out by taking the difference between the twin ions and dividing by 2. The mean values of electron mass for ibuprofen, diclofenac, and naproxen were 0.00059, 0.00065, and 0.00062 u, respectively, with an overall average of 0.00062 u and

Table 3. Calculation of Electron Mass from the Difference of the Twin Ions Based on Each Chromatographic Run (Negative - Positive) Containing Ibuprofen, Diclofenac, and Naproxen Using Hand Calculation chromatographic numbera

statistical calculations

1 2 3 4 5 6 7 8 9 10 mean electron mass, N ) 30 std dev rel std dev, % electron mass20

mean electron mass, N ) 3b 0.00057 0.00047 0.00048 0.00056 0.00055 0.00051 0.00057 0.00058 0.00057 0.00051 0.00053

% error 2.9 14.2 12.4 0.9 0.9 8.0 2.8 5.0 2.8 7.3

0.00004 7.5 0.00055

a Two chromatographic runs (one for positive ion and one for negative ion) were used to calculate the electron mass for each one of the three compounds. b The electron mass was calculated as an average of the three values obtained for the three compounds.

a standard deviation of 0.00009. The currently accepted value for electron mass is 0.000548 579 909 2 u;20 thus, at two significant figures the value is 0.00055, which is a measured error of ∼13% for the LC/TOF-MS measurements, based on 20 injections and a total of 60 calculations of electron mass. A computer software program was used to pick the accurate mass to avoid human bias. As a comparison, handpicking of the accurate mass of the first 10 samples, which is the commonly used practice in accurate mass determinations, resulted in a value of 0.000 59 with a relative error of 7%. However, for purposes of this experiment, the computer software pick was used to remove any preconceived bias for accurate mass, although there was not statistical difference between the electron mass using the two procedures. We realize, of course, that 20 injections in both positive and negative ion ESI may not be a practical way of estimating instrument accuracy. Therefore, Table 3 shows the results of electron mass measurements as a function of a single analysis in positive ion and a single analysis in negative ion for the three NSAIDs using the handpicking procedure. The range of values was from 0.00047 to 0.00058 u with an average value of 0.00053 u. The average error absolute was 5.7%, and the largest error was 14.2% (Table 3). The relative standard deviation of the 10 runs was 7.5%. Thus, we propose that a solution such as the NSAIDs be used for determination of accuracy of LC/TOF-MS in positive and negative ion and that at least three compounds be analyzed as a minimum for testing instrument accuracy using either the handpicking procedure or a computer software procedure. We realize, of course, that these measurements are the best possible instrument values in a clean system and represent the errors associated only with the precision of mass measurements by TOF mass spectrometry. Routine analyses may not reach these values (20) Werth, G.; Haffner, H.; Kluge, H. J.; Quint, W.; Valenzuela, T.; Verdu, J. Hyperfine Interactions 2001, 132, 209-213.

Table 4. Resolution in Positive and Negative Ion Electrospray for LC/TOF-MS

compound

nominal mass ion, m/z

ibuprofen twin ion ibuprofen [M - H]ibuprofen [M + H]+ naproxen twin ion naproxen [M - H]naproxen [M + H]+ diclofenac twin ion diclofenac [M - H]diclofenac [M + H]+ tune ion (+) tune ion (-)

161 205 207 185 229 231 250 294 296 922 1034

resolution (+) ion 3800 4300 3800 4600 4700

resolution (-) ion 4800 5400 5100 5600 5800 6500

5600 9700 10 200

because of interferences from sample matrixes and associated errors. However, further improvements in TOF resolution will continue and may make routine sub 1 ppm measurements possible. Resolution and Accuracy. In the accuracy data given above, the resolution at full width half-maximum (fwhm) varied from 3800 for the lowest mass twin ion (m/z 161 in positive ion) to 5800 for the largest mass twin ion (m/z 250 in negative ion), as shown in Table 4. Furthermore, it was noted that the resolution in general was 9700-10 200 for either positive or negative ion ESI at a nominal mass of m/z 922-1034 (see Table 4). The negative ion resolution was slightly greater than the positive ion resolution for each twin ion by a factor of ∼1.2 times. The mass calibration table for positive ion ESI had a tune file of mass error of 0.9-2.8 ppm, and for negative ion ESI, the mass error was from 2.2 to 5.7 ppm. These are, in our experience, standard tune files for nominal masses from m/z 113 to 1034 and they do not represent a “special case” of accurate mass measurements. It is possible to standard tune the instrument to less than 1 ppm residual error, but not necessary because the reference solution provides the final calibration, as explained later. Because the LC/TOF-MS system pulses the ions through the TOF flight tube every 100 µs in these experiments, 10 000 pulses (called transients)/s are accumulated in each “scan”. Each chromatographic peak was ∼10 s wide; thus, ∼100 000 transients were totaled for the data in these electron mass calculations for each chromatographic analysis. Finally, it is interesting to compare our data with recent papers dealing with accuracy and LC/TOF-MS.6,21-22 For example, Balogh21 in a recent review of resolution and mass accuracy stated that more resolution is needed to obtain the mass accuracies of ion cyclotron MS, which are typically from