Determination of the Nature of Naphthenic Acids Present in Crude Oils

Jan 23, 2003 - ing theoretical masses using Microsoft Excel and also using the. MassAnal ... cal isotope patterns was also made in order to further co...
1 downloads 0 Views 245KB Size
Download PDF
Anal. Chem. 2003, 75, 860-866

Determination of the Nature of Naphthenic Acids Present in Crude Oils Using Nanospray Fourier Transform Ion Cyclotron Resonance Mass Spectrometry: The Continued Battle Against Corrosion Mark P. Barrow,† Liam A. McDonnell,† Xidong Feng,† Je´re´mie Walker,‡ and Peter J. Derrick*,†

Institute of Mass Spectrometry and Department of Chemistry, University of Warwick, Coventry, CV4 7AL, United Kingdom, and Service Analyse, ATOFINA Centre de Recherche Rhoˆ ne-Alpes, Rue Henri Moissan, B.P. 63, 69493 Pierre-Be´ nite Cedex, France

Recent research has shown that the corrosivity of naphthenic acids is related to their molecular mass and that the “total acid number” (TAN), traditionally used as an indicator of the naphthenic acid content of an oil, is not as reliable as first believed. The presence of naphthenic acids in crude oils leads to the corrosion of oil refinery equipment, with the oil industry incurring costs that will ultimately be passed on to the consumer. With regard to these concerns, mass spectrometry has been increasingly applied to the investigation of the naphthenic acid content of crude oils. To ascertain the nature of the species present, however, it is necessary to utilize an ionization technique that does not result in fragmentation, ensuring the detection only of molecular species which provide useful information about the sample constitution. In the following investigation, negative ion mode nanospray Fourier transform ion cyclotron resonance (FTICR) mass spectrometry has been applied to the analysis of crude oil samples, providing insight into the different acidic species that were present. Use of the negative ion mode to allow the selective observation of the naphthenic acids and the inherent high mass accuracy and ultrahigh resolution of FTICR mass spectrometry ensure that this technique is very well suited to the characterization of naphthenic acids within a crude oil sample. Determination of the nature of the naphthenic acids present provides vital information, such as the acids’ sizes and composition, which may be used in the battle against corrosion and also used to fingerprint samples from different oil fields. One of the greatest concerns of the oil refining industry is the presence of naphthenic acids in crude oils. Naphthenic acids have been defined as carboxylic acids that include one or more saturated ring structures, where five- and six- membered rings * To whom correspondence should be addressed. Fax: +44 (0)24 76523819. E-mail: P. J. [email protected]. † University of Warwick. ‡ Service Analyse.

860 Analytical Chemistry, Vol. 75, No. 4, February 15, 2003

are most common, though the definition has become more loosely used to described the range of organic acids found within crude oil. The empirical formulas for the acids may be described by CnH2n+zO2,1-5 where z is referred to as the “hydrogen deficiency” and is a negative, even integer, and more than one isomer will typically exist for a given z homologue. Crude oil typically contains naphthenic acids in quantities of up to 4 wt % and characterization of the acids present within a sample has become a topic of great interest because of the fact that the acids corrode refinery units and, therefore, present the oil industry with significant additional costs. One important marker used to gauge the acidity of oils is the “total acid number” (TAN), which is defined as the mass of potassium hydroxide (in milligrams) required to neutralize one gram of crude oil. The TAN of an oil has frequently been used to attempt to quantify the presence of naphthenic acids, because the carboxylic acid components of oils are believed to be largely responsible for oil acidity. However, more recent research has begun to highlight deficiencies in relying upon this method for such direct correlation,6-8 and the total acid number is no longer considered to be such a reliable indicator. It is therefore essential that a more reliable method is found for characterizing the naphthenic acids within a crude oil in order that corrosion can be better understood. The naphthenic acid content of crude oils is believed to be on the rise due to the increasing use of “opportunity crude oils.”7 It is agreed that naphthenic acid corrosion of refinery units is of increasing significance to industry, and it has been claimed that savings of several dollars per barrel could be made if the (1) Dzidic, I.; Somerville, A. C.; Raia, J. C.; Hart, H. V. Anal. Chem. 1988, 60, 1318-1323. (2) Fan, T.-P. Energy & Fuels 1991, 5, 371-375. (3) Wong, D. C. L.; van Compernolle, R.; Nowlin, J. G.; O’Neal, D. L.; Johnson, G. M. Chemosphere 1996, 32, 1669-1679. (4) St. John, W. P.; Rughani, J.; Green, S. A.; McGinnis, G. D. J. Chromatogr., A 1998, 807, 241-251. (5) Hsu, C. S.; Dechert, G. J.; Robbins, W. K.; Fukuda, E. K. Energy Fuels 2000, 14, 217-223. (6) Turnbull, A.; Slavcheva, E.; Shone, B. Corrosion 1998, 54, 922-930. (7) Slavcheva, E.; Shone, B.; Turnbull, A. Br. Corr. J. 1999, 34, 125-131. (8) Meredith, W.; Kelland, S.-J.; Jones, D. M. Org. Geochem. 2000, 31, 10591073. 10.1021/ac020388b CCC: $25.00

© 2003 American Chemical Society Published on Web 01/23/2003

corrosivity of the acids could be defined properly.6 Nuclear magnetic resonance (NMR), Fourier transform infrared (FT-IR), and mass spectrometry have been used to study nature of acidic species within crude oils.9 Turnbull et al. have shown that the size and structure of naphthenic acids influence their corrosivity.6 In addition to corrosion of refinery equipment, another concern relating to naphthenic acids is their environmental impact. Wong et al. showed that naphthenic acids were toxic to fish, and the use of granular activated carbon (GAC) to effectively filter the water dramatically increased the survival rate.3 Fossil fuels and complex hydrocarbon mixtures have long been a topic of investigation for mass spectrometrists,10-17 and various separation methods have been evaluated for general use with complex hydrocarbon mixtures.18-20 One of the greatest difficulties posed when characterizing the naphthenic acids found in crude oils is that it is not possible to extract the many naphthenic acids species as separate components for analysis,1,3,21 and the resulting mass spectra are, therefore, very complex. A variety of mass spectrometric techniques have been used for the study of the acids within a sample, including gas chromatography mass spectrometry (GC/MS),1,4,21 electron ionization (EI),4 liquid secondary ion mass spectrometry (LSI-MS),3 fast atom bombardment (FAB),2 chemical ionization (CI),1,2,4,5 atmospheric pressure chemical ionization (APCI),5 and recently, electrospray ionization (ESI).22-24 Though sector and FTICR instruments have been most commonly applied to investigations of complex sample mixtures,25,26 time-of-flight (TOF) mass spectrometry has also been used.27 Electrospray ionization is the ionization technique that holds the greatest promise for the successful characterization of naphthenic acids in crude oil samples, because the lack of fragmentation should in turn provide a more accurate portrayal of the sample constituents. The analysis of such hydrocarbon mixtures can, however, pose a challenge as result of their comparatively ESI-inactive characteristics.28 Zhan and Fenn have remarked upon the lack of publications detailing the use of electrospray ionization for the analysis (9) Tomczyk, N. A.; Winans, R. E.; Shinn, J. H.; Robinson, R. C. Energy Fuels 2001, 15, 1498-1504. (10) Brown, R. A. Anal. Chem. 1951, 23, 430-437. (11) Clerc, R. J.; Hood, A.; O’Neal, M. J., Jr. Anal. Chem. 1955, 27, 868-875. (12) Gallegos, E. J.; Green, J. W.; Lindeman, L. P.; LeTourneau, R. L.; Teeter, R. M. Anal. Chem. 1967, 39, 1833-1838. (13) Aczel, T.; Allan, D. E.; Harding, J. H.; Knipp, E. A. Anal. Chem. 1970, 42, 341-347. (14) Schmidt, C. E.; Sprecher, R. F.; Batts, B. D. Anal. Chem. 1987, 59, 20272033. (15) Guan, S.; Marshall, A. G.; Scheppele, S. E. Anal. Chem. 1996, 68, 46-71. (16) Rodgers, R. P.; White, F. M.; Hendrickson, C. L.; Marshall, A. G.; Andersen, K. V. Anal. Chem. 1998, 70, 4743-4750. (17) Rodgers, R.; Blumer, E. N.; Freitas, M. A.; Marshall, A. G. Environ. Sci. Technol. 2000, 34, 1671-1678. (18) Qian, K.; Hsu, C. S. Anal. Chem. 1992, 64, 7-2333. (19) Bennett, B.; Larter, S. R. Anal. Chem. 2000, 72, 1039-1044. (20) Jones, D. M.; Watson, J. S.; Meredith, W.; Chen, M.; Bennett, B. Anal. Chem. 2001, 73, 703-707. (21) Green, J. B.; Stierwalt, B. K.; Thomson, J. S.; Treese, C. A. Anal. Chem. 1985, 57, 2207-2211. (22) Miyabayashi, K.; Suzuki, K.; Teranishi, T.; Naito, Y.; Tsujimoto, K.; Miyake, M. Chem. Lett. 2000, 172-173. (23) Miyabayashi, K.; Yasuhide, N.; Miyake, M.; Tsujimoto, K. Eur. J. Mass Spectrom. 2000, 6, 251-258. (24) Zhan, D.; Fenn, J. B. Int. J. Mass Spectrom. 2000, 194, 197-208. (25) Szulejko, J. E.; Solouki, T. Anal. Chem. 2002, 74, 3434-3442. (26) Kujawinski, E. B.; Freitas, M. A.; Zang, X.; Hatcher, P. G.; Green-Church, K. B.; Jones, R. B. Org. Geochem. 2002, 33, 171-180. (27) Dutta, T. K.; Harayama, S. Anal. Chem. 2001, 73, 864-869. (28) Roussis, S. G.; Proulx, R. Anal. Chem. 2002, 74, 1408-1414.

of fossil-fuel-based samples,24 which served to highlight the challenge posed. A variation of ESI29,30 that is more attractive for this purpose is nanoelectrospray ionization, also known as “nanospray.”31 Nanospray entails the use of a fine glass needle, coated with a conductive layer, such as gold or palladium, into which the sample is loaded. This method of ionization offers the advantages of electrospray ionization and is also the most efficient of these techniques for ion production, requires fewer mechanical components, and consumes a much lesser sample amount. In the case of crude oil analysis, the particular advantage is that the low sample throughput also minimizes the risk of contamination of the instrument. Fan stated that resolution was a limiting factor during characterization of naphthenic acids.2 Fourier transform ion cyclotron resonance mass spectrometers (FTICR-MS or FTMS)32-35 are capable of high resolving power and high mass accuracy, affording a higher confidence in the mass assignments. Perhaps the most commonly known disadvantages associated with FTICR mass spectrometry are collectively known as space-charge effects,36-42 which result from interactions between ion packets in the cell and can lead to decreased resolution and mass accuracy. Though such space-charge effects exist, they can be minimized through user expertise, maintaining FTICR mass spectrometry’s reputation as the undisputed leader in terms of mass accuracy and resolution. A 9.4-T FTICR mass spectrometer was used during the course of the following investigation in order to illustrate the advantage of routine high resolution, permitting the distinction to be made between nominally isobaric acid species. Nanospray has been selected as the ionization technique to combine the advantages of minimized sample fragmentation, the inherent sensitivity of nanospray, and minimization of the risks of instrument contamination. Two West African crude oil samples were analyzed using the negative ion mode to determine the presence of different naphthenic acid species. The names of the samples have been changed to protect industrial interests, so the samples are referred to as the samples from oil field A and oil field B throughout. Spectra obtained in the positive ion mode contained approximately three times the number of signals, as species other than the acids could be observed, and were thus more complex with regard to the data analysis stage. As further corroborated by recent work (29) Yamashita, M.; Fenn, J. B. J. Phys. Chem. 1984, 88, 4451-4459. (30) Yamashita, M.; Fenn, J. B. J. Phys. Chem. 1984, 88, 4671-4675. (31) Wilm, M.; Mann, M. Anal. Chem. 1996, 68, 1-8. (32) Amster, I. J. J. Mass Spectrom. 1996, 31, 1325-1337. (33) Marshall, A. G.; Hendrickson, C. L.; Jackson, G. S. Mass Spectrom. Rev. 1998, 17, 1-35. (34) Marshall, A. G.; Schweikhard, L. Int. J. Mass Spectrom. Ion Processes 1992, 118/119, 37-70. (35) Dienes, T.; Pastor, S. J.; Schu ¨ rch, S.; Scott, J. R.; Yao, J.; Cui, S.; Wilkins, C. L. Mass Spectrom. Rev. 1996, 15, 163-211. (36) Chen, S.-P.; Comisarow, M. B. Rapid Commun. Mass Spectrom. 1991, 5, 450-455. (37) Chen, S.-P.; Comisarow, M. B. Rapid Commun. Mass Spectrom. 1992, 6, 1-3. (38) Guan, S.; Wahl, M. C.; Marshall, A. G. Anal. Chem. 1993, 65, 3647-3653. (39) Perrung, A. J.; Kouzes, R. T. Int. J. Mass Spectrom. Ion Processes 1995, 145, 139-153. (40) Naito, Y.; Inoue, M. Int. J. Mass Spectrom. Ion Processes 1996, 157/158, 85-96. (41) Stults, J. T. Anal. Chem. 1997, 69, 1815-1819. (42) Easterling, M. L.; Mize, T. H.; Amster, I. J. Anal. Chem. 1999, 71, 624632.

Analytical Chemistry, Vol. 75, No. 4, February 15, 2003

861

by Marshall and co-workers,43 use of the negative ion mode allows analysis to be focused on the (deprotonated) acidic species. The mass accuracy associated with the aforementioned instrumentation (typically in the region of 1 ppm or less with such samples) allows more confident assignment of the different species than could be provided using other varieties of mass analyzer. The combination of the high mass accuracy, ultrahigh resolution, and selective observation of the deprotonated naphthenic acids make negative ion mode FTICR mass spectrometry the technique of choice for the characterization of the naphthenic acids within a crude oil. Once the spectra were acquired, the data was plotted in terms of relative signal intensity as a function of carbon content for a range of empirical formulas, providing an overview of the naphthenic acids present within the crude oil samples. The ability to determine the empirical formulas of the acidic species contained within a given crude oil is of relevance to both the continued fight against corrosion within the petroleum industry and also to environmental concerns. EXPERIMENTAL SECTION Throughout the course of the experiments, a BioApex II (Bruker Ltd., U.K.) 9.4-T Fourier transform ion cyclotron (FTICR) resonance mass spectrometer44 was used. The heart of the instrument was the Infinity Cell45 which was cylindrical in geometry, rather than the traditional cubic. Radio frequency (rf) potentials were applied to segmented trapping electrodes to produce more homogeneous electrostatic fields from the excitation plates, simulating an infinitely long cell. The advantage of this cell design was the minimization of ion loss arising from the effect of radial components of the excitation field; such ion loss can lead to discrimination effects during the excitation stage of an experiment. For these experiments, a trapping potential (PV1 and PV2) of -1.5 V was maintained to constrain the ions’ axial movement within the cell, and excitation was performed over a mass range of m/z 115-1500. Ions were generated using the nanospray ion source and then retained within a hexapole ion trap for a period of 4 s (D1 ) 4 s) prior to extraction. There was a delay after the extraction of ions from the hexapole during which the “SideKick” electrodes (EV2 and DEV2) were raised to higher potentials in order to increase the trapping efficiency of the cell; at the end of the set time period, the potentials of the electrodes were brought into line with the potentials on the trapping plates. This delay, referred to as P2, was set to 2400 µs. The SideKick technique was employed to deflect ions off-axis as they entered the cell (EV1 ) 2.86 V, EV2 ) 1.74 V (where EV values correspond to ICR cell extraction plates), DEV2 (ICR cell delta extraction plate 2) ) -1.05 V). Dipolar excitation was used to excite the ions to a detectable cyclotron orbit prior to the detection stage. The excitation step size was ∼5.2 kHz, the width of the frequency sweep was 1.25 MHz, the duration of each frequency step component of the rf chirp (P3) was set to 40 µs, and the rf attenuation for the excitation (PL3) was set to 5 dB, which corresponds to an excitation potential (43) Qian, K.; Robbins, W. K.; Hughey, C. A.; Cooper, H. J.; Rodgers, R. P.; Marshall, A. G. Energy Fuels 2001, 15, 1505-1511. (44) Palmblad, M.; Hakansson, K.; Hakansson, P.; Feng, X.; Cooper, H. J.; Giannakopulos, A. E.; Green, P. S.; Derrick, P. J. Eur. J. Mass Spectrom. 2000, 6, 267-275. (45) Caravatti, P.; Allemann, M. Org. Mass Spectrom. 1991, 26, 514-518.

862

Analytical Chemistry, Vol. 75, No. 4, February 15, 2003

inside the cell (6 cm in diameter) of ∼150 V. The instrument was controlled using a Silicon Graphics Indy workstation running XMASS 5.0.6 (Bruker Ltd., U.K.) under IRIX. 5.3. Data files consisted of 512 K (524288) data points and represent 128 scans. The raw data was then converted from the time domain to the frequency domain via a fast Fourier transform. Two crude oil samples labeled oil field A and oil field B were made up into solutions using the ratio of 0.1 mg of crude oil per 1 mL of acetonitrile. Ammonia solution (35 vol % in water) was added as 1% of the total crude oil solution to assist deprotonation of the naphthenic acid species. The resulting solutions were used for nanospray analysis. The nanospray apparatus was based upon the existing electrospray ion source (Analytica of Branford, U.S.A.) where a metal-coated glass needle (Protana Engineering, Denmark) replaced the arrangement of a metal needle connected in series to a syringe pump. Sample solution (10 µL) was transferred to the glass needle, the needle was secured within a mount, and a gas line was connected to this mount, which provided the propelling gas. Nitrogen was used as the gas of choice, and a pressure of 10 psi was typically used. Using a microscope to view the process, the sealed needle tip was carefully broken on a protective end cap placed over the Pyrex capillary, allowing the initiation of the spray of the sample solution. The potential at the inside end of the capillary was maintained at -123.04 V, and the skimmer was set to -1.72 V. The potential difference between the nanospray needle and the outside end of the capillary was set to ∼300 V, allowing for variation between different needles. RESULTS AND DISCUSSION The use of nanospray allowed the acquisition of mass spectra that were free of fragmentation. These spectra thus provided a more accurate portrayal of the sample contents, compared to other techniques, such as electron ionization. Naphthenic acids have frequently been studied using harsh ionization techniques that lead to increased fragmentation and, hence, less certainty about the distinction between molecular ions and fragment ions. A typical negative ion mode mass spectrum of a crude oil is shown in Figure 1, where this particular example is of the oil field A sample. The region m/z 115-1500 was excited and subsequently detected. This represents an example range of interest during investigation of the presence of naphthenic acids,5 though actually wider than typically chosen. Resolution in FTICR mass spectrometry directly relates to the choice of the lowest m/z to be detected and the dataset size chosen for an acquisition, and it is possible to calculate the theoretical maximum resolution achievable under particular conditions.33 The minimum resolution recorded was ∼50 000 fwhm, whereas the maximum resolution recorded was ∼85 000 fwhm. This indicates that even when operating in broadband mode over such a wide range due to practical reasons, FTICR mass spectrometers are suitable for the acquisition of high-resolution mass spectra of complex mixtures. One other crucial parameter to consider with regard to FTICR mass spectrometry, irrespective of the sample used, is the time during which ions may be introduced from an external ion source into the cell, referred to earlier as P2 and sometimes also referred to as the “gating time”. An unfortunate side-effect of the utilization of an external ion source arises, sometimes referred to as the “time-of-flight effect”, and this effect must be taken into consideration when examining relative signal intensities or commenting

Figure 1. A typical broadband mass spectrum of a crude oil, obtained using negative ion nanospray. The mass spectrum shown here was acquired using the oil field A sample.

on the presence or absence of signals within a spectrum encompassing a broad m/z range. Ions must be extracted from the hexapole and trapped in the cell; a discrimination effect can arise, based upon ions’ flight times from the hexapole to the cell. O’Connor et al. have suggested superimposition of spectra acquired using different gating times in order to compensate for this effect46 and, more recently, investigations by Sze and Chan have additionally taken phase coherency into account in order to compensate for gating-time-dependent differences in the phase angle of the ion signals.47 In perspective, discrimination effects are known for all varieties of ionization techniques and mass analyzers within the field of mass spectrometry, and the knowledge of such does not nullify the value of the techniques. As an additional note, rather than a parallel investigation, the average m/z for the oil field A signal distribution was calculated for the results obtained using selected P2 values. Figure 2 shows a plot of the average m/z of the distribution as a function of the square of the P2 value selected (/µs2). A value of 2400 µs was ultimately chosen, because this was in closest agreement with mass spectra obtained using a sector instrument (Micromass AutoSpec), providing a more direct comparison of results from the two instruments. The resulting mass spectra for the two crude oil samples were similar in initial appearance, though the maximums of the signal (46) O’Connor, P. B.; McLafferty, F. W. J. Am. Chem. Soc. 1995, 117, 1282612831. (47) Sze, T.-P. E.; Chan, T.-W. D. Rapid Commun. Mass Spectrom. 1999, 13, 398-406.

Figure 2. Demonstration of the time-of-flight effect phenomenon associated FTICR mass spectrometry when using an external ion source that incorporates a hexapole ion trap. Though not intended to be a separate investigation, the plot illustrates the fact that the average m/z was observed to increase as the gate timing, P2, was increased.

distributions were not located within the same m/z region. Figure 3 shows enlarged mass ranges of mass spectra for the two samples to illustrate this difference, where the maximums for both distributions can be observed within the m/z range shown, and the signal intensity of the oil field A sample was approximately three times larger than that of the oil field B sample. Though it Analytical Chemistry, Vol. 75, No. 4, February 15, 2003

863

Figure 3. Comparison of the broadband mass spectra obtained using samples from oil field A and oil field B. The signals appear at the same m/z values in both spectra, but it can be seen that the intensities differ, indicating that the abundances of the different naphthenic acid families are not the same in both samples.

appears a signal was observed at every integer m/z value, the relative intensities of each differed between the two spectra. The maximum of the distribution acquired when using the oil field A sample was found to be centered between m/z 360 and 370, but the maximum for the oil field B sample was found to be centered near m/z 350. Prior to calibration, the mass accuracies were typically better than 10 ppm. Following external calibration using a standard, mass accuracy was improved to be better than 5 ppm. Comparison with theoretical isotope patterns verified that the species observed were singly charged hydrocarbons. By calculating theoretical masses using Microsoft Excel and also using the MassAnal feature within XMASS for thoroughness, it was found that the assignments in closest agreement with the m/z values were deprotonated ions on the basis of the empirical formula (for the neutral, nondeprotonated species) CnH2n+zO2. Use of theoretical isotope patterns was also made in order to further corroborate the assignments. The high mass accuracy and high resolution of the spectra acquired and the comparison of theoretical isotope patterns with the experimental data were sufficient to conclude that the signals observed did not correlate with nitrogen-containing and sulfur-containing species. The conclusion that all species observed are indeed based upon the empirical formula CnH2n+zO2 is in agreement with the expectation that the signals obtained in the negative ion mode are due to deprotonated naphthenic acids. z is referred to as the “hydrogen deficiency” and is a negative, even number. Following unequivocal assignments of the signals after external calibration, internal calibration was performed; the 864 Analytical Chemistry, Vol. 75, No. 4, February 15, 2003

procedure of external calibration, signal assignment, and subsequent internal calibration was similar to that recently used by Marshall and co-workers.16 The majority of signals for the two crude oil samples typically had an associated mass accuracy of better than 1 ppm (79% of signals obtained using the oil field B sample and 86% of signals when using the oil field A sample). Further enlargement of the mass spectra revealed the presence of “doublets.” Comparatively weak, but genuine, signals could be observed next to their more intense neighbors, typically separated by a difference of 0.1 Th (where the Thomson, Th, is the unit of m/z). It thus becomes clear that two signals, not one, per integer m/z value were present. These weaker signals were most frequently attributed to species with a comparatively high hydrogen deficiency, such as z ) -26. Figure 4 shows one such enlarged mass region that has been enlarged to exemplify the typical observations; the spectrum shown was acquired over a period of 128 scans. The use of a comparatively large number of scans affords enhanced reproducibility, though there will always be some variation due to the replacement of nanospray needles on a regular basis. Unlabeled signals within Figure 4 are due to the presence of heavier isotopes within a given species. The relative peak heights of the 13C1 isotopomers of the weaker species correlated less well with theory (by a few percent) than the stronger species. This is attributed to the smaller statistical sample, whereby small fluctuations in peak height for the weaker signals had a greater relative significance. The inherent high resolution of FTICR mass spectrometry is immediately evident, even when

Figure 4. Enlarged m/z region from a broadband mass spectrum of the oil field A sample. It can be seen that doublets were observed throughout the spectrum. The resolution associated with FTICR mass spectrometry ensures that doublets can easily be resolved and assigned.

operating in broadband mode. It is possible to acquire highresolution mass spectra in broadband mode by lengthening the acquisition time. This can be accomplished by using a larger dataset size (greater number of data points) or increasing the lower m/z limit for the detection to be as high as possible. In addition, for this particular instrument, cell parameters such as the trapping potentials (PV1 and PV2), rf attenuation (PL3), excitation period per pulse (P3), and the SideKick parameters (EV2 and DEV2), must be tuned because they influence resolution and space-charge effects, which can in turn affect mass accuracy. The spectra obtained further illustrate that, even under the relatively adverse conditions in which a large number of different species in the cell can lead to space-charge effects and thus decreased resolution, FTICR mass spectrometry is suitable for the analysis of complex hydrocarbon mixtures in broadband mode. As stated earlier, all species detected in the negative ion mode correlated with deprotonated variants of CnH2n+zO2, the empirical formula of the naphthenic acids. It is not possible to determine whether aromatic rings are present simply from the empirical formula, for instance. As a result, it would be necessary to have access to a variety of naphthenic and naphthenoaromatic acid standards and perform tandem mass spectrometry experiments on both the standards and a given sample in order to verify the structure of the isomer found within the unknown sample. The signals were therefore categorized in terms of carbon content, with different homologues (hydrogen deficiencies) within a given category. Peaks were selected within the XMASS software using the “Peak Pick” facility. The corresponding m/z values of the signals and their correlated absolute intensities were saved using the “writepeaks” command, and the data was exported in ASCII format using the “exportpeaklist” command. The resulting data was imported to Microsoft’s Excel 97, run under Windows NT,

and sorted into corresponding z homologues and carbon contents, comparing the theoretical m/z ratio with the experimental value. The absolute intensities were then normalized so that the plots could be compared in terms of relative intensities. Because of the nature of peaks shapes of signals acquired using FTICR mass spectrometry and the fact that integration to obtain peak areas must be performed following the manual selection of boundary conditions, measuring peak intensities was considered to be less prone to random error. Miyabayashi et al.23 cited noise, baseline bias, and overlapping peak tails as sources of error when considering the areas under peaks and claimed that peak height measurement was therefore more reliable. Hence, peak intensities are referred to throughout, rather than peak areas. In addition, peak intensities are referred to rather than percentage composition of the sample (based upon the TAN), because discrimination effects (during the stages of ion formation, storage in the hexapole, and ion transfer to the cell, among others) will affect the ions’ intensities and prevent a direct correlation. Thus, to accurately translate signal intensity into percentage composition of the sample, it would be necessary to perform experiments using a mixture of equimolar amounts of a wide range of naphthenic acid standards under conditions identical to those used during the analysis of a crude oil. Once the m/z ratios had been sorted in terms of carbon content and hydrogen deficiency, graphs of the relative intensities were plotted as a function of carbon content for each of the z homologues . Using Synergy Software’s Kaleidagraph 3, run under Macintosh System version 8.6, a graph that summarized the nature of the species present and allowed comparison between the two was plotted for each sample. Figure 5 and Figure 6 show the plots for samples from oil fields A and B, respectively. Similar compounds were found to be present in the samples from oil fields Analytical Chemistry, Vol. 75, No. 4, February 15, 2003

865

Figure 5. Plot of the different naphthenic acid families observed in the oil field A mass spectrum.

Figure 6. Plot of the different naphthenic acid families observed in the oil field B mass spectrum.

A and B. As would be expected following the comparison shown in Figure 3, the oil field B signal distribution was centered at lower carbon content value than the oil field A sample. Oil field B signals were most intense at ∼23 carbon atoms, whereas oil field A signals were most intense at ∼24-26 carbon atoms, depending on the z homologue chosen. The minimum and maximum carbon contents for the oil field B sample were 19 and 33, respectively; for comparison, the minimum and maximum carbon contents were 20 and 35, respectively, for the oil field A sample. For the oil field B sample, the most pronounced signals were those associated with the CnH2n-2O2, CnH2n-4O2, and CnH2n-6O2 species, respectively. CnH2n-8O2 and CnH2n-0O2 were similar in intensity, though CnH2n-8O2 was favored over CnH2n-0O2 with increasing carbon content. It should be noted that signals could not be reliably observed for every carbon content with respect to the compounds

866

Analytical Chemistry, Vol. 75, No. 4, February 15, 2003

of greater hydrogen deficiency (due to the decreasing signal intensities at greater hydrogen deficiencies), and thus, gaps appear between some of the data points. The signals then approximately decrease in intensity with increasing hydrogen deficiency, though CnH2n-10O2 becomes favored over CnH2n-8O2 with increasing carbon content. In the case of the oil field A sample, the most intense species were associated with CnH2n-4O2, CnH2n-6O2, and CnH2n-2O2, respectively. Following these z homologues , the signals become less intense with increasing hydrogen deficiency. It is noticeable that the CnH2n-0O2 species varies in intensity and is between the CnH2n-18O2 and CnH2n-24O2 species. This is in stark contrast to the oil field B sample, in which CnH2n-0O2 signals are much more intense. Though the experimental parameters were kept as constant as possible (allowing for small differences between nanospray needles), it is apparent that the naphthenic acid content of the two oils is different. This may in turn be used as an identifying characteristic of a given sample and, hence, act as a “fingerprint” for different oilfields. CONCLUSION FTICR mass spectrometry has been demonstrated to be an effective analytical tool for the analysis of complex hydrocarbon mixtures, such as crude oils. Specifically, coupling with nanospray permitted the acquisition of spectra of minimized complexity, while greatly reducing the contamination risks to the instrument. Though operating under relatively unfavorable conditions (over a wide m/z range in broadband mode, starting with a low m/z limit, and with a large number of different ion packets in the cell), the instrument provided a routine resolution that was much greater than expected for other varieties of mass analyzer. Discrimination due to the time-of-flight effect has been observed, with the signal distribution shifting according to the P2 value chosen. This is an unavoidable consequence of using an external ion source coupled with an FTICR mass spectrometer. Doublets were observed that revealed the presence of naphthenic acids with a high degree of hydrogen deficiency. Such signals would not be observable when using instruments such as quadrupoles and some time-of-flight instruments, which are frequently used for routine analyses. The naphthenic acid content of an oil could be used as a fingerprint that would identify a particular oil sample and link it to a particular oil field. Most importantly, providing information about the composition and range of naphthenic acids present in a crude oil yields useful information that can be implemented in the study of corrosion of refinery equipment. ACKNOWLEDGMENT TotalFinaElf (France) is gratefully acknowledged for funding the research and for granting permission to publish data regarding their samples. Received for review June 12, 2002. Accepted October 11, 2002. AC020388B