Automated electrospray ionization FT-ICR mass spectrometry for petroleum analysis

Sunghwan Kim 0 2 Ryan P. Rodgers 0 1 Greg T. Blakney 0 1 Christopher L. Hendrickson 0 1 Alan G. Marshall 0 1 0 Address reprint requests to Dr. A. G. Marshall, Ion Cyclotron Resonance Program, National High Magnetic Field Laboratory , Florida State Univer- sity, 1800 E. Paul Dirac Drive, Tallahassee, FL 32310-4005, USA 1 Ion Cyclotron Resonance Program, National High Magnetic Field Laboratory, Florida State University , Tallahassee, Florida 2 Korean Basic Science Institute , Chungcheongbuk-Do, Republic of Korea Analysis of petroleum samples at the molecular level by Fourier transform ion cyclotron resonance mass spectrometry (FT-ICR MS) typically requires a prolonged accumulation of ions and/or summing up a large number of scans. Here, a chip-based micro-ESI system (Advion NanoMate, Ithaca, NY) has been successfully automated in combination with FT-ICR MS analysis of petroleum samples. A foil-sealed 96-well glass plate prevents solvent evaporation, with no visible loss of sample after 20 h of continuous operation. Mass spectra obtained from the same sample but taken from different wells after various time delays were very similar. Data from replicate samples in different wells could be combined to enhance mass spectral signal-to-noise ratio and dynamic range. Furthermore, the automated data acquisition eliminates sample carryover, and produces heteroatom class distribution, double-bond equivalents (DBE), and carbon number very similar to those from the conventional (manual) micro-ESI experiments. (J Am Soc Mass Spectrom 2009, 20, 263-268) 2009 Published by Elsevier Inc. on behalf of American Society for Mass Spectrometry - F trometry (FT-ICR MS) offers both high-resolution ourier Transform ion cyclotron resonance mass spec(mass resolving power m/ m50% 400,000, in which m50% is magnitude-mode FT-ICR mass spectral peak full width at half-maximum peak height) and high mass accuracy ( 400 ppb rms error). FT-ICR MS resolving power of 3300,000 for a 1 kDa peptide has been reported [1]. Elemental composition may be assigned uniquely from sufficiently accurate mass measurement for a well-resolved peak [2]. The identified compositions may be further sorted by various tools including hydrogen deficiency (z) based on CcH2c zNnOoSs or double-bond equivalents (DBE number of rings plus double bonds to carbon), Kendrick mass defect versus Kendrick mass [3, 4], or van Krevelen analysis (e.g., H/C ratio versus O/C or N/C or S/C ratio) [5]. Therefore, it is not surprising to find that FT-ICR MS has been successfully applied to study complex natural organic mixtures, including metabolites [6], vegetable oils [7], wine [8], explosives [9], coal extracts [10], and humic materials [11, 12]. An important application of FT-ICR MS is the emerging field of petroleomics [13], namely, the idea that the detailed chemical composition should correlate with (and ultimately predict) the properties and behavior of petroleum and its products. FT-ICR MS-based chemical composition has already enabled characterization of asphaltenes [14, 15], distillates [16, 17], and mined/steam assisted gravity drainage [18] at the molecular level. However, a single petroleum FT-ICR mass spectrum can contain up to 50,000 peaks [19], whereas detection is typically limited to 1,000,000 ions at a time. Thus, the number of ions corresponding to each resolved mass-to-charge ratio is relatively small so that it is often desirable to sum at least 100 time-domain transients to increase signal-to-noise ratio and dynamic range. Data acquisition for a single sample can thus take up to an hour. Fortunately, most petroleum samples present similar m/z range, so that the same experimental conditions apply to a series of related samples, i.e., such analyses are good candidates for automation, for improved reliability and throughput (e.g., overnight). Chip-based automated electrospray MS has been successfully applied to study carbohydrates [20], metabolites [21], peptides/proteins [22], and lipids [23]. Recently, chip-based electrospray ionization (ESI) has been used to measure total acid number (TAN) of petroleum samples by low-resolution mass analysis [24], with improved throughput and reduced sample consumption. Carryover from one sample to the next is eliminated because the automated system supplies a new spray tip and nozzle for each sample. Here, we report a fully automated chip-based micro-electrospray ultrahigh-resolution FT-ICR MS system for automated/ unattended/remote FT-ICR MS data collection for complex petroleum samples. Performance of the system has been tested for reproducibility over extended (20 h) data collection, and means for detecting and halting faulty data acquisitions are described. The system performs as well as fully attended manual data collection. Ten petroleum samples were chosen to test the ability of the automation procedure to provide high quality mass spectra from a wide range of samples: bitumens, heavy vacuum gas oils, and light (343375 C) and heavy distillates (500 525 C). All samples were dissolved to 1 mg/mL in toluene:methanol (50:50 vol/vol) and spiked with 1% (by volume) acetic acid before sample loading. Glass plates with 96 wells were purchased from Zinsser Analytic (Frankfurt, German). The petroleum samples were each placed in three or five glass wells (see below) (200 L per wellmore could be used if required) and sealed with aluminum foil by an in-house designed clamp (see the Results and Discussion section). The samples were introduced sequentially by a NanoMate 100 (Advion Biosciences, Inc., Ithaca, NY). The NanoMate was operated in positive-ion mode with 1500 2000 V spray voltage and 0.5 psi gas pressure. A custom-built FT-ICR mass spectrometer equipped with a 220 mm horizontal room-temperature bore 9.4 T magnet [25] was used in these experiments. Ions generated in the ESI source region are first accumulated in an external linear octopole ion trap, typically for 1 to 5 s, and transferred through rf-only multipoles to a 10 cm diameter, 30 cm long open cylindrical Penning ion trap [26]. Multipoles [25] were typically operated at 1.7 MHz at an rf amplitude of 170 Vp-p. After ions were excited in the trap by broadband frequency-sweep (chirp) dipolar excitation (70 641 kHz at a sweep rate of 150 Hz/ s and amplitude, 440 Vp-p), direct mode image current detection was performed to yield 4 M word timedomain data. Time domain datasets were co-added (100 acquisitions), Hanning apodized, and zero-filled once before fast Fourier transformation and magnitude calculation. Frequency was converted to m/z ratio by the quadrupolar electric trapping potential approximation [27, 28]. The instrument was controlled by an in-house Predator data station [29]. The Predator system features 17 channels of analog voltage ( 10 V, sixteen 16-bit voltages and one 12-bit voltage), 18 TTL triggers, and a 6-channel high power voltage amplifier ( 200 V). Dipolar excitation circuitry is implemented with a commerci (...truncated)