Analysis of photo-irradiated water accommodated fractions of crude oils using tandem TIMS and FT-ICR MS

Abstract

For the first time, trapped ion mobility spectrometry (TIMS) in tandem with Fourier transform ion cyclotron resonance mass spectrometry (FT-ICR MS) is applied to the analysis of the low energy water accommodated fraction (WAF) of a crude oil as a function of the exposure to light. The TIMS-FT-ICR MS analysis provided, in addition to the heteroatom series identification, new insights into the WAF isomeric complexity (e.g., [m/z; chemical formula; collision cross section] datasets) for a better evaluation of the degree of chemical and structural photo-induced transformations. Inspection of the [m/z; chemical formula; collision cross section] datasets shows that the WAF composition changes as a function of the exposure to light in the first 115 hours by initial photo-solubilization of HC components and their photo-oxidation up to O_4–5 of mainly high double bond equivalence species (DBE > 9). The addition of high resolution TIMS (resolving power of 90–220) to ultrahigh resolution FT-ICR MS (resolving power over 400k) permitted the identification of a larger number of molecular components in a single analysis (e.g., over 47k using TIM-MS compared to 12k by MS alone), with instances of over 6-fold increase in the number of molecular features per nominal mass due to the WAF isomeric complexity. This work represents a stepping stone towards a better understanding of the WAF components and highlights the need for better experimental and theoretical approaches to characterize the WAF structural diversity.

TOC image

INTRODUCTION

The complex nature of crude oil and its incorporation in aquatic systems results in complex chemical transformations mainly via bio-^1–8 and photo-degradation.^9–14 Natural and anthropogenic release of crude oil and hydrocarbons^{15, 16} to seawater is a frequent process and recent studies have highlighted the importance of characterizing released crude oil at the molecular level.¹⁷ For example, the characterization of hetero-atom (O, N, and S) poly aromatic hydrocarbons (h-PAHs) have exhibited increased levels of toxicity compared to pure hydrocarbons.¹⁸ Moreover, PAHs are photo active, undergoing oxygenation upon exposure to light, which can lead to chemical products that have increased biological accumulation and activity.^19–25 The presence of crude oil in water provides the means for the exposure of a large number of molecules to chemical and enzymatic transformation, their transportation across environments,^26–28 as well as the interaction with a variety of organisms.²⁶ Many of these chemical changes, as well as the means of exposure to organisms, occur in the water accommodated fraction, where low energy mixing introduces components of the oil into the water, without the formation of detectable emulsions.²⁹ The main analytical challenge during the analysis of the low energy water accommodated fraction (WAF) remains the identification and quantitation of both the primary molecular species, as well as the transformation intermediates and products.

Over the past years, most of the efforts for the WAF analysis has been focused on the use of gas chromatography – mass spectrometry (GC-MS)^{7, 9, 10, 13, 30–36} and two dimensional gas chromatography (GCXGC-MS) with heavy standards.^{31, 37} While progress has been made in the WAF characterization, these approaches are limited to the volatility range of molecules that can be analyzed by GC, which typically excludes large and highly polar molecules.³⁸ These analytical limitations narrow the type of studies that can be performed and our understanding of the crude oil transformations in seawater; especially, since the molecular species that are inaccessible or form unresolvable ‘humps’, known as the unresolved complex mixture (UCM), can make up most of the WAF content.^{39, 40} The analytical challenges associated with the molecular characterization of the UCM has led to the use of alternative tools in order to unravel its chemical complexity. For example, studies utilizing ultra-high mass resolution mass spectrometry (e.g., FT-ICR MS)^41–44 have enabled the identification of chemical formulas using the isotopic resolution and the high mass accuracy (sub ppm) with a variety of atmospheric pressure ionization sources (e.g., electrospray ionization, ESI,^{45, 46} atmospheric pressure chemical ionization, APCI,^47–49 atmospheric pressure photo-ionization, APPI,^50–53 and atmospheric pressure laser ionization, APLI^54–61). The use of a variety of atmospheric pressure ionization sources has enabled, in turn, the study of different molecular fractions at the molecular level and has provided evidence of the high structural diversity of the WAF components in their functional groups, aromaticity, and polarity.^62–64

The high structural diversity of the WAF samples has prompted the need to complement ultrahigh resolution mass analysis (e.g., FT-ICR MS measurements) with pre-separation techniques (e.g., liquid and gas chromatography), in order to better discriminate the components along a second axis of separation, permitting some isomeric separation and increasing the dynamic range of the FT-ICR MS measurement.^{59, 60, 65–69} However, the biggest challenge in the coupling of liquid and gas chromatography is that it limits the FT-ICR MS analysis time, and thus the ability to better separate isobaric species, due to the slow acquisition rates needed for ultrahigh resolution mass acquisitions.^{48, 65} Alternatively, other post-ionization, gas-phase separations have been proven to be a better match for FT-ICR MS analysis.^70–79 In particular, ion mobility spectrometry (IMS) presents many advantages for the analysis of complex mixtures, providing orthogonal separation to FT-ICR MS that is based on the tri-dimensional structure of the molecule.^80–82 Initial work showed the potential of IMS-MS analysis for the characterization of complex hydrocarbon mixtures using complementary IMS-MS and FT-ICR MS measurements (e.g., IMS-TOF MS and FT-ICR MS of the same sample).^83–92 More recently, with the development of trapped ion mobility spectrometry (TIMS),^93–95 several reports have shown the potential of TIMS-MS to decouple the mobility (K) separation from the MS analysis time for fast, gas-phase separation and for molecular structural elucidation.^{51, 93, 96–117} In particular, the advantages of TIMS over traditional IMS analyzers has been shown for fast screening⁹⁶ and targeted⁷⁹ analysis of molecular ions from complex chemical mixtures; the study of isomerization kinetics of small molecules,⁹⁸ peptides, DNA, proteins and their complexes in the absence of the bulk solvent;^99–103 the influence of the collision partner on the molecular structure;¹⁰⁴ and the factors that affect molecular-adduct complex lifetime and stability during TIMS measurements.¹⁰⁵ A significant feature to note is the high resolving power of TIMS analyzers (R_TIMS up to 400¹⁰⁷) and accuracy in measuring ion-neutral collision cross section (CCS, <0.6% error). In the case of crude oils and complex mixtures, their characterization by TIMS-FT-ICR MS has allowed the measurement of the CCS, accurate mass and accurate isotopic fine structure in a single experiment for a series of h-PAHs. For example, a recent report of Oversampling Selective Accumulation Trapped Ion Mobility Spectrometry (OSA-TIMS) coupled to FT-ICR MS showed high mobility resolving powers (over 250), high mass accuracy (<1 ppm), and ultrahigh mass resolution (over 1,200,000 at m/z 400) during the analysis of a complex mixture of polyaromatic hydrocarbons (PAH) from coal tar.¹⁰⁶

In the present work, for the first time, we apply tandem OSA-TIMS and FT-ICR MS for the analysis of WAF samples as a function to their exposure to light. While preliminary work has shown that WAF can undergo chemical and increase the solubilization as a function to the exposure to light,⁴³ little is known about the WAF structural variability and transformation pathways. In this study, as an initial step, the analysis focuses on the PAH compounds accessible to ionization by an APLI source (e.g., mostly conjugated molecules) which typically exhibit higher reactivity to light resulting in more hazardous byproducts.¹¹⁸ In addition to the new analytical advantages of TIMS-FT-ICR MS, a Software Assisted Molecular Elucidation (SAME) package was developed for the unsupervised processing of the OSA-TIMS-FT-ICR MS datasets. As shown below, this work highlights the need for high mobility resolution and ultra-high resolution MS for the analysis of the highly isomeric, complex WAF mixtures while providing [m/z; chemical formula; K; CCS] in a single experiment.

EXPERIMENTAL

Sample preparation

Low-energy water accommodated fractions (WAF) were generated according to the standardized protocol established by the Chemical Response to Oil Spills: Ecological Research Forum (CROSERF).^{119, 120} Briefly, WAFs were prepared at room temperature in 2-L aspirator bottles with 20% headspace by volume using artificial filtered saltwater (pore size: 0.45μm, salinity=33 parts-per-thousand) prepared with Instant Ocean^® (Aquarium Systems, Mentor OH). Oil from the Marlin Platform was added to the water surface using a gastight syringe at an oil-to-water loading of 1:1000 (1 g oil/L seawater). The bottles were covered in aluminum foil and allowed to mix for 24h at low speed (100 rpm) in the dark.

WAF exposure to light and extraction

WAF samples were irradiated up to 115h with a Suntest XLS+ Sunlight Exposure System equipped with a 1500W xenon arc lamp and light intensity of 765 W/m² (Atlas, Chicago, IL, USA). The aspirator bottles containing the WAF and the oil were placed in a water bath system set to 25℃. At specific irradiation times (0, 16, and 115h), 150 ml aliquots of the WAF were removed and subjected to liquid-liquid extraction in three 50 mL lots of methylene chloride to increase extraction efficiency. Aliquots were dried over Na₂SO₄ and concentrated down to 1 mL under a stream of nitrogen. The final samples were then diluted 1:100 in 1:1 v/v methanol/toluene for FT-ICR MS and TIMS-FT-ICR MS analysis.

FT-ICR MS analysis

WAF samples were analyzed in positive ion mode with an APLI source coupled to a custom-built TIMS – FT-ICR MS instrument based on the 7T Solarix FT-ICR MS spectrometer (Bruker Daltonics Inc., MA). Briefly, the APLI source utilizes a 266 nm excimer laser (CryLas GmbH, Berlin, Germany; Type:1HP266-50); the sample was introduced at 200 μL/h through a short nebulizer in a vaporizer set to 300 °C into the source chamber where the gas stream was ionized by the excimer laser.⁶⁵ The APLI generates radical ([M]^+·) and protonated ([M+H]⁺) ions in the source region that are introduced to the TIMS-FT-ICR MS via a 0.6 mm inner diameter, single-bore resistive glass capillary tube, allowing the nebulizer to be maintained at ground potential, while the ends of the capillary can be independently biased. Typical APLI operating conditions were 1 L/min dry gas flow rate, 2.1 bar nebulizer gas pressure, and 220 °C dry gas temperature. FT-ICR MS ion optics were optimized as follows: −900 V endcap source capillary voltage, 180 V endcap TIMS capillary voltage, 5kHz 400 peak-to-peak voltage (Vpp) segmented hexapole, 2kHz 1900 Vpp collision cell, and 4kHz 400 Vpp ion guide transfer line. The FT-ICR MS experiments were performed by co-adding 200 16 Megaword (8 second) transients, which were zero-filled to 32Megaword, processed using a half-sine apodization followed by fast-Fourier transform (FFT) and broadband phase correction (absorption spectra using absorption mode processing, AMP); an experimental MS resolving power with AMP at m/z 400 of ~2,000,000 was obtained.

TIMS-FT-ICR MS Analysis

The concept behind TIMS is the use of an electric field to hold ions stationary against a moving gas, so that the drag force is compensated by the electric field and ion packages are separated across the TIMS analyzer axis based on their mobility.^93–95 During mobility separation, a quadrupolar field confines the ions in the radial direction to increase trapping efficiency. A simplified schematic of a TIMS analyzer is shown in the Supporting Information (Figure S1). The mobility, K, of an ion in a TIMS cell is described by:

$$K=\frac{v_g}{E}=\frac{A}{\left(V_{\mathit{elution}}-V_{\mathit{out}}\right)}$$ (1)

where v_g, E, V_elution and V_out are the velocity of the gas, applied electric field, elution voltage and tunnel out voltage, respectively. TIMS separation was performed using nitrogen as a bath gas at ca. 300 K, front end P₁ = 3.0 and back end P₂ = 1.1 mbar pressures, a constant V_out = 35 V and constant RF (840 kHz and 240 Vpp) in all electrodes of the entrance funnel, mobility separating section and exit funnel. Details regarding Oversampling Selected Accumulation TIMS (OSA-TIMS) modes of operation¹⁰⁶ and specifics compared to traditional TIMS and IMS can be found elsewhere.^{93–96, 98} Briefly, OSA-TIMS experiments were performed by scanning V_in from −40 to −210 V using a 1 V ramp size and 0.2 V increments per step, accumulating 40 mobility experiments per FT-ICR MS spectrum (4Megaword, 3s transient, with six transients co-added per MS). TIMS-FT-ICR MS spectra were processed using sine-squared apodization followed by FFT, in magnitude mode resulting in an experimental MS resolving power of R ~ 400,000 at m/z 400. Mobility spectra were calibrated using a Tuning Mix calibration standard (Tunemix, G2421A, Agilent Technologies, Santa Clara, CA) with the following reduced mobility (K_o) values m/z 622 K₀=1.025, m/z 922 K₀=0.840, m/z 1222 K₀=0.724¸ m/z 1522 K₀=0.643 cm² V⁻¹s⁻¹. Mobilities were correlated with CCS (Ω) using the equation:

$$\mathrm{\Omega }=\frac{{\left(18\pi \right)}^{1/2}}{16}\frac{z}{{\left(k_{B}T\right)}^{1/2}}{\left[\frac{1}{m_I}+\frac{1}{m_b}\right]}^{1/2}\frac{1}{K}\frac{760}{P}\frac{T}{273.15}\frac{1}{N^{\ast }}$$ (2)

where z is the charge of the ion, k_B is the Boltzmann constant, N^* is the number density and m_I and m_b refer to the masses of the ion and bath gas, respectively.¹²¹ Under these conditions, the experimental TIMS resolving power for Tuning Mix (m/z 622–1522) was ~100–250 as determined by equation 3.

$$R=\frac{\mathrm{\Omega }}{\Delta \mathrm{\Omega }}$$ (3)

Data Processing

FT-ICR MS spectra were externally mass calibrated using the Tuning Mix standard. A peak list was generated allowing a signal to noise ratio of 6 and the data were internally, recalibrated (post-acquisition) using a double bond equivalence of 9 O₁ series to improve overall mass accuracy.¹²² The formulae calculations from the exact mass domain were performed using Composer software (Version 1.0.6, Sierra Analytics, CA) and confirmed with Data Analysis (Bruker Daltonics v 4.2) using formula limits of C_1–100H_1–100N_0–2O_0–7S_0–2, odd and even electron configurations were allowed, and M^+· and [M+H]⁺ ion forms. A root mean square deviation for the mass assignments of 0.3 ppm was observed. From the generated ion formulas, the double bond equivalence (DBE) was calculated by the equation:

$$\mathit{DBE}=C-\frac{H}{2}+\frac{N}{2}+1$$ (4)

where C, H and N are the number of carbons, hydrogens, and nitrogen in the chemical formula.

The peak list was used for extraction of the ion mobility spectra from the TIMS-FT-ICR MS datasets using batch processing in the Data Analysis package (Version v. 4.2, Bruker Daltonics, CA) followed by external mobility calibration using the Tuning Mix standards. The TIMS spectrum for each molecular formula was processed using a custom-built Software Assisted Molecular Elucidation (SAME) package – a specifically designed 2D TIMS-FT-ICR MS data processing script written in Python v2.7. SAME package utilizes noise removal, mean gap filling, “asymmetric least squares smoothing” base line correction,⁷⁰ peak detection by continuous wavelet transform (CWT)-based peak detection algorithm (SciPy package),^{123, 124} and Gaussian fitting with non-linear least squares functions (Levenberg-Marquardt algorithm).¹²⁵ SAME final outcome is [m/z; chemical formula; K; CCS] for each 2D TIMS-FT-ICR MS dataset. The 2D TIMS-FT-ICR MS contour plots were generated in DataAnalysis (Version v. 4.2, Bruker Daltonics, CA) and all the other plots were generated using matplotlib¹²⁶ and OriginPro 2016 (Originlab Co., MA).

RESULTS AND DISCUSSION

The analysis of the WAF samples by APLI-FT-ICR MS can be characterized by a Gaussian-like distribution, centered at m/z 300 (Figure 1). Prior to exposure to light (t-0h), ~700 peaks were observed in the FT-ICR MS spectra. After the WAF was exposed to light (t-115h), the distribution increased in size, and the center shifted to m/z 500, resulting in ~ 12,000 peaks, which represents a ~17-fold increase relative to the unexposed WAF (t-0h). The change in the MS distribution suggest an increased partitioning of the oil in the WAF (e.g., photo-solubilization) as well as potential chemical transformations within the WAF as a function of the exposure to light. The use of ultra-high resolution mass analyzers allowed the assignment of chemical formulas and to follow the WAF changes as a function of the exposure to light (see Figure S2–4). A follow up analysis using OSA-TIMS in tandem with FT-ICR MS enabled further molecular separation of the WAF content by their mobility (isomeric content) followed by ultra-high resolution mass analysis. For example, the number of molecular features increased from ~700 to ~5.2 k and from ~12k to ~47k for the t-0 and t-115h WAF samples, respectively, by adding the TIMS separation to the FT-ICR MS analysis. Moreover, the isomeric content of the WAF samples is such that TIMS separation increases the number of features in up to 6-fold at the nominal mass level (see Figure 1). That is, OSA-TIMS in tandem with FT-ICR MS enabled a more comprehensive analysis of the WAF content by increasing the peak capacity of the analysis using complementary, orthogonal TIMS and FT-ICR MS separations.

Figure 1.

Typical FT-ICR MS spectrum and 2D TIMS-FT-ICR MS contour plot for the WAF (a) at 0, and (b) 115h light exposed. The number of peaks identified per unit m/z in the MS domain (black) and TIMS-MS (blue) domains are also shown for (a) and (b). Notice the significant differences in the number of identifications between (a) and (b), as well as between the MS and TIMS MS experiments, increasing the level of molecular feautres identified per analysis.

Inspection of the 2D-TIMS-FT-ICR MS contour plots indicated that the observed chemical species from the WAF samples using the APLI source are mostly condensed/aromatic molecules (see more details on interpreting 2D-IMS-MS contour plots in reference¹⁰⁶). This observation is consistent with previous analysis using APLI sources that showed better ionization efficiencies for molecules containing conjugated bonds.⁶⁵ Closer inspection of the 2D TIMS-FT-ICR MS WAF data reveals the spectral complexity in both the mass and mobility dimensions. Particularly, the presence of multiple isobaric interferences (e.g., -C₃ to -SH₄ splits, Δm=3.4 mDa, requiring ~150,000 resolving power at m/z 489), such as C₃₁H₃₇O, C₂₈H₄₁OS, and C₂₅H₄₅OS_2, as well as multi-band ion mobility projections confirm the need for high resolution TIMS analysis during the study of the WAF samples (TIMS resolving power up to 220 is shown in Figure 2). For example, at m/z 327, 14 MS peaks are detected and chemical formulas are assigned (Table 1). For each chemical formula, an extracted ion mobility chromatogram was generated, resulting on 47 peaks with assigned chemical formula, mobility and CCS. Notice that this detailed separation is only possible due to the high resolving power of the TIMS device, the OSA-TIMS method providing sufficient points across the mobility profile, and the ultra-high resolution and mass accuracy of the FT-ICR MS. Moreover, the processing of the 2D-TIMS-FT-ICR MS data using the SAME package is able to deconvolute the mobility profile in a minimum number of isomers (see example for C₂₂H₃₁O₂⁺ and C₂₁H₂₇O₃⁺ in Figure 2). A similar complexity can be observed at m/z 489, where 17 peaks are resolved in the FT-ICR MS and 123 peaks in the 2D TIMS-FT-ICR MS. The time independent nature of the OSA-TIMS analysis permits the acquisition of high mass resolution FT-ICR MS spectra, thus maximizing the analytical potential of both techniques while providing precise collision cross section (less than 1% variability between replicates). It should be pointed out that these results provide a new reference point for the IMS-MS analysis; commonly accessible IMS-MS platforms are limited to IMS resolving power of ~30–60 (with instances up to 100) and to TOF MS detectors with MS resolving power up to 60k.^83–91

Figure 2.

Typical 2D TIMS-FT-ICR MS contour plot and MS projections for m/z a) 327 and b) 489. IMS projections for specific chemical formulas (connected scatterplot) with the unsupervised fitting by SAME package of minimum number of mobility features is shown. Note that the SAME package relies on the experimental profile of the distribution, which is able to show multiple features due to the high resolving power of the TIMS analyzer. Formula assignments are provided in table 1.

Table 1.

Table of identified ion formulas for m/z 327 and 489 as found in figure 2.

m/z		Exp. m/z	Ion Formula	Error (ppb)
327	a	327.078031	C15H21NOS3+	−159
	b	327.101557	C22H15O3+	43
	c	327.104917	C19H19O3S+	76
	d	327.137957	C23H19O2	−3
	e	327.141345	C20H23O2S+	−55
	f	327.159052	C20H23O4+	104
	g	327.174325	C24H23O+	52
	h	327.177623	C21H27OS+	275
	i	327.195425	C21H27O3+	141
	j	327.210673	C25H27+	165
	k	327.214774	C22H31S+	−2066
	l	327.231822	C22H31O2+	107
	m	327.235269	C19H35O2S	−125
	n	327.268202	C23H35O+	122
489	a	489.169661	C32H25O5+	−22
	b	489.172941	C29H29O5S+	164
	c	489.190592	C29H29O7	384
	d	489.205937	C33H29O4+	202
	e	489.209127	C30H33O4S+	572
	f	489.227101	C30H33O6+	131
	g	489.242320	C34H33O3+	206
	h	489.245496	C31H37O3S+	605
	i	489.263445	C31H37O5+	217
	j	489.266855	C28H41O5S+	137
	k	489.270311	C25H45O5S2+	−37
	l	489.278685	C35H37O2+	249
	m	489.282768	C32H41O2S+	−1206
	n	489.299809	C32H41O4+	260
	o	489.302875	C29H45O4S+	883
	p	489.336155	C33H45O3+	341
	q	489.339652	C30H49O3S+	84

Taking advantage of the high mobility resolution and ultra-high mass separation of TIMS-FT-ICR MS, the WAF [m/z; chemical formula; K; CCS] components can be followed as a function of the exposure to light for each chemical class (see example in Figure 3). The increase in the number of molecular species from the HC class as a function of the exposure to light suggest that WAF is initially subject to photo-solubilization of the surface slick into the water, in good agreement with previous observations.⁴¹ Results show that an increase in the DBE of a molecule increases is associated with a reduction in the CCS for a given carbon number. For example, for C₃₄ the lowest CCS of 210 Ų corresponds to DBE 14, while DBE 5 has the greatest CCS of 252 Ų. This indicates that the degree of condensation (e.g., rings and double bonds) imposes structural boundaries, reducing the CCS of a molecule.^{85, 127} In addition to the initial photo-solubilization, as the exposure of the WAF to light increases, a greater number of assignments with carbon numbers greater than 40 are observed, particularly comparing 16 and 115h. Furthermore, there is an increase in the number of oxygenated classes and a decrease in the HC class, particularly for compounds with DBE>9. In particular, the identified formulas for the O₄ and O₅ classes have DBE ranges between 8–20 and 9–18, respectively. These identified formulas occupy a narrow structural space in the condensed region of the mobility domain, which may indicate that these are products of the higher photosensitive and reactive aromatic HC structures.¹²⁸ Although the WAF transformation mechanisms are not well understood, our results suggest that the photo-transformation of the HC molecules in WAF leads to the observation of oxygenated species of the O_4-5 class in the first 115h of exposure to light.

Figure 3.

Typical size dependence (CCS) with carbon number for the O_0-4 chemical classes observed in the WAF samples as a function to exposure to light (t-0, 16 and 115h). The color scale corresponds to the number of rings and double bond equivalents (DBE) of a molecule. Note the large increase in assignments between t-0h and t-16h, as well as increases in oxygen content between t-16h and t-115h.

Changes in the WAF composition can be also followed by the presence of specific chemical formulas as a function of the exposure to light. For example, inspection of the WAF [m/z; chemical formula; K; CCS] components at different time points can be used to infer the degree of chemical transformations (see Figure 4). For example, the unexposed WAF, t-0h, has few identifications for the HC class (black bars); however, at t-16h (red bars) there is a significant increase in the number of assignments, 420 new identifications based on MS alone and 3000 based on TIMS-MS followed by a decrease in the number of non-oxygenated molecules (e.g., HC and N class molecules) at t-115h.⁴³ This result suggest that there are several chemical transformations occurring as a consequence of the photo-solubilization of the surface oil (e.g., indicated by the new identifications for the HC class at t-115h) and increased oxygenation of previously dissolved WAF components (e.g., the three-fold increase in the number of identifications for the O₃ class, and the appearance of the O₄ and O₅ class with 1140 and 420 identifications, respectively). The oxygenated molecules are also highly isomeric, with up to 9 ion mobility bands per chemical formula. The observation of the O_4-5 classes at later irradiation times (t-115h) suggest that these molecules were not originally present in the WAF, but are a consequence of the oxygenation process that took place over time, either by the generation of new molecules, or by a decrease in matrix effects due to a lower number of UV absorbent molecules. That is, a reduction of molecules that are highly absorbent of the 266 nm excimer laser may result in greater sensitivity for less absorbent and low concentration molecules. Note that the increase in oxygenation is in good agreement with other MS reports of WAF exposure to light.^{129, 130} Small differences were observed between the number of identifications, such as the HC class, by TIMS-MS and MS alone due to the reduced TIMS trapping efficiency for low m/z ions and low abundant ions when performing a broad range mobility analysis; however, this limitation can be overcome by performing targeted analysis for smaller PAHs (e.g., naphthalene) as previously reported.⁹⁶

Figure 4.

Total number of molecular feature assignments based on chemical formula alone from FT-ICR MS measurements and based on chemical formula and IMS profiling from TIMS-FT-ICR MS measurements . Identifications unique to t-0, t-16, and t-115h are in black, red, and blue, respectively. When using FT-ICR MS is used tracing the evolution of the chemical complexity is incomplete, because the isomeric complexity is not taken into account. Particularly, taking into account the isomeric variability indicates that the composition of the WAF at t-115h is significantly more complex, and chemically unique, than is observed by FT-ICR MS alone.

While this work showcases the unique advantages of OSA-TIMS in tandem with FT-ICR MS and represents a major step towards the analytical characterization (i.e., high mobility resolving power over 220 combined with ultrahigh mass resolution over 400k) of the WAF samples at the molecular level in a single experiment, further experiments and developments are needed. For example, a more complete characterization of the WAF content and their transformation products and intermediates will require the use of a suite of ionization sources, in both positive and negative polarities, in order to cover a larger range of chemical species during the analysis.^{62, 131} Further interpretation of the WAF [m/z; chemical formula; K; CCS] components can be made with the use of standards, theoretical calculations of candidate structures,^{79, 96, 97} and the implementation of complementary, post-ionization MS structural tools in tandem with TIMS-FT-ICR MS (e.g., MS/MS using CID, SID, IRMPD, BIRD, ExD, etc.)^132–138 Although initial attempts have been made towards the structural characterization,⁹⁷ the 2D TIMS-FT-ICR MS datasets contain a large amount of analytical data (e.g., over 50k features) in need for further development of theoretical ‘petro-informatics’ and computational approaches capable of producing more detailed structural information of the WAF photo-transformation products and intermediates.