High Field Electron Paramagnetic Resonance Characterization of

Feb 3, 2015 - High Field Electron Paramagnetic Resonance Characterization of Electronic and Structural Environments for Paramagnetic Metal Ions and Or...
1 downloads 7 Views 481KB Size
Download PDF
Article pubs.acs.org/ac

High Field Electron Paramagnetic Resonance Characterization of Electronic and Structural Environments for Paramagnetic Metal Ions and Organic Free Radicals in Deepwater Horizon Oil Spill Tar Balls Vasanth Ramachandran,† Johan van Tol,§ Amy M. McKenna,§ Ryan P. Rodgers,†,§ Alan G. Marshall,*,†,§ and Naresh S. Dalal*,†,§ †

Department of Chemistry and Biochemistry, Florida State University, 95 Chieftain Way, Tallahassee, Florida 32306, United States National High Magnetic Field Laboratory, 1800 East Paul Dirac Drive, Tallahassee, Florida 32310, United States

§

S Supporting Information *

ABSTRACT: In the first use of high-field electron paramagnetic resonance (EPR) spectroscopy to characterize paramagnetic metal−organic and free radical species from tar balls and weathered crude oil samples from the Gulf of Mexico (collected after the Deepwater Horizon oil spill) and an asphalt volcano sample collected off the coast of Santa Barbara, CA, we are able to identify for the first time the various paramagnetic species present in the native state of these samples and understand their molecular structures and bonding. The two tar ball and one asphalt volcano samples contain three distinct paramagnetic species: (i) an organic free radical, (ii) a [VO]2+ containing porphyrin, and (iii) a Mn2+ containing complex. The organic free radical was found to have a discshaped or flat structure, based on its axially symmetric spectrum. The characteristic spectral features of the vanadyl species closely resemble those of pure vanadyl porphyrin; hence, its nuclear framework around the vanadyl ion must be similar to that of vanadyl octaethyl porphyrin (VOOEP). The Mn2+ ion, essentially undetected by low-field EPR, yields a high-field EPR spectrum with well-resolved hyperfine features devoid of zero-field splitting, characteristic of tetrahedral or octahedral Mn−O bonding. Although the lower-field EPR signals from the organic free radicals in fossil fuel samples have been investigated over the last 5 decades, the observed signal was featureless. In contrast, high-field EPR (up to 240 GHz) reveals that the species is a disc-shaped hydrocarbon molecule in which the unpaired electron is extensively delocalized. We envisage that the measured g-value components will serve as a sensitive basis for electronic structure calculations. High-field electron nuclear double resonance experiments should provide an accurate picture of the spin density distribution for both the vanadyl-porphyrin and Mn2+ complexes, as well as the organic free radical, and will be the focus of follow-up studies.

T

knowledge of their concentrations and chemical identity is thus of vital importance. FTICR MS characterization of the weathered spilled oil (“tar balls”) reveals significant amounts of metalloporphyrins, in particular, Ni2+ and [V−O]2+ (vanadyl)containing porphyrins. Because [V−O]2+ contains a 3d1 unpaired electron, it is paramagnetic and hence accessible by electron paramagnetic resonance (EPR) spectroscopy. Indeed, a recent EPR study by Kiruri et al.4 of tar balls collected from the Gulf region followed several earlier studies on crude oil and petroleum asphaltenes.5−8 Biktagrov et al. published an EPR study of rotational mobility of vanadyl porphyrin complexes in crude oil asphaltenes.9 Their experiments, like those of ref 4, were conducted at X-band with the assumption of isotropic rotational diffusion. Trukhan et al. reported X-band EPR of

he 2010 Deepwater Horizon (DH) oil spill resulted in a net release of 3.2 million barrels of crude oil into the Gulf of Mexico.1 Extended environmental exposure resulted in depletion of saturated hydrocarbons by evaporation and biodegradation, as well as formation of oxygenated derivatives of the original components, based on comprehensive twodimensional gas chromatography GC × GC analysis.2 However, only ∼40% by weight of the spilled oil is sufficiently volatile to be GC-accessible. More than 30 000 elemental compositions of the larger and/or more polar organic constituents of the Macondo wellhead crude oil have recently been determined by ultrahigh-resolution positive and negative electrospray ionization 9.4 T Fourier transform ion cyclotron resonance mass spectrometry (FTICR MS).3 These materials are exposed to the marine and ambient environment weather over time and may become potential environmental and ecological hazards due to the presence of environmentally toxic species such as catalytically active metal complexes and reactive organic free radicals; a detailed © XXXX American Chemical Society

Received: November 2, 2014 Accepted: January 19, 2015

A

DOI: 10.1021/ac504080g Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry Table 1. Sample Sources and Physical Characteristics serial no.

sample name

source

1 2

Brick Burn Residue

3

Il Duomo

Elmers Island Gulf of Mexico Gulf of Mexico Provided by LSU Pacific Ocean

4 5

NIST RM2779 NIST RM8505

6

VOOEP

description

black, thick, shiny tar black, thick, tar black, dry, hard, coal-like rocks

NIST NIST

weathered asphalt volcano sample from natural petroleum seep Gulf of Mexico Crude oil; std. research material vanadium in crude oil; std. research material

Sigma-Aldrich

vanadyl octaethylporphyrin

asphaltenes at elevated temperature and pressure.10 Again, however, their analysis was limited by the low magnetic field. Although those EPR studies succeeded in identifying [V−O]2+ and its chemical bonding from asphaltene extracts, the presence of an overwhelmingly strong EPR signal from ubiquitous organic free radicals made it difficult to fully characterize the [V−O]2+ species, as well as other magnetic ions such as Mn2+, due to the limited spectral resolution of the (commercially available) instrumentation based on low microwave frequency: X-band (9.5 GHz, 0.3 T) or Q-band (35 GHz, 1.5 T). Here, we perform EPR at much higher microwave frequency (up to 240 GHz) and magnetic field (up to 10 T or higher). The resulting spectra exhibit much higher resolution, enabling complete separation of the EPR peaks from the [V−O]2+ ion from those of the organic free radical and Mn2+ ions for the first time. We also provide quantitative measurements of the vanadyl and Mn concentrations in the tar balls and marine asphalt volcano sample, based on NIST (National Institute of Standards and Technology) standards, and show that the organic radical moiety in the crude and the tar balls is not a polymerized, knotted spherical structure but rather an axially symmetric disk-shaped organic compound with the unpaired electron delocalized over the molecular frame. The introduction of high-field EPR to this field opens up a new approach for characterizing paramagnetic metal complexes and organic moieties in source and weathered petroleum crude oils, as well as other fossil fuel-related environmental5−8 samples.

■

physical characteristics

weathered oil, tar brick; not from DH oil spill tar from controlled burn related to DH oil spill

brown, oily liquid dark brown, shiny, highly viscous fluid purple powder

Sigma-Aldrich13 were also used to compare the tar ball contents. Metal concentrations were determined by inductively coupled plasma mass spectrometry (ICPMS) as described elsewhere.14 EPR Measurements. EPR measurements were made at several different fields by use of both a commercially available (Bruker E-500) spectrometer, working at fixed frequency and scanning the magnetic field, X-band (9.5 GHz, 0−0.5 T) and Q-band (34.5 GHz, 0−1.5 T), and a locally developed high frequency, high-field (240 GHz, 12.5 T) spectrometer available at the National High Magnetic Field Laboratory.15,16 Unless otherwise specified, all spectra were measured at room temperature. Owing to the limitations of the sample holder and probe setup, the standard “Gulf of Mexico Crude Oil” (NIST-RM2779) sample was frozen prior to installation inside the sample chamber maintained at 20 K and measured at 20 K. The magnetic field was calibrated with an NMR gauss meter as well as the standard organic free radical diphenylpicrylhydrazyl (DPPH)17 (g = 2.0037) (in situ). The frequency was monitored by use of built-in or Hewlett-Packard digital frequency counters. Spectral simulation was carried out with locally developed computer programs.18,19

■

RESULTS AND DISCUSSION Characterization by EPR Spectroscopy. Because the major goal of this study is the characterization of magnetic ions and organic free radicals, it is worthwhile to succinctly summarize the underlying principles. A paramagnetic system with electron spin S and nuclear spin I can be described by the following spin Hamiltonian.20−22

EXPERIMENTAL METHODS

Tar Ball and Standard Samples. Tar balls were collected from two locations in the Gulf of Mexico. The samples are labeled as Brick and Burn Residue. A marine asphalt volcano sample (extensively weathered) was collected off the coast of Santa Barbara, CA and is labeled Il Duomo. The collection location and other relevant information are listed in Table 1. Commercial grade caster oil was obtained from House of Spices, 127-40 Willets Pt Blvd, Flushing, NY 11368. The samples were analyzed for their content of paramagnetic metals and organic free radicals. Vanadyl ([V−O]2+) and manganese (Mn2+) ion oxidation states and chemical structures were determined by both commercially available (low-field) and locally developed high-field EPR spectrometry. [V−O]2+ concentrations were determined by use of vanadium standards obtained from the National Institute of Standards and Technology (NIST), NIST Standard Research Material, Vanadium in Crude Oil,11 NIST-RM8505, by comparing the EPR peak areas for the sample and the standard under essentially identical conditions. Another standard, NIST Gulf of Mexico crude oil,12 NIST-RM2779, obtained from NIST, and pure vanadyl octaethyl porphyrin (VOOEP), obtained from

⎧ S(S + 1) ⎫ ⎬ + E[Sx 2 − Sy 2] / = βH ·g ·S + S ·A ·I + D⎨Sz 2 − ⎩ ⎭ 3 (1)

The first two terms represent the electron Zeeman interaction and the electron−nuclear hyperfine interaction, and the last two are the zero-field-splitting terms, in which β is the Bohr magneton, g is the Lande g-tensor, A is the hyperfine coupling tensor, D and E are the zero-field splitting parameters, and Sx, Sy, and Sz are the components of the electron spin S, for the Zeeman field H applied in the z direction. Usually D and E derive from dipole−dipole coupling between unpaired electrons on the same metal. Hence, D and E are zero for [V−O]2+ (which contains only a single unpaired electron).20−23 In the special case of an axially symmetric system, gx = gy = g⊥, called g-perpendicular, and gz = g∥, called g-parallel, and similarly Ax = Ay = A⊥, Az = A∥. Vanadyl Octaethylporphyrin Model Compound. To illustrate the above principles and to compare low-field and high-field EPR for [V−O]2+, Figure 1 shows EPR spectra of a B

DOI: 10.1021/ac504080g Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

Figure 1. X-band (9.4 GHz, left) and high-field (240 GHz, right) EPR spectra of the model compound, polycrystalline vanadyl octaethylporphyrin (VOOEP). Simulated spectra (blue) and the stick diagrams are guides to the eye to identify various components of the anisotropic [V−O]2+ spectra. Inset: vanadyl porphyrin molecule with its VO bond aligned parallel and perpendicular to the direction of the magnetic field (H).

Table 2. Characteristic Spin Hamiltonian Parameters for the Standards, 2 Tar Ball Samples, and the Asphalt Volcano Samplea samples species VO

2+ b

parameters g∥ g⊥

Mn2+c

A∥ (mT) A⊥ (mT) [V] (μg/g) giso

free radicalb

Aiso (mT) g ΔHPP (mT)

Brick

Burn Residue

Il Duomo

RM8505

RM2779

VO-OEPe

1.9620 ±0.0005 1.9835 ±0.0005 17.1 ± 1 5.8 ± 0.5 70 ± 1 1.9985 ±0.0005 9.4 ± 0.1 2.0028 ±0.0005 0.5 ± 0.05

1.9620 ±0.0005 1.9835 ±0.0005 17.1 ± 1 5.8 ± 0.5 35 ± 1 1.9985 ±0.0005 9.4 ± 0.1 2.0028 ±0.0005 0.5 ± 0.05

1.9620 ±0.0005 1.9835 ±0.0005 17.1 ± 1 5.8 ± 0.5 500 ± 5 1.9985 ±0.0005 9.3 ± 0.1 2.0028 ±0.0005 0.6 ± 0.05

1.9620 ±0.0005 1.9835 ±0.0005 17.1 ± 1 5.8 ± 0.5 390 ± 10d 1.9985 ±0.0005 9.3 ± 0.1 2.0028 ±0.0005 0.6 ± 0.05

1.9620 ±0.0010 1.9835 ±0.0005 17.1 ± 1 5.8 ± 0.5 1 ± 0.1

1.9626 ±0.0005 1.9830 ±0.0005 17.1 ± 1 6.0 ± 0.5

2.0028 ±0.0005 0.6 ± 0.05

Note: Spectral parameters of VOOEP powder, extracted from 240 GHz spectrum: gx = 1.9834 ± 0.0005, gy = 1.9819 ± 0.0005, gz = 1.9591 ± 0.0005; Ax = Ay = 6.2 ± 0.1 mT; Az = 17.4 ± 0.1 mT. bParameters from X-band EPR spectra. cParameters from 240 GHz EPR spectra. dObtained from the NIST Report on the standard NIST-RM8505. eDilute solution of VOOEP in castor oil. a

Figure 2. X-band (9.4 GHz, left) and high-field (240 GHz, right) EPR spectra of the standard “Vanadium in Crude Oil, NIST-RM8505” sample at room temperature. Note that Mn2+ is not observable in the X-band spectrum. Right panel insets: (i) Axially symmetric [V−O]2+ hyperfine octets are denoted by a vanadyl porphyrin molecule with its V−O bond aligned parallel and perpendicular to the field direction. (ii) The isotropic hyperfine sextet arising from the Mn2+ ion in a symmetric local environment is symbolically shown by a Mn2+ ion inside a regular octahedron. (iii) A polyaromatic sheet-like molecule represents the free radical peak. The hyperfine constants and g-values are listed in Table 2.

well-characterized fossil fuel model compound, polycrystalline vanadyl octaethylporphyrin11 (VOOEP), measured at room

temperature. Figure 1a is the low-field EPR spectrum (X-band, ∼9.5 GHz), consisting of eight broad lines separated by about C

DOI: 10.1021/ac504080g Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

Figure 3. X-band (9.4 GHz, left) solution and high-field (240 GHz, right) frozen glass EPR spectra of the NIST standard reference material, “Gulf of Mexico Crude Oil, NIST-RM2779”. Peaks corresponding to the axially symmetric [V−O]2+ hyperfine structure, overlapped with a strong free radical signal are seen in X-band. The high-field spectrum reveals an additional sextet of peaks due to Mn2+. Simulated spectra (blue) and stick diagrams guide the eyes to identify the various features.

hyperfine octets5−7,20−22 are separately resolved and displaced from the FR-line; (ii) a new 6-line hyperfine structure centered at g = 1.9987 flanks the free radial line; and (iii) the free radical line is broad (ΔHp‑p ≈ 3 mT, 30 G) and (surprisingly) is slightly asymmetric (as discussed in more detail below). The parallel and perpendicular hyperfine lines of [V−O]2+ could be fitted to a simulated spectrum, thereby enabling much more accurate determination of the spin Hamiltonian parameters. Another significant novel feature of the 240 GHz spectrum is that the vanadyl perpendicular hyperfine structure of NISTRM8505 reveals two closely spaced perpendicular octets, signaled by a small shoulder just to the left side of the lowest field line. Thus, unlike the vanadium octaethyl porphyrin model compound, the vanadium in NIST-RM8505 possesses a rhombic local symmetry. This finding is quite reasonable, because some of the ethyl units could be substituted by other groups that would be expected to be present during tar formation. Moreover, the high-field EPR spectrum clearly resolves Mn2+, not seen in the X-band spectrum.4 Among the six lines of the hyperfine sextet at g = 1.9987, two lines are almost fully masked by the strong free radical line, whereas the remaining four lines are symmetric and located ∼93 G apart from each other, suggesting that the isotropic hyperfine sextet is due to the presence of Mn2+ in an essentially symmetric environment (see Figure 2, right inset). Mn2+ is a half-filled 3d5, spherically symmetric ion with electron spin S = 5/2 and has the major isotope 55Mn (100% naturally abundant) with nuclear spin I = 5/2, resulting in a six line hyperfine structure, with a peak spacing of 9.4 mT (94 gauss), in good agreement with the known isotropic hyperfine coupling constant of Mn2+ ion in octahedral or tetrahedral geometry.20−22,25 Once the high-field EPR data definitively identified the [V− O]2+ moiety in the NIST-RM8505 standard, we calibrated its EPR spectral intensity as a mass standard. For that purpose, we employed our commercially available X-band spectrometer, in view of its ready availability and routine, nondestructive analysis of tar ball and related samples. Measurements were based on direct integration of a parallel hyperfine peak of the first derivative spectrum chosen to be free from any overlap with other lines. The samples were weighed carefully to five significant figures prior to measurement, and the vanadium concentration in μg/g of sample was determined. The values are tabulated in Table 2.

17.5 mT (175 gauss) that can be ascribed to the 8-line hyperfine structure from the 51V (I = 7/2) nucleus. However, just counting the number of lines would imply that the molecule is essentially spherical rather than the disk shape known from its crystal structure. In the solid state (powder in this case), the spin concentration is high and hence the lines are broadened by the exchange interactions between the spins.20,24 The spectrum was simulated with locally developed computer programs18,19 to obtain the characteristic g and A parameters, which are listed in Table 2. The spectra could be simulated from g⊥ = 1.983, g∥ = 1.960; A⊥= 6 ± 1 mT, and A∥ = 17.5 ± 1 mT, quite reasonable for a [V−O]2+ species.23 Much more informative spectra were obtained at higher fields. Figure 1b shows the high-field (240 GHz) powder spectrum of VOOEP measured at room temperature. This spectrum presents greatly enhanced dispersion/resolution. It shows that the number of lines is actually 16 (as indicated by the analysis of Figure 1, left): eight from molecules whose V−O bond lies in the plane of the magnetic field and another eight from those whose V−O bond lies in a plane perpendicular to the magnetic field. Further analysis of the spectrum revealed that the perpendicular hyperfine octet consists of a small but significant splitting into two octets closely located, characterized by the two g-values, gx (1.9834) ≠ gy (1.9819), instead of one g⊥ (1.9826) for X-band. This effect could be due to a slight distortion of the square pyramidal geometry of the VOOEP molecule.22 NIST Crude Oil Standards. Figure 2 shows X-band (9.4 GHz) and 240 GHz EPR spectra of a standard “Vanadium in Crude Oil” NIST-8505 RM8505 sample at room temperature. NIST-RM8505 is a highly viscous, dark brown liquid, whose vanadium concentration is certified as 390 ± 10 μg/g and which serves as a calibration standard for accurately determining the content of vanadium in crude oils. The Xband spectrum (Figure 2, left) shows two distinct features: (i) a sharp symmetric line with a g-value (2.0025 ± 0.0005) very close to the free-electron g-value (2.0023) with a peak-to-peak line width of ΔHp‑p ≈ 0.6 mT (6 G), due to the presence of a hydrocarbon free radical, and (ii) a set of two hyperfine octets, with significant overlap of lines around the free radical peak. Although the spectrum could be simulated with known parameters (see Table 2), a much better resolved spectrum was obtained at 240 GHz (Figure 2, right): (i) the vanadyl D

DOI: 10.1021/ac504080g Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

Figure 4. Room temperature, X-band (9.4 GHz, left) and high-field (240 GHz, right) EPR spectra of the “Burn Residue” tar ball. The peak labeled FR is from an organic free radical, which is flanked by two overlapping octets due to the axially symmetric hyperfine structure from a [V−O]2+ species. The high-field spectrum shows signals from Mn2+ hyperfine structure that are not resolved at low field.

Figure 5. Room temperature, X-band (9.4 GHz, top) and high-field (240 GHz, bottom) EPR spectra from “Brick” (left) tar ball from the Gulf of Mexico and an “Il Duomo” (right) asphalt volcano sample collected off the coast of Santa Barbara, CA, but not related to the Deepwater Horizon Oil Spill. At X-band, both the tar ball and marine asphalt exhibit spectral features of an axially symmetric [V−O]2+ species and a free radical, as for the “Burn Residue” in Figure 4, whereas the “Il Duomo” marine asphalt also shows overlapping Mn2+ hyperfine peaks at X-band. High-field spectra exhibit separately resolved [V−O]2+ and Mn2+ features and a new unknown peak at g = 2.0121 for both samples.

Tar Balls. Burn Residue. A “Burn Residue” thick black tar ball was obtained from a controlled burn of weathered oil from the Deepwater Horizon spill. The sample agglomerated before sinking to a depth at which it adhered to a shrimp trawler. Its X-band EPR spectrum at room temperature (Figure 4, left) exhibits two major features similar to those for NIST-RM8505, namely, (i) a sharp central peak centered at g = 2.0029, arising from a free radical, and (ii) two sets of 8-line hyperfine structure with g⊥ > g∥ (1.9835 > 1.9620) and A⊥ < A∥ (58 G < 172 G), which are characteristic of an axially symmetric [V− O]2+-containing species.20,23 The 240 GHz EPR spectrum of

The X-band EPR spectrum of another standard, “Gulf of Mexico Crude Oil, NIST-RM2779”, a brown liquid, is shown in Figure 3, left. The spectrum consists of a strong central free radical peak at g = 2.0024, surrounded by several weak lines which could be assigned to [V−O]2+, at a concentration of only ∼1 μg/g of vanadium. The 240 GHz spectrum of a frozen glass sample of NIST-RM2779, at 20 K (Figure 3b), exhibits a strong free radical peak. Note that, due to the low resolution and high spectral overlap with strong peaks from the organic radical and [V−O]2+ complex, Mn2+ is not noticeable in the X-band spectrum. E

DOI: 10.1021/ac504080g Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry “Burn Residue” (Figure 4, right) exhibits three distinct features corresponding to a free radical, [V−O]2+, and a six-line multiplet with a ∼9.4 mT splitting assigned to Mn 2+ species,20,25 again similar to those observed for NIST-RM8505. Brick and Il Duomo. The other tar ball and asphalt volcano sample, unrelated to the Deepwater Horizon spill, “Brick”, collected from Elmers Island in the Gulf of Mexico, and “Il Duomo”, a heavily weathered asphalt volcano sample from natural petroleum seepage in the Pacific Ocean, were analyzed for comparison. Figure 5 shows room temperature X-band and 240 GHz EPR spectra of Brick and Il Duomo. The X-band spectra of both samples are very similar to that of Burn Residue (see Figure 4 and Table 2). One new feature from Il Duomo is a six-line hyperfine structure due to Mn2+ at X-band, indicating a high level of manganese in that sample. The high-field EPR spectrum of Brick shows clearly resolved features from free radical, [V−O]2+, and Mn2+, as for Burn Residue. Both Brick and Il Duomo show an additional asymmetrical peak of g = 2.0121. Identification of the Vanadyl Complex in the Tar Balls: Comparison with Vanadyl Octaethylporphyrin (VOOEP). To determine the chemical structure of the vanadyl complex in the tar balls and marine asphalt, we used vanadyl octaethylporphyrin13 (VOOEP) as a standard, because earlier literature has indicated the possibility of such a structure for the ubiquitously present vanadium complex in crude oil asphaltenes.6−8 It is necessary to choose a solvent in which VOOEP could be dissolved at dilute concentration comparable to that for the vanadium complex in the NIST standard and the tar balls. Simply spiking the tar balls and marine asphalt with VOOEP did not suffice, because the concentration of the paramagnetic center ([V−O]2+) in VOOEP is so high that the EPR signal is exchange-broadened and not directly comparable to the signals from the field samples. Toluene (a commonly used solvent for petroleum) is also not suitable, because the EPR spectrum of VOOEP in toluene is isotropic (8 hyperfine peaks), whereas the dissolved tar ball spectrum is anisotropic (16 peaks), like those shown in Figure 1. The molecular weight of the vanadyl complex in the asphaltenes is much larger than that of VOOEP; hence, the latter tumbles much faster in toluene and thus produces an isotropic spectrum.21 To produce slower tumbling, we used castor oil as a solvent for VOOEP, yielding the desired anisotropic spectrum of VOOEP. Figure 6 shows the EPR spectra from a tar ball “Brick” along with its computer simulation and similar spectra from VOOEP. The EPR parameters used in the simulations are listed in Table 2. It is clear that the spectra of the tar balls/marine asphalt and of VOOEP are nearly identical, even though the castor oil environment is not identical to that for a tar ball and the tar ball does not likely exhibit the symmetry of the eight ethyl groups around the porphyrin core High-Field EPR of the Organic Free Radical. Because the organic free radical peak is common to all of the samples, it is of special interest to examine its EPR spectrum at high field. Figure 7 shows low-field (X-band, left panel) and high-field (240 GHz, right panel) room temperature EPR spectra. The low-field spectrum is a single, derivative-Lorentzian, symmetric peak, of g-value 2.0028, consistent with a pure carbon centered radical, with an essentially spherical average shape, a knotted ball of a large polymeric paramagnetic species, as postulated in prior low-field EPR studies.5 The asymmetric high-field EPR spectrum gives a more correct picture, showing a clear shoulder, diagnostic for an axially symmetric spectrum. This

Figure 6. Room temperature, X-band (9.4 GHz) EPR spectra of the “Brick” tar ball and dilute vanadyl octaethylporphyrin (VOOEP) solution in castor oil. (a) Experimental spectra of Brick and VOOEP; dashed vertical lines serve as guides to the eye, showing the nearidentical locations of the vanadyl hyperfine peaks. Note that the free radical peak is truncated so as to render the vanadyl signal more visible. (b) Experimental and simulated spectra of the “Brick” tar ball. (c) Experimental and simulated spectra of “VOOEP in castor oil”. Stick diagrams provide guides to the eye and serve to identify various components of the anisotropic [V−O]2+ spectra in (a) and (b).

spectrum represents a superposition of two peaks: g⊥ value of 2.0004 and g∥ value of 1.9999. The g-value measurements reflect the local symmetry of the distribution of the spin density of the unpaired electron over the radical molecule’s electronic skeleton. An extensive EPR literature search yielded g-values of various organic free radicals, as tabulated in the Supporting Information (Table S-1). Unfortunately, there seems to be no prior report of the g-anisotropy for any hydrocarbon free radical, most likely because the anisotropy is so small that it stays undetected in low-field EPR measurements. Our study thus suggests a fruitful new application of high-field EPR, namely, measurement of the g-anisotropy of organic radicals, and their utility for understanding the electronic structure of radicals wherein hyperfine structure is not resolved, even in high-field EPR. The presence of axial symmetry is direct evidence that the unpaired electron is delocalized over at least two carbon atoms. In addition, because no proton hyperfine structure is resolved F

DOI: 10.1021/ac504080g Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

Figure 7. Room temperature, X-band (9.4 GHz, left) and high-field (240 GHz, right) EPR spectra of the organic free radical component of the “Brick” tar ball. The X-band free radical spectrum shows no structure and could be interpreted as an isotropic peak arising from a spherically symmetric molecule (left inset). The asterisk (∗) marks the location of one of the perpendicular [V−O]2+ hyperfine lines lying close to the free radical peak. The high-field (240 GHz) FR spectrum splits into two peaks, showing that the free radical molecule is not spherically symmetric (left inset) but rather a disc-like molecule such as from a fused aromatic structure (right inset).

bonding and molecular shape. In contrast, high-field (∼9 T) EPR correctly identifies both paramagnetic metals and organic free radicals and for the first time reveals the molecular symmetry of the organic radicals. Future work will focus on electron/nuclear double resonance to identify the protons nearest to the unpaired electron in the organic radicals.

in the EPR peak, in spite of the high microwave frequency or Zeeman field, the unpaired electron must be strongly delocalized over the molecular framework so as to result in little spin density at a given proton. These results show for the first time that the organic free radical is very likely a large, planar, aromatic ring, with the unpaired electron delocalized over the whole ring, and explain its stability over widely different conditions. These studies suggest that EPR measurements at still higher microwave frequencies might be fruitful and are in our near-future plans. Other Paramagnetic Species. For the present environmental samples, we have been able to identify essentially all of the observed EPR peaks and assign them to the free radical, Mn2+ hyperfine structure, or the extensive hyperfine structure of the vanadyl porphyrin complexes. High-field EPR yields qualitatively new information on the hyperfine features of the vanadyl-porphyrin-based complexes; this result opens up new avenues for characterizing many metalloenzymes and metalloproteins wherein the vanadyl ion has been found to be a sensitive spin probe of ligand structure and dynamics.23 In the context of the free radical identification, we particularly looked for oxygenated radical species or those containing sulfur substituents. Both of those species should exhibit g-values significantly higher than 2.00, and those peaks should shift in proportion to the applied Zeeman field. We were not able to identify any such peaks in the as obtained samples. We note that Kiruri et al.4 detected oxygenated free radical species in weathered tar balls, although no structural details were presented, presumably due to the absence of hyperfine structure. We conclude that the main free radical species are essentially large, multiply substituted aromatic molecules with the free electrons highly delocalized on the carbon−carbon bonded skeleton. Environmental Relevance. As evidenced by the present results, environmental samples can contain a variety of paramagnetic species, including metal ions (both free and complexed) and organic free radicals. Electron paramagnetic resonance offers the most direct probe of such species. However, the present results demonstrate that low-field (∼0.3 T) EPR is insufficient to reveal details of the electronic structure, leading to incorrect identification of molecular symmetry (spherical rather than axial) and thus the chemical

■

CONCLUSIONS Although the NIST standards are known to contain all three paramagnetic species, the vanadyl (V−O)2+ moiety, the organic free radical, and the Mn2+ complex, all three are not simultaneously detectable in the NIST standards by low-field (X-band) EPR. In particular, Mn2+ remains essentially undetectable at X-band, owing to overlap of the dominant signals from the organic free radical and the (V−O)2+ complex. Mn2+ is, however, clearly seen at 240 GHz (see Figures 2, 4, and 5). The Mn2+ peaks (hyperfine sextet) are weak and thus overwhelmed by the strong free radical and (V−O)2+ signals at X-band but are clearly discernible at 240 GHz owing to the higher dispersion at higher magnetic field. By use of the NIST vanadyl-porphyrin standard, we were able to quantitatively measure the concentration of [V−O]2+ in the tar balls. In addition, the EPR of dissolved VOOEP in castor oil as a standard enabled us to show that the chemical structure of the vanadyl species in the tar balls is quite similar to that of VOOEP. Our ability to measure the Mn2+ (also commonly denoted Mn (II)) peaks at 240 GHz demonstrates that this metal is more concentrated in the tar balls relative to the source. Biochemical effects of the additional Mn2+ metal complex must thus also be considered for evaluating the environmental impact of such tar balls. Finally, only the high field spectra at 240 GHz reveal that the g-value of the organic free radical is not just a single isotropic value of ∼2.000 ± 0.0005 but is in fact a tensor with at least two (and possibly three) components. That is important because over the past several decades, on the basis of the available (low-field) EPR data (suggesting a nearly freeelectron, isotropic g-value), this free radical peak has been attributed to some sort of a jumbled-up hydrocarbon polymer with essentially spherical symmetry (generally known as the G

DOI: 10.1021/ac504080g Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

Savory, J. J.; Kaiser, N. K.; Marshall, A. G.; Rodgers, R. P. Energy Fuels 2014, 28, 2454−2464. (15) van Tol, J.; Brunel, L.-C.; Wylde, R. J. Rev. Sci. Instrum. 2005, 76, 074101. (16) Hassan, A. K.; Pardi, L. A.; Krzystek, J.; Sienkiewicz, A.; Goy, P.; Rohrer, M.; Brunel, L.-C. J. Magn. Reson. 2000, 142, 300−312. (17) Dalal, N. S.; Kennedy, D. E.; McDowell, C. A. J. Chem. Phys. 1973, 59, 3403−3410. (18) Ozarowski, A. DOUBLET; National High Magnetic Field Laboratory: Tallahassee, FL, 2005. (19) van Tol, J. EPR-Calc; National High Magnetic Field Laboratory: Tallahassee, FL, 2006. (20) Pilbrow, J. R. Transition Ion Electron Paramagnetic Resonance; Clarendon Press: Oxford, U.K., 1990. (21) Weil, J. A.; Bolton, J. R.; Wertz, J. E. Electron Paramagnetic Resonance − Elementary Theory and Practical Applications; WileyInterscience: New York, 1994. (22) Abragam, A.; Bleaney, B. Electron Paramagnetic Resonance of Transition Ions; Oxford University Press: Oxford, England, 1970. (23) For detailed reviews on [V-O]2+ in metalloproteins, including biological samples, see (a) Chasteen, N. D. Struct. Bonding (Berlin) 1983, 53, 105−138. (b) Larsen, S. A.; Chasteen, N. D. Biol. Magn. Reson. 2010, 29, 371−409. (24) Dalal, N. S.; Millar, J. M.; Jagadeesh, M. S.; Seehra, M. S. J. Chem. Phys. 1981, 74, 1916−1923. (25) Wang, Z.; Zheng, W.; van Tol, J.; Dalal, N. S.; Strouse, G. F. Chem. Phys. Lett. 2012, 524, 73−77.

T.F. Yen model, ref 5, Figure 7 in the current manuscript). The 240 GHz data clearly show that the radical geometry/structure must be fairly rigid and essentially two-dimensional, like a resonance-stabilized, substituted aromatic hydrocarbon moiety, devoid of oxygen and heavier atoms, and containing a strongly delocalized unpaired electron. That information should provide a precise basis for testing various theoretical calculations for organic radicals in fossil fuels.

■

ASSOCIATED CONTENT

S Supporting Information *

Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Authors

*Phone: 850-644-3398. Fax: 850-644-8281. E-mail: dalal@ chem.fsu.edu. *Phone: 850-644-0529. Fax: 850-644-1366. E-mail: marshall@ magnet.fsu.edu. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We thank Professor David L. Valentine for providing the Il Duomo sample. We thank the Geochemistry group at NHMFL, especially Munir Humayan and Vincent J. Salters for access to the asher and ICPMS analyzer. This work was supported by NSF DMR-11-57490, BP/The Gulf of Mexico Research Initiative to the Deep-C Consortium, and the State of Florida.

■

REFERENCES

(1) Malakoff, D. Science 2015, DOI: 10.1126/science.aaa6358. (2) Aeppli, C.; Carmichael, C. A.; Nelson, R. K.; Lemkau, K. L.; Graham, W. M.; Redmond, M. C.; Valentine, D. L.; Reddy, C. M. Environ. Sci. Technol. 2012, 46, 8799−8807. (3) McKenna, A. M.; Nelson, R. K.; Reddy, C. M.; Savory, J. J.; Kaiser, N. K.; Fitzsimmons, J. E.; Marshall, A. G.; Rodgers, R. P. Environ. Sci. Technol. 2013, 47, 7530−7539. (4) Kiruri, L. W.; Dellinger, B.; Lomnicki, S. Environ. Sci. Technol. 2013, 47, 4220−4226. (5) Wong, G. K.; Yen, T. F. J. Pet. Sci. Eng. 2000, 28, 55−64. (6) Nasirov, R. N.; Vel’k, O. D.; Solodovnikov, S. P. Khim. Tekhnol. Topl. Masel 1991, 27, 635−637. (7) Malhotra, V. M.; Buckmaster, H. A. Fuel 1985, 64, 335−341. (8) Espinosa, M. P.; Campero, A.; Salcedo, R. Inorg. Chem. 2001, 40, 4543−4549. (9) Biktagirov, T. B.; Gafrov, M. R.; Volodin, M. A.; Mamin, G. V.; Rodionov, A. A.; Izotov, V. V.; Vakhin, A. V.; Isokov, D. R.; Orlinskii, S. B. Energy Fuels 2014, 28, 6683−6687. (10) Trukhan, S. N.; Yudanov, V. F.; Gabrienko, A. A.; Subramani, V.; Kazarian, S. G.; Martyanov, O. N. Energy Fuels 2014, 28, 6315− 6321. (11) National Bureau of Standards Report of Investigation, Vanadium in Crude Oil, Research Material 8505; National Institute of Standards and Technology: Gaithersburg, MD, 1983. (12) Certificate of Analysis, Standard Reference Material 2779; National Institute of Standards and Technology: Gaithersburg, MD, 2012. (13) 2,3,7,8,12,13,17,18-Octaethyl-21H,23H-porphine vanadium(IV) oxide, CAS Number 27860-55-5; Sigma Aldrich Co.: St. Louis, MO; http://www.sigmaaldrich.com/catalog/product/aldrich/363715?lang= en®ion=US. (14) McKenna, A. M.; Williams, J. T.; Putnam, J. C.; Aeppli, C. M.; Reddy, C. M.; Valentine, D. L.; Lemkau, K. T.; Kellerman, M. Y.; H

DOI: 10.1021/ac504080g Anal. Chem. XXXX, XXX, XXX−XXX