Organic Geochemistry 44 (2012) 21–36

Contents lists available at SciVerse ScienceDirect

Organic Geochemistry
journal homepage: www.elsevier.com/locate/orggeochem

Structural characterization of gilsonite bitumen by advanced nuclear magnetic

resonance spectroscopy and ultrahigh resolution mass spectrometry revealing
pyrrolic and aromatic rings substituted with aliphatic chains
John R. Helms a, Xueqian Kong b, Elodie Salmon a, Patrick G. Hatcher a, Klaus Schmidt-Rohr b,⇑,
Jingdong Mao a,⇑
a
Old Dominion University, Department of Chemistry and Biochemistry, 4541 Hampton Blvd., Norfolk, VA 23529, United States
b
Iowa State University, Department of Chemistry, Gilman Hall, Ames, IA 50011, United States

a r t i c l e i n f o a b s t r a c t

Article history: Gilsonite, a naturally occurring asphaltite bitumen, consists of a complex mixture of organic compounds.
Received 2 March 2011 In the present study, advanced one and two dimensional solid state and solution 1H, 13C and 15N nuclear
Received in revised form 29 November 2011 magnetic resonance (NMR) and electrospray ionization Fourier transform ion cyclotron resonance mass
Accepted 7 December 2011
spectrometry (ESI-FT-ICR-MS) were employed to investigate its composition and structure. 13C NMR
Available online 14 December 2011
yielded a carbon aromaticity of 27%. Aromatic moieties in gilsonite were primarily single rings or small
clusters of fused rings. Half of the aromatic carbons of gilsonite can be accounted for by pyrroles. 15N and
13
C cross polarization-magic angle spinning (CP-MAS) NMR showed that most nitrogen in gilsonite was
pyrrolic. The aromatic rings were heavily substituted with alkyl chains, as evidenced by 1HA13C correla-
tion spectra. Advanced solid state NMR spectral editing techniques clearly identified specific functional
groups such as CCH3, CCH2, and C@CH2 (exomethylene). 1HA13C wideline separation (WISE) NMR helped
identify mobile and non-protonated alkyl carbons. FT-ICR-MS indicated that 64% of calculated formulae
generated by ESI were aliphatic, while only about 0.8–2.5% of formulae contained possible aromatic rings.
All of the assigned formulae contained at least one heteroatom (N, O or S), indicating that ionization by
ESI was selective for the polar fraction of gilsonite and potentially less reflective of the overall chemical
character of gilsonite than NMR spectroscopy. By combining the information obtained from advanced
NMR and ultrahigh resolution MS we propose a structural model for gilsonite as a mixture of many pyr-
rolic and a few fused aromatic rings highly substituted with and connected by mobile aliphatic chains.
Ó 2011 Elsevier Ltd. All rights reserved.

1. Introduction

Bitumen refers to solid or liquid hydrocarbon deposits soluble

Abbreviations: AI, aromatic index; AImod, modified aromatic index; BC, black
carbon; CARS, condensed aromatic ring structures; CP, cross polarization; COSY,
correlation spectroscopy (2D-NMR); CSA, chemical shift anisotropy; DBE, double
bond equivalents; DEPT, distortionless enhancement by polarization transfer; DP,
direct polarization; DPEP, deoxyphylloerythroetioporphyrin ; EA, elemental anal-
ysis; ESCA, electron spectroscopy for chemical applications (X-ray photoelectron
spectroscopy or XPS); ESI, electrospray ionization; FT, Fourier transform; HETCOR,
heteronuclear correlation (2D-NMR); HH-CP, Hartmann–Hahn cross polarization;
HMQC, heteronuclear multi-quantum correlation (2D-NMR); ICR, ion cyclotron
resonance; KMDA, Kendrick mass defect analysis; LG-CP, Lee-Goldburg cross
polarization; MAS, magic angle spinning; MS, mass spectrometry; MREV-8,
Mansfield, Rhim, Elleman and Vaughn’s 8 pulse line narrowing sequence; NMR,
nuclear magnetic resonance spectroscopy; NOM, natural organic matter; ppm, parts
per million units of frequency variation from a standard; T1, spin–lattice relaxation
time; TOCSY, total correlation spectroscopy (2D-NMR); TOSS, total sideband
suppression; TPPM, two-pulse phase modulated decoupling; WISE NMR, wideline
separation nuclear magnetic resonance.
⇑ Corresponding authors. Tel.: +1 515 294 6105; fax: +1 515 294 0105 (K.
Schmidt-Rohr), tel.: +1 757 683 6874; fax: +1 757 683 4628 (J. Mao).
E-mail addresses: srohr@iastate.edu (K. Schmidt-Rohr), jmao@odu.edu (J. Mao).
in organic solvents (Killops and Killops, 2005). Gilsonite is a natu-
rally occurring bituminous asphalt and formally classified as an
asphaltite bitumen (Abraham, 1945). It occurs primarily along
the Colorado–Utah border and was found in the early 1860s in
the Uinta Basin in northeastern Utah (Bell and Hunt, 1963). Gilson-
ite has been widely used in mining and industry and sold all over
the world. For example, it has been used as an additive in oil dril-
ling fluids and for oil well cementing and has a long history of use
as both a pigment and binding agent in paints, enamels and inks
(Tripp and White, 2006). Natural bitumens are also under investi-
gation as exploitable petroleum fuels (Meyer et al., 2007). Further-
more, gilsonite and other natural bitumens are important
intermediates and/or byproducts in the geological formation of
crude oils that can be analyzed to improve our understanding of
petroleum geochemistry (Bell and Hunt, 1963; Killops and Killops,
2005).

0146-6380/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.orggeochem.2011.12.001
22 J.R. Helms et al. / Organic Geochemistry 44 (2012) 21–36
Despite the widespread industrial use and petrochemical po-
tential of gilsonite, its molecular structure has not been well char-
acterized due to its complexity and heterogeneity. Past studies
have primarily focused on the investigations of biomarkers such
as petroporphyrins (Treibs, 1934, 1936; Sugihara and McGee,
1957; Corwin, 1959; Quirke and Maxwell, 1980; Quirke et al.,
1980a, 1980b; Hajibrahim et al., 1981; Gill et al., 1985), hopanes,
steranes and carotenoids (Ruble et al., 1994; Schoell et al., 1994).
Quirke et al. (1980b) found that the aetio-type porphyrins and
deoxyphylloerythroetioporphyrin (DPEP) were products of reduc-
tive degradation of chlorophylls. The porphyrins present in gilson-
ite were most likely formed from naturally occurring chlorophylls
through this degradation (Quirke and Maxwell, 1980; Quirke et al.,
1980b; Hajibrahim et al., 1981). Compound specific isotopic analy-
sis of carotenoids indicated a strong algal or cyanobacterial source
for the diagenetic starting material for gilsonite (Ruble et al., 1994;
Schoell et al., 1994). Additionally, isotopic values for hopanes (pen-
tacyclic membrane lipids) and moretanes suggested the influence
of mid-water bacteria and potentially some methanotrophs (Scho-
ell et al., 1994).
However, biomarkers such as the porphyrin compounds only ac-
count for a minor fraction of gilsonite (Treibs, 1936). The bulk of the
gilsonite consists of aliphatic and aromatic hydrocarbons with few
olefinic groups, as indicated by the iodine number (Baker et al.,
1967; Quirke and Maxwell, 1980). Its bulk chemical characteristics
have been investigated using electron spectroscopy for chemical
applications (ESCA; usually referred to as X-ray photoelectron spec-
troscopy or XPS), infrared spectroscopy, X-ray diffraction spectros-
copy, and 1H and 13C solution NMR analysis. ESCA provided
verification of elemental analysis results. ESCA also indicated that
carbon was mainly present as aliphatic hydrocarbons, that nitro-
gen–carbon bonding environments such as aliphatic and aromatic
amines as well as nitrogen heterocycles such as pyrrole were pres-
ent, that CAO bonding environments were present and that a very
small amount of sulfur was present as organic sulfides or heterocy-
cles such as thiophenes (Clark et al., 1983). Infrared spectroscopy
showed that CAH (88% of IR signal) and CAO (10% of IR signal)
were the dominant vibrational modes present in gilsonite with
acids and esters each representing <1% of the vibrational signal
(Clark et al., 1983). Organic acids were present in gilsonite in trace
amounts (Grantham and Douglas, 1977; Clark et al., 1983).
Yen et al. (1961) used X-ray diffraction to estimate the aroma-
ticity of gilsonite as well as several petroleum asphaltenes. They
reported a much lower aromaticity for gilsonite (14%) than for
petroleum asphaltenes (22–53%). However, Bell and Hunt (1963)
described gilsonite as ‘‘predominantly aromatic’’ based on liquid
chromatography, elemental analysis and infrared measurements.
Wen et al. (1978) report the aromaticity of gilsonite asphaltene,
which is the pentane insoluble, benzene soluble fraction, as 38%
based on 1H solution NMR. Yet the perception of gilsonite as a
‘‘predominantly aromatic substance’’ (North, 1985) persisted until
Clark et al. (1983) applied solution 1H and 13C NMR to gilsonite
samples and showed a variety of aliphatic carbon environments
and a single broad aromatic peak. They reported an aromatic:ali-
phatic carbon ratio of 1:4 for a gilsonite sample with limited atmo-
spheric exposure, which was more in line with the X-ray
diffraction results (Yen et al., 1961) and the absence of a ‘shake-
up satellite’ peak, indicative of conjugated systems, in the ESCA
spectrum (Clark et al., 1983). To resolve this controversy, reliable
quantification of aromaticity is required. Direct polarization 13C
NMR, after a sufficiently long relaxation delay (Mao et al., 2000;
Mao and Schmidt-Rohr, 2004a), provides accurate aromaticity val-
ues. Gilsonite has often been described as containing condensed
polyaromatic ‘sheets’ or ‘asphaltene sheets’ (Yen et al., 1961; Bell
and Hunt, 1963; Wen et al., 1978; Yen, 2000) based on its high car-
bon content and its high refractive index, but little direct evidence
has supported this. Solid state NMR can address this question by
estimating the size of clusters of fused aromatic rings, by long
range CAH dipolar dephasing and quantification of the ring edge
carbon fractions (Brewer et al., 2009; Mao et al., 2010b).
The structural characterization of gilsonite is critical for under-
standing its properties, its applications in industry, and its geolog-
ical significance. Solid state NMR has been widely applied in the
study of natural organic matter (NOM) such as coal, oil shale, hu-
mic materials and peats (Hatcher et al., 1981, 1983; Dennis et al.,
1982; Miknis et al., 1982; Wilson, 1987; Solum et al., 1989; Ander-
son et al., 1992; Maciel et al., 1993; Wilson et al., 1993; Preston,
1996; Nanny et al., 1997; Clifford et al., 1999; Hu et al., 2000; Conte
et al., 2004; Smernik et al., 2006). Advanced NMR techniques, espe-
cially spectral editing techniques, have increased the amount of
information obtainable from NOM samples (Wu and Zilm, 1993;
Wu et al., 1994; Hu et al., 2000; Keeler and Maciel, 2000). We have
developed and applied many new, advanced solid state NMR tech-
niques, especially spectral editing (Schmidt-Rohr and Mao, 2002;
Mao and Schmidt-Rohr, 2003), for the investigations of NOM
(Mao et al., 2007a, 2007c). While the broad and heavily overlapped
bands of complex NOM allow traditional 13C NMR to identify only
about 10 types of chemical groups, our spectral editing techniques,
which selectively retain peaks of certain types of functional groups,
can identify more than two dozen different moieties (Mao et al.,
2007a, 2007c). Moreover, two dimensional 1HA13C heteronuclear
correlation (HETCOR) NMR is used to detect proximities and con-
nectivities of different functional groups. 15N CP-MAS NMR is ap-
plied to characterize forms of nitrogen in gilsonite and combined
with recoupled dipolar dephasing to determine which resonances
are from N bonded to hydrogen (Fang and Schmidt-Rohr, 2009).
Gilsonite is soluble in a range of organic solvents such as chloro-
form; therefore we have also studied it by solution NMR techniques.
Solution 13C NMR spectra appear well resolved; however, the signals
of many components are broad and almost invisible due to low
mobility of some components and resulting short spin–spin relaxa-
tion times (T2). Thus, advanced solid state and solution NMR tech-
niques provide complementary structural information.
Mass spectrometry is used to determine the molecular weight
of ions. In organic chemistry, advanced solution NMR and mass
spectrometry data have been routinely collected to elucidate the
structures of organic compounds. However, the approach has not
been frequently used to elucidate the structures of natural organic
matter in organic geochemistry. Despite the availability of high
resolution mass spectrometry for the analysis of fossil fuels for
over two decades (Comisarow and Marshall, 1974; Grigsby,
1989; Hsu et al., 1994; Amster, 1996; Marshall et al., 1998), to date,
there has been no detailed, comprehensive characterization of gil-
sonite using Fourier transform ion cyclotron mass spectrometry
(FT-ICR-MS) reported in the literature. The high mass accuracy of
FT-ICR-MS makes it possible to differentiate ions in complex mix-
tures without employing a separation method prior to infusion
into the mass spectrometer (Klein et al., 2006).
Mass separation is only carried out on charged particles (ions)
and therefore the quality of MS data strongly depends on the type
of ionization used during sample introduction. Electrospray ioniza-
tion (ESI) coupled with FT-ICR-MS yields a detailed mass spectrum
of ionic organic molecules with highly accurate mass to charge (m/
z) determination. Positive ionization and negative ionization modes
are selective in favor of chemical moieties capable of bearing posi-
tive or negative formal charges, respectively. As anticipated, posi-
tive ionization exhibits higher selectivity/sensitivity for nitrogen
containing molecules such as amines, amides and nitrogen hetero-
cycles, while negative ionization shows higher selectivity/sensitiv-
ity to acidic moieties such as fatty acids and phenols.
Smith et al. (2008) used ESI-FT-ICR-MS to characterize heavy
vacuum gas oil distillation cuts from Athabasca bitumen and found
 J.R. Helms et al. / Organic Geochemistry 44 (2012) 21–36
that the technique could rapidly and reproducibly analyze the po-
lar fraction of petroleum distillates and provide unique molecular
formulae assigned to each peak. It has been shown that chemically
isolated petroporphyrins can be analyzed by ESI-MS (Van Berkel
et al., 1993), however, it has not been determined whether they
are effectively ionized by ESI in the more complex matrix of unal-
tered bitumen.
It is the objective of this study to characterize the structure of
gilsonite bitumen by advanced NMR and FT-ICR-MS techniques.
Based on the detailed data provided by the two techniques, we
propose a structural model for gilsonite. We believe this is the
most comprehensive chemical characterization to date of a fossil
fuel sample using a combination of advanced spectroscopic tech-
niques. The protocol used here can be applied to the study of other
fossil fuels and related samples, including their extracts, coals, ker-
ogens, algaenans and other bitumens. Finally, the core structures
identified in preserved geological samples such as gilsonite will
help gain insights into the structures of potentially refractory com-
ponents in nature.
2. Materials and methods
2.1. Geological setting
The sample of the bitumen, gilsonite, used in this study was col-
lected in the Uinta Basin near the town of Bonanza, Utah and was
provided by Dr. Gary Thompson at Rocky Mountain College, Mon-
tana. The Uinta Basin covers an area of ca. 24,000 square kilome-
ters (Osmond, 1964). Over 2200 m of lacustrine sediments,
primarily siliciclastic and carbonate, were deposited there between
the late Cretaceous and middle Eocene (Picard and High, 1968; Ru-
ble et al., 1994). The source beds of gilsonite are the calcareous oil
shales of the Parachute Creek Member (Mahogany Zone) of the
Green River Formation, which are rich in organic matter (Bell
and Hunt, 1963).
2.2. NMR spectroscopy
Most solid state NMR experiments, except several spectral edit-
ing techniques and long range dipolar dephasing, were performed
on finely ground, but otherwise unaltered gilsonite using a Bruker
DSX spectrometer at 100 MHz for 13C, with magic angle spinning
(MAS) in a 7 mm double resonance probehead. The high speed
quantitative direct polarization-magic angle spinning (DP-MAS)
experiments were run in a 4 mm double resonance probehead. So-
lid state 15N cross polarization-magic angle spinning (CP-MAS) and
CP-MAS with recoupled dipolar dephasing were performed at
40 MHz for 15N in a 7 mm double resonance probehead. All the
solution NMR experiments were conducted by dissolving gilsonite
in deuterated chloroform (CDCl3) in a 5 mm triple resonance pro-
behead using a Bruker DRX 400 spectrometer, with 1H and 13C
experiments performed at 400 and 100 MHz, respectively. Chemi-
cal shifts were calculated using ACD/lab v.12.0.
2.2.1. Solid state NMR
High speed quantitative 13C direct polarization-magic angle
spinning (DP-MAS) NMR and high speed quantitative 13C DP-MAS
NMR with recoupled dipolar dephasing provide quantitative struc-
tural information. These were run at a spinning speed of 14 kHz.
The 90° 13C pulse length was 4 ls. Recycle delays were tested by
the cross polarization-spin lattice relaxation time-total suppression
of sidebands (CP-T1-TOSS) technique to ensure that all carbon sites
were fully relaxed (Mao et al., 2000). The recycle delay was 150 s
and 512 scans were collected. This technique was fully described
elsewhere (Mao and Schmidt-Rohr, 2003). In order to highlight mo-
23
bile aliphatic components, a 13C DP/MAS spectrum with a short re-
cycle delay of 1.5 s was also recorded, with 2000 scans. Recoupled
dipolar dephasing for 13C DP-MAS NMR at a spinning speed of
14 kHz was employed to obtain quantitative information on the
non-protonated carbons and carbons of mobile groups (Mao and
Schmidt-Rohr, 2003). The dipolar dephasing time was 68 ls, the re-
cycle delay was 150 s, and 512 scans were collected.
Qualitative composition information was obtained with good
sensitivity by 13C cross polarization-total sideband suppression
(CP-TOSS) NMR experiments at a spinning speed of 6.5 kHz and a
cross polarization (CP) time of 1 ms, with a 1H 90° pulse length
of 4 ls and a 2 s recycle delay. The short recycle delay for CP meth-
ods is predicated on the shorter 1H spin–lattice relaxation time
(0.5 s for gilsonite) compared to 13C and allows for many more
scans to be completed within a given time frame. Four-pulse total
suppression of sidebands (TOSS) (Dixon, 1982) was employed be-
fore detection and two pulse phase modulated (TPPM) decoupling
was applied for optimum resolution. One thousand scans were col-
lected. The corresponding subspectrum with signals of non-pro-
tonated carbons and mobile groups such as rotating CH3 was
obtained by 13C CP-TOSS combined with 40 ls dipolar dephasing.
The number of scans was 1024, with a 2 s recycle delay.
In order to separate the signals of anomeric carbons (OACAO)
from those of aromatic carbons, both of which resonate between
90 and 120 ppm, the aromatic carbon signals were selectively sup-
pressed by a five pulse 13C chemical shift anisotropy (CSA) filter
with a CSA filter time of 47 ls (Mao and Schmidt-Rohr, 2004b).
The number of scans was 1024, with a 2 s recycle delay.
The combined spectrum of immobile CH2 and CH groups was
obtained with good sensitivity in a simple spectral editing experi-
ment. First, a 13C CP-TOSS spectrum was recorded using a short CP
time of 50 ls. It showed predominantly protonated carbons in
immobile segments, but residual peaks of quaternary carbons re-
sulted from two-bond magnetization transfer. Second, a 13C CP-
TOSS spectrum was recorded using a short CP of 50 ls and 40 ls
dipolar dephasing. It contained only the residual signals of quater-
nary carbons or mobile segments (including CH3 groups with >50%
efficiency). For both spectra, 2000 scans were collected using a re-
cycle delay of 1 s. The difference of the two spectra was the spec-
trum of immobile CH2 and CH carbons, with a small CH3
contribution (Mao et al., 2007a).
Spectral editing of CH2 signals was achieved by selection of the
three spin coherence of CH2 groups, using a 13C 90° pulse and 1H
0°/180° pulses applied after the first quarter of one rotation period
with MREV-8 decoupling (Mao and Schmidt-Rohr, 2005). A total of
56,728 scans were collected with a 1 s recycle delay and the spin-
ning speed was 5.787 kHz.
For CH (methine) selection, a robust method based on CAH
multiple quantum coherence (Schmidt-Rohr and Mao, 2002) was
used at 4 kHz MAS. CH group multiple quantum coherence was
not dephased by the spin-pair CH dipolar coupling while CH2
coherence was dephased by dipolar coupling of the carbons to
the two protons. The first of a pair of recorded spectra contains sig-
nals of CH, as well as residual quaternary carbon and CH3 peaks
that were removed by taking the difference with a second spec-
trum acquired with the same pulse sequence except for additional
40 ls dipolar dephasing before detection. Each spectrum was
based on 5888 scans with a 1 s recycle delay.
Two dimensional (2D) 1HA13C heteronuclear correlation (HET-
COR) NMR experiments (Mao and Schmidt-Rohr, 2006) were per-
formed at a spinning speed of 6.5 kHz. The scale on which
1
HA13C proximities were probed was chosen by the cross polariza-
tion method and by 1H spin diffusion before cross polarization. Pri-
marily one-bond 1HA13C connectivities were revealed by 0.1 ms of
Lee-Goldburg cross polarization (LG-CP), which suppressed 1HA1H
spin diffusion during CP. A total of 512 scans were collected using a
24 J.R. Helms et al. / Organic Geochemistry 44 (2012) 21–36
1.4 s recycle delay. Prolonging LG-CP to 0.5 ms allowed for mostly
one-and two-bond 1HA13C connectivities. Scans (448) were col-
lected with a 0.8 s recycle delay. A 40 ls dipolar dephasing delay
was inserted to reveal proximities for non-protonated aromatic
carbons to alkyl substituents (Mao et al., 2007a). The number of
scans was 512, with a 1.4 s recycle delay. The correlations of the
carbons to protons within a 4 Å radius were shown with standard
Hartmann–Hahn CP (HH-CP) of 0.5 ms, which allowed for some 1H
spin diffusion. For each t1 increment, 192 scans were collected with
a recycle delay of 1.4 s. We used 96 t1 increments.
Fused aromatic rings were identified by the large number of
carbons located far from the protons on the periphery of the struc-
ture. The signals of these carbons were selected by a long range
recoupled CAH dipolar dephasing technique (Mao and Schmidt-
Rohr, 2003). In short, two 1H 180° pulses per rotation period pre-
vent magic angle spinning from averaging out weak CH dipolar
couplings. After 0.9 ms of recoupled dipolar dephasing time, the
signals of most individual aromatic rings are dephased while those
of fused rings remain at the 95% level. In order to detect non-pro-
tonated carbons with good relative efficiency, direct polarization-
total sideband suppression (DP-TOSS) as described above was used
at a spinning speed of 7 kHz. The ‘‘c-integral’’ was employed to
suppress sidebands up to the fourth order (DeAzevedo et al.,
2000). The 13C 90° and 180° pulse lengths were 4 ls and 8.1 ls,
respectively. The details of this technique have been described
elsewhere (Mao and Schmidt-Rohr, 2003). For each gilsonite spec-
trum, the number of scans and recycle delay were 1024 and 20 s,
respectively.
Two-dimensional 1HA13C wide line separation (WISE) NMR
measurements were performed with 13C decoupling during evolu-
tion (Schmidt-Rohr et al., 1992; Tekely et al., 1993; Mao and
Schmidt-Rohr, 2006). Two different CP times, 0.1 ms or 1 ms, were
used. A dipolar dephasing delay of 40 ls was employed to selec-
tively observe signals of non-protonated and mobile carbons. All
WISE experiments used 512 scans per t1 increment and 0.5 s re-
cycle delay, except the experiment with 0.1 ms CP and 40 ls dipo-
lar dephasing, which used 2000 scans. We used 96 t1 increments of
5 ls.
15
N CP-MAS was conducted with a contact time of 1 ms and
0.5 s recycle delay (262,144 scans). In order to observe non-proton-
ated and mobile nitrogen functional groups, 15N CP-MAS with a
recoupled dipolar dephasing time of 291 ls (Fang and Schmidt-
Rohr, 2009) was performed. A total of 53,760 scans were recorded
with a 0.5 s recycle delay. The spinning speed was 6.5 kHz.
2.2.2. Solution NMR
1
H NMR of gilsonite with Hahn echo was conducted with a 2 s
recycle delay. The interval s in the Hahn echo sequence (p/2–s–
p–s-detection) was varied for a series of experiments. As s in-
creased, the mobile components with long T2 relaxation time sur-
vived, but immobile components with short T2 relaxation time
were dephased. Sixteen scans were recorded for each spectrum.
A 13C solution NMR spectrum was collected consisting of 12,000
scans with a recycle delay of 1 s. Gradient enhanced heteronuclear
multi-quantum correlation (HMQC) NMR (Vuister et al., 1991) was
used to show direct 1HA13C J-couplings. Correlation spectroscopy
(COSY) was used to show 1HA1H interactions across three bonds
(Aue et al., 1976). Total correlation spectroscopy (TOCSY) NMR
(Bax and Davis, 1985) was used to show correlation of proton spins
over a relatively long range (multiple bonds) within a given spin
system.
2.3. Fourier transform ion cyclotron resonance mass spectrometry
Gilsonite was analyzed using electrospray ionization (ESI) in
both positive ionization and negative ionization modes on a 12
Tesla Bruker Apex-Qe mass spectrometer. For positive ionization,
the sample was dissolved in pyridine and diluted in 1:1 tetrahy-
drofuran (THF):methanol (MeOH) with 0.1% trifluoroacetic acid
(TFA). For negative ionization the sample was dissolved in pyridine
and diluted in 1:1 THF:MeOH with 0.1% ammonium hydroxide
(NH4OH). TFA and NH4OH were added as pH adjusting buffers
and were chosen because of their high volatility.
Prior to both analyses the mass analyzer was calibrated using a
mixture of polyethylene glycols (PEG). Positive ionization mode
mass spectra were calibrated internally using a homologous series
of aliphatic amines, while the negative ionization mode was cali-
brated internally using a homologous series of fatty acids (Sleigh-
ter et al., 2008). In both cases the internal calibrants were
naturally occurring components of the gilsonite.
Gilsonite is a complicated mixture of compounds and thus gen-
erates a very complicated spectrum. ‘Data mining’ techniques, as
outlined by several previous studies (Hsu et al., 1994; Hughey
et al., 2001; Kujawinski, 2002; Kim et al., 2003; Herniman et al.,
2005; Hertkorn et al., 2006; Hockaday et al., 2006; Koch and Ditt-
mar, 2006; Sleighter and Hatcher, 2007), were therefore used to
summarize the mass spectral results. Kendrick mass defect analy-
sis (KMDA) (Kendrick, 1963; Grigsby, 1989) was used to identify
prominent homologous series within each spectrum to internally
calibrate the exact m/z assignments. Once the spectra were cali-
brated to within 1.0 ppm error (1 ppm = 0.0001%), relatively
straightforward computational tools were used to obtain exact ele-
mental formula assignments (C, H, N, O, S) for the majority of peaks
in the spectrum (Kim et al., 2006). Elemental atom ratios (e.g. O/C,
H/C, N/C, etc.) were calculated for all of those peaks that were as-
signed formulae. From these, we plotted each peak within van
Krevelen space (Kim et al., 2003), providing graphical representa-
tions that allow a fingerprint characterization of the distribution
of ions generated by ESI. Several chemical indices were calculated,
which indicate the likely structural character of the molecules that
can be inferred from their formulae. These indices include double
bond equivalents (DBE) (Hockaday et al., 2006; Koch and Dittmar,
2006),
1
DBE ¼ 1 þ C  O  S  H ð1Þ
2
condensed aromatic ring structures (CARS) (Hockaday et al., 2006),
DBE
CARS  > 0:7 ð2Þ
C
aromatic index (AI) and modified AI (AImod) (Koch and Dittmar,
2006),
1 þ C  O  S  0:5H
AI  P 0:67;
COSNP
1 þ C  0:50O  S  0:5H
AImod  P 0:5 ð3Þ
C  0:50O  S  N  P
aliphatics (Perdue, 1984),
DBE
aliphatics  < 0:3 ð4Þ
C
and black carbon (BC) (Kim et al., 2004).
BC  H=C < 0:8 and 0:3 < O=C > 0:6 ð5Þ
Kendrick mass defect analysis (KMDA) was used to identify and
enumerate series of structural homologues and commonly ob-
served repeating units within structural compound classes. KMDA
is based on the fact that ions within a homologous series have the
same difference between the nominal mass and Kendrick mass.
The Kendrick mass of the ion was calculated by multiplying the
measured mass of the ion by the quotient of the IUPAC mass and
nominal mass of the repeating unit (Kendrick, 1963; Grigsby,
 J.R. Helms et al. / Organic Geochemistry 44 (2012) 21–36
1989; Hughey et al., 2001; Stenson et al., 2003). It is important to
note that since tandem MS was not used, FT-ICR-MS did not distin-
guish between structural isomers (Stenson, 2008) and ESI shows
considerable selectivity during ionization; thus the chemical indi-
ces and KMD series were used to simply constrain the contents of
the sample and may have given quantitatively misleading values or
reflected isomeric overlap between compound classes.
2.4. Elemental analysis
Triplicate solid gilsonite samples were analyzed for weight per-
cent carbon and nitrogen using a Thermo-Finnegan Flash 1112 ser-
ies elemental analyzer (EA). The EA was calibrated for both carbon
and nitrogen using nicotinamide.
3. Results
13
3.1. Quantitative C NMR spectra
Fig. 1a shows the 13C DP-MAS spectrum obtained at a spinning
speed of 14 kHz, which provided quantitative structural informa-
tion on the whole sample. Spin counting (Smernik and Oades,
2000) yields a 13C observability of 110 ± 15% in DP-MAS NMR,
which confirms that all gilsonite carbon is observed in this spec-
trum. Similar to the NMR spectra of other post-diagenetic organic
matter samples such as kerogen (Werner-Zwanzinger et al., 2005;
Smernik et al., 2006; Mao et al., 2010a), the DP-MAS NMR spec-
trum of gilsonite has mainly two broad bands: the signals of sp3-
hybridized alkyl C from 5–60 ppm and the band of sp2-hybridized
carbons such as aromatics around 100–165 ppm. The O-alkyls
(50–100 ppm) were almost undetectable, as expected for heavily
degraded geological samples. Fig. 1b shows the spectrum of 13C
DP-MAS with recoupled dipolar dephasing of 68 ls, providing
quantitative structural information on non-protonated carbons
and carbons of mobile groups such as CH3. Note that the highest
peak of the sp2 hybridized aromatic band around 100–150 ppm
of Fig. 1b did not match that of Fig. 1a: the former at 133 ppm
was primarily from non-protonated aromatic carbons and the lat-
ter at 128 ppm more attributed to protonated aromatic carbons
(see further discussion below). In the alkyl C region, we observed
two bands from mobile CCH3 around 17 and 22 ppm, and one from
mobile CCH2C groups around 31 ppm. In Fig. 1b, we have
Fig. 1. Quantitative spectra of (a) DP-MAS and (b) DP-MAS with 68 ls recoupled
dipolar dephasing showing non-protonated carbons and mobile segments such as
CH3. Shoulders assigned to aromatic CAO and possible pyridinic CAN are
highlighted in (a), those of non-protonated aromatic C two bonds from O or N in (b).
25
highlighted the spectral region characteristic of non-protonated
aromatic carbons two bonds from O or N.
The low shoulder from 145–165 ppm on the other side of the
aromatic peak, highlighted in Fig. 1a, can be assigned to aromatic
CAO and possibly pyridinic CAN sites. Given that two carbons
would be bonded to each pyridinic N, the spectral area fraction be-
tween 150 and 165 ppm (2%) provides an upper limit of 2%/2 = 0.01
to the ratio of all C to pyridinic N; the likely value is smaller, due to
contributions from aromatic CAO in this spectral range. Given the
3.3:80 = 0.04 C:N ratio of our sample, pyridinic N accounts for less
than 1=4 of all nitrogen. The likely fraction of pyridinic N is smaller,
between 1/5 and 1/10 of all N.
The assignments of signals and the corresponding percentages
of different functional groups are listed in Table 1. Specific assign-
ments are difficult because the bands were broad and overlapping;
however, the interpretation was assisted by spectral editing as de-
scribed below and in Table 1.
3.2. NMR spectral editing
Fig. 2 shows a series of 13C CP-MAS NMR spectra acquired with
suitably designed radio frequency pulse sequences to selectively
detect signals from specific types of chemical groups. Fig. 2a is
the 13C CP-TOSS spectrum, which shows qualitative structural
information and was used as the reference for the selected sub-
spectra (Fig. 2b–g). The corresponding 13C CP-TOSS spectrum after
40 ls of dipolar dephasing (Fig. 2b) exhibited only signals of non-
protonated carbons and carbons of mobile groups, including rotat-
ing CH3 groups, which have a reduced CAH dipolar coupling due to
their fast motions. This spectrum provided similar structural infor-
mation as in Fig. 1b but with better sensitivity for some peaks. In
particular, comparison of Fig. 2a and b (as well as Fig. 2d) showed
that the possible pyridine CAN carbons resonating near 160 ppm
are mostly not protonated. As in the spectra of Fig. 1, the maxima
of the aromatic band in Fig. 2a and b did not align exactly. Two
bands from mobile CCH3 were observed around 17 and 22 ppm,
and one from mobile CCH2 groups around 31 ppm. Some non-pro-
tonated quaternary carbons resonating around 40 ppm were also
present.
The 13C CP-TOSS spectrum after a 13C CSA filter of 47 ls, which
exhibited only sp3-hybridized carbon signals, is displayed in
Fig. 2c. This technique separates signals of anomerics (OACAO)
from overlapping bands of aromatics (Mao et al., 2007a, 2007b,
2007c). No contribution from anomeric carbon was observed for
the gilsonite sample, indicating that sugar rings were insignificant.
Fig. 2c indicates that the 13C signals above 90 ppm all belonged to
sp2-hybridized carbons.
Using a short CP time of 50 ls, primarily protonated carbons
were selected; while mobile CH2 and CH3 resonances were signif-
icantly suppressed because of their low CP efficiencies (Fig. 2d). In
order to suppress residual signals of carbons two bonds from 1H, a
spectrum after dipolar dephasing of 40 ls was also acquired. The
short CP spectrum subtracted by the spectrum with a double filter
of short CP and dipolar dephasing showed primarily signals of
immobile protonated carbons. There was a small signal from pro-
tonated aromatics or C@CH2 around 123 ppm.
Fig. 2e shows the CH2 only spectrum of gilsonite obtained by
three spin coherence selection (Mao and Schmidt-Rohr, 2005).
CH2 contributed a small amount of signal in the sp2 hybridized re-
gion, observed around 123 ppm; this band was assigned to exo-
methylene –C@CH2. Exomethylene was previously identified in
natural resins (Anderson et al., 1992; Clifford et al., 1999). Signifi-
cant CCH2C (non-polar alkyl) signals were dominant in this selec-
tive spectrum. Clearly, most of the carbons that resonated in the
alkyl region belonged to CCH2C groups. Fig. 2f shows the CH only
spectrum, acquired using dipolar distortionless enhancement by
26 J.R. Helms et al. / Organic Geochemistry 44 (2012) 21–36

Table 1
13 13 13
Quantitative integration results from C DP-MAS, C DP-MAS with recoupled dipolar dephasing, C CP-TOSS, and spectral editing.

Integration range (ppm) Predominant structural moieties DP-MAS (%) Refined using DP dipolar dephasing and spectral editing
5–27 CH3 + CH2 25
27–60 Alkyl CH2 and CH 45 38% CH2
6.2% CH (CH selection)
1.1% CQuaternary (dipolar dephasing)
60–100 Alkyl OC 0.3 Non-protonated C (dipolar dephasing)
100–165 Aromatic + alkene 29 19% Non-protonated aromatic CAC (dipolar dephasing)
5.7% Aromatic CH
2.3% Exomethylene @CH2
2.2% (150–165 ppm) Aromatic CAO and pyridine CAN
165–185 Amides, carboxylic acids, esters (N/OAC@O) 0.4 Non-protonated C (dipolar dephasing)
185–220 Ketones, quinones (C@O) 0.4 Non-protonated C (dipolar dephasing)

Fig. 2g shows a 13C DP-MAS spectrum with a short recycle delay,

which accentuated the signal from mobile alkyl moieties. The short
recycle delay selected signals with short 13C T1 (spin–lattice relax-
ation) times. Rigid segments have relatively long 13C T1 values
compared with mobile ones because mobility drives relaxation.
The spectrum of the mobile segments only exhibited signals of
sp3-hybridized carbons, indicating that many of the alkyls were
more mobile than the sp2-hybridized carbons. This conclusion
was consistent with the result of the dipolar dephased spectrum
(Fig. 1b); signals of highly mobile alkyls were also retained in this
spectrum.

3.3. Structural information from short range 1HA13C HETCOR NMR

Fig. 3 shows several 2D 1HA13C HETCOR spectra acquired under

Fig. 2. Spectral editing for identification of functional groups in 13C NMR. (a) CP-
TOSS spectrum for reference. (b) Dipolar-dephased CP-TOSS spectrum showing
non-protonated carbons and mobile segments such as CH3. (c) Selection of sp3-
hybridized carbon signals by a chemical shift anisotropy filter. (d) Selection of
immobile CH and CH2 signals with residual CH3. (e) Signals of immobile CH2 groups
selected based on three-spin coherence. (f) CH-only spectrum (by dipolar DEPT),
and (g) DP spectrum with a short, 1.5 s, recycle delay. Vertical dashed lines at 134,
127, 122, 31, 22, and 15 ppm mark important peaks or shoulders in the spectra.
polarization transfer (DEPT) (Schmidt-Rohr and Mao, 2002); small
CCH and aromatic CH signals were observed. Both bands were
broad, indicating their wide range of chemical environments. We
have highlighted the ppm region where NCH could resonate
(Fig. 2f).
various conditions. 1H slices at various 13C chemical shifts were ex-
tracted to observe the correlations more clearly. Fig. 3a is the 2D
HETCOR spectrum with LG-CP of 0.1 ms, in which primarily one-
bond connectivities were observed. In order to assign the 13C signal
between 51 and 61 ppm and at 30 ppm, we extracted 1H slices
from this 2D spectrum (Fig. 3b). We integrated over 13C chemical
shifts between 51 and 61 ppm, associated with the highest peak
in the CH-only spectrum, and took a slice at the 30 ppm resonance
position of CCH2C groups for reference. The 1H spectrum associated
with the 51–61 ppm 13C resonance showed various alkyl contribu-
tions, including a shoulder near 4 ppm that was attributed to NCH
groups (e.g. in peptides). Nevertheless, this component was not
dominant, which indicated that peptides were not the main con-
tributors to this carbon signal.
In order to observe correlations between aromatic C and alkyl H,
which are separated by at least two bonds, we recorded a 2D HET-
COR spectrum with a longer 0.5 ms LGCP contact time (Fig. 3c). It
mostly showed one-and two-bond 1HA13C connectivities, i.e. the
peaks belonged to 13C and 1H nuclei within 0.25 nm of one another.
1
HA13C 2D HETCOR successfully offered better separation between
two types of aromatic moieties, namely (i) non-protonated aro-
matics bonded to alkyls at 1H chemical shift of 0.8–1.2 ppm and
13
C chemical shift of 133 ppm, and (ii) primarily protonated aro-
matics or aromatics closer to aromatic protons than alkyl protons
with a 1H chemical shift of 7.3 ppm and a 13C chemical shift of
123 ppm. The assignment of the two types of aromatic signals
was also based on spectral editing results described above. 1H
slices extracted at the alkyl 13C chemical shifts of 17, 31 and
40 ppm showed correlations primarily with their own alkyl pro-
tons; 1H slices extracted near the aromatic 13C chemical shifts of
127 and 135 ppm showed correlations with alkyl protons, indicat-
ing that these were closely associated with alkyl side chains
(Fig. 3d). Fig. 3e displays the 2D HETCOR spectrum with 0.5 ms
LGCP and 40 ls dipolar dephasing, which provided spectra selec-
tively of the non-protonated (or mobile carbons). The 1H spectrum
associated with 13C chemical shifts of 125–145 ppm (Fig. 3f)
 J.R. Helms et al. / Organic Geochemistry 44 (2012) 21–36
Fig. 3. 1HA13C 2D HETCOR spectra of gilsonite (a) with LG-CP of 0.1 ms whose
extracted proton slices between 51–61 ppm and at 30 ppm are listed in (b); (c) with
LG-CP of 0.5 ms and its extracted proton slices at 17, 31, 40, 127, and 135 ppm in
(d); (e) with LG-CP of 0.5 ms and dipolar dephasing of 40 ls and extracted proton
slices between 125–145 ppm in (f); and (g) with HH-CP of 0.5 ms and its slices
between 115–142 ppm in (h).
showed a dominant proton contribution from alkyl H, rather than
from aromatic H, indicating that these aromatics were highly
substituted. This was further confirmed by a 2D HETCOR spectrum
with 0.5 ms Hartmann–Hahn CP with an effective spin diffusion
time of 0.125 ms (Fig. 3g); the dominant proton contribution was
also from alkyl H. Fig. 3h shows the proton cross section between
115–142 ppm. Thus, by using 1HA13C 2D HETCOR NMR, we have
shown that a significant fraction of non-protonated aromatic car-
bon signal is from substituted ring carbons connected to alkyls
(rather than from fused aromatic rings, the other possible source
of non-protonated carbon).
27
Fig. 4. (a) Series of DP-TOSS spectra of gilsonite after recoupled dipolar dephasing
of the indicated durations. (b) Corresponding long-range dipolar dephasing curves
for non-protonated aromatic carbons in gilsonite, integrated between 108 and
145 ppm. S/S0 refers to the aromatic signal from the dipolar dephased spectrum as a
fraction of the signal in the reference spectrum. Data for wood charcoal, lignin, and
type II kerogen (Mao and Schmidt-Rohr, 2003; Mao et al., 2010a) are shown for
reference. Filled diamonds: gilsonite; open hexagons: wood char; open triangles:
lignin and open squares: type II kerogen.
3.4. Information on fused aromatic rings based on 1HA13C recoupled
long range dipolar dephasing NMR
Fig. 4 shows the use of the 1HA13C recoupled long range dipolar
dephasing technique (Mao and Schmidt-Rohr, 2003) to investigate
the presence of fused ring carbons in gilsonite. Fig. 4a shows a
series of 13C direct polarization/total suppression of sidebands
(DP-TOSS) NMR spectra of gilsonite with increasing recoupled
dephasing times. In order to estimate the aromatic cluster size,
we compared the dephasing curve for the aromatic region (108–
145 ppm) of gilsonite with published data of lignin (Brewer et al.,
2009), an immature type II kerogen (Mao et al., 2010a) and char-
coal (Mao and Schmidt-Rohr, 2003) (Fig. 4b). The dephasing of gil-
sonite was faster than that of kerogen, which has been estimated to
contain an average of six rings per fused aromatic structure (Mao
et al., 2010a) but slower than that of lignin, which contains mostly
single benzene rings. This indicated small clusters of perhaps 4
fused aromatic rings in gilsonite. Three-and four-ring structures
in petroleum deposits have been attributed to aromatization of dit-
erpenoid, triterpenoid, sterol, hopane and similar moieties (Killops
and Killops, 2005).
3.5. 2D WISE NMR
1
HA13C WISE NMR can identify mobile segments, in terms of
motional narrowing of 1H wideline spectra. In gilsonite, endgroups
28 J.R. Helms et al. / Organic Geochemistry 44 (2012) 21–36
of alkyl chains are mobile. WISE NMR helped distinguish mobile
and non-protonated carbons, both of which had weak CAH dipolar
couplings and contributed to 13C spectra after dipolar dephasing.
Further, combined with 1H spin diffusion it provided information
on the distance of mobile endgroups from various spectrally re-
solved segments, in particular aromatic rings. Fig. 5 shows the 1H
wide-line patterns extracted from WISE spectra at various 13C peak
positions. Two different CP times, 0.1 and 1 ms, were used in order
to vary the spin diffusion times and also detectability of protonated
and non-protonated carbons. Dipolar dephasing of 40 ls was in-
serted to selectively observe the signals of mobile or non-proton-
ated carbons.
Fig. 5a shows the 1H wide-line spectra at 13C chemical shift of
30 ppm (thick line) and between 44 and 53 ppm (thin line), both
with 0.1 ms CP and 40 ls dipolar dephasing. The signal at
30 ppm is attributed to mobile CCH2C groups, which survived after
dipolar dephasing; the high mobility reduced HAH dipolar cou-
plings and resulted in a narrower 1H spectrum. The broad 1H spec-
trum associated with the carbons resonating between 44 and
53 ppm indicates large proton–proton dipolar couplings. There-
fore, the weak CAH dipolar coupling associated with these carbons
cannot be attributed to motional narrowing and must be explained
by assigning the 44–53 ppm signal after dipolar dephasing to non-
protonated (quaternary) carbon.
WISE NMR can differentiate overlapping contributions from ri-
gid and mobile segments. Fig. 5b shows 1H wide line spectra at
30 ppm with 0.1 ms CP (thick line), and the same after 40 ls dipo-
Fig. 5. 1H wideline spectra extracted from 2D 1HA13C wideline separation (WISE)
NMR at 13C chemical shift of (a) 53–44 ppm (thin line) and 30 ppm (thick line), both
with 0.1 ms CP and 40 ls dipolar dephasing delay; (b) 30 ppm (thick line) and 44–
53 ppm (dashed line) with 0.1 ms CP, and 30 ppm with 0.1 ms CP and 40 ls dipolar
dephasing delay (thin line; same data as that in (a) but with less line broadening);
(c) 39 ppm (thick line) and 44–53 ppm (dashed line) with 0.1 ms CP, and 39 ppm
with 0.1 ms CP and 40 ls dipolar dephasing delay (thin line); and (d) 140–
123 ppm, acquired under the conditions of 0.1 ms CP (thick line), 1 ms CP (thin
line), and 1 ms CP combined with 40 ls dipolar dephasing delay (dashed line).
lar dephasing (thin line). The difference between them shows that
the lineshape is a superposition of a motionally narrowed and a
more or less rigid component. For reference, the broad lineshape
of the rigid CH and CH2 groups that resonated between 53 and
44 ppm (with 0.1 ms CP) is also shown (dashed line). Fig. 5c shows
the corresponding set of 1H wide line spectra associated with 13C
signal at 39 ppm (thick line). Again, differences between the spec-
tra without (thick line) and with dipolar dephasing (thin line)
prove that there is a superposition of mobile and rigid components,
but with a much smaller contribution from the mobile segments.
Finally, Fig. 5d displays 1H wide line spectra of carbons resonat-
ing around 140–123 ppm with 0.1 ms CP (thick line), 1 ms CP (thin
line), and 1 ms CP and 40 ls dipolar dephasing (dashed line). The
sharp peaks of mobile alkyl segments became visible, though with
low intensity, for the non-protonated aromatic carbons at long
cross polarization times of 1 ms. This shows that aromatic cores
and mobile alkyl segments are within approximately 2 nm of one
another.
15
3.6. N NMR, with recoupled dipolar dephasing
Fig. 6a shows the 15N CP-MAS spectrum of gilsonite without
dipolar dephasing, while Fig. 6b shows the 15N CP-MAS spectrum
of gilsonite with 291 ls recoupled dipolar dephasing. The 15N CP-
MAS spectrum shows only one broad band ranging from 140–
110 ppm (Fig. 6a). Protonated pyrrole N (three bonding partners)
in five-membered rings resonates between 160 and 130 ppm in
15
N spectra, overlapping with amide N, e.g. in peptides, that reso-
nates between 140 and 110 ppm (Thorn et al., 1996). Resonances
of pyrrolic and amide N overlap heavily, but based on the very
low (<1%) amide signal in the solid state 13C NMR spectra, we
can attribute the majority of the signal observed here to pyrrolic
N. Significant 15N NMR signals of amines, which resonate at
<80 ppm, are not observed. No significant signals were detected
in the 15N CP-MAS spectrum after recoupled dipolar dephasing
(Fig. 6b), indicating that all of the detectable N forms are proton-
ated, consistent with pyrrolic NAH groups. Pyridinic N, which is
not protonated and has unfavorable chemical shift positions and
anisotropies, is not detected efficiently in these experiments. From
13
C NMR, we estimated its fraction to be between 10% and 25% of
all N (see above).
3.7. Solution NMR
Fig. 7 shows 1H NMR spectra of gilsonite after various T2 filter
times ranging from 0.2–60 ms. The interval t in the Hahn echo se-
quence (p/2–s–p–s-detection) was adjusted for each experiment.
As s increases, the signals of highly mobile components with long
T2 relaxation time will survive, but those of less mobile
Fig. 6. 15N CP-MAS spectra (a) without dipolar dephasing and (b) with 291 ls
recoupled 1HA15N dipolar dephasing, which suppresses the signals of protonated
nitrogen.
 J.R. Helms et al. / Organic Geochemistry 44 (2012) 21–36 29

On the basis of these data and those of Table 1, the carbon fractions
listed in Table 3 were obtained.
Fig. 8 shows the alkyl portion of the 13C solution NMR spectrum
of gilsonite, which matched the spectrum published by Clark et al.
(1983). We have been able to assign the 10 highest peaks to three
specific structures, namely unbranched chain ends, chain segments
with a methyl branch and chain ends with two methyl groups (see
Fig. 8a and b). Comparison with the solid state 13C NMR spectrum
in Fig. 1 and the CH only spectrum in Fig. 2f indicates that a large
fraction of the alkyl signals were invisible in 13C solution NMR. For
instance, the solid state NMR spectra showed significant intensity
above 40 ppm, which was invisible in solution NMR, probably
due to large line widths that resulted from limited mobility.
We also recorded two dimensional COSY and TOCSY 1H NMR as
well as HMQC 1HA13C NMR spectra (not shown). The 1H spectra
confirmed that various alkyl species are chemically bonded. How-
ever, the broad spectral components that account for most of the
signal are not visible and alkyl-aromatic cross peaks were not seen
in these spectra because the scalar couplings between them are too
weak.

3.8. Elemental analysis

The gilsonite sample contained 80.0 ± 0.7% carbon, and

3.3 ± 0.1% nitrogen by weight. For comparison, average values from

Table 3
Functional group quantification from combined NMR data of Tables 1 and 2.

Functional group C percentage NMR

Alkyl CH3 13
Alkyl CH2 50
Fig. 7. 1H solution NMR spectra with various T2 filter times ranging from 0.2–60 ms. Alkyl CH 7
The sharp peak at 7.3 ppm was caused by proton exchange with the solvent, @CH2 2
deuterated chloroform. Aromatic CH 6
Aromatic CAC (non-protonated) 19
Aromatic CAO 2.2
NC@O 0.4
components with short T2 relaxation time will dephase. Clearly, the C@O 0.4
sharp peaks at 1.2 and 0.8 ppm belonged to highly mobile CH2
and CH3, respectively. The broad components at around 2–3 ppm
were assigned to CH2 or CH directly attached to aromatic rings,
which have reduced flexibility. The significant intensity of this
band (17% of all H) provides evidence of abundant alkyl-aromatic
connections. The resonances of NH and CH protons of the pyrrole
and other aromatic rings overlapped at around 7–8 ppm. The sharp
peak at 7.3 ppm was due to exchanged protons on the deuterated
chloroform solvent.
The relative 1H NMR signal areas and assignments are listed in
Table 2. We also calculated the corresponding carbon signal areas.
First, the signal intensity of a CHn group was divided by n, the num-
ber of protons per carbon in the group. The resulting numbers were
then normalized to a total of 100% and finally corrected for the
fraction fnp of non-protonated carbons, which are invisible in 1H
NMR but easily determined by 13C NMR, by dividing each entry
by (1 + fnp). The results are given in the last column of Table 2.

Table 2
Signal intensities in solution 1H NMR, and corresponding carbon fractions, given the
fraction of non-protonated carbons from 13C NMR.
1
ppm Functional group H NMR intensity (%) Corresponding
C fraction (%)
0–1 Alkyl CH3 26 13
1–2 Alkyl CH1.8 52 43
2–5 CH1.8 bonded to aryl 17 14 Fig. 8. (a) Alkyl region of the 13C solution NMR spectrum of gilsonite. Assignments
6–9 Aromatic CH + NH 4.5 7 of the 10 highest peaks to three structures, shown in (b), are indicated above the
– Non-protonated C – 23 spectrum. The chemical shifts for these structures were calculated using the ACD
Lab v.12.0 software.
30 J.R. Helms et al. / Organic Geochemistry 44 (2012) 21–36

Table 4 weight for the combined positive and negative mode spectra (with
Comparison of elemental analysis with average molecular formulas determined from redundant peaks removed) was 533 Da, while the peak intensity
NMR and electrospray Fourier transform ion cyclotron mass spectra (ESI-FT-ICR-MS).
Carbon and nitrogen were measured using a Thermo-Finnegan Flash 1112 series
weighted average molecular weight was 355 Da.
elemental analyzer. N/A: Not measured. The large numbers of peaks observed for gilsonite (Fig. 9) com-
plicated our interpretation of the mass spectral data. However, we
Elemental Solid state ESI-FT-ICR-MS ESI-FT-ICR-MS
analysisa NMR number intensity
have summarized our data and used some relatively simple visual-
weighted weighted ization techniques to characterize the sample. The van Krevelen
a
Carbon wt% 80 (84.0) 83% 79.61 80.10
plot in Fig. 10a was dominated by hydrocarbon peaks (low O/C)
a
Hydrogen wt% (10.67) 11% 10.57 11.15 and lipids (H/C = 1.7–2.25; O/C = 0.02–0.18), as well as a smaller
a
Nitrogen wt% 3.3 (2.40) 3% 3.11 3.25 number of peaks in the regions associated with peptides and pro-
a
Sulfur wt% (0.50) N/A 0.88 1.04 teins (H/C = 1.5–2.0; O/C = 0.2–0.5) (Kujawinski, 2002; Sleighter
a
Oxygen wt% (3.74) 3.2% 5.83 4.32
a
and Hatcher, 2007). Because all peaks in the spectra are singly
H/C atomic ratio (1.51) 1.49 1.59 1.67
charged and have low N/C ratios (<0.2) it is likely that the few
a
Values in parentheses are averages based on literature values (Bell and Hunt, peaks in the protein region are rather degraded proteins, degraded
1963; Clark et al., 1983; Jacob, 1989).
petroporphyrins or simply nitrogen heterocycles, amides, or
amines (Fig. 10b).
Both ESI modes were selective for substituted hydrocarbons. All
previous studies (Bell and Hunt, 1963; Clark et al., 1983; Jacob, of the assigned formulae included at least one heteroatom: N, O or
1989) are reported in Table 4. This table also compiles elemental S. The average (mode) formula included two heteroatoms (mean
compositions calculated from NMR and MS. The NMR elemental value = 3.3 heteroatoms, median value = 3, and no formulae con-
data were calculated from the fractions and CHNO compositions tained more than 16). The median formula (323 of the formulae
of the functional groups in Table 3 (Mao et al., 2000). accounting for 2.7% of the total peak intensity) contained two oxy-
gens and one nitrogen (mean = 1.17 N, 1.93 O, 0.145 S; maxi-
3.9. FT-ICR-MS mum = 8 N, 10 O, 3 S). There was a clear bias in favor of oxygen
containing moieties in the negative ionization mode (Fig. 10; Ta-
Negative ionization mode ESI produced a spectrum (Fig. 9a) ble 4). We found that the most populated group of peaks and the
consisting of 2576 resolved peaks (m/z: 203–730; average highest intensity peaks in negative ionization mode contained
MW = 458), 1914 of which were assigned unique molecular formu- two oxygen atoms and likely represented carboxylic acids, which
lae (C, H, N, O, S; error < 1 ppm). Positive ionization mode ESI gen- were readily ionized by ESI, but were quantitatively negligible in
erated a spectrum (Fig. 9b) of 3663 peaks (m/z: 210–908; average the sample according to NMR. Sorting the formulae (from both
Mw = 573), 2199 of which were assigned unique molecular formu-
lae (C, H, N, O, S; error < 1 ppm). Two hundred twenty-two of the
peaks with assigned formulae (5%) were present in both positive
and negative ionization spectra, indicating that the external and
internal calibrations for positive and negative mode were both suc-
cessful and mutually compatible. The number average molecular

Fig. 9. (a) Positive and (b) negative electrospray ionization FT-ICR mass spectra for Fig. 10. (a) Van Krevelen diagram and (b) 3-D van Krevelen diagram for formula
gilsonite. Inset spectra are shown with enhanced y-axis scales. assignments from positive (black) and negative (gray) electrospray ionization.
 J.R. Helms et al. / Organic Geochemistry 44 (2012) 21–36 31

identified porphyrin molecules. However, 115 ions were observed

Fig. 11. Incidence of heteroatoms expressed as (a) number of formulae containing a
given number of heteroatoms (O, S, and N), and (b) percentage of total ESI-FT-ICR-
MS peak intensity represented by formulae with a given number of heteroatoms.
positive and negative modes) according to the number of each het-
eroatom present (Fig. 11) shows that most of the ions contained
four or fewer oxygen atoms, three or fewer nitrogen atoms and
one or no sulfur atoms. While this suggests that intensity is not
an accurate indicator of abundance within the sample, because it
is strongly impacted by variations in ionization efficiency, the bulk
elemental composition of gilsonite based on the molecular formu-
lae calculated for the ESI ionizable fraction of gilsonite accurately
reflects the bulk elemental composition determined by combus-
tion elemental analysis (Table 4).
No peaks were observed that were clearly identifiable as previ-
ously reported petroporphyrins (Quirke and Maxwell, 1980; Haji-
brahim et al., 1981; Gill et al., 1985; Clezy et al., 1989; Qian
et al., 2008; McKenna et al., 2009). Porphyrins were probably not
efficiently ionized by ESI within the complex matrix of gilsonite.
Likewise, expanding the molecular formula calculations to include
vanadium, cobalt, magnesium, and nickel yielded no previously
that had four or more nitrogen atoms, and several of these con-
tained at least 20 carbon atoms and zero, one or two metal ion(s).
These molecules likely represent partially degraded porphyrins
that have retained the tetrapyrrole metal binding core of the origi-
nal molecule. Pyrrolic compounds with alkyl or olefinic side chains
could account for most of the nitrogen containing molecules; how-
ever, they were not readily distinguishable from amines by mass
spectrometry alone and may represent structural isomers or mix-
tures of isomers.
Possible lipid biomarkers identified within ESI-FT-ICR-MS spec-
tra were suggestive of both algae and cyanobacteria. Polyunsatu-
rated fatty acids including chloroplast fatty acids were identified,
as well as some saturated and mono-unsaturated fatty acids typi-
cal of bacteria and cyanobacteria (Kenyon, 1972; Canuel et al.,
1995). Other potential biomarker ions found within the ICR-MS
spectra included phylloquinone (vitamin K; C31H46O2), brassicas-
terol (indicative of algae; C28H46O), and testosterone (indicates
bacteria or animals; C19H28O2).
Indices calculated based on the molecular formulae determined
for the MS peaks (Eqs. (1)–(5); Table 5) suggest that the predomi-
nant structural features of ionized molecules present in gilsonite
are aliphatic hydrocarbons. Nearly 65% of formulae, representing
more than 77% of the total peak intensity, satisfy the definition
of the aliphatic index. In fact, KMD analysis (Table 6) suggests that
9% of resolved peaks were part of a CH2 series and therefore con-
tained some aliphatic character while slightly less than 0.1% were
part of a CH series. The CH KMD series could have included tertiary
aliphatic carbons, as well as protonated aromatic carbons. The per-
centage of peaks identified as containing aromatic rings was be-
tween 0.08% and 2.5%, depending on whether the AI > 0.67 or the
less conservative AImod > 0.5 criterion was used. All three formulae
identified as CARS (0.08%) also satisfy the criterion for AImod > 0.5
and BC simultaneously. No formula indices were identified for
32% of formulae. As previously mentioned, oxygen containing com-
pounds were heavily selected during negative mode ESI and there-
fore COO, OCH2, H2O and O KMDA series make up a much larger
fraction of the analytical window than they represented as a frac-
tion of the gilsonite sample.
4. Discussion
4.1. Composition from NMR, FT-ICR-MS and elemental analysis
Tables 1–3 indicate that solid state NMR provided non-
destructive, comprehensive, quantitative structural information

Table 5
Summary of peaks conforming to 5 formula indices: condensed aromatic ring structures (CARS), aromatic index (AI), aliphatic, and black carbon.

CARS (Hockaday AI > 0.67 (Koch and Modified AI > 0.5 (Koch Aliphatic Black carbon
et al., 2006) Dittmar, 2006) and Dittmar, 2006) (Perdue, 1984) (Kim et al., 2004)
No. of peaks 3 3 3 2505 5
% of peaks 0.08 0.08 0.08 64 0.13
% Peak height 0.04 0.04 0.04 78 0.07
Number ave. MW 549 549 549 540 507
Height ave. MW 518 518 518 505 483

Table 6
Homologous series identified by Kendrick mass defect analysis (KMDA).

CH2 CH OCH2 COO H2 H2O O

# of series 343 3 956 633 591 1020 993
# of peaks in series 3649 6 2539 1356 3514 2624 2659
% of total peaks 8.82 0.08 24.57 16.27 15.19 26.21 25.52
# of series with >4 peaks 208 0 64 2 331 33 68
Average # of peaks per series 11 2 3 2 6 3 3
32 J.R. Helms et al. / Organic Geochemistry 44 (2012) 21–36
on gilsonite whereas 1H solution NMR had severe limitations. 13C
solid state NMR showed that 70% of carbon nuclei in gilsonite are
in alkyl groups and 27% in aromatics. By contrast, 1H solution
NMR provided little useful aromaticity information, since most
aromatic carbons in this sample are not protonated. The aromatics
also did not show up well in solution 13C NMR spectra, due to their
low mobility and resulting large line widths. Standard 1H solution
NMR could not determine the CH2 to CH ratio in gilsonite, which
was reliably estimated from solid state NMR with spectral editing.
On the other hand, the 1H NMR spectrum gave the ratio of CH3 to
CH2, which cannot be obtained from solid state NMR in gilsonite
since many CH2 groups have high mobilities similar to those of
CH3 groups. The good agreement of elemental composition be-
tween elemental analysis and atom ratios calculated from NMR
data confirms the reliability of NMR (Table 4).
Mass spectral data based on ESI for gilsonite were specific for the
polar (ionic) fraction of gilsonite, but in many respects confirmed
key NMR results. The aliphatic chemical index accounted for 64%
of assigned formulas or 78% of the total peak intensity (Table 5)
and 9% of assigned formulae were part of a CH2 homologous series
Table 6). Solid state NMR showed that aromatic CAO accounts for
most of the organically bound oxygen, which is only 2.2% of C (Ta-
ble 1). Both NMR and MS data showed that gilsonite is primarily ali-
phatic. Although the MS data suggests organic acids are present, the
NMR results show that carboxylic carbons are extremely minor
constituents of gilsonite (less than 0.5%; Table 1), highlighting the
need for a more appropriate and representative ionization method
for quantitative characterization of this type of sample, but con-
firming the conclusions of Klein et al. (2006) that indicated that
the polar fraction of petroleum samples can be successfully charac-
terized using ESI-FT-ICR-MS with no pre-infusion fractionation.
Despite the presence of mass spectral peaks with H/C and O/C
ratios typically associated with protein (Fig. 10a; Sleighter and
Hatcher, 2007), we found no quantitative NMR evidence for their
being present in the sample. If protein were present in 13C NMR
detectable quantities, it would have shown a peak near 172 ppm
due to NAC@O, which was not observed significantly. 15N and
13
C NMR indicated that much of the organic nitrogen is present
as five-membered pyrrolic rings (Fig. 6; Thorn et al., 1996). This
is consistent with the molecular formula assignments for nitrogen
containing peaks. Chlorophylls and their anoxic degradation prod-
ucts consist largely of macrocyclic moieties containing up to four
pyrrolic rings. The fact that nearly all of the FT-ICR-MS formulae
in the three-dimensional van Krevelen diagram (Fig. 10b) fell along
rays which intersect at (N/C = 0, O/C = 0, H/C = 2), suggests
that gilsonite consists largely of a complex array of intermediates
of multiple diagenetic/catagenetic processes with chemically and
structurally varied starting material (biomolecules) and a far less
diverse end point (primarily aliphatic hydrocarbons).
If we consider that elemental analysis (Table 4) is suggestive of
the molecular formula of some arbitrarily ‘average’ gilsonite mole-
cule, we calculate a formula of C40H61N1O1 with an additional sul-
fur in 10% of molecules and an additional oxygen in 35% of
molecules. This gives a DBE of approximately 9, suggesting that
no more than two aromatic ring structures are expected per mole-
cule (benzene DBE = 4). This also yields a molecular weight of
571.5 Da, which is consistent with the average MW determined
by ESI-MS, but well below the molecular weight reported in Dickie
and Yen (1967) for gilsonite and other bitumens. Similarly, the
median C, H, N, O and S values from the MS data yield a ‘typical’
formula of C34H54N1O2.
4.2. Pyrrole rings
Aromatic carbons in gilsonite can be part of five-membered
(mostly pyrrole) or ‘‘regular’’ six-membered aromatic rings. Based
on the C/N ratio and the aromatic carbon fraction determined by
quantitative 13C NMR, we determined an upper limit to the abun-
dance of pyrrole rings. There are 3.5 N per 100 C in gilsonite and for
each N in a pyrrole ring, there are four aromatic carbons. Thus,
there are no more than 3.5  4 = 14 pyrrole carbons per 100 C
(14%). Given the 27% aromatic fraction, at least 27–14 = 13% of car-
bon are in non-pyrrolic aromatics.
The preceding analysis gives only an upper limit (14%) for the
pyrrolic carbon fraction. The fraction of carbons that are in fact
pyrrolic is estimated based on the signal of non-protonated C at
<133 ppm. Aromatic carbons resonate in this range only if they
are at a two-bond distance from N or O (and not bonded to O them-
selves) (Bovey et al., 1988). Thus, the signal of non-protonated car-
bons at <133 ppm (11%, shaded in Fig. 1b) is due to C substituted
phenols, pyridines and pyrroles. There are two such carbons per
heteroatom and consequently the contributions from phenols to
the non-protonated carbon signal at <130 ppm must be twice that
of the aromatic CAO and pyridinic CAN observed between 150 and
165 ppm (2.2%, shaded in Fig. 1a). Thus, the pyrrolic carbons two
bonds from N accounts for approximately 11–4.4% = 7% of all C.
Therefore, the pyrrolic carbons accounted for 2  7% = 14% of all
C, and other aromatics for 27–14 = 13%. Further, 14% C in pyrrole
rings corresponds to 14/4 = 3.5 N per 100 C. This means that nearly
all the N is in pyrrole rings.
4.3. Highly substituted aromatic rings: refractory organic matter
The aromatic rings in gilsonite are highly substituted by alkyls.
Three independent lines of evidence supported this conclusion.
First, quantitative DP-MAS and DP-MAS results (Fig. 1; Table 1)
indicated that 77% aromatics were non-protonated and long range
dipolar dephasing data (Fig. 4) clearly showed that the aromatics in
gilsonite did not form large clusters of fused aromatic rings. Solu-
tion and solid state NMR agree that aromatic H only accounts for a
small fraction (5%) of total protons, see Tables 1–3. Specifically,
Fig. 1 showed only minor signals from aromatic CAH two bonds
from N, which typically resonate at <115 ppm. This demonstrates
that the carbons two bonds from N in the pyrrole ring are mostly
non-protonated and likely substituted by alkyl segments. Second,
1
HA13C 2D HETCOR with 40 ls dipolar dephasing (Fig. 3) showed
that the protons near non-protonated aromatics are mostly in alkyl
groups. Third, the broad band of 1H solution NMR around 2–3 ppm
could be attributed to CH2 or CH directly bonded to aromatic rings,
which accounts for 1/5 of all alkyl groups, providing further evi-
dence of abundant alkyl-aromatic connections.
Our previous studies of type II kerogen (Mao et al., 2010a) and
coal (Mao et al., 2010b) indicated that most of the aromatics in
these geologically preserved organic matter samples are non-pro-
tonated and in moderately sized clusters of fused rings. Coupled
with the results in the present study, we hypothesize that non-pro-
tonated aromatics, either in the form of fused rings or highly
substituted rings, are refractory components in nature. Further-
more, our study of a fulvic acid from Antarctica showed that dia-
genesis resulted in a significant fraction of non-protonated alkyl
and aromatic carbons (Mao et al., 2007a), consistent with our
hypothesis here. Therefore, we propose that non-protonated car-
bons can be used as a humification index.
4.4. Structural model
Our results showed that gilsonite bitumen is composed of pyr-
role and other aromatic rings highly substituted with alkyl chains.
The 1H solution NMR indicated that approximately 14/70 = 20% of
the alkyl groups (1 in 5) are bonded to aromatic rings (Table 2). If
most alkyl chains link aromatic rings (Figs. 3 and 5), the linker
between the two rings consists of 2  5 = 10 alkyl carbons on
 J.R. Helms et al. / Organic Geochemistry 44 (2012) 21–36
Fig. 12. Model of the main components of gilsonite structure. A typical ‘‘monomer unit’’ containing one aromatic ring and 21 carbons in total is outlined by the dashed ellipse.
Gilsonite exists as a highly diverse mixture of molecules of various size and structure. The model structure shown here represents the high molecular weight fraction, while
smaller fragments are also common.
average. Also, we concluded that 14/27 or about 50% of all aromatic
carbons are bonded to alkyl segments (Table 2). If we assume
about 6 aromatic carbons per isolated ring or cluster to be typical,
50% or 3 carbons are bonded to segments of 5 alkyl carbons, giving
an average ‘‘monomer’’ of 15 alkyl and 6 aromatic carbons, with al-
kyl CH3:alkyl CH2:alkyl CH:aromatic C proportions of 3:11:1.5:6
(Table 3). The majority of organic oxygen in our gilsonite sample
was apparently present as aromatic CAO with only traces of car-
boxyl and carbonyl groups detected (Fig. 1; Tables 1–3).
Based on the information obtained in this study, we propose a
structural model in which five-membered pyrrole rings and small
clusters of fused six-membered rings are connected to, and linked
by, alkyl chains of an average length of 10 carbons, see Fig. 12.
Whether the pyrrole rings are linked to the fused rings by alkyl
chains was not clear from our data. Since pyrrole rings contain
fewer carbons than fused six-membered aromatic rings, a larger
number of pyrrole rings was needed to match the number of car-
bons in the fused rings. The large fraction of CH2 and small amount
of CH groups showed that linear alkyl chains predominate. Methyl
end and side groups were detected by solution 13C and 1H NMR
(Fig. 8), and the methyl branches could account for a large fraction
of the observed alkyl CH groups.
Clearly, not every molecule of gilsonite resembles the structure
in Fig. 12, but molecules exhibiting similar structural features must
be very common. We reiterate that gilsonite is a complicated mix-
ture with a broad ranging continuum of fragments with varying
molecular weights. The model structure is depicted as a polymeric
structure because a significant number of the pyrrole and/or aro-
matic subunits appear to be connected in an amorphous polymer
or network structure. Based on the molecular weight distribution
reported here (Tables 5 and 6) and elsewhere, the degree of inter-
connection and the length of the connecting aliphatic chains are
probably highly variable. Macrocycles such as petroporphyrins
and refractory biomolecules represent a small fraction of the gil-
sonite as shown here by FT-ICR-MS and determined previously
(Quirke and Maxwell, 1980; Ruble et al., 1994; Schoell et al.,
1994). Oxygen is present mainly as aromatic CAO and some
C@O, according to NMR.
33
4.5. Comparison with literature
ESI-FT-ICR-MS detected a vanishingly small number of ions sug-
gestive of condensed hydrocarbons unlike earlier studies that sug-
gested that gilsonite contains significant quantities of condensed
aromatics or so-called asphaltene sheets (Yen et al., 1961; Wen
et al., 1978; Clark et al., 1983). Also, 13C DP-TOSS long range dipolar
dephasing curves (Fig. 4) indicated that gilsonite contains primar-
ily clusters of only a few fused aromatic rings, which shows that
large fused ring molecules such as asphaltene sheets are not pres-
ent at NMR detectable concentrations in gilsonite. This finding may
invalidate much of the structural information reported previously
(Yen et al., 1961) that assumed asphaltene sheets were a major
structural component of the mixture and indicates the need for
reinterpretation of the X-ray diffraction data (Yen et al., 1961). This
finding also differentiates gilsonite, and possibly other similar na-
tive bitumens, structurally from process asphalts that are byprod-
ucts of oil refinement. Our results do not rule out the existence of
aromatic sheets in gilsonite, but suggest they are a small fraction of
the aromatics or are largely composed of nitrogen-containing mac-
rocycles rather than carbonaceous condensed aromatics.
Earlier studies involving petroporphyrins in gilsonite suggested
an anoxic and therefore relatively rapid and comprehensive pres-
ervation of algal material prior to burial and catagenesis as the ori-
gin of gilsonite (Quirke et al., 1980b; Ruble et al., 1994). Our results
are consistent with this view. The presence of aromatic molecules
in gilsonite suggests that either (i) a more comprehensive suite of
biomolecules survived diagenesis than a typical algaenan (e.g. Bot-
ryococcus braunii), (ii) the dominant species of algae present during
sedimentation produced algaenan similar to Chlorella marina,
which were shown to yield aromatic hydrocarbons upon pyrolysis
(Derenne et al., 1996), or (iii) that the pyrrolic moieties within the
porphyrin core of algal pigments was efficiently preserved. It is not
surprising that we observed six-membered aromatic rings since it
has been commonly reported that six-membered aromatic rings
are preserved and/or produced during diagenesis and catagenesis
(North, 1985; Killops and Killops, 2005). Our results did not indi-
cate a significant contribution from lignin, preserved structural
34 J.R. Helms et al. / Organic Geochemistry 44 (2012) 21–36
features of which are observed in coal samples (Behar and Hatcher,
1995). Earlier studies have also suggested that gilsonite and other
native bitumens experienced relatively low temperature catagene-
sis based on the preservation of thermally labile biomarkers (Eglin-
ton et al., 1979). Our finding of extremely low abundance or
absence of condensed aromatics, which form during high temper-
ature catagenesis and charring, supports this as well. Exomethyl-
ene, which was detected and identified by 13C NMR, is often
associated with cyclization and cross linking reactions that occur
during diagenesis and catagenesis and remove olefinic character
from biogenic resins (Clifford et al., 1999).
Possible environmental sources of pyrrole rings include de-
graded pigments (Quirke and Maxwell, 1980; Quirke et al.,
1980b; Hajibrahim et al., 1981; Leenheer, 2003, 2009) and Mail-
lard reaction products from protein and carbohydrate starting
materials (Hodge, 1953; Hayase et al., 1989; Fang and Schmidt-
Rohr, 2009). The bulk structure of gilsonite does not resemble
that of modern chlorophylls or their degradation products in
plants or algae (Brown et al., 1984; Takamiya et al., 2000). For
chlorophylls, there are methyl or ethyl groups on the pyrrole
rings, the CH3 of which would have resonated at around 9 ppm
for 13C (2 ppm for 1H) and 16 ppm for 13C (1.3 ppm for 1H),
respectively. These signals were not observed in gilsonite. While
some chlorophyll products such as petroporphyrins may have
been preserved with the chlorin ring intact, our results suggest
that the majority of the macrocycles must have been opened
(Leenheer, 2003) and most of the unsaturated bonds must have
been hydrogenated. It might be suggested that Maillard or other
similar biomolecular reaction products might have provided sig-
nificant sources of pyrrole rings in gilsonite; however, it was
shown recently that the Maillard reaction produces mostly non-
protonated N (Fang and Schmidt-Rohr, 2009), while N in gilsonite
is mostly protonated.
Electrospray FT-ICR-MS provided an extensive database of
molecular formulae for polar constituents of gilsonite. ESI pro-
vided supporting evidence for the NMR detected presence of pri-
marily pyrrolic nitrogen compounds in gilsonite and revealed
several peaks that were likely structurally very similar to petro-
porphyrins. Petroporphyrins, hydrocarbons, and other non-polar
molecules may be more readily ionizable by employing liquid
injection field desorption ionization (Smith et al., 2008), laser in-
duced acoustic desorption (Pinkston et al., 2009), or atmospheric
pressure photo-ionization (Purcell et al., 2007; Qian et al., 2008;
McKenna et al., 2009) in place of ESI, and these modes of ioniza-
tion should be used in addition to ESI in the characterization stud-
ies of mostly hydrophobic natural organic mixtures in the future.
The lack of mass spectral differentiation between structural
isomers using FT-ICR-MS suggests that a physical or chemical sep-
aration method such as extraction, reversed phase liquid chroma-
tography, or gel permeation chromatography would be useful
prior to infusion into the mass spectrometer. The need for further
development of tandem MS methods that are compatible with FT-
ICR-MS (Sleighter, 2009) is similarly indicated. A more representa-
tive mass spectral database would provide further insights into
the molecular formulae present in natural organic mixtures,
would better complement the NMR techniques demonstrated in
this study, and would better facilitate the proposal of model
molecular structures.
The new structural information obtained using advanced NMR
techniques, especially the structural model proposed, represents
a significant step forward in the comprehensive chemical charac-
terization of gilsonite bitumen. While the use of ultrahigh resolu-
tion mass spectrometry is hampered somewhat by the selectivity
of ionization methods currently available, it is clear that the ability
to accurately calculate molecular formulae based on mass spectral
data has extraordinary potential as an analytical tool.
Acknowledgments
We would like to thank the National Science Foundation (EAR-
0843996 and CBET-0853950) and the Donors of the Petroleum Re-
search Fund, administered by the American Chemical Society Grant
46373-G2 for financial support. We also thank Dr. Gary G. Thomp-
son for technical support and Dr. Rachel Sleighter for assistance
with FT-ICR-MS. Dr. Hussain Abdulla, Dr. Sylvie Derenne, and
two anonymous reviewers provided helpful comments.
Associate Editor—Sylvie Derenne
References
Abraham, H., 1945. Asphalts and Allied Substances. Van Nostrand, New York.
Amster, J.I., 1996. Fourier transform mass spectrometry. Journal of Mass
Spectrometry 31, 1325–1337.
Anderson, K.B., Winans, R.E., Botto, R.E., 1992. The nature and fate of natural resins
in the geosphere-II. Identification, classification and nomenclature of resinites.
Organic Geochemistry 18, 829–841.
Aue, W.P., Bartholdi, E., Ernst, R.R., 1976. Two dimensional spectroscopy application
to nuclear magnetic resonance. Journal of Chemical Physics 64, 2229–2246.
Baker, E.W., Yen, T.F., Dickie, J.P., Rhodes, R.E., Clark, L.F., 1967. Mass spectrometry of
porphyrins. II. Characterization of petroporphyrins. Journal of the American
Chemical Society 89, 3631–3639.
Bax, A., Davis, D.G., 1985. MLEV-17-based two-dimensional homonuclear
magnetization transfer spectroscopy. Journal of Magnetic Resonance 65, 355–
360.
Behar, F., Hatcher, P.G., 1995. Artificial coalification of a fossilwood from brown coal
by confined system pyrolysis. Energy and Fuels 9, 984–994.
Bell, K.G., Hunt, J.M., 1963. Native bitumens associated with oil shales. In: Breger,
I.A. (Ed.), Organic Geochemistry. Elsevier, Amsterdam.
Bovey, F.A., Jelinski, L., Mirau, P.A., 1988. Nuclear Magnetic Resonance Spectroscopy.
Academic Press, San Diego, CA.
Brewer, C.E., Schmidt-Rohr, K., Satrio, J.A., Brown, R.C., 2009. Characterization of
biochar from fast pyrolysis and gasification systems. Environmental Progress
and Sustainable Energy 28, 386–396.
Brown, C.E., Spenser, R.B., Burger, V.T., Katz, J.J., 1984. Cross-polarization/magic-
angle sample-spinning 13C NMR spectroscopic study of chlorophyll a in the
solid state. Proceedings of the National Academy of Sciences of the United
States of America 81, 641–644.
Canuel, E.A., Cloern, J.E., Ringelberg, D.B., Guckert, J.B., Rau, G.H., 1995. Molecular
and isotopic tracers used to examine sources of organic matter and its
incorporation into the food webs of San Francisco Bay. Limnology and
Oceanography 40, 67–81.
Clark, D., Wilson, R., Quirke, J.M.E., 1983. An evaluation of the potential of ESCA
(electron spectroscopy for chemical applications) (and other spectroscopic
techniques) in the surface and bulk characterisation of kerogens, brown coal
and gilsonite. Chemical Geology 39, 215–239.
Clezy, P.S., Fookes, C.J.R., Prashar, J.K., 1989. The chemistry of pyrrolic compounds
LXI. Petroporphyrins from the Julia Creek Oil Shale: further evidence for the
derivation of aetiotype petroporphyrins from chlorophyll. Australian Journal of
Chemistry 42, 775–786.
Clifford, D.J., Hatcher, P.G., Botto, R.E., Muntean, J.V., Anderson, K.B., 1999. The
nature and fate of natural resins in the geosphere IX. Structure and maturation
similarities of soluble and insoluble polylabdanoids isolated from tertiary class I
resinites. Organic Geochemistry 30, 635–650.
Comisarow, M.B., Marshall, A.G., 1974. Frequency-sweep Fourier transform ion
cyclotron resonance spectroscopy. Chemical Physics Letters 26, 489–490.
Conte, P., Spaccini, R., Piccolo, A., 2004. State of the art of CPMAS 13C-NMR
spectroscopy applied to natural organic matter. Progress in Nuclear Magnetic
Resonance Spectroscopy 44, 214–223.
Corwin, A.H., 1959. Petro porphyrins. In: Proceedings of the 5th World Petroleum
Congress, New York, 5(109).
DeAzevedo, E.R., Hu, W.-G., Bonagamba, T.J., Schmidt-Rohr, K., 2000. Principles of
centerband-only detection of exchange in solid-state nuclear magnetic
resonance, and extension to four-time centerband-only detection of exchange.
Journal of Chemical Physics 112, 8988–9001.
Dennis, L.W., Maciel, G.E., Hatcher, P.G., Simoneit, B.R.T., 1982. 13C nuclear magnetic
resonance studies of kerogen from Cretaceous black shales thermally altered by
basaltic intrusions and laboratory simulations. Geochimica et Cosmochimica
Acta 46, 901–907.
Derenne, S., Largeau, C., Berkaloff, C., 1996. First example of an algaenan yielding an
aromatic-rich pyrolysate. Possible geochemical implications on marine kerogen
formation. Organic Geochemistry 24, 617–627.
Dickie, J.P., Yen, T.F., 1967. Macrostructure of the asphaltic fractions by various
instrumental methods. Analytical Chemistry 39, 1847–1852.
Dixon, W.T., 1982. Spinning-sideband-free and spinning-sideband-only NMR
spectra in spinning samples. Journal of Chemical Physics 77, 1800–1809.
 J.R. Helms et al. / Organic Geochemistry 44 (2012) 21–36
Eglinton, G., Hajibrahim, S.K., Maxwell, J.R., Quirke, J.M.E., Shaw, G.T., Volkman, J.K.,
Wardroper, A.M.K., 1979. Lipids of aquatic sediments, recent and ancient.
Transactions of the Royal Society 293, 69–91.
Fang, X., Schmidt-Rohr, K., 2009. Fate of the amino acid in glucose-glycine
melanoidins investigated by solid-state nuclear magnetic resonance (NMR).
Journal of Agricultural and Food Chemistry 57, 10701–10711.
Gill, J.P., Evershed, R.P., Chicarelli, M.I., Wolff, G.A., Maxwell, J.R., Eglinton, G., 1985.
Capillary gas chromatographic–mass spectrometric studies of the
petroporphyrins of the gilsonite bitumen (Eocene USA). Journal of
Chromatography 350, 37–62.
Grantham, P.J., Douglas, A.G., 1977. Carboxylic acids in some Uinta Basin bitumens.
In: Campos, R., Goñi, J. (Eds.), Advances in Organic Geochemistry. Enadimsa,
Madrid, pp. 193–207.
Grigsby, R.D., 1989. Processing of high resolution mass spectral data by use of
Kendrick masses in a rectangular array. In: Ashe, T.R., Wood, K.V. (Eds.), Novel
Techniques in Fossil Fuel Mass Spectrometry. American Society for Testing and
Materials, Philadelphia, pp. 172–193.
Hajibrahim, S.K., Quirke, J.M.E., Eglinton, G., 1981. Petroporphyrins V. Structurally-
related porphyrin series in bituments, shales and petroleums – evidence from
HPLC and mass spectrometry. Chemical Geology 32, 173–188.
Hatcher, P.G., Schnitzer, M., Dennis, L.W., Maciel, G.E., 1981. Aromaticity of humic
substances in soils. Soil Society of America Journal 45, 1089–1094.
Hatcher, P.G., Spiker, E.C., Szeverenyi, N.M., Maciel, G.E., 1983. Selective
preservation and origin of petroleum-forming aquatic kerogen. Nature 305,
498–501.
Hayase, F., Nagaraj, R.H., Miyata, S., Njoroge, F.G., Monnier, V.M., 1989. Aging of
proteins: immunological detection of a glucose-derived pyrrole formed during
Maillard reaction in vivo. Journal of Biological Chemistry 263, 3758–3764.
Herniman, J.M., Langley, G.J., Bristow, T.W.T., O’Connor, G., 2005. The validation of
exact mass measurements for small molecules using FT-ICRMS for improved
confidence in the selection of elemental formulas. Journal of the American
Society of Mass Spectrometry 16, 1100–1108.
Hertkorn, N., Benner, R., Frommberger, M., Schmitt-Kopplin, P., Witt, M., Kaiser, K.,
Kettrup, A., Hedges, J.I., 2006. Characterization of a major refractory component
of marine dissolved organic matter. Geochimica et Cosmochimica Acta 70,
2990.
Hockaday, W.C., Grannas, A.M., Kim, S., Hatcher, P.G., 2006. Direct molecular
evidence for the degradation and mobility of black carbon in soils from
ultrahigh-resolution mass spectral analysis of dissolved organic matter from a
fire-impacted forest soil. Organic Geochemistry 37, 501–510.
Hodge, J.E., 1953. Dehydrated foods: chemistry of browning reactions in model
systems. Journal of Agricultural and Food Chemistry 1, 928–943.
Hsu, C.S., Llang, Z., Campanas, J.E., 1994. Hydrocarbon characterization by ultrahigh
resolution Fourier transform ion cyclotron resonance mass spectrometry.
Analytical Chemistry 66, 850–855.
Hu, J.Z., Harper, J.K., Taylor, C., Pugmire, R.J., Grant, D.M., 2000. Modified spectral
editing methods for 13C CP/MAS experiments in solids. Journal of Magnetic
Resonance 142, 326–330.
Hughey, C.A., Hendrickson, C.L., Rodgers, R.P., Marshall, A.G., 2001. Kendrick mass
defect spectrum: a compact visualanalysis for ultrahigh-resolution broadband
mass spectra. Analytical Chemistry 73, 4676–4681.
Jacob, H., 1989. Classification, structure, genesis and practical importance of natural
solid oil bitument (‘‘migrabitumen’’). International Journal of Coal Geology 11,
65–79.
Keeler, C., Maciel, G.E., 2000. 13C NMR spectral editing of humic materials. Journal of
Molecular Structure 550–551, 297–305.
Kendrick, E., 1963. A mass scale based on CH2 = 14.0000 for high resolution mass
spectrometry of organic compounds. Analytical Chemistry 35, 2146–2154.
Kenyon, C.N., 1972. Fatty acid composition of unicellular strains of blue-green algae.
Journal of Bacteriology 109, 827–834.
Killops, S., Killops, V., 2005. Introduction to Organic Geochemistry. Blackwell
Publishing, Boston, MA.
Kim, S., Kramer, R.W., Hatcher, P.G., 2003. Graphical method for analysis of
ultrahigh-resolution broadband mass spectra of natural organic matter, the van
Krevelen diagram. Analytical Chemistry 75, 5336–5344.
Kim, S., Kaplan, L.A., Benner, R., Hatcher, P.G., 2004. Hydrogen-deficient molecules
in natural riverine water samples evidence for the existence of black carbon in
DOM. Marine Chemistry 92, 225–234.
Kim, S., Rodgers, R.P., Marshal, A.G., 2006. Truly ‘exact’ mass: elemental composition
can be determined uniquely from molecular mass measurements at 0.1 mDa
accuracy for molecules up to 500 Da. International Journal of Mass
Spectrometry 251, 260–265.
Klein, G.C., Angstrom, A., Rodgers, R.P., Marshal, A.G., 2006. Use of saturates/
aromatics/resins/asphaltenes (SARA) fractionation to determine matrix effects
in crude oil analysis by electrospray ionization Fourier transform ion cyclotron
resonance mass spectrometry. Energy and Fuels 20, 668–672.
Koch, B.P., Dittmar, T., 2006. From mass to structure: an aromaticity index for high-
resolution mass data of natural organic matter. Rapid Communications in Mass
Spectrometry 20, 926–932.
Kujawinski, E.B., 2002. Electrospray ionization Fourier transform ion cyclotron
resonance mass spectrometry (ESI FT-ICR MS): characterization of complex
environmental mixtures. Environmental Forensics 3, 207–216.
Leenheer, J.A., 2003. Comprehensive Characterization of Dissolved and Colloidal
Organic Matter in Waters Associated with Groundwater Recharge at the Orange
County Water District, Final Report to the Orange County Water District,
Fountain Valley, California, US Geological Survey, Washington, DC.
35
Leenheer, J.A., 2009. Systematic approaches to comprehensive analyses of natural
organic matter. Annals of Environmental Science 3, 1–130.
Maciel, G.E., Bronniman, C.E., Ridenour, C.F., 1993. H-1-NMR spectroscopy –
approaches for carbonaceous solids. Advances in Chemistry Series 229, 27–44.
Mao, J.-D., Schmidt-Rohr, K., 2003. Recoupled long-range C–H dipolar dephasing in
solid-state NMR, and its use for selection of fused aromatic rings. Journal of
Magnetic Resonance 162, 217–227.
Mao, J.-D., Schmidt-Rohr, K., 2004a. Accurate quantification of aromaticity and
nonprotonated aromatic carbon fraction in natural organic matter by 13C solid-
state nuclear magnetic resonance. Environmental Science and Technology 38,
2680–2684.
Mao, J.-D., Schmidt-Rohr, K., 2004b. Separation of aromatic-carbon 13C NMR signals
from di-oxygen alkyl bands by a chemical-shift-anisotropy filter. Solid State
Nuclear Magnetic Resonance 26, 36–45.
Mao, J.-D., Schmidt-Rohr, K., 2005. Methylene spectral editing in solid-state 13C
NMR by three-spin coherence selection. Journal of Magnetic Resonance 176, 1–
6.
Mao, J.-D., Schmidt-Rohr, K., 2006. Absence of mobile carbohydrate domains in dry
humic substances proven by NMR, and implications for organic-contaminant
sorption models. Environmental Science and Technology 40, 1751–1756.
Mao, J.-D., Hu, W.-G., Schmidt-Rohr, K., Davies, G., Ghabbour, E.A., Xing, B., 2000.
Quantitative characterization of humic substances by solid-state carbon-13
nuclear magnetic resonance. Soil Science Society of America Journal 64, 873–
884.
Mao, J.-D., Cory, R.M., McKnight, D.M., Schmidt-Rohr, K., 2007a. Characterization of
a nitrogen-rich fulvic acid and its precursor algae from solid state NMR. Organic
Geochemistry 38, 1277–1292.
Mao, J.-D., Fang, X., Schmidt-Rohr, K., Carmo, A.M., Hundal, L.S., Thompson, M.L.,
2007b. Molecular-scale heterogeneity of humic acid in particle-size fractions of
two Iowa soils. Geoderma 140, 17–29.
Mao, J.-D., Tremblay, L., Gagne, J.-P., Kohl, S., Rice, J., Schmidt-Rohr, K., 2007c. Humic
acids from particulate organic matter in the Saguenay Fjord and the St.
Lawrence Estuary investigated by advanced solid-state NMR. Geochimica et
Cosmochimica Acta 71, 5483–5499.
Mao, J.-D., Fang, X., Lan, Y., Schimmelmann, A., Mastalerz, M., Xu, L., Schmidt-Rohr,
K., 2010a. Chemical and nanometer-scale structure of kerogen and its change
during thermal maturation investigated by advanced solid-state 13C NMR
spectroscopy. Geochimica et Cosmochimica Acta 74, 2110–2127.
Mao, J.-D., Schimmelmann, A., Mastalerz, M., Hatcher, P.G., Li, Y., 2010b. Structural
features of a bituminous coal and their changes during low-temperature
oxidation and loss of volatiles investigated by advanced solid-state NMR
spectroscopy. Energy and Fuels 24, 2536–2544.
Marshall, A.G., Hendrickson, C.L., Jackson, G.S., 1998. Fourier transform ion
cyclotron resonance mass spectrometry: a primer. Mass Spectrometry
Reviews 17, 1–35.
McKenna, A.M., Purcell, J.M., Rodgers, R.P., Marshall, A.G., 2009. Identification of
vanadyl porphyrins in a heavy crude oil and raw asphaltene by atmospheric
pressure photoionization Fourier transform ion cyclotron (FT-ICR) mass
spectrometry. Energy and Fuels 23, 2122–2128.
Meyer, R.F., Attanasi, E.D., Freeman, P.A., 2007. Heavy Oil and Natural Bitumen
Resources in Geological Basins of the World, Open Form Document No. 2007-
1084. US Geological Survey, Reston, VA, pp. 1–42.
Miknis, F.P., Smith, J.W., Maughan, E.K., Maciel, G.E., 1982. Nuclear magnetic
resonance: a technique for direct evaluation of source-rock potential. American
Association of Petroleum Geologists Bulletin 66, 1396–1401.
Nanny, M.A., Minear, R.A., Leenheer, J.A., 1997. Nuclear magnetic resonance
spectroscopy in environmental chemistry. In: Topics in Environmental
Chemistry. Oxford University Press, New York.
North, F.K., 1985. Petroleum Geology. Allen & Unwin, Boston.
Osmond, J.C., 1964. Guidebook to the geology and mineral resources of the uinta
basin. In: Sabatka, E.F. (Ed.), 13th Annual Field Conference of the Intermountain
Association of Petroleum Geologists. Intermountain Association of Petroleum
Geologists, pp. 47–58.
Perdue, E.M., 1984. Analytical constraints on the structural features of humic
substances. Geochimica et Cosmochimica Acta 48, 1435–1442.
Picard, M.D., High, L.R., 1968. Sedimentary cycles in the Green River Formation
(Eocene), Uinta Basin, Utah. Journal of Sedimentary Research 38, 378–383.
Pinkston, D.S., Duan, P., Gallardo, V.A., Habicht, S.C., Tan, X., Qian, K., Gray, M.,
Mullen, K., Kenttamaa, H.I., 2009. Analysis of asphaltenes and asphaltene model
compounds by laser-induced acoustic desorption/Fourier transform ion
cyclotron resonance mass spectrometry. Energy and Fuels 23, 5564–5570.
Preston, C.M., 1996. Application of NMR to soil organic matter analysis: history and
prospects. Soil Science 16, 144–166.
Purcell, J.M., Hendrickson, C.L., Rodgers, R.P., Marshall, A.G., 2007. Atmospheric
pressure photoionization proton transfer for complex organic mixtures
investigated by Fourier transform ion cyclotron resonance mass spectrometry.
Journal of the American Society of Mass Spectrometry 18, 1682–1689.
Qian, K., Mennito, A.S., Edwards, K.E., Ferrughelli, D.T., 2008. Observation of vanadyl
porphyrins and sulfur-containing vanadyl porphyrins in a petroleum
asphaltene by atmospheric pressure photoionization Fourier transform ion
cyclotron resonance mass spectrometry. Rapid Communications in Mass
Spectrometry 22, 2153–2160.
Quirke, J.M.E., Maxwell, J.R., 1980. Petrophyrins III. Characterisation of a C32
aetioporphyrin from gilsonite as the bis[porphyrinato-mercurcy(II)
acetato]mercury(II) complex. Origin and significance. Tetrahedron 36, 3453–
3456.
36 J.R. Helms et al. / Organic Geochemistry 44 (2012) 21–36
Quirke, J.M.E., Maxwell, J.R., Eglinton, G., 1980a. Petrophyrins IV. Nuclear
overhauser enhancement 1H NMR studies of deoxophylloerythroetio
porphyrins from gilsonite. Tetrohedron Letters 21, 2987–2990.
Quirke, J.M.E., Shaw, G.J., Soper, P.D., Maxwell, J.R., 1980b. Petroporphyrins II. The
presence of porphyrins with extended alkyl substituents. Tetrahedron 36,
3261–3267.
Ruble, T.E., Bakel, A.J., Philip, R.P., 1994. Compound specific isotopic variability in
Uinta Basin native bitumens: paleoenviromental implications. Organic
Geochemistry 21, 661–671.
Schmidt-Rohr, K., Mao, J.-D., 2002. Efficient CH-group selection and identification in
13
C solid-state NMR by dipolar DEPT and 1H chemical-shift filtering. Journal of
the American Chemical Society 124, 13938–13948.
Schmidt-Rohr, K., Clauss, J., Spiess, H.W., 1992. Correlation of structure, mobility,
and morphological information in heterogeneous polymer materials by two-
dimensional wideline-separation NMR spectroscopy. Macromolecules 25,
3273–3277.
Schoell, M., Hwang, R.J., Carlson, R.M.K., Welton, J.E., 1994. Carbon isotopic
composition of individual biomarkers in gilsonites. Organic Geochemistry 21,
673–683.
Sleighter, R.L., 2009. Export of terrestrial dissolved organic matter along a river to
ocean transect of the lower Chesapeake Bay investigated by advanced analytical
techniques. In: Chemistry and Biochemistry PhD, Old Dominion University,
Norfolk, VA, p. 241.
Sleighter, R.L., Hatcher, P.G., 2007. The application of electrospray ionization
coupled to ultrahigh resolution mass spectrometry for the molecular
characterization of natural organic matter. Journal of Mass Spectrometry 42,
559–574.
Sleighter, R.L., McKee, G.A., Liu, Z., Hatcher, P.G., 2008. Naturally present fatty acids
as internal calibrants for Fourier transform mass spectra of dissolved organic
matter. Limnology and Oceanography Methods 6, 246–253.
Smernik, R.J., Oades, J.M., 2000. The use of spin counting for determining
quantitation in solid state 13C NMR spectra of natural organic matter: 2. HF-
treated soil fractions. Geoderma 96, 159–171.
Smernik, R.J., Schwark, L., Schmidt, M.W.I., 2006. Assessing the quantitative
reliability of solid-state C-13 NMR spectra of kerogens across a
gradient of thermal maturity. Solid State Nuclear Magnetic Resonance 29,
312–321.
Smith, D.F., Rahimi, P., Teclemariam, A., Rodgers, R.P., Marshall, A.G., 2008.
Characterization of athabasca bitumen heavy vacuum gas oil distillation cuts
by negative/positive electrospray ionization and automated liquid injection
field desorption ionization Fourier transform ion cyclotron resonance mass
spectrometry. Energy and Fuels 22, 3118–3125.
Solum, M.S., Pugmire, R.J., Grant, D.M., 1989. C-13 solid-state NMR of Argonne
premium coals. Energy and Fuels 3, 187–193.
Stenson, A.C., 2008. Reversed-phase chromatography fractionation tailored to mass
spectral characterization of humic substances. Environmental Science and
Technology 42, 2060–2065.
Stenson, A.C., Marshall, A.G., Cooper, W.T., 2003. Exact masses and chemical
formulas of individual suwannee river fulvic acids from ultrahigh resolution
electrospray ionization fourier transform ion cyclotron resonance mass spectra.
Analytical Chemistry 75, 1275–1284.
Sugihara, J.M., McGee, L.R., 1957. Porphyrins in gilsonite. Journal of Organic
Chemistry 22, 795.
Takamiya, K.-I., Tsuchiya, T., Ohta, H., 2000. Degradation pathway(s) of chlorophyll:
what has gene cloning revealed? Trends in Plant Science 5, 426–431.
Tekely, P., Palmas, P., Mutzenhardt, P., 1993. Elimination of hetero-nuclear dipolar
interactions from carbon-13-detected proton spectra in wide-line separation
nuclear magnetic resonance spectroscopy. Macromolecules 26, 7363–7365.
Thorn, K.A., Goldenburg, W.S., Younger, S.J., Weber, E.J., 1996. Covalent binding of
aniline to humic substances: comparison of nucleophilic addition, enzyme-, and
metal-catalyzed reactions by 15N NMR. In: Gaffney, J.S., Marley, N.A., Clark, S.B.
(Eds.), Humic and Fulvic Acids: Isolation, Structure, and Environmental Role.
American Chemical Society, Washington, DC, pp. 299–326.
Treibs, A., 1934. Organic minerals III. Chlorophyll and hemin derivatives in
bituminous rocks, petroleums, mineral waxes and asphalts. Origin of
petroleum. Justus Liebigs Annalen der Chemie 510, 42–62.
Treibs, A., 1936. Chlorophyll and hemin derivatives in organic mineral substances.
Angewandte Chemie 49, 682–686.
Tripp, B.T., White, E.R., 2006. Gilsonite. In: Kogel, J.E., Trivedi, N.C., Barker, J.M.,
Krukowski, S.T. (Eds.), Industrial Minerals and Rocks: Commodities, Markets,
and Uses. Society for Mining, Metallurgy, and Exploration, Littleton, CO, pp.
481–493.
Van Berkel, G.J., Quinones, M.A., Quirke, J.M.E., 1993. Geoporphyrin analysis using
electrospray ionization-mass spectrometry. Energy and Fuels 7, 411–419.
Vuister, G.W., Boelens, R., Kaptein, R., 1991. Gradient-enhanced HMQC and HSQC
spectroscopy – applications to N-15-labeled MNT repressor. Journal of the
American Chemical Society 113, 9688–9690.
Wen, C.S., Chilingarian, G.V., Yen, T.F., 1978. Properties and structure of bitumens.
In: Chilingarian G.V., Yen, T.F. (Eds.), Developments in Petroleum Science:
Bitumens, Asphalts, and Tar Sands, vol. 7. Elsevier, Amsterdam.
Werner-Zwanzinger, U., Lis, G., Mastalerz, M., Schimmelmann, A., 2005. Thermal
maturity of type II kerogen from the New Albany Shale assessed by 13C CP/MAS
NMR. Solid State Nuclear Magnetic Resonance 27, 140–148.
Wilson, M.A., 1987. NMR Techniques and Applications in Geochemistry and Soil
Chemistry. Pergamon Press, Oxford, UK.
Wilson, M.A., Hanna, J.V., Anderson, K.B., Botto, R.E., 1993. 1H CRAMPS NMR derived
hydrogen distributions in various coal macerals. Organic Geochemistry 20,
985–999.
Wu, X., Zilm, K.W., 1993. Complete spectral editing in CP/MAS NMR. Journal of
Magnetic Resonance A 102.
Wu, X., Burns, S.T., Zilm, K.W., 1994. Spectral editing in CP/MAS NMR. Generating
subspectra based on proton multiplicities. Journal of Magnetic Resonance 111,
29–36.
Yen, T.F., 2000. The realms and definitions of asphaltenes. In: Yen, T.F., Chilingarian,
G.V. (Eds.), Developments in Petroleum Science 40B: Asphaltenes and Asphalts,
vol. 2. Elsevier, Amsterdam, pp. 7–27.
Yen, T.F., Erdman, J.G., Pollack, S.S., 1961. Investigation of the structure of petroleum
asphaltenes by X-ray diffraction. Analytical Chemistry 33, 1587–1594.