Energy & Fuels 1997, 11, 554-560
554
High-Resolution Fourier Transform Ion Cyclotron
Resonance Mass Spectrometry of Humic and Fulvic
Acids by Laser Desorption/Ionization and Electrospray
Ionization
Anne Fievre,† Touradj Solouki,‡ Alan G. Marshall,†,‡ and William T. Cooper*,†
Department of Chemistry, Florida State University, Tallahassee, Florida 32306-3006 and
Center for Interdisciplinary Magnetic Resonance, National High Magnetic Field Laboratory,
Florida State University, 1800 Paul Dirac Drive, Tallahassee, Florida 32310
Received January 3, 1997. Revised Manuscript Received March 19, 1997X
High-field (9.4 T) Fourier transform ion cyclotron resonance (FT-ICR) mass spectra of standard
Suwannee River humic and fulvic acids have been obtained by use of laser desorption (LDI) and
electrospray (ESI) ionization. The LDI FT-ICR mass spectrum was similar to those observed
previously, with ions at essentially every nominal value, 200 e m/z e 800. In contrast, the ESI
FT-ICR mass spectrum, although still containing ions at most values in the 200 e m/z e 800
range, was dominated by a relatively few prominent species. ESI FT-ICR mass spectra of standard
humic and fulvic acid isolates were similar. Although many ionic species appeared in both fulvic
acid and humic acid ESI FT-ICR mass spectra, the fulvic acid mass spectrum contained more
highly charged species. Subfractions of the fulvic acid mixture isolated by an HPLC procedure
yielded similar mass spectra. The stability of high-mass ions produced by ESI combined with
the high-mass resolution capability of FT-ICR MS allow for precise determination of molecular
masses, from which molecular formulas may be obtained by mass alone. Future two-dimensional
FT-ICR MS2 determinations of humic and fulvic acid structures should be feasible by use of
collisionally induced and multiple-photon dissociation techniques.
Introduction
Humic Substances. Humic substances are ubiquitous in virtually all terrestrial and estuarine environments and comprise between 50 and 80% of the dissolved organic matter (DOM) in surface waters.1 These
geomacromolecular compounds are amorphous, acidic
substances of molecular weights estimated to range
from several hundred to tens of thousands. The ability
of naturally-occurring organic matter in general, and
humic substances in particular, to absorb, bind, and/or
complex environmentally-significant substances such as
pesticides, polychlorinated biphenyls, polyaromatic hydrocarbons, and other nonpolar organics has been well
documented.2 Complexation of heavy metals by humics
has also been reported. The effects of binding to organic
matter on the bioavailability of chemicals is a particularly interesting but poorly understood phenomenon.
Several studies have suggested that such binding
reduces bioavailability.3-5 At pH 1, the insoluble com†
Department of Chemistry.
National High Magnetic Field Laboratory.
Abstract published in Advance ACS Abstracts, May 1, 1997.
(1) Malcolm, R. L.; Aiken, G. R.; Bowles, E. C.; Malcolm, J. D. In
Humic Substances in the Suwanee River Georgia: Interactions, Properties, and Proposed Structures; Averett, R. C., Leenheer, J. A., McKnight, D. A., Thorn, K. A., Eds.; U.S. Geological Survey, 1989; pp 2326.
(2) Pignatello, J. J.; Boashan, X. Environ. Sci. Technol. 1996, 30,
1-11.
(3) Carlberg, G. E.; Martinsen, K.; Kringstad, A.; Gjessing, E.;
Grande, M.; Kallqvist, T.; Skare, J. U. Arch. Environ. Contam. Toxicol.
1986, 15, 543-548.
(4) Kukkonen, J.; Oikari, S.; Johnsen, E.; Gjessing, E. Sci. Total
Environ. 1989, 79, 197-207.
‡
ponent of humic substances is known as humic acid and
the soluble component is called fulvic acid.
In spite of advances in understanding the extent to
which organic matter in aquatic systems influences the
bioavailability and geochemical behavior of chemical
contaminants, little is known about the relationship
between the composition of humic substances and their
chemical and biological reactivity. It is known that
humics produced in different environments have different biogeochemical reactivity, and these differences can
be related to some extent to gross structural differences
(e.g., aromaticity, carboxyl content, etc.). However, up
to this point, analytical techniques capable of correlating
biogeochemical reactivity with specific geomacromolecular structures in humic/fulvic acid mixtures have not
been routinely available.
Mass Spectrometry of Humic Substances. Mass
spectrometry has been used extensively in the study of
humics, primarily as a means of obtaining information
about fragments after chemical6,7 or thermal8,9 degradation. Of particular interest were studies using timeresolved pyrolysis-field ionization (PFI) coupled to high-
X
S0887-0624(97)00005-4 CCC: $14.00
(5) Landrum, P. F.; Reinhold, M. D.; Nihart, S. R.; Eadie, B. J.
Environ. Toxicol. Chem. 1985, 4, 459-467.
(6) Sedlacek, J.; Kallqvist, T.; Gjessing, E. In Aquatic and Terrestrial
Humic Material; Christman, R. F., Gjessing, E. T., Ed.; Ann Arbor
Science: Ann Arbor, MI, 1983; pp 495-516.
(7) Aiken, G. R.; McKnight, D. M.; Wershaw, R. L.; MacCarthy, P.
Humics Substances in Soil, Sediment and Water; Malcolm, R. L., Ed.;
Wiley Interscience: New York, 1985; pp 181-209.
(8) Saiz-Jiminez, C.; de Leeuw, J. W. J. Anal. Appl. Pyrolysis 1986,
9, 99-119.
(9) Sorge, C.; Muller, R.; Leinweber, P.; Schulten, H.-R.; Fresenius,
J. Anal. Chem. 1993, 346, 697-703.
© 1997 American Chemical Society
LDI and ESI FT-ICR MS of Humic and Fulvic Acids
resolution mass spectrometry (Py-FIMS).10 Those
experiments provided direct molecular characterizations
and confirmed degradation products previously identified by GC-MS experiments.
Virtually all early mass spectrometric studies of
humic substances were characterized by the extensive
fragmentation produced by conventional electron impact
ionization, leading to fragment ions at virtually every
nominal mass below mass-to-charge ratio, m/z ∼ 200,
and few if any fragment ions above m/z 200. (m is ion
mass in daltons and z is the ion charge in multiples of
the elementary charge.) “Soft” ionization techniques,
such as fast atom bombardment (FAB),11 field desorption (FD),12 and in-source PFI,13 are therefore of interest. Field ionization studies were particularly promising,12 with intact fragments observed up to m/z 670.
Laser desorption (LD) and matrix-assisted laser desorption/ionization (MALDI) techniques have also been
used as soft ionization methods for humics and related
compounds such as lignins.14,15 Combining LDI with
Fourier transform ion cyclotron resonance mass spectrometry (FT-ICR MS),16-19 Srzic et al. observed from
lignins negative ions that ranged from m/z of several
hundred to above 3000.15
In addition, Novotny and Rice recently reported on
the use of low-power LDI and FT-ICR MS to characterize the number-average (Mn) molecular weights of five
fulvic acid samples.20 Their calculations were based on
the assumption that all ions observed in the FT-ICR
mass analyzer are intact, singly-charged humic ions,
and not fragments. However, their mass spectral
results did not agree with Mn values obtained by
osmometry techniques, suggesting that even at the
relatively low laser power used in that study, significant
fragmentation of parent humic acid geomacromolecules
still occurs.
Detection of such large fragments raises the possibility of reconstructing the structure of entire humic
molecules by mass analysis alone. In this initial
presentation, we report the use of what we believe to
be the “softest” ionization technique available for analysis of these macromolecules, namely, electrospray ionization (ESI). We compare LD FT-ICR MS at 3.0 T to
ESI FT-ICR-MS at the highest available magnetic field
to date (9.4 T).21 The high-field ESI FT-ICR MS
technique yields ultrahigh resolution mass spectra,22
(10) Abbt-Braun, G.; Frimmel, F. H.; Schulten, H. R. Water Res.
1989, 23, 1579-1591.
(11) Saleh, F. Y.; Chanz, D.; Frye, J. S. Anal. Chem. 1983, 55, 826.
(12) Andreux, F.; Constantin, E.; Gioia, B.; Traldi, P. Org. Mass
Spectrom. 1988, 23, 622-623.
(13) Schulten, H.-R.; Schnitzer, M. Org. Geochem. 1993, 20, 1725.
(14) Srzic, D.; Martinovic, S.; Pasa, L.; Kezele, N.; Shevchenko, S.
M. Rapid. Commun. Mass Spectrom. 1995, 9, 245-247.
(15) Srzic, D.; Martinovic, S.; Pasa-Tolic, L.; Kezele, N.; Kazazic,
S.; Senkovic, L.; Shevchenko, S. M.; Klasinc, L. Rapid. Commun. Mass
Spectrom. 1996, 10, 580-582.
(16) Marshall, A. G. Acc. Chem. Res. 1985, 18, 316-322.
(17) Buchanan, M. V.; Hettich, R. L. Anal. Chem. 1993, 65, 245A259A.
(18) Wilkins, C. L. Trends in Analytical Chemistry 13, Special Issue:
Fourier Transform Mass Spectrometry; Wilkins, C. L., Ed., 1994; pp
223-251.
(19) McLafferty, F. W. Acc. Chem. Res. 1994, 27, 379-386.
(20) Novotny, F. J.; Rice, J. A. Environ. Sci. Technol. 1995, 29, 2464.
(21) Marshall, A. G.; Guan, S. Rapid Commun. Mass Spectrom.
1996, 10, 1819-1823.
(22) Senko, M. W.; Hendrickson, C. L.; Pasa-Tolic, L.; Marto, J. A.;
White, F. M.; Guan, S.; Marshall, A. G. Rapid Commun. Mass
Spectrom. 1996, 10, 1824-1828.
Energy & Fuels, Vol. 11, No. 3, 1997 555
from which chemical formulas of various species may
be determined directly by accurate mass measurement.
In addition, after stored waveform inverse Fourier
transform (SWIFT)23,24 isolation of ions over very narrow (a few Da) mass ranges, MS/MS experiments based
on collisionally induced dissociation (CID) and infrared
multiple photon dissociation (IR/MPD) yield fragment
ions that may be used for structural characterization
of unknown molecules. Such tandem mass spectrometry experiments support the feasibility of determining
the chemical formulas (and ultimately the structures)
of primary and fragment ions, and lay the groundwork
for future mass analysis of humic substances.
Experimental Section
Samples. Standard humic and fulvic acid, SRHA and
SRFA, respectively, were supplied by the International Humic
Substances Society (Golden, CO). They were isolated from the
Suwannee River (GA), by the IHSS standard method for
extraction and isolation of aquatic humic substances.1 SFRI
is a Suwannee River (FL) fulvic acid collected and isolated by
a similar procedure described in Standard Methods of Water
and Wastewater.25
Electrospray FT-ICR Mass Spectrometry. ESI FT-ICR
mass spectra were acquired with a homebuilt FT-ICR mass
spectrometer equipped with a 9.4 T superconducting magnet,
as described elsewhere.22 Briefly, ions are produced with an
external electrospray source.26 The electrosprayed ions pass
through a 1 mm diameter skimmer prior to their entrance into
a first, 60 cm long, rf-only octupole ion guide. A second
octupole, 200 cm in length, guides the ions into a 9.4 cm
diameter cylindrical (∼30.4 cm long) open-ended three-section
Penning trap.27,28 To increase the number of trapped ions
inside the ICR cell, we allow ions to accumulate for about 10
s inside the first octupole prior to ion transfer into the second
octupole.
A Model 48-2 air-cooled carbon dioxide laser (SYNRAD,
Bothell, WA), operated at a wavelength range of 10.55-10.65
µm and a maximum output of 40 W, was used for infrared
multiphoton dissociation (IR/MPD) experiments. The CO2
laser is located outside the magnetic field and ∼135 cm away
from the center of the ICR ion trap. The laser beam (beam
diameter/divergence: 3.5 mm/4 mrad) is coaligned with the
magnetic field and is directed through a flange-mounted
barium fluoride (BaF2) window (BICRON, Solon, OH) on the
ICR cell axis. A 30 cfm rotary pump (Varian, Lexington, MA)
and three 1100 L/s hybrid turbo-drag pumps (Balzers, Hudson,
NH) provide differential pumping of the vacuum system to
maintain an operating base pressure of ∼2 × 10-8 Torr inside
the ICR ion trap. Instrumental parameters are controlled by
an Odyssey data system (Finnigan Corp., Madison, WI). Argon
collision gas was introduced into the ICR chamber at ∼2 ×
10-5 Torr, via a pulsed valve (General Valve, Fairfield, NJ).
The fulvic and humic sample solutions were electrosprayed
at a rate of 1-5 µL/min.
Laser Desorption/Ionization Mass Spectrometry. Laser desorption/ionization (LDI) FT-ICR mass spectra were
(23) Marshall, A. G.; Wang, T.-C. L.; Ricca, T. L. J. Am. Chem. Soc.
1985, 107, 7893-7897.
(24) Guan, S.; Marshall, A. G. Int. J. Mass Spectrom. Ion Processes
1996, 137/138, 5-37.
(25) Clesceri, L. S.; Greenberg, A. E.; Trussell, R. R.; Franson, M.
A. Standard Methods for the Examination of Water and Waste Water;
American Public Health Association: Washington, DC, 1989; Vol. 17,
pp 5.37-5.41.
(26) Chowdhury, S. K.; Katta, V.; Chait, B. T. Rapid Commun. Mass
Spectrom. 1990, 4, 81-87.
(27) Gabrielse, G.; Haarsma, L.; Rolston, S. L. Int. J. Mass Spectrom.
Ion Processes 1989, 88, 319-332.
(28) Beu, S. C.; Laude, D. A., Jr. Int. J. Mass Spectrom. Ion Processes
1992, 112, 215-230.
556
Energy & Fuels, Vol. 11, No. 3, 1997
acquired with an FTMS-2000 Fourier transform ion cyclotron
resonance mass spectrometer (Finnigan Corp., Madison, WI)
equipped with a 3 T superconducting magnet, dual cubic
Penning traps, and an Odyssey data system. Laser desorption/
ionization was performed with a Nd:YAG laser (Model Surelite
I-10, Continuum, Santa Clara, CA) operated at a wavelength
of 1064 nm (fundamental output) with a pulse width of 7 ns.
The laser beam is focused onto the probe tip by a 2:1 telescope
(typical laser pulse energy was 1 mJ measured before the
telescope, corresponding to an estimated laser power density
of ∼2 × 107 W cm-2 on the probe tip). The laser beam is
directed through a quartz window on the analyzer side of the
main vacuum chamber, passing through a 2 mm diameter
conductance limit to a spot size of ∼400 µm × 600 µm on the
probe tip behind the source trap plate. The fulvic and humic
sample solutions were applied to a thin stainless steel plate
(sample probe tip) and air-dried. The sample probe tip was
inserted into the vacuum chamber and mass analyzed. All
LD FT-ICR mass spectra were acquired in the source compartment of the dual trap at a background pressure of ∼2 × 10-8
Torr.
After the laser desorption/ionization event, ion z-axis translational energy was minimized by use of gated deceleration,
as previously described.29,30 The trapped ions were excited by
dipolar frequency sweep excitation (∼132 Vp-p amplitude,
1-500 kHz at a sweep rate of 1100 Hz/µs). Fourier transformation of the resulting discrete time-domain signal (32 K data,
1 MHz Nyquist bandwidth), without zero-filling, followed by
Hamming apodization, magnitude calculation, and frequencyto-mass conversion yielded an LD FT-ICR mass spectrum.
MSn Experiments. The experimental event sequences for
successive FT-ICR MSn experiments are published elsewhere.30
Briefly, ESI-generated ions of initially high kinetic energy are
decelerated and trapped. The trapped ions are allowed to relax
axially to the center of the trap for several seconds. Ion
cyclotron motions of the trapped ions are then excited by
dipolar frequency-sweep irradiation, for which the sweep rate
and radiofrequency voltage amplitude are optimized for each
sample. Fourier transformation of the resulting discrete timedomain signal (32-256 K data, 1 MHz Nyquist bandwidth),
without zero-filling and with Hamming apodization, followed
by magnitude calculation and frequency-to-mass conversion,
yields an FT-ICR parent ion mass spectrum. We used storedwaveform inverse Fourier transform (SWIFT) radial ejection
to remove parent ions of all but a selected m/z ratio(s). In CID
experiments, the mass-selected parent ions are then translationally excited to dissociate by means of collisional activation
provided by sustained off-resonance irradiation (SORI)31 at
∼800 Hz below the reduced ion cyclotron frequency of the
parent ion. In IR/MPD experiments,32 a CO2 laser beam,
coaligned with the magnetic field and directed through a
flange-mounted barium-fluoride (BaF2) window on the ion trap
axis, dissociates the selected ions inside the ICR cell.
The fragment ions are then excited by dipolar frequency
sweep excitation (∼132 Vp-p amplitude, 1-500 kHz at a sweep
rate of 700 Hz/µs). Fourier transformation of the resulting
discrete time-domain signal (32-512 K data, 1 MHz Nyquist
bandwidth), without zero-filling and with Hamming apodization, followed by magnitude calculation and frequency-to-mass
conversion yields an MS2 FT-ICR product ion mass spectrum.
Ultrahigh-Resolution Mass Spectra. A large number of
parent ions can be generated and trapped by optimizing the
ESI parameters. Following ion transfer to the ICR ion trap,
the electrostatic trapping potential is lowered adiabatically to
+0.6 V. A long time delay (5-50 s) before signal detection
(29) Castoro, J. A.; Wilkins, C. L. Anal. Chem. 1993, 65, 2621-2627.
(30) Solouki, T.; Pasa-Tolic, L.; Jackson, G. S.; Guan, S.; Marshall,
A. G. Anal. Chem. 1996, 68, 3718-3725.
(31) Gauthier, J. W.; Trautman, T. R.; Jacobson, D. B. Anal. Chim.
Acta 1991, 246, 211-225.
(32) Little, D. P.; Speir, J. P.; Senko, M. W.; O’Cornnor, P. B.;
McLafferty, F. W. Anal. Chem. 1994, 66, 2809-2815.
Fievre et al.
Figure 1. FT-ICR positive-ion mass spectra of Suwannee
River Fulvic Acid (SFRI) based on (a) laser desorption/
ionization (LDI) at 3.0 T and (b) electrospray ionization (ESI)
at 9.4 T.
alleviates unwanted interactions between ion clouds with
similar cyclotron frequencies.33 Conventional frequency-sweep
excitation from 1 to 500 kHz at a 600 Hz/µs sweep rate is
followed by heterodyne detection to yield 512 K time-domain
data points. Each ultrahigh-resolution FT-ICR mass spectrum
resulted from a single scan, based on direct FFT (and magnitude calculation) of the time-domain data. Singly and multiply
charged ions of peptides (e.g., bradykinin, [D-Pen2,5]-enkephalin, LHRH) and proteins (e.g., bovine ubiquitin) were used as
external and internal calibrants. In ultrahigh-resolution mass
spectra, observed mass measurement accuracy with external
calibration for all known peaks (at 0.6 V trapping voltage, over
the range 500 e m/z e 1200) was within 5 ppm of values
calculated from known chemical formulas; with internal
calibrants the mass measurement accuracy improved to within
2 ppm of the values calculated from known chemical formulas.
HPLC Fractionation. IHSS humic and fulvic acid solutions were prepared by dissolving 1 mg of dry standard in 10
mL of distilled, deionized water, producing a 100 ppm (w/v)
standard solution. This solution was then fractionated on a
high-performance liquid chromatography system consisting of
a Beckman 506 autosampler, Toso Haas TSK-6010 pump and
gradient controller, Toso Haas TSK-6041 UV detector operating at 245 nm, and a Hewlett-Packard 3394 integrator.
Separations were carried out with an Alltech Absorbosphere
Phenyl column, 5 µm particles, 25 cm × 4.6 mm i.d., with a
mobile phase consisting of 50% methanol and 50% water (v/v)
containing 1% acetic acid (v/v). This mobile phase mixture
had a measured pH of 3.5 and was pumped at a flow rate of
0.5 mL/min.
Results and Discussion
Comparison of Laser Desorption/Ionization and
Electrospray Ionization. Figure 1 shows FT-ICR
positive-ion mass spectra of Suwannee River (Florida)
fulvic acid (SFRI) obtained by (a) laser desorption (LDI)
and (b) electrospray (ESI) ionization. Clearly, these two
ionization techniques produce significantly different
mass spectra. The LDI mass spectrum is similar to
those observed by others by LDI for humic substances
(e.g., see ref 15), with ions at essentially every m/z value
(33) Solouki, T.; Emmett, M. R.; Guan, S.; Marshall, A. G. Anal.
Chem. 1997, 69, 1163-1168.
LDI and ESI FT-ICR MS of Humic and Fulvic Acids
Energy & Fuels, Vol. 11, No. 3, 1997 557
Figure 3. ESI FT-ICR positive-ion mass spectra of (a) humic
acid and (b) fulvic acid mixtures.
Table 1. Structural Characterization Based on ESI
FT-ICR Accurate Mass Measurement
no. of possible chemical formulas
within a given mass tolerance
Figure 2. ESI FT-ICR positive-ion mass spectra of fulvic acid
based on a single time-domain data set, truncated to just the
initial (a) 8K, (b) 16K, (c) 32K, and (d) 64K data before
Hamming apodization and Fourier transformation/magnitude
calculation.
in the range, 200 e m/z e 700. The ESI mass spectrum,
however, while still containing peaks throughout the
200 < m/z < 800 range, is dominated by a relatively
few prominent peaks. The higher molecular weight
species in the ESI mass spectrum are not seen in the
LD mass spectrum, suggesting that the higher-mass
species are fragmented by the LD process. Thus, it is
not safe to base molecular weight distributions of the
parent neutrals on LD mass spectra. The ESI mass
spectrum, on the other hand, holds out the possibility
of determining unequivocally the structures of at least
some of the individual compounds that make up a
“geopolymer” mixture of humic substances.
Low-energy electron impact and LDI spectra similar
to that of Figure 1a have previously been used to
compute number average molecular weights of complex
organic geochemical assemblages such as fulvic acids20
and crude oils.34 However, such analysis is based on
the assumption that all compounds in such mixtures
have equal ionization efficiencies and that each peak
represents a parent ion and not a product fragment ion.
Given the wide variation in compound types found in
such macromolecular mixtures and the complexities of
ion formation, we doubt that any mass spectrum,
including even those produced by “soft” ionization
techniques such as ESI (e.g. Figure 1b) is quantitatively
representative of the composition of a humic or fulvic
acid mixture before ionization.
A further difficulty with quantitation of a complex
geomacromolecular mixture by mass spectrometry alone
is demonstrated by the spectra of Figure 2. These
(34) DeCanio, S. J.; Nero, V. P.; DeTar, M. M.; Storm, D. A. Fuel
1990, 69, 1233-1236.
observed ion
10 ppm
5 ppm
1 ppm
543.0931 Da
559.1371 Da
575.1023 Da
49
52
58
25
33
29
6
6
6
spectra represent the time evolution of the ESI spectrum shown in Figure 1b. All of these mass spectra
were constructed from the same single scan data, by
truncating the time-domain data after 8K, 16K, 32K,
or 64K before Hamming apodization and FT/magnitude
calculation. It is clear that during the data acquisition
period, the charge states of some species shift (from
doubly-to-singly-charged) and ion-molecule reactions
change the identities and masses of some of the originally trapped ions. For example, the data truncated at
8K exhibit a prominent “hump” comprised of overlapping peaks in the 1000 e m/z e 1500 range. However,
this high-mass hump virtually disappears in the spectrum truncated after 64K scans, and is completely
removed in the 512K spectrum (Figure 1b). We infer
that reactive, high-mass ions and/or fragmens are
progressively lost through collisions and reactions with
background neutrals. Thus, although the final mass
spectrum obtained after averaging 512K scans includes
various stable ions and/or fragments, it is not a direct
representation of the initial chemical mixture.
Comparison of Fulvic and Humic Acid Mixtures.
Figure 3, a and b, shows ESI FT-ICR mass spectra of
IHSS standard fulvic acid and humic acid mixtures,
respectively. Electrospray conditions (e.g., pressure,
spray voltage, solvent, etc.) were identical for both
Figure 3, a and b). In contrast to spectra obtained with
laser desorption ionization, both ESI FT-ICR mass
spectra exhibit a relatively small number of prominent
peaks. Mass-to-charge ratios of a selected series of these
ions, under ultrahigh resolution conditions, are summarized in Table 1. Many of those ions appear in both
fulvic and humic acid mixtures, suggesting a certain
degree of molecular similarity between the two classes
of compounds. This result is not unexpected, given the
common origin of these macromolecules.
Fulvic acid mixtures, because of their greater solubility in acidic solutions, are thought to be composed of
smaller, more highly charged compounds than humic
acid mixtures. The ESI results shown in Figure 3 are
consistent with this view: the fulvic acid mass spectrum
558
Energy & Fuels, Vol. 11, No. 3, 1997
Figure 4. ESI FT-ICR positive-ion mass spectrum of fulvic
acid mixture with bradykinin and LHRH added as internal
standard mass markers (32K data points). The m/z scale
expansions clearly show singly and doubly charged gas-phase
ions, whose signals are spaced at integral and half-integral
m/z values, respectively.
contains more highly charged species than does humic
acid. For example, inspection of the data truncated at
32K show that electrospray produces singly and doubly
charged ions (see Figure 4). It should be noted that the
exact appearance of an ESI mass spectrum (e.g., variations in charge state relative abundances) varies with
the electrospray conditions and solvent.
Biomolecules such as peptides and proteins cospray
well with fulvic acid solutions (1% by volume) and
therefore are suitable as internal calibrants for accurate
mass measurement. Figure 4 shows a low-resolution
(32K time-domain data points) ESI FT-ICR positive-ion
mass spectrum of a sample mixture consisting of fulvic
acid compounds (HPLC purified) and small peptides.
Bradykinin (monoisotopic mass ) 530.787 98 Da) and
human luteinizing hormone releasing hormone (LHRH,
monoisotopic mass ) 591.794 06 Da) doubly-charged
ions are present in the mass spectrum. (The monoisotopic mass corresponds to the species for which all
carbons are 12C, all hydrogens are 1H, all nitrogens are
14N, all oxygens are 16O, and all sulfurs are 32S.) Mass
scale expansions (Figure 4, bottom) are shown for 760
< m/z < 780 and 1520 < m/z < 1540. Although massresolving power is not sufficient to resolve species of
different chemical formula, singly- and doubly-charged
fulvic acid species are clearly resolved. For example,
the higher mass ions in the range, 1520 < m/z < 1540,
are singly-charged, whereas doubly-charged ions are
evident in the range, 760 < m/z < 780. Without ejecting
most of the ions from the ion trap and ion isolation, it
is not possible to obtain an ultrahigh-resolution ESI FTICR mass spectrum for accurate mass measurement and
structural identification of these species (see time
evolution of the ESI FT-ICR mass spectra in Figure 2).
However, comparison of the mass spectra for the two
ranges suggests that at least some species may be
present both as singly- and doubly-charged ions.
ESI Spectra of HPLC-Fractionated Fulvic Acid.
Figure 5 is an HPLC fractionation of the IHSS standard
fulvic acid, SRFA. We believe that this separation is
based principally on charge, because the first peak
elutes well before the solvent (the negative, or “vacancy”,
peak), whereas the second peak is composed of solutes
which have been retained to some extent by the hydro-
Fievre et al.
Figure 5. HPLC chromatogram of SRFA fulvic acid mixture.
See Experimental Section for details of the separation. The
negative peak at ∼5.5 min is a “vacancy” peak caused by the
solvent (water).
Figure 6. ESI FT-ICR positive-ion mass spectra of a fractionated fulvic acid mixture: (a) spectrum for the first HPLC peak
eluting at 2-4 min, (b) spectrum of second HPLC peak eluting
at 5.5-6 min.
phobic stationary phase. Elution before the solvent
peak suggests that the first peak is a mixture of highlycharged molecules which, because of their hydrophilic
character, are excluded from the hydrophobic stationary
phase and the “stagnant” mobile phase associated with
it. Conversely, the second peak must be composed of
relatively hydrophobic molecules resulting from neutralization of their ionic sites through proton association,
ion-pairing, or possibly polymerization- or aggregationtype associations.
The ESI FT-ICR positive ion mass spectra of the
mixtures isolated by this HPLC procedure are presented
in Figure 6. Again, there does not appear to be any
difference in the molecular weight distributions of
observable ions in the two mixtures. Indeed, the spectra
are remarkably similar, with clusters of ions occurring
in similar ranges: 540 e m/z e 580, 620 e m/z e 680,
690 e m/z e 740, 760 e m/z e 800, and 840 e m/z e
880. The differences in HPLC behavior are therefore
most likely due to the nature of ionic sites on molecules
in each fraction. However, an explanation for the exact
molecular basis of this behavior will require more
sophisticated high-resolution and MSn experiments of
the type described in the next sections.
High-Resolution Mass Spectrometry of a Fulvic
Acid. The ESI FT-ICR positive-ion mass spectrum in
Figure 7 demonstrates the use of FT-ICR for ultrahigh-
LDI and ESI FT-ICR MS of Humic and Fulvic Acids
Energy & Fuels, Vol. 11, No. 3, 1997 559
Figure 7. Ultrahigh-resolution (resolving power, m/∆m50% >
450 000 for all observed ions) ESI FT-ICR positive-ion mass
spectrum of a selected m/z region for fulvic acid at 0.6 V
trapping voltage and 512K data points. Prior to ion detection,
SWIFT m/z-selective ejection removed the ions of 100 e m/z
e 550 and 580 e m/z e 3000.
resolution mass spectrometry and chemical formula
determination. The ultrahigh-resolution mass spectrum
was obtained at 0.6 V trapping voltage during detection;
mass measurement accuracy with external or internal
calibrants (see Figure 4) were within 5 and 2 ppm,
respectively. The mass-resolving power (m/∆m50%) for
all observed ions was greater than 450 000. From the
high-resolution data presented in Figure 7, we conclude
that the unknown fulvic acid compounds do not contain
any sulfur atoms (i.e., we do not see the expected 34S
species at ∼4% abundancee33). It is possible to identify
the chemical formulae for unknwon species based on
accurate mass measurement alone. Assuming that
fulvic acid compounds contain only carbon, nitrogen,
oxygen, and hydrogen atoms, we constructed Table 1
for three most abundant ions observed in Figure 7 (m/z
543.0931, 559.1371, and 575.102 28). For singlycharged ions, we subtracted the proton mass (i.e., 1.0073
Da) from the experimental data to obtain the mass of
the corresponding neutral molecule. Note that as the
mass tolerance is reduced, the number of possible
structures for unknown species decreases rapidly. For
example, at 5 ppm mass tolerance, 33 compounds with
various elemental compositions and mass of ∼558.1298
Da (i.e., 559.1371 - 1.0073) may be assigned to the
neutral compound. Moreover, the relative abundance
of the same molecule with one 13C in place of 12C (i.e.,
at the monoisotopic mass plus 1.003 35 Da) peak
provides a check as to the number of carbons in the
molecule. Although the unmatched ultrahigh-resolution
capability of FT-ICR instruments offers a great advantage over other conventional methods for fulvic acid
analysis, exact mass assignment of species from such
complex mixtures nevertheless requires careful mass
calibration. Final assignment of chemical formula
should include accurate mass measurement of product
ions formed by fragmentation of the parent ion (see
below).
MSn Spectra. Figure 8 depicts the series of events
used to obtain two-dimensional mass spectra (MS/MS
or MS2) of ions electrosprayed from a fulvic acid sample
mixture. These spectra were obtained from the firsteluted fraction of the HPLC-fractionated SRFA sample.
Briefly, the SWIFT technique is used to eject from the
Figure 8. Series of ESI FT-ICR positive-ion mass spectra
obtained in an MS/MS experiment. Proceeding from top to
bottom: full mass spectrum of fulvic acid mixture; SWIFT
waveform ejection from the ICR cell of ions of all but a narrow
m/z range; the resulting isolated parent ion mass spectrum,
and the product ion mass spectra produced by collisionallyinduced dissociation (CID).
Figure 9. MS/MS product ions spectra produced by (a) CID
and (b) infrared multiphoton dissociation (IR/MPD) of parent
ions of m/z 633.
ICR ion trap all ions in the original spectrum except
those in a very narrow m/z range (m/z 633 in this case).
The parent ions isolated in this manner are then
fragmented by the collisionally induced dissociation
(CID) process described in the Experimental Section.
The highest mass fragment ion at m/z 617 is the most
abundant ion at low CID energy (Ecenter-of-mass < 5 eV).
Thus, the lowest-energy fragmentation pathway (i.e.,
the major fragmentation pathway in a CID experiment)
is the loss of a small neutral from the parent ion.
Ultrahigh-resolution mass spectra and accurate mass
measurement identify the neutral lost from the parent
ion to be CH4 rather than an oxygen atom.
560
Energy & Fuels, Vol. 11, No. 3, 1997
CID and infrared multiphoton dissociation (IR/MPD)
product ion mass spectra derived from the m/z 633
parent ion are compared in Figure 9. The fragment ion
at m/z 617 observed by CID is also present in the IR/
MPD MS2 product ion mass spectrum. Figure 9 shows
that IR/MPD produces product ions of higher abundance
than obtainable by CID in this case. Moreover, the
degree of ion fragmentation is easily controlled by
changing the ion irradiation period or irradiating laser
power. However, CID provides complementary fragment ions that are not produced by IR/MPD; therefore,
complete identification of an unknown parent ion may
require both CID and IR/MPD to induce parent ion
fragmentation.
Summary
In this presentation we have demonstrated the ability
of FT-ICR mass spectrometry to produce high-resolution
spectra of humic and fulvic acid mixtures that include
stable, high-mass ions uniquely formed by electrospray
ionization. These high-resolution spectra can be used
Fievre et al.
to determine exact ion masses and thus unique molecular formulas. Clearly, unraveling the detailed structures of these molecules will require many additional
MSn experiments of the type described here. However,
it appears that ESI FT-ICR MS may now represent the
most direct approach for probing the molecular basis of
the environmental chemistry of humic substances.
Acknowledgment. HPLC fractionation and isolation of the SRFA sample was carried out by Tim Keefe.
The Suwannee River (FL) fulvic acid sample was
provided by William Davis, Department of Environmental Engineering Sciences, University of Florida, Gainesville, FL. This work was supported by the St. Johns
River (FL) Water Management District, N.S.F. (CHE93-22824), Florida State University, and the N.S.F.
High-Field FT-ICR Mass Spectrometry Facility (CHE94-13008) at the National High Magnetic Field Laboratory in Tallahassee, FL.
EF970005Q