Article pubs.acs.org/est
Molecular Level Description of the Sorptive Fractionation of a Fulvic Acid on Aluminum Oxide Using Electrospray Ionization Fourier Transform Mass Spectrometry Catherine Galindo* and Mirella Del Nero Institut Pluridisciplinaire Hubert Curien, UMR 7178 CNRS/UdS, 23 rue du Loess, BP 28, 67037 Strasbourg Cedex 2, France S Supporting Information *
ABSTRACT: We addressed here, by means of electrospray ionization mass spectrometry (ESI-MS) with ultrahigh resolution, the molecular level fractionation of a reference fulvic acid (SRFA) during its sorption at an alumina surface, taken as a model for surfaces of natural aluminum oxide hydrates. Examination of ESIMS spectra of a native SRFA solution and of supernatants collected in sorption experiments at acidic pH showed that the ∼5700 compounds identified in the native solution were partitioned between the solution and alumina surface to quite varying degrees. Compounds showing the highest affinity for the surface were aromatic compounds with multiple oxygenated functionalities, polycyclic aromatic compounds depleted of hydrogen and carrying few oxygenated groups, and aliphatic compounds with very high O/C values, highlighting the fact that SRFA constituents were sorbed mainly via chemical sorption involving their oxygenated functionalities. We observed an inverse correlation between the degree of sorption of a molecule within a CH2 series and its number of CH2 groups and a positive correlation between the degree of sorption and the number of CO2 groups in a COO series, which was remarkable. These correlations provide evidence at the molecular scale that molecule acidity is the key parameter governing fulvic acid (FA) sorptive fractionation, and they are useful for predicting sorption of FA at a natural oxide surface.
■
INTRODUCTION Fractionation of humic substances (HS) during their sorption at surfaces of mineral particles is a main issue in several domains of geosciences relating to the environment, including storage of organic matter in soils and the fate of metals or organic contaminants in surface systems.1,2 HS like humic acids and fulvic acids (FA) are ubiquitous in all terrestrial and aquatic surface systems as complex mixtures of organic compounds originating from various processes of degradation of plants and living organisms. They display a great variety in their composition and constituents because of their different origins and show a high degree of heterogeneity in the nature, structure, and reactivity of their compounds and functional groups, which is responsible for the preferential retention of certain HS compounds at mineral surfaces. Sorptive fractionation of HS is expected to determine the main characteristics of surface geochemical systems, including the C content of soil and water, the chemical nature of dissolved HS, which influences aqueous speciation and the mobility of metals, and the chemical nature of sorbed HS molecules, which controls the surface properties, dissolution rates, and contaminant sorptive capacities of soils, sediments, and colloidal phase. Therefore, identifying specifically the molecules that constitute HS and the mechanisms involved in HS sorptive fractionation is a scientific and technical challenge essential for the improvement of our understanding and modeling of geocycling of organic matter and metallic/organic contaminants. © 2014 American Chemical Society
Much knowledge at the molecular scale of the structural features and of the chemical nature of HS has been acquired because of recent advances in analytical chemistry, particularly in the field of ultrahigh-resolution mass spectrometry combined with electrospray ionization (ESI). HS, which were described traditionally as long-chain cross-linked molecules, are at present envisioned within the supramolecular assembly concept developed first by Piccolo and co-workers and substantiated by ESI-MS studies.3,4 The structural composition of humic solutes is described by the association of small molecular size entities that are held together by weak hydrophobic forces and hydrogen bonding and that possess similar structural functionalities and an average mass no higher than a few kilodaltons. Knowledge of the chemical nature of HS, which was previously limited to “average” characteristics as determined by elemental and spectroscopic analyses of bulk materials, has been extended by use of ultrahigh-resolution MS to the specific knowledge of the nature and chemical formula of several thousand constitutive molecules.5−7 Humic acids or fulvic acids are no longer depicted just as a combination of functional groups (carboxyl, alcohol, etc.), but as a set of distinct polycarboxylates with a specific molecular mass, Received: Revised: Accepted: Published: 7401
March 27, 2014 June 4, 2014 June 6, 2014 June 6, 2014 dx.doi.org/10.1021/es501465h | Environ. Sci. Technol. 2014, 48, 7401−7408
Environmental Science & Technology
Article
for the surface was derived from mass spectrometry analysis of FA solutions before and after sorption, using a LTQ Orbitrap XL spectrometer that provides a high resolving power in low and intermediate mass ranges (better than 100000 at m/z 400). Attention was paid to clarifying the relations existing between chemical characteristics of a molecule in a FA mixture, i.e., degree of aromaticity, acidity, functionalities, hydrophobic character, and molecular weight, and its partitioning between the solution and oxide surface. The results presented here provide the first published description at the molecular scale of the sorptive fractionation of FA, which will aid in understanding and building a more realistic model of HS−mineral surface interactions occurring in surface environments.
elemental composition, and degree of aromaticity and with limited numbers of hydroxyl or other functional groups.8−10 In contrast to the current knowledge of the chemistry of HS at the molecular level, no description of the sorption of these complex mixtures at the molecular scale exists, so that little is known about their sorptive fractionation. In many sorption studies, HS were regarded as bulk materials bearing polar groups, and many efforts were made to identify the types of functionalities involved in the macroscopic uptake of the bulk sample. Studies using Fourier transform infrared (FTIR) spectroscopy11,12 and nuclear magnetic resonance (NMR) spectroscopy13 have led to the conclusion that, under acidic conditions, the sorption on iron and aluminum oxides mostly occurs via the carboxyl functionality of NOM. ATR-FTIR spectroscopy and near-edge X-ray absorption fine structure spectroscopy (NEXAFS) analyses have provided evidence of the implication of phenol moieties in Suwannee River fulvic acid−alumina surface interactions at higher pH values.12,14 The involvement of both carboxyl and hydroxyl groups was postulated by Filius and co-workers15 to adequately fit macroscopic data on the sorption of a fulvic acid on goethite at near-neutral pH by using the CD MUSIC surface complexation model. On the basis of model compound behavior in sorption experiments and of acid titration of sorbing and nonsorbing organic molecules, Edwards et al.16 predicted that very strongly acidic groups (acid groups ionized at pH 3.0) are the key to the formation of strong surface complexes between natural organic matter and Fe oxide surfaces. In some other works, however, the sorptive fractionation of HS was studied. The effect of molecular weight was investigated by size-exclusion chromatography, and systematic changes in average molecular weight of HS were observed after the sorption, with the higher-molecular weight components sorbing preferentially at the metal oxide surface for aquatic NOM17−20 and the low-molecular weight fraction being more prone to sorption at the hematite surface for terrestrial Aldrich HA.21 Some spectroscopy studies focused on the sorption behavior of distinct classes of HS constituents. McKnight et al.22 showed that hydrophobic fractions and aromatic components carrying carboxyl groups display a higher affinity for Fe oxide surfaces than aliphatic fractions. Kaiser23 highlighted that aromatic rings carrying carboxyl or other polar functional groups are the structural units involved in the sorption process while the aromatic structure, by itself, has little influence. Reiller et al.21 analyzed by mass spectrometry terrestrial Aldrich HA solutions before and after sorption on hematite at neutral pH and deduced from MS data analysis that small aromatic structures with oxygen functional groups were more prone to sorption. The authors underlined that further analysis using a high-resolution spectrometer should be envisaged to determine the elemental compositions of sorbing and nonsorbing HS compounds and to ascertain or discuss their conclusion. From this short review, it appears that further work is needed for a detailed understanding of molecule characteristics and mechanisms governing the HS sorptive fractionation. The aim of this work was to clarify the main mechanisms and chemical characteristics governing the sorptive fractionation of fulvic acids at an aluminum oxide surface, by identifying specifically sorbing and nonsorbing organic molecules for a reference FA (Suwannee River FA, SRFA)−aluminum oxide system. The exact elemental composition of molecules quantitatively sorbed at the surface or showing a low affinity
■
EXPERIMENTAL SECTION Materials. Alumina colloids were Alfa Aesar’s α-Al2O3 crystallites (chemical purity of 99.95%, surface area of 7.6 m2/g,) with a particle size of 280 ± 20 nm as determined by applying the Contin algorithm to photon correlation spectroscopy measurements. A native solution (250 mg/L) of fulvic acid was prepared by using a SRFA reference acid supplied by the International Humic Substance Society (IHSS) and ultrapure water. The fulvic acid was isolated from the Suwannee River in Georgia by the IHSS standard method for extraction and isolation of aquatic humic substances.24 IHSS has adopted the XAD-8 resin adsorption method to isolate fulvic acids from natural waters, so that the SRFA sample does not contain the hydrophilic fraction. Sorption Experiments. Batch experiments for the sorption of SRFA (250 mg/L) onto alumina were conducted in HDPE tubes using 25 g/L α-Al2O3 suspensions at acidic pH and 298 K. No attempt was made to maintain the pH during the experiment or to fix the solution ionic strength (no background electrolyte addition), as the addition of salt was observed to considerably alter the ESI-MS response. After the tubes have been shaked in the dark for 2.5 h, the suspension was centrifuged at 8000 rpm over 2.5 h for separation of the solution and colloid (cutoff of 15 nm). The final pH of the supernatant was measured to be equal to 4.0. The final ionic strength was estimated to be ∼10−4 M. For electrospray ionization Fourier transform mass spectrometry (ESI-FTMS) analysis, the pH of supernatant was adjusted to 3.4 ± 0.1 (via addition of HCl), which is the pH of the SRFA native solution. The concentration of FA in the supernatant was determined by total organic carbon measurements using colorimetric detection at 610 nm after persulfate digestion (Hach method). ESI-FTMS Measurements. A native solution of SRFA and the supernatant were analyzed by ESI-FTMS in the negative ionization mode by using a linear ion trap Orbitrap mass spectrometer (LTQ Orbitrap XL, Thermo Scientific). The solutions were introduced directly into the ESI probe with a syringe pump, at a flow rate of 10 μL/min. Nitrogen was used as the drying and spraying gas. The temperature of the transfer tube was fixed at 275 °C. A spray voltage of 3.7 kV was applied. The voltage applied to the capillary, the multipoles, and the tube lens was automatically tuned to favor transmission of highmolecular weight ions. MS spectra were recorded using the Orbitrap analyzer, by averaging 100 scans in the ranges of m/z 120−2000 and 400−2000. Acquisition and treatment of data were conducted using Xcalibur software (version 2.1.0, Thermo Sientific). Only peaks whose intensity exceeded twice the signal-to-noise ratio were considered as peaks of the constitutive SRFA compounds. 7402
dx.doi.org/10.1021/es501465h | Environ. Sci. Technol. 2014, 48, 7401−7408
Environmental Science & Technology
Article
Figure 1. ESI-FTMS spectra obtained in the ranges of m/z 120−2000 and 490−1400 for the SRFA native solution (a and c, respectively) and for the supernatant (b and d, respectively) collected after sorption of SRFA on alumina colloids at pH 4, [SRFA]T = 250 mg/L, 25 g/L Al2O3, and a reaction time of 2.5 h. Insets are expanded sections of the ESI-MS spectra.
shows that the m/z values of the SRFA constituents do not exceed 980. ESI has the ability to generate multiply charged species, so that determination of the charge state of all detected ions is required for accurate mass assignment. When the resolving power of the MS analyzer is sufficient, the charge state of an analyte can be deduced from spacing between adjacent signals, with the spacing between the monoisotopic peak and the peak including one 13C isotope of 1.0034/z. The ESI-MS data sets for SRFA showed that each primary ion has a corresponding lower-abundance ion at m/z m + 1.0034. There was no series with mass distances related to a higher charge state. Hence, as previously noted for dissolved organic matter,6 the isotopic distribution of carbon in SRFA is consistent with the formation in ESI of singly charged constituents only. The molecular weight of the species detected in SRFA hardly exceeded 980 Da, consistent with the supramolecular assembly concept of Piccolo and co-workers,3,4 which holds that the structural composition of humic solutes resembles supramolecular-type associations of small molecules. To sort molecules on the basis of their elemental compositions, the value of the Kendrick mass defect was plotted against the nominal Kendrick mass (cf. Experimental Section) and -CH2- and -COO- homologue series were considered. The Kendrick plots highlighted the fact that SRFA compounds lie in long and relatively large bands (Figure S1 of the Supporting Information). The series of increasing CH2 units include up to 15 components, suggesting that CH2 groups are building blocks of SRFA. The COO homologue series contain at most nine components. The heterogeneous nature of the fulvic acid is substantiated by the multitude of classes of compounds (numerous vertical lines on the Kendrick mass defect plots). Taking advantage of the high resolution and high mass accuracy provided by the Orbitrap XL analyzer, a chemical formula was ascribed to 5727 peaks among the 8437 peaks detected on the ESI-FTMS spectra of SRFA. Our data assignments compare well with results of the FTICR study of Stenson et al.7 and the Q-Tof work of These et al.8 on the characterization of SRFA. All of the nine formulas derived from our Orbitrap FTMS measurements for the isobaric anions at the nominal mass of 410 Da are the same as those of Stenson et al.7 More than 95% of the chemical formulas given by These et
Kendrick mass defect (KMD) analysis of MS data sets was undertaken to find series of molecules that are separated by integer multiples of a given functional group7 (see Figure S1 of the Supporting Information). For chemical formula assignments, the Xcalibur MS calculation software was used. Odd m/z values correspond to even masses and relate to chemical formula with an even number of nitrogen atoms. As the content of nitrogen in SRFA does not exceed a few percent, we assumed that compounds peaking at odd m/z values primarily contain C, H, and O atoms. All possible formulas that can be attributed to a given odd m/z value were calculated by considering 12C, 1H, and 16O (with upper limits for numbers of atoms of 200, 600, and 50, respectively) and by rejecting all formulas whose theoretical mass differs ≥3 ppm from the measured mass. For even m/z values (odd masses), peaks whose positions coincided with those of species including one 13C isotope have been excluded. Because of the low nitrogen content of HS, residual species were considered to contain at most one 14N atom, in addition to C, H, and O atoms.
■
RESULTS AND DISCUSSION
Analysis of the Native Solution of SRFA by ESI-FTMS. First, it is to be noted that, in this study, SRFA was dissolved in 100% water to prevent the self-esterification of the constituents, which was reported in a previous ESI-MS study of humic substances prepared by dissolving HS in methanol or other alcoholic solvents.25 Figure 1a presents the full mass spectra recorded over the mass range of m/z 120−1400, using the negative ionization mode, for the native solution of SRFA. The spectrum reveals a bimodal distribution of peaks, with the first one including only a few components and lying between m/z 120 and 145 and the second one between m/z 145 and 1000. ESI-MS analysis confirmed that SRFA is a very complex mixture, because 8437 peaks were detected. The spectrum displays a typical pattern, with a first series of intense peaks at odd-integer m/z values and a second (of lower peak intensity) at even-integer m/z values. Insets in panels b and c of Figure 1 show, as an example, expanded sections around m/z 387 and 751, respectively, for the spectrum described above. Dozens of peaks are distinguishable at the same nominal mass, because of the high resolving power of the Orbitrap analyzer. The ESIFTMS spectrum recorded above m/z 500 (Figure 1c) also 7403
dx.doi.org/10.1021/es501465h | Environ. Sci. Technol. 2014, 48, 7401−7408
Environmental Science & Technology
Article
al.8 from SEC/Q-Tof-MS data in the range of m/z 190−340 were recovered in the work presented here. This is evidence of good self-consistency among different ESI-MS measurements and provides confidence in chemical formula assignments. The Orbitrap ESI-FTMS data sets recorded for SRFA were analyzed using the Van Krevelen (VK) diagram that cross-plots hydrogen/carbon and oxygen/carbon atomic ratios of compounds. The VK diagram is a useful tool that provides a clear and concise vision of the distribution of the constituent compounds among condensed aromatic structures, aromatic structures, and compounds having an aliphatic character. The SRFA compounds were classified in these three categories on the basis of the work of Koch and Dittmar,26 who proposed a mathematical formula for calculating the aromaticity index (A.I.) of a molecule (assuming that half of the oxygen is σbound), and two threshold values of A.I. as unequivocal and minimal criteria for the existence of condensed aromates (A.I. ≥ 0.67) or aromates (0.5 < A.I. < 0.67). The molecules with an A.I. of ≤0.5 thus have an aliphatic character that is more or less pronounced (aliphatic compound or compound carrying many and/or long aliphatic side chains). Figure 2 shows that the
authors cited above who used the Orbitrap analyzer27 have discarded all compounds whose intensity was below a value of relative magnitude cutoff set at 10.2, which probably led to excluding PAC (those detected in the work presented here exhibited a signal-to-noise ratio between 2 and 8). It seemed important to us not to exclude this class of compounds when characterizing the SRFA sample, because these compounds, even if present at low concentrations, are likely to affect significantly the sorptive fractionation of SRFA and the speciation of trace metals. It is recognized that different classes of biogenic compounds, e.g., lipids, carbohydrates, lignins, proteins, and condensed hydrocarbons, tend to be plotted in specific areas of the diagram, although the H/C and O/C elemental ratios are not always sufficient for structural assignment.28 For SRFA, there were few formulas matching those of proteins, carbohydrates, and lipids (H/C ratio of >1.5). This is consistent with the data of IHSS showing that SRFA contains, in particular, only a small amount of sugars. Nevertheless, aliphatic compounds may be under-represented in our ESI mass spectra as, in the ESI probe, hydrophilic species are hardly transferred into the gas phase in the presence of hydrophobic species.29 A 13C NMR spectroscopy study has shown that tannins and terpenoids are major precursors of SRFA constituents.30 Therefore, besides PAC, the major portions of the compounds that can be observed in Figure 2 are likely condensed tannins (O/C ratio of 0.23 and O/C ratios of >0.02 (bottom left) to compounds of aliphatic character with H/C ratios of
