Assessing Redox Properties of Natural Organic Matters with regard to Electron Exchange Capacity and Redox-Active Functional Groups

Redox processes in groundwater play an important role in bioavailability, toxicity, and mobility of redox-active elements and contaminants. A recent study has demonstrated that low-molecular-weight fraction (LMWF) of humic substances with great number of redox-active functional groups (RAFGs) exhibits great reducing capacity. However, whether LMWF of natural organic matter (NOM) exhibits high redox capacity still remains unclear. ,erefore, this study extracted Pahokee peat NOM (PPNOM) and Leonardite NOM (LNOM) from soils, and then LMWFs in these NOMs were collected using a dialysis method. Electron exchange capacities (EEC) and RAFGs of LMWFNOMs at different Eh were analyzed using a novel electrochemical method and a three-dimensional excitation emission fluorescence (3DEEM) spectroscopy. We found that the reducing capacity in LMWF PPNOMwas approximately 5-6 times higher than the bulk NOM, while only 7.8% LMWF PPNOMwas accounted for in the bulk NOM. An increasing in EEC (EAC+EDC, where EAC is the electron accepting capacity and EDC is the electron donating capacity) of LMWF PPNOM and LNOM with Eh reduced from −0.49V to −0.69V. Additionally, an obvious increase in fluorescent intensities of quinone-like fluorophores before and after being reduced LMWF LNOM is responsible for high EAC of LMWF LNOM. ,ese findings provide a better understanding of relationship between RAFGs Eh in LMWF of NOM, further helping in predicting and protection of groundwater environment and fate of transformation and transport for redox-active contaminants in groundwater.

Quinone/hydroquinone moieties are generally considered to be the main redox-active functional groups of natural organic matter [26]. In general, NOM can be chemical and microbial reduced, and reduced NOM facilitates electron transfer-stimulating microbial Fe [III] minerals. e extent and rate of stimulated microbial Fe [III] reduction is relating to redox property of NOM. Reducing capacities of NOM include electron accepting capacity (EAC) and electron donating capacity (EDC). NOM can be chemically reduced by H 2 in the presence of a Pd catalyst resulting in the different reducing capacities before and after being reduced. [25,26]. Alternatively, the electrochemical reduction is a more efficient method compared to the chemical reduction [10,27]. Previous electrochemical studies have revealed redox properties of NOM in varied redox potential range of −0.9 V-+1.0 V, which significantly affects transformation and electron transfer rates of pollutants in soil environments [28].
Many redox-active contaminants (such as arsenic, phosphate, and sulfate) normally are stocked in these soil nanoscale pores that have limited access to microbes [3,21,23,29,30], leading to hard transformation and transport, thus substantially arising a risk in soil groundwater health. It has been demonstrated that low-molecular-weight fractions (LMWFs) released from bulk humic substances (HSs) can exhibit great reducing capacity. e high reducing capacity of LMWF HS is attributed to LMWF HS, possessing a high amount of quinone-like fluorophores detected by threedimensional excitation-emission matrix (3DEEM) fluorescence spectroscopy [26]. ese LMWF HSs flowing into and/ or out of the soil micropores (<2 nm) are able to significantly impact fates of transport and transformation of redox-active contaminants [26]. However, still opening questions including whether LMWF exists in natural organic matter (NOM), to which extent of electron transfer capacities of LMWF in NOMs at varied E h in soil environment, and relationship of its RAFGs and E h has paid a little attention. erefore, the aims of this study were (i) to collect different LMWFs (3500 LMWF and 14000 LMWF) and retentate from bulk NOM by means of a dialysis method with dialysis bags of molecules cutoff of 3.5 and 14 KDa; (ii) to determine electron exchange capacities (EECs) including electron accepting/donating capacity (EAC/EDC) at a varied E h range of 0-−0.69 V of LMWF, retentate, and bulk Pahokee peat and Leonardite natural organic matters (PPNOM and LNOM) using chemical (by H 2 and Pt/Al) and electrochemical analysis; and (iii) finally to detect redox-active functional fluorophores by 3DEEM for different molecular weight fractions of NOM at native and varied reduced E h state (−0.49-−0.69 V). We expected to develop a potential relationship between EEC and its RAFGs for different LMWF NOMs according to gradient E h range of −0.49-−0.69 V to better understand and predict fates of electron transfer process of LMWF NOM during the biogeochemistry process in soil and groundwater environments.

Sample Preparation and Dialysis Experiments.
Pahokee peat soil and Leonardite soil were purchased from the International Humic Substances Society (IHSS). PPNOM and LNOM were extracted from Pahokee peat soil and Leonardite soil samples (3.85 g/L) that dissolved in a phosphate buffer solution (PP buffer; 50 mM, pH 7) to achieve final approximately total organic carbon (TOC) contents of 250 mg C/L. e NOM samples in sealed glass bottles with aluminum foil wrap were in continuous stirring for 5 days (200 rpm) to ensure fully release.
Dialysis bags were purchased from MWCO Regenerate (VISKASE, US). Before dialysis experiments, the chosen dialysis bags, MD34-3500 (pore size 1.25 nm) and MD34-14000 (pore size 2.5 nm), were prior soaked in deionized water for 24 hours. e dialysis bag containing 50 mL bulk sample was submerged in beaker with 200 mL PP buffer, and the whole device was wrapped in aluminum foil to prevent photochemical reaction. At the same time, the stirring rate was set to 200 rpm to simulate flowing rate of water in underground environment. e TOC (mg C/L) was determined by TOC-V scan analyzer (SHIMADZU).

Electrochemical Analysis with Direct Electrochemical Reduction (DER).
All electrochemical experiments were conducted in an anoxic glove box (100% N 2 atmosphere at 25 ± 1°C). An electrolysis cell consists of (i) a platinum disk working electrode (CHI102), (ii) an Ag/AgCl reference electrode (CHI111), and (iii) a Pt wire counter electrode (CHI115, all from CH Instruments, Inc. US). e current response I was determined by CH Instruments 660E (Austin, TX). All samples were vacuumed for 3 mins, followed by purging with N 2 for 3 mins three times before electrochemical experiment. 0.1 M KCl was used as supporting electrolyte during electrochemical experiment. Each of 4 mL sample in an electrolysis cell was directly electrochemical reduced at an electrode potential of −0.49 V, −0.59 V, and −0.69 V, respectively.

Electrochemical Analysis with Direct Electrochemical
Oxidation (DEO). All setups of DEO experiments completely followed DER experiments instead of the +0.61 V applied potential in DEO. e number of transferred electrons Q DER and Q DEO (μmol) was calculated by an integration of the reductive and oxidative currents I over the time t (s), respectively [10]. Considering the differences in different molecular weight fractions based on TOC contents, EEC (mmol e−/g C) of each sample was normalized to TOC: where F is a Faraday constant (96485 (C/mol)) and V (mL) is a volume of reaction samples. e EDC and I-t curve in DEO experiments of different molecular weight fractions of PPNOM and LNOM are shown in Figures S1 and S2 in supporting information.

Chemical Reduction.
e samples were chemically reduced by H 2 with Pd/Al catalysts (palladium-coated aluminum pellets, 0.5% Pd, Merck) as chemically reduced samples. e reducing capacities of all samples before and after reduction were determined as previously described by Kappler et al. [25] and Yang et al. [26].

3DEEM.
Fluorescence spectra of all samples before and after reduction were characterized by an F-7000 fluorescence spectrophotometer equipped with both excitation and emission monochromators (Ushio Inc. Japan) with a 1500 W Xenon arc lamp as the excitation source. e fluorophores were obtained in the form of 3DEEMs via collecting a series of emission spectra over a range of excitation wavelengths, in which fluorescence intensity was presented as a function of excitation wavelength on one axis and emission wavelength on the other. To prevent reduced samples from reoxidizing, the standard 10 mm four-way quartz cell was prior purged with N 2 for 30 mins before analysis. e fluorescence spectra were analyzed as previously described by Yang et al. [26].

Dialysis Process and TOC Contents.
Dialysis experiments were conducted within 5 days to collect different molecular weight fractions (3500 LMWF, 14000 LMWF, 3500 retentate, and 14000 retentate from bulk PPNOM and LNOM), and the results on total organic carbon TOC content among different LMWFs are shown in Table 1. e TOC results of LMWF NOM showed a continuous increase over the dialysis process (not shown), which are in agreement with our previous studies that the continuously released LMWF HS results in an increase in TOC content [19,26]. Approximately 4.1-7.8% LMWF NOMs of bulk LNOM and PPNOM, respectively, were obtained at the end of dialysis, which was obviously higher than only 2% LMWF HA of total bulk PPHA and LHA shown in our previous study. e results suggested that NOM owned high LMWF NOM, and these LMWF NOMs occupied higher carbon contents than HA. Moreover, we found that, for both PPNOM and LNOM, TOC contents of 14000 LMWF were higher than 3500 LMWF, which suggested that 14000 LMWF NOM had more C functional groups (i.e., some electron donor and acceptor moieties: phenolic and possibly substituted aromatic carboxylic acid groups, and carbonyl-containing groups. And the number or type of RAFGs of NOM such as quinone/ hydroquinone moieties [14]) that were higher in 14000 LMWF than 3500 LMWF, but not much.

Direct Electrochemical Oxidation (DEO) and Reduction (DER) of Different Molecular Weight Fractions of PPNOM and LNOM.
Quantified electron exchange capacities (EECs) including electron accepting capacity (EAC) and electron donating capacity (EDC) of different molecular weight fractions of PPNOM and LNOM were determined by electrochemical methods. e current I of different molecular weight fractions of NOM in DER and DEO experiments is shown in Figure 1 and S1, respectively. No obvious current I was detected in DEO experiments for all molecular fractions in DEO, which suggested that EDC of all LMWF, retentate, and bulk PPNOM and LNOM had no significant contribution to EEC. e I-t curve results of DER experiments showed a similar decreased tendency in current I from 77.46 μA to 11.14 μA of 3500 LMWF PPNOM/LNOM in −0.59 V DER within 50 min reduction until complete final reduction. Moreover, the I-t curve of DER experiment showed a decreased tendency for all molecular fractions of both PPNOM and LNOM with a decreased E h from −0.49 V to −0.69 V.
ese results showed EAC of PPNOM and LNOM were related to environment E h . Electron accepting moieties were distributed and concentrated in low E h .
In addition, we found that the current I of DER experiments in LMWF molecular weight fractions showed a sharp decrease compared to that of bulk NOM with a decrease in E h range. Bulk NOM showed a relatively higher Q DER than LMWF and retentate NOM due to a high TOC in bulk NOM.
is result indicated that EAC and EEC of different molecular weight fractions of NOM were highly dependent on its TOC contents.
In order to evaluate total EEC among different molecular weight fractions of NOM, EEC (EEC � EAC + EDC) of different molecular weight fractions was normalized to its TOC contents, referring to the number of electron transferred per gram carbon (meq e−/g C), as shown in Figure 2. Since no obvious contribution of EDC of NOM to EEC, the EAC was approximately equal to EEC according to equation (5). e EEC was significant higher in LMWF than in retentate and bulk NOM, suggesting a high electron transfer capacity of LMWF NOM. is is consistent with the previous study reported that the LMWF HA has a great reducing capacity using chemical reduction (H 2 and Pd/Al) [26]. Moreover, a small current I responded a low EEC in the low reducing potential of −0.49 V, while a large current I responded a high EEC in high reducing potentials. e EEC of NOM in reducing potential range of −0.59 V-−0.69 V accounted for up to 90% EEC in reducing potential range of −0.49 V-−0.69 V (Figure 2). ese results indicated that   electrons rather than donate electrons; in particular, electron accepting moieties of PPNOM were fully distributed within E h range of −0.59 V-−0.69 V. Reducing capacities of different molecular weight fractions of PPNOM were evaluated by chemical reduction (H 2 and Pt/Al), as shown in Figure 3. Reduced PPNOM showed a higher reducing capacity than that of native PPNOM, indicating that PPNOM possessed more oxidized functional groups, which is in agreement with the results shown by electrochemical analysis. LMWF PPNOM displayed a higher EAC (difference in reducing capacity between reduced state and native state) than retentate and bulk PPNOM ( Figure S3), indicating that LMWF had more electron accepting moieties compared to retentate and bulk NOM.
In order to investigate the electron accepting moieties in different NOM samples, we calculated the proportion of EEC in different LMWFs and retentate of PPNOM and LNOM at different E h of −0.49 V-−0.59 V, as shown in Table 2. Calculated EEC of PPNOM accounted for around 3.5-58.4% of total EEC for both LMWF and retentate. e result suggested that LNOM, compared to PPNOM, has higher amount of electron accepting moieties. In particular, the number of electron accepting moieties in LMWF LNOM obviously higher than in LMWF PPNOM occupied 3.5-5.4% of LNOM EEC at E h of −0.49 V. A decrease in E h from −0.49 V to −0.59 V was observed, which resulted in an obvious increase in proportion of LMWF of PPNOM and LNOM from 3.5-5.4% to 26.4-44.9% (Table 2). ese findings indicated distribution of electron accepting moieties of LMWF NOM is dependent on its E h range. In this study, we demonstrated that LMWF LNOM with great electron accepting moieties is responsible for a high EEC and EAC.

Characterization of Redox-Active Functional Fluorophore of LMWF NOM at Different Reducing Potentials by 3DEEM
Spectroscopy. In order to clarify responsibility of redoxactive functional moieties/fluorophores of LMWF LNOM to a high electron accepting capacity during the electron transfer process, RAFGs of LMWF LNOM were detected before and after reduction in different E h ranges of −0.49-−0.59 V using 3DEEM spectroscopy. Overall, in the native state, 3500 LMWF LNOM had a higher relative fluorescent intensity of humic-like fluorophore with quinonoid π-π * transition compared to 14000 LMWF ( Figure 4 and Table 3). ese results suggested that 3500 LMWF occupied more humic-like fluorophores than 14000 LMWF LNOM. Additionally, compared to 14000 LMWF, emission wavelength (Em) of 3500 LMWF showed a shorter Em wavelength position accompanied with a small blue-shifting (ΔEm � 5 nm, to a shorter wavelength; Figure 4 and Table 3). is result implied that, in the native state, 14000 LMWF had a more complicate structure of redox-active functional fluorophores than 3500 LMWF.
In reduced state with −0.49 V potential, compared to native LMWF LNOM, Em of both 3500 LMWF and 14000 LMWF LNOM shifted to a longer Em wavelength (ΔEm � 5 or 10 nm; Figure 4 and Table 3) due to a quinonoid π-π * transition particularly. ese observations indicated that, in reduced state of −0.49 V, functional groups of LMWF changed to a more complicate structure which most likely to high level condensation in functional groups. Moreover, in reduced state of −0.59 V, Em still slightly shifted to longer wavelengths (ΔEm � 5 nm) for 14000 LMWF and 3500 LMWF compared to native 14000 LMWF and 3500 LMWF and even to LMWF samples in reduced state of −0.49 V (Figure 4 and Table 3). ese results indicated that a decrease in E h from −0.49 V to −0.59 V for LMWF LNOM could lead to an increase in aromatic condensation of redox-active functional groups due to quinonoid π-π * and benzenoid   e reducing capacity of different molecular weight fractions of PPNOM in their native and reduced states. Each reduced sample was prepared by H 2 in the presence of two Pd catalysts. e reducing capacities were determined by the difference of Fe[III] before and after reacting with ferricyanide. Each data were normalized into meq e − /g C, which is equal to mmol e − /g C in numerical value.
π-π * . In addition, after being reduced to E h range of −0.49 to −0.59 V, 3500 LMWF had a higher relative fluorescent intensity compared to 14000 LMWF LNOM.
is finding suggested that 3500 LMWF possessed a relatively low aromatic condensation but high number of quinone-like fluorophore, which was consistent with the result in respect to high EAC in 3500 LMWF LNOM using electrochemical analysis.

Conclusions
Redox processes impact transformation and transport of redox-active contaminants in environment, which is dependent on E h and RAFGs of NOM. We evaluated the redox properties of NOM with respect to EECs and RAFGs. We found that the reducing capacity in LMWF PPNOM was approximately 5-6 times higher than that of the bulk NOM,   ese quinone-like fluorophores with a redshifting in emission wavelength were observed with a reduced E h range from −0.49 V to −0.59 V. e findings can provide a better understanding of relationship between RAFGs E h in LMWF of NOM and thus can help to predict fates of redox-active contaminants in LMWF NOM-containing groundwater environment.

Data Availability
e experimental data used to support the findings of this study are included in the article. And more detailed data are available from the corresponding author upon request.

Conflicts of Interest
e authors declare that there are no conflicts of interest regarding the publication of this paper.