Introduction

Redox properties of humic substances (HS) have manifold implications in biogeochemical processes in the environment (Peretyazhko and Sposito 2006; Macalady and Walton-Day 2011; Sposito 2011; Pipenbrock et al. 2014). A few examples are enhancement of Fe(III) bioreduction by electron shuttling (Nevin and Lovley 2000; Kappler et al. 2004; Rakshit et al. 2009; Pipenbrock et al. 2014), increasing the reduction rate of substituted nitrobenzene (Dunnivant et al. 1992; Hernandez-Monotoya et al. 2012), dechlorination of aliphatic chlorinated pollutants (Kappler and Haderlein 2003), and oxidation/reduction of arsenic (Jiang et al. 2009). All the above processes are known to occur because of the ability of HS to donate and accept electrons easily or in other words the reversibility of the functional groups of HS to get oxidized and reduced (Weber et al. 2006). Quinone and other aromatic moieties of HS have been identified as sole contributors of electron transfer processes (Scott et al. 1998; Chen et al. 2003; Ratasuk and Nanny 2007; Hernandez-Monotoya et al. 2012).

Quantification of redox properties of HS has recently gained enormous importance (Adhikari et al. 2016; de Melo et al. 2016; Yang et al. 2016; Pipenbrock et al. 2014; Klüpfel et al. 2014; Borch et al. 2010). Several operational procedures coupled with spectroscopic tools have been used for this purpose (Aeschbacher et al. 2010; Ratasuk and Nanny 2007; Kappler et al. 2004; Matthiessen 1995). However, owing to the complexity in structure and functional groups of HS, a general consensus regarding the most applicable methods to characterize redox properties of HS has not yet been reached. While discussing redox properties of HS, it should be noted that we will use the terminology of Peretyazhko and Sposito (2006). They defined native reducing capacity (NRC), chemical reducing capacity (CRC), and microbial reducing capacity (MRC) of HS as reducing capacity of HS that did not have any reduction pretreatment, chemically reduced HS that have undergone chemical pretreatment in the laboratory (usually by H2 gas in the presence of Pd catalyst at pH = 7), and microbially reduced HS that have been reduced by microbes, respectively. There is generally a good agreement between CRC and MRC. Thus, CRC became a standard surrogate for assessing electron transfer capacities of HS in the environment (Peretyazhko and Sposito 2006).

There are various methods available to assess CRC of HS. One of the most common methods is reduction of HS dissolved in phosphate buffer at pH 7 using H2/Pd (Visser 1964; Kappler et al. 2004) followed by incubation with ferric citrate for 15 min. Then, the suspension is extracted by 0.5 M HCl, and Fe(II) is measured colorimetrically by UV–Vis spectrophotometer at 562 nm by complexation with ferrozine reagent (Stookey 1970). There are a few assumptions made here. First, H2/Pd will cause complete reduction of all functional groups of HS responsible for electron transfer, and second, ferric citrate will react with all reduced functional groups of HS within 15 min. These assumptions have been questioned by Ratasuk and Nanny (2007) and Aeschbacher et al. (2010). Ratasuk and Nanny (2007) found that depending on pH values and the presence of neighboring functional groups (i.e., either electron donating or electron withdrawing) near the quinone groups of HS, the H2/Pd reduction method can destroy the quinone moieties in HS. They supported their hypothesis using several model compounds such as anthraquinone-2,6-disulfonate (AQDS), anthraquinone-2-carboxylate (AQC), 2,6-dihydroxyanthraquinone (AQOH), which have both electron-donating and electron-withdrawing neighboring groups close to quinone moieties. Aeschbacher et al. (2010) indicated that 15 min incubation by ferric citrate was not sufficient to measure all the reducible functional groups of HS. Therefore, both assumptions one and two may underestimate the CRC measured by this method. However, although the kinetic limitation of the second step of the process was demonstrated clearly by Aeschbacher et al. (2010), the limitation of H2/Pd reduction method, indicated by Ratasun and Nanny (2008), at variable pH, and the presence of different functional groups near quinone moieties were not obvious from the UV–visible and IR spectroscopic evidences. This is because the quinone peaks are known to shift in the blue or red regions in the presence of both electron-withdrawing and electron-donating neighboring groups at different pH values (Gautrot et al. 2007; Diaz 1990; Peters and Sumner 1953). Therefore, consideration of 290-nm peak as quinone for all the model compounds such as AQOH, AQC, 1,4-naphthoquinone-2-sulfonate (NQS), plumbagin, juglone, and 1,4-naphthoquinone (NQ) under different chemical treatments is questionable.

An alternative to the second step of the above procedure is to use a different oxidant such as potassium ferricyanide [K3Fe(CN)6] (Matthiessen 1995). The procedure relies on the ability of reduced HS to oxidize [K3Fe(CN)6] to [K4Fe(CN)6] and the loss of absorbance measured in UV–Vis at 420 nm. The incubation time for this procedure is 24 h, which is much longer than ferric citrate (i.e., 15 min). However, no efforts have been made till date to compare these two incubation methods resulting in different reported CRC and MRC values for same HS and model compounds.

Schnitzer and Riffaldi (1972) modified a method from Vasilyevskaya et al. (1971) to determine quinone contents of HS and model compounds and found very good agreement with the theoretical quinone values for model compounds. This procedure uses Sn2+ as a reductant for HS. The quinone content is determined via potentiometric titration of excess Sn2+ using potassium dichromate (K2Cr2O7). None of the recent studies used this method to investigate the relative reliability of quinone content measurements of HS. We decided to apply this method to assess its applicability in regard to redox property measurements of HS and compare with other common methods. Our main objective of this study was therefore to compare and contrast redox property measurements of HS by Schnitzer and Riffaldi (1972), ferric citrate incubation, and [K3Fe(CN)6] incubation methods using AQDS as a model compound, and standard Elliot soil humic acid (1S102H, ESHA), reference Pahokee peat (1R103H, PPHA), and Suwannee River natural organic matter (1R101N, SRNOM) as representative HS.

Materials and methods

Humic acids and model compounds

Anthraquinone-2,6-disulfonate (AQDS) was purchased from Fluka and used as received. Standard Elliot soil humic acid (1S102H, ESHA), reference Pahokee peat (1R103H, PPHA), and Suwannee River natural organic matter (1R101N, SRNOM) were purchased from International Humic Substances Society and used as received. The properties of these HS and AQDS can be found in Table 1 of Rakshit et al. (2009).

Reduction pretreatment

An aliquot of unreduced HS taken from stock aqueous suspension (200 mg L−1) in phosphate buffer at pH 7 was reduced with N2/H2 gas (95:5%; Visser 1964; Kappler et al. 2004) in the presence of Pd catalyst (5% Pd on activated charcoal 100 mg L−1; Acros Organics) for 24 h. The reduction pretreatment was performed in amber-colored sealed serum vials under the hood. After the reduction pretreatment, headspace was flushed with Ar gas to remove excess H2 and transferred to anaerobic glovebox for further analysis. A similar reduction pretreatment was conducted on AQDS with concentration of 2 g L−1.

Determination of reducing equivalent upon reduction pretreatment

Two procedures were used to determine the extent of reduction. The first one was to incubate the reduced AQDS and HS with ferric citrate for 15 min and subsequent 0.5 M HCl extraction of the suspension followed by measurement in UV–Vis at 562 nm wavelength (Lovley and Blunt-Harris 1999). The second procedure was incubation by K3Fe(CN)6 and determination of loss of absorbance at 420 nm in UV–Vis spectrophotometer. Details of the procedure are in Matthiessen (1995). Only AQDS was measured by this method.

Measurement of quinone content using potassium dichromate and stannous chloride

Procedure described in Schnitzer and Riffaldi (1972) was used for quinone measurements. All concentrations were kept exactly same as that of Schnitzer and Riffaldi (1972). However, for few experiments, a K2Cr2O7 concentration of 0.025 or 0.005 N was used instead of 0.05 N to increase precision. Here N of K2Cr2O7 was related to its M as N = M/6 considering the recommendation of IUPAC (Irving 1978) and assuming a six-electron transfer reaction between K2Cr2O7 and Sn2+. Ten milliliters of 0.05 N SnCl2, 2H2O was always used. Three types of titration data are reported here: first, titration of either HS or AQDS by dichromate in the presence of SnCl2 (called Sn(II)-reduced, which is essentially the original quinone titration by Schnitzer and Riffaldi (1972)); second, titration of H2/Pd-reduced HS or AQDS by dichromate in the absence of SnCl2 (called control, a required experiment to check potential undesirable reaction between reduced AQDS/HS and K2Cr2O7); and third, titration of H2/Pd-reduced HS/AQDS in the presence of SnCl2 (called H2-Sn(II)-reduced, which was a necessary experiment to evaluate completeness of the H2/Pd reduction pretreatment). All these titrations were accompanied by a blank, in which 10 mL 0.05 N SnCl2 was titrated with 0.05 N K2Cr2O7 using the methods of Schnitzer and Riffaldi (1972). For a majority of the titrations, two replicates of samples and blanks were used. In some cases, where all three titrations, i.e., two replicate samples and one blank, were performed on the same day, only one replicate of blank was reported.

Electrode calibration

pH electrodes were calibrated using pH 4 and 7 buffers (Fisher). Platinum redox electrodes (Orion Model 97-98) were calibrated with redox buffers containing either 0.1 M potassium ferrocyanide and 0.05 M potassium ferricyanide or 0.01 M potassium ferrocyanide, and 0.36 M potassium fluoride. The measured cell electromotive force values were 238 ± 7 and 304 ± 5 mV, respectively, as expected for the two buffers.

Results and discussion

AQDS

Charge concentrations of AQDS are listed in Table 1 and Fig. 1. The one-way analysis of variance was significant, F(3, 8) = 68.35, p < 0.05, in which total charge concentration was significantly different among four methods evaluated in this study (Fisher’s LSD, p < 0.05). It is evident from the data that the K2Cr2O7 titration by Schnitzer and Riffaldi (1972) resulted in the most accurate value of theoretical AQDS charge concentrations arising from quinone moieties (i.e., a theoretical quinone charge of 5.46 mol kg−1 AQDS). Titration of H2/Pd-reduced AQDS in the absence of SnCl2 (termed control in Fig. 1) suggests that the reduced AQDS did not react with K2Cr2O7. Experiment with reduced AQDS in the presence of SnCl2 was carried out to assess whether H2/Pd reduction method was incomplete (Fig. 2). A quinone charge of 1.88 mol kg−1 of AQDS was determined, indicating that H2/Pd reduction method did not reduce all quinone moieties of AQDS (Fig. 2). Incubation by K3Fe(CN)6 resulted in higher AQDS charge concentrations compared to the theoretical value. Additionally, precision was poor as reflected by the large standard error values. Due to very small absorbance change (see standard curves in “Appendix” Figs. 9, 10) with concentration in case of K3Fe(CN)6 compared to that of ferrozine method (Stookey 1970), the K3Fe(CN)6 method cannot be used. For example, K3Fe(CN)6 changed absorbance by only 0.46 U for a concentration change of 500 μM. In comparison, for ferrozine, absorbance changed by 2.44 U for a concentration difference of 100 μM. Hence, K3Fe(CN)6 incubation method is not sensitive enough to be used for measurements of small differences in charge concentrations.

Table 1 AQDS charge determined by different methods
Full size table
Fig. 1
figure 1

Potentiometric titration of AQDS by potassium dichromate. AQDSR1 and AQDSR2 represent two replications, and Blank represents where no AQDS added. Control indicates H2/Pd-reduced AQDS titrated with potassium dichromate in the absence of stannous chloride

Full size image
Fig. 2
figure 2

Potentiometric titration of reduced anthraquinone-2,6-disulfonate titration. Reduced AQDS R1 and R2 indicates titration of H2/Pd (phosphate buffer at pH = 7, 24 h)-reduced AQDS. Blank represents titration of SnCl2 without AQDS

Full size image
Fig. 3
figure 3

Potentiometric titration of ESHA by 0.05 N potassium dichromate. ESHA R1 and ESHA R2 represent two replications, and Blank represents where no ESHA was added. Control represents an experiment where H2/Pd-reduced ESHA was titrated by potassium dichromate in the absence of stannous chloride

Full size image
Fig. 4
figure 4

Potentiometric titration of H2-Sn(II)-reduced ESHA by 0.025 N potassium dichromate

Full size image
Fig. 5
figure 5

Potentiometric titration of PPHA by 0.05 N potassium dichromate. PPHAR1 and PPHAR2 represent two replications, and Blank represents where no PPHA added. Control indicates H2/Pd-reduced PPHA titrated with potassium dichromate in the absence of stannous chloride

Full size image
Fig. 6
figure 6

Potentiometric titration of H2-Sn(II)-reduced PPHA by 0.05 N potassium dichromate in the presence of 10 mL 0.05 N SnCl2. PPHAR1 and PPHAR2 represent two replicated experiments where H2/Pd-reduced PPHA was titrated with dichromate in the presence of 10 mL 0.05 N SnCl2. Blank represents where no reduced PPHA added

Full size image
Fig. 7
figure 7

Potentiometric titration of SRNOM by 0.025 N potassium dichromate. Blank represents where no SRNOM added. Control represents H2-reduced SRNOM titration with K2Cr2O7

Full size image
Fig. 8
figure 8

Potentiometric titration of H2-Sn(II)-reduced SRNOM by 0.025 N potassium dichromate in the presence of 10 mL 0.05 N SnCl2. Reduced_SRNOMR1 and Reduced_SRNOMR2 represent two replicated experiments where H2/Pd-reduced SRNOM were titrated with dichromate in the presence of 10 mL 0.05 N SnCl2. Blank represents where no reduced SRNOM added

Full size image
Fig. 9
figure 9

Standard curve for K4Fe(CN)6 method

Full size image
Fig. 10
figure 10

Standard curve for ferric citrate method. Ferrozine represents the colorimetric reagent used to quantify the amount of Fe(III) reduced to Fe(II)

Full size image

Ferric citrate incubation resulted in lower charge concentrations of AQDS arising from quinone moieties compared to theoretical value (Table 1). However, the precision was very good, supported by our reasoning that this method is very sensitive. The lower value of charge concentration resulted from incomplete reduction of AQDS by H2/Pd (evidenced in Fig. 2) and the short incubation time of 15 min, as suggested by Aeschbacher et al. (2010). Therefore, the method of Schnitzer and Riffaldi (1972) provides accurate charge concentrations for AQDS.

Humic substances

From the titration results of H2/Pd-reduced HS (ESHA, PPHA, SRNOM) in the absence of Sn(II), it is clear that K2Cr2O7 does not react with H2/Pd-reduced HS and does not interfere with the titration method (control in Figs. 3, 5, 7). For ESHA and PPHA, there is little reduction in redox potential at the time of titration; however, the interference occurs at a lot lower redox ladder than the actual reaction between excess Sn2+ with K2Cr2O7. Therefore, it is unlikely that H2/Pd-reduced HS will react with K2Cr2O7 to cause erroneous titration values. Hence, we can safely rule out the possibility of overestimation of charge concentrations during titration due to reaction between H2/Pd-reduced HS and K2Cr2O7 (Figs. 4, 6).

The next logical step was to analyze the conditions where H2-Sn2+-reduced HS had been reacted with K2Cr2O7. These charge concentration values would represent the reducing equivalents in HS that are not reduced by H2/Pd. Table 2 and Figs. 4 and 6 show that there remains considerable amount of reducible charge concentrations for ESHA and PPHA that cannot be reduced by H2/Pd treatment. Therefore, the hypothesis of Ratasuk and Nanny (2007), which assumed that H2/Pd treatment may destroy quinone moieties, could not be validated. This is because if quinone moieties were destroyed by H2/Pd reduction treatment at pH 7, the titration method would not produce any residual charge concentrations. Hence, we conclude that H2/Pd reduction method at pH 7 is incomplete for ESHA and PPHA. For SRNOM, we did not detect any charge concentrations after H2/Pd reduction followed by titration with K2Cr2O7 in the presence of SnCl2 (Table 2). This is most likely due to the fact that H2/Pd reduction completely reduced all the reducible moieties of SRNOM. The charge concentration values resulting from Schnitzer and Riffaldi (1972) method (or Sn2+-reduced method) are also listed in Table 2. The little difference in values, i.e., 0.25 compared to 0.205 mol charge kg−1, may arise from experimental variability. In this context, it should be mentioned that the titration method cannot distinguish between small differences in charge concentrations in the range of 0.03–0.05 mol charge kg−1 HS. For example, if one single drop from the buret changes 0.1 mL volume reading for either 0.05 or 0.025 N K2Cr2O7, the calculated charge concentrations for 0.1 mL would then be in the range of 0.03–0.05 mol charge kg−1 HS. This argument supports our conclusion that for SRNOM, H2/Pd reduction was also complete. Hence, we conclude that charge concentration determination of HS by the method of Schnitzer and Riffaldi (1972) is most accurate in comparison with the other tested methods. In fact, for ESHA and PPHA, the charge conc. values by quinone titration were 1.38 and 1.25 mol charge/kg, respectively. These values were close to the values reported by Aeschbacher et al. (2010), in which the reported charge conc. of ESHA and PPHA were 1.96 and 1.62 mol charge/kg, respectively (Fig. 8).

Table 2 Charge concentrations of ESHA, PPHA and SRNOM by different methods
Full size table

The incompleteness of H2/Pd followed by FeCit incubation (i.e., CRC method) can be corrected using the charge concentration values from H2-Sn2+-reduced HS. For example, in Table 2, for ESHA and PPHA, if one adds the charge concentration values from the CRC method and H2-Sn2+ methods, it will be in good agreement with charge concentrations obtained from Schnitzer and Riffaldi (1972) (or Sn2+ reduced) methods.

In summary, K3Fe(CN)6 incubation method should not be used because of its low sensitivity. FeCit incubation for 15 min may underestimate charge concentrations of HS. H2/Pd reduction method results in incomplete reduction pretreatment that does not destroy all quinone moieties. Quinone determination method of Schnitzer and Riffaldi (1972) results in accurate estimate of quinones for model compounds (AQDS) and estimates quinone charge concentrations of HS, which can be used for characterizing the redox properties of HS. Sensitivity of Schnitzer and Riffaldi (1972) method can be enhanced by using lower concentrations of K2Cr2O7.