Acid-Base Chemistry of Natural Organic Matter,Hydrophobic Acids,and Transphilic Acids from the Suwannee River,Georgia,as Determined by Direct Potentiometric Titration

Abstract

Potentiometric direct titrations were used to measure the carboxyl and phenolic contents of natural organic matter that was isolated using reverse osmosis, hydrophobic acids (HPOA) that were isolated by adsorption on XAD-8 resin, and transphilic acids (TPIA) that were isolated by adsorption on XAD-4 resin. All samples were isolated from the headwaters of the Suwannee River in southeastern Georgia, in May, 2012. One approximation method and two numerical models were used to quantify the concentrations and acidic strengths of carboxyl groups and phenolic groups for the samples. The approximation method was the pH 8 and 10 method. Numerical models were the Gaussian distribution model and the modified Henderson–Hasselbalch model. All measurements and models yielded a consistent trend in the carboxyl contents, with TPIA having the greatest concentration and HPOA having the least. The 2R101N sample had the lowest average log K for proton binding (4.02), indicating that this sample contains a significant concentration of highly acidic compounds that are not in HPOA and TPIA. This can be explained by the loss of amino acids, peptides, and hydrophilic polycarboxylic acids that prefer the mobile phase in the solid-phase extractions with XAD-8 and XAD-4. The approximation method and both models were used to assign the phenolic content for the 2R101N sample. Only the approximation method was used for HPOA and TPIA. Concentrations of phenolic groups from the pH 8 and 10 method for 2R101N, HPOA, and TPIA were 2.86, 3.05, and 2.77 meq/gC, respectively.

Introduction

The International Humic Substances Society (IHSS, www.humicsubstances.org) maintains and distributes a set of standard and reference samples of organic matter from natural waters, including eight samples that have been isolated from the Suwannee River, Georgia. Six fulvic acids (FA) and humic acids (HA) were isolated by the XAD-8 method (Malcolm et al., 1989) in 1983 and 2003, and two samples of natural organic matter (NOM) were isolated by reverse osmosis (Serkiz and Perdue, 1990) in 1999 and 2012. During the collection of the NOM sample in 2012 (Green and Perdue, this issue), a small volume of river water was also processed by the XAD-8/XAD-4 method (Aiken et al., 1992) to isolate hydrophobic acids (HPOA) and transphilic acids (TPIA). The purpose of this study is a comparison of the acidic functional group composition of the NOM, HPOA, and TPIA samples, all collected at the same time from the same sampling site on the Suwannee River.

Experimental

Sample preparation

Samples of NOM, HPOA, and TPIA isolated from the Suwannee River in May, 2012, were dried for 16 h in an Isotemp Vacuum Oven Model 281A at 50°C under vacuum. Each sample was weighed on a Sartorius Quintix analytical balance and dissolved at a concentration of ∼400 mg/L in 0.1 M NaCl. Preparation and transfer of solutions were conducted under an atmosphere of humidified nitrogen.

Analyses of inorganic solutes

Concentrations of inorganic solutes in all three samples were determined in the vapor pressure osmometric studies of these materials (Pavlik and Perdue, this issue).

Potentiometric titrations

A 40-mL aliquot of the sample was titrated to an endpoint of pH 10.5 with a Mettler Toledo G20 Compact Auto titrator. A Mettler Toledo DGi115-SC pH electrode was used to monitor pH during the titration. The titrant was 0.0988 M NaOH, standardized with dried potassium hydrogen phthalate. Titrant was added at a rate of 30 μL/min with constant stirring until pH reached 10.5. Each titration lasted 35–45 min. Titration data (volume of added titrant, pH, etc.) were recorded using LabX titration software (version 3.1). A contaminated pH buffer caused a calibration error in titrations of HPOA and TPIA samples. The error was discovered subsequently during analysis of the titration data. Once data for the contaminated buffer were removed from the calibration data set, it was possible to use original voltages to recalculate correct pH values at each increment of added titrant. Unfortunately, the calibration error caused these titrations to end prematurely at pH 10, thus limiting the interpretation of these titration curves at high pH. Insufficient quantities of HPOA and TPIA were available to repeat these titrations. All titrations were conducted under flowing humidified nitrogen gas, using a Haake G temperature bath to maintain the temperature of the sample at 25°C±0.1°C. All titrations were performed in triplicate.

Data treatment

The organic charge (in eq/L) at any point in the titration was calculated using Equation (1). \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*} \Sigma [ Org_{i}^{ - } ] = & [ H^{ + } ] + [ Na^{ + } ] + 2 [ Ca^{2 + } ] + 2 [ Mg^{2 + } ] +[ K^{ + } ] \\ & + [ NH_{4}^{ + } ] - [ OH^{ - } ] - [ Cl^{ - } ] - 2 [ SO_{4}^{2 - } ] \quad\quad\quad [ { \rm Eq.} \ ( 1 ) ] \end{align*} \end{document}

where the concentrations of Na⁺, Ca²⁺, Mg²⁺, K⁺, NH₄⁺, Cl⁻, and SO₄²⁻ include initial concentrations in the sample, NaCl that was added to adjust ionic strength, and NaOH titrant. All of these concentrations are corrected for dilution that is caused by addition of the titrant.

The concentrations of H⁺ and OH⁻ were calculated from pH using the Davies equation to calculate their activity coefficients. Because this calculation depends on ionic strength, as defined in Equation (2), Equations (1) and (2) must be solved iteratively. Because the sample is dissolved in 0.1 M NaCl, the ionic strength varies only slightly from 0.1 throughout the titration. \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*} I = & \frac {1} {2} \left( \Sigma [ Org_{i}^{ - } ] + [ H^{ + } ] + [ OH^{ - } ] \right.\\ & + [ Na^{ + } ] + 4 [ Ca^{2 + } ] + 4 [ Mg^{2 + } ] + [ K^{ + } ] + [ NH_{4}^{ + } ] \\& \left.+ [ Cl^{ - } ] + 4 [ SO_{4}^{2 - } ] \right) \quad\quad\quad [ { \rm Eq.} \ ( 2 ) ] \end{align*} \end{document}

The organic charge calculated in Equation (1) was divided by the dilution-corrected concentration of dissolved organic carbon to convert it into an intensive property—the experimental organic charge density (Q_exp), which has units of meq/gC. This is shown in Equation (3). \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*} Q_ { exp } = 1000 \left( \frac { \Sigma [ Org_ { i } ^ { - } ] } { DOC } \right) \quad\quad\quad [ { \rm Eq. } \ ( 3 ) ] \end{align*} \end{document}

Modeling

One approximation method and two numerical models were used to assign concentrations of carboxylic acid functional groups and phenolic groups for the samples. The approximation method was the pH 8 and 10 method (Bowles et al., 1989), in which it is assumed that all the carboxyl groups and no phenolic groups are titrated by pH 8. The carboxyl content was thus estimated to be equal to Q_exp at pH 8 (Cabaniss, 1991). From pH 8 to pH 10, it was assumed that half of the phenolic groups have been titrated (Thurman, 1985). The phenolic content was thus estimated to be double the difference in Q_exp between pH 8 and pH 10 (Santos et al., 1999).

The concentrations of acidic functional groups in the samples were estimated using the modified Henderson–Hasselbalch model. This model describes proton binding by two different types of proton-binding sites using Equation (4) (Katchalsky and Spitnik, 1947). \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*} Q_ { model } = \left( \frac { Q_ { 1 } } { 1 + ( K_ { 1 } [ H^ { + } ] ) ^ { 1 / n_ { 1 } } } \right) + \left( \frac { Q_ { 2 } } { 1 + ( K_ { 2 } [ H^ { + } ] ) ^ { 1 / n_ { 2 } } } \right) \quad\quad\quad [ { \rm Eq. } \ ( 4 ) ] \end{align*} \end{document}

One of the two different types of binding sites was assumed to be the carboxyl groups, with a concentration of Q₁ meq/gC, an average equilibrium constant of K₁ for proton binding, and width parameter (n₁) that increases with increased range of K values within the class of carboxyl groups. The second type of proton-binding sites was assumed to be the phenolic groups, with a concentration of Q₂ meq/gC, an average equilibrium constant of K₂ for proton binding, and a width parameter (n₂). The Fletcher–Powell minimization algorithm was used to optimize the model fitting parameters (Fletcher and Powell, 1963).

Concentrations of carboxylic and phenolic groups for the sample were also estimated using the Gaussian distribution model (Perdue et al., 1984). In this model, the variation of Q_exp with pH was attributed to two Gaussian distributions of acidic functional groups. Each of these distributions can be described independently by a charge density (Q, meq/gC), an average log K for proton binding (μ), and a standard deviation of log K values within the distribution of proton-binding sites (σ). The overall charge density (Q_model) was described mathematically in Equation (5). The same minimization algorithm was used to fit the data. \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*}Q_{model} =& \left( \frac {Q_1} {\sigma_1 \sqrt{2 \pi}} \int_{ - \infty}^{ \infty} \left( \frac {1} {1 + K_1 [ H^ + ] } \right) \right. { \left. e^{- \left(\frac {\mu - \log K_1} {\sigma_{2} \sqrt{2}} \right) ^2} dlog K_{1} \right)} \\ & + \left( \frac {Q_{2}} {\sigma_{2} \sqrt{2 \pi}} \int_{ - \infty}^{ \infty} \left(\frac {1} {1 + K_{2} [ H^ + ]} \right) e^{- \left(\frac {\mu - \log K_2} {\sigma_{2} \sqrt{2}} \right) ^2} d log K_{2} \right) \quad\quad\quad [ { \rm Eq.} \ ( 5 ) ] \end{align*} \end{document}

When using these three means of estimating the carboxyl content, the pH 8 and 10 method gives the lowest estimate, the modified Henderson–Hasselbalch model gives the highest estimate, and the Gaussian distribution model generally gives values between the other two approaches (Zhang et al., 2013).

Results and Discussion

Titration curves (pH versus volume of added NaOH) are given in Fig. 1 for the three samples. Nearly identical concentrations of organic carbon (179 and 181 mg C/L) were used in the titrations of the 2R101N and HPOA samples, respectively, so it is evident that the concentration of acidic functional groups is greater in 2R101N than in HPOA. Due to the available sample, a lower concentration of organic carbon (119 mg C/L) was used in the titrations of TPIA, so its titration curve cannot be compared directly with those of 2R101N and HPOA. When differences in the concentration of organic carbon are taken into account, the TPIA sample is expected to contain the highest concentration of acidic functional groups.

FIG. 1.

Acid-base titrations of 2R101N, hydrophobic acid (HPOA), and transphilic acid (TPIA) shown in pH versus mL of 0.1 M NaOH titrant added.

Charge densities (Q_exp) were calculated at every point in each titration using inorganic data, titration data, and Equations (1 )–(3). Graphs of Q_exp versus pH were plotted for each sample. These graphs were overlaid to better visualize differences in charge density (Fig. 2). All samples were shown to have the same overall shape and slightly differing Q_exp values through the pH range, with TPIA having the steepest increase in Q_exp values at low pH. Above pH 3.7, Q_exp values were greatest for TPIA and lowest for HPOA, with intermediate values for 2R101N. As noted earlier, the titrations of HPOA and TPIA were terminated prematurely at pH 10. It is nonetheless evident from Fig. 2 that the rate of increase in Q_exp with increasing pH at pH 10 is approximately the same for those samples as for 2R101N, for which the titrations reached pH 10.5. This indicates that HPOA and TPIA also contain titratable phenolic groups; however, the lack of data above pH 10 makes it unlikely that the acidic properties of phenolic groups can be determined using the Gaussian and Henderson–Hasselbalch models. For this reason, only the pH 8 and 10 method has been used here to estimate the acidic properties of phenolic groups in HPOA and TPIA.

FIG. 2.

Charge density versus pH for all three titrated samples of 2R101N, HPOA, and TPIA.

The pH 8 and 10 estimation method was applied to the charge density and pH data to estimate concentrations (in meq/gC) of the carboxylic and phenolic functional groups (Table 1). The results are consistent with the sequence of Q_exp values above pH 3.7 (Fig. 2), where TPIA has the highest carboxyl content and HPOA has the lowest. All three samples have similar phenolic contents in the range of 2.77–3.05 meq/gC. It is to be noted that the pH 8 and 10 method is just an estimation method based on Q_exp at pH 8 and pH 10. This method of estimation, unlike the two models to follow, does not interpret the pH dependence of Q_exp to provide any insight into the distribution of acidic strengths in each class of proton-binding sites.

Table 1.

Carboxyl Content (Q_exp at pH 8) and Phenolic Content (Two Times the Change in Q_exp Between pH 8 and pH 10)

Sample	Carboxyl content	Phenolic content
2R101N	10.39±0.94	2.86±0.60
HPOA	9.15±0.20	3.05±0.50
TPIA	12.66±0.10	2.77±0.11

HPOA, hydrophobic acid; TPIA, transphilic acid.

Carboxyl contents from the Gaussian distribution model are given in Table 2 and are consistent with the general trend in Figs. 1 and 2 and Table 1. TPIA has the highest carboxyl content and HPOA has the smallest. In contrast to the simple pH 8 and pH 10 method, this model also provides information regarding acidic strengths. Log K values for proton binding by carboxyl groups (μ₁) are in the order 2R101N<TPIA<HPOA, indicating that 2R101N contains a greater proportion of strongly acidic carboxylic acids than either of the samples that were obtained by solid-phase extractions. The standard deviation of log K values in a Gaussian distribution of proton-binding sites (σ) is greatest for HPOA and least for TPIA, which indicates that HPOA contains the most chemically diverse mixture of carboxylic acids.

Table 2.

Results for the Gaussian Distribution Model (See Text for Explanation of Model)

	Carboxyl groups			Phenolic groups
Sample	Q₁	μ₁	σ₁	Q₂	μ₂	σ₂
2R101N	10.70±0.90	4.02±0.18	2.10±0.21	2.23±1.09	9.94±0.13	0.60±0.18
HPOA	10.27±0.30	4.57±0.05	2.46±0.07	0±0.00	0±0.00	0±0.00
TPIA	13.18±0.08	4.25±0.03	1.46±0.02	0±0.00	0±0.00	0±0.00

Table 2 also contains the estimates of acidic properties of phenolic groups from the Gaussian distribution model for 2R101N. The phenolic content of 2.23 meq/gC is slightly lower than the estimate from the pH 8 and pH 10 method. The average log K for proton binding by phenolic groups (μ₂) was 9.94. These results compare well with other titrations of humic substances from the Suwannee River (Driver and Perdue, 2014). As noted earlier, the acidic properties of phenolic groups of HPOA and TPIA could not be estimated using the Gaussian distribution model.

Kuhn et al. (this issue) recovered 60% of organic carbon in the HPOA fraction and 18% of organic carbon in the TPIA fraction, leaving 22% of the organic carbon in a fraction that was too hydrophilic to be recovered by solid-phase extraction. The 2R101N sample, which was isolated using reverse osmosis, represents 84.2% of organic carbon, leaving 15.8% of unrecovered organic carbon (Green and Perdue, this issue). The fact that the average log K for proton binding by carboxyl groups in 2R101N (4.02) is substantially lower than the average log K values for the HPOA (4.57) and TPIA (4.25) samples confirms that 2R101N contains highly acidic compounds (compounds with low log K values for proton binding) that are not well recovered by the solid-phase extractions.

The distribution of proton-binding sites in whole NOM from the Suwannee River is assumed here to be equal to the distribution of proton-binding sites in 2R101N. That distribution is well modeled as a mixture containing 60% HPOA, 18% TPIA, and 22% missing NOM. When the Gaussian fitting parameters of 2R101N, HPOA, and TPIA from Table 2 are combined, the best Gaussian parameters for the missing NOM are Q₃=11.06 meq/gC, μ₃=2.83, and σ₃=1.78, revealing a highly acidic unrecovered fraction. Compounds that typically contain such highly acidic carboxyl groups include amino acids, peptides, and polycarboxylic acids, in which carboxyl groups are in close proximity to one another (simple examples are oxalic acid and o-phthalic acid). These compound classes are likely to be charged and/or highly water-soluble at pH 2, so they do not readily partition out of aqueous solution and are thus less likely to be recovered in solid-phase extractions.

The modified Henderson–Hasselbalch model was also fit to titration data (Table 3). This model also showed the same trend in carboxyl contents that was found by other approaches, with TPIA having the greatest carboxyl content and HPOA having the lowest. The log K values for proton binding for 2R101N, HPOA, and TPIA (values of 4.16, 4.74, and 4.26 meq/gC, respectively), followed the trend that was found using the Gaussian distribution model, again indicating the presence of strongly acidic carboxylic acids in 2R101N that are not present in HPOA and TPIA. The modified Henderson–Hasselbalch model predicts a lower phenolic content for 2R101N (1.60 meq/gC), but a very similar log K (9.99), compared to the corresponding predicted properties from the Gaussian distribution model.

Table 3.

Results for the Henderson–Hasselbalch Model (See Text for Explanation of Model)

	Carboxyl groups			Phenolic groups
Sample	Q₁	Log K₁	n₁	Q₂	Log K₂	n₂
2R101N	11.20±0.82	4.16±0.20	3.44±0.39	1.60±0.74	9.99±0.18	1.03±0.06
HPOA	10.74±0.08	4.74±0.10	3.99±0.30	0±0.00	0±0.00	1.00±0.00
TPIA	13.25±0.06	4.26±0.04	2.32±0.03	0±0.00	0±0.000	1.00±0.00

Conclusions

Potentiometric titrations were used to determine the concentrations and acidic strengths of carboxyl and phenolic groups of 2R101N, HPOA, and TPIA samples that were isolated from the Suwannee River in 2012. All measurements and models yielded a consistent trend in the concentrations of carboxyl groups (HPOA<2R101N<TPIA) and log K values for proton binding (2R101N<TPIA<HPOA). The greater overall acidic strength of carboxyl groups in 2R101N can be explained by the successful recovery of amino acids, peptides, and very polar polycarboxylic acids that are not well recovered from water by adsorption to XAD resins at pH 2. The proton-binding properties of the sample isolated using reverse osmosis (2R101N) were similar to those of NOM, HA, and FA that were isolated previously from the Suwannee River (Driver and Perdue, 2014). Average log K values for proton binding by carboxyl and phenolic groups were reasonable for those classes of compounds. The concentrations of phenolic groups were in the order of TPIA<2R101N<HPOA; however, the range of estimated concentrations (2.76–3.05 meq/gC) was rather small and probably within experimental error.

Footnotes

Acknowledgments

The authors acknowledge the financial support from the IHSS for isolation of the sample of Suwannee River NOM that was used in this study. They also thank Dr. George Aiken and Dr. Patricia Maurice for providing samples of HPOA and TPIA for this study.

Author Disclosure Statement

No competing financial interests exist.

References

Aiken

G.R.

, McKnight

D.M.

, Thorn

K.A.

, and Thurman

E.M

. (1992). Isolation of hydrophilic organic acids from water using nonionic macroporous resins. Org. Geochem., 18, 567.

Bowles

E.C.

, Antweiler

R.C.

, and MacCarthy

(1989). Acid–base titrations and hydrolysis of fulvic acid from the Suwannee River. In: Averett

R.C

, Leenheer

J.A.

, McKnight

D.M.

, and Thorn

K.A.

, Eds. Humic Substances in the Suwannee River, GA: Interaction, Properties, and Proposed Structures. USGS Open File Report 87-557, pp. 209–229.

Cabaniss

S.E.

(1991). Carboxylic-acid content of a fulvic-acid determined by potentiometry and aqueous fourier-transform infrared spectrometry. Anal. Chim. Acta, 255, 23.

Driver

S.J.

, and Perdue

E.M.

(2014). Acidic functional groups of Suwannee River natural organic matter, humic acids, and fulvic acids. In: Rosario-Ortiz

and Michael Thurman

, Eds., Advances in the Physicochemical Characterization of Organic Matter, ACS Symposium Series: 75.

Fletcher

, and Powell

M.J.D

. (1963). A rapidly convergent descent method for minimization. Comput. J., 6, 163.

Green

N.W.

, McInnis

, Hertkorn

, Maurice

P.A.

, and Perdue

E.M

. Suwannee River natural organic matter: Isolation of the 2R101N reference sample by reverse osmosis. Environ. Eng. Sci., 32, 38.

Katchalsky

, and Spitnik

. (1947). Potentiometric titrations of polymethacrylic acid. J. Polym. Sci., 2, 432.

Kuhn

K.M.

, Neubauer

, Hofmann

, von der Kammer

, Aiken

G.R.

, and Maurice

P.A

. Concentrations and distributions of metals associated with dissolved organic matter from the Suwannee River (GA, USA). Environ. Eng. Sci. 32, 54.

Malcolm

R.L.

, Aiken

G.R.

, Bowles

E.C.

, and Malcolm

J.D

. (1989). Isolation of fulvic and humic acids from the Suwannee River. In: Averett

R.C.

, Leenheer

J.A.

, McKnight

D.M.

, and Thorn

K.A.

, Eds., Humic Substances in the Suwannee River, Georgia: Interactions, Properties, and Proposed Structures, USGS Open File Report 87-557, Washington, DC: United States Government Printing Office, pp. 23–35.

10.

Pavlik

J.W.

, and Perdue

E.M

. Number-average molecular weights of natural organic matter, hydrophobic acids, and transphilic acids from the Suwannee River, Georgia, as determined using vapor pressure osmometry. Environ. Eng. Sci., 32, 23.

11.

Perdue

E.M.

, Reuter

J.H.

, and Parrish

R.S

. (1984). A statistical-model of proton binding by humus. Geochim. Cosmochim. Acta, 48, 1257.

12.

Santos

E.B.H.

, Esteves

V.I.

, Rodrigues

J.P.C.

, and Duarte

A.C

. (1999). Humic substances' proton-binding equilibria: Assessment of errors and limitations of potentiometric data. Anal. Chim. Acta, 392, 333.

13.

Serkiz

S.M.

, and Perdue

E.M

. (1990). Isolation of dissolved organic matter from the Suwannee River using reverse osmosis. Water Res., 24, 911.

14.

Thurman

E.M.

(1985). Organic Geochemistry of Natural Waters. Martinus Nijhoff/Dr. W. Junk Publishers: Dordrecht.

15.

Zhang

, Green

N.W.

, and Perdue

E.M

. (2013). Acid-base properties of dissolved organic matter from pristine and oil-impacted marshes of Barataria Bay, Louisiana. Mar. Chem., 155, 42.