Environ. Sci. Technol. 2003, 37, 5727-5731
Determination of Strongly Reducing Substances in Sediment Q I N G M A N L I , * ,†,‡ S H U P I N G B I , * ,† A N D GUOLIANG JI‡ Department of Chemistry, Nanjing University, Nanjing 210093, P. R. China, and Institute of Soil Science, Chinese Academy of Sciences, Nanjing 210008, P. R. China
Reducing substances in sediments play an important role in regulating the chemical and biological status of the sediment. Strongly reducing substances are the most active part of them, and are difficult to measure directly. In this work, we have developed a method for their determination based on the oxidation by Fe(III). Then, the produced Fe(II) was determined by colorimetry, using 2,2′-bipyridine as the chelating agent. Prior to determination, these substances were extracted from the sediment by M acetic acid. The whole process was carried out in a closed system, so that oxidation of Fe(II) and reducing substances by atmospheric oxygen could be avoided. The calibration curve between absorbance and Fe(II)-chelate concentration in extract was linear. When ascorbic acid was added to the extract, the recovery was larger than 94.0%. The effect of surplus Fe(III) on the result was discussed. With the proposed method, the concentration of strongly reducing substances in sediment samples from Dianchi Lake of China was measured with good reproducibility.
Introduction Reducing substances in sediments play an important role in regulating the chemical and biological status of the sediment, and thus are of significance in the management and restoration of aquatic ecosystems. The consumption of oxygen causes the redox potential of sediments to decrease (1). In this process, decaying organic matter is decomposed into organic reducing substances by anaerobic microbes, and the sediment changes to a reducing status, which can encourage the growth of toxin-producing microbes such as Clostridium botulinum within the eutrophic water column (2). Electrons from these organic reducing substances can transfer to oxidizing components such as ferric oxides and manganese oxides (3). As a result, the adsorption capacity of these oxides for nutrients decreases (4,5). Simultaneously, the products of reduction reactions such as ammonia, manganous ions, ferrous ions, and sulfide ions occur on the surface of sediments to hinder other elements from approaching the adsorption sites of the sediments, and cause some nutrients such as phosphate, originally adsorbed by the sediments, to release to the overlying water column (6, 7). As a consequence, the concentrations of nutrients and toxic chemicals in the overlying water column increase, and the quality of this water column is deteriorated. Furthermore, some strongly reducing substances may be directly toxic to * Corresponding authors’ phone: 86-25-6881184; fax: 86-256881000; e-mail:
[email protected] or
[email protected]; bisp@ nju.edu.cn. † Nanjing University. ‡ Chinese Academy of Sciences. 10.1021/es0343297 CCC: $25.00 Published on Web 10/29/2003
2003 American Chemical Society
the growth of microbes. Therefore, in treatment processes for improving eutrophic water column, strongly reducing substances in the sediment should be considered (5). Reducing substances in sediments are a mixture of compounds with different standard potentials. Ding and coworkers (8-10) have characterized them in soils into two classes, namely, strongly reducing substances and weakly reducing substances, on the basis of their peak potentials measured by voltammetry. If the peak potential on the i-E curve is lower than 0.35 V vs SCE measured with a graphite electrode, the reducing substances belong to strongly reducing substances; otherwise, they are weakly reducing substances. Some investigators have also distinguished reducing substances in soils or sediments by chemical methods (11,12). In these methods, substances that can reduce potassium permanganate at room temperature are called active reducing substances, and those that can reduce potassium dichromate under acid and high temperature conditions are referred to as weakly reducing substances. Obviously, strongly or active reducing substances in sediments are more important than weakly reducing substances. However, because permanganate and dichromate have standard potentials much higher than that of ferric oxides, the main acceptor of electrons in sediments, these chemical methods cannot be used to distinguish the activity of reducing substances in detail. Some methods for the determination of reducing substances have been reported. Yu et al. developed a DC voltammetric method in which the quantity of reducing substances was estimated through the current produced on a solid electrode (8). Wu and Ding have applied differential pulse voltammetry to quantify strongly reducing substances in soils and decomposing plant materials, based on the height of the peak current on i-E curves (13). However, when voltammetry is applied to sediments, many current peaks of the reducing substances may overlap on the i-E curve, unless proper separating techniques are applied beforehand. In chemical methods, because strong oxidizing agents are used, the amount of strongly reducing substances in sediments may be overestimated. To evaluate the amount of strongly reducing substances in sediments in a more precise way, a method based on the oxidation by Fe(III) was developed. In the method, acetic acid was used as the extractant, and the sediment was extracted in a closed system free from oxygen. The extract reacted with an excess of Fe(III), following the reaction
Or + Fe3+ h Ox + Fe2+ where Or denotes strongly reducing substances, and Ox represents products of strongly reducing substances oxidized by Fe(III). Then, Fe(II) was determined by colorimetry after the formation of a chelate with 2,2′-bipyridine (14). With the proposed method, the amount of strongly reducing substances in sediment samples from Dianchi Lake of China was determined.
Experimental Section 1. Reagents and Instruments. 1.1. Reagents. All chemicals were of reagent grade unless otherwise specified. Hydrochloric acid and acetic acid were from Nanjing Chemical Factory. Ferric chloride was from Shuanghuan Chemical Factory, Beijing. Ammonium iron (II) sulfate hexahydrate and 2, 2′-bipyridine were from Shanghai Chemical Factory. 1.2 Instruments. A PAR-174 polarograph (Princeton Applied Research) fitted with an X-Y recorder (2000-Recorder, Houston) was used for differential pulse voltammetric VOL. 37, NO. 24, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5727
FIGURE 1. Schematic diagram of the determination of strongly reducing substances. measurements. A glassy carbon electrode was utilized as the working electrode of the three-electrode cell, using a platinum-wire as the counter electrode and an Ag-AgCl electrode as the reference electrode. The spectrophotometer was from Shanghai Analytical Instruments. 2. Preparation of Solution. Ammonium iron (II) sulfate hexahydrate (3.9213 g)was dissolved in deionized water and diluted to 1000 mL to get a 0.01 M Fe(II) solution. To prevent the Fe(II) from oxidizing, a small amount of ascorbic acid was added. This solution was used as the stock solution throughout the study. Serial dilutions with 0.5% 2,2′bipyridine solution were made to prepare standard ferrous solutions. Ferric chloride (2.70 g) was dissolved in 1000 mL of water to get a 0.01 M solution. The concentration of Fe(III) was calibrated by AAS (Shimadzu, AA-680). Acetic acid (58.0 mL) was added into 1000 mL of water free from oxygen. The solution was approximately 1.0 M with respect to acetic acid and had a pH of about 2.0. 3. Procedure of Determination. 3.1 Sampling of Sediment. The sediment was sampled from Dianchi Lake of Yunnan Province, China. The lake had been in eutrophication. Owing to uncontrolled discharge of domestic and industrial effluents during the past decades, the lake had been seriously contaminated. A large part of the lake was choked by water hyachinth, and the sediment about 10 m beneath the water surface was in a heavily anaerobic condition, with an oxidation-reduction potential as low as -400 mV. The sampling was carried out with a gravity corer, in which the length of the cylinder was modified to gather more overlying water and sediment. When the gravity corer was pulled out of the water, the cylinder containing the sediment was unloaded and was shifted instantly with a device for transferring the sediment to brown storage bottles (shown in Figure 1). Prior to sampling, the storage bottle was treated with the following procedures to exclude oxygen: (a) filled with overlying water of the sediment, and bubbled with highpurity nitrogen gas; and (b) drained by a vacuum pump. The depth of sampled sediment was 20 cm. When not analyzed immediately, the sample was kept at 4 °C and free from light. 5728
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 24, 2003
3.2 Determination. A given volume of a mixed solution of ferric chloride with 2,2′-bipyridine, or a 2,2′-bipyridine solution alone, was added into reaction cells 1 and 2, respectively. The vacuum pump was switched on to exclude air, and high-purity nitrogen gas was bubbled through the system. The sediment in the storage bottle was stirred by a magnetic stirrer. After 10 min, 25 mL of sediment and 25 mL of acetic acid were transferred into the extraction cell with a volumetric cylinder on the extraction cell, respectively. Before addition, the acetic acid had been bubbled with nitrogen gas to exclude oxygen. At the same time, a subsample was transferred into a beaker for weighing. After extraction for 30 min, the suspension was kept standing for 30 min, and then the supernatant solution was shifted to the reaction cells through a filter (Nuclepore filter, 0.2-µm pore diameter). The strongly reducing substances in the extract were allowed to react with Fe(III) in reaction cell 1, accompanied by the formation of Fe(II)-chelate with 2,2′-bipyridine. Simultaneously, in reaction cell 2, extracted Fe(II) also reacted with 2,2′-bipyridine, which was used as the check. When the reaction between strongly reducing substances and Fe(III) was complete, Fe(II) was determined by colorimetry at a wavelength of 520 nm. The amount of strongly reducing substances was calculated as the difference between the Fe(II) determined and the Fe(II) extracted.
Results Calibration Curve of Fe(II) and Recovery of Ascorbic Acid. In the method, it is essential to know the relationship between the absorbance and the concentration of Fe(II)-chelate in the extract. The results show that the absorbance of Fe(II)chelate was proportional to the concentration of Fe(II) with a correlation coefficient (r2) of 0.9955, meaning that the Fe(II) produced from the reduction of Fe(II) can be directly determined in the extract. When Fe(III) was measured in pure solution, the correlation coefficient was 0.9997. A comparison between the two curves suggests that there might be some substances in the extract to increase the absorbance. When the concentration of Fe(II) was low, that is, when few
TABLE 1. Recovery of Ascorbic Acid in the Extract and Sediment (n ) 3)
a
determination (mol/L × 104)
recovery (%)
sample no
concentration of SRSa (mol/L × 104)
ascorbic acid added (mol/L × 104)
sediment
extract
sediment
extract
1 2 3 4 5 6
2.48 ( 0.14 2.10 ( 0.10 1.51 ( 0.11 0.73 ( 0.13 1.87 ( 0.21 1.05 ( 0.19
2.0 2.0 1.0 0.5 1.0 1.0
4.39 ( 0.23 3.95 ( 0.17 2.34 ( 0.17 1.07 ( 0.15 2.79 ( 0.26 1.80 ( 0.14
4.43 ( 0.10 4.06 ( 0.11 2.46 ( 0.06 1.20 ( 0.07 2.83 ( 0.12 2.02 ( 0.09
95.5 92.5 83.0 68.0 92.0 75.0
97.5 98.0 95.0 94.0 96.0 97.0
Note: SRS denotes strongly reducing substances.
TABLE 2. Amount of Strongly Reducing Substances in Sediments (n ) 3)
a
redox potential (mV)
strongly reducing substances (meq/kg)a
84 -128 -253 -389 -449
0.12 ( 0.11 0.97 ( 0.09 1.83 ( 0.13 2.46 ( 0.11
Reported as mean ( standard deviation.
strongly reducing substances were extracted, this influence was more distinct. Fortunately, this interference was generally negligible, and can be excluded by a blank experiment. To examine the utility of the proposed method, ascorbic acid was used as the model substance, and was added to sediments and extracts to measure its recovery. The results are given in Table 1. The recovery was related to the amount of strongly reducing substances when ascorbic acid was added to sediments, being lower when the amount of the former was low. On the other hand, the recovery in extracts was higher than that in sediments, attaining a value of larger than 94%. This means that Fe(III) can oxidize ascorbic acid quantitatively, and Fe(III) can be applied as a reagent to evaluate strongly reducing substances. Effect of Surplus Fe(III) on Determination. In the method, an excess of Fe(III) was added as the oxidant to determine the strongly reducing substances in sediments. The interference of surplus Fe(III) to the determination should be estimated. The results show that at a wavelength of 520 nm the Fe(III) affected the determination of Fe(II) to some extent. When the concentration of Fe(II) was maintained constant, the measured concentration of Fe(II) increased linearly with the amount of Fe(III) added. However, if the concentration of Fe(III) added was lower than 2.5 × 10-4 mol/L, the measured concentration of Fe(II) was in the confidence interval of 95%, meaning that the effect of surplus Fe(III) on determination was insignificant. A preliminary test for controlling the amount of surplus Fe(III) may be required in some cases. Amount of Strongly Reducing Substances in Sediments with Different Redox Potentials. The Dianchi Lake is eutrophic, and the decay of algae and macraphyte induced the sediment to an anaerobic condition. At some sites the redox potential of the sediments was below -400 mV, leading to the accumulation of strongly reducing substances. With the proposed method, the amounts of strongly reducing substances in sediments with different redox potentials were determined, and the results are given in Table 2. The table indicates that the amount of strongly reducing substances was related to the redox potential of the sediment. The higher the amount, the lower the redox potential. Because the amount of reducing substances is a capacity factor, while redox potential is only an intensity factor of oxidation-
FIGURE 2. i-E curves of reducing substances: A, extract from lake sediment; B, ascorbic acid, by differential pulse voltammetry (Redox potential of sediment was -448 mV, indifferent electrolyte was 1:1 HCl, scan rate was 1 mV/s, pulse potential was 50m V). reduction properties, a knowledge of the former parameter may be more important for evaluating the role of sediments in controlling water quality ecologically. Voltammetric Characterization of Reducing Substances. It would be of interest to characterize the reducing substances by a voltammetric means. The i-E curve of reducing substances in the extract of a sediment is shown in Figure 2 (A), and the i-E curve of ascorbic acid is given in Figure 2 (B). Under the experimental conditions, several groups of reducing substances were detected, based on the number of current peaks. Because only limited amounts of inorganic strongly reducing substances were expected to be present in the sediments, the substances appearing on the i-E curve would be largely organic in nature (9). Under identical experimental conditions, the peak potentials of some of these substances were lower than that of ascorbic acid, meaning that these substances were more apt to donate electrons than ascorbic acid during oxidation. If a criterion of 0.35 V vs SCE is adopted to define strongly reducing substances (8), these substances belong to this category. The relative amounts of these substances can also be compared through the height of the current peak. The reducing substances oxidized at -0.24 V vs SCE had the highest concentration, followed by the substances oxidized at 0.01 V vs SCE. However, it was difficult to estimate their absolute amount, because the peak potentials approached each other, resulting in the overlapping of current peaks. Stability of Reducing Substances. The stability of strongly reducing substances that can reduce Fe(III) in sediment of VOL. 37, NO. 24, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5729
FIGURE 3. Stability of reducing substances in sediments. Dianchi Lake was evaluated with the proposed method. After extraction from the sediment, the extract was exposed to air. At different intervals, the extract was sampled for determination. The results are shown in Figure 3. Within 30 min, 35% of the reducing substances could be oxidized by atmospheric oxygen. The amount decreased progressively, and no reducing substances could be detected after 24 h. This means that strongly reducing substances in sediments are very unstable, and can be oxidized readily by oxygen of the air. Therefore, it is essential to measure them immediately after extraction, and keep them free from oxygen during the whole process.
Discussion Reducing substances in sediments and soils exist in forms of different extractabilities. To extract them, many extractants have been employed in the literature. They vary in strength from water (15) to neutral salt calcium chloride (16, 17) or ammonium acetate (18), acid salt aluminum chloride (19) or aluminum sulfate (11), dilute strong acid (20) to concentrated strong acid (17). Because iron compounds are the most abundant and most important buffer in soils and sediments, methods for their detailed speciation (17, 21-24) have been developed. An integrated geochemical and microbiological approach has also been adopted (25). Bao (21, 22) emphasized the more active forms of reduced iron and manganese, distinguishing them as water-soluble, exchangeable, complexed with the solid part of soil organic matter, and precipitated, and further characterized water-soluble iron and manganese as those in ionic form and in chelated form. In the present research, because the main scheme is the transformation of Fe(III) into Fe(II), it would be feasible to keep the extracted amount of Fe(II) as low as possible, so that the blank in the subsequent determination can be small. Therefore, we chose M acetic acid as the extractant. This acid, when mixed with an equal volume of sediment suspension, had a pH value of about 4.0. Experimental results showed that the highest concentration of Fe(II) in the extract was only 1.9 × 10-4 mol/L. The concentration was even lower when the redox potential of the sediment was high. At this pH, the amount of dissolved sulfide would also be low, as shall be discussed in the next paragraph. In the extract there may be the presence of natural organic compounds capable of forming complexes with Fe(II). However, because the stability constant logK of Fe(II) with 2,2′-bipyridine is as large as 17 (14), much larger than those with naturally occurring chelating agents such as oxalate and citrate (26), it may be safely considered that these chelating agents would not compete with 2,2′-bipyridine during the color-developing process in the determination. Actually Bao (21, 22) has employed 2,2′-bipyridine to extract 5730
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 24, 2003
Fe(II) complexed with the organic matter of the solid part of submerged soils, and characterized water-soluble Fe(II) as those in ionic form and those in chelated form, utilizing the principle that 2,2′-bipyridine can replace complexed natural organic chelating agents. The composition of the strongly reducing substances is not exactly known. Apparently, Mn(II) with a standard potential higher than that of the Fe(II)-Fe(III) couple can be excluded. Sulfide ions, if present, may reduce Fe(III) (27, 28). However, theoretical considerations lead to the supposition that their contribution to the total reduction of Fe(III) in the determination would be small. Sulfide ions in sediments and soils are chiefly present as precipitates combining with iron and manganese ions. Their solubility is strongly pH-dependent (27). Measurements by Pan (27) with a sulfideselective electrode and hydrogen sulfide sensor showed that, in submerged soils at pH 4, the pH2S value ranged from 4.5 to 5.5 and the pS2- value ranged from 17 to 19. At pH 8, the pH of the investigated sediment under natural conditions, the corresponding figures were about 8.5 and 13 to 14, respectively. Besides, in the examination by voltammetric method, because the medium was acid in reaction (1:1 HCl), the concentration of free S2- ions should be so low that no corresponding signal could be detected. From the argument made above it can be concluded that these strongly reducing substances should be mainly organic in nature. Unfortunately, at present we know very little about their composition and properties (9, 29, 30), although some electrochemical characterizations have been made (9). This may be the main reason that many attempts have been made to use model substances (31) for simulating a given process. Insight into the composition and properties of organic-reducing substances, particularly the strongly reducing ones, remains an unexplored field. By definition in this work, those substances that can reduce Fe(III) ions are called strongly reducing substances. This definition seems more meaningful than that by Ding and co-workers (8-10), in which those substances which have a peak potential of lower than 0.35 V vs SCE on the i-E curve are called strongly reducing substances, and thus Fe(II) is included in this category. Actually, owing to its abundance in soils and sediments, iron mainly functions as a buffer for keeping the system in a mild reducing condition. It is these strongly reducing substances, not iron, that donate electrons to electron-acceptors, inducing oxidation-reduction reactions. Therefore, these substances play the primary role in sediments and other natural systems with respect to oxidation-reduction. Another term related to reducing substances is chemical oxygen demand (COD), a routine parameter for evaluating water quality. A similar term is TRC, total reduction capacity, in contrast to OXC, oxidation capacity (23, 32). The COD is generally determined by oxidation with potassium dichromate at a high temperature. Actually only a part of the measured value represents those substances that can react with atmospheric oxygen under natural conditions. Ding and Wang (8) determined COD with a voltammetric method in which the diffusion current at a carbon electrode was measured when a positive potential corresponding to the oxidation-reduction potential of water at natural conditions was applied. Either determined by the voltammetric method or determined by the present method, the measured value reflects a more realistic chemical oxygen demand under natural conditions than that determined by the conventional method.
Acknowledgments The support by the National Natural Science Foundation of China (projects 49971045 and 49831005) is gratefully acknowledged.
Literature Cited (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) (17) (18) (19) (20)
Ponnamperuma, F. N. Adv. Agron. 1972, 24, 29-96. Wobeser, G. J. Wildl. Dis. 1997, 33, 181-186. Lovley, D. R. Adv. Agron. 1995, 54, 175-231. Hem, J. D. Chem. Geol. 1978, 21, 199-218. Murphy, T. P.; Lawson, A.; Kumagai, M.; Babin, J. J. Aquat. Ecosyst. Health Manage. 1999, 2, 419-434. Heijs, S. K.; Azzoni, R.; Giordani, G.; Jonkers, H. M.; Nizzoli, D.; Viaroli, P.; van Gemerden, H. Aquat. Microb. Ecol. 2000, 23, 85-95. Crockford, R. H.; Willett, I. R. Hydrol. Process. 2000, 14, 23832392. Ding, C. P.; Wang, J. H. In Electrochemical Methods in Soil and Water Research; Yu, T. R., Ji, G. L., Eds.; Pergaman Press: Oxford, 1993, 366-412. Liu, Z. G.; Ding, C. P.; Wu, Y. X.; Pan, S. Z.; Xu, R. K. In Chemistry of Variable Charge Soils; Yu, T. R., Ed.; Oxford University Press: New York, 1997, 442-472. Ding, C. P.; Liu, Z. G.; Yu, T. R. Soil Sci. 1982, 134, 252-257. Liu, Z. G.; Yu, T. R. Acta Pedol. Sin. 1962, 10, 13-28. Bartlett, R. J.; James, B. R. Adv. Agron. 1993, 50, 151-208. Wu, Y. X.; Ding, C. P. Pedosphere 1991, 1, 157-167. Martell, A. E.; Calvin, M. Chemistry of Metal Chelate Compounds, Prentice-Hall: NewYork, 1952; pp 523-373. Flaig, W.; Scharrer, K.; Judel, G. K. Z. Pflanzenernaehr. Du ¨ eng. Bodenkd. 1955, 68, 203-218. Hemstock, G. A.; Low, P. F. Soil Sci. 1953, 76, 331-343. Heron, G.; Crouzet, C.; Bourg,. A. C. M.; Christensen, T. H. Environ. Sci. Technol. 1994, 28, 1698-1705. Fujimoto, C.; Sherman, G. D. Soil Sci. 1948, 66, 131-145. Ignatieff, V. Soil Sci. 1941, 51, 249-263. Starkey, R. L.; Wight, K. M. Anaerobic Corrosion of Iron in Soil, with Particular Consideration of the Soil Redox Potential as an
(21) (22) (23) (24) (25)
(26)
(27) (28) (29) (30) (31) (32)
Indicator of Corrosiveness; American Gas Association, New York, 1946. Bao, X. M. In Physical Chemistry of Paddy Soils; Yu, T. R., Ed.; Science Press-Springer-Verlag: Beijing, Berlin, 1985; pp 69-91. Bao, X. M. In Chemistry of Variable Charge Soils; Yu, T. R. Ed.; Oxford University Press: New York, 1997; pp 473-499. Heron, G.; Christensen, T. H. Environ. Sci. Technol. 1995, 29, 187-192. Chao, T. T.; Zhou, L. Soil Sci. Soc. Am. J. 1983, 47, 225-232. Cozzarelli, I. M.; Suflita, J. M.; Ulrich, G. A.; Harris, S. H.; Scholl, M. A.; Schlottmann, J. L.; Christenson, S. Environ. Sci. Technol. 2000, 34, 4025-4033. Zhang, S. W.; Tang, F. L.; Zhang, T. Handbook of Chemical Reagents (in Chinese); Chemical Industry Press: Beijing, 1987; pp 345-347, pp 370-371. Pan, S. Z. In Physical Chemistry of Paddy Soils; Yu, T. R. Ed.; Science Press-Springer-Verlag: Beijing-Berlin, 1985, 92-110. Butler, E. C.; Hayes, K. F. Environ. Sci. Technol. 2000, 34, 422-429. Nevin, K. P.; Lovley, D. R. Environ. Sci. Technol. 2000, 34, 24722478. Liang, L.; McNabb, J. A.; Paulk, J. M.; Gu, B.; McCarthy, J. F. Environ. Sci. Technol. 1993, 27, 1864-1870. Smolen, J. M.; Weber, E. J.; Tratnyek, P. G. Environ. Sci. Technol. 1999, 33, 440-445. Heron, G.; Christensen, T. H.; Tjell, J. C. Environ. Sci. Technol. 1994, 28, 153-159.
Received for review April 10, 2003. Revised manuscript received September 1, 2003. Accepted September 21, 2003. ES0343297
VOL. 37, NO. 24, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5731
