Modification of Chemical Oxygen Demand Monitoring in the Yellow River, China, With a High Content of Sediments

By Ni, Jinren Sun, Liyiing; Sun, Weiling

ABSTRACT: This paper presents a modification of chemical oxygen demand (COD) monitoring giving a better indication of the pollution level compared with the conventional COD method for rivers with a high content of sediments. The correlation between the sediment organic carbon and COD was investigated using sediments sampled in the middle Yellow River, China. Partitioning of the sediment organic carbon between the water and sediment phases was also investigated using batch experiments, with the sediment concentration varying from 20 to 400 g/L. As a result, the COD modification equations are proposed for both turbid water (mixture of water and sediment) and supernatant water (filtrate using a 0.45-[mu]m membrane). The modified COD in turbid water and supernatant water could be 40 and 10% less than the monitored COD values, respectively. These results may have a significant influence on the assessment of water quality class in the Yellow River. Water Environ. Res., 79, 2336 (2007).

KEYWORDS: chemical oxygen demand, sediment organic carbon, sediment, partitioning, Yellow River.

doi:10.2175/106143007X183790

Introduction

Chemical oxygen demand (COD) is an important constituent reflecting the oxygen demand of organic pollutants. The COD measures the oxygen equivalents consumed in the oxidation of organic compounds by strong oxidizing reagents, such as potassium dichromate, eerie sulfate, potassium iodate, and potassium permanganate (Dharmadhikari et al., 2005). Potassium dichromate has been widely used because of its high oxidizing capability to a wide range of organic pollutants (Zhao et al., 2004). Moreover, potassium dichromate has been used for COD determination in the China National Standard (GB 3838-2002).

Problems have been raised for the COD measurement using potassium dichromate in rivers with a high content of sediments, such as in the Yellow River, China. For example, the suspended sediment concentration is approximately 200 to 300 g/L and could be up to 1650 g/L in the middle Yellow River in flood season (Zhao et al., 1998). More organic pollutants would be absorbed intensively in the hyperconcentrated, sediment-laden flows, and the absorbed organic pollutants may be released to river water under certain conditions. Therefore, COD in both supernatant water (filtrate using a 0.450- [mu]m membrane) and turbid water (the mixture of water and sediment) of the sediment-laden flows should be measured, reflecting the oxygen-demanding characteristic of organic pollutants in the water and sediment phases, when the sediment concentration is greater than 0.2 g/L (Water Environmental Monitoring Center of the Yellow River Basin, 2001).

In recent years, COD has been the parameter exceeding the permissive limit most frequently among the monitored parameters for water quality of the Yellow River (Zhao et al., 1998). In fact, the monitored COD values have often misinterpreted the water quality of the Yellow River. For example, the monitored COD values of the Yellow River could be nearly 6 to IQ times of the most contaminated urban rivers in China, which obviously should have higher COD values (Chen et al., 2006).

The question is if the conventional technique applied when determining COD could be properly used for reflecting oxygen- consuming pollution in the Yellow River. Chen et al. (2006) have indicated that natural sediment organic carbon (SOC) could be oxidized by potassium dichromate when the sample is digested in a high temperature and strongly acidic solution. Thus, the increasing natural SOC concentration with increasing sediment concentration may result in the increasing COD values with increasing sediment concentration.

However, the natural SOC should not be confused with oxygendemanding pollutants. It is well-known that severe organic pollutants derived from diverse human activities may lead to rapid deoxygenating of the river water (Apsite and Klavins, 1998). The natural organic matter is decomposed very slowly under natural river conditions, because it mainly consists of nonbiodegradable compounds, such as humic substances (Dudal and Gerard, 2004; Hongve et al., 1999; Zouboulis et al., 2004).

The conventional monitored COD in turbid water is an overall indicator for the oxygen-consuming characteristic of SOC and organic pollutants in both the water and sediment phases. Moreover, the conventional monitored COD in supernatant water may also lead to exaggerated reports of organic pollution in the water phase, because the natural SOC may provide dissolved organic carbon to the river water by partitioning (Alberts and Takacs, 2004; You et al., 1999). Therefore, both the monitored COD in turbid water and the monitored COD in supernatant water using the conventional method could not be used for reflecting the true water quality of the Yellow River. The key task for modification of the monitored COD in the Yellow River is to identify the contribution of SOC to COD.

Materials and Methods

Sampling. Surface bed sediments (BS) and suspended sediments (SS) were collected at each of the representative monitoring stations during the flood season (i.e., SanMenXia [SMX], TongGuan [TG], and HuaYuanKou [HYK]) in the middle Yellow River. A total of six types of sediment samples were obtained-SMX-BS, SMX-SS, TG-BS, TG-SS, HYK- BS, and HYK-SS. For comparison, loess was also collected in the catchment area, from where most of the river sediments originate (Zhao et al., 1998). Sediments and loess were air-dried and sieved to particulates <0.063 mm and <0.1 mm. The mineral composition of the collected samples was analyzed using a Dmax 2400 X-ray diffractometer (Rigaku, Japan). The total concentration of the sediment organic carbon (SOC*) in the collected samples was analyzed by a carbon/hydrogen/nitrogen analyzer (Elemental Vario El, Hanau, Germany) using dried and homogenized sediments, which were first acidified using 10% hydrochloric acid overnight to remove carbonate and then dried at 60[degrees]C.

Partitioning Experiments. Deionized water was used for the partitioning experiments in the laboratory. Batch experiments were conducted in 150-mL bottles with 0.01 M NaNO^sub 3^ controlling the ionic strength of the solution. Sediment concentration ranged from 20 to 400 g/L. No apparent changes of pH were found in the solution because of the high carbonate content in the collected sediments or loess. The prepared samples were shaken at room temperature (25 +- 1[degrees]C) for 4 hours and left standing for 20 hours. After equilibration, supernatant water was separated from the mixture of the water-sediment system using a 0.45-um glass fiber membrane. The concentration of dissolved organic carbon in supernatant water after filtration was determined using a Multi NC/ 3000 total organic carbon analyzer (Analytik, Jena, Germany).

Chemical Oxygen Demand Measurement The COD was measured using a closed microwave digestion system (FR-100 Microwave Laboratory Systems, Chengdu, China) in the laboratory. Potassium dichromate (analytically pure) was used as the oxidation reagent, and the samples were digested for 10 minutes. The measured COD in the mixture of deionized water and sediment before filtration in the laboratory represents the contribution of SOC to COD in turbid water. The measured COD in the filtrate in the laboratory represents the contribution of SOC to COD in supernatant water, which is mainly attributed to the dissolved organic carbon originating from the SOC.

Results and Discussion

Sediment Characteristic. As shown in Table 1, SOC* in the collected sediments and loess ranged from 3.15 to 5.64 mg/g by dry weight Figure 1 shows a strong positive linear relationship between the clay mineral composition (including montmorillonite, illite, amphibole, kaolinite, and chlorite) and SOC* in the collected sediments and loess (R^sup 2^ [coefficient of determination] = 0.83). This suggests that the sediment organic matter in the collected sediments and loess is dominated by the humic substances strongly adsorbed on clay mineral surfaces (Guggenberger and Kaiser, 2003; Wang et al., 2001).

Contribution of Sediment Organic Carbon to Chemical Oxygen Demand in Turbid Water. Monitored COD in turbid water using the conventional method is an overall indicator for the oxygen- consuming characteristics of SOC and organic pollutants in both water and sediment phases, which could be described as follows:

Where

COD^sub T^ – monitored COD in turbid water using the conventional method (mg/L),

COD^sub TP^ = contribution of organic pollutants to COD^sub T^ (mg/L), and

COD^sub TS^ = contribution of SOC to COD^sub T^ (mg/L).

The parameter COD^sub TS^ could be calculated based on the relationship of SOC* and the calculated COD value per unit mass of sediment (COD^sub S^) in the laboratory. Figure 2 shows a very good correlation between COD^sub S^ and SOC* (R^sup 2^ = 0.98). Sediment concentration (5*) is 50 g/L for all runs in the experiments using six types of sediment samples with grain size <0.063 mm (SMX-BS, SMX- SS, TG-BS, TG-SS, HYK-BS, and HYK-SS). The calibrated ratio of COD^sub S^ and SOC* is 2.12, which is approximately 82% of the theoretical ratio (oxygen/carbon) of 2.6. This implies that most SOC could be oxidized by potassium dichromate. Based on the definition of COD^sub S^ and the relationship between COD^sub S^ and SOC*, COD^sub TS^ could be described as follows:

From eq 2, we can see that COD^sub TS^ will increase significantly with increasing sediment concentration. Eqs 1 and 2 suggest that COD^sub T^ could be magnified, even without any effect of organic pollutants (COD^sub TP^ = 0) on the contribution of SOC (COD^sub TS^) being significant with high sediment concentration.

Figures 3a and 3b show that the measured COD^sub TS^ in the laboratory agrees well with the calculated COD^sub TS^ from eq 2 for sediments <0.063 mm, with varying sediment concentration. Although eq 2 is obtained using sediments <0.063 mm, it could also be applied to calculate COD^sub TS^ for particulates <0.1 mm, with varying sediment concentration (see Figures 3c and 3d).

Contribution of Sediment Organic Carbon to Chemical Oxygen Demand in Supernatant Water. Monitored COD in supernatant water using the conventional method is an overall indicator for the oxygen- consuming characteristic of both dissolved organic pollutants and natural dissolved organic carbon originating from SOC, which could be described as follows:

Where

COD^sub W^ = monitored COD in supernatant water using the conventional method (mg/L);

COD^sub WP^ = contribution of dissolved organic pollutants to COD^sub W^ (mg/L); and

COD^sub WS^ = contribution of SOC to COD^sub W^, which is mainly attributed to the dissolved organic carbon originating from SOC (mg/ L).

The concentration of the dissolved organic carbon in the water phase provided by SOC could be quantified basing on the partitioning of water soluble organic carbon (WSOC) between the water and sediment phases (Tao and Lin, 2000). According to Tao and Lin (2000), WSOC is the entire pool of organic carbon in SOC available to release into surface water, regardless of being extracted or not, which could be further divided into two subpools-those sorbed on sediment and those dissolved in the pore water-when sorption- desorption equilibrium is reaching, as follows:

Where

WSOC* = total concentration of the whole pool of WSCXU (mg/g),

C^sub e^ = equilibrium WSOC concentration in water (mg/L), and

q^sub e^ = equilibrium WSOC concentration in sediment (mg/g).

The partitioning mechanism is dominated by the nonspecific hydrophobic sorption and described by the linear-type isotherm (Tao and Lin, 2000; Weber et al., 1992), which can be expressed as follows:

Where

k (L/g) = partitioning coefficient of WSOC.

With eqs 4 and 5, we obtain the following:

Considering transformation between oxygen and carbon, CODWS could be estimated as follows:

Where

alpha = ratio of oxygen and carbon (or COD^sub S^/SOC*).

Here, alpha = 2.12, according to the experimental results (see Figure 2). With varying sediment concentration (S*), the equilibrium concentration of WSOC in water (C^sub e^) could be determined through batch partitioning experiments. Thus, WSOC* and k could be obtained from eq 6 using the STATISTICA software (StatSoft, Tulsa, Oklahoma) with a 95% confidence level (see Table 2), according to the results of multiple partitioning experiments.

As shown in Table 2, the calibrated WSOC* values range from 0.012 to 0.029 mg/g for sediments from the Yellow River. For loess<0.063 mm and loess<0.1 mm, the calculated WSOC* values are 0.028 and 0.065 mg/g, respectively. The calculated k values range from 0.0003 to 0.0027 Ug for sediments from the Yellow River. For loess<0.063 mm and loess<0.1 mm, the calculated k values are close to 0.0012 L/g.

As shown in Figure 4, the COD^sub WS^ values calculated from eq 7 are well fitted with the measured COD^sub WS^ values in the laboratory for sediments <0.063 mm (see Figure 4a) and sediments <0.1 mm (see Figures 4b and 4c). From Figures 3 and 4, we can see that the COD^sub TS^ and COD^sub WS^ increase with increasing sediment concentration. It seems that COD^sub TS^ and COD^sub WS^ will be significant when the sediment concentration is extremely high in the Yellow River during the flood season.

Modification of the Monitored Chemical Oxygen Demand. The proposed method for modification of COD^sub T^ in the Yellow River is based on a deduction of COD^sub TS^, which could be expressed as follows:

Where

MCOD^sub T^ = modified COD in turbid water (mg/L).

Although the parameters WSOC* and k could be calibrated from eq 6, multiple partitioning experiments are troublesome. Empirical parameters are expected when modifying COD^sub W^. Compared with WSOC*, SOC*could be directly measured using proper analytic equipment. To replace WSOC* with SOC*. COD^sub WS^ could be further expressed as follows:

Where

theta = proportion of WSOC* to SOC* (WSOC*/SOC*) in sediment.

The theta ranges from 0.29 to 0.57% for sediments <0.063 mm and from 1.33 to 1.56% for sediments <0.1 mm (see Table 2). As shown in Table 2, sediment grain sizes may have influences on the 5OC*, k, and theta. The SOC* is higher in sediments <0.063 mm, while theta and k are lower in sediments <0.063 mm than in sediments <0.1 mm. This may imply higher SOC* and lower theta and k values in the finer size fraction of sediments in the Yellow River. Information in Table 2 also indicates that theta is much higher in loess than in sediments from the Yellow River. This characteristic of the loess should be considered when a large volume of loess is eroded into the Yellow River during the flood season.

Based on eq 9, the proposed method for modification of COD^sub W^ could be expressed as follows:

Where

MCOD^sub W^ = modified COD in supernatant water (mg/L).

The MCOD^sub T^ and MCOD^sub W^ could be used for reflecting the true oxygen-consuming characteristics of the organic pollutants in turbid water and supernatant water, respectively.

The modified results (calculated from eqs 8 and 10) are listed in Table 3, in which the monitored COD^sub T^ COD^sub W^, and biochemical oxygen demand (BOD) in turbid water before modification are taken from literature (Guo, 1990). Moreover, the water quality class (classified based on the Chinese Environmental Quality Standard for Surface Water, GB 3838-2002) is also listed in Table 3 for comparison.

The SOC*, theta, and k are key parameters to calculate MCOD^sub W^, according to eq 10. The median empirical values of theta and k for sediments <0.1 mm are taken (theta = 1.4%; k = 0.0012 L/g) when calculating the MCOD^sub W^ in this paper, because the larger size fractions of sediments may be eroded into the sediment-laden flows in the Yellow River. The SOC* is taken as 4.6 mg/g, according to the reported average SOC* in sediments of the Yellow River (Zhao et al., 1998). As shown in Table 3, COD^sub WS^ could be up to 18% of the COD^sub W^, which implies that the MCODw could be 10% less than the CO^sub W^,

The SOC* is also taken as 4.6 mg/g when calculating MCOD^sub T^, according to eq 8 in this paper. As shown in Table 3, COD^sub TS^ accounted for more than 40% of COD^sub T^ and could be up to 70% when the sediment concentration is high. This implies that the equivalent proportion should be deducted from COD^sub T^.

As shown in Table 3, COD^sub T^ has exceeded the highest permissible limit of the surface water (>V), which suggests heavy organic pollution in the Yellow River. This conflicts with the monitored BOD, which suggests that the organic pollution is controlled in the limit of the surface water (IV). Moreover, the ratio of BOD and COD^sub T^ (BOD/COD^sub T^) is less than 5% (see Table 3), and it is much lower than the empirical 40 to 80% of the organic pollutants in wastewater (Eckenfelder, 2002; EL-Rehaili, 1995) and 10 to 60% in river water with heavy organic pollution in China (Zhao et al., 1998).

It seems that MCOD^sub T^ gives a more satisfactory interpretation of water quality of the Yellow River than COD^sub T^ (see Table 3). The ratio of BOD to MCOD^sub T^ (BOD/MCOD^sub T^) becomes 2 to 5 times the BOD/COD^sub T^ and is close to 10% after modification (Table 3). Moreover, with the elimination of sediment effects, the MCOD^sub T^ is no longer linear correlated with the sediment concentration like COD^sub T^ (see Figure 5). Hence, the modified COD values are helpful for improving understanding of the real situation of organic pollution in rivers with a high content of sediments.

Conclusions

The conventional method for determination of COD in water flows with a high content of sediments could not appropriately reflect the real situation of organic pollution in some rivers, such as the Yellow River. Therefore, a modification was needed to identify the sediment effects based on the quantifying of the contribution of SOC to the monitored COD.

The new method for COD modification is proposed to be used in the Yellow River. Laboratory experiments were conducted with a mixture of deionized water and sediments collected from the Yellow River or loess. The modification for the supernatant water (filtrate using a 0.45-[mu]m membrane) is primarily based on the mechanism of the partitioning of water-soluble organic carbon between the water and sediment phases. The results show that the modified COD in supernatant water could be 10% less than the monitored COD value. For turbid water (the mixture of water and sediment), however, more than 40% of the monitored COD is attributed to sediment organic carbon. The modified COD is of great significance when studying organic pollution in sediment-laden rivers and judging their water quality class.

Nomenclature

COD = Chemical oxygen demand (mg/L)

SOC = Sediment organic carbon (mg/g)

BOD = Biochemical oxygen demand (mg/L)

WSOC = Water soluble organic carbon in sediment (mg/g)

SOC* = Total concentration of the sediment organic carbon (mg/g)

COD^sub T^ = Monitored COD in turbid water (the mixture of water and sediment) using the conventional method (mg/L)

COD^sub TP^ = Contribution of organic pollutants to COD in turbid water (mg/L) COD^sub TS^ = Contribution of sediment organic carbon to COD in turbid water (mg/L)

COD^sub W^ = Monitored COD in supernatant water (filtrate using a 0.45-[mu]m membrane) using the conventional method (mg/L)

COD^sub WP^ = Contribution of organic pollutants to COD in supernatant water (mg/L)

COD^sub WS^ = Contribution of sediment organic carbon to COD in supernatant water, which is mainly attributed to the dissolved organic carbon originating from the sediment organic carbon (mg/L)

COD^sub S^ = Calculated COD value per unit mass of sediment (mg/ g)

MCOD^sub T^ = Modified COD in turbid water (mg/L)

MCOD^sub W^ = Modified COD in supernatant water (mg/L)

C^sub e^ = Equilibrium concentration of water soluble organic carbon in water (nig/L)

q^sub e^ = Equilibrium concentration of water soluble organic carbon in sediment (mg/g)

k = Partitioning coefficient of water soluble organic carbon between water and sediment phases (L/g)

WSOC* = Total concentration of the whole pool of water soluble organic carbon in sediment (mg/g)

alpha = Ratio of oxygen and carbon (or COD^sub S^/SOC*)

theta = Proportion of the water soluble organic carbon to the total concentration of the sediment organic carbon (WSOCVSOC*)

S* = Sediment concentration (g/L)

Credits

Financial support is from the major state basic research program of the People’s Republic of China under Grant No. 1999043603.

Submitted for publication November 5,2006; revised manuscript submitted February 12, 2007; accepted for publication February 26,2007.

The deadline to submit Discussions of this paper is January 15, 2008.

References

Alberts, J. J.; Takacs, M. (2004) Total Luminescence Spectra of IHSS Standard and Reference Fulvic Acids, Humic Acids and Natural Organic Matter: Comparison of Aquatic and Terrestrial Source Terms. Org. Geochem., 35, 243-256.

Apsite, E.; Klavins, M. (1998) Assessment of the Changes’of COD and Color in Rivers of Latvia During the Last Twenty Years. Environ. Int., 24, 637-643.

Chen, J. S.; Yu, T.; Ongley, E. (2006) Influence of High Levels of Total Suspended Solids on Measurement of Cod and Bod in the Yellow River. China. Environ. Monit. Assess., 116, 321-334.

Dharmadhikari, D. M.; Vanerkar, A. P.; Barbate, N. M. (2005) Chemical Oxygen Demand Using Closed Microwave Digestion System. Environ. Sd. Technol, 39, 6198-6201.

Dudal, Y.; Gerard, F. (2004) Accounting for Natural Organic Matter in Aqueous Chemical Equilibrium Models: A Review of the Theories and Applications. Earth Sd. Rev., 66, 199-216.

Eckenfelder, W. W. Jr. (2002) Industrial Water Pollution Control, 3rd ed.; Tsinghua University Press, McGraw Hill: Beijing, China.

EL-Rehaili, A. M. (1995) Response of BOD, COD and TOC of secondary Effluents to Chlorination. Water Res., 29, 1571-1577.

Guggenberger, G.; Kaiser, K. (2003) Dissolved Organic Matter in Soil: Challenging the Paradigm of Sorptive Preservation. Geoderma., 113, 293-310.

Guo, H. C. (1990) Princlpium Study on the Effects of Suspended Sediment on Water Quality of the Middle Yellow River, Study Corpus on the Water Resource Protection of Yellow River; Peking University Press: Beijing, China, 70-76.

Hongve, D.; Baann, J.; Becher, G.; Beckmann, O. A. (1999) Experiences from Operation and Regeneration of An-Anionic Exchanger for Natural Organic Matter (NOM) Removal. Water Sd. Technol, 40, 215- 221.

Tao, S.; Lin, B. (2000) Water Soluble Organic Carbon and Its Measurement in Soil and Sediment. Water Res., 34, 1751-1755.

Wang, X. C.; Zhang, Y. X.; Robert, F. C. (2001) Distribution and Partitioning of Polycyclic Aromatic Hydrocarbons (PAHs) in Different Size Fractions in Sediments from Boston Harbor, United States. Mar. Pollut. Bull., 42,1139-1149.

Water Environmental Monitoring Center of the Yellow River Basin (2001) Technical Regulation of Sampling and Preparing on Water Environment of the Heavy Sediment-Laden River; China Water Power Press: Beijing, China.

Weber, W. J.; Jr., McGinley, P. M.; Katz, L. E. (1992) A Distributed Reactivity Model for Sorption by Soils and Sediments. 1. Conceptual Basis and Equilibrium Assessments. Environ. Sd. Technol., 26, 1955-1962.

You, S. J.; Yin, Y. J.; Alien, H. E. (1999) Partitioning of Organic Matter in Soils: Effects of pH and Water/Soil Ratio. Sc/. Total Environ., 227, 155-160.

Zhao, H. J.; Jiang, D. L.; Zhang, S. Q.; Catterall, K.; John, R. (2004) Development of a Direct Photoelectrochemical Method for Determination of Chemical Oxygen Demand. Anal. Chem., 76, 155-160.

Zhao, P. L.; Shen, X. C.; Xia, J.; Li, Q. F.; Gao, H.; Zhang, S. G. (1998) Effects of the Sediment on the Water Quality and Pollution Control on Major Reaches of the Yellow River; Yellow River Conservancy Press: Zhengzhou, China.

Zouboulis, A. L; Chai, X. L.; Katsoyiannis, I. A. (2004) The Application of Bioflocculant for the Removal of Humic Acids from Stabilized Landfill Leachates. Environ. Manage., 70, 35-41.

Jinren Ni1,2*, Liyiing Sun1,2, Weiling Sun1,2

1 Department of Environmental Engineering, Peking University, Beijing, China.

2 Key Laboratory of Water and Sediment Sciences, Ministry of Education, Beijing, China.

* Department of Environmental Engineering, Peking University, Beijing 100871, China; e-mail: nijinren@iee.pku.edu.cn.

Copyright Water Environment Federation Oct 2007

(c) 2007 Water Environment Research. Provided by ProQuest Information and Learning. All rights Reserved.