Biosensors and Bioelectronics 34 (2012) 7782

Contents lists available at SciVerse ScienceDirect

Biosensors and Bioelectronics

journal homepage: www.elsevier.com/locate/bios

A biolm reactor-based approach for rapid on-line determination of biodegradable organic pollutants
Changyu Liu a , Huijun Zhao b, , Lijie Zhong a , Chang Liu a,b , Jianbo Jia a , Xiaolong Xu a , Ling Liu a , Shaojun Dong a,
a b

State Key Laboratory of Electroanalytical Chemistry, Changchun Institute of Applied Chemistry, Chinese Academy of Sciences, Changchun 130022, China Centre for Clean Environment and Energy, and Grifth School of Environment, Gold Coast Campus, Grifth University, Queensland 4222, Australia

a r t i c l e

i n f o

a b s t r a c t
A new analytical approach utilizing a biolm reactor (BFR) for rapid online determination of biochemical oxygen demand (BOD) was proposed and experimentally validated. The BFR was fabricated via a cultivation process using naturally occurring microbial seeds from locally collected wastewaters. The resultant BFR displays a remarkable rate of biodegradation towards a wide spectrum of organic substrates, capable of degrading over 20% of biodegradable organic substrates within 100 s. More importantly, the BFR exhibits a superior indiscriminative biodegradation feature, enabling a precise prediction of BOD values of total biodegradable organics based on experimentally determined BOD values from partial degradation processes without a need for on-going calibration. The proposed approach was systematically validated using a range of individual organic substrates, their mixtures, synthetic samples and wastewaters. Highly signicant linear correlations between the BFR and the standard BOD5 methods were obtained from diversied synthetic samples (r = 0.988, p = 0.000, n = 45) and wastewaters (r = 0.983, p = 0.000, n = 40). Near unity slope values of the principal axis of the correlation ellipse were obtained from all tested samples, suggesting both methods were essentially measuring the same BOD value. The reported method could be a useful online monitoring tool for determination of biodegradable organic pollutants. 2012 Elsevier B.V. All rights reserved.

Article history: Received 25 October 2011 Received in revised form 6 January 2012 Accepted 17 January 2012 Available online 26 January 2012 Keywords: Biolm reactor Biochemical oxygen demand (BOD) Biodegradation Rapid Online

1. Introduction Analytical determination of aggregate organic pollutants in water and wastewaters is critically important for water quality assessment (APHA, 1997; Liu et al., 2005, 2009a,b,c, 2010a, 2011; Jia et al., 2003; Catterall et al., 2003; Jia and Dong, 2003; Zhao et al., 2004; Zhang et al., 2006, 2009; Morris et al., 2005). Analytically, chemical oxygen demand (COD) and biochemical oxygen demand (BOD) are commonly used methods to quantify aggregate organic pollutions (Spicer et al., 1985; Dasgupta and Petersen, 1990; Kim et al., 2000; Preininger et al., 1994; Catterall et al., 2001). Although COD method is suitable alternative for heavily polluted industrial wastewaters because of the inhibited biodegradation in such sample matrixes, BOD has been the most preferred method for environmental applications because it determines only the biodegradable components of the aggregate organic pollutants, providing useful information for biological and environmental impact assessment (Morris et al., 2001). For these reasons, BOD analysis has been regarded as one of the most widely used analytical

Corresponding author. Tel.: +61 7 55528261. Corresponding author. Tel.: +86 431 8526 2101. E-mail addresses: h.zhao@grifth.edu.au (H. Zhao), dongsj@ciac.jl.cn (S. Dong). 0956-5663/$ see front matter 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.bios.2012.01.020

methods and routinely performed by governmental and legislative authorities, water, wastewater and other industrial bodies (Guwy et al., 1999). In this regard, the standard BOD5 method has been the most widely employed method for determining biodegradable organic compounds in waters and wastewaters. The widespread uses of BOD5 are mainly due to its distinctive advantage of universal applicability to a wide spectrum of sample matrixes (Morris et al., 2001). However, the BOD5 method has disadvantages. The method requires 5 days to complete an assay, involves tedious procedures, and needs highly skillful personnel to obtain reproducible results (Zhang et al., 2004). These disadvantageous features, especially the 5-day duration assay time, make BOD5 method unsuitable for real-time, online applications. To date, considerable efforts have been devoted to develop new BOD methods with reduced assay time and simplied procedures to meet the demands of end-users (Liu et al., 2005, 2011; Jia et al., 2003; Catterall et al., 2001, 2003; Preininger et al., 1994; Morris et al., 2001; Thavarungkul et al., 2008; Wang et al., 2010; Chang et al., 2004; Chan et al., 2000; Chee et al., 2005; Pang et al., 2007; Lin et al., 2006; Sakaguchi et al., 2007; Kwok et al., 2005; Nakamura et al., 2007a,b; Chen et al., 2008; Liu and Mattiasson, 2002). To this end, biosensor approaches have been widely adopted to realize rapid BOD assay (Liu et al., 2005, 2011; Jia et al., 2003; Preininger et al., 1994; Thavarungkul et al., 2008;

78

C. Liu et al. / Biosensors and Bioelectronics 34 (2012) 7782

Wang et al., 2010; Chang et al., 2004; Chan et al., 2000). These approaches employ immobilized forms of articially selected and purely cultured microorganisms to achieve biodegradation while the measured rate of respiration are used for BOD quantication (Chan et al., 2000; Chee et al., 2005; Pang et al., 2007). Such methods exhibit an inherent limitation of narrow applicability because microorganisms employed are only capable of preferentially biodegrading a small fraction of easier biodegradable organic substrates in the sample, resulting in the measured analytical signal highly dependent on the sample matrix (Liu et al., 2000, 2011; Catterall et al., 2003). The selective degradation nature of rapid BOD sensors is the root of many problems that limits their practical use (Catterall et al., 2001, 2003; Morris et al., 2001). This is because, in practice, substrate compositions and sample matrixes can be highly diversied and variable. BOD values predicted based on analytical signals originated from the degradation of small fraction of easier degradable organic substrates can therefore be unauthentic (Catterall et al., 2003; Liu et al., 2000). In strong contrast to rapid BOD sensors, the traditional BOD5 method employs naturally occurring microorganisms for biodegradation and quanties analytical signal based on exhaustive degradation model (APHA, 1997; Catterall et al., 2003). The use of naturally occurring microorganisms allows indiscriminate degradation of wide spectrum of organic substrates while the analytical signal quantied from an exhaustive degradation process represents the total biodegradable organics in a sample. In such a fashion, BOD5 method, in theory, determines the total biodegradable organics and highly tolerant to sample composition/matrix variations, earning a reputation as universal method (APHA, 1997; Bourgeois et al., 2001). In fact, the rapidity achieved by rapid BOD sensors is at the cost of sacricing distinctive advantages of BOD5 method (Catterall et al., 2003). The development of new BOD methods capable of achieving rapidity (inherent advantage of a rapid BOD sensor) while retains the wide applicability of BOD5 method is of a scientically challenging and practical important issue. In this study, we propose a new analytical approach that embraces inherent advantages of both rapid BOD sensors (rapidity) and BOD5 method (wide applicability). This new approach utilizes a biolm reactor (BFR) cultivated with naturally occurring microorganisms to achieve rapid and indiscriminative biodegradation. Such a BFR-based method is capable of rapid online BOD determination and applicable to a wide spectrum of sample matrixes. A new analytical signal quantication/calibration method was also proposed to enable the BOD determination without the need for on-going calibration. Systematic validations were performed using a range of individual organic substrates, their mixtures, synthetic samples and wastewaters. 2. Experimental 2.1. Materials and sample preparation Multi-walled carbon nanotubes (MWNTs) were purchased from Nanoport Co., Ltd. (China). Poly(diallyldimethylammonium chloride) (PDDA, 200,000350,000 Mw) was purchased from Sigma. The positively charged PDDA functionalized MWNTs (PDWNTs) were prepared according to our previous study (Wang et al., 2007). All chemicals used in this study were of analytical reagent grade and all solutions were prepared using high-purity deionized (Milli-Q) water. All synthetic samples including the glucose-glutamic acid (GGA) solution, the Organization for Economic Cooperation and Development (OECD) recommended synthetic sample, and the articial wastewater (AWW) were prepared in accordance with reported methods (APHA, 1997; Kim et al., 2000; OECD, 1981). The BOD5 values of the as-prepared GGA, OECD, and AWW (with a

10-fold increase in concentration) stock solutions were 188 15, 1398 112, and 32.4 3.0 mg O2 L1 , respectively. The wastewaters used in this study were collected from various sites including wastewater treatment plants (WWTPs), dairy/food manufactures, sugar plants, sh farms, lakes and rivers in Jilin Province of China. All wastewaters were preserved according to the guideline of the standard method (APHA, 1997). When necessary, both synthetic samples and wastewaters were diluted to a suitable concentration prior to the analysis. After dilution, the same sample was simultaneously analyzed by both standard BOD5 (see Supplementary data) and BFR method. 2.2. Fabrication of biolm reactor The glass tube (inner diameter 1.04.0 mm, length ranged from 30 to 180 cm) was rstly treated with HF/NH4 F (1.7%/2.3%, w/w) solution, followed by thoroughly washing with water to obtain a rough inner surface. A negatively charged inner tube surface was then created by sonicating the glass tube lled with 1 M NaOH for 30 min. After thoroughly washed with water, the tube was relled with PDWNTs suspension solution, attaching PDWNTs to the inner surface of the tube. The resultant tube was washed with copious amount of water and dried with nitrogen stream. Wastewaters collected from municipal WWTPs were used as the cultivation solution for biolm formation. Air-saturated cultivation solution with added nutrients was continually pumped through the functionalized tube at a ow rate of 0.5 mL min1 under a constant temperature of 30 C. The status of biolm formation was estimated by measuring the current responses of a dissolved oxygen (DO) probe to an injected GGA solution (equivalent to 16.0 mg O2 L1 BOD) at a 4 h interval. The gradually decreased current signal with increased cultivation time indicates the progressive biolm formation process. The cultivation process was terminated when no further decrease in current signal was observed from the injections of the GGA solution in two consecutive time intervals. The resultant BFR (see Fig. S1, Supplementary data) was lled in phosphate buffer saline (PBS) and stored at room temperature before use. 2.3. Measurement procedures Fig. 1a shows the schematic diagram of the BFR online BOD measurement system. A three-electrode DO probe with a Au working electrode ( = 0.8 mm) covered by the Teon membrane (Orbisphere 2956A) was used and all current signal measurements were performed under a constant applied potential of 700 mV vs. Ag/AgCl (0.1 M KCl), controlled by an electrochemical workstation (CHI 832B, Chenhua Company, China). Unless otherwise stated, the BFR and PBS buffer solution were placed in a thermostatic chamber under a constant temperature of 30 C. A plug-ow operation mode was employed. The response of DO probe was evaluated with the steady-state method (Liu et al., 2000). Between sample injections, the air-saturated PBS was pumped through the BFR, removing any remaining organic substrates. Once the DO probe achieved a steady-state current (ib ), the air-saturated sample solution was then continuously injected into the BFR while the oxygen depletion of the efuent was simultaneously monitored by the current signal decrease until a steady-state current (is ) was achieved. The change in the steady-state current ( i = ib is ) was obtained and used as the analytical signal for BOD quantication (Fig. 1b). 3. Results and discussion 3.1. Analytical signal generation In theory, the microbial composition of the biolm should be directly related to the microbial strains or consortia of the

C. Liu et al. / Biosensors and Bioelectronics 34 (2012) 7782

79

Fig. 1. Schematic diagram of experimental set up (a) and a set of typical oxygen probe responses to GGA solutions. Insert: linear relationship of the steady-state current change ( i) and GGA concentrations (b).

cultivation medium. In this work, the BFR cultivated with microbial seeds from the wastewater should contain naturally occurring microbial strains or consortia of the wastewater that are the most robust and best suit for consumption of substrates in wastewater streams. Fig. 1b shows a set of typical DO probe responses. It requires 400500 s to achieve a near steady-state current, which is determined by the response time of the DO probe and ow rate. A linearly decreased steady-state current was observed when the tested organic substrate concentrations were increased. Such a liner decrease in the steady-state current implies a linearly increased oxygen consumption (decreased oxygen concentration in the efuent) caused by the biodegradation actions of the BFR. In fact, the linear relationship between the substrate concentration and oxygen consumption demonstrates that the BFR employed could be used for BOD determination as it is capable of degrading organic substrate, and importantly, the extent of degradation is directly proportional to the substrate concentrations 3.2. Analytical signal optimization The diameter () and length (L) of BFRs, and ow rate were collectively optimized. For a BFR ( = 2.0 mm and L = 105 cm), 2.0 mL min1 was found to be the optimal ow rate. Under such conditions, it takes 99 s for an injected sample passing through the reactor to reach the DO probe. The effect of temperature on the catabolic activity of the BFR was investigated using a GGA test solution with 16.0 mg O2 L1 BOD, under different temperatures (see Fig. S2, Supplementary data). The obtained analytical signal was found to increase as temperature was increased from 20 to 45 C, similar to the results reported by Kowk and co-workers for

activated sludge system (Kwok et al., 2005). The effect of pH on the catabolic activity of the microorganisms was examined under 30 C with pHs varied from 4.6 to 8.4 (see Fig. S3, Supplementary data). Insignicant current response changes were observed within the pH range investigated, which is conspicuously different to previous reports where the biosensor response depends strongly on pH changes (Jia et al., 2003; Pang et al., 2007). In order to prevent the possible nitrication in the biolm reactor, an adverse condition of pH 6.5 was selected in the following studies. The effect of PBS concentration was also investigated under 30 C and pH 6.5 (see Fig. S4, Supplementary data). The optimal PBS concentration range was found to be between 5 and 100 mM. The insensitivity to pH and ionic concentration, coupled to the high temperature tolerance nature make the BFR method suitable for variety of analytical environments. The linear range, detection limit and reproducibility were evaluated under the optimized conditions using GGA as the testing samples. A linear range up to 30.0 mg O2 L1 and a practical detection limit of 0.5 mg O2 L1 were obtained. A set of relative standard deviations of 3.9%, 3.2% and 2.8% were obtained from seven successive tests of 4.0, 12.0 and 25.0 mg O2 L1 GGA samples, respectively. 3.3. Analytical signal quantication and validation Theoretically, an ideal biological recognition component for BOD sensors should be able to completely degrade all biodegradable substrates in a sample regardless of their chemical structure and concentration. However, in practice, to exhaustively degrade all biodegradable substrates within a short time frame is extremely difcult. Despite this, a partial degradation is achievable within

80

C. Liu et al. / Biosensors and Bioelectronics 34 (2012) 7782

a short time frame and the degraded portion could be used to represent biodegradable substrates of a sample if an indiscriminative degradation toward all substrates can be realized. In other word, the amount of total biodegradable substrates can be estimated by the representative degradation fraction when the extent of degradation is directly proportional to the concentration of individual substrates, regardless of the presence (or absence) of others (independent of sample matrix). The biodegradation characteristics of the BFR were hence investigated. Fig. S5, Supplementary data shows the plots of the steady-state current change ( i) against substrate concentration (measured BOD5 values as reference) for a range of sample compositions including single organic substrates, their mixtures, and common synthetic samples. Within the concentration range investigated, linear relationships between i and BOD5 values were obtained for all cases. More importantly, the Pearson correlation analysis indicated that all measured data can be statistically represented by a common line of best t, having a slope of the principal axis of the correlation ellipse value of 12.46 0.48 nA L mg1 of oxygen with a highly signicant correlation coefcient r = 0.982, at a 95% condence level. This indicates the biodegradation by BFR was indiscriminative and a same extent of biodegradation towards all substrates tested was achieved, regardless of concentrations and chemical identities of substrates. Such a desirable biodegradation feature of the BFR provides a rm foundation enabling the development of a rapid BOD detection principle based on the determinable representative portion of BOD to quantify the total biodegradable BOD value of a sample. The indiscriminative degradation nature of the BFR ensures a wide applicability, overcoming the matrix dependent problem, which has been an inherent limitation for most existing rapid BOD sensors (Catterall et al., 2001, 2003; Morris et al., 2001; Liu and Mattiasson, 2002). Such a promising biodegradation capability can be attributed to the wide spectrum substrate consumption abilities of the biolm. For a given DO probe under controlled experimental conditions, the analytical signal conversion factor (e.g., steady-state current to dissolved O2 concentration) should be a constant and can be experimentally determined. The conversion factor of our DO probe was obtained under a ow rate of 2.0 mL min1 with a series of DO concentration obtained by the controlled ratio of nitrogen and oxygen-saturated puried water, and the steady-state current of the DO probe to these solutions were recorded accordingly. For the DO probe used in this work, the relationship between the steady-state current change ( i) and dissolved oxygen concentration change ([ O2 ]) was experimentally determined, e.g., i (nA) = [ O2 ] (mg L1 ) = 56.4 [ O2 ] (mg L1 ), where is the conversion factor and can be presented in a general form of: = i (nA L mg1 ) [ O2 ] (1)

7 6
[O2 ] (mg L )
-1

5 4 3 2 1 0
0

y = (0.2210.009) x r = 0.982 p =0.000 n =99

1 3 5 7 9 11

2 4 6 8 10

10

BOD5 (mg L )
Fig. 2. Correlation between oxygen consumption [ O2 ] of samples by BFR method and corresponding BOD5 values. Temperature: 30 C; pH: 6.5; PBS concentration: 5 mM. 1, glucose; 2, glutamic acid; 3, fructose; 4, glycerol; 5, histidine; 6, asparagine; 7, malic acid; 8, succinic acid; 9, citric acid; 10, GGA; 11, mixture of compounds 19 with equal BOD concentration of each.

-1

20

30

Considering the fact that the [ O2 ] values shown in Fig. 1 are the BOD values determined by the BFR method, the physical meaning for can then be interpreted as the BOD ratio of BFR and BOD5 methods. This means that when BOD5 values are used as reference, BOD values measured by BFR method represent 22.1% of BOD5 values. Furthermore, considering that the data shown in Fig. 2 were obtained from a large number of individual substrates, their mixtures and synthetic BOD testing samples, it is reasonable to predict that the highly correlated linear relationship shown in Fig. 1 is applicable to unknown samples (sharing the same value). For a given system, given that both and values are pre-determinable under controlled experimental conditions, the BFRBOD5 value of an unknown sample can then be calculated based on the measured BFRBOD value ([ O2 ]) and predetermined and values. That is: BFR[BOD5 ] = i [ O2 ] = (mg L1 ) (2)

The obtained analytical signal conversion factor () can then be used to convert the measured steady-state current change into the equivalent dissolved oxygen concentration change. Fig. 2 shows the correlation of oxygen consumption values ([ O2 ]) with corresponding samples BOD5 values. The [ O2 ] values were converted from the steady-state current values shown in Fig. S5, Supplementary data using = 56.4 nA L mg1 . The Pearson correlation analysis was used to measure the intensity of association between [ O2 ] and BOD5 values. This was done in preference to linear regression analysis because both [ O2 ] and BOD5 values involve measurement errors. A highly signicant correlation between [ O2 ] and BOD5 values was obtained (r = 0.982, p = 0.000, n = 99). At 95% condence level, the slope of the principal axis of the correlation ellipse () was found to be 0.221 0.009. This demonstrates that all measured data shown in Fig. S5, Supplementary data can be statistically represented by a linear relationship.

In the equation, i is the measured steady-state current change by BFR system and dened as i = ib is , where ib and is are the baseline and the steady-state current of sample solution. This means that BOD5 values of an unknown sample can be obtained by simply measuring i value if value is a constant for a given DO probe and value obtained from Fig. 2 is applicable to the unknown sample. The validity of Eq. (2) is therefore evaluated using widely accepted synthetic samples including GGA, OECD and AWW. The validation was carried out over a period of two weeks and each sample were analysed in three concentrations at day 1, 2, 5, 10 and 14. Fig. 2 shows the correlation between the BOD5 values determined by BFR method (BFRBOD5 ) in accordance with Eq. (2) and by the standard BOD5 method. A highly signicant correlation (r = 0.988, p = 0.000, n = 45) between the two sets of BOD5 values was obtained, suggesting the BOD5 values obtained from the two methods are well agreed. This is also supported by the slope of the principal axis of the correlation ellipse of 0.975 0.047 with 95% condence level. Considering the organic substrates in GGA (e.g. glucose and glutamic acid) are easier to be biodegraded than that of organic substrates in OECD (e.g. peptone, meat extract, urea, etc.) and AWW (e.g., lignin, humic acid, tannic acid, gum arabic and sodium dodecylbenzenesulfonate), the data shown in Fig. 3 demonstrated that the value obtained from Fig. 2 can be applied to unknown samples. In order to further conrm the applicability of Eq. (2), over the same period, the BFR method was also validated by wastewaters collected from WWTPs, dairy/food manufactures, sugar plants, sh farms, lakes and rivers. All samples were analyzed by both BFR and

C. Liu et al. / Biosensors and Bioelectronics 34 (2012) 7782

81

30

5
y = (0.9750.047) x r =0.988 p =0.000 n =45

50

BFR-BOD5 (mg L )

-1

40 30 20

BFR-BOD 5 (mg L-1)

[O 2 ] (mg L )

-1

20

3 2 1

10

1 2 3

b 10
18 0

10
-1

20

30

BOD5 (mg L )
Fig. 3. Correlation between BFRBOD5 and standard BOD5 values for a range of synthetic samples. Temperature: 30 C; pH: 6.5; PBS concentration: 5 mM. 1, GGA; 2, OECD; 3, AWW.

12

15

Month
Fig. 5. Oxygen consumption [ O2 ] by 16.0 mg O2 L1 GGA samples (curve a) and the BFRBOD5 values to 20.0 mg O2 L1 OECD samples (curve b) over 17 months. The error bars represent standard deviations of seven measurements over the monthly test.

standard BOD5 methods. The correlation between the two methods is given in Fig. 4. The obtained Pearson correlation coefcient of r = 0.983 (p = 0.000, n = 40) demonstrates a high intensity of association between the two sets of BOD values. Furthermore, the slope of the principal axis of the correlation ellipse obtained from the Pearson analysis was found to approach an almost unity value (0.969 0.056), suggesting both methods were essentially measuring the same BOD value. The results shown in Fig. 4 further conrm that the value obtained from Fig. 2 can be used for different types of wastewaters. In other word, the value determined from a set of synthetic samples can be shared by different types of wastewaters, attributing to the indiscriminative biodegradation ability of the BFR. 3.4. Long-term stability Long-term stability is critically important for practical use of an analytical method, especially for an online system. For BFR method, the long-term stability of the bioactivity determines the applicability of the method for online applications. The bioactivity of the BFR method over a 17-month period was evaluated by determining oxygen consumption ([ O2 ]) of the reactor with injections of a GGA testing solution having an equivalent BOD values of 16.0 mg O2 L1 (Fig. 5, curve a). The data shown in Fig. 5 (curve a) were obtained from 7 injections of the GGA testing solution within a monthly interval. An approximately 22% variation on the determined [ O2 ] values were observed over the 17 months testing period. The observed changes were found to be random and without a clear trend. Considering the measured variations on [ O2 ] directly represent the changes in the bioactivity (biodegradation
30

efciency) of the BFR, the data shown in Fig. 5 (curve a) indicate a 22% variation in bioactivity. Comparing to existing rapid BOD sensors, this variation in bioactivity over such a long test period suggest a noticeable improvement in stability. Nevertheless, such a variation in biodegradation efciency would affect the accuracy of BOD measurement. However, the indiscriminative degradation nature of the BFR makes it possible to reduce the effect of the biodegradation efciency changes through a simple recalibration process. To conrm this, the value of the DO probe and value of the BFR were obtained at the beginning of each monthly test via a system calibration process and used to convert the determined [ O2 ] values into BFRBOD5 values according to Eq. (2). The values were obtained from injections of a set of GGA samples having BOD5 values ranged from 0.5 to 25.0 mg O2 L1 . The measured BFRBOD5 values to 20.0 mg O2 L1 of OECD samples over the 17 months period are shown in Fig. 5 (curve b). No signicant analytical performance (in terms of reproducibility, reliability and accuracy) decay was observed over the test period. The mean value of the determined BFRBOD5 over the test period was found to be 19.7 1.0 mg O2 L1 , demonstrating a good long-term stability. This means that the long-term stability can be readily achieved by performing a monthly calibration. This combined with other demonstrated superiorities makes the BFR method an attractive alternative for rapid online BOD determination.

4. Conclusion A new analytical approach utilizing a biolm reactor as the biodegradation means has been proposed for rapid online BOD determination and experimentally validated with a wide range of individual organic substrates, their mixtures, synthetic and real wastewaters. The obtained rapidity and indiscriminative biodegradation capability demonstrate that BFRBOD5 method embraces the inherent advantages of both rapid BOD sensors and BOD5 method. It is capable of non-selectively degrading over 20% of biodegradable substrates within 99 s. These superior biodegradation characteristics make the BFR method applicable to wide range of wastewaters as demonstrated by the highly signicant correlation (r = 0.983, p = 0.000, n = 40) and almost unity slope value of the principal axis of the correlation ellipse (0.969 0.056) between the BFR and BOB5 methods. The superiority of the BFR method can has also been demonstrated by its ability to perform long-term online analysis without the need for on-going calibration. These demonstrated BFR qualities make it a very promising analytical tool for rapid online BOD determination.

BFR-BOD5 (mg L )

20

y = (0.9690.056)x r = 0.983 p=0.000 n=40

-1

1 3 5 7 9

2 4 6 8 10

10

10

20
-1

30

BOD5 (mg L )
Fig. 4. Correlation between BFRBOD5 and standard BOD5 values for wastewaters. Temperature: 30 C; pH: 6.5; PBS concentration: 5 mM. The samples were collected from: 14, wastewater treatment plants; 56, dairy/food manufactures; 7, rivers; 8, sh farms; 9, lakes; 10, sugar plants.

82

C. Liu et al. / Biosensors and Bioelectronics 34 (2012) 7782 Liu, C., Sun, T., Zhai, Y.M., Dong, S.J., 2009a. Talanta 78 (2), 613617. Liu, C., Sun, T., Xu, X.L., Dong, S.J., 2009b. Analytica Chimica Acta 641 (1-2), 5963. Liu, C.Y., Qu, J.Y., Jia, J.B., Qi, L., Dong, S.J., 2005. Chinese Journal of Analytical Chemistry 33 (5), 609613. Liu, C.Y., Ma, C., Yu, D.B., Jia, J.B., Liu, L., Zhang, B.L., Dong, S.J., 2011. Biosensors & Bioelectronics 26 (5), 20742079. Liu, J., Mattiasson, B., 2002. Water Research 36 (15), 37863802. Liu, J., Bjornsson, L., Mattiasson, B., 2000. Biosensors & Bioelectronics 14 (12), 883893. Liu, L., Shang, L., Liu, C., Liu, C.Y., Zhang, B.L., Dong, S.J., 2010a. Talanta 81 (45), 11701175. Liu, L., Shang, L., Guo, S.J., Li, D., Liu, C.Y., Qi, L., Dong, S.J., 2009c. Biosensors & Bioelectronics 25 (2), 523526. Morris, K., Catterall, K., Zhao, H., Pasco, N., John, R., 2001. Analytica Chimica Acta 442 (1), 129139. Morris, K., Zhao, H.J., John, R., 2005. Australian Journal of Chemistry 58 (4), 237245. Nakamura, H., Suzuki, K., Ishikuro, H., Kinoshita, S., Koizumi, R., Okuma, S., Gotoh, M., Karube, I., 2007a. Talanta 72 (1), 210216. Nakamura, H., Abe, Y., Koizumi, R., Suzuki, K., Mogi, Y., Hirayama, T., Karube, I., 2007b. Analytica Chimica Acta 602 (1), 94100. Organization for Economic Cooperation and Development, 1981. OECD Guideline for Testing of Chemicals, 209, pp. 110. Pang, H.L., Kwok, H.Y., Chan, P.H., Yeung, C.H., Lo, W.H., Wong, K.Y., 2007. Environmental Science & Technology 41 (11), 40384044. Preininger, C., Klimant, I., Wolfbeis, O.S., 1994. Analytical Chemistry 66 (11), 18411846. Sakaguchi, T., Morioka, Y., Yamasaki, M., Iwanaga, J., Beppu, K., Maeda, H., Morita, Y., Tamiya, E., 2007. Biosensors & Bioelectronics 22 (7), 13451350. Spicer, C.W., Joseph, D.W., Schumacher, P.M., 1985. Analytical Chemistry 57 (12), 23382341. Thavarungkul, P., Kumlanghan, A., Kanatharana, P., Asawatreratanakul, P., Mattiasson, B., 2008. Enzyme and Microbial Technology 42 (6), 483491. Wang, J.L., Zhang, Y.X., Wang, Y.Y., Xu, R.H., Sun, Z.H., Jie, Z., 2010. Biosensors & Bioelectronics 25 (7), 17051709. Wang, L., Guo, S.J., Huang, L.J., Dong, S.J., 2007. Electrochemistry Communications 9 (4), 827832. Zhao, H.J., Jiang, D.L., Zhang, S.Q., Catterall, K., John, R., 2004. Analytical Chemistry 76 (1), 155160. Zhang, S.Q., Jiang, D.L., Zhao, H.J., 2006. Environmental Science & Technology 40 (7), 23632368. Zhang, S.Q., Li, L.H., Zhao, H.J., 2009. Environmental Science & Technology 43 (20), 78107815. Zhang, S.Q., Zhao, H.J., Jiang, D.L., John, R., 2004. Analytica Chimica Acta 514 (1), 8997.

Acknowledgments We acknowledge nancial supports from the National Natural Science Foundation of China (20820102037, 20935003, 20805044), the National Basic Research Program of China (No. 2010CB933603), and Australia Research Council Discovery Projects. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.bios.2012.01.020. References
American Public Health Association, 1997. Standard Methods for the Examination of Water and Wastewater, 19th ed. American Public Health Association, Washington, DC. Bourgeois, W., Burgess, J.E., Stuetz, R.M., 2001. Journal of Chemical Technology and Biotechnology 76 (4), 337348. Catterall, K., Morris, K., Gladman, C., Zhao, H.J., Pasco, N., John, R., 2001. Talanta 55 (6), 11871194. Catterall, K., Zhao, H.J., Pasco, N., John, R., 2003. Analytical Chemistry 75 (11), 25842590. Chan, C.Y., Lehmann, M., Chan, K.M., Chan, P.Y., Chan, C.W., Gruendig, B., Kunze, G., Renneberg, R., 2000. Biosensors & Bioelectronics 15 (78), 343353. Chang, I.S., Jang, J.K., Gil, G.C., Kim, M., Kim, H.J., Cho, B.W., Kim, B.H., 2004. Biosensors & Bioelectronics 19 (6), 607613. Chee, G.J., Nomura, Y., Ikebukuro, K., Karube, I., 2005. Biosensors & Bioelectronics 21 (1), 6773. Chen, X., Chen, H.L., Ye, T.X., Qiu, B., Chen, G.N., 2008. Analytica Chimica Acta 612 (1), 7582. Dasgupta, P.K., Petersen, K., 1990. Analytical Chemistry 62 (4), 395402. Guwy, A.J., Farley, L.A., Cunnah, P., Hawkes, F.R., Hawkes, D.L., Chase, M., Buckland, H., 1999. Water Research 33 (14), 31423148. Jia, J.B., Dong, S.J., 2003. Chinese Journal of Analytical Chemistry 31 (6), 742748. Jia, J.B., Tang, M.Y., Chen, X., Li, Q., Dong, S.J., 2003. Biosensors & Bioelectronics 18 (8), 10231029. Kim, Y.C., Lee, K.H., Sasaki, S., Hashimoto, K., Ikebukuro, K., Karube, I., 2000. Analytical Chemistry 72 (14), 33793382. Kwok, N.Y., Dong, S.J., Lo, W.H., Wong, K.Y., 2005. Sensors and Actuators B: Chemical 110 (2), 289298. Lin, L., Xiao, L.L., Huang, S., Zhao, L., Cui, J.S., Wang, X.H., Chen, X., 2006. Biosensors & Bioelectronics 21 (9), 17031709.