Professional Documents
Culture Documents
Direct Determination of COD PDF
Direct Determination of COD PDF
pubs.acs.org/ac
ABSTRACT: Chemical oxygen demand (COD) was measured directly with a simple electrochemical method using a boron-
doped diamond (BDD) electrode. By applying a highly positive potential (+2.5 V vs Ag/AgCl) to an aqueous electrolyte
containing potassium hydrogen phthalate, glucose, and lactic acid or sodium dodecylbenzenesulfonate using a BDD electrode, an
anodic current corresponding to the electrolytic decomposition of these organic compounds was observed. No such current was
seen on glassy carbon or platinum electrodes due to a significant background current caused by the oxygen evolution reaction.
The electric charge for the anodic current observed at the BDD electrode was found to be consistent with the theoretical charge
required for the electrolytic decomposition of the organic compounds to CO2 and was used to calculate COD. This analysis was
performed by a simple I−t measurement at constant potential using a BDD electrode, and no calibration was needed. This new
simple indicator, “ECOD” (electrochemical oxygen demand), will be useful for continuous monitoring of industrial wastewater
with low protein concentrations and on-site instant analysis of natural water with a BDD electrode-based portable ECOD meter.
© 2014 American Chemical Society 8066 dx.doi.org/10.1021/ac500919k | Anal. Chem. 2014, 86, 8066−8072
Analytical Chemistry Article
Figure 1. (a) CV for 3 mM glucose in 0.1 M Na2SO4 using a BDD electrode, at a potential sweep rate of 100 mV s−1. (b−d) Anodic current
response to injection of glucose samples in stirred 0.1 M Na2SO4 with a BDD electrode. Initial solution volume was 2.7 mL, and a 0.3 mL sample
aliquot containing 150 nmol glucose was injected at each step indicated by an arrow. The constant potentials were (b) 2.4, (c) 2.5, and (d) 2.6 V vs
Ag/AgCl.
materials for electrolytic experiments, and its wide potential On the other hand, complete oxidation of organic compounds
window or large overpotential for the oxygen evolution reaction by reaction with oxygen can be expressed as follows:
of water decomposition is also favorable for efficient
decomposition of organic compounds.13−15 However, it is CaHbNcOd + yO2 → aCO2 + (b/2)H 2O + c NO2 (2)
difficult to make amperometric measurements in practical COD
From eqs 1 and 2, the relationship x = 4y (= 4a + b + 4c − 2d)
tests. First, the use of a reaction current as the signal for COD can be obtained. Thus, the total charge for the electro-oxidation
estimation requires a calibration curve. Furthermore, the can be used to calculate COD (denoted here as electrochemical
amperometric current does not always reflect the COD value. oxygen demand, ECOD). Practically, however, estimation of
Basically, only when the sample contains a single species can the charge is often difficult because the highly positive electrode
the COD value be estimated from the calibration curve. Lastly, potential causes water decomposition, resulting in a significant
calculating the number of electrons transferred during oxidation anodic current by the oxygen evolution reaction (OER). The
is not simple, especially for real sample solutions containing OER current at a highly positive potential often fluctuates
many chemical species. This is because the particulars of the significantly due to bubble formation on the electrode surface
oxidation reaction depend on the type of compounds present. making it difficult to separate the current due to the
A direct method for determining COD based on the decomposition of organic compounds from the total current.
complete oxidation of organic compounds in water with a In addition, application of a highly positive potential may cause
photocatalytic technique has been investigated.16,17 This should reaction/decomposition of the electrode material itself.
be useful because COD can be calculated directly, without BDD thin films have been known to be one of the most
calibration, from the charge required for photoelectrolytic stable electrode materials with respect to electrochemical
decomposition of organic compounds. For the photoelec- polarization at large overpotentials for water decomposition.
trochemical measurement, a UV light source is necessary for Thus, we expected the BDD electrode to be suitable for the
excitation and this may limit the miniaturization of the electrochemical estimation of COD. In addition, it has been
measurement system. reported that hydroxyl radicals (•OH) can be generated at the
In this work, we investigated the direct determination of BDD electrode during anodic water decomposition with high
COD from the anodic charge of electrolytic decomposition of current efficiency:18−22
organic compounds at highly positive potentials using a BDD H 2O → •OH + H+ + e− (3)
electrode. In principle, organic compounds in an aqueous
electrolyte can be oxidized completely to carbon dioxide by The hydroxyl radical is a strong oxidant capable of oxidizing
electrochemical oxidation at a highly positive potential: organic compounds to carbon dioxide:
In this study, we discovered that the electric charge for anodic AgCl for the electrolysis processes examined in this study, since
decomposition of certain organic compounds at +2.5 V versus the response current was large and stable at this potential.
Ag/AgCl can be evaluated by the use of a BDD electrode. The 2. ECOD Measurements on Organic Compounds at
amount of charge was found to correspond to the theoretical +2.5 V vs Ag/AgCl Using a BDD Electrode. Figure 2a
value, enabling direct determination of COD.
■ EXPERIMENTAL SECTION
BDD electrodes were prepared by microwave plasma-assisted
chemical vapor deposition (MPCVD) of a polycrystalline BDD
thin film on a conductive silicon wafer substrate using a
microwave power of 5000 W, a hydrogen gas flow rate of 540
sccm, a total pressure of 120 Torr, and a deposition time of 8 h.
The carbon/boron source was a mixture of acetone/methanol
(9:1, v/v) containing trimethoxyborane at a total atomic B/C
ratio of 0.5%. Detailed MPCVD conditions are described
elsewhere.23 For ECOD measurements, potassium hydrogen
phthalate (KC8H5O4, KHP), glucose (C6H12O6), lactic acid
(C3H6O3, LA), and sodium dodecylbenzenesulfonate (Na-
C18H29O3S, SDBS) were used as model organic pollutants. The
BDD electrode was placed at the bottom of an electrochemical
cell, contacting the electrolyte solution with an O-ring (11 mm
in diameter). A magnetic stirrer was put on the bottom of the
cell, parallel to the electrode, with a 4 mm thick spacer. A
platinum wire and an Ag/AgCl electrode in 3 M NaCl were
used as the counter and reference electrodes, respectively. A
Figure 2. (a) I−t curve for anodic decomposition of 200 nmol KHP in
potential of +2.5 V vs Ag/AgCl was applied to 4 mL of 0.1 M 0.1 M Na2SO4 at a BDD electrode. The constant potential was +2.5 V
Na2SO4 in the cell with stirring, and 0.3 mL of 0.1 M Na2SO4 vs Ag/AgCl. The KHP sample was injected at 1000 s. The shaded area
containing the organic compound was added to the solution was integrated to calculate ECOD. (b) Q−t curve created from the
after the background current became stable. The charge data of (a).
transferred during the anodic decomposition was calculated
from the area under the I−t curve after subtraction of the
shows the I−t curve for the electrochemical decomposition of
background. The area was calculated by integration of current
200 nmol of KHP (12.0 mg O2 L−1) in 4 mL of 0.1 M Na2SO4
with a sampling interval of 1 s. In this study, the background
at +2.5 V vs Ag/AgCl. After a stable background current was
current was determined as a linear line passing through the
obtained with the blank solution (0.1 M Na2SO4), a sample
current value just before addition of the sample and that after
solution containing KHP was added to the solution, resulting in
8000 s from the sample addition. A simple COD meter
an increase of the anodic current. The current then decreased
(photoLab 6100, WTW) and a TOC analyzer (TOC-L,
gradually to the initial background level within 10000 s. The
Shimadzu) were employed to estimate the COD and TOC,
background current was attributed to the anodic decomposition
respectively, of the model solutions for comparison. For a
of water and was basically constant over the course of each
standard addition study using a real sample, water was collected
electrolysis experiment. The increase in current upon addition
from the Tone canal, Nagareyama, Japan.
■
of the KHP sample solution was thus due to the anodic
decomposition of KHP. As the electrolysis continued, this
RESULTS AND DISCUSSION current decreased as the KHP concentration in the cell
1. Optimization of Electrode Potential. In general, BDD dropped. The electric charge transferred during KHP
electrodes exhibit large overpotentials for the anodic decom- decomposition, estimated by integrating the background-
position of organic compounds due to their poor surface subtracted anodic reaction current, was 580 mC (Figure 2b).
adsorption (or catalytic effect). Figure 1a shows a cyclic The overall electrochemical oxidation reaction of the hydrogen
voltammogram (CV) of 3 mM glucose in 0.1 M Na2SO4 at the phthalate ion is expressed as follows:
BDD electrode. From the CVs, it was determined that the C8H5O4 − + 12H 2O → 8CO2 + 29H+ + 30e− (5)
onset of the OER current was at around +1.8 V vs Ag/AgCl,
and no obvious anodic current for glucose oxidation was In accordance with Faraday’s law, using the Faraday constant
observed at less positive potentials. Next, we investigated the (96485 C mol−1), the total number of moles of electrons
operating potential for glucose oxidation at more positive required for KHP decomposition resulting from 580 mC of
potentials. At a constant electrode potential below +2.3 V vs charge passing through the cell was calculated to be 6.01 μmol.
Ag/AgCl with stirring, no obvious current response was The amount of KHP reacted was then estimated from eq 5 to
observed after addition of the glucose solution (150 nmol). be 200 nmol, identical to the actual value. This confirmed that
At potentials from +2.4 to 2.6 V vs Ag/AgCl, however, a clear the reaction current was based on the total decomposition of
current response was observed upon addition of the glucose KHP. The anodic charge Q [C] could be converted to ECOD
sample (Figure 1, panels b−d). At potentials higher than +2.6 V [mg O2 L−1] from the following relationship derived from eqs 1
vs Ag/AgCl, the current was not stable due to a significant and 2:
increase in the background (OER) current. We ultimately
decided on an optimal electrode potential of +2.5 V vs Ag/ ECOD[mg O2 L−1] = (8 × 106 /Fv)Q (6)
Figure 3. (a) Estimated ECOD for KHP, glucose (G), LA, and SDBS as a function of the number of moles of each compound. (b) Experimental
ECOD (ECODexp) vs theoretical ECOD (ECODtheor) for KHP, glucose (G), LA, SDBS, and an equimolar mixture of KHP and glucose. The solid
lines indicate theoretical values.
Figure 5. Anodic decomposition of 300 nmol KHP at +2.5 V vs Ag/AgCl at (a) GC and (b) Pt electrodes. The KHP sample was injected at 1000 s
(arrows). (c) CVs in 0.1 M Na2SO4 using BDD, GC, and Pt electrodes at a potential sweep rate of 100 mV s−1.
Figure 6. (a) COD values measured with a commercial COD meter and (b) TOC values measured with a commercial TOC analyzer, and the
corresponding values calculated from the anodic charge, along with theoretical values, for a KHP solution. Solid lines indicate the theoretical values.
completely to CO2 during anodic decomposition.28 Direct per mol of glucose. Thus, the number of electrons transferred,
anodic oxidation of 1 mol of glucose to carbon dioxide requires and hence the electric charge, can be directly used to calculate
24 mol of electrons: COD. In this regard, this method is fundamentally different
from the electrochemical estimation of COD based on the
xC6H12O6 + 6x H 2O → 6xCO2 + 24x H+ + 24x e− (7)
measurement of amperometric current, which requires an
On the other hand, in hydroxyl radical-mediated anodic experimental calibration curve.
oxidation (indirect oxidation), we have 4. Comparison with Results from a Commercial COD
Meter and TOC Analyzer. The ECOD value obtained with a
24(1 − x)H 2O → 24(1 − x)• OH + 24(1 − x)H+ BDD electrode was compared to COD data measured with a
+ 24(1 − x)e− (8) COD meter. Figure 6a shows the experimentally measured
COD versus the theoretical value for a KHP solution. The
(1 − x)C6H12O6 + 24(1 − x)• OH values obtained with the COD meter were somewhat smaller
than the theoretical values. The conventional COD test
→ 6(1 − x)CO2 + 18(1 − x)H 2O (9) includes multiple steps, namely, addition of reagent (potassium
Thus, the overall reaction from eqs 8 and 9 is dichromate), heating (for the oxidation reaction to occur), and
an absorbance measurement, thus requiring considerable time
(1 − x)C6H12O6 + 6(1 − x)H 2O and often resulting in data scattering. On the other hand, the
electrochemical method proposed in this study is more precise,
→ 6(1 − x)CO2 + 24(1 − x)H+ + 24(1 − x)e− can be performed more easily and is suitable for automation
(10) with a simple instrument. Figure 6b shows the experimental
The sum of eqs 7 and 10 gives TOC value obtained with a commercial TOC analyzer and by
calculation from the anodic charge. When the chemical
C6H12O6 + 6H 2O → 6CO2 + 24H+ + 24e− (11)
composition of the sample is known, TOC can be calculated
Even when direct and indirect oxidation occur in any from the charge transferred using eq 1. TOC measurements,
proportion x (0 ≤ x ≤ 1), 24 mol of electrons are transferred based on the complete combustion of organic compounds in
8070 dx.doi.org/10.1021/ac500919k | Anal. Chem. 2014, 86, 8066−8072
Analytical Chemistry Article
the sample solution, are known to be precise. The TOC values complete electrolytic decomposition. This measurement time
calculated from the charge, however, were found to be closer to could be reduced by using an electrochemical cell with a thin
the corresponding theoretical values than those obtained by a electrolyte layer in order to increase the electrode area relative
TOC analyzer. This indicated that the ECOD values found to the electrolyte solution volume, and this investigation will be
using a BDD electrode could be useful as a reliable indication of done in a future study. Since the ECOD measurement can be
organic water pollution, measurable with a simple instrument. performed fairly simply (I−t measurement at a constant
5. Interference and Standard Addition Studies. Effect potential) and no calibration is needed, this method should
of the presence of proteins in the sample solution to ECOD be useful for continuous monitoring of industrial wastewater
measurement was investigated as an interference study. When a with low protein concentrations and on-site instant analysis of
sample aliquot containing 200 nmol KHP and bovine serum natural water with a portable ECOD meter.
albumin (BSA, 1 wt % in the sample aliquot or 0.025 wt % in
the electrochemical cell) was injected in the stirred 0.1 M
Na2SO4, at +2.5 V vs Ag/AgCl, the current suddenly decreased
■ AUTHOR INFORMATION
Corresponding Authors
and minor current increase by electrolysis was observed (result *E-mail: t-kondo@rs.noda.tus.ac.jp.
not shown). On the other hand, in the case of lower BSA *E-mail: einaga@chem.keio.ac.jp.
concentrations (0.001 and 0.01 wt % in the sample aliquot), no
Notes
current decrease by the injection and current increase for the
The authors declare no competing financial interest.
■
electrolysis of KHP was observed. However, the anodic charge
was found to be smaller (7.8 and 5.4 mg O2 L−1 for 0.001 and
0.01 wt % BSA, respectively) than the theoretical one (12.0 mg ACKNOWLEDGMENTS
O2 L−1). These results should be because of adsorption of the This work was supported by DKK-TOA Corporation (Saitama,
protein to the electrode surface, which can disturb the Japan).
background current (oxygen evolution reaction current) and
hydroxyl radical generation, even though the BDD surface may
not adsorb proteins strongly.29,30
■ REFERENCES
(1) Korenaga, T.; Ikatsu, H. Analyst 1981, 106, 653−662.
In addition, we have demonstrated a standard addition study (2) Lee, K.-H.; Ishikawa, T.; McNiven, S. J.; Nomura, Y.; Hiratsuka,
using a canal water sample. 100, 200, and 300 nmol of KHP A.; Sasaki, S.; Arikawa, Y.; Karube, I. Anal. Chim. Acta 1999, 398, 161−
were added in the water sample, and the ECOD was estimated. 171.
When the blank canal water sample (without KHP) was (3) Westbroek, P.; Temmerman, E. Anal. Chim. Acta 2001, 437, 95−
injected to the electrolyte, no obvious increase of current was 105.
(4) Ai, S.; Gao, M.; Yang, Y.; Li, J.; Jin, L. Electroanalysis 2004, 16,
observed. However, in the case of KHP-containing canal water, 404−409.
the current increased as the sample was injected, and the (5) Li, J.; Li, L.; Zheng, L.; Xian, Y.; Jin, L. Meas. Sci. Technol. 2006,
ECOD calculated from the anodic charge was close to the 17, 1995.
theoretical one (Figure 7). Thus, this method should be useful (6) Silva, C. R.; Conceiçaõ , C. D. C.; Bonifácio, V. G.; Filho, O. F.;
for estimation of organic pollution of environmental waters Teixeira, M. F. S. J. Solid State Electrochem. 2009, 13, 665−669.
with low protein concentrations. (7) Yang, J.; Chen, J.; Zhou, Y.; Wu, K. Sens. Actuators, B 2011, 153,
78−82.
(8) Cheng, Q.; Wu, C.; Chen, J.; Zhou, Y.; Wu, K. J. Phys. Chem. C
2011, 115, 22845−22850.
(9) Li, J.; Li, L.; Zheng, L.; Xian, Y.; Ai, S.; Jin, L. Anal. Chim. Acta
2005, 548, 199−204.
(10) Zhou, Y.; Jing, T.; Hao, Q.; Zhou, Y.; Mei, S. Electrochim. Acta
2012, 74, 165−170.
(11) Yu, H.; Wang, H.; Quan, X.; Chen, S.; Zhang, Y. Electrochem.
Commun. 2007, 9, 2280−2285.
(12) Yu, H.; Ma, C.; Quan, X.; Chen, S.; Zhao, H. Environ. Sci.
Technol. 2009, 43, 1935−1939.
(13) Panizza, M.; Cerisola, G. Electrochim. Acta 2005, 51, 191−199.
(14) Martinez-Huitle, C. A.; Ferro, S. Chem. Soc. Rev. 2006, 35,
1324−1340.
(15) Fierro, S.; Abe, K.; Christos, C.; Einaga, Y. J. Electrochem. Soc.
Figure 7. Estimated ECOD for KHP-containing canal water samples 2011, 158, F183−F189.
as a function of the number of moles of KHP added. The solid line (16) Zhao, H.; Jiang, D.; Zhang, S.; Catterall, K.; John, R. Anal. Chem.
indicates theoretical values. 2003, 76, 155−160.
■
(17) Han, Y.; Qiu, J.; Miao, Y.; Han, J.; Zhang, S.; Zhang, H.; Zhao,
H. Anal. Methods 2011, 3, 2003−2009.
CONCLUSIONS (18) Michaud, P. A.; Panizza, M.; Ouattara, L.; Diaco, T.; Foti, G.;
By applying a high positive potential (+2.5 V vs Ag/AgCl) at a Comninellis, C. J. Appl. Electrochem. 2003, 33, 151−154.
BDD electrode, KHP, glucose, LA, and mixtures thereof, were (19) Marselli, B.; Garcia-Gomez, J.; Michaud, P.-A.; Rodrigo, M. A.;
Comninellis, C. J. Electrochem. Soc. 2003, 150, D79−D83.
completely decomposed in the aqueous electrolyte solution. (20) Komatsu, M.; Rao, T. N.; Fujishima, A. Chem. Lett. 2003, 32,
The electric charge required for the decomposition was found 396−397.
to be consistent with the charge required for electrochemical (21) Kapałka, A.; Fóti, G.; Comninellis, C. Electrochim. Acta 2009, 54,
oxidation of the organic compounds to CO2. In this study, we 2018−2023.
used an electrochemical cell with a solution volume of 4 mL for (22) Enache, T. A.; Chiorcea-Paquim, A.-M.; Fatibello-Filho, O.;
the experiment, requiring a long measurement time for Oliveira-Brett, A. M. Electrochem. Commun. 2009, 11, 1342−1345.