Article pubs.acs.org/ac
Electrochemical Nanocomposite-Derived Sensor for the Analysis of Chemical Oxygen Demand in Urban Wastewaters Manuel Gutiérrez-Capitán,† Antoni Baldi,† Raquel Gómez,‡ Virginia García,‡ Cecilia Jiménez-Jorquera,† and César Fernández-Sánchez*,† †
Instituto de Microelectrónica de Barcelona (IMB-CNM), CSIC, Campus de la UAB s/n, Bellaterra, Barcelona 08193, Spain Adasa Sistemas S.A., C/José Agustín Goytisolo 30-32, L’Hospitalet de Llobregat, Barcelona 08908, Spain
‡
S Supporting Information *
ABSTRACT: This work reports on the fabrication and comparative analytical assessment of electrochemical sensors applied to the rapid analysis of chemical oxygen demand (COD) in urban waste waters. These devices incorporate a carbon nanotube−polystyrene composite, containing different inorganic electrocatalysts, namely, Ni, NiCu alloy, CoO, and CuO/AgO nanoparticles. The sensor responses were initially evaluated using glucose as standard analyte and then by analyzing a set of real samples from urban wastewater treatment plants. The estimated COD values in the samples were compared with those provided by an accredited laboratory using the standard dichromate method. The sensor prepared with the CuO/AgO-based nanocomposite showed the best analytical performance. The recorded COD values of both the sensor and the standard method were overlapped, considering the 95% confidence intervals. In order to show the feasible application of this approach for the detection of COD online and in continuous mode, the CuO/AgO-based nanocomposite sensor was integrated in a compact flow system and applied to the detection of wastewater samples, showing again a good agreement with the values provided by the dichromate method.
C
these including Fourier transform infrared (FT-IR) spectroscopy combined with attenuated total reflection (ATR) technology,8 near-infrared (NIR) transmission together with UV absorbance,9 atomic absorption spectrometry (AAS),10 inductively coupled plasma optical emission spectrometry (ICP-OES),11 linear sweep polarography,12 or chemiluminescence detection systems.13 Although many of these works have implemented flow injection analysis (FIA) in order to automate the measurements, the use of benchtop laboratory instrumentation makes it difficult to conduct them on-site. An in situ online COD quantification system able to provide real-time information is very important to monitor the pollution of water and to make timely decisions. Ideally, such a decentralized system should be cheap, fully automated, simple, and reliable as well as robust and sensitive. Previously developed approaches that fulfill all these requirements are based on the use of electrochemical devices. In this context, applications using oxygen electrodes14 or voltammetric electronic tongues15 have been reported. In recent years, many efforts have been devoted to the fabrication of photoelectrochemical sensors based on the catalytic activity of TiO2.16,17 Great attention has also been paid to the development of novel materials with electrocatalytic
hemical oxygen demand (COD) is an estimation of the amount of organic compounds in water. It is expressed as the quantity of molecular oxygen (in mg of O2) necessary to decompose all the organic matter contained in 1 L of solution to carbon dioxide and water. As an indicator of contamination, it is a key parameter for the evaluation of water quality. Especially relevant is the determination of the COD value during the processing of urban and industrial waste waters. In fact, the wastewater treatment plants (WWTP) have a legal limit of COD in the effluents, set to 125 mg L−1 O2, or a minimum 75% reduction with relation to the organic load of the influent.1 The most common method for the measurement of COD is the dichromate titration,2 which is well established and standardized in many countries. However, this procedure requires a time-consuming reflux process (2−4 h) and the use of expensive (Ag),3 corrosive (H2SO4 and Cr2O72−), and highly toxic (Hg and Cr(VI)) reagents.4 In order to circumvent these drawbacks, some modifications have been introduced, such as the use of radiation-assisted methods for digesting the sample,5 the use of the nontoxic cerium(IV) instead of Cr(VI) as oxidizing agent,6 or the use of bismuth-based adsorbents instead of Hg as masking agent for the chloride interference.7 However, all these variations require the final back-titration, which confers a poor reproducibility to the analysis and, besides, it is difficult to automate. Therefore, other analytical methods have been reported for the determination of COD, © XXXX American Chemical Society
Received: September 4, 2014 Accepted: January 16, 2015
A
DOI: 10.1021/ac503329a Anal. Chem. XXXX, XXX, XXX−XXX
Article
Analytical Chemistry
oxidize 1 mol of glucose, the concentration of glucose can be directly related with the COD value measured in mg L−1 O2. Therefore, a stock solution of 25 g L−1 glucose was prepared. This solution was left for mutarotation for 24 h and then stored at 4 °C. A solution containing 0.1 M NaOH (Panreac, Barcelona, Spain) was used as background electrolyte solution for all the determinations. This alkaline medium favors the oxidation of organic molecules on the surface of metallic oxides.24 Fabrication of the Carbon Nanotube Composite Electrochemical Sensors. The different composites were deposited on thin film platinum electrodes (100 nm Pt/20 nm Ni/20 nm Ti, 4.39 mm2 geometric area), defined on 3 mm × 3.5 mm silicon chips fabricated using standard microfabrication processes. They were wire-bonded to a printed-circuit board stick and packaged using an epoxy resin.30 For the application in the flow system, the sensors were packaged so that a 2.4 × 2.6 mm2 open window around them was defined using Ebecryl photocurable polymer.31 Composites were prepared by dissolving the PS pellets in toluene. Then, carboxylated MWCNTs and the electrocatalysts were added to the solution and ultrasonicated using a bath sonicator for 3 h. Four different dispersions were prepared, one for each catalyst to be tested, that is Ni, NiCu alloy, CoO, and CuO/AgO. The mass percentages in the ternary mixture were 13.4% MWCNTs and 6% catalyst. The proportion of catalyst was optimized in preliminary studies by our group.32 In the case of the CuO/AgO, the percentage of each oxide was 5.9% CuO and 0.1% AgO, in order to attain the synergistic effect of both catalysts.26 Besides, a fifth dispersion was considered by adding just carboxylated MWCNTs without catalysts, to the PS solution, in order to study the possible electrocatalytic effect of the CNTs. In this case, the proportion of MWCNTs was equally 13.4% in the PS composite. Once dispersed in toluene, the mixture was further homogenized by gentle magnetic stirring overnight. Some 4 μL of the toluene dispersion was manually dropped on the surface of the thin film Pt electrode using a micropipette, left in air for solvent evaporation, and then cured for 30 min in an oven at 80 °C. The process was repeated seven times in order to obtain a homogeneous composite thick film over the electrode.33 Then, the sensor was immersed in a 0.1 M NaOH solution for 1 h in order to wet the nanocomposite and to stabilize its capacitive current. Devices and Equipment. For batch measurements, the electrochemical cell included one of the composite sensors as working electrode, a Pt commercial counter electrode (Radiometer Analytical, Lyon, France), and a Ag/AgCl/10% (w/v) KNO3 reference electrode (Metrohm 0726 100, Herisau, Switzerland). For in flow measurements, a gold thin-film twoelectrode device defined on a 3 × 3.5 cm2 silicon substrate and also fabricated by a standard microfabrication process was implemented.31 The device included a 2.77 mm2 counter electrode and a 1.62 mm2 pseudoreference electrode. Before its use, 5 μL of the Nafion commercial solution was cast on the gold electrodes. After solvent evaporation at room temperature, a Nafion film was generated over the two electrodes that acted as a protection barrier and prevented them from adsorption/ desorption and contamination processes of organic species presented in the water samples, while keeping the required electrical contact between the electrodes and the solution.34 A μ-Autolab potentiostat/galvanostat (Eco Chemie B.V., Uthecht, The Netherlands), controlled by a GPES 4.7 software package
properties for COD sensors with amperometric detection, which unlike TiO2, do not require any UV light source to activate them and thus act as electrocatalysts. Although a boron-doped diamond electrode has been reported for COD detection at +2.8 V (vs Ag/AgCl),18 the use of materials with catalytic properties can be more advantageous in terms of power consumption, sensor lifetime, selectivity, and reproducibility. Indeed, the direct oxidation of organic compounds using metal or carbon-based sensors is not possible given that the high potentials required are outside the potential window of the electrochemical devices developed so far, whose limit at positive potentials is defined by the oxidation of water. Up to now, many materials have been successfully used as electrocatalysts to construct amperometric COD sensors, including PbO2,19 Rh2O3,20 IrO2,21 CoO,22 Ni,23 or Cu.24 Different combinations of these and other components have been reported to show a synergistic effect25 such as CuO and AgO,26 NiCu alloy,27 or Sb-SnO2/PbO2.28 The efficient catalytic activity of these materials appears to be related to the formation and chemisorption of strong oxidant hydroxyl radicals (·OH) on the electrode surface.29 Nevertheless, the reaction mechanism, efficiency of electrochemical oxidation of organics, and resulting reaction products depend on the material and its structure. Therefore, its selection is very important for the construction of an electrochemical COD sensor. In the present work, some of the most recently reported electrocatalytic materials for COD determination have been tested for the fabrication of an electrochemical sensor to be applied in real influents of urban WWTP. Ni, NiCu alloy, CoO, and CuO/AgO in the form of nanoparticles were incorporated in a composite material based on carboxylated multiwalled carbon nanotubes (MWCNTs) as conductive material and polystyrene (PS) as matrix. The resulting nanocomposites were deposited on thin film Pt electrodes. In order to evaluate the sensor behavior, a calibration study was carried out using glucose standard solutions, followed by the COD analysis of a set of urban wastewater samples. The estimated COD values were compared with the values provided by an accredited laboratory using the dichromate standard method. The sensor approach showing the best analytical performance was further tested and eventually incorporated in a compact flow system in order to demonstrate the sensor feasible application for online monitoring of the organic load in wastewaters.
■
MATERIALS AND METHODS Reagents and Solutions. All reagents used were of high purity, analytical grade, or equivalent. All solutions were prepared with deionized water. For the fabrication of the composite-derived sensors, PS pellets (Sigma-Aldrich, Madrid, Spain), carboxylated MWCNTs with 10 nm diameter and
