Biotic and abiotic characterization of bioanodes formed on oxidized carbon electrodes as a basis to predict their performance Bibiana Cercado a , Luis Felipe Cházaro-Ruiz a , Vianey Ruiz b , Israel de Jesús López-Prieto b , Germán Buitrón b , Elías Razo-Flores a,n a Instituto Potosino de Investigación Científica y Tecnológica, A.C. División de Ciencias Ambientales, Camino a la Presa San José No. 2055, Lomas 4a Sección, 78216 San Luis Potosí, San Luis Potosí, México b Laboratorio de Investigación en Procesos Avanzados de Tratamiento de Aguas, Unidad Académica Juriquilla, Instituto de Ingeniería, Universidad Nacional Autónoma de México, Campus Juriquilla, UNAM. Blvd. Juriquilla No. 3001, 76230 Querétaro, México article info Article history: Received 26 March 2013 Received in revised form 14 June 2013 Accepted 24 June 2013 Available online 4 July 2013 Keywords: Bioanode Carbon based electrodes Oxidation treatments Electroactive biofilm Electrode characterization abstract Bioelectrochemical systems (BESs) are based on the catalytic activity of biofilm on electrodes, or the so- called bioelectrodes, to produce electricity and other valuable products. In order to increase bioanode performance, diverse electrode materials and modification methods have been implemented; however, the factors directly affecting performance are yet unclear. In this work carbon cloth electrodes were modified by thermal, chemical, and electrochemical oxidation to enhance oxygenated surface groups, to modify the electrode texture, and consequently the electron transfer rate and biofilm adhesion. The oxidized electrodes were physically, chemically, and electrochemically characterized, then bioanodes were formed at +0.1 V vs. Ag/AgCl using domestic wastewater amended with acetate. The bioanode performance was evaluated according to the current and charge generated. The efficacy of the treatments were in the order Thermal 4Electrochemical 4Untreated 4Chemical oxidation. The maximum current observed with untreated electrode was 0.152 70.026 mA (380 792 mA m 2 ), and it was increased by 78% and 28% with thermal and electrochemical oxidized electrodes, respectively. Moreover, the volatile solids correlated significantly with the maximum current obtained, and the electrode texture was revealed as a critical factor for increasing the bioanode performance. & 2013 Elsevier B.V. All rights reserved. 1. Introduction In the last decade interest in BESs has increased. BESs are based on the catalytic activity of biofilm adhered to electrodes to convert the chemical energy present in organic matter into electrical energy. The applications of BESs comprise microbial fuel cells (MFCs), microbial electrolysis cells (MECs), and more recently, MFC-based biosensors (Su et al., 2011). BESs are dependent on multiple factors, one of which is the biofilm formation on different electrode surfaces. Electrode materials and their modifications have been broadly studied (Zhou et al., 2011), as well as the electron transfer mechanisms between the microbial cells and the electrodes (Torres et al., 2010). Carbon materials are the most widely used electrodes due to their chemical stability, conductiv- ity, biocompatibility, and relatively low cost; they traditionally include carbon, graphite, activated carbon, glassy carbon, and nanotubes, in different structures such as plate, foil, rod, brush, paper, felt, cloth, mesh, or particles. To increase BESs performance, different modifications focus on changing the pore texture, increasing the surface area, and modifying the surface chemistry. Some modifications of carbon electrodes are summarized in Table 1. Among the methods that have been explored are covering with metals or polymers, heat treatments under specific atmo- spheres and electrochemical and chemical oxidation using acid soaks. Recently carbon electrodes have been modified with electro- chemically active species like RuO 2 and graphene (Liu et al., 2012; Contents lists available at SciVerse ScienceDirect journal homepage: www.elsevier.com/locate/bios Biosensors and Bioelectronics 0956-5663/$ - see front matter & 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.bios.2013.06.051 Abbreviations: ANCOVA, Analysis of covariance; ANOVA, Analysis of variance; BES, Bioelectrochemical system(s); C, Capacitance; %CE, Coulombic efficiency; CO, Chemically oxidized electrode; CODs, Soluble chemical oxygen demand; CPE, Constant phase element; CV, Cyclic voltammetry; %CV, Coefficient of variation; EAA, Electrochemically active area; EO, Electrochemically oxidized electrode; ΔEp, Potential peak separation; I, Current; Ip, Current peak; M, Diffusion element; MEC, Microbial electrolysis cell(s); MFC, Microbial fuel cell(s); OCP, Open circuit potential; PEIS, Potentiostatic electrochemical impedance spectroscopy; PZC, Point of zero charge; Q, Constant phase element; R, Resistance; R CT , Charge transfer resistance; Ro, Ohmic resistance; SSA, Specific surface area; TO, Thermal oxidized electrode; UE, Untreated electrode; VS, Volatile solids; W, Warburg's diffusion element n Corresponding author. Tel.:+52 444 8342026; fax:+52 444 8342010. E-mail addresses: [email protected], [email protected], [email protected] (B. Cercado), [email protected] (L.F. Cházaro-Ruiz), [email protected] (V. Ruiz), [email protected] (I.de J. López-Prieto), [email protected] (G. Buitrón), [email protected] (E. Razo-Flores). Biosensors and Bioelectronics 50 (2013) 373–381
Transcript
Biosensors and Bioelectronics 50 (2013) 373–381
Contents lists available at SciVerse ScienceDirect
Biotic and abiotic characterization of bioanodes formed on oxidizedcarbon electrodes as a basis to predict their performance
Bibiana Cercado a, Luis Felipe Cházaro-Ruiz a, Vianey Ruiz b, Israel de Jesús López-Prieto b,Germán Buitrón b, Elías Razo-Flores a,n
a Instituto Potosino de Investigación Científica y Tecnológica, A.C. División de Ciencias Ambientales, Camino a la Presa San José No. 2055, Lomas 4a Sección,78216 San Luis Potosí, San Luis Potosí, Méxicob Laboratorio de Investigación en Procesos Avanzados de Tratamiento de Aguas, Unidad Académica Juriquilla, Instituto de Ingeniería, Universidad NacionalAutónoma de México, Campus Juriquilla, UNAM. Blvd. Juriquilla No. 3001, 76230 Querétaro, México
a r t i c l e i n f o
Article history:Received 26 March 2013Received in revised form14 June 2013Accepted 24 June 2013Available online 4 July 2013
Keywords:BioanodeCarbon based electrodesOxidation treatmentsElectroactive biofilmElectrode characterization
63/$ - see front matter & 2013 Elsevier B.V. Ax.doi.org/10.1016/j.bios.2013.06.051
Bioelectrochemical systems (BESs) are based on the catalytic activity of biofilm on electrodes, or the so-called bioelectrodes, to produce electricity and other valuable products. In order to increase bioanodeperformance, diverse electrode materials and modification methods have been implemented; however,the factors directly affecting performance are yet unclear. In this work carbon cloth electrodes weremodified by thermal, chemical, and electrochemical oxidation to enhance oxygenated surface groups, tomodify the electrode texture, and consequently the electron transfer rate and biofilm adhesion. Theoxidized electrodes were physically, chemically, and electrochemically characterized, then bioanodeswere formed at +0.1 V vs. Ag/AgCl using domestic wastewater amended with acetate. The bioanodeperformance was evaluated according to the current and charge generated. The efficacy of the treatmentswere in the order Thermal4Electrochemical4Untreated4Chemical oxidation. The maximum currentobserved with untreated electrode was 0.15270.026 mA (380792 mA m�2), and it was increased by78% and 28% with thermal and electrochemical oxidized electrodes, respectively. Moreover, the volatilesolids correlated significantly with the maximum current obtained, and the electrode texture wasrevealed as a critical factor for increasing the bioanode performance.
& 2013 Elsevier B.V. All rights reserved.
1. Introduction
In the last decade interest in BESs has increased. BESs are basedon the catalytic activity of biofilm adhered to electrodes to convertthe chemical energy present in organic matter into electrical
ll rights reserved.
VA, Analysis of variance; BES,ulombic efficiency; CO,al oxygen demand; CPE,V, Coefficient of variation;ally oxidized electrode; ΔEp,; M, Diffusion element; MEC,(s); OCP, Open circuitnce spectroscopy; PZC, Pointnce; RCT, Charge transferarea; TO, Thermal oxidized; W, Warburg's diffusion
energy. The applications of BESs comprise microbial fuel cells(MFCs), microbial electrolysis cells (MECs), and more recently,MFC-based biosensors (Su et al., 2011). BESs are dependent onmultiple factors, one of which is the biofilm formation on differentelectrode surfaces. Electrode materials and their modificationshave been broadly studied (Zhou et al., 2011), as well as theelectron transfer mechanisms between the microbial cells and theelectrodes (Torres et al., 2010). Carbon materials are the mostwidely used electrodes due to their chemical stability, conductiv-ity, biocompatibility, and relatively low cost; they traditionallyinclude carbon, graphite, activated carbon, glassy carbon, andnanotubes, in different structures such as plate, foil, rod, brush,paper, felt, cloth, mesh, or particles. To increase BESs performance,different modifications focus on changing the pore texture,increasing the surface area, and modifying the surface chemistry.Some modifications of carbon electrodes are summarized inTable 1. Among the methods that have been explored are coveringwith metals or polymers, heat treatments under specific atmo-spheres and electrochemical and chemical oxidation usingacid soaks.
Recently carbon electrodes have been modified with electro-chemically active species like RuO2 and graphene (Liu et al., 2012;
Reduce internal resistance. Power (W m�2): Zhou et al.,(2012)792 (43%)
736 (33%)567 (3%)
Lagoon sediment Glucose One-chamberMFC
Activated carbonfiber
Oxidation in HNO3 andetylendiamine (EDA) treatments.
Reduce starting up time. Zhu et al.,(2011)Power (mW m�2):
EDA. 1641 (59%),HNO3, 2066 (100%).
a Three-electrode electrochemical cell under potentiostatic control.b Percent increase in parenthesis.
B. Cercado et al. / Biosensors and Bioelectronics 50 (2013) 373–381374
Lv et al., 2012). These modifications resulted in the creation ofnanostructures and activation centers on the carbon electrodes toincrease both the specific surface area and electron transfer kinetics.
The increases in BES performance reported are clearly depen-dent on their operational parameters, but the characteristics of theelectrodes that contribute to the reported performance are yetunclear. It has been signaled that enlarging the specific surface ofthe electrode and/or improving its electrocatalytic activity couldincrease bioanode performance (Lowy and Tender, 2008). How-ever, Heijne et al. (2008) did not find a direct correlation of powergeneration with the increase of electrode surface area. It has beenreported that chemical surface groups promote electron transfer,notably the surface oxygenated groups (Leon y Leon and Radovick,1994), although a reduction of O/C ratio after oxidation electrode
treatment and disagreement of N/C ratio with the power densitywas reported by Zhu et al. (2011). Moreover, the power increase byintroducing either nitrogen or oxygenated groups has resultedvery similar (Saito et al., 2011). In order to deepen the knowledgeof BESs and to study how bioanodes performance can beimproved, an exhaustive and systematic electrode and bioanodecharacterization should be performed, which includes physical–chemical, electrochemical, and microbiological investigations.
The aims of the present work were to select an oxidationmethod to modify carbon cloth electrodes in order to increasebioanode performance through a simple and effective treatment,and to identify the biotic and abiotic factors with the highestimpact on bioanode performance, which could be consequentlyuseful as predictive parameters.
B. Cercado et al. / Biosensors and Bioelectronics 50 (2013) 373–381 375
In order to achieve the objectives, untreated carbon clothelectrodes (UEs) were modified by thermal (TO), chemical (CO),and electrochemical oxidation (EO) to enhance surface oxygenatedgroups and modify the electrode texture. An enhancement of theelectron transfer and the bacterial adhesion was expected. Themodified electrodes were physical-chemically and electrochemi-cally characterized and therefore tested as support to formbioanodes using domestic wastewater. The characteristics of un-colonized electrodes were correlated with the bioanode perfor-mance, and the impact of biotic and abiotic factors was analyzed.
2. Materials and methods
2.1. Treatment and characterization of electrodes
Oxidation treatments were performed on graphite carbon cloth(1 m2 g�1, 0.56 mm thick, 99.9% carbon content, pH 6–10, Bruns-sen México). Chemical oxidation was achieved by heating thematerial at 80 1C for 1 h in a condenser containing nitric acidsolution (50:50 with deionized water); therefore the material waswashed exhaustively with deionized water until wash-water wasnear to pH 7. Electrochemical oxidation was performed by anodiz-ing the material for 1 h at +1.6 V (all the potentials are referred toAg/AgCl/KCl sat. system) in 120 mL of 0.5 M phosphate buffersolution composed of (g L�1): 4.33 Na2HPO4, 2.69 NaH2PO4, 2.85NaCl (Cercado-Quezada et al., 2011). Thermic oxidation wasachieved by heating the material at 636 1C in air atmosphere(Fig. 1S in Supplementary information); the temperature wasselected by previous thermogravimetric analysis (Thermo Cahn)as described by Wang et al. (2009).
N2 adsorption–desorption experiments were performed on themodified and unmodified electrodes in a Micrometrics ASA P 2000at �197.5 1C. The specific surface areas (SSA) were estimated usingthe Brunauer–Emmett–Teller (BET) equation. Prior to the mea-surements, the samples were degassed for 4 h at 100 1C. Fouriertransform infrared analyses (FTIR) were achieved in a ThermoScientific Nicolet 6700. The point of zero charge (PZC) of theelectrodes was determined following the method described byRangel-Mendez and Streat (2002).
Scanning electron microscopy was used to investigate thesurface topography of materials; the percent elemental composi-tion (C, N, O) was determined simultaneously in a FEI Quanta 200,X-Ray analyzer (EDAX) equipped microscope (Fig. 2S in Supple-mentary information).
2.2. Formation and biofilm performance
Domestic wastewater from an urban treatment plant wascollected and immediately used as inoculum. Biomass contentmeasured as volatile solids (VS) was 0.31370.116 g L�1, solublechemical oxygen demand (CODS) was 1331799 mg L�1, and pHwas 7.870.2. Sodium acetate (20 mM) was used as supplementarysubstrate. Substrate consumption was followed as CODs variationswith the reflux COD method described in APHA standard methods(Eaton et al., 2005). In brief, 2 mL sample were incubated during2 h at 550 1C in a reaction tube containing an oxidant solution(K2CrO7 in H2SO4). The chromic ion was quantified at 600 nmagainst a reaction blank in a spectrophotometer (Aquamate,Thermo Spectronic), and COD concentrations were estimated froma calibration curve. Acetate was determined by capillary electro-phoresis (Agilent Technologies G1602A) using the methoddescribed by Davila-Vazquez et al. (2008).
The biofilm was developed on carbon cloth electrodes(2�2 cm2) in duplicate under potentiostatic control at +0.1 V.The electrochemical cells were placed in a water bath at 25 1C. The
biofilm electroactivity was followed over 5–6 days by chronoam-perometry. A simple method for biofilm quantification on electro-des was developed. Bioanodes were dried at room temperatureand pulverized in a mixer mill (MM200, Retsch Inc.); a 0.02 gsample was placed in a tube containing 1 mL of 0.1 N NaOH andheated in a boiling water bath for 10 min. The mixture wascentrifuged at 9000g for 10 min, and the supernatant was usedfor protein quantification (Bradford, 1976). Colonized electrodeswere examined in a Helios NanoLab 600 ultrahigh resolutionmicroscope; the samples were dried in 30, 50, 70, 80, 90, and100% ethanol solutions and covered with a gold film beforeobservations.
2.3. Electrochemical analysis
A 150 mL three-electrode electrochemical cell was used asexperimental setup. Untreated and oxidized electrodes (2�2 cm2)were used as the working electrode; platinum mesh as the counter-electrode (3.5�2.5 cm2), and Ag/AgCl/KCl sat. as the referenceelectrode. Analyses were performed in phosphate buffer solution atpH 7. The electrochemical measurements were performed using apotentiostat (Bio-Logic SAS, EC-Lab ver. 10.23). The ohmic resistance(Ro), charge transfer resistance (RCT), and capacitance were obtainedby potentiostatic electrochemical impedance spectroscopy (PEIS).The electrochemically active area (EAA) was evaluated by cyclicvoltammetry (CV) of a typical reversible redox system composed of0.01 M K4Fe(CN)6 and 1 M KNO3 Argon saturated solution at roomtemperature. The voltammetric parameters, anodic and cathodiccurrent peaks (Ip), and potential peak separation (ΔEp¼Eap�Ecp)were determined from cycles starting at open circuit potential (OCP)towards anodic direction at 5 mV s�1 scan rate. The bioanodes wereinvestigated at starting and final test time. The OCP, CV, and PEISwere performed systematically in the indicated order. Voltammetriccycles were obtained in duplicate at 1 or 10 mVs�1 scan rate, towardsanodic direction, in the range from �1 to +1 V. PEIS was performedover a frequency range from 10 mHz to 100 kHz, or lightly modifiedwhen perturbations were observed, with a sinusoidal amplitude of10 mV. Data from PEIS were obtained as Bode and Nyquist plots andfitted to estimate the Ro, RCT, and capacitance parameters with twoapproaches: the first by adjusting the semi-circle data from Nyquistplot, the second by adjusting the PEIS data to equivalent circuitsR1+C2/R2+M3, R1+Q2/(R2+M2), R1+Q2/(R2+W) and R1+C2/R2+M3 forthe UE, TO, EO and CO electrodes, respectively.
Data were statistically analyzed using the IBMs SPSSs Ver. 20software. Pearson's correlation coefficients were obtained for thecurrent or experimental charge as dependent variable and thebioprocess operational conditions or electrochemical characteris-tics as independent variable. An ANOVA analysis, Tukey test, andANCOVA analysis were conducted to determine the significance ofthe treatments.
3. Results and discussion
3.1. Characterization of oxidized electrodes
Carbon cloth electrodes were oxidized by thermal, electroche-mical, and chemical methods, then characterized for physical–chemical and electrochemical properties. The electrodes character-ization is summarized in Table 2. Anodic oxidation was performedin NaCl electrolyte solution. The oxidation was first produced by theelectro-generated chlorine species; once the electrode was coveredwith oxygenated groups, carbon particles detached exposing newsurface oxygenated groups (Berenguer et al., 2011). Thermal oxida-tion removed the outer surface layer of the material, and then somesurface groups were eliminated while others were introduced
Table 2Physical–chemical and electrochemical characterization of oxidized carbon cloth electrodes.
Characteristic Oxidation treatment
Untreated Thermal Electrochemical Chemical
Surface area by BET (m2 g�1) 1.11 10.89 1.59 1.87Average pore size (Å) 83.36 33.57 86.11 66.08Total volume in pores 10�3 (cm3 g�1) 3.46 6.38 2.69 4.48Total area in pores (cm2 g�1) 0.56 3.54 0.99 0.87Total area pores/Total surface area 0.208 0.326 0.620 0.205C% 80.97 76.35 73.30 75.75N% 15.16 18.42 20.18 19.03O% 3.86 5.23 6.52 5.22O/C 0.048 0.069 0.089 0.069N/C 0.187 0.241 0.275 0.251Point of zero charge, PZC 8.8 10.9 2.6 11.0Open circuit potential, OCP (V) �0.01 0.07 �0.15 0.12Ohmic resistance, Roa (Ω) 12.0 16.8 12.6 9.2Rob 6.6 13.4 10.4 4.1Charge transfer resistance, RCTa (Ω) 1741 831 1279 2019RCT
a Estimated by semi-circle data adjusting from Nyquist plot.b Estimated by circuit model adjusting from PEIS data.c Determined in 0.01 M K4Fe(CN)6 solution.
B. Cercado et al. / Biosensors and Bioelectronics 50 (2013) 373–381376
during the cooling process in air atmosphere. The oxygen surfacegroups produced by air oxidation are non-homogenous and causean increase in electrode porosity (Polovina et al., 1997). Chemicaloxidation in HNO3 incorporates oxygenated surface groups such ascarboxylic, carbonyl, hydroxyl, and so on; this process has beendescribed in depth by Leon y Leon and Radovick (1994). In sum, itwas expected that oxidation treatments generate defects andoxygenated functional groups at different degrees, thus affectingthe carbon cloth properties.
During the characterization of electrodes it was observed thatthe O/C ratio in the material was increased, regardless of theoxidation method. The highest O/C and N/C ratios were observedin the EO electrode, while it was low and similar for TO and COelectrodes. The O/C values of oxidized electrodes were one order ofmagnitude lower than those reported by Régisser and Lavoie(1996) in electrochemically oxidized randomly oriented graphite,nevertheless the O/C ratio specifically in TO electrode was quitesimilar to that obtained by Wang et al. (2009) for carbon meshtreated similarly by heating. Thus, it is possible that TO treatmentis reproducible for similar carbon electrodes.
Nitrogen or oxygenated surface groups can contribute toelectron transfer and/or bacterial attachment as a function of theircharge at the pH of the medium. The PZC suggested that at the pHvalue of the buffer solution, the UE, TO, and CO electrodes werepositively charged, but EO electrode was negatively charged. Theeffect of the surface groups is synergic, the positive charges attractthe assumed negatively charged bacteria (Hori and Matsumoto,2010) and the negative charges promote the electron transferprocess, thus in both conditions the resulting effect is desirable.
Surface groups created by oxidation treatments were investi-gated by FTIR (Fig. 3S in Supplementary information). Very smallbands in the middle IR spectra were observed in the three oxidizedelectrodes but not in UE. Notably at wave numbers 2260, 2100, and2000 cm�1, and less intensely at 1500 and 1240 cm�1. The bandsaround at 2000 cm�1 matches a chloride-related compound,which was explained by the NaCl containing buffer solution usedduring the electrochemical characterization. The bands at2100 cm�1 can be related to C¼O tension, imine, and CN
isocyanide functional groups; those at 1240 cm�1 can be asso-ciated to amino, hydroxyl, C¼N and C¼O tension, aldehyde, andcarboxylic acid; finally the bands at 1500 cm�1 could be associatedto benzene, naphthalene, amino secondary, imino, and nitroconjugated compounds (Smith, 1995). Subtraction of the UEspectrum was performed to enhance the signal of the otherspectra; however the identification of chemical groups was incon-clusive. By matching the detected chemical groups with thosereported to occur on carbon electrodes, it was inferred thatcarboxylic and quinone groups were possibly formed on oxidizedelectrodes. The nitrogen-containing groups possibly correspondedto the precursor of carbon fibers, the polyacrylonitrile.
The electrode SSA and textural characterization has beentypically related to bioanode performance (Liu et al., 2010; Tsaiet al., 2009). High available surface area logically increases thedirect contact of the bacteria with the electrode; porosity texturecould have a positive effect for the occurrence of mediated redoxreactions, and to enhance exopolymeric substance anchoring.
The TO had the highest SSA, up to 10 times higher than theother electrodes as it was evidenced in the adsorption isotherm(Fig. 4S in Supplementary information). During TO treatment,oxygen in air reacts with edge carbon atoms to produce –CO and–CO2 evolving groups, which ultimately led to gasification of theelectrode and creates pores. The process proceeds until graphitesheets are completely destroyed forming crevasses and additionalsurface area to form new pores. The EO consisted of the chemicalseparation of part of the oxide layer from the carbon surface sopatches could be observed (Fig. 5S in Supplementary information).EO and CO were less aggressive oxidative treatments; therefore theircorresponding textural characteristics were closer to those of UE.
Because of its higher SSA, high performance for the TOelectrode was expected. Textural analysis showed the presenceof mesopores in both UE and oxidized electrodes. The lowestaverage pore size was quantified in TO. Besides the total volume inpores being the highest for the same electrode, this texture mighttrap biofilm components. The contribution of total area in pores tototal surface area was higher for EO; therefore, a high performancewas expected for EO electrode as well.
B. Cercado et al. / Biosensors and Bioelectronics 50 (2013) 373–381 377
SEM micrographs confirmed the previously obtained porositydata. Dust particles were observed on UE, which were removed byTO and EO treatments while it did not occur in the CO treatment.The electrode surfaces were finely wrinkled with some smoothpatches on EO, rough and flaky on TO, and smooth on both UE andCO electrodes (Fig. 5S in Supplementary information). In summary,SSA, texture porosity, and O/C ratio signaled the TO and EO as themost suitable supports for bioanode formation.
The electrochemical characterization of the electrodes indi-cated that the OCP varied from �0.15 to 0.12 V. It can be assumedthat low OCP values enhance bioanode performance, as OCPbecomes negative naturally in MFCs. In the present study theOCP was not used as a predictive parameter because the bioanodeperformance was evaluated at fixed potential. The Ro was in thesame order of magnitude in all tests as expected because anidentical electrochemical cell configuration was used. The slightvariations observed were possibly due to the variations in thewastewater composition. Two critical parameters involved inelectrode performance are RCT and capacitance; the former isindicative of the capability of the material to transfer electrons,whereas the latter is associated with the charge distributed at thesolid–liquid interface and indicates the capacity of the material tokeep charge (Bard and Faulkner, 2001). A low RCT value facilitateselectron transfer and thus an enhanced bioanode performance.The ultimate application of the electrode, either as a conductor ora capacitor, indicates the preferable capacitance value. In thisstudy the electrode was considered as capacitor, because MFCsystems can operate in a charge–discharge sequence, thus a highcapacitance signaled an improved electrode. The capacitances ofoxidized electrodes estimated by adjusting the semi-circle data ofNyquist plots were in the order TO4EO4CO, which suggests ahigh future performance with TO and a quite low performancewith the CO electrode. The typical double layer capacitance forcarbon material is around 20 μF cm�2 (Régisser and Lavoie, 1996).This value was modified with the oxidation treatments; in mostcases this value was maintained at the same magnitude except forthe TO electrode, which had the highest capacitance value.
On the other hand, the capacitance values obtained by adjustingthe CPE model circuits did not necessarily correlate with the physicalprocess because the α parameter was lower than 1; thus, CPE was anadjusting parameter limited and not indicative of the double layerelectrode capacitance (Dominguez-Benetton et al., 2012).
The RCT obtained from the Nyquist plots suggested that the TOelectrode could generate the highest current because of its low RCTvalue. The data for this parameter obtained from the circuitmodels were almost similar (CV¼16%), with a lower value forthe CO electrode. As mentioned above, the RCT values from the CPEcircuit model must be taken with caution for TO and EO electrodes.According to the Nyquist plot data, the TO could be an adequateelectrode to form electroactive bioanode.
The SSA characterization was complemented with the mea-surement of the EAA, which was estimated considering thecomplexity of non-plain electrodes as the increase of Ip, as wellas the estimation of the electron transfer rate which is inverselyproportional to ΔEp (Fig. 1A and B; Table 2). The three oxidationtreatments were effective at increasing EAA and electron transferrate; however, the best conditions to increase the bioelectrochem-ical process were observed with the EO electrode.
The formation of surface quinone/hydroquinone groups wasinvestigated by CV. Peak pairs centered around 0.0, �0.07, and�0.125 V were detected for the TO, EO, and CO, respectively, butwere not observed in UE; additionally, increasing capacitivecurrents were evidenced (Fig. 1C–F). The redox peaks attributedto surface quinone moieties superimposed onto the capacitiveenvelope correspond to those reported by Régisser and Lavoie(1996). The authors showed an increase to the surface area of the
electrode, as well quinone groups detected by XPS and CV onelectrochemically treated graphite electrodes.
3.2. Bioanodes performance
Electroactive biofilm was developed on carbon cloth at E¼+0.1 V in domestic wastewater amended with 20 mM sodiumacetate. The biofilm formation and electrochemical activity wasfollowed for about 100 h. The first current cycle was preferentiallyanalyzed in order to avoid mass transfer limitations in a thickbiofilm, perturbations due to slugging, or complex evolution ofmicrobial community due to a long-term operation. Bioanodeperformance data are summarized on Table 3.
The current production performance was in the orderTO4EO4UE4CO and was similar for the charge generated. Theexperimental charge was calculated from start-up time to peakcurrent in order to exclude final perturbations and an inaccuratefinal time. Maximum current and the total of coulombs producedin the test period did not maintain the same proportion for allbioanodes (Fig. 6S in Supplementary information). This observa-tion suggested that microbial populations on the electrodes haddifferent kinetics.
The chronoamperograms showed a current peak around 60 h(Fig. 2E), followed by a decrease in current due to acetatedepletion, as verified by capillary electrophoresis. Differences inthe current start-up time were observed over the first 10 h. Thestart-up time was slightly lower for CO than the other oxidizedelectrodes. The capacitance estimated by adjusting the Nyquistplot data was the lowest for the CO electrode, possibly causing afaster charge release and faster current registration.
The differences in current production and experimental chargebetween the bioanodes were supported statistically with a var-iance analysis (significance 0.003), the post-hoc Tukey test indi-cated that the UE, TO, and CO effects on current production weresignificantly different, but the effect due to EO was not differentfrom UE nor to TO treatments.
The variations in current profiles were investigated consideringthe microbial cells as catalysts on the electrode. The quantity and/ortype of adhered cells were explored by SEM and the adheredbiofilm was quantified as protein. The micrographs showed a largecoverage of carbon fibers for almost all the electrodes, except forthe CO. Bacterial shapes were cocci, and short and long bacilli.The electrodes predominantly colonized with cocci, generatedlower current than those colonized with the diverse cell shapes(Fig. 2A–D). The geometry of cells could also have a subtle effect oncurrent production; assuming that electron transfer is predomi-nantly achieved by direct contact, spherical cells are less efficientthan rod-shape cells. The most frequently reported electroactivemicroorganism, Geobacter sp. and Shewanella sp., are rod shapedwith a length of 1–4 μm and diameter of 0.4–0.7 μm. They aremotile by means of a single or multiple flagella (Caccavo et al., 1994;Venkateswaran et al., 1999); nevertheless, new electroactive speciesare being reported which do not match with these typical celldescriptions (Jain et al., 2012). Future studies will be achieved toinvestigate the microbial community on the oxidized electrodes.
Based on these observations, it was thought that the differentcurrent profiles were due to the electron transfer kinetics linked tothe type of microorganisms, additionally to the biofilm densities.Differences in microbial communities caused by the electrodetreatment are also reported by Liu et al. (2007) and Zhu et al.(2011). These authors suggested that different surface groupsmight have selectivity to different bacteria. In the present workon the basis of the biofilm density differences observed, it washypothesized that the composition and charge of exopolymericsubstances excreted by the bacteria had a strong influence on
Fig. 1. Cyclic voltammetry of carbon cloth electrodes in potassium ferricyanide. (A) Untreated and thermal oxidized electrodes, (B) untreated, chemically, andelectrochemically oxidized electrodes. Cyclic voltammetry of electrodes in phosphate buffer solution, (C) untreated electrode, (D) electrochemically oxidized, (E) thermaloxidized, and (F) chemically oxidized electrode.
Table 3Performance of bioanodes formed on carbon cloth in domestic wastewater amended with 20 mM sodium acetate. Poised potential E¼+0.1 V, T¼25 1C.
B. Cercado et al. / Biosensors and Bioelectronics 50 (2013) 373–381378
biofilm adhesion and thus on electron transfer phenomena asreported by Rollefson et al. (2011).
Electroactive biofilm is frequently formed at fixed polarizationpotential, though the effect of the anode potential on the biofilm
communities that can be developed remains unclear (Wagneret al., 2010). In the present research a sole potential was applied,although the biofilm community and structure followed differentoxidizing treatments of the same carbon-based electrode. Thus,
Fig. 2. SEM observations of biofilm on (A) Electrochemical oxidized electrode, (B) thermal oxidized, (C) untreated electrode, (D) chemically oxidized electrode, and(E) chronoamperograms of bioanodes formed in domestic wastewater at E¼+0.1 V and T¼25 1C. The geometric figures represent the shape and size of cells observed inthermal and electrochemically oxidized electrodes (up) and untreated and chemically oxidized electrodes (down).
B. Cercado et al. / Biosensors and Bioelectronics 50 (2013) 373–381 379
the physical–chemical and electrochemical characteristics of elec-trode materials should be considered systematically when elec-troactive biofilm is under investigation.
In recent years, attempts to introduce electroanalytical techni-ques to bioanode analysis have been reinforced. Reports on purecultures and synthetic media have revealed their applicability(Marsili et al., 2008; Ouitrakul et al., 2007); however, when thecomplexity of the system is increased by using consortia asinoculum, the bioanode analysis must be limited to the earlydeveloping biofilm stage (Ramasamy et al., 2008), otherwise theinterpretation of results becomes difficult as it has been reportedin a recent review (Dominguez-Benetton et al., 2012).
In this work an electrochemical characterization of the bioa-nodes was performed at final test time (3–5 days). The modelcircuit adjustment of PEIS data showed that the RCT of the UE, TO,and CO decreased in the low-frequency region of the Bode plot,which was related to the development of an electroactive biofilm.The bioanode on the CO electrode presented the highest impe-dance value and a maximum at low frequencies around �901which corresponded to its low current production. The TO and EObioanodes with much better performance showed the lowervalues of impedance at low frequencies and the phase anglereached values close to 01 as well (Fig. 7S in Supplementaryinformation).
Based on these results, it can be concluded that the oxidationtreatments had a significant effect on the carbon electrode proper-ties and thus on bioanode performance. The TO and EO providedincreases of 78% (680721 mA m�2 electrode geometric area) and28% (4887mA m�2) in current compared to UE. By comparingthese results with similar electrochemical systems, the currentobserved is still lower than those obtained with graphite–carbon
composite (Patil et al., 2013) and with a double pretreatment in afelt carbon electrode (Cercado-Quezada et al., 2011). On the otherhand, the negative effect of CO treatment on the bioanodeperformance for the CO bioanode, was attributed to scarcecolonization and high impedance of the CO electrode. Never-theless, chemical oxidation had led to positive results whenactivated carbon was tested (Zhu et al., 2011); very likely the highsurface area associated with activated carbons contributed tothese results.
3.3. Factors correlating with the bioanode performance
One of the aims of this research was to determine the criticalfactors to bioanode performance. Correlation analysis performedbetween the current and bioprocess parameters showed that thepH of the medium and the initial CODs had no influence oncurrent because they were almost constant (CV 3.8% and 7.4% forpH and CODs, respectively). The %COD removal was slightly higherin the electrochemical cells with oxidized electrodes than with theUE; however, the coulombic efficiency (%CE) was quite low for allthe test because mainly the planktonic microorganism consumedthe COD rather than the biofilm on the relatively small anode, thusthe %COD removal was not related to current generation, but the %CE followed the pattern TO4EO4UE4CO because %CE is directlyproportional to experimental coulombs produced.
Given that it is broadly accepted that biofilm developed onelectrodes is primarily responsible for current production, it couldbe convenient to report BES performance on biofilm biomass basis.In this work biofilm protein and VS were quantified for each test.The protein did not correlate significantly to current production; itis possible that the quantification method was not sensitive
B. Cercado et al. / Biosensors and Bioelectronics 50 (2013) 373–381380
enough. Conversely, VS was shown to influence current generationwhich was evidenced in the dispersion plot (Fig. 3A). The Pearson'scorrelation analysis indicated a coefficient of 0.768, significant at0.05 level, thus an ANCOVA test was performed with the VS asco-variable, which confirmed significant differences betweenoxidation treatments (significance 0.031). Once the significanceof the I–VS correlation was demonstrated, the yield of the currentwith respect to the volatile solids YI/VS was determined for eachelectrode, obtaining 0.84, 0.57, 0.46, and 0.14 mA (g VS L�1)�1 forEO, TO, UE, and CO, respectively. Expressing the bioanode perfor-mance in this way, the best performance corresponded to theEO bioanode.
Other parameters to be related with the bioanode performancewere the electrochemical characteristics of electrodes. The lowfrequency region of the Bode plot showed that the RCT of the TOand EO electrodes were similar and lower than those of the UE andCO electrodes (Figs. 3B and 7S in Supplementary information),indicating that the rate of electron transfer on the first twoelectrodes would be faster and thus, allows a better performancein the presence of bacteria. The electrode characteristics that canbe related to electron transfer process, the O/C ratio, EAA, andelectron transfer rate, were higher for the EO electrode, but thecorresponding bioanode was second in performance. Analyzingthe physical characteristics, it was noted that the SSA was thehighest for the TO but their EAA was rather low, even though thiselectrode produced the highest current (Fig. 3C). These findingssuggest that the surface area was a more important factor than thesurface oxygenated groups, EAA and electron transfer rate forcurrent production. Then, if the surface area is sufficiently high,this condition should produce a high bioanode performance, but ithas not always been clear as reported by Heijne et al. (2008) andLiu et al. (2010). An electrode with high EAA seems to be usefulonly if it is completely colonized. Although the UE had the lowestSSA, EAA, and electron transfer rate, this electrode was abundantly
Fig. 3. Biotic and abiotic factors correlating with current generation from bioanodes oninoculum (biological tests performed in duplicate), (B) experimental and fitted impedancand (D) comparison of electrode texture between UE (left) and TO electrode (right) at 6
covered and it was third in performance; this suggests that surfacecoverage is also a critical factor for bioanode performance. Theconditioning film, i.e. the ubiquitous particles on surfaces thatpromote the biofilm formation (Hori and Matsumoto, 2010), waspresent in the UE that had no modification; therefore UE allowed amore dense and homogenous biofilm development. The relativelylow current observed in the UE may have been caused bydiffusional limitations and/or by the adhesion of bacteria withrather low electroactivity. This hypothesis was supported by theobservation of scarce colonization of CO electrode with similarcoccus-shaped bacteria; thus, this bacteria type had fairly lowelectroactivity.
It is broadly reported that biofilm formation is initiated bybacterial adhesion. The texture of supports has been shown as acritical factor in this process (Habouzit et al., 2011). In the presentwork it was observed that the ratio of total pore area to surfacearea, and the texture influenced the biofilm structure. Bacilli andcocci colonized the TO and EO electrodes, while only coccicolonized the UE and CO electrodes. Moreover, bacilli seemed toproduce more exopolymeric substances, since aggregates (Fig. 2Aand B) and union structures between carbon fibers were observed(Fig. 8S in Supplementary information). The EO electrode had thehighest pore size, and the TO had the highest total volume inpores, thus the texture created with the TO allowed the anchorageof exopolymeric substances (Fig. 3D), additionally to pili or flagella,as has also been suggested by Lowy and Tender (2008) foranodized carbon electrodes.
The effort to correlate biotic and abiotic parameters involved inbioanode formation with their performance revealed that thesystem biofilm-material support is complex, noticeably becauseof biofilm heterogeneity. The attempt to identify some factors withhigh weight seems to be unrealistic because they are interconnectedto lead an observable bioanode current; nevertheless, in the presentwork the biomass in the inoculum source, the electrode charge
oxidized carbon cloth electrodes. (A) Volatile solid content in wastewater used ase spectra for the electrodes in buffer solution, (C) specific surface area of electrodes,5,000X augmentation.
B. Cercado et al. / Biosensors and Bioelectronics 50 (2013) 373–381 381
transfer resistance, and the surface texture seemed to be directlyrelated with current production.
4. Conclusions
A critical condition to improve bioanode performance is biofilmformation on the electrode and the electron transfer at the inter-face. These phenomena can be controlled by the modification ofelectrode surfaces. In this work three oxidation methods oncarbon cloth were compared, in which thermal oxidation showedto be the most effective, increasing current production by 78%;however, all the treatments affected the electrode characteristicsand thus the electroactive biofilm developed. The analysis of bioticand abiotic factors involved in bioanode formation revealed the VSas a factor correlating directly to current production; moreover,the electrode RCT, SSA and textural porosity had a high impact onbiofilm structure. The bioanode performance is the result of manysynergic factors; however the previous parameters could be usefulto preselect electrode materials for bioanode formation.
Acknowledgments
We thank the technical assistance of Dulce Partida, GuillermoVidriales, Juan Pablo Rodas and Gladys Labrada from IPICYT andJaime Pérez from II-UNAM. We also thank J.M. Olvera fromPROAGUA-Potosí for providing the wastewater samples and J.Eckerly Goss from IPICYT for proofreading the manuscript. Thisresearch was financially supported by SEP-CONACYT project132483 and DGAPA-UNAM through Project PAPIIT IN104710.
Appendix A. Supporting information
Supplementary data associated with this article can be found inthe online version at http://dx.doi.org/10.1016/j.bios.2013.06.051.
References
Bard, A.J., Faulkner, L.R., 2001. Electrochemical methods. Fundamentals andapplication, second ed. John Wiley & Sons, New Jersey.
Berenguer, R., Marco-Lozar, J.P., Quijada, C., Cazorla-Amorós, D., Morallón, E., 2011.Carbon 50 (3), 1123–1134.
Bradford, M., 1976. Analytical Biochemistry 72, 248–254.Caccavo, F., Lonergan, D.J., Lovley, D.R., Davis, M., Stolz, J.F., McInerney, M.J., 1994.
Applied and Environmental Microbiology 60 (10), 3752–3759.Cercado-Quezada, B., Délia, M.L., Bergel, A., 2011. Electrochemistry Communication
13, 440–443.Cheng, S., Logan, B.E., 2007. Electrochemistry Communication 9, 492–496.Davila-Vazquez, G., Alatriste-Mondragón, F., de León-Rodríguez, A., Razo-Flores, E.,
2008. International Journal of Hydrogen Energy 33, 4989–4997.
Dominguez-Benetton, X., Sevda, S., Vanbroekhoven, K., Pant, D., 2012. ChemicalSociety Reviews 41 (21), 7228–7246.
Eaton, A.D., Clesceeri, L.S., Rice, E.W., Greenberg, A.E., 2005. Standard methods forthe examination of water & wastewater, 21st ed. APHA, Washington DC.
Bioengineering 101 (1), 101–108.Rangel-Mendez, J.R., Streat, M., 2002. Water Research 36, 1244–1252.Régisser, F., Lavoie, M., 1996. Journal of Electroanalytical Chemistry 415, 47–54.Rollefson, J.B., Stephen, C.S., Tien, M., Bond, D.R., 2011. Journal of Bacteriology 193
