Title page
Title: Biofouling and in situ electrochemical cleaning of a boron-doped diamond free chlorine sensor
Robert E. Wilsona, Ivan Stoianova & Danny O’Hareb*
Authors contact details:
aRobert Euan
aIvan Stoianov
b*Danny O’Hare
Corresponding author
aDepartment of Civil & Environmental Engineering
Imperial College London
Exhibition Road
LONDON
SW7 2AZ
UK
bDepartment of Bioengineering
Imperial College London
Prince Consort Road
LONDON
SW7 2BP
UK
Abstract:
Biofouling presents a significant obstacle to the long-term use of electrochemical sensors in complex media. Drinking water biofilms reduce performance of sensors by insulating electrode surfaces by inter alia inhibiting mass transport. Boron-doped diamond (BDD) electrodes arerelatively resistant to biofouling and inert at high potentials. These qualities can be exploited to create a drinking water quality sensor that resists biofouling to meet performance criteria for longer,and to enable electrochemical cleaning of the sensor surface in situusing high potentials without disconnecting or disassembling the sensor.
A purpose-built BDD wall-jet sensor was compared with a glassy carbon (GC) sensor in ability to determine free chlorine, detect biofilm and remove biofilm in situ. It was found that the BDD produced accurate, and reliable readings with a 4.86% standard error and a LOD of 0.18 ppm. The BDD could be electrochemically cleaned in situ whereas this was less successful with the GC electrode. The BDD electrode could also detect electroactive pyocyanin, secreted in the biofilm of the drinking water biofilm indicator organismP.aeruginosa, potentially enabling biofouling and non-biological fouling such as scaling to be distinguished. Observed changes in flow sensitivity and current-voltage curves that correspond to fouling provide multiple fouling detection methods, resulting in an accurate, sensitive, water quality sensor that can be cleaned without disassembly or replacement of parts and can identify when cleaning is required.
Keywords: water, chlorine, BDD, biofouling, sensor, electroanalysis
1-Introduction
Boron-doped diamond (BDD) has been proven to be a versatile electrochemical material, with applications in biosensing[1–4]; disinfection[5–8] and in electrochemical cleaning[9–12]. The most commonly reported advantages of BDD electrodes are having the widest solvent window of all materials, low capacitance, reduced fouling and resistance to high potentials, temperatures and extremes of pH[13].The innate resistance to fouling and high potentials of BDD electrodes can be exploited for use in complex media where fouling inhibits long-term use and in situations where replacement of antifouling consumables such as filters or mechanical wipers is not possible. A particular challenging medium for these reasons is drinking water distribution networks, a medium where on-line continuous and widely distributed monitoringis desired for safe and efficient operation.
Safe levels of residual disinfectant as free chlorine in drinking water distribution systems are between 0.2 and 2 ppm[14]. Underchlorination results in inadequate disinfection, causing pathogenic microorganisms to proliferate[15–17]. Continuous underchlorination can cause chlorine resistance, allowing pathogens to adapt[18–20]. Overchlorination of drinking water produces disinfection by-products such as chloroform, low levels of which are suspected to cause adverse reproductive and developmental effects in humans[21,22].The quantity of chlorine required to maintain a safe and effective residual is calculated from the chlorine demand at terminal points of water distribution systems. The most effective disinfectant form of chlorine that remains after the chlorine demand is met,is referred to as free chlorine which in the water industry is the sum of concentrations of hypochlorite, hypochlorous acid and dissolved chlorine, though the later species is relatively low abundance at the pH values commonly found in drinking water. Free chlorine is most commonly measured by the N,N diethyl-1,4 phenylenediamine sulfate (DPD) or sodium thiosulfate titration colourimetric tests – neither of these tests are on-line methods suited for continuous monitoring[23].
Amperometric chlorine sensors are also available, but are susceptible to biofouling[24].Drinking water organisms form biofilms to create a stable environment to reproduce, retain nutrients, resist disinfectants and to produce colonies[25–28],and can render a sensor inoperable in 10 days[29].Microorganisms are very difficult to remove from water completely and so methods of cleaning must be developed to restore sensor function; usually this involves electrochemical acid cleaning of the sensor or the replacement of consumables, such as filters. BDD electrodes are tolerant of high potentials and so an attractive approach is to use a high potential to generate reactive oxygen species to remove films from the sensor surface without having to disconnect the sensor for maintenance. This in situ electrochemical cleaning method has been studied previously for biofilm removal and for disinfection of bulk water[7,9,30]. It is also possible to detect the presence of biofilm on sensor surfaces by deterioration in performance, detection of characteristic electroactive compounds in biofilm[31,32], and changes in voltammetric curves.This communication describescombination these elements to create a continuous, free chlorine sensor for drinking water using BDD that is capable of detecting biofouling and can be reconditioned within situ electrochemical cleaning.
2-Materials and methods
The Unijet flow cell (BASI), with wall-jet geometry hydrodynamics, was chosen due to its simple, scalable design. 6 mm Glassy carbon (GC) working blocks (PEEK) were included with the flow cellhowever BDD blocks were made by drilling a 5.2 mm diameter hole through the centre in which to push-fit 5.2 mm disks of BDD(Element 6, electroanalytical grade[33]). The non-detecting surface was coated with silver epoxy and a silver wire was attached. The hole left in the block was filled with low-viscosity resin (Robnor).The existing silver wire quasi-reference electrode was unmodified. Its potential relative to a commercial AgAgCl KClaq, 3 M) (CH Instruments) was -170 mV. Custom-made 1 mm-thick PTFE gaskets were used to vary the separation between nozzle outlet and electrode surface.
All electrochemistry was carried out with a CHI1030a potentiostat (CH Instruments). Electrodes were prepared for use by polishing with alumina powder, sonication in a detergent solution (Decon), rinsing with ethanol solution and electrochemical cleaning at a constant potential of either +2.0 V or +5.0 V in 0.5 M sulfuric acid for 5 minutes[13]. Electrodes prepared in this way are referred to as “pristine”.
Free chlorine is a technical term used in the water industry to refer to the sum of concentrations of OCl-, HOCl and Cl2 (aq)[23]. Standard solutions were prepared fromanalytical grade sodium hypochlorite solution (Sigma) and deionised water (> 15 M cm, Purite). The concentrations of stock solutions were standardised by the approved sodium thiosulfate method used by UK drinking water utilities[23]. Solution gradients for calibration curves were prepared by mixing a 10 ppm and a 0 ppm solution of free chlorine by HP1050 HPLC pumps (Hewett-Packard) to enable automated generation of typically 900-point calibration working curves. A range of 0 – 3.0 ppm was chosen as the WHO recommended range is 0 – 2.0 ppm[34]. The free chlorine content of tap water was determined by DPD test (Palintest). Tap water samples were taken from the laboratory sink. Volume flow rate for all experiments was 0.722 ml/s unless otherwise stated. An optimisation process was previously carried out to determine 1.052 V and 1.237 V as the optimal potentials for detection of free chlorine with the BDD and GC electrodes respectively.
Limits of detection (LOD) were calculated as the value of the blank signal plus 3 standard deviations of the mean of the blank signal. Error is expressed as the relative standard error.
3-Results and Discussion
3.1-Calibration of sensors for free chlorine detection
Figure 1 shows calibration working curves for the GC and BDD electrodes. Detection limits for both materials are below the WHO minimum recommended value of 0.2 ppm: 0.18 ppm and 0.06 ppm for BDD and GC respectively. The deionised water solution was not buffered since samples were unbuffered tap water. The pH of samples of taken during the concentration gradient used for calibration resulted in a small increase in pH from 7.2 to 7.7 across a concentration range of 0 to 3.0 ppm. The GC electrode was more sensitive and with a lower limit of detection than the BDD, but the variation between individual calibration curves was greater – this is represented in the relative standard error of 12.43% and 4.86% for GC and BDD respectively.
3.2-Tap water fouling
The long-term performance of the electrodes was tested by continuously circulating tap water through the flow cells for 31 days. The same potential and flow rate used to calibrate the sensors was used in this experiment however consistently lower currents were recorded in the tap water samples for both GC and BDD.At a free chlorine concentration of 0.21 ppm, the resultant current was 15 µA and 24 µA. This could be due to fouling or interference from other compounds in the tap water sample.
3.31-Detection of fouling – pyocyanin detection
Pyocyanin is produced by Pseudomonas aeruginosa, a commonly detected organism in drinking water biofilms[32,35].P.aeruginosa and the electroactive compound pyocyaninthat it secretes was used as a biomarker for the presence of biofilm. Figure 2 shows that peaks were detected at -0.4 V and +0.1 V on the BDD electrode after fouling which were absent before the experiment began and disappeared after in situ cleaning in the flowing tap water sample following electrochemical cleaning at +5.0 V for 30 minutes. These peaks correspond to the reduction and oxidation of pyocyanin, which is present on the surface of the fouled electrode due to attached biofilm produced by P.aeruginosa.The background signal for the in situ-cleaned BDD is noticeably lower than the pristine electrode. In situcleaning conditions are however different and are likely to lead to different surface functionalisation. The GC signal has more noise and so the results are less clear but it is also important to note that +5.0V for extended periods would cause damage to the GC and so a cleaning potential of +2.0 V was used. This could account for a less effective in situ cleaning and this is confirmed with less satisfactory recovery of performance post-clean for the GC.
Pyocyanin detection potentially allows the BDD electrode to differentiate between biofilm and scale, as only bacterial biofilm willproduce pyocyanin.Webster et al[32] detected P. aeruginosa in tripticase soy broth after 45 hours with this technique which resulted in peaks at -0.15 V and 0.1 V with peak heights of 4.75 µA and 0.5 µA.
3.32-Detection of fouling – cyclic voltammetry
In Figure 3, the resultant currentsat high oxidation and reduction potentials were also considered as indicators of the presence of biofilm. The currents at +2.0 V and -2.0 V increased after fouling, and again after in situ cleaning.Again, the GC electrode was not cleaned as aggressively as it is less resistant to high potentials and recovery of clean i-V characteristics was less satisfactory. The BDD electrode also showed an increased current at +2.0 Vwith fouling and a return to a value similar to that of the pristine electrode after in situ cleaning. This would appear to show that in situ cleaning was as effective as electrochemical cleaning in acid.
3.33-Detection of fouling – flow sensitivity
Sensitivity to flow is related to the diffusion-limited current by the following expression[36]:
(Eqn. 1)
Where V is volume flow rate. Simplifying and taking logs of both sides gives:
(Eqn. 2)
Plotting log(ilim) against log(V) would result in a slope of 0.75. Slopes of 0.75 are not observed for either of the pristine electrodes (Figure 4) althoughfouling could have caused the value of these slopes to change by changing the mass transport boundary layer at the surface, changing the mass transport rate constant.
Values of 0.296 and 0.052 were recorded for the slopes of pristine GC and BDD. After one month of continuous circulation with tap water, the slope for GC had decreased to 0.13 and increased for the BDD electrode to 0.229. After in situ cleaning, the values recovered to 0.310 and 0.439. It appears that, for this experiment,in situ cleaning was successful for GC as the decrease in slope seen with the fouled electrode was reversed with cleaning. The effect on fouled BDD was less clear as the slope observed for the pristine electrode was low, increased with fouling and again increased with in situ cleaning.
Another possible relationship is the value of the offset in Figure 4 which decreases with fouling, but is restored towards the value of the pristine electrodes following in situ cleaning. This is consistent with the principal effect being on the effective diffusion coefficient which will be manifest in the term k in Equation 2. This holds for both electrodes and suggests a rapid means of detecting fouling in situ.
Figure 1: Calibration of sensors for the detection of free chlorine in solutions of sodium hypochlorite in deionised water at room temperature. Both electrodes were pristine prior to use. The concentration of chlorine in each preparation of the stock solution was verified by standard titration with sodium thiosulfate (relative standard error in chlorine concentration of 1.24%). The data shown is the mean of 7 curves for each material, A) GC B) BDD. 900 data points were recorded for each curve. Amperometric i-T parameters were: potential 1.052 V (BDD), 1.237 V (GC); sample rate, 1 point per second. Limits of detection were 0.18 ppm (BDD) and 0.06 ppm (GC). Data points are shown in black, 9% prediction bounds are shown as dotted lines and lines of best fit is shown in grey.
Figure 2: Pyocyanin detection of fouled electrodes, A) GC, B) BDD, in tap water by square wave voltammetry. Amplitude = 50 mV, frequency = 15 Hz. Solid line = pristine electrode, dashed line = fouled electrode, dotted line = in situ cleaned electrode. 2B shows two peaks at -0.4 and 0.1 V with heights of 10.5 µA and 6 µA µA.
Figure 3: Biofilm detection by cyclic voltammetry of fouled electrodes in tap water from changes to background current. A) GC, B) BDD. Scan rate = 50 mV. Solid line = pristine electrode, dashed line = fouled electrode, dotted line = in situ cleaned electrode.
Figure 4:Calculation of flow rate sensitivity to detect biofilm on biofouled electrodes in tap water, A) GC B) BDD. Flow rates were set by HPLC pump and held at specific values until a steady current was observed. Solid line, plus signs = pristine electrode; dashed line, crosses = fouled electrode; dotted line, circles = in situ cleaned electrode.
4 -Conclusions
BDD and GC can both be used to detect free chlorine in drinking water. Although BDD is less sensitive, the results are assessed as more reliable from calibration curves and after long-term use. Square wave voltammetry detection pyocyanin, voltammetric determination of solvent limits, and changes in the relationship between volume flow rate and diffusion-limited current couldall be used to detect biofilm on the surface of the BDD electrode.The fouled electrodes were fouled in tap water for one month. However as the free chlorine content remained above 0.2 ppm, this limited the rate of fouling which would be expected to be more severe in longer pipe runs where lower free chlorine residuals are expected due to chlorine decay.
We have demonstrated a multifactorial approach to the detection of biofilm formation and demonstrated that in situ cleaning in drinking water can restore function of BDD and, to a lesser extent, GC free chlorine sensors.
5-Acknowledgements
The authors would like to thank the Engineering and Physical Sciences Research Council (EPSRC), Cla-Val and Bristol Water for funding this research.
References
[1]E. Majid, K.B. Male, J.H.T. Luong, Boron doped diamond biosensor for detection of Escherichia coli, J. Agric. Food Chem. 56 (2008) 7691–7695. Doi 10.1021/Jf8013958.
[2]C.E. Nebel, B. Rezek, D. Shin, H. Uetsuka, N. Yang, Diamond for bio-sensor applications, J. Phys. D. Appl. Phys. 40 (2007) 6443–6466. doi:10.1088/0022-3727/40/20/S21.
[3]B.C. Lourenção, R.A. Medeiros, R.C. Rocha-Filho, L.H. Mazo, O. Fatibello-Filho, Simultaneous voltammetric determination of paracetamol and caffeine in pharmaceutical formulations using a boron-doped diamond electrode, Talanta. 78 (2009) 748–752. doi:10.1016/j.talanta.2008.12.040.
[4]N. Spãtaru, B. V. Sarada, E. Popa, D. a. Tryk, a. Fujishima, Voltammetric determination of L-cysteine at conductive diamond electrodes, Anal. Chem. 73 (2001) 514–519. doi:10.1021/ac000220v.
[5]M. Rajab, C. Heim, T. Letzel, J.E. Drewes, B. Helmreich, Electrochemical disinfection using boron-doped diamond electrode - The synergetic effects of in situ ozone and free chlorine generation., Chemosphere. 121 (2015) 47–53. doi:10.1016/j.chemosphere.2014.10.075.
[6]C. Heim, M. Ureña de Vivanco, M. Rajab, E. Müller, T. Letzel, B. Helmreich, Rapid inactivation of waterborne bacteria using boron-doped diamond electrodes, Int. J. Environ. Sci. Technol. 12 (2015) 3061–3070. doi:10.1007/s13762-014-0722-9.
[7]P.M. Natishan, W.E. O’Grady, F.J. Martin, P.L. Hagans, H. Martin, B.R. Stoner, Electrochemical Oxidation of Organic Compounds using Boron-Doped Diamond Electrodes, ECS Trans. 45 (2013) 19–30. doi:10.1149/04517.0019ecst.
[8]B. Marselli, J. Garcia-Gomez, P. -a. Michaud, M. a. Rodrigo, C. Comninellis, Electrogeneration of Hydroxyl Radicals on Boron-Doped Diamond Electrodes, J. Electrochem. Soc. 150 (2003) D79. doi:10.1149/1.1553790.
[9]J. Iniesta, P. a. Michaud, M. Panizza, G. Cerisola, a. Aldaz, C. Comninellis, Electrochemical oxidation of phenol at boron-doped diamond electrode, Electrochim. Acta. 46 (2001) 3573–3578. doi:10.1016/S0013-4686(01)00630-2.
[10]M. Panizza, P.A. Michaud, G. Cerisola, C. Comninellis, Anodic oxidation of 2-naphthol at boron-doped diamond electrodes, J. Electroanal. Chem. 507 (2001) 206–214. doi:10.1016/S0022-0728(01)00398-9.
[11]P.A. Michaud, Preparation of Peroxodisulfuric Acid Using Boron-Doped Diamond Thin Film Electrodes, Electrochem. Solid-State Lett. 3 (1999) 77. doi:10.1149/1.1390963.
[12]F. Montilla, P. A. Michaud, E. Morallon, J.L. Vazquez, C. Comninellis, Electrochemical oxidation of benzoic acid at boron-doped diamond electrodes, Electrochim. Acta. 47 (2002) 3509–3513. doi: 10.1016/S0013-4686(02)00318-3
[13]J. V Macpherson, A practical guide to using boron doped diamond in electrochemical research., Phys. Chem. Chem. Phys. 17 (2015) 2935–49. doi:10.1039/c4cp04022h.
[14]World Health Organization (WHO), Disinfectants, (2014).
[15]US EPA, Health risks from microbial growth and biofilms in drinking water distribution systems, Rep. Water Off. United State Environ. (2002). (accessed May 12, 2015).
[16]WHO, Microbial fact sheets, WHO Organ Guidelines for Drinking Water Quality (2004) 221–296.
[17]U. Szewzyk, R. Szewzyk, W. Manz, S. KH, Microbiological Safety of Drinking Water, Annu. Rev. Microbiol. 54 (2000) 81–127.
[18]D. Berry, C. Xi, L. Raskin, Microbial ecology of drinking water distribution systems., Curr. Opin. Biotechnol. 17 (2006) 297–302. doi:10.1016/j.copbio.2006.05.007.
[19]F.H. Yildiz, G.K. Schoolnik, Vibrio cholerae O1 El Tor: identification of a gene cluster required for the rugose colony type, exopolysaccharide production, chlorine resistance, and biofilm formation., Proc. Natl. Acad. Sci. U. S. A. 96 (1999) 4028–4033. doi:10.1073/pnas.96.7.4028.
[20]L.A.Carson, N.J. Petersen, M.S. Favero, S.M. Aguero, Growth characteristics of atypical mycobacteria in water and their comparative resistance to disinfectants, Appl. Environ. Microbiol. 36 (1978) 839–846.