Quasi-Real Time Quantification of Uric Acid in Urine Using BoronDoped Diamond Microelectrode with in Situ CleaningRaphael Kiran,*,† Emmanuel Scorsone,† Pascal Mailley,‡ and Philippe Bergonzo†

†CEA-LIST, Diamond Sensors Laboratory, 91191 Gif-sur-Yvette, France‡CEA-INES, Laboratoire de Stockage de l’Electricite, 73377, Le Bourget du Lac, France

*S Supporting Information

ABSTRACT: We report herein an innovative electrochemical (EC)technique based on boron doped diamond (BDD) microelectrodeswhich enable the fast determination of uric acid (UA) concentrationsin urine. On the basis of fast cyclic voltammetry (CV), the techniquewas assessed in human urine samples and compared successfully usingroutine spectrophotometric diagnosis. The approach relies on the useof BDD’s superior properties such as low background current, lowadsorption of species, long-term stability, and antifouling capabilitiesusing electrochemical reactivation. Moreover, the article also describesan in situ activation technique, where the electrodes were reactivated within human urine, thereby opening the way towardautomatic quantification of UA with in situ cleaning. The time taken to quantify UA concentration and cleaning remains below0.5 s. Two analytic models were derived, based on different concentrations of ascorbic acid (AA) and uric acid, consisting of 2 sorder calibration curves. Solving the second order equation enables the direct estimation of UA concentration, and valuesdemonstrated good accuracy when compared with spectrophotometric measurements.

Uric acid (UA) is the principal breakdown product ofpurine metabolism,1 and the normal UA concentration in

human urine is around 2 mM.2 Hyperuricemia, associated withrenal disease, can cause gout,3 cardiovascular disease,4 kidneystones,5 etc., whereas hypouricemia can be due to Fanconisyndrome,6 nephritis,7 and other kidney disorders. Hence it isimportant to monitor UA level in bodily fluids such as bloodand urine. Nowadays the determination of uric acid in urine isperformed in medical laboratories using mainly spectrophoto-metric analysis methods. However there is also an interest forcontinuous monitoring of UA in urine in particular for patientsadmitted in intensive care units (ICU), where the earlydiagnostic of acute renal failure (ARF), observed in up to 25%of the patients admitted in ICUs, can have a major impact onthe survival rate of those patients.8 In this context, electro-chemical (EC) detection techniques are seen as a promisingalternative to conventional optical methods due to their goodsensitivity, fast measuring time, portability, low powerconsumption, and cost effectiveness, thus enabling directbedside monitoring. Various electrochemical approaches suchas polymer modified electrode,9−11 chemically modifiedelectrode,12−15 enzyme modified electrode,16 and electro-chemical pretreatment17 were developed to detect UA.However UA coexists with ascorbic acid (AA) in biologicalfluids and has got nearly the same oxidation potential.10

Although modified electrodes show good selectivity to UA,complications like adsorption, fouling, etc. are associated withthose techniques.13 Simultaneous determination of dopamine,AA, and UA were also investigated using fast cyclicvoltammetry (CV) and differential pulse voltammetry (DPV)

without any electrode modification.18 However the analytemust be diluted to prevent the surface from fouling, using thistechnique. We report on a simple electrochemical method forselective quantification of UA using boron doped diamond(BDD) microelectrode, without any special modification or useof any specific reagents or analyte dilution. The technique canbe performed directly in urine and is combined with an in situcleaning/activation technique which brings back the reactivityof the electrode that has been lost due to fouling. Thus it opensthe possibility to automate the urine UA monitoring techniquewith minimal power consumption especially when compared toin situ cleaning by sonoelectrodes.The BDD electrode possesses unique electrochemical

properties viz. low capacitive background current, widepotential window in aqueous media, poor adsorption of polarmolecules, and corrosion resistance in harsh environmentsmaking them highly promising as an electrochemical sensorwhen compared to other conventional electrode materials.19−21

Bioinertness and long-term stability of BDD material makesthem ideal candidates for biomedical applications.22 Because ofits carbon nature, diamond opens the way to immobilization offunctional groups such as DNA, proteins, enzymes, etc. ontothe electrode surface,23 thereby bringing high selectivity tosensing process. Popa et al. has achieved UA and AA peakseparation by anodizing the diamond surface.2 However, at highpH values the peak separation was diminished thus making UA

Received: May 3, 2012Accepted: November 5, 2012Published: November 5, 2012

Article

pubs.acs.org/ac

© 2012 American Chemical Society 10207 dx.doi.org/10.1021/ac301177z | Anal. Chem. 2012, 84, 10207−10213

quantification difficult in the presence of AA. Moreover, theanalyte has to be diluted by several thousand folds to obtainreliable results. Fast CV demonstrates the production of a veryreducible electro-active species that has been produced as aresult of oxidation of UA.24

The typical CV response of UA shows one broad irreversiblepeak at slow potential sweeping (down to 1 V s−1) on classicalmacro-electrodes. By using fast CV (1 V s−1), Dryhurstdemonstrated the existence of a weak reduction peak resultingfrom UA oxidation.24 Indeed, the oxidation of C4C5 bond ofUA gives readily reducible bis-imine (on C4 atom) that mayundergo further irreversible chemical hydration reaction if notquickly electrochemically reduced (Figure 1). Conversely, AAelectrochemical oxidation is known to be highly irreversible.25

Hence fast CV may be used to selectively determine theconcentration of UA in the presence of AA.

The present study reports on the fabrication and character-ization of the BDD microelectrode sensors for the selective andsensitive detection of UA in the presence of low and highquantities of AA. Since the technique used for quantification isfast CV, the low double layer capacitance of diamond reducesthe background current and increases the signal to backgroundratio. Microelectrodes show a decreased ohmic drop, hemi-spherical diffusion layer, and fast establishment of a steady-statesignal when compared to macro-electrodes.26 Moreover, thedouble layer capacitance and the volume required to analyze isfurther decreased by microstructuring such an electrode. Thelong-term goal of our work is to develop BDD bioMEMSplatforms for quantifying different constituents in urine thatcould be directly incorporated in bedside monitoring devices.“As grown” BDD electrodes were used for the quantificationtaking advantage of diamond’s high electrochemical reactivity.When compared to other electrochemical detection techniques,no groups have yet reported the possibility to perform UA

quantification directly in urine without dilution. Otherapproaches based on high-performance liquid chromatography,solid phase extraction, or permselect membranes techniquesrequire sample preparation further to result in the contami-nation of the detection module (membrane). Here ourinnovative detection technique directly enables UA quantifica-tion without particular sample preparation and thus could beimplemented easily in automatic systems, e.g., for home patientmonitoring.

■ EXPERIMENTAL SECTION

BDD BioMEMS Fabrication. A novel technological processwas developed for the fabrication of diamond microelectrodeswhich involves the selective growth of diamond over siliconsubstrates, followed by the deposition of metal contacts andpassivation layers. This process includes initially the fixing ofdetonation diamond nanoparticles over a preoxidized 4 in.silicon wafer using a protocol described elsewhere.27 Then analuminum hard mask consisting of a disk of 100 μm in diameterwas selectively deposited, over the areas where the BDDmicroelectrodes have to be grown. Photolithography enabled topartially pattern the Al mask thus used to protect areas wherediamond nanoparticles were protected from being etched in areactive ion etching (RIE) process in oxygen. The Al hard maskwas finally removed and the diamond electrode was grown. Thedimensions of the resulting diamond disk electrodes weretypically of 300 nm in thickness and 100 μm in diameter. Thenan assembly of Ti (50 nm)/Pt (150 nm) metal tracks wasdeposited over the substrate with a metal ring going around theedge of the diamond disk in order to take electrical contactfrom the diamond electrode. Then a silicon nitride (Si3N4)passivation layer was deposited by CVD over the substrate.Finally an opening of 40 μm was made over the diamondelectrodes by using local etching of the Si3N4 layer by RIE withSF6 gas.

Apparatus, Chemicals and EC Measurement. AnAutolab PGSTAT 302 potentiostat was used to perform allthe EC characterizations and experiments. The preparedmicroelectrodes were characterized by cyclic voltammetry(CV) and electrochemical impedance spectroscopy (EIS). Athree electrode setup consisting of the BDD microelectrode asthe working electrode, a platinum wire mesh as the counterelectrode, and an Ag/AgCl (3 M KCl) reference electrode wasemployed for the CV experiments. The Ag/AgCl electrode wasreplaced by a Pt pseudoreference for the EIS experiment. EISwas recorded over a frequency range of 50 kHz−0.1 Hz withlogarithmic point spacing and potential amplitude of 0.01 Vrms, while the BDD electrode was maintained at an opencircuit potential. The electron transfer rate k0 was determinedfrom the Nyquist plot fit using the ZSimWin 3.21 software.Ultrapure deionized (DI) water (Millipore Direct Q3) was usedto prepare all the solutions. EIS was performed in an aqueoussolution containing 0.5 M potassium chloride (AcrosOrganics), 1 mM of potassium ferricyanide(III) and ofpotassium hexacyanoferrate(II)·trihydrate (both from AcrosOrganics). The 0.5 M lithium perchlorate (Sigma Aldrich)aqueous solution was used as the electrolyte for themeasurement of the potential window. The electrodes wererinsed in DI water and dried under a flow of argon gas prior toeach experiment. The steady state limiting current plateau wasobserved in a 0.5 M potassium chloride aqueous solutioncontaining 1 mM Fe(CN)6

4− ion while the electrodes were

Figure 1. Proposed schematics for UA oxidation in the presence/absence of AA where (I) is UA, (II) bis-imine compound, (III) imine-alcohol compound, and (IV) uric acid-4,5-diol.

Analytical Chemistry Article

dx.doi.org/10.1021/ac301177z | Anal. Chem. 2012, 84, 10207−1021310208

scanned at 100 mV s−1 from 0.05 to 0.45 V vs Ag/AgCl. Theelectrode electron transfer rate k0 is defined as

=k RT n F SR C/02 2

T 0 (1)

where R is the universal gas constant, T the absolutetemperature (K), S the surface area of the electrode (cm2), FFaraday’s constant (96 500 C mol−1), RT the electron transferresistance of the electrode (ohm), C0 the concentration ofredox couple (mol cm−3), and n the number of electronstransferred. The limiting current ilim recorded at the BDDelectrode is given by the following equation:

=i nrFDC4lim 0 (2)

where r the radius of the electrode, D the diffusion coefficientof Fe(CN)6

4− (6.67 × 10−6 cm2 s−1),28 and C the bulkconcentration of the species. Unless stated otherwise, thepotential is given versus an Ag/AgCl reference electrodethroughout the paper.UA and AA solutions were prepared daily by dissolving in a

phosphate buffer saline (PBS) solution of pH 7.2 (all threechemicals are from Sigma Aldrich). Oxidation and reductionpeak were obtained from fast CV (scan rate = 20 V s−1) fordifferent concentration of UA and AA. The electrode wasscanned from −0.3 to 1.4 V vs Ag/AgCl. Calibration curves ofUA were plotted from the obtained value. Urine samples werecollected from volunteers and were used without anypretreatment or dilution.

■ RESULTS AND DISCUSSIONEC Characterization of BDD Film. CV in LiClO4 solution

has demonstrated that the accessible potential window of theBDD film is about 3.4 V with a background current of 30 pA at0.2 V s−1 and a steady state limiting current ilim of 4.7 nA in a0.5 M potassium chloride aqueous solution containing 1 mMFe(CN)6

4− ion. The EIS results demonstrated that theelectrodes exhibit a very fast electron transfer rate (k0) up to0.02 cm s−1. The potential window corresponds to that ofdiamond and the fast electron transfer rate, and the lowbackground current demonstrates that neither cracks norpinholes were detected in the passivation layer.CV of AA and UA. Figure 2a shows the CV of AA and UA

recorded at low scanning rate (0.1 V s−1) under steady statehemispherical diffusion in PBS buffer. Both AA and UA exhibitwell-defined oxidation waves with half-wave potentials around0.5 V vs Ag/AgCl that correspond to the irreversible exchangeof two electrons and two protons.24,25 These CVs clearly showthe overlapping of the electrochemical oxidation waves of bothspecies at the same potential range. More precisely, the ECoxidation of UA gives, through a reversible two electron-twoproton exchange, an unstable bis-imine that could exist in twotautomeric forms (species IIa and IIb, Figure 1).29 This bis-imine compound is then decomposed, at the used experimentalpH of 7.2, into uric acid-4,5-diol (species IV, Figure 1), throughtwo fast irreversible hydration steps. By increasing the scanningrate from 0.1 to 20 V s−1, one can see, as already mentioned byDryhurst, the appearance of a weak reduction peak (for whichthe amplitude increases with the scan rate) at a potential of−0.1 V vs Ag/AgCl (Figure 2b). Studies have been carried outto evaluate the signal to background ratio (S/B) as a function ofthe scan rate. The S/B was found to increase with the scan rateup to an optimal value of 20 V s−1 and to decrease beyond thisvalue. Indeed, provided the sweep rate is fast enough, the

primary oxidized products IIa or IIb can be electrochemicallyreduced owing to the reversibility of the first oxidation step dueto the instability of the imine compound. At slower scan rates,oxidation products of UA are rapidly hydrated to irreversibleuric acid-4,5-diol. Conversely to UA and as expected, noreduction wave of the electrochemically produced dehydroas-corbic acid (DHAA) was observed in the scanned potentialwindow (Figure 2b), even at the highest investigated scanningrate of 20 V s−1. However, DHAA is known to strongly adsorbat electrochemical interfaces and may induce diamond foulingalong UA determination. Hence, according to the aforemen-tioned electrochemical behavior, fast CV can be used toselectively determine the concentration of UA.Elsewhere, one can distinguish from Figure 2a for UA, the

existence of a second flat wave (oxidation half-wave potential of0.8 V vs Ag/AgCl). Indeed, uric acid-4,5-diol undergoessubsequent chemical rearrangement at pH 7 leading to theformation of allantoin and urea as the main products. However,parabanic acid can be also produced from uric acid-4,5-diol,through complex chemical rearrangement to dihydroxyimida-zole and subsequent two electron-two proton electrochemicaloxidation. Indeed, Dryhurst has clearly shown that such latterdecomposition follows a minority path which increases in yieldwith acidity of the media.24,29 Moreover, the yield of thissecondary path strongly depends on the electrode material asreported by Struck et al.30 who detected parabanic acid

Figure 2. Cyclic voltammograms of uric acid (dashed line) andascorbic acid in phosphate buffer solution scanned at (a) 0.1 V s−1

indicating overlapping oxidation potential and at (b) 20 V s−1.Although peak A is observed for both uric acid and ascorbic acid, peakB is observed only for uric acid reduction.

Analytical Chemistry Article

dx.doi.org/10.1021/ac301177z | Anal. Chem. 2012, 84, 10207−1021310209

reduction using polarography following UA oxidation at aspectroscopic graphite electrode in place of pyrolitic graphite.In such a way, according to the local pH decrease at the vicinityof the electrode owing to UA oxidation and to the acidiccomportment and the nature of hydrogenated diamondelectrodes, parabanic acid may be produced with a yield ofaround 6% (obtained from ratio of the plateau limiting currentsthat involve both two electrons and two protons). One cannotice the disappearance of such a second electrochemical stepin fast CV (Figure 2b), and this behavior can be ascribed to theenhanced electrochemical recycling of UA that decreases theyield of production of parabanic acid which is producedthrough a slow chemical and electrochemical pathway.Moreover, this second electrochemical oxidation peak is hardlyvisible due to the Cottrell evolution of the anodic current, itselfbeing attributed to the primary oxidation wave of UA.Calibration Curves for UA Concentration. CV of UA (1

mM), at the BDD electrode in a solution that contains varyingconcentrations of AA, is shown in Figure 3. In order to plot the

calibration curve, the concentration of the UA solution wasvaried from 0, 500, 1000, 1500, 2000 to 2500 μM and AAconcentration from 0, 250 to 500 μM. CVs of differentcombinations of UA and AA mixtures were done, and the peakoxidation current (iA) and peak reduction current (iB) wereobserved at 20 V s−1. Two 3D curves were plotted with UA andAA concentrations (CUA and CAA) on the X and Y axes and iA oriB values on the Z axis, respectively. Two second orderequations were plotted from the curve which corresponds tomodel 1:

= + + − ×

− × − ×

−

− −

i C C C

C C C

4.24 0.07 0.06 (2.9 10 )

(2.36 10 ) {(1.82 10 )( )}

A UA AA6

UA2

5AA

2 5UA AA

(3)

= + − − ×

− × − ×

−

− −

i C C C

C C C

8.2 0.02 0.009 (1.9 10 )

(2.67 10 ) {(8.52 10 )( )})

B UA AA6

UA2

7AA

2 6UA AA

(4)

By solving the eqs 3 and 4, the concentration of UA and AA(CUA and CAA) can be obtained. It was observed that iAincreases with the concentration of UA as well as that of AA,whereas iB decreases with concentrations of AA but increases

with concentrations of UA. AA can deactivate the BDDelectrode due to deposition of its oxidation product. It wasobserved from Nyquist plots that the reactivity of the activeBDD electrode was reduced by 10% after a few CVs insolutions containing AA. This could be one explanation for thedecrease in iB values when the concentration of AA is increased.The other assumption is derived from the antioxidant nature ofAA. AA is known to reduce quinone imines.31 In the presenceof AA, the bis-imines might have been attacked or quicklyreduced to UA chemically. Hence, as the concentration of AA isincreased, the amplitude of iB is decreased.When the concentration of AA was increased beyond 1 mM,

the peak B has disappeared completely for UA concentrationsbelow 2 mM. Hence model 1 cannot be used to determine theconcentration of UA. However, a third peak (peak C) appearsat 0.8 V vs Ag/AgCl as seen in Figure 4. Thus, a second model

(model 2) is proposed using the two peaks (peaks A and C)and their corresponding oxidation peak currents (iA and iC).For different concentrations of UA and AA mixtures, the peakoxidation current (iA) and second peak oxidation current (iC)were extracted from the CV measurements at 20 V s−1. Theconcentration of the UA solution was varied from 0, 500, 1000,1500, 2000 to 2500 μM and AA concentration from 1, 3 to 5mM. Like model 1, two 3D curves were plotted with UA andAA concentrations (CUA and CAA) on the X and Y axes and iA oriC values on the Z axis, respectively. The equations of model 2are

= + + + ×

− × − ×

−

− −

i C C C

C C C

27.93 0.02 0.07 (1.36 10 )

(3.54 10 ) {(3.14 10 )( )}

A UA AA6

UA2

9AA

2 6UA AA

(5)

= + + + ×

+ × + ×

−

− −

i C C C

C C C

25.2 0.017 0.03 (3.89 10 )

(2.43 10 ) {(2.41 10 )( )}

C UA AA6

UA2

6AA

2 6UA AA

(6)

The concentration of UA and AA (CUA and CAA) can thus beobtained by solving eqs 5 and 6. Both iA and iC increase withthe concentration of UA and that of AA. For higher

Figure 3. Cyclic voltammogram of 1 mM uric acid and ascorbic acid(0, 250, 500 μM) in phosphate buffer solution scanned at 20 V s−1.Peak A increases with rising concentrations of both uric acid andascorbic acid, whereas peak B decreases with rising concentrations ofascorbic acid.

Figure 4. Cyclic voltammograms of 1.5 mM uric acid and ascorbic acid(0, 2, 4 mM) in phosphate buffer solution scanned at 20 V s−1. PeaksA and C increase with rising concentrations of both uric acid andascorbic acid.

Analytical Chemistry Article

dx.doi.org/10.1021/ac301177z | Anal. Chem. 2012, 84, 10207−1021310210

concentrations of AA (>5 mM), it was observed that the valueiA does not depend much on UA concentration. As proposedearlier, the AA catalyzes the production of imine-alcohol, andno peak B is observed at higher concentrations of AA. On BDDelectrodes, the peak C was not observed during the CV insolely AA even at higher concentrations. The simultaneousoxidation of UA in the presence of high concentrations of AApresents a complex mechanism. At a higher concentration ofAA, the adsorbed oxidation products of AA influence theelectro-kinetics of UA oxidation. A possible explanation isrelated to the fouling properties of DHAA. Thereby, because ofthe blockade of electroactive sites by DHAA, UA oxidation maytake place through two different pathways: one via stillelectroactive DHAA-free diamond-surface sites and one viaDHAA blocked sites through the adsorbed species. Oxidationof UA at the active sites and at the fouled surface could causethe peak separation. On the other hand, one can note that peakC potential fits quite well with the potential of the secondplateau observed at slow scan rate, thus suggesting the presenceof parabanic acid. Thereby, the increase in the peak Camplitude with both UA and AA can be explained first by ahigher generation of parabanic acid with UA concentration andsecond to the possible recycling of parabanic acid owing to theantioxidant nature of AA. Indeed, some complementary studiesare on the way to examine more in depth the contributions ofparabanic acid and/or DHAA fouling.The pH of the UA solution was varied from 2 to 13, and the

model 1 was tested in a solution containing 1 mM UA and 0.5mM AA. It was observed that the pH variation does not impacton the peak values of iA and iB. Also there was no interferenceobserved with other compounds and including glucose, NaCl,and KCl.Proposed Model vs Spectrophotometric Quantifica-

tion. Human urine samples were selected as real samples foranalysis by the proposed models and were compared with thespectrophotometric method. In spectrophotometric method,uric acid is oxidized to allantoin in the presence of the enzymeuricase, which leads to formation of H2O2, which reacts with 4-amino phenazone in the presence of peroxidase to formquinone-diimine. The intensity of the color of the quinonediimine is directly proportional to the concentration of uricacid. Human urine samples were neither diluted nor pretreatedfor EC quantification by the proposed models. CV was done oneach urine sample and based on the nature of the peaks, usingmodel 1 or model 2; the equations were solved to identify theUA value. The results are presented in Table 1. The measuredvalues using the proposed model were observed to be veryclose to the spectrophotometric results with a maximumdifference of 13%. The spectrophotometric measuringtechnique, although widely used in the field of UA estimation,has an error percentage of 5%. UA and AA are fast antioxidants,

and the time lag between the spectrophotometric and ECmeasurement might have also influenced the difference in theresults. The other parameter which could affect the measure-ment is the variation in electron transfer rate k0 of the electrode.Only one electrode was used to verify the comparison

between the spectroscopic and electrochemical (using thepresented models) results. It was observed that iA and iB/iCvalues were the same for any BDD electrode prepared using thesame technique (or from the same batch) after electrochemicalactivation. It was also observed that after activation, the CVcurves overlap and iA and iB/iC values with a maximumdeviation less than 1%. Furthermore, the reactivity of theelectrode (k0) has an influence on the peak current values. Ithas been already shown by our group that electrodes from amicroelectrode array exhibit different k0 values because of therelative variations in boron intake in the diamond grains as wellas the associated difference in surface area due to the diamondroughness.32 The fouling of the electrode is another parameterthat also affects the reactivity of the electrode. As the analyte isa complex biological fluid, electrodes were systematically andautomatically cleaned between each measurement.

Automation of Quantification Procedure and in SituCleaning. When electrodes are used continuously in abiological fluid, they lose their reactivity because of fouling.33

Electrode fouling can be due to adsorption or adhesion ofbiomolecules such as proteins, enzymes, cells, intermediateproducts of oxidation of organic compounds, etc.34 Althoughhydrogen terminated BDD exhibit high reactivity, continuoususe in urine leads to deactivation of electrode reactivity becauseof fouling. This would lead to difficulty in automation ofquantification process. Earlier, we have reported on an in situactivation process of BDD electrode which has been tried onseveral biological and synthetic fluids.34 Human urine sampleswere diluted 2-fold in PBS solution. From this solution, fiveother solutions were prepared by adding different quantities ofUA (400, 800, 1200, 1600, and 2000 μM). CVs at 20 V s−1

were done in solution 1 (diluted urine) through 6 (dilutedurine + 2000 μM UA). iA and iB values were obtained from eachscan, and from the CV of solution 1 the UA concentration wasestimated to be 2012 μM. It was observed that the values of iAand iB were not increased as expected, after each scan (solutions1−6) despite the increase in concentration of UA (Figure 5).This is due to fouling of the electrode. The electrodes were ECcleaned and CVs were done from solutions 1 (diluted urine)through 6 (diluted urine + 2000 μM UA). Between each CV,the electrodes were activated in the previous solution. The ECactivation was done directly in the solution containing urine, byapplying a train of negative current pulses consisting of 3 pulseswhere each pulse has amplitude of −100 mA cm−2 and durationof 100 ms and duty cycle of 90%. The values of iA and iB wereincreased as expected, after each scan (Figure 6). Values of iAand iB were extracted from these graphs and are termed as iAO1,iBO1 and iAO2, iBO2 for trails with and without activation betweenCVs. The concentration of UA was estimated using eqs 3 and 4.Also the theoretical values of iA and iB were calculated using themodels and are termed iAC and iBC. The close comparisonbetween the calculated value iAC and observed value iAO1 (trialswith activation) and iAO2 (trials without activation) demon-strated that there was a negligible difference between iAC andiAO1 when compared with the difference between iAC and iAO2. Adetailed comparison between the different values is depicted inTable 2 and Figure 7, where Table 2 shows the expected andmeasured concentration of UA. Because of heavy fouling of the

Table 1. Comparison of Uric Acid Concentration Measuredin Different Urine Sample Using Models 1 and 2 and theSpectrophotometric Technique

sample no. spectrophotometer (mM) proposed model (mM) % error

1 3.66 4.15 132 4.76 5.10 7.33 3.01 2.68 −114 5.90 5.15 −125 2.78 2.74 −1.4

Analytical Chemistry Article

dx.doi.org/10.1021/ac301177z | Anal. Chem. 2012, 84, 10207−1021310211

electrode, the electrode was deactivated and hence thepercentage recovery was as low as 67.5%, whereas thepercentage recovery for activated trials were close to 100%. Itis clearly demonstrated that a simple EC activation procedurecan restore the loss of reactivity of the electrode for theelectrode to be reused in the sample solution for consecutivemeasurements.

■ CONCLUSIONS

Selective determination of UA in the presence of AA wasachieved using a BDD microelectrode without any furthermodification. Comparison of the EC quantification techniqueand that of the spectrophotometric technique shows that anaccurate measurement can be done using the two proposedmodels. This technique highlights the potential of BDDelectrodes as a biosensor owing to its low double layercapacitance, robustness at high current density, and corrosionresistance. The EC treatment retrieves the lost reactivity of anelectrode either aged in air or fouled by a medium withoutusing any specific reagent or solution. The advantage of thistechnique is to enhance the reusability of the BDD electrode byactivating in urine itself. This demonstrates the possibility ofautomation of UA quantification as the electrode can beactivated directly in urine and hence it can be used forcontinuous monitoring for a long period of time. The timetaken for activation is 300 ms, and the time taken for CV at 20V s−1 is less than 200 ms.

■ ASSOCIATED CONTENT

*S Supporting InformationSecond order curve fitting results and cyclic voltammogram ofuric acid. This material is available free of charge via theInternet at http://pubs.acs.org.

Figure 5. Cyclic voltammogram of urine diluted by 2-fold and that ofdiluted urine containing added uric acid (250−1250 μM) scanned at20 V s−1. Both peaks A and B are expected to increase steadily, but dueto fouling of the electrode the amplitudes of peaks A and B are notincreased as expected.

Figure 6. Cyclic voltammogram of urine diluted by 2-fold and that ofdiluted urine containing added uric acid (250−1250 μM) scanned at20 V s−1. The electrodes were activated electrochemically in the samesolution in between two successive scans.

Table 2. Comparison between the Measured Concentrationand Expected Concentration for Uric Acid for BothActivated and Nonactivated Trials

expectedconcentration

measuredconcentration(activation)

%recovery

measuredconcentration (no

activation)%

recovery

2012 2012 100.00 2012 100.002412 2496 103.48 2250 93.282812 2924 103.98 2375 84.463212 3330 103.67 2479 77.183612 3566 98.72 2751 76.164012 4010 99.95 2708 67.50

Figure 7. Comparison between the calculated value (iAC, iBC),observed value with activation between CVs (iAO1, iBO1), and observedvalue without activation between CVs (iAO2, iBO2) for the peak currentsiA and iB extracted from CV in urine solutions 1−6.

Analytical Chemistry Article

dx.doi.org/10.1021/ac301177z | Anal. Chem. 2012, 84, 10207−1021310212

■ AUTHOR INFORMATIONCorresponding Author*E-mail: raphael.kiran@cea.fr.NotesThe authors declare no competing financial interest.

■ ACKNOWLEDGMENTSThe authors would like to give special thanks to the SMTmedical division of CEA Saclay, and namely, CatherineBourcier and her colleagues who performed the samplepreparation and the spectrophotometric measurements. Theauthors also acknowledge our colleagues who volunteered toprovide urine samples for the tests.

■ REFERENCES(1) Alvarez-Lario, B.; Macarron-Vicente, J. Rheumatology (Oxford,England) 2010, 49, 2010−2015.(2) Popa, E.; Kubota, Y.; Tryk, D. A.; Fujishima, A. Anal. Chem. 2000,72, 1724−1727.(3) Martinon, F.; Petrilli, V.; Mayor, A.; Tardivel, A.; Tschopp, J.Nature 2006, 440, 237−241.(4) Daoussis, D.; Kitas, G. D. Rheumatology (Oxford, England) 2011,50, 1354−1355.(5) Sun, Q.; Shen, Y.; Sun, N.; Zhang, G. J.; Chen, Z.; Fan, J. F.; Jia,L. Q.; Xiao, H. Z.; Li, X. R.; Puschner, B. EurJ. Pediatr. 2010, 169,483−489.(6) Shaikh, A.; Wiisanen, M. E.; Gunderson, H. D.; Leung, N. Ann.Pharmacother. 2009, 43, 1370−1373.(7) Dinour, D.; Gray, N. K.; Campbell, S.; Shu, X.; Sawyer, L.;Richardson, W.; Rechavi, G.; Amariglio, N.; Ganon, L.; Sela, B.-A.;Bahat, H.; Goldman, M.; Weissgarten, J.; Millar, M. R.; Wright, A. F.;Holtzman, E. J. J. Am. Soc. Nephrol.: JASN 2010, 21, 64−72.(8) Lameire, N.; Van Biesen, W.; Vanholder, R. Nephrol. Dial.Transplant. 1999, 14, 2570−2573.(9) Li, Y.; Lin, X. Sens. Actuators, B 2006, 115, 134−139.(10) Lin, X.; Zhang, Y.; Chen, W.; Wu, P. Sens. Actuators, B 2007,122, 309−314.(11) Lin, L.; Chen, J.; Yao, H.; Chen, Y.; Zheng, Y.; Lin, X.Bioelectrochemistry (Amsterdam, Netherlands) 2008, 73, 11−17.(12) Zen, J.; Tang, J. Anal. Chem. 1995, 67, 1892−1895.(13) Zen, J.-M.; Chen, P.-J. Anal. Chem. 1997, 69, 5087−5093.(14) Wang, Z.; Wang, Y.; Luo, G. Analyst 2002, 127, 1353−1358.(15) Fernandez, L.; Carrero, H. Electrochim. Acta 2005, 50, 1233−1240.(16) Nakaminami, T.; Ito, S.; Kuwabata, S.; Yoneyama, H. Anal.Chem. 1999, 71, 1928−1934.(17) Strochkova, E. M.; Tur’yan, Y. I.; Kuselman, I.; Shenhar, A.Talanta 1997, 44, 1923−1928.(18) Safavi, A.; Maleki, N.; Moradlou, O.; Tajabadi, F. Anal. Biochem.2006, 359, 224−229.(19) Yano, T.; Tryk, D. A.; Hashimoto, K.; Fujishima, A. J.Electrochem. Soc. 1998, 145, 1870−1876.(20) Tian, R.; Zhi, J. Electrochem. Commun. 2007, 9, 1120−1126.(21) Swain, G. M. Anal. Chem. 1993, 65, 345−351.(22) Panizza, M.; Cerisola, G. Electrochim. Acta 2005, 51, 191−199.(23) Agnes, C.; Ruffinatto, S.; Delbarre, E.; Roget, A.; Arnault, J.-C.;Omnes, F.; Mailley, P. IOP Conf. Ser.: Mater. Sci. Eng. 2010, 16, 1−11.(24) Dryhurst, G. J. Electrochem. Soc. 1972, 119, 1659−1664.(25) Hu, I.-F.; Kuwana, T. Anal. Chem. 1986, 58, 3235−3239.(26) Lawrence, N. S.; Pagels, M.; Meredith, A.; Jones, T. G. J.; Hall,C. E.; Pickles, C. S. J.; Godfried, H. P.; Banks, C. E.; Compton, R. G.;Jiang, L. Talanta 2006, 69, 829−834.(27) Scorsone, E.; Saada, S.; Arnault, J. C.; Bergonzo, P. J. Appl. Phys.2009, 106, 014908.(28) Konopka, S. J.; Mcduffie, B. Anal. Chem. 1970, 42, 1741−1746.(29) Fry, A. J.; Dryhurst, G. Organic Electrochemistry; Springer-Verlag:Berlin, Germany, 1972; p 85.

(30) Struck, W. A.; Elving, P. J. Anal. Chem. 1964, 36, 1374−1375.(31) Kang, P.; Dalvie, D.; Smith, E.; Renner, M. Chem. Res. Toxicol.2009, 22, 106−117.(32) Kiran, R.; Rousseau, L.; Lissorgues, G.; Scorsone, E.; Bongrain,A.; Yvert, B.; Picaud, S.; Mailley, P.; Bergonzo, P. Sensors 2012, 12,7669−7681.(33) Iniesta, J.; Michaud, P. A.; Panizza, M.; Comninellis, C.Electrochem. Commun. 2001, 3, 346−351.(34) Wang, J.; Hutchins, L. D. Anal. Chem. 1985, 57, 1536−1541.

Analytical Chemistry Article

dx.doi.org/10.1021/ac301177z | Anal. Chem. 2012, 84, 10207−1021310213