ANALYTIC4 CHIMICA ACTA
ELSEVIER
Analytica
Chimica Acta 332 (1996) 131-138
Hydrogen peroxide amperometric peroxidase-graphite-epoxy
biosensor based on a biocomposite
A. Moralesa, F. Ckspedes”, J. Mufiozb, E. Martinez-F$bregasa,
S. Alegret”**
“Grup de Sensors & Biosensors, Departament de Quimica, Universitat Autcinoma de Barcelona, 08193 Bellaterra, Catalonia, Spain bCentre National de Microelecttinica, Universitat Audnoma de Barcelona, 08193 Bellaterra, Catalonia, Spain Received
19 October
1995; revised 23 April 1996; accepted
1 May 1996
Abstract An amperometric biosensor sensitive to hydrogen peroxide has been built, using a HRP-graphite-epoxy biocomposite. The proximity of the redox centers of the enzyme and the conductive sites at the surface of the electrode permits a direct regeneration of the enzyme. The response characteristics of this biosensor have been improved by adding powdered platinum to the matrix of the biocomposite. When Pt is absent, the working potential is -300 mV vs. Ag/AgCl and the response is linear between 0.03 rnM and 7 mM. When 3% wt platinum is added to the biocomposite, a stable response is produced at -50 mV vs. Ag/AgCl and the linearity range lies between 0.09 mM and 9 mM. The response time is approximately 2 s for both the cases. Keywords: Biosensors;
Hydrogen
peroxide;
Horseradish
peroxidase;
1. Introduction Hydrogen peroxide measurement is interesting in a large variety of samples of industrial and environmental fields as well as a by-product of several analytical enzymatic reactions. Hydrogen peroxide has been measured electrochemically by direct oxidation at the electrode (Eq. (I)) [l]. The potential (I?) used for measuring the peroxide is dependent on the nature of the working electrode (platinum, gold, graphite, graphite-polymer composite, etc.). At these potentials (650-1200 mV)
* Corresponding author. Tel.: 343 581 1017; fax: 343 581 2477; e-mail:
[email protected].
Graphite-epoxy
biocomposite
other species present in the sample solution can act as interferents [2]. H202 3 02 + 2H+2e
.
(1)
The integration of new materials, new fabrication strategies and conventional voltammetric techniques has enabled the appearance of new amperometric biosensors that are simpler, more robust, more reliable and less costly. The use of amperometric biosensors based on horseradish peroxidase (HRP) is an attractive alternative for the measurement of hydrogen peroxide. The analytical signal of a HRP biosensor is normally based on the monitoring of the reduction of an oxidized redox mediator (Eq. (4)) produced during the enzymatic reaction
A. Morales
132
et al./Analytica Chimica Acta 332 (1996) 131-138
(Eqs. (2) and (3)) H202 + 2H+ + HRPred 4 HRP,, + MED,,d MEDo, 3
2H20 + HRP,,,
--+ HRP,d + MED,,,
MED,ai,
(2) (3) (4)
where HRPred, HRP,,, MEDred and MED,, are the reduced and oxidized forms of the enzyme and the mediator, respectively. The working potential is determined by the mediator being used, and is generally below 600mV. However, these mediators [3-51 in their oxidized or reduced form, or both are very soluble in the electrolytic medium, which is often aqueous [6-81, and this is an inconvenience. Some of the mediators used frequently include ferrocene, tetrathiafulvalene and tetracyanoquinodimethane [9]. The loss of the chemical mediator implies a rapid loss of sensitivity of the biosensor. Recent studies have proposed the direct regeneration of HRP on the electrode surface. This regeneration is attained by direct electron transfer from the redox centers of the enzyme (associated to the ferroheme/ferriheme pair) and the conducting sites on the biosensor surface (Eq. (5)) as HRP,, + 2e- -% HRPred.
(5)
Electronic transfer between the HRP and the electrode is relatively faster when compared with other reductases. This phenomenon is due to the tridimensional configuration of this protein, where the hemin group is in the outer region, facilitating the electron transfer [ 10-161. This direct transfer has been achieved by adsorbing the enzyme on the surface of an electrode made of carbon black [ 1 l] or spectroscopic graphite [17]. The direct mixing of HRP within the bulk of carbon paste has been proposed as well [6,18]. Unfortunately, these approaches to enzyme immobilization are inherently weak which is translated as a decrease in sensitivity as the enzyme is lost from the surface of the electrode. Our working group has a long experience in the design and construction of rigid matrix materials using enzyme-graphite-polymer composites. This design approach is very simple and the results have
been satisfactory [ 19-241. The location of the enzyme within the conducting composite matrix forces a close electrical contact between the enzyme and the conducting sites in the bulk and therefore on the surface of the electrode. Additionally, the rigidity of the biocomposite ensures a simple and lasting immobilization of the enzyme. On the other hand, if enzyme loss or denaturation occurs by external causes, the biosensor surface can be regenerated by a simple polishing operation [25-271. Often, the active redox center of the biological component is located in the less accessible parts of the molecule. This slows down the transfer of electrons between the enzyme and the conducting sites of the electrode, thus slowing down response times. The working potential is also displaced to inconvenient values. To get around this obstacle, many researchers have resorted to the use of a metallic species that catalyzes this electron transfer. Metals such as rhodium [9], mercury or platinum [28] modifying the electrode surface have been used for this purpose. Cardosi [15] reports the possibility of an electrocatalytic reduction in HRP when it is immobilized in platinized carbon. The preparation of the biocomposites as described above permits the addition of these catalysts to the matrix of the material following a very simple procedure. The main objective of the present work is to build and characterize a hydrogen peroxide sensor based on a new conductive biocomposite. This material is made of graphite as the conductor, HRP as the biocatalyst and an epoxy resin acts as a binding agent. The effect on the signal when platinum is added as a catalyst was also studied. These HRP composites represent base materials which can be modified with hydrogen peroxide producing enzymes (oxidases) thus allowing the detection of a large variety of substrates.
2. Experimental 2.1. Apparatus
Cyclic voltammograms were obtained with a PGSTAT 20 Autolab potentiostat (Echochemie). Calibration curves were obtained with a 641
A. Morales et al./Analytica
amperometric VA detector (Metrohm) E506 Polarecord plotter (Metrohm).
coupled
Chimica Acta 332 (1996) 131-138
to a
2.2. Reagents Biocomposites were prepared using graphite powder (Merck) with a particle size of less than 50pm, epoxy resin (Epotek H77, Epoxy Technology, Billerica, MA, USA), powder platinum (32,716-g, Aldrich) with a particle size of 0.5-1.2 urn and Horseradish peroxidase (type VI-A EC 1.11.1.7, Sigma). Hydrogen peroxide (Merck) solutions were prepared daily and standardized by an indirect volumetric method [29]. 2.3. Auxiliary
133
epoxy resin [21]. For each gram of this graphiteepoxy composite, 20mg of peroxidase enzyme was added. The mixture was blended thoroughly and placed in the hollow end of a tube to form the body of the electrode (Fig. 1). The biocomposite was cured at 40°C during four days. After curing, the surface was polished starting with medium-sized sand paper and progressing to alumina paper with a particle size of 3 pm (polishing strips Orion 301044-001). This last step is a wet procedure using bidistilled water. When platinum was added to the biocomposite, it was present in a weight proportion of 3%. The blending and filling procedures were identical to those described above.
electrodes 3. Results and discussion
Three electrodes formed the electrochemical cell. The auxiliary electrode was a platinum electrode. A double junction Ag/AgCl reference electrode (Orion 92-02-00) with an external 0.1 M KC1 solution was used. Three working electrodes were studied: one based on a HRP-Pt-graphite-epoxy biocomposite, one featuring a HRP-graphite-epoxy biocomposite and one amperometric transducer based on a graphiteepoxy composite. 2.4. Biosensor
construction
The base conducting composite was prepared by dispersing one part wt. of graphite in four parts wt. of
2cm
i I
4
3.1. HRP reduction potential in the unmodi$ed biocomposite The biosensor featuring the HRP-graphite-epoxy biocomposite was submerged for 30 min in a 0.01 M hydrogen peroxide solution. After soaking, the sensor was washed with bidistilled water, and then used for performing a cyclic voltammetry in an inert nitrogen atmosphere. This treatment was necessary to determine the reduction potential of the HRP enzyme in the biocomposite. The support electrolyte was a 0.1 M phosphate and 0.1 M KC1 solution buffered to pH 7. Fig. 2 shows the cyclic voltammogram obtained for the HRP-graphite-epoxy biosensor. The voltammogram shown in a broken line corresponds to a graphite-epoxy electrode and no reduction or oxidation peaks can be observed. The HRP-graphite-epoxy biosensor shows a reduction peak at -300mV that corresponds to the HRP enzyme. There are several problems associated with the regeneration of the enzyme to its reduced form at -3OOmV. The presence of oxygen causes an instability in the current signal. Furthermore, practically any interfering substance in the oxidized state can be easily reduced at this potential. The addition of metals in the biocomposite can eliminate this inconvenience if the sensitivity is enhanced allowing the applied potential to decrease.
u---N
Fig. 1. Schematic representation of the electrode: (1) HRPgraphite-epoxy biocomposite material, (2) copper contact, (3) electrical contact and (4) electrode body.
134
a t
A. Morales et al./Analytica
-2.00 1 %.
-2.50 -
Chimica Acta 332 (1996) 131-138
-40 -30 i -50 i
-3.00 I
-4.00
~
HRP-Ptgraphite-apoxy
-4.50 -100
,-----I_-
-5.00 -800
-600
-400
-200
0
CT-
i-7-
-800
E/mV
7
---_-‘I’---
-600
-400
-200
0
200
400
Et mV
Fig. 2. Cyclic voltammograms using electrodes based on a graphite-epoxy composite and the HRP-graphite-epoxy biocomposite. The support electrolyte is a 0.1 M phosphate and 0.1 M KC1 solution at pH=7.00. Sweep rate is 50 mV/s in a N2 atmosphere.
Fig. 3. Cyclic voltammograms using electrodes based on a Ptgraphite-epoxy composite and the HRP-Pt-graphite-epoxy hiocomposite. The support electrolyte is a 0.1 M phosphate and 0.1 M KC1 solution at pH=7.00. Sweep rate is 50mVls in a N2 atmosphere.
3.2. HRP reduction potential modified with platinum
will be noted. Consequently, a more stable signal will be obtained. As can be observed in Fig. 4, the intensities for the HRP-Pt-graphite-epoxy biosensor vary between -4.5 pA and -5 pA for the same concentration of hydrogen peroxide. The potential range was between -2OOmV and OmV vs. Ag/AgCl. Oxygen interference is present for cathodic potentials more positive than -1OOmV to higher cathodic potential. At anodic potentials higher than OmV (not shown), intensities are around 0.3 PA. In the potential range from - 100 mV to 0 mV, the maximum intensity is at -5OmV. Therefore, it was decided to work at -5OmV for the biosensor with platinum in the biocomposite. For the same potential range, using HRP-graphiteepoxy biosensors, intensities are constant and close to -0.5 PA.
in the biocomposite
A biosensor with platinum in the biocomposite was built. Fig. 3 shows the cyclic voltammogram of the biosensor in an inert nitrogen atmosphere. The supporting electrolyte was a 0.1 M phosphate and 0.1 M KC1 solution buffered at pH 7. Before recording the cyclic voltammogram, the biosensor was placed for 30 min in a 0.01 M hydrogen peroxide solution. The enzyme regeneration voltage was found to be -400 mV. This value is comparable to the value when no platinum is added to the biocomposite. Since no decrease in the reduction potential of HRP in the presence of platinum is observed, it can be deduced that electron transfer between the hemin group and the electrode is not facilitated. However, an important increase in current (175%) is observed when platinum is introduced in the biocomposite compared with the case when platinum is not present. Therefore, it is possible to work with the platinum biocomposite at a lower cathodic potential. Current intensity will be even higher, and a decrease of oxygen interference
3.3. pH effects The concentration of the buffer solution was 0.1 M at all times. Only the proportion between sodium dihydrogen phosphate and sodium hydrogen phos-
A. Morales et al./Analytica
Chimicu Acta 332 (1996) 131-138
1.0 .5 l
0.0 .
-.5
a 2
’
.
’
.
’
.
.
l
1
-1.0 1 -4.0 . -4.5
.
.
.
. .
.
T
a1
1
l
55 50
.
8
f
.
-5.0 -5.5
8
35i
/m
1
-6.0 -250
!
30 -200
.
60 -I
-150
-100
-50
0
6
50
ElmV Fig. 4. Hydrodynamic voltammograms using one electrode based on HKP-graphite-epoxy and another one modified with Pt. The support electrolyte is a 0.1 M phosphate and 0.1 M KC1 solution at pH=7.00. The hydrogen peroxide concentration was 5 x10-s M.
Fig. 5. Current changes vs. pH. The buffer solution is 0.1 M phosphate. For the HKP-P&graphite-epoxy biosensor, a hydrogen peroxide concentration of 5.6x 10e4 M was used and 5 x lo-‘M for the HRP-graphite-epoxy biosensor. n E=-300mV vs. Ag/AgCl l E=-50mV vs. AgtAgCl.
3
3.4. Linear range of the response
2 s : f? -II
0
-1
-5
I
I
I
-4
-3
-2
Log
3.5. Reproducibility Reproducibility was studied by making calibration runs every 12 h. This was done in two stages. In the
I 6
PH
phate was varied. These experiments are performed using the working potential chosen previously. The effect of pH on both types of biosensor is shown in Fig. 5. A maximum current is obtained at pH 7. At an acidic pH the biocatalytic reaction is diminished. When the pH is alkaline the enzyme is denaturated and the biomaterial is rapidly destroyed.
Fig. 6 shows the calibration curves for both the biosensors. The linearity range for the HRP-graphiteepoxy biosensor lies between 0.03 mM and 7 mM. For the HRP-Pt-graphite-epoxy biosensor the response is linear between 0.09 mM and 9 n&I. Table 1 shows the sensitivities for both the types of biosensors. The biosensor not containing Pt has the higher sensitivity. This is a result of working at an optimal potential where the regeneration of the HRP enzyme occurs.
, 7
-1
P-PJM
Fig. 6. Calibration curves for both biosensor types. The support electrolyte is a 0.1 M phosphate and 0.1 M KCI solution at a pH=7.00. n E=-300mV vs. Ag/AgCl l E=-50mV vs. Ag/ AgCl.
A. Morales et al./Analytica
136 Table 1 Sensitivity of the HRP-graphite-epoxy epoxy biosensors tested
Calibration
and the HRP-Pt-graphite-
Sensitivity (nA M-‘) HRP-graphite-epoxya
Sensitivity (nAM_‘) HRP-Pt-graphite-epoxyb
95499 99456 93325 96180 98569 96606 2.5
16218 16587 17024 16325 16698 16570 2.0
2 3 4 5 Average R.S.D=
The support electrolyte is a 0.1 M phosphate solution at pH=7.00. aWorking potential -300 mV. *orking potential -50 mV. ‘R.S.D: Relative standard deviation (%).
parameters= for biocomposite
HRPgraphite-epoxy polished surface
and 0.1 M KC1
time
For the HRP-graphite-epoxy biosensor, hydrogen peroxide concentration was increased from 0 to 4x 10p4M. The biosensor attained 95% of the response in 2 s. A similar result was obtained for the HRP-Pt-graphite-epoxy biosensor when the hydrogen peroxide concentration varied from 0 to 6.22x 10-4M. In order to study the effect of the enzymatic immobilization over the response time, we repeated the same experiment using a graphite-epoxy transducer and putting the enzyme into the solution (0.1 M phosphate and 0.1 M KC1 buffer solution at pH 7.00). In this case, the response time was less than 2 s when the hydrogen peroxide concentration goes from 0 to 4x 10e4M. This result is an indication that the enzyme is not hindered by being confined to the composite matrix.
HRPPt-graphite-epoxy polished St&ace 1 2 3 unpolished surface 4 5 6
based H20~ biosensors
A
B
R2
4.9910 4.9921 4.9913
0.93 11 0.9300 0.93 12
0.9995 0.9992 0.9994
4.8956 4.8997 4.9012
0.9325 0.9326 0.9318
0.9991 0.9994 0.9991
4.2199 4.2156 4.2160
0.9422 0.9563 0.9568
0.9996 0.9996 0.9995
4.2189 4.2089 4.2146
0.9362 0.9487 0.9400
0.9994 0.9993 0.9995
biocompositeb
2 3 unpolished surface 4 5 6
first stage, the biosensor was polished with alumina paper after each calibration run. In the second part of the study, no polishing was involved and the biosensor was washed in bidistilled water after each calibration. A different biosensor was used in each part of the study. Table 2 shows the results of these studies. The enzyme is not lost by dissolution and enzyme activity is stable when the enzyme is confined to the epoxy matrix. When the surface of the biosensor is polished, a fresh active surface appears. 3.6. Response
Table 2 Calibration
1
plot
1
Chimica Acta 332 (1996) 131-138
biocomposi&
The support electrolyte is a 0.1 M phosphate solution at a pH=7.00. “logl=A+B log[Hz02]. bWorking potential -300mV vs. Ag/AgCl. “Working potential -5OmV vs. Ag/AgCl.
and 0.1 M KC1
3.7. Interferences The number of interferent species depends on the working potential and the nature of the sample. The interferents selected for this study include those most likely to appear in biological and food samples. As mentioned in the introduction, the final application of the composites reported here is the construction of bi-enzymatic biosensors including oxidases that produce hydrogen peroxide and a peroxidase (HRP) to measure, for instance, glucose in different types of samples. In food industries it is necessary to determine glucose in juices and sparkling drinks. Antioxidants such as citric acid, ascorbic acid, benzoic acid are common possible interferents. In glucose blood analysis, frequent interferents are drugs such as acetaminophen and chloramphenicol [30]. Furthermore, uric acid is a common metabolite in blood samples [31]. Considering all these facts, it was decided to study the interferents mentioned above using the HRP biosensors with and without platinum in the biocomposite.
A. Morales et al.LAnalytica Chimica Acta 332 (19%) 131-138 Table 3 Relative accuracy
in the determination
of HsOa in aqueous solutions containing
Molar ratio lnterferent species(M):HaOa(M)
Relative accuracy Citric acid:HaOa 4.3 X 10-55 x 10-4 5x1o-4:5x1o-4 1.5x1o-3:5x1o-4 Ascorbic acid:HaOs 4.3x 10-5:5x 10-4 8x1o-4:5x1o-4 1.5x1o-3:5x1o-4 Uric acid: Hz02 4.3x lo-5:5 x 1o-4 8x10-4:5x10~4 1.5x1o-3:5x1o-4 Cloramphenicol:HaOOa 4.3 x 10-5:5x 1o-4 8~10-~:5xlO-~ 1.5x 10-3:5x 10-4 Benzoic acid:HsOa 4.3 x 10-5:5x 1o-4 8x1o-4:5x1o-4 1.5x 1o-3:5x1o-4 Acetaminophen:HzOz 4.3x1o-5:5x1o-4 8x 1O-4:5x lop4 1.5x1o-3:5x1o-4
some common interferent
HRP-graphite-epoxya
137
species using the biosensors
tested
HRP-Pt-graphite-epoxyb (%)
Relative accuracy
Imperceptible Imperceptible Imperceptible
Imperceptible Imperceptible Imperceptible
2.0 7.5 30.4
15.0 26.5 33.0
Imperceptible Imperceptible Imperceptible
Imperceptible Imperceptible Imperceptible
4.3 17.4 42.1
9.5 23.4 34.0
0.6 1.5 2.2
(%)
1.3 3.6 5.6
Imperceptible Imperceptible ImDerceDtible
Imperceptible Imperceptible Imperceptible
The supportelectrolyte is a 0.1 M phosphate and 0.1 M KC1 solution at pH=7.00 in a Na atmosphere. “Working potential bWorking potential
1300 mV vs. &/A&l. -5OmV vs. Ag/AgCl.
Table 3 shows the results of this interference study at different working potentials. From the results it is concluded that only ascorbic acid, chloramphenicol and benzoic acid interfere with the analytical signal. 3.8. Determination of the Michaelis-Menten apparent constant. The Lineweaver-Burk graphic representation was used to calculate the Michaelis-Menten apparent constant KEP. It was found that KmapP=26mM for the biosensor containing platinum. For the biosensor without platinum KZP was 37.6mM. The apparent constant is dependent on the working potential. Therefore, the two apparent constants cannot be compared since they were obtained at different potentials.
4. Conclusions A new type of amperometric hydrogen peroxide biosensor has been built. The biosensor is based on a HRP-graphite-epoxy biocomposite which is low cost and is simple to prepare. A homogeneous dispersal of the HRP enzyme in the conducting rigid matrix gives several desirable features to the sensor. The direct contact of the enzyme with the conducting sites in the surface of the electrode permits a direct regeneration of the enzyme so redox mediators are not needed. Also, a simple polishing produces a fresh surface that gives reproducible results. The regeneration potential of the HRP enzyme present in the biocomposite is -3OOmV. At this potential, enzyme denaturation was not observed. When platinum is added to the matrix of the biocomposite, the regeneration potential of the enzyme is -400 mV. However, an important increase
138
A. Morales et al./Analytica
in current is observed when platinum is introduced in the biocomposite. Because of this, lower working potentials (-50 mV) can be used. At this lower potential, the analytical signal is reproducible, the stability is better than at -300 mV and measurements can be made even in the presence of oxygen.
Acknowledgements Adriana Morales acknowledges a fellowship from the Direction General de Asuntos de1 Personal Academico (DGAPA) of the Universidad National Autonoma de Mexico (UNAM). The present work was partially funded by CICYT, Madrid (project BI093-0635 and BI095-1196-CE).
Chimica Acta 332 (1996) IJI-I38 [111 WI [I31 [I41 [I51 [I61
[I71 WI L191 Lw PII
References
WI ~31
G.G. Guibault, Analytical uses of immobilized enzymes, Marcel Dekker, New York, 1984. p. 1. 121 M.F. Cardosi and A.P.F. Turner, in A.P.F. Turner (Ed.), Advances in Biosensors, Vol. 1, 1991, p. 125. [31 J. Wang, L.H. Wu and L. Angnes, Anal. Chem., 63 (1991) 2993. [41 T. Ikeda, T. Shihata and S. Senda, J. Electroanal. Chem., 261 (1989) 351. [51 J.J. Kulys and N.K. Cenas, Biochim. Biophys. Acta, 744 (1983) 57. 1 (1989) 465. [61 G. Jonsson and L. Gorton, Electroanalysis, 171 Y. Degani and A. Heller, J. Am. Chem. Sot., 111 (1989) 2357. WI M. Green and M.A. Hill, J. Chem. Sot., Faraday Trans. 1, 82 (1986) 1237. 191 J. Janata, M. Josowick and M. Devaney, Anal. Chem., 66 ( 1994) 207R. K.G. Welinder, Eur. J. B&hem., 96 (1979) 483. WI HI
v41
~251 1261 ~271 WI ~291 [301 r311
A. Yaparov and V Malovick, Dokl. Akad. Nauk. SSSR., 249 (1979) 1399. R.M. Paddock and F. Bowden, J. Electroanal. Chem., 260 (1989) 487. J. Kulys and R.D. Schmid, Bioelectrochem. Bioenerg., 24 (1990) 305. V. Wollenberger and V. Bogdanouskaya, Anal. Lett., 23 (1990) 1795. M.F. Cardosi, Electroanalysis, 6 (1994) 89. L. Gorton, G. Jonsson-Pettersson, E. Csoregi, K. Johansson, E. Dominguez and G. Marko-Varga, Analyst, 117 (1992) 1235. L. Gorton, G. Bremle, E. Csoregi, B. Persson and G. Jonsson-Pettersson, Anal. Chim. Acta, 249 (1991) 43. G. Jonsson-Pettersson, Electroanalysis, 3 (1991) 741. E Cespedes, E. Martinez-Rbregas and S. Alegret, Anal. Chim. Acta, 284 (1993) 21. F. Ctspedes, E. Martinez-Fabregas and S. Alegret, Electroanalysis, 6 (1994) 759. F. Ctspedes, E. Martinez-Fabregas, J. Bartroli and S. Alegret, Anal. Chim. Acta, 273 (1993) 409. D. Martorell, F. Ctspedes, E. Martinez-Fabregas and S. Alegret, Anal Chim. Acta, 290 (1994) 343. F. Cespedes, F. Valero, E. Martinez-Fabregas and S. Alegret, Analyst, 120 (1995) 2255. S. Alegret, F. Cespedes, E. Martinez-Fabregas, D. Martorell, A. Morales, E. Centelles and J. Mutioz, Biosensors Bioelectronics, 11 (1996) 35. J. Wang and K. Varughese, Anal. Chem., 62 (1990) 318. J. Wang and M. Ozsoz, Electroanalysis, 2 (1990) 647. J. Wang, E. Gonzalez-Romero and M. Ozsoz, Electroanalysis, 4 (1992) 539. V.A. Bogodanauskaya, E. Khorozova and M.R. Tarasevich. Elektrokhimiya, 26, 5 (1990) 573. I.M. Kolthoff (Ed.), Quimica Analitica Cuantitativa, Omega Barcelona, Spain, 1980, p. 721. T. Moore, G. Gina Nam, L.C. Pipes and L.A. Coury, Jr.Anal. Chem., 66 (1994) 3158. D. Giorgio, in R.J. Henry, D.C. Cannon and J.W. Winkleman, Harper and Row (Eds.), Clinical Chemistry Principles and Practice, New York, 2nd edn., 1988, p. 1974.
