Letter pubs.acs.org/ac
Noiseless Performance of Prussian Blue Based (Bio)sensors through Power Generation Maria A. Komkova, Elena E. Karyakina, and Arkady A. Karyakin* Chemistry Faculty, M.V. Lomonosov Moscow State University, 119991, Moscow, Russia S Supporting Information *
ABSTRACT: In contrast to “self-powered” (bio)sensors aiming to generate maximum energy output, we propose the systems with the lowest potential difference between the working and the counter electrodes, which in galvanic mode would provide achievement of the best analytical performance characteristics. Prussian Blue based (bio)sensors known to operate at 0.00 V versus Ag|AgCl reference, in the short-circuit regime generate the current proportional to analyte concentration. Sensitivity and dynamic range of Prussian Blue based (bio)sensors in power generation mode are, respectively, even slightly higher and wider compared to the same (bio)sensors operated in the conventional three-electrode regime powered by a potentiostat. Selectivity of the (bio)sensors in power generation mode is similarly high relative to both oxygen, allowing H2O2 detection by its reduction, and reductants. Among the most important advantages of the proposed power generation mode is an order of magnitude decreased noise compared to performance in a conventional three-electrode setup powered by a potentiostat. Noiseless performances of Prussian Blue based (bio)sensors would open new horizons for electrochemical analysis. he concept of the so-called “self-powered” electrochemical (bio)sensors introduced at the threshold of the millennium (see ref 1 for a review) has become fashionable nowadays. Starting from ref 2, the “self-powered” (bio)sensors comprise the two electrode electrochemical cells operated in galvanic mode3,4 similarly to fuel cells. Hence, the “selfpowered” (bio)sensor concept is a replica of fuel cell5 or biofuel cell6 type sensors introduced 25 years earlier. Despite their attractive title, “self-powered” (bio)sensors are able to generate power sufficient for low-power electronics only close to the upper limit of their dynamic range (see ref 1 for a review). This makes such analytical devices hardly applicable in complete absence of an external power supply. Elaborating “self-powered” (bio)sensors, researchers aimed to gain maximum power output. Accordingly, the electrodes with large potential difference have been chosen.1,4,7−9 However, the realistic analytical application of “self-powered” (bio)sensors without an external power supply is amperometry in a short-circuit regime. This regime, considering a large potential difference between electrodes, forces operation of the working electrode far from its optimal potential, which dramatically affects the analytical performance characteristics. In the early 90s, we proposed Prussian Blue as a selective electrocatalyst for reduction of hydrogen peroxide (H2O2) allowing its low-potential detection in the presence of oxygen.10 The advantages of Prussian Blue modified electrodes over platinum, the material most widely used for detection of hydrogen peroxide, are (i) 3 orders of magnitude higher activity in H2O2 reduction and oxidation in terms of 1000 times higher
T
© XXXX American Chemical Society
electrochemical rate constants,11 which obviously provide dramatically improved sensitivity and (ii) 3 orders of magnitude higher selectivity in hydrogen peroxide reduction relative to oxygen reduction.12 These unique properties of the electrocatalyst allowed elaboration of sensors12,13 and biosensors14,15 with advantageous analytical characteristics. Among crucial characteristics of modified electrodes is their poor stability upon current pulses. Moreover, analytical performance characteristics of the corresponding sensors suffer from noises. Noise in a conventional three-electrode chronoamperometry powered by a potentiostat is generated by the operational amplifier enhancing noises from its inputs. The latter are fluctuations of both voltage generator and the potential difference between the working and reference electrodes, particularly upon stirring. To overcome these disadvantages we propose to exclude a potentiostat from the measurement system forcing thus operation of (bio)sensors in galvanic or power generation mode. In contrast to the reported “self-powered” biosensors aiming to generate maximum power output (above), we propose to use the system with the lowest potential difference between electrodes. The proposed system in short-circuit regime would obviously provide operation of the working electrode close to its optimal potential achieving the most Received: March 28, 2017 Accepted: May 25, 2017 Published: May 25, 2017 A
DOI: 10.1021/acs.analchem.7b01142 Anal. Chem. XXXX, XXX, XXX−XXX
Letter
Analytical Chemistry advantageous analytical performance characteristics. Here we report that power generation mode provides noiseless performance of Prussian Blue based (bio)sensors.
■
EXPERIMENTAL SECTION Materials. Experiments were carried out with Millipore Milli-Q water. All inorganic salts and hydrogen peroxide (30 vol %) were obtained at the highest purity from Reachim (Russia). Nafion solution, γ-aminopropyltriethoxysilane, Triton X-100, Dglucose, glucose oxidase (EC 1.1.3.4, 270 IU mg−1) from Aspergillus niger were purchased from Sigma-Aldrich. Sodium Llactate (40% solution) was from ICN. Lactate oxidase (EC 1.1.3.2, 72 IU mg−1) from Pediococcus sp. was from Sorachim (Switzerland). Planar sensors were made on the basis of three electrode screen printed structures with the 2 mm in diameter carbon working electrode, produced by Ltd. Rusens (Russia). Instrumentation. Voltammetry and amperometry were carried out using a PalmSens 3 potentiostat (PalmSens BV, The Netherlands). Current and voltage were measured with a digital multimeter Tektronix DMM4020 (Tektronix Inc.). Methods. Interfacial synthesis of Prussian Blue16 was made by dipping a droplet of 2−5 mM K3[Fe(CN)6] and 2−5 mM FeCl3 in 0.1 M HCl and 0.1 M KCl and initiated by addition of H2O2 to a final concentration of 50−200 mM. Stabilized films were synthesized via layer-by-layer deposition of the Prussian Blue and the stabilizing Ni hexacyanoferrate layers.17 Growing solution for the latter was 0.5−2 mM of NiCl2 and 0.1−1 mM K3[Fe(CN)6] in 0.5 M KCl containing 0.1 M HCl. After deposition, the modified electrodes were annealed at 100 °C for 1 h. Biosensors were made casting an enzyme containing drop (2 μL) onto the transducer surface with subsequent drying for 1 h. Glucose oxidase casting mixture was prepared suspending aqueous enzyme (10 mg/mL) by 0.3% Nafion analogue in 85% isopropanol.14 Lactate oxidase was suspended by 2% γaminopropyltriethoxysilane in 90% isopropanol.15 Glucose test strips were made forming a 2 μL capillary with the use of a double sided adhesive onto the surface of a glucose biosensor. As a standard method for glucose detection in blood, the electroanalytical system with a glucose biosensor14 validated previously for analysis of whole blood was used. The mean values and error bars in plots were calculated from not less than 3 measurements (n = 3) taking P = 0.95.
Figure 1. Steady-state current−voltage dependencies for Prussian Blue modified screen-printed electrodes in air-saturated 0.05 M phosphate with 0.1 M KCl in power generation mode: working electrode and Ag| AgCl reference connected through a resistance kit; hydrogen peroxide concentrations are 0 μM (○) and 10 μM H2O2 (▲) and 20 μM (Δ), 100 μM (■), and 200 μM (●) H2O2. Inset: semi-logarithmic plots.
Current−voltage dependencies in Figure 1 are fit to a simple exponential equation (i = C/(1+ B exp(A · U))) valid for polarographic waves (solid lines), where A−C are coefficients, and U is voltage (Figure 1). The evaluated half-wave potential almost coincides with the Prussian Blue/Prussian White redox potential found from cyclic voltammograms of the corresponding modified electrodes (118 mV, Ag|AgCl in 0.1 M KCl). Such a coincidence of the half-wave potential of H2O2 reduction with the Prussian Blue/Prussian White redox potential has been found previously in the three-electrode regime powered by a potentiostat.18 Of particular importance is sensor selectivity. Among the main advantages of Prussian Blue based electrocatalyst is the selective reduction of hydrogen peroxide in the presence of oxygen. In a certain potential range, optimal for sensor operation, the current of hydrogen peroxide reduction by Prussian Blue modified electrodes is 2 orders of magnitude higher than the current of oxygen reduction.18 Selectivity in power generation mode can be assessed from the current− voltage dependencies in semilogarithmic plots (Figure 1, inset). As seen, the difference between decimal logarithms of cathodic currents in air-saturated buffer and in the presence of similar concentration of H2O2 (0.2 mM) is always above 2. This means that also in power generation mode the selectivity of Prussain Blue modified electrodes toward hydrogen peroxide reduction relative to oxygen reduction is more than 100. As mentioned, the only realistic operation mode of (bio)sensors without an external power supply is amperometry in a short-circuit regime. Accordingly, by a simple short circuiting of the Prussian Blue modified electrode with the Ag| AgCl reference through an ammeter, we’ve expected to generate current proportional to hydrogen peroxide concentration. Short-circuiting obviously equilibrates electrode potentials. Since the reference electrodes are by definition characterized by much higher exchange current, the potential of the working electrode would be maintained close to the Ag| AgCl open circuit potential. This would force operation of the Prussian Blue modified electrode close to its optimal potential. For brevity we define “power generation” regime as galvanic mode with short-circuiting the working and the reference electrodes through an ammeter. As seen in Figure 2, in this regime the Prussian Blue based three-electrode hydrogen
■
RESULTS AND DISCUSSION As mentioned, to achieve the best analytical performance characteristics in galvanic mode, it is necessary to provide operation of the working electrode close to its optimal potential. Considering short-circuit regime, this means that the potential of the counter electrode should be as close as possible to the optimal potential of the working electrode. Prussian Blue based (bio)sensors operated around 0.00 V versus Ag|AgCl in 0.1−3 M KCl reference12 are the best candidates for such devices. Figure 1 displays the current−voltage dependencies of the three-electrode sensor structures with the working electrode modified with Prussian Blue. As seen, connection of the working electrode with the silver reference through a resistor (and an ammeter in series) indeed generates current output. In air-saturated buffer the only minor current is observed. As hydrogen peroxide concentration is increased, the raise in current generated by the sensor is observed. B
DOI: 10.1021/acs.analchem.7b01142 Anal. Chem. XXXX, XXX, XXX−XXX
Letter
Analytical Chemistry
(Figure S1, Supporting Information). Sensor response to paracetamol in both regimes is even an order of magnitude lower. We note that the potential of Ag|AgCl reference in 0.1 KCl is not enough cathodic to achieve the highest selectivity relative to reductants. Hence, Prussian Blue based sensors in power generation mode retain high selectivity relative to both oxygen, allowing detection of H2O2 by its reduction and reductants practically avoiding their influence on analysis. Both precision of measurements and lower limit of the calibration range are highly affected by noise. In the threeelectrode constant potential setup, the output noise is generated by the operational amplifier enhancing noise from its inputs. One source of such input noise is a voltage generator. Another source of input noise is due to measurement of the potential difference between the working and reference electrode, especially in case of solution turbulence (for instance, in batch cell upon stirring). Both sources of input noise mentioned above should disappear if a (bio)sensor is operated in power generation mode because of the absence of the corresponding operational amplifier. Hence, one can expect a significant decrease of noise. We note that for the conventional three-electrode setup the worldwide accepted low-noise potentiostat (PalmSens, The Netherlands) has been used. The frequency of measurements, which also affects the noise level, was the same (1 measurement per second) in power generation and three-electrode regimes. The noise of Prussian Blue based sensors in power generation mode is almost an order of magnitude lower compared to it in a conventional three-electrode regime (insets in Figure 2). Particularly for low hydrogen peroxide concentrations, most important for analytical applications, even in batch cell upon stirring the current after analyte injection remains nearly constant (Figure 2, insets). This allows a conclusion on noiseless performance of Prussian Blue based sensors in power generation mode. Since Ag|AgCl is used as reference electrode, performance characteristics in power generation mode could be affected by concentration of chloride ions, [Cl¯]. Figure 3 displays the dependence of sensor sensitivity on concentration of the latter. As seen, the sensitivity in power generation mode is decreased at lower chloride ion concentrations. The dependence has been successfully fit to the simple hyperbolic equation (solid line in Figure 3), allowing calculation of the half-wave chloride ion concentration of ≈1.2 mM. We note that according to the
Figure 2. Calibration graph of the Prussian Blue based three-electrode hydrogen peroxide planar sensor in a conventional three-electrode setup (○) and in power generation mode (□); air-saturated 0.05 M phosphate with 0.1 M KCl, pH 6.0, upon stirring. Insets: responses toward addition of H2O2 when powered by a potentiostat (1) and in power generation mode (2).
peroxide planar sensor generates current proportional to H2O2 concentration. The linear calibration range in double logarithmic plots is prolonged over almost 4 orders of magnitude of hydrogen peroxide concentration (from 2 × 10−7 to 1 × 10−3 M). For comparison, a calibration graph of the same Prussian Blue based sensor, but in a conventional threeelectrode setup powered by a potentiostat, is added to Figure 2. Its dynamic range is even narrower (from 5 × 10−7 M to 1 × 10−3 M). Sensor response in power generation mode is always slightly higher compared to the response in a conventional threeelectrode setup. Sensitivity determined as a slope in Figure 2 reaches the value of 0.65 A M−1 cm−2, which is also slightly higher than in the three-electrode system. Linearity in double logarithmic plots in an entire concentration range presumes that the current density response is always proportional to the concentration with only a minor free term. The linear relationship between the current density (in A cm−2) and concentration of H2O2 (in moles per liter) is given by the equation i = 0.65[H2O2]. We thus conclude that in power generation mode Prussian Blue based sensors display advantageous performance characteristics in terms of high sensitivity (0.65 A·M−1·cm−2) in batch cell and wide dynamic range (from 2 × 10−7 to 1 × 10−3 M), which are, respectively, even slightly higher and wider compared to those in a conventional three-electrode setup. Sensor selectivity in power generation mode relative oxygen reduction exceeds 2 orders of magnitude allowing detection of H2O2 by its reduction in the presence of oxygen. A possibility to detect H2O2 by its reduction using Prussian Blue modified electrodes obviously decreases sensor sensitivity to reductants known to dramatically affect performance characteristics of the noble metal based electrocatalysts.19 Among reductants the ascorbate is the most powerful one. Accordingly, selectivity of Prussian Blue based sensors relative to ascorbate has been investigated. As found, sensor response to micromolar concentrations of hydrogen peroxide can be masked only by 20-fold excess of ascorbate both in power generation mode and in a conventional three-electrode setup
Figure 3. Sensitivity of Prussian Blue based hydrogen peroxide sensor in power generation (short-circuit) mode as a function of chloride ion concentration; 0.05 M phosphate, pH 6.0, upon stirring. C
DOI: 10.1021/acs.analchem.7b01142 Anal. Chem. XXXX, XXX, XXX−XXX
Letter
Analytical Chemistry
a potential for low voltage read-out methods, for example, for printable devices.
Nernstian equation, the potential difference of Ag|AgCl electrode in 1.2 mM and 0.1 M Cl¯ is approximately 113 mV. This value with the precision of 5 mM coincides with the Prussian Blue|Prussian White redox potential relative to Ag| AgCl, 0.1 M KCl reference (above). Hence, the observed sensitivity decrease is due to the lost of reduction ability of the Prussian Blue|Prussian White couple. Another concern of the reference electrode in power generation mode is its stability in the course of monitoring of high hydrogen peroxide concentrations. This problem, however, can be solved by increasing surface area of the Ag| AgCl reference. Indeed, Prussian Blue based sensor equipped with 2 mm2 reference remains stable response for less than 10 min under 1 mM H2O2. An increase of the reference electrode surface area to 40 mm2 provides in similar conditions a stable response for more than 1 h (Figure S2, Supporting Information). Monitoring of high hydrogen peroxide levels requires stabilized sensors: Prussian Blue covered with nickel hexacyanoferrate.17 Glucose and lactate biosensors have been elaborated by immobilization of, respectively, glucose oxidase in perfluorosulfonated ionomer14 and lactate oxidase in alkoxysilane gel15 membranes onto the surface of Prussian Blue modified working electrodes. Performance characteristics of the biosensors in power generation (short-circuited) mode are also similar to them in conventional three-electrode setup (Figures S3 and S4, Supporting Information). For the glucose biosensor, the dynamic range is prolonged from 5 × 10−6 to 2 × 10−2 M. Lactate detection with the use of the corresponding biosensor is possible from 5 × 10−7 to 2 × 10−3 M. Sensitivities in power generation mode are slightly higher than in a conventional three-electrode setup. The glucose biosensor displays sensitivity of 0.043 ± 0.004 A M−1 cm−2 in power generation mode. Sensitivity of lactate biosensor in similar conditions encounters 0.18 ± 0.04 A M−1 cm−2. Both glucose and lactate biosensors in power generation mode display an order of magnitude reduced noise (Figures S3 and S4, Supporting Information) like the hydrogen peroxide transducer (above). Glucose test strips on the basis of Prussian Blue glucose biosensors have been operated in power generation mode with coulometric detection. For more than 20 different blood samples, the Pearson correlation coefficient with the standard method has reached the value of 0.82 for 1 min of analysis with test strips and the value of 0.96 for 20 min of analysis (Figure S5, Supporting Information). This clearly shows the validity of Prussian Blue based biosensors in power generation mode for analysis of real samples, such as complex ones like whole undiluted blood.
■
ASSOCIATED CONTENT
S Supporting Information *
The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.analchem.7b01142. Responses of Prussian Blue-based sensor to ascorbate and H2O2 in a conventional and power generation modes; monitoring of high H2O2 level with sensors having different Ag|AgCl reference electrode areas; calibration graphs and responses of Prussian Blue-based glucose and lactate biosensors in a conventional and power generation modes; and validation of glucose biosensor for analysis of whole blood (PDF)
■
AUTHOR INFORMATION
Corresponding Author
*E-mail:
[email protected]. ORCID
Arkady A. Karyakin: 0000-0002-0457-7638 Notes
The authors declare no competing financial interest.
■ ■
ACKNOWLEDGMENTS Financial support through Russian Science Foundation Grant No. 16-13-00010 is greatly acknowledged. REFERENCES
(1) Arechederra, R. L.; Minteer, S. D. Anal. Bioanal. Chem. 2011, 400, 1605−1611. (2) Katz, E.; Buckmann, A. F.; Willner, I. J. Am. Chem. Soc. 2001, 123, 10752−10753. (3) Liu, Z.; Cho, B.; Ouyang, T.; Feldman, B. Anal. Chem. 2012, 84, 3403−3409. (4) Zhang, L.; Zhou, M.; Dong, S. Anal. Chem. 2012, 84, 10345− 10349. (5) Huck, H. Fresenius' Z. Anal. Chem. 1974, 270, 266−273. (6) Mele, M.; Cardos, M. J.; Videla, H. A. An. Asoc. Quim. Argent. 1979, 67, 125−138. (7) Sekretaryova, A. N.; Beni, V.; Eriksson, M.; Karyakin, A. A.; Turner, A. P. F.; Vagin, M. Y. Anal. Chem. 2014, 86, 9540−9547. (8) Hickey, D. P.; Reid, R. C.; Milton, R. D.; Minteer, S. D. Biosens. Bioelectron. 2016, 77, 26−31. (9) Jeerapan, I.; Sempionatto, J. R.; Pavinatto, A.; You, J. M.; Wang, J. J. Mater. Chem. A 2016, 4, 18342−18353. (10) Karyakin, A. A.; Gitelmacher, O. V.; Karyakina, E. E. Anal. Chem. 1995, 67, 2419−2423. (11) Karyakin, A. A.; Karyakina, E. E.; Gorton, L. J. Electroanal. Chem. 1998, 456, 97−104. (12) Karyakin, A. A. Electroanalysis 2001, 13, 813−819. (13) Karyakin, A. A.; Puganova, E. A.; Bolshakov, I. A.; Karyakina, E. E. Angew. Chem., Int. Ed. 2007, 46, 7678−7680. (14) Karyakin, A. A.; Kotel’nikova, E. A.; Lukachova, L. V.; Karyakina, E. E.; Wang, J. Anal. Chem. 2002, 74, 1597−1603. (15) Yashina, E. I.; Borisova, A. V.; Karyakina, E. E.; Shchegolikhina, O. I.; Vagin, M. Y.; Sakharov, D. A.; Tonevitsky, A. G.; Karyakin, A. A. Anal. Chem. 2010, 82, 1601−1604. (16) Borisova, A. V.; Karyakina, E. E.; Cosnier, S.; Karyakin, A. A. Electroanalysis 2009, 21, 409−414. (17) Karpova, E.; Karyakina, E. E.; Karyakin, A. A. RSC Adv. 2016, 6, 103328−103331. (18) Karyakin, A. A.; Karyakina, E. E.; Gorton, L. Electrochem. Commun. 1999, 1, 78−82.
■
CONCLUSION In the present report, we are discussing an advantage of Prussian Blue modified electrodes, which have not been taken into account before. It is a possibility for operation at 0.00 V relative to Ag|AgCl electrode, most often used for elaboration of electrochemical (bio)sensors. First, it may simplify elaboration of the controlling potentiostat, because there is no need for a voltage generator: the corresponding inputs of the key operational amplifier could be short-circuited. Second, as has been clearly shown in the article, Prussian Blue based (bio)sensors successfully operate without a potentiostat by a simple short-circuiting of the working and the reference electrodes. The achieved noiseless performances of Prussian Blue based (bio)sensors in power generation mode would have D
DOI: 10.1021/acs.analchem.7b01142 Anal. Chem. XXXX, XXX, XXX−XXX
Letter
Analytical Chemistry (19) Scheller, F. W.; Pfeifer, D.; Schubert, F.; Reneberg, R.; Kirstein, D. In Biosensors: Fundamental and Applications; Turner, A. P. F., Karube, I., Wilson, J. S., Eds.; Oxford University Press: Oxford, U.K., 1987.
E
DOI: 10.1021/acs.analchem.7b01142 Anal. Chem. XXXX, XXX, XXX−XXX
