Insight into the Electro-Oxidation Mechanism of Glucose and Other

Subscriber access provided by UNIVERSITY OF TOLEDO LIBRARIES

Article

A NEW INSIGHT INTO THE ELECTRO-OXIDATION MECHANISM OF GLUCOSE AND OTHER CARBOHYDRATES BY CuO-BASED ELECTRODES Jose Tiago Barragan, Sergio Kogikoski Jr, Everson T. S. G. da Silva, and Lauro T. Kubota Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.7b04963 • Publication Date (Web): 09 Feb 2018 Downloaded from http://pubs.acs.org on February 10, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Analytical Chemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 9 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

A NEW INSIGHT INTO THE ELECTRO-OXIDATION MECHANISM OF GLUCOSE AND OTHER CARBOHYDRATES BY CuO-BASED ELECTRODES. José T. C. Barragan,† Sergio Kogikoski Jr.,† Everson T. S. G. da Silva,† and Lauro T. Kubota.*,† † Department of Analytical Chemistry, Institute of Chemistry, State University of Campinas (UNICAMP), P.O. Box 6154, Zip code 13083970, Campinas-SP (Brazil).* Tel: (+55) 19 35213127. E-mail: [email protected]

ABSTRACT: In this work, a new hypothesis for the electrocatalytic behavior of CuO electrodes is presented. Different from the established mechanism, here we discuss why CuIII species do not participate in the oxidation mechanism of carbohydrates. We show that hydroxyl ion adsorption and the semi-conductive properties of the material play a more significant role in this process. The relationship between the flat band potential and the potential that begin oxidation suggests that the concentration of vacancies in the charge region acts upon the reactivity of the adsorbed hydroxyl ions through a partial charge transfer reaction. In the presence of carbohydrate molecules, the electron transfer is facilitated and involves the transfer of electrons from the adsorbed hydroxyl ions to the CuO film. This mechanism is fundamentally relevant since it helps the understanding of several experimental misleads. The results can also lead to obtaining better catalysts, since improvements in the material should focus on enhancing the semi-conductive properties rather than the CuII/CuIII redox transition. The results shed light on different aspects of carbohydrate molecules oxidation that could lead to novel applications and possibly a better description of other semiconductor mechanisms in electrocatalysis. KEYWORDS: copper oxide, copper oxyhydroxide, CuOOH, semiconductor, electrocatalytic oxidation.

INTRODUCTION Copper has on several occasions been present in advances that contributed to the development of human cultures1. Although the use of copper oxides began thousands of years ago, they are still involved in new technologically important advances such as hydrogen photogeneration,2 CO2 photoreduction,3 supercapacitors,4,5 high temperature superconductors,6 and new microelectronic elements called memristors.7 Some properties of CuO are well known, it has a monoclinic structure and behaves like a ptype semiconductor.8,9 The high availability of the raw material, easy preparation, and the possibility of different morphologies are also important advantages.10 Electrodes based on CuO have promising electrocatalytic properties, principally for the sensitive detection of carbohydrates.11 Recently, intense research has focused on the use of CuO-based electrodes for non-enzymatic glucose sensing.12-26 However, the development of better sensors depends on adequate understanding of the chemical and electrochemical processes involved in the detection. By now, many CuIII oxyhydroxides have been reported to participate in glucose oxidation mechanism.27-29 In these cases, the oxidations may occur through an electrocatalytic mechanism, where initially the CuII is oxidized to the CuIII species. In a second step, the CuIII that reacts with the target molecule, regenerating the CuII and leading to the oxidized products.28 The formation of CuIII species during the potential sweep was pointed out in 1969 by Miller, due to the appearance

of a pair of peaks in potential regions near the oxygen evolution.30 Later, the probable formation of CuIII was reported, relating the formation of a yellowish transient coloration that rapidly changed to blue, making it difficult to characterize by spectroscopy.31 The presence of the CuIII species in the oxidation mechanism of carbohydrates was first pointed out in 1992 by Marioli and Kuwana, as an attempt to explain ring-disk electrode results.27 After that, the participation of CuIII is part of the interpretation for the results of copper-based sensors.12-26 In contrast, few authors have realized that the hypothesis of CuIII species participation appears to be inconsistent with the results.32 A similar finding has been observed when analyzing glucose sensor results. In this case, the glucose oxidation peak occurs in more negative potentials than that attributed to the formation of CuIII species.33 Other information that also corroborates these controversies is that there are few discrete molecules in which the existence of CuIII is accepted, and in these cases, the ligands play a fundamental role in their stabilization.34, 35 Electron affinity and ionization energy data show that the energy required to remove 1 electron from O(g)2- (+844 KJ mol-1) is smaller than the energy required to remove 1 electron from Cu(g)2+ (+3554.6 KJ mol-1).36 CuIII species are strong oxidant and show an equilibrium potential more positive than the water stability limit, this led some authors to remove CuIII species from Pourbaix diagrams. 37 However, these thermodynamic notes are no longer important if there are kinetic limitations in reactions. This

ACS Paragon Plus Environment

Analytical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

reinforces the necessity to further investigate interfacial processes of this material. Over the years, our research group has studied the properties of Ni and Cu electrodes as an alternative to Au electrodes for the detection of carbohydrates,11,38 as well as new detection approaches based on the properties of charge accumulation on their surfaces.39 In a search for a better understanding of the processes involved, in this work, we show that the CuIII species do not occur in the used experimental conditions and also do not participate in the oxidation processes of carbohydrate molecules in general. We show that the electrocatalytic behavior is attributed to the semi-conductive properties of the CuO film and the adsorption of hydroxyl ions. We understand that this new approach has shed light on the understanding of the electrocatalytic behavior and the closure of the inconsistencies previously found that could lead to novel applications of semi-conductive metal oxides. EXPERIMENTAL SECTION Chemicals. D-Glucose (Sigma), sodium sulfate (Merck), ammonium persulfate (Sigma) and sodium hydroxide (Sigma) were used without further purification. All solutions were prepared with deionized water from a DirectQ®3UV Millipore system with resistivity > 18.2 MΩ⋅cm and subjected to ultraviolet radiation to prevent the growth of microorganisms and organic compounds. Instruments. A PGSTAT 12 potentiostat with high input impedance of electrometer (> 100 GΩ; < 8 pF) and containing a Frequency Response Analyzer-FRA module (Metrohm-AUTOLAB) was used for all electrochemical measurements. The potentiostat was connected to a Dell Latitude computer and controlled by NOVA 1.11 software. An electrochemical cell with three electrodes was used for cyclic voltammetry (CV), linear sweep voltammetry (LSV), zero-current potential (ZCP) and Mott-Schottky (MS) assays. In these tests, a Cu electrode with 3 mm diameter, a Pt grid with 20 mm diameter and an Ag|AgCl|KCl 3 mol L-1 electrode were used as the working electrode (WE), the counter-electrode (CE) and the reference electrode (RE), respectively. This electrochemical system was used in all experiments except for Raman spectral measurements. The surface morphology of the electrodes was characterized using a Scanning Electron Microscope (SEM) JEOL JSM 6360LV, operated at 10 kV. Characterization by X-ray diffraction (XRD) was performed using a Shimadzu XRD 7000 with λ = 1.5406 Å referring to the Kα copper emission line, 40 mA and 40 kV. Diffractograms were obtained between 10°< 2θ< 100°, at 2° min-1. Diffuse reflectance characterization (DR) was performed using an Ocean Optics HR2000 UV-VIS-NIR Spectrophotometer. Curves that were obtained were converted into absorption spectra using a polytetrafluoroethylene standard. The Kubelka-Munk (KM) model was employed to determine the materials band-gap. Raman spectra (RS) were obtained by using a confocal Raman T64000 from Horiba Scientific, a 632.8 nm helium–neon laser of 1.25 mW and a 50x objective mounted on an Olympus optical

microscope. The electrochemical cell used for the in situ measurements was similar to the cell described previously by our group, but with platinum wire as the counter electrode40. A 520 nm indium-galium diode laser (5 mW) was used for the controlled light incidence experiments. Solutions. NaOH stock solution (0.1 mol L-1) was prepared in a 2 L volumetric flask and filtered using a 0.22 μm membrane and a vacuum system (Quimis) before its use as an electrolyte. To minimize the presence of dissolved gas, the solution was transferred to a Schott bottle and subjected to a vacuum treatment in an ultrasound for 60 minutes. Glucose stock solution was prepared and used fresh. For the preparation of copper|cupric hydroxide (Cu|Cu(OH)2) and copper|cupric oxide (Cu|CuO) electrodes, an alkaline and an oxidizing solution of NaOH (2.50 mol L-1) and (NH4)2S2O8 (0.125 mol L-1), respectively, were prepared using deionized water. These solutions were also used on the same day they were prepared. Electrodes. Cu|CuO electrodes were prepared as previously described by our group11. Briefly, Cu electrodes (> 99%) with a 3-mm diameter were mechanically pretreated using sandpaper with granulometry varying from 300 to 1200. Then, they were polished with alumina slurry, followed by washing with deionized water and ultrasound for 10 minutes to remove adsorbed alumina. Next, they were subjected to a chemical oxidation by immersion in NaOH 2.5 mol L-1 and (NH4)2S2O8 0.125 mol L-1 for 10 min. These electrodes were then rinsed, dried and subjected to a thermal treatment at 150 °C for 60 minutes.

RESULTS AND DISCUSSION Electrode characterization. The electrochemical behavior of Cu in alkaline media is quite complex, since a variety of species of different nature can be formed.41,42 For this reason, electrochemical studies were carried out on the electrodes that were modified with CuO. As predicted by Pourbaix diagram (Figure S1) this species is more stable under the experimental conditions. Figure 1A shows the Cu electrode before modification and its typical reddish brightness. After the first chemical treatment, its color changed to blue (Figure 1B) and then to brown (Figure 1C). The electrode surface was evaluated at each stage by SEM and EDX (Figure S2), which showed nanostructured CuO after the modifications, as previously reported.11 The Figure 1D shows the diffractograms obtained for Cu powder, which presented a characteristic crystalline lattice of metallic Cu (JCPDS 04-0836). After the chemical treatment, the diffractogram (Figure 2E) showed the diffraction pattern of the Cu(OH)2 (JCPDS 04-0836). And after the thermal treatment (Figure 2F), the material exhibited diffraction in accordance with the JCPDS 48-1548 and JCPDS 44- 0706, characteristic of crystalline CuO. Electrochemical characterization was carried out in alkaline solution, in the absence and presence of glucose, and the third cycle of the CV was recorded. The results for the Cu, Cu(OH)2 and CuO electrode are shown in Figures 1G, 1H and 1I, respectively. In the absence of glucose, the typical redox transition peaks of this analyte were not


Page 2 of 9

Page 3 of 9 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry observed for all three electrodes. This observation is even more evident for the Cu electrode, which is quite different from the Ni electrodes. The Ni electrode shows a clear peak pair, referring to the NiII(OH)2/ NiIIIOOH transition and the glucose electro-oxidation mechanism (see Figure S3).38 Instead of peaks, however, it is possible to observe a current plateau in all three studied cases, with anodic and cathodic current values increasing in the following order: Cu < Cu(OH)2 < CuO. In these CVs, an exponential increase in current at potentials higher than 0.6 V vs Ag|AgCl is also observed, which is probably related to the oxidation of OH- on the electrode surface, forming O2.36

In the CVs shown in Figure 1 and also for other carbohydrates (Figure S4), it is evident that glucose oxidation occurs above the current plateau observed in the absence of glucose. This argument leads to the hypothesis that oxidant species are formed in this region, and up to date, they were assumed to be CuIII. Some representations most commonly found in the literature for CuIII species are: Cu3+, Cu(OH)3, CuO2-, Cu2O3, Cu(OH)4 and CuOOH. 37,43 Although CuOOH is frequently used to explain the experimental results, 12-26 we do not find the equilibrium potential of this reaction. In addition, as seen (Figure 1) no voltammetric peaks, that would allow calculating the formal potential, are observed. However for CuIII species in which the equilibrium potential is obtained, as Cu2O3 (1), the potentials are higher than the water stability limit (2): 43 2 → 2 2 (1)

1.648 0.0591

4 → 2 4

1.229 0.0591 (2)

(2)

III

Figure 1. Photographs: (A) Cu, (B) Cu(OH)2 and (C) CuO electrodes. XRD obtained for the: (D) Cu; (E) Cu(OH)2 and (F) CuO materials. CVs: (G) Cu, (H) Cu(OH)2 and (I) CuO. All CVs were obtained in 0.1 M NaOH at 10 mV s-1 in the absence and presence of 1 g L-1 glucose. In the absence of glucose, the film current increase in the following order Cu < Cu(OH)2 < CuO is observed. At 0.7 V, the current increases in the same order, Cu < Cu(OH)2 < CuO, suggesting that the CuO film presents better electrocatalytic properties for O2 formation. CVs obtained in the presence of glucose showed even more interesting results. In all cases, at least one oxidation peak of glucose is observed, and both peak current and the initial current increase are shifted to less anodic values in the following order: Cu < Cu(OH)2 < CuO. The thicker CuO film was expected to act as an ohmic barrier, shifting the glucose oxidation peak to more anodic potentials. However, an inverse behavior was observed, with a difference in peak potential between the Cu and CuO electrodes of 0.22 V. This difference is much higher than expected if the process were mediated by the CuIII species. These observations are intriguing and lead to the suspicion that the redox CuII/CuIII couple is not involved in this electrocatalytic reaction. Although some researchers have already mentioned these suspicions,32,33 the mechanism of glucose oxidation mediated by the CuIII species is considered in almost all works,12-26 and it is considered an established consensus.

This is the reason why these Cu species are considered strong oxidant agents and unstable in aqueous solutions, 37 . Recently, CuIII species was detected by Raman in very anodic potentials where O2 formation occurs (+1.62 V vs NHE). 44 These results seem to be in agreement with the predictions of the Pourbaix diagram (Figure S1) and Equation (1), which indicates that CuIII are formed in potentials higher than the water stability. On the other hand, although the thermodynamic aspects bring some note, any statement cannot be conclusive without knowing the real processes that occur at the interface of the electrode especially if there are kinetic limitations involving CuIII species. Due to these factors, the following studies were dedicated to understanding the processes involved in the current plateau region, at first in the absence of glucose. Current plateau studies for Cu|CuO. To understand the Cu|CuO electrode behavior, the need to understand the current plateau presented in the absence of electroactive molecules is evident. Some preliminary ZCP studies in Na2SO4 0.33 mol L-1 (Figure S5) shows that the OH- readily interacts with the electrode surface, and this can be described by equation (3): H O!"# OH ⇌ OH!"# H O

(3)

This equation also shows that a reasonable amount of OH- is readily adsorbed on the electrode surface, even before the polarization by an external potential. EQCM assays (Figure S6) shown that the frequency variation obtained experimentally is only 34.6% of the variation calculated by CV charge. This suggests that if OH- ions are the species that undergo the charge transfer a large part of these species must readily adsorbed on electrode surface. This preliminary adsorption of OH- also appears to be in agreement with found in other semi-conductive materials such as TiO2 (anatase).45



CVs at different scan rates were obtained under two different sets of conditions: increasing the scan rate in the anodic direction and maintaining the cathodic scan rate at 20 mV s-1 (Figure 2A); and maintaining the scan rate in the anodic direction at 20 mV s-1, followed by increases in the cathodic scan rate (Figure 2B). In these CVs, although the resulting currents are dependent on the scan rate, the charge is approximately constant and depends on the applied potential, reaching a value of 0.89 mC at 0.6 V (Figure S7), which agrees with an electro-adsorptive process. This result also indicates the existence of a pseudocapacitive behavior, already reported for this material.5

Figure 2. Study of the electrochemical behavior of the Cu|CuO: (A) CVs at different scan rates for the anodic scan and maintaining the cathode scan at 0.020 V s-1; (B) CVs with scan rate in the anode direction at 0.020 V s-1 followed by different scan rates in the cathode direction; (C) pH influence at 0.01 V s-1 (D) Temperature influence at 0.1 V s-1. Voltammograms obtained during 220 cycles (Figure S8A) and at different rotation rate of electrode (Figure S8B) did not show significant changes in the voltammetric profile. In the first case (Figure S8A), it suggests that the process does not provide significant changes in the surface due to a phase transition or dissolution. However it was reasonable to expect that a transition in the oxidation state of the CuII/CuIII into oxide would be accompanied by a change in the structural characteristics of the film and, thus in the voltammetric profile, such as those observed for the CuI/CuII (cubic to monoclinic system).8 In the second case (Figure S8B) the results indicate that current is not limited by the diffusion of species. Other parameters showed greater changes in the voltammetric profile. As shown in Figure 2C, the increase in the pH of the Na2SO4 0.33 mol L-1 solution promotes a CV shifted to the cathodic direction. In this case, in accord to Equation 2, O2 formation is favored and consequently occurs at a less anodic potential. However, the current plateau does not exhibit significant changes, suggesting that the concentration of OH- in this pH range is high enough and does not limit the current. Another parameter that promotes greater changes in the CV profile is the temperature. Figure 2D shows CVs obtained at three different temperatures: 0 oC, 40 oC and 80

o

C. In this case it is possible to see that the temperature promotes a more significant change and this suggest that the current plateau is governed mainly by the semiconductor properties of the film.

Another behavior noticed in this system was its ability to remain polarized after an applied potential. Figure 3A shows ZCP results after the electrode polarization for 60 s at different potentials. It is possible to observe that with an increase in the applied potential, the time that the electrode remains charged also increases (≈ 8 hours for 0.6 V). These results do not show a connection with the transient nature reported for the CuIII.31 Figure 3B shows the ZCP after electrode polarization for 60 s at different potentials, but with additional steps “a” at 30 s and “b” at 60 s. In "a", the potential register was paused, the electrode was removed from the cell, rinsed with deionized water, reinserted into the cell, and the potential register was restarted. In “b”, the potential register was paused, the electrode was removed from the cell, inserted into a glucose solution, rinsed with water, reinserted into the cell, and the potential register was restarted. The obtained results show that in "a", the discharge of the electrode is considered negligible. However, in "b", the discharge of the electrode occurs in all charging potentials evaluated, except for -0.2 V. Similar results were reported recently by us39 showing that electrode charging and glucose oxidation can be performed in individual steps. The existence of an intermediate, which is formed even in the absence of glucose, does not agree with the early suggestion of the formation of a glucose-CuIII complex that leads to subsequent glucose oxidation.27 To obtain more information about the CuIII species, in situ Raman was performed. Figure 3C shows the results obtained at different applied potentials (between -0.2 V and 0.6 V), in the same conditions used for glucose oxidation. In these cases, it is possible to see the presence of a Raman peak at 298 cm-1 and others with lower intensity at 347 cm-1 and 591 cm-1, attributed to the vibration of the CuO network. These peaks are observed only at potentials higher than 0 V vs Ag|AgCl. The disappearance of the Raman peaks in the spectra at 0 V and -0.2 V suggests that there is a surface conversion from CuO to metallic Cu in the region where the laser is focused. Although there is no information in the literature that explains this behavior, we suggest that it is attributed to the cathodic photocurrent generated by the p-type semiconductor CuO. Moreover, in all cases, an acute peak at 603 cm-1 related to the stretching vibration of the typical Cu-O bond in CuIII is not observed. These results are in agreement with those found in the literature for O2 formation, showing that the detection of CuIII occurs only at more positive potentials (1.62 V vs NHE).44 We believe that the above results lead to an undeniable confirmation that the CuIII species is absent in this case and does not participate in the oxidation mechanism of carbohydrates. Raman spectra obtained in the absence and presence of glucose 1 g L-1 (Figure 3D) with applied potential at 0.4 V also does not show CuIII peaks, suggesting that it do not appear under the conditions in which glucose is oxidized.


Page 4 of 9

Page 5 of 9 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry potentials smaller than 0 V (negative potentials), there is an electron-vacancy separation, with a higher concentration of electrons on the surface of the electrode, explaining the appearance of a cathodic photocurrent in the CV (Figure 3F). So, in the absence of species with a higher tendency of reduction, the excessive electrons can cause the reduction of CuO to Cu, explaining the Raman results obtained at 0 and -0.2 V (Figure 3C), where the laser caused the reduction of CuO to Cu, leading to the peakless spectra obtained. The disturbance caused by light is an argument for the existence of species with semiconductor properties. For the Cu electrode, the light also has influence when the electrode potential reaches the CuII formation region (Figure S9), allowing us to state that even the recently formed CuO or hydrated CuO has also semi-conductive properties.

Figure 3. Characteristics of the Cu|CuO in alkaline media: (A) ZCP profile measurements after different applied E per 60 s. (B) ZCP obtained after different potentials applied followed by removal of the electrode from the cell and washing -1 with water (“a”) and in (“b”) immersion in glucose 1 g L and washing. (C) Raman spectra at different applied potentials without glucose. (D) Raman spectra obtained at 0.4 V vs Ag|AgCl in the presence of glucose. (E) ZCP before, during and after the radiation incidence (λ = 520 nm). (F) CVs experiments before and during the incidence of the laser.

As observed in the Raman spectra, the CuO electrode is strongly influenced by light (λ=632.8 nm), suggesting again a semiconductor behavior. Although the semiconductor behavior of copper oxide species has been known for decades, little attention is given to this characteristic when it comes to electrocatalytic oxidations. To improve the understanding of this behavior, a green coherent light beam (λ= 520 nm) was focused on the CuO electrode during the electrochemical experiment. Figure 3E shows the disturbance caused in the ZCP profile, which demonstrates that with the incidence of light, the potential of the electrode increases by ~0.080 V in the anodic direction. After the irradiation ceased, a potential decrease was observed for ten minutes, until it approached the initial value of -0.080 V vs Ag|AgCl. In this experiment, it is evident that the light affects the equilibrium potential of the system, due to its semiconducting properties. Figure 3F shows the CVs obtained at 10 mV s-1 before and after the illumination of the electrode with the green laser. The laser promoted perturbations in the current profile, mainly in the cathodic potentials (E < 0 V), which agrees with the Raman results. In potentials more positive than 0.05 V, no noticeable change in the CV was observed. This result agrees with the characteristics of a ptype semiconductor46 for light with energy equal to or higher than the material band-gap (hν > Eg). In applied

CuO semiconductor behavior. Due to the semiconductor character of the electrodes, complementary characterizations were also performed. Thus, 3 electrodes, obtained the same way as described here, were characterized by UV-Vis to obtain the band-gap energy.47 The results, which are shown in Figure 4A, provided a value of Eg = 1.482 eV for the CuO electrode. This value is in agreement with the expected for CuO (1.3 eV to 2.1 eV).8 48 The Mott-Schottky method was also employed to obtain more detailed information, such as the flat band potential (Efb) and density carriers (see SI Equation 1). Figure 4B show Mott-Schottky plot for the CuO electrode. The negative slope confirms that the majority charge carriers are vacancies represented by h+, and thus, the electrode presents p-type semiconductive properties. The density of charge carriers obtained was 5.09 1021 cm-3. This is similar to the 5 1021 cm-3 found for CuO thin film48 and higher than obtained using chronoamperometry (2.41 1018 cm-3).49 In addition Efb, was also obtained (-0.048 V vs Ag|AgCl). This value seems to be consistent with the observed initial cathodic photocurrent in the CV experiments (Figure 3F) and Raman spectroscopy data (Figure 3C). With the information obtained from the Mott-Schottky method along with the reflectance results, and considering that 0 V vs SHE is equal to the Fermi level energy in vacuum (4.5 eV),46 it was possible to construct the energy diagram for the CuO electrode as shown in Figure 4C. We consider that the semiconductor characteristics of the Cu-based electrodes are of great importance for the interpretation of the carbohydrate oxidation processes. For a p-type semiconductor, an applied potential higher than the flat band potential (E > Efb) leads to the accumulation of vacancies in the interface region, and for (E < Efb) represents a depletion. Thus, it is reasonable to suggest that the accumulation of vacancies at the interface of the electrode is mainly responsible for modulating the glucose and others carbohydrate oxidation reactions (Figure S4). LSV at a slow scan rate (1 mV s-1) was obtained to avoid influences of mass transport and prove the importance of semiconductor properties on the oxidation of glucose. Figure 4D shows the LSV curves obtained for the CuO electrode in the presence and absence of glucose. The



potential where the glucose oxidation start coincides with the Efb. The same correlation was observed for the Cu after electro-formation of CuII thin film (Figure S10). In this case, the Efb found was higher than that obtained for the CuO electrode (Efb 0.193 V vs Ag|AgCl) and the carrier density (1.32 1021 cm-3) was 4 times smaller. The difference in the flat-band potential (0.24 V) apparently also shows an agreement with the displacement of the glucose peak potentials on the CVs in Figure 1 (0.22 V).

Page 6 of 9

the I-E profile in the same region of glucose electrooxidation is not expected. Thus, the behavior of the CuO film was also studied in the absence of electrolyte, replacing it for a copper metal conductor (Cu|CuO|Cu). Instead of a linear ohmic-type behavior, an exponential semiconductor I-E LSV profile was observed for the Cu|CuO|Cu system (Figure 5A). For this reason, it is trivial to expect that the CuO electrode does not behave like a conductor. As far as we know, the understanding and interpretation of the electrocatalytic processes involved in carbohydrate oxidation through the semiconducting properties have not been reported to date. However, contributions to the understanding of the solid system based on copper oxides have been reported by Brattain.51 In this case, experimental curves can be adjusted by an equivalent to the thermionic emission equation (4): ()

I I& I& eη*+

(4)

where e is the electron elementary charge (1.60 10-19 C), E is the applied potential (V), k is the Boltzmann constant (1.38 10-23 J K-1), η is the ideality factor, T is the temperature (K), and I0 represents the carrier current at equilibrium (A). Figure 4. Semiconductor characterization of the Cu|CuO electrode. (A) the Kubelka-Munk plot obtained by UV-Vis. (B) Mott-Schottky plot obtained at 100 Hz. (C) Energy dia-1 gram for the CuO electrode. (D) LSV obtained in 1 mV s in -1 the absence and in the presence of glucose 1 g L .

We consider that this is a consistent indication that the accumulation of vacancies (h+) at the electrode interface or, more broadly, the semiconductor properties of the CuO-based electrodes is closely linked to the carbohydrates oxidation. Semiconducting properties also help us to understand why CuIII species do not effectively participate in the glucose oxidation mechanism. The conduction band of transition metal oxides mainly consisted of d orbitals or d and p hybridized orbitals of the transition metal atoms, whereas the valence band in these oxides is composed mainly of O 2p orbitals.50 This affirmation agrees with this new hypothesis, because of the O 2p orbitals of the semiconductor CuO being responsible for the valence band, it becomes logical to think that the oxygen atoms of the hydroxyl ions adsorbed on the surface are more likely to transfer electrons, showing an agreement with the electro-affinity, and explaining why there is no experimental evidence on the effective participation of CuIII species in the mechanism. Improvements in the semiconductor properties also seem to favor the reaction of O2 formation (Figure 1). In addition measures performed in flow,11 show a decrease in the oxidation current of glucose at the region that the formation of O2 increases, suggesting a probable competition for electro-adsorbed hydroxyls. Current-potential profile. A mechanism considering the semiconductor character of copper oxide-based electrodes appears to be more suitable for interpretation of the results. However, if a CuII/CuIII redox transition does not occur under the conditions studied, the appearance of

Figure 5A shows that Equation 4 offers good precision and fits the results very well. The results showed that I0 = 2.64·10-7 A and e/ηkT = 4.48 CJ-1. The exponential profile observed Cu|CuO|Cu system is very similar to the anodic profile observed in the solution, and it is expected that this behavior is correlated to the semiconductor behavior and not to a redox transition. Figure 5B shows the I-E profile obtained in potentials higher than the Efb for the Cu|CuO|NaOH(aq) system. Here, we also observe the presence of an exponential behavior, which indicates that the current observed in the I-E profile for this system, is due mostly to the conduction process through the semiconductor material. However, in potentials above 0.3 V, a linear behavior becomes more accentuated. Despite the greater complexity of this system, which involves adsorption of species, it is expected that at higher applied potentials, the resulting current should be limited by the flow of species in the electrolyte, and so predicted by the Nernst-Planck equation46 (see SI Equation 2). Despite the complexity involved in solving this equation, some simplifications can be considered, especially if we consider that the experiments were carried out with a stationary electrode. So, it is expected that the charge through the electrolyte is carried mostly by the migration of the ionic species. In this way, considering E’ = E-Efb and the existence of a uniform electric field between the auxiliary electrode and the working electrode separated by a distance x, the Equation (5) below can be described: (),

I I& I& eη*+

-. /01 234

∑ 6 7

(5)

Figure 5 shows that Equation 5 fits precisely to the experimental results, showing that I0 = 2.25·10-3 A; e/ηkT = 1.04 CJ-1; F2AΣz2DC/RTx = 2.37·10-3 AV-1. Despite all approximations, Equation 5 provides the basis for a more encompassing equation to be developed in future. In addition, it seems to be the first time that attempts have been made


Page 7 of 9 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry to explain the LSV for CuO electrodes, and the findings shown here appear to be more suitable when compared to the attempt involving a CuII/CuIII.

reactions. Thus, carbohydrate molecules can react with these surfaces, oxidizing by the abstraction of a labile hydrogen atom, similar to that previously suggested for carbohydrate molecules27 and methanol:54

>?@ A RC ;

Insight into the Electro-Oxidation Mechanism of Glucose and Other

Recommend Documents