Electrochemical Glucose Oxidation Using Glassy Carbon Electrodes Modified with Au-Ag Nanoparticles: Influence of Ag Content

1Centro de Investigación y Desarrollo Tecnológico en Electroquı́mica, Parque Tecnológico Querétaro, s/n, Sanfandila, 76703 Pedro Escobedo, QRO, Mexico 2Centro de Investigación en Quı́mica Aplicada, Bulevar Enrique Reyna 140, Colonia San José de los Cerritos, 25294 Saltillo, COAH, Mexico 3Laboratorio de Electroquı́mica, Facultad de Ciencias Quı́micas, Universidad Autónoma de Nuevo León, Guerrero y Progreso, s/n, Colonia Treviño, 64570 Monterrey, NL, Mexico

On the other hand, Gluconic acid (GA) is a high value added chemical used as intermediate in the food, pharmaceutical, and beverage industries.The only commercial/industrial method to produce GA is by glucose biochemical oxidation [35], whose main disadvantages are slow reaction rate, low space-time, and problematic enzymeproduct separation.In the past few years, heterogeneous catalysis using Pt, Pd, or Au nanoparticles has attracted much attention [36,37].The selective glucose oxidation to GA has been reported using Au clusters as catalyst and O 2 as oxidant giving a yield of 98% at mild conditions, that is, 50 ∘ C and atmospheric pressure.Indeed, higher costs, electrode poisoning, and hard reuse of nanocatalyst make this process far from competitive GA production [38].Some studies have been oriented to the use of bi-and trimetallic nanoparticles to overcome the noneconomic electrochemical GA production, at least to elucidate the best nanoparticle composition for the electrooxidation of glucose.In this sense, bimetallic Au-Ag nanoparticles show important synergistic effects in the catalytic properties; these properties depend upon metals ratio of nanoparticle, nanoparticle synthesis, nanoparticle size, form, and crystallographic orientation [39,40]; indeed, it has been shown that the surface morphology and particularly the size of nanoparticles can be easily controlled by experimentally adjusting the synthesis conditions [41,42].
In our previous work, we reported that Au nanoparticles supported on glassy carbon presented a catalytic activity and selectivity towards glucose oxidation, depending on the particle size and on the crystallographic orientation.Results also suggest that oxidation process in these conditions is taking place with lower poisoning of the surface in the case of the Au nanoparticles than for massive gold that the process is irreversible, and perhaps there are some chemical reactions involved in the overall oxidation process [25].
The aim of this work is to evaluate the catalytic properties of Au x -Ag y NPs towards the glucose electrooxidation; hence, we studied mainly the Au-Ag nanoparticles voltammetric response of different Au-Ag ratio.

Preparation of Au x -Ag y Nanoparticles (Au x -Ag y NPs).
Au x -Ag y NPs were prepared according to a previously published procedure [25].The procedure has been modified in order to prepare bimetallic nanoparticles.Briefly, 30 mL of aqueous solutions containing the respective amount of HAuCl 4 and AgNO 3 to give nanoparticles Au 100 (M1), Au-Ag 90/10 (M2), Au-Ag 80/20 (M3), and Au-Ag 70/30 (M4) was prepared (see Table 1) and was added to 10 mL of toluene (Aldrich) containing 0.34 mM of tetraoctylammonium bromide (TOAB 98%, Fluka) as a phase-transfer agent.Dodecanethiol (Aldrich) was incorporated to these different solutions as a stabilizing agent in an Au-Ag : thiol 2 : 1 ratio followed by the addition of an excess of NaBH 4 as an aqueous reducing agent.The reaction was allowed to proceed under constant stirring at controlled temperature (80 ∘ C), for 3 h.Finally, colored dispersions were obtained and purified several times in ethanol (J.T. Baker).The resulting Au-Ag nanoparticles were characterized by transmission electron microscopy (TEM), UV-Visible spectroscopy (UV-Vis), Xray diffraction spectroscopy (XRD), and cyclic voltammetry (CV).

Preparation of Glassy Carbon Modified Electrodes.
To remove the thiol stabilizing agent layer from Au x -Ag y NPs, the NPs were mixed with Vulcan XC-72 and were heated at 300 ∘ C for 2 hours under air atmosphere; the temperature was controlled within 2 ∘ C.
Then, Au x -Ag y NPs-modified electrodes were prepared as follows: 1 L aliquot of thiol-Au x -Ag y NPs in hexane and Nafion (5% Electrochem) were mixed (1 : 10 ratio) and cast onto a carbon disk (CD) followed by natural evaporation at room temperature.

Characterization of Nanoparticles. Synthesized Au
x -Ag y NPs were characterized using TEM, UV-Vis, XRD, and cyclic voltammetry (CV).TEM characterizations were performed on a TITAN 80-300 FEI microscope and the particle size distribution was measured from about 200 particles using the TEM images.Nanoparticles samples dissolved in hexane were cast onto a carbon-coated copper grid sample holder followed by natural evaporation at room temperature.UV-Vis measurements were carried out on HP spectrophotometer model 8453.XRD measurements were obtained using a Siemens D5000 diffractometer.Spectra were collected from 10 to 80 ∘ at a speed of 0.002 ∘ ⋅seg −1 .Cyclic voltammetric measurements were performed in BioLogic VSP Modular 5 channels Potentiostat/Galvanostat using a three-electrode conventional cell.A separated compartment and an Ag/AgCl electrode were used as reference and a Pt wire as the counter electrode; all potentials were referred to this electrode.Glucose 0.01 M in KOH 0.3 M solution was used for the electrochemical characterization.The solution was purged for twenty minutes with high-purity nitrogen before taking measurements.ranging from 1 to 7 nm.Figures 2(a)-2(d) show the Au x -Ag y NPs supported in Vulcan after heat treatment for thiol elimination.It can be seen that Au x -Ag y NPs keep their spherical form but at the same time there is an increase in size and distribution.The final size ranges from 6 to 40 nm; in spite of the size increase, the Au x -Ag y NPs are still small compared to other synthesis methods [33].

UV-Visible.
It is well known that nanoparticles have surface plasmon (SP) resonance absorption bands in the visible region.SP resonance bands are strongly dependent on the size, shape, composition, and medium dielectric properties [43][44][45][46].
Figure 3 shows UV-Visible spectra from Au x Ag y nanoparticles dispersed in hexane.Au and Ag nanoparticles are known to have plasmon bands in the visible region located at 520 and 420 nm, respectively.In the figure, distinct peaks are observed clearly at 530, 520, 512, and 506 nm; all these peaks are located at intermediate positions between the Au and Ag plasmon bands.The plasmon band is blue-shifted as the silver amount increases.
Two plasmon bands would be expected for a physical mixture of Au and Ag nanoparticles and the formation of an Au-Ag alloy could be deduced from the fact that the optical absorption spectrum shows only one plasmon band for all the samples [47,48].
The formation of spherical Au x -Ag y NPs was demonstrated by TEM (Figure 1).Indeed, upon heat treatment for thiol elimination, the Au x -Ag y NPs, in spite of slight size increase, keep the spherical form as shown in Figure 2. Results are in agreement with the literature [25].

XRD Characterization.
Figure 4 shows the XRD patterns of the different NPs synthesized.XRD analysis revealed in all the samples the presence of four representative diffraction peaks of face-centered cubic (fcc) structure that correspond to (111), ( 200), (220), and (311) planes.The angles of these diffraction peaks in all the samples were greater than those for pure Ag (38.11, 44.27The intensity of each signal respect to the other signals means the proportion of each orientation present in that particular nanoparticle.For nanoparticles synthesized by our method, the main factor influencing the orientation plane is the Au-Ag : thiol ratio [25].In this case, we used a constant Au-Ag : thiol ratio of 2 : 1 in order to obtain a (111) : (200) plane ratio of 2.3-2.8.

Electrochemical Characterization.
We have used cyclic voltammetry for the Au x -Ag y NPs electrochemical characterization.Figure 5 shows the response of both electrodes, Au polycrystalline (Poly-Au) (Figure 5 We can see the typical Au voltammetric response, that is, the formation and reduction of Au oxides.The surface area of the Au 100 NPs electrode was larger than that of a Poly-Au electrode by integrating the reduction charge of gold oxide monolayer.It indicated that there was nanostructured gold on the Au NPs electrode.The Au 100 NPs electrode exhibited the same behavior of Poly-Au electrode, having higher current densities of relative peaks compared to that of Poly-Au electrode.Poly-Au and Au 100 NPs electrodes exhibit the anodic peak located at 0.28 V corresponding to the oxidation of gold.The cathodic peak is located at 0.07 V for Au 100 NPs and 0.1 V for Poly-Au corresponding to the subsequent reduction of the gold oxides.Nevertheless, there was a weak cathodic peak located at −0.2 V for Au 100 NPs and −0.18 V for Poly-Au attributed to desorption of hydroxide anions. As the content of Ag in nanoparticles increases (Au 80 Ag 20 and Au 70 Ag 30 , Figures 5(c) and 5(d)), the voltammogram changes to take the form of that of Ag nanoparticles in alkaline media (Figures 5(e) and 6(f)) [49].The cyclic voltammograms of the silver disk electrode or silver nanoparticlemodified electrode in alkaline media show three anodic peaks (Ox 1 , Ox 2 , and Ox 3 ) appearing at 0.25, 0.33, and 0.65 V, respectively (Figures 5(e) and 6(f)).During a reverse scan, two cathodic peaks (R 1 and R 2 ) appeared at 0.38 and 0.06 V, respectively, as reported in previous studies [49].The first anodic peak (Ox 1 ) is due to the electroformation of the monolayer of Ag 2 O.The second anodic peak (Ox 2 ) is due to the formation of the multilayer of Ag 2 O.The third peak (Ox 3 ) is due to the oxidation of Ag 2 O to AgO or direct oxidation of Ag to AgO.
The cathodic current peak R 1 is due to the reduction of AgO to Ag 2 O, and R 2 corresponds to the reduction of Ag 2 O to Ag.A small anodic peak (Ox 4 ) at 0.63 V appears during the cathodic scan, but it was dependent on the scan rate and occurred only when the scan rate is faster than 50 mV/s.Stonehart has suggested that peak Ox 4 is associated with autocatalytic oxidation process to convert Ag 2 O to AgO [50].When the nuclei of AgO formed at the Ag 2 O surface during the anodic scan, further formation of AgO became easy and occurred at a lower potential than the initial formation of AgO nuclei.Therefore, during the cathodic scan, further formation of AgO resulted in the appearance of a small anodic peak.
It can be seen in Figure 5(c) (Au 80 Ag 20 ) that the anodic scan shows three peaks probably due to the oxidation of gold (Ox 1 ), the electroformation of the monolayer of Ag 2 O (Ox 2 ), and the oxidation of Ag 2 O to AgO (Ox 3 ).The cathodic scan shows also three peaks probably corresponding to the reduction of AgO to a multilayer of Ag 2 O (R 1 ), the reduction of Ag 2 O to Ag (R 2 ), and the reduction of Au oxides (R 3 ).
As a consequence of the increase in Ag content, the anodic scan of Au 70 Ag 30 NPs shows two peaks.The first one (Ox 1 ) is clearly due to the simultaneous formation of Au oxides and the formation of the Ag 2 O monolayer because practically these two processes take place at the same potential resulting in one peak; the second anodic peak (Ox 2 ) is due to the oxidation of Ag 2 O to AgO.The cathodic scan shows four peaks: R 1 is due to the reduction of AgO to a multilayer of Ag 2 O, R 2 is due to the reduction of Ag 2 O to Ag, and  peaks R 3 and R 4 are due to the reduction of the gold oxides and the desorption of hydroxide anions.These results are in agreement with others works [49,51].There is also a shift of the glucose potential oxidation on the Au x -Ag y NPs M1, M2, and M3 electrodes with respect to the Poly-Au electrode (see Table 2); the shift is to less-negative potentials and reveals that the Au x -Ag y NPs electrodes could catalyze the oxidation of glucose.Indeed, we show in Table 2, for comparison purposes, the current density under the same potential for the electrode tested.These results are similar to that of Qiaofan's work on glucose oxidation in an alkaline electrolyte obtained at bimetallic Au-Ag/RGO/GC (bimetallic Au-Ag/reduced grapheme oxide/glassy carbon) nanoparticles modified electrode [51].

Electrochemical Oxidation of Glucose on AuNPs and
The CV curves of the Au x -Ag y NPs electrode presented a typical two-step oxidation process and three obvious oxidation peaks in the positive scan were observed.As the mechanism for the electrocatalytic oxidation of glucose at Au x -Ag y NPs electrodes is a multistep one, the first step mediated the chemisorptions of hydroxide anions onto the electrode surface, leading to the formation of hydrous goldsilver oxide (Au x -Ag y OH ads ) which was believed to be the catalytic component of Au x -Ag y NPs electrodes [52].
The glucose oxidation highly depends on the quantity of Au x -Ag y OH ads and the premonolayer oxidation was enhanced at Au x -Ag y NPs surface.The first peak, Ox 1 , located between −0.54 and −0.39 V could be attributed to the electrosorption of glucose to form adsorbed intermediate gluconolactone, releasing one proton per glucose molecule.At around these potentials, there could be a very limited number of Au x -Ag y OH ads sites on the Au x -Ag y NPs surfaces due to the reductive nature of the electrodes potential.Glucose molecules interacted with the Au x -Ag y OH ads sites to give rise to the intermediate, gluconolactone [53,54].Nevertheless, the accumulation of gluconolactone blocks the active sites of the Au x -Ag y NPs electrode surface inhibiting the direct oxidation of glucose, leading to the decrease of current density.The second oxidation peak Ox 2 appeared at a more positive potential located between 0.18 and 0.26 V depending on the Ag content in the electrode.When the potential reaches those values, the population of Au x -Ag y OH ads sites on the Au x -Ag y NPs electrode increased; hence, catalytic oxidation of the intermediates takes place.A further positive increase in the potential to around 0.6 V gives rise to the formation of goldsilver oxides on the surface of Au x -Ag y NPs.The formation of Au-Ag oxides results in a decrease of the number of the Au x -Ag y OH ads sites on the surface Au x -Ag y NPs and impedes the electrocatalytic oxidation of glucose and by-products.The    current density, then, decreases dramatically at more positive values than 0.60 V, due to formation of gold and silver oxides.
In the negative potential scan, the reduction of the surface gold and silver oxides would occur at the potential more negative than 0.16 V.Then, enough surface active sites would be available for the direct oxidation of glucose, resulting in a sharp increase in anodic current showing a peak (Ox 3 ), located between 0.001 and 0.15 V, depending on the Ag content in the electrode.Tang's group had reported a similar result; that is, there is a sharp increase in anodic current with a peak at 0.1 V [53].The surface structure of a gold electrode influenced the electrooxidation of glucose via the influence on the adsorption of OH − anions and gold oxide surface monolayer formation [55].The as-prepared Au x -Ag y NPs electrode exhibited better performance on electrocatalytic oxidation toward glucose than a Poly-Au electrode.
To evaluate the kinetics of the glucose oxidation on the Au x -Ag y NPs, the CVs of the electrodes in a 0.3 M KOH solution containing 10 mM glucose at different scan rates were recorded (figure not shown).The peak Ox 1 in the positive scan, located between −0.54 and −0.39 V, corresponds to the direct oxidation of glucose and the current density is proportional to the square root of scan rate for all the electrodes used (Poly-Au, M1, M2, M3, and M4, Figure 7(a)); this fact suggests a linear diffusion-controlled process.Indeed, it is in agreement with the proposed mechanism for peak Ox 1 , in which sufficient active sites on the hydroxyl coated gold surface allowed direct oxidation of glucose to generate gluconolactone [25,56].We can see that the order from better to worst with respect to the glucose oxidation is M3 > M1 > M2 > Poly-Au > M4.As expected, the current density of all electrodes increases as scan rate rises but that of Poly-Au remains practically unchanged.Furthermore, the slope of Au  -Ag  NPs increases depending on the Ag content in the NPs.The Au 80 -Ag 20 electrode has the highest slope, that is, the highest glucose oxidation rate.
In Figure 7(b), we can see the plot of current density versus square root of scan rate corresponding to the further oxidation of gluconolactone located between 0.18 and 0.26 V (Ox 2 ) on the same electrodes except that of M4. Figure 7(b) suggests again that the oxidation of gluconolactone is also a diffusion-controlled process.The order from better to worst is M3 > M1 > Poly-Au > M2.This time, the slope of Poly-Au increases as expected; we are working on the matter.
Finally, Figure 7(c) shows the plot of current density versus square root of scan rate for the oxidation peak Ox 3 .Interestingly, as could be seen, peak Ox 3 corresponds to the direct oxidation of glucose in cathodic scan and is also linear versus scan rates.Indeed, this is a typical adsorptioncontrolled process corresponding to a fast electron-transfer behavior.In the negative potential scan, the reduction of the gold surface oxides provides enough surface active sites for the direct oxidation of glucose; they could account for the appearance of peak Ox 3 .The order from better to worst is M3 > Poly-Au > M1 > M2.This time, the slope of Poly-Au remains again unchanged.
None of the straight lines of plots in Figures 7(a), 7(b), and 7(c) go through origin indicating a previous chemical step in the oxidation process; as noted above, this step is probably due to the chemisorption of hydroxide anions onto the gold and silver surface, leading to the formation of hydrous oxides (Au x -Ag y OH ads ), which are believed to be the catalytic component of Au x -Ag y NPs electrodes, particularly Au 80 -Ag 20 NPs electrode.

Conclusion
The nanoparticles applications, for instance, glucose oxidation, critically depend on their specific catalytic activity; unfortunately, it is still poorly understood [57].Experimental evidence that can be found in literature points towards several variables affecting it, that is, nanoparticle size [57], nanoparticle crystal orientation [56], and composition of nanoparticle [51].It is admitted that the glucose oxidation depends on the presence of what is known as monolayer or premonolayer of AuOH ads or AgOH ads [48]; at higher potential, the formation of Au and Ag oxides impedes the glucose oxidation.Nevertheless, Mikhlin et al. [57] found that the formation of AgO proceeded readily at the "primary" oxide for nanoparticles of about 12 nm but was retarded at the smaller ones.These findings are in agreement with our previous work where we have found a maximum in the catalytic activity for Au nanoparticles of about 6 nm and could explain the present findings; that is, smaller nanoparticles have the highest catalytic activity towards glucose oxidation.
Regarding nanoparticle composition, the experimental evidence strongly suggests that the best composition is around 80% Au and 20% Ag.
Indeed, we have performed the present study for a (111) : (200) plane ratio of 2.3-2.8.Studies to optimizing the NPs size, Au-Ag ratio, and crystallographic orientation are underway.

3. 1 .
Characterization of Au x -Ag y NPs 3.1.1.TEM Characterization.Figures 1(a)-1(d) show TEM micrographs of the synthesized Au x -Ag y nanoparticles (labeled as M1 to M4) and their particle size distribution.Synthesis of Au x -Ag y NPs can be easily controlled to obtain narrow particle size distribution of spherical nanoparticles

Table 1 :
Composition of Au-Ag nanoparticles synthesized.
, 64.42, and 77.47 ∘ ) and lower than those for pure Au(38.18,44.39,64.57, and77.54∘ ) according to the standards for Ag [JCPDS 04-0783] and Au [JCPDS 04-0784].These results confirm again the formation of an alloy of Au-Ag.No other impurity peak is detected in the samples.

Table 2 :
Oxidation peak and current density as a function of Ag content.