Remarkable bismuth-gold alloy decorated on MWCNT for glucose electrooxidation: the effect of bismuth promotion and optimization via response surface methodology

In this study, the carbon nanotube supported gold, bismuth, and gold-bismuth(Au/MWCNT, Bi/MWCNT, and Au-Bi/MWCNT) nanocatalysts were prepared with NaBH4 reduction method at varying molar atomic ratio for glucose electrooxidation (GAEO). The synthesized nanocatalysts at different Au: Bi atomic ratios are characterized via x - ray diffraction (XRD), transmission electron microscopy (TEM), and N2 adsorption-desorption. For the performance of AuBi/MWCNT for GAEO, electrochemical measurements are performed by using different electrochemical techniques namely cyclic voltammetry (CV), linear sweep voltammetry (LSV), chronoamperometry (CA), and electrochemical impedance spectroscopy (EIS). Monometallic Au/MWCNT exhibits higher activity than Bi/MWCNT with 256.57 mA/mg (0.936 mA/cm2) current density. According to CV results, Au80Bi20/MWCNT nanocatalyst has the highest GAEO activity with the mass activity of 320.15 mA/mg (1.133 mA/cm2). For Au80Bi20/MWCNT, central composite design (CCD) is utilized for optimum conditions of the electrode preparation. Au80Bi20/MWCNT nanocatalysts are promising anode nanocatalysts for direct glucose fuel cells (DGFCs).

Herein, we aim to investigate the effect of Bi addition to Au in terms of GAEO activity. Thus, Au/MWCNT, Bi/ MWCNT, and Au-Bi/MWCNT nanocatalysts were prepared via NaBH 4 reduction method, and these nanocatalysts were characterized by XRD, BET, and TEM. To investigate the effect of Bi promotion, GAEO activities of these nanocatalysts are measured via cyclic voltammetry (CV), linear sweep voltammetry (LSV), chronoamperometry (CA), and electrochemical impedance spectroscopy (EIS). For Au 80 Bi 20 /MWCNT nanocatalyst, central composite design (CCD) was utilized for determining the optimum conditions of electrode preparation. The volume of nanocatalyst slurry (V c , A), ultrasonication time of the nanocatalyst slurry (t u , B), and the drying time of the electrode (t d , C) were determined as independent variables.
All electrochemical properties of Au/MWCNT, Bi/MWCNT, and Au-Bi/MWCNT nanocatalysts were determined by CV, LSV, CA, and EIS in 0.5 M GA. A nanocatalyst ink was obtained by dispersing 3 mg nanocatalyst in 1 mL of Nafion. Then, 5 mL of nanocatalyst ink was transferred to glassy carbon electrode and dried. CV measurements were performed at -0.6 V to 0.8 V potentials at 50 mv s -1 scan rate. Stability measurements were conducted by CA during 1000 s. CCD was utilized for optimum conditions of the electrode preparation. The volume of nanocatalyst slurry (V c , A), ultrasonication time of the nanocatalyst slurry (t u , B), and the drying time of the electrode (t d , C) are determined as independent variables. The maximum current density values obtained for GAEO were identified as the response. The error for the value of response was determined by 6 experiments at the middle levels of the parameters, and 20 sets of experiments were performed in total. Table 2 depicts the experimental points determined by Design Expert 7.0 and their corresponding response values, where, ˗ 1, 0, and + 1 represent the lowest, central, and highest levels of the parameters. Interactions between independent parameters were statistically evaluated with analysis of variance (ANOVA), and the suitability of the proposed model was tested with the coefficient of determination (R 2 ).

Characterization results
Characterizations of Au/MWCNT, Bi/MWCNT, and Au 80 Bi 20 /MWCNT nanocatalysts were performed with XRD, BET, and TEM. XRD patterns of MWCNT supported Au, Bi, and Au 80 Bi 20 nanocatalysts were given in Figure 1. The diffraction peaks of Au and Bi were clearly seen in Figure 1. The diffraction peaks of C (0 0 2) and C (1 0 0) planes were observed at around 25.5° and 42.8° for all nanocatalysts, respectively. The presence of the C (0 0 2) indicates that the carbon in the structure is hexagonal carbon [44]. The (1 1 1), (2 0 0), (2 2 0), (3 1 1), and (2 2 2) facets of Au were obtained at 38.3°, 44.4°, 64.5°, 77.3°, and 81.9° 2θ values for Au/MWCNT. These peaks are specific crystallographic planes of the face-centered cubic (fcc) Au [45,46] (Figure 1). The average diameter of the nanocatalysts can be achieved by using Scherrer's Equation [47,48]. The crystal size of Au/MWCNT was found as 19.   (43) active sites of Bi. The crystal size of Au 80 Bi 20 /MWCNT nanocatalyst was found as 21.96 nm. Moreover, the average interplanar distances of Au/CNT, Bi/CNT, and Au 80 Bi 20 /MWCNT nanocatalysts were calculated using Bragg's Law [49,50]. The 2θ values of Au (111) and Bi (110) peaks, which are the most intense peaks in the XRD patterns, were used. The average interplanar distance for Au/CNT, Bi/CNT, and Au 80 Bi 20 /MWCNT nanocatalysts was calculated as 2.35, 2.95, and 2.35 nm, respectively. N 2 adsorption-desorption were used to determine pore size, BET surface area, and pore volume of Au/MWCNT, Bi/ MWCNT, and Au 80 Bi 20 /MWCNT. BET surface area, pore size, and pore volume of nanocatalysts were given in Figure 2 and Table 3. In this study, all of the used nanocatalysts were exhibited the V-type adsorption-desorption isotherm with H1 type hysteresis loop [51]. This indicates that the catalysts are mesoporous in according to International Union of Pure and Applied Chemistry (IUPAC) categorization. BET surface areas of Au/MWCNT, Bi/MWCNT, and Au 80 Bi 20 /MWCNT were found as 159.0, 225.1, and 221.6 m 2 /g, respectively. As can be clearly seen from Table 3, the use of Bi and Au together increased the BET surface area. Likewise, the increase in pore volume and pore size of the nanocatalyst were observed (Table 3). According to the pore size and pore volume, the nanocatalysts are sorted Au 80 Bi 20 /MWCNT > Bi/MWCNT > Au/MWCNT.
The morphological and particle size of the Au/MWCNT, Bi/MWCNT, and Au 80 Bi 20 /MWCNT nanocatalysts was determined with TEM and were depicted in Figure 3. It is explicit that Au and Bi nanoparticles were agglomerated for Au/MWCNT and Bi/MWCNT nano-catalysts. However, it can be clearly seen that such a situation is not observed for Au 80 Bi 20 /MWCNT nanocatalyst and that there is a homogeneous distribution. This can be explained by the fact that Bi nanoparticles on the MWCNT surface have a positive effect by entering between the Au nanoparticles. The increase in BET surface area of Au 80 Bi 20 /MWCNT compared to Au/CNT could support this positive effect. The average particle size for Au/MWCNT, Bi/MWCNT, and Au 80 Bi 20 /MWCNT nanocatalysts was found as 26.8, 23.2, and 19.38 nm, respectively. The particle size for Au 80 Bi 20 /MWCNT was found to be consistent with crystal sizes obtained from the XRD result.

Electrochemical assessment
Au/MWCNT, Bi/MWCNT, and Au-Bi/MWCNT nanocatalysts were prepared via the NaBH 4 reduction method to investigate their GAEO activity.  The hydroxide (OH -) adsorption-desorption peak was observed for Au and Bi between 0.4 V and 0.6 V, while these peaks were not visible for Au 80 Bi 20 (Figure 4a). As described in the literature, the Au desorption peak is obtained due to the reduction of the oxidative gold layer [52]. Due to the dispersion of Bi nanoparticles in the Au layer formed on the surface of the AuBi/MWCNT nanocatalyst, it could prevent oxidation at the positive forward direction peak of the Au layer. As seen, electrooxidation peaks were obtained for all nanocatalysts prepared. When using Bi together with Au, it is observed that the current density is clearly increased. Au 80 Bi 20 /MWCNT nanocatalyst exhibited the highest performance among prepared nanocatalysts with 1.133 mA/cm 2 (320.1 mA/mg Au) for GAEO ( Figure 5 and Table 4). These results are consistent with CA and EIS measurements.     The mass activities of Au/MWCNT, Bi/MWCNT, and Au 80 Bi 20 /MWCNT catalysts were examined via LSV technique at a scan rate of 50 mV s -1 . LSV profile of these nanocatalysts in 1 M KOH + 0.5 M GA solution were given in Figure 6. As could be seen in Figure 6, Au 80 Bi 20 /MWCNT nanocatalyst exhibited a higher mass activity compared to these of Au/ MWCNT and Bi/MWCNT nanocatalysts toward GA electrooxidation. Mass activities over the total potential for Au/ MWCNT, Bi/MWCNT, and Au 80 Bi 20 /MWCNT were determined as 1024.60, 200.82, and 1601.64 mA/mg Au, respectively. These results were consistent with the results from CV, CA, and EIS.
Electrode preparation parameters for maximum glucose electrooxidation, namely V c , t u , and t d were optimized by using RSM. In optimization studies, the working electrode was modified with the AuBi nanocatalyst. Table 2 (Table  5). Besides, the determination of coefficient and adequate precision values of the model were found to be 0.89 and 10.7, respectively. The fact that the lack of fit value was statistically insignificant indicates that the model depicts a good agreement with the experimental data. Accordingly, the proposed model can be utilized to navigate the design space [53]. Figure 8 depicts the response surface plots for t u , V c and t d parameters. The interaction between t u and V c for specific activity toward GAEO was presented in Figure 8a. Specific activity for GAEO decreases when the V c value is increased from 0.5 to 7.75 µL. An increase in specific activity was observed for nanocatalyst loads higher than 7.75 µL. Figure 8b depicts that the specific activity increases up to about 15 min of t d and begins to decrease after this maximum point. It was observed that the AuBi nanocatalyst could not attach enough to the electrode surface at very low t d values, and some of the nanocatalysts were removed from the electrode surface. At higher t d values, lower specific activities were observed as a result of oxidation of metals and vaporizing of Nafion in the nanocatalyst slurry. It was determined from Figure 8c that the relation of the specific activity with t u depicts a volcano shape. The specific activity of AuBi for GAEO increased up to about 45 min of t u , and a decrease was observed after this value. This may be due to the sonification time affecting the crystal structures of the nanoparticles. Pollet et al. emphasized that during high sonification periods, the crystallinity of nanoparticles can be disrupted, and the formation of amorphous structures could be observed [54]. Design-Expert software was used to determine optimum conditions for GAEO, and related results were summarized in Table 6. The V c of 0.5 µL, t u of 44.87 min, and t d of 11.49 min were obtained as an optimum condition for electrode preparation toward GAEO on AuBi/MWCNT. It could be seen in Table 6 that specific activity under optimum conditions was predicted by the obtained model as 1.40971 mA/cm 2 . The experiment was conducted under optimum conditions to verify the specific activity value derived from the model, and the specific activity was found to be 1.62 mA/cm 2 . It was determined that the obtained model was close to the experimental value with an error of 13%, indicating that the predicted value was in harmony with the observed value.   MWCNT and Bi/MWCNT, respectively. Moreover, Au 80 Bi 20 /MWCNT has the best stability and highest GAEO activity in the long term. Figure 10 depicts the Nyguist plot for Au/MWCNT, Bi/MWCNT, and AuBi/MWCNT in 0.5 M GA. The shape of the Nyguist plots is generally semicircle, and the diameter of these semicircles has a significant effect on the charge transfer resistance of catalyst. Accordingly, when the diameter of the semicircles decreases, the charge transfer resistance decreases and the GAEO activity of the nanocatalyst increases. According to Figure 10, the charge transfer resistance can be listed as Au 80 Bi 20 /MWCNT <Au/MWCNT <Bi/MWCNT. The fitted EIS profile of Au/MWCNT, Bi/MWCNT, and AuBi/MWCNT were given in Figure S1, Figure S2, and Figure S3, respectively. The charge transfer resistance of Au/MWCNT, Bi/MWCNT, and AuBi/MWCNT were determined as 2.502, 3.733, and 2.279 Ω, respectively. As a result, it was found that Au 80 Bi 20 / MWCNT nanocatalyst has the highest GAEO activity, and these results are in agreement with CV and CA results.

Conclusion
Au/MWCNT, Bi/MWCNT, and bimetallic Au-Bi/MWCNT were synthesized via NaBH 4 reduction method, characterized by advanced surface analytical methods. The electrocatalytic performance of prepared catalyst was investigated with EIS, CV, CA, and LSV toward GAEO. Following results and insights were obtained: Ø Au/MWCNT, Bi/MWCNT, and Au-Bi/MWCNT at varying Au:Bi ratios could be easily prepared from corresponding Au and Bi precursors via NaBH 4 reduction method.
Ø According to XRD and TEM results, particle sizes of Au 80 Bi 20 /MWCNT were compatible with each other. It was observed that the BET surface area of Au/MWCNT increased with the addition of Bi.
Ø Electrochemical measurement was revealed that Bi addition improves the electrochemical activity of Au/ MWCNT. This situation can be explained by electronic effect. Ø According to CV results, Au 80 Bi 20 /MWCNT showed the highest GAEO performance. The optimum metal molar ratio is the basis for this performance.
Ø CCD was utilized for optimum conditions of the electrode preparation. The volume of nanocatalyst slurry (V c , A), ultrasonication time of the nanocatalyst slurry (t u , B), and the drying time of the electrode (t d , C) are determined as independent variables. The maximum current density values obtained for GAEO were identified as the response. The V c of 0.5 µL, t u of 44.87 min, and t d of 11.49 min were obtained as an optimum condition for electrode preparation toward GAEO on AuBi/MWCNT. Ø CA and EIS results revealed that AuBi nanocatalyst has a high stability and fast oxidation kinetics. Ø The data obtained from this study depicts that Au 80 Bi 20 /MWCNT nanocatalyst is a good candidate as anode nanocatalyst for DGFC.

Acknowledgments
Hilal Demir Kıvrak would like to thank for the financial support for the Scientific and Technological Research Council of Turkey (TUBITAK) projects (project no: 114M879, 114M156, 116M004) for chemicals and characterization. Scholarships were provided from YOK 100/2000 and TUBITAK 2211 A for Ömer Faruk ER.