Introduction

Heterogeneous thermocatalytic oxidation and hydrogenation reactions over metals contribute to a wide range of chemical transformations for compounds vital to societal production or environmental protection1,2,3,4,5,6,7. The Langmuir-Hinshelwood (L-H) mechanism, which requires co-chemisorption of reactants on the catalyst surfaces and elementary reactions between adjacent adsorbates, is one of the most widely accepted pathways for the oxidation or hydrogenation of targeted molecules5,8,9,10. However, when a reaction is transferred from a solid/gas phase to an aqueous/solid phase, the catalytic process may be impacted by the presence of water and ions. For instance, pH affects the binding energy of hydrogen on a catalyst11, and the ionic strength with consequent local electric field alters the selectivity of catalysis12. Adequate electrochemical elements, such as electrical contact, ionically-conductive media and resultant electrical potential gradient, can even modify the reaction mechanism to an alternate style involving electrochemical charge transfer, distinct from the one at solid/gas interphases13,14.

To deepen understanding of the entire catalytic process and further guide rational catalyst design, in situ spectroscopies have served as powerful tools to capture key intermediates15. The use of computational tools such as density functional theory (DFT) calculations also aids in elucidating reaction mechanisms16,17. Equally, the past few decades have witnessed the growing attention towards using electrochemical means to reveal the inherent properties of thermocatalytic reactions18,19,20. On the one hand, previous studies have shown that the open-circuit potential (OCP) of a catalyst, i.e., the electrochemical potential without current flow, significantly depends on the identities and coverages of adsorbates on its surface, which subsequently dictates the resultant catalytic activity21,22. From another perspective, some researchers have emphasized the importance of the mixed potential theory (MPT) in explicating thermal catalysis18. This theory separates a catalytic reaction into two independent electrochemical half-reactions that are coupled together by internal charge transfer through the catalyst20,23,24. Specifically, at a certain potential, the half-reactions will present identical absolute current, and this mixed potential and current indicate the operando OCP and thermocatalytic rates, respectively. We note that, in this manuscript, we define the operando OCP as the electrode potential measured during thermocatalytic reactions without any current flowing through the potentiostat. Successful analyses have been applied to multiple catalytic reactions, including the aerobic oxidation of small organic molecules13,20, the hydrogenation of a nitro group24, and the selective synthesis of H2O2 from H2 and O223. This fundamental understanding of thermal catalysis, in turn, guides catalyst design by incorporating different active sites that excel at different half-reactions, maximizing the apparent thermocatalytic activity of a given reaction25,26. Despite this progress, the applications of MPT primarily focus on reactions at OCP and cases with an external bias where thermal- and electro-catalysis occur simultaneously have not gained much attention from the community.

The importance of using electrochemical tools to directly alter the catalyst potential for improved thermocatalytic performance has also been highlighted in previous studies, a concept known as electrochemical promotion of catalysis (EPOC) or non-Faradaic electrochemical modification of catalytic activity (NEMCA)27. It was first reported that in an electrochemical system, where a solid electrolyte was sandwiched between a working electrode (WE) containing the catalyst and a counter electrode (CE), applying a bias to the WE would lead to directed migration of spill-overs27. This changed the work function of the catalyst through the formation of a promoter layer, significantly increasing the reaction rate mainly in a non-Faradaic manner27. Such a methodology was later applied to liquid-phase systems for promoted thermal catalysis and was found to be effective even without promoter layer formation19. The potential applied to the catalyst was mainly aimed to optimize the adsorption of surface intermediates directly or suppress/remove the coverage of poisons, thereby facilitating thermocatalytic activity22,28,29,30,31. Despite these advancements, the criteria for whether a thermodynamically favored catalytic reaction can be electrochemically promoted remains unclear. Furthermore, the connection between MPT and EPOC has not been adequately clarified thus far.

In this study, we use the aqueous aerobic oxidation/hydrogenation of the hydroquinone (HQ)/benzoquinone (BQ) redox couple over Pt catalysts as model reactions, allowing us to extend MPT from open-circuit to closed-circuit conditions and further reveal its connection with EPOC. Although a previous report recognized the aerobic oxidation of HQ over Co/N-doped carbon catalyst as a combination of the oxygen reduction reaction (ORR) on Co and HQ oxidation reaction (HQOR) on N-doped carbon32, there remains uncertainty regarding whether competitive adsorption of reactants occurs, and if the reaction mechanism changes when the metal shifts from a single atom to supported nanoparticles or sole metal. Our experimental findings indicate that thermal HQ/BQ redox reactions exhibit distinct reactant adsorption routes and relationships with MPT, dependent on whether Pt/C or Pt foil is utilized. Hence, the (electro)catalytic behavior of the reaction diverged over different Pt catalysts. We analyze the relationship between MPT and EPOC, providing insights into connecting them.

Results

Coupling of independent half-reactions for aerobic hydroquinone oxidation over supported Pt/C catalyst

Previous investigations on HQ suggest that it has strong adsorption on Pt33,34,35, and that the horizontal orientation may be the dominant one on a flat Pt surface36, which likely causes competitive activation between O2 and HQ when its aerobic oxidation is catalyzed by Pt nanoparticles. Figure 1 displays the possible reaction pathways for aqueous aerobic oxidation of HQ over Pt/C. If the overall reaction occurs via two independent half-reactions and the Pt surface is the active phase for both HQOR and ORR, internal charge transfer likely proceeds within a Pt nanoparticle or between Pt nanoparticles via the carbon support (Fig. 1a). Since both oxygen37,38 and HQ39 are known to exhibit high binding energy on the Pt surface, the competitive adsorption of reactants may cause a disparity between the observed thermocatalytic rate and the predicted rate derived from MPT. This is because MPT assumes that each half-reaction has unhindered access to the Pt surface. One possibility that can result in consistency with MPT is shown in Fig. 1b. If the carbon support acts as the active site for HQOR, as indicated by previous electrochemical research40, and Pt is solely responsible for ORR, the overall thermal reaction will be completed through electron transfer from carbon to Pt and the observed rate should align with the MPT prediction. On the other hand, Bates et al. reported that HQ oxidation occurred over a Co-N-C catalyst via a so-called “direct inner-sphere reaction”, which involved a non-Faradaic redox reaction between HQ and formed hydro(pero)xyl intermediate(s)41. Since ORR over Pt generates hydro(pero)xyl intermediate(s)42 and these species are generally considered active for catalytic oxidation of phenolic compounds43,44, a reaction mechanism of aerobic oxidation of HQ over Pt/C involving such a route cannot be excluded, either. Figure 1c shows a possible mechanism of this type. HQOR occurs on either Pt or carbon support and the released electrons reduce O2 on Pt to form OH* intermediates, which proceeds to oxidize additional HQ in a non-Faradaic manner. We note here that the occurrence of two half-reactions over different active sites (Fig. 1b) is not a hard requirement for agreement between the MPT-predicted rate and experimental measurement. With the presence of only one active phase, as is the case in Fig. 1a (Pt with inert carbon), kinetic competition for active sites cannot be assumed. For example, facile kinetics, weak adsorption, or poor mass transport can result in low coverages and prevent competition. To better understand the catalytic process of aerobic HQ oxidation over Pt/C, detailed analyses were first conducted to reveal its reaction mechanism.

Fig. 1: Possible reaction pathways for aqueous aerobic hydroquinone (HQ) oxidation over Pt/C.
Fig. 1: Possible reaction pathways for aqueous aerobic hydroquinone (HQ) oxidation over Pt/C.
Full size image

a Both oxygen reduction reaction (ORR) and HQ oxidation reaction (HQOR) occur on Pt surface with internal charge transfer inside Pt or between Pt nanoparticles. b ORR occurs on Pt, and HQOR occurs on carbon support with internal charge transfer from carbon support to Pt. c ORR over Pt and HQOR over either Pt or carbon occur, generating hydro(pero)xyl intermediate(s), which further oxidizes additional HQ in a non-Faradaic manner.

Commercial Pt/C (37.5 wt%; mean size: 2.4 nm, TEM and particle size distribution in Supplementary Fig. 1; X-ray diffraction (XRD) pattern in Supplementary Fig. 2) was first employed as the catalyst and was loaded on carbon paper (CP) as the WE. The thermocatalytic rate of aerobic HQ oxidation over Pt/C and the corresponding operando OCP with varying reactant pressure/concentration was measured in a two-compartment cell (Supplementary Fig. 3a) and is summarized in Table 1. Generally, increasing the O2 partial pressure (PO2) led to a positive shift of operando OCP, while increasing the initial HQ concentration (CHQ) resulted in a negative shift. BQ formation rate reached plateaus with increased CHQ and fixed PO2 (NO. 6–8) or increased PO2 and fixed CHQ (NO. 2, 4 and 7), which might be attributed to mass transport limitation of reactant(s) or saturated coverage of adsorbate(s) on the catalyst surface. Figure 2a shows the separate current dependence of HQOR (under Ar) and ORR half-reactions on the catalyst potential, indicating diffusion limitation of reactants with low concentration at high overpotentials. Using MPT, the operating potential and thermal rates of reaction can be predicted by identifying potentials where the absolute current produced by the two half reactions are equal. If MPT is obeyed, the working potential (OCP) during catalysis will correspond to this mixed potential, and the observed reaction rate (rmix in μmol min−1) will be equal to the absolute current (|I|mix in mA) at the mixed potential, as below:

$${r}_{{{{\rm{mix}}}}}=\frac{{\left|I\right|}_{{{{\rm{mix}}}}}\times 60\times {10}^{3}}{{nF}}$$
(1)

where n is the number of electron transfers involved in the reaction, i.e., 2 for HQ oxidation as well as BQ hydrogenation, and F is the Faradaic constant in C mol−1. Figure 2b, c compare the measured BQ formation rate and operando OCP with the predictions from MPT, respectively. Good consistency was observed for both OCP and BQ formation rate, excluding a mechanism involving non-Faradaic redox elementary step(s) (Fig. 1c)41,45. This result lends support to the separate action of HQOR and ORR.

Table 1 Benzoquinone (BQ) formation rate over Pt/C and the corresponding open-circuit potential (OCP) with varied initial hydroquinone (HQ) concentration and O2 partial pressure
Full size table
Fig. 2: Examination of the mixed potential theory (MPT) for aerobic HQ oxidation over Pt/C.
Fig. 2: Examination of the mixed potential theory (MPT) for aerobic HQ oxidation over Pt/C.
Full size image

a The current dependence of HQOR under Ar and ORR on the catalyst potential. Comparisons of (b) measured BQ formation rate and (c) operando open-circuit potential (OCP) with the prediction based on MPT. Conditions: Working electrode (WE): 0.75 mg Pt/C on 1 × 1 cm2 on CP; pressure: 101 kPa; reference electrode (RE): Hg/Hg2Cl2 (sat. KCl); counter electrode (CE): carbon rod; 295 K; 0.1 M HClO4; pH 1.2; stirring rate: 450 rpm.

Subsequent experiments further corroborated the presence of independent active sites for HQOR and ORR. Supplementary Fig. 4 displays cyclic voltammograms (CVs) for Pt/C under Ar or (diluted) CO with or without 10 mM HQ. Notable CO oxidation occurred at potentials higher than approximately 0.7 V vs. reversible hydrogen electrode (RHE; see green dashed line in Supplementary Fig. 4), and the high current observed at potentials lower than 0.7 V was associated with HQOR in the presence of aqueous HQ (refer to the black, red, and blue lines in Supplementary Fig. 4). Regardless of the presence of Pt-poisoning CO, similar CV curves were registered for all scenarios with 10 mM HQ. This likely implies that HQOR occurs on the surface of the carbon support independently of Pt nanoparticles. This assertion was further supported as carbon black (CB) on CP was shown to exhibit a comparable HQOR voltammetric response in place of Pt/C (Supplementary Fig. 4b). Furthermore, CO gas was introduced to block the Pt sites. (Supplementary Fig. 5). This caused a shift in the OCP from 0.686 V to 0.371 V vs. RHE, and only a trace formation of CO2 was recorded (less than 1 nmol s−1), indicative of a CO-poisoned Pt surface22. As a result, the HQ oxidation rate dwindled to a negligible amount compared to experiments run without CO due to the loss of anodic overpotential contributed by ORR over Pt. This implies that carbon alone cannot catalyze aerobic HQ oxidation.

A modified two-compartment cell with a reactant-separated configuration (Supplementary Fig. 3b) was employed to better study the half-reactions contributing to overall aerobic HQ oxidation. Different combinations of Pt/C and CB electrodes were placed into the cathode chamber containing 20 kPa O2 (Ar-diluted) or the anode chamber containing 10 mM HQ (Ar-purged). A conductive wire connected the electrodes for a “short-circuit” (SC) condition, and a potentiostat recorded the current. The current flow (converted to the unit of μmol e− min−1) and corresponding formation rate of BQ in the anodic chamber are summarized in Fig. 3a, which revealed good consistency between the BQ formation rate and recorded current (2 electron transfer for HQOR). Additionally, since the BQ formation rate and measured current were comparable regardless of the anode material, the carbon in CB and Pt/C was the active phase for the HQOR, as long as Pt/C was utilized for the ORR. The negligible activity from the combination of a Pt/C anode and a CB cathode signified the necessity of Pt for ORR. When the potential and current response of this electrode combination was measured at SC over time (Fig. 3b), only a minor overpotential was measured at the Pt/C electrode when ORR was occurring at the CB electrode, accounting for the negligible overall HQ oxidation activity. When Pt/C was used for ORR (Fig. 3c, d), larger overpotentials were observed for both HQOR and ORR, although substantial IR loss (I for current and R for resistance) was present due to the reactor’s high resistance. The large IR loss existing in the above configuration would lead to smaller overpotentials applied to Pt and carbon active sites (Fig. 3c) compared to the case of the original three-electrode configuration (Fig. 2a) and resultant smaller HQ oxidation rate. From the derivation in the Supplementary Information, reaction orders in CHQ and PO2 during the overall thermal catalysis were derived to be 0.31 and 0.69, respectively, based on the voltammetry curves of the electrochemical half-reactions46. The actual measurement of the BQ formation rate with various PO2 and CHQ (Supplementary Table 1) suggested a reaction order of 0.71 in PO2 and a reaction order of 0.40 in CHQ (Supplementary Fig. 6), consistent with the derivation. The above analyses support the assertion that aqueous aerobic HQ oxidation over Pt/C proceeds via two independent electrochemical half-reactions, i.e., HQOR over carbon (reaction 2) and ORR (reaction 3) over Pt, separately.

$${{{\rm{HQ}}}}{\mathop{\longrightarrow }^{{{\rm{Carbon}}}}}{{{\rm{BQ}}}}+2{{{{\rm{H}}}}}^{+}+2{{{{\rm{e}}}}}^{-}$$
(2)
$$\frac{1}{2}{{{{\rm{O}}}}}_{2}+2{{{{\rm{H}}}}}^{+}+2{{{{\rm{e}}}}}^{-}{\mathop{\to }^{{{\rm{Pt}}}}}{{{{\rm{H}}}}}_{2}{{{\rm{O}}}}$$
(3)
Fig. 3: Contribution of two electrochemical half-reactions over different active sites to the overall aerobic HQ oxidation.
Fig. 3: Contribution of two electrochemical half-reactions over different active sites to the overall aerobic HQ oxidation.
Full size image

a Comparison of the BQ formation rate in the anodic camber and the current flow (in μmol min−1) between electrodes containing different catalysts (cell configuration in Supplementary Fig. 3b). Time-evolution profiles of electrode potentials and current when (b) CB as the cathode and Pt/C as the anode, (c) Pt/C as the cathode and CB as the anode, or (d) Pt/C as both the cathode and the anode. Conditions: Total pressure: 101 kPa; 295 K; 0.1 M HClO4; pH 1.2; stirring rate: 450 rpm; OC: open-circuit; SC: short-circuit.

In the three-electrode system, this condition is achieved by the rapid oxidation of HQ catalyzed by abundant carbon sites in the presence of O2. This suppresses the accumulation of HQ on Pt sites and allows the ORR to proceed unhindered. The significant influence of HQ accumulation at Pt sites on the ORR activity is shown in Supplementary Fig. 14.

Validity of the mixed potential theory with applied external potential

The HQ oxidation performance over Pt/C with different reactant concentrations/pressures was also investigated with an applied external potential. Figure 4 shows the BQ formation rate as a function of applied anodic potential with 10 mM initial CHQ and varying PO2. Since the Faradaic efficiency (FE) of HQOR between 0.65 V and 0.78 V vs. RHE approaches unity (Supplementary Fig. 7a), it was assumed that all of the measured current contributed to the electrocatalytic HQOR when thermal- and electro-catalytic HQ oxidation was performed in parallel. Accordingly, the electrochemical (re in μmol min−1) and thermal (rt in μmol min−1) contributions to the overall HQ oxidation rate (rtotal) were calculated by Eqs. (4) and (5), respectively:

$${r}_{{{{\rm{e}}}}}=\frac{\left|I\right|\times 60\times {10}^{3}}{{nF}}$$
(4)
$${r}_{{{{\rm{t}}}}}={r}_{{{{\rm{total}}}}}-{r}_{{{{\rm{e}}}}}$$
(5)

where |I| was the absolute current in mA for HQOR (or benzoquinone reduction reaction (BQRR) in the following sections), and n is the number of electron transfers involved in the reaction, i.e., 2 for HQOR (as well as BQRR). With a fixed CHQ and varying PO2, it was found that the overall reaction rate increased with a larger applied potential. However, the thermal contribution decreased at higher potentials, made up for by the higher electrocatalytic contribution. Similar trends appeared for HQ oxidation with varying PO2 and CHQ (Supplementary Fig. 8). The effect of cathodic potentials was also investigated; only a decreased overall HQ oxidation rate was observed. For example, with 10 mM HQ and 20 kPa O2, rates of 0.313 μmol min−1 and 0.513 μmol min−1 were detected at 0.666 V and 0. 678 V vs. RHE, respectively, both smaller than that at OCP (0.923 μmol min−1). It seems likely that applying an external potential can only induce suppression rather than promotion of the thermal HQ oxidation rate.

Fig. 4: Thermal and electro-catalytic rate distribution during aqueous HQ oxidation with applied external potentials.
Fig. 4: Thermal and electro-catalytic rate distribution during aqueous HQ oxidation with applied external potentials.
Full size image

Conditions: WE: 0.75 mg Pt/C on 1 × 1 cm2 on CP; total pressure: 101 kPa; RE: Hg/Hg2Cl2 (sat. KCl); CE: carbon rod; 295 K; 10 mM HQ; 0.1 M HClO4; pH 1.2; stirring rate: 450 rpm.

Since thermal HQ oxidation over Pt/C occurs through two separate electrochemical half-reactions—the ORR and the HQOR—it is hypothesized that applying an external anodic potential to the WE will disrupt both the ORR and HQOR rates. For a positive applied potential, the excess charges generated from HQOR that cannot be consumed in the ORR will travel through the external circuit and be recorded by the potentiostat (Fig. 5a). More quantitatively, as demonstrated in Fig. 5b, when a certain anodic potential is applied (the vertical dash line as the independent variable), both the ORR and HQOR currents adjust to corresponding values based on voltammetric properties distinct from those at OCP. This results in a discrepancy between their current values, with only the overlapping part of cathodic and anodic currents being consumed within Pt/C as the apparent thermal catalysis (the blue part that is equal to ORR current based on voltammetry). The excess anodic current of HQOR in purple must be transferred through the external circuit, representing the electrocatalytic contribution. Similarly, when a cathodic potential is applied, an increased absolute ORR current cannot be fully compensated by the reduced absolute HQOR current. This leads to a suppressed overall HQ oxidation rate equaling the HQOR current based on the half-reaction voltammetry, and an undetectable formation of H2O that draws current from the potentiostat. Accordingly, thermal (rt, prediction in μmol min−1) and electro-catalytic (re, prediction in μmol min−1) contributions to the overall HQ oxidation rate can be predicted as long as the applied potential is set based on the following equations (take anodic potential application as an example here):

$${r}_{{{{\rm{t}}}},{{{\rm{prediction}}}}}=\frac{\left|{I}_{{{{\rm{ORR}}}}}\right|\times 60\times {10}^{3}}{{nF}}$$
(6)
$${r}_{{{{\rm{e}}}},{{{\rm{prediction}}}}}=\frac{\left(\left|{I}_{{{{\rm{HQOR}}}}}\right|-\left|{I}_{{{{\rm{ORR}}}}}\right|\right)\times 60\times {10}^{3}}{{nF}}$$
(7)

where IORR and IHQOR are the current values of ORR and HQOR half-reactions at a given potential (in mA), based on their respective voltammetric responses. This hypothesis is further demonstrated in Fig. 5c, d, which compare the measured and predicted thermal and electrochemical contributions, respectively, to the overall HQ oxidation rate under applied external bias. High consistency was observed between measurement and prediction at the same potential. This suggests that the foundational principle of MPT is applicable not only for thermal catalysis at OCP but also for the cases where thermal- and electro-catalysis occur in parallel with an external potential.

Fig. 5: Prediction of both thermal- and electro-catalytic rate for HQ oxidation over Pt/C with an external bias applied.
Fig. 5: Prediction of both thermal- and electro-catalytic rate for HQ oxidation over Pt/C with an external bias applied.
Full size image

a Schematic representation of charge transfer from HQOR to ORR and external CE when WE is connected to an external close-circuit. b Quantification example of thermal- and electro-catalytic HQ oxidation rate occurring in parallel based on the electrochemical properties of ORR and HQOR. Comparisons of (c) electrochemical contribution and (d) thermal contribution to the overall HQ oxidation rate with prediction when external potentials are applied. WE: 0.75 mg Pt/C on 1 × 1 cm2 on CP; total pressure: 101 kPa; RE: Hg/Hg2Cl2 (sat. KCl); CE: carbon rod; 295 K; 0.1 M HClO4; pH 1.2; stirring rate: 450 rpm.

External bias is usually applied to thermal catalysts in pursuit of enhanced thermocatalytic performance19,27. However, the analysis here suggests that electrochemical promotion does not work for thermocatalytic reactions that follow MPT and are made up of two electrochemical half-reactions without competitive reactant adsorption. Given that the half-reactions always adhere to their separate voltammetric responses, an external bias will simply cause a decline in their overlapping currents from the maximum at OCP. More specifically, taking HQ oxidation as an example (Fig. 5b), the application of an anodic potential will result in a larger anodic overpotential (driving force) for HQOR but a smaller cathodic overpotential for ORR; that is, higher HQOR rate and lower ORR rate compared to open-circuit conditions. The electrons transferred from HQOR to ORR over Pt/C account for the thermal contribution to the overall HQ oxidation rate, which is equal to the reduced ORR rate over Pt sites. Therefore, the apparent thermal reaction rate is suppressed with the external potential application for reactions following MPT, even though the overall reaction rate (thermal + electrochemical) should be increased.

Investigation of benzoquinone hydrogenation: the reverse reaction

The reverse reaction of HQ oxidation, i.e., aqueous BQ hydrogenation, was also examined with MPT. Figure 6a summarizes the current dependence of BQRR and hydrogen oxidation reaction (HOR) on catalyst potential over commercial Pt/C. Due to the large potential gap between H+/H2 (0.00 V vs. RHE) and BQ/HQ (0.75 V vs. RHE) redox couples and limited concentrations of reactant in the solution, the position where the current intersection appeared was always accompanied by mass transport issues. Taking the current intersections and corresponding potentials as predictions based on MPT, the results were compared with measurement (Supplementary Table 2) and summarized in Fig. 6b, c. Both the predicted HQ formation rate and potential showed good uniformity with experimental rate and OCP, indicating the validity of MPT for BQ hydrogenation reaction over Pt/C. The introduction of CO into the reactor significantly suppressed the BQ hydrogenation rate, which resulted from an OCP shift from 0.460 V to 0.711 V vs. RHE (Supplementary Fig. 5). Negligible CO2 formation (less than 1 nmol s−1) suggested the reduction of BQ was still driven by H2 rather than CO. The reactant-separated configuration (Supplementary Fig. 3b) was employed to study the individual contributions of half-reactions, i.e., HOR and BQRR, to the overall thermal catalysis. As shown in Supplementary Fig. 9a, b, carbon displayed low activity for HOR, and only a small current flowed between the electrodes when CB and Pt/C were respectively used for HOR and BQRR, resulting in a low overall HQ formation rate. However, when Pt/C was used for HOR, a similar high HQ formation rate was obtained regardless of the identity of the cathode for BQRR (Supplementary Fig. 9a, c, d). Therefore, it can be concluded that the aqueous BQ hydrogenation reaction is composed of electrochemical BQOR over carbon support and HOR over Pt.

Fig. 6: Examination of MPT for BQ hydrogenation over Pt/C.
Fig. 6: Examination of MPT for BQ hydrogenation over Pt/C.
Full size image

a Current dependence of benzoquinone reduction reaction (BQRR) under Ar and hydrogen oxidation reaction (HOR) as a function of the catalyst potential. Comparisons of (b) measured HQ formation rate and (c) operando OCP with the prediction based on MPT. Conditions: WE: 0.75 mg Pt/C on 1 × 1 cm2 on CP; pressure: 101 kPa; RE: Hg/Hg2Cl2 (sat. KCl); CE: carbon rod; 295 K; 0.1 M HClO4; pH 1.2; stirring rate: 450 rpm.

The influence of external bias on BQ hydrogenation performance was also examined with varying PH2 and CBQ. Unity FE of BQRR to HQ was confirmed under Ar from 0.02 V to 0.65 V vs. RHE (Supplementary Fig. 7b). Similarly, as observed for HQ oxidation, the application of a cathodic potential led to an increase in the overall HQ formation rate but resulted in a decrease in the thermal hydrogenation rate (Supplementary Fig. 10). Figure 7 compares the rates predicted by MPT to the separate contributions from thermal- and electro-catalysis making up the total BQ hydrogenation rate. These comparisons reveal that MPT remains valid for HQ hydrogenation with an applied potential when thermal- and electro-catalytic pathways proceed in parallel. Overall, the BQ hydrogenation results also support the hypothesis that an MPT-obeying reaction proceeding via two independent electrochemical half-reactions without reactant competition for active sites cannot be electrochemically promoted for higher thermal activity.

Fig. 7: Prediction of both thermal- and electro-catalytic rate for BQ reduction over Pt/C with an external bias applied.
Fig. 7: Prediction of both thermal- and electro-catalytic rate for BQ reduction over Pt/C with an external bias applied.
Full size image

Comparisons of (a) electrochemical contribution and (b) thermal contribution to the overall BQ hydrogenation rate with prediction when external potentials are applied. WE: 0.75 mg Pt/C on 1 × 1 cm2 on CP; total pressure: 101 kPa; RE: Hg/Hg2Cl2 (sat. KCl); CE: carbon rod; 295 K; 0.1 M HClO4; pH 1.2; stirring rate: 450 rpm.

Enhanced thermal hydroquinone oxidation over platinized Pt foil with anodic bias application

The oxidation of HQ was further examined over a platinized Pt foil catalyst, where the competitive activation of HQ and O2 would take place. Electrochemical deposition of Pt was conducted on the surface of commercial Pt to prepare platinized Pt foil with higher surface area and (electro)catalytic activity, which was supported by electrochemical characterizations (Supplementary Fig. 11), scanning electron microscope (SEM) images (Supplementary Fig. 12) and XRD patterns (Supplementary Fig. 2). With a given O2 partial pressure (Supplementary Fig. 13a), the decrease of CHQ from 10 mM to 4 mM led to negligible changes in HQ oxidation rate and a slight positive shift in OCP, indicating that HQ likely dominated the surface of platinized Pt foil. Further decrease of CHQ to 2 mM caused a slight change in BQ formation rate and a significant shift of OCP to more positive values, which was attributed to the reduced coverage of HQ on the Pt surface due to its low concentration. From another perspective, with a given initial CHQ (Supplementary Fig. 13b), the increase in O2 partial pressure led to an almost linear increase of HQ oxidation rate but a limited change in OCP. This also supports the claim that the Pt surface is dominated by HQ and that there is competitive adsorption between O2 and HQ for active sites. In Fig. 8a, the current dependence of HQOR and ORR over platinized Pt used to predict the BQ formation rate and working potential during overall aerobic oxidation of HQ. The comparisons between the predicted and measured rates and potentials prediction are compiled in Fig. 8b, c. Unlike the case with Pt/C, the absence of carbon led to significant discrepancies in reaction rate and working potential between measurement and prediction. The measured rates were consistently lower than the predicted rates, likely caused by the competitive adsorption of different reactants at the active sites compared to the isolated half-reactions in Fig. 8a. The predicted working potentials were also more positive than the measured OCPs, as shown in Fig. 8c. Since the Pt surface was likely dominated by HQ during the overall aerobic HQ oxidation based on previous analysis, the adsorption of HQ caused a discrepancy of working potential between measurement and prediction here.

Fig. 8: Examination of MPT for aerobic HQ oxidation over platinized Pt foil.
Fig. 8: Examination of MPT for aerobic HQ oxidation over platinized Pt foil.
Full size image

a Current dependence of HQOR under Ar and ORR as a function of the potential of platinized Pt foil. Comparisons of (b) measured BQ formation rate and (c) operando OCP with the prediction based on MPT. Conditions: WE: Platinized Pt foil (1 × 1 cm2) on CP; total pressure: 101 kPa; RE: Hg/Hg2Cl2 (sat. KCl); CE: carbon rod; 295 K; 0.1 M HClO4; pH 1.2; stirring rate: 450 rpm.

The results over platinized Pt foil and Pt/C catalysts further indicate an additional requirement for MPT to be valid for the prediction of thermocatalytic reaction rates. Not only is it required for two completely independent electrochemical half-reactions to exist without non-Faradaic redox elementary step(s)45,47, but the reaction must also proceed with non-competitive activation of reactant species on the catalyst surface. As a further example of this, the ORR activity of Pt/C was observed to decrease significantly when HQ is allowed to adsorb to Pt without the presence of oxygen (Supplementary Fig. 14). Under these conditions, the altered overpotentials required to drive the ORR are expected to cause deviations from MPT.

Furthermore, the effect of an external bias was examined for HQ oxidation over platinized Pt foil, and the resultant (electro)catalytic behavior was found to differ from that of Pt/C. PO2 was first fixed at 100 kPa, and initial CHQ was varied (Fig. 9a–c). The promotional effect of external bias on thermal HQ oxidation rate was quantified by the following equation:

$${{{\rm{Promotion}}}}=\frac{{r}_{{{{\rm{t}}}}}}{{r}_{0}}\times 100\%$$
(8)

where r0 is the aerobic HQ oxidation rate at OCP under the given reactant concentrations and pressures. It was found that with an initial HQ concentration of 2 mM, applying anodic potentials led to the increase of total HQ oxidation rate but the decrease of thermocatalytic rate (promotion <100%). It was noticed that the overall rate here was close to the diffusion-limited value (ca. 0.65 μmol min−1 based on Fig. 8a), which might conceal the real effects of the external bias. When CHQ was increased to 4 mM, the thermocatalytic rate exhibited a volcanic behavior as a function of catalyst potential with a maximum promotion of ~140% (Fig. 9b). Since the reaction order for aerobic HQ oxidation over platinized Pt foil was nearly 0 for CHQ between 4 and 10 mM (Supplementary Fig. 13a), a similar trend was observed for CHQ = 10 mM (Fig. 9c). The influence of PO2 is further analyzed based on Fig. 9c–e. Generally, the enhancement of thermal HQ oxidation as a function of catalyst potential retained its volcanic behavior, ranging from ~150%, ~180%, and 140% at PO2 values of 50 kPa, 70 kPa and 100 kPa, respectively. On one hand, increased PO2 improved its competitiveness against HQ, which favored a higher thermal contribution to the overall HQ oxidation. On the other hand, aerobic HQ oxidation shows a positive reaction order in PO2 (Supplementary Fig. 13b), resulting in higher OCP reaction rates when PO2 was increased. This meant that the promotional effect was lower when calculated as a percentage. Regardless of the extent of the promotional effect, thermal HQ oxidation over platinized Pt foil was enhanced by external bias.

Fig. 9: HQ oxidation performance over platinized Pt foil with applied external potentials.
Fig. 9: HQ oxidation performance over platinized Pt foil with applied external potentials.
Full size image

The reactant concentration/partial pressure: (a) 100 kPa O2 + 2 mM HQ; (b) 100 kPa O2 + 4 mM HQ; (c) 100 kPa O2 + 10 mM HQ; (d) 50 kPa O2 + 10 mM HQ; (e) 70 kPa O2 + 10 mM HQ. Conditions: WE: Platinized Pt foil (1 × 1 cm2) on CP; pressure: 101 kPa; RE: Hg/Hg2Cl2 (sat. KCl); CE: carbon rod; 295 K; 0.1 M HClO4; pH 1.2; stirring rate: 450 rpm.

The external potential applied to the platinized Pt alters the coverage of adsorbates and overpotentials for electrochemical half-reactions. As seen in Fig. 8a, adsorption of HQ and O2 on platinized Pt in isolation led to OCPs of ~0.65 V and ~0.90 V vs. RHE, respectively. Consistent with previous investigation using infrared spectroscopy48, an application of anodic potential lowered the coverage of HQ on Pt and created more active sites for O2 to adsorb, perturbing thermocatalytic rate. The external applied potential also changed the overpotentials of the electrochemical half-reactions; applying an anodic potential increases the overpotential for electrochemical HQOR and reduces it for ORR. As mentioned previously, the thermal contribution to the overall rate arises from the balanced ORR and HQOR rates. When the anodic potential was applied, HQOR was enhanced while ORR was suppressed electrocatalytically, leading to a high overall HQ oxidation rate but a lower thermal contribution (consisting only of the balanced rate between the two half reactions). In other words, electrocatalytic rate surpasses thermal catalysis, leading to the volcano trend observed for thermal contribution22. The above experiments demonstrate that when a thermocatalytic reaction involves competitive or inhibitory adsorption of reactants and deviates from the prediction of MPT, there may be a chance to enhance the thermocatalytic rate by optimizing the adsorption of species on the catalyst surface with external potential application22,31.

Discussion

The thermocatalytic oxidation/hydrogenation of the HQ/BQ redox couple over Pt/C was thoroughly investigated utilizing electrochemical methods. The MPT provided a precise prediction of the thermocatalytic rate as well as the working potential at open-circuit under given reaction conditions. It was shown that thermocatalytic oxidation/hydrogenation of HQ/BQ consists of two independent electrochemical half-reactions, each catalyzed by different active sites: the ORR/HOR over Pt and HQOR/BQRR over the carbon support. The impacts of an externally applied potential on HQ oxidation and BQ hydrogenation were explored, revealing that static anodic and cathodic potentials increased the formation rates of BQ and HQ, respectively. However, under such conditions where thermal- and electro-catalysis occurred in parallel, the thermal contribution to the overall rate decreased. The distribution of thermal- and electro-catalysis strictly adhered to the voltammetric properties of the half-reactions involved, suggesting the applicability of MPT for catalysis even when an external bias was applied. In contrast, when a platinized Pt foil was utilized as the catalyst for HQ oxidation, the absence of a carbon support led to dominant adsorption of HQ on Pt. Consequently, discrepancies between experimental measurements and MPT-predicted catalytic rate and working potential were found. The application of an appropriate anodic overpotential to the platinized Pt foil surprisingly accelerated the thermal HQ oxidation rate, possibly by suppressing the coverage of HQ on Pt and facilitating O2 adsorption.

The study suggests that for MPT-obeying thermocatalytic reactions involving two independent coupled half-reactions, the promotion of thermal catalysis through static potential applications is likely improbable. Where a discrepancy exists between the overall catalytic reaction and MPT prediction, the thermocatalytic rate may be enhanced by optimizing the coverage of the adsorbate on the catalyst surface with external bias. This investigation using the HQ/BQ couple as a model serves to bridge a gap between the electrochemical promotion of thermal catalysis and MPT predictions.

Methods

Materials

Carbon-supported Pt with a mass fraction of 37.5% was obtained from Tanaka Kikinzoku Kogyo. NafionTM 117 proton-exchange membrane and Toray carbon paper 060 value pack (wet proofed) were purchased from Fuel Cell Store. Vulcan® XC-72 carbon black was procured from Cabot Corporation. The Pt foil (99.95%, 0.05 mm thickness) was purchased from Nilaco Corporation. NafionTM 117 containing solution (~5% in a mixture of lower aliphatic alcohols and water), sulfuric acid (95.0 ~ 98.0%, ACS reagent), ethanol (99.5%, JIS special grade), hydroquinone (SAJ first grade), and acetone (SAJ first grade) were purchased from Sigma-Aldrich. Perchloric acid (70%) maleic acid (99.0 + %), D2O (99.8%), nitric acid (1.42), hydrogen hexachloroplatinate(IV) hexahydrate (98.5 + %), and p-benzoquinone (98.0 + %) purchased from Fujifilm Wako Chemicals. All chemicals were used without further purification. Ultrapure water with a resistivity of 18.2 MΩ cm was obtained from a water purification system (Purelab Flex 3, ELGA LabWater) and used for all experimental procedures. Gas cylinders of pure CO (99.95%), pure H2 (99.99999%), pure Ar (99.9999%), and pure O2 (99.99995%) were purchased from Taiyo Nippon Sanso Corporation.

Catalyst characterization

The powder XRD patterns were measured in the air using Cu-Kα radiation on a Rigaku MinFlex 600. The signal was collected over an angular range of 20–80° (2θ) with a scan rate of 10° min−1. The transmission electron microscopy (TEM) image was obtained using a JEOL JEM-2100 microscope at an accelerating voltage of 200 kV. SEM images were taken with a JEOL JSM-IT800 setup at a beam voltage of 2 kV.

Catalytic measurement

The CP was cut into a size of 1 × 1 cm2, subsequently immersed into 0.2 M H2SO4, ethanol and acetone for precleaning and then dried in air. A catalyst ink of 5 mg ml−1 was prepared by dispersing CB or Pt/C particles and 0.5 wt% Nafion 117 dispersed in ethanol. A WE was fabricated by drop-casting the catalyst ink on the CP with a target amount of 0.75 mg cm−2 for metal/C or 0.45 mg cm−2 for CB. A two-compartment cell (Supplementary Fig. 3a) separated by one piece of Nafion membrane was employed as the reactor, where the WE together with a KCl-saturated Hg/Hg2Cl2 reference electrode (RE) was placed in one chamber and a carbon rod as CE was placed in the other chamber unless otherwise stated. The reactor containing 0.1 M HClO4 was first placed under the Ar or (Ar-diluted) O2 (for HQ oxidation) or H2 (for BQ hydrogenation) for stabilization, concentrated aqueous HQ or BQ solution (containing 0.1 M HClO4) was injected into the reaction chamber to start the catalysis with target initial concentrations of reactants. After certain reaction time (12 min for HQ oxidation over Pt/C or 8 min for BQ hydrogenation over Pt/C or 16 min HQ oxidation over platinized Pt foil; conversion less than 15%), the product was quantified by 1H nuclear magnetic resonance (NMR) spectroscopy. Specifically, a 500 μl liquid sample was mixed with 100 μl D2O and 20 μl 0.1 M maleic acid (internal standard), and the resultant solution was tested on a JEOL ECS 400 MHz NMR spectrometer (calibration curve in Supplementary Fig. 15). The overall product formation rate, rtotal in μmol min−1, was calculated based on the equation below:

$${r}_{{{{\rm{total}}}}}=\frac{{cV}}{t}$$
(9)

where c in mM was the product concentration, V in ml (typically 25 ml) was the liquid volume in the reaction chamber, and t in min was the reaction time. Operando OCP was monitored with a VMP3 potentiostat (BioLogic) and presented on the RHE scale. A time average of OCP was taken for every catalyst test. External bias was applied to the WE during catalysis by chronopotentiometry technique, and the time average of the potential was used for further analyses. Any gas product was analyzed by a Micro GC Fusion gas analyzer 2-module chassis (Inficon) equipped with thermal conductivity detectors (TCDs).

Electrochemical characterization

The series resistance of the cell was measured with potentiostatic electrochemical impedance spectroscopy (PEIS) technique, and all the potential values were fully corrected based on the IR drop (\({E}_{{{{\rm{corrected}}}}}={E}_{{{{\rm{measured}}}}}-{IR}\)). The current dependence of electrochemical HOR, ORR, HQOR, and BQRR on catalyst potential was determined by applying static potentials to the catalyst, incrementing by 0.05 V (3 min for each potential). The FE of the electrocatalytic process was calculated based on the following equation:

$${{{\rm{FE}}}}=\frac{{cV}\times {10}^{-6}}{Q/{nF}}\times 100\%$$
(10)

where Q was the total charge amount in C during the reaction recorded by the potentiostat.

Preparation of platinized Pt foil49,50

A piece of Pt foil was also cut into a size of 1 × 1 cm2 and immersed in 10% nitric acid for 15 min and washed with ultrapure water. The Pt foil was then placed into an aqueous solution containing 0.008 M H2PtCl6 and 0.023 M HClO4, and a cathodic potential of −0.1 V vs. Hg/Hg2Cl2 (sat. KCl) was applied to it for 10 min. The resultant platinized Pt foil was again washed with ultrapure water.