Improving Catalytic Activity towards the Direct Synthesis of H 2 O 2 through Cu Incorporation into AuPd Catalysts

: With a focus on catalysts prepared by an excess-chloride wet impregnation procedure and supported on the zeolite ZSM-5(30), the introduction of low concentrations of tertiary base metals, in particular Cu, into supported AuPd nanoparticles can be observed to enhance catalytic activity towards the direct synthesis of H 2 O 2 . Indeed the optimal catalyst formulation (1%AuPd (0.975) Cu (0.025) /ZSM-5) is able to achieve rates of H 2 O 2 synthesis (115 mol H 2 O 2 kg cat − 1 h − 1 ) approximately 1.7 times that of the bi-metallic analogue (69 mol H 2 O 2 kg cat − 1 h − 1 ) and rival that previously reported over comparable materials which use Pt as a dopant. Notably, the introduction of Cu at higher loadings results in an inhibition of performance. Detailed analysis by CO-DRFITS and XPS reveals that the improved performance observed over the optimal catalyst can be attributed to the electronic modiﬁcation of the Pd species and the formation of domains of a mixed Pd 2+ /Pd 0 oxidation state as well as structural changed within the nanoalloy.

Thi s v e r sio n is b ei n g m a d e a v ail a bl e in a c c o r d a n c e wit h p u blis h e r p olici e s. S e e h t t p://o r c a . cf. a c. u k/ p olici e s. h t ml fo r u s a g e p olici e s. Co py ri g h t a n d m o r al ri g h t s fo r p u blic a tio n s m a d e a v ail a bl e in ORCA a r e r e t ai n e d by t h e c o py ri g h t h ol d e r s .

Introduction
The direct synthesis of H 2 O 2 from molecular H 2 and O 2 (Scheme 1) represents an attractive alternative to the current means of large-scale production of this environmentally benign, powerful oxidant, the anthraquinone oxidation (AO) process. Indeed, the direct approach would allow for the production of appropriate concentrations of H 2 O 2 at the point of final use, avoiding the substantial economic and environmental drawbacks associated with the industrial route. Due to production costs, H 2 O 2 production via the AO process is typically centralised, with H 2 O 2 transported at concentrations in excess of that required by the end-user [1]. The subsequent dilution of the oxidant prior to use effectively wastes a significant amount of energy associated with the initial purification and concentration steps [2]. Additionally, H 2 O 2 is relatively unstable, decomposing readily to H 2 O in the presence of mild temperatures or weak bases and, as such, requires acid stabilizers to prolong its shelf-life [3], which results in complex product streams and can deleteriously affect the reactor lifetime [4]. These cumulative drawbacks associated with off-site H 2 O 2 production pass on significant costs to the end user, which would be greatly reduced or removed altogether via a direct synthesis approach to H 2 O 2 production. In particular, the direct route may find the greatest application for oxidative transformations where the synthesised H 2 O 2 is readily utilised for chemical valorisation or generated in situ [5][6][7].
A Langmuir-Hinshelwood mechanism involving the successive hydrogenation of molecular O 2 has been often proposed for the direct synthesis reaction [8,9]. However, in recent years Flaherty and co-workers have advanced an alternative, non-Langmuirian mechanism [10], which involves a water-mediated proton-electron transfer and have further reported the role of protic solvents in the formation of surface-bound intermediates that shuttle both the protons and electrons to active sites [11]. Indeed, detailed theoretical studies have further demonstrated that energy barriers associated with a solvent-mediated protonation of adsorbed O 2 are not prohibitive and, indeed, are as low as O 2 hydrogenation [12]. Pd-based catalysts have been well studied for their application in the direct synthesis reaction [13][14][15] since the first patent was granted to Henkel and Weber in 1914 [16]. However, a major challenge associated with catalytic selectivity has prevented the development of an industrial-scale direct synthesis process [17,18]. This can be understood as the formation of H 2 O is thermodynamically more favourable than that of H 2 O 2 , with H 2 O formation driven via combustion or through the subsequent degradation of H 2 O 2 (via decomposition and hydrogenation pathways).
Measures to improve catalytic selectivity have often focused on the introduction of secondary metals into supported Pd catalysts, with AuPd systems being perhaps the most extensively studied [19][20][21]. In recent years, significant attention has been placed on the alloying of Pd with a range of earth-abundant metals [22][23][24][25][26][27]. Further investigations have focused on the introduction of dopant levels of precious metals, such as Pt, into supported Pd [28][29][30][31][32][33] and AuPd [34][35][36][37] catalysts. The resulting improved catalytic activity towards H 2 O 2 production is often attributed to a combination of the electronic promotion of Pd and the isolation of contiguous Pd domains, widely considered to be key in promoting the cleavage of O-O bonds (in *O 2 , *H 2 O 2 , or *OOH) and the resultant formation of H 2 O [38][39][40].
The use of zeolitic and zeo-type materials as catalyst supports for use in the direct synthesis of H 2 O 2 has received significant attention, with such catalysts typically offering improved activity and selectivity compared to oxide-supported analogues [5,36,[41][42][43]. This has often been attributed to the improved dispersion of metal species and the acidic nature of the support materials, with supports of low isoelectric points (i.e., the pH at which the surface has zero net charges and an indication of catalyst acidity/basicity) and are reported to offer enhanced catalytic performance compared to those with a high isoelectric point [20].
Recently, we have demonstrated that significant improvements in catalytic performances can be achieved through the introduction of low concentrations of base metals into AuPd nanoalloys when supported on TiO 2 (P25) [44]. Indeed, the catalysts developed within this earlier work offered comparable H 2 O 2 synthesis rates to those previously observed through Pt incorporation while avoiding the additional cost associated with the precious metal dopant [44]. With these earlier studies in mind, we now investigate the efficacy of base metal-incorporated AuPd catalysts supported on the zeolite ZSM-5 (30) towards the direct synthesis of H 2 O 2 .
A series of bi-and tri-metallic 1%AuPdX/ZSM-5 (X = Cu, Ni, Zn) catalysts were prepared by an excess chloride wet co-impregnation procedure, based on the methodology previously reported in the literature, which has been shown to result in the enhanced dispersion of precious metals, in particular Au, when compared to conventional wet coimpregnation methodologies [45]. The procedure to produce 1%AuPd (0.975) Cu (0.025) /ZSM-5 (1 g) is outlined below where the total metal loading was 1 wt.%, the combined weight loading of Au and Pd was 0.975 wt.%, and that of Cu was 0.025 wt.%, and in all cases the Au:Pd ratio was 1:1 (mol/mol). A similar methodology to that outlined below was utilised for all catalysts investigated, with the exact quantities of metal precursor used to synthesise the key catalysts used within this work reported in Table S1.
Aqueous solutions of HAuCl 4 .3H 2 O (0.322 mL, 12.25 mgmL −1 , Strem Chemicals), PdCl 2 (0.356 mL, 6 mgmL −1 , 0.58 M HCl, Sigma Aldrich, Burlington, MA, USA), and CuCl 2 (106 µL, 2.36 mgmL −1 , Sigma Aldrich) were mixed in a 50 mL round bottom flask and heated to 60 • C with stirring (1000 rpm) in a thermostatically controlled oil bath, with the total volume fixed to 16 mL using H 2 O (HPLC grade). Upon reaching 65 • C, ZSM-5 (0.99 g, SiO 2 :Al 2 O 3 = 30:1, Alfa Aesar) was added over the course of 10 min with constant stirring. The resultant slurry was stirred at 60 • C for a further 15 min, and following this, the temperature was raised to 95 • C for 16 h to allow for the complete evaporation of water. The resulting solid was ground prior to heat treatment in a reductive atmosphere (5%H 2 /Ar, 400 • C, 4 h, 10 • Cmin −1 ).  [45]. In regard to the role of the CO 2 diluent, this was found to act as an in-situ promoter of H 2 O 2 stability through its dissolution in the reaction solution and the formation of carbonic acid. We have previously reported that the use of the CO 2 diluent has a comparable promotive effect to that observed when acidifying the reaction solution to a pH of 4 using HNO 3 [46].

Direct Synthesis of H 2 O 2
Hydrogen peroxide synthesis was evaluated using a Parr Instruments stainless steel autoclave with a nominal volume of 100 mL, equipped with a PTFE liner so that the total volume was reduced to 66 mL and a maximum working pressure of 2000 psi. To test each catalyst for H 2 O 2 synthesis, the autoclave liner was charged with a catalyst (0.01 g), solvent (methanol (5.6 g, HPLC grade, Fischer Scientific, Waltham, MA, USA), and H 2 O (2.9 g, HPLC grade, Fischer Scientific)). The charged autoclave was then purged three times with 5%H 2 /CO 2 (100 psi) before filling with 5%H 2 /CO 2 to a pressure of 420 psi, followed by the addition of 25%O 2 /CO 2 (160 psi), with the pressure of 5%H 2 /CO 2 and 25%O 2 /CO 2 given as gauge pressures. The reactor was not continually fed with reactant gas. The reaction was conducted at a temperature of 2 • C for 0.5 h with stirring (1200 rpm). The H 2 O 2 productivity was determined by titrating aliquots of the final solution after the reaction with acidified Ce(SO 4 ) 2 (0.0085 M) in the presence of a ferroin indicator. Catalyst productivities are reported as mol H 2 O 2 kg cat −1 h −1 . The catalytic conversion of H 2 and selectivity towards H 2 O 2 were determined using a Varian 3800 GC fitted with TCD and equipped with a Porapak Q column. (1)) and H 2 O 2 selectivity (Equation (2)) are defined as follows:

H 2 conversion (Equation
The total autoclave capacity was determined via water displacement to allow for the accurate determination of H 2 conversion and H 2 O 2 selectivity. When equipped with the PTFE liner, the total volume of an unfilled autoclave was determined to be 93 mL, which included all available gaseous space within the autoclave.

Gas Replacement Experiments for the Direct Synthesis of H 2 O 2
An identical procedure to that outlined above for the direct synthesis reaction was followed for a reaction time of 0.5 h. After this, stirring was stopped, and the reactant gas mixture was vented prior to replacement with the standard pressures of 5% H 2 /CO 2 (420 psi) and 25% O 2 /CO 2 (160 psi). The reaction mixture was then stirred (1200 rpm) for a further 0.5 h. To collect a series of data points, as in the case of Figure 5, it should be noted that individual experiments were carried out, and the reactant mixture was not sampled online.

Catalyst Reusability in the Direct Synthesis and Degradation of H 2 O 2
In order to determine catalyst reusability, a similar procedure to that outlined above for the direct synthesis of H 2 O 2 was followed utilising 0.05 g of catalyst. Following the initial test, the catalyst was recovered by filtration and dried (30 • C, 16 h, under vacuum); from the recovered catalyst sample 0.01 g and was used to conduct a standard H 2 O 2 synthesis or degradation test.

Degradation of H 2 O 2
Catalytic activity towards H 2 O 2 degradation was determined in a similar manner to the direct synthesis activity of a catalyst. The autoclave liner was charged with a solvent (methanol (5.6 g, HPLC grade, Fischer Scientific), H 2 O (2.9 g, HPLC grade, Fischer Scientific)), and H 2 O 2 (50 wt. % 0.69 g, Sigma Aldrich), with the resultant solvent composition equivalent to a 4 wt. % H 2 O 2 solution. From the solution, two 0.05 g aliquots were removed and titrated with acidified Ce(SO 4 ) 2 solution using ferroin as an indicator to determine an accurate concentration of H 2 O 2 at the start of the reaction. Subsequently, a catalyst (0.01 g) was added to the reaction media, and the autoclave was purged with 5%H 2 /CO 2 (100 psi) prior to being pressurised with 5%H 2 /CO 2 (420 psi). The reaction medium was cooled to a temperature of 2 • C prior to stirring (1200 rpm) for 0.5 h. After the reaction was complete, the catalyst was removed from the reaction mixture, and two 0.05 g aliquots were titrated against the acidified Ce(SO 4 ) 2 solution using ferroin as an indicator. The degradation activity is reported as mol H 2 O 2 kg cat −1 h −1 .

Note 3.
In all cases, the reactor temperature was controlled using a HAAKE K50 bath/circulator and an appropriate coolant. The reactor temperature was maintained at 2 ± 0.2 • C throughout the course of the H 2 O 2 synthesis and degradation reaction.
In all cases, the reactions were run multiple times, over multiple batches of catalysts, with the data being presented as an average of these experiments. The catalytic activity toward the direct synthesis and subsequent degradation of H 2 O 2 was found to be consistent to within ±3% on the basis of multiple reactions.

Characterisation
A Thermo Scientific K-Alpha + photoelectron spectrometer was used to collect XP spectra utilising a micro-focused monochromatic Al K α X-ray source operating at 72 W.
Samples were pressed into a copper holder and analysed using the 400 µm spot mode at pass energies of 40 and 150 eV for high-resolution and survey spectra, respectively. Charge compensation was performed using a combination of low-energy electrons and argon ions, which resulted in a C(1s) binding energy of 284.8 eV for the adventitious carbon present in all the samples and all samples also showed a constant Ti(2p 3/2 ) of 458.5 eV. All data were processed using CasaXPS v2.3.24 (Casa Software Ltd., Teignmouth, UK) with a Shirley background, Scofield sensitivity factors, and an electron energy dependence of −0.6, as recommended by the manufacturer. Peak fits were performed using a combination of Voigttype functions and models derived from the bulk reference samples where appropriate.
The bulk structure of the catalysts was determined by powder X-ray diffraction using a (θ-θ) PANalytical X'pert Pro powder diffractometer with a Cu K α radiation source, operating at between 40 keV and 40 mA. Standard analysis was carried out using a 40 min run with a backfilled sample, between 2θ values of 5 and 75 • . Phase identification was carried out using the International Centre for Diffraction Data (ICDD). Figure S1 (and accompanying text) with no reflections associated with active metals, indicative of the relatively low total loading and high dispersion of the immobilised metals.

Note 4. X-ray diffractograms of key as-prepared catalysts are reported in
Transmission electron microscopy (TEM) was performed on a JEOL JEM-2100 (Tokyo, Japan) operating at 200 kV. Samples were prepared through their dispersion in ethanol by sonication, and they were deposited on 300 mesh copper grids coated with holey carbon film. Energy dispersive X-ray spectroscopy (XEDS) was performed using an Oxford Instruments (Abingdon, UK) X-Max N 80 detector, and the data analysed used Aztec software (Abingdon, UK). Aberration corrected scanning transmission electron microscopy (AC-STEM) was performed using a probe-corrected Hitachi (Brisbane, Australia) HF5000 S/TEM, operating at 200 kV. The instrument was equipped with bright field (BF), high angle annular dark field (HAADF), and secondary electron (SE) detectors for high spatial resolution STEM imaging experiments. This microscope was also equipped with a secondary electron detector and dual Oxford Instruments (Abingdon, UK) XEDS detectors (2 × 100 mm 2 ) with a total collection angle of 2.02 sr.
Total metal leaching from the supported catalyst was quantified via inductively coupled plasma mass spectrometry (ICP-MS). Post-reaction solutions were analysed using an Agilent (Santa Clara, CA, USA) 7900 ICP-MS equipped with an I-AS auto-sampler. All samples were diluted by a factor of 10 using HPLC grade H 2 O (1%HNO 3 and 0.5% HCl matrix). All calibrants were matrix matched and measured against a five-point calibration using certified reference materials purchased from Perkin Elmer and certified internal standards acquired from Agilent.
Fourier-transform infrared spectroscopy (FTIR) was carried out with a Bruker (Hanau, Germany) Tensor 27 spectrometer fitted with a HgCdTe (MCT) detector and was operated with OPUS software (Ettinger, Germany). Figure S2 (and accompanying text) and indicates no discernible changes in the structure of the HZSM-5 support upon metal immobilisation and exposure to a reductive heat treatment. N 2 isotherms were collected on a Micromeritics 3-Flex. Samples (ca. 0.070 g) were degassed (350 • C, 9 h) prior to analysis. Analyses were carried out at 77 K, with P 0 measured continuously. Free space was measured post-analysis with He. Data analyses were carried out using the Micromeritics 3-Flex software with the non-local density functional theory (NLDFT), Tarazona model. Table S2 and Figure S3. The immobilisation of active metals can be seen to lead to a general decrease in both the total surface area and pore volume in comparison to the bare ZSM-5 support. This is ascribed to the deposition of metal nanoparticles inside the zeolitic pore structure.

Results and Discussion
The introduction of small concentrations of precious dopants, in particular Pt, [38,39] into the supported AuPd nanoalloys has been extensively reported to offer improved catalytic activity towards the direct synthesis of H 2 O 2 , when compared to the bimetallic analogue. We recently demonstrated that comparable enhancements in performance could result from the incorporation of dopant concentrations of base meals into AuPd nanoalloys [44]. In keeping with this earlier work, our initial investigations identified the promotive effect that can result from the introduction of Cu, Ni, and Zn at low concentrations (0.025 wt.%) into a 1%AuPd (1.00) /ZSM-5 catalyst (Figure 1). In particular, the introduction of Cu, which is known to be readily incorporated into AuPd alloys [47], was observed to significantly increase activity towards H 2 O 2 synthesis, with this metric approximately 1.7 times greater (115 mol H 2 O 2 kg cat −1 h −1 ), than that observed for the bi-metallic 1%AuPd ( [48] or Pd-only [49] catalysts, can inhibit catalytic activity, with DFT studies indicating that the formation of the intermediate hydroperoxyl species (OOH*) and subsequently H 2 O 2 is thermodynamically unfavoured over Cu-containing precious metal surfaces [50]. However, notably, these prior works have focused on the incorporation of Cu at much higher concentrations than that utilised within this study. In comparison, the introduction of Ni (81 mol H 2 O 2 kg cat −1 h −1 ) and Zn (77 mol H 2 O 2 kg cat −1 h −1 ) resulted in only a minor improvement in the catalyst performance compared to the bi-metallic AuPd analogue, although the improved selectivity of the 1%AuPd (0.975) Ni (0.025) /ZSM-5 catalyst is noteworthy, with H 2 O 2 degradation rates (281 mol H 2 O 2 kg cat −1 h −1 ) significantly lower than that observed over the 1%AuPd (1.00) /ZSM-5 catalyst (320 mol H 2 O 2 kg cat −1 h −1 ) or, indeed, the other trimetallic formulations. The enhanced performance of the 1%AuPd (0.975) Cu (0.025) /ZSM-5 catalyst is further evidenced by the comparison of initial reaction rates, where the contribution of competitive H 2 O 2 degradation pathways is considered to be negligible (Table S3). An evaluation of the as-prepared 1%AuPd (0.975) X (0.025) /ZSM-5 catalysts by XPS can be seen in Figure 2 (additional data reported in Table S4). Interestingly, despite exposure to a high-temperature reductive heat treatment (5%H 2 /Ar, 400 • C, 4 h, 10 • Cmin −1 ), the 1%AuPd (1.00) /ZSM-5 catalyst was found to consist of a relatively high proportion of Pd 2+ . Such an observation is in keeping with our previous investigations into AuPd systems, where the introduction of Au has been found to modify Pd speciation [20]. Upon the introduction of low quantities of Ni, Cu, and Zn, a significant shift in Pd speciation was observed, towards Pd 2+ , with the formation of mixed domains of the Pd oxidation state, which is well known to offer improved activity compared to Pd 0 or Pd 2+ rich analogues [51]. The shift in the Pd oxidation state towards Pd 2+ upon the introduction of Ni was found to be the greatest, which aligned well with the observed selectivity of the 1%AuPd (0.975) Ni (0.025) /ZSM-5 catalyst. However, it should be noted that the Pd speciation of the fresh catalyst is likely to be not representative of those under direct synthesis reaction conditions. With the improved activity of the 1%AuPd (0.975) Cu (0.025) /ZSM-5 catalyst towards H 2 O 2 production established, we were subsequently motivated to determine the effect of Cu loading on catalytic activity while maintaining the total metal loading at 1 wt.% ( Figure 3A,B, with initial reaction rates reported in Table S5). The introduction of low concentrations of Cu (< 0.025 wt.%) was observed to significantly increase the catalytic activity towards both the direct synthesis and subsequent degradation of H 2 O 2 , compared to the bimetallic AuPd parent material. However, both these metrics decreased considerably at higher loadings of Cu (75 and 287 mol H 2 O 2 kg cat −1 h −1 , respectively, for H 2 O 2 direct synthesis and degradation pathways at a Cu loading of 0.037%, which is equivalent to 3.7% of the total metal loading). This is in keeping with previous works, which demonstrated a deleterious effect on performance with the introduction of high loadings of Cu into precious metal nanoparticles [48,50] and suggested a high sensitivity towards tertiary metal content. While the evaluation of catalytic activity towards H 2 O 2 synthesis alone ( Figure 3A) may suggest that there is very little difference in the performance over a range of Cu loadings (H 2 O 2 synthesis rates between 111 and 116 mol H 2 O 2 kg cat −1 h −1 observed for Cu loadings of 0.012-0.025 wt.%), the determination of H 2 conversion rates and H 2 O 2 selectivity indicates that a substantial reduction in catalytic selectivity towards H 2 O 2 coincides with the introduction of Cu ( Figure 3B), which would align with determination of trends in H 2 O 2 degradation activity ( Figure 3A). Indeed, these observations imply that the enhanced activity of the Cu-containing catalysts is associated with increased reactivity (i.e., the rate of H 2 conversion) rather than H 2 O 2 selectivity. However, it is important to consider that such evaluations are not made at comparable rates of H 2 conversion and, notably, the high H 2 selectivity of the 1%AuPd (1.00) /ZSM-5 catalyst (81%) can be related to the low rates of conversion (10%) observed. With the introduction of Cu at concentrations greater than 0.025 wt.%, a substantial decrease in H 2 conversion rates was observed, which correlates well with the observed loss in catalytic activity towards both the direct synthesis and subsequent degradation of H 2 O 2 .
--−1  With our XPS analysis revealing a modification in the Pd oxidation state as a result of the incorporation of dopant metals (Figure 2), we were subsequently motivated to investigate the 1%AuPdCu/ZSM-5 catalytic series via CO-DRIFTS (Figure 4). CO-DRIFTS is a technique that has been extensively utilised to probe the surface of supported precious metal catalysts [51][52][53][54]. For each catalyst, spectra were measured in the 1750-2150 cm −1 range, which contains the stretching modes associated with the CO adsorbed on Pd and Au surfaces. The DRIFTS spectra of all the catalysts were dominated by Pd-CO bands. The peak observed at approximately 2080 cm −1 represents the CO bound in a linear manner to low co-ordination Pd sites (i.e., corner or edge sites), while the broad feature, centred around 1950 cm −1 , represents the bi-and tri-dentate bridging modes of CO on Pd [55]. Upon the introduction of small concentrations of Cu into the AuPd nanoalloy, a small blue shift in the band relates to the linearly bonded CO on Pd, which can be observed. This aligns well with previous investigations by Wilson et al. into the formation of AuPd alloys [51]. In particular, such a shift can be attributed to the segregation of Pd at the nanoparticle surface and a corresponding occupation of lower coordination sites. It is, therefore, possible to propose that the introduction of Cu into AuPd nanoalloys at low concentrations results in a similar modification of the nanoparticle composition. With the evident improvement upon the incorporation of Cu into a supported 1%AuPd (1.00) /ZSM-5 catalyst, we subsequently set out to further contrast the catalytic performance of the optimal catalyst and the bimetallic AuPd analogue. Time-on-line studies comparing H 2 O 2 synthesis rates are reported in Figure 5A, where a significant difference in catalytic performance is observed. Indeed, the enhanced reactivity of the 1%AuPd (0.975) Cu (0.025) /ZSM-5 catalyst is clear, achieving H 2 O 2 concentrations (0.35 wt.%) far greater than that of the AuPd analogue (0.21 wt.%) over a 1 h H 2 O 2 synthesis reaction. Further investigation of catalytic performance over several successive H 2 O 2 synthesis reactions can be seen in Figure 5B, where a marked enhancement in H 2 O 2 concentration can be observed over the 1%AuPd (0.975) Cu (0.025) /ZSM-5 catalyst (0.60 wt.%) compared to that achieved by the 1%AuPd (1.00)/ ZSM-5 catalyst (0.27 wt.%). Notably, the concen-tration of H 2 O 2 achieved over the AuPdCu catalyst was found to be comparable to that achieved when utilising identical concentrations of Pt as a catalytic promoter for AuPd nanoalloys [44]. Numerous works have elucidated the dependence between catalytic performance towards H 2 O 2 synthesis and particle size, with studies by Tian et al., in particular, revealing that particle size in the sub-nanometre range is crucial for achieving optimal catalytic performance, at least in the case of monometallic Pd catalysts [56]. Comparisons of the mean particle size of the as-prepared 1%AuPd (1.00)) /ZSM-5 and 1%AuPd (0.975) Cu (0.025) /ZSM-5 catalysts, as determined from the bright field transmission electron micrographs presented in Figure S4, are reported in Table 1 with a negligible variation in particle size observed between the AuPd and optimal AuPdCu catalysts (3.7-4.1 nm). As such, it is possible to conclude that the enhanced catalytic activity towards H 2 O 2 synthesis achieved through the introduction of Cu into AuPd nanoparticles is not associated with increased metal dispersion. Rather, it can be considered that the electronic modification of the Pd species is a result of dopant introduction, as indicated by our CO-DRIFTS ( Figure 4) and XPS analysis (Figure 2), as well as possible structural changes which were indicated by our CO-DRIFTS analysis ( Figure 4)and are responsible for the observed reactivity improvements. For any heterogeneous catalyst operating in a liquid phase reaction, the possibility of catalyst deactivation via the leaching of supported metals and the resultant homogeneous contribution to the observed catalytic activity is a major concern. This is particularly true given the ability of colloidal Pd to catalyse the direct synthesis reaction [57]. It was found that for both the AuPd and AuPdCu catalysts, catalytic activity toward H 2 O 2 production decreased upon second use (Table 2). However, the ICP-MS analysis of H 2 O 2 reaction solutions ( Table 2) and TEM analysis of spent materials ( Figure S4) indicated that such a loss in catalyst performance could not be attributed to either the leaching of active metals or nanoparticle agglomeration, with negligible levels of active metals detected through the analysis of H 2 O 2 synthesis reaction solutions and the mean particle size determined to be comparable for both the as-prepared and used materials. Similar observations have been recently reported for AuPd-supported catalysts prepared by an identical excess-chloride impregnation procedure, with the loss in catalytic performance upon use found to be associated with a significant loss of surface Cl species, a known promoter of activity in the H 2 O 2 synthesis reaction [21]. With these earlier observations in mind, we set out to determine the extent of surface Cl loss (if any) after use in the H 2 O 2 direct synthesis reaction via XPS ( Figure S5). Interestingly, negligible concentrations of Cl species were observed in either the fresh or used AuPd and AuPdCu catalysts. This is in stark contrast to our earlier investigations into AuPd/TiO 2 catalysts prepared via an analogous synthesis technique and possibly highlights the key role of the support in retaining halide species [21]. Regardless, such an observation excludes the possibility of Cl loss as the cause for the observed loss in catalytic performance upon reuse. However, further investigation by XPS ( Figures S6 and S7) does reveal a modification in the Au: Pd ratio for both catalysts after use, which may be indicative of nanoalloy restructuring and perhaps, more importantly, a total shift in the Pd speciation towards Pd 0 , which can be attributed to the presence of H 2 within the reaction. As such, it is possible to conclude that while the catalytic materials developed within this work represent a promising basis for future study, there is still a need to address stability concerns, particularly around Pd speciation.

Conclusions
The introduction of low concentrations of earth-abundant metals (Ni, Cu, Zn) into supported AuPd nanoparticles has been demonstrated to improve catalytic activity towards the direct synthesis of H 2 O 2 , with the inclusion of Cu in particular, found to offer an enhancement compared to that previously reported upon the use of Pt as a promoter for AuPd nanoalloys. Indeed, the activity of the optimal AuPdCu catalyst is shown to outperform the bimetallic analogue by a factor of 1.7. The underlying cause for the increase in H 2 O 2 synthesis activity can be attributed to the electronic modification of the Pd species and changes in the surface composition of the nanoalloys as a result of Cu inclusion, as evidenced by XPS and CO-DRIFTS investigations. While catalytic stability is of concern, with deactivation attributed to in situ reductions in Pd species, it can be considered that these materials represent a promising basis for future exploration in a range of reactions, particularly where the in situ supply of H 2 O 2 is required.  Table S1. Synthesis details of the precursors used in the preparation of key bi-and tri-metallic 1%AuPd/ZSM-5 catalysts. Table S2. Summary of porosity and surface area of key 1%AuPd (0.0975) X (0.025) /ZSM-5 catalysts and HZSM-5 (30). Table S3. Comparison of initial H 2 O 2 synthesis rates over 1%AuPd (0.0975) X (0.025) /ZSM-5 (X = Cu, Ni, Zn) catalysts, as a function of Cu loading. Table S4. The effect of tertiary metal introduction upon the surface atomic composition of 1%AuPd (0.975) X (0.025) /ZSM-5 catalysts (X = Cu, Ni, Zn), as determined by XPS. Table S5. Comparison of initial H 2 O 2 synthesis rates over 1%AuPdCu/ZSM-5 catalysts, as a function of Cu loading. Figure Figure S3. BET analysis plots for key 1%AuPd (0.0975) X (0.025) /ZSM-5 catalysts and HZSM-5 (30). Key: ZSM-5(30) (red triangles), 1%AuPd (1.00) /ZSM-5 (blue squares), 1%AuPd (0.975) Cu (0.025) /ZSM-5 (green circles). Note: ZSM-5 support exposed to calcination prior to metal immobilisation (flowing air, 550 • C, 3 h, 20 • Cmin −1 ). Figure S4. Representative bright field transmission electron micrographs and corresponding particle size histograms of as-prepared (A) 1%AuPd (1.

Data Availability Statement:
The data presented in this study are fully available within the manuscript and supporting information.