Cu Promoted the Dynamic Evolution of Ni-Based Catalysts for Polyethylene Terephthalate Plastic Upcycling

Upcycling plastic wastes into value-added chemicals is a promising approach to put end-of-life plastic wastes back into their ecocycle. As one of the polyesters that is used daily, polyethylene terephthalate (PET) plastic waste is employed here as the model substrate. Herein, a nickel (Ni)-based catalyst was prepared via electrochemically depositing copper (Cu) species on Ni foam (NiCu/NF). The NiCu/NF formed Cu/CuO and Ni/NiO/Ni(OH)2 core–shell structures before electrolysis and reconstructed into NiOOH and CuOOH/Cu(OH)2 active species during the ethylene glycol (EG) oxidation. After oxidation, the Cu and Ni species evolved into more reduced species. An indirect mechanism was identified as the main EG oxidation (EGOR) mechanism. In EGOR, NiCu60s/NF catalyst exhibited an optimal Faradaic efficiency (FE, 95.8%) and yield rate (0.70 mmol cm–2 h–1) for formate production. Also, over 80% FE of formate was achieved when a commercial PET plastic powder hydrolysate was applied. Furthermore, commercial PET plastic water bottle waste was employed as a substrate for electrocatalytic upcycling, and pure terephthalic acid (TPA) was recovered only after 1 h electrolysis. Lastly, density functional theory (DFT) calculation revealed that the key role of Cu was significantly reducing the Gibbs free-energy barrier (ΔG) of EGOR’s rate-determining step (RDS), promoting catalysts’ dynamic evolution, and facilitating the C–C bond cleavage.


Electrocatalytic EGOR
EGOR was carried out in an H-typed cell (divided by a Nafion membrane) by Chronoamperometry (i-t) under continuous stirring.Herein, EG was only added into the anodic chamber.

Chemical oxidation of ethylene glycol
The as-prepared electrodes NiCu60s/NF were activated by 10 CV cycles followed by a LSV scan (0 to 0.8 V vs. Ag/AgCl, 10 mV s -1 ) in 1.0 M KOH solution to form R-NiCu60s/NF.After LSV scan, the R-NiCu60s/NF was immediately immersed into 1.0 M KOH with the addition of 0.3 M EG.

EGOR products identification and quantification
The products of EGOR were identified and quantified by 1 H NMR based on the standard chemical calibration curves (Figure S20). 1 H NMR spectra were collected on a 400 MHz Varian spectrometer.The NMR samples were prepared by mixing 540 µL solution after electrolysis and 60 µL D2O.The water peaks in all 1 H NMR spectra were removed from around 4 to 6 ppm by adding a break to clearly demonstrate product peaks.
1.9 Pretreatment of commercial polyethylene terephthalate (PET) powder and water bottle.
6.3 g PET powder (from GoodFellow) was soaked into 100 mL 2.0 M KOH solution in a sealed flask with stirring (500 rpm) for 18 hours at 60 °C.After that, the solid and liquid mixture was separated by centrifuging.The clear aqueous solution was denoted as PET powder hydrolysate and employed for further analysis and PET upcycling.5 g Aquafina PET water bottle was cut into small pieces and soaked into 100 mL 2.0 M KOH with stirring (500 rpm) for 48 hours at 60 °C.After that, the solid and liquid mixture was separated by centrifuging.The clear aqueous solution was denoted as PET water bottle hydrolysate and employed for further analysis and PET upcycling.

Products analysis after PET waste pretreatment
The hydrolysis products of PET hydrolysates were further analyzed by 1 H NMR and LC-MS.LC-MS: 100 μL PET powder or PET water bottle hydrolysate was diluted to 10 mL with methanol and analyzed by LC-MS (Agilent 6530B accurate-MASS Q-tof LC/MS)

Electrochemical upcycling of PET hydrolysate
For LSV, 5 mL PET-hydrolysate was diluted with the addition of 5 mL H2O (2 M KOH was diluted to 1 M KOH) to exclude the influence of pH differences.
For electrolysis, 8 mL PET-hydrolysate was diluted with 8 mL H2O (2 M KOH was diluted to 1 M KOH) in the anodic chamber while 16 mL 1.0 M KOH was added into the cathodic chamber.Electrolysis was conducted at 1.47 V vs. RHE for 5 hours.The products were quantified by 1 H NMR at each one-hour interval.

Electrochemical upcycling of PET water bottle.
PET water bottle hydrolysate was directly used as substrate without dilution.After electrolysis at 1.52 V vs. RHE for 1 h, the anodic solution was concentrated by the rotary evaporator and acidified with pure formic acid to result in the precipitation of TPA.Further, the solid and liquid parts were isolated by centrifuging.The solid product was dried in an oven at 60 °C.Following that, the solid product was dissolved into DMSO-d6 for the NMR analysis.

Calculation Equations
The Faradaic efficiency (FE) and yield rate were calculated using the following equations.
Where n is numbers of electron transferred for each product.Herein, n is 3 for formate, while it is 4 for glycolate.96485 C mol -1 is used as the Faraday constant.
The moles of products were calculated based on the standard chemicals 1 H NMR calibration curves as shown in Figure S20.

Computational methods
All calculations were performed using the CASTEP Package.The GGA-PBE exchange-correlation potential and ultrasoft pseudopotentials were used.The energy cutoff was set to 450.0 eV and the 2 × 2 × 1 MonkhorstPack mesh k-point was employed for the surface calculations.
The DFT+D dispersion correction was employed to account for the Van Der Waals (vdW) interactions.The (010) crystal plane of NiOOH is first constructed to ensure that Ni atoms located at the octahedral position serve as the effective adsorption sites for reaction intermediates.To simulate NiCu/NF active species, a layer of NiOOH was replaced by a layer of CuOOH.At the same time, a 15a Å vacuum layer was constructed to ensure that the adsorbent is not disturbed in the Z direction.The convergence tolerances were set to 2 × 10 −5 eV per atom for energy, 2 × 10 −3 Å for maximum displacement, and 0.05 eV Å −1 for maximum force during the geometry optimization.In the entire geometry optimization, the bottom three layers of atoms are fixed to simulate the bulk phase of the crystal.
The Gibbs free energy barrier (ΔG) calculations of each elementary step were based on the calculated hydrogen electrode model proposed by Nørskov et al., 1 which can be determined as where ΔE and ΔS are the reaction energy and entropy change and ΔEZPE is the difference in zero-point energy.The ZPE and entropic corrections were performed through frequency calculations.At the same time, all data in our article were discussed in the case of V vs. SHE = 0 V.        S7).Herein, all electrodes display three similar peaks.The absorption band at 1653 cm -1 is attributed to the bending vibration of water adsorbed on the electrodes' surface.Additionally, the bands at 3221 cm -1 and 1362 cm -1 can be assigned to -OH stretching and deformation vibrations, respectively.These characteristic peaks correspond to the formation of metal hydroxides after the reconstruction process.  The molality of aqueous species was 1´10 -6 mol/kg.

The impact of air oxidation on metallic species.
All electrodes were prepared by electrodepositing at -0.956 V vs. Ag/AgCl in 0.1 M Cu(NO3)2‧2.5 H2O solution (pH, 4.15).Under this condition, Cu 2+ ions will be reduced to Cu 0 and deposited on Ni foam according to the Pourbaix diagram (Figure S8a). 4 However, all ex-situ results indicate the existence of CuO species on the surface.Therefore, it is reasonable to assume the deposited metallic Cu 0 could be oxidized to CuO when exposed to air and forming a Cu/CuO core-shell structure.
To confirm this, XPS measurements and analysis were further performed.

The importance of shake-up satellite peak in Cu species XPS spectrum.
It is necessary to mention the importance of shake-up satellite peaks in Cu 2p XPS spectrum in identifying the paramagnetic Cu 2+ species (d 9 configuration).
As pointed out by the work of Mark C. Biesinger, 5 shake-up satellite peaks may occur when the outgoing photoelectron simultaneously interacts with a valence electron, exciting it to a higher-energy level.Following that, the kinetic energy of the core electron will be slightly reduced, according to the equation: KE = hv-BE -Φ, giving a satellite peak at a higher binding energy than the main lines.
As a classic example, shake-up peaks are present in the 2p3/2 spectra of d 9 Cu(II) species (Cu(OH)2 and CuO) but are absent in the d 10 Cu (0) or Cu(I) (Cu2O) spectra. 6Therefore, the presence of satellite peak is an important sign to qualitative or semi-quantitatively identify the Cu 2+ species. 5

XPS experimental section
To illustrate the effects of air oxidation during the characterization processes, two samples, NiCu60s/NF and NiCu180s/NF, where NiCu60s/NF and NiCu180s/NF were prepared by electrodepositing for 60 s and 180 s, respectively.CuNi60s/NF and CuNi180s/NF samples were dried under 60 °C for 3 h and 45 h prior to the XPS measurements.
After exposed to air for 3 h, in the Cu 2p3/2 spectrum (Figure S10a) of CuNi60s/NF, Cu 2+ (933.6 eV) and Cu + /Cu 0 (932.5 eV) species were identified while only Cu + /Cu 0 (932.7 eV) can be curve-fitted in the CuNi180s/NF sample.Meanwhile, smaller satellite peaks were observed in both samples, suggesting the presence of CuO.It is worth mentioning that Cu 2+ is present in CuNi180s/NF even though its Cu 2p3/2 cannot be fitted.This is ascribed to the amount of oxidized Cu 2+ is relatively smaller compared to Cu + /Cu 0 in the CuNi180s/NF, causing the fitting of Cu 2p3/2 peak by Cu 2+ and Cu + /Cu 0 to be difficult.
In comparison with air exposure for 3 h, stronger satellites, and a much higher Cu 2+ /(Cu 1+ +Cu 0 ) ratio were observed when both samples were exposed to air for 45 h.Mark C. Biesinger developed a mathematical method to quantitatively determine the ratio of Cu 0 /Cu 2+ and Cu + /Cu 2+ . 5Based on this method, the percent of CuO species can be derived by calculating the ratio of the satellite peak area to the total peak area (Table S3).Specifically, the percent of CuO increases from 10.63% to 34.43% and 0.49% to 19.10% in the CuNi60s/NF and CuNi180s/NF, respectively, when the air exposure time was increased from 3 h to 45 h.This result suggests that air oxidation is the main reason for the formation of CuO.In addition, when the air exposure time was 3 h, Ni 0 species was observed in the NiCu60s/NF while it was absent in the NiCu180s/NF, possibly resulting from the coverage of thicker Cu overlayer due to the longer electrodeposition time.In addition, the increased M-O species percent in the O 1s spectra when the air exposure time was increased (3 h to 45 h) further supports the fact that Cu can be spontaneously oxidized to CuO in air.

XPS depth-profiling analysis
To confirm the formation of the core-shell structure and estimate the depth of the oxidized layer (shell), Ar + depth profile was conducted with the following specifications: Ar + ion source operates with the 2 kV Ar + beam energy and the sample slit was set to 30 degrees.Sample raster and sputter rate were set to 7.5 mm ´ 7.5 mm and 0.01 nm/min, respectively.As shown in Figure S11, with an increased time (0 to 5 min) of Ar + sputtering, O 1s peak intensity was apparently reduced, suggesting the removal of the oxidized metal species from the electrode's surface.
As displayed in Figure S12 and Table S4, for Cu 2p spectra, after 5 mins of Ar + sputtering, the percent of CuO species was reduced from 34.43% to 15.32% (calculated from satellite peak area to total peak area) and a higher percentage of Cu + /Cu 0 was observed.
Meanwhile, NiO (at 853.9 eV) species were observed from the Ni 2p spectra after Ar + sputtering, indicating the NiO layer is under the Ni(OH)2 (at 855.4 eV) layer.Ar + depth profile sputtering results indicate that the metal on the electrodes' surface was oxidized to higher oxidation states when the sample was exposed to air while the core metals remained at their lower oxidation states.To investigate oxidation states of the core metals after reconstruction, 10 mins of Ar + sputtering was further conducted on the R-NiCu60s/NF sample to remove the surface oxide/hydroxide layer.As shown in Figure S13b, after 10 mins of Ar + sputtering, lowoxidation state Cu species and NiO were present, which indicates the surface metals were reconstructed to their oxidized forms as CuO and Ni(OH)2 while the core metals remain in the Cu 0 and NiO forms.As shown in CVs (Figure 3a) and in-situ Raman spectroscopy (Figure 4b-4e) of NiCu60s/NF, the characteristic Cu species peaks are absent due to its small deposition amount compared with the strong Ni substrate peak.Therefore, to exclude the interference of Ni in identifying the Cu active species, Cu foam was used as the substrate.

Energy dispersive X-ray spectroscopy (EDS) analysis
To verify if the Cu foam is a proper substrate to identify the active Cu species, carbon fiber paper (CFP) and fluorine-doped tin oxide (FTO) were used as substrates for Cu deposition with the same deposition condition as that of NiCu60s/NF.The corresponding samples were denoted as Cu60s/CFP and Cu60s/FTO.As shown in Figure S29a, three anodic peaks and three cathodic peaks were identified when Cu foam was employed.The peak a at ~-0.4 V vs. Ag/AgCl is attributed the Cu 0 to Cu 1+ , while the peak b (-0.2 V to 0 V vs. Ag/AgCl) might be assigned to two possible oxidation reactions: Cu 0 to Cu 2+ and Cu 1+ to Cu 2+ .The following oxidation peak c at ~0.5 V vs. Ag/AgCl is associated with the conversion of Cu 2+ to Cu 3+ .The Cu 3+ species can be ascribed to the formation of CuOOH, confirming CuOOH is the Cu active species for EGOR and OER. 7,8 he corresponding anodic and cathodic reactions are displayed in Figure S29b.In comparison, the CV curves of Cu60s/CFP (Figure S29c) and Cu60s/FTO (Figure S29d) exhibited similar redox peaks, suggesting that Cu species on different substates possess the same active species during the EGOR and OER.

In-situ Raman Section
In-situ Raman studies Cu foam Subsequently, Cu foam was directly applied for the in-situ Raman measurement to investigate the active species of Cu during OER and EGOR.
It is worth noting that although previous works demonstrate that CuOOH is the active Cu 3+ species for OER and biomass oxidation reaction via CV and XPS, 7-9 none of them provided direct evidence to confirm the existence of CuOOH under in-situ electrochemical conditions.This causes the assignment of CuOOH in the Raman spectroscopy is still a mystery.Thus, DFT calculation of CuOOH was conducted here to provide the necessary theoretical support.
Most previous works assigned peak-c to Cu 2+ to Cu 3+ conversion to form a CuOOH active species with formal Cu 3+ (3d 8 ) under the alkaline condition. 7,8 herefore, in this work, the active Cu 3+ species for OER and EGOR are still ascribed to CuOOH (Cu 3+ , 3d 8 ).Under OER condition (Figure S30a), with applied potentials between 1.32 V and 1.62 V vs. RHE, a small peak at ca. 483 cm -1 was observed, which is ascribed to the formation of Cu(OH)2. 10,11 ith the applied potential increased over 1.62 V vs. RHE, a new peak at 570 cm -1 , corresponding to the formation of CuOOH 12 species, was detected.This result is supported by the DFT calculation as shown in Figure S30c where vibration peaks at ~500 cm -1 and ~550 cm -1 were noticed.Under the EGOR condition, besides the peak from Cu(OH)2 at 491 cm -1 , the characteristic peak of CuOOH was observed (at ~565 cm -1 with an applied bias of 1.52 V vs. RHE > Cu 2+ /Cu 3+ redox potential of 1.50 V vs. RHE ).Therefore, in-situ formed Cu(OH)2 and CuOOH simultaneously served as the active species for OER and EGOR.

Figure S30.
In-situ Raman spectra of (a) Cu foam in OER and (b) EGOR.The peak at 867 cm -1 is ascribed to C-C stretching vibration from EG. 13 (c) Simulated CuOOH Raman spectra by DFT calculation.

Indirect oxidation mechanism and direct oxidation mechanism
Indirect oxidation mechanism (Scheme S1): in the indirect oxidation process, Ni(OH)2 will be electrochemically oxidized to NiOOH (1), serving as a chemical oxidant and abstracting the α-hydrogen from alcohol, which is the rate-determining step (RDS) (3).Following the RDS, NiOOH will be converted back to Ni(OH)2 (4).This process will be cycled by applying a bias to regenerate NiOOH from Ni(OH)2.For indirect oxidation, as long as the applied potential is sufficiently positive to regenerate NiOOH, the rate of indirect alcohol oxidation is potential-independent because the RDS is a chemical step without the involvement of electricity.Scheme S1.[16] Direct oxidation mechanism: A few previous works reported that alcohol oxidation via the NiOOH catalyst is potential-dependent (PD) when the applied potentials are more positive than the potentials that enable indirect oxidation.This PD oxidation was referred to as direct oxidation in contrast to indirect oxidation.
As reported by Kyoung-Shin Choi's work, 16 when the dominant alcohol reaction is shifted from indirect to direct oxidation with an increased potential, an increased average Ni oxidation state is observed.For example, the average Ni valence is +2.57at 0.45 V vs. Ag/AgCl for furfural oxidation (10 mM in pH = 13 KOH), but the average Ni valence is increased to +3.63 at 0.625 V vs. Ag/AgCl.The increased average Ni valence is attributed to the accumulation of Ni 4+ .Therefore, the direct oxidation mechanism was proposed via hydride transfer on NiO2 sites as shown in Scheme S2.Scheme S2.Proposed PD oxidation mechanism of alcohols to aldehydes via hydride transfer. 17onfirm EG was indirectly oxidized by R-NiCu/NF without applied potential.Firstly, the NiCu60s/NF catalyst was reconstructed to R-NiCu60s/NF.Subsequently, R-NiCu60s/NF was immersed into a 1.0 M KOH solution with the addition of 0.3 M EG.After certain reaction time, the products were identified by 1 H NMR as shown in Figure S31a.To quantitively decouple the contribution of indrect and direct oxidation, a three-step electrochemical procedure (Figure S32) was applied.

Quantitative methods description:
In the quantitative method (Figure S32), the first step is to apply an oxidation potential (E1) to Ni(OH)2 to oxidize it to NiOOH within a certain time to achieve a steady-state current (I1).Herein, E1 is less positive than the onset potential of OER to avoid any OER influence.In this step, Ni(OH)2 is converted to NiOOH, and both indirect and direct reactions occur, thus,  !=  "#$"%&'( +  $"%&'( (1).Once the I1 is established, the applied oxidation potential (E1) is withdrawn and the electrolyte remains under stirring (step 2) for a t time, where only indirect oxidation will occur due to the direct oxidation is a potential-dependent reaction.In this step, NiOOH will chemically oxidize alcohol while itself will be reduced back to Ni(OH)2.In the third step, a reducing potential is applied to reduce the remaining Ni 3+ in the second step to Ni(OH)2.The passed charge number (stored charge, Q) in step 3 is equivalent to the charge needed to reduce all remaining Ni 3+ to Ni 2+ .By repeating this 3-step process with different stirring time (t) under the open circuit condition, the disappearance of positive charge as a function of time (t) in step 2 can be obtained.
In the reported work with NiOOH as catalyst, the plot of 1/C vs t displays a linear relationship, therefore, the rate of disappearance of charge from catalyst at time t is calculated by a pseudo-second order rate law: − $)(() $( =  , () (2), where the C(t) is the charge stored in the catalyst at time t and k is the 1/C vs. t plot slope.Furthermore, the disappeared charge from the catalyst during the open circuit condition is used to oxidize alcohol, therefore, the instantaneous rate of the disappearance of charge from catalysts is equal to the rate of indirect oxidation at that time t: − $)(() $( =  "#$"%&'( () =  "#$"%&'( () (3).Here, the instantaneous rate of indirect process at t = 0 s during step 2 corresponds to the steady state partial current for indirect oxidation (Iindirect) at the applied potential in step 1 due to the unit of rate is coulombs per second (C/s).Therefore, once Iindirect is determined by solving equation 2, the Idirect can be calculated by solving equation 1.

Indirect and direct oxidation contribution for EGOR over NiCu60s/NF:
In this work, the developed NiCu60s/NF displayed a FE > 90% at 1.52 V vs. RHE (0.50 V vs. Ag/AgCl.Therefore, at this potential, this method will be accurate enough to distinguish the contribution of the indirect and direct oxidation for EGOR.If the FE is low, the partial current generated from OER cannot be ignored and equation 1 will be  !=  "#$"%&'( +  $"%&'( +  -./ .

Electrochemical method details:
Electrochemical experiments were conducted by using Metrohm Autolab potentiostat (software version: Nova 2.1.4).The experimental program is shown in Figure S33a, an oxidation potential of 0.50 V vs. Ag/AgCl was applied for 20 s to reach the steady state current.After waiting for a certain time t, a 0 V vs. Ag/AgCl was applied to reduce the remaining Ni 3+ .
For step 3, previous work 16 holds the reduction potential for 20 s to reduce all the remaining Ni 3+ to Ni 2+ .However, for NiCu60s/NF, when applying 0 V vs. Ag/AgCl, a positive current density was generated.Therefore, if the Ni 3+ reduction to Ni 2+ in less than 20 s, the accumulative charge number in step 3 is not the real charge number for Ni 3+ to Ni 2+ reduction.To address this issue, an extra command was set for step 3 (Figure S33b), this cutoff can help to find the accurate charge number for reducing Ni 3+ to Ni 2+ .n.a. is the abbreviation of "not available" from the published literature.

Figure S1 .
Figure S1.The optical photography of the prepared electrodes, where NiCu60s/NF represents Cu species deposited on a NF electrode after 60 s.R-NiCu60s/NF represents the catalyst after reconstruction.Post OER sample (NiCu60s/NF-Post OER) was received by electrolysis at 1.62 V vs. RHE for 2 h in 1.0 M KOH solution.Post EGOR sample (NiCu60s/NF-Post EGOR) was received at 1.42 V vs. RHE for 2 h in 1.0 M KOH with the addition of 0.3 M EG.

Figure S2 .
Figure S2.SEM images of (a) commercial Ni and (b) Cu foams.Prior to the SEM characterizations, Ni foam, and Cu foam were cleaned by sonicating in 4 M HCl for 10 mins to remove the oxidized layer and contaminants.

Figure S5 .
Figure S5.CV scans of (a) NiCu60s/NF, (b) Ni foam, and (c) Cu foam in 1.0 M KOH at a scan rate of 100 mV s -1 .

Figure S6 .
Figure S6.XRD patterns of Cu electrodeposition on (a) Ni foam and (b) carbon fiber paper.

Figure S7 .
Figure S7.FTIR spectra of (a) Ni foam, (b) Cu foam, and (c) NiCu60s/NF before and after the reconstruction process.R-Ni foam, R-Cu foam, and R-NiCu60s/NF represent Ni foam, Cu foam, and NiCu60s/NF after the reconstruction process, respectively (FigureS7).Herein, all electrodes display three similar peaks.The absorption band at 1653 cm -1 is attributed to the bending vibration of water adsorbed on the electrodes' surface.Additionally, the bands at 3221 cm -1 and 1362 cm -1 can be assigned to -OH stretching and deformation vibrations, respectively.These characteristic peaks correspond to the formation of metal hydroxides after the reconstruction process.2

Figure S9 .
Figure S9.XPS survey spectra of NiCu60s/NF and R-NiCu60s/NF.The survey scan of NiCu60s/NF (Figure S9) revealed the presence of Cu, Ni, O, C, N, Cl and C elements.The presence of Cl is attributed to the use of Ag/AgCl reference electrode in the electrodeposition process.For the reconstructed sample, R-NiCu60s/NF, the presence of K might originate from the KOH electrolyte.

Figure S11 .
Figure S11.XPS survey spectra of NiCu60s/NF (exposed to air for 45 h) as a function of Ar + sputtering time from 0 to 5 mins.
Figure S13.XPS survey spectra of R-NiCu60s/NF (a) as a function of Ar + sputtering time from 0 to 10 min and (b) Cu 2p, Ni 2p, and O 1s spectra after 10 min Ar + sputtering.

Figure S15 .
Figure S15.EDS analysis of NiCu/NF.(a) STEM image of the NiCu/NF.The green line squared area was selected as the EDS scanning area.(b) The EDS spectrum of existing elements in the sample.Herein, Au originated from Au TEM grids.(c) The corresponding elemental mapping.In this NiCu/NF sample, the electrodeposition time was set to 30 mins in order to collect enough catalyst for analysis.

Figure S16 . 2 . 5
Figure S16.EDS analysis of R-NiCu/NF.(a) STEM image of the NiCu/NF for EDX analysis.The green line squared area was selected as the EDS scanning area.(b) The EDS spectrum of the existing elements in the sample.Herein, Au originated from the Au TEM grids.(c) The corresponding elemental mapping.In this R-NiCu/NF sample, the electrodeposition time was set to 30 mins in order to collect enough catalyst for analysis.

Figure S18 .
Figure S18.(a) 1 H NMR and (b) 13 C NMR before and after electrolysis of 0.3 M EG in 1.0 M KOH for 24 h at 1.47 V vs. RHE, the concentration of HCOOH standard was 500 mM.

Figure S19. 1 H
Figure S19. 1 H NMR spectra of (a) anodic and (b) cathodic chamber products of NiCu/NF with different electrodeposition times.Electrolysis conditions: at 1.47 V vs. RHE for 2 h in 0.3 M EG solution.

Figure S21 .
Figure S21.Ethylene glycol potential-dependent electrolysis of NiCu60s/NF electrode. 1H NMR spectra of (a) anodic and (b) cathodic chamber products at different potentials.Electrolysis conditions: electrolysis of 0.3 M EG in 1 M KOH solution for 1 h.

Figure S22 .Figure S23 .
Figure S22.Products FEs and yield rates of (a) Ni foam and (b) Cu foam electrodes for EGOR over 1 h electrolysis at different applied potentials.ECSA measurements

Figure S25 .
Figure S25.Cu 2p, Ni 2p, and O 1s XPS spectra of NiCu60s/NF after OER (a) and EGOR (b), respectively.OER condition: at 1.62 V vs. RHE for 2 h in 1.0 M KOH solution EGOR condition: at 1.42 V vs. RHE for 2 h in 1.0 M KOH with the addition of 0.3 M EG.

Figure S26 .
Figure S26.Chronoamperometry at 1.42 V vs. RHE for 24 h in 0.3 EG solution.In the 24 h stability test, a significant decrease in current density was observed, This phenomenon was mainly attributed to the consumption of EG.

Figure S31 .
Figure S31.(a) 1 H NMR spectra of indirect EGOR products after different reaction times.(b) Formate concentration as a function of indirect oxidation time.

Figure S32 .
Figure S32.Schematic illustration of the 3-step electrochemical method to deconvolute direct and indirect processes.16

Figure S33 .
Figure S33.(a) Nova electrochemical experiment program for indirect oxidation and direct oxidation and (b) the cutoff when working electrode current > 0 is added for step 3, which is helpful for finding the accurate charge number for Ni 3+ to Ni 2+ reduction.The 1/C vs. waiting time (t) was displayed in Figure S34a, and the current density for indirect and direct oxidation was calculated by solving equations 1 to 3. As shown in Figure 34b, 91.2% of the total current density originated from indirect oxidation, suggesting indirect oxidation is the dominant reaction in EGOR over NiCu60s/NF.

Figure S34 .
Figure S34.(a) 1/Charge vs. time plot for measuring the amount of remaining positive charge in NiCu60s/NF after a given stirring time in 300 mM EG solution (1.0 M KOH).(b) Contribution of the indirect and direct oxidations for EGOR at 1.52 V vs. RHE.Favorable reaction pathway

Figure S39 .
Figure S39.Liquid chromatography-mass spectroscopy (LC-MS, negative ion mode) of the commercial Aquafina PET water bottle hydrolysate

Figure S41 .
Figure S41.Chronoamperometry (i-t) curves at 1.52 V and 1.62 V vs. RHE in PET water bottle hydrolysate with the NiCu60s/NF electrode.

Table S1 .
XPS peak binding energies and assignments of NiCu60s/NF and R-NiCu60s/NF.

Table S2 .
Quantitative analysis of different O species obtained by curve fitting the XPS peaks.

Table S6 .
A literature survey of ethylene glycol oxidation and PET electrochemical upcycling with non-noble metal electrocatalysts in the alkaline electrolyte.