Tuning the Selectivity of Electrocatalytic CO2 Reduction Reaction via Anion-driven Nanostructuring of Copper Electrodeposits

Electrochemical CO2 reduction (ECR) to value-added products is one of the potential ways to utilise CO2 as a feedstock, thereby decreasing its level in the atmosphere as it has harmful repercussions on planet Earth. Copper (Cu)-nanostructures have demonstrated a great potential to convert CO2 into valuable higher-end hydrocarbons electrochemically but with poor selectivity. Therefore, novel strategies to tune Cu-based electrocatalysts’ activity and selectivity toward multi-carbon products, particularly at low overpotential, are highly desirable. In the present work, we report an atom-economic strategy to tune the physicochemical properties and the electrocatalytic activity of Cu-nanostructures towards ECR. The Cu-nanostructures synthesized via pulse electrodeposition from an electrolyte bath containing Cu-precursor salts with varying anions (viz. acetates, nitrates, sulphates, and chlorides) are investigated for their effect on the physicochemical properties and the ECR performance. The Cu-electrodeposits from Cu-chloride, having cubic morphology, exposed Cu(100) facets, higher Cu+ content and enhanced electrochemical active surface area demonstrated the best ECR performance depicting good selectivity for ethylene formation.

An exponential increase in CO 2 concentration in the atmosphere 1 due to anthropogenic activities has called out urgent mitigation measures. At the same time, CO 2 is a valuable carbon feedstock for value-added chemicals. Electrochemical CO 2 reduction (ECR) is a viable approach to achieve this goal as it can be carried out under ambient conditions, have easily controllable parameters, and can be scaled up to suit industrial-scale processes. 2,3 Particularly, Cu has been the focal point of research as an electrocatalyst for ECR to hydrocarbons 4-8 since the work of Hori et al. 9 The ECR activity and selectivity can be tuned by varying physicochemical properties of Cu including size, 10 shape, 11 oxide content, [12][13][14][15] crystal phase and facets, 16 and surface roughness 17,18 through judiciously conceived synthesis approaches. Electrodeposition is one such simple and effective synthesis method with reasonable control over the physicochemical properties of Cu nano-electrocatalysts for desired applications. 13,19 The properties of the electrodeposits are tuneable with the constituents of the electrolytic bath, 20 precursors concentration, 21 applied current, 22 and deposition time. 23 Several researchers have electro-synthesized well-defined Cunanoarchitectures and have studied the effect of electrodeposition parameters on the physicochemical properties (morphology, composition, and crystal facets) and ECR activity of Cu-electrodeposits. 19,22,24 Cook et al. noted the profound impact of supporting electrolytes used for CO 2 electrolysis on the ECR activity of the in situ electrodeposited Cu-nanostructures from CuSO 4 solution. 25 The nature of anions in the supporting electrolyte (catholyte for ECR) has also been reported to influence the ECR product distribution. [26][27][28] Frese and Summers noted that Cl − impurities in aqueous Na 2 SO 4 could lower CH 4 and increase CO yields. Gao et al 29 reported that the presence of halides like I − , Br − , Cl − , and CO 3 − affect the selectivity of the Cu-nano electrodeposits towards ethylene and other multi-carbon alcohols. The superior selectivity towards ethylene has been attributed to the electronic and chemical effects of stabilizing subsurface oxygen and cationic Cuspecies due to surface-adsorbed halides. Wang et al. 30 demonstrated that the wet chemical halide (Cl − , Br − , and I − ) treatment of Cu-foil yields cuprous halide microcrystals that exhibit distinct ECR activity and selectivity, attributed to the halide-derived morphology effects in Cu. A comprehensive comparison of halide-assisted electrode surface restructuring and its ECR activity was also studied by Tang et al. 31 Their study distinctly shows that the varied surface morphology of the electrocatalyst obtained by electrochemical treatment of Cu foil in the presence of KX (X = Cl − , Br − , and I − ) in KHCO 3 is crucial for increased C 2+ product yield.
However, most of the literature reported studies have utilized externally added ions, like supporting electrolytes and halides, for making Cu-nanostructured ECR catalysts, where the presence of adsorbed anions play a significant role towards ECR performance. Considering the concerns related to environmental impact of the synthetic procedures and the atom-economy among chemists, synthesis methods employing reagents with multiple characteristics are essential to ensure minimal use of the undesirable and excess reagents. In this work, for the first time to be reported, we present the precursor Cu-salt anions driven tuning of physicochemical properties and the ECR performance of Cu-nanoelectrodeposits. In line with the principles of green and sustainable chemistry, the presence of different anions was ensured in the electrolyte bath directly from the Cu-precursor salts rather than from any additional reagents to make this synthesis method atom-economical. 32 Various commonly used copper salts were chosen (Cu-acetate, Cu-chloride, Cu-nitrate, and Cu-sulphate) as the copper source for electrodeposition. Among the halides, we purposely chose Cl − as it has the lowest tendency to get adsorbed on the electrode surface than the others 29 (CuI > CuBr > CuCl) so that the activity enhancement due to adsorbed halide can be ruled out. The galvanostatic pulse electrodeposition method was used to synthesize the Cu electrocatalysts as this approach has a periodical time off, which allows careful etching of the nanostructures and homogeneous distribution of ions during electrodeposition. 33 It should be noted that several studies discuss the effect of anions in the catholyte for electroreduction of CO 2 . [26][27][28]34 However, this work focuses on the effect of anions towards the effective synthesis of Cu nanostructured electrocatalysts with tuneable ECR activity. It was found that Cu-electrodeposits made from Cu-chloride depicts the best ECR performance with the highest total faradaic efficiency and good selectivity for ethylene formation of 30% at −0.93 V vs RHE.  2 and Ar gases were passed through a filter equipped with an oxygen trap, gas purifier, and moisture trap to remove the traces of impurities and moisture (Sigma Gases, New Delhi) before using for electrochemical measurements. SmartTrak 100 -Digital Mass Flow Controller (Sierra Instruments, Model No.: C100L-DD-2-OV1-SV1-PV2-V1-SO) was used for controlling the flow of gases at the desired rate. Carbon fiber paper used as the electrode substrate was purchased from Nickunj Pvt. Ltd. Double distilled water has been utilised to make all the solutions.
Synthesis of Cu-nanostructured electrodeposits.-The Cu electrodeposits from various Cu-salts were obtained by galvanostatic pulse electrodeposition (GPED) method using a three-electrode setup: Ag/AgCl/Cl − as reference, Pt gauze as counter and Cu-foil (1 × 1 cm 2 ) as the working electrode. The Cu-foil was cleaned with 5 M HNO 3 and distilled water, prior to use. The electrolyte bath contained freshly prepared 0.1 M ascorbic acid, 5 mM Cu 2+ salt solution along with their corresponding anions CH 3 COO − , Cl − , NO 3 − , and SO 4 2− . Autolab 302 N was used to provide the current pulses (−50 mA cm −2 ) at the stipulated time intervals (5 s) with a cut-off time of 5 s. This cycle was repeated 40 times to give a total charge of 10 coulombs for the electrodeposition of Cu. The brownish Cu-electrodeposits were then scraped off from the Cu-foil surface very carefully by bath sonication, rinsed in double distilled water, centrifuged, and dried. The Cu-electrodeposits from copper sulphate, copper nitrate, copper acetate, and copper chloride are named as CS, CN, CA, and CC, respectively. These dried electrodeposits were weighed and drop-casted onto CFP and used as the working electrode. We believe that the routinely used bath sonication is a mild technique that would not remove any Cu-powder from Cu-foil substrate causing any interference with the ECR activity . Nevertheless, to verify if the Cu-foil has any interference towards ECR activity, we compared the cyclic voltammetry measurements of Cu-deposits obtained by the process of scrapping off (which was then drop casted on carbon fibre paper (CFP)) viz-a-viz direct electro-deposition on CFP ( Fig. SI-1). It shows similar Cu-redox peaks and current responses, which suggest the insignificant interference from the Cu-foil or Cu-powder scrapping off from the Cufoil. However, the method of drop-casting the pre-synthesized Cudeposits on CFP was chosen to gravimetrically normalize the ECR activity and to avoid ambiguity that could arise due to non-uniform mass loading of Cu-electrodeposits during direct electrodeposition because of the varying electrodeposition conditions, like, different anion-containing salts. Scheme 1 illustrates the synthesis protocol and the obtained morphology of the electrodeposits.
Material characterisation.-The crystal phases and structures of the materials were studied by X-ray diffractometer (Bruker D8 Advance) fitted with Ni-filtered Cu Kα (λ = 0.154 nm) radiation. Raman spectra were recorded (wavelength: 532 nm) with HORIBA, Olympus BX 41 Raman spectrometer. To understand the surface morphologies, scanning electron microscopy (SEM) and Transmission Electron Microscopy (TEM) images were taken from Zeiss EVO50: ZEISS Germany, and Thermo Fischer Scientific US (operating voltage 120 kV), respectively, while Field emission SEM (FESEM) images were taken from FESEM TESCAN Instrument. The surface elemental composition was characterised by energy dispersive X-ray spectroscopy (EDX) and an in-depth study of its oxidation states was done using X-ray photoelectron spectroscopy (XPS, PHI 5000 Versa Prob II, FEt Inc.). Deconvolution and spectral fitting of the XPS curves was done with the help of CASA XPS software.
Electrochemical measurements.-All the electrochemical measurements were performed in a custom-made gas-tight 2-chambered H-type cell using AUTOLAB 302 N Potentiostat/Galvanostat. A three-electrode system with Ag/AgCl/Cl − (E 0 = 0.197 V vs RHE) reference electrode and catalyst-modified carbon fiber paper (CFP) working electrode fixed in the cathodic chamber, and Pt foil counter electrode placed in the anodic chamber was used. Each chamber contains 40 ml of 0.2 M KHCO 3 separated by a proton exchange membrane. A schematic of the complete setup is shown in Scheme SI-1. The CFP (2 cm × 2 cm) was hand-painted with catalyst-ink made by dispersing 2.0 mg of dried Cu-electrodeposits in 180 μl of isopropanol and 20 μl of Nafion binder solution, followed by ultrasonication for 15 min A small part of CFP from the top was used to make electrical connections and masked properly with resin. Prior to electrochemical measurements, Ar was purged for 30 min to remove any residual O 2 followed by CO 2 purging for 30 min at 20 sccm to saturate the electrolyte. Linear sweep voltammetry (LSV) and cyclic voltammetry (CV) curves were recorded within −0.23 to −1.23 V vs RHE, iR corrected, and the potentials measured using Ag/AgCl/Clare converted to RHE using Eq. 1.
Electrochemical impedance spectroscopy, particularly Nyquist plot, measurements were done at open circuit potential (OCP) and −0.83 V vs RHE in the frequency range of 0.1 to 10 6 Hz. ECR gaseous products were collected in a Tedlar bag connected to the gas-tight Teflon cell during 2 h of CO 2 electrolysis under continuous purging (in both chambers), at a flow rate of 5 sccm, to maintain the back pressure. The gaseous samples were analysed using Gas chromatography (GC) (Thermo Fischer Trace 1100) instrument equipped with a Methanizer, flame ionization detector (FID) and thermal conductivity detector (TCD). Ar was used as a carrier gas. The liquid products were also collected (after 2 h of CO 2 electrolysis) for analysis using 1 H NMR spectrometer at 400 MHz Larmor frequency by sampling aliquots of the catholyte in D 2 O as solvents. The solvent suppression method (detailed in Supporting Information) was followed to reduce the overwhelming background response from water. The faradaic efficiency (F.E.) of ECR products was calculated using Eq. 2.
Where, N product is the total number of electrons required for forming a particular product, while N total is the total number of electrons given for the reaction. N product is obtained (Eq. 3) from the quantity of ECR product (X in ppm), the number of gas molecules (n in moles) present in the injected volume (1 ml), and the electrons (n e ) needed to produce one molecule of the specific product, N A is Avogadro's number (6.022 × 10 23 mol −1 ). N total (Eq. 4) is estimated from current (I in A) used during CA for duration, t, of CO 2 electrolysis and charge on electron, e.
The selectivity towards a particular product, 35 e.g., the main Carbonproducts (C-products), is determined at −0.83 V vs RHE as per the Eq. 5.

Results and Discussion
Material characterisation.- Figure 1a depicts powder XRD patterns of the Cu-electrodeposits obtained with the different Cu salts, showing diffraction peaks for both Cu(0) at 43°(111) and 74°( 200), and Cu 2 O at 36°(111), 42°(200), 61°(220) and 73°( 311). 10,36 The peak positions match well with the JCPDS File Nos. of Cu (01-085-1326) and Cu 2 O (00-005-0667). The (111) facets of Cu 2 O and Cu 0 at 36°and 43°, respectively, are grown well in all the Cu-electrodeposits, as noted from the respective high peak intensities. The Cu 2 O to Cu ratio calculated from the integrated peak areas of (111) plane is highest for CC, followed closely by CS (Table I). The presence of (220) Cu 2 O plane in only CC and CS also shows growth of Cu 2 O lattices in these samples. Raman spectroscopy, which is highly surface-sensitive technique, has been used to dissect the oxide phases of copper. 37 Figure 1b shows the overlay of Raman spectra for all the Cu-electrodeposits, which reveal the presence of CuO in addition to Cu 2 O, unlike XRD. The absence of CuO in XRD points out that CuO could be present only as an amorphous thin film on the surface of the electrodeposits. 38,39 Raman peaks at 294, 345, and 632 cm −1 can be assigned to A g , B 1g and B 2g modes of CuO phonon while the peaks at 148.7, 218.3, and 520 cm −1 are attributed to the T 1u , E u and T 2g modes of Cu 2 O. 40 The effect of Cu salt anions on the morphology of the obtained Cu-electrodeposits is visible from the SEM and TEM images shown in Fig. 2 and Fig. SI-2, respectively. The possible reactions responsible for the distinct morphology of the Cu electrodeposits are given in Scheme 1. Due to the low pH of the electrolytic bath (∼2.5), the Cu-electrodeposition is accompanied by hydrogen evolution reaction (HER), 41 which results in the formation of nondendritic structures as seen in the case of CS and CA. The presence of NO 3 − in the electrolyte bath causes Cu-catalyzed nitrate reduction reaction, a side reaction, as Cu is a well-known catalyst for nitrate reduction. The simultaneous redox reactions at the Cu-surface cause an additional dissolution of the Cu nuclei 42 and result in a smoother morphology for CN, where dendrite formation is overridden. The presence of Cl − during the electrodeposition of CC results in the formation of nanocubes by reducing Cu 2+ via the formation of CuCl which is known to exist in cubic form when it precipitates. 43   Cu-electrodeposition involves the reduction of Cu 2+ (Eq. 6) to Cu, the presence of Cu 2 O is reasoned due to the presence of AsH, a chelating ligand, forming Cu(II)-ascorbate complex that undergo reduction as Eq. 7. The electrodeposits without chelating agents forms Cu(OH) 2 , which is more vulnerable to get reduced to Cu 0 as per Eqs. 8 Figure 3 shows the ex-situ core-level (CL) photoelectron spectra of the Cu 2p region of all the Cu electrodeposits. The background of the curves was accommodated by linear fit, and the curves were deconvoluted to assess the presence of Cu, Cu 2 O, and CuO in these samples. For all electrodeposits, the Cu 2p band split into Cu 2p 1/2 (952 eV) and Cu 2p 3/2 (932 eV) with a spin-orbital splitting energy of ∼20 eV. 46 The XPS also confirms the presence of CuO due to the peaks at 952.6 ± 0.3 eV and 932.8 ± 0.5 eV in all the samples in accordance with the Raman spectra. The satellite features at 944--941 eV evince the presence of Cu 2+ 47 . Since the binding energies of Cu 0 and Cu(I) lie very close to each other, with <0.1 eV difference between them, it is difficult to deconvolute these peaks. 38,48 Nevertheless, the Cu(I) presence is proved from the Cu L 3 M 4,5 M 4,5 Auger peak at 571 eV 49 (Fig. SI-3, right panel). The integrated peak area ratios under Cu 2+ to Cu 0 -Cu + (Fig. 3) points out the relatively higher percentage of copper oxides in CC than other samples (Table I)    catalysed oxidation of ascorbic acid (H 2 Asc) on the catalyst surface (Eq. 10). 50 The Cl − ion-containing solutions (especially up to 0.1 mol l −1 ) are well-reported catalyst for ascorbic acid oxidation compared to other anions. 51 The presence of trapped DDHA can be proved from TGA analysis of these electrodeposits (Fig. SI-4). The weight loss at lower temperatures of 200°C-300°C evolves from the removal of adventitious H 2 O molecules on the surface of the electrocatalysts. For CC, the weight loss at ∼450°C is due to the removal of molecules like CO 2 from disintegrating DDHA. The elemental analysis (Fig. SI-5) shows very low percentages of anionic elements left on the surface, indicating negligible adsorption of anions on the surface of Cu-electrodeposits. Lesser anionic adsorption is advantageous because it can block the active sites on the catalyst surface.
Electrochemical reduction of CO 2 and product collection.-The comparison of LSV curves recorded on CC-modified CFP in Ar, and CO 2 saturated 0.2 M KHCO 3 solution (for 30 min of purging at 20 sccm influx for both gases) is depicted in Fig. 5a. It shows an increased current density and lower onset potential for CO 2 than for Ar purged solution. This points to the possibility of ECR activity in addition to HER, which is the only possible reaction in an Ar-purged solution. An overlay of LSV curves recorded on Cu-nanoarchitectures prepared with different Cu-precursors is shown in Fig. 5c. It indicates the best ECR activity by CC, showing the highest current density after the onset potential of −0.5 V vs RHE (marked at current density of 5.5 mA cm −2 ). Figure 5b compares the Nyquist plots for CC under Ar and CO 2 purged electrolyte conditions at −0.83 V vs RHE, demonstrating a considerable decrease in charge transfer resistance (R ct ) for CO 2 saturated solution than for Arsaturated one. The comparison of Nyquist plots measured at OCP is depicted in Fig. 5d, which gives an insight into the material characteristics such as the type of charged species and the conductivity of the reducible species at the electrode surface, charge transfer resistance (R ct ), and contact resistance (R s ). The lowest R ct indicates the best catalytic activity of CC, possibly, due to higher number of active sites for CO 2 adsorption and activation.
To verify this, the roughness factors of all (four) electrodeposits was estimated from the double-layer capacitance (C dl ) measurements. For this, EIS was taken at OCP in 0.1 M HClO 4 at 0.1 to 10 5 Hz for each electrocatalyst (Fig. SI-6(a)). Due to the depressed semi-circles, an indication of non-homogeneity of the double layer structure, C dl is expressed as constant phase element (CPE) 52 and the "true" capacitance was calculated from CPE using Cole-Cole relation (details in Supporting Information and Table SI-2). The roughness factor, RF, was calculated (Eq. 11) by taking the ratio of C dl to the specific capacitance, C s (assigned a value of 28 μF cm −2 as per the reported literature 12  To understand the intrinsic catalytic activity of the synthesized Cu samples, the LSV curves over different catalyst showing current densities as a function of the potentials were normalised by electrochemically active surface area. For this, the ECSA of the four catalysts-modified CFP electrodes were also found out from the ratio of C dl values (independent of the activity of CFP 54 ) and C s . The LSV curves normalised by ECSA are given in Fig. SI-6(b). The ECR products, collected after two hours of CO 2 electrolysis at −0.83 V vs RHE, were qualitatively and quantitatively analysed   with the help of GC and NMR (Fig. SI-10). As depicted in Fig. 6a, the presence of H 2 , CO, HCOOH, CH 4 , and C 2 H 4 as the main products for all the catalysts is underlined. Figure 6b depicts the comparison of selectivity towards main Carbon-products (C-products), determined at −0.83 V vs RHE using Eq. 5, for all the samples. CC shows the largest FE, 66%, for C-products (besides FE of ∼30% for H 2 ) along with the highest selectivity for ethylene formation (FE of ∼26%). Moreover, the highest HER suppression was demonstrated by CC amongst the studied Cu-electrodeposits. CS catalyst shows the maximum FE of 49% for C-products with the formation of CO (FE ∼20%), CH 4 (FE ∼15%), HCOOH (FE ∼12%) and C 2 H 4 (FE ∼11%), depicting the second-best performance. While CO is the major carbon-product formed over CA, CS, and CN catalyst, formation of ethylene over CO is favoured over CC.
Further, the effect of applied potential, from −0.63 to −1.03 V vs RHE (interval of 100 mV), on the distribution of ECR products was studied for CC and CS. As shown in Fig. 6c, CC shows an increased FE for ethylene; highest at −0.93 V (FE ∼30%), with the applied potential. The formation of small amounts of ethane is also seen on CC after −0.73 V. The formation of these hydrocarbons (methane, ethane, and ethylene) increases with the potential applied till −1.03 V. At higher negative potential, the FE for hydrocarbons decreases due to an increase in the competitive HER. A similar trend is seen on CS catalyst, but without the formation of ethane and with the higher selectivity towards CO, particularly at lower negative potentials. The highest FE for CO (FE ∼15%) was observed at −0.83 V for CC and at −0.73 V (FE ∼20%) for CS, which decreases as the FE for the hydrocarbons increases for CC. The higher selectivity towards multi-carbon ECR products on CC is attributed to the optimum adsorption of *CO intermediate on its surface at these potentials. It gives an adequate residence time for C-C coupling reaction to take place that is essential for higher end hydrocarbons. Further, the ECR activity of the presented catalyst, CC, was compared with the Cu nanocubes and oxide-derived Cu electrocatalyst reported in the recent literature studies and it shows comparable ECR activity by CC towards ethylene formation. (Table  SI-3).
To check the stability of the CC,7 hrs of continuous CO 2 electrolysis along with gaseous product quantification (i.e., CO, CH 4 , and C 2 H 4 ) and their faradaic efficiency calculation was done at the potentials where each of the catalysts depict their best ECR activity, i.e., −0.93 V and −0.73 V vs RHE for CC and CS respectively, as shown in Figs. 7a and 7c. For this, the catalystmodified CFP electrodes were analysed for the morphological and crystallographic changes in CC. The crystal phases corresponding to Cu 2 O are seen only with very low intensity in the XRD patterns depicted in Figs. 7b and 7d. While the peaks for metallic Cu have increased intensity, suggesting in situ transformation of Cu 2 O into Cu 0 at the applied negative potentials. The Cu + /Cu 2+ oxide films are known to be unstable in the high negative potential range of −0.2 to −1.1 V that is commonly used for ECR. The post-ECR ratio of the Cu to Cu 2 O content on the working electrodes for two of the best performing catalysts viz. CC and CS was calculated from XRD patterns. CC had a higher Cu/Cu 2 O (111) ratio of 79 while for CS the same ratio was only 5 (Figs. 7b and 7d).
A comparison of the ratio of Cu 2+ to combined Cu 0 -Cu + in the spent and fresh CC catalyst was also calculated from XPS as given in Fig. SI-7, which corroborate the results from XRD analysis. A similar decrease in the percentage of copper oxide components is revealed also by the decrease in the peak area percentage corresponding to lattice oxygen (Cu-O) in the O 1s XPS. The dissolution of DDHA present in the matrix of the electrodeposit might have also taken place during CO 2 electrolysis as the peak corresponding to O c is diminished for the spent catalyst. Traces of Cu 2 O might be present due to the air oxidation of samples during ex situ XRD and XPS measurements. This is in accordance with the literature reports that suggest the strong effect of Clin the electrolyte bath on the structure of Cu-electrodeposits, which accelerate the deposition and dissolution process via an electron-bridge mechanism. 55 FESEM images suggest considerable changes in the morphology of the catalyst post-ECR ( Fig. SI-8), where the cubic morphology of CC was noted to transform into roughened surfaces with blunt edges. Grosse et al. reported that these dynamic changes of the nanocubes in the composition, chemical state, and the morphology arising from the reduction of copper oxide to copper coupled with CO 2 reduction trigger better electrochemical reaction due to the presence of more roughened surfaces in the place of smooth and flat topographies. 56 Since, the catalytic activity of any material is a manifestation of its structure-property relationship, treatment of electrodes in the presence of Cl − ions is expected to play an important role in the formation of catalytic sites for ECR. 57 Particularly, the in situ oxidederived Cu (either from CuO or Cu 2 O or both) is known as an efficient catalyst for the electron transfer towards the generation of hydrocarbons via ECR. 15,58 Besides, the electroreduction of Cu is mostly accompanied by the formation of various lattice defects and the dissolved O atoms tend to occupy these defects and are termed residual or subsurface oxygen. The presence of subsurface oxygen formed during in situ reduction of Cu-oxides is also reasoned to be one of the factors towards enhanced selectivity for ethylene on CC. 59 Especially, these residual subsurface oxygen species can reduce σ-repulsion towards CO (an intermediate) adsorption, thereby increasing the CO coverage on the catalyst surface favouring C-C coupling reactions. 60 Moreover, the single crystals of Cu(100) are reported for highly selective ethylene formation at low overpotentials due to lower barrier for Cu-CO* (* represent the surface adsorbed species) that leads to greater CO coverage favouring CO*-CO* coupling at low bias. 8 The morphology and composition of catalyst also play a vital role towards ECR In addition to the cubic morphology, CC catalyst also depicts large number of Cu(100) faces 61 and hence would have larger number of sites selective for ethylene. As per theoretical calculations reported elsewhere, 62 Cu(100) square sites show significantly stronger binding of *OCCO and *OCCHO intermediates  than the other facets like (111) and stepped sites of (211) are proven. These observations support the selective formation of ethylene on the CC catalyst over the other electrodeposits studied in the present work.
Additionally, surface roughness gives a measure of the uncoordinated active sites which is also known to be related to electrocatalytic activity. A catalyst with high surface roughness, which can be calculated from the roughness factors, will have increased number of under-coordinated Cu sites, and thus possess higher adsorption capability than the highly coordinated sites. 18 The reported studies have also shown that the highly porous surfaces of Cu electrocatalysts are more selective towards C 2+ species and towards lowering the HER. 63 Moreover, an in situ reduction of copper oxides to metallic Cu during ECR develops concave defects on Cu(100), which serves as active site for C-C coupling reactions. Figure 8 gives a schematic representation of the proposed mechanism.
The CS electrodeposits also gives comparable ECR activity to that of CC due to the dendritic structures exhibiting increased electrochemical surface areas thus exposing large number of uncoordinated active sites. This structural effect contributes to the mass transfer and electron transport enhancement and increases the adsorption sites which are capable of stabilising *CO leading to C 1 or C 2 pathway. 64 The CN and CA electrodeposits depict poor ECR activity as seen in Fig. 6a, where CO is the major product and hydrocarbons are formed with only very low FEs. The clustered morphology and low ECSA can be argued to be the reasons for the same. Thus, an enhanced ECR selectivity of ethylene on CC can be attributed to the higher Cu + /Cu ratio, cubic morphology, presence of subsurface oxygen, and enhanced surface roughness of the catalyst. It is pertinent to mention here that the organic compound could exert profound impacts on the catalysts' activity, as has been widely reported in the literature. For example, Sargent and co-workers 65 reported that the functionalization of the electrocatalyst surface with organic molecules can stabilize intermediates for more selective CO 2 reduction to ethylene conversion. Therefore, it is necessary to study if the presence of trapped DHAA have any profound effect on the catalysts' activity used in the present study. For this, the relatively better performing, CC and CS catalysts, among the four studied electrocatalysts in the present work were studied in detail (Fig SI-9). The FTIR spectra on both CC and CS show peaks at 1030, 1125, and 1750 cm −1 corresponds to C-O-C, C-O-H, and C=O, respectively, which are characteristic peaks for ascorbic acid. Similarly, C 1s XPS spectra also shows the characteristic peaks for C-C and C=O at 285.7 and 288.2 eV, respectively. 66 Further, the TGA analysis clearly show that the quantity of DHAA that gets precipitated out along with the Cu electrodeposition is much higher for CC than CS (Fig. SI-3). This combined analysis via FTIR, XPS, and TGA measurements suggests the presence of larger content of DHAA on the catalysts surface. On the contrary, the ECR performance of the two catalysts CC and CS does vary significantly, at least not proportional to DHAA content in the catalyst. Thus, the considerable impact from the trapped DHAA on the catalytic activity of CC can be ruled out in the present study.

Conclusions
In summary, Cu-nanoarchitectures electrodeposited from Cu-salt precursors containing varying anions (viz., acetates, chlorides, nitrates, and sulphates) using pulse galvanostatic electrodeposition method were studied for their electrocatalytic CO 2 reduction (ECR) activity. The electrocatalyst obtained copper chloride (CC) demonstrate the best ECR activity, showing high FE for CO, CH 4 , C 2 H 4 , and formic acid and the highest selectivity for ethylene formation. The detailed analysis, viz-a-viz, physicochemical properties and ECR activity, suggest that the presence of Clions play an important role in generating more catalytic active sites for CO 2 adsorption and activation. The formation of well-defined facets of Cu 2 O and Cu (as seen from XRD), subsurface oxygen, and cubic morphology has been attributed to these results. CS catalyst electrodeposited from the copper sulphate also show good ECR activity, due to Cu 2 O facet growth and dendritic structure giving high surface area, but with a poor selectivity for ethylene formation.