Photocatalytic CO2 reduction to syngas using nickel phosphide-loaded CdS under visible light irradiation

Photocatalytic CO2 reduction is a sustainable pathway to produce syngas (H2 + CO), which is a key feedstock for the production of several important liquid fuels on the industrial scale. However, achieving an appropriate tunable ratio of H2:CO in syngas for commercial purposes is a challenging task. In this work, we present a low-cost and non-noble metal, phosphide-based co-catalyst—Ni2P-loaded cadmium sulfide (CdS) photocatalyst system, for photocatalytic CO2 reduction. As a co-catalyst, Ni2P fosters an efficient charge separation of photoexcited charges generated in the CdS production of syngas. In total, 3 wt.% Ni2P/CdS exhibited exceptional performance of 50.6 µmol g−1 h−1 in the CO evolution rate and 115 µmol g−1 h−1 in the H2 evolution rate, with a syngas composition varying from 2 to 4 in the H2:CO ratio. Furthermore, first-principles density functional theory calculations were performed to study the surface energetics of the catalyst system and the results are found to be consistent with our experimental findings. Indeed, they establish that the composite favors CO2 photoreduction into syngas more efficiently than pure surfaces.


Introduction
In an ideal scenario, the CO 2 production rate in nature should be balanced by its consumption rate to retain and stabilize the CO 2 level in our environment.Unfortunately, with the intensification of human industrialization and the rapid combustion of fossil fuels to fullfill the energy demands, the balance of CO 2 is and continues to be disrupted.Consequently, the global warming caused by the excessive emission of CO 2 has become a pressing issue and a severe threat to the environment [1].Therefore, the immediate mitigation of the increasing CO 2 level is a matter of prime concern, which is required to be addressed and resolved.
Among the available technologies, CO 2 utilization and conversion are considered to be the most attractive and promising solutions.Photocatalytic conversion of CO 2 into value-added fuels is the most sustainable and dual pathway to mitigate the issues of environmental pollution and global climate change, along with the depletion of natural resources in one shot [2][3][4][5].Among the acquired fuels from the photoreduction of CO 2 , the most valuable fuel is the syngas (mixture of CO and H 2 ) because it is a key feedstock or intermediate for the industrial production of chemicals, fertilizers, pharmaceuticals, and liquid fuels such as CH 4 , CH 3 OH, HCHO, HCOOH, and synthetic gas [6].However, the above process has two major limitations in precisely controlling and tuning the suitable composition of CO/H 2 as per the requirement.For example, 2:1 H 2 /CO mixture is required for methanol [7] and Fischer-Tropsch hydrocarbon synthesis [8], whereas 1:1 H 2 /CO is required for the production of aldehydes via the hydroformylation of alkenes [9].Hence, tailoring the ratio of H 2 /CO mixtures is critical, as it limits its practical applicability in industries for large-scale production.Hence, a rational designing of appropriate catalysts with suitable composition and surface properties is the greatest objective.Furthermore, the fundamental understanding of the catalytic process and related energetics are also equally important to elucidate for further advancement in catalytic system design.
Until now, only a few materials have been explored for syngas production.For example, Akhter et al [10] reported syngas production (H 2 : 30.0 µmol g −1 h −1 and CO: 5.8 µmol g −1 h −1 ) from titanium dioxide nanoparticles of size 11 nm under 300 W UV lamp, whereas Lee et al [11] reported dye-sensitized TiO 2 hybrid system under visible light irradiation.More recently, Pollak et al [12] designed visible light photocatalyst BP/Co@C 3 N 4 , which exhibited 5.625 µmol g −1 h −1 of H 2 and 2.5 µmol g −1 h −1 of CO.The search for much more efficient materials is clearly required for improved syngas production.Cadmium sulfide (CdS) is a potentially interesting candidate in this respect, with its narrow band gap (2.3 eV), visible-light responsive characteristic, and nanosize-tunable electronic structure [13][14][15].Unfortunately, the high recombination rate of photogenerated charge carriers in CdS limits its photo conversion efficiency; hence, various strategies have to be adopted to inhibit the prevailing recombination and therefore enhance the efficacy of the process.Co-catalyst loading is known as one of the competent approaches to increase the lifetime of photogenerated charge carriers by promoting the separation and transfer of photoinduced electron-hole pairs [16].Along these lines, various co-catalysts have been explored to enhance the photocatalytic performance of CdS for CO 2 reduction (table of comparison is given in detail in supporting information table 1).For example, Peng et al [17] employed the most promising noble metal, platinium (Pt), as a co-catalyst for CO 2 photoreduction on CdS but it presented a very slight increment in CO production from 0.12 µmol g −1 h −1 to 0.18 µmol g −1 h −1 .Moreover, the scarcity and high costs of noble metals clearly restrict their applicability.Zhu et al [18] deposited Ag metal on CdS by the photodeposition method, resulting in enhancement in the CO production rate by a factor of 3 as compared to pristine CdS, but it did not exhibit significant enhancement in the H 2 production.Amine-functionalized rGO/CdS showed 0.25 µmol g −1 h −1 of CO production along with methane [19].However, no co-catalyst along with CdS has particularly shown the production of syngas.
Metal-rich phosphides are interesting catalysts due to their promising properties, such as high thermal and electrical conductivity and high thermal stability [20].In this work, we report on the use of nickel phosphide (Ni 2 P), as a low-cost and non-noble co-catalyst, to prevent recombination prevalent in CdS for the production of syngas.To our knowledge, this is the first report on CO 2 photoreduction to syngas using TMP as a co-catalyst for CdS.This work describes the production of syngas from photocatalytic CO 2 reduction with 'tunable' CO/H 2 ratio via adjustment of the components of Ni 2 P on CdS with spatially separated co-catalysts to promote charge separation.In total, 3 wt.%Ni 2 P-CdS exhibited an exceptional performance of 50.9 µmol g −1 h −1 of CO evolution rate and 115 µmol g −1 h −1 of H 2 with a syngas composition varying from 2 to 4 in the H 2 :CO ratio.Furthermore, we theoretically investigated the surface structure and energetics to elucidate the mechanism of CO 2 photoreduction to syngas in our photocatalytic system.

Synthesis of Ni 2 P
A uniform solution containing 6 mmol nickel chloride and 30 mmol red phosphorous was prepared by dissolving them in 160 ml ethylenediamine solvent.The resultant solution was transferred to 200 ml teflon-lined autoclave and kept at 200 • C for 24 h.After completion of the reaction, the product was washed with distilled water and ethanol, and then dried.

Synthesis of CdS
For CdS synthesis, 0.1 M of the sodium sulfide solution was added dropwise to 0.1 M cadmium chloride solution and allowed to precipitate for 2 h.After completion of the reaction, a yellow-colored product was obtained, which was washed with distilled water and ethanol and then dried.

Synthesis of CdS/Ni 2 P
For the synthesis of CdS/Ni 2 P composite sonochemical method was applied, where 200 mg of finely ground CdS powder and different wt.% of Ni 2 P were dispersed in ethanol and kept for ultrasonication for 2 h.Dispersed material was then dried at 80 • C and used for further studies.Depending upon the wt.% of Ni 2 P, the samples were labeled as 0.5 wt.%, 1 wt.%, 2 wt.%, 3 wt.%, 5 wt.%, and 10 wt.% Ni 2 P/CdS.Ni 2 P/CdS, in most of the following discussion, refers to 3 wt.%Ni 2 P/CdS, otherwise stated.

Characterization techniques
The synthesized samples were examined by various techniques, such as powder x-ray diffraction (XRD) measurements using a Philips X'Pert PRO diffractometer with nickel-filtered Cu Kα radiation, field emission-scanning electron microscopy (FE-SEM) with Hitachi S-4200, and transmission electron microscopy (TEM, FEI Tecnai F20 FEG with 200 KV).X-ray photoelectron spectroscopy (XPS) measurements were performed using a Thermo K-Alpha XPS instrument with Al-Kα (1486.7 eV) radiation source at room temperature under ultra-high vacuum (10 −8 Pa).XPS data were carbon corrected with the standard C1s peak (284.8 eV).The ultraviolet-visible diffuse reflectance spectra (UV-Vis DRS) of the samples were recorded using a UV-3600, Shimadzu (UV-vis NIR spectrophotometer) equipped with an integrating sphere, and BaSO 4 was used as a reference.Photoluminescence (PL) spectra of the samples were recorded at room temperature on a steady state spectrofluorometer (SHIMADZU RF-6000) with an excitation wavelength of 365 nm.Time resolved PL was recorded in the HORIBA delta flex time-correlated single photon counting system with 375 nm excitation as a source.

Photoelectrochemical measurements
Photoelectrochemical measurements were performed on the IVIUM Technologies electrochemical workstation with a three-electrode system.Photocatalyst-coated FTO glass substrate served as a working electrode, the counter electrode was a platinum wire, and Ag/AgCl was used as a reference electrode.All three electrodes were dipped in 0.1 M Na 2 SO 4 solution.In preparation for the working electrode, a slurry of as-synthesized samples was prepared in which 20 mg of the sample was dispersed in 0.5 ml of ethanol and 20 µl of Nafion solution (5 wt.%).A uniform slurry is prepared by thorough mixing.The obtained slurry is then coated on an FTO glass substrate and dried at 80 • C to obtain a uniform film.On-off transient photocurrent measurements were conducted using LEDs of wavelength 365 nm (power density: 323 µWcm −2 ).

Photocatalytic activity measurements
Photocatalytic activity of the as-prepared samples for CO 2 photoreduction was subjected in 250 ml stainless steel photoreactor with a quartz window in the gaseous phase.A xenon arc lamp with 300 W was used with a UV cutoff filter of wavelength less than 420 nm.In the photocatalytic process, 50 mg of the sample was placed in a reactor.Before irradiation, the reactor was vacuum treated and purged with high purity CO 2 gas (5 ml min −1 ) for 1 h to remove air from the system.CO 2 reduction experiments were performed in the liquid phase with Na 2 S and Na 2 SO 3 as sacrificial agents.Gaseous products obtained during irradiation were analyzed using gas chromatography (Shimadzu Tracera GC-2010 Plus) with Barrier Ionization Detector and He as a carrier gas.For the stability/reusability test, the photocatalyst under test was collected after each run, refreshed by washing and heating at 100 • C and then tested again for its photocatalytic activity.

Computational details
The full density functional theory (DFT) calculations were performed using the Vienna ab-initio simulation package [21] with the projected augmented wave method [21].The exchange and correlation potential is described with the Perdew-Burke-Ernzerhof (PBE) of the generalized gradient approximation together with the D3 Grimme's dispersion correction method [22] to describe the van der Waals interactions.The plane wave cut-off energy for electronic wave functions was set to 500 eV.The structures were optimized until the energy and force components were less than 10 −6 eV and 0.02 eV Å −1 [23].Brillouin zone integration was performed with a Γ-centered k-point mesh.We have considered CdS (110) and Ni 2 P (0001) surfaces for this study, which are found to be stable and nonpolar surfaces [24,25].In addition, the CdS (110) surface was used to construct the Ni 2 P/CdS heterostructure.In the present study, a Ni 2 P/CdS heterostructure is constructed by placing a single layer Ni 2 P at the top of the CdS (110) surface.Different stacking patterns and rotation angles were created to identify the most stable Ni 2 P/CdS heterostructure configuration (figure S11).To decouple the system from its periodic images, 15 Å of vacuum space was adopted along the z-axis.To represent the bulk, the bottom two layers of the system were fixed, and the above layers were relaxed during the optimization.The charge density difference for the heterostructure was calculated as the difference between the charge density of Ni 2 P/CdS, Ni 2 P, and CdS structures and visualized by VESTA.Reactive energy adsorption analysis was used to identify the most active facets, providing an estimate for the peak activity that may be observed experimentally.The electronic binding energies of adsorption for all intermediates are referenced to their gas-phase counterparts.The hydrogen evolution reaction (HER) of the pure heterostructure was calculated using the free energy difference of hydrogen adsorption (∆H), which is given as: where ∆E H refers to the binding energy (BE) of hydrogen and it is calculated as follows: where E slab * H and E slab represent the energies of the pure slab and heterostructure with one H adsorption and the heterostructure complex without H adsorption and E H2 refers to the energy of the hydrogen molecule.

Results and discussion
The XRD patterns of CdS, Ni 2 P, and different hybrid composites of Ni 2 P loaded CdS (0.5 wt.%, 1 wt.%, 3 wt.%, 5 wt.%, and 10 wt.% Ni 2 P) are presented in figure 1. Figure 1(a) shows three prominent peaks at 2θ values, namely 27.0 • , 44.5 • and 51.7 • (PCPDF# 800019), which correspond to the cubic phase of CdS.The presence of broad peaks indicates that CdS has a small crystallite size in the nano domain.Figure 1(b) shows distinct and sharp XRD peaks at 40.94 • , 44.84 • , 47.62 • , and 54.38 • (PCPDF#741385), which confirm Ni 2 P with hexagonal crystal phase [26].The XRD pattern (c) of 0.5 wt% Ni 2 P loaded CdS shows only the peaks corresponding to CdS, and no noticeable peaks corresponding to Ni 2 P were observed.This result may be attributed to the small amount of Ni 2 P loading on CdS and diminished peak intensity relative to the intensity of CdS diffraction peaks.A similar condition was observed in 1 wt.%Ni 2 P-loaded CdS (figure 1(d)).However, as the loading of Ni To study the microstructure of different samples, characterizations using FESEM and TEM were performed.Figure S1(a) indicates the perforated sheet-like morphology of Ni 2 P, and figure S1(b) indicates the nanoparticle-like morphology of CdS.The highest photocatalytic performance was shown by 3 wt.%composite; therefore, it was characterized thoroughly and was studied in greater detail.The FESEM images in figure S1(c) give mixed morphology in the 3 wt.%composite but the presence of Ni 2 P is not clearly discernible, which may be due to low loading amount of Ni 2 P over CdS or due to the resolution limit of FESEM to distinguish between nanosize particles of CdS and Ni 2 P.However, the presence of Ni 2 P is clearly demonstrated by the energy disperse x-ray spectroscopy (EDS) and elemental mapping.Figure 2 exhibits elemental mapping of the Ni 2 P-CdS composite, which indicates that the sample comprises elements Cd, S, Ni, and P with the distribution of Ni 2 P in CdS nanoparticles.EDS showed ∼1:1 stoichiometric ratio of Cd and S, whereas a higher phosphorous content was observed with ∼2:1.47 ratio in Ni 2 P.This can be attributed to the surface enhancement of phosphorus, which is quite apparent in such kind of synthesis process and also seen in the literature [27,28].
Morphology and microstructure of the samples were also characterized using TEM, and the images are shown in figure 3.In figure 3(a), Ni 2 P seems to exhibit irregular sheet-like structure, whereas CdS exhibits aggregates of nano particles (figure 3(b)) and the size of 3-5 nm is seen from the HRTEM images (figures S1(d) and (e)).Figures 3(c  Furthermore, to investigate the surface chemical bonds of composite photocatalysts, the catalysts were characterized using XPS. Figure 4(a) shows the Cd XPS spectra from pristine CdS and Ni 2 P/CdS composites.The peaks with BE of 405 eV and 411.80 eV represent the Cd 3d 5/2 and Cd 2p 3/2 signatures, respectively.These are seen to have been slightly shifted from the locations for pristine CdS to lower BE in the composite case, indicating the nature of interface chemistry and the electron interactions between the Ni 2 P and CdS in the composite case [26,29].Figure 4(b) (red curve) shows the peaks at 161.7 eV and 162.5 eV representing the S 2p 3/2 and S 2p 1/2 energy levels corresponding to sulfur in pristine CdS, which are shifted (blue curve) to the higher BE in the composite case, reflecting the interaction between the CdS and Ni 2 P and strong electron     The above results imply that the Ni-P bond is covalent in nature with charge transfer features, which are favorable for the fast charge transfer from the surface of the CdS sample [29,30].
To investigate the optoelectronic properties of the photocatalysts, ultraviolet-visible diffused reflectance spectroscopy (UV-vis DRS) was performed as the photoactivity is relevant to the light response characteristics of the as-prepared samples.The UV-DRS data for CdS and 3 wt.%Ni 2 P/CdS are shown in figure 5(a).Pristine CdS demonstrated a strong absorption edge located at 532 nm, which corresponds to a band gap of 2.36 eV.Remarkably, the introduction of Ni 2 P in CdS as in a composite is viewed to enhance and extend the absorption (note that the inverse of %R represents absorption) in the region from 540 nm to 800 nm.This increased photo-absorption property is attributed to the low band gap of Ni 2 P (0.5 eV) and related broadband absorption across the visible range loading of Ni 2 P. The improved visible light absorption of the composite ensures the generation of more photoexcited charges required to execute the photocatalytic activity, which can result in enhanced performance.The loading of Ni 2 P in the CdS for all weight ratios shows enhanced absorption in the visible region (absorption spectra given in the supporting information S3(a)); however, it exhibits a very slight change in the band gap values (Tauc' s plot given in S3(b)), indicating a negligible change in the band gap of the host material.The band gap obtained from the Tauc plot for Ni 2 P is 0.5 eV as given in the supporting information S3(c).
We also conducted density of states (DOS) calculations and band structure calculations using DFT to analyze the electronic structure of the materials.Figure S4 illustrates the DOS for both CdS and Ni 2 P-loaded CdS.Our findings reveal a subtle reduction in the bandgap upon the introduction of Ni 2 P into CdS, aligning closely with the experimental results.This agreement reinforces the influence of Ni 2 P loading on the bandgap of CdS.Moreover, CdS shows a direct bandgap at the gamma point, as shown in figure S5, whereas Ni 2 P shows a metallic character because the bandgap calculations are performed at the PBE level, which is in agreement with the theoretical calculations [31].

Photocatalytic activity of Ni 2 P/CdS
The photocatalytic activity of CO 2 reduction over pristine Ni 2 P, CdS, and Ni 2 P-loaded CdS samples was investigated under visible light irradiation and the results are shown in figure 6.In total, 50 mg of the sample was dispersed in aqueous solution of 0.5 M Na 2 S, and Na 2 SO 3 and used as a hole scavenger.After removing air from the setup, high purity CO 2 and water vapor mixture was purged.After every 1 h, the obtained products are analyzed by GC.CO and H 2 are the main products and no other carbonaceous products were obtained.Controlled blank experiments (figure S6) in the absence of light irradiation and catalysts are performed to ensure that the products are solely originated from CO 2 photoreduction and only the photocatalyst is responsible for the output.
Figure 6(a) illustrates the evolution of CO and H 2 in CdS and Ni 2 P/CdS composites under light irradiation.Pristine CdS shows a total CO evolution of ∼1.93 µmol after 5 h, whereas pristine Ni 2 P shows no product formation.Notably, all the composites, where CdS is loaded with different percentages of Ni 2 P, exhibited remarkable enhancement in the photoactivity compared to pristine CdS.The highest CO evolution is seen in the 3 wt.%Ni 2 P-CdS case.As the amount of Ni 2 P in CdS increases, it triggers the evolution of CO almost six times higher in 3 wt.%Ni 2 P-CdS (50.6 µmol g −1 h −1 ) case, as compared to pristine CdS (7.64 µmol g −1 h −1 ).In addition, a H 2 evolution rate of 115 µmol g −1 h −1 is seen in the case of 3 wt.%Ni 2 P-CdS, which is 15 times higher than that of pristine CdS (7.41 µmol g −1 h −1 ).Furthermore, figure 6(a) also indicates that the ratio of CO to H 2 in the syngas mixture produced from the composite can be tuned easily to different ratios from 2 to 4 by just changing the loading of Ni 2 P on CdS.However, the highest syngas with perfect 2.27:1 H 2 /CO ratio is achieved in the case of the 3 wt.%Ni 2 P/CdS composite.Initially, 0.5 wt.% of Ni 2 P loaded on CdS showed low photoactivity for CO production, and the probable reason may be the low loading amount of the co-catalyst.As the Ni 2 P loading increases, more H + ions are easily adsorbed and reduced on its surface, increasing the H 2 evolution rate as compared to CO 2 reduction because proton reduction and CO 2 reduction are competitive processes and Ni 2 P favors H 2 evolution reaction [20].However, with increased loading of 3 wt.% of Ni 2 P, more active sites are available for CdS to transfer photogenerated electrons to Ni 2 P for simultaneous reduction of both CO 2 and H 2 O. Therefore, enhancement in syngas (CO and H 2 both) production is seen in the case of 3 wt.%Kumar et al [26] and Liu et al [32] reported unique structural arrangement associated with 'ensemble effect' in Ni 2 P decreasing the Ni metal sites due to the presence of P. Subsequently, it facilitates moderate bonding of intermediates and products with the surface, resulting in an easy release of H 2 .With further loading (5 wt.% and 10 wt.%) of Ni 2 P, the activity reduces.A possible reason may be the less exposure of light to the photocatalyst and the coverage of surface-active sites due to the high loading of Ni 2 P, which results in a small number of photoexcited electron-hole pairs.Furthermore, the photostability of the catalyst was tested for three cycles using fresh water and CO 2 in each cycle.After three successive cycles (∼15 h), photocatalytic CO evolution decreases from 2.6 µmol to 2.0 µmol retaining 77% of initial photoactivity and H 2 evolution decreases from 7.91 µmol to 5.92 µmol retaining 75% of initial photoactivity.The small decrease in activity may be attributed to the inevitable recovery loss of samples during successive runs.The XRD and SEM images of the spent sample are also presented in supporting information figures S7 and S8, respectively.In particular, there is slight dissolution of Ni 2 P as seen from the spent XRD pattern; however, no obvious change is seen in the morphology.The overall solar conversion efficiency obtained for the composite is 0.79% at a wavelength of 475 nm under monochromatic light, which is higher than the traditional oxide-and sulfide-based systems [33][34][35].

Photocatalytic mechanism of Ni 2 P/CdS
To scrutinize the mechanism of photocatalytic CO 2 reduction over CdS/Ni 2 P, a study of electron-hole pair separation and transfer between the band structures of the two is required.The PL was recorded and the data are presented in figure 7(a).It shows two broad peaks, a tall and broad peak around 534 nm and a broad peak around 650-720 nm.The peak around 534 nm corresponds to the free (band-edge) and bound excitons in CdS; its band gap IS 2.33 eV from the UV data [33].The contribution around 550-570 nm indicates the presence of interstitial sulfur.The broad PL peak at 650-720 nm has been and can be attributed to the sulfur vacancies [36][37][38].The intense PL peak seen in the pristine CdS at 534 nm corresponds to the severe recombination of charges prevailing in it.In contrast, Ni 2 P-loaded CdS shows a drastic quenching in the PL intensity, indicating the role of Ni 2 P in the separation of charges.To establish the charge transfer in some detail, a time-resolved PL spectroscopy study was undertaken.The PL data were fitted using a bi-exponential decay function and the best fit data is provided in table 2. As shown in the figure 7(b), a faster PL decay is observed in Ni 2 P/CdS as compared to pristine CdS, which evidently reflects the fast charge transfer at the interface.There is a decrease in the average lifetime from 2 ns to 0.12 ns, implying an efficient electron transfer from the excited CdS to Ni 2 P.Moreover, the exponential factor increases from 43% for CdS to 68% for Ni 2 P-CdS.This establishes the high efficacy of Ni 2 P as a co-catalyst for prohibiting the recombination  arising in CdS photocatalysts and fast photoinduced charge transfer process in Ni 2 P /CdS for CO 2 photoreduction.
To provide further evidence to prove the efficient charge transfer system, photoelectrochemical measurements were performed.The transient photocurrent response was measured under on-off light illumination for both CdS and CdS/Ni 2 P (figure 8(a)).photo response is seen in the case of CdS/Ni 2 P as compared to pristine CdS, evidently indicating a well-formed interfacial contact between CdS and Ni 2 P thereby improving the charge transfer rate.This shows the promotional effect of Ni 2 P on charge separation and on the resulting photoactivity.Photoresponse measurements were also taken for all the wt.% of Ni 2 P-loaded CdS composites (given in the supporting information figure S9).
Similar results were obtained from electrochemical impedance spectra as shown in figure 8(b).The charge transfer resistance (R ct ) of the generated charge carriers in the Nyquist plot is another important characterization, which elucidates the catalytic activity.The R ct value is seen to decrease from 324 Ω in CdS to 124 Ω in 3 wt.%Ni 2 P/CdS.This indicates that charge carriers face less resistance at the interface between Ni 2 P-CdS and the electrolyte.This result is consistent with the transient photo-response data (figure 8(a)) obtained on coatings of the two materials to elucidate the sweeping of charge.The enhancement in the photo-response is the result of enhanced lifetime of charge carriers due to the efficient separation along with the decreased charge resistance at the interface of composites.The decrease in the PL lifetime and charge transfer resistance in Ni 2 P-loaded CdS composites indicates that Ni 2 P acts as a sink for electron capture, which results in superior photocatalytic activity.
To verify the possibility of charge transfer from CdS to Ni 2 P energy-level diagram was obtained with the help of valence band spectra and DRS.To examine the valence band levels, the valence band spectra were acquired for CdS photocatalyst and Ni 2 P and the same are shown in figures 9(a) and (b), respectively.The system is located at −1.1 eV and −0.9 eV for CdS and Ni 2 P, respectively.From the CB and VB edges, the energy-level diagram of both the materials is obtained as shown in figure 10.
It clearly indicates that the CB of CdS is much more negative than the reduction potential of CO 2 ; indeed, it is a thermodynamically viable photocatalyst.In addition, the Ni 2 P-loaded CdS is much more favorable than the pristine CdS for the photoreduction of CO 2 .In terms of energy levels and its positioning, Ni 2 P energy levels are well matched with CdS in such a way that the photogenerated electrons in CB of CdS are hastily captured by Ni 2 P due to the negative reduction potential of Ni 2 P CB w.r.t. to CdS CB.In contrast, the photo-generated holes produced in the VB of CdS are consumed by the scavenger.Furthermore, the Ni 2 P captured electrons are efficiently and promptly utilized in the process of CO 2 reduction, which would otherwise get recombined with holes in CdS and be unproductive.The reason lies in the negatively placed CB of Ni 2 P w.r.t.CO 2 reduction potentials.Consequently, Ni 2 P-loaded CdS ensures the rapid transfer of photoinduced charge carriers and subsequently its effective utilization in redox reactions, promoting the separation efficiency of photocatalyst carriers and in turn the overall solar-to-fuel conversion efficiency.

Theoretical calculations and interpretation
It is understood from the above mentioned characterizations that Ni 2 P/CdS has thermodynamically and kinetically viable capability for efficient CO 2 reduction.However, to investigate possible mechanisms for syngas production, surface studies of Ni 2 P-loaded CdS photocatalysts are obligatory because photocatalysis is a surface phenomenon.DFT calculations were therefore performed to estimate the stable surface and surface energetics involved in bending the very stable and straight CO 2 molecules for reduction on their surface.In the very first step of CO 2 reduction, the CO 2 molecule adsorbs on the catalyst surface.In the second step, either it directly decomposes into CO and O or is protonated into COOH and subsequently produces CO [34].In both these cases, a moderately exergonic binding of CO 2 on the catalytic surface is desired.Very strong binding or very weak binding of CO 2 to the catalyst surface has the disadvantage of making the subsequent steps to be endergonic that are thermodynamically unfavorable [39,40].Hence, we first studied the adsorption of CO 2 on the respective surfaces-CdS (110) and Ni 2 P (0001), as given in figure S12.It may be noted that we have not calculated any possible barriers for reactions via transition state calculations and the following discussion only reflects the thermodynamic aspects pertaining to end states of the possible reaction pathways.
Our calculations show that the adsorption of CO 2 on the pristine CdS and Ni 2 P surfaces is endergonic.The adsorption energy of CO 2 was calculated to be about 2.8 eV and 1.1 eV, respectively, for CdS and Ni 2 P, as shown in figure 11(a).The thermodynamically unfavorable binding of CO 2 on the pristine CdS and Ni 2 P surfaces explains why we do not observe any notable amount of CO production in these catalysts.A distortion of the stable linear CO 2 molecule into a bent geometry is observed, which is attributed to the strong binding of CO 2 to the pristine Ni 2 P surface.However, the energy gained due to the interaction of CO 2 with the surface is not large enough in comparison to the energy cost to go from linear to further bent geometry (table 3).Whereas in CdS, the CO 2 molecule is weakly physisorbed with no significant distortion.The poor binding of the CO on pristine CdS catalysts can be understood from their narrow and deeply bound d electrons that are not suitable for the chemisorption of the CO 2 molecule.
As evident from the XPS results, there is a direct electronic coupling between Ni 2 P and CdS in the Ni 2 P/CdS composite (figures 4(a) and (b)), which significantly alters the electron density of the binding sites.The same is demonstrated by the theory of charge transfer iso-surfaces (figure 12(b)).As seen from the charge transfer iso-surfaces, because of strong chemical interaction, there is a significant amount of charge depletion/accumulation around Ni/P atoms.As a result, the Ni sites favor the adsorption of CO 2 with no significant distortion in the linear structure, and the adsorption of CO 2 on the Ni 2 P/CdS composite is found to be exergonic at about −0.22 eV.Hence, the direct decomposition of CO 2 into CO and O is found to be exergonic in both pristine heterostructures, thus making the CO 2 adsorption a critical step.
In the case of indirect decomposition of COOH-mediated processes, the protonation of CO 2 into COOH is also exergonic on all these catalyst surfaces as given in figure 11(b).However, further reduction of COOH to produce CO is found to be exergonic on the pristine surfaces, whereas in the heterostructure, this step is slightly endergonic at about 0.23 eV.Thus, the composite catalyst offers suitable sites for the favorable adsorption of CO 2 and other intermediates.Moreover, the reaction mechanism in the heterostructure involves all the reaction steps to be exergonic except only one slightly endergonic step (CO formation step) in comparison to the large endergonic adsorption of CO 2 on the pristine catalyst surface.
HER is an inevitable competing reaction during CO 2 reduction.A more favorable adsorption/desorption of hydrogen to the catalyst surface results in significantly higher hydrogen production than CO 2 reduction.On the pristine catalyst and the composite catalyst, about three times more hydrogen relative to CO is produced due to the favorable interactions of hydrogen with the catalyst surface.On the pristine catalyst surfaces, the H interaction is slightly endergonic, whereas on the composite the H adsorption is more favorable with slightly exergonic binding as given in figure 11(c).
Overall, the pristine Ni 2 P and CdS catalysts do not favor the adsorption of CO 2 (the most important step for the CO 2 reduction reaction to occur) and also suffer from the recombination of the electron-hole pairs generated by the photons.Hence, the pristine catalyst is poor toward the reduction of CO 2 into CO.In contrast, the composite Ni 2 P/CdS overcomes both these issues to some extent to enhance the reactivity of the catalyst.Chemical bonding between Ni 2 P and CdS modifies the charge density of the reaction sites for suitable adsorption of the reactant and intermediate species.Moreover, the heterostructure also stabilizes the electron-hole pairs that are crucial for electron transfer.

Conclusion
In summary, we successfully engineered a hybrid structure (Ni 2 P/CdS) with a metal phosphide-based co-catalyst for the visible light responsive photocatalyst, which delivered the precious fuel-syngas.The critical issue of recombination of photogenerated charges in CdS is taken care of by the loading of Ni 2 P nanosheets, which act as a sink for the captured photoelectrons generated from CdS, ultimately enhancing the photoconversion efficiency.The composite demonstrated a remarkable production of CO (50.6 µmol g −1 h −1 ) and H 2 (115 µmol g −1 h −1 ) in the varying ratio of syngas from 2 to 4 depending on the co-catalyst loading.Hence, replacing expensive noble metal co-catalyst with Ni 2 P not only facilitated the enhancement of the syngas production but also displays control over H 2 /CO syngas ratio.Furthermore, using DFT calculations, we reveal the most stable surface of the two components and the mechanism pathway by which the very stable CO 2 molecule can be adsorbed and reduced to the surface.
This work opens new opportunities to develop carbon neutral selective syngas production by photoreduction of CO 2 by a sustainable pathway with the rational design of photocatalysts.Such studies demonstrate the future of low-cost non-precious element based systems with increased efficiencies for the syngas production.Future work may also involve low-and mixed-dimensional catalysts with engineered interfaces via predictive modeling.
2 P is increased from 3 wt.% to 10 wt.% (figures 1(e)-(g)) in the composite, prominent peaks of Ni 2 P at 2θ 40.94 • , 44.84 • , 47.62 • , and 54.38 • with increased intensity are clearly visible.In the case of 10 wt.% Ni 2 P/CdS (figure 1(g)), all the peaks of the Ni 2 P phase are visible along with the CdS peaks without any shift in the peaks.This confirms the presence of Ni 2 P in Ni 2 P/CdS composites and indicates the unchanged structure of CdS with loading of Ni 2 P.
)-(e) show the morphology of Ni 2 P/CdS composite, where Ni 2 P is covered with CdS nanoparticles.
Figure 3(f) shows TEM d-spacing analysis, which exhibits d-spacing of 0.22 nm (111) and 0.34 nm (111) corresponding to Ni 2 P and CdS, respectively.Morphological information confirms the dispersion of Ni 2 P in CdS including the possibility of formation of junctions between CdS-Ni 2 P.

Figure 4 (
c) shows the Ni 2p spectrum of Ni 2 P with BE of 852.7, 855.5 eV, and 861.48 eV corresponding to the Ni δ+ , Ni 2+ , and satellite peaks of Ni 2p 3/2 , respectively.Furthermore, in figure 4(c) BE of 869.7 eV, 873.9 eV, and 880 eV represent the Ni δ+ , Ni 2+ , and satellite peaks of Ni 2p 1/2, respectively, from the pristine Ni 2 P sample [26, 29].The peak intensities are very weak in the case of the composite as the loading amount of Ni 2 P is very low.The P 2p spectrum (figure 4(d)) shows the peaks at 129.7 eV, 130.8eV, and 133.1 eV, which represent the metal phosphorus bond and oxidized P species due to air exposure, respectively.Similar to Ni peaks, P peaks are also much lower in intensity in the composite.For better clarity about the composite formation, the XPS data of 10 wt.% Ni 2 P-CdS were also recorded.These data are shown is figure S2 wherein the presence of Ni and P in the composite can be seen clearly.This indicates the presence of Ni 2 P in the composite, which is in good agreement with the aforementioned results.

Figure 10 .
Figure 10.Energy-level diagram of the Ni2P-CdS system.

Figure 11 .
Figure 11.The energy diagrams of the CO2 reduction process through (a) direct and (b) through COOH-mediated process to CO on CdS, Ni2P, and Ni2P-CdS and (c) hydrogen evolution reaction process (HER) reduction reaction.

Figure 12 .
Figure 12.The optimized atomic configuration and charge-density difference of the iso-surfaces of the bonded Ni2P/CdS composite.The purple, yellow, light pink, and blue balls represent Cd, S, P and Ni atoms, respectively.Yellow and cyan regions correspond to electron accumulation and depletion, respectively (iso-surface of 0.0025 e Å −3 ).