Unraveling the role of phase engineering in tuning photocatalytic hydrogen evolution activity and stability

Jinjin Yi,Zhou Zhou,Yu Xi,Gnghu Zhou,Guoxing Zhng,Li Li,Xuyu Wng,Xingwng Zhu,Xiozhi Wng,e,f,?,Hun Png

a College of Environmental Science and Engineering,Yangzhou University,Yangzhou 225127,China

b School of Chemistry and Chemical Engineering,Yangzhou University,Yangzhou 225127,China

c Institute for Energy Research,Jiangsu University,Zhenjiang 212013,China

d School of Environmental and Chemical Engineering,Jiangsu University of Science and Technology,Zhenjiang 212003,China

e Joint International Research Laboratory of Agriculture and Agri-Product Safety of Ministry of Education of China,Yangzhou University,Yangzhou 225127,China

f Jiangsu Collaborative Innovation Center for Solid Organic Waste Resource Utilization,Nanjing 210095,China

Keywords:Phase engineering Hydrogen evolution Photocatalysis Cocatalyst Surface reconstruction

ABSTRACT In this work,taking NiSe2 as a prototype to be used as cocatalyst in photocatalytic hydrogen evolution,we demonstrate that the crystal phase of NiSe2 plays a vital role in determining the catalytic stability,rather than activity.Theoretical and experimental results indicate that the phase structure shows negligible influence to the charge transport and hydrogen adsorption capacity.When integrating with carbon nitride(CN) photocatalyst forming hybrids (m-NiSe2/CN and p-NiSe2/CN),the hybrids show comparable photocatalytic hydrogen evolution rates (3.26 μmol/h and 3.75 μmol/h).Unlike the comparable catalytic activity,we found that phase-engineered NiSe2 exhibits distinct stability,i.e.,m-NiSe2 can evolve H2 steadily,but p-NiSe2 shows a significant decrease in catalytic process (～57.1% decrease in 25 h).The factor leading to different catalytic stability can be ascribed to the different surface conversion behavior during photocatalytic process,i.e.,chemical structure of m-NiSe2 can be well preserved in catalytic process,but partial p-NiSe2 tends to be converted to NiOOH.

Recent studies have demonstrated that,in addition to the composition,size,dimensionality,defect,and facet,crystal phase of nanomaterials plays a crucial role to their catalytic functions [1–3].Tuning the phase structure of nanomaterials can change the atom arrangement,leading to the regulation of their physical/chemical properties and optimization of catalytic performance [4,5].With the deepening of relevant understanding,phase engineering is now considered as an effective strategy to optimize materials for efficient catalysis.However,it is still urgent to do research work regarding the following aspects: (1) developing general synthetic methods for the controllable synthesis of desired phases,(2) unmasking the mechanism about the phase transition process,and(3) uncovering the inner mechanism for phase-dependent catalytic activity,stability and selectivity [3,6].

Great efforts have been devoted into material design,controllable synthesis and mechanism exploration,to boost the development of phase engineering toward catalysis.For example,semiconducting 2H-MoS2and 2H-MoSe2have been widely explored as catalysts for hydrogen evolution reaction,but their performances are limited by the poor conductivity and limited active sites at edges [7–9].Phase engineering strategy can prepare metallic 1TMoS2and 1T-MoSe2,which shows increased charge transfer capacity and active basal plane (chemical insert in 2H phase) [10,11].As such,improved hydrogen evolution catalysis can be realized.Except for phase transition to metallic 1T phase,phase engineering of MoS2from crystalline to amorphous phase can also improve the catalytic hydrogen evolution performance [12].CO2reduction requires more than efficiency,high selectivity to a specific product is of great significance [13].Phase engineering is also verified that it can affect the selectivity in CO2reduction reaction.It was found that,Au nanorods with a well-defined fcc-2H-fcc phase structure show improved CO production selectivity compared to fcc phase Au nanorods in electrocatalytic CO2reduction reaction [14].Deep mechanism study revealed that the reaction energy barrier in the formation of?COOH is decreased at 2H surface and 2H/fcc interface compared to fcc surface,resulting in improved selectivity and activity.Previous case studies have highlighted the fascinating characteristics of phase engineered nanomaterials,and how the catalytic functions are influenced by phase engineering.Despite many phase-engineered nanomaterials have been explored,the understanding of the phase-dependent catalytic performance including activity,selectivity and stability is nowadays still under development.

Photocatalytic hydrogen evolutionviawater splitting is considered as a good research orientation,aiming to mitigating energy and environmental problems in the present century [15–17].Similar to other typical photocatalytic reactions including CO2reduction and N2activation,the realization of efficient and stable water splitting is determined by the rational design of catalyst structure,to promote charge separation and molecule adsorption/activation[18,19].Transitional metal dichalcogenides (TMDs) can serve as cocatalyst on semiconductor surface to boost the photocatalytic hydrogen evolution performance,but their performances are hard to meet the practical requirement regarding both activity and stability,compared to noble metals [6,20,21].Phase engineering of TMDs offers a feasible avenue to tame the electronic structures and thus improving the catalytic performance.Nickel selenide (NiSe2),a neglected material in TMDs family,has shown potential to catalyze water splitting [22,23].There are many reported strategies that can tune the property or functions for improved catalysis,for example nanostructure engineering [24],defect engineering [23],constructing heterostructure [22].However,phase engineering of NiSe2lacks systematic research.Theoretically,there are two different crystal phases according to the atom arrangement (marcasite phase with orthorhombic structure namely m-NiSe2,and pyrite phase with cubic structure namely p-NiSe2),as presented in Fig.1a.However,systematic research about the specific role of phase structure for NiSe2in photocatalytic hydrogen evolution lacks systematic research.

Fig.1.Theoretical simulation and experimental characterization studies of NiSe2 with different crystal phases.(a) Structure models.(b) The calculated hydrogen adsorption Gibbs free energy (Pt is referred from Ref.[31]).(c,d) The calculated band structures.XRD patterns (e) and Raman spectra (f).

In this communication,we report the phase engineering of NiSe2and the application in photocatalytic hydrogen evolution,with aims to: (1) develop a general approach for the controllable synthesis of phase-engineered NiSe2,(2) explore the feasibility of improving catalytic activity/stabilityviaphase engineering,and (3)uncover the origin of phase-dependent catalytic behaviors.We first developed a simple wet-chemical method to prepare m-NiSe2and annealing treatment to obtain p-NiSe2,followed by fine characterization of chemical structures.Then,the catalytic hydrogen evolution performance of NiSe2as cocatalyst was examined by integrating with two-dimensional carbon nitride (2D g-C3N4).Finally,we found that the crystal phase structure of NiSe2plays a vital role in determining the reaction stability,rather than the activity,and discussed the role of phase structure in affecting the activity/stability.

To gain insight into the phase dependent catalytic hydrogen evolution properties of NiSe2,the hydrogen adsorption Gibbs free energy (ΔGH?),density of state (DOS),and band structure and of NiSe2based catalysts were theoretically calculated.Fig.1a and Fig.S1 (Supporting information) illustrate the crystal structures of m-NiSe2and p-NiSe2,clearly showing the different structures at atomic level.TheΔGH?of catalyst’s surface can provide strong information about the interaction between proton and catalyst’s surface.A zero-approachingΔGH?with low energy barrier is beneficial for hydrogen evolution reaction [25,26].In our case,the optimizedΔGH?values of m-NiSe2and p-NiSe2are determined to be 0.48 and 0.49 eV respectively,as shown in Fig.1b and Fig.S2 (Supporting information).These results indicate that both m-NiSe2and p-NiSe2possess appropriate and comparable hydrogen adsorption surfaces,and the hydrogen evolution capacity could not be effectively tuned by phase engineering.

Apart from theΔGH?,charge delivery ability is another important parameter affecting the hydrogen evolution performance.As shown in Figs.1c and d and Fig.S3 (Supporting information),the band structure and DOS of m-NiSe2and p-NiSe2were further analyzed.It can be observed that both pure m-NiSe2and p-NiSe2show electronic states across the Fermi level,demonstrating their metallic characteristics [27,28].As such,it can be theoretically predicted that both m-NiSe2and p-NiSe2possess good electrochemical conductivity from the theoretical perspective.It can also be noticed that,p-NiSe2shows only slightly increased numbers of states at Fermi level compared to m-NiSe2,suggesting the comparable intrinsic conductivity theoretically.In summary,theoretical analysis shows that NiSe2with different crystal phases could both extract the electrons from electrode or light absorber,and possess comparable proton to hydrogen molecule conversion efficiency.

Experimental investigation of the catalytic hydrogen performance for NiSe2relies on the synthesis of NiSe2with different crystal phases.The idea of the synthesis is to synthesize metastable state NiSe2(m-NiSe2)viawet chemical method firstly,followed by phase transition to stable state NiSe2(p-NiSe2) by extra energy input.In this work,we synthesized m-NiSe2using a simple wet chemical oil bath method.Phase transition from m-NiSe2to p-NiSe2can be realized by a general high-temperature annealing treatment.X-ray diffraction (XRD) was firstly employed to provide structure information of NiSe2samples,as shown in Fig.1e.XRD patterns of m-NiSe2and p-NiSe2can both be well indexed to the standard PDF cards (JCPDS 18-0886 and JCPDS 88-1711),which preliminarily confirm the successful preparation of phase engineered NiSe2.The phase transition from m-NiSe2to p-NiSe2can be further evidenced by Raman analysis.As illustrated in Fig.1f,two sharp peaks located at 196.2 cm-1and 214.2 cm-1for Se-Se stretching mode were detected.The active peak shift of m-NiSe2to higher wavelength position for p-NiSe2is powerful evidence for the changed Se-Se bond length,further illustrating the phase transition from pyrite phase to marcasite phase [29,30].The surface chemical states and elemental composition were also explored using X-ray photoelectron spectroscopy (XPS) as shown in Fig.S4(Supporting information),evidencing the fine chemical structure of synthesized NiSe2samples (detailed illustration in Supporting information).

Upon finely characterizing the chemical structures of NiSe2,the next concern of this work is to investigate the cocatalytic function of phase engineered NiSe2in photocatalytic hydrogen evolution.In this work,the evaluation of photocatalytic hydrogen evolution performance is based on the selection of carbon nitride nanosheets (2D g-C3N4) with stable chemical structure,visible light response and flat surface characteristics,as hosting semiconductor [15,16,32–37].The large lateral size and specific area are beneficial for the loading of NiSe2,while the ultrathin thickness can effectively shorten the charge migration pathway from bulk to surface,thereby reducing the charge recombination.It can be observed in transmission electron microscopy (TEM,Fig.2a)image that pristine 2D g-C3N4shows transparent and wrinkled nanosheet morphology.Atomic force microscope (AFM) image and corresponding height profile (Figs.2b and c) show the lateral size(～5 μm) and thickness corresponding to ～2 atomic layers (～1.08 nm) of 2D g-C3N4[38].The construction of a typical semiconductor/cocatalyst reaction system is based on the effective charge migration of photogenerated charge carriers from semiconductor to cocatalyst.

Fig.2.Structural characterization of 2D g-C3N4 and work function simulation.(a) TEM image.AFM image (b) and corresponding height profile (c).(d–f) The calculated work functions of 2D g-C3N4,m-NiSe2 and p-NiSe2.

Considering that the charge transfer between two components relies on the work function difference,the work functions (Φ) of 2D g-C3N4,m-NiSe2and p-NiSe2were theoretically calculated to ensure the potential of presented systems (Figs.2d–f).On the basis that both m-NiSe2(4.56 eV) and p-NiSe2(4.89 eV) have larger work function values than pristine 2D g-C3N4(4.47 eV),the charge migration from 2D g-C3N4to NiSe2after contact can be ensured thermodynamically.

The fabrication process of hybrid photocatalysts is illustrated in Fig.3a.m-NiSe2isin-situgrown on 2D g-C3N4surface by oil bath reaction using NiCl2and Se powder as precursors,namely m-NiSe2/CN.As contrast,p-NiSe2/CN can be obtained by a general annealing treatment on m-NiSe2/CN to induce phase transition.XRD patterns and Fourier transform infrared spectra (FT-IR) of the hybrids in Fig.S5 (Supporting information) confirm the coexistence of NiSe2and 2D g-C3N4,and the introduction of NiSe2does not affect the structure of 2D g-C3N4.In the meantime,the morphology and phase structure of as-obtained m-NiSe2/CN and p-NiSe2/CN were analyzed using TEM and high-resolution TEM (HRTEM).TEM images in Fig.S6 (Supporting information) depict that m-NiSe2and p-NiSe2with similar nanoribbon morphology are loaded on the surface of sheet-like 2D g-C3N4.To disclose the fine phase structure of presented hybrids,HRTEM was performed as shown in Figs.3b and c.The observed lattice spacings of 0.296 and 0.265 nm in Fig.3b correspond to (101) and (111) crystal plane of m-NiSe2,while lattice spacings of 0.297 and 0.266 nm in Fig.3d can be ascribed to (200) and (210) crystal plane of p-NiSe2.The distribution of NiSe2on 2D g-C3N4surface was also analyzed by scanning TEM (STEM) and corresponding elemental mapping (Figs.3d and e).The homogeneous distribution of C,N,Ni and Se elements in the catalysts can be observed,clearly showing that NiSe2with different phases in nanoribbon morphology are anchored uniformly on 2D g-C3N4surface.As contrast,both pristine m-NiSe2and p-NiSe2show seriously aggregated morphology as presented in Fig.S7 (Supporting information),suggesting 2D g-C3N4can serve as a good platform for the uniform loading of NiSe2.XPS spectra of m-NiSe2/CN and p-NiSe2/CN were also analyzed as illustrated in Fig.S8 (Supporting information).The Ni 2p spectra and Se 3d spectra are in good agreement with pristine NiSe2samples,while the C 1s and N 1s spectra are consistent with previous reported g-C3N4[9,15].

Fig.3.Structural characterization of m-NiSe2/CN and p-NiSe2/CN.(a) Schematic illustration of construction process.(b,c) HRTEM images.(d,e) STEM and elemental mapping images.

Considering the charge migration and highly active surface of NiSe2with different crystal phases,we are in position to examine whether NiSe2can serve as cocatalyst advancing photocatalytic hydrogen evolution,and investigate the role of crystal phase structures in catalytic hydrogen evolution.The examination of photocatalytic hydrogen evolution performance was then performed using 2D g-C3N4as light absorber,NiSe2as cocatalysts and TEOA as sacrificial agent under visible light irradiation.As presented in Fig.4a,it can be firstly found that pristine 2D g-C3N4displays a hydrogen formation rate of 0.75 μmol/h.After merging with NiSe2cocatalysts,the hydrogen evolution rates can be greatly promoted to 3.26 μmol/h for m-NiSe2/CN and 3.75 μmol/h for p-NiSe2/CN,highlighting the advantages of NiSe2in promoting charge migration and providing surface active site [20,39].Despite that many previous reports about phase engineered nanomaterials for hydrogen evolution generally show crystal dependent performance [40–43],in our case,the crystal phase structure of NiSe2does not play an important role in determining the activity.The reason for this phenomenon is owing to that the phase engineering on NiSe2leads to negligible optimization on the hydrogen adsorption and charge delivery capacity.Although there is a big gap between NiSe2and other benchmark cocatalysts (e.g.,Pt or MoS2) for the catalytic performance,the significance and finding for this work can be summarized as below: (i) A universal method is developed for the synthesis of phase-engineered NiSe2;(ii) the effects of phase engineering to catalytic stability is demonstrated over NiSe2;and (iii)the origin for phase-dependent stability is uncovered,which can be owing the different surface reconstruction processes.

Fig.4.Hydrogen evolution performance of the catalysts.(a) A comparison of photocatalytic hydrogen evolution rates.(b) Wavelength-dependent hydrogen evolution rates.(c) Transient state fluorescence spectra of the catalysts.(d) Photocatalytic stability evaluation.

Fig.5.Morphology and composition analysis of the NiSe2/CN catalysts before and after photocatalytic reaction.(a,b) HRTEM images of m-NiSe2/CN.(c,d) HRTEM images of p-NiSe2/CN catalyst.(e,f) Schematic illustration of the structural transformation of NiSe2 and the effect to photocatalytic processes.

To determine the specific role of 2D g-C3N4and NiSe2in photocatalytic process,the wavelength-dependent hydrogen evolution performance was measured (Fig.4b).The experimental results illustrate that,H2can be evolved when 2D g-C3N4can effectively absorb light irradiation (380 nm monochromatic light) and thus generate charge carriers.Although the introduction of NiSe2can improve the light absorption in visible light region,not any gas product can be detected under 420 nm and 550 nm monochromatic lights irradiation since NiSe2with metallic characteristic cannot be excited upon light illumination.The results mentioned above strongly demonstrate the light absorber role of 2D g-C3N4and cocatalyst role of NiSe2.

To uncover the origin of the comparable photocatalytic H2evolution activity in NiSe2/CN with different NiSe2crystal phases,the physicochemical properties of the catalysts were performed to provide insightful information,focusing on light absorption,charge separation and interfacial proton reduction.The light absorption capacity can be evaluated by UV-visible diffuse reflectance spectra(DRS),as can be viewed in Fig.4b and Fig.S9 (Supporting information).After the introduction of NiSe2,the light absorption in visible light region can be improved due to the black color characteristic of NiSe2.Given the similar light absorption capacity observed in DRS spectra,the charge generation ability based on light harvesting of two NiSe2/CN catalysts are determined to be comparable.

The charge separation efficiency was further determined by transient-state fluorescence (FL),steady state fluorescence spectra(PL) and photocurrent response.Shorter intensity-average charge lifetimes for m-NiSe2/CN and p-NiSe2/CN compared to pristine CN,can illustrate the enhanced charge separation of 2D g-C3N4after introducing NiSe2as electron trap (Fig.4c) [44].At a meantime,the quenched PL emission (Fig.S10a in Supporting information) and enhanced photocurrent intensity (Fig.S10b in Supporting information) can also support that the loading of NiSe2can extract photogenerated charge carriers from 2D g-C3N4.It should be pointed out that,the charge separation efficiency is also comparable between m-NiSe2/CN and p-NiSe2/CN,based on the characterizations above and theoretical simulations.

The interfacial proton reduction process can be determined by usingΔGH?as descriptor.As discussed in Fig.1b,we have demonstrated that both m-NiSe2and p-NiSe2can serve as active sites for proton reduction to hydrogen,with calculated ofΔGH?of 0.48 eV and 0.49 eV respectively.Taken together,it can be concluded that,the introduction of NiSe2can promote hydrogen evolution since NiSe2can improve charge separation and provide effective active sites.For photocatalytic hydrogen evolution process,the factors affecting performance includes light absorption,charge separation and surface reactivity.In our case,NiSe2with different phases can exactly promote the charge separation and provide active site for surface proton reduction,resulting in improved catalysis compared to pristine CN.Since the charge separation efficiency and surface hydrogen adsorption capacity is comparable,the presented m-NiSe2/CN and p-NiSe2/CN exhibited close hydrogen evolution rates.

The stability of m-NiSe2/CN and p-NiSe2/CN was investigated by performing 5 cycles of photocatalytic reaction (25 h).As presented in Fig.4d,unlike the comparable catalytic activity,it is interesting that m-NiSe2/CN and p-NiSe2/CN show different durability.The hydrogen evolution rate goes steady over m-NiSe2/CN,but almost 50% decrease can be observed in p-NiSe2/CN.Taken the photocatalytic rate measurement result together,it can be demonstrated that,crystal phase engineering on NiSe2holds the key affecting the catalytic stability,rather than the generally mentioned activity.In general,the most common causes of significant stability changes are the loss or deterioration active components.It has been widely reported that,the surface reconstruction of active sites is a common phenomenon using transitional metal dichalcogenides/nitrides/phosphides as catalysts during catalytic process.For example,it has been demonstrated that the sulfides sometimes can only be considered as pre-catalysts,which will be completely converted to (hydrogen) oxide,or form a sulfide-(hydrogen) oxide core-shell structure [45].Moreover,it has been found that,metallic Co4N surface will be transformed to cobalt oxides/hydroxides in electrocatalytic oxygen evolution process [46].The surface structure reconstruction would thus result in different catalytic activity/stability.

To investigate whether surface structures of NiSe2were changing with the ongoing of photocatalytic reaction or not,the surface structure transformation behavior of NiSe2/CN catalysts was investigated using HRTEM based on the p-NiSe2/CN and m-NiSe2/CN samples before and after photocatalytic reactions.Fig.S11 (Supporting information) firstly illustrate that the nanosheet structure of 2D g-C3N4can be preserved without significant loss of NiSe2.As can be further viewed from Figs.5a and b,m-NiSe2layers on 2D g-C3N4surface could be will preserved after photocatalytic process,without significant observation of other components.As contrast,complete structure of p-NiSe2is hardly to be seen after reaction,partial lattice spacing corresponding to NiOOH layers can be found around the surface of p-NiSe2(Figs.5c and d).The results above reveal that,the formation of NiOOH on p-NiSe2surface destroyed the pristine active sites,leading to the significant decrease of photocatalytic hydrogen evolution activity.According to the stability test result,it could also be deduced that,the surface reconstruction from p-NiSe2to NiOOH could reach the steady state after 5 h reaction,since the hydrogen evolution rate becomes relatively stable in 5 h to 25 h (Fig.4d).To provide more evidence for phase engineering induced stability change,cyclic voltammetry (CV) tests of m-NiSe2and p-NiSe2were also performed (Fig.S12 in Supporting information),observing the reduction peak in p-NiSe2corresponding to Ni4+to Ni0.As contrast,not any reduction/oxidation peak can be found in m-NiSe2.The CV results can also indirectly support the stable characteristic of m-NiSe2rather than p-NiSe2in catalytic process.

Taken together,the role of crystal phase engineering toward photocatalytic hydrogen evolution can be uncovered (Figs.5e and f).NiSe2with charge delivery capacity and surface proton reduction ability can serve as an effective cocatalysts for boosting the photocatalytic hydrogen evolution of 2D g-C3N4.Upon light illumination,2D g-C3N4will be excited to generate electron-hole pairs.The photogenerated electrons tend to transfer across the interface to NiSe2surface,and react with adsorbed protons to evolve H2gas product.In our case,phase engineering of NiSe2exhibits negligible influence to charge transport and hydrogen adsorption capacity,leading to comparable photocatalytic activity between m-NiSe2/CN and p-NiSe2/CN.However,experimental results demonstrate that the crystal phase of NiSe2plays a vital role in determining the catalytic stability.In photocatalytic process,the surface structure of m-NiSe2can be preserved,whilst p-NiSe2surface tends to be transformed to NiOOH,resulting in different stability.

In summary,a facile and effective method is developed to synthesize phase engineered NiSe2for application in photocatalytic hydrogen evolution.The hydrogen evolution activity/stability for these different-phased NiSe2catalysts is also systematically investigated.The results show that m-NiSe2and p-NiSe2shows comparable catalytic activity in boosting hydrogen evolution of 2D g-C3N4,owing to the similar light harvesting,charge transport and hydrogen adsorption capacity.m-NiSe2surface is stable during catalytic process while p-NiSe2tends to be converted to NiOOH,so that m-NiSe2exhibits better stability than that of p-NiSe2.These results highlight the role of phase engineering in tuning catalytic stability,not limited to activity,and also,suggest that in the process of preparation and screening of catalysts,the materials with high activity/stability can be selected through the regulation of crystalline phase structure.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgments

The authors appreciate for the financial support by Natural Science Foundation of Jiangsu Province (No.BK20210827),China Postdoctoral Science Foundation (No.2021M700117),National Natural Science Foundation of China (Nos.U1904215 and 41977085),Program for Young Changjiang Scholars of the Ministry of Education (No.Q2018270),Six Talent Peaks Project in Jiangsu Province(No.TD-JNHB-012),and 333 Project in Jiangsu Province (No.BRA2020300).

Supplementary materials

Supplementary material associated with this article can be found,in the online version,at doi:10.1016/j.cclet.2023.108328.