Manipulating Ferroelectric Polarization and Spin Polarization of 2D CuInP2S6 Crystals for Photocatalytic CO2 Reduction

Manipulating electronic polarizations such as ferroelectric or spin polarizations has recently emerged as an effective strategy for enhancing the efficiency of photocatalytic reactions. This study demonstrates the control of electronic polarizations modulated by ferroelectric and magnetic approaches within a two-dimensional (2D) layered crystal of copper indium thiophosphate (CuInP2S6) to boost the photocatalytic reduction of CO2. We investigate the substantial influence of ferroelectric polarization on the photocatalytic CO2 reduction efficiency, utilizing the ferroelectric-paraelectric phase transition and polarization alignment through electrical poling. Additionally, we explore enhancing the CO2 reduction efficiency by harnessing spin electrons through the synergistic introduction of sulfur vacancies and applying a magnetic field. Several advanced characterization techniques, including piezoresponse force microscopy, ultrafast pump–probe spectroscopy, in situ X-ray absorption spectroscopy, and in situ diffuse reflectance infrared Fourier transformed spectroscopy, are performed to unveil the underlying mechanism of the enhanced photocatalytic CO2 reduction. These findings pave the way for manipulating electronic polarizations regulated through ferroelectric or magnetic modulations in 2D layered materials to advance the efficiency of photocatalytic CO2 reduction.


SUPPLEMENTARY NOTES 1. Synthesis and crystal growth of CIPS sample
CIPS crystals were grown by CVT method using iodine as the transport agent.The growth of high-quality layered CIPS was achieved by preparation of the powdered elements of Cu (99.99% purity), In (99.99% purity), P (99.999% purity), and S (99.999% purity) with stoichiometry (Cu:In:P:S=1:1:2:6) together with an appropriate amount of I 2 (10 mg/cm 3 ) were put into a quartz ampoule (20 cm in length and 3 cm in inner diameter).Total 10 gram of the powdered mixture with the weight Cu = 1.4686 g, In = 2.6535 g, P = 1.4317 g, and S = 4.4462 g were used.The quartz ampoule with powdered mixture was directly cooled with liquid nitrogen and then sealed in a vacuum environment at ~10 -6 Torr.Two-step heating process of the layer compound was used for firstly placing two quartz ampoules in a horizontal three-zone furnace at a constant temperature of 600 C for two days to get synthesized reaction of the starting material, and then setting as 660 C (heating zone) and 600 C (growth zone) with a gradient of -2 C cm -1 to the two quartz ampoules for the single-crystal growth.The role of transport agent I 2 is to facilitate the vapor transportation of CIPS from high-temperature end to lower-temperature end for nucleation and growing crystals.The growth reaction kept 360 hours for growing crystals.After the growth, some big and orange-yellow (or brown) like CIPS layered single crystals with area size up to ~ 2-3 cm 2 and a thickness up to 300 μm was obtained.
The other batch of small powered-like CIPS crystals was also formed together with the big area crystals.The small crystals have a size area about tens micrometer and a thickness about hundred nanometer.Powdered XRD experiment confirmed that all the as-grown crystals are crystallized in the monoclinic structure of C c symmetry.The obtained lattice constants of CIPS are a = 6.09Å, b = 10.56 Å, c = 13.62 Å, and  = 107.1,respectively.

Preparation of V S -CIPS
V S -CIPS is prepared in the furnace with a quartz tube and controlled argon atmosphere by a flowmeter and vacuum pump station.The annealing temperature is set at 450 ℃ and held for 1 hour.To preserve the ferroelectric polarization, the annealing temperature decreases to 380 ℃ for 10 minutes.

Corona poling
A homemade DC needle-plate configured corona poling apparatus is constructed with an array of needles and an aluminum plate.The needle-to-plate gap is set to 2 cm, and the entire setup is leveled via level meters.An external high-voltage source is connected to the needle electrodes while the plate electrode is attached to a grounded microammeter to monitor the received current.The system is confined in a grounded box to minimize atmospheric fluctuations and collect excess ions.During poling, samples are placed beneath the emitting needles to ensure the successful deposition of charged ions.The samples are poled for 30, 45, and 60 minutes with the voltage source (~15 kV) adjusted to maintain the current value at around 80 μA.

Material characterizations
The XRD is performed by Bruker D8 Venture X-ray diffractometer (Cu Kα1) with each step of 0.01 degrees and a scan rate of 0.5 s per step.The TEM sample is prepared by exfoliating CIPS bulk crystal into micro-flake on a copper grid.The HRTEM image and SAED pattern are performed in the FEI Tecnai G2-F-20 system.Raman is performed by a homemade system equipped with continuous Nd:YAG 532 nm laser, Olympus microscope (50X objective lens), Andor Kymera 193i-B2 spectrometer, and Andor iDus416 low-noise detector.UV-Vis absorbance spectroscopy is performed by PerkinElmer Lambda 365 UV-Vis spectrophotometer with double-beam design and deuterium and tungsten-halogen light sources.The EPR sample is prepared in powder form, ground from bulk crystal, and characterized by Bruker EMXplus-10/12.The Agilent 7800 ICP-MS is used for vacancy concentration determination.

Piezoresponse force microscopy
The PFM sample is exfoliated with tape (SPV-224SRB, Nitto) on the heavily doped silicon substrate.PFM measurements were acquired using a commercial scanning probe microscope system (Multimode 8, Bruker) with a Nanoscope Controller V. PFM images, offfield hysteresis loops, and cKPFM curves were all carried out under contact-resonance mode with the commercial Pt/Ir-coated tips with spring constant of 2.8 Nm -1 (NANOSENSORS PPP-EFM).The tip was driven with an AC voltage amplitude of about 0.5 V and was working at a contact-resonance frequency of about 300 kHz.The off-field hysteresis data and cKPFM curves were obtained via the switching spectroscopic technique with an arbitrary waveform generator (G5100A, Picotest).The temperature-dependent PFM images and quantitative piezoelectric coefficient d 33 were collected under off-resonance mode by the commercial Pt/Ir-coated tips with spring constant of 7.4 Nm -1 (NANOSENSORS PPP-NCSTPt).The thermal application module was used to achieve temperature-dependent measurements with a thermal application controller (TAC, Bruker).During the heating process, an excitation signal of amplitude 1 V working at 7 kHz was applied to capture the temperature-dependent PFM images.A conductive tip was calibrated by a standard sample to extract the quantitative d 33 values of the sample at individual temperatures.

Pump-probe spectroscopy
Time-resolved pump-probe spectroscopy in this study was executed using a dual-color pump-probe system, with a 400 nm (3.1 eV) pump and an 800 nm (1.55 eV) probe.The laser light source was a Ti:sapphire laser with a 5.2 MHz repetition rate, 800 nm wavelength, and 70 fs pulse duration.The pump and probe fluences were set at 77 and 8 μJ cm -2 , respectively.The temperature-dependent measurements were carried out in a standard cryostat at ~10 -3 torr.

In situ DRIFT spectroscopy
The in situ DRIFT spectra were recorded using a Bruker Tensor 27 FTIR spectrometer with a HgCdTe detector for CO 2 adsorption process and photocatalytic CO 2 reduction reaction. 1,2 Al spectra were acquired with 64 scans with a spectral resolution of 4 cm -1 . 3The acquisition time is ~30 s for each spectrum.A mixture of 4 mg of catalyst and 196 mg of dried KBr powders was used for in situ DRIFT measurement.The sample composite was purged with Ar gas for

Figure S3 .
Figure S3.(a) AFM topography image, (b) OP PFM amplitude image, and (c) OP PFM phase image of the CIPS.

Figure S4 .
Figure S4.cKPFM curves carried out on the CIPS as a function of V read under a series of V write .

Figure S5 .
Figure S5.The calibration curves for (a) CO and (b) CH 4 gas concentrations determined by GC-MS.

Figure S7 .
Figure S7.XRD pattern of the V S -CIPS.

Figure S8 .
Figure S8.Raman spectrum of the V S -CIPS.

Figure S9 .
Figure S9.The comparison in CO and CH4 yields of pristine CIPS without and with a magnetic field of 300 mT.(a) CO and (b) CH4 yields after 1, 2, 4, and 6-hour reactions.

Figure S10 .
Figure S10.XAS Cu K-edge spectra with various Cu-oxides references (without applying a magnetic field).

Figure S11 .
Figure S11.Synergistic manipulation of ferroelectric polarization and spin polarization of V S -CIPS.(a) OP PFM image and (b) CH 4 yield of of the CIPS after annealing and post-poling processes.