Tuning the Photoresponse of Nano‐Heterojunction: Pressure‐Induced Inverse Photoconductance in Functionalized WO3 Nanocuboids

Abstract Inverse photoconductivity (IPC) is a unique photoresponse behavior that exists in few photoconductors in which electrical conductivity decreases with irradiation, and has great potential applications in the development of photonic devices and nonvolatile memories with low power consumption. However, it is still challenging to design and achieve IPC in most materials of interest. In this study, pressure‐driven photoconductivity is investigated in n‐type WO3 nanocuboids functionalized with p‐type CuO nanoparticles under visible illumination and an interesting pressure‐induced IPC accompanying a structural phase transition is found. Native and structural distortion induced oxygen vacancies assist the charge carrier trapping and favor the persistent positive photoconductivity beyond 6.4 GPa. The change in photoconductivity is mainly related to a phase transition and the associated changes in the bandgap, the trapping of charge carriers, the WO6 octahedral distortion, and the electron–hole pair recombination process. A unique reversible transition from positive to inverse photoconductivity is observed during compression and decompression. The origin of the IPC is intimately connected to the depletion of the conduction channels by electron trapping and the chromic property of WO3. This synergistic rationale may afford a simple and powerful method to improve the optomechanical performance of any hybrid material.


S1. Electron-hole pairs generation:
Here, we used a 532 nm laser with a photon energy of (E=hc/, 2.34 eV) that is little smaller than the band gap of WO 3 (2.6 eV) but much higher than that of CuO (1.2-1.4 eV), and can easily generate electron-hole pairs, introducing a positive or negative photoelectric response depending on a variation of the band gap and hetero-junction with pressure. As CuO has a band gap smaller than the laser energy, the electron-hole pairs are mainly generated in CuO at ambient conditions. By increasing pressure, the band gap of WO 3 decreases 1 but the material remains as a semiconductor at least up to 25 GPa. 21,25 On the other hand, by comparing CuO with isomorphic AgO, which has similar compressibility, one can predict that the band gap of CuO would not collapse within the pressure range of our interest 2 , and might decrease slowly. Thus, changes in the electronic structure of WO 3 will dominate changes in the photoresponse. In particular, the above described subtle change in W-O-W interatomic bonding and WO 6 octahedral tilting, can favor a pressure-induced band-gap reduction in WO 3 that is consistent with the dramatic increase in the photoresponse (270%) at 2.3 GPa.

S2. PPC and trapping charge carrier:
The high-pressure phase of WO 3 adopted a highly distorted structure where the appearance of a new peak and elongation of the W-O bonds were confirmed by XAS spectra (Table 1: G2). Such an elongation helps to increase the oxygen vacancy concentration by decreasing the vacancy formation energy, which leads to the slow recovery of charge carriers and agrees very well with our observed PrPPC (Figure 1 The band at 950 cm -1 is due to W 5+ -O bonds at 12 GPa under compression and 6 GPa under decompression, however, at 0 GPa after decompression it is mainly due to the W 6+ -O bonds. The broadening of this band at 0 GPa after decompression correlates well with the Bragg's peak broadening in Fig. S1(b).

S5. X-ray absorption spectroscopy measurements
The pressure dependence of the W L 3 -edge XANES spectra is shown in Fig. S3, and the experimental W L 3 -edge EXAFS spectra and their Fourier transforms are presented in Fig. S4.
The contribution from the first coordination shell of the tungsten atoms ( Fig. 5(a) Manuscript) was isolated by the Fourier filtering procedure in the R-space range from 0.5 to 2.4 Å, and the true radial distribution function (RDF) g W-O (R) for the W-O atom pairs ( Fig. 5(b) Manuscript) was obtained using the regularization method, 4 which has an advantage in cases of arbitrary structural disorder. 61 The calculations were performed using theoretical scattering amplitude and phase shift functions for the W-O atom pair obtained by the ab initio real-space multiplescattering code FEFF8.5L. 63 The photoelectron inelastic losses were accounted for within oneplasmon approximation using the complex exchange-correlation Hedin-Lundqvist potential. 64 The best-fit of the experimental W L 3 -edge EXAFS signals from the first coordination shell ( Fig.   5(a) Manuscript) was performed in the k-space range from 1.5 to 13 Å -1 . The amplitude of the theoretical EXAFS spectra was scaled by the constant factor S 0 2 =0.69.
The obtained RDFs g W-O (R) were decomposed into two or three Gaussian contributions ( Fig.   5(b) Manuscript), which were used to estimate the type of WO 6 octahedra distortion and its variation upon compression. The obtained results allowed us to separate all RDFs into three groups G1-G3 (Table 1   . Figure S4. Pressure dependence of the W L 3 -edge EXAFS spectra of WO 3 /CuO and their Fourier transforms. It is known that the oxidation state of tungsten in WO 3 is 6+ but at high pressure, it can turn to a bluish color that is due to the formation of W 5+ color centers by gaining electrons 5,6 .

S8: Relation between the high pressure and the carrier density of WO3
Upon compression, band structures and band gap changes, so it is obvious the carrier density will change. Because the band gap was reduced, the effective mass is modified. Mostly in high pressure, with change of band gap the carrier density change. In WO3, resistivity decreases with pressure as we have discussed in main text and we have added new result (See Figure S9), so obviously carrier density will be change. The pressure induced resistivity drop is caused by an increase in carrier concentration, which is related to the increase of the additional energy levels in energy band gap.