Unravelling the Surface Oxidation-Induced Evolution of the Electronic Structure of Gallium

Gallium is widely used in liquid metal catalyst fabrication, and its oxidized species is a well-known dielectric material. In the past decades, these two species have been well studied separately. However, the surface oxide layer-induced impact on the chemical and electronic structure of (liquid) gallium is still mostly unclear because of the extreme fast formation of thermodynamically stable surface Ga2O3. In this study, we used a combination of direct and inverse photoemission complemented by scanning electron microscopy to examine the surface properties of Ga and Ga oxide (on a SiOx/Si support) and the evolution of the surface structure upon stepwise oxidation and subsequent reduction at an elevated temperature. We find oxidation time-dependent self-limited formation of a substoichiometric Ga2O3−δ surface layer on the Ga nanoparticles. The valence band maximum (conduction band minimum) for this Ga2O3−δ is located at −3.8 (±0.1) eV [1.4 (±0.2) eV] with respect to the Fermi level, resulting in an electronic surface band gap of 5.2 (±0.2) eV. Upon annealing in ultrahigh vacuum conditions, the Ga2O3−δ surface layer can efficiently be removed when using temperatures of 600 °C and higher. This study reveals how the surface properties of Ga nanoparticles are influenced by stepwise oxidation–reduction, providing detailed insights that will benefit the optimization of this material class for different applications.


Energy resolution
In all XPS measurements with Mg Ka source, the pass energy for the core level detail spectra measurements was set to 20 eV, resulting in a total energy resolution of approximately 1.2 eV.For the He II-UPS measurements, a pass energy of 5 eV was used, resulting in a total energy resolution of 0.2 eV.The total energy resolution of the IPES setup is determined as 1.

Surface coverage of Ga NPs on SiOx/Si support
The fitting of Ga 3p and Si 2p core level peaks give some insights on the surface coverage of the SiOx/Si support by the Ga NPs.2] The IMFP for both, the Ga 3p3/2 and Si 2p3/2 photoelectrons from Ga and SiOx is 19.85 Å. [3][4][5] The intensity ratio of SiOx/Si is determined as 0.09 by Si 2p core level peaks of the bare SiOx/Si support (Fig. S1a), and we assume this ratio is the same in the subsequent fitting of the overlapping Ga 3p/Si 2p core level region (Fig. S1b) -it can be expected to have only (if at all) a minor impact on the results of the calculation even though this assumption is not applicable to other samples.The surface coverage of Ga is calculated by the fitted peak area of Ga 3p peak from metallic Ga and Si 2p peak from non-oxidized Si.In the calculation, we assume the substrate is only observable in the region without Ga coverage.The dewetting behavior of Ga NPs is expected to diminish the Ga thin layer (≦20 Å) on substrate.For surface oxidized sample, due to the IMFP of Ga 3p and Si 2p core level electrons are at least 2 times larger than the film thickness of Ga2O3-δ and SiOx (≦9 Å), only the peak of metallic Ga peak and non-oxidized Si is taken into account to avoid overestimating the Ga coverage.

Ga-O stoichiometry
The fitting of Ga 3d and O 1s core level peaks was used to derive the stoichiometry of the formed gallium oxide.2] The inelastic mean free path (IMFP) for both, the Ga 3d5/2 ][5] Determination of Ga2O3-δ film thickness In this study, a simple overlayer model is utilized to discuss the Ga2O3-δ layer formation on top of metallic Ga, assuming a mechanism of homogeneous, closed packed oxide film growth. 6The following equation can be used to calculate the Ga2O3-δ film thickness, D: λi, Ga2O3-δ and λi,Ga are the IMFP values in Ga2O3-δ and metallic Ga, respectively for core level I (calculated using the TPP2-M equation 7 with the density and electron configuration of stoichiometric Ga2O3 as the absorbing layer -we considered this the best approximation available due to the lack of reliable parameters for Ga2O3-δ ).The λi,Ga2O3-δ and λi,Ga of Ga 3d photoelectrons are 20.9 and 27.2 Å; of Ga 3p photoelectrons are 19.7 and 25.6 Å; of Ga 2p photoelectrons are 5.3 and 6.3 Å, respectively. 1, 4-5, 8-9Ii, Ga2O3-δ and Ii,Ga are the intensities (i.e., areas) of the Ga2O3-δ and Ga peak contributions, respectively, derived for the core level i (obtained by XPS data fitting, see Figs.S4 and S5).N(Ga)Ga2O3-δ and N(Ga)Ga are the atomic densities of Ga in Ga2O3 (0.038 Atoms per cubic Å) and Ga (0.053 Atoms per cubic Å), respectively. 10It's noted that the formula for D assumes a uniform, closed capping oxide layer, and thus the discrepancies in the oxide thicknesses calculated using Ga 2p (Table S3) and using Ga 3d (Table S4) are related to the observed incomplete coverage of the Ga, which will cause an underestimation of the layer thickness in both cases, with the effect being more pronounced for the more surface sensitive data.
As the oxide coverage of Ga increases, the results of the two calculations converge, with the remaining disagreement possibly attributable to the influence of morphology -i.e., the differing relative surface/bulk contributions of the nanoparticles compared to the smooth layer assumed in the calculation.

S24
Table S1.Analyzer transmission function corrected peak area of the Ga 3p and Si 2p core level peaks of Ga NPs on SiOx/Si support after different treatments steps.The molar ratio of Ga/Si and surface coverage of Ga NPs on SiOx/Si support are derived by correcting the peak area by the respective photoionization cross sections (σ) 2 and the inelastic mean free path (IMFP) [4][5]8 . Thefitting results of Ga 3p/Si 2p core level peaks are shown in Fig. S3-4.

S25
Table S2.Analyzer transmission function corrected peak area of the Ga2O3-δ contributions to the Ga 3d and O 1s core level peaks of the sample oxidized in 1×10 -6 mbar O2 for 240 min and in ambient conditions for 1 month.Photoionization cross sections (σ) 2 and inelastic mean free path (IMFP) values applied for intensity correction are listed.The SiOx and Ga2O3-δ peak contributions are obtained from the fitting result depicted in Fig. S10.
Table S3.Ga2O3-δ/metallic Ga ratio of the respective spectral contributions to the Ga 2p core level and calculated Ga2O3-δ film thickness for the Ga nanoparticles oxidized in 1×10 -6 O2 for different time.
Table S4.Ga2O3-δ/metallic Ga ratio of the respective spectral contributions to the Ga 3d core level and calculated Ga2O3-δ film thickness for the Ga nanoparticles oxidized in 1×10 -6 O2 for different time.
Table S5.Ga2O3-δ/metallic Ga ratio of the respective spectral contributions to the Ga 2p, Ga 3p, and Ga 3d core levels of the sample oxidized in ambient condition for 1 month together with the corresponding film thickness.The peak area is examined by the fitting results shown in Fig. S4 and Fig. S6-7.

7 ×
3 eV via fitting of a measured Fermiedge (EF) of a clean gold film by the following fit function with correction of temperature (T = 300 K) employing the Boltzmann constant (Kb•T = 25 meV):  • ) +  . = 2 × √2 • 2 ×  σtotal is the total Gaussian broadening including instrumental and thermal (Kb•T) broadening, Ef denotes the energy of Fermi-edge.a, b, c, d are dependent variables in the fit function.Quantification All XPS data were fitted and quantified by Winspec (LISE, Université de Paix, Namur), The metallic Ga peak is fitted by an asymmetric (Doniach-Sunjic) profile and the Ga2O3-and SiOx features are fitted by a Voigt profile.All spectra are fitted subsequently with the same constraints (distance and intensity ratio of doublet peaks, FWHM, ratio of Lorentzian and Gaussian contribution, etc.).The peak area of all core level peaks are corrected by the transmission function of the electron analyzer: I and I0 denote to the corrected peak area and original peak area, respectively.Ex denotes to the excitation energy (Mg K = 1253.56eV), Eb refers to the binding energy of the core level peak.The quantitative analyses of certain core level peaks or elements in this study are then processed by following equation: IA and IB denote to transmission function corrected peak area of core level A and B. ωA and ωB denote to respective photoionization cross section of core level A and B. IMFPA and IMFPB refer to the inelastic mean free path of photoelectrons of core level A and B.

Figure S1 .
Figure S1.XPS survey spectra recorded with Mg Ka excitation of Ga/SiOx/Si samples before (black) and after 10 min (red), 30 min (blue), 60 min (green), and 240 min (purple) exposure to 1×10 -6 mbar O2.After the 240 min oxidation, the sample was taken out from UHV condition and oxidized at ambient condition for 1 month.

Figure S2 .S9Figure S3 .
Figure S2.C 1s XPS detail spectra recorded with Mg Ka excitation of Ga/SiOx/Si samples before and after different oxidation times in 1×10 -6 mbar O2.After the 240 min oxidation, the sample was taken out from UHV condition and oxidized at ambient condition for 1 month.

Figure S4 .
Figure S4.Fit analysis of the Ga 3p and Si 2p peaks of the Ga NPs on SiOx/Si support after oxidation in ambient condition for 1 month.The SiOx peak is composed of a spectral contribution of the oxide support and of surface contaminants (presumably a silicate).

Figure S5 .
Figure S5.SEM image of Ga nanoparticles on the SiOx/Si substrate after 240 min oxidation in 1×10 -6mbar O2 and after transfer to the SEM in ambient atmosphere, thus further surface oxidation is expected.

Figure S6 .S13Figure S7 .
Figure S6.Size distribution of deposited Ga particles based on the statistical evaluation of 206 particles in the SEM image shown in Figure S5.

Figure S8 .
Figure S8.Fitting results of the Ga 3d XPS data collected by Mg Kexcitation for Ga/SiOx/Si samples oxidized in 1×10 -6 mbar O2 for different times (0-240 min) and in ambient conditions for 1 month.Note that the Ga 3d doublet cannot be resolved by the used experimental setup and thus one peak (representing the sum of the Ga 3d3/2 and Ga 3d5/2 line) is used to fit each peak contribution.The metallic Ga peak is fitted by an asymmetric (Doniach-Sunjic) profile and the Ga2O3- feature is fitted by a Voigt profile.The broader peak shape used to fit the Ga2O3-δ contribution is tentatively attributed to different oxide environments and/or due to the formed oxide being a less ordered material (compared to the metallic Ga) resulting in varying bond lengths and bond angles -all of which causing BE variations that may increase the FWHM of the Gaussian contribution of the Voigt profile used to fit this spectral component.The high BE background of the spectrum collected for the sample exposed to ambient conditions for one months is dominated by the increased contribution of the O 2s line.

Figure S9 .
Figure S9.Fitting results of the O 1s detail spectrum recorded with Mg K excitation of the as prepared Ga NPs on SiOx/Si support.

Figure S10 .
Figure S10.Fitting results of the O 1s detail spectra recorded with Mg K excitation of Ga NPs on SiOx/Si support after oxidation at (a) 1×10 -6 mbar O2 for 240 min and (b) in ambient condition for 1 month.

Figure S11 .
Figure S11.Examination of the above-VBM spectral feature related to oxygen vacancy derived surface defect states in the valence band region.The linear extrapolation is utilized to derive their approximate position with respect to the Fermi level (EF).

Figure S12 .
Figure S12.UPS (He I) spectra of Ga NPs on SiOx/Si support before (black spectrum) and after surface oxidation at 1×10 -6 mbar O2 for 240 min (red spectrum) and oxidation in ambient condition for 1 month.

Figure S13 .
Figure S13.UPS (He II, red) and IPES (black) data (on a common energy scale with "0" indicating the position of the Fermi level) of a Ga/SiOx/Si sample after 1 month oxidation in ambient conditions.The linear extrapolation to derive valence band maximum (VBM) and conduction band minimum (CBM) positions, respectively, together with the derived values are also indicated.The VBM and CBM valueswere derived using the UPS and IPES spectra from which the metallic Ga contribution had been subtracted (dashed lines) and have an experimental uncertainty of ±0.1 and ±0.2 eV, respectively.

Figure S14 .
Figure S14.Fitting results of Ga 2p3/2 peak of Ga NP on SiOx/Si substrate that had been oxidized at 1×10 -6 mbar O2 for 240 min before and after annealing in UHV (1×10 -9 for 30 min) at temperatures between 400 and 700°C.The metallic Ga peak is fitted by an asymmetric (Doniach-Sunjic) profile and

Figure S15 .
Figure S15.UPS (He II) data of the valence band (VB) of a Ga/SiOx/Si sample oxidized in 1×10 -6 mbar of O2 for 240 min before and after annealing in UHV at different temperatures (400-700°C).All measurements were taken at room temperature (i.e., after sample cool down).

Figure S16 .
Figure S16.SEM image of Ga NP on SiOx/Si substrate that had been oxidized in 1×10 -6 mbar of O2 for 240 min after annealing at 700 o C for 30 min in UHV conditions (1×10 -9 mbar).

Figure S17 .
Figure S17.Size distribution of the Ga particles derived by statistical evaluation of 148 particles exhibited in the SEM image depicted in Fig. S16.