Aqueous Phase Reforming over Platinum Catalysts on Doped Carbon Supports: Exploring Platinum–Heteroatom Interactions

A series of platinum catalysts supported on carbon nanofibers with various heteroatom dopings were synthesized to investigate the effect of the local platinum environment on the catalytic activity and selectivity in aqueous phase reforming (APR) of ethylene glycol (EG). Typical carbon dopants such as oxygen, nitrogen, sulfur, phosphorus, and boron were chosen based on their ability to bring acidic or basic functional groups to the carbon surface. In situ X-ray absorption spectroscopy (XAS) was used to identify the platinum oxidation state and platinum species formed during APR of EG through multivariate curve resolution alternating least-squares analysis, observing differences in activity, selectivity, and platinum local environment among the catalysts. The platinum-based catalyst on the nitrogen-doped carbon support demonstrated the most favorable properties for H2 production due to high Pt dispersion and basicity (H2 site time yield 22.7 h–1). Direct Pt–N–O coordination was identified by XAS in this catalyst. The sulfur-doped catalyst presented Pt–S contributions with the lowest EG conversion rate and minimal production of the gas phase components. Boron and phosphorus-doped catalysts showed moderate activity, which was affected by low platinum dispersion on the carbon support. The phosphorus-doped catalyst showed preferential selectivity to alcohols in the liquid phase, associated with the presence of acid sites and Pt–P contributions observed under APR conditions.


XPS
The N 1s deconvolution was constrained with a set value of the full width at half maximum (FWHM) to 1.5 ± 0.1 eV, and the shape factors were maintained at approximately 0.5, equal for all functions during the fitting process.A comprehensive list of the species present in the spectra and their corresponding binding energies can be found in Table S1.
The S 2p contribution were deconvoluted considering 2p1/2 and 23/2 S species with doubles of two pseudo-Voigt profiles with a fixed intensity ratio of 0.55±0.5 and a set relative binding energy of 1.18 eV.The FWHM was limited within the range of 1.2±0.2eV.The shape factors were consistently kept at around 0.5 during the fitting process across all tasks.
The P 2p contribution included 2p1/2 and 3P3/2, represented by two pseudo-Voigt profiles with a fixed intensity ratio of 0.65 ± 0.5 and set relative binding energy of 0.84 eV.The FWHM was constrained to 1.5f ± 0.1 eV.The shape factors were maintained equal for all functions during the fit and were typically close to a value of 0.5.
The B 1s spectrum consisted of four components.The FWHM was constrained to 0.75 ± 0.5 eV.The shape factors were kept equal for all functions during the fit and were typically close to a value of 0.5.
For the analysis of Pt species, the distance between the spin-orbit splitting is set to 3.3 eV to represent 4f7/2 and 4f5/2, and the intensity ratio was constrained to 0.8 ± 0.05.FWHM and the shape factor of Pt 4f were set to be equal for all the functions during the fitting.a Total carbon present in the gas phase b Stoichiometric reforming ratio for EG  =  2  2 ⁄ = 5 2 ⁄ c Total carbon in each gaseous product (i) d Moles of H2 produced divided by the duration of the experiment (120 min) e Molecular weight of Pt f Pt dispersion estimated by CO chemisorption (%) g Amount of catalyst (g) h Weight fraction of Pt in the catalyst (wt%) j Total carbon in each liquid product (i) k Total carbon contained in the feed stream h Besides alkanes (C1 -C6), ethene was also occasionally detected and included.

Figure S 1 .
Figure S 1. Schematic diagram of the setup used for gasification-assisted heteroatom doping (GAHD) of carbon nanofibers (CNF).

Figure S 3 .
Figure S 3. Scheme of the high pressure cell used for the APR experiments at the CAT-ACT beamline at the KIT Light Source, based on reaction cell in[9,10]

Figure S 5 . 10 *
Figure S 5. [a] N2 physisorption isotherms of CNF subjected to heat treatment, N, S, P and B doping.[b]BET specific surface area and total pore volume for all CNF supports estimated by single-point desorption at p/p0 = 0.9

Figure S 6 .
Figure S 6. Raman spectra showing the D and G bands of the CNF subjected to heat treatment, N, S, P and B doping.

Figure S 7 .Figure S 8 .
Figure S 7. [a] Raman intensity ratio between the D and G bands [c] Raman shift of the G band of CNF samples subjected to heat treatment, N, S, P and B doping.

Figure S 9 .
Figure S 9. EDS mapping for [a] reduced and [b] spent Pt/CNF-HT catalyst containing HAADF-STEM images and Pt, C, and O signals.

Figure S 10 .
Figure S 10. EDS mapping for [a] reduced and [b] spent Pt/N-CNF catalyst containing HAADF-STEM images and Pt, N, and O signals.

Figure S 11 .
Figure S 11.EDS mapping for [a] reduced and [b] spent Pt/S-CNF catalyst containing HAADF-STEM images and Pt, S, and O signals.

Figure S 12 .
Figure S 12. EDS mapping for [a] reduced and [b] spent Pt/P-CNF catalyst containing HAADF-STEM images and Pt, P, and O signals.

Figure S 13 .
Figure S 13.EDS mapping for [a] reduced and [b] spent Pt/B-CNF catalyst containing HAADF-STEM images and Pt, B, and O signals.

Figure S 14 .
Figure S 14. High-resolution XPS [a, b, c] N 1s of N-CNF, Pt/ N-CNF reduced and Pt/ N-CNF spent [d, f, g] S 2p of S-CNF, Pt/ S-CNF reduced and Pt/ S-CNF spent [h, i, j] P 2p of P-CNF, Pt/ P-CNF reduced and Pt/ P-CNF spent [k, l, m] B 1s of B-CNF, Pt/ B-CNF reduced and Pt/ B-CNF spent.

Figure S 20 .
Figure S 20.Pt L3-edge EXAFS fitting curves of Pt/CNF-HT after catalyst reduction [a] in R space [b] in k space.After APR of ethylene glycol at 225°C and 30 bars (WHSV = 9h -1 ) [c] in R space [d] in k space.

Figure S 21 .
Figure S 21.Pt L3-edge EXAFS fitting curves of Pt/N-CNF data after catalyst reduction [a] in R space [b] in k space.After APR of ethylene glycol at 225°C and 30 bars (WHSV = 9h -1 ) [c] in R space [d] in k space.

Figure S 22 .
Figure S 22. Pt L3-edge EXAFS fitting curves of Pt/S-CNF data after catalyst reduction [a] in R space [b] in k space.After APR of ethylene glycol at 225°C and 30 bars (WHSV = 9h -1 ) [c] in R space [d] in k space.

Figure S 23 .
Figure S 23.Pt L3-edge EXAFS fitting curves of Pt/P-CNF data after catalyst reduction [a] in R space [b] in k space.After APR of ethylene glycol at 225°C and 30 bars (WHSV = 9h -1 ) [c] in R space [d] in k space.

Figure S 24 .
Figure S 24.Pt L3-edge EXAFS fitting curves of Pt/B-CNF data after catalyst reduction [a] in R space [b] in k space.After APR of ethylene glycol at 225°C and 30 bars (WHSV = 9h -1 ) [c] in R space [d] in k space.

Table S 2
. Parameters and equations applied to evaluate APR catalyst performance.
[6]leS1.Band assignments and fitting parameters for the XPS analysis a Triphenylphosphin oxide as model compound[6]In situ XAS-XRD setup for APR

Table S 3
. APR of ethylene glycol (6 %wt) at 250 °C and 26 bar initial pressure for 2h at batch conditions.The carbon balance was confirmed to a degree of 88-100 % for all measurements.Based on the CO uptake before APR of ethylene glycol (6 %wt) at 250 °C and 26 bar initial pressure for 2h b Based on the CO uptake after APR of ethylene glycol (6 %wt) at 250 °C and 26 bar initial pressure for 2h

Table S 4
. Surface concentration of individual heteroatom species determined by deconvolution of high-resolution XPS spectra.
* Surface concentrations without considering Pt.

Table S 5
. Fractions of Pt species of reduced and spent catalyst as determined by deconvolution of the Pt 4f spectra in XPS.
*Spent -After 2hours of aqueous phase reforming of EG at 250°C and 26 bar of initial pressure.