Waveguide photoreactor enhances solar fuels photon utilization towards maximal optoelectronic – photocatalytic synergy

The study addresses the challenge of maximizing solar-to-fuel conversion for the reverse water–gas shift (RWGS, CO2 + H2 ⇌ H2O + CO) reaction by rethinking light management on photocatalysts. Conventional strategies concentrate light intensity to increase reaction rates, but this can oversaturate photogenerated charge carriers relative to available surface active sites, causing high recombination, semi-saturation behavior, and poor photon-to-yield efficiencies at high photon flux. The authors propose a principle of light maximization: distributing light below saturation across more catalyst area to balance optical absorption with charge carrier utilization at surface catalytic sites. Using defect-engineered In2O3-x(OH)y nanorod coatings and an annular glass cylindrical rod photoreactor, they aim to overcome semi-saturation limits, leverage omnidirectional scattering and internal reflection to illuminate the entire catalyst coating, and exploit sub-gap trap states to steer carrier recombination through surface catalytic pathways rather than bulk recombination, thereby enhancing RWGS performance.

Catalyst and coatings: Defect-engineered In2O3-x(OH)y nanorods (arrays of necked nanoparticles, 10–20 nm in diameter) were coated as 5 ± 0.3 µm-thick films on (i) planar quartz substrates and (ii) cylindrical quartz rod waveguides (diameter 5 mm, length 6 cm). Rods were encapsulated with grease-free aluminum foil along the length and back end to enhance side and back propagation; the front end remained exposed. Coating method employed a thin-film colloidal solution of nanorods.

Reactor configurations and operating conditions:

• Batch reactor: Vacuum to 2×10⁻² mbar at set temperature and lowest photo-intensity, then introduce H2 to 26–28 psi for 5 min followed by CO2 to 48–52 psi (start of reaction cycle). Maintain set temperature and photo-illumination for 1 h before GC measurement; collect ≥3 hourly data points per condition. Increase photo-intensity in 15–20 mW/cm² steps; reset to vacuum between runs. Rods centered via a cap with a central hole.
• Flow reactor: Continuous flow of 5 sccm H2 and 1 sccm CO2 at 14.1 psi using custom flow controllers. A viewport allowed illumination of the rod front; a thermocouple contacted the coating at the back end. For persistent yield tests, equilibrate under flow and temperature for 3–4 h before light exposure.
• Temperature range: 90–200 °C depending on test.
• Illumination: Broadband lamp with controlled photo-intensity (typical range ~26–160+ mW/cm²); optical filters used to probe spectral response: UV bandpass absorbing λ < 410 nm; yellow filter (λ > 500 nm); red filter (λ > 620 nm).

Electrical and spectroscopic characterization:

• Photocurrent measurements on coated planar substrates under varying photo-intensities to evaluate external quantum efficiency (EQE) and saturation trends.
• Gas sensitivity electrical conductivity: Modified cryostat with 4-probe on square Au electrodes (0.5 mm spacing). After pump-down to ~1×10⁻² mbar, flows of Ar (up to 1000 sccm), H2 (1 sccm), and CO2 (1 sccm) were introduced; Ar flow was reduced to vary (H2+CO2)/Ar ratio. Measurements at various temperatures and photo-intensities; separate 0.5 mg powder aliquots per condition.
• XPS (O 1s) under glovebox protection after exposures to H2:CO2 (1:1) atmospheres at set temperatures and photo-intensities. Quantified OH shoulder area ratio as area between 531–534.5 eV divided by total O 1s area 528–534.5 eV; monitored binding energy shifts.
• UPS and transient absorption pump–probe spectroscopy (supporting evidence for sub-gap states and long-lived carriers at low fluence).

Optical waveguiding and scattering:

• Laser coupling at ~60° incidence using blue (405 nm), green (535 nm), and red (732 nm) beams to visualize propagation through coated rods.
• Integrating sphere measurements of diffused transmittance to quantify side leakage vs end transmission for coated/uncoated rods, with and without an aluminum cap on the rod end.
• Electromagnetic simulations (Mie-type) of scattering by arrays of 60 nm-diameter cylindrical rods at incident wavelengths of 400 and 600 nm and various incident angles to assess near- and far-field differential scattering cross sections and angular redistribution.
• Lambertian limit estimation of optical pathlength enhancement and resulting absorbance A(λ) = 1 − e^(−4π a(λ) d), using a(600 nm) = 277 cm⁻¹ and d = 5 µm.

Silicon tandem layers:

• PECVD amorphous silicon (a-Si:H) on quartz rods: 200 nm p+ doped a-Si:H and optional 40 nm n-type a-Si:H deposited in two steps to cover the cylindrical surface (rotation between depositions). Conditions: p-type: 160 °C, SiH4 8 sccm, 0.175 Torr, 10 W RF, B2H6 (1.98% in H2) 20 sccm, 20 min; n-type: 180 °C, SiH4 8 sccm, 0.175 Torr, 2 W RF, PH3 (1.88% in H2) 20 sccm, 25 min. In2O3-x(OH)y nanorods were then coated atop the silicon.
• Electrical characterization of a-Si:H on glass via transmission line method with Al (p-type) and Ag (n-type) contacts to estimate carrier concentrations (p: 8.5×10¹⁸ cm⁻³, n: 3.9×10¹⁸ cm⁻³) and mobilities (p: 5.3×10⁻⁵ cm² V⁻¹ s⁻¹, n: 1.5×10⁻³ cm² V⁻¹ s⁻¹) from prior PECVD doping efficiency correlations.
• Band diagram simulations and EQE simulations for n–p a-Si:H stacks coupled to In2O3-x(OH)y.

Persistent photoconductivity (PPC) and EPR:

• EPR under 405 nm illumination and during post-illumination decay, tracking spin centers associated with singly ionized oxygen vacancies; monitored spin concentration vs time.
• Persistent CO production monitored in batch and flow reactors after turning off illumination; fitted decay constants and compared to thermal-only baselines.

Outdoor solar concentrator test:

• Coated, Al-encapsulated waveguide integrated into a rooftop solar batch concentrator without tracking; monitored CO yield during daylight (~9 h) and overnight post-illumination.

• Light maximization via waveguiding outperforms planar illumination: An In2O3-x(OH)y-coated quartz rod waveguide surpasses a coated planar substrate by a factor of 2.2 in CO production under distributed sub-saturation illumination.
• Batch reactor, 150 °C: Planar CO rate increases to ~20 µmol g⁻¹ cat h⁻¹ below ~65 mW cm⁻² and semi-saturates to ~27 µmol g⁻¹ cat h⁻¹ above ~81 mW cm⁻²; quantum efficiency drops from ~0.6% to ~0.2% at higher intensities. The coated waveguide exhibits a near-linear rate increase with relatively constant quantum efficiency (~0.6–0.7%). At 200 °C, the rod maintains higher photon-to-yield efficiency than the planar by ~+0.25% absolute across intensities.
• Flow reactor: Dark activation energy is 48.1 kJ mol⁻¹, decreasing to 23.8 kJ mol⁻¹ under 139 mW cm⁻² illumination. At 200 °C, CO rate rises from 118.9 to 246.2 µmol g⁻¹ cat h⁻¹ with increasing photo-intensity; photon-to-yield efficiency remains ~0.7% at high intensities (150 °C).
• Photocurrent and gas sensitivity: Planar photocurrent rises sharply below ~60–100 mW cm⁻² and semi-saturates at higher intensities; EQE ~5.9 ± 0.3% at low intensities, decreasing to 4.7–2.4% at high intensities. Electrical conductivity sensitivity to (H2+CO2)/Ar increases with photo-intensity, saturating near ~104 mW cm⁻²; H2 is more strongly adsorbed than CO2, especially at 150 °C with illumination.
• XPS (O 1s) under H2+CO2: Photo-thermal conditions increase the OH shoulder area ratio by 190–230% over thermal/dark at 75–200 °C, with In binding energy shift of ~+0.15 ± 0.05 eV and OH peak shift ~+0.3 eV at higher intensities; OH shoulder expansion reflects H+OH sub-peak growth (H2 dissociation on OH). For rods, maximum OH growth under illumination is 294% above dark. Planar OH shoulder growth slows above ~66 mW cm⁻²; rod OH growth remains more linear beyond ~95 mW cm⁻², consistent with lower effective local intensity on the rod.
• Optical waveguiding and scattering: Coated rods guide green (535 nm) and red (732 nm) light with significant internal reflection; blue shows attenuation along length. Integrating sphere: Coated rod side leakage ~30–37% for λ > 500 nm; end transmittance ~40–45%; with end cap, side leakage drops to 18–27%. Uncoated rods show ~90% total transmission (side + end). Nanorod arrays (60 nm diameter) exhibit strong, wide-angle scattering at 400–600 nm, relatively angle-insensitive at 600 nm; enhanced pathlength can raise 600 nm absorbance of a 5 µm film from 0.13 to ~0.76 (Lambertian limit, a(600 nm)=277 cm⁻¹).
• Visible-tail harvesting: With UV filter (blocking λ < 410 nm), planar yields 7.75 ± 0.32 µmol g⁻¹ h⁻¹ vs rod 35.6 ± 0.21 µmol g⁻¹ h⁻¹ (4.6×). With yellow (λ > 500 nm) and red (λ > 620 nm) filters, planar yields 0.6–1.02 ± 0.2 µmol g⁻¹ h⁻¹, while rod yields 9.68 ± 0.52 µmol g⁻¹ h⁻¹ (8.7×) and 4.87 ± 0.34 µmol g⁻¹ h⁻¹ (8.1×), respectively, evidencing sub-bandgap tail utilization.
• Tandem with amorphous silicon: p+ a-Si (200 nm)/In2O3-x(OH)y on rods increases visible-range CO rates: red 4.31 ± 0.35 µmol g⁻¹ h⁻¹ and yellow 12.54 ± 0.33 µmol g⁻¹ h⁻¹; photo-action shows 19% higher CO in 440–520 nm vs nanorod-only rods. Adding a 40 nm n-type layer (p+/n a-Si) further boosts production across 400–560 nm; overall UV–visible increase of 27% vs nanorod-only rods.
• Persistent photocatalysis (PPC): Photocurrent saturates within ~2 h and decays over ≥7 h; EPR indicates photoionization of oxygen vacancies with slow post-illumination decay. Batch at 200 °C, 107 mW cm⁻²: CO increases from 0.7 to 114.6 ppm in 3 h (~38 ppm h⁻¹). After lights off: ~23.4 ppm h⁻¹ for 2 h, then ~11.9 ppm h⁻¹; dark thermal baseline at 200 °C is ~14.1 ppm h⁻¹. Persistent period yields are ~1.62× thermal. Flow reactor PPC decay periods: ~50–55 min at 90 °C and ~35–45 min at 120 °C; average yields during decay exceed thermal by factors of ~1.44 and ~1.57, respectively; decay is sub-exponential (fit constants ~5.10 and 1.73 min), consistent with a continuum of trap states.
• Outdoor solar concentrator: In a day–night cycle without tracking, 9 h daylight produced 20.45 ± 0.43 ppm CO (rate 8.89 ± 0.25 µmol g⁻¹ h⁻¹); overnight, yield increased by an additional 24.05 ± 0.22 ppm (~19% over the daylight-only period), demonstrating practical PPC utility.
• Mechanistic insight: Low-intensity illumination favors long-lived carriers and surface SFLP (In–OH/Ovac) activation, enabling H2 heterolytic splitting (H+OH and In–H) and CO2 activation via intermediate carbonate; illumination lowers apparent activation energies and shifts recombination toward surface catalytic pathways.

Distributing light below saturation across an In2O3-x(OH)y-coated waveguide enhances photon-to-yield efficiency and overall RWGS rates compared to conventional light concentration on planar catalysts. The waveguide geometry spreads the incident flux, preventing overpopulation of charge carriers relative to surface active sites and mitigating recombination-driven semi-saturation. Omnidirectional scattering by nanorods and internal reflection within the quartz rod extend optical pathlengths, particularly for weakly absorbed green/red wavelengths, effectively populating sub-gap trap states. This promotes recombination through catalytically productive surface states (SFLP sites) rather than bulk, which, together with increased OH formation and H2 heterolytic dissociation evidenced by XPS, accelerates CO formation. The approach sustains high quantum efficiencies over a broad intensity range and lowers apparent activation energies under illumination in both batch and flow reactors. Coupling with thin doped amorphous silicon further improves visible-light harvesting via built-in fields that drive electrons into the In2O3-x(OH)y, broadening spectral response and increasing CO production by up to 27%. Persistent photoconductivity from oxygen-vacancy-related traps maintains reactive carrier populations after illumination ceases, enabling significant post-sunset CO production and higher cumulative yields, as validated by lab reactors and outdoor concentrator tests. Collectively, the optoelectronic–photocatalytic synergy addresses the core hypothesis that light maximization, not concentration, better aligns photophysics with surface catalysis to improve solar fuel generation.

The work introduces a light maximization strategy using an In2O3-x(OH)y-nanorod-coated quartz rod waveguide to overcome semi-saturation at high intensities and enhance RWGS photocatalysis. Key contributions include: (i) distributed sub-saturation illumination that preserves photon-to-yield efficiency and elevates CO rates versus planar geometries; (ii) nanorod-driven scattering and waveguiding that extend optical pathlengths to harness sub-gap tail states and drive surface catalytic recombination; (iii) tandem integration with doped amorphous silicon, boosting visible-range activity and overall production by up to 27%; and (iv) exploitation of persistent photoconductivity to sustain CO generation after illumination. Future directions proposed include multi-layer/tandem photocatalyst coatings along waveguides to harvest different spectral bands, compact high surface-to-volume waveguide reactor assemblies (e.g., shell-and-tube) for enhanced light reutilization among adjacent guides, and integration with solar concentrators to combine spectral utilization with high quantum efficiencies for scalable solar-fuels production.