Skip to content

Photoactive and Ultrastable Silicon/Boron Doped Diamond Photoanodes Enable Longstanding Photoinduced Electrochemiluminescence

FieldValue
JournalAdvanced Optical Materials
Published2026-10-01
AuthorsChanglin Zhou, Clément Briquet‐Laugier, Bertrand Goudeau, Sirine Ben Khemis, Zonghua Wang
DOI10.1002/adom.71943

An n-type silicon photoanode coated with boron-doped diamond sustained photoinduced electrochemiluminescence for more than 100 h under 850 nm illumination in a neutral electrolyte, with periodic solution renewal. The 240 nm junction produced a 0.4 V photovoltage and supported visible upconversion at 630 nm with Ru(bpy)₃²⁺/TPrA and 450 nm with luminol/H₂O₂. The device regenerated its signal after alkaline fouling and imaged individual 12 µm Ru-labeled magnetic beads.

The device was a front-illuminated n-Si/BDD heterojunction. BDD protected photoactive silicon from aqueous oxidation and provided a conductive path for interfacial hole transfer. The authors used dark, nonphotoactive p++-Si/BDD as a control. Without illumination, n-Si/BDD produced negligible anodic current and no emission in either ECL electrolyte.

Photogenerated holes in silicon supplied a photovoltage of up to 0.4 V, reducing the external potential needed to drive oxidation. Impedance measurements from 100 Hz to 2 kHz gave a flat-band potential near -0.71 V for n-Si. Mott-Schottky analysis yielded a dopant density of 1.5 × 10¹⁵ cm⁻³ and a resistivity of 3.1 Ω cm. The measured resistivity fell within the 1 to 10 Ω cm range of the wafer.

The n-type silicon was phosphorus-doped, single-side polished, and 500 ± 25 µm thick, with a resistivity of 1 to 10 Ω cm. Cleaning preceded oxidation in 7:3 concentrated H₂SO₄/30% H₂O₂ at 105 °C for 30 min. This step produced a 1.5 to 2 nm SiOx layer. The substrate was then seeded with diamond nanoparticles before boron-doped diamond growth.

Microwave-plasma-assisted chemical vapor deposition used 1000 W, 40 mbar, and a substrate temperature of 750 °C. The gas mixture contained 2% methane in hydrogen and 1000 ppm trimethylboron in hydrogen. A 50 min growth produced a 240 nm film, as measured by optical interferometry and SEM.

Cross-sectional SEM showed a void-free Si/BDD interface. Top-view SEM showed dense polycrystalline coverage without visible cracks or pinholes, with an average grain size of about 100 nm. Raman analysis found a 1315 cm⁻¹ diamond-related peak, a broad band near 1225 cm⁻¹, and a weak G band near 1500 cm⁻¹. The weak G band was interpreted as low graphite content.

Two ECL chemistries tested whether the photovoltage could drive different oxidative reactions. All potentials were referenced to Ag/AgCl with 3 M NaCl.

ECL systemElectrolytePECL peakPotential result
Ru(bpy)₃²⁺/TPrA1 mM Ru(bpy)₃Cl₂ and 100 mM TPrA in 0.1 M PBS at pH 7.4630 nmIlluminated n-Si/BDD maximum at 1.1 V; 0.4 V cathodic onset shift
Luminol/H₂O₂10 mM luminol and 33 mM H₂O₂ in 0.1 M KOH at pH 13450 nmIlluminated onset at -0.25 V versus 0.2 V for dark p++-Si/BDD; 0.45 V shift

The Ru system showed a dark oxidation peak current at 1.29 V on p++-Si/BDD. Under illumination, n-Si/BDD reached maximum emission at 1.1 V. The onset shifted cathodically by 0.4 V, matching the measured photovoltage.

For luminol/H₂O₂, the illuminated onset moved from 0.2 V to -0.25 V, a 0.45 V shift. The PECL onset shifted by 0.4 V. IPCE measured at 1.2 V reached 26% at 850 nm and showed response from 550 to 1000 nm.

An identical BDD film on glass transmitted about 33% at 850 nm, which still admitted sufficient light to drive the silicon. The 850 nm excitation and visible emission peaks demonstrate the reported upconversion process.

Long-term tests used galvanostatic operation at 0.6 mA under 850 nm illumination. The Ru/TPrA cell ran for more than 100 h with 5 mM Ru(bpy)₃Cl₂ and 100 mM TPrA in 0.1 M PBS at pH 7.4. The cell held 7.6 mL, and a peristaltic pump maintained a 0.7 mL/s flow. The solution was replaced completely eight times.

After each renewal, PECL reached its maximum for 2 to 3 h and then declined to about 50% of that peak. Recovered intensity after fresh solution showed that solution chemistry, rather than electrode damage, caused most decay. SEM after 100 h found no detectable topographic modification, contamination, or byproduct deposition.

The luminol test covered 30 h and three cycles. Its signal fell to zero after 8 h in each cycle, and the electrolyte darkened after more than 15 h. The authors attributed this behavior to H₂O₂ depletion and dark oxidative byproducts from luminol. Replacing the electrolyte and rinsing the photoanode with PBS restored the initial signal for the next cycle.

Conventional silicon photoanodes in this field used Ni, Ir, or Pt coatings, which the authors describe as vulnerable to anodic passivation, poisoning, and structural degradation. Supporting Table S1 reports that some metal coatings had higher initial photovoltages but short lifetimes and no regeneration. The BDD device combined extended operation with full signal recovery under alkaline fouling.

IR laser ablation removed BDD in a “PECL” letter pattern using a 1064 nm laser at 1 W and a 100 mm/s scan speed. Ablated regions emitted no PECL because exposed silicon passivated rapidly in water. The authors treated this result as direct evidence that BDD mediates hole transfer.

Ru-labeled 12 µm magnetic polystyrene beads were immobilized on the smooth BDD surface with a magnet. Labels were attached through amide bonds. Each bead was imaged in Procell M with TPrA under 850 nm, 30 mW LED illumination and applied bias.

Neither applied bias alone nor LED illumination alone produced detectable luminescence. These controls linked the bead images to simultaneous photoexcitation and charge transfer. The beads were an immunoassay model only. The study performed no antigen-antibody recognition or quantitative biomarker detection.

BDD served as a protective layer, charge-transfer interface, and partially transparent window in one photoanode. The measured 0.4 to 0.45 V onset reduction shows that a silicon junction can operate the tested ECL reactions at lower external potentials.

Designers should treat electrolyte renewal and surface cleaning as operating requirements, not exceptional recovery events. Both tests restored full signal after renewal, but neither maintained constant output throughout its test period. Laser patterning must retain BDD over active areas because bare silicon stopped emission during the test.

For bead readers, illumination and bias should be independently controlled because each gate alone produced no signal. These results support BDD when chemical stability and regeneration matter more than maximizing initial photovoltage.

The optical penalty remains material. The glass-supported BDD transmitted about 33% at 850 nm. The authors propose halving film thickness and lowering boron concentration tenfold. They predict 60% to 80% transmittance under those conditions, but did not test the proposed combination.

The long Ru/TPrA operation also depended on eight full replacements and periodic loss of roughly half the signal. The alkaline system fouled, required PBS rinsing, and returned to zero emission before renewal. External illumination adds optical background relative to conventional ECL, as the paper notes.

Reproducibility requires attention to two reported variations. The measured film was 240 nm, although the discussion refers to around 200 nm. SIMS gave a doping order of 5 × 10²¹ atoms cm⁻³, while the discussion describes about 2 × 10²¹ atoms cm⁻³. Finally, bead imaging established spatial readout, not diagnostic performance.

The paper’s 850 nm PECL photoanode requires a conductive, transparent, chemically stable boron-doped diamond film integrated with n-type silicon. For 6CCVD, the diamond materials catalog lists MPCVD-grown single-crystal and polycrystalline diamond along with boron-doped diamond. The engineering support can define the BDD-on-silicon coating and electrode contact requirements.

View Original Abstract

ABSTRACT The lifetime of narrow bandgap photoanodes, employed in harsh photoelectrochemical energy conversion processes, relies on the protection layers that shield the semiconductor from severe oxidative degradation and passivation caused by the electrolyte. Here, we report ultrastable n ‐Si‐based photoanodes, coated with boron‐doped diamond (BDD) thin films, as an alternative to the transition metal‐coated Si photoelectrodes employed so far for upconversion photoinduced electrochemiluminescence (ECL). Under near‐infrared (λ exc = 850 nm) excitation, the ( n ‐Si/BDD) heterojunction generates a stable 0.4 V photovoltage, driving efficient optical upconversion to visible photoinduced ECL (PECL) emission with both model ECL systems: luminol/H 2 O 2 (pH 13, λ PECL = 450 nm) and [Ru(bpy) 3 ] 2+ /TPrA (pH 7.4, λ PECL = 630 nm), both at remarkably low potentials. Benefiting from the structural integrity and chemical inertness of the BDD surface, the photoanode sustains continuous operation for more than 100 h in the neutral electrolyte with periodic renewal and can be regenerated after fouling in the alkaline environment. The n ‐Si/BDD photoanodes were demonstrated for localized PECL imaging of [Ru(bpy) 3 ] 2+ ‐functionalized magnetic microbeads as an immunoassay‐relevant model system. This work establishes a chemically robust platform for reliable PECL imaging of luminophore‐functionalized beads, highlighting its potential for future bead‐based immunoassay applications and advanced clinical diagnostic technologies.