Thermophotovoltaic (TPV) energy conversion — Wikipedia

URL: https://en.wikipedia.org/wiki/Thermophotovoltaic_energy_conversion Fetched: 2026-07-01 Reliability: mixed — Wikipedia may reflect outdated or unverified information; cross-check quantitative claims against primary sources. Why it’s here: TPV converts heat to electricity — the inverse of moving heat out of a space — but it lives on the same physics this KB tracks: black-body/selective radiation, spectral filtering, and the radiative thread. The direct tie-in is thermoradiative / “negative-emission” cells (photodiodes on the hot side of a heat engine that generate power by radiating to a cold sink — a “radiative” cousin of night-sky cooling), plus overlap with the KB’s waste-heat-recovery and heat-battery thermal-storage threads.

Overview

Thermophotovoltaic (TPV) energy conversion is a direct process for transforming heat into electricity via photons. A basic TPV system is a heated object emitting thermal radiation plus a photovoltaic cell — like a solar cell but calibrated for the heat source’s spectrum. Because TPV runs at lower temperatures than solar cells, it typically achieves lower efficiencies, often offset with expensive non-silicon multi-junction cells. Current applications are niche: spacecraft power and industrial waste-heat recovery.

How TPV works

Basic PV principles. A p–n junction near a semiconductor surface: photons above the bandgap free electrons, which the junction field drives to electrodes. Sub-bandgap photons do nothing; above-bandgap photons dump their excess as heat (thermalization). Higher bandgap → higher voltage but fewer usable photons; electrical power peaks at an intermediate bandgap.

TPV-specific challenges. The AM1.5 solar spectrum delivers ~1,000 W/m² at an apparent 5,780 K, optimal capture near 1.4 eV — close to silicon’s 1.1 eV, which is why solar PV is cheap. Lower-temperature sources (industrial burners 900–1,300 °C; welding torch ~3,400 K) emit a broader spectrum (Wien’s displacement law), so TPV nearly always uses multi-junction cells with different bandgaps to reach double-digit efficiency.

System design. Typically a heat source, an emitter (often a metal block), PV cells, and a waste-heat rejection system. PV efficiency drops as the cell heats, so keeping the converter cool is a major engineering challenge. Related: thermoradiative / negative-emission cells put photodiodes on the hot side of a heat engine; hybrid setups have been proposed to extract power from both hot and cold photodiodes.

Efficiency and limits

Theoretical ceiling follows Carnot: η = 1 − (T_cell / T_emit). With T_cell ≈ 300 K and T_emit ≈ 1,800 K, the max is ~83%. Practical losses are large:

As of 2021: lab max ~35% at 1,773 K emitter (heat-input → electrical).
System-level (fuel → electricity): ~5% is a “world record”; real losses come from heat-transfer degradation, low output voltage, and active cooling.

Loss mechanisms: emitter (non-ideal black-body, deep-IR phonon resonances), filter (imperfect reflection back to emitter → waste heat), converter (non-radiative recombination, Ohmic, Fresnel reflection, unabsorbed light), geometric (incomplete converter coverage around the emitter).

Materials

Emitters

Polycrystalline SiC — most common burner emitter, stable to ~1,700 °C, cheap; long-wavelength emission recycled by selective filters / back mirrors.
Tungsten — common refractory selective emitter, emissivity 0.45–0.47 (vis/near-IR), 0.1–0.2 (IR); oxidizes readily.
Rare-earth oxides — Yb₂O₃ peaks 1.29 eV (Si cells), Er₂O₃ 0.827 eV (GaSb/InGaAs); selective emission significant only above ~1,100 °C. ~13% with Yb₂O₃ + Si.
Photonic crystals — engineered photonic bandgaps; Sandia predicted 34% with tungsten photonic crystals, but manufacturing is commercially infeasible.

PV cells

Silicon — cheap/scalable but 1.1 eV bandgap needs >2,000 K emitters; no efficient Si TPV realized.
Germanium — 0.66 eV but poor performance (high carrier mass, hard passivation).
GaSb — invented 1989, basis of most modern TPV; 0.72 eV, high power density; ~20% at a 1,000 °C black-body spectrum (radiative limit 52%); GaAs/GaSb bilayer hit 35% solar efficiency.
InGaAsSb — 0.5–0.6 eV; internal QE approaching 90–95% (MBE), but phase separation and cost limit commercialization.
InGaAs — lattice-matched to InP → 0.74 eV, ~15% efficiency; bandgap-engineered (0.4–1.4 eV) with strain-graded layers; lattice-matches germanium substrates (low defects).
InPAsSb — 0.3–0.55 eV, understudied; longest spectral response 4.3 μm.
PbSnSe/PbSrSe quantum wells — IV-VI on Si substrates, 0.3–0.6 eV, low Auger recombination (>10× smaller than comparable III-V).

History

1956 — Henry Kolm builds an elementary TPV system at MIT.
1960–61 — Pierre Aigrain (credited inventor) MIT lectures catalyze sustained R&D.
1980s — efficiency approaches 30%.
1997 — Western Washington University “Viking 29” TPV hybrid automobile.
2022 — MIT/NREL: 41% device, multiple III-V layers (UV/visible/IR), gold reflector recycling, tungsten emitter to 2,400 °C.
May 2024 — 44% efficiency using SiC as heat-storage emitter at 1,435 °C (2,615 °F); semiconductor captures 20–30% of photons, plus air and gold-reflector layers.

Applications

RTGs — replacing inefficient thermocouples on spacecraft; Emcore/Creare/Oak Ridge/NASA Glenn demoed 15–20%; U. Houston reached 30% (3–4× thermocouples).
Thermal energy storage — off-peak electricity heats carbon / phase-change blocks; TPV cells (with reflectors + insulation) around the mass stay “off” (photons reflected back) until demand, then connect to load.
Waste-heat collection — auxiliary conversion of otherwise-lost heat (steam turbines, solar cells).
Spacecraft power — solar (orbital concentrators) and radioisotope; a converter hit 20% with a 1,350 K tungsten emitter + tandem filters + 0.6 eV InGaAs cooled to room temp. Heat rejection in vacuum (no convective sinks) is the hard problem.
Off-grid generators & CHP — TPV augments solar with alternative fuels; best value from combined heat and power. JX Crystals gas stove/generator: SiC emitter 1,250 °C + GaSb → 25,000 BTU/hr (7.3 kW heat) + 100 W (1.4%). Proposed boiling-coolant CHP: SiC 1,425 °C + GaSb → 85,000 BTU/hr (25 kW heat) + 1.5 kW, est. 12.3% (calc suggests 6%), ~0.08 €/kWh vs 0.12 (gas-engine) / 0.16 (fuel-cell). Never commercialized (market too small).
Portable power — battlefield generators; 2001 JX Crystals Army battery charger: 230 W on propane, SiC 1,250 °C + GaSb, ~0.5 m tall, 2.5% — too low for field use.
Grid-scale storage — spare electricity → heat for long-term storage; graphite medium, molten tin heat transfer at ~2,000 °C; companies claim costs below lithium-ion.
Recreational vehicles — proposed (multi-fuel, silent for “quiet hours”), but required emitter temps make it unlikely.

Black-body radiation

Planck’s law governs emission: I’(λ,T) = (2hc²/λ⁵) · 1/(e^(hc/λkT) − 1). Peak via Wien: λ_max = b/T. Most materials cap at ~1,800 °C → peak λ ≈ 1,600 nm (~0.75 eV); 1,200 °C → ~0.5 eV. Traditional bandgaps (Si 1.1 eV, GaAs 1.4 eV) are impractical for realistic-temperature emitters.

Summary

TPV is direct heat-to-electricity, distinct from solar PV. Theoretical limit ~83%; lab 35–44%; system (fuel→electricity) ~5%. Advances in narrow-bandgap semiconductors and selective emitters are pushing toward aerospace, waste-heat, and energy-storage viability.