Ambient energy harvesting is the extraction of microwatt-to-milliwatt electrical power from light, heat, vibration and radio waves already present in an environment, and its practical role is not grid generation but the elimination of batteries from very large numbers of sensors.
Summary
Ambient energy harvesting covers a family of transducers that convert whatever energy flux happens to be crossing a device into usable electricity: indoor light via organic or perovskite photovoltaics, temperature differences via thermoelectric or pyroelectric materials, mechanical vibration via piezoelectric elements, and incident radio waves via rectifying antennas. The physics differs entirely between branches, but the engineering problem is common: the available flux is small and intermittent, so the harvester, the storage element and the load must be co-designed around a power budget measured in microwatts.
Further out sit mechanisms that are still physics rather than engineering. Thermoradiative diodes, which generate power by emitting infrared into a cold sky, are at the point where 1x1 mm2 MBE-grown InAs devices show breakdown above 0.3 V but reverse saturation current densities 200 times the radiative limit. Magnet-free nonreciprocal thermal emitters based on magnetic Weyl semimetals such as Co3Sn2S2 are predicted from first principles to outperform InAs, a near-field photon Nernst effect has been proposed with efficiency bounded by Carnot, and an in-plane anomalous Nernst effect has been observed in ultrathin ferromagnetic oxide films, removing the orthogonality constraint on device geometry.
The decisive parameters are not the transducer figures of merit alone. They are power density per unit area or volume under realistic ambient conditions, the quiescent draw of the electronics being powered, and cost per cm2 at volume. A hardware-orchestrated power management architecture that cuts quiescent drain to 452 nA by fully power-gating the microcontroller illustrates the point: in low-light conditions the sleep current, not the harvester, sets autonomy.
Evidence base
Open questions
- Does the Fe2VAl module deliver competitive conversion efficiency and cost per watt in a real waste-heat application, or does the average zT of about 0.3 confine it to demonstration status?
- Can InAs or Weyl-semimetal thermoradiative devices close the two-orders-of-magnitude gap between measured reverse saturation current and the radiative limit, and if so at what device area and cost?
- Does Dracula’s stated 600M cm2/yr capacity translate into shipped volume and a cost per cm2 that undercuts a coin cell over the sensor’s service life 2025 2026 European Next Gen Pv Cohort?
- Will unified SWIPT receivers that rectify power and demodulate information in the same analogue hardware reach a power budget compatible with a 452 nA quiescent node, or does the harvesting-plus-decoding overhead cancel the gain?
Assessment drafted 2026-08-31 from up to 18 KB sources using the technology-scorecard framework; scores are a draft read pending review.