Photodiode
A photodiode is a semiconductor device that converts light into electric current. It is built around a semiconductor p-n junction, and when photons of sufficient energy are absorbed near that junction they generate electron-hole pairs that the device collects as a measurable current.[1][2] Photodiodes are also called photodetectors or light sensors, and they are among the most common ways an electronic system senses light.
In virtual reality (VR) and augmented reality (AR) hardware, photodiodes are the light-sensing element behind several subsystems rather than a single feature. They decode the infrared laser sweeps in Lighthouse base-station tracking, register reflected infrared light in some eye tracking systems, detect a user's presence through proximity sensors that turn the display on and off, and read the pulsing light in optical heart-rate sensors on companion wearables. Because a bare photodiode and an infrared LED cost far less and draw far less power than an image sensor and its processing pipeline, photodiode-based designs recur wherever a headset or pair of smart glasses needs to sense light cheaply and at low power.[3][4]
How a photodiode works
A photodiode is a p-n junction or p-i-n structure operated so that incoming light controls the current through it. Light absorbed in the depletion region (the carrier-depleted zone at the junction) creates electron-hole pairs; the junction's electric field sweeps the electrons toward the n-side and the holes toward the p-side, producing a photocurrent that is proportional to the optical power falling on the device.[1][2] A p-i-n photodiode inserts a wide, lightly doped or intrinsic layer between the p and n regions so that most absorption happens in that thick depletion zone, which improves efficiency and speed.[1][5]
A photodiode can run in two modes. In photovoltaic mode it is zero-biased, so it builds up a voltage like a tiny solar cell; this keeps dark current low but the response is slower because junction capacitance is higher and the voltage response is nonlinear. In photoconductive mode a reverse bias is applied, which widens the depletion region, lowers the capacitance, and gives a fast, very linear current response over many orders of magnitude of optical power, at the cost of slightly higher dark current. Photoconductive operation is the usual choice when fast optical pulses must be detected.[1][6]
Several figures of merit describe a photodiode. Responsivity is the photocurrent produced per unit of optical power and is tied to quantum efficiency, the fraction of incident photons that contribute a carrier. Dark current is the small current that flows with no light present and sets the noise floor, so it must be kept low to detect faint signals. Speed is limited by junction capacitance and the load resistance: a high capacitance forces a trade-off between bandwidth and sensitivity, which is why fast detectors use small, low-capacitance junctions.[1][6]
Types
| Type | Structure and operation | Typical use |
|---|---|---|
| PN photodiode | Simple p-n junction; modest speed and sensitivity | General light sensing |
| PIN photodiode | Intrinsic layer widens the depletion region for higher efficiency and faster, more linear response | High-speed detection, optical communications, optical tracking sensors |
| Avalanche photodiode (APD) | Operated near reverse breakdown so each photo-generated carrier is multiplied by impact ionization, giving internal gain | Detecting very weak light; LiDAR receivers |
| Single-photon avalanche diode (SPAD) | APD biased above breakdown (Geiger mode); a single photon triggers a detectable avalanche | Photon counting, time-of-flight depth sensing |
PIN photodiodes are the workhorse for fast, linear detection. Avalanche photodiodes add internal gain, with reported responsivity roughly 5 to 100 times that of a PIN device, which helps when the returning signal is faint, although the multiplication process adds noise.[7] Pushing the reverse bias above the breakdown voltage turns an APD into a single-photon avalanche diode, where one absorbed photon can set off a strong avalanche current, which is the basis of photon-counting and many time-of-flight depth sensors.[7]
The semiconductor material sets the wavelength range a photodiode can detect.
| Material | Approximate detection range | Notes |
|---|---|---|
| Silicon (Si) | about 400 to 1000 nm, best around 800 to 900 nm | Covers visible light and near-infrared; the common low-cost choice |
| Gallium arsenide (GaAs) | about 400 to 870 nm | Visible to near-infrared |
| Germanium (Ge) | about 900 to 1600 nm | Longer near-infrared, higher dark current |
| Indium gallium arsenide (InGaAs) | about 900 to 1700 nm, best around 1300 to 1600 nm | Telecom wavelengths and short-wave infrared |
Silicon's sensitivity through the near-infrared, including the 850 nm band, is why silicon photodiodes pair naturally with the 850 nm infrared LEDs and lasers used throughout VR and AR tracking and sensing.[1][8]
History
The physics behind the photodiode predates the semiconductor era. In 1839 Alexandre Edmond Becquerel observed the photovoltaic effect, a voltage produced when light struck an electrode in an electrolyte.[9] In 1873 Willoughby Smith reported photoconductivity in selenium, finding that its resistance dropped under light.[9] Practical light detection in solid-state form followed the development of the p-n junction diode through the 1940s, and photodiode technology was refined during the 1950s. The PIN photodiode, whose wide intrinsic absorption layer underlies most modern fast detectors, is credited to Jun-Ichi Nishizawa and colleagues around 1950, with the absorption behavior of the wide depletion layer analyzed later that decade.[5][10]
Role in VR and AR
Lighthouse base-station tracking
The clearest VR use of photodiodes is the Lighthouse tracking system that Valve designed for SteamVR and that shipped with the HTC Vive. Each tracked object, a head-mounted display, a controller, or a standalone tracker, carries an array of photodiodes spread across its surface, typically between 20 and 32 of them.[3][11] Stationary base stations flood the room with a synchronized infrared flash and then sweep infrared laser lines across it on the horizontal and vertical axes. Each photodiode records the precise time at which the flash arrives and the time at which each laser sweep passes over it; because the geometric layout of the photodiodes on the object is known in advance, an onboard chip turns those timings into the object's position and orientation.[11][12]
The sensing element behind each Lighthouse photodiode is a small "light-to-digital" integrated circuit. Triad Semiconductor's TS3633 takes the raw signal from a photodiode and converts it into data the SteamVR tracking algorithms can process; a later part, the TS4231, decodes the on-beam data of the second-generation base stations. The first-generation TS3633 sold for roughly $0.92 each in lots of 50 and about $0.49 at 1,000 units, so a 32-sensor headset used on the order of $15 of these chips.[3] The SteamVR Base Station 2.0 covers a field of view of about 150 degrees horizontally by 110 degrees vertically with an effective range near 7 m.[12] This is a form of outside-in tracking: the moving headset and controllers sense light from fixed emitters, which keeps high-precision tracking off the headset's own processor.
Eye tracking
Some eye tracking systems use photodiodes instead of, or alongside, cameras. An infrared LED illuminates the eye, and a photodiode measures the intensity of the light reflected back from the cornea and sclera; as the eye moves, the reflected intensity changes, and that changing photocurrent encodes eye position.[13] The FOVE headset, an early consumer VR product with built-in eye tracking, used Osram SFH 4053 infrared ChipLEDs emitting at 850 nm with a wide beam angle of about plus or minus 70 degrees to light the eyes evenly, which let the system feed gaze data into foveated rendering.[8]
A more recent photodiode-based approach comes from AdHawk Microsystems, which replaced the camera with a microelectromechanical (MEMS) micromirror that sweeps a beam of infrared light across the eye while inexpensive photodiodes measure the reflection.[4][14] Removing the image sensor and its processing lets the system run at a high update rate for very little power: AdHawk has cited tracking at 250 Hz wirelessly and 500 Hz tethered, latency under 4 ms, and accuracy near 1 degree, with power on the order of 300 mW per eye and a target below 50 mW.[4][14] In March 2025 several outlets reported that Google was acquiring AdHawk, with figures cited around $100 million plus about $15 million in performance-based payments, to fold its low-power eye-tracking technology into the Android XR effort; coverage framed the deal as reported rather than formally confirmed at the time.[4][15][16] The relative strength of photodiode eye tracking is power and speed; camera-based systems generally reach finer accuracy, measured in tenths of a degree.[4]
Proximity and presence sensing
Most headsets need to know whether they are being worn. The common solution is an infrared proximity sensor seated between the lenses: an infrared LED emits light, and a photodiode detects how much of it reflects off the wearer's face, with the reflected intensity indicating how close an object is. When the headset is put on, the display wakes; when it is removed, the screen turns off to save power and reduce burn-in.[17] The Samsung Gear VR placed a small infrared sensor between the eyes for exactly this purpose, and the same proximity-detection pattern continues on standalone headsets such as the Meta Quest 3S.[17]
Optical heart-rate sensing on companion wearables
Photodiodes also drive the optical heart-rate sensors on the smartwatches and fitness bands that often accompany VR and AR use for health and fitness tracking. The technique, photoplethysmography (PPG), shines an LED into the skin, commonly green light around 525 to 565 nm because hemoglobin absorbs it strongly, and a photodiode next to the LED measures the light that reflects back. Each heartbeat changes the blood volume in the vessels and thus the amount of light absorbed, so the photodiode's signal pulses in time with the heart, and processing turns those pulses into a heart rate.[18][19]
Why photodiodes recur in XR hardware
Across these uses the appeal is consistent. A photodiode plus an infrared LED is small, inexpensive, and low-power, and it produces a simple analog signal rather than a full image that must be processed frame by frame. In Lighthouse this keeps per-sensor cost under a dollar and offloads tracking math from the headset.[3] In eye tracking it removes the camera's image-processing burden, which is what lets MEMS-and-photodiode systems hit high refresh rates at a fraction of a camera's power, a property that matters most for lightweight AR glasses and all-day wear.[4][14] The trade-off is that a photodiode reports intensity, not a picture, so designs that need rich spatial detail (full pupil imaging, hand tracking, passthrough) still rely on image sensors; photodiodes fill the roles where speed, power, and cost outweigh the need for an image.
References
- ↑ 1.0 1.1 1.2 1.3 1.4 1.5 "Photodiodes". https://www.rp-photonics.com/photodiodes.html.
- ↑ 2.0 2.1 "How Photodiodes Work and Their Applications". https://www.electrical4u.com/photodiode/.
- ↑ 3.0 3.1 3.2 3.3 "These Tiny Sensors Will Let You Build Lighthouse Tracked Headsets and Peripherals". 2016. https://www.roadtovr.com/triad-chips-lighthouse-steamvr-tracking-ts3633-cm1/.
- ↑ 4.0 4.1 4.2 4.3 4.4 4.5 "Google Set To Acquire Ultra-Low-Power Eye Tracking Startup". 2025-03-12. https://www.uploadvr.com/google-set-to-acquire-adhawk/.
- ↑ 5.0 5.1 "PIN diode". https://en.wikipedia.org/wiki/PIN_diode.
- ↑ 6.0 6.1 "Photodiode biasing (Photoconductive or Photovoltaic mode)". https://tavotech.com/photodiode-biasing-photoconductive-or-photovoltaic-mode/.
- ↑ 7.0 7.1 "Avalanche Photodiodes". https://www.meetoptics.com/academy/avalanche-photodiodes.
- ↑ 8.0 8.1 "Tiny infrared LEDs power new eye-tracking system for VR". https://optics.org/news/8/8/35.
- ↑ 9.0 9.1 "Back to Basics, Part 5: A Brief Timeline of Photovoltaics". https://www.pagerpower.com/news/back-to-basics-part-5-a-brief-timeline-of-photovoltaics/.
- ↑ "Understand Photodiode Technology". https://www.electronics-notes.com/articles/electronic_components/diode/photodiode-detector-technology.php.
- ↑ 11.0 11.1 "SteamVR tracking". http://www.xvrwiki.org/wiki/SteamVR_tracking.
- ↑ 12.0 12.1 "VR Trackers and Virtual Reality Tracking Explained - VR 101: Part III". https://blog.vive.com/us/tracking-in-virtual-reality-and-beyond-vr-101-part-iii/.
- ↑ "Power-efficient and shift-robust eye-tracking sensor for portable VR headsets". https://www.researchgate.net/publication/333502149_Power-efficient_and_shift-robust_eye-tracking_sensor_for_portable_VR_headsets.
- ↑ 14.0 14.1 14.2 "AdHawk Microsystems launches cameraless eye-tracking sensors for AR/VR". https://gamesbeat.com/adhawk-microsystems-launches-camera-less-eye-tracking-sensors-for-ar-vr/.
- ↑ "Google Reportedly Set to Acquire Eye-tracking Startup to Bolster Android XR Hardware Efforts". 2025-03-11. https://www.roadtovr.com/google-acquire-eye-tracking-adhawk-android-xr-glasses-rerpot/.
- ↑ "Google reportedly negotiating $115M deal for eye-tracking startup AdHawk Microsystems". 2025-03-11. https://siliconangle.com/2025/03/11/google-reportedly-negotiating-115m-deal-eye-tracking-startup-adhawk-microsystems/.
- ↑ 17.0 17.1 "VR Infrared Proximity Sensors: Technology, Applications, and Future Trends". https://www.sense-the-world.com/tech-hub/9482.html.
- ↑ "Monitoring Heart Rate with LED? Photoplethysmography (PPG)". https://global.samsungdisplay.com/29237/.
- ↑ "PPG Technology in Smartwatches: How Your Wearable Tracks Health Through Light". https://www.poalarhealth.com/blogs/news/ppg-technology-in-smartwatches-how-your-wearable-tracks-health-through-light.