Photodetector Types and Technologies for PPG Sensing: A Comprehensive Guide

The photodetector is the critical signal transduction element in every PPG system, converting optical information about pulsatile blood volume into an electrical signal that can be amplified, digitized, and processed. While the LED illumination source receives significant design attention, the photodetector's characteristics -- responsivity, noise performance, spectral response, bandwidth, and active area -- fundamentally limit the achievable signal-to-noise ratio and determine the analog front-end architecture required to extract the cardiac signal from the raw photocurrent.

This guide provides a detailed technical treatment of photodetector technologies used in PPG systems, from the dominant silicon photodiode to emerging alternatives, with quantitative performance data and design guidance for sensor engineers and researchers working with PPG technology.

Silicon PIN Photodiodes: The PPG Workhorse

Silicon PIN photodiodes are used in over 99% of commercial PPG sensors, from clinical pulse oximeters to consumer wrist-worn wearables. Their dominance reflects an excellent combination of performance, cost, reliability, and manufacturability.

Operating Principle

A PIN photodiode consists of a p-type semiconductor layer, an intrinsic (undoped) layer, and an n-type layer. When reverse-biased, the intrinsic region becomes a depletion zone where incident photons with energy above the silicon bandgap (1.12 eV, corresponding to wavelengths below approximately 1100 nm) generate electron-hole pairs. These charge carriers are swept by the electric field across the depletion zone, producing a photocurrent proportional to the incident optical power.

The photocurrent I_ph is given by: I_ph = R * P_opt, where R is the responsivity in A/W and P_opt is the incident optical power. For silicon photodiodes, responsivity ranges from approximately 0.2 A/W at 500 nm to a peak of 0.5-0.6 A/W near 850-950 nm, then drops sharply above 1050 nm as the photon energy falls below the silicon bandgap. This spectral response conveniently spans the entire PPG operating range.

Key Performance Parameters for PPG

Responsivity and quantum efficiency. The external quantum efficiency (QE) of a silicon photodiode represents the fraction of incident photons that generate collected electron-hole pairs. Well-designed silicon photodiodes achieve QE of 70-90% in the 600-900 nm range, dropping to 50-70% at 525 nm (green) and 30-50% at 940 nm. The lower QE at green wavelengths means that green PPG requires higher LED drive power to achieve the same photodetector signal level as infrared PPG, contributing to the higher power consumption of green-wavelength PPG systems.

Dark current. Dark current represents thermally generated carriers that flow even without illumination, contributing DC offset and shot noise. For silicon PIN photodiodes at 25 degrees Celsius, typical dark current densities are 0.05-1 nA/mm^2, depending on manufacturing process and reverse bias voltage. A 3 mm^2 photodiode thus has dark current of 0.15-3 nA. This is small compared to typical PPG photocurrents (1-100 microamperes under LED illumination), but the associated shot noise contributes to the overall noise floor.

Dark current approximately doubles for every 10 degrees Celsius temperature increase. For wearable PPG sensors worn against the skin (local temperature 32-36 degrees Celsius), dark current can be 2-4 times higher than the room-temperature specification. This temperature sensitivity must be accounted for in noise budget calculations.

Junction capacitance. The capacitance of the photodiode junction directly affects the bandwidth and noise of the transimpedance amplifier (TIA) used to convert photocurrent to voltage. Junction capacitance is proportional to active area and inversely proportional to the square root of the reverse bias voltage. Typical values for PPG photodiodes are 5-50 pF for 1-7 mm^2 active areas at 1-3 V reverse bias.

Higher capacitance reduces TIA bandwidth and increases the TIA's noise gain at higher frequencies, degrading SNR. For PPG, the required bandwidth is modest (DC to approximately 25 Hz for the cardiac signal, or up to 200 Hz if resolving the waveform morphology for blood pressure estimation), so junction capacitance is rarely bandwidth-limiting. However, its effect on TIA noise performance is significant and must be considered in the analog front-end design.

Noise equivalent power (NEP). NEP represents the minimum detectable optical power, defined as the optical power that produces a signal equal to the noise floor. For silicon PIN photodiodes in PPG applications, NEP is typically 10^(-13) to 10^(-14) W/sqrt(Hz), well below the optical powers encountered in PPG measurement. The practical noise floor in PPG systems is dominated by TIA noise and shot noise from the ambient light photocurrent rather than the photodiode's intrinsic NEP.

Photodiode Configurations for PPG

Single photodiode. The simplest configuration uses a single photodiode positioned adjacent to the LED(s). LED-detector separations of 3-8 mm are typical for reflectance-mode wrist PPG, balancing signal amplitude (which decreases with distance) against signal modulation depth (which can increase with distance as the light samples deeper tissue). Mendelson and Ochs (1988) systematically studied the effect of LED-photodiode separation on PPG signal quality and found optimal separations of 4-7 mm for reflectance-mode forehead PPG (doi: 10.1109/10.7254).

Multi-photodiode arrays. Advanced PPG sensors use multiple photodiodes at different distances from the LED to sample different tissue depths simultaneously. The Apple Watch sensor module, for example, uses an array of four photodiode clusters surrounding multiple LED wavelengths. Multi-detector configurations enable differential measurement techniques that can suppress common-mode noise including ambient light interference and motion artifacts. The closer detector receives a stronger but shallower PPG signal, while the farther detector receives a weaker but deeper signal. Combining these signals can improve measurement robustness and enable multi-depth tissue analysis.

Back-to-back photodiode pairs. Some PPG designs use two photodiodes with one facing the tissue (signal + ambient light) and one facing away (ambient light only). Subtracting the ambient-only photodiode signal from the tissue-facing signal provides hardware-level ambient light rejection without requiring time-multiplexed LED switching.

Common PPG Photodiode Components

Several photodiode models are widely used in PPG sensor designs.

The Vishay VEMD5510 is a silicon PIN photodiode in a compact 0805 surface-mount package with 0.65 mm^2 active area, 15 pF capacitance, and 0.3 A/W responsivity at 940 nm. Its small size makes it suitable for miniaturized wearable sensors.

The OSRAM SFH 2201 offers a larger 1.0 mm^2 active area with an integrated daylight-blocking filter, providing 0.55 A/W responsivity at 870 nm and only 3 nA dark current. The integrated filter simplifies ambient light rejection for infrared-only PPG designs.

The Hamamatsu S1337 series provides larger active areas (1-6 mm^2) with excellent linearity and low dark current (0.1-1 nA), commonly used in clinical pulse oximeter designs where signal quality takes priority over miniaturization. Hamamatsu's technical documentation provides comprehensive guidance on photodiode selection for biomedical applications.

Integrated PPG sensor modules like the Maxim MAX30102 and TI AFE4404 combine the photodiode, LEDs, and analog front-end in a single package, simplifying design at the cost of reduced flexibility in component selection.

Phototransistors

Phototransistors provide internal current gain (typically 100-1000x compared to a photodiode), which can simplify the analog front-end by reducing TIA gain requirements. However, this gain comes with significant tradeoffs that limit phototransistor use in modern PPG systems.

Advantages and Limitations

The primary advantage is higher photosensitivity: a phototransistor with a gain of 500 produces 500 times the photocurrent of a photodiode with the same active area for a given optical power. This can enable PPG measurement with simpler, lower-gain amplification circuits.

The disadvantages are substantial. Phototransistors have significantly higher noise (the gain amplifies both signal and noise, and the transistor adds its own 1/f and shot noise), slower response times (10-100 microseconds versus under 1 microsecond for photodiodes, limiting the time-multiplexing speed for ambient light subtraction), higher dark current (10-100 nA), and poorer linearity. The gain is also temperature-dependent and varies between devices, making calibrated measurements difficult.

For these reasons, phototransistors were used in early PPG designs when amplifier performance was more limited, but have been almost entirely replaced by PIN photodiodes with high-performance TIAs in modern PPG systems. Lee et al. (2012) compared photodiode and phototransistor performance in a reflectance-mode PPG system and found that the photodiode configuration achieved 8-12 dB higher SNR despite the phototransistor's higher raw signal amplitude, due to the phototransistor's excess noise.

Avalanche Photodiodes (APDs)

Avalanche photodiodes operate at high reverse bias voltages (50-200 V for silicon) near the breakdown point, where impact ionization provides internal gain of 10-200x. APDs combine the speed of photodiodes with internal current multiplication.

Potential PPG Applications

APDs could theoretically enable PPG measurement at extremely low optical power levels, reducing LED drive current and extending battery life, or operating through thicker tissue or at larger LED-detector separations. They could also enable PPG measurement through highly absorbing tissue (dark skin at green wavelengths) where photodiode signals may be marginal.

However, APDs have several practical limitations for PPG. The high reverse bias voltage (50-200 V) requires a boost converter, adding circuit complexity, power consumption, and electromagnetic interference. Gain is highly sensitive to temperature (changing 2-5% per degree Celsius), requiring active bias voltage adjustment with temperature feedback. Excess noise from the avalanche multiplication process reduces the SNR advantage below the theoretical gain factor. APDs are also significantly more expensive than standard photodiodes.

For these reasons, APDs remain uncommon in PPG applications. Their primary use in biomedical optics is in diffuse optical tomography (DOT) and time-resolved spectroscopy where extremely weak optical signals must be detected. For standard PPG, the combination of high-efficiency LEDs and low-noise TIAs provides sufficient sensitivity without the complexity of APD bias circuits.

CMOS Image Sensors for Remote PPG

Camera-based remote photoplethysmography (rPPG) uses standard CMOS image sensors to detect the subtle skin color changes caused by pulsatile blood volume variations. This contactless approach has enabled PPG measurement from video without any dedicated optical hardware.

Operating Principle for rPPG

The pulsatile blood volume changes that drive contact PPG also modulate the reflected light from skin surfaces. These modulations are extremely small, typically 0.1-0.5% of the reflected light intensity, requiring careful signal processing to extract the cardiac signal from camera noise, ambient light variations, and motion.

CMOS image sensor pixels have sufficient sensitivity to detect these small modulations when spatial averaging across a region of interest (typically the face or forehead) reduces pixel-level noise. With a camera resolution of 640x480 and a facial region of interest covering 10,000-50,000 pixels, the effective spatial averaging reduces random noise by 100-220x (square root of pixel count), bringing the cardiac signal above the noise floor.

Verkruysse et al. (2008) demonstrated the first rPPG measurement using an ambient-light camera, achieving heart rate detection from facial video under normal indoor lighting. Subsequent work by Poh et al. (2010) applied independent component analysis (ICA) to the RGB color channels of facial video to separate the cardiac signal from noise and ambient light variations, achieving heart rate estimation with mean absolute errors of 2-5 BPM under controlled conditions (doi: 10.1364/OE.18.010762).

CMOS Sensor Requirements for rPPG

Frame rate. The cardiac signal occupies 0.5-4 Hz, requiring a minimum frame rate of 8 fps by the Nyquist criterion. In practice, 30 fps provides adequate temporal resolution with margin for anti-aliasing, and most cameras operate at 30-60 fps.

Bit depth. The small modulation depth of rPPG (0.1-0.5%) requires sufficient ADC resolution to avoid quantization noise. At 8-bit depth (256 levels), a 0.5% modulation on a half-scale signal corresponds to only 0.6 levels -- near the quantization noise floor. 10-bit or higher ADC resolution significantly improves rPPG signal quality. Many modern CMOS sensors provide 10-12 bit raw output even when the processed image is delivered at 8 bits.

Spectral response. Standard CMOS sensors with Bayer color filter arrays provide red, green, and blue channels. The green channel typically provides the strongest rPPG signal because hemoglobin absorption is highest at green wavelengths, consistent with contact PPG findings. The availability of separate color channels enables multi-wavelength analysis analogous to multi-LED contact PPG, supporting both heart rate and crude SpO2 estimation from camera video.

Noise performance. Read noise, fixed pattern noise, and temporal noise of the CMOS sensor all limit rPPG signal quality. Modern CMOS sensors with correlated double sampling achieve read noise levels of 1-5 electrons per pixel, which is sufficiently low for rPPG when spatial averaging is applied. For deeper understanding of how noise impacts PPG signal quality, see our guide on SNR improvement techniques.

Organic Photodetectors (OPDs)

Organic photodetectors based on organic semiconductor materials represent an emerging technology with unique properties for next-generation PPG sensors.

Material Systems and Properties

Organic photodetectors use thin films (100-500 nm) of organic semiconducting polymers or small molecules that absorb light and generate photocurrent through exciton generation and charge separation at donor-acceptor interfaces. Common material systems include P3HT:PCBM (poly(3-hexylthiophene) with phenyl-C61-butyric acid methyl ester) and newer non-fullerene acceptor blends that provide tunable spectral response.

Key advantages of OPDs for PPG include mechanical flexibility (enabling conformal skin-contact sensors that maintain better optical coupling), large-area fabrication on flexible substrates at low cost using printing techniques, tunable spectral response through material selection (enabling narrowband photodetection without external optical filters), and semitransparency (enabling novel sensor architectures where light passes through the detector).

Performance Comparison with Silicon

Current organic photodetectors have lower performance than silicon in several critical metrics. External quantum efficiency of 40-70% (versus 70-90% for silicon). Dark current density of 1-100 nA/cm^2 (versus 0.05-1 nA/mm^2 for silicon, which is 0.5-10 nA/cm^2 after unit conversion, so comparable for the best OPDs). Response times of 1-100 microseconds (versus under 1 microsecond for silicon). Operational stability is a significant concern, with OPD performance degrading over weeks to months under continuous use due to photo-oxidation and morphological changes.

Yokota et al. (2016) demonstrated a flexible organic PPG sensor on a 1-micrometer-thick polymer substrate that achieved heart rate measurement comparable to a commercial pulse oximeter, demonstrating the feasibility of ultra-thin conformable PPG sensors (doi: 10.1126/sciadv.1501856). Khan et al. (2016) developed a flexible organic optoelectronic sensor incorporating organic LEDs and organic photodetectors for pulse oximetry, achieving SpO2 measurement accuracy within 2% of a commercial pulse oximeter.

Timeline for Adoption

Organic photodetectors are unlikely to replace silicon in mainstream PPG applications in the near term due to performance, stability, and manufacturing maturity gaps. However, they enable form factors and applications that silicon cannot address, such as large-area patch-type PPG sensors, disposable clinical sensors, and fully flexible wearable sensors. As material stability improves and manufacturing processes mature, OPDs may find niche applications in next-generation PPG-based health monitoring systems.

Quantum Dot Photodetectors

Colloidal quantum dot (CQD) photodetectors use semiconductor nanocrystals whose bandgap is tunable through quantum confinement effects (controlled by nanocrystal size). This enables photodetection at specific wavelengths selected by choosing the appropriate quantum dot size, without external optical filters.

Advantages for PPG

Quantum dot photodetectors offer wavelength-selective absorption that can be precisely tuned during manufacturing, solution-processable fabrication compatible with flexible substrates, potential for multi-spectral detection using layers of different-sized quantum dots, and spectral response extending into the short-wave infrared (SWIR, 1000-2000 nm) using lead sulfide (PbS) or mercury telluride (HgTe) quantum dots, which silicon photodiodes cannot reach.

For PPG, the ability to create a photodetector that is inherently sensitive only to the LED wavelength (for example, a quantum dot detector tuned to 525 nm for green PPG) would eliminate the need for external optical bandpass filters, simplifying sensor module design and potentially improving ambient light rejection.

Current State of Development

Quantum dot photodetectors remain in the research stage for PPG applications. Current limitations include lower quantum efficiency than silicon (30-60% for the best CQD detectors versus 70-90% for silicon), higher dark currents, concerns about the toxicity of lead- and cadmium-based quantum dots (relevant for skin-contact devices), and limited long-term stability. Konstantatos and Sargent (2010) reviewed the state of colloidal quantum dot photodetectors and projected that performance parity with silicon could be achieved for specific niche applications, though broad replacement of silicon remains unlikely (doi: 10.1021/nl903451y).

Photodetector Selection Guidelines for PPG System Design

Key Selection Criteria

When selecting a photodetector for a PPG system, the following parameters should be evaluated in order of priority.

Spectral response match. The detector responsivity must be adequate at the LED wavelength(s). For multi-wavelength PPG systems using green (525 nm), red (660 nm), and infrared (940 nm) LEDs, a single silicon photodiode covers all three wavelengths with good responsivity. Check the specific responsivity at each operating wavelength to calculate the required LED drive current for adequate signal level.

Active area. Larger area increases signal but also increases capacitance, ambient light collection, and physical size. For wrist-worn wearables, 1-3 mm^2 is typical. For finger-clip pulse oximeters, 3-7 mm^2 is common. For earlobe sensors, 1-2 mm^2 is sufficient.

Dark current. Lower is better. Select photodiodes with dark current specifications at the expected operating temperature (skin temperature of 32-36 degrees Celsius), not just at the standard 25 degrees Celsius specification. Temperature-compensated designs may require additional circuit complexity.

Package and integration. Surface-mount packages (0805, QFN) are preferred for wearable assembly. Integrated sensor modules (photodiode + LEDs + AFE in one package) simplify design but limit flexibility. Consider optical aperture design, lead configuration, and compatibility with optical filter attachment.

Cost and availability. For consumer wearable products with volumes exceeding 100,000 units, component cost is significant. Standard silicon PIN photodiodes cost $0.10-0.50 in volume, while integrated modules cost $1-5. APDs and specialty detectors are $5-50 each, limiting them to clinical or research applications.

Application-Specific Recommendations

Consumer wrist wearable: Silicon PIN photodiode, 1-3 mm^2 active area, in integrated module with multi-wavelength LEDs and AFE. Examples: Maxim MAX30102, TI AFE4404 module.

Clinical pulse oximeter: Silicon PIN photodiode, 3-7 mm^2 active area, discrete component with external optical filter, paired with precision analog front-end. Example: Hamamatsu S1337 series.

Camera-based rPPG: Standard CMOS image sensor, minimum 30 fps, 10-bit or higher ADC, RGB Bayer filter. Leverage spatial averaging over region of interest for noise reduction.

Research/prototype: Start with integrated evaluation modules (MAX30102 EVKIT, AFE4404 EVM) for rapid prototyping, then migrate to discrete components for optimized performance. Document all detector specifications for reproducibility.

Understanding photodetector characteristics is essential for interpreting PPG signal quality and designing effective signal processing pipelines. The photodetector defines the fundamental signal quality ceiling that no amount of downstream processing can exceed.

Frequently Asked Questions

What type of photodetector is used in PPG sensors?

The vast majority of PPG sensors use silicon PIN photodiodes. Silicon photodiodes offer excellent responsivity across the PPG operating range (500-1000 nm), with peak quantum efficiency of 70-90% near 800-900 nm. They provide low dark current (0.1-10 nA), fast response times (under 1 microsecond), and are available in compact surface-mount packages compatible with wearable device assembly. Some specialized PPG applications use phototransistors for higher gain, avalanche photodiodes for extreme low-light conditions, or CMOS image sensor pixels for camera-based remote PPG.

How does photodetector active area size affect PPG signal quality?

Larger photodetector active areas collect more light, increasing the PPG signal amplitude proportionally. However, larger detectors also have higher junction capacitance (typically 5-50 pF per mm^2), which limits bandwidth and increases noise from the transimpedance amplifier. They also collect more ambient light, requiring higher dynamic range from the analog front-end. The optimal active area balances signal collection, noise, ambient light tolerance, and physical size constraints. Typical PPG photodiodes use active areas of 1-7 mm^2, with wearable sensors trending toward 1-3 mm^2 for compact integration.

Can camera sensors be used for PPG measurement?

Yes, standard CMOS camera sensors can detect PPG signals, enabling remote photoplethysmography (rPPG) from video of a person's skin. The camera pixels detect the subtle color changes caused by pulsatile blood volume variations beneath the skin surface. This works because CMOS sensors have sufficient sensitivity and frame rates (30-60 fps) to capture the 0.5-4 Hz cardiac signal. Accuracy is lower than contact PPG due to weaker signal amplitude, ambient light variations, and motion, but rPPG has been demonstrated for heart rate estimation with mean absolute errors of 2-5 BPM under controlled conditions (Verkruysse et al., 2008).

What is dark current and why does it matter for PPG?

Dark current is the electrical current that flows through a photodetector even when no light is incident on it, caused by thermally generated charge carriers in the semiconductor. For PPG, dark current adds a DC offset and shot noise to the measurement. At room temperature, typical silicon photodiode dark currents are 0.1-10 nA for 1-5 mm^2 active areas. This is generally small compared to the photocurrents in PPG operation (1-100 microamperes), but dark current doubles approximately every 10 degrees Celsius temperature increase, which can become relevant for sensors worn against warm skin (32-36 degrees Celsius).

References

• Single photodiode.* The simplest configuration uses a single photodiode positioned adjacent to the LED(s). LED-detector separations of 3-8 mm are typical for reflectance-mode wrist PPG, balancing signal amplitude (which decreases with distance) against signal modulation depth (which can increase with distance as the light samples deeper tissue). Mendelson and Ochs (1988) systematically studied the effect of LED-photodiode separation on PPG signal quality and found optimal separations of 4-7 mm for reflectance-mode forehead PPG (doi: 10.1109/10.7254).
• Verkruysse et al. (2008) demonstrated the first rPPG measurement using an ambient-light camera, achieving heart rate detection from facial video under normal indoor lighting. Subsequent work by Poh et al. (2010) applied independent component analysis (ICA) to the RGB color channels of facial video to separate the cardiac signal from noise and ambient light variations, achieving heart rate estimation with mean absolute errors of 2-5 BPM under controlled conditions (doi: 10.1364/OE.18.010762).
• Yokota et al. (2016) demonstrated a flexible organic PPG sensor on a 1-micrometer-thick polymer substrate that achieved heart rate measurement comparable to a commercial pulse oximeter, demonstrating the feasibility of ultra-thin conformable PPG sensors (doi: 10.1126/sciadv.1501856). Khan et al. (2016) developed a flexible organic optoelectronic sensor incorporating organic LEDs and organic photodetectors for pulse oximetry, achieving SpO2 measurement accuracy within 2% of a commercial pulse oximeter.
• Quantum dot photodetectors remain in the research stage for PPG applications. Current limitations include lower quantum efficiency than silicon (30-60% for the best CQD detectors versus 70-90% for silicon), higher dark currents, concerns about the toxicity of lead- and cadmium-based quantum dots (relevant for skin-contact devices), and limited long-term stability. Konstantatos and Sargent (2010) reviewed the state of colloidal quantum dot photodetectors and projected that performance parity with silicon could be achieved for specific niche applications, though broad replacement of silicon remains unlikely (doi: 10.1021/nl903451y).