LED and Photodiode Interview Questions

Q1. How does an LED emit light and what determines its emission wavelength?

An LED emits light through direct-bandgap radiative recombination — when electrons injected from the N-side recombine with holes in the P-region near the junction, the energy difference is released as a photon with wavelength λ = hc/Eg. A GaN-based blue LED has Eg ≈ 3.4 eV, emitting at λ = 1240/3.4 ≈ 365 nm (ultraviolet edge), while InGaN with indium content tuning shifts it to 450–470 nm blue. The narrow spectral output (FWHM 20–30 nm) of LEDs compared to incandescent bulbs is a direct consequence of the discrete bandgap energy determining photon wavelength.

Follow-up: Why can silicon not be used to make a practical LED?

Q2. What is the forward voltage of an LED and how does it relate to the bandgap?

The LED forward voltage Vf is approximately equal to the photon energy Eg/q in volts — a red AlGaInP LED at 630 nm has Vf ≈ 1.8–2.0V, a blue GaN LED at 470 nm has Vf ≈ 3.0–3.5V, and a UV LED at 365 nm has Vf ≈ 3.8V. When driving an LED from a 3.3V supply, a red LED needs only a 470 Ω resistor (R = (3.3-2.0)/20mA ≈ 65 Ω to 100 Ω), while a blue LED barely has enough headroom (3.3-3.2 = 0.1V) and requires a different drive approach. Using the wrong current-limiting resistor calculated for a red LED to drive a blue LED at 3.3V causes either the LED to not turn on or to overdrive it to failure.

Follow-up: What is the effect of temperature on LED forward voltage?

Q3. How do you drive an LED from a microcontroller GPIO — what circuit is needed?

The standard circuit is a series current-limiting resistor R = (Vsupply - Vf) / If, where Vf is the LED forward voltage and If is the desired current; the GPIO sources or sinks current through the resistor and LED. For a STM32 GPIO at 3.3V driving a red LED (Vf=2.0V) at 10 mA: R = (3.3-2.0)/0.010 = 130 Ω, use 120 Ω standard value. For high-brightness LEDs requiring 20–350 mA, a BJT (2N2222A) or MOSFET (2N7002) driver stage is required because GPIO source/sink current is limited to 8–25 mA depending on the MCU — direct connection at 20 mA can latch up or damage the GPIO pin.

Follow-up: What is the advantage of constant-current LED driving over constant-voltage with a resistor?

Q4. What is a photodiode and in what modes can it operate?

A photodiode is a reverse-biased P-N junction where incident photons with energy above the bandgap create electron-hole pairs in the depletion region, generating a photocurrent proportional to optical power. It operates in two modes: photoconductive mode (reverse biased, faster response, larger depletion region, some dark current) and photovoltaic mode (zero bias, zero dark current, slower, used for precision low-light measurement). The BPW34 silicon photodiode in photoconductive mode with -5V reverse bias has a depletion width of about 200 µm and a rise time of 50 ns, while in photovoltaic mode the depletion width is smaller and rise time increases to ~500 ns.

Follow-up: What is dark current in a photodiode and how does temperature affect it?

Q5. What is the responsivity of a photodiode and how is it measured?

Responsivity R = Iph / P_optical (A/W) describes how efficiently the photodiode converts optical power to photocurrent; a silicon photodiode like BPW34 has peak responsivity of about 0.62 A/W at 850 nm. To measure it: illuminate the photodiode with a calibrated 1 mW laser at 850 nm and measure the short-circuit current (or photocurrent through a precision transimpedance amplifier) — a 0.62 A/W device produces 620 µA. Responsivity falls off at short wavelengths (photons absorbed near the surface where recombination is fast) and at long wavelengths above the bandgap cutoff (~1100 nm for silicon).

Follow-up: What is quantum efficiency and how does it relate to responsivity?

Q6. What is a transimpedance amplifier (TIA) and why is it used with a photodiode?

A TIA converts the photodiode's current output to a voltage by placing a feedback resistor Rf across an op-amp configured as an inverting amplifier; the output voltage is Vout = -Iph × Rf. For a photodiode generating 10 µA at full illumination with Rf = 100 kΩ: Vout = 1V — easily measurable by an ADC. The TIA's virtual ground at the inverting input keeps the photodiode at nearly zero bias (photovoltaic mode) if the non-inverting input is at ground, eliminating dark current while providing low input impedance that minimizes noise from the large feedback resistor and photodiode capacitance.

Follow-up: How do you select Rf and the feedback capacitor Cf in a TIA for a given bandwidth?

Q7. What is the junction capacitance of a photodiode and how does it affect measurement bandwidth?

Junction capacitance Cj of a photodiode arises from the charge stored in the depletion region; it decreases with reverse bias as Cj ∝ 1/√Vr and directly limits bandwidth in a transimpedance circuit as BW = 1/(2π×Rf×Cj). A BPW34 at zero bias has Cj ≈ 60 pF, giving bandwidth of 1/(2π×100kΩ×60pF) ≈ 26 kHz with a 100 kΩ TIA — too slow for a 100 kHz optical link. Applying -5V reverse bias reduces Cj to ~20 pF and raises bandwidth to ~80 kHz; for 100 Mbps fiber optic receivers, InGaAs PIN photodiodes with Cj < 1 pF at -5V achieve bandwidth exceeding 1 GHz.

Follow-up: What is a PIN photodiode and how does the intrinsic layer help bandwidth?

Q8. What is an avalanche photodiode (APD) and in what applications is it used over a standard photodiode?

An APD operates at a high reverse voltage (50–300V) near breakdown where photogenerated carriers trigger avalanche multiplication, providing internal gain of 10–100× before the signal reaches the circuit, effectively amplifying weak optical signals. In fiber optic receivers at 10 Gbps, a Hamamatsu S8664-10 InGaAs APD with gain M=10 allows detection of optical powers as low as -30 dBm (1 µW) that a standard PIN photodiode would bury in circuit noise. The trade-off is excess noise factor F(M) that increases with gain, and stringent bias voltage regulation (±0.1V) needed to control the gain — making APD receiver circuits significantly more complex than PIN-based designs.

Follow-up: What is the excess noise factor in an APD and how does it differ between silicon and InGaAs?

Q9. What is the spectral response range of silicon photodiodes and what covers longer wavelengths?

Silicon photodiodes respond from about 400 nm (visible blue) to 1100 nm (near-infrared), with peak responsivity around 800–900 nm, because photons below 400 nm are absorbed in the surface passivation layer and photons above 1100 nm have energy below silicon's 1.12 eV bandgap. For wavelengths from 900 nm to 1700 nm (telecom O and C bands at 1310 nm and 1550 nm), InGaAs photodiodes with Eg ≈ 0.75 eV are used — all fiber optic transceivers use InGaAs receivers. Germanium photodiodes cover a similar range to InGaAs but with higher dark current, making them a lower-cost option for non-telecom NIR applications.

Follow-up: What semiconductor material is used for mid-infrared (3–5 µm) photodetectors?

Q10. What is the difference between a photodiode and a phototransistor — when do you choose each?

A phototransistor is a BJT where the base-collector photocurrent is amplified by the transistor's current gain (hFE typically 100–500), providing built-in gain that makes it more sensitive to moderate light levels; a photodiode has no gain but is much faster and more linear. For a simple object detection sensor in a printer using a BPT23 phototransistor, the output current at room light is large enough to directly drive a comparator without amplification. For a high-speed optical fiber link or LiDAR pulse detection requiring 10 ns response, a PIN photodiode into a TIA is mandatory because phototransistor bandwidth is limited to 100 kHz–1 MHz by the base charge storage time.

Follow-up: What is the bandwidth limitation of a phototransistor compared to a photodiode?

Q11. What is ambient light rejection in photodiode sensor design and how do you implement it?

Ambient light rejection is the technique of preventing steady-state background illumination from saturating or disturbing a photodiode sensor system, implemented by modulating the light source at a frequency outside the ambient light bandwidth and using an AC-coupled or bandpass receiver. In an IR proximity sensor (like VCNL4010 or APDS-9900), the IR LED is pulsed at 390 kHz and the receiver detects only the 390 kHz component, rejecting 50/100 Hz fluorescent flicker and DC sunlight. Without modulation, even 1 mW of sunlight on a BPW34 generates 620 µA of photocurrent that saturates the 100 kΩ TIA output, making the 1 µW proximity signal completely undetectable.

Follow-up: What is a lock-in amplifier and how does it relate to modulated light sensing?

Q12. How does a solar cell work and how is it related to the photodiode?

A solar cell is a large-area photodiode operated in photovoltaic mode — unbiased — where photogenerated electron-hole pairs are separated by the junction field and collected by external contacts, producing an open-circuit voltage Voc and short-circuit current Isc. A 156 mm × 156 mm silicon solar cell with peak illumination (1000 W/m²) generates Isc ≈ 8A and Voc ≈ 0.6V, with maximum power point at about 4.8W corresponding to 77% fill factor. The key difference from a signal photodiode is area (solar cell: 243 cm², signal photodiode: 1–10 mm²) and the absence of reverse bias, and both devices are modeled by the same Shockley diode equation with an added photocurrent term.

Follow-up: What is the fill factor of a solar cell and what determines it?

Q13. What is forward biasing versus reverse biasing a photodiode — what happens in each case?

Under forward bias a photodiode conducts like a normal diode (dominated by injection current), and any incident light only slightly modifies the forward current, making forward bias useless for linear light detection. Under reverse bias, only a small reverse saturation current flows in the dark, and photogenerated carriers easily cross the depletion region and add a photocurrent proportional to optical power — this is the useful operating mode. The BPW34 in the dark at -5V has Idark ≈ 2 nA; at 1 mW illumination at 870 nm, Iph ≈ 620 µA, giving a signal-to-dark ratio of 310,000:1, demonstrating why reverse bias is always used for linear detection.

Follow-up: What determines the maximum reverse bias for a photodiode before breakdown occurs?

Q14. What is optical isolation and how does an optocoupler use an LED and photodiode together?

Optical isolation passes a signal between two circuits with no electrical connection, using an LED on the input side and a phototransistor or photodiode on the output side separated by a transparent insulating barrier rated for 1–5 kV isolation voltage. In a PC817 optocoupler, a 20 mA LED input current produces a phototransistor collector current of 8–40 mA (CTR = 40–200%), providing both isolation and some current gain. Optocouplers are used to cross ground potential differences in motor drive gate isolation, medical device patient isolation, and industrial RS-485 interface protection — the isolation barrier's creepage distance determines the working voltage rating.

Follow-up: What is Current Transfer Ratio (CTR) in an optocoupler and how does it degrade over lifetime?

Q15. What is LED efficiency droop and what causes it at high current?

Efficiency droop is the phenomenon where LED luminous efficacy (lumens per watt) decreases as forward current increases above the optimal operating point, typically dropping 20–40% from peak to rated current. In high-power GaN blue LEDs like Cree XP-L, efficiency peaks around 50 mA and droops significantly by 700 mA, caused primarily by Auger recombination (CHSH process) where electron-hole recombination energy is transferred to a third carrier as heat rather than emitted as photons. Droop is why high-brightness LED drivers use large-area chips at moderate current density rather than a single small chip at maximum current — distributing 1A over 4 chips at 250 mA each maintains higher efficiency than 1A through one chip.

Follow-up: What is thermal droop and how does it compound with efficiency droop at high LED drive currents?