17-Phys-B2 Electro-Optical Engineering · December 2018
Question 3 of 7: PIN Photodiode — Load Line, Saturation, Output Characteristic and Bandwidth Limit
Nivaar worked solution (AI-drafted; not reviewed by a licensed engineer)
Notes on this paper
Paper format. 17-Phys-B2 Electro-Optical Engineering,
National Examination December 2018 — a three-hour closed-book examination (one
8.5×11 inch double-sided handwritten note sheet permitted). The cover page states any
five of the seven questions constitute a complete paper and only the first
five as they appear in the answer book are marked; every question is nonetheless answered in
full below so the paper remains a complete study resource.
Reference texts. G. Keiser, Optical Fiber Communications, 4th ed.
(fiber modes, dispersion and bend loss; laser-diode longitudinal modes and DFB gratings; PIN
photodiode responsivity, noise and receiver design; link power and dispersion budgets); B. E. A.
Saleh and M. C. Teich, Fundamentals of Photonics, 2nd ed. (semiconductor optical gain,
LED spontaneous-emission linewidth, avalanche-photodiode noise); A. Yariv, Quantum
Electronics / E. Hecht, Optics, 5th ed. (Mach–Zehnder interferometry and
electro-optic modulators).
Find. (a) circuit and loop equation, (b) load line, (c) saturation optical
power, (d) $V_{\text{out}}(P)$ from 5–50 μW, (e) illuminated I-V and operating
mode, (f) bandwidth-limiting mechanism.
Approach. The photocurrent $I_{\text{ph}}=RP+I_d$ flows through $R_L$,
so $V_{\text{out}}=I_{\text{ph}}R_L$ traces a load line that clamps once the diode voltage
runs out of headroom; bandwidth is then set by whichever of the carrier transit time or the
$R_LC_j$ time constant is slower.
Part (a) — circuit and loop equation.
PIN detector bias circuit and loop equation.
The reverse-biased PIN
diode acts as a light-controlled current source in series with the bias supply and the load
resistor. Kirchhoff's voltage law around the loop gives
$$V_{\text{bias}}=I(R_L)\,R_L+V_D,\qquad I(R_L)=I_{\text{ph}}+I_d=RP+I_d,$$
where $V_D$ is the (reverse) voltage left across the diode itself.
Part (b) — load line. Solving the loop equation for $I$ in terms of
$V_D$ gives the straight line $I=(V_{\text{bias}}-V_D)/R_L$, running from
$(V_D,I)=(V_{\text{bias}},0)=(10\ \text{V},0)$ down to
$(0,V_{\text{bias}}/R_L)=(0,5.0\ \mu\text{A})$, slope $-1/R_L$.
Part (c) — saturation optical power. The diode saturates when all of
$V_{\text{bias}}$ has been dropped across $R_L$ and the diode itself is left with essentially
zero reverse voltage — the load-line endpoint on the current axis:
$$I_{\text{sat}}=\frac{V_{\text{bias}}}{R_L}=\frac{10\ \text{V}}{2\times10^{6}\ \Omega}
=5.0\ \mu\text{A},$$
$$P_{\text{sat}}=\frac{I_{\text{sat}}}{R}=\frac{5.0\ \mu\text{A}}{0.25\ \text{A/W}}
=\boxed{20.0\ \mu\text{W}}.$$
Part (d) — output voltage vs. optical power. Below saturation
$V_{\text{out}}=I_{\text{ph}}R_L=RPR_L$, a straight line of slope
$RR_L=0.25\times2\times10^{6}=5\times10^{5}\ \text{V/W}=0.5\ \text{V}/\mu\text{W}$; above
$P_{\text{sat}}=20.0\ \mu\text{W}$ it clamps at $V_{\text{bias}}=10\ \text{V}$.
$P$ (μW)
$I_{\text{ph}}$ (μA)
$V_{\text{out}}$ (V)
5
1.25
2.5
10
2.50
5.0
15
3.75
7.5
20 (=$P_{\text{sat}}$)
5.00
10.0 (clamped)
25–50
6.25–12.5
10.0 (clamped, saturated)
Part (e) — illuminated I-V and operating mode.
I-V load line (red) and illuminated diode characteristic (blue, reverse bias); Q-point at saturation.
Under
constant illumination and reverse bias, the diode's I-V characteristic is approximately a
horizontal line at $I\approx I_{\text{ph}}$ across the whole reverse-bias range (the
photocurrent is essentially independent of $V_D$ once fully depleted), intersecting the load
line at the operating (Q-) point. Because the diode is operated under an applied reverse bias
that sweeps out photo-generated carriers rather than at zero bias with an open-circuit
photovoltage, it is operating in photoconductive mode (as opposed to the
zero-bias photovoltaic mode used e.g. in solar cells).
Part (f) — bandwidth-limiting mechanism. The transit-time-limited
bandwidth follows from the depletion-region crossing time,
$$\tau_{\text{tr}}=\frac{d}{v_d}=\frac{30\times10^{-6}\ \text{m}}{5\times10^{4}\ \text{m/s}}
=0.60\ \text{ns},\qquad
f_{\text{tr}}\approx\frac{0.44}{\tau_{\text{tr}}}=\boxed{0.733\ \text{GHz}}.$$
The $R_LC_j$ time constant, using the given $2\ \text{M}\Omega$ load, is
$$\tau_{RC}=R_LC_j=(2\times10^{6}\ \Omega)(0.45\times10^{-12}\ \text{F})=900\ \text{ns},
\qquad f_{RC}=\frac{1}{2\pi\tau_{RC}}=\boxed{176.8\ \text{kHz}}.$$
Since $f_{RC}\ (176.8\ \text{kHz})$ is more than three decades below
$f_{\text{tr}}\ (0.733\ \text{GHz})$, the huge $2\ \text{M}\Omega$ load resistor —
not the depletion transit time — is what limits the usable bandwidth: this detector is
firmly RC-limited.