1. Theory

Definition: A pinned photodiode (PPD) is a p $^{+}$ /n/p photodiode in which a thin p $^{+}$ layer at the silicon surface is tied to the substrate potential, so that the n storage region is fully depleted when it is empty, and the surface potential is pinned at 0 V; it is the photodetector of most CCD and CMOS image sensors.
Four-transistor pixel: The PPD is read through a transfer gate (TG) into a floating diffusion (FD), which is reset (RS), buffered by a source follower (SF) and selected (SEL), so that the charge of the photodiode is moved completely to the node of small capacitance and read there.
Why it is used: The p $^{+}$ layer keeps the interface of Si and SiO $_{2}$ filled with holes and suppresses the generation of dark current at the surface; the complete transfer removes the lag and allows correlated double sampling, which cancels the reset noise of the floating node.
Full well: The number of electrons that the empty diode holds is set by its capacitance and by the potential between the pinning and the barrier of the transfer gate; 3000 electrons is the lowest value for a good image, 40000 electrons or more are held by a pixel of 6 μm.
Noise and dynamic range: Read noise below 3 electrons rms is typical, so that a full well of 40000 electrons gives more than $80 dB$ .
Dark current: State-of-the-art pixels of 1.4 μm have less than 15 electrons per second at 60 °C; the remaining sources are the surface defects at the Si-SiO $_{2}$ interface and the region near the transfer gate.
History: Buried photodiodes (Noble, 1968; Hagiwara, 1975; Koike, 1977) and the low-lag structure of Teranishi (1980 to 1982) led to the name pinned photodiode (Burkey, Kodak, 1984), to the CMOS pixel with intra-pixel transfer (JPL, 1993) and to the four-transistor CMOS pixel with a low-voltage PPD (1995).
Scaling: Pixel pitches of 10 to 20 times the technology node, back-side illumination with a thickness of 2 μm to 5 μm, shared readout (a transistor count of 1.375 per pixel for 8-way sharing) and global shutters with pinned storage nodes.
Figures of merit: The full well, the conversion gain, the read noise, the dynamic range, the dark current, the lag, the quantum efficiency and the pixel pitch.

2. Working Principle

Integration: With the transfer gate off, photoelectrons are collected in the n region of the diode, its potential falls until it reaches the barrier of the transfer gate, and the excess charge spills over; this sets the full well.
Reset and read: The floating diffusion is reset and sampled, the transfer gate is pulsed, the whole charge moves to the floating diffusion, which is sampled again, and the difference is the signal, free of reset noise.
Pinning: The n layer is depleted between the p $^{+}$ layer and the substrate; when the diode is empty its potential is the pinning potential $V_{p}$ , and the surface stays at 0 V, so that the interface states do not generate current into the storage region.
Transfer: The electrons move to the transfer gate by diffusion and by the built-in fields, in 12 ns for 5.6 μm and 600 ns for 40 μm on average, and a potential barrier or pocket between the photodiode and the gate channel gives incomplete transfer and image lag.
Negative gate bias: Holding the transfer gate off at -0.6 V instead of 0 V accumulates holes under the gate, which raised the well by $4.4 dB$ and lowered the dark current by $6.1 dB$ in a pixel of 3.2 μm.

3. Applications

Consumer and mobile cameras: Over 2.2 billion CMOS image sensors were shipped in 2013, with PPD pixels of 1.4 μm in back-side-illuminated sensors.
Global-shutter imaging: Pinned storage nodes and pump-gate pixels (2009 to 2013) for machine vision and high-speed imaging.
Scientific and low-light imaging: Pixels with a read noise of a few electrons and a conversion gain of 50 μV per electron or more, and charge-coupled imagers that use the same pinned structure.

Photodiode, Storage and Pinch-off

Name	Formula	Related Formulae	Variables/Constants
Photocurrent and charge integration	$I_{p h} = q \int Φ (λ) η (λ) d λ$	$\frac{d V}{d t} = \frac{I _{p h}}{C ( V )}$ $Q_{p h} = I_{p h} t_{in t} (electrons : N = \frac{I _{p h} t _{in t}}{q})$	$I_{p h} : p h o t oc u r r e n t$ $Φ : p h o t o n f l ux p er u ni t w a v e l e n g t h$ $η : q u an t u m e f f i c i e n cy (r e f l ec t i o n, ab sor pt i o n an d co l l ec t i o n in c l u d e d)$ $C (V) : v o l t a g e - d e p e n d e n t c a p a c i t an ce o f t h e f l o a t in g j u n c t i o n$ $t_{in t} : in t e g r a t i o n t im e$ $N : s i g na l e l ec t r o n s$ $(t h e f l o a t in g r e v er se - bia se d j u n c t i o n t ha t d i sc ha r g es a t t h e r a t e I_{p h} / C i s t h e in t e g r a t in g p h o t o d e t ec t or o f W ec k l er, 1965)$
Full well capacity and storage barrier	$N_{F W} = \frac{C _{P P D} ( V _{p} - V _{B} )}{q}$	$φ_{i} = V_{t} ln (\frac{N _{d} N _{a}}{n _{i}^{2}}), V_{t} = \frac{k _{B} T}{q}$ $q Δ V_{B} \approx 20 k_{B} T (Δ V_{B} \approx 0.5 V)$	$N_{F W} : f u l l w e l l in e l ec t r o n s$ $C_{P P D} : c a p a c i t an ce o f t h e p inn e d p h o t o d i o d e$ $V_{p} : p innin g p o t e n t ia l o f t h e e m pt y p h o t o d i o d e$ $V_{B} : p o t e n t ia l o f t h e ba r r i er a t w hi c h t h e c ha r g e s p i l l s o v er$ $φ_{i} : b u i l t - in v o l t a g e o f a p - n j u n c t i o n$ $N_{d}, N_{a} : d o n or an d a cce pt or d e n s i t i es$ $n_{i} : in t r in s i c d e n s i t y$ $Δ V_{B} : e x t r a ba r r i er h e i g h t t ha t h o l d s t h e e l ec t r o n s$ $k_{B} : B o l t z mann co n s t an t$ $(N_{F W} = 40000 n ee d s C_{P P D} (V_{p} - V_{B}) = 6.4 f F V an d 3000 n ee d s 0.48 f F V; 20 k_{B} T = 0.517 e V a t 300 K; φ_{i} = 0.83 V f or 1 0^{17} c m^{- 3} o n b o t h s i d es (a ss u m e d), a l l co m p u t e d)$
Pinch-off voltage from the pixel output	$V_{F D} = V_{r ese t} - \frac{V _{o u t}}{G _{S F}}$	$V_{o u t} = G_{S F} Δ V_{F D}$ $V_{p in} \approx V_{F D} at the change of slope of V_{o u t} (Q)$	$V_{F D} : v o l t a g e o f t h e f l o a t in g d i f f u s i o n$ $V_{r ese t} : r ese t v o l t a g e$ $V_{o u t} : p i x e l o u tp u t v o l t a g e$ $G_{S F} : g ain o f t h e so u r ce f o l l o w er$ $V_{p in} : p in c h - o f f (p innin g) v o l t a g e a t w hi c h t h e p h o t o d i o d e b eco m es e m pt y$ $(V_{r ese t} = 3.3 V, V_{o u t} = 2 V an d G_{S F} = 0.91 g i v e 1.10 V, co m p u t e d, t h e m e a s u r e d p in c h - o f f o f 1.1 V; o t h er p i x e l s : 1.2 V an d 1.55 V, an d a s hi f t o f ab o u t 1 V b y t h e d o p in g o f t h e t r an s f er g a t e)$

Gain, Noise and Dark Current

Name	Formula	Related Formulae	Variables/Constants
Conversion gain	$C G = \frac{q}{C _{F D}}$	$C G_{o u t} = G_{S F} \frac{q}{C _{F D}}$ $\frac{d σ _{V}^{2}}{d V _{o u t}} = \frac{q G _{S F}}{C _{F D}}$	$C G : co n v er s i o n g ain a t t h e f l o a t in g d i f f u s i o n$ $C_{F D} : c a p a c i t an ce o f t h e f l o a t in g d i f f u s i o n$ $G_{S F} : g ain o f t h e so u r ce f o l l o w er$ $σ_{V}^{2} : v a r ian ce o f t h e o u tp u t v o l t a g e, t h e l a s t f or m i s t h e p h o t o n - t r an s f er r e l a t i o n f or a P o i sso n s i g na l$ $(C_{F D} = 2.3 f F g i v es 69.7 μ V / e^{-}; 50 μ V / e^{-} i s t y p i c a l; 40 μ V / e^{-} a t t h e o u tp u t w i t h G_{S F} = 0.85 (a ss u m e d) m e an s 3.4 f F, co m p u t e d)$
Reset noise and correlated double sampling	$σ_{k_{B} T C} = \frac{k _{B} T C _{F D}}{q}$	$σ_{r e a d}^{2} = σ_{S F}^{2} + σ_{A D C}^{2}$ $S_{C D S} = V_{r ese t} - V_{s i g na l}$	$σ_{k_{B} T C} : r ese t n o i se o f t h e f l o a t in g d i f f u s i o n in e l ec t r o n s$ $σ_{r e a d} : r e a d n o i se a f t er cor r e l a t e d d o u b l e s am pl in g$ $σ_{S F}, σ_{A D C} : n o i se o f t h e so u r ce f o l l o w er an d o f t h e co n v er t er$ $S_{C D S} : d i f f er e n ce o f t h e r ese t an d s i g na l s am pl es, w hi c h c an ce l s t h e r ese t n o i se b ec a u se t h e c ha r g e t r an s f er l e a v es t h e f l o a t in g d i f f u s i o n r ese t$ $k_{B} : B o l t z mann co n s t an t$ $(C_{F D} = 2.3 f F an d 300 K g i v e 19.3 e^{-} r m s, co m p u t e d; t h e r e a d n o i se o f co mm er c ia l se n sor s i s b e l o w 3 e^{-} r m s)$
Shot noise, signal-to-noise ratio and dynamic range	$S N R = \frac{N}{N + N _{d} + σ _{r e a d}^{2}} \approx N$	$D R = 20 lo g_{10} (\frac{N _{F W}}{σ _{r e a d}})$ $S N R_{d B} = 20 lo g_{10} (S N R)$	$N : s i g na l e l ec t r o n s$ $N_{d} : d a r k e l ec t r o n s$ $σ_{r e a d} : r e a d n o i se in e l ec t r o n s$ $N_{F W} : f u l l w e l l$ $D R : d y nami c r an g e, t h e a pp r o x ima t e f or m h o l d s f or a s h o t - n o i se - l imi t e d s i g na l$ $(3000 e^{-} w i t h 3 e^{-} g i v es 60 d B an d 40000 e^{-} g i v es 82.5 d B; 3000 = 54.8, i . e . 34.8 d B, an d 40000 = 200, co m p u t e d)$
Dark current of the pixel	$N_{d} = \frac{I _{d} t _{in t}}{q}, σ_{d} = N_{d}$	$J_{d} = \frac{I _{d}}{A _{p}}, t_{e q} = \frac{σ _{r e a d}^{2}}{I _{d} / q}$ $Δ N_{d} = \frac{C _{F D} Δ V _{d}}{q}$	$N_{d} : d a r k e l ec t r o n s$ $I_{d} : d a r k c u r r e n t o f t h e p i x e l$ $A_{p} : p i x e l a r e a$ $σ_{d} : s h o t n o i se o f t h e d a r k c ha r g e$ $t_{e q} : t im e a t w hi c h t h e d a r k s h o t n o i se e q u a l s t h e r e a d n o i se$ $Δ V_{d} : d r o p o f t h e f l o a t in g d i f f u s i o n v o l t a g e in t h e d a r k$ $(15 e^{-} / s in a p i x e l o f 1.4 μ m a t 60^{\circ} C i s 2.4 \times 1 0^{- 18} A an d 1.2 \times 1 0^{- 10} A/c m^{2}, an d t_{e q} = 0.6 s f or 3 e^{-}; 3.77 mV an d 1.87 mV o n 2.3 f F a r e 54 an d 27 e^{-}, co m p u t e d)$

Charge Transfer

Name	Formula	Related Formulae	Variables/Constants
Diffusion transfer time	$t_{t r} \approx \frac{l ^{2}}{D _{n}}, D_{n} = \frac{k _{B} T}{q} μ_{n}$	$\frac{t _{2}}{t _{1}} = (\frac{l _{2}}{l _{1}})^{2}$ $t_{co m pl e t e} \approx (5 to 10) t_{t r}$	$t_{t r} : m e an t r an s f er t im e$ $l : l a t er a l d i s t an ce f r o m t h e p h o t o d i o d e t o t h e t r an s f er g a t e$ $D_{n} : e l ec t r o n d i f f u s i o n coe f f i c i e n t$ $μ_{n} : m o bi l i t y, t h e l a s t f or m i s t h e s t a t e m e n t o f t h e so u r ce f or co m pl e t e t r an s f er$ $k_{B} : B o l t z mann co n s t an t$ $(12 n s f or 5.6 μ m an d 600 n s f or 40 μ m a r e a r a t i o o f 50 a g ain s t 51 f or t h e s q u a r e o f t h e l e n g t h s; t h ey im pl y D_{n} = 26 c m^{2} /s, co m p u t e d)$
Retention over the barrier	$\frac{N _{o v er}}{N} \propto exp (- \frac{q Δ V _{B}}{k _{B} T})$	$L a g = \frac{Q _{l e f t}}{Q _{s i g na l}}$ $\frac{q N}{C _{P P D}} ≶ b (regimes of the transfer)$	$N_{o v er} : e l ec t r o n s t ha t cr oss t h e ba r r i er$ $Δ V_{B} : ba r r i er h e i g h t$ $Q_{l e f t} : c ha r g e l e f t in t h e p h o t o d i o d e a f t er t h e t r an s f er$ $b : ba r r i er p a r am e t er o f t h e so u r ce (n o t d e f in e d in t h e t e x t t ha t w a s r e a d), t h e f i r s t f or m i s t h e B o l t z mann f a c t or an d n o t a m e a s u r e d l a w$ $k_{B} : B o l t z mann co n s t an t$ $(20 k_{B} T g i v es a f a c t or exp (- 20) = 2.1 \times 1 0^{- 9}, co m p u t e d)$
Negative off-voltage of the transfer gate	$Δ N_{F W} (dB) = 20 lo g_{10} (\frac{N _{F W, 2}}{N _{F W, 1}})$	$N_{F W} = \frac{C _{F D} Δ V _{F D}}{q}$ $Δ J_{d} (dB) = 20 lo g_{10} (\frac{Δ V _{d, 1}}{Δ V _{d, 2}})$	$N_{F W, 1}, N_{F W, 2} : f u l l w e l l a t tw o o f f - v o l t a g es o f t h e t r an s f er g a t e$ $Δ V_{F D} : o u tp u t s w in g o n t h e f l o a t in g d i f f u s i o n (a ss u m e d t o b e t h e n o d e v o l t a g e)$ $Δ V_{d} : d a r k d r o p o f t h e f l o a t in g d i f f u s i o n$ $(1.04 V a t - 0.6 V a g ain s t 0.629 V a t 0 V i s 4.4 d B, an d 3.77 mV a g ain s t 1.87 mV i s 6.1 d B; w i t h 2.3 f F t h e w e l l s a r e 14900, 9030 an d 3560 e^{-} a t - 0.6, 0 an d + 0.6 V, co m p u t e d)$

Important Questions

Ques: Why does pinning lower the dark current and allow complete transfer?

The p $^{+}$ layer fixes the surface at 0 V and keeps holes at the Si-SiO $_{2}$ interface, so that interface states do not generate electrons into the storage region; the source lists the surface defects and a boron-implanted region above the transfer gate as the remaining sources. The n region is depleted when empty, so the whole charge can move to the floating diffusion, and the reset noise of the floating node can be removed by sampling it before and after the transfer. A transfer gate held at -0.6 V gave a floating-node drop in the dark of 1.87 mV instead of 3.77 mV, i.e. $6.1 dB$ (computed), 27 against 54 electrons for 2.3 fF (computed), but below -0.6 V the dark current rose steeply through a direct path to the floating diffusion; the integration time is not given. Reference: Fossum and Hondongwa, A Review of the Pinned Photodiode for CCD and CMOS Image Sensors, IEEE J. Electron Devices Soc. (2014); Mheen, Song and Theuwissen, Negative Offset Operation of Four-Transistor CMOS Image Pixels for Increased Well Capacity and Suppressed Dark Current, IEEE Electron Device Lett. 29, 347 (2008).

Ques: What fixes the full well, and how many electrons is it?

$N_{F W} = C_{P P D} (V_{p} - V_{B}) / q$ : 40000 electrons need $C_{P P D} (V_{p} - V_{B}) = 6.4$ fF·V and 3000 need 0.48 fF·V (computed). The source gives neither $V_{p}$ nor $V_{B}$ . In a pixel of 3.2 μm the full well was 1.04 V at -0.6 V, 0.629 V at 0 V and 0.248 V at +0.6 V of gate off-voltage; if this is the swing on the floating diffusion of 2.3 fF, it is 14900, 9030 and 3560 electrons (computed), but the table of the electron numbers was not in the text read, and the output may include the gain of the follower. A simulation that was not calibrated overestimated the well by 4 to 6 thousand electrons. Reference: Fossum and Hondongwa, A Review of the Pinned Photodiode for CCD and CMOS Image Sensors, IEEE J. Electron Devices Soc. (2014); Mheen, Song and Theuwissen, Negative Offset Operation of Four-Transistor CMOS Image Pixels for Increased Well Capacity and Suppressed Dark Current, IEEE Electron Device Lett. 29, 347 (2008); Van Sieleghem et al., Efficient Methods for Analyzing the Floating Diffusion, Photodiode, and Transfer Gate of a 4T Pixel and Calibrating TCAD, International Image Sensor Workshop (2025).

Ques: How large is the reset noise, and why does the four-transistor pixel remove it?

The reset of a floating node of 2.3 fF leaves a noise charge $k_{B} T C / q = 19.3$ electrons rms at 300 K (computed), six times the read noise below 3 electrons of commercial sensors, and it would limit the dynamic range of 40000 electrons to $66 dB$ (computed) instead of $82.5 dB$ . In the four-transistor pixel the node is reset and sampled first, then the photodiode charge is transferred and the node is sampled again, and the difference does not contain the random charge of the reset, because the pinned photodiode is separate from the node. The source states that correlated double sampling suppresses the reset noise on the floating diffusion and gives no equation for it; the value above is the standard $k_{B} T C$ formula. Reference: Fossum and Hondongwa, A Review of the Pinned Photodiode for CCD and CMOS Image Sensors, IEEE J. Electron Devices Soc. (2014); Mheen, Song and Theuwissen, Negative Offset Operation of Four-Transistor CMOS Image Pixels for Increased Well Capacity and Suppressed Dark Current, IEEE Electron Device Lett. 29, 347 (2008).

Ques: What is the dark current of a good pixel, and when does it matter?

The state-of-the-art value of less than 15 electrons per second for a 1.4 μm pixel at 60 °C is a current of $2.4 \times 1 0^{- 18}$ A and a density of $1.2 \times 1 0^{- 10} A/c m^{2}$ (computed). The shot noise of this charge equals a read noise of 3 electrons after 0.6 s (computed), and it is 0.7 electron in 33 ms, so that at video rates the dark current of such a pixel is negligible against the read noise and matters for long exposures and for high temperature. The review states the value as the state of the art without a specific measurement, and the pixel of 3.2 μm was characterized only by the voltage drop of the floating node. Reference: Fossum and Hondongwa, A Review of the Pinned Photodiode for CCD and CMOS Image Sensors, IEEE J. Electron Devices Soc. (2014); Mheen, Song and Theuwissen, Negative Offset Operation of Four-Transistor CMOS Image Pixels for Increased Well Capacity and Suppressed Dark Current, IEEE Electron Device Lett. 29, 347 (2008).

Ques: Is the transfer time of 12 ns and 600 ns consistent with diffusion?

The two pixels differ by a factor of 50 in time (600/12) and 51 in the square of the pitch, $(40/5.6)^{2}$ (computed), as for diffusion with $t \approx l^{2} / D_{n}$ , and the implied coefficient is 26 $c m^{2} /s$ for both (computed), a mobility of about 1000 $c m^{2} /Vs$ by the Einstein relation, plausible for doped silicon (the mobility is an inference). The time is the average, the complete transfer takes 5 to 10 times longer, and the review gives a shorter time, below 10 ns, for pixels with fields from the layout and a power law of the distance for field-assisted transfer whose exponent was not legible. Lag appears when a potential barrier remains between the diode and the gate channel. Reference: Fossum and Hondongwa, A Review of the Pinned Photodiode for CCD and CMOS Image Sensors, IEEE J. Electron Devices Soc. (2014); Cao et al., Measurement of charge transfer potential barrier in pinned photodiode CMOS image sensors, J. Semicond. 37, 054007 (2016); Li et al., An approach to obtain the pinch-off voltage of 4-T pixel in CMOS image sensor, J. Semicond. 31, 074010 (2010).

Ques: Which conversion gain do the sources give, and are they consistent?

A floating diffusion of 2.3 fF gives $q / C = 69.7 μ V/ e^{-}$ at the node (computed), and the review quotes about 50 μV per electron as typical, which corresponds to 3.2 fF (computed) or to the gain of the source follower, 0.72. The paper on the pinch-off voltage prints ‘40 V/e’ for its pixel, which is evidently 40 μV per electron at the output; with a follower gain of 0.85 (assumed) it means a floating capacitance of 3.4 fF (computed). The same paper obtains the gain from the variance of the output, $d σ^{2} / d V_{o u t} = q G / C_{F D}$ , which requires a shot-noise-limited signal. A prototype array reported 0.35 digital number per electron, a unit that depends on the converter. Reference: Fossum and Hondongwa, A Review of the Pinned Photodiode for CCD and CMOS Image Sensors, IEEE J. Electron Devices Soc. (2014); Mheen, Song and Theuwissen, Negative Offset Operation of Four-Transistor CMOS Image Pixels for Increased Well Capacity and Suppressed Dark Current, IEEE Electron Device Lett. 29, 347 (2008); Li et al., An approach to obtain the pinch-off voltage of 4-T pixel in CMOS image sensor, J. Semicond. 31, 074010 (2010); Cao et al., Measurement of charge transfer potential barrier in pinned photodiode CMOS image sensors, J. Semicond. 37, 054007 (2016).

Ques: Why do the two equations for the pinch-off voltage differ?

Eq. (1) of the paper gives $V_{s i g na l} = (V_{r ese t} - V_{o u t}) / G$ , which is 1.44 V for $V_{r ese t} = 3.3$ V, $V_{o u t} = 2$ V and $G = 0.9$ , while Eq. (8), $V_{F D} = V_{r ese t} - V_{o u t} / G$ , gives 1.08 V (both computed). The result of 1.1 V requires $G = 0.91$ with Eq. (8), slightly above the stated range of 0.8 to 0.9, and Eq. (8) is consistent only if $V_{o u t}$ is the swing at the output and not an absolute voltage. The measured 1.1 V is larger than the 0.6 V to 0.7 V built-in voltage of a silicon junction, and the paper does not explain the difference; another pixel had 1.2 V and 1.55 V, and a change of the doping of the transfer gate shifted the barrier by about 1 V. Reference: Li et al., An approach to obtain the pinch-off voltage of 4-T pixel in CMOS image sensor, J. Semicond. 31, 074010 (2010); Van Sieleghem et al., Efficient Methods for Analyzing the Floating Diffusion, Photodiode, and Transfer Gate of a 4T Pixel and Calibrating TCAD, International Image Sensor Workshop (2025).

Ques: How should the signal-to-noise ratios in the review be read?

The review writes ‘3000/√3000 = 34’, but $3000 = 54.8$ (computed); the number 34 matches $34.8 dB$ , i.e. $20 lo g_{10} 54.8$ (computed), so it is probably the ratio in decibels without the unit. For 40000 electrons the ratio is 200, or $46 dB$ (computed). The dynamic range of $60 dB$ for 3000 electrons and 3 electrons of read noise, and ‘more than $80 dB$ ’ for 40000 electrons ( $82.5 dB$ , computed), are correct. Shot noise limits the image to $N$ , and the signal shot noise equals a read noise of 3 electrons at 9 electrons (computed), so that a sensor with a read noise below 3 electrons is shot-noise-limited above 9 signal electrons. Reference: Fossum and Hondongwa, A Review of the Pinned Photodiode for CCD and CMOS Image Sensors, IEEE J. Electron Devices Soc. (2014).

Ques: How does the full well scale as the pixel shrinks?

The review quotes at least 3000 electrons for good image quality and about 40000 electrons for a pixel of 6 μm, a density of 1110 electrons per $μ m^{2}$ , while 3000 electrons in a pixel of 1.4 μm is 1530 per $μ m^{2}$ (both computed), so a small pixel needs a larger density; a negative off-voltage of the transfer gate raised the well by $4.4 dB$ in the 3.2 μm pixel. The pitch has halved every four years, from 10 to 20 times the technology node, and the ratio of the pitch to the absorption length has fallen well below one, which goes with back-side illumination and a silicon thickness of 2 μm to 5 μm. The depths of the doped layers printed in the review (an n region of 2.5 μm to 5 μm) look large for a 180 nm process and were not used. Reference: Fossum and Hondongwa, A Review of the Pinned Photodiode for CCD and CMOS Image Sensors, IEEE J. Electron Devices Soc. (2014); Van Sieleghem et al., Efficient Methods for Analyzing the Floating Diffusion, Photodiode, and Transfer Gate of a 4T Pixel and Calibrating TCAD, International Image Sensor Workshop (2025).

Quartz 5

Explorer

Pinned Photodiode (CMOS Image Sensor Pixel)