---
title: 'Photon Avalanche: Mechanisms & Applications'
url: https://www.emergentmind.com/topics/photon-avalanche-pa
type: topic
---

# Photon Avalanche: Mechanisms & Applications

Searching arXiv for recent and foundational papers on photon avalanche across detector and luminescent-material contexts.
{"query":"all:\"photon avalanche\" OR title:\"photon avalanche\"","max_results":10,"sort_by":"relevance"}
Photon avalanche (PA) is a polysemous term in modern optics and photodetection. In reverse-biased avalanche photodiodes, single-photon avalanche diodes, and silicon photomultipliers, it denotes the self-sustaining impact-ionization process by which a single absorbed photon can trigger a macroscopic current pulse in Geiger mode. In lanthanide-doped upconverting media, it denotes a thresholded nonlinear population-dynamics mechanism based on excited-state absorption and cross-relaxation, with long rise times at threshold and very large power-law exponents. More recent usage extends the term to broadband emission in Yb³⁺/Er³⁺–trimesate metal-organic frameworks and to single-photon-stimulated switching between metastable states of a coherently driven Kerr cavity [1202.1591][2007.10551][2507.16293][2606.27555].

## 1. Terminological scope and conceptual separation

In semiconductor detector physics, PA refers specifically to avalanche multiplication of charge carriers in a high-field p–n junction operated above breakdown. A single photo-generated electron–hole pair can trigger a self-sustaining avalanche, the output is effectively digital, and quenching plus recovery define the detector dead time. In this usage, the term is explicitly distinguished from “photon avalanche” in upconversion or luminescent materials [2308.13603][1202.1591].

In lanthanide-doped solids and nanoparticles, the same term names a different microscopic process. There the defining ingredients are dominant excited-state absorption over ground-state absorption, a clear excitation threshold, and anomalously long rise time near threshold; in Tm³⁺-doped NaYF₄ core/shell nanoparticles, fitting yields an ESA-to-GSA ratio \(R_2/R_1 \approx 13{,}000\), and the observed emission can scale with the 26th power of pump intensity [2007.10551]. In Yb³⁺/Er³⁺–trimesate MOFs, the central conceptual shift is that the avalanche occurs in the organic molecule part rather than in the lanthanide manifold, and relies on a cooperative process involving multiple emission centers [2507.16293].

A third usage has now emerged in driven-dissipative quantum optics. In a coherently driven Kerr cavity biased near optical bistability, a single incident photon can stimulate a jump from a low-photon state to a high-photon state, yielding a macroscopic and long-lived increase of intracavity photon number and transmitted intensity [2606.27555]. This suggests a family resemblance across the literature—thresholded positive feedback producing disproportionate output—while the microscopic mechanisms remain categorically different.

## 2. Semiconductor photon avalanche in APDs, SPADs, and SiPMs

In Geiger-mode semiconductor detectors, the operational condition is reverse bias above breakdown. For a silicon avalanche photodiode, the overbias is defined as
\[
V_\text{over} = V_\text{bias} - V_\text{br},
\]
and operating with \(V_\text{over}>0\) makes the device sensitive to single photons because any carrier injected into the high-field region can trigger a self-sustaining avalanche with high probability [1202.1591]. Once started, the avalanche must be quenched by lowering the bias below breakdown; the subsequent recharge interval is the dead time. Trap release after an avalanche produces afterpulses, and thermally or tunneling generated carriers produce dark counts [1202.1591][1505.04407].

This carrier avalanche can be engineered at very high repetition rates. In InGaAs avalanche photodiodes at telecom wavelengths, 2 GHz gating with a tunable self-differencing circuit allows extremely weak avalanches to be sensed while suppressing afterpulsing; the reported afterpulse probabilities are \(4.84\%\) and \(1.42\%\) for photon detection efficiencies of \(23.5\%\) and \(11.8\%\), respectively [1001.0934]. A different control strategy replaces the fixed quenching resistor with a bias-dependent adaptive resistive switch; with a Pt/Al\(_2\)O\(_3\)/Ag resistor and a commercial silicon SPAD, avalanche pulse widths as small as \(\sim 30\) ns were demonstrated, about \(10\times\) smaller than in a passively quenched approach [2105.11454]. At the low-afterpulsing extreme, a commercial reach-through silicon APD operated in Geiger mode exhibited afterpulsing systematically less than \(0.05\) percent at \(10\) V excess voltage [1505.04407].

In SiPMs, the avalanche-triggering probability is not only bias dependent but also depth dependent. A physics-motivated parameterization shows that in p-on-n SiPMs the induced avalanches are electron-driven in the ultra-violet and near-ultra-violet ranges, while they become increasingly hole-driven toward the near-infra-red range [1904.05977]. The avalanche itself is also optically emissive: measurements on Hamamatsu VUV4 and FBK VUV-HD3 SiPMs found \(40\pm 9\) and \(61\pm 11\) photons produced per avalanche at the source at 4 volts over-voltage, with no significant temperature dependence observed within measurement uncertainties [2402.09634]. Those secondary photons underlie internal and external optical cross-talk.

## 3. Statistical descriptions, inversion methods, and number-state information

A central feature of semiconductor PA is that the device output is intrinsically binary over a dead-time interval, even though the underlying optical field need not be. For a pulse containing exactly \(n\) photons in the active window, a standard SPAD model uses
\[
P(\text{click}\mid n)=1-(1-\eta)^n(1-p_\mathrm{dc}),
\]
with detection efficiency \(\eta\) and dark-count probability \(p_\mathrm{dc}\). On this basis, number-state reconstruction with a single SPAD becomes a statistical inversion problem rather than direct photon-number resolution [2308.13603].

Using maximum-likelihood reconstruction with a detector model whose parameters are measurable, photon number-state reconstruction has been demonstrated with only one SPAD. The reported agreement is excellent for coherent states with up to \(\approx 10\) photons and peak input photon rates up to several Mcounts/s, and when detector imperfections are small, good agreement is maintained for coherent pulses with peak input photon rates of over 40 Mcounts/s, greater than one photon per detector dead time. For anti-bunched light, reconstructed and independently measured pulse-averaged values of \(g^{(2)}(0)\) are consistent, provided the pulse width and correlation time scales are both at least a few detector dead times [2308.13603]. This is significant because it reframes PA-based binary detection as a computationally invertible measurement process.

Avalanche detectors can also be used in a photon-number-resolving regime when the avalanche pulse height retains information about the number of initiating carriers. With two high-speed avalanche photodiodes, higher-order photon correlations were probed through the quantity
\[
\gamma(n_1+n_2)=\frac{g(n_1+n_2)}{g(n_1)g(n_2)}.
\]
For a filtered multimode laser near threshold, the measured higher-order correlation reached \(\sim 47.1 \pm 4.5\), whereas a coherent-like laser above threshold remained close to unity [1107.3089]. This demonstrates that avalanche detection can expose correlation structure far beyond standard click-based \(g^{(2)}\) measurements.

At a more microscopic level, the statistics of electron-hole avalanches are governed by a two-species branching process. For constant fields, the avalanche parameter
\[
A=\frac{\alpha v_e N_e^0 + \beta v_h N_h^0}{\alpha v_e+\beta v_h}
\]
controls both gain fluctuations and timing; the intrinsic timing jitter at large thresholds saturates to
\[
\sigma_T^\infty = \frac{\sqrt{\psi_1(A)}}{\alpha v_e+\beta v_h}.
\]
This connects avalanche build-up, excess noise, and time resolution in SPADs, SiPMs, APDs, and LGADs within a single stochastic framework [2012.11285].

## 4. Photon-avalanche upconversion in lanthanide systems

In lanthanide-doped materials, PA is a nonlinear upconversion mechanism rather than an electronic breakdown process. In Tm³⁺-doped NaYF\(_4\) core/shell nanoparticles, the defining loop is
\[
{}^3\mathrm{F}_4 \xrightarrow{\text{ESA}} {}^3\mathrm{H}_4 \xrightarrow{\text{CR}} 2\times {}^3\mathrm{F}_4,
\]
so population in the intermediate reservoir state increases the excited-state absorption rate, which in turn feeds cross-relaxation and positive feedback. The fitted ratio \(R_2/R_1 \approx 13{,}000\) satisfies the usual PA criterion that ESA dominate GSA, the rise time near threshold reaches \(\tau_{\text{rise,max}} \approx 608\) ms, nearly \(400\tau_2\), and the nonlinear exponent reaches \(s=26.0\) in optimized 8% Tm³⁺ particles [2007.10551]. Thresholds of about 20 kW/cm\(^2\) were observed for ensemble 8% Tm³⁺ particles at 1064 nm, and improved core/shell geometries reduced the threshold to below 10 kW/cm\(^2\) [2007.10551].

The same extreme nonlinearity enables photon-avalanche single-beam superresolution imaging. PASSI achieves sub-70 nm spatial resolution using simple scanning confocal microscopy and before any computational analysis; near the PA threshold, single-particle images showed short-axis and long-axis FWHM values of \(65 \pm 7\) nm and \(81 \pm 9\) nm, respectively [2007.10551]. The mechanism is therefore both spectroscopic and functional: the avalanche reshapes the effective point-spread function.

PA is also unusually susceptible to quenching. In LiYF\(_4\):3%Tm\(^{3+}\) co-doped with Nd\(^{3+}\), resonant energy transfer from the looping level increases \(I_{\mathrm{th}}\) and \(S_{\max}\), whereas energy transfer from the emitting level diminishes \(S_{\max}\) and the final emission intensity [2405.05764]. Experimentally, at fixed high pump power the PA 800 nm emission decreases by four orders of magnitude when going from 0% to 1% Nd\(^{3+}\), while the corresponding Stokes 800 nm emission decreases only by about a factor of \(\sim 4\) [2405.05764]. The relative sensitivity \(S_R\) reaches up to \(\sim 900\%\) per 1% Nd\(^{3+}\), which is why PA is being considered as an ultra-sensitive fluorescence-based reporting mechanism [2405.05764]. A common misconception is that all strong upconversion nonlinearity is PA; the data instead indicate that PA requires threshold behavior, dominant ESA, and long rise-time anomalies, whereas ordinary upconversion usually shows \(S\approx2\text{–}4\) [2405.05764][2007.10551].

## 5. Hybrid material and cavity realizations

A recent extension of PA to porous hybrid materials is the Yb³⁺/Er³⁺–trimesate metal-organic framework. Here infrared photons are first absorbed and upconverted by the lanthanide ions, but the avalanche itself proceeds in the organic molecule part and relies on a cooperative process involving multiple emission centers. The resulting emission is spectrally broadband and arises mainly from ligand triplet states rather than from conventional sharp lanthanide lines [2507.16293]. The nonlinearity factor is reported to be comparable with well-established PA inorganic nanoparticles, and the effect is strongly related to crystallinity: not well-formed frameworks support only the characteristic Er\(^{3+}\) emission with only linear increase as a function of excitation power [2507.16293]. This broadens the PA materials landscape beyond a limited set of lanthanide ions.

A still more abstract implementation is theoretical: a single photon incident on a coherently driven Kerr-nonlinear cavity near optical bistability can stimulate a jump from the low-photon-number to the high-photon-number state of the bistability loop [2606.27555]. The cavity is treated quantum mechanically, with quantum fluctuations included through a Lindblad master equation and the single photon introduced through a cascaded-system coupling. In the optimal regime, the single-photon-stimulated jump probability \(\bar{P}_j\) is found to be on the order of \(0.1\text{–}0.3\), the ratio \(Q/\Gamma_{LH}\) can exceed \(\sim 10\), and the avalanche gain can range from tens to thousands of emitted photons beyond the trigger [2606.27555]. This reinterprets PA as single-photon-triggered switching across a first-order-like dynamical transition and suggests an all-optical single-photon avalanche detector.

## 6. Applications, limitations, and persistent misconceptions

In semiconductor instrumentation, PA underlies quantum optics measurements, quantum communication, QKD, time-correlated single-photon counting, fluorescence, LiDAR, and transient imaging [1202.1591][1001.0934][1703.02635]. In SiPM-based large-area detectors, avalanche-induced secondary photons are now directly relevant to external cross-talk budgets in cryogenic experiments [2402.09634]. In upconversion media, PA supports superresolution imaging, temperature and pressure transduction, neuromorphic computing, and quantum optics, while the MOF results point toward combined porous-host, sensing, and energy-conversion functionalities [2007.10551][2507.16293].

The limitations are domain specific. In SPADs and APDs, dead time, afterpulsing, saturation, cross-talk, and imperfect detector models constrain dynamic range and count rate; at 3 GHz InGaAs gating, for example, the jitter of \(\sim 380\) ps exceeds the 333 ps clock period and becomes prohibitively high [1001.0934]. Single-SPAD number-state reconstruction is demonstrably useful but is applicable only when pulse width and correlation time scales are at least a few detector dead times, and the demonstrated reconstruction range is up to roughly 10 photons [2308.13603]. In passive or adaptive quenching architectures, faster reset can come with degradation, higher dark counts, or afterpulsing trade-offs [2105.11454]. In upconversion PA, threshold intensities of \(\sim 10\text{–}40\) kW/cm\(^2\) remain high for many applications, rise times can reach hundreds of milliseconds, and material synthesis must control surface quenching, dopant concentration, and crystallinity [2007.10551][2507.16293].

Two misconceptions recur. The first is terminological: PA is not a single universal microscopic mechanism. In SPADs and SiPMs it is an electronic carrier avalanche, in lanthanide systems it is a nonlinear excitation-and-cross-relaxation loop, and in Kerr cavities it is a metastable jump in a driven quantum optical phase space [2308.13603][2606.27555]. The second is instrumental: a SPAD avalanche does not directly preserve photon-number information. The output is binary over a dead-time interval, and photon-number sensitivity arises only through specialized pulse-height discrimination or through statistical inversion with an explicit detector model [2308.13603][1107.3089]. This distinction remains essential for interpreting both detector data and the broader PA literature.

Source: https://www.emergentmind.com/topics/photon-avalanche-pa