Real quantum-dot blinking is not simple exponential (Poissonian) switching — the ON and OFF dwell times both follow a power-law distribution, P(t) ∝ t−α (typically α ≈ 1.4–1.7), a well-documented anomalous statistic first reported by Nirmal et al. (1996) and modeled by Kuno, Frantsuzov and others as diffusion-controlled tunneling to a fluctuating trap. This dot samples every dwell time directly from a truncated power law via inverse-transform sampling — no exponential is used anywhere.
P(t) ∝ t^-α, t ∈ [t_min, t_max]
t = ( a + u·(b−a) )^(1/(1-α)), a=t_min^(1-α), b=t_max^(1-α), u~U(0,1)
The ON-state cutoff t_max(ON) is tied to the same Auger-ionization rate used by the 3D companion sim:
k_ion = k0 · ρ_trap · exp(−d_shell / λ) · (R_pump / R_ref)
t_max(ON) = t_max,ref · (k_ref / k_ion) [longer cutoff ⇒ heavier tail ⇒ longer bright runs]
- Shell thickness — a thicker CdS/ZnS shell exponentially suppresses k_ion, which stretches the ON power-law's cutoff to much longer times. The distribution shape (exponent α) is unchanged — only how far the heavy tail reaches — exactly how real "giant-shell" non-blinking QDs work.
- OFF (dark, trion) dwell times keep a fixed cutoff — real detrapping kinetics are set by trap depth, not shell thickness, so growing the shell barely touches how long a blink lasts once it starts, only how rarely it starts.
- The log-log histogram below accumulates every completed dwell time. A power law is a straight line on log-log axes; a simple exponential would curve downward — watch the ON histogram's line get flatter (heavier tail) as you thicken the shell.
Real-world relevance: blinking suppression via thick shells is the actual strategy behind non-blinking QDs used as single-molecule fluorescent tags in super-resolution microscopy, where every dark period is a missed detection.