Exponential Decay

Functions & Advanced Algebra

Exponential decay occurs when a quantity decreases by a constant percentage rate over equal time intervals, shrinking rapidly at first then more slowly.

Formula

A(t) = A_0 \cdot b^t, 0 < b < 1
Visualization

Definition

Exponential decay happens when something keeps shrinking by the same fraction each time period, decreasing quickly at first and then more and more slowly, never quite reaching zero. It follows $A(t) = A_0 \cdot b^t$ where $0 < b < 1$, or equivalently $A(t) = A_0 \cdot e^{-rt}$ where $r > 0$ is the decay rate, with half-life $T_{1/2} = \ln(2)/r$ the time to reduce to half. It solves the ODE $dA/dt = -rA$ ($r > 0$): $A(t) = A_0 e^{-rt}$; in quantum mechanics, unstable particles decay exponentially, the probability of surviving to time $t$ being $e^{-t/\tau}$ where $\tau$ is the mean lifetime, and radioactive decay follows this law exactly at the level of individual nuclei, probabilistically.

Example

A hot cup of coffee losing $10\%$ of its remaining heat each minute, starting at $80$ degrees, cools as $80, 72, 64.8, 58.3, \ldots$, each minute losing less actual heat than before. A $200$ mg drug dose decaying at $r = 0.1$ per hour, $A(t) = 200e^{-0.1t}$, drops to about $121$ mg after $5$ hours, with a half-life of $\ln(2)/0.1 \approx 6.93$ hours. Carbon-14 decay, $A(t) = A_0 e^{-\lambda t}$ with $\lambda = \ln(2)/5730$ yr$^{-1}$, is the basis of radiocarbon dating, which measures the ratio $A(t)/A_0$ to determine $t$, assuming decay follows this law and the initial ratio is known.

Key Insight

Exponential decay is the mathematical description of "gradually fading away," seen in radioactive decay, medicine in the body, and battery discharge, and its hallmark is that the same percentage (not the same amount) is lost each unit of time, fundamentally different from linear decay. This memorylessness is exact: given a nucleus has survived to time $t$, the distribution of its remaining lifetime is identical to the original distribution, the continuous-time analog of the geometric distribution in discrete probability.