Calculus › Section G · Differential equations › Topic G.2

Exponential growth and decay

The simplest differential equation of all, rate proportional to current value, governs bacteria, radioactivity and bank balances alike. Its solution is one line long, and the same few formulas answer questions in biology, physics and finance.

The simplest model of all

"The rate of change is proportional to the current amount" describes a remarkable range of processes. Write it as an equation and you get

\frac{dy}{dt} = ky

the exponential model. A colony where doubling cell mass doubles breeding speed, an account where interest accrues on interest, a lump of ore whose decay events are proportional to the atoms present: all live here. The parameter \(k\), the proportionality constant, carries the model's entire character, its sign decides growth versus decay and its size sets the speed. This topic solves the equation once, then reads every application off the solution.

The solution, one line

The solutions of \(y' = ky\) are exactly the exponential family

y = Ce^{kt}

where \(C = y(0)\), the starting amount. Verification, as G.1 taught, is a substitution: \(y' = kCe^{kt} = k y\), identically. And there are no other solutions hiding: if \(y' = ky\), the product rule gives \(\frac{d}{dt}\big(y e^{-kt}\big) = (y' - ky)e^{-kt} = 0\), so \(y e^{-kt}\) is a constant \(C\) and \(y = Ce^{kt}\). Reading it: whatever the units, the quantity multiplies by \(e^k\) per unit time, and every fixed chunk of time multiplies it by a fixed factor. That constant-factor-per-time behaviour is the fingerprint of proportionality models.

Growth versus decay

Everything hinges on the sign of \(k\). With \(k > 0\), the curve grows, steepening as it goes, since larger sizes earn larger rates. With \(k < 0\), it decays toward zero, flattening as it approaches, since smaller sizes earn smaller rates; physicists often write the decay case as \(k = -\lambda\) with \(\lambda > 0\), keeping the sign explicit. A quick discipline to hold onto: before computing anything, read the sign of \(k\), state whether the process grows or decays, and check your final answer against that reading. Most modelling errors in this topic announce themselves to that check.

Doubling time and half-life

How long does a growing quantity take to double? Solve \(e^{kT} = 2\): \(T = \frac{\ln 2}{k}\). For decay with rate \(\lambda > 0\), the half-life is \(T = \frac{\ln 2}{\lambda} \approx \frac{0.693}{\lambda}\). Two things deserve notice. First, these times never depend on the current amount, halving from a gram takes exactly as long as halving from a tonne, which is the deep content of the constant-factor property. Second, the units work out: \(k\) is per unit time, so its reciprocal is a time. Both are the doubling-time computation from F.5, now applied to any quantity that follows the model.

Worked example: a bacterial colony

A colony starts at 200 cells and doubles every 3 hours. What is it after 10 hours? The doubling time pins the rate: \(k = \frac{\ln 2}{3} \approx 0.231\) per hour, so \(P(t) = 200e^{0.231t}\). At \(t = 10\): \(P(10) = 200e^{2.31}\), and since \(10\) hours is \(\frac{10}{3}\) doubling periods, the answer is also readable directly as \(200 \cdot 2^{10/3} = 200 \cdot 8 \cdot 2^{1/3} \approx 200 \cdot 10.08 \approx 2{,}016\), about 2,000 cells, assuming unrestricted resources the whole way. The two-computation agreement, exponent form and doubling form, is the check to imitate on every growth problem.

Worked example: carbon dating

Carbon-14 decays with a half-life of about 5,730 years, so its decay constant is \(\lambda = \frac{\ln 2}{5730} \approx 1.21 \times 10^{-4}\) per year. An artefact shows \(\frac{2}{5}\), that is 40%, of the carbon-14 activity of fresh organic material: how old is it? Solve \(e^{-\lambda t} = 0.4\) with logs: \(t = \frac{\ln 0.4}{-\lambda} = \frac{-0.916}{-1.21 \times 10^{-4}} \approx 7{,}600\) years. The model assumes that activity is proportional to carbon-14 content and that the organism began with the same carbon-14 level as living material today. The method in two steps: half-life to decay constant, then measured fraction to time with one logarithm.

Interest as a growth process

The continuous-compounding formula of F.5 is the same model wearing a bank's clothes. An account whose instantaneous return rate is \(r\) obeys \(\frac{dA}{dt} = rA\), so \(A(t) = Pe^{rt}\) for a deposit \(P\): 1,000 dollars at 6% compounded continuously becomes \(1000e^{0.9} \approx 2{,}460\) dollars after 15 years. The vocabulary carries over exactly: the doubling time \(\frac{\ln 2}{r}\) is the time to double your money, whatever the balance. Since \(100\ln 2 \approx 69.3\), the rule of 72 from F.5 is this formula rounded to a number that divides easily. One equation now spans biology, physics and finance.

Modelling honesty

The exponential model is a claim, not a truth: it asserts a constant per-unit rate. Colonies obey it while food is unlimited and stop when it runs out, which is G.4's entire subject; radioactive decay fits it extremely well for large samples, because each atom has the same chance of decaying in a given time however old it is; a fixed-rate account with continuous compounding obeys it exactly, by contract. When you see an exponential answer, ask whether the constant-rate assumption holds over the whole time range, and say so.

Common mistakes to avoid

The recurring slips: using a percentage as \(k\) directly, 5% per year is \(k = 0.05\), not \(k = 5\); putting a negative \(k\) straight into \(\frac{\ln 2}{k}\) and getting a negative half-life, when the half-life is \(\frac{\ln 2}{|k|}\); forgetting \(\ln 2\) and writing the doubling time as \(\frac{1}{k}\); solving \(e^{kt} = \text{fraction}\) without a logarithm; misreading a decay constant's sign, so a decaying answer comes out exploding; and forgetting units, since \(k\) is per-time and every answer's time unit must match the one \(k\) was quoted in. The discipline of reading sign, then scale, then units, catches the whole list.

Key takeaways

The equation \(y' = ky\) says rate proportional to size, and its entire solution family is \(y = y_0 e^{kt}\), growth for positive \(k\), decay for negative. Doubling time and half-life are \(\frac{\ln 2}{|k|}\), independent of the amount present, which is the constant-factor fingerprint. The colony and the artefact showed the two computational directions: rate from a doubling time, and time from a surviving fraction. Interest is the same model in a suit. When the constant-rate assumption fails, the model needs a brake, and the logistic equation of G.4 is the answer.

Flashcards

Flashcard

y′ = ky

Rate proportional to amount

The sign of k decides growth or decay.

Test your understanding

Question 1 of 5

What is the general solution of y′ = −4y?

Question 2 of 5

A substance decays with rate constant 0.04 per day. What is its half-life?

Question 3 of 5

A colony of 100 cells doubles every 2 hours. How many cells after 6 hours?

Question 4 of 5

A sample keeps 1/8 of its fresh carbon-14 activity (half-life 5,730 yr). How old is it?

Question 5 of 5

An account holds $1,000 at 6% compounded continuously. About how much is it worth after 15 years?

Standards: 7

← Previous topicNext topic →