AP® CALCULUS AB & BC · UNIT 7

Exponential Models with Differential Equations

Connect a proportional rate to an exponential model. Find exact growth and decay constants, interpret units, and check predictions against the given data.

Topic 7.8 study notes8 worked examples · 8 practice questionsUpdated 12 November 2026
Calculus notebook with limit and integral symbols, a compass and pencil on an olive and cream study desk

1. Translate proportional change into a model

An exponential model describes a quantity whose instantaneous rate of change is proportional to its current value, with a constant proportionality coefficient. Translate that statement into a differential equation before substituting data.

dydt=ky
AP Calculus AB and BC · Topic 7.8. Learning objectives FUN-7.F and FUN-7.G, essential knowledge FUN-7.F.1–2 and FUN-7.G.1, and suggested skill 3.G connect differential equations with context, general and particular solutions, and checks of whether a model is accurate and appropriate.

The family and the initial-value model

y(t)=Aekt
y(0)=y0
y(t)=y0ekt

The coefficient A is arbitrary in the general family; the initial value determines it. For a given starting time t0, use the equivalent anchored form

y(t)=y(t0)ek(t−t0)

The exponential function is defined for every real input. A physical model often uses only t≥0, or another stated observation interval. Do not interpret backward extrapolation as a measured initial value.

Interpret the sign and units correctly

For a positive initial quantity, k>0 gives growth, k<0 gives decay, and k=0 gives a constant. The zero initial value gives the zero solution for every constant coefficient. If an abstract initial height is negative, a positive coefficient does not mean the height increases.

The coefficient has units of inverse time: for example, h−1 when time is in hours. The product kt is dimensionless. The derivative has quantity-per-time units, while the coefficient is a relative rate.

1ydydt=k

2. Derive the solution and find the coefficient

The zero solution satisfies the original equation and should be checked before dividing by the quantity. For a nonzero branch, separate and integrate:

1ydy=kdt
ln|y|=kt+C
|y|=eCekt
y=Aekt

For nonzero branches, A=±eC. Adding the checked zero solution allows any real coefficient. With a positive initial population or mass, the selected coefficient is positive. Substitution of y(0)=y0 gives A=y0.

Use exact data to find the coefficient

For positive values y1 and y2 at distinct times t1 and t2,

k=ln(y2y1)t2−t1

The ratio inside the logarithm is dimensionless. If the starting input is zero, this becomes

k=1t1ln(y1y0)

Require a nonzero time difference and a positive logarithm argument. Keep exact logarithms until the final numerical answer. A model with zero initial quantity cannot later become positive under this constant-coefficient equation.

3. Plan the calculation and interpret the time scales

A reliable exponential-model workflow
StageWhat to do
Define the quantitiesName the dependent quantity, time variable and their units. Identify the actual starting time.
Translate the rate statementUse y′=ky for proportional change in the quantity itself. For cooling, identify the temperature difference that follows an exponential law.
Calibrate the modelUse an initial value and an exact coefficient, a second observation, a half-life or a doubling time.
Write the full modelUse y(t)=y0ekt, or anchor at the actual starting time.
Answer the questionEvaluate the quantity or rate, solve for a target time, or interpret the given coefficient as requested.
Verify and interpretCheck the differential equation and supplied data. Check units, sign, time interval, assumptions and final rounding.

Half-life and doubling time

For a positive decaying quantity, the half-life is the time to multiply the quantity by 12. For a positive growing quantity, the doubling time multiplies it by 2.

H=ln2|k|(k<0)
D=ln2k(k>0)

Use ln2≈0.693 only as a final approximation, not as an exact equality. After n half-lives, the fraction remaining is (12)n; after n doublings, the multiplier is 2n.

4. Eight colorful worked examples

Example 1 · Calibrate a bacteria-growth model

A model population starts with 1000 cells and reaches 8000 cells after 3 hours. Find the exponential model and the time it reaches 50000 cells.

1 · Use the second observation without early rounding

8000=1000e3k
k=ln83=ln2

The coefficient is ln2h−1. Thus the population doubles each hour.

2 · State the exact model

P(t)=1000e(ln2)t

Time is measured in hours, and the output is a modeled cell count. The model gives P(0)=1000 and P(3)=8000 exactly.

3 · Solve for the target and check it

50=e(ln2)t
t=ln50ln2
t≈5.644h

The time is after the observation at three hours, as expected. Substituting the exact time gives the target population. The forecast assumes the same proportional rate remains appropriate.

Example 2 · Use the supplied radioactive half-life

Assume a Carbon-14 sample follows constant-rate decay with half-life 5730 years. Starting with 100 grams, how much remains after 10000 years?

1 · Use the half-life to determine the signed coefficient

k=−ln25730

With time in years, the coefficient has units yr−1.

2 · Keep the coefficient exact in the model

M(t)=100e−ln25730t

The formula gives M(0)=100 grams and M(5730)=50 grams.

3 · Evaluate with one final rounding step

M(10000)=1002−100005730
M(10000)≈29.829g

To one decimal place, approximately 29.8 grams remain. This is between the amounts after one and two half-lives.

Example 3 · Model continuous compounding

A mathematical account model starts at USD 5000, with a constant 6% annual rate compounded continuously and no deposits or withdrawals. Find the balance after 10 years.

1 · Translate the specified continuous rate

dBdt=0.06B
B(0)=5000

Time is in years, so the coefficient is 0.06yr−1.

2 · Evaluate the exponential balance

B(t)=5000e0.06t
B(10)=5000e0.6

3 · Report the amount in context

B(10)≈9110.59USD

The model balance is USD 9,110.59 to the nearest cent. Its derivative equals the stated proportional rate, and the initial value is the given principal.

Example 4 · Distinguish continuous and annual percentage change

Compare a 5% continuous relative growth rate with an amount that increases by exactly 5% each year. Both start at 1000 units.

1 · Translate the continuous-rate statement

Q(t)=1000e0.05t
Q(1)≈1051.271

The one-year multiplier is e0.05, not exactly 1.05.

2 · Write the annual-update model

Qn=1000(1.05)n
Q1=1050

The index counts completed years. A continuous exponential interpolation matching these annual values uses k=ln(1.05), rather than 0.05.

Example 5 · Use the temperature difference in a cooling model

A model object starts at 80∘C in a constant 20∘C environment. After 10 minutes its temperature is 50∘C. Find the model and the time it reaches 40∘C.

1 · Model the excess above the environment

dTdt=−a(T−20)

Here a>0. The excess temperature U=T−20 obeys U′=−aU, with U(0)=60.

T(t)=20+60e−at

2 · Use the measured temperature

30=60e−10a
a=ln210

The coefficient has units min−1. The model approaches the ambient temperature rather than zero.

3 · Solve for the requested temperature

20=60e−at
t=10ln3ln2
t≈15.850min

The temperature stays above 20∘C for every finite nonnegative time. The target is reached after the observation at ten minutes.

Example 6 · Anchor at a nonzero starting time

An exponential model has P(2)=300 and P(5)=600, with time in hours. Write the model and find its implied value at zero.

1 · Use the elapsed time between observations

k=ln(600300)5−2
k=ln23

2 · Anchor the model at the first observation

P(t)=300eln23(t−2)

The model gives both supplied observations exactly.

3 · Distinguish an inferred value from an observation

P(0)=3002−23
P(0)≈188.988

The given value at time two is not the value at time zero. This earlier value is a backward extrapolation under the stated model.

Example 7 · Interpret a proportional-rate motion model

A particle has velocity v(0)=10ms and satisfies v′=−0.2v, with time in seconds. Find its velocity, acceleration and displacement over the first five seconds.

1 · Solve for velocity and interpret the sign

v(t)=10e−t5

The velocity remains positive while decreasing. The particle keeps moving in the positive direction.

2 · Differentiate to get acceleration

v′(t)=−2e−t5

Acceleration has units ms2. Negative acceleration here means slowing down, not negative velocity.

3 · Accumulate velocity to get displacement

Δx=∫0510e−t5dt
Δx=50(1−e−1)
Δx≈31.606m

This is displacement from the starting position, not an absolute position without an initial position value.

Example 8 · Compare an instantaneous rate with an average rate

A modeled quantity satisfies Q′=0.08Q, Q(0)=100. Time is in minutes and quantity in milligrams. Compare the instantaneous rate at ten minutes with the average rate over the first ten minutes.

1 · Write and evaluate the quantity model

Q(t)=100e0.08t
Q(10)=100e0.8

2 · Evaluate the instantaneous derivative

Q′(10)=8e0.8
Q′(10)≈17.804mgmin

3 · Use a difference quotient for the average

Q(10)−Q(0)10−0=10(e0.8−1)
average rate≈12.255mgmin

The increasing slopes of this positive, concave-up exponential explain why the ending instantaneous rate is greater than the average over the full interval.

5. Six verified graphs with contextual units

These graphs use exact model coefficients. Filled dots mark the specified observations or computed landmarks. Line styles supplement color. Each plot has labeled quantity and time units; ordinary function plots use independent axis scales, so visual angles are not measurements of slope.

Match both population observations

image/svg+xml Matplotlib v3.11.2, https://matplotlib.org/ 0 1 2 3 4 5 6 Time (h) 0 20000 40000 60000 Population (cells)
The model P(t)=1000e(ln2)t passes through (0,1000) and (3,8000). The third marked point reaches 50000 cells at t=ln50ln2 hours.

Check the given half-life before forecasting

image/svg+xml Matplotlib v3.11.2, https://matplotlib.org/ 0 5000 10000 15000 20000 Time (years) 0 25 50 75 100 Mass (g)
The mass model M(t)=1002−t5730 gives 50 grams after 5730 years and about 29.829 grams after 10000 years. The time and mass axes use years and grams.

A continuous-compounding balance curve

image/svg+xml Matplotlib v3.11.2, https://matplotlib.org/ 0 3 6 9 12 Time (years) 4000 6000 8000 10000 Balance (USD)
The model B(t)=5000e0.06t starts at USD 5000 and reaches approximately USD 9110.59 after 10 years. A constant relative rate produces an increasing absolute derivative.

Two half-lives leave one quarter

image/svg+xml Matplotlib v3.11.2, https://matplotlib.org/ 0 10 20 30 40 50 60 Time (days) 0.00 0.25 0.50 0.75 1.00 Fraction remaining
The dimensionless fraction F(t)=2−t20 falls to 12 at 20 days and 14 at 40 days. Half-life multiplies the current amount, not a fixed number of units.

The excess temperature decays toward zero

image/svg+xml Matplotlib v3.11.2, https://matplotlib.org/ 0 10 20 30 40 Time (min) 20 40 60 80 Temperature (°C) Model Ambient
The temperature curve T(t)=20+602−t10 passes through (0,80) and (10,50). The horizontal dashed line is the constant ambient temperature, not the zero-temperature axis. The target 40∘C is marked at the exact computed time.

Separate the two meanings of rate

image/svg+xml Matplotlib v3.11.2, https://matplotlib.org/ 0 2 4 6 8 10 12 Time (min) 100 150 200 250 Quantity (mg) Model Average Tangent
The blue quantity curve is Q(t)=100e0.08t. The orange dashed secant joins the quantity values at zero and ten minutes. The green dash-dot tangent at ten minutes has the larger slope 8e0.8. Both slopes have units mgmin.

All modeled time domains shown here use t≥0. The initial points are included. Window edges and the drawn endpoints of the secant or tangent are display choices, not extra solution-domain boundaries. Model predictions remain conditional on the stated constant-rate assumptions.

6. Formula reference and model distinctions

Exponential model formula reference
PurposeFormulaCondition or meaning
Proportional ratedydt=kyConstant coefficient; inverse-time units.
Initial-value modely(t)=y0ektFor y(0)=y0. A general family uses an arbitrary coefficient.
Two observationsk=ln(y2y1)t2−t1Positive observations at distinct times; keep exact ratios.
Half-lifeH=ln2|k|For k<0 and a positive quantity.
Coefficient from half-lifek=−ln2HUse a positive half-life in consistent time units.
Doubling timeD=ln2kFor k>0 and a positive quantity.
Coefficient from doubling timek=ln2DUse a positive doubling time.
Target timet=ln(Yy0)kFor positive target and initial values with k≠0; check whether the result lies in the model interval.

Continuous versus discrete change

A continuous relative rate of 5% per year means k=0.05yr−1. A yearly multiplier of 1.05 instead gives a discrete annual-update model; its matching continuous coefficient is ln(1.05)yr−1. State which rate interpretation the question specifies.

Cooling is a shifted exponential model

dTdt=−a(T−Tenv)
T(t)=Tenv+(T0−Tenv)e−at

Here the environment is constant and a>0. The difference from the environment follows a decay equation. An object above the environment cools; one below it warms. The temperature itself does not generally satisfy a zero-baseline exponential-decay equation.

Check model assumptions and scope

Population growth, an idealized elimination or radioactive-decay process, and continuous-compounding models can use proportional-rate equations when their assumptions fit the situation. Constant relative rate is a modeling assumption, not a guarantee of indefinite real-world growth or perfect data fit. Resource limits motivate the later BC logistic topic.

For positive initial quantity, exponential growth is increasing and concave up. Positive exponential decay is decreasing and concave up, tending to zero without reaching it at a finite time. Constant-rate motion models require interpreting whether the modeled quantity is position, velocity or acceleration.

7. Eight practice questions with complete solutions

1. Find a population model from two values

A population model grows from 500 to 2000 in 4 years. Find the exponential model.

Show complete solution

1 · Calculate the coefficient exactly

k=ln(2000500)4
k=ln44

The coefficient has units yr−1.

2 · State and check the complete model

P(t)=500eln44t

The model gives the two supplied values exactly. Its doubling time is 2 years. The decimal coefficient 0.347 is only a rounded approximation.

2. Use the fraction remaining after repeated half-lives

A substance has half-life 20 days. How long until 25% of its initial amount remains?

Show complete solution

1 · Translate the remaining percentage into a fraction

14=(12)2

2 · Count the half-life intervals

t=2(20)=40days

Equivalently, solving 2−t20=14 gives the same exact time.

3. Interpret a continuous doubling rate

An account model doubles in 12 years with continuous compounding. Find the continuous annual rate.

Show complete solution

1 · Use the doubling condition

2=e12k
k=ln212

2 · Label the rate interpretation

k≈0.057762yr−1

The continuous annual rate is approximately 5.78%. The effective one-year percentage increase is different:

100(ek−1)%≈5.946%

4. Evaluate a specified initial-value model

If dydt=0.08y and y(0)=100, find y(10).

Show complete solution

1 · Use the given initial value

y(t)=100e0.08t

2 · Evaluate at the requested input

y(10)=100e0.8
y(10)≈222.55

The derivative is the specified proportional rate, and the initial value is 100. No physical units were specified in this question.

5. Find a coefficient from an instantaneous loss rate

A modeled mass is 40 grams at time zero and is then decreasing at 6gh. Assuming a constant proportional decay model, find the model and its half-life.

Show complete solution

1 · Use the signed derivative at the same instant

−6=40k
k=−320

The coefficient has units h−1.

2 · State the model and half-life

M(t)=40e−320t
H=20ln23
H≈4.621h

The initial derivative is −6gh, so the interpretation and sign agree.

6. Keep the actual observation times

A positive quantity has A(3)=80 and A(7)=20, with time in days. Find its constant-rate exponential model.

Show complete solution

1 · Divide by the elapsed time

k=ln(2080)7−3
k=−ln22

2 · Anchor at the first supplied observation

A(t)=80e−ln22(t−3)

The coefficient has units day−1. Substitution gives A(3)=80 and A(7)=20. The model’s value at zero would be a separate extrapolation.

7. Recognize warming in a shifted decay equation

A temperature model satisfies T′=−0.2(T−30), T(0)=10, with temperature in degrees Celsius and time in minutes. Find the model and describe its behavior.

Show complete solution

1 · Model the temperature difference

T(t)=30−20e−0.2t

2 · Interpret the sign of the temperature derivative

T′(t)=4e−0.2t>0

The temperature increases toward 30∘C. Its difference from the environment tends to zero even though the temperature is warming.

8. Test whether all data fit one exponential law

Treat these observations as exact: P(0)=100, P(1)=150, P(2)=250. Can one constant-coefficient exponential model fit all three?

Show complete solution

1 · Calibrate using the first two values

k=ln(1.5)
P(t)=100e(ln1.5)t

2 · Test the unused observation

P(2)=100(1.5)2=225

The prediction is not the stated value 250, so no single constant-rate exponential model fits all three exact observations. Approximate fitting of noisy data is a different task.

8. Common mistakes and clear written work

  • Rounding the coefficient too soon. Retain logarithms or exact ratios until the final result.
  • Losing the decay sign. A positive decaying quantity has a negative coefficient in the signed-rate convention used here.
  • Confusing continuous and annual updates. A continuous coefficient and an effective yearly percentage are different quantities.
  • Mixing time units. The coefficient’s inverse-time units must match the time in the exponent.
  • Calling the first observation the value at zero. Use the actual reference time or an anchored model.
  • Using a quantity value as a rate. Distinguish the amount, its derivative and its constant relative rate.
  • Using the initial slope for an entire changing interval. Exponential derivatives change with the quantity.
  • Applying zero-baseline decay directly to temperature. Model the difference from a constant environment.
  • Ignoring the remaining observations. A model fitted to two points should be checked against additional data.
  • Leaving the question unfinished. After finding the model, evaluate the requested amount, rate or target time and include units.

Explain what the calculation means

State the rate equation when requested, show how the coefficient is determined, write the complete model, and answer the requested question in context. Check the differential equation, supplied values and reasonableness. AP scoring requirements depend on the individual question and its scoring guidelines.

9. Frequently asked questions

What wording indicates a proportional exponential model?

A rate proportional to the current quantity, with constant proportionality coefficient, gives y′=ky. A constant absolute rate instead gives a linear model.

What are the units of the coefficient?

They are inverse time, such as h−1. The derivative has quantity-per-time units, and the product in the exponential is dimensionless.

Is a five-percent continuous rate the same as five-percent annual growth?

No. A continuous coefficient of 0.05 gives one-year multiplier e0.05. A yearly multiplier 1.05 matches the continuous coefficient ln(1.05) when time is in years.

Do half-life and doubling time depend on the initial amount?

For this constant-coefficient model and positive quantities, they depend on the coefficient rather than the starting amount. The amount is multiplied by the same factor over each equal interval.

Can I use a data point taken after time zero?

Yes. Use the elapsed time between observations and an anchored model. Do not silently interpret a later observation as the value at zero.

Why does a cooling model approach ambient temperature?

The quantity that decays exponentially is the temperature difference from the constant environment. An initially cooler object warms while that difference tends to zero.

How do I know whether the model is appropriate?

Check the supplied observations, units, signs and assumptions. A constant relative rate may be reasonable over a stated interval but need not describe unlimited long-term behavior.

10. Continue through Unit 7

Exponential models assume a constant relative rate. Logistic models add a limiting population or carrying capacity and are the next BC-only topic.

Use the differential equation to explain what is proportional, verify the calibrated model, and interpret every final value with the right units and time interval.