Skip to content
Family Table Math
Auto

Derivatives of Exponential Functions

You’ve already met the rule ddx[ex]=ex\dfrac{d}{dx}\big[e^x\big] = e^x on the derivatives of sine, cosine, eˣ, and ln x page. This page shows where it comes from: every exponential function has a slope that is a fixed multiple of its own height, and ee is the one base where that multiple is exactly 11. From there you’ll get the derivative of any exponential, axa^x or A⋅bktA \cdot b^{kt}, and use it to find how fast populations, investments, and radioactive samples are changing.

Take f(x)=axf(x) = a^x, with a>0a \gt 0 and a≠1a \ne 1. Use the definition of the derivative and the exponent law ax+h=ax⋅aha^{x + h} = a^x \cdot a^h:

f′(x)=lim⁡h→0ax+h−axh=lim⁡h→0ax(ah−1)h=ax⋅lim⁡h→0ah−1hf'(x) = \lim_{h \to 0} \frac{a^{x + h} - a^x}{h} = \lim_{h \to 0} \frac{a^x\big(a^h - 1\big)}{h} = a^x \cdot \lim_{h \to 0} \frac{a^h - 1}{h}

The limit at the end doesn’t contain xx at all. It’s just a number that depends on the base aa. Call it kk:

f′(x)=k⋅ax,where k=lim⁡h→0ah−1hf'(x) = k \cdot a^x, \qquad \text{where } k = \lim_{h \to 0} \frac{a^h - 1}{h}

So the slope of an exponential graph at any point is a constant times its height. In other words, the ratio f′(x)f(x)\dfrac{f'(x)}{f(x)} is the same constant kk everywhere. Since f′(0)=k⋅a0=kf'(0) = k \cdot a^0 = k, the constant is also the slope of the graph where it crosses the yy-axis.

You can’t simplify ah−1h\dfrac{a^h - 1}{h} algebraically, but a calculator can estimate the limit. Use smaller and smaller values of hh:

hh0.10.10.010.010.0010.0010.00010.0001
2h−1h\dfrac{2^h - 1}{h}0.71770.71770.69560.69560.69340.69340.69320.6932
3h−1h\dfrac{3^h - 1}{h}1.16121.16121.10471.10471.09921.09921.09871.0987

(Values rounded to 4 decimal places.) So for 2x2^x, k≈0.693k \approx 0.693, and for 3x3^x, k≈1.099k \approx 1.099:

ddx[2x]≈0.693⋅2x,ddx[3x]≈1.099⋅3x\frac{d}{dx}\big[2^x\big] \approx 0.693 \cdot 2^x, \qquad \frac{d}{dx}\big[3^x\big] \approx 1.099 \cdot 3^x

For base 22 the constant is less than 11, and for base 33 it’s more than 11. Somewhere between 22 and 33 there’s a base whose constant is exactly 11. That base is the number

e≈2.718 28e \approx 2.718\,28

If you graph y=axy = a^x with a slider for aa (Desmos or GeoGebra work well) and compare the slope of the tangent with the height of the graph, they match only when a=ea = e. For f(x)=exf(x) = e^x:

f′(x)=exf'(x) = e^x

In words: the slope of the tangent at any point equals the height of the graph at that point. At (1,e)(1, e) the slope is ee; at (2,e2)(2, e^2) the slope is e2≈7.389e^2 \approx 7.389.

The curves y = 2 to the x, y = e to the x and y = 3 to the x with their tangent lines at (0, 1) −1 1 1 2 3 (0, 1) y = 3ˣ y = eˣ y = 2ˣ
At (0,1)(0, 1) the tangent slopes are about 0.6930.693 for y=2xy = 2^x, exactly 11 for y=exy = e^x, and about 1.0991.099 for y=3xy = 3^x.

Compare the constants with natural logarithms: ln⁡2≈0.6931\ln 2 \approx 0.6931 and ln⁡3≈1.0986\ln 3 \approx 1.0986. They match the table. This works for every base:

aa1.71.7222.52.5ee333.53.5
k=ln⁡ak = \ln a0.5310.5310.6930.6930.9160.916111.0991.0991.2531.253

Here’s why. Since a=eln⁡aa = e^{\ln a}, you can write ax=e(ln⁡a)xa^x = e^{(\ln a)x}. The chain rule (inside function (ln⁡a)x(\ln a)x, whose derivative is the constant ln⁡a\ln a) gives

ddx[ax]=e(ln⁡a)x⋅ln⁡a=axln⁡a\frac{d}{dx}\big[a^x\big] = e^{(\ln a)x} \cdot \ln a = a^x \ln a

This includes exe^x as a special case, because ln⁡e=1\ln e = 1.

What the graph of the derivative looks like

Section titled “What the graph of the derivative looks like”

Since f′(x)=(ln⁡a)⋅f(x)f'(x) = (\ln a) \cdot f(x), the graph of f′f' is the graph of ff stretched vertically by a factor of ln⁡a\ln a. So f′f' is also an exponential function with the same base:

  • If a>ea \gt e, then ln⁡a>1\ln a \gt 1: f′f' sits above ff (an expansion).
  • If 1<a<e1 \lt a \lt e, then 0<ln⁡a<10 \lt \ln a \lt 1: f′f' sits below ff (a compression), as for 2x2^x in the figure.
  • If 0<a<10 \lt a \lt 1, then ln⁡a<0\ln a \lt 0: f′f' is a reflection in the xx-axis as well. That makes sense, because a decaying exponential is always decreasing.
The graphs of y = 2 to the x and its derivative, y = ln 2 times 2 to the x −2 −1 1 2 1 2 3 4 5 6 (2, 4) (2, 2.77) y = 2ˣ y′ = (ln 2)·2ˣ
The derivative of y=2xy = 2^x is y=(ln⁡2)⋅2xy = (\ln 2) \cdot 2^x: the same curve, compressed vertically by a factor of ln⁡2≈0.693\ln 2 \approx 0.693.

Combine the rule with the chain rule whenever the exponent is more than just xx:

FunctionDerivative
ekxe^{kx}kekxk e^{kx}
eg(x)e^{g(x)}eg(x)⋅g′(x)e^{g(x)} \cdot g'(x)
ag(x)a^{g(x)}ag(x)ln⁡a⋅g′(x)a^{g(x)} \ln a \cdot g'(x)
A⋅bktA \cdot b^{kt}A⋅bkt⋅kln⁡bA \cdot b^{kt} \cdot k \ln b

The last line is the general growth or decay model. Notice that its derivative is (kln⁡b)(k \ln b) times the original function: the rate of change is proportional to the amount present. That’s the defining feature of exponential growth and decay. A population of 10 00010\,000 grows twice as fast (in people per year) as the same kind of population of 50005000.

Example 1: Estimating a derivative numerically

Section titled “Example 1: Estimating a derivative numerically”

Let f(x)=3xf(x) = 3^x. Use the table of values of 3h−1h\dfrac{3^h - 1}{h} above to estimate f′(2)f'(2). Then check your estimate with the rule f′(x)=axln⁡af'(x) = a^x \ln a.

Solution. From the definition,

f′(2)=lim⁡h→032+h−32h=32⋅lim⁡h→03h−1h≈9(1.0987)≈9.888f'(2) = \lim_{h \to 0} \frac{3^{2 + h} - 3^2}{h} = 3^2 \cdot \lim_{h \to 0} \frac{3^h - 1}{h} \approx 9(1.0987) \approx 9.888

Check with the rule: f′(x)=3xln⁡3f'(x) = 3^x \ln 3, so

f′(2)=9ln⁡3≈9(1.098 61)≈9.888f'(2) = 9\ln 3 \approx 9(1.098\,61) \approx 9.888

The two answers agree (to 3 decimal places). The exact value is 9ln⁡39\ln 3.

Example 2: Differentiating exponential functions

Section titled “Example 2: Differentiating exponential functions”

Find the derivative of each function.

  • (a) y=5xy = 5^x
  • (b) y=e−4xy = e^{-4x}
  • (c) y=3ex2y = 3e^{x^2}
  • (d) y=23x+1y = 2^{3x + 1}

Solution.

(a) Use ddx[ax]=axln⁡a\dfrac{d}{dx}\big[a^x\big] = a^x \ln a: dydx=5xln⁡5\quad \dfrac{dy}{dx} = 5^x \ln 5.

(b) The inside function is −4x-4x, with derivative −4-4: dydx=−4e−4x\quad \dfrac{dy}{dx} = -4e^{-4x}.

(c) The inside function is x2x^2, with derivative 2x2x:

dydx=3ex2⋅2x=6xex2\frac{dy}{dx} = 3e^{x^2} \cdot 2x = 6xe^{x^2}

(d) The inside function is 3x+13x + 1, with derivative 33:

dydx=23x+1ln⁡2⋅3=3ln⁡2⋅23x+1\frac{dy}{dx} = 2^{3x + 1} \ln 2 \cdot 3 = 3\ln 2 \cdot 2^{3x + 1}

Find the equation of the tangent line to y=2xy = 2^x at x=3x = 3. Where does the tangent line cross the xx-axis?

Solution. The point is (3,23)=(3,8)(3, 2^3) = (3, 8). The slope is

dydx=2xln⁡2⇒m=23ln⁡2=8ln⁡2≈5.545\frac{dy}{dx} = 2^x \ln 2 \quad\Rightarrow\quad m = 2^3 \ln 2 = 8\ln 2 \approx 5.545

The tangent line is

y−8=8ln⁡2 (x−3)y - 8 = 8\ln 2\,(x - 3)

It crosses the xx-axis where y=0y = 0:

−8=8ln⁡2 (x−3)⇒x−3=−1ln⁡2⇒x=3−1ln⁡2≈1.557-8 = 8\ln 2\,(x - 3) \quad\Rightarrow\quad x - 3 = -\frac{1}{\ln 2} \quad\Rightarrow\quad x = 3 - \frac{1}{\ln 2} \approx 1.557

A hospital patient receives 200200 mg of iodine-131, which has a half-life of 88 days. The amount remaining after tt days is

A(t)=200(12)t/8A(t) = 200\left(\frac{1}{2}\right)^{t/8}
  • (a) Find the rate of decay at t=0t = 0 and at t=8t = 8 days.
  • (b) Show that A′(t)A'(t) is proportional to A(t)A(t).
  • (c) When is the amount decreasing at a rate of 22 mg per day?

Solution.

(a) Use ag(t)a^{g(t)} with a=12a = \frac{1}{2} and g(t)=t8g(t) = \frac{t}{8}. Remember ln⁡12=−ln⁡2\ln \frac{1}{2} = -\ln 2:

A′(t)=200(12)t/8⋅ln⁡12⋅18=−25ln⁡2(12)t/8A'(t) = 200\left(\frac{1}{2}\right)^{t/8} \cdot \ln\frac{1}{2} \cdot \frac{1}{8} = -25\ln 2\left(\frac{1}{2}\right)^{t/8} A′(0)=−25ln⁡2≈−17.329,A′(8)=−25ln⁡2⋅12=−12.5ln⁡2≈−8.664A'(0) = -25\ln 2 \approx -17.329, \qquad A'(8) = -25\ln 2 \cdot \frac{1}{2} = -12.5\ln 2 \approx -8.664

At the start, the iodine is decaying at about 17.32917.329 mg per day. After one half-life, half as much is left, and it’s decaying half as fast: about 8.6648.664 mg per day. (The negative sign means the amount is decreasing.)

(b) From part (a),

A′(t)=ln⁡128⋅200(12)t/8=−ln⁡28A(t)≈−0.0866 A(t)A'(t) = \frac{\ln \frac{1}{2}}{8} \cdot 200\left(\frac{1}{2}\right)^{t/8} = -\frac{\ln 2}{8} A(t) \approx -0.0866\,A(t)

So at any moment, the amount decreases at about 8.66%8.66\% of the current amount per day.

(c) Set A′(t)=−2A'(t) = -2. Using part (b), −ln⁡28A(t)=−2-\dfrac{\ln 2}{8} A(t) = -2, so A(t)=16ln⁡2≈23.083A(t) = \dfrac{16}{\ln 2} \approx 23.083 mg. Now solve for tt:

200(12)t/8=16ln⁡2(12)t/8=225ln⁡2t8ln⁡12=ln⁡(225ln⁡2)take ln of both sidest=8ln⁡(225ln⁡2)ln⁡12≈24.921\begin{aligned} 200\left(\frac{1}{2}\right)^{t/8} &= \frac{16}{\ln 2} \\ \left(\frac{1}{2}\right)^{t/8} &= \frac{2}{25\ln 2} \\ \frac{t}{8}\ln\frac{1}{2} &= \ln\left(\frac{2}{25\ln 2}\right) && \text{take ln of both sides} \\ t &= \frac{8\ln\left(\frac{2}{25\ln 2}\right)}{\ln \frac{1}{2}} \approx 24.921 \end{aligned}

After about 24.92124.921 days (just under 2525 days), the iodine is decaying at 22 mg per day.

Using the power rule on an exponential. ddx[2x]\dfrac{d}{dx}\big[2^x\big] is not x⋅2x−1x \cdot 2^{x - 1}. The power rule is for a variable base and a constant exponent (x2x^2). Here the base is constant and the exponent is the variable, so you need axln⁡aa^x \ln a.

Forgetting the ln a. ddx[5x]=5xln⁡5\dfrac{d}{dx}\big[5^x\big] = 5^x \ln 5, not 5x5^x. Only base ee has no extra factor, because ln⁡e=1\ln e = 1. A quick check: the slope of 5x5^x at x=0x = 0 should be about 1.6091.609, not 11.

Forgetting the chain rule factor. ddx[e−4x]=−4e−4x\dfrac{d}{dx}\big[e^{-4x}\big] = -4e^{-4x}, not e−4xe^{-4x}. Whenever the exponent is anything other than plain xx, multiply by the derivative of the exponent.

Changing the exponent. The derivative of ex2e^{x^2} is 2xex22xe^{x^2}. The exponent stays x2x^2; it does not become 2x2x or drop by one. The 2x2x comes out in front as a factor.

Dropping the negative sign for decay. For b<1b \lt 1, ln⁡b\ln b is negative, so the derivative is negative. If your “rate of decay” comes out positive, check ln⁡12=−ln⁡2\ln \frac{1}{2} = -\ln 2. In words, you can then say “decreasing at 17.32917.329 mg per day”.

Rounding too early. Keep ln⁡2\ln 2 or the full calculator value until the last step. Rounding ln⁡2\ln 2 to 0.70.7 in Example 4 would give A′(0)≈−17.5A'(0) \approx -17.5, which is wrong to 3 decimal places.

1. (Warm-up) Differentiate each function.

  • (a) y=7xy = 7^x
  • (b) y=e5xy = e^{5x}
  • (c) y=4e−0.2xy = 4e^{-0.2x}
Solution

(a) dydx=7xln⁡7\dfrac{dy}{dx} = 7^x \ln 7

(b) dydx=5e5x\dfrac{dy}{dx} = 5e^{5x}

(c) dydx=4e−0.2x⋅(−0.2)=−0.8e−0.2x\dfrac{dy}{dx} = 4e^{-0.2x} \cdot (-0.2) = -0.8e^{-0.2x}

2. (Warm-up) Let f(x)=5xf(x) = 5^x.

  • (a) Evaluate 5h−1h\dfrac{5^h - 1}{h} for h=0.01h = 0.01, h=0.001h = 0.001 and h=0.0001h = 0.0001, to 4 decimal places.
  • (b) Use your values to estimate the constant kk in f′(x)=k⋅5xf'(x) = k \cdot 5^x, and compare it with ln⁡5\ln 5.
Solution

(a) 50.01−10.01≈1.6225\dfrac{5^{0.01} - 1}{0.01} \approx 1.6225, 50.001−10.001≈1.6107\quad \dfrac{5^{0.001} - 1}{0.001} \approx 1.6107, 50.0001−10.0001≈1.6096\quad \dfrac{5^{0.0001} - 1}{0.0001} \approx 1.6096.

(b) The values are settling down near 1.6091.609, so k≈1.609k \approx 1.609 and f′(x)≈1.609⋅5xf'(x) \approx 1.609 \cdot 5^x. This matches ln⁡5≈1.6094\ln 5 \approx 1.6094.

3. (Warm-up) Explain, without a calculator, why the slope of the tangent to y=exy = e^x at the point (2,e2)(2, e^2) is e2e^2. Then give that slope to 3 decimal places.

Solution

For f(x)=exf(x) = e^x, f′(x)=exf'(x) = e^x: the slope at any point equals the height of the graph there. At x=2x = 2 the height is e2e^2, so the slope is e2≈7.389e^2 \approx 7.389.

4. (Core) Differentiate each function.

  • (a) y=32x−1y = 3^{2x - 1}
  • (b) y=10x2y = 10^{x^2}
  • (c) y=x2e−3xy = x^2 e^{-3x}
Solution

(a) The exponent 2x−12x - 1 has derivative 22:

dydx=32x−1ln⁡3⋅2=2ln⁡3⋅32x−1\frac{dy}{dx} = 3^{2x - 1} \ln 3 \cdot 2 = 2\ln 3 \cdot 3^{2x - 1}

(b) The exponent x2x^2 has derivative 2x2x:

dydx=10x2ln⁡10⋅2x=2xln⁡10⋅10x2\frac{dy}{dx} = 10^{x^2} \ln 10 \cdot 2x = 2x\ln 10 \cdot 10^{x^2}

(c) Use the product rule, with the chain rule on e−3xe^{-3x}:

dydx=2x⋅e−3x+x2⋅(−3e−3x)=xe−3x(2−3x)\frac{dy}{dx} = 2x \cdot e^{-3x} + x^2 \cdot \big(-3e^{-3x}\big) = xe^{-3x}(2 - 3x)

5. (Core) Find the equation of the tangent line to y=6−2ex−1y = 6 - 2e^{x - 1} at x=1x = 1.

Solution

Point: y=6−2e0=4y = 6 - 2e^0 = 4, so (1,4)(1, 4).

Slope: dydx=−2ex−1\dfrac{dy}{dx} = -2e^{x - 1}, so at x=1x = 1 the slope is −2e0=−2-2e^0 = -2.

y−4=−2(x−1)⇒y=−2x+6y - 4 = -2(x - 1) \quad\Rightarrow\quad y = -2x + 6

6. (Core) The population of a town is modelled by P(t)=12 000(1.025)tP(t) = 12\,000(1.025)^t, where tt is the number of years after 2026.

  • (a) Find P′(t)P'(t).
  • (b) How fast is the population growing in 2026 and in 2036? Round to the nearest person per year.
  • (c) Why is the growth rate larger in 2036?
Solution

(a) P′(t)=12 000(1.025)tln⁡1.025P'(t) = 12\,000(1.025)^t \ln 1.025

(b) In 2026, t=0t = 0: P′(0)=12 000ln⁡1.025≈296.3P'(0) = 12\,000\ln 1.025 \approx 296.3, so about 296296 people per year.

In 2036, t=10t = 10: P′(10)=12 000(1.025)10ln⁡1.025≈379.3P'(10) = 12\,000(1.025)^{10}\ln 1.025 \approx 379.3, so about 379379 people per year.

(c) The rate of growth is proportional to the population: P′(t)=(ln⁡1.025)P(t)≈0.0247 P(t)P'(t) = (\ln 1.025)P(t) \approx 0.0247\,P(t). By 2036 the population has grown to about 15 36115\,361, so it grows by more people each year.

7. (Core) You invest $2000 at 6%6\% per year, compounded annually, so the value after tt years is A(t)=2000(1.06)tA(t) = 2000(1.06)^t dollars.

  • (a) Find the rate at which the investment is growing at t=0t = 0, to the nearest cent per year.
  • (b) When is the investment growing at $150 per year? Round to 2 decimal places.
Solution

(a) A′(t)=2000(1.06)tln⁡1.06A'(t) = 2000(1.06)^t \ln 1.06, so A′(0)=2000ln⁡1.06≈116.54A'(0) = 2000\ln 1.06 \approx 116.54. The investment is growing at about $116.54 per year at the start.

(b) Solve A′(t)=150A'(t) = 150:

2000ln⁡1.06 (1.06)t=150(1.06)t=1502000ln⁡1.06≈1.287 13t=ln⁡(1.287 13)ln⁡1.06≈4.33\begin{aligned} 2000\ln 1.06\,(1.06)^t &= 150 \\ (1.06)^t &= \frac{150}{2000\ln 1.06} \approx 1.287\,13 \\ t &= \frac{\ln(1.287\,13)}{\ln 1.06} \approx 4.33 \end{aligned}

After about 4.334.33 years, the investment is growing at $150 per year.

8. (Challenge) Let f(x)=e2x−6xf(x) = e^{2x} - 6x. Find the point where the graph has a horizontal tangent, and decide whether it’s a local maximum or minimum.

Solutionf′(x)=2e2x−6=0⇒e2x=3⇒x=ln⁡32≈0.549f'(x) = 2e^{2x} - 6 = 0 \quad\Rightarrow\quad e^{2x} = 3 \quad\Rightarrow\quad x = \frac{\ln 3}{2} \approx 0.549

The yy-coordinate is f(ln⁡32)=eln⁡3−6⋅ln⁡32=3−3ln⁡3≈−0.296f\left(\frac{\ln 3}{2}\right) = e^{\ln 3} - 6 \cdot \frac{\ln 3}{2} = 3 - 3\ln 3 \approx -0.296.

f′′(x)=4e2xf''(x) = 4e^{2x}, and f′′(ln⁡32)=4(3)=12>0f''\left(\frac{\ln 3}{2}\right) = 4(3) = 12 \gt 0, so by the second derivative test the point (ln⁡32, 3−3ln⁡3)≈(0.549,−0.296)\left(\dfrac{\ln 3}{2},\ 3 - 3\ln 3\right) \approx (0.549, -0.296) is a local minimum.

9. (Challenge) For f(x)=axf(x) = a^x, the ratio f′(x)f(x)\dfrac{f'(x)}{f(x)} is constant.

  • (a) For which base aa is f′(x)=2f(x)f'(x) = 2f(x) for every xx?
  • (b) For which base aa is f′(x)=−f(x)f'(x) = -f(x) for every xx? Describe how the graphs of ff and f′f' are related in this case.
Solution

(a) f′(x)f(x)=ln⁡a\dfrac{f'(x)}{f(x)} = \ln a, so we need ln⁡a=2\ln a = 2, which gives a=e2≈7.389a = e^2 \approx 7.389. (Check: e2xe^{2x} has derivative 2e2x2e^{2x}.)

(b) We need ln⁡a=−1\ln a = -1, so a=e−1=1e≈0.368a = e^{-1} = \dfrac{1}{e} \approx 0.368. Then f(x)=e−xf(x) = e^{-x} and f′(x)=−e−xf'(x) = -e^{-x}: the graph of f′f' is the reflection of the graph of ff in the xx-axis.