What Are Exponential Functions?
Exponential functions are functions in which the variable sits in the exponent, written $f(x) = a^x$ where the base $a$ is a positive constant with $a > 0$ and $a \neq 1$. The base is fixed and the input $x$ moves in the exponent, which is the reverse of a power function like $x^2$ where the base moves and the exponent is fixed.
One base matters more than all the others in calculus: the natural exponential function $f(x) = e^x$, whose base is the constant $e \approx 2.7183$. It earns its place because of a single property that no other base shares exactly, and that property is the reason exponential functions run through every calculus course.
The whole family shares a fixed set of features:
Domain: all real numbers, so $a^x$ is defined for every $x$.
Range: all positive numbers, $a^x > 0$ for every $x$, so the curve never touches or crosses the $x$-axis.
Anchor point: every curve passes through $(0, 1)$, since $a^0 = 1$.
Horizontal asymptote: the line $y = 0$, which the curve approaches on one side without ever reaching it.
How Do You Differentiate An Exponential Function?
The derivative of the natural exponential function is the headline result of the whole topic: it gives itself back.
$$\frac{d}{dx}(e^x) = e^x$$
Geometrically this says something striking. The derivative is the slope of the tangent line, so at every point on $y = e^x$ the slope of the curve equals the height of the curve at that point. Where the graph is at height $7$, it is also climbing at a rate of $7$. That self-matching slope is what "its own derivative" means, and it is the fact the rest of calculus leans on.
For a general base, the derivative carries an extra constant factor:
$$\frac{d}{dx}(a^x) = a^x \ln a$$
Here is why the natural logarithm appears. Rewrite any base through $e$ using $a = e^{\ln a}$, so $a^x = e^{x \ln a}$. Differentiating with the chain rule multiplies by the derivative of the inside, $\ln a$:
$$\frac{d}{dx}(a^x) = \frac{d}{dx}\big(e^{x \ln a}\big) = e^{x \ln a} \cdot \ln a = a^x \ln a$$
When $a = e$ the factor $\ln e = 1$, so the general rule collapses back to $\frac{d}{dx}(e^x) = e^x$. The natural base is precisely the base whose extra factor is $1$.
The self-derivative fact also follows straight from the limit definition of a derivative:
$$\frac{d}{dx}(e^x) = \lim_{h \to 0} \frac{e^{x+h} - e^x}{h} = e^x \lim_{h \to 0} \frac{e^h - 1}{h} = e^x \cdot 1 = e^x$$
The number $e$ is defined so that $\lim_{h \to 0} \frac{e^h - 1}{h} = 1$, which is exactly the limit that makes this line work.
Example 1: Differentiate $y = 3^x$.
Apply the general-base rule with $a = 3$:
$$\frac{d}{dx}(3^x) = 3^x \ln 3$$
Final answer: $\dfrac{dy}{dx} = 3^x \ln 3 \approx 1.0986 \cdot 3^x$.
Example 2: Differentiate $y = e^{-x^2}$.
The exponent is now a function, so the chain rule applies: differentiate the outside $e^{\square}$ and multiply by the derivative of the inside $-x^2$, which is $-2x$.
$$\frac{d}{dx}\big(e^{-x^2}\big) = e^{-x^2} \cdot (-2x) = -2x,e^{-x^2}$$
Final answer: $\dfrac{dy}{dx} = -2x,e^{-x^2}$.
How Do You Integrate An Exponential Function?
Because $e^x$ differentiates to itself, it also integrates to itself, up to the constant of integration:
$$\int e^x , dx = e^x + C$$
Check it by differentiating the antiderivative back: $\frac{d}{dx}(e^x + C) = e^x$, which is the original integrand. For a general base, undo the $\ln a$ factor from the derivative by dividing:
$$\int a^x , dx = \frac{a^x}{\ln a} + C$$
Differentiating the right side returns $\frac{a^x \ln a}{\ln a} = a^x$, so the formula is correct. The geometric reading is the partner of the derivative: an integral is the area under the curve. A clean fact makes this vivid for the natural base, since the area beneath $y = e^x$ from far left up to a point $x = b$ equals the height there:
$$\int_{-\infty}^{b} e^x , dx = \big[e^x\big]_{-\infty}^{b} = e^b - 0 = e^b$$
So $e^x$ is the rare curve whose slope, whose height, and whose accumulated area all read the same number at any point.
Example 3: Evaluate $\int 2^x , dx$.
Use the general-base rule with $a = 2$:
$$\int 2^x , dx = \frac{2^x}{\ln 2} + C$$
Verify by differentiating back: $\frac{d}{dx}\left(\frac{2^x}{\ln 2}\right) = \frac{2^x \ln 2}{\ln 2} = 2^x$.
Final answer: $\dfrac{2^x}{\ln 2} + C \approx 1.4427 \cdot 2^x + C$.
Not every exponential integral has such a tidy answer. The Gaussian integrand $e^{-x^2}$, the bell curve behind statistics, has no antiderivative expressible with elementary functions. Its integral is defined instead through the special error function $\operatorname{erf}(x)$, and definite versions are found numerically or by series. Being honest about this matters: writing a fake closed form for $\int e^{-x^2}, dx$ is a real error, not a shortcut.
Table: Derivatives and integrals of the exponential functions.
Function | Derivative | Integral |
|---|---|---|
$e^x$ | $e^x$ | $e^x + C$ |
$a^x$ | $a^x \ln a$ | $\dfrac{a^x}{\ln a} + C$ |
$e^{kx}$ | $k,e^{kx}$ | $\dfrac{1}{k},e^{kx} + C$ |
The third row is the workhorse of growth and decay, and it also follows from the chain rule and its reverse.
What Are The Properties Of Exponential Functions?
The shape of the graph is decided by one thing: whether the base is bigger or smaller than $1$.
Growth ($a > 1$): the curve rises, and it rises faster the higher it gets, because the slope tracks the height. Examples are $2^x$ and $e^x$.
Decay ($0 < a < 1$): the curve falls toward the $x$-axis, halving repeatedly without reaching zero. Examples are $\left(\tfrac{1}{2}\right)^x$ and $e^{-x}$.
Table: How the base decides growth versus decay.
Case | Base | Behaviour as $x$ increases | Sign of $k$ in $e^{kx}$ |
|---|---|---|---|
Growth | $a > 1$ | Rises, steeper over time | $k > 0$ |
Decay | $0 < a < 1$ | Falls toward zero | $k < 0$ |
A decay curve is just a growth curve read backwards, since $\left(\tfrac{1}{2}\right)^x = 2^{-x}$. The two cases are one family reflected across the vertical axis.
What Is The Number e?
The constant $e \approx 2.7183$ is the natural base of exponential functions, an irrational number like $\pi$. It first showed up in a very practical place: compound interest. If one unit of money earns 100% interest split into $n$ equal steps across the year, the year-end total is $\left(1 + \tfrac{1}{n}\right)^n$, and squeezing the steps to be continuous drives that total to a limit:
$$e = \lim_{n \to \infty} \left(1 + \frac{1}{n}\right)^n \approx 2.71828$$
The same constant has a series form, where each term is a power of $x$ divided by that power's factorial:
$$e^x = 1 + x + \frac{x^2}{2} + \frac{x^3}{6} + \frac{x^4}{24} + \frac{x^5}{120} + \cdots$$
Setting $x = 1$ recovers $e = 1 + 1 + \tfrac{1}{2} + \tfrac{1}{6} + \tfrac{1}{24} + \cdots$. This series is also the quickest proof that $e^x$ is its own derivative: differentiate it term by term and every term shifts down onto the one before it, reproducing the same series.
How Do You Model Exponential Growth And Decay?
Whenever a quantity changes at a rate proportional to how much of it there is, it follows an exponential model. In calculus that sentence is a differential equation:
$$\frac{dy}{dt} = k y \quad \Longrightarrow \quad y = y_0 e^{kt}$$
Here $y_0$ is the starting amount at $t = 0$ and $k$ is the growth constant. You can confirm the solution by differentiating it: $\frac{dy}{dt} = y_0 k e^{kt} = k,(y_0 e^{kt}) = k y$, and at $t = 0$ it gives $y_0 e^{0} = y_0$. When $k > 0$ the quantity grows; when $k < 0$ it decays.
Two constants describe the speed of the process:
Doubling time (growth): set $2 y_0 = y_0 e^{kt}$, so $\ln 2 = kt$ and $t_{\text{double}} = \dfrac{\ln 2}{k}$.
Half-life (decay): set $\tfrac{1}{2} y_0 = y_0 e^{kt}$, so $t_{1/2} = \dfrac{\ln 2}{|k|}$.
Example 4: A culture starts at 500 cells and doubles every 3 hours. How many cells are there after 10 hours?
The doubling time gives the constant: $k = \dfrac{\ln 2}{3} \approx 0.2310$ per hour. Substitute into the model:
$$y = 500, e^{(0.2310)(10)} = 500, e^{2.3105} \approx 500 \times 10.0794$$
Final answer: about $5{,}040$ cells. The same result comes from the base-2 form $y = 500 \cdot 2^{10/3} \approx 5{,}040$, a useful cross-check.
Why Do Exponential Functions Work?
The behaviour is not a coincidence of the formula. It comes from one idea: growth feeding on itself.
Proportional change. In an exponential process the rate of change is a fixed multiple of the current amount. More money earns more interest, more cells divide into more cells, so the increments get bigger exactly as the total gets bigger.
Self-similar shape. Scaling the input by a constant multiplies the output by a constant, so the curve looks the same at every zoom level. This is why the graph has no flat stretch and no sudden corner.
Why $e$ is special. Among all bases, $e$ is the one where the proportionality constant is exactly $1$: the curve's slope equals its value with no correction factor. Every other base carries the correction $\ln a$, which is the price of not using the natural base.
The natural exponential is the fixed point of differentiation, the function calculus keeps returning to because operating on it changes nothing.
Who Discovered The Number e And Exponential Functions?
Exponential functions were used long before the constant behind them was named, and the naming came from the study of continuous growth.
The derivative and integral machinery that acts on these functions came from Isaac Newton (1643–1727, England) and Gottfried Wilhelm Leibniz (1646–1716, Germany), whose calculus gave a way to measure exactly how fast an exponential rises and how much area it sweeps.
Where Are Exponential Functions Used In The Real World?
The same $y = y_0 e^{kt}$ model, and the derivatives and integrals that go with it, run under a wide range of fields.
Biology and medicine: populations of cells and bacteria grow exponentially while resources last, and drug concentrations in the bloodstream decay exponentially, which sets dosing schedules.
Physics: radioactive material decays as $y_0 e^{-kt}$, and carbon dating reads an object's age from how much has decayed.
Finance and economics: continuously compounded interest and inflation both follow the exponential model, with $e$ built into the formula.
Engineering: the voltage across a discharging capacitor and the cooling of a hot object (Newton's law of cooling) are exponential decays solved with the same integral.
Machine learning: the sigmoid and softmax functions at the heart of neural networks are built from $e^x$, and their derivatives drive how the models learn.
One curve, differentiated and integrated the same way each time, describes cells, atoms, money, circuits, and learning machines.
What Are The Most Common Mistakes With Exponential Functions?
These four errors account for most lost marks, verified against MIT OpenCourseWare calculus notes, Purdue's MA161 differentiation lesson, and the OpenStax treatment of growth and decay.
Applying the power rule to an exponential.
Where it slips in:
A student sees $a^x$ and reaches for $\frac{d}{dx}(x^n) = n x^{n-1}$, writing something like $\frac{d}{dx}(3^x) = x \cdot 3^{x-1}$.
Don't do this:
Do not use the power rule when the variable is in the exponent. The power rule is for a variable base with a constant exponent, which is the opposite setup.
The correct way:
Use the exponential rule $\frac{d}{dx}(a^x) = a^x \ln a$. For base three, $\frac{d}{dx}(3^x) = 3^x \ln 3$.
Dropping the $\ln a$ factor.
Where it slips in:
A student remembers that $e^x$ differentiates to itself and assumes every base does the same, writing $\frac{d}{dx}(a^x) = a^x$.
Don't do this:
Do not omit the natural-log factor. Only the base $e$ has a factor of $1$; every other base carries $\ln a$.
The correct way:
Attach $\ln a$ on differentiation and divide by $\ln a$ on integration: $\frac{d}{dx}(a^x) = a^x \ln a$ and $\int a^x , dx = \frac{a^x}{\ln a} + C$.
Forgetting the constant of integration.
Where it slips in:
A student writes $\int e^x , dx = e^x$ and moves on, losing the family of antiderivatives that differ by a constant.
Don't do this:
Do not leave off the $+C$ on any indefinite integral, and do not multiply by $\ln a$ when you should divide.
The correct way:
Write $\int e^x , dx = e^x + C$ and $\int a^x , dx = \frac{a^x}{\ln a} + C$. Divide by $\ln a$, never multiply.
Confusing growth with decay through the sign of $k$.
Where it slips in:
A student sets up a decay problem, such as radioactive material, with a positive $k$, and the model grows without bound instead of fading.
Don't do this:
Do not assume $k$ is positive. A decaying quantity needs $k < 0$ (or the explicit $e^{-kt}$ form with $k > 0$).
The correct way:
Read the situation first. Growth uses $k > 0$; decay uses $k < 0$. A half-life is $\frac{\ln 2}{|k|}$, so a negative sign there is a signal, not an error.
Practice Problems On Exponential Functions
Answers follow each problem. Verify each derivative or integral by reversing the operation.
Differentiate $y = e^x + x^2$.
(Answer: $e^x + 2x$.)Differentiate $y = 5^x$.
(Answer: $5^x \ln 5$.)Differentiate $y = e^{4x}$.
(Answer: $4 e^{4x}$.)Evaluate $\int e^x , dx$.
(Answer: $e^x + C$.)Evaluate $\int 3^x , dx$.
(Answer: $\dfrac{3^x}{\ln 3} + C$.)A 200 mg sample has a half-life of 5 years. How much remains after 15 years?
(Answer: $15 \div 5 = 3$ half-lives, so $200 \times \left(\tfrac{1}{2}\right)^3 = 25$ mg.)
Where Should You Go Next After Exponential Functions?
Exponential functions sit at a crossroads between algebra and calculus, and several natural doors open from here.
Logarithmic functions. The inverse of the exponential, and the source of the $\ln a$ factor in every general-base derivative.
Derivative formulas. A reference for the differentiation rules, including the chain rule used on $e^{-x^2}$ and $e^{kx}$.
Integration formulas. The reverse operations, where the area-under-the-curve reading of $\int e^x , dx$ comes to life.
Calculus overview. How derivatives and integrals fit together across the whole subject.
If your child is meeting exponential functions for the first time, a live Bhanzu trainer builds them from the "why", the self-matching slope and the growth model, through Bhanzu's online math tutoring.
Was this article helpful?
Your feedback helps us write better content