Calculus › Section F · Logarithmic, exponential, and other transcendental functions › Topic F.4
Exponential functions: derivatives and integrals
The logarithm was defined first, by an integral. Its inverse is the exponential, and it turns out to be the function whose slope always equals its height.
Inverting the logarithm
F.1 built \(\ln x\) from an integral and learned it is strictly increasing, so by F.3 it has an inverse. Name that inverse the natural exponential, written \(\exp(x)\) or
The inverse identities are the definition itself: \(\ln(e^x) = x\) for every real \(x\), and \(e^{\ln x} = x\) for \(x > 0\). The graph is the logarithm's reflected in \(y = x\): it passes through \((0, 1)\), is always positive, rising and concave up, races to \(\infty\) as \(x \to \infty\) and flattens toward \(0\) far to the left. The base is the number from F.1 whose logarithm is \(1\), so \(e^1 = e\), consistent with ordinary power notation.
The exponent laws, inherited
Every exponential law is a logarithm law run through the inversion. The product law for logs becomes the addition law for exponents: \(e^{a + b} = e^a \cdot e^b\). Powers of powers: \((e^a)^b = e^{ab}\). Negatives reciprocate: \(e^{-a} = \frac{1}{e^a}\), since \(e^a \cdot e^{-a} = e^0 = 1\). Each is provable in one line by taking \(\ln\) of both sides and applying F.1's laws: for the first, \(\ln(e^a e^b) = \ln e^a + \ln e^b = a + b\). These laws are what make exponential growth models easy to split into periods and combine.
The derivative, from the inverse rule
Now the famous result, proved properly this time. Let \(y = e^x\), so inverting gives \(x = \ln y\). Differentiate that with respect to \(y\): \(\frac{dx}{dy} = \frac{1}{y}\). Flip it, which is F.3's inverse rule: \(\frac{dy}{dx} = y\). Rephrased:
The exponential's slope everywhere equals its own height. At the point \((0, 1)\) the slope is exactly \(1\), the property that singles out the base \(e\) among all exponentials. Only its constant multiples \(Ce^x\) share this property, and "rate proportional to current value" is why the exponential keeps turning up whenever things grow like populations or balances.
Worked example: chain forms
Composite exponentials all follow from the chain rule: \(\frac{d}{dx} e^{u} = e^{u}\, u'\). First: \(\frac{d}{dx} e^{3x} = 3e^{3x}\). Second, the bell curve of statistics: \(\frac{d}{dx} e^{-x^2} = -2x\,e^{-x^2}\). Check the second's sign: the bell curve is symmetric, rising to its peak at \(0\), so its derivative is indeed negative for \(x > 0\) and positive for \(x < 0\), which \(-2x\) delivers. The pattern to internalise is "the exponential keeps its shape, times the inside's derivative".
The integral side
Reading the derivative rule backwards, \(\int e^x\,dx = e^x + C\): the exponential is also its own antiderivative. With a linear inside, E.5's adjustment applies:
So \(\int e^{-2t}\,dt = -\frac{1}{2}e^{-2t} + C\). The chain form in reverse also absorbs composites, as the next worked example shows. The whole exp toolkit is three lines: same derivative, same integral, divide by the inside's slope when the inside is linear.
Worked example: substitution with an exponential
Evaluate \(\int x\,e^{x^2}\,dx\). The inner \(x^2\) has derivative \(2x\), nearly the \(x\) in the integrand: with \(u = x^2\), \(x\,dx = \frac{1}{2}\,du\), so the integral is \(\frac{1}{2}\int e^u\,du = \frac{1}{2}e^{x^2} + C\). Differentiate to verify: \(\frac{1}{2} e^{x^2} \cdot 2x = x e^{x^2}\), exact. And a definite one for the road: \(\int_0^1 e^x\,dx = \big[e^x\big]_0^1 = e - 1\). So the area under \(y = e^x\) from \(x = 0\) to \(x = 1\) is \(e - 1 \approx 1.718\); note that the lower limit contributes \(e^0 = 1\), not 0.
Why e earns the name natural
Ask what law governs a quantity whose growth rate is proportional to its current size, like a bacterial colony with more breeders breeding faster. In Section G that sentence will become the equation \(y' = ky\), and its solution is exactly \(y = Ce^{kt}\): no other family fits. The same structure, with \(k < 0\), drives radioactive decay, and it also governs the mathematics of interest that compounds ever more frequently, converging to continuous \(e^{rt}\) growth. So \(e\) is "natural" in a concrete sense: nature's proportionality-to-self rule, whenever it appears, writes its answers in base \(e\), and every other base is just \(e\) re-scaled.
Common mistakes to avoid
The classics, and most are old friends in new costumes: applying the power rule to get \(x e^{x-1}\), but the exponent varies and the base is fixed, so the power rule does not apply; dropping the chain factor and writing \(\frac{d}{dx} e^{3x} = e^{3x}\); integrating \(e^{ax}\) to \(a e^{ax}\) instead of dividing; writing the slope of \(e^x\) at \(x = 0\) as if it were about the origin, the curve passes through \((0, 1)\) with slope \(1\), not through \((0,0)\); and forgetting that the exponential of anything is positive, so equations like \(e^x = -3\) have no solution. Every one of these is caught by differentiating your last line.
Key takeaways
The exponential is the inverse of the built-from-an-integral logarithm, and the two identities \(\ln(e^x) = x\), \(e^{\ln x} = x\) hold respectively for all \(x\) and for \(x > 0\). Its laws are the log laws inverted, and its calculus is one pair: \(\frac{d}{dx} e^x = e^x\) and \(\int e^x\,dx = e^x + C\), with chain factors for composed insides. The graph passes through \((0, 1)\) with slope \(1\), and slope-equals-height is the form of proportional growth, which Section G will solve into \(Ce^{kt}\). Next: other bases, and how every one of them is secretly \(e\) wearing a different hat.
Flashcards
Flashcard
The inverse function of ln x
Test your understanding
Question 1 of 5
What is the derivative of e^x?
Question 2 of 5
What is the derivative of e^(−x²)?
Question 3 of 5
Evaluate ∫ e^(2x) dx.
Question 4 of 5
At the point where the graph of y = e^x crosses the y-axis, what is its slope?
Question 5 of 5
Evaluate ∫ from 0 to 1 of e^x dx.
Standards: 2, 6
