Differentiating ln x and e^x
The exponential function is the one function that is its own derivative, and its inverse has the simplest possible derivative, . Together with the chain rule, these two facts let you differentiate every exponential and logarithmic expression on the syllabus. They appear on every P3 paper: inside stationary-point questions, in tangents and normals, in rates of growth and decay, and as the starting point for integrating and .
Why e^x is its own derivative
Look at any exponential curve with . It gets steeper as it climbs, and the gradient at each point turns out to be proportional to the height at that point: for some constant that depends on .
- For , the constant is about : the gradient is a bit less than the height.
- For , the constant is about : the gradient is a bit more than the height.
Somewhere between 2 and 3 there is a base for which the constant is exactly 1. That base is , and that is its definition for calculus. For , the gradient at every point equals the -coordinate. At the gradient is ; at the gradient is ; at the gradient is .
In particular .
So differentiating leaves the exponential exactly as it was and multiplies it by the derivative of the power.
Why ln x has derivative 1/x
is the inverse of : means exactly . Differentiate this with respect to :
and turn it upside down:
Geometrically, is the reflection of in the line . Reflection swaps the roles of and , so the gradient at a point becomes the reciprocal of the gradient at the mirror-image point. The gradient of at is , so the gradient of at is ; the gradient of at is , so the gradient of at is .
The curves (solid) and (dashed) are reflections of each other in . Both have gradient where they cross the axes.
In particular .
The general form is worth reading aloud: the derivative of of something is the derivative of the something, divided by the something.
The surprising case ln(kx)
This looks wrong at first, but it is right: , and is a constant, which differentiates to zero. The graphs of , and are vertical translations of one another, so they have the same gradient at every .
Use the log laws before you differentiate
The laws of logarithms turn products, quotients and powers inside a log into sums, differences and multiples outside it. Since sums are much easier to differentiate than quotients, always expand first.
Without the expansion you would need the quotient rule and the chain rule nested inside the chain rule.
does not split. There is no law for the log of a sum. must be differentiated with the chain rule as it stands: .
Also distinguish:
| Expression | Means | Derivative |
|---|---|---|
| the cube of | ||
| log of a log |
Simplifying exponentials first
Index laws do the same job for exponentials.
- , derivative .
- , derivative .
- , derivative .
- and : simplify these on sight.
The syllabus only requires , but occasionally appears. Write , so , and differentiate with the chain rule:
Equations that come out of these derivatives
Setting a derivative to zero frequently gives an equation in or . These standard moves finish the job:
- , valid only for . If there is no solution, and that is often the point of the question.
- .
- An equation in and is a quadratic in , because .
- : multiply through by to get .
See logarithms and Euler's number for more on solving them.
Worked examples
Differentiate (a) (b) (c) (d) .
Solution
(a) .
(b) The power has derivative : .
(c) Rewrite as : derivative .
(d) Write as . Chain rule:
Differentiate (a) (b) (c) , giving (b) as a single fraction.
Solution
(a) .
(b) Expand first: .
(c) Outer function is a cube: .
Find the exact coordinates of the stationary point of and determine its nature.
Solution
Set to zero and multiply by : , so , (since ), and .
. The point is .
for all , so it is a minimum.
Show that the curve has no stationary points, and state whether is increasing or decreasing.
Solution
Complete the square: for all . The denominator is also positive.
So for all : it is never zero, so there are no stationary points, and is decreasing for all .
The mass, grams, of a radioactive substance after days is .
(a) Find the rate at which the mass is decreasing when .
(b) Find the value of at which the mass is decreasing at a rate of grams per day.
Solution
(a) . At : . The mass is decreasing at grams per day (3 s.f.).
(b) days (3 s.f.).
Note the sign: "decreasing at a rate of " means .
The tangent to the curve at the point where passes through the point . Find , and the equation of this tangent.
Solution
At : and . Tangent:
It passes through :
Divide by , which is never zero: , so .
The tangent is , i.e. .
- . The power rule does not apply to exponentials. The power stays the same; multiply by its derivative: .
- . The inner derivative is missing.
- Splitting . .
- Losing solutions or inventing them. has no solution; do not write . gives only.
- Forgetting that needs . A stationary point at of a function containing is not on the curve.
- The phrase "exact" with an exponential or log answer means leave it as , , . Simplify and when they appear.
- When you multiply an equation through by or divide by , say why it is allowed (""). It costs nothing and protects the method mark.
- Context questions: "rate of decrease" is the size of a negative derivative. State the final answer with units and a sentence.
- Expanding a log with the log laws first is not just faster, it is safer; examiners routinely report quotient-rule errors on functions that should have been simplified.
- ; .
- ; .
- because is a constant.
- Expand logs with the log laws before differentiating. does not split.
- , but needs the chain rule.
- Equations from : take logs, or use for quadratics in . Reject .
- (via ).
Practice
- Differentiate (a) (b) (c) .
- Differentiate (a) (b) (c) .
- Find the equation of the tangent to at the point where .
- Find the exact coordinates of the stationary point of and show that it is a minimum.
- Find the exact coordinates of the stationary point of , for , and determine its nature.
- Find the point on at which the tangent passes through the origin.
- The curve passes through , where its gradient is . Find and , and the exact value of at which the gradient is .
- Show that the curve has no stationary points.
- The curve , where is a constant, has a stationary point at . Find , the exact -coordinate of the stationary point, and determine its nature.
Answers
-
(a) . (b) The power has derivative : . (c) Expand: , derivative .
-
(a) . (b) , so , which simplifies to . (c) .
-
At : , . Tangent: , i.e. .
-
. Then . , a minimum at .
-
, . : a minimum at .
-
At the tangent is . Through : , so , . The point is .
-
gives . , and at this is , so . Then , so .
-
with . Since , the derivative is always positive and never zero. No stationary points.
-
. At : . Then . ; at : , so a minimum at .