Differentiating e^x, a^x and ln x
- Two derivatives to memorise, then one habit.
- is the function that is its own gradient; differentiates to .
- Everything else is just the chain rule: multiply by the derivative of the inside.
The two results to memorise
and . (Here means natural log, base . It always does, in 9709.)
- being its own gradient is the whole reason the number matters. (Key Points 4.3, 4.5)
Why d/dx(ln x) = 1/x
Start from , which means . Differentiate that with respect to : . Now flip it: . So the log rule is just the rule read backwards. (textbook §4.4)
The chain-rule forms (what papers test)
A plain or almost never shows up alone; what you get is something inside:
: keep the , then multiply by the derivative of the inside. And : the derivative of the inside goes on top.
| Function | Inside derivative | Answer |
|---|---|---|
- In particular and . (Key Points 4.4, 4.6)
Common mistake
Why d/dx(ln 3x) = 1/x
By the rule, , the same as . The number out front drops out because and is a constant. Spotting this saves you a chain-rule step. (textbook Ex4D)
Use ln-laws first
Before you differentiate a complicated log, simplify it with ln-laws first. A power moves to the front. A product or quotient splits into a sum or a difference. A long quotient-rule problem turns into a one-line sum.
Worked example
Differentiate
- Bring the power down first: .
- Now it is a simple chain-rule log (inside derivative ): .
No quotient rule, no product rule needed. (textbook Worked Example 4.7)
Worked example
Differentiate : split it before you touch it
- Split the product, then drop each power: .
- Differentiate each piece (second is a chain-rule log, inside derivative ): .
The other way, a product rule inside a chain-rule log, is far more work and far more likely to go wrong. (textbook Worked Example 4.8)
And a^x: rewrite it as e to a power
For a base that is not , turn it into one: , then use the chain rule (inside derivative ).
. The is part of the exact answer, so keep it (e.g. ).
- This is rare, so one idea is enough. Underneath, everything is just . (textbook Ex4C)
Where it shows up
- Built into bigger questions, not on their own: an exp or log with the product or quotient rule (§4.0), then a stationary point with exact coordinates, or a tangent/normal.
- The same e-derivatives also turn up in implicit (§4.3) and parametric (§4.4) questions.
Your turn— tap to reveal the worked answer (9709-style: y = x² ln x, exact stationary point)
Product rule: . Zero when , i.e. :