The natural cases d/dx[ex] = ex and d/dx[ln x] = 1/x are the simplest. For general bases, the derivatives pick up an extra ln a factor: d/dx[ax] = ax ln a, and d/dx[logax] = 1/(x ln a). Both proved via logarithmic differentiation — the technique of taking ln of both sides and differentiating implicitly.
The “extra” ln a: when a = e, ln a = 1 and the formulas collapse to the natural case. The general formula always carries this factor.
Geometric meaning of ln a: it’s the slope of y = ax at x = 0. Different bases give different “steepness” at the curve’s intersection with the y-axis.
Chain rule for linear arguments: d/dx[apx+q] = p · apx+q · ln a. For logs: d/dx[loga(px+q)] = p / ((px+q) · ln a).
Logarithmic differentiation: take natural log of both sides, simplify with log laws, differentiate implicitly. Required for things like y = xx (both base and exponent are variable).
Log laws can simplify before differentiating — e.g., ln(x²/sin x) = 2 ln x − ln(sin x) makes the derivative cleaner.
Domain reminder: ax requires a > 0 (and usually a ≠ 1); logax requires x > 0.
The four derivatives — special and general
Formula booklet — exp and log derivativesddx[ex] = ex |
ddx[ln x] = 1x |
ddx[ax] = ax ln a |
ddx[logax] = 1x ln a
The general formulas reduce to the natural ones when a = e, because ln(e) = 1. Geometrically, the “extra” ln a factor controls how steeply y = ax rises:
Both curves pass through (0, 1) — because a0 = 1 for every base. At that point, the BLUE tangent (slope ≈ 0.69) is shallower than the ORANGE tangent (slope ≈ 1.61). The slopes are exactly ln 2 and ln 5 — matching the formula d/dx[ax]|x=0 = a0 · ln a = ln a.
Why ln a appears: starting from y = ax, take ln of both sides → ln y = x ln a. Differentiate implicitly: (1/y) · dy/dx = ln a. So dy/dx = y · ln a = ax · ln a. The ln a comes from the constant in front of x.
Chain rule for compound arguments
Exponential — argument u(x)
ddx[au] = au · ln a · dudx
multiply by ln a AND by the inner derivative
Logarithm — argument u(x)
ddx[logau] = du/dxu · ln a
inner derivative in numerator; u and ln a in denominator
🧭 Recipe — differentiate any exp or log expression
Identify the function type: ex, ax, ln x, or logax. Note whether the argument is just x or something more complex.
Quote the formula-booklet derivative for the base form. Remember: general bases pick up a ln a factor.
Apply chain rule if the argument u is not just x: for exp → multiply by du/dx; for log → put du/dx in the numerator.
For products, quotients, or composite expressions, apply the relevant rule (product / quotient / chain) with the formula-booklet derivative plugged in.
For expressions like xx or (sin x)x where both base and exponent depend on x, use LOGARITHMIC DIFFERENTIATION: take ln of both sides, simplify, then differentiate implicitly.
Worked examples
WE 1
Prove d/dx[ax] = ax ln a via logarithmic differentiation
Show, using logarithmic differentiation, that the derivative of y = ax is ax · ln a, where a is a positive constant.
Step 1 — take natural log of both sidesy = a^xln y = ln(a^x) = x · ln a (power-law for logs)Step 2 — differentiate both sides with respect to x (implicit)d/dx[ln y] = d/dx[x · ln a](1/y) · (dy/dx) = ln a (ln a is a CONSTANT, not differentiated)Step 3 — solve for dy/dxdy/dx = y · ln aStep 4 — substitute back y = a^xdy/dx = a^x · ln a ✓∴ d/dx[a^x] = a^x · ln aln a is a fixed number (since a is constant), so when differentiating x · ln a it acts just like a coefficient. The whole proof rests on log laws + implicit differentiation.
WE 2
Chain rule on an exponential with NONLINEAR argument
Find dydx for y = 2x² + 1.
Step 1 — identify the inner function uu = x² + 1 ⟹ du/dx = 2xStep 2 — apply chain rule with d/dx[a^u] = a^u · ln a · du/dxdy/dx = 2^(x² + 1) · ln 2 · d/dx[x² + 1] = 2^(x² + 1) · ln 2 · 2xStep 3 — tidy up = 2x · 2^(x² + 1) · ln 2dy/dx = 2x · 2^(x² + 1) · ln 2three multiplicative factors: the original exponential 2^(x² + 1), the ln 2 constant, and the inner derivative 2x. Don’t drop any of them.
WE 3
Chain rule on a logarithm with linear argument
Find dydx for y = log3(4x + 1).
Step 1 — identify u and du/dxu = 4x + 1 ⟹ du/dx = 4Step 2 — apply chain rule with d/dx[log_a u] = (du/dx) / (u · ln a)dy/dx = 4 / ((4x + 1) · ln 3)dy/dx = 4(4x + 1) · ln 3the inner derivative (4) lives in the NUMERATOR; the inner expression (4x + 1) and ln 3 both live in the DENOMINATOR. A common mistake is to put ln 3 in the wrong place.
WE 4
Product rule with general-base exponential
Find dydx for y = x · 2x.
Step 1 — identify the two factorsu = x, u’ = 1v = 2^x, v’ = 2^x · ln 2Step 2 — apply product rule: (uv)’ = u’v + uv’dy/dx = (1)(2^x) + (x)(2^x · ln 2) = 2^x + x · 2^x · ln 2Step 3 — factor 2^x for cleaner form = 2^x · (1 + x · ln 2)dy/dx = 2^x (1 + x ln 2)factor out the common 2^x at the end if asked for “simplified” form. The natural exp case y = x · e^x would give e^x(1 + x) — much tidier because ln(e) = 1.
WE 5
Logarithmic differentiation on y = xx
Find dydx for y = xx, where x > 0.
Why neither power rule nor exp rule alone worky = x^x has VARIABLE base AND VARIABLE exponentpower rule needs constant exponent; exp rule needs constant baseStep 1 — take natural log of both sidesln y = ln(x^x) = x · ln x (power-law for logs)Step 2 — differentiate both sides implicitly with respect to xd/dx[ln y] = (1/y) · (dy/dx) (chain rule on left)d/dx[x · ln x] = 1 · ln x + x · (1/x) = ln x + 1 (product rule on right)⟹ (1/y) · (dy/dx) = ln x + 1Step 3 — solve for dy/dxdy/dx = y · (1 + ln x)Step 4 — substitute back y = x^xdy/dx = x^x · (1 + ln x)dy/dx = x^x (1 + ln x)the technique works for any “tower” expression like x^x, x^(2x), (sin x)^x — take ln of both sides, use log laws to bring the exponent down to a coefficient, then differentiate implicitly.
WE 6
Equation of tangent to y = x ln x at x = 1
Find the equation of the tangent to the curve y = x ln x at the point where x = 1.
Step 1 — find the y-coordinatey at x = 1: (1) · ln(1) = 1 · 0 = 0point of tangency: (1, 0)Step 2 — find dy/dx using product ruleu = x, u’ = 1; v = ln x, v’ = 1/xdy/dx = 1 · ln x + x · (1/x) = ln x + 1Step 3 — gradient at x = 1dy/dx |x = 1 = ln 1 + 1 = 0 + 1 = 1Step 4 — tangent y – y₀ = m(x – x₀)y – 0 = 1 · (x – 1)y = x – 1tangent: y = x − 1at x = 1 (where the curve crosses the x-axis), the tangent is y = x − 1 — a straight line of slope 1. Geometrically, the curve y = x ln x crosses the x-axis at x = 1 with gradient exactly 1.
💡 Top tips
Quote the formula booklet for ALL four core derivatives. Even d/dx[ln x] = 1/x — examiners notice the citation.
The ln a factor never disappears for general bases. Forgetting it is the single most-common error in this topic.
For logs with compound arguments, use log laws to simplify FIRST when possible: loga(x²y) = 2 logax + logay is much easier to differentiate.
Logarithmic differentiation is the right tool whenever you have a variable raised to a variable power (like xx), or a really messy product/quotient where ln converts it to a sum.
ex is self-derivative — d/dx[ex] = ex. Unique among all functions. That’s WHY e is called the “natural” base.
⚠ Common mistakes
Forgetting the ln a factor on general-base derivatives. d/dx[2x] = 2x · ln 2, NOT just 2x. (The natural-base shortcut only works when a = e.)
Confusing the power rule with the exp rule: d/dx[x3] = 3x² (power rule, fixed exponent), but d/dx[3x] = 3x · ln 3 (exp rule, fixed base).
Putting ln a in the wrong place for logs: d/dx[log5x] = 1/(x · ln 5), NOT (ln 5)/x or 1/(x · log 5).
Treating ln a as a function of x. ln a is a CONSTANT (since a is a constant), so it differentiates to 0 — not to 1/a.
Trying to apply the power rule to xx: that gives x · xx−1 = xx, which is WRONG. The actual derivative is xx(1 + ln x), found via logarithmic differentiation.
That’s the full Further Differentiation chapter — first principles, inverse functions, implicit differentiation, related rates, reciprocal trig, inverse trig, and now exp/log. Up next: Further Integration. With every derivative pattern in your toolkit, you’ll meet their integration counterparts (integration by substitution and integration by parts), and HL-only techniques like partial fractions. Every derivative becomes an integral you can reverse — and that’s the bridge to the Fundamental Theorem of Calculus.
Need help with Calculus?
Get 1-on-1 help from an IB examiner who knows exactly what Paper 1 & 2 are looking for.