Intuitional
← Back to blog
Math

O-Level Additional Mathematics: Differentiation — Rules, Tangents, Stationary Points and Rates of Change

By Intuitional Team12 min read

Differentiation is the engine behind some of the most marks-rich questions in O-Level Additional Mathematics — from finding the gradient of a curve at a specific point, to locating maximum and minimum values, to solving connected-rates problems involving spheres and cones. Students who learn the four rules as a coherent system, rather than isolated tricks, pick up method marks consistently even when arithmetic slips occur. This post walks through the power, chain, product and quotient rules and then shows exactly how Singapore examiners apply them.

O-Level Additional Mathematics: Differentiation — Rules, Tangents, Stationary Points and Rates of Change

Why Differentiation Is Worth Mastering Properly

In the O-Level Additional Mathematics syllabus, calculus is split into two halves: differentiation and integration. Differentiation comes first, and for good reason — it is the foundation everything else builds on. Integration is, at its core, the reverse of differentiation, which is why fluency here pays dividends across the entire calculus portion of the paper.

Differentiation earns its place on its own merits too. A typical A-Math paper will allocate 20–30 marks across questions that involve:

  • Finding the gradient of a curve at a specific point
  • Writing the equation of a tangent or normal
  • Locating stationary points and determining whether they are maximum, minimum, or points of inflection
  • Solving connected-rates problems (the classic balloon or conical flask question)
  • Showing that a function is increasing or decreasing on a given interval

Students who learn differentiation as a collection of disconnected rules to memorise tend to fall apart when examiners combine two ideas in one question — for example, asking for a stationary point and then a connected rate involving the same curve. The goal here is a single coherent framework so that any differentiation question feels like a recognisable type.

The Core Idea: What a Derivative Actually Means

The derivative dy/dx at a point tells you the instantaneous gradient of the curve at that point — in other words, the gradient of the tangent line drawn at that point. You do not need to reproduce the limit definition in an O-Level exam, but holding this geometric picture in mind matters: if you zoom in on any smooth curve at a point, it looks more and more like a straight line. The gradient of that limiting straight line is dy/dx.

Two notations appear in Singapore papers. If your curve is written as y = f(x), the derivative is written as dy/dx or f′(x). Differentiating again gives the second derivative, written as d2y/dx2 or f′′(x). Both notations appear in O-Level questions, sometimes within the same question.

The Four Rules You Must Own

1. The Power Rule

This is the foundation. For any term of the form axn:

d/dx (axn) = anxn−1

Multiply by the power, then reduce the power by one. This applies to positive, negative, and fractional powers equally:

  • d/dx (x5) = 5x4
  • d/dx (3x−2) = −6x−3
  • d/dx (4x1/2) = 2x−1/2
  • d/dx (7) = 0 — the derivative of any constant is zero

Critical habit before you start: Rewrite every term in the form axn first. Convert fractions like 1/x to x−1, and roots like √x to x1/2. If you attempt to differentiate 1/(x2+1) using the power rule directly, you will get the wrong answer — that requires the chain or quotient rule.

2. The Chain Rule

The chain rule handles functions of functions. Expressions like (3x + 1)8 or √(x2 − 5) have one function nested inside another.

d/dx [f(g(x))] = f′(g(x)) · g′(x)

In plain terms: differentiate the outer function while leaving the inner function unchanged, then multiply by the derivative of the inner function. A reliable three-step approach:

  1. Identify the inner function. Call it u.
  2. Differentiate the outer function with respect to u.
  3. Multiply by du/dx.

Example: Differentiate (3x2 + 1)5.

  • Inner function: u = 3x2 + 1, so du/dx = 6x
  • Outer function: u5, derivative with respect to u is 5u4
  • Result: 5(3x2 + 1)4 · 6x = 30x(3x2 + 1)4

The chain rule extends to every function type in the A-Math syllabus:

  • d/dx [sin(f(x))] = f′(x) cos(f(x))
  • d/dx [ef(x)] = f′(x) ef(x)
  • d/dx [ln(f(x))] = f′(x) / f(x)

3. The Product Rule

When two expressions are multiplied together — such as x2 sin x or (x + 3)(2x − 1)4 — you use the product rule.

d/dx (uv) = u′v + uv′

Label the two factors u and v, differentiate each separately, then add the two cross-terms.

Example: Differentiate x3 ln x.

  • u = x3, so u′ = 3x2
  • v = ln x, so v′ = 1/x
  • Result: 3x2 · ln x + x3 · (1/x) = 3x2 ln x + x2 = x2(3 ln x + 1)

4. The Quotient Rule

When one expression is divided by another, the quotient rule applies.

d/dx (u/v) = (u′v − uv′) / v2

The order is fixed: top derivative × bottom, minus top × bottom derivative, all over bottom squared. The minus sign means swapping the subtraction gives a different — wrong — answer.

Example: Differentiate (2x + 1) / (x2 − 3).

  • u = 2x + 1, so u′ = 2
  • v = x2 − 3, so v′ = 2x
  • Numerator: 2(x2 − 3) − (2x + 1)(2x) = 2x2 − 6 − 4x2 − 2x = −2x2 − 2x − 6
  • Result: −2(x2 + x + 3) / (x2 − 3)2

Standard Derivatives to Know

Beyond the power rule, the A-Math syllabus expects these results to be known and applied with the chain rule:

  • d/dx (sin x) = cos x
  • d/dx (cos x) = −sin x
  • d/dx (tan x) = sec2x
  • d/dx (ex) = ex
  • d/dx (ln x) = 1/x

One point that catches students out: x must be in radians for the trigonometric derivatives above to hold. This is always the case in A-Math calculus questions, but it is worth making a habit of noting it.

The Main Question Types and How to Approach Each

Type 1 — Gradient, Tangent, and Normal

If a question gives you a curve and a point, and asks for the tangent or normal, the method is always the same four steps:

  1. Differentiate to find dy/dx.
  2. Substitute the x-coordinate of the given point to find the gradient of the tangent, m.
  3. For the tangent: the line has gradient m and passes through the given point. Use y − y1 = m(x − x1).
  4. For the normal: the gradient is −1/m (perpendicular condition: m1 × m2 = −1). Then use y − y1 = (−1/m)(x − x1).

Worked example: The curve C is defined by y = x3 − 4x + 1. Find the equation of the normal to C at the point where x = 2.

  1. dy/dx = 3x2 − 4. At x = 2: dy/dx = 12 − 4 = 8.
  2. At x = 2: y = 8 − 8 + 1 = 1. The point is (2, 1).
  3. Gradient of normal = −1/8.
  4. Equation of normal: y − 1 = −(1/8)(x − 2), which gives 8y − 8 = −x + 2, or x + 8y = 10.

Type 2 — Stationary Points and Their Nature

A stationary point is a point on the curve where dy/dx = 0 — the tangent is horizontal. There are three types:

  • Maximum point: the curve rises to the peak, then falls.
  • Minimum point: the curve falls to a trough, then rises.
  • Stationary point of inflection: dy/dx = 0 but the curve does not turn — it changes concavity instead.

The standard method:

  1. Set dy/dx = 0 and solve for x.
  2. Find the corresponding y-values by substituting back into the original equation.
  3. Use the second derivative test: compute d2y/dx2 and substitute each x-value.
    • If d2y/dx2 < 0: maximum point (curve is concave downward).
    • If d2y/dx2 > 0: minimum point (curve is concave upward).
    • If d2y/dx2 = 0: the test is inconclusive — use a sign change of dy/dx instead.

Worked example: Find and classify the stationary points of y = 2x3 − 9x2 + 12x − 3.

  1. dy/dx = 6x2 − 18x + 12 = 6(x2 − 3x + 2) = 6(x − 1)(x − 2).
  2. Setting dy/dx = 0: x = 1 or x = 2.
  3. At x = 1: y = 2 − 9 + 12 − 3 = 2. Point: (1, 2).
  4. At x = 2: y = 16 − 36 + 24 − 3 = 1. Point: (2, 1).
  5. d2y/dx2 = 12x − 18.
    • At x = 1: d2y/dx2 = −6 < 0 → maximum point at (1, 2).
    • At x = 2: d2y/dx2 = 6 > 0 → minimum point at (2, 1).

Type 3 — Increasing and Decreasing Functions

A function is increasing on an interval when dy/dx > 0 throughout that interval, and decreasing when dy/dx < 0. Questions often ask: “Find the range of values of x for which f(x) is decreasing.”

Method: find dy/dx, then solve the inequality dy/dx < 0 (or > 0 as required). Using the example above, dy/dx = 6(x − 1)(x − 2) < 0 when 1 < x < 2. So the function is decreasing on the interval 1 < x < 2, which matches the local maximum and minimum we identified.

Type 4 — Connected Rates of Change

This question type describes a physical situation and asks how fast one quantity changes, given information about the rate at which another related quantity changes. The tool is the chain rule written in rate form:

dA/dt = (dA/dB) × (dB/dt)

The typical setup: a geometric relationship exists between two quantities (for example, the volume and radius of a sphere), and you are told the rate at which one changes. You must find the rate at which the other changes.

Worked example: A spherical balloon is inflated so that its volume increases at a constant rate of 50 cm3/s. Find the rate of increase of the radius when the radius is 5 cm.

  1. Geometric formula: V = (4/3)πr3.
  2. Differentiate with respect to r: dV/dr = 4πr2.
  3. We want dr/dt. Apply the chain rule: dr/dt = (dr/dV) × (dV/dt) = (1 / 4πr2) × 50.
  4. At r = 5: dr/dt = 50 / (4π × 25) = 50 / (100π) = 1/(2π) cm/s.

Notice the structure is always the same: write the geometric formula → differentiate to get dV/dr → take the reciprocal to get dr/dV → multiply by the given rate. Write the chain rule symbolically before substituting any numbers — this prevents setting it up backwards.

Exam Traps That Cost Students Marks

Trap 1: Forgetting the Chain Rule Multiplier

This is the single most common differentiation error. Students write d/dx [(3x+1)4] = 4(3x+1)3 and forget to multiply by 3 — the derivative of the inner function. Before writing down any derivative involving a bracketed expression, ask: is there an inner function? If yes, the final step is always to multiply by its derivative.

Trap 2: Using the Tangent Gradient for the Normal

A tangent has the same gradient as the curve at that point. A normal is perpendicular — its gradient is −1/m. Under exam pressure, students copy the tangent gradient into the normal equation. Labelling “gradient of tangent = m” and “gradient of normal = −1/m” explicitly before writing either equation is a simple guard against this.

Special cases: if dy/dx = 0 (horizontal tangent), the normal is vertical and its equation is x = constant. If dy/dx is undefined (vertical tangent), the normal is horizontal: y = constant. These appear in harder questions.

Trap 3: Concluding Nature When d2y/dx2 = 0

If the second derivative equals zero at a stationary point, the second derivative test gives no information. You must then check the sign of dy/dx on either side of the stationary point. A frequent error is automatically writing “point of inflection” whenever d2y/dx2 = 0 — this is not always true. Check the first derivative sign change to be certain.

Trap 4: Setting Up the Rate Equation Backwards

In connected rates, students sometimes write dV/dr instead of dr/dV when they need the latter. A simple check: write out the chain rule with all three quantities labelled before substituting. Confirm that the intermediate quantities cancel correctly — dA/dt = (dA/dB) × (dB/dt) works because the dB terms cancel symbolically. If they do not cancel in your setup, the chain is wrong.

Trap 5: Not Simplifying dy/dx Before Setting It to Zero

When finding stationary points, students sometimes attempt to solve a complicated unsimplified expression for zero. Factor out common terms from dy/dx first. An expression like 6x2 − 18x + 12 becomes far easier to solve once you extract the common factor of 6 and then factorise the quadratic. The factorised form also makes it immediately clear how many stationary points exist.

Trap 6: Applying Only One Rule When Two Are Needed

An expression like x · sin(x2) needs both the product rule (because x and sin(x2) are multiplied) and the chain rule (to differentiate sin(x2)). Before differentiating any expression, identify all the rules required. Writing u = and v = labels before starting, and identifying the inner function before applying the chain rule, prevents mid-working confusion.

A Systematic Method for Every Differentiation Question

Run through this checklist before touching any differentiation question:

  1. Identify the structure: Sum of terms (power rule, term by term)? Composition (chain rule)? Product (product rule)? Quotient (quotient rule)? Many expressions need more than one rule.
  2. Rewrite if needed: Convert surds and fractions to index notation. This single step prevents a large category of errors.
  3. Label clearly: For product and quotient rules, explicitly write u =, v =, u′ =, v′ = before substituting into the formula. These labels earn method marks even if the final arithmetic is wrong.
  4. Simplify dy/dx fully before moving to the next part of the question. A factorised dy/dx is far easier to set to zero or to substitute into.
  5. Answer the question asked: If the question asks for the equation of a tangent, write it in the exact form the question specifies — often “ax + by + c = 0” or “y = mx + c”. Leaving it in point-gradient form loses the final mark.

How to Practise Effectively

Differentiation rewards methodical repetition more than creative problem-solving. Here is a progression that works:

  1. Fluency drills first. Spend a focused period doing nothing but differentiating expressions — no applications yet. Cover every combination: polynomials, negative and fractional powers, trigonometric functions, exponentials, logarithms, and nested versions of all of these. The goal is that (sin x + 1)3 immediately triggers “3(sin x + 1)2 × cos x” without deliberate thought.
  2. Applications in isolation. Once fluency is solid, practise each application type separately: a set of tangent-and-normal questions, then a set of stationary-point questions, then a set of connected-rates questions. Mixing types too early before the mechanics are automatic creates confusion about which tool to use.
  3. Mixed questions from past papers. The final stage is working from Singapore O-Level A-Math past papers. Before starting each question, note in the margin which rules it requires — “chain + stationary” or “product + normal”. This trains question-type recognition, which is the skill that earns marks under time pressure.
  4. Time connected-rates questions specifically. These are reliably worth 6–8 marks and reliably take longer than expected. Practise them under time pressure until identifying the geometric formula and setting up the chain is automatic.

One final point worth noting: differentiation and integration are inverses. If you differentiate a polynomial expression and then integrate your result, you recover the original expression (up to a constant). Building this as a checking habit early will catch errors before they cost marks — and it reinforces the deeper understanding that these two operations are two sides of the same coin.

#O-Level#Additional Mathematics#Differentiation#Chain Rule#Stationary Points#Rates of Change#Tangent and Normal#A-Math