This module teaches one big idea in three steps: limits → derivatives → using derivatives.
Limits tell us what value a function is going towards near a point, even if it never reaches that point.
Derivatives are a special kind of limit. They tell us how fast a function is changing. This is the same as the slope of its graph.
Applications use derivatives for real work: tangent and normal lines, straight-line approximation (linearization) and the bending of a curve (concavity).
What you will be able to do
Find limits by substitution, factorising, rationalising, the sandwich theorem, and "divide by the highest power".
Check whether a function is continuous at a point, and find where it is not.
Differentiate from first principle and with the standard rules, including the chain rule and implicit differentiation.
Write the equation of a tangent line and a normal line.
Find an approximate value of a function near a point using linearization.
Use the second derivative to find where a curve bends up or down.
How to use these notes
Every topic has the same shape: idea → rule → worked example → common mistakes.
Try each worked example yourself first. Cover the answer, solve it, and then compare.
Green boxes hold key ideas. Amber boxes marked Watch out show the mistakes that make you lose marks.
The last page is a one-page formula sheet. Use it for revision before the exam.
Where each topic sits
Part
Topics
Limits
Limits, limit laws, techniques for 0/0, sandwich theorem, limits at infinity
Continuity
Continuity at a point, one-sided limits, where a function breaks
Derivatives
First principle, rules, higher orders, rate of change, chain rule, implicit differentiation
Applications
Tangent and normal lines, linearization, concavity
Before you start. You should know how to factorise quadratics, the formula (a−b)(a+b)=a2−b2, laws of indices (e.g. x−2=1/x2, x=x1/2), and the values of sin, cos, tan at 0,π/6,π/4,π/2.
Page 2 of 19
Limits of function values
The idea
A limit answers this question: "As x gets very close to c, what value is f(x) getting close to?"
If f(x) comes as close as we like to a number L when x is close enough to c (but x=c), we write
x→climf(x)=L
and say "f approaches L as x approaches c".
Key idea. A limit is about what happens nearc, not atc. The function need not even be defined at c. That is why limits help us when substitution gives the meaningless 0/0.
Example: a hole in the graph
Find x→1limx−1x2−1.
If we put x=1, we get 0/0, so the function is not defined at 1. For x=1 we can cancel the common factor:
x−1x2−1=x−1(x+1)(x−1)=x+1(x=1)
So near x=1 the function behaves exactly like x+1, and
x→1limx−1x2−1=1+1=2
The graph is the line y=x+1 with one hole at (1,2). The limit is the height of that hole.
Two simple limits to remember
Identity functionf(x)=x: as x→c, the output is x itself, so
Constant functionf(x)=k: the output never changes, so
x→climk=k
When a limit does not exist
Consider the unit step function
u(x)={0,1,x<0x≥0
When x comes to 0 from the left (x<0), u(x) stays at 0.
When x comes to 0 from the right (x>0), u(x) stays at 1.
There is no single number that u(x) goes to, so x→0limu(x)does not exist. The graph jumps.
Rule of thumb. A limit exists only when the function goes to the same value from both sides.
Page 3 of 19
Limit laws
Suppose limx→cf(x)=L and limx→cg(x)=M, where L, M, c, k are real numbers. Then:
Rule
Statement
Sum
lim[f(x)+g(x)]=L+M
Difference
lim[f(x)−g(x)]=L−M
Constant multiple
lim[kf(x)]=kL
Product
lim[f(x)g(x)]=L⋅M
Quotient
limg(x)f(x)=ML, provided M=0
Power
lim[f(x)]n=Ln
Root
limnf(x)=nL (for even n, need L≥0)
In simple words: you can take the limit inside +,−,×,÷, powers and roots, as long as each part has its own limit (and you never divide by a limit equal to 0).
Watch out. The root and power rules work on the whole expression inside. The root of a sum is not the sum of the roots:
16+3=16+3. First find the limit of what is inside, then take the root.
Polynomials and rational functions: just put the value
After cancelling one factor, the numerator goes to 7 and the denominator goes to 0. The expression becomes very large (towards +∞ from the right of 5 and towards −∞ from the left), so the limit does not exist.
Watch out. First cancel and remove the 0 from the denominator, and only then decide. Here a 0 is still left in the denominator, so there is no limit. Compare this with question 14, where the 0 was removed.
If a function is stuck between two other functions, and both of them go to the same value, then the function in the middle must also go to that value.
Sandwich theorem. Suppose g(x)≤f(x)≤h(x) for all x near c (except possibly at c itself). If x→climg(x)=x→climh(x)=L, then x→climf(x)=L.
Use it when f is difficult (for example it has sinx1, which keeps oscillating), but you can put simple functions below and above it.
Example 1: bounds are given
A function u satisfies
1−4x2≤u(x)≤1+2x2(x=0).
Find limx→0u(x).
x→0lim(1−4x2)=1,x→0lim(1+2x2)=1
Both bounds tend to 1, so x→0limu(x)=1.
Example 2: find the bounds yourself
Find x→0limx2sinx21.
We cannot substitute, because sinx21 is not defined at 0. But the sine of any number lies between −1 and 1:
−1≤sinx21≤1
Multiply every part by x2 (which is ≥0, so the inequality signs do not change):
−x2≤x2sinx21≤x2
Since limx→0(−x2)=0 and limx→0x2=0:
x→0limx2sinx21=0
Steps for "find the bounds yourself" questions. (1) Start with a known bound like −1≤sin(⋅)≤1 or −1≤cos(⋅)≤1. (2) Multiply by the term outside. (3) Take the limit of both ends. (4) If both limits are equal, you are done.
Note on the source. The handwritten page used this for 2−2cosx2sinx. That function does not satisfy these bounds (it becomes very large near 0), so it was most likely a copying mistake. The sandwich argument above is correct for any u(x) that really lies between the two bounds.
Page 7 of 19
Limits at infinity
The idea
x→∞limf(x) asks: what value does f(x) come close to when x becomes very, very large?
You need only one fact:
x→∞limx1=0,x→∞limx21=0,and in generalx→∞limxnk=0(n>0)
A constant divided by a very large number is almost zero.
Method for rational functions
Divide the top and the bottom by the highest power of x in the denominator. Then every remaining fraction like x2 or x21 goes to 0.
Check Example 1: both degrees are 2, leading coefficients 3 and 2 → 23 ✓.
Watch out.∞ is not a number. Writing ∞2=0 is a short form, but the correct statement is x2→0 as x→∞.
Page 8 of 19
Continuity
The idea
A function is continuous at a point if you can draw its graph through that point without lifting your pen. There is no jump, no hole and no break.
Definition
f is continuous at x=a if all three hold:
f(a) exists (the function is defined at a);
x→alimf(x) exists;
x→alimf(x)=f(a) (the limit equals the actual value).
If any one fails, f is discontinuous at a.
Example 1: f(x)=x2 at x=3
f(3)=9 and limx→3x2=9, so the limit equals the value. Continuous at 3.
Example 2: f(x)=x1 at x=1 and at x=0
At x=1: f(1)=1 and limx→1x1=1, so it is continuous.
At x=0: f(0)=01 is not defined, so condition 1 fails. Not continuous at 0.
A continuous function
A function is continuous (with no "at a point") if it is continuous at every point of its domain.
f(x)=x2 is continuous everywhere.
f(x)=x1 is continuous at every point where it is defined. It has a break at x=0, which is not in its domain.
Where is a function continuous?
Polynomials, sin, cos, ex and ∣x∣ are continuous everywhere. Sums, products and compositions of continuous functions are continuous. Quotients are continuous except where the denominator is 0. So the main work is: find where the denominator is 0.
Practice
1.y=x−21−3x. Denominator x−2=0 at x=2. Discontinuous only at x=2.
2.y=x2−4x+3x+1. Solve x2−4x+3=(x−1)(x−3)=0 → x=1,3. Continuous except at x=1 and x=3.
3.y=∣x+1∣+sinx. Both parts are continuous everywhere. Continuous at every point.
4.y=∣x∣+11−2x2. The denominator ∣x∣+1≥1 is never 0. Continuous at every point.
5.y=xcosx. Denominator is 0 at x=0. Continuous except at x=0.
6.y=tan2πx. Since tanθ=cosθsinθ, it breaks wherever cos2πx=0, i.e. when 2πx is an odd multiple of 2π. That happens at x=±1,±3,±5,…Continuous at every point except the odd integers.
Page 9 of 19
One-sided limits and one-sided continuity
Notation
x→a−limf(x) is the left-hand limit: x comes to a from values less thana.
x→a+limf(x) is the right-hand limit: x comes to a from values greater thana.
Key idea.x→alimf(x) exists if and only if both one-sided limits exist and are equal.
(This is exactly why the step function from page 2 has no limit at 0: its left limit is 0 and its right limit is 1.)
One-sided continuity
Type
Condition at x=a
Continuous from the left
x→a−limf(x)=f(a)
Continuous from the right
x→a+limf(x)=f(a)
f is continuous at a exactly when it is continuous from both sides.
How to check in 3 steps
Take the formula that is valid on the side from which x is coming, and find its limit.
Find f(a) using the formula that includes a (check the ≤ and ≥ signs carefully).
Compare the two.
Example 1. Left-continuity at x=2
f(x)={x2+3,5,x<2x≥2
Left limit: use x2+3 → 22+3=7.
Value: f(2)=5 (since x≥2 uses the second line).
7=5, so not continuous from the left at 2.
(From the right, the limit is 5=f(2), so it is continuous from the right.)
Example 2. Right-continuity at x=−1
g(x)={sinx,x2,x≤−1x>−1
Right limit: use x2 → (−1)2=1.
Value: g(−1)=sin(−1)=−sin1≈−0.84.
1=−sin1, so not continuous from the right at −1.
Example 3. Right-continuity at x=3
f(x)={3x−9,x2−6x+9,x≤3x>3
Right limit: 32−6(3)+9=0.
Value: f(3)=3(3)−9=0.
Both are equal, so continuous from the right at 3. (The left limit is also 0, so f is fully continuous at 3.)
Example 4. Left-continuity at x=1
f(x)={2x+1,3,x<1x≥1
Left limit: 2(1)+1=3.
Value: f(1)=3.
Both are equal, so continuous from the left at 1.
Page 10 of 19
The derivative
The idea
Take two points on a curve and draw a straight line through them (a secant). Now move the second point closer and closer to the first point. The secant becomes the tangent line, and its slope is the derivative.
The slope of the secant between x=a and x=a+h is the difference quotient
hf(a+h)−f(a)
When the gap h becomes 0, we get the derivative:
f′(a)=h→0limhf(a+h)−f(a)
This is the first principle (or definition) of the derivative.
Four ways to say the same thing.f′(a) is
(1) the slope of the graph of y=f(x) at x=a,
(2) the slope of the tangent line to the curve at x=a,
(3) the instantaneous rate of change of f at a,
(4) the derivative of f at a.
Method for first-principle questions
Write f(a+h) and f(a).
Write hf(a+h)−f(a) and simplify until h cancels from the top and the bottom.
Then put h→0.
You will see 0/0 before simplifying. This is normal. Cancelling removes it.
Watch out. Many students use L'Hôpital's rule here (differentiating top and bottom). This is wrong here, because you would be using dhdsinh=cosh, and that is the very result we are trying to prove. Use the standard limit instead.
6. f(x)=lnx at x=1
f′(1)=h→0limhln(1+h)−ln1=h→0limhln(1+h)=1
using the standard limit h→0limhln(1+h)=1.
Page 11 of 19
Rules of differentiation
First principle takes a lot of time. These rules help you differentiate almost anything in one line.
Basic rules
Rule
Formula
Constant
dxd(c)=0
Power
dxd(xn)=nxn−1
Constant multiple
dxd(kf(x))=kf′(x)
Sum / difference
dxd(f±g)=f′±g′
Product
dxd(fg)=fg′+gf′
Quotient
dxd(gf)=g2gf′−fg′
Memory trick for the quotient rule. "Low d-high minus high d-low, over low squared". The order on top is important: gf′−fg′. If you swap them, the sign changes.
Standard derivatives to remember
f(x)
f′(x)
xn
nxn−1
x
2x1
x1
−x21
ex
ex
lnx
x1
sinx
cosx
cosx
−sinx
tanx
sec2x
Worked examples
1. Power rule.f(x)=5x4
f′(x)=5⋅4x3=20x3
2. Product rule.f(x)=2x3sinx. Take f=2x3, g=sinx:
f′(x)=2x3cosx+sinx⋅6x2=2x3cosx+6x2sinx
3. Product rule.f(x)=excosx
f′(x)=ex(−sinx)+cosx⋅ex=ex(cosx−sinx)
4. Quotient rule.f(x)=x2ex
f′(x)=(x2)2x2⋅ex−ex⋅2x=x4xex(x−2)=x3ex(x−2)
5. Quotient rule.f(x)=x3lnx
f′(x)=(x3)2x3⋅x1−lnx⋅3x2=x6x2−3x2lnx=x41−3lnx
6. Quotient rule.f(x)=sinxx2
f′(x)=sin2xsinx⋅2x−x2⋅cosx=sin2x2xsinx−x2cosx
7. Slope of a curve. Find the slope of y=x2 at x=3.
dxdy=2x⇒dxdyx=3=6
(First principle gives the same answer: limh→0h(3+h)2−9=limh→0(6+h)=6.)
8. Slope of a curve. Find the slope of y=sinx at x=4π.
dxdy=cosx⇒cos4π=21
Page 12 of 19
Higher-order derivatives
The idea
The derivative of a function is also a function, so we can differentiate it again.
Order
Notation
1st
y′ , f′(x) , dxdy
2nd
y′′ , f′′(x) , dx2d2y
3rd
y′′′ , f′′′(x) , dx3d3y
nth
y(n) , f(n)(x) , dxndny
Differentiate step by step, using the previous answer each time.
Meaning. If f(x) is position, then f′ is velocity and f′′ is acceleration. The second derivative also tells how a curve bends (page 18).
Examples
1. Polynomial, 4th derivative.y=3x5
y′=15x4,y′′=60x3,y′′′=180x2,y(4)=360x
2. Reciprocal.f(x)=x1=x−1
f′(x)=−x21,f′′(x)=x32
3. Third derivative.f(x)=x5−4x3+x
f′=5x4−12x2+1,f′′=20x3−24x,f′′′=60x2−24
4. All derivatives.y=2x4−23x2−x
y′=2x3−3x−1,y′′=6x2−3,y′′′=12x,y(4)=12,y(5)=0
A polynomial of degree n becomes a constant after n derivatives, and becomes 0 after n+1 derivatives.
Here dxdx2+1=2x2+12x=x2+1x comes from the chain rule (page 14).
Page 13 of 19
Instantaneous rate of change
The idea
The average rate of change over an interval is "change in output ÷ change in input". The instantaneous rate of change at one moment is what we get when the interval becomes very, very small. This is exactly the derivative.
Steps. (1) Differentiate. (2) Put the given value. (3) Write the units: units of the output per unit of the input.
Examples
1. Rate of change of f(x)=x2 at x=5:
f′(x)=2x⇒f′(5)=10
2. Rate of change of f(x)=4x3−x2+2x−7 at x=1:
f′(x)=12x2−2x+2⇒f′(1)=12−2+2=12
3. Velocity. The position of a car is s(t)=4t2 metres at time t seconds. Its velocity at t=5 s:
s′(t)=8t⇒s′(5)=40m/s
4. Stress in a material.σ(t)=100t−5t2, with t in seconds. Rate of change of stress at t=8 s:
σ′(t)=100−10t⇒σ′(8)=100−80=20
in units of σ per second (for example Pa/s if σ is measured in pascals).
The sign tells the story. A positive rate means the quantity is increasing at that moment. A negative rate means it is decreasing. Zero means it is not changing at that moment.
Page 14 of 19
The chain rule
The idea
When one function is inside another function, like sin(4x3), differentiate from the outside to the inside:
Chain rule. If y=f(u) and u=u(x), then dxdy=dudy⋅dxdu.
In simple words: derivative of the outside function (keep the inside as it is) × derivative of the inside function.
Method
Identify the inner function u and the outer function.
Differentiate the outer function, and keep u as it is.
Multiply by dxdu.
Examples
1.y=(3x+2)3. Outer: u3. Inner: u=3x+2.
dxdy=3(3x+2)2⋅3=9(3x+2)2
2.y=sin(4x3). Outer: sinu. Inner: u=4x3.
dxdy=cos(4x3)⋅12x2=12x2cos(4x3)
3.y=esinx. Outer: eu. Inner: u=sinx.
dxdy=esinx⋅cosx=cosxesinx
4.y=5x2+3x. Outer: u. Inner: u=5x2+3x.
dxdy=25x2+3x1⋅(10x+3)=25x2+3x10x+3
5.y=tan(2x+1).
dxdy=sec2(2x+1)⋅2=2sec2(2x+1)
Chain rule combined with other rules
6.y=xsin4x+xcos−2x. Use the product rule on each term and the chain rule inside.
The numerator of the inner derivative becomes simple using sin2θ+cos2θ=1:
(1+cosθ)cosθ+sin2θ=cosθ+cos2θ+sin2θ=1+cosθ
so the inner derivative is (1+cosθ)21+cosθ=1+cosθ1 and
dθdf=(1+cosθ)22sinθ
8.y=xtan(2x)+7. Use the product rule, and the chain rule for tan(2x).
dxdy=xsec2(2x)⋅dxd(2x)+tan(2x)⋅1+0
with dxd(2x)=x1, so
dxdy=xsec2(2x)+tan(2x)
Page 15 of 19
Implicit differentiation
The idea
Sometimes y is not given as "y=…". For example, x2+y2=r2 connects x and y without giving y separately. To find dxdy you need not solve for y. Differentiate both sides with respect to x, and treat y as a function of x.
The only new point
Whenever you differentiate a term having y, the chain rule gives an extra factor dxdy:
dxd(y2)=2ydxdy,dxd(siny)=cosydxdy
and a product like xy needs the product rule:
dxd(xy)=xdxdy+y
Method
Differentiate both sides with respect to x.
Take all terms having dxdy to one side, and all other terms to the other side.
7.x2(x−y)2=x2−y2. Use the product rule on the left side and the chain rule for (x−y)2.
2x(x−y)2+2x2(x−y)(1−dxdy)=2x−2ydxdy
Collect the terms having dxdy:
dxdy[2y−2x2(x−y)]=2x−2x(x−y)2−2x2(x−y)
Divide the top and the bottom by 2:
dxdy=y−x2(x−y)x[1−(x−y)2−x(x−y)]
Note on the source. The handwritten page first tried x2(x−y2)=x2−y2, then cut it and continued with the (x−y)2 version. We have solved the version that was kept.
8.x3=x+3y2x−y. Use the quotient rule on the right side.
3x2=(x+3y)2(x+3y)(2−dxdy)−(2x−y)(1+3dxdy)
Expand the numerator. The 2x and the 3ydxdy terms cancel, and we are left with 7y−7xdxdy:
3x2(x+3y)2=7y−7xdxdy⇒dxdy=7x7y−3x2(x+3y)2
Watch out. In the quotient rule, subtract the whole second product, with brackets. If you miss the bracket, the signs change and the final answer becomes wrong. Students often make this mistake in problem 8.
Page 16 of 19
Tangent and normal lines
The idea
The tangent line to a curve at a point touches the curve at that point and has the same slope as the curve.
The normal line is perpendicular to the tangent at that point.
Both pass through the point (x0,y0) on the curve.
Formulas
Let m=f′(x0) be the slope of the curve (the derivative at x0).
Line
Slope
Equation
Tangent
m
y−y0=m(x−x0)
Normal
−m1
y−y0=−m1(x−x0)
For perpendicular lines the product of the slopes is −1, so the slope of the normal is the negative reciprocal.
Method
Find the point: y0=f(x0) (if only x0 is given).
Differentiate and find m=f′(x0).
Put the values in the tangent formula. Then use −m1 for the normal.
Write the answer in y=… form.
Special case. If m=0, the tangent is horizontal (y=y0) and the normal is vertical (x=x0). The formula −m1 would need division by zero, so write the vertical line directly.
Examples
1.y=x2 at (1,1). y′=2x⇒m=2.
Tangent: y−1=2(x−1)⇒y=2x−1
Normal: y−1=−21(x−1)⇒y=−2x+23
2.y=x2+3x+2 at x=1. Point: y0=1+3+2=6, so (1,6). y′=2x+3⇒m=5.
Tangent: y−6=5(x−1)⇒y=5x+1
Normal: y−6=−51(x−1)⇒y=−5x+531
3.y=ex at x=0. Point (0,1). y′=ex⇒m=1.
Tangent: y−1=1(x−0)⇒y=x+1
Normal: y−1=−1(x−0)⇒y=−x+1
4. Projectile (horizontal tangent).h(t)=−5t2+20t+25 at t=2 s.
Point: h(2)=−20+40+25=45, so (2,45). Slope: h′(t)=−10t+20⇒h′(2)=0.
Tangent: slope 0, so h=45 (horizontal; this is the highest point of the flight).
Normal: a line perpendicular to a horizontal line is vertical, so t=2.
5.y=sinx at x=4π. Point (4π,21). y′=cosx⇒m=21.
Tangent: y−21=21(x−4π)⇒y=2x−4π+1
Normal: y−21=−2(x−4π)⇒y=−2x+42π+21
When the curve is given implicitly
6. Tangent and normal to 6x2+3xy+2y2+17y−6=0 at (−1,0).
Check that the point lies on the curve: 6+0+0+0−6=0 ✓.
If we zoom in a lot on any smooth curve, it looks like a straight line, that is, its tangent line. So near a point x=a we can replace a difficult function by its tangent line, which is easy to calculate.
Linearization of f at x=a:L(x)=f(a)+f′(a)(x−a).
For x close to a: f(x)≈L(x).
This is the tangent-line equation from the last page, now used to calculate approximate values. It is the best straight-line approximation of f near a. The farther x is from a, the less accurate the approximation.
Method
Compute f(a).
Differentiate and compute f′(a).
Write L(x)=f(a)+f′(a)(x−a) and simplify.
Examples
1. Linearize f(x)=x at a=4.
f(4)=2,f′(x)=2x1⇒f′(4)=41L(x)=2+41(x−4)=4x+1
Example:4.1≈L(4.1)=44.1+1=2.025 (true value 2.0248…).
So near 25∘C, the resistance increases by about 1.22 units for every 1∘C rise.
7. Population of a city.P(t)=100e0.02t thousand people, with t in years since 2020. Linearize at t=0 and estimate the population in 2021.
P(0)=100,P′(t)=2e0.02t⇒P′(0)=2L(t)=100+2t
For 2021, t=1: L(1)=102. About 102 thousand people (the exact value is 100e0.02≈102.02 thousand, so the approximation is very close).
Page 18 of 19
Concavity
The idea
Concavity tells how a curve bends.
Concave up (like a cup, ∪): the slope is increasing as you move to the right.
Concave down (like an upside-down cup, ∩): the slope is decreasing as you move to the right.
The point where the bending changes direction is called an inflection point.
The second derivative test for concavity
Let f be twice differentiable on an interval I.
Sign of f′′(x) on I
Shape of the graph
f′′(x)>0
concave up
f′′(x)<0
concave down
Why.f′′ is the rate of change of the slope f′. If f′′>0, the slope is increasing, so the curve bends upward.
Watch out. If f′′(x)=0 (or is not defined), the point is only a possible inflection point. It is a real inflection point only if f′′changes sign there. For f(x)=x4, f′′(0)=0 but f′′ is positive on both sides, so 0 is not an inflection point.
Method
Find f′′(x).
Solve f′′(x)=0 (and note where f′′ is not defined) to get the points where the sign may change.
Check the sign of f′′ in each interval between these points (a sign chart on a number line helps here).
Write the concavity, and mark the inflection points where the sign changes.