8.2 Definite Integrals and their Properties
The definite integral $\int_a^b f(x)\,dx$ adds up $f(x)\,dx$ across $[a,b]$. It is a limit of sums; the fundamental theorem of calculus evaluates it as $F(b)-F(a)$; and a handful of properties — King's rule $x\to a+b-x$, even and odd functions, periodicity — evaluate integrals whose antiderivatives are hopeless.
A water tank is filled by a pump whose flow rate changes during the morning: $r(t)$ litres per minute at time $t$. If the rate were constant, the volume pumped would be rate × time. It is not constant, so cut the morning into short intervals of length $\Delta t$; on each one the rate hardly changes, and the volume pumped is close to $r(t)\,\Delta t$. Add these up and let the intervals shrink: the sums settle down to an exact total. The same recipe gives the distance covered from a speed record, the charge passed from a current, the work done by a varying force — and the area under a curve, where $r(t)\,\Delta t$ is the area of a thin rectangle.
That limit of sums is the definite integral. The surprise — the fundamental theorem of calculus, discovered independently by Newton and Leibniz — is that it can be computed with an antiderivative from Section 8.1: $\int_a^b f(x)\,dx=F(b)-F(a)$. This section defines the definite integral, proves the fundamental theorem, and then develops the properties that make definite integrals the favourite calculus topic of JEE Main: limits of sums, the derivative of an integral with variable limits, King's rule, even, odd and periodic integrands, and integrands with $|x|$ and $[x]$. Expect two or three questions in every paper.
The Definite Integral as a Limit of Sums
Let $f$ be continuous on $[a,b]$. Divide $[a,b]$ into $n$ equal parts of width $h=\tfrac{b-a}{n}$ with division points $x_r=a+rh$ ($r=0,1,\dots,n$), and choose a point $x_r^*$ in each piece $[x_{r-1},x_r]$. The Riemann sum is
$$S_n=\sum_{r=1}^{n}f(x_r^*)\,h,$$and the definite integral of $f$ from $a$ to $b$ is its limit:
$$\int_a^b f(x)\,dx=\lim_{n\to\infty}\sum_{r=1}^{n}f(x_r^*)\,h.$$The numbers $a$ and $b$ are the lower and upper limits of integration. We also set $\int_a^a f(x)\,dx=0$ and, for $a\gt b$, $\int_a^b f(x)\,dx=-\int_b^a f(x)\,dx$.
For a continuous $f$ this limit always exists and does not depend on how the points $x_r^*$ are chosen — a fact from advanced analysis that we accept. Choosing right ends ($x_r^*=x_r$), left ends or midpoints gives the right, left and midpoint sums. If $f\ge0$, each term $f(x_r^*)h$ is the area of a thin rectangle, so the integral is the area under the curve; where $f\lt0$ the terms are negative, so in general the integral is a signed area — area above the axis minus area below it. That is why Leibniz's $\int$ is an S: $f(x)\,dx$ is one thin rectangle and $\int$ adds them up.
If $f$ is continuous on $[0,p]$, where $p$ is a positive integer, then
$$\lim_{n\to\infty}\frac1n\sum_{r=1}^{pn}f\!\left(\frac rn\right)=\int_0^p f(x)\,dx.$$The recipe: write the sum as $\tfrac1n\sum f\left(\tfrac rn\right)$; replace $\tfrac rn$ by $x$ and $\tfrac1n$ by $dx$; the limits are the values of $\tfrac rn$ at the first and last terms as $n\to\infty$.
Proof. On $[0,p]$ with $pn$ equal parts, $h=\tfrac{p}{pn}=\tfrac1n$ and the right ends are $x_r=\tfrac rn$. So the sum is exactly the right Riemann sum of Definition 8.4, and its limit is the integral. $\blacksquare$ Similarly, terms running from $r=n+1$ to $r=2n$ give $\int_1^2$, because $\tfrac rn$ runs from just above $1$ to $2$.
Let $f$, $g$ be continuous on $[a,b]$, $a\lt b$.
- If $f(x)\le g(x)$ on $[a,b]$, then $\int_a^b f(x)\,dx\le\int_a^b g(x)\,dx$. In particular $f\ge0$ gives $\int_a^b f\ge0$.
- If $m\le f(x)\le M$ on $[a,b]$, then $m(b-a)\le\int_a^b f(x)\,dx\le M(b-a)$.
- $\left|\int_a^b f(x)\,dx\right|\le\int_a^b|f(x)|\,dx$.
Proof. Each inequality holds term by term for the Riemann sums with the same points — for (2), $m\,h\le f(x_r^*)\,h\le M\,h$ summed over $r$ gives $m(b-a)\le S_n\le M(b-a)$ — and non-strict inequalities survive the limit. (3) follows from (1) applied to $-|f|\le f\le|f|$. $\blacksquare$
Use right-endpoint sums for $\int_0^2x^2\,dx=\tfrac83$. When $n$ is doubled from $10$ to $20$, the error $S_n-\tfrac83$ is
$\displaystyle\lim_{n\to\infty}\frac1n\sum_{r=1}^{n}e^{r/n}$ equals
The Fundamental Theorem of Calculus
Let the upper limit move. For $f$ continuous on $[a,b]$, the accumulation function $A(x)=\int_a^x f(t)\,dt$ is the signed area from $a$ up to $x$. (The variable inside is renamed $t$ so that $x$ can be the moving end.) How fast does $A$ grow?
If $f$ is continuous on $[a,b]$, then $A(x)=\int_a^x f(t)\,dt$ is differentiable on $[a,b]$ and
$$A'(x)=\frac{d}{dx}\int_a^x f(t)\,dt=f(x).$$In particular every continuous function has an antiderivative.
Proof. For $h\gt0$, $A(x+h)-A(x)=\int_x^{x+h}f(t)\,dt$, the area of a thin strip (additivity, Theorem 8.17 below — it follows directly from the sums). Let $m_h$ and $M_h$ be the least and greatest values of $f$ on $[x,x+h]$. By Theorem 8.12(2),
$$m_h\,h\le A(x+h)-A(x)\le M_h\,h,\qquad\text{so}\qquad m_h\le\frac{A(x+h)-A(x)}{h}\le M_h.$$As $h\to0^+$, both $m_h$ and $M_h$ tend to $f(x)$ because $f$ is continuous, so the quotient tends to $f(x)$ by the sandwich theorem (Section 7.1). The case $h\lt0$ is the same. $\blacksquare$ The rate at which area accumulates equals the height of the curve at the moving edge.
If $f$ is continuous on $[a,b]$ and $F$ is any antiderivative of $f$ on $[a,b]$, then
$$\int_a^b f(x)\,dx=F(b)-F(a)=\Big[F(x)\Big]_a^b.$$Proof. $A$ and $F$ are both antiderivatives of $f$ on $[a,b]$ (Theorem 8.13), so $A(x)=F(x)+C$ by Theorem 8.1. Putting $x=a$: $0=F(a)+C$. Putting $x=b$: $\int_a^bf=A(b)=F(b)+C=F(b)-F(a)$. $\blacksquare$ The constant cancels, so no $+C$ is written in a definite integral — and the answer is a number, not a function.
If $f$ is continuous and $g$, $h$ are differentiable, then
$$\frac{d}{dx}\int_{g(x)}^{h(x)}f(t)\,dt=f\big(h(x)\big)\,h'(x)-f\big(g(x)\big)\,g'(x).$$Proof. With $F$ an antiderivative of $f$, the integral is $F\big(h(x)\big)-F\big(g(x)\big)$ by Theorem 8.14; differentiate with the chain rule. $\blacksquare$ ⏱ In words: integrand at the top × derivative of the top, minus integrand at the bottom × derivative of the bottom. The integrand itself is never integrated.
Substitution and integration by parts carry over to definite integrals, with one improvement: after a substitution, change the limits and never return to $x$.
(1) If $g'$ is continuous on $[a,b]$ and $f$ is continuous on the set of values of $g$, then $\displaystyle\int_a^b f\big(g(x)\big)g'(x)\,dx=\int_{g(a)}^{g(b)}f(u)\,du.$
(2) If $u'$, $v'$ are continuous on $[a,b]$, then $\displaystyle\int_a^b u\,v'\,dx=\Big[uv\Big]_a^b-\int_a^b u'\,v\,dx.$
Proof. Both sides of (1) equal $F\big(g(b)\big)-F\big(g(a)\big)$ by Theorems 8.4 and 8.14; (2) is Theorem 8.7 evaluated between the limits. $\blacksquare$
$\dfrac{d}{dx}\displaystyle\int_0^{x^2}\cos t\,dt$ equals
Properties of Definite Integrals
Many definite integrals in JEE Main have no antiderivative you could write down, or one so long that finding it would eat the whole time limit. They are evaluated with properties instead — above all, by a substitution that reflects the interval.
For $f$ continuous on the intervals involved:
in particular $\int_0^a f(x)\,dx=\int_0^a f(a-x)\,dx$
Proofs. P1–P3 follow from Theorem 8.14: e.g. $\big(F(c)-F(a)\big)+\big(F(b)-F(c)\big)=F(b)-F(a)$. P4: in $\int_a^bf(a+b-x)\,dx$ put $t=a+b-x$, $dx=-dt$; $x=a$ gives $t=b$ and $x=b$ gives $t=a$:
$$\int_a^bf(a+b-x)\,dx=-\int_b^af(t)\,dt=\int_a^bf(t)\,dt.$$The substitution reflects the interval about its midpoint $\tfrac{a+b}{2}$, and reflecting a region does not change its area. P5: split at $a$ and put $x=2a-t$ in $\int_a^{2a}$. P6: split at $0$ and put $x=-t$ in $\int_{-a}^0f(x)\,dx$ to get $\int_0^af(-t)\,dt$, which is $+\int_0^af$ for even $f$ and $-\int_0^af$ for odd $f$. P7: $\int_a^{a+T}=\int_a^T+\int_T^{a+T}$, and $x=t+T$ turns the last into $\int_0^af(t)\,dt$; the two pieces add to $\int_0^T$. $\blacksquare$
Write $I=\int_a^bf(x)\,dx$ and also $I=\int_a^bf(a+b-x)\,dx$, and add the two. This pays off when $f(x)+f(a+b-x)$ is simple — typically a constant — as with
- $\dfrac{g(x)}{g(x)+g(a+b-x)}$: the sum is $1$, so $I=\dfrac{b-a}{2}$;
- $x\,g(x)$ on $[0,\pi]$ with $g(\pi-x)=g(x)$, e.g. $g$ built from $\sin x$: adding gives $\pi\,g(x)$, so $I=\dfrac\pi2\int_0^\pi g(x)\,dx$ and the troublesome $x$ disappears;
- $\ln$ of a trigonometric expression on $[0,\tfrac\pi2]$ or $[0,\tfrac\pi4]$, where $\sin\leftrightarrow\cos$ or $\tan\left(\tfrac\pi4-x\right)=\tfrac{1-\tan x}{1+\tan x}$.
For $f(x)=\dfrac{\sqrt x}{\sqrt x+\sqrt{7-x}}$ on $[2,5]$, what is $f(x)+f(7-x)$, and hence $\int_2^5f(x)\,dx$?
$\displaystyle\int_{-2}^{2}x^3e^{x^2}\,dx$ equals
Integrands with $|x|$ and $[x]$, and Hard Evaluations
An integrand built with a modulus or the greatest integer function $[x]$ is given by different formulas on different pieces. Use additivity (P3): split $[a,b]$ at every point where the formula changes, and integrate each piece separately.
- Modulus: find the zeros of the expression inside $|\ |$; on each piece between them it is $+$ or $-$ the expression.
- Greatest integer: $[g(x)]=k$ where $k\le g(x)\lt k+1$; split at the points where $g(x)$ is an integer. For $[x]$ itself, $\int_0^n[x]\,dx=0+1+\dots+(n-1)=\tfrac{n(n-1)}{2}$, and $x-[x]=\{x\}$ has period $1$ with $\int_0^1\{x\}\,dx=\tfrac12$.
1. Limits after substitution. Change the limits together with the variable ($u=g(a)$ to $u=g(b)$), or return to $x$ — never mix the two. When the new lower limit exceeds the upper, keep the order and let P2 handle the sign.
2. Signed area. $\int_0^{2\pi}\sin x\,dx=0$, but the area between $y=\sin x$ and the axis is $4$. A definite integral counts parts below the axis as negative.
3. The upper limit in a limit of a sum. In $\tfrac1n\sum_{r=1}^{2n}f\left(\tfrac rn\right)$ the upper limit is $2$, not $1$. Find it from the last term: $\tfrac rn=\tfrac{2n}{n}$.
4. Leibniz rule. Multiply by the derivative of each limit, and subtract the lower-limit term. $\tfrac{d}{dx}\int_{x}^{x^2}f(t)\,dt=2xf(x^2)-f(x)$ — not $f(x^2)-f(x)$.
5. Odd and even need a symmetric interval. P6 works on $[-a,a]$ only, and P4 must reflect the whole interval: for $\int_2^5$ use $7-x$, not $5-x$. And after "adding $I$ to itself", remember to halve.
$\displaystyle\int_0^{3}[x]\,dx$, where $[x]$ is the greatest integer $\le x$, equals
Exercises
Give exact answers where you can, and decimals to 4 significant figures in the answer boxes. $[x]$ is the greatest integer $\le x$.
- 1 $\int_0^3(x^2-2x+2)\,dx$
- 2 $\int_1^2\left(\dfrac1x+e^x\right)dx$
- 3 $\int_0^{\pi/3}\sec x\tan x\,dx$
- 4 $\int_0^1\dfrac{dx}{\sqrt{4-x^2}}$
- 5 Evaluate $\int_0^1(2x+3)\,dx$ as the limit of right Riemann sums, and check with Theorem 8.14.
- 6 $\displaystyle\lim_{n\to\infty}\sum_{r=1}^{n}\frac{1}{n+r}$
- 7 $\displaystyle\lim_{n\to\infty}\frac{1^4+2^4+\dots+n^4}{n^5}$
- 8 $\displaystyle\lim_{n\to\infty}\sum_{r=1}^{n}\frac{r}{n^2+r^2}$
- 9 $\displaystyle\lim_{n\to\infty}\frac1n\sum_{r=1}^{3n}\sqrt{\frac rn}$
- 1 $\dfrac{d}{dx}\displaystyle\int_1^{x}\sqrt{1+t^2}\,dt$
- 2 Show that $F(x)=\displaystyle\int_x^{2x}\frac{dt}{t}$ ($x\gt0$) is constant, using Theorem 8.15, and find the constant.
- 3 $f$ is continuous and $\int_0^xf(t)\,dt=x\sin x$ for all $x$. Find $f(\pi)$.
- 4 $G(x)=\int_0^x(t^2-4t+3)\,dt$. Find the point of local minimum of $G$.
- 5 $\int_0^1x(1-x)^9\,dx$ (use P4 first)
- 6 $\int_0^{\pi/2}\sin^3x\,dx$
- 7 $\int_0^{\pi/2}x\cos x\,dx$
- 8 $\int_1^ex\ln x\,dx$
- 9 $\int_0^2x\sqrt{x^2+5}\,dx$
- 10 $\int_0^2\sqrt{4-x^2}\,dx$ — by Theorem 8.9, and then by geometry.
- 1 $\int_{-3}^{3}(x^5-4x^3+x)\,dx$
- 2 $\int_{-2}^{2}(x^3\cos x+3x^2)\,dx$
- 3 $\displaystyle\int_0^{\pi/2}\frac{\cos^5x}{\sin^5x+\cos^5x}\,dx$
- 4 $\displaystyle\int_1^4\frac{\sqrt[3]{x}}{\sqrt[3]{x}+\sqrt[3]{5-x}}\,dx$
- 5 $\int_0^{\pi}x\sin x\,dx$ — by King's property
- 6 $\displaystyle\int_0^{\pi/2}\frac{\sin x-\cos x}{1+\sin x\cos x}\,dx$
- 7 $\int_0^{\pi}\cos^3x\,dx$ (use P5)
- 8 $\int_0^{4\pi}|\cos x|\,dx$
- 9 $\int_0^{100}(x-[x])\,dx$
- 10 $\int_0^{\pi/2}\ln\tan x\,dx$
- 1 $\int_0^3|x-1|\,dx$
- 2 $\int_{-1}^{2}|x^3-x|\,dx$
- 3 $\int_0^{1.5}[2x]\,dx$
- 4 $\int_0^{\sqrt3}[x^2]\,dx$
- 5 Show that $1\le\int_0^1e^{x^2}\,dx\le e$.
- 6 Water enters a tank at $r(t)=40+10\sin\dfrac{\pi t}{12}$ litres per hour, $t$ in hours. How much enters in $24$ hours? In the first $12$ hours?
- 7 A car's speed is $v(t)=|t-4|+2$ m/s. How far does it travel from $t=0$ to $t=6$ s?
- 8 $\displaystyle\lim_{n\to\infty}\frac1n\big[(n+1)(n+2)\cdots(2n)\big]^{1/n}$
E1. JEE Main · MCQ $\displaystyle\int_0^{\pi/2}\frac{dx}{1+\tan^3x}$ is equal to (A) $\dfrac\pi2$ (B) $\dfrac\pi4$ (C) $0$ (D) $1$
E2. JEE Main · Numerical Evaluate $\displaystyle\int_0^9\big[\sqrt x\big]\,dx$, where $[t]$ is the greatest integer $\le t$. Give the answer to the nearest integer.
E3. JEE Main · MCQ Let $F(x)=\displaystyle\int_{x^2}^{x^3}\ln t\,dt$ for $x\gt0$. Then $F'(e)$ is equal to (A) $3e^2-2e$ (B) $9e^2+4e$ (C) $9e^2-4e$ (D) $5e^2$
E4. JEE Main · Numerical Find $\displaystyle\lim_{n\to\infty}\sum_{r=1}^{2n}\frac{3r^2}{n^3}$. Give the answer to the nearest integer.
E5. JEE Main · MCQ $\displaystyle\int_0^{\pi/2}\ln(\tan x+\cot x)\,dx$ is equal to (A) $\pi\ln2$ (B) $\dfrac\pi2\ln2$ (C) $-\pi\ln2$ (D) $0$
E6. JEE Main · Numerical Evaluate $\displaystyle\frac{1}{\sqrt2}\int_0^{50\pi}\sqrt{1-\cos2x}\,dx$. Give the answer to the nearest integer.
Research question. Medicine regulators compare a generic drug with the original by the area under the concentration–time curve (AUC), computed from a handful of blood samples. How accurate is that area, and how much does the choice of sampling times matter?
Why it matters. The AUC measures the total exposure of the body to a drug, and bioequivalence studies across the world accept a generic only if its AUC is close to the original's. The AUC is a definite integral, but the curve is known only at sampling times, so it is estimated by sums of trapezoids — Riemann-type sums with uneven widths.
What to do.
- Use the model concentration $C(t)=20\big(e^{-0.2t}-e^{-1.5t}\big)$ mg/L, $t$ in hours. Find the exact $\int_0^{24}C(t)\,dt$ with the fundamental theorem.
- Take the sampling times $0, 0.5, 1, 2, 4, 6, 8, 12, 24$ h. Compute $C$ at each and the trapezoid estimate $\sum\tfrac12\big(C_{k}+C_{k+1}\big)(t_{k+1}-t_{k})$. Find the percentage error.
- Repeat with only $0, 2, 6, 12, 24$ h. Which part of the curve causes most of the error — the fast rise or the slow decay? Explain using the shape of the graph.
- Find the exact AUC from $0$ to $\infty$ (the limit as the upper limit grows). How much of it is missed by stopping at $24$ h?
Tools: A spreadsheet (one row per sample) or a few lines of Python.
Hand in: Your exact values, both trapezoid estimates with their errors, and a recommendation — in three sentences — on where to put extra sampling times if a study can afford two more.
Going further. Is the trapezoid rule an over- or under-estimate where the curve is concave down? Where it is concave up? Relate your answer to the picture of the chords.
🧪 Self-check quiz
Q1. If $\int_0^2f(x)\,dx=5$ and $\int_0^5f(x)\,dx=3$, then $\int_2^5f(x)\,dx$ is
Q2. $\dfrac{d}{dx}\displaystyle\int_1^{x}\ln t\,dt$ ($x\gt0$) is
Q3. $\displaystyle\int_{-\pi}^{\pi}x^2\sin x\,dx$ equals
Q4. $\displaystyle\lim_{n\to\infty}\frac1n\sum_{r=1}^{n}\left(\frac rn\right)^2$ equals
Q5. $\displaystyle\int_0^{\pi/2}\frac{\sin x}{\sin x+\cos x}\,dx$ equals