Skip to main content
Logo image

Coordinated Calculus

Section 5.11 Comparison of Improper Integrals

In the previous section, we learned how to compute improper integrals -- integrals involving certain functions over unbounded integrals, as well as functions that become infinite at a point within or at the endpoint of the interval of integration.
Recall the important classes of improper integrals we examined in Section 5.10.

Convergence Behavior of Important Classes of Improper Integrals.

  • 11xp dx converges if p>1 and diverges if p1
  • 011xp dx converges if p<1 and diverges if p1
  • 01eax dx converges if a>0 and diverges if a0
In this section, we will learn how to use improper integrals that we know converge or diverge (such as those above) to determine the convergence of more complex integrals.

Subsection 5.11.1 Comparing Improper Integrals

Sometimes we may encounter an improper integral for which we cannot easily evaluate the limit of the corresponding proper integral. For instance, consider 111+x3dx. While it is hard (or perhaps impossible) to find an antiderivative for 11+x3, we can still determine whether or not the improper integral converges or diverges by comparison to a simpler one. Observe that for all x>0, 1+x3>x3, and therefore
11+x3<1x3.
It therefore follows that
1b11+x3dx<1b1x3dx
for every b>1. We know that limb1b1x3dx=11x3dx converges, since it is an improper integral of the form 11xpdx with p>1. By the inequality above, if limb1b1x3+1dx exists, it must be less than or equal to 11x3dx; in particular, if limb1b1x3+1dx exists, it is finite. By making some technical considerations, it can be proved that in fact limb1b1x3+1dx must exist, and therefore 11x3+1dx is a convergent improper integral.
The comparison of these improper integrals can be visualized by comparing the area under each function. The area under the curve 1x3 is shown in green, and the area under the curve 1x3+1 is shown in blue.
Figure 5.106. Comparing the area under the curves 1x3 and 1x3+1
In particular, 111+x3dx must converge, even though we never explicitly evaluated the corresponding limit of proper integrals. We use this idea and similar ones in the examples that follow.

Example 5.107.

  1. Explain why x2+x+1>x2 for all x1, and hence determine if 11x2+x+1dx converges or diverges by comparison to 11x2dx.
  2. Observe that for each x>1, ln(x)<x. Explain why
    2b1xdx<2b1ln(x)dx
    for each b>2. Why must it be true that 2b1ln(x)dx diverges?
  3. Explain why x4+1x4>1 for all x>1. Then, determine whether or not the improper integral
    11xx4+1x4dx
    converges or diverges.
Answer.
  1. converges
  2. diverges
  3. diverges
Solution.
  1. Since x1, it follows that x+11>0, so adding x2 to both sides of this inequality shows that
    x2+x+1>x2
    for all x1. Taking reciprocals (which reverses the inequality),
    1x2+x+1<1x2
    and thus for any b>1,
    1b1x2+x+1dx<1b1x2dx.
    Taking the limit as b, it follows that
    11x2+x+1dx<11x2dx,
    and since we know that 11x2dx converges, we conclude that 11x2+x+1dx also converges.
  2. Since for each x>1, ln(x)<x, it also holds that ln(x)<x for all x>2. Taking reciprocals,
    1x<1ln(x)
    and thus
    2b1xdx<2b1ln(x)dx
    for each b>2. Taking the limit as b, we have
    21xdx21ln(x)dx
    and since we know that 11xdx diverges, it follows that 2b1ln(x)dx also diverges.
  3. Since x4+1>x4 for all x, it follows that
    x4+1x4>1
    for all x>1, and thus taking square roots,
    x4+1x4>1
    for all x>1. Integrating from 1 to b and then letting b, we can next say that
    1x4+1x4dx11dx
    It is obvious that 11dx diverges since the integrand does not tend to 0. By the last inequality above, we can conclude that
    11xx4+1x4dx
    diverges.
Part (b) demonstrates that if an improper integral can be bounded below by one that diverges, then the improper integral in question also diverges. The same logic can be used to show that if an improper integral can be bounded above by one that converges, it also converges. We formalize these principles as the Comparison Test for Improper Integrals.

Comparison Test for Improper Integrals.

If f(x)g(x)0 and af(x)dx converges, then ag(x)dx converges.
If f(x)g(x)0 and ag(x)dx diverges, then af(x)dx diverges.
For Part (b), we claimed that 21xdx diverges, but we’ve technically only discussed that 11xdx diverges. However, since 11xdx=21xdx+121xdx and 121xdx is finite, 21xdx must diverge; otherwise, the right-hand side of the equation would be finite. In fact, a1xdx diverges for any real number a1 for the same reason. More generally:

.

Suppose f(x) is continuous on the interval [a,b]
  • If af(x)dx converges, then bf(x)dx also converges.
  • If af(x)dx diverges, then bf(x)dx also diverges.
We’ll finish this section with a few guided examples of applications of the Comparison Test. A useful strategy for these types of problems goes as follows: first, make an educated guess about whether the given integral converges or diverges; then, based on this guess, bound the integrand above or below by a simpler function whose integral over the same interval converges or diverges, respectively.

Example 5.108.

Use the Comparison Test for Improper Integrals to determine whether the following improper integrals converge or diverge. Be sure to justify any inequalities used when applying the Test.
  1. 1sin2(x)x2dx
  2. 11x3+6dx
  3. 103x(x+4)2dx
  4. 12+cos(x)x+12dx
Hint.
  1. Since 1sin(x)1 for any xR, we have 0sin2(x)1 for any xR.
  2. When x is large, observe that x3+6x3=x3/2.
  3. When x is large, 3x(x+4)23xx2=3x.
  4. Since cos(x)1 for all xR, the numerator is at least 1. For the denominator, when x is large we have x+12x.
Answer.
  1. Converges, via comparison with e.g. 1x2.
  2. Converges, via comparison with e.g. 1x3/2.
  3. Diverges, via comparison with e.g. 34x.
  4. Diverges, via comparison with e.g. 12x on the interval [4,).
Solution.
  1. From the hint, we know 0sin2(x)1 for any xR. Therefore, 0sin2(x)x21x2 for any xR. (In particular, this inequality holds on the interval [1,).) Since 11x2dx converges, by the Comparison Test it follows that 1sin2(x)x2dx converges.
  2. From the hint, we should expect our improper integral to behave like 11x3/2dx, which converges. Indeed, we can compare our integral directly to this one.
    For any xR, we have x3+6>x3. Therefore, x3+6>x3=x3/2 whenever both x3+6 and x3 are non-negative, i.e., for all x0. This, along with the assumption that we use the positive square root, yields the inequality 1x3/2>1x3+6>0 for all x>0. Specifically, the inequality holds on the interval [1,). Because 11x3/2dx converges, by the Comparison Test we conclude that 11x3+6dx converges.
  3. From the hint, we should expect our imroper integral to behave like 103xdx, which diverges. However, since (x+4)2>x2 for all xR it’s actually the case that 3x(x+4)2<3xx2=3x for xR. So we cannot compare our function to 3x to conclude divergence. More care is needed to find an appropriate lower bound for 3x(x+4)2.
    Observe that if x4, we have 2xx+4. (In fact, these inequalities are equivalent.) So (2x)2(x+4)2 for x4 and thus we may write 3x(x+4)23x(2x)2=34x0. Now, 1034xdx=34101xdx diverges, so it follows from the Comparison Test that 103x(x+4)2dx diverges.
  4. From the hint, for large enough x the function 2+cos(x)x+12 is approximately bounded below by 1x. Since 11xdx diverges, we should expect our integral to diverge as well. However, a direct comparison to 1x won’t show divergence, since 1x+12<1x.
    Note, however, that for x4 we have x+124x. (These inequalities are indeed equivalent.) So when x4, we know 2+cos(x)x+1214x=12x>0. Since 412xdx=1241xdx diverges, by the Comparison Test 42+cos(x)x+12dx diverges. It follows that 12+cos(x)x+12dx diverges as well.

Subsection 5.11.2 Summary

  • There are three important classes of improper integrals discussed in this section.
    • One important class of improper integrals is given by
      11xpdx
      where p is a positive real number. We can show that this improper integral converges whenever p>1 and diverges whenever p1.
    • A related class of improper integrals is
      011xpdx,
      which converges for p<1 and diverges for p1.
    • Another class of improper integrals is
      0eaxdx,
      which converges when a>0 and diverges when a0.
  • These important classes of improper integrals are used for comparisons in the Comparison Test for Improper Integrals. The Comparison Test for Improper Integrals allows us to determine if an improper integral converges or diverges without having to calculate the antiderivative. The actual test states the following:
    If f(x)g(x)0 and af(x)dx converges, then ag(x)dx converges.
    If f(x)g(x)0 and ag(x)dx diverges, then af(x)dx diverges.

Exercises 5.11.3 Exercises

1. Determining convergence or divergence of various improper integrals.

Determine, with justification, whether each of the following improper integrals converges or diverges.
  1. eln(x)xdx
  2. e1xln(x)dx
  3. e1x(ln(x))2dx
  4. e1x(ln(x))pdx, where p is a positive real number
  5. 01ln(x)xdx
  6. 01ln(x)dx