Idea of trigonometric substitution: almost the reverse of -substitution, we want to add trigonometry to a problem
so that the integral simplifies, using trigonometric identities. It’s best to see this with an example:
This is the area under a…..semicircle! Since the curve is related to circles, we could try to replace with a trigonometric
function, which measures the -coordinate of a circle . Then , and so
This square root simplifies to (using ), and so all together the substitution results in
Which, when is positive (on , say) is just the integral
Which is one we have already learned to do with trigonometric identities!
What substitutions are useful?
This is the overall goal:
Choose a substitution trigonometric function
Compute , substitute all into the integral
Use trigonometric identities to simplify the result, so you can integrate.
The difficulty in this technique is not the substitution or integration itself (its just a form of -substitution, after all!)
but rather making a good choice of which substitution to make in the first place. We want the choice to lead to some nice
simplification using trigonometric identities, which means we need to be thinking ahead!We derive three useful substitutions
from two identities and , by re-arranging and taking square roots.
- means that the substitution will convert expressions like to .
- means that the substitution will convert expressions like to .
- means that the substitution will convert expressions like to .
Substituting
When an integral contains (where ranges over a subset of ), we may wish to substitute
(where ranges over a subset of ). This converts into and .
Examples:
Then so finally
Substituting
When an integral contains (where ranges over any subset of the real line), we may wish to substitute
(where ranges over a subset of ). This converts into and .
$$\int \frac{1}{x^2\sqrt{x^2+4}dxx=2\tan\theta$$
Then so finally
Substituting
To do these integrals, its often useful to know the integral of secant
This can be confirmed by differentiation; though we will derive it soon when we do partial fractions!
When an integral contains (where ranges is always greater in absolute value than ), we may wish to substitute
(where ranges over a subset of or ). This converts into and .
Let .
Substitution for Definite Integrals
If an integral is a definite integral measuring the area under a curve, we can convert the bounds of integration instead of
transforming back to the original variable, and then evaluating at the original bounds.
While in ‘standard’ -sub both approaches take the same amount of work, converting the bounds here feels significantly easier
than going through with the whole substitution first!
Here’s an example:
Making the substitution on the indefinite integral results in
Substituting the bounds to get our new definite integral, we need to figure out what equals when and .
If and , then .
If then , which has the unique solution for .
Thus our new definite integral is
Recognizing a ‘Hidden’ Trigonometric Substitution
Complete the square underneath the integral before proceeding!
Thus our integral is
Substituting will make the square root into a form we know:
This we can attack with the trigonometric substitution !