Inverse Substitution

In the substitution method, we introduce u = g ( x ) . Inverse substitution reverses this direction: we write x as a function of a new variable t , setting x = ϕ ( t ) , so that the integrand simplifies dramatically.

Inverse Substitution

In the substitution method, we introduce a new variable u as a function of x , writing u = g ( x ) , and replace the integral in x with one in u . Inverse substitution reverses this direction: we write x as a function of a new variable t , setting x = ϕ ( t ) , and transform the integral accordingly. If x = ϕ ( t ) , then dx = \phi'(t)\,dt, and the integral becomes

\int f(x)\,dx = \int f\bigl(\phi(t)\bigr)\,\phi'(t)\,dt.

The substitution x = ϕ ( t ) must be invertible on the interval of integration, so that once we evaluate the right-hand side, we can return to the original variable x .

The power of inverse substitution lies in choosing ϕ so that the transformed integrand simplifies dramatically. A radical that resists direct attack often collapses into a trigonometric or hyperbolic identity after a well-chosen substitution. But the technique is not limited to radicals. Any time a change of the dependent variable x simplifies the structure of the integrand, inverse substitution applies.

Worked Examples

Using the substitution x = tan t , evaluate d x 1 + x 2 .

Solution

Many students recall that d d x arctan x = 1 1 + x 2 , and simply write down arctan x + C . Here we derive that result from scratch using inverse substitution, so the method is clear.

Set x = tan t , so that d x = sec 2 t d t . The denominator becomes

1 + x 2 = 1 + tan 2 t = sec 2 t .

Substituting,

\begin{aligned} \int\frac{dx}{1+x^2} &= \int\frac{\sec^2 t}{\sec^2 t}\,dt \\ &= \int dt \\ &= t + C. \end{aligned}

Since x = tan t , we have t = arctan x , also written tan 1 x , and so

d x 1 + x 2 = arctan x + C .

Using the substitution x = sin t , evaluate d x 1 x 2 .

Solution

One may recall that d d x arcsin x = 1 1 x 2 , and write down arcsin x + C directly. Again, we derive this using inverse substitution.

Set x = sin t with t ( π 2 , π 2 ) , so that d x = cos t d t . On this interval cos t > 0 , so

1 x 2 = 1 sin 2 t = | cos t | = cos t .

Substituting,

\begin{aligned} \int\frac{dx}{\sqrt{1-x^2}} &= \int\frac{\cos t}{\cos t}\,dt \\ &= \int dt \\ &= t + C. \end{aligned}

Since x = sin t , we have t = arcsin x (also written sin 1 x ), and so

d x 1 x 2 = arcsin x + C .

Using the substitution x = tanh t , evaluate d x 1 x 2 , | x | < 1 .

Solution

This integral can be handled by partial fractions, but it also yields cleanly to inverse substitution with a hyperbolic function, and parallels the first example closely.

Set x = tanh t , so that d x = sech 2 t d t . The denominator becomes

1 x 2 = 1 tanh 2 t = sech 2 t .

Substituting,

\begin{aligned} \int\frac{dx}{1-x^2} &= \int\frac{\operatorname{sech}^2 t}{\operatorname{sech}^2 t}\,dt \\ &= \int dt \\ &= t + C. \end{aligned}

Since x = tanh t , we have t = arctanh x , and so

d x 1 x 2 = arctanh x + C .

Using the substitution x = sinh t , evaluate d x 1 + x 2 .

Solution

This is the hyperbolic analogue of the second example. Set x = sinh t , so that d x = cosh t d t . Since cosh t > 0 for all t ,

1 + x 2 = 1 + sinh 2 t = cosh t .

Substituting,

\begin{aligned} \int\frac{dx}{\sqrt{1+x^2}} &= \int\frac{\cosh t}{\cosh t}\,dt \\ &= \int dt \\ &= t + C. \end{aligned}

Since x = sinh t , we have t = arcsinh x , and so

d x 1 + x 2 = arcsinh x + C .