Rationalizing Fractional Powers with x = t^n

Why Use a Power Substitution?

A fractional power of \(x\) can make an algebraic integral awkward to simplify. If the fractional exponents have denominators that can all be cleared by a single integer, the substitution \(x=t^n\) converts those powers into integer powers of \(t\). In favorable cases, the entire integrand, including the differential, becomes a rational function of \(t\).

The substitution does not automatically make every algebraic integral rational. Its success depends on the powers present, and the real domain matters: an even power \(t^n\) cannot represent negative \(x\), and fractional powers such as \(\sqrt{x}\) require a careful choice of branch.

Choosing the Exponent

Write each fractional exponent in reduced form. If the powers of \(x\) include \(x^{p_1/q_1},\ldots,x^{p_k/q_k}\), choose \(n\) to be a common multiple of the denominators \(q_1,\ldots,q_k\). The least common multiple is often the smallest useful choice. Then \(n(p_i/q_i)\) is an integer for each exponent.

For example, the powers \(x^{1/2}\) and \(x^{1/3}\) have denominators \(2\) and \(3\). Taking \(n=6\) gives \(x=t^6\). On the positive branch \(t>0\), this means \(x^{1/2}=t^3\) and \(x^{1/3}=t^2\). The differential is obtained by differentiating \(x=t^n\):

Consequently, every power \(x^r\) with \(r=p/q\) becomes \(t^{np/q}\), an integer power when \(q\) divides \(n\). Negative integer powers are allowed; they give powers in the denominator of a rational function. If the original integrand is built from rational operations on these powers, the transformed integrand is rational.

Derivation of the Transformed Integral

The calculation follows directly from substituting the expressions for \(x\), its powers, and its differential. For a positive interval, choose \(t>0\), so that \(x=t^n\) has the inverse \(t=x^{1/n}\) and \(x^{p/q}=t^{np/q}\). The resulting integral is:

This formula is a change of variable: it requires the chosen \(t\)-interval to map to the \(x\)-interval and the integrand to be defined there. To see why the differential factor is necessary, suppose \(G(t)\) is an antiderivative of the transformed integrand. The chain rule gives:

Since \(G'(t)=f(t^n)nt^{n-1}\), it follows that:

Thus the transformed antiderivative, expressed back in terms of \(x\), differentiates to the original integrand wherever the substitution is reversible and the expressions are defined.

Worked Example: Clearing Two Fractional Powers

When the Substitution Produces a Rational Function

Consider an expression made from powers \(x^{p_i/q_i}\) using addition, subtraction, multiplication, and division. If \(n\) is a common multiple of their denominators, then substituting \(x=t^n\) changes each power into an integer power of \(t\). Addition and multiplication of integer powers produce Laurent polynomials (finite sums that may include negative powers). A quotient of Laurent polynomials can be written as a quotient of ordinary polynomials after multiplying numerator and denominator by a suitable power of \(t\). Including \(dx=nt^{n-1}\,dt\) therefore still gives a rational function, except at points where a denominator vanishes.

This provides a practical recognition rule. A mixture of rational powers of \(x\) is a candidate for this substitution; a single radical is not the only possible case. Once the transformed integrand is rational, the rational-function integration methods of the earlier material apply, subject to checking for poles on the interval.

Worked Example: A Rational Function After Substitution

Sign, Range, and Real Powers

The allowed values of \(t\) are part of the substitution, not an optional detail. If \(x>0\) and \(n\) is even, use \(t>0\) to make \(x=t^n\) one-to-one. Allowing both signs of \(t\) would represent each positive \(x\) twice and would make \(\sqrt{x}=|t^{n/2}|\), not always \(t^{n/2}\). If \(n\) is odd, \(t\mapsto t^n\) is one-to-one on all real numbers and preserves sign.

Negative \(x\) requires special care. An even power \(t^n\) is never negative, so it cannot cover an interval where \(x<0\). For negative \(x\), a real fractional power \(x^{p/q}\) is defined in the usual real sense only when the reduced denominator \(q\) is odd. If all such denominators are odd, an odd common multiple \(n\) can be chosen, and \(x=t^n\) maps negative \(t\) to negative \(x\). The corresponding range of \(t\) must remain negative.

For definite integrals, transform both endpoints using the chosen branch. If \(x\) runs through an interval where the substitution is one-to-one, the new limits are the corresponding values of \(t\); the factor \(dx=nt^{n-1}\,dt\) already accounts for orientation. If the original interval crosses \(x=0\), a singularity, or a point where the chosen formula changes branch, handle the relevant pieces separately.

Worked Example: Using an Odd Power on a Negative Interval

A Common Branch Error

A frequent mistake is to use an even power substitution on a positive interval while allowing \(t\) to be negative, then replace a root by an unsigned power of \(t\). For example, with \(x=t^6\), the claim \(\sqrt{x}=t^3\) is wrong for \(t<0\): in fact, \(\sqrt{t^6}=|t^3|\). The incorrect calculation \(\sqrt{t^6}=t^3\) would turn a nonnegative quantity into a negative one whenever \(t<0\).

The correction is to specify the branch. For \(x>0\), take \(t>0\), so \(\sqrt{x}=t^3\) is valid. Alternatively, if both signs of \(t\) are retained, write \(\sqrt{x}=|t^3|\) and split the work by sign. Restricting to the positive branch is usually simpler and keeps the substitution one-to-one.

Practical Procedure

The substitution \(x=t^n\) is useful because it can convert several interacting fractional powers into ordinary integer powers at once. The decisive checks are whether the exponent denominators are cleared, whether the transformed expression is genuinely rational, and whether the selected real branch covers the original interval.