Why Use a Substitution to Factor?
A polynomial can have a high degree while using only a few powers of its variable. For instance, \(x^4-5x^2+4\) has degree four, but its terms involve only \(x^2\), \(x^4\), and a constant. If we temporarily write \(u=x^2\), the expression becomes a quadratic in \(u\). We can factor that quadratic using the methods for quadratic trinomials, then replace \(u\) with \(x^2\).
The substitution is a way to reveal the structure already present in the polynomial. It does not change the expression or its value: it gives us a simpler form to factor before we return to the original variable.
The Substitution and Its Conditions
To use the method, first identify a repeated power in the terms. If the exponents are \(2n\), \(n\), and \(0\), set \(u=x^n\). Factor the resulting quadratic in \(u\), if it factors over the coefficient system you are using, and then substitute \(x^n\) back for every \(u\).
For example, when \(n=2\), a polynomial in \(x^4\), \(x^2\), and a constant becomes a quadratic in \(u=x^2\). When \(n=3\), a polynomial in \(x^6\), \(x^3\), and a constant becomes a quadratic in \(u=x^3\). The same reasoning works for any positive integer \(n\).
Here, “factors” depends on the coefficients allowed in the answer. A quadratic might factor over the real numbers but not over the integers or rationals. In each example, the coefficient system will be stated where it matters. After substituting back, also check whether the resulting factors can be factored further using familiar identities, such as the difference of squares or the sum and difference of cubes.
Why the Method Works
The reasoning is based on substitution into an identity. Suppose the quadratic factors as \(a u^2+b u+c=(p u+q)(r u+s)\), with constants \(p\), \(q\), \(r\), and \(s\). Replacing \(u\) by \(x^n\) gives an identity in \(x\):
To justify it, expand the right-hand side after substitution. Its product is \(prx^{2n}+(ps+qr)x^n+qs\), exactly the result of substituting \(x^n\) for \(u\) in the original quadratic product. Thus each value of \(x\) gives the same value on both sides. This is why the substitution preserves the factorization.
It is useful to think of \(u\) as a temporary name, not a new restriction on \(x\). In particular, when \(u=x^2\), we do not need to solve an equation for \(x\) or restrict \(x\) to make the algebraic identity valid. We are factoring an expression, not solving for values of a variable.
Worked Examples
Worked Example: Factor a Quartic and Continue Factoring
Factor \(x^4-5x^2+4\) into linear factors over the integers.
The terms have powers \(x^4=(x^2)^2\), \(x^2\), and a constant, so set \(u=x^2\). Then
We need two numbers with product \(4\) and sum \(-5\). They are \(-1\) and \(-4\), so
Substitute \(x^2\) back for \(u\):
Each factor is a difference of squares. Using \(A^2-B^2=(A-B)(A+B)\), factor both:
Therefore the complete factorization over the integers is
Check by multiplying the paired factors: \((x-1)(x+1)=x^2-1\) and \((x-2)(x+2)=x^2-4\). Their product is \(x^4-4x^2-x^2+4=x^4-5x^2+4\), as required.
Worked Example: Factor a Quartic with a Leading Coefficient
Factor \(2x^4-7x^2+3\) over the integers as far as this substitution method allows.
Set \(u=x^2\), so the expression becomes
To factor, split the middle term. The product of the leading coefficient and constant is \(2\cdot3=6\). The numbers \(-6\) and \(-1\) have product \(6\) and sum \(-7\):
Now substitute \(x^2\) for \(u\):
Neither factor splits into factors with rational coefficients: their roots would involve \(\sqrt{1/2}\) and \(\sqrt{3}\), respectively. So this is the factorization over the rationals obtained here. Check it by multiplying:
Worked Example: Factor Further Using Difference of Cubes
Factor \(x^6-9x^3+8\) over the integers, and continue factoring after substituting back.
The powers are \(x^6=(x^3)^2\), \(x^3\), and a constant. Set \(u=x^3\):
The numbers \(-1\) and \(-8\) have product \(8\) and sum \(-9\), giving
Substitute \(x^3\) back:
Both factors are differences of cubes. The identity \(A^3-B^3=(A-B)(A^2+AB+B^2)\) gives
Thus
Check: the first pair of factors multiplies to \(x^3-1\), and the second pair to \(x^3-8\). Multiplying those two results gives \(x^6-8x^3-x^3+8=x^6-9x^3+8\).
A Reliable Sequence of Steps
Use the following sequence when the exponents suggest that a polynomial is quadratic in a power of its variable.
Check whether the exponents are \(2n\), \(n\), and \(0\), and name \(x^n\) as \(u\).
Use the appropriate quadratic factoring method, and keep the coefficients in the form required by the question.
Replace every \(u\) in every factor with the original power \(x^n\).
Check the substituted factors for familiar patterns, including differences of squares and sums or differences of cubes.
Multiply the final factors, or multiply in stages, to confirm that they reproduce the original polynomial.
When Substitution Does Not Give a Factorization
The method depends on the quadratic in \(u\) factoring in the chosen coefficient system. For example, \(x^4+3x^2+5\) becomes \(u^2+3u+5\) when \(u=x^2\). Its discriminant is \(3^2-4(1)(5)=9-20=-11\), so it has no real linear factors in \(u\). Consequently, this substitution does not produce a factorization into two linear factors in \(x^2\) over the real numbers.
This conclusion is deliberately limited: it says what the quadratic-substitution method yields, not that no other kind of factorization of the quartic is possible. Always distinguish “this substitution did not factor it” from “the polynomial cannot be factored in any way.”
Common Mistake: Substituting Back Incorrectly
A frequent error is to factor the quadratic in \(u\) correctly but then replace \(u\) with \(x\) instead of \(x^2\). For example, starting with \(x^4-5x^2+4\), one might write \(u=x^2\), obtain \((u-1)(u-4)\), and mistakenly claim
This is wrong: multiplying the claimed factors gives \(x^2-5x+4\), which is not the original quartic. The correct substitution replaces each \(u\) with \(x^2\), not with \(x\):
Multiplying the corrected expression gives \(x^4-5x^2+4\). Keeping track of exactly what \(u\) represents prevents this mistake.
Why This Pattern Matters
Recognizing a polynomial in a power lets you use a familiar quadratic structure even when the polynomial has a higher degree. It can also reveal factors that are not obvious from the original expression. The substitution is most useful when it shortens the algebra without hiding the original variable: write down what \(u\) means, factor, substitute back, and then inspect the result for further structure.
Three checks help keep the work precise: verify that the exponents really match the required pattern; replace every \(u\) with the entire power used in the substitution; and multiply the final factors to check the original expression. If the quadratic in \(u\) does not factor over the required coefficients, do not force factors that are not there.
Check Your Understanding
For each question, show the substitution and the algebra needed to support your answer.
- Write \(x^8-6x^4+8\) as a quadratic in a suitable temporary variable. What is that variable?
- Factor \(x^4-10x^2+9\) as far as possible over the integers. Verify your result by multiplying.
- Factor \(3x^4-10x^2+3\) by setting \(u=x^2\), and check the product after substituting back.
- For \(x^6-5x^3+4\), factor first as a quadratic in \(x^3\), then check whether either resulting factor can be factored further over the integers.
- Explain why a factorization of \(u^2+3u+5\) into real linear factors is unavailable. What does that tell you—and not tell you—about using \(u=x^2\) on \(x^4+3x^2+5\)?
- Correct this proposed step for \(x^4-5x^2+4\): with \(u=x^2\), \((u-1)(u-4)\) becomes \((x-1)(x-4)\).
