Quadratic Equations — Advanced

The quadratic formula solves ax^2 + bx + c = 0. But what about x^4 - 5x^2 + 4 = 0? That is a fourth-degree equation — the formula does not apply directly. Yet if you notice that x^4 = (x^2)^2 and set t = x^2, the equation becomes t^2 - 5t + 4 = 0, which is a quadratic in t. Solve for t, then solve for x. A degree-four problem reduced to a degree-two problem in one line.

This is the central idea of this article: many equations that are not quadratic in the original variable become quadratic after a substitution. The trick is spotting the substitution. Once you see it, the rest is routine.

Equations reducible to quadratic form

An equation is "reducible to quadratic form" if there exists some expression t = f(x) such that the equation can be rewritten as at^2 + bt + c = 0. The most common patterns fall into a few recognisable families.

Pattern 1: Bi-quadratic equations

A bi-quadratic (or quartic with only even powers) has the form

Set t = x^2. Then x^4 = t^2, and the equation becomes at^2 + bt + c = 0. Solve for t, then take x = \pm\sqrt{t} for each positive value of t.

Pattern 2: Equations in x^n

Any equation of the form a\bigl(f(x)\bigr)^2 + b\bigl(f(x)\bigr) + c = 0 becomes quadratic with t = f(x). Common instances:

• a(x^3)^2 + b(x^3) + c = 0 — set t = x^3
• a \cdot 4^x + b \cdot 2^x + c = 0 — set t = 2^x (since 4^x = (2^x)^2)
• a\sin^2\theta + b\sin\theta + c = 0 — set t = \sin\theta

The key observation is the same each time: look for a sub-expression that appears both squared and unsquared. That sub-expression is your t.

Pattern 3: Fractional equations

Consider

This does not look quadratic. But set t = x/(x-1). Then 1/t = (x-1)/x, and the equation becomes t + 1/t = 5/2, which gives 2t^2 - 5t + 2 = 0 after multiplying through.

Alternatively — and often more cleanly — set u = x - 1/x or some other combination that captures both fractions. The right substitution depends on the structure. The guiding principle: find the expression that, when squared, produces the highest-degree term.

Reciprocal equations

A reciprocal equation is a polynomial equation where the coefficients read the same forwards and backwards. The simplest quadratic example is ax^2 + bx + a = 0 — note c = a.

For a fourth-degree reciprocal equation like

the standard trick is to divide the entire equation by x^2 (valid since x = 0 is never a root of a reciprocal equation — check: a \cdot 0 + \ldots + a = a \neq 0):

Now set t = x + 1/x. Then x^2 + 1/x^2 = t^2 - 2 (verify: (x + 1/x)^2 = x^2 + 2 + 1/x^2). The equation becomes

A quadratic in t. Solve for t, then solve x + 1/x = t — which itself is a quadratic: x^2 - tx + 1 = 0.

The substitution t = x + 1/x is worth memorising. It appears in reciprocal equations, in trigonometric identities (since \sec\theta + \cos\theta has this form), and in many competition problems. Its key property is that x^2 + 1/x^2 = t^2 - 2, x^3 + 1/x^3 = t^3 - 3t, and so on — all higher powers of x + 1/x can be expressed as polynomials in t.

Substitution methods: a systematic approach

Here is a checklist for identifying substitutions. Given an equation that is not obviously quadratic:

1. Look for a repeated sub-expression. If you see f(x) and \bigl(f(x)\bigr)^2 in the equation, set t = f(x).

2. Look for paired terms. If the equation involves both x and 1/x (or f(x) and 1/f(x)), try t = x + 1/x or t = x - 1/x.

3. Look for exponential patterns. If you see a^{2x} and a^x, set t = a^x. If you see a^x + a^{-x}, try t = a^x.

4. Look for trigonometric patterns. Equations in \sin^2\theta and \sin\theta are quadratic in t = \sin\theta. Use \cos^2\theta = 1 - \sin^2\theta to reduce mixed equations to a single function.

5. Look for radical patterns. If you see \sqrt{x} and x (which is (\sqrt{x})^2), set t = \sqrt{x}.

After solving for t, you must back-substitute to find x, and you must check for extraneous solutions — the substitution may have introduced values that do not satisfy the original equation.

Parametric quadratic equations

A parametric quadratic is one where the coefficients depend on a parameter k (or m, or \lambda). The question is usually not "solve for x" but "for what values of k does the equation have roots with a specific property?"

The toolkit for these problems comes from the discriminant and Vieta's formulas:

• Both roots real: require D \geq 0
• Both roots positive: require D \geq 0, sum > 0, product > 0
• Both roots negative: require D \geq 0, sum < 0, product > 0
• Roots of opposite sign: require product < 0 (this automatically gives D > 0)
• One root greater than k_0: analyse f(k_0) and the position of the vertex
• Both roots in an interval [p, q]: require D \geq 0, f(p) \geq 0, f(q) \geq 0, and p \leq -b/(2a) \leq q

Each condition translates into an inequality in the parameter. Solving the system of inequalities gives the range of the parameter.

An interactive parameter explorer

Drag the point to change k in the equation x^2 - 2kx + (k + 2) = 0. The readouts show the discriminant, the sum, and the product of the roots. Watch how the nature of the roots changes as k moves.

Two worked examples

Common confusions

• "Setting t = x^2 means t is always positive." If you are solving for real x, then t = x^2 \geq 0, so yes — any negative value of t you get from the quadratic must be discarded. This is why a bi-quadratic equation with four expected roots might give only two real ones (or none).

• "Back-substitution always works." It does, but it can introduce extraneous solutions. For instance, if you square both sides during a substitution (as when clearing a radical), you might get extra values that do not satisfy the original equation. Always verify.

• "A fourth-degree equation always has four roots." Over the complex numbers, yes (counting multiplicity). Over the reals, a fourth-degree equation can have 0, 2, or 4 real roots. The substitution t = x^2 reduces the problem to a quadratic in t, but not every t-root gives real x-values.

• "Reciprocal equations are rare." They appear more often than you might expect — especially in trigonometry (where \sin\theta and \csc\theta are reciprocals) and in problems involving x + 1/x. The JEE is particularly fond of them.

• "I need to memorise the substitution for each type." The underlying principle is always the same: find the expression that appears both linearly and squared in the equation. That expression is your t. Once you internalise this, the specific patterns (bi-quadratic, reciprocal, exponential) are just instances of the same idea.

Going deeper

Why reciprocal equations always reduce

A polynomial p(x) = a_n x^n + a_{n-1}x^{n-1} + \ldots + a_1 x + a_0 is called self-reciprocal (or palindromic) if a_k = a_{n-k} for all k. If \alpha is a root of such a polynomial, then 1/\alpha is also a root. The proof: p(x) = x^n p(1/x) when the coefficients are palindromic, so p(\alpha) = 0 implies \alpha^n p(1/\alpha) = 0, and since \alpha \neq 0 (because a_0 = a_n \neq 0), we get p(1/\alpha) = 0.

This means the roots come in reciprocal pairs. For a degree-4 palindromic polynomial, the four roots are \alpha, 1/\alpha, \beta, 1/\beta. The substitution t = x + 1/x groups each reciprocal pair into a single t-value, halving the degree. This is not a trick — it is exploiting the symmetry of the equation.

The general principle: symmetry reduces degree

Every substitution method in this article is an instance of a broader principle: if an equation has a symmetry, you can exploit that symmetry to reduce the degree. The bi-quadratic ax^4 + bx^2 + c = 0 has the symmetry x \to -x (replacing x with -x does not change the equation). The substitution t = x^2 "quotients out" this symmetry, halving the degree. The reciprocal equation has the symmetry x \to 1/x, and t = x + 1/x quotients that out.

In the language of Galois theory (which you will meet much later), these substitutions correspond to intermediate fields in the splitting field of the polynomial. But you do not need Galois theory to use them — the practical observation is simple: look for a transformation that leaves the equation unchanged, and use it to define your substitution variable.

Equations with common roots

If two quadratic equations a_1x^2 + b_1x + c_1 = 0 and a_2x^2 + b_2x + c_2 = 0 share a common root \alpha, then \alpha satisfies both equations. Subtracting one from the other (after appropriate scaling) eliminates x^2 and gives a linear equation in \alpha, from which you can find \alpha explicitly. Then substituting \alpha back gives a condition on the coefficients.

The condition for two quadratics to have a common root can be expressed as a determinant (the resultant of the two polynomials) being zero:

This is a topic that connects the algebra of quadratic equations to the broader theory of polynomial resultants, which you will explore in Common Roots.

Where this leads next

• Quadratic Formula — the formula that powers all the solutions in this article, derived from first principles.
• Sum and Product of Roots — Vieta's formulas, which appear throughout parametric problems as the conditions on roots.
• Discriminant and Nature of Roots — the discriminant is the gatekeeper for every parametric problem: is D \geq 0?
• Polynomial Operations — how polynomial algebra extends the techniques of this article to higher degrees.
• Common Roots — when two polynomial equations share a root, and what conditions that imposes on their coefficients.