1.1 - Equations in quadratic form

1.1 - Equations in quadratic form

An equation can hide a quadratic inside powers, square roots or reciprocals. You will identify a useful replacement variable, solve the resulting quadratic and return carefully to the original input. Conditions on the replacement variable will explain why some algebraic candidates must be rejected.

Choosing a new variable

How can x4−13x2+36=0x^4-13x^2+36=0 be solved without learning a new formula for fourth powers? Notice that x4=(x2)2x^4=(x^2)^2: the equation is quadratic in x2x^2.

Set u=x2u=x^2. The equation becomes u2−13u+36=0u^2-13u+36=0, which uses the methods you already know. The job has two stages: solve for uu, then solve x2=ux^2=u. Because xx is real, remember from the start that u≥0u\ge0.

Quadratic in an expression

An equation is quadratic in an expression when it can be written au2+bu+c=0au^2+bu+c=0, with a≠0a\ne0, after replacing that repeated expression by uu. A biquadratic equation such as ax4+bx2+c=0ax^4+bx^2+c=0 is quadratic in x2x^2.

Choose the repeated expression, not its square: replacing x4x^4 by uu would leave an awkward x2x^2 term rather than a quadratic in uu.

Worked example 1: four real answers

Question. Solve x4−13x2+36=0x^4-13x^2+36=0 for real xx.

Plan. Put u=x2u=x^2, solve the quadratic in uu, then take both signs for each positive value.

u2−13u+36=0(u−4)(u−9)=0u=4 or 9\begin{aligned} u^2-13u+36&=0\\ (u-4)(u-9)&=0\\ u&=4\text{ or }9 \end{aligned}

Thus x2=4x^2=4 gives x=±2x=\pm2, and x2=9x^2=9 gives x=±3x=\pm3.

Answer. x=−3,−2,2,3x=-3,-2,2,3.

Check. At x=±2x=\pm2, the original is 16−52+36=016-52+36=0. At x=±3x=\pm3, it is 81−117+36=081-117+36=0. Only even powers occur, so each sign pair gives the same value.

Keep the first two terms but change the constant. One of the new values of uu will violate u≥0u\ge0.

Worked example 2: an impossible value of the square

Question. Solve x4−13x2−14=0x^4-13x^2-14=0 for real xx.

Plan. Use the same replacement and check each transformed root before returning to xx.

u2−13u−14=0(u−14)(u+1)=0u=14 or −1\begin{aligned} u^2-13u-14&=0\\ (u-14)(u+1)&=0\\ u&=14\text{ or }-1 \end{aligned}

Since u=x2≥0u=x^2\ge0, reject u=−1u=-1. From x2=14x^2=14, take both signs.

Answer. x=±14x=\pm\sqrt{14}.

Check. Both proposed roots have x2=14x^2=14 and x4=196x^4=196, giving 196−182−14=0196-182-14=0. No real number has square −1-1.

A fourth-degree equation can therefore have fewer than four distinct real roots. The transformed equation alone does not decide which values return to real inputs.

Watch out

Stopping at u=4,9u=4,9 has not solved an equation in xx. Write the substitution explicitly, then return to xx; remember both signs when solving x2=x^2= a positive number.

Use a new quadratic in x2x^2.

Returning to the original variable

The substitution tells you how to return to the original variable. If u=xu=\sqrt x, then x=u2x=u^2; if u=1/xu=1/x, then x=1/ux=1/u. These are different return steps even if the quadratic in uu looks identical.

Write the restrictions alongside the substitution. For u=xu=\sqrt x, the original requires x≥0x\ge0, and the square-root symbol requires u≥0u\ge0. For u=1/xu=1/x, x≠0x\ne0 and u≠0u\ne0. A transformed root outside these restrictions cannot represent an original input.

Worked example 3: replacing a square root

Question. Solve 2x−3x+1=02x-3\sqrt x+1=0.

Plan. Put u=x≥0u=\sqrt x\ge0, so x=u2x=u^2, and solve the resulting quadratic.

2u2−3u+1=0(2u−1)(u−1)=0u=12 or 1\begin{aligned} 2u^2-3u+1&=0\\ (2u-1)(u-1)&=0\\ u&=\frac12\text{ or }1 \end{aligned}

Both roots satisfy u≥0u\ge0. Squaring each gives the original input.

Answer. x=1/4x=1/4 or x=1x=1.

Check. At x=1/4x=1/4, 2x−3x+1=1/2−3/2+1=02x-3\sqrt x+1=1/2-3/2+1=0. At x=1x=1, it is 2−3+1=02-3+1=0.

Now use the same quadratic coefficients but replace the repeated expression by a reciprocal. The return step changes from squaring to taking a reciprocal.

Worked example 4: replacing a reciprocal

Question. Solve 2/x2−3/x+1=02/x^2-3/x+1=0.

Plan. Put u=1/x≠0u=1/x\ne0, so 1/x2=u21/x^2=u^2, and solve for uu.

2u2−3u+1=0(2u−1)(u−1)=0u=12 or 1\begin{aligned} 2u^2-3u+1&=0\\ (2u-1)(u-1)&=0\\ u&=\frac12\text{ or }1 \end{aligned}

Both values are nonzero. Since x=1/ux=1/u, they give x=2x=2 and x=1x=1, respectively.

Answer. x=1x=1 or x=2x=2.

Check. At x=2x=2, the original is 2/4−3/2+1=02/4-3/2+1=0; at x=1x=1, it is 2−3+1=02-3+1=0. Neither answer makes a denominator zero.

In both examples the intermediate roots were 1/2,11/2,1. The original answers differed because the substitutions differed. This is why the substitution must remain visible throughout the solution.

Watch out

If u=xu=\sqrt x, do not write x=±u2x=\pm u^2: xx is simply u2u^2. The two-sign step belongs to solving x2=ux^2=u, which is a different substitution.

Use a reciprocal substitution in this check.

Rejecting impossible values

Solving the transformed quadratic gives candidates, rather than an automatic list of original solutions. Check the substitution's restrictions before reversing it, then substitute the remaining answers into the original equation.

In particular, u=xu=\sqrt x cannot be negative. Squaring a negative candidate would produce a non-negative xx, but x\sqrt{x} would then be its positive magnitude, not that negative candidate. The original equation exposes the mistake.

Worked example 5: rejecting a negative square-root candidate

Question. Solve x+x−6=0x+\sqrt x-6=0.

Plan. Use u=x≥0u=\sqrt x\ge0, solve for uu, and reject any negative candidate before squaring.

u2+u−6=0(u+3)(u−2)=0u=−3 or 2\begin{aligned} u^2+u-6&=0\\ (u+3)(u-2)&=0\\ u&=-3\text{ or }2 \end{aligned}

Reject u=−3u=-3, because u=x≥0u=\sqrt x\ge0. The valid candidate u=2u=2 gives x=4x=4.

Answer. x=4x=4 only.

Check. 4+4−6=04+\sqrt4-6=0. Squaring the rejected value would give x=9x=9, but 9+9−6=6≠09+\sqrt9-6=6\ne0: it is not a solution.

Change only the sign before the square root. The quadratic's linear coefficient changes, so the accepted and rejected magnitudes exchange roles.

Worked example 6: a different valid candidate

Question. Solve x−x−6=0x-\sqrt x-6=0.

Plan. Use the same non-negative substitution and test the new roots against it.

u2−u−6=0(u−3)(u+2)=0u=3 or −2\begin{aligned} u^2-u-6&=0\\ (u-3)(u+2)&=0\\ u&=3\text{ or }-2 \end{aligned}

Reject u=−2u=-2; u=3u=3 gives x=9x=9.

Answer. x=9x=9 only.

Check. 9−9−6=09-\sqrt9-6=0. The rejected candidate would give x=4x=4, but 4−4−6=−4≠04-\sqrt4-6=-4\ne0.

The original equation is the final authority. A valid return step plus substitution into the original checks that no denominator is zero, no square-root sign has changed and the stated variable is the one you have actually solved for. The same habit will be useful when trigonometric expressions become the replacement variable later.

Problem-solving

Write three lines of decisions: define the substitution and its restrictions; solve for the new variable; return to the original variable and check. Do not square a forbidden value merely to obtain a real-looking answer.

The opening fourth-power equation needed only a quadratic in x2x^2. Its four real answers came from two permitted positive values of that square, each giving two signs.

Quick Recap

  • A repeated expression can hide a quadratic; choose it as a new variable.
  • State its restrictions, such as x2≥0x^2\ge0, x≥0\sqrt x\ge0 or 1/x≠01/x\ne0.
  • Reverse the actual substitution: square, take square roots or take reciprocals as appropriate.
  • Check the original equation after rejecting impossible transformed values.

The opening equation has x2=4x^2=4 or 99, so its real roots are −3,−2,2,3-3,-2,2,3.