Solve for :
Substitute so the radicand factors. The expression is a difference of squares in . With (so and ) the equation becomes
Writing as is what makes the left side factor — that is the whole reason this equation has a clean root.
Fix the domain and kill one branch immediately. The radical needs , i.e. or .
If , the left side is negative while the right side has a positive factor times a non-negative root, so the right side is . No solution there — which also rules out every .
So , i.e. . On that branch the left side is , so we also need : the search narrows to .
Set and read off the boundary root. Then and
Taking satisfies this exactly: both sides are . That gives and
Check it in the original equation: . ✓ This is the root that a purely numerical solver most often misses, because it sits exactly on the edge of the domain where the square root turns on.
Handle by squaring, then divide out the factor . Squaring gives , and dividing by :
Its three roots are , and . Only the tiny positive one satisfies ; the other two make negative while is positive (or vice versa), so squaring introduced them.
Convert back and verify both answers numerically. From :
Substituting into the original equation gives , residual to machine precision. Together with the boundary root:
Why the popular shortcut answer is wrong. A common route squares first and then tries to factor the quartic as . Expanding that product gives an coefficient of , not , so the factorization is simply false and the values and it produces are not roots. Substituting into the original equation returns about , not ; and the companion value is not even in the domain, since .
Need to solve a different problem like this? Open the solver →