Unit 2, lesson 26

Rational Zeros Theorem and Factoring

How the coefficients of a polynomial with integer coefficients limit its possible rational zeros to a short list, how to test that list quickly, and why the list can miss irrational and complex zeros.

  • Theory part 51
  • Practice part 52
  • Polynomial Functions

A cubic like 2x3+x2−13x+62x^3 + x^2 - 13x + 6 has no obvious factoring pattern, and guessing zeros at random could take forever. Its coefficients, though, already narrow things down. If it has a rational zero, written as a reduced fraction, the numerator must divide 66 and the denominator must divide 22. That leaves a short list to check, and the Factor Theorem does the rest.

Key terms

Reduced fraction

Definition A rational number is a number of the form pq\frac{p}{q} with integers pp and qq and q≠0q \ne 0. The fraction is in lowest terms, or reduced, when q>0q > 0 and pp and qq have no common positive divisor other than 11. Every rational number has exactly one such reduced form. An integer mm has the reduced form m1\frac{m}{1}.

In plain words Cancel every common factor and keep the minus sign in the numerator. What is left is the one standard way to write that number.

Example The fraction 46\frac{4}{6} reduces to 23\frac{2}{3}, the fraction 9−12\frac{9}{-12} reduces to −34\frac{-3}{4}, and the integer 55 is 51\frac{5}{1}.

The Rational Zeros Theorem

Definition Let

P(x)=anxn+an−1xn−1+⋯+a1x+a0\begin{aligned} P(x) &= a_n x^n + a_{n-1} x^{n-1} \\[0.4em] &\quad + \cdots + a_1 x + a_0 \end{aligned}

have integer coefficients, with n≥1n \ge 1, an≠0a_n \ne 0 and a0≠0a_0 \ne 0. If the reduced fraction pq\frac{p}{q} is a zero of PP, then pp divides a0a_0 and qq divides ana_n.

The reason fits in a few lines. Multiplying P(p/q)=0P(p/q) = 0 by qnq^n gives

anpn+an−1pn−1q+⋯+a1pqn−1+a0qn=0.\begin{aligned} &a_n p^n + a_{n-1} p^{n-1} q + \cdots \\[0.4em] &\quad + a_1 p q^{n-1} + a_0 q^n = 0. \end{aligned}

Every term except the last contains the factor pp, so pp divides a0qna_0 q^n. Since pp and qq have no common prime factor, neither do pp and qnq^n, so pp divides a0a_0. In the same way every term except the first contains qq, so qq divides anpna_n p^n and therefore ana_n.

In plain words A rational zero cannot be just any fraction. Its numerator must divide the constant term and its denominator must divide the leading coefficient. The fraction has to be reduced first, or the test says nothing.

Example The polynomial 6x2−x−2=(3x−2)(2x+1)6x^2 - x - 2 = (3x - 2)(2x + 1) has the zeros 23\frac{2}{3} and −12-\frac{1}{2}. In reduced form the numerators 22 and −1-1 divide −2-2, and the denominators 33 and 22 divide 66, as promised. Written as 46\frac{4}{6}, the same zero has numerator 44, which does not divide −2-2. That is no contradiction, because 46\frac{4}{6} is not reduced.

Factors of the constant and leading coefficients

Definition For integers d≠0d \ne 0 and mm, the number dd divides mm when m=dkm = dk for some integer kk. For the theorem you need all divisors of the constant term a0a_0 and of the leading coefficient ana_n. A nonzero integer mm has finitely many divisors, and they come in pairs ±d\pm d where dd runs through the positive divisors of ∣m∣\lvert m \rvert. Two cases need a step first.

  1. If a0=0a_0 = 0, then xx is a factor. Write P(x)=xkR(x)P(x) = x^k R(x) with R(0)≠0R(0) \ne 0 and apply the theorem to RR.
  2. If every coefficient is rational but some are not integers, multiply PP by a common denominator of the coefficients. This gives a polynomial with integer coefficients and the same zeros.

In plain words List the positive divisors of the last coefficient and of the first coefficient. Signs are handled once, at the end, with a ±\pm in front of each candidate.

Example For 2x3+x2−13x+62x^3 + x^2 - 13x + 6 the constant term 66 has positive divisors 1,2,3,61, 2, 3, 6, and the leading coefficient 22 has positive divisors 1,21, 2. For x4−5x3+6x2x^4 - 5x^3 + 6x^2 the constant term is 00, so first write it as x2(x2−5x+6)x^2(x^2 - 5x + 6) and work with x2−5x+6x^2 - 5x + 6.

Rational root candidates

Definition Under the hypotheses of the theorem, the rational root candidates of PP are all numbers

±pq,p a positive divisor of a0,q a positive divisor of an,\begin{gathered} \pm \frac{p}{q}, \\[0.8em] p \text{ a positive divisor of } a_0, \\[0.4em] q \text{ a positive divisor of } a_n, \end{gathered}

with repeated values listed once. Every rational zero of PP is on this list. The list is finite.

In plain words Form every fraction with a numerator from the first list and a denominator from the second, put ±\pm in front, and cross out duplicates.

Example For P(x)=2x3+x2−13x+6P(x) = 2x^3 + x^2 - 13x + 6, the denominator 11 gives ±1,±2,±3,±6\pm 1, \pm 2, \pm 3, \pm 6. The denominator 22 gives ±12,±1,±32,±3\pm \frac{1}{2}, \pm 1, \pm \frac{3}{2}, \pm 3, and only ±12\pm \frac{1}{2} and ±32\pm \frac{3}{2} are new. So there are twelve candidates,

±1,±2,±3,±6,±12,±32.\begin{gathered} \pm 1, \quad \pm 2, \quad \pm 3, \quad \pm 6, \\[0.4em] \pm \frac{1}{2}, \quad \pm \frac{3}{2}. \end{gathered}

Testing candidates

Definition A candidate rr is a zero of PP exactly when P(r)=0P(r) = 0. By the Remainder Theorem this is the same as the remainder of P(x)P(x) divided by x−rx - r being 00, so either direct evaluation or synthetic division decides it. Two values are quick to get. The value P(1)P(1) is the sum of the coefficients, and P(−1)P(-1) is the same sum after switching the sign of each coefficient of an odd power of xx.

In plain words Try the candidates one by one and stop at the first one that gives zero. Start with the easy ones.

Example For P(x)=2x3+x2−13x+6P(x) = 2x^3 + x^2 - 13x + 6,

CandidateValue of PPZero?
112+1−13+6=−42 + 1 - 13 + 6 = -4no
−1-1−2+1+13+6=18-2 + 1 + 13 + 6 = 18no
2216+4−26+6=016 + 4 - 26 + 6 = 0yes

So 22 is a zero and x−2x - 2 is a factor.

Synthetic division and the depressed polynomial

Definition Synthetic division divides P(x)P(x) by x−rx - r using the coefficients alone, with 00 for each missing power. Bring down the leading coefficient, then repeatedly multiply the newest bottom entry by rr and add the product to the next coefficient. The last bottom entry is the remainder P(r)P(r), and the entries before it are the coefficients of the quotient. When rr is a zero the remainder is 00, so

P(x)=(x−r) Q(x),P(x) = (x - r)\,Q(x),

where the quotient QQ, called the depressed polynomial, has degree one less than PP. The remaining zeros of PP are exactly the zeros of QQ, so the search continues on QQ. Every rational zero of QQ is also a rational zero of PP, so it is already on the list for PP, and candidates that failed for PP can be skipped. A zero can repeat, so test rr itself again on QQ. When QQ reaches degree 22, finish with factoring or the quadratic formula.

In plain words Each zero you find shrinks the problem by one degree. Keep dividing until a quadratic is left.

Example Continue with 2x3+x2−13x+62x^3 + x^2 - 13x + 6 and the zero 22.

221−136410−625−30\begin{array}{r|rrrr} 2 & 2 & 1 & -13 & 6 \\ & & 4 & 10 & -6 \\ \hline & 2 & 5 & -3 & 0 \end{array}

The quotient is 2x2+5x−3=(2x−1)(x+3)2x^2 + 5x - 3 = (2x - 1)(x + 3), so

2x3+x2−13x+6=(x−2)(2x−1)(x+3),\begin{aligned} &2x^3 + x^2 - 13x + 6 \\[0.4em] &\quad = (x - 2)(2x - 1)(x + 3), \end{aligned}

and the zeros are 22, 12\frac{1}{2} and −3-3.

A repeated zero shows up when the same candidate works twice. For x4−3x3−3x2+11x−6x^4 - 3x^3 - 3x^2 + 11x - 6 the candidate 11 works, and the quotient is x3−2x2−5x+6x^3 - 2x^2 - 5x + 6. The candidate 11 works again on that quotient, leaving x2−x−6=(x−3)(x+2)x^2 - x - 6 = (x - 3)(x + 2). So

x4−3x3−3x2+11x−6=(x−1)2(x−3)(x+2).\begin{aligned} &x^4 - 3x^3 - 3x^2 + 11x - 6 \\[0.4em] &\quad = (x - 1)^2 (x - 3)(x + 2). \end{aligned}

The theorem gives candidates, not all roots

Definition The Rational Zeros Theorem says only that every rational zero is on the candidate list. It does not say that any candidate is a zero, and it says nothing about zeros that are irrational or not real. A polynomial with integer coefficients may have no rational zeros at all, even when it has real zeros.

In plain words The list is a set of suspects, not a list of answers. When it runs out, the remaining zeros have to come from the quadratic formula, from a graph or from a numerical method.

Example For P(x)=2x3−4x2−7x+3P(x) = 2x^3 - 4x^2 - 7x + 3 the candidates are ±1,±3,±12,±32\pm 1, \pm 3, \pm \frac{1}{2}, \pm \frac{3}{2}, and only 33 works. Dividing by x−3x - 3 leaves 2x2+2x−12x^2 + 2x - 1, whose zeros are

x=−2±4+84=−1±32.x = \frac{-2 \pm \sqrt{4 + 8}}{4} = \frac{-1 \pm \sqrt{3}}{2}.

These two zeros are irrational, so they were never on the list. As another case, the only candidates for x2−2x^2 - 2 are ±1\pm 1 and ±2\pm 2, and none of them is a zero. So x2−2x^2 - 2 has no rational zeros, which proves that 2\sqrt{2} is irrational.

Check yourself

  1. List the rational root candidates of P(x)=3x3+x2−8x+4P(x) = 3x^3 + x^2 - 8x + 4.
Show the answer

The positive divisors of 44 are 1,2,41, 2, 4 and those of 33 are 1,31, 3. The twelve candidates are

±1,±2,±4,±13,±23,±43.\begin{gathered} \pm 1, \quad \pm 2, \quad \pm 4, \\[0.4em] \pm \frac{1}{3}, \quad \pm \frac{2}{3}, \quad \pm \frac{4}{3}. \end{gathered}
  1. Factor P(x)=3x3+x2−8x+4P(x) = 3x^3 + x^2 - 8x + 4 completely.
Show the answer

The sum of the coefficients is 3+1−8+4=03 + 1 - 8 + 4 = 0, so 11 is a zero. Synthetic division by x−1x - 1 gives the quotient 3x2+4x−4=(3x−2)(x+2)3x^2 + 4x - 4 = (3x - 2)(x + 2), so

3x3+x2−8x+4=(x−1)(3x−2)(x+2).\begin{aligned} &3x^3 + x^2 - 8x + 4 \\[0.4em] &\quad = (x - 1)(3x - 2)(x + 2). \end{aligned}
  1. Explain why P(x)=x3+x+1P(x) = x^3 + x + 1 has no rational zeros, even though it has a real zero.
Show the answer

The only candidates are 11 and −1-1, and P(1)=3P(1) = 3 and P(−1)=−1P(-1) = -1 are not zero. Since P(−1)=−1<0<1=P(0)P(-1) = -1 < 0 < 1 = P(0) and the graph of a polynomial has no breaks, it crosses the xx-axis between −1-1 and 00. That real zero is irrational.

How this lesson connects