Unit 2, lesson 30

Vertical Asymptotes and Holes

How to tell whether an excluded input of a rational function gives a vertical asymptote or a hole, and how signs show which way the graph runs on each side of an asymptote.

  • Theory part 59
  • Practice part 60
  • Rational Functions

Give the function 1x−1\frac{1}{x - 1} the inputs 1.11.1, 1.011.01 and 1.0011.001, and the outputs are 1010, 100100 and 10001000. Near some excluded inputs a rational function runs away like this, while near others it stays calm and the graph only loses a single point. One factorization tells the two cases apart, and the difference shapes the whole graph.

Key terms

One-sided unbounded behavior

Definition Let ff be defined on an open interval (a,a+d)(a, a + d) for some d>0d > 0. The statement

f(x)→∞asx→a+f(x) \to \infty \quad \text{as} \quad x \to a^+

means that for every number KK there is a number δ\delta with 0<δ≤d0 < \delta \le d such that f(x)>Kf(x) > K whenever a<x<a+δa < x < a + \delta. The statement f(x)→−∞f(x) \to -\infty as x→a+x \to a^+ means the same with f(x)<Kf(x) < K in place of f(x)>Kf(x) > K. For the left side, written x→a−x \to a^-, use an interval (a−d,a)(a - d, a) and the inputs with a−δ<x<aa - \delta < x < a. In limit notation these statements read lim⁡x→a+f(x)=∞\lim_{x \to a^+} f(x) = \infty, lim⁡x→a−f(x)=−∞\lim_{x \to a^-} f(x) = -\infty and so on.

In plain words As xx creeps toward aa from one side, the outputs eventually pass any bound you name and stay past it. Each side is checked on its own, and the two sides can head in different directions.

Example Here are some values of f(x)=1x−1f(x) = \frac{1}{x - 1}.

xx0.90.90.990.990.9990.9991.0011.0011.011.011.11.1
f(x)f(x)−10-10−100-100−1000-1000100010001001001010

The table only suggests the pattern. To prove the right side, take any K>0K > 0, which is enough because a δ\delta that works for KK also works for every smaller bound. If 1<x<1+1K1 < x < 1 + \frac{1}{K}, then 0<x−1<1K0 < x - 1 < \frac{1}{K}, and so f(x)>Kf(x) > K. Thus f(x)→∞f(x) \to \infty as x→1+x \to 1^+. The same argument with the signs reversed gives f(x)→−∞f(x) \to -\infty as x→1−x \to 1^-.

Vertical asymptote

Definition The line x=ax = a is a vertical asymptote of the graph of ff if at least one of the following four statements is true.

lim⁡x→a−f(x)=∞,lim⁡x→a−f(x)=−∞,lim⁡x→a+f(x)=∞,lim⁡x→a+f(x)=−∞.\begin{gathered} \lim_{x \to a^-} f(x) = \infty, \\[0.4em] \lim_{x \to a^-} f(x) = -\infty, \\[0.4em] \lim_{x \to a^+} f(x) = \infty, \\[0.4em] \lim_{x \to a^+} f(x) = -\infty. \end{gathered}

For a rational function, aa is then an excluded value, so the graph never meets the line x=ax = a.

In plain words Near the line the graph runs up or down along it without bound, on at least one side. The line is a guide for the eye, not part of the graph.

Example For g(x)=3x+4g(x) = \frac{3}{x + 4} the numerator is always 33, while the denominator is a tiny positive number just to the right of −4-4 and a tiny negative number just to the left. So g(x)→∞g(x) \to \infty as x→−4+x \to -4^+ and g(x)→−∞g(x) \to -\infty as x→−4−x \to -4^-, and the line x=−4x = -4 is a vertical asymptote. For instance g(−3.99)=300g(-3.99) = 300 and g(−4.01)=−300g(-4.01) = -300.

Denominator zero after cancellation

Definition Let r=pqr = \frac{p}{q} be a rational function with reduced form PQ\frac{P}{Q}, so that PP and QQ share no factor of positive degree. If Q(a)=0Q(a) = 0 for a real number aa, then P(a)≠0P(a) \ne 0, because otherwise x−ax - a would be a common factor. Near aa the numerator stays close to the nonzero number P(a)P(a) while the denominator approaches 00. The denominator has only finitely many zeros, so it keeps one sign just left of aa and one sign just right of aa. Hence on each side of aa the values r(x)r(x) tend to ∞\infty or to −∞-\infty, and x=ax = a is a vertical asymptote. If Q(a)≠0Q(a) \ne 0, the values r(x)r(x) approach the finite number P(a)Q(a)\frac{P(a)}{Q(a)} and there is no vertical asymptote at aa. So the vertical asymptotes of rr are exactly the lines x=ax = a where aa is a real zero of the reduced denominator QQ.

In plain words Cancel first, then look at the bottom. A zero that survives the cancellation gives a vertical asymptote, and a zero that cancels away does not.

Example Take

f(x)=x2+x−2x2−3x+2=(x+2)(x−1)(x−2)(x−1).\begin{aligned} f(x) &= \frac{x^2 + x - 2}{x^2 - 3x + 2} \\[0.8em] &= \frac{(x + 2)(x - 1)}{(x - 2)(x - 1)}. \end{aligned}

The original domain excludes 11 and 22. After cancelling x−1x - 1 the reduced form is x+2x−2\frac{x + 2}{x - 2}, whose denominator is zero only at 22. So x=2x = 2 is the only vertical asymptote. Near 22 the numerator is close to 44 while the denominator shrinks, and indeed f(1.99)=−399f(1.99) = -399 and f(2.01)=401f(2.01) = 401.

Removable hole

Definition With the same notation, let aa be excluded from the domain of rr while Q(a)≠0Q(a) \ne 0. Then the graph of rr has a removable hole at the point

(a,P(a)Q(a)).\left(a, \frac{P(a)}{Q(a)}\right).

The values of rr stay bounded near aa and approach the height of the hole from both sides. A vertical asymptote and a hole can never occur at the same aa, since one needs Q(a)=0Q(a) = 0 and the other needs Q(a)≠0Q(a) \ne 0.

In plain words At a hole the graph is calm on both sides and only one point is missing. At a vertical asymptote the graph runs off the page.

Example The function ff above has a hole at x=1x = 1, at the height 1+21−2=−3\frac{1 + 2}{1 - 2} = -3. The values near 11, rounded to four decimals, settle toward −3-3 from both sides.

xx0.90.90.990.99111.011.011.11.1
f(x)f(x)−2.6364-2.6364−2.9604-2.9604undefined−3.0404-3.0404−3.4444-3.4444

Compare them with the values next to x=2x = 2, which are in the hundreds. The hole is the point (1,−3)(1, -3).

Sign analysis near an excluded value

Definition Let x=ax = a be a vertical asymptote of rr and write its reduced form as

P(x)Q(x)=P(x)(x−a)m S(x),\frac{P(x)}{Q(x)} = \frac{P(x)}{(x - a)^m \, S(x)},

where m≥1m \ge 1 is the multiplicity of aa as a zero of QQ and S(a)≠0S(a) \ne 0. Polynomials have no jumps, so PP and SS keep the signs of P(a)P(a) and S(a)S(a) on a small interval around aa. On that interval the sign of r(x)r(x) is the sign of P(a)S(a)\frac{P(a)}{S(a)} times the sign of (x−a)m(x - a)^m.

  1. If mm is odd, the factor (x−a)m(x - a)^m is negative to the left of aa and positive to the right, so the two sides go in opposite directions, one toward ∞\infty and the other toward −∞-\infty.
  2. If mm is even, the factor (x−a)m(x - a)^m is positive on both sides, so both sides go toward ∞\infty when P(a)S(a)>0\frac{P(a)}{S(a)} > 0 and both go toward −∞-\infty when P(a)S(a)<0\frac{P(a)}{S(a)} < 0.

In plain words Close to the asymptote only the signs matter. Find the sign of each factor just left and just right of aa, multiply them, and the result tells you whether that side of the graph shoots up or down.

Example In the reduced form x+2x−2\frac{x + 2}{x - 2} of ff the numerator is positive near 22, and 22 is a zero of multiplicity 11 of the denominator. Just left of 22 the denominator is negative, so f(x)→−∞f(x) \to -\infty as x→2−x \to 2^-. Just right of 22 it is positive, so f(x)→∞f(x) \to \infty as x→2+x \to 2^+. For an even case take

g(x)=−x(x−3)2.g(x) = \frac{-x}{(x - 3)^2}.

Near 33 the numerator is close to −3-3 and the denominator is positive on both sides, so g(x)→−∞g(x) \to -\infty as x→3−x \to 3^- and as x→3+x \to 3^+. For instance g(2.99)=−29900g(2.99) = -29900 and g(3.01)=−30100g(3.01) = -30100.

Infinity is not a function value

Definition The symbols ∞\infty and −∞-\infty do not stand for real numbers. The statement f(x)→∞f(x) \to \infty as x→a+x \to a^+ describes the values of ff at inputs near aa and says nothing about f(a)f(a). At a vertical asymptote of a rational function, aa is not in the domain, so f(a)f(a) is undefined, and writing f(a)=∞f(a) = \infty or 10=∞\frac{1}{0} = \infty is incorrect. A one-sided limit that equals ∞\infty or −∞-\infty does not exist as a real number, and the notation only records the way in which it fails to exist.

In plain words Infinity is a direction the outputs head in, not a number they reach or a number you can plug in. At the asymptote the function simply has no value.

Example Let h(x)=1x2h(x) = \frac{1}{x^2}. Every output is a positive real number, and h(0)h(0) is undefined. Still h(x)→∞h(x) \to \infty as x→0x \to 0 from either side, which means for instance that h(x)>106h(x) > 10^6 whenever 0<∣x∣<0.0010 < \lvert x \rvert < 0.001, since then x2<10−6x^2 < 10^{-6}. The range of hh is the interval (0,∞)(0, \infty), which contains no number called ∞\infty.

Check yourself

  1. Find the vertical asymptotes and the holes of the function below.
r(x)=x2−9x2+x−6r(x) = \frac{x^2 - 9}{x^2 + x - 6}
Show the answer

Factoring gives

r(x)=(x−3)(x+3)(x−2)(x+3),r(x) = \frac{(x - 3)(x + 3)}{(x - 2)(x + 3)},

so the original domain excludes 22 and −3-3. The reduced form is x−3x−2\frac{x - 3}{x - 2}, whose denominator is zero at 22, so x=2x = 2 is a vertical asymptote. The factor x+3x + 3 cancels completely and the reduced form equals −6−5=65\frac{-6}{-5} = \frac{6}{5} at −3-3, so the graph has a hole at (−3,65)(-3, \frac{6}{5}).

  1. Describe the behavior of the function below on both sides of each vertical asymptote.
r(x)=x−4x(x−2)2r(x) = \frac{x - 4}{x(x - 2)^2}
Show the answer

The numerator is not zero at 00 or at 22, so nothing cancels and both x=0x = 0 and x=2x = 2 are vertical asymptotes. Near 00 the numerator is close to −4-4 and (x−2)2(x - 2)^2 is close to 44, so r(x)r(x) has the sign of −x-x. Hence r(x)→∞r(x) \to \infty as x→0−x \to 0^- and r(x)→−∞r(x) \to -\infty as x→0+x \to 0^+. Near 22 the numerator is close to −2-2, the factor xx is close to 22 and (x−2)2(x - 2)^2 is positive, so r(x)→−∞r(x) \to -\infty on both sides of x=2x = 2.

  1. A classmate says that the function below has a hole at x=1x = 1 because its numerator is 00 there. Is that right?
g(x)=x2−1(x−1)2g(x) = \frac{x^2 - 1}{(x - 1)^2}
Show the answer

No. Cancelling one factor x−1x - 1 gives g(x)=x+1x−1g(x) = \frac{x + 1}{x - 1}, and the reduced denominator is still 00 at x=1x = 1. So x=1x = 1 is a vertical asymptote. The reduced numerator is close to 22 there, so g(x)→−∞g(x) \to -\infty as x→1−x \to 1^- and g(x)→∞g(x) \to \infty as x→1+x \to 1^+.

How this lesson connects