Unit 3, lesson 43

Solving Logarithmic Equations

How to solve equations with the unknown inside a logarithm by rewriting them in exponential form, and why every candidate has to be checked against the domain of the original equation.

  • Theory part 85
  • Practice part 86
  • Exponential and Logarithmic Equations

If the pH of a solution is 4.5, its hydrogen ion concentration xx in moles per liter satisfies −log⁡x=4.5-\log x = 4.5. From log⁡x=−4.5\log x = -4.5 the definition of the logarithm gives x=10−4.5≈3.2×10−5x = 10^{-4.5} \approx 3.2 \times 10^{-5}. Many logarithmic equations are this quick. The catch is that the usual steps can also produce numbers that do not belong, so every answer needs a check.

Key terms

Exponential form conversion

Definition Let b>0b > 0 with b≠1b \ne 1, let cc be a real number and let AA be an expression in xx. Then

log⁡bA=cexactly whenA=bc.\begin{gathered} \log_b A = c \\[0.4em] \text{exactly when} \quad A = b^c. \end{gathered}

This is the definition of the logarithm, so the conversion is an equivalent step. Since bc>0b^c > 0, every solution of A=bcA = b^c automatically makes the argument AA positive.

In plain words Once a single logarithm equals a number, rewrite the equation as a power. The base of the logarithm becomes the base of the power, and the number on the other side becomes the exponent.

Example The equation log⁡3(2x+1)=4\log_3(2x + 1) = 4 becomes 2x+1=34=812x + 1 = 3^4 = 81, so x=40x = 40. As a check, log⁡381=4\log_3 81 = 4. In the same way ln⁡(x−2)=1.5\ln(x - 2) = 1.5 becomes x−2=e1.5x - 2 = e^{1.5}, so

x=2+e1.5≈6.4817.x = 2 + e^{1.5} \approx 6.4817.

Domain validation

Definition The domain of a logarithmic equation is the set of real numbers xx at which every part of the original equation is defined. For each logarithm this means that its argument is positive and its base is positive and not 11. Domain validation means finding this set before solving and accepting a candidate only if it lies in the set and satisfies the original equation.

In plain words Write down the allowed values of xx first. A number outside them is never a solution, however it was found.

Example In log⁡2(x−1)+log⁡2(x+2)=2\log_2(x - 1) + \log_2(x + 2) = 2 the arguments need x−1>0x - 1 > 0 and x+2>0x + 2 > 0. Both hold exactly when x>1x > 1, so the domain is (1,∞)(1, \infty). Bases count too. In log⁡x(12−x)=2\log_x(12 - x) = 2 the base needs x>0x > 0 and x≠1x \ne 1, and the argument needs x<12x < 12. Converting gives x2=12−xx^2 = 12 - x, with roots 33 and −4-4, and only 33 lies in the domain. Indeed log⁡39=2\log_3 9 = 2.

Equal logarithms

Definition Let b>0b > 0 with b≠1b \ne 1. For positive AA and BB,

log⁡bA=log⁡bBexactly whenA=B,\begin{gathered} \log_b A = \log_b B \\[0.4em] \text{exactly when} \quad A = B, \end{gathered}

because log⁡b\log_b is one-to-one on (0,∞)(0, \infty). The equation A=BA = B by itself does not require AA and BB to be positive, so its solutions are only candidates until the domain is checked.

In plain words When one logarithm equals another with the same base, drop the logarithms and set the arguments equal. Then make sure both arguments are positive at each answer.

Example Solve ln⁡(x2−3)=ln⁡(2x)\ln\left(x^2 - 3\right) = \ln(2x). Setting the arguments equal gives x2−2x−3=0x^2 - 2x - 3 = 0, so x=3x = 3 or x=−1x = -1. At x=−1x = -1 both arguments equal −2-2. They match, yet neither logarithm exists, so −1-1 is rejected. At x=3x = 3 both sides are ln⁡6\ln 6, so the only solution is x=3x = 3.

Condensing first

Definition When an equation contains several logarithms with the same base, the product, quotient and power rules combine them into a single logarithm, which can then be converted to exponential form. For M>0M > 0 and N>0N > 0,

log⁡bM+log⁡bN=log⁡b(MN),log⁡bM−log⁡bN=log⁡bMN.\begin{aligned} \log_b M + \log_b N &= \log_b(MN), \\[0.4em] \log_b M - \log_b N &= \log_b \frac{M}{N}. \end{aligned}

The combined logarithm log⁡b(MN)\log_b(MN) is defined whenever MN>0MN > 0, which also allows M<0M < 0 and N<0N < 0, and the same is true of log⁡bMN\log_b \frac{M}{N}. So condensing can enlarge the domain.

In plain words Gather the logarithms into one, rewrite it as a power and solve. Then go back to the domain of the original equation, not the condensed one.

Example Return to log⁡2(x−1)+log⁡2(x+2)=2\log_2(x - 1) + \log_2(x + 2) = 2, with domain x>1x > 1. Condensing and converting give

(x−1)(x+2)=22=4,x2+x−6=0,(x−2)(x+3)=0.\begin{gathered} (x - 1)(x + 2) = 2^2 = 4, \\[0.4em] x^2 + x - 6 = 0, \\[0.4em] (x - 2)(x + 3) = 0. \end{gathered}

The candidate x=2x = 2 lies in the domain, and log⁡21+log⁡24=0+2=2\log_2 1 + \log_2 4 = 0 + 2 = 2, so it is a solution. The candidate x=−3x = -3 lies outside the domain.

Extraneous solutions

Definition An extraneous solution is a number produced by the solving steps that satisfies a rewritten equation but not the original one. With logarithms the usual cause is a step that enlarges the domain, such as condensing log⁡bM+log⁡bN\log_b M + \log_b N into log⁡b(MN)\log_b(MN) or rewriting 2log⁡bM2 \log_b M as log⁡b(M2)\log_b\left(M^2\right), which is defined for every M≠0M \ne 0. Such a step keeps every solution but may add new ones outside the original domain.

In plain words An extraneous candidate fits the simplified equation, so the algebra gives no warning. Only a check in the original equation catches it.

Example At x=−3x = -3 the condensed equation above is true, since (−4)(−1)=4(-4)(-1) = 4, but the original equation needs log⁡2(−4)\log_2(-4), which is undefined. On a graph, y=log⁡2((x−1)(x+2))y = \log_2\big((x - 1)(x + 2)\big) has two branches, one for x<−2x < -2 and one for x>1x > 1, and both meet the line y=2y = 2. The original left side has only the right branch, so it meets the line once. The power rule can cause the same thing. In 2log⁡x=log⁡(3x+4)2 \log x = \log(3x + 4) rewriting the left side as log⁡(x2)\log\left(x^2\right) leads to x2−3x−4=0x^2 - 3x - 4 = 0, so x=4x = 4 or x=−1x = -1. The original equation needs x>0x > 0, so −1-1 is extraneous, and x=4x = 4 checks, since 2log⁡4=log⁡162 \log 4 = \log 16.

Steps that lose solutions

Definition A step that shrinks the domain can lose solutions, and a lost solution never comes back in the check. The main example is expanding with the power rule, replacing log⁡b(x2)\log_b\left(x^2\right), which is defined for every x≠0x \ne 0, by 2log⁡bx2 \log_b x, which is defined only for x>0x > 0. The safe form is

log⁡b(x2)=2log⁡b∣x∣,x≠0.\log_b\left(x^2\right) = 2 \log_b \lvert x \rvert, \qquad x \ne 0.

In plain words Extra candidates can be checked and thrown away, but missing ones cannot be checked at all. Keep the original form, or use absolute values, when a step would cut the domain down.

Example The equation ln⁡(x2)=2\ln\left(x^2\right) = 2 converts to x2=e2x^2 = e^2, so x=ex = e or x=−ex = -e. Both work, since (−e)2=e2(-e)^2 = e^2. Rewriting the equation first as 2ln⁡x=22 \ln x = 2 gives only x=ex = e and loses the solution −e≈−2.718-e \approx -2.718.

Different bases in one equation

Definition The product, quotient and power rules combine logarithms only when they share a base. If the bases differ, rewrite the logarithms in one base with the change of base formula

log⁡bx=log⁡axlog⁡ab,\log_b x = \frac{\log_a x}{\log_a b},

valid for positive bases aa and bb other than 11 and for x>0x > 0.

In plain words Bring every logarithm to one base, often the smallest base in the equation, and then solve as before.

Example Solve log⁡4x=log⁡2(x−2)\log_4 x = \log_2(x - 2), whose domain is x>2x > 2. Change of base gives

log⁡4x=log⁡2xlog⁡24=12log⁡2x,\log_4 x = \frac{\log_2 x}{\log_2 4} = \frac{1}{2} \log_2 x,

so the equation becomes

log⁡2x=2log⁡2(x−2)=log⁡2((x−2)2)\begin{aligned} \log_2 x &= 2 \log_2(x - 2) \\[0.4em] &= \log_2\big((x - 2)^2\big) \end{aligned}

and then x=(x−2)2x = (x - 2)^2. That is x2−5x+4=0x^2 - 5x + 4 = 0, with roots 11 and 44. The root 11 lies outside the domain. It slipped in when 2log⁡2(x−2)2 \log_2(x - 2) became log⁡2((x−2)2)\log_2\big((x - 2)^2\big), a step that is correct for x>2x > 2 but also allows x<2x < 2. The root 44 checks, since log⁡44=1=log⁡22\log_4 4 = 1 = \log_2 2.

Check yourself

  1. Solve log⁡5(3x−1)=2\log_5(3x - 1) = 2.
Show the answer

Converting gives 3x−1=52=253x - 1 = 5^2 = 25, so x=263x = \frac{26}{3}. At this value the argument is 2525, which is positive, and log⁡525=2\log_5 25 = 2.

  1. Solve log⁡x+log⁡(x−21)=2\log x + \log(x - 21) = 2.
Show the answer

The domain is x>21x > 21. Condensing and converting give x(x−21)=100x(x - 21) = 100, so x2−21x−100=0x^2 - 21x - 100 = 0 and (x−25)(x+4)=0(x - 25)(x + 4) = 0. The candidate −4-4 lies outside the domain. The solution is x=25x = 25, and indeed log⁡25+log⁡4=log⁡100=2\log 25 + \log 4 = \log 100 = 2.

  1. Solve log⁡(x2)=4\log\left(x^2\right) = 4.
Show the answer

Converting gives x2=104x^2 = 10^4, so x=100x = 100 or x=−100x = -100. Both are in the domain x≠0x \ne 0, and both satisfy the equation. Rewriting the left side as 2log⁡x2 \log x first would lose the solution −100-100.

How this lesson connects

Where this leads in Calculus