In the previous lesson we got stuck on 2ˣ = 5: 2² = 4 is too small, 2³ = 8 too big, so x lies between 2 and 3, yet it cannot be written as a “nice” number. Aysel asks a similar question: at 10% a year, how many years does it take her deposit to double? In 1.1ⁿ = 2 the unknown is again in the exponent. For such numbers mathematicians created a new name and symbol, the logarithm: x = log₂5 ≈ 2.32 and n = log₁.₁2 ≈ 7.27.
In this lesson we study the logarithm from its definition: how to compute it, what its properties are, what the graph of y = logₐx looks like and where logarithms are used in real life. Logarithmic equations and inequalities are solved in the next lesson, “Logarithmic equations and inequalities”.
The definition and the basic identity
LogarithmDefinition
The logarithm of a positive number b to base a (logₐb) is the number c such that aᶜ = b; here a > 0 and a ≠ 1. In other words, a logarithm answers the question “to what power must a be raised to get b?” For example, log₂8 = 3, because 2³ = 8.
Where
- a
- the base: a > 0, a ≠ 1
- b
- the argument: b > 0
- c
- the value of the logarithm, any real number (it may be negative)
A logarithm is an exponent. The same equality can be read in two ways: 2³ = 8 ⇔ log₂8 = 3. Special cases: logₐ1 = 0, because a⁰ = 1; logₐa = 1, because a¹ = a.
Put simply, a logarithm answers one question: how many copies of the base must I multiply together to get the given number? log₂8 asks: “how many 2s multiplied together make 8?” 2 · 2 · 2 = 8 — three of them, so log₂8 = 3. In the same way log₃9 = 2, because 3 · 3 = 9, and log₁₀1000 = 3, because 10 · 10 · 10 = 1000.
Work it out by counting
Problem
Evaluate: a) log₅25; b) log₂16; c) log₇7.
Show solutionHide solution
Solution
b) 2 · 2 · 2 · 2 = 16 — four 2s, so log₂16 = 4.
c) One 7 is just 7, so log₇7 = 1. This holds for every base: logₐa = 1.
The reasons come from the exponential function. Since aᶜ is always positive, negative numbers and zero have no logarithm: log₂(−8) and log₂0 make no sense. For a = 1 we have 1ᶜ = 1, so for log₁5, say, no c exists. The key consequence: the expression under a logarithm must always be positive, and that is why we will find domains in the next lesson.
Remember: the number under the logarithm is always positive, and the base is positive and not equal to 1. The reason is simple: whatever power you raise a positive base to, the result is positive — you never get a negative number or zero. And every power of 1 is 1, so 1 cannot be a base.
Evaluating from the definition
Problem
Evaluate:
1) log₂32
2) log₃(1/9)
3) log₅√5
4) log₁₀0.001
5) log₀.₅8
6) log₄8
Show solutionHide solution
Solution
2) 3⁻² = 1/9 ⇒ log₃(1/9) = −2.
3) √5 = 5^(1/2) ⇒ log₅√5 = 1/2.
4) 10⁻³ = 0.001 ⇒ log₁₀0.001 = −3.
5) 0.5⁻³ = 2³ = 8 ⇒ log₀.₅8 = −3.
6) 4 = 2², 8 = 2³: (2²)ᶜ = 2³ ⇒ 2c = 3 ⇒ log₄8 = 3/2.
Common and natural logarithmsDefinition
The logarithm to base 10 is the common logarithm, written log b (some books write lg b): log b = log₁₀b. The logarithm to base e ≈ 2.718 is the natural logarithm: ln b = logₑb. The log and ln keys of a calculator compute exactly these two: log 1000 = 3, log 0.01 = −2, ln e = 1, ln 1 = 0.
Where
- a
- the base: a > 0, a ≠ 1
- b
- a positive number
- c
- any real number
The basic logarithmic identity: raising a to the logarithm of b gives back b. Raising to a power and taking a logarithm “undo” each other, like addition and subtraction.
Using the basic identity
Problem
Evaluate:
1) 2^(log₂7)
2) 3^(2 + log₃5)
3) 10^(1 + log 3)
4) 25^(log₅3)
5) ln e³ and e^(ln 5)
Show solutionHide solution
Solution
2) Split the power into two factors: 3² · 3^(log₃5) = 9 · 5 = 45.
3) 10¹ · 10^(log 3) = 10 · 3 = 30.
4) 25 = 5²: 25^(log₅3) = (5^(log₅3))² = 3² = 9.
5) ln e³ = 3, e^(ln 5) = 5.
Many logarithms are irrational numbers: no fraction equals them exactly, and a calculator only shows an approximation. The classic way to prove this is to assume the opposite and reach a contradiction through parity or divisibility.
log₂3 is irrational
Problem
Prove that log₂3 is not a rational number.
Show solutionHide solution
Solution
The same method shows that lg 2 is irrational: in 10ᵖ = 2^q the left side is divisible by 5 and the right side is not.
Laws of logarithms
A logarithm is an exponent, so its laws follow from the laws of powers. Let x = aᵐ and y = aⁿ, that is, m = logₐx and n = logₐy. Then xy = aᵐ⁺ⁿ, x/y = aᵐ⁻ⁿ and xᵖ = a^(pm). Reading these equalities through the definition gives the three main laws: multiplication turns into adding logarithms, division into subtracting them, and raising to a power into multiplying.
You can check these properties by counting. 4 = 2 · 2 (two 2s) and 8 = 2 · 2 · 2 (three 2s). In the product 4 · 8 = 32 the numbers of 2s add up: 2 + 3 = 5, and indeed 2⁵ = 32. So log₂(4 · 8) = log₂4 + log₂8 = 2 + 3 = 5. In a division the 2s cancel and their counts are subtracted: log₂(32 ÷ 4) = 5 − 2 = 3. In short: multiplying becomes adding, dividing becomes subtracting.
Where
- x, y
- positive numbers
The logarithm of a product is the sum of the logarithms of the factors. It also reads from right to left: a sum of logarithms with the same base is the logarithm of the product.
The logarithm of a product
Problem
1) log₆4 + log₆9
2) log 25 + log 4
3) log₃18 + log₃1.5
4) Write log₂(8x) in terms of log₂x (x > 0).
Show solutionHide solution
Solution
2) log(25 · 4) = log 100 = 2.
3) log₃(18 · 1.5) = log₃27 = 3.
4) log₂8 + log₂x = 3 + log₂x.
Where
- x, y
- positive numbers
The logarithm of a quotient is the logarithm of the dividend minus the logarithm of the divisor. Special case: logₐ(1/y) = −logₐy.
The logarithm of a quotient
Problem
1) log₃54 − log₃2
2) log 5000 − log 5
3) log₂24 − log₂3
4) log₅(1/125)
Show solutionHide solution
Solution
2) log(5000 ÷ 5) = log 1000 = 3.
3) log₂(24 ÷ 3) = log₂8 = 3.
4) log₅(1/125) = −log₅125 = −3.
Where
- x
- a positive number
- p
- any real number
- k
- the exponent of the base, k ≠ 0
The exponent of the argument comes out in front as a factor, the exponent of the base comes out as its reciprocal: log₈x = log_(2³) x = ⅓ · log₂x.
The logarithm of a power
Problem
1) log₂(1/32)
2) log₃√27
3) 2 log₅10 − log₅4
4) Write log₉(x²) in terms of log₃x (x > 0).
Show solutionHide solution
Solution
2) √27 = 3^(3/2) ⇒ (3/2) · log₃3 = 3/2.
3) 2 log₅10 = log₅100, then the quotient: log₅(100 ÷ 4) = log₅25 = 2.
4) 9 = 3²: log₉(x²) = ½ · log₃(x²) = ½ · 2 log₃x = log₃x.
Olympiad and hard test problems often ask you to compare two logarithms without a calculator. The main method is an intermediate number: pick a simple number (for example 3/2) such that one logarithm is greater than it and the other is smaller. By the definition, each comparison turns into a comparison of powers.
The intermediate-number method
Problem
Which is greater: log₂3 or log₃5?
Show solutionHide solution
Solution
log₂3 > 3/2 ⇔ 3 > 2^(3/2) = 2√2. Square both sides: 9 > 8 — true.
log₃5 < 3/2 ⇔ 5 < 3^(3/2) = 3√3. Square both sides: 25 < 27 — true.
So log₃5 < 3/2 < log₂3, which means log₂3 is greater (about 1.585 and 1.465).
Change of base
Let c = logₐb, that is, aᶜ = b. Take the logarithm to base k of both sides: c · logₖa = logₖb, so c = logₖb / logₖa. This formula lets us compute any logarithm with the log or ln key of a calculator and bring logarithms with different bases to one base.
Why do we need this formula? A calculator usually has only two logarithm keys: lg (base 10) and ln. To find log₂5, write it as a ratio of two lg values: log₂5 = lg 5 / lg 2 ≈ 0.699 / 0.301 ≈ 2.32. Check: 2^2.32 ≈ 5. So a logarithm to any base is a fraction: the logarithm of the number on top, the logarithm of the base underneath.
Where
- k
- the new base: k > 0, k ≠ 1 (often 10 or e)
- x
- a positive number; in the second formula also x ≠ 1
The change-of-base formula and two of its consequences. On a calculator: logₐx = log x / log a = ln x / ln a.
Changing the base
Problem
1) log₄8
2) log₂5 (with a calculator)
3) log₂3 · log₃4
4) log₈2
5) If log₂3 = a, express log₆12 in terms of a.
Show solutionHide solution
Solution
2) log 5 / log 2 ≈ 0.69897 / 0.30103 ≈ 2.32. Check: 2² = 4 < 5 < 8 = 2³.
3) The “chain”: log₂3 · log₃4 = log₂4 = 2.
4) 1 / log₂8 = 1/3.
5) log₆12 = log₂12 / log₂6 = (log₂4 + log₂3) / (log₂2 + log₂3) = (2 + a) / (1 + a).
Use one base
If the bases differ, bring all of them to one base (often 2, 3, 5 or 10) with the change-of-base formula.
Move coefficients inside
Move a number in front of a logarithm inside as an exponent: 2 log₅10 = log₅100.
Combine
Collect a sum into the logarithm of a product and a difference into the logarithm of a quotient.
Evaluate and check
Evaluate the result from the definition or with the basic identity, then check the answer with an estimate.
Exam-style expressions
Problem
Evaluate:
1) (log 8 + log 18) / (2 log 2 + log 3)
2) log₃5 · log₂₅27
3) 3^(log₉16)
Show solutionHide solution
Solution
2) log₂₅27 = log_(5²)(3³) = (3/2) · log₅3. Since log₃5 · log₅3 = 1, the product is 3/2.
3) log₉16 = log_(3²)(4²) = log₃4 ⇒ 3^(log₃4) = 4.
The swap identity. Both sides have the same logarithm to base k: logₖc · logₖa. The number inside the logarithm in the exponent may swap places with the base of the power.
Using the swap identity
Problem
Evaluate 3^(log₂5) − 5^(log₂3).
Show solutionHide solution
Solution
Check: both numbers have the base-2 logarithm log₂5 · log₂3 ≈ 2.32 · 1.585 ≈ 3.68, so each is about 2^3.68 ≈ 12.8.
| Law | Example |
|---|---|
| logₐ1 = 0, logₐa = 1 | log₇1 = 0, log₇7 = 1 |
| a^(logₐb) = b | 2^(log₂7) = 7 |
| logₐ(xy) = logₐx + logₐy | log₆4 + log₆9 = log₆36 = 2 |
| logₐ(x/y) = logₐx − logₐy | log₃54 − log₃2 = log₃27 = 3 |
| logₐ(xᵖ) = p · logₐx | log₂(1/32) = log₂2⁻⁵ = −5 |
| logₐx = logₖx / logₖa | log₉27 = log₃27 / log₃9 = 3/2 |
| logₐx = 1 / logₓa | log₈2 = 1 / log₂8 = 1/3 |
- 1.log₂32 =
- 2.log 0.01 =
- 3.log₃(1/27) =
- 4.log₄2 =
- 5.5^(log₅7) =
- 6.log₆2 + log₆3 =
The logarithmic function y = logₐx
The function y = aˣ takes each positive value only once, so it has an inverse: we solve for x, x = logₐy, and swap the letters: y = logₐx. Graphs of inverse functions are symmetric about the line y = x, so the properties of the logarithmic function mirror those of the exponential one, with x and y swapped.
An easy way to remember the graph: the graph of y = logₐx is the mirror image of the graph of y = aˣ, and the mirror is the line y = x. The graph of aˣ passes through (0, 1), so the graph of the logarithm passes through (1, 0). aˣ never touches the x-axis, so the logarithm never touches the y-axis. That is why the graph exists only for positive x.
Where
- a
- the base
- x
- the argument, positive numbers only
The inverse of y = aˣ. Domain — (0, +∞), range — all real numbers.
- The graph passes through (1, 0) and (a, 1), because logₐ1 = 0 and logₐa = 1.
- For a > 1 the function is increasing, for 0 < a < 1 decreasing.
- The y-axis (x = 0) is a vertical asymptote: as x approaches zero, the graph goes down (a > 1) or up (0 < a < 1) without bound.
- Sign: if a and x are on the same side of 1, logₐx > 0; if they are on different sides, logₐx < 0.
Interactive
Where
- x₁, x₂
- the positive arguments being compared
Comparing logarithms. If the bases differ, compare each logarithm with a “benchmark” number such as 0 or 1: 1 = logₐa, 2 = logₐa².
Comparing logarithms and finding a domain
Problem
1) log₂5 and log₂7
2) log₀.₃5 and log₀.₃7
3) log₂3 and log₃2
4) log₃10 and 2
5) Find the sign of log₀.₅3.
6) Find the domain of y = log(4 − x²).
Show solutionHide solution
Solution
2) The base 0.3 < 1, the function is decreasing, 5 < 7 ⇒ log₀.₃5 > log₀.₃7.
3) The bases differ, so compare with 1: log₂3 > log₂2 = 1, log₃2 < log₃3 = 1 ⇒ log₂3 > log₃2.
4) 2 = log₃9 and 10 > 9 ⇒ log₃10 > 2.
5) The base 0.5 < 1 and the argument 3 > 1 are on different sides of 1 ⇒ log₀.₅3 < 0.
6) 4 − x² > 0 ⇒ x² < 4 ⇒ −2 < x < 2.
At university the logarithm is seen from another side. The natural logarithm is defined as the area under the graph of 1/t, which gives its derivative: (ln x)′ = 1/x. From this point of view, multiplication turning into addition follows from adding areas. It is also worth knowing that ln x grows more slowly than any positive power, even √x — which is why algorithms with a “log n” running time count as very fast.
The last formula follows from the change of base: logₐx = ln x / ln a, and ln a is a constant.
Logarithms in real life
A logarithmic scale describes a quantity that varies over a huge range with convenient numbers: a difference of one unit on the scale means a 10-fold change in the quantity (on the decibel scale this is 10 dB). Such scales are used in chemistry, acoustics and seismology.
Where
- [H⁺]
- the hydrogen ion concentration (mol/L)
- L
- the sound level in decibels (dB)
- I, I₀
- the sound intensity and the threshold of hearing
In pure water [H⁺] = 10⁻⁷ mol/L, so pH = 7; pH < 7 means acidic, pH > 7 basic. Earthquake magnitude is a common logarithm too: one unit more means 10 times the amplitude recorded by a seismograph and about 32 times the energy released.
pH, decibels and magnitude
Problem
1) A solution has [H⁺] = 10⁻³ mol/L. What is its pH? Is it acidic or basic?
2) Find the pH when [H⁺] = 2 · 10⁻⁵ mol/L (log 2 ≈ 0.301).
3) How many times larger is [H⁺] in a solution with pH = 4 than in one with pH = 6?
4) A sound is 10⁶ times more intense than the threshold of hearing. What is its level? What is the level if the intensity grows another 100 times?
5) How many times larger is the amplitude of a magnitude 7 earthquake than that of a magnitude 5 one?
Show solutionHide solution
Solution
2) pH = −log(2 · 10⁻⁵) = −(log 2 − 5) = 5 − 0.301 ≈ 4.7.
3) [H⁺] = 10⁻⁴ and 10⁻⁶: 10⁻⁴ ÷ 10⁻⁶ = 100 times.
4) L = 10 · log 10⁶ = 60 dB; with another 100 times the intensity, 10 · log 10⁸ = 80 dB (+20 dB).
5) The difference is 2 units ⇒ 10² = 100 times.
Back to Aysel’s question: 1.1ⁿ = 2. Taking the common logarithm of both sides and using the power law gives n · log 1.1 = log 2: the unknown “comes down” from the exponent and an ordinary linear equation is left. Taking logarithms of both sides will be one of the main tools for solving equations in the next lesson.
Where
- p
- the growth rate per period, in %
- n
- the number of periods needed to double
The doubling time. The approximate rule of 72 (n ≈ 72/p) works well for small rates and lets you estimate in your head.
Doubling time
Problem
1) At 10% a year, how many years does it take Aysel’s deposit to double?
2) How many years at 8% a year? Check with the rule of 72.
3) A number of bacteria grows by 50% every hour. After how many hours has it grown 10 times?
Show solutionHide solution
Solution
2) n = log 2 / log 1.08 ≈ 9.0 years; 72 ÷ 8 = 9, a match.
3) 1.5ᵗ = 10 ⇒ t = log 10 / log 1.5 = 1 / log 1.5 ≈ 1 / 0.176 ≈ 5.7 hours.
Remember
- logₐb = c ⇔ aᶜ = b (a > 0, a ≠ 1, b > 0); a^(logₐb) = b, logₐ1 = 0, logₐa = 1.
- log b = log₁₀b is the common logarithm, ln b = logₑb the natural logarithm.
- logₐ(xy) = logₐx + logₐy, logₐ(x/y) = logₐx − logₐy, logₐ(xᵖ) = p · logₐx; there is no rule for the logarithm of a sum.
- Change of base: logₐx = logₖx / logₖa = log x / log a; logₐx = 1 / logₓa.
- y = logₐx is the inverse of y = aˣ: domain (0, +∞), the graph passes through (1, 0), increasing for a > 1 and decreasing for 0 < a < 1.
- Logarithmic scales: pH = −log[H⁺], L = 10 · log(I/I₀); doubling time n = log 2 / log(1 + p/100) ≈ 72/p.