1.5 Exponential and Logarithmic Functions

The form and graph of an exponential function, growth against decay, compound interest and where the number e comes from, the logarithm as the exponential's inverse, the three laws of logarithms and why they are exponent laws read backwards, change of base, solving exponential and logarithmic equations, and the hyperbolic functions.

Subject: Calculus I · 75 slides · symbolic lesson

Open the interactive version of this deck

What this lesson covers

The lesson, slide by slide

1. Exponential and Logarithmic Functions

Title

Calculus I · Section 1.5

The function that multiplies, the function that undoes it, and the number e

2. What this lesson gives you

Objectives

Linear functions add a fixed amount per step. The functions in this section multiply by a fixed factor, and they model everything that grows or decays in proportion to its own size. Calculus will differentiate and integrate them in every chapter that follows.

OpenStax Calculus Volume 1, §1.5 Exponential and Logarithmic Functions §1.5, p. 80 — learning objectives 1.5.1 to 1.5.7

Most of the functions you have met so far change by adding. A line climbs the same amount for every step to the right. The functions in this section change by multiplying: every step to the right multiplies the output by the same factor. That one difference makes them the natural model for anything whose growth depends on how much of it there already is, such as money earning interest, bacteria dividing, or a drug leaving the bloodstream.

You will also meet their inverses, the logarithms, which turn multiplication into addition and let you pull an unknown down out of an exponent. And you will meet the number e, which looks arbitrary at first and turns out to be the one base that calculus finds simplest.

The last part builds the hyperbolic functions out of e. They look like strangers, but they behave like cousins of sine and cosine, and they will return when you integrate in later chapters.

3. Before anything new: undoing a function

Warm-up

Discussion prompt

The equation 2 to the x equals 8 has the answer 3. Now try 2 to the x equals 10. Between which two whole numbers must the answer lie, and what property of 2 to the x (from Section 1.4) guarantees there is exactly one answer?

Write your answer before revealing it. The first half is pure estimation: 2 cubed is 8 and 2 to the fourth is 16, and 10 sits between them, so the exponent you want sits between 3 and 4.

The second half is the important one. Section 1.4 said a function has an inverse exactly when it is one-to-one, and a function that only ever increases can never hit the same output twice. So there is exactly one exponent that turns 2 into 10.

The trouble is that none of the tools you have so far can name that exponent. Roots and powers will not produce it. This lesson gives it a name, the logarithm base 2 of 10, and shows you how to compute it: about 3.32.

4. Exponential functions

Section

Part 1

5. A population that grows 2 percent a year

Concept

Each year the population gains 2 percent of whatever it has. Watch what that does to the formula.

\[ P(1) = P_0 + 0.02P_0 = P_0(1.02) \]

\[ P(2) = P(1)(1.02) = P_0(1.02)^2 \]

\[ P(t) = P_0(1.02)^t \]

Every year multiplies by the same factor, 1.02. A constant multiplier per unit of input is what makes a function exponential.

exponential function — A function of the form f(x) = b to the x, with a constant base b that is positive and not equal to 1, and a variable exponent x.

OpenStax Calculus Volume 1, §1.5 Exponential and Logarithmic Functions §1.5, p. 80 — population growth and the definition

Follow the three lines down the slide. After one year the population is its starting size plus 2 percent of it, and factoring out the starting size shows that this is the same as multiplying by 1.02. The second year does the same thing to the new population, so you multiply by 1.02 again.

After t years you have multiplied by 1.02 exactly t times, which is 1.02 raised to the power t. The variable has climbed into the exponent. That is the defining feature of an exponential function: the base is a fixed number and the input is the exponent.

Notice the two restrictions on the base. It must be positive, because a negative base raised to a fraction such as one half is not a real number. And it must not be 1, because 1 to any power is just 1, which is a flat line and not worth the name.

6. Base or exponent: which one varies?

Concept

\[ \underbrace{x^{2}}_{\text{power: variable base}} \qquad\text{vs}\qquad \underbrace{2^{x}}_{\text{exponential: variable exponent}} \]

x−3−10123456
x²910149162536
2ˣ1/81/21248163264

Figure (svg): The parabola y equals x squared and the exponential y equals 2 to the x on the same axes from x equals minus 6 to 7. They cross three times: near x equals minus 0.77, at x equals 2 and at x equals 4. After x equals 4 the exponential shoots above the parabola.

Near the origin the parabola is ahead. From x equals 4 onward the exponential wins, and the gap grows without bound. To the left, the parabola climbs while the exponential sinks toward zero.

They tie at 2 and at 4. After that the exponential pulls away for good, and to the left it sinks toward zero while the parabola climbs.

OpenStax Calculus Volume 1, §1.5 Exponential and Logarithmic Functions §1.5, pp. 80-81 — Table 1.10 and Figure 1.43

These two functions are easy to confuse because both contain a 2 and an x. The difference is which one moves. In x squared the base changes and the exponent is fixed; that is a power function. In 2 to the x the base is fixed and the exponent changes; that is an exponential.

Read the table from left to right. For negative inputs the parabola is large while the exponential is a small fraction. The two agree at x equal to 2 and again at x equal to 4. From 5 onward the exponential is ahead, and each step doubles it while the parabola adds only a little.

The graph shows the same story. Near the origin the two curves wind around each other, but on the right the exponential leaves the parabola behind for good. That is the first sign of a fact you will prove later in the course: exponential growth eventually beats every power.

7. Who wins in the end: x to the tenth or 2 to the x?

Prediction

\[ x^{10} \quad\text{vs}\quad 2^{x}, \qquad x^{10} = 10^{10} \text{ and } 2^{10} = 1024 \text{ at } x = 10 \]

Predict first

At x = 10 the power is ten million times bigger. What happens as x keeps growing?

  • x¹⁰ stays ahead forever
  • 2ˣ overtakes it and stays ahead
  • They keep swapping the lead
  • They grow closer and become equal

Correct: 2ˣ overtakes it and stays ahead

Why: The exponential catches up near x equal to 58.8 and is ahead from then on. Each step of 1 multiplies 2 to the x by 2, while x to the tenth gains a factor that shrinks toward 1 as x grows. A fixed doubling eventually beats any fixed power.

\[ 2^{x} > x^{10} \quad\text{for all } x > 58.77 \]

Commit to an answer before revealing. At x equal to 10 the power is ahead by a factor of about ten million, so it is natural to guess it stays ahead.

Think about what one step to the right does to each function. The exponential always doubles. The power x to the tenth gets multiplied by x plus 1 over x, all to the tenth, and that factor shrinks toward 1 as x grows. At x equal to 100 it is only about 1.1. Something that keeps doubling must eventually overtake something whose growth factor is sinking toward 1.

The crossover is near x equal to 58.8, and after that the exponential is ahead forever. This is why computer scientists fear exponential running times: any power of the input, however large, is eventually smaller.

8. What 2 to the root 2 means

Concept

Whole-number exponents count factors; negative ones take reciprocals; fractional ones take roots. None of that says what an irrational exponent is.

\[ b^{p/q} = \sqrt[q]{b^{p}}, \qquad 9^{3/2} = \sqrt{9^3} = 27 \]

x1.41.411.4141.41421.414211.414213
2ˣ2.6392.657372.664752.6651192.6651382.665143

Figure (svg): A zoomed-in graph of y equals 2 to the x for x between 1.395 and 1.418. Dots sit on the curve at x equals 1.4, 1.41, 1.414 and 1.4142, crowding toward a dashed vertical line at x equals root 2, where the curve meets a dashed horizontal line at height 2.665144.

Each rational exponent can be computed with roots and powers. As the exponents crowd toward root 2, the heights crowd toward one number, and that number is what 2 to the root 2 means.

As the rational exponents close in on root 2, the values close in on one number. That limit is the definition.

\[ 2^{\sqrt2} = L \approx 2.665144 \]

OpenStax Calculus Volume 1, §1.5 Exponential and Logarithmic Functions §1.5, p. 81 — Table 1.11

Look at the chain of definitions first. Whole-number exponents count factors, negative exponents take reciprocals, and fractional exponents take roots, so 9 to the three halves is the square root of 9 cubed, which is 27. Every rational exponent has a meaning built from arithmetic you already trust.

An irrational exponent such as root 2 has no such meaning, because root 2 is not a fraction. The fix is a limit, and the table shows it happening. Replace root 2 by better and better decimal approximations, each of which is a fraction, and compute 2 to each of them. The values settle down: 2.639, 2.657, 2.6647, and so on.

The picture zooms right in on the curve so you can see the dots crowding toward the dashed lines. The height they approach, about 2.665144, is by definition 2 to the root 2. Filling in the irrational exponents this way is what makes the graph of 2 to the x an unbroken curve.

9. Example 1.33: bacteria that double every 4 hours

Worked example

A culture starts with 1000 bacteria and doubles every 4 hours. Find the count after 6, 10 and 24 hours.

\[ n(4) = 1000\cdot 2, \quad n(8) = 1000\cdot 2^2, \quad n(4m) = 1000\cdot 2^{m} \]

Rewrite in hours

Why: After t hours there have been t over 4 doublings, so put m equal to t over 4.

\[ n(t) = 1000\cdot 2^{t/4} \]

Six hours

Why: One and a half doublings.

\[ n(6) = 1000\cdot 2^{6/4} = 1000\cdot 2^{1.5} \approx 2828 \]

Ten hours

Why: Two and a half doublings.

\[ n(10) = 1000\cdot 2^{10/4} = 1000\cdot 2^{2.5} \approx 5657 \]

Twenty-four hours

Why: Exactly six doublings, so the answer is exact.

\[ n(24) = 1000\cdot 2^{24/4} = 1000\cdot 2^{6} = 64{,}000 \]

Figure (svg): The curve n of t equals 1000 times 2 to the t over 4, in thousands, for t from 0 to 26 hours. Grey stems mark the doublings at 4, 8, 12, 16, 20 and 24 hours (2, 4, 8, 16, 32 and 64 thousand). Highlighted dots mark 2828 at 6 hours, 5657 at 10 hours and 64,000 at 24 hours.

Every four hours the height doubles. In-between times such as 6 and 10 hours land on the same smooth curve, which is why a fractional exponent is the right model.

Check with the doubling period

Why: Ten hours is six hours plus one doubling period, so the count at 10 hours must be exactly twice the count at 6.

\[ 2\times 2828.43 = 5656.85 = n(10) \]

OpenStax Calculus Volume 1, §1.5 Exponential and Logarithmic Functions §1.5, p. 81 — Example 1.33

Start from the pattern in the problem: after 4 hours there has been one doubling, after 8 hours two, and after 4m hours m doublings. The only real modelling step is to turn hours into doublings, which you do by dividing by 4. That puts t over 4 in the exponent.

Six hours is one and a half doublings, so you need 2 to the 1.5, which is 2 times root 2, about 2.83. Ten hours is two and a half doublings. Twenty-four hours is exactly six doublings, so that answer, 64,000, needs no calculator at all.

The check uses the structure of the model rather than recomputing. Ten hours is six hours plus one full doubling period, so the count at 10 hours must be exactly double the count at 6. It is. Look at the figure too: the grey stems are the whole doublings, and the highlighted in-between times sit on the same smooth curve.

10. Checkpoint 1.27: evaluate 100 times 3 to the x over 2

Fill the middle

\[ f(x) = 100\cdot 3^{x/2} \]

Fill in the blanks

f(4) = 100\cdot 3^900 = 24300, \qquad f(10) = 100\cdot 3^___ = ___

Why: Halve the input first: 4 over 2 is 2 and 10 over 2 is 5. Then 3 squared is 9 and 3 to the fifth is 243, so the values are 900 and 24,300.

OpenStax Calculus Volume 1, §1.5 Exponential and Logarithmic Functions §1.5, p. 81 — Checkpoint 1.27

Fill in both blanks before checking. The only trap is order of operations: the exponent is x over 2, so halve the input first and only then raise 3 to that power.

At x equal to 4 the exponent is 2, and 3 squared is 9, so the value is 900. At x equal to 10 the exponent is 5, and 3 to the fifth is 243, so the value is 24,300.

A common slip is to multiply 100 by 3 before raising to the power, which would give 300 squared, a much larger number. The exponent applies only to the 3, never to the coefficient in front of it.

11. The graph of b to the x: domain, range, asymptote

Concept

Figure (svg): Four exponential curves through the point (0, 1): 2 to the x and 4 to the x rise to the right, one half to the x and one quarter to the x rise to the left. All four stay above the x-axis, which is their horizontal asymptote.

A base above 1 climbs to the right; a base between 0 and 1 is the mirror image and climbs to the left. Every one of them passes through (0, 1) and never touches the x-axis.

\[ f(x) = b^{x}: \quad \text{domain } (-\infty,\infty), \quad \text{range } (0,\infty) \]

basedirectionas x → ∞as x → −∞
b > 1increasing (growth)bˣ → ∞bˣ → 0
0 < b < 1decreasing (decay)bˣ → 0bˣ → ∞

\[ \left(\tfrac12\right)^{x} = 2^{-x} \]

OpenStax Calculus Volume 1, §1.5 Exponential and Logarithmic Functions §1.5, p. 82 — Figure 1.44

Every exponential shares a skeleton, and it is worth being able to draw it from memory. The domain is every real number, because any real exponent now has a meaning. The range is only the positive numbers, because a positive base raised to any power stays positive. And every curve passes through the point 0 comma 1, because any base to the power 0 is 1.

The base decides the direction. Above 1, the function grows to the right and sinks toward the x-axis on the left. Between 0 and 1 it is the other way round. The last line on the slide explains why: one half to the x is the same as 2 to the minus x, so a decay curve is just a growth curve reflected in the y-axis.

The x-axis is a horizontal asymptote on one side of every exponential. The curve gets as close to it as you like but never touches it, because the output is never zero.

12. Slide the base b

Tweak it

Parameter explorer

Move b from 0.2 up to 4. Where does the curve always pass, when does it flatten into a horizontal line, and when does it switch from rising to falling?

\[ y = {b}^{x} \]

  • b — from 0.2 to 4: base b

Start the slider at 2 and move it slowly upward. The curve gets steeper on the right and hugs the axis more tightly on the left, but it never leaves the point 0 comma 1.

Now bring it down toward 1. The curve flattens, and at exactly 1 it becomes the horizontal line at height 1. That is why 1 is excluded as a base: it produces a constant function, which is not one-to-one and has no inverse.

Push it below 1 and the curve turns around and falls from left to right. Compare b equal to 0.5 with b equal to 2: they are mirror images in the y-axis. The smaller the base, the faster the decay.

13. Growth or decay?

Sorting

Sort into buckets

Rewrite each as a single base if you need to, then sort it.

Growth (increasing)
f(x) = 3ˣ; f(x) = (5/4)ˣ; f(x) = (1/3)^(−x)
Decay (decreasing)
f(x) = (0.4)ˣ; f(x) = 2⁻ˣ; f(x) = e^(−0.2x)
grow
3, 5/4 are bases above 1; and (1/3)^(−x) equals 3ˣ, since a negative exponent flips the base.
decay
0.4 is below 1; 2⁻ˣ equals (1/2)ˣ; and e^(−0.2x) equals (e^(−0.2))ˣ, a base of about 0.819.

Sort each item before checking, and for the tricky ones rewrite the function with a single base and a plain exponent x. Once it looks like some number to the x, the rule is simple: above 1 means growth, between 0 and 1 means decay.

Three of these hide their base. Two to the minus x is one half to the x, so it decays. The function e to the minus 0.2x is e to the minus 0.2, all to the x, and e to the minus 0.2 is about 0.82, which is less than 1, so it decays too. And one third to the minus x flips the fraction to give 3 to the x, which grows.

The lesson is that a minus sign in the exponent reverses the direction. Two minus signs, one in the exponent and one hidden in a fractional base, cancel.

14. Reading the laws of exponents

Notation

Annotate

On: \( b^{x}b^{y} = b^{x+y}, \quad (b^{x})^{y} = b^{xy}, \quad (ab)^{x} = a^{x}b^{x} \)

  • Multiplying powers of the same base adds the exponents: x factors of b followed by y more make x + y factors.
  • Raising to a further power multiplies the exponents: y copies of x factors each.
  • An exponent on a product goes onto each factor. There is no such rule for a SUM: (a + b)ˣ does not split.
  • Quotients work the same way with subtraction: bˣ/bʸ = bˣ⁻ʸ and aˣ/bˣ = (a/b)ˣ.

\[ \frac{b^{x}}{b^{y}} = b^{x-y}, \qquad \frac{a^{x}}{b^{x}} = \left(\frac ab\right)^{x}, \qquad a > 0,\; b > 0 \]

OpenStax Calculus Volume 1, §1.5 Exponential and Logarithmic Functions §1.5, p. 82 — Rule: Laws of Exponents

Step through the annotations and try to justify each law by counting factors before reading the explanation. If you multiply x factors of b by y more factors of b, you have x plus y factors. If you take y copies of a block of x factors, you have x times y factors.

The third law says an exponent on a product lands on each factor. The note warns about the matching law that does not exist: an exponent on a sum does not split. Two plus three, squared, is 25, not 4 plus 9.

These laws are stated for positive bases, and that condition is what lets them hold for every real exponent, not just whole numbers. You will use them constantly, both to simplify expressions and, in a few slides, to prove the laws of logarithms.

15. Example 1.34(a): fractional exponents

Worked example

\[ \frac{\left(2x^{2/3}\right)^3}{\left(4x^{-1/3}\right)^2} \]

Put each outer power on each factor

Why: Law 4: a power of a product is the product of the powers.

\[ = \frac{2^3\left(x^{2/3}\right)^3}{4^2\left(x^{-1/3}\right)^2} \]

Multiply the exponents

Why: Law 3: two thirds times 3 is 2, and minus one third times 2 is minus two thirds.

\[ = \frac{8x^{2}}{16x^{-2/3}} \]

Divide the numbers and subtract the exponents

Why: Law 2 on the powers of x; 8 over 16 is one half.

\[ = \frac{x^{2-(-2/3)}}{2} \]

Simplify the exponent

Why: Two plus two thirds is eight thirds.

\[ = \frac{x^{8/3}}{2} \]

Check at x equal to 2

Why: Evaluate the original expression and the answer on a calculator; they must agree.

\[ \frac{\left(2\cdot 2^{2/3}\right)^3}{\left(4\cdot 2^{-1/3}\right)^2} \approx 3.17480, \qquad \frac{2^{8/3}}{2} \approx 3.17480 \]

OpenStax Calculus Volume 1, §1.5 Exponential and Logarithmic Functions §1.5, pp. 82-83 — Example 1.34a

The strategy is always the same: first push the outer exponents onto every factor inside the brackets, then deal with each base separately. Doing it in this order keeps you from multiplying an exponent into one factor and forgetting another.

The numbers become 2 cubed, which is 8, and 4 squared, which is 16. The powers of x need the power-of-a-power rule: two thirds times 3 is 2, and minus one third times 2 is minus two thirds. Then dividing powers of x subtracts exponents, and subtracting a negative adds it, so the exponent becomes eight thirds.

The check evaluates the original and the final answer at x equal to 2. They agree to every digit shown. A numerical check like this catches almost every sign error in exponent work, so it is worth the ten seconds.

16. Example 1.34(b): negative exponents

Worked example

\[ \frac{\left(x^3y^{-1}\right)^2}{\left(xy^2\right)^{-2}} \]

Put each outer power on each factor

Why: Law 4, top and bottom.

\[ = \frac{\left(x^3\right)^2\left(y^{-1}\right)^2}{x^{-2}\left(y^2\right)^{-2}} \]

Multiply the exponents

Why: Law 3 on every factor.

\[ = \frac{x^6y^{-2}}{x^{-2}y^{-4}} \]

Bring the denominator up

Why: Dividing by a power is multiplying by the power with the opposite sign.

\[ = x^6x^{2}y^{-2}y^{4} \]

Add the exponents

Why: Law 1, one base at a time.

\[ = x^{8}y^{2} \]

Check at x equal to 2 and y equal to 3

Why: The original is 64 over 9, divided by 18 to the minus 2; the answer is 256 times 9.

\[ \frac{64}{9}\cdot 18^{2} = 2304 = 2^{8}\cdot 3^{2} \]

OpenStax Calculus Volume 1, §1.5 Exponential and Logarithmic Functions §1.5, p. 83 — Example 1.34b

This is the same routine with more negative signs, which is where errors creep in. Push the outer powers in: x cubed squared is x to the sixth, y to the minus 1 squared is y to the minus 2, and on the bottom every exponent gets multiplied by minus 2.

The step that moves the denominator up is the key idea. Dividing by x to the minus 2 is the same as multiplying by x to the 2, because a negative exponent means a reciprocal, and dividing by a reciprocal multiplies. The same goes for y to the minus 4.

Now just add exponents base by base: 6 plus 2 is 8 for x, and minus 2 plus 4 is 2 for y. The check at x equal to 2 and y equal to 3 gives 2304 both ways, which is 256 times 9.

17. Checkpoint 1.28: simplify a quotient of powers

Fill the middle

Fill in the blanks

\frac3y^___}___y^___} = \frac___}___} = \frac______} }}

Why: 6 over 12 is one half. The x exponents give minus 3 minus minus 4, which is 1, so x stays upstairs. The y exponents give 2 minus 5, which is minus 3, so y cubed moves downstairs.

\[ \text{check at } x = 2,\; y = 3: \quad \frac{6\cdot\frac18\cdot 9}{12\cdot\frac1{16}\cdot 243} = \frac{1}{27} = \frac{2}{2\cdot 27} \]

OpenStax Calculus Volume 1, §1.5 Exponential and Logarithmic Functions §1.5, p. 83 — Checkpoint 1.28

Work out the blank before checking. Handle the numbers and each variable separately. Six over twelve is one half, which puts a 2 in the denominator.

For x, subtract the bottom exponent from the top one: minus 3 minus minus 4 is plus 1, so a single x stays on top. For y, the top exponent 2 minus the bottom exponent 5 is minus 3, and a negative exponent means the power belongs downstairs, so y cubed goes in the denominator.

The numerical check under the formula uses x equal to 2 and y equal to 3, and both sides come out to one twenty-seventh. If you ever lose track of which way a negative exponent sends a factor, a quick number like this settles it.

18. Trap: multiplying the bases

Trap

The trap

A tempting shortcut:

\[ 2\cdot 3^{x} = 6^{x} \]

Wrong. The exponent belongs to the 3 alone.

The fix

Test it at x equal to 2. The left side is 2 times 9, the right side is 36. Bases combine only when they carry the same exponent.

\[ 2\cdot 3^{2} = 18 \]

\[ 6^{2} = 36 \]

\[ 2^{x}\cdot 3^{x} = 6^{x} \;\checkmark \]

This error comes from half-remembering the law that combines bases. That law needs both bases to carry the same exponent: 2 to the x times 3 to the x really is 6 to the x. In 2 times 3 to the x, the 2 has no exponent at all, or if you like, exponent 1, while the 3 has exponent x.

The fastest defence is to test a small value. At x equal to 2 the left side is 18 and the right side is 36. At x equal to 0 the left side is 2 and the right side is 1. The two functions are not even close. Whenever you are unsure whether an exponent rule applies, plug in a number.

19. The number e

Section

Part 2

20. Compounding more often

Concept

Put P dollars in an account at annual rate r. Compounded once a year:

\[ A(1) = P + rP = P(1+r), \qquad A(t) = P(1+r)^{t} \]

Compounded twice a year, half the rate is paid each half year:

\[ A\left(\tfrac12\right) = P\left(1 + \tfrac r2\right), \qquad A(1) = P\left(1 + \tfrac r2\right)^{2} \]

\[ A(t) = P\left(1 + \frac r2\right)^{2t} \]

Compounded n times a year, the rate per period is r over n and there are nt periods:

\[ A(t) = P\left(1 + \frac rn\right)^{nt} \]

OpenStax Calculus Volume 1, §1.5 Exponential and Logarithmic Functions §1.5, p. 83 — compound interest

Follow the account one period at a time. Compounded yearly, each year multiplies the balance by 1 plus r, so t years multiply it by 1 plus r to the power t. Nothing new there; it is the population model again.

Compounded twice a year, the bank pays half the annual rate every six months. Each half year multiplies the balance by 1 plus r over 2, and there are two half years in a year, so after t years you have multiplied by that factor 2t times.

The general formula follows the same pattern: n periods a year, rate r over n per period, nt periods in total. Keep this formula in mind, because the question on the next slide is what happens to it as n grows without bound, and the answer produces the most important constant in calculus.

21. Where e comes from

Concept

What happens as n grows without bound? Substitute m equal to n over r, so that n equals mr.

\[ \left(1 + \frac rn\right)^{nt} = \left(1 + \frac 1m\right)^{mrt} = \left[\left(1 + \frac 1m\right)^{m}\right]^{rt} \]

Figure (svg): The values of one plus one over m, all to the power m, plotted against m on a logarithmic axis from 1 to a million. The curve starts at 2, passes 2.613 at m equals 12 and 2.7146 at m equals 365, and flattens under a dashed line at e, about 2.718282.

One dollar at 100 percent for a year. Compounding more often always helps, but by less and less: the values climb toward a ceiling instead of running away. The ceiling is e.

the number e — The limit of one plus one over m, all to the power m, as m grows without bound: about 2.718282. Euler named it in the 1720s.

\[ e = \lim_{m\to\infty}\left(1 + \frac1m\right)^{m} \approx 2.718282 \quad\Longrightarrow\quad A(t) = Pe^{rt} \]

OpenStax Calculus Volume 1, §1.5 Exponential and Logarithmic Functions §1.5, pp. 83-84 — Table 1.12 and continuous compounding

The substitution looks like a trick, but its purpose is simple: it separates the part that depends on n from the part that does not. Setting m equal to n over r turns the rate over n into one over m and the exponent nt into m times rt. The bracket, one plus one over m all to the m, is now the only thing that changes as compounding becomes more frequent.

The graph shows what that bracket does. Its horizontal axis is logarithmic, so each tick is ten times the last. Compounding once gives 2, monthly gives 2.613, daily 2.7146, and the values keep rising but by less and less, flattening under the dashed line at about 2.718282. That ceiling is the number e.

Replacing the bracket by its limit gives continuous compounding: the balance is P times e to the rt. The same formula describes any quantity that grows continuously at a rate proportional to its size.

22. A dollar compounded every second

Estimation

\[ \left(1 + \frac{1}{31{,}536{,}000}\right)^{31{,}536{,}000} = \;? \]

Predict first

One dollar at 100 percent a year, compounded every second (31,536,000 times). About how much after one year?

  • About $2.00
  • About $2.72
  • About $100
  • Millions of dollars

Correct: About $2.72

Why: That is the expression whose limit defines e, with m about 31.5 million: it comes out at 2.7182818, agreeing with e to seven digits. Compounding more often can never push past e dollars, which is why continuous compounding is a sensible idea rather than an infinite one.

\[ \left(1 + \tfrac1m\right)^{m}\Big|_{m = 31{,}536{,}000} \approx 2.7182818 \approx e \]

Pick an answer before revealing. Many people expect compounding every second to produce enormous amounts of money, because it sounds like interest on interest on interest without end.

Compounding every second means m is about 31.5 million in the expression that defines e, and the result agrees with e to seven digits. So the dollar grows to about 2 dollars and 72 cents, and no amount of extra compounding can push it past e dollars.

Two effects fight each other. More periods means more compounding, but each period pays a proportionally smaller rate. They nearly cancel, and the small amount that is left over converges. That convergence is exactly what makes continuous growth a sensible model rather than an infinite one.

23. The one base whose slope at zero is 1

Concept

Figure (svg): The curve y equals e to the x with its tangent line y equals x plus 1 touching it at the point (0, 1). The line stays below the curve everywhere else.

Of all the exponential curves through (0, 1), only the one with base e leaves that point at exactly 45 degrees. That is the property that makes e the base calculus prefers.

Estimate each curve's slope at x equal to 0 with a tiny step h of one ten-thousandth:

\[ \text{slope of } b^{x} \text{ at } 0 \approx \frac{b^{h} - b^{0}}{h} = \frac{b^{h} - 1}{h} \]

base b(bʰ − 1)/h with h = 0.0001
20.69317
e ≈ 2.718281.00005
31.09867

Base 2 leaves the y-axis too gently, base 3 too steeply. Base e is exactly in between: slope 1. The function e to the x is called the natural exponential function.

OpenStax Calculus Volume 1, §1.5 Exponential and Logarithmic Functions §1.5, p. 84 — Figure 1.45

Every exponential passes through 0 comma 1, but they leave that point at different steepnesses. The table estimates each slope the way you will define slope in Chapter 2: rise over run across a tiny step.

Base 2 has slope about 0.69 at that point, less than 1. Base 3 has slope about 1.10, more than 1. Somewhere between 2 and 3 there must be a base whose slope there is exactly 1, and that base is e. The figure shows the natural exponential with its tangent line y equals x plus 1, touching at 0 comma 1 and rising at exactly 45 degrees.

It will not surprise you later that the slopes 0.69 and 1.10 are the natural logarithms of 2 and 3. This slope-1 property is why calculus uses base e almost exclusively: every formula involving rates of change is simplest there.

24. Example 1.35: 500 dollars compounded continuously

Worked example

500 dollars is invested at 5.5 percent a year, compounded continuously. Find A(t), then the balance after 10 and 20 years.

Recall the continuous model

Why: Principal times e to the rate times the time.

\[ A(t) = Pe^{rt} \]

Substitute the principal and the rate

Why: 5.5 percent as a decimal is 0.055.

\[ A(t) = 500e^{0.055t} \]

After 10 years

Why: The exponent is 0.055 times 10.

\[ A(10) = 500e^{0.055\cdot 10} = 500e^{0.55} \approx \$866.63 \]

After 20 years

Why: The exponent is 0.055 times 20.

\[ A(20) = 500e^{0.055\cdot 20} = 500e^{1.1} \approx \$1502.08 \]

Check against yearly compounding

Why: Continuous compounding must beat compounding once a year, but only by a little.

\[ 500(1.055)^{10} \approx \$854.07 < \$866.63 \]

OpenStax Calculus Volume 1, §1.5 Exponential and Logarithmic Functions §1.5, p. 84 — Example 1.35

The model comes straight from the previous slides: principal times e to the rate times time. The only care needed is to write the rate as a decimal, 0.055 rather than 5.5, since the formula multiplies by it.

For part (b) the exponents are 0.55 and 1.1, and a calculator does the rest: about 866 dollars after 10 years and about 1502 after 20. Notice that the balance does not double from 10 years to 20. It is multiplied by e to the 0.55, about 1.73, which is what the first 10 years did to the original 500.

The check compares with compounding once a year. Continuous compounding must give more, since it is the limit of an increasing sequence, but only a little more: about 866.63 against 854.07. If your continuous answer ever comes out smaller than a discrete one, something has gone wrong.

25. Checkpoint 1.29: 750 dollars at 4 percent

Worked example

750 dollars at 4 percent a year, compounded continuously. Find A(t) and the balance after 30 years.

Write the model

Why: P is 750 and r is 0.04.

\[ A(t) = 750e^{0.04t} \]

Put t equal to 30

Why: The exponent is 0.04 times 30.

\[ A(30) = 750e^{0.04\cdot 30} = 750e^{1.2} \]

Evaluate the exponential

Why: A calculator gives e to the 1.2.

\[ e^{1.2} \approx 3.320117 \]

Multiply

Why: 750 times that factor.

\[ A(30) \approx 750\times 3.320117 \approx \$2490.09 \]

Check against yearly compounding

Why: Yearly compounding must give a little less.

\[ 750(1.04)^{30} \approx \$2432.55 < \$2490.09 \]

OpenStax Calculus Volume 1, §1.5 Exponential and Logarithmic Functions §1.5, p. 84 — Checkpoint 1.29

This is the same pattern with different numbers, so let the steps carry you. The principal is 750, the rate as a decimal is 0.04, and the exponent after 30 years is 0.04 times 30, which is 1.2.

E to the 1.2 is about 3.32, so the balance is a little more than three times the deposit: about 2490 dollars. A rate of 4 percent sounds small, but over 30 years it more than triples the money, which is the practical meaning of exponential growth.

The check again compares with yearly compounding, which gives about 2432.55. The continuous balance is larger, as it must be, and the gap after 30 years is about 58 dollars.

26. Find the error: the rate was not divided by 12

Error analysis

Someone computes 1000 dollars at 5 percent for 10 years, compounded monthly.

Annotate

On: \( A = 1000\left(1 + 0.05\right)^{120} \approx \$348{,}912 \)

  • 120 is right: 12 periods a year for 10 years.
  • Wrong. 5 percent is the ANNUAL rate; each month pays one twelfth of it, 0.05/12. This line pays 5 percent every month.
  • 5 percent a year for 10 years cannot turn 1000 dollars into 348,912. Even yearly compounding gives only about 1628.89.
  • 1000(1 + 0.05/12)¹²⁰ ≈ 1647.01 dollars, just below the continuous 1000e^0.5 ≈ 1648.72.

Look at the calculation and decide what is wrong before revealing the notes. The number of periods is correct: monthly for 10 years is 120 periods.

The error is in the rate. Five percent is an annual rate, so each month should pay one twelfth of it. Using 0.05 per month pays 5 percent every single month, which is about 80 percent a year once compounded. No wonder the balance is over three hundred thousand dollars.

The fastest way to catch this kind of mistake is a sanity check against something simple. Five percent a year compounded yearly gives about 1629 dollars after 10 years, so any monthly answer must be close to that, and the correct one, about 1647 dollars, is.

27. Logarithms as inverses

Section

Part 3

28. The logarithm is an exponent

Concept

Each exponential with base b is one-to-one, from all real numbers onto the positive numbers, so it has an inverse. That inverse is the logarithm with base b.

\[ \log_b(x) = y \iff b^{y} = x \]

logarithmbecause
log₂ 8 = 32³ = 8
log₁₀ (1/100) = −210⁻² = 1/10² = 1/100
log_b 1 = 0b⁰ = 1 for every base b

logarithm base b — For b positive and not 1, the logarithm of x is the exponent you must put on b to get x. Domain: the positive numbers. Range: all real numbers.

OpenStax Calculus Volume 1, §1.5 Exponential and Logarithmic Functions §1.5, p. 85 — the logarithmic function with base b

Read the definition as a question. The logarithm base b of x asks: what power must b be raised to in order to give x? The logarithm base 2 of 8 is 3 because 2 cubed is 8. The logarithm base 10 of one hundredth is minus 2 because 10 to the minus 2 is one hundredth.

The logarithm exists because the exponential is one-to-one. Its domain and range are the exponential's range and domain swapped. The exponential accepts every real number and outputs only positives, so the logarithm accepts only positive inputs and can output any real number.

The last row of the table is worth remembering on its own: the logarithm of 1 is 0 in every base, because every base to the power 0 is 1. You will see that fact again as the point every logarithm graph passes through.

29. Logarithmic form to exponential form

Matching

Match the pairs

  • a. log₃ 81 = 4
  • b. log₈ 2 = 1/3
  • c. log 0.1 = −1
  • d. ln(1/e³) = −3
  • e. log₉ 3 = 0.5
  • p. 3⁴ = 81
  • q. 8^(1/3) = 2
  • r. 10⁻¹ = 0.1
  • s. e⁻³ = 1/e³
  • t. 9^0.5 = 3

Why: In every pair the base of the logarithm becomes the base of the power, the logarithm's value becomes the exponent, and the argument becomes the result. The cube root of 8 is 2 and the square root of 9 is 3, which is why fractional logarithms appear.

OpenStax Calculus Volume 1, §1.5 Exponential and Logarithmic Functions §1.5, p. 95 — Exercises 246 to 252

Match each pair before checking. In every conversion, three numbers swap roles in the same way: the base of the logarithm stays the base, the value of the logarithm becomes the exponent, and the thing inside the logarithm becomes the result.

The fractional ones are the most instructive. The logarithm base 8 of 2 is one third because the cube root of 8 is 2, and 8 to the one third means the cube root. The logarithm base 9 of 3 is one half for the same reason with square roots.

The item with the natural logarithm, ln of one over e cubed equal to minus 3, becomes e to the minus 3 equal to one over e cubed, which is just the meaning of a negative exponent. If a logarithm statement ever looks mysterious, rewrite it in exponential form and it usually becomes obvious.

30. Undo, then redo: the inverse identities

Concept

Composing a function with its inverse gives back the input, in either order.

\[ \log_b\left(b^{x}\right) = x \quad\text{for all } x, \qquad b^{\log_b x} = x \quad\text{for } x > 0 \]

The logarithm with base e is the natural logarithm, written ln.

\[ \ln e = 1, \qquad \ln\left(e^3\right) = 3, \qquad \ln 1 = 0 \]

\[ \ln\left(e^{x}\right) = x, \qquad e^{\ln x} = x \]

OpenStax Calculus Volume 1, §1.5 Exponential and Logarithmic Functions §1.5, p. 85 — the natural logarithm

These two identities are the definition of inverse functions written out for the pair exponential and logarithm. Put x into the exponent, take the logarithm, and you get x back. Take the logarithm first, use it as an exponent, and you get x back again.

There is one asymmetry to notice. The first identity holds for every real x, because b to the x always exists and is always positive. The second needs x positive, because the logarithm of x has to exist before you can use it.

The natural logarithm lines are the ones you will use most. The natural log of e is 1, the natural log of e cubed is 3, and the natural log of 1 is 0. And e to the ln x is x, which is the identity that will let you rewrite any exponential in base e.

31. Mirror images across y = x

Picture it

Figure (svg): The curves y equals e to the x and y equals ln x, drawn with equal scales, mirror images across the dashed line y equals x. Dotted segments join matching points: (0, 1) with (1, 0), (1, e) with (e, 1), and (minus 2, e to the minus 2) with (e to the minus 2, minus 2).

Swap the coordinates of any point on one curve and you land on the other. The exponential's horizontal asymptote, the x-axis, becomes the logarithm's vertical asymptote, the y-axis.

Inverse functions swap inputs and outputs, so their graphs swap coordinates: reflect the exponential in the diagonal line and you get the logarithm. Each dotted segment crosses the diagonal at right angles.

OpenStax Calculus Volume 1, §1.5 Exponential and Logarithmic Functions §1.5, p. 85 — Figure 1.46

The figure uses the same scale on both axes, which matters: only with equal scales does the dashed diagonal act like a true mirror. Pick any dot on the exponential curve and follow its dotted segment across the diagonal. It lands on the matching dot on the logarithm curve, with its two coordinates swapped.

So 0 comma 1 on the exponential becomes 1 comma 0 on the logarithm, and 1 comma e becomes e comma 1. The shapes swap too. The exponential lies flat along the negative x-axis; the logarithm plunges down along the negative y-axis. A horizontal asymptote in one becomes a vertical asymptote in the other.

This is the picture to keep in your head whenever you need a fact about logarithms that you cannot remember. Reflect the corresponding fact about exponentials and you will have it.

32. Every logarithm passes through (1, 0)

Concept

Figure (svg): Three logarithm curves, log base 2, ln, and log base 10, for x from 0 to 10. All three pass through (1, 0), plunge toward the y-axis on the left, and rise ever more slowly on the right; the larger the base, the flatter the curve.

Every logarithm sends 1 to 0, because every base to the power 0 is 1. The base only changes how steeply the curve climbs.

\[ g(x) = \log_b x: \quad \text{domain } (0,\infty), \quad \text{range } (-\infty,\infty) \]

Each curve plunges down the y-axis, its vertical asymptote, and climbs forever on the right, but more and more slowly. To reach a height of 10, the natural logarithm needs x to be more than 22,000.

\[ \ln x = 10 \iff x = e^{10} \approx 22{,}026 \]

OpenStax Calculus Volume 1, §1.5 Exponential and Logarithmic Functions §1.5, pp. 85-86 — Figure 1.47

Here are three logarithms side by side: base 2, base e and base 10. All three pass through 1 comma 0, all three have the y-axis as a vertical asymptote, and all three rise forever to the right. What the base changes is how quickly they rise: the larger the base, the flatter the curve.

The rise is extremely slow. The natural logarithm does not reach 10 until x is about 22,000, and it does not reach 20 until x is nearly half a billion. Logarithms grow without bound, but more slowly than any positive power of x, which is the mirror image of exponentials outgrowing every power.

That slowness is exactly why logarithmic scales are useful. Quantities that span many orders of magnitude, such as earthquake sizes, sound levels and acidity, become manageable numbers between about 0 and 14 once you take a logarithm.

33. Is the logarithm of a negative number defined?

Prediction

\[ \ln(-4) = \;? \]

Predict first

Which statement is correct?

  • ln(−4) = −ln 4
  • ln(−4) is undefined: no power of e is negative
  • ln(−4) = 0
  • ln(−4) is defined but negative

Correct: ln(−4) is undefined: no power of e is negative

Why: The natural log of minus 4 would be a number y with e to the y equal to minus 4. But e to any real power is positive, so no such y exists. Negative OUTPUTS are fine, as in ln of one half, about minus 0.693; negative INPUTS are not.

Choose an option before revealing. The tempting wrong answer copies a rule from odd functions, but the logarithm is not odd, and it has no rule for negative inputs at all.

Go back to the definition. The natural log of minus 4 would be a number y with e to the y equal to minus 4. But e to any real power is positive; the graph of the exponential never dips below the x-axis. So no such y exists and the logarithm of a negative number is undefined. The same reasoning rules out the logarithm of 0.

Keep inputs and outputs separate in your mind. A logarithm can happily be negative: the natural log of one half is about minus 0.693. What it cannot do is accept a negative or zero input. That domain restriction is what produces extraneous roots later in the lesson.

34. The laws of logarithms are exponent laws

Concept

Name the two logarithms, rewrite them as powers, and multiply:

\[ u = \log_b a, \quad v = \log_b c \iff b^{u} = a, \quad b^{v} = c \]

\[ ac = b^{u}b^{v} = b^{u+v} \]

\[ \log_b(ac) = u + v = \log_b a + \log_b c \]

propertylogarithm lawexponent law it comes from
productlog_b(ac) = log_b a + log_b cbᵘbᵛ = bᵘ⁺ᵛ
quotientlog_b(a/c) = log_b a − log_b cbᵘ/bᵛ = bᵘ⁻ᵛ
powerlog_b(aʳ) = r log_b a(bᵘ)ʳ = bᵘʳ

OpenStax Calculus Volume 1, §1.5 Exponential and Logarithmic Functions §1.5, p. 86 — Rule: Properties of Logarithms

Follow the derivation of the product law. Give the two logarithms names, u and v, which by definition means b to the u is a and b to the v is c. Multiplying a and c then multiplies two powers of b, and the exponent law adds the exponents. Reading the result back as a logarithm says the logarithm of ac is u plus v.

The table puts each logarithm law beside the exponent law it comes from. Products of powers add exponents, so logarithms of products add. Quotients subtract. A power of a power multiplies exponents, so a power inside a logarithm comes out as a multiplier in front.

The power property is the workhorse of the next part of the lesson. It is the only tool that can bring an unknown down from an exponent, and every exponential equation you solve will use it.

35. A slide rule adds logarithms

Picture it

Figure (svg): Two rulers marked on a logarithmic scale, 1 to 10. The lower ruler is slid right so its 1 sits under the upper ruler's 2; then the lower ruler's 3 sits exactly under the upper ruler's 6. Arrows underneath show the length log 2 plus the length log 3 making the length log 6.

On a slide rule, distance is the logarithm. Laying one length after another adds logarithms, and the scale reads off the product. Multiplication has become addition.

Before calculators, engineers multiplied by sliding two logarithmic rulers. Lengths add, logarithms add, numbers multiply.

\[ \log 2 + \log 3 = 0.30103 + 0.47712 = 0.77815 = \log 6 \]

Logarithms were invented in the early 1600s for exactly this purpose: to turn hard multiplications into easy additions. A slide rule makes the idea physical. Its scales are marked so that the distance from 1 to any number is proportional to that number's logarithm.

Look at the figure. The sliding ruler has been moved so its 1 sits under the fixed ruler's 2, which offsets it by the length log 2. The sliding 3 is then a further length log 3 along, so it sits under the point at distance log 2 plus log 3 from the start. The fixed ruler reads 6 there, because log 2 plus log 3 is log 6.

For three centuries engineers multiplied this way. The product law on the previous slide is the reason it works, and the arithmetic under the figure checks it to five decimal places.

36. Trap: the logarithm of a sum

Trap

The trap

An invented law:

\[ \ln(a + c) = \ln a + \ln c \]

Wrong. The product law is about products.

The fix

Try a equal to c equal to 1. The left side is ln 2, the right side is 0. There is no law for the logarithm of a sum.

\[ \ln(1 + 1) \approx 0.693 \]

\[ \ln 1 + \ln 1 = 0 \]

\[ \ln(ac) = \ln a + \ln c \;\checkmark \]

This invented law looks so much like the product law that it is the most common error in the whole topic. The real law says the logarithm of a product is the sum of the logarithms. It says nothing about the logarithm of a sum, and in fact there is no simplification for that at all.

A single number exposes it. With a and c both equal to 1, the left side is ln 2, about 0.693, and the right side is ln 1 plus ln 1, which is 0. Whenever you want to split a logarithm, check that what is inside is a product, a quotient or a power. If it is a sum or a difference, the logarithm stays whole.

37. Is ln of x squared always 2 ln x?

Counterexample

\[ \ln\left(x^{2}\right) \overset{?}{=} 2\ln x \]

Discussion prompt

The power property seems to say so. Find a value of x where the left side makes sense but the right side does not, and write the version that is true for every nonzero x.

Write your counterexample before revealing. The power property does seem to say that the logarithm of x squared equals 2 times the logarithm of x, so it is worth asking what the property assumed.

Try x equal to minus 1. The left side is the natural log of 1, which is 0. The right side is 2 times the natural log of minus 1, which does not exist. So the two sides are not the same function: the left side is defined for every nonzero x, the right side only for positive x.

The power property was proved for a positive argument, and the fix is the absolute value: the logarithm of x squared equals 2 times the logarithm of the absolute value of x, for all nonzero x. The same care is needed whenever a power property changes the domain, which is why Example 1.37(c) records its domain first.

38. Solving equations

Section

Part 4

39. Example 1.36(a): the unknown in the exponent

Worked example

\[ 5^{x} = 2 \]

Take the natural log of both sides

Why: Equal positive numbers have equal logarithms.

\[ \ln\left(5^{x}\right) = \ln 2 \]

Apply the power property

Why: The exponent comes down as a multiplier; this is the whole point of taking logs.

\[ x\ln 5 = \ln 2 \]

Divide by ln 5

Why: ln 5 is a nonzero constant.

\[ x = \frac{\ln 2}{\ln 5} \]

Evaluate

Why: Both logarithms from a calculator.

\[ x \approx \frac{0.693147}{1.609438} \approx 0.430677 \]

Check by substituting back

Why: Five to that power should be 2.

\[ 5^{0.430677} \approx 2.000001 \]

OpenStax Calculus Volume 1, §1.5 Exponential and Logarithmic Functions §1.5, p. 86 — Example 1.36a

The unknown sits in the exponent, where ordinary algebra cannot reach it. The logarithm is the tool that brings it down. Taking the natural log of both sides is legitimate because both sides are the same positive number, and equal numbers have equal logarithms.

The power property then turns the natural log of 5 to the x into x times ln 5. Now x is an ordinary unknown multiplied by a constant, and dividing by ln 5 finishes the job. You could have used any base; the natural log is simply the one on every calculator.

The answer, about 0.43, makes sense: 5 to the 0 is 1 and 5 to the 1 is 5, so an exponent a little under a half should give something near the square root of 5, which is about 2.2. The check substitutes back and gets 2.

40. Example 1.36(b): a quadratic in disguise

Worked example

\[ e^{x} + 6e^{-x} = 5 \]

Multiply both sides by e to the x

Why: e to the x times e to the minus x is 1, so the negative exponent disappears.

\[ e^{2x} + 6 = 5e^{x} \]

Move everything to one side

Why: Subtract 5 e to the x.

\[ e^{2x} - 5e^{x} + 6 = 0 \]

Read it as a quadratic

Why: e to the 2x is the square of e to the x; call e to the x u.

\[ u^{2} - 5u + 6 = 0, \qquad u = e^{x} \]

Factor

Why: Two numbers with product 6 and sum 5.

\[ (u - 3)(u - 2) = 0 \]

Return to x

Why: Both values are positive, so both are possible values of e to the x.

\[ e^{x} = 3 \quad\text{or}\quad e^{x} = 2 \]

Take logarithms

Why: ln undoes the exponential.

\[ x = \ln 3 \quad\text{or}\quad x = \ln 2 \]

Figure (svg): A zoomed graph of y equals e to the x plus 6 e to the minus x, for x between 0.2 and 1.6. The curve dips to a minimum of about 4.899 at x about 0.896 and crosses the horizontal line y equals 5 twice, at x equals ln 2 and x equals ln 3.

The curve dips only a tenth below 5, so it meets the line exactly twice. Two crossings, two solutions: the algebra's two factors are these two points.

Check both roots

Why: e to the ln 2 is 2 and e to the minus ln 2 is one half; likewise for 3.

\[ 2 + 6\cdot\tfrac12 = 5, \qquad 3 + 6\cdot\tfrac13 = 5 \]

OpenStax Calculus Volume 1, §1.5 Exponential and Logarithmic Functions §1.5, pp. 86-87 — Example 1.36b

Taking logarithms straight away would fail here, because the left side is a sum and the logarithm of a sum does not split. The trick is to clear the negative exponent: multiply every term by e to the x, and since e to the x times e to the minus x is 1, the equation becomes e to the 2x plus 6 equals 5 e to the x.

Now look at it with fresh eyes. E to the 2x is the square of e to the x, so if you call e to the x u, you have an ordinary quadratic in u. It factors, giving u equal to 3 or 2. Both are positive, so both are genuine values of e to the x, and taking natural logs gives x equal to ln 3 or ln 2.

The figure shows why there are exactly two answers: the curve dips just below 5 and comes back up, crossing the line twice. The check substitutes each root and uses e to the ln 2 equal to 2 to get 5 exactly.

41. Checkpoint 1.30: what is the first move?

Step zero

\[ \frac{e^{2x}}{3 + e^{2x}} = \frac12 \]

Discussion prompt

Before touching a logarithm, decide the first move. Why would taking ln of both sides right now be a poor choice?

Write your answer before revealing. The instinct in an exponential equation is to take logarithms immediately, and this slide is here to slow that instinct down.

If you take the natural log of both sides now, the left side becomes the logarithm of a fraction whose denominator is 3 plus e to the 2x. The quotient property splits off the denominator, but you are left with the logarithm of a sum, and nothing simplifies that.

Clearing the fraction first changes everything. After cross-multiplying, every term is a multiple of e to the 2x or a constant, and the exponential terms collect into one. Only then is taking logarithms useful. The general rule: isolate a single exponential before you reach for the logarithm.

42. Checkpoint 1.30: clear the fraction first

Worked example

\[ \frac{e^{2x}}{3 + e^{2x}} = \frac12 \]

Cross-multiply

Why: The denominator 3 plus e to the 2x is positive, so this is safe.

\[ 2e^{2x} = 3 + e^{2x} \]

Collect the exponential terms

Why: Subtract e to the 2x from both sides.

\[ e^{2x} = 3 \]

Take the natural log

Why: ln undoes e.

\[ 2x = \ln 3 \]

Divide by 2

Why: Then evaluate.

\[ x = \frac{\ln 3}{2} \approx 0.549306 \]

Check by substituting back

Why: At that x, e to the 2x is exactly 3.

\[ \frac{3}{3 + 3} = \frac12 \]

OpenStax Calculus Volume 1, §1.5 Exponential and Logarithmic Functions §1.5, p. 87 — Checkpoint 1.30

Cross-multiplying is safe here because the denominator, 3 plus e to the 2x, is always positive, so you are not multiplying by zero. The result, 2 e to the 2x equals 3 plus e to the 2x, has the exponential on both sides.

Subtracting e to the 2x from both sides isolates a single exponential, e to the 2x equals 3. Now the natural log does its job in one step, and dividing by 2 gives x equal to half of ln 3, about 0.549.

The check is almost free: at this x, e to the 2x is exactly 3, so the fraction is 3 over 6, which is one half. Notice the answer happens to equal the inverse hyperbolic tangent of one half that appears at the end of the lesson; that is not a coincidence, as the tanh formula is built from exactly this kind of equation.

43. Example 1.37(a): ln of one over x equals 4

Prediction

\[ \ln\left(\frac1x\right) = 4 \]

Predict first

Convert to exponential form. What is x?

  • x = e⁴
  • x = e⁻⁴
  • x = −4
  • x = 1/4

Correct: x = e⁻⁴

Why: ln of one over x equals 4 means e to the 4 equals one over x, so x is one over e to the 4, which is e to the minus 4, about 0.0183. Check: one over x is then e to the 4, whose natural log is 4.

\[ \ln\left(\tfrac1x\right) = 4 \iff e^{4} = \frac1x \iff x = e^{-4} \approx 0.0183 \]

OpenStax Calculus Volume 1, §1.5 Exponential and Logarithmic Functions §1.5, p. 87 — Example 1.37a

Choose your answer before revealing. The move is to convert from logarithmic form to exponential form: ln of something equal to 4 means e to the 4 equals that something.

So e to the 4 equals one over x, and flipping both sides gives x equal to one over e to the 4, which is e to the minus 4, about 0.0183. The most common wrong answer is e to the 4, which comes from forgetting that the argument was one over x rather than x.

Check it: if x is e to the minus 4 then one over x is e to the 4, and the natural log of that is 4. The answer is positive, so it is in the domain. You could also have used the quotient property first to write minus ln x equals 4, which lands in the same place.

44. Example 1.37(b): combine, then exponentiate

Worked example

\[ \log_{10}\sqrt{x} + \log_{10}x = 2 \]

Product property

Why: A sum of logarithms is the logarithm of a product.

\[ \log_{10}\left(\sqrt{x}\cdot x\right) = 2 \]

Combine the powers of x

Why: The square root is the one-half power, and one half plus one is three halves.

\[ \log_{10}x^{3/2} = 2 \]

Power property

Why: The exponent comes out in front.

\[ \frac32\log_{10}x = 2 \]

Solve for the logarithm

Why: Multiply both sides by two thirds.

\[ \log_{10}x = \frac43 \]

Convert to exponential form

Why: Base 10 to the value gives the argument.

\[ x = 10^{4/3} = 10\sqrt[3]{10} \approx 21.5443 \]

Check by substituting back

Why: x is positive, so both logarithms exist; add them.

\[ \log_{10}4.64158 + \log_{10}21.5443 \approx 0.66667 + 1.33333 = 2 \]

OpenStax Calculus Volume 1, §1.5 Exponential and Logarithmic Functions §1.5, p. 87 — Example 1.37b

When the unknown sits inside logarithms, the plan runs in reverse: gather everything into a single logarithm, then exponentiate to release the unknown. Here the product property combines the two logarithms, and root x times x is x to the three halves.

The power property brings the three halves out in front, and dividing gives log base 10 of x equal to four thirds. Now convert to exponential form: x is 10 to the four thirds, which you can write as 10 times the cube root of 10, about 21.54.

The check substitutes back. The logarithm of the square root of x is about 0.667, the logarithm of x itself is about 1.333, and they add to 2. The answer is positive, so both logarithms in the original equation exist.

45. Example 1.37(c): check for extraneous solutions

Worked example

\[ \ln(2x) - 3\ln\left(x^{2}\right) = 0 \]

Record the domain first

Why: ln of 2x needs 2x positive.

\[ x > 0 \]

Power property

Why: 3 times ln of x squared is ln of x to the sixth.

\[ \ln(2x) - \ln\left(x^{6}\right) = 0 \]

Quotient property

Why: Then cancel one power of x.

\[ \ln\frac{2x}{x^{6}} = \ln\frac{2}{x^{5}} = 0 \]

Exponentiate

Why: ln of something is 0 exactly when that something is 1.

\[ \frac{2}{x^{5}} = 1 \]

Solve

Why: Take the fifth root.

\[ x^{5} = 2 \;\Longrightarrow\; x = \sqrt[5]{2} \approx 1.148698 \]

Check the domain and the equation

Why: The root is positive, so it is not extraneous; substitute it.

\[ \ln(2.297397) - 3\ln(1.319508) \approx 0.831777 - 0.831777 = 0 \]

OpenStax Calculus Volume 1, §1.5 Exponential and Logarithmic Functions §1.5, p. 87 — Example 1.37c

The first step writes down the domain before any algebra. The natural log of 2x needs 2x to be positive, so only positive x can be answers. Recording this first means you already know what to throw away at the end.

The power property turns 3 times the natural log of x squared into the natural log of x to the sixth, and the quotient property combines the two logarithms into the log of 2 over x to the fifth. A natural log equals 0 exactly when its argument is 1, so 2 over x to the fifth is 1 and x is the fifth root of 2.

The check does two things. It confirms the root is positive, so it is in the domain, and it substitutes the root into the original equation, where both logarithms come out to about 0.8318 and cancel. The book's phrase, check for extraneous solutions, is a habit, not a formality: the next exercise shows a root that fails.

46. Checkpoint 1.31: two logarithms of x

Fill the middle

Fill in the blanks

\ln\left(x^e^{-1}\right) - 4\ln x = 3\ln x - 4\ln x = -\ln x = 1 \;\Longrightarrow\; x = ___

Why: The power property turns ln of x cubed into 3 ln x, and 3 ln x minus 4 ln x is minus ln x. So ln x is minus 1 and x is e to the minus 1, one over e, about 0.368. It is positive, so it is in the domain.

\[ \text{check: } \ln\left(e^{-3}\right) - 4\ln\left(e^{-1}\right) = -3 + 4 = 1 \]

OpenStax Calculus Volume 1, §1.5 Exponential and Logarithmic Functions §1.5, p. 87 — Checkpoint 1.31

Fill in the blank before checking. Both terms are logarithms of powers of x, so the power property turns them into multiples of ln x, and they combine like like terms.

Three ln x minus four ln x is minus ln x, so ln x is minus 1 and x is e to the minus 1, one over e, about 0.368. The answer is positive, so the logarithms exist there.

The check under the cloze substitutes back: the natural log of e to the minus 3 is minus 3, and minus 4 times the natural log of e to the minus 1 is plus 4. Together they make 1.

47. Exercise 291: a root that must be thrown away

Worked example

\[ \ln x + \ln(x - 2) = \ln 4 \]

Record the domain

Why: Both x and x minus 2 must be positive, and the second condition is stronger.

\[ x > 2 \]

Combine the left side

Why: Product property.

\[ \ln\big(x(x - 2)\big) = \ln 4 \]

Drop the logarithms

Why: ln is one-to-one, so equal logarithms mean equal arguments.

\[ x^{2} - 2x = 4 \]

Solve the quadratic

Why: Quadratic formula on x squared minus 2x minus 4 equal to 0.

\[ x = \frac{2 \pm \sqrt{4 + 16}}{2} = 1 \pm \sqrt5 \]

Discard the root outside the domain

Why: 1 minus root 5 is negative, so ln x would not exist there.

\[ 1 - \sqrt5 \approx -1.236 \quad\Longrightarrow\quad x = 1 + \sqrt5 \approx 3.236 \]

Figure (svg): The curve y equals ln x plus ln of x minus 2, defined only for x greater than 2, crosses the line y equals ln 4 once, at x equals 1 plus root 5. A dashed curve y equals ln of x squared minus 2x, defined also for negative x, crosses the same line a second time at x equals 1 minus root 5, marked with a hollow dot. The region x at most 2 is shaded as outside the domain.

Combining the two logarithms into one quietly enlarges the domain: the dashed combined curve lives on the left too, and that is where the second root comes from. The original equation never had it.

Check the survivor

Why: Both logarithms exist and add to ln 4.

\[ \ln 3.236068 + \ln 1.236068 \approx 1.174359 + 0.211935 = 1.386294 = \ln 4 \]

OpenStax Calculus Volume 1, §1.5 Exponential and Logarithmic Functions §1.5, p. 96 — Exercise 291

This exercise from the section's problem set shows why the domain step matters. Both logarithms need positive arguments, so x must exceed 0 and exceed 2; the second condition wins.

Combining the logarithms, dropping them because ln is one-to-one, and rearranging gives a quadratic that does not factor nicely, so use the quadratic formula. It gives two roots, 1 plus root 5 and 1 minus root 5. The second is about minus 1.24, which is not greater than 2, so it is thrown away.

The figure explains where the phantom root comes from. The solid curve is the original left side, which only exists to the right of 2. The dashed curve is the combined logarithm of x squared minus 2x, which also exists for negative x. The combined equation has a crossing on the left that the original equation never had.

48. Trap: keeping a root outside the domain

Trap

The trap

After solving x squared minus 2x equal to 4:

\[ x = 1 + \sqrt5 \text{ and } \]

\[ x = 1 - \sqrt5 \]

Wrong as an answer to the logarithm equation.

The fix

Combining logarithms produced an equation with a bigger domain than the one you started with. Substitute each root into the ORIGINAL equation.

\[ \ln(1 - \sqrt5) \text{ is undefined} \]

\[ x = 1 + \sqrt5 \text{ only} \]

Solving the quadratic correctly and then reporting both roots is the most common way to lose marks on logarithm equations. The algebra was fine; the final answer was not checked against the original equation.

The reason extra roots appear is that the step which combined the logarithms quietly widened the domain. The logarithm of x times x minus 2 exists whenever that product is positive, which happens for negative x too. So the combined equation can have solutions that the original cannot accept. Substitute every root into the original, and reject any that make an argument zero or negative.

49. Change of base

Section

Part 5

50. Deriving the change-of-base formula

Concept

Calculators have only log and ln. To reach any other base, name the three logarithms involved:

\[ u = \log_b a, \quad v = \log_a x, \quad w = \log_b x \]

\[ b^{u} = a, \quad a^{v} = x, \quad b^{w} = x \]

\[ b^{uv} = \left(b^{u}\right)^{v} = a^{v} = x = b^{w} \]

Exponentials are one-to-one, so the exponents agree: uv equals w.

\[ \log_a x = \frac{\log_b x}{\log_b a} \qquad \text{and with } b = e: \quad \log_a x = \frac{\ln x}{\ln a} \]

\[ \text{the same idea for powers: } a^{x} = b^{x\log_b a} = e^{x\ln a} \]

OpenStax Calculus Volume 1, §1.5 Exponential and Logarithmic Functions §1.5, pp. 87-88 — Rule: Change-of-Base Formulas and proof

Calculators offer only base 10 and base e, so you need a way to compute any other logarithm from those. The derivation names the three logarithms involved, u, v and w, and rewrites each as an exponential statement.

The middle line is the heart of it. Raise b to the product uv. By the power-of-a-power law this is b to the u, raised to the v. But b to the u is a, and a to the v is x, and x is also b to the w. So b to the uv equals b to the w, and because exponentials are one-to-one, uv equals w. Solving for v gives the formula.

The last line is the first change-of-base formula, for powers rather than logarithms: any exponential can be rewritten with base e. In Chapter 3 this is exactly how you will differentiate 2 to the x: rewrite it as e to the x ln 2.

51. Reading the change-of-base formula

Notation

Annotate

On: \( \log_a x = \frac{\ln x}{\ln a} \)

  • a, the awkward base, ends up in the DENOMINATOR.
  • x, the number whose logarithm you want, goes on top.
  • log x / log a gives the same answer, because every logarithm is a constant multiple of every other.
  • If x = a the fraction is 1, and log_a a is 1. Swapping top and bottom fails this check.

Step through the notes. The only thing to get right is which logarithm goes where, and the rule is that the base you want to get rid of goes underneath.

The sanity check in the last note is the best way to remember it. If the argument equals the base, the logarithm must be 1. With ln x over ln a, putting x equal to a gives 1, as it should. The flipped fraction gives 1 as well in that case, so try a second test: the logarithm base 2 of 8 must be 3, and ln 8 over ln 2 is 3, while ln 2 over ln 8 is one third.

The third note says any base can play the role of e here. Every logarithm is a constant multiple of every other, so the ratio comes out the same whichever base you use to compute it.

52. Example 1.38: log base 3 of 7

Worked example

\[ \log_3 7 \]

Use the second change-of-base formula with base e

Why: Natural logarithms are on every calculator.

\[ \log_a x = \frac{\ln x}{\ln a} \]

Put a equal to 3 and x equal to 7

Why: The base goes underneath.

\[ \log_3 7 = \frac{\ln 7}{\ln 3} \]

Evaluate

Why: Divide the two natural logarithms.

\[ \frac{1.945910}{1.098612} \approx 1.77124 \]

Checkpoint 1.32 the same way

Why: Base 4, argument 6.

\[ \log_4 6 = \frac{\ln 6}{\ln 4} = \frac{1.791759}{1.386294} \approx 1.29248 \]

Check by raising the base to the answer

Why: Each power should return the argument. (The book's solution says to use a equal to 3 and e equal to 3; it means a equal to 3 with base e.)

\[ 3^{1.77124} \approx 7.0000, \qquad 4^{1.29248} \approx 6.0000 \]

OpenStax Calculus Volume 1, §1.5 Exponential and Logarithmic Functions §1.5, p. 88 — Example 1.38 and Checkpoint 1.32

This is the change-of-base formula used as intended: to get a number out of a calculator that has no base 3 button. The base goes underneath, so the answer is ln 7 over ln 3, about 1.771.

A quick reasonableness check before the formal one: 3 to the 1 is 3 and 3 to the 2 is 9, and 7 is between them, closer to 9. So the answer should be between 1 and 2, closer to 2. It is. The checkpoint, log base 4 of 6, goes exactly the same way and lands between 1 and 2 as well.

The final check raises each base to its computed logarithm and recovers 7 and 6. One more note about the book: its solution says to use a equal to 3 and e equal to 3, which is a misprint. It means a equal to 3 with the natural logarithm, base e, as the calculator's base.

53. Example 1.39: Japan 2011 against Haiti 2010

Worked example

The Richter magnitudes and seismograph amplitudes of two earthquakes satisfy the relation below. Compare Japan, magnitude 9, with Haiti, magnitude 7.3.

Write the Richter relation

Why: The difference of magnitudes is the base-10 logarithm of the amplitude ratio.

\[ R_1 - R_2 = \log_{10}\left(\frac{A_1}{A_2}\right) \]

Substitute the magnitudes

Why: Japan first.

\[ 9 - 7.3 = \log_{10}\left(\frac{A_1}{A_2}\right) \]

Simplify

Why: The difference is 1.7.

\[ \log_{10}\left(\frac{A_1}{A_2}\right) = 1.7 \]

Convert to exponential form

Why: Base 10 to the 1.7.

\[ \frac{A_1}{A_2} = 10^{1.7} \approx 50.12 \]

Figure (svg): Left: a magnitude scale from 6 to 10 lined up with an amplitude scale from 10 to the 6 to 10 to the 10, so each step of 1 multiplies the amplitude by 10; Haiti at 7.3 and Japan at 9.0 are marked, 1.7 apart. Right: bars drawn to true scale for the amplitudes of magnitudes 7, 7.3, 8 and 9, heights 1, 2, 10 and 100; the first two bars are almost invisible.

The left scale is logarithmic, so 7.3 and 9.0 look close together. Drawn to true size on the right, the magnitude 9 amplitude dwarfs the magnitude 7.3 one by a factor of about 50.

Check by splitting the exponent

Why: 1.7 is 1 plus 0.7, and 10 to the 0.7 is about 5, so the ratio is about 10 times 5.

\[ 10^{1.7} = 10^{1}\cdot 10^{0.7} \approx 10\times 5.012 = 50.12 \]

OpenStax Calculus Volume 1, §1.5 Exponential and Logarithmic Functions §1.5, pp. 88-89 — Example 1.39

The Richter relation says the difference of two magnitudes is the base 10 logarithm of the ratio of their amplitudes. So every one-unit step on the scale multiplies the amplitude by 10, and a difference of 1.7 multiplies it by 10 to the 1.7.

Converting from logarithmic to exponential form gives the ratio as about 50. The check splits the exponent into 1 plus 0.7: 10 to the 1 is 10, and 10 to the 0.7 is about 5, so the product is about 50. That split is how you can estimate such powers without a calculator.

Look at the two panels. On the left, the logarithmic scale makes 7.3 and 9.0 look like neighbours. On the right the same amplitudes are drawn to true size, and the magnitude 9 bar towers over the others while the 7.3 bar barely shows. That contrast is the whole reason for using a logarithmic scale.

54. Checkpoint 1.33: magnitude 8.4 against 7.4

Estimation

Predict first

How many times the amplitude is a magnitude 8.4 earthquake compared with a magnitude 7.4 one?

  • About 1.14 times
  • 10 times
  • 84/74 times
  • 100 times

Correct: 10 times

Why: Only the difference of magnitudes matters: 8.4 minus 7.4 is 1, and 10 to the 1 is 10. The ratio 84 over 74, about 1.14, would be right for a linear scale, which the Richter scale is not.

\[ \log_{10}\frac{A_1}{A_2} = 8.4 - 7.4 = 1 \;\Longrightarrow\; \frac{A_1}{A_2} = 10 \]

OpenStax Calculus Volume 1, §1.5 Exponential and Logarithmic Functions §1.5, p. 89 — Checkpoint 1.33

Choose an answer before revealing. The tempting choice treats the magnitudes like ordinary sizes and divides one by the other, but the Richter scale is logarithmic, so ratios of amplitude come from differences of magnitude.

The difference is 1, so the amplitude ratio is 10 to the 1, which is 10. Notice that the actual magnitudes did not matter, only their difference. An earthquake of magnitude 3.4 compared with one of magnitude 2.4 also has a ratio of 10.

This is the general feature of any logarithmic scale: equal steps on the scale mean equal factors in the quantity. You will see the same feature in the pH scale on the next slide.

55. pH is a logarithm too

Real world

\[ \text{pH} = -\log_{10}[\text{H}^{+}] \]

Discussion prompt

Exercise 303: find the pH of eggs, beer and tomato juice, whose hydrogen-ion concentrations are 1.6 × 10⁻⁸, 3.16 × 10⁻³ and 7.94 × 10⁻⁵ mol/L, and say which are acids (pH below 7) and which are bases.

Figure (svg): A pH number line from 0 to 14 with water marked at 7. Beer is placed at 2.5 and tomato juice at 4.1 on the acid side, eggs at 7.8 on the base side.

Each step left on this line means ten times as many hydrogen ions. Beer, at 2.5, is about 40 times as acidic as tomato juice at 4.1.

OpenStax Calculus Volume 1, §1.5 Exponential and Logarithmic Functions §1.5, p. 97 — Exercise 303

Work out the three values before revealing. The pH is minus the base 10 logarithm of the hydrogen-ion concentration, and for numbers written in scientific notation the exponent does most of the work: a concentration of about 10 to the minus 8 gives a pH of about 8.

The more careful values are 7.8 for eggs, 2.5 for beer and 4.1 for tomato juice. Water sits at 7, so eggs are slightly basic and the other two are acids. The minus sign in the definition is why a larger concentration of hydrogen ions gives a smaller pH.

The number line in the answer shows the practical meaning. Beer and tomato juice are 1.6 apart on the scale, so beer has about 10 to the 1.6, roughly 40, times the hydrogen-ion concentration. Once again, equal steps on the logarithmic scale are equal factors in the underlying quantity.

56. The hyperbolic functions

Section

Part 6

57. Cosh and sinh are built from e

Concept

\[ \cosh x = \frac{e^{x} + e^{-x}}{2}, \qquad \sinh x = \frac{e^{x} - e^{-x}}{2} \]

Figure (svg): Two panels. Left: y equals cosh x, a U-shaped curve with lowest point (0, 1), lying just above the dashed guides one half e to the x and one half e to the minus x, which it averages. Right: y equals sinh x, an S-shaped curve through the origin, squeezed between the dashed guides one half e to the x and minus one half e to the minus x.

Far to the right both functions are essentially one half e to the x. Far to the left, cosh follows one half e to the minus x upward while sinh follows its negative downward.

Cosh is the average of e to the x and e to the minus x, so it is even with lowest value 1. Sinh is half their difference, so it is odd and passes through the origin.

\[ \cosh(-x) = \cosh x, \qquad \sinh(-x) = -\sinh x, \qquad \cosh x + \sinh x = e^{x} \]

OpenStax Calculus Volume 1, §1.5 Exponential and Logarithmic Functions §1.5, pp. 90-92 — Definition and Figure 1.51

The two definitions are built from the same two ingredients, e to the x and e to the minus x. Hyperbolic cosine is their average; hyperbolic sine is half their difference. That is all there is to them.

The figure shows each one riding on its ingredients. On the left, cosh lies just above both dashed curves, since it averages them, and its lowest point is at 0 comma 1, where the two ingredients are equal. On the right, sinh follows one half e to the x upward and minus one half e to the minus x downward, passing through the origin.

The identities under the figure follow straight from the definitions. Replacing x by minus x swaps the two ingredients, which leaves the average alone and flips the sign of the difference; so cosh is even and sinh is odd, like cos and sin. And adding cosh and sinh cancels the e to the minus x terms, leaving e to the x.

58. Tanh and the other three

Concept

\[ \tanh x = \frac{\sinh x}{\cosh x} = \frac{e^{x} - e^{-x}}{e^{x} + e^{-x}} \]

functiondefinitionin exponentials
csch x1/sinh x2/(eˣ − e⁻ˣ)
sech x1/cosh x2/(eˣ + e⁻ˣ)
coth xcosh x/sinh x(eˣ + e⁻ˣ)/(eˣ − e⁻ˣ)

Figure (svg): The curve y equals tanh x rising from near minus 1 on the left, through the origin, to near 1 on the right, between dashed asymptotes y equals 1 and y equals minus 1; and the bell-shaped curve y equals sech x with peak 1 at x equals 0.

Tanh is a ratio of sinh to cosh; since the two become nearly equal far out, tanh flattens toward 1 on the right and toward minus 1 on the left. Sech is one over cosh: a bump of height 1.

Say them as cosh like gosh, sinh like cinch, and tanch, seech, coseech, cotanch.

OpenStax Calculus Volume 1, §1.5 Exponential and Logarithmic Functions §1.5, pp. 90-92 — Definition and graphs of the hyperbolic functions

The remaining four hyperbolic functions are defined from cosh and sinh exactly as tangent, secant, cosecant and cotangent are defined from cos and sin. The table rewrites each in terms of exponentials.

The figure shows the two you will meet most. Tanh is the ratio of sinh to cosh, and far to the right those two are almost equal, so tanh approaches 1; far to the left sinh is almost minus cosh, so tanh approaches minus 1. It passes through the origin and never reaches either asymptote. Sech is one over cosh, so it is a bump with its peak of 1 at the origin, shrinking toward zero on both sides.

Unlike tan, tanh has no vertical asymptotes, because cosh is never zero. The pronunciations on the slide are the ones the book gives; you will hear others, and that is fine.

59. Why hyperbolic: cosh squared minus sinh squared

Concept

Square the two definitions and subtract.

\[ \cosh^{2}t = \frac{e^{2t} + 2 + e^{-2t}}{4}, \qquad \sinh^{2}t = \frac{e^{2t} - 2 + e^{-2t}}{4} \]

\[ \cosh^{2}t - \sinh^{2}t = \frac{2 - (-2)}{4} = 1 \]

Figure (svg): Two panels with equal scales. Left: the unit circle x squared plus y squared equals 1, with the point (cos 1, sin 1). Right: the right branch of the unit hyperbola x squared minus y squared equals 1, between dashed asymptotes y equals x and y equals minus x, with the point (cosh 1, sinh 1), about (1.543, 1.175).

Cos and sin put a point on the circle; cosh and sinh put a point on the hyperbola. The identity with the minus sign is the hyperbola's equation, and that is where the name comes from.

So the point with coordinates cosh t and sinh t always lies on the unit hyperbola, just as cos t and sin t put a point on the unit circle.

OpenStax Calculus Volume 1, §1.5 Exponential and Logarithmic Functions §1.5, pp. 90-91 — the identity and Figure 1.50

The derivation squares each definition. Squaring e to the t plus e to the minus t gives e to the 2t, plus twice their product, plus e to the minus 2t, and their product is 1. The same happens for sinh except that the middle term is minus 2. Subtracting, everything cancels except 2 minus minus 2, which is 4, and 4 over 4 is 1.

Now look at the figure. The trigonometric identity puts the point cos t, sin t on the unit circle. The hyperbolic identity puts the point cosh t, sinh t on the curve x squared minus y squared equals 1, which is the unit hyperbola. That is where the name comes from.

Because cosh t is always at least 1, the point always lies on the right-hand branch of the hyperbola. As t runs from minus infinity to infinity the point sweeps up that branch, hugging the dashed asymptotes at both ends.

60. The hanging chain

Picture it

Figure (svg): A chain hanging between two posts at x equals 0 and x equals 4, drawn as y equals cosh of x minus 2, from height 1 at the middle x equals 2 to 3.762 at the posts, with a dashed parabola through the same lowest point and the same two ends. The parabola runs visibly above the chain between the middle and the posts.

Galileo guessed a hanging chain was a parabola. It is not: the true shape is a hyperbolic cosine, which sags a little lower between the lowest point and the supports.

A chain or cable hanging under its own weight takes the shape of a hyperbolic cosine, called a catenary. Physics gives the height as below, for constants a and c.

\[ h(x) = a\cosh\left(\frac xa\right) + c \]

OpenStax Calculus Volume 1, §1.5 Exponential and Logarithmic Functions §1.5, pp. 89-90 — Figure 1.49 and the catenary

Hang a chain between two supports and it sags into a curve. For centuries people assumed that curve was a parabola; Galileo thought so. In 1691 several mathematicians showed independently that it is not. The true shape, called a catenary, is a hyperbolic cosine.

The figure compares the two. The solid curve is cosh x, shifted so its lowest point sits at x equal to 2, hanging between posts at 0 and 4. The dashed parabola passes through the same lowest point and the same two ends. They are close, but the parabola runs visibly above the chain between the middle and the posts; one unit from the middle the gap is about 0.15.

The constants a and c in the formula let the same shape fit any chain: a controls how tightly it hangs and c shifts it vertically. Suspension-bridge cables, power lines and the silk in a spider web all follow this curve, which is why the book's opening image is a web.

61. Trigonometric and hyperbolic identities side by side

Comparison

Comparison matrix

identitytrigonometrichyperbolic
c² ? s² = 1: plus or minus?plusminus
cos and cosh: even or odd?eveneven
sin and sinh: even or odd?oddodd
1 ? tan² = sec²: plus or minus?plusminus
cos and cosh: bounded?boundedunbounded
trigonometrichyperbolic
cos²t + sin²t = 1cosh²t − sinh²t = 1
1 + tan²x = sec²x1 − tanh²x = sech²x
cot²x + 1 = csc²xcoth²x − 1 = csch²x
sin(x ± y) = sin x cos y ± cos x sin ysinh(x ± y) = sinh x cosh y ± cosh x sinh y

OpenStax Calculus Volume 1, §1.5 Exponential and Logarithmic Functions §1.5, p. 92 — Rule: Identities Involving Hyperbolic Functions

Fill in the blank cells before checking. Every hyperbolic identity looks like its trigonometric partner, and the differences are almost always just signs.

The Pythagorean identity is the model: a plus for the circle, a minus for the hyperbola. The tangent-secant identity changes sign the same way. Both cos and cosh are even and both sin and sinh are odd, so those rows match exactly. The one real difference in character is boundedness: cos stays between minus 1 and 1 forever, while cosh grows without bound like one half e to the x.

The table underneath lists the full identities from the book for reference. The addition formulas carry over with no sign change at all for sinh, and you can verify any of these by writing everything in exponentials and simplifying, as you did for the Pythagorean one.

62. Example 1.40(a): sinh of 5 ln x

Worked example

\[ \sinh(5\ln x), \qquad x > 0 \]

Apply the definition of sinh

Why: Replace x in the definition by 5 ln x.

\[ \sinh(5\ln x) = \frac{e^{5\ln x} - e^{-5\ln x}}{2} \]

Move the coefficients inside the logarithms

Why: Power property, backwards.

\[ = \frac{e^{\ln(x^{5})} - e^{\ln(x^{-5})}}{2} \]

Let the exponential undo each logarithm

Why: e to the ln of anything positive is that thing.

\[ = \frac{x^{5} - x^{-5}}{2} \]

Check at x equal to 2

Why: 5 ln 2 is about 3.465736.

\[ \sinh(3.465736) \approx 15.984375 = \frac{32 - \frac{1}{32}}{2} \]

OpenStax Calculus Volume 1, §1.5 Exponential and Logarithmic Functions §1.5, p. 93 — Example 1.40a

Whenever a hyperbolic function has a logarithm inside it, go straight to the exponential definition, because the exponential will undo the logarithm. Replacing x in the definition by 5 ln x gives e to the 5 ln x minus e to the minus 5 ln x, all over 2.

The power property, used backwards, moves each coefficient inside its logarithm: 5 ln x is the log of x to the fifth, and minus 5 ln x is the log of x to the minus fifth. Then e to the ln of anything is that thing, so the logarithms disappear and you are left with a plain algebraic expression.

The check at x equal to 2 evaluates sinh of 5 ln 2 directly on a calculator and compares with the simplified form, 32 minus one thirty-second, all over 2. Both give 15.984375.

63. Example 1.40(b): from sinh x to the other five

Worked example

\[ \sinh x = \frac34 \]

Use the hyperbolic Pythagorean identity

Why: cosh squared is 1 plus sinh squared.

\[ \cosh^{2}x = 1 + \left(\frac34\right)^{2} = \frac{25}{16} \]

Take the positive root

Why: cosh x is at least 1 for every x, so the negative root is impossible.

\[ \cosh x = \frac54 \]

Divide for tanh

Why: Sinh over cosh.

\[ \tanh x = \frac{3/4}{5/4} = \frac35 \]

Take reciprocals for the other three

Why: Csch, sech and coth are the reciprocals of sinh, cosh and tanh.

\[ \operatorname{csch} x = \frac43, \quad \operatorname{sech} x = \frac45, \quad \coth x = \frac53 \]

Check with the actual x

Why: Inverting sinh gives x equal to ln 2, and cosh of ln 2 is the average of 2 and one half.

\[ x = \ln\left(\tfrac34 + \tfrac54\right) = \ln 2, \quad \cosh(\ln 2) = \frac{2 + \frac12}{2} = \frac54 \]

OpenStax Calculus Volume 1, §1.5 Exponential and Logarithmic Functions §1.5, p. 93 — Example 1.40b

This mirrors a familiar trigonometry exercise: given one function's value, find the rest. The hyperbolic Pythagorean identity gives cosh squared as 1 plus sinh squared, which is 25 over 16.

Choosing the root is where you need a fact about the hyperbolic functions rather than trigonometry. Cosh is at least 1 for every x, so the negative root is impossible and cosh x is five quarters. In trigonometry you would have needed to know the quadrant; here there is no choice to make. Tanh is sinh over cosh, three fifths, and the other three are reciprocals.

The check finds the actual x. Using the inverse sinh formula from later in the lesson, x is ln of three quarters plus five quarters, which is ln 2. Cosh of ln 2 is the average of 2 and one half, which is five quarters, confirming the value.

64. Checkpoint 1.34: cosh of 2 ln x

Fill the middle

Fill in the blanks

\cosh(2\ln x) = \fracx^{-2} + e^___}___ = \frac___ + ___}___

Why: Exactly as in Example 1.40(a): minus 2 ln x is the logarithm of x to the minus 2, and e undoes the logarithm. So cosh of 2 ln x is x squared plus x to the minus 2, all over 2.

\[ \text{check at } x = 2: \quad \cosh(2\ln 2) = \cosh(1.386294) \approx 2.125 = \frac{4 + \frac14}{2} \]

OpenStax Calculus Volume 1, §1.5 Exponential and Logarithmic Functions §1.5, p. 93 — Checkpoint 1.34

Fill in the blank before checking. This is Example 1.40(a) again with cosh instead of sinh and 2 instead of 5, so follow the same two moves: the definition, then let e undo the logarithm.

E to the 2 ln x is x squared, and e to the minus 2 ln x is x to the minus 2, which you can also write as one over x squared. Because cosh adds rather than subtracts, the answer is x squared plus x to the minus 2, all over 2.

The check at x equal to 2 gives 4 plus one quarter, over 2, which is 2.125, and cosh of 2 ln 2 on a calculator agrees.

65. Inverse hyperbolic functions are logarithms

Concept

Restrict cosh and sech to x at least 0 and every hyperbolic function is one-to-one. Invert sinh by solving for y:

\[ y = \sinh^{-1}x \iff x = \sinh y = \frac{e^{y} - e^{-y}}{2} \]

\[ e^{y} - 2x - e^{-y} = 0 \;\Longrightarrow\; e^{2y} - 2xe^{y} - 1 = 0 \]

\[ e^{y} = \frac{2x \pm \sqrt{4x^{2} + 4}}{2} = x \pm \sqrt{x^{2} + 1} \]

e to the y is positive and the minus sign gives a negative number, so only the plus sign survives:

\[ \sinh^{-1}x = \ln\left(x + \sqrt{x^{2} + 1}\right) \]

inverseas a logarithm
cosh⁻¹x (x ≥ 1)ln(x + √(x² − 1))
tanh⁻¹x (|x| < 1)½ ln((1 + x)/(1 − x))
coth⁻¹x (|x| > 1)½ ln((x + 1)/(x − 1))
sech⁻¹x (0 < x ≤ 1)ln((1 + √(1 − x²))/x)
csch⁻¹x (x ≠ 0)ln(1/x + √(1 + x²)/|x|)

OpenStax Calculus Volume 1, §1.5 Exponential and Logarithmic Functions §1.5, p. 93 — Definition: Inverse Hyperbolic Functions

Since the hyperbolic functions are built from exponentials, their inverses should be built from logarithms, and the derivation shows how. Start from y equal to the inverse sinh of x, rewrite it as x equal to sinh y, and use the definition.

Multiplying through by e to the y turns the equation into a quadratic in e to the y, the same trick as Example 1.36(b). The quadratic formula gives e to the y as x plus or minus the square root of x squared plus 1. That square root is always bigger than the absolute value of x, so the minus sign gives a negative number, and e to the y cannot be negative. Only the plus sign survives, and taking the natural log gives the formula.

The table lists all six inverses with their domains. Cosh and sech had to be restricted to x at least 0 before inverting, because they are even and fail the horizontal line test otherwise. You will meet these formulas again as the answers to certain integrals.

66. Example 1.41: evaluating inverse hyperbolic functions

Worked example

\[ \sinh^{-1}(2), \qquad \tanh^{-1}\left(\tfrac14\right) \]

Recall the formula for inverse sinh

Why: Derived on the previous slide.

\[ \sinh^{-1}x = \ln\left(x + \sqrt{x^{2} + 1}\right) \]

Put x equal to 2

Why: Two squared plus one is 5.

\[ \sinh^{-1}(2) = \ln\left(2 + \sqrt5\right) \approx 1.4436 \]

Recall the formula for inverse tanh

Why: Valid for x strictly between minus 1 and 1.

\[ \tanh^{-1}x = \frac12\ln\left(\frac{1 + x}{1 - x}\right) \]

Put x equal to one quarter

Why: Five quarters over three quarters is five thirds.

\[ \tanh^{-1}\tfrac14 = \frac12\ln\frac{5/4}{3/4} = \frac12\ln\frac53 \approx 0.2554 \]

Checkpoint 1.35 the same way

Why: Three halves over one half is 3.

\[ \tanh^{-1}\tfrac12 = \frac12\ln\frac{3/2}{1/2} = \frac12\ln 3 \approx 0.5493 \]

Check by applying sinh and tanh

Why: Each should return the input.

\[ \sinh(1.443635) \approx 2.0000, \qquad \tanh(0.255413) \approx 0.2500 \]

OpenStax Calculus Volume 1, §1.5 Exponential and Logarithmic Functions §1.5, p. 94 — Example 1.41 and Checkpoint 1.35

With the logarithm formulas in hand, evaluating an inverse hyperbolic function is just substitution. For inverse sinh of 2, the square root of 4 plus 1 is root 5, so the answer is the natural log of 2 plus root 5, about 1.4436.

For inverse tanh the formula is half the natural log of 1 plus x over 1 minus x. With x equal to one quarter the fraction is five quarters over three quarters, which is five thirds, giving about 0.2554. The checkpoint with x equal to one half gives half of ln 3, about 0.5493, which is the same number you found in Checkpoint 1.30.

The check applies sinh and tanh to the answers and recovers 2 and one quarter. That is the definition of an inverse function in action, and it is always the quickest way to confirm an inverse calculation.

67. Putting it together

Section

Part 7

68. Pattern: solving exponential and logarithmic equations

Pattern

Figure (svg): A flow diagram with two rows. Top row, for an unknown in an exponent: isolate the exponential, take ln of both sides, bring the exponent down with the power law, solve and substitute back. Bottom row, for an unknown inside a logarithm: write down the domain, combine into one logarithm, exponentiate both sides, solve and discard any root outside the domain.

Both routes do the same thing: apply the inverse function to peel the unknown out. The logarithm route has one extra duty, which is to check the domain at both ends.
  1. Unknown in an exponent: isolate the exponential (clear fractions, or treat e to the x as u when you see e to the 2x), then take ln of both sides.
  2. Bring the exponent down with the power property, and solve the resulting linear or quadratic equation.
  3. Unknown inside a logarithm: write the domain FIRST, combine into a single logarithm, then exponentiate.
  4. Check every root in the original equation; discard any that make an argument zero or negative.
  5. Other bases: convert with log_a x = ln x / ln a.

The diagram shows the two routes this lesson has used again and again. If the unknown is in an exponent, isolate the exponential and apply the logarithm, which brings the unknown down. If the unknown is inside a logarithm, combine the logarithms and apply the exponential, which releases it.

The two routes are mirror images, because exponential and logarithm are inverses. But they are not equally safe. The logarithm route changes the domain when you combine logarithms, so it has two extra duties, highlighted in the diagram: write the domain down first, and check every root against it at the end.

The list adds the details. Clear fractions and spot hidden quadratics before taking logarithms, never split the logarithm of a sum, and use change of base whenever a base other than e or 10 appears.

69. Put the solving steps in order

Ranking

\[ \log_2 x + \log_2(x - 2) = 3 \]

Put in order

Order the steps of a complete solution.

  1. Note the domain: x > 2
  2. Combine: log₂(x(x − 2)) = 3
  3. Exponentiate: x² − 2x = 2³ = 8
  4. Solve: (x − 4)(x + 2) = 0, so x = 4 or x = −2
  5. Discard x = −2; the answer is x = 4

Why: The domain comes first so that you know what to discard at the end. Combining and exponentiating produce a quadratic whose domain is larger than the original, which is exactly why minus 2 appears and must be thrown out. Check: log base 2 of 4 plus log base 2 of 2 is 2 plus 1, which is 3.

Put the steps in order before checking. The one real decision is where the domain goes, and the answer is first, so that you know in advance which roots to reject.

The quadratic has two roots, 4 and minus 2, and only 4 is in the domain. Check it in the original: log base 2 of 4 is 2 and log base 2 of 2 is 1, which add to 3. The rejected root would have required the logarithm of minus 2, which does not exist.

70. Check: reading a logarithm

Check

Check your understanding

What is log₂(1/8)?

  • A. −3 (correct)
  • B. 3
  • C. 1/3
  • D. −1/3

Answer: A

Why: Ask which power of 2 gives one eighth. Since 2 cubed is 8, 2 to the minus 3 is one eighth, so the logarithm is minus 3. Convert to exponential form and the answer reads off directly.

Why B tempts people
2³ = 8, not 1/8. A reciprocal needs a negative exponent.
Why C tempts people
2^(1/3) is the cube root of 2, about 1.26, not 1/8.
Why D tempts people
2^(−1/3) is about 0.79. The minus sign is right, but the exponent must be 3, not one third.

Turn the logarithm into a question about powers: which power of 2 gives one eighth? Because one eighth is a reciprocal, the exponent must be negative, and because 8 is 2 cubed, it must be minus 3.

The distractors each capture a specific slip: dropping the minus sign, or confusing a reciprocal with a root. Converting to exponential form avoids all of them.

71. Check: the laws of logarithms

Check

Check your understanding

For positive x and y, which expression equals ln(x³/y)?

  • A. 3 ln x / ln y
  • B. 3 ln x − ln y (correct)
  • C. (ln x)³ − ln y
  • D. 3 ln(x − y)

Answer: B

Why: The quotient property splits the logarithm of a quotient into a difference, and the power property brings the 3 in front of ln x. So the answer is 3 ln x minus ln y.

Why A tempts people
The logarithm of a quotient is a DIFFERENCE of logarithms, not a quotient of them.
Why C tempts people
The power property moves the exponent out in front as a multiplier; it does not cube the logarithm.
Why D tempts people
Nothing turns a quotient into a difference inside the logarithm; that invents a law.

Apply the laws one at a time. The quotient inside the logarithm becomes a difference of two logarithms, and the power on x becomes a coefficient of 3 in front of ln x.

Each wrong option invents a law that does not exist: a logarithm of a quotient as a quotient of logarithms, a power property that cubes the logarithm, and a quotient that turns into a difference inside the logarithm. If you are ever unsure, test with numbers such as x equal to e and y equal to 1.

72. Check: hyperbolic values

Check

Check your understanding

If cosh x = 5/3 and x > 0, what is sinh x?

  • A. −4/3
  • B. 2/3
  • C. 4/3 (correct)
  • D. 3/5

Answer: C

Why: Sinh squared is cosh squared minus 1, which is 25 ninths minus 1, or 16 ninths. Sinh has the same sign as x, and x is positive, so sinh x is 4 thirds. Indeed x is ln 3: cosh of ln 3 is 5 thirds and sinh of ln 3 is 4 thirds.

Why A tempts people
Minus four thirds has the right size, but sinh x is positive when x is positive.
Why B tempts people
Subtracting 1 from cosh itself, not from its square, misuses the identity.
Why D tempts people
Three fifths is sech x, the reciprocal of cosh x.

Use the hyperbolic Pythagorean identity: sinh squared equals cosh squared minus 1, which is 25 ninths minus 9 ninths, or 16 ninths. The square root is 4 thirds up to sign.

The sign comes from the behaviour of sinh: it is odd and increasing, so it is positive exactly when x is positive. Since x is positive here, sinh x is 4 thirds. The values come from x equal to ln 3, where cosh is 5 thirds and sinh is 4 thirds.

73. Explain where extraneous roots come from

Explain it to yourself

Discussion prompt

In Exercise 291, x = 1 − √5 appeared and then had to be thrown away. In two or three sentences, explain which step created it and why checking the original equation is not optional.

Write your two or three sentences before revealing. Being able to explain this is what stops you from reporting extraneous roots in the first place.

The key point is that some algebra steps are not reversible with respect to the domain. The original left side, ln x plus ln of x minus 2, needs both x and x minus 2 to be positive. The combined logarithm only needs their product to be positive, which also happens when both are negative. Every solution of the original equation solves the combined one, but not the other way round.

So solving the combined equation gives a list of candidates, not a list of answers. Checking each candidate in the original equation turns the list into answers.

74. Exit ticket

Exit ticket

\[ Q(t) = Q_0e^{-0.08664t}, \quad t \text{ in days} \]

Discussion prompt

Iodine-131 decays according to this model (Exercise 304). How many days, to the nearest day, until 95 percent of a sample has decayed?

OpenStax Calculus Volume 1, §1.5 Exponential and Logarithmic Functions §1.5, p. 97 — Exercise 304

This brings the lesson together in one real problem. The key translation is the first line: if 95 percent has decayed, then 5 percent remains, so the exponential factor must equal 0.05.

Taking the natural log of both sides brings the unknown down, and dividing by minus 0.08664 gives t equal to ln 20 over 0.08664, about 34.58 days. Notice that minus ln 0.05 is ln 20, since one over 0.05 is 20.

The rounding needs a moment's thought. At 34 days a little more than 5 percent is left, and at 35 days a little less, so the first whole day on which 95 percent has decayed is day 35. When a count comes from an inequality, check the whole numbers on either side.

75. Recap

Recap

ideathe key line
exponential functionf(x) = bˣ, b > 0, b ≠ 1; domain all reals, range (0, ∞)
the number ee = lim (1 + 1/m)ᵐ ≈ 2.718282; continuous growth A = Pe^(rt)
logarithmlog_b x = y ⟺ bʸ = x; domain (0, ∞)
laws of logarithmsproducts → sums, quotients → differences, powers → multipliers
change of baselog_a x = ln x / ln a
hyperbolic functionscosh x = (eˣ + e⁻ˣ)/2, sinh x = (eˣ − e⁻ˣ)/2, cosh² − sinh² = 1

\[ \log_b\left(b^{x}\right) = x, \qquad b^{\log_b x} = x \]

Next, Chapter 2 asks how fast these functions change, and e to the x will turn out to be its own rate of change.

Stewart, Calculus: Early Transcendentals 8e, §1.4 Exponential Functions §1.4, pp. 45-54 — exponential functions and the number e in Stewart

Stewart, Calculus: Early Transcendentals 8e, §1.5 Inverse Functions and Logarithms §1.5, pp. 55-67 — inverse functions, logarithms and the inverse hyperbolic functions in Stewart

The table is the lesson on one screen. Exponential functions multiply by a fixed factor per step, are always positive, and pass through 0 comma 1. The number e is the limit that continuous compounding produces, and it is the base whose graph leaves the y-axis with slope 1.

Logarithms are the inverses of exponentials, so every logarithm fact is an exponential fact read backwards: the definition, the graph reflected in the diagonal, and the three laws that turn products, quotients and powers into sums, differences and multipliers. Change of base lets you compute any logarithm from ln. When you solve equations, the domain decides which roots survive.

The hyperbolic functions are combinations of e to the x and e to the minus x that satisfy identities like the trigonometric ones, with a minus sign where the circle had a plus. In Chapter 2 you will begin asking how fast these functions change, and e to the x will turn out to be its own rate of change.

Sources

  1. OpenStax Calculus Volume 1, §1.5 Exponential and Logarithmic Functions — Strang & Herman, OpenStax / Rice University, CC BY-NC-SA 4.0, pp. 80-91
  2. Stewart, Calculus: Early Transcendentals 8e, §1.4 Exponential Functions — James Stewart, Cengage Learning, 2016, pp. 45-54
  3. Stewart, Calculus: Early Transcendentals 8e, §1.5 Inverse Functions and Logarithms — James Stewart, Cengage Learning, 2016, pp. 55-67

Want this taught 1-on-1? Alexander tutors Calculus I — $55/session, free consultation.

Book on Wyzant · Text (657) 465-8108