1.6 Exponential and Logarithmic Functions

Gordon Moore (1929–). Moore, who later became chairman of Intel Corporation, predicted that in the decades following 1965, the number of transistors per integrated circuit would grow “exponentially.” This prediction has held up for nearly five decades and may well continue for several more years. Moore has said, “Moore’s Law is a term that got applied to a curve I plotted in the mid-sixties showing the increase in complexity of integrated circuits versus time. It’s been expanded to include a lot more than that, and I’m happy to take credit for all of it.”

An exponential function is a function of the form \(f(x) = b^{x}\), where \(b > 0 \) and \(b \neq 1\). The number \(b\) is called the base. Some examples are \(2^{x}\), \((1.4)^{x}\), and \(10^{x}\). The case \(b = 1\) is excluded because \(f(x) = 1^{x}\) is a constant function. Calculators give good decimal approximations to values of exponential functions:

\[ 2^4=16,\quad 3^{-3}=0.125,\quad (1.4)^{0.8}\approx1.309,\quad 10^{4.6}\approx39,810.717 \]

Three properties of exponential functions should be singled out from the start (see Figure 1.76 for the case \(b = 2\)):

Figure 1.76: ---

If \(b > 1\), the exponential function \(f(x) = b^{x}\) is not merely increasing but is, in a certain sense, rapidly increasing. Although the term “rapid increase” is perhaps subjective, the following precise statement is true: \(f(x) = b^{x}\) increases more rapidly than the power function \(x^{n}\) for all \(n\) (we will prove this in Section 4.5). For example, Figure 1.77 shows that \(f(x) = 3^{x}\) eventually overtakes and increases faster than the power functions \(x^{3}\), \(x^{4}\), and \(x^{5}\). Table 1.8 compares \(3^{x}\) and \(x^{5}\).

Figure 1.77: Comparison of 3x and power functions.
\(x\) \(x^{5}\) \(3^{x}\)
\(1\) \(1\) \(3\)
\(5\) \(3125\) \(243\)
\(10\) \(100,000\) \(59,049\)
\(15\) \(759,375\) \(14,348,907\)
\(25\) \(9,765,625\) \(847,288,609,443\)
Table 1.8: ---

We now review the laws of exponents. The most important law is

\[ \boxed{b^xb^y=b^{x+y}} \]

In other words, under multiplication, the exponents add, provided that the bases are the same. This law does not apply to a product such as \(3^{2} \cdot 5^{4}\).

44

Laws of Exponents (\(b > 0\;\))

Be sure you are familiar with the laws of exponents. They are used throughout this text.

\[ \begin{array}{lll} \hline &\quad\text{Rule}&\quad\text{Example}\\ \hline \text{Exponent zero}&\quad b^0 = 1&\\[5pt] \text{Products}&\quad b^xb^y=b^{x+y}& \quad2^5\cdot2^3=2^{5+3}=2^8\\[5pt] \text{Quotients}&\quad\frac{b^x}{b^y}=b^{x-y}& \quad\frac{4^7}{4^2}=4^{7-2}=4^5\\[5pt] \text{Negative exponents}&\quad b^{-x}=\frac{1}{b^x}& \quad3^{-4}=\frac{1}{3^4}=\frac{1}{81}\\[5pt] \text{Power to a power}&\quad(b^x)^y=b^{xy}& \quad(3^2)^4=3^{2(4)}=3^8\\[5pt] \text{Roots}&\quad b^{\frac{1}{n}}\sqrt[n]{b}& \quad5^{\frac{1}{2}}=\sqrt{5} \end{array} \]

EXAMPLE 1

Rewrite as a whole number or fraction:

(a) \(16^{-\frac{1}{2}}\)

(b) \(27^{\frac{2}{3}}\)

(c) \(4^{16} \cdot 4^{-18}\)

(d) \(\frac{9^3}{3^7}\)

Solution

(a) \(16^{-\frac{1}{2}}=\frac{1}{16^{1/2}}=\frac{1}{\sqrt{16}}=\frac{1}{4}\).

(b) \(27^{\frac{2}{3}} = (27^{\frac{1}{3}})^{2} = 3^{2} = 9\).

(c) \(4^{16} \cdot 4^{-18}=4^{-2}=\frac{1}{4^2}=\frac{1}{16}\).

(d) \(\frac{9^3}{3^7} = \frac{(3^2)^3}{3^7}=\frac{3^6}{3^7}=3^{-1}=\frac{1}{3}\).

In the next example, we use the fact that \(f(x) = b^{x}\) is one-to-one. In other words, if \(b^{x} = b^{y}\), then \(x = y\).

EXAMPLE 2

Solve for the unknown:

(a) \(2^{3x+1} = 2^{5}\)

(b) \(b^{3} = 5^{6}\)

(c) \(7^{t+1} = \left(\frac{1}{7}\right)^{2t}\)

Solution

(a) If \(2^{3x+1} = 2^{5}\), then \(3x + 1 = 5\) and thus \(x=\frac{4}{3}\).

(b) Raise both sides of \(b^{3} = 5^{6}\) to the \(\frac{1}{3}\) power. By the “power to a power” rule, \[ b=(b^3)^{\frac{1}{3}}=(5^6)^{\frac{1}{3}}=5^{\frac{6}{3}}=5^2=25 \]

(c) Since \(\frac{1}{7}=7^{-1}\), the right-hand side of the equation is \((\frac{1}{7})^{2t}=(7^{-1})^{2t}=7^{-2t}\). The equation becomes \(7^{t+1} = 7^{-2t}\). Therefore, \(t + 1 = -2t\), or \(t=-\frac{1}{3}\).

45

1.6.1 The Number e

Although written references to the number \(\pi\) go back more than 4000 years, mathematicians first became aware of the special role played by \(e\) in the seventeenth century. The notation \(e\) was introduced by Leonhard Euler, who discovered many fundamental properties of this important number.

In Chapter 3, we will use calculus to study exponential functions. One of the surprising insights of calculus is that the most convenient or “natural” base for an exponential function is not \(b = 10\) or \(b = 2\), as one might think at first, but rather a certain irrational number, denoted by \(e\), whose value is approximately \(e \approx 2.718\). A calculator is used to evaluate specific values of \(f(x) = e^{x}\). For example,

\[ e^3\approx20.0855,\quad e^{-\frac{1}{4}}\approx0.7788 \]

In calculus, when we speak of the exponential function, it is understood that the base is \(e\). Another common notation for the exponential function \(e^{x}\) is \(\exp(x)\).

How is \(e\) defined? There are many different definitions, but they all rely on the calculus concept of a limit. We shall discuss one way of defining \(e\) in Section 3.2. Another definition is described in Example 4 of Section 1.7. For now, we mention the following two graphical descriptions.

Figure 1.78: ---

From these descriptions it is not clear why \(e\) is important. As we will learn, however, the exponential function \(e^{x}\) plays a fundamental role because it behaves in a particularly simple way with respect to the basic operations of calculus: differentiation and integration.

1.6.2 Logarithms

Renato Solidum, director of the Philippine Institute of Volcanology and Seismology, checks the intensity of the October 8, 2004, Manila earthquake, which registered 6.2 on the Richter scale. The Richter scale is based on the logarithm (to base 10) of the amplitude of seismic waves. Each whole-number increase in Richter magnitude corresponds to a 10-fold increase in amplitude and approximately 31.6 times more energy.

Logarithmic functions are inverses of exponential functions. More precisely, if \(b > 0\) and \(b \neq 1\), then the logarithm to the base b, denoted \(\log_{b} x\), is the inverse of \(f(x) = b^{x}\). By definition, \(y = \log_{b} x\) if \(b^{y} = x\), so we have

\[ \boxed{b^{\log_bx}=x\quad\text{and}\quad\log_b(b^x)=x} \]

In other words, \(\log_{b} x\) is the number to which \(b\) must be raised in order to get \(x\). For example,

\[ \begin{array}{rcl} \log_28=3&\text{because}&2^3=8\\ \log_{10}(1)=0&\text{because}&10^0=1\\ \log_3\left(\frac{1}{9}\right)=-2&\text{because}& 3^{-2}=\frac{1}{3^2}=\frac{1}{9} \end{array} \]

46

The logarithm to the base \(e\), denoted \(\ln x\), plays a special role and is called the natural logarithm. We use a calculator to evaluate logarithms numerically. For example,

\[ \ln17\approx2.83321\quad\text{because}\quad e^{2.83321}\approx17 \]

Recall that the domain of \(b^{x}\) is \(\mathbb{R}\) and its range is the set of positive real numbers \(\{x : x > 0\}\). Since the domain and range are reversed in the inverse function,

In this text, the natural logarithm is denoted \(\ln x\). Other common notations are \(\log x\) and \(\text{Log}\,x\).

If \(b \gt 1\), then \(\log_{b} x\) is positive for \(x \gt 1\) and negative for \(0 \lt x \lt 1\). Figure 1.79 illustrates these facts for the base \(b = e\). Keep in mind that the logarithm of a negative number does not exist. For example,\(\log_{10}(-2)\) does not exist because \(10^{y} = -2\) has no solution.

Figure 1.79: \(y = ln x\) is the inverse of \(y = e^{x}\).

For each law of exponents, there is a corresponding law for logarithms. The rule \(b^{x+y} = b^{x}b^{y}\) corresponds to the rule

\[ \boxed{\log_b(xy)=\log_bx+\log_by} \]

In words: The log of a product is the sum of the logs. To verify this rule, observe that

\begin{align*} b^{\log_b(xy)}=xy&=b^{\log_bx}\cdot b^{\log_by}\\ &=b^{\log_bx+\log_by} \end{align*}

The exponents \(\log_{b}(xy)\) and \(\log_{b} x + \log_{b} y\) are equal as claimed because \(f(x) = b^{x}\) is one-to-one. The remaining logarithm laws are collected in the following table.

Laws of Logarithms

\[ \begin{array}{lll} \hline &\text{Law}&\text{Example}\\ \hline \text{Log of }1\quad&\log_b(1)=0\quad&\\[5pt] \text{Log of }b\quad&\log_b(b)=1\quad&\\[5pt] \text{Products}\quad&\log_b(xy)=\log_bx+\log_by \quad&\log_5(2\cdot3)=\log_52+\log_53\\[5pt] \text{Quotients}\quad& \log_b\left(\frac{x}{y}\right)=\log_bx-\log_by\quad& \log_2\left(\frac{3}{7}\right)=\log_23-\log_27\\[5pt] \text{Reciprocals}\quad& \log_b\left(\frac{1}{x}\right)=-\log_bx\quad& \log_2\left(\frac{1}{7}\right)=-\log_27\\[5pt] \text{Powers (any \(n\))}\quad& \log_b(x^n)=n\log_bx\quad& \log_{10}(8^2)=2\cdot\log_{10}8 \end{array} \]

We note also that all logarithmic functions are proportional. More precisely, the following change-of-base formula holds (see Exercise 49):

\[ \boxed{\log_bx=\dfrac{\log_ax}{\log_ab},\quad\log_bx=\dfrac{\ln x}{\ln b}}\tag{1} \]

47

EXAMPLE 3 Using the Logarithm Laws

Evaluate:

(a) \(\log_69+\log_64\)

(b) \(\ln\left(\frac{1}{\sqrt{e}}\right)\)

(c) \(10\log_b(b^3)-4\log_4(\sqrt{b})\)

Solution

(a) \(\log_69+\log_64=\log_6(9\cdot4)=\log_6(35)=\log_6(6^2)=2\)

(b) \(\ln\left(\frac{1}{\sqrt{e}}\right)=\ln(e^{-\frac{1}{2}})=-\frac{1}{2}\ln(e)=-\frac{1}{2}\)

(c) \(10\log_b(b^3)-4\log_4(\sqrt{b})=10(3)-4\log_b(b^{\frac{1}{2}})=30 - 4\left(\frac{1}{2}\right)=28\)

EXAMPLE 4 Solving an Exponential Equation

The bacteria population in a bottle at time \(t\) (in hours) has size \(P(t)=1000e^{0.35t}\). After how many hours will there be \(5000\) bacteria?

Solution We must solve \(P(t)=1000e^{0.35t}=5000\) for \(t\) (Figure 1.80):

\begin{align*} e^{0.35t}&=\frac{5000}{1000}=5\\ \ln(e^{0.35t})&=\ln5&&\text{(take logarithms of both sides)}\\ 0.35t&=\ln5\approx1.609&&\text{[because \(\ln(e^a)=a\)]}\\ t&\approx\frac{1.609}{0.35}\approx4.6\text{ hours} \end{align*}

Figure 1.80: Bacteria population as a function of time.

1.6.3 Section 1.6 Summary

(i) \(\log_{b}(xy) = \log_{b} x + \log_{b} y\)

(ii) \(\log_b\left(\frac{x}{y}\right)=\log_bx-\log_by\)

(iii) \(\log_{b}(x^{n}) = n \log_{b}x\)

(iv) \(\log_{b} 1 = 0\) and \(\log_{b} b = 1\)