209
When you finish this section, you should be able to:
So far we have differentiated only functions \(y=f( x) \) where the dependent variable \(y\) is expressed explicitly in terms of the independent variable \(x\). There are functions that are not written in the form \(y=f( x) \), but are written in the implicit form \(F( x, y) =0\). For example, \(x\) and \(y\) are related implicitly in the equations \[ xy-4=0\qquad y^{2}+3x^{2}-1=0\qquad e^{x^{2}-y^{2}}-\cos~ \left( xy\right) =0 \]
The implicit form of a function is discussed in Section P.1, p. 4.
The implicit form \(xy-4=0\) can easily be written explicitly as the function \(y=\dfrac{4}{x}\). Also, \(y^{2}+3x^{2}-1=0\) can be written explicitly as the two functions \(y_{1}=\sqrt{1-3x^{2}}\) and \(y_{2}=-\sqrt{1-3x^{2}}\). In the equation \(e^{x^{2}-y^{2}}-\cos~ \left( xy\right) =0,\) it is impossible to express \(y\) as an explicit function of \(x\). In this case, and in many others, we use the technique of implicit differentiation to find the derivative.
The method used to differentiate an implicitly defined function is called implicit differentiation. It does not require rewriting the function explicitly, but it does require that the dependent variable \(y\) is a differentiable function of the independent variable \(x\). So throughout the section, we make the assumption that there is a differentiable function \(y=f( x) \) defined by the implicit equation. This assumption made, the method consists of differentiating both sides of the implicitly defined function with respect to \(x\) and then solving the resulting equation for \(\dfrac{dy}{dx}\).
210
Find \(\dfrac{dy}{dx}\) if \(xy-4=0\).
Solution (a) To differentiate implicitly, we assume \(y\) is a differentiable function of \(x\) and differentiate both sides with respect to \(x\). \[ \begin{eqnarray*} \begin{array}{rl@{\qquad}l} \dfrac{d}{dx}(xy-4) &=\dfrac{d}{dx}0 & {\color{#0066A7}{\hbox{Differentiate both sides with respect to}\ {x}.}}\\ \dfrac{d}{dx}(xy) -\dfrac{d}{dx}4 &= 0 & {\color{#0066A7}{\hbox{Use the Difference Rule.}}}\\ x\cdot \dfrac{d}{dx}y+\left(\dfrac{d}{dx}x\right) y-0 &= 0 & {\color{#0066A7}{\hbox{Use the Product Rule.}}}\\ x\dfrac{dy}{dx}+y & =0 & {\color{#0066A7}{\hbox{Simplify}.}} \\ \dfrac{dy}{dx} & = -\dfrac{y}{x} & {\color{#0066A7}{\hbox{Solve for}\, \dfrac{dy}{dx}.}} & (1) \end{array} \end{eqnarray*} \]
(b) Solve \(xy-4=0\) for \(y\), obtaining \(y=\dfrac{4}{x}=4x^{-1}\). Then \[ \begin{equation*} \dfrac{dy}{dx}=\dfrac{d}{dx}(4x^{-1}) =-4x^{-2}=-\dfrac{4}{x^{2}}\tag{2} \end{equation*} \]
(c) At first glance, the results in (1) and (2) appear to be different. However, since \(xy-4=0,\) or equivalently, \(y=\dfrac{4}{x},\) the result from (1) becomes \[ \begin{eqnarray*} && \dfrac{dy}{dx}\underset{\underset{\color{#0066A7}{(1)}}{\color{#0066A7}{\uparrow}}}{=}-\dfrac{y}{x}\underset{\underset{\color{#0066A7}{y=\tfrac{4}{x}}}{\color{#0066A7}{\uparrow}}}{=} -{\dfrac{\dfrac{4}{x}}{x}}=-\dfrac{4}{x^{2}} \end{eqnarray*} \] which is the same as (2).
In most instances, we will not know the explicit form of the function (as we did in Example 1) and so we will leave the derivative \(\dfrac{dy}{dx}\) expressed in terms of \(x\) and \(y\) [as in (1)].
Problem 17.
The Power Rule for Functions is \[ \dfrac{d}{dx}[f(x)] ^{n}=n[f(x)] ^{n-1}f^\prime (x) \] where \(n\) is an integer. If \(y=f( x)\), it takes the form \[\bbox[5px, border:1px solid black, #F9F7ED]{ \dfrac{d}{dx}y^{n}=ny^{n-1}\dfrac{dy}{dx}} \]
This is convenient notation to use with implicit differentiation when \(y^n\) appears. \[ \dfrac{d}{dx}y\underset{\underset{\color{#0066A7}{\hbox{\(n\)=1}}}{\color{#0066A7}{\uparrow}}}{=}1\cdot \dfrac{dy}{dx}=\dfrac{dy}{dx}\quad\quad \dfrac{d}{dx}y^{2}\underset{\underset{\color{#0066A7}{\hbox{\(n\)=2}}}{\color{#0066A7}{\uparrow}}}{=}2y\dfrac{dy}{dx}\quad\quad \dfrac{d}{dx}y^{3}\underset{\underset{\color{#0066A7}{\hbox{\(n=3\)}}}{\color{#0066A7}{\uparrow}}}{=}3y^{2}\dfrac{dy}{dx} \]
211
To differentiate an implicit function:
Find \(\dfrac{dy}{dx}\) if \(3x^{2}+4y^{2}=2x\).
Solution We assume that \(y\) is a differentiable function of \(x\) and differentiate both sides with respect to \(x\). \[ \begin{eqnarray*} \begin{array}{rl@{\qquad}l} \dfrac{d}{dx} ( 3x^{2}+4y^{2}) &= \dfrac{d}{dx}(2x) & {\color{#0066A7}{\hbox{Differentiate both sides with respect to}\; x.}}\\ \dfrac{d}{dx} (3x^{2}) +\dfrac{d}{dx}( 4y^{2}) &= 2 & {\color{#0066A7}{\hbox{Sum Rule}.}} \\ 3\dfrac{d}{dx}x^{2}+4\dfrac{d}{dx}y^{2} &=2 &{\color{#0066A7}{\hbox{Constant Multiple Rule}.}}\\ 6x+4\!\left( 2y\dfrac{dy}{dx}\right) &= 2 &{\color{#0066A7}{\dfrac{d}{dx}{y}^{{ 2}}{=2y} \dfrac{dy}{dx}}}\\ 6x+8y\dfrac{dy}{dx} &= {2} & {\color{#0066A7}{\hbox{Simplify.}}}\\ \dfrac{dy}{dx} &= \dfrac{2-6x}{8y}=\dfrac{1-3x}{4y} & {\color{#0066A7}{\hbox{Solve for}\; \dfrac{dy}{dx}.}} \end{array} \end{eqnarray*} \] provided \(y\ne 0\).
Notice in Example 2 that \(\dfrac{dy}{dx}\) is expressed in terms of \(x\) and \(y\).
Problem 15.
Find an equation of the tangent line to the graph of the ellipse \(3x^{2}+4y^{2}=2x\) at the point \(\left( \dfrac{1}{2},-\dfrac{1}{4}\right)\).
Solution First we find the slope of the tangent line. We use the result from Example 2, and evaluate \(\dfrac{dy}{dx}=\dfrac{1-3x}{4y}\) at \(\left( \dfrac{1}{2},-\dfrac{1}{4}\right) \). \[ \begin{eqnarray*} && \dfrac{dy}{dx}=\dfrac{1-3x}{4y} \underset{\underset{\color{#0066A7}{x=\tfrac{1}{2},y=-\dfrac{1}{4}}}{\color{#0066A7}{{\left\uparrow{\vphantom{\vrule width0pc height12.5pt depth0pt}}\right.}}}}{=} \dfrac{1-3\cdot \dfrac{1}{2}}{4\,{\cdot}\,\!\left(-\dfrac{1}{4}\right) }=\dfrac{1}{2}\\ \end{eqnarray*} \]
The slope of the tangent line to the graph of \(3x^{2}+4y^{2}=2x\) at the point \(\left( \dfrac{1}{2},-\dfrac{1}{4}\right) \) is \(\dfrac{1}{2}\). An equation of the tangent line is \[ \begin{eqnarray*} y+\dfrac{1}{4} &=&\dfrac{1}{2}\!\left( x-\dfrac{1}{2}\right) \\ y &=&\dfrac{1}{2}x-\dfrac{1}{2} \end{eqnarray*} \]
212
Figure 2 shows the graph of the ellipse \(3x^{2}+4y^{2}=2x\) and the graph of the tangent line \(y=\dfrac{1}{2}x-\dfrac{1}{2}\) at the point \(\left( \dfrac{1}{2},-\dfrac{1}{4}\right) \).
Problem 63.
Solution (a) We use implicit differentiation: \[ \begin{eqnarray*} \begin{array}{rl@{\qquad}l} \dfrac{d}{dx}( e^{y}\cos x) &= \dfrac{d}{dx}(x+1) \\ e^{y}\cdot \dfrac{d}{dx}( \cos x) +\left( \dfrac{d}{dx} e^{y}\right) \cdot \cos x &= 1 & {\color{#0066A7}{\hbox{Use the Product Rule.}}} \\ e^{y}( -\sin x) +e^{y}y^\prime \cdot \cos x &= 1 \\ ( e^{y}\cos x) y^\prime &= 1+e^{y}\sin x \\ y^\prime &= \dfrac{1+e^{y}\sin x}{e^{y}\cos x} \end{array} \end{eqnarray*} \]
(b) At the point \((0,0) \), the derivative \(y^\prime \) is \(\dfrac{1+e^{0}\sin 0}{e^{0}\cos 0}=\dfrac{1+1\cdot 0}{1\cdot 1}=1\), so the slope of the tangent line to the graph at \((0,0) \) is \(1\). An equation of the tangent line to the graph at the point \((0,0)\) is \(y=x.\)
The graph of \(e^{y}\cos x=x+1\) and the tangent line to the graph at \((0,0)\) are shown in Figure 3.
Implicit differentiation can be used to find higher-order derivatives.
Use implicit differentiation to find \(y^\prime \) and \(y^{\prime\prime} \) if \(y^{2}-x^{2}=5\). Express \(y^{\prime\prime} \) in terms of \(x\) and \(y\).
Solution First, we assume there is a differentiable function \(y=f( x) \) that satisfies \(y^{2}-x^{2}=5\). Now we find \(y^\prime \).
provided \(y≠ 0\).
213
Equations (3) and (4) both involve \(y^\prime \). Either one can be used to find \(y^{\prime\prime} .\) We use (3) because it avoids differentiating a quotient. \[ \begin{eqnarray*} \dfrac{d}{dx} ( 2yy^\prime -2x) &=&\dfrac{d}{dx}0 \nonumber\\ \dfrac{d}{dx} ( yy^\prime ) -\dfrac{d}{dx}x &=&0 \nonumber\\ y\cdot \dfrac{d}{dx}y^\prime +\left( \dfrac{d}{dx}y\right) y^\prime -1 &=&0 \nonumber\\ yy^{\prime\prime} +( y^\prime ) ^{2}-1 &=&0 \nonumber\\ y^{\prime\prime} &=&\dfrac{1-( y^\prime ) ^{2}}{y}\tag{5} \end{eqnarray*} \] provided \(y≠ 0\). To express \(y^{\prime\prime} \) in terms of \(x\) and \(y,\) we use (4) and substitute for \(y^\prime \) in (5).
Problem 59.
In Section 2.4, we showed that the Simple Power Rule \[ \dfrac{d}{dx}x^{n}=nx^{n-1} \] is true if the exponent \(n\) is any integer. We use implicit differentiation to extend this result for rational exponents.
If \(y=x^{p/q}\), where \(\dfrac{p}{q}\) is a rational number, then \[ \begin{equation*} \bbox[5px, border:1px solid black, #F9F7ED]{y^\prime =\dfrac{d}{dx}x^{p/q}=\dfrac{p}{q}\cdot x^{\left(p/q\right) -1}}\tag{6} \end{equation*} \] provided that \(x^{p/q}\) and \(x^{(p/q) -1}\) are defined.
We begin with the function \(y=x^{p/q},\) where \(p\) and \(q>0\) are integers. Now, we raise both sides of the equation to the power \(q\) to obtain \[ y^{q}=x^{p} \]
This is the implicit form of the function \(y=x^{p/q}.\) Assuming \(y\) is differentiable*, we can differentiate implicitly, obtaining \[ \begin{eqnarray*} \frac{d}{dx}y^q &=& \frac{d}{dx} x^p\\ qy^{q-1}y^\prime &=& px^{p-1} \end{eqnarray*} \]
*In Problem 113, you are asked to show that \(y=x^{p/q}\) is differentiable.
214
Now solve for \(y^\prime\). \[ \begin{eqnarray*} && y^\prime =\dfrac{px^{p-1}}{qy^{q-1}}\underset{\underset{\color{#0066A7}{y=x^{p/q}}}{\color{#0066A7}{\uparrow}}}{=}\dfrac{p}{q}\cdot \dfrac{x^{p-1}}{( x^{p/q}) ^{q-1}}=\dfrac{p}{q} \cdot \dfrac{x^{p-1}}{x^{p-(p/q) }}=\dfrac{p}{q}\cdot x^{p-1-[ p-(p/q)] }=\dfrac{p}{q}\cdot x^{(p/q)-1}\\ \end{eqnarray*} \]
Problem 31.
If \(u\) is a differentiable function of \(x\) and \(r\) is a rational number, then \[\bbox[5px, border:1px solid black, #F9F7ED] {\dfrac{d}{dx}[u(x)]^{r}=r[u(x)] ^{r-1}u^\prime (x)} \] provided \(u^{r}\) and \(u^{r-1}\) are defined.
Problem 39.
Inverse functions are discussed in Section P.4, pp. 32-37.
Suppose \(f\) is a function and \(g\) is its inverse function. Then \[ g( f( x) ) =x \] for all \(x\) in the domain of \(f.\)
215
If both \(f\) and \(g\) are differentiable, we can differentiate both sides with respect to \(x\) using the Chain Rule. Then \[ \begin{eqnarray*} && \dfrac{d}{dx}[g(f(x))] \underset{\underset{\underset{\underset{\color{#0066A7}{\hbox{on the left}}}{\color{#0066A7}{\hbox{Chain Rule}}}}{\color{#0066A7}{\hbox{Use the}}}}{\color{#0066A7}{\uparrow}}}{=} g^\prime (f(x)) \cdot f^\prime ( x) \underset{\underset{\underset{\color{#0066A7}{\hbox{on the right}}}{\color{#0066A7}{\hbox{\(\tfrac{d}{dx}x=1\)}}}}{\color{#0066A7}{\uparrow}}}{=} 1\\ \\ \end{eqnarray*} \]
Since the product of the derivatives is never \(0\), each function has a nonzero derivative on its domain.
Conversely, if a one-to-one function has a nonzero derivative, then its inverse function also has a nonzero derivative as stated in the following theorem.
Let \(y=f( x) \) and \(x=g( y) \) be inverse functions. Suppose \(f\) is differentiable on an open interval containing \(x_{0}\) and \(y_{0}=f( x_{0}) .\) If \(f^\prime ( x_{0}) ≠ 0\), then \(g\) is differentiable at \(y_{0}=f( x_{0}) \) and \[ \begin{equation*} \bbox[5px, border:1px solid black, #F9F7ED]{ \dfrac{d}{dy}g(y_{0}) =g^\prime (y_{0}) =\dfrac{1}{f^\prime (x_{0})}}\tag{7} \end{equation*} \] where the notation \(\dfrac{d}{dy}g( y_{0})=g^\prime(y_{0}) \) means the value of \(\dfrac{d}{dy}g( y) \) at \(y_{0},\) and the notation \(f^\prime ( x_{0}) \) means the value of \(f^\prime ( x) \) at \(x_{0}\).
In Leibniz notation, formula (7) has the simple form \[\bbox[5px, border:1px solid black, #F9F7ED]{ \dfrac{dx}{dy}=\dfrac{1}{\dfrac{dy}{dx}}} \]
We have two comments to make about this theorem, which is proved in Appendix B:
The function \(f( x) =x^{5}+x\) has an inverse function \(g\). Find \(g^\prime (2) \).
Solution Using (7) with \(y_{0}=2\), we get \[ g^\prime (2) =\dfrac{1}{f^\prime (x_{0})}\qquad \hbox{where}\ 2=f( x_{0}) \]
A solution of the equation \[ f( x_{0}) =x_{0}^{5}+x_{0}=2 \] is \(x_{0}=1\). Since \(f^\prime ( x) =5x^{4}+1\), then \( f^\prime (x_{0})= f^\prime (1) =6\) and \[ g^\prime (2) =\dfrac{1}{f^\prime (1) }=\dfrac{1}{6} \]
Observe in Example 8 that the derivative of the inverse function \(g\) at \(2\) was evaluated without actually knowing a formula for \(g\).
Problem 71.
216
Table 1 lists the inverse trigonometric functions and their domains.
Inverse trigonometric functions are discussed in Section P.7, pp. 58-61.
\(f\) | Restricted Domain | \(f^{-1}\) | Domain |
---|---|---|---|
\(f(x) =\sin x\) | \(\left[ -\dfrac{\pi }{2},\dfrac{\pi }{2}\right] \) | \(f^{-1}( x) =\sin ^{-1}x\) | \([-1,1] \) |
\(f(x) =\cos x\) | \([0,\pi] \) | \(f^{-1}( x) =\cos ^{-1}x\) | \([-1,1] \) |
\(f(x) =\tan x\) | \(\left( -\dfrac{\pi }{2},\dfrac{\pi }{2}\right) \) | \(f^{-1}( x) =\tan ^{-1}x\) | \((-\infty ,\infty) \) |
\(f(x) =\csc x\) | \(\left(-\pi, -\dfrac{\pi}{2}\right]\cup \left(0, \dfrac{\pi}{2}\right]\) | \( f^{-1}( x) =\csc ^{-1}x\) | \(|x| ≥ 1\) |
\(f(x) =\sec x\) | \(\left[0, \dfrac{\pi}{2}\right)\cup \left[\pi, \dfrac{3\pi}{2}\right)\) | \( f^{-1}(x) =\sec ^{-1}x\) | \(|x| ≥ 1\) |
\(f(x) =\cot x\) | \((0,\;\pi) \) | \(f^{-1}( x) =\cot ^{-1}x\) | \((-\infty ,\infty) \) |
To find the derivative of \(y=\sin ^{-1}x,\) \(-1≤ x≤ 1,\) \(-\dfrac{\pi }{2}≤ y≤ \dfrac{\pi }{2},\) we write \(\sin y=x\) and differentiate implicitly with respect to \(x\). \[ \begin{eqnarray*} \dfrac{d}{dx}\sin y &=&\dfrac{d}{dx}x \\ \cos y\cdot \dfrac{dy}{dx} &=&1 \\ \dfrac{dy}{dx} &=&\dfrac{1}{\cos y} \end{eqnarray*} \] provided \(\cos y≠ 0\). Since \(\cos y=0\) if \(y=-\dfrac{\pi }{2}\) or \(y=\dfrac{\pi }{2}\), we exclude these values. Then \(\dfrac{d}{dx}\sin^{-1} x=\dfrac{1}{\cos y}\), \(-\dfrac{\pi }{2}< y< \dfrac{\pi }{2}.\) Now, if \(-\dfrac{\pi }{2}< y< \dfrac{\pi }{2}\), then \(\cos y> 0\). Using a Pythagorean identity, we have \[ \begin{eqnarray*} \cos ^{2}y &=&1-\sin ^{2}y \\ \cos y &\underset{\underset{\color{#0066A7}{\hbox{\(\cos y>0\)}}}{\color{#0066A7}{\uparrow}}}{=}& \sqrt{1-\sin ^{2}y} \underset{\underset{\color{#0066A7}{\hbox{\(\sin y=x\)}}}{\color{#0066A7}{\uparrow}}}{=} {\sqrt{1-x^{2}}}\\ \end{eqnarray*} \]
The derivative of the inverse sine function \(y=\sin ^{-1}x\) is \[\bbox[5px, border:1px solid black, #F9F7ED]{ y^\prime = \dfrac{d}{dx}\sin ^{-1}x=\dfrac{1}{\sqrt{1-x^{2}}} \qquad -1< x< 1} \]
Find \(y^\prime \) if:
(a) \(y=\sin ^{-1}(4x^{2}) \) \( \quad \) (b) \(y=e^{\sin ^{-1}x}\)
Solution (a) If \(y=\sin ^{-1}u\) and \(u=4x^{2}\), then \(\dfrac{dy}{du}=\dfrac{1}{\sqrt{1-u^{2}}}\) and \(\dfrac{du}{dx}=8x.\) By the Chain Rule, \[ \begin{eqnarray*} && y^\prime =\dfrac{dy}{dx}=\dfrac{dy}{du}\cdot \dfrac{du}{dx}=\left( \dfrac{1}{ \sqrt{1-u^{2}}}\right) (8x) = \dfrac{8x}{\sqrt{1-16x^{4}}}\qquad {\color{#0066A7}{\hbox{\(u=4x^{2}\)}}} \end{eqnarray*} \]
217
(b) If \(y=e^{u}\) and \(u=\sin ^{-1}x\), then \(\dfrac{dy}{du}=e^{u}\) and \(\dfrac{du}{dx}=\dfrac{1}{\sqrt{1-x^{2}}}\). By the Chain Rule, \[ \begin{eqnarray*} && y^\prime =\dfrac{dy}{dx}=\dfrac{dy}{du}\cdot \dfrac{du}{dx}=e^{u}\cdot \dfrac{ 1}{\sqrt{1-x^{2}}} \underset{\underset{\color{#0066A7}{\hbox{\(u=\sin ^{-1}x\)}}}{\color{#0066A7}{\uparrow}}}{=} \dfrac{ e^{\sin ^{-1}x}}{\sqrt{1-x^{2}}} \\ \end{eqnarray*} \]
Problem 51.
To derive a formula for the derivative of \(y\,{=}\,\tan^{-1}x,\) \(-\infty\,{< }\,x\,{< }\,\infty ,\) \(-\dfrac{\pi }{2}\,{< }\,y\,{<}\,\dfrac{\pi }{2},\) we write \(\tan y=x\) and differentiate with respect to \(x\). Then \[ \begin{eqnarray*} \dfrac{d}{dx}\tan y &=&\dfrac{d}{dx}x \\ \sec ^{2}y\dfrac{dy}{dx} &=&1 \\ \dfrac{dy}{dx} &=&\dfrac{1}{\sec ^{2}y} \end{eqnarray*} \]
Since \(-\dfrac{\pi }{2}< y< \dfrac{\pi }{2},\) then \(\sec y≠ 0.\)
Now we use the Pythagorean identity \(\sec ^{2}y=1+\tan ^{2}y,\) and substitute \(x=\tan y\). Then \[ y^\prime =\dfrac{d}{dx}\tan ^{-1}x=\dfrac{1}{1+x^{2}} \]
The derivative of the inverse tangent function \(y=\tan ^{-1}x\) is \[\bbox[5px, border:1px solid black, #F9F7ED]{ y^\prime =\dfrac{d}{dx}\tan ^{-1}x=\dfrac{1}{1+x^{2}}} \]
The domain of \(y^\prime \) is all real numbers.
Find \(y^\prime \) if:
(a) \(y=\tan ^{-1}( 4x)\) \( \quad \) (b) \(y=\sin ( \tan ^{-1}x)\)
Solution (a) Let \(y=\tan ^{-1}u\) and \(u=4x.\) Then \(\dfrac{dy}{du}=\dfrac{1}{1+u^{2}}\) and \(\dfrac{du}{dx}=4.\) By the Chain Rule, \[ y^\prime =\dfrac{dy}{dx}=\dfrac{dy}{du}\cdot \dfrac{du}{dx}=\dfrac{1}{1+u^{2}} \cdot 4=\dfrac{4}{1+16x^{2}} \]
(b) Let \(y=\sin u\) and \(u=\tan ^{-1}x\). Then \(\dfrac{dy}{du}=\cos u\) and \(\dfrac{du}{dx}=\dfrac{1}{1+x^{2}}\). By the Chain Rule, \[ y^\prime =\dfrac{dy}{dx}=\dfrac{dy}{du}\cdot \dfrac{du}{dx}=\cos u\cdot \dfrac{1}{1+x^{2}}=\dfrac{\cos ( \tan ^{-1}x) }{1+x^{2}} \]
Problem 55.
Finally to find the derivative of \(y=\sec ^{-1}x,\) \(\vert x\vert ≥ 1,\) \(0≤ y< \dfrac{\pi }{2}\) or \(\pi ≤ y< \dfrac{3\pi }{2},\) we write \(x=\sec y\) and use implicit differentiation. \[ \begin{eqnarray*} \dfrac{d}{dx}x &=&\dfrac{d}{dx}\sec y \\ 1 &=&\sec y\tan y\dfrac{dy}{dx} \end{eqnarray*} \]
218
We can solve for \(\dfrac{dy}{dx}\) provided \(\sec y≠ 0\) and \(\tan y≠ 0,\) or equivalently, \(y≠ 0\) and \(y≠ \pi.\) That is, \[ \dfrac{dy}{dx}=\dfrac{1}{\sec y\tan y},\qquad \hbox{provided}\; 0< y< \dfrac{\pi }{2} \ \hbox{or}\ \pi < y< \dfrac{3\pi }{2}. \]
With these restrictions on \(y,\) \(\tan y>0\).* Now we use the Pythagorean identity \(1+\tan ^{2}y=\sec ^{2}y\). Then \(\tan y = \sqrt{\sec^2 y-1}\) and \[ \begin{eqnarray*} && \dfrac{1}{\sec y\tan y}=\dfrac{1}{\sec y\sqrt{\sec ^{2}y-1}} \underset{\underset{\color{#0066A7}{\hbox{\(\sec y=x\)}}}{\color{#0066A7}{\uparrow}}}{=} \dfrac{1}{x\sqrt{x^{2}-1}}\\ \end{eqnarray*} \]
The derivative of the inverse secant function \(y=\sec ^{-1}x\) is \[\bbox[5px, border:1px solid black, #F9F7ED]{ y^\prime =\dfrac{d}{dx}\sec ^{-1}x=\dfrac{1}{x\sqrt{x^{2}-1}}\qquad \vert x\vert > 1 } \]
Notice that \(y=\sec ^{-1}x\) is not differentiable when \(x=\pm 1\). In fact, as Figure 4 shows, at the points \(\left( -1,\pi \right) \) and \((1,0)\), the graph of \(y=\sec ^{-1}x\) has vertical tangent lines.
The formulas for the derivatives of the three remaining inverse trigonometric functions can be obtained using the following identities:
*The restricted domain of \(y=\sec x\) was chosen as \(\left\{x|0≤ x< \dfrac{\pi }{2}\hbox{ or } \pi ≤ x< \dfrac{3\pi }{2}\right\}\) so that \(\tan x > 0\).