7.3: Inverse Trigonometric Functions
( \newcommand{\kernel}{\mathrm{null}\,}\)
Inverse functions were studied in Section 2.4. We now take up the topic again and apply it to trigonometric functions. A binary relation on the real numbers is any set of ordered pairs of real numbers. Thus a real function f of one variable is a binary relation such that for each x, either there is exactly one y with (x,y) in f or there is no y with (x,y) in f. (Other important relations are x<y,x≤y,x≠y,x=y.)
Let S be a binary relation on the real numbers. The inverse relation of S is the set T of all ordered pairs (y,x) such that (x,y) is in S. If S and T are both functions they are called inverse functions of each other.
The inverse of a function f may or may not be a function. For example, the inverse of y=x2 is the relation x=±√y, which is not a function (Figure 7.3.1). But the inverse of y=x2,x≥0, is the function x=√y (Figure 7.3.2).
Figure 7.3 .1
Geometrically, the graph of the inverse relation of y=f(x) can be obtained by flipping the graph of y=f(x) about the diagonal line y=x (the dotted line in Figures 7.3.1 and 7.3.2). This flipping interchanges the x - and y-axes. This is because f(x)=y means (x,y) is in f, and g(y)=x means (y,x) is in g. It follows that:
If f and g are inverse functions then the range of f is the domain of g and vice versa.
Which functions have inverse functions? We can answer this question with a definition and a simple theorem.
A real function f with domain X is said to be one-to-one if f never takes the same value twice, that is, for all x1≠x2 in X we have f(x1)≠f(x2).
f has an inverse function if and only if f is one-to-one.
PROOF
The following statements are equivalent.
- f is a one-to-one function.
- For every y, either there is exactly one x with f(x)=y or there is no x with f(x)=y.
- The equation y=f(x) determines x as a function of y.
- f has an inverse function.
Every function which is increasing on its domain I has an inverse function. So does every function decreasing on its domain I.
PROOF Let f be increasing on I. For any two points x1≠x2 in I, the value of f at the smaller of x1,x2 is less than the value of f at the greater, so f(x1)≠ f(x2).
For example, the function y=x2 is not one-to-one because (−1)2=12, whence it has no inverse function. The function y=x2,x≥0, is increasing on its domain [0,∞) and thus has an inverse.
Now let us examine the trigonometric functions. The function y=sinx is not one-to-one. For example, sin0=0,sinπ=0,sin2π=0, etc. We can see in Figure 7.3.3 that the inverse relation of y=sinx is not a function.
y=sinx
not a function
Figure 7.3.3
However, the function y=sinx is increasing on the interval [−π/2,π/2], because its derivative cosx is ≥0. So the sine function restricted to the interval [−π/2,π/2],
y=sinx,−π/2≤x≤π/2
has an inverse function shown in Figure 7.3.4. This inverse is called the arcsine
y=sinx,−π2≤x≤π2
Figure 7.3.4
function. It is written x=arcsiny. Verbally, arcsin y is the angle x between −π/2 and π/2 whose sine is y.
The other trigonometric functions also are not one-to-one and thus do not have inverse functions. However, in each case we obtain a one-to-one function by restricting the domain to a suitable interval, either [−π/2,π/2] or [0,π]. The resulting inverse functions are called the arccosine, arctangent, etc.
The inverse trigonometric functions are defined as follows.
x=arcsiny is the inverse of y=sinx,−π/2≤x≤π/2
x=arccosy is the inverse of y=cosx,0≤x≤π
x=arctany is the inverse of y=tanx,−π/2≤x≤π/2
x=arccoty is the inverse of y=cotx,0≤x≤π
x=arcsecy is the inverse of y=secx,0≤x≤π
x=arccscy is the inverse of y=cscx,−π/2≤x≤π/2
The graphs of these functions are shown in Figure 7.3.5. The domains of the inverse trigonometric functions can be read off from the graphs, and are shown in the table below.
Table 7.3.1
Function | Domain |
---|---|
arcsiny | −1≤y≤1 |
arccosy | −1≤y≤1 |
arctany | whole real line |
arccoty | whole real line |
arcsecy | y≤−1,y≥1 |
arccscy | y≤−1,y≥1 |
We can prove the inverse trigonometric functions have these domains (i.e., the figures are correct) using the Intermediate Value Theorem. As an illustration we prove that arcsin y has domain [−1,1].
arcsiny is undefined outside [−1,1] because −1≤sinx≤1 for all x. Suppose y0 is in [−1,1]. Then
sin(−π/2)=−1≤y0≤1=sin(π/2)
sinx is continuous, so by the Intermediate Value Theorem there exists x0 between −π/2 and π/2 such that sinx0=y0. Thus
arcsiny0=x0
and y0 is in the domain of arcsin y.
Find arccos(√2/2). From Table 7.1.1, cos(π/4)=√2/2. Since 0≤π/4≤π,
Solution
arccos(√2/2)=π/4
Find arcsin(−1). From Table 7.1.1, sin(3π/2)=−1. But 3π/2 is not in the interval [−π/2,π/2]. Using sin(θ+2nπ)=sinθ, we have
Solution
sin(−π/2)=sin(3π/2)=−1arcsin(−1)=−π/2
Find arctan(−√3). We must find a θ in the interval [−π/2,π/2] such that tanθ=−√3. From Table 7.1.1, sin(π/3)=√3/2,cos(π/3)=1/2. Then sin(−π/3)=−√3/2,cos(−π/3)=1/2. So
Solution
tan(−π/3)=−√3/21/2=−√3arctan(−√3)=−π/3
. Find cos(arctany). Let θ=arctany. Thus tanθ=y. Using
Solution
sin2θ+cos2θ=1sinθcosθ=y
we solve for cosθ.
sinθ=ycosθ,(ycosθ)2+cos2θ=1cos2θ(y2+1)=1,cos2θ≐1y2+1cosθ=±1√y2+1
By definition of arctan y, we know that −π/2≤θ≤π/2. In this interval, cosθ≥0. Therefore
cosθ=1√y2+1
Show that arcsiny+arccosy=π/2 (Figure 7.3.6). Let θ=arcsiny. We have y=sinθ=cos(π/2−θ). Also, when −π/2≤θ≤π/2, we have
Solution
π/2≥−θ≥−π/2,π≥π/2−θ≥0
Thus
π/2−θ=arccosy
arcsiny+arccosy=θ+(π/2−θ)=π/2
Figure 7.3.6
We shall next study the derivatives of the inverse trigonometric functions. Here is a general theorem which tells us when the derivative of the inverse function exists and gives a rule for computing its value.
INVERSE FUNCTION THEOREM
Suppose a real function f is differentiable on an open interval I and f has an inverse function g. Let x be a point in I where f′(x)≠0 and let y=f(x). Then
- g′(y) exists,
- g′(y)=1f′(x).
We omit the proof that g′(y) exists. Intuitively, the curve y=f(x) has a nonhorizontal tangent line, so the curve x=g(y) should have a nonvertical tangent line and thus g′(y) should exist.
The Inverse Function Rule from Chapter 2 says that (ii) is true if we assume (i). The proof of (ii) from (i) is an application of the Chain Rule:
g(f(x))=x,g′(f(x))f′(x)=1,g′(y)f′(x)=1,g′(y)=1f′(x)
The Inverse Function Theorem shows that all the inverse trigonometric functions have derivatives. We now evaluate these derivatives.
(i) d(arcsinx)=dx√1−x2 (where −1<x<1.
d(arccosx)=−dx√1−x2( where −1<x<1)
(ii) d(arctanx)=dx1+x2.
d(arccotx)=−dx1+x2
(iii) d(arcsecx)=dx|x|√x2−1 (where |x|>1 ).
d(arccscx)=−dx|x|√x2−1( where |x|>1)
PROOF
We prove the first part of (i) and (iii). Since the derivatives exist we may use implicit differentiation.
(i) Let y=arcsinx. Then
x=siny,−π/2≤y≤π/2dx=cosydy
From sin2y+cos2y=1 we get
cosy=±√1−sin2y=±√1−x2
Since −π/2≤y≤π/2,cosy≥0. Then
cosy=√1−x2
Substituting,
dx=√1−x2dy,dy=dx√1−x2
(iii) Let y=arcsecx.
Then
x=secy;0≤y≤π,dx=secytanydy.
From tan2y+1=sec2y we get tany=±√sec2y−1=±√x2−1.
Since 0≤y≤π,tany and secy=1cosy have the same sign.
Therefore
secytany′≥0
and
dx=|secy||tany|dy=|x|√x2−1dydy=dx|x|√x2−1
When we turn these formulas for derivatives around we get some surprising new integration formulas.
- ∫1√1−x2dx=arcsinx+C=−arccosx+C.(Provided that |x|<1 ).
- ∫dx1+x2=arctanx+C=−arccotx+C.
- ∫dx|x|√x2−1=arcsecx+C=−arccscx+C. (Provided that |x|>1 ).
From part (i), arcsin x and -arccos x must differ only by a constant. We already knew this from Example 5,
arcsinx=−arccosx+π/2
Before now we were not able to find the area of the regions under the curves
y=1√1−x2,y=11+x2,y=1x√x2−1
It is a remarkable and quite unexpected fact that these areas are given by inverse trigonometric functions.
(a) Find the area of the region under the curve
y=11+x2
for −1≤x≤1.
(b) Find the area of the region under the same curve for −x<x<x. The regions are shown in Figure 7.3.7.
Solution
(a) A=∫1−111+x2dx=arctanx]1−1=π4−(−π4)=π2.
Figure 7.3.7
(b) A=∫∞−∞11+x2dx=∫0−∞11+x2dx+∫∞011+x2dx
=lima→−∞∫0a11+x2dx+limb→∞∫b011+x2dx
=lima→−∞(arctan0−arctana)+limb→∞(arctanb−arctan0)=−lima→−∞arctana+limb→∞arctanb
From the graph of arctanx we see that the first limit is −π/2 and the second limit is π/2, so
A=−(−π2)+π2=π
Thus the region under y=1/(1+x2) has exactly the same area as the unit circle, and half of this area is between x=−1 and x=1.
Find ∫−√2−21x√x2−1dx.
Solution
∫−√2−21x√x2−1dx=∫−√2−2−1|x|√x2−1dx=−arcsecx]−√2−2=−(arcsec(−√2)−arcsec(−2))=−(3π4−2π3)=−π12
PROBLEMS FOR SECTION 7.3
In Problems 1-9, evaluate the given expression.
1 | arcsin(√3/2) | 2 | arcsin(−1/2) |
3 | arctan(−1) | 4 | sec(arctan(−1)) |
5 | arcsec2 | 6 | arcsin(cosπ) |
7 | sin(arccosx) | 8 | cot(arcsecx) |
9 | arcsin(cosx),0≤x≤π |
10 Prove the identity arctan(−x)=−arctanx.
11 Prove arctan (1/x)=arccotx, for 0<x.
12 Prove arccos(−x)=π−arccosx
13 Prove arctanx+arccotx=π/2.
Find the derivatives in Problems 14-25.
14y=arcsin(x/2) 15
16y=(arcsinx)2
18y=arctan√x
20y=xarcsecx
22y′=arccosx+(x/√1−x2)
24u=arcsect+√t2−1
\boldsymbol{\begin{aligned} & y=\operatorname{arcsec}(5 x-2) \\ & y=\arcsin \left(x^{2}\right) \tag{17}\\ & s=t \arcsin t \\ & y=\arcsin x+\sqrt{1-x^{2}} \\ & y=x \arcsin x+\sqrt{1-x^{2}} \\ & y=\arctan (1 / \sqrt{x}) \tag{25}\end{aligned}}
26 Evaluate limx→∞arccscx.
27 Evaluate limx→−∞arctanx.
28 Evaluate limx→0arcsinxx.
29 Evaluate limx→∞arccotxarccscx.
In Problems 30-47 evaluate the integrals.
30 ∫dx1+4x2
31 ∫dx9+x2
32∫dx√4−x2
33∫dx√x−x2
34∫cosx1+sin2xdx
35
∫dxx√4x2−1,x>1
36∫xdxx4+1
∫xdx√1−x4
38∫dx(1+x)√x
∫dxx√x−1
40
∫arctanx1+x2dx
∫arcsinx√1−x2dx
42∫√3−√3dx1+x2
∫211x√x2−1dx
44
∫−1−√21x√x2−1dx
∫1/201√1−x2dx
46
∫∞0dx25x2+1
47
∫∞−∞dxa2+x2
48 Find the area of the region bounded by the x-axis and the curve y=1/√1−x2, −1<x<1.
49 Find the area of the region under the curve y=1/(x√x2−1),1≤x<∞.
50 Find the area of the region bounded below by the line y=12 and above by the curve y=1/(x2+1).