From the graph of y=sinhx (see Figure 1.8), it is clear from the horizontal line test that the function is bijective. Hence, its inverse function exists, and we define it as follows:x=sinh−1y⟺y=sinhx
Worked example
Example 10
Prove that sinh−1x=ln(x+x2+1).
Given: y=sinhx is a bijection of R onto R, so sinh−1x exists for every real
Answer: sinh−1x=ln(x+x2+1)
(ii) Inverse Hyperbolic cosine Function
From the graph of y=coshx (see Figure 1.9), it is clear from the horizontal line test that the function is not one-to-one on the domain (−∞,∞). However, if we restrict the domain to [0,∞), the range set is not disturbed. It follows that y=coshx becomes one-to-one for the restricted domain [0,∞) and thus bijective. Therefore, on the domain [0,∞), the inverse hyperbolic cosine function exists and we define it as follows:x=cosh−1y⟺y=coshx
Worked example
Example 11
Prove that cosh−1x=ln(x+x2−1) for all x∈[1,∞).
Given: coshx restricted to [0,∞) is a bijection onto [1,∞), so cosh−1x
Answer: cosh−1x=ln(x+x2−1)
x
.
To prove:
sinh−1x=ln(x+x2+1)
.
Method: put
y=sinh−1x
, so
x=sinhy
. Write
sinhy
with
ey
, which gives a quadratic equation in
ey
; solve it, reject the root that cannot be
ey
, then take
ln
.
Change the inverse into the function
Let y=sinh−1x, where x,y∈]−∞,∞[. Thenx=sinhyWhy: by definition, y=sinh−1x if and only if x=sinhy. We know how to handle sinh, not sinh−1.
Write it with exponentialsx=2ey−e−y2x=ey−ey1Why: the definition of sinh, then e−y=ey1.
Clear the fraction2x=eye2y−12xey=e2y−1(ey)2−2xey−1=0Why: multiplying by ey (never zero) removes the fraction. Since e2y=(ey)2, this is a quadratic equation in the unknown ey with a=1, b=−2x, c=−1.
Use the quadratic formulaey=22x±4x2+4=x±x2+1Why: b2−4ac=4x2+4, and 4x2+4=2x2+1; the 2 cancels.
Choose the sign
Note that x−x2+1<0∀x∈]−∞,∞[. Since ey>0, we take the positive sign:ey=x+x2+1Why: x2+1>x2=∣x∣≥x, so the minus root is always negative, but an exponential is always positive.
Take lny=ln(x+x2+1)Hence, sinh−1x=ln(x+x2+1).
Why: ln undoes e: if ey=A with A>0, then y=lnA.
Check: take x=43: 169+1=45, so the formula gives y=ln(43+45)=ln2.sinh(ln2)=22−21=43It returns x. Also x=0 gives ln1=0, and sinh0=0.
for every real
x
. (Hence proved.)
exists for
x≥1
and its values are
≥0
.
To prove:
cosh−1x=ln(x+x2−1)
,
x≥1
.
Method: the same as Example 10:
x=coshy
, a quadratic in
ey
, then
ln
. This time both roots are positive, so the condition
y≥0
decides the sign.
Change the inverse into the function
For y∈[0,∞) = restricted domain of coshx and x∈[1,∞) = range of coshx, lety=cosh−1xx=coshyWhy: definition of the inverse on the restricted domain (Remember!: the domain of f is the range of f−1).
Write it with exponentialsx=2ey+e−y2x=ey+ey1Why: the definition of cosh and e−y=ey1.
Make a quadratic in ey2x=eye2y+12xey=e2y+1(ey)2−2xey+1=0Why: multiply by ey=0; here a=1, b=−2x, c=1.
Use the quadratic formulaey=22x±4x2−4ey=x±x2−1where x≥1, y≥0 and ey>0. Soy=ln(x±x2−1)Why: 4x2−4=2x2−1, which is real because x≥1.
Choose the sign
It can easily be observed that 0<x−x2−1<1 for all x≥1. Thus, y=ln(x−x2−1)<0
Concludey=ln(x+x2−1)cosh−1x=ln(x+x2−1)Why: replace y by cosh−1x.
Check: take x=45: 1625−1=43, so y=ln(45+43)=ln2≥0.cosh(ln2)=22+21=45It returns x. At x=1: ln1=0 and cosh0=1.
for all
x∈[1,∞)
. (Hence proved.)
. But
y≥0
, which is possible only if the positive sign is taken.
Why: the two roots multiply to 1:
(x−x2−1)(x+x2−1)
=x2−(x2−1)=1
For
x>1
the plus root is bigger than 1, so the minus root is between 0 and 1 and its