1.5 Trigonometric and Hyperbolic Functions

1.5.1 Definitions

We recall Euler’s formula for the sine and cosine,

\begin{eqnarray} \mathop{exp}\nolimits (ix) =\mathop{ cos}\nolimits (x) + i\mathop{sin}\nolimits (x),& & %&(1.59) \\ \end{eqnarray}

where x is a real number. From this, we can express sine and cosine as

\begin{eqnarray} \mathop{cos}\nolimits (x)& :=& {1\over 2}\left ({e}^{ix} + {e}^{−ix}\right ) %& \\ \mathop{sin}\nolimits (x)& :=& {1\over 2i}\left ({e}^{ix} − {e}^{−ix}\right ).%&(1.60) \\ \end{eqnarray}

We now define the hyperbolic functions ‘hyperbolic cosine’ and ‘hyperbolic sine’ as

\begin{eqnarray} \mathop{cosh}\nolimits (x)& :=& {1\over 2}\left ({e}^{x} + {e}^{−x}\right ) %& \\ \mathop{sinh}\nolimits (x)& :=& {1\over 2}\left ({e}^{x} − {e}^{−x}\right ),%&(1.61) \\ \end{eqnarray}

i.e. analogous to cosine and sine but without the imaginary unit i. Using {i}^{2} = −1, we recognise that

\begin{eqnarray} \mathop{cosh}\nolimits (x) =\mathop{ cos}\nolimits (ix),\quad \mathop{sinh}\nolimits (x) = −i\mathop{sin}\nolimits (ix),& & %&(1.62) \\ \end{eqnarray}

which means that trigonometric and hyperbolic functions are closely related. Their behaviour as a function of x, however, is different: while sine and cosine are oscillatory functions, the hyperbolic functions \mathop{cosh}\nolimits (x) and \mathop{sinh}\nolimits (x) are not oscillatory, because they are just linear combinations of {e}^{x} and {e}^{−x} which are not oscillatory. We have the following properties:

\begin{eqnarray} \mathop{cosh}\nolimits (0)& =& 1,\quad \mathop{cosh}\nolimits (x) =\mathop{ cosh}\nolimits (−x) %&(1.63) \\ \mathop{cosh}\nolimits (x →∞)& →& {1\over 2}{e}^{x} {→}_{ x→∞}∞,\quad \mathop{cosh}\nolimits (x →−∞) → {1\over 2}{e}^{−x} {→}_{ x→−∞}∞ %& \\ \mathop{sinh}\nolimits (0)& =& 0,\quad \mathop{sinh}\nolimits (x) = −\mathop{sinh}\nolimits (−x) %& \\ \mathop{sinh}\nolimits (x →∞)& →& {1\over 2}{e}^{x} {→}_{ x→∞}∞,\quad \mathop{sinh}\nolimits (x →−∞) →−{1\over 2}{e}^{−x} {→}_{ x→−∞}−∞.%& \\ \end{eqnarray}

from which we already can sketch the two hyperbolic functions, see Fig. 1.2.


trigh1


Figure 1.2: Hyperbolic functions

In addition, one defines the hyperbolic tangent and cotangent

\begin{eqnarray} \mathop{tanh}\nolimits (x) := {\mathop{sinh}\nolimits (x)\over \mathop{cosh}\nolimits (x)},\quad \mathop{coth}\nolimits (x) := {\mathop{cosh}\nolimits (x)\over \mathop{sinh}\nolimits (x)} .& & %&(1.64) \\ \end{eqnarray}


trigh2


Figure 1.3: Hyperbolic tangent and cotangent.

1.5.2 Inverse hyperbolic functions

Inverting

\begin{eqnarray} y =\mathop{ sinh}\nolimits (x) → x ={\mathop{ sinh}\nolimits }^{−1}(y),& & %&(1.65) \\ \end{eqnarray}

we find the inverse hyperbolic sine {\mathop{sinh}\nolimits }^{−1} by setting

\begin{eqnarray} y = {{e}^{x} − {e}^{−x}\over 2} ⇝ {e}^{2x} − 2y{e}^{x} − 1 = 0.& & %&(1.66) \\ \end{eqnarray}

This is a quadratic equation in u = {e}^{x} with the solutions

\begin{eqnarray}{ u}_{±} = y ±\sqrt{{y}^{2 } + 1}.& & %&(1.67) \\ \end{eqnarray}

Since u = {e}^{x} > 0 is positive, we must take the positive solution {u}_{+} and must discard the negative solution {u}_{−}. Therefore,

\begin{eqnarray}{ e}^{x} ≡ u = y + \sqrt{{y}^{2 } + 1} ⇝ x =\mathop{ ln}\nolimits \left (y + \sqrt{{y}^{2 } + 1}\right ),& & %&(1.68) \\ \end{eqnarray}

which means that

\begin{eqnarray} {\mathop{sinh}\nolimits }^{−1}(y) =\mathop{ ln}\nolimits \left (y + \sqrt{{y}^{2 } + 1}\right ).& & %&(1.69) \\ \end{eqnarray}

Similarly, one obtains

\begin{eqnarray} {\mathop{tanh}\nolimits }^{−1}(y) = {1\over 2}\mathop{ln}\nolimits \left [{1 + y\over 1 − y}\right ].& & %&(1.70) \\ \end{eqnarray}

The {\mathop{cosh}\nolimits }^{−1} is a bit more tricky.

1.5.3 Derivatives

These are obtained by going back to the definitions of the hyperbolic functions.

\begin{eqnarray} \mathop{sinh}\nolimits '(x) =\mathop{ cosh}\nolimits (x),\quad \mathop{cosh}\nolimits '(x) =\mathop{ sinh}\nolimits (x),\quad \mathop{tanh}\nolimits '(x) = 1 −{\mathop{ tanh}\nolimits }^{2}(x).& & %&(1.71) \\ \end{eqnarray}

1.5.4 Hyperbolic Identities

These also are obtained by using the definitions of \mathop{cosh}\nolimits and \mathop{sinh}\nolimits :

\begin{eqnarray} {\mathop{cosh}\nolimits }^{2}(x) −{\mathop{ sinh}\nolimits }^{2}(x)& =& 1,\quad \mathop{cosh}\nolimits (2x) = 1 + 2{\mathop{sinh}\nolimits }^{2}(x)%& \\ \mathop{sinh}\nolimits (2x)& =& 2\mathop{sinh}\nolimits (x)\mathop{cosh}\nolimits (x). %&(1.72) \\ \end{eqnarray}