logo

The Second Series Solution of the Bessel Equation: Bessel Functions of the Second Kind, Neumann Functions, Weber Functions 📂Odinary Differential Equations

The Second Series Solution of the Bessel Equation: Bessel Functions of the Second Kind, Neumann Functions, Weber Functions

Definition[^1]

A second solution of the Bessel equation is called the Neumann function, denoted by Nν(x)N_{\nu}(x) or Yν(x)Y_{\nu}(x). For non-integer ν\nu,

Nν(x)=Yν(x)=cos(νπ)Jν(x)Jν(x)sin(νπ) N_{\nu}(x)=Y_{\nu}(x)=\frac{\cos (\nu \pi)J_{\nu}(x)-J_{-\nu}(x)}{\sin (\nu\pi)}

For integer ν\nu, it is defined by the limit. For nZn\in \mathbb{Z}, νRZ\nu \in \mathbb{R}\setminus\mathbb{Z},

Nn(x)=limνnNν(x) N_{n}(x)=\lim \limits_{\nu \rightarrow n}N_{\nu}(x)

Here, J±ν(x)J_{\pm \nu}(x) is the first kind Bessel function. Thus, the general solution of the Bessel equation is as follows.

y(x)=AJν(x)+BNν(x) y(x)=AJ_{\nu}(x)+BN_{\nu}(x)

Here AA, BB are arbitrary constants.

Explanation

x2y+xy+(x2ν2)y=0 x^{2}y^{\prime \prime} + xy^{\prime} +(x^{2}-\nu^{2})y=0

The series solution of the above Bessel equation is denoted as J±ν(x)J_{\pm\nu}(x) and is called the ν\nuth first kind Bessel function.

Jν(x)=n=0(1)nΓ(n+1)Γ(n+ν+1)(x2)2n+ν J_{\nu}(x)=\sum \limits_{n=0}^{\infty} \frac{(-1)^{n} }{\Gamma (n+1) \Gamma (n+\nu+1)} \left(\frac{x}{2} \right)^{2n+\nu}

Jν(x)=n=0(1)nΓ(n+1)Γ(nν+1)(x2)2nν J_{-\nu}(x)=\sum \limits_{n=0}^{\infty}\frac{(-1)^{n}}{\Gamma (n+1)\Gamma (n-\nu+1)} \left( \frac{x}{2} \right)^{2n-\nu}

As you can see, since the two solutions are independent, the general solution is as follows.

y(x)=AJν(x)+BJν(x) y(x)=AJ_{\nu}(x)+BJ_{-\nu}(x)

However, when ν\nu is an integer, the two solutions are not linearly independent. Therefore, a second solution independent of Jν(x)J_{\nu}(x) must be found when ν\nu is an integer.

Consider briefly sinx\sin x and cosx\cos x. The two functions are linearly independent. However, any linear combination of sinx\sin x and cosx\cos x is also linearly independent of sinx\sin x. With this idea, any linear combination of Jν(x)J_{\nu}(x) and Jν(x)J_{-\nu}(x) is considered the second solution of the Bessel equation.

Nν(x)=cos(νπ)Jν(x)Jν(x)sin(νπ) \begin{equation} N_{\nu}(x)=\frac{\cos (\nu \pi)J_{\nu}(x)-J_{-\nu}(x)}{\sin (\nu\pi)} \label{eq1} \end{equation}

Nν(x)N_{\nu}(x) is independent of Jν(x)J_{\nu}(x) regardless of the condition of ν\nu. However, a problem arises again if ν\nu is an integer, then

Nν(x)=cos(νπ)Jν(x)Jν(x)sin(νπ)=(1)νJν(x)(1)νJν(x)0=00 N_{\nu}(x)=\frac{\cos (\nu \pi)J_{\nu}(x)-J_{-\nu}(x)}{\sin (\nu\pi)}=\frac{(-1)^{\nu}J_{\nu}(x)-(-1)^{\nu}J_{\nu}(x)}{0}=\frac{0}{0}

it becomes undefined. Therefore, when ν\nu is an integer, it is defined using the limit as follows.

Nn(x)=limνnNν(x)for nZ, νRZ N_{n}(x)=\lim \limits_{\nu \rightarrow n}N_{\nu}(x)\quad \text{for }n\in \mathbb{Z},\ \nu \in \mathbb{R}\setminus \mathbb{Z}

At this time, the above limit exists for arbitrary x0x \ne 0.

Theorem

For integer ν\nu, the Bessel function J±ν(x)J_{\pm \nu}(x) satisfies the following equation. In other words, it is not independent.

Jν(x)=(1)νJν(x) J_{-\nu}(x)=(-1)^{\nu}J_{\nu}(x)

Proof

Jν(x)=n=0(1)nΓ(n+1)Γ(nν+1)(x2)2nν J_{-\nu}(x)=\sum \limits_{n=0}^{\infty}\frac{(-1)^{n}}{\Gamma (n+1)\Gamma (n-\nu+1)} \left( \frac{x}{2} \right)^{2n-\nu}

By substituting n=k+νn=k+\nu, then

Jν(x)=k=ν(1)k+νΓ(k+ν+1)Γ(k+1)(x2)2k+ν J_{-\nu}(x)=\sum \limits_{k=-\nu}^{\infty}\frac{(-1)^{k+\nu}}{\Gamma (k+\nu+1)\Gamma (k+1)} \left( \frac{x}{2} \right)^{2k+\nu}

Since the gamma function diverges for 00 and negative integers, when k=ν,ν+1,,1k=-\nu,-\nu+1,\cdots,-1, the denominator’s Γ(k+1)\Gamma (k+1) diverges, making Jν(x)=0J_{-\nu}(x)=0. Therefore,

Jν(x)=k=ν(1)k+νΓ(k+ν+1)Γ(k+1)(x2)2k+ν=k=0(1)k+νΓ(k+ν+1)Γ(k+1)(x2)2k+ν=(1)νk=0(1)kΓ(k+ν+1)Γ(k+1)(x2)2k+ν=(1)νJν(x) \begin{align*} J_{-\nu}(x)&=\sum \limits_{k=-\nu}^{\infty}\frac{(-1)^{k+\nu}}{\Gamma (k+\nu+1)\Gamma (k+1)} \left( \frac{x}{2} \right)^{2k+\nu} \\ &=\sum \limits_{k=0}^{\infty}\frac{(-1)^{k+\nu}}{\Gamma (k+\nu+1)\Gamma (k+1)} \left( \frac{x}{2} \right)^{2k+\nu} \\ &=(-1)^{\nu}\sum \limits_{k=0}^{\infty}\frac{(-1)^{k}}{\Gamma (k+\nu+1)\Gamma (k+1)} \left( \frac{x}{2} \right)^{2k+\nu} \\ &=(-1)^{\nu}J_{\nu}(x) \end{align*}