logo

Bessel Functions as Solutions to Differential Equations 📂Odinary Differential Equations

Bessel Functions as Solutions to Differential Equations

Theorem1

Theorem 1

Given a differential equation slightly different from the Bessel equation as follows:

y+12axy+[(bcxc1)2+a2ν2c2x2]y= 0orx2y+(12a)xy+[b2c2x2c+(a2ν2c2)]y= 0 \begin{equation} \begin{aligned} && y^{\prime \prime}+\frac{1-2a}{x}y^{\prime}+\left[ (bcx^{c-1})^{2}+\frac{a^{2}-\nu^{2}c^{2}}{x^{2}} \right]y =&\ 0 \\ \text{or} && x^{2}y^{\prime \prime}+(1-2a)xy^{\prime}+\left[ b^{2}c^{2}x^{2c}+(a^{2}-\nu^{2}c^{2}) \right]y =&\ 0 \end{aligned} \label{1} \end{equation}

And let Zν(x)Z_{\nu}(x) be any linear combination of Jν(x)J_{\nu}(x) and Nν(x)N_{\nu}(x). Then, the solution to the given differential equation is as follows:

y=xaZν(bxc)=xa[AJν(bxc)+BNν(bxc)] y=x^{a}Z_{\nu}(bx^{c})=x^{a}[AJ_{\nu}(bx^{c})+BN_{\nu}(bx^{c})]

ν\nu, aa, bb, cc, AA, BB are constants.

Theorem 2

x2y+xy+(K2x2ν2)y=0 x^{2}y^{\prime \prime} + xy^{\prime}+(K^{2}x^{2}-\nu^{2})y=0

The general solution to the differential equation is as follows:

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

Explanation

Bessel’s Equation

x2y+xy+(x2ν2)y= 0x(xy)+(x2ν2)y= 0y+1xy+(1ν2x2)y= 0 \begin{align*} x^2 y^{\prime \prime} +xy^{\prime} +(x^2-\nu^2)y =&\ 0 \\ x(xy^{\prime})^{\prime}+(x^2- \nu ^2) y =&\ 0 \\ y^{\prime \prime}+\frac{1}{x} y^{\prime} + \left( 1-\frac{\nu^{2}}{x^{2}} \right)y =&\ 0 \end{align*}

Even if a differential equation is not exactly like the Bessel equation, the fact that its solutions can be expressed using Bessel functions is of significant help in solving various differential equations. For example,

3xy+y+12y= 0xy+2y+4y= 0y+9xy= 0 \begin{align*} 3xy^{\prime \prime}+y^{\prime}+12y =&\ 0 \\ xy^{\prime \prime}+2y^{\prime}+4y =&\ 0 \\ y^{\prime \prime}+9xy =&\ 0 \end{align*}

Although this differential equation is not exactly in the form of Bessel’s equation, its solution can still be expressed by Bessel functions. The solutions are as follows:

y= x1/3Z2/3(4x1/2)=x1/3[AJ2/3(4x1/2)+BN2/3(4x1/2)]y= x1/2Z1(4x1/2)=x1/2[AJ1(4x1/2)+BN1(4x1/2)]y= x1/2Z1/3(2x3/2)=x1/2[AJ1/3(2x3/2)+BN1/3(2x3/2)] \begin{align*} y =&\ x^{1/3}Z_{2/3}(4x^{1/2} )=x^{1/3}\left[A J_{2/3}(4x^{1/2} )+BN_{2/3}(4x^{1/2} ) \right] \\ y =&\ x^{-1/2}Z_{1}(4x^{1/2} )=x^{-1/2}\left[A J_{1}(4x^{1/2} )+BN_{1}(4x^{1/2} ) \right] \\ y =&\ x^{1/2}Z_{1/3}(2x^{3/2} )=x^{1/2}\left[A J_{1/3}(2x^{3/2} )+BN_{1/3}(2x^{3/2} ) \right] \end{align*}

Proof

Proof 1

It is sufficient to show that it holds for y=xaJν(bxc)y=x^{a}J_{\nu}(bx^{c}). First, obtaining yy^{\prime}, yy^{\prime \prime} gives the following:

y= xaJν(bxc)y= axa1Jν(bxc)+bcxa+c1Jν(bxc)y= a(a1)xa2Jν(bxc)+abcxa+c2Jν(bxc)+(a+c1)bcxa+c2Jν(bxc)+b2c2xa+2c2Jν(bxc) \begin{align*} y =&\ x^{a}J_{\nu}(bx^{c}) \\ y^{\prime} =&\ ax^{a-1}J_{\nu}(bx^{c})+bcx^{a+c-1}J_{\nu}^{\prime}(bx^{c}) \\ y^{\prime \prime} =&\ a(a-1)x^{a-2}J_{\nu}(bx^{c}) +abcx^{a+c-2}J_{\nu}^{\prime}(bx^{c}) \\ & +(a+c-1)bcx^{a+c-2}J_{\nu}^{\prime} (bx^{c})+b^{2}c^{2}x^{a+2c-2}J_{\nu}^{\prime \prime}(bx^{c}) \end{align*}

Substituting this into (1)\eqref{1} results in:

[a(a1)xaJν(bxc)+abcxa+cJν(bxc)+(a+c1)bcxa+cJν(bxc)+b2c2xa+2cJν(bxc)]+[(12a)axaJν(bxc)+(12a)bcxa+cJν(bxc)]+[b2c2xa+2cJν(bxc)+(a2ν2c2)xaJν(bxc)]=0 \begin{align*} & \Big[a(a-1)x^{a}J_{\nu}(bx^{c}) +abcx^{a+c}J_{\nu}^{\prime}(bx^{c})+(a+c-1)bcx^{a+c}J_{\nu}^{\prime} (bx^{c}) +b^{2}c^{2}x^{a+2c}J_{\nu}^{\prime \prime}(bx^{c}) \Big] \\ & +\Big[(1-2a)ax^{a}J_{\nu}(bx^{c})+(1-2a)bcx^{a+c}J_{\nu}^{\prime}(bx^{c}) \Big] \\ & +\Big[ b^{2}c^{2}x^{a+2c}J_{\nu}(bx^{c})+(a^{2}-\nu^{2}c^{2})x^{a}J_{\nu}(bx^{c})\Big] = 0 \end{align*}

After organizing according to the differential coefficients:

(a2a+a2a2+b2c2x2c+a2ν2c2)xaJν(bxc)+(abc+abc+bc2bc+bc2abc)xa+cJν(bxc)+(b2c2x2c)xaJν(bxc)=0 \begin{align*} & (a^{2}-a+a-2a^{2}+b^{2}c^{2}x^{2c}+a^{2}-\nu^{2}c^{2})x^{a}J_{\nu}(bx^{c}) \\ & +(abc+abc+bc^{2}-bc+bc-2abc)x^{a+c}J_{\nu}^{\prime}(bx^{c}) \\ & +(b^{2}c^{2}x^{2c})x^{a}J_{\nu}^{\prime \prime}(bx^{c}) = 0 \end{align*}

Multiplying both sides by 1xa\dfrac{1}{x^{a}} and organizing the coefficients gives:

c2(b2x2cν2)Jν(bxc)+c2(bxc)Jν(bxc)+c2(b2x2c)Jν(bxc)= 0    (b2x2cν2)Jν(bxc)+(bxc)Jν(bxc)+(b2x2c)Jν(bxc)= 0 \begin{align*} && c^{2}(b^{2}x^{2c}-\nu^{2})J_{\nu}(bx^{c}) + c^{2}(bx^{c})J_{\nu}^{\prime}(bx^{c})+c^{2}(b^{2}x^{2c})J_{\nu}^{\prime \prime}(bx^{c}) =&\ 0 \\ \implies && (b^{2}x^{2c}-\nu^{2})J_{\nu}(bx^{c}) + (bx^{c})J_{\nu}^{\prime}(bx^{c})+(b^{2}x^{2c})J_{\nu}^{\prime \prime}(bx^{c}) =&\ 0 \end{align*}

Let’s denote by bxc=z bx^{c}=z, Jν(z)=yJ_{\nu}(z)=y, then from the above equation, we obtain the following:

(z2ν2)y+zy+z2y= 0    x2y+xy+(x2ν2)y= 0 \begin{align*} && (z^{2}-\nu^{2})y+zy^{\prime}+z^{2}y^{\prime \prime}=&\ 0 \\ \implies && x^{2}y^{\prime \prime}+xy^{\prime}+(x^{2}-\nu^{2})y =&\ 0 \end{align*}

This is a Bessel equation, and since Bessel functions are solutions to the Bessel equation, the above statement is valid.

Proof 2

x2d2ydx2+xdydx+(x2ν2)y=0 x^{2}\frac{ d ^{2}y}{ dx^{2} }+x\frac{ d y}{ d x }+(x^{2}-\nu^{2})y=0 The general solution to the above Bessel differential equation is y=AJν(Kx)+BNν(Kx)y=AJ_{\nu}(Kx)+BN_{\nu}(Kx). By substituting x=Kxx=Kx into the equation and general solution, we obtain the following:

K2x2d2yd(Kx)2+Kxdyd(Kx)+(K2x2ν2)y=0 K^{2}x^{2}\frac{ d ^{2}y }{ d (Kx)^{2} } +Kx\frac{ d y}{ d(Kx) } +(K^{2}x^{2}-\nu^{2})y=0

The general solution to this differential equation is y=AJν(Kx)+BNν(x)y=AJ_{\nu}(Kx)+BN_{\nu}(x). By organizing the constants of the differential equation, we get:

x2d2ydx2+xdydx+(K2x2ν2)y=0 x^{2}\frac{d^{2}y}{dx^{2}}+x\frac{ d y}{ dx }+(K^{2}x^{2}-\nu^{2})y=0


  1. Mary L. Boas, Mathematical Methods in the Physical Sciences (3rd Edition, 2008), p608-609 ↩︎