[go: up one dir, main page]

10 Bessel FunctionsBessel and Hankel Functions

§10.24 Functions of Imaginary Order

With z=x and ν replaced by i⁢ν, Bessel’s equation (10.2.1) becomes

10.24.1 x2⁢d2wdx2+x⁢dwdx+(x2+ν2)⁢w=0.

For ν∈ℝ and x ∈(0,∞) define

10.24.2 J~ν⁡(x) =sech⁡(12⁢π⁢ν)⁢ℜ⁡(Ji⁢ν⁡(x)),
Y~ν⁡(x) =sech⁡(12⁢π⁢ν)⁢ℜ⁡(Yi⁢ν⁡(x)),

and

where γν is real and continuous with γ0=0; compare (5.4.3). Then

10.24.4 J~−ν⁡(x) =J~ν⁡(x),
Y~−ν⁡(x) =Y~ν⁡(x),

and J~ν⁡(x), Y~ν⁡(x) are linearly independent solutions of (10.24.1):

As x→+∞, with ν fixed,

10.24.6 J~ν⁡(x) =2/(π⁢x)⁢cos⁡(x−14⁢π)+O⁡(x−32),
Y~ν⁡(x) =2/(π⁢x)⁢sin⁡(x−14⁢π)+O⁡(x−32).

As x→0+, with ν fixed,

and

where γ denotes Euler’s constant §5.2(ii).

In consequence of (10.24.6), when x is large J~ν⁡(x) and Y~ν⁡(x) comprise a numerically satisfactory pair of solutions of (10.24.1); compare §2.7(iv). Also, in consequence of (10.24.7)–(10.24.9), when x is small either J~ν⁡(x) and tanh⁡(12⁢π⁢ν)⁢Y~ν⁡(x) or J~ν⁡(x) and Y~ν⁡(x) comprise a numerically satisfactory pair depending whether ν≠0 or ν=0.

For graphs of J~ν⁡(x) and Y~ν⁡(x) see §10.3(iii).

For mathematical properties and applications of J~ν⁡(x) and Y~ν⁡(x), including zeros and uniform asymptotic expansions for large ν, see Dunster (1990a). In this reference J~ν⁡(x) and Y~ν⁡(x) are denoted respectively by Fi⁢ν⁡(x) and Gi⁢ν⁡(x).