BESSEL EQUATIONS AND BESSEL FUNCTIONS

Bessel functions form a class of the so called special functions. They are im-
portant in math as well as in physical sciences (physics and engineering). They
are especially important in solving boundary values problems in cylindrical coordi-
nates. First we define another important function: the Gamma function which is
used in the series expansion of the Bessel functions, then we construct the Bessel
functions Jα and Yα .

1. The Gamma Function

The Gamma function (also called Euler’s integral ) is the function defined for
x > 0 by ∫ ∞
Γ(x) = e−s sx−1 ds .
0
The improper integral defining Γ is convergent for x > 0. To see why, note that for
every x > 0,
e−s sx−1
lim =0.
s→∞ s−2
−s x−1
Thus there exists M > 0 such that e s ≤ s−2 for s > M . This implies that
∫ ∞ ∫ ∞
ds 1
e−s sx−1 ds ≤ 2
= .
M M s M
Also for s ∈ (0, M ), e−s sx−1 ≤ sx−1 and
∫ M ∫ M [ x ]s=M
s Mx
e−s sx−1 ds ≤ sx−1 ds = = .
0 0 x s=0 x
We have then
∫ ∞ ∫ ∞ ∫ M
1 Mx
e−s sx−1 ds = e−s sx−1 ds + e−s sx−1 ds ≤ + .
0 M 0 M x
This shows that Γ(x) is well defined for x > 0.
The most important property of the Gamma function is given in the following
lemma.
Lemma 1. The function Γ satisfies the following
Γ(x + 1) = xΓ(x) , ∀x > 0 .

Proof. The proof is simply an integration by parts

∫ A ∫ A ∫
[ ]A A
e−s sx ds = −e−s sx 0 + x e−s sx−1 ds = x e−s sx−1 ds − Ax e−A .
0 0 0
By taking the limit as A → ∞, we get Γ(x + 1) = xΓ(x)

Date: March 31, 2014.

1
2 BESSEL EQUATIONS AND BESSEL FUNCTIONS

It can be shown that Γ has derivatives of all orders for x > 0 and that Γ has
a unique extremum (global minimum) on the interval (0, ∞). The minimum is
reached at a number x0 ∈ (1, 2) and Γ(x0 ) < 1. Furthermore, Γ satisfies
lim Γ(x) = ∞, and lim Γ(x) = ∞
x→0+ x→∞

Graph of Γ(x) over (0, ∞)

1 2

We can use the fundamental property to extend Γ as a smooth functions to

R\{0, −1, −2, · · · } (the whole real line except 0 and the negative integers). First
we extend Γ to the interval (−1, 0) by defining
Γ(x + 1)
Γ(x) = for x ∈ (−1, 0)
x
(note the above definition makes sense since x+1 ∈ (0, 1) and Γ(x+1) is defined by
the integral). Once, Γ is defined on (−1, 0), we extend it to the interval (−2, −1) by
using the same property. More precisely, if Γ is defined on the interval (−j, −(j −
1)) with j ∈ Z+ , then we extend it to the interval (−(j + 1), −j) by using the
fundamental property. We have in particular that limx→k |Γ(x)| = ∞ for k =
0, −1, −2, · · ·
Now we compute some values of the Gamma function.
∫ ∞
Γ(1) = e−s ds = 1 .
0
By using the fundamental property of Γ, we get easily its values at the positive
integers.
Γ(2) = Γ(1 + 1) = 1Γ(1) = 1
Γ(3) = Γ(2 + 1) = 2Γ(2) = 2 = 2!
Γ(4) = Γ(3 + 1) = 3Γ(3) = 3 · 2! = 3!
..
.
Γ(n + 1) = n! ∀n ∈ Z+
The Gamma function appears as an interpolation of the factorial function.
 BESSEL EQUATIONS AND BESSEL FUNCTIONS 3

Graph of Γ(x)

−3 −2 −1

∫ ∞ √
e−t dt =
2
To compute Γ(1/2) we use the value of the Gaussian integral π /2
0
(you have probably encountered this integral in Multivariable Calculus (MAC2313)
or in Prob./Statistics
√ class). In the following calculation, we have made the substi-
tution t = s.
( ) ∫ ∞ ∫ ∞
1 −s −1/2
√
e−t dt = π .
2
Γ = e s ds = 2
2 0 0

The Gamma function satisfies many other identities such:

π
Reflection formula : Γ(x)Γ(1 − x) = (x ̸= 0, ±1, ±2, · · · )
2x−1
sin πx( )
2 1
Duplication formula : Γ(2x) = √ Γ(x)Γ x + (2x ̸= 0, −1, −2, · · · )
π 2

2. Bessel’s Equation
Bessel’s equation of order α (with α ≥ 0) is the second order differential equation
(1) x2 y ′′ + xy ′ + (x2 − α2 )y = 0
In order to find all solutions we need two independent solutions. We are going to
construct the independent solutions for x > 0.

2.1. Construction of a first solution. Note that x = 0 is a singular point of

the equation. More precisely, it is a regular singular point (see your notes from the
first differential equations class, MAP2302). For such differential equations, we can
use the method of Fröbenius to construct series solutions. We seek a (formal) series
solution
∑∞ ∞
∑
(2) y = xr ck xk = ck xk+r
k=0 k=0
4 BESSEL EQUATIONS AND BESSEL FUNCTIONS

of equation (1), with ck ∈ R, and c0 ̸= 0. The substitution of this series and its
(formal) derivatives into equation (1) gives
∑∞ ∞
∑ ∞
∑
x2 (k + r)(k + r − 1)ck xk+r−2 + x (k + r)ck xk+r−1 + (x2 − α2 ) ck xk+r = 0
k=0 k=0 k=0
We rewrite this as
∞
∑ ∞
∑ ∞
∑ ∞
∑
(k + r)(k + r − 1)ck xk+r + (k + r)ck xk+r + ck xk+r+2 − α2 ck xk+r = 0
k=0 k=0 k=0 k=0
then as
∑∞ ∞
∑ ∞
∑ ∞
∑
(k + r)(k + r − 1)ck xk+r + (k + r)ck xk+r + ck−2 xk+r − α2 ck xk+r = 0
k=0 k=0 k=2 k=0
After grouping the like terms and simplifying, we obtain
∞
∑ [ ]
(r2 − α2 )c0 xr + ((r + 1)2 − α2 )c1 xr+1 + ((r + k)2 − α2 )ck + ck−2 xk+r = 0
k=2
In order for this series to be identically zero, each coefficient must be zero. We have
then
( − α 2)c0 =2 )0 ,
2 2
(r
((r + 1)2 − α2 ) c1 = 0 ,
(r + j) − α cj + cj−2 = 0 , j = 2, 3, 4, · · ·
Since c0 ̸= 0, then the first equation implies that r must satisfy
r 2 − α2 = 0 .
This is the indicial equation of the Bessel equation. The indicial roots are
r=α and r = −α .
Consider the case r = α. The second equation becomes
(2α + 1)c1 = 0 ⇒ c1 = 0 (since α > 0) .
For j ≥ 2 the recurrence relation becomes
( ) −cj−2
(α + j)2 − α2 cj + cj−2 = 0 ⇒ cj = .
j(2α + j)
Since c1 = 0, the above relation gives
−c1 −c3
c3 = = 0, c5 = = 0, c7 = 0, · · ·
3(2α + 3) 5(2α + 5)
That is, all coefficients with odd indices are 0 (codd = 0). For the coefficients with
even indices, we have
−c0 −c0
c2 = =
2(2α + 2) 4(1 + α)
−c2 (−1)2 c0
c4 = = 4
4(2α + 4) 2 (2!)(1 + α)(2 + α)
−c4 (−1)3 c0
c6 = = 6
6(2α + 6) 2 (3!)(1 + α)(2 + α)(3 + α)
A proof by induction gives
(−1)j c0
c2j = , j = 1, 2, 3, · · ·
j!22j (1 + α)(2 + α) · · · (j + α)
 BESSEL EQUATIONS AND BESSEL FUNCTIONS 5

A formal solution is therefore

∞
∑ ∞
∑ (−1)j c0
y= c2j x2j+α = x2j+α
j=0 j=0
j!22j (1 + α)(2 + α) · · · (j + α)
We are going to select c0 and use the Gamma function to rewrite the series solution
in a more compact form. It follows from the fundamental property of the Gamma
function that
Γ(j + 1 + α) = (j + α)Γ(j + α)
= (j + α)(j − 1 + α)Γ(j − 1 + α)
..
.
= (j + α)(j − 1 + α) · · · (1 + α)Γ(1 + α) .
Equivalently,
Γ(j + 1 + α)
(1 + α)(2 + α) · · · (j + α) = .
Γ(1 + α)
We select c0 as
1
c0 = .
2α Γ(1 + α)
With this choice of c0 , the particular series solution becomes
∑∞
(−1)j ( x )2j+α
Jα (x) = .
j=0
j!Γ(j + 1 + α) 2
This solution is known as the Bessel function of the first kind of order α.
Now we determine the domain where the series converges. Note that
( x )α ∑∞
(−1)j ( x )2j
Jα (x) = .
2 j=0
j!Γ(j + 1 + α) 2

The last series is a power series in (x/2)2 . To find its radius of convergence, we can
use the ratio test:
 
 (−1)j+1 / ((1 + j)!Γ(j + 2 + α))  1
lim   = lim
j→∞  (−1)j / (j!Γ(j + 1 + α))  j→∞ (j + 1)(j + 1 + α) = 0 .
The radius of convergence is infinite (the power series converges to an analytic
function on R). The function Jα (x) is defined for x ≥ 0.
2.2. Construction of a second solution. Recall that the indicial roots of the
Bessel equation are r = ±α. We have used r = α to construct the solution Jα (x).
We can redo the above construction with r = −α. However, this can be done only
if α ̸∈ Z+ . In this case a second independent solution of Bessel’s equation is
∑∞
(−1)j ( x )2j−α
J−α (x) = .
j=0
j!Γ(j + 1 − α) 2
Note that J−α is not defined at x = 0. We have
lim |J−α (x)| = ∞ .
x→0+
The general solution of equation in (0, ∞) is
y(x) = AJα (x) + BJ−α (x) ,
with A and B constants.
6 BESSEL EQUATIONS AND BESSEL FUNCTIONS

When α = n ∈ Z+ , the situation is a little more involved. The first solution is

(−1)j ( x )2j+n
∑∞
Jn (x) = .
j=0
j!(j + n)! 2
If we try to define J−n by using the recurrence relations for the coefficients, then
starting with c0 ̸= 0, we can get
−c0 −c0
c2 = =
2(2 − 2n) 4(1 − n)
−c2 (−1)2 c0
c4 = = 4
4(4 − 2n) 2 (2!)(1 − n)(2 − n)
..
.
(−1)n−1 c0
c2(n−1) =
(n − − n)(2 − n) · · · 2 · 1
1)!22(n−1) (1
At the order 2n however we get
0c2n − c2(n−1) = 0 .
This is a contradiction since c2(n−1) ̸= 0. Thus, the recurrence relations will not
lead to a series solution.
Another attempt to define J−n is to define it as
J−n (x) = lim J−α (x).
α→n
In this case, we get back either J−n = ±Jn and J−n and Jn are dependent solutions
of the equations. More precisely, we have the following lemma.
Lemma. We have
J−n (x) = (−1)n Jn (x)

Proof. For α ̸∈ Z+ (and α close to n), we have

∑∞
(−1)j
J−α (x) = x−α 2j−α Γ(j + 1 − α)
x2j .
j=0
j!2

Recall that limz→−p |Γ(z)| = ∞ for p = 0 or p ∈ Z+ . When α → n, (j + 1 − α)

tends to 0 or a negative integer for j = 0, 1, 2, · · · , (n − 1). For such values of j,
the coefficients of x2j in the series above approaches 0:
(−1)j
lim =0.
α→n j!22j−α Γ(j + 1 − α)
We get then,
∞
∑ (−1)j
J−n (x) = lim J−α (x) = x−n x2j .
α→n
j=n
+ 1 − n)
j!22j−n Γ(j
and after using the fundamental property of the Gamma function we obtain
∑∞ ∞
∑
(−1)j x2j−n (−1)k+n x2k+n
J−n (x) = 2j−n (j − n)!
= 2k+n k!
= (−1)n Jn (x) .
j=n
j!2 (k + n)!2
k=0

Now we indicate how to construct a second independent solution of equation (1)

when α = n. Consider α = n + ϵ with 0 < ϵ < 1 (hence such α ̸∈ Z+ ). The
 BESSEL EQUATIONS AND BESSEL FUNCTIONS 7

corresponding Bessel equation has two independent solutions Jn+ϵ and J−(n+ϵ) .
The function Yn+ϵ defined by
Jn+ϵ (x) − (−1)n J−(n+ϵ) (x)
Yn+ϵ (x) = .
ϵ
Since function Yn+ϵ is a linear combination of Jn+ϵ and J−(n+ϵ) , then Yn+ϵ is also
a solution of the corresponding Bessel’s equation of order α = n + ϵ. We define Yn
as:
Jn+ϵ (x) − (−1)n J−(n+ϵ) (x)
Yn (x) = lim Yn+ϵ (x) = lim .
ϵ→0 ϵ→0 ϵ
It can be proved that the function Yn is a solution of the Bessel equation of order
n and that Yn and Jn are independent (see for example R.Courant and D. Hilbert,
Method of Mathematical Physics, vol. 2, or H. Sagan, Boundary and Eigenvalue
Problems of Mathematical Physics). This solution Yα is called the Bessel function
of the second kind of order n. It can also be proved that
lim Yn (x) = −∞ .
x→0+

Another method to obtain a second solution of the Bessel equation in the excep-
tional case is to seek it in the form
∑
y(x) = Jn (x) ln x + Cj xj .
j

The coefficients Cj are then found by a recurrence relation.

The explicit expression of the Yn (x) is given below. Its derivation can be found
in advanced texts about special function. For n ∈ Z + , we have
2 ( x ) xn ∑ (−1)j (cj+n + cj ) 2j
∞
Yn (x) = Jn (x) γ + ln − x
π 2 π j=0 22j+n (j!)(n + j)!
1 ∑ (n − j − 1)! 2j
n−1
− x ,
πxn j=0 22j−n (j!)

where, for j = 0, 1, 2, · · · , the constants cj are given by

1 1 1
c0 = 0, c1 = 1, c2 = 1 + , · · · , cj = 1 + + · · ·
2 2 j
and where γ is the Euler constant given by
( )
1 1
γ = lim 1 + + · · · − ln j , γ ≈ 0.57721....
j→∞ 2 j
For n = 0, we have
2 ( x ) 2 ∑ (−1)j cj 2j
∞
Y0 (x) = Jn (x) γ + ln − x
π 2 π j=0 22j (j!)2

2.3. General solution of the Bessel equation. We summarize the above dis-
cussions in the following theorem.
Theorem. Given the Bessel equation of order α ≥ 0,
x2 y ′′ + xy ′ + (x2 − α2 )y = 0
then we have the followings:
8 BESSEL EQUATIONS AND BESSEL FUNCTIONS

• If α ̸∈ Z+ ∪ {0}, the equation has two independent solutions Jα (x) and

J−α (x) (Bessel functions of the first kind) and the general solution is

y(x) = AJα (x) + BJ−α (x) ,

where A and B are constants.

• If α = n with n = 0 or n ∈ Z+ , the equation has only one Bessel function
of the first kind Jn (x), another independent solution is the Bessel function
of the second kind Yn (x). The general solution of the equation is

y(x) = AJn (x) + BYn (x) .

3. Remarks on Bessel functions

The expansions of the functions J0 and J1 are

(−1)j ( x )2j
∑∞
x2 x4 x6
J0 (x) = 2
=1− 2 + 4 2
− 6 + ···
j=0
(j!) 2 2 2 (2!) 2 (3!)2
(−1)j ( x )2j+1
∑∞
x x3 x5 x7
J1 (x) = = − 3 + 5 − 7 + ···
j=0
j!(j + 1)! 2 2 2 (2!) 2 (2!)(3!) 2 (3!)(4!)

The graphs of J0 and of J1 resemble those of cosine and sine with a decreasing

Graphs of J0 and J1
1.5
J0(x)
J1(x)

0.5

−0.5

−1
0 2 4 6 8 10 12 14 16 18 20

amplitude. Notice how the zeros of J0 and J1 behave. Between two consecutive
zeros of J0 there is exactly one zero of J1 . The following table lists the approximate
values of the first 9 positive zeros of J0 and J1

j 1 2 3 4 5 6 7 8 9
J0 2.405 5.520 8.654 11.792 14.931 18.071 21.212 24.353 27.494
J1 3.832 7.016 10.174 13.324 16.471 19.616 22.760 25.904 29.047
 BESSEL EQUATIONS AND BESSEL FUNCTIONS 9

For n large, the n-th zero of J0 is approximately nπ − (π/4) and the n-th zero
of J1 is approximately nπ + (π/4). It is shown that for x large we have
√ ( √ ( )
2 π) 2 3π
J0 (x) ≈ cos x − and J1 (x) ≈ cos x −
πx 4 πx 4
In fact the m-th Bessel function Jm has the following behavior
√ ( )
2 (2m + 1)π
Jm (x) ≈ cos x − for x large.
πx 4
This approximation shows that Jm has infinitely many positive zeros that tends
to infinity. More precisely, we have following proposition about the zeros of Bessel
functions.
Proposition 1. For every α ∈ R, the positive zeros of Jα form an increasing
unbounded sequence. That is, the solution set of the equation

Jα (x) = 0, x>0,

forms a sequence

0 < x1 < x2 < x3 < · · · < xn < · · · , with lim xn = ∞

n→∞

The proof of this proposition is beyond the aim of this course.

1 3 5 2k + 1
For α = ± , ± , ± , · · · , ± , · · · , the Bessel functions Jα are elementary
2 2 2 2
functions. This means that Jα (x) can be expressed algebraically in terms of sin x,
cos x and x. The following proposition gives the expressions of some Bessel functions
with such indices.
Proposition 2. We have the following relations
√
2
J1/2 (x) = sin x ,
√πx
2
J−1/2 (x) = cos x ,
√ πx( )
2 sin x
J3/2 (x) = − cos x ,
πx
√ x
2 ( cos x )
J−3/2 (x) = − + sin x .
πx x

Proof. We prove the first relation and leave the others as an exercise. Recall that
the Taylor expansion of sin x is
∞
∑ (−1)k 2k+1
sin x = x .
(2k + 1)!
k=0
√
We need the value of Γ(j + (3/2)). We have Γ(1/2) = π (see section about the
Gamma function). By using the fundamental property of the Gamma function, we
10 BESSEL EQUATIONS AND BESSEL FUNCTIONS

get ( ) ( ) ( ) √
3 1 1 1 π
Γ =Γ +1 = Γ = ,
2
( ) ( 2 ) 2 2
( ) 2 √
5 3 3 3 (3 · 1) π
Γ =Γ +1 = Γ = ,
(2) (2 ) 2 (2) 22 √
7 5 5 5 (5 · 3 · 1) π
Γ =Γ +1 = Γ = .
2 2 2 2 23
We prove by induction that for j ∈ Z+ ,
( )
3 (2j + 1)(2j − 1) · · · 3 · 1 √
Γ j+ = π.
2 2j+1
We can simplify the product of the odd integers above as
(2j + 1)! (2j + 1)!
(2j + 1)(2j − 1) · · · 5 · 3 · 1 = = .
(2j)(2j − 2) · · · 4 · 2 2j (j!)
Hence, ( )
3 (2j + 1)! √
Γ j+ = 2j+1 π.
2 2 (j!)
Now we use these to show the first relation of the proposition.
∑∞
(−1)j ( x )2j+(1/2)
J1/2 (x) =
j!Γ(j + (3/2)) 2
√
j=0
∞
2∑ (−1)j
= x2j+1
x j=0 22j+1 (j!)Γ(j + (3/2))
√ ∞
2∑ (−1)j 22j+1 (j!)
= 2j+1
√ x2j+1
x j=0 2 (j!)(2j + 1)! π
√ ∞
2 ∑ (−1)j 2j+1
= x
πx j=0 (2j + 1)!
√
2
= sin x
πx

Analogous results about the behaviors of the Bessel functions of the second kind
can be obtained.
 BESSEL EQUATIONS AND BESSEL FUNCTIONS 11

Bessel functions of the second kind: Y , Y , Y

0 1 2

0.4 Y0(x)
Y1(x)
Y2(x)
0.2

−0.2

−0.4

−0.6

−0.8

−1

−1.2

−1.4

−1.6

0 2 4 6 8 10 12 14 16 18 20

4. Some properties of the Bessel functions Jα

The bessel functions satisfy a large number of properties. We limit ourself here
to list the following.
Properties of Jα .

(1) J0 (0) = 1 and Jα (0) = 0 if α > 0.

(2) Jn (x) is an even function if n ∈ Z+ is even and Jn (x) is an odd function in
n is odd.
(3) J−n (x) = (−1)n Jn (x) for n ∈ Z+ .
d ( −α )
(4) x Jα (x) = −x−α Jα+1 (x).
dx
d
(5) (xα Jα (x)) = xα Jα−1 (x).
dx
d 1
(6) (Jα (x)) = (Jα−1 (x) − Jα+1 (x)) .
dx 2
2α
(7) Jα+1 (x) + Jα−1 (x) = Jα (x).
∫ x
(8) x−α Jα+1 (x)dx = −x−α Jα (x) + C.
∫
(9) xα Jα−1 (x)dx = xα Jα (x) + C.

The first two properties are easy to obtain from the series representation of Jα
and the third has already been verified.
12 BESSEL EQUATIONS AND BESSEL FUNCTIONS

Proof of 4. Multiply the series representation of Jα by x−α and differentiate

∞
∑
d ( −α ) (−1)j 2j x2j−1
x Jα (x) =
dx j=1
j!Γ(j + α + 1) 22j+α
∞
∑ (−1)j x2j−1
=
j=1
(j − 1)!Γ(j + α + 1) 22j−1+α
∞
∑ (−1)j ( x )2j−1+α
= x−α
j=1
(j − 1)!Γ(j + α + 1)
2
∑∞
(−1)k+1 ( x )2k+(α+1)
= x−α
k!Γ(k + 1 + (α + 1)) 2
k=0
−α
= −x Jα+1 (x)

Proof of 5. Left as an exercise

Proof of 6. We have (take into account properties 4 and 5)
d d ( α ( −α )) ( ) d ( −α )
(Jα (x)) = x x Jα (x) = αxα−1 x−α Jα (x) + xα x Jα (x)
dx dx−1 ( ) dx
= αx Jα (x) + xα −x−α Jα+1 (x)
= αx−1 Jα (x) − Jα+1 (x)
Similarly
d d ( −α α ) d
(Jα (x)) = x (x Jα (x)) = −αx−α−1 (xα Jα (x)) + x−α (xα Jα (x))
dx dx −1 dx
= −αx Jα (x) + x−α (xα Jα−1 (x))
= −αx−1 Jα (x) + Jα−1 (x)
By adding the two expressions we get
d
2 (Jα (x)) = Jα−1 (x) − Jα+1 (x)
dx
Proof of 7. It follows from the proof of 6. that
d
(Jα (x)) + αx−1 Jα (x) = Jα−1 (x)
dx
d
(Jα (x)) − αx−1 Jα (x) = −Jα+1 (x)
dx
We get, by subtraction,
2αx−1 Jα (x) = Jα−1 (x) + Jα+1 (x)

Proof of 8. It follows from property 4 that

∫ ∫
−α d ( −α )
x Jα+1 (x)dx = − x Jα (x) dx = −x−α Jα (x) + C
dx

Proof of 9. Left as an exercise.

√
2
Example 1. We have proved in Proposition 2 that J1/2 (x) = sin x. In one
√ πx
2
of the exercises you will be asked to prove that J−1/2 (x) = cos x We can use
πx
 BESSEL EQUATIONS AND BESSEL FUNCTIONS 13

property 7 with α = 1/2 to deduce that

1
J3/2 (x) + J−1/2 (x) = J1/2 (x)
x
Thus,
( √ )
2 sin x
J3/2 (x) = − cos x .
πx x
Similar arguments can be used to prove that Jk+(1/2) is an elementary function.
Example 2. We can use property 5 with α = 1 to get
(xJ1 (x))′ = xJ0 (x) ⇔ xJ1′ (x) + J1 (x) = xJ0 (x)
or
J1 (x)
J1′ (x) = J0 (x) −
x

The following table lists the values J0 (x) and of J1 (x) for some values of x
between 0 and 10.
x 0 0.5 1.0 1.5 2.0 2.5 3.0
J0 (x) 1.0000 0.9385 0.7652 0.5118 0.2239 -0.0484 -0.2601
J1 (x) 0 0.2423 0.4401 0.5579 0.5767 0.4971 0.3391
x 3.5 4 4.5 5 5.5 6 6.5
J0 (x) -0.3801 -0.3971 -0.3205 -0.1776 -0.0068 0.1506 0.2601
J1 (x) 0.1374 -0.0660 -0.2311 -0.3276 -0.3414 -0.2767 -0.1538
x 7 7.5 8 8.5 9 9.5 10
J0 (x) 0.3001 0.2663 0.1717 0.0419 -0.0903 -0.1939 -0.2459
J1 (x) -0.0047 0.1352 0.2346 0.2731 0.2453 0.1613 0.0435

By repeated use of property 7, we can get Jn (x) for any integer n once J0 (x)
and J1 (x) are known.
Example 3. Let use the table to find J4 (3.5). We have J0 (3.5) = −0.3801 and
J1 (3.5) = 0.1374. By using property 7 with α = 1, then α = 2, 3, and 4, we get
2
J2 (3.5) + J0 (3.5) = J1 (3.5) J2 (3.5) = 0.4586
3.5
4
J3 (3.5) + J1 (3.5) = J2 (3.5) J3 (3.5) = 0.3868
3.5
6
J4 (3.5) + J2 (3.5) = J3 (3.5) J4 (3.5) = 0.2044
3.5
Example 4. We use the integral property 8 and integration by parts to find the
following integral
∫ ∫
x J5 (x)dx = x2 (x−4 J5 (x))dx
−2
∫
= x2 (−x−4 J4 (x)) + (2x)x−4 J4 (x)dx
∫
= −x J4 (x) + 2 x−3 J4 (x)dx
−2

= −x−2 J4 (x) + 2x−3 J3 (x) + C

14 BESSEL EQUATIONS AND BESSEL FUNCTIONS

5. An integral representation of Jn (x)

There is an interesting representation of the Bessel functions of the first kind with
integer order n in terms of a definite integral. We have the following proposition.
Proposition 3. For n ∈ Z, we have
∫
1 π
Jn (x) = cos (nθ − x sin θ) dθ .
π 0

Proof. Recall the Taylor expansion of the exponential function

∞
∑ zj
ez = ∀z ∈ C .
j=0
j!

(the series converges uniformly and absolutely for |z| ≤ R for every R > 0). We
have then
∞ j ( )j
(−1)j ( x )j
∑ ∑∞
t x
ext/2 = and e−x/2t = .
j=0
j! 2 j=0
j!tj 2

The product is
∞ j ( )j
t x (−1)k ( x )k ∑ ∑ (−1)k ( x )j+k j−k
∑∞ ∑ ∞ ∞
ext/2 e−x/2t = = t
j=0
j! 2 k!tk 2 j=0
(j!)(k!) 2
k=0 k=0
n
We rewrite this relation as a power series in t (so the coefficients will depend on
x).
∑∞ ∑∞
1
e(x/2)(t−(1/t)) = C−n (x) n + C0 (x) + Cn (x)tn .
n=1
t n=1
We need to show that Jm (x) = Cm (x). The coefficient Cn (x) is obtained from the
double series by grouping all the coefficients of tm . Thus all term with j − k = m:
∑ (−1)k ( x )j+k
Cm (x) =
(j!)(k!) 2
j−k=m, j,k≥0

or equivalently (by setting k = j − m),

(−1)j−m ( x )2j−m (−1)j−m ( x )2j−m
∑ ∞ ∑∞
m
Cm (x) = = (−1)
j=0
(j!)(j − m)! 2 j=0
(j!)(j − m)! 2

The last series is precisely J−m (x). We have then

Cm (x) = (−1)m J−m (x) = Jm (x) .
The expansion of e(x/2)(t−(1/t)) is therefore
∞
∑ [ ]
(−1)n
e(x/2)(t−(1/t)) = J0 (x) + Jn (x) tn + .
n=1
tn
Now we evaluate the left side and the right side of the above expression for t =
eiθ = cos θ + i sin θ. For n ∈ Z+ , we have
{
n (−1)n inθ n −inθ 2 cos(nθ) if n = 2p is even
t + = e + (−1) e =
tn 2i sin(nθ) if n = 2p + 1 is odd
 BESSEL EQUATIONS AND BESSEL FUNCTIONS 15

and
e(x/2)(t−(1/t)) = eix sin θ = cos(x sin θ) + i sin(x sin θ) .
It follows that
∞
∑ ∞
∑
cos(x sin θ) + i sin(x sin θ) = J0 (x) + 2 J2p (x) cos(2pθ) + 2i J2p+1 (x) sin(2p + 1)θ
p=1 p=0

By equating the real and imaginary parts, we get

∞
∑
cos(x sin θ) = J0 (x) + 2 J2p (x) cos(2pθ)
p=1
∞
∑
sin(x sin θ) =2 J2p+1 (x) sin(2p + 1)θ
p=0

Recall the orthogonality of the trigonometric system

∫ ∫ {
2 π 2 π 1 if j = k ,
cos(jθ) cos(kθ) = cos(jθ) cos(kθ) =
π 0 π 0 0 ̸ k.
if j =

By using these orthogonality relations and the above series, we get

∫ ∫ [ ∑∞
]
1 π 2 π
cos(x sin θ) cos(nθ)dθ = J0 (x) + J2p (x) cos(2pθ) cos(nθ)dθ
π 0 π 0
{ p=1
Jn (x) if n is even,
=
0 if n is odd.

Similarly,
∫ ∫ [∞ ]
1 π
2 π∑
sin(x sin θ) sin(nθ)dθ = J2p+1 (x) sin(2p + 1)θ sin(nθ)dθ
π 0 π 0 p=0
{
Jn (x) if n is odd,
=
0 if n is even.

By adding these relations we get for n ∈ Z that

∫
1 π
[cos(x sin θ) cos(nθ) + sin(x sin θ) sin(nθ)] dθ = Jn (x)
π 0
which proves the proposition.
A immediate consequence of the integral representation is the following
Corollary. For every n ∈ Z, we have

|Jn (x)| ≤ 1, ∀x ∈ R

and
lim Jn (x) = 0 .
x→∞
16 BESSEL EQUATIONS AND BESSEL FUNCTIONS

6. Exercises

Exercise 1. The table bellow lists approximate values of the Gamma function
for values of x in the interval [0, 1]. Use the table together with the fundamental
property of the Gamma function to find the following values
Γ(9.45), Γ(23.10), Γ(6.05), Γ(4.85), Γ(8.85),
Γ(−0.75), Γ(−4.65), Γ(−0.01), Γ(−2, 85), Γ(−3.75).
x 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50
Γ(x) 19.470 9.513 6.220 4.591 3.626 2.992 2.546 2.218 1.968 1.773
x 0.55 0.60 0.65 0.70 0.75 0.80 0.85 0.90 0.95 1.00
Γ(x) 1.616 1.489 1.385 1.298 1.225 1.164 1.113 1.069 1.032 1.00
Exercise 2. The aim of this exercise is to establish the formulas
∫ π/2
Γ(x)Γ(y)
=2 cos2x−1 θ sin2y−1 θdθ x > 0, y > 0 (∗)
Γ(x + y) 0

1. Show that
∫ ∞ ∫ ∞ ∫ ∞ ∫ ∞
−s x−1 −t y−1
e−(u
2
+v 2 ) 2x−1 2y−1
Γ(x)Γ(y) = e s ds e t dt = u v dudv
0 0 0 0

(Hint: consider the substitutions s = u2 and t = v 2 )

2. Use polar coordinates u = r cos θ, v = r sin θ to establish formula (∗).
3. Use formula (∗) to establish the following formula (j, k ∈ Z+ )
∫ π/2
(j − 1)! (k − 1)!
cos2j−1 θ cos2k−1 θdθ =
0 2 (k + j − 1)!
(2j − 1)! √
4. Use formula (∗) together with Γ(j + (1/2)) = π. to establish
22j−1 (j − 1)!
∫ π/2
(2j − 1)!(k − 1)!(k + j − 1)!
cos2j θ cos2k−1 θdθ =
0 22j−1 (j − 1)!(2k + 2j − 1)!
(Hint: Use x = j + (1/2) and y = k in formula (∗).)
5. Use the table of values of the Gamma function given in exercise 1 to find an
approximation of the integral
∫ π/2
cosπ θ sine θdθ
0

Exercise 3. The Psi function is defined as the logarithmic derivative of Γ:

Γ′ (x)
Ψ(x) =
Γ(x)
Use the fundamental property of Γ to show that Ψ satisfies
1
Ψ(x + 1) = Ψ(x) + .
x

Exercise 4. Write the first five terms of the series representation of J0 ; J1 ; J2 ;

J−3 ; J3/4 ; J1/5 .
 BESSEL EQUATIONS AND BESSEL FUNCTIONS 17

Exercise 5. Use the series expansion of J−1/2 to establish

√
2
J−1/2 (x) = cos x .
πx
You can also
√ establish this formula by using property (5) with α = 1/2 and
2
J1/2 (x) = sin x
πx
Exercise 6. Repeat the steps of example 1 to show that
√
2 ( cos x )
J−3/2 (x) = − + sin x .
πx x

Exercise 7. Find the expressions of J5/2 and of J−5/2 .

Exercise 8. Use the table of values of J0 and J1 to find the following values

J2 (.5), J3 (5), J4 (8.5)

∫ x
Exercise 9. Prove that sJ0 (s)ds = xJ1 (x).
0
Exercise 10. Find the integrals
∫ ∫ ∫
x9 J8 (x)dx, x−3/2 J5/2 (x)dx, x5 J2 (x)dx

Exercise 11. Find the integrals

∫ ∫ ∫
x2−α Jα+1 (x)dx, J1 (x)dx, J2 (x)dx

Exercise 12. Find the integrals

∫ ∫ x
[J3 (x) − J5 (x)] dx, s4 J1 (s)ds
0

Exercise 13. Show that

∫ R
Rα
xα Jα−1 (λx)dx = Jα (λR)
0 λ

Exercise 14. Show that

x2 Jα′′ (x) − (α2 − α − x2 )Jα (x) − xJα+1 (x) = 0

(Hint: Use Bessel’s equation and property 4)

Exercise 15. Show that
∫ x
J1 (x)
J3 (s)ds = 1 − J2 (x) − 2
0 x
(Hint: Start with J3 (s) = s2 (s−2 J3 (s)) and use integration by parts)
18 BESSEL EQUATIONS AND BESSEL FUNCTIONS

Exercise 16. Use the expansion of cos(x sin θ) involved in the proof of Proposition
3 to show that
∞
∑
cos x = J0 (x) + 2 (−1)j J2j (x)
j=1
∞
∑
sin x = 2 (−1)j J2j+1 (x)
j=0
∞
∑
1 = J0 (x) + 2 J2j (x)
j=1

Exercise 17. Use the integral representation of Jn (x) to show that

∫
′ 1 π
Jn (x) = sin(nθ − x sin(θ)) sin θ dθ
π 0