August 7, 2012 21:04 c05 Sheet number 1 Page number 247 cyan black

CHAPTER

Series Solutions of
Second Order
Linear Equations

Finding the general solution of a linear differential equation depends on determin-

ing a fundamental set of solutions of the homogeneous equation. So far,we have given
a systematic procedure for constructing fundamental solutions only if the equation
has constant coefficients. To deal with the much larger class of equations that have
variable coefficients, it is necessary to extend our search for solutions beyond the
familiar elementary functions of calculus. The principal tool that we need is the rep-
resentation of a given function by a power series. The basic idea is similar to that in
the method of undetermined coefficients: we assume that the solutions of a given dif-
ferential equation have power series expansions, and then we attempt to determine
the coefficients so as to satisfy the differential equation.

5.1 Review of Power Series

In this chapter we discuss the use of power series to construct fundamental sets of
solutions of second order linear differential equations whose coefficients are func-
tions of the independent variable. We begin by summarizing very briefly the pertinent
results about power series that we need. Readers who are familiar with power series
may go on to Section 5.2. Those who need more details than are presented here
should consult a book on calculus.
∞
!
1. A power series an (x − x0 )n is said to converge at a point x if
n=0
m
"
lim an (x − x0 )n
m→∞
n=0

247
August 7, 2012 21:04 c05 Sheet number 2 Page number 248 cyan black

248 Chapter 5. Series Solutions of Second Order Linear Equations

exists for that x. The series certainly converges for x = x0 ; it may converge for all x, or it
may converge for some values of x and not for others.
!∞
2. The series an (x − x0 )n is said to converge absolutely at a point x if the series
n=0

∞
" ∞
"
|an (x − x0 )n | = |an ||x − x0 |n
n=0 n=0

converges. It can be shown that if the series converges absolutely, then the series
also converges; however, the converse is not necessarily true.
3. One of the most useful tests for the absolute convergence of a power series is the ratio
test. If an ̸ = 0, and if, for a fixed value of x,
# # # #
# an+1 (x − x0 )n+1 #
# = |x − x0 | lim # an+1 # = |x − x0 |L,
# #
lim ## # #
n→∞ a (x − x )
n
n
0 n→∞ a # n

then the power series converges absolutely at that value of x if |x − x0 |L < 1 and diverges
if |x − x0 |L > 1. If |x − x0 |L = 1, the test is inconclusive.

For which values of x does the power series

EXAMPLE
1
∞
"
(−1)n+1 n(x − 2)n
n=1

converge?
To test for convergence, we use the ratio test. We have
# #
# (−1)n+2 (n + 1)(x − 2)n+1 #
lim ## # = |x − 2| lim n + 1 = |x − 2|.
n→∞ (−1)n+1 n(x − 2)n # n→∞ n

According to statement 3, the series converges absolutely for |x − 2| < 1, or 1 < x < 3, and
diverges for |x − 2| > 1. The values of x corresponding to |x − 2| = 1 are x = 1 and x = 3. The
series diverges for each of these values of x since the nth term of the series does not approach
zero as n → ∞.

∞
!
4. If the power series an (x − x0 )n converges at x = x1 , it converges absolutely for
n=0
|x − x0 | < |x1 − x0 |; and if it diverges at x = x1 , it diverges for |x − x0 | > |x1 − x0 |.
5. For a typical power series, such as the one in Example 1, there is a positive number ρ, called
!∞
the radius of convergence, such that an (x − x0 )n converges absolutely for |x − x0 | < ρ
n=0
and diverges for |x − x0 | > ρ. The interval |x − x0 | < ρ is called the interval of convergence;
it is indicated by the hatched lines in Figure 5.1.1. The series may either converge or diverge
when |x − x0 | = ρ. Many important power series converge for all values of x. In this case
it is customary to say that ρ is infinite and the interval of convergence is the entire real
line. It is also possible for a power series to converge only at x0 . For such a series we say
that ρ = 0 and the series has no interval of convergence. When these exceptional cases are
taken into account, every power series has a nonnegative radius of convergence ρ, and if
ρ > 0, then there is a (finite or infinite) interval of convergence centered at x0 .
August 7, 2012 21:04 c05 Sheet number 3 Page number 249 cyan black

5.1 Review of Power Series 249

Series
Series Series
converges
diverges diverges
absolutely
x0 – ρ x0 x0 + ρ x

Series may
converge or diverge
FIGURE 5.1.1 The interval of convergence of a power series.

Determine the radius of convergence of the power series

EXAMPLE ∞
" (x + 1)n
2 n2n
.
n=1

We apply the ratio test:

# #
# (x + 1)n+1 n2n ## |x + 1| n |x + 1|
lim ## = lim = .
n→∞ (n + 1)2n+1 (x + 1)n # 2 n→∞ n+1 2
Thus the series converges absolutely for |x + 1| < 2, or −3 < x < 1, and diverges for
|x + 1| > 2. The radius of convergence of the power series is ρ = 2. Finally, we check the
endpoints of the interval of convergence. At x = 1 the series becomes the harmonic series
"∞
1
,
n=1
n

which diverges. At x = −3 we have

∞
" "∞
(−3 + 1)n (−1)n
= ,
n=1
n2n n=1
n

which converges but does not converge absolutely. The series is said to converge condition-
ally at x = −3. To summarize, the given power series converges for −3 ≤ x < 1 and diverges
otherwise. It converges absolutely for −3 < x < 1 and has a radius of convergence 2.

∞
! ∞
!
Suppose that an (x − x0 )n and bn (x − x0 )n converge to f (x) and g(x), respec-
n=0 n=0
tively, for |x − x0 | < ρ, ρ > 0.
6. The two series can be added or subtracted termwise, and
∞
"
f (x) ± g(x) = (an ± bn )(x − x0 )n ;
n=0

the resulting series converges at least for |x − x0 | < ρ.

7. The two series can be formally multiplied, and
$∞ %$ ∞ % ∞
" " "
n n
f (x)g(x) = an (x − x0 ) bn (x − x0 ) = cn (x − x0 )n ,
n=0 n=0 n=0

where cn = a0 bn + a1 bn−1 + · · · + an b0 . The resulting series converges at least for

|x − x0 | < ρ.
August 7, 2012 21:04 c05 Sheet number 4 Page number 250 cyan black

250 Chapter 5. Series Solutions of Second Order Linear Equations

Further, if g(x0 ) ̸ = 0, the series for f (x) can be formally divided by the series for g(x), and
∞
f (x) "
= dn (x − x0 )n .
g(x) n=0

In most cases the coefficients dn can be most easily obtained by equating coefficients in
the equivalent relation
∞
$∞ %$ ∞ %
" " "
n n n
an (x − x0 ) = dn (x − x0 ) bn (x − x0 )
n=0 n=0 n=0
∞
& n '
" "
= dk bn−k (x − x0 )n .
n=0 k=0

In the case of division, the radius of convergence of the resulting power series may be less
than ρ.
8. The function f is continuous and has derivatives of all orders for |x − x0 | < ρ. Moreover,
f ′ , f ′′ , . . . can be computed by differentiating the series termwise; that is,

f ′ (x) = a1 + 2a2 (x − x0 ) + · · · + nan (x − x0 )n−1 + · · ·

∞
"
= nan (x − x0 )n−1 ,
n=1

f (x) = 2a2 + 6a3 (x − x0 ) + · · · + n(n − 1)an (x − x0 )n−2 + · · ·

′′

∞
"
= n(n − 1)an (x − x0 )n−2 ,
n=2

and so forth, and each of the series converges absolutely for |x − x0 | < ρ.
9. The value of an is given by
f (n) (x0 )
an = .
n!
The series is called the Taylor1 series for the function f about x = x0 .
∞
! ∞
!
10. If an (x − x0 )n = bn (x − x0 )n for each x in some open interval with center x0 , then
n=0 n=0
∞
!
an = bn for n = 0, 1, 2, 3, . . . . In particular, if an (x − x0 )n = 0 for each such x, then
n=0
a0 = a1 = · · · = an = · · · = 0.
A function f that has a Taylor series expansion about x = x0
∞
" f (n) (x0 )
f (x) = (x − x0 )n ,
n!
n=0

1 Brook Taylor (1685–1731), English mathematician, received his education at Cambridge University. His
book Methodus incrementorum directa et inversa, published in 1715, includes a general statement of the
expansion theorem that is named for him.This is a basic result in all branches of analysis,but its fundamental
importance was not recognized until 1772 (by Lagrange). Taylor was also the first to use integration by
parts, was one of the founders of the calculus of finite differences, and was the first to recognize the
existence of singular solutions of differential equations.
August 7, 2012 21:04 c05 Sheet number 5 Page number 251 cyan black

5.1 Review of Power Series 251

with a radius of convergence ρ > 0, is said to be analytic at x = x0 . All of the familiar

functions of calculus are analytic except perhaps at certain easily recognized points.
For example, sin x and ex are analytic everywhere, 1/x is analytic except at x = 0, and
tan x is analytic except at odd multiples of π/2. According to statements 6 and 7, if f
and g are analytic at x0 , then f ± g, f · g, and f /g [provided that g(x0 ) ̸ = 0] are also
analytic at x = x0 . In many respects the natural context for the use of power series
is the complex plane. The methods and results of this chapter nearly always can be
directly extended to differential equations in which the independent and dependent
variables are complex-valued.

Shift of Index of Summation. The index of summation in an infinite series is a dummy

parameter just as the integration variable in a definite integral is a dummy variable.
Thus it is immaterial which letter is used for the index of summation. For example,
∞
" ∞
"
2n xn 2j xj
= .
n! j!
n=0 j=0

Just as we make changes of the variable of integration in a definite integral, we find

it convenient to make changes of summation indices in calculating series solutions of
differential equations. We illustrate by several examples how to shift the summation
index.

∞
!
Write an xn as a series whose first term corresponds to n = 0 rather than n = 2.
n=2
EXAMPLE
Let m = n − 2; then n = m + 2, and n = 2 corresponds to m = 0. Hence
3 ∞
" ∞
"
an xn = am+2 xm+2 . (1)
n=2 m=0

By writing out the first few terms of each of these series, you can verify that they contain
precisely the same terms. Finally, in the series on the right side of Eq. (1), we can replace the
dummy index m by n, obtaining
∞
" ∞
"
an xn = an+2 xn+2 . (2)
n=2 n=0

In effect, we have shifted the index upward by 2 and have compensated by starting to count
at a level 2 lower than originally.

Write the series ∞

EXAMPLE "
(n + 2)(n + 1)an (x − x0 )n−2 (3)
4 n=2

as a series whose generic term involves (x − x0 )n rather than (x − x0 )n−2 .

Again, we shift the index by 2 so that n is replaced by n + 2 and start counting 2 lower. We
obtain ∞
"
(n + 4)(n + 3)an+2 (x − x0 )n . (4)
n=0

You can readily verify that the terms in the series (3) and (4) are exactly the same.
August 7, 2012 21:04 c05 Sheet number 6 Page number 252 cyan black

252 Chapter 5. Series Solutions of Second Order Linear Equations

Write the expression

EXAMPLE ∞
"
x2 (r + n)an xr+n−1
5 n=0
(5)

as a series whose generic term involves xr+n .

First, take the x2 inside the summation, obtaining
∞
"
(r + n)an xr+n+1 . (6)
n=0

Next, shift the index down by 1 and start counting 1 higher. Thus
∞
" ∞
"
(r + n)an xr+n+1 = (r + n − 1)an−1 xr+n . (7)
n=0 n=1

Again, you can easily verify that the two series in Eq. (7) are identical and that both are exactly
the same as the expression (5).

Assume that
∞
" ∞
"
EXAMPLE
nan xn−1 = an xn (8)
6 n=1 n=0

for all x, and determine what this implies about the coefficients an .
We want to use statement 10 to equate corresponding coefficients in the two series. In order
to do this, we must first rewrite Eq. (8) so that the series display the same power of x in their
generic terms. For instance, in the series on the left side of Eq. (8), we can replace n by n + 1
and start counting 1 lower. Thus Eq. (8) becomes
∞
" ∞
"
(n + 1)an+1 xn = an xn . (9)
n=0 n=0

According to statement 10, we conclude that

(n + 1)an+1 = an , n = 0, 1, 2, 3, . . .
or
an
an+1 = , n = 0, 1, 2, 3, . . . . (10)
n+1
Hence, choosing successive values of n in Eq. (10), we have
a1 a0 a2 a0
a1 = a0 , a2 = = , a3 = = ,
2 2 3 3!
and so forth. In general,
a0
an = , n = 1, 2, 3, . . . . (11)
n!
Thus the relation (8) determines all the following coefficients in terms of a0 . Finally, using the
coefficients given by Eq. (11), we obtain
∞
" ∞
" xn
an xn = a0 = a0 ex ,
n=0 n=0
n!

where we have followed the usual convention that 0! = 1.

August 7, 2012 21:04 c05 Sheet number 7 Page number 253 cyan black

5.1 Review of Power Series 253

PROBLEMS In each of Problems 1 through 8, determine the radius of convergence of the given power
series.
∞
! !∞ n
1. (x − 3)n 2. xn
n=0 2
n
n=0

!∞ x2n ∞
!
3. 4. 2n xn
n=0 n! n=0

!∞ (2x + 1)n !∞ (x − x )n
0
5. 6.
n=1 n2 n=1 n
!∞ (−1)n n2 (x + 2)n !∞ n!xn
7. 8.
n=1 3n n=1 n
n

In each of Problems 9 through 16, determine the Taylor series about the point x0 for the given
function. Also determine the radius of convergence of the series.
9. sin x, x0 = 0 10. ex , x0 = 0
2
11. x, x0 = 1 12. x , x0 = −1
1
13. ln x, x0 = 1 14. , x0 = 0
1+x
1 1
15. , x0 = 0 16. , x0 = 2
1−x 1−x
∞
!
17. Given that y = nxn , compute y′ and y′′ and write out the first four terms of each series,
n=0
as well as the coefficient of xn in the general term.
∞
!
18. Given that y = an xn , compute y′ and y′′ and write out the first four terms of each
n=0
series, as well as the coefficient of xn in the general term. Show that if y′′ = y, then the
coefficients a0 and a1 are arbitrary, and determine a2 and a3 in terms of a0 and a1 . Show
that an+2 = an /(n + 2)(n + 1), n = 0, 1, 2, 3, . . . .

In each of Problems 19 and 20, verify the given equation.

∞
! ∞
!
19. an (x − 1)n+1 = an−1 (x − 1)n
n=0 n=1
∞
! ∞
! ∞
!
20. ak+1 xk + ak xk+1 = a1 + (ak+1 + ak−1 )xk
k=0 k=0 k=1

In each of Problems 21 through 27, rewrite the given expression as a sum whose generic term
involves xn .
∞
! ∞
!
21. n(n − 1)an xn−2 22. an xn+2
n=2 n=0
∞
! ∞
! ∞
!
23. x nan xn−1 + ak xk 24. (1 − x2 ) n(n − 1)an xn−2
n=1 k=0 n=2
∞
! ∞
! ∞
! ∞
!
25. m(m − 1)am xm−2 + x kak xk−1 26. nan xn−1 + x an xn
m=2 k=1 n=1 n=0
∞
! ∞
!
27. x n(n − 1)an xn−2 + an xn
n=2 n=0
August 7, 2012 21:04 c05 Sheet number 8 Page number 254 cyan black

254 Chapter 5. Series Solutions of Second Order Linear Equations

28. Determine the an so that the equation

∞
" ∞
"
nan xn−1 + 2 an xn = 0
n=1 n=0

∞
!
is satisfied. Try to identify the function represented by the series an xn .
n=0

5.2 Series Solutions Near an Ordinary Point, Part I

In Chapter 3 we described methods of solving second order linear differential equa-
tions with constant coefficients. We now consider methods of solving second order
linear equations when the coefficients are functions of the independent variable. In
this chapter we will denote the independent variable by x. It is sufficient to consider
the homogeneous equation
d2 y dy
P(x)
2
+ Q(x) + R(x)y = 0, (1)
dx dx
since the procedure for the corresponding nonhomogeneous equation is similar.
Many problems in mathematical physics lead to equations of the form (1) having
polynomial coefficients; examples include the Bessel equation
x2 y′′ + xy′ + (x2 − ν2 )y = 0,
where ν is a constant, and the Legendre equation
(1 − x2 )y′′ − 2xy′ + α(α + 1)y = 0,
where α is a constant. For this reason,as well as to simplify the algebraic computations,
we primarily consider the case in which the functions P, Q, and R are polynomials.
However, as we will see, the method of solution is also applicable when P, Q, and R
are general analytic functions.
For the present, then, suppose that P, Q, and R are polynomials and that there is
no factor (x − c) that is common to all three of them. If there is such a factor (x − c),
then divide it out before proceeding. Suppose also that we wish to solve Eq. (1) in
the neighborhood of a point x0 . The solution of Eq. (1) in an interval containing x0
is closely associated with the behavior of P in that interval.
A point x0 such that P(x0 ) ̸= 0 is called an ordinary point. Since P is continuous, it
follows that there is an interval about x0 in which P(x) is never zero. In that interval
we can divide Eq. (1) by P(x) to obtain
y′′ + p(x)y′ + q(x)y = 0, (2)
where p(x) = Q(x)/P(x) and q(x) = R(x)/P(x) are continuous functions. Hence,
according to the existence and uniqueness Theorem 3.2.1, there exists in that inter-
val a unique solution of Eq. (1) that also satisfies the initial conditions y(x0 ) = y0 ,
y′ (x0 ) = y′0 for arbitrary values of y0 and y′0 . In this and the following section, we
discuss the solution of Eq. (1) in the neighborhood of an ordinary point.
On the other hand, if P(x0 ) = 0, then x0 is called a singular point of Eq. (1). In
this case at least one of Q(x0 ) and R(x0 ) is not zero. Consequently, at least one of