MATH0042: Mathematical Methods in Chemistry AY 2019–2020

Problem Sheet 1
Lecturer: Matteo Capoferri† Due: 2 PM on 11 October 2019

Solutions coming in multiple pages must be stapled. Please make sure you write name and/or student ID on your
script.
Assessed questions: 1, 2 and 4. The bonus question gives extra marks if answered correctly.

Question 1
Consider a second order ODE with constant coefficients of the form

a y 00 (x) + b y 0 (x) + c y(x) = 0, a 6= 0. (1)

(a) Show that y(x) = Ceλ x is a solution for (1) if and only if aλ2 + bλ + c = 0.
(b) Using Euler’s formula, show that a solution to (1) of the form

y(x) = Aeλ1 x + Beλ2 x (2)

with λ1 = λ2 = α + i β, β 6= 0, can be equivalently recast as

h i
y(x) = eαx à cos(βx) + B̃ sin(βx) . (3)

Write down à and B̃ in terms of A and B.

Solution: (a) Substituting y(x) = Ceλx into (1) one obtains

aλ2 Ceλx + bλCeλx + cCeλx = 0. (4)

We observe that Ceλx is always nonzero. So we can safely divide both sides of (4) by it, thus obtaining

aλ2 + bλ + c = 0.

(b) As λ1 = α + iβ, Euler’s formula tells us that

eλ1 x = eαx+iβx = eαx [cos(βx) + i sin(βx)]. (5)

Similarly, as λ2 = α − iβ, Euler’s formula tells us that

eλ2 x = eαx−iβx = eαx [cos(βx) − i sin(βx)]. (6)

Substituting (5) and (6) into (2) we get

y(x) = eα(x) [(A + B) cos(β(x)) + i(A − B) sin(βx)] ,

which corresponds to (3) with

à = A + B, B̃ = i(A − B).

†
[email protected], http://www.ucl.ac.uk/~ucahmca/.

PLEASE TURN OVER −→

1
Question 2
Solve the first order ordinary differential equation

(x + 3) y 0 (x) + (x + 2) y(x) = 0

using power series methods. Check that the resulting power series can be written as

y(x) = A (x + 3) e−x ,

where A is a constant. You may take into account terms up to x5 only in the series solution.

Solution: We seek a power series solution about x = 0,

∞
X
y(x) = an xn = a0 + a1 x + a2 x2 + a3 x3 + . . .
n=0

The derivative of y is obtained by differentiating the series term by term:

∞
X
0
y (x) = n an xn−1 = a1 + 2a2 x + 3 a3 x2 + . . .
n=1

Substituting y and y 0 into the ODE gives

∞
X ∞
X
(x + 3) n an xn−1 + (x + 2) an xn = 0,
n=1 n=0

∞
X ∞
X ∞
X ∞
X
n an xn + 3 n an xn−1 + an xn+1 + 2 an xn = 0,
n=1 n=1 n=0 n=0
∞
X ∞
X ∞
X ∞
X
n n n
n an x + 3 (n + 1) an+1 x + an−1 x + 2 an xn = 0.
n=0 n=0 n=1 n=0

Summing up the coefficients of xn and equating the result to zero we obtain

2
n=0 : 3 a1 + 2 a0 = 0 ⇒ a1 = − a0 ,
3
1
n=1 : a1 + 6a2 + a0 + 2a1 = 0 ⇒ a2 = a0 ,
6
n=2 : 2a2 + 9a3 + a1 + 2a2 = 0 ⇒ a3 = 0,
1
n=3 : 3a3 + 12a4 + a2 + 2a3 = 0 ⇒ a4 = − a0 ,
72
1
n=4 : 4a4 + 15a5 + a3 + 2a4 = 0 ⇒ a5 = a0 .
180
So  
2 1 2 1 4 1 5
y(x) = a0 1 − x + x − x + x + ... .
3 6 72 180

Now, using the Maclaurin expansion of the exponential function, we obtain that
 
1 1 1 1
A(x + 3)e−x = A(x + 3) 1 − x + x2 − x3 + x4 − x5 + . . .
2! 3! 4! 5!
    
3 1 1
= A 3 + (1 − 3)x + − 1 x2 + − x3
2 2 2
    
1 1 1 1
+ − x4 + − x5 + . . .
8 6 24 40

2
  
1 2 1 4 1 5
= A 3 − 2x + x − x + x + . . .
2 24 60
 
2 1 2 1 4 1 5
= 3A 1 − x + x − x + x + ... .
3 6 72 180

If we let 3A = a0 we end up exactly with our series solution, as required.

Question 3
Solve the first order ordinary differential equation

y 0 (x) − 2x y(x) = 0

using power series methods. Verify that the resulting power series solution can be written in the form
2
y(x) = A ex ,

where A is a constant. You may take into account terms up to x6 only in the power series solution.

Solution: We look for a power series solution about x = 0,

∞
X
y(x) = an xn .
n=0

The derivative of y is obtained by differentiating the series term by term:

∞
X
y 0 (x) = n an xn−1 = a1 + 2a2 x + 3 a3 x2 + . . . .
n=1

Substituting y and y 0 into our ODE we obtain

∞
X ∞
X
nan xn−1 − 2an xn+1 = 0,
n=0 n=0

and, after shifting the summation indices so that x appears in both series with the same power, we arrive
at
X∞ X∞
(n + 1)an+1 xn − 2an−1 xn = 0,
n=0 n=1
∞
X
a1 + [(n + 1)an+1 − 2an−1 ] xn = 0.
n=1

Hence we must have

a1 = 0

and
2
(n + 1)an+1 − 2an−1 = 0 ⇔ an+1 = an−1 , for n ≥ 1.
n+1
Since a1 is vanishing, the recurrence relation tells us that all an for n odd vanish as well. As far as even
coefficients are concerned, we can use the recurrence relation to get

2
a2 = a0 = a0 ,
2

PLEASE TURN OVER −→

3
 2 1 1
a4 = a2 = a2 = a0 ,
4 2 2!
2 1 1
a6 = a4 = a4 = a0 .
6 3 3!
So we have  
1 4
2 1 6
y(x) = a0 1+x + x + x .
2! 3!

Now, using the Maclaurin expansion of the exponential function, we obtain that

(x2 )2 (x2 )3
 
2
Aex = A 1 + (x2 ) + + + ...
2! 3!
4 6
 
x x
= A 1 + x2 + + + ... .
2! 3!

If we let A = a0 we end up exactly with our series solution, as required.

Question 4
Consider the Chebyshev’s equation

(1 − x2 ) y 00 (x) − x y 0 (x) + p2 y(x) = 0, −1 < x < 1,

where p is a real constant.

P∞
(a) Seeking a solution in the form y(x) = n=0 an xn , write down the recurrence relation relating the
coefficient an+2 to an .
(b) Write down the first three nonzero terms in each of the linearly independent series solutions of the
differential equation.
(c) Let m be a positive integer. Use the recurrence relation found in (a) to show that if p = m the
differential equation has a polynomial solution.
(d) For the case p = m, use the substitution x = cos θ to show that the Chebyshev’s equation transforms
into a second-order ODE with constant coefficients. Find the general solution of this ODE.
(e) (Bonus) The Chebyshev polynomials Tm (x) are defined to be the even solutions of the above ODE such
that Tm (1) = 1. Deduce from (d) that

Tm (x) = cos(m arccos(x)).

Solution: (a) Substituting

∞
X
y(x) = an xn
n=0

into Chebyshev’s equation gives

∞
X ∞
X ∞
X
(1 − x2 ) an n(n − 1)xn−2 − x an nxn−1 + b2 an xn = 0,
n=2 n=1 n=0

∞
X ∞
X ∞
X ∞
X
an n(n − 1)xn−2 − an n(n − 1)xn − an nxn + p2 an xn = 0,
n=2 n=2 n=1 n=0
∞
X ∞
X ∞
X ∞
X
(n + 2)(n + 1)an+2 xn − an n(n − 1)xn − an nxn + p2 an xn = 0,
n=0 n=0 n=0 n=0

4
 ∞
X
(n + 2)(n + 1) n + 2 − (n2 − p2 ) an xn = 0.
 
n=0

Therefore, by equating to zero the coefficient of xn we obtain that the recursion relation for the coefficients
is
n2 − p2
an+2 = an , for all n ≥ 0.
(n + 2)(n + 1)

(b) Since the recurrence relation connects coefficients whose indices differ by 2, the two linearly indepen-
dent solutions to Chebyshev’s equation are give by

y0 (x) = a0 + a2 x2 + a4 x4 + . . . ,
y1 (x) = a1 x + a3 x3 + a5 x5 + . . . .

We use the recurrence relation found in (a) to compute the coefficients of y0 and y1 .

• For n even:
p2
a2 = − a0 ,
2!
(22 − p2 )(−p2 ) p2 (p2 − 4)
a4 = a0 = .
4! 4!
Therefore we get
p2 p2 (p2 − 4) 4
 
y0 (x) = a0 1 − x2 + x + ... .
2! 4!
• For n odd:
1 − p2 p2 − 1
a3 = a1 = − a1 ,
3! 3!
(32 − p2 )(1 − p2 ) (p2 − 9)(p2 − 1)
a5 = a1 = a1 .
5! 5!
Therefore we get

p2 − 1 3 (p2 − 9)(p2 − 1) 5
 
y1 (x) = a1 x − x + x + ... .
3! 5!

(c) When p = m, the recurrence relation established in (a) tells us that

m2 − m2
am+2 = = 0.
(m + 2)(m + 1)

Consequently, am+4 = am+6 = . . . = 0. Thus, when p is an integer, one of the two linearly independent
solutions becomes a polynomial. More precisely, for p = m, y0 (x) is a polynomial if m is even, y1 (x) is
a polynomial if m is odd.
(d) When p = m, Chebyshev’s equation is

d2 y dy
(1 − x2 ) −x + m2 y = 0.
dx2 dx
Substituting x = cos θ, we get
dy dy dθ 1 dy
= = .
dx dθ dx − sin θ dθ

d2 y dθ d2 y
 
1 d −1 dy
= =−
dx2 dx dθdx sin θ dθ sin θ dθ
1 d2 y 1 d2 y
      
1 d 1 dy 1 − cos θ dy
= + = +
sin θ dθ sin θ dθ sin θ θ2 sin θ sin2 θ dθ sin θ θ2

PLEASE TURN OVER −→

5
 dy d2 y
  
1 − cos θ
= + 2
sin2 θ sin θ dθ θ

Substituting the above relations into Chebyshev’s euqation, we obtain

dy d2 y
    
2 1 − cos θ 1 dy
(1 − cos θ) 2 + 2 − cos θ + m2 y = 0,
sin θ sin θ dθ dθ − sin θ dθ

sin2 θ − cos θ dy d2 y
 
cos θ dy
+ + + m2 y = 0,
sin2 θ sin θ dθ dθ2 sin θ dθ
d2 y
+ m2 y = 0.
dθ2
This is a second-order ordinary differential equation with constant coefficients, whose general solution
reads
y(θ) = Am cos(mθ) + Bm sin(mθ),

where Am and Bm are arbitrary constants for each choice of m.

(e) (Bonus) Since Tm (x) are defined to be the even polynomial solutions and sine is odd, when looking
at Chebyshev polynomials the coefficients Bm in the general solution in (d) are all zero. Hence,

Tm (x) = Am cos(mθ) = Am cos(m arccos(x)).

We use the normalisation condition to determine the coefficients Am :

1 = Tm (1) = Am cos(m arccos(1)) = Am cos(m · 0) = Am .

Therefore, we conclude that

Tm (x) = cos(m arccos(x)).