Griffiths Quantum Mechanics 3e: Problem 2.

52 Page 1 of 12

Problem 2.52
Consider the potential
~2 a2
V (x) = −sech2 (ax),
m
where a is a positive constant, and “sech” stands for hyperbolic secant.
(a) Graph this potential.

(b) Check that this potential has the ground state

ψ0 (x) = A sech(ax),

and find its energy. Normalize ψ0 , and sketch its graph.

(c) Show that the function  

ik − a tanh(ax)
ψk (x) = A eikx ,
ik + a
√
(where k ≡ 2mE/~, as usual) solves the Schrödinger equation for any (positive) energy E.
Since tanh z → −1 as z → −∞,

ψk (x) ≈ Aeikx , for large negative x.

This represents, then, a wave coming in from the left with no accompanying reflected wave
(i.e. no term exp(−ikx)). What is the asymptotic form of ψk (x) at large positive x? What
are R and T , for this potential? Comment: This is a famous example of a reflectionless
potential—every incident particle, regardless its energy, passes right through.62

Solution

Part (a)

62
R. E. Crandall and B. R. Litt, Annals of Physics, 146, 458 (1983).

www.stemjock.com
Griffiths Quantum Mechanics 3e: Problem 2.52 Page 2 of 12

Part (b)

The goal here is merely to verify that ψ0 (x) = A sech(ax) satisfies the time-independent
Schrödinger equation (TISE),

~2 d2 ψ ~2 a2
− − sech2 (ax)ψ(x) = Eψ(x),
2m dx2 m
and to determine the corresponding energy E. Evaluate the left side with ψ = ψ0 (x).

~2 d2 ψ0 ~2 a2 2 ~2 d2 ~2 a2
− − sech (ax)ψ0 (x) = − [A sech(ax)] − sech2 (ax)[A sech(ax)]
2m dx2 m 2m dx2 m
~2 d ~2 a2
=− [−Aa sech(ax) tanh(ax)] − A sech3 (ax)
2m dx m
~2 ~2 a2
=− A[−a2 sech3 (ax) + a2 sech(ax) tanh2 (ax)] − A sech3 (ax)
2m m
~2 a2 ~2 a2 ~2 a2
=A sech3 (ax) − A sech(ax) tanh2 (ax) − A sech3 (ax)
2m 2m m
~2 a2 ~2 a2
= −A sech3 (ax) − A sech(ax) tanh2 (ax)
2m 2m
~2 a2 ~2 a2
= −A sech3 (ax) − A sech(ax)[1 − sech2 (ax)]
2m 2m
~2 a2 3
 ~2 a2 ~2 a2 3

= −A sech (ax) − A sech(ax) + A sech (ax)
 2m 2m  2m

~2 a2
= −A sech(ax)
2m
~2 a2
=− [A sech(ax)]
2m
~2 a2
=− ψ0 (x)
2m
Because the left side yields a constant times ψ0 (x), ψ0 (x) is a solution to the TISE. The constant
in front is the energy associated with ψ0 (x).

~2 a2
E0 = −
2m
Normalize ψ0 (x) now.
ˆ ∞ ˆ ∞ ˆ ∞ ∞
dz A2 A2
1= |ψ0 (x)|2 dx = A2 sech2 (ax) dx = A2 sech2 z = (tanh z) = (2)
−∞ −∞ −∞ a a −∞ a

Solve for A. r
a
A=
2

www.stemjock.com
Griffiths Quantum Mechanics 3e: Problem 2.52 Page 3 of 12

Therefore, r
a
ψ0 (x) = sech(ax).
2

Part (c)

The goal here is merely to verify that

     
ik − a tanh(ax) ikx ik ikx a
ψk (x) = A e =A e −A tanh(ax)eikx
ik + a ik + a ik + a

satisfies the time-independent Schrödinger equation (TISE),

~2 d2 ψ ~2 a2
− − sech2 (ax)ψ(x) = Eψ(x).
2m dx2 m
Evaluate the left side with ψ = ψk (x).

~2 d2 ψk ~2 a2 ~2 d2
     
2 ik ikx a ikx
− − sech (ax)ψk (x) = − A e −A tanh(ax)e
2m dx2 m 2m dx2 ik + a ik + a
~2 a2
     
2 ik ikx a ikx
− sech (ax) A e −A tanh(ax)e
m ik + a ik + a
~2
   2   2 
ik d ikx a d ikx
=− A e −A tanh(ax)e
2m ik + a dx2 ik + a dx2
~2 a2
     
2 ik ikx a ikx
− sech (ax) A e −A tanh(ax)e
m ik + a ik + a

www.stemjock.com
Griffiths Quantum Mechanics 3e: Problem 2.52 Page 4 of 12

Continue the simplification.

~2 d2 ψk ~2 a2 ~2
  
2 ik
− − sech (ax)ψ k (x) = − A (ik)2 eikx
2m dx2 m 2m ik + a
  
a d
−A [a sech2 (ax)eikx + ik tanh(ax)eikx ]
ik + a dx
2 2
     
~ a 2 ik ikx a ikx
− sech (ax) A e −A tanh(ax)e
m ik + a ik + a
~2
    
ik 2 ikx a
=− −A k e −A [−2a2 sech2 (ax) tanh(ax)eikx
2m ik + a ik + a
+ 2ika sech2 (ax)eikx

2 ikx
− k tanh(ax)e ]

~2 a2
     
2 ik ikx a ikx
− sech (ax) A e −A tanh(ax)e
m ik + a ik + a
~2

(((
=A ik 3 eikx − 2a(3
sech 2 (((((
((ax) tanh(ax)e ikx
2m(ik + a) ( ((
hhhh i
+2ika2 sechh2h hhikx
(ax)e h
h − ak 2
tanh(ax)e ikx

~2
 
hhhh (((
2 2h ikx 3 2 ((((( ikx
−A 2ika sech (ax)e
h hhh h − 2a(sech
((( (ax) tanh(ax)e
2m(ik + a) (

~2
=A k 2 [ik − a tanh(ax)]eikx
2m(ik + a)
~2 k 2
   
ik − a tanh(ax) ikx
= A e
2m ik + a
~2 k 2
= ψk (x)
2m
Because the left side yields a constant times ψk (x), ψk (x) is a solution to the TISE. The constant
in front is the energy associated with ψk (x):

~2 k 2 ~2 2mE
 
= = E.
2m 2m ~2

Since tanh(ax) → 1 as x → ∞,
   
ik − a tanh(ax) ik − a
ψk (x) = A eikx ≈ A eikx
ik + a ik + a

for large positive x. Like before, this is a plane wave going from left to right with no
corresponding reflected wave. The reflection and transmission coefficients are defined by

reflected probability current transmitted probability current

R= T = ,
incident probability current incident probability current

www.stemjock.com
Griffiths Quantum Mechanics 3e: Problem 2.52 Page 5 of 12

where the probability current is [noting that Ψ(x, t) = ψ(x)e−iEt/~ ]

∂Ψ∗
 
i~ ∗ ∂Ψ
J(x, t) = Ψ −Ψ
2m ∂x ∂x
 
i~ ∂ ∂
= [ψ(x)e−iEt/~ ] [ψ ∗ (x)eiEt/~ ] − [ψ ∗ (x)eiEt/~ ] [ψ(x)e−iEt/~ ]
2m ∂x ∂x
∗
 
i~ −iEt/~ dψ iEt/~ ∗ iEt/~ dψ −iEt/~
= [ψ(x)e ] e − [ψ (x)e ] e
2m dx dx
∗
 
i~ dψ dψ
= ψ − ψ∗ .
2m dx dx

Therefore,

reflected probability current

R=
incident probability current

0
= nh   i h   i h   i h   io
i~
A ik−a
e ikx d A −ik−a e−ikx − A −ik−a e−ikx d A ik−a eikx
2m ik+a dx −ik+a −ik+a dx ik+a

=0

and
transmitted probability current
T =
incident probability current
nh   i h   i h   i h   io
i~
A ik−a
e ikx d A −ik−a e−ikx − A −ik−a e−ikx d A ik−a eikx
2m ik+a dx −ik+a −ik+a dx ik+a
= nh   i h   i h   i h   io
i~ ik−a ikx d −ik−a −ikx − A −ik−a −ikx d ik−a ikx
2m A ik+a e dx A −ik+a e −ik+a e dx A ik+a e

= 1.

www.stemjock.com
Griffiths Quantum Mechanics 3e: Problem 2.52 Page 6 of 12

Part (d)

The Schrödinger equation will be solved here on the whole line with V (x) = −(~2 a2 /m) sech2 (ax)

∂Ψ ~2 ∂ 2 Ψ ~2 a2
i~ =− − sech2 (ax)Ψ(x, t), −∞ < x < ∞, t > 0
∂t 2m ∂x2 m
and the usual Dirichlet boundary conditions, Ψ → 0 as x → ±∞. Because the Schrödinger
equation is linear and homogeneous, the method of separation of variables can be applied.
Assume a product solution of the form Ψ(x, t) = ψ(x)φ(t) and plug it into the PDE.

∂ ~2 ∂ 2 ~2 a2
i~ [ψ(x)φ(t)] = − [ψ(x)φ(t)] − sech2 (ax)[ψ(x)φ(t)]
∂t 2m ∂x2 m
Evaluate the derivatives.
~2 00 ~2 a2
i~ψ(x)φ0 (t) = − ψ (x)φ(t) − sech2 (ax)ψ(x)φ(t)
2m m
Divide both sides by ψ(x)φ(t) to separate variables.

φ0 (t) ~2 ψ 00 (x) ~2 a2
i~ =− − sech2 (ax)
φ(t) 2m ψ(x) m

The only way a function of t can be equal to a function of x is if both are equal to a constant E.

φ0 (t) ~2 ψ 00 (x) ~2 a2
i~ =− − sech2 (ax) = E
φ(t) 2m ψ(x) m

As a result of applying the method of separation of variables, Schrödinger’s equation has reduced
to two ODEs—one in each of the independent variables, x and t.

φ0 (t)

i~ = E

φ(t) 
~2 ψ 00 (x) ~2 a2
sech2 (ax) = E 

− − 
2m ψ(x) m

Values of E for which the boundary conditions of this second equation are satisfied are known as
eigenvalues (or eigenenergies in this context), and the nontrivial solutions ψ(x) that satisfy this
second equation are known as eigenfunctions (or eigenstates in this context). Multiply both sides
of the TISE by (−2m/~2 )ψ(x).

d2 ψ 2mE
2
+ 2a2 sech2 (ax)ψ(x) = − 2 ψ(x) (1)
dx ~
Make the substitution, z = ax. Use the chain rule to write the second derivative of ψ in terms of
this new variable.
dψ dψ dz dψ
= =a
dx dz dx dz
d2 ψ d2 ψ
   
d dψ dz d dψ
2
= = a = a2 2
dx dx dx dx dz dz dz

www.stemjock.com
Griffiths Quantum Mechanics 3e: Problem 2.52 Page 7 of 12

Consequently,
d2 ψ 2mE
a2 2
+ 2a2 (sech2 z)ψ(z) = − 2 ψ(z).
dz ~
2 2 2
Divide both sides by a , let E represent 2mE/(~ a ), and bring all terms to the left side.
d2 ψ
+ (E + 2 sech2 z)ψ(z) = 0 (2)
dz 2
Now make the substitution, u = sech2 z. Use the chain rule again to write this equation in terms
of u.
dψ dψ du dψ
= = (−2 sech2 z tanh z)
dz du dz du
d2 ψ
   
d dψ d dψ 2
= = (−2 sech z tanh z)
dz 2 dz dz dz du
 
d dψ dψ d
= (−2 sech2 z tanh z) + (−2 sech2 z tanh z)
dz du du dz
 
du d dψ dψ
= (−2 sech2 z tanh z) + (4 sech2 z tanh2 z − 2 sech4 z)
dz du du du
d2 ψ dψ
= (−2 sech2 z tanh z)2 + (4 sech2 z tanh2 z − 2 sech4 z)
du2 du
d2 ψ dψ
= 2
(4 sech4 z tanh2 z) + (4 sech2 z tanh2 z − 2 sech4 z)
du du
d2 ψ dψ
= 2
[4 sech4 z(1 − sech2 z)] + [4 sech2 z(1 − sech2 z) − 2 sech4 z]
du du
d2 ψ 2 dψ
= 2
[4u (1 − u)] + [4u(1 − u) − 2u2 ]
du du
d2 ψ dψ
= 4u2 (1 − u) 2 + 2u(2 − 3u)
du du
As a result, equation (2) becomes
d2 ψ dψ
4u2 (1 − u) 2
+ 2u(2 − 3u) + (E + 2u)ψ(u) = 0.
du du
Check to see if zero is an eigenvalue: E = 0.
d2 ψ dψ
4u2 (1 − u) 2
+ 2u(2 − 3u) + 2uψ(u) = 0
du du
Divide both sides by 4u.
d2 ψ
 
3 dψ 1
u(1 − u) 2 + 1 − u + ψ(u) = 0 (3)
du 2 du 2
This is the hypergeometric equation with a = −1/2 and b = 1 and c = 1, so one of its solutions is
√
 
1
ψ(u) = 2 F1 − , 1; 1; u = 1 − u.
2
Just for reference, the hypergeometric equation is

x(1 − x)y 00 + [c − (a + b + 1)x]y 0 − aby = 0,

www.stemjock.com
Griffiths Quantum Mechanics 3e: Problem 2.52 Page 8 of 12

and one of its two linearly independent solutions is denoted by

∞
Γ(c) X Γ(a + n)Γ(b + n) xn
y(x) = 2 F1 (a, b; c; x) = .
Γ(a)Γ(b) Γ(c + n) n!
n=0

Use the √
method of reduction of order to obtain the general solution: Substitute
ψ(u) = 1 − u w(u) into equation (3).

d2 √ d √ 1√
 
3
u(1 − u) 2 [ 1 − u w(u)] + 1 − u [ 1 − u w(u)] + [ 1 − u w(u)] = 0
du 2 du 2

Simplify the left side.

1√ d2 w
 
dw
− 1 − u 2u(u − 1) 2 + (5u − 2) =0
2 du du
This is a first-order linear ODE for dw/du, which can be solved with an integrating factor.
√
 
1
− tanh−1


w(u) = C1 + C2 √ 1−u
1−u
Consequently, √ √ √
ψ(u) = C1 1 − u + C2 1 − 1 − u tanh−1


1−u .
Changing back to the original variables yields
q  q q 
2 2 −1 2
ψ(x) = C1 1 − sech (ax) + C2 1 − 1 − sech (ax) tanh 1 − sech (ax)

q  q q 
2 2 −1 2
= C1 tanh (ax) + C2 1 − tanh (ax) tanh tanh (ax)

= D1 tanh(ax) + D2 [1 − ax tanh(ax)] .

In order to satisfy the boundary conditions, ψ → 0 as x → ±∞, D1 and D2 must be set to zero.
This leads to the trivial solution, ψ(x) = 0, so zero is not an eigenvalue. Now check to see if there
are negative eigenvalues: E = −γ 2 .

d2 ψ dψ
4u2 (1 − u) + 2u(2 − 3u) + (−γ 2 + 2u)ψ(u) = 0
du2 du
As it is, this equation can’t be solved in terms of hypergeometric functions like before because
−γ 2 is not multiplied by u. Make the substitution, ψ(u) = ur f (u), then. This is motivated from
the u2 factor in front of the second derivative and the u factor in front of the first derivative.
dψ df
= rur−1 f (u) + ur
du du
d2 ψ r−2 r−1 df
2
rd f
= r(r − 1)u f (u) + 2ru + u
du2 du du2
Plug these formulas into the ODE.
2
   
2 r−2 r−1 df rd f r−1 r df
4u (1 − u) r(r − 1)u f (u) + 2ru +u + 2u(2 − 3u) ru f (u) + u + (−γ 2 + 2u)ur f (u) = 0
du du2 du

www.stemjock.com
Griffiths Quantum Mechanics 3e: Problem 2.52 Page 9 of 12

Divide both sides by ur and simplify the left side.

d2 f df
4u2 (1 − u) + (4u + 8ru − 6u2 − 8ru2 ) + (4r2 − γ 2 + 2u − 2ru − 4r2 u)f (u) = 0
du2 du
Choose r so that
γ
4r2 − γ 2 = 0 → r= .
2
As a result,
d2 f df
4u2 (1 − u) 2
+ (4u + 4γu − 6u2 − 4γu2 ) + (2u − γu − γ 2 u)f (u) = 0.
du du
Divide both sides by 4u.
d2 f 1 γ γ2
   
3 df
u(1 − u) 2 + 1 + γ − u − γu + − − f (u) = 0 (4)
du 2 du 2 4 4
This ODE for f is a hypergeometric equation with a = −1/2 + γ/2, b = 1 + γ/2, and c = 1 + γ.
That means one solution is
 
1 γ γ
f (u) = 2 F1 − + , 1 + ; 1 + γ; u .
2 2 2
Use the wag-the-dog method to determine appropriate values of γ: Plot f (u) versus u for several
values of γ and see where the graph flips over.

The graph only flips over once at γ = 1. Note that negative values of γ are not considered because
u = sech2 z, and ψ(u) = uγ/2 f (u) would blow up as a result. With γ = 1, equation (4) becomes

d2 f
 
5 df
u(1 − u) 2 + 2 − u = 0,
du 2 du
which is a first-order linear ODE for df /du, solvable with an integrating factor. The general
solution is √
√

1−u −1


f (u) = C3 + C4 + tanh 1−u ,
u

www.stemjock.com
Griffiths Quantum Mechanics 3e: Problem 2.52 Page 10 of 12

which means
"r #
√ 1−u √ √
ψ(u) = u1/2 f (u) = C3 u + C4 + u tanh−1


1−u .
u

Changing back to the original variables yields

"s #
2 q
1 − sech (ax)
q q
ψ(x) = C3 sech2 (ax) + C4 + sech2 (ax) tanh−1 1 − sech2 (ax)
sech2 (ax)
"s #
2 q
tanh (ax)
q q
= C3 sech2 (ax) + C4 + sech2 (ax) tanh−1 tanh2 (ax)
sech2 (ax)

= D3 sech(ax) + D4 [sinh(ax) + ax sech(ax)].

In order to satisfy the boundary conditions, ψ → 0 as x → ±∞, D4 must be set to zero.

Normalization determines D3 as shown in part (b).
r
a
ψ(x) = sech(ax)
2
Remarkably, this is the only bound state. The corresponding energy is

2mE ~2 a2
E= = −1 → E=− .
~2 a2 2m

Solving the ODE in t yields φ(t) = e−iEt/~ . Therefore, one solution to Schrödinger’s equation is

~a2
r  
a
Ψ(x, t) = φ(x)φ(t) = sech(ax) exp i t .
2 2m

Return to equation (1), now assuming that E is positive.

d2 ψ 2mE
2
+ 2a2 sech2 (ax)ψ(x) = − 2 ψ(x) (1)
dx ~
√
Use k for 2mE/~ and bring all terms to the left side.

d2 ψ
+ [k 2 + 2a2 sech2 (ax)]ψ(x) = 0
dx2
Observe from the graph in part (a) that the potential energy has a dip around x = 0 and goes to
zero as |x| becomes large. Because V (x) is similar to the delta-function well, we expect there to
be plane wave solutions, C5 e−ikx and C6 eikx , when |x| is large. But unlike the delta-function well,
V (x) is a bounded and continuous function, meaning ψ(x) has the same formula for x < 0 and
x > 0. That means the incident and transmitted probability currents are equal, or T = 1 and
R = 0. Check for a solution of the form ψ(x) = q(x)eikx .

d2
[q(x)eikx ] + [k 2 + 2a2 sech2 (ax)][q(x)eikx ] = 0
dx2

www.stemjock.com
Griffiths Quantum Mechanics 3e: Problem 2.52 Page 11 of 12

Simplify the left side.

d2 q
 
ikx dq 2 2
e + 2ik + 2a sech (ax)q(x) = 0
dx2 dx
Divide both sides by eikx .
d2 q dq
+ 2ik + 2a2 sech2 (ax)q(x) = 0
dx2 dx
Make the substitution z = ax. The derivatives become
dq dq
=a
dx dz
d2 q 2
2d q
= a .
dx2 dz 2
As a result,
d2 q dq
a2 2
+ 2iak + 2a2 sech2 (z)q(z) = 0.
dz dz
Divide both sides by a2 .
d2 q 2ik dq
+ + 2 sech2 (z)q(z) = 0 (5)
dz 2 a dz
This ODE is signficantly harder to solve because of the presence of the first derivative, but that’s
okay. Since V (x) = 0 > −(~2 a2 /m) sech2 (ax) for all finite x, the eigenfunction corresponding to
the zero eigenvalue is a scattering state. Whatever formula we find for q(x) should then simplify
to this result for the case that E = 0. The eigenfunction corresponding to the zero eigenvalue
therefore serves as a hint for q(x). Plug in a trial solution of the form
q(z) = B0 + B1 tanh z + B2 z tanh z into equation (5).

d2 2ik d
(B0 + B1 tanh z + B2 z tanh z) + (B0 + B1 tanh z + B2 z tanh z)
dz 2 a dz
+ 2 sech2 (z)(B0 + B1 tanh z + B2 z tanh z) = 0
Simplify the left side.
2 2ik 2ik
(B0 a + B2 a + ikB1 ) sech2 z + B2 z sech2 z + B2 tanh z = 0
a a a
Match the coefficients on both sides.
B0 a + B2 a + ikB1 = 0
2ik
B2 = 0
a
2ik
B2 = 0
a
Set B2 = 0 to satisfy these last two equations. The first equation then reduces to B0 a + ikB1 = 0.
Choose B0 = ik and B1 = −a. As a result, one solution to equation (5) is
q(z) = ik − a tanh z.
Use the method of reduction of order now to determine the general solution: Plug in
q(z) = (ik − a tanh z)p(z) into equation (5).
d2 2ik d
2
[(ik − a tanh z)p(z)] + [(ik − a tanh z)p(z)] + 2 sech2 (z)[(ik − a tanh z)p(z)] = 0
dz a dz

www.stemjock.com
Griffiths Quantum Mechanics 3e: Problem 2.52 Page 12 of 12

Simplify the left side.

d2 p 2  2  dp
(ik − a tanh z) 2
− a sech2 z + k(k + ia tanh z) =0
dz a dz
This is a first-order linear ODE for dp/dz, so it can be solved with an integrating factor. The
general solution is
ik cosh z + a sinh z
p(z) = C5 + C6 e−2ikz/a .
k cosh z + ia sinh z
That means

q(z) = (ik − a tanh z)p(z) = C5 (ik − a tanh z) + C6 e−2ikz/a (ia tanh z − k).

Change back to the original variable x.

q(x) = C5 [ik − a tanh(ax)] + C6 e−2ikx [ia tanh(ax) − k]

Consequently,

ψ(x) = q(x)eikx
= C5 [ik − a tanh(ax)]eikx + C6 e−2ikx [ia tanh(ax) − k]eikx

= D5 [ik − a tanh(ax)]eikx + D6 [ik + a tanh(ax)]e−ikx .

Since neither term blows up as x → ±∞, both of them must be kept. Solving the ODE in t yields
φ(t) = e−iEt/~ . Therefore, another solution to Schrödinger’s equation is

Ψ(x, t) = φ(x)φ(t)
= D5 [ik − a tanh(ax)]eikx e−iEt/~ + D6 [ik + a tanh(ax)]e−ikx e−iEt/~
   
iEt iEt
= D5 [ik − a tanh(ax)] exp ikx − + D6 [ik + a tanh(ax)] exp −ikx −
~ ~
     
Et Et
Ψ(x, t) = D5 [ik − a tanh(ax)] exp ik x − + D6 [ik + a tanh(ax)] exp −ik x + ,
~k ~k
√
where E > 0 and k = 2mE/~. If |x| is large, then this is a linear combination of plane waves
travelling to the right and to the left, respectively, as predicted. For a plane wave incident from
the left only, D6 is set to zero; and for a plane wave incident from the right only, D5 is set to zero.

Again, it must be emphasized that V (x) = −(~2 a2 /m) sech2 (ax) is a reflectionless potential
because ψ(x) has the same formula for all x. And this is due to the fact that V (x) is bounded
and continuous unlike all the cases considered previously in the chapter.

www.stemjock.com