3 The Probability Current

Let’s take a closer look at the integrand of equation (2.10). Using the explicit expression for
the Hamiltonian we have
∂ρ i ˆ ∗ ˆ
= ((HΨ) Ψ − Ψ∗ (HΨ))
∂t ~
(3.1)
~2 ∂ 2 Ψ∗ 2
   
i ∗∂ Ψ ∗ ∗
= − Ψ−Ψ + V (x, t)Ψ Ψ − Ψ V (x, t)Ψ .
~ 2m ∂x2 ∂x2

The contributions from the potential cancel and we then get

 2 ∗ 2

i ∗ ∗ ˆ ~ ∂ Ψ ∗∂ Ψ
((ĤΨ) Ψ − Ψ (HΨ)) = Ψ−Ψ . (3.2)
~ 2im ∂x2 ∂ x2

The only chance to get to show that the integral of the right-hand side is zero is to show that
it is a total derivative. Indeed, it is!

~ ∂Ψ∗
  
i ˆ ∗ ∗ ˆ ∂ ∗ ∂Ψ
((HΨ) Ψ − Ψ (HΨ)) = Ψ−Ψ
~ ∂x 2im ∂x ∂x
∂Ψ∗
  
∂ ~ ∗ ∂Ψ
= − Ψ − Ψ
∂x 2im ∂x ∂x
   (3.3)
∂ ~ ∗ ∂Ψ
= − 2i Im Ψ
∂x 2im ∂x
  
∂ ~ ∂Ψ
= − Im Ψ∗ ,
∂x m ∂x

where we used that z − z ∗ = 2i Im(z). Recall that the left-hand side we have evaluated is
actually ∂ρ
∂t
and therefore the result obtained so far is
  
∂ρ ∂ ~ ∗ ∂Ψ
+ Im Ψ = 0. (3.4)
∂t ∂x m ∂x

This equation encodes charge conservation and is of the type

∂ρ ∂J
+ = 0, (3.5)
∂t ∂x
where J(x, t) is the current associated with the charge density ρ. We have therefore identified
a probability current
~  ∂Ψ 
J(x, t) ≡ Im Ψ∗ . (3.6)
m ∂x

There is just one component for this current since the particle moves in one dimension. The
units of J are one over time, or probability per unit time, as we now verify.

5
 −1/2
R For 2one spatial dimension, [Ψ] = L , which is easily seen from the requirement that
dx|Ψ| is unit free. (When working with d spatial dimensions the wavefunction will have
units of L−d/2 ). We then have

ML2 L2
   
∗ ∂Ψ 1 ~
Ψ = 2 , [~] = , = , (3.7)
∂x L T m T
1
=⇒ [J] = = probability per unit time (3.8)
T
We can now show that the time derivative of N is zero. Indeed, using (3.5) we have
Z ∞ Z ∞
dN ∂ρ ∂J
= dx = − dx = −(J(∞, t) − J(−∞, t)) . (3.9)
dt −∞ ∂t ∞ ∂x

The derivative vanishes if the probability current vanishes at infinity. Recalling that

∂Ψ∗
 
~ ∗ ∂Ψ
J= Ψ −Ψ , (3.10)
2im ∂x ∂x

we see that the current indeed vanishes because we restrict ourselves to wavefunctions for which
limx→±∞ Ψ = 0 and limx→±∞ ∂∂xΨ remains bounded. We therefore have

dN
= 0, (3.11)
dt

as we wanted to show.
To illustrate how probability conservation works more generally in one dimension, focus
on a segment x ∈ [a, b]. Then the probability Pab to find the particle in the segment [a, b], is
given by Z b
Pab = ρ(x, t) dx . (3.12)
a
If we now take the time derivative of this and, as before, use current conservation we get
Z b
dPab ∂J(x, t)
=− dt = −J(b, t) + J(a, t) . (3.13)
dt a ∂x

This is the expected result. If the amount of probability in the region [a, b] changes in time, it
must be due to probability current flowing in or out at the edges of the interval. Assuming the
currents at x = b and at x = a are positive, we note that probability is flowing out at x = b
and is coming in at x = a. The signs in the above right-hand side correctly reflect the effect of
these flows on the rate of change of the total probability inside the segment.