Вы находитесь на странице: 1из 4

We want to solve the Schroedinger equation - the full, time dependent Schroedinger

equation - for the wavefunction Ψ(x, t). If the potential is constant in time
then the wavefunction is separable in space and time. we find the spatial bit
by solving the time independent Schroedinger equation - and in general there
are multiple solutions, Hψn = En ψn . Then a particular solution is Ψn (x, t) =
ψn (x)e−iEn t/h̄ and the general solution is Ψ(x, t) = n cn ψn (x)e−iEn t/h̄ .

3.5 Time dependence of multiple states

each individual energy eigenfunction is a stationary state. But if we have a
sum of them as in the general solution

cn ψn (x)e−iEn t/h̄
Ψ(x, t) =

then the contribution of the different energy states changes with time
We can see this explicitally by looking at the probability for a two state
system Ψ(x, t) = c1 ψ1 (x)e−iE1 t/h̄ + c2 ψ2 (x)e−iE2 t/h̄ . and lets make it simple
by choosing states that are real so ψ ∗ = ψ and cn =c∗n

P (x, t) = Ψ∗ Ψ = (c∗1 ψ1∗ eiE1 t/h̄ + c∗2 ψ2∗ eiE2 t/h̄ )(c1 ψ1 e−iE1 t/h̄ + c2 ψ2 eiE2 t/h̄ )

= |c1 |2 ∗ ψ12 + |c2 |2 ψ22 + c1 c2 ψ1 ψ2 eiE1 t/h̄ e−iE2 t/h̄ + c1 c2 ψ1 ψ2∗ e−iE1 t/h̄ eiE2 t/h̄
so the last two terms contain e±i(E2 −E1 )t/h̄ i.e. they are oscillatory in time.
Let E2 − E1 = h̄ω and these go to e±iωt = cos ±iωt + i sin ±iωt.

= |c1 |2 ∗ ψ12 + |c2 |2 ψ22 + c1 c2 ψ1 ψ2 (e−iωt + eiωt )

= |c1 |2 ∗ ψ12 + |c2 |2 ψ22 + c2 ψ1 ψ2 (cos ωt − i sin ωt + cos ωt + i sin ωt)
= |c1 |2 ∗ ψ12 + |c2 |2 ψ22 + 2c1 c2 ψ1 ψ2 cos ωt
so the probability density IS a function of time.
it is still normalised so probablity of finding the particle SOMEWHERE
P (x, t) = 1 as the orthoganality of eigenfunctions kills the time dependent

term in the integral. but when we start asking for expectation values for
systems which are NOT in an energy eigenstate then these CAN be time

3.6 Expectation values of mixed states
Becasue there is this time dependence in P (x, t) then expectation values can
also be time dependent.

< x(t) >= xP (x, t)dx = x(|c1 |2 ψ12 + |c2 |2 ψ22 + 2c1 c2 ψ1 ψ2 cos ωt)dx
= c21 < x(n = 1) > +c22 < x(n = 2) > +2c1 c2 cos ωt xψ1 ψ2 dx
Consider an electron in the ground state of an atom. This is a single eigen-
function, so < x > is not dependent on time as its a standing wave. since an
electron is charged, then this is saying that the charge distribution is constant
also, so there is no radiation. so QM provides a way to resolve the paradox
you get by thinking of electrons as particles which orbit around atoms. in
this picture the electron is accelerating (circular motion) so should radiate
and hence the orbit should decay. But in QM the probability is stationary
so the charge is not ’moving’ so it doesn’t radiate.
But atoms in their excited states do eventually make a transition down to
the ground state. So if we have an atom in state n = 2 then it has some
probability of being in state 1 as well - of decaying down. If its a mix of two
states then its probability density and hence charge distribution oscillate in
time with frequency h̄ω = E2 − E1 which is precisely the frequency of photon
which is emitted to carry away the energy difference. This can’t happen for
the ground state as there is no lower state for this to mix with.
Schroedingerwrote ’It is hardly necessary to point out how much more grat-
ifying it would be to concieve of a quantum transition as an energy change
from one vibrational mode to another, than to regard it as a jumping of an
Wavefunctions which are mixtures of eigenstates give expectation values
which depend on time! pure eigenstates do not.

3.7 How to find the cn for any function

we have f (x) = n cn ψn so multiply by ψm and integrate

∗ ∗ ∗
ψm f (x)dx = ψm cn ψn = cn ψm ψn dx = cn δmn = cm
n n n

so for any functional form f (x) we can decomposeR it into a weighted sum of
energy eigenfunction cn ψn , calculating each cn = ψn∗ f (x)dx
e.g. ψ(x, t = 0) = Ax(L − q x) in an infinite square well potential 0 < x <
L. normalise to get A = 30/L5 . Then we can decompose it into the
sum of energy eigenfunctions for this potential ψn (x) = 2/L sin nπx/L =
N sin nπx/L
cn = AN sin nπx/Lx(L−x)dx = ANL x sin nπx/Ldx−AN x2 sin nπx/Ldx

standard integrals x sin(ax)dx = 1/a2 (sin(ax)−ax cos(ax)) and x2 sin(ax)dx =


1/a3 (2ax sin(ax) − (a2 x2 − 2) cos(ax), where in our case a = nπ/L so

= [sin(nπx/L) − nπx/L cos(nπx/L)]L0 −
n2 π 2 /L2
[2nπx/L sin(nπx/L) − (n2 π 2 x2 /L2 − 2) cos(nπx/L)]L0
n3 π 3 /L3
ANL3   ANL3 
2 2

= 0−nπ cos(nπ)−0+0 − 0−(n π −2) cos(nπ)−0+(0−2)
n2 π 2 n3 π 3
= 3 3 −n2 π 2 cos(nπ) + (n2 π 2 − 2) cos(nπ) + 2)

√ 3
√ √
2.30L 2 60 4 15
=√ (2 − 2 cos(nπ)) = 3 3 (1 − cos(nπ)) = 3 3 (1 − cos(nπ))
L.L5 n3 π 3 nπ nπ
√ √
at n = 1 c1 = 8 π315 , c2 = 0, c3 = 827π153 etc.... all the even n are zero as this is
a function which is symmetic about L/2.

4 1 D potentials
4.1 infinite square well
Ψn (x, t) = ψn (x)e−iEn t/h̄ where the energy eigenfunctions ψn (x) = L2 sin(nπx/L)
corresponding to energy En = n2 E1 and E1 = π 2h̄2 /(2mL2 ). This is for
V = 0 for 0 < x < L, and v = ∞ elsewhere.
BUT!! this is actually a bit inconvienient - we are wanting to do Hydrogen -
we will want potentials which are symmetric about 0 rather than L/2. so we

could simply change the variable to x′ = x − L/2 and we are in a centrally
symmetric potential. Energy doesn’t change, this is just a shift in axis - so
instead of sines we get a bunch of sine and cosine functions depending on
whether n is even or odd....
′ 2
ψn (x ) = sin(nπ(x′ + L/2)/L)
s s
2 2
= sin(nπx′ /L+nπ/2) = (sin nπx′ /L cos(nπ/2)+cos(πx′ /L) sin(nπ/2))
but this depends on n. n=1, then sin π/2 = 1 and cos π/2 = 0
ψ1 (x′ ) = cos(πx′ /L)
similarly for n = 2 then sin π = 0 and cos(π) = −1 so
ψ2 (x′ ) = − sin(2πx′ /L)
′ 2
ψ3 (x ) = − cos(3πx′ /L)
ψ4 (x′ ) = sin(4πx′ /L)