Documente Academic
Documente Profesional
Documente Cultură
Chris Clark
Spicy Lifestyle Academy, Department of Physics, Tokyo, Japan∗
(Dated: September 17, 2010)
A pedagogically-oriented historical introduction to the theoretical aspects of wave mechanics
followed by a discussion of real-valued wave functions in the Klein-Gordon equation.
themselves could not be represented as sim- are traveling through the wave form as it is oscillating in
ple corpuscles either, but that a periodicity place. The actual frequency of the wave in our new refer-
had also to be assigned to them too.” [3] ence frame is still given by the Planck-Einstein equation.
Thus, the relation between the frequency of the wave, ω1 ,
He may have also been encouraged to take this leap by and the apparent frequency of its internal oscillation, ω00 ,
Einstein’s work on the photoelectric effect. is
“...I had a sudden inspiration. Einstein’s γ1 mc2
wave-particle dualism was an absolutely gen- ω1 = = γ1 ω0 = γ12 ω00
h̄
eral phenomenon extending to all physical na-
ture.” [4] Now consider the value of the wave at the center of the
light quantum, ψ0 (t), still observing from our moving
The most famous equation to emerge from de Broglie’s
frame. As implied by the definition of ω00 , the value will
work was his relation between the momentum of a parti-
oscillate sinusoidally according to
cle and the wavelength of its matter wave. Such a rela-
tion was already known for light quanta. By combining ψ0 (t) = A sin(ω00 t)
the Planck-Einstein equation, E = hν, with the phase
velocity formula, c = λν, and the result from electro- provided a suitable choice for the zero point of time.
magnetism. E = pc (based on Poynting’s Theorem), one Even though the wave has to be confined to a finite
can derive region, the closer we zoom in on the center, the more
E hν h it will look like an infinite monochromatic plane wave,
p= = = since we are assuming a light quantum with a definite
c λν λ
frequency. The frequency of this monochromatic wave
In 1923, de Broglie suggested that if matter waves ex- is what we measure by counting passing crests, namely
isted, they would also obey this equation. [5] He started ω1 . Therefore, we can approximate the value of the wave
by making the assumption, now believed to be false, that near the center of the quantum as
light quanta have some very small but non-zero mass.
Under this assumption, he equated the energy of a light ψ(x, t) = A sin(ω1 t − kx)
quantum given by the Planck-Einstein equation, E = hν,
with the relativistic energy of a massive particle given by with coordinates chosen such that the center of the quan-
E = γmc2 . In terms of the constants h̄ = h/2π and tum passes x = 0 at time t = 0, and k being an unde-
ω = 2πν, this can be written as termined wave vector. The value at the center of the
particle, ψ0 (t), can be evaluated from this expression by
h̄ω = γmc2 plugging in the distance that we have traveled from the
center at time t i.e. x = v1 t.
According to theory of relativity, this equation should
apply in all inertial reference frames. So in the rest frame ψ0 (t) = ψ(v1 t, t) = A sin(ω1 t − kv1 t)
of the particle, where γ = 1, there should be a frequency
of oscillation ω0 such that Equating the two expressions for ψ0 (t) gives
h̄ω0 = mc2 A sin(ω00 t) = A sin(ω1 t − kv1 t)
In this frame the particle is not moving, so this oscil-
For this to be true, the parameters of the sine functions
lation would presumably look like a localized standing
must be equal.
wave. Now suppose we boost to a reference frame that is
moving at velocity v1 with respect to the particle. The ω00 = ω1 − kv1
particle remains untouched, so its matter wave cannot
have changed. However, due to relativistic time-dilation,
the apparent frequency of oscillation of the standing wave
v12
will be reduced to 1− ω1 = ω1 − kv1
c2
ω0
ω00 =
γ1
v12
If we were to measure the frequency of the wave from ω1 = kv1
c2
this frame by counting the number of crests that pass by
per second, we would not get ω00 . The reason is that the
wave crests will be passing by at a higher rate due to ω1 c2
our motion relative to the standing wave. Essentially we =
k v1
3
This tells us the phase velocity u = ω/k of the matter perhaps, be overcome in an exactly similar
wave for any velocity, which automatically gives us the way?”
wavelength since u = λν. “As stated above, the wave-phenomena must
u h ω h c h 2
h in this case be studied in detail. This can
λ= = = = = only be done by using an “equation of wave
ν γmc2 k γmc2 v γmv p
propagation.” Which one is this to be? In
Strictly speaking, this derivation only applied to light the case of a single material point, moving in
quanta because there was no scientific evidence at the an external field of force, the simplest way is
time to suggest that other types of particles obeyed the to try to use the ordinary wave equation...”
Planck-Einstein equation. So when de Broglie concluded [10]
the derivation with the statement,
The wave equation, already well-known from electromag-
“We are then inclined to admit that any mov- netism, is
ing body may be accompanied by a wave and
that it is impossible to disjoin motion of body ∂2ψ
= u 2 ∇2 ψ
and propagation of wave” [6] ∂t2
he was taking a radial step. where ψ is the wave function and u is the phase veloc-
ity of the wave. The phase velocity is exactly what de
Broglie calculated, so we just need to use de Broglie’s re-
THE SCHRÖDINGER EQUATION sults to plug in for u (though Schrödinger actually used
a different method to obtain the phase velocity that was
De Broglie summarized his work in his doctoral thesis. based on the analogy between Hamilton’s principle and
It is said that his thesis examiners were unsure about Fermat’s principle). The only catch is that we can’t just
giving it their blessing, so they sent it to Einstein for his plug in the result u = c2 /v because v is an unknown vari-
opinion. [7] Einstein endorsed it, which got de Broglie able. Since we only have one equation, we can only allow
his PhD, and also led Einstein to promote de Broglie’s one unknown, which is ψ; otherwise the equation won’t
idea. Erwin Schrödinger, who had already been working be solvable. The solution is to express the phase velocity
on the quantum theory of electrons in 1922 [8], was ex- in terms of the total energy of the particle and the ex-
posed to deBroglie’s ideas through Einstein. [9] In 1926, ternal potential, since the total energy is a constant and
Schrödinger published a paper that introduced his fa- the external potential is known by assumption.
mous equation governing the behavior of particle wave Initially, Schrödinger used the correct relativistic en-
functions. The introduction began, ergy and found the relativistic wave equation. However,
the predictions that came out of the relativistic wave
“The theory which is reported in the following equation contradicted the experimental data on the fine
pages is based on the very interesting and fun- structure lines of the Hydrogen spectrum. The reason
damental researches of L. de Broglie on what was that the equation neglected the spin of the electron,
he called ”phase waves” and thought to be as- which he correctly suspected, but he did not know how
sociated with the motion of material points, to fix the problem. [11] So instead he swept the issue
especially with the motion of an electron or under the rug by taking the non-relativistic approxima-
proton. The point of view taken here... is tion, which chops off the fine structure term. This places
rather that material points consist of, or are the focus on the gross structure predictions, which did
nothing but, wave-systems.” [10] match well with experiment. The consequence is that
In taking this leap to a fully wave-based view of matter, Schrödinger’s famous equation is only an approximation,
Schrödinger was compelled by the similarity between Fer- not a true law of physics.
mat’s principle, pertaining to wave paths, and Hamilton’s In the non-relativistic approximation, the total energy
principle, pertaining to particle paths. With this rela- can be split into the kinetic (K) and potential (V ) com-
tionship in mind, Schrödinger considered the inadequacy ponents as follows.
of classical physics to explain atomic emission spectra
p2
and asked, E =K +V = +V
2m
“...is one not greatly tempted to investigate
whether the non-applicability of ordinary me- p2
chanics to micro-mechanical problems is per- =E−V
2m
haps of exactly the same kind as the non-
applicability of geometrical optics to the phe- p
nonema of diffraction or interference and may, p= 2m(E − V )
4
Using this expression and de Broglie’s result λ = h/p complex conjugate obeys the equation with the other
gives sign. Despite the fact that this equation is only a non-
relativistic approximation, Schrödinger was able to use
hE E it to derive the energy levels of the Bohr atom, which
u = λν = =p
ph 2m(E − V ) provided convincing evidence for the wave interpretation
of electrons. Some mark this as the birth of wave me-
Notice we have assumed that the energy in the Planck-
chanics.
Einstein equation is the total energy, even though we
have not ruled out the possibility that it should be the
kinetic energy since the kinetic energy and total energy of THE SCHRÖDINGER CONTINUITY EQUATION
photons are equal. Schrödinger used a different method
to obtain u, based on the analogy between Hamilton’s
Shortly after publication of his famous wave equation,
principle and Fermat’s principle that had inspired de
Schrödinger added another important piece to the puz-
Broglie. For more on this issue, see Appendix 1. Plug-
zle that would aid in establishing an interpretation of
ging this u into the wave equation,
the wave function. [12] His discovery was that the wave
∂2ψ E2 function could be fit into the continuity equation familiar
2
= ∇2 ψ from electromagnetism,
∂t 2m(E − V )
Now we restrict the possible solutions to those with fre- ∂ρ
= −∇ · j
quency ν = E/h by declaring the time-dependence of ψ ∂t
so that with the proper definitions of ρ and j in terms of ψ. The
±iωt ±iEt/h̄ Schrödinger continuity equation can be derived directly
ψ(x, t) = ψ(x, 0)e = ψ(x, 0)e
from the Schrödinger equation. Multiplying the whole
Note that here Schrödinger introduces complex num- equation by −i/h̄,
bers without justification or comment. This was per-
∂ψ ih̄ 2 i
haps a seemingly harmless assumption because electro- = ∇ ψ− Vψ
magnetic theory taught physicists that waves are more ∂t 2m h̄
conveniently written in complex notation. The first two The complex conjugate of this is,
time derivatives of ψ are
∂ψ ∗ ih̄ 2 ∗ i
∂ψ iE iE =− ∇ ψ + V ψ∗
= ± ψ(x, 0)e±iEt/h̄ = ± ψ(x, t) ∂t 2m h̄
∂t h̄ h̄ ∗
Therefore, the quantity ψ ∗ ∂ψ ∂ψ
∂t + ψ ∂t is
∂2ψ E2 E2
= − 2 ψ(x, 0)e±iEt/h̄ = − 2 ψ(x, t) ∂ψ ∂ψ ∗ ih̄
ψ∗ ψ ∗ ∇2 ψ − ψ∇2 ψ ∗
∂t2
h̄ h̄ +ψ =
∂t ∂t 2m
With this we can eliminate the second time derivative
Equivalently,
E2 E2
− 2ψ= ∇2 ψ ∂ ∗ ih̄
h̄ 2m(E − V ) (ψ ψ) = ∇ · (ψ ∗ ∇ψ − ψ∇ψ ∗ )
∂t 2m
This has the form of the continuity equation, where
h̄2 2
(E − V )ψ = − ∇ ψ
2m ρ = |ψ|2
h̄2 2 ih̄
Eψ = − ∇ ψ+Vψ j=− (ψ ∗ ∇ψ − ψ∇ψ ∗ )
2m 2m
This is known as the time-independent Schrödinger equa-
Naturally, Schrödinger was led to interpret ρ and j as
tion. To get the time-dependent Schrödinger equation we
the electric charge and current densities for the particle
just substitute the first time-derivative back in
being modeled. For the case of an electron he suggested,
∂ψ h̄2 2 “...the charge of the electron is not concen-
±ih̄ =− ∇ ψ+Vψ
∂t 2m trated in a point, but is spread out through
The sign ambiguity, as Schrödinger points out, is not the whole space, proportional to the quantity
a problem; ψ obeys the equation for one sign and its ψ ψ̄.” [10]
5
THE BORN RULE function is known as the Born rule. Einstein, De Broglie,
and Schrödinger all disliked this interpretation. [14] One
By comparing the theoretical predictions of the wave reason was that it seemed counter-intuitive that the uni-
equation to the experimental data on Hydrogen’s spec- verse would operate randomly with no underlaying de-
trum, Schrödinger was able to further justify his hypothe- terministic behavior since ordinarily randomness emerges
sis about the interpretation of the wave function ψ. How- from outcome counting of deterministic processes, reduc-
ever, by the end of the year this interpretation of the wave ing randomness to an illusion of ignorance. Another issue
function would be overturned by Max Born. [13] Back in was that the probabilistic interpretation created a para-
1925, a year before Schrödinger published his wave equa- doxical wave-particle duality in which quantized particles
tion, one of Born’s assistants named Werner Heisenberg still exist, but mysteriously act like waves in unspecified
had introduced a statistical form of quantum theory that circumstances. The wave-particle duality issue can be
would come to be known as matrix mechanics. Heisen- nearly resolved by assuming that there are no particles;
berg’s theory was motivated by the problem of accurately just waves obeying the stipulation that interactions are
predicting atomic emission spectra, which are determined quantized. In perhaps the most famous book on optics
by atomic state transitions. In 1916, Einstein had es- ever written, Max Born and Emil Wolf corroborate this
tablished the idea of assigning transition probabilities to view for the case of light.
each atomic state transition. As a result of these origins,
Heisenberg’s theory was inherently probabilistic. Born “Light propagates as if it were an an elec-
revised Heisenberg’s theory, including reformulating it in tromagnetic wave, but interacts with matter
terms of matrices. Therefore, Born had a predisposition as if its energy were concentrated in photons,
against Schrödinger’s non-probabilistic interpretation of each with an energy quantum.” [15]
the wave function.
If we extend this view to apply to all species of “parti-
“To us in Göttingen [Schrödinger’s] interpre- cles”, the wave-particle duality ceases to be so paradox-
tation seemed unacceptable in face of well es- ical. However, the mechanism of quantization and prob-
tablished experimental facts... I had there- abilistic behavior in interactions still remains an open
fore, as early as the end of 1925, made an at- question.
tempt to extend the matrix method... Again
an idea of Einstein’s gave me the lead. He
had tried to make the duality of particles - THE KLEIN-GORDON EQUATION
light quanta or photons - and waves com-
prehensible by interpreting the square of the When de Broglie read Schrödinger’s first publication
optical wave amplitudes as probability den- on the wave equation, he immediately objected to the
sity for the occurrence of photons. This con- non-relativistic approximation that Schrödinger had em-
cept could at once be carried over to the ψ- ployed. [16] De Broglie’s concept of matter waves had
function: |ψ|2 ought to represent the prob- been heavily based on the relativistic phenomenon of
ability density for electrons (or other parti- time dilation, so he could not be satisfied with a non-
cles).” [14] relativistic theory. Schrödinger had already worked out
the relativistic wave equation in December 1925, and it is
Born chose to consider an atomic scattering problem in believed that he was the first to do so, but as discussed
which a swarm of electrons collide with a heavy atom. He earlier, he didn’t publish it. Many others also saw the
solved Schrödinger’s equation for this problem and made need for a relativistic equation, and during 1926 several
the following observations about the wave function. researchers independently reached the same result. The
“The square of the amplitude of this wave first to publish was Oskar Klein in April 1926, where the
at a great distance from the scattering centre equation played a subordinate role in his five-dimensional
determines the relative probability of scatter- theory of gravitation and electromagnetism that would
ing as a function of direction. Moreover, if the later develop into Kaluza-Klein theory. Walter Gordon
scattering atom itself is capable of existing in independently arrived at the equation a few months later
different stationary states, then Schrödinger’s in a paper addressing Compton scattering. Despite nu-
wave equation gives automatically the prob- merous other discoverers, the equation eventually came
ability of excitation of these states...” [14] to be known as the Klein-Gordon equation. [16]
The Klein-Gordon equation can be obtained in the
These observations led Born to a probabilistic interpre- same manner as the Schrödinger equation; we simply in-
tation of the wave function, in which |ψ|2 represents the sert the proper relativistic expressions for energy and mo-
probability density for the particle’s position. The gen- mentum. Again we start by expressing the momentum in
eral form of this probabilistic interpretation of the wave terms of the total energy and potential. In order In order
6
to do that, we need to eliminate β using the relativistic There are three fundamental problems with this equa-
energy expression. tion as it currently stands. The first, as we have already
mentioned, is that it does not account for spin. Since all
E = γmc2 + V fundamental particles discovered to date have non-zero
spin, this equation can only be experimentally tested on
mc2 spineless composite composite particles like the spinless
E−V = p
1 − β2 pion. Such experiments have validated the Klein-Gordon
equation. [17] However this doesn’t necessarily mean the
Klein-Gordon is inapplicable to free fundamental parti-
m2 c4
1 − β2 = cles since spin might only be relevant during interactions.
(E − V )2 The other two problems are that the Klein-Gordon
equation seems to permit (1) negative probability den-
m2 c4 sities and (2) negative energies. We will encounter these
β2 = 1 − two problems in the upcoming sections.
(E − V )2
m2 c2
p2 = γ 2 m2 β 2 c2 = β 2 THE KLEIN-GORDON CONTINUITY
1 − β2 EQUATION
2 4
(E − V )2
m c
= 1−
(E − V )2 c2 A continuity equation can be found for the Klein-
1 Gordon equation by following a procedure similar to that
= 2 (E − V )2 − m2 c4
c used to get the Schrödinger continuity equation. Divid-
ing the Klein-Gordon equation by −h̄2 and rearranging,
Now that we have the momentum expressed in terms of ∂2ψ 2i ∂ψ V 2 2 2 m2 c4
E and V , we can easily get the phase velocity in terms = − V + 2 ψ + c ∇ ψ − ψ
∂t2 h̄ ∂t h̄ h̄2
of the same parameters.
The complex conjugate is
2 h2 E 2
2 2 c2 E 2
u =λ ν = 2 2 =
p h (E − V )2 − m2 c4 ∂ 2 ψ∗ 2i ∂ψ ∗ V2 m2 c4 ∗
= V + 2 ψ ∗ + c2 ∇2 ψ ∗ − ψ
Finally, we can insert the phase velocity into the wave
∂t 2 h̄ ∂t h̄ h̄2
equation. 2 2 ∗
So the quantity ψ ∗ ∂∂tψ2 − ψ ∂∂tψ2 is
∂2ψ c2 E 2
= ∇2 ψ ∂2ψ ∂ 2 ψ∗
∂t2 (E − V )2 − m2 c4 ψ∗ − ψ
∂t2 ∂t2
∂2ψ
(E − V )2 − m2 c4 = c2 E 2 ∇2 ψ 2i
∂ψ ∗
∂t2 ∗ ∂ψ
+ c2 ψ ∗ ∇2 ψ − ψ∇2 ψ ∗
=− V ψ +ψ
h̄ ∂t ∂t
Now since the time-dependence of ψ is of the form e±iωt ,
2 2
we can substitute ∂∂tψ2 = − E
h̄2
ψ(x, t).
∂ψ ∗
∂ ∂ψ 2i
E2 ψ∗ −ψ + V ψ∗ ψ = c2 ∇·(ψ ∗ ∇ψ − ψ∇ψ ∗ )
− (E − V )2 − m2 c4 2 ψ = c2 E 2 ∇2 ψ ∂t ∂t ∂t h̄
h̄
To make the units consistent with the Schrödinger con-
Canceling the factor E 2 and rearranging,
tinuity equation, we multiply this equation by ih̄/2mc2 .
(E − V )2 ψ = −h̄2 c2 ∇2 ψ + m2 c4 ψ
∂ψ ∗
ih̄ ∂ ∗ ∂ψ 2i ∗
ψ − ψ + V ψ ψ
If the potential is time-independent, we can write 2mc2 ∂t ∂t ∂t h̄
E 2 ψ − 2EV ψ + V 2 ψ = −h̄2 c2 ∇2 ψ + m2 c4 ψ
ih̄
Inserting time derivatives to eliminate E yields the = ∇ · (ψ ∗ ∇ψ − ψ∇ψ ∗ )
2m
time-dependent Klein-Gordon equation for the time-
independent potential V . This has the form of the continuity equation,
∂2ψ ∂ψ ∂ρ0
−h̄2 2
− 2ih̄V + V 2 ψ = −h̄2 c2 ∇2 ψ + m2 c4 ψ = −∇ · j0
∂t ∂t ∂t
7
follow the standard prescription, but use V in place of equation. The textbook complex wave function’s proba-
the usual complex conjugate. bility flux is
∂2U m2 c4
V = c2 V ∇2 U − VU h̄
∂t2
h̄2 j0 = (ψ ∗ ∇ψ − ψ∇ψ ∗ )
2mi
h̄
= ((ψR − iψI ) ∇ (ψR + iψI )
∂2V 2 2 m2 c4 2mi
U = c U ∇ V − UV
∂t2 h̄2 − (ψR + iψI ) ∇ (ψR − iψI ))
h̄
Subtracting the second from the first, = (ψR ∇ψR + iψR ∇ψI − iψI ∇ψR + ψI ∇ψI
2mi
∂2U ∂2V − (ψR ∇ψR − iψR ∇ψI + iψI ∇ψR + ψI ∇ψI ))
V − U = c2 V ∇2 U − c2 U ∇2 V h̄
∂t2 ∂t2 = (2iψR ∇ψI − 2iψI ∇ψR )
2mi
h̄ h̄ h̄
∂
∂U ∂V
= U ∇V − V ∇U
V −U = c2 ∇ · (V ∇U − U ∇V ) m 2mc2 2mc2
∂t ∂t ∂t
1 h̄2
= (U ∇V − V ∇U )
We multiply both sides by h̄2 /2mc2 to give this quantity 2 m2 c2
units of energy.
which is exactly equal to j1 /mc2 for real wave functions.
1 h̄2 ∂ 1 h̄2
∂U ∂V
V −U = ∇ · (V ∇U − U ∇V ) The textbook complex wave function’s probability den-
2 mc2 ∂t ∂t ∂t 2m
sity is
This is the continuity equation, having the form
∂ψ ∗
∂ρ1 ih̄ ∗ ∂ψ ih̄ ∗ ∂ψ
= −∇ · j1 ρ0 = ψ −ψ = Im ψ
∂t 2mc2 ∂t ∂t mc2 ∂t
ih̄ h̄
where = Im U − i V
2mc2 2mc2
1 h̄2
∂U h̄ ∂V
∂U ∂V
ρ1 = V −U × +i 2
2 mc2 ∂t ∂t ∂t mc ∂t
1 h̄2
∂U ∂V
= V − U
2 m 2 c4 ∂t ∂t
1 h̄2
j1 = (U ∇V − V ∇U )
2m
1 2 2 1 h̄2 2 2 1 2 2 2
ρ2 = mc U0 (x) sin2 (ωt) + ω U0 (x) cos2 (ωt) ρ2 = mc A sin (k · x − ω(k)t)
2 2 mc2 2
1 h̄2 1 h̄2 2
+ |∇U0 (x)|2 sin2 (ωt) + ω (k)A2 cos2 (k · x − ω(k)t)
2m 2 mc2
1 2 2 1 h̄2 2 2
ρ2 = mc U0 (x) sin2 (ωt) + k A cos2 (k · x − ω(k)t)
2 2m
1 1 1 2 2 2
+ (h̄2 k 2 c2 + m2 c4 )U02 (x) cos2 (ωt) ρ2 = mc A sin (k · x − ω(k)t)
2 mc2 2
1 h̄2 1 1
+ |∇U0 (x)|2 sin2 (ωt) + (h̄2 k 2 c2 + m2 c4 )A2 cos2 (k · x − ω(k)t)
2m 2 mc2
1 2 2 1 h̄2 k 2 2 1 h̄2 2 2
ρ2 = mc U0 (x) + U (x) cos2 (ωt) + k A cos2 (k · x − ω(k)t)
2 2 m 0 2m
1 h̄2 1 2 2 h̄2 2 2
+ |∇U0 (x)|2 sin2 (ωt) ρ2 = mc A + k A cos2 (k · x − ω(k)t)
2m 2 m
h̄2 h̄2
j2 = − ωU0 (x) cos(ωt) sin(ωt)∇U0 (x) j2 = kω(k)A2 cos2 (k · x − ω(k)t)
m m
∂2U 2 2 m2 c4
= c ∇ U − U
∂t2 h̄2
1 h̄2 2 we obtain,
ρ1 = ω (k)A2
2 mc2
× cos2 (k · x − ω(k)t) + sin2 (k · x − ω(k)t) −ω 2 (k) A sin(k · x − ω(k)t) =
A2 h̄2 m2 c4
j1 = kω(k) ω 2 (k) = k 2 c2 +
2 m h̄2
11
r
m2 c4 The surface integral goes to zero when integrating over a
ω(k) = ± k 2 c2 +
h̄2 region bounded by a surface where U = 0 so
Z Z
Indeed, this is the same dispersion relation as that of the 3
ρ1 d x = ρ2 d3 x
complex Klein-Gordon equation. V V
Now, the group velocity vector is given by
This means that in the classical limit, where quanta
vg (k) = ∇k ω(k) look like point particles, ρ1 and ρ2 are indistinguishable.
Given that this is the case, the connection between j1 and
where the subscript indicates that derivatives should be classical momentum mentioned in the last section makes
taken with respect to the parameter k instead of the po- more sense because ρ1 does represent the energy from a
sition variable. Plugging in the Klein-Gordon dispersion classical perspective.
relation, the x-component is,
r
∂ ∂ m2 c4 A REVISED INTERPRETATION OF THE WAVE
(vg )x (k) = ω(k) = ± kx2 c2 + ky2 c2 + kz2 c2 + FUNCTION
∂kx ∂kx h̄2
So for we have found that
kx c2 kx c2
= ±q =± 1. ρ2 represents a non-negative energy density.
kx2 c2 + ky2 c2 + kz2 c2 + m2 c4 ω(k)
h̄2
2. ρ1 and ρ2 are indistinguishable in the classical limit.
The other components are similar, so
3. j1 = 0 for a set of solutions with zero momentum,
kc 2 whereas j2 6= 0.
vg (k) = ±
ω(k) 4. j1 is constant for a set of solutions with constant
translational motion, whereas j2 is not.
Therefore, combining this with the result from the last
section, we see that 5. j1 satisfies an equation that is analagous to the clas-
sical definition of momentum for monochromatic
j1 = ρ1 vg (k) plane wave solutions.
where j1 and ρ1 are the flux and density for a monochro- Based on this evidence, we propose that ρ2 is the energy
matic plane wave with wave vector k. This is some- density, ρ2 /mc2 is the probability density, and g = j1 /c2
what reminiscent of the classical definition of momen- is the momentum density. So the first continuity equa-
tum, p = mv, if we consider ρ1 to be the energy density tion can be thought of as the momentum continuity equa-
and g = j1 /c2 to be the momentum density. However, tion and the second as the energy/probability continuity
we already saw that ρ1 can go negative and that ρ2 rep- equation.
resents the actual energy density. In the next section we This is consistent with the use of the momentum op-
will examine the relationship between ρ1 and ρ2 further. erator in standard quantum mechanics. Recalling that
j1 is proportional to the complex form according to
j1 = mc2 j0 ,
INTEGRATING OVER A WAVE PACKET
j1 ih̄
g= = mj0 = − (ψ ∗ ∇ψ − ψ∇ψ ∗ )
If we compute the densities over a complete wave c2 2
packet, bounded by a surface where U = 0, we find The net momentum in the volume V is then
1 h̄2
Z Z Z Z
ρ1 d 3 x = ρL − U ∇2 U d3 x pnet = g d3 x = Re ψ ∗ (−ih̄∇)ψ d3 x
V V 2m V V
∂2U
Z we find
(x, t) ' −ω (k0 ) a(k)Uk (x, t) d3 k
2
∂t2 1 2 1 h̄2
|ψ|2 = U + V2
2 2 m2 c4
∂2U Therefore, for non-relativistic wave packets
(x, t) ' −ω 2 (k0 )U (x, t)
∂t2
ρ2
Now we can apply the non-relativistic approximation |ψ|2 '
mc2
to the dispersion relation to get an appropriate expression
for ω. In the non-relativistic approximation, v c so where the right hand side is the predicted probability
γ ' 1. Therefore density. This explains why the magnitude squared of the
wave function works for the non-relativistic Schrödinger
γmv mc equation.
13
APPENDIX 1: WHY IS THE FREQUENCY Mag. 47, 446 (1924). See p. 449-450.
PROPORTIONAL TO TOTAL ENERGY? [7] Wikipedia, Louis de Broglie, retrieved 23:24, July 31,
2010.
[8] E. Schrödinger, Ober eine bemerkenswerte Eigenschaft
Why is hν equal to the total energy instead of the ki-
der Quantenbahnen eines einzelnen Elektrons, Z. Phys.
netic energy? Consider photons since we have good data 12, 1, 13 (1922).
on their frequency. The only force they feel is gravity so [9] V. V. Raman and Paul Forman, Why Was It Schrdinger
we use gravitational red shift as our inspiration. Grav- Who Developed de Broglie’s Ideas? Hist. Stud. Phys. Sci.
itational red shift does not affect the signal itself; it is 1, 291 (1969). See p. 311.
an observational effect caused by different clock rates at [10] E. Schrödinger, An undulatory theory of the mechanics
different gravitational potentials. If clocks actually run of atoms and molecules, Phys. Rev. 28, 6, 1049 (1926).
See p. l049, 1053, 1056-1057, 1066.
slower in time-dilation situations, as ether theory sug-
[11] S. Weinberg, The Quantum Theory of Fields, Vol. 1:
gests, then the frequency never changes. So both total Foundations (Cambridge University Press, Cambridge,
energy and frequency are conserved, thus preserving the 1995). See p. 5.
relation E = hν where E is the total energy. [12] E. Schrödinger, Quantisierung als Eigenwertproblem
(Vierte Mitteilung), Ann. Phys. 81, 109 (1926).
[13] M. Born, Zur Quantenmechanik der Stovorgnge, Z. Phys.
37, 863 (1926).
[14] M. Born, The statistical interpretation of quantum me-
∗
URL: http://www.dfcd.net chanics, Nobel Lecture (1954). See p. 261-262, 256.
[1] A. Michelson, Light waves and their uses (The University [15] M. Born, E. Wolf, and A. B. Bhatia, Principles of Optics
of Chicago Press, Chicago, 1903). See p. 23-25. (Cambridge University Press, Cambridge, 1999) 7th ed.
[2] A. Hermann, The Genesis of the Quantum Theory (1899- See p. 879.
1913) (MIT Press, Cambridge, 1971). See p. 23. [16] H. Kragh, Equation with the many fathers. The Klein-
[3] L. de Broglie, The wave nature of the electron, Nobel Gordon equation in 1926, Am. J. Phys. 52, 11, 1024
Lecture (1929). See p. 247. (1984). See p. 1025.
[4] L. de Broglie, New Perspectives in Physics (Basic Books, [17] Wikipedia, Klein-Gordon equation, retrieved 22:06, July
New York, 1962). See p. 138. 20, 2010.
[5] L. de Broglie, Ondes et quanta Comptes Rendes 177, 507 [18] D. Bohm, Quantum Theory (Prentice-Hall, Englewood
(1923). Cliffs, 1951). See section 3.8.
[6] L. de Broglie, A tentative theory of light quanta, Phil.