NEUTRONS
AS SCHROEDINGER WAVES AND CATS
Helmut Rauch
Atominstitut der Österreichischen
Universitäten
Stadionallee 2, A-1020 Wien, Austria
Abstract
Neutrons are proper tools for testing quantum mechanics because they are massive, they couple to electromagnetic fields due to their magnetic moment and they are subject to all basic interactions and they are sensitive to topological effects, as well. Related experiments will be discussed. Recent neutron interferometry experiments based on postselection methods renewed the discussion about quantum nonlocality and the quantum measuring process. It has been shown that interference phenomena can be revived even when the overall interference pattern has lost its contrast. This indicates a persisting coupling in phase space even in cases of spatially separated Schrödingercat-like situations. These states are extremely fragile and sensitive against any kind of fluctuations and other decoherence processes. More complete quantum experiments also show that a complete retrieval of quantum states behind an interaction volume becomes impossible in principle.
1 Introduction
- basic relations
Different kinds of neutron interferometers
based on wavefront and amplitude division have been tested in the past [1-4].
The perfect crystal interferometer - first tested in 1974 at our 250 kW
TRIGA reactor - provides highest intensity and became the most frequently used
neutron interferometer due to its wide beam separation and its universal
availability for fundamental-, nuclear- and solid-state physics [5]. It
represents a macroscopic quantum device with characteristic dimensions of
several centimeters (Fig. 1).

Fig. 1.
Sketch of a symmetric perfect crystal neutron interferometer and a
typical interference pattern.
The basis for this kind of neutron
interferometry is provided by the undisturbed arrangement of atoms in a
monolithic perfect silicon crystal [6,2]. An incident beam is split coherently
at the first crystal plate, reflected at the middle plate and coherently superposed
at the third plate. From general symmetry considerations follows immediately
that the wave functions in both beam paths, which compose the beam in the
forward direction behind the interferometer, are equal (
), because they are transmitted-reflected-reflected (TRR) and
reflected-reflected-transmitted (RRT), respectively. The theoretical treatment
of the diffraction process from the perfect crystal is described by the
dynamical diffraction theory [7,8]. To preserve the interference properties
over the length of the interferometer, the dimensions of the monolithic system
have to be accurate on a scale comparable to the so-called Pendellösung length
(~50 mm). The whole
interferometer crystal has to be placed on a stable goniometer table under
conditions avoiding temperature gradients and vibrations. A phase shift between
the two coherent beams can be produced by nuclear, magnetic or gravitational
interactions. In the first case, the phase shift is most easily calculated
using the index of refraction [9,10]:
(1)
where bc is the coherent
scattering length, sr the reaction cross section, and N is the particle
density of the phase shifting material. The different k-vector inside the phase
shifter causes a spatial shift
of the wave packet
which depends on the orientation of the sample surface
and which is related
to the scalar phase shift c by
(2)
where c can be written as a path
integral of the canonical momentum kc along the beam paths
[11]. Therefore, the
intensity behind the interferometer becomes
(3)
The intensity of the beam in the deviated
direction IH follows from particle conservation
. Thus, the intensities behind the interferometer vary as a
function of the thickness D of the phase shifter, the particle density N or the
neutron wavelength .
Neutron optics is a part of
quantum optics and many phenomena can be described properly in that terminology
where the coherence function plays an important role [12,13]
(4)
which is the autocorrelation function of the
wave function. Using a wave packet description for the wave functions
(amplitudes
) one obtains
(5)
where
and
denote the phase
shifts at the mean momentum
. This gives:
. (6)
Thus the absolute value of the coherence
function can be obtained from the fringe visibility
= (IMax - IMin)/(IMax + IMin) or as the Fourier
transform of
.
The mean square distance
related to
defines the coherence
length
which is for Gaussian
distribution functions directly related to the minimum uncertainty relation
. Similar relations can be obtained for time-dependent
phenomena where the spectral distribution g(w) and the temporal
coherence function come into play.
Any experimental device
deviates from the idealized situations: the perfect crystal can have slight
deviations from its perfectness, and its dimensions may vary slightly; the
phase shifter contributes to such deviations by variations in its thickness and
inhomogeneities; and even the neutron beam itself contributes to a deviation
from the idealized situation because of its momentum spread k. Therefore, the experimental interference patterns have to be
descibed by a generalized relation
(7)
where A,
B and are characteristic
parameters of a certain set-up. It should be mentioned, however, that the
idealized behaviour described by eq. (3) can nearly be approached by a well
balanced set-up (Fig. 1). Phase shifts can be applied in the longitudinal,
transverse and vertical directions and the related coherence properties can be
measured [14]. In the transverse direction the phase shift becomes wavelength
independent (cT = -2dhklNbcD0; dhkl
... reflecting lattice planes), which implies a much larger coherence length in
that direction.
All the results of
interferometric measurements, obtained up until now can be explained well in
terms of the wave picture of quantum mechanics and the complementarity
principle of standard quantum mechanics. Nevertheless, one should bear in mind
that the neutron also carries well defined particle properties, which have to
be transferred through the interferometer. These properties are summarized in
Table 1 together with a formulation in the wave picture. Both particle and wave
properties are well established and, therefore, neutrons seem to be a proper
tool for testing quantum mechanics with massive particles, where the
wave-particle dualism becomes very obvious.
All neutron interferometric
experiments pertain to the case of self-interference, where during a certain
time interval, only one neutron is inside the interferometer, if at all.
Usually, at that time the next neutron has not yet been born and is still
contained in the uranium nuclei of the reactor fuel. Although there is no
interaction between different neutrons, they have a certain common history
within predetermined limits which are defined, e.g., by the neutron moderation
process, by their movement along the neutron guide tubes, by the monochromator
crystal and by the special interferometer set-up. Therefore, any interferometer
pattern contains single particle and ensemble properties together.
2 Classic
neutron interference experiments
2.1 Gravity
experiments
The
gravitational interaction of neutrons for usual laboratory conditions is of a
comparable order of magnitude with the mean nuclear and magnetic interaction,
and, therefore, a measurable interference signal is to be expected. The
interaction Hamiltonian in this case reads as
(8)
where
means the
gravitational acceleration directed towards the center of the earth and w the angular rotation frequency of the earth.
The phase shift within the interferometer is
calculated by using the path integral with the canonical momentum as mentioned
earlier. In this way, after several intermediate steps, one gets the gravitational phase shift
(9)
where A is the area enclosed by the coherent beam
trajectories in the interferometer, F is
the angle at which the interferometer is turned out of the horizontal plane, FL is the latitude of the
point at which the experiment takes place, and e is
the angle of rotation around the vertical axis. The first expression in the
above equation describes the familiar gravitational term and was proven by
Colella, Overhauser and Werner [15] by rotating the
interferometer around a horizontal axis (COW-experiment).
This phase shift can be understood as the difference in gravitational
potentials of the two coherent beams, as one travels higher than the other.
Table
1. Properties of neutrons
PARTICLE PROPERTIES WAVE
PROPERTIES
m = 1.6748220(25) x 10-27
kg CONNECTION
= 1.319695(20) x 10-15 m
s =
de
Broglie
= - 9.6491783(18) x 10-27
J/T
for
thermal neutrons: = 1.8 A, v = 2200 m/s
= 882.6 (2.7) s Schrödinger
= 1.8 x 10-10 m
R = 0.7 fm
10-8 m
= 12.0 (2.5) x10-4
fm3 &
10-2 m
u - d - d - quark structure boundary
conditions
1.942(5) x 106 m
0
2 (4)
m mass, s spin, magnetic moment, c Compton wavelength, B de Broglie
-decay lifetime,
R (magnetic) con- -mB __________________ wavelength, c coherence length,
finement radius, electric
polarizability; ô two level system p packet length, k momentum width,
all other measured quantities
like electric mB __________________ t chopper opening
time, v group velocity,
charge magnetic monopole and magnetic phase.
dipole moment are
compatible with zero.
The Coriolis
or Sagnac term in eq. (9) was observed experimentally for the first
time by Werner et al. [16,17], by directing a neutron beam vertically upwards and
by turning a perfect crystal interferometer around this vertical axis (see
Fig. 2). The result gives an impression of how sensitive the
interferometric measuring method actually is. The easiest way to visualize this
effect is by imagining the area encompassed by the two coherent beams as a
differently oriented flag on the rotating earth. A more detailed discussion can
be found by Greenberger [18].
A complimentary investigation to the
gravitational measurements was performed by Bonse and Wroblewski [19],
who brought the interferometer in a slightly oscillatory motion and, in so doing,
also observed a phase shift, this being proportional to the respective
acceleration of the interferometer plates. In summary, this proved the validity
of the classical transformation laws for non-inertial
frames of reference in the quantum limit.
2.2 Neutron
Fizeau effect
An additional phase shift arises when there is a
relative motion between the beam and the phase shifter. The calculation of this
effect can be done on the basis of Galileian transformation because the
velocity of the neutron (
) is much smaller than the velocity of light.
Therefore, the momentum of a particle
inside a material
which moves with the velocity
is given as
. (10)
The Fizeau phase shift arises from the
different phase shifts of a static and moving phase shifter [20].
. (11)

Quantum
mechanics predicts that a Fizeau-phase shift occurs only when the boundary is
moving relative to the neutron beam which is quite a difference to the optical
Fizeau effect which depends on the motional state of the phase shifter
material. The first observation of this effect was achieved by Klein et al.
[21] with a double slit interferometer and a moving quartz phase shifter.
Fig. 2.
Results of the Earth rotational experiment [16,17].
2.3 4p-spinor symmetry
This is probably one of the most discussed
interference experiments. Based on elementary principles of quantum mechanics,
the propagation of a wave function can be described by a unitary
transformation, given by the relevant Hamiltonian. For magnetic interaction,
, the propagation of the two-component spinor wave function, which describes the neutron as a fermion, can
be represented as follows:
(12)
where a
means the Larmor precession angle
. (13)
When inserting the Pauli spin operators, one can
easily show that y(a) has a 4p-symmetry,
and not the 2p-symmetry which we
are used to with respect to expectation values and within the scope of
classical physics,
. (14)
These facts, which were not previously regarded as
verifiable, can be elucidated very easily with neutron interferometry by
observing the intensity modulations, while one of the coherent beams passes
through a magnetic field,
. (15)

Fig. 3. First
verification of the 4p-symmetry
of a spinor wavefunction [22].
The above relation is valid for polarized as well as
for unpolarized neutrons, which points to the inner symmetry properties of
fermions. From eqs. (14) and (15) one recognizes that only for a = 4p is
the original state reproduced. This was verified, nearly simultaneously, in
measurements by Rauch et al. [22] and by Werner et al. [23] (see Fig. 3).
Afterwards, this effect was also proven through several other methods and for a
series of other fermion systems.
2.4 Spin-superposition
Spin superposition is an often used principle of
quantum mechanics. Its curiousity value has been stressed by Wigner [24]. The
wave function of both coherent beams is originally polarized in |z>-direction. One beam is then
inverted to a polarization in |-z>-direction,
whereas the other remains unchanged. Both beams are then superimposed. This
spin flip can be produced, for example, by Larmor precession around a magnetic
field perpendicular to z-direction.
The result for superposition of these two beams, thus prepared, can be obtained
by applying the rotation operator (equ. (12)) to the spin-flipped beam for a
rotation of 180° in y-direction. If we also allow for a
nuclear phase shift, one gets
. (16)
The total wave function
leads
to the following polarization of the out-going beam
. (17)
Consequently, this polarization lies in the x,y-plane,
and is perpendicular to the polarizations of the two superimposed, coherent
beams. This implies that a pure quantum state in |z>-direction, e.g. for c = 0, has been transformed
into a quantum state in |x>-direction,
and, in the sense of self-interference, which definitively applies here, it
seems that each neutron has information about the physical situation in both of
the widely separated coherent beams. The experiment by Summhammer et al. [25]
has fully confirmed this process. Intensity modulations appear only when the
polarization analysis is done in the x,y-plane.
This above-mentioned experiment was repeated with a Rabi resonance flipper, which is also routinely used as a spin
flipping device in polarized neutron physics. Along with the spin flip, a
simultaneous energy exchange is taking place between the resonator system and
the neutrons (DE = hwr = 2mB0,
where B0 is the strength
of the guide field). Here one has to use the time-dependent Schrödinger
equation and to take into consideration the exchange of the total energy of the
neutron system
(18)
which
leads to the following polarization, when the other unchanged coherent beam |z> is superimposed,
. (19)
The
polarization is also in the x,y-plane, however it rotates within this plane
synchroneously with the resonance flip-field. It was possible to demonstrate
this effect with a stroboscopic measurement, where the polarization in a given
direction was measured synchronously with the phase of the flip-field [26].
In
connection with these results, the obvious question arises whether the
measurement of the energy transfer makes a determination of the beam path
possible. One can, however, show that this is impossible, because interference
vanishes in the presence of a measurable energy shift (i.e. larger than the
energy width of the beam), and because the measurement of the energy change of
the flip-field is impossible due to the photon
number-phase uncertainty relationship (DfDN > 1).
These energy exchange measurements have been extended by Summhammer et
al. [27] to multiphoton exchange experiments. In this case an oscillating
magnetic field with a frequency of 7.534 kHz was inserted into one beam and up
to five photon emission and absorption processes have been identified from the
time resolved interference pattern.
2.5 Neutron
Josephson effect
A double coil
arrangement can be used for the observation of a new quantum beat effect, which
is the magnetic analog to the well-known superconducting electric Josephson
effect. If the frequencies of the two coils are chosen to be slightly
different, the energy transfer becomes different too
. The flipping efficiencies for both coils are always very
close to unity (better than 99%). Now, the wave functions change according to
(20)
Therefore, the intensity behind the
interferometer exhibit a typical quantum beat effect, given by
. (21)
Thus, the intensity behind the interferometer
oscillates between the forward and deviated beam without any apparent change
inside the interferometer [28]. The time constant of this modulation can reach
a macroscopic scale which is correlated to the uncertainty relation
. Figure 4 shows the result of an experiment, where the
periodicity of the intensity modulation, T
= 2/
, amounts to T =
(47.90 0.15)s caused by a
frequency difference of about 0.02 Hz. This corresponds to a mean energy
transfer difference E between the two beams, E = 8.610-17 eV, and to an energy sensitivity of 2.710-19 eV, which is by many orders of
magnitudes larger than that of other advanced spectroscopic methods. This high
resolution is strongly decoupled from the monochromaticity of the neutron beam,
which was EB = 5.510-4 eV around a mean
energy of the beam EB = 0.023 eV in this case. The quantum beat effect
can also be interpreted as the magnetic Josephson effect analog where the phase
difference D(t) is driven by the magnetic energy whereas in the
well-known Josephson effect in superconducting tunnel junctions [29], the phase
of the

Cooper-pairs
in both superconductors is driven by the electrical energy.
Fig. 4.
Quantum beat effect observed when the frequencies of the two flipper
coils differ by about 0.02 Hz around 71.899 79 kHz [28].
2.6 Stochastic
versus deterministic beam path detection
A certain beam
attenuation can be achieved either by a semi-transparent material or by a
proper chopper or slit system. The transmission probalility in the first case
is defined by the attenuation cross section a of the material [t = I/I0 = exp(-aN D)]. The change of the wave function is
obtained directly from the complex index of refraction (eq.1):
(22)
Therefore, the beam modulation behind the
interferometer is obtained in the following form
(23)
On the other hand, the transmission
probability of a chopper wheel or another shutter system is given by the open
to closed ratio, t = topen/(topen + tclosed), and one obtains after
straightforward calculations
(24)
i.e. the contrast of the interference pattern
is proportional to
, in the first case, and proportional to t in the second
case, although the same number of neutrons are absorbed in both cases. The
absorption represents a measuring process in both cases, i.e. a beam path
detection, because compound nuclei are produced with an excitation energy of
several MeV, which are usually deexcited by capture gamma rays. The measured
contrast lies along the lines ”stochastic“ and ”deterministic“ of Fig. 5
[30,31]. The different contrast becomes especially obvious for low transmission
probabilities. The discrepancy diverges for t
but it has been shown
that in this regime the variations of the transmission due to variations of the
thickness or of the density of the absorber plate have to be taken into account
which shifts the points below the
- (”stochastic“) curve [32].

Fig. 5.
Lattice absorber in the interferometer approaching the classical limit
when the slits are oriented horizontally and the quantum limit when they are
oriented vertically [34].
The region between the
linear and the square root behaviour can be reached by very narrow chopper
slits or by narrow transmission lattices, where one starts to loose information
of through which individual slit the neutron went. This is exactly the region
which shows the transition between a deterministic and a stochastic situation
and, therefore, it can be formulated by a Bell-like inequality (
, [33]).
The stochastic limit
corresponds to the quantum limit when one does not know anymore through which
individual slit the neutron went. Which situation exists depends how the slit
widths l compares to the coherence
lengths in the related direction. In case that the slit widths become
comparable to the coherence lengths, the wave function behind the slits show
distinct diffraction peaks which correspond to new quantum states (n 0). The creation of the new
quantum states means that those labbeled neutrons carry information about the
chosen beam path and, therefore, do not contribute to the interference
amplitude [34] (Fig. 5). A related experiment has been carried out by
rotating an absorption lattice around the beam axis where one changes from
(vertical slits) to
(horizontal slits).
Thus, the attenuation factor t
has to be generalized including not only nuclear absorption and scattering
processes but also lattice diffraction effects if they remove neutrons from the
original phase space.
The partial absorption and
coherence experiments are closely connected to the quantum duality principle
which states that the observation of an interference pattern and the
acquisition of which-way informations are mutually exclusive. Various
inequalities have been formulated to describe this mutual exclusion principle
[35,36]. The most concise formulation reads as
(25)
where V denotes the fringe visibility (equ.
(5)) and P is the predictability of the way through the interferometer, which
is a quantitative measure of the a priori which-way knowledge.

Fig. 6.
Sketch of various postselection methods.
3 Postselection
experiments
Various post-selection measurements in
neutron interferometry have shown that interference fringes can be restored by
proper filtering methods even in cases when the overall beam does not exhibit
any interference fringes due to spatial phase shifts larger than the coherence
lengths of the interfering beams [37-38]. Postselection procedures can be
applied to various parameters of an experiment:
· spatial
postselection
· momentum
postselection
· counting
statistic postselection
· phase
postselection
· topology
postselection
Figure 6 shows some of
them schematically. Here we discuss momentum postselection and phase-echo
experiments and refer the reader for other methods to the literature [39-41].
3.1 Postselection
of momentum states
The experimental arrangement with an
indication of the wave packets at different parts of the interference
experiment is shown in Figure 6. An additional monochromatization is
applied behind the interferometer by means of a single crystals brought into
Bragg-position or by time-of-flight systems. Adding the plane wave interference
pattern (equ. (3)), the momentum-dependent intensity reads as:
. (26)

The spatial
phase shift-dependent intensity is given by eq. (5). The formula show that the
overall interference fringes disappear for spatial phase shifts much larger
than the coherence lengths
. The surprising feature is that I0(k) becomes
oscillatory for large phase shifts where the interference fringes described by
eq. (5) disappear ([37], see Fig. Fig. 7). This indicates that interference in
phase space has to be considered [42].
Fig. 7.
Interference pattern as a function of the relative phase shift (middle)
and related wave packets and momentum spectra behind the interferometer for
different values of the phase shift [37].
The amplitude function of
the packets arising from beam paths I and II determines the spatial shape of
the packets behind the interferometer
(27)
which separates for large phase shifts into
two peaks (Fig. 7). For an appropriately large displacement (>>c), the related state can be interpreted as a
superposition state of two macroscopically distinguishable states, that is a
stationary Schrödinger cat-like state [43,44], but here first for massive
particles. These states - separated in ordinary space and oscillating in
momentum space - seem to be notoriously fragile and sensitive to dephasing
effects [44-48].
Measurements of the
wavelength spectrum were made with a silicon crystal with a rather narrow
mosaic spread which reflects in the parallel position a rather narrow band of
neutrons only (k´/k0 0.0003) causing a restored visibility at large phase
shifts [38] (Fig. 8). This feature shows that an interference pattern can
be revived even behind the interferometer by means of a proper postselection
procedure. In this case the overall beam does not show interference fringes
anymore and the wave packets originating from the two different beam paths do not
overlap. The momentum distribution has been measured by scanning the analyzer
crystal through the Bragg-position. These results clearly demonstrate that the
predicted spectral modulation (eq. (26)) appears when the interference
fringes of the overall beam disappear. The modulation is somehow smeared out
due to averaging processes across the beam due to various imperfections,
unavoidably existing in any experimental arrangement. The contrast of the empty
interferometer was 60%.

Fig. 8.
Interference pattern of the unfiltered overall beam (dk/k0 = 0.012,
middle) and the filtered beam reflected from a nearly perfect crystal analyzer
in the antiparallel position (dk'/k0 = 0.0003,
left) and the observed spectral modulation (right) of the outgoing beam for
different phase shifter thicknesses [38].
The new quantum states
created behind the interferometer can be analyzed with regard to their
uncertainty properties. Analogies between a coherent state behavior and a free
but coherently coupled particle motion inside the interferometer have been addressed
[31]. In such cases, the dynamical conjugate variables x and p minimize the
uncertainty product with identical uncertainties (x)2 = (k)2 = 1/2 (in dimensionless
units). Simple calculations show that for (k)2 a value below the coherent
state value can be achieved, which in quantum optic terminology means squeezing
[49-52]. One emphasizes that a single coherent state does not exhibit
squeezing, but a state created by superposition of two coherent states can
exhibit a considerable amount of squeezing. Thus highly nonclassical states are
made by the power of the quantum mechanical superposition principle.
3.2 Contrast retrieval by phase echo
A large phase shift ( > c) can be applied
in one arm of the interferometer, which can be compensated by a negative phase
shift acting in the same arm or by
the same phase shift applied to the second beam path [53]. Because the phase
shift is additative, the coherence function depends on the net phase shift
only. Thus, the interference pattern can be restored as it is shown in form of
an experimental example in Fig. 9. The phase-echo method can also be
applied behind the interferometer loop when multiplate interferometers are used
[41]. In this case, the situation becomes even more similar to the situation
discussed in the previous section. The experimental results completely
confirmed that behaviour. Phase echo is a similar technique to spin echo [3],
which is routinely used in neutron spectroscopy and which represents an
interference experiment as well.

Fig. 9.
Loss of contrast at high interference and its retrieval by an opposite
phase shifter inserted into the same beam [53].
4 Phase
space coupling
In the previous sections one noticed that
coherence phenomena can be exchanged between various parts of the phase space.
The appearing modulation of the momentum distribution when the spatial
interference pattern disappears may be the most direct evidence of this
phenomena (Fig. 8) [38]. In quantum optics, many phenomena are visualized
by Wigner quasi-distribution functions, which are defined as [54,13]
(28)
where in our case
(29)
we arrive at
![]()
(30)
![]()

Integration over the momentum variable gives
the spatial distribution (eq. (5)) and integration over the spatial variable
gives the momentum distribution (eq. (26)). Typical results are shown in Figure
10 [55].

Fig. 10.
Wigner functions for various phase shifts without (left) and with
(right) fluctuations in the phase shifter [55].
When fluctuations of the phase shifter (dN or dD0) are
included, one notices that the wiggle structure between the separated peaks is
more sensitive at high interference order than at low order. This causes a
decrease of the coherence and a transition from a pure quantum state to a
mixture. As a result, upper limits for the separation of massive
(Schrödinger-cat) systems due to unavoidable zero-point fluctuations can be
derived. It also indicates why the retrieval of a quantum state from one phase
space to another one becomes intrinsically more difficult the larger the
separations in one phase space happened.
5 Topological
effects
Topological and geometrical effects appear in
the solution of the Schrödinger equation due to special geometric forms of the
interaction [56-58]. Thus they are part of quantum mechanics but they are
easily overlooked by a pure intensity experiment. It also shows that a
wavefunction often carries more information than those extracted in a standard
experiment. A typical example is the spin superposition experiment discussed in
Chapt. 2.4 where the exacted result also depends around which axis the spin has
been rotated into the opposite direction. Wagh et al. [59] did recently a
related experiment and showed clearly the existence of the topological phase.
In a similar sense the scalar and the vector Aharonov-Bohm effects of neutrons
have been verified by neutron interferometric methods [60,61]. The geometric
nature can be seen in Fig. 11 in comparison to the well-known electron
Aharonov-Bohm effects.
The concept of topological
phases can be extended to the description of absorption phenomena as they are
discussed in Chapt. 2.6. Increasing absorption can be attributed to a more
particle-like behaviour (see equ. (25)) and the state moves away from the equator
to a latitude circle of the Poincaré sphere with a solid angle W related to the geometrical
phase fg
. (31)

Fig. 11.
Comparison of various Aharonov-Bohm phenomena for electrons and
neutrons.
This can be detected by a four-plate double
loop interferometer as done by Hasegawa et al. [62] (Fig. 12). Good
agreement between the theoretical predictions and the experimental results has
been achieved.
6 Discussion
It has been shown in the previous sections
that more information about a quantum system can be extracted when more
experimentally accessible parameters are measured. It becomes obvious that a
system remains coupled in phase space even when it becomes separated in any
parameter space. Thus, interference properties can be shifted from one
parameter space to another one and back again. Related bands of plane wave
components which compose the wave packets may be considered as a responsible
factor for the understanding of the coupling and non-locality phenomena in quantum
mechanics. It looks like that these plane wave components of the wave packets,
i.e. narrow bands, interact over much larger distances than the size of the
packets. This interaction guides neutrons of certain momentum bands to the 0-
or H-beam, respectively. These phenomena throw a new light on the discussion on
Schrödinger-cat-like situations in quantum mechanics [63,64]. Spatially
separated packets remain entangled in phase space and nonlocality appears as a
result of this entanglement. The analogy with optical experiments performed in
the time-frequency domain is striking [65].
The summaries drawn for the different
experimental situations discussed in this article are followed by a statement
that the retrieval of the interference properties by several post-selection
procedures become increasingly more difficult the wider the separation of the
quantum system happened before. This is also demonstrated by means of the
Wigner distributions (Sec. 4, Fig. 10), where it has been shown that the
transition from a quantum state to a mixture is always related to some
statistical features of the interaction acting on the quantum system. Such
fluctuations are, in principle, unavoidable due to residual quantum
fluctuations inherent to any physical system.


Fig. 12.
Poincaré presentation of the absorption topological phase and the
experimental arrangement of the related experiment [62].
Unavoidable fluctuations (even zero-point
fluctuations) cause an irreversibility effect which becomes more influential
for widely separated Schrödinger-cat like states. All these effects can be
described by an increasing entropy inherently associated with any kind of
interaction. This also supports the idea that irreversibility is a fundamental
property of nature and reversibility, an approximation only, as stated by
several authors [66-69,47].
All the results of the
neutron interferometric experiments are well described by the formalism of
quantum mechanics. According to the complementarity principle of the Copenhagen
interpretation, the wave picture has to be used to describe the observed
phenomena. The question how the well-defined particle properties of the neutron
are transferred through the interferometer, is not a meaningful one within this
interpretation, but from the physical point of view it should be an allowed
one.
More complete quantum experiments show that a
complete retrieval of all wave components behind an interaction the quantum
system experienced becomes impossible, in principle. This implies on a high
accuracy level a basic non-commutivity of operators A.B B.A.
ACKNOWLEDGEMENTS
Most of the experimental results discussed in
detail have been obtained by our Dortmund-Grenoble-Vienna interferometer group
working at the high flux reactor in Grenoble, and some recent ones stem from
our cooperation with the Columbia-Missouri group working at the MURR-reactor.
The cooperations with these groups and especially the cooperation with
colleagues from our Institute, which are cited in the references, are
gratefully acknowledged.
REFERENCES
1. Maier-Leibnitz,
H., Springer, T., Z. Physik 167 (1962).
2. Rauch,
H., Treimer, W., Bonse, U., Phys. Lett. A47, 369 (1974).
3. Mezei,
F., Z. Physik 25, 146 (1972).
4. Ioffe,
A.I., Zabiyankan, V.S., Drabkin, G.M., Phys. Lett. 111, 373 (1985).
5. Rauch,
H., Werner, S.A. (Editors), "Neutron Interferometry" (Clarendon
Press, Oxford 2000).
6. Bonse,
U., Hart, M., Appl. Phys. Lett. 6, 155 (1965).
7. Bauspiess,
W., Bonse, U., Graeff, W., J. Appl. Cryst. 9, 68 (1976).
8. Rauch,
H., Petrascheck, D., "Neutron Diffraction" (Edited by H. Dachs),
(Springer Verlag, Berlin 1978), Chap.9.
9. Goldberger,
M.L., Seitz, F., Phys Rev. 71, 294 (1947).
10. Sears, V.F., "Neutron Optics"
(Oxford Univ. Press 1989).
11. Feynman, R.P., Leighton, R.B., Sands, M.,
"The Feynman Lectures on Physics", Vol. III (Addison-Wesley 1965).
12. Glauber, R.J., Phys. Rev. 130, 2529
and 131, 2766 (1963).
13. Mandel, L., Wolf, E., "Optical
Coherence and Quantum Optics" (Cambridge Univ. Press 1995).
14. Rauch, H., Wölwitsch, H., Kaiser, H.,
Clothier, R., Werner, S.A., Phys. Rev. A53, 902 (1996).
15. Colella, R., Overhauser, A.W., Werner,
S.A., Phys. Rev. Lett. 34, 1472 (1975).
16. Werner, S.A., Staudenmann, J.L., Colella,
R., Phys. Rev. Lett. 42, 1103 (1979).
17. Staudenmann, J.L., Werner, S.A., Colella,
R., Overhauser, A.W., Phys. Rev. A21, 1419 (1980).
18. D.M. Greenberger, Rev. Mod. Phys. 55,
875 (1983).
19. Bonse, U., Wroblewski, T., Phys. Rev. Lett.
51, 1401 (1983).
20. Horne, M.A., Zeilinger, A., Klein, A.G.,
Opat, G.I., Phys. Rev. A28, 1 (1983).
21. Klein, A.G., Opat, G.I., Cimmino, A.,
Zeilinger, A., Treimer, W., Gähler, R., Phys. Rev. Lett. 46, 1551
(1981).
22. Rauch, H., Zeilinger, A., Badurek, G.,
Wilfing, A., Bauspiess, W., Bonse, U., Phys. Lett. A54, 425 (1975).
23. Werner, S.A., Colella, R., Overhauser,
A.W., Eagen, C.F., Phys. Rev. Lett. 35, 1053 (1975).
24. Wigner, E.P., Am. J. Phys. 31, 6
(1963).
25. Summhammer, J., Badurek, G., Rauch, H.,
Kischko, U., Zeilinger, A., Phys. Rev. A27, 2523 (1983).
26. Badurek, G., Rauch, H., Summhammer, J.,
Phys. Rev. Lett. 51, 1015 (1983).
27. Summhammer, J., Hamacher, K.A., Kaiser, H.,
Weinfurter, H., Jacobson, D.L., Werner, S.A., Phys. Rev. Lett. 75, 3206
(1995).
28. Badurek, G., Rauch, H., Tuppinger, D.,
Phys. Rev. A34, 2600 (1986).
29. Josephson, B.D., Rev. Mod. Phys. 46,
251 (1974).
30. Summhammer, J., Rauch, H., Tuppinger, D.,
Phys. Rev. A36, 4447 (1987).
31. Rauch, H., Summhammer, J., Zawisky, M.,
Jericha, E., Phys. Rev. A42, 3726 (1990).
32. Namiki, M., Pascazio, S., Phys. Lett. 147A,
430 (1990).
33. Bell, J., Physics 1: 195 (1965).
34. Rauch, H., Summhammer, J., Phys. Rev. 46,
7284 (1992).
35. Wooters, W.K., Zurek, W.H., Phys. Rev. D19,
473 (1979).
36. Englert, B.-G., Phys. Rev. Lett. 77,
2154 (1996).
37. H. Rauch, Phys. Lett. A173, 240
(1993).
38. Jacobson, D.L., Werner, S.A., Rauch, H.,
Phys. Rev. A49, 3196 (1994).
39. Bauspiess, W., Bonse, U., Rauch, H., Nucl.
Instr. Meth. 157, 495 (1978).
40. Zawisky, M., Rauch, H.,
Hasegawa, Y., Phys. Rev. A50, 5000 (1994).
41. Heinrich, M., Petrascheck,
D., Rauch, H., Z. Physik B72, 357 (1988).
42. Schleich, W., Wheeler, J.A., Nature 326,
574 (1987).
43. Legett, A., Proc. Found. Quantum Mechanics,
(edited by S. Kamefuchi) (Phys. Soc. Japan, 1984), p.74.
44. Schleich, W., Pernigo, M., Fam Le Kien,
Phys. Rev. A44, 2172 (1991).
45. Walls, D.F., Milburn, G.J., Phys. Rev. A31,
2403 (1985).
46. Zurek, H., Physics Today, Oct. 1991, p.36.
47. Blanchard, P., Jadczyk,
A., Phys. Lett. A175, 157 (1993).
48. Buzek, V., Keitel, C.H., Knight, P.L.,
Phys. Rev. A51, 2594 (1995).
49. Walls, D.F., Nature 306, 141 (1983).
50. Loudon, R., Knight, P.L., J. Mod. Opt. 34,
709 (1987).
51. Janski, J., Vinogradov, A.V., Phys. Rev.
Lett. 64, 2771 (1990).
52. Wolf, E., Phys. Rev. Lett. 63, 2220
(1989).
53. Clothier, R., Kaiser, H., Werner, S.A.,
Rauch, H., Wölwitsch, H., Phys. Rev. A44, 5357 (1991).
54. Wigner, E.P., Phys. Rev. 40,
749 (1932).
55. Rauch H., and Suda, M.,
Appl. Phys. B60, 181 (1995).
56. Pancharatram, S., Proc.
Indian Acad. Sci. A44, 247 (1956).
57. Berry, M.V., Proc. Royal
Soc. London A392, 45 (1984).
58. Aharonov, Y., Anandan, J.,
Phys. Rev. Lett. 58, 1593 (1987).
59. Wagh, A.G., Rakhecha,
V.C., Summhammer, J., Badurek, G., Weinfurter, H., Allman, B.M., Kaiser, H.,
Hamacher, K., Jacobson, D.L., Werner, S.A., Phys. Rev. Lett. 78, 755
(1997).
60. Allman, B.E., Cimmino, A.,
Klein, A.G., Opat, Kaiser, H., G.I., Werner, S.A., Phys. Rev. Lett. 68,
2409 (1992).
61. Cimmino, A., Opat, G.I.,
Klein, A.G., Kaiser, H., Werner, S.A., Arif, M., Clothier, R., Phys. Rev. Lett.
68, 380 (1989).
62. Hasegawa, Y., Zawisky, M.,
Rauch, H., Ioffe, A.I., Phys. Rev. A53, 2406 (1996).
63. Einstein, A., Podolsky,
B., Rosen, N., Phys. Rev. 47, 777 (1935).
64. Greenberger, D.M., Horne,
M.A., Zeilinger, A., in: "Bell's Theorem, Quantum Theory and Conceptions
of the Universe"; (edited by M. Kafatos) (Kluwer Publ. 1989) p.69.
65. Zou, X.Y., Grayson, T.P., Mandel, L., Phys.
Rev. Lett. 69, 3041 (1992).
66. Lorentz, H.A.,
"Theorie der Strahlung", (Akad. Verlagsges. Leipzig 1927).
67. Prigogine, I., Proc. Ecol.
Phys. Chem., (Siena, Elsevier, Amsterdam 1991) p. 8.
68. Haag, F., Comm. Math.
Phys. 123, 245 (1990).
69. Komo, N., Machida, K.,
Namiki, M., Pascazio, S., Phys. Rev. A54, 1064 (1996).