In a paper from 2006, Couder and Fort [1] describe a version of the famous double
slit experiment performed with drops bouncing on a vibrated fluid surface, where
interference in the particle statistics is found even though it is possible to determine
unambiguously which slit the “walking” drop passes. It is one of the first papers in
an impressive series, showing that such walking drops closely resemble de Broglie
waves and can reproduce typical quantum phenomena like tunneling and quantized
states [2–13]. The double slit experiment is, however, a more stringent test of
quantum mechanics, because it relies upon superposition and phase coherence. In
the present comment we first point out that the experimental data presented in
[1] are not convincing, and secondly we argue that it is not possible in general to
capture quantum mechanical results in a system, where the trajectory of the particle
is well-defined.
In the double slit experiment [1], 75 drop passages of the slits are recorded (their
Fig. 3). This small number is increased by symmetrization, which, however, does
not improve the statistics. Submitting the data to a standard χ
2
-test, a fit to
a Gaussian distribution is found to be just as good as the fit to the Fraunhofer
interference pattern presented in the paper. In addition the blue envelope curve
(single slit result) shown in their Fig. 3 is not backed up by data because the single
slit results presented in the paper (their Fig. 2) are for slits of different widths than
those of their Fig. 3. We have tried to reproduce their results experimentally with
our own double slit set-up, but without success.
The walking drops are reminiscent of de Broglie waves. In his later years de
Broglie [14] took his wave idea further and imagined that particles could be described
as moving singularities in a field, which, in addition to the probabilistic
Schr¨odinger wave function, had a new “physical” component excited locally by the
particle - just like what happens in the experiment. We have tried to implement
this idea by introducing a source term in the standard Schr¨odinger equation, i.e.,
i h ∂/∂t ¯ − Hˆ
w(r, t) = J(r − R(t)) where R(t) is the position of the particle, and
the source term J is a complex function. Due to linearity, it is sufficient to choose
the source term equal to a δ-function, i.e., J(r − R(t)) = δ(r − R(t)). In addition,
the particle is guided by the wave according to the standard Madelung-Bohm equa-