This paper Bouncing Alkaline Batteries: A Basic Solution confirms the causal link between battery bounce characteristics and state of charge for selected battery chemistries and describes the mechanism imvolved. They say:
- The coefficient of restitution of alkaline batteries has been shown to increase as a function of
depth of discharge. In this work, using non-destructive mechanical testing, the change in coefficient
of restitution is compared to in situ energy-dispersive x-ray diffraction data to determine the cause
of the macroscopic change in coefficient of restitution. The increase in coefficient of restitution
correlates to the formation of a percolation pathway of ZnO within the anode of the cell, and that
the coefficient of restitution saturates at a value of 0.63 ± .05 at 50% state of charge when the
anode has densified into porous ZnO solid. Of note is the sensitivity of coefficient of restitution to
the amount of ZnO formation that rivals the sensitivity of in situ energy-dispersive x-ray diffraction
spectroscopy.
and
- In this paper the coefficient of restitution (COR) of an alkaline AA battery is measured at various
depths of discharge by dropping the battery in a controlled fashion and observing the subsequent bouncing. Our
measurements show that this simple bounce test can provide a considerable amount of information of the structure of
the battery’s Zn anode, rivaling the sensitivity of in situ energy-dispersive x-ray spectroscopy (EDXRD) in detection
of ZnO formation. This discovery shows that non-destructive acoustic testing of batteries can provide valuable
information about a battery’s health and state of charge (SOC).
And
- The densification of the anode affects the mechanical properties of the battery. The COR, which measures
the elasticity of a collision between an object and a rigid object, is one such mechanical property that can be determined
as shown in Eq. 5.
COR = sqrt(h2/h1)
where h1 is the drop height, and h2 is the maximum bounce height determined from Eq. 1. Using the bounce test
described previously, the COR of alkaline batteries was measured through full discharge
Authors institutions: Andlinger Center for Energy and The Environment,
Princeton University, , Department of Physics, Rutgers University, Piscataway, New Jersey, University of New York Energy Institute, New York, New York, Department of Chemical Engineering, City College of New York, New York, New York, Sustainable Technologies Division, Brookhaven National Laboratory, Upton, New York, National Synchrotron Light Source, Brookhaven National Laboratory, Upton, New York 11973, USA††
______________________________________________________
It definitely works for Alkaline batteries, which is what the detailed explanation relates to. THe probable reasons are well explained by the information provided by georgechalhoub. I'm just adding the observations of an experienced competent professional observer [tm] :-).
I am an electrical engineer with a Masters degree in EE and 50+ years experience with electronics. (Yes, we did have electronics "way back then"). I, too, was extremely sceptical when I read of this phenomenon. However, testing with dozens (probably ~= 50) of AA Alkaline batteries of a number of brands and states of charge showed and extremely large difference between fully charged and well used batteries. I did not carry out extensive tests with varying degrees of use - batteries here tend to be new, moderately well used or well used. Flattened to within a few mAh of their life tend to be thrown out asap.
As a means of sorting through a bag of mixed batteries to find unused ones this method works superbly. It is probably no quicker than using a test meter, bt hard flat surfaces are often easier to come by than multimeters "in the field".
Here are a series of reasonably well done experiments and battery dis-assemblies in a you tube video. This appears convincing but, of course, could be "cooked" if anyone wanted to spend the very extensive effort required to falsify results. There is no good reason to disbelieve what is shown here, but it's not peer reviewed and doesn't have a recognised institution name on it.