[EDIT 1]
The following is my progression as I attempt to get the answer. Some of it is me thinking out loud.
I've referenced these two documents and only found the following information:
BASF Snap Fit Design Manual
http://web.mit.edu/2.75/resources/random/Snap-Fit%20Design%20Manual.pdf
(Page I1) Although snap-fits can be designed with many materials,
the ideal material is thermoplastic because of its high
flexibility and its ability to be easily and inexpensively
molded into complex geometries. Other advantages
include its relatively high elongation, low coefficient of
friction, and sufficient strength and rigidity to meet the
requirements of most applications.
So I'm aware that thermosplastics are better. But which ones?
(Page III-1) Rigidity can be increased either by using a
higher modulus material (E) or by increasing the cross
sectional moment of inertia (I) of the beam.
Ok so I need to find a list of plastic materials with higher modulus to make it stiffer. But does that correlate with repetitive loading?
(Page III-2) However, as the beam deflection increases, the beam
stress also increases. This will result in a failure if the beam
stress is above the yield strength of the material... The calculated stress or strain value should
be less than the yield strength or the yield strain of the
material in order to prevent failure.
So the ideal plastic needs a high yield strength.
(Page VI-2) Fatigue, or repetitive loading, is the third major cause
of failure. Fatigue concerns primarily apply if hundreds or
thousands of cycles are anticipated. While the design
stress level might be well within the strength of the
material, the repeated application of this stress can
result in fatigue failure at some point in the future. Some polymers perform better than others in this regard,
making them ideal candidates for snap-fits or living hinges
that must flex repeatedly. The first way to avoid a fatigue
failure is to choose a material known to perform well in
fatigue.
What are these ideal candidates?
It continues with the following:
This can be done by comparing the so-called S-N
curves of the materials, which show the expected number
of cycles to failure at various stress levels and at different
temperatures of exposure. The second way, still using the
S-N curves, is to choose a design stress level, at the
correct temperature, that results in the required number of
load applications prior to failure. This method will usually
be conservative since S-N curves are typically generated at much higher
frequencies than would be anticipated for repeated
application of a snap-fit assembly.
Ok so I need to find these curves and compare.
Page IV-4 also contains this table showing allowable strain value.
This table is unclear to me as it appears to mix 70% and 100% yield strain values. If I'm reading it correctly, if the PEI value is 70%, it would be the best. If the PEI value is 100%, Acetal would be the best
Bayer Material Science Snap Fit Joints for Plastics
http://fab.cba.mit.edu/classes/S62.12/people/vernelle.noel/Plastic_Snap_fit_design.pdf
(Page 4) In view of their high level of flexibility,
plastics are usually very suitable materials
for this joining technique.
Yes, plastics.
(Page 11) The permissible deflection y (permissible
undercut) depends not only on the shape but
also on the permissible strain E for the material
used.
In general, during a single, brief snap-fitting
operation, partially crystalline materials may
be stressed almost to the yield point, amorphous
ones up to about 70% of the yield strain.
So assuming a partially crystalline and amorphous plastic have the same yield point, the partially crystalline one would be better because it can flex further.