Thermoluminescence (TL) methods might perturb the crystal lattice at a level high enough to displace some of the molecular units in the unit cells by sheer increased thermal vibration.  Recall that the crystalline state depends upon the strict ordered repetition of unit cell molecules - if such order and repetition are perturbed, crystallinity will be lost.  I am less familiar with the optically stimulated luminescence (OSL) method and will not hazard a response here.

Please note that at doses of few Gy or even higher from photons or betas, crystal structure remain unperturbed and same can be verified using XRD prior and after irradiation. Neutrons may cause some damage etc. For photons at  say 100 kGy or above, permanent color centre/lattice damage etc occurs. It varies from material to material.

There are many advantages in OSL as no heating so no IR background, no thermal quenching, even polymers can be used for dosimeter manufacturing etc. 

Generally blue light or green etc is used.Blue has  energy of 2.65 eV. So deep traps highly responsible for residual dose may not be erased and also has (if) less photoionization cross-section but will compete generally at higher doses. So not a problem in individual radiation dosimetry but surely for dating etc. where sensitization occurs.

In a nutshell:

- the OSL detector can be erased using a different wavelenght light than for reading. This allows you to use optical bleaching for your detectors. I guess your interest is to bleach a number of probes in order to use them for autocalibration for dating.

- TL techniques can, indeed, damage the crystal structure if you use inapropriate heating. But this can be avoided if you establish a proper heating protocol that will not go above a certain temperature.

Generally, for dating it is better to use the two techniques in parallel, if you have the possibility (specialised readers, like the one from RISOE, allow you to use both OSL and TL)

I use Aluminum Oxide. It is possible released all of  electron. Very high deep trap do nor release Photon energy. But it is not touch to dosimetry. 

TLD is very old technology, it is first generation solid dosimetry. OSL is second generation. I recommend if do you possible use OSL, you should use OSL.

TLD reader control is very difficult, but OSL reader easy control to light  source. You carefully select optically filters.  

And OSL was damage the material.

Agreed with Ikou. OSL gives flexibility.