How about channelling contrast of backscattered electrons? depends of how big your crystalline areas are if this could work. But requires only SEM with BSE detector and not TEM or even something more...
Small angle scattering should reveal the presence of scattering density contrast, although it will not tell you directly if crystals are present.
Synchrotron x-ray diffraction may allow you to increase the signal-to-noise enough to detect a small >~1vol% crystalline fraction.
Optical light scattering may also help.
See e.g.
Wang Z-G, Hsiao BS, Sirota EB, Agarwal P, & Srinivas S (2000) Probing the Early Stages of Melt Crystallization in Polypropylene by Simultaneous Small- and Wide-Angle X-ray Scattering and Laser Light Scattering. Macromolecules 33(3):978-989.
I agree with Oliver that synchrotron X-ray diffraction might help. But even using simple XRD equipment, we have been able to detect small volume fractions well below 1% by limiting measurements to narrow angular domains close to the one where diffraction peaks are expected and then integrating for long time or averaging many measurements (see e.g. K. Chamma, J. Lumeau, L. Glebova, and L. Glebov, "X-ray diffraction study of NaF nano-crystals in photo-thermo-refractive glass ", Journal of Non-Crystalline Solids 405, 188–195 (2014)).
Regarding optical scattering measurement, this can be a good indirect measurement (that is also presented in the mentioned paper) at the condition there is large enough contrast between the glass matrix and the crystals refractive indices.
I've used neutron diffraction in combination with XRD and TEM for looking at nano-domains of certain crystals. Whilst XRD didn't detect the domains, the neutron diffraction did, although the data was still hard to refine. By far the best approach I think is TEM for this sort of thing. Obviously this isn't immediately available to you otherwise you wouldn't be asking the question, but I'd recommend chasing up an institute that has a TEM to help you out.
TEM is available for the external users at the resource centre of SPbSU. If you are interested in the synchrotron measurements, we could discuss it during my visit to St. Petersburg in August.
Hi If you just want to confirm the presence of nanocrystalline droplets (rather than characterise them) then I was able I to detect them by conventional PXRD but needed to run the samples for at least 1.5 hrs (5 - 40 degrees 2 theta, spinning, reflectance, small step size) in aluminosilicates with a d50 of ~ 5 micron to see them on top of the amorphous halo. TEM would be the next method that I would have considered but, as suggested above, if your Tg is low enough (< 700 dC) the delta Cp of a fully amorphous vs partially cry may be able to be detected by DSC at high heating rates (> 100 dC/min) - I haven't tried this but theoretically it should work if you have a fully amorphous material to compare with.
Concerning to quantitative XRD analysis of amorphous and crystalline content in powdered samples by the internal standard method; cf.: https://www.researchgate.net/post/Could_anyone_explain_how_the_amorphous_content_of_a_material_is_analyzed_using_the_internal_standard_method_by_XRD https://www.researchgate.net/post/How_can_I_determine_amorphous_to_crystalline_ratio_in_an_XRD_pattern