Assuming you already have the optical setup built around, aiming for the least amount of expensive chemicals, you may try semicrystalline extruded polymers, e.g. polyethylene foils, or PET bottles.
The «trick» is that you do not melt them, but that you reach just the glass transition temperature ($T_g$) which is below melting. Here, crystalline parts in the polymer break up, rendering the material isotropic and in transmittent observation with crossed polarizer / analyzer black. Depending on material and degree of polymerization, this solid-solid transition already may happen in a range of 60 to 80 Celsius. Put the material on a glass slide which is heated gently from beneath with an low-powered heating plate. With a tooth-pick you may test that the material becomes softer.
Upon cooling you will see how spherulites re-appear, e.g.
Size and shape of them will depend on the rate of cooling, for example if let cool at air (or on the table of your microscope), or if you cool the glass slide by putting it on an ice bag.
If not overheated this may be repeated over and over easily for a dozen of times. Apart from the heating plate, there is little potential danger.
A low-cost set up including a heating stage, where you could continue the observation of the sample's recrystallization was described in the February issue of the Journal of Chemical Education (J. Chem. Educ., 2019, 96, 823-826, doi 10.1021/acs.jchemed.8b00879).