One of the suitable methods is to prepare the cross section by chemical etching. In this case you will get smooth surface that you can observe by the electron microscope.

If the etch rate is different across the cross section of the layers it will automatically delineate the layers. Otherwise in order to increase the contrast you can bias it by a DC voltage.

Best wishes

How about a diamond wire saw?

I think diamond wire saw will damage the surface, and remain too much dirt.

Usually we use diamond cutter pan: scratching by glass side, and breaking. It gives sometimes good results, but not as beautiful as in publications:)

Oh sure it does damage the surface, the samples cut with a saw have to be polished for microscopy. If you have the equipment for TEM preparation, this could at least attempted - you stack several pieces of your sample on top of each other, embed it in epoxy. Then cut the embedded stack in halves with a wire saw and polish the cutting plane. If available, you could even finish the surface with an ion mill to remove the polishing artifacts.

We usually prepare cross-sections of final devices by FIB milling.This preparation method works both for organic as well as perovskite devices. Of course, it necessitates access to such a type of instrument.

Another suitable and very clean alternative might be the simple cleavage of the glass substrate. Therefore you should scratch the substrate from the back and brake it in a way so that tensile stress is applied to the device side of the substrate: just put the substrate with the device facing upwards onto another glas substrate, so that one half protrudes the edge of the underlying substrate + press.

We usually use the 2nd procedure, but it seems FIB milling is more encouraging. Thank you!

One technique that I understand gives good results in some cases is to freeze the sample in liquid nitrogen and to crack it using simple cleavage of the glass substrate. The low temperature hardens everything and reduces the difference in mechanical deformation between layers. Although if you have access to a FIB then it's probably the most reliable method with the best results.

This is true, I forgot to mention that. Works, if the layered system withstands the thermal shock

There are couple of ways to observe cross section of multilayered systems. Basically if you have no sophisticated tool you can use liquid nitrogen. Put the sample inside the liquid nitrogen wait for 1 min. then put it out and immediately push from the top of the layers with a tweezer or diamond knife (not to thick substrate side) and you will have a clear cross section.

Second option is using FIB.  Make a cut through the layers and then switch ion to electron for imaging. It will take time to cut and also you can see only a small part of it. Besides you need to optimize gun power current etc. I wouldn't use that option.

I have many more suggestions but lastly  i will explain couple more examples. If you can not make the liquid nitrogen breaking work, then embed your samples in a "bakelite" or what else you have. The trick is place the sample vertical to the ground and put couple of samples side by side touching to each other from the same material surface. Grind the surface rough  to fine using a rotating grinding machine or by hand. Don't forget to rotate the sample frequently, use IPA as lubricant. Then, lap the sample using polishing cloths and polishing paste (pick small particle solutions). Check under optical microscope in each stage. Wash it with IPA and dry it with nitrogen. Do not ever touch on the surface.

If you have cryo-ultra-microtome, break the sample in pieces put a tiny piece in teknovoit 7100 resin. Use UV light for rapid curing or wait for couple of days. Depending on the feature thicknesses arrange the cutting speed and ratio. Check the surface under optical microscope in each stage.

These are only examples how can you characterize the cross section of thin film devices such as solar cell. When you put the sample under investigation in SEM I strongly suggest that u take a reference measurement. For instance measure the thicker layers in advance with a different technique. Put the sample back in the SEMDchamber. First measure the known thickness and tilt the stage untill you will get the right dimension. After calibration you are ready to measure thickness of your layers without changing the tilt and working distance.

Any further question will be answered with pleasure.

Regards,

Mehmet Kanik

Microtome surface smoothing and breaking in liquid nitrogen can be seen in my paper (http://pubs.acs.org/doi/abs/10.1021/nn503269b)

But as for perovskite solar cell, I am wondering that whether the liquid nitrogen freezing method is feasible? The extreme temperature not only will hardern everything, but also induce the phase transition of the perovskite materials. You can see the apparent color change of the perovskite from black to yellow and then recover to black. At last, the film is quite bad and rough.

One of the suitable methods is to prepare the cross section by chemical etching. In this case you will get smooth surface that you can observe by the electron microscope.

If the etch rate is different across the cross section of the layers it will automatically delineate the layers. Otherwise in order to increase the contrast you can bias it by a DC voltage.

Best wishes