The most crusial part is the sample preaperation for TEM as its important to know the absorption edge for the electro beam energy and accordingly the thickness can only bdecided.
So it is rather better to go for Smal Angle X-ray scattering (SAXS0 analysis which does not require any such care towards the sample preparation.
For TEM measurements, the sample should be prepared such as the partilces remain dispersed. If your particles are monodispersed and you want well arranged or self-assempled particles, dpending on the natrure of the particles (e.g. metallic, metal oxide, hydrophilic, hydrofobic, etc.) you have to select solvent in which you disperse them NPs before puting them on a TEM microgrid.
First, you need truly monodisperse NP, dispersion lower than10%, better when lower than 5%. Prepare a clear dispersion of your NPs in a not very volatile solvent, e. g., xylene or toluene for hydrophobic NPs. Place a small drop on a TEM grid and ensure that it does not evaporate too quickly (cold room or put grid in a vial containing a piece of paper soaked in the same solvent or ...). You may want to prepare 2/3 grids using dispersion with different NP concentration. You should observe nice multi-layer ordered NP arrays.
Practical tips: here is what I do usually to obtain high quality images for size measurement purpose. Put a drop of NP (dilute) solution (in toluene), for gold, very pale pink, CdSe, almost colorless on the TEM grid. Hold the very edge of TEM grid with a reverse tweezer, blow air from a quite distance with a heat gun at low-mid temperature. Or you can skip the heat gun part, and just use a very thin kimwipe edge to suck the solvent out slowly , circling the TEM grid, but you need to increase the NP conc.
The above mentioned methods for TEM investigation is for QDs which can be dispersed in a liquid. But if you are referring to QDs on a solid surface (with importance to interface) whose cross sectional image you want to see then its best to try FIB preparation (which is fast) or you can use conventional cross sectional sample preparation technique also. Its not clear to me why you need an ordered array of QDs if you are looking for dispersion in a liquid? To me ordered array only looks good for a publication but even if it is not an ordered array you still can get all the information that you want from TEM and publish it