You coat your nanoparticles with a gel. The DLS method is not suitable for determining the size of such nanoparticles, and here is why. Gels as solutions of ionic micelles of a system in which nuclear quantum effects are present. The proton of alcohol and the hydroxyl group of silica tunnel through a chain of hydrogen bonds and create an unpredictable quantum picture of the diffusion of your particles, and hence the size.
Preprint Nuclear quantum effect in aqueous micellar surfactant solutions
Preprint Supplementary information for Nuclear quantum effect in aque...
To level this phenomenon, it is necessary to measure in solution, 0.5 M NaCl. To do this, it is necessary to measure the refractive index of a solution of alcohol with salt. It is better not to measure the size of nanoparticles using DLS. Each method has its drawbacks.
Usually such results come either from dust (open cuvette, dirty pipette etc) or from aggregation. Adsorption of NPs on a solid surface and analysis by EM or AFM is not ideal... the operator might select a "nice spot" and ignore a big mountain of aggregated NPs! DLS, in contrast, counts all your sample. Generally, my recommendation is to first understand the correlation curve (the raw data in DLS), and then decide. Peaks in the micron range in the zetasizer are in all cases that I have seen artefacts!
A very short note. Before applying the DLS, it is necessary to determine the shape and size of the particles by independent methods, since the Mie theory (the assumption of particle sphericity) is usually used in the algorithm for processing the results in standard instruments.
First disperse NPs of desire concentration in suitable solvent like Ethanol or DMF using ultrasonicator and then measure particles size using DLS. Before measuring sample size, first check in blank solvent. You may also used FE-SEM and XRD for the same measurement.