Structure and characterisation of hydroxyethylcellulose–silica nanoparticles

Functionalising nanoparticles with polymers has gained much interest in recent years, as it aids colloidal stability and manipulation of surface properties. Here, polymer-coated thiolated silica nanoparticles were synthesised by self-condensation of 3-mercaptopropyltrimethoxysilane in the presence of hydroxyethylcellulose. These nanoparticles were characterised by dynamic light scattering, small angle neutron scattering, Nanoparticle Tracking Analysis, Raman spectroscopy, FT-IR spectroscopy, thermogravimetric analysis, Ellman's assay, transmission electron microscopy and cryo-transmission electron microscopy. It was found that increasing the amount of hydroxyethylcellulose in the reaction mixture increased the nanoparticle size and reduced the number of thiol groups on their surface. Additionally, by utilising small angle neutron scattering and dynamic light scattering, it was demonstrated that higher concentrations of polymer in the reaction mixture (0.5–2% w/v) resulted in the formation of aggregates, whereby several silica nanoparticles are bridged together with macromolecules of hydroxyethylcellulose. A correlation was identified between the aggregate size and number of particles per aggregate based on size discrepancies observed between DLS and SANS measurements. This information makes it possible to control the size of aggregates during a simple one-pot synthesis; a prospect highly desirable in the design of potential drug delivery systems.


TGA analysis
To calculate the grafting density, the weight of polymer in the sample was divided by the surface area of the raw silica particle (based on the size determined by TEM/SANS analysis; 2827 nm 2 ). It should be noted that due to the presence of free HEC in the sample, there may be some error in this calculation.

Small Angle Neutron Scattering
Following collection and data reduction, all SANS data was modelled using the SASview software package (www.sasview.org).
Initially, data were modelled to a spherical form factor (Equation S1), however the results yielded a poor fit at higher q-values, suggesting a core-shell type structure. The likely reason for this is the presence of either free HEC or a shell consisting of HEC.

Equation S1
( ) = Where scale is volume fraction, V is the volume of the scattering object, r is the radius, bkg is the background, and Δρ is the scattering length density contrast factor (i.e. difference in SLD between the sphere and surrounding solvent).
To account for this addition, a core-shell sphere model was used (Equation S2).

Equation S2
( Where scale is a scaling factor, Vs is the volume of the outer shell, V c is the volume of the core, r s is the radius of the shell, r c is the radius of the core, ρ c is the SLD of the core, ρ s is the SLD is the shell, ρ solv is the SLD of the solvent, and bkg is the background. The fits for this model can be found in Fig  SI2, and a table for the parameters in Table SI1. Given the uncertainty of how the HEC is interacting with the silica nanoparticles, the SLD parameter for the core and shell were left as floating variables, and SLD for solvent was fixed at 6.35x10 -6 Å 2 . 1.38x10 -6 4.6x10 -6 4.6x10 -6 3.7x10 -6 1.38x10 -6 SLD shell (Å 2 ) 1.38x10 -6 7.9x10 -6 6.7x10 -6 6.0x10 -6 5.2x10 -6 SLD solvent (Å 2 ) 6.35x10 -6 6.35x10 -6 6.35x10 -6 6.35x10 -6 6.35x10 -6

Estimation of number of particles per aggregate.
Using the aggregate size from DLS which includes the hydration shell around all of the particles in the aggregate, the volume of the hydrated aggregate can be calculated.
To estimate the number of particles in each aggregate, the hydrated aggregate volume from DLS is simply divided by the hydrated particle volume. Individual particle size data for each HEC concentration is available from the SANS or TEM diameters; SANS data was selected. However, SANS provides the sizes of the particle cores and so 20 nm (as the mean hydration layer thickness) was added to the SANS size data to account for the hydration shell we expect around each particle in the aggregate. From this, the volume of each hydrated particle was calculated. Dividing the hydrated aggregate volumes by hydrated particle volumes gave an estimate of the number of particles in the aggregate.
Clearly there are numerous assumptions in this approach. We assume the hydration shells of the particles within the aggregates remain at 20 nm thick and that the particles don't "share" the layer in the aggregates (i.e. the hydration shells may not be 20 nm between particles), we assume perfect dense packing and have taken SANS values rather than sizes from TEM which could provide some minor discrepancies. Not withstanding these caveats, the estimates illustrate that the higher HEC concentrations allows greater number of particles to aggregate.