Schematic of the surface reactions involved in the preparation of the

polypeptide composite particles having a homopolypeptide and a heteropolypeptide shell.

Control over the shell thickness could be assured by the sequential addition of the NCA monomer. The Heise group expanded the work on stimuli-responsive polypeptide-based silica hybrids and investigated the optimal conditions of the polymerization to produce uniform shells.²⁴⁸ Reactions carried out at 0°C seemed to give a core-shell structure whose hydrodynamic radius increased proportionally with the BLG-NCA monomer addition. The same polypeptide was grown on silica enclosing a magnetic (magnetite) nugget. Magnetic inclusions broaden the responsiveness spectrum of the hybrid particle.

Even though the polypeptide loading on the particles is much higher than click particles, the grafting from method suffers from several constrains which will be further addressed. First,

50

the control over the grafting density still needs to be improved. For this reason, the Russo group used the tandem passivator/activator (e.g. methyltrimethoxy silane, MTMS/aminoethyl-3- aminopropyltrimethoxy silane, AEAPTMS, MTMS/ 3-aminopropyltrimethoxy silane, APTMS) in the effort to avoid a densely polypeptide packed surface. Two sets of particles were prepared:

one carrying a shell consisting of PCBL and the other had PBLG.²¹⁸ Methods to demonstrate the packing nature of the shell are limited because the molecular weight of the polymer is not easily accessible unless the hybrid particle is subjected to a “haircut”. This aspect will be developed in a future paragraph. Information about the “densification” of the shell was gathered by a combination of light scattering and computation methods based on the core-shell form factor.

555-558

In this equation, R is the outer radius of the polypeptide silica core shell particle, x= qR, q is the scattering vector magnitude defined as q = 4πn•sin(θ/2)/λo (n is the solvent refractive index, θ the scattering angle and λo the wavelength in vacuum), j1 is the first-order spherical Bessel function given by sin(x)/x²- cos(x)/x. The relative refractive indices were expressed as m₁ = n₁/n₀ and m₂ = n2/n0 with n0, n1 and n2 the refractive indices for the solvent, the shell and the core, respectively.

The term f designates the fraction belonging to the core from the total dimension of the particle: f

= (R-t)/R (t-is the thickness of the shell). Well-solvated particles can be modeled using the refractive index contrast between the solvent, the shell and the core.^556,559 Suspensions of PCBL silica particles in pyridine showed a near agreement between their radius measured by DLS and SLS suggesting good solvation of a shell which was not densely packed. Setting the values of the refractive indices to 1.507 for the pyridine solvent,^560,561 1.544 for the PCBL⁵⁶² and 1.445 for the

51

silica⁵⁶³ adequate contrast was ensured to detect both entities: the core and the shell. In the Rayleigh-Gans-Debye limit^556,564 at zero scattering angle the scattering intensity is given by:

In this expression ν is the number density of particles, dn/dc is the differential refractive index increment and M is the mass. Under these conditions and assuming a hypothetical suspension containing only PCBL shell without core, light scattering at finite angles should return different values for both R_SLS (core-shell) and R_core (silica). The results didn’t reflect this trend pointing once again to a good solvation of the shell, in contradiction with the situation of a highly packed one. Yet, this qualitative assessment couldn’t associate the apparent thickness of the shell with its length even under the assumption that polypeptide is oriented perpendicular to the core surface.

The continuous reactivity of the chain ends was proven by sequential addition of the NCA (BLG-NCA) monomer.²¹⁸ The direct assay of the shell was facilitated by the thermogravimetric analysis. The weight loss at each step increased, indeed not linearly, with the monomer concentration and the results were confirmed by DLS measurements. The lower grafting efficiency could be explained by the formation of the untethered polymer in the polymerization mixture. This aspect was also noted by the Heise group.²⁴⁸ The growth of the copolymers made of homopolypeptides on silica spheres was assessed. The importance of washings to remove the free, untethered polypeptide was reflected in the thermogravimetric curves. Approximately 25% weight loss difference was recorded between the unwashed and the washed sample. The reactions conducted at various temperatures (0°C and 20°C) showed good control over the shell thickness, especially at 0°C. This is in agreement with the findings that at 0°C the NCA doesn’t react with moisture.

52

A sensitive and debatable issue is the determination of the molecular weight of the homopolypeptide grown on the surface. The most popular technique is etching the silica cores by hydrofluoric acid, HF. Soto Cantu et al. reported the ability to control the size of the silica when HF etchant was employed.²⁴⁶ The etching method was adapted to the polypeptide composites from the extensively used treatment of the silica wafers.^565,566 Yet, no detailed study was conducted so far to determine the effect of etching conditions on the polypeptide shell. Questions if the degradation of the polypeptide can occur were asked and answers still wait. The Heise group was the first to combine molecular weights from GPC and mass of shell polymer from TGA to determine the number of polypeptide chains on a particle (0.4 chains/nm² or 2.5 nm²

“parking area” per chain).²⁴⁸ The obtained molecular weights of the PBLG ranged from 7,700 to 20,100 Da for polymerizations conducted at 0°C while the values for polymerizations performed at 20°C varied from 8,700 to 20,100 Da. The polydispersity indexes were 1.4 (0°C) and mostly 1.3 (20°C). The authors claimed that lower temperatures are better but a net difference in the molecular weights values cannot be distinguished. Moreover, the polymerization conducted at 20°C seemed to favor slightly lower PDI’s. Some research groups have shown that polymerization in solvents such as DMF as Heise used can initiate the polymerization of the NCA, consequently competing with the initiation by the amino groups. The DMF initiation, as shown in the Scheme 1.17 was also studied by Kricheldorf.^285,363

This competition arrests the formation of the high molecular weight polymer and yields polydisperse polymers, because the growth of the nascent chains occurs before, during and after the addition of initiator. Yet, the Heise group brought an important contribution to this field by engineering hybrids decorated with a copolypeptide shell. These results demonstrated the potential adaptability of the polymerization to a broader range of amino acids. The continuous

53

reactivity of the polypeptidic chains was consequently proven, which agreed with the Russo group reports.