Stimuli‐Responsive Particle‐Based Amphiphiles as Active Colloids Prepared by Anisotropic Click Chemistry†

Abstract Amphiphiles alter the energy of surfaces, but the extent of this feature is typically constant. Smart systems with amphiphilicity as a function of an external, physical trigger are desirable. As a trigger, the exposure to a magnetic field, in particular, is attractive because it is not shielded in water. Amphiphiles like surfactants are well known, but the magnetic response of molecules is typically weak. Vice‐versa, magnetic particles with strong response to magnetic triggers are fully established in nanoscience, but they are not amphiphilic. In this work colloids with Janus architecture and ultra‐small dimensions (25 nm) have been prepared by spatial control over the thiol‐yne click modification of organosilica‐magnetite core–shell nanoparticles. The amphiphilic properties of these anisotropically modified particles are proven. Finally, a pronounced and reversible change in interfacial stabilization results from the application of a weak (<1 T) magnetic field.

The prototypes for amphiphilic compounds originate from molecular chemistry, like surfactants or block-copolymers. Unique properties and important applications result from the combination of reluctant parts such as hydrophobic and hydrophilic in one molecular entity. The latter is responsible for interfacial activity within amphiphiles and the capability for complex selforganization. [1] The degree of interfacial activity is constant for the vast majority of molecular systems, because it is difficult to alter the molecular configuration to such an extent, amphiphilic properties may become stimuli-responsive. However, amphiphiles that can be turned on and off by demand would be highly interesting.
The higher sensitivity of nanoparticles towards external triggers is only one argument, why particle-based amphiphiles have raised interest. [2] While only few publications can be found for a literature survey on particle-based amphiphiles, an impressive amount of publications exist for particles with so-called Janus design. [3] The classic examples for a Janus particle (JP) involve colloidal beads with the two hemispheres differing in chemical composition. [4] Therefore, the successful preparation of a JP requires some sort of symmetry break. One can roughly differentiate between two preparation approaches. The JP can be assembled from two individual and chemically different species, which are joint together. This approach is also described as compartimentalization. [5] The alternative strategy is a partial (!) modification of the surface of the original particle. This can be achieved by restricting the accessibility of one hemisphere by immobilization of the particle(s) on a surface or by imbedding them in a matrix like a polymer or wax. [6] Clearly, these methods work better the larger and less mobile the particles are. This is one of the reasons, why JPs smaller than 50nm in diameter are hard to find in literature. [7] However, thinking of stimuli-responsive, particle-based amphiphiles, minimization of the size/ mass of the JPs is desirable. A smaller size and mass would increase the sensitivity towards the external trigger and, thus, lead to more versatile active colloids than currently known. [8] As amphiphiles operate in water, it would also be an advantage, if the system could react to a magnetic field, because it is not shielded in an aqueous electrolyte. [9] The latter arguments define the target system for the current work: A nano-sized (< 50nm) JP-based amphiphile, which reacts on an externally applied magnetic field. The blueprint of the target JPs is shown in Scheme 1. Scheme 1. Stimuli-responsive particle-based amphiphiles. Silica encapsulated magnetite nanoparticles are covered with a shell of alkyne-modified organosilica. The anisotropic modification of the alkyne groups by click chemistry leads to JPs. Amphiphilicity originates from suitable selection of R (= hydrophobic group). The application of a magnetic field is expected to orient the magnetic dipoles of the superparamagnetic cores, which induces an additional force. This could have an influence on the way, the particles stabilize the interface.
Two, preliminary achievements are important to mention. Others have published impressive work on the encapsulation of single-domain magnetite nanoparticles in silica. [10] The advantage of a silica shell is, it is chemically robust, the particles become dispersible in polar solvents and, last but not least, many routes are known for the modification of silica surfaces. [11] Our group could recently establish an organosilica system, which is suitable by modification using the so-called click chemistry. [11a, 12] The advantages of click chemistry have been discussed in several review articles and could not be better represented than by the Nobel Prize in 2001. [13] Monodomain magnetite (Fe3O4) nanoparticles (NPs) with a diameter of 9.8 nm (polydispersity index (PDI) = 6%) were prepared according to a method published by Sturm et al. in 2017. [14] The unambiguous characterization is summarized in Supporting Information Fig. S1. Next, we adapted the method published by Ding et al. for encapsulating the Fe3O4-NPs with a shell of silica. [11c] The resulting particles are shown in Supporting  Information Fig. S2. The thickness of the SiO2 shell can be minimzed by controlling the amount of tetraethyl orthosilicate (TEOS) added to the synthesis mixture. The smallest value of the SiO2 layer is only 1 nm.
The alkyne-containing shell was prepared by sol-gel process with a novel, phenylacetylene-bridged silsesquioxane precursor (4) shown in Scheme 2 (see also the experimetal section). We start from the phenylbromide-bridged sol-gel precursor (1), which can be transformed to multiple other groups exploiting the full potential of aromatic substitution chemistry. [11a, 11b, 12a, 12b, 15] Here, the bromide is substituted by a trimethylsilyl-protected acetylene (2) catalyzed by copper(I) iodide/ 1,1´-bis(diphenylphosphino) ferrocenedichloropalladium(II). [16] The trimethylsilyl group is removed from (3) by treatment with AgNO3 leading to the new solgel precursor (4). (4) was characterized by 1 H-, 13 C-and 29 Si-NMR spectroscopy (Supporting Information Fig. S3). The NMR spectra prove the purity of the compound, in particular the presence of only one 29 Si-signal at -62.93 ppm. The signals at 3.08 ppm/ 84.51 ppm and 77.36 ppm ( 1 H, 13 C) show the presence of the acetylene function. The successful synthesis is also demonstrated by electro spray ionization mass spectrometry (ESI-MS) (Supporting Information Fig. S3). Hydrolysis and condensation lead to the phenylacetylene containing organosilica material (AlkySil). The chemical nature of the material was characterized by 13 C-and 29 Si-NMR, ATR-IR, FT Raman spectroscopy and thermogravimetric analysis (TGA) (see Supporting Information The final, core-shell-shell particles were prepared as described in detail in the experimental section (Fig. 1a). The size and morphology of the synthesized Fe3O4/SiO2/AlkySil particles were characterized by TEM (see Fig. 1b). The particle diameter increased from 20.8 nm (Fe3O4/SiO2 particles, PDI = 8%) to 22.9 nm (PDI = 5%). The successful formation of the AlkySil shell is confirmed by the different imaging contrast seen in TEM. In TEM the particles seem to be agglomerated. However, this can be excluded by analytical ultracentrifugation (AUC) measurements. The resulting particle size distribution curve (dh,max= 26.1nm) is shown in Fig. 1c, and is consistent with the TEM data when compared to isolated particles. Complete characterization of the particles is summarized in Supporting  Information Fig. S5. The alkyne-functionalized core-shell-shell particles can be further modified via two different types of click reactions, the Huisgen cycloaddition and the Thiol-Yne click reaction. [17] Both types were used for the isotropic modification of the particles as a proof-of-concept. An example for the Cu-catalysed Huisgen cycloaddition of the dye Cumarin-343 is shown in Supporting  Information Fig. S6. The photochemical Thiol-Yne variant discussed is demonstrated using mercaptoundecanoic acid (Fig. 2). As it is not possible by electron microscopy to visualize the presence of the organic constituents directly, Zn 2+ ions were coordinated to the carboxylic acid groups to increase the imaging contrast. A bright field STEM image is shown in Fig. 2b. Compared to the Fe3O4/SiO2/AlkySil particles (Fig. 1b), a thin, dark rim at the exterior surface can be clearly identified. An energy-dispersive X-ray spectroscopy (EDX) linescan (see Fig. 2c) confirms that this dark rim consists of Zn. Additional data is shown in Supporting Information Fig. S7. The reference experiment, where Fe3O4/SiO2/AlkySil particles are treated with Zn 2+ (Supporting Information Fig. S8) proves that the coordination of the metal at the surfaces is due to the successful click reaction with mercaptoundecanoic acid. The second advantage using the photochemical variant of the Thiol-Yne reaction is that particles deposited as a monolayer on a surface, which is immersed in a solution containing mercaptoundecanoic acid, can easily be radiated only from one side (Fig. 3a). A representative SEM image of such a monolayer is shown in Supporting Information Fig. S9. The additional ZnOlayer (50nm thickness) on the silicon wafer facilitates the detachment of the particles after click-modification. The outcome of the described experiment was analyzed and the results are shown in Fig. 3b, c. It can be seen that the dark rim indicating the presence of Zn is not isotropic in space anymore. Only a segment of 1/4-1/3 has been modified now with mercaptoundecanoic acid, obviously.
After successfully proving the preparation of ultra-small Janus nanoparticle, we can (anisotropically) modify the particles with alternative clickable compounds like pentafluorothiophenol next. The existence of the pentafluorothiophenol entity can be proven by IR-and EDX spectroscopies (Supporting Information Fig. S10). The corresponding JPs were prepared using the methodology described above. For demonstration of the amphiphilic properties of the JPs, the pendant drop method was used and the liquidvapor surface tension (glv = 20.4 mN/m) was calculated (see Supporting Information Fig. S11). However, more interesting is, if and how the amphiphilicity changes in the presence of a magnetic field. Experiments were performed using the sessile drop method (Fig. 4a). A magnet (< 1T) is placed beneath the substrate and the difference in the contact angle Q is recorded. Surface tensions solid-gas (constant), liquid-gas and solid-liquid enter Q. Thus, the spreading of the drop is indicative for a change in the interfacial energy. As nothing has changed, except for exposure to the magnet, this change in surface energy has to originate in a difference in amphiphilicity of the JPs (Scheme 1). The phenomenon can be described as follows: The magnetite cores of the JPs are superparamagnetic. Thus, in absence of an external magnetic field, the moments are randomly distributed. In the presence of a magnetic field, the domains become oriented resulting in a force on the particles (see also Supporting Information Fig. S6). Because one interpretation of the surface tension, is the force per distance acting at the interface, the additional, magnetic influence can eventually change g.  One also has to consider, there are gradients in the magnetic field applied in the experiment described in Fig. 4a. Because particles migrate in such a gradient (Supporting Information Fig.  S6), it can not be excluded there are also local changes of the composition of the interface. Migration does not occur in a homogeneous magnetic field. Instead, rotation happens. The reaction to a homogeneous magnetic field was probed by determination of the Cotton Mouton effect (CME), which is the emergence of birefringence caused by a magnetic field. The CME curve (Fig. 4b) show a rapid and strong answer of the dispersion of the JPs exposed to a weak field (0.05 mT). In addition, dynamic light scattering (DLS) indicates the presence of aggregates of the JPs (Fig. 4c). The diffusion coefficient and the resulting hydrodynamic radius RH were determined paralell and perpendicular to the field. For B = 0T both values are almost equal, which speaks for spherical aggregates like vesicular structures. A distinct anisotropy is found for B = 0.5T, with the extension of the aggregates almost doubled in direction of the field. This speaks for a deformation and/or alignment of the aggregates.
Starting from core-shell-shell nanoparticles, we showed that the Thiol-Yne click reaction can be used for the spatially resolved modification leading to a Janus architecture. The anisotropic modification using hydrophobic moieties leads to particle-based amphiphilic properties. Addressing the magnetite core by an external field changed the inter-particle interaction, which had a direct influence on the degree of amphiphilicity. The system presented herein is not only a versatile platform for new particlebased amphiphiles, considering click chemistry will allow to modify the surfaces with almost any group and the remaining hemisphere of the JPs (Scheme 1) can still be modified differently.
The microwave was heated to 110°C over 20 min, the temperature was held for 45 min before cooling down to 55 °C. The obtained precursor 1,3-Bis(isopropoxysilyl)-benzene-5trimethylsilylacetylene (3) was filtrated and purified via column chromatography. Preparation of Fe3O4/SiO2 core-shell particles A total of 5.0 g Igepal CO-520 dissolved in 110 mL cyclohexane was sonicated for 15 min before adding 10 mL (3.4 mg/mL in cyclohexane) Fe3O4 cores and 0.9 mL conc. ammonia.

S-4
Under continuous stirring 0.5 mL tetraethylorthosilicat (TEOS) were added with 50 mL/h rate of addition. The reaction was stopped by addition of 30 mL methanol and purification by centrifugation in ethanol. The Fe3O4/SiO2 core-shell particles were redispersed in ethanol.

S-26
The negative control experiment (click reaction with Fe3O4/SiO2 particles) shows that the dye is covalently-bound and not just adsorbed on the surface of the particles.