Self-Assembly and Surface Patterning of Polyferrocenylsilane-Functionalized Gold Nanoparticles. Macromolecular Rapid

Chemical and topographic surface patterning of inorganic polymer-functionalized nanoparticles (NPs) and their self-assembly in nanostructures with controllable architectures enable the design of new NP-based materials. Capping of NPs with inorganic polymer ligands, such as metallopolymers, can lead to new synergetic properties of individual NPs or their assemblies and enhance NPs processing in functional materials. Here, for gold NPs functionalized with polyferrocenylsilane, we used two distinct triggers to induce attraction between the polymer ligands and achieve NP self-assembly or topographic surface patterning of individual polymer-capped NPs. Control of polymer-solvent interactions was achieved by either changing the solvent composition, or by the electrooxidation of polyferrocenylsilane ligands. These results expand the range of polymer ligands used for NP assembly and patterning and can be used to explore new self-assembly modalities. The utilization of electrochemical polymer oxidation stimuli at easily accessible potentials broadens the range of stimuli leading to NP self-assembly and patterning.


Introduction
Polymer-functionalized nanoparticles (NPs) have a broad range of applications, including colloidal stabilization, [1] chemical and biological sensing, [2] imaging, [3] and medical diagnostics and therapeutics. [4]In comparison with individual NPs coated with a uniform layer of polymer ligands, the self-assembly (SA) and/or the surface patterning (SP) of polymer-capped NPs provide additional versatility and level of control in the design of NPbased functional materials.Patterning of NPs with chemically and/or topographically distinct surface regions renders directionality in their interactions and self-assembly and can be used to explore new SA modes, generate colloidal surfactants [5] and produce templates for the synthesis of multicomponent, hybrid NPs. [6]Self-assembly enables the organization of NPs into nanostructures with increasingly complex architectures, which exhibit new collective optical, magnetic or electronic properties [7] due to the coupling of properties of individual NPs.
Both SA and SP can be realized for NPs capped with end-tethered polymer chains by tuning the relationship between the polymer conformational entropy and the enthalpy of its interactions with a solvent. [8]When a polymer-functionalized NP is transferred from a good to a poor solvent, polymer ligands collapse and associate with each other to reduce the free energy of the system.Figure 1 illustrates the impact of this effect for the SA and SP approaches.In both cases, the initial state of the system are NPs coated with a uniformly-thick layer of strongly solvated end-tethered polymer ligands in a good solvent.At a sufficiently high NP concentration, a reduction in solvent quality leads to the association of polymer ligands on proximal NPs and the formation of interparticle physical bonds, thereby producing self-assembled nanostructures. [6,9,10]Tuning the molecular weight and the grafting density of the polymer ligands, as well as the quality of the solvent enables control over interparticle spacing in the nanostructures, the characteristics that is important in chemical and biological sensing. [11,12]he very same effectthe association of polymer ligands in a poor solventcan be used for the SP of individual NPs.When the probability of NP collisions is reduced in dilute NP solutions, polymer association between molecules attached to the same NP dominates. [13]For polymers that are strongly end-grafted to the surface, surface segregation (also called constrained dewetting) leads to the formation of pinned micelles.This effect is particularly beneficial for the SP of small (~10-20 nm) NPs, which is particularly challenging to achieve.
The number of pinned micelles (forming "surface patches") and their dimensions are thermodynamically controlled and can be changed by varying polymer molecular weight and grafting density, the NP size and the quality of the solvent.
Exploring the range of polymer ligands that can be used for SA and SP with a particular focus on their functionality can lead to a new and broadened range of applications of polymerfunctionalized NPs in nanoscience and nanotechnology.In particular, combining optical and electrochemical properties of inorganic polymer ligands [14] with optical, electronic or magnetic properties of inorganic NPs can lead to new synergetic properties of individual NPs or their assemblies.
In addition, expanding the range of external stimuli for SA or SP can enhance the functionality of such systems.Currently, the reduction in solvent quality for polymer ligands is achieved by either adding a non-solvent to the NP solution in a good solvent, or by changing solution pH. [11]An alternative, potentially useful approach would be a stimulusresponsive change in the hydrophilic/hydrophobic properties of the polymer.
In the present work, we explored the SA and SP approaches for NPs capped with a metallopolymer ligand which combines the properties of metals with the desirable ease of polymer processing. [15][17][18] These polymers have been used in sensing, [19] as an electroactive component in photonic crystals, [20] as ligands for the preparation of colloidal preceramic materials [21], and as an oxygen-plasma resistant block in block-copolymer based lithography. [22,23]Thus, we aimed to exploit the electroactive properties of new PFDMS ligands for the SA and SP of NPs.

Results and Discussion
The SP and SA strategies were explored for gold NPs capped with thiol-terminated PFDMS (see Figure 1).The latter was prepared in two steps which involved i) termination of living anionic PFDMS with a chloro(vinyl)silane followed by ii) attachment of the thiol group via a photochemically mediated thiol-ene "Click" reaction.Full details of the synthesis and characterization of the polymer are provided in Supporting Information.In the present work, two approaches were used to trigger attraction between the PFDMS ligands.In the first method, to reduce the quality of the solvent for the polymer, a non-solvent was added to the solution of PFDMS-capped NPs in a good solvent.In the second method, the solvophilicity of the PFDMS ligands was changed by applying an electrochemical potential to the NP solution in a good solvent and in this manner, changing the oxidation state of the ferrocene groups. [24]e parameter space for the SA and SP for PFDMS-functionalized gold NPs included the change in temperature, the time, the solvent composition and the magnitude of the electrochemical potential.The PFDMS-functionalized NPs exhibited both SA and SP behavior, which yielded single-patch Janus NPs and self-assembled NP clusters.
Gold NPs with dimensions of either 20, or 40 nm were synthesized using a procedure reported elsewhere. [25]Following NP synthesis, the surfactant cetyltrimethylammonium bromide was replaced with thiol-terminated PFDMS with a weight average molecular weight, Mw, of 36 100 g/mol (Figures S1-3, Supporting information) via a ligand exchange procedure.
An aqueous solution of as-synthesized gold NPs was concentrated from 1.5 mL solution to  30 μL using 15 min centrifugation at 5000 g at 27°C and subsequent removal of the supernatant.The concentrated NP solution was sonicated for 5 s and added to 1.5 mL of the dilute solution of thiol-terminated PFDMS in tetrahydrofuran (THF).The resulting solution was maintained undisturbed at room temperature overnight.Then, the NPs were separated from free (non-attached) PFDMS via one cycle of centrifugation of the solution (15 min, 5000 g, 20 °C), removal of the supernatant, and dilution of the solution with THF.Successful exchange of cetyltrimethylammonium bromide with PFDMS was verified by the colloidal stability of the NPs in THF, a poor solvent for cetyltrimethylammonium bromide.In addition, an increase in the apparent hydrodynamic radius of the NPs after functionalization with PFDMS was observed due to the engulfment of the NPs with a polymer corona (Figure S4, Supporting Information).The extinction spectra of the PFDMS-capped NPs did not exhibit a significant shift in the surface plasmon resonance after ligand exchange (Figure S5,

Supporting information).
The PFDMS-capped NPs were transferred from THF to chloroform, a good solvent for PFDMS (a Hildebrand solubility parameter, δ, of 19.0 MPa 1/2 of chloroform is close to that of PFDMS of 18.7 MPa 1/2 ). [26]To reduce the solvent quality, cyclohexane with δ = 16.8MPa 1/2 (a non-solvent for PFDMS), [27][28][29] was added dropwise to the solution of PFDMScapped NPs in chloroform.The volume fraction, φ, of cyclohexane in the solution was varied from 0 to 0.75.Subsequently to incubation of PFDMS-capped NPs in a poor solvent, a droplet of the NP solution was deposited on a carbon-coated copper mesh grid.The NPs were imaged using a Hitachi H-7000 transmission electron microscope.Figure 2a shows transmission electron microscopy (TEM) images of the arrays of PFDMS-capped 20 nm-diameter NPs deposited on a carbon-coated grid from chloroform/cyclohexane solutions with varying φ.In chloroform (φ=0), a uniform 13 ± 3 nm distance between the NPs served as the indication of a PFDMS "shell" acting as a spacer between the gold core.The grafting density, , of PFDMS on these NPs was estimated as [30] -7 - where h is the average height of the brush measured from TEM images (taken as half of the inter-particle distance, h6.5 nm), ρ is the density of PFDMS (ρ1.3 g/mL), [31] and NA is Avogadro's number.For 20 nm-diameter NPs, the estimated  was 0.14 polymer chains/nm 2 .
With increasing φ, the inter-particle distance decreased, consistent with a reducing solvent quality.This effect is quantitatively presented in Figure 2b, where the interparticle distance reduced from 13 to 5 nm with φ increasing from 0 to 0.75, respectively.The reduction in interparticle distance served as an indication of the collapse of the PFDMS ligands in a poor solvent, with a degree of collapse increasing with reducing solvent quality.At high values of , brush collapse in a poor solvent resulted in a uniformly thick polymer shell around NPs and regular interparticle distances.
The results shown in Figure 2c, in conjunction with the results of dynamic light scattering experiments (showing the formation of NP clusters in a poor solvent) (Fig. S6, Supporting Information) pointed to the capability of PFDMS-functionalized NPs to undergo selfassembly, as shown in Figure 1, top and consistent with our previous results for the SA of polystyrene-capped gold NPs in a poor solvent. [8]e SP and SA approaches of PFDMS-capped NPs in the chloroform/cyclohexane solutions was explored at 20, 40, and 60°C.Moderate temperatures 80 o C were required to avoid irreversible NP agglomeration and evaporation of the solvent.For temperatures <60 °C, PFDMS segregation on the NP surface was not observed, even after 12 h-long incubation time (Figure S7, Supporting Information).The optimized conditions -heating the NP solution in a chloroform/cyclohexane mixture at φ0.5 at 60°C for 15 min -were used, similar to the methodology developed in SA experiments for PFDMS-containing block copolymers. [31]e interplay between poor solvent-induced SA and SP was further explored for 40 nmsize of PFDMS-stabilized NPs in solutions with ten-fold difference in NP concentration, CNP.
Nanoparticle concentrations in the solution were determined in ultraviolet-visible spectroscopy experiments, whereby the solution extinction was related to NP concentration by using previously determined molar extinction coefficients for gold NPs. [10]Using Eq. 1, the PFDMS grafting density, , for 40 nm-diameter NPs was estimated to be 0.11 chains/nm 2 , that is, lower than for PFDMS-capped 20 nm NPs.
Figure 3a shows representative TEM images of the self-assembled nanostructures formed by PFDMS-functionalized NPs in the chloroform/cyclohexane mixture at φ=0.5 at CNP=10 nM.In solutions, based on the results of dynamic light scattering, the NPs formed clusters with the average size of 337 ± 40 nm (Figure S6, Supporting Information).This self-assembly mode was similar to the SA of polystyrene-capped gold NPs in solvents with a low dielectric constant. [8] dilute solutions of PFDMS-functionalized NP at CNP=1 nM and φ=0.5 we observed PFDMS segregation into a patch on the surface of individual NPs, as shown in a representative TEM image in Figure 3b.Qualitatively, the morphology of NPs with a patterned surface was similar to that observed for polystyrene-capped NPs in a poor solvent at polymer grafting density below a threshold value. [11]We note that the intrinsic condition of SP experiments was a low value of CNP and thus low-magnification TEM images for NPs with a patterned surface were not helpful due to the low NP density.The average hydrodynamic diameter of patchy NPs was 79 ± 5 nm, comparable to the hydrodynamic diameter of as-prepared PFDMS-tethered NPs.Importantly, due to the lower polymer grafting density on the surface of 40 nm NPs than on 20 nm NPs, PFDMS ligands formed patches, rather than undergoing uniform collapse in a poor solvent, as shown in Figure 2a, in agreement with our earlier work. [13] the next step, to explore the SA and SP strategies for PFDMS-functionalized NPs we utilized an electrochemical stimulus.We rationalized that when an appropriate -9 -overpotential is applied to solutions of PFDMS-capped NPs, the iron centers on the backbone of PFDMS ligands would be oxidized, thereby changing the solvophobicity of PFDMS.
A three-electrode cell comprising a platinum counter electrode, a silver wire reference electrode and a disposable carbon paper working electrode was custom-made.The PFDMScapped NPs were dispersed at a concentration of ~1 nM in dichloromethane with 0.1 M tetraethylammonium tetrafluoroborate (TEATFB) used as a supporting electrolyte.This solution is commonly utilized as a supporting electrolyte solution in electrochemistry experiments of PFDMS. [34]Cyclic voltammetry and chronoamperometry experiments were performed using an Autolab potentiostat with magnetic stirring.As a control system, we used a solution of free PFDMS polymer in 0.1 M solution of TEATFB in DCM.
In Figure 4a, a blue-color cyclic voltammogram (CV) curve displays the reversible twostep oxidation of PFDMS. [22]Analogously to ferrocene, under an applied potential the iron (II) centers in PFDMS underwent one electron oxidation to form iron (III) at the anode.The first and the second oxidation peaks corresponded to the oxidation of alternating and adjacent ferrocene centres, respectively, indicating repulsive interaction between the electroactive centers in the polymer [20,33] The oxidation of iron centers occurred step-wise, because a ferrocenium group increases the oxidation potential of the neighboring unoxidized ferrocene centers and thus make its oxidation energetically unfavourable.Similarly, the reduction of the positive iron centers occurred in two steps and was reversible.
Similarly to free PFDMS, the CV curve of PFDMS-functionalized NPs in 0.1 M TEATFB solution in DCM exhibited the characteristic reversible two-peak oxidation of PFDMS (red trace).The positions of oxidation peak maxima for PFDMS-NPs were similar to those for free PFDMS polymer, however, the reduction peaks were shifted to higher potentials.Further investigation is required to explore the origin of this shift.
In the next step, we conducted chronoamperometric (CA) experiments for both free PFDMS polymer and for the PFDMS-coated NPs in 0.1 M TEATFB solution in DCM (Figure 4b, blue and red traces, respectively).Experiments were carried out at the oxidation potential of +0.7 V, that is, at the maximum for the second oxidation peak.For both systems, the initial current was on the order of tens of microamperes.In the course of the experiment, the current gradually decreased due to the PFDMS oxidation, thereby leading to the loss of colloidal stability of the NPs and their subsequent deposition on the working electrode.This process was marked by a change of the solution colour from light pink to colourless.
After 10 min, the CA process was stopped, a small section of the carbon paper electrode was cut, mounted on a stub and imaged using SEM.Both individual NPs with a PFDMS patch (Figure 4d) and three-dimensional NP clusters (Figure 4c S3 and the caption for peak assignments which are fully consistent with the expected structure).

Effect of temperature on SP and SA of PFDMS-capped NPs
The effect of temperature, T, on PFDMS-capped NPs suspended in the 1:1 vol/vol cyclohexane/chloroform mixture was explored for T of 20, 40, and 60°C for incubation times, t of 15 min, 30 min, 1 h and 12 h.Experiments were conducted as described in Section 4. The samples for TEM imaging were prepared immediately after the completion of the incubation period.
Figure S7 shows TEM images of PFDMS-capped NSs in 1:1 vol/vol cyclohexane/chloroform after 12 h of incubation at 20 and 40°C.Polymer surface segregation did not occur under these conditions, as can be seen from the images.
-26 - ) were observed, being adhered to the carbon fibers of the electrode.At the beginning of the CA experiment, individual patchy PFDMS-coated NPs were present in larger amount, however as the PFDMS oxidation proceeded, a larger number of clusters was observed, as individual PFDMS-capped NPs became colloidally unstable in the electrode vicinity.After 30 min, large-scale NP aggregation dominated.

FigureFigure 1 .
Figure4cshows NP clusters formed by the association of PFDMS-capped NPs due to the loss in solubility of electrooxidized PFDMS next to the electrode and subsequent cluster precipitation of the electrode surface.The appearance of these clusters was similar to that of the assemblies of PFDMS-tethered NPs formed in the cyclohexane/chloroform mixture.The average size of NP clusters formed by electrooxidation-induced SA was 50 ± 13 nm.Figure4dshows single-patch PFDMS-capped NPs.Patches were formed due to the formation of pinned micelles by electrooxidized PFDMS.

Figure 2 .Figure 3 .
Figure 2. Effect of solvent quality on interparticle distance.(a) TEM images of PFDMStethered NPs deposited of the grid from chloroform/cyclohexane solution with varying volume fraction of cyclohexane, ϕ.Scale bars are 50 nm.(b) Variation in the average interparticle distance with ϕ.The 10 nM solution of PFDMS-capped NPs was subjected to heating at 60°C for 15 min.

Figure 4 .
Figure 4. Electrooxidation-induced SA and SP of PFDMS-capped gold NPs (a) Cyclic voltammograms of PFDMS polymer (blue) and PFDMS-functionalized gold NPs (red) in 0.1M TEATFB electrolyte solution in DCM.Sweep rate 200 mV/s.(b) PFDMS oxidation at +0.7 V over 30 min for PFDMS (blue) and PFDMS-capped gold NPs (red) in 0.1M TEATFB electrolyte solution in DCM.(c, d) Representative SEM images of (c) PFDMS-stabilized Au NPs on the carbon paper electrode coated with clusters of PFDMS-NPs after 30 min oxidation at +0.7 V applied potential relative to Ag/Ag+ and (d) individual PFDMS-stabilized Au NPs on the carbon paper electrode after 10 min oxidation at +0.7 V applied potential relative to Ag/Ag+.Scale bars are 250 nm in (c) and 100 nm in (d).

Figure
Figure S3.MALDI-TOF mass spectrum of thiol-terminated PFDMS147.The spacing between the major peaks corresponds to 242, the repeat unit of PFDMS.The spectrum consists of 3 sets of peaks assigned as (on moving to lower molar mass): PFDMS-SiMe2CH2CH2SCH2CH2SH (low intensity), PFDMS-SiMe2CH2CH2S (major peak, loss of 61 amu) and PFDMS-SiMe2 (very low intensity, further loss of 60 amu).

3 .
Figure S4 shows the normalized intensity of dynamic light scattering for CTAB-coated and PFDMS-coated gold NPs dispersed in water and THF, respectively.The hydrodynamic diameter of the NPs increased upon functionalization with PFDMS, indicating a successful ligand exchange.

Figure S4 .
Figure S4.Hydrodynamic diameter of CTAB-and PFDMS-coated gold NPs measured in water and THF, respectively.

Figure S7 .
Figure S7.Representative TEM images of PFDMS-capped NPs after incubation in the 1:1 vol/vol cyclohexane/chloroform mixture at T = 20°C for t = 12 h (left) and at T = 40°C for t = 12 h (right).Scale bars are 500 nm.The insets show high-magnification images of representative PFDMS-coated NPs with no segregation in the polymer shell after incubation.Insets: scale bars are 50 nm.