Synthesis and Characterization of Tunable, pH-Responsive Nanoparticle–Microgel Composites for Surface-Enhanced Raman Scattering Detection

Abstract

The synthesis of microgels with pH-tunable swelling leads to adjustable and pH-responsive substrates for surface-enhanced Raman scattering (SERS)-active nanoparticles (NPs). Sterically stabilized and cross-linked latexes were synthesized from random copolymers of styrene (S) and 2-vinylpyridine (2VP). The pH-dependent latex-to-microgel transition and swellability were tuned based on their hydrophobic-to-hydrophilic content established by the S/2VP ratio. The electrostatic loading of polystyrene/poly(2-vinylpyridine) microgels [PS_xP2VP_y (M)] with anions such as tetrachloroaurate (AuCl₄^–) and borate-capped Ag NPs was quantified. The PS_xP2VP_y (M) can load ∼0.3 equiv of AuCl₄^– and the subsequent photoreduction results in Au NP-loaded PS_xP2VP_y (M) with NPs located throughout the structure. Loading PS_xP2VP_y (M) with borate-capped Ag NPs produces PS_xP2VP_y (M) with NPs located on the surface of the microgels, where the Ag content is set by S/2VP. The pH-responsive SERS activity is also reported for these Ag NP-loaded microgels. Analytical enhancement factors for dissolved crystal violet are high (i.e., 10⁹ to 10¹⁰) and are set by S/2VP. The Ag NP-loaded microgels with ∼80 wt % 2VP exhibited the most stable pH dependent response.

Introduction

Surface-enhanced Raman scattering (SERS) is a highly sensitive spectroscopic technique for analyzing and detecting simple to complex collections of molecules or biological structures.^1,2 The sensitivity arises from a SERS effect that results from a large enhancement of the Raman scattering cross section of molecules that are in close proximity to metal surfaces.^3,4 Many theoretical and experimental studies indicate that the excitation of surface plasmons in such metals leads to a localized electromagnetic field that greatly enhances the sensitivity in SERS measurements.⁵⁻⁹ The excitation of surface plasmons depends on factors such as metal type, surface roughness, and for nanostructures, their size, shape, proximity, and packing architecture.^10,11 In certain static arrangements with nano- to molecular-scale interparticle gaps, “hot spots” have been reported which have extremely high SERS sensitivity¹²⁻¹⁵ and the ability to detect single molecules.¹⁶⁻¹⁸ The concept of a SERS substrate capable of dynamically tuning SERS-active particle distances is, therefore, attractive. Example systems include the following: the swelling of gels with embedded nanoparticles (NPs),¹⁹⁻²¹ mechanical stretching of elastomeric substrates bearing silver (Ag) NPs,²² wrinkling of metal and composite films on thermally contracting polymers,²³⁻²⁵ inducing morphology changes of interparticle ligands through pH adjustments,^26,27 and aggregation of NPs within magnetic fields.^28,29

The simplest SERS experiments are performed by delivering analytes to a suspension of metal NPs of Ag or gold (Au). The main drawback for such experiments is that the NPs are prone to aggregation, which reduces the reproducibility of the SERS effect. Accordingly, anchoring SERS-active NPs to water-dispersible polymer substrates has been explored as a mode for stabilizing the SERS sensitivity. Platforms based on polystyrene (PS),³⁰ polyurethane,³¹ polyacrylic acid, and polyallylamine³² have shown promise in this regard. Loading stimuli-responsive microgels with NPs has resulted in interesting composites with new properties such as catalytic activity, optical, magnetic, and pH response.³³⁻⁴³ The use of microgels to stabilize the SERS activity of Ag NPs is less well-explored.⁴⁴ Previously, Armes and Nakamura et al. have studied the pH-responsiveness and Au NP loading of lightly cross-linked and sterically stabilized poly(2-vinylpyridine) (P2VP) microgels.^45,46 The Au NP-loaded P2VP microgels studied in their work demonstrated a reversible pH-sensitive plasmon absorbance. Homopolymers, block copolymers, and random copolymers of 4-vinylpyridine and 2VP are widely exploited for their ability to carry a positive charge at a pH lower than 5.0.⁴⁷⁻⁴⁹ Microgels of P2VP are promising candidates for such a role owing to the multifunctionality of the pyridine residues located throughout the macromolecular network. Aqueous protonation of pyridinylenes in lightly cross-linked P2VP results in significant polymer solvation and swelling, while electrostatic and ligand coordination to metal ions can also occur in P2VP materials.⁵⁰

We present herein the synthesis of sterically stabilized latexes consisting of random copolymers of PS and P2VP [PS_xP2VP_y (L)]. The dimensions of the latex and its transition to a microgel are studied by titration with monitoring by dynamic light scattering (DLS). The loading of the microgel state with Au ions for subsequent Au NP synthesis is explored. The loading of the copolymer microgels with borate-capped Ag NPs is also described. The Ag NP-loaded microgels exhibit SERS activity demonstrated by detection of pyridinyl groups in the polymer support as well as the detection of a dissolved analyte, crystal violet (CV). A reversible pH-dependent response for the intensity of the SERS spectra for CV is found. The most pH-responsive Ag NP-loaded microgel contains nominal P2VP and PS content values of 80 and 20%, respectively, and operates with a high analytical enhancement factor (AEF) of 1.08 × 10¹⁰ and a greater stability than the analogous microgel that lacks the PS content.

Results and Discussion

Synthesis of PS_xP2VP_y Latexes

An emulsion polymerization approach was selected for the synthesis of sterically stabilized pH-responsive microgels^45,46,51 with a tuned hydrophilic: hydrophobic ratio. Scheme 1 outlines the emulsion polymerization that leads to a copolymer latex of S, 2VP, divinylbenzene (DVB), and poly(ethylene glycol)methyl methacrylate (PEGMA). Previous reports have shown that analogous latexes without the PS component are water-dispersible over a large pH range because of the PEGMA-derived PEG-corona on individual particles.^45,46 This PEG corona stabilizes the latex and any resulting microgels from aggregation or precipitation by providing a steric repulsion between particles.⁴⁶ At 60 °C, the monomer reactivity ratios (r) for styrene and 2VP are r_s = 0.55 and r_2VP = 1.14, respectively.⁵² The product of r_sr_2VP is 0.63. Commonly, the range for r₁r₂ is 0–3, with an ideal copolymerization having r₁r₂ = 1. The copolymerization of S and 2VP is therefore classified as moderately ideal.⁵³ Also, because the two r values are similar, the two different monomers can be expected to show equal reactivity toward the propagating radical chain ends in the latex.⁵³ Because the reactivity of the monomers is relatively equivalent, the S/2VP ratio in the latex should closely match the nominal S/2VP monomer synthesis ratio, and these two units will likely be randomly included during polymerization.⁵⁴ Entries 1–5 in Table 1 outline the effect of varying certain synthetic parameters on the mean D_H and polydispersity as determined by DLS. Solvent-free diameters, as determined from scanning electron microscopy (SEM) analyses are also reported in this table. Generally, as the nominal weight percent of 2VP is tuned from 40 to 100%, the D_H values of the latex (D_H,L) at a pH condition of approximately 5 are found to increase from ∼400 to ∼540 nm, while the polydispersity values remained within the range of 0.017–0.063. The increase in diameter with increasing 2VP content suggests that the polarity of neutral P2VP units, and any conjugate acids (P2VP⁺) units, require additional volume in the colloid to hydrate and accommodate additional counter ions. The D_H,L and polydispersity values for these entries in Table 1 are also in agreement with previous and analogous sterically stabilized P2VP latexes.^45,46 An interesting deviation in the trend of increasing D_H,L with increasing nominal P2VP content was found for the PS₄₀P2VP₆₀ latexes [PS₄₀P2VP₆₀ (L), Table 1, entry 3], which may be because of a latex with polymer chains in an unrelaxed conformation owing to the kinetic conditions of the polymerization. After a transition to a microgel, the diameter of the PS₄₀P2VP₆₀ particle is in agreement with the observed trend for others in the series. The SEM images of solvent-free packed colloids [some exhibiting face-centered cubic (FCC) packing] for the PS₀P2VP₁₀₀ (L), PS₂₀P2VP₈₀ (L), PS₄₀P2VP₆₀ (L), PS₅₀P2VP₅₀ (L), and PS₆₀P2VP₄₀ (L) are shown in Figure 1a–e. The characteristic for low polydispersity in the diameter of the particles is also evident in these images where efficient hexagonal packing of highly uniform particles occurs for all of the materials in this series. Because latex particles are known to pack in an FCC arrangement, the layers in Figure 1a–e represent a (111) plane of the FCC lattice. The number average diameters for the dried latexes (D_N,L) were obtained from the SEM images by averaging 5 separate distance measurements, which came from five different and central particles in a close packed seven-particle hexagon. This value is, therefore, determined by measuring the distance between contact points between particles that were in the same plane and row in the packed arrays. In an FCC lattice, these measurements are along the ⟨110⟩ direction (or other geometric equivalents), and hence we provide diameter measurements along this direction with a D_{N,L,⟨110⟩} label. As expected, the D_{N,L,⟨110⟩} values were lower than the D_H,L values found by DLS. The trend for increasing D_{N,L,⟨110⟩} with increasing P2VP content was also found in these data (Table 1).

Scheme 1. Generalized Chemical Equation for the Synthesis of Sterically Stabilized pH-Responsive Latex from a Random Copolymer of Styrene and 2-Vinylpyridine [PS_xP2VP_y (L)] (at Right) with a Tuned a Hydrophilic/Hydrophobic Ratio.

In this work, the molar amounts of the PEGMA steric stabilizer, Aliquat 336 surfactant, and the DVB cross-linker were held constant and in proportion to the total molar amount of the remaining monomers, styrene (S) and 2VP. The hydrophilic/hydrophobic ratio is tuned by adjusting the molar amounts of S and 2VP in the emulsion polymerization.

Table 1. Summary of the Nominal Composition of the PS_xP2VP_y Particles and the Resulting Hydrodynamic Diameter (D_H), Polydispersity, Solvent-Free Number Average Diameter (D_N,L), and the Inorganic Loading Values for the Latex/Microgel.

		compositiona		latex (L) condition			microgel (M) condition
entry	sample name	S (wt %)	2VP (wt %)	number average diameter, DN,L in nmb	intensity average diameter DH,L in nm (pH)c	polydispersityc (pH)	intensity average diameter DH,M in nm (pH)c	polydispersityc (pH)	inorganic loading (wt %)d
1	PS0P2VP100 (L/M)	0	100	338	542 (5.1)	0.063 (5.1)	2409 (3.0)	0.876 (3.0)	0
2	PS20P2VP80 (L/M)	20	80	265	476 (5.0)	0.017 (5.0)	1897 (2.5)	0.095 (2.5)	0
3	PS40P2VP60 (L/M)	40	60	192	318 (5.3)	0.040 (5.3)	1498 (2.6)	0.283 (2.6)	0
4	PS50P2VP50 (L/M)	50	50	235	424 (5.0)	0.049 (5.0)	815 (1.4)	0.028 (1.4)	0
5	PS60P2VP40 (L/M)	60	40	247	396 (5.0)	0.058 (5.0)	531 (1.0)	0.139 (1.0)	0
6	PS20P2VP80 (L/M)/AuCl4–	20	80		713 (7.0)	0.168 (7.0)	959 (2.5)	0.096 (2.5)	42.8
7	PS40P2VP60 (L/M)/AuCl4–	40	60		567 (7.0)	0.135 (7.0)	588 (2.5)	0.149 (2.5)	31.5
8	PS60P2VP40 (L/M)/AuCl4–	60	40		459 (7.0)	0.012 (7.0)	485 (1.1)	0.010 (1.1)	19.6
9	PS20P2VP80 (L/M)/Au NPs	20	80		653 (4.2)	0.291 (4.2)	638 (2.5)	0.301 (2.5)	38.3
10	PS40P2VP60 (L/M)/Au NPs	40	60	228	270 (7.0)	0.130 (7.0)	375 (2.5)	0.218 (2.5)	24.4
11	PS60P2VP40 (L/M)/Au NPs	60	40	264	428 (7.0)	0.068 (7.0)	510 (1.1)	0.106 (1.1)	18.9
12	PS0P2VP100 (L/M)/Ag NPs	0	100	402	423 (10.5)	0.274 (10.5)	843 (2.0)	0.309 (2.0)	39.7
13	PS20P2VP80 (L/M)/Ag NPs	20	80	385	599 (9.1)	0.247 (9.1)	896 (1.8)	0.250 (1.8)	24.6
14	PS40P2VP60 (L/M)/Ag NPs	40	60	268	582 (9.1)	0.238 (9.1)	709 (1.9)	0.276 (1.9)	15.1

^aAll latexes were nominally synthesized with additional PEGMA (9.0 wt %), Aliquat 336 (3.1 wt %), and DVB (1.0 wt %).

^bThe values were estimated by averaging 5 particles in an SEM image. In the case of entries 1–5, the D_N,L values correspond to the diameter along the ⟨110⟩ direction of an FCC lattice (D_N,L⟨110⟩) and were determined by averaging five central and hexagonally packed particle diameters along the ⟨110⟩ direction of an FCC lattice.

^cThe values were estimated from DLS measurements at 20 °C. The polydispersity = |μ₂/Γ²|, where μ₂ is the second cumulant and Γ is the average decay rate determined by the CONTIN analysis.

^dThe values were estimated from the char mass at 775 °C in the TGA plot. In the case of Au- and Ag NP-loaded microgels, the mass is assumed to represent the metal-loading values for the composites.

Figure 1.

SEM images of the dried (a) PS₀P2VP₁₀₀ (L), (b) PS₂₀P2VP₈₀ (L), (c) PS₄₀P2VP₆₀ (L), (d) PS₅₀P2VP₅₀ (L), and (e) PS₆₀P2VP₄₀ (L) samples. (f) Plot for the ratio of the number average diameter (D_N,L) distance along the ⟨121⟩ to ⟨110⟩ directions versus the nominal P2VP weight content. Representative [110] and [121] directions are shown in (a). The scale bars in (a–e) are 2 μm.

The balance of hydrophobic and hydrophilic (PS and P2VP units, respectively) properties control the water content in the latex particles and consequently may lead to plastic deformation in the dried particles. Cross-linked latexes rich in PS retain their spherical shape when dried because they do not incorporate significant amounts of water at a neutral pH. Latex particles rich in P2VP incorporate small amounts of water that act as a plasticizer to permit polymer flow. This polymer flow can cause a distortion from a spherical shape during the drying process. Inspection of the SEM images in Figure 1 reveals an interesting trend in the evolution of the shape of the solvent-free PS_xP2VP_y (L). These images depict near-monodisperse particles from dried latexes that were synthesized with a similar amount of PEGMA and Aliquat 336 (9.0 and 3.1 wt %, respectively). As the nominal weight percent of 2VP is tuned from 40 to 100%, the dried and hexagonally packed particles evolve away from a spherical morphology and take on a distorted shape that arises from gradual occupation of the octahedral and tetrahedral void space in the packed monolayer lattices. This distorted shape is a common characteristic of soft or low glass transition (T_g) polymer latexes that dry in packed arrays.⁵⁵ It is clear that as the nominal weight content of the P2VP increases, the extent of distortion of the spherical latex also increases. This distortion is greatest along the direction that extends from the center of a particle and points toward the voids (for FCC-type packing, the ⟨121⟩ direction bisects octahedral and tetrahedral voids). Conversely, the distortion is minimal along the ⟨110⟩ direction of the lattice where packed spheres contact each other. The distortion is assumed to arise from strong capillary stresses acting on the colloids as water dries from unoccupied space between 3-D packed colloids. The ratio of D_{N,L,⟨121⟩} to D_{N,L,⟨110⟩} represents a metric for quantifying the extent of distortion in the packed spheres. The D_{N,L,⟨121⟩} to D_{N,L,⟨110⟩} ratio is plotted versus the nominal P2VP weight content and shown in Figure 1f. The trend suggests that when the nominal P2VP content exceeds 60 wt %, distortion along the ⟨121⟩ direction becomes significant and can be tuned according to the P2VP/PS ratio. The uniform trend in the D_{N,L,⟨121⟩} to D_{N,L,⟨110⟩} ratio suggests that the synthesis of the copolymer latex is one that permits a high degree of stoichiometric control, which can be used to adjust the ability of the particles to undergo plasticization by water.

pH-Response of the PS_xP2VP_y Latex Particles

The tuned content of pyridine groups in the latexes contributes to the amount of pH-triggered swelling in these lightly cross-linked materials. Protonation of these groups can result in significant electrostatic forces and an increased volume due to solvation of water and ions within the polymer particles. A representation of the molecular transformation that results from the treatment of the highly neutral PS_xP2VP_y (L) with a general Brønsted–Lowry acid (HX) is shown in Scheme 2. Measurements by DLS were used to monitor the D_H of selected PS_xP2VP_y (L) in water with 0.1 M NaCl over the course of a titration with 0.1 M HCl. These D_H versus pH titration curves are found in Figure 2a. The initial D_H values of the latexes are largely in agreement with data reported in Table 1. Most latexes exhibited stable D_H values until pH ∼4. Each PS_xP2VP_y (L) eventually underwent a significant size change upon treatment with acid and consequently transitioned to a microgel state with a larger diameter. The diameter of the microgel also appears to be independent of pH. Previous work has shown that analogous P2VP latexes undergo this swelling transition very rapidly, roughly requiring only 100–150 ms.⁴⁶ The polydispersity values for the resulting microgels all remained within a range of 0.028–0.876 (Table 1). The D_H associated with the microgel state is listed as D_H,M in Table 1, where P2VP-rich microgels had the largest D_H,M and those that were more PS-rich had the smallest D_H,M. A linear correlation was found between the D_H,M and the nominal P2VP weight content of the microgels (Figure 2b), which emphasizes the degree of synthetic control for generating a uniform distribution of P2VP⁺ units throughout the microgels. It is also evident that as the P2VP content in the latex is increased, the degree of change in D_H of the resulting microgel also increases. The evolution in the D_H,M to D_H,L ratio versus the nominal weight percent of P2VP in the latex is reported in Figure 2c. Again, a linear correlation was found for this plot, further emphasizing the degree of synthetic control on the swellability in the microgels made possible by including tailored amounts of PS in the macromolecular network. A volumetric swelling ratio can be calculated from the D_H,M and D_H,L values to quantify this characteristic. The volumetric swelling factor for the latex with no PS content (PS₀P2VP₁₀₀) was 140, while that for the most PS-rich in the series (PS₆₀P2VP₄₀) was 2.5. The titration curves in Figure 2a can also be used to determine an average pK_a value for the pyridinyl groups in each microgel. All pK_a values were found to be less than those for P2VP homopolymer (pK_a ≈ 4.92).^56,57 Armes et al. have previously observed a linear dependence for the pK_a of P2VP with a DVB cross-linker.⁴⁶ The decreasing pK_a with an increasing cross-linker content was suggested to arise from a combination of the “polyelectrolyte effect” and macromolecular swelling strain.⁵⁸ Unique to this work, we have observed that the pK_a of the P2VP can be tuned by adjusting the content of hydrophobic PS. Shown in Figure 2d is the plot for the dependency of the pK_a versus the nominal weight percent of P2VP in the latex. In this case, a linear trend was found for the correlation of pK_a with P2VP content. Enthalpic penalties associated with exposure of PS segments to water are reflected in the overall pK_a of the microgel. Utilizing hydrophobic PS content for tuning the pH-responsive properties of pyridinyl groups in microgels represents a novel method and one that may allow for better matching of the conditions and specificity needed for various applications such as pH-reporting,^59,60 catalysis,⁶¹⁻⁶³ and molecular detection via SERS.^44,64

Scheme 2. Generalized Chemical Equation for the Reversible Transition of Sterically Stabilized PS_xP2VP_y (L) (at Left) into Sterically Stabilized PS_xP2VP_y (M) (at Right) upon Addition of a Brønsted–Lowry Acid (HX).

Figure 2.

(a) Titration curves relating D_H (by DLS) to pH for PS_xP2VP_y samples. The large size change in D_H in each curve represents the transition from a latex (higher pH) to a microgel (lower pH). (b) Plot for D_H,M vs the nominal P2VP weight content in the PS_xP2VP_y series. The D_H,M value was determined by averaging the data for lower pH plateaus in (a). Plots for D_H,M to D_H,L ratio (c) and pK_a (d) vs the nominal P2VP weight content in the PS_xP2VP_y series.

Loading of PS_xP2VP_y Microgels with AuCl₄^– Ions

The electrostatic loading of the selected PS_xP2VP_y (M) with anionic species such as AuCl₄^– and borate-caped Ag NPs was investigated. Experiments were designed to probe the effect of loading anionic materials in an aqueous condition with a pH below the pK_a values of the microgel. Two routes to metal NP microgels (e.g.,, PS_xP2VP_y (M)/Au NPs and PS_xP2VP_y (M)/Ag NPs) are possible, and these are distinguished in Scheme 3. To achieve PS_xP2VP_y (M)/Au NPs, aurate-loaded microgels (e.g., PS_xP2VP_y (M)/AuCl₄^–) were prepared using acidic conditions. Solutions of PS₂₀P2VP₈₀ (M) and PS₄₀P2VP₆₀ (M) had their pH set to 2.5 using 0.1 M HCl, while that for PS₆₀P2VP₄₀ (M) required a pH of 1.2. Measurements by DLS were used to monitor the size of the PS₂₀P2VP₈₀ (M) as the solution was titrated with KAuCl₄. In this experiment, the pH of the KAuCl₄ was adjusted to match the pH of the PS_xP2VP_y (M) solution. The plot in Figure 3a reports the evolution in the D_H of PS₂₀P2VP₈₀ (M) as increasing amounts of KAuCl₄ are introduced (shown as equivalents of KAuCl₄ to 2VP units). The initial D_H of the microgel was ∼1980 nm and was found to decrease exponentially. The introduction of only 0.05 equiv of KAuCl₄ results in a decrease in D_H to ∼1420 nm, which corresponds to a volume that is only 37% of that of the initial state. This dramatic contraction confirms a strong electrostatic cross-linking throughout the macromolecular network. This volume contraction continues as more KAuCl₄ is delivered, with an ultimate plateau value for a D_H of 960 nm once ∼0.30 equivalents are bound (D_H,M in entry 6 in Table 1). The final pH of the PS₂₀P2VP₈₀ (M)/AuCl₄^– solution was similar to that of the initial condition. McPhee et al. have also found that an ultimate D_H is achieved in the titration of poly(N-isopropylacrylamide) microgels with a sub-equimolar amount of sodium dodecylsulfate.⁶⁵ Stable PS₄₀P2VP₆₀ (M)/AuCl₄^– and PS₆₀P2VP₄₀ (M)/AuCl₄^– were also produced from the introduction of 0.3 molar equivalents of KAuCl₄. The D_H values of these samples were also observed to decrease in a similar manner as the KAuCl₄ is incorporated into the microgels (entries 7 and 8 in Table 1). The polydispersity values of the PS_xP2VP_y (M)/AuCl₄^– samples remained relatively low. Conversion of the aurate-loaded microgels to their latex state [termed PS_xP2VP_y (L)/AuCl₄^–] was investigated by raising the pH with 0.1 M NaOH. Only modest adjustments in the D_H values were found (Table 1, entries 6–8), further confirming the extensive electrostatic cross-linking present in the samples.

Scheme 3. Loading of the PS_xP2VP_y (M) with (i) KAuCl₄ and the Subsequent Photoreduction (Top Pathway), or (ii) Loading with Borate-Capped Ag NPs (Bottom Pathway).

Figure 3.

(a) Titration curve monitored by DLS for PS₂₀P2VP₈₀ loaded with 24 mM KAuCl_4(aq). The pH of the PS₂₀P2VP₈₀ solution and the KAuCl₄ solution were both adjusted to 2.5 and the pH of the PS₂₀P2VP₈₀ (M) solution remained 2.5 throughout the titration. (b) TGA curves for unloaded and aurate- and Au NP-loaded PS₂₀P2VP₈₀ (M). (c) TGA curves for Au NP-loaded PS₂₀P2VP₈₀ (M), PS₄₀P2VP₆₀ (M), and PS₆₀P2VP₄₀ (M). Inset: M_775°C vs the nominal P2VP weight content for the aurate-loaded (squares) and Au NP-loaded (circles) samples. (d) High-resolution X-ray photoelectron spectroscopy (HRXPS) for PS₂₀P2VP₈₀ (M)/AuCl₄^–, PS₂₀P2VP₈₀ (M)/Au NPs, PS₄₀P2VP₆₀ (M)/Au NPs, and PS₆₀P2VP₄₀ (M)/Au NPs. The dash vertical line indicates average binding energy of the maxima in the Au 4f_7/2 signals for PS_xP2VP_y (M)/Au NPs. The HRXPS plot has been normalized and offset for clarity. Tracer plots for the absorbance at 1585 cm^–1 (e) vs temperature for PS₄₀P2VP₆₀ (M), PS₄₀P2VP₆₀ (M)/AuCl₄^–, and PS₄₀P2VP₆₀ (M)/Au NPs. In (e), the bottom trace is for a UV-treated PS₄₀P2VP₆₀ (M). (f) Tracer plots for the absorbance at 1585 cm^–1 versus temperature for PS₂₀P2VP₈₀ (M)/Au NPs, PS₄₀P2VP₆₀ (M)/Au NPs, and PS₆₀P2VP₄₀ (M)/Au NPs. The absorbance at 1585 cm^–1 is associated with pyridine-containing volatiles from P2VP. In (e,f), the absorbance scale has been normalized and plots are offset for clarity.

Suspensions of all PS_xP2VP_y (M)/AuCl₄^– were purified by dialysis and subsequently studied using thermogravimetric analysis (TGA)–Fourier transform infrared (FTIR). The TGA data for the unloaded microgels (solid traces in Figure 3b and the Supporting Information, Figure S1a–c) largely exhibit similar mass loss profiles with a single dominant mass loss event at 420 °C and other characteristics such as a temperature for 5% mass loss (T_D5%) in the range of 360–380 °C and a char mass at T = 775 °C (M_775°C) of ∼0% for the PS₂₀P2VP₈₀ (M). The mass loss at 420 °C arises from a depolymerization degradation mechanism for the PS and P2VP components in the random copolymer network.^66,67 The TGA of a AuCl₄^– compound has been previously published⁶⁸ and exhibits a single mass loss event at 245 °C with a T_D5% of ∼240 °C and a M_775°C of ∼35%. The corresponding TGA data for the AuCl₄^–-loaded microgels are different and are a function of the nominal P2VP weight content (Figures 3b and S1a–c). These PS_xP2VP_y (M)/AuCl₄^– materials exhibit a two-step mass-loss profile with main degradation events at ∼260–280 and 400–405 °C. It is expected that as the microgel is enriched in P2VP content that the capacity for loading KAuCl₄ should increase. As the nominal P2VP content in the microgel increases in the series of samples, and consequently as the amount of AuCl₄^– increases, the mass loss associated with the first degradation event increases. On the basis of the mass loss data, and the close correspondence in the temperature of this event to that of the native AuCl₄^– material, we assign the first degradation event as a mass loss associated with incorporated aurate ions. The second mass loss event would, therefore, correspond to the depolymerization mechanism mentioned above. The M_775°C of the AuCl₄^–-loaded microgels was also found to increase linearly with the nominal P2VP weight content (see inset plot in Figure 3c). The inorganic content of the AuCl₄^–-loaded microgels is largely responsible for mass at 775 °C.⁶⁸ The inorganic content of the AuCl₄^–-loaded microgels is listed in Table 1 (entries 6–8). As the P2VP content is tuned from 40 to 80%, the inorganic content increases from ∼20 to 43%. The high inorganic content and the linear relationship in M_775°C with P2VP weight content suggests that the electrostatic binding of aurate into the network proceeds efficiently and utilizes most P2VP⁺ sites in the microgel. Furthermore, the uniformity in the trend also speaks to a highly uniform distribution of PS and P2VP residues throughout the polymer network.

Microgel Au NP Composites Prepared from AuCl₄^–-Loaded PS_xP2VP_y Microgels

The AuCl₄^–-loaded microgels were investigated as precursors for the preparation of Au NP-loaded PS_xP2VP_y microgels [termed PS_xP2VP_y (M)/Au NPs]. Previous investigations on the chemical reduction of similar aurate-loaded P2VP-based microgels using dimethylamineborane as a chemical reducing agent led to microgels with only surface-loaded Au NPs.⁴⁵ The method of photoreduction of metal ion/polymer solutions takes advantage of ligands coordinated to the metal cation for rapid reduction and NP growth and is, therefore, a reduction method that is not limited by the mass transport of a chemical reducing agent.⁶⁹ Solutions of yellow AuCl₄^–-loaded microgels were exposed to 254 nm UV light and produced pink to purple colored solutions, which is indicative of the formation of Au NPs. A control sample of PS₂₀P2VP₈₀ (M) was treated to the same UV exposure and showed similar characteristics to that of the native microgel, confirming that the polymeric scaffold is stable to the UV treatment (see Figure S1d; and later FTIR discussion). The recovered PS_xP2VP_y (M)/Au NPs were purified by dialysis prior to characterization. The DLS studies revealed that the D_H,M values of PS_xP2VP_y (M)/Au NPs are similar or smaller than the starting PS_xP2VP_y (M)/AuCl₄^– materials. Both the PS₂₀P2VP₈₀ (M)/Au NPs and the PS₄₀P2VP₆₀ (M)/Au NPs contracted to 29 and 26% of the original volume, respectively. The PS₆₀P2VP₄₀ (M)/Au NPs had a similar volume to the precursor microgel. We hypothesize that the lower aurate loading and the lower pK_a of the PS₆₀P2VP₄₀ (M) is responsible for the different degree of volumetric change when compared to the other two Au NP-loaded microgels. For PS₂₀P2VP₈₀ (M)/Au NPs and PS₄₀P2VP₆₀ (M)/Au NPs, the decrease in D_H,M is likely the result of the reduction of Au ions to neutral Au(0). The presence of Au(0) reduces ionic solvation and electrostatic repulsion forces previously at play in the larger D_H,M values found for the aurate loaded precursors. Access to the equivalent latex state [termed PS_xP2VP_y (L)/Au NPs] was attempted with a pH-adjustment using 0.1 M NaOH. Similar to the parent aurate-loaded microgels, relatively small changes in D_H were found. A more detailed D_H versus pH titration curve for PS₂₀P2VP₈₀ (M)/Au NPs was performed and monitored by DLS, which confirmed this observation (Figure S2a). Riedinger et al. have studied the volume change associated with the latex-to-microgel transition for poly(2-vinylpyridine-co-divinylbenzene) polymer spheres loaded with Au NPs and reported similar stability in D_H.⁷⁰ It was proposed that the NPs induced steric hindrance in the spheres and led to stable diameter values in the more rigid structure. The SEM investigations on PS₄₀P2VP₆₀ (M)/Au NPs and PS₆₀P2VP₄₀ (M)/Au NPs demonstrated that the shape of the Au NP composites were spherical and indicated that the Au NPs were ∼7 nm in diameter (Figure S2b,c). The D_N,L values for the solvent-free PS₄₀P2VP₆₀ (M)/Au NPs and PS₆₀P2VP₄₀ (M)/Au NPs samples are also reported in Table 1 (entries 10–11). Similarly, these D_N,L values were smaller than any D_H value for the Au NP-loaded particles, confirming solvent-based swelling in the Au NP-loaded latex and microgel states.

The TGA mass loss profile for PS₂₀P2VP₈₀ (M)/Au NPs is shown in Figure 3b (dotted trace), and several differences are apparent when the data for the PS₂₀P2VP₈₀ (M)/AuCl₄^– are compared to that for the parent PS₂₀P2VP₈₀ (M) (dashed and solid traces, respectively). Generally, all Au NP-loaded microgels show a single mass loss event at ∼390–395 °C, which can be assigned to the depolymerization of the microgel polymer network. In NP-loaded samples, the amount of mass loss due to degradation of AuCl₄^– was negligible (previously discussed mass loss event at ∼T = 250 °C), suggesting that the aurate ions have successfully reduced into the more thermally stable Au(0) metallic state. The exposure of 254 nm UV light to AuCl₄^– causes multiple electron transfer steps from chloride ligands to the Au(III) center, resulting in a reduction to Au(0) and the liberation of chloride ligands as a byproduct.^71,72 In all cases, the char mass of the Au NP-loaded microgels, indicated herein by the M_775°C, was found be less than that of the metal anion-loaded precursor. The decrease in char mass can be attributed to the successful reduction of AuCl₄^– ions to Au NPs and the associated loss of Cl^– ions. The TGA mass loss profiles for PS₄₀P2VP₆₀ (M)/Au NPs and PS₆₀P2VP₄₀ (M)/Au NPs are shown alongside that of the PS₂₀P2VP₈₀ (M)/Au NPs in Figure 3c for comparison. As the nominal P2VP weight content increases, no significant differences in the T_D5% were found, yet the M_775°C increases linearly from 19 to 38% for the Au NP-loaded microgels (see the inset in Figure 3c). The evidence for the loss of Cl^– supports the assignment of the M_775°C as mass derived primarily from Au NPs and hence inorganic-loading entries 9–11 in Table 1 equate to the Au NP mass content of PS_xP2VP_y (M)/Au NPs samples. The increasing linear trend in M_775°C with increasing P2VP content in the microgels (see the inset plot in Figure 3c) confirms that the P2VP content specifies the ultimate Au NP loading in the PS_xP2VP_y (M)/Au NPs composites.

The size, location, and the Au-oxidation state of the Au NPs were determined using transmission electron microscopy (TEM) and XPS. The charge-corrected HRXPS plots for the 4f region for Au acquired for PS₂₀P2VP₈₀ (M)/AuCl₄^– and selected PS_xP2VP_y (M)/Au NPs are shown in Figure 3d. For the aurate-loaded microgel, the plot reveals two peaks for the 4f core electrons of Au with peak binding energy values for the 4f_5/2 and 4f_7/2 peaks at 89.8 and 86.1 eV, respectively, and with a peak separation of 3.7 eV. These values confirm that the Au in the ion-loaded microgels has an oxidation state of (III).⁷³ Corresponding plots for PS₂₀P2VP₈₀ (M)/Au NPs, PS₄₀P2VP₆₀ (M)/Au NPs, and PS₆₀P2VP₄₀ (M)/Au NPs reveal the peak binding energy values for the 4f_5/2 and 4f_7/2 peaks at 87.6 and 84.0 eV, respectively, with a peak separation of 3.6 eV, all of which confirm that the Au in these systems has an oxidation state of (0).^74,75 The tomography video (Movie S1) and TEM images (Figure S3a) identify Au NPs in the PS₀P2VP₁₀₀ (M)/Au NP sample. The TEM-estimated diameter for the Au NPs is 7 nm, where ∼150 Au NPs are present in a PS₀P2VP₁₀₀ (M)/Au NP nanosphere. Frames from the 3D tomography dataset have been extracted (Figure S3b) and used to determine that the Au NPs are located throughout the P2VP microgel, not just on their surfaces as previously seen by others.⁴⁵

To probe at the nature of the bonding associated with pyridinyl groups in PS_xP2VP_y (M)/AuCl₄^– and PS_xP2VP_y (M)/Au NPs, TGA–FTIR studies were conducted. The intensity maps from FTIR for temperature versus wavenumber of the TGA–FTIR datasets are shown in Figure S4a–n. In comparing the data from PS₂₀P2VP₈₀ (M), PS₂₀P2VP₈₀ (M)/AuCl₄^–, and PS₂₀P2VP₈₀ (M)/Au NPs, all samples show evidence for a methyl-containing volatile (ν_CH3 = 2973 cm^–1) that is liberated with the onset and peak temperatures of 465 and 480 °C, respectively. Similarly, each sample evolves an ether-containing volatile (detected with a ν_CO at 1136 cm^–1; likely from the PEGMA stabilizer) with the onset and peak temperatures of 450 and 475 °C, respectively. Volatiles associated with a ν_aromatic CH at 3048–3111 cm^–1 are liberated at similar onset and peak temperatures (∼445 and ∼480 °C, respectively) as well. The notable differences in the TGA–FTIR studies are found for the vibrational ring-stretching modes for phenyl and pyridinyl groups that are associated with evolved gases from the samples. The intensity tracer plots from the FTIR data relate the intensity of absorbance for the phenyl ring stretch of PS (ν_benz = 1635 cm^–1) and the pyridinyl ring stretch of P2VP (ν_pyr = 1585 cm^–1); they are reported in Figures S4o and 3e, respectively. For the inert conditions applied, the thermal degradation of PS and P2VP proceeds by a depolymerization mechanism that evolves S and 2VP monomers at temperatures of 436 °C.⁷⁶ The signals shown in Figures S4o and 3e can therefore be assigned to specific degradation products for PS and P2VP, respectively. From the data in Figure 3e, the flux of pyridine-containing volatiles from PS₄₀P2VP₆₀ (M) occurs with an onset and peak temperature of 410 and 478 °C, respectively. This same flux of pyridine-containing volatiles from PS₄₀P2VP₆₀ (M)/AuCl₄^– and PS₄₀P2VP₆₀ (M)/Au NPs occurs at lower temperatures (top two traces in Figure 3e). For PS₂₀P2VP₈₀ (M)/Au NPs, the onset and peak temperature (T_onset = 390 °C and T_peak = 455 °C, respectively) were also ∼20 °C lower than the metal-free analogue. The profiles of the flux of pyridine-containing volatiles from PS₂₀P2VP₈₀ (M)/Au NPs and PS₆₀P2VP₄₀ (M)/Au NPs were also the same as that for PS₄₀P2VP₆₀ (M)/Au NPs (Figure 3f). A control experiment where unloaded PS₄₀P2VP₆₀ (M) was treated to the photoreduction step (bottom trace in Figure 3e) indicates that the chemical changes in the macromolecular network are likely trivial and are not responsible for the decrease in temperature required to volatilize pyridinyl compounds. The decreased thermal stability of P2VP in the aurate- and the Au NP-loaded microgels suggests that there is relatively weak bonding to these pyridinyl groups, which permit a volatilization of the depolymerization products at a lower temperature. However, it is possible that other kinetic factors are at play as well. For example, it is possible that more efficient heat transfer into the microgel network arises from an increase in the thermal conductivity in the samples imparted by the presence of metallic species.⁷⁷ The temperature required to volatilize degradation products from PS, indicated by the signal associated with the phenyl ring stretch, was also reduced by ∼20 °C (Figure S4o). The lack of significant bonding interactions between pyridine rings, which would have otherwise increased the temperature required to volatilize pyridine-containing molecules, suggests that the surface of the Au NPs is only weakly bonded to the microgel network and is, therefore, available for the binding of analytes for SERS detection.

SERS from PS_xP2VP_y (M)/Au NPs

The high loading of Au NPs in PS_xP2VP_y (M)/Au NP composites and the relatively uncoordinated Au surface suggest that the samples may be a candidate material for SERS detection. Figure S5a,b shows the SERS spectra acquired on solutions of PS₂₀P2VP₈₀ (M)/Au NPs, PS₄₀P2VP₆₀ (M)/Au NPs, and PS₆₀P2VP₄₀ (M)/Au NPs using 633 nm excitation. The data suggest that the Raman active vibration modes in the microgel network cannot be excited and/or detected under the conditions tested. Similarly, when the soluble and representative CV analyte was introduced, the SERS spectra did not show any detectable signals for this molecule. The high-resolution TEM data (Figure S3) indicated that the diameter of the Au NPs in this system was ∼7 nm, and hence, their plasmon-excitation absorbance, a requirement for the SERS technique, may be unlikely with the excitation conditions used herein.⁷⁸ We have further confirmed this by showing that free Au NPs are SERS active for the detection of CV using our experimental conditions provided that the Au NP diameter is greater than or equal to ∼30 nm (Figure S5). While no SERS spectra were acquired using PS_xP2VP_y (M)/Au NPs, it is possible that the composites would be excellent SERS substrates, provided that a correctly matched plasmon excitation source was employed or that a secondary growth step was used to enlarge the Au NPs. The SERS-activity study that focuses on the application of PS_xP2VP_y (M)/Au NPs is a topic of future work and is beyond the scope of this current report.

Ag NP Microgel Composites from Loading Borate-Capped Ag NPs onto PS_xP2VP_y Microgels

The low-sensitivity SERS detection nanocomposites from aurate-loaded microgels led us to consider larger, Ag-based NPs that exhibit a red-shifted absorbance instead. The immobilization of Ag NPs onto PS_xP2VP_y (M) was designed to utilize the cationic sites that are generated at the protonated pyridinyl residues. Ligand-based pyridine–Ag interactions that stabilize bound Ag NPs at the microgel are also possible.⁷⁹ Borate-capped Ag NPs were selected because of their ease of synthesis, stability, and an inherent negative surface charge.⁸⁰Figure 4a shows the SEM image of the borate-capped Ag NPs that were used in this work. The Ag NPs are roughly spherical in shape and have a diameter of ∼55 nm, a value that agrees with the DLS data (see the Experimental Section). As described above, the pH of the solution of PS_xP2VP_y (M) was adjusted below that of the pK_a. Introduction of a dilute solution of borate-capped Ag NPs to a solution of PS_xP2VP_y (M) at pH = 1–2 resulted in the self-assembly of the PS_xP2VP_y (M)/Ag NP composite. The PS_xP2VP_y (M)/Ag NP composites were purified by dialysis prior to characterization. The SEM images in Figure 4b–d depict the PS₀P2VP₁₀₀ (M)/Ag NPs, PS₂₀P2VP₈₀ (M)/Ag NPs, and PS₄₀P2VP₆₀ (M)/Ag NPs, respectively. Inspection of these images confirmed that the immobilization procedure led to a single nanostructure composite in each case with neither free Ag NPs on the underlying SEM substrate nor any unloaded PS_xP2VP_y (M). The SEM-estimated number of Ag NPs per PS_xP2VP_y (M) was determined by averaging five different PS_xP2VP_y (M) in each corresponding SEM image (any Ag NPs beneath the colloids are, therefore, omitted). The average number of Ag NPs per PS₀P2VP₁₀₀ (M), PS₂₀P2VP₈₀ (M), and PS₄₀P2VP₆₀ (M) sphere was 29 ± 12, 15 ± 5 and 4 ± 3, respectively (see inset in Figure 5b). On the basis of the diameter values for the Ag NPs and those for the dried latexes, and ignoring any NP–NP overlap, the Ag NP footprints are ∼17, 13, and 9% on the surface of their polymer colloid substrates. This correlation confirms that adjusting the S/2VP ratio at the synthetic stage is a means for controlling the number of Ag NPs per microgel and, therefore, the surface density of Ag NPs for SERS.

Figure 4.

SEM images of (a) Ag NPs, (b) PS₀P2VP₁₀₀ (M)/Ag NPs, (c) PS₂₀P2VP₈₀ (M)/Ag NPs, and (d) PS₄₀P2VP₆₀ (M)/Ag NPs. The scale bars in (b–d) are 2 μm.

Figure 5.

(a) TGA curves of PS_xP2VP_y (L)/Ag NPs. (b) Mass percent remaining at 775 °C, indicating Ag NP loading. (c) Raman spectra of the PS_xP2VP_y (M)/Ag NPs. (d) Plot for the intensity of the signal at 1011 cm^–1 versus the nominal weight content of the P2VP in PS_xP2VP_y (M)/Ag NPs.

Further quantification of the Ag content for the immobilization of borate-capped Ag NPs on PS_xP2VP_y (L) was conducted using TGA. Figure 5a shows the overlaid mass loss profiles for PS₀P2VP₁₀₀ (L)/Ag NPs, PS₂₀P2VP₈₀ (L)/Ag NPs, and PS₄₀P2VP₆₀ (L)/Ag NPs. Each dataset shows that the char plateau is less well-defined relative to the parent and unloaded microgel analogues. Previous TGA analyses on the related borate-loaded polymers attributed the more gradual mass loss in the char region to a slow breakage of B–O bonds for the mass loss.⁸¹ Nonetheless, the M_775°C metrics can still be extracted from the plots for the Ag NP composites and used for further discussion. The M_775°C values for PS₀P2VP₁₀₀ (L)/Ag NPs, PS₂₀P2VP₈₀ (L)/Ag NPs, and PS₄₀P2VP₆₀ (L)/Ag NPs were 40, 25, and 15%, respectively, and based on the near-zero M_775°C values for the parent microgels can be assigned as inorganic mass due to Ag. These data are also found in entries 12–14 in Table 1. Figure 5b shows the plot for the M_775°C versus the nominal mass percent of P2VP. The increasing trend in Ag mass with P2VP content correlates with conclusions based on the SEM where increased Ag loading occurred as the P2VP content increased. On the basis of the TGA-mass estimate for Ag, the spherical shape and diameter information gleaned from SEM, and the bulk density of Ag,⁸² the number of Ag NPs per microgel sphere can be calculated using eq 1

1

where the mass of a single PS_xP2VP_y colloid (mass_PSxP2VPy) was determined from the D_N,L of the PS_xP2VP_y (L) and the weighted-average of the bulk density of PS and P2VP;⁸³ the mass of a single Ag NP (mass_Ag NPs) was determined using the SEM estimated diameter for the Ag NPs and the bulk density for Ag. As such, the number of Ag NPs per microgel sphere for PS₀P2VP₁₀₀ (M)/Ag NPs, PS₂₀P2VP₈₀ (M)/Ag NPs, and PS₄₀P2VP₆₀ (M)/Ag NPs was found to be 26, 12, and 5, respectively (see the inset in Figure 5b). These values are in agreement with those determined by SEM particle-counting. The controlled loading of Ag NPs onto sterically stabilized PS_xP2VP_y (M) offers a potential advantage for tuned SERS activity. Contrary to the Au NP-loaded samples, the PS_xP2VP_y (M)/Ag NPs analogues exhibit increased pH sensitivity as characterized by DLS (Table 1, entries 12–14). The D_H,M of the PS_xP2VP_y (M)/Ag NP samples are less than the values of their metal-free parent microgels (Table 1, entries 1–3), possibly indicating a contraction due to bonding to the Ag NPs. Interestingly, the D_H,M and D_H,L for the PS₂₀P2VP₈₀ (M)/Ag NPs subtly deviate from the expected trend of decreasing values with decreasing P2VP content. It is possible that the PS content in this system allows for a reduction in the number of contacts between bound Ag NPs and allows for more flexibility in the polymer network, thus favoring a subtly larger size than the PS₀P2VP₁₀₀ (L/M)/Ag NPs. This effect may be offset by energetic penalties for PS-contacts with water that may dominate the PS₄₀P2VP₆₀ (M)/Ag NP sample. The volumetric swelling factors for the series of samples, however, followed a similar and expected trend to the parent microgels, where the values for PS₀P2VP₁₀₀ (M)/Ag NPs, PS₂₀P2VP₈₀ (M)/Ag NPs, and PS₄₀P2VP₆₀ (M)/Ag NPs, were 7.9, 3.3, and 1.8, respectively. The combination of the tuned Ag NP content and the high pH-sensitivity in the PS_xP2VP_y (M)/Ag NPs series offer a unique experimental landscape to maximize SERS properties such as sensitivity and stability.

The immobilization of SERS-active NPs onto a polymer structure suggests that a Raman spectrum of that polymer substrate could be acquired. Previous studies on the Raman spectrum of copolymers of PS and P2VP have been reported.^79,84 Solutions of dispersed PS₀P2VP₁₀₀ (M)/Ag NPs, PS₂₀P2VP₈₀ (M)/Ag NPs, and PS₄₀P2VP₆₀ (M)/Ag NPs were studied using a Raman spectrometer, and the baseline-corrected 633 nm excitation SERS spectra are presented in Figure 5c. The most prevalent peaks in PS₀P2VP₁₀₀ (M)/Ag NP spectra are found at (i) 1224, (ii) 1049, (iii) 1011, (iv) 751, and (v) 629 cm^–1. These correspond to modes relating to (i) the stretching of the C–C bond between the pyridine ring and the P2VP backbone, (ii) a mode for the in-plane C–H bending P2VP, (iii) a symmetric ring breathing mode in P2VP, (iv) the B–O–B bending mode in the borate anions, and (v) in-plane bending of the pyridine ring, respectively.⁸⁴⁻⁸⁶ The signal for the ring-breathing mode at 1011 cm^–1 indicates a high content of coordinated pyridine at the silver surface due to the adsorption through its nitrogen atoms.⁸⁴ The confirmation of the Raman spectra from the PS₀P2VP₁₀₀ (M)/Ag NPs further supports the structure proposed that is shown in Scheme 3. The SERS spectra for PS₂₀P2VP₈₀ (M)/Ag NPs and PS₄₀P2VP₆₀ (M)/Ag NPs are very similar in composition to that of PS₀P2VP₁₀₀ (M)/Ag NPs. The signal at 1049 cm^–1 represents a composite C–H bending signal for PS and P2VP. Notably, the signal for the ring-breathing mode in P2VP shifts subtly to 1002 cm^–1 once the nominal P2VP is decreased to 60%. This suggests that there is less N-coordinated-pyridine and more free pyridine sites in the P2VP in this system. Figure 5d shows the plot for the intensity for the signal at 1011 cm^–1 versus the nominal mass percent of P2VP in the microgel. The exponential decrease in the signal with decreasing P2VP content also reflects this diminished N-coordination. These data suggest that by tailoring the P2VP content in this system, a balanced amount of free P2VP results, which may favor improved acid-based swelling performance in the Ag NP-containing composites.

To assess the potential for PS_xP2VP_y (M)/Ag NPs to act as substrates for SERS-based detection, CV was used as a model analyte.⁸⁷Figure 6a (entry i) shows the 633 nm excitation Raman spectra for CV (1.1 × 10^–5 M in water) acquired without a SERS method. Because CV has electronic transitions at 633 nm, the observed CV spectra contain an additional resonance Raman enhancement contribution. The spectrum is consistent with previous literature studies⁸⁸ and consists of five main peaks at (a) 795, (b) 905, (c) 1175, (d) 1385, and (e) 1620 cm^–1, which correspond to (i) out-of-plane phenyl C–H bending, (ii) ring-breathing, (iii) asymmetric in-plane phenyl C–H bending, (iv) asymmetric phenyl C–N stretching, and (v) asymmetric phenyl C–C stretching modes for CV, respectively.⁸⁹ Many of these signature peaks for the analyte do not overlap with those for the PS_xP2VP_y (M) and, therefore, the background interference is minimal (top most spectra in Figures 5c and S6 in for comparisons). Previous SERS studies with CV under similar conditions concluded that the molecule is normally bonded to the Ag surface by Coulombic and van der Waals interactions.^90,91 The Ag NPs used in the synthesis of the PS_xP2VP_y (M)/Ag NPs are good substrates for SERS-based detection and as such, the 633 nm excitation SERS spectra of 1 × 10^–9 M Ag NPs with 1.1 × 10^–8 M CV in water are also included in Figure 6a (entry ii). Entries (iii–v) in Figure 6a represent the 633 nm excitation SERS spectra for CV acquired for 1.1 × 10^–8 M CV exposed to 1 × 10^–10 M PS₀P2VP₁₀₀ (M)/Ag NPs, PS₂₀P2VP₈₀ (M)/Ag NPs, and PS₄₀P2VP₆₀ (M)/Ag NPs, respectively. After comparing these spectra to that acquired using only Ag NPs (entry ii), it is apparent that the Ag NP-loaded microgels function with a similar bonding mode for CV but are more sensitive. Comparison of the intensity from the spectra for the Ag NP-loaded microgels shows that the SERS intensity increases with P2VP content. This increased intensity is presumably due to the increased Ag NP loading that was correlated from TGA and SEM investigations. The signal-to-noise (S/N) ratio for the peak found at 1620 cm^–1 was used to further characterize the correlation in intensity with P2VP content, where the S/N value for this peak is plotted versus the nominal P2VP mass percent in the microgel (Figure 6b). A linear dependence of the S/N at 1620 cm^–1 with P2VP content was found. Under the conditions investigated herein, the linear trend suggests that the PS_xP2VP_y (M)/Ag NPs are uniform in their sensitivity and enhancement mechanism. AEFs were calculated using eq 2(87) where

2

I_SERS and I_RS are the intensity of the 1620 cm^–1 peak for the background-corrected SERS spectra and the non-SERS spectra of CV only, respectively. Parameters c_SERS and c_RS are the associated concentrations of CV for each test. For free Ag NPs, the AEF value was calculated to be 5.21 × 10⁹. The AEF results are outlined in Figure 6c for the three PS_xP2VP_y (L)/Ag NP nanocomposite systems (Table 2 also lists the calculated AEFs for the PS_xP2VP_y (L)/Ag NPs). The highest AEF value was found for the PS₀P2VP₁₀₀ (L)/Ag NPs whose value of 1.11 × 10¹⁰ is similar to other systems,^87,92,93 but 2.1 fold higher than that found for the control sample containing only Ag NPs (bottom entry in Table 2). As the P2VP content of the system decreased, the AEF values decreased marginally, but remained higher than that of the control sample containing only Ag NPs (Figure 6c). The AEF for the PS₂₀P2VP₈₀ (L)/Ag NPs was 1.08 × 10¹⁰, while that for PS₄₀P2VP₆₀ (L)/Ag NPs was 7.08 × 10⁹. Previous work for Au NP-decorated P2VP microgels reported an enhancement factor of 1.3 × 10⁶ for CV with an excitation source of 532 nm.⁵⁰ The high AEF values reported herein are a consequence of the application of the more SERS-sensitive Ag NPs, additional resonance Raman enhancement at 633 nm excitation for CV, and the controlled loading of these NPs onto well-defined copolymer microgels.

Figure 6.

(a) Initial Raman spectra of the PS_xP2VP_y (M)/Ag NPs and control comparisons with CV. (b) S/N ratio at 1620 cm^–1 for CV for spectra acquired using PS_xP2VP_y (M)/Ag NPs versus the nominal P2VP weight content. (c) AEFs (for pH cycle 0) of PS_xP2VP_y (L)/Ag NPs versus the nominal P2VP weight content. (d) Tracking of the S/N of the 1620 cm^–1 signal for CV as the PS_xP2VP_y/Ag NPs are cycled between the latex (integer pH cycles) and microgel (half integer pH cycles) states.

Table 2. Summary of Calculated Analytical Enhancement Factors for the 1620 cm^–1 Raman-Active Mode of CV.

PSxP2VPy (M)/Ag NPs + CV ([CV] = 1.1 × 10–8 M)	average AEF
PS0P2VP100 (M)/Ag NPs + CV	1.11 × 1010
PS20P2VP80 (M)/Ag NPs + CV	1.08 × 1010
PS40P2VP60 (M)/Ag NPs + CV	7.08 × 109
Ag NPs + CV	5.15 × 109

The demonstration of a successful transition of the PS_xP2VP_y (M) hybrids to their latex-like state suggests that the sensitivity of the SERS-substrate may be pH-tuned. Previous work on Au NPs immobilized on grafted P2VP brushes correlated the pH-dependent SERS activity to the interparticle spacing.⁹⁴ The SERS-based spectra for CV were therefore iteratively acquired using PS₀P2VP₁₀₀ (M)/Ag NPs, PS₂₀P2VP₈₀ (M)/Ag NPs, and PS₄₀P2VP₆₀ (M)/Ag NPs upon shifting the pH of the media from values of 7 to 2. We, therefore, define a pH cycle value to track the performance of the SERS substrate. Integer values for the pH cycle value represent the number of times that a PS_xP2VP_y (M)/Ag NP sample was cycled back to pH = 7 for the acquisition of the SERS spectra of CV. Half-integer values of the pH cycle value represent intermediate experiments where the SERS spectra of CV were acquired at the intermediate condition of pH = 2. In all cases, the average intensity of the spectra increased with an increase in pH. The S/N value for the asymmetric phenyl C–C stretching modes at 1620 cm^–1 was used as a representative metric for this intensity change and is plotted against the pH cycle number, n, (Figure 6d). During the period of n = 0–4, the PS₀P2VP₁₀₀ (M)/Ag NP system showed an overall evolution to a condition where the sets of S/N values become more self-consistent for the high and low pH conditions. After this 4 cycle break-in period, the S/N values alternate between ∼35 and ∼20 for high and low pH, respectively. Interestingly, the PS₂₀P2VP₈₀ (M)/Ag NP and PS₄₀P2VP₆₀ (M)/Ag NP systems function with a reduced break-in period (Figure 6d). The number of pH cycles required to stabilize the S/N at 1620 cm^–1 for the SERS spectra of CV for PS₂₀P2VP₈₀ (M)/Ag NP and PS₄₀P2VP₆₀ (M)/Ag NP systems was 1.5 and 2.5, respectively. After break-in, the S/N values for the SERS spectra of CV for PS₂₀P2VP₈₀ (M)/Ag NPs alternate between ∼77 and ∼30 for high and low pH, respectively. The corresponding values for the PS₄₀P2VP₆₀ (M)/Ag NP system were 44 and 18, respectively. The PS₂₀P2VP₈₀ (M)/Ag NP system, therefore, appears to have the greatest sensitivity for pH-tuned SERS activity and operates with the least amount of break-in.

The stability of the intensity of the SERS spectra of CV was evaluated using PS₀P2VP₁₀₀ (L)/Ag NPs and PS₂₀P2VP₈₀ (L)/Ag NPs. Spectra were acquired over a multiday period with pH cycling on each day, and the value of the S/N for the 1620 cm^–1 vibration mode is plotted versus time and reported in Figure S7. When PS₀P2VP₁₀₀ (L)/Ag NPs were employed, the S/N for the 1620 cm^–1 for CV dropped from ∼50 to an ultimate value of 4 over three days. The corresponding CV SERS signal derived from the PS₂₀P2VP₈₀ (L)/Ag NPs statistically increased during the first 3 days (from 82 to 237) followed by a decline to ultimate values within the range of 13 to 30 after 1 week. The observation that the CV SERS signal derived from PS₂₀P2VP₈₀ (L)/Ag NPs is greater than that derived from PS₀P2VP₁₀₀ (L)/Ag NPs further emphasizes the importance of the PS component in the microgel/latex. We hypothesize that the decline in the CV SERS signal in the PS₂₀P2VP₈₀ (L)/Ag NP system is due to oxidation of the Ag NPs with ongoing exposure to the ambient atmosphere. Confirming this and further stabilizing the SERS active NPs by way of the addition of secondary metals and/or refining the size and shape control of the NPs represents the subject of current work in this area.

We hypothesize that the inclusion of some PS into the structure of the PS_xP2VP_y (M)/Ag NPs improves the pH-sensitive SERS performance by balancing swellability with interfacial bonding strain. By reducing the number of pyridine sites in the microgel, the swellability is reduced (Figure 2b) and the bonding strain is reduced at the Ag NP–microgel interface during the pH-triggered iterative expansion and contraction is reduced. If this interfacial bonding strain is too high, the bonds to the Ag NPs break and the SERS activity and pH-sensitivity will decrease because of the loss of Ag NPs and an overall loss in the number of SERS-active interparticle gaps. Figure S8 shows the STEM images of the PS₀P2VP₁₀₀ (M)/Ag NPs and PS₂₀P2VP₈₀ (M)/Ag NPs prior to any pH cycling and after 3 pH cycles, respectively. It is clear that the number of Ag NPs on the PS₀P2VP₁₀₀ (M)/Ag NP sample decreases during the first 3 pH cycles, a finding that explains the decrease in the intensity of the Raman spectra (i.e., the decrease in the S/N at 1620 cm^–1 in Figure 6d). The population of Ag NPs on the PS₂₀P2VP₈₀ (M)/Ag NPs after 3 pH cycles is similar to that of uncycled analogue. This supports the theory that a composite with balanced swellability and a high number of pyridine sites for anchoring Ag NPs will exhibit improved pH-sensitive SERS performance. To further confirm the relationship between the change in the SERS signal and a related variation of the distance of the gap between Ag NPs, a UV–vis study was also conducted. The UV–vis spectra for PS₀P2VP₁₀₀ (M/L)/Ag NPs and PS₂₀P2VP₈₀ (M/L)/Ag NPs in Figures S9 and S10 were acquired over the course of the first three pH cycles. In the initial microgel condition (spectra I), both materials only show a broad plasmon absorbance band at ∼405 nm that is derived from roughly spherical and nonagglomerated Ag NPs. Upon treatment with acid, the latex condition (spectra II) evolved and included an additional plasmon-coupled absorbance feature at ∼610 nm because of a decreased distance between the Ag NPs for both the PS₀P2VP₁₀₀ (L)/Ag NPs and PS₂₀P2VP₈₀ (L)/Ag NPs. In both systems, the plasmon-coupled absorbance value reversibly alternates in size as the two PS_xP2VP_y (M/L)/Ag NP samples are taken through pH cycling until the final test with n = 3 (spectra VI). This strongly suggests that the alternating SERS signal intensity reported in Figure 6d is related to the number and the size of the nanoscale gaps between the Ag NP. Additionally, it is also apparent that the ultimate plasmon-coupled absorbance for PS₂₀P2VP₈₀ (L)/Ag NPs is stronger than that found in PS₀P2VP₁₀₀ (L)/Ag NPs (spectra VI in Figures S9 and S10). This finding also confirms that the PS₂₀P2VP₈₀ (L)/Ag NPs are the more SERS-sensitive material after break-in.

Conclusions

A controlled methodology for the synthesis of microgels with programmed properties such as pH sensitivity, swellability, and capacity for metal-loading is of great importance to many fields. In this work, sterically stabilized and covalently cross-linked latexes based on random copolymers of S and 2VP are synthesized. The latexes were found to be monodisperse by DLS and by SEM and exhibited a pH-triggered latex-to-microgel transition in the range of pH = 1.2–3.8, a characteristic that is set by the S/2VP ratio that is defined at the synthesis stage. The volumetric swelling factor was tuned from 2.5 to 140 with this ratio. The charged PS_xP2VP_y (M) was loaded with metal-containing anions such as AuCl₄^–- and borate-capped Ag NPs. Loading of PS_xP2VP_y microgels by DLS with KAuCl₄ showed a diameter reduction with increased binding where a loading limit of ∼0.3 molar equivalents of AuCl₄^– to pyridinyl groups was determined. The aurate-loaded PS_xP2VP_y (M) was converted to Au NP-loaded PS_xP2VP_y (M) by photoreduction. Tomography by TEM showed that Au NPs were located throughout the microgel structure. The loading of PS_xP2VP_y (M) with borate-capped Ag NPs produced PS_xP2VP_y (M) with surface-located Ag NPs. The Ag NP content of this nanocomposite was determined by the molecular ratio of S/2VP corresponding to a range of ∼5–30 NPs per microgel particle when the P2VP content ranged from 60 to 100%. The Ag NP-loaded PS_xP2VP_y (M) composites exhibited SERS activity for the macromolecular structure in the microgel, as well as for the representative solution-phase analyte, CV. A pH-responsive SERS activity for the Ag NP-loaded PS_xP2VP_y (M) was found where the AEF for dissolved CV was tunable from ∼5.21 × 10⁹ to 1.11 × 10¹⁰. The Ag NP-loaded PS_xP2VP_y (M) with ∼80 wt % P2VP exhibited the most stable pH response and a high AEF of 1.08 × 10¹⁰.

Experimental Section

Materials

Styrene (S; 99% purity), 2VP (97% purity), and 2,2′-azobis(2-methylpropionamidine) dihydrochloride (AIBA; 98% purity) were acquired from Acros Organics. PEGMA (average M_n = 2000, 50 wt % in H₂O) and DVB (55% purity) were purchased from Sigma-Aldrich. N-Methyl-N,N,N-trioctyloctan-1-ammonium chloride (Aliquat 336) was used as received from Alfa Aesar. Potassium tetrachloroaurate (KAuCl₄), hydrogen tetrachloroaurate (HAuCl₄), sodium borohydride (NaBH₄), silver nitrate (AgNO₃), and dry alumina oxide were obtained through Fischer Scientific, Inc. Hydrochloric acid (HCl) was used as received from Macron Fine Chemicals. Ultrapure water was produced using a Barnstead Nanopure water purifier and had an overall resistivity of 18 MΩ·cm. Regenerated cellulose dialysis tubing with a MWCO of 8000 or 12 000–14 000 were used as received from Fisher Scientific, Inc. The purification of S, 2VP, and DVB was done under Schlenk conditions by passing each monomer through dry alumina oxide powder to remove the polymerization inhibitors. The purified monomers were used within 24 h, and if short-term storage was required, they were sealed in round bottom flasks with an inert atmosphere and stored in the dark at −18 °C.

Synthesis of PS_xP2VP_y Latexes

The preparation of latex particles based on random copolymers of PS and P2VP (PS_xP2VP_y) follows a modified procedure outlined by Dupin et al.⁴⁶ The emulsion polymerization was carried out in a 100 mL single-necked round-bottom reaction flask. The PEGMA stabilizer (0.420 mL) and Aliquat 336 surfactant (0.15 g) were both dissolved in ultrapure water (40.00 g) in the reaction flask and sealed with a rubber septum. The reaction flask contents were deaerated with 30 min of nitrogen bubbling or five freeze, pump, thaw cycles. The flask, containing an inert atmosphere, was placed in a hot oil bath, and the contents stirred at 500 rpm until the temperature stabilized at 60 °C. The monomer mixture of 2VP (2.00–5.00 g), S (0.00–3.00 g) and DVB (0.03 g) was added to a scintillation vial under inert atmosphere for a total targeted amount of 5.03 g. The contents of the monomer vial were transferred to the reaction flask using a syringe equipped with a filter (PTFE, 0.45 μm, Millex). An emulsified condition was established by continual stirring (500 rpm, 60 °C) for approximately 30 min. Deaerated, ultrapure water (5.00 g) was used to pre-dissolve the AIBA initiator (0.050 g) in a separate scintillation vial with an inert atmosphere. Once dissolved, the AIBA solution was rapidly delivered to the reaction flask via a syringe to initiate the polymerization reaction. Stirring continued for 24 h. Purification of the resulting latex was achieved by centrifuging the crude latex for 15 min at 19 000 rpm in high-density polycarbonate tubes (Thermo Scientific). After decanting the supernatant liquid, an addition of ultrapure water was made to bring the volume of the latex solution back to its original level. Redispersion was carried out by sonication and/or stirring until the solution appeared uniform. This centrifugation/decanting process was repeated a minimum of three times. The final solution contained 8–9 wt % solids (with yields of 72–81%).

Preparation of Au NP-Loaded PS_xP2VP_y Microgels

The procedure for the preparation of Au ion-loaded colloids was adapted from Nakamura et al.⁴⁵ The pH of the PS_xP2VP_y (L) was adjusted to transform the latex into its microgel state, (i.e., to pH = 2.5 for PS₀P2VP₁₀₀, PS₂₀P2VP₈₀, and PS₄₀P2VP₆₀, and pH = 1.2 for PS₆₀P2VP₄₀) using 0.1 M HCl_(aq). The PS_xP2VP_y (M) was further diluted with HCl_(aq) (1:20) using the appropriate HCl solution. The KAuCl₄ was dissolved in 0.0032 M HCl_(aq) (i.e., pH = 2.5 for PS₂₀P2VP₈₀ and PS₄₀P2VP₆₀) and 0.063 M HCl_(aq) (i.e., pH = 1.2 for PS₆₀P2VP₄₀) to make a 24 mM KAuCl₄ solution. This pH-matched KAuCl₄ solution was added slowly to the diluted PS_xP2VP_y (M) until the desired level of ion-loading was achieved. In some cases, the loading was monitored by DLS. The final solution was stirred for 30 min and purified by dialysis with a MWCO of 8000 for 48 h. This afforded the ion-doped colloidal microgels. The ion-doped microgels were subjected to photolysis with stirring (500 rpm) for ∼18 h using a Rayonet model RPR-600 UV photochemical reactor (Southern New England Ultraviolet, Inc.) equipped with eight light sources, each with a 253.7 nm maximum emission and a power of 8 W. The resulting Au NP-loaded PS_xP2VP_y (M) were purified by three sequential 8 h dialysis steps using a MWCO of 12 000–14 000 membrane and fresh ultrapure water.

Synthesis of Borate-Capped Ag NPs

The borate-capped Ag NPs were synthesized according to modified procedures.^80,95 The reducible Ag ion solution (5 mM) was prepared by dissolving 21.2 mg of AgNO₃ in 25 mL of ultrapure water. This solution was added to an ice bath-chilled 75 mL aqueous solution of NaBH₄ (2 mM) and stirred for a minimum of 1 h. The reaction mixture gradually turned yellow-green over the first hour. This solution was aged in the dark for a minimum of 2 weeks prior to use. An average diameter of 59.8 nm (polydispersity = 0.361) for a pH = 8 solution of borate-capped Ag NPs was determined by DLS.

Complexation of PS_xP2VP_y Microgels with Ag NPs

A 10 μL volume of the purified PS_xP2VP_y (L) was diluted to a final volume of 10 mL with ultrapure water. The latexes were converted to the microgel state as indicated above. The borate-capped Ag NP solution was evaporated to half the volume (3 mL) and then was added slowly to the stirred microgel solution. The pH of the final solution was adjusted back to the starting value with the addition of HCl_(aq). The Ag NP-loaded PS_xP2VP_y microgels were purified by three consecutive 8 h dialysis steps using a MWCO of 12 000–14 000 membrane and fresh ultrapure water.

Particle Size Characterization by DLS

DLS measurements were conducted using a DelsaNano HC instrument (Beckman Coulter, Inc., Fullerton, CA) equipped with a 658 nm laser diode. The photomultiplier and aperture attenuator were set to collect scattered light at an angle of 15° with a photon counting rate of ∼10 kcps. Size measurements were conducted on ∼2–3 mL of the sample in disposable PS cuvettes (4 mL capacity). The properties of pH-adjusted solutions were assumed to be the same as water (n = 1.3329, η = 0.890 cP at 25 ± 1 °C). Samples (0.2 wt % solids) were filtered using 5 μm cellulose syringe filters (Millex) and allowed to equilibrate at 25 °C for ∼5 min before each measurement, followed by a 35–60 s acquisition time (1 s per acquisition). The calculation of the particle size and size distribution was performed using CONTIN particle size analysis routines. The intensity distribution plots were averaged, and the data are reported in Table 1 as the average hydrodynamic diameter (D_H) of the particles. A DLS-monitored titration with 0.1 M HCl was conducted on diluted PS_xP2VP_y colloids (original solution/ultrapure water = 1:20 v/v), with DLS size measurements occurring every 0.2 pH units after 1 min of equilibration.

Thermal Analysis

A TA Instruments Q500 TGA instrument equipped with an inline Thermo Scientific iS10 FTIR gas cell spectrometer was used for both the TGA and TGA–FTIR experiments.⁹⁶ To monitor the mass changes during thermal degradation and to determine the NP-loading, TGA experiments were conducted with a 15 min isotherm at 100 °C followed by a 20 °C/min ramp from 100 to 800 °C. An isotherm before and after this temperature program was held for 5 min to monitor the background signal of FTIR spectra. The effluent from the sample was carried to the FTIR spectrometer via an isothermal transfer line (225 °C) with dry nitrogen gas flowing at 90 mL/min. The FTIR spectra (4 scans at 8 cm^–1 resolution per spectra) were acquired at a 2.7 s interval for the duration of the TGA and were corrected using a background dataset created from a time averaged signal (5 min) from the blank nitrogen carrier gas.

General Instrumentation

XPS data were acquired using a Kratos M-Probe instrument equipped with an Al Kα monochromatic X-ray source and a hemispherical analyzer (with an outer diameter of 145 mm and an inner diameter of 88 mm). The analysis spot size was 15.71 μm² in area and the take-off angle was 55° with respect to the vertical direction of samples. The spectra were fit using 30% Gaussian–Lorentzian components and a Shirley background subtraction. All binding energies were corrected to the C 1s peak for the background hydrocarbon component (C–C/C–H_x) at 284.9 eV.⁹⁷ Electron tomography was conducted using a Talos F200X transmission electron microscope. The tilted view inspections of the samples were conducted with Inspect 3D software. SEM images were acquired using either Vega TS 5136MM operated with an accelerating voltage of 15 kV or JEOL JSM-7200F/LV operated with an accelerating voltage of 30 kV. With an exception of the Ag NPs, samples for SEM analysis were coated with a thin layer of Au and palladium (Pd) to dissipate charge (nominal thickness ≈ 1 nm). Scanning TEM (STEM) images were acquired with the JEOL JSM-7200F/LV instrument. Samples investigated by STEM were applied to a lacey carbon copper grid (Ted Pella) and not coated with charge dissipation materials prior to microscopy. The UV–vis spectroscopy was performed with a JASCO V670 spectrophotometer. The wavelength resolution was set to 1.0 nm and the scan speed was 400 nm/min. Absorbance was monitored for Ag-loaded PS_xP2VP_y solutions following dialysis and a 1:50 v/v dilution with pH = 2.5 aqueous HCl. The UV–vis spectra were acquired at each pH condition for a total of three pH cycles. The pH was cycled with the addition of either 82 μL of 3.0 M NaOH or 41 μL of 1.71 M HCl to 2.5 mL of the diluted Ag-loaded PS_xP2VP_y. The UV–vis spectra were normalized to their peak maxima.

Surface-Enhanced Raman Scattering

The SERS data were acquired using a DeltaNu Advantage 200A spectrometer equipped with a 633 nm helium–neon laser (5 mW). The spectral acquisition time was 60 s. The system was calibrated using bulk PS and neat cyclohexane as standards for the Raman frequency shifts (peaks for PS at 1001 and 621 cm^–1; peaks for cyclohexane at 1444 and 1266 cm^–1). All of the spectra were interpreted after background correction to exclude interference due to fluorescence.⁹⁸ Error bars correspond to the 95% confidence interval for 4–5 replicates.

Supporting Information Available

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsomega.8b01561.

• TGA–FTIR data for selected PS_xP2VP_y (M) samples and their AuCl₄^– and Au NP-loaded equivalents; TGA–FTIR data for a control PS_xP2VP_y (M) sample treated to a UV photoreduction condition; TGA–FTIR data for selected PS_xP2VP_y (M)/Ag NPs; DLS-monitored pH titration for PS₂₀P2VP₈₀ (M)/Au NPs; SEM images for PS₄₀P2VP₆₀ (M)/Au NPs and PS₆₀P2VP₄₀ (M)/Au NPs; TEM-tomography video (AVI), micrographs, and tilt frames for PS₀P2VP₁₀₀ (M)/Au NPs; Raman spectra for selected PS_xP2VP_y (M)/Au NPs and equivalents exposed to solution state CV; S/N of the 1620 cm^–1 signal for CV versus the diameter of free Au NPs; background comparison from an overlay plot for the Raman spectra of the PS₂₀P2VP₈₀ (M)/Ag NPs without and with CV in the solution; study for the S/N at 1620 cm^–1 for CV versus time using selected PS_xP2VP_y (M)/Ag NPs; STEM images of PS₀P2VP₁₀₀/Ag NPs (L) and PS₂₀P2VP₈₀/Ag NPs (L) after 0 pH cycles and after 3 pH cycles; and UV–vis spectra for selected PS_xP2VP_y (L)/Ag NPs during sequential pH cycling (PDF)

Author Present Address

^∥ Department of Chemistry, Simon Fraser University, 8888 University Drive, Burnaby, Canada, V5A 1S6.

Author Present Address

^⊥ Department of Chemistry, Indiana University, 800 E. Kirkwood Avenue, Bloomington, IN 47405-7102.

The authors declare no competing financial interest.