Salt Plays a Pivotal Role in the Temperature-Responsive Aggregationand Layer-by-Layer Assembly of Polymer-Decorated GoldNanoparticlesZhiyue Zhang,†,∥ Samarendra Maji,‡,∥ Andre ́ Bruno da Fonseca Antunes,† Riet De Rycke,§ Qilu Zhang,‡

Richard Hoogenboom,*,‡,⊥ and Bruno G. De Geest*,†,⊥

†Department of Pharmaceutics and ‡Supramolecular Chemistry Group, Department of Organic Chemistry, Ghent University, 9000Ghent, Belgium§Department of Molecular Biomedical Research, VIB, Zwijnaarde, 9000 Ghent, Belgium

*S Supporting Information

ABSTRACT: We report that the aqueous self-assembly behavior of citratebased gold nanoparticles decorated with the temperature responsive RAFT-based polymer poly(N-isopropylacrylamide) critically depends on thepresence of salt in the medium. Both for temperature induced reversibleagglomeration and for hydrogen bonding based layer-by-layer assembly withtannic acid, the presence of salt dramatically promotes the assemblybehavior. We attribute this to a combination of ionic screening of theremaining citrate groups on the nanoparticle surface and a salting out effectwhich increases the contribution of hydrophobic interactions in the self-assembly process. These findings provide new insights into an attractive classof polymer/gold hybrid nanomaterials that can find application inbiotechnology, catalysis, and biomedicine.

KEYWORDS: gold nanoparticles, temperature responsive polymers, N-isopropylacrylamide, layer-by-layer, hydrogen bonding

1. INTRODUCTION

Nanomaterials hold great promise for a wide variety ofapplications including in the biomedical field, e.g., as imagingtools, as, phototherapy agents, or as drug carriers.1 In thisregard, their surface properties play a critical role indetermining the outcome of their interactions with thephysiological environment. Among many different classes ofnanomaterials, gold nanoparticles (goldNP)2 functionalizedwith a responsive polymeric coating that is sensitive to externalstimuli, including biological cues,3 light irradiation,4 andchanges in pH or temperature, have gained considerableinterest for applications in a diverse range of areas embracingcatalysis, imaging, diagnostics, and therapeutic purposes.Envisioning biomedical applications, gold nanoparticles havethe advantage of being chemically inert, having a track record inclinical use for the treatment of rheumatoid arthritis and theversatile option to modify the goldNP surface “a ̀ la carte” byvirtue of the quasi covalent interaction between gold and(functional) thiols or more generic, sulfur-based compounds(Scheme 1).2b Importantly, goldNP exhibit interesting opticalproperties, owing to surface plasmon resonance of thecollective oscillation of electrons, that can be fine-tuned byaltering the size and the shape of the goldNP.5

The synthesis routes toward polymer coated gold nano-particles include either the use of presynthesized polymers orsurface-initiated polymerization from gold nanoparticles. The

use of preformed polymers is more versatile as it allows athorough characterization of the polymers combined with ahigher degree of freedom regarding reaction conditions,introduction of functionalities, or using mixtures of polymers.Using presynthesized polymers, one can either reduce gold saltsin the presence of sulfur end-capped polymers or graft suchpolymers onto preformed goldNP via ligand exchange. Thelatter option offers a higher degree of versatility regarding thesize and shape of the gold nanoparticles. Well-defined polymerssynthesized via reversible addition−fragmentation chain trans-

Received: July 19, 2013Revised: October 14, 2013Published: October 15, 2013

Scheme 1. (A) Ligand Exchange of Citrate Stabilized goldNPwith Thiol End-Group Containing Polymer and (B) Salt andTemperature Induced Aggregationa

aNote that this is only a 2D graphical impression with aspect ratiosand numbers of ligands diverging from reality.

Article

pubs.acs.org/cm

© 2013 American Chemical Society 4297 dx.doi.org/10.1021/cm402414u | Chem. Mater. 2013, 25, 4297−4303

fer (RAFT) polymerization6 have inherent gold bindingabilities originating from the use of chain transfer agents(CTA) bearing a di- or trithio carbonate or xanthate group.7

These sulfur containing compounds can either adsorb assuch onto metallic gold or can be reduced to thiols at the goldsurface; both mechanisms have been suggested in literature.8

Additionally, heterofunctional polymers can be easily designedvia the judicious selection of RAFT chain transfer agents, whichcould be used for further conjugation, thereby widening thescope of applications.9

The preparation of temperature-responsive gold nano-particles has been previously reported via a variety ofmethods.10 Ligand exchange with the sulfur-containing endgroup of RAFT polymers, including temperature-responsivepolymers (Scheme 1A), has definitely emerged as one of themost simple routes to obtain well-defined hybrid polymer@goldNP constructs. Surprisingly, not much has been reportedon how the solvent quality influences the temperature-responsive properties of such polymer@goldNP while saltingin and salting out following the Hofmeister series has beenreported to strongly influence the phase behavior of temper-ature-responsive polymers.11 To our knowledge there is onlyone paper12 describing the influence of salt on the aqueousphase transition behavior of gold nanoparticles coated with atemperature responsive polymer. However, in this study, theaqueous medium contained an excess of free polymer whichwill induce important cooporative effects with the goldNPgrafted polymers, as described by Gibson and co-workers.10c

Therefore, there is still a need to fully understand the influenceof salt on the assembly behavior of pure polymer-coatedgoldNP into higher order structures. In this paper weinvestigate the effect of salt on goldNP coated with atemperature-responsive polymer with respect to temperature-triggered phase transition and hydrogen bonded layer-by-layerassembly.13

2. METHODS2.1. Materials. All chemical were purchased from Sigma-Aldrich

unless otherwise stated. The CTA with methyl ester group wassynthesized by esterification of 3-(((butylthio)carbonothioyl)thio)-2-methylpropanoic acid with methanol.14 Purified Milli-Q grade waterwas used for all experiments.2.2. Polymerization of NIPAm. The RAFT polymerization of N-

isopropyl acrylamide (NIPAm) was performed in a 25 mL Schlenktube under nitrogen atmosphere. In a typical run, NIPAm (2 g, 17,67mmol), methyl 2-(butylthiocarbonothioylthio)propanoate (MBTTC)(44.61 mg, 0.18 mmol), 2,2′-azobisisobutyronitrile (AIBN) (14.51 mg,0.09 mmol), and DMF (8.8 mL) were charged in the 25 mL Schlenktube at a molar ratio of 100:1:0.5. The solution was deoxygenated bythree freeze−pump−thaw cycles. Then the reaction mixture wasplaced in a preheated oil bath at 60 °C to initiate the polymerization.After 50 min of polymerization, the reaction vessel was removed fromthe oil bath and opened to the air to stop the polymerization. Theresulting polymer was precipitated by dropping the polymer solutioninto a large amount of a 50:50 mixture of hexane and diethyl ether.After decantation of the solvent, the polymer was dissolved intetrahydrofuran (THF) and precipitated again in diethyl ether. Thisprecipitation procedure was repeated three times. The resultingpowdery light yellow polymer was dried overnight under vacuum at 50°C.By determining the monomer conversion via gas chromatography

(GC), a theoretical molecular weight, Mn,th, of 6.9 kDa was measured.1H NMR analysis indicated a Mn,NMR of 7.6 kDa. Size exclusionchromatography (SEC) in dimethylacetamide (DMA) showed Mn,SEC= 15.4 kDa against PMMA standards and a dispersity (Đ) of 1.07.

Cloud points were measured on a Cary 300 Bio UV−visiblespectrophotometer at a wavelength of 600 nm with a temperaturecontroller. Aqueous polymer solutions (5 mg/mL) were heated from10 to 40 °C with a heating rate of 1.0 °C/min followed by cooling to10 °C at a cooling rate of 1.0 °C/min while stirring. This cycle wasrepeated two times. The cloud points are reported as the 50%transmittance temperature in the second heating run.

2.3. Synthesis of Citrate Stabilized Gold Nanoparticles.Citrate stabilized gold nanoparticles were synthesized according to thereported literature procedures.15 All glassware was first washed withaqua regia and then rinsed with Milli-Q water several times prior tosynthesis. Briefly, 20 mL of 1 mM HAuCl4 was refluxed for 30 min.Then 2 mL of 1 wt % sodium citrate was quickly added and the colorof solution changed from yellow to wine red within 5 min. Aftercooling, the reaction solution was stored at 4 °C for further research.

2.4. PolyNIPAm Coating of Gold Nanopartciles. A total of 9mL of a citrate stabilized gold nanoparticles solution was mixed with200 μL of an aqueous solution containing 8 mg of polyNIPAm andstirred overnight at room temperature. The resulting conjugates werethree times purified by centrifugation at 4 °C with 10 000 g for 30 minfollowed be redispersion in pure water.

2.5. Characterization. TEM (Transmission Electron Microscopy).A drop of gold nanoparticle solution was allowed to air-dry onto aFormvar-carbon-coated 200 mesh copper grid and visualized using 80keV TEM (Jeol 1010, Japan). TEM images of the goldNPs wereprocessed via ImageJ to determine the number average sizedistribution in the dry state.

UV−Vis Spectroscopy. UV−vis spectra were acquired with aShimadzu UV-1650PC spectrophotometer. Gold nanoparticles withdifferent coatings were placed in plastic cuvettes, and spectral analysiswas performed in the 200−800 nm range at room temperature. In thecase of temperature-dependent measures (gold concentration of 0.09mg/mL), a Cary 300 Bio UV−visible spectrophotometer was utilizedto monitor the absorbance via changing the temperatures from 20 to40 °C.

DLS (Dynamic Light Scattering). The size and zeta potential of theparticles (gold concentration of 0.18 mg/mL) were measured byphoton correlation spectroscopy using a Malvern Zetasizer NanoSeries operating a 4mW He−Ne laser at 633 nm. All samples werefiltered through a 0.02 μm filter prior to analysis. Analysis wasperformed at an angle of 90° and at constantly changing temperaturesfrom 20 to 40 °C at a heating rate of 1.0 °C/min.

2.6. Layer-by-Layer assembly of PolyNIPAm@goldNP andTA. QCM Monitoring of LbL Assembly. Quartz crystal microbalancemonitoring of the Layer-by-Layer assembly was performed using aGamry eQCM 10 M quartz crystal microbalance equipped with anALS flow cell. Gold coated quartz chips with a nominal resonancefrequency of 10 MHz were first precoated by 1 h immersion in anaqueous solution of mercaptosuccinic acid (2 mg/mL) followed byextensive rising with water. Second, the quartz chip was immersed intoan aqueous poly(ethyleneimine) 22 kDa PEI solution (2 mg/mL) for1 h and again extensively washed with water and dried under a gentlenitrogen stream. Next, the chip was mounted into the flow cell andwater was injected and the measurement was started and continueduntil a flat baseline was obtained. Then the measurement was restartedand after 100 s; 0.3 mL of a 2 mg/mL TA solution was injected intothe flow cell. After 100 s, the signal leveled off, indicating thatsaturation of the surface was obtained and longer incubation timeswould not lead to further adsorption of material. The flow cell wasthen rinsed by injection of 0.7 mL of water. Next 0.3 mL ofpolyNIPAm@goldNP solution was injected followed by 0.7 mL ofwater after 100 s. This procedure was repeated for five TA/polyNIPAm@goldNP deposition cycles.

The coverage of the surface upon deposition of the first layer ofpolyNIPAm@goldNP was calculated from the Sauerbrey equation.16

ρ μΔ = − Δf

f

Am

2

.0

2

q q

Chemistry of Materials Article

dx.doi.org/10.1021/cm402414u | Chem. Mater. 2013, 25, 4297−43034298

With f 0 = 10 MHz the fundamental resonance frequency of the quartzcrystal, A = 0.2047 cm2 the electrode area, and ρq (∼2.65 g·cm−2) andμq (∼2.95 g × 1011·cm−2·s−2) the shear modulus and density of quartz,respectively. From the obtained Δf of 253 Hz, Δm/A = 1124 ng·cm−2

was calculated as the mass change per area. Considering the diameterof the goldNP to be 13 nm (as measured by TEM) and the density ofgold to be 19.3 g·cm−3, the number of goldNP per cm2 was calculatedto be 5.066 × 1010 which corresponds to a surface coverage of 6.72%.Atomic Force Microscopy (AFM). The morphology of the

polyNIPAm@goldNP/TA hybrid films was investigated by tappingmode AFM. Experiments were performed on air-dried films depositedonto silicon wafers. Images were obtained in tapping mode underambient conditions in air with Bruker Innova equipped with a 100 μmscanner. Antimony doped silicon AFM probes were used with anominal spring constant of 3 N/m and a resonance frequency of 75kHz. From the number of particles on a 1 μm2 image and a meanparticle diameter of 13 nm (as determined by TEM), a surfacecoverage of 9.32% was calculated, which is in fairly good agreementwith the value obtained by QCM.UV−Vis Monitoring of LbL Assembly. Quartz slides (Hellma

Optics) were cleaned in piranha solution (a 3:1 mixture of sulfuric acidand hydrogen peroxide; Caution, piranha is extremely corrosive andshould be handled with extreme care and not be stored in closedcontainers!) for 10 min, followed by rinsing with water. After thoroughcleaning, the quartz substrate was immersed in a 2 mg/mL ofpolyethylene imine (22 kDa; PEI) solution for 10 min to facilitatedeposition of the subsequent layer. After exhaustive rinsing with water,the slide was immersed in a tannic acid (2 mg/mL) solution for 5 min.After rinsing with water, the substrate was transferred to 2 mg/mL(gold content) polyNIPAm@goldNP solution for 5 min. Thisprocedure was repeated for 10 TA/polyNIPAm@goldNP depositioncycles. After the deposition of each layer, the quartz slide was driedand the UV−vis absorption spectrum was measured using a ShimadzuUV-1650PC spectrophotometer.

3. RESULTS AND DISCUSSION3.1. Synthesis and Characterization of Polymer

Coated goldNP. Poly(N-isopropylacrylamide) (polyNIPAm)was chosen as a model temperature-responsive polymer in thisstudy. PolyNIPAm is one of the most documented temper-ature-responsive polymers with a cloud point temperature (TCP,i.e., the temperature at which the polymer precipitates fromsolution during heating due to entropic reasons) of around 32°C which remains, although being slightly affected by molecularweight and polymer end-group, relatively constant over a wideconcentration range. Well-defined polyNIPAm with Mn,th of 6.9kDa and a dispersity (Đ) of 1.07 was synthesized via RAFTpolymerization according to the reaction scheme depicted inScheme 2. To avoid pH effects caused by the end group of the

RAFT chain transfer agent (CTA), a CTA was synthesized witha methyl ester end group.14 Turbidity measurements on 5 mg/mL aqueous solutions of the synthesized polyNIPAm revealeda TCP of 29.7 °C, which is slightly lower than the typical valueof 32 °C found in literature and can be attributed to theinfluence of the hydrophobic end groups on a relatively lowmolecular weight polyNIPAm.Citrate stabilized goldNP were synthesized following the

Turkevich method by direct reduction of the HAuCl4 salt with

citrate in aqueous medium under reflux.15 A deep red solutionwas obtained containing goldNP with a mean diameter of 13nm ± 3 nm as verified by TEM (processed by automated imageanalysis via ImageJ) and a surface plasmon peak of 525 nm asmeasured by UV−vis spectrophotometry.Mixing of aqueous solutions of citrate goldNP and

polyNIPAm was subsequently performed in the absence andpresence of n-propylamine. n-Propylamine aminolyzes thetrithiocarbonate group of the RAFT agent into a free thiolwhich could facilitate ligand exchange with citrate. On the otherhand also uncleaved RAFT agents have been reported toefficiently adsorb on gold. In this regard, the issue of whetherthe RAFT agent is cleaved on the gold surface or adsorbs assuch is still a matter of debate. In our hands we observed noinfluence of the addition of propylamine. In both cases stablecolloidal solutions were obtained that could be centrifuged andredispersed in pure water without color change.17 This is a firstindication that the goldNP were successfully decorated withpoyNIPAm as bare citrate stabilized goldNP irreversiblyaggregate during centrifugation and removal of the excesscitrate. In total the polyNIPAm@goldNP were centrifuged andredispersed three times to completely remove the excess ofunbound polyNIPAm. We also verified that the supernatantafter the second centrifugation and washing step did not exhibitany phase transition anymore upon heating, indicating effectiveremoval of the polyNIPAm.Clear proof that the goldNP were indeed coated with

polyNIPAm was gained by transmission electron microscopy(TEM). As shown in Figure 1A, bare goldNP are aggregated

and form piled assembliesdue to drying upon samplepreparationon the TEM grid. By contrast, the goldNP thatwere coated with polyNIPAm are well dispersed and clearlyshow the presence of an approximately 3 nm thick coronasurrounding each individual particle (Figure 1B). Next, weanalyzed the goldNP with and without polyNIPAm coating bydynamic light scattering (DLS) and UV−vis spectroscopy.Uncoated goldNP show a mean (volume average) diameter of15 nm on DLS measurements while after coating a hydro-dynamic diameter of 49 nm is measured (Figure 1C). This

Scheme 2. Reaction Scheme of the RAFT Polymerization ofpolyNIPAm

Figure 1. Transmission electron microscopy (TEM) images of (A)uncoated and (B) polyNIPAm coated goldNP. The zoomed image inpanel B2 clearly indicates the presence of a polymer coatingsurrounding the goldNP. Size distribution (C) and UV−vis spectra(D) of uncoated and polyNIPAm coated goldNP obtained via DLSand UV−vis spectroscopy, respectively.

Chemistry of Materials Article

dx.doi.org/10.1021/cm402414u | Chem. Mater. 2013, 25, 4297−43034299

increase in diameter can be attributed to tethering of thepolyNIPAm to their surface as the length of a 6.9 kDa polymeris estimated to be around 12 nm in the fully stretched state.However, from these data the presence of a minor fraction ofaggregates that are formed due to the repetitive centrifugationsteps cannot be completely ruled out. The UV−vis spectrarecorded of the particles in solution reveal a slight red shift ofthe plasmon peak from 525 to 529 nm upon polyNIPAmcoating (Figure 1D). The observation of a slight red shift of theplasmon peak is in accordance with earlier reports by Li et al.18

and Gibson et al.10c stating that red shifts smaller than 20 nmare attributed to polymer grafting rather than the formation ofagglomerates. In this regard it is also noteworthy that TEManalysis of over 1000 polyNIPA@goldNP did not reveal thepresence of aggregates.3.2. Temperature-Responsive Properties of Polymer

Coated goldNP. In a next series of experiments we aimed atinvestigating the temperature responsive properties of thepolyNIPAm coated goldNP, at first in deionized water. Wefound that simply heating a polyNIPAm@goldNP solution(0.18 mg/mL on gold base) above the TCP of the polyNIPAmdid not induce any color change of the goldNP solutions, evenup to the boiling temperature of the aqueous medium. Also nochange in turbidity of the polyNIPAm@goldNP solutions wasobserved. Even very concentrated (1 mg/mL on gold base;achieved via centrifugation) polyNIPAm@goldNP solutionsdid not exhibit any temperature-responsive behavior. Whenthese experiments were repeated with polyNIPAm that wastreated with propylamine to ensure the presence of a thiol endgroup on the polyNIPAm, still no temperature responsivebehavior could be observed. These findings already rule out theinfluence of the polymer end group, i.e., thiol vs trithiocar-bonate, on the synthesis of temperature responsive poly-NIPAm@goldNP. These observations further add to theunresponsive behavior of the nanoparticles themselves. DLSand UV−vis measurements confirmed these findings, yieldingconstant values of particle size and position of the plasmonpeak as function of temperature. Control experiments withfreely soluble polyNIPAm (1 mg/mL) added to goldNP doshow phase transition above the TCP of the polyNIPAm whichis expressed as an increase in turbidity with a slight shift inplasmon peak of the goldNP.However, when the polyNIPAm@goldNP were dispersed in

0.1 M sodium chloride (NaCl), rather than deionized water, thenanoparticles did exhibit temperature-responsive behavior.First, visual proof was gained by simply heating thepolyNIPAm@goldNP solutions until 40 °C, which induced acolor change from red to bluish of the polyNIPAm@goldNPsolution. After cooling back to room temperature the solutionregained its red color. Figure 2 shows a series of snapshotstaken from polyNIPAm@goldNP at different temperatureswith (A) and without (B) salt added to the medium. Fromliterature19 it is also known that salt affects the TCP ofpolyNIPAm. We confirmed this (Figure S1 in SupportingInformation) for the polyNIPAm used in our present study andobserved in the presence of 0.1 M NaCl a decrease in TCP of∼3 °C and a broadening of the temperature interval over whichthe phase transition takes place.To investigate the phase transition behavior of the polymer

coated gold nanoparticles and the effect on their opticalproperties in detail, we monitored their size and plasmon peakas a function of temperature by DLS and UV−vis, respectively.Figure 3A shows that in pure water uncoated citrate stabilized

goldNP remain invariant in size and the plasmonic propertiesalso remain constant (Figure 3B1,C1 over the temperaturerange from 25 to 40 °C, which is in agreement with our earliervisual observations in this paper. However, in the presence of0.1 M NaCl, temperature-responsive behavior emerges. BothDLS and UV−vis indicate the onset of nanoparticle aggregationto be 29 °C, followed by a steady increase in size of theaggregates (Figure 3). At this transition temperature of 29 °C,also the plasmon peak undergoes a red shift from 530 nm at 25°C to 565 nm at 40 °C accompanied by a decrease in absolutevalue of the absorbance (Figure 3B,C).20 Figure S2 inSupporting Information depicts a zoomed view on the intervalof the phase transition. Note that data of uncoated goldNP in0.1 M NaCl are not shown because of immediate precipitationfrom solution of the goldNP. In salt solutions, the electrostaticcharges on the surface of the goldNP are screened and do notprovide sufficient repulsion to keep the goldNP colloidallystable. The substantial red shift in plasmon peak of thepolyNIPAm@goldNP in salt above the TCP of the polyNIPAmcan unambiguously be attributed to electronic dipole−dipoleinteraction between plasmons of goldNP that come into closeproximity with each other.21

Next, we measured the zeta-potential values (Figure S3 inSupporting Information) of the polyNIPAm@goldNP in thepresence and absence of salt. A strongly negative value of −40mV was measured for polyNIPAm@goldNP in the absence ofsalt which is only slightly less negative than the zeta-potential of−47 mV measured for bare citrate@goldNP. This suggests thata certain fraction of citrate is still bound to the goldNP surfaceafter ligand exchange with PNIPAm, providing the negativezeta-potential. However, when we measure the zeta-potential inthe presence of 0.1 M of NaCl, a near neutral value of −1 mV isobserved, which we attribute to screening of the negative citratecharges by Na+ ions. Thus, in absence of salt, when thepolyNIPAm chains collapse above the TCP, there is stillelectrostatic repulsion between the gold nanoparticles due tothe remaining citrate groups leading to a stable colloidal

Figure 2. Photographs of plastic cuvettes containing polyNIPAm@goldNP in water (A) and in 0.1 M NaCl (B). The numbers (1→ 3) inthe upper left corner refer to polyNIPAm@gold solutions (1) at roomtemperature, (2) heated to 40 °C, and (3) cooled back to roomtemperature.

Chemistry of Materials Article

dx.doi.org/10.1021/cm402414u | Chem. Mater. 2013, 25, 4297−43034300

solution.22 By contrast, in the presence of salt, the charges arescreened, thereby strongly reducing electrostatic repulsion andallowing the dehydrated polyNIPAm chains to agglomerateabove the TCP leading to agglomeration of the AuNPs.Additionally, the effect of salt on the TCP of polyNIPAm(Figure S2 in Supporting Information; 0.1 M NaCl induces ashift of the TCP to lower temperature and a broadening of thephase transition interval) indicates that more pronounceddehydration of the polyNIPAm chains in the presence of NaCl(i.e., salting out effect) can contribute to enhanced interparticlehydrophobic interactions by enhancing dehydration of thepolyNIPAm chains. Scheme 1B schematically represents theinterplay between salt and temperature responsive properties ofthe polyNIPAm@goldNP.3.3. Hydrogen Bonded Layer-by-Layer Assembly with

Polymer Coated goldNP. In a final series of experiments weaimed at elucidating whether salt also plays a crucial role in theassembly of higher order structures containing polyNIPAm@goldNP. As water-soluble amide-containing polymers, such aspolyNIPAm, are known to form donor−acceptor complexeswith hydrogen bond donors such as tannic acid (TA),23 weattempted to assemble the polyNIPAm@goldNP together withTA in a layer-by-layer (LbL) fashion. This means that TA andpolyNIPAm@goldNP will be alternately adsorbed on asubstrate while hydrogen bonding between the phenol groupsof TA and the amide bonds of polyNIPAm, likely aided byhydrophobic interactions as well, will act as driving force formultilayer. Scheme 3 schematically represents the interactionbetween the amide groups of the polyNIPAm and the phenolgroups of TA. In the framework of the present research weaimed at elucidating the effect of salt on the formation of TA/polyNIPAm@goldNP multilayers.Figure 4A shows UV−vis spectra recorded from TA/

polyNIPAm@goldNP multilayers absorbed on quartz slides inthe absence and presence of 0.1 M NaCl salt. Additionally alsothe evolution of the absorbance at 279 nm (i.e., adsorption

maximum of TA) and 560 nm (i.e., the plasmon peak of thepolyNIPAm@goldNP in the multilayer film) is represented as afunction of the number of deposition cycles (Figure 4B). Inboth conditions, i.e., with and without salt, alternatingdeposition of TA and polyNIPAm@goldNP onto quartz slidesleads to an increase in absorbance. However, in the presence of0.1 M of NaCl salt, a dramatic increase in the amount ofadsorbed material is observed. The same trend is also witnessedby real time quartz crystal microbalance (QCM) monitoring ofthe assembly of TA and polyNIPAm@goldNP. Indeed, asshown in Figure 4C2, a larger drop in resonance frequency ofthe quartz chip is depicted when TA and polyNIPAm@goldNPwere alternately flown over the quartz chip in the presence ofsalt versus in the absence of salt. The influence of saltconcentration on layer-by-layer assembly of oppositely chargedpolymers via electrostatic interaction is well-known.24 Salt ionsscreen the electrostatic charges along the polymer chains

Figure 3. (A) Evolution of particle size as function of temperature measured by DLS. (B) UV−vis absorbance as function of temperature ofpolyNIPAm@goldNP in (B1) water and (B2) 0.1 M NaCl. (C) Evolution of the wavelength of maximum absorbance (red curve) and theabsorbance at this wavelength (blue curve) as a function of temperature of polyNIPAm@goldNP in (C1) water and (C2) 0.1 M NaCl.

Scheme 3. Schematic Representation of the HydrogenBonding Interaction between TA and polyNIPAm@goldNP

Chemistry of Materials Article

dx.doi.org/10.1021/cm402414u | Chem. Mater. 2013, 25, 4297−43034301

thereby inducing a rod to coil conformational change whichallows a higher mass to be deposited during each depositionstep. However, in the case of hydrogen bonded layer-by-layerassembly, electrostatic interaction does not play a role and theinfluence of salt on the assembly behavior is, contrary to theinfluence of pH,25 unexplored. In accordance to our findings inSection 3.2 of this work, the role of salt will be a combination ofreducing the electrostatic repulsion between the polyNIPAm@goldNP and enhancing hydrophobic interactions. From boththe UV−vis and QCM measurements the most significantgrowth in layer thickness is witnessed for the first depositioncycle of polyNIPAm@goldNP. This suggests that likely nodistinct strata of TA and nanoparticles are obtained.To further characterize the film growth, we performed

atomic force microscopy (AFM) to characterize the TA/polyNIPAm@goldNP multilayers formed in 0.1 M aqueous

NaCl solution. Figure 5 shows 1 × 1 μm scans of siliconsubstrates that were coated with respectively 1, 5, and 10 cylcesof TA/polyNIPAm@goldNP. After the first deposition cycle ofTA/polyNIPAm@goldNP, the substrate is clearly coated with alayer of nanoparticles. From this AFM image and taking intoaccount a diameter of the goldNP of 13 nm (measured byTEM), we calculated a surface coverage of 9.3% (on gold base).We also calculated the surface coverage based on the shift inresonance frequency Δf = 253 Hz measured by QCM uponadsorption of the polyNIPAm@gldNP onto a TA layer (seeSection 2.6 in the Methods section for the detailed calculation)and obtained a value of 6.7%, which is in fairly good agreementwith the values derived from the AFM data. The AFM imagescorresponding to 5 and 10 deposition cycles reveal featuresbeyond the dimensions of a single particle and correspond topiled assemblies of particles. These observations confirm the

Figure 4. UV−vis spectra recorded during layer-by-layer assembly of polyNIPAm@goldNP and tannic acid in water (A1) and in 0.1 M sodiumchloride (A2). Absorption values of multilayers of polyNIPAm@goldNP and tannic acid at 279 nm (B1) and at 560 nm (B2). (C) QCM data oflayer-by-layer assembly of polyNIPAm@goldNP and tannic acid in water and 0.1 M sodium chloride. Panel C1 exemplifies the evolution of Δf as afunction of time during the assembly of TA/polyNIPAm@goldNP and panel C2 depicts the evolution of Δf as a function of the number of layers.

Figure 5. AFM images in tapping mode (topography channel) of TA/polyNIPAm@goldNP films after 1 (A), 5 (B), and 10 (C) assembly steps. Theheight profile was measured along the X axis in the middle of the image, as indicated by the dotted line.

Chemistry of Materials Article

dx.doi.org/10.1021/cm402414u | Chem. Mater. 2013, 25, 4297−43034302

UV−vis and QCM data suggesting a smaller film growth afterthe first TA/polyNIPAm@goldNP deposition cycle.

4. CONCLUSIONSIn summary, we have demonstrated that salt plays a crucial roleregarding the temperature responsive behavior of purifiedcitrate stabilized goldNP that were grafted through ligandexchange with thiol-functional polyNIPAm synthesized viaRAFT polymerization. Whereas in pure water these particleswere found insensitive to temperature, the polyNIPAm@goldNP did show temperature-responsive behavior in aqueous0.1 M sodium chloride solutions. Additionally, the presence ofsalt also strongly enhances layer-by-layer build-up of poly-NIPAm@goldNP with tannic acid. We attribute theseobservations to a combination of ionic screening of remainingcitrate groups on the nanoparticle surface and a salting outeffect on the polyNIPAm, causing a stronger dehydration of thepolyNIPAm chains, thereby enhancing hydrophobic interac-tions. We believe our findings will contribute to a betterunderstanding of the self-assembly behavior of polymer@goldNP hybrid nanostructures. In forthcoming work we willexplore multiple types of temperature-responsive systems whichalso appear to yield salt-dependent responsive metal nano-particles. This is important in view of the design of functionalhigher order structure for a wide variety of applications rangingfrom diagnostics to catalytically and optically active assemblies.

■ ASSOCIATED CONTENT*S Supporting InformationAdditional experimental data. This information is available freeof charge via the Internet at http://pubs.acs.org/.

■ AUTHOR INFORMATIONCorresponding Authors*E-mail: richard.hoogenboom@ugent.be (R.H.).*E-mail: br.degeest@ugent.be (B.G.D.G.).Author Contributions∥These authors equally contributed as first author (Z.Z. andS.M.).Author Contributions⊥These authors equally supervised the work (R.H. andB.G.D.G.).NotesThe authors declare no competing financial interest.

■ ACKNOWLEDGMENTSZ.Z. and Q.Z. gratefully acknowledge the Chinese ScholarshipCouncil for a PhD scholarschip and Ghent University for BOFcofunding. S.M. is grateful to FWO for the Pegasus Marie Curiepostdoctoral fellowship. R.H. thanks the Belgian Program onInteruniversity Attraction Poles initiated by the Belgian State,the Prime Minister’s office (P7/541 05), and the EuropeanScience Foundation − 542 Precision Polymer Materials (P2M)program for financial support. B.G.D.G. acknowledges GhentUniversity (BOF-GOA) and the FWO Flanders for funding.

■ REFERENCES(1) De, M.; Ghosh, P. S.; Rotello, V. M. Adv. Mater. 2008, 20, 4225−4241.(2) (a) Dreaden, E. C.; Alkilany, A. M.; Huang, X. H.; Murphy, C. J.;El-Sayed, M. A. Chem. Soc. Rev. 2012, 41, 2740−2779. (b) Sperling, R.

A.; Rivera gil, P.; Zhang, F.; Zanella, M.; Parak, W. J. Chem. Soc. Rev.2008, 37, 1896−1908.(3) (a) Mirkin, C. A.; Letsinger, R. L.; Mucic, R. C.; Storhoff, J. J.Nature 1996, 382, 607−609. (b) Elghanian, R.; Storhoff, J. J.; Mucic,R. C.; Letsinger, R. L.; Mirkin, C. A. Science 1997, 277, 1078−1081.(4) (a) Radt, B.; Smith, T. A.; Caruso, F. Adv. Mater. 2004, 16,2184−2189. (b) Skirtach, A. G.; Javier, A. M.; Kreft, O.; Kohler, K.;Alberola, A. P.; Mohwald, H.; Parak, W. J.; Sukhorukov, G. B. Angew.Chem., Int. Ed. 2006, 45, 4612−4617.(5) Foss, C. A.; Hornyak, G. L.; Stockert, J. A.; Martin, C. R. J. Phys.Chem. 1992, 96, 7497−7499.(6) Moad, G.; Rizzardo, E.; Thang, S. H. Aust. J. Chem. 2005, 58,379−410.(7) Sumerlin, B. S.; Lowe, A. B.; Stroud, P. A.; Zhang, P.; Urban, M.W.; McCormick, C. L. Langmuir 2003, 19, 5559−5562.(8) Slavin, S.; Soeriyadi, A. H.; Voorhaar, L.; Whittaker, M. R.; Becer,C. R.; Boyer, C.; Davis, T. P.; Haddleton, D. M. Soft Matter 2012, 8,118−128.(9) Boyer, C.; Bulmus, V.; Davis, T. P.; Ladmiral, V.; Liu, J. Q.;Perrier, S. Chem. Rev. 2009, 109, 5402−5436.(10) (a) Boyer, C.; Whittaker, M. R.; Luzon, M.; Davis, T. P.Macromolecules 2009, 42, 6917−6926. (b) Shan, J.; Zhao, Y. M.;Granqvist, N.; Tenhu, H. Macromolecules 2009, 42, 2696−2701.(c) Gibson, M. I.; Paripovic, D.; Klok, H. A. Adv. Mater. 2010, 22,4721−4725.(11) (a) Weber, C.; Hoogenboom, R.; Schubert, U. S. Prog. Polym.Sci. 2012, 37, 686−714. (b) Zhang, Y. J.; Furyk, S.; Bergbreiter, D. E.;Cremer, P. S. J. Am. Chem. Soc. 2005, 127, 14505−14510.(12) Yusa, S. I.; Fukuda, K.; Yamamoto, T.; Iwasaki, Y.; Watanabe,A.; Akiyoshi, K.; Morishima, Y. Langmuir 2007, 23, 12842−12848.(13) (a) Decher, G. Science 1997, 277, 1232−1237. (b) Ferreira, M.;Rubner, M. F. Macromolecules 1995, 28, 7107−7114. (c) Quinn, J. F.;Johnston, A. P. R.; Such, G. K.; Zelikin, A. N.; Caruso, F. Chem. Soc.Rev. 2007, 36, 707−718. (d) Kharlampieva, E.; Sukhishvili, S. A. Polym.Rev. 2006, 46, 377−395.(14) Zhang, Q.; Vanparijs, N.; Louage, B.; De Geest, B. G.;Hoogenboom, R. Polym. Chem. 2013, DOI: 10.1039/C3PY00971H.(15) Turkevich, J.; Stevenson, P. C.; Hillier, J. Discuss. Faraday Soc.1951, 11, 55−75.(16) (a) Sauerbrey, G. Z. Phys. 1959, 155, 206−222. (b) Ko, Y.;Baek, H.; Kim, Y.; Yoon, M.; Cho, J. ACS Nano 2013, 7, 143−153.(17) Schneider, G.; Decher, G. Nano Lett. 2004, 4, 1833−1839.(18) Li, D. X.; He, Q.; Yang, Y.; Mohwald, H.; Li, J. B.Macromolecules 2008, 41, 7254−7256.(19) Baltes, T.; Garret-Flaudy, F.; Freitag, R. J. Polym. Sci., Part A:Polym. Chem. 1999, 37, 2977−2989.(20) (a) Itoh, H.; Naka, K.; Chujo, Y. J. Am. Chem. Soc. 2004, 126,3026−3027. (b) Raula, J.; Shan, J.; Nuopponen, M.; Niskanen, A.;Jiang, H.; Kauppinen, E. I.; Tenhu, H. Langmuir 2003, 19, 3499−3504.(21) Heath, J. R.; Knobler, C. M.; Leff, D. V. J. Phys. Chem. B 1997,101, 189−197.(22) Stocco, A.; Chanana, M.; Su, G.; Cernoch, P.; Binks, B. P.;Wang, D. Angew. Chem., Int. Ed. 2012, 51, 9647−9651.(23) (a) Shutava, T.; Prouty, M.; Kommireddy, D.; Lvov, Y.Macromolecules 2005, 38, 2850−2858. (b) Erel-Unal, I.; Sukhishvili, S.A. Macromolecules 2008, 41, 8737−8744. (c) Kozlovskaya, V.;Kharlampieva, E.; Drachuk, I.; Cheng, D.; Tsukruk, V. V. Soft Matter2010, 6, 3596−3608. (d) Lomas, H.; Johnston, A. P. R.; Such, G. K.;Zhu, Z. Y.; Liang, K.; van Koeverden, M. P.; Alongkornchotikul, S.;Caruso, F. Small 2011, 7, 2109−2119.(24) Dubas, S. T.; Schlenoff, J. B. Macromolecules 2001, 34, 3736−3740.(25) (a) Stockton, W. B.; Rubner, M. F. Macromolecules 1997, 30,2717−2725. (b) Khutoryanskiy, V. V.; Dubolazov, A. V.; Nurkeeva, Z.S.; Mun, G. A. Langmuir 2004, 20, 3785−3790.

Chemistry of Materials Article

dx.doi.org/10.1021/cm402414u | Chem. Mater. 2013, 25, 4297−43034303