Facile synthesis of core-shell organic–inorganic hybrid nanoparticles with amphiphilic polymer shell by one-step sol–gel reactions

Facile Synthesis of Core-Shell Organic–Inorganic HybridNanoparticles with Amphiphilic Polymer Shell byOne-Step Sol–Gel Reactions

M. MARINI,1,2 M. TOSELLI,2,3 S. BORSACCHI,4 G. MOLLICA,4 M. GEPPI,4 F. PILATI1,2

1Dipartimento di Ingegneria dei Materiali e dell’Ambiente, Universita di Modena e Reggio Emilia,Via Vignolese 905/A, 41100 Modena, Italy

2NIPLAB-INSTM Consortium, Reference Centre, Via Giusti 9, 50121 Firenze, Italy

3Dipartimento di Chimica Applicata e Scienza dei Materiali, Universita di Bologna, Viale Risorgimento 2,40136 Bologna, Italy

4Dipartimento di Chimica e Chimica Industriale, Universita di Pisa, Via Risorgimento 35, 56126 Pisa, Italy

Received 3 September 2007; accepted 31 October 2007DOI: 10.1002/pola.22511Published online in Wiley InterScience (www.interscience.wiley.com).

ABSTRACT: Organic–inorganic hybrid core-shell nanoparticles with diameters rangingfrom 100 to 1000 nm were prepared by a one-pot synthesis based on base catalyzedsol–gel reactions using tetraethoxysilane and a triethoxysilane-terminated polyethyl-ene-b-poly(ethylene glycol) as reactants. Data from TEM, TGA, and solid-state NMRanalysis are in agreement with the formation of core-shell nanoparticles with an inor-ganic-rich core and an external shell consisting of an amphiphilic block copolymermonolayer. The influence of the organic–inorganic ratio, solution concentration, andpostcuring temperature on core and shell dimensions of the nanospheres were inves-tigated by TEM microscopy. VVC 2008 Wiley Periodicals, Inc. J Polym Sci Part A: Polym Chem

46: 1699–1709, 2008

Keywords: amphiphiles; nanocomposites; nanoparticles; organic–inorganic; sol–gel

INTRODUCTION

Core-shell nanoparticles have recently been thefocus of a lot of scientific efforts because of thecombination of different properties in one parti-cle based on different compositions of the coreand the shell.1–3 The core often shows the rele-vant property while the shell helps to stabilizethe core, to create compatibility between the

core and the environment, to promote the hydro-carbon dispersion,4 or can change the charge,functionality, or reactivity of the surface.5 Thisis especially important when the shell is gener-ated by an alkylsiloxane, so that the nanosphereresults completely dispersible in hexane, orwhen it is a polymer, so that the final core-shellparticles can be homogeneously dispersed in apolymer matrix at the nanoscale level.6

Core-shell architectures are generally theresult of either a two-step approach, consistingof preparation of nanoparticles and then modifi-cation of their surface, or an in situ approachduring particle formation.5 In particular, inor-ganic–organic hybrid core-shell nanoparticles

Correspondence to: F. Pilati (E-mail: pilati.francesco@unimore.it)

Journal of Polymer Science: Part A: Polymer Chemistry, Vol. 46, 1699–1709 (2008)VVC 2008 Wiley Periodicals, Inc.

1699

were typically prepared by a two-step synthesisstarting with the inorganic core preparation bymicroemulsion technique,5 that allows to pre-pare a variety of particles such as SiO2,

7–10

TiO2,11–13 ZrO2.

14 Then the polymer shell can begrafted onto the surface of inorganic particleseither by using trialkoxy-substitued silane cou-pling agents,15,16 or by polymerization of mono-mers via a surface-immobilized initiator,17–20 orby reacting a preformed functional polymercontaining reactive groups with the surfaceof the inorganic particle,21–23 or by polymerencapsulation.24,25

Indeed, it is a common practice to soak thesilica surface with functionalized siloxanes26

which can act as coupling agents: the siloxaneportion can interact/react with the surface of thefiller, while the organic groups can interact/reactwith the polymer matrix. Silica treated with c-aminopropyl triethoxysilane, for example, wasmixed with maleated-polyolefins, to obtain agood interfacial interaction as a result of chemi-cal reaction between amino and anhydridegroups to give cyclic imides.27 Silica particlessurface-modified by grafting of octenyl- or octyl-silanes have been used in catalytic emulsion po-lymerization of ethylene with nickel catalysts,giving rise to stable dispersions of silica/polyeth-ylene nanocomposite particles.28 However thegrafting of organic products onto the surface offillers is not an easy task, in fact a nonhomoge-neous surface morphology with the formation ofagglomerates is often obtained.16,29

Although there is a wide literature on thisargument, only few works are focused on thesol–gel preparation of core-shell inorganic–organic hybrid nanoparticles30–33 and, at thebest of our knowledge, no articles describe thepreparation of hybrid core-shell nanoparticles byone-pot reaction, except for a recent article ofFei et al.,33 where organosilica-chitosan nano-spheres were prepared by concurrent graftingpolymerization and sol–gel reaction using analkyloxosilane methacrylate simultaneouslypolymerized with an hydroperoxide.

In this article, we report results about thepreparation of organic–inorganic hybrid core-shell nanoparticles (with hybrid core and anamphiphilic polymer shell), obtained by one-potreaction using the sol–gel chemistry34 under ba-sic catalysis starting from tetraethyl orthosili-cate, TEOS, as SiO2 precursor, and a-triethoxy-silane-terminated polyethylene-b-poly(ethyleneglycol). The resulting products were character-

ized through TEM, 13C and 29Si solid-stateNMR, TGA, and DSC.

EXPERIMENTAL

Materials

High-purity TEOS (Aldrich), 3-isocyanatopropyl-triethoxysilane (ICPTES, Fluka), sodium hy-droxide in pellets (NaOH, Carlo Erba), ethanol(EtOH, Carlo Erba), tetrahydrofuran (THF, CarloErba), a-hydroxy-terminated poly(ethylene)-block-poly(ethylene glycol) (PE–PEG 20/80 wtratio, Mn 2250, Aldrich) were used as receivedwithout further purification.

Preparation of Triethoxysilane-TerminatedCopolymer (PE–PEG–Si)

The reaction of PE–PEG–OH copolymer withICPTES was carried out as reported in a previ-ous article.35 In a 50-mL glass flask equippedwith magnetic stirring, ICPTES was directlyadded at 1.1:1 molar ratio to the copolymer pre-viously heated at 120 8C. The progress of thereaction between hydroxyl groups of the copoly-mer and isocyanate groups of ICPTES wasmonitored using FTIR spectroscopy, by followingthe formation of the strong absorption band ofthe carbonyl of the urethane groups (at about1700 cm�1) and the disappearance of the bandrelated to isocyanate groups (at 2270 cm�1). Itwas found that, under the experimental condi-tions used in this study, the reaction goes tocompletion within 3 h. The molecular structureof the resulting product was confirmed to be thea-triethoxysilane-terminated copolymer (PE–PEG–Si) by 1H NMR.

Preparation of PE–PEG–Si/Silica Hybrids

The PE–PEG–Si and TEOS were dissolved intetrahydrofuran in the desired organic–inor-ganic ratio at the concentration of about 30% w/v.Then water (for the hydrolysis reaction), EtOH(to make the system homogeneous), and NaOH(as catalyst) were added at the following molarratios with respect to ethoxide groups of bothPE–PEG–Si and TEOS: EtO�:H2O:EtOH: NaOH¼ 1:1:1:0.05 (pH solution about 11).

As described in a previous article,35,36 a typi-cal preparation of PE–PEG–Si/SiO2 50:50 wasas follows (with NaOH replacing HCl): 0.70 g of

1700 MARINI ET AL.

Journal of Polymer Science: Part A: Polymer ChemistryDOI 10.1002/pola

a PE–PEG–Si and 2.30 g of TEOS were addedto 10 mL of tetrahydrofuran and mixed until ahomogeneous solution was obtained. Then 2.92 gof a mixture of EtOH, water, and NaOH (inmolar ratio 1:1:0.05) were added under stirring.The final hybrids were coded as PE–PEG–Si/SiO2 x:y, in which x:y is the weight ratiobetween the organic and inorganic components,assuming the occurrence of complete hydrolysisand condensation reactions according to the typ-ical reaction scheme of sol–gel processes.37

Hybrid nanoparticles with 80:20, 50:50, and20:80 nominal organic–inorganic weight ratioswere prepared. The solutions were dropped intoTeflon Petri dishes, and cured at 60 or 100 8Cfor 24 h when no presence of solvent was detect-able. The final hybrids were ground by using anagate pestle and mortar. To extract theunreacted PE–PEG–Si and the organic phasenot bonded to silica-rich nanoparticles, sampleswere dispersed in warm THF and then a centri-fuge (at 4000 rpm for 50) was used to separatethe solid nanoparticles from the supernatant liq-uid containing the organic-rich products thatare either soluble or highly swollen in warmTHF. This procedure was repeated until nochanges in the TGA curves were observed.

Sample Characterization

To characterize the nanoparticles (core andshell, size, and composition) an Electron Trans-mission Microscope Jeol JEM 2010 (200 kV)equipped with a GIF Gatan Camera and amicroanalyzer Link Inca 100 was used. ForTEM measurements, samples were dispersed inwarm toluene to form a diluted solution anddropped on carbonated nets. EDS spectra wereperformed with a X-EDS Oxford INCA350equipped on the TEM microscope.

An environmental scanning electron micro-scope ESEM Quanta-200 was used to obtainSEM micrographs. An accelerating voltage of20 kV and a magnification of 100k3 were used.

A Perkin Elmer TGA 7 was used to estimatethe organic and inorganic fractions in the vari-ous samples. TGA experiments were performedin air with a heating rate of 20 8C/min.

Differential scanning calorimetry (DSC) wasperformed with a TA DSC2010 instrument inthe range of �100 to þ150 8C with a heatingrate of 10 8C/min. Melting temperature andmelting enthalpy were determined on the basis

of the curves recorded during the second heatingscan.

Solid-state NMR was used to get a moredetailed information about the molecular struc-ture of the products resulting from sol–gel reac-tions. NMR experiments were carried out on adouble-channel Varian InfinityPlus 400 spec-trometer, equipped with a 7.5-mm Cross Polar-ization/Magic Angle Spinning (CP/MAS) probe-head, working at a Larmor frequency of 399.89MHz for proton, 100.75 MHz for carbon-13, and79.44 MHz for silicon-29. 13C and 29Si spectrawere recorded under high-power proton decou-pling conditions, with 908 pulse lengths between4 and 7 ls. Quantitative 13C-Direct Excitation(DE)/MAS spectra were recorded at a MAS fre-quency of 6 kHz, using a recycle delay of 400 s,collecting 600 scans, using a depth pulsesequence for background suppression.38 29Si-CP/MAS spectra were acquired at a MAS frequencyof 5.5 kHz, with a contact time of 3 ms, a recycledelay of 4 s, and collecting 16,000 scans. Quantita-tive 29Si-Single Pulse Excitation (SPE)/MAS spec-tra were recorded at a MAS frequency of 5.5 kHz,with a recycle delay of 300 s, and collecting 620scans. All the spectrawere acquired at 256 0.1 8C.

RESULTS AND DISCUSSION

Formation of Core-Shell Particles bySol–Gel Chemistry

The classical sol–gel process consists in a two-step acid or base catalyzed reaction, startingfrom metal alkoxides M(OR)4, typicallyTEOS.34,39 Even though the reaction mechanismmay be quite complex and until now it is notcompletely understood, it is well-known thatseveral reactions can occur involving metal alk-oxides hydrolysis and formation of intermediatespecies of the metal hydroxide type. These spe-cies then undergo a stepwise polycondensationinvolving both hydroxy and alkoxy groups withthe formation of a metal oxide three-dimensionalnetwork. The sol–gel process has proved to beflexible enough for an efficient incorporation oforganic oligomers or polymers; in particular thisapproach is interesting for organic productsbearing reactive groups able to participate incondensation reactions, so that they can be cova-lently linked to silica domains. It is also impor-tant to note that different morphologies oforganic and inorganic domains are expected

CORE-SHELL ORGANIC–INORGANIC HYBRID NANOPARTICLES 1701

Journal of Polymer Science: Part A: Polymer ChemistryDOI 10.1002/pola

using different catalysts (acidic or basic).40–43

Linear and branched siloxane chains grow pref-erably under acidic conditions before formingthe final network, whereas, under basic cataly-sis, particles first grow and then link themselvesto form a three-dimensional network of particlesin the final part of the process. When sphericalnanoparticles are the desired product of the sol–gel process, basic catalysis and high water/Siratios (from 7 to 25 mol/mol) are typicallyused.42

When amphiphilic block copolymers with re-active groups just at one end of the chains areused under basic catalysis along with metal alk-oxides in a suitable organic/inorganic ratio, itcould in principle be possible to obtain taperedcore-shell particles in one-pot synthesis, accord-ing to the scheme reported in Figure 1. More-over, if the polymer chains are able to behavelike surfactants it should be possible to obtainnanoparticles even using a H2O/Si molar ratiotypical of sol–gel processes that aim at reachinggelation, instead of the higher H2O/Si molar ra-tio suggested in the literature42 for the prepara-tion of all inorganic spherical nanoparticles. Toverify this hypothesis, TEOS and PE–PEG–Si,as precursors of inorganic and organic phase,respectively, were reacted in the presence ofNaOH as catalyst (pH about 11), at H2O/Si

molar ratio of 4, and under the reaction condi-tions reported in Table 1.

The resulting crude products were dispersedin toluene, and a drop of this solution was puton a carbonated net; the residue remaining aftersolvent evaporation was investigated by TEM.Typical TEM pictures are reported in Figure 2.As it appears from Figure 2(a), the crude prod-uct resulting from the sol–gel process is an ag-gregate of spherical particles (black) envelopedby a shapeless gel-like mass (grey). Isolated par-ticles were also present in the dispersion of thecrude product showing a black core and a greyshell. EDS analysis confirmed that the nanopar-ticle core is rich in silicon, while no significantamount of silicon was present in the shell andin the gel-like phase as reported in the followingFigure 3. While the shell should be formed byPE–PEG–Si covalently bonded to the inorganic-rich core, it seems reasonable to suppose thatthe gel-like phase is constituted by unreactedorganic material and/or PE–PEG–Si that hasreacted to form clusters or micelles with a smallsilica content.

Characterization of Core-Shell Particles

Whereas PE–PEG–Si bonded to silica (eitherpresent in the core or on the shell) is not

Figure 1. Scheme of the nanoparticles formation.

Table 1. Effect of Reaction Conditions on the Dimensions of PE–PEG–Si/SiO2 Nanoparticles

Samples

Organic–InorganicRatio

Type ofCatalyst

Concentration(% w/v)

CuringTemp. (8C)

ParticleDiameter

(nm)

CoreDiameter

(nm)

ShellThickness

(nm)

A 80/20 Basic 30 60 No particles formationB 20/80 Basic 30 60 1600–2000 1500–1800 20–30C 50/50 Basic 30 60 300–700 250–650 20–30D 50/50 Basic 10 60 300–700 200–600 20–30E 50/50 Basic 30 100 200–280 160–220 20–30F 50/50 Acid 30 100 No particles formation

1702 MARINI ET AL.

Journal of Polymer Science: Part A: Polymer ChemistryDOI 10.1002/pola

extractable by solvents, both unreacted PE–PEG–Si and PE–PEG–Si in clusters or micellesshould be either miscible or deeply swollen bywarm THF. Therefore, to get information aboutthe fraction of PE–PEG bonded to silica in thenanoparticles, and for a better characterizationof nanoparticles, the solid crude product result-ing from sol–gel reactions was soaked withwarm THF. The warm suspension was then cen-trifuged and the supernatant liquid, whichshould contain both unreacted PE–PEG–Si andPE–PEG–Si clusters or micelles (as the densityof PE–PEG–Si clusters or micelles swollen inTHF is expected to be much lower than that ofsilica-rich particles) was removed before submit-ting the residual solid to further extractions andthen to further investigation.

TGA curves of various products, namely PE–PEG–Si, the crude solid product resulting fromthe sol–gel process, and the solid recovered aftercentrifugation, are reported in Figure 4 forthe 50:50 organic–inorganic ratio composition(Sample E). As it appears, the crude product hasa residue of 52% above 500 8C, in good agree-

ment with the amount of SiO2 expected from theinitial composition after complete hydrolysis andcondensation of the alkoxy groups. In the 100–250 8C temperature range, there is a significantweight decrease that can be attributed either toa completion of the hydrolysis/condensationreactions (with evolution of water or ethanol) orto the removal of volatile unreacted products.When the solid product resulting after THFextraction was submitted to TGA, the residueabove 500 8C rised to 63%; this small but signifi-cant increase of the final SiO2, as well as thedisappearance of weight loss in the initial partof the TGA curve, suggest that a fraction oforganic phase has been removed by warmTHF. TEM and SEM pictures taken on samplesafter THF extraction are reported also in Fig-ures 2(d–f) and 5, respectively. As it appearsfrom the comparison of Figure 2(a,d), most ofthe intraparticles gel-like organic phase hasbeen removed by the THF treatment, eventhough a residual fraction bounded or strictlyentangled with the nanoparticle shell wasretained. According to the TEM/SEM micro-

Figure 2. TEM micrographs of Samples E, B, and C before extraction (a, b, and c,respectively) and Sample E after extraction (d, e, and f).

CORE-SHELL ORGANIC–INORGANIC HYBRID NANOPARTICLES 1703

Journal of Polymer Science: Part A: Polymer ChemistryDOI 10.1002/pola

graphs in Figures 2 and 5, most of the nanopar-ticle cores obtained for the 50/50 organic/inor-ganic composition have a diameter in the regionof 160–220 nm, and a shell about 20–30 nmthick. Figure 5 reveals also the presence of someunusual ordered particle aggregations, consistingof hexagonally arranged particles with one parti-cle fitting the centre. This feature is difficult toexplain within the bounds of this work. From acomparison of the data in Figures 2 and 5 it wouldseem, however, that this may be associated withthe way the solution has been deposited on theTEM carbonated nets, rather than organized par-ticles within the bulk of the sample.

To have a rough idea of the distribution of theorganic and inorganic components into the nano-particles, from the TEM images of an isolatedrepresentative particle, reported in Figure2(d,e), we estimated core and shell volumes(taking a core diameter of 206 nm and a 20-nm

shell thickness), and, assuming that the densityis close to 1 and 2 g/cm3 for organic and inor-ganic phase, respectively, it could be inferredthat about 60–70% of the PE–PEG–Si block co-polymer is present in the organic shell. Theremaining part of the copolymer is surely pres-ent as gel-like phase remaining after the THFextraction, and might also be partially includedin the core (hybrid core). It is also worth notingthat the average thickness of the organic shell isvery close to the chain-length of the PE–PEG co-polymer in its extended conformation (26 nm,calculated assuming that C��C and C��O bondsare 1.54 and 1.43 A, respectively, and consider-ing the C��C��C and C��O��C dihedral anglesof 1098 and 113.58, respectively) suggesting thatthe sol–gel process leads to the formation of a

Figure 3. EDS analysis of a PE–PEG–Si/SiO2 core-shell nanosphere.

Figure 4. TGA analysis of PE–PEG–Si copolymer(a) and of Sample E before (b) and after extraction (c)with warm THF.

Figure 5. SEM micrographs of Sample E afterextraction.

1704 MARINI ET AL.

Journal of Polymer Science: Part A: Polymer ChemistryDOI 10.1002/pola

shell consisting of a monolayer of PE–PEG blockcopolymer chains in the extended conformation.It is also interesting to note that the sol–gel pro-cess leads to a quite narrow size distribution(see Table 1) when a 50:50 organic–inorganic ra-tio composition was postcured at 100 8C startingfrom a solution concentration of 30% w/v.

Polymer Chain Conformation in the Particles

To obtain information about the conformationalproperties of the organic fraction of the hybridnanoparticles, quantitative 13C DE/MAS NMRspectra were recorded on the THF-washed andunwashed sample E (Fig. 6). For both samples,peaks at 70–72 and 30–33 ppm are present, at-tributable to PEG and PE methylene carbons,respectively. Signals arising from carbon nucleiin different chemical environments (for example,those of the urethanic as well as terminalmethyl groups) can be hardly distinguishedfrom the baseline noise. By applying a suitabledeconvolution procedure, it has been possible toquantitatively analyze both the PEG and PEsignals. For both samples, the PEG signalresulted to be constituted by a narrower peak(linewidth of about 120 Hz) resonating at 71.0ppm, and a broader one (linewidth of about 520Hz) resonating at 71.6 ppm; also the PE signalwas actually the superposition of two peaks, onewith a chemical shift of 32.7–32.9 ppm and alinewidth of about 290 Hz, the other with achemical shift of 30.5–30.7 ppm and a linewidthof about 230 Hz. According to literature datathese peaks arise from either ordered (‘‘crys-talline’’) or disordered (‘‘amorphous’’) conforma-

tions for both PEG and PE segments. In particu-lar, the broad signal at 71.6 ppm and the narrowone at 71.0 ppm can be attributed to the ‘‘crys-talline’’ (ordered chain extended conformation)and ‘‘amorphous’’ PEG segments, respec-tively,44,45 whereas the signals at 32.7–32.9 and30.5–30.7 ppm are ascribed to ‘‘crystalline’’ (all-trans conformation) and ‘‘amorphous’’ PE seg-ments, respectively.46,47 The percentages of the‘‘crystalline’’ and ‘‘amorphous’’ fractions of thePE and PEG blocks in both samples could bedetermined from the spectral deconvolutions. Inthe unwashed E sample the crystalline fractionresulted to be about 48% (46 and 52% for PEGand PE, respectively), while it increased up toabout 62% after the THF washings (61 and 65%for PEG and PE, respectively). This increase ofcrystallinity suggests that, as it is reasonable toexpect, the gel-like phase, which is mostlyremoved by the THF washings, has an amor-phous character. By assuming that the gel-likephase is completely amorphous, from the com-parison between the crystalline and amorphouspercentages before and after the THF treat-ment, it is possible to estimate a 22–24% weightloss of the organic fraction. Considering the ini-tial 50/50 organic–inorganic weight ratio, thiscorresponds to a final ratio of 38–39/62–61, inexcellent agreement with the TGA results.Moreover it is interesting to compare the per-centage of copolymer in ordered ‘‘crystalline’’conformation as obtained from NMR, about 63%,with the fraction of copolymer constituting thenanoparticles shell, roughly estimated fromTEM images reported in Figure 2(b,c) to be 60–70%. The similarity between these two resultswould support the above-reported hypothesisthat the nanoparticles shell is mostly consti-tuted by a monolayer of PE–PEG–Si copolymerarranged in the ordered all-trans conformation,even if the presence of a small ‘‘amorphous’’ co-polymer fraction cannot be ruled out. On thecontrary both the gel-like phase and the copoly-mer possibly present in the hybrid core wouldshow a disordered ‘‘amorphous’’ character.According to this picture, the prepared hybridparticles consist of a silica-rich core with amonolayer PEG–PE amphiphilic shell; thetapered shell composition, ranging from highlyhydrophilic close to the silica surface to com-pletely lipophilic in the very outer surface,seems particularly interesting for applications inmedicine as drug carriers. Figure 7 reports DSCcurves of: (a) PE–PEG–OH, (b) PE–PEG–Si,

Figure 6. Quantitative 13C DE/MAS spectra of Sam-ple E before (upper trace) and after (lower trace) THFwashings, acquired at a MAS frequency of 6 kHz.

CORE-SHELL ORGANIC–INORGANIC HYBRID NANOPARTICLES 1705

Journal of Polymer Science: Part A: Polymer ChemistryDOI 10.1002/pola

(c) the crude hybrid product resulting from sol–gel process, and (d) the same product recoveredafter THF treatment. As it appears both thePEG and the PE segments are able to crystallizein PE–PEG–OH copolymer [Fig. 7(a)], whileonly the PEG segments are able to self organizein crystalline domains in PE–PEG–Si, eventhough to a minor extent [Fig. 7(b)]. Evidence ofa limited PEG crystallization is still present inthe 50:50 nominal hybrid washed sample, whileno evidence of crystalline domains can beobserved before THF extraction. The apparentdiscrepancy with the NMR data, which indi-cated that about a half of copolymer in the or-dered ‘‘crystalline’’ conformation is present alsobefore the THF treatment, can be explained tak-ing into account the different sensitivity of NMRand DSC to the size of the crystalline domains:while the 13C chemical shift and/or linewidthare sensitive to the molecular conformation,that is, to nanocrystalline domains, much largercrystalline domains must be present to give calo-rimetric effects. Nevertheless the increase ofcrystallinity after the THF washings as detectedby DSC is in agreement with the increased frac-tion of copolymer chains in the ordered all-transconformation as observed from 13C DE/MASspectra.

Investigation of Silicon ConnectivityWithin Particles

Further information about the molecular struc-ture of the nanoparticles was obtained by 29Sisolid-state NMR. In Figure 8 the 29Si CP/MAS

and SPE/MAS spectra of Sample E after THFtreatment are reported. It can be useful toremind that silicon nuclei forming four andthree Si��O bonds are conventionally indicatedas Qn and Tn respectively, where n is the num-ber of oxygen atoms further bonded to other sili-con atoms. In both spectra intense signals aris-ing from Q4 (� �112 ppm), Q3 (� �101 ppm),and Q2 (� �92 ppm) silicon nuclei derived fromTEOS are observed; moreover in the region from�60 to �70 ppm a broad signal ascribable to T3

and T2 silicon nuclei belonging to the PE–PEG–Si copolymer is present. It is worth to noticethat neither T0 and Q0, nor T1 and Q1 signalsare observed, indicating the absence of bothunreacted and ‘‘monofunctionalized’’ TEOS andPE–PEG–Si copolymer.

The 29Si CP/MAS and SPE/MAS spectra areclearly different: this is easily explained by con-sidering that the CP process strongly enhancesthe signals of the silicon nuclei spatially close toprotons,48 the most evident effects concerning,in this case, the Q3 and T2 � T3 signals, the lat-ter being almost undetectable in the SPE/MASspectrum. If on the one hand the CP/MAS spec-trum gives a relatively fast qualitative response,on the other hand quantitative information canbe extracted only from the SPE/MAS spectrum.By applying a suitable deconvolution to theSPE/MAS spectra of Sample E before (spectrumnot reported) and after THF washings, it hasbeen possible to obtain the relative amounts ofQ4, Q3, and Q2 silicon nuclei (from TEOS), aswell as its ‘‘condensation degree’’ (TEOS c.d.%),calculated as the percentage of condensed

Figure 7. DSC curves of PE–PEG–OH copolymer(a), of functionalized PE–PEG–Si copolymer (b), andof THF extracted (c) and unextracted (d) Sample Ecore-shell nanoparticles.

Figure 8. 29Si CP/MAS (upper trace) and quantita-tive SPE/MAS spectra of Sample E after THF wash-ings, acquired at a MAS frequency of 5.5 kHz.

1706 MARINI ET AL.

Journal of Polymer Science: Part A: Polymer ChemistryDOI 10.1002/pola

Si��O��Si groups with respect to the total num-ber of Si��O groups:

TEOS c:d:% ¼ 4Q4 þ 3Q3 þ 2Q2

4ðQ4 þQ3 þQ2Þ 3 100 ð1Þ

All the results are reported in Table 2. It isworth to notice that, as expected, THF wash-ings, extracting the unreacted organic fractionof the sample, do not significantly influence nei-ther the distribution of the different Qn siliconsites of the mostly inorganic core, nor the TEOSdegree of condensation, which is about 90% forboth the washed and unwashed samples.

Effect of Some Variables on the Particle Formation

It is also interesting to consider the effect of dif-ferent reaction conditions on the formation ofparticles and on their dimensions. As it isreported in Table 1, no particles were obtainedwhen the organic–inorganic ratios were equal orhigher than 80:20. Analogously overall gelationrather than spherical core-shell particles wasobtained when HCl was used as catalyst(Sample F). The average particle diameter de-creases by increasing the organic/inorganic ratioand by increasing the reaction temperature,whereas particle diameters are scarcely affectedby the initial reactant concentration. It is alsointeresting to note that the shell thickness issubstantially unaffected by these parameters, afurther suggestion that in all cases the shellconsists of a PE–PEG monolayer.

Comments About the Mechanism ofParticle Formation

The mechanism of formation of nanoparticles bysol–gel processes has beenwidely investigated, anda clear picture of the structural evolution under dif-ferent reaction conditions has been reported.37,42

In particular, a pertinent particles-growth modelhas been proposed to explain the formation of nano-particles by the sol–gel process of metal alkoxidesalone. It assumes that there is a random distribu-tion of reactants in the initial stage of the processand that micelles, clusters, and then particles areformed as the reactions proceed.

Computer simulation has clearly shown thatrandom monomers-cluster growth results insmooth uniform particles when condensation israte limiting.43 The most accepted model todescribe particle formation is a mechanism ofnucleation and growth; nuclei are formed at acritical degree of super-saturation, and subse-quently grow by the addition of low-molecularweight products (probably mainly monomers). Itis also assumed that the monomers, which areinitially present at high concentration in the so-lution, are continuously formed by depolymer-ization of oligomers in the latest stages of theprocess. Because of the different relative ratesof hydrolysis, condensation, and depolymeriza-tion in different pH ranges, and to the require-ments of high water concentration to have asuitable depolymerization rate, it is reportedthat the particles formation is favored at highpH and in the presence of relevant amount ofwater (H2O/Si molar ratio 7–25).42

Also in this case, we can assume that a ran-dom distribution of the reactants (including tri-alkoxysilane-terminated polymers) is present inthe initial solution, and that the reactantsorganize themselves into micelles, clusters, par-ticles, or networks as the reaction proceeds,depending on the relative rates of different reac-tions/phenomena, such as hydrolysis, condensa-tion, depolymerization, and phase separation.

Whereas for metal alkoxides alone the par-ticles are electrostatically stabilized by the pres-ence of repulsive charges on their surfaces thatprevent particle coalescence (formation of an‘‘electrical double layer’’ around the nano-sphere), in the presence of reactive polymers,such as PE–PEG–Si, that can behave like sur-factants, a significant contribution to the parti-cle formation/stabilization can derive also by asurface energy balance. Contributions from the

Table 2. Chemical Shifts and Areas of the Qn SiliconSignals (from TEOS) in Sample E Before and AfterTHF Washings, as Obtained from Suitable SpectralDeconvolution of Quantitative 29Si SPE/MAS Spectra

Sample Silicon SitesChemical

Shift d (ppm) Areas (%)

E before THFwashings

Q4 �112.3 67.1Q3 �102.1 28.9Q2 �93.1 4.0

TEOS c.d.% 90.8E after THFwashings

Q4 �112.2 65.9Q3 �101.3 29.3Q2 �92.1 4.8

TEOS c.d.% 90.3

The TEOS condensation degree (TEOS c.d.%), calculatedfollowing eq 1, is also reported for the two samples.

CORE-SHELL ORGANIC–INORGANIC HYBRID NANOPARTICLES 1707

Journal of Polymer Science: Part A: Polymer ChemistryDOI 10.1002/pola

rate of diffusion of the different species (reac-tants and clusters probably move with Browniantrajectories but with a velocity inversely propor-tional to the square root of their mass), andprobably a preferential segregation of the cata-lyst close to the growing particles have to beconsidered to explain why mainly nanoparticlesare formed at H2O/Si molar ratio ¼ 4 (wellbelow the molar ratio typically used for all-inor-ganic particles), and why most of the PE–PEG–Si is found on the surface of the nanoparticlesrather than in the core particle.

Tentatively we can assume that TEOS andthe related products (monomers and oligomerseither hydrolized or not) react faster than PE–PEG–Si (either with alkoxy or silanol groups).Therefore, in the first part of the process noPE–PEG–Si (or only a very limited part of it)has reacted, while silica nanoparticles, electro-statically stabilized against coalescence underbasic conditions, are still growing.

The reason why the PE–PEG–Si copolymer isless reactive may be ascribed to the fact that theSi��OR/Si��OH groups, located at the end of thecopolymer chains, are sterically hindered by therelatively long organic chain and/or because ofinductive effects.37 A lower translational mobilityof polymer chains with respect to low-molecularweight silanes/silanols may further contribute toreduce the reactivity of the polymer component.As a consequence of this lower reactivity only asmall fraction of copolymer molecules will havereacted before most of the TEOS derivatives havereacted to build up silica nanoparticles. ThereforePE–PEG–Si will react preferentially with thesurface of the particles only in the last part of theprocess, when most of the initial alkoxides havealready been included in the particles. Whenbonded to the particle surfaces, the copolymer willcontribute to stabilize the particles against fur-ther growth, coalescence, and gelation.

The higher particle diameter observed at loworganic–inorganic ratio [Fig. 2(b) and Sample Bin Table 1] seems a support to this picture. Onthe contrary, an increase of temperature, whichis expected to increase the polymer chain mobil-ity relatively much more than for the low-molec-ular weight reactants, should lead to a decreaseof the particle diameter, as actually it was foundfor Sample E in comparison with Sample C [Fig.2(c) and Table 1].

The concentration of reagents in the solventsseems to have no influence, as it is also reportedin Table 1 for Samples C and D. Finally, under

acidic conditions, a continuous network, ratherthan particles, is formed (see Sample F), asexpected from the literature.41–43

CONCLUSIONS

This work demonstrated that core-shell nano-particles with an organic PE–PEG amphiphilictapered shell and an inorganic-rich core can beprepared in a high yield by a simple and not ex-pensive method using a novel one-pot synthesisbased on sol–gel chemistry. TEOS and a a-tri-ethoxysilane-terminated polyethylene-b-poly(eth-ylene glycol) were used as starting reactantsalong with THF as solvent, NaOH as catalyst,and a H2O/Si molar ratio of 4, well below thattypically used to prepare all inorganic particles.The nanoparticles resulting from this processwere investigated by TEM microscopy (combinedwith EDS analysis), 13C and 29Si solid-stateNMR, TGA, and DSC. The nanoparticles arecharacterized by an inorganic-rich core with a90% TEOS condensation degree, with a diame-ter depending on the initial organic–inorganicratio and reaction conditions, and a shell, with athickness of about 20–30 nm, mostly formed bya monolayer of PE–PEG–Si chains in a ‘‘crys-talline’’, fully extended conformation. Moreover, agel-like phase, constituted by amorphous organicphase, is present among the nanoparticles, whichis partially removed bywarm THFwashings.

The influence of organic–inorganic ratio, solu-tion concentration, temperature, and kind of cat-alyst on the particles dimensions were investi-gated. Particles were obtained only under basiccatalysis and at suitable solvent concentrationand organic/inorganic ratios. The results suggestthat many phenomena occur at the same timeduring the process in a complex way, and theaverage diameter of the particles resulting fromthe sol–gel process is expected to depend on theorganic/inorganic ratio and on the relative rateof phenomena, such as hydrolysis and condensa-tion reactions and diffusion of reactants withinthe solution, and on thermodynamic features suchas miscibility of reactants in the medium and sur-factant properties of the organic polymer chains.

This work is partially supported by PRRIITT (RegioneEmilia Romagna), Net-Lab ‘‘Surface & Coatings forAdvanced Mechanics and Nanomechanics’’ (SUP&RMAN). The authors acknowledge the contribution of

1708 MARINI ET AL.

Journal of Polymer Science: Part A: Polymer ChemistryDOI 10.1002/pola

M. Zapparoli at the CIGS Centre of the University ofModena and Reggio Emilia in producing the TEMmicrographs reported in the text.

REFERENCES AND NOTES

1. Kichelbick, G.; Liz-Marzan, L. M. In Encyclopediaof Nanoscience and Nanotechnology; Nalwa, H. S.,Ed.; American Scientific Publishers: Los Angeles,2004; Vol. 2, p 199.

2. Zhong, C. J.; Maye, M. M. Adv Mater 2001, 13,1507.

3. Dias, M. L.; Pedroso, M. A. S.; Mothe, C. G.;Azuma, C. J Metastable Nanocryst Mater 2004,22, 83.

4. Pedroso, M. A. S.; Dias, M. L.; Azuma, C.; SanGil, R. M.; Mothe, C. G. Colloid Polym Sci 2003,281, 19.

5. Burgeat-Lami, E. J Nanosci Nanotechnol 2001,2, 1.

6. Kichelbick, G. Prog Polym Sci 2003, 28, 83.7. Arriagada, F. J.; Osseo-Assare, K. Colloids Surf

1992, 69, 105.8. Arriagada, F. J.; Osseo-Assare, K. J Colloid Inter-

face Sci 1995, 170, 8.9. Esquena, J.; Tadros, T. F.; Kostarelos, K.; Solans,

C. Langmuir 1997, 13, 6400.10. Arriagada, F. J.; Osseo-Assare, K. J Colloid Inter-

face Sci 1999, 211, 210.11. Chabra, V.; Pillai, V.; Mishra, B. K.; Morrone, A.;

Shah, D. O. Langmuir 1995, 11, 3307.12. Li, G. L.; Wang, G. H. Nanostruct Mater 1999,

11, 663.13. Wu, M.; Long, J.; Huang, A.; Luo, Y.; Feng, S.;

Xu, R. Langmuir 1999, 15, 8822.14. Schmidt, H.; Kropf, C.; Schiestel, T.; Schirra, H.;

Sepeur, S.; Lesniak, C. Ceram Trans 1998, 95, 49.15. Van Bladereen, A.; Vrij, A. J Colloid Interface Sci

1993, 156, 1.16. Duchet, J.; Chabert, B.; Chapel, J. P.; Gerard, J.

F.; Chovelon, J. M.; Renault, N. J. Langmuir1997, 13, 2271.

17. Boven, G.; Oosterling, M. L. C. M.; Chella, G.;Schouten, A. J. Polymer 1990, 31, 2377.

18. Prucker, O.; Ruhe, J. Macromolecules 1998, 31,602.

19. Bottcher, H.; Hallensleben, M. L.; Nuss, S.;Wurm, H. Polym Bull 2000, 44, 223.

20. Mori, H.; Seng, D. C.; Zhang, M.; Muller, A. H. E.Langmuir 2002, 18, 3682.

21. Yoshinaga, K.; Tani, Y.; Tanaka, Y. Colloid PolymSci 2002, 280, 85.

22. Vincent, B. Chem Eng Sci 1993, 48, 249.23. Yoshinaga, K. Surf Sci Ser (Fine Part) 2000, 92,

626.

24. Hergeth, W. D.; Starre, P.; Schmutzler, K. Poly-mer 1988, 29, 1323.

25. Hergeth, W. D.; Steinau, U. J.; Bittrich, H. J.;Simon, G.; Schmutzler, K. Polymer 1989, 30, 254.

26. Scott, C.; Ishida, H.; Maurer, F. H. J. J Mater Sci1987, 22, 3963.

27. Bayer, T.; Eichorn, K. J.; Grundke, K.; Jacobasch,H. J. Macromol Chem Phys 1999, 200, 852.

28. Monteil, V.; Stumbaum, J.; Thomann, R.; Meck-ing, S. Macromolecules 2006, 39, 2056.

29. Mai, K.; Mader, E.; Myhle, M. Compos: Part A1998, 29, 1111.

30. See, K. H.; Mullins, M. E.; Mills, O. P.; Heiden, P.A. Nanotechnology 2005, 16, 1950.

31. Arkhireeva, A.; Hay, J. N.; Lane, J. M.; Manzano,M.; Masters, H.; Oware, W.; Shaw, S.J. J Sol–GelSci Technol 2004, 31, 31.

32. Epifani, M.; Leo, G.; Lomascolo, M.; Vasanelli, L.;Manna, L. J Sol–Gel Sci Technol 2003, 26, 441.

33. Fei, B.; Lu, H.; Xin, J. H. Polymer 2006, 47, 947.34. Brinker, C. J.; Scherer, G. W. In Sol–Gel Science:

The Physics and Chemistry of Sol–Gel Processing;Klein, L., Ed.; Academic Press: San Diego, 1990;p 97.

35. Toselli, M.; Marini, M.; Fabbri, P.; Messori, M.;Pilati, F. J Sol–Gel Sci Technol 2007, 43, 73.

36. Geppi, M.; Mollica, G.; Borsacchi, S.; Marini, M.;Toselli, M.; Pilati, F. J Mat Res 2007, 22, 3516.

37. Brinker, C. J. J Non-Cryst Solids 1988, 100, 31.38. Cory, D. G.; Ritchey, W. M. J Magn Reson 1988,

80, 128.39. Tian, D.; Dubois, P.; Jerome, R. J Polym Sci Part

A: Polym Chem 1997, 35, 2295.40. Brinker, C. J.; Keefer, K. D.; Schaefer, D. W.; Ash-

ley, C. S. J Non-Cryst Solids 1982, 48, 47.41. Brinker, C. J.; Keefer, K. D.; Schaefer, D. W.;

Assink, R. A.; Kay, B. D.; Ashley, C. S. J Non-Cryst Solids 1984, 63, 45.

42. Brinker, C. J.; Tallant, D. R.; Roth, E. P.; Ashley,C. S. J Non-Cryst Solids 1986, 82, 117.

43. Keefer, K. D. In Better Ceramics Through Chem-istry, II; Brinker, C. J.; Clark, D. E.; Ulrich, D.R., Eds.; Proceedings of the Materials ResearchSociety: Pittsburg, 1986; Vol. 73, p 295.

44. Bartolotta, A.; Forte, C.; Geppi, M.; Minniti, D.;Visalli, G. Solid State Nucl Magn Reson 1997, 8,231.

45. Dechter, J. J. J Polym Sci Polym Lett Ed 1985,23, 261.

46. McBrierty, V.; Paker, K. J. Nuclear Magnetic Res-onanance in Solid Polymers; Cambridge Univer-sity Press: Cambridge, 1993.

47. Borsacchi, S.; Geppi, M.; Ricci, L.; Ruggeri, G.;Veracini, C. A. Langmuir 2007, 23, 3953.

48. Borsacchi, S.; Geppi, M.; Veracini, C. A.; Fallani,F.; Ricci, L.; Ruggeri, G. J Mater Chem 2006, 16,4581.

CORE-SHELL ORGANIC–INORGANIC HYBRID NANOPARTICLES 1709

Journal of Polymer Science: Part A: Polymer ChemistryDOI 10.1002/pola