at SciVerse ScienceDirect

Polymer 54 (2013) 1388e1396

Contents lists available

Polymer

journal homepage: www.elsevier .com/locate/polymer

Hybrid Janus particles based on polymer-modified kaolinite

Stephan Weiss a,2, Dunja Hirsemann b,2, Bernhard Biersack c, Mazen Ziadeh b, Axel H.E. Müller a,*,1,Josef Breu b,*

aMacromolecular Chemistry II, University of Bayreuth, 95440 Bayreuth, Germanyb Inorganic Chemistry I, University of Bayreuth, 95440 Bayreuth, GermanycOrganic Chemistry I, University of Bayreuth, 95440 Bayreuth, Germany

a r t i c l e i n f o

Article history:Received 21 September 2012Received in revised form11 December 2012Accepted 17 December 2012Available online 3 January 2013

Keywords:Polymereclay hybridsPolymer blendJanus particles

* Corresponding authors. Tel.: þ49 921 552530; faxE-mail addresses: axel.mueller@uni-mainz.de (A.

bayreuth.de (J. Breu).1 New address: Institute of Organic Chemistry, Joh

55099 Mainz, Germany.2 These authors contributed equally to this project

0032-3861/$ e see front matter � 2012 Elsevier Ltd.http://dx.doi.org/10.1016/j.polymer.2012.12.041

a b s t r a c t

Janus particles are superior to ordinary Pickering emulsifiers. Widespread industrial usage is howeverhampered by the restricted accessibility of these polar colloids. Whereas established synthesis protocolsthat allow for breaking the symmetry of colloid surfaces are laborious and expensive, here we use Janusparticles that are based on the ubiquitous layered silicate kaolinite. As a consequence of its crystalstructure, the two opposing basal planes, the tetrahedral (TS) and the octahedral (OS) surface, are cappedby distinct functional groups. We show that these chemically different basal surfaces allow for facile se-lective modification making them compatible with e.g. polystyrene (PS) and poly(methyl methacrylate)(PMMA), respectively. The TS was selectively modified by simple cation exchange with poly(2-(dime-thylamino)ethyl methacrylate) polycations attached to a polystyrene block while on the OS PMMA chainswere covalently anchored via statistically distributed catechol groups. Solid-state NMR proved successfulmodification, while TGA was provided the amount of polymer bound to both external basal surfaces ofkaolinite. Moreover, the selective nature of the modification was proven indirectly by qualitatively com-paring the sedimentation stability of suspensions of differently modified samples. These Janus particleswith fine-tuned surface tension of the opposing hemispheres were tested on their interfacial activity insolvent-cast films of PSePMMA blends. The assembly of the Janus particles at the interfaces of the blendwas proven by transmission electron microscopy.

� 2012 Elsevier Ltd. All rights reserved.

1. Introduction

In the last decades polymer blends have become technicallyimportant in the field of high performance polymers. Tailoringproperties by blending homopolymers is simply economicallymore attractive than the synthesis of an appropriate new mono-mer/homopolymer [1]. Heterogeneous blends often combine themechanical properties of homopolymers like the storage and lossmoduli in an additive way and in some cases even in a synergisticway [2,3]. In any case, the mechanical properties critically dependon the microstructure of the blends and control of the metastablemorphologies is therefore highly desirable [4,5]. In contrast, thehigh surface tension at the polymerepolymer interface isdetrimental for other properties like strength and toughness [6].

: þ49 921 552788.H.E. Müller), josef.breu@uni-

annes Gutenberg University,

.

All rights reserved.

These negative effects can at least partially be remedied by com-patibilization [6,7].

Prominent examples of compatibilizers are random copolymers[36], block co- or terpolymers [8,35] while solid particles have beenexplored as cheap alternatives [10,11]. Much similar to amphiphilicmolecules and block copolymers, Pickering particles with surfacetensions intermediate between the two immiscible polymer phasesare arranged at the interface [12,13]. Janus particles, colloids withchemically distinct hemispheres, combine the best of the twoworlds of compatibilizers e amphiphilicity and Pickering effect[9,14]. A perfect compatibility with both polymer phases can beachieved by tailoring the Janus character of the particles. Ideally,Janus particles whose hemispheres are comprised of the twopolymers of the blend, should achieve a particleepolymer surfacetension of zero on both sides of the blend interface. A proof ofconcept of such tailor-made Janus particles in a polymer blend waspreviously delivered using soft Janus particles made from ABC tri-block terpolymers [15].

2:1-Layered silicates, like laponite or montmorillonite, havebeen used for some time as fillers and compatibilizers [10,16e21].

HO OH

O

O

O O

* 3

O

N

O20

O O

*

90

stat

block

*

115

H

(H2C)3

Fig. 2. Structure of poly(3-(2,3-dihydroxybenzoyloxy)propylmethacrylate)-stat-(methylmethacrylate) (PCM) (top) and poly(2-(dimethylamino)ethyl methacrylate)-block-poly-styrene (PDPS) cations (bottom).

S. Weiss et al. / Polymer 54 (2013) 1388e1396 1389

Next to using the inherent Pickering effect, clay has been applied toimprove several material properties of the polymeric matrix, likehigh moduli, increased strength, heat resistance, decreased gaspermeability and flammability [37e40]. In most of the studies,commercially available organophilized natural montmorilloniteshave been used [16,17,20,22]. The structure of these commercialbentonites is symmetrical and displays no Janus character. Thusthe emulsification is purely of Pickering character. Furthermore, thesimple modification of the clay with alkyl ammonium cations isinsufficient to tune the surface tension to fall between the inter-facial tensions of the two polymer phases. Consequently, the Pick-ering emulsification is far from optimum.

Whereas established synthesis protocols that allow for break-ing the symmetry of colloid surfaces are laborious and expensive[23e25], in this paper we explore the emulsification efficiency oflayered silicates that are of genuine Janus character. These Janusparticles are easily manufactured from a cheap, non-symmetric,natural 1:1-layered silicate, (Al2Si2O5(OH)4) kaolinite from anAmazonas deposit.

Amazone kaolinite is a fine-grained material with a com-paratively small aspect ratio in the range of 15 and a lateral dimen-sion smaller than 2 mm (Fig. 1, left) [26]. The dioctahedral layeredsilicate consists of polar lamellae which are stacked and heldtogether by hydrogen bonds (Fig. 1, centre). Since, usually no twin-ning has been observed [27], the two opposing external basal planesof the kaolinite platelets consist of an AlO2(OH)4 octahedral layer(octahedral surface,OS),which is cappedbym-hydroxyl groups at theexternal surface, anda SiO4 tetrahedral layer (tetrahedral surface, TS)(Fig. 1, right). Natural kaolinite, though formally neutral in charge,possesses a small cation exchange capacity (CEC) due to an iso-morphous substitution in the tetrahedral layers, which can only becounterbalanced at the outer TS by hydrated cations [28] that renderthe surface hydrophilic. Although even pristine kaolinite displaysa Janus character, the surface tension of the twounmodified externalbasal surfaces is similar. However, these different chemical basalplanes can be addressed selectively as published previously [29].

As sketched in Fig. 1, due to its polar crystal structure, the twoopposing external basal planes of kaolinite, TS and OS are trun-cated by distinct functional groups and may selectively be modi-fied by simple cation exchange and covalent grafting via catecholgroups [41], respectively. As an example we chose poly((2-dimethylamino)ethyl methacrylate)-block-polystyrene (PDPS) andpoly(3-(2,3-dihydroxybenzoyloxy)propyl methacrylate)-stat-(methyl methacrylate)) (PCM).

PCM is a statistical copolymer, while PDPS is a block copolymer(Fig. 2). Consequently, both modifiers interact with the kaolinitebasal planes in a different manner (Fig. 3). PCM will likely be closeto the OS, forming short loops or lying flat. In contrast, the poly-styrene block might arrange brush-like on the TS of the kaolinite.

The resulting Janus particles are tailored for compatibilizing PSePMMA or industrially more relevant PPEeSAN (poly(2,6-dimethyl-

Fig. 1. SEM top view image of a typical kaolinite platelet (left), schematic picture of kaolinchemical functions at the basal TS and OS (right).

1,4-phenylene ether) (PPE), poly(styrene-co-acrylonitrile) (SAN))blends [11,18]. TEM images of solvent-cast films of PSePMMAblends were used as a proof of concept for the fine-tunedinterfacial tension of the opposing hemispheres of the hybrid Ja-nus particles.

2. Experimental

2.1. Materials

Kaolinite (Amazone 88/90) from Brazil was provided by Vale In-ternational S.A. (Saint-Prex, Switzerland). The mineral was size frac-tioned by a hydrocyclone but no dispersing agent and nosedimentation agent were added. The kaolinite was further purifiedby removal of calcium- and magnesium-carbonates with ethyl-enediaminetetraacetic acid (EDTA), followed by deferrationvia the dithioniteecitrateebicarbonate (DCB)-method. Moreover

ite platelets (centre) and crystal structure of three kaolinite lamellae with the specific

Fig. 3. Schematic picture of a) pristine kaolinite, b) modified with PDPS on the tetrahedral surface (TS), c) further modified with PCM on the opposite octahedral surface (OS) and d)embedding of the final hybrid particle at the interface in a PSePMMA blend.

S. Weiss et al. / Polymer 54 (2013) 1388e13961390

ozonization was applied for 2 h to remove organic impurities.The particle size of the material was fractionated to <2 mm by theAtterberg procedure to remove traces of agglomerates. This materialwas used in characterization experiments. Size fractionatedkaolinite < 500 nmwas used in blending experiments. Purity of thekaolinite was confirmed applying powder X-ray diffraction (PXRD),solid-state nuclearmagnetic resonance (NMR) spectroscopy, infrared(IR) spectroscopy, and energy dispersive X-ray spectroscopy (EDX).

PMMA (Mw¼ 120,000 g/mol) and PS (Mw¼ 192,000 g/mol)werepurchased from Sigma Aldrich and used without any furthertreatment. THF was dried by distillation over Na, and DCM wasdried by distillation over CaH2 under an argon atmosphere. Solventsof analytical grade and the starting compounds were purchasedfrom the usual sources and were used without further purification.

2.2. Characterization methods

IR spectra were recorded on a PerkineElmer One FT-IR spec-trophotometer. Liquid NMR spectra were recorded under ambientconditions on a Bruker Avance 300 spectrometer. Chemical shifts (d)are given in parts per million downfield from TMS as internal stan-dard. The 13C solid-state NMR measurements were performed ona Bruker Avance 2 spectrometer operating at 7.05 T with a resonancefrequency n0 of 75.468 MHz under magic angle spinning condition(nrot¼ 10 kHz) via cross polarization. The 13C spectrawere referencedrelative toTMS.Mass spectrawere recorded using a VarianMAT 311A(EI). The polymer/hybrid thin films were cut with a Leica EM UC7microtome and transmission electron microscope (TEM) images wererecorded with a Zeiss EM 922 Omega microscope at 200 kV. For col-umn chromatographyMerck silica gel 60 (230e400 mesh) was used.

The molecular weights and molecular weight distribution of thepolymers were measured by gel permeation chromatography (GPC).

OBn

OBn

CO2H

OBn

OBn

O O

1 2

(i)

Scheme 1. Synthesis of the catechol monomer. Reagents and conditions: (i) SOCl2, CH2(C

Column set: 5 mm SDV gel, 102, 103, 104, and 105�A, 30 cm each (PSS,Mainz). Used detectors are refractive index (RI) and ultraviolet (UV)operated at 254 nm. Polystyrene standards (PSS, Mainz) with nar-row molecular weight distribution were used for calibration of thecolumn set, and THF and DMAc were used as eluents at a flow rateof 1 ml/min.

Thermogravimetric analysis (TGA) was carried out using a Met-tler Toledo TGA/SDTA 85 at a heating rate of 5 K/min between 30and 700 �C under a nitrogen-flow of 60 ml/min. The typical sampleweight was between 8 and 15 mg.

The stability measurements were performed in a LUMiFuge� 114(LUM) with a variable rotation frequency of 300, 600, 900 rpm(rounds per minute) and different time intervals of 200 s, 300 s,and 900 s, respectively. Kaolinite suspensions (0.25 wt%) in THF andwater were placed in tubes in horizontal positions on the disc of theLUMiFuge�. During the horizontal rotation of this disc the trans-parencies of the suspensions were measured in the area betweenthe menisci and the sediment. The mean transparency of the wholearea was determined. The transparency was measured in time in-tervals of 10 s while increasing rotation speed stepwise. High tur-bidity, even after applying centrifugal forces indicates a stablesuspension.

2.3. Synthesis of the catechol-modified poly(methyl methacrylate)copolymer (PCM)

Compounds 1 and 2 were prepared following literature pro-cedures (Scheme 1) [30].

2.3.1. 3-Methacryloyloxypropyl-3,4-dibenzoxybenzoate 3Compound 2 (2.11 g, 5.38 mmol) was dissolved in dry DCM

(20 ml) and cooled in an ice bath. Et3N (1.12 ml, 8.07 mmol) and

OH

OBn

OBn

O O O O

3

(ii)

H2OH)2, Et3N, THF/DCM, r.t., 5 h, 51%; (ii) CH2C(CH3)COCl, Et3N, DCM, r.t., 3 h, 67%.

S. Weiss et al. / Polymer 54 (2013) 1388e1396 1391

methacryloyl chloride (626 ml, 6.47 mmol) were added and thereaction mixture was stirred at room temperature for 3 h. Afterwashing with water the aqueous phase was extracted with DCMand the combined organic phases were dried over Na2SO4, filteredand concentrated in vacuum. The residue was purified by columnchromatography (silica gel 60, ethyl acetate/n-hexane 1:2, v/v).Yield: 1.64 g (3.57 mmol, 67%); colourless oil; Rf ¼ 0.63 (ethylacetate/n-hexane 1:2); nmax (ATR)/cm�1: 3032, 2963, 1711, 1636,1599,1510,1454,1427,1380,1321,1266,1204,1163,1130,1104,1038,1006, 944, 815, 761, 734, 695; 1H NMR (300 MHz, CDCl3): d 1.92(3 H, s), 2.0e2.2 (2H, m), 4.28 (2H, t, 3J 6.3 Hz), 4.36 (2H, t, 3J 6.3 Hz),5.18 (2H, s), 5.21 (2H, s), 5.5e5.6 (1H, m), 6.0e6.1 (1H, m), 6.91 (1H,d, 3J 9.0 Hz), 7.3e7.5 (10H, m), 7.6e7.7 (2H, m); 13C NMR (75.5 MHz,CDCl3): d 18.3, 28.2, 61.3, 61.4, 70.8, 71.2, 113.2, 115.6, 123.0, 124.0,125.6, 127.1, 127.4, 127.9, 128.0, 128.5, 128.6, 128.9, 136.2, 136.5,136.8, 148.3, 153.0, 166.1, 167.3; m/z (%) 461 (13) [Mþ], 460 (47)[Mþ], 369 (6), 317 (8), 225 (17), 181 (27), 127 (12), 91 (100).

2.3.2. Copolymer PCBM 4Methyl methacrylate (650 mg, 6.52 mmol), compound 3

(100 mg, 0.22 mmol) and dodecanethiol (26 mg, 0.13 mmol) weredissolved in dry THF (3 ml) under argon atmosphere and AIBN(10 mg) was added to the reaction mixture, which was stirredunder reflux for 5 h. The solution was poured into cyclohexane(100ml) and the appearing colourless precipitatewas collected andprecipitated once more from an acetone/cyclohexane mixture.Yield: 710 mg; colourless solid; nmax (ATR)/cm�1: 2996, 2952, 1722,1601, 1484, 1448, 1432, 1385, 1363, 1268, 1241, 1189, 1144, 989, 965,911, 842, 761, 748, 698; 1H NMR (300 MHz, acetone-d6): d 0.8e1.0(33H, m), 1.8e2.0 (24H, m), 3.61 (30H, s), 4.1e4.2 (2H, m), 4.3e4.4(2H, m), 5.2e5.3 (4H, m), 7.1e7.7 (13H, m).

2.3.3. Copolymer PCM 5Compound 4 (580mg)was dissolved in dioxane/methanol (40ml,

1:1), flushed with argon and 10% Pd/C (80 mg) was added. The argonatmosphere was replaced by hydrogen gas and the reaction mixturewas stirred at room temperature for 5 h. The suspensionwas filteredover celite and the filtrate was concentrated in vacuum. The oilyresidue was triturated with n-hexane and dried in vacuum. Yield:500 mg; off-white solid; nmax (ATR)/cm�1: 3392, 2996, 2950, 1725,1605,1480,1444,1386,1270,1239,1191,1146,1121, 988, 965, 889, 873,842, 765, 750; 1HNMR (300MHz,DMSO-d6): d 0.5e0.9 (33H,m),1.6e2.0 (24H,m), 3.55 (30H, s), 4.0e4.1 (2H,m), 4.2e4.3 (2H,m), 6.81 (1H,d, 3J 7.9 Hz), 7.2e7.4 (2H, m), 9.32 (1H, s), 9.81 (1H, s); 13C NMR(75.5 MHz, DMSO-d6): d 16.1, 18.4, 27.5, 43.9, 51.6, 53.7, 60.8, 115.2,116.3, 120.5, 121.8, 145.0, 150.4, 165.5, 176.2, 176.9, 177.3.

2.3.4. Copolymer PDPS 6DMAEMA (1.4 g, 9.05mmol), the RAFT-agent 2-cyanopropan-2-yl

benzodithioate (100 mg, 0.45 mmol), trioxane (10 mg) as internalstandard and AIBN as initiator were added to benzene (8 ml) andflushed with argon. After 4 h at 70 �C the reaction mixture wastransferred to an already heated and argon flushed solution of sty-rene (18.8 g, 181 mmol) in benzene (30 ml). The polymerizationwasstopped after 2 days. The copolymerwas precipitated in isopropanol,the appearing pink precipitate was collected, dissolved in THF andprecipitated from isopropanol again. Yield: 7 g; pink solid.

2.4. Surface modification

It is possible to modify each side (TS and OS) of the kaolinitespecifically and individually as shown in a previous publication [29]without influencing the other side. The order of modificationchosen, starting with PDPS has a purely practical purpose, as theDMAEMA block of the PDPS is charged at pH 6 and thus kaolinite

can be modified in aqueous suspension where it is dispersed best.After cation exchange the unilaterally modified kaolinite can bedispersed in THF more easily than unmodified kaolinite, as seen inSection 3.2.3 (Stability measurements). PCM is soluble in THF, butnot in water. Nevertheless pristine kaolinite can be modified byPCM as first step as well, but for that kaolinite has to be dispersed inTHF by vigorous stirring first.

2.4.1. Modification of TS100 mg PDPS were dissolved in 30 ml THF. 400 mg of the kao-

linite was suspended in 30 ml of water (pH w 5.5, degree of pro-tonation of the DMAEMA blockw80% [33]). After 20 min of stirringa complete flocculation of the kaolinite was achieved and the sus-pension was washed ten times with THF to remove the excess ofPDPS. The hybrid was dispersed and stored in THF to preventdrying.

2.4.2. Modification of OS100 mg of PCM was dissolved in 20 ml dry THF under argon

atmosphere in a Schlenk flask. 400 mg kaolinite (pristine or alreadyone-sided modified hybrid from 2.4.1) was dispersed in the PCMTHF solution by vigorous stirring over night at 60 �C. After the re-action the kaolinite was washed ten times with THF to remove theexcess of PCM and then dispersed in THF.

2.5. Preparation and characterization of the polymer blend

PS and PMMA in the ratio of 1:2 were dissolved in THF. Thepolymer content of the solutions was 10 wt%. 50 mg ofPDPS/PCM-kaolinite was suspended in 10 ml of the PS/PMMA so-lutions by 15 min of strong shearing (Heidolph Silent Crusher,16.000 rpm) at 30 �C.

A film was cast by letting the solvent evaporate slowly from themixture in a glass vial. The resulting dry film was cut with an ultramicrotome and examined via TEM.

3. Results and discussion

3.1. Synthesis of the copolymers

3.1.1. Synthesis of the catechol-modified poly(methyl methacrylate)copolymer (PCM)

Initially, a suitable catechol-modified methacrylate monomerwas prepared for copolymerization with methyl methacrylate(MMA). 3-Hydroxypropylbenzoate 2 was obtained from3,4-dibenzoxybenzoic acid 1 [30]. Reaction of 2 with methacryloylchloride gave the mixed diester 3 (Scheme 1).

Monomer 3 was copolymerized with a 30-fold excess of MMAby free radical polymerization using AIBN as initiator and dodeca-nethiol as transfer agent to gain control and reduce molecularweight, giving copolymer 4. 1H NMR spectroscopy and integrationof the proton signals of the copolymer 4 confirmed the applied ratioof 1:30 of monomer 3 and MMA in the copolymer. Subsequentcatalytic hydrogenation removed the benzyl protecting groupsfrom the copolymer 4 giving the catechol-modified copolymer 5(Scheme 2). Both FT-IR and 1H NMR spectroscopy gave evidence forthe free catechol moieties of the copolymers 5 and revealed newsignals for the catechol hydroxy groups of the copolymer (5:nmax ¼ 3392 cm�1; dH ¼ 9.32, 9.81). The synthesized copolymer hasa number-average molecular weight Mn ¼ 10,000 g/mol (based onPMMA calibration) with a PDI of 1.7.

3.1.2. Synthesis of the PDMAEMA-b-PS block copolymer (PDPS)The selected block copolymerwas synthesized via sequential RAFT

polymerization, has a number-average molecular weight of

OBn

OBn

O

O

O

3

CO2MeO

MMA

(i)CO2Me

OO

O

O

OR

OR

(ii)4: R = Bn

5: R = H

30+

Scheme 2. Synthesis of PCM. Reagents and conditions: (i) AIBN, THF, reflux, 5 h; (ii) H2, Pd/C (10%), MeOH/dioxane, r.t., 5 h.

S. Weiss et al. / Polymer 54 (2013) 1388e13961392

16,000 g mol�1 and exhibits a narrow molecular weight distributionwith Mw/Mn w 1.13 (determined by DMAc GPC with PS calibration).DMAEMA was polymerized as a first block achieving a conversionnear 99% (as determined by 1H NMR), anMw of 3200 g/mol and a PDIof 1.20. This PDMAEMA was utilized as a macro-RAFT agent withoutfurther purification and for the polymerization of styrene (Fig. 4). Thefinal composition was determined as PDMAEMA20ePS115 by the 1HNMR integral ratios between characteristic signals of the two blocks(PS: 5H, aromatic, d 6.8e7.3 ppm; PDMAEMA: 2H, eCH2e, d 4 ppm).

3.2. Modification of the kaolinite basal planes

It is possible to modify each side (TS and OS) of the kaolinitespecifically and individually [29] without influencing the otherside. The chosen order of modification, starting with PDPS hasa practical purpose, as the DMAEMA block of the PDPS is charged atthe pH of 6 and thus kaolinite can be modified in aqueous sus-pension, where it is dispersed best. After that, the one-sidedlymodified kaolinite can be dispersed in THF more easily than un-modified kaolinite. PCM is soluble in THF, but not in water. Never-theless pristine kaolinite can be modified by PCM as a first step aswell, but for that it has to be dispersed in THF by vigorous stirring.

20 22 24 26 28 30 32 34

elution volume / mL

Fig. 4. SEC-traces of poly(DMAEMA) precursor (continuous curve, right) and PDPSdiblock copolymer (dotted curve, left).

Successful surface modification of kaolinite with both polymericmodifiers was proven by 13C solid-state MAS (magic angle spin-ning) NMR, while TGA (thermogravimetric analysis) was performedto estimate the amount of polymer bound to both external basalsurfaces of kaolinite. Moreover, the selective nature of the mod-ification was confirmed indirectly by qualitatively comparing thesedimentation stability (as determined by LUMiFuge�) of suspen-sions of differently modified samples.

We used a natural fine-grained kaolinite with typical di-mensions of the ideally hexagonal platelets that were <2 mm indiameter and up to 70 nm in height. The specific surface area wasapproximately 4 m2 g�1, and about 80% of this area could beattributed to the external basal surfaces. Nevertheless, the detec-tion of monolayer coverage of the external surfaces required highlysensitive analytical methods, and the proof of the selective mod-ification is inherently difficult, but selectivity of anchoring groupswas already shown in our previous publication [29].

3.2.1. 13C solid-state NMR13C solid-state NMR spectra of kaolinite samples solely modified

at the OS with PCM (Fig. 5a, PCM-kaolinite) or solely modified withPDPS at the TS (Fig. 5b, PDPS-kaolinite) as well as dually modifiedkaolinite (Fig. 5c, PDPS/PCM-kaolinite) were recorded in order todemonstrate the modification. Please note that according to theTGA results the total weight fraction of modifiers is less than 5%.Therefore the noise is high.

The spectrum of the PCM-kaolinite (Fig. 5a) featured a signal at16.5 ppm indicating the presence of a CH3 group of the MMA mon-omer. Moreover, the eOCH3 group and the CH2 polymer backbone ofPCM could be identified at 44.8 ppm and at 51.5 ppm. Additionally, at177.1 ppm a signalwas detectedwhich is caused by the C]O group ofthe MMA ester function. The aromatic ring of the catechol functionwas hardly detectable at 127.1 ppmwith small signal-to-noise ratio.

PDPS-kaolinite (Fig. 5b) featured all the characteristic signals ofthe neat PDPS polymer. At 39.8 ppm and at 44.5 ppm the character-istic CH2 backbone signals and CH2 signals of the DMAEMA wererecorded. At 127.1 ppmand 146.7 ppm the aromatic CH1 groups of thestyrene could be observed while only a very small signal caused byC]O ester function of the DMAEMA block at 177.1 ppmwas detected.

The spectrum of the PDPS/PCM-kaolinite (Fig. 5c) represents anoverlay of the specific signals of both modifiers indicating thata successful modificationwith both surfacemodifiers is feasible. Thesignals at 16.5 ppm, 51.5 ppm and 177.1 ppm are only or mainlycaused by the modificationwith PCMwhile the signals at 39.8 ppm,127.1 ppm and 146.7 ppm could only or mainly be assigned to PDPS.Please note that the low loading required very long measurementtimes (17 h). Therefore, cross-polarization measurement [31] wasperformed and quantification of signals by integration is impossible.

Fig. 5. 13C solid-state MAS NMR spectra of a) PCM-kaolinite b) PDPS-kaolinite and c)PCM/PDPS-kaolinite.

S. Weiss et al. / Polymer 54 (2013) 1388e1396 1393

However, the relative intensities of the different signals of each in-dividual modifier (Fig. 5a, b) do not change (Fig. 5c) indicating thatno significant structural changes of the polymer chains occur uponadsorption of the second modifier. This fact in turn would indicatethat the adsorption of the two modifiers occurs in a spatially seg-regated mode as expected given the Janus character.

In summary, modification with PCM and PDPS can unequivo-cally be proven by 13C solid-state NMR, however, due to the verylow loading neither quantification nor determination of the ratiobetween PDPS and PCM can be achieved.

3.2.2. Thermogravimetric analysis (TGA)Since quantification by NMR was impossible estimating the

amount of polymer adsorbed was attempted by TGA. The TGA ofpristine kaolinite (Fig. 6, pink) features a dehydroxylation to met-akaolinite above 410 �C [32] which is accompanied by amass loss ofapproximately 13.0 wt%. At 410 �C the modifiers PDPS and PCM are,however, already removed as indicated by TGA experiments ofmixtures of polymer with inert quartz. Below 410 �C pristine kao-linite showed only a minute mass loss of about 0.3%, probably dueto physically adsorbed water which is expected to be completely

Fig. 6. TGA of pristine kaolinite (pink), PCM-kaolinite (blue), PDPS-kaolinite (red) and PDPSfigure legend, the reader is referred to the web version of this article.)

removed by the surface modification. Consequently, the completemass loss observed up to a temperature of 410 �C may be attributedto the modifiers.

In the case of the PDPS-kaolinite (Fig. 6, red) a one-stepdecomposition with a mass loss of 2.7 wt% starting at 275 �C wasobserved. In contrast the modifier the PCM-kaolinite (Fig. 6, blue)featured a comparatively slowmass loss of 2.7 wt% starting at about100 �C. The similar overall mass loss for PDPS-kaolinite and PCM-kaolinite is of course just coincidental.

The PDPS/PCM-kaolinite (Fig. 6, black) shows a mass loss of 5%in total, close to the sum of the weight losses of PDPS-kaolinite andPCM-kaolinite, indicating that the adsorption of the two modifiersdoes not influence each other. Moreover, the shape of the TGA-curve of PDPS/PCM-kaolinite (Fig. 6, black) shows the prominentfeatures of both, PDPS-kaolinite and PCM-kaolinite curves.

The experimentally observed loading of PDPS-kaolinite may becompared to the loading expected from the CEC of the kaolinitewhich was around 2.0 mval/100 g. Assuming charge neutrality andgiven that at a pH of w5.5 approximately 80% of the DMAEMAmonomer units are protonated [33], aweight content of the PDPS inthe range of 2.1 wt% of the PDPS-kaolinite is expected, slightly lessthan the experimentally observed mass loss of 2.7 wt%. This dif-ference might be caused either by a lower degree of protonation ofthe adsorbed PDPS.

3.2.3. Stability measurementsTo verify the selective nature of the dual modification of

PDPS/PCM-kaolinite indirectly, the stability of the suspension wascompared with pristine kaolinite (Fig. 7, pink), PCM-kaolinite(Fig. 7, blue), and PDPS-kaolinite (Fig. 7, red). Since the wettabilityof the samples differs significantly, the experiments were per-formed in water and in THF.

Please note that in Fig. 7 dimeric aggregates are shown ratherthan intercalation compounds. Kaolinite does not form inter-calation compounds with polycations, the ion exchange is limitedto the tetrahedral external basal surface.

With pristine kaolinite both, TS and OS are highly hygroscopicand consequently suspensions in water were quite stable (Fig. 7a,pink), while in THF the stability was very low (Fig. 7b, pink).

In contrast, with PDPS-kaolinite (Fig. 7a, red) the TS was ren-dered hydrophobic, while the OS remains hydrophilic. Finally, withPDPS/PCM-kaolinite the dual modification renders both, TS and OShighly hydrophobic (Fig. 7a, black). The aqueous suspensions ofboth, PDPS-kaolinite and PDPS/PCM-kaolinite were very unstable.The transparency values increase rapidly at early stages of themeasurement for both suspensions indicating a high in-compatibility between the solvent and the modified kaoliniteparticles. Apparently, modification of the TS plane makes thewhole

/PCM-modified kaolinite (black). (For interpretation of the references to colour in this

Fig. 7. Integrated transparency of 0.25 wt% suspensions in a) water and b) THF of pristine kaolinite (pink), PCM-kaolinite (blue), PDPS-kaolinite (red) and PDPS/PCM-kaolinite(black) under time dependent centrifugal forces of 300 rpm, 600 rpm, and 900 rpm. (For interpretation of the references to colour in this figure legend, the reader is referredto the web version of this article.)

S. Weiss et al. / Polymer 54 (2013) 1388e13961394

platelet so hydrophobic, that a ranking regarding the stability ofthese two suspensions in water, cannot be made because differ-ences observed with these highly unstable suspensions areinsignificant.

Surprisingly, in case of the OS-modified PCM-kaolinite (Fig. 7a,blue) a high stability of the suspension in water can be achievedwhich is even comparable with the pristine kaolinite (Fig. 7a, pink,blue) despite the hydrophobization of the OS. Contrary to PDPS-kaolinite where the long PS-brushes extend into the solvent, thestatistical copolymer PCM stays comparatively close to the surface.Moreover, the hydrated inorganic cations residing at unmodified TScontribute to an efficient electrostatic stabilization of PCM-kaolinite in water. Alternatively, the stability of PCM-kaolinitemight be explained by the formation of sandwich structures asdepicted in Fig. 7a (blue framed inset). For such sandwich struc-tures only the hydrophilic TS are exposed to the aqueous media.Such polymer-bridged sandwich structures would not be expectedfor PDPS-kaolinite because the long PS-brushes will hamper dimer-formation sterically.

In THF (Fig. 7b) both, the PCM-kaolinite (Fig. 7b, blue) as well asthe PDPS-kaolinite (Fig. 7b, red), showed good stability which inturn is comparable to that of the dually modifiedPDPS/PCM-kaolinite (Fig. 7b, black). This suggests that even theshort PCM loops at the OS were able to assure a good stability inTHF and expectedly the longer chains perform as well. Moreover, itwould be expected that sandwich structures of PDPS- andPCM-kaolinite are formed (Fig. 7b, red and blue squares).

In summary, the stabilities in water- and THF-suspensionsobserved for the different kaolinite samples are in line with aspecific modification of TS and OS by PDPS and PCM, respectively,and strongly support the Janus character of PDPS/PCM-kaolinite.

3.3. Compatibilization of a PS/PMMA blend

Dually modified PDPS/PCM-kaolinite, where the surface ten-sions of the opposing basal surfaces are fine-tuned to match PS andPMMA, respectively, was tested as compatibilizer in films of PSePMMA blends cast from THF, a good solvent for both polymers.For comparison and to be able to estimate the effect of the Januscharacter in excess of the pure Pickering effect additional blendswith unilaterally modified and blends with unmodified kaolinitewere prepared by solvent casting under the same conditions. APS/PMMA ratio of 1:2 (wt/wt) was chosen. The FloryeHugginsparameter for a blend of this molecular weight is cSM ¼ 0.041 at20 �C [34] indicating its incompatibility.

The samples for transmission electron microscopy (TEM) wereprepared by casting the polymer solution with dispersed clay into

a glass vial followed by slow drying and microtome cutting. Thedifference between PS and PMMA is clearly visible in the TEMimages (Fig. 8) even without selective staining. Dark grey areasresult from stronger electron contrast of PS and light grey areasfrom PMMA, which is more easily damaged by the electron beam.The kaolinite particles appear even darker, almost black, and theirshapes are clearly visible due to their strong contrast, the com-pletely white regions are holes in the film, introduced during ultramicrotome cutting.

We are aware of the fact that solvent evaporation will trap thesystem in a metastable state and such prepared films can onlyshow the qualitative aspect of compatibilization achieved by ourhybrid particles. To determine industrially relevant quantitativeeffects, like mechanical properties of compatibilized blends, it isnecessary to conduct extrusion experiments and mechanical tests,which are in preparation, but practically beyond the scope of thispublication.

For pure PS/PMMA blend films of comparable molecular weightthat contain no compatibilizer, it is known, that large (several mmin diameter) spherical domains of the minority phase insidea matrix formed by the majority phase result from phase segre-gation [18].

With unmodified kaolinite we observe macrophase separation(Fig. 8a). No dispersion is achieved, only large aggregates of clayparticles canbe found, separating fromthematrix, trapped inside thepolymer phase where they happen to be upon drying (Fig. 8a). Thisbehaviour is expected due to the clay’s hydrophilic nature (chargedon one side and polar hydroxy groups on the other side) and the factthat it does not form stable dispersions in THF (and thus is hard todisperse in thePolymermixture to startwith). In anotherexperimentwe modified the TS of kaolinite with dodecylamine, which is com-parable in structure to the alkyl ammonium salts used to preparecommercial organoclay like the widely used Cloisite 20A. Here wecan observe clustering in the PMMA phase (Fig. 7b). Like in theLumifuge experiments we expect the kaolinite to form sandwichstructures with the alkyl chains of the organophilized TS aggregatedvia hydrophobic interactions in the inside and the polar OS at theoutside or the other way round. In none of the cases we could finda surface which has high compatibility with any of the polymerphases and thus is not dispersed homogeneously. This observation isin good agreement with literature about other organoclay (e.g.Cloisite 20A),whichdisperses only in the PMMAphaseof a PS/PMMAblend and forms strong clusters in PS Homopolymer blends [24].Obviously modification of one side is not sufficient to align the par-ticles at the interface under these conditions.

In contrast, in the film prepared with the Janus-type PDPS/PCM-kaolinite (Fig. 8c, d) the kaolinite particles are assembled exactly at

Fig. 8. TEM images of 3:7 (wt/wt) PS/PMMA blend films. a) With pristine kaolinite, b) with unilaterally organophilized kaolinite, c) with PDPS/PCM-kaolinite, and d) close up at aninterface. The fraction of the clay is 5 wt% and the scale bar represents 500 nm.

S. Weiss et al. / Polymer 54 (2013) 1388e1396 1395

the interface between both polymer phases. A nearly full coverage ofthe interface by compatibilizer is realized. Due to the Janus characterof themodified kaolinite the interfacial tension of the platelets in theblend interface should be very low. Therefore, the assembly of theparticles at the interface is energetically highly favoured. As a con-sequence, the PS domains are no longer spherical but appear pol-ygonal following the shape of the clay platelets (Fig. 8c, d).

4. Conclusions

A synopsis of all experimental data presented confirms that theexternal basal planes of kaolinite platelets canbe selectivelyaddressedby polystyrene and PMMA, similar to what has been studied in detailfor the molecular modification with Ru(bpy)32þ and a Phosphorous-labelled catechol (3-diphenylphosphinyloxypropyl-3,4-dihydrox-ybenzoate) in [29]. Janus-type PDPS/PCM-kaolinite platelets obtainedby dual modification showed interfacial activity in a solvent-castPS/PMMA blend film. Obviously, blend preparation via melt extru-sionwould be advantageous. Work in that direction is on theway butthe results obtained by solvent casting already give a strong indicationon the efficiency of the hybrid Janus particles as blend compatibilizers.

While the presented work represents a proof of principle, theapproach is, of course, highly modular and should allow for facileand affordable fine-tuning of appropriate compatibilizers fora broad range of blend systems.

The intrinsically polar structure of kaolinite serves as versatilecore of these Janus platelets. Adjustment of the surface tensions ofboth basal planes can easily and selectively be tailored for eachspecific blend composition. Moreover, particle size distribution andmorphology (aspect ratio) may be varied over a wide range by thechoice of the kaolinite source. Furthermore, that concept is notrestricted to kaolinite but can be transferred to any other inorganicmaterial which possesses a polar crystal structure and whereopposing crystal faces are truncated by chemically different func-tional groups, paving the way to selective modification.

An additional advantage of the concept should be an inherentreinforcement of the blend by the inorganic filler, which, moreover,is concentrated at the blends interfaces. This should create a syn-ergistic effect stretching far beyond a pure Pickering effect andshould boost the mechanical properties of the blend.

Acknowledgements

We thank the Deutsche Forschungsgemeinschaft (SFB 840,project A1), the Elite Study Program “Macromolecular Science”, andthe International Graduate School “Structure, Reactivity and Prop-erties of Oxide Materials” within the Elite Network Bavaria forfinancial support. The authors thank Prof. Dr. Jürgen Senker forproviding the NMR equipment, Jasmin Schmid for help in con-ducting practical experiments and Annika Pfaffenberger for per-forming the TEM measurements.

References

[1] Koning C, van DuinM, Pagnoulle C, Jerome R. Prog Polym Sci 1998;23(4):707e57.[2] Kleiner LW, Karasz FE, Macknight WJ. Polym Eng Sci 1979;19(7):519e24.[3] Hara M, Sauer JA. J Macromol Sci Rev Macromol Chem Phys 1998;C38(2):

327e62.[4] Ruckdäschel H, Rausch J, Sandler JKW, Altstadt V, Schmalz H, Müller AHE.

Polym Eng Sci 2008;48(11):2111e25.[5] Paul DR, Barlow JW. J Macromol Sci Rev Macromol Chem Phys 1980;C18(1):

109e68.[6] Barlow JW, Paul DR. Polym Eng Sci 1984;24(8):525e34.[7] BrownHR, Char K, DelineVR, Green PF.Macromolecules 1993;26(16):4155e63.[8] Ruckdäschel H, Sandler JK, Altstaedt V, Schmalz H, Abetz V, Müller AH.

Polymer 2007;48(9):2700e19.[9] Walther A, Müller AHE. Soft Matter 2008;4(4):663e8.

[10] Ray SS, Pouliot S, Bousmina M, Utracki LA. Polymer 2004;45(25):8403e13.[11] Khatua BB, Lee DJ, Kim HY, Kim JK. Macromolecules 2004;37(7):2454e9.[12] Binks BP, Lumsdon SO. Langmuir 2000;16(23):8622e31.[13] Binks BP. Curr Opin Colloid Interface Sci 2002;7(1e2):21e41.[14] De Gennes PG. Rev Mod Phys 1992;64(3):645e8.[15] Walther A, Matussek K, Müller AHE. ACS Nano 2008;2(6):1167e78.[16] Ray SS, Bousmina M. Macromol Rapid Commun 2005;26(6):450e5.[17] Chen GX, Kim HS, Kim ES, Yoon JS. Polymer 2005;46(25):11829e36.

S. Weiss et al. / Polymer 54 (2013) 1388e13961396

[18] Li YJ, Shimizu H. Polymer 2004;45(22):7381e8.[19] Wang Y, Zhang Q, Fu Q. Macromol Rapid Commun 2003;24(3):231e5.[20] Chen GX, Yoon JS. Polym Int 2005;54(6):939e45.[21] Si M, Araki T, Ade H, Kilcoyne ALD, Fisher R, Sokolov JC, et al. Macromolecules

2006;39(14):4793e801.[22] Li YJ, Shimizu H. Macromol Rapid Commun 2005;26(9):710e5.[23] Perro A, Reculusa S, Ravaine S, Bourgeat-Lami EB, Duguet E. J Mater Chem

2005;15(35e36):3745e60.[24] Yang SM, Kim SH, Lim JM, Yi GR. J Mater Chem 2008;18(19):2177e90.[25] Yu H, Chen M, Rice PM, Wang SX, White RL, Sun SH. Nano Lett 2005;5(2):

379e82.[26] Neder RB, Burghammer M, Grasl T, Schulz H, Bram A, Fiedler S. Clays Clay

Miner 1999;47(4):487e94.[27] Kogure T, Elzea-Kogel J, Johnston CT, Bish DL. Clays Clay Miner 2010;58(1):

62e71.[28] Brady PV, Cygan RT, Nagy KL. J Colloid Interface Sci 1996;183(2):356e64.[29] Hirsemann D, Shylesh S, De Souza RA, Diar-Bakerly B, Biersack B, Mueller DN,

et al. Angew Chem Int Ed 2012;51(6):1348e52.

[30] Wang WL, Chai SC, Ye QZ. Bioorg Med Chem Lett 2009;19(4):1080e3.[31] Levitt MH, Raleigh DP, Creuzet F, Griffin RG. J Chem Phys 1990;92(11):

6347e64.[32] Rocha J, Klinowski J. Angew Chem Int Ed 1990;29(5):553e4.[33] Plamper FA, Ruppel M, Schmalz A, Borisov O, Ballauff M, Müller AHE.

Macromolecules 2007;40(23):8361e6.[34] Russell TP, Hjelm RP, Seeger PA. Macromolecules 1990;23(3):890e3.[35] Tighilt F, Gabouze N, Sam S, Belhousse S, Beldjilali K. Surf Sci 2007;601(18):

4217e21.[36] Gersappe D, Balazs AC. Phys Rev E 1995;52:5061e4.[37] Ray S, Okamoto M. Prog Polym Sci 2003;28:1539e641.[38] Fischer B, Ziadeh M, Pfaff A, Breu J, Altstädt V. Polymer 2012;53:3230e7.[39] Möller MW, Kunz DA, Lunkenbein T, Sommer S, Nennemann A, Breu J. Adv

Mater 2012;24:2142e7.[40] Schütz MR, Kalo H, Lunkenbein T, Groschel AH, Müller AHE, Wilkie CA, et al.

J Mater Chem 2011;21:12110e6.[41] Tulevski GS, Miao Q, Fukuto M, Abram R, Ocko B, Pindak R, et al. J Am Chem

Soc 2004;126:15048.