Ternary Graft Copolymers and Their Use in Nanocapsule PreparationFeng Liu,†,‡ Jiwen Hu,*,†,‡ Guojun Liu,*,†,§ Chengmin Hou,† Shudong Lin,† Hailiang Zou,†

Ganwei Zhang,† Jianping Sun,† Hongsheng Luo,† and Yuanyuan Tu†

†Guangzhou Institute of Chemistry, Chinese Academy of Sciences, Guangzhou, P. R. China 510650‡University of Chinese Academy of Sciences, Beijing, P. R. China 100049§Department of Chemistry, Queen’s University, 90 Bader Lane, Kingston, Ontario, Canada K7L 3N6

ABSTRACT: Grafting three types of polymer chains onto thebackbone of a fourth polymer yielded a ternary graftcopolymer. The copolymer dispersed oil droplets in waterwith one type of graft stretching into the oil phase, the secondtype forming a thin membrane separating the two phases, andthe third type stretching into the water phase. Since the secondtype was also photo-cross-linkable, shining UV light on the system produced permanent nanocapsules. To produce the graftcopolymer, the backbone polymer used was poly(3-azido-2-hydroxypropyl methacrylate), P(GMA-N3). The grafts used were allend-functionalized by alkyne groups, and the polymers were poly(ethylene glycol) methyl ether (MPEG), polystyrene (PS), andpoly(2-cinnamoyloxyethyl methacrylate) (PCEMA), respectively. Evidently, MPEG was water-soluble, PS was soluble in the usedoil decahydronaphthalene (DN), and PCEMA was photo-cross-linkable and soluble in neither water nor DN. The grafts denotedas MPEG−CCH, PS−CCH, and PCEMA−CCH were coupled to P(GMA-N3) via click chemistry between the azide andalkyne units. Under the used conditions, the one-pot grafting reactions were quantitative.

I. INTRODUCTION

A linear polymer chain bearing three types of pendant polymerchains is a ternary graft copolymer.1−3 If the total density of thegrafted chains is high so that they strongly repel one anotherdue to steric hindrance and the backbone chain is long relativeto the grafts, the copolymer is called a ternary cylindrical brushor heterografted ternary cylindrical brush. While there havebeen many reports on heterografted binary cylindrical brushesin the past two decades,4−13 ternary cylindrical brushes or graftcopolymers have not caught the attention of the polymercommunity. This situation is not warranted as ternary graftcopolymers or brushes are interesting and potentially useful. Toillustrate this point, we have developed a facile one-pot methodto synthesize a family of ternary graft copolymers. We showthat these polymers are useful in the making of nanocapsulesthat may find applications in controlled release applica-tions.14−17

Graft copolymers are normally synthesized from the “graft-through”, “graft-from”, and “graft-onto” methods.1,18 In thegraft-through method, macromonomers are prepared first. Thepolymerization of one type of macromonomer yields acylindrical homopolymer brush.19,20 Heterograft binary cylin-drical brushes are prepared from the random copolymerizationof two types of macromonomers.21,22 The sequential polymer-ization of two or more types of macromomers yields blockycylindrical brushes.22 Core−shell and core−shell−coronacylindrical brushes are prepared from the polymerization ofdiblock and triblock macromonomers, respectively.23,24 In thegraft-from method, the grafts are prepared from polymerizingmonomers using initiating sites on a polymer backbone. If twotypes of initiating sites are used to grow two types of polymer

chains, a heterograft binary brush or a binary graft copolymer isprepared.4 Binary graft copolymers are prepared in the graft-onto method by attaching two different types of polymer chainsonto a third polymer backbone.25,26 Compared to the other twomethods, the graft-onto method may not achieve high graftingdensities. An advantage is the easy characterization of thecomponents before they are linked. Further, this method issuited for the one-pot synthesis of multicomponent graftcopolymers. For applications where a high grafting density isnot desired, this method yields versatile multicomponentcopolymers.The graft-onto method was used in this study to produce the

desired ternary graft copolymers. The backbone polymers usedwere two poly(3-azido-2-hydroxypropyl methacrylate), P-(GMA-N3), samples. The grafts used were end-functionalizedpoly(ethylene glycol) methyl ether (MPEG), polystyrene (PS),and poly(2-cinnamoylethyl methacrylate) (PCEMA), respec-tively. The precursory grafts MPEG−CCH, PS−CCH,and PCEMA−CCH were coupled to P(GMA-N3) via Cu-catalyzed alkyne−azide cycloaddition (CuAAC). At a [CCH] to [N3] molar ratio of ≤23/100 or when x + y + z ofScheme 1 was ≤23%, the one-pot grafting reactions werequantitative. The residual N3 groups were then deactivated byreaction with propargyl alcohol.To prepare nanocapsules, two of the synthesized copolymers

were each used to disperse decahydronaphthalene (DN) inwater by emulsification. Stretching into the emulsified DN

Received: December 28, 2012Revised: February 20, 2013

Article

pubs.acs.org/Macromolecules

© XXXX American Chemical Society A dx.doi.org/10.1021/ma302663u | Macromolecules XXXX, XXX, XXX−XXX

droplets were the DN-soluble PS chains. The water-solubleMPEG chains stretched away from the droplets into theaqueous phase. Since PCEMA is soluble in neither water norDN, it formed a thin membrane separating the two phases. Thismembrane was photo-cross-linked by shining UV light on thesystem to yield “permanent” capsules.As far as the applications of cylindrical brushes are

concerned, many reports have appeared on the applicationsof binary cylindrical brushes, core−shell and core−shell−corona cylindrical brushes, and blocky cylindrical brushes. Forexample, binary cylindrical brushes have been used to disperseoil in water to make miniemulsions.27 Binary cylindrical brusheshave also been assembled in selective solvents into rodlikeaggregates,28 spherical micelles,29,30 and vesicles.31 In the caseof core−shell cylindrical brushes, their cores have been used totemplate metal and metal oxide nanoparticles32,33 or conduct-ing polymer34 and for drug loading.35 Drug has also beenloaded into capsules or tubes made from core−shell−coronacylindrical brushes.36 To make these hollow structures, the coreblock of the core−shell−corona brushes was degraded after theshell block was cross-linked.24,37,38 A remarkable application ofthe blocky cylindrical brushes has been in the making ofphotonic crystals.39,40

II. EXPERIMENTAL SECTIONMaterials. CuBr (98%), CuCl (98%), CuCl2 (99%), CuSO4·5H2O

(99%), anhydrous aluminum chloride (99%), sodium ascorbate (SA,99%), disodium ethylenediamine tetraacetate (EDTA, 99%), 2-hydroxyethyl methacrylate (HEMA, 98%), styrene (St, 99%), glycidylmethacrylate (98%), tetrabutylammonium fluoride (TBAF, 98%), 1-(3-(dimethylamino)propyl)-3-ethylcarbodiimide hydrochloride(EDC·HCl, 98%), decahydronaphthalene (DN, 99%), dichloro-methane (DCM, 99%), pyridine (99%), N,N-dimethylformamide(DMF, 99%), and diphenyl ether (99%) were purchased from AladdinReagent of China. CuBr and CuCl were purified by rinsing with glacialacetic acid, methanol, and diethyl ether before they were dried undervacuum, while CuSO4·5H2O, sodium ascorbate, and EDTA were usedas received. CuCl2 was dried under vacuum before use. HEMA waspurified according to a procedure described in the literature.41 Theprocedure involved washing an aqueous solution (25 vol % HEMA) ofmonomer with hexanes (4 × 200 mL), salting the monomer out of theaqueous phase by addition of NaCl, drying over MgSO4, and distillingunder reduced pressure. Styrene was washed thrice with equalamounts of 10% NaOH aqueous solution and then washed with wateruntil the water tested was neutral. It was dried over anhydrous sodiumsulfate and then distilled under reduced pressure. Glycidylmethacrylate was purified via distillation under reduced pressure.Poly(ethylene glycol) methyl ether (MPEG, Mn = 5000 g/mol) waspurchased from Aldrich and used as received. Pyridine was refluxedover CaH2 overnight and distilled prior to use. Diphenyl ether andmethylbenzene (Aldrich, 99%) were refluxed over a sodium wire anddistilled before use. N,N-Dimethylformamide (DMF) was dried overanhydrous magnesium sulfate for 3 days and distilled before use.Dimethylaminopyridine (DMAP, Aldrich, 99.9%) was purified via

recrystallization from toluene. TBAF, anhydrous aluminum chloride,EDC·HCl, N,N,N′,N″,N″-pentamethyldiethylenetriamine (PMDETA,Aldrich, 99%), DN, DCM, cinnamoyl chloride (Aldrich, 98%), and2,2-dipyridyl (bpy, Aldrich, 99%) were all used as received. 2-Methoxyethyl 2-bromoisobutyrate, propargyl 2-bromopropanoate, 4-oxo-4-(prop-2-yn-1-yloxy)butanoic acid, and 3-(trimethylsilyl)-prop-argyl 2-bromoisobutyrate were synthesized according to literatureprocedures.25,42 All other reagents and solvents were used as receivedunless otherwise indicated.

PGMA and P(GMA-N3). PGMA41 was prepared via atom transferradical polymerization by using 2-methoxyethyl 2-bromoisobutyrate asthe initiator and CuCl/PMDETA as the catalyst system. Diphenylether (20.0 mL), 2-methoxyethyl 2-bromoisobutyrate (0.675 g, 3.0mmol), GMA (17.1 g, 0.120 mol), CuCl (0.315 g, 3.0 mmol), and amagnetic stirring bar were added into a 100 mL round-bottom flask.The flask was subjected to an “evacuate and argon backfill” processthrice before it was further deoxygenated by a “freeze, evacuate, thaw,and argon fill” process thrice. PMDETA (0.522 g, 3.0 mmol) wasinjected into the flask using a degassed syringe, and the flask was thenimmersed in a preheated oil bath at 30 °C for 30 min for GMApolymerization. This was followed by immersing the flask into liquidnitrogen and then exposing the contents to air. The resultant viscousreaction mixture was diluted with DCM (100 mL) and passed throughan activated neutral alumina column. The filtrate was concentrated to∼40 mL via rotary evaporation and subsequently added into 500 mLof hexane to precipitate the polymer. It was redissolved in ∼40 mL ofDCM and precipitated into 500 mL of hexane again. The precipitatewas dried under vacuum for 24 h to yield 16.9 g of PGMA in a 99%yield.

PGMA102 was prepared analogously, except the use of a differentGMA to initiator molar ratio. The yield of the final polymer was 96%.

To attach azide groups, PGMA41 or PGMA102 (6.50 g, 0.046 mol ofepoxide groups) was dissolved in DMF (150 mL). Sodium azide (6.5g, 0.10 mol) and anhydrous aluminum chloride (0.10 g, 0.75 mmol)was then added to polymeric DMF solution, which was subsequentlystirred at 50 °C for 25 h. After the reaction, insoluble impurities wereremoved by filtration. After most of the DMF had been evaporated,P(GMA-N3) was precipitated in water (500 mL). The polymer wasredissolved in ∼20 mL of DMF and precipitated into 500 mL of wateragain. The product was filtrated, washed with water, and vacuum-driedto give 6.4 g of white solid at a 75% yield.

PHEMA−CCH. Alkyne end-functionalized poly(2-hydroxyethylmethacrylate) (PHEMA−CCH) was prepared via ATRP using 3-(trimethylsilyl)propargyl 2-bromoisobutyrate as the initiator. Meth-anol (8.0 mL), methyl ethyl ketone (12.0 mL), 3-(trimethylsilyl)-propargyl 2-bromoisobutyrate (0.535 g, 1.93 mmol), HEMA (20.00 g,0.154 mol), CuCl2 (25 mg, 0.19 mmol), and CuCl (0.191 g, 1.93mmol), along with a magnetic stir bar, were placed inside a 100 mLround-bottom flask. The flask was subjected to an “evacuate and argonbackfill” process thrice before it was further deoxygenated by a “freeze,evacuate, thaw, and argon fill” process thrice. Bpy (0.599 g, 4.0 mmol)dissolved in 0.5 mL of methanol was deoxygenated by a “freeze,evacuate, thaw, and argon fill” process thrice and then injected into theflask using a degassed syringe. The flask was immersed into an oil bathpreheated to 50 °C, and the polymerization was allowed to go at thistemperature for 2.5 h before the flask was removed and immersed intoliquid nitrogen and opened to introduce air. The mixture was dilutedwith 100 mL of methanol and was subsequently passed through anactivated neutral alumina column before it was concentrated via rotaryevaporation to ∼30 mL. The concentrate was added into 500 mL ofwater to precipitate the polymer. The polymer was redissolved in ∼30mL of methanol and precipitated into 500 mL of water again. Theprecipitate was dried under vacuum for 24 h, yielding 13.0 g of theproduct as a white powder in a 65% yield.

To remove the trimethylsilyl protecting group in the initiating unit,the product obtained above (10.0 g), DMF (30 mL), andtetrabutylammonium fluoride (5.0 g, 0.019 mol) were stirred togetherat room temperature for 72 h. After most of the solvent was removedunder reduced pressure, the polymer solution (∼10 mL) wasprecipitated in water (500 mL) to remove residual salts. The obtained

Scheme 1. Structure of PGMA-g-(PS-r-PCEMA-r-MPEG)

Macromolecules Article

dx.doi.org/10.1021/ma302663u | Macromolecules XXXX, XXX, XXX−XXXB

product was dried under vacuum for 24 h, thus generating 9.3 g ofPHEMA−CCH as a white powder in a 93% yield.Synthesis of PCEMA−CCH. To cinnamate the hydroxyl groups

of PHEMA−CCH, PHEMA−CCH (4.0 g) was mixed withcinnamoyl chloride (7.68 g) and dissolved in freshly distilled pyridine(160 mL). The mixture was stirred overnight before it was centrifugedto remove the pyridinium salt. After most of the solvent was removedunder reduced pressure, the polymer solution was added to an excessof methanol to precipitate PCEMA−CCH. The polymer wasredissolved in ∼15 mL of DMF and precipitated into 500 mL ofmethanol again. The polymer was then dried at room temperatureunder vacuum for 24 h, generating 7.63 g of product as white powderin a 93% yield.MPEG−CCH. MPEG−CCH was synthesized by esterification

of MPEG with an excess of 4-oxo-4-(prop-2-yn-1-yloxy)butanoic acidusing EDC·HCl and DMAP as catalysts. In particular, MPEG (20.0 g,4.0 mmol, Mn = 5000 g/mol), 4-oxo-4-(prop-2-yn-1-yloxy)butanoicacid (1.87 g, 0.012 mol), DMAP (1.95 g, 0.016 mol), and EDC·HCl(2.29 g, 0.012 mol) were dissolved in 100 mL of DCM. The solutionwas then stirred at room temperature for 72 h before it was washedtwice in sequence with each of the following liquids: 2 M HCl (5 mL),saturated NaHCO3 (10 mL), and distilled water (20 mL). The organiclayer was collected and dried over anhydrous magnesium sulfate for 12h and subsequently filtered. The filtrate was concentrated by rotaryevaporation to ∼30 mL and then added into anhydrous diethyl ether(500 mL) to precipitate the polymer. The polymer was redissolved in∼30 mL of DCM and precipitated into 500 mL of diethyl ether again.

The precipitate was dried under vacuum for 24 h, generating 16.0 g ofMPEG−CCH as a white powder in an 80% yield.

PS−CCH. PS−CCH was prepared by ATRP using propargyl2-bromopropanoate as the initiator. Methylbenzene (15.0 mL),propargyl 2-bromopropanoate (0.210 g, 1.0 mmol), St (26.5 g,0.255 mol), and CuBr (0.143 g, 1.0 mmol) and a magnetic stir barwere placed inside a 100 mL round-bottom flask. The flask wassubjected to an “evacuate and argon backfill” process thrice before itwas further deoxygenated by a “freeze, evacuate, thaw, and argon fill”process thrice. PMDETA (0.173 g, 1.0 mmol) was injected into theflask using a degassed syringe, and the flask was then immersed in apreheated oil bath at 90 °C. The mixture turned slightly greenimmediately and slightly yellow over time. The reaction was performedfor 5.5 h before the flask was immersed into liquid nitrogen and thecontent was exposed to air. The mixture was diluted by adding 100 mLof tetrahydrofuran (THF). This solution was subsequently passedthrough an activated neutral alumina column before the filtrate wasconcentrated to ∼40 mL via rotary evaporation and added into 500mL of methanol to precipitate the polymer. The polymer wasredissolved in ∼30 mL of THF and precipitated into 500 mL ofmethanol again. The precipitate was dried under vacuum for 24 h,yielding 13.5 g of PS−CCH as a white powder in a 51% yield.

Preparation of Ternary Graft Copolymers. In an examplepreparation, DMF (16.0 mL), P(GMA-N3)41 (0.11 g, 0.59 mmol ofazide groups), PCEMA−CCH (0.22 g, 7.05 μmol of alkynegroups), PS−CCH (0.40 g, 0.030 mmol of alkyne groups), and anaqueous sodium ascorbate solution (90 mg, 4.54 mmol, dissolved into0.20 mL of water) were mixed in a 50 mL round-bottomed flask and

Scheme 2. Synthetic Route toward PGMA-g-(PS-r-PCEMA-r-MPEG)

Macromolecules Article

dx.doi.org/10.1021/ma302663u | Macromolecules XXXX, XXX, XXX−XXXC

deoxygenated via bubbling with argon for 50 min. Then, 0.20 mL of anaqueous solution of CuSO4·5H2O (50 mg, 0.20 mmol) was added.This was followed by stirring the reaction mixture at roomtemperature for 24 h. Subsequently, 6.0 mL of a degassed DMFsolution of MPEG−CCH (0.40 g, 0.080 mmol of alkynyl groups)was introduced into the flask using a syringe. The reaction was allowedto go for another 48 h. Lastly, propargyl alcohol (0.80 g, 1.4 mmol)was injected into the flask, and the reaction mixture was stirred for 16h to deactivate the residual azide groups. After all of the solvent wasremoved under reduced pressure, the product was dissolved with 100mL of DCM and extracted with 10 mL of a saturated aqueous EDTAsolution to remove the catalyst. The organic layer was then collectedand dried with anhydrous sodium sulfate for 5 h. After most of thesolvent was removed via rotary evaporation, the residue (∼2.0 mL)was added into diethyl ether to precipitate the polymer. The obtainedproduct was dried under vacuum for 24 h, yielding 1.04 g of theproduct as a white powder in a 92% yield.Capsule Formation. Water (20 mL) and DN with volume

between 0.10 and 0.20 mL were added into a 100 mL round-bottomflask immersed in a water bath. The mixture was stirred mechanicallyat 1600 rpm with a hemispherically shaped Teflon blade attached tothe end of a stirring shaft. To it was added dropwise 0.50 mL DCMcontaining 20 mg of a graft copolymer. This was followed by stirring atroom temperature for 30 min, at 32 °C for 2 h, and at 50 °C for 1 h toremove DCM.To cross-link the PCEMA layer, the emulsion was diluted with 1

volume of water and then irradiated in a quartz round-bottom flaskunder magnetic stirring at 25 °C for 1−2 h. The UV light was from a100 W Hg lamp powered by a universal system. To determine thedegree of PCEMA double-bond conversion, the sample was dilutedwith DMF and the absorbance decrease at 274 nm was monitored.The typical CEMA double conversion used was ∼60%.Size Exclusion Chromatography. The number-average molec-

ular weight (Mn) and polydispersity index (Mw/Mn) of each polymerwere determined at 35 °C using a Waters 1515 size exclusionchromatograph (SEC) equipped with a Waters 2414 refractive index(RI) detector. DMF containing tetrabutylammonium bromide (0.05mg/L) or THF was used as the eluant and the columns used were thestyragel HR3 and HR4 columns calibrated by narrow PS standards.

NMR and Dynamic Light Scattering. 1H NMR spectra wereobtained on a Bruker DMX-400 spectrometer. Deuterated chloroform(CDCl3), deuterated water (D2O), or deuterated dimethyl sulfoxide(DMSO-d6) was used as the solvent. The hydrodynamic diameters(Dh) of the capsules and their polydispersity indices (PDI) weredetermined by dynamic light scattering (DLS) on a Malven ZetasizerNano System. The emulsions were passed through 1.0 μm filtersbefore DLS measurements. The measurements were conducted in a3.0 mL quartz cuvette, using an 800 nm diode laser at 25 °C, and thescattering angle used was 90°. Each set of Dh and PDI values was theaverage from five measurements.

Transmission Electron Microscopy. To visualize the nano-capsules by transmission electron microscopy (TEM), dispersions ofthe capsules were aerosprayed43 onto nitrocellulose-coated 200 meshcopper grids before they were dried under high vacuum to remove anyvolatile residues. The samples were subsequently stained with RuO4for 30 min before TEM observation on a JEM-100CX II microscopeoperated at 80 kV.

Atomic Force Microscopy (AFM). Dispersions of the cross-linkednanocapsules were aero-sprayed onto freshly cleaved mica surfacesbefore they were dried under high vacuum to remove any volatileresidues. The dried nanocapsules were observed using a MultiMode 8SPM AFM system (Bruker) using a ScanAsyst mode.

III. RESULTS AND DISCUSSIONP(GMA-N3), MPEG−CCH, PS−CCH, and PCEMA−CCH were first synthesized, and then the latter threepolymers were grafted onto the P(GMA-N3) backbone to yieldPGMA-g-(PS-r-PCEMA-r-MPEG). Scheme 2 shows thereactions used to prepare the individual components and thefinal graft copolymers.

P(GMA-N3). According to Scheme 2, PGMA was theprecursor to P(GMA-N3). PGMA was synthesized by ATRPfollowing a modified literature method:44 using diphenyl etheras the solvent, 2-methoxyethyl 2-bromoisobutyrate as theinitiator, CuCl as the catalyst, and PMDETA as the ligand. 2-Methoxyethyl 2-bromoisobutyrate was used as the initiator

Figure 1. 1H NMR spectra and peak assignments for PGMA41 and (PGMA-N3)41 (a), PHEMA−CCH and PCEMA−CCH (b), MPEG−CCH (c), and PS−CCH (d).

Macromolecules Article

dx.doi.org/10.1021/ma302663u | Macromolecules XXXX, XXX, XXX−XXXD

mainly to facilitate the degree of polymerization determinationby NMR because the initiator’s −OCH3 group in the 1H NMRspectrum did not overlap with resonances of PGMA protons.Different halides were used in the initiator and the catalyst toslow down the rate of polymerization relative to initiation andthus to produce polymers with low polydispersity indices.44

Under these conditions, high monomer conversions wereachieved within 30 min.Two PGMA homopolymers (PGMA41 and PGMA102 with

GMA repeat units of 41 and 102) were synthesized using themonomer to initiator molar ratios [M]0/[I]0 of 40 and 100,respectively. The resultant polymers were analyzed by 1H NMRin CDCl3 solvent. Shown in Figure 1 is a 1H NMR spectrumfor PGMA41. The repeat unit number of 41 was obtained fromcomparing the peak area of the initiator’s −OCH3 group at δ3.35 ppm with that of the epoxide CH protons 3.21 ppm. Thenumber 102 was determined analogously. These numberscompared well with the targeted repeat unit numbers and thehigh GMA conversions.The samples were also analyzed by size exclusion

chromatography (SEC) using THF as the eluant. Thepolydispersity indices Mw/Mn were low at 1.21 and 1.22 forthe two polymers based on PS calibration standards (Table 1).Coincidentally, the SEC number-average molecular weights Mn

agreed with those calculated from the GMA repeat unitnumbers determined by 1H NMR (Table 1).The azide groups were introduced by reacting the oxirane

rings of GMA with sodium azide.25 Matyjaszewski and co-workers confirmed that the azide anion attacked exclusively theless substituted carbon atom of the epoxide rings.45 Thecompletion of this reaction was confirmed by 1H NMR andFTIR results. The signals of the CH and CH2 protons of theepoxide ring at 2.62, 2.82, and 3.21 ppm disappeared in theP(GMA-N3)41 spectrum shown in Figure 1a after the reactionbetween PGMA and NaN3. This event was also accompanied

by the disappearance of a characteristic infrared (IR) absorptionpeak at 909 cm−1 for the epoxide ring and the appearance ofcharacteristic absorption peaks at 2104 cm−1 for the azidegroup and at 3500 cm−1 for hydroxyl group (Figure 2a,b).The P(GMA-N3) samples were also analyzed by SEC using

THF as the eluant. Compared with their precursory PGMA, the“apparent” molecular weights of P(GMA-N3) increased, asexpected. The polydispersity indices increased slightly as welland were 1.26 and 1.25 for P(GMA-N3)41 and P(GMA-N3)102,respectively. These peaks broadened probably because of theenhanced interaction between P(GMA-N3) and the SECcolumns. For example, we previously observed very broadSEC peaks for P(GMA-N3) when DMF was used as the eluant,and the peaks narrowed after the P(GMA-N3) hydroxyl groupswere reacted with acetic anhydride or were masked.25

PCEMA−CCH. According to Scheme 2, PCEMA−CCH was synthesized in three steps. First, reacting 3-(trimethylsilyl)propargyl alcohol with 2-bromoisobutyric bro-mide following a literature procedure yielded 3-(trimethylsilyl)-propargyl 2-bromoisobutyrate.25,42 The latter was then used toinitiate HEMA polymerization to yield PHEMA−CCH.Reacting PHEMA−CCH with cinnamoyl chloride eventuallyproduced PCEMA−CCH.The alkyne proton in 3-(trimethylsilyl)propargyl 2-bromoi-

sobutyrate was replaced by a trimethylsilyl group because it wasdifficult to obtain well-defined PHEMA−CCH whenpropargyl 2-bromopropanoate was used as the initiator. Thisdifficulty should be due to side reactions between the radicalsand the terminal alkyne group during the polymerization.46

HEMA polymerization in methanol using 3-(trimethylsilyl)-propargyl 2-bromoisobutyrate as the initiator has beenreported.47 At the ratios of [HEMA]0/[initiatoir]0/[CuCl]0/[bpy]0 of 80/1/1/2, we found that the polymerization inmethanol was too fast to produce well-defined polymers. Thus,a modified literature method was used to synthesize PHEMA−

Table 1. Preparation Conditions and Molecular Characteristics of the Precursory Polymers

sample [M]0/[I]0 yielda (%) NMR DP NMR Mn, (kg/mol) SEC Mn (kg/mol) SEC Mw/Mn

PGMA41 40:1 99 41 5.8 5.4 1.21P(GMA-N3)41 41 7.5 7.9 1.26PGMA102 100:1 96 102 14.4 13.8 1.22P(GMA-N3)102 102 18.8 19.5 1.25MPEG−CCH 114 5.0 8.7 1.02PS−CCH 255:1 51 130 13.5 15.6 1.11PCEMA−CCH 80:1 65 121 31.5 16.4 1.16

aDetermined by gravimetric analysis.

Figure 2. IR spectra of PGMA41 (a), P(GMA-N3)41 (b), and ternary graft copolymers (GP 1) before (c) and after (d) reaction with excess propargylalcohol.

Macromolecules Article

dx.doi.org/10.1021/ma302663u | Macromolecules XXXX, XXX, XXX−XXXE

CCH using 2-butanone and methanol (v/v = 3/2) as thesolvent and CuCl/CuCl2 as the catalyst.

48 The polymerizationwas stopped at a 65% HEMA conversion.The trimethylsilyl protecting group prevented the side

reactions and was readily removed by stirring the resultantpolymer in DMF with tetrabutylammonium fluoride. Theresultant PHEMA−CCH was characterized by 1H NMR,and a spectrum for it is shown in Figure 1b together with peakassignments. The signal of the trimethylsilyl protecting group at0.14 ppm was definitely absent in the spectrum, confirming thefull removal of the protecting group.The PHEMA−CCH spectrum of Figure 1b also allowed

the determination of the number of repeat units for PHEMA.Comparing the peak area of the initiator’s methylene protons(−CC−CH2−) at 4.60 ppm (Figure 1b) with those of theethyl groups of the hydroxyethyl group of HEMA at 3.88 ppmyielded a repeat unit number of 120. This number wassubstantially larger than 52 calculated from [HEMA]0/[Initiator]0 and HEMA conversion but should be correctbecause the initiation efficiency has been shown to be low inATRP of HEMA in the past.41

Reacting PHEMA−CCH with cinnamoyl chloride yieldedPCEMA−CCH. This cinnamation reaction was shown in thepast to be quantitative.49 Its quantitative occurrence here can beconcluded by comparing the 1H NMR spectra of PHEMA−CCH and PCEMA−CCH shown in Figure 1b. The c andd peaks of the PHEMA−CCH peaks at 3.56 and 3.88 ppmtotally disappeared and were replaced by the c and d peaks ofPCEMA−CCH at 4.06 and 4.25 ppm after the cinnamationreaction.The PCEMA−CCH spectrum of Figure 1b also allowed

the comparison of integral of the methylene protons (HCCCH2−) of the terminal propargyl group at 4.6 ppm with thatof the cinnamate aromatic protons between 7.26 and 7.57 ppm.This operation yielded a number-average degree of polymer-ization of 121 for PCEMA−CCH. This repeat unit numberagreed with that determined at the PHEMA−CCH stage andrendered us confidence in our determined numbers.Aside from the low initiation efficiency, the polymerization

seemed to be well behaved. The SEC polydispersity index (Mw/Mn) of PCEMA was low at 1.16.MPEG−CCH. According to Scheme 2, MPEG−CCH

was prepared in two steps. First, propargyl alcohol was reactedwith succinic anhydride to produce 4-oxo-4-(prop-2-yn-1-yloxy)butanoic acid.25 The latter was then reacted with theterminal hydroxyl group of a commercial MPEG polymer witha nominal molecular weight of 5000 to yield MPEG−CCH.42

Literature procedures were used to perform the reactions,25

and the resultant MPEG−CCH was carefully characterizedby SEC and 1H NMR. Shown in Figure 1c is a 1H NMRspectrum of MPEG−CCH together with peak assignments.A comparison of the proton integrals corresponding to theMPEG terminal CH3−O− group (CH3−O−CH2− at 3.26ppm), and that of the two methylene groups of 4-oxo-4-(prop-2-yn-1-yloxy)butanoic acid (−OCCH2CH2COO− at 2.66ppm) yielded a ratio of 3.00/4.03, which was the same, withinexperimental error, as 3/4. Thus, the esterification wasquantitative.As expected for PEG prepared from anionic ring-opening

polymerization, the polydispersity of this sample was low at1.02. The SEC molar mass of this sample was apparent and was

higher than that provided by the supplier because the SECsystem was calibrated by PS rather than PEG standards.

PS−CCH. The initiator used for ATRP of styrene wasprepared by reacting propargyl alcohol with 2-bromopropionylbromide. A literature method was followed for this synthesis aswell as the synthesis of PS−CCH.25 The resultant PS−CCH was again carefully characterized by 1H NMR and SEC.Shown in Figure 1d is a 1H NMR spectrum of PS−CCH

together with peak assignments. Comparing the integral of thea peak of the −CH3 protons occurring between 0.8 and 1.0ppm and that of the e peaks of the benzene ring between 6.2and 7.2 ppm yielded an average degree of polymerization (DP)of 130 for PS−CCH. This DP compares well with thatcalculated from the styrene to initiator molar feed ratio of 255and the monomer conversion of 51% determined gravimetri-cally. This suggested that ATRP was well behaved. However,the number-average molecular weight of 1.36 × 104 calculatedfrom this DP value was 14% lower than 1.56 × 104 determinedfrom SEC. The difference could be due to the errors associatedwith SEC and 1H NMR measurements, and the more likelyvalue should be between 1.36 × 104 and 1.56 × 104. The moreinteresting result from SEC was the low polydispersity index of1.11 for the polymer.

Ternary Graft Copolymers. PGMA-g-(PS-r-PCEMA-r-MPEG) samples were synthesized by coupling P(GMA-N3)with MPEG−CCH, PS−CCH, and PCEMA−CCH.Since MPEG−CCH were known to readily graft to P(GMA-N3),

26,48 we started by reacting the more hindered PS−CCHand PCEMA−CCH chains with P(GMA-N3) for 24 h beforeMPEG−CCH addition. This was followed by another 48 hof reaction before an excess of propargyl alcohol was added toexhaust the residual azide groups.Three graft copolymers (GP) denoted as GP 1, 2, and 3 were

prepared by grafting PS−CCH, PCEMA−CCH, andMPEG−CCH. While P(GMA-N3)41 was used as thebackbone for GP 1 and 2, the backbone used for GP 3 wasP(GMA-N3)102. The recipes used to prepare the copolymersare listed in Table 2.

The grafting reactions were followed, and the end productswere analyzed by SEC, 1H NMR, and FTIR. Figure 3 comparesSEC traces of the precursors except P(GMA-N3)102 andproducts made using the recipes shown in Table 2. HereDMF rather than THF was used as the eluant because theternary graft copolymers were better solubilized in DMF. A sideeffect of using DMF was that the variation trend of the apparentmolecular weights measured in THF and reported in Table 1was missing by the data shown in Figure 3. For example, theP(GMA-N3)41 peak in Figure 3 had an abnormally short

Table 2. Preparation Conditions and MolecularCharacteristics of Ternary Graft Copolymers

samplefeed massratioa feed molar ratiob

Mn,theorc

(kg/mol)

Mn,SEC(×106g/mol) Mw/Mn

GP 1 1.1:4.0:2.2:4.0 41:2.1:0.48:5.5 79 2.7 1.22GP 2 1.0:4.0:4.5:4.0 41:2.3:1.1:6.1 104 3.5 1.20GP 3 1.1:4.0:2.5:4.0 102:5.2:1.4:13.7 202 5.2 1.23

aMass ratio between P(GMA-N3), PS−CCH, PCEMA−CCH,MPEG−CCH. b[−N3]:[PS−CCH]:[ PCEMA−CCH]:[MPEG−CCH]. cMn,theor = Mn(P(GMA-N3)) + x × n × 13500 +y × n × 31500 + z × n × 5000.

Macromolecules Article

dx.doi.org/10.1021/ma302663u | Macromolecules XXXX, XXX, XXX−XXXF

retention time, probably due to the excessive swelling of theP(GMA-N3) chains in DMF.25,45 These should not be aconcern as the SEC molecular weights were not absolute butapparent anyhow.An important result of Figure 3 was that no SEC peaks for

the precursors were observed for the reacted mixtures preparedusing the recipes shown in Table 2. This was due to the lowmolar ratios used for the polymer alkyne to azide groups usedduring the reactions. For the samples listed in Table 2, thehighest molar ratio used between MPEG−CCH, PS−CCH, and PCEMA−CCH and the azide groups of P(GMA-N3) was 23%. At these low feed ratios, all the polymers addedwere quantitatively grafted under our reaction conditions. Wealso used a feed ratio of 30% between alkyne and azide. Thefinal reaction mixture in this case showed peaks for theprecursor polymers, and the polymers were thus not fullygrafted at this high alkyne to azide feed ratio.The molecular weights of the graft copolymers should

increase from GP 1 to GP 2 and GP 3. This trend was followedby the SEC traces for these samples shown in Figure 3.However, the apparent SEC molecular weights in column 5 ofTable 2 for these samples were much higher than thetheoretical values again because the SEC values were apparent.Unlike linear polymers which are randomly coiled, the graftcopolymers were most likely stretched. Molecular weightdetermination for these samples based on PS standards shouldbe far off.

The FTIR study yielded the FTIR spectra of Figure 2b−d forP(GMA-N3)41 and the GP 1 sample before and after its residualazide groups were reacted with excess propargyl alcohol. Theazide peak at 2104 cm−1 relative to the carbonyl peak at 1680cm−1 decreased significantly for two reasons. First, some azidepeaks were consumed during the Cu-catalyzed alkyne−azidecycloaddition between polymeric backbone and polymeric sidechains. Second, PCEMA bearing two pendant carbonyl groupsper unit was grafted increasing the intensity of the peak at 1680cm−1. More interestingly, the peak 2104 cm−1 totallydisappeared after reaction of the graft copolymer with excesspropargyl alcohol, in agreement with the capping of the residualazide groups by propargyl alcohol.A NMR study yielded Figure 4, showing two 1H NMR

spectra of purified GP 1 measured in CDCl3 and DMSO-d6 andthe assignments for the observed peaks. All the protons of thegrafted MPEG, PS, and PCEMA chains were observed in thespectrum measured in CDCl3. A quantitative comparison of theintegrals at 3.66, 7.08, and 7.66 ppm yielded a molar ratio of45:12.7:1.0 for the EG, St, and CEMA repeat units,respectively. These values compared well with the expectedvalues of 46:13.1:1.0 if the numbers of grafted MPEG, PS, andPCEMA chains per P(GMA-N3)41 chain were 5.5:2.1:0.48,which were calculated from the polymer feed ratios used forgraft copolymer synthesis.Another interesting observation was the presence of the

signals at 7.9 and 5.2 ppm for the t and s protons of the triazolelinkage in the spectrum measured in DMSO-d6. These peaksprovided direct evidence for the desired click chemistry.However, these signals were not seen in CDCl3 probablybecause of the lower mobility of the triazole linkage in thissolvent. This mobility difference was possible because theP(GMA-N3) backbone is solvated in DMSO-d6 but not CDCl3.This reduced solubility of the P(GMA-N3) backbone in CDCl3could have affected the mobility of the triazole linkage.

Graft Copolymers vs Cylindrical Brushes. Graftcopolymers become cylindrical brushes if the grafted sidechains are sufficiently dense so that they repel one another andalso the backbone chain is much longer than the grafts. At agrafting density of 20%, the average spacing between two graftswas 5 monomer units or 10 C−C bonds, which had a fullystretched length of 1.26 nm. A revisit of Table 2 revealed thatthe majority of the grafts in a PGMA-g-(PS-r-PCEMA-r-MPEG) chain consisted of MPEG. To a first approximation, we

Figure 3. SEC traces of polymer precursors and ternary graftcopolymers.

Figure 4. 1H NMR spectra and peak assignments for GP 1.

Macromolecules Article

dx.doi.org/10.1021/ma302663u | Macromolecules XXXX, XXX, XXX−XXXG

considered a homograft copolymer PGMA-g-MPEG andcalculated the radius of gyration RG of a MPEG chain in theunperturbed state. The root-mean-square end-to-end distanceof a MPEG chain with a molar mass of 5000 g/mol and in theunperturbed state was calculated using a literature formula tobe 6.0 nm.50 Assuming random coil conformation, RG wascalculated to be 2.5 nm. It was evidently difficult to squeeze in aMPEG chain with a diameter of 5.0 nm into a spacing of 1.26nm. Thus, the grafts in our copolymer were crowded. Despitethis, we will continue to call our polymers graft copolymersbecause the polymer backbone was substantially shorter thanthe graft chains.Nanocapsule Preparation. A literature method using a

linear triblock copolymer for making capsules was modified andused to make the graft copolymer capsules.51 This methodinvolved first dissolving PGMA-g-(PS-r-PCEMA-r-MPEG) indichloromethane (DCM), which solubilized all three types ofgrafts. This solution was then slowly added into a stirredmixture of water and DN (A, Scheme 3), where DN solubilized

only the PS chains. After the DCM solution was fully added,the DCM to DN volume ratio in the droplets should reachbetween 5/1 and 5/2 depending on the recipe used, and thedroplet phase should solubilize both PCEMA and PS asillustrated in B of Scheme 3. Results of a previous studysuggested that the solubilized PCEMA chains would collapsefrom the DN phase after DCM was preferentially evaporatedvia gentle heating (B → C).51 We further imagined that thePCEMA chains would form a continuous membrane separatingthe water and DN phases if the PCEMA chains were longenough and present in sufficient numbers. A membrane wouldform mainly to minimize interfaces between PCEMA and thetwo liquids. Shining light on the system would photo-cross-linkPCEMA, yielding DN-filled stable nanocapsules (C→ D).52−54

GP 1, 2, and 3 were used individually to preparenanocapsules. The reported procedure did not work well forthe high-molecular-weight GP 3 because it did not dissolve wellin DCM. The preparation proceeded smoothly when either GP1 or 2 was used. At stage B, a strong scattering whitish emulsionwas obtained. The emulsion turned less scattering or whitish atstage C, suggesting the shrinkage of the original larger oildroplets. No visual changes were noticed for the samples beforeand after photolysis. However, our UV−vis spectrophotometricanalysis detected decreases at 274 nm for CEMA double-bond

absorbance.51 Typically, a CEMA double-bond conversion of60% was used to lock in the capsule structure.Our experiments suggested that the amount of DCM used

was important for proper emulsification. When 20 mg of GP 1,0.10 mL of DN, and 20 mL of water were used together with0.50 mL of DCM, a proper emulsion both before and afterDCM evaporation was obtained. However, the emulsion atstage B in Scheme 3 turned unstable when the DCM amountwas increased to 1.0 mL under otherwise identical conditions.After stirring stopped, the DCM/DN phase separated from theaqueous phase and settled at the bottom of the vial as shown inFigure 5a.

There are several possible reasons for the instability of theemulsion at the higher DCM amount of 1.0 mL. First, the usedGP 1 amount was probably insufficient to stabilize so much ofthe discrete phase consisting of DCM/DN. Second, the densityand the size of the droplets probably increased as the DCMamount increased. While DN has a density of 0.90 g/cm3 at 25°C, DCM’s density is 1.33 g/cm3, which is substantially higherthan 1.00 g/cm3 for water. Both a high density and large size forthe oil droplets favored droplet settlement. Third, an increase inDCM volume fraction in the droplets would increase thesolubility of MPEG chains in them, and the solubilization ofMPEG chains in the oil phase would decrease the hydrophilicto hydrophobic balance of GP 1 and its function as a surfactant.A proper DN amount was also important. At 20 mg of GP 1,

20 mL of water, and 0.50 mL of DCM, the use of 0.10 mL ofDN yielded a stable homogeneous emulsion after DCMevaporation. The use of 0.20 mL of DN yielded a creamy phaseconsisting probably of large capsules floating on the top of ahomogeneous emulsion as shown in Figure 5b. In this study,only capsules prepared using 20 mg of GP 1 or 2, 0.50 mL ofCH2Cl2, 0.10 mL of DN, and 20 mL of water werecharacterized.Capsules prepared from the above protocol were irradiated

by UV light to cross-link the PCEMA layer before they wereaero-sprayed onto nitrocellulose-coated grids and stained byRuO4 for TEM observations. Figures 6a and 6b show TEMimages of cross-linked capsules prepared from GP 1 and 2,respectively.The structures in Figure 6a,b could be divided into three

types. The first type consisted of a gray circle enclosed by adark ring. Another dark ring close to the center existed in thesecond type of structures. This central ring turned into an ill-defined dark dot or crease in the third type of structure.

Scheme 3. Schematic Illustration of the Preparation Processfor the Capsules

Figure 5. Photographs of emulsions prepared using 20 mg of GP 1 butdifferent amounts of DCM or DN at stage B (a) and stage C (b),respectively.

Macromolecules Article

dx.doi.org/10.1021/ma302663u | Macromolecules XXXX, XXX, XXX−XXXH

Further, the type 1 structure dominated in Figure 6b and was

only a minor product in Figure 6a.The type 1 structure corresponded to round capsules because

of the following considerations. First, the emulsification

procedure should produce capsules. Second, a gray circle with

an outer dark ring is the projection expected of a capsule. SinceRuO4 stained the PCEMA and PS chains, the dark rings musthave been projections of the standing PCEMA walls bearing thedried PS chains. These walls appeared dark because of the largepath lengths of the electron beam in them. Third, the average

Figure 6. TEM images of cross-linked capsules aero-sprayed from water immediately after their preparation using GP 1 (a) or 2 (b) as thedispersant. Shown in (c) and (d) are TEM images of the corresponding capsules after their dialysis against DMF.

Figure 7. AFM topography images of cross-linked capsules aero-sprayed from water immediately after their preparation using GP 1 (a) or 2 (b) asthe dispersant. Shown in (c) and (d) are AFM images of the corresponding capsules after their dialysis against DMF.

Macromolecules Article

dx.doi.org/10.1021/ma302663u | Macromolecules XXXX, XXX, XXX−XXXI

diameters of the structures in Figures 6a and 6b were 60 ± 10and 57 ± 12 nm, respectively. These structures were too largeto be solid because a fully stretched PS chain of 130 repeatunits would be only 33 nm long and the PS chains would beunlikely to be fully stretched in the core. A central ring was seenin the type 2 structure due to formation of a crater and thus aPCEMA cliff or wall bearing PS chains around the crater. Whenthe craters were not so well developed or when only localdimples or folds were formed, dark dots or creases were seenclose to the center of these particles or type 3 particles.We believe that the craters, folds, and creases were formed in

the type 2 and 3 particles due to the collapsing of the capsulesduring DN evaporation before TEM analysis. This belief agreeswith the observation that more type 1 structures were seen inFigure 6b than in Figure 6a because the capsules made from GP2 containing more PCEMA chains were more robust andcollapsed less than those made from GP 1. However, we cannotrule out the possible existence of the type 2 and 3 particles inthe emulsion already because the sample has not been analyzedby an in situ technique such as cryo-TEM. Despite this, we canconclude with confidence that all types of structures observedwere hollow consisting of either round or deformed capsules.We further note that the outer rings of the gray objects in

Figure 6a were not uniformly dark or were not as welldeveloped as in Figure 6b. This difference might again be dueto the higher number of PCEMA and PS chains in GP 2 than inGP 1. From Figure 6b, we determined an average ring thicknessof 6 ± 1 nm. Since this layer contained a contribution from thedried PS chains as well, the PCEMA wall should be thinnerthan 6 nm.Solutions of the cross-linked capsules were also aero-sprayed

onto freshly cleaved mica surfaces and dried under high vacuumto remove DN before AFM observations. Figures 7a and 7bshow the AFM topography images of the sprayed capsulesprepared from GP 1 and 2, respectively. Round and bowl-shaped particles coexisted in these two images. The averagediameters of the particles in Figures 7a and 7b were 84 ± 10and 80 ± 6 nm, respectively. Also, more bowl-shaped particlesexisted in Figure 7a than in Figure 7b.The AFM observations supported the TEM results. The

AFM diameters were larger than the TEM diameters because oftwo reasons. First, AFM probed the size of the whole particlesincluding the MPEG layer not seen by TEM. Second, the AFMsize contained a contribution from the finite size of the AFMtip.Capsule Formation Mechanism. The emulsion droplets

stabilized by GP 1 at different stages of capsule formation werestudied by DLS. Figure 8 shows the droplet size distributionsthus determined by DLS at stages B, C, and D of Scheme 3.From these distributions we also obtained the averagehydrodynamic diameters Dh of 315, 198, and 196 nm,respectively. The Dh decrease from 315 to 198 nm fromstage B to C corresponded to a 75% decrease in thehydrodynamic volume of the emulsion droplets after DCMevaporation. At a feed volume ratio of 5/1 for DCM/DN,DCM should account for 83% of the droplet volume beforeDN evaporation. These two values, 75% and 83%, agreed wellgiven that the PEG corona thickness contributed to thedetected Dh value, and this thickness might not change muchbefore and after droplet shrinkage. Thus, the data suggestedthat DCM evaporation did not change the number of theemulsion droplets but decreased the size of the nanocapsules.This reasonable size change from B to C and the insignificant

Dh change from stage C to D supported the capsule formationmechanism proposed in Scheme 3.The DLS diameters were substantially larger than the AFM

and TEM diameters for the capsules for several reasons. First,DLS probed the solvated particles and the other techniquesprobed the dried particles. Since the PCEMA layer was verythin, these particles should shrink substantially when solventevaporated. Second, the MPEG chains should assume a morestretched conformation in the solvated than in the dried state.Third, the DLS diameter was a z-average, and the otherdiameters were the number-average values. The z-average valueemphasized contributions from the larger capsules.

Permanent Capsules. The capsular structure got locked inonly if the PCEMA grafts of the different copolymer chainsoverlapped and photo-cross-linked properly. The 1,4-cyclo-addition reaction of two double bonds of different CEMA unitshas been used extensively by the Liu group to lock in blockcopolymer micellar structures49 and solid structures52 andshould work equally well here if the PCEMA grafts of differentcopolymer chains overlapped. To ensure the latter, the PCEMAchains were designed to be longer than the PGMA backbone.Also, the number of PCEMA chains per graft copolymer wasincreased from 0.48 to 1.1 from GP 1 to GP 2.The cross-linked capsules of GP 1 were dialyzed against

DMF changed several times over 2 days, and DLS was thenused to analyze the capsule sample. The sample had a Dh valueof 198 nm, which was essentially the same as 196 nm, Dh forthe sample before dialysis against DMF. Thus, the structuralintegrity was retained even for capsules prepared from GP 1.TEM and AFM images were also obtained of the cross-linked

capsules after their dialysis against DMF and aero-spraying. Theintegrity of the particles was clearly retained in Figures 6c,d and7c,d despite the possible extraction of some of the PGMA-g-(PS-r-MPEG) chains particularly from the GP 1 capsules. Thus,the capsules were permanent structures.

IV. CONCLUSIONSATRP has been used to prepare PGMA, PS−CCH, andPHEMA−CCH that was cinnamated to PCEMA−CCH.Reacting PGMA with sodium azide yielded P(GMA-N3).Further, MPEG has been end-functionalized to MPEG−CCH. Grafting PS−CCH, MPEG−CCH, and PCEMA−CCH to P(GMA-N3) via click chemistry yielded ternary

Figure 8. DLS size emulsion droplets at different stages nanocapsulepreparation using GP 1 as the dispersant.

Macromolecules Article

dx.doi.org/10.1021/ma302663u | Macromolecules XXXX, XXX, XXX−XXXJ

graft copolymers PGMA-g-(PS-r-PCEMA-r-MPEG). The graft-ing reactions were facile and achieved in a one pot. At molarratios ≤23% for the polymer terminal alkyne groups to azidegroups, these polymer chains with repeat unit numbers between110 and 130 units were grafted quantitatively. While thismodular or graft-onto approach has been used to prepare onetype of ternary graft copolymers, it should also be useful in thepreparation of other types of ternary graft copolymers or graftcopolymers with even more types of grafts.Copolymers prepared with a polymer grafting densities of

∼23% and P(GMA-N3)41 readily solubilized in CH2Cl2. Such asolution was added into a stirred water/DN mixture, and thepolymer was concentrated at the DN/water interface tostabilize the DN/CH2Cl2 emulsion droplets. Evaporation ofCH2Cl2 from the droplets yielded DN droplets that most likelycontained PS chains stretching from the copolymer at the DN/water interface. These droplets were stabilized by the MPEGchains solubilized in the aqueous phase. The PCEMA chainsprobably formed a membrane or wall separating the DN andwater phases. Photolysis of this system yielded permanentcapsules. This general method should be useful for preparingcapsules from other graft copolymers or capsules with loadsother than DN in the core phase, and the resultant capsulesmay find applications in controlled release applications.

■ AUTHOR INFORMATIONCorresponding Author*E-mail: hjw@gic.ac.cn (J.H.); gliu@chem.queensu.ca (G.L.).NotesThe authors declare no competing financial interest.

■ ACKNOWLEDGMENTSWe thank the National Natural Science Foundation of China(No. 20474068, 51173204, 51203191), the OutstandingOverseas Chinese Scholars Funds of the Chinese Academy ofSciences, Guangdong Natural Science Foundation(S2012010009063), and the Leading Talents Program ofGuangdong Province for providing financial support.

■ REFERENCES(1) Sheiko, S. S.; Sumerlin, B. S.; Matyjaszewski, K. Prog. Polym. Sci.2008, 33, 759.(2) Chen, Y. M. Macromolecules 2012, 45, 2619.(3) Ishizu, K. Polym. J. 2004, 36, 775.(4) Neugebauer, D.; Zhang, Y.; Pakula, T.; Matyjaszewski, K. Polymer2003, 44, 6863.(5) Neugebauer, D.; Zhang, Y.; Pakula, T.; Matyjaszewski, K.Macromolecules 2005, 38, 8687.(6) Neugebauer, D.; Theis, M.; Pakula, T.; Wegner, G.;Matyjaszewski, K. Macromolecules 2006, 39, 584.(7) Hans, M.; Keul, H.; Heise, A.; Moeller, M. Macromolecules 2007,40, 8872.(8) Hans, M.; Xiao, Y.; Keul, H.; Heise, A.; Moeller, M. Macromol.Chem. Phys. 2009, 210, 736.(9) Bo, Z. S.; Rabe, J. P.; Schluter, A. D. Angew. Chem.., Int. Ed. 1999,38, 2370.(10) Ishizu, K.; Satoh, J.; Tsubaki, K. J. Mater. Sci. Lett. 2001, 20,2253.(11) Tsubaki, K.; Kobayashi, H.; Sato, J.; Ishizu, K. J. Colloid InterfaceSci. 2001, 241, 275.(12) Zhang, B.; Zhang, S. J.; Okrasa, L.; Pakula, T.; Stephan, T.;Schmidt, M. Polymer 2004, 45, 4009.(13) Rathgeber, S.; Pakula, T.; Wilk, A.; Matyjaszewski, K.; Beers, K.L. J. Chem. Phys. 2005, 122, 124904.(14) Meier, W. Chem. Soc. Rev. 2000, 29, 295.

(15) Mora-Huertas, C. E.; Fessi, H.; Elaissari, A. Int. J. Pharm. 2010,385, 113.(16) De Koker, S.; Hoogenboom, R.; De Geest, B. G. Chem. Soc. Rev.2012, 41, 2867.(17) Esser-Kahn, A. P.; Odom, S. A.; Sottos, N. R.; White, S. R.;Moore, J. S. Macromolecules 2011, 44, 5539.(18) Zhang, M. F.; Muller, A. H. E. J. Polym. Sci., Polym. Chem. 2005,43, 3461.(19) Vazaios, A.; Hadjichristidis, N. J. Polym. Sci., Polym. Chem. 2005,43, 1038.(20) Heroguez, V.; Breunig, S.; Gnanou, Y.; Fontanille, M.Macromolecules 1996, 29, 4459.(21) Ishizu, K.; Shen, X. X. Polymer 1999, 40, 3251.(22) Xia, Y.; Olsen, B. D.; Kornfield, J. A.; Grubbs, R. H. J. Am. Chem.Soc. 2009, 131, 18525.(23) Cheng, C.; Qi, K.; Germack, D. S.; Khoshdel, E.; Wooley, K. L.Adv. Mater. 2007, 19, 2830.(24) Cheng, C.; Qi, K.; Khoshdel, E.; Wooley, K. L. J. Am. Chem. Soc.2006, 128, 6808.(25) Sun, J. P.; Hu, J. W.; Liu, G. J.; Xiao, D. S.; He, G. P.; Lu, R. F. J.Polym. Sci., Polym. Chem. 2011, 49, 1282.(26) Zhao, P.; Yan, Y. C.; Feng, X. Q.; Liu, L. X.; Wang, C.; Chen, Y.M. Polymer 2012, 53, 1992.(27) Li, Y. K.; Zou, J.; Das, B. P.; Tsianou, M.; Cheng, C.Macromolecules 2012, 45, 4623.(28) Ishizu, K.; Sawada, N.; Satoh, J.; Sogabe, A. J. Mater. Sci. Lett.2003, 22, 1219.(29) Yin, J.; Ge, Z. S.; Liu, H.; Liu, S. Y. J. Polym. Sci., Polym. Chem.2009, 47, 2608.(30) Han, D. H.; Tong, X.; Zhao, Y. Macromolecules 2011, 44, 5531.(31) Lian, X. M.; Wu, D. X.; Song, X. H.; Zhao, H. Y. Macromolecules2010, 43, 7434.(32) Djalali, R.; Li, S. Y.; Schmidt, M. Macromolecules 2002, 35, 4282.(33) Mullner, M.; Lunkenbein, T.; Breu, J.; Caruso, F.; Muller, A. H.E. Chem. Mater. 2012, 24, 1802.(34) Kang, E. H.; Lee, I. H.; Choi, T. L. ACS Macro Lett. 2012, 1,1098.(35) Zhao, P.; Liu, L. X.; Feng, X. Q.; Wang, C.; Shuai, X. T.; Chen,Y. M. Macromol. Rapid Commun. 2012, 33, 1351.(36) Huang, K.; Jacobs, A.; Rzayev, J. Biomacromolecules 2011, 12,2327.(37) Huang, K.; Canterbury, D. P.; Rzayev, J. Chem. Commun. 2010,46, 6326.(38) Huang, K.; Rzayev, J. J. Am. Chem. Soc. 2011, 133, 16726.(39) Miyake, G. M.; Weitekamp, R. A.; Piunova, V. A.; Grubbs, R. H.J. Am. Chem. Soc. 2012, 134, 14249.(40) Sveinbjornsson, B. R.; Weitekamp, R. A.; Miyake, G. M.; Xia, Y.;Atwater, H. A.; Grubbs, R. H. Proc. Natl. Acad. Sci. U. S. A. 2012, 109,14332.(41) Beers, K. L.; Boo, S.; Gaynor, S. G.; Matyjaszewski, K.Macromolecules 1999, 32, 5772.(42) Opsteen, J. A.; van Hest, J. C. M. Chem. Commun. 2005, 57.(43) Ding, J. F.; Liu, G. J. Macromolecules 1999, 32, 8413.(44) Canamero, P. F.; de la Fuente, J. L.; Madruga, E. L.; Fernandez-Garcia, M. Macromol. Chem. Phys. 2004, 205, 2221.(45) Tsarevsky, N. V.; Bencherif, S. A.; Matyjaszewski, K.Macromolecules 2007, 40, 4439.(46) Mansfeld, U.; Pietsch, C.; Hoogenboom, R.; Becer, C. R.;Schubert, U. S. Polym. Chem. 2010, 1, 1560.(47) Fan, X. S.; Wang, G. W.; Huang, J. L. J. Polym. Sci., Polym. Chem.2011, 49, 1361.(48) Gao, H. F.; Matyjaszewski, K. J. Am. Chem. Soc. 2007, 129, 6633.(49) Henselwood, F.; Liu, G. J. Macromolecules 1997, 30, 488.(50) Ozdemir, C.; Guner, A. J. Appl. Polym. Sci. 2006, 101, 203.(51) Zheng, R. H.; Liu, G. J. Macromolecules 2007, 40, 5116.(52) Zhou, R. L.; Zhou, N.; Yan, X. H.; Liu, G. J. Chem. J. Chin. Univ.2002, 23, 158.(53) Hoppenbrouwers, E.; Li, Z.; Liu, G. J. Macromolecules 2003, 36,876.

Macromolecules Article

dx.doi.org/10.1021/ma302663u | Macromolecules XXXX, XXX, XXX−XXXK

(54) Zhou, Z. H.; Liu, G. J.; Hong, L. Z. Biomacromolecules 2011, 12,813.

Macromolecules Article

dx.doi.org/10.1021/ma302663u | Macromolecules XXXX, XXX, XXX−XXXL