crystallizable diluent-templated polyacrylonitrile foams for macroporous carbon monoliths

8
Crystallizable diluent-templated polyacrylonitrile foams for macroporous carbon monoliths Qing-Yun Wu, Ling-Shu Wan, Zhi-Kang Xu * MOE Key Laboratory of Macromolecular Synthesis and Functionalization, Department of Polymer Science and Engineering, Zhejiang University, Hangzhou 310027, China article info Article history: Received 6 March 2012 Received in revised form 6 November 2012 Accepted 7 November 2012 Available online 10 November 2012 Keywords: Crystallization Carbonization Templates abstract Polyacrylonitrile (PAN) foams with different pore structures were prepared for the fabrication of mac- roporous carbon monoliths. The foams were prepared through thermally induced phase separation (TIPS) method using dimethyl sulfone (DMSO2) as a crystallizable diluent. Honeycomb-like porous foam is obtained from PAN/DMSO2 mixture containing about 5 wt.% PAN, and those with channel-like pores are resulted from the mixtures with 10e40 wt.% PAN. However, they only have few mesopores and the porosity is as low as 30e47% for the foams prepared from those mixtures containing 50e60 wt.% PAN. Real-time observation with polarized optical microscopy reveals that the channel-like structure stems from the spherulitic orientation of DMSO2 crystals in the polymer matrix. Taking into account this morphology, DMSO2 crystals are capable of acting as in situ formed templates, which subsequently enable to shape the nal pore structure of PAN foams. Macroporous carbon monoliths with honeycomb- or channel-like pores were constructed from PAN foams by oxidative stabilization and carbonization. Their graphitic structure and specic surface areas were analyzed by wide-angle X-ray diffraction and BrunauereEmmetteTeller measurement. This TIPS method using crystallizable diluent provides a new route to control the porous structure of PAN foams for carbon materials. Ó 2012 Elsevier Ltd. All rights reserved. 1. Introduction Carbon materials with unique mechanical properties and chemical inertness have attracted considerable attention for elec- trochemical double-layer capacitors [1e3], catalyst supports [4], electrode materials in Li-ion batteries [5,6], water purication lters [7], and chemical adsorbents [8,9]. Despite various precur- sors, polyacrylonitrile (PAN) is the most suitable one for carbon bers with excellent performance because of its high melting point, considerable carbon yield, and relatively cheap price [10e13]. It is accepted that fast pyrolysis of PAN generates a graphitic structure by thermal stabilization and subsequent carbonization. Recently, intensive studies have focused on the design and construction of porous structures for carbon materials with controllable pore number, pore size and functional surface in an effective fashion [6,14]. Various methods have been adopted to prepare porous carbon materials. Physical and chemical activation enables to create pores by subjecting the carbon to steam at high temperature, but is complex, time-consuming, and difcult to control the pore struc- ture. Templating method is capable of constructing ordered porous carbons with well-dened structures depending on the framework of the used templates [15e19]. Inorganic materials (e.g., silica or silica particles, zeolites, and anodic alumina oxide (AAO) membrane) are often used as hard-templates. For example, meso- porous carbons were prepared with silica as the template [18,19]. However, for a polymer precursor with high viscosity, it is difcult to completely diffuse and ll into the space of templates [20]. Unexpected defects will be left in the resulted porous carbon materials. Therefore the templating method is more suitable for thin lms or bers than bulky materials. Given these, another method has been developed based on phase separation and then carbonization from block copolymer or polymer blend with a sacricial component [21e24]. This component serves as soft- template, allowing for replication of the phase separated struc- ture into carbon materials. Microdomains of the sacricial component can be selectively removed during pyrolysis. Accord- ingly, the morphology and scale of the microdomains must be bound to conne the pore structures. Thomassin et al. [25] devel- oped a templating compression molding process using a poly- acrylonitrile-poly(vinyl acetate) block copolymer to prepare mesoporous carbon bers. As the sacricial block, poly(vinyl * Corresponding author. Fax: þ86 571 8795 1773. E-mail addresses: [email protected], [email protected], xuzk@ ipsm.zju.edu.cn (Z.-K. Xu). Contents lists available at SciVerse ScienceDirect Polymer journal homepage: www.elsevier.com/locate/polymer 0032-3861/$ e see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.polymer.2012.11.025 Polymer 54 (2013) 284e291

Upload: zhi-kang

Post on 06-Dec-2016

225 views

Category:

Documents


3 download

TRANSCRIPT

Page 1: Crystallizable diluent-templated polyacrylonitrile foams for macroporous carbon monoliths

at SciVerse ScienceDirect

Polymer 54 (2013) 284e291

Contents lists available

Polymer

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

Crystallizable diluent-templated polyacrylonitrile foams for macroporous carbonmonoliths

Qing-Yun Wu, Ling-Shu Wan, Zhi-Kang Xu*

MOE Key Laboratory of Macromolecular Synthesis and Functionalization, Department of Polymer Science and Engineering, Zhejiang University, Hangzhou 310027, China

a r t i c l e i n f o

Article history:Received 6 March 2012Received in revised form6 November 2012Accepted 7 November 2012Available online 10 November 2012

Keywords:CrystallizationCarbonizationTemplates

* Corresponding author. Fax: þ86 571 8795 1773.E-mail addresses: [email protected], xuzk

ipsm.zju.edu.cn (Z.-K. Xu).

0032-3861/$ e see front matter � 2012 Elsevier Ltd.http://dx.doi.org/10.1016/j.polymer.2012.11.025

a b s t r a c t

Polyacrylonitrile (PAN) foams with different pore structures were prepared for the fabrication of mac-roporous carbon monoliths. The foams were prepared through thermally induced phase separation(TIPS) method using dimethyl sulfone (DMSO2) as a crystallizable diluent. Honeycomb-like porous foamis obtained from PAN/DMSO2 mixture containing about 5 wt.% PAN, and those with channel-like poresare resulted from the mixtures with 10e40 wt.% PAN. However, they only have few mesopores and theporosity is as low as 30e47% for the foams prepared from those mixtures containing 50e60 wt.% PAN.Real-time observation with polarized optical microscopy reveals that the channel-like structure stemsfrom the spherulitic orientation of DMSO2 crystals in the polymer matrix. Taking into account thismorphology, DMSO2 crystals are capable of acting as in situ formed templates, which subsequentlyenable to shape the final pore structure of PAN foams. Macroporous carbon monoliths with honeycomb-or channel-like pores were constructed from PAN foams by oxidative stabilization and carbonization.Their graphitic structure and specific surface areas were analyzed by wide-angle X-ray diffraction andBrunauereEmmetteTeller measurement. This TIPS method using crystallizable diluent provides a newroute to control the porous structure of PAN foams for carbon materials.

� 2012 Elsevier Ltd. All rights reserved.

1. Introduction

Carbon materials with unique mechanical properties andchemical inertness have attracted considerable attention for elec-trochemical double-layer capacitors [1e3], catalyst supports [4],electrode materials in Li-ion batteries [5,6], water purificationfilters [7], and chemical adsorbents [8,9]. Despite various precur-sors, polyacrylonitrile (PAN) is the most suitable one for carbonfibers with excellent performance because of its highmelting point,considerable carbon yield, and relatively cheap price [10e13]. It isaccepted that fast pyrolysis of PAN generates a graphitic structureby thermal stabilization and subsequent carbonization. Recently,intensive studies have focused on the design and construction ofporous structures for carbon materials with controllable porenumber, pore size and functional surface in an effective fashion[6,14].

Various methods have been adopted to prepare porous carbonmaterials. Physical and chemical activation enables to create poresby subjecting the carbon to steam at high temperature, but is

@zjuem.zju.edu.cn, xuzk@

All rights reserved.

complex, time-consuming, and difficult to control the pore struc-ture. Templating method is capable of constructing ordered porouscarbons with well-defined structures depending on the frameworkof the used templates [15e19]. Inorganic materials (e.g., silica orsilica particles, zeolites, and anodic alumina oxide (AAO)membrane) are often used as hard-templates. For example, meso-porous carbons were prepared with silica as the template [18,19].However, for a polymer precursor with high viscosity, it is difficultto completely diffuse and fill into the space of templates [20].Unexpected defects will be left in the resulted porous carbonmaterials. Therefore the templating method is more suitable forthin films or fibers than bulky materials. Given these, anothermethod has been developed based on phase separation and thencarbonization from block copolymer or polymer blend witha sacrificial component [21e24]. This component serves as soft-template, allowing for replication of the phase separated struc-ture into carbon materials. Microdomains of the sacrificialcomponent can be selectively removed during pyrolysis. Accord-ingly, the morphology and scale of the microdomains must bebound to confine the pore structures. Thomassin et al. [25] devel-oped a templating compression molding process using a poly-acrylonitrile-poly(vinyl acetate) block copolymer to preparemesoporous carbon fibers. As the sacrificial block, poly(vinyl

Page 2: Crystallizable diluent-templated polyacrylonitrile foams for macroporous carbon monoliths

Q.-Y. Wu et al. / Polymer 54 (2013) 284e291 285

acetate) favored the penetration of the copolymer into AAOtemplate and created nanopores in the final carbon fibers bypyrolysis at 600 �C. This technique combined the advantage of thetemplating method and the microphase separation of blockcopolymer. Nevertheless, PAN-based block copolymers are limitedand their phase separation is not well understood.

It should be promising to design the pore structures of carbonmonoliths a priori through the preparation of porous polymerprecursors [26e29]. Thermally induced phase separation (TIPS) isuseful to fabricate polymer foams and separation membranes [30e36]. Several phase separation processes can be thermodynamicallydriven by temperature in polymerediluent systems. After removingthe diluent, cellular, lacy, bi-continuous, spherulitic, or metallicstructures are usually obtained by varying the preparation condi-tions such as polymer concentration and/or by modulating thepolymerediluent interactions (quantified as interaction param-eter). Generally, if a system possesses strong polymerediluentinteraction (small interaction parameter), solideliquid phaseseparation takes place via polymer crystallization; otherwise, thepolymer/diluent mixture undergoes liquideliquid phase separationand shows an upper-critical temperature behavior upon cooling[30,31]. For a specific polymer, the polymerediluent interaction canalso be modulated by adding a non-solvent [37]. Olivier et al. [38]reported PAN foams for carbon materials using maleic anhydrideas the diluent. This study proposed an ingenious way to design theporous framework. However, it was pointed out that maleicanhydride is only suitable for PAN with a very low concentrationbecause of the low dissolubility [39].

We report here a method to control the pore structures of PANfoams using crystallizable dimethyl sulfone (DMSO2) as the diluent.As a polar solvent with high melting point, DMSO2 can dissolvePAN in a wide range up to 70 wt.%. PAN foams with honeycomb- orchannel-like pores were prepared by varying the polymerconcentration, the background of which was particularly studiedfrom the perspective of DMSO2 crystallization in the polymermatrix. These PAN foams were then used as precursors for mac-roporous carbon monoliths by the oxidative stabilization andcarbonization processes.

2. Experimental

2.1. Materials

Polyacrylonitrile (PAN, Mn ¼ 115,000, Mw ¼ 251,000,MWD ¼ 2.19) was kindly supplied by Anqing Petroleum ChemicalCorporation of China. It is a copolymer of acrylonitrile (AN) andvinyl acetate (VA), and the molar ratio of AN:VA is 24:1. The poly-mer was crushed finely and dried at 60 �C in an oven before use.Dimethyl sulfone (DMSO2, 99% purity) was purchased fromDakangChemicals Co., Ltd of China. De-ionized water was chosen as theextractant. Ethanol and hexane (AR grade) were commercially ob-tained from Sinopharm Chemical Reagent Co., Ltd. All chemicalswere used without further purification.

2.2. Preparation of PAN foams

PAN and DMSO2 were premixed with various compositions andsealed in glass tubes (8 mm in diameter and 1 mm thick). PANconcentration ranged from 5 to 60 wt.%. These tubes wereimmersed into an oil bath at 160 �C until the mixtures transformedinto homogeneous solutions. After quenching the solutions into anoil bath at 30 �C, the obtained solid mixtures were immersed intode-ionized water with gentle vibration for 4 days at 30 �C to extractDMSO2 completely. To prevent collapse of pores during drying,water in the pores was exchanged by ethanol and hexane for 1 day

in consequence. The resulted foams were dried in vacuum oven at30 �C overnight.

2.3. Preparation of porous carbon monoliths

Porous carbon monoliths from PAN precursors are usuallyprepared through two steps: (a) oxidative stabilization in air ata temperature between 200 and 300 �C, and (b) final carbonizationin an inert atmosphere at temperatures between 600 and 1000 �C.In this work, the pristine PAN foams were placed in a quartz tubehoused within an electric tube furnace, and heated to 250 �C ata rate of 1 �C/min and maintained at 250 �C for 1 h. Subsequently,the oxidized samples were conducted under the protection ofnitrogen, heated to 800 �C at a rate of 2 �C/min and then held for 1 hat 800 �C for carbonization.

2.4. Determination of phase diagram

Differential scanning calorimeter (DSC, Pyris-1, PerkineElmer,USA) was used to determine the melting and crystallizationtemperatures of PAN/DMSO2mixtures. After hermetically sealed inan aluminumDSC pan, a sample of 5e8mg was heated from 0 �C to160 �C at 20 �C/min, maintained at 160 �C for 3 min to eliminatethermal history, cooled to 10 �C at 5 �C/min, and then reheated to160 �C at 5 �C/min.

2.5. Wide-angle X-ray diffraction (WAXD)

WAXD was carried out on a Rigaku D/Max-2550PC X-raydiffractometer (Panalytical, Netherlands). The radial scans ata voltage of 40 kV and a current of 40 mA using a Cu:Ka radiationwere employed on PAN/DMSO2 mixtures, PAN foams, and porouscarbon monoliths. Data were collected at 0.0167� interval withcounting for 10 s at each step.

2.6. Scanning electron microscopy (SEM)

FESEM (Sirion-100, FEI, USA) was used to observe themorphologies of the porous materials. Samples were frozen andfractured in liquid nitrogen to obtain tidy cross-section. Aftersputtered with gold using an ion sputter JFC-1100, samples wereimaged by FESEM at an acceleration voltage of 25 kV.

2.7. Polarized optical microscopy (POM)

Polarized optical micrographswere recorded on anOlympus BX-5 POM equipped with a Linkam hot stage (CSS450, Linkam, UK) anda temperature controller (Linksys32, Linkam, UK). PAN/DMSO2mixtures used here and the experimental parameters were as sameas those in DSC measurements. Each sample was sliced into smallpieces and placed between a pair of glasses, the edges ofwhichweresealed with Teflon tape to prevent losses caused by evaporation.Photographs were taken by a CCD camera or a digital camera.

2.8. Pore size and porosity measurements

Dilatometerwas employed tomeasure the apparent volume (Va)of the porous samples, which was used to calculate the apparentdensity (da) and the porosity (r). A dilatometer filled with mercurywas immersed in awater bath at 30 �C and recorded the height (h0)of the mercury in the capillary with a diameter (D). Thereafter,a piece of porous materials with weight (m0) was put into theconical beaker and record the height (h1). The apparent volume, theapparent density, and the porosity can be defined as equations (1)e(3), respectively. Table 1 lists the density of PAN bulk (dPAN).

Page 3: Crystallizable diluent-templated polyacrylonitrile foams for macroporous carbon monoliths

Table 1Typical properties of PAN and DMSO2.

Parameters Units PAN DMSO2

Molecular weight g/mol 115,000 (Mn);251,000 (Mw)

94

Degree of polymerization 2169 N/AMolar volume cm3/mol 46.5a 75Density g/cm3 1.14 1.17Heat of fusion J/mol 6230a,b 18,300b

Melting temperature K (�C) 593 (320)c 381.9 (108.9)Crystallization temperature K (�C) N/A 365.2 (92.2)Solubility parameter MPa1/2 31.5c 29.7b

a Acrylonitrile unit.b Representative value ([62]).c Representative value ([63]).

Fig. 1. PAN concentration-dependent melting point (,) and crystallization tempera-ture (-) of PAN/DMSO2 mixtures. The solid curve indicates the theoretical meltingpoint of DMSO2 in the mixtures calculated from equation (4). The arrow denotes PAN/DMSO2 mixture was cooled from liquid phase (a) to solid phase (b).

Q.-Y. Wu et al. / Polymer 54 (2013) 284e291286

Va ¼ 14pD2ðh1 � h0Þ (1)

da ¼ m0

Va(2)

r ¼ dPAN � dadPAN

(3)

The porosity and pore size distribution of PAN foams weredirectly measured by mercury intrusion porosimetry (MIP, IV9500,Micromeritics, USA), the data of which were analyzed by Porowin-32 software. Two pressure ranges were applied: low-pressurerange (0.5e50 psi) to measure pores >10 mm, and high-pressurerange (20e34,000 psi) to measure pore size <10 mm. Both intru-sion and extrusion runs were conducted on the sample introducedin a penetrometer having a stem volume 0.39 cm3 and first kept ina low-pressure chamber. After completion of the experiment thepenetrometer was transferred to the high-pressure chamber, andthe pressure was applied by hydraulic means. Each data was anaverage result of at least three parallel measurements.

2.9. BrunauereEmmetteTeller (BET) analysis

BET method qualifiedly confirmed the specific surface area ofthe obtained carbon monoliths. Nitrogen adsorption at 77 K wasperformed on an adsorption analyzer (ASIC-2, QuantachromeInstruments, USA) after degassing each sample at 473 K for 2 h ina vacuum.

3. Results and discussion

3.1. Phase behavior of the PAN/DMSO2 binary system

PAN/DMSO2 mixtures can form homogeneous solutions in widecomposition when they are heated above the melting point ofDMSO2, because the melted DMSO2 is a good solvent for PAN. Thesolutions solidify as cooled to a relatively low temperature. Themelting point and crystallization temperature were obtained forthe PAN/DMSO2 binary system based on the experimental resultsof DSC. As shown in Fig. 1, the melting point of DMSO2 decreases by18 �C from neat DMSO2 to the mixture with 60 wt.% PAN. Itsuggests dipoleedipole interaction exists between C^N groups ofPAN and O]S groups of DMSO2 [40]. However, a miscible liquidcannot be formed when the polymer concentration is beyond70 wt.%.

Generally, equilibrium phase diagram can be calculated for thepolymer/diluent binary system based on the FloryeHuggins theoryof melting point depression [41,42]. Herein, the theoretical melting

temperature depression of DMSO2 was calculated using equation(4):

1Tm; 1

¼"1þ Rbf2

2DH1

#�1"1

T0m; 1

� RDH1

��1� 1

N

�f2 þ lnð1�f2Þ

�#

(4)

The subscripts 1 and 2 refer to the diluent and the polymer,respectively. T0m is the melting temperature of pure component andTm represents the melting temperature of the mixture. R, 4, DH, andV are the gas constant, the volume fraction, the molar enthalpy offusion, and the molar volume fraction, respectively. N is the degreeof polymerization, and b is a constant which is calculated from thesolubility parameters of diluent (d1) and polymer (d2) along withthe molar volume of diluent (V1), using the equationb ¼ (d1 � d2)2V1/R. The parameters used for calculation are given inTable 1. It shows that the experimental values are in line with thetheoretical results (Fig. 1).

Upon cooling, PAN/DMSO2 solutions solidify accompanied withDMSO2 crystallization in the total miscible region. The crystalliza-tion temperature of DMSO2 decreases with the polymer concen-tration, and it is 20e65 �C lower than the corresponding meltingpoint. Both PAN and DMSO2 crystallize in the mixtures with 5e60 wt.% PAN as confirmed by WAXD analysis, which will be dis-cussed in the following text. Accordingly, the PAN/DMSO2 binarysystem goes through solidesolid TIPS. However, no exotherm refersto the crystallization of DMSO2 in the mixture containing 60 wt.%PAN (see Supporting information, Fig. S1). It may be due to the lowcontent of DMSO2 and the effect of inhibition of viscous solution.

3.2. Structure and properties of PAN foams

PAN/DMSO2mixtures with polymer concentration of 5e60wt.%were then used to prepare PAN foams. Different pores were formedin the polymer matrix via TIPS and extraction of DMSO2. Fig. 2presents the cross-sections of PAN foams obtained from PAN/DMSO2 mixtures with various compositions. The foam fabricatedfrom 5 wt.% PAN exhibits honeycomb-like morphology with largepores. As the PAN concentration increases to 10 wt.%, channel-likestructure appears, and the pores well align along a direction. Thiswell-patterned channel-like morphology remains in the foams

Page 4: Crystallizable diluent-templated polyacrylonitrile foams for macroporous carbon monoliths

Fig. 2. SEM images for the cross-sections of PAN foams with different polymer concentrations: (a) 5 wt.%, (b) 10 wt.%, (c) 15 wt.%, (d) 20 wt.%, (e) 25 wt.%, (f) 30 wt.%, (g) 40 wt.%, (h)50 wt.%, and (i) 60 wt.%. The scale bar in inset of (a) is 100 mm.

Fig. 3. Dependence of the porosity of PAN foam on PAN concentration.

Q.-Y. Wu et al. / Polymer 54 (2013) 284e291 287

derived from the mixtures with 15e40 wt.% PAN. Moreover, thechannel-like pores overlap to each other accompanied by the nar-rowing of pores as the polymer concentration increases. However,pores are not clearly discernible when the PAN concentration hasbeen further increased to 50e60 wt.%. From the above results, wecan obtain two kinds of ordered macroporous structures, i.e.,honeycomb- and channel-like morphologies. Experimentally, thepore structures of PAN foam are significantly controlled by thepolymer concentration. Similar results have been reported in ice-templated materials [43]. However, the pore size is not onlyaffected by the polymer concentration but also by the coolingrate (See Supporting information, Fig. S2). We found that smallcooling rate results in large pore size, which may be ascribed tothe deep degree of phase separation under slight supercooling.This phenomenon is in agreement with the results reported inliterature [44].

Fig. 3 shows that the porosity calculated from the apparentdensity decreases from 95% to 30% when the PAN concentrationincreases from 5 wt.% to 60 wt.%. Moreover, the porosity measuredby MIP is slightly lower than that determined by a dilatometer,especially for samples from low PAN concentration. This differencemay be due to the collapse of pores by the intrusion of mercury.Fig. 4 compares the MIP histogram (both high pressure and lowpressure) of PAN foams prepared from those mixtures with 5e40 wt.% PAN. The foams fabricated from PAN/DMSO2 mixturescontaining 5e20 wt.% PAN have large amount of macropores withdiameter of 45e66 mm; and the mixtures with 25 wt.% and 30 wt.%PAN lead to macropores with diameter of 5e20 mm. When the PANconcentration increases to 40 wt.%, the foams show few macro-poreswith pore size larger than 0.1 mm. Apart from themacropores,mesopores with a maximum pore size diameter of 28 nm exist inthe foam prepared from the mixture containing 30 wt.% PAN.Similarly, most other samples also have a few mesopores.

To further elucidate effects of the polymer concentration on thepore structures, we monitored the cooling process of PAN/DMSO2mixtures by real-time POM. Neither crystallization nor liquideliquid phase separation takes place at the initial stage. As coolingto the crystallization temperature of DMSO2 in the mixture, thesolution solidifies in a few seconds. Fig. 5 shows POM images ofsamples with different PAN concentrations. PAN/DMSO2 mixturecontaining 5 wt.% PAN displays mosaic-like morphology, whereasthat with 10 wt.% PAN exhibits a certain degree of orientation fromside to interior of the sample. When the PAN concentration is in therange of 15e40 wt.%, a spherulitic morphology with a giantMaltese-cross presents in each sample, indicating a radial

Page 5: Crystallizable diluent-templated polyacrylonitrile foams for macroporous carbon monoliths

Fig. 4. Pore size distribution of PAN foams fabricated from the mixtures with differentPAN concentrations: (a) 5 wt.%, (b) 10 wt.%, (c) 15 wt.%, (d) 20 wt.%, (e) 25 wt.%, (f)30 wt.%, and (g) 40 wt.%. The inset is the magnification of the range indicated by therectangle.

Q.-Y. Wu et al. / Polymer 54 (2013) 284e291288

alignment of crystals. The spherulitic morphology occupies thewhole vision, and its centre locates randomly. Sometimes, ring-bands present around the centre. The spherulitic morphology alsoexists in the mixture containing 50 wt.% PAN, but the size is smalland its appearance is stochastic. No such morphology appears inthe mixture with 60 wt.% PAN.

The spherulitic morphology of PAN/DMSO2 mixtures accountsfor the pore structure of PAN foams prepared from different

Fig. 5. POM images of PAN/DMSO2 mixtures with different PAN concentrations: (a) 5 wt.%, ((i) 60 wt.%. Samples were cooled from 160 to 30 �C at a rate of 5 �C/min. Scale bars are 10

compositions as mentioned above (Fig. 2). The honeycomb-likepores originate from the isotropic morphology, whereas thechannel-like pores result from the anisotropic morphology ofspherulite. SEM images further reveal the spherulitic morphologyof PAN foams fabricated from the mixtures with 20, 25 and 30 wt.%PAN (Fig. 6).

WAXD was used to confirm the crystallization of both compo-nents in typical PAN/DMSO2 mixtures. Fig. 7(a) shows that eachpattern of the mixture contains diffraction peaks at 2q ¼ 16.4�,20.4�, 24.3�, 25.4�, 29.4�, and 33.1�, corresponding to (011), (111),(200), (120), (211), and (220) planes, respectively. They are as sameas that of neat DMSO2. It indicates that DMSO2 crystallizes aftermixing with PAN, even in the case with high PAN concentration(60 wt.%). After removing DMSO2, the characteristic peaks ofDMSO2 are absent, whereas a peak at 17� arises in the patterns ofPAN foams (Fig. 7(b)). This peak is well matched with that of PANcrystals having a hexagonal polymorph [45]. Combining with theresults of DSC and POM,we can conclude that both PAN and DMSO2crystallize in their mixtures with 5e60 wt.% PAN.

As mentioned above, the PAN/DMSO2 binary system undergoessolidesolid TIPS, and both the polymer and the diluent crystallizein the process. Obviously, DMSO2 crystals act as templates to shapethe pores of PAN foams. In fact, this phenomenon is well known inseveral polymerediluent binary systems, such as polypropyleneepentaerythrityl tetrabromide [46], polypropyleneehexamethyl-benzene [47], high density polyethylenee1,2,4,5-tertrachloro-benzene [48], linear low-density polyethyleneehexamethylb-enzene [49], and poly(L-lactic acid)eparahydroxybenzoic acid[50]. In all these systems, the shape and distribution of diluentcrystals depend on the mixture composition and the polymerediluent interactions. However, polyethylene, polypropylene, andpoly(L-lactic acid) undergo a typical crystallization through chain

b) 10 wt.%, (c) 15 wt.%, (d) 20 wt.%, (e) 25 wt.%, (f) 30 wt.%, (g) 40 wt.%, (h) 50 wt.%, and0 mm.

Page 6: Crystallizable diluent-templated polyacrylonitrile foams for macroporous carbon monoliths

Fig. 6. SEM images for the surfaces of PAN foams fabricated from PAN/DMSO2 mixtures with different PAN concentrations: (a) 20 wt.%, (b) 25 wt.%, and (c) 30 wt.%.

Q.-Y. Wu et al. / Polymer 54 (2013) 284e291 289

folding [51e56], whereas the crystallization of PAN has moredependence on the intra- or inter-molecular interaction amongC^N groups [57]. In our cases, both PAN and DMSO2 crystallize intheir mixtures with awide range of 5e60 wt.% PAN. This means thesolidification process is governed by the mutual diffusion of bothcomponents across the solideliquid interface [46]. Experimentally,DMSO2 forms isolated polyhedral crystals when PAN is entirelyabsent. However, a certain amount of PAN raises the systemviscosity (see Supporting information, Fig. S3), which may affects

Fig. 7. WAXD patterns of PAN/DMSO2 mixtures with different PAN concentrationsbefore (a) and after (b) removing DMSO2.

the crystal growth of DMSO2 from their mixtures. Previous workhas shown that small molecule organic crystals can promotedendrite formation or even can grow into spherulites in the pres-ence of polymers by increasing the viscosity and creating a diffu-sion-limited crystal growth regime [58e60]. In our cases, themixtures with PAN concentration of 10e50 wt.% can providea suitable viscosity for spherulite formation. If the mixture with toolow PAN concentration, the diluent molecules easily aggregate intolarge crystals; as for the mixture with high PAN concentration, thediluent molecules would be fixed at an adjacent nucleus prioragglomeration.

3.3. Pyrolysis of PAN foams for porous carbon monoliths

Based on the above studies, we synthesized PAN foams withhoneycomb- or channel-like pores by the crystallizable diluent-templating method. These foams were then adopted to prepareporous carbon monoliths through stabilization at 250 �C followedby carbonization at 800 �C. It is important to trace chemical andphysical changes of the foams during the stabilization and thecarbonization procedures. The chemical structures of the pristine,stabilized and carbonized samples were determined by FTIR (seeSupporting information, Fig. S4). Moreover, WAXD pattern indi-cates the main component of the porous carbon monolith is ingraphitic structure, and the corresponding distance between twosuperimposed planes is 3.71 Å (see Supporting information,Fig. S5). The spacing is slightly larger than that in puregraphite (3.35 Å), which may be due to the disordered structure ofcarbon [61].

The PAN foam turns from white to dark brown under stabili-zation and then changes to black after carbonization. Its weight lossand volume shrink are slight in the first step but dramatic in thesecond step. This is reasonable that the density of carbon (1.5e2.0 g/cm3) is larger than that of PAN (1.14 g/cm3). The carbonyield is above 60% and increases slightly with the PAN concentra-tion for foam synthesis (63.9, 64.8, and 67% for the foams derivedfrom PAN/DMSO2 mixtures with 5, 10, and 20 wt.% PAN, respec-tively). As illustrated in Fig. 8(a), the pores collapse seriously whenthe PAN foam with honeycomb-like structure is carbonized. Thefibrils in the channel-like structure stack much more closelycompared with the pristine foam fabricated from the mixturecontaining 10 wt.% PAN (Fig. 8(b)). In this case, the fibrils remain tobe oriented in a distance, although there is a certain degree offracture. More importantly, the spherulitic morphology still main-tains after the pyrolysis procedure (Fig. 8(c)). The lamellar poresalmost do not collapse in the carbon monoliths, which may be dueto the channel-like architecture in the foams with high mechanicalproperty. On the other hand, the pore walls are relatively smooth,and it seems there are no micropores. BET measurements showthat macroporous carbon monoliths with specific surface area of0.025e0.263 m2/g can be prepared from PAN foams synthesizedwith 5e20 wt.% PAN in the mixtures.

Page 7: Crystallizable diluent-templated polyacrylonitrile foams for macroporous carbon monoliths

Fig. 8. SEM images for the cross-sections of porous carbon monoliths prepared from PAN foams from different PAN concentrations: (a) 5 wt.%, (b) 10 wt.%, and (c) 20 wt.%.(Up, �500; down, �2000).

Q.-Y. Wu et al. / Polymer 54 (2013) 284e291290

4. Conclusion

PAN foams with controllable pore structures can be prepared bythe solidesolid TIPS of PAN/DMSO2 mixtures, in which the diluentcrystals act as templates to shape the pores. Foams with honey-comb- and/or channel-like pores depend on the polymer concen-tration in the mixtures. These pore structures originate from thespecific crystallization of DMSO2 during the TIPS process. Forexample, the channel-like structures stem from the spheruliticorientation of DMSO2 crystals in the polymer matrix. The resultantPAN foams are useful as precursors for pyrolyzing into carbonmonoliths with different macroporous structures. This workprovides an approach to controllable PAN foams and porous carbonmaterials based on TIPS assisted with crystallizable diluent-templating. The macroporous carbon monoliths are promising invarious fields such as redoxase immobilization, filtration, adsorp-tion, and catalyst support.

Acknowledgments

We are grateful to the financial support by the National NaturalScience Foundation of China (Grant no. 21174124). The authors alsothank Dr. Xiao-Jun Huang for his helpful discussion during the earlystages of this work.

Appendix A. Supplementary data

Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.polymer.2012.11.025.

References

[1] Kim C, Ngoc BTN, Yang KS, Kojima M, Kim YA, Kim YJ, et al. AdvancedMaterials 2007;19(17):2341e6.

[2] Kim B-H, Yang KS, Woo H-G. Electrochemistry Communications 2011;13(10):1042e6.

[3] Ra EJ, Raymundo-Piñero E, Lee YH, Béguin F. Carbon 2009;47(13):2984e92.[4] Rikkinen E, Santasalo-Aarnio A, Airaksinen S, Borghei M, Viitanen V, Sainio J,

et al. The Journal of Physical Chemistry C 2011;115(46):23067e73.[5] Wei L, Chen Q, Kong X. Journal of the American Ceramic Society 2011;94(9):

3078e83.[6] Ji L, Lin Z, Medford AJ, Zhang X. Carbon 2009;47(14):3346e54.[7] Kim E-S, Deng B. Journal of Membrane Science 2011;375(1e2):46e54.[8] Shim WG, Kim C, Lee JW, Yun JJ, Jeong YI, Moon H, et al. Journal of Applied

Polymer Science 2006;102(3):2454e62.

[9] Grzyb B, Albiniak A, Broniek E, Furdin G, Marêché JF, Bégin D. Microporous andMesoporous Materials 2009;118(1e3):163e8.

[10] Bahl OP, Manocha LM. Carbon 1974;12(4):417e23.[11] Rahaman MSA, Ismail AF, Mustafa A. Polymer Degradation and Stability 2007;

92(8):1421e32.[12] Wangxi Z, Jie L, Gang W. Carbon 2003;41(14):2805e12.[13] Deng W, Lobovsky A, Iacono ST, Wu T, Tomar N, Budy SM, et al. Polymer 2011;

52(3):622e8.[14] Xiao B, Thomas KM. Langmuir 2005;21(9):3892e902.[15] Knox JH, Kaur B,Millward GR. Journal of Chromatography A 1986;352(0):3e25.[16] Kruk M, Dufour B, Celer EB, Kowalewski T, Jaroniec M, Matyjaszewski K. The

Journal of Physical Chemistry B 2005;109(19):9216e25.[17] Li Z, Jaroniec M. Carbon 2001;39(13):2080e2.[18] Ryoo R, Joo SH, Jun S. The Journal of Physical Chemistry B 1999;103(37):

7743e6.[19] Lee J, Yoon S, Hyeon TM, Oh S, Bum Kim K. Chemical Communications 1999;

21:2177e8.[20] Chen JT, Shin K, Leiston-Belanger JM, Zhang M, Russell TP. Advanced Func-

tional Materials 2006;16(11):1476e80.[21] Peng M, Li D, Shen L, Chen Y, Zheng Q, Wang H. Langmuir 2006;22(22):

9368e74.[22] Kim C, Jeong YI, Ngoc BTN, Yang KS, Kojima M, Kim YA, et al. Small 2007;3(1):

91e5.[23] Ho R-M, Wang T-C, Lin C-C, Yu T-L. Macromolecules 2007;40(8):2814e21.[24] Stefik M, Sai H, Sauer K, Gruner SM, DiSalvo FJ, Wiesner U. Macromolecules

2009;42(17):6682e7.[25] Thomassin J-M, Debuigne A, Jérôme C, Detrembleur C. Polymer 2010;51(14):

2965e71.[26] Cohen N, Silverstein MS. Polymer 2011;52(2):282e7.[27] Okada K, Nandi M, Maruyama J, Oka T, Tsujimoto T, Kondoh K, et al. Chemical

Communications 2011;47(26):7422e4.[28] Gitli T, Silverstein MS. Polymer 2011;52(1):107e15.[29] David D, Silverstein MS. Journal of Polymer Science Part A: Polymer Chemistry

2009;47(21):5806e14.[30] Lloyd DR, Kinzer KE, Tseng HS. Journal of Membrane Science 1990;52:239e61.[31] Lloyd DR, Kim S, Kinzer KE. Journal of Membrane Science 1991;64:1e11.[32] Liu S, Zhou C, Yu W. Journal of Membrane Science 2011;379(1e2):268e78.[33] Shibutani T, Kitaura T, Ohmukai Y, Maruyama T, Nakatsuka S, Watabe T, et al.

Journal of Membrane Science 2011;376(1e2):102e9.[34] Shang MX, Matsuyama H, Teramoto M, Lloyd DR, Kubota N. Polymer 2003;

44(24):7441e7.[35] Matsuyama H, Takida Y, Maki T, Teramoto M. Polymer 2002;43(19):5243e8.[36] Vanegas ME, Quijada R, Serafini D. Polymer 2009;50(9):2081e6.[37] Wu Q-Y, Wan L-S, Xu Z-K. Journal of Membrane Science 2012;409e410(0):

355e64.[38] Olivier BJ, Lagasse RR, Schaefer DW, Barnes JD, Long GG. Macromolecules

1996;29(27):8615e21.[39] Sadlik B. The role of viscosity in banded-spherulitic growth. Physics. Vol.

Master of Science: Simon Fraser University: 2004.[40] Wu Q-Y, Chen X-N, Wan L-S, and Xu Z-K. Journal of Physical Chemistry B, in

press.[41] Flory PJ. Principles of polymer chemistry. Ithaca: Cornell University Press;

1953.[42] Burghardt WR. Macromolecules 1989;22(5):2482e6.[43] Gutiérrez MC, Ferrer ML, del Monte F. Chemistry of Materials 2008;20(3):

634e48.

Page 8: Crystallizable diluent-templated polyacrylonitrile foams for macroporous carbon monoliths

Q.-Y. Wu et al. / Polymer 54 (2013) 284e291 291

[44] Zhou J, Zhang H, Wang H, Du Q. Journal of Membrane Science 2009;343(1e2):104e9.

[45] Bashir Z. Journal of Polymer Science Part B: Polymer Physics 1994;32(6):1115e28.

[46] Smith P, Pennings AJ. Journal of Polymer Science: Polymer Physics Edition1977;15(3):523e40.

[47] Alwattari AA, Lloyd DR. Polymer 1994;35(13):2710e5.[48] Smith P, Pennings AJ. Polymer 1974;15(7):413e9.[49] Yoon J, Lesser AJ, McCarthy TJ. Macromolecules 2009;42(22):8827e34.[50] Zwiers RJM, Gogolewski S, Pennings AJ. Polymer 1983;24(2):167e74.[51] Li J, Li W, Cheng H, Zhang L, Li Y, Han CC. Polymer 2012;53(12):2315e9.[52] Mileva D, Androsch R, Zhuravlev E, Schick C, Wunderlich B. Polymer 2012;

53(2):277e82.[53] Stoclet G, Seguela R, Vanmansart C, Rochas C, Lefebvre JM. Polymer 2012;

53(2):519e28.

[54] Tsuji H, Tashiro K, Bouapao L, Hanesaka M. Polymer 2012;53(3):747e54.[55] Yoon JG, Zin W-C, Kim JH. Polymer 2012;53(13):2744e50.[56] Polec IA, Fujiwara T, Kanaya T, Deng C. Polymer (0).[57] Henrici-Olivé G, Olivé S. Molecular interactions and macroscopic properties of

polyacrylonitrile and model substances chemistry, vol. 32. Berlin/Heidelberg:Springer; 1979. p. 123e152.

[58] Ul�cinas A, Butler MF, Heppenstall-Butler M, Singleton S, Miles MJ. Journal ofCrystal Growth 2007;307(2):378e85.

[59] Zawko SA, Schmidt CE. Acta Biomaterialia 2010;6(7):2415e21.[60] Shtukenberg AG, Punin YO, Gunn E, Kahr B. Chemical Reviews 2011.[61] Setnescu R, Jipa S, Setnescu T, Kappel W, Kobayashi S, Osawa Z. Carbon 1999;

37(1):1e6.[62] Lide DR. CRC handbook of chemistry and physics. CRC Press (Taylor & Francis);

2010.[63] Brandrup J, Immergut EH, Grulke EA. Polymer handbook. Wiley; 1999.