Journal of Colloid and Interface Science 436 (2014) 77–82
Contents lists available at ScienceDirect
Journal of Colloid and Interface Science www.elsevier.com/locate/jcis
Hollow S-doped carbon spheres from spherical CT/PEDOT composite particles and their CO2 sorption properties Jin-Yeon Hong, Seong Huh ⇑ Department of Chemistry and Protein Research Center for Bio-Industry, Hankuk University of Foreign Studies, Yongin 449-791, Republic of Korea
a r t i c l e
i n f o
Article history: Received 15 July 2014 Accepted 2 September 2014 Available online 16 September 2014 Keywords: Porous carbon S-doping Conducting polymers Carbonization Carbon dioxide capture
a b s t r a c t Chemically functionalizable shape-controlled poly(3,4-ethylenedioxythiophene) (PEDOT)-derived conducting copolymers, C1(C4)-CT/PEDOT/PSS-20APS and C1(C4)-CT/PEDOT/PSS-10APS, were prepared through oxidative polymerization of 3,4-ethylenedioxythiophene (EDOT) and 3-thiophenecarboxylic acid (C1-CT) or 4-(3-thienyl)butyric acid (C4-CT) in the presence of acid-labile mesoporous ZnO/Zn(OH)2 hard template. The mesoporous ZnO/Zn(OH)2 hard template could be removed by mild acid etching. The morphology of these polymeric microparticles was dependent on the concentration of ammonium persulfate (APS, (NH4)2S2O8) catalyst and the type of CT monomers. Hollow capsular C1-CT/PEDOT/PSS-20APS copolymer spheres with a large surface opening were obtained when the amount of oxidizing agent APS was 20 mmol under the same experimental conditions. Because C1(C4)-CT/PEDOT/PSS-xAPS copolymers contain S-rich moieties in the polymer backbone, they are suited for the preparation of S-doped carbonaceous materials. Therefore, we carbonized C1-CT/PEDOT/PSS-20APS at several different temperatures in a high-purity nitrogen atmosphere to easily prepare hollow S-doped carbon spheres (HSCSs). The level of S-dopants in carbon spheres was strongly dependent on the carbonization temperature. Lower carbonization temperature led to a higher content of S-dopants but lower BET surface area. These carbon spheres were further analyzed by TEM, SEM, PXRD, and XPS. Gas sorption analyses were also performed to study gas sorption with different amount of S-dopants. Ó 2014 Elsevier Inc. All rights reserved.
1. Introduction Nanoporous carbon materials are important in modern materials chemistry because of their utility in a wide variety of applications. There are numerous reports dealing with porous carbon materials with distinct porosity and pore functionality. The effectiveness of using porous carbons for various advanced applications such as H2 storage [1], CO2 capture [2], encapsulation of catalytically active metallic nanoparticles [3], electrode materials in energy-storage devices [4], and supercapacitors [5] is mainly dependent on the porosity and functionality of the carbon materials. The surface area, pore size, and pore volume are the primary factors governing the efficiency of gas sorption and encapsulation of functional nanoparticles. Among these porous carbonaceous materials, heteroatomdoped carbons such as N-doped carbon materials having Lewis basic sites in the framework often possess enhanced surface functionalities especially in many applications related to CO2 capture [6], metal-free oxygen reduction reaction (ORR) [7], photocatalysis ⇑ Corresponding author. Fax: +82 31 330 4566. E-mail address: [email protected] (S. Huh). http://dx.doi.org/10.1016/j.jcis.2014.09.003 0021-9797/Ó 2014 Elsevier Inc. All rights reserved.
[8], and catalytic organic transformations [9]. In general, the preparation of N-doped carbon materials inevitably requires N-rich organic polymer precursors for carbonization in inert atmosphere. Many research groups reported on novel N-rich organic-based precursors for the carbonization. For instances, we developed N-doped porous carbons having good CO2 sorption abilities [10]. The carbonization precursors were prepared by the azide-alkyne 1,3dipolar Huisgen cycloaddition reaction, also known as the Click reaction, between 1,4-bis(azidomethyl)benzene and 1,3,5-ethynylbenzene. N-doped carbon monolith with Brunauer–Emmett–Teller (BET) surface area of 467 m2 g 1 was also obtained by carbonizing polypyrrole precursor at 500 °C by Hao and co-workers [11]. Sevilla and co-workers prepared highly N-doped carbon material by using polypyrrole as a carbon source and KOH activation of the resulting carbon [12]. Mechanically stable N-doped porous carbons were also prepared from poly(benzoxazine-co-resol) precursor and subsequently tested for CO2 separation from a flue gas [13]. Most of these N-doped porous carbons are being actively investigated to fine-tune the content of N atoms and porosity for enhanced functionalities. In contrast, relatively few studies have described for the preparation and characterization of S-doped porous carbon materials
78
J.-Y. Hong, S. Huh / Journal of Colloid and Interface Science 436 (2014) 77–82
[14–21]. The CO2 sorption properties of S-doped porous carbons have only very recently been explored [16,19–21]. Both the Schmidt and the Böttger-Hiller groups independently prepared S-doped microporous carbon materials using thiophene-containing organic polymer precursors [14,15]. S-doped microporous carbon materials might have comparable properties as N-doped carbon materials in gas sorption applications. Xia and co-workers recently reported on S-doped microporous carbons prepared through nanocasting of zeolite EMC-2 [16]. The resulting S-doped microporous carbons exhibited superior adsorption abilities for both H2 and CO2. Notably, the low surface coverage heat of CO2 adsorption had a high value of 59 kJ mol 1. Moreover, hierarchically structured S-doped carbon materials were demonstrated as good cathode materials for rechargeable Li–S batteries [17]. We envision that the utilization of S-doped porous carbon materials will be widely expanded. In this study, we present a new simple method to prepare shape-controlled S-containing carbon precursors from poly(3,4ethylenedioxythiophene) (PEDOT)-derived conducting polymers which were prepared in the presence of poly(4-ammonium styrenesulfonic acid) (NH4PSS) and mesoporous ZnO/Zn(OH)2 composite hard template. One of the resulting carbon precursors, C1-CT/PEDOT/PSS-20APS, was carbonized at several different temperatures to produce shape controlled hollow S-doped carbon spheres (HSCSs). The doping level of S atoms into HSCSs upon change of carbonization temperature was investigated by elemental analysis and X-ray photoelectron spectroscopy (XPS). Gas sorption analyses using N2 and CO2 were also performed to study the textural properties and CO2 sorption abilities of HSCSs. The relationship between the textural properties of HSCSs and the S-contents for CO2 sorption will be discussed.
2. Experimental 2.1. Synthesis of mesoporous ZnO/Zn(OH)2 Mesoporous ZnO/Zn(OH)2 composite material was prepared according to the method previously reported [22]. 2.2. Synthesis of C1-CT/PEDOT/PSS-20APS and C1-CT/PEDOT/PSS-10APS These products were prepared by modifying a method in the literature, employing increased amount of ammonium persulfate polymerization catalyst [23]. 3,4-Ethylenedioxythiophene (0.947 mL, 8.87 mmol, Sigma–Aldrich), 3-thiophenecarboxylic acid (0.032 g, 0.25 mmol, Sigma–Aldrich), and a 30 wt% aqueous solution of poly(4-ammonium styrenesulfonic acid) (2 mL, Sigma–Aldrich) were dissolved in 20 mL of distilled water. Ammonium persulfate (4.592 g, 20 mmol for C1-CT/PEDOT/PSS-20APS; 2.296 g, 10 mmol for C1-CT/PEDOT/PSS-10APS) was added after 5 min and the reaction mixture was stirred constantly (rpm = 750). After 30 min of reaction, 500 mg of mesoporous ZnO/Zn(OH)2 was added. The reaction mixture was further stirred at room temperature for one week. The solid product was separated by centrifugation and washed with distilled water and methanol. In order to remove the ZnO/Zn(OH)2 template, the solids were etched with 10 mL of 1.0 M HCl for 1 h with stirring. Finally, the resulting solids were washed with water and methanol and dried at 80 °C for 2 h. 2.3. Synthesis of C4-CT/PEDOT/PSS-20APS and C4-CT/PEDOT/PSS10APS These materials were prepared using the same method for C1-CT/PEDOT/PSS-20APS and C1-CT/PEDOT/PSS-10APS (above)
except 4-(3-thienyl)butyric acid (0.043 g, 0.25 mmol, Rieke Fine Chemicals) was used instead of 3-thiophenecarboxylic acid. 2.4. Physical measurements Powder X-ray diffraction patterns were recorded on a Rigaku MiniFlex (30 kV, 15 mA). FE-SEM images were recorded on an FEI Nova 200 NanoSEM (accelerating voltage = 10 kV). The water suspension of the sample was dropped and dried on a Cu grid supported by a carbon film for TEM measurement, for which a JEOL JEM-3000F (accelerating voltage = 300 kV) equipped with an EDAX for the EDS spectrum was used. XPS data were collected on a K-Alpha X-ray photoelectron spectrometer (Thermo VG, UK). The X-ray source was a monochromated Al Ka line (hm = 1486.6 eV), and the X-ray power was 12 kV at 3 mA. The pass energy of the survey spectrum was fixed at 100 eV with a step size of 1 eV while the pass energy for the high-resolution spectrum was fixed at 50 eV with a step size of 0.1 eV. The sampling area was 400 lm in diameter. The binding energies were referenced to 284.8 eV (C 1s peak for C–C bonds). The cryogenic volumetric N2 adsorption– desorption analysis was performed on a Belsorp-miniII at 77 K (BEL Japan). The samples were evacuated at 423 K under high vacuum for 2 h. Low pressure CO2 adsorption measurements were also performed on a Belsorp-miniII at 196 K (2-propanol/dry ice bath) and 273 K (ice bath). 3. Results and discussion We recently reported on hierarchically structured mesoporous PEDOT-derived conducting polymers, C1-CT/PEDOT/PSS-5APS and C4-CT/PEDOT/PSS-5APS, containing 3-thiophenecarboxylic acid (C1-CT) or 4-(3-thienyl)butyric acid (C4-CT) in the polymer backbone [23]. The amount of ammonium persulfate (APS, (NH4)2S2O8) catalyst was 5 mmol in each case. Interestingly, the porous materials were large particles with microscale dimensions having cheese-like wall structures. The porous walls contained small polymeric fibers. As a result, the BET surface areas of C1-CT/ PEDOT/PSS-5APS and C4-CT/PEDOT/PSS-5APS were 60.1 m2 g 1 and 44.5 m2 g 1, respectively. In order to accelerate the polymerization reaction in the same system, the concentration of the catalyst, APS, was increased twofold or fourfold compared with the previous condition for C1(C4)-CT/PEDOT/PSS-5APS [23]. Surprisingly, the simple increase of concentration of APS catalyst not only accelerated the polymerization rate, but also led to discrete smaller solid particles with distinct particle morphology (Fig. 1). Four different products were prepared in this study by using 20 mmol or 10 mmol of APS in the presence of C1-CT or C4-CT along with EDOT. C1-CT/PEDOT/ PSS-20APS is a sample obtained from 20 mmol APS with C1-CT while C1-CT/PEDOT/PSS-10APS is a sample from 10 mmol APS with C1-CT. The particle morphology dramatically varied depending on the ratio of APS and the type of CT monomers. As shown in Fig. 1a and b, the type of CT monomers induced quite different particle morphologies. The C1-CT/PEDOT/PSS-20APS sample had spherical particles with a large surface hole while C4-CT/PEDOT/PSS-20APS displayed much smaller particles with less featured shape. Both C1-CT/PEDOT/PSS-10APS and C4-CT/PEDOT/PSS-10APS indicated similar particle dimensions compared with C1(C4)-CT/PEDOT/ PSS-20APS but their particle morphologies were also less featured. Among the four samples, C1-CT/PEDOT/PSS-20APS was chosen as a carbonization precursor due to its uniformity in particle shape and dimension. Furthermore, most particles in the C1-CT/PEDOT/ PSS-20APS had hollow, spherical geometry which is ideal for the generation of hollow capsular carbon materials. C1-CT/PEDOT/ PSS-20APS was carbonized at 600 °C, 700 °C, 800 °C, and 900 °C
J.-Y. Hong, S. Huh / Journal of Colloid and Interface Science 436 (2014) 77–82
79
Fig. 1. FE-SEM images of the nanostructured samples on the same magnifications (20,000). (a) C1-CT/PEDOT/PSS-20APS, (b) C4-CT/PEDOT/PSS-20APS, (c) C1-CT/PEDOT/ PSS-10APS, and (d) C4-CT/PEDOT/PSS-10APS.
to give HSCS-600, HSCS-700, HSCS-800, HSCS-900, respectively. Transmission electron microscopy (TEM) investigation of HSCSs confirmed that the majority of hollow carbon precursors were successfully converted into HSCSs (Fig. 2). As shown in Fig. 3, the low magnification TEM image of the representative sample, HSCS-700, clearly indicates that most of the particles have hollow capsular geometry and their sizes are similar to each other. The diameters of particles are about 250–500 nm. The powder X-ray diffraction patterns of the HSCS materials are shown in Fig. 4. All HSCS samples exhibited two very broad diffraction peaks centered at around 24° and 44°, which indicated the disordered graphitic nature of the carbons [24]. The corresponding d-spacing values are 0.37 nm and 0.21 nm, respectively. Very interestingly, upon increase of the carbonization temperature from 600 °C to 900 °C, the diffraction peaks corresponding to the wurtzite-type ZnS were clearly revealed. The complete indexing results of wurtzite-type ZnS (JCPDS card No. 36-1450) in HSCS-900 is shown in Fig. 4d [25]. The source of Zn could be from the incompletely removed ZnO/Zn(OH)2 hard template after acid etching. Therefore, ZnS was formed during the carbonization of C1-CT/ PEDOT/PSS-20APS. Only HSCS-900 showed well crystallized wurtzite-type ZnS. As a result, HSCS-900 can be considered to contain nanoparticulate semiconducting ZnS in S-doped carbon structure. The S-contents of HSCS samples were estimated from elemental analysis (Table 1). Upon increase of the carbonization temperature from 600 °C to 900 °C, the content of S atoms gradually decreased from 21.6% for HSCS-600 to 8.71% for HSCS-900. However, it must be considered that some of S atoms were not directly incorporated in the carbon network, but contained as a form of ZnS nanoparticles. Nevertheless, very limited amounts of ZnS were detected on XPS surface analysis based on Zn 2p orbital signals (data not shown). All HSCS samples exhibited C-bound S 2p orbital doublet
at around 163.7 eV and 165.0 eV in XPS spectra (Fig. 5) [14]. These signals were previously assigned as sulfide bridge (–C–S–C–) species [18]. There were relatively small amounts of several oxidized forms of S atom which gave rise to signals at 166.5 eV, 167.8 eV, and 169.0 eV. The atomic percentages of oxidized S species among S-dopants estimated by XPS were 11% (HSCS-600), 15% (HSCS700), 20% (HSCS-800), and 16% (HSCS-900). The intensities corresponding to S 2p orbital signals gradually decrease upon increase of carbonization temperature, indicating lower S-dopant content at higher temperature. In contrast to S-dopant contents, the BET surface areas were proportional to the carbonization temperature. HSCS-900 exhibited the highest BET surface area of 315 m2 g 1 with pore volume of 0.21 cm3 g 1. Compared with HSCS-900, HSCS-600 showed very low levels of BET surface area and pore volume. The pore volume for HSCS-600 was just 0.082 cm3 g 1. N2 adsorption–desorption isotherms (Fig. 6a) clearly demonstrates that both HSCS-800 and HSCS-900 have better N2 sorption abilities than both HSCS-600 and HSCS-700. The shapes of isotherms for HSCS-800 and HSCS-900 are type I [26]. The heteroatom substituted carbon materials are thought to have better CO2 sorption abilities than pure carbonaceous materials due to the enhanced interactions between quadrupolar CO2 molecules and heteroatoms in the network. S-doped carbons, as well as N-doped carbons, have better CO2 sorption characteristics due to the Lewis base sites than conventional carbon materials [16]. The CO2 uptake abilities of the four HSCS samples were volumetrically measured at 196 K (Fig. 6b). To our surprise, despite the large differences in BET surface areas and pore volumes, all HSCS samples exhibited a very similar range of CO2 uptake values at 196 K: 90.2–98.4 cm3 g 1 (4.0–4.4 mmol g 1). Considering the large differences in BET surface areas and pore volumes of each HSCS sample, these results are quite astonishing. One plausible
80
J.-Y. Hong, S. Huh / Journal of Colloid and Interface Science 436 (2014) 77–82
Fig. 2. TEM images of HSCS samples. (a) HSCS-600, (b) HSCS-700, (c) HSCS-800, and (d) HSCS-900.
Fig. 4. PXRD patterns of the samples. (a) HSCS-600, (b) HSCS-700, (c) HSCS-800, and (d) HSCS-900. The wurtzite-type ZnS phase was indexed in (d). Fig. 3. Low magnification TEM image of HSCS-700.
account of these unexpected behaviors may be the role of S-dopant in the carbon frameworks. HSCS-600, having the smallest BET surface area, actually contains the highest content of S-dopant. As a result, the highest S-dopant content in HSCS-600 sufficiently offset the lowest BET surface area and pore volume to lead to similar levels of CO2 sorption compared with HSCS-800 and HSCS-900. The isosteric heats of adsorption (Qst) for all HSCS samples were
estimated by using the Clausius–Clapeyron equation based on the adsorption data measured at 196 K and 273 K (Fig. 7). The values of low surface coverage Qst are increasing in the order of HSCS-900 > HSCS-800 > HSCS-600 > HSCS-700. Therefore, despite the lowest level of S-content, HSCS-900, exhibited the highest Qst and uptake capacity for CO2. The range of Qst values, 23.1– 32.8 kJ mol 1, is comparable to the commonly observed values for N-doped porous carbons (18.9–36.0 kJ mol 1) [12,19].
81
J.-Y. Hong, S. Huh / Journal of Colloid and Interface Science 436 (2014) 77–82 Table 1 Contents of S-dopants, textural properties, and CO2 sorption data of HSCS materials. Samples
S content (%)
BET surface area (m2 g
HSCS-600 HSCS-700 HSCS-800 HSCS-900
21.6 17.0 8.92 8.71
117 161 280 315
1
)
Pore volume (cm3 g 0.082 0.093 0.24 0.21
1
)
CO2 uptake at 196 K (cm3 g 90.2 98.4 90.7 94.1
1
)
Heats of CO2 adsorption (kJ mol
1
)
24.5 23.1 27.8 32.8
Fig. 5. High-resolution XPS spectra of S 2p orbital region for HSCSs. (a) HSCS-600, (b) HSCS-700, (c) HSCS-800, and (d) HSCS-900.
Fig. 7. Isosteric heats of CO2 adsorption for HSCS samples upon increase of CO2 uptake amount.
Currently, the detailed role of S-dopant in the carbon network for CO2 sorption application has not been as understood as N-doped carbon materials. Xia and co-workers reported relatively high heat of CO2 adsorption for S-doped microporous carbons derived from EMC-2 zeolite hard template [16]. They concluded that despite the high heat of adsorption the magnitude of textural properties (surface areas and pore volumes) of S-doped microporous carbons was a primary governing factor for enhanced CO2 sorption. Kim and co-workers pointed out that the content of oxidized S species as well as the textural properties of S-doped microporous carbon was important for enhancing the CO2 sorption [20]. Bandosz and co-workers considered both S atoms incorporated in aromatic rings and oxidized forms of S atom, such as sulfonic acids, sulfoxides, and sulfones, to be determining factors for efficient interactions between the carbon network and CO2 gaseous molecules [21]. Since our HSCS materials are mainly composed of sulfide bridge (–C–S–C–) types of S-dopants based on XPS analysis,
our results suggest that the overall content of various forms of Sdopant is also an important factor not only to increase the heat of adsorption compared with S-free carbons, but also to enhance the uptake capacity for CO2. 4. Conclusion We have developed a simple preparation method for submicron-scale functionalizable PEDOT-derived hollow spherical polymeric particles, which could be carbonized into hollow S-doped carbon spheres. The uniformly shaped hollow S-doped carbon spheres exhibited a temperature dependence on the S contents. The content of S-dopants ranged from 8.71% to 21.6% based on elemental analysis. The higher the carbonization temperature, the lower the content of S-dopants. In contrast, the BET surface areas of the S-doped carbons are directly proportional to the carbonization temperature. The carbon samples carbonized above at 800 °C
Fig. 6. (a) N2 adsorption–desorption isotherms measured at 77 K and (b) CO2 adsorption–desorption isotherms measured at 196 K. Solid symbols and open symbols represent adsorption and desorption isotherms, respectively.
82
J.-Y. Hong, S. Huh / Journal of Colloid and Interface Science 436 (2014) 77–82
were microporous materials judged from BET surface areas and N2 sorption isotherms. The hollow S-doped carbon spheres were good sorbents for CO2 gas at cryogenic conditions. All hollow S-doped carbon spheres, which were carbonized at 600–900 °C, exhibited a wide range of BET surface areas and S-dopant contents. Nevertheless, all four samples showed a very similar level of CO2 sorption abilities despite their differences in BET surface areas and the content of S-dopants. Importantly, the surface area and the overall content of S-dopants played a significant role the ability to take up CO2 gas. Acknowledgment This work was supported by the Hankuk University of Foreign Studies Research Fund of 2014. References [1] H. Nishihara, T. Kyotani, Adv. Mater. 24 (2012) 4473. [2] M.S. Shafeeyan, W.M.A.W. Daud, A. Houshmand, A. Shamiri, J. Anal. Appl. Pyrol. 89 (2010) 143. [3] S.H. Joo, S.J. Choi, I. Oh, J. Kwak, J. Liu, O. Terasaki, R. Ryoo, Nature 412 (2001) 169. [4] L.L. Zhang, X.S. Zhao, Chem. Soc. Rev. 38 (2009) 2520. [5] M. Inagaki, H. Konno, O. Tanaike, J. Power Sources 195 (2010) 7880. [6] N.P. Wickramaratne, J. Xu, M. Wang, L. Zhu, L. Dai, M. Jaroniec, Chem. Mater. 26 (2014) 2820.
[7] N. Li, C. Hu, L. Ren, M. Cao, Cryst. Eng. Commun. 15 (2013) 8504. [8] M. Shalom, S. Inal, C. Fettkenhauer, D. Neher, M. Antonietti, J. Am. Chem. Soc. 135 (2013) 7118. [9] F. Goettmann, A. Fischer, M. Antonietti, A. Thomas, Angew. Chem. Int. Ed. 45 (2006) 4467. [10] J.-M. Gu, W.-S. Kim, Y.-K. Hwang, S. Huh, Carbon 56 (2013) 208. [11] G.-P. Hao, W.-C. Li, D. Qian, A.-H. Lu, Adv. Mater. 22 (2010) 853. [12] M. Sevilla, P. Valle-Vigón, A.B. Fuertes, Adv. Funct. Mater. 21 (2011) 2781. [13] G.-P. Hao, W.-C. Li, D. Qian, G.-H. Wang, W.-P. Zhang, T. Zhang, A.-Q. Wang, F. Schüth, H.-J. Bongard, A.-H. Lu, J. Am. Chem. Soc. 133 (2011) 11378. [14] J.P. Paraknowitsch, A. Thomas, J. Schmidt, Chem. Commun. 47 (2011) 8283. [15] F. Böttger-Hiller, A. Mehner, S. Anders, L. Kroll, G. Cox, F. Simond, S. Spange, Chem. Commun. 48 (2012) 10568. [16] Y. Xia, Y. Zhu, Y. Tang, Carbon 50 (2012) 5543. [17] G. Xu, B. Ding, P. Nie, L. Shen, H. Dou, X. Zhang, ACS Appl. Mater. Interfaces 6 (2014) 194. [18] M. Sevilla, A.B. Fuertes, Micropor. Mesopor. Mater. 158 (2012) 318. [19] Y. Xia, R. Mokaya, G.S. Walker, Y. Zhu, Adv. Energy Mater. 1 (2011) 678. [20] H. Seema, K.C. Kemp, N.H. Le, S.-W. Park, V. Chandra, J.W. Lee, K.S. Kim, Carbon 66 (2014) 320. [21] M. Seredych, J. Jagiello, T.J. Bandosz, Carbon 74 (2014) 207. [22] Y.-K. Hwang, T.-H. Kwon, S.S. Park, Y.-J. Yoon, Y.S. Won, S. Huh, Eur. J. Inorg. Chem. (2010) 4747. [23] J.-Y. Hong, Y.-K. Hwang, Y.M. Kye, H. Myung, Y.S. Won, S. Huh, Mater. Lett. 96 (2013) 181. [24] K.K.R. Datta, V.V. Balasubramanian, K. Ariga, T. Mori, A. Vinu, Chem. Eur. J. 17 (2011) 3390. [25] E. Omurzak, T. Mashimo, S. Sulaimankulova, S. Takebe, L. Chen, Z. Abdullaeva, C. Iwamoto, Y. Oishi, H. Ihara, H. Okudera, A. Yoshiasa, Nanotechnology 22 (2011) 365602. [26] M. Kruk, M. Jaroniec, Chem. Mater. 13 (2001) 3169.
Contents lists available at ScienceDirect
Journal of Colloid and Interface Science www.elsevier.com/locate/jcis
Hollow S-doped carbon spheres from spherical CT/PEDOT composite particles and their CO2 sorption properties Jin-Yeon Hong, Seong Huh ⇑ Department of Chemistry and Protein Research Center for Bio-Industry, Hankuk University of Foreign Studies, Yongin 449-791, Republic of Korea
a r t i c l e
i n f o
Article history: Received 15 July 2014 Accepted 2 September 2014 Available online 16 September 2014 Keywords: Porous carbon S-doping Conducting polymers Carbonization Carbon dioxide capture
a b s t r a c t Chemically functionalizable shape-controlled poly(3,4-ethylenedioxythiophene) (PEDOT)-derived conducting copolymers, C1(C4)-CT/PEDOT/PSS-20APS and C1(C4)-CT/PEDOT/PSS-10APS, were prepared through oxidative polymerization of 3,4-ethylenedioxythiophene (EDOT) and 3-thiophenecarboxylic acid (C1-CT) or 4-(3-thienyl)butyric acid (C4-CT) in the presence of acid-labile mesoporous ZnO/Zn(OH)2 hard template. The mesoporous ZnO/Zn(OH)2 hard template could be removed by mild acid etching. The morphology of these polymeric microparticles was dependent on the concentration of ammonium persulfate (APS, (NH4)2S2O8) catalyst and the type of CT monomers. Hollow capsular C1-CT/PEDOT/PSS-20APS copolymer spheres with a large surface opening were obtained when the amount of oxidizing agent APS was 20 mmol under the same experimental conditions. Because C1(C4)-CT/PEDOT/PSS-xAPS copolymers contain S-rich moieties in the polymer backbone, they are suited for the preparation of S-doped carbonaceous materials. Therefore, we carbonized C1-CT/PEDOT/PSS-20APS at several different temperatures in a high-purity nitrogen atmosphere to easily prepare hollow S-doped carbon spheres (HSCSs). The level of S-dopants in carbon spheres was strongly dependent on the carbonization temperature. Lower carbonization temperature led to a higher content of S-dopants but lower BET surface area. These carbon spheres were further analyzed by TEM, SEM, PXRD, and XPS. Gas sorption analyses were also performed to study gas sorption with different amount of S-dopants. Ó 2014 Elsevier Inc. All rights reserved.
1. Introduction Nanoporous carbon materials are important in modern materials chemistry because of their utility in a wide variety of applications. There are numerous reports dealing with porous carbon materials with distinct porosity and pore functionality. The effectiveness of using porous carbons for various advanced applications such as H2 storage [1], CO2 capture [2], encapsulation of catalytically active metallic nanoparticles [3], electrode materials in energy-storage devices [4], and supercapacitors [5] is mainly dependent on the porosity and functionality of the carbon materials. The surface area, pore size, and pore volume are the primary factors governing the efficiency of gas sorption and encapsulation of functional nanoparticles. Among these porous carbonaceous materials, heteroatomdoped carbons such as N-doped carbon materials having Lewis basic sites in the framework often possess enhanced surface functionalities especially in many applications related to CO2 capture [6], metal-free oxygen reduction reaction (ORR) [7], photocatalysis ⇑ Corresponding author. Fax: +82 31 330 4566. E-mail address: [email protected] (S. Huh). http://dx.doi.org/10.1016/j.jcis.2014.09.003 0021-9797/Ó 2014 Elsevier Inc. All rights reserved.
[8], and catalytic organic transformations [9]. In general, the preparation of N-doped carbon materials inevitably requires N-rich organic polymer precursors for carbonization in inert atmosphere. Many research groups reported on novel N-rich organic-based precursors for the carbonization. For instances, we developed N-doped porous carbons having good CO2 sorption abilities [10]. The carbonization precursors were prepared by the azide-alkyne 1,3dipolar Huisgen cycloaddition reaction, also known as the Click reaction, between 1,4-bis(azidomethyl)benzene and 1,3,5-ethynylbenzene. N-doped carbon monolith with Brunauer–Emmett–Teller (BET) surface area of 467 m2 g 1 was also obtained by carbonizing polypyrrole precursor at 500 °C by Hao and co-workers [11]. Sevilla and co-workers prepared highly N-doped carbon material by using polypyrrole as a carbon source and KOH activation of the resulting carbon [12]. Mechanically stable N-doped porous carbons were also prepared from poly(benzoxazine-co-resol) precursor and subsequently tested for CO2 separation from a flue gas [13]. Most of these N-doped porous carbons are being actively investigated to fine-tune the content of N atoms and porosity for enhanced functionalities. In contrast, relatively few studies have described for the preparation and characterization of S-doped porous carbon materials
78
J.-Y. Hong, S. Huh / Journal of Colloid and Interface Science 436 (2014) 77–82
[14–21]. The CO2 sorption properties of S-doped porous carbons have only very recently been explored [16,19–21]. Both the Schmidt and the Böttger-Hiller groups independently prepared S-doped microporous carbon materials using thiophene-containing organic polymer precursors [14,15]. S-doped microporous carbon materials might have comparable properties as N-doped carbon materials in gas sorption applications. Xia and co-workers recently reported on S-doped microporous carbons prepared through nanocasting of zeolite EMC-2 [16]. The resulting S-doped microporous carbons exhibited superior adsorption abilities for both H2 and CO2. Notably, the low surface coverage heat of CO2 adsorption had a high value of 59 kJ mol 1. Moreover, hierarchically structured S-doped carbon materials were demonstrated as good cathode materials for rechargeable Li–S batteries [17]. We envision that the utilization of S-doped porous carbon materials will be widely expanded. In this study, we present a new simple method to prepare shape-controlled S-containing carbon precursors from poly(3,4ethylenedioxythiophene) (PEDOT)-derived conducting polymers which were prepared in the presence of poly(4-ammonium styrenesulfonic acid) (NH4PSS) and mesoporous ZnO/Zn(OH)2 composite hard template. One of the resulting carbon precursors, C1-CT/PEDOT/PSS-20APS, was carbonized at several different temperatures to produce shape controlled hollow S-doped carbon spheres (HSCSs). The doping level of S atoms into HSCSs upon change of carbonization temperature was investigated by elemental analysis and X-ray photoelectron spectroscopy (XPS). Gas sorption analyses using N2 and CO2 were also performed to study the textural properties and CO2 sorption abilities of HSCSs. The relationship between the textural properties of HSCSs and the S-contents for CO2 sorption will be discussed.
2. Experimental 2.1. Synthesis of mesoporous ZnO/Zn(OH)2 Mesoporous ZnO/Zn(OH)2 composite material was prepared according to the method previously reported [22]. 2.2. Synthesis of C1-CT/PEDOT/PSS-20APS and C1-CT/PEDOT/PSS-10APS These products were prepared by modifying a method in the literature, employing increased amount of ammonium persulfate polymerization catalyst [23]. 3,4-Ethylenedioxythiophene (0.947 mL, 8.87 mmol, Sigma–Aldrich), 3-thiophenecarboxylic acid (0.032 g, 0.25 mmol, Sigma–Aldrich), and a 30 wt% aqueous solution of poly(4-ammonium styrenesulfonic acid) (2 mL, Sigma–Aldrich) were dissolved in 20 mL of distilled water. Ammonium persulfate (4.592 g, 20 mmol for C1-CT/PEDOT/PSS-20APS; 2.296 g, 10 mmol for C1-CT/PEDOT/PSS-10APS) was added after 5 min and the reaction mixture was stirred constantly (rpm = 750). After 30 min of reaction, 500 mg of mesoporous ZnO/Zn(OH)2 was added. The reaction mixture was further stirred at room temperature for one week. The solid product was separated by centrifugation and washed with distilled water and methanol. In order to remove the ZnO/Zn(OH)2 template, the solids were etched with 10 mL of 1.0 M HCl for 1 h with stirring. Finally, the resulting solids were washed with water and methanol and dried at 80 °C for 2 h. 2.3. Synthesis of C4-CT/PEDOT/PSS-20APS and C4-CT/PEDOT/PSS10APS These materials were prepared using the same method for C1-CT/PEDOT/PSS-20APS and C1-CT/PEDOT/PSS-10APS (above)
except 4-(3-thienyl)butyric acid (0.043 g, 0.25 mmol, Rieke Fine Chemicals) was used instead of 3-thiophenecarboxylic acid. 2.4. Physical measurements Powder X-ray diffraction patterns were recorded on a Rigaku MiniFlex (30 kV, 15 mA). FE-SEM images were recorded on an FEI Nova 200 NanoSEM (accelerating voltage = 10 kV). The water suspension of the sample was dropped and dried on a Cu grid supported by a carbon film for TEM measurement, for which a JEOL JEM-3000F (accelerating voltage = 300 kV) equipped with an EDAX for the EDS spectrum was used. XPS data were collected on a K-Alpha X-ray photoelectron spectrometer (Thermo VG, UK). The X-ray source was a monochromated Al Ka line (hm = 1486.6 eV), and the X-ray power was 12 kV at 3 mA. The pass energy of the survey spectrum was fixed at 100 eV with a step size of 1 eV while the pass energy for the high-resolution spectrum was fixed at 50 eV with a step size of 0.1 eV. The sampling area was 400 lm in diameter. The binding energies were referenced to 284.8 eV (C 1s peak for C–C bonds). The cryogenic volumetric N2 adsorption– desorption analysis was performed on a Belsorp-miniII at 77 K (BEL Japan). The samples were evacuated at 423 K under high vacuum for 2 h. Low pressure CO2 adsorption measurements were also performed on a Belsorp-miniII at 196 K (2-propanol/dry ice bath) and 273 K (ice bath). 3. Results and discussion We recently reported on hierarchically structured mesoporous PEDOT-derived conducting polymers, C1-CT/PEDOT/PSS-5APS and C4-CT/PEDOT/PSS-5APS, containing 3-thiophenecarboxylic acid (C1-CT) or 4-(3-thienyl)butyric acid (C4-CT) in the polymer backbone [23]. The amount of ammonium persulfate (APS, (NH4)2S2O8) catalyst was 5 mmol in each case. Interestingly, the porous materials were large particles with microscale dimensions having cheese-like wall structures. The porous walls contained small polymeric fibers. As a result, the BET surface areas of C1-CT/ PEDOT/PSS-5APS and C4-CT/PEDOT/PSS-5APS were 60.1 m2 g 1 and 44.5 m2 g 1, respectively. In order to accelerate the polymerization reaction in the same system, the concentration of the catalyst, APS, was increased twofold or fourfold compared with the previous condition for C1(C4)-CT/PEDOT/PSS-5APS [23]. Surprisingly, the simple increase of concentration of APS catalyst not only accelerated the polymerization rate, but also led to discrete smaller solid particles with distinct particle morphology (Fig. 1). Four different products were prepared in this study by using 20 mmol or 10 mmol of APS in the presence of C1-CT or C4-CT along with EDOT. C1-CT/PEDOT/ PSS-20APS is a sample obtained from 20 mmol APS with C1-CT while C1-CT/PEDOT/PSS-10APS is a sample from 10 mmol APS with C1-CT. The particle morphology dramatically varied depending on the ratio of APS and the type of CT monomers. As shown in Fig. 1a and b, the type of CT monomers induced quite different particle morphologies. The C1-CT/PEDOT/PSS-20APS sample had spherical particles with a large surface hole while C4-CT/PEDOT/PSS-20APS displayed much smaller particles with less featured shape. Both C1-CT/PEDOT/PSS-10APS and C4-CT/PEDOT/PSS-10APS indicated similar particle dimensions compared with C1(C4)-CT/PEDOT/ PSS-20APS but their particle morphologies were also less featured. Among the four samples, C1-CT/PEDOT/PSS-20APS was chosen as a carbonization precursor due to its uniformity in particle shape and dimension. Furthermore, most particles in the C1-CT/PEDOT/ PSS-20APS had hollow, spherical geometry which is ideal for the generation of hollow capsular carbon materials. C1-CT/PEDOT/ PSS-20APS was carbonized at 600 °C, 700 °C, 800 °C, and 900 °C
J.-Y. Hong, S. Huh / Journal of Colloid and Interface Science 436 (2014) 77–82
79
Fig. 1. FE-SEM images of the nanostructured samples on the same magnifications (20,000). (a) C1-CT/PEDOT/PSS-20APS, (b) C4-CT/PEDOT/PSS-20APS, (c) C1-CT/PEDOT/ PSS-10APS, and (d) C4-CT/PEDOT/PSS-10APS.
to give HSCS-600, HSCS-700, HSCS-800, HSCS-900, respectively. Transmission electron microscopy (TEM) investigation of HSCSs confirmed that the majority of hollow carbon precursors were successfully converted into HSCSs (Fig. 2). As shown in Fig. 3, the low magnification TEM image of the representative sample, HSCS-700, clearly indicates that most of the particles have hollow capsular geometry and their sizes are similar to each other. The diameters of particles are about 250–500 nm. The powder X-ray diffraction patterns of the HSCS materials are shown in Fig. 4. All HSCS samples exhibited two very broad diffraction peaks centered at around 24° and 44°, which indicated the disordered graphitic nature of the carbons [24]. The corresponding d-spacing values are 0.37 nm and 0.21 nm, respectively. Very interestingly, upon increase of the carbonization temperature from 600 °C to 900 °C, the diffraction peaks corresponding to the wurtzite-type ZnS were clearly revealed. The complete indexing results of wurtzite-type ZnS (JCPDS card No. 36-1450) in HSCS-900 is shown in Fig. 4d [25]. The source of Zn could be from the incompletely removed ZnO/Zn(OH)2 hard template after acid etching. Therefore, ZnS was formed during the carbonization of C1-CT/ PEDOT/PSS-20APS. Only HSCS-900 showed well crystallized wurtzite-type ZnS. As a result, HSCS-900 can be considered to contain nanoparticulate semiconducting ZnS in S-doped carbon structure. The S-contents of HSCS samples were estimated from elemental analysis (Table 1). Upon increase of the carbonization temperature from 600 °C to 900 °C, the content of S atoms gradually decreased from 21.6% for HSCS-600 to 8.71% for HSCS-900. However, it must be considered that some of S atoms were not directly incorporated in the carbon network, but contained as a form of ZnS nanoparticles. Nevertheless, very limited amounts of ZnS were detected on XPS surface analysis based on Zn 2p orbital signals (data not shown). All HSCS samples exhibited C-bound S 2p orbital doublet
at around 163.7 eV and 165.0 eV in XPS spectra (Fig. 5) [14]. These signals were previously assigned as sulfide bridge (–C–S–C–) species [18]. There were relatively small amounts of several oxidized forms of S atom which gave rise to signals at 166.5 eV, 167.8 eV, and 169.0 eV. The atomic percentages of oxidized S species among S-dopants estimated by XPS were 11% (HSCS-600), 15% (HSCS700), 20% (HSCS-800), and 16% (HSCS-900). The intensities corresponding to S 2p orbital signals gradually decrease upon increase of carbonization temperature, indicating lower S-dopant content at higher temperature. In contrast to S-dopant contents, the BET surface areas were proportional to the carbonization temperature. HSCS-900 exhibited the highest BET surface area of 315 m2 g 1 with pore volume of 0.21 cm3 g 1. Compared with HSCS-900, HSCS-600 showed very low levels of BET surface area and pore volume. The pore volume for HSCS-600 was just 0.082 cm3 g 1. N2 adsorption–desorption isotherms (Fig. 6a) clearly demonstrates that both HSCS-800 and HSCS-900 have better N2 sorption abilities than both HSCS-600 and HSCS-700. The shapes of isotherms for HSCS-800 and HSCS-900 are type I [26]. The heteroatom substituted carbon materials are thought to have better CO2 sorption abilities than pure carbonaceous materials due to the enhanced interactions between quadrupolar CO2 molecules and heteroatoms in the network. S-doped carbons, as well as N-doped carbons, have better CO2 sorption characteristics due to the Lewis base sites than conventional carbon materials [16]. The CO2 uptake abilities of the four HSCS samples were volumetrically measured at 196 K (Fig. 6b). To our surprise, despite the large differences in BET surface areas and pore volumes, all HSCS samples exhibited a very similar range of CO2 uptake values at 196 K: 90.2–98.4 cm3 g 1 (4.0–4.4 mmol g 1). Considering the large differences in BET surface areas and pore volumes of each HSCS sample, these results are quite astonishing. One plausible
80
J.-Y. Hong, S. Huh / Journal of Colloid and Interface Science 436 (2014) 77–82
Fig. 2. TEM images of HSCS samples. (a) HSCS-600, (b) HSCS-700, (c) HSCS-800, and (d) HSCS-900.
Fig. 4. PXRD patterns of the samples. (a) HSCS-600, (b) HSCS-700, (c) HSCS-800, and (d) HSCS-900. The wurtzite-type ZnS phase was indexed in (d). Fig. 3. Low magnification TEM image of HSCS-700.
account of these unexpected behaviors may be the role of S-dopant in the carbon frameworks. HSCS-600, having the smallest BET surface area, actually contains the highest content of S-dopant. As a result, the highest S-dopant content in HSCS-600 sufficiently offset the lowest BET surface area and pore volume to lead to similar levels of CO2 sorption compared with HSCS-800 and HSCS-900. The isosteric heats of adsorption (Qst) for all HSCS samples were
estimated by using the Clausius–Clapeyron equation based on the adsorption data measured at 196 K and 273 K (Fig. 7). The values of low surface coverage Qst are increasing in the order of HSCS-900 > HSCS-800 > HSCS-600 > HSCS-700. Therefore, despite the lowest level of S-content, HSCS-900, exhibited the highest Qst and uptake capacity for CO2. The range of Qst values, 23.1– 32.8 kJ mol 1, is comparable to the commonly observed values for N-doped porous carbons (18.9–36.0 kJ mol 1) [12,19].
81
J.-Y. Hong, S. Huh / Journal of Colloid and Interface Science 436 (2014) 77–82 Table 1 Contents of S-dopants, textural properties, and CO2 sorption data of HSCS materials. Samples
S content (%)
BET surface area (m2 g
HSCS-600 HSCS-700 HSCS-800 HSCS-900
21.6 17.0 8.92 8.71
117 161 280 315
1
)
Pore volume (cm3 g 0.082 0.093 0.24 0.21
1
)
CO2 uptake at 196 K (cm3 g 90.2 98.4 90.7 94.1
1
)
Heats of CO2 adsorption (kJ mol
1
)
24.5 23.1 27.8 32.8
Fig. 5. High-resolution XPS spectra of S 2p orbital region for HSCSs. (a) HSCS-600, (b) HSCS-700, (c) HSCS-800, and (d) HSCS-900.
Fig. 7. Isosteric heats of CO2 adsorption for HSCS samples upon increase of CO2 uptake amount.
Currently, the detailed role of S-dopant in the carbon network for CO2 sorption application has not been as understood as N-doped carbon materials. Xia and co-workers reported relatively high heat of CO2 adsorption for S-doped microporous carbons derived from EMC-2 zeolite hard template [16]. They concluded that despite the high heat of adsorption the magnitude of textural properties (surface areas and pore volumes) of S-doped microporous carbons was a primary governing factor for enhanced CO2 sorption. Kim and co-workers pointed out that the content of oxidized S species as well as the textural properties of S-doped microporous carbon was important for enhancing the CO2 sorption [20]. Bandosz and co-workers considered both S atoms incorporated in aromatic rings and oxidized forms of S atom, such as sulfonic acids, sulfoxides, and sulfones, to be determining factors for efficient interactions between the carbon network and CO2 gaseous molecules [21]. Since our HSCS materials are mainly composed of sulfide bridge (–C–S–C–) types of S-dopants based on XPS analysis,
our results suggest that the overall content of various forms of Sdopant is also an important factor not only to increase the heat of adsorption compared with S-free carbons, but also to enhance the uptake capacity for CO2. 4. Conclusion We have developed a simple preparation method for submicron-scale functionalizable PEDOT-derived hollow spherical polymeric particles, which could be carbonized into hollow S-doped carbon spheres. The uniformly shaped hollow S-doped carbon spheres exhibited a temperature dependence on the S contents. The content of S-dopants ranged from 8.71% to 21.6% based on elemental analysis. The higher the carbonization temperature, the lower the content of S-dopants. In contrast, the BET surface areas of the S-doped carbons are directly proportional to the carbonization temperature. The carbon samples carbonized above at 800 °C
Fig. 6. (a) N2 adsorption–desorption isotherms measured at 77 K and (b) CO2 adsorption–desorption isotherms measured at 196 K. Solid symbols and open symbols represent adsorption and desorption isotherms, respectively.
82
J.-Y. Hong, S. Huh / Journal of Colloid and Interface Science 436 (2014) 77–82
were microporous materials judged from BET surface areas and N2 sorption isotherms. The hollow S-doped carbon spheres were good sorbents for CO2 gas at cryogenic conditions. All hollow S-doped carbon spheres, which were carbonized at 600–900 °C, exhibited a wide range of BET surface areas and S-dopant contents. Nevertheless, all four samples showed a very similar level of CO2 sorption abilities despite their differences in BET surface areas and the content of S-dopants. Importantly, the surface area and the overall content of S-dopants played a significant role the ability to take up CO2 gas. Acknowledgment This work was supported by the Hankuk University of Foreign Studies Research Fund of 2014. References [1] H. Nishihara, T. Kyotani, Adv. Mater. 24 (2012) 4473. [2] M.S. Shafeeyan, W.M.A.W. Daud, A. Houshmand, A. Shamiri, J. Anal. Appl. Pyrol. 89 (2010) 143. [3] S.H. Joo, S.J. Choi, I. Oh, J. Kwak, J. Liu, O. Terasaki, R. Ryoo, Nature 412 (2001) 169. [4] L.L. Zhang, X.S. Zhao, Chem. Soc. Rev. 38 (2009) 2520. [5] M. Inagaki, H. Konno, O. Tanaike, J. Power Sources 195 (2010) 7880. [6] N.P. Wickramaratne, J. Xu, M. Wang, L. Zhu, L. Dai, M. Jaroniec, Chem. Mater. 26 (2014) 2820.
[7] N. Li, C. Hu, L. Ren, M. Cao, Cryst. Eng. Commun. 15 (2013) 8504. [8] M. Shalom, S. Inal, C. Fettkenhauer, D. Neher, M. Antonietti, J. Am. Chem. Soc. 135 (2013) 7118. [9] F. Goettmann, A. Fischer, M. Antonietti, A. Thomas, Angew. Chem. Int. Ed. 45 (2006) 4467. [10] J.-M. Gu, W.-S. Kim, Y.-K. Hwang, S. Huh, Carbon 56 (2013) 208. [11] G.-P. Hao, W.-C. Li, D. Qian, A.-H. Lu, Adv. Mater. 22 (2010) 853. [12] M. Sevilla, P. Valle-Vigón, A.B. Fuertes, Adv. Funct. Mater. 21 (2011) 2781. [13] G.-P. Hao, W.-C. Li, D. Qian, G.-H. Wang, W.-P. Zhang, T. Zhang, A.-Q. Wang, F. Schüth, H.-J. Bongard, A.-H. Lu, J. Am. Chem. Soc. 133 (2011) 11378. [14] J.P. Paraknowitsch, A. Thomas, J. Schmidt, Chem. Commun. 47 (2011) 8283. [15] F. Böttger-Hiller, A. Mehner, S. Anders, L. Kroll, G. Cox, F. Simond, S. Spange, Chem. Commun. 48 (2012) 10568. [16] Y. Xia, Y. Zhu, Y. Tang, Carbon 50 (2012) 5543. [17] G. Xu, B. Ding, P. Nie, L. Shen, H. Dou, X. Zhang, ACS Appl. Mater. Interfaces 6 (2014) 194. [18] M. Sevilla, A.B. Fuertes, Micropor. Mesopor. Mater. 158 (2012) 318. [19] Y. Xia, R. Mokaya, G.S. Walker, Y. Zhu, Adv. Energy Mater. 1 (2011) 678. [20] H. Seema, K.C. Kemp, N.H. Le, S.-W. Park, V. Chandra, J.W. Lee, K.S. Kim, Carbon 66 (2014) 320. [21] M. Seredych, J. Jagiello, T.J. Bandosz, Carbon 74 (2014) 207. [22] Y.-K. Hwang, T.-H. Kwon, S.S. Park, Y.-J. Yoon, Y.S. Won, S. Huh, Eur. J. Inorg. Chem. (2010) 4747. [23] J.-Y. Hong, Y.-K. Hwang, Y.M. Kye, H. Myung, Y.S. Won, S. Huh, Mater. Lett. 96 (2013) 181. [24] K.K.R. Datta, V.V. Balasubramanian, K. Ariga, T. Mori, A. Vinu, Chem. Eur. J. 17 (2011) 3390. [25] E. Omurzak, T. Mashimo, S. Sulaimankulova, S. Takebe, L. Chen, Z. Abdullaeva, C. Iwamoto, Y. Oishi, H. Ihara, H. Okudera, A. Yoshiasa, Nanotechnology 22 (2011) 365602. [26] M. Kruk, M. Jaroniec, Chem. Mater. 13 (2001) 3169.
