Microcrystalline Composite Particles of Carbon Nanotubes and

Feb 15, 2008 - Materials Science Laboratory, Stuttgart Technology Center, Sony ... 61, D-70327 Stuttgart, Germany, and Sony Corporation Gotenyama ...
1 downloads 0 Views 1MB Size
Download PDF
Langmuir 2008, 24, 3479-3485

3479

Microcrystalline Composite Particles of Carbon Nanotubes and Calcium Carbonate William E. Ford,*,† Akio Yasuda,‡ and Jurina M. Wessels† Materials Science Laboratory, Stuttgart Technology Center, Sony Deutschland GmbH Hedelfingerstrasse 61, D-70327 Stuttgart, Germany, and Sony Corporation Gotenyama Technology Center, Life Science Laboratory, 5-1-12 Kitashinagawa, Shinagawa-ku, Tokyo, 141-0001 Japan ReceiVed August 30, 2007. In Final Form: December 11, 2007 Coprecipitation of urea-melt modified carbon nanotubes and calcium carbonate from an aqueous solution by two methods yielded microcrystalline composite particles. Powders obtained by colloidal crystallization from a supersaturated solution that were isolated and dried soon after precipitation were a mixture of raspberry-shaped and rhombohedral particles. These were shown by infrared and X-ray diffraction analyses to be mainly calcite. Particles that were kept wet for 1 day or longer before being isolated were typically entirely rhombohedral with edge lengths in the range of 5-30 µm. Scanning electron microscopy investigations revealed that the nanotubes were adsorbed on the particle surface and also incorporated into the interior matrix. Removal of the calcium carbonate component by treating the particles with acid yielded nanotube shells whose size and shape reflected those of the original particles.

Introduction Carbon nanotubes (CNTs) are investigated for a variety of biomedical applications that utilize their high tensile strength, electrical conductivity, and chemical stability.1 One particular application of interest is the use of composite materials containing CNTs for bone tissue engineering as scaffolds to sustain bone cell growth or to replace bone tissue.2 Besides providing mechanical strength to the composite materials, the inherent electrical conductivity of CNTs also makes it possible to deliver electrical stimulation to bone-forming cells (osteoblasts) to accelerate bone repair.3 Materials that are used as scaffolds or replacements for bone tissue are often coated with the calcium phosphate minerals hydroxyapatite or carbonated hydroxyapatite because of their biocompatibility, nontoxicity, and bioactivity (i.e., ability to promote interaction with living tissue such as bone, which is largely a composite of calcium phosphate and collagen). Composites of CNTs with these two minerals have been the subject of recent * Corresponding author. E-mail: [email protected]; tel.: +49 711 5858723; fax: +49 711 5858484. † Stuttgart Technology Center. ‡ Sony Corporation Gotenyama Technology Center. (1) (a) Mattson, M. P.; Haddon, R. C.; Rao, A. M. J. Mol. Neurosci. 2000, 14, 175-182. (b) Correa-Duarte, M. A.; Wagner, N.; Rojas-Chapana, J.; Morsczeck, C.; Thie, M.; Giersig, M. Nano Lett. 2004, 4, 2233-2236. (c) Gheith, M. K.; Sinani, V. A.; Wicksted, J. P.; Matts, R. L.; Kotov, N. A. AdV. Mater. 2005, 17, 2663-2670. (d) Hu, H.; Yu, A.; Kim, E.; Zhao, B.; Itkis, M. E.; Bekyarova, E.; Haddon, R. C. J. Phys. Chem. B 2005, 109, 11520-11524. (e) Ni, Y.; Hu, H.; Malarkey, E. B.; Zhao, B.; Montana, V.; Haddon, R. C.; Parpura, V. J. Nanosci. Nanotechnol. 2005, 5, 1707-1712. (f) Sorkin, R.; Gabay, T.; Blinder, P.; Baranes, D.; Ben-Jacob, E.; Hanein, Y. J. Neural Eng. 2006, 3, 95-101. (g) Wang, K.; Fishman, H. A.; Dai, H.; Harris, J. S. Nano Lett. 2006, 6, 2043-2048. (h) Liopo, A. V.; Stewart, M. P.; Hudson, J.; Tour, J. M.; Pappas, T. C. J. Nanosci. Nanotechnol. 2006, 6, 1365-1374. (i) Gheith, M. K.; Pappas, T. C.; Liopo, A. V.; Sinani, V. A.; Shim, B. S.; Motamedi, M.; Wicksted, J. P.; Kotov, N. A. AdV. Mater. 2006, 18, 2975-2979. (j) Harrison, B. S.; Atala, A. Biomaterials 2007, 28, 344-353. (k) Jan, E.; Kotov, N. A. Nano Lett. 2007, 7, 1123-1128. (2) (a) MacDonald, R. A.; Laurenzi, B. F.; Viswanathan, G.; Ajayan, P. M.; Stegemann, J. P. J. Biomed. Mater. Res., Part A 2005, 74, 489-496. (b) Wang, W.; Omori, M.; Watari, F.; Yokoyama, A. Dent. Mater. J. 2005, 24, 478-486. (c) Shi, X.; Hudson, J. L.; Spicer, P. P.; Tour, J. M.; Krishnamoorti, R.; Mikos, A. G. Biomacromolecules 2006, 7, 2237-2242. (d) Meng, J.; Song, L.; Meng, J.; Kong, H.; Zhu, G.; Wang, C.; Xu, L.; Xie, S.; Xu, H. Biomed. Mater. Res., Part A 2006, 79, 298-306. (e) Zanello, L. P.; Zhao, B.; Hu, H.; Haddon, R. C. Nano Lett. 2006, 6, 562-567. (f) Marrs, B.; Andrews, R.; Rantell, T.; Pienkowski, D. J. Biomed. Mater. Res., Part A 2006, 77, 269-276. (g) Firkowska, I.; Olek, M.; Pazos-Perez, N.; Rojas-Chapana, J.; Giersig, M. Langmuir 2006, 22, 54275434. (3) Supronowicz, P. R.; Ajayan, P. M.; Ullmann, K. R.; Arulanandam, B. P.; Metzger, D. W.; Bizios, R. J. Biomed. Mater. Res. 2002, 59, 499-506.

investigations.4 Calcium carbonate, which is the major mineral component of the exoskeletons of many animals and comprises approximately 95% of the dry weight of birds’ egg shells, can be converted into hydroxyapatite or carbonated hydroxyapatite and so can be used as a precursor for these minerals.5 Calcium carbonate also has good biocompatibility and is used as a component in composites for biomedical applications.6 Despite the interest in both CNT- and CaCO3-based composites noted previously, composite materials comprising CNTs and CaCO3 have so far been scarcely reported.7 Anderson and Barron observed that the precipitation of CaCO3 in the presence of hydroxy- or carboxy-functionalized single-walled CNTs (SWNTs) yielded amorphous CaCO3-coated SWNTs.7a Liu et al. produced CNTs coated with CaCO3 nanoparticles by precipitating CaCO3 in the presence of nonfunctionalized multiwalled CNTs (MWNTs), but the phase state of CaCO3 was not reported.7b Most recently, Li and Gao investigated CNT-CaCO3 composites obtained by precipitating calcium carbonate in the presence of various types (4) (a) Zhao, L.; Gao, L. Carbon 2004, 42, 423-426. (b) Zhao, B.; Hu, H.; Mandal, S. K.; Haddon, R. C. Chem. Mater. 2005, 17, 3235-3241. (c) Kealley, C.; Elcombe, M.; van Riessen, A.; Ben-Nissan, B. Physica B 2006, 385-386, 496-498. (d) Aryal, S.; Bhattarai, S. R.; Bahadur, K. C. R.; Khil, M. S.; Lee, D.-R.; Kim, H. Y. Mater. Sci. Eng., A 2006, 426, 202-207. (e) Aryal, S.; Bahadur, K. C. R.; Dharmaraj, N.; Kim, K.-W.; Kim, H. Y. Scr. Mater. 2006, 54, 131-135. (f) Akasaka, T.; Watari, F.; Sato, Y.; Tohji, K. Mater. Sci. Eng., C 2006, 26, 675-678. (g) Tasis, D.; Kastanis, D.; Galiotis, C.; Bouropoulos, N. Phys. Status Solidi B 2006, 243, 3230-3233. (h) Balani, K.; Anderson, R.; Laha, T.; Andara, M.; Tercero, J.; Crumpler, E.; Agarwal, A. Biomaterials 2007, 28, 618-624. (i) Chen, Y.; Zhang, T. H.; Gan, C. H.; Yu, G. Carbon 2007, 45, 998-1004. (5) (a) Jinawath, S.; Polchai, D.; Yoshimura, M. Mater. Sci. Eng., C 2002, 22, 35-39. (b) Ni, M.; Ratner, B. D. Biomaterials 2003, 24, 4323-4331. (c) Kamiya, M.; Hatta, J.; Shimada, E.; Ikuma, Y.; Yoshimura, M.; Monma, H. Mater. Sci. Eng., B 2004, 111, 226-231. (d) Yoshimura, M.; Sujaridworakun, P.; Koh, F.; Fujiwara, T.; Pongkao, D.; Ahniyaz, A. Mater. Sci. Eng., C 2004, 24, 521-525. (e) Rocha, J. H.; Lemos, A. F.; Agathopoulos, S.; Valerio, P.; Kannan, S.; Oktar, F. N.; Ferreira, J. M. Bone 2005, 37, 850-857. (f) Rocha, J. H.; Lemos, A. F.; Agathopoulos, S.; Kannan, S.; Valerio, P.; Ferreira, J. M. J. Biomed. Mater. Res., Part A 2006, 77, 160-168. (6) (a) Serizawa, T.; Tateishi, T.; Akashi, M. J. Biomater. Sci., Polym. Ed. 2003, 14, 653-663. (b) Kasuga, T.; Maeda, H.; Kato, K.; Nogami, M.; Hata, K.; Ueda, M. Biomaterials 2003, 24, 3247-3253. (c) Ogomi, D.; Serizawa, T.; Akashi, M. J. Biomed. Mater. Res., Part A 2003, 67, 1360-1366. (d) Maeda, H.; Kasuga, T. Acta Biomater. 2006, 2, 403-408. (7) (a) Anderson, R. E.; Barron, A. R. Main Group Chem. 2005, 4, 279-289. (b) Liu, Y.; Wang, R.; Chen, W.; Chen, X.; Hu, Z.; Cheng, X.; Xin, H. J. Chem. Lett. 2006, 35, 200-201. (c) Li, W.; Gao, C. Langmuir 2007, 23, 4575-4582. (d) Tasis, D.; Pispas, S.; Galiotis, C.; Bouropoulos, N. Mater. Lett. 2007, 61, 5044-5046.

10.1021/la7026797 CCC: $40.75 © 2008 American Chemical Society Published on Web 02/15/2008

3480 Langmuir, Vol. 24, No. 7, 2008

of carboxy-functionalized and pristine CNTs.7c They concluded that carboxy-functionalized CNTs, both multiwalled and singlewalled, nucleate and stabilize micrometer-scale spherical crystals of vaterite, which is the least thermodynamically stable crystalline phase of CaCO3. Stabilization of vaterite occurred as a result of the crystals being embedded and covered by the CNTs.7c We recently reported a method for solubilizing oxidatively purified CNTs by reacting them with molten urea.8 The process converts carboxy groups on the CNTs into amido and ureido groups, thereby making the nanotubes highly soluble in water (>1 g/L). Utilizing this property, we precipitated CaCO3 from aqueous solutions of the CNTs and examined the product, primarily by scanning electron microscopy (SEM). As detailed next, the product was comprised of microcrystalline composite particles that generally displayed the rhombohedral shape characteristic of calcite and that were shown by infrared and X-ray diffraction analyses to be mainly calcite. The SEM investigations revealed that the nanotubes were not simply adsorbed to the particle surface but incorporated into the interior matrix as well. This result contrasts that of Li and Gao, who found that carboxy-functionalized CNTs interact strongly with vaterite but not with calcite.7c We also demonstrated that the removal of the CaCO3 component by treating the particles with hydrochloric acid yielded nanotube shells whose size and shape reflected those of the original particles. Considering the potential biomedical applications of composite materials comprising CNTs and biocompatible minerals outlined previously, as well as an interest in CNT-based microcapsules,9 we believe that our findings are of interest to a wide scientific community. Experimental Procedures Materials. Ammonium carbonate (Fluka, #09716), calcium chloride dihydrate (Aldrich, #22,350-6), and sodium carbonate (Aldrich, #22,353-0) were used as received. Water was purified with a Milli-Q Synthesis A10 system and had a resistivity of 18.2 MΩ cm at 25 °C. The preparation of the soluble SWNTs from acid-purified nanotubes produced by a modified electric-arc method (Carbon Solutions, Inc.) was described previously,8 and the preparation of the soluble MWNTs is described in the Supporting Information. In some cases, the CNTs were solubilized using urea alone, and will be referred to as U-SWNTs or U-MWNTs, or using a combination of urea and anisaldehyde, and will be referred to as UA-SWNTs. We observed no significant differences in composites prepared from UA-SWNTs as compared to those prepared from U-SWNTs. The concentrations of the CNTs in the solutions were estimated, after dilution with water, by UV-vis absorption spectroscopy, assuming the extinction coefficient at 800 nm to be 19 (g/L)-1 cm-1.8 Preparation of U-SWNT-CaCO3 Composites by Diffusion of (NH4)2CO3 Vapor into CaCl2 Solution. This procedure is based on the method of Addadi et al.10 In a typical preparation, a solution of calcium chloride (0.030 M, 0.3 mL) in water was mixed with a solution of U-SWNTs in water (0.7 mL) in a polypropylene vial. The approximate concentration of SWNTs before being mixed with the calcium chloride solution was 4-5 mg/L, based on the absorbance of the solution at 800 nm.8 Ammonium carbonate powder (∼25 mg) was placed at the bottom of a 25 mL glass beaker together with the open vial containing the solution, and the beaker was sealed tightly (8) Ford, W. E.; Jung, A.; Hirsch, A.; Graupner, R.; Scholz, F.; Yasuda, A.; Wessels, J. M. AdV. Mater. 2006, 18, 1193-1197. (9) (a) Sano, M.; Kamino, A.; Okamura, J.; Shinkai, S. J. Nano Lett. 2002, 2, 531-533. (b) Panhuis, M. I. H.; Paunov, V. N. Chem. Commun. (Cambridge, U.K.) 2005, 1726-1728. (c) Paunov, V. N.; Panhuis, M. I. H. Nanotechnology 2005, 16, 1522-1525. (d) Salgueirino-Maceira, V.; Hoppe, C. E.; Correa-Duarte, M. A. J. Phys. Chem. B 2007, 111, 331-334. (e) Yi, H.; Song, H.; Chen, X. Langmuir 2007, 23, 3199-3204. (10) Addadi, L.; Moradian, J.; Shay, E.; Maroudas, N. G.; Weiner, S. Proc. Natl. Acad. Sci. U.S.A. 1987, 84, 2732-2736.

Ford et al. with a layer of Parafilm and left at room temperature. Formation of a precipitate in the vial was evident within 30 min. After 1.5-2.5 h, the vial was removed from the beaker, capped, and stored at room temperature. To compact the precipitate, the sample was centrifuged (5000 rpm for 5 min), yielding a black solid and colorless supernatant. To remove excess salts, the colorless supernatant was removed, and the solid was suspended in water (1 mL) by vortex mixing and again compacted by centrifugation. This washing step was repeated once. The SWNT content in the resulting composite was estimated as 0.35% by weight by assuming complete conversion of Ca2+ into CaCO3 and complete incorporation of the SWNTs into the composite.11 Preparation of CNT-CaCO3 Composites by Mixing CaCl2 and Na2CO3 Solutions. This procedure began with the precipitation of uniform, nearly spherical microparticles of CaCO3 by colloidal crystallization from a supersaturated solution, as described by Volodkin et al.12 In a typical preparation, a solution of sodium carbonate (0.33 M, 100 µL) in water was added to a solution of calcium chloride (0.33 M, 100 µL) in water, while being stirred with a vortex mixer, in a polypropylene vial, producing a milky suspension of CaCO3 microparticles. After 1 min, a solution of UA-SWNTs in water (6 µL) was added to the suspension while being stirred with a vortex mixer. The approximate concentration of SWNTs before being mixed with the calcium chloride solution was 1700-1800 mg/L, based on the absorbance of the solution at 800 nm.8 After 10-20 min, the suspension was compacted by centrifugation (5000 rpm for 10 min), yielding a black solid and colorless supernatant. To remove excess salts from the composite, the supernatant was removed, and the solid was suspended in water (200 µL) using an ultrasonic bath briefly (approximately 2 s) and collected by centrifugation (5000 rpm for 10 min). This washing process was repeated once. Alternatively, composites were prepared with the order of mixing of the sodium carbonate and calcium chloride solutions being reversed from what was stated previously. To prepare the sample for Fourier transform infrared (FTIR) and X-ray diffraction (XRD) analyses, after the composite was washed with water, it was washed with ethanol (200 µL) to remove the water and dried under ambient conditions. The mass of the dried powder obtained was 2.4 ( 0.1 mg.13 The SWNT content in the resulting composite was estimated as 0.32% by weight using the same assumptions as stated previously. Composites were prepared analogously with U-SWNTs or U-MWNTs having similar nanotube contents and with U-SWNTs having a 30 times greater CNT content.11 Measurements. SEM images of the solids on conductive substrates (F-doped SnO2-coated glass (FTO), silicon, or gold-coated glass) were obtained using a Leo Gemini 1530 field emission microscope operating with a beam voltage of 4 or 5 kV and a working distance of 4-5 mm. No coating of the samples with a conductive layer was applied. FTIR and XRD analyses were performed by Dr. Thomas Hatzl (Consulting Geologist, Pu¨rgen, Germany). The FTIR analysis was performed with a pellet made from the CNT-CaCO3 composite powder (1.0 mg) and KBr (199 mg), using a Mattson 3020 spectrometer (4000-400 cm-1). XRD analysis of the CNT-CaCO3 composite powder on a glass objective carrier was carried out with a Philips PW-Basis-X-ray diffractometer (adjusting parameters: (11) Estimated CNT contents (wt %) in the composites are subject to uncertainties in the masses of both the CaCO3 and the CNT components. Uncertainty in the mass of CaCO3 arises from the fact that it is slightly soluble in water. Uncertainty in the mass of CNTs arises from two factors: (1) the CNT concentrations of the aqueous solutions used to prepare the composites are uncertain because the extinction coefficient at 800 nm is an approximate one and (2) an unknown proportion of the CNTs exist outside the composite particles as films (mats) rather than being integrated into the particles. (12) (a) Volodkin, D. V.; Petrov, A. I.; Prevot, M.; Sukhorukov, G. B. Langmuir 2004, 20, 3398-3406. (b) Sukhorukov, G. B.; Volodkin, D. V.; Gu¨nther, A. M.; Petrov, A. I.; Shenoy, D. V.; Mo¨hwald, H. J. Chem. Mater. 2004, 14, 2073-2081. (c) Volodkin, D. V.; Larionova, N. I.; Sukhorukov, G. B. Biomacromolecules 2004, 5, 1962-1972. (d) Petrov, A. I.; Volodkin, D. V.; Sukhorukov, G. B. Biotechnol. Prog. 2005, 21, 918-925. (13) This mass is less than the predicted mass of 3.3 mg calculated assuming that Ca2+ was completely converted into CaCO3. The difference is assumed to be due to the fact that crystalline as well as amorphous forms of CaCO3 are slightly soluble in water.

Microcrystalline CNT-CaCO3 Composite Particles

Langmuir, Vol. 24, No. 7, 2008 3481

Figure 2. SEM image of the surface of a U-SWNT-CaCO3 composite particle prepared by mixing solutions of CaCl2 and Na2CO3. The SWNTs (bundles) appear in either black or white contrast due to charging effects. The round particles are impurities associated with the CNTs. The calculated CNT content in the composite is 0.32 wt %. Magnification: 40 000×.

Figure 1. SEM image of U-SWNT-CaCO3 composite particles prepared by the vapor diffusion method and deposited onto a silicon substrate. The CNTs are mainly visible as films on the sides of the particles and between them. The films extend onto the substrate and are responsible for the dark gray domains. The even darker patches are due to impurities associated with the CNTs. The calculated CNT content in the composite is 0.35 wt % (see text). (A) Magnification: 300× and (B) magnification: 1000×. detector, Cu; divergence blend, 1/12°; receiving aperture, 0.2; and goniometer speed, 3°/min). Graphical interpretation of the diffraction diagrams was carried out manually, corresponding to a petrographic standard procedure.

Results SWNT-CaCO3 Composites Obtained by Diffusion of (NH4)2CO3 Vapor into a CaCl2 Solution. Precipitates appeared within 30 min after exposing solutions of the SWNTs and CaCl2 to vapor from (NH4)2CO3. Although CaCl2 alone can cause the nanotubes to precipitate, no precipitation was evident under the conditions employed until exposure to (NH4)2CO3. Precipitation of the nanotubes was complete (>95%), based on UV-vis analysis of the supernatant. The precipitates obtained in some cases were comprised of mainly individual, well-formed rhombohedral particles with sizes in the 10 µm range, such as those shown in the SEM image in Figure 1, where the darker patches are films of nanotubes (mats). In general, however, the precipitates obtained by this method had less regular morphologies than seen in Figure 1, although they still exhibited rhombohedral features. CNT-CaCO3 Composites Obtained by Mixing CaCl2 and Na2CO3 Solutions. This procedure involved the rapid mixing of equal volumes of equimolar (0.33 M) aqueous solutions of

CaCl2 and Na2CO3 to form CaCO3.12 To avoid precipitation of CNTs by either the CaCl2 or the Na2CO3 solutions prior to CaCO3 formation, aqueous solutions of the CNTs were added 1 min after precipitating the CaCO3, before it had crystallized into calcite. In general, composite particles that were kept wet for 1 day or longer were rhombohedral with edge lengths in the range of 5-30 µm. There was no obvious dependence of the shape or size on the sequence used to precipitate the CaCO3 (i.e., adding the CaCl2 solution to a stirred Na2CO3 or vice versa). Experiments in which nonstoichiometric (by 10%) amounts of CaCl2 and Na2CO3 were used yielded similar results. Composite particles obtained with MWNTs were less regular in shape and had rougher surfaces as compared to those with SWNTs (Figure S1, Supporting Information). Figure 2 is an SEM image of the surface of an as-prepared rhombohedral particle demonstrating the intermingling of SWNTs with the CaCO3 matrix. The nanotubes14 are visible as fibers whose contrast appears as either dark or light depending on whether they are inside or outside the CaCO3 phase, respectively, due to charging phenomena.15 The SWNT-CaCO3 composite that was dried immediately after being isolated by centrifugation and washed with water, within about 30 min after precipitation of CaCO3, was mainly a mixture of raspberry-shaped and rhombohedral particles, as shown in the SEM image in Figure 3A. Comparison of the image with the one acquired 2 weeks earlier at the same position on the substrate (Figure S2A, Supporting Information) shows that the particle morphology did not change when stored under dry conditions (ambient air). These prepared particles were comprised of 70-80% calcite and 5-10% vaterite according to FTIR analysis (Figure S2B, Supporting Information), which was confirmed by XRD.16 In contrast to the particles stored under dry (14) SWNTs visible in this figure and other SEM images are bundles of nanotubes with diameters of ca. 10 nm. (15) Homma, Y.; Suzuki, S.; Kobayashi, Y.; Nagase, M.; Takagi, D. Appl. Phys. Lett. 2004, 84, 1750-1752. (16) FTIR and XRD analyses are considered to be qualitative due to the small quantities of samples. The FTIR spectra in Figures S2B and S3B have similar absorbances in the 2800-3000 cm-1 range due to C-H stretching vibrations, even though the composite represented by Figure S2B was prepared with 30 times less SWNTs than the composite represented by Figure S3B. These two samples are not directly comparable, however, because the SWNTs used in the latter case were U-SWNTs while in the former case they were UA-SWNTs and thus contained anisaldehyde condensates with the amido and ureido groups (ref 8).

3482 Langmuir, Vol. 24, No. 7, 2008

Figure 3. SEM images of UA-SWNT-CaCO3 composite particles prepared by mixing solutions of CaCl2 and Na2CO3 and depositing onto an FTO-glass substrate. The calculated CNT content in the composite is 0.32 wt %. (A) Particles were stored dry on the substrate for 2 weeks before this image was acquired. Magnification: 2000×. (B) Same batch of particles was stored wet for 2 weeks before being deposited onto the substrate. Magnification: 2000×.

conditions, those that were under wet conditions (in contact with water) for 2 weeks were entirely rhombohedral (Figure 3B). In an experiment with a smaller (by a factor of 7.5) proportion of SWNTs relative to CaCO3 than the subsequent ones, particles that were isolated and dried within a few minutes after the solutions of CaCl2 and Na2CO3 were mixed were comprised of porous spheres (∼10 µm in diameter) together with particles exhibiting rhombohedral features; the nanotubes were visible by SEM imaging on the surfaces of the spherical particles as fuzzy bright lines, but contrast was poor due to electron charging (Figure S3, Supporting Information). An experiment with a larger (by a factor of 30) proportion of CNTs (U-SWNTs) relative to CaCO3 was also performed. After the composite was precipitated and washed, it was kept in contact with water for 6 days before being dried and analyzed by FTIR and XRD. The FITR analysis showed that the sample consisted of mainly calcite (∼95 wt %).16 The XRD analysis confirmed that the major component was calcite. No vaterite was detected by either FTIR or XRD. SEM examinations of the composite 1, 3, and 6 days after preparation indicated that the particle morphology obtained after 1 day was stable. The SEM images showed the particles to be rhombohedral with edge lengths of ∼10, ∼22, and ∼30 µm; the particles were coated and connected by dense films of nanotubes (Figure S4, Supporting Information).

Ford et al.

Figure 4. SEM images of fractured U-SWNT-CaCO3 composite particles. The particles were prepared by mixing solutions of CaCl2 and Na2CO3, depositing onto an FTO substrate, and then fracturing with forceps. The calculated CNT content in the composite is 0.32 wt %. (A) Image of the side of a particle showing a layered structure with SWNTs (bundles) pointing outward. Magnification: 30000×. (B) Image of adjacent fragments of a particle showing SWNTs (bundles) bridging the gap between them. Magnification: 20 000×.

Interestingly, the faces of the larger-sized particles appeared to be comprised of either four (∼22 µm) or five (∼30 µm) of the smaller (∼10 µm) ones. Fractured Particles. To address the question as to whether the CNTs are located inside as well as outside of the particles, they were also examined after being mechanically crushed on the substrate. The resulting SEM images provide clear evidence of CNTs being inside the particles, whether they were prepared by vapor diffusion (Figure S5, Supporting Information) or by mixing solutions of CaCl2 and Na2CO3 (Figure 4 and Figure S6, Supporting Information). The layered nature of the composite is evident in these images. Acid-Treated Particles. Treating the composite particles with acids (e.g., HCl solutions or vapor) caused the CaCO 3 component to decompose, leaving shells of the CNTs. In some cases, the shells were thick and rigid enough to maintain the shape of the particle upon drying (Figure 5). In most cases, however, the shells collapsed, leaving mats of randomly intertwined CNTs that often had well-defined edges and shapes derived from the parent particles (Figure 6 and Figure S7, Supporting Information). The thickness of these mats from composites with SWNTs was typically that of one to several SWNT bundles (i.e., in the range of 10-100 nm). The CNTs in mats derived from composites

Microcrystalline CNT-CaCO3 Composite Particles

Langmuir, Vol. 24, No. 7, 2008 3483

Figure 5. (A) SEM image of U-SWNT-CaCO3 composite particles prepared by the vapor diffusion method and deposited onto a gold substrate. The CNTs are mainly visible as films on the sides of the particles and between them. The films extend onto the substrate. The dark patches are due to impurities associated with the CNTs. The calculated CNT content in the composite is 0.32 wt %. Magnification: 1500×. (B) An image of the sample following treatment with 1 M HCl solution to remove the CaCO3 component. Magnification: 1400×. The field of view in this image is approximately the same as the one in panel A.

Figure 6. (A) SEM image of U-SWNT-CaCO3 composite particles prepared by mixing solutions of CaCl2 and Na2CO3 and depositing onto an FTO-glass substrate. The calculated CNT content in the composite is 0.35 wt %. Magnification: 1000×. (B) An image of the sample following treatment with 1 M HCl solution to remove the CaCO3 component. The dark patches are films of the CNTs from shells (microcapsules) that collapsed upon drying. Magnification: 1000×.

with MWNTs appeared to be more loosely packed (Figure S8, Supporting Information). Because of the hygroscopic nature of CaCl2, CNT mats remaining on substrates after treatment with HCl gas collected water from the atmosphere and tended to diffuse away from their original positions. By drying the samples immediately after a short exposure to HCl gas, the mats adhered well enough to the substrate to allow the CaCl2 to subsequently be removed in a water bath. In this way, it was possible to place the composite particles in particular locations on substrates, such as between electrodes, and to remove the CaCO3 to provide CNT mats in the same locations (see Supporting Information for details). Adsorption of Soluble CNTs to Calcite Powder. The previous experiments suggested that soluble CNTs adsorb strongly to CaCO3. As a further test, experiments were also performed by mixing aqueous solutions of the SWNTs with powders of precipitated calcium carbonate. Most of the nanotubes were adsorbed and were not removed by washing (see Supporting Information for details). SEM images showed that the washed powders were comprised of loose networks of nanotubes adsorbed to and interlinking CaCO3 crystals (Figure S9, Supporting Information).

It is clear from the present study that CNTs can be incorporated inside calcite crystals, as well as on their surfaces, by precipitating CaCO3 together with water-soluble CNTs. The two methods that we used to precipitate CaCO3 gave qualitatively similar results, but our work focused on the crystalline composites obtained by the mixing of solutions of CaCl2 and Na2CO3 since this process provides better experimental control and has been more intensively studied in the literature12,17 than the vapor diffusion method. The stoichiometric precipitation of CaCO3 by rapidly mixing equal volumes of equimolar aqueous solutions of CaCl2 and Na2CO3 results in the immediate formation of amorphous calcium carbonate (ACC), which crystallizes within a few minutes to a mixture of kinetically stable vaterite and thermodynamically stable calcite in the temperature range of 14-30 °C.17 The thermodynamically unstable vaterite subsequently transforms into calcite predominantly through a solution-mediated mechanism.17 Gen-

Discussion

(17) (a) Kralj, D.; Brecevic, L.; Kontrec, J. J. Cryst. Growth 1997, 177, 248257. (b) Kitamura, M. J. Colloid Interface Sci. 2001, 236, 318-327. (c) Kawano, J.; Shimobayashi, N.; Kitamura, M.; Shinoda, K.; Aikawa, N. J. Cryst. Growth 2002, 237-239, 419-423. (d) Andreassen, J.-P. J. Cryst. Growth 2005, 274, 256-264. (e) Shen, Q.; Wei, H.; Zhou, Y.; Huang, Y.; Yang, H.; Wang, D.; Xu, D. J. Phys. Chem. B 2006, 110, 2994-3000.

3484 Langmuir, Vol. 24, No. 7, 2008

Ford et al.

Figure 7. Left: schematic representation of the formation of CNT-CaCO3 composite particles by mixing solutions of CaCl2 and Na2CO3, and the formation of CNT shells (microcapsules) by removing the CaCO3 component. The structure indicated by I is a metastable porous spherical particle of CaCO3 to which the CNTs are adsorbed (II) before CaCO3 transforms to give a composite particle comprised of the CNTs and the thermodynamically stable calcite phase of CaCO3 (III). Subsequent removal of CaCO3 with acid generates a shell of the CNTs (IV), which often collapses into a mat upon drying on a substrate (V). Right: schematic representation of possible hydrogen-bonding interactions between a ureido group appended to the wall of a CNT and the hydrated surface of calcite.

erally, the vaterite and calcite crystals have spherical and rhombohedral shapes, respectively, while the transformation of vaterite into calcite yields polycrystalline particles as intermediates with more complex shapes.17e The shapes, degrees of aggregation, and rates of transformation of the CaCO3 particles depend on various experimental factors, including the concentrations of the CaCl2 and Na2CO3 solutions and the rate at which they are mixed.12,17 The conditions that we employed (i.e., 0.33 M concentrations and rapid mixing) were determined by Volodkin and co-workers to be optimal for producing individual porous vaterite microspheres with an average diameter of around 5 µm.12 Our results are in accord with the literature cited previously and thus indicate that the CNTs do not significantly influence the morphology or phase transformation kinetics of CaCO3. The reason for the difference between our results and those reported previously for CNT-CaCO3 composites7 is not clear. The proportion of CNTs relative to CaCO3 in most of our samples was relatively low (10% by weight.18 Li and Gao attributed the stabilization of spherical crystals of vaterite in their composites to the crystals being embedded and covered by the CNTs.7c However, calcite was the predominant phase in our composites even when they were prepared with a 30 times higher proportion (18) We estimated the wt % CNTs from the data presented in ref 7c by assuming complete conversion of CaCl2 into CaCO3.

of CNTs than usual, so steric effects do not appear to be influential in the present composites.11 Natural and synthetic organic molecules and macromolecules with amido groups are known to influence the nucleation and growth of CaCO3 crystals.19 The fact that the CNTs in our experiments have no apparent influence on the CaCO3 phase behavior implies no influence of the urea-derived amido and ureido groups that are appended to the CNTs. However, these groups may promote interactions between CNTs and CaCO3, thereby facilitating blending of the two materials in the resulting composite particles. A good indication that such interactions can occur is the observation that mixing aqueous solutions of CNTs with CaCO3 (calcite) powder results in rapid, nearly quantitative adsorption of the CNTs (Supporting Information). While the details of the interaction are not known, they presumably include hydrogen bonding between amido and ureido groups on the CNTs and water molecules in the hydration layer on the CaCO3 surface.20 On the basis of the previous discussion, the stages in the formation of crystalline CNT-CaCO3 composite particles are (19) (a) Wei, H.; Shen, Q.; Zhao, Y.; Wang, D.-J.; Xu, D.-F. J. Cryst. Growth 2003, 250, 516-524. (b) Jimenez-Lopez, C.; Rodriguez-Navarro, A.; DominguezVera, J. M.; Garcia-Ruiz, J. M. Geochim. Cosmochim. Acta 2003, 67, 16671676. (20) (a) Stipp, S. L. S. Geochim. Cosmochim. Acta 1999, 63, 3121-3131. (b) Magdans, U.; Torrelles, X.; Angermund, K.; Gies, H.; Rius, J. Langmuir 2007, 23, 4999-5004.

Microcrystalline CNT-CaCO3 Composite Particles

shown schematically in Figure 7. The CNTs are initially adsorbed to the surface of spherical particles of vaterite. When these particles subsequently transform into rhombohedral calcite particles, the CNTs become incorporated internally as well as externally. This result contrasts with that of Li and Gao, who found that carboxy-functionalized CNTs were excluded from the CaCO3 phase when composites with vaterite transformed into calcite.7c Our results may differ due to more favorable interactions between the CNTs and the CaCO3 matrix that is provided by the amido and ureido groups on the CNTs; our system may be more comparable to calcite crystals grown in hydrogel media, wherein the polymeric gelator molecules, gelatin or agarose, become incorporated into the calcite matrix.21 When the calcite matrix is decomposed with acid, a shell, or microcapsule, of CNTs is formed whose shape matches that of the original composite particle (Figure 7). We believe that the shell initially is comprised of the CNTs located at the surface of the particle. CNTs that are freed from the CaCO3 matrix as it decomposes diffuse to the shell or are carried there by the evolved CO2 bubbles, and are trapped, thereby becoming incorporated into the shell wall. Light microscopic observations of free-floating CNTs after complete removal of CaCO3 from rhombohedral particles showed them to be shell-like with the shapes of the parent particles, although the shells had burst open presumably due to the generation of gaseous CO2 during decomposition of CaCO3 (Figure S10, Supporting Information). (21) (a) Grassmann, O.; Mu¨ller, G.; Lo¨bmann, P. Chem. Mater. 2002, 14, 4530-4535. (b) Estroff, L. A.; Addadi, L.; Weiner, S.; Hamilton, A. D. Org. Biomol. Chem. 2004, 2, 137-141. (c) Li, H.; Estroff, L. A. J. Am. Chem. Soc. 2007, 129, 5480-5483.

Langmuir, Vol. 24, No. 7, 2008 3485

These shells, which are comprised of mats of intertwined nanotubes, are flexible and tend to collapse upon drying (Figure 7). This method of preparing CNT shells is simpler than previously reported methods.9 As noted in the Introduction, the composites described herein are of interest for several possible biomedical applications. We also envision possible applications for the CNT shells derived from the composites as, for example, microelectrode supports for bone-forming cells or neurons.1,3

Conclusion Two methods of precipitating calcium carbonate together with water-soluble CNTs yielded composite particles having a crystalline CaCO3 matrix, mainly calcite, with the CNTs incorporated into the matrix as well as on the particle surface. The mechanism by which the composite particles are formed likely involves the interaction of CaCO3 with amido and ureido groups on the CNTs. Removal of the CaCO3 matrix from the composite particles by treating them with acid results in the formation of shells (microcapsules) of randomly intertwined CNTs with shapes derived from the parent particles. The composite particles or CNT shells derived from this method have potential biomedical applications. Supporting Information Available: Additional experimental details and SEM and optical microscopic images, as discussed in the text. This material is available free of charge via the Internet at http://pubs.acs.org. LA7026797