Availableonlineatwww.sciencedirect.com ° ScienceDirect surface science ELSEVIER Applied Surface Science 254(2008)4925-4929 www.elsevier.com/locate/apsusc Electrospun ceramic fibers: Composition, structure and the fate of precursors R.W. Tuttle a, A. Chowdury a, E. T. Bender b, R D. Ramsier a, ., c, k, J.L. Rapp M.P. Espe of Physics, Ayer Hall, 302 Buchtel Common, The University of Akron, Akron, OH 44325-4001, US Department of Chemistry, Knight Chemical Laboratory, 302 Buchtel Common, The University of Akron, Akron, OH 44325-3601, USA The Institute for Teaching and Leaming, Leigh Hall, 302 Buchtel Common, The University of Akron, Akron, OH 44325-6236, USA Received 19 June 2007: accepted 27 December 2007 Available online 7 February 2008 Abstract Fibers are electrospun from aluminum acetate/polymer mixtures and characterized by an array of techniques before and after annealing at 1200C. We demonstrate that sodium and boron present in the initial starting materials as adducts and stabilizers remain incorporated into the resulting fibers after annealing and pyrolysis of the host polymer. The influence of these minor constituents on the surfaces of the fibers is suggested by infrared and X-ray photoelectron spectroscopic data. The presence of these species may impact potential chemical applications of small diameter ceramic fibers, such as their use as catalytic supports or for chemical decomposition C 2008 Elsevier B V. All rights reserved. Keywords: Electrospinning: Fibers; Ceramics: NMR; XPS; SEM; FTIR 1. Introduction aspect ratio of individual fibers[19-24, as well as their relative orientation within the as-formed nanofiber structures [25-28] The field of electrospinning has witnessed a recent upsurge even for polymer-only systems. In the case of metal oxide in interest, particularly involving the production of metal oxide nanofibers, there are often additional complications. Most of nanofibers [1-18]. There has been significant effort on these materials require heating to pyroline the polymer and tanium-containing materials systems in our laboratory form the desired ceramic crystal structure, so heating and 14,5, 14, 17] and others [2, 7, 16], as well as on fibers formed cooling rates, and annealing temperatures and atmospheres, are from aluminum [1, 6, 18] and zirconium [9, 12] precursors. also variables that need to be controlled. In the work reported However, there are a wide variety of other ceramic nanofibers here, we focus on electrospinning solutions containing being produced in laboratories around the world, such as In3 O2 aluminum acetate. Since our end-use goal is to form [15], to name only a few. These materials have many potential are most interested in the fate of the precursor constituents a [3], wO3 [8], Cuo [101, NaCo2O4 [11, BaTiO3[13] and Sio2 alumina-based catalytic or chemically reactive structures, applications in such areas as photovoltaics, catalysis, photo- hemistry, sensors and photonics. They are studied using 2. Experimental different types of characterization techniques, but predomi- nantly by electron micro scopy and X-ray diffraction 2.1. Electrospinning It is well established that electrospinning parameters such as on viscosity, applied voltage, gap distance and collector Aluminum acetate stabilized with boric acid ( sTREM netry significantly affect the resulting morphology and chemicals), polyvinylpyrrolidone(PVP, Sigma-Aldrich), and absolute ethanol(Pharmco Products) were used as received The concentration of the boric acid adduct present in the Hall, 302 Buchtel Common, The University of Akron, Akron, OH 44325-6236, aluminum acetate was approximately 2.7wt % and there was USA.Tel:+13309728834;fax:+13309725728 also sodium present at approximately 2 wt %. For comparison E-mail address: rex(@ uakron edu(R D. Ramsier) to the synthesized fibers, aluminum oxide powder(a-Al2O sevier B. v. All rights reserved
Electrospun ceramic fibers: Composition, structure and the fate of precursors R.W. Tuttle a , A. Chowdury a , E.T. Bender b , R.D. Ramsier a,b,c, *, J.L. Rapp b , M.P. Espe b a Department of Physics, Ayer Hall, 302 Buchtel Common, The University of Akron, Akron, OH 44325-4001, USA b Department of Chemistry, Knight Chemical Laboratory, 302 Buchtel Common, The University of Akron, Akron, OH 44325-3601, USA c The Institute for Teaching and Learning, Leigh Hall, 302 Buchtel Common, The University of Akron, Akron, OH 44325-6236, USA Received 19 June 2007; accepted 27 December 2007 Available online 7 February 2008 Abstract Fibers are electrospun from aluminum acetate/polymer mixtures and characterized by an array of techniques before and after annealing at 1200 8C. We demonstrate that sodium and boron present in the initial starting materials as adducts and stabilizers remain incorporated into the resulting fibers after annealing and pyrolysis of the host polymer. The influence of these minor constituents on the surfaces of the fibers is suggested by infrared and X-ray photoelectron spectroscopic data. The presence of these species may impact potential chemical applications of small diameter ceramic fibers, such as their use as catalytic supports or for chemical decomposition. # 2008 Elsevier B.V. All rights reserved. Keywords: Electrospinning; Fibers; Ceramics; NMR; XPS; SEM; FTIR 1. Introduction The field of electrospinning has witnessed a recent upsurge in interest, particularly involving the production of metal oxide nanofibers [1–18]. There has been significant effort on titanium-containing materials systems in our laboratory [4,5,14,17] and others [2,7,16], as well as on fibers formed from aluminum [1,6,18] and zirconium [9,12] precursors. However, there are a wide variety of other ceramic nanofibers being produced in laboratories around the world, such as In3O2 [3], WO3 [8], CuO [10], NaCo2O4 [11], BaTiO3 [13] and SiO2 [15], to name only a few. These materials have many potential applications in such areas as photovoltaics, catalysis, photochemistry, sensors and photonics. They are studied using different types of characterization techniques, but predominantly by electron microscopy and X-ray diffraction. It is well established that electrospinning parameters such as solution viscosity, applied voltage, gap distance and collector geometry significantly affect the resulting morphology and aspect ratio of individual fibers [19–24], as well as their relative orientation within the as-formed nanofiber structures [25–28], even for polymer-only systems. In the case of metal oxide nanofibers, there are often additional complications. Most of these materials require heating to pyrolize the polymer and form the desired ceramic crystal structure, so heating and cooling rates, and annealing temperatures and atmospheres, are also variables that need to be controlled. In the work reported here, we focus on electrospinning solutions containing aluminum acetate. Since our end-use goal is to form alumina-based catalytic or chemically reactive structures, we are most interested in the fate of the precursor constituents. 2. Experimental 2.1. Electrospinning Aluminum acetate stabilized with boric acid (STREM chemicals), polyvinylpyrrolidone (PVP, Sigma–Aldrich), and absolute ethanol (Pharmco Products) were used as received. The concentration of the boric acid adduct present in the aluminum acetate was approximately 2.7 wt.%, and there was also sodium present at approximately 2 wt.%. For comparison to the synthesized fibers, aluminum oxide powder (a-Al2O3, www.elsevier.com/locate/apsusc Available online at www.sciencedirect.com Applied Surface Science 254 (2008) 4925–4929 * Corresponding author at: The Institute for Teaching and Learning, Leigh Hall, 302 Buchtel Common, The University of Akron, Akron, OH 44325-6236, USA. Tel.: +1 330 972 8834; fax: +1 330 972 5728. E-mail address: rex@uakron.edu (R.D. Ramsier). 0169-4332/$ – see front matter # 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.apsusc.2007.12.068
R.W. Tutle et al./Applied Surface Science 254(2008)4925-4929 <10 um, Sigma-Aldrich) was also used as received. The 2.5. X-ray diffraction(XRD) aluminum acetate solution was made with a 5.0 g: 10 ml: 13 ml ratio of aluminum acetate, water and ethanol and kept overnight X-ray diffraction patterns were collected with a Philips to let the solution become homogeneous. A polymer solution PW1710-based instrument using a copper anode normally was made with a 1.0 g: 10 ml ratio of PVP to ethanol. These operating at 40 kV and 35 mA. Samples were ground to a solutions were combined to yield a solution with a 1: I molar powder then mixed with acetone to make a paste that would ratio of aluminum acetate and pvp monomer adhere to a substrate. A sample-spinner was not used The electrospinning apparatus consisted of aluminum foil wrapped around a conducting plate that was held at ground 2.6. X-ray photoelectron spectroscopy(XPS) potential. The PVP/aluminum acetate solution was placed into a pipette and a copper wire was placed into the solution and XPS measurements were performed using a VG ESCALAB connected to a power supply which was set to a potential in the Mk II system under high vacuum conditions. The aluminum range of 25-30 kV. The resulting electrospun fibers were anode on a dual anode(Mg/Al)X-ray source was used at a removed from the aluminum foil and annealed under ambient power of 180 W, and the analyzer had a fixed transmission tmosphere For the annealing process, the samples were placed energy of 100 eV. in a room temperature furnace, which was then slowly ramped to 1200C. The fibers spent a total of 15 min at 1200C( the 3. Results entire heating process was 2 h in duration), and then were removed from the oven and allowed to cool in air An SEM image of fibers spun from a water/ethanol lution of PvP and aluminum acetate is shown in Fig 1(A 2.2. Scanning electron microscopy (SEM) and an image after annealing is shown in Fig. 1(B) The diameters of the fibers following annealing at 1200C SEM was performed on a JOEL JEM-5310 instrument appear, on average, smaller than the as-spun fibers. However, perating at 25 kV and 60 mA. A sample of the fibers was the aspect ratios are reduced significantly, and the fibrous placed on double-sided conducting carbon tape and attached to material is considerably more brittle and delicate than the a sample holder. It was then coated, via sputtering with a thin as-spun materials. The decrease in the flexibility of the layer of Pd alloy, to reduce charging effects fibers may result from a high fraction of crystallinity 2.3. Solid-state nuclear magnetic resonance spectroscopy MR) 27Al spectra were collected on a Varian Unityplus 750 MHz (176T) spectrometer using a Doty Scientific DSI-971 narrow bore cross-polarization magic-angle spinning CP/MAS) probe. The samples were packed into 4 mm silicon nitride rotors with aurum end caps. One pulse Bloch decay experiments were performed with a spinning speed of 13 kHz and the- Al chemical shifts were referenced to an aluminum hexahydrate solution(0 ppm). TheC spectra were collected on a Varian Unityplus-200(4.7 T)spectrometer using a Doty Scientific variable temperature magic-angle spinning (VTMAS) Probe Cross-polarization and magic-angle spinning were used to obtain the C spectra, using a 2 ms cross (B) polarization time and a spinning speed of 5 kHz. The chemical shifts were referenced to HMB(17.3 ppm for th methyl carbons)and the samples were packed into 7 mm silicon nitride rotors with Kel-F end caps 2.4. Diffuse reflectance infrared spectroscopy(DRIFTS) Fourier transform infrared spectroscopy was perform Bruker IFS 66v/s spectrometer, using a Harrick "praying mantis"diffuse reflectance accessory. A broadband mercury- cadmium-telluride detector. SiC source and a KBr beam platter were used. The spectra were recorded using 2 cm- resolution and 1024 scans were averaged. All experiments were Fig. I SEM images of (A)as-spun aluminum acetate/PVP fiber, and(B) done under vacuum, with a pressure of less than 5 mbar. electrospun fibers after 1200C annealing
<10 mm, Sigma–Aldrich) was also used as received. The aluminum acetate solution was made with a 5.0 g:10 ml:13 ml ratio of aluminum acetate, water and ethanol and kept overnight to let the solution become homogeneous. A polymer solution was made with a 1.0 g:10 ml ratio of PVP to ethanol. These solutions were combined to yield a solution with a 1:1 molar ratio of aluminum acetate and PVP monomer. The electrospinning apparatus consisted of aluminum foil wrapped around a conducting plate that was held at ground potential. The PVP/aluminum acetate solution was placed into a pipette and a copper wire was placed into the solution and connected to a power supply which was set to a potential in the range of 25–30 kV. The resulting electrospun fibers were removed from the aluminum foil and annealed under ambient atmosphere. For the annealing process, the samples were placed in a room temperature furnace, which was then slowly ramped to 1200 8C. The fibers spent a total of 15 min at 1200 8C (the entire heating process was 2 h in duration), and then were removed from the oven and allowed to cool in air. 2.2. Scanning electron microscopy (SEM) SEM was performed on a JOEL JEM-5310 instrument operating at 25 kV and 60 mA. A sample of the fibers was placed on double-sided conducting carbon tape and attached to a sample holder. It was then coated, via sputtering with a thin layer of Pd alloy, to reduce charging effects. 2.3. Solid-state nuclear magnetic resonance spectroscopy (NMR) 27Al spectra were collected on a Varian Unityplus 750 MHz (17.6T) spectrometer using a Doty Scientific DSI-971 narrow bore cross-polarization magic-angle spinning (CP/MAS) probe. The samples were packed into 4 mm silicon nitride rotors with aurum end caps. One pulse Bloch decay experiments were performed with a spinning speed of 13 kHz and the 27Al chemical shifts were referenced to an aluminum hexahydrate solution (0 ppm). The 13C spectra were collected on a Varian Unityplus-200 (4.7 T) spectrometer using a Doty Scientific variable temperature magic-angle spinning (VTMAS) probe. Cross-polarization and magic-angle spinning were used to obtain the 13C spectra, using a 2 ms crosspolarization time and a spinning speed of 5 kHz. The 13C chemical shifts were referenced to HMB (17.3 ppm for the methyl carbons) and the samples were packed into 7 mm silicon nitride rotors with Kel-F end caps. 2.4. Diffuse reflectance infrared spectroscopy (DRIFTS) Fourier transform infrared spectroscopy was performed on a Bruker IFS 66v/s spectrometer, using a Harrick ‘‘praying mantis’’ diffuse reflectance accessory. A broadband mercury– cadmium–telluride detector, SiC source and a KBr beam splitter were used. The spectra were recorded using 2 cm1 resolution and 1024 scans were averaged. All experiments were done under vacuum, with a pressure of less than 5 mbar. 2.5. X-ray diffraction (XRD) X-ray diffraction patterns were collected with a Philips PW1710-based instrument using a copper anode normally operating at 40 kV and 35 mA. Samples were ground to a powder then mixed with acetone to make a paste that would adhere to a substrate. A sample-spinner was not used. 2.6. X-ray photoelectron spectroscopy (XPS) XPS measurements were performed using a VG ESCALAB Mk II system under high vacuum conditions. The aluminum anode on a dual anode (Mg/Al) X-ray source was used at a power of 180 W, and the analyzer had a fixed transmission energy of 100 eV. 3. Results An SEM image of fibers spun from a water/ethanol solution of PVP and aluminum acetate is shown in Fig. 1(A), and an image after annealing is shown in Fig. 1(B). The diameters of the fibers following annealing at 1200 8C appear, on average, smaller than the as-spun fibers. However, the aspect ratios are reduced significantly, and the fibrous material is considerably more brittle and delicate than the as-spun materials. The decrease in the flexibility of the fibers may result from a high fraction of crystallinity Fig. 1. SEM images of (A) as-spun aluminum acetate/PVP fibers, and (B) electrospun fibers after 1200 8C annealing. 4926 R.W. Tuttle et al. / Applied Surface Science 254 (2008) 4925–4929
R.W. Tutle et al. /Applied Surface Science 254(2008)4925-4929 4927 Fibers annealed at 1200C D Bulk a-alumina Fibers annealed at 1200C Bulk PVp Fibers as-spun ulk aluminum acetate Bulk aluminum acetate 40120100806040200-20 Fig.2.51 MHzC CPMAS NMR spectra of(A)aluminum acetate, ( B) Fig 3. 195 MHZ"Al MAS NMR polyvinylpyrrolidone(PvP). (C)electrospun fibers, and(D)fibers annealed at electrospun fibers.(C)fibers annealed at 1200"C, and ()a-alumina powder. 1200C. All samples were spun at 5 kHz All samples were spun at 13 kHz. The features labeled with an asterisk are pinning side band and/or from cross-linking/sintering between fibers during annealing The Bragg peaks observed in the XRD data of Fig. 5 from Characterization of the fibers by NMR indicates that, the annealed fibers predominantly result from the a-alumina other than fiber formation, there is no additional chemistry phase [33, 34]. However, there are other phases present which occurring during the electrospinning process. TheC solid- contain sodium and boron from the aluminum acetate state NMR spectrum from as-spun fibers shown in Fig. 2(C) precursor. The boron appears in the form of the borate is the same as the weighted sum of the C NMR spectra 2Al2O3-B2O3 31], while the sodium leads to the formation of from aluminum acetate and PVP presented in Fig. 2(A)and the sodium B-alumina phase diaoyudaoite, NaAl1O7 (35 (B), respectively with the exception of DRIFTS, the data described thus far Annealing the fibers at 1200C pyrolizes the polymer and have mainly provided insight into the bulk structure and acetate groups, as confirmed by Fig. 2(D), where the only signal morphology of the electrospun materials. However, it is the observed is from the NMR probe. In addition to the loss of the composition of the surfaces of these fibers that is critical in polymer, annealing at this temperature converts some of the determining their chemical reactivity for our end-use goals aluminum acetate to a-alumina as discussed below Therefore, the elemental composition of the surface has been The fibers have also been studied using-'Al solid-state investigated using XPS NMR, a technique that can readily identify the different Comparing Fig. 6(A)from as-spun materials to that of aluminum coordination environments based on chemical shifts: Fig. 6(B)after annealing, it is evident that there is a substantial 15 ppm for octahedral sites and 70 ppm for tetrahedral sites. reduction in the carbon and nitrogen signatures, as expected he"Al solid-state NMR spectrum from as-spun fibers shown from the pyrolysis of the PVp noted in the NMR and Ir data. in Fig. 3(B)is nearly identical to that of Fig. 3(A) from After annealing, the atomic %o composition determined by XPS aluminum acetate, again indicating that no reactive chemistry occurs during the electrospinning process After annealing, the -Al NMR spectrum from the fibers (Fig. 3(C) is very similar to that from bulk a-alumina (Fig. 3(D), where in both cases nearly all of the aluminum sites In addition to the characterization of the bulk structure of the 3M0 are 6-coordinate with a chemical shift of 15 ppm. Alumina powder as-received studied by DRIFTS In Fig 4(A), as-spun fibers exhibit IR bands of both PVP [29]and aluminum Fibers annealed at 1200.C acetate(neat spectra not shown), consistent with the NMr results discussed above indicating that no reactions occur during electrospinning. After annealing, the fibers produce an Fibers as-spun IR spectrum(Fig 4(B) that differs from a-alumina powder (Fig 4(C) Fiber modes in the 1200-1400 cm region are due to borates [18,30-32], and hydroxyl bands near 3300 cm that 10001500 0002500 300035004000 are not present in the powder appear in the spectrum of the Wavenumber(cm.) annealed fibers. These differences and their potential implie Fig 4. DRIFTS spectra of (A)as-spun fibers, (B)fibers annealed at 1200C, tions will be discussed in Section 4 and(C) alpha alumina powder(commercial)
and/or from cross-linking/sintering between fibers during annealing. Characterization of the fibers by NMR indicates that, other than fiber formation, there is no additional chemistry occurring during the electrospinning process. The 13C solidstate NMR spectrum from as-spun fibers shown in Fig. 2(C) is the same as the weighted sum of the 13C NMR spectra from aluminum acetate and PVP presented in Fig. 2(A) and (B), respectively. Annealing the fibers at 1200 8C pyrolizes the polymer and acetate groups, as confirmed by Fig. 2(D), where the only signal observed is from the NMR probe. In addition to the loss of the polymer, annealing at this temperature converts some of the aluminum acetate to a-alumina as discussed below. The fibers have also been studied using 27Al solid-state NMR, a technique that can readily identify the different aluminum coordination environments based on chemical shifts: 15 ppm for octahedral sites and 70 ppm for tetrahedral sites. The 27Al solid-state NMR spectrum from as-spun fibers shown in Fig. 3(B) is nearly identical to that of Fig. 3(A) from aluminum acetate, again indicating that no reactive chemistry occurs during the electrospinning process. After annealing, the 27Al NMR spectrum from the fibers (Fig. 3(C)) is very similar to that from bulk a-alumina (Fig. 3(D)), where in both cases nearly all of the aluminum sites are 6-coordinate with a chemical shift of 15 ppm. In addition to the characterization of the bulk structure of the fibers, they have also been studied by DRIFTS. In Fig. 4(A), the as-spun fibers exhibit IR bands of both PVP [29] and aluminum acetate (neat spectra not shown), consistent with the NMR results discussed above, indicating that no reactions occur during electrospinning. After annealing, the fibers produce an IR spectrum (Fig. 4(B)) that differs from a-alumina powder (Fig. 4(C)). Fiber modes in the 1200–1400 cm1 region are due to borates [18,30–32], and hydroxyl bands near 3300 cm1 that are not present in the powder appear in the spectrum of the annealed fibers. These differences and their potential implications will be discussed in Section 4. The Bragg peaks observed in the XRD data of Fig. 5 from the annealed fibers predominantly result from the a-alumina phase [33,34]. However, there are other phases present which contain sodium and boron from the aluminum acetate precursor. The boron appears in the form of the borate 2Al2O3–B2O3 [31], while the sodium leads to the formation of the sodium b-alumina phase diaoyudaoite, NaAl11O7 [35]. With the exception of DRIFTS, the data described thus far have mainly provided insight into the bulk structure and morphology of the electrospun materials. However, it is the composition of the surfaces of these fibers that is critical in determining their chemical reactivity for our end-use goals. Therefore, the elemental composition of the surface has been investigated using XPS. Comparing Fig. 6(A) from as-spun materials to that of Fig. 6(B) after annealing, it is evident that there is a substantial reduction in the carbon and nitrogen signatures, as expected from the pyrolysis of the PVP noted in the NMR and IR data. After annealing, the atomic % composition determined by XPS Fig. 2. 51 MHz 13C CP/MAS NMR spectra of (A) aluminum acetate, (B) polyvinylpyrrolidone (PVP), (C) electrospun fibers, and (D) fibers annealed at 1200 8C. All samples were spun at 5 kHz. Fig. 3. 195 MHz 27Al MAS NMR spectra of (A) aluminum acetate, (B) electrospun fibers, (C) fibers annealed at 1200 8C, and (D) a-alumina powder. All samples were spun at 13 kHz. The features labeled with an asterisk are spinning side bands. Fig. 4. DRIFTS spectra of (A) as-spun fibers, (B) fibers annealed at 1200 8C, and (C) alpha alumina powder (commercial). R.W. Tuttle et al. / Applied Surface Science 254 (2008) 4925–4929 4927
R.W. Tutle et al./Applied Surface Science 254(2008)4925-4929 lower aspect ratio coarse fibers after annealing. Their final annealing temperature was 1300C vS. 1200C in our work, and the final fibers in their case appear to be smaller in diameter on average than ours it is unclear what annealing times and heating rates were used in their previous study. Although no spectroscopic analysis or precursor stabilizers/adducts were discussed, the XRD data shown in Ref [1] appear to be pure a- alumina and do not contain the B and borate phases that our fibers exhibit. Therefore. either sodium and boron were not present in their aluminum acetate, or the higher temperature annealing removed these constituents Azad [6] started with aluminum 2, 4-pentane donate mixed with PVP to arrive at very small diameter a-alumina fibers after ATW MV ww\ electrospinning and annealing at 1500C for I h. Energy dispersive spectroscopy(EDS) was used to determine the bulk 1015202530354045505560 Al: o ratio. which was consistent with diffraction data indicating only the a-alumina phase. Recognizing that residual Fig. 5. XRD from electrospun fibers after 1200C annealing. Alpha-phase carbon would be difficult to identify by EDS or XRD, Azad o, umina(a) diffraction peaks dominate, however borate(B)and sodium beta- used laser Raman spectroscopy to demonstrate that carbon was umina(B) phases are also detected. The feature labeled with an asterisk is instrumental artifact completely removed from the fibers due to the high temperature annealing. This is consistent with our NMr results however the lack of signal in our studies may result from is 37.49 Al, 2.7%0 Na, 4.5%0 B, 47.0%O, and 7.8% C.(The significant line broadening due to the presence of a wide variety uncertainty for each of these numbers is approximately 0. 25 of carbon species generated during the annealing process. Note times the corresponding percentage). The atomic percentages that these data are in contrast to our XPS results, which indicate of sodium and boron which would be expected based on the that carbon remains on the fibers. Adventitious carbon from the composition of the starting materials would be 5.5%o and 2.8%, atmosphere, to which only XPS would be sensitive, could be the respectively, assuming that all of the aluminum, sodium, and reason for this difference. However, we have also identified boron present in the starting materials remain in the annealed CO2 trapped within other ceramic nanofibers (4, 5], so the fibers source of the carbon that we identify here in this work could also be from the precursors 4. Discussion Dai et al. [18] used electrospinning to form fibers from aluminum acetate stabilized with boric acid (no mention of Panda and Ramakrishna performed electrospinning from sodium) mixed with PVA. It is interesting to note that their XRD luminum acetate- and aluminum nitrate nano hydrate- data indicate the presence of Al4 B2Og and Alig O33 phases, and containing mixtures with poly vinyl alcohol(PVA)and essentially no a-alumina, after 1200C annealing for 2 h. It is polyethylene oxide polymers(PEo)[1]. Their SEM images unclear as to why our XRD data differ substantially from those in seem to indicate a trend similar to that reported here, namely Ref. [18]. One possibility is that the atomic percentage of boron that higher aspect ratio smooth fibers transform into somewhat in the aluminum acetate boric acid adduct used by Dai et al differed significantly from the starting material that we used, or the annealing conditions were different. The a-alumina phase (Kvv) was only detected in Ref [18]after annealing to 1400C. Finally although the Ftir data presented here after annealing appear to O(KLL) Na( Is) contain the same borate modes shown in Ref [18, our fibers have Fibers Al(2s) a significant hydroxyl component absent in the previous work. annealed at I200℃ n Na(KLL) This may point to the role of sodium in our fibers, or to different annealing conditions as well Our DRIFTS results show that hydroxyl modes not present in a-alumina appear in spectra for fibers annealed at 1200C. El-Hakam and El-Sharkawy 3l] have shown that the presence Fibers as-spun of borates in a-alumina can dramatically increase the surface ity of the material, so it extra"hydroxyl modes present in the annealed fibers are due to an increased surface acidity. However, the hydroxyl modes Binding Energy(ev) petra of the Al4B2Og and Al18B4 O3 Fig. 6.XPS spectra from(A) as-spun fibers, and(B) fibers after annealing at materials made by Dai et al. [18] are different than those 200°C reported in this work(Fig 4(B)). Since sodium has also been
is 37.4% Al, 2.7% Na, 4.5% B, 47.0% O, and 7.8% C. (The uncertainty for each of these numbers is approximately 0.25 times the corresponding percentage). The atomic percentages of sodium and boron which would be expected based on the composition of the starting materials would be 5.5% and 2.8%, respectively, assuming that all of the aluminum, sodium, and boron present in the starting materials remain in the annealed fibers. 4. Discussion Panda and Ramakrishna performed electrospinning from aluminum acetate- and aluminum nitrate nano hydratecontaining mixtures with poly vinyl alcohol (PVA) and polyethylene oxide polymers (PEO) [1]. Their SEM images seem to indicate a trend similar to that reported here, namely that higher aspect ratio smooth fibers transform into somewhat lower aspect ratio coarse fibers after annealing. Their final annealing temperature was 1300 8C vs. 1200 8C in our work, and the final fibers in their case appear to be smaller in diameter on average than ours. It is unclear what annealing times and heating rates were used in their previous study. Although no spectroscopic analysis or precursor stabilizers/adducts were discussed, the XRD data shown in Ref. [1] appear to be pure aalumina and do not contain the b and borate phases that our fibers exhibit. Therefore, either sodium and boron were not present in their aluminum acetate, or the higher temperature annealing removed these constituents. Azad [6] started with aluminum 2,4-pentane dionate mixed with PVP to arrive at very small diameter a-alumina fibers after electrospinning and annealing at 1500 8C for 1 h. Energy dispersive spectroscopy (EDS) was used to determine the bulk Al:O ratio, which was consistent with diffraction data, indicating only the a-alumina phase. Recognizing that residual carbon would be difficult to identify by EDS or XRD, Azad used laser Raman spectroscopy to demonstrate that carbon was completely removed from the fibers due to the high temperature annealing. This is consistent with our 13C NMR results, however the lack of signal in our studies may result from significant line broadening due to the presence of a wide variety of carbon species generated during the annealing process. Note that these data are in contrast to our XPS results, which indicate that carbon remains on the fibers. Adventitious carbon from the atmosphere, to which only XPS would be sensitive, could be the reason for this difference. However, we have also identified CO2 trapped within other ceramic nanofibers [4,5], so the source of the carbon that we identify here in this work could also be from the precursors. Dai et al. [18] used electrospinning to form fibers from aluminum acetate stabilized with boric acid (no mention of sodium) mixed with PVA. It is interesting to note that their XRD data indicate the presence of Al4B2O9 and Al18B4O33 phases, and essentially no a-alumina, after 1200 8C annealing for 2 h. It is unclear as to why our XRD data differ substantially from those in Ref. [18]. One possibility is that the atomic percentage of boron in the aluminum acetate boric acid adduct used by Dai et al. differed significantly from the starting material that we used, or the annealing conditions were different. The a-alumina phase was only detected in Ref.[18] after annealing to 1400 8C. Finally, although the FTIR data presented here after annealing appear to contain the same borate modes shown in Ref.[18], our fibers have a significant hydroxyl component absent in the previous work. This may point to the role of sodium in our fibers, or to different annealing conditions as well. Our DRIFTS results show that hydroxyl modes not present in a-alumina appear in spectra for fibers annealed at 1200 8C. El-Hakam and El-Sharkawy [31] have shown that the presence of borates in a-alumina can dramatically increase the surface acidity of the material, so it is reasonable to propose that these ‘‘extra’’ hydroxyl modes present in the annealed fibers are due to an increased surface acidity. However, the hydroxyl modes reported for the spectra of the Al4B2O9 and Al18B4O33 materials made by Dai et al. [18] are different than those reported in this work (Fig. 4(B)). Since sodium has also been Fig. 5. XRD from electrospun fibers after 1200 8C annealing. Alpha-phase alumina (a) diffraction peaks dominate, however borate (B) and sodium betaalumina (b) phases are also detected. The feature labeled with an asterisk is an instrumental artifact. Fig. 6. XPS spectra from (A) as-spun fibers, and (B) fibers after annealing at 1200 8C. 4928 R.W. Tuttle et al. / Applied Surface Science 254 (2008) 4925–4929
R.W. Tutle et al. /Applied Surface Science 254(2008)4925-4929 shown to significantly alter the hydroxyl modes seen in IR [31 Y Zhang, J.Li,Q Li, L. Zhu, X Liu, X Zhong, J.Meng, X Cao, Scripta spectra of aluminas [36,37, this may explain why the data of Mater.56(2007)409. Fig 4(B)differ from those shown by Dai et al. [18]. (4(a)ET Bender, P. Katta, G.G. Chase, R D Ramsier, Surf Interface Anal he XPs results. it (2006)1252 as if the surface (b) E.T. Bender, P Katta, G.G. Chase, R.D. Ramsier, Surf Interface Anal. concentration of sodium is enhanced while that for boron is 9(2007)374 depleted. However, because the concentrations of sodium and [5 E.T. Bender, P. Katta, us.SJJ. Park, G.G. Chase. R D. Ramsier ting mate rials as reported by the manufacturer Chem. Phys. Lett. 423(2006)302. are only approximate, and because of the relatively large [7 X Lu, QZhao, X Liu, D Ig.A435-436(2006)468 6] A.M. Azad, Mater. Sci. En experimental uncertainty in percentages calculated from XPS w. Zhang, C. Wang, Y. Wei, Macromol Rapid Commun. 27(2006)430. spectra, we cannot determine whether or not the apparent [8] X. Lu, X. Liu, W.Zhang, C. Wang. Y. Wei, J Colloid Interf.Sci.298 discrepancies in the sodium and boron percentages are 2006)996 significant. Nevertheless, these results raise the question as 19.M. Azad, Mater. Lett. 60(2006)67. to whether or not surface segregation of the different elements (11)S. Maensiri, W.Nuansing, Mater.Chem.Phys.99(2006)104 is occurring during annealing. This may be an important issue [121 A.-M. Azad, T. Matthews, J Swary, Mater. Sci Eng B 123(2005)252. relevant for potential applications of these fibers for catalysis or [13] J. Yuh, J.C. Nino, W.M. Sigmund, Mater. Lett. 59(2005)3645 for decomposition of chemical spec [14] V. Tomer, R. Teye-Mensah, J.C. Tokash, N. Stojilovic, W. Kataphinan, E.A. Evans. GG. Chase. R D. Ramsier D. Smith. D.H. Reneker, Solar 5. Summary Energy Mater Solar Cells 85(2005)477 [15] G. Zhang, w. Kataphinan, R. Teye-Mensah, P. Katta, L. Khatri, E.A. Evans. GG. Chase. R D. Ramsier. D.H. Reneker. Mater Sci Eng B 116 In this work we have used bulk and surface characterization 2005)353 techniques to investigate the fate of sodium, boron and carbon [16] D. Li, Y Xia, Nano Lett. 4(2004)933 in fibers produced by electrospinning and annealing from ar [17] R. Teye-Mensah. V. Tomer, w. Kataphinan, J.C. Tokash, N. aluminum acetate/polymer mixture. It is clear that no chemistry GG. Chase. E.A. Evans. R D. Ramsier D. Smith. D.H. Rene Condens. Mater. 16(2004)7557 occurs during the electrospinning process, and that minor [18] H. Dai, J. Gong, H. Kim, D. Lee, Nanotechnology 13(2002)674 constituents(Na and B in this case)in the initial starting [19](a)A L Yarin, S. Koombhongse, D H Reneker, J Appl. Phys. 89(2001 materials can be incorporated into the resulting fibers even after annealing. The possibility of selective segregation of these b)AL Yarin, S. Koombhongse, D H Reneker. J Appl. Phys. 90(2001) 4836. species to the surfaces of the fibers cannot be ruled out, thereby (20) D.H. Reneker, A. Yarin, E. Zussman, S. Koombhongse,W.Kataphinan, potentially impacting the ability of these fibers to catalyze or Polym. Nanofibers ACS Sympos. Series 918(2006)7 engage in reactive chemistry in certain applications. As the field [21] C Wang, C-H. Hsu, J-H Lin, Macromolecules 39(2006)7662 of electrospinning continues to expand into different cera [22] J. Zheng, A. He, J. Li, J r47(2006)7095 materials systems, work of the type reported here will become 23) M E Helgeson, NJ. Wagner, AIChE 3.53(2007)51 more important since it demonstrates that attention to detail [24] C. Wang, w. Zhang, Z.H. Huang. E Y. Yan, Y.H. Su, Pigment Resin Technol.35(2006)278. concerning elemental composition of the surfaces of the fibers [25) P Katta, M. Alessandro, R.D. Ramsier, G.G. Chase, Nano Lett. 4(2004) Is necessary. [26 S. Zhong, W.E. Teo,, X. Zhu, R.w. Beuerman, S Ramakrishna, L.Y.L. Acknowledgements Yung, J. Biomed Mater Res. A 79(2006)456 [27] L M. Bellan, H G. Craighead, J. Vac. Sci. Technol. B 24(2006)3179 [28 M.V. Kakade, S Givens, K Gardner, K.H. Lee, D B Chase, J F. Rabolt, J. The authors would like to thank the National science Am.Chem.Soc.129(2007)2777 Foundation NIRT program(DMI-0403835) for partial support [29] H. Schmiers, J. Friebel, P Streubel, R Hesse, R. Kopsel, Carbon 37(1999) with the Xrd data collection. and to Mr. Vivek Tomer. Dr. W. [30]FH. El Batal. A.H. Ashour, Mater. Chem. Phys. 77(2002)677 [31] S.A. El-Hakam, E.A. Al-Sharkawy, Mater. Lett. 36(1998)167. Tony"Kataphinan, and Dr. Darrell Reneker for their initial [32]G. Lefevre. Adv. Colloid Interf. Sci. 107(2004)109 assistance with this project 33 ICDD card number 43-1484 [34] F.R. Feret, D. Roy. C. Boulanger, Spectrochim. Acta B 55(2000) References 135 ICDD card number 32-1033 l1 P.K. Panda, S. Ramakrishna, J. Mater. Sci. 42(2007)2189 [36]S. Srinivasan, C.R. Narayanan, A.K. Datye, AppL. Catal. A: Gen. 132 [2] K. Nakane, K. Yasuda, T. Ogihara, N. Ogata, S. Yamaguchi, J. Appl. 1995)289 Polym.Sci.104(2007)1232 37 D H. Lee, R.A. Condrate Sr, Mater. Lett. 23(1995)241
shown to significantly alter the hydroxyl modes seen in IR spectra of aluminas [36,37], this may explain why the data of Fig. 4(B) differ from those shown by Dai et al. [18]. From the XPS results, it appears as if the surface concentration of sodium is enhanced while that for boron is depleted. However, because the concentrations of sodium and boron in the starting materials as reported by the manufacturer are only approximate, and because of the relatively large experimental uncertainty in percentages calculated from XPS spectra, we cannot determine whether or not the apparent discrepancies in the sodium and boron percentages are significant. Nevertheless, these results raise the question as to whether or not surface segregation of the different elements is occurring during annealing. This may be an important issue relevant for potential applications of these fibers for catalysis or for decomposition of chemical species. 5. Summary In this work we have used bulk and surface characterization techniques to investigate the fate of sodium, boron and carbon in fibers produced by electrospinning and annealing from an aluminum acetate/polymer mixture. It is clear that no chemistry occurs during the electrospinning process, and that minor constituents (Na and B in this case) in the initial starting materials can be incorporated into the resulting fibers even after annealing. The possibility of selective segregation of these species to the surfaces of the fibers cannot be ruled out, thereby potentially impacting the ability of these fibers to catalyze or engage in reactive chemistry in certain applications. As the field of electrospinning continues to expand into different ceramic materials systems, work of the type reported here will become more important since it demonstrates that attention to detail concerning elemental composition of the surfaces of the fibers is necessary. Acknowledgements The authors would like to thank the National Science Foundation NIRT program (DMI-0403835) for partial support of this project. We are grateful to Mr. Tom Quick for his help with the XRD data collection, and to Mr. Vivek Tomer, Dr. W. ‘‘Tony’’ Kataphinan, and Dr. Darrell Reneker for their initial assistance with this project. References [1] P.K. Panda, S. Ramakrishna, J. Mater. Sci. 42 (2007) 2189. [2] K. Nakane, K. Yasuda, T. Ogihara, N. Ogata, S. Yamaguchi, J. Appl. Polym. Sci. 104 (2007) 1232. [3] Y. Zhang, J. Li, Q. Li, L. Zhu, X. Liu, X. Zhong, J. Meng, X. Cao, Scripta Mater. 56 (2007) 409. [4] (a) E.T. Bender, P. Katta, G.G. Chase, R.D. Ramsier, Surf. Interface Anal. 38 (2006) 1252; (b) E.T. Bender, P. Katta, G.G. Chase, R.D. Ramsier, Surf. Interface Anal. 39 (2007) 374. [5] E.T. Bender, P. Katta, A. Lotus, S.J. Park, G.G. Chase, R.D. Ramsier, Chem. Phys. Lett. 423 (2006) 302. [6] A.-M. Azad, Mater. Sci. Eng. A 435–436 (2006) 468. [7] X. Lu, Q. Zhao, X. Liu, D. Wang, W. Zhang, C. Wang, Y. Wei, Macromol. Rapid Commun. 27 (2006) 430. [8] X. Lu, X. Liu, W. Zhang, C. Wang, Y. Wei, J. Colloid Interf. Sci. 298 (2006) 996. [9] A.-M. Azad, Mater. Lett. 60 (2006) 67. [10] H. Wu, D. Lin, W. Pan, Appl. Phys. Lett. 89 (2006) 133125. [11] S. Maensiri, W. Nuansing, Mater. Chem. Phys. 99 (2006) 104. [12] A.-M. Azad, T. Matthews, J. Swary, Mater. Sci. Eng. B 123 (2005) 252. [13] J. Yuh, J.C. Nino, W.M. Sigmund, Mater. Lett. 59 (2005) 3645. [14] V. Tomer, R. Teye-Mensah, J.C. Tokash, N. Stojilovic, W. Kataphinan, E.A. Evans, G.G. Chase, R.D. Ramsier, D.J. Smith, D.H. Reneker, Solar Energy Mater Solar Cells 85 (2005) 477. [15] G. Zhang, W. Kataphinan, R. Teye-Mensah, P. Katta, L. Khatri, E.A. Evans, G.G. Chase, R.D. Ramsier, D.H. Reneker, Mater. Sci. Eng. B 116 (2005) 353. [16] D. Li, Y. Xia, Nano Lett. 4 (2004) 933. [17] R. Teye-Mensah, V. Tomer, W. Kataphinan, J.C. Tokash, N. Stojilovic, G.G. Chase, E.A. Evans, R.D. Ramsier, D.J. Smith, D.H. Reneker, J. Phys. Condens. Mater. 16 (2004) 7557. [18] H. Dai, J. Gong, H. Kim, D. Lee, Nanotechnology 13 (2002) 674. [19] (a) A.L. Yarin, S. Koombhongse, D.H. Reneker, J. Appl. Phys. 89 (2001) 3018; (b) A.L. Yarin, S. Koombhongse, D.H. Reneker, J. Appl. Phys. 90 (2001) 4836. [20] D.H. Reneker, A. Yarin, E. Zussman, S. Koombhongse, W. Kataphinan, Polym. Nanofibers ACS Sympos. Series 918 (2006) 7. [21] C. Wang, C.-H. Hsu, J.-H. Lin, Macromolecules 39 (2006) 7662. [22] J. Zheng, A. He, J. Li, J. Xu, C.C. Han, Polymer 47 (2006) 7095. [23] M.E. Helgeson, N.J. Wagner, AIChE J. 53 (2007) 51. [24] C. Wang, W. Zhang, Z.H. Huang, E.Y. Yan, Y.H. Su, Pigment Resin Technol. 35 (2006) 278. [25] P. Katta, M. Alessandro, R.D. Ramsier, G.G. Chase, Nano Lett. 4 (2004) 2215. [26] S. Zhong, W.E. Teo, X. Zhu, R.W. Beuerman, S. Ramakrishna, L.Y.L. Yung, J. Biomed. Mater. Res. A 79 (2006) 456. [27] L.M. Bellan, H.G. Craighead, J. Vac. Sci. Technol. B 24 (2006) 3179. [28] M.V. Kakade, S. Givens, K. Gardner, K.H. Lee, D.B. Chase, J.F. Rabolt, J. Am. Chem. Soc. 129 (2007) 2777. [29] H. Schmiers, J. Friebel, P. Streubel, R. Hesse, R. Kopsel, Carbon 37 (1999) 1965. [30] F.H. El Batal, A.H. Ashour, Mater. Chem. Phys. 77 (2002) 677. [31] S.A. El-Hakam, E.A. Al-Sharkawy, Mater. Lett. 36 (1998) 167. [32] G. Lefevre, Adv. Colloid Interf. Sci. 107 (2004) 109. [33] ICDD card number 43-1484. [34] F.R. Feret, D. Roy, C. Boulanger, Spectrochim. Acta B 55 (2000) 1051. [35] ICDD card number 32-1033. [36] S. Srinivasan, C.R. Narayanan, A.K. Datye, Appl. Catal. A: Gen. 132 (1995) 289. [37] D.H. Lee, R.A. Condrate Sr., Mater. Lett. 23 (1995) 241. R.W. Tuttle et al. / Applied Surface Science 254 (2008) 4925–4929 4929