Rapid preparation of high surface area iron oxide and alumina nanoclusters through a soft templating approach of sol–gel precursors

Abstract

A facile, one-pot and organic-solvent free strategy for the synthesis of Fe₂O₃/Fe₃O₄ and Al₂O₃ nanoparticles has been explored. Our synthesis combines the simple preparation of small particles from an epoxide assisted, sol–gel method with the rapid through-put of soft-template processing to yield highly porous metal oxides. The role of the supporting dextran template is investigated and is found to be crucial in tuning surface area and morphological shape. In all, we show this process to be a simple, low-cost approach that could be amended to other metal systems. The resulting porous Fe and Al materials was investigated using gas adsorption/desorption analysis, scanning electron microscopy and transmission electron microscopy. The annealed materials were analyzed using powder X-ray diffraction.

---

Introduction

Synthesis of porous metal oxides with high surface area is of particular interest due to their applicability in catalysis,^1–4 adsorption,^5–7 energy storage,⁸ optics,^9,10 electrochemistry,^11,12 capacitance¹³ and conductivity.¹⁴ With the current increasing energy costs and decreasing resources, substantial research effort has focused on developing more cost efficient and less time consuming methods for the preparation of high surface-area metal oxides, such as thermal degradation,¹⁵ polymer-matrix assisted synthesis,¹⁶ spray-pyrolysis,^17,18 and surfactant-mediated syntheses.^19,20 Nevertheless, many methods that produce high surface areas suffer from slow turn-over and difficulty in scalability. For example, the preparation of aerogels via low pressure/temperature critical point drying of sol–gel precursors affords highly porous nanomaterials because of the preservation of the small colloidal particles.^21–24 Regrettably, this approach entails days of washing the gels and solvent-exchange with liquid CO₂. In addition, the critical point drying itself poses a challenge for scalability, since high-throughput would require large autoclaves capable of withstanding high pressures.

In this study, we report efforts to circumvent these practical disadvantages by replacing the critical point drying of sol–gel precursors with direct annealing of the sol–gel in the presence of dextran as a sacrificial template. Ordinarily, the annealing of a solvent-filled gel results in severe loss of surface area due to collapse of the pore network under the capillary forces of the evaporating liquid. Such structural collapse from capillary forces is precisely the reason that aerogels and xerogels are dried of solvent before annealing. Our method still relies on a sol–gel synthesis to afford nanostructured inorganic precursors. However, we prepare the gels in the presence of a template which supports the gel network through the wet-annealing process as the solvent leaves the pores. Once the pores have been evacuated, continued heating removes the template. During this annealing step, the inorganic sol–gel precursors are simultaneously transformed into the desired phase. Elimination of the stepwise solvent exchange, drying, and annealing steps required in traditional aerogel/xerogel preparations greatly accelerates the acquisition of high surface area inorganic metal oxides.

Dextran has been previously used as a soft, sacrificial template due to its clean-burning property; however, the resulting materials possessed low surface areas around 0.5 m² g⁻¹.²⁵ In that work, metal and/or metal oxide particles were formed during the calcination step from their salts in solution with no control over aggregation. Here, by pre-forming the nanostructured colloidal particles by sol–gel synthesis prior to calcination, higher surface areas can be attained. The well-established epoxide-driven polycondensation of metal salt precursors was chosen for the preparation of inorganic colloids because of its proven efficacy in a wide variety of transition metal systems, its simplicity, and its tolerance of a wide range of preparative parameters. In this method, an epoxide is added to a solution of metal salt to promote the steady formation of colloidal oxy–hydroxy metallate polymers. The mechanism has been previously described,²⁶ and relies on the epoxide acting as a proton-sink for the system, thereby irreversibly driving forward the olation and oxolation reactions that form the inorganic polymer. Because the epoxide mediates the polymerization, very small colloidal particles are formed. Iron and aluminum sol gel precursors were chosen as test subjects for this proof-of-concept study in order to establish the scope of the sol–gel/soft template approach, since both iron and aluminum are known to gel at different rates and require different annealing temperatures to attain the desired metal oxide phase.^27–30 Additionally, these metals were chosen because of their ability to undergo the sol–gel transition in water, since dextran is soluble in water but not alcohols. Different amounts of dextran solution were used during the preparation of the gels to determine the effect of the template: colloid ratio over the surface area and porosity of the resulting material.

Experimental

Preparation of iron oxide and alumina nanoclusters

Iron chloride hexahydrate, FeCl₃·6H₂O (Mallinckrodt) and aluminum chloride hexahydrate, AlCl₃·6H₂O (Spectrum) were used as precursors. Propylene oxide (ACROS Organics, 99.6%), glycidol (ACROS Organics, 96%), and Dextran (75 000 MW, Spectrum) were all used as received without further purification.

In a typical experiment 2 mL of a 1 M solution of FeCl₃·6H₂O was mixed with 0, 1, 2, or 4 mL (Fe0, Fe1, Fe2, and Fe4, respectively) of a homogeneous and aqueous solution of dextran (2.5 g dissolved in 8 mL of DI water), followed by addition of 1.39 mL of propylene oxide (PO). The composite was then vortexed for ∼10 s and then set aside to gel undisturbed for 24 h. Gels were obtained within 1–3 hours, with the longest gelation process (18 h) observed when 4 mL of dextran solution was used. The red monolithic gels were then annealed at 500 °C (2 °C min⁻¹) for 3 h in static air to afford opaque red powders for Fe0, Fe1 and Fe2, and red metallic looking flakes for Fe4.

A similar procedure was followed using 2 mL of 2 M solutions of AlCl₃·6H₂O, 0.79 mL of glycidol (GLY), and 0, 1, 2, or 4 mL (Al0, Al1, Al2 and Al4, respectively) of dextran solution. A temperature of 800 °C was used to anneal the samples in order to increase the crystallinity of the material.

Physical characterization

The powder X-ray diffraction patterns of the samples were recorded with a Rigaku Ultima III diffractometer using Cu Kα radiation. To prepare the samples, they were first finely ground before being packed in a zero-background Pt sample holder. The collection was taken at a 2θ range of 20–80° at a step-width of 0.03° s⁻¹. The diffraction patterns were then identified by comparison to the phases in the International Centre for Diffraction Data (ICDD) powder diffraction file (PDF) database.

The surface morphology of the iron and aluminum nanomaterials was studied using a high-resolution scanning electron microscope (SEM) Hitachi S-4300. The powders were mounted on an aluminum stub with carbon tape. The microscope was operated between 2 and 10 kV to minimize charging effects.

Transmission electron microscopy (TEM) analyses were conducted using a Hitachi H-8100 at an accelerating voltage of 150 kV. The samples were prepared by suspending a small amount of the material into 200 proof ethanol, sonicating for 5 min and then placing one drop onto an ultrathin carbon, 400 mesh Cu grid and permitting to dry.

The specific surface area, pore volume and average pore sizes were all obtained from the N₂ adsorption/desorption analyses conducted at 77 K on a NOVA 4200e model Surface area/pore size analyzer (Quantachrome Instrument Corp.). Brunauer–Emmett–Teller (BET) specific surface areas were calculated from data points taken at relative pressures between 0.05 and 0.30 of the adsorption branch of the isotherm. The pore volume, pore sizes and average size distribution was taken from the desorption branch of the isotherm. Prior to physisorption analysis, the samples were placed on degas for 24 h at 80 °C. Each set of measurements was taken with an equilibrium time of 600 s, which resulted in a total experimentation time of 9–18 h per sample.

Results and discussion

Gel formation studies

Preparation of iron oxide and alumina nanoclusters was tested using both PO and GLY as the gelation agent. However, it was observed that a faster gelation time resulted with the Fe/PO and Al/GLY combinations, thus propylene oxide and glycidol were used to synthesize the Fe and Al oxide materials, respectively, reported in this work.

As previously reported, the nature of the epoxide affects the morphology and surface area of the resulting material.¹³ However, its effect on these systems was not investigated since the primary purpose of this study was to establish a streamlined, rapid synthetic methodology. It was also found that the optimum amount of dextran solution to use for the formation of Fe and Al gels was 4 mL, as larger amounts hinder the gelation process, probably due to the excess dilution of the metal salt and the amount of water present in the system. A temperature of 800 °C was used to anneal the Al samples in order to obtain the desired metal oxide phase.²⁸

Powder X-ray diffraction

The crystallinity of the annealed samples was investigated using powder X-ray diffraction (PXRD). Fig. 1 shows the PXRD pattern observed for the different iron-containing annealed samples with highly crystalline reflections corresponding mainly to Fe₂O₃ (PDF# 97-008-2135). Additionally, there are several other peaks corresponding to the formation of Fe₃O₄ (PDF# 97-009-8088) that become more apparent in Fe4 when 4 mL of dextran solution is used. This observation is in accordance with the amount of dextran present in Fe4, as more reductive aldehyde groups are present compared to Fe1 and Fe2, causing in situ reduction of the metal ions.

Fig. 1 Powder X-ray diffraction patterns for the annealed iron oxide samples (squares denote Fe₂O₃ while circles indicate Fe₃O₄).

The average particle size as indexed for the (104) reflection of Fe₂O₃ was found to be ∼19 nm for Fe0, Fe1 and Fe2 using Scherrer’s equation, while the average crystallite size for the (110) reflection of Fe₃O₄ was found to be ∼36 nm.

The diffraction pattern for the annealed aluminum samples (Fig. 2) shows consistently weak reflections that can be best assigned as Al₂O₃ (PDF# 00-046-1131).

Fig. 2 Powder X-ray diffraction patterns for the annealed alumina samples.

Electron microscopy

Fig. 3A and B illustrates the morphological differences between Fe0 and Fe4. In general, the iron specimens increased in roughness with increasing dextran content. Additionally, the SEM images of all the Fe specimens exhibit varied microstructures, each including domains of spongy texture and of well-faceted crystallites (ESI,† Fig. S1). Such morphological heterogeneity of the iron specimens is not surprising, given the high annealing temperature (500 °C) required to remove the dextran template, relative to typical annealing temperatures used for iron aerogel materials (300 °C).¹ The fine, spongy texture is consistent with typical sol–gel microstructures, while the well-facetted crystallites arise as a result of the high annealing temperature, which promotes grain growth.

Fig. 3 (A–D): SEM images of Fe0, Fe4, Al0 and Al4, left-to-right, respectively. (E–H): TEM images of Fe0, Fe4, Al0 and Al4, left-to-right, respectively.

Fig. 3C illustrates the SEM images of Al0, showing a homogeneous surface in both morphology and composition that could be described as a brittle, finely textured foam. This smooth morphology reflects the fact that Al₂O₃ is a refractory material, impervious to grain growth even at 800 °C. On the other hand, Al4 (Fig. 3D) shows a rather creased and roughened texture, as a trend that is observed when larger amounts of dextran solution have been used for the aluminium specimen. The retained roughness of the specimens with high template loading indicates that large quantities of template are necessary to counteract the densification introduced by the evaporation of water from the system. This is, perhaps, intuitive given the large volume percent of solvent in the precursor gels that is lost upon calcination. Backscatter secondary electron (BSE) imaging at low and high magnification showed that both Fe and Al specimens were compositionally homogeneous and clean of by-products, indicative that the dextran was fully degraded under the annealing conditions (ESI,† Fig. S2).

The internal morphology was examined using TEM. It was found that sample Fe0 (Fig. 3E) contains clusters of spherical particles. Upon the addition of 4 mL of dextran solution (Fig. 3F), the particles began to take on a cubic type structure with crystallite size increasing (>30 nm), which correlates to the PXRD crystallite size as determined by Scherrer's equation. As expected, the TEM of the aluminum samples further illustrates the amphorous nature of the materials (Fig. 3G and H). Upon the addition of dextran solution from 0–4 mL, there is no real change in the internal morphology since no real morphological shapes can be observed.

Nitrogen adsorption/desorption analyses

The N₂ adsorption–desorption isotherms of the Fe and Al (ESI,† Fig. S3 and S4, respectively) specimens exhibit H3 and H2 types hysteresis, respectively. The BET surface areas and pore radii of the iron oxide shows a direct dependence with the amount of dextran solution used during the synthesis (Table 1). Accordingly, the surface area of the iron samples shows a significant increase when using 0 to 4 mL of dextran.

Table 1 Physical properties of the annealed iron and aluminum samples

Sample	BET surface area (m^2 g^−1)	BJH pore volume (cm^3 g^−1)	BJH pore radius (nm)
Fe0/Al0	14/174	0.16/0.54	17.70/3.25
Fe1/Al1	21/222	0.15/0.65	10.64/4.28
Fe2/Al2	33/231	0.14/0.66	7.75/4.26
Fe4/Al4	73/242	0.18/0.44	1.78/3.66

---

As expected, with a larger amount of dextran, pore collapse of the gel network is partially counteracted which results in less densification and higher retained surface areas. The pore volume of all the samples averages 0.16 cm³ g⁻¹, while the average pore radii decrease upon increasing the amount of dextran. This decrease in pore size is most probably an indication of the material undergoing a sintering process with reduction of the average pore size due to the repacking of the material to fill inter-particle voids left by the dextran.

Gas sorption analyses of the Al nanoparticles also show an increasing trend with surface areas that range between 174 and 242 m² g⁻¹, when using 0 to 4 mL of dextran. No apparent change was observed in the pore volume and pore radius with respect to the amount of dextran used. Compared with the Fe₂O₃, an average annealing temperature of 800 °C was used for the Al specimens, which may have caused the materials to densify and collapse uniformly, resulting in comparable pore size values among the Al₂O₃ samples. Encouragingly, the surface area for Al4 (242 m² g⁻¹) approaches the surface area attained in the case of an alumina aerogel (289 m² g⁻¹) after annealing under similar conditions.²⁸

Overall, the differences in the isotherms of the Fe (H3 hysteresis) and Al (H2 hysteresis) samples can be then attributed to different repacking processes that occurs to the Fe and Al specimens when annealed at different temperatures. Accordingly, when the Fe samples are annealed at 500 °C, a sintering process may occur based on the physisorption data, while densification of the Al materials occurs at 800 °C, thus causing different pore structures.

While there is room for improvement, these preliminary results demonstrate the merit of this approach of single-step drying and annealing of sol–gels with template supports.

The importance of dextran in this synthetic method was determined by the preparation of control Fe and Al samples in the absence of template (Fe0 and Al0, respectively). The resulting red and white powders exhibited surface areas of 14 and 174 m² g⁻¹ for Fe and Al, respectively, which further confirms the role of dextran as a support matrix that prevents increased collapse of the gel network. Similarly, control samples were also prepared in the absence of epoxide. The resulting Fe and Al specimens also illustrate decreased surface areas (7 and 82 m² g⁻¹ for Fe and Al, respectively) compared to any of the materials reported herein where formation of a gel preceded the annealing process. Hence, the use of the epoxide creates nanoscale colloidal precursors in the presence of the dextran, which results in the formation of fine structures prior to the annealing process and affords higher surface areas.

Conclusions

In conclusion, this method provides significant benefits over existing synthetic procedures for the preparation of porous metal oxides with high surface area. The pre-formation of colloidal particles provides higher surface areas than attained for sacrificial templating of solutions of metal salt precursors. Evaluation of PXRD and BSE images indicates that the samples are homogenously the desired phase and clean of by-products without the need for time consuming, waste-generating washing procedures. Additionally, it is observed by TEM that varying the amounts of dextran, affects the internal shape of the material. Even though the surface areas are lower than those of the analogous aerogels materials, this approach is advantageous due to lower cost, time and the amenability to scale-up.

Acknowledgements

For the financial support provided by the Office of Naval Research (Grant #: N00014-1101-0424), the authors are exceedingly grateful.

Notes and references

1. S. Bali, F. E. Huggins, G. P. Huffman, R. D. Ernst, R. J. Pugmire and E. M. Eyring, Energy Fuels, 2009, 23, 14.
2. Z. Yu, O. Borg, D. Chen, B. C. Enger, V. Froseth, E. Rytter, H. Wigum and A. Holmen, Catal. Lett., 2006, 109, 43.
3. C. L. Carnes and K. J. Klabunde, J. Mol. Catal. A: Chem., 2003, 194, 227.
4. M. Davis, W. M. Hikal, C. Gümeci and L. J. Hope-Weeks, Catal. Sci. Technol., 2012, 2, 922.
5. J. S. Hu, L. S. Zhong, W. G. Song and L. J. Wan, Adv. Mater., 2008, 20, 2977.
6. W. Cai, J. Yu, C. Anand, A. Vinu and M. Jaroniec, Chem. Mater., 2011, 23, 1147.
7. W. Cai, J. Yu and M. Jaroniec, J. Mater. Chem., 2010, 20, 4587.
8. A. Chidembo, S. H. Aboutalebi, K. Konstantinov, M. Salari, B. Winton, S. A. Yamini, I. P. Nevirkovetsa and H. K. Liuab, Energy Environ. Sci., 2012, 5, 5236.
9. X. D. Wang, C. J. Summers and Z. L. Wang, Nano Lett., 2004, 4, 423.
10. T. Gao, Q. Li and T. Wang, Chem. Mater., 2005, 17, 887.
11. D. R. Rolison and B. Dunn, J. Mater. Chem., 2001, 11, 963.
12. W. Deng, X. Ji, Q. Chena and C. E. Banks, RSC Adv., 2011, 1, 1171.
13. M. Davis, C. Gumeci, B. Black, C. Korzeniewski and L. Hope-Weeks, RSC Adv., 2012, 2, 2061.
14. M. Takashima, S. Yonezawa, K. Horita, K. Ohwaki and H. Takahashi, J. Mater. Chem., 1996, 6, 795.
15. S.-W. Bian, J. Baltrusaitis, P. Galhotra and V. H. Grassian, J. Mater. Chem., 2010, 20, 8705.
16. S. F. Liu, C. Y. Wu and X. Z. Han, Chinese J. Inorg. Chem., 2003, 19, 624.
17. A. I. Y. Tok, F. Y. C. Boey and X. L. Zhao, J. Mater. Process. Technol., 2006, 178, 270.
18. O. Milosevic, B. Jordovic and D. Uskokovic, Mater. Lett., 1994, 19, 165.
19. Y. Wang, C. Ma, X. Sun and H. Li, Inorg. Chem. Commun., 2002, 5, 751.
20. Y. D. Wang, C. L. Ma, X. D. Sun and H. D. Li, Inorg. Chem. Commun., 2002, 5, 751.
21. S. G. Hattenbergst, J. Non-Cryst. Solids, 1986, 115.
22. Y. P. Gao, C. N. Sisk and L. J. Hope-Weeks, Chem. Mater., 2007, 19, 6007.
23. A. Khaleel and A. Al-Marzouqi, Mater. Lett., 2012, 68, 385.
24. M. Davis, C. Gumeci, C. Kiel and L. J. Hope-Weeks, J. Sol-Gel Sci. Technol., 2011, 58, 535.
25. D. Walsh, L. Arcelli, T. Ikoma, J. Tanaka and S. Mann, Nat. Mater., 2003, 2, 386.
26. A. E. Gash, T. M. Tillotson, J. H. Satcher, J. F. Poco, L. W. Hrubesh and R. L. Simpson, Chem. Mater., 2001, 13, 999.
27. A. E. Gash, J. H. Satcher Jr and R. L. Simpson, Chem. Mater., 2003, 15, 3268.
28. T. F. Baumann, A. E. Gash, S. C. Chinn, A. M. Sawvel, R. S. Maxwell and J. H. Satcher Jr, Chem. Mater., 2005, 17, 395.
29. L. Durãesa, A. Moutinhoa, I. J. Seabraa, H. C. de Sousaa and A. Portugala, Mater. Chem. Phys., 2011, 130, 548.
30. W. Dong, S. Wu, D. Chen, X. Jiang and C. Zhu, Chem. Lett., 2000, 496.

---

Footnote

† Electronic supplementary information (ESI) available: Additional electron microscopy and gas sorption data. See DOI: 10.1039/c2nj40781g

---