- Email: [email protected]
Sintering and characterization of magnesium oxide macroporous membranes
Available online at www.sciencedirect.com
CERAMICS INTERNATIONAL
Ceramics International 42 (2016) 3317–3321 www.elsevier.com/locate/ceramint
Sintering and characterization of magnesium oxide macroporous membranes Leydi Silva, Roberto R. de Avillezn Departament of Chemical and Materials Engineering, Pontificia Universidade Católica, Rio de Janeiro, Brazil Received 3 July 2015; received in revised form 15 October 2015; accepted 25 October 2015 Available online 30 October 2015
Abstract This paper describes a method for obtaining a magnesium oxide macroporous structure with a homogeneous pore size using a reproducible technique. The proposed method used MgO nanoparticles synthesized by a sol-gel/nanotemplating technique as a starting material. The MgO nanopowder was pressed to form a pellet and sintered. The studied parameters were the pressure employed to make the pellets and the sintering temperature. The samples were characterized by scanning electron microscopy, nitrogen adsorption analyzed using BET and BJH analyses, X-ray diffraction and mercury intrusion porosimetry. MgO nanoparticles with crystallite sizes of approximately 9 nm were synthesized and used to make sintered membranes with 50–150 nm macroporous diameters, total pore areas above 8 m2 =g and open porosities above 42% for the pressure of 173.4 MPa and sintering temperatures of 1173 and 1273 K. & 2015 Elsevier Ltd and Techna Group S.r.l. All rights reserved.
Keywords: A. Sol-gel; Magnesium oxide; Nanoparticles; Sintered membrane; Macroporous
1. Introduction The search for cheaper processes to separate oxygen from the air has been triggered by environmental concerns related to the reforming of methane to produce the synthesis gas and combustion using pure oxygen [1,2]. These processes may require the removal of nitrogen from the flue gas to make the CO2 capture more efficient. Mixed ionic/electronic conductive ceramics have emerged as good candidates; however, these materials must be thin to improve the flux [2], therefore, they are prone to failure by fracture. One proposed solution is the use of a ceramic membrane to support the conducting thin ceramic [3]. The ceramic membrane must have high temperature thermal stability, good mechanical resistance, good chemical stability, long life and good defouling properties [4]. Magnesium oxide has some these properties and is relatively cheap; therefore, it is an excellent material candidate for use as a porous support for thin film oxygen transport membranes used to purify oxygen in industrial processes [5,6]. Indeed, it has already been used as the support membrane for La0.2 Sr0.8 Co O3 x and Ce0.9 Gd0.1 O1.95 δ conductive membranes [7,8]. n
Corresponding author. E-mail addresses: [email protected] (L. Silva), [email protected] (R.R. de Avillez). http://dx.doi.org/10.1016/j.ceramint.2015.10.125 0272-8842/& 2015 Elsevier Ltd and Techna Group S.r.l. All rights reserved.
Porous magnesium oxide support membranes have already been made from the mixtures of MgO powder and an organic material that was eliminated before the sintering process at temperatures above 1250 1C [7,5,6], which are typical of ceramic sintering. Hong et al. [7] and Ramachandran et al. [5] used the exotemplate method [9,4,10] to restrain the sintering coalescence at the very high process temperatures. Hong et al. [7] reported a porosity range from 28% to 55% and pore diameters from 0.7 μm to 2.4 μm for membranes made from a mixture of carbon black and commercial MgO powder, pressed and sintered at 1873 K. Smaller pore sizes were obtained by Ramachandran et al. [5] with similar porosity. They were able to get pore sizes from about 75 nm up to 315 nm, with open porosity of 46.8% for the sintering temperature of 1250 1C. Their porous distribution was clearly bimodal for their lower temperature and changed to a monomodal distribution when the temperature was above 1350 1C. Lipińska-Chwalek [6] used commercial MgO powder to prepare their sintered membranes. They did not report the MgO powder size and used as a binder a mixture of methylhydroxypropylcellulose, hydroxypropylmethylcellulose and water for most membranes, except for a few membranes that they also added bohemite. Their sintering temperatures ranged from 1300 to 1700 1C. Their membranes had a porosity of 36% measured by optical, confocal and scanning electron microscopies. They
3318
L. Silva, R.R. de Avillez / Ceramics International 42 (2016) 3317–3321
observed that the fracture stress was around 45 MPa and almost independent of the sintering temperature, however the room temperature elastic modulus increased for higher sintering temperature. The recent researches on producing MgO membranes have used commercial MgO powder sizes, however, if MgO nanoparticles are used, the sintering may be performed at much lower temperatures as observed by Kleiman and Chaim [11,12]. The coalescence kinetics will be controlled by both time and the initial size of the nanoparticles [12]. Nonetheless, the present authors do not know of any previous study on the sintering of nanoparticles to produce a ceramic MgO support membrane. This paper reports the use of MgO nanoparticles produced by a modified sol-gel method [13] to fabricate sintered MgO membranes with very small pores. Membrane disks were made at two pressures for green densification and four temperatures for sintering. The sintering time was kept constant. The sintered membranes were characterized by scanning electron microscopy (SEM), X-ray diffraction (XRD), textural analysis by nitrogen adsorption (BET) and mercury intrusion porosimetry (MIP). 2. Materials and methods MgO nanoparticles were prepared by a modified sol-gel method that employed polyvinyl alcohol (PVA) as a gelling agent and template of the nanoporous structure [13]. Five grams of PVA were dissolved in 50 mL of water and heated at 373 K for approximately 3 h. Then 5 g of MgðNO3 Þ2 6H2 O, 98% purity ISOFAR, dissolved in 50 mL of water was slowly added to the PVA solution under constant stirring. The system was maintained at the same temperature until it reached rubbery consistency. This final material was calcined at 473 K for one hour followed by a second heat treatment at 873 K for another hour to burn up the polymer residues. The resulting material was a very porous white aggregate of MgO nanoparticles that was easily reduced to a powder. The crystallite size measured by XRD was in the range of 7– 11 nm, which was consistent with the previous research [13]. The membrane was a disk made by compressing the MgO powder without additives in a cylindrical steel matrix with 12.92 mm diameter. Two different pressures were used to obtain the green disks: 173.4 (P1) and 260.1 (P2) MPa. The pressure was maintained for 60 s. The disks were sintered in a tubular furnace under an air atmosphere. Four temperatures were employed: 1173 (T1), 1273 (T2), 1373 (T3) and 1473 (T4) K. The temperature was kept constant within 5 K. The samples were slowly placed inside the hot furnace, heat treated for 1 h and then slowly removed from the furnace to avoid thermal shock. The sintered samples were characterized using scanning electron microscopy (SEM), X-ray diffraction (XRD), nitrogen adsorption at liquid nitrogen temperature using the Brunauer– Emmett–Teller (BET) technique and mercury intrusion porosimetry (MIP). The sintered disks were either polished or fractured before taking SEM images obtained from secondary electrons using a Jeol JSM-6390LV microscope. Phase
identification and crystallite size determination were conducted using X-ray diffraction with a Siemens D-5000 instrument, Cu Kα radiation, a Ni filter, and Bragg–Bretano geometry. Rietveld analyses of the diffraction patterns using the fundamental parameters approach were least-squares fitted using TOPAS 3 software [14]. The surface area and porosity were measured using a gas sorption analyzer (Micromeritics ASAP 2010). The samples were dried at 573 K and weighted before the measurement. High-pressure mercury intrusion porosimetry (Micromeritics AutoPore IV 9500) was performed according to ASTM Standard D4404 [15,16]. The material was previously dried at 373 K for one hour. Thrice distilled mercury was used, and the pressure was gradually increased to 413.5 MPa. 3. Results and discussion Table 1 shows the diameter of the disks after sintering. The 2.9% diameter reduction observed at the lowest sintering temperature shows that sintering had started. As the temperature was increased, the disk diameter was reduced due to the densification of the particles (Table 2). Fig. 1 shows the fractured surface of a disk pressed at the lowest pressure, 173.4 MPa, and sintered at 1273 K. The pores showed open and fully interconnected microstructure, which was consistent with a process between the initial and the intermediate stages of the sintering process [17]. The particle size range was between 200 and 600 nm. Fig. 2 shows the polished surface cross-section of a disk pressed at 173.4 MPa and sintered at 1473 K, which was the highest temperature used. The sintering process seemed to be in the intermediate stage; however, the pore structure was still open and interconnected. Similar results were also obtained by Kleiman and Chaim [11] who studied the annealing of cold isostatically pressed compacts between 600 1C and 900 1C. They observed particle growth that was consistent with surface diffusion, a porous microstructure and sintering that occurred at temperatures one third of the melting point for MgO. However, they did not characterize the porosity, which is important for membrane applications. X-ray diffraction showed that the sintered disks were essentially made of MgO; however, a very small amount of MgO2 was also observed, usually less than 2% mass. Table 3 summarizes the experimental data from the X-ray diffraction, BET analysis and mercury intrusion porosimetry. The lattice parameter of the MgO increased as the sintering temperature was increased. This result suggested that the lower sintering Table 1 Diameter (mm) of the sintered disks as a function of sintering temperature and compressing pressure. Pressure (MPa)
173.4 (P1) 260.1 (P2)
Sintering temperature (K) 1173 (T1)
1273 (T2)
1373 (T3)
1473 (T4)
12.54 12.54
12.18 12.22
11.52 11.64
10.88 11.06
L. Silva, R.R. de Avillez / Ceramics International 42 (2016) 3317–3321 Table 2 Experimental results obtained by X-ray diffraction. Sample name
P1T1 P1T2 P1T3 P1T4 P2T2 P2T3 P2T4
Mass
Cryst. size (nm)
% MgO
% MgO2
98.8 98.7 97.0 98.2 89.7 98.8 98.4
1.2 1.3 3.0 1.8 10.3 1.2 1.6
126 155 169 154 226 158 245
_ Lattice param. ðAÞ
4.2136 4.2157 4.2160 4.2176 4.2133 4.2164 4.2166
Fig. 1. Scanning electron microscopy image of the fracture surface of a disk pressed at 173.4 MPa and sintered at 1273 K (sample P1T2).
Fig. 2. Scanning electron microscopy image of a polished cross-section of a disk pressed at 173.4 MPa and sintered at 1473 K (sample P1T4).
temperatures were not enough to produce defect-free grains. The fitted crystallite sizes were consistent with the temperatures used for sintering. Ab initio calculations showed that the
3319
Table 3 Experimental results obtained from the nitrogen adsorption and desorption isotherms (BET) and mercury intrusion porosimetry. The surface area errors are 1.1%. Sample name
P1T1 P1T2 P1T3 P1T4 P2T2 P2T3 P2T4
BET
Mercury intrusion
Surface area ðm2 =gÞ
Average pore radius (nm)
Surface area ðm2 =gÞ
Average pore radius (nm)
Porosity (%)
– 15.3 12.1 8.53 12.1 9.14 7.66
– 8.6 7.0 9.7 8.5 8.6 9.6
12.5 8.07 6.07 2.77 5.76 – –
71.6 106.3 109.0 136.1 100.3 – –
42.2 47.5 38.0 26.2 35.5 – –
Fig. 3. Nitrogen sorption isotherm for a disk pressed at 260.1 MPa and sintered at 1273 K (sample P2T2).
magnesium peroxide, MgO2 was stable at very high pressures [18]; however, experimental synthesis showed that it was possible to be form MgO2 at normal pressures and slightly above room temperature [19]. Because MgO2 was not present in the nanocrystalline MgO precursor powder [13], it was formed due to the compaction and sintering heat treatment. The nitrogen adsorption and desorption isotherms of all the samples showed a mixing of type II and IV isotherms [20] for wide range of open mesoporous and nonporous or macroporous adsorbents. Fig. 3 show the isotherm for the sample pressed with 260.1 MPa and sintered at 1273 K. This isotherm corresponded to monolayer or multilayer adsorption following the BET mechanism [21]. Kaneko [22] stated that the adsorption isotherm near P=Po ¼ 1 provided important information concerning macropores; however, this isotherm was unreliable for accurate measurements. Kaneko suggested the use of MIP to characterize macroporous solids. Table 3 shows that the surface area decreased as the sintering temperature was increased. The surface area also decreased as the pre-sintering pressure was increased for the same sintering temperature. Both behaviors were expected for the sintering process.
3320
L. Silva, R.R. de Avillez / Ceramics International 42 (2016) 3317–3321
binder, parafin wax as a plasticizer and stearic acid as a dispersant, that was sintered at 1523 K as the lowest temperature. However, two sintering conditions of the proposed method provided BET surface areas greater than those obtained by Ramachandran et al. The open porosity measured by MIP ranged from 26.2 to 47.4%, and the total pore area ranged from 2:777 0:03 m2 =g to 12:5 7 0:1 m2 =g. The samples P1T1 and P1T2 had open porosities greater than 42.2% and total areas greater than 8:077 0:08 m2 =g. The average pore volume may be underestimated because the mercury intrusion method considers only the size of the largest connection (pore channel) from the surface towards the pore [16]. 4. Conclusion Fig. 4. Correlation between the BET surface area and the MIP surface area for the sintered samples.
Fig. 5. Pore size distribution determined by MIP.
MIP was used to measure the pore size distribution. MIP is a method complementary to the BJH method that measures the connected porosity. Indeed, the plot of the BET surface area versus the MIP surface area showed a linear behavior (Fig. 4), which was a strong indication that the surface areas were correlated. Fig. 5 shows the distributions for four samples. The distributions were clearly asymmetric and showed that the pores were below 200 nm for all of the analyzed sintering conditions. The sample P2T2, corresponding to 260.1 MPa and 1273 K, showed an approximately symmetric monomodal distribution with most pores below 100 nm. For the same temperature 1273 K (T2), the greater pre-sintering pressure reduced the pore size after the sintering. The sample P1T1, corresponding to 173.4 MPa and 1173 K, was not symmetric but also had a majority of pores below 100 nm and a wider range of pore sizes. Nonetheless, because the maximum pressure used for mercury injection (413.7 MPa) was almost twice the pre-sintering pressure used to prepare the membranes, it was possible that the MgO grains may have collapsed during the measurements, modifying the observed sizes and distributions of the pores [22]. The BET surface areas obtained by the proposed method and the observed pore sizes were similar to those obtained by Ramachandran et al. [5] for membranes produced from a complex mixture of MgO powder, graphite, a thermoplastic
MgO nanoparticles, with crystallite sizes approximately 9 nm and produced by a facile and reproducible modified sol-gel technique using PVA and magnesium nitrate, were used to fabricate porous sintered membranes with controllable porosity. The sintering temperature was approximately one third of the MgO melting point; therefore, the sintering may be performed in the temperature range from 1173 to 1473 K, which are temperatures easily obtained with typical electrical furnaces. The sintered membranes had pore diameter between 50 and 150 nm, total pore areas above 8 m2/g and open porosities above 42% when using the lowest pressure and the two lowest temperatures. These properties are consistent with the membranes being developed for porous support of thin film oxygen transport membranes but require lower sintering temperatures and fewer preparation steps. Acknowledgment The authors thank the CNPq Agency for the partial support with the Grants 474428/2012-5 and 303379/2012-0. The authors are grateful to Dr. Sonia Letichevisky for fruitful discussions. References [1] H.J.M. Bouwmeester, Dense ceramic membranes for methane conversion, Catal. Today 82 (2003) 141–150, http://dx.doi.org/10.1016/S0920-5861 (03)00222-0. [2] P.V. Hendriksen, P.H. Larsen, M. Mogensen, F.W. Poulsen, K. Wiik, Prospects and problems of dense oxygen permeable membranes, Catal. Today 56 (2000) 283–295, http://dx.doi.org/10.1016/S0920-5861(99)00286-2. [3] A. Julbe, D. Farrusseng, C. Guizard, Porous ceramic membranes for catalytic reactors overview and new ideas, J. Membr. Sci. 181 (2001) 3–20. [4] R. Del Colle, C.a. Fortulan, S.R. Fontes, Manufacture and characterization of ultra and microfiltration ceramic membranes by isostatic pressing, Ceram. Int. 37 (4) (2011) 1161–1168, http://dx.doi.org/10.1016/j.ceramint.2010.11.0 39 URL 〈http://www.linkinghub.elsevier.com/retrieve/pii/S0272884210005 183〉. [5] D.K. Ramachandran, F. Clemens, a.J. Glasscock, M. Sogaard, A. Kaiser, Tailoring the microstructure of porous MgO supports for asymmetric oxygen separation membranes: optimization of thermoplastic feedstock systems, Ceram. Int. 40 (7) (2014) 10465–10473, http://dx.doi.org/10.1016/j.ceramint.2014.0 3.017.
L. Silva, R.R. de Avillez / Ceramics International 42 (2016) 3317–3321 [6] M. Lipińska-Chwalek, L. Kiesel, J. Malzbender, Mechanical properties of porous MgO substrates for membrane applications, J. Eur. Ceram. Soc. 34 (2014) 2519–2524, http://dx.doi.org/10.1016/j.jeurceramsoc.2014.03.009. [7] L. Hong, X. Chen, Z. Cao, Preparation of a perovskite La0.2Sr0.8CoO3-x membrane on a porous MgO substrate, J. Eur. Ceram. Soc. 21 (2001) 2207–2215, http://dx.doi.org/10.1016/S0955-2219(00)00320-4. [8] D. Ramachandran, M. Søgaard, F. Clemens, J. Gurauskis, A. Kaiser, Fabrication and performance of a tubular ceria based oxygen transport membrane on a low cost MgO support, Sep. Purif. Technol. 147 (2015) 422–430, http://dx.doi.org/10.1016/j.seppur.2015.02.037 URL 〈http:// www.sciencedirect.com/science/article/pii/S138358661500129X〉. [9] M. Bhagiyalakshmi, J.Y. Lee, H.T. Jang, Synthesis of mesoporous magnesium oxide: its application to CO2 chemisorption, Int. J. Greenh. Gas Control 4 (1) (2010) 51–56, http://dx.doi.org/10.1016/j.ijgg c.2009.08.001 URL 〈http://www.linkinghub.elsevier.com/retrieve/pii/ S1750583609000826〉. [10] Y. Suzuki, P.E.D. Morgan, Meso- and Macroporous ceramics by phase separation and reactive sintering, MRS Bull. 34 (August) (2009) 587–591. [11] S. Kleiman, R. Chaim, Thermal stability of MgO nanoparticles, Mater. Lett. 61 (23–24) (2007) 4489–4491, http://dx.doi.org/10.1016/j.matlet.20 07.02.032 URL 〈http://www.linkinghub.elsevier.com/retrieve/pii/ S0167577X07001796〉. [12] Z.Z. Fang, H. Wang, Densification and grain growth during sintering of nanosized particles, Int. Mater. Rev. 53 (6) (2008) 326–352, http://dx.doi.org/10.1179/174328008X353538 URL 〈http:// www.openurl.ingenta.com/content/xref?genre¼article&issn ¼0950-6608 &volume¼53&issue ¼6&spage¼ 326〉. [13] A.R. Bueno, R.F. Oman, P.M. Jardim, N.A. Rey, R.R. de Avillez, Kinetics of nanocrystalline MgO growth by the solgel combustion method, Microporous Mesoporous Mater. 185 (2014) 86–91, http://dx.doi.org/10.1016/j.micr omeso.2013.10.021 URL 〈http://www.sciencedirect.com/science/article/pii/ S1387181113005386〉 〈http://www.linkinghub.elsevier.com/retrieve/pii/ S1387181113005386〉. [14] TOPAS, General Profile and Structure Analysis Software for Powder Diffraction Data, Bruker AXS GmbH, Karlsruhe, Germany. [15] ASTM Standards D4404-10, ASTM D4404-10 standard test method for determination of pore volume and pore volume distribution of soil and rock
[16]
[17]
[18]
[19]
[20]
[21]
[22]
3321
by mercury intrusion porosimetry, in: Annual Book of ASTM Standards, vol. Di, 2010, pp. 1–7 〈http://www.astm.org/Standards/D4404.htm. H. Giesche, Mercury porosimetry: a general (practical) overview, Part. Part. Syst. Charact. 23 (September (2005)) (2006) 9–19, http://dx.doi.org/10.1002/ppsc.200601009. K. Lu, Sintering of nanoceramics, Int. Mater. Rev. 53 (1) (2008) 21–38, http://dx.doi.org/10.1179/174328008X254358 URL 〈http://www. openurl.ingenta.com/content/xref?genre¼ article&issn¼ 0950-6608& volume¼ 53&issue¼ 1&spage¼ 21〉. Q. Zhu, A. Oganov, A. Lyakhov, Novel stable compounds in the MgO system under high pressure, Phys. Chem. Chem. Phys. 15 (2013) 7696–7700, http://dx.doi.org/10.1039/c3cp50678a URL 〈http://www. pubs.rsc.org/EN/content/articlehtml/2013/cp/c3cp50678a〉. A. Kjekshus, T. Rakke, Preparations and properties of magnesium, copper, zinc and cadmium dichalcogenides, Acta Chem. Scand. 33 (1979) 617–620. K.S.W. Sing, D.H. Everett, R.A.W. Haul, L. Moscou, R.A. Pierotti, J. Rouquérol, T. Siemieniewska, Reporting physisorption data for gas/ solid systems with special reference to the determination of surface area and porosity, Pure Appl. Chem. 57 (4) (1985) 603–619. E.P. Barrett, L.G. Joyner, P.P. Halenda, The determination of pore volume and area distributions in porous substances. I. Computations from nitrogen isotherms, J. Am. Chem. Soc. 73 (1948) (1951) 373–380, http://dx.doi.org/10.1021/ja01145a126. K. Kaneko, Determination of pore size and pore size distribution: 1. Adsorbents and catalysts, J. Membr. Sci. 96 (1994) 59–89 〈http://www.sciencedirect.com/science/article/pii/037673889400126X〉.
CERAMICS INTERNATIONAL
Ceramics International 42 (2016) 3317–3321 www.elsevier.com/locate/ceramint
Sintering and characterization of magnesium oxide macroporous membranes Leydi Silva, Roberto R. de Avillezn Departament of Chemical and Materials Engineering, Pontificia Universidade Católica, Rio de Janeiro, Brazil Received 3 July 2015; received in revised form 15 October 2015; accepted 25 October 2015 Available online 30 October 2015
Abstract This paper describes a method for obtaining a magnesium oxide macroporous structure with a homogeneous pore size using a reproducible technique. The proposed method used MgO nanoparticles synthesized by a sol-gel/nanotemplating technique as a starting material. The MgO nanopowder was pressed to form a pellet and sintered. The studied parameters were the pressure employed to make the pellets and the sintering temperature. The samples were characterized by scanning electron microscopy, nitrogen adsorption analyzed using BET and BJH analyses, X-ray diffraction and mercury intrusion porosimetry. MgO nanoparticles with crystallite sizes of approximately 9 nm were synthesized and used to make sintered membranes with 50–150 nm macroporous diameters, total pore areas above 8 m2 =g and open porosities above 42% for the pressure of 173.4 MPa and sintering temperatures of 1173 and 1273 K. & 2015 Elsevier Ltd and Techna Group S.r.l. All rights reserved.
Keywords: A. Sol-gel; Magnesium oxide; Nanoparticles; Sintered membrane; Macroporous
1. Introduction The search for cheaper processes to separate oxygen from the air has been triggered by environmental concerns related to the reforming of methane to produce the synthesis gas and combustion using pure oxygen [1,2]. These processes may require the removal of nitrogen from the flue gas to make the CO2 capture more efficient. Mixed ionic/electronic conductive ceramics have emerged as good candidates; however, these materials must be thin to improve the flux [2], therefore, they are prone to failure by fracture. One proposed solution is the use of a ceramic membrane to support the conducting thin ceramic [3]. The ceramic membrane must have high temperature thermal stability, good mechanical resistance, good chemical stability, long life and good defouling properties [4]. Magnesium oxide has some these properties and is relatively cheap; therefore, it is an excellent material candidate for use as a porous support for thin film oxygen transport membranes used to purify oxygen in industrial processes [5,6]. Indeed, it has already been used as the support membrane for La0.2 Sr0.8 Co O3 x and Ce0.9 Gd0.1 O1.95 δ conductive membranes [7,8]. n
Corresponding author. E-mail addresses: [email protected] (L. Silva), [email protected] (R.R. de Avillez). http://dx.doi.org/10.1016/j.ceramint.2015.10.125 0272-8842/& 2015 Elsevier Ltd and Techna Group S.r.l. All rights reserved.
Porous magnesium oxide support membranes have already been made from the mixtures of MgO powder and an organic material that was eliminated before the sintering process at temperatures above 1250 1C [7,5,6], which are typical of ceramic sintering. Hong et al. [7] and Ramachandran et al. [5] used the exotemplate method [9,4,10] to restrain the sintering coalescence at the very high process temperatures. Hong et al. [7] reported a porosity range from 28% to 55% and pore diameters from 0.7 μm to 2.4 μm for membranes made from a mixture of carbon black and commercial MgO powder, pressed and sintered at 1873 K. Smaller pore sizes were obtained by Ramachandran et al. [5] with similar porosity. They were able to get pore sizes from about 75 nm up to 315 nm, with open porosity of 46.8% for the sintering temperature of 1250 1C. Their porous distribution was clearly bimodal for their lower temperature and changed to a monomodal distribution when the temperature was above 1350 1C. Lipińska-Chwalek [6] used commercial MgO powder to prepare their sintered membranes. They did not report the MgO powder size and used as a binder a mixture of methylhydroxypropylcellulose, hydroxypropylmethylcellulose and water for most membranes, except for a few membranes that they also added bohemite. Their sintering temperatures ranged from 1300 to 1700 1C. Their membranes had a porosity of 36% measured by optical, confocal and scanning electron microscopies. They
3318
L. Silva, R.R. de Avillez / Ceramics International 42 (2016) 3317–3321
observed that the fracture stress was around 45 MPa and almost independent of the sintering temperature, however the room temperature elastic modulus increased for higher sintering temperature. The recent researches on producing MgO membranes have used commercial MgO powder sizes, however, if MgO nanoparticles are used, the sintering may be performed at much lower temperatures as observed by Kleiman and Chaim [11,12]. The coalescence kinetics will be controlled by both time and the initial size of the nanoparticles [12]. Nonetheless, the present authors do not know of any previous study on the sintering of nanoparticles to produce a ceramic MgO support membrane. This paper reports the use of MgO nanoparticles produced by a modified sol-gel method [13] to fabricate sintered MgO membranes with very small pores. Membrane disks were made at two pressures for green densification and four temperatures for sintering. The sintering time was kept constant. The sintered membranes were characterized by scanning electron microscopy (SEM), X-ray diffraction (XRD), textural analysis by nitrogen adsorption (BET) and mercury intrusion porosimetry (MIP). 2. Materials and methods MgO nanoparticles were prepared by a modified sol-gel method that employed polyvinyl alcohol (PVA) as a gelling agent and template of the nanoporous structure [13]. Five grams of PVA were dissolved in 50 mL of water and heated at 373 K for approximately 3 h. Then 5 g of MgðNO3 Þ2 6H2 O, 98% purity ISOFAR, dissolved in 50 mL of water was slowly added to the PVA solution under constant stirring. The system was maintained at the same temperature until it reached rubbery consistency. This final material was calcined at 473 K for one hour followed by a second heat treatment at 873 K for another hour to burn up the polymer residues. The resulting material was a very porous white aggregate of MgO nanoparticles that was easily reduced to a powder. The crystallite size measured by XRD was in the range of 7– 11 nm, which was consistent with the previous research [13]. The membrane was a disk made by compressing the MgO powder without additives in a cylindrical steel matrix with 12.92 mm diameter. Two different pressures were used to obtain the green disks: 173.4 (P1) and 260.1 (P2) MPa. The pressure was maintained for 60 s. The disks were sintered in a tubular furnace under an air atmosphere. Four temperatures were employed: 1173 (T1), 1273 (T2), 1373 (T3) and 1473 (T4) K. The temperature was kept constant within 5 K. The samples were slowly placed inside the hot furnace, heat treated for 1 h and then slowly removed from the furnace to avoid thermal shock. The sintered samples were characterized using scanning electron microscopy (SEM), X-ray diffraction (XRD), nitrogen adsorption at liquid nitrogen temperature using the Brunauer– Emmett–Teller (BET) technique and mercury intrusion porosimetry (MIP). The sintered disks were either polished or fractured before taking SEM images obtained from secondary electrons using a Jeol JSM-6390LV microscope. Phase
identification and crystallite size determination were conducted using X-ray diffraction with a Siemens D-5000 instrument, Cu Kα radiation, a Ni filter, and Bragg–Bretano geometry. Rietveld analyses of the diffraction patterns using the fundamental parameters approach were least-squares fitted using TOPAS 3 software [14]. The surface area and porosity were measured using a gas sorption analyzer (Micromeritics ASAP 2010). The samples were dried at 573 K and weighted before the measurement. High-pressure mercury intrusion porosimetry (Micromeritics AutoPore IV 9500) was performed according to ASTM Standard D4404 [15,16]. The material was previously dried at 373 K for one hour. Thrice distilled mercury was used, and the pressure was gradually increased to 413.5 MPa. 3. Results and discussion Table 1 shows the diameter of the disks after sintering. The 2.9% diameter reduction observed at the lowest sintering temperature shows that sintering had started. As the temperature was increased, the disk diameter was reduced due to the densification of the particles (Table 2). Fig. 1 shows the fractured surface of a disk pressed at the lowest pressure, 173.4 MPa, and sintered at 1273 K. The pores showed open and fully interconnected microstructure, which was consistent with a process between the initial and the intermediate stages of the sintering process [17]. The particle size range was between 200 and 600 nm. Fig. 2 shows the polished surface cross-section of a disk pressed at 173.4 MPa and sintered at 1473 K, which was the highest temperature used. The sintering process seemed to be in the intermediate stage; however, the pore structure was still open and interconnected. Similar results were also obtained by Kleiman and Chaim [11] who studied the annealing of cold isostatically pressed compacts between 600 1C and 900 1C. They observed particle growth that was consistent with surface diffusion, a porous microstructure and sintering that occurred at temperatures one third of the melting point for MgO. However, they did not characterize the porosity, which is important for membrane applications. X-ray diffraction showed that the sintered disks were essentially made of MgO; however, a very small amount of MgO2 was also observed, usually less than 2% mass. Table 3 summarizes the experimental data from the X-ray diffraction, BET analysis and mercury intrusion porosimetry. The lattice parameter of the MgO increased as the sintering temperature was increased. This result suggested that the lower sintering Table 1 Diameter (mm) of the sintered disks as a function of sintering temperature and compressing pressure. Pressure (MPa)
173.4 (P1) 260.1 (P2)
Sintering temperature (K) 1173 (T1)
1273 (T2)
1373 (T3)
1473 (T4)
12.54 12.54
12.18 12.22
11.52 11.64
10.88 11.06
L. Silva, R.R. de Avillez / Ceramics International 42 (2016) 3317–3321 Table 2 Experimental results obtained by X-ray diffraction. Sample name
P1T1 P1T2 P1T3 P1T4 P2T2 P2T3 P2T4
Mass
Cryst. size (nm)
% MgO
% MgO2
98.8 98.7 97.0 98.2 89.7 98.8 98.4
1.2 1.3 3.0 1.8 10.3 1.2 1.6
126 155 169 154 226 158 245
_ Lattice param. ðAÞ
4.2136 4.2157 4.2160 4.2176 4.2133 4.2164 4.2166
Fig. 1. Scanning electron microscopy image of the fracture surface of a disk pressed at 173.4 MPa and sintered at 1273 K (sample P1T2).
Fig. 2. Scanning electron microscopy image of a polished cross-section of a disk pressed at 173.4 MPa and sintered at 1473 K (sample P1T4).
temperatures were not enough to produce defect-free grains. The fitted crystallite sizes were consistent with the temperatures used for sintering. Ab initio calculations showed that the
3319
Table 3 Experimental results obtained from the nitrogen adsorption and desorption isotherms (BET) and mercury intrusion porosimetry. The surface area errors are 1.1%. Sample name
P1T1 P1T2 P1T3 P1T4 P2T2 P2T3 P2T4
BET
Mercury intrusion
Surface area ðm2 =gÞ
Average pore radius (nm)
Surface area ðm2 =gÞ
Average pore radius (nm)
Porosity (%)
– 15.3 12.1 8.53 12.1 9.14 7.66
– 8.6 7.0 9.7 8.5 8.6 9.6
12.5 8.07 6.07 2.77 5.76 – –
71.6 106.3 109.0 136.1 100.3 – –
42.2 47.5 38.0 26.2 35.5 – –
Fig. 3. Nitrogen sorption isotherm for a disk pressed at 260.1 MPa and sintered at 1273 K (sample P2T2).
magnesium peroxide, MgO2 was stable at very high pressures [18]; however, experimental synthesis showed that it was possible to be form MgO2 at normal pressures and slightly above room temperature [19]. Because MgO2 was not present in the nanocrystalline MgO precursor powder [13], it was formed due to the compaction and sintering heat treatment. The nitrogen adsorption and desorption isotherms of all the samples showed a mixing of type II and IV isotherms [20] for wide range of open mesoporous and nonporous or macroporous adsorbents. Fig. 3 show the isotherm for the sample pressed with 260.1 MPa and sintered at 1273 K. This isotherm corresponded to monolayer or multilayer adsorption following the BET mechanism [21]. Kaneko [22] stated that the adsorption isotherm near P=Po ¼ 1 provided important information concerning macropores; however, this isotherm was unreliable for accurate measurements. Kaneko suggested the use of MIP to characterize macroporous solids. Table 3 shows that the surface area decreased as the sintering temperature was increased. The surface area also decreased as the pre-sintering pressure was increased for the same sintering temperature. Both behaviors were expected for the sintering process.
3320
L. Silva, R.R. de Avillez / Ceramics International 42 (2016) 3317–3321
binder, parafin wax as a plasticizer and stearic acid as a dispersant, that was sintered at 1523 K as the lowest temperature. However, two sintering conditions of the proposed method provided BET surface areas greater than those obtained by Ramachandran et al. The open porosity measured by MIP ranged from 26.2 to 47.4%, and the total pore area ranged from 2:777 0:03 m2 =g to 12:5 7 0:1 m2 =g. The samples P1T1 and P1T2 had open porosities greater than 42.2% and total areas greater than 8:077 0:08 m2 =g. The average pore volume may be underestimated because the mercury intrusion method considers only the size of the largest connection (pore channel) from the surface towards the pore [16]. 4. Conclusion Fig. 4. Correlation between the BET surface area and the MIP surface area for the sintered samples.
Fig. 5. Pore size distribution determined by MIP.
MIP was used to measure the pore size distribution. MIP is a method complementary to the BJH method that measures the connected porosity. Indeed, the plot of the BET surface area versus the MIP surface area showed a linear behavior (Fig. 4), which was a strong indication that the surface areas were correlated. Fig. 5 shows the distributions for four samples. The distributions were clearly asymmetric and showed that the pores were below 200 nm for all of the analyzed sintering conditions. The sample P2T2, corresponding to 260.1 MPa and 1273 K, showed an approximately symmetric monomodal distribution with most pores below 100 nm. For the same temperature 1273 K (T2), the greater pre-sintering pressure reduced the pore size after the sintering. The sample P1T1, corresponding to 173.4 MPa and 1173 K, was not symmetric but also had a majority of pores below 100 nm and a wider range of pore sizes. Nonetheless, because the maximum pressure used for mercury injection (413.7 MPa) was almost twice the pre-sintering pressure used to prepare the membranes, it was possible that the MgO grains may have collapsed during the measurements, modifying the observed sizes and distributions of the pores [22]. The BET surface areas obtained by the proposed method and the observed pore sizes were similar to those obtained by Ramachandran et al. [5] for membranes produced from a complex mixture of MgO powder, graphite, a thermoplastic
MgO nanoparticles, with crystallite sizes approximately 9 nm and produced by a facile and reproducible modified sol-gel technique using PVA and magnesium nitrate, were used to fabricate porous sintered membranes with controllable porosity. The sintering temperature was approximately one third of the MgO melting point; therefore, the sintering may be performed in the temperature range from 1173 to 1473 K, which are temperatures easily obtained with typical electrical furnaces. The sintered membranes had pore diameter between 50 and 150 nm, total pore areas above 8 m2/g and open porosities above 42% when using the lowest pressure and the two lowest temperatures. These properties are consistent with the membranes being developed for porous support of thin film oxygen transport membranes but require lower sintering temperatures and fewer preparation steps. Acknowledgment The authors thank the CNPq Agency for the partial support with the Grants 474428/2012-5 and 303379/2012-0. The authors are grateful to Dr. Sonia Letichevisky for fruitful discussions. References [1] H.J.M. Bouwmeester, Dense ceramic membranes for methane conversion, Catal. Today 82 (2003) 141–150, http://dx.doi.org/10.1016/S0920-5861 (03)00222-0. [2] P.V. Hendriksen, P.H. Larsen, M. Mogensen, F.W. Poulsen, K. Wiik, Prospects and problems of dense oxygen permeable membranes, Catal. Today 56 (2000) 283–295, http://dx.doi.org/10.1016/S0920-5861(99)00286-2. [3] A. Julbe, D. Farrusseng, C. Guizard, Porous ceramic membranes for catalytic reactors overview and new ideas, J. Membr. Sci. 181 (2001) 3–20. [4] R. Del Colle, C.a. Fortulan, S.R. Fontes, Manufacture and characterization of ultra and microfiltration ceramic membranes by isostatic pressing, Ceram. Int. 37 (4) (2011) 1161–1168, http://dx.doi.org/10.1016/j.ceramint.2010.11.0 39 URL 〈http://www.linkinghub.elsevier.com/retrieve/pii/S0272884210005 183〉. [5] D.K. Ramachandran, F. Clemens, a.J. Glasscock, M. Sogaard, A. Kaiser, Tailoring the microstructure of porous MgO supports for asymmetric oxygen separation membranes: optimization of thermoplastic feedstock systems, Ceram. Int. 40 (7) (2014) 10465–10473, http://dx.doi.org/10.1016/j.ceramint.2014.0 3.017.
L. Silva, R.R. de Avillez / Ceramics International 42 (2016) 3317–3321 [6] M. Lipińska-Chwalek, L. Kiesel, J. Malzbender, Mechanical properties of porous MgO substrates for membrane applications, J. Eur. Ceram. Soc. 34 (2014) 2519–2524, http://dx.doi.org/10.1016/j.jeurceramsoc.2014.03.009. [7] L. Hong, X. Chen, Z. Cao, Preparation of a perovskite La0.2Sr0.8CoO3-x membrane on a porous MgO substrate, J. Eur. Ceram. Soc. 21 (2001) 2207–2215, http://dx.doi.org/10.1016/S0955-2219(00)00320-4. [8] D. Ramachandran, M. Søgaard, F. Clemens, J. Gurauskis, A. Kaiser, Fabrication and performance of a tubular ceria based oxygen transport membrane on a low cost MgO support, Sep. Purif. Technol. 147 (2015) 422–430, http://dx.doi.org/10.1016/j.seppur.2015.02.037 URL 〈http:// www.sciencedirect.com/science/article/pii/S138358661500129X〉. [9] M. Bhagiyalakshmi, J.Y. Lee, H.T. Jang, Synthesis of mesoporous magnesium oxide: its application to CO2 chemisorption, Int. J. Greenh. Gas Control 4 (1) (2010) 51–56, http://dx.doi.org/10.1016/j.ijgg c.2009.08.001 URL 〈http://www.linkinghub.elsevier.com/retrieve/pii/ S1750583609000826〉. [10] Y. Suzuki, P.E.D. Morgan, Meso- and Macroporous ceramics by phase separation and reactive sintering, MRS Bull. 34 (August) (2009) 587–591. [11] S. Kleiman, R. Chaim, Thermal stability of MgO nanoparticles, Mater. Lett. 61 (23–24) (2007) 4489–4491, http://dx.doi.org/10.1016/j.matlet.20 07.02.032 URL 〈http://www.linkinghub.elsevier.com/retrieve/pii/ S0167577X07001796〉. [12] Z.Z. Fang, H. Wang, Densification and grain growth during sintering of nanosized particles, Int. Mater. Rev. 53 (6) (2008) 326–352, http://dx.doi.org/10.1179/174328008X353538 URL 〈http:// www.openurl.ingenta.com/content/xref?genre¼article&issn ¼0950-6608 &volume¼53&issue ¼6&spage¼ 326〉. [13] A.R. Bueno, R.F. Oman, P.M. Jardim, N.A. Rey, R.R. de Avillez, Kinetics of nanocrystalline MgO growth by the solgel combustion method, Microporous Mesoporous Mater. 185 (2014) 86–91, http://dx.doi.org/10.1016/j.micr omeso.2013.10.021 URL 〈http://www.sciencedirect.com/science/article/pii/ S1387181113005386〉 〈http://www.linkinghub.elsevier.com/retrieve/pii/ S1387181113005386〉. [14] TOPAS, General Profile and Structure Analysis Software for Powder Diffraction Data, Bruker AXS GmbH, Karlsruhe, Germany. [15] ASTM Standards D4404-10, ASTM D4404-10 standard test method for determination of pore volume and pore volume distribution of soil and rock
[16]
[17]
[18]
[19]
[20]
[21]
[22]
3321
by mercury intrusion porosimetry, in: Annual Book of ASTM Standards, vol. Di, 2010, pp. 1–7 〈http://www.astm.org/Standards/D4404.htm. H. Giesche, Mercury porosimetry: a general (practical) overview, Part. Part. Syst. Charact. 23 (September (2005)) (2006) 9–19, http://dx.doi.org/10.1002/ppsc.200601009. K. Lu, Sintering of nanoceramics, Int. Mater. Rev. 53 (1) (2008) 21–38, http://dx.doi.org/10.1179/174328008X254358 URL 〈http://www. openurl.ingenta.com/content/xref?genre¼ article&issn¼ 0950-6608& volume¼ 53&issue¼ 1&spage¼ 21〉. Q. Zhu, A. Oganov, A. Lyakhov, Novel stable compounds in the MgO system under high pressure, Phys. Chem. Chem. Phys. 15 (2013) 7696–7700, http://dx.doi.org/10.1039/c3cp50678a URL 〈http://www. pubs.rsc.org/EN/content/articlehtml/2013/cp/c3cp50678a〉. A. Kjekshus, T. Rakke, Preparations and properties of magnesium, copper, zinc and cadmium dichalcogenides, Acta Chem. Scand. 33 (1979) 617–620. K.S.W. Sing, D.H. Everett, R.A.W. Haul, L. Moscou, R.A. Pierotti, J. Rouquérol, T. Siemieniewska, Reporting physisorption data for gas/ solid systems with special reference to the determination of surface area and porosity, Pure Appl. Chem. 57 (4) (1985) 603–619. E.P. Barrett, L.G. Joyner, P.P. Halenda, The determination of pore volume and area distributions in porous substances. I. Computations from nitrogen isotherms, J. Am. Chem. Soc. 73 (1948) (1951) 373–380, http://dx.doi.org/10.1021/ja01145a126. K. Kaneko, Determination of pore size and pore size distribution: 1. Adsorbents and catalysts, J. Membr. Sci. 96 (1994) 59–89 〈http://www.sciencedirect.com/science/article/pii/037673889400126X〉.