- Email: [email protected]
Synthesis of porous hierarchical geopolymer monoliths by ice-templating
Microporous and Mesoporous Materials 215 (2015) 206e214
Contents lists available at ScienceDirect
Microporous and Mesoporous Materials journal homepage: www.elsevier.com/locate/micromeso
Synthesis of porous hierarchical geopolymer monoliths by ice-templating Elettra Papa a, Valentina Medri a, *, Patricia Benito b, Angelo Vaccari b, Simone Bugani b, Jakub Jaroszewicz c, Wojciech Swieszkowski c, Elena Landi a a b c
National Research Council of Italy, Institute of Science and Technology for Ceramics (CNR-ISTEC), via Granarolo 64, 48018 Faenza, Italy “Toso Montanari” Department of Industrial Chemistry, Alma Mater Studiorum, University of Bologna, Viale Risorgimento 4, 40136 Bologna, Italy Faculty of Materials Science and Engineering, Warsaw University of Technology, 141 Woloska Str., 02-507 Warsaw, Poland
a r t i c l e i n f o
a b s t r a c t
Article history: Received 15 September 2014 Received in revised form 12 May 2015 Accepted 26 May 2015 Available online 9 June 2015
Hierarchical geopolymers with mesoporous matrices and lamellar macro-porosities were produced for the first time by ice-templating (freeze-casting) technique in a water-based solegel system. Starting from an aqueous solution of metakaolin and potassium silicate, geopolymerization was triggered through a maturation step without reaching complete consolidation. Different amounts of water (20, 50, 70 vol.%) were then mixed with the geopolymer paste to induce lamellar ice growth by unidirectional freezing. The consolidation was completed during freeze-casting and drying. Ice-templated hierarchical geopolymers had 53e83% total porosity and N2 BET specific surface area values from 4 to 46 m2 g1, depending on the amount of water added. A broad mesopore Barret Joyner Halenda distribution was detected between 4 and 100 nm with maxima from 5 to 7 nm, while macropores in the 1-to-100 mm range were identified by Hg intrusion porosimetry. Lamellar monoliths (height 25 mm) were produced with a selected composition (50 vol.% of additional water) in order to assess the reproducibility of the ice-templating technique combined with geopolymerization. © 2015 Elsevier Inc. All rights reserved.
Keywords: Ice-templating Freeze-casting Geopolymer Alkali-bonded ceramics Porosity
1. Introduction Geopolymers are alkali-bonded ceramics [1], constituting a family of materials with properties varying among those characteristic of ceramics, cements, zeolites, or refractories, depending on their formulation. Geopolymers were produced by reacting alumino-silicate powders (metakaolins, blast furnace slags, fly ashes, pozzolana, etc.) with an aqueous alkali hydroxide and/or alkali silicate solution [2]. At the atomic scale, the geopolymer amorphous network is made by SiO4 and AlO4 tetrahedra connected by oxygen corners [2]. Recent results indicate that the tetrahedra form SieOeAleO rings of various sizes in the network and endow the geopolymer matrix with ion exchange properties [3]. From some standpoints, geopolymers can be regarded as either the amorphous counterpart or the precursor of crystalline zeolites.
* Corresponding author. Tel.: þ39 0546 699751; fax: þ39 0546 46381. E-mail address: [email protected] (V. Medri). http://dx.doi.org/10.1016/j.micromeso.2015.05.043 1387-1811/© 2015 Elsevier Inc. All rights reserved.
The microstructure of a metakaolin-based geopolymer consists of nano-particulates separated by micro- and mesopores [4,5]. Since geopolymers are intrinsically mesoporous, a hierarchical pore system can be constructed by combining mesopores with macroporosity. Geopolymers could be used to develop porous materials covering pore sizes ranging from a few tenths of nanometers to a few millimeters, as well as a total pore volume from 30% up to 90% [5]. Because of the high accessibility of pores, porous ceramics with three-dimensionally interconnected and distributed open pores (3D structures) are useful as catalysts, catalyst supports, filters, scaffolds, and adsorbents [6,7]. For all the above-mentioned applications, it is absolutely necessary to control the pore size dimension and distribution, as well as the pore amount and structure (shape, morphology, orientation, and surface properties). The required pore dimensions for catalysts and supports range from 1 nm to few tenths of microns, for filtering and purification systems from a few nanometers to a few millimeters, and for dust collecting from a few micrometers to a few millimeters [6]. The geopolymer production process in an aqueous medium makes it possible to obtain a custom-tailored porosity. Different
E. Papa et al. / Microporous and Mesoporous Materials 215 (2015) 206e214
techniques were used in order to meet porosity requirements for various purposes. Water contents in the starting mixtures affected the intrinsic mesoporosity of the geopolymer matrix, since water acted as a pore former during the polycondensation stage [5]. Moreover, ultra-macro-porosity was obtained by direct foaming, using hydrogen peroxide as a blowing agent [8], or exploiting the redox reaction of metallic Al [9] or Si powders [5,10] in an alkaline solution to induce porosity through H2 evolution. The water-based freeze-casting (i.e. ice-templating) is a useful technique for fabricating porous materials with main unidirectionally oriented pores and a high open porosity, where the final component results in a lamellar or laminate-like structure. This technique is particularly interesting for the peculiar structures and properties shown by porous freeze-cast ceramics, thus opening new opportunities in the field of cellular ceramics [11]. For example, unidirectional porous ceramics have a higher permeability than conventional porous ceramics, and are potentially good thermal insulators [6]. The ice-templating technique consists of freezing an aqueous suspension, followed by ice sublimation under reduced pressure. The resulting green body is generally sintered to consolidate the overall architecture as well as to obtain the desired mechanical and functional properties, while maintaining a high level of porosity. The unidirectional channel-like (lamellar) porosity is obtained in the case of unidirectional freezing, where pores are the replica of ice crystals [11e13]. The template can be easily removed through simple thawing and drying; this is a unique and beneficial feature of this method, overcoming the problems inherent in other templating techniques requiring the use of expensive templates removed through either high temperature or extremely high or low pH conditions [14]. The process is environmentally friendly, since it uses water as a removable template, highly versatile, and the resulting structures are highly tunable by changing process conditions. In detail, the final porosity content can be tuned by varying the particle content within the slurry, and the size of the pores is affected by freezing kinetics. Furthermore, since the solidification is often directional, a very anisotropic morphology of pores can be seen in the solidification plane, where porous channels run from the bottom to the top of samples [13]. Ice-templating has been applied to a large variety of materials such as alumina [15,16], hydroxyapatite [17,18], polymeric materials [19], zirconium diborides ultra-high temperature ceramics [20], zeolite monolith [21], and so on. While literature reports ice-templating of colloidal inert ceramic suspensions, the aim of the present study is to apply this technique to a water-based solegel system capable of producing alkalibonded ceramics, i.e. metakaolin-based geopolymers. To the author's best knowledge, this is the first time ice-templating has been applied to geopolymers. Again, it is important to point out that the systems generally used for freeze-casting [13] are colloidal suspensions of inert ceramic particles with the addition of organic dispersants and binders with the aim of fostering ice-lamellar growth [22], and final samples are usually consolidated by sintering. Conversely, geopolymer slurries are based on a sol/gel reactive system with no addition of any organic dispersant or binder, while the consolidation is of a chemical type, thus avoiding any high-temperature thermal treatment. The aim of the present work is to promote the simultaneous formation of geopolymer intrinsic mesoporosity and lamellar macroporosity by unidirectional ice growth, together with a final chemical consolidation. Starting from a geopolymer mixture of metakaolin and potassium silicate aqueous solution, geopolymerization was triggered through a maturation step, but without reaching a complete consolidation.
207
Different amounts of water were then mixed with the geopolymer paste for ice templating. Chemical consolidation proceeded during freeze-casting and drying to avoid a sintering step. The ice-template metakaolin-based geopolymers were fully characterized in terms of macro- and micro-structure, intrinsic and induced porosity size distribution, and specific surface area. The most promising geopolymer composition was selected to produce lamellar monoliths. 2. Experimental 2.1. Sample preparation raux. Metakaolin grade M1200S was purchased from AGS Mine The composition of the metakaolin has been reported elsewhere [23]. The activating alkali aqueous solutions were potassium disilicate solutions with molar ratio SiO2:K2O ¼ 2 and H2O:K2O ¼ 13.5. Solutions were prepared by dissolving KOH pellets (purity >85% from SigmaeAldrich) in distilled water and adding fumed silica powder (99.8% from SigmaeAldrich) under magnetic stirring as described in a previous work [5]. The slurry, with a theoretical Si/Al molar ratio equal to 2, was prepared by the mechanical mixing of metakaolin and the potassium di-silicate solution for 20 min at 100 r.p.m. A reference geopolymer, coded G13 (Table 1), was prepared by casting the slurry into a plastic cylindrical mold and curing for 24 h at room temperature and 24 h at 80 C in a heater. These process conditions made it possible to obtain a 97% geopolymerized material [5]. The production process of freeze-cast geopolymers is schematically reported in Fig. 1, while compositions and codes can be found in Table 1. The water addition is reported as vol.% over the theoretical volume of the geopolymer solid matrix plus the added water. Solid loadings refer to the wt% of the starting metakaolin in slurries. After the preparation, the G13 slurry underwent a maturation phase of 4 h at room temperature, which was set by a trial-and-error approach in order to avoid the consolidation of the slurry. After maturation, distilled water was added to the slurry (20, 50 or 70 vol.%) and mechanically mixed for 8 min. The mixture was cast in cylindrical rubber molds, and pre-cooled on the freeze dryer shaft, set at 40 C, to produce monoliths reaching a height of 10 or 25 mm. Cast slurries were frozen at 40 C and the solidified phase was sublimated at P ¼ 10 Pa in 24 h (Edwards Mod. MFD01, Crawley, UK). After demolding, samples were rinsed in deionized water to remove any residues of unreacted potassium silicate and then dried in a heater at 100 C for 5 h. At least five samples for each formulation were produced for the reproducibility assessment of the icetemplating process. 2.2. Characterization and analytical techniques The morphological and microstructural features of freeze-cast geopolymers were examined by environmental Scanning Electron Microscopy (E-SEM FEI Quanta 200, FEI Company).
Table 1 Freeze-cast and reference geopolymer compositions. Sample code
K Silicate dilution, H2O/K2O
Additional H2O for freeze casting (vol %)
Solid loading wt%
Maturation step
G13 G13-20 G13-50 G13-70
13.5 13.5 13.5 13.5
e 20 50 70
63 53 36 23
24 h RT þ 24 h 80 C 4 h RT 4 h RT 4 h RT
208
E. Papa et al. / Microporous and Mesoporous Materials 215 (2015) 206e214
calculated by the BrunauereEmmeteTeller (BET) method. The total pore volume was obtained at p/p0 ¼ 0.995. Pore size distributions were obtained by the BJH method using the desorption branch. Powders belonging to ground and 600 mm sieved samples were analyzed. The measurement error is related to the accuracy of Hg intrusion porosimetry (<4%) and N2 adsorption/desorption techniques (<1%). Si/Al and K/Al molar ratios in the final consolidated samples were determined by X-Ray Fluorescence (XRF) analysis. Measurements were performed in a PANalytical Axios Advanced WD-XRF (wave length dispersive x-ray fluorescence) Spectrometer equipped with x-ray tube (Rh target), working at 4 kW. The pellets to be analyzed (diameter 13 mm) were prepared by mixing 0.300 g of the sample with 0.100 g wax (binder) at 100 kN for 120 min. The three-dimensional structure of a selected sample was studied by m-Computer Tomography. The sample was scanned using the Skyscan Micro-CT system model 1172 (Skyscan Bruker, Kontich, Belgium). The SkyScan 1172 scanner was operated at 100 kV and 100 mA, and the exposure time was set to 240 ms. Scanning was performed by 180 rotation around the vertical axis and with a rotation step of 0.2 . During scanning, a 0.5-mm aluminum filter was used. The necessary field of view (FOV) for the sample was determined, resulting in an optimal image pixel size of 9.8 mm. The projected images were reconstructed in 2000 x 2000 pixel-sized cross-sectional images using a modified Feldkamp cone-beam reconstruction algorithm [24] (NRecon, v.1.6.9 software; Skyscan Bruker). The slices were converted into an 8-bit BMP output format and the values within the dynamic range were mapped into gray levels 0e255. This output format was suitable for further processing and the structural information, such as volume fraction and structure orientation, was obtained from binary images using commercially available software CTan, v. 1.14.4; Skyscan Bruker. The solid/pore threshold was set by both trying different values and visually inspecting the appearance of the cross sections [25]. A variation in the threshold ±2 results in a variation of structural parameters, such as porosity, of less than 1%.
3. Results and discussion Fig. 1. Flow-chart of the freeze-casting process of geopolymers.
3.1. Effect of water addition
The true density e i.e. mass/volume of the solid e of the reference G13 was 2.35 g cm3; it was determined by helium pycnometry (Multivolume pycnometer 1305 by Micrometrics) [5]. The bulk density of samples was determined by weight-to-volume ratio. The volume was geometrically measured by using a caliper (accuracy ± 0.05 mm). The percent values of sample total porosity were then calculated according to the equation (1):
Total porosity ð%Þ ¼ ½1 ðbulk density=true densityÞ 100 (1) The ultra-macro-porosity was investigated by the image analysis of photographs with a resolution of 1200 dpi (scanner Sharp JX330, Japan) and scanning electron micrographs of cross sections. The pore size distribution in the range 0.0058e100 mm was analyzed by mercury porosimetry (surface tension ¼ 0.48 N/m and contact angle ¼ 140 , Thermo Finnigan Pascal 140 and Thermo Finnigan Pascal 240). Measurements of specific surface areas, pore volumes, and pore size distributions in the 2e500 nm range were carried out in a Micromeritics ASAP 2020 instrument by N2 adsorption/desorption at 196 C. Samples were previously degassed under vacuum, heated up to 250 C, and maintained for 60 min at a pressure below 30 mm Hg. The specific surface area was
3.1.1. Macro- and micro-structures of freeze-cast geopolymers Freeze-cast geopolymers were obtained by casting mixtures in molds up to a height of 10 mm. They showed different macro- and micro-structures depending on the amount of additional water (Table 1) targeted for ice-templating (Figs. 2 and 3). Fig. 2 shows high-resolution photos of top surfaces and cross sections of samples with 20, 50, and 70 vol.% of added water. In all samples a non-unidirectional lamellar structure was obtained. It is well known that the water amount in the starting mixture plays an important role in lamellae formation [13]. In fact, icetemplating is a segregation-induced templating of a second phase (in the present case, the newly formed geopolymer particles) by a solidifying liquid medium (the water) which is then removed by sublimation [11]. G13-20 samples had the highest solid loading, while the available volume of water for ice-templating was the lowest (Table 1). As a result, the lamellar macro-structure was completely lost (Fig. 2a, d) because of the entrapment of geopolymer particles within the growing ice crystals and the fast freezing of the liquid medium [13]. Although lamellae and channels were observed on the top of both G13-50 (Fig. 2b) and G13-70 (Fig. 2c), lamellae orientation over the cross section was tilted in the former (Fig. 2e) and completely random in the latter (Fig. 2f). Moreover, G13-70 samples
E. Papa et al. / Microporous and Mesoporous Materials 215 (2015) 206e214
209
Fig. 2. Top surfaces (a, b, c) and longitudinal cross sections (d, e, f) of samples G13-20 (a, d), G13-50 (b, e) and G13-70 (c, f).
having the lowest solid loading (Table 1) were weak and difficult to handle. The SEM micrographs of fracture surfaces (Fig. 3) show the typical geopolymeric microstructures [5] together with some characteristic features deriving from the ice-templating process [11]. In G13-20 a dendritic pattern was evidenced on the fracture surface (Fig. 3a, b) even though lamellae were not developed within the material (G13-20 in Fig. 2d). A finer dendritic pattern was observed in the microstructure of G13-50 (Fig. 3c, d), where lamellae started to develop on the top surface and within the sample (Fig. 2b, e). The lamellar surface in G13-70 (Fig. 3e, f) exhibited a particular topography with jagged dendritic-like features, running in the solidification direction. Cracks are present on the lamellar surfaces, indicating the weakness of the sample as previously mentioned. 3.1.2. Porosity and accessibility of the geopolymer inner volume The total porosity is reported in Table 2. Passing from the reference material to the freeze-cast samples, the total porosity increases proportionally with the volume of water added for icetemplating (Table 1). In detail, the pore size distributions measured by Hg intrusion porosimetry of reference and freeze-cast geopolymers are shown in Fig. 4. The results obtained account for the intrinsic porosity of the geopolymer matrix and for the smallest macro-pores due to icetemplating. As previously observed [5], pores ranging from 0.0058 to 1 mm are related to the intrinsic porosity of the matrices, since water does not enter into the geopolymer framework [2] and gives rise to a steric hindrance acting as a pore-forming agent upon its removal during setting [26,27]. Pores ranging from 1 to 100 mm were detected in all the freeze-cast samples and are due to icetemplating, since they are not present in G13. The pore distribution is monomodal in G13 and located in the range 10e40 nm (Fig. 4a). In freeze-cast samples the pore distribution becomes bimodal, with main peaks detected at 10 nm and 400 nm, when 20 vol.% of water is added (Fig. 4b). The pore size distribution is still bimodal in the samples prepared using 50 vol.% of additional water; however, the contribution of macropores increases with a main peak at 20 mm, even if pores of 10 nm are still
detected. When using 70 vol.% of extra water to perform icetemplating, the resulting material was characterized by a monomodal pore size distribution with its peak located at an even higher macropore size (80 mm). N2 adsorption/desorption measurements (Fig. 5) were performed on ground and 600 mm sieved samples, in order to obtain some complementary information on the intrinsic porosity of the samples lying outside the Hg porosimetry detection range. Specific surface area and pore volume values are summarized in Table 2. SBET and Vp again depend on the additional water used during freeze-casting. The extra water slows the completion of the geopolymerization, while the consolidation taking place during the freezing process results in a decrease in the surface area and pore volume of the geopolymer matrix, which is related to the water content. N2 adsorption/desorption isotherms of G13 and ice-template samples are shown in Fig. 5. Isotherms are IUPAC-classified as type II [28], a type which is characteristic of macroporous materials regardless of their chemical composition and water content. All the samples show hysteresis loops, the shape of which is slightly dependent on the water content in the freeze-drying. The reference G13 compound has a rather narrow hysteresis loop with almost parallel adsorption/desorption branches, and the hysteresis ends with a small plateau at high p/p0; however, this is a limit situation. Enlarged loops and the plateau at high p/p0 were not observed in ice-templated materials. They might be considered a mix of H2 and H3 (according to IUPAC classification [28]), which may be related to interparticle pores, i.e. pores between geopolymer nanoprecipitates. The modified hysteresis loop shape depends on the amount of water. The larger the additional water content for freezecasting, the broader the loop in all the p/p0 ranges, in accordance with the greater delay in the completion of the geopolymerization reaction. The changes in the shape of hysteresis loops suggest that the number of pores and pore dimensions are modified during the freeze-casting. The BJH method was used to obtain pore size distributions in the mesopore range and to study these behaviors. The results were compared with those obtained by Hg intrusion in the same pore size range. The reference G13 sample has a broad pore distribution between 4 and 100 nm (Fig. 7B) with a maximum at about 13 nm. The contribution of pores with width in the range 20e100 nm is low, in
210
E. Papa et al. / Microporous and Mesoporous Materials 215 (2015) 206e214
Fig. 3. SEM micrographs of fracture surfaces of samples G13-20 (a, b), G13-50 (c, d); lamellae surface of samples G13-70 (e, f). Table 2 Bulk density, total porosity (calculated by equation (1) or measured by mercury porosimetry*), total pore volume measured by mercury porosimetry and specific surface area (SBET) and pore volume (Vp) obtained through N2 adsorption/desorption measurements. Sample
Bulk density g$cm3
Total porosity (%)
Total pore volume mm3 g1
SBET m2 g1
Vp cm3 g1
G13 G13-20 G13-50 G13-70
e 1.1 0.7 0.4
43* 53 70 83
362 418 1404 1516
60.3 46.0 39.4 4.5
0.343 0.147 0.115 0.014
accordance with Hg porosimetry data. The evolution of the pore size during freeze-casting also follows a trend similar to that observed for Mercury Intrusion Porosimetry MIP. The maximum of the pore size distribution plots are placed at 7 and 5 nm for G13-20 and G13-50 respectively, while for sample G13-70 mesopores are almost absent. These results confirm that during freeze-drying the geopolymer matrix consolidates, thus leading to the formation of smaller pores. The amount of water added affects the process by contrasting polycondensation and, consequently, mesopore formation.
3.1.3. Chemical composition A weight loss of 17 ± 1% is measured after the washing in deionized water and drying at 100 C for all the freeze-cast samples (Table 3). This weight loss originates from both the material leakage and the removal of unreacted potassium silicate by water rinsing. A preliminary investigation into the chemical composition of freeze-cast geopolymers was performed by means of XRF analyses. The elemental compositions (i.e. the molar percentages of Si, Al, and K) of the reference material G13 and freeze-cast G13-20 and G13-50 are shown in Table 3. While G13 exhibited almost the
E. Papa et al. / Microporous and Mesoporous Materials 215 (2015) 206e214
211
Fig. 4. Pore size distributions measured by Hg intrusion porosimetry of samples G13 (a), G13-20 (b), G13-50 (c) and G13-70 (d).
expected composition, i.e. molar ratio Si/Al ¼ 1.99 (nominal 2.00) and K/Al ¼ 0.80, G13-50 had a Si/Al molar ratio equal to 1.87 and a very low K/Al ¼ 0.36. In the case of G13-20, the Si/Al molar ratio decreases down to 1.62 and K/Al ¼ 0.35. Therefore, XRF data indicate a loss of potassium and silicon. While the starting solution contained potassium di-silicate (K2O$2SiO2), after freezing the composition of the salt removed by water rinsing was roughly estimated at K2O$1.6SiO2 for G13-20, K2O$0.6SiO2 for G13-50 and K2O$0.9SiO2 for G13-70. Potassium silicate solution is as a polymer solution of silicate oligomers, stabilized by K2O (i.e. KOH:H2O) [29]. Since a maturation step is performed to trigger the geopolymerization without a final consolidation, the geopolymer paste contains a potassium silicate aqueous solution that continuously modifies its chemical composition. When water is added and mixed to the paste, the remaining potassium silicate solution is also diluted. When freezing at 40 C, the water component of the solution starts to crystallize into ice, and that leaves the remaining liquid portion over-saturated with respect to SiO2. Consequently, silica polymerizes into a gel [29,30].
The development of large chains of silicate oligomers is supposed to be “quenched“ by faster freezing in G13-20, where silicate remains entrapped within ice crystals and is then washed away during rinsing. In the case of G13-50 and G13-70 samples, the main part of the polymerized (amorphous) SiO2 is segregated by lamellar ice crystals and embedded in the geopolymer matrix, while the soluble K2O is then removed by water rinsing. It follows that the discrepancy between the Si/Al and K/Al molar ratios of G13, G13-20, G13-50 and G13-70 may be related to both the amount of added water and the resulting freezing rates. 3.2. Production and characterization of lamellar monoliths Since an intermediate value of water addition of 50 vol.% held the most promise for success in lamellar ice-templating, a G13-50 mixture was used to produce monoliths with a diameter of 10 mm and height of 25 mm. Highly reproducible lamellar monoliths were obtained as shown in Fig. 6a. The top surface showed parallel lamellae grouped in domains characterized by different orientation of the lamellae in the plane (Fig. 6b, c). The average
212
E. Papa et al. / Microporous and Mesoporous Materials 215 (2015) 206e214
Fig. 5. N2 isotherms (a) and BJH pore size distributions (b) of G13, G13-20, G13-50 and G13-70 samples.
Fig. 6. G13-50 monoliths with 25 mm height and 10 mm diameter (a) and SEM micrographs of their lamellar top surface (b, c).
lamellae thickness was around 50 mm on the top surface, while the average inter-distance between lamellae was around 15 mm (Fig. 6c). A weight loss of around 17 ± 1% was again measured after washing in deionized water and drying at 100 C. 3.2.1. Structural characterization by m-Computed Tomography The three-dimensional pore structure of the G13-50 freeze-cast geopolymer monoliths was studied by m-Computed Tomography (m-CT). Fig. 7 shows a 2D vertical slice from the middle of the cylinder and 2D horizontal slices of three separate zones of the 25 mm high geopolymer monolith. They are virtual cross sections in the xz and xy planes, i.e. perpendicular and longitudinal to the growth direction. Compared to 10 mm-high samples, with the increase of the mixture volume in the mold the freezing rate decreased, thus
creating a temperature gradient that promoted the unidirectional growth of the ice, observable in the vertical section of the 10 mmhigh sample. The vertical cross section can be divided into three zones; in the first zone, random pores could be seen close to the bottom, where the mold is in contact with the cold substrate of the freeze-dryer. Here, the freezing rate was very fast and the ice engulfed the particles, creating a dense layer. At first, the interface speed was too high and particles did not have enough time to rearrange or redistribute through the diffusion mechanism [31]. The freezing in the first zone was fast, while the supercooling degree was high; thus the ice crystals had no time to orient massively and grow unidirectionally [13]. Afterwards, a cellular porous microstructure formed, thus indicating that the columnar ice front was rejecting the particles; the speed of the liquid front decreased
E. Papa et al. / Microporous and Mesoporous Materials 215 (2015) 206e214
213
Fig. 7. Representative mCT slices of the G13-50 monoliths visualized in the “XZ” plane (longitudinal section; the freezing direction is from the sample bottom zone 1 toward the sample top zone 3) and “XY” plane (AeC). The color image C* highlights the different orientations of the lamellae domains. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
Table 3 Molar percentages of Si, Al and K, and Si/Al and K/Al molar ratios of the reference material G13 and freeze-cast G13-20, G13-50, and G13-70 samples measured by XRF. Sample
Si mol%
Al mol%
K mol%
Si/Al
K/Al
G13 G13-20 G13-50 G13-70
52.55 54.61 57.78 56.91
26.36 33.72 30.98 31.84
21.09 11.67 11.25 11.26
1.99 1.62 1.87 1.79
0.80 0.35 0.36 0.35
and a lamellar morphology started to appear. In zone 2, some lamellar ice crystals started to grow tilted toward the center where the ice fronts crossed and formed a herringbone porous structure. In zone 3, a progressive lamellar ordering was present, running toward the top of the sample with an almost constant thickness. These microstructural features are also seen in the 2D horizontal slices. In the 2D slice A, taken from the solid which was closer to the base of the mold surface, random pores coexisted with some lamellar ones. Aligned lamellae and random pores were still found in the middle slice B of the cylinder. A well-defined lamellar microstructure was evident in the surface close to slice C and at the cylinder tops. Horizontal slices showed that lamellae are arranged in domains kept over a great distance, giving rise to a spiral-like structure. The color image C* in Fig. 7 highlights the different orientations of the lamellae domains. The orientation of each domain may be related to the original nucleation conditions [32] and it is evident that ice crystals do not grow only perpendicularly to the bottom of the mold, but also diagonally, due to both the inhomogeneous temperature and the use of pre-cooled molds. Discontinuities in the microstructure were observed in the vertical and horizontal planes which might seem voids or cracks. They mainly run up to the boundaries between lamellar domains, so they might simply evidence such boundaries, but they might also be considered ice-templating defects, such as ice-lenses [22]. The porosity of the cylinder was obtained by image analysis even though it might differ from the results obtained by Hg intrusion porosimetry due to either the intrinsic spatial resolution of m-CT or the bottleneck effect of MIP [33]. An estimate of the homogeneity of the porosity developed during ice-templating can be obtained. The values obtained in the bottom, middle, and top slices by excluding the largest pores, which may be related to some cracks rather than to the freeze-casting, are 51, 48, and 48%,
respectively. Thus it could be stated that the porosity is rather constant along the cylinder in term of total amount, although it changes in terms of morphology, shape, size, and distribution, as is evidenced by the sample micro-macrostructures. 4. Conclusions Unidirectional lamellar geopolymers can be synthesized by the ice-templating method, producing highly porous materials. Even though ice-templating generally concerns colloidal ceramic suspensions, this technique can also be successfully applied to a water-based reactive solegel system to produce metakaolin-based geopolymers, without the addition of any organic dispersants and binders. The simultaneous formation of geopolymer intrinsic mesoporosity and lamellar macro-porosity by unidirectional ice growth, together with a final chemical consolidation, was obtained by properly combining the maturation steps of the geopolymeric reactive system with additional water amounts for ice-templating. The ice-templated macro-structures depend on both the height of the sample and the water addition targeted for ice-templating. The microstructure of the ice-templated samples shows geopolymer nano-precipitates, thus confirming that ice-templating and geopolymerization can be successfully combined. The Si/Al molar ratio of the most promising sample G13-50 is 1.87, revealing the notecomplete reaction of potassium silicate solution due to freezing. Studies are ongoing to further improve process conditions and to better correlate them with both micro- and macrostructures and geopolymer stoichiometry. Acknowledgements The activities were partially funded by the Flag project RITMARE e La Ricerca Italiana per il Mare e coordinated by the National Research Council of Italy and funded by the Ministry of University and Research within the National Research Programme 2011e2013. References [1] M. Gordon, J. Bell, W.M. Kriven, Ceram. Trans. 175 (2006) 215e224. [2] J. Davidovits, in: J. Davidovits (Ed.), Geopolymers Chemistry and Applications, Institut Geopolymere, Saint-Quentin, France, 2008. [3] J. Dedecek, Z. Tvaruzkova, Z. Sobalik, J. Am. Ceram. Soc. 91 (9) (2008) 3052e3057.
214
E. Papa et al. / Microporous and Mesoporous Materials 215 (2015) 206e214
[4] W.M. Kriven, J.L. Bell, M. Gordon, Ceram. Trans. 153 (2003) 227e250. [5] E. Landi, V. Medri, E. Papa, J. Dedecek, P. Klein, P. Benito, A. Vaccari, Appl. Clay Sci. 73 (2013) 56e64. [6] K. Okada, T. Isobe, K. Katsumata, Y. Kameshima, A. Nakajima, K.J.D. MacKenzie, Sci. Technol. Adv. Mater. 12 (2011) 11, 064701. [7] P. Colombo, Philos. Trans. R. Soc. A 364 (2006) 109e124. [8] W. Van Bonin, U. Nehen, U. Von Gizycki, US Patent 3,864,137, 1975. [9] J.L. Bell, W.M. Kriven, in: H.T. Lin, K. Koumoto, W.M. Kriven, E. Garcia, I.E. Reimanis, D.P. Norton (Eds.), Developments in Strategic Materials: Ceramic Engineering and Science Proceedings, 29, 2009, pp. 97e111 (10). [10] V. Medri, E. Papa, J. Dedecek, H. Jirlova, P. Benito, A. Vaccari, E. Landi, Ceram. Int. 39 (2013) 7657e7668. [11] S. Deville, Adv. Eng. Mater. 10 (3) (2008) 155e169. [12] W.L. Li, K. Lu, J.Y. Walz, Int. Mat. Rev. 57 (1) (2012) 37e60. [13] S. Deville, E. Saiz, A.P. Tomsia, Acta Mater. 55 (2007) 1965e1974. [14] S.R. Mukai, H. Nishihara, H. Tamon, Catal. Surv. Asia 10 (3e4) (2006) 161e171. [15] T. Fukasawa, Z.Y. Deng, M. Ando, T. Ohji, Y. Goto, J. Mater. Sci. 36 (10) (2001) 2523e2527. [16] K. Araki, J.W. Halloran, J. Am. Ceram. Soc. 88 (5) (2005) 1108e1114. [17] S. Deville, E. Saiz, A. Tomsia, Biomaterials 27 (32) (2006) 5480e5489. [18] E. Landi, F. Valentini, A. Tampieri, Acta Biomater. 4 (6) (2008) 1620e1626. [19] H. Zhang, I. Hussain, M. Brust, M.F. Butler, S.P. Rannard, A.I. Cooper, Nat. Mater. 4 (10) (2005) 787e793. [20] E. Landi, D. Sciti, C. Melandri, V. Medri, J. Eur. Ceram. Soc. 33 (2013) 1599e1607.
[21] H. Mori, K. Aontani, N. Sano, H. Tamon, J. Mater. Chem. 21 (2011) 5677e5681. [22] A. Lasalle, C. Guizard, E. Maire, J. Adrien, S. Deville, Acta Mater. 60 (2012) 4594e4603. [23] V. Medri, S. Fabbri, J. Dedecek, Z. Sobalik, Z. Tvaruzkova, A. Vaccari, Appl. Clay Sci. 50 (2010) 538e545. [24] L.A. Feldkamp, L.C. Davis, J.W. Kress, J. Opt. Soc. Am. 1 (1984) 612e619. [25] S. Bugani, M. Camaiti, L. Morselli, E. Van de Casteele, K. Janssens, Anal. Bioanal. Chem. 391 (2008) 1343e1350. [26] Z. Zuhua, Y. Xiao, Z. Huajun, C. Yue, Appl. Clay Sci. 43 (2009) 218e223. [27] K. Okada, A. Ooyama, T. Isobe, Y. Nakajima, K.J.D. MacKenzie, J. Eur. Ceram. Soc. 29 (2009) 1917e1923. [28] K.S.W. Sing, D.H. Everett, R.A.W. Haul, L. Moscou, R.A. Pierotti, J. Rouquerol, T. Siemieniewska, Pure Appl. Chem 57 (1985) 603e619. [29] J.D. Hunt, A. Kavner, E.A. Schauble, D. Snyder, C.E. Manning, Chem. Geol. 283 (2011) 161e170. [30] T.T. Trinh, A.P.J. Jansen, R.A. van Santen, E.J. Meijer, J. Phys. Chem. C 113 (2009) 2647e2652. [31] S. Deville, E. Maire, A. Lasalle, A. Bogner, C. Gauthier, J. Leloup, C. Guizard, J. Am. Ceram. Soc. 93 (9) (2010) 2507e2510. [32] S. Deville, E. Maire, A. Lasalle, A. Bogner, C. Gauthier, J. Leloup, C. Guizard, J. Am. Ceram. Soc. 92 (11) (2009) 2497e2503. [33] S. Bugani, M. Camaiti, L. Morselli, E. Van de Casteele, K. Janssens, X-Ray Spectrom. 36 (2007) 316e320.
Contents lists available at ScienceDirect
Microporous and Mesoporous Materials journal homepage: www.elsevier.com/locate/micromeso
Synthesis of porous hierarchical geopolymer monoliths by ice-templating Elettra Papa a, Valentina Medri a, *, Patricia Benito b, Angelo Vaccari b, Simone Bugani b, Jakub Jaroszewicz c, Wojciech Swieszkowski c, Elena Landi a a b c
National Research Council of Italy, Institute of Science and Technology for Ceramics (CNR-ISTEC), via Granarolo 64, 48018 Faenza, Italy “Toso Montanari” Department of Industrial Chemistry, Alma Mater Studiorum, University of Bologna, Viale Risorgimento 4, 40136 Bologna, Italy Faculty of Materials Science and Engineering, Warsaw University of Technology, 141 Woloska Str., 02-507 Warsaw, Poland
a r t i c l e i n f o
a b s t r a c t
Article history: Received 15 September 2014 Received in revised form 12 May 2015 Accepted 26 May 2015 Available online 9 June 2015
Hierarchical geopolymers with mesoporous matrices and lamellar macro-porosities were produced for the first time by ice-templating (freeze-casting) technique in a water-based solegel system. Starting from an aqueous solution of metakaolin and potassium silicate, geopolymerization was triggered through a maturation step without reaching complete consolidation. Different amounts of water (20, 50, 70 vol.%) were then mixed with the geopolymer paste to induce lamellar ice growth by unidirectional freezing. The consolidation was completed during freeze-casting and drying. Ice-templated hierarchical geopolymers had 53e83% total porosity and N2 BET specific surface area values from 4 to 46 m2 g1, depending on the amount of water added. A broad mesopore Barret Joyner Halenda distribution was detected between 4 and 100 nm with maxima from 5 to 7 nm, while macropores in the 1-to-100 mm range were identified by Hg intrusion porosimetry. Lamellar monoliths (height 25 mm) were produced with a selected composition (50 vol.% of additional water) in order to assess the reproducibility of the ice-templating technique combined with geopolymerization. © 2015 Elsevier Inc. All rights reserved.
Keywords: Ice-templating Freeze-casting Geopolymer Alkali-bonded ceramics Porosity
1. Introduction Geopolymers are alkali-bonded ceramics [1], constituting a family of materials with properties varying among those characteristic of ceramics, cements, zeolites, or refractories, depending on their formulation. Geopolymers were produced by reacting alumino-silicate powders (metakaolins, blast furnace slags, fly ashes, pozzolana, etc.) with an aqueous alkali hydroxide and/or alkali silicate solution [2]. At the atomic scale, the geopolymer amorphous network is made by SiO4 and AlO4 tetrahedra connected by oxygen corners [2]. Recent results indicate that the tetrahedra form SieOeAleO rings of various sizes in the network and endow the geopolymer matrix with ion exchange properties [3]. From some standpoints, geopolymers can be regarded as either the amorphous counterpart or the precursor of crystalline zeolites.
* Corresponding author. Tel.: þ39 0546 699751; fax: þ39 0546 46381. E-mail address: [email protected] (V. Medri). http://dx.doi.org/10.1016/j.micromeso.2015.05.043 1387-1811/© 2015 Elsevier Inc. All rights reserved.
The microstructure of a metakaolin-based geopolymer consists of nano-particulates separated by micro- and mesopores [4,5]. Since geopolymers are intrinsically mesoporous, a hierarchical pore system can be constructed by combining mesopores with macroporosity. Geopolymers could be used to develop porous materials covering pore sizes ranging from a few tenths of nanometers to a few millimeters, as well as a total pore volume from 30% up to 90% [5]. Because of the high accessibility of pores, porous ceramics with three-dimensionally interconnected and distributed open pores (3D structures) are useful as catalysts, catalyst supports, filters, scaffolds, and adsorbents [6,7]. For all the above-mentioned applications, it is absolutely necessary to control the pore size dimension and distribution, as well as the pore amount and structure (shape, morphology, orientation, and surface properties). The required pore dimensions for catalysts and supports range from 1 nm to few tenths of microns, for filtering and purification systems from a few nanometers to a few millimeters, and for dust collecting from a few micrometers to a few millimeters [6]. The geopolymer production process in an aqueous medium makes it possible to obtain a custom-tailored porosity. Different
E. Papa et al. / Microporous and Mesoporous Materials 215 (2015) 206e214
techniques were used in order to meet porosity requirements for various purposes. Water contents in the starting mixtures affected the intrinsic mesoporosity of the geopolymer matrix, since water acted as a pore former during the polycondensation stage [5]. Moreover, ultra-macro-porosity was obtained by direct foaming, using hydrogen peroxide as a blowing agent [8], or exploiting the redox reaction of metallic Al [9] or Si powders [5,10] in an alkaline solution to induce porosity through H2 evolution. The water-based freeze-casting (i.e. ice-templating) is a useful technique for fabricating porous materials with main unidirectionally oriented pores and a high open porosity, where the final component results in a lamellar or laminate-like structure. This technique is particularly interesting for the peculiar structures and properties shown by porous freeze-cast ceramics, thus opening new opportunities in the field of cellular ceramics [11]. For example, unidirectional porous ceramics have a higher permeability than conventional porous ceramics, and are potentially good thermal insulators [6]. The ice-templating technique consists of freezing an aqueous suspension, followed by ice sublimation under reduced pressure. The resulting green body is generally sintered to consolidate the overall architecture as well as to obtain the desired mechanical and functional properties, while maintaining a high level of porosity. The unidirectional channel-like (lamellar) porosity is obtained in the case of unidirectional freezing, where pores are the replica of ice crystals [11e13]. The template can be easily removed through simple thawing and drying; this is a unique and beneficial feature of this method, overcoming the problems inherent in other templating techniques requiring the use of expensive templates removed through either high temperature or extremely high or low pH conditions [14]. The process is environmentally friendly, since it uses water as a removable template, highly versatile, and the resulting structures are highly tunable by changing process conditions. In detail, the final porosity content can be tuned by varying the particle content within the slurry, and the size of the pores is affected by freezing kinetics. Furthermore, since the solidification is often directional, a very anisotropic morphology of pores can be seen in the solidification plane, where porous channels run from the bottom to the top of samples [13]. Ice-templating has been applied to a large variety of materials such as alumina [15,16], hydroxyapatite [17,18], polymeric materials [19], zirconium diborides ultra-high temperature ceramics [20], zeolite monolith [21], and so on. While literature reports ice-templating of colloidal inert ceramic suspensions, the aim of the present study is to apply this technique to a water-based solegel system capable of producing alkalibonded ceramics, i.e. metakaolin-based geopolymers. To the author's best knowledge, this is the first time ice-templating has been applied to geopolymers. Again, it is important to point out that the systems generally used for freeze-casting [13] are colloidal suspensions of inert ceramic particles with the addition of organic dispersants and binders with the aim of fostering ice-lamellar growth [22], and final samples are usually consolidated by sintering. Conversely, geopolymer slurries are based on a sol/gel reactive system with no addition of any organic dispersant or binder, while the consolidation is of a chemical type, thus avoiding any high-temperature thermal treatment. The aim of the present work is to promote the simultaneous formation of geopolymer intrinsic mesoporosity and lamellar macroporosity by unidirectional ice growth, together with a final chemical consolidation. Starting from a geopolymer mixture of metakaolin and potassium silicate aqueous solution, geopolymerization was triggered through a maturation step, but without reaching a complete consolidation.
207
Different amounts of water were then mixed with the geopolymer paste for ice templating. Chemical consolidation proceeded during freeze-casting and drying to avoid a sintering step. The ice-template metakaolin-based geopolymers were fully characterized in terms of macro- and micro-structure, intrinsic and induced porosity size distribution, and specific surface area. The most promising geopolymer composition was selected to produce lamellar monoliths. 2. Experimental 2.1. Sample preparation raux. Metakaolin grade M1200S was purchased from AGS Mine The composition of the metakaolin has been reported elsewhere [23]. The activating alkali aqueous solutions were potassium disilicate solutions with molar ratio SiO2:K2O ¼ 2 and H2O:K2O ¼ 13.5. Solutions were prepared by dissolving KOH pellets (purity >85% from SigmaeAldrich) in distilled water and adding fumed silica powder (99.8% from SigmaeAldrich) under magnetic stirring as described in a previous work [5]. The slurry, with a theoretical Si/Al molar ratio equal to 2, was prepared by the mechanical mixing of metakaolin and the potassium di-silicate solution for 20 min at 100 r.p.m. A reference geopolymer, coded G13 (Table 1), was prepared by casting the slurry into a plastic cylindrical mold and curing for 24 h at room temperature and 24 h at 80 C in a heater. These process conditions made it possible to obtain a 97% geopolymerized material [5]. The production process of freeze-cast geopolymers is schematically reported in Fig. 1, while compositions and codes can be found in Table 1. The water addition is reported as vol.% over the theoretical volume of the geopolymer solid matrix plus the added water. Solid loadings refer to the wt% of the starting metakaolin in slurries. After the preparation, the G13 slurry underwent a maturation phase of 4 h at room temperature, which was set by a trial-and-error approach in order to avoid the consolidation of the slurry. After maturation, distilled water was added to the slurry (20, 50 or 70 vol.%) and mechanically mixed for 8 min. The mixture was cast in cylindrical rubber molds, and pre-cooled on the freeze dryer shaft, set at 40 C, to produce monoliths reaching a height of 10 or 25 mm. Cast slurries were frozen at 40 C and the solidified phase was sublimated at P ¼ 10 Pa in 24 h (Edwards Mod. MFD01, Crawley, UK). After demolding, samples were rinsed in deionized water to remove any residues of unreacted potassium silicate and then dried in a heater at 100 C for 5 h. At least five samples for each formulation were produced for the reproducibility assessment of the icetemplating process. 2.2. Characterization and analytical techniques The morphological and microstructural features of freeze-cast geopolymers were examined by environmental Scanning Electron Microscopy (E-SEM FEI Quanta 200, FEI Company).
Table 1 Freeze-cast and reference geopolymer compositions. Sample code
K Silicate dilution, H2O/K2O
Additional H2O for freeze casting (vol %)
Solid loading wt%
Maturation step
G13 G13-20 G13-50 G13-70
13.5 13.5 13.5 13.5
e 20 50 70
63 53 36 23
24 h RT þ 24 h 80 C 4 h RT 4 h RT 4 h RT
208
E. Papa et al. / Microporous and Mesoporous Materials 215 (2015) 206e214
calculated by the BrunauereEmmeteTeller (BET) method. The total pore volume was obtained at p/p0 ¼ 0.995. Pore size distributions were obtained by the BJH method using the desorption branch. Powders belonging to ground and 600 mm sieved samples were analyzed. The measurement error is related to the accuracy of Hg intrusion porosimetry (<4%) and N2 adsorption/desorption techniques (<1%). Si/Al and K/Al molar ratios in the final consolidated samples were determined by X-Ray Fluorescence (XRF) analysis. Measurements were performed in a PANalytical Axios Advanced WD-XRF (wave length dispersive x-ray fluorescence) Spectrometer equipped with x-ray tube (Rh target), working at 4 kW. The pellets to be analyzed (diameter 13 mm) were prepared by mixing 0.300 g of the sample with 0.100 g wax (binder) at 100 kN for 120 min. The three-dimensional structure of a selected sample was studied by m-Computer Tomography. The sample was scanned using the Skyscan Micro-CT system model 1172 (Skyscan Bruker, Kontich, Belgium). The SkyScan 1172 scanner was operated at 100 kV and 100 mA, and the exposure time was set to 240 ms. Scanning was performed by 180 rotation around the vertical axis and with a rotation step of 0.2 . During scanning, a 0.5-mm aluminum filter was used. The necessary field of view (FOV) for the sample was determined, resulting in an optimal image pixel size of 9.8 mm. The projected images were reconstructed in 2000 x 2000 pixel-sized cross-sectional images using a modified Feldkamp cone-beam reconstruction algorithm [24] (NRecon, v.1.6.9 software; Skyscan Bruker). The slices were converted into an 8-bit BMP output format and the values within the dynamic range were mapped into gray levels 0e255. This output format was suitable for further processing and the structural information, such as volume fraction and structure orientation, was obtained from binary images using commercially available software CTan, v. 1.14.4; Skyscan Bruker. The solid/pore threshold was set by both trying different values and visually inspecting the appearance of the cross sections [25]. A variation in the threshold ±2 results in a variation of structural parameters, such as porosity, of less than 1%.
3. Results and discussion Fig. 1. Flow-chart of the freeze-casting process of geopolymers.
3.1. Effect of water addition
The true density e i.e. mass/volume of the solid e of the reference G13 was 2.35 g cm3; it was determined by helium pycnometry (Multivolume pycnometer 1305 by Micrometrics) [5]. The bulk density of samples was determined by weight-to-volume ratio. The volume was geometrically measured by using a caliper (accuracy ± 0.05 mm). The percent values of sample total porosity were then calculated according to the equation (1):
Total porosity ð%Þ ¼ ½1 ðbulk density=true densityÞ 100 (1) The ultra-macro-porosity was investigated by the image analysis of photographs with a resolution of 1200 dpi (scanner Sharp JX330, Japan) and scanning electron micrographs of cross sections. The pore size distribution in the range 0.0058e100 mm was analyzed by mercury porosimetry (surface tension ¼ 0.48 N/m and contact angle ¼ 140 , Thermo Finnigan Pascal 140 and Thermo Finnigan Pascal 240). Measurements of specific surface areas, pore volumes, and pore size distributions in the 2e500 nm range were carried out in a Micromeritics ASAP 2020 instrument by N2 adsorption/desorption at 196 C. Samples were previously degassed under vacuum, heated up to 250 C, and maintained for 60 min at a pressure below 30 mm Hg. The specific surface area was
3.1.1. Macro- and micro-structures of freeze-cast geopolymers Freeze-cast geopolymers were obtained by casting mixtures in molds up to a height of 10 mm. They showed different macro- and micro-structures depending on the amount of additional water (Table 1) targeted for ice-templating (Figs. 2 and 3). Fig. 2 shows high-resolution photos of top surfaces and cross sections of samples with 20, 50, and 70 vol.% of added water. In all samples a non-unidirectional lamellar structure was obtained. It is well known that the water amount in the starting mixture plays an important role in lamellae formation [13]. In fact, icetemplating is a segregation-induced templating of a second phase (in the present case, the newly formed geopolymer particles) by a solidifying liquid medium (the water) which is then removed by sublimation [11]. G13-20 samples had the highest solid loading, while the available volume of water for ice-templating was the lowest (Table 1). As a result, the lamellar macro-structure was completely lost (Fig. 2a, d) because of the entrapment of geopolymer particles within the growing ice crystals and the fast freezing of the liquid medium [13]. Although lamellae and channels were observed on the top of both G13-50 (Fig. 2b) and G13-70 (Fig. 2c), lamellae orientation over the cross section was tilted in the former (Fig. 2e) and completely random in the latter (Fig. 2f). Moreover, G13-70 samples
E. Papa et al. / Microporous and Mesoporous Materials 215 (2015) 206e214
209
Fig. 2. Top surfaces (a, b, c) and longitudinal cross sections (d, e, f) of samples G13-20 (a, d), G13-50 (b, e) and G13-70 (c, f).
having the lowest solid loading (Table 1) were weak and difficult to handle. The SEM micrographs of fracture surfaces (Fig. 3) show the typical geopolymeric microstructures [5] together with some characteristic features deriving from the ice-templating process [11]. In G13-20 a dendritic pattern was evidenced on the fracture surface (Fig. 3a, b) even though lamellae were not developed within the material (G13-20 in Fig. 2d). A finer dendritic pattern was observed in the microstructure of G13-50 (Fig. 3c, d), where lamellae started to develop on the top surface and within the sample (Fig. 2b, e). The lamellar surface in G13-70 (Fig. 3e, f) exhibited a particular topography with jagged dendritic-like features, running in the solidification direction. Cracks are present on the lamellar surfaces, indicating the weakness of the sample as previously mentioned. 3.1.2. Porosity and accessibility of the geopolymer inner volume The total porosity is reported in Table 2. Passing from the reference material to the freeze-cast samples, the total porosity increases proportionally with the volume of water added for icetemplating (Table 1). In detail, the pore size distributions measured by Hg intrusion porosimetry of reference and freeze-cast geopolymers are shown in Fig. 4. The results obtained account for the intrinsic porosity of the geopolymer matrix and for the smallest macro-pores due to icetemplating. As previously observed [5], pores ranging from 0.0058 to 1 mm are related to the intrinsic porosity of the matrices, since water does not enter into the geopolymer framework [2] and gives rise to a steric hindrance acting as a pore-forming agent upon its removal during setting [26,27]. Pores ranging from 1 to 100 mm were detected in all the freeze-cast samples and are due to icetemplating, since they are not present in G13. The pore distribution is monomodal in G13 and located in the range 10e40 nm (Fig. 4a). In freeze-cast samples the pore distribution becomes bimodal, with main peaks detected at 10 nm and 400 nm, when 20 vol.% of water is added (Fig. 4b). The pore size distribution is still bimodal in the samples prepared using 50 vol.% of additional water; however, the contribution of macropores increases with a main peak at 20 mm, even if pores of 10 nm are still
detected. When using 70 vol.% of extra water to perform icetemplating, the resulting material was characterized by a monomodal pore size distribution with its peak located at an even higher macropore size (80 mm). N2 adsorption/desorption measurements (Fig. 5) were performed on ground and 600 mm sieved samples, in order to obtain some complementary information on the intrinsic porosity of the samples lying outside the Hg porosimetry detection range. Specific surface area and pore volume values are summarized in Table 2. SBET and Vp again depend on the additional water used during freeze-casting. The extra water slows the completion of the geopolymerization, while the consolidation taking place during the freezing process results in a decrease in the surface area and pore volume of the geopolymer matrix, which is related to the water content. N2 adsorption/desorption isotherms of G13 and ice-template samples are shown in Fig. 5. Isotherms are IUPAC-classified as type II [28], a type which is characteristic of macroporous materials regardless of their chemical composition and water content. All the samples show hysteresis loops, the shape of which is slightly dependent on the water content in the freeze-drying. The reference G13 compound has a rather narrow hysteresis loop with almost parallel adsorption/desorption branches, and the hysteresis ends with a small plateau at high p/p0; however, this is a limit situation. Enlarged loops and the plateau at high p/p0 were not observed in ice-templated materials. They might be considered a mix of H2 and H3 (according to IUPAC classification [28]), which may be related to interparticle pores, i.e. pores between geopolymer nanoprecipitates. The modified hysteresis loop shape depends on the amount of water. The larger the additional water content for freezecasting, the broader the loop in all the p/p0 ranges, in accordance with the greater delay in the completion of the geopolymerization reaction. The changes in the shape of hysteresis loops suggest that the number of pores and pore dimensions are modified during the freeze-casting. The BJH method was used to obtain pore size distributions in the mesopore range and to study these behaviors. The results were compared with those obtained by Hg intrusion in the same pore size range. The reference G13 sample has a broad pore distribution between 4 and 100 nm (Fig. 7B) with a maximum at about 13 nm. The contribution of pores with width in the range 20e100 nm is low, in
210
E. Papa et al. / Microporous and Mesoporous Materials 215 (2015) 206e214
Fig. 3. SEM micrographs of fracture surfaces of samples G13-20 (a, b), G13-50 (c, d); lamellae surface of samples G13-70 (e, f). Table 2 Bulk density, total porosity (calculated by equation (1) or measured by mercury porosimetry*), total pore volume measured by mercury porosimetry and specific surface area (SBET) and pore volume (Vp) obtained through N2 adsorption/desorption measurements. Sample
Bulk density g$cm3
Total porosity (%)
Total pore volume mm3 g1
SBET m2 g1
Vp cm3 g1
G13 G13-20 G13-50 G13-70
e 1.1 0.7 0.4
43* 53 70 83
362 418 1404 1516
60.3 46.0 39.4 4.5
0.343 0.147 0.115 0.014
accordance with Hg porosimetry data. The evolution of the pore size during freeze-casting also follows a trend similar to that observed for Mercury Intrusion Porosimetry MIP. The maximum of the pore size distribution plots are placed at 7 and 5 nm for G13-20 and G13-50 respectively, while for sample G13-70 mesopores are almost absent. These results confirm that during freeze-drying the geopolymer matrix consolidates, thus leading to the formation of smaller pores. The amount of water added affects the process by contrasting polycondensation and, consequently, mesopore formation.
3.1.3. Chemical composition A weight loss of 17 ± 1% is measured after the washing in deionized water and drying at 100 C for all the freeze-cast samples (Table 3). This weight loss originates from both the material leakage and the removal of unreacted potassium silicate by water rinsing. A preliminary investigation into the chemical composition of freeze-cast geopolymers was performed by means of XRF analyses. The elemental compositions (i.e. the molar percentages of Si, Al, and K) of the reference material G13 and freeze-cast G13-20 and G13-50 are shown in Table 3. While G13 exhibited almost the
E. Papa et al. / Microporous and Mesoporous Materials 215 (2015) 206e214
211
Fig. 4. Pore size distributions measured by Hg intrusion porosimetry of samples G13 (a), G13-20 (b), G13-50 (c) and G13-70 (d).
expected composition, i.e. molar ratio Si/Al ¼ 1.99 (nominal 2.00) and K/Al ¼ 0.80, G13-50 had a Si/Al molar ratio equal to 1.87 and a very low K/Al ¼ 0.36. In the case of G13-20, the Si/Al molar ratio decreases down to 1.62 and K/Al ¼ 0.35. Therefore, XRF data indicate a loss of potassium and silicon. While the starting solution contained potassium di-silicate (K2O$2SiO2), after freezing the composition of the salt removed by water rinsing was roughly estimated at K2O$1.6SiO2 for G13-20, K2O$0.6SiO2 for G13-50 and K2O$0.9SiO2 for G13-70. Potassium silicate solution is as a polymer solution of silicate oligomers, stabilized by K2O (i.e. KOH:H2O) [29]. Since a maturation step is performed to trigger the geopolymerization without a final consolidation, the geopolymer paste contains a potassium silicate aqueous solution that continuously modifies its chemical composition. When water is added and mixed to the paste, the remaining potassium silicate solution is also diluted. When freezing at 40 C, the water component of the solution starts to crystallize into ice, and that leaves the remaining liquid portion over-saturated with respect to SiO2. Consequently, silica polymerizes into a gel [29,30].
The development of large chains of silicate oligomers is supposed to be “quenched“ by faster freezing in G13-20, where silicate remains entrapped within ice crystals and is then washed away during rinsing. In the case of G13-50 and G13-70 samples, the main part of the polymerized (amorphous) SiO2 is segregated by lamellar ice crystals and embedded in the geopolymer matrix, while the soluble K2O is then removed by water rinsing. It follows that the discrepancy between the Si/Al and K/Al molar ratios of G13, G13-20, G13-50 and G13-70 may be related to both the amount of added water and the resulting freezing rates. 3.2. Production and characterization of lamellar monoliths Since an intermediate value of water addition of 50 vol.% held the most promise for success in lamellar ice-templating, a G13-50 mixture was used to produce monoliths with a diameter of 10 mm and height of 25 mm. Highly reproducible lamellar monoliths were obtained as shown in Fig. 6a. The top surface showed parallel lamellae grouped in domains characterized by different orientation of the lamellae in the plane (Fig. 6b, c). The average
212
E. Papa et al. / Microporous and Mesoporous Materials 215 (2015) 206e214
Fig. 5. N2 isotherms (a) and BJH pore size distributions (b) of G13, G13-20, G13-50 and G13-70 samples.
Fig. 6. G13-50 monoliths with 25 mm height and 10 mm diameter (a) and SEM micrographs of their lamellar top surface (b, c).
lamellae thickness was around 50 mm on the top surface, while the average inter-distance between lamellae was around 15 mm (Fig. 6c). A weight loss of around 17 ± 1% was again measured after washing in deionized water and drying at 100 C. 3.2.1. Structural characterization by m-Computed Tomography The three-dimensional pore structure of the G13-50 freeze-cast geopolymer monoliths was studied by m-Computed Tomography (m-CT). Fig. 7 shows a 2D vertical slice from the middle of the cylinder and 2D horizontal slices of three separate zones of the 25 mm high geopolymer monolith. They are virtual cross sections in the xz and xy planes, i.e. perpendicular and longitudinal to the growth direction. Compared to 10 mm-high samples, with the increase of the mixture volume in the mold the freezing rate decreased, thus
creating a temperature gradient that promoted the unidirectional growth of the ice, observable in the vertical section of the 10 mmhigh sample. The vertical cross section can be divided into three zones; in the first zone, random pores could be seen close to the bottom, where the mold is in contact with the cold substrate of the freeze-dryer. Here, the freezing rate was very fast and the ice engulfed the particles, creating a dense layer. At first, the interface speed was too high and particles did not have enough time to rearrange or redistribute through the diffusion mechanism [31]. The freezing in the first zone was fast, while the supercooling degree was high; thus the ice crystals had no time to orient massively and grow unidirectionally [13]. Afterwards, a cellular porous microstructure formed, thus indicating that the columnar ice front was rejecting the particles; the speed of the liquid front decreased
E. Papa et al. / Microporous and Mesoporous Materials 215 (2015) 206e214
213
Fig. 7. Representative mCT slices of the G13-50 monoliths visualized in the “XZ” plane (longitudinal section; the freezing direction is from the sample bottom zone 1 toward the sample top zone 3) and “XY” plane (AeC). The color image C* highlights the different orientations of the lamellae domains. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
Table 3 Molar percentages of Si, Al and K, and Si/Al and K/Al molar ratios of the reference material G13 and freeze-cast G13-20, G13-50, and G13-70 samples measured by XRF. Sample
Si mol%
Al mol%
K mol%
Si/Al
K/Al
G13 G13-20 G13-50 G13-70
52.55 54.61 57.78 56.91
26.36 33.72 30.98 31.84
21.09 11.67 11.25 11.26
1.99 1.62 1.87 1.79
0.80 0.35 0.36 0.35
and a lamellar morphology started to appear. In zone 2, some lamellar ice crystals started to grow tilted toward the center where the ice fronts crossed and formed a herringbone porous structure. In zone 3, a progressive lamellar ordering was present, running toward the top of the sample with an almost constant thickness. These microstructural features are also seen in the 2D horizontal slices. In the 2D slice A, taken from the solid which was closer to the base of the mold surface, random pores coexisted with some lamellar ones. Aligned lamellae and random pores were still found in the middle slice B of the cylinder. A well-defined lamellar microstructure was evident in the surface close to slice C and at the cylinder tops. Horizontal slices showed that lamellae are arranged in domains kept over a great distance, giving rise to a spiral-like structure. The color image C* in Fig. 7 highlights the different orientations of the lamellae domains. The orientation of each domain may be related to the original nucleation conditions [32] and it is evident that ice crystals do not grow only perpendicularly to the bottom of the mold, but also diagonally, due to both the inhomogeneous temperature and the use of pre-cooled molds. Discontinuities in the microstructure were observed in the vertical and horizontal planes which might seem voids or cracks. They mainly run up to the boundaries between lamellar domains, so they might simply evidence such boundaries, but they might also be considered ice-templating defects, such as ice-lenses [22]. The porosity of the cylinder was obtained by image analysis even though it might differ from the results obtained by Hg intrusion porosimetry due to either the intrinsic spatial resolution of m-CT or the bottleneck effect of MIP [33]. An estimate of the homogeneity of the porosity developed during ice-templating can be obtained. The values obtained in the bottom, middle, and top slices by excluding the largest pores, which may be related to some cracks rather than to the freeze-casting, are 51, 48, and 48%,
respectively. Thus it could be stated that the porosity is rather constant along the cylinder in term of total amount, although it changes in terms of morphology, shape, size, and distribution, as is evidenced by the sample micro-macrostructures. 4. Conclusions Unidirectional lamellar geopolymers can be synthesized by the ice-templating method, producing highly porous materials. Even though ice-templating generally concerns colloidal ceramic suspensions, this technique can also be successfully applied to a water-based reactive solegel system to produce metakaolin-based geopolymers, without the addition of any organic dispersants and binders. The simultaneous formation of geopolymer intrinsic mesoporosity and lamellar macro-porosity by unidirectional ice growth, together with a final chemical consolidation, was obtained by properly combining the maturation steps of the geopolymeric reactive system with additional water amounts for ice-templating. The ice-templated macro-structures depend on both the height of the sample and the water addition targeted for ice-templating. The microstructure of the ice-templated samples shows geopolymer nano-precipitates, thus confirming that ice-templating and geopolymerization can be successfully combined. The Si/Al molar ratio of the most promising sample G13-50 is 1.87, revealing the notecomplete reaction of potassium silicate solution due to freezing. Studies are ongoing to further improve process conditions and to better correlate them with both micro- and macrostructures and geopolymer stoichiometry. Acknowledgements The activities were partially funded by the Flag project RITMARE e La Ricerca Italiana per il Mare e coordinated by the National Research Council of Italy and funded by the Ministry of University and Research within the National Research Programme 2011e2013. References [1] M. Gordon, J. Bell, W.M. Kriven, Ceram. Trans. 175 (2006) 215e224. [2] J. Davidovits, in: J. Davidovits (Ed.), Geopolymers Chemistry and Applications, Institut Geopolymere, Saint-Quentin, France, 2008. [3] J. Dedecek, Z. Tvaruzkova, Z. Sobalik, J. Am. Ceram. Soc. 91 (9) (2008) 3052e3057.
214
E. Papa et al. / Microporous and Mesoporous Materials 215 (2015) 206e214
[4] W.M. Kriven, J.L. Bell, M. Gordon, Ceram. Trans. 153 (2003) 227e250. [5] E. Landi, V. Medri, E. Papa, J. Dedecek, P. Klein, P. Benito, A. Vaccari, Appl. Clay Sci. 73 (2013) 56e64. [6] K. Okada, T. Isobe, K. Katsumata, Y. Kameshima, A. Nakajima, K.J.D. MacKenzie, Sci. Technol. Adv. Mater. 12 (2011) 11, 064701. [7] P. Colombo, Philos. Trans. R. Soc. A 364 (2006) 109e124. [8] W. Van Bonin, U. Nehen, U. Von Gizycki, US Patent 3,864,137, 1975. [9] J.L. Bell, W.M. Kriven, in: H.T. Lin, K. Koumoto, W.M. Kriven, E. Garcia, I.E. Reimanis, D.P. Norton (Eds.), Developments in Strategic Materials: Ceramic Engineering and Science Proceedings, 29, 2009, pp. 97e111 (10). [10] V. Medri, E. Papa, J. Dedecek, H. Jirlova, P. Benito, A. Vaccari, E. Landi, Ceram. Int. 39 (2013) 7657e7668. [11] S. Deville, Adv. Eng. Mater. 10 (3) (2008) 155e169. [12] W.L. Li, K. Lu, J.Y. Walz, Int. Mat. Rev. 57 (1) (2012) 37e60. [13] S. Deville, E. Saiz, A.P. Tomsia, Acta Mater. 55 (2007) 1965e1974. [14] S.R. Mukai, H. Nishihara, H. Tamon, Catal. Surv. Asia 10 (3e4) (2006) 161e171. [15] T. Fukasawa, Z.Y. Deng, M. Ando, T. Ohji, Y. Goto, J. Mater. Sci. 36 (10) (2001) 2523e2527. [16] K. Araki, J.W. Halloran, J. Am. Ceram. Soc. 88 (5) (2005) 1108e1114. [17] S. Deville, E. Saiz, A. Tomsia, Biomaterials 27 (32) (2006) 5480e5489. [18] E. Landi, F. Valentini, A. Tampieri, Acta Biomater. 4 (6) (2008) 1620e1626. [19] H. Zhang, I. Hussain, M. Brust, M.F. Butler, S.P. Rannard, A.I. Cooper, Nat. Mater. 4 (10) (2005) 787e793. [20] E. Landi, D. Sciti, C. Melandri, V. Medri, J. Eur. Ceram. Soc. 33 (2013) 1599e1607.
[21] H. Mori, K. Aontani, N. Sano, H. Tamon, J. Mater. Chem. 21 (2011) 5677e5681. [22] A. Lasalle, C. Guizard, E. Maire, J. Adrien, S. Deville, Acta Mater. 60 (2012) 4594e4603. [23] V. Medri, S. Fabbri, J. Dedecek, Z. Sobalik, Z. Tvaruzkova, A. Vaccari, Appl. Clay Sci. 50 (2010) 538e545. [24] L.A. Feldkamp, L.C. Davis, J.W. Kress, J. Opt. Soc. Am. 1 (1984) 612e619. [25] S. Bugani, M. Camaiti, L. Morselli, E. Van de Casteele, K. Janssens, Anal. Bioanal. Chem. 391 (2008) 1343e1350. [26] Z. Zuhua, Y. Xiao, Z. Huajun, C. Yue, Appl. Clay Sci. 43 (2009) 218e223. [27] K. Okada, A. Ooyama, T. Isobe, Y. Nakajima, K.J.D. MacKenzie, J. Eur. Ceram. Soc. 29 (2009) 1917e1923. [28] K.S.W. Sing, D.H. Everett, R.A.W. Haul, L. Moscou, R.A. Pierotti, J. Rouquerol, T. Siemieniewska, Pure Appl. Chem 57 (1985) 603e619. [29] J.D. Hunt, A. Kavner, E.A. Schauble, D. Snyder, C.E. Manning, Chem. Geol. 283 (2011) 161e170. [30] T.T. Trinh, A.P.J. Jansen, R.A. van Santen, E.J. Meijer, J. Phys. Chem. C 113 (2009) 2647e2652. [31] S. Deville, E. Maire, A. Lasalle, A. Bogner, C. Gauthier, J. Leloup, C. Guizard, J. Am. Ceram. Soc. 93 (9) (2010) 2507e2510. [32] S. Deville, E. Maire, A. Lasalle, A. Bogner, C. Gauthier, J. Leloup, C. Guizard, J. Am. Ceram. Soc. 92 (11) (2009) 2497e2503. [33] S. Bugani, M. Camaiti, L. Morselli, E. Van de Casteele, K. Janssens, X-Ray Spectrom. 36 (2007) 316e320.