Doped iron oxide scaffolds with gradient porosity fabricated by freeze casting: Pore morphology prediction and processing parameters

Depósito de Investigación de la Universidad de Sevilla https://idus.us.es/ This is an Accepted Manuscript of an article published by Sage: Lloreda-Jurado PJ, Pérez-Soriano EM, Paúl A, Herguido J, Peña JA, Sepúlveda R. Doped iron oxide scaffolds with gradient porosity fabricated by freeze casting: Pore morphology prediction and processing parameters. Materials Science and Technology. 2020;36(11):1227-1237. doi:10.1080/02670836.2020.1765096 © 2020 Institute of Materials, Minerals and Mining En idUS Licencia Creative Commons CC BY-NC-ND 1 Doped iron oxide scaffolds with gradient porosity fabricated by freeze casting: pore morphology prediction and processing parameters Lloreda-Jurado P.J.a; Pérez-Soriano, E.M.a; Paúl, A.a; Herguido, J.b; Peña, J.A.b; Sepúlveda R.a aDepartment of Materials Engineering and Science, E.T.S. de Ingenieros, Universidad de Sevilla, Avda. Camino de los Descubrimientos s/n., 41092 Sevilla, Spain. bAragon Institute of Engineering Research (I3A), Universidad de Zaragoza, Mariano Esquillor, s/n, Ed. “I+D”, E50.018 Zaragoza, Spain ABSTRACT Iron oxide scaffolds with gradient porosity were fabricated using the freeze-casting technique. Doped iron oxide particles were employed to generate stable particle suspensions in camphene using the ball-milling technique. Viscosity measurements were performed to determine the effect of the milling time and the additive incorporation. The freeze-casting process implemented in this work created a wide diversity of solidification front velocities, 𝑉 (2-80 m/s), and thermal gradients, 𝐺 (0.2- 20 C/cm) during sample fabrication. The solidification problem of liquid/solid systems has often been reviewed, but usually only under steady-state conditions. This study presents a much simpler approach for the correlation of the average dendrite arm spacing (𝜆) and pore morphology with different processing parameters. List of symbols and units TO Doped iron oxide, Triple Oxide FC Freeze Casting PS Polystyrene PTFE Polytetrafluoroethylene SOEC Solid oxide electrochemical cell DSC Differential Scanning Calorimetry 2 OM Optical Microscopy FESEM Field Emission Scanning Electron Microscopy ℎ Sample height (mm) 𝑉 Solidification front velocity (cm·s-1) 𝐺 Average thermal gradient (ᵒC·cm-1) 𝐶𝑅 Cooling rate (ᵒC·min-1) 𝜆 Average dendrite arm spacing (µm) 𝑇𝑜 Temperature at the beginning of the freezing process (ᵒC) 𝑇𝑠𝑠 Solidification temperature of the TO/camphene suspension (ᵒC) 𝑡𝑠𝑠 Time at 𝑇𝑠𝑠 in each thermocouple during freezing process (min) 𝛼 Mid solidification time between thermocouples (min) 𝑇𝛼 Temperature on thermocouple measured at time 𝛼 (ᵒC) 𝐿𝑙 Latent heat for camphene (J·g-1) 𝑇𝑠𝑙 Solidification temperature of camphene (ᵒC) 𝐿𝑠 Latent heat of the TO/camphene suspension (J·g-1) 𝜃 Particle loading (%) 𝜌𝑙 Density of the liquid camphene (g·cm-3) 𝜌𝑠𝑙𝑢𝑟𝑟𝑦(𝑙) Density of the TO/camphene suspension (g·cm-3) 𝜌𝑇𝑂 Density of TO powder (g·cm-3) 𝜏 Shear stress (Pa) 𝜏𝑜 Yield stress (Pa) 𝜂 Viscosity (mPa·s) 𝛾 Shear rate (s-1) 𝑉𝑐𝑟 Critical velocity for particle engulfment (cm·s-1) 𝑎𝑜 Interatomic distance (cm) ∆𝜎𝑜 Difference in interfacial particle/liquid energy (J) 𝜇 Ratio of particle to liquid thermal conductivity (J·s-1) 𝑅 Particle radius (cm) 1. Introduction Porous materials show distinctive properties, such as low density, high surface area, unique mechanical performance, damping capability, and high chemical reactivity. These properties render them ideal for applications, including biomedical prostheses and implants [1], sensors [2], pharmaceuticals [3], batteries [4], and catalytic applications [5], etc. Numerous techniques produce porous materials, ranging from powder metallurgy with space-holders [6], gas injection into liquid metals (GASAR) [7], to 3D printing techniques [8]. The Freeze Casting (FC) technique has acquired increasing 3 relevance in recent decades in fabricating porous materials, because not only is it costeffective and, environmentally friendly, but it is also, scalable to an industrial operation. The FC technique consists of four sequential steps: 1) creation of a stable powder suspension, usually with water or camphene among other additives; 2) directional solidification of the previous suspension, where the solvent dendrite grows and pushes the powder into the interdendritic spaces; 3) sublimation of the solid solvent; and 4) heat treatment. The most common raw materials are ceramic powders since submicrometric particles with low contamination are easy to obtain. Furthermore, Ti powder [9], polysiloxane polymer particles [10], and nanoparticles of Fe2O3 [11] are currently being employed to create porous material by FC. However, no industrial or commercial application has been developed yet. In this regard, Xu et al [12] created a battery whereby a solid oxide electrochemical cell (SOEC) was integrated to a porous redox unit made with Fe2O3 particles, since the Fe2O3 redox process constitutes a feasible and valid alternative for the production of hydrogen [13]. Unfortunately, the redox unit could reduce its efficiency with the redox cycles, since particles might over-sinter with time, thereby reducing its specific surface area. In 2016, comparable redox performance was found for camphene-based hematite pellets manufactured by FC with an optimal lab-made solid of doped Fe2O3 [14]. After three redox cycles, the macroporosity produced by the FC seemed to compensate the higher oxygen mobility enhanced by the ceria. More recently, water-based Fe foams manufactured by FC have shown significant modifications at their pore structure, having been subjected to ten redox cycles [15], whereas the formation of a denser outer shell decreased the open porosity. Clearly, redox performance of the porous Fe-based materials seems to be influenced by the pore structure and, indeed, by the dispersion medium. Camphene-based porous materials are able to develop diverse pore microstructures due to morphological modifications in dendrites during solidification. The growth velocity [16] and final microstructure parameters [17] of the camphene dendrites have been previously examined. Miller et al [18] developed a mathematical model based on the two-phase Stefan problem to predict final pore size (and tortuosity) on alumina particles suspensions in camphene or water. Additionally, Stefanescu [19] focused his research 4 on the interaction between particles and the moving solidification front of a transparent organic material, and established a critical velocity at which the pushing-to-engulfment transition occurs. These studies were based on steady-state conditions, modelled using intensive mathematical computing, and through ascertaining the particle surface properties. However, steady-state conditions are difficult to scale, computing simulation could require a significant amount of time, and certain specific material properties are difficult to obtain due to the lack of valid or precise information. To the best of our knowledge, scatter data is available on the particle or additives interaction with camphene. In this work, a fully disperse camphene slurry with doped iron oxide particles was first achieved and subsequently characterized to fabricate porous materials by FC with various pore morphology. The influence of the temperature gradient applied, the sonication, and the dispersant agent were correlated with the final pore size and morphology. The aim of this research is to fabricate porous materials suitable for redox performance, review the processing parameters and to present an easier method for the prediction of the final pore microstructure. 2. Experimental procedures 2.1 Materials Doped iron oxide, called Triple Oxide (TO), was used as starting powder material. TO was synthesized using the gel-citrate method [20]. The final composition was approximately 98 wt% Fe2O3, 1.75 wt% Al2O3, and 0.25 wt% CeO2. Particle size was maintained under 160 m in diameter, after being milled and sieved. Camphene (C10H16, 95 % purity) was chosen as a solvent. Zephrym, an oligomeric polyester provided by CRODA, was used as a powder dispersant agent. Polystyrene (PS, Mw = 350,000 g/mol, Sigma Aldrich) was used as a binder. 2.2 Preparation of suspensions and freeze-casting procedure 5 The TO powder was ball-milled for 24 h to further reduce its particle size. A 786 ml stainless steel jar was filled with 25 g of powder and 140 ml distilled water, with a ball- to-powder ratio of 64:1 in weight using steel balls 5 mm in diameter. The particle size distribution was measured by a laser diffractometer (Hydro QS-M, MALVERN Instrument), on samples extracted after 2, 4, 8, 16 and 24 h of milling time. Suspensions of 5 vol% of TO with 3 wt% of dispersant agent and 20 vol% of PS, according to the initial powder load, were prepared by ball-milling in a 30 ml sealed glass container at 60 C. All materials were sequentially incorporated to reduce undesired interactions. First, camphene and dispersant were mixed for 30 min. Subsequently, the TO powder was incorporated and milled for 8 h. Finally, PS was added and stirred for 3 more hours. A solidification mould of 30 mm in diameter and 18 mm in height was made with a water-cooled copper base and PTFE walls. The mould ensemble was then placed inside an incubator (INCU-Line 68R, VWR International Eurolab S.L.) and heated up to 60 C. Four thermocouples were uniformly located along the central mould axis to monitor the temperature evolution during sample fabrication. Figure 1a shows a diagram of the complete experimental setup. The TO/camphene suspension was then poured and prerested for 15 minutes before the freezing process began. The freezing process was directionally promoted by: 1) cooling the copper base with a constant flow of water at 30, 35, 40, or 42.5 C; and 2) gradually reducing the temperature inside the incubator until complete solidification (approximately 0.2 C/min). Samples were frozen from 60 C to 30, 35, 40 and 42.5 C, and named as 30, 25, 20 and 17.5 respectively regarding the initial temperature gradient. In addition, in order to determine the influence of various processing parameters, an extra set of samples were frozen down to 42.5 C as follows: a) the samples were sonicated for 15 minutes in an ultrasonic bath (USC-TH, International Eurolab S.L.) before manual pouring (𝑆), b) only 1 wt% of dispersant agent (𝐷) was added; and c) the preheated mould temperature was reduced to 50 C (sample 7.5). All samples were referenced according to the initial temperature gradient applied, and the processing conditions employed. Table 1 summarizes the freezing and processing parameters employed in this work. 6 The samples were subsequently subjected to forced airflow at room temperature (20 C) to allow camphene sublimation during 72 h. Finally, samples were air sintered up to 1100 C using a tubular furnace (Thermolyne 21100, ThermoFisher Scientific), following a two-step cycle described previously [11]. 2.3 Rheological measurements A rotational viscometer Myr VR 300 (Viscotech Hispania S.L.) was employed to determine the viscosity evolution on the TO/camphene suspension during the ballmilling procedure. Viscosity measurements were performed on two suspension batches at different milling times: • Batch 1: 4 h and 8 h after the powder had been incorporated into the camphenedispersant mixture and after 3 h once PS had been incorporated. • Batch 2: 4 h, 8 h, and 12 h after the addition of powder with no PS incorporation. In order to reduce the presence of air bubbles, the TO/camphene suspensions were first sonicated for 15 min, and then kept inside the viscometer chamber at 60 C for another 15 min before each measurement. 2.4 Thermal behaviour of TO/camphene suspensions. Differential scanning calorimetry (DSC) tests were performed over raw camphene and the TO/camphene suspension, to determine their latent heats (𝐿𝑙 and 𝐿𝑠) and solidification temperatures (𝑇𝑠𝑙 and 𝑇𝑠𝑠). The DSC test consisted of a two heating and cooling cycle from 25 C to 60 C, with 1 C/min ramps. A 5 minute dwell time at every peak temperature was implemented to ensure thermal stability. The temperature evolution was monitored during the solidification process for each sample using 4 thermocouples (Figure 1b). The first thermocouple was placed in contact with the mould base (ℎ = 0 mm), and the rest were arranged with a 4 mm span ( ℎ = 4, 8, and 12 mm). Therefore, the recorded temperature profile allowed the determination 7 of: cooling rate (𝐶𝑅ℎ), solidification front velocity (𝑉ℎ), and the average thermal gradient (𝐺ℎ), through the sample height according to the following equations: 𝐶𝑅ℎ= 𝑇𝑜−𝑇𝑠𝑠 𝑡𝑠𝑠 ℎ [ C 𝑚𝑖𝑛] ,( ℎ=0,4,8,12 𝑚𝑚) (1) 𝑉ℎ= 0.4 𝑐𝑚 𝑡𝑠𝑠 ℎ− 𝑡𝑠𝑠 ℎ−4 [𝑐𝑚 𝑠] ,(ℎ =4,8,12 𝑚𝑚) (2) 𝐺ℎ= 𝑇𝛼 ℎ− 𝑇𝛼 ℎ−4 0.4 𝑐𝑚 [ ℃ 𝑐𝑚] 𝑤𝑖𝑡ℎ 𝛼 =(𝑡𝑠𝑠 ℎ−𝑡𝑠𝑠 ℎ−4) 2 ⁄ , (ℎ=4,8,12 𝑚𝑚) (3) where To is the temperature of the slurry at the beginning of the solidification process, 𝑇𝑠𝑠 is the slurry solidification temperature, 𝑡𝑠𝑠 ℎ is the time in minutes to reach 𝑇𝑠𝑠 at ℎ, and 𝑇𝛼 ℎ is the temperature measured on ℎ at the time 𝛼. 2.5 Pore morphology characterization Optical Microscopy (OM) and Field Emission Scanning Electron Microscopy (FESEM) images were taken at the central axial plane across the sample height for pore size and morphology characterization. Average dendrite arm spacing (𝜆) was calculated as the sum of the mean pore size and cell wall thickness, and those were estimated by image analysis using the non-redundant maximum-sphere-fitted technique [21]. Figure 2 summarizes this concept for sample 17.5. An area in the range between 0.03 to 0.8 mm2 was analysed on the section of each sample to ensure accurate statistic calculations. 3. Results and discussion 3.1 Initial ball-milling 8 Figure 3 shows the evolution of the particle size distribution at different milling times. After 2 h of milling, TO particles were rapidly reduced in size to an average of 0.2 m in diameter. No significant changes were observed until 24 h, where TO particles appeared to aggregate, which could be attributed to particle cold fusion during extended milling times. The FESEM images inserted in Figure 3 reveal the morphology changes in the TO powder. The sharp edges of the raw TO particles were completely rounded after 8 h of milling. This period was therefore selected to produce the necessary TO powder for the fabrication of the porous samples. 3.2 Thermal characterization Table 2 shows the experimental values of latent heat (𝐿𝑙) and solidification temperature (𝑇𝑠𝑙) for camphene and for TO/camphene suspension (𝐿𝑠, 𝑇𝑠𝑠) respectively. 𝐿𝑙 determined at 20.13 J/g is in good agreement with the literature [22], while 𝐿𝑠 exhibited a lower value (14.93 J/g) due the TO particle addition. Indeed, 𝐿𝑠 can be estimated according to: 𝐿𝑠=(1−𝜃)𝜌𝑙𝐿𝑙 𝜌𝑠𝑙𝑢𝑟𝑟𝑦(𝑙) [𝐽𝑔 ⁄] (4) where 𝜃 is the particle loading, 𝜌𝑙 is the density of the liquid, 𝐿𝑙 is the latent heat of solidification for the liquid, and 𝜌𝑠𝑙𝑢𝑟𝑟𝑦(𝑙) is the density of the working suspension. Therefore, by using the 𝐿𝑙 as the camphene latent heat obtained in this research, taking the 𝜌𝑙=0.803 𝑔 𝑐𝑚3 ⁄ reported elsewhere [23], and the calculated 𝜌𝑠𝑙𝑢𝑟𝑟𝑦(𝑙)= 1.03 𝑔 𝑐𝑚3 ⁄ (TO powder density was determined by helium picnometry in 𝜌𝑇𝑂 = 5.31 𝑔 𝑐𝑚3 ⁄), it was found that the 𝐿𝑠 of the TO/camphene suspension with a 5 vol% TO particles could be estimated in 14.93 J/g according to Equation 4. Thus, the 𝐿𝑠 experimentally obtained in this investigation is consistent and shows a negligible influence of the additive. Therefore, the 𝑇𝑠𝑠 obtained in this work for the TO/camphene suspension (43.31 C) could be positively used as the 𝑇𝑠𝑠 described in Section 2.4. 15 Acknowledgements Financial support for this work has been provided by the Spanish Ministerio de Economía, Industria y Competitividad (MINECO), through the project MAT2016-76713- P. Lloreda-Jurado P.J. also thanks to the Universidad de Sevilla for the financial support (grant PIF II.2A, through VI Plan Propio de Investigación). 16 5. REFERENCES 1. Naleway SE, Fickas KC, Maker YN, et al (2016) Reproducibility of ZrO2-based freeze casting for biomaterials. Mater Sci Eng C 61:105–112. https://doi.org/10.1016/j.msec.2015.12.012 2. Hou C, Wang H, Zhang Q, et al (2014) Highly conductive, flexible, and compressible all-graphene passive electronic skin for sensing human touch. Adv Mater 26:5018–5024. https://doi.org/10.1002/adma.201401367 3. Szepes A, Ulrich J, Farkas Z, et al (2007) Freeze-casting technique in the development of solid drug delivery systems. Chem Eng Process Process Intensif 46:230–238. https://doi.org/10.1016/j.cep.2006.06.004 4. Zhai H, Xu P, Ning M, et al (2017) A Flexible Solid Composite Electrolyte with Vertically Aligned and Connected Ion-Conducting Nanoparticles for Lithium Batteries. Nano Lett 17:3182–3187. https://doi.org/10.1021/acs.nanolett.7b00715 5. Kim K, Lee T, Kwon Y, et al (2016) Lanthanum-catalysed synthesis of microporous 3D graphene-like carbons in a zeolite template. Nature 535:131–135. https://doi.org/10.1038/nature18284 6. Chen Y, Kent D, Bermingham M, et al (2017) Manufacturing of graded titanium scaffolds using a novel space holder technique. Bioact Mater 2:248–252. https://doi.org/10.1016/j.bioactmat.2017.07.001 7. Zhang X, Li Y, Liu Y, Zhang H (2013) Fabrication of a bimodal micro/nanoporous metal by the Gasar and dealloying processes. Mater Lett 92:448–451. https://doi.org/10.1016/j.matlet.2012.10.090 8. Ran Q, Yang W, Hu Y, et al (2018) Osteogenesis of 3D printed porous Ti6Al4V implants with different pore sizes. J Mech Behav Biomed Mater 84:1–11. https://doi.org/10.1016/j.jmbbm.2018.04.010 17 9. Chino Y, Dunand DC (2008) Directionally freeze-cast titanium foam with aligned, elongated pores. Acta Mater 56:105–113. https://doi.org/10.1016/j.actamat.2007.09.002 10. Zhang H, D’Angelo Nunes P, Wilhelm M, Rezwan K (2016) Hierarchically ordered micro/meso/macroporous polymer-derived ceramic monoliths fabricated by freezecasting. J Eur Ceram Soc 36:51–58. https://doi.org/10.1016/j.jeurceramsoc.2015.09.018 11. Sepúlveda R, Plunk AA, Dunand DC (2015) Microstructure of Fe2O3 scaffolds created by freeze-casting and sintering. Mater Lett 142:56–59. https://doi.org/10.1016/j.matlet.2014.11.155 12. Xu N, Li X, Zhao X, et al (2011) A novel solid oxide redox flow battery for grid energy storage. Energy Environ Sci 4:4942–4946. https://doi.org/10.1039/C1EE02489B 13. Romero E, Soto R, Durán P, et al (2012) Molybdenum addition to modified iron oxides for improving hydrogen separation in fixed bed by redox processes. Int J Hydrogen Energy 37:6978–6984. https://doi.org/10.1016/j.ijhydene.2011.11.066 14. Durán P, Lachén J, Plou J, et al (2016) Behaviour of freeze-casting iron oxide for purifying hydrogen streams by steam-iron process. Int J Hydrogen Energy 41:19518– 19524. https://doi.org/https://doi.org/10.1016/j.ijhydene.2016.06.062 15. Wilke SK, Dunand DC (2019) Structural evolution of directionally freeze-cast iron foams during oxidation/reduction cycles. Acta Mater 162:90–102. https://doi.org/10.1016/j.actamat.2018.09.054 16. Rubinstein ER, Glicksman ME (1991) Dendritic growth kinetics and structure II. Camphene. J Cryst Growth 112:97–110. https://doi.org/10.1016/0022-0248(91)90915- R 17. Çadirli E, Marasli N, Bayender B, Gündüz M (2000) Dependency of the microstructure parameters on the solidification parameters for camphene. Mater Res Bull 35:985–995. https://doi.org/10.1016/S0025-5408(00)00285-3 18 18. Miller SM, Xiao X, Faber KT (2015) Freeze-cast alumina pore networks: Effects of freezing conditions and dispersion medium. J Eur Ceram Soc 35:3595–3605. https://doi.org/10.1016/j.jeurceramsoc.2015.05.012 19. Stefanescu DM, Phalnikar R V., Pang H, et al (1995) A Coupled Force Field- Thermal Field Analytical Model for the Evaluation of the Critical Velocity for Particle Engulfment. ISIJ Int 35:700–707. https://doi.org/10.2355/isijinternational.35.700 20. Plou J, Durán P, Herguido J, Peña JA (2015) Hydrogen from synthetic biogas by catalyzed MDR and SIP: Screening of catalyst and iron oxide mixtures. Fuel 140:470– 476. https://doi.org/http://dx.doi.org/10.1016/j.fuel.2014.09.116 21. Hildebrand T, Rüegsegger P (1997) A new method for the model-independent assessment of thickness in three-dimensional images. J Microsc 185:67–75. https://doi.org/10.1046/j.1365-2818.1997.1340694.x 22. Štejfa V, Fulem M, Růžička K, Červinka C (2014) Thermodynamic study of selected monoterpenes II. J Chem Thermodyn 79:272–279. https://doi.org/10.1016/j.jct.2013.12.012 23. Yaws CL (2014) Chapter 3 Density of Liquid – Organic Compounds 24. Zhou L, Zhang H, Zhang H, Zhang Z (2013) Homogeneous nanoparticle dispersion prepared with impurity-free dispersant by the ball mill technique. Particuology 11:441–447. https://doi.org/10.1016/j.partic.2013.01.001 25. Felix M, Perez-Puyana V, Paúl A, Sepúlveda R (2019) Camphene/polystyrene solutions: A rheological approach for material processing industry. J Appl Polym Sci 136:. https://doi.org/10.1002/app.47953 26. Yoon BH, Choi WY, Kim HE, et al (2008) Aligned porous alumina ceramics with high compressive strengths for bone tissue engineering. Scr Mater 58:537–540. https://doi.org/10.1016/j.scriptamat.2007.11.006 27. Çadırlı E, Maraslı N, Bayender B, Gündüz M (2000) Dependency of the microstructure parameters on the solidification parameters for camphene. Mater Res Bull 35:985–995. https://doi.org/http://dx.doi.org/10.1016/S0025-5408(00)00285-3 19 28. Kurz W, Fisher DJ (1981) Dendrite growth at the limit of stability: Tip radius and spacing. Acta Me 29:11–20. https://doi.org/0001-6160/81/0101-0011S02.00/0 29. Trivedi R, Kurz W (1988) Metals Handbook, Ninth Edit 30. Sen S, Kaukler WF, Curreri P, Stefanescu DM (1997) Dynamics of solid/liquid interface shape evolution near an insoluble particle - An X-ray transmission microscopy investigation. Metall Mater Trans A Phys Metall Mater Sci 28:2129–2135. https://doi.org/10.1007/s11661-997-0170-y 31. Rubinstein ER, Glicksman ME (1991) Dendritic growth kinetics and structure II. Camphene. J Cryst Growth 112:97–110. https://doi.org/http://dx.doi.org/10.1016/0022-0248(91)90915-R 32. Almansoori Z, Khorshidi B, Sadri B, Sadrzadeh M (2018) Parametric study on the stabilization of metal oxide nanoparticles in organic solvents: A case study with indium tin oxide (ITO) and heptane. Ultrason Sonochem 40:1003–1013. https://doi.org/10.1016/j.ultsonch.2017.09.012 33. Deville S (2008) Freeze-Casting of Porous Ceramics: A Review of Current Achievements and Issues. Adv Eng Mater 10:155–169. https://doi.org/10.1002/adem.200700270 34. LI Y min, LIU X quan, LUO F hua, YUE J ling (2007) Effects of surfactant on properties of MIM feedstock. Trans Nonferrous Met Soc China (English Ed 17:1–8. https://doi.org/10.1016/S1003-6326(07)60039-9 35. Moloney VMB, Parris D, Edirisinghe MJ (1995) Rheology of Zirconia Suspensions in a Nonpolar Organic Medium. J Am Ceram Soc 78:3225–3232. https://doi.org/10.1111/j.1151-2916.1995.tb07Wide 958.x 20 Figures Figure 1. Experimental setup of the freeze casting system (a), and temperature evolution profile recorded during the freezing of sample 25 (b). The solidification mould configuration includes the thermocouple distribution (T1-T2) inside the poured sample. Solidification front velocity (𝑉ℎ) and average thermal gradient (𝐺ℎ) were measured along the sample height at the solidification temperature of the TO/camphene suspension (𝑇𝑠𝑠), as shows in detail for T1 and T2. 21 Figure 2. Sequential Image processing performed by the ImageJ/BoneJ software to obtain the pore size distributions of sample 17.5: a) grayscale OM image, b) binary OM image, c) coloured spheres for pore size calculation and d) pore size histogram. 22 Figure 3. Particle size distributions of the TO powder for different ball-milling times. 23 Figure 4. Shear stress (𝜏) evolution of TO/camphene suspension as a function of the shear rate (𝛾) for different ball milling time. Viscosity values were calculated from the Bingham model. 24 Figure 5. Evolution of the pore morphology over the sample height as a function of the initial temperature different (samples 30, 25, 20 and 17.5) and the processing parameters (samples 17.5, 17.5S, 7.5S, 17.5SD). Red arrows in the figure shows the extension of the elongated dendritic pore morphology. FESEM micrographs were taken using the same magnification (400x). Scale bar placed at the bottom right of the figure represents 100 m.