Novel B4C supports for ceramic membrane filtration

Novel B 4 C supports for ceramic membrane filtration Juan R. Madrigal a , Manuel A. García-Gal´ an a , Fernando Guiberteau a , Victor M. Candelario b , Angel L. Ortiz a,* a Departamento de Ingeniería Mec´ anica, Energ´ etica y de los Materiales, Universidad de Extremadura, 06006, Badajoz, Spain b Departamento de Ingeniería y Ciencia de los Materiales y del Transporte, Escuela Polit´ ecnica Superior, Universidad de Sevilla, Calle Virgen de ´ Africa 7, 41011, Sevilla, Spain ARTICLE INFO Handling Editor: Dr P. Vincenzini Keywords: Porous ceramics B 4 C Ceramic membrane supports Ceramic processing Filtration applications ABSTRACT Novel complex-shaped porous B 4 C supports were fabricated for use in ceramic membrane filtration applications, the processing of which is described in detail. Firstly, an extrudable ceramic paste was formulated and homogenised, containing a high content of B 4 C of super-coarse and ultrafine grades, plus water as liquid medium and various organic additives as thickener, binder, plasticiser, and lubricant. Secondly, tubular honeycomb parts (i.e., cylinders of ~26 mm outer diameter with 30 inner channels of 3 mm diameter each) were extruded at ~25 bar and dried in air at 120 ◦C for 24 h, resulting in robust green B 4 C membrane supports without macro- or microdefects. Thirdly, suitable debinding conditions were identified by thermogravimetry of the paste and its components, and the dry membrane supports were debinded at 700 ◦C for 1.5 h in high vacuum (<0.1 Pa) confirming that they retain their shape with minimal isotropic shrinkage (~0.13–0.15 %). And fourthly, the debinded membrane supports were pressureless sintered at 2000 ◦C for 3 h in inert atmosphere undergoing only an additional 1 % isotropic shrinkage, resulting in B 4 C membrane supports with low apparent density (~0.89(1) g/ cm 3 ), high porosity (~38–39 %), open interconnected micrometre pores (in the range ~0.5–16 μ m, and d 50 ~1.5 μ m), and low skeletal density (~2.455(1) g/cm 3 ). Importantly, their only partial densification makes them both sufficiently strong mechanically (~68(2) MPa compressive strength) and highly permeable to water (~13131 (418) l/(m 2 ⋅h⋅bar) permeability at ~1.3 bar). These attributes, together with the intrinsic lightness and durability of B 4 C ceramics, make these novel supports in principle very appealing for ceramic membrane filtration and other applications. 1. Introduction Water, a human right according to the United Nations, is becoming increasingly scarce, to such an extent that it is recognised that there is an imminent risk of a global water crisis if widespread policies of sustainable water consumption are not adopted [1–3]. These policies must necessarily include solutions for wastewater treatment and recycling, and an emerging technology for this is ceramic membrane filtration [1, 4–9]. Briefly, this involves the physical separation of contaminants from water using porous solids with high open and interconnected porosity, typically consisting of a thin, fine-grained, small-pored ceramic coating, which is the active layer for filtration, deposited on a thick, coarse-grained, large-pored ceramic support, which provides mechanical integrity to the part and dictates how filtration occurs (i.e., inside-out, dead-end, or outside-in filtrations). The commonest support designs are monotubes, honeycomb tubulars, and multi-channeled flat-sheets and discs. Ceramic filters are more expensive and difficult to scale up industrially than polymeric filters, but have higher performance, operability, and durability [1,6,7,10–12]. Ceramic filters made of some oxides (e.g., Al 2 O 3 , TiO 2 , and ZrO 2 ) and SiC are currently commercially available [13,14], many of which are “hybrid”(i.e., support and membrane layer of different ceramics) [13–16]. SiC filters are the most durable and suitable, especially in harsh filtration conditions, because SiC is harder, stronger, and more refractory, as well as more hydrothermally stable, hydrophilic, and fouling resistant than oxides [6,7,17]. It is also lighter, an attribute particularly relevant to reducing the spurious weight of the support (which accounts for almost the entire weight of the ceramic filter). In this respect, fabricating membrane supports from B 4 C, a harder and lighter carbide than SiC, would reduce the spurious weight by another ~21 % (since the absolute densities of B 4 C and SiC are ~2.52 and 3.21 g/cm 3 , respectively), with no loss of mechanical durability or * Corresponding author. E-mail address: [email protected] (A.L. Ortiz). Contents lists available at ScienceDirect Ceramics International journal homepage: www.elsevier.com/locate/ceramint https://doi.org/10.1016/j.ceramint.2024.09.389 Received 26 July 2024; Received in revised form 25 September 2024; Accepted 28 September 2024 Ceramics International 50 (2024) 50436–50443 Available online 29 September 2024 0272-8842/© 2024 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license ( http://creativecommons.org/licenses/bync-nd/4.0/ ). even a gain. To the best of the authors’knowledge, there are no previous studies on the fabrication of ceramic filters made entirely of B 4 C. However, although they are not commercially available products, hybrid ceramic filters consisting of a B 4 C membrane layer deposited on a commercial flat-sheet SiC support have already been fabricated [16]. Thus, what has not yet been fabricated are porous B 4 C supports for ceramic membrane filtration, the development of which requires a different approach than the current research on B 4 C ceramics focused on the fabrication of fully dense parts with fine-grained microstructures for contact-mechanical and tribological applications (e.g., wear parts and armour ceramics). Fortunately, the known poor solid-state sinterability of B 4 C, which makes its complete densification without sintering additives very difficult, if not impossible, except by hot pressing or especially spark plasma sintering, is on the contrary a desirable attribute for the fabrication of porous ceramic filters by pressureless sintering. Fabricating ceramic membrane supports, which are complex-shaped parts, is however in itself a processing challenge with many critical steps [18]. Thus for example, the paste must be formulated to be rheologically suitable for plastic forming (usually extrusion, and sometimes tape casting or injection moulding). Drying and debinding must be sufficiently gentle and uniform to prevent shrinkage stresses from causing deformation and/or fracture of the extrudate. And sintering must be adjusted to only partially densify the microstructure thus gaining mechanical strength while retaining high permeability. With this in mind, this study addresses for the first time the challenging fabrication of B 4 C membrane supports with a tubular honeycomb shape (which is probably the most demanded shape), focusing on aspects of ceramic processing and microstructural–textural characterisation complemented by preliminary tests of mechanical strength and permeability. 2. Experimental procedure The novel B 4 C membrane supports were fabricated from a custommade ceramic paste, which was extruded to the desired shape and the resulting extrudates were dried, debinded, and pressureless sintered. Briefly, two commercial B 4 C powders, one super-coarse (d mean ~22–59 μ m; Grade F280, abcr GmbH, Germany) and the other ultrafine (d 50 ~0.3–0.6 μ m; Grade HD20, H¨ ogan¨ as AB, Sweden), were acquired and characterised by scanning electron microscopy (SEM; S-3600N, Hitachi, Japan) and X-ray diffractometry (XRD; D8 Advance, Bruker AXS, Germany). Two B 4 C powders with very different particle sizes were used because the ultrafine particles can fill the interstices between the coarse particles, which in turn increases the number of initial contacts between particles. On the one hand this results in a better green packing density during extrusion and thus predictably in mechanically more robust green parts, and on the other hand the recrystallisation of the ultrafine particles is expected to favour neck growth during sintering, predictably resulting in a more mechanically robust porous solid skeleton. The other components of the ceramic paste were distilled water as liquid medium, methylcellulose (Culminal™MHPC-1014, Ashland Industries Deutschland GmbH, Germany) as thickener [19], poly(vinyl) alcohol (PVA; Optapix PAF 35, Zschimmer &Schwarz, Germany) with 35 % active matter as binder [20], polyethylene glycol (PEG; Panreac, Spain) with 400 g/mol molecular weight as plasticiser [20], and paraffin oil (Ceragi, Spain) as lubricant. Several ceramic pastes were formulated, all with the super-coarse and ultrafine B 4 C powders combined in a 3:1 ratio by weight but with different ratios by weight of B 4 C:water plus organics, which were homogenised with a spiral kneader (GHR-30 2V, Seven Italy, Italy) for 3 min. The resulting ceramic pastes were re-kneaded by extruding them twice through a rectangular die of 28 mm width by 17 mm height, and finally extruded through a circular die of ~26 mm outer diameter and 30 inner circular channels of 3 mm diameter each. A ceramic paste with adequate extrudability (considered as such when extruded as honeycomb tubes at 30 bar or less with shape retention) had ~77 wt% B 4 C (specifically ~57.8 wt% super-coarse powder plus ~19.2 wt% ultrafine powder), 15.2 wt% water, 3.6 wt% methylcellulose, 1.2 wt% PVA, 1.2 wt% PEG, and 1.8 wt% paraffin oil. The extrudates were dried in air, first at room temperature for 48 h and then gradually in an oven up to a maximum of 120 ◦C for 24 h, after which they were debinded. Suitable debinding conditions were determined by characterising both the (wet) ceramic paste and the supercoarse B 4 C, ultrafine B 4 C, methylcellulose, PVA, PEG, and paraffin oil used in its formulation by thermogravimetry (TG; Labsys Evo, Setaram, France). The TG cycle involved a first heating at 10 ◦C/min in air up to 100 ◦C with holding for 4.5 h (to roughly simulate drying), and a second heating at 10 ◦C/min in an inert atmosphere up to 1000 ◦C. In practice, debinding was performed in a retort furnace at 700 ◦C for 1.5 h under a high vacuum of <0.1 Pa, with heating and cooling at ~1.7 ◦C/min. The pieces were dimensionally measured (i.e., multiple measurements of the length and diameter of the honeycomb tubes) and weighed both before and after debinding to assess axial and radial shrinkages and mass loss, respectively. Finally, the debinded membrane supports were pressureless sintered in a graphite furnace (1000-3560-FP20, Thermal Technology Inc., USA) at 2000 ◦C for 3 h, with heating at 10 ◦C/min and cooling at 20 ◦C/min (which are in fact one order of magnitude faster than the typical ramps applied in large industrial graphite furnaces used for mass production). The sintering cycle involved isothermal holdings at 1450 ◦C and 1650 ◦C for 30 min to remove H 3 BO 3 /B 2 O 3 impurities typical in B 4 C powders, and was carried out under vacuum until the end of the 1650 ◦C holding and under flowing Ar thereafter. The pieces were also dimensionally measured and weighed both before and after sintering to assess shrinkages and mass loss, respectively. The dry, debinded, and sintered membrane supports were characterised by SEM to examine microstructures, by XRD to determine phase compositions, and by He pycnometry (Stereopycnometer SPY-D160-E, Quantachrome Instruments, UK) to measure skeletal densities. Additionally, the sintered membrane supports were characterised in much more detail, first textually by both distilled water immersion porosimetry (Archimedes method) and capillary flow porometry (CFP; Quantachrome 3G zh, Anton Paar, Austria) using Porofil™as wetting fluid to measure their porosity (including open, closed, and total porosities) and open interconnected pore size distribution, respectively, then mechanically by compressive tests (5969, Instron Corp., UK) to evaluate their strength, and finally functionally by cross-flow filterability tests (DeltaE S. r.l., Italy) at ~1.3 bar with distilled water to determine their permeability. Lastly, the sintered membrane supports were further microstructurally characterised by optical profilometry (OP; Profilm 3D, Filmetric, USA) to measure the surface roughness of their longitudinal inner channels. A commercially available SiC tubular honeycomb support (LiqTech A/S, Denmark) was also characterised by OP, and used as a reference for comparison. 3. Results and discussion 3.1. Characterisation of the starting B 4 C powders Figs. 1 and 2 show SEM images and the XRD patterns, respectively, of the two B 4 C powders used for fabricating the novel membrane supports. On the one hand, it can be seen in Fig. 1 that both consist of equiaxed particles and that, in accordance with the corresponding manufacturers’ specifications of d mean ~22–59 μ m and d 50 ~0.3–0.6 μ m, the supercoarse powder has particle sizes of several tens of microns and the ultrafine powder in the submicrometre range. On the other hand, it can be seen in Fig. 2 that both are mainly B 4 C with the typical C impurities, and that the ultrafine powder also has the usual H 3 BO 3 impurities of this grade. J.R. Madrigal et al. Ceramics International 50 (2024) 50436–50443 50437 3.2. Fabrication of the B 4 C membrane supports Fig. 3 shows photographs of two selected extrusions. It can be seen that Fig. 3A demonstrates the feasibility of continuously shaping the desired tubular honeycomb membrane supports without external macro-defects by extrusion of the ceramic paste formulated with ~77 wt % B 4 C and 15.2 wt% water, which has the plasticity required for extrudability at ~25 bar and ~1.4 cm/s with shape retention. On the other hand, it can be seen in Fig. 3B that pastes with relatively lower B 4 C:water ratios (i.e., ~72.5 wt% B 4 C and 18.1 wt% water) were too wet and soft to retain shape, while pastes with relatively higher B 4 C:water ratios were discarded as too dry and hard to be extrudable at conventional pressures. Note however that other support designs (e.g., flatsheet supports) may require pastes with different extrudability, and therefore formulation. Importantly, these pastes are water-based and therefore environmentally friendly, and the extrusion process is simple and directly scalable industrially, a combination that is very appealing for the mass production of these novel ceramic membrane supports. The extrudates were next dried in air at 120 ◦C for 24 h to remove the forming water, resulting in the dry membrane supports shown in Fig. 4 (where for direct comparison the corresponding debinded and sintered Fig. 2. XRD patterns, acquired with pure CuK α 1 incident radiation, of the super-coarse and ultrafine B 4 C starting powders, as indicated. Peak assignations, performed using the PDF2 database, are included. Fig. 3. Photographs showing the extrusion of tubular honeycomb membrane supports, in the form of cylinders of ~26 mm outer diameter with 30 inner channels of 3 mm diameter each, from (A) the B 4 C ceramic paste with adequate extrudability formulated with ~77 wt% B 4 C and 15.2 wt% water and (B) a paste with inadequate extrudability that is too wet and soft. The inset in (A) is a higher-magnification photograph of the cross-section of the extrudate. Fig. 1. SEM micrographs, taken with secondary electrons at 15 kV, representative of the (A) super-coarse and (B) ultrafine B 4 C starting powders. J.R. Madrigal et al. Ceramics International 50 (2024) 50436–50443 50438 membrane supports are also presented). It can be seen in Fig. 4A that even in their green condition these membrane supports are robust and safe to handle (and store), and were also found to be easily trimmed dry. Fig. 5 shows the TG curves as a function of time of the adequately extruded (wet) ceramic paste and of the super-coarse B 4 C, ultrafine B 4 C, methylcellulose, PVA, PEG, and paraffin oil used in its formulation, measured to identify an optimal debinding temperature for the dry membrane supports. Fig. S1 in supplementary material shows the corresponding TG curves as a function of temperature. It can be seen in Fig. 5A and S1A that the paste shows two stages of mass loss, one up to ~100 ◦C where it loses ~16.6 wt% and another between ~275 ◦C and 500 ◦C where it loses another ~4.3 wt% for a total of ~20.9 wt%. By ingredients, it can be seen in Fig. 5B and S1B that super-coarse B 4 C loses no mass and that ultrafine B 4 C loses ~0.5 wt%, attributable to the removal of all or part of its H 3 BO 3 impurities. As for the organics, it can be seen in Fig. 5B and S1B that paraffin oil is rapidly removed below ~100 ◦C, that PVA is first rapidly but partially removed (i.e., ~62 wt%) up to ~100 ◦C and gradually thereafter leaving little residue (i.e., ~1.3 wt%) from ~500 ◦C, that PEG is abruptly removed between ~200 ◦C and 400 ◦C leaving no residue, and finally that methylcellulose is abruptly removed between ~275 ◦C and 400 ◦C leaving ~13 wt% residue. Thus, the first stage of mass loss of the paste is due to its dehydration (i.e., removal of the forming water, the aqueous base of the PEG, and the liquid paraffin oil) and the second to the removal (i.e., PVA and dry PEG) and pyrolysis (i.e., methylcellulose) of the remaining organics. Also importantly, it can be seen in Fig. 5A and S1A that TG indicates that in principle any temperature above ~500 ◦C would already be suitable for the debinding of the dry membrane supports, although for large parts it is preferable in practice to use higher furnace debinding temperatures. Fig. 4 also shows a photograph of the membrane supports debinded at 700 ◦C for 1.5 h under high vacuum (Fig. 4B). It can be seen that they retained their external tubular and internal multi-channel shapes, with a mass loss of only ~5.2 wt%. The latter is consistent with the TG analysis. Fig. 4. Photographs of the (A) dry, (B) debinded, and (C) sintered B 4 C membrane supports. Fig. 5. TG curves, measured as a function of time under heating at 10 ◦C/min in air up to 100 ◦C, holding at 100 ◦C for 4.5 h, and heating at 10 ◦C/min in flowing Ar thereafter, of (A) the adequately extruded B 4 C ceramic paste and (B) its components except water, as indicated. The TG cycle is the same in (A) and (B). J.R. Madrigal et al. Ceramics International 50 (2024) 50436–50443 50439 The axial and radial shrinkages during debinding were ~0.13 % and ~0.15 %, respectively, indicating that they shrank only minimally and isotropically. This is because the greatest shrinkage occurs during drying as the extrudate dehydrates, which is the critical stage where stresses can arise leading to deformation or fracture [18], while debinding, once the so-called leather-hard condition is reached, results in only a slight mass loss with negligible shrinkage. Finally, Fig. 4 also shows a photograph of some membrane supports pressureless sintered at 2000 ◦C for 3 h under inert atmosphere (Fig. 4C). It can be seen that they also retained their external tubular and internal multi-channel shapes, with a mass loss of only ~3.1 wt% attributable to the removal of the H 3 BO 3 and C impurities (by the reactions 2H 3 BO 3 →B 2 O 3 +3H 2 O(g)→B 2 O 3 (g) and 2B 2 O 3 +7C→B 4 C+6CO(g)). The axial and radial shrinkages during sintering were ~1.0 % and ~1.3 %, respectively, indicating that they again shrank only minimally and isotropically. This suggests that, as desired, the membrane supports will have gained mechanical strength while maintaining both their internal channel morphology and their solid porous skeleton. Fig. 6 shows SEM images representative of the microstructure of the dry, debinded, and sintered membrane supports, evincing that the three are free of macro- or micro-cracks. Moreover, it can be seen in Fig. 6A and B that the dry and debinded membrane supports have a similar microstructure, although compositionally the former still contain certain organics while the latter does not, both consisting of a wellpacked network of super-coarse particles embedded in a porous mat of ultrafine particles, so that neither has the desired macro-/meso-porosity. It can also be seen in Fig. 6C that the microstructure of the sintered membrane support consists however of a porous network of supercoarse particles with fine (no longer ultrafine) particles in between, thus indicating the occurrence of partial recrystallisation during sintering with formation of the desired macro-/meso-porosity. The skeletal densities measured by He pycnometry are ~2.359(2), 2.460(2), and 2.455(1) g/cm 3 for the dry, debinded, and sintered membrane supports, respectively, values which are very close, although slightly lower probably due to H 3 BO 3 and/or C impurities, to the theoretical density of B 4 C (i.e., ~2.52 g/cm 3 ) from which they are made. Fig. 7 shows the XRD patterns of the dry, debinded, and sintered membrane supports. It can be seen that the first are B 4 C with the C and H 3 BO 3 impurities expected from the super-coarse and ultrafine B 4 C starting powders, indicating that B 4 C did not oxidise during the preparation and extrusion of the aqueous paste or during the drying in air of the extrudates up at 120 ◦C. It can also be seen that the second contain Fig. 6. SEM micrographs, taken at different magnifications with secondary electrons at 15 kV, representative of the microstructure of the (A) dry, (B) debinded, and (C) sintered B 4 C membrane supports. J.R. Madrigal et al. Ceramics International 50 (2024) 50436–50443 50440 comparatively more H 3 BO 3 impurities than the first, indicating that B 4 C slightly oxidised during debinding at 700 ◦C due, given that the vacuum in the retort furnace is initially high (i.e.,<0.1 Pa), to the oxidising species released during the removal/pyrolysis of certain organics. Whatever the case, this is irrelevant because the XRD pattern of the third shows that they are B 4 C with few C impurities and no H 3 BO 3 impurities, indicating that the former are partially consumed and the latter completely removed during sintering at 2000 ◦C in the graphite furnace (again by the reactions 2H 3 BO 3 →B 2 O 3 +3H 2 O(g)→B 2 O 3 (g) and 2B 2 O 3 +7C→B 4 C+6CO(g)). 3.3. Assessment of mechanical strength and permeability The sintered membrane supports have, with the internal channels excluded, an absolute density measured from their mass and dimensions of ~1.52(2) g/cm 3 , corresponding to a relative density of ~61 % (i.e., ~39 % porosity). The absolute and relative densities measured by the Archimedes method are ~1.56(2) g/cm 3 and ~62 % (i.e., ~38 % porosity), which are in good agreement with the above. The apparent density is even lower, of only ~0.89(1) g/cm 3 . Importantly, the closed and open porosities determined by the Archimedes method are ~1.1 % and 36.9 %, respectively, indicating that essentially the sintered membrane supports have no dead volume (i.e., all porosity is available for filtration). Fig. 8 shows the pore size distribution determined by CFP for the sintered membrane supports. It can be seen that there are open interconnected pores ranging in size from ~0.5 to 16 μ m, but that almost all have sizes between ~0.8 (d 10 ) and 3.4 μ m (d 90 ) and mostly ~1.5 μ m (d 50 ). Also importantly, the mechanical characterisation tests revealed that the sintered membrane supports have a compressive strength of ~68(2) MPa, and the filtration tests at ~1.3 bar that their permeability to distilled water is ~13131(418) l/(m 2 ⋅h⋅bar). Therefore, as desired for filtration applications, these novel sintered membrane supports reached the boundary between the initial and intermediate sintering regimes [21], microstructurally developing sufficient interparticle necks to be mechanically strong while retaining an abundant open interconnected porosity that makes them very permeable. It is difficult to compare the strength and permeability of different ceramic membrane supports because these properties vary a lot with overall porosity and pore size, but to put the above numbers at least into perspective note for example that commercial SiC flat-sheet supports with ~45 % porosity and pore sizes in the range ~0.25–15 μ m have a compressive strength and water permeability of ~30 MPa and ~4500 l/ (m 2 ⋅h⋅bar), respectively [22], and that commercial SiC tubular honeycomb supports with ~40 % porosity and mean pore size of ~15 μ m have a flexural strength and water permeability of ~30 MPa and ~12000 l/(m 2 ⋅h⋅bar), respectively [16]. Thus, the present B 4 C supports compare similarly, or even favourably, with extrusion-formed SiC supports already on the market (which are also the most similar compositionally). Extrusion-formed ceramic supports made of Al 2 O 3 , ZrO 2 , and TiO 2 are also commercially available, but their microstructural characteristics and properties without a membrane layer are not usually reported. On a laboratory scale, a wider variety of extrusion-formed tubular ceramic supports have been fabricated [23], for example from mineral coal fly ash (with 51 % porosity, 4.5 μ m pore size, and 19 MPa tensile strength), pozzolan (with 41 % porosity, 8.2 μ m pore size, and 15 MPa flexural strength), kaolin (with 46 % porosity, 1.4 μ m pore size, and 18 MPa flexural strength), Al 2 O 3 (with 42 % porosity, 1 μ m pore size, and 30 MPa flexural strength), clay (with 34–50 % porosity, 9.4–10.8 μ m pore size, and 1.5–12 MPa flexural strength), quartz +calcite (with 35–55 % porosity, 1.9–12.3 μ m pore size, and 8–18 MPa flexural strength), and SiC (with 44–47 % porosity, 31–36 μ m pore size, and 12–21 MPa flexural strength). Other non-commercial, tubular and non-tubular ceramic supports have been fabricated without extrusion forming, but are difficult to scale up industrially and to compare due to the wide disparity in designs, compositions, starting powders, organic and inorganic additives, and forming and sintering conditions. Importantly, the microstructural characteristics of the present B 4 C supports make them in principle interesting, either with this or other designs, for use directly in different applications. Some are water purification applications, such as the removal of particulate matter from secondary effluents and from swimming pool and spa water (in these Fig. 7. XRD patterns, acquired with pure CuK α 1 incident radiation, of the dry, debinded, and sintered B 4 C membrane supports, as indicated. Peak assignations, performed using the PDF2 database, are included. Fig. 8. Cumulative and differential distributions of pore sizes measured by CFP using Porofil™as wetting fluid for the sintered membrane support, as indicated. J.R. Madrigal et al. Ceramics International 50 (2024) 50436–50443 50441 cases as a substitute for sand filters). Others are food recovery applications, such as the fractionation of fat globules in milk [24,25] and the separation of proteins in milk, and others are gas purification applications, such as particulate filters for diesel and gasoline engines and turbines. Logically, the range of applications widens when the ceramic membrane layer is deposited in the inner channels. In this respect, Fig. 9 shows 3D OP images of the longitudinal inner channels of the present B 4 C supports and of reference SiC supports on which uniform and defect-free commercial SiC ultrafiltration membranes are currently deposited. Both have comparatively similar surface roughness (arithmetical mean high of ~8 μ m for the former vs ~9 μ m for the latter), indicating that the present B 4 C supports should also be amenable to coating with ceramic membrane layers. 4. Conclusions A study was conducted on the fabrication of novel complex-shaped porous B 4 C supports for ceramic membrane filtration applications. Based on the experimental results and analyses, the following conclusions can be drawn: 1. A concentrated B 4 C paste (~77 wt%) can be formulated with water (~15.2 wt%) and various organic additives (~7.8 wt%) that is extrudable into complex shapes (e.g., tubular honeycomb parts) at moderate pressure (~25 Pa in this case). 2. The resulting extrudates can be dried (120 ◦C for 24 h in this case) and debinded (~700 ◦C for 1.5 h) to remove the forming water and the organics with little residue without fracturing and retaining their complex external and internal shapes. 3. Pressureless sintering (2000 ◦C for 3 h in this case) results in minimal densification with sufficient development of interparticle neck, leading to sufficiently mechanically strong (~68(2) MPa compressive strength) and highly water permeable (~13131(418) l/ (m 2 ⋅h⋅bar) at ~1.3 bar) B 4 C membrane supports. 4. The textural (i.e., abundant open interconnected porosity), structural (i.e., low skeletal density), functional (i.e., high permeability), and mechanical (i.e., sufficient strength) attributes of these novel B 4 C supports make them in principle very appealing for ceramic membrane filtration applications. CRediT authorship contribution statement Juan R. Madrigal: Writing –review &editing, Methodology, Investigation, Formal analysis, Data curation. Manuel A. García-Gal´ an: Writing –review &editing, Methodology, Investigation, Formal analysis. Fernando Guiberteau: Writing –review &editing, Supervision, Methodology, Conceptualization. Victor M. Candelario: Writing –review &editing, Supervision, Methodology, Investigation, Formal analysis, Conceptualization. Angel L. Ortiz: Writing –original draft, Supervision, Methodology, Funding acquisition, Formal analysis, Conceptualization. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Fig. 9. 3-D OP images, taken with white light, representative of the longitudinal inner channels of (A) the present novel B 4 C supports and (B) commercial SiC supports. J.R. Madrigal et al. Ceramics International 50 (2024) 50436–50443 50442 Acknowledgements This work was supported by the Spanish Ministry of Science and Innovation under Grant no. TED2021-129188B–I00. Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi. org/10.1016/j.ceramint.2024.09.389. References [1] Z. He, Z. Lyu, Q. Gu, L. Zhang, J. Wang, Ceramic-based membranes for water and wastewater treatment, Colloids Surf., A 578 (2019) 123513. [2] Water for prosperity and peace, the united Nations world water development report. https://unesdoc.unesco.org/ark:/48223/pf0000388948, 2024. [3] SDG 6 Synthesis Report 2018 on Water and Sanitation, United Nations; http s://www.unwater.org/app/uploads/2018/12/SDG6_SynthesisReport2018_Watera ndSanitation_04122018.pdf. [4] P.S. Goh, A.F. Ismail, A review on inorganic membranes for desalination and wastewater treatment, Desalination 434 (2018) 60–80. [5] M. Kamali, D.P. Suhas, M.E. Costa, I. Capela, T.M. Aminabhavi, Sustainability considerations in membrane-based technologies for industrial effluents treatment, Chem. Eng. J. 368 (2019) 474–494. [6] E. Eray, V. Boffa, M.K. Jørgensen, G. Magnacca, V.M. Candelario, Enhanced fabrication of silicon carbide membranes for wastewater treatment: from laboratory to industrial scale, J. Membr. Sci. 606 (2020) 118080. [7] E. Eray, V.M. Candelario, V. Boffa, H. Safafar, D.N. Østedgaard-Munck, N. Zahrtmann, H. Kadrispahic, M.K. Jørgensen, A roadmap for the development and applications of silicon carbide membranes for liquid filtration: recent advancements, challenges, and perspectives, Chem. Eng. J. 414 (2021) 128826. [8] U. Mueller, G. Biwer, G. Baldauf, Ceramic membranes for water treatment, Water Supply 10 (6) (2010) 987–994. [9] A. Farsi, S.H. Jensen, P. Roslev, V. Boffa, M.L. Christensen, Inorganic membranes for the recovery of effluent from municipal wastewater treatment plants, Ind. Eng. Chem. Res. 54 (13) (2015) 3462–3472. [10] B. Hofs, J. Ogier, D. Vries, E.F. Beerendonk, E.R. Cornelisssen, Comparison of ceramic and polymeric membrane permeability and fouling using surface water, Sep. Purif. Technol. 79 (3) (2011) 365–374. [11] S. Geno Lehman, L. Liu, Application of ceramic membranes with pre-ozonation for treatment of secondary wastewater effluent, Water Res. 43 (7) (2009) 2020–2028. [12] Z. Yang, Y. Zhou, Z. Feng, X. Rui, T. Zhang, Z. Zhang, A review on reverse osmosis and nanofiltration membranes for water purification, Polymers 11 (8) (2019) 1252. [13] C. Li, W. Sun, Z. Lu, X. Ao, S. Li, Ceramic nanocomposite membranes and membrane fouling: a review, Water Res. 175 (2020) 115674. [14] F.E. Bortot Coelho, G. Magnacca, V. Boffa, V.M. Candelario, M. Luiten-Olieman, W. Zhang, From ultra to nanofiltration: a review on the fabrication of ZrO 2 membranes, Ceram. Int. 49 (6) (2023) 8683–8708. [15] F.E. Bortot Coelho, N.N. Kaiser, G. Magnacca, V.M. Candelario, Corrosion resistant ZrO 2 /SiC ultrafiltration membranes for wastewater treatment and operation in harsh environments, J. Eur. Ceram. Soc. 41 (15) (2021) 7792–7806. [16] C. Ojalvo, M. Jim´ enez-Fuentes, W. Zhang, F. Guiberteau, V.M. Candelario, A. L. Ortiz, Fabrication of B 4 C ultrafiltration membranes on SiC supports, J. Eur. Ceram. Soc. 42 (7) (2022) 3118–3126. [17] E. Eray, V.M. Candelario, V. Boffa, Ceramic processing of silicon carbide membranes with the aid of aluminum nitrate nonahydrate: preparation, characterization, and performance, Membranes 11 (9) (2021) 714. [18] J.S. Reed, Principles of Ceramics Processing, second ed., Jong Wiley &Sons Inc., New York (US), 1995. [19] R. Bayer, M. Knarr, Thermal precipitation or gelling behaviour of dissolved methylcellulose (MC) derivatives—behaviour in water and influence on the extrusion of ceramic pastes. Part 1: fundamentals of MC-derivatives, J. Eur. Ceram. Soc. 32 (5) (2012) 1007–1018. [20] C. Ojalvo, M. Ayll´ on, A.L. Ortiz, R. Moreno, Aqueous tape casting of super-hard B 4 C laminates with rGO-enriched reinforcing interlayers, J. Eur. Ceram. Soc. 41 (11) (2021) 5457–5465. [21] R.M. German, Sintering Theory and Practice, Wiley, New York, 1996. [22] M.A. García-Gal´ an, V.M. Candelario, F. Guiberteau, A.L. Ortiz, Evaluating the mechanical integrity and reliability of multi-channelled flat-sheet ceramic membranes for filtration applications, Ceram. Int. 50 (5) (2024) 7441–7450. [23] D. Liang, J. Huang, H. Zhang, H. Fu, Y. Zhang, H. Chen, Influencing factors on the performance of tubular ceramic membrane supports prepared by extrusion, Ceram. Int. 47 (8) (2021) 10464–10477. [24] T. Dons, V. Candelario, U. Andersen, L. Ahrn´ e, Separation of milk fat using silicon carbide support ceramic membranas with different pore sizes, Innovat. Food Sci. Emerg. Technol. 94 (2024) 103671. [25] T. Dons, J.J.K. Kirkensgaard, V. Candelario, U. Andersen, L. Ahrn´ e, Structural and physical-chemical properties of milk fat globules fractionated by a series of silicon carbide membranes, Food Res. Int. 192 (2024) 114680. J.R. Madrigal et al. Ceramics International 50 (2024) 50436–50443 50443