Increased Microporosity in Ceramic and Metal Foams: A Novel Processing Combination

Copper, aluminum, and alumina foams are prepared by a combination of reticulated foaming and freezing processing. This results in all of these foams in an additional strut porosity consisting of lamellar pores and material lamellae; the lamellar pores enable designed access to the inner volume of the struts. To increase the functionality in terms of enlarging the total surface area and increasing the adsorption activity, the foams are loaded with microporous materials by direct crystallization: the copper foams are combined with the metal organic framework HKUST‐1, the aluminum foams are combined with the zeolite SAPO‐34, and the alumina foams are combined with the zeolite silicalite‐1. All microporous materials are shown to access the inner strut pore surface area. Because of different crystal sizes of the different microporous materials, the load of the inner strut surface area varies significantly.


Introduction
A major part of our present challenges within social transformation is energy revolution.While there are numerous suggestions and solutions to adapt electricity generation and to transform it from fossil to renewable energy sources, reduction and saving of heat remain challenging.This might be attributed to energy densities of different storage systems or to temperature levels needed for efficient conversion.Heat is stored in different forms: latent, which makes use of high melting or evaporation enthalpies; sensible, which means a steady temperature change in the storage system accompanied by temperature-dependent storage efficiencies; or chemically.The latter may be divided in systems which generate heat of combustion or which make use of sorption processes.
Within the last-mentioned heat storage principle, zeolites and metal-organic framework compounds (MOFs) are discussed.
Both compounds are fine powders after synthesis and have to be supported on a rigid, preferably macroporous support for fast diffusion processes and heat transportation from/to the microporous entities.
Seijger et al. were one of the first researchers who demonstrated that zeolites can be fixed on the surface of a ceramic foam. [1]Even if this functional composite system, the ceramic foam is the support, the zeolite provides catalytic properties, was demonstrated for catalytic purposes in the gas phase, it is adaptable for sorption-driven heat transformation processes and systems.Within these, the microporous material adsorbs small molecules from the gas phase, such as water, and releases heat in the order of the heat of adsorption combined with the heat of hydration of the adsorptive.For both, zeolites and MOFs, and for ceramic foams, a great variety of preparation routes are available.An overview of zeolite and MOF crystallization processes on cellular structures is given in ref. [2].
For the manufacturing of open-cellular rigid, that is, ceramic and metal foams, the so-called Schwartzwalder process, or reticulated foaming process, is the preferred tool due to the generation of open cells; those open structures give access to the inner surface area of the foam cells.In addition, the foams made by the Schwartzwalder process are characterized by cavities inside the ceramic struts, which are suitable for further functionalization. [3,4]The process is described in a patent from 1961 [5] and was often adapted to specific material systems and demands such as a great number of ceramic materials, [6,7] metals, [8,9] improvement of mechanical properties, [3,10,11] and increase of the specific surface area. [12]s a consequence of the manufacturing process of reticulated ceramic and metal foams, they possess a large interior surface area which is not controlled accessible: because of the use of a polyurethane template foam (PU) covered with a ceramic or metal particle containing slurry, drying, PU burn-out, and sintering, hollow cavities remain where the PU was formerly located.Access is possible only by undesired cracks in the material or by the foam piece surfaces where the struts show undesired openings; both in an uncontrolled manner.With respect to a combination of rigid reticulated foams as support materials with microporous sorption-active materials, this additional but unused space may also be covered with zeolites or MOFs in order to increase the load of the rigid foams with active components.This will actually result in a higher-energy storage density or heat transformation performance of these types of composite materials, when used in those applications.
Direct foamed and reticulation foamed materials with different compositions have been shown to be suitable for direct crystallization of zeolites and/or MOFs.In ref. [13], a polymerderived ceramic foam consisting, among the ceramic residue of the polymer-to-ceramic transformation, also of silicon and silicon carbide fillers, was used for direct crystallization of the zeolite silicalite-1; part of the ceramic foam support was dissolved during the crystallization process and incorporated into the zeolite framework.The total zeolite load (weight ratio of the zeolite to the weight ratio of the zeolite/foam composite) was %50%.It was shown that the total energy density of a zeolite LSX, which is more relevant as a heat storage sorption material, fixed on a polymer-derived ceramic foam, amounts to 99.1 kWh m À3 ; this is already half the amount of the energy density of demonstrators with zeolite/pellet composites in fixed-bed reactors.
MOFs have been demonstrated to be suitable for direct crystallization on metallic foams, for example, HKUST-1 on copper foams [14] or nickel foams. [15]Ceramic substrates were also demonstrated to be suitable supports for the crystallization coating with MOFs, for example, with the aluminum isopthalate CAU-10 grown on alumina ceramic foams. [16]MOFs are potential candidates as active materials for adsorption-driven cooling devices, [17] and a theoretical cooling performance of up to 57 kWh m À3 has been estimated for the CAU-10@alumina composites.This represents an increase by a factor of 4-11 compared to conventional lamellar heat exchangers coated with CAU-10.Silicalie-I is used in our investigations as a reference zeolite material having no active (sorption) sites.
In recent works, it was demonstrated that alumina foams, [12] aluminum foams, [18] and copper foams [19] can be manufactured by the reticulation process in combination with a freezing step followed by freeze drying.This novel combination of processing steps results in the formation of additional pores in the material lamellae of the rigid-foam struts.These lamellar strut pores connect the foam cells with the hollow strut cavities formed by the PU template.The material lamellae have a typical thickness of 20-60 μm and the lamellar strut pores generated by the freezing technique are in the same dimensional range.
The aim of this work is to demonstrate that part of the lamellar strut pores and the inner strut space of reticulated and subsequently freezing-processed ceramic and metal foams can be covered by zeolites and MOFs.By this, the amount of active microporous material crystallized onto the cellular support will be increased due to the larger specific surface area of foams with additional lamellar strut pores.This concept is tested with different combinations of metal/ceramic support and MOF/zeolite adsorbent.The obtained micro-/macroporous composites are of potential interest for adsorption-driven heat managing devices.

Experimental Section
Alumina and metal foams (Al, Cu) were manufactured by the combined Schwartzwalder sponge replication/freezing technique adapted from the literature. [12,18,19]The metal foams were prepared using a freezing temperature of À20 °C (household freezer).For the alumina foams, a freezing temperature of À20 °C and of À80 °C was applied and the results were compared.For all foams, no oriented freezing technique was applied, the specimens were always subjected to the respective freezing temperature by placing the coated PU templates into the freezer.In addition, for the alumina foams, the solid loading of the ceramic dispersion was varied between 30 vol% and 40 vol% in order to vary the volume of the lamellar pores within the struts; for details please see the respective studies.These supports were used without further pretreatment for the crystallization coating with HKUST-1 (@Cu foams), SAPO-34 (@Al foams), and silicalite-1(@Al 2 O 3 foams).The crystallization coating was performed under hydro-/solvothermal conditions in sealed PTFE/ PFA reaction vessels.Details on the crystallization coating procedures are given below.The main focus of characterization was set on the amount of microporous material deposited on and in the foam struts, its water uptake capacity as a function of the crystallization parameters (except for silicalite-I@Al 2 O 3 foam), and the characteristics of the cellular supports.In order to investigate the effect of the additional strut porosity generated by the freezing processing on the crystallization of zeolites and MOFs, conventional replica foams made by the Schwartzwalder technique, but without a subsequent freezing step, were used as reference supports.

HKUST-1 Crystallization on Copper Foams
The coating with the MOF structure HKUST-1 was carried out by direct crystallization, placing the foam samples into a synthesis gel containing copper nitrate and trimesic acid as described in a previous work. [20]The crystallization process was carried out in a PTFE-lined stainless steel autoclave and/or in a polyvinylfluoride (PVF) container with a total volume of 60 mL.The mass ratio between synthesis gel and copper foam was %20:1.The synthesis gel was obtained by mixing 1.957 g (8.1 mmol) Cu(NO 3 ) 2 ⋅3 H 2 O dissolved in 15 mL distilled water with a solution of 0.968 g (4.6 mmol) trimesic acid and 15 mL ethanol.The reaction temperature was 80 °C, and it was kept for 6, 14, 24, and 96 h in a circulating air oven.After cooling to room temperature, the coated samples were washed with demineralized water.Traces of trimesic acid in the HKUST-1 framework were removed by an activation procedure in ethanol in the autoclave/container at 70 °C for 24 h.After that, samples were dried at 110 °C for 24 h.

SAPO-34 Crystallization on Aluminum Foams
For the in situ crystallization of the silico-alumino-phosphate SAPO-34 a recipe from the literature as described in ref. [21]  was adapted.As starting materials, 2.41 g Catapal B (Sasol) as an alumina source, 3.93 g orthophosphoric acid (85 %), 2.97 g morpholine (Sigma-Aldrich, ≥99%) as an organic template, and 1.02 g fumed silica (Aerosil 200, Evonik Industries) were used.Phosphoric acid and alumina were mixed with 8.11 g distilled water and stirred for 6 h to form a homogeneous gel.A second mixture was prepared by adding fumed silica and morpholine to 8.11 g distilled water while stirring.This mixture was then added dropwise to the first mixture while stirring.The resulting gel was aged for 24 h at 27 °C.The molar composition of the initial reaction mixture was given as Al 2 O 3 :SiO 2 :P 2 O 5 :2 morpholine:60 H 2 O.The mass ratio of the synthesis gel to the aluminum foam was %45:1.
For the direct crystallization process, the aluminum foams were placed in a PTFE-lined stainless steel autoclave using a Teflon distance holder on the bottom of the autoclave.The reaction mixture was filled into the autoclave to cover all of the foam`s surface.The autoclaves were placed in an oven and the reaction was performed under autogenic pressure at 200 °C for 5, 15, and 24 h.Afterward, the coated foams as well as the solid part that accumulated at the bottom of the autoclave were washed with water and dried at 110 °C for 24 h.

Silicalite-1 Crystallization on Alumina Foams
For the preparation of the zeolite@foam composites with zeolite crystals grown directly in the hollow cavities of the struts as well as between the lamellae in the struts (and also of the outer strut surface), a reaction mixture was prepared following the description in ref. [22].The synthesis gel contained 6.5 g tetraethyl orthosilicate (TEOS) as silicon source, 10.76 g water, and 8.12 g tetrapropyl ammonium hydroxide (TPAOH, 25 wt% solution in H 2 O) as structure-directing agent in the molar ratio of TEOS:TPAOH:H 2 O = 1:0.32:30; the mass ratio of the synthesis gel to the alumina foams was %12:1.
Two foam pieces at a time were placed into a stainless steel autoclave with PTFE inlay and covered with the reaction mixture.The autoclave was heated up to 150 °C and kept at this temperature for 72 h.After cooling down and removing of the supernatant solution, the foam samples were thoroughly rinsed with water, two times ultrasonicated in water to remove loose material, and finally dried for 24 h at 110 °C.The excess silicalite-1 powder settled at the bottom of the autoclave was collected as well.
A processing scheme covering the different materials combinations and crystallization parameters is given in Figure 1.

Characterization of the Coated Foams
The microstructure of particular struts of the coated foams was characterized by scanning electron microscopy using backscattered electron (BSE) and secondary-electron (SE) detectors on a XL30 microscope (FEI/Philips, Hillsboro/OR, USA).For cross-section microsections, the foam samples were embedded in an epoxy resin, ground, polished, and sputter coated with gold prior to the measurement.
The qualitative phase analysis of the obtained excess powders of HKUST-1 and SAPO-34 was performed by powder X-ray diffraction analysis with a D8 Discover diffractometer (Bruker-AXS GmbH, Karlsruhe, Germany) operated in Bragg-Brentano reflection geometry with Co Kα 1/2 radiation.The obtained powder patterns were compared against HKUST-1 and SAPO-34 reference diffractograms calculated using structural data obtained from the literature. [23,24]he cellular HKUST-1@Cu and SAPO-34@Al composite materials were characterized gravimetrically with respect to the amount of deposited MOF/zeolite and the water adsorption behavior.First, the samples were dried and subsequently weighed.Afterward, the composite materials were stored over saturated brine (75% r. h.) for 96 h and weighed again.From both weightings, the amount of deposited HKUST-1 or SAPO-34 in the dry (0% r. h.) and the water-loaded state (75% r. h.) was calculated after subtraction of the mass of the respective initial foam.From the mass difference between the dried and the water-saturated composite material, the water uptake capacity was determined.

HKUST-1@Cu Foams
The MOF HKUST-1 (copper trimesate, Cu 3 [C 6 H 3 (COO) 3 ] 2 ) has been grown successfully in the form of a blue coating on the struts of the Cu foams (Figure 2c).The HKUST-1 load ranged between 1 and 34 wt% and increased with increasing crystallization duration (Figure 2a).By crystallization coating with HKUST-1, the total porosity of the cellular composite is reduced compared to the initial support.The highest porosity reduction of 35 vol% on an absolute scale is observed for the sample containing the highest MOF loading of 34 wt%.Interestingly, the MOF load was higher for the Cu supports without additional lamellar pores within the strut material as generated by freezing processing.Most likely, this can be explained by the fact that the HKUST-1 crystals grow to a comparatively large size of up to 100 μm (see Figure 3) and the additional surface area provided inside the lamellar pores of the RP/FP supports can only be covered partially, or not at all (Figure 4).These pores have an average diameter of 35 AE 21 μm, see ref. [19] for details.Due to the presence of lamellar pores by freezing processing, the outer strut surface area, which is easily accessible for HKUST-1 crystallization, is reduced significantly.This lowers the total HKUST-1 coverage for these RP/FP foams.Nevertheless, it has been proven that HKUST-1 is capable of growing partially inside the lamellar pores (Figure 4).
The water uptake capacity of the HKUST-1@Cu composites naturally correlates to the total amount of MOF deposited on the Cu supports; it ranges between 0.005 g H2O /g comp.and 0.02 g H2O /g comp.for the samples crystallization coated for 6 h, and it increases to 0.14 g H2O /g comp.for the samples coated for 96 h (Figure 2b).The dashed line in Figure 2b represents the expected composite water uptake based on the HKUST-1 load and the theoretical water uptake capacity of pure HKUST-1 (0.55 g H2O /g MOF ; ref. [25]).It becomes apparent that for some samples the observed water uptake is higher than the expected water uptake.This is most pronounced for composites with lower HKUST-1 load obtained at shorter crystallization durations.A first hypothesis is that the Cu foams contain small amounts of copper oxides on the strut surface, which dissolve in the synthesis gel.This reduces the weight fraction of the support material in the final composite, on the one hand side, and provides an additional copper source for HKUST-1 crystallization, on the other.Both effects lead to an underestimation of the total MOF load and thus to a higher observed water uptake capacity.For longer crystallization duration, the total amount of HKUST-1 is larger, and this effect is less pronounced.
For composites containing larger quantities of HKUST-1 obtained after crystallization processing of 96 h, the water uptake is lower than the expected value.For these samples, the formation of nonporous copper trimesate byproducts is expected, which is a common problem in HKUST-1 syntheses. [22,26]owder X-ray diffraction (XRD) measurements of the HKUST-1 excess powders indicate the formation of phase-pure HKUST-1 (Figure 5).As a first approximation, a similar composition of the material deposited on the foam struts is expected, which is supported by the water uptake capacity of most of the respective composites.

SAPO-34@Al Foams
The zeolite SAPO-34 adapting the chabazite framework has been grown successfully on the struts of aluminum foams made by the sponge replication and the sponge replication combined with the freezing technique, respectively.In analogy to the HKUST-1@Cu composites, the amount of microporous adsorbent deposited on the foam struts increased with prolonged crystallization time (Figure 5).Thus, the SAPO-34 weight fraction ranges between 10 wt% (RP foams, 5 h) and 43 wt% (RP/FP foams, 24 h).Due to the higher density of SAPO-34, the reduction of the total porosity in the SAPO-34@Al composites is only 13 vol% (absolute) for the sample with the highest zeolite load of 43 wt%.In contrast to the HKUST-1 crystallization experiments, a slightly positive effect of the lamellar pores within the struts of the Al support on the amount of crystallized SAPO-34 has been observed (Figure 6a).The water uptake capacity of the SAPO-34@Al composites ranges between 0.08 g H2O /g comp.and 0.15 g H2O /g comp.; as expected, the composite water uptake is governed by the SAPO-34 load, essentially (Figure 6b).However, for all composite samples, the observed water uptake is significantly increased compared to the expected value based on a water uptake of 0.32 g H2O /g SAPO-34 for pure SAPO-34. [27]his effect is most pronounced for the samples obtained from shorter crystallization experiments, and it is also more pronounced for the RP/FP supports.
The reason for this increased composite water uptake capacity is the much higher reactivity of the Al supports toward the SAPO-34 synthesis gel compared to the HKUST-1 crystallization onto Cu.Aluminum is more reactive than copper, and the synthesis gel of SAPO-34, containing orthophosphoric acid, is more aggressive (with respect to the support dissolution) than the starting materials for the HKUST-1 synthesis.Consequently, the Al strut material is partially dissolved and transformed into SAPO-34, which is observed in scanning electron microscopy (SEM) images obtained on the laid open-strut surface area of SAPO-34-coated Al foams (Figure 7).Therefore, the weight fraction of the support material in the final composite material is lowered, which results in an underestimation of the SAPO-34 content and thus in an increased water uptake capacity.This dissolution effect is more pronounced during the initial state of the SAPO-34 crystallization process as the ongoing growth of a zeolite layer on the Al surface leads to a passivation effect.Thus, the discrepancy between the observed   and expected water uptake capacity is most pronounced for the samples obtained after 5 h of crystallization.This process is usually addressed as partial support transformation and is used for the SAPO-34 coating of Al heat exchangers of adsorption-driven cooling devices as well. [28]s the specific surface area of the RP/FP aluminum foams is considerably higher than for the conventional RP foams, [18] the reactivity toward the SAPO-34 synthesis gel is increased for the RP/FP specimens.As a consequence, this results in an increase of the SAPO-34 load, on the one hand side, and an increased rate of support transformation, which is expressed in the larger discrepancy between observed and expected water uptake capacity (Figure 5b).
From SEM investigations of strut cross section microsections, it became apparent that the SAPO-34 crystals are significantly smaller with %20 μm compared to HKUST-1 crystals (Figure 8).Consequently, the lamellar pore space within the RP/FP aluminum supports is much more accessible for SAPO-34 crystallization processing resulting in the observed increase in zeolite load for these supports.
Powder XRD measurements of the SAPO-34 excess powders obtained from the bottom of the autoclaves indicate a phase-pure SAPO-34 phase (Figure 9).As a first approximation, a similar material composition is assumed for the zeolite material deposited on the struts of the Al supports, which is corroborated by the water uptake results.

Silicalite-1@Al 2 O 3 Foams
Silicalite-1 (MFI framework) was successfully grown on the surface of all alumina supports in form of the typical coffin-shaped and twinned crystals, as shown in Figure 10.In contrast to SAPO-34, and especially to HKUST-1, the silicalite-1 crystals were significantly smaller having a typical size of 2 μm.The MFI load on the alumina foams is, however, much lower compared to the other investigated composite materials and ranges between %0.1 and 4.9 wt% (Figure 11).This is mostly attributed to the low thickness of the coating layers of only a few micrometers as a consequence of the small crystallite size of silicalite-1.Consequently, the effect of MFI coating on the total porosity is marginal.Therefore, the positive effect of additional strut porosity and strut surface area is much more pronounced for the MFI@Al 2 O 3 samples: For alumina supports made from dispersions containing 40 vol% Al 2 O 3 , the MFI load is 1.0 wt% for RP foams made without freezing processing, but 2.2 wt% for the corresponding RP/FP foams made with a freezing step at À80 °C.The observed increase in silicalite-1 load with a factor of 2.2 is in good agreement with the increase in the specific surface area by additional freezing processing (factor 2.8, see ref. [12]).By increasing the amount of dispersant (i.e., decreasing the solid content) in the alumina dispersions applied for the  manufacturing of the RP/FP foams, the freezing processing generated strut porosity can be increased, leading to a higher strut surface area being accessible for MFI deposition.Consequently, the silicalite-1 load is increased from 2.2 wt% for the RP/FP foams made from 40 vol% alumina dispersions, to 4.9 wt% for the supports made from 30 vol% alumina dispersions (both freeze processed at À80 °C).For the supports freeze processed at À20 °C, the amount of crystallized MFI is 1.8 wt%, on average, and does not depend significantly on the solid load of the dispersion applied in the foam manufacturing.This is connected to the formation of significantly larger lamellar pores compared to the RP/FP supports made at À80 °C, but at a similar level of strut porosity (please see refs.[12,29] for further detail).Consequently, the increase in strut surface area accessible for the MFI crystallization is less pronounced.

Conclusion
Aim of this work was to demonstrate that the inner strut surface of ceramic and metal foams, processed by a combination of reticulated foaming and freezing processing, is accessible for direct crystallization processes of microporous materials such as MOFs and zeolites.The foams used for this work possess open cells and hollow struts.In order to have regular access to the inner space of the hollow struts, freeze processing was introduced as an additional step in foam manufacturing to generate lamellar pores in the strut material; they showed a gallery height of 10-25 μm and reached from the open foam cells through the struts into the hollow cavities of the foam struts.
In order to demonstrate the accessibility of microporous materials to the inner strut cavities by direct crystallization, copper foams were combined with the MOF material HKUST-1 (HKUST-1@Cu), aluminum foams were combined with the zeolite SAPO-34 (SAPO-34@Al), and alumina foams were combined with the zeolite silicalite-1 (silicalite-1@Al 2 O 3 ).HKUST-1@Cu showed the lowest content of MOF in the inner struts, the amount of SAPO-34 in the inner struts of Al foams was moderate, and silicalite-1 was found in major amounts within the inner struts of the alumina foams, all in comparison to reticulate-processed-only (without freezing processing) foams used as reference.These differences were attributed to the very  different crystallite sizes of the microporous components: HKUST-1 crystals were in the range of up to 100 μm.This resulted in a very low amount of crystals in the inner space of the struts, but also on the outer surface of the copper foams because of the reduced outer surface area due to material lamellae and lamellar pores formation.SAPO-34 showed a crystal size of around 20 μm, which is in the range of the lamellar pore gallery heights, and significantly more SAPO-34 was able to crystallize in the inner strut space.The highest increase in surface area was found with the silicalite-1@Al2O3, which was explained by a crystal size of %2 μm.
These findings indicate that the final crystallite size of the microporous materials control the accessibility of the inner strut cavities during direct crystallization processing.The effects behind this behavior are still in the dark.However, even if the use of the inner strut surface area for hosting microporous components was shown for the very first time, we believe that there is a huge potential for the development of high-capacity heat conversion materials based on sorption processes.

Figure 1 .
Figure 1.Processing scheme with the crystallization parameters of the direct crystallization of HKUST@Cu foam, SAPO-34@Al foam, and silicalite-I@Al 2 O 3 foam.

Figure 2 .
Figure 2. a) HKUST-1 load on cellular Cu as a function of the crystallization duration; RP indicates Cu foams made by conventional sponge replication processing.RP/FP stands for the combination of sponge replication and freezing processing; b) Water uptake capacity of the HKUST-1@Cu composites as a function of the MOF load; c) HKUST-1-coated Cu RP/FP foam.

Figure 3 .
Figure 3. HKUST-1 grown on the strut surface of Cu RP foam after a crystallization time of 96 h.

Figure 4 .
Figure 4. HKUST-1 grown on and inside Cu RP/FP foam after a crystallization time of 96 h (cross section microsection).

Figure 5 .
Figure 5. XRD powder patterns of the HKUST-1 excess powders obtained after different crystallization times.

Figure 6 .
Figure 6.a) SAPO-34 load on cellular Al as a function of the crystallization duration; RP indicates Al foams made by conventional sponge replication processing, RP/FP stands for the combination of sponge replication and freezing processing; b) Water uptake capacity of the SAPO-34@Al composites as a function of the zeolite load.

Figure 7 .
Figure 7. SAPO-34 coating on RP/FP foam after a crystallization time of 5 h.The dissolution of Al is clearly visible by the pits in the strut material.

Figure 8 .
Figure 8. Cross section microsections of the RP (top) and RP/FP (bottom) aluminum foams after crystallization of SAPO-34 for a,d) 5 h, b,e) 15 h, and c,f ) 24 h; all samples from PU template foams with a cell size of 20 ppi.

Figure 9 .
Figure 9. Powder XRD patterns of the SAPO-34 excess powders obtained after different crystallization times.

Figure 11 .
Figure 11.Silicalite-1 load on cellular alumina processed at different freezing temperatures and with different solid loadings of the alumina dispersions for foam manufacturing.