Open-celled zinc oxide ceramic foams were prepared by the polymer sponge replication (Schwartzwalder) technique from aqueous ZnO dispersions with Sb2O3 and Bi2O3 as sintering additives, and mechanically stable ZnO foams with an average porosity of 93.6% were obtained. Their microstructure consists of ZnO grains with a Bi-containing grain boundary phase together with a Zn-Sb-O secondary phase with spinel structure. The obtained ZnO ceramic foams were characterized with respect to their morphology by computed tomography; in addition, the compressive strength and the thermal conductivity were determined, and the data were applied for modelling of the mechanical and thermal properties of the bulk ZnO strut material.
Open-celled ceramic foams are used within several technological fields; with respect to the quantity, the most prominent applications are filter materials for metal melts in casting or catalyst supports [
Zinc oxide is a semiconductor material, which crystallizes in the wurtzite structure type in analogy to aluminum nitride or silicon carbide, for example. Therefore, it belongs to the class of adamantine compounds with a basic all-tetrahedral coordination of cations and anions [
A typical application for ZnO-based ceramics is varistors, which have a distinct nonohmic electrical conductivity, i.e., a very low conductivity below a characteristic breakthrough voltage and can be used for overvoltage protection [
In the present work, the established polymer sponge replication—Schwartzwalder—process has been adopted for the manufacturing of zinc oxide ceramic foams. As sintering aids, Sb2O3 and Bi2O3 were used in a fixed molar ratio of 2 : 1. The obtained foams were characterized with respect to their microstructure (SEM and phase composition by XRD) and their macroscopic properties mechanical strength and thermal conductivity as a function of the total porosity and morphology. Finally, the bulk properties of the ZnO strut material were extrapolated from the obtained data by applying established structure (porosity)-property relations.
The ceramic raw powder was prepared by ball-milling a dispersion of 200 g zinc oxide (
The ceramic dispersion for foam manufacturing was prepared by adding 100 g of the ZnO-Sb2O3-Bi2O3 oxide mixture and 1.0 g ethanolammonium citrate deflocculant (Dolapix CE64, Zschimmer & Schwarz Chemie GmbH, Lahnstein, Germany) to 26.5 mL distilled water. The mixture was homogenized for 15 min using a planetary centrifugal mixer (THINKY Mixer ARE-250, THINKY Corp., Tokyo, Japan) operated at 2000 rpm. Afterward, 1.5 g polyvinylalcohol binder (Optapix PA 4G, Zschimmer & Schwarz Chemie GmbH) and 0.1 g polyalkylene glycolether defoamer (Contraspum K1012, Zschimmer & Schwarz Chemie GmbH) were added to the dispersion followed by a second mixing step for 15 min at 2000 rpm. The resulting dispersion has a solid content of 77.5 wt.%, which is 37.9 vol.%.
For the ZnO foam, manufacturing reticulated polyester polyurethane (PU) foams (SP30P20R, Koepp Schaum GmbH, Oestrich-Winkel, Germany) with 20 pores per linear inch (ppi) and a cubic geometry with 20 mm × 20 mm × 20 mm were used as template structure. The PU foams were completely immersed into the ZnO dispersion and subsequently freed from the excess amount by manual squeezing of the foam template until its weight reached approximately 2.4 g (corresponding to ≈ 93.5% porosity in the final foam piece after sintering). After drying under ambient conditions, the PU template was removed thermally in three steps (110°C/2 h, 250°C/3 h, 400°C/3 h, heating/cooling rate 1 K·min−1) in a circulating air furnace (KU 40/04/A, THERMCONCEPT Dr. Fischer GmbH, Bremen, Germany). Afterward, the samples were densified at 1100°C for 3 h in air (heating rate of 3 K·min−1) using a 30 L sintering furnace (LH 30/14, Nabertherm GmbH, Lilienthal, Germany). For thermal conductivity measurements, rectangular PU templates with a dimension of 50 mm × 50 mm × 20 mm were coated with the ZnO dispersion; the excess of slurry was extruded with a roller press to reach a total weight of 15 g for the coated PU foam. Template removal and sintering were performed as described above.
The total porosity of the foams (
The quantitative phase composition of the ZnO strut material was determined by powder X-ray diffraction (PANalytical X’Pert Pro Bragg-Brentano diffractometer, Co-K
The thermal conductivity of the ZnO foams was determined using the transient plane source (TPS) technique and a TPS 2500 S device (Hotdisk SE, Gothenburg, Sweden) by placing the sensor in between two 42 mm × 42 mm × 17 mm foam samples with previously sanded surfaces [
The compressive strength was determined using a TIRAtest 2825 universal testing machine and circular loading plates with 150 mm in diameter and a crosshead speed of 1 mm·min−1 (TIRA GmbH, Schalkau, Germany). To ensure a more homogeneous load on the samples, a cardboard piece with 1 mm thickness was placed between the foam and the loading plates. From the obtained data, the maximum force was extracted and used for the calculation of the compressive strength. The results of 35 specimens were evaluated using a two-parameter Weibull distribution and the Visual-XSel 14 program package [
The microstructure of selected specimens was characterized by scanning electron microscopy using a XL30 ESEM-FEG microscope (FEI/Philips, Hillsboro/OR, USA) equipped with a secondary electron (SE) and backscattered electron (BSE) detector. The grain size distribution in the ZnO strut material was determined from the BSE micrographs by manually measuring the dimension of 150 individual grains. The elemental composition of the ZnO strut material was analyzed by energy dispersive X-ray spectroscopy (EDAX-AMETEK GmbH, Weiterstadt, Germany). Beforehand, the grinded sample material was uniaxially pressed at 30 MPa for 2 min into slabs, which were used for the EDS characterization. Nine EDS spectra were recorded at different positions on the sample and were used for the subsequent elemental analysis.
Micro-computed tomography (
The dispersion of the ZnO-Sb2O3-Bi2O3 powder mixture in water using an ethanolammonium citrate-based deflocculant was successful with respect to the rheological behavior up to a solid content of 77.5 wt.%/37.9 vol.%. The obtained dispersion possessed the desired shear-thinning flow behavior and a viscosity suitable for the successful manufacturing of cellular ZnO ceramics by the polymer sponge replication technique [
After sintering, mechanically stable ZnO ceramic foams were obtained which showed an intense yellow color (Figure
Open-cellular zinc oxide ceramic foams with a pore count of 20 ppi and 93.6% total porosity after sintering at 1100°C. The yellow color is a consequence of structural defects in the ZnO lattice (oxide vacancies and interstitial zinc atoms).
In the SEM micrographs of the strut material, a dense microstructure of well-sintered ZnO grains was observed (Figures
SEM micrographs of the strut fracture surface in zinc oxide ceramic foams: (a) SE image, (b, c) BSE images revealing the Bi-containing secondary phase in the grain boundaries, (d) grain size distribution in the ZnO strut material, and (e, f) BSE micrograph and EDS spectra of the strut surface (black ○) and of individual grains of a Sb-containing secondary phase (red ○).
Moderate ZnO grain growth from 0.2 ± 0.07
Powder X-ray diffraction reveals 89.0 wt.% of hexagonal wurtzite-ZnO as the main phase being present in the strut material (Figure
Powder XRD pattern of the ZnO strut material and the corresponding Rietveld fit revealing the individual phase contents of hexagonal ZnO, cubic
Phase and elemental composition of open-celled zinc oxide ceramic foams determined by powder XRD with Rietveld analysis and a comparative EDS measurement.
Phase composition (Rietveld) | |||||
Phase (wt.%) | ZnO |
|
|
||
ZnO 20 ppi | 89.0 ± 0.14 | 10.3 ± 0.04 | 0.7 ± 0.01 | ||
|
|||||
Elemental composition (calculated from Rietveld results) | |||||
Element (wt.% ‖ at.%) | Zn | O | Sb | Bi | Al |
ZnO 20 ppi | 76.6 ‖ 47.6 | 20.0 ‖ 50.8 | 2.4 ‖ 0.8 | 0.6 ‖ 0.1 | 0.4 ‖ 0.7 |
Starting powder# | 75.4 ‖ 48.2 | 19.3 ‖ 50.3 | 2.8 ‖ 1.0 | 2.5 ‖ 0.5 | -/- |
|
|||||
Elemental composition (EDS measurement) | |||||
ZnO 20 ppi | 78.1 ‖ 52.2 | 16.5 ‖ 45.0 | 2.9 ‖ 1.1 | 1.7 ‖ 0.4 | 0.8 ‖ 1.3 |
#Calculated for the initial powder mixture of 93.9 wt.% ZnO, 3.4 wt.% Sb2O3, and 2.7 wt.% Bi2O3.
In this context, the polymorphism of Zn7Sb2O12 is known, but not fully understood yet. In recent studies, the
However, the temperature at which this phase transformation occurs can be drastically affected by doping of
In the case of the ZnO ceramic foams, a possible impurity is Al3+ originating from alumina abrasions during the ball-milling procedure of the initial ZnO-Sb2O3-Bi2O3 powder mixture. As the Al3+ and Cr3+ ions are similar in size and preferred coordination environment, a similar stabilization mechanism as described for the doping with Cr3+ may be expected. An indication for this hypothesis can be found in the lattice constant of
Approximation of the Al3+ content in the cubic Zn7−2/3
The Bi2O3 amount of 0.7 wt.%, corresponding to 0.1 at.% Bi (Table
EDS analysis averaged from 9 separate EDS spectra recorded on an uniaxially pressed slab of the ZnO strut material.
The total porosity of the obtained ZnO ceramic foams is high with a value of 93.6 ± 0.4% (Table
Properties of open-celled zinc oxide ceramic foams. The total porosity and cell porosity values refer to the geometric foam volume and the ratio
ZnO, 20 ppi | ||
---|---|---|
Foam porosity ( |
Total | 93.6 ± 0.4% |
Cella | 93.5 ± 0.9% | |
Strut porosity ( |
Totalb | 32 ± 5% |
Hollow strutb | 32 ± 3% | |
Strut materialc | <0.5% | |
Shrinkage | Linear | 15.8 ± 1.1% |
Volumetric | 41 ± 4% | |
Morphology (CT) | Strut thickness | 0.38 ± 0.15 mm |
Cell size | 2.69 ± 0.23 mm | |
Thermal conductivity | Foam | 0.82 ± 0.07 W·m−1·K−1 |
Bulk materiald | 37 W·m−1·K−1 | |
Compressive strength | Average strength | 0.15 ± 0.03 MPa |
Weibull modulus | 5.6 |
aIncluding the cavities resulting from the PU template burnout. bRelated to the overall strut volume;
An isotropic linear shrinkage of 16 ± 1%, equivalent to a volumetric shrinkage of 41 ± 4% in relation to the initial foam template dimensions, has been observed during the sintering process at 1100°C. This is in a good agreement to sintering studies of dense ceramics in the ZnO-Sb2O3-Bi2O3 system, for which a linear shrinkage of 17% is reported for the same powder composition as used within this work [
The open porous structure of the ZnO ceramic foams is confirmed by computed tomography; a total porosity of 94% and a closed porosity < 0.1% have been determined, which is in good agreement to the results of the Archimedes measurements. In addition, no significant pore window blocking is present resulting in a thoroughly open-cellular structure (Figure
(a) Strut thickness and cell size distributions calculated from a three-dimensional
The compressive strength data obtained on all 35 ZnO foam specimens could be satisfactorily modelled by a two-parameter Weibull distribution [
Compressive strength data of 20 ppi ZnO ceramic foams (□) and the corresponding fit with the two-parameter Weibull function (red line).
For the evaluation of the strength-porosity correlation, the compressive strength data obtained for the cellular ZnO specimens were modelled with the Gibson–Ashby (GA) relation for the description of the crushing behavior of brittle, cellular materials (equation (
According to equation (
Compressive strength data of 20 ppi ZnO foams (□) and the corresponding fit with the Gibson–Ashby model for the strength-porosity correlation (red line).
The thermal conductivity of the obtained ZnO ceramic foams determined by the transient plane source technique was 0.82 ± 0.07 W·m−1·K−1 at 93.6% total porosity. In order to evaluate the thermal properties of the ZnO strut material, the bulk thermal conductivity
The parameter
The manufacturing of cellular ZnO ceramic foams with the established polymer sponge replication technique and an aqueous ZnO dispersion and Sb2O3 and Bi2O3 as sintering additives has been demonstrated. The obtained open-celled ZnO foams show a high total porosity of 93.6% including the characteristic hollow strut cavities typical for this processing technique. The strut material itself is almost fully densified; its microstructure consists of ZnO grains with an amorphous Bi-containing grain boundary phase. In addition, an Al-doped spinel secondary phase with the composition Zn6.17Sb1.58Al1.25O12 was formed. The dopant Al3+ originates from a contamination during the ball-milling process of the raw powders and leads to a stabilization of the thermodynamically metastable spinel phase at room temperature.
The mechanical and thermal properties of the obtained ZnO ceramic foams were evaluated, and an average compressive strength of 0.15 MPa and a thermal conductivity of 0.82 W·m−1·K−1 were measured. The strength-porosity correlation was modelled with the Gibson–Ashby law for brittle, cellular structures; a bending strength of 54 MPa for the bulk ZnO strut material being in good accord to published data on comparable ZnO-based varistor ceramics was estimated. From the thermal conductivity data, a bulk thermal conductivity of 37 W·m−1·K−1 was determined for the ZnO strut material, which is in good agreement with the thermal conductivity of ZnO varistor ceramics of similar composition.
The data used to support the findings of this study are included within the article.
The authors declare that they have no conflicts of interest.
Funding for this research was supplied by the federal state of Saxony-Anhalt.