Electrical resistivity of ceramic±metal composite materials: application in crucibles for induction furnaces - PDF

Ceramics International 26 (2000) 325±335 Electrical resistivity of ceramic±metal composite materials: application in crucibles for induction furnaces J.R. Martinelli*, F.F. Sene Energy and Nuclear Research

Please download to get full document.

View again

of 11
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.

Travel & Places

Publish on:

Views: 18 | Pages: 11

Extension: PDF | Download: 0

Ceramics International 26 (2000) 325±335 Electrical resistivity of ceramic±metal composite materials: application in crucibles for induction furnaces J.R. Martinelli*, F.F. Sene Energy and Nuclear Research Institute, Brazilian Nuclear Energy Commission, C.P Pinheiros, S. Paulo, SP, Brazil Received 2 March 1999; received in revised form 9 March 1999; accepted 5 May 1999 Abstract A study was conducted of the electrical properties of partially-stabilized zirconia ceramic matrix composites with 0 to 40 vol% of titanium, niobium, nickel, and titanium and cristobalite, over temperatures ranging from 25 to 800 C. Composites containing metallic inclusions above 25 vol% show the predominance of electronic type conduction. For metallic inclusions bellow 25 vol%, a typical ionic conduction behavior has been observed. Due to the percolation e ect, partially-stabilized zirconia containing 25 vol% of titanium, or nickel, shows an insulator to conductor transition, and cristobalite containing 30 vol% of titanium shows a conductor to insulator transition as the temperature is raised. Crucibles of composites of partially stabilized zirconia containing titanium were made by slip casting and sintering. Tests performed in an induction furnace showed that both pellets and crucibles containing metallic inclusions above 25 vol% had self-heated. The maximum temperature reached by xing the power to 2 kw was 1350 C without any charge. X-ray di raction, energy dispersive spectroscopy, scanning electron and optical microscopy were performed to provide a full microstructure characterization. Continuous paths formed by metallic clusters can be seen for materials containing metallic concentration over 25 vol%. # 2000 Elsevier Science Ltd and Techna S.r.l. All rights reserved. Keywords: B. Composites; C. Electrical conductivity; D. ZrO 2 1. Introduction The addition of ceramic particles to metals has been done since the thirties in order to improve the mechanical properties of metal alloys. For instance, tungsten carbide was added to cobalt, and consequently this cermet could be used as cutting tools [1]. Recently, metal alloys and ceramic materials have been reinforced with ceramic particles and whiskers in order to improve their mechanical properties too. These materials have been named composites, and they are attractive to be used as structural materials because of their relatively low density, high toughness, and improved mechanical strength [2]. Electrical properties of composite materials are also relevant and should be investigated more extensively, which could lead to a wide range of applications. One of the main phenomenon to be investigated is the insulating to conducting transition that happens in a narrow * Corresponding author. Tel.: ; fax: address: (J.R. Martinelli). percolation range. This phenomenon depends on the concentration, shape, dimension and crystallographic direction of the embedded particles. The increase of the electrical conductivity of insulating materials containing metallic inclusions during the exposition to electromagnetic elds was rst reported by Brandy in New electrical conducting paths were observed when materials were exposed to radio frequency elds. Those paths were responsible for a drastic increase of the conductivity. Composite materials that combine insulating and conducting phases are similar to the ones studied by Brandy. Electrical properties have been previously measured at room temperature for ceramic±metal composites such as Al 2 O 3 ±Ti [3], ZrO 2 ±TiCN [4], ZrO 2 ±Ti and ZrO 2 ±Nb [5]. Generally, a reduction of electrical resistivity has been observed as the concentration of metallic inclusions is raised. Zirconia has been used in high temperature applications ( 1000 C) due to its high melting point and chemical inertness. Zirconia is used as heating elements, cathodes for plasma torches and electrodes for MHD /00/$20.00 # 2000 Elsevier Science Ltd and Techna S.r.l. All rights reserved. PII: S (99) 326 J.R. Martinelli, F.F. Sene / Ceramics International 26 (2000) 325±335 energy conversion. Furthermore, partially- stabilized zirconia (PSZ) is extensively used as oxygen sensors and as electrodes in H 2 production devices based on the thermohydrolysis of water [6]. Composite materials based on insulating matrix and conducting inclusions can be potentially used in a wide range of technological applications such as in the manufacture of crucibles for vacuum induction furnaces. Crucibles made of carbon free materials can be used in melting processes that require no carbon contamination. Thermal etching of ceramic materials can also be performed by using composite materials in this same heating process. When composite materials show electrical resistivity behavior with positive temperature coe cient, devices sensible to temperature variation can be built [7]. If the material shows non-linear current± voltage behavior in the percolation region, varistors can be produced [8]. Finally, electrodes for fuel cells made by zirconia±ni have been reported [9]. In the present work the mechanisms of electrical conduction in PSZ±Ti, PSZ±Nb, PSZ±Ni were investigated through the measurement of electrical resistivity as a function of temperature. It was also checked the selfheating behavior of these materials when exposed to electromagnetic elds in the range of radio frequency (r.f.). The system a-sio 2 ±Ti was also investigated because cristobalite shows a relatively low temperature phase transition with volume change which can provide some information on percolation related to phase transition. Crucibles made of ZrO 2 ±Ti by slip casting were checked in an induction furnace for self heating. 2. Experimental procedure Partial stabilized zirconia containing 3 mol% of ytria was prepared by coprecipitation of zirconium hydroxide from an aqueous solution of ZrOCl 2 and YCl 3. Ammonium hydroxide was used as the precipitation agent. The precipitated material was ltered, washed in distilled water, dehydrated in ethyl alcohol, washed with acetone, ltered again and nally dried at 100 C for 24 h. The material was then calcined in air at 850 C for 2 h, ground in a mortar with a pestle and nally ball milled with alumina media for 3 h. Impurity levels for all raw materials were determined by ame spectroscopy. A spectrograph Jarrel model S 800 was used for that purpose. The particle size distribution was determined by using a sedigraph Micromeritics Model The yttrium oxide molar fraction in the PSZ after the coprecipitation process was determined by X-ray uorescence. The PSZ tetragonal phase content and the amount of all other crystalline phases were determined by X-ray di raction. A Rigaku Model DMAX-100 di ractometer was used for this purpose. The amount of monoclinic and tetragonal zirconia phases were determined from the appropriated X-ray di raction peak intensities [10]. Mixtures of PSZ with Nb, Ti, or Ni powders in the range of 10 to 45 vol% were used to prepare 14 mm diameter pellets by uniaxial pressing at 100 MPa for 15 s. Pellets of cristobalite±ti were also prepared following the same procedure. These pellets were nally isostatically pressed at 200 MPa for 60 s at room temperature. Samples of PSZ±Nb and PSZ±Ti were sintered in a tubular electric furnace in argon at 1600 C for 1.5 h. The heating rate was 10 C/min. To sinter PSZ±Ni and SiO 2 ±Ni pellets the same procedure was adopted except for the sintering temperature that was 1450 C. Densities for all samples before and after sintering were determined by the Archimedes' method. Microstructure characterization was performed by using a light microscope Olimpus and a scanning electron microscope Philips Model XL30 with capacity for energy-dispersive X-ray spectroscopy. Samples were prepared by cutting and polishing with diamond pastes from 15 to 1 mm. A thermal etching was performed on PSZ samples containing Ti or Nb at 1500 C for 20 min in argon in order to reveal the grain boundaries. Samples were nally coated with gold by sputtering before performing the SEM analyses. A preliminary electrical evaluation of the samples was performed at room temperature in order to classify them as insulators or conductors. A multimeter Fluke model 8050A and an electrometer Keithley model 610C were used to measure the electrical resistance of these samples. Electrical resistivity measurements in the temperature range of 25±800 C under argon ow of 1 l/min were also performed on samples 2 to 5 mm thick by using the two probe d.c. method. Electrodes were built on the sample surfaces by previously coating them with platinum by sputtering. Fig. 1 shows a schematic diagram of the experimental setup. A thermocouple was set close to the sample to provide a precise temperature measurement. Platinum wires were used to connect the electrodes to the electrometer. A reversible switch was used to reverse the electrodes and consequently to eliminate any artifact caused by parasitic currents due to thermal gradients. Composite materials were exposed to r.f. and the temperature was measured as a function of power and voltage of the r.f. generator. An induction furnace HWG was used for that purpose. Temperatures were measured by a pyrometer Minolta model S40. The experimental setup is described elsewhere [11]. Crucibles of PSZ containing 10±30 vol% of Ti were made by slip casting. Powders of Ti and PSZ were mixed in water. A commercial product named Lambra was added as de occulator. A mold made of plaster of Paris was used to cast the slurry in the shape of crucibles. J.R. Martinelli, F.F. Sene / Ceramics International 26 (2000) 325± Fig. 1. Experimental arrangement for measurement of electrical resistivity as a function of temperature. The material was pulled out from the mold, dried at 100 C for 24 h, pre-sintered at 1000 C in argon for 1 h, polished and nally sintered in argon at 1600 C for 1.5 h. These crucibles were also exposed to r.f. and the temperature was monitored following the same procedure described above. 3. Results and discussion Table 1 shows the main impurities in the raw materials. Aluminum and silicon in PSZ are originated from the milling process. A relatively high concentration of Ta is found in Nb powders. It is quite di cult to remove Ta from Nb during the puri cation process. However, we believe that these impurities will not substantially a ect the percolation of electrical paths in our composite materials when compared to the e ects expected by metallic inclusions. In fact, impurities could change the electrical resistivity background. Others factors, such as metal particle oxidation and the contact geometry would far overshadow the e ects due to impurities in the metal phase. The molar fraction of yttria in PSZ was determined to be 2.8 mol%, which is enough to keep the zirconia partially stabilized. From the X-ray di raction the tetragonal phase content was determined to be 96.4%. Fig. 2 shows a representative example of X-ray diffraction patterns for PSZ±10 vol% Nb, and for SiO 2 ±10 vol% Ti after sintering. Others di raction patterns show similar characteristics related to the material's constituents. Di raction peaks related to the metallic inclusions and ceramic matrix are also shown. From the di raction patterns it was possible to calculate the amount of crystalline phases shown in Table 2. The medium equivalent spherical diameters for PSZ, cristobalite, Nb, Ti, and Ni are 2.8, 9, 35, 38, and 25 mm, respectively. It is noticed that the metallic particle average sizes are one order of magnitude higher than the ceramic ones. Therefore it can be assumed that mixtures of ceramic and metal particles described in the present work are suitable for the preparation of ceramic±metal composites, since the probability to have metallic inclusions inside ceramic grains will be low [12]. In order to compare the composite density to a reference value, the rule of mixture was used to calculate the ``theoretical'' density value (TD) considering the volumetric Table 1 Impurities detected by ame spectroscopy for: PSZ, niobium, titanium, nickel, and cristobalite (mg/g) Element PSZ Nb Ti Ni -SiO 2 Fe Cr Ni ± 45 Zn Si Al Mn Mg Pb Sn Bi Cu 5 ± Na 30 ± ± ± Ba ± ± ± Co ± 6 ± Ca ± Sb ± ± ± V Ta ± ± Ti 45 ± 328 J.R. Martinelli, F.F. Sene / Ceramics International 26 (2000) 325±335 fractions of metallic inclusions and the ceramic matrix. Density for green samples varied from 57 to 62% TD and for sintered samples, from 72 to 92% TD depending on the volume and type of metallic inclusions. The density after sintering decreases as the amount of metallic inclusions is raised. This result is in agreement with the theoretical model that explains the densi cation of ceramic materials in the presence of embedded metallic particles, considering the di erence between the ceramic and metal sintering rates, and the relatively low di usion of their constituents [12]. In the other hand density of green samples increases as the amount of metallic inclusions is raised. Metal particle inclusions lead to higher ductility and mechanical deformation during the compaction. 4. Microstructure analyses Fig. 3 shows the microstructure of PSZ±10Ti, PSZ± 20Ti, PSZ±30Ti and PSZ±45Ti samples taken from a light microscope (LM). The metallic inclusions are located in the white regions and the ceramic matrix is the gray region in those pictures. Pores are represented by black spots. From Fig. 3A and B (samples that are insulators at room temperature) it can be noticed that there is an increase of metallic inclusions without formation of continuous paths. In Fig. 3C (sample that shows electrical conduction at room temperature) few continuous paths can be seen formed by metallic clusters. In the sample showed by the Fig. 3D (electrical conductor sample) these paths are saturated. Samples containing niobium and nickel show similar microstructure characteristics. Fig. 4 shows a microstructure from PSZ±Ti and PSZ± Nb obtained by SEM. By comparing the Nb and Ti dot mapping built by EDS and the microstructures of corresponding samples, it is quite possible to assume that most of the metallic inclusions are located at the grain boundaries. Few metallic grain clusters have been observed. The amount of clusters increases as the metallic particle inclusion is raised. When these clusters touch themselves electronic conduction paths are formed. The ceramic grain medium size is 30 5 mm and it does not depend on the amount of Nb and Ti. The porosity of the material increases as the metal particle amount is raised leading to lower density values as it was earlier reported. 5. Electrical properties Fig. 2. X-ray di raction patterns for: (a) PSZ±10 vol% Nb, and (b) SiO 2 ±10 vol% Ti. Table 2 Amount of detected phases in the sintered samples: (a) PSZ±Ti, (b) PSZ±Nb, (c) PSZ±Ni and (d) SiO 2 ±Ti Phase Amount (%vol) Phase Amount (%vol) a. PSZ±Ti b. PSZ±Nb ZrO 2 (T) 92.0 ZrO 2 (T) 85.0 ZrO 2 (M) 0.2 ZrO 2 (M) 4.2 Ti 7.5 Nb 9.0 TiO 2 1.0 Nb 2 O 5 1.0 c. PSZ±Ni d. SiO 2 ±Ti ZrO 2 (T) 89.0 SiO 2 (a) 92.0 ZrO 2 (M) 0.2 Ti 7.5 Ni 9.4 TiO 2 ± NiO ± ± ± A typical electrical behavior of composite materials containing metallic particles embedded in a ceramic matrix is shown in Fig. 5. The electrical resistivity is plotted as a function of Ni concentration in a PSZ matrix. As the amount of metal inclusion is raised, the originally insulating material becomes conductor. The critical Ni volumetric fraction is 28 2%. This result is in agreement with the percolation model established by Landauer [13]. All other materials investigated in this work show a similar electrical behavior. Fig. 6 shows the electrical resistivity as a function of temperature for PSZ±Ti, PSZ±Nb, PSZ±Ni, and asio 2 ± Ti containing 30 vol% of metallic inclusions. The electrical resistivity increases linearly showing that the electrical conduction is electronic type and probably it will occur through the metallic conduction paths. The electrical resistivity of these composites are three orders of magnitude higher than the ones related to pure metals. This di erence is probably caused by impurities in the metal crystalline lattice, the cross section and length of J.R. Martinelli, F.F. Sene / Ceramics International 26 (2000) 325± conducting paths, which were not determined in the present work. As it is known, some metal alloys show resistivity values up to three orders of magnitude higher then pure metals [14]. Moreover, other factors can contribute in the reduction of electrical conductivity such as reduced contact area among conducting particles due to oxidation. Fig. 7 shows the electrical resistivity as a function of temperature in an Arrhenius plot for PSZ samples containing 10 and 20 vol% of Nb. The activation energies are 0.78 and 0.72 ev, respectively. The electrical resistivity of these samples, originally insulators at room temperature, decreases exponentially as the temperature is raised. The conduction mechanism is similar to the one established for pure PSZ, but the activation energy (E a ) and the pre-exponential factor are di erent [15]. The conduction is primarily ionic and occurs through the PSZ matrix. The PSZ±Nb composites show higher resistivity values than pure PSZ and this fact can be explain if it is considered that Zr +4 ions are replaced by Nb +5 in the PSZ lattice. To keep the electrical neutrality, each positive charge originated from the aliovalent ion in the lattice should be compensated by an electron and an oxygen atom. The electron will contribute for the electronic conduction but the oxygen atom will occupy an oxygen vacancy usually located in the lattice of a nonstoichiometric oxide, leading to higher resistivities. On the other hand a lower value for the activation energy has been observed in the PSZ±Nb, therefore it can be assumed that the mobility of oxygen ions is higher. The charge carrier mobility depends on the ceramic microstructure and the existence of grain boundaries leads to an increase of mobility. Niobium inclusions can lead to microstructure defects [16]. However the d.c. electrical measurements do not allow to separate grain boundary from bulk e ects. Eq. (1) shows the electrical balance due to the replacement of Zr +4 by Nb +5. 2Nb Nb 5O O! 2Nb Zr 4O O V Nb O 2 Therefore, as a consequence of this neutrality, niobium vacancy and oxygen atoms are generated. Fig. 8 shows the electrical resistivity as a function of temperature in an Arrhenius plot for PSZ samples containing 10 and 20 vol% of Ti. The activation energies are 0.98 and 0.93, respectively. These values are close to the ones for pure PSZ. The ion Ti +4 has a similar valence compared to zirconium in the PSZ lattice, and even if a replacement of Zr +4 by Ti +4 happens, there is no electrical unbalance in the 1 Fig. 3. Micrographical pictures taken by an optical microscope for: (A) PSZ±Ti 10; (B) PSZ±Ti 20; (C) PSZ±Ti 30; and (D) PSZ±Ti 45. 330 J.R. Martinelli, F.F. Sene / Ceramics International 26 (2000) 325±335 crystalline lattice, so no defects are generated to keep the electrical neutrality. Fig. 9 shows the electrical resistivity as a function of temperature in an Arrhenius plot for PSZ samples containing 10 and 20 vol% of Ni. The activation energies are 1.00 and 0.98 ev, respectively. The resistivity of PSZ±Ni is lower than the one for pure PSZ, and this fact can be explained considering that Zr+4 in the PSZ lattice is replaced by Ni+2. To keep the electrical neutrality, a doubly ionized oxygen vacancy is formed. The ion mobility in this material is similar to the one expected for the nominally pure PSZ. The following equation shows the electrical charge balance corresponding to the replacement of Zr+4 by Ni+2: NiNi OO! NiZr00 OO VO 2 Fig. 10 shows the electrical resistivity of cristobalite nominally pure and containing 20 vol% Ti as a function of temperature in an Arrhenius plot. The activation energies are 0.
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks