• 沒有找到結果。

Sintering Behavior and Interfacial Analysis of Ni/Cu Electrode with BaTiO3 Particulates

N/A
N/A
Protected

Academic year: 2021

Share "Sintering Behavior and Interfacial Analysis of Ni/Cu Electrode with BaTiO3 Particulates"

Copied!
12
0
0

加載中.... (立即查看全文)

全文

(1)

Sintering Behavior and Interfacial Analysis of Ni/Cu Electrode with

BaTiO

3

Particulates

CHIH-LUNG CHEN,1WEN- HSI LEE2& WEN-CHENG J. WEI1,∗ 1Institute of Mat’s Sci. & Eng., National Taiwan University, Taipei, Taiwan 106, Republic of China

2Phycomp Taiwan Ltd., P.O. Box 35-106, Kaohsiung, Taiwan, Republic of China

Submitted September 1, 2003; Revised March 1, 2004; Accepted March 9, 2004

Abstract. The sintering behavior of Ni electrode alloyed with Cu and the interfacial structure between Ni/Cu to BaTiO3 (BT) have been investigated. The quantitative properties, which include thermal shrinkage, thermal expansion, wetting behaviors of Ni/Cu alloys on BT sheet, and composition distribution were measured by several thermal analysis techniques (TGA/DTA/TMA) and microstructural techniques (SEM/TEM/ HRTEM) with energy-dispersive spectroscopy (EDS). The shrinkage of the Ni/Cu/ BaTiO3 composite tested in 5%H2/N2 atmosphere showed strong influence by the addition of Cu, and retarded slightly due to the addition of the BT particulates. The Cu alloyed with Ni improves the continuity of the electrode and does not trigger mutual reaction between Ni and BT.

Keywords: Ni, Cu, BaTiO3, electrode, interfacial reaction

1. Introduction

In order to keep pace up with the trend in the miniatur-ization of electronic components, concentrated effort has been put in forced on the development of small mul-tilayer ceramic capacitors (MLCCs) to meet the rapid development of integrated circuit and surface mounting technology. The demand of the MLCCs in high quality has quickly increased in the past decade. However, the high sintering temperature of BaTiO3 (BT) suggested the adoption of high melting point internal electrodes, such as Pd and Ag/Pd alloy. To confront the require-ments on smaller size, higher performance and lower cost, the manufacturing techniques of the MLCCs have closely engaged the development of base metal inter-nal electrode (BME) and thinner dielectric ceramic lay-ers [1–6]. Therefore, the fabrication of MLCCs lately has concentrated on the sintering process of BME and on the adjustment of dielectric ceramic composition, which are considered as the solutions of above critical issues [5, 7].

To whom all correspondence should be addressed. E-mail:

wjwei@ntu.edu.tw

Another challenging issue in manufacturing Ni-based MLCCs is cofiring the electrode/dielectric multilayer structure. Many essential issues in cofiring technology of multilayer films include not only the for-mation mechanism of interfacial defects, but also the modification of the mismatch of sintering shrinkage. Whereas the different sintering kinetics and mecha-nisms between the Ni electrode and BT dielectrics are monitored, various types of cofiring defects, includ-ing cracks, delamination, and warpage come into im-portant issues [3–9]. Those defects are often observed in the final products. In addition to those defects, sil-ver migration form inner electrode into ceramics in the cofiring process of MLCCs is another critical factor, which affects the reliability and dielectric characteris-tics [9, 10]. Moreover, the influence of carbon residue in the capacitor is also concerned. The residual carbon could increase the dielectric constant if the organics or polymer in the electrode couldn’t be burned-out com-pletely [11]. However, the investigation on the wetting and sintering behaviors of carbon with BaTiO3couldn’t be found in literature.

Metal-ceramic multilayers are prone to failure owing to the difference in the thermal expansion

(2)

coefficient. Besides, a difference in the sintering rates between the metal electrode and the ceramic dielectric layers, consequently, the thermal mismatch problem has been recognized in the past reports [11, 12]. A basic understanding of the process resulting in these defects is needed to optimize the processing parameters and material compositions that minimize the formation of the defects and thus increase the production yield.

Several reports [13, 14] of MLCCs have focused on making a thinner dielectric layer, since capacitance per unit volume is approximately inversely proportional to the square of the thickness. However, the interfacial re-action of electrode (metal) elements into the dielectrics (ceramic) on the electrical properties of MLCCs will be more critical with decreasing the thickness of dielec-tric layer [15]. On the other hand, some technological difficulties have been experienced, such as delaminat-ing of layer structure and discontinuity of the inner electrode of MLCCs. A multi-layer structure cofired metal-containing electrodes with BT may appear inter-diffusion between BT and metal layer [16, 17]. It is not desirable because of the reduction of breakdown voltage and dielectric quality.

Cu as the second metallic additive is selected as a potential alloying phase to Ni [18, 19]. The crystalline structure of Cu is face-centered cubic (FCC), which is the same as the structure of Ni. However, the Cu has an atomic radius just a few percents different to that of Ni. Thermodynamically, the interfacial energy of Ni al-loy system will be different from pure Ni. Cu is a high electrical and thermal conductive material, and Ni is very good for corrosion resistance. For these reasons, a mixture of Ni and Cu is proposed for the use as the elec-trode of MLCCs [20]. Besides, BT is often added to the metal electrode so to reduce the coefficient mismatch of thermal expansion, and to increase the sintering match with BT dielectric layers. Most importantly, interfacial reaction may occur between the internal metal (Ni/Cu) electrode and dielectric ceramic layer.

One study about the reaction between Cu and BaTiO3was reported by Richard et al. [21]. Two newly formed Cu/BaTiO3 compounds, Ba8Cu2Ti22O54 and Ba4Ti13O30, were identified in a mixture of BaCO3, CuO and TiO2(1:1:2 ratio) after sintered at 1380◦C for 48 h. The phase reflections were hardly found in the pat-tern of X-ray diffraction analysis. Interfacial reaction between BT to metallic electrodes was recognized to be an important role in the degradation of MLCCs [22]. In this study, the sintering behavior changing with the ad-dition of Cu was investigated. The interfacial analysis

Table 1. Exact content of the Ni/Cu/BT composites tested in

this study.

Exact content of each phase (mass%)

Mark Ni (%) BT (%) Cu (%) N15B 85 15 0 N15B10C 76.5 15 8.5 N15B30C 59.5 15 25.5 N30B30C 49 30 21 N60B30C 28 60 12

of Ni/Cu with dielectric ceramic layer is also carefully characterized in order to clarify the role of Cu.

2. Experimental Procedures

2.1. Sample Preparation

The original BaTiO3 powder (grade PR, Ba/Ti = 0.995, Degussa Uden, Japan) in average particle size of 0.67µm was used for the preparation of the plates in diameter of 5 mm and thickness of 3 mm. The BT samples were calcined and sintered at 1200◦C in air. Besides, two kinds of metallic powders were studied, including Ni (99.0%, median particle size is 0.3µm) and Cu (99.0%, median particle size is 0.6µm) sup-plied by Phycomp Co., Taiwan). The Ni and Cu can also be a form of paste in the solid content of either 71.2 mass% or 85.0 mass%.

BT ceramic tape with inner electrodes was also made by tape-casting for the multilayer specimens. In the following section, the abbreviations, N, B and C mean the Ni, BaTiO3and Cu phases, respectively. The number in front of the abbreviation is the content of the material in mass%, e.g. N20B representing Ni with 20 mass% BaTiO3phase. The exact composition of the composites is shown in Table 1.

2.2. Analyses of Phases, Composition and Microstructure

The phases of sintered bulks were identified by an X-ray diffractometer (Philips PW1710, Philips Co., Netherlands). The composition analysis of impuri-ties in the raw powder was analyzed by the Oxygen-Nitrogen and Carbon-Sulfur gaseous analyzers (LECO CS-136 and CS-244, LECO Co., USA). The mor-phologies of the polished and cross-section surfaces

(3)

of Ni/Cu/BaTiO3 composites were analyzed by scan-ning electron microscopy (SEM, Philips XL30, Philips Co., USA) equipped with an X-ray energy disper-sive spectroscopy (EDS, DX-4, EDAX Co., USA), and field-emission SEM (Leo Instrument 1530, England). Thin foil samples for TEM observation were cut to thin layers by Isomet (Buehler Ltd., USA), then was ground and reduced to <50 µm by using Minimet grinder (Buehler Ltd., USA). Afterward, the sample was milled by an ion miller (PIPS, Gatan Co., USA) until a thin perforated sample was obtained. Transmis-sion electron microscopes (TEM, 100CXII, JEOL Co., Jpn. or FETEM, Tecnai 300 kV, Philips, Netherlands) equipped with EDS were employed to analyze the microstructure and composition distribution near the interfaces.

2.3. Thermal and Shrinkage Analysis

The sintering behavior of the Ni/Cu/BT composites was analyzed by using a thermal mechanical analyzer (TMA, SETSYS TMA16/18, SETRAM Co., France), which could operate to 1700◦C. The linear expansion or shrinkage of these specimens was detected. Besides, TGA test was also conducted (Thermal Analyst 2000, Dupont Co., USA). About 10 mg of sample was placed in Al2O3crucible in each test. Laminated samples be-fore the thermal tests were always going through a suit-able binder-burn-out (BBO) step.

3. Results and Discussion

3.1. Dilatometric Behavior of Ni with BT

The effects of BT on the sintering of Ni are reported in Fig. 1(a). The submicron Ni powder starts to densify as low as 500◦C. As 5–20 mass% of BT powder is added, the threshold of sintering shifts to higher temperature. It is conspicuous that the densification of the pure Ni paste is almost complete before the temperature reaches the densification of pure BT powder compact. In addition, the Ni with BT powder shows a linear shrinkage of 15.5 to 17.0% when sintered at 1300◦C. The shrinkage is nearly 5% more than that of a pure Ni due to different packing efficiency of Ni to the binary Ni/BT powder systems.

The sintering behavior of the Ni is slightly altered by the addition of BT. The threshold sintering temperature

Fig. 1. (a) Dilatometric sintering curves showing the shrinkage

be-havior of Ni sintered with various amounts of BT at a rate of 10◦C/min to 1300◦C, and held for 10 min in 5%H2/N2atmosphere, (b) the

shrinkage properties of Ni with BT as a function of BT content.

(onset of sintering) is varied gradually. The addition of BT powders retards the shrinkage of the Ni electrode, leading to better interfacial matching. The results of the shrinkage and threshold temperature of each Ni/BT sample are summarized in Fig. 1(b). It is also noted that the shrinkage curves of the Ni/BT composites hold-ing at 1300◦C depict 2–4% additional shrinkage (a straight line), which might imply a liquid phase sin-tering. The phenomena will be investigated in the later section.

Figure 2 indicates the microstructure of layer, either pure Ni or N15B, between BT dielectric layers. The flaws are found at the interface. That is the lack of a good physical match between the layers resulted from inferior interfacial bonding (or called “de-bonding”). This kind of interfacial condition at the interface is

(4)

Fig. 2. SEM micrographs between BT ceramics and electrode layer of (a) pure Ni; and (b) 15 mass% BT-doped Ni. The samples were sintered

at 1250◦C for 1 h in 5%H2/N2atmosphere.

easily damaged, so to cause the interfacial delamina-tion. Compared to N15B, the electrode layer hardly shows improvement on the bonding conditions, as shown in Fig. 2(b). Besides, the continuity of the metal electrode is still poor. A good adhesion mechanism be-tween electrode layer and BT ceramics should be active unless the interface of thin electrode/BT layers always shows the defects. In the next section, the discontinuity of the electrode will be shown greatly improved by the addition of Cu.

3.2. Thermal Behavior of Ni/BT with Cu

A series of TGA and TMA tests of NBC composite samples was conducted in reducing atmospheres. The humidity in the flowing gas was pre-determined. The dew-point was−63◦C, which corresponds to a H2O content of 7 ppm, or an oxygen partial pressure of PO2∼ 5 × 10−11atm. The result indicated that the sin-tering condition in 5%H2/N2might provide an appro-priate reducing atmosphere protecting the metal phase from oxidation.

TGA tests of the samples with and without Cu were conducted. The results are shown in Figs. 3(a) and (b), illustrating that the oxidation behavior of the dry pastes depends on the content of the Cu and the atmosphere.

The mass gain of N15B sample occurs at≥360◦C im-plying the oxidation of Ni. In comparison, the oxidation curve of N15B20C shows more mass gain (Fig. 3(a)) and occurs from 250◦C. The occurrence of additional mass gain (ca. 2%) is due to the oxidation of Cu.

Figure 3(b) shows the TGA tests in reducing at-mosphere. The mass-change curves of N15B and N15B10C show a small reduction taking place from 250 and 290◦C, respectively. The amount of mass loss (0.3% or 0.6%) is proportional to the amount of Cu addition. Most importantly, the reduction of the Cu ox-ides is responsible form the mass reduction carried out in a dried 5% H2/N2 up to 310◦C. Nevertheless, the mass gain of Ni/Cu/BT above 300◦C of TGA result (Fig. 3) is possibly either the oxidation of copper or a background drifting of the TGA. However, based on the XRD analysis of the calcined samples, it depicted that only a single FCC phase appeared. No crystalline Ni- or Cu-oxides are detected. That means native ox-ide scale in amorphous condition could exist on the Cu powder (or bulk) surface.

3.3. Effect of Cu on Sintering of Ni-BT

Various amounts of Cu were added in the N15B bulk in order to investigate the densification behavior of

(5)

Fig. 3. TGA results of dried paste N15B (after BBO) with various

contents of Cu. The testing conditions are 5◦C/min (a) in air (b) in flowing 5%H2/N2atmosphere.

the electrode material. The amount of BT is fixed to 15 mass%. The dilatometric data are shown in Fig. 4, in which the curves can be categorized into two groups. One is the samples that contain less than 30 mass% Cu. The other is more than 30 mass% Cu.

Samples in the first-group (Fig. 4(a)) start sintering at 800◦C, and undergo faster shrinkage above 1080◦C. It is concerned that the Cu may melt near 1080◦C and will quickly alloy with Ni. That, especially the N15B30C, may result in a small volume expansion in the range from 1080 to 1200◦C. The amount of expan-sion gets more apparent, even higher than 1200◦C as the Cu content is≥30%. Every sample of low Cu-content (≥20%) shows similar sintering threshold tempera-tures between 700–800◦C, and a rapid rate of shrinkage above 1100◦C. A shrinkage phenomenon takes place as the sintering temperature holding at 1300◦C for 60 min (Fig. 4(a)).

Fig. 4. Dilatometric sintering curves of Ni/BT/Cu composites

sin-tered to (a) 1300◦C, and (b) 1200◦C, then held for 10 min in 5%H2/N2

atmosphere.

The second-group samples (Fig. 4(b)) show a dif-ferent shrinkage behavior from the samples in the first-group. First of all, small amount (1%) of the shrinkage in the second-group samples was measured in the tem-perature range from 300 to 450◦C, which might be the reduction of Cu-oxide. The surface of the submicron Cu particle in the original metal paste is likely in an oxidized condition. During the reduction of CuO in 5%H2/N2, the sample experiences volume shrinkage as below.

CuO→ Cu, V = −44% (1)

Therefore, the more the Cu content, the greater the shrinkage within the range of 300–450◦C.

(6)

Fig. 5. Dilatometric sintering curves of pure Cu and several

N15B50C composites sintered or treated at various temperatures in 5%H2/N2atmosphere.

It’s also noted that two expansions occur between 450 to 700◦C and 1080 to 1200◦C, respectively. In or-der to investigate the reasons causing the expansions, we prepared two sets of samples for comparisons. The results of the sintering tests are shown in Fig. 5. The first set includes two samples, one is pure Cu com-pact made by die-pressing, and the other is pretreated N15B50C which has been calcined at 900◦C for 10 min. The second set includes two samples as well, one is green N15B50C compact and the other N15B50C has been sintered at 1200◦C. That means that the organics (e.g. binder) in the paste or the impurity in Cu particles after 900◦C or 1200◦C treatment is cleaned from the sample. However, the states of Ni alloyed with Cu are also different after 1200◦C treatment.

The result of first set sample clearly reveals the dif-ference of the Cu powder compacts with calcinations, so to reveal the effects of any impurity in the raw Cu powder. The raw Cu powder had the impurity contents of O: 0.57%, N: 0.08%, C: 0.51%, and S: 0.004%. It is apparent that the carbon content is high possibly due to polymeric residue coming from Cu paste. Of course, it is concerned that the raw Cu and N15B50C (which shows C content of 0.09%) powders adsorbed carbonaceous residue on the particle surface. It is con-cerned that the carbon content may affect the wetting

behavior of Ni (or Ni/Cu) electrode. In this case, the carbon content of N15B50C was analyzed to be≤0.5%. The amount of carbon residue was typical value in the Ni/BT samples. It shouldn’t be the critical factor to the sintering characteristics. But, the thermal expansion in the range from 450 to 700◦C of both samples could be the results of evaporation of volatile gas (CO or CO2). The concentration of carbonaceous gas is enough to result in the volume expansion of the Ni/Cu/BT com-posites above 450◦C. It is consistent to the curve of pretreated N15B50C in Fig. 5 that shows no shrinkage up to 700◦C. The expansion didn’t occur in the pre-treated N15B50C specimen, in which the Cu did not have any gaseous residue after the pre-treatment.

The second expansion starting from 1080◦C in Fig. 4(b) can be the causes of Cu melting and/or al-loying effect. Liquid phase formation is close relation to the melting of Cu (Tm = 1065◦C). This is why the N15B50C sample pre-sintered at 1200◦C hardly shows any expansion, except an expansion aft ca. 1150◦C. That is the reason why the pre-sintered N15B50C does not show extensive linear shrinkage comparing to the original N15B50C which experiences 18% lin-ear shrinkage at 1200◦C.

The final shrinkage of N15B30C and N15B40C are 10 and 11.7% at 1200◦C for 1 h, respectively. But the sample of N15B50C experiences a greater shrinkage possibly because the sintering temperature is close to the melting temperature of Ni50Cu alloy. The wetting of Ni/Cu matrix on solid particles (BT) is more conspic-uous during sintering. Therefore, the surface tension of the Ni/Cu drives the particles together, resulting in the largest linear shrinkage (∼20%) of N15B50C among the NBC specimens.

3.4. Microstructural Continuity of Ni/Cu Electrode

The microstructures of various NBC composites sin-tered at 1300◦C for 1 h in 5%H2/N2 atmosphere are shown in Fig. 6. The Ni/Cu alloy appears darker con-trast, and the BT particles are in white contrast due to the effects of different electric conductivities for each phase. Figure 6(a) shows that the BT particles form clusters in the Ni matrix. According to percola-tion threshold, the BT clusters are still isolated (lower than 16 vol% inclusion). Besides, the average size of BT isolated particle is 1.43µm as shown in Fig. 6(a). The electric conductivity of the electrode (N15B) is still good as the Ni matrix is dense. However, the continuity

(7)

Fig. 6. SEM micrographs of the polished surfaces of (a) N15B, (b) N15B30C, and (c) N30B30C composites after sintered at 1300◦C for 1 h in 5%H2/N2atmosphere. (a) and (b) were imaged by secondary electron (SE) mode, (c) is by back-scattering (BS) mode.

(8)

of the electrode depend strongly on the wetting condi-tion of Ni on BT. Figure 6(b) shows the N15B alloyed with 30 mass% of Cu. The density of metal matrix of N15B30C composite is higher than the sintering con-dition of N15B composite in Fig. 6(a). The sintered N15B30C composite shows almost fully densification (≥99% T.D.) after this heat-treatment.

N15B30C composite shows a good wetting and the Ni/Cu alloy apparently cover the BT particle surface with the Ni/Cu alloy. As the contents of BT increased to 30% in the matrix, e.g. the N30B30C in Fig. 6(c), the BT phase obviously dispersed in the matrix. No

Fig. 7. SEM micrographs of (a) N15B and (b) N15B50C layers sandwiched between BT plates, and sintered at 1250◦C for 1 h in 5%H2/N2

atmosphere.

doubt the 30% BT clusters form continuous phase due to the content higher than percolation threshold (higher than 16 vol%). Besides, the porosity of the N30B30C composite becomes more owing to unfinished sinter-ing of BT particle cluster. Although BT particles were distributed in the mixture, they tended to coalesce as the increase of BT content. The effect of electrode sur-face area on dielectric constant will be also considered because of the additional surface area is formed in the Ni/Cu electrode with BT particles. That is, too much BT addition might reduce the dielectric properties of MLCC due to the decrease of the conductive area in

(9)

the Ni/Cu electrode. We have to take the densification and electric properties of the electrode into account. Therefore, the optimal formulation, N15B30C, for the composite is proposed in this study.

Figure 7 is a set of SEM images of inner N15B electrode layers without and with Cu additive. The im-ages show two important phenomena. One is the shape of pure Ni grains (Fig. 7(a)) which are nearly spher-ical, but irregular for N50C grains. Most of Ni parti-cles shrink and result in the dewetting due to surface tension and become spherical at sintering temperature. The other is that the average Ni grain size of N50C is slightly larger than that of pure Ni grains. Obviously, the pure Ni layer has the weakest interface and poorly bonded with dielectric ceramic layer. Additionally, the evidences tell that the difference of interfacial energies of Ni on BT is larger than N50C on BT, which depicts good wetting and bonding in Fig. 7(b). We think that liquid-phase sintering of Cu in Ni/Cu alloy may play an important rule during the sintering process. The images strongly suggest that the framework and structure of BT

Fig. 8. SEM and EDS (mapping) results of N15B50C electrode illustrating no obvious inter-diffusion between Ni/BT and Cu/BT co-fired at

1250◦C for 1 h in 5%H2/N2atmosphere. The partially agglomerated BT in the electrode layer causes some defects, including pores, fine cracks

and discontinuity between metal layer and BT plates.

ceramics are improved by the rearrangement of the BT particles due to the wetting of Ni/Cu alloy. Therefore, the continuity of the electrode would be enhanced by the addition of Cu. But the melting of Cu is a transition state and show volume expansion observed from the TMA curves shown in Fig. 4.

3.5. Interfacial Characteristics Between Ni/Cu and BT

A detail analysis of previous N15B50C samples is shown in Fig. 8. The inclusions in the electrode layer are BT (as analyzed by EDS spectrum). The BT grains form isolated colonies in the electrode. The Ni/Cu al-loy shows a well solid-state mixing in the electrode. In addition to this, adverse interfacial reactions may de-grade the electronic conductivity. The X-ray mapping images on the bottom of Fig. 8 appear a clear interface. No evidence of inter-diffusion between Ba and Ni/Cu is observed.

(10)

Fig. 9. (a) AEM micrograph of the 90%BT-10%Ni sintered at a rate

of 10◦C/min to 1300◦C, held for 1 h in 5%H2/N2atmosphere, and

(b) the qualitative composition analysis near the BT/Ni interface.

Figure 9 also depicts the ceramic phase (BT inclu-sion) shows a brighter contrast and uniform distribu-tion in the matrix. Composidistribu-tion analyses revealed that neither interdiffusion of elements nor interfacial reac-tion between the Ni and the BT phases occurred. The TEM/EDS results (Fig. 9(b)) appear the elaborate data for the interfacial behavior. The composition of BT and Ni is almost constant in each grain themselves. The interface between Ni and BT is clear, implying no inter-reaction occurred at the interface.

The solubility of NiO in BaTiO3had not been pre-cisely determined before. Some investigators [23–25] suggested that Ni ion would occupy Ba site because

Fig. 10. (a) AEM micrograph of the N15B10C sintered at 10◦C/min to 1300◦C for 1 h at 5%H2/N2atmosphere, and (b) the qualitative

composition analysis near the BT/Ni/Cu interface.

of the same valence state. The Ba site of BaTiO3 lat-tice replaced by Ni ion is still in dispute. However, low partial pressure of oxygen offered by 5%H2/N2is nec-essary to prevent the Ni from oxidation. Figure 10 also shown a clear interface between BT and the Ni/Cu al-loy imaged by AEM. No apparent defect, porosity, or reaction layers are found at the BT/NC interface. The Ni/Cu alloy shows a texture of metal phase. On the other hand, the detail EDS analysis is shown in Fig. 10(b). Herein the closest analysis point from the interface is in a distance of 10 nm. The composition of Ni/Cu atoms

(11)

Fig. 11. TEM images between Ni/Cu and BT interface in high resolution. The sample sintered at a rate of 10◦C/min to 1300◦C, held for 1 h in 5%H2/N2atmosphere.

abruptly reduces to nearly nothing across the interface. The residual Ni/Cu intensity (3–10 at%) found on the side of BT might imply almost no interfacial diffusion (or beam broadening).

A high-resolution TEM image of the interfacial structure between BT and Ni/Cu measured immedi-ately is presented in Fig. 11, indicating the clean and no defects including pores, crack, glassy phase and re-action layers at the interface. The diffrre-action patterns show both crystalline phases [030] zone pattern of BT phase, and the Ni phase [01-1] zone pattern, respec-tively. In the other hand, Fig. 11(b) shows the high-resolution TEM micrographs illustrating the atomic structures near the interface between intragranular Ni/Cu grains. The HRTEM micrograph appears a tilted interface of which only one or two layers of atoms are randomly arranged. The interface is as thin as 2–3 nm and depicts a non-coherent character. This layer has an average thickness of 2 to 3 atomic planes, and is uniform along the boundary. There is no glassy or in-termediate phase at the NC/BT interface during the sort of sintering condition. Most of the observed interfaces of the NC/BT belong to incoherent junction. Further-more, the BT layer exhibited a lattice structure with a spacing of 6.3 ˚A and the lattice plane aligned along the 101 axis.

4. Conclusion

The shrinkage behavior and microstructure develop-ment of Ni/Cu electrode with BaTiO3particulates has been investigated. The results indicate several abnor-mal changes of the dilatometric curves. The linear shrinkage for the NBC composite experienced the re-duction of Cu-oxide from 200 to 450◦C. Besides, the first volume expansion of the NBC composite from 450 to 700◦C was resulted from the evaporation of carbona-ceous impurity in the Cu powder. Sintering shrinkage started either above 800◦C for the case without Cu ad-dition, or induced volume expansion above 1080◦C due to the melting of Cu or alloying effect in Ni/Cu alloy.

Furthermore, 30 to 50% Cu addition into N15B ex-hibited excellent densification and wetting behavior in the Ni/Cu electrode comparing to the pure Ni elec-trode. Good interfacial stability between BT particles and metallic electrode were obtained in the Ni/Cu elec-trode. Besides, differential sintering behavior between pure Ni and BaTiO3ceramics could be reduced by the addition of Cu as the BT amount of 5 to 20 mass% in Ni electrode.

The studies of SEM/EDS and AES revealed that there was almost no inter-diffusion or reaction be-tween the Ni/Cu and pure Ni electrodes with BaTiO3

(12)

dielectric layer. Most importantly, the interface be-tween BaTiO3and Ni/Cu (or pure Ni) was very clean. The interfaces between BaTiO3 and Ni/Cu boundary are incoherent. No defects and no glassy phase could be found at the interface. The Cu does not only improve the continuity of Ni electrode layer, but also induce no interfacial reaction with BT ceramics.

Acknowledgment

The authors like to thank the kindly offer of some pro-cessing equipments by Phycomp Taiwan Ltd., and the funding given by National Science Council (NSC92-2212-E002-097) in Taiwan.

References

1. J. Yamamatsu, N. Kawano, and T. Arashi, J. Power Source, 60, 199 (1996).

2. H. Saito, H. Chazono, H. Kishi, and N. Yamaoka, Jpn. J. Appl.

Phys., 30, 2307 (1991).

3. G. With, J. Europ. Ceram. Soc., 12, 323 (1993).

4. Y. Park, Y.H. Kim, and H.G. Kim, Mater. Lett., 28, 101 (1996). 5. S. Sato, Y. Nakano, A. Sato, and T. Nomura, J. Europ. Ceram.

Soc., 19, 1061 (1999).

6. L. Paulsen and E.K. Reed, Microelectr. Reliab., 42, 815 (2002). 7. Y. Masuda, T. Koumura, T. Okawa, and K. Koumoto, J. Colloid

Interf. Sci., 26, 190 (2003).

8. S. F. Wang, J. Am. Ceram. Soc., 76(2), 474 (1993).

9. Y.I. Shin, K.M. Kang, Y.G. Jung, J.G. Yeo, S.G. Lee, and U. Paik, J. Europ. Ceram. Soc., 23, 1427 (2003).

10. R. Zuo, L. Li, and Z. Gui, Mat. Chem. Phys., 74, 182 (2002). 11. R.R. Tummala, J. Am. Ceram. Soc., 74, 895 (1991).

12. S.G. Lee, U. Paik, Y.I. Shin, J.W. Kim, and Y.G. Jung, Mater.

Design., 24, 169 (2003).

13. G.W. Howatt, R.G. Breckenridge, and J.M. Brownlow, J. Am.

Ceram. Soc., 30, 237 (1947).

14. R.E. Mistler, Am. Ceram. Soc. Bull., Oct., 82 (1998). 15. R. Chen, X. Wang, L. Li, and Z. Gui, Mater. Sci. Eng., B99, 298

(2003).

16. H. Hyuga, Y. Hayashi, T. Sekino, and K. Niihara, Nanostruct.

Mater., 9, 547 (1997).

17. H. Emoto and J. Hojo, J. Ceram. Soc. Jpn., 100, 553 (1992). 18. W.D. Callister, Materials Science and Engineering An

Introduc-tion, 3rd ed., 1994.

19. D.F.K. Hennings, J. Europ. Ceram. Soc., 21, 1637 (2001). 20. D.R. Lide, CRC Handbook of Chemistry and Physics, 72th ed.,

1911 (1992).

21. D.A. Richard and R. Layland, Polyhedron, 16(12), 2105 (1997). 22. K.M. Kang, Y.G. Jung, U. Paik, and J.H. Kim, Mater. Res. Bull.,

38, 555 (2003).

23. J.P. Bonsack, Am. Cream. Soc. Bull., 50, 488 (1971).

24. H.J. Hagemann and D. Hannings, J. Am. Ceram. Soc., 64, 590 (1981).

25. O. Parkash, H.S. Tewari, L. Pandey, R. Kumar, and D. Kumar,

數據

Table 1. Exact content of the Ni/Cu/BT composites tested in this study.
Fig. 1. (a) Dilatometric sintering curves showing the shrinkage be- be-havior of Ni sintered with various amounts of BT at a rate of 10 ◦ C/min to 1300 ◦ C, and held for 10 min in 5%H 2 /N 2 atmosphere, (b) the shrinkage properties of Ni with BT as a funct
Fig. 2. SEM micrographs between BT ceramics and electrode layer of (a) pure Ni; and (b) 15 mass% BT-doped Ni
Fig. 4. Dilatometric sintering curves of Ni/BT/Cu composites sin- sin-tered to (a) 1300 ◦ C, and (b) 1200 ◦ C, then held for 10 min in 5%H 2 /N 2
+7

參考文獻

相關文件

“Tests of an American Option Pricing Model on the Foreign Currency Options Market.” Journal of Financial and Quantitative Analysis, 22, No.. Bogle on

The product of the luminescent nano materials on the film were illuminated with LED and luminance value were measured with incident angles of 0 o 、30 o 、45 o 、60 o 、90 o..

The analysis of biomechanics after the total hip replacement can be reached by comparing with the stress and strain distribution of the intact femur and “Tripot Stemless Total

Therefore, the purpose of this study is to perform a numerical analysis on the thermal effect of shape-stabilized PCM plates as inner linings on the indoor air temperature

The mechanical properties of organic solder ability preservatives (OSP) Cu substrate with a Sn-3Ag-0.5Cu-8In-1Zn Pb-free solder have been studied.. For comparison, a

If we cannot understand or appreciate the human behavior, we will feel depressed and confused with the behaviors which are not coincidence with our world view.. These kinds

Carrying on the analysis by the expert fuzzy Delphi method regarding this investigation results, the research seeks and induces the most suitable weight of the leading behavior

Chan, “Effect of Intermetallic Compounds on the Thermal Fatigue of Surface Mount Solder Joints,” IEEE Transactions on Components, Packaging, and Manufacturing Technology B, Vol.