HAL Id: jpa-00224667
https://hal.archives-ouvertes.fr/jpa-00224667
Submitted on 1 Jan 1985
HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.
MICROSTRUCTURE AND INTERFACIAL
REACTIONS IN COPPER-CERAMIC MATERIALS
J. Poetzinger, S. Risbud
To cite this version:
J. Poetzinger, S. Risbud. MICROSTRUCTURE AND INTERFACIAL REACTIONS IN COPPER- CERAMIC MATERIALS. Journal de Physique Colloques, 1985, 46 (C4), pp.C4-147-C4-152.
�10.1051/jphyscol:1985418�. �jpa-00224667�
Colloque C4, supplément au n04, Tome 46, avril 1985 page C4-147
MICROSTRUCTURE AND I N T E R F A C I A L R E A C T I O N S I N COPPER-CERAMIC M A T E R I A L S
J.E. Poetzinger and S.H. Risbud
University of IZZinois at Urbana-Champaign, Department of Ceramic Engineering and Mater-iaZs Research Laboratory, Urbana, Illinois 61801, U.S.A.
Résumé
-
Les réactions ayant lieu entre les phases, la microstructure et la micro-chimie du système de la céramique cordierite-cuivre ont été examinées.Les résultats de diffusion et dissolution du CU' ions dans le verre, la préci- pitation de Cu20 et la décomposition du cuivre métallique sont présentés.
Abstract
-
Interfacial reactions, microstructure and microchemistry in a cordierite ceramic-copper system were examined. The results of diffusion and dissolution of CU+ ions in the glass, precipitation of Cu20 and decomposition to metallic copper are presented.INTRODUCTION
Because copper thick films provide enhanced conductivity and cost savings, there is considerable interest in employing thick film copper metallization in multilayer dielectrics and in microcircuit structures. During firing, the thick f lm copper reacts with the ceramic dielectric and forms a pink reaction zone.'-' The purpose of this study was to characterize the copper/ceramic reactions responsible for the pink zone formation in a cordierite type gla ceramic/thick film copper system fired in a reducing atmosphere, po2 = 10-
78
atm. Samples were examined after firing at 800°C and at 1000°C.
Microstructure and composition of the pink zone were characterized by optical microscopy, cathodoluminescence, electron microprobe, x-ray diffraction and transmission electron microscopy.
PROCEDURE
Samples were prepared with a commercial copper thick film paste containing less than 1% copper oxide. The glass ceramic composition was similar to that of Corning 9606 (cordierite type) with additions of several percent of P205 as a nucleating agent. The crystallizatin temperature for this glass is
approximately 975OC. Samples were fired in a reducing atmosphere (p -10-l0 atm) at 800°C (below the crystallization temperature) and at 1000°C ?above the crystallization temperture). A Nuclid Corporation ELM-2A cathodoluminoscope was used to observe the blue fluorescence characteristic of CU+ in solution in a glass. Operating conditions for the luminoscope were an accelerating voltage of 10 Kev and beam current of 100 microamperes.
A JEOL mode1 JXA-50A scanning electron microprobe was employed for microprobe analysis. Intensity profiles for Cu-K a radiation were obtained using a LiF crystal and a wavelength dispersive spectrometer and were graphically
recorded. Operating conditions for profiling were an accelerating voltage of 20Kev and a beam current of 40-50 nano amps. Quantitative point analyses across the copper-ceramic reaction zone were accomplished by collecting counts for Si, Al, Mg, P and Cu for 120 seconds every 10 p
.
The results were corrected for atomic number, absorption and fluorescence (ZAF) and oxygen was calculated by dif f erence.
Powder x-ray patterns were taken with a Philips Norelco automatic scanning vertical powder diffractometer. Cu-K a radiation was employed for powder patterns. Optimum signal/noise ratios were attained by para focusing the radiation to increase sample exposure and by using a curved graphite
monochromator after the sample. Operating conditions for the diffractometer were an accelerating voltage of 40keV and filament current of 12mA.
Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphyscol:1985418
C4-148 JOURNAL DE PHYSIQUE
Microstructural and mlcrochemical analyses of the reaction zone were performed with a Philips 400T transmission electron microscope. Thin sections were prepared by thinning to 80 p with Sic, then ball cratering with a 30mm diameter steel ball coated Irith 1 diamond paste to a Crater depth of 60 p.
and finally ion milling with 4kV argon ions.
RESULTS AND DISCUSSION
~eactions and Microstructure at 800°C
No pink reaction zone wae observed after firing at 800°C. However, Cu-K a intensity profiles from electron microprobe analysis indicated that copper had penetrated about 120 1i into the cordierite base glass. Further, cathodo luminescence examination revealed that the glass near the copper thick film, in the area found to be copper containing, fluoresced a bright blue.
Fig. 1
-
Cu-K a intensity profile (a) and cathodoluminescence micrograph (b) of the thick ilm copper-glass reaction zone after firing at 800°C, po2 -10-l5 atm. The Cu-K a intensity profile indicated that copper had diffused 120 p into the glass. Fluorescence of the glass in this area revealed that the copper speciesdif fusing was Cu+.
It has been established that glasses containing coppe n solution as a Cu4 ion are both colorless and exhibit blue fluorescence.'' As the samples fired at 800°C did not show any pink color and evidenced blue fluorescence near the copper thick film, it is reasonable to conclude that at 800°C copper diffused into the gl ss as a CU+ ion. Thermodynamic studies by other authors confirm that the Cuqf
-
Cu+-
Cu0 equi bria should fav r he CU+ ion under the reducing atmosphere,
p O ~ = Io-
''
atm, at ~ O O O C . ? - ~ ~ R o m the copperconcentration profile s own in figure 2 a dif fugion goef f icient for CU+ in the cordierite base glass was estimated to be -10- cm /sec.
3
1.00-
à?
3f
0.50-
1
O 20 40 60 80 100 120
Distance From Interface ( p l
Fig. 2 Copper concentration profile across the copper-glass interface for a sample fired at 800°C for -20 hr. indicates that less than -1.5 wtX copper is present in the lass The profile was used to estimate (at 800'~) rn cm2/sec.
formation of both small droplets (less than 100 nm in size), and tiny dark particles approximately 20 nm in size (Fig. 3a). X-ray spectra showed that the droplets, most likely phase separated areas, were depleted in silica compared to the bulk glass (Fig. 3b). The tiny particles, too small to analyze, may have been heterogeneous nucleation sites.
Fig. 3 TEM micrograph (a) and x-ray spe tra (b) of the bulk glass for a sample fired at 800°C atm. Micrograph shows both phase separated areas'(iy2and microheterogenieties ( 8 ) . X-ray spectra indicated that the droplets were depleted in silica compared to the base glass.
Reactions and Microstructure at 1000°C
Reaction between the thick-film copper and the glass ceramic after firing at 1000°C resulted in a pink reaction-zone extending -170 into the ceramic. A Cu-Ka intensity profile confirmed the presence of copper in the pink area.
Cathodoluminescence revealed a dim blue fluorescence (fig. 4) that was much less intense than the fluorescence observed after firing at 800°C (fig. 1). TEM micrographs showed the presence of copper-rich precipitates less than 100nm in size in the pink area (Fig. 5a). These precipitates were identified from microdiffraction patterns to be copper metal (Fig. 5b). The decrease in fluorescence in the 1000°C samples compared to the 800°C samples indicates that copper precipitates formed at the expense of CU+ ions in solution. Such a decrease in fluorescence accompanying red color formation has been
documented for copper ruby glasses.
Fig. 4 Cu-K a intensity profile (a) and cathodoluminescence micrograph (b) for a sample fired at 1000°C. Fluorescence after firing at 1000°C was substantially less than that noted on 800°C
samples. (Fig. 1).
C4-150 JOURNAL DE PHYSIQUE
Fig. 5 TEM micrograph (a) of copper precipitate responsible for red coloring. Indexing of the microdiffraction pattern (b) of the particle revealed that it was metallic copper. (152
Orientation).
Red or pink coloring of glass due to the precipitation of copper rich
particles has been extensively studied.14-l8 There has been some controversy as to whether the red coloring centers are Cu20 or metallic copper
particles. Mechanisms proposed for red color formation are shown in Table 1. Only two of
TABLE 1
Mechanisms proposed for the formation of red coloring centers for copper ruby glass.
Coloring Center
1
Mechanism Proposed 1) n Cu0 + (Cu), (ref. 4)2)
CU+ +
sn2+ + 2Cu0+
sn4+ (ref. 4) 3) 2 ~ u + + CU'+
cu2+ (ref. 4)4) ; S i - O - & + C u - O - S i +
Cu,O
+
z Si-
O-
Si 5 (ref. 17) these mechanisms, 3 and 4, can account for the decreased CU+ concentration (decreased fluorescence) which accompanied the pink color formation for samples fired at 1000°C. Mechanism 3, coloring by metallic copper would result in the formation of cu2+ ions .4,pf2+ in solution in glass in amounts as small as .O5 wt% yields a blue color. The samples fired at 1000°Ccontained approximately 1 wt% copper in the pink zone. Even if only half of the copper precipitated, 0.5 wt% cu2+ ion would be formed and a bluish or purple color should be observed. Because no blue color is apparent, mechanism 4, coloring by Cu20, is more satisfactory. However, copper rich precipitates contained in the pink zone were identified as copper metal by
microdiffraction, as shown in Fig. 5.
Simple thermodynamic calculations for the reaction i) ~ C U ~ O ( ~ ) 4 C u ( , )
+
log T
+
1.1 x 1 0 - ~ T2+
.1 x 105/T-
21.93 T," indicates that at 1000°C under po2 = 10-l0 atm the reduction of Cu20 to copper metal is favorable ( AGlOo00 =-39,300 cal). Thus, it is possible that Cu20 particles formed at the expense of Cu+ in solution in the glass resulted in the observed decrease in
fluorescence and that these particles then decomposed to copper metal by reaction i) explaining the identification of metallic copper. Atma et al reported such dec position for ruby glasses melted under p = 1O-letm when held above 550°C.' The decomposition was not noted at 520'8. Thermodynamic calculations predict their results. At 520°C, AGT = 1,100 cal and the decomposition is not favored while at 550°C AGT = -300 cal and the decomposition should occur.
The cordierite-type glass ceramic was predominantly crystalline after firing at 1000°C. Powder x-ray diffraction indicated that the major phase was a -cordierite. However, cordierite peaks were shifted to lower d spacings. An Al O3 contaminant which acted as an interna1 standard did not show such a shift. A shift to lower d-spacings has been reported for p-CO dierite and was found to be caused by excess Si in the cordierite structure5'. Tt is likely that the p-cordierite formed contains excess Si.
TEM analysis indicates that there is -5% of a second phase, a magnesium silicate enriched in phosphorus. This phase could not be positively
identified by x-ray diffraction because of swamping out of the spectra by the cordierite peaks and because there was only a small amount of P.
CONCLUSIONS
During firing at 800°C in a reducing atmosphere of po2 = 10-l0 atm CU+
diffused into the cordierite base glass resulting in a blue fluorescence adjacent to the copper-thick film. The glass ceramic i-emained amorphous but some phase separation was noted. After firing at 1000°C a pink reaction zone was observed and fluorescence was less than that of the 800°C samples. TEM analysis revealed metallic copper particles less than 100 nm in size in the pink area. The metallic copper particles may have formed by precipitation of Cu20 followed by decomposition to metallic copper. The glass ceramic
crystallized after firing at 1000°C forming predominantly a-cordierite containing excess silica and 5% of a magnesium silicate enriched in phosphorus.
JOURNAL DE PHYSIQUE
REFERENCES
1. P i t k a n e n , D. E., Cwmnings, J. P., S a r t e l l , J. A., Proc. ISHM, (1979) 148.
2. P i t k a n e n , D. E., Cummings, J. P., S a r t e l l , J. A., Proc. Isham, (1980) 49.
3. Val, C. M., P r i b a t , D., C o t t o , D., Proc. ISHM, (1980) 37.
4. Weyl, W. A., Coloured Glasses, ( S h e f f i e l d , U. K.: The S o c i e t y of Glass Technology), pp. 154-67, 421-32.
5. Karapetyan, G. O., I z v . Akad. Nauk SSSR Ser. Fiz.,
25,
(1961) 535.6. Rodriquez, A. R., Parmelee, C. W., Badger, A. E., J; Am. Ceram. Soc.,%, (1943) 137.
7. Parke, S., Webb, R. S., Phys. & Chem. G l a s s e s ,
2,
(1972) 157.8. L e s t e r , W. R., Glass Ind.,
3,
(1931) 83.9. Douglas, R. W., I n t . Congress on G l a s s , 10th Proceedings, Kyoto, JPN, J u l y 8-12, P a r t 1 (Tokyo: Ceram. Soc. of JPN, 19741, pp. 45-70.
10. J o h n s t o n , W. D., Chelko, A., J. Am. Ceram. S o c . , s , ( 1 9 6 6 ) 562.
11. Singh, S. P., P r a s a d , G., Nach, P., J. Am. Ceram. SOC.,^, (1978) 377.
12. Nath, P., v a g i , B. S., Glass and Ceram. B u l l . , x , (1972) 80.
13. Douglas, R. W., Nath, P., P a u l , A., Phys. & Chem. G l a s s e s , 6 (1965) 216.
14.. Brown, S. F., Norton, F. H., J. Am. Ceram. Soc.,
42,
(1959)-499.15. Atma, Ram, P r a s a d , S. N., Vaisch, V. K., Verma, S. S., Upadhyaya, V. G., Cent. Glass & Ceram. Res. I n s t . Bull.,
6,
(1959) 34.16. B a n e r j e e , S., Paul, A., J. Am. Ceram. Soc.,=, (1974) 286.
17. Atma, Ram, P r a s a d , S. N., Advances i n G l a s s Technology (New York: Plenum P r e s s , 1962), pp. 256-69.
18. Atma, Ram, P r a s a d , S. N., S r i v a s t a v a , K. P., S p r e c h s a a l
102,
(1969) 315.19. Atma, Ram, P r a s a d , S. N., Vaish, V. K., Cent. Glass and Ceram. Res. I n s t . B u l l . ,
6,
(1959) 86.20. G a s k e l l , D. R., I n t r o d u c t i o n t o M e t a l l u r g i c a l Thermodynamics, (Washington, D. C.: S c r i p t a P u b l i s h i n g Co., 1973), pp. 386.
21. Zdaniewski, W., Ph. D. T h e s i s , Chalmers Univ. and Univ. of Gothenburg, Goteborg, Sweden, 1975.
