Characterization of Interfaces in Diamond/Metal Composites by Raman Scattering Spectroscopy

Abstract

The interfaces between diamond and metal (Al or Ag) in the diamond/50 vol%Al and diamond/50 vol%Ag composites were characterized by Raman scattering spectroscopy. In the diamond/50 vol%Al composite, graphitization of diamond has been partly detected in the diamond region close to the interface between diamond and Al. On the other hand, a fairly thick graphitized layer of about 8 μm has been formed in the diamond/50 vol%Ag composite. It is concluded that the thick graphitized layer is attributed to the higher sintering temperature in the diamond/50 vol%Ag composite.

1 Introduction

Up to the present time, much attention has been focused in the SPS processing as an innovative sintering of various powder materials. For example, the successful results of synthesizing highly transparent ceramics^1–4) and functionally-graded ceramics/metal or ceramics/ceramics composites^5–8). In order to obtain the highly thermal-conductive composites by SPS processing, we have developed a new sintering technique based on the co-existence of liquid and solid phases and we have succeeded to synthesize the ceramics/metal composites with high thermal conductivity including diamond/metal^9–11). According to our previous results, the high thermal conductivities of 552 W/(m·K) and 717 W/(m·K) have been realized for the diamond/50 vol%Al¹⁰⁾ and diamond/50 vol%Ag¹¹⁾. Fully densification and ideal interfacial structures are necessitated for decreasing thermal resistance between the constitutional materials. Though the attainment of ideal interfacial structures for thermal transfer is suggested by the successful synthesis of the diamond/metal composites with very high thermal conductivity, it is further necessary to investigate the interfacial structures of these composites in details. In the present study, it is tried to clarify the interfacial states by examining the state of carbon in the interfacial region between diamond and Al or Ag in the diamond/50 vol%Al and diamond/50 vol%Ag composites by using Raman scattering spectroscopy.

2 Experimental Procedures

Both the diamond/50 vol%Al and diamond/50 vol%Ag composites were synthesized by the SPS processing based on the co-existence of liquid and solid phases. Single-crystal diamond (Tomei Diamond Co. Ltd., average size; 310 μm), pure Al powder (Mitsuwa Chemical Co. Ltd., 99.9 %, 104 μm under (150 mesh)) and Al-5 mass%Si alloy (Hikari Sozai Kyogyo Co., average size; 70 μm) were used as the starting materials to synthesize the diamond/50 vol%Al. The mixing ratio of Al to Al-5 mass%Si alloy was fixed to 9/1, resulting in the addition of 0.5 mass%Si in total matrix. After mixing these raw materials well, the mixed powder was SPSed at a temperature of 876 K (603 °C) under a pressure of 80 MPa. Heating rate was selected to be 65 K/min in the temperature range from RT to 798 K (525 °C) and 3 K/min in the temperature range from 798 K (525 °C) to 876 K (603 °C).

In the synthesis of the diamond/50 vol%Ag, the same single-crystal diamond, pure Ag powder (Kohjundo Kagaku Co., Ltd., 99.9 %, average size; 2 μm) and pure Si powder (Kohjundo Kagaku Co., Ltd., 99 %, 74 μm under (200 mesh)) were used as the starting materials. The amount of Si added to Ag was fixed to 0.28 mass% so that the ratio of liquid phase could become 10 % at the eutectic point(3 mass%Si) in the Ag-Si system. The mixed powders of these raw materials were sintered at a temperature range from 1073 K (525 °C) to 1178 K (603 °C) under 80 MPa, similar to the case of diamond/50 vol%Al composite. Heating rate was selected to be 100 K/min in the range from RT to 1073 K (800 °C) and 5 K/s in the range from 1073 K (525 °C) to 1178 K (603 °C). Details on the syntheses of the diamond/50 vol%Al and diamond/50 vol%Ag composites are described in the our previous papers^10,11).

The specimens for the measurement of Raman spectra were prepared by polishing the diamond/50 vol%Al and diamond/50 vol%Ag composites by using composite-type polishing papers composed of diamond particles and Ni matrix. By selecting a suitable diamond particle in the flattened surfaces of these composites, Raman spectra were measured by scanning the He-Ne laser beam (632.8 nm) linearly at the step of 0.5 μm so as to cross the diamond-Al (or Ag) interface, operating a spectrometer of LabRAM-ARAMIS (Horiba Co. Ltd.).

3 Results and Discussion

Firstly, the appearance of the polished surfaces of the diamond/50 vol%Al and diamond/50 vol%Ag composites are shown in Fig. 1. Both photos obtained from optical and SEM observations are shown in this figure. From SEM observation, it is found that diamond particles in both the composites are distributed homogenously. Enough densification are also observed from these photos and, according to our previous results, it has been reported that nearly full densification was reached in the diamond/50 vol%Al composite¹⁰⁾ and slightly lower densification was observed in the diamond/50 vol%Ag composite¹¹⁾. Further, it has indicated that the experimental value of thermal conductivity (552 W/(m·K)) in the diamond/50 vol%Al composite is close to the value predicted by Maxwell-Eucken equation, while the experimental value (717 W/(m·K)) corresponds to the 80 % value of the predicted one in the diamond/50 vol%Ag composite, though the value of 717 W/(m·K) is quite high thermal conductivity.

Fig. 1

Appearances of the polished surfaces of the diamond/50 vol%Al and diamond/50 vol%Ag composites. (a and b : optical photographs of diamond/50 vol%Al and diamond/50 vol%Ag composites, c and d : SEM photographs of diamond/50 vol%Al and diamond/50 vol%Ag composites)

Figs. 2 and 3 show the Raman spectra obtained from the diamond/50 vol%Al composite and the diamond/50 vol%Ag composite. The numerical values written in these figures show the distance from the interface between diamond and metal phase (Al or Ag). The sign of plus or minus means that the spectrum is obtained from diamond region or metal region, respectively. The position of interface was determined by the drastic change in the dependence of the intensity on the relative distance, which will be indicated later. In the spectra obtained from the diamond/50 vol%Al composite, as shown in Fig. 2, a sharp peak is observed at 1331 cm⁻¹ and the peak is assigned by diamond D-peak¹²⁾, which is attributed to the diamond particle. While two broad peaks observed near 1330 cm⁻¹ and 1600 cm⁻¹ are attributed to the thermally-changed diamond and assigned to the disordered graphite (notated by D(graph) and G(graph)), though these peaks are observed at a slightly higher and lower positions than those (1350 cm⁻¹ and 1582 cm⁻¹) observed in the standard disordered graphite, respectively¹²⁾. The positions of the two peaks and their broadness probably depend on the degree of disorder, graphitization and/or collapse of diamond structure¹³⁾. Therefore, it will be important factor, especially for thermal conductivity, whether the thermally-changed diamond at the outermost surfaces of diamond particles is graphite-like or diamond-like. In the spectra obtained from the diamond/50 vol%Ag composite, the sharp diamond D-peak is observed in the inner region of diamond particle and the intense D(graph) and G(graph) peaks due to disordered graphite are detected after disappearance of the sharp diamond D-peak, and thereafter these peaks are suddenly weakened in the region far from the diamond particle.

Fig. 2

Raman spectra obtained from the diamond/50 vol%Al composite. (The numerical figures attached in each spectrum mean the relative distance from the diamond/Al interface. See text.)

Fig. 3

Raman spectra obtained from the diamond/50 vol%Ag composite. (In this figure, the ordinate scale is described by cps so as to understand the high intensity of the peaks due to disordered graphite. The numerical figures attached in each spectrum mean the relative distance from the diamond/Ag interface. See text.)

Examining the Raman spectra near the diamond-Al interface, the intensity of the peak near 1600 cm⁻¹ increases with approaching the interface as shown in Fig. 4. However, the intensity is weak, compared with the diamond D-peak. The result shows the partial transformation of diamond to disordered graphite and the amount of disordered graphite gradually increases with approaching to the interface. The gradual increase of disordered graphite is attributed to the attacking effect of Al (or Si) on the diamond during sintering. Further, irregular detection of weak G(graph) and D(graph) peaks is attributed to the residual disordered graphite which may be produced by peeling the disordered graphite formed at the metal-diamond interface and/or by the deterioration of diamond in the composite when the composite is polished by diamond pastes. However, there is no evidence on the formation of aluminum carbide from Raman spectrometry.

Fig. 4

Raman peaks near 1600 cm⁻¹ obtained from the neighbor of the diamond-Al interface.

Subsequently dependence of the sum of intensities of the diamond D-peak, D(diamond), and graphite D-peak, D(graph), on the relative distance in the diamond/50 vol%Al and diamond/50 vol%Ag composites are shown in Figs. 5 and 6, respectively, where the abscissa scale is given by taking the starting point of Raman spectrum measurement in the metal side as the origin and the ordinate scale is given by the sum of intensities of D(diamond) and D(graph) peaks, though the ordinate scale corresponds substantially to D(diamond) or D(graph) in the region without coexistence of diamond and disordered graphite. Further, it was judged that the detection of a lower intensity than 2000 cps is due to the adhered graphite on metal. In the diamond/50 vol%Al, somewhat decrease in the sum of intensities of the two peaks is observed with occurring the partial transformation of diamond, in other words, in the coexisting region of diamond and disordered graphite and thereafter the intensity drastically decreases. On the other hand, in the diamond/50 vol%Ag, the higher intensity is observed in the disordered graphite region than in the diamond one and the drastic decrease appears in the region of existing only disordered graphite. The drastic decrease shows that fairly thick disordered graphite layer is formed between diamond particle and Ag metal in the diamond/50 vol%Ag composite. The formation of graphite layer cannot be attributed to the attacking effect by Ag metal because Ag is not reactive to diamond and carbide-forming metal. The thick graphite layer is formed in such a form as wrapping a diamond particle, so that thermodynamic decomposition of diamond into graphite may cause to form the thick graphite layer. In general, the thermal stability of diamond decreases with the increase of the atmospheric temperature and diamond degrades abruptly into graphite when the temperature exceeds about 800 °C. Therefore, the difference in the formation of graphite layer is simply attributed to the difference in sintering temperature because the maximum sintering temperatures are 603 °C and 910 °C in the diamond/50 vol%Al and diamond/50 vol%Ag composites, respectively.

Fig. 5

Dependence of the sum of the intensities of the diamond D-peak and graphite D-peak (D(diamond) and D(graph)) on the relative distance in the diamond/50 vol%Al composite.

Fig. 6

Dependence of the sum of the intensities of the diamond D-peak and graphite D-peak (D(diamond) and D(graph)) on the relative distance in the diamond/50 vol%Ag composite.

From the Raman spectroscopic analysis, the thicknesses of disordered graphite layers changed from diamond are estimated to be about 5 μm and 8 μm in the diamond/50 vol%Al and diamond/50 vol%Ag composites, respectively. Though the difference in thickness is not so larger than that expected from the difference in the sintering temperature, large discrepancy is found in the degree of degradation of diamond. As indicated above, the outermost surface layer of diamond particle partly degrades in the diamond/50 vol%Al composite and it is judged from the coexistence of diamond D-peak and graphite D(graph) peak that diamond partially changes into the disordered graphite in the composite. On the other hand, in the diamond/50 vol%Ag composite, the outermost surface layer of diamond particle degrades and changes completely into disordered-graphite with 8 μm in thickness. Thus the degradation of diamond in the diamond/50 vol%Ag composite is very severe on account of a higher sintering temperature than 900 °C.

In the present study, silicon was used as the alloying element for Ag in order to form liquid phase during sintering. In the diamond/50 vol%Ag composite, the addition of Si was carried out in the elemental form, unlike in the diamond/50 vol%Al composite. Because the eutectic reaction between Ag and Si is somewhat retarded by the addition of elemental Si, unlike in the case of alloy form, the effect of liquid phase on densification may be somewhat decreased, compared with the addition of Si in the form of Ag-Si alloy.

According to the phase diagram, the eutectic temperature in the Ag-Si system is 835 °C and the temperature is fairly higher than the eutectic temperature (577 °C) in the Al-Si system. Therefore, other alloying elements should be selected in order to form a liquid phase at a lower temperature by lowering the eutectic temperature in the second alloy of Ag. For example, Ag-Al alloy is a candidate one for the formation of a liquid phase at a lower temperature than 835 °C by using an Ag-Al alloy because the peritectic temperature is 778 °C in the Al-rich side of the Ag-Al system. Further lowering of the peritectic temperatures to 611 °C and 695 °C can be obtained when Ga or In is used for second alloying element instead of Al. Thus the improvement of thermal conductivity in diamond/50 vol%Ag composite can be expected by the decrease of sintering temperature due to the lowering of liquidus temperature of second alloy and the derivative effect on the improvement of densification.

4 Summary

The interfaces of two diamond/metal composites (diamond/50 vol%Al and diamond/50 %Ag) with high thermal conductivity have been analyzed by Raman spectroscopy. Graphitized layers have been detected in both composites and the thickness are estimated to be 5 μm and 8 μm in diamond/50 vol%Al and diamond/50 %Ag composites, respectively. Though the thickness is not so large between two composites, a large difference is found in the degree of degradation into graphitization. In the diamond/50 vol%Al composite, partial degradation of diamond has been detected in the surface layer of diamond particle, while severe degradation of diamond into graphite has been detected in the surface layer of diamond particle in the diamond/50 vol%Ag composite. The difference is simply attributed to the difference in the sintering temperature in both composites. The possible improvement of thermal conductivity in the diamond/50 vol%Ag composite is suggested by the lowering the sintering temperature based on the design of the additional second alloy, though it should be considered to improve slightly lower densification than that in diamond/50 vol%Al composite.

References

• 1)   B.-N.  Kim,  K.  Hiraga,  K.  Morita,  H.  Yoshida,  T.  Miyazaki,  Y.  Kagawa: “Microstructure and optical properties of transparent alumina”, Acta Materialia, 57 (2009) 1319–1326.
• 2)   Y.  Makino,  M.  Kawahara,  M.  Sakaguchi,  Y.  Kaiyama,  K.  Akamatsu,  T.  Yasuno: “Synthesis of Nano-structured Alumina with High Transparency by SPS Processing”, J. Jpn. Soc. Powder Powder Metallurgy, 59 (2012) 532–537.
• 3)   S.  Grasso,  B.-N.  Kim,  C.  Hu,  G.  Maiizza,  Y.  Sakka: “Highly Transparent Pure Alumina Fabricated by High-Pressure Spark Plasma Sintering”, J. Am. Ceram. Soc., 93[9] (2010) 2460–2462.
• 4)   Y.  Futami,  T.  Yanagida,  Y.  Fujimoto,  J.  Pejchal,  M.  Sugiyama,  S.  Kurosawa,  Y.  Yokota,  A.  Ito,  A.  Yoshikawa,  T.  Goto: “Optical and scintillation properties of Sc₂O₃, Y₂O₃ and Lu₂O₃ transparent ceramics synthesized by SPS method”, Radiation Measurements, 55 (2013) 136–140.
• 5)   M.  Tokita,  M.  Kawahara,  K.  Mizuuchi,  Y.  Makino: “Characterization of Zirconia(3Y)/ SUS410L Functionally-Graded Material Fabricated by SPS Processing”, J. Jpn. Soc. Powder Powder Metallurgy, 56 (2009) 383–388.
• 6)   W.  Que,  K. A.  Khor,  J. L.  Xu,  L. G.  Yu: “Hydroxyapatite/titania nanocomposites derived by combining high-energy ball milling with spark plasma sintering processes”, J. Euro. Ceram. Soc., 28 (2008) 3083–3090.
• 7)   F.  Watari,  A.  Yokoyama,  M.  Omori,  T.  Hirai,  H.  Kondo,  M.  Uo,  T.  Kawasaki: “Biocompatibility of materials and development to functionally graded implant for bio-medical application”, Composite Sci. & Technol., 64 (2004) 893–908.
• 8)   M.  Omori,  M.  Kawahara,  H.  Sakai,  A.  Okubo,  T.  Hirai: “Preparation of ZrO₂(3Y)SUS410 Functionally Gradient Material by Thermal Gradient Sintering”, J. Jpn. Soc. Powder Powder Metallurgy, 41 (1994) 649–652.
• 9)   K.  Mizuuchi,  K.  Inoue,  Y.  Agari,  Y.  Morisada,  M.  Sugioka,  M.  Tanaka,  T.  Takeuchi,  M.  Kawahara,  Y.  Makino: “Thermal conductivity of diamond particle dispersed aluminum matrix composites fabricated in solid-liquid co-existent state by SPS”, Composites: Part B42 (2011) 1029–1034.
• 10)   K.  Mizuuchi,  K.  Inoue,  Y.  Agari,  Y.  Morisada,  M.  Sugioka,  M.  Tanaka,  T.  Takeuchi,  J.  Tani,  M.  Kawahara,  Y.  Makino: “Processing of diamond particle dispersed aluminum matrix composites in continuous solid-liquid co-existent state by SPS and their thermal properties”, Composites: Part B42 (2011) 825–831.
• 11)   K.  Mizuuchi,  K.  Inoue,  Y.  Agari,  M.  Sugioka,  M.  Tanaka,  T.  Takeuchi,  M.  Kawahara,  Y.  Makino,  M.  Ito: “Processing of diamond-particle-dispersed silver-matrix composites in solid-liquid co-existent state by SPS and their thermal properties”, Composites: Part B43 (2012) 1445–1452.
• 12)   M. S.  Dresselhous,  M. A.  Pimenta,  P. C.  Eklund,  G.  Dresselhous: “Raman Scattering in Materials Science”, First Edition, Eds. by  W. H.  Weber and  R.  Merlin, Chap. 8, Springer (2000) 314–359.
• 13)   P. W.  Huong: “Structural studies of diamond films ultrahard materials by Raman and micro-Raman spectroscopies”, Diamond and Related Materials, 1 (1991) 33–41.