Reduction-assisted sintering of micron-sized copper powders at low temperature by ethanol vapor

Abstract

The low temperature sintering of micron-sized Cu powders is achieved by ethanol vapor annealing. A Cu-pancake is formed and has enough mechanical strength to sustain the gravitational pull. The electrical resistivity of the Cu-pancake formed by flaky powders is lower than that by spheroidal ones because more contact area of the former facilitates the sintering process. The resistivity of the Cu-pancake grows with decreasing the annealing temperature but it is still about 10⁻³ Ω cm at 120 °C. The sintered Cu-pancake is characterized by X-ray diffraction and X-ray photoelectron spectroscopy to investigate the sintering mechanism. The low temperature sintering is attributed to the reduction of the native oxide on surfaces of Cu powders by the ethanol vapor. The reduced Cu is very active and tends to sinter with each other to lower the surface energy. This reduction-assisted sintering may be useful in the fabrication of conductive patterns on flexible substrates. The prepared Cu pattern on polyethylene naphthalate exhibits repeatable flexibility and acceptable conductivity.

---

1. Introduction

Recently, the development of copper (Cu) conductive lines has attracted a lot of attention due to its relatively low cost compared to silver. Because of the growing demand for flexible electronics, methods such as ink-jet printing^1–5 and screen printing^6–8 have been investigated. In these approaches, Cu particles dispersed in droplets are printed on the polymer-based substrate. In order to obtain a better electrical property of the Cu conductive line, sintering of deposited Cu particles is required for the formation of Cu interconnect. Owing to the poor thermal stability of the polymer-based substrate, the sintering process has to be conducted at low temperature (typically below 150 °C for poly(ethylene terephthalate)).

Sintering involves fusion of powders to form a more dense mass by heat or pressure. In general, a sintering process is conducted at about 2/3 of the melting temperature of the powder materials.⁹ Particles are fused together because of the diffusion of the atoms across the boundaries between touched particles. Necks are formed along the contact regions and a grain boundary develops within each neck. Due to the increment of the continuous solid phase, various properties such as mechanical strength, electrical conductivity and thermal conductivity are enhanced after sintering. Cu sintering is generally controlled by the surface and grain boundary diffusion, which vary with the temperature significantly.^10–12 As a result, high temperature treatment is usually employed for Cu sintering but it is unsuitable for polymer-based substrates associated with flexible electronics. Low temperature sintering technologies are thereby developed, such as the local laser sintering with sophisticated 3D gantry system,¹³ spark plasma sintering,^14–16 microwaves sintering,¹⁷ and intense pulsed light sintering.¹⁸ Nevertheless, these methods are expensive, complicated, and difficult to be used for mass production.

Two key factors influencing the sintering temperature are the oxidation degree and particle size of the metal powders. The presence of metal oxide on the surfaces of the metal powders not only lowers the electrical conductivity but also elevates the required sintering temperature.¹⁹ The effect of the particle size on the sintering temperature is associated with the dependence of the melting temperature on the particle size. It is known that the melting temperature of metal nanoparticles declines as their size decreases.²⁰ This fact reveals that the sintering temperature can be reduced by decreasing the particle size. As the size of Cu particles is greater than 30 nm, the melting temperature is close to that of the bulk material. Only when the particles size is smaller than 10 nm, the melting temperature of Cu is significantly lowered from 1084 °C to 900 °C. It seems that Cu nanoparticles can be used to resolve the problem of low temperature sintering. Unfortunately, Cu nanoparticles have very poor oxidative resistance and thus their sintering at low temperature cannot be achieved.^21–23

In this study, low temperature sintering of micron-sized Cu powders which can resist oxidation is explored. The Cu patterns are reduced and sintered by ethanol vapor at various temperatures (120–220 °C). The electrical conductivity, surface morphology, and chemical composition are analyzed by four-point probe measurement, scanning electron microscopy (SEM), X-ray diffraction (XRD), and X-ray photoelectron spectroscopy (XPS).

2. Materials and experimental methods

The flaky and spheroidal Cu powders (99%, USA) are purchased from Sigma-Aldrich. The anhydrous ethanol (99.5%, Japan) is obtained from Showa. All chemicals in this work are analytical grade and used as received without further purification. The glass dishes are the products of Kwo-Yi Co. (Taiwan). The polyethylene naphthalate (PEN) films are purchased from Teijin DuPont Films (Japan).

The Cu-pancake can be made from Cu powders at low temperatures. The Cu powders are uniformly placed on a glass dish and then preheated at a specific temperature for 20 min. Afterward, the sample and a glass cup filled with ethanol (30 ml) are set on a hot plate and covered in a glass hood. The apparatus is shown in Fig. 1. The sample is annealed in ethanol vapor until ethanol liquid is evaporated completely. Typically, it takes about 5–20 min, depending on the annealing temperature. The pattern of Cu powders on a PEN substrate is made. It is achieved simply by placing the drops of the Cu powder suspension on the pattern created by the polyimide tape on the substrate. After the patterned Cu is dried, it is sintered by ethanol vapor and the tape is removed later on.

Fig. 1 Schematic diagram of the apparatus for ethanol-assisted sintering. The copper powders are annealed on the hot plate in ethanol vapor at low temperature.

The electrical resistivity of the Cu-pancake is measured by a Keithley 2400 source/meter with a four-point probe measurement to avoid the error from contact resistance. X-Ray diffraction (XRD) measurements are performed using a BRUKER D8AXRD diffractometer with Cu K_α radiation. The X-ray photoelectron spectroscopy (XPS) is conducted using a Sigma Probe spectrometer (Thermo VG Scientific Co. Ltd.) equipped with monochromatic aluminum anode X-ray source (Al K_α 1486.6 eV). The surface morphology of the Cu powders is observed via the field-emission scanning electron microscopy (FE-SEM, JEOL JSM-6500F). Prior to the XRD and XPS analyses, the samples before and after sintering are dried in vacuum for about 12 h.

3. Results and discussion

Fig. 2(a) shows the flaky Cu powders in a glass dish before sintering. After in the environment of ethanol vapor at 120 °C for about 1 h, the flaky Cu powders exhibit the shiny, lustrous appearance of metal, as demonstrated in Fig. 2(b). By shaking the dish, the Cu powders can be detached as a whole and seem like a pancake, indicating sintering of micron-sized Cu powders. This Cu-pancake with radius about 2.5 cm can be picked up by a stick easily, as illustrated in Fig. 2(b). This result shows that the free-standing Cu-pancake possesses enough mechanical strength to resist the gravitational pull and also reveals the formation of Cu interconnection associated with sintering.

Fig. 2 (a) The photos of flaky Cu powders before sintering. (b) The photos of Cu-pancakes made by flaky Cu powders after annealing in ethanol vapor at 120 °C.

3.1 Electrical resistivity: annealing temperature and particle shape

The electrical resistivity of Cu powders before the ethanol vapor treatment has been measured and its value is about 10⁵ Ω cm, indicating a semiconductor-like behavior. However, after alcohol-assisted annealing, the electrical resistivity drops drastically and a conductor-like behavior appears. Fig. 3 shows the variation of the electrical resistivity with the annealing temperature for flaky and spheroidal Cu powders. In general, the electrical resistivity decreases as the annealing temperature is increased. Evidently, the resistivity of flaky Cu powders is significantly less than that of spheroidal ones under the same annealing temperature. It should be noted that as the annealing temperature is less than 160 °C, the resistivity of spheroidal Cu powders is very large and the Cu-pancake cannot be formed. However, even at 120 °C, the resistivity of flaky Cu powders can be as small as 10⁻³ Ω cm and the Cu-pancake is obtained.

Fig. 3 The electrical resistivity of spheroidal and flaky copper powders after annealing in ethanol vapor at various temperatures.

According to Fig. 3, a tremendous drop of the resistivity (about eight order of magnitude) occurs at about 120 °C for flaky powders and 160 °C for spheroidal ones. The decrease of the resistivity is still significant (about one order of magnitude) from 120 to 140 °C for flaky powders and from 160 to 180 °C for spheroidal ones. As the annealing temperature is further increased, the decrement of the resistivity becomes gradual. The great decline in the resistivity corresponds to the formation of Cu-pancake, which takes place at the temperature above 120 °C for flaky powders and 160 °C for spheroidal ones. Evidently, these critical temperatures associated with the Cu-pancake formation can be regarded as the sintering temperature, below which the electrical conductivity is extremely small. Note that good adhesion between powder particles is unable to provide good electrical conductivity. When the annealing temperature is slightly higher than the sintering temperature (within 20 °C), the degree of sintering can still be elevated significantly and thereby the conductivity rises accordingly. As the annealing temperature is well above the sintering temperature, the variation of the conductivity with temperature is just like that of metals. Clearly, the sintering temperature of flaky powder is lower than that of spheroidal ones. Although the volumes of the two types of powders are essentially the same, the flaky powder has the thickness (about 1 μm) smaller than the diameter of spheroidal powder (about 15 μm). In addition, the contact area of flaky Cu powders with face-to-face arrangement is greater than that of spheroidal powders, facilitating the diffusion of Cu atoms in the sintering process.

3.2 Reduction-assisted sintering

In general, the surface of Cu is oxidized very easily in ambient condition. Because of the high electrical resistance associated with copper oxides on the Cu surface, the overall resistivity of Cu powders is relatively high. In addition, the presence of the thin layer of copper oxides also significantly hinders sintering among micron-sized Cu particles. As a result, it is anticipated that the removal of this copper oxide layer will facilitate both electrical conductivity and sintering. The success of this low temperature annealing process in ethanol vapor can be attributed to the reduction of Cu powders by ethanol. The redox reaction is given by

6CuO(s) + C₂H₅OH(g) → 6Cu(s) + 3H₂O(g) + 2CO₂(g)

During the annealing process in which the temperature is higher than the boiling point of ethanol (78.4 °C), the ethanol vapor reduces the native copper oxide on the surface of Cu powders continuously. The reduced, outermost Cu layer on a particle is very active and tends to merge together with another Cu particle so that the total surface free energy is lowered.

In order to examine the sintering mechanism associated with the reduction of copper oxides, the chemical compositions of Cu powders and Cu-pancake are characterized by XRD and XPS. As shown in Fig. 4(a), presence of a small amount of copper oxide in Cu powders is demonstrated by the peaks, (110), (002), and (113). However, these peaks vanish for the Cu-pancake, revealing the reduction of CuO by the ethanol vapor. The evidence of copper oxide in Cu powders is also provided by XPS spectrum as illustrated in Fig. 4(b). After the annealing process, the satellite peaks of the Cu-pancake are significantly weakened. Moreover, the Cu 2p_3/2 binding energy is right-shift from 934 to 932 eV. Both changes indicate the decrease of the amount of CuO due to the reduction and sintering processes.

Fig. 4 (a) The XRD patterns of flaky Cu powders before and after sintering by ethanol vapor. (b) The XPS patterns of flaky Cu powders before and after sintering by ethanol vapor. “S” means the Cu²⁺ satellite peak which can be evidently observed in copper(II) oxide.

3.3 Surface morphology of Cu-pancake

The Cu-pancake is formed at the annealing temperature much lower than the typical sintering temperature. As a result, it is anticipated that the structure of Cu-pancake is different from that sintered at high temperature. Fig. 5 demonstrates the SEM images of surface morphology of Cu particles and Cu-pancakes. The particle sizes of spheroidal and flaky Cu powders are about 15 and 50 μm, respectively. The fusion of spheroidal Cu particles at 180 °C is evidently observed but the shapes of these particles are still shown. Similar phenomena are also seen for flaky Cu particles sintered at 120 °C. Note that the formations of necks between particles in the typical sintering process cannot be clearly identified. Nonetheless, the morphology of the flaky Cu powders illustrated in Fig. 5(f) for sintering at 120 °C shows differences. For comparison, the SEM images after sintering at 180 °C and 220 °C are provided in Fig. 6. The interconnection between flaky Cu powders can be obviously seen as the magnification exceeds 10 000 (Fig. 6(f)).

Fig. 5 The SEM images of (a) spheroidal Cu powders before sintering, (b) Cu-pancake formed from spheroidal powders at 180 °C with magnification of 1000× and (c) 3000×, (d) flaky Cu powders before sintering, (e) Cu-pancake formed from flaky powders at 120 °C with magnification of 1000× and (f) 4500×.

Fig. 6 The SEM images of Cu-pancake formed from flaky powders at 180 °C with magnification of (a) 1000×, (b) 3000×, and (c) 20 000× and at 220 °C with magnification of (d) 1000×, (e) 3000×, and (f) 10 000×.

Fig. 7 The design patterns of flaky Cu powders on PEN substrate.

3.4 Flexible conductive patterns

Our reduction-assisted sintering approach can be carried out at the annealing temperature as low as 120 °C. Consequently, it may be useful in the production of Cu conductive lines onto flexible polymer substrates for flexible electronics. Fig. 7 demonstrates a simple conductive Cu pattern on a polymer substrate of PEN with the deflection temperature about 155 °C.²⁴ The PEN substrate remains intact after sintering at 120 °C. The electrical resistivity of the Cu pattern is about 10⁻³ Ω cm. The conductive pattern on PEN with the thickness 50 μm is easily flexed by tweezers.

4. Conclusions

The sintering of micron-sized Cu powders by ethanol vapor is achieved at low temperatures. After the ethanol-assisted sintering process, the Cu-pancake is formed and exhibits the mechanical strength strong enough to resist the gravitational pull. At the same annealing temperature, the electrical resistivity of Cu-pancake formed by flaky powders is lower than that by spheroidal ones. It is because the former has more contact area due to the face-to-face arrangement, which is helpful in the diffusion of Cu atoms associated with the sintering process. As the annealing temperature is decreased, the electrical resistivity of Cu-pancake increases but it is still about 10⁻³ Ω cm even when the temperature is as low as 120 °C. The mechanism of this low temperature sintering is the reduction of the outermost, native oxide on Cu powders by the ethanol vapor. The fresh surface of reduced Cu powders is very active and tends to sinter with each other in order to lower the total surface energy. The reduction reaction is verified by the XRD and XPS analyses. Owing to the low annealing temperature, the shape of Cu powder is still shown after sintering and the formation of necks is not seen. This reduction-assisted sintering may be used for the fabrication of conductive lines on various flexible substrates. The conductive pattern on PEN possesses good flexibility and electrical conductivity.

Acknowledgements

This research work is supported by Ministry of Science and Technology of Taiwan and Industrial Technology Research Institute of Taiwan.

References

1. B. K. Park, D. Kim, S. Jeong, J. Moon and J. Kim, Thin Solid Films, 2007, 515, 7706–7711.
2. S. Jeong, K. Woo, D. Kim, S. Lim, J. S. Kim, H. Shin, Y. Xia and J. Moon, Adv. Funct. Mater., 2008, 18, 679–686.
3. Y.-I. Lee and Y.-H. Choa, J. Mater. Chem., 2012, 22, 12517–12522.
4. D. Li, D. Sutton, A. Burgess, D. Graham and P. D. Calvert, J. Mater. Chem., 2009, 19, 3719–3724.
5. S. Jeong, H. C. Song, W. W. Lee, S. S. Lee, Y. Choi, W. Son, E. D. Kim, C. H. Paik, S. H. Oh and B.-H. Ryu, Langmuir, 2011, 27, 3144–3149.
6. Y. N. Xia and G. M. Whitesides, Annu. Rev. Mater. Sci., 1998, 28, 153–184.
7. D. Qin, Y. N. Xia and G. M. Whitesides, Nat. Protoc., 2010, 5, 491–512.
8. F. C. Krebs, Sol. Energy Mater. Sol. Cells, 2009, 93, 394–412.
9. M. N. Rahaman, Sintering of ceramics, SRC Press, Boca Raton, 2008, pp. 45–176.
10. P. Zeng, S. Zajac, P. Clapp and J. A. Rifkin, Mater. Sci. Eng., A, 1998, 252, 301–306.
11. F. Wakai, J. Am. Ceram. Soc., 2006, 89, 1471–1484.
12. P. Song and D. Wen, J. Nanopart. Res., 2010, 12, 823–829.
13. S. H. Ko, H. Pan, C. P. Grigoropoulos, C. K. Luscombe, M. J. Frechet and D. Poulikakos, Appl. Phys. Lett., 2007, 90, 141103.
14. D. M. Hulbert, A. Anders, D. V. Dudina, J. Andersson, D. Jiang, C. Unuvar, U. Anselmi-Tamburini, E. J. Lavernia and A. K. Mukherjee, J. Appl. Phys., 2008, 104, 3305.
15. O. Guillon, J. Gonzalez-Julian, B. Dargatz, T. Kessel, G. Schierning, J. Räthel and M. Herrmann, Adv. Eng. Mater., 2014, 16, 830–849.
16. R. Castro and K. van Benthem, Sintering: Mechanisms of convention nanodensification and field assisted processes, Springer Science & Business Media, 2012.
17. J. Perelaer, R. Abbel, S. Wunscher, R. Jani, T. van Lammeren and U. S. Schubert, Adv. Mater., 2012, 24, 2620–2625.
18. H.-S. Kim, S. R. Dhage, D.-E. Shim, H. T. Hahn and H. Thomas, Appl. Phys. A, 2009, 97, 791–798.
19. S. Magdassi, M. Grouchko and A. Kamyshny, Materials, 2010, 3, 4626–4638.
20. P. A. Buffat and J. P. Borel, Phys. Rev. A, 1976, 13, 2287–2298.
21. C.-J. Wu, S.-M. Chen, Y.-J. Sheng and H.-K. Tsao, J. Taiwan Inst. Chem. Eng., 2014, 45, 2719–2724.
22. S.-H. Tu, H.-C. Wu, C.-J. Wu, S.-L. Cheng, Y.-J. Sheng and H.-K. Tsao, Appl. Surf. Sci., 2014, 316, 88–92.
23. F.-M. Chang, S.-L. Cheng, S.-J. Hong, Y.-J. Sheng and H.-K. Tsao, Appl. Phys. Lett., 2010, 96, 114101.
24. http://www.teijindupontfilms.jp/english/product/pen_teo.html.

---