The films obtained at a precursor ratio of 1:1 were rich in both copper and nickel colors (whitish copper); the films obtained at 5:1 possess a slight metallic shiny
copper color, whereas the films obtained at 1:5 were rich in shiny nickel color. These visual observations agree with my expectation derived from the Cu/Ni concentration ratios.
The Auger depth profiles of the Cu-Ni films deposited at 240ºC with different precursor ratios are shown in Fig. 2.1. The Auger depth profile for a precursor ratio of 1:1 [Fig. 2.1(a)] shows that the surface is Ni-rich and becomes Cu-rich at the interface.
Figure 2.1(b) shows the Auger depth profiles of a Cu-Ni film with a precursor ratio of 5:1. It was found that the surface is Ni-rich and the interface was enriched with Cu. It was also found that both Cu and Ni deposited at the same time between the surface and the interface. For the precursor ratio of 1:5 [Fig. 2.1(c)], the film surface was enriched with only Ni, and near the interface, Cu was enriched. Both F and O impurities were below the detection limit. It is interesting to note that Cu-Ni solid solutions did not grow. Cu deposited initially on the substrate, then Ni started to grow.
One reason is that the decomposition temperature of Cu is lower than that of Ni [29], and thus, the Ni nucleation preceded the growth of Cu. Another reason is that Cu and Ni are immiscible at a lower temperature (240ºC) and diffusion is slow, which agree with the phase diagram of Cu-Ni [9]. This tendency is in accordance with a previous study on Cu-Ag formation [3].
Although the films had a bilayer structure, the average Cu-Ni composition ratios of the films fairly agreed with the precursor mixing ratios. Figure 2.2 shows the Cu/Ni molar ratios for different precursor compositions of 1:1, 5:1, and 1:5 determined by EDS. The deposition temperature was 240ºC. Quantitative compositional analyses of the films show that Cu and Ni deposits are only slightly less than the actual precursor ratio, whereas a very good linearity is observed.
To make an alloy or solid solution of Cu and Ni, annealing at a higher temperature was performed. The depth profiles of the Cu-Ni film with a precursor ratio of 1:1 after annealing at 300°C for 30 min in an Ar environment. Figure 2.3 show a change in the concentration slopes at the position of the Cu-Ni interface, and a considerable amount of Ni diffuses into Cu, and Cu diffuses into Ni, showing that quasi-mixing occurred.
The depth profile of Cu-Ni was also confirmed that diffusion of Cu in Ni is less than the diffusion of Ni in Cu.
Fig. 2.1 (a) AES depth profiles of a Cu-Ni film with a precursor composition of 1:1.
Fig. 2.1(b) AES depth profiles of a Cu-Ni film with a precursor composition of 5:1.
Intensity (arb. unit)
Etching cycle
0 20 40 60 80 100 120
Cu Ni Ti N F O (a)
Intensity (arb. unit)
Etching cycle
0 10 20 30 40 50 60
Cu Ni Ti N F O (b)
Fig. 2.1(c) AES depth profiles of a Cu-Ni film with a precursor composition of 1:5.
Fig. 2.2 EDS Cu/Ni molar ratio of films with different precursor Cu/Ni molar ratios.
0 20 40 60 80 100
Cu Ni Ti N F O
Intensity (arb. unit)
Etching cycle
(c)
0 1 2 3 4 5 6
0 1 2 3 4 5 6
EDS C/Ni molar ratio
Precursor Cu/Ni molar ratio
Whereas atomic diffusion takes place across the Cu-Ni interface, the mixing was not completed at this annealing temperature of 300 ºC.
Fig. 2.3 AES depth profiles of ex situ annealed (at 300 ºC for 30 min) Cu-Ni film with a precursor composition of 1:1.
Fig. 2.4 (a) Surface and (b) cross-sectional SEM images of Cu-Ni films with a precursor composition of 1:1. The deposition temperature was 240ºC.
The surface morphology of Cu-Ni films obtained at a precursor ratio of 1:1 is shown in Fig. 2.4. It is evident that the film obtained by SFCD exhibits uniform growth throughout the film surface [Fig. 2.4(a)]. A cross-sectional SEM image [Fig. 2.4 (b)]
revealed that the Cu and Ni were deposited as separate films as confirmed by AES
0 20 40 60 80 100 120 140
Cu Ni Ti N F O
Intensity (arb. unit)
Etching cycle
(a) (b)
analyses. However, for the other precursor concentrations such as 5:1 and 1:5, the films were also bilayers (data not shown) but the thicknesses were different.
Figure 2.5 shows the ex situ X-ray diffraction patterns of Cu-Ni films with a precursor ratio 1:1 deposited at different temperatures (200−240ºC). Both Cu and Ni peaks were observed. However, at a lower temperature of 200ºC, the Cu peaks were predominant, and above this temperature, the Ni peaks become predominant; also, the peak intensity increases at 240ºC. Again, it is said that the Cu precursor decomposes faster than the Ni precursor in the scCO2 environment and decomposes predominantly at lower temperatures. As the temperature increases, both precursors were decomposed.
Fig. 2.5 Ex situ X-ray diffraction patterns of a Cu-Ni film with a precursor composition of 1:1 at different temperatures.
As it was seen in Fig. 2.3, Cu and Ni form alloys at an elevated temperature. To observe the temperature dependence of the Cu-Ni alloy formation, an in situ X-ray diffraction measurement was performed. Figure 2.6(a) shows X-ray diffraction patterns of a Cu-Ni film with a precursor composition of 1:1 during in situ vacuum annealing. The ramp rate was 10ºC/min and annealing was performed in an Ar atmosphere. Peaks corresponding to Cu(111) and Ni(111) planes were clearly observed. With increasing annealing temperature, shifts in the Cu peak position to a larger 2θ angle and in the Ni peak position to a smaller 2θ angle were observed. The
Intensity (arb. unit)
42 44 46 48 50 52 54 56
Ni(111) Cu(111)
Cu(200)
Ni(200)
200 oC 240 oC
220 oC
200°C 220°C 240°C
2Theta (deg.)
Ni peak gradually disappeared above 300 ºC. These results demonstrate the formation of a single phase. The critical solution temperature of Cu-Ni is 354.5ºC [9], and the decrease in the Ni intensity above 300 ºC agrees with this temperature. The Cu-Ni film with a precursor composition of 1:1 deposited at 240 ºC after annealing at higher temperature indicates the formation of a Cu-rich solid solution, owing to the intermixing enhancements of Cu and Ni.
Fig. 2.6(a) In situ X-ray diffraction patterns of a Cu-Ni film with a precursor composition of 1:1 at different vacuum annealing temperatures.
The intensity ratio between the Cu(111) and Ni(111) peaks gradually increases with annealing temperature [Fig. 2.6 (b)], which indicates that the diffusion was much faster at higher annealing temperatures, leading to larger grains, which further increases the Cu peaks. Obviously, more work should be conducted to elucidate the alloying mechanism to produce alloys in a more controlled manner.
500ºC
Intensity (arb. unit)
2Theta (deg.) 300ºC
100ºC 700ºC 1000ºC
Fig. 2.6 (b) Intensity ratios between Cu (111) and Ni (111) peaks estimated from (a).
It has been known that the adhesion strength (9.4 mN) at the Cu/substrate is weak when Cu is directly deposited on the substrate by SFCD [30]. The adhesion strength of the Cu-Ni film stack with a precursor composition of 1:1 was measured using a scratch tester, and the average adhesion strength was found to be 45.5 mN, which is higher than that of the Cu/substrate. Therefore, adhesion was improved upon alloying.
The sheet resistances of the Cu-Ni film stacks were measured by the four-point method. The resistivities of Cu-Ni films as-deposited at 240ºC and vacuum-annealed with a precursor composition of 1:1 were found to be 7.40 and 4.55 μΩ cm, respectively. Note that the resistivities of pure Cu and pure Ni are 2.1 and 13.3 μΩ cm, respectively [14,31].
One advantage of SFCD is the ability to grow films with conformability and filling capability at a high aspect ratio (HAR) through holes and trenches. Figure 2.7(a) shows a cross-sectional scanning electron microscopy image taken after a trial filling into 250-nm-wide narrow features with a Cu-Ni precursor composition of 1:1 on a TiN-coated SiO2/Si trench substrate. The deposition temperature was 200ºC. A good filling in HAR features is clearly demonstrated.
ICu(111)/INi(111)
0 5 10 15 20
0 100 200 300 400 500 600
Temperature (°C)
Compared with the Cu filling in 250-nm-wide trenches published in the literature [Fig.
2.7(b)] at the same deposition temperature [30], it is evident that Ni did not affect the filling of nanotrenches with Cu-Ni.
Fig. 2.7 SEM images of 250-nm-wide trenches filled with (a) Cu-Ni at a precursor composition of 1:1 and (b) only Cu at a deposition temperature of 200ºC.
(a) (b)