Summary

This chapter investigates the Cu nucleation on the TaN substrate with and without the C2H5I catalyst surface treatment. It was found that the C2H5I treatment enhanced the chemisorption of Cu-containing adspecies, and thus reduced the incubation time and accelerated the Cu film formation. For the Cu grains nucleated on the C₂H₅I-treated substrate, the wetting angle increased from 63^o to 94^o, implying that the C2H5I-treated substrate has a reduced surface energy and/or increased interfacial energy; this might promote the vertical growth of nucleated Cu grains, degraded the adhesion to substrate, and enhanced (200) orientation packed configuration.

Table 3-1 Major parameters and processing conditions of Cu CVD used in this study.

Cu CVD processing conditions

Substrate temperature (^oC) 160

Operation pressure (mTorr) 150

Cu precursor flow rate (ml/min) 0.4

Carrier gas (He) flow rate (sccm) 25 Substrate holder rotation speed (rpm) 10 Gas-injector/susceptor distance (cm) 2 CEM temperature (^oC) 70 Precursor’s delivery line temperature (^oC) 72 Reactor (reaction chamber) wall temperature (^oC) 45 C2H5I storage temperature (^oC) 25 Chamber pressure during catalyst treatment (mTorr) 150

(c) (b) (a)

Fig. 3-1 Top view SEM micrographs showing Cu grain nucleation on control TaN substrate

(a)

(b)

(c)

Fig. 3-2 Top view SEM micrographs showing Cu grain nucleation on C2H5I-treated TaN substrate for (a) 1 min, (b) 2 min, and (c) 3 min of Cu-CVD.

σ

_c

Substrate Cu grain σ

_i

θ σ

_s

θ : wetting angle (contact angle) σ

s

: surface energy of substrate σ

c

: surface energy of Cu grain

σ

i

: interfacial energy between Cu grain and substrate

cos

^{s i}

c

σ σ

θ σ

= −

Fig. 3-3 Schematic illustration of Cu nucleation and Young’s equation.

200nm (a)

θ～63

^o

(b)

θ～94

^o

Fig. 3-4 Oblique view SEM micrographs showing Cu grain nucleation for (a) 3 min on as-deposited and (b) 1 min on C2H5I-treated TaN substrates.

Chapter 4

CVD Cu Films on C

₂

H

₅

I –Treated TaN Substrate

4.1 Introduction

The properties of CVD Cu films are closely related to the initial nucleation of Cu on the substrates [26-28]. Variation of nucleation and growth mechanism of CVD Cu on different substrate conditions leads to different properties of the deposited Cu films. In chapter 3, we have found that C2H5I treatment on TaN substrate resulted in different Cu nucleation and film-forming process. In this chapter, properties of Cu films deposited on the C2H5I-treated TaN substrate, such as growth rate, surface morphology, film resistivity and film texture, are investigated.

4.2 Experimental Details

CVD Cu films were deposited on the same TaN surface of the substrate wafer TaN(25nm)/SiO₂(500nm)/Si, as was used in the study of chapter 3. The TaN substrate wafer (TaN/SiO2/Si) placed on the substrate holder was loaded into the load-locked chamber of the Cu CVD system. When the pressure of all relevant chambers was pumped down to 2×10^-6 torr, the substrate wafer was transferred to the reaction chamber via the transfer chamber. Then, helium (He) gas was introduced into the reaction chamber so as to maintain the chamber pressure at 150 mtorr, while the heating element was turned on to heat the substrate holder to the desired reaction temperature. As the temperature of the substrate holder reached the desired reaction temperature, which usually took

about one hour, the reaction chamber was pumped down to 10 mtorr. Then, the control valve to the catalyst container was opened to guide the vapor phase catalyst of C₂H₅I into the reaction chamber. As the chamber pressure rose to 1 torr in a short time, it was pumped down again and maintained at 150 mtorr until the valve was closed to turn off the catalyst supply. In this study for the effect of catalyst, the control valve to the catalyst container was opened for 3 min. Then, the chamber pressure was pumped down to 10 mtorr again, and He gas was introduced into the chamber. When the chamber pressure was stabilized at 150 mtorr, the control valve to the Cu precursor was opened, and the liquid precursor was propelled by the pushing nitrogen gas through the LFM. It was then vaporized in the CEM and mixed with the He carrier gas. The precursor-saturated carrier gas was introduced into the reaction chamber through the gas injector to proceed with the Cu CVD. In this study, Cu CVD was usually conducted for 10 min. The general processing conditions of Cu CVD for the study in this chapter are the same as those used in chapter 3, as shown in Table 3-1.

The thickness of Cu films was measured using a DekTek profiler and was further verified by cross-sectional scanning electron microscopy (SEM). SEM was also used to observe the surface morphology of the Cu films. A four-point probe was employed to measure the sheet resistance. The X-ray diffraction (XRD) analysis was used for phase identification. The surface roughness of the Cu films deposited was evaluated using atomic force microscopy (AFM).

4.3 Effect of TaN Substrate Treatment by C

₂

H

₅

I on CVD Cu Films

Figure 4-1 shows the effective deposition rate of Cu films on the TaN substrate with and without C2H5I treatment as a function of substrate temperature

(Arrhenius plot). The effective deposition rate was determined using the thickness of Cu films deposited in a duration of 10 min. Apparently the C2H5I treatment on TaN substrate resulted in increase of the effective deposition rate as well as the activation energy. This is presumably due to the shorter incubation time and the much easier nucleation of Cu on the C₂H₅I-treated TaN substrate.

Iodine atoms on the TaN substrate promote the dissociation of Cu species in the precursor and accelerate the reaction, resulting in the faster growth rate. The larger wetting angle of Cu grains nucleated on the C2H5I-treated TaN substrate, as shown in chapter 3, promotes the three-dimensional growth of Cu grains; this might also be a factor of faster Cu film growth rate. At temperatures above 180^oC, the Cu film deposited on the C₂H₅I-treated TaN substrate turned out to be porous, nonuniform, and poorly adhesive that the film easily peeled off. Figure 4-2 illustrates the SEM micrograph for the Cu film deposited on the C₂H₅I-treated TaN substrate at 200^oC, showing the nonuniform grain size and the plenty of voids in the Cu film. Figure 4-3 shows the resistivity of Cu films as a function of deposition temperature. The film resistivity is closely related to the impurity content and the film’s microstructure [29, 30]. For the films on the control TaN substrate, the slightly higher resistivity at temperatures below 160^oC is presumably due to higher contamination of residual impurities from the reaction by-products, while the higher resistivity at temperatures above 160^oC is presumably also due to higher contamination of impurities in the film in addition to the porous film structure [30]. The Cu films deposited on the C2H5I-treated TaN substrate exhibited lower film resistivity than those deposited on the control TaN substrate, presumably because the iodine atoms on the C2H5I-treated substrate weakened the [Cu⁺-(hfac)^-] ionic bond, promoting and accelerating the dissociation of the [Cu(hfac)] species that is adsorbed on the substrate surface

[13], thus the residual impurities from the reaction by-products had enough time to be pumped away from the film surface. Figure 4-4 and Fig. 4-5 illustrate the SEM micrographs showing the surface morphology of Cu films deposited on the control and C2H5I-treated TaN substrates, respectively. On the control substrate, although the Cu films deposited at higher temperatures generally have a larger grain size, they also contain larger voids and the grains appear to be loosely connected. On the other hand, it appears that the Cu films deposited on the C2H5I-treated substrate contain fewer voids in comparison with the corresponding Cu films deposited on the control substrate.

Figure 4-6 and Fig. 4-7 show the XRD spectra for the Cu films deposited on the control and C₂H₅I-treated TaN substrates, respectively, at various temperatures. Apparently, C2H5I-treatment on TaN substrate had an adverse effect on the growth of Cu film in (111) orientation, in particular at low deposition temperatures. Nonetheless, higher temperatures tended to promote the growth of (111) texture, either on the control or the C₂H₅I-treated TaN substrate.

Figure 4-8 and Fig 4-9 show the AFM images for the Cu films deposited on the control and C₂H₅I-treated TaN substrate, respectively. The RMS roughness of the Cu films deposited in a duration of 10 min on the C2H5I-treated TaN substrate is comparable with that deposited on the control TaN substrate. However, it should be noted that the thicknesses of the Cu films are widely different, although they were all deposited in a duration of 10 min.

The adhesion of Cu film to the underlayer substrate was evaluated by a simple Scotch tape pulling test. Table 4-1 shows the results of the test for the adhesion of Cu films to the TaN substrate with and without C2H5I treatment. The testing results for the samples of Cu film deposited on C₂H₅I-treated TaN substrate followed by thermal annealing at 400^oC for 30 min in N2 ambient, are

also included in the table for comparison. For the Cu films deposited at and above 160^oC on the control TaN substrate, they all passed the Scotch tape pulling test. However, the Cu films deposited on the C₂H₅I-treated TaN substrate all failed to pass the test. The large wetting angle of the nucleated Cu grains on the C₂H₅I-treated TaN substrate at the nucleation stage, because of the reduced substrate surface energy, might be responsible for the degraded adhesion.

However, thermal annealing (at 400^oC) resulted in slight improvement on the adhesion, although the improvement was not adequate to meet the requirement of practical applications.

4.4 Effect of Post-Deposition Thermal Annealing

Cu films deposited on the TaN substrates with/without C₂H₅I treatment were thermally annealed at 400^oC for 30 min in N2 ambient. It was found that the thermal annealing resulted in the decrease of film resistivity, as shown in Table 4-2, and a slight improvement on the adhesion of Cu film to the C2H5I-treated TaN substrate, as shown in Table 4-1. Figure 4-10 and Fig. 4-11 show, respectively, the surface morphology of 400^oC-annealed Cu films deposited on the control and C₂H₅I-treated TaN substrates at various temperatures. Compared with Fig. 4-4 and Fig. 4-5, which show the surface morphology of the as-deposited Cu films, it can be seen that the thermal annealing at 400^oC resulted in obvious change in surface morphology for all Cu films except the one deposited at 240^oC [Fig. 4-10 (f)], which was deposited in the flow-rate-controlled regime. The annealed Cu films exhibited milky or fluid-like surface, showing better connected Cu grains, diminished voids and smoother surface. The AFM images shown in Fig. 4-12 and Fig. 4-13 confirm the smoother

surface of the annealed Cu films. Figure 4-14 shows the improvement of the the as-deposited and 400^oC-annealed Cu films, respectively. Figure 4-15 and Fig.

4-16 illustrate, respectively, the XRD spectra for the 400^oC-annealed Cu films deposited on the control and C2H5I-treated TaN substrates at different temperatures. No obvious change in the XRD intensity peak ratio of Cu(111) to Cu(200) reflections was detected in comparison with that of the as-deposited Cu film (Fig.4-6 and Fig. 4-7).

4.5 Summary

This chapter explores the effects of TaN substrate pretreatment by C₂H₅I catalyst on the CVD-Cu films as well as the effect of post-deposition thermal annealing of Cu/TaN samples. The Cu film deposited on the C₂H₅I-treated TaN substrate exhibited higher deposition rate and lower electrical resistivity in comparison with that deposited on the control TaN substrate, presumably because the iodine atoms on the C2H5I-treated substrate surface promoted the dissociation of Cu species. However, it also enhanced the growth of (200) oriented texture and degraded the adhesion of the deposited Cu films to the TaN substrate. Thermal annealing at 400^oC resulted in obvious change of the surface morphology as well as reduced surface roughness. Moreover, thermal annealing also resulted in reduced film resistivity and slightly improved Cu film adhesion to C₂H₅I-treated TaN substrate.

Table 4-1 Results of Scotch tape pulling test on the adhesion of CVD Cu films to TaN substrate.

Deposition Temperature of Cu Film (^oC) Substrate

140 160 180 200 220 240

Control TaN F* P P P P P

C₂H₅I-treated TaN F F F

C2H5I-treated TaN (after 400^oC anneal) F* F* F*

P: Cu film passed the Scotch tape pulling test.

F: Cu film peeled off after Scotch tape pulling test.

F*: Cu film partially peeled off after Scotch tape pulling test.

Table 4-2 Effect of thermal annealing (400^oC in N2 for 30 min) on resistivity for Cu films deposited on TaN substrates with/without C2H5I treatment.

Resistivity (µΩ-cm)

Control substrate C2H5I-treated substrate Deposition

temperature

(^oC) as-deposited annealed as-deposited annealed 140 3.35 3.07 1.95 1.90 160 2.26 2.08 2.30 2.25 180 2.77 2.44 2.37 2.27

200 3.24 2.94

220 3.50 3.07

240 7.83 7.44

Effective deposition rate vs. substrate temperature (Arrhenius plot) for Cu films deposited at a pressure of 150 mtorr on TaN substrate with and without C2H5I treatment.

Fig. 4-1

Fig. 4-2 SEM micrograph showing surface morphology of Cu film deposited on C2H5I-treated TaN substrate at 200^oC.

Fig. 4-3 Resistivity vs. deposition temperature for Cu films deposited at a pressure of 150 mtorr on TaN substrate with and without C2H5I treatment.

Fig. 4-4

(e) (f) (c) (d)

(a) (b)

SEM micrographs showing surface morphology of Cu films deposited on control TaN substrate at (a) 140^oC, (b) 160^oC, (c) 180^oC, (d) 200^oC, (e) 220^oC, and (f) 240^oC.

Fig. 4-5

(c) (b) (a)

SEM micrographs showing surface morphology of Cu films deposited on C2H5I-treated TaN substrate at (a) 140^oC, (b) 160^oC, and (c) 180^oC.

Fig. 4-6 XRD spectra of Cu films deposited on control TaN substrate at different temperatures.

Fig. 4-7 XRD spectra of Cu films deposited on C2H5I-treated TaN substrate at different temperatures.

@ 160

^o

C

RMS = 23.670nm Thickness = 276nm

@ 140

^o

C

RMS = 19.599nm Thickness = 207nm

@ 180

^o

C

RMS = 30.746nm Thickness = 422nm

AFM image showing surface roughness for Cu films deposited in a duration of 10 min on control TaN substrate at different temperatures.

Fig. 4-8

Fig. 4-9

@ 180

^o

C

RMS = 30.636nm Thickness = 577nm

@ 160

^o

C

RMS = 25.132nm Thickness = 345nm

@ 140

^o

C

RMS = 21.119nm Thickness = 209nm

AFM image showing surface roughness for Cu films deposited in a duration of 10 min on C2H5I-treated TaN substrate at different temperatures.

Fig. 4-10

(e) (f) (c) (d)

(a) (b)

SEM micrographs showing surface morphology of 400^oC-annealed Cu films d on control TaN substrate at (a) 140

eposited

oC, (b) 160^oC, (c) 180^oC, (d) 200^oC, (e) 220^oC, and (f) 240^oC.

Fig. 4-11

(c) (b) (a)

SEM micrographs showing surface morphology of 400^oC-annealed Cu films deposited on C2H5I-treated TaN substrate at (a) 140^oC, (b) 160^oC, and (c) 180^oC

@ 180

^o

C

RMS = 25.68nm

@ 140

^o

C

RMS = 18.53nm

@ 160

^o

C

RMS = 20.47nm

Fig. 4-12 AFM image showing surface roughness of 400^oC-annealed Cu films deposited on control TaN substrate at different temperatures.

@ 180

^o

C

RMS = 25.57nm

@ 160

^o

C

RMS = 20.43nm

@ 140

^o

C

RMS = 15.13nm

Fig. 4-13 AFM image showing surface roughness of 400^oC-annealed Cu films deposited on C2H5I-treated TaN substrate at different temperatures.

Fig. 4-14 Improvement of surface roughness vs. deposition temperature for Cu films deposited on TaN substrate with and without C2H5I treatment after 400^oC thermal annealing.

Fig. 4-15 XRD spectra of 400^oC-annealed Cu films deposited on control TaN substrate at different temperatures.

Fig. 4-16 XRD spectra of 400^oC-annealed Cu films deposited on C2H5I-treated TaN substrate at different temperatures.

Chapter 5

Via-Filling of Cu CVD on Via-Patterned Substrate with TaN Liner

5.1 Introduction

The Cu damascene scheme has been developed to implement the multilevel interconnect of Cu lines in integrated circuits. The Cu ECD combined with the ionized metal plasma (IMP) PVD of a thin Cu seed layer and barrier layer provides a viable solution for IC technologies above 0.13µm; however, the technique of this method will face more stringent challenge because of the requirement of a more stringent conformal and continuous thin barrier as well as conformal and void-free Cu film filling into deep sub-quarter micron vias for the IC technology beyond 0.13µm [32]. Although the traditional Cu CVD method easily results in a seam inside the vias and trenches due to the conformal deposition, it has been reported that vias/trenches could be filled with Cu in a bottom-up fashion (superfilling), resulting in a seam-free filling [13, 14]. This technique used iodine as catalytic surfactant to carry out the bottom-up superfilling. In this chapter, we investigate the via-filling behavior of Cu CVD using iodoethane (C2H5I) as catalyst surfactant.

5.2 Experimental Details

In this study, 25 nm-thick TaN layers were used as the diffusion barrier for the CVD of Cu films. To evaluate the via-filling capability of Cu CVD, two sets of vias of different feature sizes were patterned on 1-µm-thick fluorine doped

silicon glass (FSG) covered Si wafer. An IMP sputtering system was used to deposit conformal TaN layers of 25 nm thickness into the high aspect ratio (AR) vias. The diameters of the vias are 0.16 and 0.25µm. In the real pattern, the patterned and measured error will be about several nanometers. The general processing conditions of Cu CVD for the study in this chapter are the same as those used in chapter 3, as shown in Table 3-1, except the deposition pressure and the flow rate of carrier gas (helium). The deposition pressures used for via-filling of Cu film were 60 and 150 mtorr. In order to pump the pressure in the reaction chamber down to 60 mtorr, the flow rate of carrier gas was reduced to 13 sccm while the liquid precursor flow rate was maintained at 0.4 ml/min. For the Cu CVD run at 150 mtorr, the flow rate of He carrier gas was 25 sccm and the Cu precursor flow rate was maintained at the same value of 0.4 ml/min.

5.3 Via-Filling on Via-Patterned Substrate with TaN Liner

Via-filling behavior of Cu CVD was investigated at temperatures ranging from 140 to 200^oC. At temperatures above 200^oC, the Cu film deposition belongs to the mass-flow-controlled mechanism, and it is disadvantageous to via-filling [33]. Post via-filling thermal annealing was performed at 400^oC for 30 min in N₂ ambient to improve the adhesion of Cu films to the TaN barrier layer and to eliminate the microvoids between Cu grains in the vias. Figure 5-1 illustrates the SEM micrographs showing cross-sectional views of via-filling for the 0.25-µm-diameter vias under 150 mtorr at various temperatures. It can be seen that at the deposition temperature of 160^oC, the bottom of the vias can be filled.

At the higher deposition temperatures of 180 and 200^oC, presumably the growth rate of Cu film became too fast such that the top of the via was quickly obstructed

by the Cu film, resulting in failure of via-filling. For the case of via-filling at 140^oC, it can be seen that Cu was able to reach the bottom of the via; however, the via-filling was apparently incomplete for the 10 min run presumably because of low deposition rate and long incubation time at the low deposition temperature of 140^oC. Figure 5-2 illustrates the cross-sectional view SEM micrographs for the via-filling of the 0.16-µm-diameter vias under 150 mtorr at various temperatures.

At the higher temperature of 180 and 200^oC, Cu was not able to deposite inside these smaller vias; even at the lower temperatures of 160 and 140^oC, Cu was not able to fill the vias at the bottom. It was reported that for the Cu CVD at high temperatures and high pressures, the reactive sticking coefficient of the precursor species is presumably too high such that the Cu containing species cannot reach the bottom of the smaller vias [31], such as the via-filling with Cu-CVD at 180 and 200^oC and 150 mtorr studied in this work. At the lower temperatures of 160 and 140^oC, the results of via-filling were apparently improved to some extent.

Figure 5-3 shows the cross-sectional view SEM micrographs for the via-filling of the 0.16-µm-diameter vias under 60 mtorr at 140 and 160^oC. By comparing the results of via-filling shown in Fig. 5-2 and Fig. 5-3, we found clearly that reducing the deposition pressure (from 150 to 60 mtorr) and increasing the Cu precursor concentration in the carrier gas (from He flow rate of 25 to 13 sccm mixed with the same liquid Cu precursor of 0.4 ml/min) improved the via-filling capability, particularly at the lower deposition temperature of 140^oC. From these experimental results, we may conclude that low pressure, low temperature, and high concentration of Cu containing species in the gas phase of CVD facilitate the via-filling.

5.4 Effect of C

₂

H

₅

I treatment on Via-Filling

Figure 5-4 illustrates the cross-sectional view SEM micrographs for the via-filling of the 0.16-µm-diameter vias under 150 mtorr at various temperatures.

The C₂H₅I treatment on the patterned substrate did improve the via-filling capability, as evidenced by comparing Fig. 5-4 and Fig. 5-2; the most obvious improvement can be found for the via-filling at the deposition temperature of 140^oC. Figure 5-6 shows the cross-sectional view SEM micrographs for the via-filling of the 0.16-µm-diameter vias under 60 mtorr at 140 and 160^oC. Better result was obtained by using lower pressure (60 mtorr) and lower temperature (140^oC) deposition.

5.5 Summary

Via-filling capability of Cu CVD on 0.16-µm and 0.25-µm vias with TaN

Via-filling capability of Cu CVD on 0.16-µm and 0.25-µm vias with TaN

在文檔中 碘乙烷催化劑對銅化學氣相沉積特性之影響 (頁 38-80)