Interfacial reactions and electrical properties of hafnium-based thin films in Cu/barrier/n+-p junction diodes

(1)

Interfacial reactions and electrical properties of

hafnium-based thin ﬁlms in Cu/barrier/n

+

–p junction diodes

Keng-Liang Ou

a

, Ming-Hung Tsai

b

, Haw-Ming Huang

a

, Shi-Yung Chiou

c

,

Che-Tong Lin

d

, Sheng-Yang Lee

d,e,*

a

Graduate Institute of Oral Sciences, College of Oral Medicine, Taipei Medical University, Taipei 110, Taiwan, ROC b

Department of Dentistry, En Chu Kong Hospital, Taipei Hsien 237, Taiwan, ROC c

Department of Mold and Die Engineering, National Kaohsiung University of Applied Sciences, Kaohsiung 807, Taiwan, ROC d

School of Dentistry, College of Oral Medicine, Taipei Medical University, Taipei 110, Taiwan, ROC e

Dental Department of Wan-Fang Hospital, Taipei Medical University, Taipei 110, Taiwan, ROC Received 2 September 2004; accepted 26 October 2004

Available online 18 November 2004

Abstract

In this study, the barrier properties of Hf and nitrogen incorporated Hf films were investigated by Cu/Hf–N/Si struc-ture. Hafnium and hafnium nitride films were prepared by reactive rf-magnetron sputtering on blank silicon wafers. The barrier properties were evaluated by sheet resistance, X-ray diffraction, transmission electron microscopy, scanning electron microscopy, and X-ray photoelectron spectroscopy. The as-deposited Hf film has a hexagonal close packed structure and a low resistivity of 100.98 lX-cm. With increasing nitrogen concentration of Hf–N film, phase transfor-mations are identified as hcp-Hf! fcc-HfN. The thermal stability of Cu/Hf/Si and Cu/HfN0.47/Si contact system is

evaluated by thermal stressing at various annealing temperatures. Nitrogen incorporated Hf ﬁlms possess better barrier performance than sputtered Hf ﬁlms. For the Cu/Hf/Si contact system, the interfacial reaction between the Hf barrier layer and the Cu layer is observed after annealing at 550C for 30 min, and copper–hafnium compounds form. Highly resistive copper silicide forms after annealing at 600C for 30 min. The Hf barrier fails due to the reaction of Cu and the Hf barrier, in which Cu atoms penetrate into the Si substrate after annealing at high temperature. The Cu/HfN0.47/Si is

fairly stable up to annealing at 650C for 30 min. In addition, no copper–hafnium and copper silicide compounds are found. Diﬀusion resistance of nitrogen-incorporated Hf barrier is more eﬀective. The thermal stabilities of Cu/HfN0.47/

n+–p junction diodes are enhanced by nitrogen incorporation. The Cu/Hf/n+–p junction diodes result in large reverse-biased junction leakage currents after annealing at 500C for 30 min. On the other hand, Nitrogen incorporated Hf diffusion barriers retained the integrity of junction diodes up to 550C with lower reverse current densities. Phase trans-formation of hafnium-based barrier films with nitrogen incorporation are believed to impede Cu diffusion into the Si

*

Corresponding author. Tel.: +886 2 27361661x5128; fax: +886 2 27362295.

E-mail address:[email protected](S.-Y. Lee).

(2)

substrate and hence improve the barrier performance. Nitrogen incorporated hafnium diﬀusion barrier can suppress the formation of copper–hafnium compounds and copper penetration, and thus improve the thermal stability of barrier layer.

Keywords: Copper; Hafnium; Nitrides; Sputtering; Junction diodes

1. Introduction

A high performance interconnection network on a chip is becoming increasingly important for ultralarge-scale integration (ULSI) of Si integrated circuits. The use of copper in on-chip metallization of microelectronic devices has recently attracted considerable attention due to its lower electrical resistivity and higher electromigration resistance than aluminum. The use of copper, however, raises several problems. For instance, Cu cannot adhere well to most dielectric substrates and is highly reactive with most metals and semiconductors. Hence, thin film adhesion promoters and diffusion barriers must be used to enhance the adhesion and inhibit diffusion in the Cu-based metallization. Refractory metals have been investigated for met-allization applications. Among the barrier films, tantalum-based film has been proven to be one

of the most useful barrier materials [1–3].

How-ever, since the resistivities of TaN and Ta–Si–N ﬁlms are about 200 and 625 lX-cm, respectively, these materials are deemed unfavorable for use

as low-resistivity diﬀusion barrier[3–5]. Therefore,

new barrier materials with high thermal stability and low electrical resistivity are needed to develop. Other refractory metals probably exhibit very favorable properties, and in particular sputtered hafnium ﬁlms have been subjected to preliminary

evaluation [6–9]. In this article, thermal stability

of hafnium-based barrier layers (50 nm) was stud-ied in the Cu metallization system. Furthermore, properties of barrier layers were evaluated by elec-trical measurements and material analyses.

2. Experimental procedure

The barrier performace of Hf–N ﬁlms against Cu diﬀusion was investigated using a structure of

Cu/Hf–N/n+–p junction diodes. The key feature

of this experiment is the various nitrogen ﬂow rate during sputtering of Hf–N ﬁlm formation. First, p-type (1 0 0)-oriented Si wafers with a resistivity 6–9 X-cm were used in this study. After standard RCA cleaning, the wafers were administered the LOCOS

process to deﬁne active regions. The n+–p

junc-tions were formed by As+ implantation at 60

keV with a dose of 5· 1015

cm2followed by the

rapid thermal annealing (RTA) process at 1050

C for 30 s in N2ambient. After the contact

win-dows were cleaned by dipping sample in HF, a reactively sputtered Hf–N film (50 nm) was depos-ited onto the active regions with various nitrogen flow ratio. In this paper, nitrogen flow ratio is

de-fined as a ratio of N2partial flow to total gas flow

(N2+ Ar). Then Cu ﬁlm with a thickness of 300

nm was deposited subsequently in the same sput-tering system without break vacuum. During the sputtering, gas pressure was maintained at 0.8 Pa with a power selected at 500 and 1500 W for Hf– N and Cu, respectively. Finally Cu and Hf–N

layers were patterned by dilute HNO3 and Cl2

plasma, respectively, for the formation of Cu/Hf–

N/n+–p junction diodes.

To estimate the barrier capability of Hf–N ﬁlms

against Cu diﬀusion, the devices (Cu/Hf–N/n+–p

junction diode) were thermally annealed at a

tem-perature ranging from 400 to 650C for 30 min in

a vacuum of 1.33 Pa. For electrical analyses, leak-age current of the diodes was measured by HP4145B semiconductor parameter-analyzer at a

reverse bias of 5 V. After annealing at various

temperatures for 30 min, the diode leakage current was measured. In addition, surface morphologies of the Hf–N ﬁlms were analyzed by a Nanoscope III atomic force microscope (AFM) with a Si probe. AFM probe was scanned over an area of

5· 5 lm2 with 512 scans at 1 Hz scanning rate

(3)

were taken using a four-point probe system. Graz-ing incidence X-ray diffractometry (GIXRD) was used to identify the phases of the films. The compositions of the films were analyzed by X-ray photoemission spectroscopy (XPS) with a monoc-hcrromatic Mg Ka source. The X-ray power was 250W (15 kV at 16.7 mA). The XPS energy scale was calibrated by setting the binding energy of

Ag3d5/2line on clean silver to exactly 368.3 eV

ref-erenced to the Femi level. The angle of incidence of the X-ray beam with the specimen normal was 45. High-resolution scans were run for Hf and N using X-ray beam with about a 15 nm diameter. Furthermore, 300 nm thick Cu films were sput-tered onto Hf–N films to investigate their ability to resist Cu diffusion. Cu/Hf–N/Si samples were

annealed from 450 to 700 C in vacuum for 30

min to evaluate their barrier stability. Surface morphology of annealed Cu/barrier/Si was ob-served by scanning electron microscopy (SEM). Compositions of failure sites were analyzed by en-ergy dispersive spectrometry (EDXS) after remov-ing both copper and barrier layers by wet-chemical solution. The interface microstructure was exam-ined by transmission electron microscopy (TEM) and energy dispersive spectrometry (EDXS). Cross-sectional TEM samples were prepared elec-tron transparency by mechanical thinning fol-lowed by ion milling in a precision ion polishing system (PIPS).

3. Results and discussion

3.1. Properties of the Hf and Hf–N thin ﬁlms

Fig. 1 shows a series of X-ray diﬀraction

pat-terns for different Hf and Hf–N films, deposited on bare silicon substrates under various nitrogen flow rates. The X-ray diffraction pattern of the

hafnium ﬁlm is denoted as (a) in Fig. 1. The

(1 0 1) and (1 1 0) peaks conﬁrm that the hafnium ﬁlms on silicon substrates have a hexagonal close packed structure (a-Hf) (JCPDS 05-0670). The

ﬁrst nitride phase to from is HfN0.4 (JCPDS

40-1277) when a small amount of nitrogen is added,

as indicated in spectrum (b) of Fig. 1. When the

nitrogen content increases to 22.1 at. %, the

dif-fraction pattern of the HfN0.28ﬁlm contains both

narrow and broad peaks, as shown inFig. 1,

de-noted as (c). This indicates that at least two phases with diﬀerent grain sizes are present. It is reported that diﬀerent phases, a-Hf, e-Hf3N2, f-Hf4N3, and

HfN can exist in a limited composition range[10].

The narrow peaks belong to e-Hf3N2(JCPDS

24-0466) phase. The broad scanning XRD pattern

implies that the HfN0.28ﬁlm contains some

amor-phous-like or ﬁne-grained materials. The relatively sharp peaks of the fcc-HfN (JCPDS 33-0592)

phase (denoted as (d) inFig. 1) are observed for

the HfN0.47ﬁlm. The XRD results inFig. 1clearly

show that a-Hf, HfN0.4, e-Hf3N2and HfN phases

form successively as nitrogen ﬂow rate increases from 0 to 3.0 sccm.

To determine the nitrogen concentrations of the Hf–N films deposited at various nitrogen flow rates, chemical compositions are evaluated by XPS. The nitrogen content of the Hf–N film in-creases with the amount of nitrogen in the sputter-ing gas. The compositions of Hf–N films deposited at nitrogen flow rates of 1, 2, and 3 sccm, are HfN0.1, HfN0.28, and HfN0.47, respectively.Fig. 2 offers consistent results regarding the critical

nitro-gen ﬂow rates to form nitronitro-gen-saturated HfN0.1

(1 sccm), HfN0.28(2 sccm) and HfN0.47(3 sccm).

Fig. 3 also shows the electrical resistivity of Hf–

N ﬁlm as a function of nitrogen ﬂow rate. The

20 30 40 50 60 70 80 (102) (110) (103) (112) (202) (110) (311) (220) (200) (101) (d)HfN_0.47 (c)HfN_0.28 (b)HfN_0.1 (a)Hf (111) (110) Intensity (arb.unit) 2θ(degree) (002) (101) (104) (009) ♦ ♦ ♦ ♦ ♦ ♦ (101) ∆ α-Hf HfN0.4 ε-Hf3N2 ∆ HfN ∆ _∆ _∆ (100) (b)HfN0.1

Fig. 1. XRD patterns of the Hf–N ﬁlms with various nitrogen contents.

(4)

resistivity of the as-deposited film shows several interesting features. The electrical resistivity of the Hf film is 100.98 lX-cm. Resistivity increases slightly when a small amount of nitrogen is added to the sputtering gas. The resistivity of the film in-creases to 292.17 lX-cm when the nitrogen flow rate is 2 sccm. When more nitrogen is incorporated

into the Hf ﬁlm, the resistivity of the HfN0.47ﬁlm

increases to 356.28 lX-cm. For comparison, the nitrogen ﬂow rates (1, 2, and 3 sccm) dividing the four regions closely correspond to the ﬂow

rates of ﬁnding HfN0.4, e-Hf3N2, and fcc-HfN

from XRD patterns. Moreover, the variations in

resistivity are attributed to microstructure and phase transformation.

3.2. Thermal stability of Cu/Hf/Si and Cu/HfN0.47/

Si systems

Barrier capability of thin Hf and HfN0.47ﬁlms

was investigated by evaluating the thermal

stabil-ity of Cu/barrier (50 nm)/n+–p junction diodes

using electrical measurements. In this measure-ment, the reverse current densities were obtained from an average value of 25 samples and the diode

area was 1000· 1000 lm2. Fig. 4 illustrates the

statistical distributions of reverse bias reverse

cur-rent density for Cu/barrier (50 nm)/n+–p junction

diodes annealed at various temperatures. For the diodes without annealing, the reverse current

den-sities remain stable (below 10 nA/cm2) as nitrogen

ﬂow ratio is increased. However, the diode leakage increases with increasing the annealing tempera-ture and most diodes are degraded after annealing

at 600C. As shown inFig. 4, the reverse current

densities initially decrease for 400 C annealing,

and then increase with increasing the annealing temperatures. If a failure criterion is deﬁned as

106 A/cm2, the Cu/Hf/n+–p diodes remained

sta-ble after annealing at temperatures up to 450 C

but suﬀered degradation at 500 C for 30 min. It

is reported that the barrier properties can be im-proved by adding impurities, such as nitrogen

0.0 0.5 1.0 1.5 2.0 2.5 3.0 0.0 0.1 0.2 0.3 0.4 0.5 N/Hf ratio

Nitrogen flow rate (sccm)

Fig. 2. Composition (N/Hf atomic ratio) plots of Hf-based ﬁlms deposited on silicon dioxide under various nitrogen ﬂow rates. 0.0 0.5 1.0 1.5 2.0 2.5 3.0 100 150 200 250 300 350 Film resistivity ( µΩ – cm)

Nitrogen flow rate (sccm)

Fig. 3. The electrical resistivity of Hf–N ﬁlms as a function of nitrogen ﬂow rate.

300 350 400 450 500 550 600 10-10 10-9 10-8 10-7 10-6 1x10-5 1x10-4 10-3 10-2 10-1 100 Annealing temperature (˚C)

Leakage Current Density (A/cm

2)

Cu/hcp-Hf/n+-Si

Cu/fcc-HfN_0.47/n+-Si

Fig. 4. Variation in reverse current density of Cu/barrier/Si as a function of annealing temperature.

(5)

and oxygen[3,13,14,16,17]. It was similar reported results by Tsai et al. It was found that the diodes with 60 nm PVD and CVD metal nitride barriers

would begin to deteriorate at 500 and 550 C for

30 min. [15]. It is shown that barrier capability

of Hf–N ﬁlm is better than that of Hf ﬁlm without nitrogen incorporation. As nitrogen concentration

increases, HfN0.47 barrier ﬁlm was formed. The

Cu/HfN0.47/n

+

–p diodes retained better electrical

integrity after annealing at 550 C with lower

re-verse current densities. The HfN0.47 ﬁlms have

much better barrier performance than Hf barrier film. The improved barrier capability is attributed to finer crystallization and interstitial effect is thought to result in microstructural variation and hence improve barrier capability. It is reported that the microstructure within the barrier layer strongly affects the barrier performance because Cu diffuses through fast diffusion paths such as

grain boundaries within the barrier layer[3,15,18].

Fig. 5plots the GIXRD patterns of the Cu/Hf/

Si samples before and after annealing at 550C for

30 min. Strong Cu (1 1 1) and weak Cu (2 0 0) peaks are observed in unannealed Cu/Hf/Si sam-ple, implying that the Cu ﬁlms prefer the (1 1 1) crystal orientation. Copper with a high (1 1 1) tex-ture has been reported to exhibit well resistance to

electromigration [11]. The diﬀraction peaks of

a-Hf (1 0 1) and Cu (1 1 1) clearly disappear and

the CuHf2phase appears for Cu/Hf/Si sample

an-nealed at 550 C for 30 min. These results show

that the interdiﬀusion of Cu and Hf induces the

formation of Cu–Hf and Cu–Hf–Si compounds.

Fig. 6 shows XRD patterns of the Cu/Hf/Si and

Cu/HfN0.47/Si samples after annealing at 600 C

for 30 min. Strong Cu3Si peaks are observed for

annealed Cu/Hf/Si sample. It reveals that high-resistance copper silicide compounds were formed

after 600C annealing. It also indicates the

degra-dation of the Hf barrier after annealing. The as-deposited Hf film consists of fine columnar grains. The degradation of the Hf barrier is attributed to the diffusion of Cu into the Si substrate through the columnar Hf barrier. Strong Cu (1 1 1), weak Cu (2 0 0), and HfN (1 1 1) peaks are found in

the XRD spectrum of annealed Cu/HfN0.47/Si

sample. The intensity of the HfN peak is lower. That is, it has a higher full width at half maximum

(FWHM). The as-deposited HfN0.47 barrier ﬁlm

has an amorphous-like structure [12]. Neither

Cu–Si nor Cu–Hf and Cu–Hf–Si compounds are

observed after annealing at 600 C for 30 min.

The results reveal that the HfN0.47barrier is more

effective in preventing Cu diffusion than the Hf barrier film. The Cu–Hf and Cu–Hf–Si

com-pounds are not observed in Cu/HfN0.47/Si sample

after annealing. The diﬀraction peaks of HfN phases are observed. These results show that nitro-gen incorporation in the Hf ﬁlm can suppress the

formation of Cu–Hf(–Si) compound.Fig. 7shows

XRD patterns of the Cu/HfN0.47/Si samples after

annealing at 650C and 700 C for 30 min. The

Fig. 5. XRD patterns of Cu/Hf/Si systems before and after annealing at 550C for 30 min.

Fig. 6. XRD patterns of of Cu/Hf/Si and Cu/HfN0.47/Si systems after annealing at 600C for 30 min.

(6)

intensity of the HfN (1 1 1) peak increases

obvi-ously after 650 C-annealing. This result reveals

that annealing causes development of HfN crys-tals. Copper silicide phases are found obviously

for Cu/HfN0.47/Si samples annealed at 700 C for

30 min.

Fig. 8 shows the SEM-EDXS and

cross-sec-tional TEM micrographs of Cu/Hf/Si after

anneal-ing over the failure temperature. As shown inFig.

8(a), protrusions were observed on the surface,

indicating a severe reaction of Cu/Hf/Si.Fig. 8(a)

shows EDXS spectrum of the protrusion (denoted as A). It revealed that the protrusion consisted of copper and silicon element and was a copper rich region. These protrusions were presumably caused by Cu diﬀusion through the localized weak points in the barrier and reacting with underlying Si to

form Cu3Si. Similar phenomena were observed

for Ta diﬀusion barriers by Wu et al. [18]. Fig.

8(b) andFig. 9are shown the local phase

determi-nation by analytical methods of cross-sectional

TEM and EDXS. Fig. 8(b) shows the

cross-sec-tional TEM image of Cu/Hf/Si sample after

annealing at 600C for 30 min. It is obvious that

hafnium silicide and copper silicide were observed after annealing, and the interface of the sample was not clear. It indicates the degradation of the

Hf barrier after annealing at 600 C for 30 min.

The as-deposited Hf film consists of fine columnar grains. The degradation of the Hf barrier is attrib-uted to the diffusion of Cu into the Si substrate through the columnar Hf barrier. Trapezoidal

copper silicide spikes bounded by Si {1 1 1} and

Si {001} planes were observed [13]. Fig. 9 shows

cross-sectional TEM micrograph of Cu/HfN0.47/

Si systems after annealing at various temperatures.

Fig. 9(a) shows the TEM image of the Cu/HfN0.47/

Si sample after annealing at 650C for 30 min. The

multilayered structure is obvious. No copper sili-cides are observed at the interface, demonstrating

the excellent barrier properties. Fig. 9(b) shows

cross-sectional TEM micrograph of 700

C-an-nealed Cu/HfN0.47/Si. Trapezoidal copper silicide

compound is observed, demonstrating the Cu dif-fusion into Si substrate. These investigations of TEM are in agreement with the XRD results, which indicate the diﬀusion of Cu atoms through

Fig. 7. XRD patterns of Cu/HfN0.47/Si contact systems before and after annealing at 650 and 700C for 30 min.

Fig. 8. (a) SEM micrograph and EDS spectrum obtained from the precipitate of Cu/Hf/Si sample annealed at failure temper-ature, and (b) cross-sectional TEM micrograph of Cu/Hf/Si annealed at 550C for 30 min.

(7)

HfN0.47 layer and the formation of compounds containing Cu, Hf and Si are the main causes of

failure for Cu/HfN0.47/Si barrier layer.

3.3. Failure mechanism of Cu/hcp-Hf/Si and Cu/fcc-HfN/Si system

The GIXRD patterns, SEM-EDS, XTEM and reverse current density clearly reveal a diﬀerence between the failures of hafnium and hafnium

nitride barriers. Nitrogen incorporated hafnium film has improving barrier capability against cop-per diffusion. Nitrogen is incorporated into the hafnium and it induces phase transformation to hafnium nitride. The interstitial effect is thought to induce microstructural variation and thereby

improve barrier performance. Fig. 10

schemati-cally reveals cross-sections of the interfacial

struc-tures of the Cu/Hf/Si and Cu/HfN0.47/Si systems

before and after annealing. Cu ﬁlms on Hf–N bar-riers have a preferred {1 1 1} orientation. The as-deposited Hf barrier has an hcp-Hf structure with

columnar grains. The formation of CuHf2and

haf-nium silicide is observed after annealing at 550C

for 30 min, revealing barrier degradation. The mechanism by which the barrier fails involves the sacriﬁcial reaction of Hf with Cu and the motion

of Cu through columnar Hf barrier to form Cu3Si.

The barrier capability of Hf film against Cu diffu-sion can be improved by incorporating nitrogen into Hf films using reactive sputtering. Adding impurities, such as nitrogen and oxygen has been reported to improve the barrier properties of

refractory metals [3,13,14]. The as-deposited

HfN0.47 barrier has the ﬁne-grained fcc-HfN

Fig. 10. Schematic illustrations of the microstructures of: (a) Cu/Hf/Si, (b) Cu/HfN0.1/Si, (c) Cu/HfN0.28/Si, and (d) Cu/ HfN0.47/Si contact systems before and after annealing. Fig. 9. The cross-sectional TEM micrograph of Cu/HfN0.47/Si

(8)

phases. Neither Cu–Si nor Cu–Hf(–Si) compounds

are observed for the Cu/HfN0.47/Si sample

an-nealed at 550C for 30 min, revealing better

bar-rier performance than hafnium barbar-rier. However,

enhancing crystalline structure of the HfN0.47

bar-rier is found after annealing at 600C for 30 min.

Cu3Si compounds are not observed for the Cu/

HfN0.47/Si contact system annealed at 650 C.

The microstructural transition of nitrogen-incor-porated hafnium ﬁlm alleviates the copper diﬀu-sion and, therefore, enhances the stability of the barrier and restrains the formation of copper–haf-nium compounds. Cu/Hf–N/Si contact system with high thermal stability is obtained.

4. Conclusion

Microstructure variations of hafnium barrier film with various nitrogen concentrations were investigated by Hf/Si and Hf–N/Si. Furthermore, barrier performances against Cu diffusion were evaluated for Cu/Hf/Si and Cu/Hf–N/Si contact systems. A thermally stable Cu/Si contact system, with a low resistivity Hf–N diffusion barrier, is successfully demonstrated. The as-deposited Hf film has a hexagonal close packed structure and a low resistivity of 100.98 lcm. Phase

transfor-mations are identiﬁed as a-Hf! HfN0.4!

e-Hf3N2! fcc-HfN with increasing nitrogen

concentration of Hf–N ﬁlm. Cu3Si compounds

are found for Cu/Hf/Si contact systems after

annealing at 600 C for 30 min. The mechanism

by which the Hf barrier fails involves the sacriﬁcial reaction of Hf with Cu and the motion of Cu

through columnar Hf barrier to form Cu3Si. The

Cu/HfN0.47/Si contact system tolerates annealing

at 650C for 30 min without any reaction. Cu3Si

compounds are observed after annealing at 700 C due to accelerating Cu diﬀusion through crys-talline Hf–N barrier. The thermal stabilities of

Cu/HfN0.47/n+–p junction diodes are enhanced

by nitrogen incorporation. The Cu/Hf/n+–p

junc-tion diodes result in large reverse-biased juncjunc-tion

leakage currents after annealing at 500C for 30

min. HfN0.47barrier retained the integrity of

junc-tion diodes up to 550C with low reverse current

densities. Phase transformation of hafnium-based

barrier films with nitrogen incorporation are be-lieved to impede Cu diffusion into the Si substrate and hence improve the barrier performance. Nitro-gen incorporated hafnium diffusion barrier can suppress the formation of copper–hafnium com-pounds and copper penetration, and thus improve the thermal stability of barrier layer.

Acknowledgements

The work was ﬁnancially supported by the Na-tional Science Council of the Republic of China under Contract No. NSC92-2314-B-038-012 and this study was sponsored by the Taipei Medical University (TMU92-AE1-B02).

References

[1] D.S. Yoon, H.K. Baik, S.M. Lee, J. Vac. Sci. Technol. B 17 (1) (1999) 174–181.

[2] H. Ono, T. Nakano, T. Ohta, Appl. Phys. Lett. 64 (12) (1994) 1511–1513.

[3] C.E. Ramberg, E. Blanquet, M. Pons, C. Bernard, R. Madar, Microelectron. Eng. 50 (2000) 357–368.

[4] E. Kolawa, J.S. Chen, J.S. Reid, P.J. Pokela, M.A. Nicolet, J. Appl. Phys. 70 (1991) 1369–1373.

[5] J.O. Olowolafe, I. Rau, K.M. Unruh, C.P. Swann, Z.S. Jawad, T. Alford, Thin Solid Films 365 (2000) 19–21. [6] I. Suni, M. Ma¨enpa¨a¨, M.A. Nicolet, M. Luomaja¨rvi, J.

Electrochem. Soc. 130 (1983) 1215–1218.

[7] S. Shinkai, K. Sasaki, Jpn. J. Appl. Phys. 38 (1999) 3646– 3650.

[8] B.O. Johansson, J.E. Sundgren, U. Helmersson, J. Appl. Phys. 58 (8) (1985) 3112–3117.

[9] S. Shinki, K. Sasaki, Jpn. J. Appl. Phys. 38 (1999) 2097– 2102.

[10] B.O. Johansson, J.E. Sundgren, U. Helmersson, M.K. Hibbs, Appl. Phys. Lett. 44 (7) (1984) 670–672.

[11] D. Save, F. Braud, J. Torres, F. Binder, C. Mu¨ller, J.O. Weidner, W. Hasseand, Microelectron. Eng. 33 (1997) 75– 84.

[12] S.Y. Chiou, K.L. Ou, W.F. Wu, C.P. Chou, Y.M. Chang. Eﬀect of nitrogen addition on microstructure and thermal stability of hafnium nitride diﬀusion barriers in copper metallization, in: Proceedings of Symposium on Nano Device Technology, 2002, p. 65.

[13] K. Holloway, P.M. Fryer, C. Cabral Jr., J.M.E. Harper, P.J. Bailey, K.H. Kelleher, J. Appl. Phys. 71 (11) (1992) 5433–5444.

[14] X. Sun, E. Kolawa, J.S. Chen, J.S. Reid, M.A. Nicolet, Thin Solid Films 236 (1993) 347–351.

(9)

[15] M.H. Tsai, S.C. Sun, C.E. Tsai, S.H. Chuang, H.T. Chiu, J. Appl. Phys. 79 (9) (1996) 6932–6938.

[16] K.H. Min, K.C. Chun, K.B. Kim, J. Vac. Sci. Technol. B 14 (5) (1996) 3263–3269.

[17] G.S. Chen, S.T. Chen, J. Appl. Phys. 87 (12) (2000) 8473– 8482.

[18] W.F. Wu, K.L. Ou, C.P. Chou, C.C. Wu, J. Electrochem. Soc. 150 (2) (2003) G83.