Continuous CoSi2 layers in silicon synthesized by Co-ion implantation

Continuous CoSi2 layers in silicon synthesized by Co-ion implantation

August 1997 Materials Letters 32 (1997) 121-126 EIBVIER Continuous CoSi, layers in silicon synthesized by Co-ion implantation ’ Jizhong Zhang a,c,*...

481KB Sizes 0 Downloads 26 Views

August 1997

Materials Letters 32 (1997) 121-126


Continuous CoSi, layers in silicon synthesized by Co-ion implantation ’ Jizhong Zhang a,c,*, Xiaoyan Ye b, Jun Chang ‘, S. Bernard d a International

Centre for Materials Physics, Academia Sinica, Shenyang 110015, and Chinese Centre of Advanced Science and Technology (World Laboratory), P.O. Box 8730, Beijing, China b Department of Chemistry, Tsinghua University, Beijing 100084, China ’ Department of Materials Science and Engineering, Tsinghaa University, Beijing 100084, China d Lahoratoire d’Electronique, Ecole Centrale a’e Lyon, Lyon 69131, Ecully Cedex, France Received 5 November

1996; revised 26 December

1996; accepted 30 December


Abstract CoSi, thin layers were fabricated by Co-ion implantation into Si(100) and subsequent annealing. The ion implantation was carried out in a metal vapour vacuum arc (MEVVA) source implanter. Various doses of Co ions were implanted using doses of 1.6 X 10” to 7.5 X 10” Co/cm’ at an extraction voltage of 40 kV and ion current densities of 60 and 150 PA/cm’. It was found that the thicknesses of continuous CoSi, layers after annealing within Si(100) increased with increasing implantation dose, and ranged from 43 to 60 nm for the implantation doses used in this work. The specific resistivity of implanted samples after annealing was between 12.1 and 13.8 p!J cm at room temperature. The results indicated that the ion current density played an important role in the structure of the CoSi, layer, and annealing can significantly improve the electrical property of the CoSi, layers. PACS: 68.55.LM Keywords:

Silicide; Cobalt; CoSi,;

Ion implantation

1. Introduction Metal silicides have been studied for more than 20 years because of the growing interest for the fabrication of microelectronic devices, and more recently for synthesis of high-temperature silicides [l-

* Corresponding author. Tel.: 86-10-6278-4546; fax: 86-lo6256-2768; e-mail: [email protected] ’ This project was supported by National Natural Science Foundation of China, and Radiation Beams and Materials Engineering Laboratory, Beijing Normal University, Beijing 100875, China. 00167-577X/97/$17.00 Copyright PII SOl67-577X(97)00031-1

41. The transition model silicides have been the object of extensive investigation due to their potential application for ultra large scale integration (ULSI) because of their low resistivities, good thermal stability, and ease of processing. They are increasingly used for contact and gate metallizations. These silitides may also be useful for use as surface or as buried interconnection conductors, for three-dimensional integration [5]. Among the transition metal silicides, CoSi, is an important silicide [6-81. Soref

et al. developed a new class of infrared waveguide devices in which light is trapped in silicon by one or

0 1997 Elsevier Science B.V. All rights reserved.


J. Zhang et al./Muterial~

more layers of cobalt disilicide [9]. Hermanns et al. reported an ultrafast vertical metal-semiconductormetal photodetector. A CoSi, layer in silicon acts both as a bottom Schottky contact and a buried light reflector [lo]. MoSi, has also been an attractive candidate for electronic, heating, and high-temperature structural applications. An attempt has been made to understand the reaction mechanism involved in synthesis of high-temperature silicides by combustion synthesis of MoSi, [4]. A chromium silicide alloy is being considered for aerospace applications [ 1 l]. Some oxidation-resistant boron- and germanium-doped silitide coatings for refractory metals at high temperature have also been investigated lately [12]. Two different methods are usually used for epitaxial growth and for silicide formation in general, i.e. the conventional ultra-high vacuum (UHV) and ion beam synthesis (IBS) techniques. The ion beam synthesis is a newly developing technology and is attracting much interest in materials science. IBS uses high dose implantation to add the transition metal ions to a silicon substrate and under proper conditions, followed by subsequent high-temperature annealing which leads to the formation of a silicide layer by coalescence. This new growth method, also called mesotaxy, was successfully used for the first time in 1957 for synthesizing silicides [13]. In this paper, we report the preparation and investigation of layers of ion beam synthesized CoSi, by use of a metal vapour vacuum arc (MEVVA) source implanter.

2. Experimental The wafers used in this work were n-type Si(100) with a resistivity of 2-4 fi cm. These wafers were cut into 30 X 30 mm2 squares. The Co ion implantation was carried out by use of a MEVVA source implanter at an extraction voltage of 40 kV and beam current densities of 60 and 150 FA/cm’, respectively. The beam spot size was 100 mm in diameter. The Co ion beam. was composed of 47% Co+, 49% Co2+ and 4% Co3+, and the corresponding ion energies were therefore 40, 80 and 120 keV from 1.5 X 10” to 7.5 X 10” Co/cm’. The implantation doses of 1.5 X lo”, 2.5 X lo”, 3.5 X 1017,

Letter-s 32 (1997) 121-126

5.5 X lo”, and 7.5 X 10” Co/cm’ were used at an ion current density of 60 kAA/cm2. The dose of 3.5 X 1OJ7 Co/cm2 was also employed at an ion current density of 150 p,A/cm”, in order to understand the effect of temperature resulting from ion implantation. The sample temperature rise during implantation was only due to ion beam heating, and was measured by a thermocouple in close contact with the sample’s surface. The thermocouple was protected from the beam current and thermally insulated with a piece of mica in order to prevent the thermocouple from being heated by parts other than the sample. The temperature of the samples was increased quickly from room temperature to a stable value. Under our experimental conditions, the stable values of temperature resulting from ion bombardment were 462 and 628°C corresponding to ion current densities of 60 and 150 PA/cm*, respectively. It took 7 and 12 min to reach the above stable temperatures for each ion current density as mentioned above. The total implantation times were measured to be 14, 23, 32, 50 and 68 min corresponding to each implantation dose (from 1.5 X lOI to 7.5 X 1017 Co/cm2) at the ion current density of 60 pA/cm2, and 13 min at the ion current density of 150 pA/cm2, respectively. After implantation, a two-step annealing procedure was carried out in a vacuum furnace, with a vacuum of better than 1 X 10m3 Pa. The samples were annealed at 550°C for 1 h, followed by heating at 950°C for half an hour. The as-implanted samples were analyzed by X-ray diffraction (XRD), Rutherford backscattering spectrometry (RBS), and scanning Auger microprobe (SAM) to determine the structure and Co ion depth profile. RBS spectra were measured with a collimated 2.1 MeV He2’ beam produced by a 3SDH type pelletron accelerator. Resistivity measurements were taken at room temperature to characterize the electrical transport of the CoSi, layers.

3. Results and discussion Fig. la shows the XRD spectrum of implanted with Co ions of 60 pA/cm2 to 2.5 X 10’7/cm2. With the exception of the the substrate, Si(lOO), three diffraction

a sample a dose of pattern of peaks of

J. Zhang et al./Materials

Letters 32 (1997) 121-126


5 ‘\






Si ‘.

.... --__

‘C ::a



: I h

60 Two Theta














:’ i,

,c___/ , .__. 100



,,I I










200 Channel





with Co and (b)

Fig. 2. RBS spectra of the sample in Fig. la. (a) dash line: before annealing and (b) solid line: after annealing.

Co%, appeared, corresponding to interplanar spacings for CoSi,(lll), CoSi,(220) and CoSi,(400). This means that the CoSi, phase was formed within Si(100). The measured interplanar spacings are in agreement with that listed in the Joint Committee on Powder Diffraction Standards (JCPDS) Card NO. 38-1449, as shown in Table 1. The similar results were also obtained from the samples implanted at other implantation doses. Fig. lb shows the XFUI

spectrum of the sample implanted with Co ions of ion current density of 60 p,A/cm’ to a dose of 3.5 X 10”/cm2. The pattern of Si(100) disappeared, and three reflections of CoSi,(ll l), CoSi,(331) and Si(5 11) were identified. For all the above implanted samples, only the CoSi, phase and no other compounds was detected. Fig. 2 shows two RBS spectra of the same sample as shown in Fig. la. The pattern in Fig. 2a corresponds to the as-implanted sample before annealing, and Fig. 2b is the after annealing pattern. Fig. 2a shows that a diffusion-like tail on the Co profile is evident, and annealing made this Co diffusion-like tail shorter, as shown in Fig. 2b. Fig. 3 shows the SAM depth profile of the same sample as shown in Fig. la. This result is consistent with the RBS data, and also confirms the presence of a Co diffusion-like tail. Before annealing, the Co distribution (dash line) extended much deeper than the Co ion projected range predicted from the Ziegler et al. [14] tables, i.e., 66 nm for 80 keV Co ions. The shape of the Co distribution curve is also different

Fig. 1. XRD spectra ions of 60 PA/cm* 3.5 X 10’7/cm2.

of the Si(100) samples implanted to doses of: (a) 2.5X 10’7/cm2

Table 1 Comparison of the measured interplanar spacings with the published values for CoSi, (sample implanted with 2.5 X lOI7 Co/cm*) 2 8 (deg)

28.84 47.94 69.20 70.23


CoSi, (111) CoSi, (220) Si (400) CoSi, (400)



measured (nm)

published (JCPDS) (nm)

0.3093 0.1896 0.1357 0.1340

0.30960 0.18967 0.13577 0.13409

J. Zhang et al./Materials



before after

Letters 32 (19971 121-126 Table 2 Specific resistivity tion dose

annealing annealing

( p) of Co%,

layers as a function of implanta-

Sample No.








Depth (run) Fig. 3. SAM profile of Co concentration for the same sample as shown in Fig. la; dash line: before annealing and solid line: after annealing.

compared to the standard transportation range of ions in matter (TRIM) simulation for 80 keV Co ions implanted into Si. It was not a standard Gaussian distribution, and the cobalt profile decreased much slowly. The Co distribution was about half of the maximum concentration value at a depth of nearly 100 nm. After taking account of the range straggeling, a diffusion layer of about 15 nm was measured. The long Co profile tail may be attributed to radiation enhanced diffusion of the implanted Co ions in the sample. The Co atom concentration distribution (solid line) of the annealed sample was obviously different with that (dash line) of the as-implanted sample without annealing. The implanted Co ions are present with a maximum concentration value of 28% for the as-implanted sample and 33% for the annealed sample, respectively. The Co stoichiometric concentration in CoSi, is 33%. After annealing, the thick-


0.5 1.5


Implantation Fig. 4. Thickness dose.

of CoSi,


I. 5


Dose (~10~~ Co/cm’) layers as a function

of implantation











dose (X 10’7/cm2) sheet resistivity (as-implanted) (fi / 0) sheet resistivity (annealed) (a / 0 ) thickness (nm)






3.0 43

2.7 45

2.6 48

2.5 55

2.3 60

p, (p.a cm) (as-implanted) p2 (~a cm) (annealed)

19.4 12.9

27.0 12.1

34.6 12.5

46.8 13.7

120.0 13.8

ness of buried CoSi, thin layer was 45 nm determined from the RBS and SAM data. For implantations of 3.5 X 1017, 5.5 X 1017 and 7.5 X 1017 Co/cm’, the Co distribution before annealing had leveled off at the stoichiometric concentration value of 33%. Apparently the density of the CoSi, precipitates increased with increasing dose. The thickness of the buried CoSi, thin layer also increased with increasing implantation dose, and was linear with implantation does, as shown in Fig. 4. All the implanted samples shown in Fig. 4 were implanted at the same current density of 60 kA/cm2. For microelectronic applications, the electrical properties of the silicide are of critical importance. The sheet resistance of the samples was measured with a four-point probe. The thickness of buried CoSi, layers was obtained from the RBS and SAM data. Hence the specific resistivity of these CoSi, layers was calculated, as listed in Table 2. All the samples listed in Table 2 were implanted at the same ion current density of 60 p A/cm*. It is shown clearly in Table 2 that the specific resistivity of the as-implanted samples increased with increasing implantation dose. For example, the specific resistivity of sample No. 5, i.e., 120.0 p,fl cm, is six times that of sample No. 1. A possible explanation is as follows: the higher implantation dose resulted in heavier radiation damage. The annealing was more efficient in removing the damage and decreasing specific resistivity. The thickness of CoSi, layer formed in samples No. 2 and No. 5 was measured to be 45 and 60 nm, respectively. The measured sheet resistivities of 2.7 n/Cl and 2.3

J. Zhang et al./Materials Letters 32 (1997) 121-126

a/ 0 for these two samples were converted to the specific resistivities of 12.1 and 13.8 pfl cm respectively. Nevertheless, with the exception of sample No. 1, the effect of implantation dose on specific resistivity of annealed samples is still obvious. The higher the implantation dose, the higher the specific resistivity of the annealed sample. A dose range from 2.5 X lOI to 3.5 X 10 is suitable to form a buried CoSi, thin layer with lower electrical resistivity under our experimental conditions. An additional experiment was carried out in order to understand the effect of ion current density on the structure and resistivity of the synthesized silicide. A piece of Si sample was implanted at an ion current density of 150 kA/cm’ to a dose of 3.5 X 1017 Co/cm’. The sample’s temperature increased to 628°C within 12 min. The sheet resistance of the sample before annealing was 52 a/ 0. It was impossible to measure the sheet resistance of this sample after annealing, because the microstructure of this sample displayed a clear network-like porous feature. It is found that ion current density plays an important role in the structure and electrical properties of CoSi, layers. A lower ion current density is better in order to synthesize a higher quality CoSi, layer. As mentioned above, the electrical property of CoSi, layers synthesized at an ion current density of 60 PA/cm* was improved significantly after annealing. This is much different compared to some previous results reported in the literature [15]. It has been observed earlier for binary silicides synthesized using lower ion current densities that the electrical properties get slightly improved upon annealing. Under our experimental conditions, the dependence of resistivity on beam current density correlated with the uniformity of the CoSi, layer. The CoSi, thin layer formed directly after implantation is not uniform. An uniform CoSi, layer was obtained only after annealing, and its electrical properties were also greatly improved.

4. Conclusion In summary, we have reported the results of an experimental study of CoSi, thin layers prepared by use of the MEVVA ion implanter. A buried CoSi,


layer was formed after direct Co ion implantation and subsequent annealing. The thickness of the CoSi, thin layer increased with increasing implantation dose, and ranged between 43 and 60 nm for implantation doses of 1.5 X 10t7-7.5 X 1017 Co/cm2. The specific resistivity of as-implanted samples without annealing was affected strongly by implantation dose and ion current density. The two-step annealing procedure was efficient in removing the radiative damage and improving electrical characteristic of the buried CoSi, thin layer. The specific resistivity of annealed samples ranged from 12.1 to 13.8 JL~ cm. The dose between 2.5 X 10” and 3.5 X 1017 Co/cm2 is better for forming a buried CoSi, thin layer with lower specific resistivity under the above experimental conditions. The ion current density of 150 PA/cm* is not suitable for preparing a high quality CoSi, layer.


This work is also supported by Beijing Zhongguancun Associated Center of Analysis and Measurement, Beijing, China.

References [I] S. Ma&, Nucl. Instr. Meth. Phys. Research B 80/81 (1993) 895. [2] A. Terrasi, S. Ravesi, M.G. Grimaldi and C. Spinella, Appl. Phys. L.&t. 62 (1993) 2102. [3] D.J. Oostra, C.W.T. Bulle-Lieuwma, D.E.W. Vandenhoudt, F. Felten and J.C. Jam, J. Appl. Phys. 74 (1993) 4347. [4] Seetharama C. Deevi, Naresh N. Tbadhani, Mater. Sci. Eng. A 192/193 (1995) 604. [5] M.F. Wu, J. De Wachter, A.M. Van Bavel et al., J. Appl. Phys. 75 (1994) 1201. [6] R. Jebasinski, S. Mantl and Chr. Dieker, Thin Solid Films 223 (193) 298. [7] A. Vantomme, M.F. Wu, G. Langouche, H. Vanderstraeten and Y. Bruynseraede, Appl. Surface Sci. 53 (1991) 278. [81 Neil M. Zimmerman, J. Alex Liddle, Alice E. White and Ken T. Short, Appl. Phys. I&t. 62 (1993) 387. [9] R.A. Soref, F. Namavar, N.M. Kalkhoran and D.M. Koker, Optics Lett. 19 (1994) 1319. [lo] J.P. Hermanns, F. Ruders, E. Stein von Kamienski, H.G. Roskos, H. Kurz, 0. Hollricher, C. Buchal and S. Mantl, Appl. Phys. L&t. 66 (1995) 866.


J. Zhang et al./Materials

[ll] S.V. Ray, Mater. Sci. Eng. A 192/193 (1995) 583. 1121 B.V. Cockeram and R.A. Rapp, Mater. Sci. Eng. A 192/193 (1995) 980. [13] A.F. White, K.T. Short, R.C. Dynes, J.P. Garno and J.M. Gibson, Appl. Phys. Lett. 50 (1987) 95.

Letters 32 (19971 121-126 [14] J.F. Ziegler, J.P. Biersack and U. Littmark, The stopping and range of ions in matter, Vol. 1 (Pergamon Press, New York, 1985). 1151 K. Maex, J. Vanhellemont, S. Petersson and A. Lauwers, Appl. Surface Sci. 53 (1991) 273.