Characterization of the low temperature dopant activation behavior at NiSi/silicon interface formed by implant into silicide method

(1)

Characterization of the low temperature dopant activation behavior

at NiSi/silicon interface formed by implant into silicide method

Kow-Ming Chang

*

, Jian-Hong Lin, Cheng-Yen Sun

Department of Electronics Engineering and Institute of Electronics, National Chiao-Tung University, 1001 Ta Hsueh Road, Hsinchu 30050, Taiwan, ROC

1. Introduction

Both shallow junction and high-k dielectric formation are important issues for manufacture of CMOS devices beyond the 45 nm generation[1]. For ohmic contact consideration, traditional junction formation technology requires high-temperature anneal-ing (i.e. RTA 1000 8C 30 s), but this high-temperature process will affect not only the original dopant concentration profile but also make some high-k dielectric crystallized[2,3]. For device integra-tion consideraintegra-tion, low temperature activaintegra-tion technique provides an alternative possibility to give solutions to these two difficulties. Solid phase epitaxial regrowth (SPER)[4,5]is the main idea to achieve the low temperature activation. However, preamorphized layer, especially for implantation formed amorphous layer, still needs well thickness control to achieve the goal of the shallow junction formation. On the other hand, IIS method[6]can take the advantages of both SPER (a thin amorphous Si layer will be formed at the silicide/silicon interface during silicidation process) and fine junction depth control (determined by silicide thickness). In addition, dopant segregation [7] phenomenon can pile up the dopant concentration at the silicide/silicon (M/S) interface, and thus has a potential to form a better ohmic contact. SIMS has frequently been adapted to show the dopant segregation phenomena at the silicide/silicon interface and the spreading resistance profiling (SRP) method has been used to analysis the dopant concentration at

the bulk activated region [4,5,8]. However, few literatures have discussed the dopant activation behavior at the silicide/silicon interface. In this work, we use the metal/semiconductor (MS) diode structure combined with I–V (current–voltage), C–V (capacitance– voltage), and SIMS measurements to explore the high activation ability at the silicide/silicon interface made by the IIS method. 2. Device fabrication

Stating material is (1 0 0) n-type (1–10 V cm, phosphorous doped) and (1 0 0) p-type (1–10 V cm, boron doped) wafer. A 200 nm thick silicon dioxide was thermally grown for isolation. Nickel silicide was formed by two-step rapid thermal annealing (RTA) process. A 20 nm thick nickel ﬁlm was deposited by sputter system. All samples were under ﬁrst step RTA at the condition 400 8C 30 s. After removing the unreacted Ni, p-type substrate samples were implanted with BF2 (1 1013cm2, 20 keV) and n-type

substrate samples were implanted with arsenic (1 1013_cm2_,

10 keV). Before thermal coated Al as the back contact, samples were treated with 2nd RTA step 30 s at different temperatures from 400 to 650 8C, 50 8C per step (400, 450, 500, 550, 600, and 650 8C). 3. Results and discussions

3.1. SIMS measurement

SIMS proﬁle of the BF2doped sample with 2nd RTA 400 8C 30 s

is shown inFig. 1, and it provides two messages. First, the peak boron concentration is appeared inside the nickel silicide ﬁlm

Applied Surface Science 254 (2008) 6151–6154

A R T I C L E I N F O Article history:

Available online 18 March 2008 PACS:

61.72.Tt 81.20.n Keywords: Nickel silicide Implant into silicide Solid phase epitaxial regrowth Rapid thermal anneal

A B S T R A C T

Process temperature and thermal budget control are very important for high-k dielectric device manufacturing. This work focuses on the characteristics of low temperature activated nickel silicide/ silicon (M/S) interface formed by implant into silicide (IIS) method. By combining SIMS, C–V, I–V, and AFM measurements in this work, it provides a clear picture that the high dopant activation ratio can be achieved at low temperature (below 600 8C) by IIS method. From SIMS and C–V measurements, high dopant activation behavior is exhibited, and from I–V measurement, the ohmic contact behavior at the M/ S junction is showed. AFM inspection displays that under 2nd RTA 700 8C 30 s no agglomeration occurs. These results suggest that IIS method has the potential to integrate with high-k dielectric due to its low process temperature. It gives an alternate for future device integration.

* Corresponding author. Tel.: +886 3 5712121x54205; fax: +886 3 5724361. E-mail address:[email protected](K.-M. Chang).

C o n t e n t s l i s t s a v a i l a b l e a tS c i e n c e D i r e c t

Applied Surface Science

j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m / l o c a t e / a p s u s c

(2)

suggesting that the projected range at this study is in the silicide, which implies the implantation damage in the silicon is depressed. And secondly, dopant segregation phenomenon can be found at the silicide/silicon interface even at the short period with low temperature annealing (2nd RTA 400 8C 30 s). The silicide thickness from SIMS measurement is about 23 nm, and this value is conﬁrmed by the SEM inspection.

3.2. C–V measurement

At this study, we adapted the capacitance–voltage measure-ment[9]to analysis the silicide/silicon interface. With depletion approximation, depletion width (W) in the silicon is estimated by W ¼Kse0A

C (3.1)

where Ksis dielectric constant of silicon and A is the MS diode’s

contact area. And the effective doping density at depth W is given by

NAðWÞ ¼

2 qKse0A2dð1=C2Þ=dV

(3.2) Averaged boron doping density within the depletion region extracted form C–V measurement is illustrated inFig. 2. It shows that with higher 2nd RTA temperature, higher doping density is obtained. From C–V measurement results, two observations can be

found. First, even at the lowest activation temperature (400 8C), the effective doping density (Neff) is higher than 1018cm3, this

convinces that IIS is an effective method to achieve low temperature activation. And secondly, at higher temperature annealing, Neffbecomes higher than 1019cm3. Silicon begins to

become the degenerate substance at this high doping level, and this property is good for silicide/silicon interface to form an ohmic contact. Furthermore, we compared the integral of the dopant concentration over depths from SIMS and C–V measurement results (nSIMS and nCV) in Fig. 3. It shows that the activation

percentage (nCV/nSIMS) is very high. Over 50% dopant was activated

over 2nd RTA 450 8C. The summation process was taken from Vr= 0

to 0.2 V and dopant within the depletion region at V = 0 was not summed in the C–V measured data. Samples with high Neffmight

not be fully depleted in this voltage range; as a result, under-estimated nCVmight occur in these high Neffsamples. There exist an

abnormal activation calculation occurred at 2nd RTA 650 8C sample. This may be due to the incorrect high doping density obtained from Eqs. (3.1) and (3.2) that using Maxwell–Boltzmann distribution. The Maxwell–Boltzmann distribution will over-estimate the Neff when the material becomes degenerate [10]

and as the consequence, the activated dopant obtained using Neff

will also be over-estimate. At this study, we only use the Maxwell– Boltzmann distribution to estimate the doping density. Although it may not good enough to extract the real concentration of the sample, it does reﬂect the high activation ability at the silicide/ silicon interface, and shows that near the M/S interface, silicon is undoubtedly turning into the degenerate material.

Fig. 4. Phosphorous concentration (cmS3

) calculated from C–V measurement, and activated dopant (cmS2

) as the integral of the concentration over depths from C–V measurement at different 2nd RTA temperatures.

Fig. 1. SIMS proﬁle of the BF2doped p-type substrate, with 2nd RTA 400 8C 30 s. The

dopant segregation phenomenon can be found at about 23 nm from surface. The shadow area from 23 nm to about 50 nm is about 5 T 1012_cmS2_{, which is deﬁned}

as dopant implanted inFig. 3.

Fig. 3. Dopant (cmS2

) as the integral of the concentration over depths from C–V and SIMS proﬁle (dopant activated and dopant implanted) versus different 2nd RTA temperatures.

Fig. 2. Boron concentrations at different 2nd RTA temperatures extracted by C–V method.

K.-M. Chang et al. / Applied Surface Science 254 (2008) 6151–6154 6152

(3)

Similarly, phosphorous doped samples exhibit the high activa-tion ability. It can be found inFig. 4that with above 500 8C 2nd RTA, phosphorous doped samples also achieve high effective doping densities more than 1 1019cm3_{, which make the silicon at the M/}

S interface to become ohmic contact. 3.3. I–V measurement

For direct determine if the M/S junction is ohmic contact or not, I– V measurement can be a good reference. The corresponding I–V characteristics of boron doped samples are illustrated inFig. 5all

samples with 2nd RTA have linear I–V characteristics instead of Schottky diodes regulation characteristics, which means that all samples behave more likely to be ohmic contacts than Schottky contacts. In addition, the slopes of the I–V curves are increasing when raising the 2nd RTA temperature from 400 to 550 8C. Since the silicide after 2nd RTA temperatures at the range between 400 and 650 8C is at the nickel mono silicide phase[11], the sheet resistance is the same at these temperatures. The different characteristics of I–V curves with different 2nd RTA temperatures should be contributed to the differences in contact resistance. This indicates that the contact resistance is decreasing with increasing 2nd RTA tempera-ture. This is in good agreement to C–V measurement result that more boron is activated at higher 2nd RTA temperature and higher 2nd RTA temperature can form better ohmic contact. Phosphorous doped samples have similar I–V behaviors. Both BF2doped p-type

substrate and P doped n-type substrate show ohmic contact behavior at low temperature activation with nickel silicide as the contact metal using IIS method.

3.4. AFM inspection

Morphology inspection is required to further to clarify that previous C–V and I–V measurement results are not biased by silicide deformation. Samples (n-type substrate, phosphorous doped) without 2nd RTA and with 2nd RTA 30 s at different temperatures (400, 500,600, and 700 8C) are prepared. Fig. 6 shows the AFM inspection results for these samples (area: 10 mm 10 mm), and all RMS values are around 1 nm. No agglomerations are discovered in these samples even at RTA 700 8C 30 s, which are similar to other groups’ results[11,12]. Since ﬂuorine can retard nickel silicide ﬁlm’s

Fig. 6. AFM images show NiSi surface morphology with 30 s 2nd RTA at (a) 400 8C, (b) 500 8C, (c) 600 8C and (d) 700 8C. Fig. 5. I–V characteristics of boron doped samples. All samples with 2nd RTA behave

more likely to ohmic contacts than Schottky contacts.

(4)

agglomeration[13], it is believed that BF2doped samples can also

sustain RTA 700 8C 30 s without agglomeration occur. Since the experiments from 3.1 to 3.3 are taken under 700 8C, it is believed that these results are not biased by silicide agglomeration effect. 4. Conclusion

From SIMS measurement, it exhibits that there are high dopant dosages pill up at the M/S interface by dopant segregation phenomenon even under low thermal budget 2nd RTA control. This can limit the implantation damage mainly occurred in the silicide layer and provide a better M/S interface. C–V measurement shows that with higher 2nd RTA temperatures at the range 400–650 8C, higher activation can be achieved at the M/S interface for both boron and phosphorous doped samples. Ohmic contact property can be found in I–V measurements, with higher 2nd RTA temperatures between 400 and 550 8C, lower contact resistances are present. AFM inspections show that no agglomeration of silicide ﬁlm is occurred under 700 8C 30 s. From these results, 2nd RTA 500 to 600 8C 30 s may be a good process window for IIS method to future integration with high-k dielectric device.

References

[1] ITRS Roadmap 2006 update.

[2] G.D. Wilk, R.M. Wallace, J.M. Anthony, J. Appl. Phys. 89 (2001) 5243. [3] C.S. Kang, H.J. Cho, R. Choi, Y.H. Kim, C.Y. Kang, S.J. Rhee, C. Choi, M.S. Akbar, J.C.

Lee, Electron Dev., IEEE Trans. 51 (February (2)) (2004) 220.

[4] B.J. Pawlak, R. Lindsay, R. Surdeanu, P.A. Stolk, K. Maex, in: Proceedings of the 14th International Conference on Ion Implantation Technology, 2002, p. X. [5] M.J.P. Hopstakena, Y. Tamminga, M.A. Verheijen, R. Duffy, V.C. Venezia, A. Heringa,

Appl. Surf. Sci. 231–232 (2004) 688.

[6] D.L. Kwong, T.H. Ku, S.K. Lee, E. Louis, N.S. Alvi, P. Chou, J. Appl. Phys. 61 (1987) 5084.

[7] J. Kedzierski, D. Boyd, C. Cabral Jr., P. Ronsheim, S. Zafar, P.M. Kozlowski, J.A. Ott, M. Ieong, Electron Dev., IEEE Trans. 52 (January (1)) (2005) 220.

[8] R. Lindsay, K. Henson, W. Vandervorst, K. Maex, B.J. Pawlak, R. Duffy, R. Surdeanu, P. Stolk, J.A. Kittl, S. Giangrandi, X. Pages, K. van der Jeugd, J. Vac. Sci. Technol. B 22 (January/February (1)) (2004) 306.

[9] Dieter K. Schroder, Semiconductor material and device characterization, second edition, John Wiley & Sons, Inc., 1998 (chapter 2).

[10] J.S. Blakemore, Solid-State Electron. 25 (1982) 1067.

[11] Lauwer Anne, Jorge A. Kittl, Mark J.H. Van Dal, Oxana Chamirian, Malgorzata A. Pawlak, Muriel de Potter, Richard Lindsay, Toon Raymakers, Xavier Pages, Bench-erki Mebarki, Tushar Mandrekar, Karen Maex, Mater. Sci. Eng. B 114–115 (2004) 29.

[12] Chao-Chun Wang, Mao-Chieh Chen, Jpn. J. Appl. Phys. 45 (2006) 1582. [13] A.S. Wong, D.Z. Chi, M. Loomans, D. Ma, M.Y. Lai, W.C. Tjiu, S.T. Chua, C.W. Lim, J.E.

Greene, Appl. Phys. Lett. 81 (2002) 5138. K.-M. Chang et al. / Applied Surface Science 254 (2008) 6151–6154