RegularIssuePaper
J o u r n a l o fE l e c t r o n i c Materials, Vol. 24, No. 6, 1995
A Study of Self-Aligned Formation of C54 Ti(Sil_y,? ,)2 to p÷ and n÷ Sio.TGeo.a Alloys Using Rapid Therma
nealing
S.P. A S H B U R N and M.C. OZTURK Department of Electrical and C o m p u t e r Engineering, N o r t h Carolina S t a t e University, Raleigh, NC 27695-7911 G. H A R R I S and D.M. M A H E R D e p a r t m e n t of Materials Science and Engineering, N o r t h Carolina S t a t e University, Raleigh, NC 27695-7916 In this paper, solid s t a t e reactions of t i t a n i u m with b o r o n and phosphorus d o p e d Sio.7Ge9. 3 alloys have been investigated for application in a self-aligned germanosilicide process. W e t chemical etching of the germanosilicide with respect to unreacted T i in a solution of 1:1:5 NHtOH:H202:H20 has been investigated. Characterization was performed u s i n g four-point probe s h e e t resistance measurements, x - r a y diffraction, cross-sectional transmission electron microscopy, Nomarski optical imaging, and scanning electron microscopy. The C54 Ti(SiI yGe.)2 p h a s e was observed to form for reactions on both b o r o n and phosphorus doped'Sio.7Geo. 3 alloys. G r a i n structures of the C54 phases were f o u n d to be similar to g r a i n structures of intrinsic alloy reactions with l a t e r a l g r a i n dimensions on the o r d e r of 0.3 ~m. Resistivities of 22 ~ - c m have been determined for the b o r o n and phosphorus reactions.A l t h o u g h the germanosilicide p h a s e s were observed to etch slowly in 1:1:5 NH4OH:H202:H20 , w h i c h is conventionally used in the self-aligned t i t a n i u m silicide process, the much h i g h e r etch rate of t i t a n i u m n i t r i d e compounds and unreacted T i provided for a self-aligned germanosilicide process. A f i r s t a n n e a l in a n i t r o g e na m b i e n t was f o u n d to be necessary to eliminate lateral silicidation over s u r r o u n d i n g oxide d u r i n g self-aligned germanosilicide formation. K e y w o r d s : Chemical v a p o r deposition (CVD), germanide, germanosilicide, LPCVD, r a p i d t h e r m a l annealing (RTA), RTCVD, self-aligned silicide, silicide, t i t a n i u m
INTRODUCTION The scaling of m e t a l oxide semiconductor field effect transistors (MOSFETs) to smaller physical dimensions causes a n increase in interconnect and parasitic series resistances of the devices.1 Silicides were initially used in microelectronic devices to lower the interconnect resistance a t the gate level u s i n g a polycide (poly-Si plus silicide) stacked gate struct u r e .2~ As device scaling is continued, increases in parasitic source/drain series resistance, due to h i g h e r s h e e t resistances, l i m i t the c u r r e n td r i v e capability of the devices.5,6A self-aligned silicide process was then proposed, w h i c h enabled the simultaneous formation of silicide on the gate, source, and d r a i n of the MOS(Received April 11, 1994; revised February 14, 1995)
FET, thus reducing both the interconnect and series resistances. 5,7 The f i r s t self-aligned silicide process was proposed by S h i b a t a et. al. u s i n g P t silicide.5Selfaligned t i t a n i u m and cobalt silicide processes quickly followed due to t h e i r l o w e r resistivity and h i g h e r temperature t h e r m a l stability necessary for post silicidation processing in comparison to o t h e r silicides.7-13 Today, self-aligned t i t a n i u m and cobalt silicides are commonplace and are used extensively in industry. However, a discrepancy s t i l l exists as to the b e t t e r choice of silicide due to various electrical and process-related tradeoffs between t h e s e two silicides.9 Recently, the use of Ge and Sil_xGeXalloys in conventional Si processing has g a i n e d interest for application in novel device structures relying on b a n d g a p engineering. These structures include high e m i t t e r efficiency heterojunction bipolar transistors, 14 high
774
Ashburn, Ozttirk, Harris, and Maher
mobility quantum-well channel MOS transistors, ~5,16 and MOS gate electrodes for threshold voltage adjustm e n t ? 7,1s O t h e r applications of Si~_ Ge~ alloys are b a s e d on the l o w e r temperature processing capabilities of the material in comparison to Si for applicat i o n s such as low t h e r m a l b u d g e t fabrication of thin film transistors, a9,2° Also, we have recently s h o w n t h a t Sil_~G% alloys c a n be deposited selectively on Si with respect to SiO2 a t temperatures as low as 600°C u s i n g r a p i d t h e r m a l chemical v a p o r deposition.21 Deposition r a t e s a t this temperature are on the o r d e r of 500A/min., w h i c h provide a process that is compatible with single wafer manufacturing. A s c a n n i n g electron micrograph (SEM) of selectively deposited Sio.7Geo.3 deposited onto exposed Si a r e a s defined in a thermally g r o w n
Fig. 1. Selective Sio.7Geo.3 deposition on exposed Si defined in SiO2 windows by rapid thermal chemical vapor deposition.
40
~
f
40
I3ooT2ooos,o7Oeo3,Si
35 [ 30
SiO 2 l a y e r is s h o w n in Fig. 1. As observed in this figure, very smooth films with excellent deposition selectivity c a n be obtained. The selectively deposited Sil_xGe ~ alloys have also been investigated as a solid diffusion source for f o r m i n g ion-implantation-damage-free, ultra-shallow junctions in Si,22 and as a sacrificial l a y e r for consumption d u r i n g silicidation for application in raised source/drain t r a n s i s t o r technologies. 23 Many of the above-mentioned structures will req u i r e low resistivity contacts similar to t h o s e obt a i n e d from silicides. Therefore, a n understanding of reactions between m e t a l s and Sil_xGex alloys is necessary. In o r d e rto implement t i t a n i u m germanosilicide materials in the fabrication oft h e s e devices, processing issues concerning the reactions of T i with doped Si~_ Ge alloys and the effect of various wet chemicals on the germanosilicide need to be investigated. For some device applications, i m p u r i t i e s m a y be incorporated into the Si~_xGe~ alloys to minimize the resistivity of the material as is done in Si. We have previously reported on the solid s t a t e reactions of T i with intrinsic Si1_~Gex alloys;24,25 however, the incorp o r a t i o n of d o p a n t s in high concentrations m a y a l t e r the reactions. Also, wet chemical e t c h i n g is used in microelectronics fabrication in a v a r i e t y of processes. These chemical etchants include NH4OH:H202:H20, w h i c h is commonly used for cleaning or in the selfaligned silicide process. In this paper, we p r e s e n t r e s u l t s on the s h e e t resistance characteristics and m i n i m u m resistivity p h a s e formation for solid s t a t e reactions of T i with b o r o n and phosphorus d o p e d Sio.TGeo. 3alloys. Also, we have identified a self-aligned germanosilicide process for forming self-aligned cont a c t s to Si0.TGeo.3 alloys.
L___
intrinsic Si0.TGe0.3 RTA: 10 seconds ~ Ar Ambient
- ~
~
25
"11" -
' \
Nitrogen Ambient
20
N 20
,= 10
@=lxl016 cm-2 boron E=10 keV DopantAnneal." 950 °C, 1 hr.
5 I
0 a
20 s e c o n d s
[
, .o , ~ ' j
"~ 25
RTA:
35
RTA: 2 0 s e c o n d s "
I'Nitrogen A m b i e n t
3 0 0 .~ T i / 2 0 0 0 .~ Si0.7Ge0.3/Si
I
I
i
I
I
I
i
,
,
,= 10
i (3-^~ - ' O .
I
,
.
,
O=lxl016 cm-2 phosphorus "~ E=20 keV \ ,, , JD Dopant Anneal: ~ ~ , 1000°C'10sec" L : _~
r~
O .t
' i
5
~ .
200 400 600 800 RTA Temperature (°C)
.
Ar Ambient 0
,
,
0
1000
,
~
I
,
,
,
I
,
,
t
~ I
,
~ ,
,
I
,
200 400 600 800 RTA Temperature (°C)
,
,
1000
b
Fig. 2. Sheet resistance as a function of R T A temperature in both argon and nitrogen ambients for solids t a t e reactions of T i with (a) boron doped Sio.TGeo. 3, and (b) phosphorus doped Sio.TGeo 3.
A Study of Self-Aligned Formation of C54 Ti(Sil_yGey) 2 to p+ and n ÷ S i o . T G e o . 3 Alloys Using RTA
775
EXPERIMENTAL Intrinsic, four-inch, Si wafers of (100) orientation were used in this experiment. U n p a t t e r n e d w a f e r samples were used to investigate the s h e e t resistance behavior of the reactions between T i and doped Sio.7Geo. 3 alloys. P a t t e r n e dw a f e r samples were used to investigate a self-aligned germanosilicide process. A 4 0 0 0 At h i c k oxide l a y e r was t h e r m a l l y g r o w n onto some of the wafers following a n RCA clean. Various p a t t e r n s were then formed in the oxide, including v a r y i n g dimension l i n e s and spaces, circles, and regions for s h e e t resistance measurements. Both the p a t t e r n e d wafers and u n p a t t e r n e d wafers were then exposed to a n RCA clean p r i o r to Si0.7Geo. 3 deposition. E a c h w a f e r was then individually subjected to a 10 s dip in a 5:1 r a t i o of H 2 0 : H F followed by a 30 s deionized (DI) w a t e r r i n s e immediately before loading into the Sil_xGex deposition system. Depositions were performed in a L E I S KTM r a p i d t h e r m a l processor used in the r a p i d t h e r m a l chemical v a p o r deposition mode. A detailed description of the reactor c a n be f o u n d in a previous publicationY6 Intrinsic Sio.TGeo.3 alloy depositions (2500A thick) were performed a t a pressure of approximately 4 Torr u s i n g SiH2C12 and 8% GeH4 premixed in a H2 carrier gas. Subsequent to Si0.TGeo. 3 deposition, d o p a n t s were i m p l a n t e d into the wafers with a dose of 1 × 10TM cm-2, and energies of 10 and 20 keV for b o r o n and phosphorus, respectively. A 1000A t h i c k low temperature oxide was then deposited onto the substrates to minimize d o p a n t outdiffusion. An a n n e a lwas then performed to uniformly distribute the d o p a n t s t h o u g h o u t the polycrystalline Sio.TGeo. ~alloy.22A 950°C, 1 h furnace a n n e a l was used for b o r o n implanted wafers and a 1000°C, 10 s r a p i d t h e r m a l a n n e a l was used for phosphorus i m p l a n t e d wafers. E a c h wafer was then exposed to a backdoor etch, consisting of NH4F/NHaH2PO4/I-I20, immedia t e l y p r i o r to t i t a n i u m deposition to remove interfacial oxide. A b l a n k e t l a y e r of T i with a thickness of 300A was then deposited onto the substrates by evaporation in a resistively h e a t e d Balzaders TM BAK 760 evaporation system. The wafers were then diced into 1 × 1 cm samples and annealed in a n AG Associates Heatpulse TM 210T r a p i d t h e r m a l annealer. Ann e a l s were performed in e i t h e r a n a r g o n or n i t r o g e n a m b i e n t a t atmospheric pressure. The temperature was monitored d u r i n g the a n n e a l with a 2 × 2 cm thermocouple chip located in the vicinity of the samples. The samples were annealed on the backside of a four inch silicon wafer. This m e t h o d has been s h o w n to accurately m i m i c w h o l e wafer anneals for i m p l a n t activationY We believe t h a t in our case the e r r o r due to the different absorptivity of T i and its compounds will be minimized due to t h e r m a l conduct i o n . Nevertheless, some e r r o r is expected and hence the temperatures referred to s h o u l d be t r e a t e d not as absolute, b u t as relative numbers. Characterization of the films was accomplished u s i n g a MAGNETRON TM I n s t r u m e n t s model M - 7 0 0 resistivity/conductivity test system, a R i g a k uTM x - r a y diffractometer
C54 Ti(SiGe) 2 (022) Si (2OO) C54 Ti(SiGe) 2 (3 ~ )
~
,tf
PhosphorusDoped 5i0.7Ge0.3 Boron Doped Sio7Geo3 intrinsic
30 32 34 36 38 40 42 44 46 48 50 2 - T h e t a (degrees) Fig. 3. X-ray diffraction spectraof minimum resistivity C54 Ti(Si,~,Ge~)2 p h a s e formation for solid state reactions of Ti with undoped,bor6n doped, and phosphorus doped SioTGeo3 alloys.
( C u K a radiation), Phillips T M 430T transmission electron microscope (TEM) operating a t 300 kV, Nomarski optical imaging, and a JEOL TM 6400 field emission scanning electron microscope (FE SEM). EXPERIMENTAL RESULTS AND DISCUSSION Solid State Reactions of Ti with Boron and Phosphorus D o p e d Si0.7Ge0.3 Alloys S h e e t resistance characteristics p l o t t e d as a function of annealing temperature for solid s t a t e silicide formation serve as indicators of temperatures a t w h i c h various phases form d u r i n g the reactions. P l o t s of the s h e e t resistance as a function of RTA temperature for solid s t a t e reactions of T i with b o r o n and phosphorus doped Sio.7Geo. 3 alloys are s h o w n in Figs. 2a and 2b, respectively. The anneals were performed in a n a r g o n or n i t r o g e na m b i e n t for 20 s. Included for comparison in Fig. 2a is the s h e e t resistance characteristics for reactions between T i and intrinsic Sio.7Geo. 3. As observed in Fig. 2a, the t r a n s i t i o n to the m i n i m u m s h e e t resistance p h a s e for b o r o n doped alloys annealed in e i t h e ra r g o n or nitrogen occurs a f t e ra n RTA temperature of approximately 700°C for 20 s. However, the transition to the m i n i m u m s h e e t resistance p h a s e for t i t a n i u m reactions with the intrinsic alloy occurs a t approximately 100°C l o w e r in temperature for a s h o r t e r RTA d u r a t i o n of only 10 s. These r e s u l t s suggest t h a t the g r o w t h rate of the m i n i m u m resist i v i t y germanosilicide p h a s eis l o w e ron heavily b o r o n doped alloys t h a n on u n d o p e d alloys. Annealing in n i t r o g e n is observed to cause a n increase of approxim a t e l y 40% in the m i n i m u m s h e e t resistance v a l u e obtained. In a previous report, we have s h o w n that a surface t i t a n i u m oxynitride f o r m s on the surface of
776
Ashburn, Ozttirk, Harris, and Maher
Si 0.1 um a
.iii
S! 0.1vm b
Fig. 4. Cross-sectional T E M micrographs of the C54 Ti(Sil_yG%) 2 reacted layersformed on (a) boron doped Sio 7Geo. 3 and (b) phosphorus doped Sio.TGeo. 3 alloys.
the germanosilicide when nitrogen is w i t h i n the reacting ambient. 24 Therefore, the increase in s h e e t resist a n c e observed here is a t t r i b u t e d to the competing reactions of t i t a n i u m germanosilicide and t i t a n i u m n i t r i d e formation, w h i c h causes the formation of a t h i n n e r (and thus h i g h e r s h e e t resistance) germanosilicide layer. The reactions with phosphorus d o p e d Sio.TGeo. 3 alloys, s h o w n in Fig. 2b, are s i m i l a rto t h o s e with b o r o n d o p e d alloys. A g a i n this suggests t h a t the germanosilicide reaction rate is decreased for both t i t a n i u m reactions with b o r o n and phosphorus d o p e d alloys. O t h e r investigators have observed t h a t the g r o w t h rate ofTiSi2 was lower on heavily phosphorus or arsenic d o p e d Si;28 however, no such dependence of g r o w t h rate on heavily b o r o n doped Si has been observed-29,3° The increases in s h e e t resistances
observed in Figs. 2a and 2b a f t e r a 9 0 0 ° C a n n e a l are a t t r i b u t e dto the o n s e tof agglomeration fort h e s e thin film reactions. In a previous report, we investigated the t h e r m a l stability of the m i n i m u m resistivity germanosilicide p h a s e and f o u n d that agglomeration of the germanosilicide and not the formation of a new p h a s e causes the s h e e t resistance to increase a f t e r high temperature anneals. 25 U s i n g x - r a y diffraction analysis, we have analyzed the p h a s e formation t h a t corresponds to the m i n i m u m s h e e t resistance m a t e r i a l indicated in Figs. 2a and 2b. The diffraction spectra as a function of the B r a g g diffraction angle, presented in t e r m s of 2-theta, are s h o w n in Fig. 3. Spectra obtained for reactions of T i with intrinsic, b o r o n doped, and phosphorus d o p e d Sio.TGeo ~ alloys are shown. The m i n i m u m resistivity
A Study of Self-Aligned Formation of C54 Ti(Sil_yGey) 2 to p+ and n÷ Sio.7Geo.3 Alloys Using RTA p h a s e that forms for t i t a n i u m reactions with Sil_xGex alloys, with Ge concentrations r a n g i n g from 0 to 100%, has been determined in a previous r e p o r t to be Ti(Si I • G e )2 with the C54 structure. 24 The characteri s t i c C-~4 (311) and (022) p e a k s were observed to s h i f t from t h o s e of C54 TiSi2 to C54 TiG% as the germanium concentration w i t h i n the reacting alloy was increased. The C54 Ti(Sil_yGey) 2p h a s e also forms for t i t a n i u m reactions with b o r o n and phosphorus doped Sio7Geo.3 alloys with characteristic C54 p e a k s corresponding to t h o s e of intrinsic SioTGeo.3 reactions as s h o w n in Fig. 3. We have used cross-sectional TEM to analyze the g r a i n structures of the C54 phases formed on both b o r o n and phosphorus d o p e d Sio7Geo.3 alloys. The r e s u l t s are s h o w n in Fig. 4 for samples annealed in a n a r g o n a m b i e n t a t 800°C. Figures 4a and 4b corres p o n d to b o r o n and phosphorus d o p e d sample r e a c tions, respectively. A uniformly reacted germanosilicide r e g i o nis observed in e a c h of the reactions with thicknesses of approximately 550A. This thickness is slightly less than t h a t observed for reactions with intrinsic a l l o y reactions (630A).24 On the surface of e a c h of the C54 reacted layers is a t h i n (approxim a t e l y 1 0 0 k ) layer, w h i c h was also observed in int r i n s i c alloy reactions annealed in argon, u U s i n g A u g e r surface s c a n analysis, this l a y e r was determ i n e d to be a t i t a n i u m oxynitride. The formation of this oxynitride l a y e ris a t t r i b u t e d to n i t r o g e nb e i n g a m a j o r i m p u r i t y in the a r g o n annealing gas used d u r i n g the r a p i d t h e r m a l anneal. From s h e e t resistance measurements t o g e t h e r with thicknesses obt a i n e d from the cross-sectional TEM images, a resist i v i t y of approximately 22 ~ 2 - c m was determined for e a c h of the reactions on b o r o n and phosphorus doped alloys. Lateral g r a i n dimensions for e a c h of the r e a c t i o n s were observed to be on the o r d e rof 0.3 pm, w h i c h are comparable to the l a t e r a l g r a i n dimensions for u n d o p e d alloy for reactions (0.4 ~m).24 Self-AlignedGermanosilicide Formation T i t a n i u m reacts with oxide layers to form t i t a n i u m oxides a t temperatures g r e a t e r t h a n approximately 700°C. 7 In the self-aligned t i t a n i u m silicide process, t i t a n i u m is f i r s tb l a n k e t deposited onto the substrate with p a t t e r n s defined in SiO2. Due to the t i t a n i u m b e i n g in contact with the SiO2 on the substrate, a twostep a n n e a l process is r e q u i r e d for the self-aligned t i t a n i u m silicide process. A f i r s ta n n e a l is performed a t a temperature lower than 700°C (typically 600°C) for a time sufficient to o b t a i n a u n i f o r m and complete reaction of T i with exposed Si. 7 This f i r s t a n n e a l is also performed in a nitrogen a m b i e n t to eliminate l a t e r a l silicidation t h a t c a n occur due to Si b e i n g a d o m i n a n t diffusing species in the t i t a n i u m silicide process.9 Unreacted T i and t i t a n i u m n i t r i d e p h a s e s are then selectively removed from the p a t t e r n e d SiO2 and surface of the silicide formed on exposed Si. This is typically performed u s i n g wet chemical etching in e i t h e r 1 : 1 : 5 N H 4 O H : H 2 0 2 H 2 0 a t 21°C or 5:1 H2SOt:H202 a t 100°C.9 Following the wet etch, the
777 substrates are annealed a t h i g h e r temperatures ( 7 0 0 900°C) to convert the reacted silicide into the m i n i mum resistivity C54 TiSi2 phase.7 For the case of t i t a n i u m germanosilicide contacts to Sil_xGe~ alloys, the substrates are also likely to be p a t t e r n e d with a n SiO2 layer. Therefore, a two-step a n n e a l process is also necessary for self-aligned germanosilicide form a t i o n to eliminate reaction of T i with SiO2. In addit i o n , we have observed t h a t l a t e r a l silicidation over s u r r o u n d i n g oxide regions occurs a f t e r the f i r s t low temperature (600°C, 120 s RTA) a n n e a l when performed in a n a r g o na m b i e n t as s h o w n in the Nomarski optical i m a g e of Fig. 5a). As s h o w n in this figure, l a t e r a l silicidation over the oxide c a n easily lead to electrical s h o r t i n g of adjacent lines. This lateral germanosilicide formation is suggestive of Si a n d / o r Ge or both b e i n g d o m i n a n t diffusing species d u r i n g the t i t a n i u m germanosilicide reaction, w h i c his currently b e i n g investigated. However, as i n the case with the self-aligned t i t a n i u m silicide process, annealing in a nitrogen a m b i e n tc a n eliminate this lateral silicidation as s h o w n in Fig. 5b, due to g r a i n b o u n d a r y stuffing of
b Fig. 5. Nomarski optical images of selective germanosilicide formation on p÷ Sio.~Geo. 3a f t e r a 6 0 0 ° C , 120 s RTA in (a) an argon ambient, and (b) a nitrogen ambient.
778
Ashburn, 0zttirk, Harris, and Maher
140 ~ T /
overSi02]''lx~ -.Completely--[ - removed / -after 6min/
120 100
60
~ ,1= r~
40
21 °C
boronimplanted material
[
intermediate germanosilicide phase overSi07Geo3
~5~TiSi2Phase
y
80
"~
1:1:5 NH4OH:H202:H20
20 ~ I
0" 0
2
4 6 8 10 Etch T i m e ( m i n u t e s )
140
12
1:1:5
NH4OH:H202:H20
~00,~ Ti 21 °C overSi02 ~ I phosphorus implanted C;?Ploe~y ? material
120 ~100~ ~
80
"~ 6 0
¢l~ ~
intermediate germanosilicide phase overSio.7Geo.3
40
20 2
-
o ?,,, ..., 0 b
2
4 6 8 10 Etch Time (minutes)
12
Fig. 6. Sheet resistance as a function of etch time in a solution of 1:1:5 NH4OH:H202:H20 of titanium compounds f o r m e d over SiO2 and of various germanosilicide phases f o r m e d over Sio.TGeo. 3a f t e r an anneal in a nitrogen ambient f o r (a) boron implanted substrates, and (b) phosphorus implanted substrates.
the unreacted T i l a y e rwith nitrogen. 9The a n n e a lwas performed for the same time and temperature (600°C, 120 s RTA) as the a r g o n annealed sample in Fig. 5a, with the only difference b e i n g the gas used in the annealing ambient. A f t e r the f i r s ta n n e a l performed a t 6 0 0 ° C for 120 s, s h e e t resistance measurements t o g e t h e r with x - r a y diffraction analysis indicated t h a t both the h i g h e r r e s i s t i v i t y i n t e r m e d i a t e p h a s e a n d t h e C54 Ti(Sia yGey)2 were formed. Therefore, to investigate selective removal of unreacted T i and t i t a n i u m ni-
t r i d e phases with respect to the germanosilicide m a terial, it is necessary to determine the selectivity with respect to both the intermediate and m i n i m u m resist i v i t y phases. P a t t e r n e d w a f e r samples were therefore annealed a t e i t h e r 500 or 600°C in a nitrogen a m b i e n t to minimize the reaction of T i with SiO2, and to form e i t h e r intermediate or m i n i m u m resistivity C54 germanosilicide p h a s e s as determined by s h e e t resistance measurements. These samples were then exposed to NH4OH:H202:H20 i n a r a t i o of 1:1:5 a t a temperature of 2 I°C for v a r y i n gt i m e s a f t e rw h i c h the s h e e t resistances were a g a i n measured. Due to the s h e e t resistance b e i n g proportional to the inverse of the m a t e r i a lthickness, etching of the germanosilicide c a n be monitored. The r e s u l t s obtained for intermediate and m i n i m u m u m r e s i s t i v i t y germanosilicide p h a s e s formed on both b o r o n and phosphorus doped Sio.vGeo3 alloys are s h o w n in Fig. 6. As s h o w n in Fig. 6a, removal of the 300A ofunreacted T i and t i t a n i u m n i t r i d e compounds over the SiO2 is achieved a f t e r a 6 min etch. The intermediate germanosilicide p h a s e (approximately 20 f~/sq) b e g i n s etching a f t e r this 6 min d u r a t i o n w h i l e the m i n i m u m resistivity C54 Ti(Si1. Gey)2 p h a s e is more r e s i s t a n t to the solution with s-l~ighte t c h i n g observed for t i m e s i n the r a n g e of 8 to 12 min. Therefore, selective removal ofunreacted T i and t i t a n i u m n i t r i d e p h a s e s with respect to the germanosilicide is possible; however, care m u s t be t a k e n due to the process not b e i n g completely selective. For comparison, the s h e e t resistance o f C 5 4 TiSi2 is s h o w n as a function of etch time indicating that no e t c h i n g occurred d u r i n g the 12 min investigated. S i m i l a r r e s u l t s are s h o w n in Fig. 6b for p h o p h o r u s d o p e d Sio.7Geo. 3 alloys. I t is interesting to note t h a t removal of the unreacted T i and t i t a n i u m n i t r i d e compounds from the SiO2 t a k e s 10 min in this c a s e , w h i c h is 4 min l o n g e r t h a n the time required to remove the material from b o r o n d o p e d wafers. The intermediate and m i n i m u m resistivity C54Ti(Sil Gev)2 p h a s e s are observed to b e g i ne t c h i n gfor t i m e s ex-'cee~ling approximately 8 min, t h u s allowing for a selfaligned germanosilicide process on phosphorus d o p e d Sio.7Geo. 3 alloys. As s h o w n in Figs. 6a and 6b, t i t a n i u m silicide is completely selective to the e t c h a n t used in this study. On the o t h e rh a n d , t h e r e e x i s t s a n incubation time on the o r d e r of 10 min for Ti(Si1 yGev)2 etching. We believe t h a t the presence of this incubation time c a n be a t t r i b u t e d to the t i t a n i u m oxynitride l a y e r observed on Ti(Si~ yGey)2. As discussed above, the oxyn i t r i d e forms du-~ to oxygen a t the i n i t i a l Ti/SixGel_ x (or Ti/Si) interface as well as oxygen in Si Ge~_x (or Si) and nitrogen t h a t exists as a contaminant in the a r g o n gas used for annealing. The oxynitride has a l a r g e resistivity due to the presence of oxygen. As a result, d u r i n g four-point p r o b e measurements, cond u c t i o n is primarily t h r o u g h the germanosilicide or silicide layers. In Figs. 6a and 6b, the s h e e t resistances of the germanosilicide do not initially increase due to the chemical e t c h i n g of the surface oxynitride w h i c h is not the p r i m a r y path for con-
A Study of Self-Aligned Formation of C54 Ti(Sil_yGey) 2 to p+ and n ÷ Si0.7Geo.3 Alloys Using RTA duction. A f t e r the oxynitride is removed with the etchant, the germanosilicide etches slowly (not completely selective to the etchant). The silicide, however, is completely selective to the e t c h a n t and therefore when the oxynitride is removed from the surface of the silicide, etching stops, and the p r i m a r y conductive silicide l a y e r remains. U s i n g the measured s h e e t resistance d a t a , we c a n estimate the etch rate of the germanosilicide l a y e r a f t e r the low temperature anneal. A s s u m i n g a resist i v i t y of 150 ~ 2 - c m for the intermediate germanosilicide p h a s e and u s i n g the w o r s t case condition in Fig. 6a, one c a n calculate the germanosilicide thicknesses for the etch time (x-axis) a t 6 and 10 min anti determine the etch rate to be approximately 120A/ min. Similarly, the etch rate of the C54 germanosilicide is approximately 30A/min. The etch rate of the TiN and unreacted T i over the oxide is expected to be on the o r d e r of 300PJ6 min. = 50~Jmin. This suggests t h a t the intermediate p h a s e etches faster t h a n the T i and TiN w h i l e the C54 p h a s e etches with a comparable r a t e . However, because the germanosilicide is t h i c k e r than the T i and TiN, complete etching c a n not take place. Following selective removal of the unreacted T i and t i t a n i u m n i t r i d e phases, h i g h e r temperature r a p i d t h e r m a l a n n e a l a t 750°C for 15 s is performed to transform any non-C54 p h a s e into the m i n i m u m resistivity phase. A SEM micrograph of self-aligned C54 Ti(Si1-y Ge),)2 formation on selectively deposited, b o r o n d o p e d Slo.7Geo3 is s h o w n i n Fig. 7. The process for the self-aligned germanosilicide formation cons i s t e d of a low temperature RTA a t 600°C for 120 s in a nitrogen ambient, followed by a 7 min etch in 1:1:5 NH4OH:H202:H20 a t 21°C to remove unreacted T i and t i t a n i u m n i t r i d e phases. A 750°C, 15 s r a p i d t h e r m a l a n n e a l then converted the m a t e r i a l into the m i n i m u m resistivity phase. As observed in Fig. 7, a relatively s m o o t h germanosilicide surface has been produced with no s i g n s ofl a t e r a l silicidation over the s u r r o u n d i n g oxide. CONCLUSIONS S o l i d s t a t e reactions of T i with Sio7Geo.~ alloys h e a v i l yd o p e d with e i t h e rb o r o n or phosphorus (dose of I x 1018 cm-2 producing a concentration of approxim a t e l y 3 x 1020 cm-~) have been investigated for application in contacting device structures u s i n gt h e s e materials. The reactions were f o u n d to be s i m i l a r to reactions with u n d o p e dSio.TGeo3 alloys with the form a t i o n of a Ti(Sil_yGey)2 m i n i m u m resistivity p h a s e with the C54 structure. Resistivities of approximately 22 ~t~-cm were determined for the reactions with both b o r o n and phosphorus alloys with l a t e r a l g r a i n dimensions on the o r d e r of 0.3 pm. Selective etching of t i t a n i u m n i t r i d e compounds and unreacted T i with respect to germanosilicide phases were investigated in a 1:1:5 NH4OH:H20:H20 solution. A l t h o u g h both the i n t e r m e d i a t e and m i n i m u m r e s i s t i v i t y C54 germanosilicide phases etch in this solution, the etch rate was very slow a l l o w i n g for a self-aligned
779
Fig. 7. Scanning electron micrograph of self-aligned germanosilicide formation on selectively deposited SioTGeo3 using a two-step RTA processof 600°C, 120 s in nitrogen and 750°C, 15 s in argon separated by selective chemical etching of titanium nitride compounds and unreacted T i in a solution of 1:1:5 NH4OH:H202:H20.
germanosilicide process. However, more selective wet etching techniques and possibly dry e t c h i n g techn i q u e s need to be developed for this process to e n s u r e repeatability. ACKNOWLEDGMENTS This work has been partially supported by the NSF Engineering Research Centers P r o g r a m t h r o u g h the C e n t e r for Advanced Electronic Materials Processing ( G r a n t CDR-8721505) and SRC Microstructures Sciences P r o g r a m( G r a n t 93-SJ-081). The a u t h o r sw o u l d also like to acknowledge Dr. W.T. L y n c h (SRC), Dr. W.C. H o l t o n (SRC), and Dr. J.J. W o r t m a n (NCSU) for valuable discussions. The a u t h o r s w o u l d also like to acknowledge J. C l a r k e ( M C N C ) and C.J. B e r r y (MCNC) for m e t a l evaporation, and J. O'Sullivan (NCSU), Dr. R. K u e h n (NCSU), and H. Taylor (NCSU) for laboratory assistance. Also, the a u t h o r sw o u l d like to t h a n k M u h s i n ~elik, Archie Lee, Xiao-Wei Ren, M a h e s h Sanganeria, K u r t Seastrand, K a t h e r i n e Violette, and I b r a h i m Ban for many helpful discussions in weekly g r o u p meetings. REFERENCES 1 . S.P. Murarka, Silicides f o r V L S I Applications (New York: Academic P r e s s , Inc., 1983). 2 . B.L. C r o w d e r a n d S. Zirinsky, I E E E Trans. Electron Dev. ED26 (4), 3 6 9 , (1974). 3 . S.P. M u r a r k a , D.B. F r a s e r , A.K. S i n h a a n d H.J. Levenstein, I E E E Trans. Electron Dev. ED-27, 1409 (1980). 4 . E.M. King a n d K.E. Gsteiger, J. Vac. Sci. Tech. A 1 (2), 614 (1983). 5 . T.Shibata, K. H i e d a , M . Sato, R. L. M . Dang a n d H. Iizuka, I E D M Tech.l Digest 647 (1981). 6 . C.M. O s b u r n , J . Electron. Mater. 1 8 , 67 (1990). 7 . M.E. Alperin, T.C. Hollaway, R.A. Haken, C.D. Gosmeyer, R.V. K a r n a u g h a n d W.D. Parmantie, I E E E Trans. Electron Dev. ED-32 (2), 141 (1985). 8 . K. Tsukamoto, T . Okamoto, M . Shimizu, T . M a t s u k a w a a n d H. Nakata, I E D M Tech. Digest 130 (1984). 9 . L. V a n d e n hove, P h D. thesis, Katholieke Universiteit,
780 Leuven, Belgium (1988). 10. L. Van den hove, R. Wolters, K. Maex, R.D. Keersmaecker and G. Declerck, J. Vac. Sci. Technol. B 4 (6), 1358 (1986). 11. L. Van den hove, R . Wolters, K. Maex, R.F.D. Keersmaecker and G.J. Declerck, IEEE T r a n s . Electron Dev. ED-34 (3), 554 (1987). 12. A.E. Morgan, E.K. Broadbent, M . Delfino, B . Coulman and D.K. Sadana, J. Electrochem. Soc. 134 (4), 925 (1987). 13. F . Yang and M . Chen, Thin Solid Films 297, 75 (1992). 14. J.C. Sturm, E.J. P r i n z and C.W. Magee, IEEE Electron Dev. L e t t . EDL-12, 303 (1991). 15. E. Murakami, K. Nakagawa, H. Etoh, A. Nishida and M . Miyao, IEDM Tech. Digest 375 (1990). 16. P.M. Garone, V. Venkataraman and J.C. Sturm, IEDM Tech. Digest 383 (1990). 17. T.J. King, J.R. Pfiester, J.D. S h o t t , J.P. McVittie and K.C. Saraswat, IEDM Tech. Digest 253 (1990). 18. D.T. Grider, M.C. ()zttirk, M . Sanganeria and J.J. Wortman, Electron. M a t e r . Conf. (1991). 19. T.J. King, K.C. Saraswat and J.R. Pfiester, IEEE Electron Dev. L e t t . EDL-12, 584 (1991).
Ashburn, 0zttirk, Harris, and M a h e r 20. T.J. K i n g a n dK.C. Saraswat, IEDMTech. Digest 567 (1991). 21. Y.L. Zhong, M.C. 0zttirk, D.T. Grider, J.J. W o r t m a n and M.A. Littlejohn,Appl. Phys. L e t t . 57 (20), 2092 (1990). 22. D.T. Grider, P h D. thesis, N o r t h Carolina S t a t e University (1993). 23. S . P . Ashburn, D.T. Grider and M.C. ()ztiirk, to be published (1994). 24. S.P. Ashburn, M.C. {)zttirk, G. Harris, D.M. M a h e r , D.B. Aldrich and R.J. Nemanich, to be published (1994). 25. S . P . Ashburn, M.C. ()zttirk, G. H a r r i s and D.M.M a h e r , to be published (1994). 26. S.P. Ashburn, M.C. ()zttirk, J.J. Wortman, G. Harris, J . Honeycutt and D.M. M a h e r , J. Electron. Mater. 21 (1), 81 (1992). 27. G.A. Ruggles, S.N.H o n g ,J.J. Wortman, F.Y. Sorrell and M.C. ()ztiirk, J. Vac. Sci. Technol. B 8 , 122 (1990). 28. R. Beyers, D. Coulman and P . Merchant, J. Appl. Phys. 61 (11), 5110 (1987). 29. T.P. Chow, W. Katz and G. Smith,Appl. Phys. L e t t . 46 (1), 41 (1985). 30. T.P. C h o w , W. K a t z , R. Goehner and G. Smith, J. Electrochem. Soc. August, 1985, 1914 (1985).