Differential Hall analysis of ultrashallow carrier profiles
using X-ray photoelectron spectroscopy
for nanometer depth resolution
Yu-Ting Ling,1,a) Wan-Ting Su,1 Tun-Wen Pi,2 and Ruey-Dar Chang1
1
2
Chang Gung University, 259 Wen-Hwa 1st Road, Kwei-Shan Tao-Yuan 33302, Taiwan
National Synchrotron Radiation Research Center, 101 Hsin-Ann Road, Hsinchu Science Park,
Hsinchu 30076, Taiwan
Abstract. In this study, differential Hall measurement (DHM) was developed to measure the carrier profiles in
phosphorus doped ultrashallow junctions (USJs). Experiments using uniform phosphorus profiles in silicon on insulator
(SOI) wafers demonstrated that the growth rate of the native oxide strongly depends on the phosphorus doping level.
Therefore, the thickness of native oxide was monitored by X-ray photoelectron spectroscopy (XPS) to achieve
nanometer depth resolution during DHM. The DHM method was applied to investigate the deactivation of phosphorus in
laser annealed USJs. The DHM results indicate carrier profile redistribution near the surface due to uphill diffusion
caused by phosphorus deactivation.
Keywords: Differential Hall measurement (DHM), X-ray photoelectron spectroscopy (XPS), ultrashallow junctions.
PACS: 61.72.uf, 61.72.sh, 72.20.My, 82.80.Pv
INTRODUCTION
EXPERIMENTS
The series resistance of metal-oxide-semiconductor
field effect transistors (MOSFETs) is critical to the
speed of integrated circuits. Therefore, heavily doped
ultrashallow junctions (USJ) are required to achieve
low resistance in source and drain regions [1]. Rapid
annealing at high temperatures is used in dopant
activation. However, dopant deactivation occurs
during subsequent low-temperature processes.
Influence of dopant deactivation on the carrier density
needs to be characterized. Differential Hall
measurement (DHM) has been used to measure active
dopant profiles. A continuous anodic oxidation
technique (CAOT) has been developed for DHM
[2][3]. Hall measurement is performed after the
removal of the anodic oxidized silicon. Carrier profiles
are then calculated based on the residual carrier
concentration obtained from each Hall measurement.
Native oxide was also suggested for etching silicon
layers [4]. However, an average oxidation rate is
usually assumed for DHM. X-ray photoelectron
spectroscopy (XPS) provides nanometer resolution in
measuring the thickness of native oxide. In this study,
we developed a DHM technique using XPS to achieve
nanometer resolution for active dopant profiles. With
the developed DHM technique, carrier redistribution
due to the uphill diffusion caused by phosphorus
deactivation was observed near the surface in
ultrashallow junctions.
In order to examine the dependence of doping level
on the growth rate of native oxide, (100) oriented ptype silicon-on-insulator (SOI) wafers were used. SOI
wafers were first oxidized to form a screen oxide layer
with a thickness of 15 nm. Then phosphorus doping
was introduced by ion implantation with doses of
1.5×1015, 5×1015 and 1×1016 cm-2 at 50 keV. Samples
were annealed at 1100 °C for 2 hr to obtain uniform
profiles and eliminate implantation damages. Rapid
thermal annealing (RTA) at 1100 °C for 10 s was used
to activate the phosphorus. After the screen oxide was
removed by dilute HF, samples were exposed to air
from 1 hr to 21 days with a humidity of around 40±5
% to grow a native oxide layer. The thickness of the
native oxide was determined by XPS with the light
source from synchrotron radiation. Phosphorus
concentration was analyzed by secondary ion mass
spectroscopy (SIMS).
The experimental procedure for DHM analysis of
phosphorus deactivation in USJs is illustrated in Fig. 1.
(100) oriented p-type silicon wafers were
preamorphized by germanium with a dose of 1×1015
cm-2 and at the energy of 20 keV. Phosphorus was then
implanted into the amorphous layer at 2 keV with a
dose of 1×1015 cm-2. Laser annealing was carried out
for dopant activation. Subsequent furnace annealing at
500 °C for 640 min was performed to cause
deactivation of phosphorus. Samples were first
analyzed by Hall measurement. Then the thickness of
native oxide was measured using XPS. After the native
oxide was removed by dilute HF, samples were
exposed to air for 24 hr to grow a new layer of native
oxide. The process of measurement and oxidation
repeated until the sheet resistance increased
significantly. The phosphorus signal in XPS spectrums
was also used to monitor the redistribution of
phosphorus near the surface.
beginning of exposure. The growth rate of native oxide
depends on the doping level. Higher doping level
leaded to higher growth rate. After exposure to the air
for more than one day, the oxide thickness almost
saturated. The increase of the oxide thickness was less
than 0.5 nm with more exposure time for 7 days.
Because of the saturation of oxide thickness, the
exposure time for the growth of native oxide was set to
be 24 hr for DHM analysis.
Oxide Thickness (nm)
3
FIGURE 1. Experimental procedure of DHM for laser
annealed USJ samples with and without phosphorus
deactivation.
RESULTS & DISCUSSION
2
be extracted as following:
I Si
=
(
4x1020 cm-3
2x1020 cm-3
6x1019 cm-3
SOI
2
1
0
0
7
14
21
Exposure Time (days)
FIGURE 2. Growth rate of native oxide in uniform doped
SOI samples with various phosphorus doping levels.
Figure 2 shows the thickness of native oxide
measured by XPS for SOI samples that were exposed
to the air at room temperature for different times. The
phosphorus profiles in SOI wafers are uniformed and
highly activated after annealing at 1100 °C. Based on
the XPS signals of Si 2p, the oxide thickness tSiO can
I SiO2
P concentration
)
nSiO2 λSiO2 1 − exp  −tSiO2 / λSiO2 cos θ 
, (1)
⋅
⋅
nSi λSi
exp  −tSi / ( λSi cos θ )
where I SiO and I Si refer to the signals from the oxide
2
layer and the silicon substrate. Parameters nSiO and
2
nSi are densities of silicon dioxide and crystalline
silicon. Inelastic mean free path (IMFP) of
photoelectrons escaping from silicon dioxide and
crystalline silicon is represented by λSiO and λSi ,
2
respectively. The IMFP value is a function of kinetic
energy. The XPS light source used for SOI samples
was 400 eV obtained from synchrotron radiation. The
IMFP values for silicon and silicon dioxide are 1.01
and 1.54 nm, respectively. Parameter θ is the tack-off
angle from the axis normal to the sample surface. A
rapid growth rate of native oxide was observed at the
Figure 3 shows the thickness of native oxide after
each oxide etching step during differential Hall
measurement of laser annealed phosphorus USJs. The
XPS light source was 1486.6 eV with Al mono. The
IMFP values for silicon and silicon dioxide are 2.16
and 3.3 nm, respectively. The initial oxide thickness
was thicker than that during DHM because wafers had
been exposed to air for a long time. After deactivation
at 500 °C, the initial oxide became thicker due to the
additional annealing. However, the oxide thickness
during DHM was less than that in samples before
deactivation. This implies that the carrier density
indeed affected the growth rate of native oxide on
heavily phosphorus doped USJs. The oxide thickness
was used to estimate the consumption of the silicon
thickness after the growth of native oxide. The depth
from the surface after each etching step during DHM
was therefore obtained. The DHM results in Figure 4
show sheet resistance, sheet concentration and
mobility as a function of the etching depth for USJ
samples with and without deactivation annealing. The
sheet carrier concentration decreased after deactivation
annealing. This leaded the increase of the Hall
mobility due to the reduction of the ionized impurity
scattering. When the sheet concentration of carriers
decreased to1×1013 cm-2, the sheet resistance increased
Laser
3.0
o
Laser + 500 C deactivation
2.5
2.0
1.5
1.0
-2
0.0
2.5
init. 1st 2nd 3rd 4th 5th 6th 7th
Grown Oxide Layers
FIGURE 3. Native oxide thickness measured by XPS
during each DHM step for phosphorus USJ samples before
and after deactivation.
significantly. The Hall mobility obtained from the last
Hall measurement was contributed by the tail profile
after high concentration phosphorus near surface was
removed. Therefore, a high Hall mobility was
observed. For Hall measurement, the correlation
between sheet resistance ρ s and sheet Hall coefficient
1.0
0.5
0.0
(2)
(3)
where σ s is sheet conductivity and N s is sheet
concentration. Parameters q and µc refer to
electronic charge and conductivity mobility,
respectively. Effect of circulating currents can be
described by the ratio r = µh / µc where µh is Hall
mobility. Based on two Hall measurements before and
after the ith removed oxide layer, conductivity
mobility µci and carrier concentration ni can be
calculated by:
2
 1   ∆ ( Rhs ⋅ σ s )i ∆ (σ s )i  ,
/
µci =   ⋅
∆ti
∆ti 
 r  

(
 ∆ σ  ∆ Rhs ⋅ σ s 2
and n =  r  ⋅  ( s )i  /
i
 
∆ti
 q   ∆ti 
2
(4)
)  .
i


(5)
where ∆ti is the depth of silicon consumed by
oxidation [5][6]. For USJs, assumption of an average
etching depth was not suitable for such abrupt profile.
10
0
1
2
3
4
5
Depth (nm)
×1014
Laser
o
Laser + 500 C deactivation
8
6
4
2
0
0
1
2
3
4
5
Depth (nm)
300
2
ρ s = 1/ σ s = 1/ q ⋅ N s ⋅ µc ,
and Rhs = r / q ⋅ N s .
Laser
o
Laser + 500 C deactivation
1.5
(b)
Mobility (cm /V-s)
Rhs can be described as:
×104
2.0
(a)
0.5
Sheet Concentration (cm )
Oxide Thickness (nm)
Sheet Resistance (ohm/sqr.)
3.5
Laser
o
Laser + 500 C deactivation
250
200
150
100
50
0
(c)
0
1
2
3
4
5
Depth (nm)
FIGURE 4. DHM results showing sheet resistance, sheet
carrier concentration and Hall mobility for phosphorous USJ
before and after deactivation.
XPS provided an accurate depth resolution around 0.5
nm.
According to the DHM results using XPS for
monitoring the oxide thickness, the carrier
concentration as a function of the depth are shown in
Fig. 5. An activation level around 1×1021 cm-3 was
obtained after laser annealing near the surface of the
phosphorus junction. After deactivation, the carrier
concentration of the whole profile decayed. Carrier
profiles before and after deactivation decreased
sharply at 4.5 and 3.5 nm, respectively. The sharp
decrease of the carrier profile is consistent with the
projected range of phosphorus after implantation.
However, the difference in the depth profile is around
1 nm between the carrier and chemical profiles after
annealing. The shape of carrier profile remained
similar except for a dip near the surface. This suggests
some profile redistribution near the surface during
deactivation annealing at 500 °C.
Laser
o
Laser + 500 C deactivation
22
P 2p
init.
o
Laser + 500 C deactivation
21
10
20
Intensity (a.u.)
Laser
-3
Carrier Concentration (cm )
10
1st
2nd
10
19
10
0
1
2
3
4
5
Depth (nm)
FIGURE 5. Carrier concentrations obtained by DHM using
XPS analysis for samples before and after deactivation.
Figure 6 compares normalized XPS signals before and
after deactivation. Similar background Si plasma
signals were obtained for all profiles. Clear P 2p
signals were only observed in samples with initial
oxide. This indicates that most of the phosphorus is
near the surface. Interestingly, evident increase of the
intensity of the P 2p signal was observed after
deactivation. The increase of the XPS signal suggests a
high chemical concentration due to phosphorus uphill
diffusion to the surface. This confirms evident profile
redistribution as that observed in DHM results during
deactivation annealing 500 °C. Point defect generation
and enhanced tail diffusion during phosphorus
deactivation was reported [7]. However, DHM and
XPS results indicate that the redistribution of active
and chemical profiles at the peak region cannot be
ignored during phosphorus deactivation even at 500 °C.
CONCLUSION
A DHM technique was developed using XPS to
monitor the thickness of native oxide grown on
phosphorus USJs. The growth rate of native oxide on
uniform doped SOI samples depends on the doping
level of phosphorus. The initial growth rate was rapid
while the thickness saturated for long exposure times.
Therefore, the thickness of native oxide was measured
by XPS to obtain nanometer depth resolution in each
etching step during DHM analysis. The DHM method
was applied to measure phosphorus deactivation in
laser annealed USJs. DHM results indicated
redistribution of carrier profile near the surface due to
phosphorus deactivation. The redistribution is caused
140
136
132
128
Binding energy (eV)
FIGURE 6. Comparison of normalized XPS signals for
samples with and without deactivation with two oxide
etching steps.
by uphill diffusion of phosphorus which resulted in the
increase of surface phosphorus signals after
deactivation.
ACKNOWLEDGMENTS
The XPS system used for analyzing SOI samples
was supported by National Synchrotron Radiation
Research Center (NSRRC), Taiwan.
REFERENCES
1. International Technology Roadmap for Semiconductors
(ITRS), http://www.itrs.net/, (2009).
2. S. Qin, S. Prussin, J. Reyes, Y. J. Hu and A. Mcteer,
IEEE Trans. on Plasma Sci., 37 1754-1759 (2010).
3. S. Prussin, S. Qin, J. Reyes and J. Reyes, in Proc. 17th
Int. Conf. IIT, AIP Conf. Proc., 1066 75-78 (2008).
4. N. S. Bennett, A. J. Smith, B. Colombeau, R. Gwilliam,
N. E. B. Cowern and B. J. Sealy, Mat. Sci. Eng. B, 124125 305-309 (2005).
5. N. G. E. Johansson and J. W. Mayer, Solid-State
Electron., 13, 317-335 (1970).
6. Y. K. Yeo, R. L. Hengehold and D. W. Elsaesser, J. Appl.
Phys., 61, 5070 (1987).
7. Y. Takamura, P. B. Griffin and J. D. Plummer, J. Appl.
Phys., 92, 235 (2002).