Formation of NiSi-Silicided p+n Shallow Junctions Using Implant-Through-Silicide and Low-Temperature Furnace Annealing

Article abstract of DOI:10.1149/1.1599851

NiSi-silicided p⁺n shallow junctions are fabricated using BF₂⁺ implantation into/through thin NiSi silicide layer (implant-through-silicide technology) followed by low-temperature furnace annealing (from 550 to 800°C). The NiSi film agglomerates following a thermal annealing at 600°C and may result in the formation of discontinuous islands at a higher temperature. The incorporation of fluorine atoms in the NiSi film can retard the formation of film agglomeration and thus improving the film's thermal stability. The forward ideality factor of about 1.02 and the reverse current density of about 1 nA/cm² can be attained for the NiSi (310A)/p⁺n junctions fabricated by BF₂ ⁺ implantation at 35 keV to a dose of 5 × 10¹⁵ cm^-2 followed by a 650°C thermal annealing; the junction formed is about 60 nm measured from the NiSi/Si interface. Activation energy measurement indicates that the reverse bias junction currents are dominated by the diffusion current, indicating that most of the implanted damages can be recovered after annealing at a temperature as low as 650°C.

Full text of DOI:10.1149/1.1599851

Journal of The Electrochemical Society, 150 ͑9͒ G557-G562 ͑2003͒
G557
0013-4651/2003/150͑9͒/G557/6/$7.00 © The Electrochemical Society, Inc.
Formation of NiSi-Silicided p¿n Shallow Junctions Using
Implant-Through-Silicide and Low-Temperature
Furnace Annealing
z
*
Chao-Chun Wang, Chiao-Ju Lin, and Mao-Chieh Chen
Department of Electronics Engineering, National Chiao-Tung University, Hsinchu 300, Taiwan
NiSi-silicided pϩn shallow junctions are fabricated using BFϩ2 implantation into/through thin NiSi silicide layer ͑implant-through-
silicide technology͒ followed by low-temperature furnace annealing ͑from 550 to 800°C͒. The NiSi film agglomerates following
a thermal annealing at 600°C and may result in the formation of discontinuous islands at a higher temperature. The incorporation
of fluorine atoms in the NiSi film can retard the formation of film agglomeration and thus improving the film’s thermal stability.
The forward ideality factor of about 1.02 and the reverse current density of about 1 nA/cm2 can be attained for the NiSi
(310 Å)/pϩn junctions fabricated by BFϩ2 implantation at 35 keV to a dose of 5 ϫ 1015 cmϪ2 followed by a 650°C thermal
annealing; the junction formed is about 60 nm measured from the NiSi/Si interface. Activation energy measurement indicates that
the reverse bias junction currents are dominated by the diffusion current, indicating that most of the implanted damages can be
recovered after annealing at a temperature as low as 650°C.
© 2003 The Electrochemical Society. ͓DOI: 10.1149/1.1599851͔ All rights reserved.
Manuscript submitted October 21, 2002; revised manuscript received March 25, 2003. Available electronically July 24, 2003.
Metal silicides have been extensively used in submicron devices
for lowering the series and contact resistance of source/drain and
polysilicon gate. Moreover, with the continuing scaling of device
dimension into the deep submicrometer regime, the silicide-as-
diffusion source ͑SADS͒ process has been proposed for the forma-
tion of silicide-contacted shallow junctions. In this process, silicide-
contacted shallow junctions are formed by implanting dopants into
the silicide layer followed by low-temperature annealing; thus, the
process is also designated as implant-into/through-silicide ͑ITS͒
technology. Shallow junction formation, low-temperature process-
ing, and low contact resistance are the major advantages of the
SADS process.1-7
also shows neither adverse linewidth dependence nor creep-up
phenomenon.23,27 In addition, NiSi has a lower formation tempera-
ture ͑ϳ400°C͒ than TiSi2 and CoSi2 , making it suitable for the
low-temperature process required for future device fabrication. The
further advantage of NiSi relies on its smaller consumption of Si
͑0.82 Å Si for 1 Å NiSi͒ compared to CoSi2 ͑1.04 Å Si for 1 Å
CoSi2),28 which enables a shallow junction depth formation at the
metal/Si contact. Besides, NiSi has the properties of low NiSi/Si
contact resistance,23,29 wide process window ͑400-750°C͒, and low
film stress.20-26 These peculiarities make NiSi suitable for usage in
the low-temperature process for sub-100 nm CMOS technology.30,31
In this work, formation of NiSi-silicided pϩn shallow junction
using ITS technology is investigated with respect to various implan-
tation energies and doses. Feasibility of low-temperature processing
and high-temperature stability of the NiSi/pϩn junction are evalu-
ated by analyzing the electrical characteristics and material proper-
ties of the silicided junction. Effects of implantation conditions on
the junction characteristics and high-temperature stability of the
NiSi films are also investigated.
Among the metal silicides, titanium disilicide (TiSi2) and cobalt
disilicide (CoSi2) are now widely used in salicide processes across
the industry because of their good thermal stability and low electri-
cal resistivity. However, some critical drawbacks limit their applica-
tions to future ultralarge-scale integrated ͑ULSI͒ technology. It has
been reported that Ti may react with implanted dopant to form com-
pounds such as Ti-B and Ti-As, making TiSi2 a very ineffective
diffusion source for boron or arsenic.8,9 Moreover, the processing
temperature window of TiSi2 is relatively narrow due to the high-
temperature requirement for the high-resistivity C49 to low-
resistivity C54 phase transition ͑м800°C͒ and the silicide agglom-
eration temperature limit ͑ϳ950°C͒.10 More importantly, incomplete
transformation from C49 to C54 TiSi2 phase occurs when the line-
width is scaled down below 2 ␮m because of the lack of nucleation
center in the narrow lines, especially the triple-C49 grain
boundaries.11-14 In addition, the creep-up phenomenon during the
formation of TiSi2 silicide may form a bridge between the gate and
source/drain regions in submicrometer devices, causing device
failure.15 Unlike TiSi2 , CoSi2 has neither adverse linewidth depen-
dence nor creep-up phenomenon. However, high Si consumption
during the formation of CoSi2 silicide is a major drawback for the
submicrometer process. This restricts the vertical scaling for CoSi2
to achieve shallow junction.16 Moreover, junction spiking of CoSi2
contacted shallow junction due to sensitivity to native oxide, and the
oxygen-containing environment needs a more complex silicidation
process, such as capping a passivation film during silicidation.17-20
Recently, nickel monosilicide ͑NiSi͒ has been recognized as a
promising candidate for a contact metal for deep submicrometer
device applications.15,21-33 NiSi has a low electrical resistivity of
14-20 ␮m cm, which is comparable to those of TiSi2 and CoSi2 ; it
Experimental
The NiSi/pϩn junction diodes were fabricated on n-type ͑100͒-
oriented silicon wafers with 2.7-4 ⍀ cm nominal resistivity. After
standard RCA cleaning, 5500 Å thick SiO2 was thermally grown by
pyrogenic oxidation at 1050°C. Active regions with areas of 1100,
580, 270, and 120 ␮m2 were defined by photolithography followed
by wet etching. A Ni film of 150 Å thickness was sputter deposited
in a dc sputtering system with a base pressure of less than 2.5
ϫ 10Ϫ8 Torr, using a Ni target in Ar ambient at a pressure of 2
ϫ 10Ϫ3 Torr with a deposition rate of about 10 Å/s. After the Ni
film deposition, the samples were rapid thermal annealed ͑RTA͒ at
500°C for 30 s in a N2 ambient to form NiSi. The unreacted Ni film
was selectively etched using a solution of H2SO4 :H2O2 ϭ 3:1 at
75-85°C. The formed NiSi film was about 310 Å thick as deter-
mined by transmission electron microscopy ͑TEM͒ analysis. The
pϩn junction diodes were formed by BFϩ2 implantation into/through
the NiSi silicide at an energy of 20-35 keV to a dose of 2 ϫ 1015 or
5 ϫ 1015 cmϪ2, followed by thermal annealing at a temperature
ranging from 550 to 800°C in a N2 ambient for 30 min. Finally, a
5000 Å thick Al layer was deposited on the back side of Si substrate
for all samples for a better contact in electrical measurements.
The thickness of the as-deposited Ni film and the formed NiSi
film were determined with chemical delineation and TEM observa-
tion. Sheet resistance was measured by four-point probe on the un-
patterned area. The pϩn junction depth was determined by spreading
* Electrochemical Society Active Member.
z E-mail: wangchauchiung.ee88g@nctu.edu.tw
Downloaded on 2015-03-08 to IP 130.15.241.167 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

G558
Journal of The Electrochemical Society, 150 ͑9͒ G557-G562 ͑2003͒
Figure 1. As-implanted boron profiles obtained by TRIM simulation for
BFϩ2 implantation into a 310 Å thick NiSi film on Si substrate to a dose of
5 ϫ 1015 cmϪ2 at various energies.
resistance profiling ͑SRP͒ measurement. Surface morphology was
observed by scanning electron microscopy ͑SEM͒. Secondary ion
mass spectrometry ͑SIMS͒ was used to determine the dopant con-
centration profiles. The current-voltage ͑I-V͒ characteristics of the
NiSi/pϩn junction diodes were measured by a semiconductor param-
eter analyzer HP-4145B. The open-circuit leakage current of the
measuring system was kept below 0.5 pA. The forward bias from 0
to 1 V at a step of 0.01 V and the reverse bias from 0 to Ϫ5 V at a
step of 0.05 V were used for the I-V characteristics measurement.
Temperature dependence of the reverse junction current was mea-
sured from room temperature to 200°C on a thermal vacuum chunk.
Figure 2. Sheet resistance vs. annealing temperature for NiSi͑310 Å͒/Si
samples implanted with BFϩ2 at various energies to a dose of ͑a͒ 2 ϫ 1015
and ͑b͒ 5 ϫ 1015 cmϪ2. The sample without ion implantation is designated
as control sample.
Results and Discussion
Transport of ions in matter simulation.—Before making the BF2ϩ
implantation, the as-implanted dopant distributions in NiSi film and
Si substrate were predicted by transport of ions in matter ͑TRIM͒
simulation. For BFϩ2 ion implantation, the BFϩ2 ions are believed to
be dissociated upon their first atomic scattering. Therefore, the bo-
ron energy is obtained by multiplying the BFϩ2 energy by the mass
plantation are also included for comparison. The RS of the control
sample starts to show drastic increase after annealing at tempera-
tures above 600°C. This increase of RS was apparently due to the
formation of islands at high temperatures, resulting in discontinuous
structure of the NiSi film, as confirmed by the SEM micrographs to
be shown later. The RS of the BFϩ2 -implanted samples decreased
slightly with increasing temperature up to 700°C. For the sample
implanted with a lower dose of 2 ϫ 1015 cmϪ2, annealing at 750°C
resulted in NiSi film agglomeration and thus the increase of RS ,
while the sample implanted with a higher dose of 5 ϫ 1015 cmϪ2
was able to remain stable at temperatures up to 750°C. The corre-
sponding electrical resistivity of the NiSi film is about 18 ␮⍀ cm,
which is in agreement with the value reported in the literature.16 The
different behavior of RS value vs. annealing temperature for the
sample with and without BFϩ2 implantation is attributed to the incor-
poration of fluorine in the NiSi film.35-37 Moreover, a larger amount
of fluorine retarded the NiSi film agglomeration to a higher tempera-
ture. Figure 3 shows the secondary ion counts of fluorine profiles in
the sample of NiSi ͑310 Å͒/Si implanted with BFϩ2 at 35 keV to a
dose of 2 ϫ 1015 as well as 5 ϫ 1015 cmϪ2 followed by 750°C
thermal annealing. The fluorine concentration in the NiSi film and at
the NiSi/Si interface is much higher for the higher dose
BFϩ2 -implanted samples. The slightly higher value of RS after an-
nealing at 800°C is attributed to the formation of NiSi2 phase.
ratio of Bϩ to BFϩ2 , which is 11/49. In this work, boron ions are to
be implanted into a sample which is composed of a 310 Å thick NiSi
film and an underlying Si substrate. Figure 1 shows the as-implanted
boron profiles obtained by TRIM simulation for BFϩ2 implantation at
various energies to a dose of 5 ϫ 1015 cmϪ2. For the 20 keV im-
plantation, nearly all implanted boron ions are located inside the
silicide layer. In this case, the silicide film serves as a diffusion
source of boron for the pϩn junction formation during the subse-
quent thermal annealing process. Boron in the silicide layer has to
diffuse into the silicon substrate and become electrically active to
form a good junction. Because boron can be electrically activated at
a temperature as low as 550°C,34 using NiSi as a boron diffusion
source is possible at low processing temperature provided that a
sufficient amount of boron atoms are diffused into the Si substrate.
For the 25, 30, and 35 keV implantations to a dose of 5
ϫ 1015 cmϪ2, the boron concentrations at the NiSi/Si interface are
all higher than 1 ϫ 1020 cmϪ3, while all projection ranges are kept
within the silicide. These interface boron concentrations are high
enough to form a good silicided pϩn junction provided that suffi-
cient amount of dopants are electrically activated and most of the
implantation damage can be annealed out.
SEM analysis.—Figure 4 shows the surface morphology of the
NiSi ͑310 Å/Si͒ control sample annealed at various temperatures.
Film agglomeration occurred at temperatures as low as 600°C, and
the NiSi film was turned into isolated islands after annealing at and
above 700°C, resulting in a drastic increase in the measured sheet
resistance ͑Fig. 2͒. Figure 5 shows the surface morphology of the
Sheet resistance.—Figure 2 shows the sheet resistance (RS) as a
function of annealing temperature for the BFϩ2 -implanted NiSi ͑310
Å͒/Si samples; RS data for the control sample without any ion im-
Downloaded on 2015-03-08 to IP 130.15.241.167 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 150 ͑9͒ G557-G562 ͑2003͒
G559
Figure 3. Secondary ion counts of fluorine ions in the sample of NiSi ͑310
Å͒/Si implanted with BF2ϩ at 35 keV to a dose of 2 ϫ 1015 as well as 5
ϫ 1015 cmϪ2 followed by 750°C thermal annealing.
BFϩ2 ͑35 keV͒ implanted NiSi ͑310 Å͒/Si samples annealed at vari-
ous temperatures. For the lower dose (2 ϫ 1015 cmϪ2) implanted
sample, the NiSi film remained smooth after annealing at 700°C, but
some minor agglomeration of NiSi film was observed after anneal-
ing at 750°C, which might cause NiSi/Si interface roughness and
degrade the NiSi/pϩn shallow junction characteristics. For the
higher dose (5 ϫ 1015 cmϪ2) implanted samples, the NiSi film re-
mained stable at temperatures up to 750°C. Similar results were
observed for the samples implanted with BFϩ2 at lower energies of
25 and 30 keV. Apparently, incorporation of fluorine in NiSi film in
a larger amount promoted the retardation of the NiSi film agglom-
eration. We presume that the segregation of the implanted fluorine
species at the NiSi grain boundaries resulted in the change of grain
boundary energy and interfacial energy, leading to retardation of the
film agglomeration. After annealing at 800°C, localized sinking was
observed on the silicide surface of both low- and high-dose im-
Figure 5. Top-view SEM micrographs showing the surface morphology of
BFϩ2 ͑35 keV͒ implanted NiSi ͑310 Å͒/Si samples annealed at ͑a͒ 700, ͑b͒
750, and ͑c͒ 800°C. The micrographs of the low-dose (2 ϫ 1015 cmϪ2) and
high-dose (5 ϫ 1015 cmϪ2) implanted samples are shown on the left and
right columns, respectively.
planted samples, and the degree of film agglomeration was relieved
to some extent. Because NiSi2 phase was found to appear after
annealing at 800°C ͓X-ray diffraction ͑XRD͒ result not shown͔ and
it consumes more Si (Ni:Si:NiSi2 ϭ 1 Å:3.65 Å:3.63 Å) than the
formation of NiSi phase (Ni:Si:NiSi ϭ 1 Å:183 Å:2.34 Å),28 the
localized newly formed NiSi2 surface would sink. Thus, we presume
that the formation of NiSi2 relaxed the agglomeration of silicide film
and was able to maintain the continuity of the silicide film. We may
conclude that the BFϩ2 implantation improves the thermal stability of
NiSi film as confirmed by the film’s sheet resistance behavior and
the smoother surface morphology.
Junction depth measurement.—SRP measurement was used to
determine junction depth in this work with an SSM-150 SRP sys-
tem. All samples prepared for SRP measurements were first capped
with a 3000 Å plasma-enhanced chemical vapor deposited ͑PECVD͒
oxide and then were polished to a small beveling angle of 17° for
better resolution. Table I lists the junction depths measured from the
silicide/silicon interface for the NiSi/pϩn samples studied in this
work. For the sample processed with BFϩ2 implantation at 20 keV,
the junction depth was determined to be about 25 nm after a
650°C/30 min thermal annealing. Since nearly all implanted boron
ions are confined in the silicide film for the BFϩ2 implantation at 20
keV, the junction must be formed by the diffusion of boron from the
silicide during the thermal annealing process. Although the diffusiv-
ity of boron in silicon is negligibly small at 650°C, it is presumed
that the diffusivity of boron in silicon of this study could be en-
hanced by the presence of silicide film.3,38 The enhancement in dif-
fusivity is attributable to the vacancy injection from the silicide. For
the implantation at higher energies ͑25, 30, and 35 keV͒, the implan-
tation damage in silicon may also play a role for boron diffusion,
Figure 4. Top-view SEM micrographs showing the surface morphology of
NiSi ͑310 Å͒/Si control sample annealed at ͑a͒ 600, ͑b͒ 700, ͑c͒ 750, and ͑d͒
800°C.
Downloaded on 2015-03-08 to IP 130.15.241.167 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

G560
Journal of The Electrochemical Society, 150 ͑9͒ G557-G562 ͑2003͒
Table I. Junction depths „in unit of nanometers… of
NiSi „310 Å…Õp¿n junctions formed by ITS scheme with BF¿2 im-
plantation at various energies to a dose of 5 ϫ 1015 cmÀ2 fol-
lowed by a 30 min thermal annealing.
Implantation energy ͑keV͒
Annealing
temperature
20
25
30
35
600°C
650°C
700°C
750°C
23
25
28
35
24
36
40
45
40
45
49
59
56
60
65
70
resulting in deeper junctions. Higher implantation energy and higher
annealing temperature resulted in an even deeper junction.
Electrical characteristics.—The electrical properties of the sili-
cide contacted pϩn shallow junction diodes fabricated by the ITS
scheme are dependent on a number of factors, including the dopant
activation level, implantation damage recovery, silicide/silicon inter-
face roughness, and the distance between the silicide/silicon inter-
face to the junction position. All of these are closely related to the
energy and dosage of the dopant ion implantation as well as the
dopant activation ability and the drive-in diffusion during the sub-
sequent annealing process; this is especially important for the case
of low-thermal-budget and low-energy implantation for the ITS
scheme. All measurements were performed at room temperature,
and each data point was obtained by averaging the data measured
from six randomly chosen samples.
Forward ideality factor.—The forward ideality factor ␩ of a junc-
tion diode can be extracted from the basic I-V relation
I ϭ IS exp qV/␩kT͒ Ϫ 1
͔
͓
͑
Figure 7. Reverse bias current density vs. annealing temperature for the
NiSi/pϩn junction diodes fabricated with BFϩ2 implantation at various ener-
gies to a dose of ͑a͒ 2 ϫ 1015 and ͑b͒ 5 ϫ 1015 cmϪ2
.
where IS is the reverse saturation current, q is the electronic charge,
k is the Boltzmann constant, and T is the temperature at measure-
ment. An ideality factor of unity indicates that diffusion current
predominates while a factor of two indicates that depletion recom-
bination current is dominant. The ideality factor ␩ can be deter-
mined from the slope of the linear segment of the I-V curve plotted
on semilogarithmic coordinates. Figure 6 shows the ideality factor
vs. annealing temperature for the NiSi/pϩn junction diodes fabri-
cated with BFϩ2 implantation at various energies to a dose of 5
ϫ 1015 cmϪ2. All ideality factors are below 1.06 for the samples
annealed at and below a temperature of 750°C. The data in Fig. 6
indicates that a 30 min low-temperature ͑600-700°C͒ annealing is
capable of incorporating sufficient amount of activated boron atoms
in Si substrate as well as recovering the implantation damage to
obtain a good NiSi/pϩn shallow junction with an ideality factor
below 1.04.
Reverse bias current.—Figure 7 shows the reverse bias current den-
sities (JR) vs. annealing temperature for the NiSi/pϩn junction di-
odes with an area of 0.0121 cm2 (1100 ϫ 1100 ␮m) measured at a
reverse bias of Ϫ5 V. The JR is determined by directly dividing the
measured current by the diode’s area. Roughness of the silicide/Si
interface in a shallow junction may lead to the formation of local-
ized Schottky contacts or the agglomeration-induced local silicide
spiking, resulting in the increase of reverse bias current. For the
lower dose (2 ϫ 1015 cmϪ2) implanted samples, JR of less than 2
nA/cm2 was easily achieved for the samples fabricated with a BF2ϩ
implantation at 25-35 keV followed by a thermal annealing at 550-
700°C. For the samples annealed at 750°C, the 25 and 30 keV im-
planted samples exhibited drastic increase in reverse bias leakage
current, while the 35 keV implanted sample revealed only a slight
increase in leakage current. This increase in reverse bias leakage
current is consistent with the behavior of the sample’s sheet resis-
tance (RS), which in turn is related to the extent of agglomeration of
the NiSi film. The agglomeration of the NiSi film induces roughness
of the silicide/Si interface, leading to the penetration of silicide
through the shallow junction and thus the degradation of the junc-
tion characteristics. The shallower the junction, the more susceptible
the junction to the silicide/Si interface roughness. Thus, the JR and
the thermal stability of the pϩn junction fabricated with BFϩ2 im-
plantation at 35 keV are better than those of the shallower junctions
Figure 6. Forward ideality factor vs. annealing temperature for the NiSi/pϩn
junction diodes fabricated with BFϩ2 implantation at various energies to a
dose of 5 ϫ 1015 cmϪ2
.
Downloaded on 2015-03-08 to IP 130.15.241.167 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 150 ͑9͒ G557-G562 ͑2003͒
G561
Figure 9. The JЈR vs. P/A plot of the NiSi/pϩn junctions fabricated with BFϩ2
implantation at 20 keV to a dose of 5 ϫ 1015 cmϪ2 followed by annealing at
550-750°C.
Figure 8. Arrhenius plots of the NiSi (310 Å)/pϩn junctions fabricated by
BFϩ2 implantation at 20 and 35 keV to a dose of 5 ϫ 1015 cmϪ2 followed by
650-750°C annealing for 30 min. The measurement was conducted at 1 V
reverse bias.
Area and peripheral current.—The reverse bias leakage current (IR)
of a pϩn junction consists of the reverse area leakage current (IRA
)
and the reverse peripheral leakage current (IRP
)
fabricated with BFϩ2 implantation at 25 and 30 keV. With the anneal-
ing temperature raised to 800°C, all samples exhibited an even
higher reverse bias leakage current. This is attributed to the forma-
tion of NiSi2 silicide phase, the increased volume of which de-
creased the distance between the silicide and the junction position,
leading to easier formation of localized Schottky contacts and thus
the increase of reverse bias current. Moreover, since NiSi2 is formed
via a nucleation-controlled mechanism which induces a corrugated
silicide/Si interface due to random nucleation on the original inter-
face, large reverse bias current can be easily induced for shallow
silicided junction.35,39 As for the NiSi/pϩn samples fabricated with a
higher dose (5 ϫ 1015 cmϪ2) BF2ϩ implantation, the JR of less than
2 nA/cm2 can be easily achieved with a postimplant thermal anneal-
ing at a temperature of 550-750°C. With a higher dose of BF2ϩ
implantation, the pϩn junction formed becomes deeper and a larger
amount of fluorine atoms is incorporated in the NiSi film, improving
the thermal stability of the thin NiSi film.36,37 With the annealing
temperature raised to 800°C, formation of NiSi2 phase started to
occur, resulting in a drastic increase in reverse bias leakage current,
similar to those observed in the samples fabricated with a lower
dose (2 ϫ 1015 cmϪ2) of BFϩ2 implantation.
I
R ϭ IRA ϩ IRP ϭ A ϫ JRA ϩ P ϫ JRP
where A is the junction area, P is the length of junction perimeter,
RA is the junction area leakage current density, and JRP is the junc-
tion peripheral leakage current density. A simple arrangement gives
J
JЈ ϭ JRA ϩ JRP P/A͒
͑
R
where J R ϭ I /A. Thus, by measuring the I of junctions with
different P/A ratio, the slope of JЈR vs. P/A plot gives the JRP and
the Y axis intersection gives the JRA . Figure 9 shows the JЈR vs. P/A
Ј
R
R
plot of the NiSi/pϩn junctions fabricated with BFϩ2 implantation at
20 keV to a dose of 5 ϫ 1015 cmϪ2 followed by thermal annealing
at a temperature from 550 to 750°C. It is found that the JRA de-
creases from 1.38 ϫ 10Ϫ9 to 6.57 ϫ 10Ϫ10 A/cm2 and the JRP in-
creases from 1.59 ϫ 10Ϫ11 to 4.28 ϫ 10Ϫ11 A/cm as the annealing
temperature was increased from 550 to 750°C. For the junction di-
ode with a size of 1100 ␮m2 formed by annealing at 700°C, the JRA
is 0.79 nA/cm2 and the JRP is 46.6 pA/cm, and the corresponding
area and peripheral current are 9.5 and 20.5 pA, respectively. This
indicates that more than 68% of the total reverse current leaks
through the junction’s perimeter. For a smaller junction with an area
of 120 ␮m2, the peripheral component accounts for 95% of the total
reverse current, indicating the major role of the peripheral leakage
component. Presumably, most implanted defects were confined
within the silicide layer; thus, very few extended defects are located
beyond the silicide layer and the impact of these defects is minimal
on the junction leakage. The peripheral leakage is sensitive to the
interfacial behavior of Si/SiO2 along the junction perimeter because
the junctions formed in this study are all surrounded by field oxide.
It is notable that the distance from the silicide/Si interface to the
junction at the diode perimeter is much shorter than that at the diode
bottom area after the postimplant thermal annealing. Moreover, the
mechanical stress-induced defects along the diode perimeter is pre-
sumed to cause a higher peripheral leakage.40 Thus, the junction
characteristics are very sensitive to the silicide/Si interface property
and the reverse junction current is dominated by the silicide-induced
current along the periphery. Figure 10 illustrates a typical I-V char-
acteristic for the NiSi/pϩn junction fabricated by BFϩ2 implantation
at 35 keV to a dose of 5 ϫ 1015 cmϪ2 followed by a 700°C thermal
annealing for 30 min. It is clear that NiSi can serve as an effective
Activation energy measurement.—The temperature dependence of
reverse bias junction current can provide insight into the junction
leakage mechanism. The temperature dependence of reverse current
IR is given by
IR ϰ T3 exp ϪE /kT͒
͑
a
where Ea is the activation energy of the junction, k is the Boltzmann
constant, and T is the temperature at measurement. The value of Ea
is close to the bandgap of silicon Eg when the reverse current is
dominated by the diffusion current and is close to Eg/2 when the
reverse current is dominated by the generation current. Figure 8
shows the Arrhenius plots for the NiSi (310 Å)/pϩn samples fabri-
cated with various implantation and annealing conditions. The mea-
surement was conducted at 1 V reverse bias. The activation energy
was found to be close to the silicon bandgap of 1.12 eV for all
samples investigated. This result clearly indicates that the reverse
current was dominated by the minority carrier diffusion current at
temperatures from 40 to 200°C.
Downloaded on 2015-03-08 to IP 130.15.241.167 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

G562
Journal of The Electrochemical Society, 150 ͑9͒ G557-G562 ͑2003͒
National Chiao-Tung University assisted in meeting the publication costs
of this article.
References
1. V. Probst, H. Schaber, A. Mitwalsky, H. Kabza, L. Van den hove, and K. Maex, J.
Appl. Phys., 70, 708 ͑1991͒.
2. B. Y. Tsui, J. Y. Tsai, and M. C. Chen, J. Appl. Phys., 69, 4352 ͑1991͒.
3. B. S. Chen and M. C. Chen, J. Appl. Phys., 72, 4619 ͑1992͒.
4. B. S. Chen and M. C. Chen, IEEE Trans. Electron Devices, ED-43, 258 ͑1996͒.
5. C. T. Lin, K. P. Ma, P. F. Chou, and H. C. Cheng, J. Electrochem. Soc., 142, 1579
͑1995͒.
6. F. La Via and E. Rimini, IEEE Trans. Electron Devices, ED-44, 526 ͑1997͒.
7. J. S. Park, D. K. Sohn, J. U. Bae, C. H. Han, and J. W. Park, IEEE Trans. Electron
Devices, ED-47, 994 ͑2000͒.
8. V. Probst, H. Schaber, P. Lippens, L. Van den Hove, and R. F. Keersmacker, J.
Appl. Phys., 52, 1803 ͑1988͒.
9. K. Maex, R. F. De Keersmaecker, G. Ghosh, L. D. Delaey, and V. Probst, J. Appl.
Phys., 66, 5327 ͑1989͒.
10. H. Jeon, C. A. Sukow, J. W. Honeycutt, G. A. Rozgonyi, and R. J. Nemanich, J.
Appl. Phys., 71, 4269 ͑1992͒.
11. S. Motakef, J. M. E. Harper, F. M. d’Heurle, T. A. Gallo, and N. Herbots, J. Appl.
Phys., 70, 2660 ͑1991͒.
12. J. B. Lasky, J. S. Nakos, O. J. Cain, and P. J. Geiss, IEEE Trans. Electron Devices,
ED-38, 262 ͑1991͒.
13. Y. Matsubara, T. Horiuchi, and K. Okumura, Appl. Phys. Lett., 62, 2634 ͑1993͒.
14. P. S. Lee, K. L. Pey, D. Mangelinck, J. Ding, D. Z. Chi, and L. Chan, IEEE
Electron Device Lett., EDL-22, 568 ͑2001͒.
Figure 10. Typical I-V characteristic for the NiSi/pϩn junction fabricated
with BFϩ2 implantation at 35 keV to a dose of 5 ϫ 1015 cmϪ2 followed by a
30 min thermal annealing at 700°C.
15. F. Deng, R. A. Johnson, P. M. Asbeck, S. S. Lau, W. B. Dubbelday, T. Hsiao, and
J. Woo, J. Appl. Phys., 81, 8047 ͑1997͒.
16. J. P. Gambion and E. G. Colgan, Mater. Chem. Phys., 52, 99 ͑1998͒.
17. J. K. Kittl, Q. Z. Hong, H. Yang, N. Yu, S. B. Samavedam, and M. A. Gribelyuk,
Thin Solid Films, 332, 404 ͑1998͒.
18. Q. Z. Hong, W. T. Shiau, H. Yang, J. A. Kittl, C. P. Chao, H. L. Tsai, S. Krishnan,
I. C. Chen, and R. H. Havemann, Tech. Dig. Int. Electron Devices Meet., 1997,
107.
diffusion source for boron diffusion during the low-temperature
thermal annealing, resulting in the formation of NiSi/pϩn shallow
junction with excellent electrical characteristics.
Conclusion
This work investigates the material properties of the NiSi film
and the electrical characteristics of the NiSi contacted pϩn shallow
junctions fabricated using BFϩ2 implantation into/through thin NiSi
silicide layer ͑ITS technology͒ followed by low-temperature furnace
annealing. The NiSi film agglomerates following a thermal anneal-
ing at 600°C and may result in the formation of discontinuous is-
lands at a higher temperature. The incorporation of fluorine atoms in
the NiSi film can retard the formation of film agglomeration and
thus improve the film’s thermal stability. For the sample with a BF2ϩ
implantation dose of 5 ϫ 1015 cmϪ2, which is the highly doped
sample studied in this work, the NiSi film is able to remain stable at
temperatures up to 750°C. It is found that a larger amount of fluo-
rine incorporation would result in a higher thermal stability tempera-
ture. Upon annealing at 800°C, however, the film agglomeration is
relieved to some extent and localized sinking on the surface of Ni-
silicide is observed, presumably due to the formation of NiSi2 phase.
The junction depth of the NiSi/pϩn junction diodes fabricated in this
work ranges from 23 to 70 nm measured from the NiSi/Si interface.
The activation energy measurements indicate that the reverse bias
currents of the NiSi/pϩn junctions studied in this work are all domi-
nated by the diffusion current. The reverse bias current is composed
of the area current and the peripheral current components. For the
diode’s size of 120 ␮m2 or smaller, more than 95% of the total
reverse current would be contributed by the peripheral current. The
reverse bias current density of 1 nA/cm2 can be easily achieved for
the NiSi/pϩn junctions studied in this work. To be more specific, the
NiSi/pϩn junction fabricated with a 35 keV BFϩ2 implantation to a
dose of 5 ϫ 1015 cmϪ2 followed by a 30 min thermal annealing at
650°C has a forward ideality factor of 1.02, a reverse bias current
density ͑at Ϫ5 V͒ of less than 1 nA/cm2, and a junction depth of 60
nm.
19. K. Maex, A. Lauwers, P. Besser, E. Kondoh, M. de Potter, and A. Steegen, IEEE
Trans. Electron Devices, ED-46, 1545 ͑1999͒.
20. A. Lauwers, P. Besser, M. De Potter, E. Kondoh, N. Roelandts, A. Steegen, M.
Stucchi, and K. Maex, IEEE International Interconnect Technology Conference
(IITC), p. 99 ͑1998͒.
21. T. Morimoto, H. S. Momose, T. Iinuma, I. Kunishima, K. Suguro, H. Okana, I.
Katakabe, H. Nakajima, M. Tsuchiaki, M. Ono, Y. Katsumata, and H. Iwai, Tech.
Dig. Int. Electron Devices Meet., 1991, 653.
22. T. Ohguro, S. Nakamura, M. Koike, T. Morimoto, A. Nishiyama, Y. Ushiku, T.
Yoshitomi, M. Ono, M. Saito, and H. Iwai, IEEE Trans. Electron Devices, ED-41,
2305 ͑1994͒.
23. T. Ohguro, S. Nakamura, E. Morifuji, M. Ono, T. Yoshitomi, M. Saito, H. S.
Momose, and H. Iwai, Tech. Dig. Int. Electron Devices Meet. 1995, 453.
24. T. Morimoto, T. Ohguro, S. Momose, T. Iinuma, I. Kunishima, K. Suguro, I. Ka-
takabe, H. Nakajima, M. Tsuchiaki, M. Ono, Y. Katsumata, and H. Iwai, IEEE
Trans. Electron Devices, ED-42, 915 ͑1995͒.
25. D. X. Xu, S. R. Das, C. J. Peters, and L. E. Erickson, Thin Solid Films, 326, 143
͑1998͒.
26. M. C. Poon, F. Deng, M. Chan, W. Y. Chan, and S. S. Lau, Appl. Surf. Sci., 157, 29
͑2000͒.
27. A. Lauwers, A. Steegen, M. de Potter, R. Lindsay, A. Satta, H. Bender, and K.
Maex, J. Vac. Sci. Technol. B, B19, 2026 ͑2001͒.
28. S. P. Murarka, Silicides for VLSI Applications, Academic Press, New York ͑1983͒.
29. Y. Tsuchiya, A. Tobioka, O. Nakatsuka, H. Ikeda, A. Sakai, S. Zaima, and Y.
Yasuda, Jpn. J. Appl. Phys., Part 1, 41, 2450 ͑2002͒.
30. Q. Xiang, C. Woo, E. Paton, J. Foster, B. Yu, and M. R. Lin, Symposium on VLSI
Technology Digest, p. 76 ͑2000͒.
31. R. Chau, J. Kavalieros, B. Roberds, R. Schenker, D. Lionberger, D. Barkage, B.
Doyle, R. Arghavani, A. Murtht, and G. Deewy, Tech. Dig. Int. Electron Devices
Meet., 2000, 45.
32. P. S. Lee, K. L. Pey, D. Mangelinck, J. Ding, A. T. S. Wee, and L. Chan, IEEE
Electron Device Lett., EDL-21, 566 ͑2000͒.
33. R. Mukai, S. Ozawa, and H. Yagi, Thin Solid Films, 270, 567 ͑1995͒.
34. M. Y. Tsai and B. G. Streetman, J. Appl. Phys., 50, 183 ͑1979͒.
35. S. S. Lau and N. W. Cheung, Thin Solid Films, 71, 117 ͑1980͒.
36. B. Y. Tsui, J. Y. Tsai, T. S. Wu, and M. C. Chen, IEEE Trans. Electron Devices,
ED-40, 54 ͑1993͒.
37. T. P. Nolan, R. Sinclar, and R. Beyers, J. Appl. Phys., 71, 720 ͑1992͒.
38. A. A. Brown, P. B. Moynagh, and P. J. Rosser, Semiconductor Silicon, 13, 280
͑1989͒.
39. B. A. Julies, D. Knoesen, R. Pretorius, and D. Adams, Thin Solid Films, 347, 201
͑1999͒.
40. K. Goto, J. Watanabe, A. Fushida, T. Sakuma, and T. Sugii, IEEE International
Reliability Physics Symposium (IRPS), p. 363 ͑1998͒.
Acknowledgment
This work was supported by the National Science Council, ROC,
under contract no. NSC-90-2215-E009-067.
Downloaded on 2015-03-08 to IP 130.15.241.167 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).