The effects of oxygen-insertion technology and low-energy fluorine (F) implantation on the Schottky barrier height (ΦBp) of a Pt/Ti/p-type Si metal–semiconductor contact are studied via electrical characterizations of Schottky diodes, physical analyses by secondary ion mass spectrometry , and chemical analyses by x-ray photoelectron spectroscopy. It is found that both oxygen-inserted (OI) layers and F can reduce ΦBp due to Ti 2p and Si 2p binding energy shifts before forming gas anneal (FGA) and due to retarded Pt diffusion into Si (facilitating low-ΦBp Pt mono-silicide formation) during FGA. The experimental findings also suggest that OI layers are more effective than F for reducing ΦBp.

As CMOS transistor dimensions are scaled down toward the atomic limit, reductions in the size of the source and drain (S/D) contact areas result in increased contact resistance (RCON), which degrades transistor on-state current and hence integrated circuit (IC) performance.1,2RCON is sensitive to both the S/D active dopant concentration (NSD,active) and the Schottky barrier height (SBH) (ΦB). Therefore, research efforts to reduce RCON have pursued multiple approaches: (1) increase the S/D active dopant concentration, e.g., by solid phase epitaxial (SPE) recrystallization of dopant-implanted semiconductor S/D regions;3–5 (2) maximize the real S/D contact area by removing highly resistive residues using an optimized reactive-ion etching (RIE) cleaning process;6 and (3) de-pin the Fermi level to reduce the effective Schottky barrier height (SBH), by inserting an ultra-thin interfacial layer (forming a metal–insulator–semiconductor structure)7–10 and/or by tuning the metal work function.11,12

To reduce RCON and thereby provide for improved CMOS IC performance, the preferred choice of metal for S/D contact formation has changed over the past few decades. In the earliest CMOS technology, aluminum (Al) was commonly used to form Al/Si contacts.13 It was replaced by TiSi2 due to the Al spiking problem.14 Lower resistivity silicides such as CoSi2 and Ni–Pt monosilicide subsequently replaced TiSi2.15 With the advent of FinFET structures at the 22 nm technology node, the semiconductor industry returned to Ti-based contacts.16 Ti has a low work-function value of 4.33 eV,17 resulting in a larger SBH (ΦBp ≅ 0.70 eV) for p-type silicon contacts and a smaller SBH (ΦBn ≅ 0.42 eV) for n-type silicon contacts.18 In this work, the effects of oxygen-inserted (OI) layers and of F (introduced by shallow implantation followed by a thermal spike anneal) on the SBH of a Pt/Ti/p-type Si system (ΦBp) are investigated experimentally. The effects of OI layers and of F on chemical-bond energy, metal/semiconductor intermixing, and metal-silicide formation during forming gas anneal (FGA) are discussed.

To experimentally extract the value of ΦBp, back-to-back Schottky diodes were fabricated on top of (100) p-type control and OI silicon wafer substrates. The OI samples were fabricated by growing 10 nm thick silicon epitaxial layers on blanket silicon wafers. During epitaxy, sample surfaces were intermittently exposed to a dilute oxygen source to form a OI layered structure similar to that shown in Fig. 1. As illustrated in Fig. 2, OI samples comprise four partial monolayers of oxygen inserted into the crystalline Si substrate, with a 1.6 nm-thick Si capping layer intended to protect the OI layers from the strong O gettering effects of Ti.19 

FIG. 1.

Schematic atomistic structure of the OI layers (yellow: silicon atom and red: oxygen atom). Each one of the incorporated O atoms occupies an interstitial site and is covalently bonded with two adjacent Si atoms.

FIG. 1.

Schematic atomistic structure of the OI layers (yellow: silicon atom and red: oxygen atom). Each one of the incorporated O atoms occupies an interstitial site and is covalently bonded with two adjacent Si atoms.

Close modal
FIG. 2.

Schematic cross-sectional view of the 6 nm-thick OI region in the OI silicon samples used in this work.

FIG. 2.

Schematic cross-sectional view of the 6 nm-thick OI region in the OI silicon samples used in this work.

Close modal

Some wafers received a 2.5 keV, 1 × 1015 cm−2 F implant to study the effects of F. To avoid channeling effects, a 7° tilt angle between the incident ion beam direction and the normal direction of the wafer [the (100) direction] was used. The rotation angle between the projection of the incident ion beam on the wafer and the (110) direction was kept at 0°. The samples were first cleaned in a sulfuric peroxide mixture (H2SO4: H2O2) bath at 120 °C for 10 min followed by a dip in dilute hydrofluoric acid (DHF) solution (10:1 H2O:49% HF) until the surface became hydrophobic. Afterward, the samples were immediately rinsed with de-ionized (DI) water. To recrystallize the amorphized Si lattice, F-implanted samples received a 1050 °C spike anneal in inert N2 ambient following the chemical cleaning.

Figure 3 illustrates the fabrication process used to form back-to-back Schottky diode test structures on the sample surface, with r1 ranging from 100 μm to 160 μm and r2 ranging from 120 μm to 200 μm, respectively. First, a low-temperature (350 °C) 230 nm-thick SiO2 layer was deposited by plasma-enhanced chemical vapor deposition (PECVD) onto the cleaned substrate surfaces. This oxide layer subsequently serves to protect the Si surface from being etched by TMAH-containing photoresist developer solution. Next, a 400 nm-thick lift-off photoresist (LOR-3A) layer and a 2 μm-thick g-line photoresist layer were coated onto the sample surface in a HeadwayTM photoresist spinner. Afterward the bi-layer photoresist stack was patterned by photolithography. Then, the metal/Si contact regions were exposed by removing the low-temperature oxide (LTO) layer in DHF (10:1 H2O:49% HF) solution for 7 min. Next, metal films (3 nm Ti followed by 10 nm Pt) were deposited by e-beam evaporation. The 3 nm Ti layer has superior interfacial adhesion, while the Pt layer has ultrahigh conductivity and serves as a probe layer. Thermal annealing in forming gas (a gaseous mixture of N2 and H2) is commonly used to passivate interface states for S/D contact formation to help reduce RCON.20,21 In this work, samples were subjected to a 5-min, 300 °C anneal in 10% H2, 90% N2 forming gas in a conventional rapid thermal processing (RTP) system (AccuThermoTM 610).

FIG. 3.

Fabrication process flow for Schottky diode test structures.

FIG. 3.

Fabrication process flow for Schottky diode test structures.

Close modal

Based on thermionic emission theory,22 the reverse saturation current density (J0) for a Schottky diode is given by

J0=A**T2exp(qΦB/kBT),
(1)

where A** is Richardson’s constant, ΦB is the SBH, and T is the absolute temperature in kelvin.

It can be seen from Eq. (1) that J0 is exponentially dependent on ΦB. This correlation allows ΦB to be extracted from measurements of the temperature-dependent electrical current–voltage (IV) characteristics of Schottky diodes,

ΦB=kBTqlnA**lnJ0T2.
(2)

Figure 4 illustrates the experimental setup used for the electrical measurement of Schottky diode IV characteristics in this work.

FIG. 4.

Back-to-back Schottky diode IV measurement: (a) the applied voltage on probe 2 was swept from 0 V to 0.5 V with probe 1 grounded and (b) example measured IV curves for r1 = 120 μm and r2 = 140 μm measured at 300 K.

FIG. 4.

Back-to-back Schottky diode IV measurement: (a) the applied voltage on probe 2 was swept from 0 V to 0.5 V with probe 1 grounded and (b) example measured IV curves for r1 = 120 μm and r2 = 140 μm measured at 300 K.

Close modal

Figure 5 shows Richardson plots [ln(IrT2) vs 1000T] for control and OI samples within the temperature range of 220 K (−53.15 °C)–310 K (+36.85 °C). The experimental data were fitted to straight lines using the least squares method; then, the slope values were used to extract ΦBp. The extracted ΦBp value for the control sample (691 meV) is consistent with the previously reported values for Ti/p-type Si contacts (ΦBp ≅ 700 meV).18 This suggests that the 3 nm-thick Ti layer continuously covers the semiconductor surface in the contact region. A comparison between the extracted ΦBp values for control and OI samples indicates that the OI layers result in lower ΦBp by 27 meV. Because of this, the Schottky diode saturation current (Ir) is higher for the OI sample. These results indicate that by inserting partial oxygen layers near to the Si substrate surface, RCON of a p-type Ti/Si contact can be lowered.

FIG. 5.

Richardson plots for OI and control samples.

FIG. 5.

Richardson plots for OI and control samples.

Close modal

Figures 6 and 7 show Richardson plots for un-implanted and F-implanted samples after FGA treatment. It can be seen that OI samples have lower ΦBp than control samples regardless of whether they had undergone F implant or FGA. Table I provides a summary of ΦBp values for all samples. A comparison of the values before vs after FGA indicates that both OI layers and F can enhance FGA-induced ΦBp reduction. However, FGA can reduce ΦBp by more than 50% (664 meV → 330 meV) in the presence of OI layers, whereas FGA only reduces ΦBp by 3.7% (563 meV → 542 meV) in the presence of F. The OI FGA sample has the lowest ΦBp (330 meV) among all the samples, suggesting that the OI technology is more beneficial than F implantation for reducing p-type contact resistance.

FIG. 6.

Richardson plots for OI and control samples after FGA.

FIG. 6.

Richardson plots for OI and control samples after FGA.

Close modal
FIG. 7.

Richardson plots for F-implanted OI and F-implanted control samples after FGA.

FIG. 7.

Richardson plots for F-implanted OI and F-implanted control samples after FGA.

Close modal
TABLE I.

Comparison of extracted ΦBp values for samples before and after FGA.

SampleΦBp (meV)
Control 691 
OI 664 
Control FGA 589 
OI FGA 330 
F-implanted control FGA 542 
F-implanted OI FGA 423 
SampleΦBp (meV)
Control 691 
OI 664 
Control FGA 589 
OI FGA 330 
F-implanted control FGA 542 
F-implanted OI FGA 423 

To gain insight into the mechanisms for lower ΦBp in the presence of OI layers before FGA, secondary ion mass spectrometry (SIMS) and x-ray photoelectron spectroscopy (XPS) analyses were performed to elucidate the effects of oxygen on Ti. The dynamic SIMS data were obtained using the Cameca IMS-6f SIMS instrument equipped with both Cs and O2+ ion guns. Low bombardment energies below 1 keV were utilized to achieve a high depth resolution, and the depth scale was determined using the average sputtering rate for the films studied. The XPS analyses were carried out using a ThermoFisher K-alpha + system with monochromatic aluminum x-rays of 1468 eV.

Ti is a strong oxygen getter and is typically used to remove O contaminants in standard CMOS fabrication processes.19 At room temperature, Ti reacts with SiO2 to form TiO2: Ti + SiO2 → TiO2 + Si. The gettering process is accelerated at elevated temperatures. Figure 8 plots the O concentration profiles in OI samples before and after FGA. An O peak was observed in the top 3 nm-thick Ti layer even for the unannealed OI sample. This is likely due to the gettering of background O contamination during the e-beam evaporation process. A comparison between unannealed and annealed OI samples reveals that the OI layers lost some O atoms during FGA due to O gettering by Ti.

FIG. 8.

Oxygen concentration profiles in OI samples before and after FGA, from SIMS analyses. The vertical green line delineates the location of the Ti/Si interface.

FIG. 8.

Oxygen concentration profiles in OI samples before and after FGA, from SIMS analyses. The vertical green line delineates the location of the Ti/Si interface.

Close modal

Figure 9 compares the O concentration profiles in control and OI samples before FGA. The O peak in the Ti layer is similar for the control and OI samples.

FIG. 9.

Oxygen concentration profiles in OI and control samples before FGA, from SIMS analyses. The vertical green line delineates the location of the Ti/Si interface.

FIG. 9.

Oxygen concentration profiles in OI and control samples before FGA, from SIMS analyses. The vertical green line delineates the location of the Ti/Si interface.

Close modal

Figure 10 compares the O concentration profiles in control and OI samples after FGA. Due to FGA-enhanced O gettering, the O level in the Si substrate is much lower for the control sample than for the OI sample. This suggests that Si–O bonds in the OI layers can help retain O atoms in the silicon substrate, which can modify ΦBp.

FIG. 10.

Oxygen concentration profiles in OI and control samples after FGA, from SIMS analyses. The vertical green line shows the location of the Ti/Si interface.

FIG. 10.

Oxygen concentration profiles in OI and control samples after FGA, from SIMS analyses. The vertical green line shows the location of the Ti/Si interface.

Close modal

Previous studies identified the important role of interfacial chemistry to determine the SBH.23–26 In this work, XPS analyses were performed for the Pt/Ti/p-type Si system to study the interfacial chemical state and composition. Table II lists the Ti 2p3/2 binding energies of common chemical states.27–29 Higher values of Ti 2p3/2 binding energy are associated with Ti–O and Ti–N bonds. Therefore, it can be expected that O gettering of Ti (Ti–O bonding) results in a shift in the Ti 2p3/2 core energy level. Figures 11 and 12 plot the XPS spectrum data of Ti 2p peaks at the Ti/Si interface for control and OI samples before FGA. Because of O gettering, the Ti 2p3/2 peak energy levels are higher than the ideal reference value (454.1 eV for pure Ti metal)27 by 0.4 eV and 0.9 eV for control and OI samples, respectively. These values of the Ti 2p3/2 core energy level are much smaller than that for TiO2, indicating TiOx (x < 2) formation at the Ti/Si interface. It can be expected that OI layers facilitate TiOx formation since they are a source of additional O atoms so that the OI technology helps changing the Ti chemical state at the Ti/Si interface.

TABLE II.

Ti 2p3/2 binding energy values for common chemical states of Ti.

Chemical stateBinding energy Ti 2p3/2 (eV)
Pure Ti metal 454.1 
TiN 454.9 
TiO2 458.5 
Chemical stateBinding energy Ti 2p3/2 (eV)
Pure Ti metal 454.1 
TiN 454.9 
TiO2 458.5 
FIG. 11.

XPS spectrum data for Ti 2p peaks (Ti 2p3/2 and Ti 2p1/2) for the control sample before FGA.

FIG. 11.

XPS spectrum data for Ti 2p peaks (Ti 2p3/2 and Ti 2p1/2) for the control sample before FGA.

Close modal
FIG. 12.

XPS spectrum data of Ti 2p peaks (Ti 2p3/2 and Ti 2p1/2) for the OI sample before FGA.

FIG. 12.

XPS spectrum data of Ti 2p peaks (Ti 2p3/2 and Ti 2p1/2) for the OI sample before FGA.

Close modal

Previous work also discovered a correlation between the Ti chemical state and Ti/Si SBH.18 It was hypothesized that an increase in the Ti 2p3/2 core energy level corresponds to the work function of Ti moving closer to the Si valence band edge (Ev), which is beneficial for reducing ΦBp. The reduction in ΦBp by 27 meV (cf. Fig. 5) is consistent with the observed shift in the Ti 2p3/2 energy level due to Ti–O bond formation.

The electrical measurements presented above show that FGA treatment can reduce ΦBp significantly. The FGA thermal budget (5-min, 300 °C) is insufficient to cause Ti silicidation for which the lowest reported reaction temperature is 400 °C.30 At room temperature, the 3 nm-thick Ti layer can serve as an effective barrier to Pt diffusion into the Si substrate to prevent Pt silicidation. During FGA, Pt diffusion across the thin Ti interlayer into the Si substrate can occur via a grain boundary enhanced diffusion mechanism,31 causing Pt silicide formation.32 To understand the impacts of OI layers and F on Pt diffusion and Pt silicide formation, SIMS and XPS analyses were performed on samples that underwent FGA treatment.

Figure 13 plots the Pt concentration profiles for control and OI samples after FGA. It can be seen that FGA causes a significant amount of Pt diffusion into the Si substrate for both samples. Previous work on Pt diffusion showed that O contamination (in the form of a 2 nm-thick discontinuous SiO2 layer) at the Pt/Si interface can reduce Pt diffusion into the Si.32 It was hypothesized that the thin oxide layer serves as a Pt diffusion barrier because the dominant Pt diffusion mechanism at low temperatures (200–325 °C) is grain boundary diffusion. The difference in Pt concentration profiles for the control sample vs the OI sample shows that OI layers serve to retard Pt diffusion into the Si substrate, i.e., O atoms inserted interstitially into the crystalline Si lattice can also form an effective Pt diffusion barrier.

FIG. 13.

Pt concentration profiles (log scale) from SIMS analyses, for control and OI samples after FGA. Note that for the OI sample, Pt diffusion into Si is reduced as compared with the control sample.

FIG. 13.

Pt concentration profiles (log scale) from SIMS analyses, for control and OI samples after FGA. Note that for the OI sample, Pt diffusion into Si is reduced as compared with the control sample.

Close modal

Previous studies showed that the presence of F helps reduce Pt diffusion into Si because of the “fluorine buffer” effect during high temperature (400–850 °C) anneals.33,34 It is, therefore, expected that the presence of F reduces Pt diffusion into Si at lower temperatures. A comparison of Pt concentration profiles for un-implanted vs F-implanted samples is shown in Fig. 14. The F implant was found to reduce Pt diffusion into Si during FGA for both the control and OI samples. However, the F-induced reduction in Pt diffusion is enhanced significantly in the presence of OI layers (F-implanted OI FGA sample). To elucidate why OI layers enhance the “fluorine buffer” effect, F concentration profiles before and after thermal anneals (the 1050 °C spike anneal prior to Schottky diode fabrication and the 300 °C FGA) are plotted in Fig. 15. Due to F clustering with point defects, the F concentration profile for the control FGA sample shows a typical “double-peak” consistent with the previous studies of low-energy F implants.35 However, this “double-peak” is not seen in the OI sample because F diffusion into the Si substrate is blocked by the OI layers.36,37 Previous work found that the F concentration must exceed a certain level in order for F to effectively reduce Pt diffusion into Si.34 Since the OI layers result in higher F concentration near the silicon surface, the OI technology enhances the “fluorine-buffer” effects. This explains why the F-implanted OI FGA sample has the shallowest Pt depth profile among all F-implanted samples after FGA (cf. Fig. 14).

FIG. 14.

Pt concentration profiles (log scale) from SIMS analyses for unimplanted (solid lines) and F-implanted (dashed lines) samples after FGA. Pt diffusion into the Si substrate is dramatically reduced in the presence of OI layers.

FIG. 14.

Pt concentration profiles (log scale) from SIMS analyses for unimplanted (solid lines) and F-implanted (dashed lines) samples after FGA. Pt diffusion into the Si substrate is dramatically reduced in the presence of OI layers.

Close modal
FIG. 15.

F concentration profiles (log scale) from SIMS analyses before and after annealing (1050 °C recrystallization spike anneal and 5-min 300 °C FGA).

FIG. 15.

F concentration profiles (log scale) from SIMS analyses before and after annealing (1050 °C recrystallization spike anneal and 5-min 300 °C FGA).

Close modal

XPS analyses were performed to study the effects of OI layers and F on Pt silicidation during FGA. Table III provides a summary of common Pt 4f7/2 chemical states. Upon bonding with O (oxidation) and Si (silicidation), the Pt 4f7/2 core energy level shifts to larger values. Figures 16 and 17 plot XPS spectrum data for Pt 4f peaks at the original Si substrate surface for un-implanted samples after FGA. Based on the extracted Pt 4f7/2 binding energy shifts (+1.11 eV for the control FGA sample and +1.25 eV for the OI FGA sample), it can be deduced that FGA caused Pt2Si formation in the un-implanted control sample, while it caused PtSi formation in the un-implanted OI sample. Previous research reported a PtSi/Si ΦBp of 320 meV,38 consistent with the extracted ΦBp of 330 meV for the OI FGA sample.

TABLE III.

Pt 4f7/2 binding energy values for common chemical states of Pt.

Chemical stateBinding energy Pt 4f7/2 (eV)
Pure Pt metal 71.0 
Pt2Si 72.11 
PtSi 72.25 
PtO 72.4 
PtO2 74.9 
Chemical stateBinding energy Pt 4f7/2 (eV)
Pure Pt metal 71.0 
Pt2Si 72.11 
PtSi 72.25 
PtO 72.4 
PtO2 74.9 
FIG. 16.

XPS spectrum data of Pt 4f peaks for the control FGA sample.

FIG. 16.

XPS spectrum data of Pt 4f peaks for the control FGA sample.

Close modal
FIG. 17.

XPS spectrum data of Pt 4f peaks for the OI FGA sample.

FIG. 17.

XPS spectrum data of Pt 4f peaks for the OI FGA sample.

Close modal

Figures 18 and 19 plot the XPS spectrum data of Pt 4f peaks at the original Si substrate surface for F-implanted samples after FGA. For the control sample, F increases the Pt 4f7/2 core energy level by 20 meV, from +1.11 eV to +1.13 eV. In contrast, F reduces the Pt 4f7/2 core energy level from +1.25 eV to +1.18 eV for the OI sample. Based on Table III, these results suggest PtxSi (1 < x < 2) formation. It was shown in previous studies that a larger shift of the Pt 4f7/2 core level indicates more low-ΦBp Pt monosilicide formation, whereas a smaller shift indicates more high-ΦBp Pt-rich silicide formation.38 Consistent correlations between Pt 4f7/2 core level shifts and ΦBp were observed in this work. For the control sample after FGA, F reduces ΦBp by 47 meV (589 meV → 542 meV); in contrast, for the OI sample after FGA, F increases ΦBp from 330 meV to 423 meV. These results show that ΦBp is very sensitive to the Pt silicide phase.

FIG. 18.

XPS spectrum data of Pt 4f peaks for F-implanted control FGA sample.

FIG. 18.

XPS spectrum data of Pt 4f peaks for F-implanted control FGA sample.

Close modal
FIG. 19.

XPS spectrum data of Pt 4f peaks for F-implanted OI FGA sample.

FIG. 19.

XPS spectrum data of Pt 4f peaks for F-implanted OI FGA sample.

Close modal

It is worth noting here that previous work found that the normal sequence of Pt silicidation is Pt → Pt2Si → PtSi and that 300 °C is insufficient to form PtSi.38,39 Pt2Si forms first and converts to PtSi upon sufficient supply of Pt atoms at high temperature (∼700 °C). It was also shown that a thin oxide layer at the Pt/Si interface can reduce Pt diffusivity, enabling Pt2Si → PtSi conversion at a temperature much lower than 700 °C.32 Because both F and OI layers reduce Pt diffusion, there is insufficient supply of Pt atoms for PtSi formation in the F-implanted OI FGA sample as compared with the OI FGA sample.

The effects of OI layers and F on ΦBp for a Pt/Ti/p-type Si metal–semiconductor contact are studied in this work. For unannealed samples, OI layers facilitate Ti–O and Si–O bond formation, which can enhance Ti 2p and Si 2p core energy level shifts. FGA can reduce ΦBp by more than 50% (664 meV → 330 meV) in the presence of OI layers, whereas FGA only reduces ΦBp by 3.7% (563 meV → 542 meV) in the presence of F. XPS analyses of Pt 4f7/2 core level shifts reveal that Pt diffusion during FGA critically affects the Pt silicidation phase. Because OI layers can suppress Pt diffusion more effectively than F, OI layers promote low-ΦBp Pt monosilicide formation. It is found that the co-existence of OI layers and F blocks Pt diffusion almost completely, thus preventing Pt silicide formation. This finding suggests that care should be taken when using BF2 implants and OI layers in advanced contacts. In conclusion, the OI technology is demonstrated to be very promising for reducing p-type silicon contact resistance with Ti-based contacts.

The data that support the findings of this study are available from the corresponding author upon reasonable request.

1.
K. J.
Kuhn
,
IEEE Trans. Electron Devices
59
,
1813
(
2012
).
2.
C.
Lavoie
,
P.
Adusumilli
,
A. V.
Carr
,
J. S. J.
Sweet
,
A. S.
Ozcan
,
E.
Levrau
,
N.
Breil
, and
E.
Alptekin
,
ECS Trans.
77
,
59
(
2017
).
3.
H.
Niimi
,
Z.
Liu
,
O.
Gluschenkov
,
S.
Mochizuki
,
J.
Fronheiser
,
J.
Li
,
J.
Demarest
,
C.
Zhang
,
B.
Liu
, and
J.
Yang
,
IEEE Electron Device Lett.
37
,
1371
(
2016
).
4.
O.
Gluschenkov
,
Z.
Liu
,
H.
Niimi
,
S.
Mochizuki
,
J.
Fronheiser
,
X.
Miao
,
J.
Li
,
J.
Demarest
,
C.
Zhang
, and
C.
Niu
, “
FinFET performance with Si: P and Ge: Group-III-Metal metastable contact trench alloys
,” in
2016 IEEE International Electron Devices Meeting (IEDM)
(
IEEE
,
2016
), p.
17.2. 1
.
5.
Y.
Yang
,
N.
Breil
,
C.
Yang
,
J.
Hsieh
,
F.
Chiang
,
B.
Colombeau
,
B.
Guo
,
K.
Shim
,
N.
Variam
, and
G.
Leung
, “
Ultra low p-type SiGe contact resistance FinFETs with Ti silicide liner using cryogenic contact implantation amorphization and solid-phase epitaxial regrowth (SPER)
, in
2016 IEEE Symposium on VLSI Technology
(
IEEE
,
2016
), p.
1
.
6.
V.
Kamineni
,
M.
Raymond
,
S.
Siddiqui
,
F.
Mont
,
S.
Tsai
,
C.
Niu
,
A.
Labonte
,
C.
Labelle
,
S.
Fan
, and
B.
Peethala
, “
Tungsten and cobalt metallization: A material study for MOL local interconnects
,” in
2016 IEEE International Interconnect Technology Conference / Advanced Metallization Conference (IITC/AMC)
(
IEEE
,
2016
), p.
105
.
7.
A.
Agrawal
,
J.
Lin
,
M.
Barth
,
R.
White
,
B.
Zheng
,
S.
Chopra
,
S.
Gupta
,
K.
Wang
,
J.
Gelatos
, and
S. E.
Mohney
,
Appl. Phys. Lett.
104
,
112101
(
2014
).
8.
D.
Connelly
,
C.
Faulkner
,
P.
Clifton
, and
D.
Grupp
,
Appl. Phys. Lett.
88
,
012105
(
2006
).
9.
J.-Y. J.
Lin
,
A. M.
Roy
,
A.
Nainani
,
Y.
Sun
, and
K. C.
Saraswat
,
Appl. Phys. Lett.
98
,
092113
(
2011
).
10.
R.
Islam
,
G.
Shine
, and
K. C.
Saraswat
,
Appl. Phys. Lett.
105
,
182103
(
2014
).
11.
J.
Hu
,
K. C.
Saraswat
, and
H.-S. P.
Wong
,
J. Appl. Phys.
107
,
063712
(
2010
).
12.
J.
Padilha
,
A.
Fazzio
, and
A. J.
da Silva
,
Phys. Rev. Lett.
114
,
066803
(
2015
).
13.
J. D.
Plummer
,
Silicon VLSI Technology: Fundamentals, Practice and Modeling
(
Pearson Education India
,
2009
).
14.
C.
Ting
and
B.
Crowder
,
J. Electrochem. Soc.
129
,
2590
(
1982
).
15.
J.
Kittl
,
A.
Lauwers
,
O.
Chamirian
,
M.
Van Dal
,
A.
Akheyar
,
M.
De Potter
,
R.
Lindsay
, and
K.
Maex
,
Microelectron. Eng.
70
,
158
(
2003
).
16.
C.-N.
Ni
,
X.
Li
,
S.
Sharma
,
K.
Rao
,
M.
Jin
,
C.
Lazik
,
V.
Banthia
,
B.
Colombeau
,
N.
Variam
, and
A.
Mayur
, “
Ultra-low contact resistivity with highly doped Si: P contact for nMOSFET
,” in
2015 Symposium on VLSI Technology
(
IEEE
,
2015
), p.
T118
.
17.
M.-A.
Nicolet
and
S.
Lau
, in
VLSI Electronics Microstructure Science
(
Elsevier
,
1983
), Vol. 6, p.
329
.
18.
C.
Lal
,
R.
Dhunna
, and
I.
Jain
,
Vacuum
83
,
931
(
2009
).
19.
F.
Meli
,
Z.
Sheng
,
I.
Vedel
, and
L.
Schlapbach
,
vacuum
41
,
1938
(
1990
).
20.
E.
Simoen
,
A.
Rothschild
,
B.
Vermang
,
J.
Poortmans
, and
R.
Mertens
,
Electrochem. Solid-State Lett.
14
,
H362
(
2011
).
21.
R.
Carter
,
E.
Cartier
,
A.
Kerber
,
L.
Pantisano
,
T.
Schram
,
S.
De Gendt
, and
M.
Heyns
,
Appl. Phys. Lett.
83
,
533
(
2003
).
22.
E. H.
Rhoderick
,
IEE Proc., Part I: Solid-State Electron Devices
129
,
1
(
1982
).
23.
M. S.
Arshad
,
Š.
Trafela
,
K. Ž.
Rožman
,
J.
Kovač
,
P.
Djinović
, and
A.
Pintar
,
J. Mater. Chem. C
5
,
10509
(
2017
).
24.
J.-S.
Jang
and
T.-Y.
Seong
,
J. Appl. Phys.
88
,
3064
(
2000
).
25.
P.
Grunthaner
,
F.
Grunthaner
, and
A.
Madhukar
,
Physica B+C
117-118
,
831
(
1983
).
26.
J.
Sun
,
K.
Rickert
,
J. M.
Redwing
,
A.
Ellis
,
F.
Himpsel
, and
T.
Kuech
,
Appl. Phys. Lett.
76
,
415
(
2000
).
27.
M.
Kuznetsov
,
J. F.
Zhuravlev
, and
V.
Gubanov
,
J. Electron Spectrosc. Relat. Phenom.
58
,
169
(
1992
).
28.
Q.
Wang
and
R. J.
Madix
,
Surf. Sci.
474
,
L213
(
2001
).
29.
Q.
Sun
,
Y.
Fu
,
J.
Liu
,
A.
Auroux
, and
J.
Shen
,
Appl. Catal., A
334
,
26
(
2008
).
30.
L.
Brillson
,
M.
Slade
,
H.
Richter
,
H.
VanderPlas
, and
R.
Fulks
,
J. Vac. Sci. Technol., A
4
,
993
(
1986
).
31.
C.-K.
Lee
,
C.-D.
Hsieh
, and
B.-H.
Tseng
,
Thin Solid Films
303
,
232
(
1997
).
32.
C.
Crider
,
J.
Poate
,
J.
Rowe
, and
T.
Sheng
,
J. Appl. Phys.
52
,
2860
(
1981
).
33.
B. Y.
Tsui
and
M. C.
Chen
,
J. Appl. Phys.
76
,
1995
(
1994
).
34.
B.-Y.
Tsui
,
J.-Y.
Tsai
,
T.-S.
Wu
, and
M.-C.
Chen
, and
IEEE Trans. Electron Devices
40
,
54
(
1993
).
35.
A.
Vanderpool
and
M.
Taylor
,
Nucl. Instrum. Methods Phys. Res., Sect. B
237
,
142
(
2005
).
36.
X.
Zhang
,
D.
Connelly
,
H.
Takeuchi
,
M.
Hytha
,
R. J.
Mears
,
L. M.
Rubin
, and
T. J. K.
Liu
,
J. Appl. Phys.
123
,
125704
(
2018
).
37.
X.
Zhang
,
D.
Connelly
,
H.
Takeuchi
,
M.
Hytha
,
R. J.
Mears
,
L. M.
Rubin
, and
T.-J. K.
Liu
,
J. Vac. Sci. Technol., B
36
,
061211
(
2018
).
38.
R. T.
Fryer
and
R. J.
Lad
,
J. Alloys Compd.
682
,
216
(
2016
).
39.
Y.
Kumar
,
M.
Yang
,
K.
Ihm
,
K.-J.
Lee
, and
C. C.
Hwang
,
Chem. Phys. Lett.
608
,
324
(
2014
).