Contact Model Based on TCAD-Experimental Interactive Algorithm

Peijie Feng Jiseok Kimdagger Jin Cho Shesh Mani Pandey Sudarshan Narayanan Michelle TngDagger Bingwu Liu Edmund

Banghart Baofu Zhu Pei Zhao Muhammad Rahman Yumi Park Liu Jiang and Francis BenistantDagger GLOBALFOUNDRIES Malta NY 12020 USA PeijieFengglobalfoundriescom

daggerGLOBALFOUNDRIES Albany NY 12203 USA DaggerGLOBALFOUNDRIES Singapore 528830 Singapore

AbstractmdashThis work demonstrated a novel method utilizing Sentaurus Technology Computer Aided Design simulation along with experiments to intermediately extract Schottky barrier height and contact resistance in FinFETs The proposed algorithm can automatically calibrate contact model based on measurement data This interactive contact model is also capable of prediction of contact resistance sensitivity including key process features such as implant energy dose and thermal process based on a design of experiment splits This robust physical and efficient contact model provides insightful understandings of the metal-semiconductor contact in FinFETs It can be easily implemented in simulation tools for device design in state-of-art semiconductor technology development

KeywordsmdashSchottky Contact Barrier Height FinFET TCAD Interactive Algorithm

I INTRODUCTION With the scaling of MOSFET transistors the source and

drain (SD) contact size is aggressively reduced and the contact silicide formation faces many process challenges Due to the small effective contact area and the difficulties in lowering the silicide resistivity the SD contact resistance RC is becoming one of the major bottlenecks limiting device performance for future technology node 3-D devices [1-4] To drive the FinFET transistor performance following Moorersquos law a robust and physical contact model and good understanding of the metal-semiconductor Schottky barrier is crucial in the early stage technology development to capture RC sensitivity to process flow towards device performance optimization [5] However it is a difficult task to estimate the Schottky barrier height (SBH) in the device level with conventional contact model due to the very complex nature of the SD interfaces [6] And the barrier lowering model accounting for image force and tunneling effects can cause convergence issues in 3-D transistors simulations In addition the conventional contact model lacks an automatic calibration technique to hardware data

In this study we present a novel method which utilizes Sentaurus Technology Computer Aided Design (TCAD) simulation along with experiments to extract SBH in FinFETs The proposed algorithm can automatically calibrate contact model based on measurement data and then predicts contact resistance TCAD simulation results show that the interactive contact model is able to reproduce RC sensitivity by intended design of experiment (DoE) splits

II ALGORITHM As shown in Fig 1 SD contact resistance RC

Exp is obtained with a normalized value around 40 Ωmiddotmicrom by Kelvin contact resistance measurement in the first step Then the calibrated process simulation extracts the surface doping density NC

TCAD over the contact area ATCAD Fig 2 illustrates that TCAD dopant profile is consistent with as-implanted phosphorus dopant data from SIMS measurement It should be noted that during the silicide contact formation more than 5-nm-thick silicon will be consumed and therefore a small doping concentration variation over the contact surface can be considered after silicidation

Next an iterative computation of ρCTCAD via device

simulation is performed until that a match of the TCAD result RC

TCAD which is a product of ρCTCAD and ATCAD to the

experimental data RCExp is obtained

Fig 1 The flow chart of the TCAD-Experimental contact model algorithm extracts the resistivity dopant dependent equation for TCAD simulations

SISPAD 2015 September 9-11 2015 Washington DC USA

SISPAD 2015 - httpwwwsispadorg

238978-1-4673-7860-415$3100 copy2015 IEEE

Having obtained ρCTCAD and NC

TCAD SBHthe fourth step using the theoretically characteristics For this calculation one-dimeconsistent Schottky barrier potential profile aheight in the semiconductor region is assumdensity across the barrier is calculated usinformula wherein barrier tunneling probabilitythe transfer matrix method [6-7] Then the conobtained from the current voltage characteristicapplied bias By repeating the procedure abdependent ρC characteristics are computed avalues With the theoretical calculated SBH-can map the interested NFET and PFET (NC

TCAD ρCTCAD) and P (NC

TCAD ρCTCAD) to ext

interface barrier height The examples are shand 32 NFET SBH is estimated to be 02 eVeV

In the fifth step a simple dopant concentρC fitting equation suitable for device electricextracted within specific range of doping dcontact surface doping level The equation follmiddot exp

where ρO and NO are fitting parameters interactive fine tuning to ensure RC

TCAD equalperformed on those two parameters to obtainand NO

Cal through the equation below in the sixmiddot exp

The TCAD-experimental interactive andescribed above directly captures the sensitivitcontact surface doping concentration Fimplementation of this simple equation in Tconvergence challenges from the direct modelof image force tunneling and dipole effects

Fig 2 Calibrated as-implanted phosphorus profmatches SIMS data The silicidation consumes silicon at the Epi surface and therefore the kink ccontact modelling

H is extracted in y calculated ρC ensional non-self-at a given barrier med The current ng the Tsu-Esaki y is evaluated by ntact resistivity is cs at a very small bove the doping at different SBH curve group one contact data N

tract their contact hown in Fig 31

V while PFET 06

tration dependent cal simulation is

density related to lows the form

The automatic l to RC

Exp will be n calibrated ρO

Cal xth step

nalytic equation ty of resistivity to

Furthermore the TCAD eliminates ing consideration

file in Epi region around 5nm thick can be ignored for

Fig 31 Map the calibrated the NFET cto theoretically computed SBH curve gSBH as 02 eV

Fig 32 Map the calibrated the PFET cotheoretically computed SBH curve grouas 06 eV

III RESULT

Fig 4 shows the TEM of the Eand the contact interface in the simulation mimics the silicide shacontact surface doping concentratio1e20 cm-3

An average surface doping concanalysis method over the contact surlocal doping density to determine Nfact that (1) the contact surface dapproximated as a uniform distributas stated previously and (2) the ssimulation efficiency and robustness

Experimental results are used asverify the proposed model The Do

coordinate N (NC

TCAD ρCTCAD)

group and then extract NFET

oordinate P (NC

TCAD ρCTCAD) to

up and then extract PFET SBH

TS Epi (the diamond) structure

FinFET TCAD process ape (green region) and the on which is of the order of

centration by finite element rface is extracted instead of NC

TCAD This is due to the doping level can be fairly tion around the contact area simplification can improve s

s a reference to interact and Es are on SD implantation

239

with dose and energy variation and SD dbudget to examine the contact sensitivity to doping concentration For the DoEs the S0 trstandard NFET and PFET devices which aTCAD to hardware data for both linearoperation region (see Fig 51 and 52) wresistance matching measurement results Sanalytic contact model with certain SBH is accbased on the S0 transistor

Fig 4 TCAD process simulation mimic TEMs Epdiamond pink region is for SD Epi region the greyellow for metal the grey for STI and the brown fo

Fig51 Calibrated NFET simulation results agree w

diffusion thermal contact interface ransistors are the are calibrated in r and saturation

with their contact Subsequently the curately extracted

pi and Contact the een for silicide the or nitride

with electrical data

Table 11 and Table 12 summPFET front end of line (FEOL) Dtreated as a constant This assumptcontact electrode materials do notsurface doping density is higher significant tunneling current throaccording to the criterion compariwith E00 defined as

2where q is the electron charge ħ thεS the silicon dielectric constant andmass [8]

The review of the table showsextracted contact resistance values ntrend among experiments but alsodata with an error less than 2

The DoE sensitivity provides dFinFETs It suggests that (1) shaconcentration around the contactresistance (2) contact surface dopiSD implant dose and (3) increadoping distribution and lowers concentration The reliable simusensitivity reminds the designer ocontact the junction and other devic

Fig52 Calibrated PFET simulation resu

marize results of NFET and DoEs where SBH can be tion is valid as long as the t change and the contact than 1e20 cm-3 allowing

ugh the contact interface ing the thermal energy kT

e reduced Plankrsquos constant d mT the tunneling effective

s that the predicted TCAD not only shows a consistent o well match experimental

device design guideline for allow implant retains peak t interface to reduce the ing level is proportional to ased thermal budget alters

contact surface doping ulated contact resistivity

of the trade-off among the ce factors

ults agree with electrical data

240

NFET Contact Related Condition TCAD ExpS0 Standard As 3e15cm-2 3keV Imp w RTA 401 402 S1 Energy As 3e15cm-2 2keV Imp w RTA 395 396 S2 Dose As 5e15cm-2 3keV Imp w RTA 353 350 S3 Thermal As 3e15cm-2 3keV Imp wo RTA 371 377

Table 11 NFET DoE comparison results show that TCAD simulation capture experimental sensitivity on Epi implantation dose and energy and rapid thermal anneal (RTA) the normalized contact resistance result has a unit of Ωμm

PFET Contact Related Condition TCAD ExpS0 Standard B 4e15cm-2 2keV Imp w RTA 473 477 S1 Energy B 4e15cm-2 15keV Imp w RTA 462 468 S2 Dose B 5e15cm-2 2keV Imp w RTA 435 428 S3 Thermal B 4e15cm-2 2keV Imp wo RTA 400 392

Table 12 PFET DoE comparison results show that TCAD simulation capture experimental sensitivity on Epi implantation dose and energy and rapid thermal anneal (RTA) the normalized contact resistance result has a unit of Ωμm

IV CONCLUSION We have demonstrated a novel contact model based on

TCAD-Experimental interactive algorithm with theoretical analysis and DoE results verification The proposed model is capable of extracting SBH and it accurately predicts sensitivity of contact resistivity to variation in doping concentration By interactive calibration method in the model our TCAD simulation results accurately agree with hardware data based on FinFET technology The implementation of the model is simple and will improve both simulation efficiency and

robustness This model can be extended to nanowire and future 3-D devices

ACKNOWLEDGMENT The authors would like to thank Brendon Murphy in

Synopsys Inc for valuable discussion on Sentaurus TCAD

REFERENCES

[1] E Yu L Chang S Ahmed H Wang S Bell C Yang C Tabery et al ldquoFinFET scaling to 10 nm gate lengthrdquo Electron Devices Meeting 2002 IEDM02 International pp 251-254 IEEE 2002

[2] ldquoInternational technology roadmap for semiconductorsrdquo 2013 [Online] Available httpwwwitrsnet

[3] P Feng ldquoDesign Modeling and Analysis of Non-classical Field Effect Transistorsrdquo (2012)

[4] S Narayanan et al ldquoExtraction of Parasitic and Channel Resistance Components in FinFETs using TCAD toolsrdquo Electron Devices IEEE Transactions on to appear

[5] R T Tung ldquoThe physics and chemistry of the Schottky barrier heightrdquo Applied Physics Reviews vol 1 no 1 (2014) 011304

[6] B Jonsson and S T Sverre ldquoSolving the Schrodinger equation in arbitrary quantum-well potential profiles using the transfer matrix methodrdquo Quantum Electronics IEEE Journal of vol 26 no 11 (1990) 2025-2035

[7] J Kim B Lee Y Park K Murali and F Benistant ldquoSpecific contact resistivity of n-type Si and Ge M-S and M-I-S contactsrdquo Simulation of Semiconductor Processes and Devices (SISPAD) 2015 International Conference on Sep 2015

[8] F A Padovani and R Stratton ldquoField and thermionic-field emission in Schottky barriersrdquo Solid-State Electronics vol 9 no 7 (1966) 695-707

241

Having obtained ρCTCAD and NC

TCAD SBHthe fourth step using the theoretically characteristics For this calculation one-dimeconsistent Schottky barrier potential profile aheight in the semiconductor region is assumdensity across the barrier is calculated usinformula wherein barrier tunneling probabilitythe transfer matrix method [6-7] Then the conobtained from the current voltage characteristicapplied bias By repeating the procedure abdependent ρC characteristics are computed avalues With the theoretical calculated SBH-can map the interested NFET and PFET (NC

TCAD ρCTCAD) and P (NC

TCAD ρCTCAD) to ext

interface barrier height The examples are shand 32 NFET SBH is estimated to be 02 eVeV

In the fifth step a simple dopant concentρC fitting equation suitable for device electricextracted within specific range of doping dcontact surface doping level The equation follmiddot exp

where ρO and NO are fitting parameters interactive fine tuning to ensure RC

TCAD equalperformed on those two parameters to obtainand NO

Cal through the equation below in the sixmiddot exp

The TCAD-experimental interactive andescribed above directly captures the sensitivitcontact surface doping concentration Fimplementation of this simple equation in Tconvergence challenges from the direct modelof image force tunneling and dipole effects

Fig 2 Calibrated as-implanted phosphorus profmatches SIMS data The silicidation consumes silicon at the Epi surface and therefore the kink ccontact modelling

H is extracted in y calculated ρC ensional non-self-at a given barrier med The current ng the Tsu-Esaki y is evaluated by ntact resistivity is cs at a very small bove the doping at different SBH curve group one contact data N

tract their contact hown in Fig 31

V while PFET 06

tration dependent cal simulation is

density related to lows the form

The automatic l to RC

Exp will be n calibrated ρO

Cal xth step

nalytic equation ty of resistivity to

Furthermore the TCAD eliminates ing consideration

file in Epi region around 5nm thick can be ignored for

Fig 31 Map the calibrated the NFET cto theoretically computed SBH curve gSBH as 02 eV

Fig 32 Map the calibrated the PFET cotheoretically computed SBH curve grouas 06 eV

III RESULT

Fig 4 shows the TEM of the Eand the contact interface in the simulation mimics the silicide shacontact surface doping concentratio1e20 cm-3

An average surface doping concanalysis method over the contact surlocal doping density to determine Nfact that (1) the contact surface dapproximated as a uniform distributas stated previously and (2) the ssimulation efficiency and robustness

Experimental results are used asverify the proposed model The Do

coordinate N (NC

TCAD ρCTCAD)

group and then extract NFET

oordinate P (NC

TCAD ρCTCAD) to

up and then extract PFET SBH

TS Epi (the diamond) structure

FinFET TCAD process ape (green region) and the on which is of the order of

centration by finite element rface is extracted instead of NC

TCAD This is due to the doping level can be fairly tion around the contact area simplification can improve s

s a reference to interact and Es are on SD implantation

239

with dose and energy variation and SD dbudget to examine the contact sensitivity to doping concentration For the DoEs the S0 trstandard NFET and PFET devices which aTCAD to hardware data for both linearoperation region (see Fig 51 and 52) wresistance matching measurement results Sanalytic contact model with certain SBH is accbased on the S0 transistor

Fig 4 TCAD process simulation mimic TEMs Epdiamond pink region is for SD Epi region the greyellow for metal the grey for STI and the brown fo

Fig51 Calibrated NFET simulation results agree w

diffusion thermal contact interface ransistors are the are calibrated in r and saturation

with their contact Subsequently the curately extracted

pi and Contact the een for silicide the or nitride

with electrical data

Table 11 and Table 12 summPFET front end of line (FEOL) Dtreated as a constant This assumptcontact electrode materials do notsurface doping density is higher significant tunneling current throaccording to the criterion compariwith E00 defined as

2where q is the electron charge ħ thεS the silicon dielectric constant andmass [8]

The review of the table showsextracted contact resistance values ntrend among experiments but alsodata with an error less than 2

The DoE sensitivity provides dFinFETs It suggests that (1) shaconcentration around the contactresistance (2) contact surface dopiSD implant dose and (3) increadoping distribution and lowers concentration The reliable simusensitivity reminds the designer ocontact the junction and other devic

Fig52 Calibrated PFET simulation resu

marize results of NFET and DoEs where SBH can be tion is valid as long as the t change and the contact than 1e20 cm-3 allowing

ugh the contact interface ing the thermal energy kT

e reduced Plankrsquos constant d mT the tunneling effective

s that the predicted TCAD not only shows a consistent o well match experimental

device design guideline for allow implant retains peak t interface to reduce the ing level is proportional to ased thermal budget alters

contact surface doping ulated contact resistivity

of the trade-off among the ce factors

ults agree with electrical data

240

NFET Contact Related Condition TCAD ExpS0 Standard As 3e15cm-2 3keV Imp w RTA 401 402 S1 Energy As 3e15cm-2 2keV Imp w RTA 395 396 S2 Dose As 5e15cm-2 3keV Imp w RTA 353 350 S3 Thermal As 3e15cm-2 3keV Imp wo RTA 371 377

Table 11 NFET DoE comparison results show that TCAD simulation capture experimental sensitivity on Epi implantation dose and energy and rapid thermal anneal (RTA) the normalized contact resistance result has a unit of Ωμm

PFET Contact Related Condition TCAD ExpS0 Standard B 4e15cm-2 2keV Imp w RTA 473 477 S1 Energy B 4e15cm-2 15keV Imp w RTA 462 468 S2 Dose B 5e15cm-2 2keV Imp w RTA 435 428 S3 Thermal B 4e15cm-2 2keV Imp wo RTA 400 392

Table 12 PFET DoE comparison results show that TCAD simulation capture experimental sensitivity on Epi implantation dose and energy and rapid thermal anneal (RTA) the normalized contact resistance result has a unit of Ωμm

IV CONCLUSION We have demonstrated a novel contact model based on

TCAD-Experimental interactive algorithm with theoretical analysis and DoE results verification The proposed model is capable of extracting SBH and it accurately predicts sensitivity of contact resistivity to variation in doping concentration By interactive calibration method in the model our TCAD simulation results accurately agree with hardware data based on FinFET technology The implementation of the model is simple and will improve both simulation efficiency and

robustness This model can be extended to nanowire and future 3-D devices

ACKNOWLEDGMENT The authors would like to thank Brendon Murphy in

Synopsys Inc for valuable discussion on Sentaurus TCAD

REFERENCES

[1] E Yu L Chang S Ahmed H Wang S Bell C Yang C Tabery et al ldquoFinFET scaling to 10 nm gate lengthrdquo Electron Devices Meeting 2002 IEDM02 International pp 251-254 IEEE 2002

[2] ldquoInternational technology roadmap for semiconductorsrdquo 2013 [Online] Available httpwwwitrsnet

[3] P Feng ldquoDesign Modeling and Analysis of Non-classical Field Effect Transistorsrdquo (2012)

[4] S Narayanan et al ldquoExtraction of Parasitic and Channel Resistance Components in FinFETs using TCAD toolsrdquo Electron Devices IEEE Transactions on to appear

[5] R T Tung ldquoThe physics and chemistry of the Schottky barrier heightrdquo Applied Physics Reviews vol 1 no 1 (2014) 011304

[6] B Jonsson and S T Sverre ldquoSolving the Schrodinger equation in arbitrary quantum-well potential profiles using the transfer matrix methodrdquo Quantum Electronics IEEE Journal of vol 26 no 11 (1990) 2025-2035

[7] J Kim B Lee Y Park K Murali and F Benistant ldquoSpecific contact resistivity of n-type Si and Ge M-S and M-I-S contactsrdquo Simulation of Semiconductor Processes and Devices (SISPAD) 2015 International Conference on Sep 2015

[8] F A Padovani and R Stratton ldquoField and thermionic-field emission in Schottky barriersrdquo Solid-State Electronics vol 9 no 7 (1966) 695-707

241

with dose and energy variation and SD dbudget to examine the contact sensitivity to doping concentration For the DoEs the S0 trstandard NFET and PFET devices which aTCAD to hardware data for both linearoperation region (see Fig 51 and 52) wresistance matching measurement results Sanalytic contact model with certain SBH is accbased on the S0 transistor

Fig 4 TCAD process simulation mimic TEMs Epdiamond pink region is for SD Epi region the greyellow for metal the grey for STI and the brown fo

Fig51 Calibrated NFET simulation results agree w

diffusion thermal contact interface ransistors are the are calibrated in r and saturation

with their contact Subsequently the curately extracted

pi and Contact the een for silicide the or nitride

with electrical data

Table 11 and Table 12 summPFET front end of line (FEOL) Dtreated as a constant This assumptcontact electrode materials do notsurface doping density is higher significant tunneling current throaccording to the criterion compariwith E00 defined as

2where q is the electron charge ħ thεS the silicon dielectric constant andmass [8]

The review of the table showsextracted contact resistance values ntrend among experiments but alsodata with an error less than 2

The DoE sensitivity provides dFinFETs It suggests that (1) shaconcentration around the contactresistance (2) contact surface dopiSD implant dose and (3) increadoping distribution and lowers concentration The reliable simusensitivity reminds the designer ocontact the junction and other devic

Fig52 Calibrated PFET simulation resu

marize results of NFET and DoEs where SBH can be tion is valid as long as the t change and the contact than 1e20 cm-3 allowing

ugh the contact interface ing the thermal energy kT

e reduced Plankrsquos constant d mT the tunneling effective

s that the predicted TCAD not only shows a consistent o well match experimental

device design guideline for allow implant retains peak t interface to reduce the ing level is proportional to ased thermal budget alters

contact surface doping ulated contact resistivity

of the trade-off among the ce factors

ults agree with electrical data

240

NFET Contact Related Condition TCAD ExpS0 Standard As 3e15cm-2 3keV Imp w RTA 401 402 S1 Energy As 3e15cm-2 2keV Imp w RTA 395 396 S2 Dose As 5e15cm-2 3keV Imp w RTA 353 350 S3 Thermal As 3e15cm-2 3keV Imp wo RTA 371 377

Table 11 NFET DoE comparison results show that TCAD simulation capture experimental sensitivity on Epi implantation dose and energy and rapid thermal anneal (RTA) the normalized contact resistance result has a unit of Ωμm

PFET Contact Related Condition TCAD ExpS0 Standard B 4e15cm-2 2keV Imp w RTA 473 477 S1 Energy B 4e15cm-2 15keV Imp w RTA 462 468 S2 Dose B 5e15cm-2 2keV Imp w RTA 435 428 S3 Thermal B 4e15cm-2 2keV Imp wo RTA 400 392

Table 12 PFET DoE comparison results show that TCAD simulation capture experimental sensitivity on Epi implantation dose and energy and rapid thermal anneal (RTA) the normalized contact resistance result has a unit of Ωμm

IV CONCLUSION We have demonstrated a novel contact model based on

TCAD-Experimental interactive algorithm with theoretical analysis and DoE results verification The proposed model is capable of extracting SBH and it accurately predicts sensitivity of contact resistivity to variation in doping concentration By interactive calibration method in the model our TCAD simulation results accurately agree with hardware data based on FinFET technology The implementation of the model is simple and will improve both simulation efficiency and

robustness This model can be extended to nanowire and future 3-D devices

ACKNOWLEDGMENT The authors would like to thank Brendon Murphy in

Synopsys Inc for valuable discussion on Sentaurus TCAD

REFERENCES

[1] E Yu L Chang S Ahmed H Wang S Bell C Yang C Tabery et al ldquoFinFET scaling to 10 nm gate lengthrdquo Electron Devices Meeting 2002 IEDM02 International pp 251-254 IEEE 2002

[2] ldquoInternational technology roadmap for semiconductorsrdquo 2013 [Online] Available httpwwwitrsnet

[3] P Feng ldquoDesign Modeling and Analysis of Non-classical Field Effect Transistorsrdquo (2012)

[4] S Narayanan et al ldquoExtraction of Parasitic and Channel Resistance Components in FinFETs using TCAD toolsrdquo Electron Devices IEEE Transactions on to appear

[5] R T Tung ldquoThe physics and chemistry of the Schottky barrier heightrdquo Applied Physics Reviews vol 1 no 1 (2014) 011304

[6] B Jonsson and S T Sverre ldquoSolving the Schrodinger equation in arbitrary quantum-well potential profiles using the transfer matrix methodrdquo Quantum Electronics IEEE Journal of vol 26 no 11 (1990) 2025-2035

[7] J Kim B Lee Y Park K Murali and F Benistant ldquoSpecific contact resistivity of n-type Si and Ge M-S and M-I-S contactsrdquo Simulation of Semiconductor Processes and Devices (SISPAD) 2015 International Conference on Sep 2015

[8] F A Padovani and R Stratton ldquoField and thermionic-field emission in Schottky barriersrdquo Solid-State Electronics vol 9 no 7 (1966) 695-707

241

NFET Contact Related Condition TCAD ExpS0 Standard As 3e15cm-2 3keV Imp w RTA 401 402 S1 Energy As 3e15cm-2 2keV Imp w RTA 395 396 S2 Dose As 5e15cm-2 3keV Imp w RTA 353 350 S3 Thermal As 3e15cm-2 3keV Imp wo RTA 371 377

Table 11 NFET DoE comparison results show that TCAD simulation capture experimental sensitivity on Epi implantation dose and energy and rapid thermal anneal (RTA) the normalized contact resistance result has a unit of Ωμm

PFET Contact Related Condition TCAD ExpS0 Standard B 4e15cm-2 2keV Imp w RTA 473 477 S1 Energy B 4e15cm-2 15keV Imp w RTA 462 468 S2 Dose B 5e15cm-2 2keV Imp w RTA 435 428 S3 Thermal B 4e15cm-2 2keV Imp wo RTA 400 392

Table 12 PFET DoE comparison results show that TCAD simulation capture experimental sensitivity on Epi implantation dose and energy and rapid thermal anneal (RTA) the normalized contact resistance result has a unit of Ωμm

IV CONCLUSION We have demonstrated a novel contact model based on

TCAD-Experimental interactive algorithm with theoretical analysis and DoE results verification The proposed model is capable of extracting SBH and it accurately predicts sensitivity of contact resistivity to variation in doping concentration By interactive calibration method in the model our TCAD simulation results accurately agree with hardware data based on FinFET technology The implementation of the model is simple and will improve both simulation efficiency and

robustness This model can be extended to nanowire and future 3-D devices

ACKNOWLEDGMENT The authors would like to thank Brendon Murphy in

Synopsys Inc for valuable discussion on Sentaurus TCAD

REFERENCES

[1] E Yu L Chang S Ahmed H Wang S Bell C Yang C Tabery et al ldquoFinFET scaling to 10 nm gate lengthrdquo Electron Devices Meeting 2002 IEDM02 International pp 251-254 IEEE 2002

[2] ldquoInternational technology roadmap for semiconductorsrdquo 2013 [Online] Available httpwwwitrsnet

[3] P Feng ldquoDesign Modeling and Analysis of Non-classical Field Effect Transistorsrdquo (2012)

[4] S Narayanan et al ldquoExtraction of Parasitic and Channel Resistance Components in FinFETs using TCAD toolsrdquo Electron Devices IEEE Transactions on to appear

[5] R T Tung ldquoThe physics and chemistry of the Schottky barrier heightrdquo Applied Physics Reviews vol 1 no 1 (2014) 011304

[6] B Jonsson and S T Sverre ldquoSolving the Schrodinger equation in arbitrary quantum-well potential profiles using the transfer matrix methodrdquo Quantum Electronics IEEE Journal of vol 26 no 11 (1990) 2025-2035

[7] J Kim B Lee Y Park K Murali and F Benistant ldquoSpecific contact resistivity of n-type Si and Ge M-S and M-I-S contactsrdquo Simulation of Semiconductor Processes and Devices (SISPAD) 2015 International Conference on Sep 2015

[8] F A Padovani and R Stratton ldquoField and thermionic-field emission in Schottky barriersrdquo Solid-State Electronics vol 9 no 7 (1966) 695-707

241