leakage monitoring and control with an advanced e-beam

16
Leakage Monitoring and Control with an Advanced e-beam Inspection System Hermes Liu, J H Yeh, Chan Lon Yang, S C Lei, J Y Kao, Y D Yang, Mingsheng Tsai, S F Tzou United Microelectronics Corp. Hong Xiao and Jack Jau Hermes Microvision, Inc. USA,

Upload: others

Post on 29-Nov-2021

3 views

Category:

Documents


0 download

TRANSCRIPT

Leakage Monitoring and Control

with an Advanced e-beam

Inspection System

Hermes Liu, J H Yeh, Chan Lon Yang, S C Lei, J Y Kao, Y D Yang, Mingsheng Tsai, S F Tzou

United Microelectronics Corp.

Hong Xiao and Jack Jau

Hermes Microvision, Inc. USA,

List of Topics

� Introduction to EBI

�Principle of detection

�Experiment arrangement

�Experiment results

�Conclusions

Primary Beam

Secondary and

Back Scattered

Electrons

Illustration of eScan®300

LE (eV)

Back Scattered Electron

Secondary

Electron

Secondary Electron and

Back Scattered Electron

Positive Mode

Poly

Poly

N+ P+

P N

Dark Bright Dark

eGate Oxide

Normal

Bright

Poly

N+

N

Bright

e P+

P

Bright

e

N+ P+

P N

Bright Bright Bright

eGate Oxide

Leakage

(BVC

defects)

Bright

Poly

N+

N

Bright

e P+

P

Bright

eee

e

This study

W-Plugs & VC in Positive Mode

Experiment 1

� 6 wafers

� 3 ion implantation conditions

� After WCMP

�One wafer from each condition send to

inspection

� Another one continue to M1 for WAT

� Inspected and measured test keys and

SRAM arrays

Leak

age C

urren

tBVC Defect C

ount

Split 1 Split 2 Split 3

BVC Defect

Correlation of WAT and eScan®

Correlation of WAT and eScan®Defect count

Split 1 Split 2 Split 3

Split 1 Split 2 Split 3

GLV Binning

� Strong leakage causes high GLV

� Auto binning GLV of all BVC defects

� High GLV BVC map matches with high

leakage current map very well.

Bin 1

WAT Leakage Map and eScan BVC Map

GLV1 GLV2

GLV3 GLV4

Extrusion Length vs eScan GLV

Extrusion length GL vs NISI piping extruded length

L3 L4

L2L1

0

10

20

30

40

50

60

GLV4 GLV3 GLV2 GLV1

NiSi Diffusion

Leakage is caused by NiSi diffusion

SEM TEM

Experiment 2

� 13 wafers

� 3 ion implantation conditions

� eScan inspection on two test keys

�WAT after M1 CMP

BVC counts vs. WAT Leakage

Leakage current

BVC counts

R2 = 0.8079

R2 = 0.7464Test key 1 BVC Counts

Test key 2 BVC Counts

Test

key 1

Test

key 2Leakage current

WAT

Conclusion

� BVC defects of W-plugs that connect to NiSi on N+/P-well junction on both test key structures and SRAM array

� BVC defect counts correlate with WAT leakage current on both test key and SRAM

� FA delayer showed that N+/P-well leakage is caused by nickel silicide diffusion

� Extrusion length of the nickel silicide strongly correlates with GLV of the W-plug

� BVC counts strongly correlate to WAT leakage current

Development

� More EBI systems:� Leap Scan EBI : eScan®310, eScan®315 (Improve

resolution)

� Continue scan EBI : eScan®Lite (Improve throughput)

� Hotspot EBI: eP2

� More EBI applications� Monitoring hole non-open at contact AEI and WCMP.

� Qualifying the contact lithography process window.

� Negative ModeTM EBI for P+/N-well leakage detection.

� Monitoring NMOS and gate leakage right after NiSi formation.