Sensors and Metrology - EECS Instructional Support Group Home Page

Lecture 17: Sensors and Metrology I

SpanosEE290H F05

1

Sensors and Metrology

A Survey


SpanosEE290H F05

2

Outline

• General Issues & the SIA Roadmap• Post-Process Sensing (SEM/AFM, placement)• In-Process (or potential in-process) Sensors

– temperature (pyrometry, thermocouples, acoustic waves)– pressure and flow (manometers, momentum devices)– composition (OES, LIF, RGA, mass Spectroscopy Actinometry)– thickness (reflectometry, ellipsometry, scatterometry)– smart-dummy wafers and smart substrates


SpanosEE290H F05

3

Introduction

• Sensors (and actuators) are key limiting factors in application of control techniques to semiconductor manufacturing

• sources of difficulty– implementation environment (vacuum, clean facilities, etc.)– perception that in-situ sensors affect process– ex-situ sensors can reduce throughput– cost of ownership– traditional resistance in industry


SpanosEE290H F05

4

General Remarks on Sensors

• modeling is often key part of sensing– physical quantity of interest may not be directly measured (ex:

OES indirectly contains info about etch process state)– thus, sensors are based on a model of the underlying physical

process

sensors = data + model

• signal processing– needed to reduce noise, improve bandwidth– difference between data and information

• problems– sensors require calibration– must account for drift

• other issues- sensor fusion- data compression


SpanosEE290H F05

5

Key Issues in Sensors

• some key tradeoffs– non-invasive vs. invasive– non-destructive vs. destructive– in-situ vs. ex-situ– speed vs. accuracy– noise

• bias (accuracy) vs. precision (repeatability + reproducibility)– a sensor could be inaccurate, (thermocouple readings are off by 4o K)– but the sensor might have good precision, (it is consistently off)– precision is often more important for process control

• modern filtering and estimation methods can be of great use in improved sensing software.


SpanosEE290H F05

6

Accuracy vs. Precision


SpanosEE290H F05

7

SIA Road Map Challenges above 45nm (through 2009)


SpanosEE290H F05

8

SIA Road Map Challenges below 45nm (beyond 2009)


SpanosEE290H F05

9

The 2004 update Metrology Road Map – near term


SpanosEE290H F05

10

The 2004 update Metrology Road Map – long term


SpanosEE290H F05

11

The 2004 update Lithography Metrology Road Map –near term


SpanosEE290H F05

12

The 2004 update Lithography Metrology Road Map – long term


SpanosEE290H F05

13

CD Metrology

• CD-SEM is today’s pre-eminent technique…• Electron yield in interaction volume is a function of

surface topography (Secant effect) and atomic number.• Extracting CD is not so simple...


SpanosEE290H F05

14

Various Edge Detection Algorithms are in use...


SpanosEE290H F05

15

CD-SEM is often calibrated with AFM

Atomic force Microscopy uses either repulsive forces (sub nm proximity), or weaker attractive forces (a few nm away).

Tip tracks surfaces using feedback control.

Shape and size of tip is the critical source of errors.

Standard features are use to calibrate and “de-convolve” the tip profile from the measurements.

AFM is 100-1000 times slower than a CD-SEM.

AFM is sensitive to line-edge roughness.


SpanosEE290H F05

16

CD-SEM vs. AFM standards


SpanosEE290H F05

17

CD-SEM cannot “see” the actual profile, so AFM is used for this purpose...


SpanosEE290H F05

18

AFM/SEM reading of Dense vs. Isolated Lines

dense

iso

AFM SEM


SpanosEE290H F05

19

Comparison for Contact holesAFM SEM


SpanosEE290H F05

20

More CD-SEM vs. AFM comparisons


SpanosEE290H F05

21

Optical CD MeasurementI

• Not very repeatable

• Limited spatial resolution

• Relatively inexpensive


SpanosEE290H F05

22

Electrical CD Measurement

Measurement is very repeatable and fast.Can only be used in conductive layers. Need at least one conductive layer and one insulator.Can be extended for misalignment measurements.

SEMLines

ID: XXNGratingHorz Iso

W/S= 180/1000

GratingVert Iso

W/S= 180/1000

GratingHorz Iso

W/S= 180/360

GratingVert Iso

W/S= 180/360

GratingHorz

MediumW/S = 180/

240

GratingVert MediumW/S = 180/

240

GratingHorz DenseW/S = 180/

180

GratingVert DenseW/S = 180/

180

DUT1Vert

DUT2Vert

DUT3Horz

DUT4Horz VDP

DUT5Horz

DUT6Horz

DUT7Vert

DUT8Vert

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

31

30 29 28 27 26 25 24 23 22 21 20 19 18 17 16

32

SEMLines

ID: XXO

DUT1Vert

DUT2Vert

DUT3Horz

DUT4Horz VDP

DUT5Horz

DUT6Horz

DUT7Vert

DUT8Vert

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

31

30 29 28 27 26 25 24 23 22 21 20 19 18 17 16

32

SEMLines

CF_LENS coma/flare

CD_LIN linearity

MEFH

MEF


SpanosEE290H F05

23

Basic SEM Structure


SpanosEE290H F05

24

The Many Modes of SEM


SpanosEE290H F05

25

The Issue of Spatial Resolution


SpanosEE290H F05

26

Chrome on Silicon Example


SpanosEE290H F05

27

Signal Depends Strongly on Material


SpanosEE290H F05

28

Voltage Contrast SEM


SpanosEE290H F05

29

CD-SEM Resolution

Scanning Resolution shown to 1-5nm.

CD metrology on resist has 5-10nm precision.

Other solutions:

ATF, has (theoretically) atomic resolution.

Problem:

What is CD??


SpanosEE290H F05

30

CD-SEM Conclusions

• Accuracy is several (many) nm• Precision (1-σ repeatability + reproducibility) is

1-2nm today• CD-SEM is stand-alone (i.e. expensive)• CD-SEM measurements are available only after

patterning, and data integration with control systems is difficult at best.

• AFM-based calibration will not be possible for trenches less than ~100nm wide.


SpanosEE290H F05

31

State of the Art CD/Imaging-SEM

Documents

Sensors and Metrology - EECS Instructional Support Group Home Page