Metryx Mass Metrology For Tsv (Icep2009)

1Metryx Copyright ©

Characterisation of Through Silicon Via (TSV) processes

utilising Mass Metrology

Liam Cunnane, Adrian Kiermasz PhD, Gary DitmerMetryx Ltd., Bristol UK

2Metryx Copyright ©

Mass MetrologyPrinciplesMethodology

Through-Silicon Via (TSV) Process SequenceDeep Silicon Etch & Polymer cleanOxide LinerBarrier & Seed DepositionCopper ECP & CMP

Summary

Outline

3Metryx Copyright ©

Normal Distribution

x

dens

ity

43 45 47 49 51 53 55 57 59 61 630

0.1

0.2

0.3

0.4

Normal Distribution

x

dens

ity

43 45 47 49 51 53 55 57 59 61 630

0.1

0.2

0.3

0.4

Normal Distribution

x

dens

ity

43 45 47 49 51 53 55 57 59 61 630

0.1

0.2

0.3

0.4

Normal Distribution

x

dens

ity

43 45 47 49 51 53 55 57 59 61 630

0.1

0.2

0.3

0.4

ALD

CMP

ETCHNormal Distribution

x

dens

ity

43 45 47 49 51 53 55 57 59 61 630

0.1

0.2

0.3

0.4

PECVD PVD

Normal Distribution

x

dens

ity

43 45 47 49 51 53 55 57 59 61 630

0.1

0.2

0.3

0.4

ETCHΔ

Mas

s

Process Step

Principle of Mass Metrology

All Microelectronic devices are manufactured through a series ofprocess steps which add or remove material.

Accurate measurement of mass change allows production monitoring orsupports development in determining of a layer’s physical parameters.

Mass metrology provides the advantages of On-product measurement, an atomic-level sensitivity, and total flexibility in application.

4Metryx Copyright ©

Mass ≠ WeightWeight Measurement

Unstable, irreproducible, not designed for semiconductor measurement use

Mass MeasurementLoad-cell utilising complex force measurement Real-time corrections for internal and external forces influencing measurementFully automatic wafer handling and host communication compliant

5Metryx Copyright ©

Expressed as thickness on a 300mm wafer

Standard Mass error (80µg 1σ)

Detection capability:~1Å for a dense material such as Ta (TaN)<5Å for silicon, silicon oxide / nitride.

Advanced structures, with increased complexity, actually improves the sensitivity of mass metrology.

6Metryx Copyright ©

Mass Metrology is a non-invasive method for In-line monitoring‘On-Product-Wafers’ Measurement, Backside Contact only

Mass change is a direct representation of process performance Mass excursions outside the normal distribution represent process problems

Pre-MeasurementM1

M2Post-Measurement

ΔM = | M1-M2 |

Simple Measurement

STI Oxide Fill deposition

110000

115000

120000

125000

130000

135000

140000

0 50 100 150 200 250 300

NumberM

ass

chan

ge (µ

g)

+ 5%

- 5%

7Metryx Copyright ©

TSV - Process Steps

TSV Etch

TSV Cleaning

Oxide Liner

Barrier/Seed

Cu Plate

CMP

8Metryx Copyright ©

Through Si Via (TSV) Study

Two Via types were studied. -TSV A: near straight walled

via with a AR of 5:1-TSV B: highly tapered via

with a AR of ~35:1

Mass change characteristics of each via type investigated

Exposed Area %

Asp

ect R

atio

0

0.5 1

1.5 2

2.5 3

0

10

20

30

40

Mass (mg)

0.0-30.0

30.0-60.0

60.0-90.0

90.0-120.0

120.0-150.0

150.0-180.0

180.0-210.0

210.0-240.0

240.0-270.0

TSV B

TSV A

RA

DA

TSV B

TSV A

Via Diameter ~ 5 um

9Metryx Copyright ©

TSV By = -0.0024x2 + 1.1901x + 4.7453

R2 = 0.995

TSV Ay = -0.0022x2 + 1.2293x + 0.0903

R2 = 0.9992

0

50

100

150

0 50 100 150Etch Time % (Cycles)

Dep

th %

of P

OR

Depth TSV ADepth TSV B

Process requirements of each via type are unique.

Via etch process change and response is compared by normalizing to the related Process of Record (PoR).

Rate of change in etch rate is very similar, albeit slightly more pronounced on TSV B.

TSV B

TSV A

RA

DA

Etch Depth v Cycle time

10Metryx Copyright ©

TSV By = -0.0038x2 + 1.3509x + 3.0548

R2 = 0.9893

TSV Ay = 0.9672x + 2.2678

R2 = 0.9999

0

50

100

150

0 50 100 150Etch Time % (cycles)

Mas

s Lo

ss %

of P

oR

Mass TSV AMass TSV B

Mass vs etch cycle time, more clearly shows the difference in behavior between TSV A & B.

TSV A exhibits a linear loss in mass vs cycle time suggesting the transport rate of species and by products in and out of the Via remains constant.

However, in the case of TSV B, the mass loss is rapidly slowing as the feature becomes deeper.

TSV B

TSV A

RA

DA

Mass Loss v Cycle time

11Metryx Copyright ©

Production Monitoring of Si Etch

Chambers 1 & 2 are well matched

Reducing Si mass loss as the wafers are processed is due to polymer loading in the chamber

If a sudden shift of similar magnitude occurs in both chambers, it is known that this is related to incoming material

PhotolithographyHard-mask issue

-69000

-67000

-65000

-63000

-61000

-59000

1 26 51 76 101 126 151 176Measurement Number

Mas

s Lo

ss (u

g)

Chm 1 Chm 2 USL Target LSL

12Metryx Copyright ©

Mas

s Lo

ss (u

g)Etch Chm

Stab

le

Uns

tabl

e

-2800

-2300

-1800

-1300

-800

-300Mass measurement of the polymer removed in the wet strip process, provides a clear quantitative measurement

Chamber B is exhibits a high level of variability in mass loss during the wet clean

Variability is related to the degree of polymer formed during the etch

Chamber A Chamber B

Data randomized across wet clean stations

Polymer Removal – Etch Chamber Sort

13Metryx Copyright ©

Polymer Removal – Wet Station Sort

When post etch wafers are randomized between Wet Stations #1 & #2, the same mass should be removed

Mass removed in Station #1 is greater than Station #2, indicating a more aggressive chemistry

Mass offers a simple and direct method to monitor via wet cleaning and wet chemistry stability M

ass

Rem

oved

(ug)

System

Sys

tem

1

Sys

tem

2

-1400

-1200

-1000

-800

-600More Aggressive Less Aggressive

14Metryx Copyright ©

0

5000

10000

15000

20000

25000

30000

1 2 3 4 5Aspect Ratio (AR)

Mas

s (u

g)

SidewallTop Surface

1:1 5:1 10:1 20:1 50:1

Mass of Oxide Liner over topography increases rapidly with exposed area and Aspect Ratio, as compared with the same nominal surface film thickness

40 nm TEOS Liner / 3 % Exposed Area

ITRS roadmap calls for AR values as high as 20:1 in the future.

Oxide Liner – (HAR, Surface Area & Mass)

15Metryx Copyright ©

Liner / Barrier deposition, including sidewall coverage monitored non-destructively

Similar to the Oxide Liner, sidewall contribution of mass added increases on higher Aspect Ratio features

Line

r Mas

s (u

g)

Lot No.1 2 3 4 5 6 7 8

8500

9500

10500

11500

LinerBarrier

Liner / Barrier Sidewall Coverage

16Metryx Copyright ©

Monitor Mass increase v. Top surface thickness

Selective bottom-up, void free fill, results in the mass increasing rapidly as compared to top surface thickness

If Voids are present then the mass will increase more slowly compared to increase in top surface thickness increase

Developing Effective Cu-Fill with Mass

Surface thickness ‘x’

Tota

l Mas

s (m

g)

Voiding

Fill Marginal

Fill Superior

Process C

Process A

Process B

Thickness ‘xa’

Thickness ‘xb’

Thickness ‘ya’

Thickness ‘yb’

17Metryx Copyright ©

Mass removed during CMP reflects the mass of the Barrier & Cu Fill

This correlation of mass removed to mass deposited affirms the CMP process stability

-20.00

-15.00

-10.00

-5.00

0.00

5.00

10.00

15.00

20.00

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

Data No

Mas

s D

evia

tion

(mg)

DEPCMP

Data is plotted in mg deviation from target

Metal Fill & CMP

18Metryx Copyright ©

CMP Stability

Mass removed during CMP is well correlated to the mass added by Barrier + ECP

Therefore the CMP is observed to be well controlled. Where there is an undesirable CMP under- or over-polish, scatter will be seen in the data

This ‘compensation’ results in a mass stability of the wafers relative to post contact etch which is excellent

y = -1.0001x + 2E-13R 2 = 0.9997

-17.00

-12.00

-7.00

-2.00

3.00

8.00

-8.00 -3.00 2.00 7.00 12.00 17.00

DEP Mass (Barrier + ECP) Added

CM

P M

ass

(Bar

rier +

EC

P)

Rem

oved

Data is plotted in mg deviation from Target

LinerBarrier

Cu

LinerBarrier

19Metryx Copyright ©

-15

-10

-5

0

5

10

15

TSV

Etch

Liner Barrier

Cu ECP CMP Final

Stability

Mass variation is shown relative to the mean for the group

Box-Whisker plots indicate the Mass stability relative to the previous step

The CMP process accommodates variance in the Fill process by adjusting the CMP polish

Final Mass Stability shown is relative to the post contact etch

Mas

s (m

g)

TSVLinerBarrier

Cu

LinerBarrier

Mass Stability

20Metryx Copyright ©

Mass added Post TSV CMP

Mass measured post Etch and post CMP reflects the mass filling the TSV.

Process excursions related to Via Etch, Via Fill and CMP are all effectively monitored by the stability in this mass delta.

156000

157000

158000

159000

160000

161000

0 20 40 60 80 100Measurement Number

Mas

s A

dded

(ug)

TSVLinerBarrier

21Metryx Copyright ©

Through-Silicon Via (TSV) Etch in Silicon Measure etched silicon volumeConfirms etch depth and profile meet process definition

Oxide LinerConfirm step coverage of liner

Barrier/SeedAccurate measurement of multi-stack layerDetermine effective sidewall coverage

Copper – ECD Fill and CMP Cleaning of Copper Oxide from Seed prior to ECPOptimization of Bottom Fill and Void preventionMonitor, affirm process stability

TSV Metrology – Summary

22Metryx Copyright ©