application of plasma parameter measurement by … amat emax chamber michael hartenberger...

11.03.2002

MichaelHartenbergerUniversity of CottbusPeter MollDRAM InnovationsAndreas SteinbachCenter forDevelopment andInnovationInfineonTechnologiesDresden

3rd AEC/APC Conference Europe, April 10th-12th 2002, Dresden

CDI

ProjectSENSOR

APCAPC

Application of plasma parameter measurementby SEERS on oxide etch process development

for new DRAM shrink generationsin AMAT eMax chamber

Michael Hartenberger [email protected] of Cottbus www.tu-cottbus.dePeter Moll [email protected] Steinbach [email protected] Technologies Dresden www.infineon.com

11.03.2002



CDI

ProjectSENSOR

APCAPC

Acknowledgementq The results presented in this paper were supported partially:

Ø by the EFRE fund of the European CommunityØ and by funding of the State Saxony

of the Federal Republic of Germany, project SENSOR.q The authors are responsible for the content of the paper.

Acknowledgement

11.03.2002



CDI

ProjectSENSOR

APCAPC

Outlineq Description of process and experimentq Process developmentØ Characterisation of plasma parameters by variation of B-Field,

pressure and RF-powerØ Verification of plasma parameter impact on product wafersØ Investigation of conditioning and warm up procedure

q Benefit of plasma parameter measurement during processdevelopment

q Summary

Outline

11.03.2002



CDI

ProjectSENSOR

APCAPC

SixNxSi

Poly Si

BSGStep 1BSG etch(endpoint)

Step 2nitride etch(by time)

Process description

Typical oxide etch SEM picture

q Oxide etch in two steps:Ø First BSG is etched using a Poly-Si mask and an optical endpoint systemØ Second Nitride is etched for a fixed time

q The result is a high aspect ratio trench

11.03.2002



CDI

ProjectSENSOR

APCAPC

Aim of experimentsq Characterisation of process window by use of plasma

parametersè Assessment of optimisation potential for existing process

q Enhancement of conditioning proceduresq Investigation of warm up effects

Experiment description

11.03.2002



CDI

ProjectSENSOR

APCAPC

Experimental set up ofcharacterisation experimentq Before experiment:

Chamber- conditioning and warm up with 3 conditioning wafersq Design of experiment::

Ø Power variation using nine wafers (unpatterned Si-Wafer) in three stepsØ Pressure variation in three steps for every power settingØ B- field variation in 6 steps for every wafer

Parameter variation

11.03.2002



CDI

ProjectSENSOR

APCAPC

Electron collision rate vs. process time for threepower settings

q At 20 mTorr electroncollision rate indicatesvery unstable process atmedium and higher B-Fieldè pressure too low

Parameter variation

Electron Collision Rate vs. Time

1,0E+07

1,1E+08

2,1E+08

3,1E+08

4,1E+08

5,1E+08

6,1E+08

0 60 120 180 240 300 360time

Elec

tron

Col

lisio

n R

ate

[s-1

]

20 mTorr 40 mTorr 60 mTorr


1,0E+07

1,1E+08

2,1E+08

3,1E+08

4,1E+08

5,1E+08

6,1E+08

0 60 120 180 240 300 360time

Elec

tron

Col

lisio

n R

ate

[s-1

]



1,0E+07

1,1E+08

2,1E+08

3,1E+08

4,1E+08

5,1E+08

6,1E+08

0 60 120 180 240 300 360time

Elec

tron

Col

lisio

n R

ate

[s-1

]


20 40 60 80 100 120

20 40 60 80 100 12020 40 60 80 100 120H [Gauss]

1400 W

H [Gauss]

1800 W

H [Gauss]

2000 W

11.03.2002



CDI

ProjectSENSOR

APCAPC

Electron collision rate vs.B-Field and Pressure (1400W)

Pressure [mTorr]H [Gauss]

Electron collision rate [s-1]

20

40

60

040

80120

1e8

2e8

3e8

4e8

5e8



Electron collision rate [s -1]

20

40

60

040

80120

1e8

2e8

3e8

4e8

5e8

Median electron collision rate vs. B-field andpressure for RF power 1400, 1800 and 2000 W



Electron collision rate [s-1]

20

40

60

040

80120

1e8

2e8

3e8

4e8

5e8

q 20mTorr processes were very unstable and should not beconsidered

q Baseline recipe has a comparatively low electron collision rate èpotential for optimisation: going up to 80 Gauss at 40 mTorr couldresult in an increasing number of reactive species in the plasma

•

Baseline recipe

Parameter variation

11.03.2002



CDI

ProjectSENSOR

APCAPC

Electron Density vs.B-Field and Pressure (2000W)


Electron Density [cm-3]

20

40

60

040

80120

02e94e96e98e9

1e101,2e101,4e10

Median Electron Density vs. B-Field and Pressurefor RF Power 1400, 1800 and 2000 W

q There is a small dependence of electron density on pressure anda big one on magnetic field strength

q Higher electron density should lead to higher ion density andtherefore could cause an increasing ion bombardment on thewafer surface




20

40

60

040

80120

02e94e96e98e9

1e101,2e101,4e10

•

Baseline recipe




20

40

60

040

80120

0

2e9

4e9

6e9

8e9

1e10

Parameter variation

11.03.2002



CDI

ProjectSENSOR

APCAPC

Parameter variation

Electron collision rate [median] vs. pressure

2E+08

3E+08

4E+08

5E+08

10 20 30 40 50 60 70pressure [mTorr]

med

ian

elec

-tr

on c

ollis

ion

rate

[s-1

]

Experiment 1 Experiment 2

Electron density [median] vs. pressure

2E+09

3E+09

4E+09

5E+09

6E+09

7E+09

10 20 30 40 50 60 70

pressure [mTorr]

med

ian

elec

tron

de

nsity

[cm

-3]

Experiment 1 Experiment 2

Repeatability test with unpatterned Si-wafers:Charts of electron density and collision rate

q There is a goodreproducibility ofexperimental results

q Discussion on nextpage

11.03.2002



CDI

ProjectSENSOR

APCAPC

Repeatability test with unpatterned Si-wafers:explanationq Results of electron collision rate and electron density are

reproducible for 40 and 60 mTorrq Median electron density and electron collision rate for 20 mTorr

process different between experiments because of unstableprocess

è Plasma parameters allow a reliable prediction of chamberconditions

Parameter variation

11.03.2002



CDI

ProjectSENSOR

APCAPC

Experiments on product wafersq Baseline process was found to be at

an unfavourable point: There was astrong rise in electron collision ratewhen changing B-field to slightlyhigher valuesè process susceptible tofluctuations in process parameters ?

q Impact of 3 B- field settings ongeometric properties of patternedproduct wafers was investigated

q Magnetic Field was set to 40, 60 and 80 Gauss in BSG step at1800W and 40 mTorr

q There was no magnetic field in the nitride stepImpact on product wafers



Electron collision rate [s -1]

20

40

60

040

80120

1e8

2e8

3e8

4e8

5e8

11.03.2002



CDI

ProjectSENSOR

APCAPC

Impact on product wafers

Tests on patterned product wafers: Electrondensity and El. collision rate for B-Field Variation

q 80 Gauss processproduces differentresults in electroncollision rate

q Electron densityincreases with risingmagnetic field, asobserved in previousexperiments on testwafers

Electron Density vs. time

1,0E+08

1,1E+09

2,1E+09

3,1E+09

4,1E+09

5,1E+09

6,1E+09

7,1E+09

0 50 100 150 200 250

process time [s]

elec

tron

dens

ity[c

m-3

]

40 Gauss 60 Gauss 80 Gauss

Electron Collision Rate vs. time

1,0E+06

4,1E+07

8,1E+07

1,2E+08

1,6E+08

2,0E+08

0 50 100 150 200 250process time [s]

elec

tron

colli

sion

ra

te [s

-1]


80 G

60 G 40 G

BSG step nitride step80 G

60 G 40 G

11.03.2002



CDI

ProjectSENSOR

APCAPC

Tests on patterned product wafers: Explanationq Electron collision rate and electron density measured during B-

field variation on product wafers behave similar tomeasurements on unpatterned Si-wafers

q Again high electron collision rate and electron density in BSGstep for an 80 Gauss setting is observed

q In nitride step, where no magnetic field is present, all values areat the same level


11.03.2002



CDI

ProjectSENSOR

APCAPC

SEM cross-sections of patterned wafers - etchedwith different B- field settings

q Differences between left andright trench width are a normalphenomenon, caused bylithography tolerances

q Poly hard mask height (redlines) is CD of interest

40Gauss

Poly

BSG

SiNSi

60Gauss

80Gauss


11.03.2002



CDI

ProjectSENSOR

APCAPC

remaining Poly-Hardmask-height vs. Median electron collision rate

180

185

190

195

200

205

1,00E+07 4,00E+07 7,00E+07 1,00E+08 1,30E+08median electron collision rate (BSG step) [s-1]

poly

hei

gth

remaining Poly-Hardmask-height vs. Magnetic Field Strength

180

185

190

195

200

205

40 60 80H [Gauss]

poly

hei

gth

Discussion of REM cross sections

q Etch rates on test wafers were nearly equal for 60 and 80 Gauss,but selectivity of BSG to other layers was about 10% higher at 60Gauss è Poly hard mask height is smaller for 80 Gauss

q This correlates to plasma parameters: Electron collision ratecorrelates to poly mask height

q This is not critical, because mask is thick enough to resist etchingand all other dimensions are nearly equal for different B- fields


11.03.2002



CDI

ProjectSENSOR

APCAPC

Investigation of Conditioning and Warm up- Part 1

q Experiment in upperchart started with warmplasma chamber

q experiment in lowerchart started with coldchamber

q there are cleardifferences in(temperature sensitive)nitride step

q it needs two additionalwafers until conditionsare stable again

Conditioning and warm up


1.0E+06

5.1E+07

1.0E+08

1.5E+08

2.0E+08

2.5E+08

3.0E+08

0 50 100 150 200 250process time [s]

elec

tron

col

lisio

n ra

te [s

-1]


Electron Collision Rate vs. time

1,0E+06

4,1E+07

8,1E+07

1,2E+08

1,6E+08

2,0E+08

0 50 100 150 200 250process time [s]

elec

tron

colli

sion

ra

te [s

-1]


Warm chamber

Cold chamber

11.03.2002



CDI

ProjectSENSOR

APCAPC

Investigation of Conditioning and Warm up- Part 2q Warm up is insufficient: 3 conditioning wafers are not enough to heat up

chamber - at least 4 are necessary (nitride etch is very temperature sensitive !)


1.0E+065.1E+071.0E+081.5E+082.0E+082.5E+083.0E+083.5E+08

0 100 200 300 400 500 600 700 800process time

elec

tron

colli

sion

rate

[s

-1]

without pre-process with pre-process

Conditioning procedure

q Conditioning procedure with 3 conditioning wafers is sufficient - no first wafereffect observed for conditioning wafers and first (40 Gauss) product wafer

q First experiment (upper graph, previous page): other processes beforeexperiment

q No impact of pre- process on electron collision rate for following experimentalwafers

wafer1 wafer2 wafer3

Conditioning and warm up

11.03.2002



CDI

ProjectSENSOR

APCAPC

Benefit of plasma parameter measurementq Plasma parameters were used to characterise oxide etch process by

parameter variation and showed potential for process developmentand optimisation.

q They indicated temperature and conditioning effects è they are apowerful tool to verify conditioning and warm up procedures withminimal effort in real time.

q It was shown that plasma parameters give reproducible results andcan be used for stability control and failure detection.

q Additional measurement techniques are necessary for investigationof chemical reactions, e.g., optical emission spectroscopy.

q Correlation to selectivity was shown è plasma parameters can beused to optimise processes

Benefit

11.03.2002



CDI

ProjectSENSOR

APCAPC

Summaryq An oxide etch process has been characterised using plasma

parameters electron collision rate and electron density byvariation of pressure, power and B- field.

q Unstable processes have been found for a pressure of 20mTorr- this pressure should be avoided for wafer processing.

q Experiments on product wafers showed a correlation of plasmaparameters to remaining Poly-Si thickness and selectivity.

q 60 Gauss process is recommended, because its selectivity ishighest.

q Conditioning procedure with two conditioning wafers is sufficientto eliminate pre-process influence on products.

q Warm up procedure with two conditioning wafers is insufficient -strong temperature influence has been found.

Summary

application of plasma parameter measurement by … amat emax chamber michael hartenberger...

Documents