TEA Systems Corp. Confidential LithoWorks PEB PEB and CD analysis of two plates August 14, 2003 Thermal analysis of two PEB plates: 093-08_peb_bake7.csv.

TEA Systems Corp. Confidential

LithoWorks PEBLithoWorks PEBLithoWorks PEBLithoWorks PEB

PEB and CD analysis of two plates

August 14, 2003

Thermal analysis of two PEB plates:

093-08_peb_bake7.csv

093-08_peb_bake8.csv

August 2003 LithoWorks PEB - Two plate study Page -2-TEA Systems Corp. Confidential

Summary• These are preliminary results

• The first half of this analysis addresses PEB thermal analysis for PEB’s 7 & 8 Investigates Temperature cycle Energy installed into the wafer (Temp * Time) Variation in energy installed Temperature over time contour plots

• The second half of the analysis examines two Nanometrics OCD data sets Fixed focus analysis of wafers 12,16

• Wafer 12 = PEB 8

• Wafer 16 = PEB 7

The data has been modeled for across wafer variation Residuals to the wafer model illustrate the anticipated, non-PEB influenced performance

• Statement based upon expected whole-wafer variation of PEB temperature

• The last slide touches on Wafer 13 & 17 These are FEM arrays and warrant further study.


Initial Observations• PEB Thermal Analysis

The Bake cycle was examined as a function of:• Time-Temperature curve

• Time-temp variations across wafer.

• Total energy (integrated time * temperature) delivered to the wafer

• Energy variation across wafer due to non-uniform heating.

PEB 7 is more uniform than PEB 8• Not observed in the temperature or the total energy delivered

• Can be seen in the plots of temperature variation from wafer mean over time and in the Energy Variation plots.

Neither bake plate is allowing enough time for the bake temperature to achieve steady state.• See slides 26 to 28

• Feature Response Note: The On-Wafer sensor orientation has not been corrected or adjusted for this analysis.


Feature Response and Correlation• Feature Response

Note: The On-Wafer sensor orientation has not been corrected or adjusted for this analysis. Examined wafers 12 (PEB 8) and wafer 16 (PEB 7) Each feature’s variation was modeled for whole-wafer variations

• Wafer-model should correlate to PEB energy variations

• Residuals to wafer model should be approximate equal in response – if both wafers were exposed on same scanner

• Residuals to wafer model should show exposure tool performance.

MSE Behavior• Differs significantly for Vertical (group 1) and Horizontal (group 2) features.

– May be the result of scanner slit vs scan sensitivity of feature edges.

– May also be from metrology tool algorithm.

There is no strong and obvious visual correlation of PEB energy and feature response. There is a suggestion of PEB energy variation and SWA

• Further studies More thought must be given to these correlations. FEM wafers need to be studied.


PEB #7


PEB #7 overview

• Left graphic Details location of sensor points

• Right graphic Summary of rise/fall times vs temperature


Sensor variation from wafer mean

• Each site contains a plot of delta-temp versus time

• Area product of (degrees * seconds), summed over

all sites. This is proportional to the energy entered into the wafer during the bake.

• Area Variance Energy variation across the wafer Summed deg-sec’s over all sites reporting

total variation from mean-temperature over time.


Corresponding points on the graph

• Moving the mouse to a point on the temp-time curve will highlight the corresponding points on the delta-temp wafer plot

(50.1 sec, 113 C)


On-Wafer thermal loading?

• Notice how the 7 wafer-center sites are the first to rise in temperature.• Note also the two sites located over the On-Wafer sensor packed lag in heading and also lag in

cooling. This area also does not reach uniformity until the end of the heading cycle. This is most probably due to the thermal-mass loading of the sensor packet.

• The red-dot on each plot marks the same relative point on each site.

Area of sensor packet


PEB 7; Thermal Energy Delivered

• Plot of thermal area (deg-sec) across wafer.

• Notice that there is a range of 168,714 deg-sec with the upper left of the wafer receiving the most energy

• Thermal-package sensor area is relatively cool

• No activation threshold was set for this plot

thermal sensor area


PEB 7; Thermal Energy Delivered

• Plot of area (deg-sec) integrated over time

• Notice that there is a range of 168,174 deg-sec with the upper left of the wafer receiving the most energy


PEB 7: Thermal Energy Variance across wafer

• Variation in energy (integrated temperature*time) delivered

• Variation is greatest at wafer’s upper left corner, least in center & right.


Thermal Energy Variance Across Wafer - Contour


PEB 8 Data


Plate PEB 8, Bake Curve• Top: PEB 8

• Bottom: PEB 7

• Curves appear very similar in shape.


PEB 8; Delta-Temperature Response

(50.1 sec , 113 C)

• Same point is marked on the time-temp curve. Distribution in upper left of wafer

is now very uniform.


Comparison PEB 8 vs PEB 7• Time-Temp Area

Differences

• PEB 8 has a higher energy

area. Exhibits greater

excursion range and variation across the plate.

PEB 8 PEB 7


Comparison: Delta temperature with time

• PEB 8 exhibits more variation at the bottom of the wafer.

PEB 8 PEB 7


PEB 8; Thermal energy delivered


Thermal Energy Delivered – Bake 8 Contour


PEB 8 Thermal Energy


PEB 8; Thermal Energy Variation


Comparison: Thermal Energy Delivered – PEB 8/ PEB 7

• Both wafers as plotted on the same scale.• More thermal energy delivered to PEB 7

PEB 8 PEB 7


Thermal Energy Variation – Bake 8 Contour

• Energy delivered as a delta from the wafer average temperature.• Energy = Sum((Temperature-Activation Temp.) * DeltaTime)

DeltaTime = time interval of sample


PEB 8: Greatest variation in thermal energy


Comparison: Thermal Energy Variation – PEB 8/ PEB 7

• Scales vary but characteristic is same• PEB 7 has has significantly better energy uniformity

PEB 8 PEB 7


Comparison energy variation statistics

• PEB 8 exhibits A greater mean temperature, range and variance of temperature than PEB 7

PEB 8 PEB 7


PEB 8/ PEB 7; thermal variation

• PEB 8 and 7 variations plotted to the same scale.

PEB 8 PEB 7


Thermal range study Bake8

• Temperature range about the average temperature of the wafer at each time-slice.

• The cycled-temperature rose a total of 98.07 degrees.

• Steady state uniformity is 0.18 degree across wafer.

• Notice that the wafer at 121 degrees (max. temp at 134 sec.) did not reach steady-state uniformity.


Thermal range study Bake7

• Same study with PEB 7

• This study shows a red square at each transition start/stop point on the curve.


Range Uniformity Summary

• Plots exhibit the thermal range across the wafer at each time-slice• Both plates exhibit the same steady-state thermal resolution.

Thermal range and maximum temperatures differs by approximately one degree

• Neither plate reaches steady state while at the maximum temperature.• Bake 8 has

greater across-wafer variations (9.6 v 7.0 degrees) Greater range at the final temperature (0.9 v 0.6 degree) Smoother cooling curves than 7 Near identical rise/cool thermal slopes and time at the maximum temperature compared with Bake 7

• Conclusion: Bake 8 has more thermal variation across the wafer.

Bake 8 Bake 7

Bake 8


Thermal Uniformity at any point in time• Generate single or matrix

graphics of all of the plots.

• A “movie” video will also be added to allow you to watch wafer heating.


PEB 8 – Steady State Temperature


PEB 8 Steady State Temp


Thermal changes during the up-ramp


CD data review


Data Setup and Organization

Wafer 12: processed on PEB plate 8, const. exp/focus 18.5/0

Wafer 13: processed on PEB plate 8, FEM 18.5/0 +- 0.5/0.05

Wafer 16: processed on PEB plate 7, const. exp/focus 18.5/0

Wafer 17: processed on PEB plate 7, FEM 18.5/0 +- 0.5/0.05

• The measurements are taken on a nominal 150:150 grating, somewhat overexposed here (~140nm average linewidth).

• All wafers have PAB step on plate 1, also attached. • For the PEB data, there are several measurements for each plate, each corresponding to a different

orientation of the sensor wafer. '3_oclock', for example, means that the wafer module is at the 3o'clock position relative to the cassette. 1-2 o'clock is the position between 1 and 2, which appears to be 3/4 of a full rotation from the orientation of the

process wafer when it's on the PEB plate-since OnWafer formats the data such that the module is situated at 3 o'clock, this means we just have to rotate the 1-2 data 180 degrees and should have the right orientation.

• Within the Nanometrics data set, there are two "groups" of data for each wafer group 1 corresponds to measurements made on gratings with lines of vertical orientation group 2 corresponds to the horizontal orientation.


Wafers 13,15; Assumed Dose and Focus layout

• Wafers 12 & 16 are fixed focus/dose


MSE behavior (12 & 16)

• MSA can be used to determine the measurement quality

• Variation is in fact a little better for the FEM wafers than the fixed exposure. Probably due to the bi-modal

distribution of the fixed-exposure data.

Layout, all sites


MSE; Wafer 12 (PEB 8)• Horizontal and vertical features are

responding differently to MSE


Wafer 12, TCD (fixed focus, Group #1)

• Vertical FeaturesAll features


Wafer 12, TCD group 2 (Horizontal)

• Field layout is shown on the left.


Wafer #12 –T3 (PR)

• Photoresist thickness behaves quite differently for vertical vs horizontal feature orientation.

Group 1Group 2


Wafer 12, BCD

• Have about 3 nm of difference in mean values.

• Range of values is the same to within 0.7 nm


Wafer #12, BCD Residuals to Wafer variation

• Aberrations of scan or slit are influencing the CD variation

• Left = vertical features so edges are influenced by the scan

• Right = horizontal features so there is more slit influence.


Wafer 12; SWA


Wafer #12; SWA Residuals to wafer model

• The horizontal line behavior (right) is very surprising. Metrology Algorithm?


Wafer #12 model; BARC T2

• Good, repeatable to about 0.1 nm


Wafer #12; BARC residuals

• Vertical features (#1) have characteristic ring. Ring is not as evident in the horizontal features on the left. Scale on left is wider because of the difference in noise.


Wafer #16 (PEB 7)


Wafer 16 BARC and PR


Wafer #16; Modeled wafer SWA (Group 1)


Wafer #16; Modeled wafer BCD & TCD (Group 1)


Wafer 16; Residuals to wafer BCD & TCD (Group 1)


Wafer #13 FEM; in-die groups 1 & 2• Different behavior and sampling

of groups 1 & 2

Vertical

Horizontal

TEA Systems Corp. Confidential LithoWorks PEB PEB and CD analysis of two plates August 14, 2003 Thermal analysis of two PEB plates: 093-08_peb_bake7.csv.

Documents

peb plates

wafer energy variation

wafer model

wafer variations wafermodel

wafer temp

wafer sensor orientation

tea systems

plots of temperature