Process Quality Control at CMS

Download Report

Transcript Process Quality Control at CMS

Quality Assurance of Silicon Strip
Detectors and Monitoring of
Manufacturing Process
Thomas Bergauer
Institute f. High Energy Physics
HEPHY, Vienna
SiLC meeting @ ILC Workshop
Vienna, Nov 18th, 2005
Outline of Talk
1.
2.
Characterization of Silicon Strip Detectors for
Quality Assurance
Characterization of “standardized” teststructures to monitor manufacturing process
6” wafer:
Characterization of Strip Detector
 global measurements (IV, CV)
 strip-by-strip tests
(Ileak, Cac, Rpoly and Idiel)
Characterization of test structures
with 9 different measurements
Thomas Bergauer, HEPHY Vienna
2
1. Quality Assurance of Silicon
Strip Detectors
Sensor Characterization Basics


Silicon Sensors for future
high-energy experiments will
have many strips to achieve
a high spatial resolution.

Global parameters:

Large Tracker will use
enormous area of silicon
sensors


What do we test?

Efficient Quality Assurance
mandatory
Automated test system is
necessary to determine the
electrical parameters of each
strip.

IV-Curve:

Dark current

Breakthrough
CV-Curve:

Depletion voltage

Total Capacitance
Strip Parameters

strip leakage current Istrip

poly-silicon resistor Rpoly

coupling capacitance Cac

dielectric current Idiel
Thomas Bergauer, HEPHY Vienna
4
AC-coupled Silicon Strip Detector

What do we test?



Si Strip Sensors for the CMS
Tracker
n bulk
p+ implanted strips



connected to bias ring via
polysilicon resistors
AC-coupled Aluminium
readout strips
Dielectric Oxide SiO2 + Si3N4
Thomas Bergauer, HEPHY Vienna
5
AC-coupled Silicon Strip Detector
Strip Pitch
80-170μm
Corner of a typical CMS
Silicon Strip Detector
Thomas Bergauer, HEPHY Vienna
6
Sensor Test Setup


Light-tight Box, Instruments,
Computer
vacuum support carrying the
sensor


Needles to contact sensor bias
line


Mounted on freely movable
table in X, Y and Z
fixed relative to sensor
Needles to contact:



DC pad (p+ implant)
AC pad (Metal layer)
Can contact ever single strip
while table with sensor is
moving
Thomas Bergauer, HEPHY Vienna
7
Sensor Test Schematics

Instruments (HV source, Amp-Meter, LCR-Meter,…) on
the left are connected via a cross-point switching
matrix to the needles which contact the sensor to
perform different measurements
Thomas Bergauer, HEPHY Vienna
8
Example Measurements: CV, IV

Combined voltage
ramp up to 500800V



Dark current (blue) and
total capacitance (red,
plotted 1/C2) is
recorded.
Depletion voltage is
extracted
Thomas Bergauer, HEPHY Vienna
9
Example Measurements: Stripscan



After IV-CV ramp, bias voltage
is adjusted to stable value (e.g.
400 V) and stripscan is started
4 parameters tested for each
strip:

dielectric current Idiel

coupling capacitance Cac

poly-silicon resistor Rpoly

strip leakage current Istrip
For each test, the switching
matrix has to be reconfigured

Full characterization of detector
with 512 strips: 3h
Thomas Bergauer, HEPHY Vienna
10
Example Results: Depletion Voltage, Dark
current (Sensors for CMS)
Depletion Voltage
Dark current @ 450V
Thomas Bergauer, HEPHY Vienna
11
Results: Stripscan

Total number of bad strips




Total = sum of Istrip , Rpoly,
Cac, Idiel )
Bad = outside specified
cuts
CMS requires less than 1%
of strips are outside cuts for
at least one of the strip
parameters
Average bad strips per
sensor:

0,37
Thomas Bergauer, HEPHY Vienna
12
2. Monitoring of Manufacturing
Process
“PQC”…. Process Quality Control
Motivation and Assumptions

Full Characterization of Strip Detector has some
disadvantages




takes a lot of time (if every strip is checked)
Only sample tests possible (assumption that production batch
behave similar)
Some interesting parameters are not accessible on standard
detector or would require destructive tests
Remedy: Doing similar measurements on standardized teststructures




Assumption: Test structures behave identical to main sensor,
since produced on the same wafer
Measure many parameter, each on a dedicated test structure
Destructive Tests possible
Fast measurement possible
Thomas Bergauer, HEPHY Vienna
14
Company test-structures
Standardized Set of Test Structures
“Standard Half moon”


TS-CAP
baby
GCD
sheet
9 different structures
individually described in the next slides
CAP-TS-AC
CAP-TS-AC
Thomas Bergauer, HEPHY Vienna
diode
MOS out
MOS in
15
Test Structures

TS-CAP:



Coupling capacitance CAC to
determine oxide thickness
IV-Curve: breakthrough voltage of
oxide



Aluminium resistivity
p+-impant resistivity
Polysilicon resistivity




Gate Controlled Diode
IV-Curve to determine surface
current Isurface
Characterize Si-SiO2 interface
Inter-strip Resistance Rint
Diode:



IV-Curve for dark current
Breakthrough
CAP-TS-DC:


Inter-strip capacitance Cint
Baby-Sensor:

GCD:

CAP-TS-AC:

Sheet:



CV-Curve to determine depletion
voltage Vdepletion
Calculate resistivity of silicon bulk
MOS:

CV-Curve to extract flatband voltage
Vflatband to characterize fixed oxide
charges
(details on next slide)
Thomas Bergauer, HEPHY Vienna
16
MOS


Metal Oxide Semiconductor
Oxide composition represents configuration of




Thick dielectric in inter-strip region
Thin dielectric underneath strips (right)
Extraction of flatband voltage Vfb

Seen by sharp decrease of Capacitance
(between accumulation and inversion)

to determine fixed positive charges in Oxide
Limit defined experimental after test beam
Thomas Bergauer, HEPHY Vienna
17
Setup Description

Probe-card with 40 needles
contacts all pads of test
structures in parallel




Instruments





Source Measurement Unit (SMU)
Voltage Source
LCR-Meter (Capacitance)
Heart of the system: Crosspoint
switching box


Half moon fixed by vacuum
Micropositioner used for
Alignment
In light-tight box with humidity
and temperature control
Used to switch instruments to
different needles
PC with Labview used to control
instruments and switching
system
GPIB Bus for communication
Thomas Bergauer, HEPHY Vienna
18
PQC Setup
Thomas Bergauer, HEPHY Vienna
19
After run:
Software
Before run:
Yellow Fields:
Limits and cuts for
qualification
Blue Fields:
Obtained results
extracted from graph
by linear fits
(red/green lines)


Self-developed LabVIEW program
Fully automatic measurement procedure (~30 minutes)


Except alignment of Half moon and placement of probecard
Automatic extraction of parameters
Thomas Bergauer, HEPHY Vienna
20
Examples of identified problems
Interstrip Resistance





Limit: Rint > 1GΩ to have
a good separation of
neighbouring strips
Value started to getting
below limit
We reported this to the
company
Due to the long production
pipeline, a significant
amount of ~1000 sensors
were affected
These sensors will not be
used for CMS Tracker!
Thomas Bergauer, HEPHY Vienna
21
Examples of identified problems (2)
Flatband Voltage



Limit of 10V determined
during irradiation
campaign
We observed values up
to 40V for early
deliveries
Some batches from later
deliveries suffer from
contamination of
production line
Thomas Bergauer, HEPHY Vienna
22
Examples of identified problems (3)
Aluminium resistivity




Aluminium resistivity too
high for some delivered
batches
Limit: <30mΩ/sq.
Affects noise behaviour
of readout chip
After discovery of this
issue we requested to
increase thickness of Al
layer
=> Problem disappeared
Thomas Bergauer, HEPHY Vienna
23
Summary

Future experiments with a large tracker will require a huge
number of silicon strip sensors.

Compare with CMS Silicon Strip Tracker: 206 m2 is equal to




Its fabrication will last many months (years) and a stable
production during the whole production time is mandatory.
Strip-by-strip test of detectors is necessary but not sufficient


24.244 pieces of sensors and
9.316.352 channels
Slow, reduced set of parameters to test
Measurements on dedicated test-structures is a powerful
possibility to monitor the fabrication process




During a long production time
Also on parameters which are not accessible on the main sensor
(e.g. MOS, GCD,... )
Destructive tests possible
Fast measurement allows high throughput
Thomas Bergauer, HEPHY Vienna
24
Outlook and Future Plans

We have to optimize our test structures


We learned during the CMS QA that some things
can be improved:
 Smaller structures
 Better design of some structures
(e.g. diode, sheet)
We want to offer this standardized set of test
structures to all interested groups in the
future

To put it on unused space of their wafer design
Thanks.
Thomas Bergauer, HEPHY Vienna
25
The End.
Backup Slides
TS CAP

Array of 26 AC-coupled strips

Test of Coupling Capacitance


Oxide Thickness can calculated
Test of dielectric breakdown

Destructive !
Thomas Bergauer, HEPHY Vienna
28
Sheet

Combination of





Three polysilicon resistors
Three Aluminium Strips (10, 20, 50 um
thickness)
Three p+ Strips (10, 20, 50 um
thickness)
Used to determine resistivity of
implant, Aluminium and polysilicon
These Parameters have influence on
noise behavior of readout chip
Thomas Bergauer, HEPHY Vienna
29
GCD

Gate controlled diodes



Used to extract surface
current





Two circles ones (not used)
Two squares ones with combshaped p+-Diodes and combshaped MOS structures
alternately arranged
by applying a constant reverse
bias voltage through the diode
while varying the gate voltage
of the MOS structure.
Sharp decrease of dark
current in the inversion region
gives the surface current
Important Parameter to
monitor oxide and Si-SiO2
interface quality
Limit determined experimental
by irradiation
Thomas Bergauer, HEPHY Vienna
30
CAP-TS-AC

Measurement of inter-strip
capacitance




Between single central strip and
two neighboring ones
Outer strips on top and bottom
are shorted and connected to
ground (directly on the
structure)
While biasing of structure is
mandatory
Parameter related to noise and
SNR of readout chip
Thomas Bergauer, HEPHY Vienna
31
Baby Sensor



Structure with 192 AC-coupled strips
Identical to main detector
Used to measure IV-curve up to 700 V

Breakthrough voltage is determined
Thomas Bergauer, HEPHY Vienna
32
CAP-TS-DC


Used to determine inter-strip
resistance
Similar structure like CAP-TS-AC
(used for C_int) but with exceptions



no polysilicon resistor (strips do not
have a connection to bias ring)
p+ strips are directly connected to
Aluminium strips
High value of inter-strip resistance
necessary to have a good electrical
separation of strips
Thomas Bergauer, HEPHY Vienna
33
Diode


Simple square diode
Voltage scan is used to measure
Capacitance and to extract

total bulk thickness
A
Cdepl  0r
d



Bulk resistivity
d 2 nominal

2 0r eVdepl
Thomas Bergauer, HEPHY Vienna
34