SOLA2020/SOLA9006

Photovoltaic Technology
and Manufacturing
Lecture 8 – Quality Control

Dr Alison Lennon
 H
Housekeeping
k i
• No tutorials this week.
• Major assignment posted on BB
• Electronic submission.
submission
• Work in pairs – submit 1 assignment per pair.
• Must also submit VPL batch file of most efficient batch.
• D 5 pm F
Due Friday
id off W
Weekk 12
12.
 Thi Lecture
This L t
• Overview: In-line QC
• Lifetime measurement techniques
• Process
P QC
• Incoming wafer
• Wet chemistry
• Post diffusion
• Post metallisation
• Cell characterisation
• Summary of typical QC processes
 Quality Control Overview
• In-line: Integrated into the production line therefore
must:
• Process at the production rate (e.g.,
(e g 2400 wafers/hr);
• Be compatible with automated wafer / cell handling; and
• 24/7 operation.
• 100% in-line
i li not always
l possible
ibl
• Remove wafers and measure off-line then modify
production parameters accordingly.
accordingly
• But for promise of in-line, inspection/analysis time
must be fast.
 Lifetime Measurements
• Lif
Lifetime
ti measurements t can be
b done
d att many points
i t in
i
the manufacturing process.
g (L
• Lifetime/diffusion length ( τ)) measurement
( = √(D
techniques:
– PL/EL
– Photoconductance decay (PCD)
– Microwave PCD
– LBIC
• Needs to be fast for in
in-line
line QC!
 PCD Measurements – Lecture 4
• Two ways of measuring carrier lifetimes (τ):
– Maintain SS generation rate and use G = Δn/τ to estimate τ.
– Abruptly
p y terminate generation
g and measure rate as which
carriers decay: dΔn/dt = -Δn/τ.
= Photoconductance decay (PCD)
• In both cases need to measure Δn.
Place wafer over an inductor which is part of
an rff circuit.
i i Induced
I d d voltage
l is
i proportional
i l
to conductivity of wafer.
σL = qW(μn+μp) Δn
W = wafer thickness
μn, μp = mobilities of electrons and holes and
p
depend on the doping
p g and injection
j level.
Can iterate to find both Δn and (μn+μp) .
PCD Measurements – Lecture 4 (cont)
 PCD Lifetime Measurements
• C
Can measure lifetimes
lif ti off passivated
i t d andd non-
passivated wafers.
– Lifetimes of non-passivated (“out of the box”) wafers are
very low
l since
i photogenerated
h t t d carriers
i quickly
i kl diffuse
diff to
t
the surface and recombine.
– Can use infra-red excitation for non-passivated wafers →
carriers are generated further from the surface and so are
more sensitive to bulk lifetime.
• Can make measurements on ingots and bricks.
 Microwave PCD
• Mi
Microwave PCD (μ-PCD)
( PCD) iis usedd to
t generate
t
maps which show spatial variations in lifetime
across a surface.
– Direct a small microwave excitation signal and
scan reflected microwave power which is a
measure of conductance.
– Unlike
U lik Si Sinton
t PCD values,
l μ-PCD
PCD values
l are curve
fitting parameters (i.e., arbitrary units) and hence
do not relate to actual physical quantities .
– But quick and non-contact.
non contact
 Diffusion Length Mapping (LBIC)
• LBIC = light
li ht beam
b induced
i d d current. t
• Locally irradiate the solar cell and measure ISC. At
different wavelengths
g → spectral
p variation of the
photo-current → L.
• Useful for measuring EQE/IQE (see later).
LBIC vs EL: Diffusion Length Comparison
 Incoming Wafer QC
• IIncoming
i waferf checking
h ki systems t that
th t can sortt wafers
f
according to thickness, surface characteristic and
electrical characteristics (e.g., resistivity, lifetime)
– Need wafers with uniform properties for consistent processing.

• Lifetime tests can be done using PCD or PL.

PL
• PL: Just need to capture a single image.
• PCD & μ-PCD: Point-to-point measurement
 Resistivity Tests
• T
Typically
i ll incorporated
i t d as partt off incoming
i i wafer
f tests.
t t
• Preferred methods are contact free.
– E.g.,
g , Eddyy current flow method implemented
p byy SemiLab
 Wafer Thickness Measurements
• Thickness
Thi k measurements
t can be:
b
– Contact
• Hold the bottom surface of the wafer against a reference plane and
th move a small
then ll spherical
h i l surface
f or stylus
t l into
i t contact
t t with
ith the
th
top surface.
• Not desirable because of scratching/breakage of thin wafers.
– Non-Contact
Non Contact
• Typically use capacitive sensors or optical sensors.
• Capacitive and optical sensors are located a fixed height above the
wafer.
• Two sensors can be used: one above and one below.
• Measurements can be in-line or off-line.
• Many systems can also measure:
– Total thickness variation (TTV)
– Bow
– Warp
W
• Also called metrology tools.
 Non-Contact Metrology Systems
Many different non-contact
systems are available.
 Saw Damage Etch
• P
Purpose: R
Remove surface
f regions
i th t have
that h been
b
damaged by sawing.
• Q
Quality
y Control:
– Periodically measure thickness of processed wafers.
• Bath maintenance: Some level of build-up of Na
silicate is desirable.
desirable
– Slows etching rate and prevents formation of polished surface.
– Need to prevent solidification of bath.

In-situ monitoring of
wafer thickness but
maybe not cost effective?
 Alkaline Texturing QC
• Inspection of pyramid height, % coverage (SEM,
Optical microscope)
– Check ~ 1 from each batch
– Pyramid coverage test – want 100%
– Average
g pyramid
py size – want 3-8 μ
μm
• Reflectance measurements: Measure before ARC
– Spectrophotometer
– Spectral Response tester – newer models also enable the
measurement of reflectance.
E.g., Tester Model QEX7 from PV Measurements, Inc.
 Acidic Si Etching QC
• Acidic Si etching may be used for:
– Acidic texturing of mc-Si wafers.
– Rear
Rear-side
side etch edge junction isolation
isolation.
• Reaction is very fast, autocatalytic and exothermic
→ very hard to control.
• If [HF] >> [HNO3] can get porous Si forming.
– Can be detected by visual inspection (coloured films).
• Cost effective inline monitoring of HF/HNO3 etching
baths hot topic at the moment.
• Can take samples of solution and perform various
spectrophotometric
t h t t i andd chemical
h i l tests.
t t
– Off-line but doesn’t involve stopping production.
 Acidic Si Etching QC (cont)
• Detection of H2SiF6 with ion chromatography.
– Separate F and Si by ion chromatography.
– Si complex is detected by UV-VIS
UV VIS spectrophotometer.
spectrophotometer
– F concentration is measured by ion selective electrode or
conductivity.
 Expected Reflectances
• Mono-crystalline wafers
– Start with 33% then saw-damage etch.
– 10
10-11%
11% after alkaline texturing
– < 1% after ARC (at minimum λ).
• Multi-crystalline wafers
– N
No saw ddamage etch
t h
– [HNO3] >> [HF] (acidic texturing) → 20-23%
– [HF] >> [HNO3] (isotexturing) → 10-15% but poorer cell
performance
f ((see L
Lecture
t 7).
7)
– <1% after ARC (at minimum λ) – Avg 5-7%
• Expect
p extra shadingg of 5-6% from metal fingers.
g
 Post Texturing Cleaning
• Typical
T i l cleaning
l i process:
– HF acid ~ 10% for 5 min + Rinse (remove silicate/oxides)
– HCl acid ~ 10% for 5 min + Rinse + Dry
y (remove
( metals))
• Cleanliness extremely important!
• Tests
– Purity of water Æ resistivity test Æ 18 MΩ cm
– Visual inspection Æ no water marks
• Water marks may result from dissolved O2 in the water which
oxidises the Si surface leaving SiO2 particles.
• Spin-drying tends to result in more watermarks than IPA drying
• Concern about some IPA remaining on the surface.
– After HF step surface is hydrophobic
 Application of Phosphorus Source
• Important for in-line processes
– Spray – phosphoric acid using ultrasonic spray systems.
– Higher frequency
→ smaller particles.

• Visual
Vis al Inspection
– After drying
– Uniform coating
– No ‘white bits’ → too thick, non-uniform
 SemiLab: Emitter Uniformity
• Laser excitation of carriers.
carriers
• Detect changed surface photo voltage (SPV) a distance
awayy from the excitation ppoint usingg a capacitive
p probe.
p
• Can also do 4-pt probe measurements.
 Diffusion (Dopant) Profiles
• Need to know your diffusion profile so when you “fire
through” Ag it does not penetrate too deep.
– Worst case – can shunt the junction
– But also can metal that has penetrated deep into the junction
can restrict the path for electrons → current crowding.
• Methods of obtaining diffusion profiles
profiles.
– SIMS = Secondary Ion Mass Spectrometry
– ECV = Electrochemical Capacitance Voltage
– SPR = Spreading
S di resistance
it profiling
fili
 Secondary Ion Mass Spectrometry (SIMS)
• Expensive!
E i !
– ~US$350 per test
• Measures all dopants.
p
• Large SA – difficult for textured surfaces

Source: Bentzen, A. Phosphorus diffusion and gettering in silicon solar cells, PhD Thesis, Oslo University
 Secondary Ion Mass Spectrometry (SIMS)

High
Hi h energy iions are supplied
li d by
b an ion
i gun
(1 or 2) and focused on to the target sample
(3), which ionizes and sputters some atoms off
th surface.
the f The
Th secondary
d ions
i are then
th
collected by ion lenses (5) and filtered
according to atomic mass (6), then projected
onto an electron multiplier (7,
(7 top),
top) Faraday
cup (7, bottom), or CCD screen (8).
 Electrochemical Capacitance Voltage (ECV)
• Capacitance
C it readings
di as a function
f ti off V.
V
– Etch, measure, etch, measure… time consuming but
equipment manufacturers have automated this process.
– Only
O l electrically-active
l i ll i dopants
d are detected.
d d

Source: Bentzen, A. Phosphorus diffusion and gettering in silicon solar cells, PhD Thesis, Oslo University
 ECV Profilers

http://www.wepcontrol.com/cv-profiler/comparison/index.htm
 Spreading Resistance Profiling (SRP)
• Spreading resistance (SR) is determined by applying a known I
between the electrodes on the surface of the sample and
measuring the voltage drop between the probes.
• Dopant concentration profile is measured by bevelling the sample
and making a series of SR measurements along the bevel.
• Profile depth is calculated as a function of the bevel angle.
angle
• Correction factors are used to convert the spreading resistance
values to carrier concentration levels.
 Pseudo IV Curve – Suns-VOC
• This is a post-diffusion test.
• Can probe the n+ or p+ layer directly but usually
measurement is done via metallisation layers.
layers
• Illumination is varied and the JSC and VOC is measured
for each illumination.
• Instead of having to measure current a monitor cell can be used
to determine the current.
• Flash with bright light.
• Measure wafers photoconductivity as light decays
(PCD)
• Contactless,
Contactless measure VOC.
• Implied IV curve with NO series resistance.
 Sinton’s Suns-VOC Flash Tester
• Represent the results as:
• Light-intensity vs V
• Pseudo
P d IV curve

http://www.sintonconsulting.com/Brochures/SVOC_0808.pdf
 Sinton’s Suns-VOC Flash Tester

http://www.sintonconsulting.com/Brochures/SVOC_0808.pdf
 PSG Removal
• Visual inspection
– Looks different
– Hydrophilic
y p Æ Hydrophobic
y p
• Schmid/Rena
– PSG removal + edge junction isolation + rear etch
 Edge Junction Isolation
• Need
N d tto ddetect
t t shunting.
h ti
• Voc Contour Test
– Need to p
perform after rear Al
• Also PL & Dark LIT.
 Shunt Detection
• Process-induced shunts.
– Any cracks, scratches that damage the pn junction
– Processes such as diffusion which can result in leakage
current at edges.
• Material-induced shunts
– Due to material properties (e.g.,
(e g precipitates in multi-Si
multi Si
wafers).
• Because shunts are local sites of increased current
they can be detected using thermal methods.
methods
– Thermal contrasts are very weak (< 1 mK).
– Due to the large
g heat conductivity y of Si the thermal contrasts
tend to "smear out" laterally, leading to a poor spatial
resolution resulting image.
 Shunts – Lock-in Thermography
• So
S use llock-in
k i thermography
th h
– A repetitively-pulsed bias is applied to the wafer and the
resulting
g pperiodic temperature
p modulations are detected and
averaged.
• Non-destructive but quite expensive.
• DLIT = Dark Lock-in Thermography
– Pulsed voltage is applied without illumination.
– O l dark
Only d k current flows.
fl
– Can distinguish between linear (ohmic) and nonlinear shunts.
– If DLIT signals are the same strength at 0.5
0 5 V (fwd bias) and
-0.5 V (reverse bias) → linear shunt.
• ILIT = Illuminated Lock-in Thermography
– Enables imaging of series resistance as well.
 Shunts – Lock-in Thermography

• Phase iimages show

Ph h hhow fast
f t heat
h t is
i spreading
di from
f different
diff t
points in the sample → can resolve smaller shunts.
 Shunts – Lock-in Thermography

Good general reference on LIT.

 ARC Quality Control
• Lifetime
– Should have improved through passivation
• Visual Inspection
– Uniform
– Colour ~ blue / black
• Ellipsometry
Elli
– Thickness
– Refractive index
PECVD SiNx Colour Chart

http://cleanroom.byu.edu/color_chart.phtml
 Ellipsometry
• Measures the change of polarisation upon reflection.
• Polarisation change is determined by the sample's
properties (thickness,
(thickness complex refractive index or
dielectric function tensor).
• Analysis of reflected light → refractive index and
thi k
thickness off a thin
thi film.
fil
 Ag/Al Screen Set-Up, Print, Dry
• Visual
Vi l iinspection
ti
– Check for BREAKS or SMEARING
– Check belts for metal (especially for separate firing)
– People vs Machines
 Firing
• Temperature Profile
– Uniform throughout furnace
• Ag
A
– Metal line series resistance
• Al
– Visual inspection – no pimples ?
– Warping / bowing
– Colour
– Voltage
g is the best
indication of BSF
quality.
 Series Resistance Measurements
• Series resistance due to:
– Series resistance of Ag fingers
– Lateral resistance to current flow in the fingers.
– Contact resistance
• Multimeter
– Ok for one-off tests
• IV Curve

• Imaging better because can see all defects at once.

• Corescan, LIT: Longer acquisition times
• PL andd EL: Fast - just
j needd to recordd a single
i l image.
i
Series Resistance: PL Imaging
 Imaging Metallisation Defects

• EL + Optical → detecting breaks in fingers

• EL image also shows handling-induced micro-cracks
andd scratches.
h
Contact Resistance: Corescan
• Operate the cell at short circuit.
circuit
• Scan cell with a focussed light source → local
current.
• Probe (centred in beam) measures the local
potential between metal line and adjacent
surface.
• Can detect local variations but too slow for in-
line measurement.
 Contact Resistance: TLM
• Transmission Line Model (TLM) of measuring contact
resistance.
– See Q4 of minor assignment.
– Specific contact resistivity – ρc (Ω cm2)
– Typically ~ 3 ×10-3 Ω cm2 for screen-printed cells.
– Measure resistance between metal conductors spaced
different distances apart, L. L
Rm = (2 × Rc) + Rlateral
30

25
ance (Ω)

20
Resista

15

10

5
0 2 4 6 8 10
Distance (mm)
 IV Testing
• Measure IV curve under standard test conditions
(STC)
– Irradiance of 1,000 W/m², solar spectrum of AM 1.5 and
module temperature at 25 °CC.
• Calibrate the light with a reference cell.
• Cells tested as part of cell sorting before being
i
incorporatedd into
i modules.
d l

From PVCDROM
 IV Testing: Illumination Sources
• Need to be able to simulate the illumination of the sun
(intensity and spectrum).
• The most common light source is a Xenon arc lamp
with
ith filters
filt installed
i t ll d to
t approximate
i t the
th AM1.5G
AM1 5G
spectrum.

From PVCDROM
 IV Testing - Probing
• Use 4-pt
4 pt probes to contact the cell.
cell
• Metal of the block acts as the rear current probe and a
voltage pin is inserted within the block.
• Typically more than one probe is used for the front
contacts.

From PVCDROM
 IV Testing - Scanning
• Measure ISC and VOC separately.
separately
• Then scan in two sections:
– From 0 to 70 % of VOC.
– More closely spaced V values from 70% to VOC (contains the
MPP, the VOC and has a much higher slope.

From PVCDROM
 Dark IV Testing
• Record an IV curve in the dark - useful to determine diode
properties.
• Slope (log plot) → ideality factor.
• Different parts of the curve are dominated by different loss
mechanisms. See simulation in PVCDROM (Ideality factor).
• Note: I flows in the opposite direction and the current paths are
g in the current path
different. The change p causes a lower Rs in the
dark measurements to the light measurements.
Dark

Light

From PVCDROM
 Spectral Response (EQE & IQE)
• Spectral response is one of the richest sources of
information for fabricated solar cells.
• External Quantum Efficiency (EQE) = ratio of the
number of charge carriers collected by the solar cell to
the number of photons of a given energy shining on the
solar cell from outside.
• Internal Quantum Efficiency (IQE) = ratio of the
number
b off charge
h carriers
i collected
ll t d by
b the
th solar
l cell
ll to
t
the number of photons of a given energy that shine on
the solar cell from outside and are not reflected back by
the cell, nor penetrate through.
• Measure the photo-generated current (ISC) at each
wavelength.
l h
 Spectral Response (EQE & IQE)
• Most testers compare the photo-generated current
with that of a calibrated reference cell.
• Measure EQE andd R andd calculate
l l IQE.

IQE = EQE(λ)/(1-R(λ))

(See Lecture 7)
 Spectral Response Measurements
• Need to make sure you measure EQE and R in similar
regions (so both similarly affected by finger shading)

Typically
yp y IQE Q (λ))/(1-R(
Q = EQE( ( (λ))

Where M is the metallised (shaded)

area fraction; and
Rcell is the measured reflectance of the
cell.
ll
IQE (SemiLab’s LBIC Method)
Process Parameter Technique
Crystal growth (ingot quality) Lifetime, resistivity PCD/μ-PCD
Incoming wafer inspection Surface roughness Optical
Thickness Metrology
Electronic quality of wafers Lifetime/resistivity PL /PCD/μ-PCD/LBIC
Saw damage etch Thickness Metrology systems
Texturing Reflectance Spectrophotometry
Pyramid height/% Coverage SEM/Optical

Diffusion Sheet resistance p pprobe, Emitter imaging

4-pt g g
Dopant profile SIMS/ECV/SPR
Edge junction isolation Shunt resistance VOC contour map, PL
ARC coating ARC thickness Ellipsometry/Optical (colour)
Refractive index Ellipsometry
Metallisation Line width/profile Optical/SEM
Shunt resistance PL/EL LIT
PL/EL,
Series resistance PL/EL
JSC,VOC, FF, η IV Tester
IQE, EQE, R Spectral response tester
Implied VOC Suns VOC
Lifetime imaging PL /PCD/μ-PCD/LBIC
Process Parameter Technique
Crystal growth (ingot quality) τ PCD
Wire sawing quality/wafer cleanliness Surface roughness Optical microscopy

El t i quality
Electronic lit off wafers
f τ,L PL or PCD + Optical
O ti l
Saw damage etch Thickness Metrology systems
Texturingg Reflectance Spectrophotometry
p p y
Pyramid height SEM/Optical
% Coverage

Diffusion Sheet resistance 4 pt probe, Emitter imaging

4-pt
Dopant profile SIMS/ECV/SPR

Edge junction isolation Shunt resistance VOC contour map

ARC coating ARC thickness Ellipsometry/Optical colour
Refractive index Ellipsometry
Metallisation Line width/profile Optical/SEM
Shunt resistance PL, LIT,
Series resistance PL/EL
JSC,VOC, FF, η IV Tester
IQE, EQE, R Spectral response tester
Pseudo IV Suns VOC
Lifetime imaging PL
 Next Week
• High Efficiency Solar Cells