Characterization Issues for Bifacial Solar Cells

Ronald A. Sinton Sinton Instruments, Inc., Boulder, Colorado USA

Characterization of bifacial solar cells.

-Device Physics and IV-curves-Low-efficiency vs. high-efficiency-Complex IV-curve interpretation-2D Effects: So-called “electrical shading”-Passivation stability (Al BSF was stable, opaque, and eff limiting)

-Measurement Strategies-Ambiguous Voltage: Bifacial cells have shadow loss due to

2

-Ambiguous Voltage: Bifacial cells have shadow loss due to back metalization. Therefore metal is minimized.-Small contact pads at high current densities-Cheating

-Capacitance issues for high-efficiency and n-type solar cells

NASA Helios goes to 96863 ft (29.7 km) on Bifacial Solar Cells

Oops!

Bummer

SunPower Backside-Contact Bifacial Cells

Reduced metalization fraction on back reduces front efficiency.

Zhou, Verlinden, Crane, Swanson, Sinton PVSC 1997.

Differences Between Traditional &High-Efficiency-Bifacial CellsHigh-Efficiency-Bifacial Cells

Classic Al BSF Production Solar Cell

heavy n+ diffusion (40 ohms/sq)

Fully 1DIdeality factor = 1, enforced by Al BSF and heavy n+ diffusion

[ ] ( )( )[ ]2,,

i

Abackoefrontoe

bulkGen

n

npNJJ

nqWJJ

∆∆++−∆−=τ

Al BSF

Classic Al BSF Production Solar Cell

heavy n+ diffusion (40 ohms/sq)

Al BSF

Fully 1D !

Ideality factor = 1, enforced by Al BSF and heavy n+ diffusion

Lsh

ssJ

R

IRV

kT

IRVqJJ −−+

−

−= )(1

)(exp0

One-diode model!

Suns-Voc Curves, Al BSF

0,02

0,03

0,04

Cu

rre

nt

De

nsi

ty (

A/c

m2) FZ (n=1)

5 x 1011 cm-3 Fe

0

0,01

0,0 0,2 0,4 0,6

Cu

rre

nt

De

nsi

ty (

A/c

m

Voltage (V)

B:O CZ

1.6 ohm-cm p-type substrate, 250 µm thick

Prototypical Bifacial Point-Contact Cell

Light n+ diffusion (120 ohms/sq), with nitride passivation, selective emitter

Ideality factor = ?

[ ] ( )( )[ ]2,,

i

Abackoefrontoe

bulkGen

n

npNJJ

nqWJJ

∆∆++−∆−=τ

Al2O3 with light B diffusion, selective emitter

Suns-Voc Curves, Passivated Backside

0,015

0,02

0,025

0,03

0,035

0,04

Cu

rre

nt

De

nsi

ty (

A/c

m2)

FZ

Fe

0

0,005

0,01

0 0,2 0,4 0,6

Cu

rre

nt

De

nsi

ty (

A/c

m

Voltage (V)

B:O CZ

Lsh

ssJ

R

IRV

nkT

IRVqJJ −−+

−

−= )(1

)(exp0

Completely General Analysis (Good Physics)

Parameters

Eff pEff

Jsc

Jmp

Voc

Vmp Voc (0.1 sun)Vmp Voc (0.1 sun)

FF pFF

Rs Rs

Rsh

Suns-Voc curve

Injection-dependence: Both Bulk and Surface

Injection dependence needs to be taken into account.

This has been pointed out many times:

Fe:B (Macdonald PIP 2000)B:O (Bothe 2006)Al2O3 vs. SiNx (Schmidt EPSEC 2008)Highly-injected substratesHighly-injected substrates

1-diode or 2-diode models almost never apply real-world solar cells. It is best to avoid forcing data into inappropriate models.

Suns-Voc analysis automatically incorporates the full injection dependence.

Passivation stability

Both front and back surface passivations need:Both front and back surface passivations need:-UV stability-Temperature/humidity stability

To be measured in a flash tester:-Light-soaking, time-dependent effects need to be absent or characterized (Fe, B:O, passivation charging*)

*Seiffe et al, JAP 109 064505 2011

2D Effects: N-type IBC Cell

Light n+ diffusion (200 ohms/sq), with nitride passivation

Ideality factor = ? 2D, 3D effects, high-level injection.

[ ] ( )( )[ ]2,,

i

Dbackoefrontoe

bulkGen

n

nnNJJ

nqWJJ

∆∆++−∆−=τ

Light IV vs. Suns-Voc vs. Dark IV

Follow the holes at V = Vmp: (for example 620 mV)

Light IV

Suns-Voc

Dark IV

Dark IV weights the shortest current path between electrodes. May be irrelevant.+-

Light IV vs. Suns-Voc vs. Dark IV

Follow the holes at V = Vmp: (for example 620 mV)

Light IV

Suns-Voc

Dark IV

Dark IV weights the shortest current path between electrodes. May be irrelevant.+-

Shading

n+nn+ n+np+n+nn+ n+np+

“Electrical Shading” ???

n+nn+ n+np+

V

n+nn+ n+np+

“Shadow-like losses” Dross et al. 2005“Shading effects” Hermle et al. 2008“Electrical Shading” Granek 2010

So-Called Electrical Shading: Diagram at Short Circuit

n+nn+ n+np+

0,04

Pile-up of holes due to 2-D current crowding

Local implied Voc of 625 mV, even at short circuit!

0

0,005

0,01

0,015

0,02

0,025

0,03

0,035

0 0,2 0,4 0,6

Cu

rre

nt

De

nsi

ty (

A/c

m2)

Voltage (V)

-Distance from junction-Current crowding-High local voltage (pn product)-Recombination loss

Shading is not a good physical analogy.

n-type, 3e15 cm-2 doping, 1- mm n-type stripes, 180-µm thick wafer

So-Called Electrical Shading: Diagram at Short Circuit

n+nn+ n+np+

Pile-up of holes due to 2-D current crowding

Local implied Voc of 625 mV, even at short circuit!

Stranded regions?Stranded regions?Carrier-diffusion bottlenecks?High-voltage islands?Carrier diffusion constriction?Lateral carrier-collection bottlenecks?

Seitliche Überschüssigeladungsträgerdichte Samlung Engpässe?

Measurement Challenges

Cells with Contact Pads, High Current Per Pad

SunPower A-300 specification sheet, circa 2003.

Multi-busbar cells have ambiguous voltage prior to interconnection

-Cu ribbons are an essential part of the cell metalization

-Bifacial cells need minimal metalization shadowing (on back)

-Trends to minimize Ag also lead in this direction

-Less metal means higher series-resistance losses:-ambiguous “cell voltage”.-ambiguous “cell voltage”.

-New designs sometimes lack redundant metalization connections-Almost-isolated cells in parallel, prior to ribbon connection.

The cell voltage of cell without ribbons can vary widely across the cell. Be careful!

Strategy: Pull fixed current from each pin

This makes the result independent of pin and harness resistance, and balances both the current and voltage on the cell.

Su

I-I+

Place the voltage pin close, but not too close to the current pin.

In more complex designs, engineer the measurement so that the current from each pin is proportional to the area of the cell that it “drains”. This is generally a good emulation of what will happen for the cell with interconnects.

Voltage pins can be averaged, or usually one is enough if you have enforced a single operating point across the cell.

Size of future metal tab

Current PinVoltage Pin

Measuring Short-circuit current with no shadow? Is it valid?

This seems likely to lead to “cheating”:

Designing cells to test high prior to interconnect.This leads to disappointing results at the module level.

Also, in standard measurements, placement of voltage probes can be used to “optimize” the FF.

Capacitance

High efficiency cells, especially n-type, are difficult to measure with flash testers.

Ramp-rate Artifacts

Hypothetical 720mV 200 µµµµm Silicon CellIV Curves for Different Ramp Rates

0.03

0.035

0.04

0.045

Steady State100 ms50 ms20 ms10 ms5 ms2 ms

Voltage Ramp Rates

0.6

0.7

0.8

50 20 105

0

0.005

0.01

0.015

0.02

0.025

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8

0

0.1

0.2

0.3

0.4

0.5

0.6

0 20 40 60 80 100Time (ms)

Vol

tage

per

Cel

l

100 ms ramprate

High-Efficiency n-type Cells: 100X Higher Capacitance!

SunPower,Sanyo HIT

Junction Voltage (mV)

Conventional Devices

Brute-Force Works, Use 200+ ms, or 1mV/ms Ramp

Hypothetical 720mV 200 µµµµm Silicon CellIV Curves for Different Ramp Rates

0.025

0.03

0.035

0.04

0.045

Steady State100 ms50 ms20 ms10 ms5 ms2 ms

Voltage Ramp Rates

0.6

0.7

0.8

Vol

tage

per

Cel

l

50 20 105

0

0.005

0.01

0.015

0.02

0.025

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8

0

0.1

0.2

0.3

0.4

0.5

0 20 40 60 80 100Time (ms)

Vol

tage

per

Cel

l

100 ms ramprate

Novel Approach to Neutralize Capacitance Artifacts

Intensity and Module Voltage vs Time

1.5

2.0

2.5

Ligh

t Int

ensi

ty (

suns

)

50.0

52.0

54.0

56.0

58.0

60.0

Mod

ule

Vol

tage

(V

)

0.0

0.5

1.0

0.0000 0.0020 0.0040 0.0060 0.0080Time (s)

Ligh

t Int

ensi

ty (

suns

)

40.0

42.0

44.0

46.0

48.0

50.0

Mod

ule

Vol

tage

(V

)

Sinton, De Ceuster, Wilson, Barbosa, EPSEC, Barcelona 2005 US patents 7309850 (2007), 7696461 (2010).

Short Pulses

Intensity vs Time

2.0

2.5

3.0

3.5

Ligh

t Int

ensi

ty (

suns

)

1/4

0.0

0.5

1.0

1.5

0.000 0.002 0.004 0.006 0.008Time (s)

Ligh

t Int

ensi

ty (

suns

)

1/2

Full

Capacitance Artifacts in Solar Cell IV Curves.

6

8

10

12

14

16

Err

or in

Pow

er M

easu

rem

ent (

%) Terminal Voltage = Vmp

NEW METHOD: V = Constant - (K*current)

1/2

1/4

0

2

4

6

- 180 suns/sec - 340 suns/sec - 660 suns/sec

Light pulse slew rate at One Sun

Err

or in

Pow

er M

easu

rem

ent (

%)

Full pulse

1/2

0.2%0.1%

Use Engineered Cell Measurements: Only Module Eff Counts

Cell Measurements are NOT comparable between labs, unless you duplicate thethe chuck, the current and voltage probe placements, and other details.

FF matching between labs is difficult. Naked cells have ambiguous voltages.

Know what you are (and are not) measuring. (Cell design and device physics)

Pull the cell down to a conducting chuck? -information is lost about lateral back metalization conductance.

Spectrum:-Class A does not solve all issues.

-900-1100 band (light trapping)-300-400 band (EVA vs. Silicone, glass quality, selective emitters)-measurements on naked cells do not see AR, texture, encapsulant effects.

It is best to understand your cells, and use detailed spectral response to optimize current from the cell. A simulator is a crude instrument for this optimization.

Target measurements to determine the critical parameters.

Conclusions

Characterization issues for Bifacial Cells.

-Pay attention to 2-D effects and the injection dependence of substrates

and surface passivations. Be wary of ideal-diode curve fits.

-Bifaciality requires a compromise between metalization shadowing and

series resistance

-ambiguous voltages during measurement!

39

-High-efficiency cells can have capacitance artifacts during measurements

-especially n-type.

-Use an engineered approach to cell measurement. Know your cells, what

you are measuring, control the measurement, and avoid the temptation to

cheat, (since this will only cause disappointment at the module level).

This work was supported by the customers of Sinton

Instruments that are so kind as to buy stuff

Special thanks for Keith Forsyth and Adrienne Blum at Sinton

Instruments.