Long term stability of dye solar cells – meeting IEC 61646 ......Dyesol >25,000 h quasi-continuous illumination 25-40 years life time extrapolated, depending on location R.Harikisun,

Commercial In Confidence Copyright Dyesol 2012 1

Long term stability of dye solar cells –

meeting IEC 61646 requirements

Hans Desilvestro, Nancy Jiang,

Martin Berry and Paul Murray

Presenter: Ben Wilkinson (Dyesol UK)

4 April 2012

Outline

� Motivation

� The challenge

� Disentangling degradation mechanisms through DSC

accelerated testing


− Hydrophilic vs hydrophobic dye

− IV data

− IPCE (Incident Photon-to-Electron Conversion Efficiency)

− EIS (Electrochemical Impedance Spectroscopy)

� Towards larger devices

� Summary

Motivation

� Develop integrated understanding of all materials

and design aspects of dye solar cells (DSC)

‒ To best serve and advise our commercial partners

‒Optimum focus on most promising materials and

technologies


‒

‒

technologies

� Achieve optimised LCOE for any given application

‒ Performance

‒ Stability

‒ Cost

Stability

Performance

1/Cost Stability

Performance

1/Cost

The challenge

� To achieve grid parity under a moving target of

declining PV prices

� Performance – Stability - Cost

PerformanceHero cells

Performance

Industrial design today


� 20+ years life time in building applications

� Standards: PV-specific + Building Standards

Stability1/Cost Stability1/Cost

20-year product life goal

System’s requirements

At the molecular level

>100 million turnovers �� at least for certain dyes/electrolytes

~No dye desorption �� at least for certain dyes/electrolytes

Loss of NCS-, substitution by ?

Lund et al vs Falaras et al. Low 80-


Loss of NCS-, substitution by I-, nitriles, imidazoles, etc.

?Lund et al vs Falaras et al. Low 80-85oC stability with nitrile solvents

Isomerisation, e.g. N- to S-bound SCN -

?Not significant according to micro-Raman (Falaras et al)

Decomposition of electrolyte components

?Little is known, more in situ spectroscopic work required

Stability of electrocatalyst? ��At least for Pt and certain electrolytes

At the cell level (Dyesol up to 250 mm length): Seals



Suppress ingress of O2, H2O �� Dyesol, Fujikura, Fraunhofer ISE

Suppress egress of solvent �� Dyesol, Fujikura, Fraunhofer ISE

Performance stability, 85oC/1,000h; -40oC/+85oC

at least for certain dyes/electrolytes, 85oC/85% r.h.: requires glass-based


85oC/1,000h; -40oC/+85oC thermal cycling; 85oC/85% r.h./1,000h

��85oC/85% r.h.: requires glass-based encapsulation (i.e. similar to CIGS or CdTe), best assessed at the panel level

� Excellent DSC durability under light soaking conditions (~60oC). E.g. Dyesol >25,000 h quasi-continuous illumination �� 25-40 years life time extrapolated, depending on locationR. Harikisun, H. Desilvestro, Sol. Energy, 85, 1179 (2011), “Long-term stability of dye solar cells”

� IEC 61646 85oC/1,000h and thermal cycling tests remain challenging for DSC

At the module level

- Cell-to-cell interconnects

- Corrosion protection of current collectors

- Potential shunt paths




- Potential shunt paths

- Sealing

At the system’s level

- Environmental, temperature extremes, hail, etc.

- Building code requirements

- Maximum power point tracking independent of

any ageing phenomena

Cell chemistry

NN

HO C

NN

HO2C

RuN

CS

NC

S

CO2TBA Z907N719


� N719 (hydrophilic) vs Z907 (hydrophobic)

� High-boiling solvent, non-nitrile

� Pt catalyst based on Dyesol Platinum Paste PT1

� 8××××11 or 8××××168 mm active area

NHO2C

CO2TBA

S

Cell Efficiency

Effectively meeting IEC 61646 for most practical light levels

Loss mainly due to loss of Jsc - Why?0

1

2

3

4

5

6

7

8

0 200 400 600 800 1,000 1,200

Eff

icie

nc

y (

%)

Time (hours)

1,000h, 85oC

1/3 sun, Z907

1/3 sun N719

1 sun, N719

1 sun, Z907

-10%-10%

-18%-18%


• Virtually no loss of Voc for

either dye (≤≤≤≤3.5%)

• At 1 sun N719 loses less

Jsc, but some ff (-2%)

while Z907 loses more Jscand gains some ff (+3%)

0

2

4

6

8

10

12

14

0 200 400 600 800 1,000 1,200

Jsc

(mA

/cm

2)

Time (hours)

1,000h, 85oC

1/3 sun, Z907

1/3 sun N719

1 sun, N719

1 sun, Z907

-13%

-19%

-7%-8%

30%

40%

50%

60%E

QE

(%

) N719, initial N719, 300h 85oC N719, 1000h 85oC Z907, initial Z907, 300h 85oC Z907, 1000h 85oC

IPCE (after 85oC storage)Incident Photon-to-Electron Conversion Efficiency


0%

10%

20%

30%

400 500 600 700 800 900

EQ

E (

%)

Wavelength (nm)

Some loss of IPCE over large part of spectrum over the first 300 h at 85oC, then

almost complete recovery in the 400 to ~570 nm region (due to decreasing

conduction band level?)

IPCE

� No major dye desorption or decomposition such as ligand exchange

� Loss of IPCE under low intensity monochromatic light is only partly responsible for loss of Jsc under higher intensity light, particularly

• after 1,000 h 85oC storage


• after 1,000 h 85 C storage

• at the higher light levels of 1 sun

• for Z907

∆∆∆∆(IPCE) *) ∆∆∆∆Jsc (0.33 sun) ∆∆∆∆Jsc (1 sun) ∆∆∆∆(IPCE) *) ∆∆∆∆Jsc (0.33 sun) ∆∆∆∆Jsc (1 sun)

N719 -6.4% -3.2% -7.4% -3.7% -6.6% -13.1%

Z907 -7.3% -5.8% -14.6% -3.9% -8.0% -18.9%

300 h 1,000 h

Jsc as a function of light level

and storage time at 85oC

10

15(m

A/c

m2)

N719, initial

N719, 300h

N719, 1000h


Serious current limitation at the higher light levels as a

result of thermal stress testing, particularly for Z907

0

5

0.0 0.2 0.4 0.6 0.8 1.0

Jsc

(mA

/cm

Sun level

Z907, initial

Z907, 300h

Z907, 1000h

100

200

-Zi(O

hm

cm

2)

Initial400h 85oC1000h 85oC

Z907

0.1Hz

100

200

(Oh

m c

m2)

Initial400h 85oC1000h 85oC

N719

0.1Hz

EIS Electrochemical Impedance Spectroscopy, 0.4V, 0.33 sun


00 100 200 300

Zr (Ohm cm2)

00 100 200 300

-Zi(O

hm

cm

Zr (Ohm cm2)

• Significant decrease of Rbr with prolonged

85oC storage, mainly from 400 to 1,000h

• Significant increase of Rd with prolonged

85oC storage from 400 to 1,000h

• Significant increase of CE Rct at 0.4 V, vs

decrease at 0.7 or 0.8 V

• Initial increase of Rbr then decrease from

400 to 1,000h

• More significant increase of Rd with

prolonged 85oC storage compared to N719

• Significant increase of CE Rct at 0.4 V, vs

decrease at 0.7 or 0.8 V

EIS (AC impedance)Transmission line model

rbr rbr rbr

Transport resistance, TiO2 resistance


Electron collection efficiency = Rbr/(Rbr+Rt)

Electron diffusion constant De = d2 / (RtCc)

Electron diffusion length Ln = d (Rbr/Rt)1/2 d = TiO2 layer thickness

a) F. Fabregat-Santiago, et al Solar Ener. Mat. and Solar Cells 87, (2005) 117-131. b) Hoshikawa, et al. J. Electroan. Chem., 588 (2006) 59

Cc: chemical capacitance = dQ/dVRbr: electron backtransfer resistance

4

5

N719, initial

Conclusions from EIS

1) 85oC: Rbr ��, Rt: no major change →→→→ ηηηηcoll �� →→→→ Jsc ��

- to a larger extent for Z907 than N719 (various batches)

- in contrast to light soaking (LS) where ηcoll hardly decreases


0

1

2

3

4

0.6 0.7 0.8 0.9 1.0

j sc

(m

A/c

m2)

ηηηη coll: electron collection efficiency at 0.4V

Z907, initial

N719, LS

Z907, LS

N719, 85oC

Z907, 85oC

Conclusions from EIS

1) 85oC: Rbr ��, Rt: no major change →→→→ ηηηηcoll �� →→→→ Jsc ��

- to a larger extent for Z907 than N719

- in contrast to light soaking (LS) where ηcoll hardly decreases

2) 85oC: evidence of conduction band downward shift by ~50-100

mV. Possibly reason for increased IPCE over large part of


spectrum. In contrast to light soaking with no significant shift of

Vcb over 1,000h

3) 85oC: Rd ��, due to diffusion polarisation under photogeneration,

particularly for Z907 →→→→ Jsc ��, diffusion limitation →→→→ Vmpp �� (Vocis much less affected, no diffusion polarisation)

I3- (CE) �� →→→→ Rct(photogeneration) ��

while Rct(0.7 or 0.8V) �� due to ‘standard’ Pt activation

Conclusions from EISI3

- diffusion polarisation

-4

-2

0

2

4I3

- diffusion resistance @ 0.33sun

Initial

400h 85oC

1000h 85oC

400h light soaking

1000h light soaking

j (m

A c

m-2

)N719


-6

0 20 40 60 80 100 120 140 160 180

1000h light soaking

-8

-6

-4

-2

0

2

4

0 20 40 60 80 100 120 140 160 180Rd (Ohm)

Initial

400h 85oC

1000h 85oC

400h light soaking

1000h light soaking

j (m

A c

m-2

)

Z907

Conclusions from EISI3

- diffusion polarisation

� Mainly occurs under increasing photocurrents, due to I3

- concentration polarisation

� Much less diffusion polarisation under net negative currents, due to high [I-]/[I3

-] concentration ratio

− most notable with Z907 as a result of 1,000 h at 85oC

� Much more pronounced for Z907:


� Much more pronounced for Z907:

− significant increase of Rd as a result of light soaking

− significant increase of Rd as a result of 400 h at 85oC

− very significant increase of Rd as a result of 1,000 h at 85oC

� Much less pronounced for N719:

− almost no increase of Rd as a result of light soaking

− some increase of Rd as a result of 400 h at 85oC

− very significant increase of Rd as a result of 1,000 h at 85oC

Thermal cycling – IEC 61646


60%

80%

100%R

ela

tive

eff

icie

ncy (

1 s

un

)

Aft

er

30 m

in i

llu

min

ati

on

at

1 s

un

Thermal cycling – IEC 616468×168 mm, MPN-based electrolyte


0%

20%

40%

0 50 100 150 200 250

Rela

tive

eff

icie

ncy (

1 s

un

)

# Cycles

Aft

er

30 m

in i

llu

min

ati

on

at

1 s

un

� No visual changes

� No electrolyte leaks

� After 200 thermal cycles with 168mm long cells

− temporary loss of 20% efficiency, probably due to

nitrile-based solvent

Thermal cycling – IEC 61646Results / towards larger devices


nitrile-based solvent

− largely recoverable after 30 min illumination (@ 1 sun)

� Similar tests for larger multi-cell devices and with

electrolytes offering improved high temperature stability

� Very promising chemical and mechanical stability achievable under IEC 61646 85oC storage and thermal cycling

� Loss of Jsc (rather than Voc or ff) under 85oC storage with the

specific cell chemistry is due to

− decreasing electron collection efficiency ηcoll− increasing I - diffusion resistance

Conclusions


− increasing I3- diffusion resistance

� Chemical reasons for increased diffusion resistance upon high temperature exposure are presently not known

− more in situ spectroscopic and electrochemical work

required

� Strategies in place to further improve DSC high temperature stability


Thank you for your attention!

www.dyesol.com

Documents

Long term stability of dye solar cells – meeting IEC 61646 ......Dyesol >25,000 h quasi-continuous illumination 25-40 years life time extrapolated, depending on location R.Harikisun,