Outdoor Performance of Organic Phot - PowerFilm Solar Inc.

J. Renewable Sustainable Energy 7, 013111 (2015); https://doi.org/10.1063/1.4906915 7, 013111

© 2015 AIP Publishing LLC.

Outdoor performance of organicphotovoltaics: Diurnal analysis, dependenceon temperature, irradiance, and degradationCite as: J. Renewable Sustainable Energy 7, 013111 (2015); https://doi.org/10.1063/1.4906915Submitted: 06 October 2014 . Accepted: 17 January 2015 . Published Online: 29 January 2015

N. Bristow, and J. Kettle

ARTICLES YOU MAY BE INTERESTED IN

Effects of concentrated sunlight on organic photovoltaicsApplied Physics Letters 96, 073501 (2010); https://doi.org/10.1063/1.3298742

Temperature dependence for the photovoltaic device parameters of polymer-fullerene solarcells under operating conditionsJournal of Applied Physics 90, 5343 (2001); https://doi.org/10.1063/1.1412270

Two-layer organic photovoltaic cellApplied Physics Letters 48, 183 (1986); https://doi.org/10.1063/1.96937

Outdoor performance of organic photovoltaics: Diurnalanalysis, dependence on temperature, irradiance,and degradation

N. Bristow and J. Kettlea)

School of Electronic Engineering, Bangor University, Dean St., Bangor, Gwynedd, Wales,United Kingdom

(Received 6 October 2014; accepted 17 January 2015; published online 29 January 2015)

The outdoor dependence of temperature and diurnal irradiance on inverted organic

photovoltaic (OPV) module performance has been analysed and benchmarked

against monocrystalline-silicon (c-Si) photovoltaic technology. This is first such

report and it is observed that OPVs exhibit poorer performance under low light

conditions, such as overcast days, as a result of inflexion behaviour in the current-

voltage curves, which limits the open-circuit voltage (VOC) and fill factor. These

characteristics can be removed by photo-annealing at higher irradiance levels,

which occur diurnally as irradiance increases after sunrise. We also report the first

temperature coefficients for OPVs from outdoor data; the OPV modules showed a

positive temperature coefficient, which compared to a negative coefficient from the

c-Si modules. Overall, the cell degradation outdoors appears very severe for these

modules and highlights the need for improved barrier. VC 2015 AIP Publishing LLC.

[http://dx.doi.org/10.1063/1.4906915]

I. INTRODUCTION

Remarkable improvements in organic photovoltaic (OPV) performance have been achieved

by the use of bulk heterojunction OPV devices, in which conjugated polymer donors are blended

with electron acceptors such as fullerene derivatives in solution. Currently, record power conver-

sion efficiencies (PCEs) of over 10% has been demonstrated in laboratory conditions by a number

of companies1 and 11% has been reported in the literature.2 Whilst such efficiencies are impres-

sive, understanding how such devices operate in the real world under a variety of atmospheric

conditions is imperative in order to fully commercialise the technology.

To support the development of the technology, ISOS consensus standards have been devel-

oped to standardise testing protocols between international laboratories.3 Included in these ISOS

standards are three levels of outdoor performance measuring protocols. These tests require addi-

tional test facilities not commonly available in laboratories, including suitable measurement equip-

ment and complex data analysis software packages. Due to the complexity of this, few groups

have developed test facilities for conducting outdoor tests and therefore limited data exists on the

outdoor performance of OPVs. However, these are some of the most useful tests as they allow for

performance of materials and systems to be understood in an environment where the modules are

likely be deployed and to understand how the behaviour of OPVs during outdoor testing is

affected by the climatic conditions of the dynamic testing environment.4,5 Despite these challenges,

there have been some comprehensive studies. Katz et al. published a study of outdoor degradation

based on OPVs, encapsulated with glass and aluminium, with different semiconductors in the

active layer6 and showed that the main losses occur due to falls in short circuit current density

(JSC) and fill factor (FF), which were equated to degradation of the back electrode. Overall, the

cells were relatively unstable and showed after 90 h, the best cell dropped to 50% of the original

a)Email: [email protected]. Tel.: 01248 382471.

1941-7012/2015/7(1)/013111/12/$30.00 VC 2015 AIP Publishing LLC7, 013111-1

JOURNAL OF RENEWABLE AND SUSTAINABLE ENERGY 7, 013111 (2015)

efficiency, despite the glass encapsulation. Another approach was adopted by Hauch et al., who

used an additional barrier film to encapsulate flexible Poly(3-hexylthiophene-2,5-diyl):Phenyl-C61-

butyric acid methyl ester (P3HT:PCBM) modules on Polyethylene terephthalate (PET)7 and

showed over 1 year, the performance dropped by only 20% of the original efficiency. Other work

includes that of Krebs who showed how edge sealing plays a significant role in preserving the

OPV from environmental degradation by using glass-fibre reinforced thermosetting epoxy (in this

case, prepreg8) as an edge sealant, which demonstrated a reduction in efficiency of only 65% over

a year.9 In addition there have been a number of inter-laboratory stability tests; Gevorgyan et al.conducted an inter-laboratory monitoring programme of flexible modules of P3HT:PCBM on PET,

which were encapsulated by a barrier film from Amcor Flexibles and studied at different outdoor

locations.10 On average, the performance dropped by 40% of the original efficiency after approxi-

mately 1000 h (�42 day) of outdoor exposure. Few of these reports present data showing how an

OPV performs relative to other modules positioned at the same site or have presented measure-

ments to explain how climatic conditions affect the OPVs performance.

In this work, data are presented for two separate outdoor monitoring campaigns performed

on OPV modules made by the Technical University of Denmark (DTU), supplied as part of

their freeOPV programme.11–13 The focus of this work is benchmarking the performance of

OPV modules relative to c-Si modules and identifying how climatic conditions affect solar cell

performance. For the first time in literature, the temperature coefficient of OPV modules is

reported based on outdoor performance data. Results are also shown for how module degrada-

tion is affected by seasonal variation during the summer and winter.

II. EXPERIMENTAL

A. Organic photovoltaic modules

Roll-to-roll coated OPV modules produced at the DTU were used for the tests. The fabrication

of these modules is reported by Krebs et al.12 and were encapsulated using flexible Amcor packag-

ing barrier foil and epoxy adhesive. The devices had an ITO free structure of Ag-grid/Poly(3,4-

ethylenedioxythiophene)-poly(styrenesulfonate) (PEDOT:PSS)/ZnOX/P3HT:PCBM/PEDOT:PSS/

Ag-grid/PET-substrate. The device terminals were connected to easily accessible electric plugs,

which were connected to a switch matrix and Source Measurement Unit (SMU). The modules have

a PET/SiOX barrier layer with a refractive index which is similar to that of glass (�1.57). Prior to

monitoring, the modules were fixed on a platform so that air could circulate behind the panels.

B. Outdoor monitoring setup

The OPV modules were measured in at the School of Electronics, Bangor, Gwynedd,

North Wales, which has latitude and longitude of 53.2280� N, 4.1280� W and is located at low

altitude (20 m above sea level) and 250 m from the Menai Straits (Irish Sea). Long term cli-

matic average temperatures for the winter are 4.7 �C and for the summer are 14 �C. However,

the first of these tests was conducted during the summer of 2013, which corresponded to a pe-

riod of abnormally hot and humid conditions, with an average mean temperature of 16.5 �C and

an average maximum temperature of 20.2 �C. Data are also supplied for a winter measurement

campaign during which climatic conditions were milder than usual with an average mean tem-

perature of 8.8 �C, an average minimum of 3.1 �C, and an average maximum temperature of

11.7 �C. The humidity levels for both summer and winter were very similar: with an average

mean of 79%, an average maximum of 90%, and the average minimum being 61% in summer

and 65% in winter. UV indices were very different with an average of 1.07 in the summer (av-

erage maximum 5.45) and 0.43 in the winter (average maximum 2.38).

The outdoor monitoring system at Bangor University is on the roof of the School of

Electronic Engineering. Two 185 Wp silicon modules (manufactured by Pure Wafer Solar Ltd.,

Swansea) which are monitored using an Egnitec PVMS250 PV measurement system (manufac-

turer Egnitec Ltd., Caernarfon, Gwynedd, UK). The modules are kept at maximum power point

in between periodic current-voltage (IV) sweeps (once every minute) and each has a PT100

013111-2 N. Bristow and J. Kettle J. Renewable Sustainable Energy 7, 013111 (2015)

temperature sensor fixed to its backplane. Current and voltage at the maximum power point

(IMPP, VMPP) and PT100 measurements are taken every 15 s.

The OPV measurement system comprises an 8-channel measurement unit with switch matrix and

two separately adjustable mounting boards on a rack (Figure 1). Each board can have up to nine OPV

modules mounted on it and an IMT GmbH solar silicon reference cell for irradiance measurements.

There are two PT100 temperature measurement channels: one was mounted on a horizontal module

and the other on an inclined module. For the purpose of this experiment, one silicon module and one

OPV mounting board were mounted horizontally and the other silicon module and OPV mounting

board were inclined at 35�, which is the optimum inclination for maximising solar power over a year

in Bangor, Gwynedd. The outdoor measurement setup conforms to the ISOS-O-1 outdoor measuring

protocol.3 The data were analysed using a combination of MySQL, MS Access, and MS Excel.

III. RESULTS

Eight DTU modules were monitored for eleven weeks over the summer of 2013 (start date:

09/07/2013) and for fourteen weeks over winter 2014 (start date: 14/01/2014). For the summer

campaign eight modules were monitored with four mounted on the horizontal rack and four

mounted on the inclined rack. For the winter campaign, only three modules were monitored, all

mounted on the inclined rack.

A. Comparison with c-Si modules

Initially, two contrasting days were used to compare relative OPV performance with that of

c-Si. The 3rd of August (corresponding to a sunny day) and the 17th August (overcast day)

FIG. 1. (a) Schematic of the DTU module used for outdoor performance monitoring showing individual layers in the cell

and (b) outdoor performance monitoring setup established in Bangor, Gwynedd, North Wales, which has a latitude and lon-

gitude of 53.2280� N 4.1280� W.

013111-3 N. Bristow and J. Kettle J. Renewable Sustainable Energy 7, 013111 (2015)

were selected and daily temperature and irradiance is shown in Figs. 2(a)–2(c), allowing us to

draw a comparison between the performance of c-Si and OPV modules under diffuse and direct

irradiation. Optical measurements indicate that the irradiance on the 17th August is

FIG. 2. Diurnal module temperatures of OPV and c-Si modules and ambient temperature for the (a) 3rd and (b) 17th

August 2013 and (c) diurnal irradiance on 3rd and 17th August 2013.

013111-4 N. Bristow and J. Kettle J. Renewable Sustainable Energy 7, 013111 (2015)

predominantly diffuse. As the days were reasonably close to one another, the effects of module

degradation are not significant. On the 3rd August, c-Si module temperature reaches 38.1 �C,

whilst the OPV reaches 32.7 �C, despite the ambient temperature only reaching 21.2 �C; indicat-

ing the maximum variation in temperature between cell and ambient (TDelta) is þ16.8 �C for

the c-Si and þ11.6 �C for the OPV. The OPV and c-Si modules were subjected to similar levels

of ambient cooling, so the data indicate that OPV modules are less likely to heat up under out-

door operation. The variation in heating of the modules is likely to be due to the c-Si absorbing

greater infrared radiation, as silicon absorbs up to 1100 nm.13

The following graphs [Figs. 3(a)–3(h)] show open circuit voltage (VOC), short circuit cur-

rent (ISC), FF, and PCE plotted for the two days. Considering first of all the data for the 3rd

August, this represents a day with much stronger direct irradiation and where daily irradiance

reaches 1000 W/m2. During the afternoon intermittent full sun was observed, leading to cloud

lensing affects. It can be seen that the value for ISC tracks the diurnal irradiance level almost

exactly for both OPV and c-Si modules, although some shading is seen in the c-Si in the early

morning (before 8:00 am). However, FF and VOC show some relative variation in output

between the c-Si and OPV modules. At sunrise, the FF of the c-Si module rises up very quickly

to the maximum value (0.65) and reduces at sunset without much variation during the day,

other than a small gradual decrease to 0.58 at 2:00 pm due to module temperature rise. Some

large spikes are present in the data for Figs. 3(c) and 3(d) due to mismatching with the irradi-

ance data caused by cloud lensing. Conversely, the FF and VOC of the OPV module rise more

slowly to their limiting values after sunrise, indicating that FF and VOC for the OPV modules

are more affected by low irradiance than for the c-Si module. The cloud lensing mismatch is

not as obvious with the OPV modules as the IV measurement sweep takes longer (�8 s) than

the c-Si module (�1 s).

Considering data from the 17th August, where the irradiation levels are much lower (and

mainly diffuse) and ambient temperatures are also slightly lower. As with data for the brighter

day (3rd August), the ISC of both the OPV and c-Si track the diurnal irradiance level very

closely and the shading effects seen on the 3rd are removed due to the higher proportion of dif-

fuse light. For FF and VOC, the c-Si module shows similar trends with the brighter day, rising

close to the maximum value early in the morning and staying reasonably flat thereafter until

sunset. Significantly though, the OPV shows quite large variation between the data obtained on

the 3rd and that obtained on the 17th August; the FF reaches a maximum value of 0.33, sub-

stantially lower than the 3rd August value, and the VOC does not appear to reach the maximum

value, reaching a limit of 3.49 V.

B. Effect of irradiance on OPV performance

The following graphs [Figs. 4(a)–4(d)] show the relationship of ISC, VOC, FF, and effi-

ciency to irradiance for both days (3rd and 17th August). As anticipated, ISC has a linear

FIG. 3. Diurnal c-Si and OPV performance from the 3rd August 2014 (sunny day) showing (a) VOC, (b) ISC, (c) fill factor,

and (d) PCE, with corresponding data from the 17th August (overcast day) showing (e) VOC, (f) ISC, (g) FF, and (h) PCE.

013111-5 N. Bristow and J. Kettle J. Renewable Sustainable Energy 7, 013111 (2015)

relationship to irradiance for both OPV and c-Si modules and the rate in change of ISC as a

function of irradiance is constant for both days. Likewise VOC shows expected behaviour with

a logarithmic relationship to irradiance for both c-Si and OPV modules. However, there is a

FIG. 4. The effect of irradiance on (a) ISC, (b) VOC, (c) FF, and (d) PCE for OPV and c-Si modules on the 3rd and 17th August.

013111-6 N. Bristow and J. Kettle J. Renewable Sustainable Energy 7, 013111 (2015)

clear difference in the low light behaviour of VOC and FF for the OPV and c-Si modules,

observed from the data obtained at low irradiance levels [Figs. 4(b) and 4(c)]. As irradiance is

increased, the c-Si modules reaches close to its limiting value for VOC at low levels of irradi-

ance and this is consistent for both sunny and cloudy days: 90% of its maximum VOC value is

reached at an irradiance of 64 W/m2. However, in the case of VOC for the OPV module it can

be seen that the curves for the two days do not overlay and compared to c-Si, VOC rises much

more slowly with increasing irradiance. Furthermore, on the diffuse day the rate is even slower

and maximum VOC is not reached. On the bright day VOC reaches 90% of its maximum value

at �200 W/m2, whereas on the diffuse day this level is reached at �400 W/m2. This poorer low

light behaviour of the OPV is also observed in the FF characteristics.

Figure 4(d) shows that the PCE dependence on irradiance is determined primarily by VOC

and FF, since ISC is linear with irradiance. As the OPV exhibits poor VOC and FF at lower irra-

diance, the overall PCE at these levels is also low and only reaches a peak at 600 W/m2. As the

low light performance of the c-Si is much better, the peak PCE value is reached at much lower

irradiances (150 W/m2).

Overall, the data from Figs. 3 and 4 suggest that the OPV module is not well suited for

performance under low irradiances, based on the current device architecture. The primary rea-

son for the poorer VOC and FF under low light conditions is the appearance of inflexion

behaviour in the IV curves. The majority of the improvement in VOC and FF with increasing

irradiance is as a result of the removal of the inflexion characteristic in the IV-curve. This

behaviour has been reported in various papers for indoor experiments and several theoretical

explanations have been devised to account for this behaviour.14,15 The most commonly

reported cause is attributed to an energy barrier, caused by a poor carrier transport in one of

the layers or interfaces, which prevents charge extraction from the device, leading to

decreased VOC and FF. These reports have shown the inflexion characteristics can be over-

come by exposing the cells to a combination of illumination, temperature, and with UV light

with wavelengths between 360 and 400 nm.16

The data in Figs. 4(a)–4(d) suggest OPV modules underperform c-Si modules at low irradi-

ance levels as the conditions affecting the photo-annealing (temperature, irradiance, and UV

index) at these times are low. The OPV only reaches its maximum PCE when irradiance and

temperature is at its highest (for example, at 11 am on 3rd August). Therefore, over the course

of a day, the OPV modules are subjected to a “temporal photo-annealing” effect; at early morn-

ing and late evening or other times of low irradiation, the photo-annealing effect shows low in-

tensity, limiting VOC, FF, and thus PCE from the cell, whilst on days such as 10 am–5 pm on

3rd August, the OPV module experiences a high intensity photo-anneal and reaches its maxi-

mum possible PCE. This supports the view of Lilliedal et al. who studied how various anneal-

ing processes affected the cell performance.16 The report showed a “low intensity” treatment

had a small impact on photovoltaic performance after treatment though a “high intensity” treat-

ment had a much stronger effect reducing the time taken for removal of the inflexion points in

the IV characteristics. However, our data indicates that when the modules are left under low

irradiance for a few minutes, the behaviour will re-emerge in the IV curves.

C. Effect of temperature on OPV performance

Figures 5(a)–5(d) show the effect of temperature on OPV performance in outdoor environ-

ments. One issue with using outdoor data to understand how variations in temperature affect

PV performance is that temperature often increases in line with irradiance and the change in

PV performance is often a combination of the two variations. Therefore, the data are filtered

for fixed irradiance ranges (600 6 20 W/m2 and 1000 6 20 W/m2) which allows for the effect

of temperature alone to be studied. Data are shown for a 10 day period so we assume no signif-

icant OPV degradation.

VOC as a function of temperature is shown in Fig. 5(a) and shows the least variation out of

the four performance parameters investigated. It appears to increase slightly with increasing

temperature, but the level of increase is relatively low. ISC is also shown to increase with

013111-7 N. Bristow and J. Kettle J. Renewable Sustainable Energy 7, 013111 (2015)

temperature [Fig. 5(b)]. In OPVs, charge carrier transport in polymers is governed by “carrier

hopping” and is therefore thermally assisted. Mobility and thus charge-carrier transport should

improve with increasing temperature, in agreement with indoor measurements.17 In addition,

temperature increases will lead to more efficient dissociation of electrons and holes. This posi-

tive temperature coefficient of ISC is observed also in c-Si, but here the main contribution is

due to thermally excited intrinsic charge carriers and narrowing of the semiconductor’s bandgap

with increasing temperature. The FF of the OPV also appears also to rise with temperature

[Fig. 5(c)] and this is consistent with other reports and has been attributed to a decrease in se-

ries resistance as shunt resistance remains relatively unchanged with temperature.17

Ordinarily, for temperature coefficient calculations, the PV module is first shaded to lower

the temperature to ambient conditions and as soon as the device is uncovered, it rises in tem-

perature, several I-V curves are acquired at different temperatures.18–20 This is impractical for

OPVs as the temperature variation [see Figs. 2(a) and 2(b)] in an OPV over the course of a day

is relatively low. However, the temperature coefficient can also be obtained through outdoor

performance monitoring. To obtain the temperature coefficient of the OPV and c-Si module,

data were fitted to the following equation:18–20

gG ¼ gref ½1� bref ðTG � Tref Þ�; (1)

where gG is efficiency at an elevated/reduced temperature, (TG), gref is efficiency at the refer-

ence temperature, (Tref ), and bref is temperature coefficient. Due to the strongly temperature-

dependent ISC and FF, the OPV possesses a positive temperature coefficient. Based on Eq. (1),

a temperature coefficient of þ0.007%/K can be extracted for the OPV device whereas a value

of �0.34%/K is obtained for the c-Si module at Bangor University (data not shown). Previous

temperature coefficient studies using the approach adopted by Refs. 18–20 have shown that the

power of c-Si PV modules decreases at a rate of �0.400%/K (Ref. 20) based on outdoor data,

so it shows this approach gives a reasonably close value for c-Si and gives some confidence in

the OPV temperature coefficient value. The absolute temperature coefficient of the OPV is low

compared to that of the c-Si, however the PCE at Standard test Conditions (STC) conditions is

also low (see Table I). Therefore, the temperature coefficient values have been normalised to

FIG. 5. Temperature dependence of (a) VOC, (b) ISC, (c) fill factor, and (d) PCE measured at fixed irradiance of 600 W/

m2 6 10 W/m2 and 1000 W/m2 6 10 W/m2. Linear fitted curves are applied to measure the temperature coefficient of the

module.

013111-8 N. Bristow and J. Kettle J. Renewable Sustainable Energy 7, 013111 (2015)

the individual performance at Standard Test Conditions, STC, (1000 W/m2, 25 �C) to compare

the relative changes seen in OPV modules and c-Si modules. The normalised increased in the

OPV is þ0.014/K whereas the c-Si sees a normalised decrease of �0.028/K (see Table I).

As a comparison with other thin film technologies, CdTe modules possess a measured tem-

perature coefficient of around �0.25%/K (Ref. 21) and a-Si shows a power temperature coeffi-

cient of up to approximately �0.17%/K (Ref. 22), so the OPV appears to be one of the few PV

technologies that possess a positive temperature coefficient.

D. Degradation of OPV modules

The graphs in Fig. 6 show how the performance parameters Isc, VOC, FF, and PCE change

with time. Two sets of data have been acquired; one set follows OPV degradation during the

summer months (July–September 2013) and the second set follows degradation of the same

type of modules during winter (January–April 2014). Data were selected for a fixed irradiance

(600 6 15 W/m2).

Considering first of all the summer data, this data is represented in Figs. 6(a)–6(d) with

data shown from one module (“DTU2”). Overall, 8 modules were tested and showed relatively

similar initial PV performance (PCE� 0.7% 6 0.2%) at 600 W/m2, demonstrating the high pre-

cision and control of modules developed in the freeOPV programme. Out of the 8 modules

tested, they all exhibited one of two degradation patterns. The first degradation pattern sees

TABLE I. Temperature coefficient of OPV and c-Si obtained from the outdoor monitoring campaigns and typical values of

other technologies obtained from literature.

Temperature

coefficient, bref (%/K)

PCE at STC

(1 sun, 25 �C)

Normalised temperature

coefficient (AU/K)a

OPV þ0.007 0.69 þ0.010

c-Si (Bangor) �0.341 12.89 �0.026

c-Si20b �0.400 18.20 �0.022

CdTe20,21 �0.25 No data

a-Si20b �0.165 No data

CIGs20 �0.320 No data

aData is normalised to the performance of each solar cell under STC (25 �C) conditions.bData is averaged from 2 (a-Si) and 9 (c-Si) modules.

FIG. 6. Change in (a) VOC, (b) Isc, (c) fill factor, and (d) PCE over time for two selected “average” modules at an irradiance

of 600( 6 15) W/m2 for summer 2013 (DTU2) and winter 2013-14(DTU35) demonstrating the increase lifetime observed

by cells.

013111-9 N. Bristow and J. Kettle J. Renewable Sustainable Energy 7, 013111 (2015)

steady decreases in all performance parameters (VOC, FF, and ISC), leading to an exponential

decay in PCE with time. The second degradation pattern [shown in Figures 6(a)–6(d)] showed

PCE initially relatively stable for the first 2 weeks and a sharper linear decrease thereafter. In

this case, the degradation falls mostly as a result in FF and ISC decreasing, as VOC remains rela-

tively steady for the initial 2 weeks. This degradation pattern was observed by half of all mod-

ules and showed better overall stability.

Overlaid on the summer degradation data are results from similar cells monitored over a

14 week period over the winter of early 2014. This allows the examination of differences in

degradation between summer and winter. It is clear that the module degradation is greatly

reduced owing to the reduced UV index, temperature and irradiance seen over this period. One

“hero” module exhibits only a 40% drop in PCE over the 14 week period. The average half-life

(t1=2) observed for winter modules was over 9 weeks, compared with a half-life of only 3 weeks

for the summer modules.

Overall, the outdoor lifetime of these cells was much lower than other technologies, This is

because the data obtained in this paper is from roll-to-roll manufactured OPVs and the substrate

used in this case was PET coated with e-beam evaporated SiO2, which acted as a barrier layer;

water vapour transmission rate are estimated at around 10�3 g m�2 d�1g/m2/day, which is

FIG. 7. Microscope photographs (magnification �2.5) showing the edge of an OPV module (a) before and (b) after outdoor

performance monitoring, demonstrating degradation at the module edges. Laser cutting is used to isolate modules, giving

rise to the burnt appearance at the edges.

013111-10 N. Bristow and J. Kettle J. Renewable Sustainable Energy 7, 013111 (2015)

insufficient for maintaining performance outdoors. In addition, the silicon dioxide barrier layer

was the only UV filter applied to these cells. Generally, the cells did not alter significantly in

appearance during ageing; the silver metal grid, electrical contacts, and active layer showing no

obvious changes in colour or appearance. One major source of degradation appeared at the

edge of devices [see Figures 7(a) and 7(b)]. It is very likely that oxygen and water can pene-

trated sideways through the device, leading to chemical degradation of the active layer, interfa-

cial layers, or electrodes.

IV. CONCLUSIONS

The outdoor dependence of temperature and illumination on OPV performance has been

analysed in this paper. The performance was benchmarked against c-Si modules and the OPVs

are shown to exhibit relatively poorer performance under low light conditions. The VOC is

shown to possess a logarithmic dependence on irradiation, but the low light performance of

VOC as well as FF is limited by inflexion points in the IV curves. These can be removed by

photo-annealing at higher irradiance; however, this effect has negative consequences for OPV

performance on overcast days. For the first time in literature, outdoor OPV evaluation has

yielded a temperature coefficient, which was found to be positive of þ0.007%/K was found,

which compared to that of �0.341%/K for c-Si. Overall, the cell degradation outdoors for these

modules appeared very severe and highlights the need for better barrier layers, UV filters and

edge sealants for this technology.

ACKNOWLEDGMENTS

The authors would particularly like to thank Morten Madsen and Frederik Krebs of the

Technical University of Denmark for their advice, support and for supplying the OPV modules

through the generous “freeOPV programme” (http://plasticphotovoltaics.org/free-opv). In addition,

the authors would like to thank the Wales Ireland Network for Innovative Photovoltaic

Technologies (WIN-IPT) project, funded through Interreg IVA, for financially supporting this work

and finally to Egnitec Ltd., Gwynedd, Wales, UK (www.egnitec.com) for supporting the work with

state of the art measurement facilities.

1L. Dou, J. You, Z. Hong, Z. Xu, G. Li, R. A. Street, and Y. Yang, “25th anniversary article: A decade of organic/poly-meric photovoltaic research,” Adv. Mater. 25(46), 6642–6671 (2013).

2J. You, L. Dou, K. Yoshimura, T. Kato, K. Ohya, T. Moriarty, K. Emery et al., “A polymer tandem solar cell with 10.6%power conversion efficiency,” Nat. Commun. 4, 1446 (2013).

3M. O. Reese, S. A. Gevorgyan, M. Jørgensen, E. Bundgaard, S. R. Kurtz, D. S. Ginley, D. C. Olson et al., “Consensus sta-bility testing protocols for organic photovoltaic materials and devices,” Sol. Energy Mater. Sol. Cells 95(5), 1253–1267(2011).

4T. Ishii, T. Takashima, and K. Otani, “Long-term performance degradation of various kinds of photovoltaic modulesunder moderate climatic conditions,” Prog. Photovoltaics: Res. Appl. 19(2), 170–179 (2011).

5M. Jørgensen, K. Norrman, and F. C. Krebs, “Stability/degradation of polymer solar cells,” Sol. Energy Mater. Sol. Cells92(7), 686–714 (2008).

6E. A. Katz, S. Gevorgyan, M. S. Orynbayev, and F. C. Krebs, “Out-door testing and long-term stability of plastic solarcells,” Eur. Phys. J.: Appl. Phys. 36(03), 307–311 (2006).

7J. A. Hauch, P. Schilinsky, S. A. Choulis, R. Childers, M. Biele, and C. J. Brabec, “Flexible organic P3HT: PCBM bulk-heterojunction modules with more than 1 year outdoor lifetime,” Sol. Energy Mater. Sol. Cells 92(7), 727–731 (2008).

8F. C. Krebs, “Air stable photovoltaic device,” U.S. patent application 12/918,282 (17 February, 2009).9F. C. Krebs, “Encapsulation of polymer photovoltaic prototypes,” Sol. Energy Mater. Sol. Cells 90(20), 3633–3643(2006).

10S. A. Gevorgyan, A. J. Medford, E. Bundgaard, S. B. Sapkota, H.-F. Schleiermacher, B. Zimmermann, U. W€urfel et al.,“An inter-laboratory stability study of roll-to-roll coated flexible polymer solar modules,” Sol. Energy Mater. Sol. Cells95(5), 1398–1416 (2011).

11R. R. Søndergaard, T. Makris, P. Lianos, A. Manor, E. A. Katz, W. Gong, S. M. Tuladhar et al., “The use of polyurethaneas encapsulating method for polymer solar cells—An inter laboratory study on outdoor stability in 8 countries,” Sol.Energy Mater. Sol. Cells 99, 292–300 (2012).

12F. C. Krebs, M. H€osel, M. Corazza, B. Roth, M. V. Madsen, S. A. Gevorgyan, R. R. Søndergaard, D. Karg, and M.Jørgensen, “Freely available OPV—The fast way to progress,” Energy Technol. 1(7), 378–381 (2013).

13A. V. Shah, J. Meier, E. Vallat-Sauvain, N. Wyrsch, U. Kroll, C. Droz, and U. Graf, “Material and solar cell research inmicrocrystalline silicon,” Sol. Energy Mater. Sol. Cells 78(1), 469–491 (2003).

14M. Zhang, H. Wang, and C. W. Tang, “Hole-transport limited S-shaped IV curves in planar heterojunction organic photo-voltaic cells,” Appl. Phys. Lett. 99(21), 213506 (2011).

013111-11 N. Bristow and J. Kettle J. Renewable Sustainable Energy 7, 013111 (2015)

15M. T. Lloyd, C. H. Peters, A. Garcia, I. V. Kauvar, J. J. Berry, M. O. Reese, M. D. McGehee, D. S. Ginley, and D. C.Olson, “Influence of the hole-transport layer on the initial behavior and lifetime of inverted organic photovoltaics,” Sol.Energy Mater. Sol. Cells 95(5), 1382–1388 (2011).

16M. R. Lilliedal, A. J. Medford, M. V. Madsen, K. Norrman, and F. C. Krebs, “The effect of post-processing treatments oninflection points in current–voltage curves of roll-to-roll processed polymer photovoltaics,” Sol. Energy Mater. Sol. Cells94(12), 2018–2031 (2010).

17I. Riedel, J. Parisi, V. Dyakonov, L. Lutsen, D. Vanderzande, and J. C. Hummelen, “Effect of temperature and illumina-tion on the electrical characteristics of polymer–fullerene bulk-heterojunction solar cells,” Adv. Funct. Mater. 14(1),38–44 (2004).

18K. Akhmad, A. Kitamura, F. Yamamoto, H. Okamoto, H. Takakura, and Y. Hamakawa, “Outdoor performance of amor-phous silicon and polycrystalline silicon PV modules,” Sol. Energy Mater. Sol. Cells 46(3), 209–218 (1997).

19D. L. King, J. A. Kratochvil, and W. E. Boyson, “Measuring solar spectral and angle-of-incidence effects on photovoltaicmodules and solar irradiance sensors,” in Conference Record of the Twenty-Sixth IEEE Photovoltaic SpecialistsConference (IEEE, 1997), pp. 1113–1116.

20G. Makrides, B. Zinsser, G. E. Georghiou, M. Schubert, and J. H. Werner, “Temperature behaviour of different photovol-taic systems installed in Cyprus and Germany,” Sol. Energy Mater. Sol. Cells 93(6), 1095–1099 (2009).

21K. Emery, J. Burdick, Y. Caiyem, D. Dunlavy, H. Field, B. Kroposki, and T. Moriarty et al., “Temperature dependenceof photovoltaic cells, modules and systems,” in Conference Record of the Twenty Fifth IEEE Photovoltaic SpecialistsConference (IEEE, 1996), pp. 1275–1278.

22S. Hegedus, “Thin film solar modules: The low cost, high throughput and versatile alternative to Si wafers,” Prog.Photovoltaics: Res. Appl. 14(5), 393–411 (2006).

013111-12 N. Bristow and J. Kettle J. Renewable Sustainable Energy 7, 013111 (2015)