This is an electronic reprint of the original article.This reprint may differ from the original in pagination and typographic detail.

Powered by TCPDF (www.tcpdf.org)

This material is protected by copyright and other intellectual property rights, and duplication or sale of all or part of any of the repository collections is not permitted, except that material may be duplicated by you for your research use or educational purposes in electronic or print form. You must obtain permission for any other use. Electronic or print copies may not be offered, whether for sale or otherwise to anyone who is not an authorised user.

Voroshilov, Pavel; Ovchinnikov, Victor; Papadimitratos, Alexios; Zakhidov, Anvar; Simovski,ConstantinLight Trapping Enhancement by Silver Nanoantennas in Organic Solar Cells

Published in:ACS Photonics

DOI:10.1021/acsphotonics.7b01459

Published: 10/04/2018

Document VersionPeer reviewed version

Please cite the original version:Voroshilov, P., Ovchinnikov, V., Papadimitratos, A., Zakhidov, A., & Simovski, C. (2018). Light TrappingEnhancement by Silver Nanoantennas in Organic Solar Cells. ACS Photonics, 5(5), 1767-1772.https://doi.org/10.1021/acsphotonics.7b01459

Light Trapping Enhan ement by Silver

Nanoantennas in Organi Solar Cells

Pavel M. Voroshilov,

∗,†,‡Vi tor Ov hinnikov,

‡Alexios Papadimitratos,

¶Anvar A.

Zakhidov,

†,§,¶and Constantin R. Simovski

†,‡

†ITMO University, Kronverkskiy pr. 49, 197101 St. Petersburg, Russia

‡Aalto University, P.O. Box 15500, 00076 Aalto, Finland

¶The University of Texas at Dallas, Ri hardson 75080, USA

§National University of S ien e and Te hnology MISiS, Mos ow 119049, Russia

E-mail: p.voroshilov�metalab.ifmo.ru

Abstra t

In this paper, we study experimentally the en-

han ement of parameters of a typi al organi

thin �lm solar ell (OSC) a hieved by an orig-

inally proposed light trapping stru ture (LTS).

Our LTS is an array of silver nanoantennas

supporting olle tive modes, what provide sub-

wavelength light enhan ement in the substrate.

Low losses in the metal result from the advanta-

geous opti al �eld distribution that is a hieved

in the broad range of wavelengths. We study

the enhan ement of all the main photovoltai

(PV) parameters observed in OSC with our

LTS, su h as power onversion e� ien y (PCE),

whi h is e�e tively in reased by 18%, �ll fa -

tor (FF), that an also be e�e ted positively

and short- ir uit urrent Isc improvement. Al-though the Ag nanoantenna o upies up to 40%

of the photoa tive OSC area, it provides the

nearly twi e enhan ed opti al absorption, re-

sulting in overall higher Isc and FF. Improve-

ments of the opti al and ele tri al design of

OSC ar hite tures are dis ussed for further en-

han ement of performan e.

Keywords

Metasurfa e, Nanoantennas, Small mole ules,

Organi solar ells, Light trapping, Domino-

modes

The photovoltai industry regularly requests

new s ienti� and te hni al solutions for the

su essful global ommer ialization of renew-

able energy.

1

Considerable e�orts have been

made to enhan e the e� ien y of thin �lm so-

lar ells whi h an �nd numerous appli ations

(see e.g. in

1�4

). There is a variety of strate-

gies to in rease the e� ien y of thin �lm solar

ells, whi h in lude material development,

5,6

in-

terfa e engineering,

7

improvements in fabri a-

tion and pro essing,

8

devi e ar hite ture mod-

ernization

9

and optimization of harge trans-

fer.

10

Photon management by light-trapping nanos-

tru tures represents one of the promising ways

to improve the performan e of su h solar ells,

whose e� ien y is limited due to insu� ient ab-

sorption of light in the a tive (PV) layer. Ab-

sorption is limited by PV layer, whi h thi kness

lays in subwavelength range at frequen ies of

operation. Organi solar ells of ex iton type

(we do not refer dye-sensitized and perovskite

solar ells as OSCs) have a fundamental limi-

tation on the thi kness of the PV layer � be-

low 100-150 nm � due to small di�usion length

of Frenkel's ex itons.

11

Opti al absorption of

organi PV materials is also low. Therefore,

the larger part of the in ident sunlight passes

through the PV layer. An anti-re�e ting oat-

ing does not resolve the issue of low absorption.

1

Even if the thi kness of the PV layer e�e tively

doubled by a polished rear ele trode it is insuf-

� ient. Moreover, many OSCs imply a trans-

parent ( ondu ting oxide) or absorbing (e.g. Al

foil) rear ele trode that does not ompletely re-

�e t light ba k to the a tive layer. In this situa-

tion, one has to apply LTSs, usually performed

as plasmoni nanostru tures.

12

In

13,14

we suggested a on ept of light trap-

ping employing a metasurfa e of metal nanoan-

tennas, whose resonan es over a very broad

band of wavelengths. These resonan es orre-

spond to olle tive os illations that keep even

for a perfe tly ondu ting metal. The �eld is

lo ally enhan ed in the gaps between the metal

elements and the losses in the metal are almost

absent. Su h advantageous distribution origins

from leaky modes

15

supported by the metasur-

fa e suggested in.

13

We all them leaky domino-

modes by analogy with guided domino-modes

whi h initially were revealed in far-infrared

16

and visible

17

ranges for arrays of metal nano-

bars. Leaky domino-modes were studied theo-

reti ally in

15

and experimentally in.

18

The aim of the present paper is to experimen-

tally demonstrate the enhan ement of a typi al

OSC by LTS with leaky domino-modes. Sin e

the pioneering work by Ching W. Tang

19

on

two-layer organi OSC based on a mole ular

donor-a eptor stru ture with a PCE of about

1%, the signi� ant progress has been a hieved

in the improvement of OSC systems, leading

to a PCE of over 11%

20

in single-jun tion and

over 13%

21

in tandem small-mole ule based de-

vi es with a polished rear ele trode. However,

the main drawba k inherent to typi al OSCs by

C.W. Tang keeps � the absorption of light is in-

su� ient even in these advan ed OSCs. There-

fore, high opti al losses still determine the om-

paratively low e� ien y of organi solar photo-

voltai s ompared to sili on photovoltai s.

Usually, in papers on LTSs, the in rease of op-

ti al absorption in the a tive layer is laimed as

a su� ient eviden e of the e� ien y enhan e-

ment be ause it results in the orresponding in-

rease of the short- ir uit urrent Isc.22

How-

ever, the total PCE of a solar ell is a�e ted

not only by Isc but also by other parameters

su h as FF and open ir uit voltage Voc:

PCE =Voc · Isc · FF

Pin

, (1)

where Pin is the power of the in ident light (re-

ferred to the surfa e of our thin �lm solar ell

and integrated over the spe trum). In a or-

dan e with Eq. 1 an LTS in order to grant the

enhan ement must not harm both FF and Voc,

otherwise, the overall e� ien y may de rease.

In this paper, we study the e�e t of proposed

LTS on all mentioned parameters. Our target

is the enhan ement of PCE.

An expli it example of an OSC that an be

drasti ally improved by our LTS is a small-

mole ule OSC

14

operating in the near-infrared

and transparent for the visible light. In that

work, we have hosen a bulk-heterojun tion

small-mole ule OSC with phthalo yanine as a

donor and fullerene as an a eptor. We pre-

ferred this design as the typi al one that is suf-

� iently simple in fabri ation.

In our theoreti al work,

14

we have shown

that proposed LTS in reased the absorption of

infrared solar light in the tin phthalo yanine

(SnP , donor) � fullerene (C60, a eptor) lay-

ers of the OSC more than triple. That OSC had

two indium tin oxide (ITO) ele trodes (both an-

ode and athode), that made it transparent for

visible light, though by a pri e of very low e�-

ien y.

3

Our LTS with its triple enhan ement of

the useful absorption promises a nearly double

in rease of the short- ir uit urrent and, per-

haps, a twi e higher PCE (though this ques-

tion was not studied). Meanwhile, the trans-

mittan e of visible light is kept su� ient for the

e�e tive transparen y of a window into whi h

su h the enhan ed OSC ould be in orporated.

For estimation of the e�e tive in rease of the

short- ir uit urrent, both proper opti al and

ele tri al simulations should be performed, as

in has been done e.g. in.

23

It is not easy to pre-

di t how both opti al absorption and photo ur-

rent an be optimally improved simultaneously.

So in reased opti al absorption (and number

of photogenerated arriers) an be a hieved by

in reasing photoa tive layer thi kness, however

this may lead to overall de rease of the expe ted

2

PV layersPV layers

ITO cathode Al cathode

0 1E field, arb. unit-

λ = 840 nmλ = 480 nm

0 1E-field, arb. unit

glass

ITO anode

CuPc

SnPc

C60

BCP

Al cathode

Ag

glass

ITO anode

CuPc

SnPc

C60

BCP

ITO cathode

Ag

a b c

x

z

yx

z

y

x

z z

y x y

z z

Transparent device

Opaque device

P1P 1P 2

P2

Figure 1: Normalized ele tri �eld intensity distributions in the horizontal planes P1 and P2 (top

panel) and in the slightly tilted ross-se tions (bottom panel) for an OSC ontaining ITO anode

and athode, λ =840 nm (a) and for a similar OSC with ITO anode and Al athode, λ =480 nm

( ). A tive layer boundaries are shown as white dotted lines (bottom panel). Ag nanoantennas are

lo ated on top of ITO anode, embedded into the PV layer. In panel (b) we show the OSC s hemati s

for these two ases. Horizontal planes P1 and P2 are shown as bla k dashed lines. Orange arrows

show the wave in iden e.

PCE enhan ement due to poor harge olle tion

from thi ker layer (due to e.g. short di�usion

length of arriers, and higher series resistan e

of thi ker layer) and thus overall de reased �ll

fa tor. Therefore in future, we will qualitatively

validate the numeri al model of nanoantennas

in orporated in a solar ell with the experimen-

tal results for further development of the e�-

ient light-trapping on ept for high PCE solar

ells.

Unfortunately, in the present work, we did not

manage to fabri ate a transparent ITO ath-

ode. Therefore, the experimental realization of

the exa tly same stru ture as in

14

turned out to

be impossible. We have repla ed ITO for the

athode by Al and dropping the fun tionality

of the transparen y on entrated on the e�e t

of light trapping and its impli ations. Sin e in

this ase the transparen y of the PV layer is

not required, we have deviated from the design

of

14

adding a nanolayer of opper phthalo ya-

nine (CuP ) as an additional donor and hole

transport layer (HTL). It allowed us to reate

a double heterojun tion and to in rease the ini-

tial e� ien y of the prototype. Our new design

is typi al and orresponds to works.

9,11

In or-

der to demonstrate the advantageous operation

of our LTS, we do not modify the design pa-

rameters of the prototype. In this meaning, our

study fairly demonstrates the real enhan ement

granted by the LTS to a typi al OSC.

The HTL of CuP has maximal absorption

in the visible region. Therefore, onversion of

solar radiation o urs in the extended spe tral

range, whi h in ludes both the near infrared

(750-900 nm) part where the absorbing mate-

rial is SnP and the visible part (550-750 nm),

where CuP absorbs. One more di�eren e of

the present design from that of

14

is the slight

shift of the metasurfa e with respe t to the ITO

anode. Instead of embedding the nanoantennas

into ITO, we lo ate them on the ITO surfa e

that allowed us to strongly simplify the fabri-

ation.

3

Glass

ITO

LTS

25 mm

2 mm

hITO

Glass

Dx

Dy

c

ab

e

z

xy

300 400 500 600 700 800 900 1000

Wavelength, nm

0

20

40

60

80

100

Tran

smit

tan

ce o

r re

flec

tan

ce, %

T (substrate)

T (LTS)

R (substrate)

R (LTS)

a b c

Figure 2: (a) Photo of the sample with our LTS. (b) S heme of the metasurfa e geometry. The

geometri al parameters are as follows: Dx = Dy = 1000 nm, a = 300 nm, b = 150 nm, c = 300

nm, e = 250 nm, h = 35 or 50 nm (two di�erent samples). ( ) Transmission and re�e tion spe tra

were measured for 50-nm sample (LTS) and for bare ITO area (substrate).

400 nm

Figure 3: S anning ele tron mi ros ope image

of our LTS.

Although the design modi� ations listed

above may seem insigni� ant, they an dras-

ti ally a�e t the performan e of the LTS. It

be omes ne essary to study the in�uen e of the

Al athode on the lo al ele tri �eld distribu-

tion inside the PV ell. We have performed

full-wave (CST) ele tromagneti simulations of

two PV devi es: one with a transparent ITO

athode and another with an opaque Al ath-

ode. Complex refra tive indexes were taken

from.

24�26

Ele tri �eld intensity distributions

for these two ases are presented in Fig. 1.

Indeed, the opaque metal �lm fully hanges

the hot spots arrangement in the bulk of the

stru ture at all wavelengths. Obviously, in

presen e of a transparent athode, an e�e tive

Fabry-Perot avity is formed in the stru ture

that works together with domino-modes ex-

ited in our nanoantennas and leads to the

more advantageous lo ation of hot spots in the

near-infrared range (700-850 nm). Meanwhile,

in this on�guration, these spots do not appear

at other wavelengths. The situation hanges

when Al repla es ITO as a athode material:

hot spots now also appear in the visible range

exa tly in favorable lo ations. However, in the

near-infrared band hot spots shift away from

the PV layer. Anyway, this does not result

in noti eable parasiti losses be ause these hot

spots do not interse t with the metal. In gen-

eral, simulations on�rmed our expe tations �

the enhan ement of the useful absorption keeps

after the repla ement of ITO by an aluminum.

More details on this modeling an be found in

the Supporting information.

We have prepared two samples with 50 nm

and 35 nm thi k Ag nanoantennas on top of

the ITO-glass substrate. These samples on-

tain 1.5 · 106 and 3 · 106 unit ells of our

LTS, respe tively. Fabri ation was done using

ele tron-beam lithography (EBL) and ele tron-

beam evaporation (see Methods and Support-

ing information). SEM analysis showed a very

high quality of our nanoantenna arrays, as an

be guessed from Figs. 2 and 3. We measured

transmittan e T and re�e tan e R for a 50-nm

thi k LTS (in iden e from the glass side) and

ompared them with those obtained for the bare

4

BCP [10 nm]C70 [55 nm]SnPc:C70 [30 nm]CuPc [15 nm]

Glass substrate

ITO anode

Al cathode50 nm

PV layers

Solar light

ITO(an

od

e)

Cu

Pc

SnPc

C70

BC

P

Al(cath

od

e)

-4.8- 34.

-5.2

-3.5-4.0

-5.1- 24.

-6.1

-3.0

-6.4

Cu

Pc

SnPc

C70

Active layers

Ener

gy

rela

tive

to

vac

uu

m le

vel,

eV

e-

h+

Device Isc [mA/cm2] Voc [V] FF [%] PCE [%]

W/o NAs 4.6 41 0.366 41.6 0.702

W/ NAs 4.955 0.276 42.9 0.587

Cu

rren

t (m

A/c

m)²

Without NAsWith NAs

Voltage (V)

a b c

Silver

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4

0

-1

-2

-3

-4

-5

Figure 4: (a) Cross-se tion s hemati s of the OSC with C70 as an a eptor layer and 50-nm thi k

silver nanoantennas. (b) Energy-level band diagram for the urrent on�guration. ( ) IV urves

measured under illumination for two ases: with (red line) and without (blue line) nanoantennas.

substrate (glass and ITO). The result is shown

in Fig. 2 ( ). The averaged re�e tan e growth

is about 10%. Noti e, that this negative e�e t

keeps for the OSC, where it is over ompensated

by the light-trapping e�e t.

The target of these simulations is twofold.

Firstly, we may estimate the in rease of the

parasiti absorption A introdu ed by our LTS.

Sin e the losses in Ag elements are negligibly

small, parasiti absorption an be attributed

to opti al losses in ITO. From energy balan e,

we have A = 1 − R − T . For integral absorp-

tion in the operational range λ =400-900 nm

we have nearly 5% in absen e of our LTS and

nearly 8% in its presen e. This is an a ept-

able harm. Se ondly, we may see that the re-

�e tan e grows mu h less that it an be esti-

mated using the geometri al opti s. Really, Ag

nanoantennas over almost 40% of the area, but

the re�e tan e is λ-dependent and mu h lower

than 0.4. This is so be ause the eigenmodes ofour LTS are ex ited by the normally in ident

plane wave, what makes geometri al opti s in-

adequate.

Further, inspe ting the opaque design de-

pi ted in Fig. 1 we de ided to additionally fab-

ri ate a variant of the OSC with C70 fullerene

derivative instead of C60. If the impa t of

our LTS will be more signi� ant for the stru -

ture with C70, it will on�rm the shift of light-

trapping regime (predi ted by our simulations)

from the IR range to the visible range for Al

athode. This should be so be ause C70 has

higher absorption than C60 below λ =700 nm,

whereas C60 has higher absorption at longer

waves.

Our �rst sample with 50 nm-thi k nanoan-

tennas has the following stru ture above the

glass substrate: ITO anode (150 nm) / CuP

(15 nm) / SnP :C70 (30 nm) / C70 (55 nm)

/ BCP (10 nm) / Al (70 nm). The or-

responding s hemati is shown in Fig. 4

(a). This on�guration with o-evaporated ph-

thalo yanine and fullerene leads to the for-

mation of two donor-a eptor heterojun tions:

CuP /C70 and SnP /C70. We illuminated our

sample with an AM 1.5G solar simulator from

the side of glass through a 1 mm aperture.

We measured the IV dependen ies of our sam-

ple illuminating either the 1 mm areas om-

prising our nanoantennas or 1 mm areas free

of them. In this way, we have olle ted su�-

ient statisti s for a reliable plotting two IV-

urves � one for an OSC enhan ed by our LTS

and another for a bare OSC (with only an-

tire�e tive oating). These urves are shown

in Fig. 4 and demonstrate a noti eable in-

rease of Isc aused by our LTS. Photo urrent

generated in the nanoantenna area rea hes the

value 5 mA/cm2, whereas the stru ture with-

out nanoantennas gives 4.6 mA/cm2. This 8%

in rease of the photo urrent orresponds to a

higher in rease of the PV absorban e due to a

nonlinear urrent response of su h OSC.

27

Due to 40% of the area overed by Ag nanoan-

tennas, the photoa tive area that photogener-

ates arriers and provides Isc is only 60% of the

similar area in a referen e OSC. Therefore the

5

a tual enhan ement of Isc, provided by the LTSis not 5 mA/cm2

, but 5/0.6 = 8.3 mA/cm2.

The Ag nanoantenna array o upied area is

not reating arriers but is reating on entra-

tion and enhan ement of the photoni ele tri-

al �eld whi h results in the nearly twi e en-

han ement of the opti al absorption of light,

resulting in 8.3 mA / 4.6 mA = 1.8 times en-

han ed on entration of arriers. We assume

here that the edges of sharp Ag stru tures do

not provide better harges olle tion (sin e the

nanoantennas an olle t only ele trons by low

work fun tion), but on the ontrary, are reat-

ing leakage of arriers and quen hing of ex i-

tons. Despite all those negative e�e ts of Ag

nanoantennas: shading the useful area, wrong

work fun tion, re ombination of ele trons and

holes on Ag nanostru tures, we have still ob-

served the overall sizable enhan ement of Isc =5 mA (8% more than in the referen e ell) at a

nearly twi e small photoa tive area.

Fill fa tor al ulated from IV hara teristi s

is improved by 2.6% for our LTS (see inset of

Fig. 4 ( )). Unfortunately, the open- ir uit

voltage Voc of the area with nanoantennas de-

reased by 0.09 V, i.e. 25% ompared to the

bare OSC. This harm overshadows the posi-

tive gain in photo urrent and negatively a�e ts

the overall PCE. The possible reason for the

fall in Voc ould be quite substantial dimen-

sions of metal nanoantennas ompared with ul-

trathin a tive layers. Really, Voc in OSCs is

dire tly proportional to the di�eren e between

the Fermi levels, and the ITO ele trode having

the bandgap in the ultraviolet has an impa t on

Voc. Silver elements having the ohmi onta t

with ITO may modify the e�e tive bandgap of

the ele trode. As a result, the di�eren e of the

Fermi levels squeezes and Voc be omes smaller.

Parameters of the OSC with nanoantennas

and without nanoantennas are shown in the in-

set of Fig. 4 ( ). These data are averaged over

several sample areas, as explained above.

Our next example is totally identi al to the

previous one ex lude repla ement of C70 by

C60 and thinner Ag elements (35 nm instead

of 50 nm). The stru ture is shown in Fig. 5

(a). It onsists of the following layers on the

glass plate: ITO (150 nm) / CuP (15 nm)

/ SnP :C60 (30 nm) / C60 (25 nm) / BCP

(10 nm) / Al (70 nm). Sin e C60 is less ab-

sorptive material in the visible range and its

thi kness de reased, the impa t of our LTS in

the photo urrent also de reased ompared to

the previous ase � 4.6% instead of 8%, as

an be seen from Fig. 5 ( ). However, thin-

ner metal elements improved FF by 19% versus

2.6%, whereas Voc almost kept un hanged. To-

tal averaged value of the PCE enhan ement is

18% that is a quite ompetitive result and is

higher than the gain provided by known plas-

moni ounterparts of our LTS for similar OSCs

(see e.g.

28

).

It should also be noted, that our samples were

not en apsulated. Therefore, in order to pre-

vent the degradation of organi materials and

oxidation of Ag, our measurements were per-

formed at room temperature and in an inert

atmosphere. In order to prevent Ag nanoanten-

nas tarnishing during transportation from the

lean room to the laboratory, where the solar

ell was prepared, we used a polymer prote -

tive oating. This oating was removed in the

laboratory with a etone. However, all samples

were treated by gentle UV-ozone for additional

leaning before the organi layer deposition.

29

So, a ertain oxidation of the thin silver ele-

ments might take pla e during this treatment.

This is another reason, why the in rease of the

PV absorption granted by our LTS in reality

turned out to be lower than that predi ted by

numeri al simulations. Hopefully, the experi-

mental enhan ement o�ered by our nanoanten-

nas will in rease on e more, if the fabri ation

fa ilities will allow lithography pro ess and so-

lar ell preparation in the same lean room.

To summarize, in this paper, we have the-

oreti ally and experimentally proved that our

original LTS from silver nanoantennas an suf-

� iently in rease the e� ien y of OSCs. First

of all, our numeri al simulations have shown ad-

vantageous �eld distributions that predi ted en-

han ement of the useful absorption. Se ondly,

we have fabri ated several OSCs with di�erent

materials and measured gain in PV parameters

aused by nanostru tures. Sin e total PCE of

a solar ell is a�e ted not only by Isc, but alsoby �ll fa tor FF and open ir uit voltage (Voc),

6

Solar light

ITO(an

od

e)

Cu

Pc

SnPc

C06

BC

P

Al(cath

od

e)

-4.8- 34.

-5.2

-3.5-4.0

-5.1 - 54.

- 26.

-3.0

-6.4

Cu

Pc

SnPc

C06

Active layers

Ener

gy

rela

tive

to

vac

uu

m le

vel,

eV

e-

h+

Cu

rren

t (m

A/c

m)²

Voltage (V)

a b c

Silver

BCP [10 nm]C60 [25 nm]SnPc:C60 [30 nm]CuPc [15 nm]

Glass substrate

ITO anode

Al cathode35 nm

PV layers

Without NAsWith NAs

Device Isc [mA/cm2] Voc [V] FF [%] PCE [%]

W/o NAs 2 869. 0.396 36 5. 0.415

W/ NAs 2 997. 0. 763 4 .3 4 0.489

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45

0

-0.5

-1

-1.5

-2

-2.5

-3

-3.5

Figure 5: (a) Cross-se tion s hemati s of the OSC with C60 as an a eptor layer and 35-nm thi k

silver nanoantennas. (b) Energy-level band diagram for the urrent on�guration. ( ) IV urves

measured under illumination for two ases: with (red line) and without (blue line) nanoantennas.

we have onsidered several design solutions and

dis ussed the in�uen e of the devi e ar hite -

ture on LTS performan e. Experimental data

for our OSCs based on C60 fullerene and en-

han ed by 35 nm thi k nanoantennas show the

enhan ement 18% that qualitatively �ts the re-

sults of simulations. Other experimental data

are also in line with theoreti al expe tations.

We believe that the present results are helpful

for further development of LTSs for thin �lm

solar ells. In our future designs, we will hoose

nanoantennas with proper work fun tion on-

tributing into olle tion of holes at ITO side

and with better lo ation allowing us to avoid

the 40% shading of the PV layer.

Methods

E-beam exposurePolymer spin-coating Al sputteringSubstrate cleaning

Al etchingDevelopmentAg evaporationLift-off

Glass

ITO

MMA

PMMA

Al

Ag

Figure 6: Pro ess �ow diagram illustrating fab-

ri ation steps of LTS.

Several samples with silver nanoantennas

were fabri ated on top of 150 nm-thi k ITO

oated glass substrates pur hased from Lumi-

nes en e Te hnology Corp by EBL and lift-o�

pro ess in the Mi ronova Nanofabri ation Cen-

tre of Aalto University. Pro ess �ow diagram is

shown in Fig. 6. Substrates were preliminary

leaned with a etone and isopropanol (IPA) in

an ultrasoni bath at 40

◦C and rinsed in DI-

water. A opolymer resist (MMA) with on-

entration of solids 11% in Ethyl La tate and

polymer resist (950PMMA A) with on entra-

tions of solids 2.25% in Anisole (both from Mi-

ro hem) were sequentially deposited by spin-

oating. Samples were dried on a hot plate

for 3 minutes at a temperature of 160 degrees

after ea h spin- oating pro edure resulting in

formation of 400 and 100 nm layers of MMA

and PMMA, respe tively. A thin (20 nm) sa -

ri� ial layer of aluminum was sputtered. The

purpose of this layer is to solve the problem

with e-beam fo using on a diele tri layer dur-

ing EBL. E-beam exposure was done using Vis-

te EBPG 5000Plus ES tool at 100 keV with

4 nA urrent beam and 1000 µC/cm2e-beam

dose. Then, samples were pro essed in et hing

mixture PWS 80-16-4 (65) from Honeywell on-

taining phosphori a id and nitri a id to om-

pletely remove the sa ri� ial aluminum layer.

Next, PMMA and MMA were developed in so-

lution of 1:3 MIBK:IPA during 30 se and then

rinsed in IPA during 30 se to terminate de-

velopment and prevent s umming. Silver was

deposited by using ele tron beam evaporation

at a rate of 2 Å/s and operating pressure of

10−7mbar. Finally, non-exposed areas of poly-

7

mer layers were washed out together with sil-

ver on their surfa e by a etone in the ultra-

soni bath. Samples were hara terized by us-

ing AFM (DI3100), SEM (Zeiss Supra 40), and

multifun tional tool for re�e tion (R), trans-

mission (T) and quantum e� ien y (QE) mea-

surements (QEX10).

Small mole ule OSCs were prepared in the

multi-sour e high-va uum system (Angstrom

Engineering In ., Canada) by sequential ther-

mal evaporation onto ITO-glass substrates with

our nanostru tures of the following �lms: a

15 nm thi k opper(II) phthalo yanine (from

H.W. Sanders Corp) HTL, a 30 nm thi k o-

evaporated in equal quantity SnP :fullerene

mixed D:A layer, an additional ontinuous

fullerene layer (25 or 55 nm), BCP (10 nm)

and Al (70 nm) layers. We used C60 and

C70 fullerenes (both >98% from Nano-C). Base

pressure of the hamber was kept around 10−8

mbar during evaporation. Evaporation rate

for separate layers was 0.5 Å/s, while o-

evaporation rate was equal to 0.25 Å/s for ea h

parti ular material. OSCs were hara terized

with an AM 1.5G solar simulator alibrated to

one sun (100 mW/cm2) through aperture holes

of 1 mm in diameter, whi h provided sele tive

illumination of nanostru tured regions only.

A knowledgement This work was par-

tially supported by the Ministry of Edu ation

and S ien e of the Russian Federation (Grant

14.Y26.31.0010) and Nokia Foundation. The

authors thank Mikhail Omelyanovi h, Kristina

Ian henkova, and Matthias Mes hke for useful

dis ussions. We also a knowledge the provision

of fa ilities and te hni al support by Aalto Uni-

versity at Mi ronova Nanofabri ation Centre.

Supporting information

The Supporting Information (SI) is available

free of harge on the ACS Publi ation website.

The SI in ludes details of 3D ele tromagneti

modeling and additional simulated ele tri �eld

distributions at di�erent wavelengths. More-

over, the SI ontains omplementary SEM and

AFM analysis of samples.

Referen es

(1) Bosio, A.; Romeo, A. Thin Film Solar

Cells: Current Status and Future Trends;

Energy s ien e, engineering and te hnol-

ogy series; Nova S ien e Publishers, 2011.

(2) S hmidt-Mende, L.; Wei kert, J. Organi

and Hybrid Solar Cells: An Introdu tion;

De Gruyter Textbook; De Gruyter, 2016.

(3) Lunt, R. R.; Bulovi , V. Transparent,

near-infrared organi photovoltai solar

ells for window and energy-s avenging

appli ations. Applied Physi s Letters

2011, 98, 61.

(4) Lipomi, D. J.; Bao, Z. Stret hable, elasti

materials and devi es for solar energy on-

version. Energy & Environmental S ien e

2011, 4, 3314�3328.

(5) Sun, S.-S.; Sari ift i, N. S. Organi

photovoltai s: me hanisms, materials,

and devi es; CRC press, 2005.

(6) Brédas, J.-L.; Norton, J. E.; Cornil, J.;

Corop eanu, V. Mole ular understanding

of organi solar ells: the hallenges.

A ounts of hemi al resear h 2009, 42,

1691�1699.

(7) Steim, R.; Kogler, F. R.; Brabe , C. J.

Interfa e materials for organi solar ells.

Journal of Materials Chemistry 2010, 20,

2499�2512.

(8) Krebs, F. C. Fabri ation and pro essing

of polymer solar ells: a review of print-

ing and oating te hniques. Solar energy

materials and solar ells 2009, 93, 394�

412.

(9) Heremans, P.; Cheyns, D.; Rand, B. P.

Strategies for in reasing the e� ien y of

heterojun tion organi solar ells: ma-

terial sele tion and devi e ar hite ture.

A ounts of hemi al resear h 2009, 42,

1740�1747.

(10) Deibel, C.; Strobel, T.; Dyakonov, V. Role

of the harge transfer state in organi

8

donor�a eptor solar ells. Advan ed

Materials 2010, 22, 4097�4111.

(11) Riede, M.; Mueller, T.; Tress, W.; S huep-

pel, R.; Leo, K. Small-mole ule solar ells

� status and perspe tives. Nanote hnology

2008, 19, 424001.

(12) Atwater, H. A.; Polman, A. Plasmoni s

for improved photovoltai devi es. Nature

materials 2010, 9, 205�213.

(13) Simovski, C.; Morits, D.; Voroshilov, P.;

Guzhva, M.; Belov, P.; Kivshar, Y.

Enhan ed e� ien y of light-trapping

nanoantenna arrays for thin-�lm solar

ells. Opti s express 2013, 21, A714�

A725.

(14) Voroshilov, P. M.; Simovski, C. R.;

Belov, P. A. Nanoantennas for enhan ed

light trapping in transparent organi solar

ells. Journal of Modern Opti s 2014, 61,

1743�1748.

(15) Voroshilov, P. M.; Simovski, C. R.

Leaky domino-modes in regu-

lar arrays of substantially thi k

metal nanostrips. Photoni s and

Nanostru tures-Fundamentals and

Appli ations 2016, 20, 18�30.

(16) Martin-Cano, D.; Nesterov, M.;

Fernandez-Dominguez, A.; Gar ia-

Vidal, F.; Martin-Moreno, L.; Moreno, E.

Plasmons for subwavelength terahertz

ir uitry. Opti s Express 2010, 18,

754�764.

(17) Simovski, C.; Luukkonen, O. Tapered

plasmoni waveguides with e� ient and

broadband �eld transmission. Opti s

Communi ations 2012, 285, 3397�3402.

(18) Sinev, I. S.; Voroshilov, P. M.;

Mukhin, I. S.; Denisyuk, A. I.;

Guzhva, M. E.; Samusev, A. K.;

Belov, P. A.; Simovski, C. R. Demon-

stration of unusual nanoantenna array

modes through dire t re onstru tion of

the near-�eld signal. Nanos ale 2015, 7,

765�770.

(19) Tang, C. W. Two-layer organi photo-

voltai ell. Applied Physi s Letters 1986,

48, 183�185.

(20) Deng, D.; Zhang, Y.; Zhang, J.; Wang, Z.;

Zhu, L.; Fang, J.; Xia, B.; Wang, Z.;

Lu, K.; Ma, W.; Wei, Z. Fluorination-

enabled optimal morphology leads to

over 11% e� ien y for inverted small-

mole ule organi solar ells. Nature

ommuni ations 2016, 7, 13740.

(21) Heliatek sets new Organi Photo-

voltai world re ord e� ien y of 13.2%.

2016; http://www.heliatek. om/

en/press/press-releases/details/

heliatek-sets-new-organi -photovoltai -world-re ord-effi ien y-of-13-2.

(22) Lasnier, F. Photovoltai engineering

handbook; CRC Press, 1990.

(23) Mirsafaei, M.; Fallahpour, A. H.; Lugli, P.;

Rubahn, H.-G.; Adam, J.; Madsen, M.

The in�uen e of ele tri al e�e ts on de-

vi e performan e of organi solar ells

with nano-stru tured ele trodes. S ienti�

reports 2017, 7, 5300.

(24) O'Connor, B. Organi Ele troni s on

Fibers for Energy Conversion Appli a-

tions. Ph.D. thesis, University of Mi hi-

gan, 2009.

(25) El-Nahass, M.; Yaghmour, S. E�e t of an-

nealing temperature on the opti al prop-

erties of thermally evaporated tin ph-

thalo yanine thin �lms. Applied Surfa e

S ien e 2008, 255, 1631 � 1636.

(26) Rand, B. P.; Li, J.; Xue, J.; Holmes, R. J.;

Thompson, M. E.; Forrest, S. R. Organi

Double-Heterostru ture Photovoltai

Cells Employing Thi k Tris (a etyla et-

onato) ruthenium (III) Ex iton-Blo king

Layers. Advan ed Materials 2005, 17,

2714�2718.

(27) Koster, L.; Mihailet hi, V.; Xie, H.;

Blom, P. Origin of the light intensity

dependen e of the short- ir uit urrent

of polymer/fullerene solar ells. Applied

Physi s Letters 2005, 87, 203502.

9

(28) Ahn, S.; Rourke, D.; Park, W. Plas-

moni nanostru tures for organi photo-

voltai devi es. Journal of Opti s 2016,

18, 033001.

(29) Chen, C.-W.; Hsieh, P.-Y.; Chiang, H.-H.;

Lin, C.-L.; Wu, H.-M.; Wu, C.-C. Top-

emitting organi light-emitting devi es us-

ing surfa e-modi�ed Ag anode. Applied

physi s letters 2003, 83, 5127�5129.

10