Copyright WILEY‐VCH Verlag GmbH & Co. KGaA, 69469 Weinheim, Germany, 2014. 

Supporting Information   

for Adv. Mater., DOI: 10.1002/adma.201305196  

Universal Formation of Compositionally Graded Bulk Heterojunction for Efficiency Enhancement in Organic Photovoltaics Zhengguo Xiao, Yongbo Yuan, Bin Yang, Jeremy VanDerslice, Jihua Chen, Ondrej Dyck, Gerd Duscher, and Jinsong Huang*

1

Supporting Document

Universal Formation of Compositionally Graded Bulk Heterojunction for

Efficiency Enhancement in Organic Photovoltaics

Zhengguo Xiao1, Yongbo Yuan1, Bin Yang1, Jeremy VanDerslice1,2, Jihua Chen3 and Jinsong

Huang1*

1Department of Mechanical and Materials Engineering & Nebraska Center for Materials and Nanoscience,

University of Nebraska-Lincoln, Lincoln, NE 68588-0526

2J.A.Wollam Co., Inc, Lincoln, NE 68508

3Oak Ridge National Laboratory, Oak Ridge, Tennessee, 37831-6494;

* To whom correspondence should be addressed: jhuang2@unl.edu

1. Chemical structures of different polymer donors and fullerene-derivative acceptors.

Figure S1 Chemical structures of donors and acceptors

FTQ

R=2-EthylhexylPBDTTT-CT

[60]ICBA [70]PCBM

P3HTPCDTBT

2

The chemical structures of the chemicals used in this study are shown in Fig.S1: poly[4,8-

bis-(2-ethyl-hexyl-thiophene-5-yl)-benzo[1,2-b:4,5-b′] dithiophene-2,6-diyl]-alt-[2-(2′-ethyl-

hexanoyl)-thieno [3,4-b]thiophen-4,6-diyl (PBDTTT-CT), poly[N-9′-hepta-decanyl-2,7-

carbazole-alt-5,5-(4′,7′-di-2-thienyl-2′,1′,3′-enzothiadiazole)] (PCDTBT), Poly[6-fluoro-2,3-bis-

(3-octyloxyphenyl) quinoxaline-5,8-diyl-alt-thiophene-2,5-diyl] (FTQ), poly(3-hexylthiophene)

(P3HT), Phenyl-C71-Butyric-Acid-Methyl Ester ([70]PCBM), and indene-C60 bisadduct

([60]ICBA).

PBDTTT-CT was purchased from Solarmer Energy Inc (molecular weight: around 29

KDa), PCDTBT was provided by Konarka Technologies (molecular weight: approximately

100K), P3HT was purchased from Rieke Metals Inc. (molecular weight: approximately 50-70K),

FTQ was synthesized according the previous reported method[1]

(molecular weight: around 69

KDa), [60]ICBA was provided by Solaris Inc. and [70]PCBM was purchased from Nano Carbon

Inc. They were used as received.

2. Fabrication details of polymer: fullerene-derivative blend films

The details of Polymer: fullerene-derivative blend films preparing are summarized in Table S1.

Table S1 Different polymer: fullerene-derivative blend films fabrication details (DIO(1,8-

diiodooctane))

Materials Polymer and fullerene-

derivative weight ratio

(polymer concentration)

Working

solvent

DIO

volume

ratio

Spin

coating

parameter

Annealing

condition

PBDTTT-CT: 1:1.5 (10 mg/ml) DCB 3% 900 rpm No

3

[70]PCBM for 60 sec. annealing

PCDTBT:

[70]PCBM

1:4 ( 4 mg/ml) DCB:CB

(3:1 V/V)

0.4% 2400 rpm

for 14 sec.

No

annealing

P3HT:

[60]ICBA

1:1 (17.5 mg/ml) DCB 3% 800 rpm

for 20 sec.

150 °C

for 10 min

FTQ:

[70]PCBM

1:1 (15 mg/ml) DCB 0.4% 1000 rpm

for 50 sec.

110 °C

for 1 min

The PBDTTT-CT:[70]PCBM and PCDTBT:[70]PCBM solution were kept at 50 °C to

60 °C and 40 °C to 50 °C respectively. All other solutions were spin-coated at room temperature.

PBDTTT-CT:[70]PCBM, PCDTBT:[70]PCBM and FTQ:[70]PCBM were fast-dried, and

P3HT:[60]ICBA was slowly-dried with the films covered by a glass petri dish.

3. Ellipsometric characterization of the vertical composition profile in the blend films

Ellipsometry measures two parameters, Psi and Delta, which can be attributed to the

change in polarization that occurs when light interacts with materials. The change in

polarization can be described mathematically by the ratio of reflection amplitudes of p-polarized

light (rp) to s-polarized light (rs) as described in Eq 1. The ratio of reflection amplitudes contains

information about the dielectric function and thickness of materials as predicted by the Fresnel

reflection equations.

(1)

4

Mathematical models representing the optical response of the sample were built and compared to

the experimental data. The difference between the modeled and experimental data was quantified

by the Mean Squared Error (MSE), which represents the suitability of the model. The

Levenberg-Marquardt regression algorithm was used to vary select parameters in a manner that

provided the best fit to the experimental data. It is beneficial to reduce the complexity and

ambiguity involved in modeling by reducing the number of layers in the sample structure. For

the ellipsometric investigation, the PBDTTT-CT:[70]PCBM and PCDTBT:[70]PCBM blends

were deposited on a silicon wafer coated with poly(3,4-ethylenedioxythiophene):

poly(styrenesulfonate) (PEDOT:PSS) to maintain consistency with the working device while

reducing complexity.

PEDOT:PSS, PBDTTT-CT, and PCDTBT deposited using a spin-coat procedure are

typically uniaxial anisotropic materials with the c-axis oriented along the sample normal. To

enhance the sensitivity to the out-of-plane dielectric function of these materials, they were

deposited on thick thermal-oxide-coated silicon wafers. Further, to increase the uniqueness of

the solution, each sample was deposited on two or more thermal oxides where the oxide

thickness was varied between 300 nm-1000 nm. A multi-sample analysis approach was then

utilized to fit the measured data of each sample simultaneously. The optical properties of these

materials, shown in Fig. S2, were later held fixed when calculating the vertical phase separation

of the PBDTTT-CT:[70]PCBM and PCDTBT:[70]PCBM blends.

5

Figure S2 Optical properties for (a) PEDOT:PSS, (b) [70]PCBM, (c) PBDTTT-CT, (d) PCDTBT

determined individually as single layer films.

The optical model used to describe the measured data is shown in Fig. S3. The thickness of

the thermal oxide, PEDOT:PSS and the blend BHJ layers were allowed to vary during the fit

procedure. The compositional profile of the film was determined using a linear Bruggeman

Effective Medium Approximation where the percentage of polymer at the top and bottom of the

BHJ layer was allowed to vary to provide the best fit to the experimental data. The measured

ellipsometric data is shown in Fig. S4 in combination with the model results.

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0

0.5

1

1.5

2

2.5

0 500 1000 1500 2000

Exti

nct

ion

Co

eff

icie

nt,

k

Ind

ex

of

refr

acti

on

, n

Wavelength (nm)

nxnzkxkz

0

0.2

0.4

0.6

0.8

1

0

0.5

1

1.5

2

2.5

0 500 1000 1500 2000

Exti

nci

on

Co

eff

icie

nt,

k

Ind

ex

of

refr

acti

on

, n

Wavelength (nm)

nx, nz

kx, kz

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0

0.5

1

1.5

2

2.5

3

0 500 1000 1500 2000 Exti

nct

ion

Co

eff

icie

nt,

k

Ind

ex

of

refr

acti

on

, n

Wavelength (nm)

nxnzkxkz

0

0.2

0.4

0.6

0.8

1

1.2

0

0.5

1

1.5

2

2.5

0 500 1000 1500 2000 Exti

nct

ion

Co

eff

fici

en

t, k

Ind

ex

of

refr

acti

on

, n

Wavelength (nm)

nxnzkxkz

a b

c d

6

Figure S3 Models used to characterize (a) PBDTTT-CT:[70]PCBM and (b)

PCDTBT:[70]PCBM blends. The thickness (t) was allowed to vary along with the percentage of

the polymer in the blend at the top and bottom interface (%top, %bot, respectively).

Figure S4 Experimental data (solid lines) and model results (dashed lines) for (a) Vacuum-dried

PBDTTT-CT:[70]PCBM, (b) Fluxed PBDTTT-CT:[70]PCBM, (c) Vacuum-dried

PCDTBT:[70]PCBM, and (d) Fluxed PCDTBT:[70]PCBM

a b

Silicon Substrate

500 nm SiO2 (t)

60 nm PEDOT:PSS (t)

95 nm PCDTBT:PCBM (t, %top, %bot)

Silicon Substrate

300 nm SiO2 (t)

65 nm PEDOT:PSS (t)

72 nm PBDTTT:PCBM (t, %top,%bot)

0

5

10

15

20

25

30

35

40

45

200 700 1200 1700

Psi

Wavelegnth (nm)

55 deg

65 deg

75 deg0

5

10

15

20

25

30

35

40

45

200 700 1200 1700

Psi

Wavelength (nm)

55 deg

65 deg

75 deg

0

10

20

30

40

50

60

70

80

90

200 700 1200 1700

Psi

Wavelength (nm)

55 deg65 deg75 deg

0

10

20

30

40

50

60

70

80

90

200 700 1200 1700

Psi

Wavelength (nm)

55 deg65 deg75 deg

a b

c d

7

The modeling procedure utilizes fixed optical properties to determine the compositional

profile of the blend. This is required to avoid direct correlation between the optical properties of

the film and the compositional profile of the BHJ layer. In doing this, the optical properties of

the materials combined in the blend were assumed to be consistent with their corresponding pure

material. This is a reasonable assumption as noted by the relatively good fits to the experimental

data, however, slight variations between the modeled and experimental data might be attributed

to the variation in the polymer structure occur when forming the blended BHJ layers. A robust

model was built such that the compositional profile of the blended films could be uniquely

determined independent of the user-defined initial conditions provided to the regression

algorithm. The uniqueness of the model was estimated by varying the percentage of the polymer

at the air and anode interface while monitoring the fit quality. It is evident in Fig. S5 that the

polymer percentage at the air and anode interface provides the lowest MSE over a small

percentage range, which suggests that the solutions provided by the model are unique.

8

Figure S5 Uniqueness of fit parameters for (a) vacuum-dried PBDTTT-CT:[70]PCBM, (b)

fluxed PBDTTT-CT:[70]PCBM, (c) vacuum-dried PCDTBT:[70]PCBM, (d) fluxed

PCDTBT:[70]PCBM

4. Device performance with different time interval for solvent-fluxing

An appropriate time interval between the spin coating of the polymer:fullerene-derivative

blend films and methanol-fluxing is critical to increase the device performance. Fig. S6 shows

the photocurrents of different polymer:fullerene-derivatives with various time interval between

deposition of blend films and solvent-fluxing.

25

26

27

28

29

30

31

0 10 20 30 40

M S

E

% PCDTBT in Vacuum dried BHJ

Air Interface

Anode Interface

20

21

22

23

24

25

26

27

0 10 20 30 40

M S

E

%PCDTBT in Fluxed BHJ

Air Interface

Anode Interface

45

47

49

51

53

55

20 30 40 50 60 70

M S

E

% PBDTTT-CT in Vacuum dried BHJ

Air Interface

Anode Interface

39

41

43

45

47

49

51

53

5 25 45 65 85

M S

E

% PBDTTT-CT in Fluxed BHJ

Air Interface

Anode Interface

a b

c d

9

Figure S6 Photocurrents of PBDTTT-CT:[70]PCBM (a), PCDTBT:[70]PCBM (b),

P3HT:[60]ICBA (c), and FTQ:[70]PCBM (d) with different time intervals between deposition of

blend films and solvent-fluxing.

5. Verification of the versatile and universal of the solvent-fluxing method

0.0 0.2 0.4 0.6 0.8

-12

-10

-8

-6

-4

-2

0

PCDTBT:[70]PCBM

Vacuum-dried

1 min

10 min

30 min

Cu

rren

t d

en

sit

y (

mA

/cm

2)

Voltage (V)

0.0 0.2 0.4 0.6 0.8

-10

-8

-6

-4

-2

0

FTQ:[70]PCBM Vacuum-dried

1 min

10 min

30 min

Cu

rren

t d

en

sit

y (

mA

/cm

2)

Voltage (V)

0.0 0.2 0.4 0.6 0.8

-10

-8

-6

-4

-2

0P3HT:[60]ICBA

Vacuum-dried

10 min

30 min

60 min

Cu

rren

t d

en

sit

y (

mA

/cm

2)

Voltage (V)

0.0 0.2 0.4 0.6 0.8-18

-16

-14

-12

-10

-8

-6

-4

-2

0

Vacuum-dried

1 min

10 min

30 min

PBDTTT-CT:[70]PCBMC

urr

en

t d

en

sit

y (

mA

/cm

2)

Voltage (V)

ba

c d

0.0 0.2 0.4 0.6 0.8-18

-16

-14

-12

-10

-8

-6

-4

-2

0

Vacuum dried

Methanol fluxing

Ethanol fluxing

Isoprapanol fluxing

Cu

rren

t d

en

sit

y (

mA

/cm

2)

Voltage (V)

10

Figure S7 Photocurrents of PBDTTT-CT:[70]PCBM devices vacuum-dried and fluxed with

solvents of methanol, ethanol and isopropanol.

As shown in Fig. S7, all of the low boiling point solvents resulted in almost the same

device performance enhancement compared to the vacuum-dried device, which demonstrated the

versatile and universal application of the solvent-fluxing method.

References

[1] Y. Lu; Z. Xiao; Y. Yuan; H. Wu; Z. An; Y. Hou; C. Gao; J. Huang, J. Mater. Chem. C

2013, 1, 630.