Gaining Further Insight into the Effects of Thermal Annealing and Solvent Vapor

Annealing on Time Morphological Development and Degradation in Small Molecule

Solar Cells

Jie Min1,2*，Nusret S. Güldal2, Jie Guo1, Chao Fang1, Xuechen Jiao3, Huawei Hu3, Thomas

Heumüller2, Harald Ade3, Christoph J. Brabec2,4

1The Institute for Advanced Studies, Wuhan University, Wuhan 430072, China

2Institute of Materials for Electronics and Energy Technology (i-MEET), Friedrich-Alexander-University Erlangen-Nuremberg, Martensstraße 7, 91058 Erlangen, Germany

3Department of Physics and Organic and Carbon Electronics Laboratory (ORaCEL), North Carolina State University, Raleigh, North Carolina 27695, United States

4Bavarian Center for Applied Energy Research (ZAE Bayern), Haberstraße 2a, 91058 Erlangen, Germany

*E-mail: Min.Jie@whu.edu.cn (J. Min)

Electronic Supplementary Material (ESI) for Journal of Materials Chemistry A.This journal is © The Royal Society of Chemistry 2017

0.0 0.2 0.4 0.6 0.8 1.0-14

-12

-10

-8

-6

-4

-2

0

WO TA SVA

Curr

ent d

ensi

ty (m

A cm

-2)

Voltage (V)

Figure S1. Current density-voltage (J-V) characteristics of the relevant cells evaluated under illumination with a solar simulator.

Table S1. Photovoltaic parameters of DRCN5T:PC70BM BHJ solar cells fabricated under different processing conditions. Cells were tested in air without encapsulations, illumination intensity was 100 mW cm-2.

Processing conditions Voc [mV] Jsc [mA cm-2] FF [%] PCEmax[%]

WO 982 7.53 48.24 3.57

TA 932 12.27 58.81 6.73

SVA 952 12.58 55.45 6.64

400 500 600 700 8000,0

0,1

0,2

0,3

0,4

Abso

rban

ce (a

.u.)

Wavelength (nm)

WO TA SVA

Figure S2. Absorption spectrum of films without any treatments and with TA and SVA treatments, respectively.

WO RMS = 1.0 nm TA RMS = 0.8 nm SVA RMS = 1.4 nm

Figure S3. AFM images (5×5 μm2; 1×1 μm2) of DRCN5T:PC70BM BHJ layers without and with TA and SVA treatments.

0.1 1

109

1010

1011

Inte

nsity

(a.u

.)

Q vector [A-1]

WO TA SVA

Figure S4. Circular 1D GIWAXS profiles of DRCN5T:PC70BM films under various annealing conditions. The curves are vertically shifted for better comparison about PC70BM scattering peak around 1.3 Å-1.

0.01 0.10

10

20

30

40

50

60

Inte

nsity

q2 (a

.u.; 1

0-21 )

q (nm-1)

WO TA SVA

Figure S5. Lorentz-corrected and thickness normalized circular RSoXS profiles of DRCN5T:PC70BM films with different processing conditions. All data were takend under 284.2 Ev, which maximizes inter-domain scattering and eliminates fluorescence background.

WO TA SVA

(a) (b) (c)

Figure S6. Multi-peak fitting results of films with various processing conditions, implemented to all RSoXS profiles.

WO TA SVA(a) (b) (c)

Figure S7. Quantitative comparison between RSoXS and AFM images. The power spectral density (PSD) of AFM images is converted from FFT of the direct images, weighted by q2. The low-q peak from 270 eV RSoXS originates from mass-thickness variation, while the high q peak in 270 eV RSoXS correlated well with the dominant peak of 284.2 eV RSoXS. The PSD of AFM phase images matche well with 284.2 eV RSoXS, indicating that the phase images reflect the material contrast between donor-rich domains and acceptor-rich domains within the bulk of the films. The AFM height images exhibit large discrepancy with RSoXS results due to the fact that AFM height measurement is only sensitive the topography of films.

Figure S8. Schematic illustration of the experimental chamber for in situ measurements of thin film drying from the side view; the horizontal purple line on the sample represents the X-ray; the red line indicates the measurement sequence of PL and LS.

Figure S9. Schematic of fitting model used for the thickness calculation of swollen films.

A layer (dtotal) under a solvent vapor which is a good solvent for one or all components in the

layer would induce a swelling behavior. The vapor starts to be absorbed by the film surface,

and creates a ‘second’ layer (dblend+solvent) on the dry film (dblend). This solvent-soaked layer can

be modelled with effective medium approximation (EMA).[1] Three assumptions were made for

the thickness calculations: (1) the effect of re-evaporation is omitted. (2) Constituent molecules

are randomly distributed and are smaller than the wavelength. (3) It is assumed that the EMA

layer has high fraction of solvent molecules. Hence, the fraction of solvent molecules in the

EMA layer is taken as 90 v% and kept at this value for all calculations. The film optical

constants were determined by NIKA software,[2] which is a calculation method based on the

reflection and transmission behavior of a flat layer. Since SVA has induced changes in

crystallinity (i.e. change in refractive index), 2 sets of film optical constants were used in the

EMA and dblend layer modelling: (1) optical constants from the film without any treatment and

(2) optical constants from the film with SVA treatment. Hence, the optical constants were

changed accordingly for the thickness calculation when reflection spectrum of the layer has

indicated crystallinity changes. The optical constants for chloroform molecules used in EMA

layer were modelled with a Cauchy dispersion[3] due to their transparency.

400 500 600 700 800

0.00

0.05

0.10

0.15

0.20

0.25

0.30

Refle

ctio

n (a

.u.)

Wavelength (nm)

Fit Spectrum Reflectometry data

Si Wafer = 650 µm

SiO2 = 2 nm

Blend

EMA

400 500 600 700 800

0.00

0.05

0.10

0.15

0.20

0.25

0.30

Fit Spectrum Reflectometry data

Refle

ctio

n (a

.u.)

Wavelength (nm)

Si Wafer = 650 µm

SiO2 = 2 nm

Blend (dry layer, 111 nm)

400 500 600 700 800

0.00

0.05

0.10

0.15

0.20

0.25

0.30

Refle

ctio

n (a

.u.)

Wavelength (nm)

Fit Spectrum Reflectometry data

Si Wafer = 650 µm

SiO2 = 2 nm

EMA (blend + 5%CF, 116 nm)

400 500 600 700 8000.00

0.05

0.10

0.15

0.20

0.25

0.30

Refle

ctio

n (a

.u.)

Wavelength (nm)

Fit Spectrum Reflectometry data

Si Wafer = 650 µm

SiO2 = 2 nm

EMA (blend + 10%CF, 122 nm)

(a)

(b)

(c)

(d)

Figure S10. Schematic of four different fitting models, including (a) the bilayer model as mentioned above, (b) the single model with respect to the film layer without any chloroform, (c) the single model with respect to the completely swelled film layer with 5% chloroform after 60 seconds, and (d) the single model with respect to the completely swelled film layer with 10% chloroform after 60 seconds. While the right column presents the relevant fits of the reflection spectrums.

Here Figure S10a is the EMA model as mentioned in Figure S9. Absorption peak (right column)

fits much better as compared to the other three models. As shown in Figure S10b, optical

constants of the model without any chloroform in film layer does not fit to the measured

reflectometry data of SVA 60th second. Absorption peak is blue-shifted, while n and k data of

the dray layer do not fit to that of SVA treated layer. We also used another model, showing that

the DRCN5T:PC70BM layer is swelled completely with 5% chloroform after 60 seconds SVA.

Theoretically, a dry non-porous layer with 111 nm dry thickness would be 116 nm after soaking

5 vol.% chloroform. Further checking the corresponding graph shows that the absorption peak

fits good from the position, but the intensity is too different between fit spectrum and

reflectometry data (see Figure S10c). In addition, when the film layer is swelled completely

with 10 vol.% chloroform, the thicknesses of swelled layer changes from 111 nm to 122 nm

after soaking. Compared with the reflectometry data, the absorption peak from the fit starts to

red shift with this model (see Figure S10d). In short, we selected this bilayer model (Figure S9)

to analysis the in situ PL data, resulting from the better fitting curve as compared to the other

three models.

400 500 600 700 8000,00

0,05

0,10

0,15

0,20

0,25

550 600 650 700

Ref

lect

ion

[%]

Wavelength [nm]

Ref

lect

ion

[%]

Wavelength [nm]

0 s 10 s 20 s 30 s 40 s 50 s 60 s 120 s

Figure S11. The reflection spectra of TA film are plotted from 0 seconds till 120 seconds. The inset is a zoom to the wavelength range of 550-750 nm for a better clarity of the spectral changes.

400 500 600 700 8000,00

0,05

0,10

0,15

0,20

0,25

0,30

0,35

600 650 700

Ref

lect

ion

[%

]

Wavelength [nm]

Ref

lect

ion

[%]

Wavelength [nm]

0 s 60 s 90 s 100 s 120 s 150 s 180 s 210 s 240 s

Figure S12. The reflection spectra of SVA film are plotted from 0 seconds till 240 seconds. The inset is a zoom to the wavelength range of 575-725 nm for a better clarity of the spectral changes.

0.0 0.2 0.4 0.6 0.8 1.00

10

20

30

Mas

s (*1

03 a.u

.)

t /tmax

WO TA SVA

Figure S13. SIMS ion yield as a function of sputtering time for (a) WO sample, (b) TA sample and (c) SVA sample, the depth distribution of DRCN5T molecules is shown here. The blue dashed line is an indicator for the bottom interface between active layer and cathode.

0 1 2 3 4 5 6 7 8

100

101

102

103

104

Electron Only Devicee = cm2V-1s-1 Hole Only Device h = cm2V-1s-1

J (A

/m2 )

Vin (V)0 1 2 3 4 5 6 7

10-1

100

101

102

103

104

J (A

/m2 )

Vin (V)

Hole Only Device h = cm2V-1s-1

Electron Only Devicee = cm2V-1s-1

(a) (b)

Figure S14. The dark J-V characteristics of devices with TA and SVA treatments, including hole-only devices and electron-only devices, respectively. The inset mobility data are average values of each sample within 6 devices.

0 100 200 300 400 5000.80

0.85

0.90

0.95

1.00

norm

. Voc

Time (h)

TA SVA

0 100 200 300 400 5000.80

0.85

0.90

0.95

1.00

no

rm. J

sc

Time (h)

TA SVA

0 100 200 300 400 5000.80

0.85

0.90

0.95

1.00

norm

. FF

Time (h)

TA SVA

0 100 200 300 400 5000.65

0.70

0.75

0.80

0.85

0.90

0.95

1.00

norm

. PCE

Time (h)

TA SVA

(a)

(c) (d)

(b)

Figure S15. Changes of normalized (a) Voc, (b) Jsc, (c) FF and (d) PCE losses over illumination time for TA and SVA treated solar cells.

400 500 600 700 8000.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

No

rmal

ized

abs

orba

nce

(a.c

.)

Wavelength (nm)

TA--fresh TA--aged

400 500 600 700 8000.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

Norm

aliz

ed a

bsor

banc

e (a

.c.)

Wavelength (nm)

SVA--fresh SVA--aged

(a) (b)

Figure S16. Changes of normalized absorption spectra over illumination time for TA and SVA treated films.

0.0 0.2 0.4 0.6 0.8 1.0

101

102

103

104

105

Coun

ts

t /tmax

WO-Sample---Fresh SiO2 Si DRCN5T DRCN5T:PC70BM

0.0 0.2 0.4 0.6 0.8 1.0

101

102

103

104

105

WO-Sample---Aged SiO2 Si DRCN5T DRCN5T:PC70BM

Coun

ts

t /tmax

0.0 0.2 0.4 0.6 0.8 1.0

101

102

103

104

105

TA-Sample---Aged SiO2 Si DRCN5T DRCN5T:PC70BM

Coun

ts

t /tmax

0.0 0.2 0.4 0.6 0.8 1.0

101

102

103

104

105

SVA-Sample---Fresh SiO2 Si DRCN5T DRCN5T:PC70BM

Coun

ts

t /tmax0.0 0.2 0.4 0.6 0.8 1.0

101

102

103

104

105

SVA-Sample---Aged SiO2 Si DRCN5T DRCN5T:PC70BM

Coun

ts

t /tmax

(a)

0.0 0.2 0.4 0.6 0.8 1.0100

101

102

103

104

105

TA-Sample---Fresh SiO2 Si DRCN5T DRCN5T:PC70BM

Coun

ts

t /tmax

(b)

(c) (d)

(e) (f)

Figure S17. SIMS ion yield as a function of sputtering time for (a) fresh sample without any treatments, (b) aged sample without any treatments, (c) fresh sample with TA treatment, (d) aged sample with TA treatment, (e) fresh sample with SVA treatment and (f) aged sample with SVA

treatment.

0.0 0.2 0.4 0.6 0.8 1.00

5

10

15

20 WO-Sample Fresh Aged

Mas

s (*1

03 a.u

.)

t /tmax

Figure S18. SIMS ion yield as a function of sputtering time for (a) fresh and (b) aged samples without any treatments, the depth distribution of DRCN5T molecules is shown here. The blue dashed line is an indicator for the bottom interface between active layer and cathode.

References:

1. K. Hinrichs and K. J. Eichhorn: Ellipsometry of Functional Organic Surfaces and Films, 1st ed. (Springer-Verlag Berlin Heidelberg, 2014).

2. A. Campa: NIKA-Model for extracting refractive indices. 48th Int. Conf. Microelectron. Devices Mater. Work. Ceram. Microsystems (2012).

3. M. Losurdo and K. Hingerl: Ellipsometry at the Nanoscale, 1st ed. (Springer-Verlag Berlin Heidelberg, 2013).