We demonstrate that a single-walled carbon nanotube network noncovalently coupled with a pyrene-modified azo-benzene chromophore functions as a host matrix for a broad range of photo-orientation and photomechanical effects. The chromophore could be efficiently reoriented through repeated trans-cis-trans isomerization under linearly polarized 480 nm light, with Δn of 0.012 at 650 nm and fast characteristic rise-times of 0.12 s. Erasable phase diffraction gratings could also be written, with permanent surface relief gratings forming at sufficiently long irradiation times. In addition to demonstrating a mechanism for photo-manipulation of single-walled carbon nanotubes, these results show photo-orientation of chromophores in azo-functionalized single-walled carbon nanotube networks as a path towards the photosensitive tuning of the electrostatic environment of the nanotube.

Hybrid materials composed of nanocarbons, such as two-dimensional graphene or one-dimensional nanotubes with organic chromophores and/or polymers, have emerged as light-responsive materials for next-generation optoelectronics. Initial research has focused primarily on electronic phenomena such as doping effects1–4 and changes in electronic structure upon hybridization, with purely optical effects remaining relatively unexplored. To date, the optical studies have focused on multi-walled nanotube (MWNT) or single-walled carbon nanotube (SWCNT) as minority additives to a majority host consisting of organic optical materials such as azobenzene-doped polymers,5 liquid crystals (LC),6 and functionalized azopolymers.7 These studies have shown enhanced light-induced chromophore ordering and subsequently faster and higher-contrast photoinduced birefringence. The enhancement from nanotube additives has been attributed to a range of factors such as increased π-π stacking of the LC molecules on the nanotubes, to increased availability of free volume for cis-trans isomerization in azo-polymers wrapped on nanotubes.

Here, we describe a material system where nanotubes serve as hosts to dock light-switchable molecules. We define this system as “functionalized nanocarbon solid film” as the nanotubes here are not minority components but rather are hosts for the chromophores. We demonstrate a reversible, photoinduced birefringence in a 200 nm thick film consisting of SWCNTs functionalized with pseudostilbene azo chromophores. Photoinduced birefringence Δn ≈ 0.012 at 650 nm was reached using low intensity 480 nm light and followed biexponential dynamics, with characteristic rise times as fast as 0.12 s. Illuminating the film with 480 nm intensity and polarization gratings resulted in dynamic phase diffraction gratings. A permanent periodic surface modulation of the film was also observed at longer irradiation times, providing direct evidence that azobenzene photomechanical effects can be used to spatially redistribute SWCNTs. Independent of photonics applications linked to photoinduced birefringence, the observed coupling of azobenzene photomechanics directly to the SWCNT is a unique advantage as it provides wavelength and polarization sensitive functionality for spatially addressing and manipulating SWCNTs.

The chromophore in our study is a pseudostilbene Disperse Red 1 (DR1), modified with a pyrene tether (DR1P) that couples with SWCNT; the resulting hybrid nanomaterial is referred to as DR1P-SWCNT. This and other similar azo-functionalized SWCNT systems have been utilized in optoelectronic applications8–11 such as photodetection,12,13 light-gated transistors,11,14,15 and energy-storage.16 The optical functionality arises from the reversible photoisomerization of the chromophore between its trans and cis states. The trans-cis photostationary state dynamics serves as a sensitive probe of the local chromophore microenvironment or as a photoswitch that can tune the electrostatic environment of the SWCNT11,13 via the large change in chromophore dipole moment and orientation that occurs on repeated photoisomerization.

The essential molecular feature that enables photoinduced birefringence in the DR1P-SWCNT system is the pyrene tether, which forms a noncovalent flexible link that couples the chromophore to the SWCNT via π-π stacking interactions. This particular interaction promotes stable chromophore coverage on the nanotube, while also avoiding the harsh electronic perturbations of the SWCNT electronic states17 that occur with covalent coupling schemes.18 Upon illumination of DR1P-SWCNT with linearly polarized light, the trans-cis-trans isomerization of the chromophore eventually rotates the trans isomer perpendicular to the field vector of the polarized light, as shown schematically in Fig. 1. Experimentally, this effect is manifested as an isotropic to anisotropic transition in linear optical properties, easily detected through photoinduced birefringence experiments. There is considerable interest in this phenomenon,19–21 as light-induced manipulation of chromophore order is central to holographic applications,22–24 in which dynamic diffraction gratings are optically written with intensity and/or polarization gratings. More recently, it has also been observed in the context of photomechanics,25,26 in which the reversible shape change of the chromophore upon photoisomerization is exploited to effect surface mass transport. However, neither photoinduced birefringence nor the associated phenomena of surface mass transport have been observed in non-polymeric, nanocarbon solid films.

FIG. 1.

(a) Chromophore DR1 (red) with pyrene tether (green) noncovalently linked to SWCNT. At 25 °C and with no external illumination, chromophore is predominantly in trans configuration and is oriented randomly. (b) Illumination of DR1P-SWCNT with linearly polarized 480 nm light drives trans→cis→trans isomerization. (c) Flexible pyrene tether facilitates light-driven reorientation of chromophore to be perpendicular to 480 nm polarization. (d) Bulk solid-film sample exhibits optical anisotropy (nx ≠ ny) due to oriented chromophores.

FIG. 1.

(a) Chromophore DR1 (red) with pyrene tether (green) noncovalently linked to SWCNT. At 25 °C and with no external illumination, chromophore is predominantly in trans configuration and is oriented randomly. (b) Illumination of DR1P-SWCNT with linearly polarized 480 nm light drives trans→cis→trans isomerization. (c) Flexible pyrene tether facilitates light-driven reorientation of chromophore to be perpendicular to 480 nm polarization. (d) Bulk solid-film sample exhibits optical anisotropy (nx ≠ ny) due to oriented chromophores.

Close modal

A typical DR1P-SWCNT sample was fabricated by adding DR1P to a SWNCT liquid suspension and drop casting onto glass slides to a typical circular sample size of 1.5 cm diameter. Detailed information can be found in the supplementary material. Atomic force microscopy (AFM) of the resulting solid films showed thickness of 200–230 nm and a roughness of ∼2–3 nm, while scanning electron microscopy revealed a well-dispersed network of SWCNT as shown in the supplementary Fig. S1.27 Raman spectroscopy confirmed the binding of the DR1P to the SWCNT, with approximately 2–3 DR1P chromophores per 100 C atoms as measured by X-ray photoelectron spectroscopy.17 DR1P-SWCNT film showed λmax near 490 nm and is essentially transparent in the red (Fig. S2).27 

To study the trans-cis dynamics, the films were irradiated with unpolarized 480 nm laser light while being probed with a weaker 543 nm laser. The probe was an unpolarized green HeNe laser, with output intensity stable to within 1%. The intensity of the 480 nm pump ranged from 4.0 mW/cm2 to 30.0 mW/cm2. Assuming the initially unexposed DR1P is in the trans state, pumping near λmax isomerizes a fraction of the trans population to the cis conformation. For pseudostilbenes, such as DR1P, there is spectral overlap between the cis and trans isomers, with slightly differing λmax and more significant differences in absolute absorbance. Steady 480 nm illumination will therefore generate a photostationary state with the cis/trans ratio determined by the 480 nm intensity and the relative cis/trans absorbance.

Typical results are shown in Fig. 2. The transmission of the film increases as the photoconversion of trans to cis proceeds under constant 480 nm illumination, due to the weaker absorbance of the cis isomer at 543 nm. This cis-trans interconversion is driven by the substantial overlap in the absorption spectrum of the two isomers as well as the thermal back-isomerization with a rate that depends on the temperature and microenvironment of the cis isomer. Upon removal of the 480 nm pump, the cis-trans back isomerization proceeds thermally, returning the system to its initial trans population. Beer's law provides a straightforward interpretation of the change in 543 nm transmission, resulting from 480 nm irradiation. Specifically, ln(Ti − Tss) is proportional to the cis population, where Ti and Tss are the initial and steady state 543 nm transmitted signals, respectively. Fig. 2(b) shows ln(Ti − Tss) vs the 480 nm pump intensity, demonstrating that the cis population in a given photostationary state scales with the 480 nm pump intensity used to reach that state.

FIG. 2.

(a) DR1P-SWCNT thin film transmittance of 543 nm probe laser (unpolarized) as 480 nm pump laser (unpolarized) is switched on (blue shading) and off. (b) Magnitude of ln(Ti − Tss) vs. 480 nm pump intensity, where Ti and Tss refer to the initial and steady state transmittance of DR1P-SWCNT film at 543 nm.

FIG. 2.

(a) DR1P-SWCNT thin film transmittance of 543 nm probe laser (unpolarized) as 480 nm pump laser (unpolarized) is switched on (blue shading) and off. (b) Magnitude of ln(Ti − Tss) vs. 480 nm pump intensity, where Ti and Tss refer to the initial and steady state transmittance of DR1P-SWCNT film at 543 nm.

Close modal

The photostationary state behavior in Fig. 2 is prototypical of azobenzene systems and is a prerequisite for photoinduced birefringence. Hence, we investigated the photoinduced birefringence using the setup shown in supplementary Fig. S3.27 A 480 nm laser beam is divided into a linearly polarized write beam and a circularly polarized erase beam. A linearly polarized 650 nm diode laser probe beam passes through the sample at normal incidence and through a Glan-Thompson polarizer, where the transmitted light is detected with an amplified photodetector. With the write beam off, the initially isotropic sample is between crossed polarizers, and the transmission at 650 nm is zero. With the write beam on, photoinduced reorientation of the chromophores generates in-plane birefringence, which is detected as an increase in transmission. For small optical retardance (applicable here), the birefringence Δn is related to the transmittance T according to Δn=(λπd)T, where λ = 650 nm and d is the sample thickness. A separate circularly polarized 480 nm beam is used to explore the erasure dynamics.

Fig. 3(a) summarizes the dynamic photoinduced birefringence properties of the DR1P-SWCNT system. The rapid increase in relative birefringence (i.e., T) during the writing phase 3 s < t < 35 s shows that the repeated trans-cis-trans isomerization cycle drives DR1P reorientation perpendicular to the write beam polarization; the sample becomes optically anisotropic, with an in-plane birefringence. With the writing beam off for 35 s < t < 53 s, the cis population thermally back-isomerizes to trans. The portion of the trans population resulting from this back-isomerization will not have a preferred in-plane orientation, causing the photoinduced birefringence to decay to a quasi-steady state birefringence that is approximately 45% of the initial steady state value. This residual birefringence will decay to near zero over 12 h under ambient conditions or it can be rapidly and completely erased with circularly polarized light (as shown for t > 53 s), which effectively randomizes the chromophore orientation. The cycle depicted in Fig. 3(a) could be repeated at least 103 times without significant change, indicating that repeated isomerization does not induce significant photodegradation of the DR1P chromophore.

FIG. 3.

(a) Rise, thermal decay, and erasure of the relative birefringence of DR1P-SWCNT film. (b) Maximum absolute photoinduced birefringence (black) and characteristic risetime (blue) for DR1P-SWCNT as a function of the linearly polarized 480 nm write intensity.

FIG. 3.

(a) Rise, thermal decay, and erasure of the relative birefringence of DR1P-SWCNT film. (b) Maximum absolute photoinduced birefringence (black) and characteristic risetime (blue) for DR1P-SWCNT as a function of the linearly polarized 480 nm write intensity.

Close modal

Fig. 3(b) focuses on the writing phase (i.e., 3 s < t < 35 s) and summarizes the dynamics and absolute steady-state birefringence as the 480 nm write beam intensity varies from 0.15 to 32.0 mW/cm2. Two regimes were observed: (I) for I480 < 1.0 mW/cm2, the birefringence increased monoexponentially with a time constant of approximately 10 s, reaching a steady state Δn ≈ 0.007 with 1.0 mW/cm2. (II) for I480 > 1.0 mW/cm2, the birefringence was biexponential, with Δn reaching a maximum of ≈0.013 at 32 mW/cm2. The biexponential dynamics consisted of a fast component (with time constant shown in Fig. 3(b)) and a smaller contribution from a slower component. The slow time constant is approximately a factor of 10 longer than the fast constant and is not displayed in Fig. 3(b). The most significant changes in steady state birefringence and time constant occur over a relatively small intensity range of 0.1–2 mW/cm2. Here, the overall degree of chromophore alignment is minimal, suggesting that cooperative effects among aligned chromophores play little, if any, role in the dynamics. At these intensities, the birefringence is primarily driven by the trans-cis-trans isomerization and rotation of individual chromophores. Above the 1–2 mW/cm2 threshold, however, the steady state birefringence and its accompanying time constant show a weaker dependence on the 480 nm pump intensity. This corresponds to the emergence of a second exponential term in the birefringence dynamics, suggesting that dipolar interactions among chromophores (accompanied possibly by movement of the nanotubes themselves) are contributing factors.

The observation of photoinduced birefringence in the DR1P-SWCNT film indicates that spatially modulated 480 nm pump light would optically inscribe a phase diffraction grating. Fig. 4(a) shows the experimental setup. Two s-polarized mutually coherent 480 nm beams overlap in the DR1P-SWCNT film, creating a spatially modulated intensity grating. The grating periodicity is given by Λ=λ2sinϕ, where λ is the pump beam wavelength and ϕ is the half-angle between the two beams. The grating is probed by a 632.8 nm HeNe laser beam incident normally on the film. Gratings from 0.75 to 3.25 μm were inscribed (corresponding to 4.0°< φ < 18.5°) and the angular location θexp of the m = 1 (first order) diffracted mode was measured. Using the grating equation λ=dsinθ with λ = 632.8 nm and d being the calculated grating spacing, the predicted locations θpred of the m = 1 mode were determined. Fig. 4(b) shows excellent agreement between the predicted and measured locations of the m = 1 diffracted mode.

FIG. 4.

(a) Experimental setup for optically defined birefringence gratings in DR1P-SWCNT film. (b) Predicted (black) and observed (red) locations of m = 1 diffracted order as a function of optically induced grating spacing Λ. Inset shows m = 1 intensity as 480 nm write intensities ranged from 4 mW/cm2 (black trace) to 40 mW/cm2 (dark blue trace). Write beams are turned on at t = 0 s.

FIG. 4.

(a) Experimental setup for optically defined birefringence gratings in DR1P-SWCNT film. (b) Predicted (black) and observed (red) locations of m = 1 diffracted order as a function of optically induced grating spacing Λ. Inset shows m = 1 intensity as 480 nm write intensities ranged from 4 mW/cm2 (black trace) to 40 mW/cm2 (dark blue trace). Write beams are turned on at t = 0 s.

Close modal

The inset in Fig. 4(b) shows the diffraction efficiency (i.e., IdiffractedIincident) as the 480 nm pump intensity is varied from 4 mW/cm2 to 40 mW/cm2 in six successive steps, with a single circularly polarized beam used to erase the grating after each exposure. At t = 0, two 480 nm pump beams overlap in the film, and the 633 nm diffracted probe increases until the phase grating reaches steady state. The maximum steady state diffraction efficiency was of order 0.01% using 40 mW/cm2. Additional experiments using p-polarized, orthogonal, and circularly polarized write beams revealed similar behavior with the intensity of the diffracted mode dependent on the polarization configuration. The orthogonal and counter-rotating circular configurations resulted in the highest diffraction efficiency. These configurations generate a grating of polarization (not intensity) in the film, demonstrating clearly that the observed diffraction phenomena originated from a spatially varying orientation of the DR1P chromophore.

The gratings described in Fig. 4(b) could be completely erased by exposing the DR1P-SWCNT film to a single circularly polarized 480 nm beam, except in those cases when the film was exposed longer than ∼20 min. In these cases, the diffraction efficiency increased twofold and the first order diffracted mode in transmission became accompanied by a similar mode in reflection. This indicates that the longer exposure re-structured the film's topography, creating a permanent surface relief grating. Such gratings could be produced with intensity or polarization gratings, with counter-rotating circularly polarized 480 nm beams being the most effective. Evidence for this mass re-distribution effect is shown in the AFM image (Fig. 5), taken at the surface of a 220 nm thick DR1P-SWCNT film where two counter-rotating circularly polarized 480 nm beams intersected. The surface is modulated ±40 nm with a periodicity of approximately 1.8 μm, in good agreement with the predicted grating spacing of 2.0 μm. In some samples, the presence of a surface relief grating was visually evident in the diffraction of white light from the film, as shown in supplementary Fig. S4;27 identical images could be obtained 1 year after exposure, indicating that the gratings are permanent and stable.

FIG. 5.

AFM images of DR1P-SWCNT film surface (a) before illumination with optical polarization grating and after ((b)–(d)). AFM height (b) and phase (c) images show evidence of surface relief grating, with ±40 nm of surface modulation as shown in (d).

FIG. 5.

AFM images of DR1P-SWCNT film surface (a) before illumination with optical polarization grating and after ((b)–(d)). AFM height (b) and phase (c) images show evidence of surface relief grating, with ±40 nm of surface modulation as shown in (d).

Close modal

Light-induced mass transport has been widely studied in azobenzene-polymer systems,28–35 but has not been observed in non-polymeric nanocarbon solid films such as DR1P-SWCNT. Some models in polymeric systems attribute mass migration to the free-volume differential between the chromophore cis and trans states, while others are based on photoinduced stress resulting from anisotropic molecular alignment.36–38 At this stage, it is unclear how such mechanisms may (or may not) apply to chromophore functionalized nanotubes and are the subject of ongoing investigations. What can be concluded is that photo-orientation of the DR1P chromophore produces a collective motion of the chromophore-SWCNT that ultimately results in a redistribution of material at the film surface and which points towards a promising technique for the photomanipulation of SWCNTs.

In conclusion, we have demonstrated that a non-polymeric nanocarbon solid film of chromophore-functionalized SWCNTs functions as a host matrix for a broad range of photo-orientation and photomechanical effects. The flexible linker binding the chromophore non-covalently enables efficient reorientation through repeated trans-cis-trans isomerization under linearly polarized 480 nm light, with Δn of 0.012 at 650 nm and fast characteristic rise times of 0.12 s. Erasable phase diffraction gratings could also be written in the film, with permanent surface relief gratings forming with sufficiently long irradiation times. In addition to demonstrating a mechanism for photo-manipulation of SWCNTs mediated by light switchable molecules, these results show an alternate route for the photosensitive tuning of the electrostatic environment of SWCNTs, with implications for optoelectronic applications such as photonic switching, detection, and possibly in SWCNT processing.

P.G. and D.J.M. acknowledge support from the Division of Materials Sciences and Engineering, Office of Basic Energy Science, U.S. Department of Energy under Award No. ER46590 for the synthesis, fabrication, and characterization of the hybrids.

1.
J.
Huang
,
A.
Datar
,
S.
Somu
, and
A.
Busnaina
,
Nanotechnology
22
(
45
),
455202
(
2011
).
2.
D. M.
Guldi
and
R. D.
Costa
,
J. Phys. Chem. Lett.
4
(
9
),
1489
(
2013
).
3.
C.
Sciascia
,
R.
Castagna
,
M.
Dekermenjian
,
R.
Martel
,
A. R.
Srimath Kandada
,
F.
Di Fonzo
,
A.
Bianco
,
C.
Bertarelli
,
M.
Meneghetti
, and
G.
Lanzani
,
J. Phys. Chem. C
116
(
36
),
19483
(
2012
).
4.
V.
Schneider
,
T.
Strunskus
,
M.
Elbahri
, and
F.
Faupel
,
Carbon
90
,
94
(
2015
).
5.
G. D.
Costanzo
,
L.
Ribba
,
S.
Goyanes
, and
S.
Ledesma
,
J. Phys. D: Appl. Phys.
47
(
13
),
135103
(
2014
).
6.
M. E.
Abbasov
,
S.
Ghosh
,
A.
Quach
, and
G. O.
Carlisle
,
J. Mater. Sci.: Mater. Electron.
21
(
8
),
854
(
2010
).
7.
C.
Vijayakumar
,
B.
Balan
,
M.
Kim
, and
M.
Takeuchi
,
J. Phys. Chem. C
115
(
11
),
4533
(
2011
).
8.
P.
Avouris
,
M.
Freitag
, and
V.
Perebeinos
,
Nat. Photonics
2
,
341
(
2008
).
9.
W.
Feng
,
W.
Luo
, and
Y.
Feng
,
Nanoscale
4
(
20
),
6118
(
2012
).
10.
V.
Sgobba
and
D. M.
Guldi
,
Chem. Soc. Rev.
38
(
1
),
165
(
2009
).
11.
J. M.
Simmons
,
I.
In
,
V. E.
Campbell
,
T. J.
Mark
,
F.
Léonard
,
P.
Gopalan
, and
M. A.
Eriksson
,
Phys. Rev. Lett.
98
(
8
),
086802
(
2007
).
12.
M.
Freitag
,
Y.
Martin
,
J. A.
Misewich
,
R.
Martel
, and
P.
Avouris
,
Nano Lett.
3
(
8
),
1067
(
2003
).
13.
X.
Zhou
,
T.
Zifer
,
B. M.
Wong
,
K. L.
Krafcik
,
F.
Leonard
, and
A. L.
Vance
,
Nano Lett.
9
(
3
),
1028
(
2009
).
14.
J.
Borghetti
,
V.
Derycke
,
S.
Lenfant
,
P.
Chenevier
,
A.
Filoramo
,
M.
Goffman
,
D.
Vuillaume
, and
J. P.
Bourgoin
,
Adv. Mater.
18
(
19
),
2535
(
2006
).
15.
D. S.
Hecht
,
L.
Hu
, and
G.
Irvin
,
Adv. Mater.
23
(
13
),
1482
(
2011
).
16.
A. M.
Kolpak
and
J. C.
Grossman
,
Nano Lett.
11
(
8
),
3156
(
2011
).
17.
C. S.
Huang
,
R. K.
Wang
,
B. M.
Wong
,
D. J.
McGee
,
F.
Leonard
,
Y. J.
Kim
,
K. F.
Johnson
,
M. S.
Arnold
,
M. A.
Eriksson
, and
P.
Gopalan
,
ACS Nano
5
(
10
),
7767
(
2011
).
18.
J. L.
Bahr
and
J. M.
Tour
,
J. Mater. Chem.
12
(
7
),
1952
(
2002
).
19.
P.
Rochon
,
J.
Gosselin
,
A.
Natansohn
, and
S.
Xie
,
Appl. Phys. Lett.
60
(
1
),
4
(
1992
).
20.
Z.
Sekkat
and
W.
Knoll
,
Photoreactive Organic Thin Films
(
Academic Press
,
2002
).
21.
Z.
Sekkat
,
D.
Morichère
,
M.
Dumont
,
R.
Loucif-Saïbi
, and
J. A.
Delaire
,
J. Appl. Phys.
71
(
3
),
1543
(
1992
).
22.
R.
Hagan
and
T.
Bieringer
,
Adv. Mater.
13
(
23
),
1805
(
2001
).
23.
T.
Todorov
,
L.
Nikolova
, and
N.
Tomova
,
Appl. Opt.
23
(
23
),
4309
(
1984
).
24.
W.
Zhang
,
S.
Bian
,
S. I.
Kim
, and
M. G.
Kuzyk
,
Opt. Lett.
27
(
13
),
1105
(
2002
).
25.
C. J.
Barrett
,
J.
Mamiya
,
K. G.
Yager
, and
T.
Ikeda
,
Soft Matter
3
(
10
),
1249
(
2007
).
26.
J. J.
Wie
,
D. H.
Wang
,
K. M.
Lee
,
L.
Tan
, and
T. J.
White
,
Chem. Mater.
26
(
18
),
5223
(
2014
).
27.
See supplementary material at http://dx.doi.org/10.1063/1.4934864 for additional experimental details and characterization of DR1P/SWCNT film.
28.
C. J.
Barrett
,
A. L.
Natansohn
, and
P. L.
Rochon
,
J. Phys. Chem.
100
,
8836
(
1996
).
29.
J. A.
Delaire
and
K.
Nakatani
,
Chem. Rev.
100
,
1817
(
2000
).
30.
X. L.
Jiang
,
L.
Li
,
J.
Kumar
,
D. Y.
Kim
,
V.
Shivshankar
, and
S. K.
Tripathy
,
Appl. Phys. Lett.
68
(
19
),
2618
(
1996
).
31.
M. L.
Juan
,
J.
Plain
,
R.
Bachelot
,
P.
Royer
,
S. K.
Gray
, and
G. P.
Wiederrecht
,
ACS Nano
3
(
6
),
1573
(
2009
).
32.
D. Y.
Kim
,
S. K.
Tripathy
,
L.
Li
, and
J.
Kumar
,
Appl. Phys. Lett.
66
(
10
),
1166
(
1995
).
33.
J. E.
Koskela
,
J.
Vapaavuori
,
J.
Hautala
,
A.
Priimagi
,
C. F. J.
Faul
,
M.
Kaivola
, and
R. H. A.
Ras
,
J. Phys. Chem. C
116
(
3
),
2363
(
2012
).
34.
P.
Rochon
,
E.
Batalla
, and
A.
Natansohn
,
Appl. Phys. Lett.
66
(
2
),
136
(
1995
).
35.
Z.
Mahimwalla
,
K. G.
Yager
,
J.
Mamiya
,
A.
Shishido
,
A.
Priimagi
, and
C. J.
Barrett
,
Polym. Bull.
69
(
8
),
967
(
2012
).
36.
O. N.
Oliveira
,
L.
Li
,
J.
Kumar
, and
S. K.
Tripathy
,
Photoreactive Organic Thin Films
(
Academic Press
,
2002
), p.
429
.
37.
M.
Saphiannikova
and
D.
Neher
,
J. Phys. Chem. B
109
(
41
),
19428
(
2005
).
38.
V.
Toshchevikov
,
M.
Saphiannikova
, and
G.
Heinrich
,
J. Phys. Chem. B
113
(
15
),
5032
(
2009
).

Supplementary Material