Photoinduced crystal-to-liquid phase transitions of azobenzene derivatives and their application in photolithography processes through a solid-liquid patterning

Article abstract of DOI:10.1021/ol502223u

The direct and reversible transformation of matter between the solid and liquid phases by light at constant temperature is of great interest because of its potential applications in various manufacturing settings. We report a simple molecular design strategy for the phase transitions: azobenzenes having para-dialkoxy groups with a methyl group at the meta-position. The photolithography processes were demonstrated using the azobenzene as a photoresist in a single process combining development and etching of a copper substrate.

Full text of DOI:10.1021/ol502223u

Letter
pubs.acs.org/OrgLett
Photoinduced Crystal-to-Liquid Phase Transitions of Azobenzene
Derivatives and Their Application in Photolithography Processes
through a Solid−Liquid Patterning
Yasuo Norikane,*,† Emi Uchida,† Satoko Tanaka,† Kyoko Fujiwara,† Emiko Koyama,† Reiko Azumi,†
Haruhisa Akiyama,‡ Hideyuki Kihara,‡ and Masaru Yoshida‡
†Electronics and Photonics Research Institute, National Institute of Advanced Industrial Science and Technology (AIST), Central 5,
1-1-1 Higashi, Tsukuba, Ibaraki 305-8565, Japan
‡Nanosystem Research Institute, National Institute of Advanced Industrial Science and Technology (AIST), Central 5, 1-1-1 Higashi,
Tsukuba, Ibaraki 305-8565, Japan
S
* Supporting Information
ABSTRACT: The direct and reversible transformation of
matter between the solid and liquid phases by light at constant
temperature is of great interest because of its potential
applications in various manufacturing settings. We report a
simple molecular design strategy for the phase transitions:
azobenzenes having para-dialkoxy groups with a methyl group
at the meta-position. The photolithography processes were
demonstrated using the azobenzene as a photoresist in a single
process combining development and etching of a copper
substrate.
hase transitions of matter are generally induced by heat. In
about azobenzene-based systems; thus, molecular design
1
Pcontrast, some materials, such as azobenzene, undergo criteria need to be established. Earlier, the trans-to-cis
photoisomerization of azobenzene had been long believed
not to take place in the crystal bulk phase.6 Later, it was found
to take place on the surface of the crystals because of the crystal
packing and steric hindrance.7,8 More recently, we and others
found that several azobenzene-based compounds undergo
crystal-to-liquid phase transitions,2,9 but their complex
molecular structures obfuscate the molecular design criteria.
Thus, we set our target to find simpler molecules that show the
photoinduced phase transition and can be used as a photoresist.
Here, we report a simple molecular design strategy that is
very effective for the crystal-to-liquid phase transition.
Introduction of a methyl group at the 3-position on rod-
shaped azobenzene derivatives drastically alters the photo-
induced phase transition behavior (Figure 1). Despite their
simple molecular structures, these are new compounds. These
materials could also be used as photoresists to pattern copper
substrates, such as those used in printed circuit boards. The
liquefied region of a pattern can be readily removed by wiping,
blowing, or washing, and the substrate can then undergo
etching. Notably, the proposed process provides a new concept
in application of phase transition materials because this process
does not require the acids and/or bases currently necessary to
remove conventional photoresists.
photochromic reactions that result in photoinduced phase
transitions involving liquid crystal phases,1f,i,j sol−gel tran-
sitions,1g,i and fluidity changes in thin film surfaces.1d,h In
particular, the photoinduced transition between the solid and
liquid phases for azobenzene induced by light at rt2,3 has
attracted significant interest from the application viewpoint.
The direct phase conversions of bulk materials and the
reversibility of such compounds suggest simple and resource-
saving applications in reusable smart materials. An intriguing
application has been proposed for use as photocontrollable
adhesives3 that can perform photoattaching and detaching.
Another possible application of such phase-transition materials
is as photoresists that enable solid/liquid patterning. Currently,
conventional photoresists4 are not reusable because they use
irreversible photopolymerization, degradation, or cross-linking
reactions. In addition, significant amounts of acids and/or bases
are used to remove the photoresists in the development/
washing processes.4 By using phase transition materials, the
isothermal solid/liquid patterning provides simple development
processes such as wiping, blowing, or washing by solvents, and
thus it is expected to reduce the chemical waste and to simplify
the photolithography process. However, until now, there has
been no report of even a simple demonstration of photolitho-
graphic processes using such phase-transition materials.
Although solid-state (crystalline) organic photochemistry has
Received: August 3, 2014
Published: September 12, 2014
been extensively studied in various compounds,5 little is known
© 2014 American Chemical Society
5012
dx.doi.org/10.1021/ol502223u | Org. Lett. 2014, 16, 5012−5015

Organic Letters
Letter
Figure 2. UV−vis absorption spectra of (a) 1b, (b) 2b, and (c) 3b in
the crystalline phase upon photoirradiation.
(Figure 1b,c), it took less than 1 min for a thin film (Figure 2b).
Crystalline films show slight scattering, but it does not
significantly affect the optical measurements (Figures 1b, S4−
S6). A crystalline film of 2b showed absorption bands around
390 nm; these diminished upon irradiation using 365 nm light,
accompanied by the appearance of an absorption band around
450 nm (Figure 2b). Polarization optical microscopy (POM)
showed a loss in birefringence of the crystal on irradiation at
365 nm, indicating that a phase transition occurred due to
photoisomerization (Figure S7). At the photostationary state,
the content of the cis isomer in this film is 90% estimated by
Figure 1. (a) Chemical structures and (b) photographs of the
crystalline powders of the azobenzene derivatives 1a−b, 2a−c, and
3a−b. (c) The same powders after irradiation with 365 nm light. The
irradiation period was 30 min at an intensity of 100 mW cm−2.
We selected rod-shaped molecules because of their facile
synthesis and simple structure−property relationships. Methyl
groups were introduced at the 3-position (2a−c) or the 3- and
3′-positions (3a−b), anticipating reduced intermolecular π−π
interactions that could potentially retard photoisomerization in
the solid phase. The parent azobenzenes (1a−b)10 were also
investigated as controls. The syntheses of these compounds
were achieved via two conventional reactions in fair yields. In
the case of 2a, for example, the first step is a synthesis of 4,4′-
dihydroxy-3-methyl-azobenzene by diazo coupling between 4-
aminophenol and o-cresol with a 53% yield. The second step
involves alkylation of the hydroxyl groups by dodecyl bromide
with a 96% yield. Compounds with three different alkoxy chain
lengths (C12, C6, and, for 2 only, C10) were synthesized (see
Supporting Information (SI) for further details). The thermal
phase transition temperatures of the final products were
determined by differential scanning calorimetry (DSC). All
compounds were crystalline at rt (Table 1 and SI).
1
dissolving the film into a solvent and measuring H NMR and
absorption spectra.
The viscosity of photoirradiated 2b was 460 mPa·s, clearly
revealing its liquid state.11 The liquefied samples remained in
their liquid phase for ∼1 h at rt after ceasing the irradiation
(Figure S8). Reverse photoisomerization (cis-to-trans), and the
consequent crystallization, was observed on irradiation with
visible (436 nm) light (Figure S7). An exceptional case was
observed in 3b; although its crystals exhibited a slight change in
the absorption spectrum on irradiation (Figure 2c), no phase
transition was observed (Figure 1b and 1c) at rt. Although 3a
has a lower melting point (mp, 83 °C) than that of 2b (87 °C),
3a does not show the phase transition. The temperature
increase of the crystal powder surface, measured using an
infrared radiation thermometer, was about 3 K during
irradiation at 200 mW cm−2. To confirm that the heating
effects are not the main reason for the liquefaction, an
irradiation experiment at lower intensity of light (4.6 mW
cm−2) was carried out, and we observed the phase transition
(Figure S9). Since the deactivation process of photoexcited
molecules involves dissipation of energy as heat, a temperature
increase cannot be avoided. However, the above-mentioned
results show that the heating effects are not the main reason for
the phase transition.
To understand the crystal-to-liquid phase transition, let us
consider the isomerization and melting steps separately. First, it
is presumed that the photoisomerization takes place at the
surface of the crystals.7,8 The intensity of the incoming light is
highest here, and thus, photons are primarily absorbed at the
outer surface. The molecular packing and dynamic nature of the
crystal might affect its photoisomerization ability. To study this,
single crystal X-ray analyses were performed to determine the
crystal structures of 1b, 2a−c, and 3b (Figure 3; see SI for
details). Comparing the crystal structures of 2b and 3b, their
single molecular conformations are very close to each other and
similar molecular assembly structures are formed in the ab
planes. However, within a single layer, molecules of 2b form a
zigzag molecular assembly, whereas 3b molecules are linearly
assembled with interdigitated alkyl chains.
a
Table 1. Thermal Phase Transition Temperatures of the
Azobenzene Derivatives
b
phase transition temperature (°C)
molecule
heating
cooling
1a
1b
2a
2b
2c
3a
3b
CrII 97 CrI 107 I
Cr 102 N 114 I
Cr 66 N 72 I
I 105 N 99 CrI 88 CrII
I 113 N 99 Cr
I 72 N 53 Cr
Cr 87 I
I 78 Cr
CrII 59 CrI 60 N 75 I
CrII 54 CrI 83 I
CrII 94 CrI 105 I
I 74 N 51 CrII
I 62 CrI 51 CrII
I 92 CrII
a
b
Determined by DSC at scan rate of 2 K min−1. Cr: crystal. N:
nematic liquid crystal phase. I: isotropic liquid phase. Subscripts (I and
II) indicate polymorphism.
In the solid state at rt, photoinduced phase transitions were
only observed for 2a−c (Figures 1b, 1c, 2, and SI video),
whereas the photoisomerization of all the compounds was
confirmed in chloroform (Figure S4−S6). The irradiation time
required for complete liquefaction strongly depended on the
sample thickness: although it required 30 min (at 100 mW
cm−2) to complete the liquefaction of a pile of crystal powder
5013
dx.doi.org/10.1021/ol502223u | Org. Lett. 2014, 16, 5012−5015

Organic Letters
Letter
removal. We could also selectively remove the liquid material
by immersing the film in aqueous propanol (H2O/2-propanol =
2:3 v/v) for 10 min at rt (Figure S20).
The chemical etching of copper is widely used in art and
industry. It is an important photolithographic process in the
production of printed circuit boards, the substrates on which
electronic parts are placed.13,14 Generally, a photoresist film is
formed on a copper substrate and exposed to light through a
mask. Then, the exposed/modified resist film is developed to
produce a pattern. An etching process removes the copper that
is not protected by the remaining resist. In the final step, the
remaining resist is stripped, leaving the patterned features of the
substrate.4,13
Figure 3. Crystal structures of 1b, 2b, and 3b. Top: single molecule
structures. Middle: crystal faces parallel to the molecular plane of (a)
Etching of a copper substrate was successfully achieved using
2c as a photoresist (the sequence (1) → (2) → (3) → (4) →
(6) in Figure 4a). A spin-coated film of 2c on a copper
the (101) face of 1b, (b) the (001) face of 2b, and (c) the (001) face
̅
of 3b, displayed as space-filling models. Bottom: the same crystal faces
as in the middle images showing two molecular layers, where the front
and rear layers are shown as stick and wireframe models, respectively.
A single molecule in each image is highlighted for clarity.
Second, the cis isomers generated by light irradiation at the
crystal surface should affect the macroscopic phase of the
sample. When the cis isomer has an mp sufficiently higher than
rt, the crystal does not melt. The solid cis isomer prevents
further isomerization because of steric hindrance.12 When the
mp is lower than rt, the liquid cis isomer appears and results in
further isomerization at the crystal−liquid interface. To
estimate the mp of the corresponding cis isomers, the thermal
behavior of the photoirradiated samples was investigated using
POM. It was revealed that the mp of the cis isomer of 2b is
around −6 °C (Figure S15), whereas that of 3b is around 90 °C
(Figures S16 and S17). These results show that the mp of the
cis isomer changes drastically because of the rather small
molecular modification upon the introduction of a methyl
group.
Since the phase transition occurs at the crystal surface, the
irradiation time strongly depends on the sample thickness for
the complete liquefaction. Therefore, it requires less photons
when the sample is thinner (Figure S9). In addition, the melted
state (liquid) covers the crystal and absorbs the incoming
photons resulting in slowing down the transition. Melting
occurs when the conversion at the crystal surface reaches the
ratio at which the mp of the isomer mixture is below rt.
For photolithography, a photoresponsive material should
possess film-forming ability, sensitivity to light exposure, spatial
resolution, compatibility with development chemicals, robust-
ness against etching reagents, low cost, low toxicity, and so on.
We fabricated spin-coated films (thickness: ∼1 μm) on a cover
glass and observed that 2c was able to form a good film (Figure
S18). The X-ray diffraction profile of the 2c thin film was
similar to that of the crystalline powder, indicating that it is also
crystalline (Figure S19).
When exposed to 365 nm light through a mask, the spin-
coated 2c exhibited a photoinduced phase transition to liquid
selectively in the irradiated regions (Figure S18). It should be
noted that the liquefaction of the thin film occurred within 60 s
of irradiation, much faster than the transition in the crystal
powder (Figure 1) at the same light intensity. We found no
significant dependence of the phase transition property on the
sample preparation method (crystalline films prepared by
cooling from the melt or by spin coating). Blowing, wiping, or
blotting can remove the liquid material (viscosity 425 mPa·s),
although each method must be optimized for complete
Figure 4. (a) Photolithographic process for the selective etching of
copper substrates. (b) Optical photomicrograph of a patterned
azobenzene 2c on a copper substrate (state D in (a)). (c) Photograph
of a patterned copper substrate being subjected to the processes (1) →
(2) → (5) → (6) (obtaining state F). (d) SEM image of the substrate
shown in (b). (e) SEM image of the copper substrate shown in (c).
substrate (commercially available copper-clad) was irradiated at
365 nm through a mask, and the substrate was immersed in a
propanol solution (Figure 4b and 4d). Then the substrate was
placed in an etching solution (23% FeCl3 in H2O). The
substrate was rinsed with water, and then with acetone, to
remove the unexposed 2c on top of the unetched copper layer
(Figure S22b). Thus, the spin-coated layer of 2c resists the
etching solution.
In addition, the developing and etching steps were combined
and carried out simultaneously by simply adding 2-propanol
(2.5% v/v) to the etching solution (the sequence (1) → (2) →
(5) → (6) in Figure 4a). A photograph and an SEM image of
the final product are shown in Figure 4c and 4e, respectively.
In summary, we have shown that simple rod-shaped
azobenzenes having a methyl group at the 3-position and
para-alkoxy chains exhibited crystal-to-liquid photoinduced
phase transitions at rt as revealed by optical observations,
spectroscopic analysis, and viscosity measurements. The
molecular packing and melting points of the corresponding
5014
dx.doi.org/10.1021/ol502223u | Org. Lett. 2014, 16, 5012−5015

Organic Letters
Letter
R.; Nozawa, S.; Tomita, A.; Sato, T.; Adachi, S.; Koshihara, S. J. Am.
Chem. Soc. 2014, 136, 9158−9164.
(3) (a) Akiyama, H.; Yoshida, M. Adv. Mater. 2012, 24, 2353−2356.
(b) Akiyama, H.; Kanazawa, S.; Okuyama, Y.; Yoshida, M.; Kihara, H.;
Nagai, H.; Norikane, Y.; Azumi, R. ACS Appl. Mater. Interfaces 2014, 6,
7933−7941.
(4) (a) Okoroanyanwu, U. Chemistry and Lithography; SPIE
PressJohn Wiley & Sons: Bellingham, WA, Hoboken, NJ, 2010.
(b) Madou, M. J. Fundamentals of Microfabrication; CRC Press: Boca
Raton, FL, 2002.
(5) (a) Keating, A. E.; Garcia-Garibay, M. A. In Organic and Inorganic
Photochemistry, Vol. 2; Ramamurthy, V., Schanze, K., Eds.; Marcel
Dekker: New York, 1998; pp 195−248. (b) Garcia-Garibay, M. A. Acc.
Chem. Res. 2003, 36, 491−498. (c) de Loera, D.; Stopin, A.; Garcia-
Garibay, M. A. J. Am. Chem. Soc. 2013, 135, 6626−6632.
(6) Tsuda, M.; Kuratani, K. Bull. Chem. Soc. Jpn. 1964, 37, 1284−
1288.
cis isomers provide information on the phase transition. Further
studies including observing the dynamic behavior of molecules
in the crystal phase, obtaining a complete phase diagram of the
phase transition, and improving their photosensitivity are
currently ongoing. One of the compounds, possessing an
excellent film-forming ability, can be used as a photoresist, and
the selective etching of copper was successfully achieved using a
simple photolithographic method. The azobenzene is easily
removed using common organic solvents such as acetone and
2-propanol. No acids or bases are required to develop or to
remove the resist. These new materials also can be used to
fabricate patterns on solid surfaces sensitive to acids/bases. In
principle, the azobenzene can be recovered and reused. Our
molecular design strategy is simple but very effective for the
development of phase-transitionable materials. The results
described here provide significant knowledge of azobenzene-
based materials and a novel concept of their application that
will contribute to energy and resource saving technologies.
(7) (a) Nakayama, K.; Jiang, L.; Iyoda, T.; Hashimoto, K.; Fujishima,
A. Jpn. J. Appl. Phys. 1997, 36, 3898−3902. (b) Ichimura, K. Chem.
Commun. 2009, 1496−1498.
(8) (a) Koshima, H.; Ojima, N.; Uchimoto, H. J. Am. Chem. Soc.
2009, 131, 6890−6891. (b) Nakano, H. J. Mater. Chem. 2010, 20,
2071−2074. (c) Bushuyev, O. S.; Singleton, T. A.; Barrett, C. J. Adv.
Mater. 2013, 25, 1796−1800. (d) Bushuyev, O. S.; Tomberg, A.;
ASSOCIATED CONTENT
* Supporting Information
Experimental procedures; synthesis and characterization data of
4,4′-dihydroxy-3-methylazobenzene, 2a−c, and 3b; Figures
S1−S24; crystallographic information files (CIF) for 1b, 2a−
c, and 3b, video of phase transition. This material is available
free of charge via the Internet at http://pubs.acs.org.
■
S
̌ ̌ ́
Friscic, T.; Barrett, C. J. J. Am. Chem. Soc. 2013, 135, 12556−12559.
(e) Nath, N. K.; Pejov, L.; Nichols, S. M.; Hu, C.; Saleh, N.; Kahr, B.;
Naumov, P. J. Am. Chem. Soc. 2014, 136, 2757−2766.
(9) Okui, Y.; Han, M. Chem. Commun. 2012, 48, 11763−11765.
(10) (a) Kidowaki, M.; Moriyama, M.; Wada, M.; Tamaoki, N. J.
Phys. Chem. B 2003, 107, 12054−12061. (b) Ruslim, C.; Ichimura, K.
J. Mater. Chem. 1999, 9, 673−681.
AUTHOR INFORMATION
Corresponding Author
■
(11) Viscosity of trans isomer of 2b observed at 100 °C was 12 mPa·
s.
*E-mail: y-norikane@aist.go.jp.
Notes
(12) Sworakowski, J.; Janus, K.; Nespurek, S. Adv. Colloid Interface
̌
̊
Sci. 2005, 97−110.
(13) Khandpur, R. S. Printed Circuit Boards: Design, Fabrication,
Assembly and Testing; Tata McGraw-Hill: New Delhi, 2005; pp 365−
383.
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
(14) The etching process of the copper substrate has been adopted as
educational experiment programs. For example, see: Outreach
resources, The National HE STEM Programme, Royal Society of
Chemistry. http://www.rsc.org/Education/HESTEM/Outreach/
photolithography.asp (accessed July 27th, 2014).
■
This work was supported in part by JSPS/MEXT KAKENHI
(23760680), Iketani Science and Technology Foundation, and
The Canon Foundation. We acknowledge Dr. Midori Goto
(AIST) for the X-ray crystallography studies. We thank Dr.
Masayuki Chikamatsu (AIST) for the thin film X-ray diffraction
experiments. We thank Dr. Masashi Kuwahara and Dr.
Takayuki Shima (AIST) for providing the SEM and stylus
profiler. We thank Ms. Tomoko Kobayashi and Ms. Naomi
Noguchi (AIST) for technical assistance. We also thank Dr.
Masahiro Aoyagi (AIST) for helpful discussions and
suggestions.
REFERENCES
■
(1) Reviews of azobenzenes: (a) Bandara, H. M. D.; Burdette, S. C.
Chem. Soc. Rev. 2012, 41, 1809−1825. (b) Mahimwalla, Z.; Yager, K.
G.; Mamiya, J.; Shishido, A.; Priimagi, A.; Barrett, C. J. Polym. Bull.
2012, 69, 967−1006. (c) Merino, E.; Ribagorda, M. Beilstein J. Org.
Chem. 2012, 8, 1071−1090. (d) Lee, S.; Kang, H. S.; Park, J.-K. Adv.
Mater. 2012, 24, 2069−2103. (e) Beharry, A. A.; Woolley, G. A. Chem.
Soc. Rev. 2011, 40, 4422−4437. (f) Yu, H.; Ikeda, T. Adv. Mater. 2011,
23, 2149−2180. (g) Yagai, S.; Kitamura, A. Chem. Soc. Rev. 2008, 37,
1520−1529. (h) Seki, T. Bull. Chem. Soc. Jpn. 2007, 80, 2084−2109.
(i) Kato, T.; Hirai, Y.; Nakaso, S.; Moriyama, M. Chem. Soc. Rev. 2007,
36, 1857−1867. (j) Ikeda, T.; Mamiya, J.; Yu, Y. Angew. Chem., Int. Ed.
2007, 46, 506−528.
(2) (a) Norikane, Y.; Hirai, Y.; Yoshida, M. Chem. Commun. 2011,
47, 1770−1772. (b) Uchida, E.; Sakaki, K.; Nakamura, Y.; Azumi, R.;
Hirai, Y.; Akiyama, H.; Yoshida, M.; Norikane, Y. Chem.Eur. J. 2013,
19, 17391−17397. (c) Hoshino, M.; Uchida, E.; Norikane, Y.; Azumi,
5015
dx.doi.org/10.1021/ol502223u | Org. Lett. 2014, 16, 5012−5015