S1

Supporting Information for:

Photoinduced Macroscopic Morphological Transformation of an Amphiphilic Diarylethene Assembly: Reversible Dynamic Motion

Kenji Higashiguchi,*,†,‡ Genki Taira,† Jun-ichiro Kitai,† Takashi Hirose,† and Kenji Matsuda*,†

†Department of Synthetic Chemistry and Biological Chemistry, Graduate School of Engineering,

Kyoto University, Katsura, Nishikyo-ku, Kyoto 615-8510, Japan

‡PRESTO, Japan Science and Technology Agency, Kawaguchi 332-0012, Japan

S2

A. EXPERIMENTAL SECTION

1. Preparation of Materials.

General. All reactions were monitored by thin-layer chromatography carried out on 0.2 mm Merck

silica gel plates (60F-254). Column chromatography was performed on silica gel (Nacalai, 70-230

mesh). Some compounds were purified by a JAI recycling preparative GPC LC-908 with

JAIGEL-1H and 2H column. HPLC was carried out on a HITACHI LC System LaChrom. An

analytical (Kanto Chemical, Mightysil RP-18(H) GP250-4.6 (5 m)) column and CH3CN was used

for purification of the synthesized compounds and separation of the closed-ring isomer of

diarylethene 1. 1H and 13C NMR spectra were recorded on a JEOL JNM-ECS400, a JNM-Alpha500,

and a JNM-ECA600 instrument. Samples were dissolved in CDCl3. Mass spectra were obtained by

a ThermoFisher Scientific EXACTIVE mass spectrometer. 1-Bromo-4-nonyloxybenzene is known

compound.[S1]

Synthesis.

Bis(2-(2-(2-methoxyethoxy)ethoxy)ethyl)-5-iodoisophthalate. To a THF (40 mL) solution of

5-iodoisophthalic acid[S2] (4.5 g, 15 mmol) were added DCC (15 g, 73 mmol), DMAP (0.31 g, 1.5

mmol), and triethylene glycol monomethyl ether (6.0 mL, 6.6 g, 36 mmol) at room temperature and

then the mixture was stirred for 12 h. Purification with column chromatography (silica gel,

CH2Cl2:acetone = 10:0-0:10) gave bis(2-(2-(2-methoxyethoxy)ethoxy)ethyl) 5-iodoisophthalate (2.4

g, 27%, colorless oil).

1H NMR (600 MHz, CDCl3): 8.64 (t, J = 1.5 Hz, 1H), 8.55 (d, J = 1.5 Hz, 2H), 4.51 (t, J =

5.0 Hz, 4H), 3.85 (t, J = 5.0 Hz, 4H), 3.71-3.72 (m, 4H), 3.66-3.69 (m, 4H), 3.64-3.65 (m, 4H),

3.53-3.55 (m, 4H), 3.37 (s, 6H); 13C NMR (100 MHz, CDCl3) 59.1, 64.7, 69.0, 70.59, 70.61, 70.7,

71.9, 76.7, 93.4, 130.1, 132.2, 142.6, 164.3ESI-HRMS (m/z) [M+H]+ calcd for C22H34IO10+

585.1191; found: 585.1185.

S3

1-(2-Methyl-5-(4-nonyloxyphenyl)-3-thienyl)-2-(5-(3,5-bis(2-(2-(2-methoxyethoxy)ethoxy)etho

xycarbonyl)phenyl)-2-methyl-3-thienyl)hexafluorocyclopentene 1. An m-xylene (4 mL) solution

of 2,2'-bipyridine (38 mg, 0.24 mmol) and PdCl2 (42 mg, 0.24 mmol) was heated at 60 °C for 30

min under a N2 atmosphere, after which it was allowed to cool to room temperature.

1,2-Bis(2-methyl-3-thienyl)hexafluorocyclopentene (1.06 g, 2.88 mmol),

1-bromo-4-nonyloxybenzene (1.0 g, 2.9 mmol), bis(2-(2-(2-methoxyethoxy)ethoxy)ethyl)

5-iodoisophthalate (1.7 g, 2.9 mmol), and Ag2CO3 (2.0 g, 7.2 mmol) in m-xylene (4 mL) were

added to the above solution at room temperature, and the mixture was heated at 140 °C overnight

under a N2 atmosphere. The precipitates were removed by silica gel bed filtration with acetone, and

the filtrate was purified by column chromatography (silica gel, CH2Cl2/acetone = 1:9 to 2:8).

1-(2-Methyl-5-(4-nonyloxyphenyl)-3-thienyl)-2-(5-(3,5-bis(2-(2-(2-methoxyethoxy)ethoxy)ethoxyc

arbonyl)phenyl)-2-methyl-3-thienyl)hexafluorocyclopentene 1 (490 mg, 16%) was obtained as a

colorless amorphous product.

1H NMR (500 MHz, CDCl3): 8.60 (s, 1H), 8.37 (s, 2H), 7.46 (d, J = 8.0 Hz, 2H), 7.44 (s,

1H), 7.16 (s, 1H), 6.90 (d, J = 8.0 Hz, 2H), 4.55 (t, J = 5.0 Hz, 4H), 3.97 (t, J = 5.0 Hz, 2H),

3.85-3.90 (m, 4H), 3.70-3.75 (m, 4H), 3.67-3.69 (m, 4H), 3.65-3.66 (m, 4H), 3.53-3.55 (m, 4H),

3.35 (s, 6H), 1.98 (s, 3H), 1.96 (s, 3H), 1.76-1.80 (m, 2H), 1.55-1.59 (m, 2H), 1.25-1.35 (m, 10H),

0.88 (t, J = 7.5 Hz, 3H); 13C NMR (150 MHz, CDCl3): 14.0, 14.45, 14.49, 22.5, 25.9, 29.09, 29.13,

29.3, 29.4, 31.8, 58.9, 64.5, 68.0, 69.0, 70.45, 70.53, 70.6, 71.8, 110.9 (m), 114.8, 116.0 (m),117.7

(m), 120.9, 123.8, 125.4, 125.7, 126.1, 126.8, 129.6, 130.5, 131.4, 134.1, 135.3 (t), 136.7 (t), 139.8,

140.1, 142.4, 142.5, 159.0, 165.3; ESI-HRMS (m/z) [M+NH4]+ calcd for C52H68F6O11S2N

+:

1060.4132; found: 1060.4116.

S4

NMR charts.

1H NMR of 1 (500 MHz, CDCl3)

S5

13C NMR of 1 (150 MHz, CDCl3)

S6

1H NMR of bis(2-(2-(2-methoxyethoxy)ethoxy)ethyl)-5-iodoisophthalate (600 MHz, CDCl3).

S7

13C NMR of bis(2-(2-(2-methoxyethoxy)ethoxy)ethyl)-5-iodoisophthalate (100 MHz, CDCl3).

S8

2. Photochromic Reaction.

The suspensions of diarylethene 1 for absorption spectroscopy were prepared as follows:

pure water (4 mL) was added to the acetonitrile solution (500 L) of 1 (7.9 mg) in one portion at

room temperature. The obtained suspension was diluted 1/10 and the resulting suspension showed

absorbance lower than 2 around 300 nm.

UV-vis spectra were measured with a JASCO V-670 spectrometer equipped with ETCS-761

Peltier temperature controller for cuvette. The temperature control was carried out as follows; the

cooling and heating rate was 0.1 °C/min and the sample was kept at suitable temperature for 15 min

with stirring at 60 rpm before measurement. For photoirradiation, an USHIO 500 W super

high-pressure mercury lamp was employed as a light source. The light passed through some ATG

band-pass filters and a monochromator (Ritsu MC-20L) was irradiated to induce the

photoisomerization of the molecules in the solution.

Closed-ring isomer 1b was isolated by passing the solution containing open- and closed-ring

isomers through HPLC (a HITACHI L-7100 pump and a L-7400 UV detector) with silica gel

column (Mightysil RP-18(H), Kanto, for reversed phase).

3. Photochromic Reaction under Microscope.

The turbid suspensions of diarylethene 1 were prepared by the addition of pure water (90

L) to the acetonitrile solution (15 L) of 1 (25 mg) in one portion at room temperature.

Optical microscope observation of the aggregates of diarylethene 1 in water was performed

in transmission geometry using a Nikon ECLIPSE LV100 instrument coupled with a Nikon DS-Fi1

CCD camera and a SONY HDR-CX590 camcorder. The camcorder was used to observe the fast

division of microspheres as Figures 4(e), 6, S6, and S7 despite loss of the scale information. Figures

4(c) and 4(d) were also obtained by the camcorder to observe the fast movement by irradiation of

visible light, however the scale was calibrated by another image with scale information.

The objective lenses were Nikon Plan-Fluor 10×/0.30, 20×/0.50, and 40×/0.75. A glass cell

with a shallow basin was filled with an aqueous suspension at room temperature. Temperature

control was achieved using a Peltier stage (Tokai Hit, MATS-555S).

UV irradiation of a selected area was performed using an epifluorescence microscope with a

130 W mercury lamp (G-HGFI, Nikon) and filter set (UV-1A, Nikon) for the photochromic reaction

S9

of the aggregates of diarylethene 1. The expected conversion ratio was around 60% at the

photostationary state upon irradiation with 365 nm light. The decoloration by the irradiation of

visible light was achieved using a band-pass filter (Y-50, ATG) or a filter set (G-2A, Nikon).

The luminescent spectra of light source used for irradiation of the selected supramolecular

assemblies were measured by a miniature fiber-optic spectrometer Ocean Optics S2000 with a 600

m optical fiber as shown in Figure S1.

Figure S1. Luminescent spectra of (a) UV light, (b) visible light (cf. Figure 4b), and (c) visible light (cf. Figure 4d).

The irradiation power was measured by a thermal power sensor OPHIR 3A-FS. The power

of UV (Figures 4a and S1a) and visible (Figures 4b and S1b) light under an epifluorescence

microscope was measured from the slope of Figure S2 as 6.2 × 10-5 and 2.5 × 10-4 W/m2,

respectively. The number of photon of UV and visible light was estimated as 1.9 × 10-4 and 1.2 ×

10-3 einstein m-2 s-1, respectively.

0

2

4

6

8

10

12

14

16

0 10 20 30 40 50 60 70Area / 10-5 cm2

Ener

gy / μW

Figure S2. Relationship between the radiation energy and the area under epifluorescence microscope: blue open circles and green solid circles denote UV and visible light, respectively.

S10

4. Viscosity Change of the Suspension.

The change of viscosity of the suspension was measured by a differential pressure

viscometer Viscotech VISC-S (Figure S3). The flow rate was 0.45 mL/min and the temperature

was 23 °C. The concentrated suspension was prepared by addition of 1.5 mL of water to a 0.3 mL

solution of 1 in acetonitrile (50 mg/1 mL).

1st irradiationwith vis. light

1st irradiationwith UV light

2nd irradiationwith vis. light

Visc

osity

/ m

Pa

s

Conversion / %

Figure S3. Viscosity change of suspension of 1 along with photochromism.

5. DLS Measurement.

Particle size distribution was measured on a DLS instrument Nicomp 380ZLS particle sizer

equipped with a 785 nm red laser as light source, using a fixed angle (90°) at room temperature.

The DLS measurement was carried out using the concentrated aqueous suspension showing strong

scattering (Figure 3). The autocorrelation functions were also shown in Figure S4. The particle sizes

were determined using the summation of volume-weighted size distributions. The viscosity change

(Figure S3) was not considered in this measurement.

S11

τ / ms

0

0.1

0.2

0.3

0.4

0.5

0 0.5 1 1.5 2

Inte

nsity

/ 10

6 cou

nt

τ / ms

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0 0.5 1 1.5 2

Inte

nsity

/ 10

6 cou

nt

τ / ms

UV vis.

(a) (b)

0

1

2

3

4

5

300 400 500 600 700 800

Abs

orba

nce

Wavelength / nm

0

1

2

3

4

5

300 400 500 600 700 800

Wavelength / nm

Abs

orba

nce

0

0.5

1

1.5

2

2.5

3

3.5

0 2 4 6 8 10 12 14

Inte

nsity

/ 10

6 cou

nt(c) (d) (e)

Diameter / nm1 10 100 1000

0

0.5

1.0

Rel

ativ

e A

mou

nt

Diameter / nm1 10 100 1000

0

0.5

1.0

Rel

ativ

e A

mou

nt

Diameter / nm1 10 100 1000

0

0.5

1.0

Rel

ativ

e A

mou

nt(f) (g) (h)

0

0.05

0.1

0.15

0.2

700 720 740 760 780 800Wavelength / nm

Abso

rban

ce

0

0.05

0.1

0.15

0.2

700 720 740 760 780 800

Abso

rban

ce

Wavelength / nm

Figure S4. Absorption spectral change of concentrated aqueous suspension of 1 used for DLS

measurement upon irradiation with (a) UV (313 nm) and (b) visible (578 nm) light. Insets show the

change of scattering around 800 nm. Autocorrelation function upon irradiation with UV light in the

aqueous suspension: (c) before irradiation, (d) at the PSS, and (e) after subsequent irradiation with

visible light. (f)-(h) Volume size distribution of the particles at the respective states. Upon

irradiation with UV light, autocorrelation function was changed by the increase of viscosity (Figure

S3) and the morphological transformation from sphere to fiber (Figure 8), and then restored by

irradiation with visible light. Size distribution, which was calculated by assuming that the viscosity

does not change, also shows the corresponding change. Because the photoinduced change of DLS

includes the change of the viscosity and the morphology, the change of size distribution does not

mean the simple change of particle size.

S12

6. XRD.

Small-angle and wide-angle X-ray diffractions were recorded on a Rigaku NANO-Viewer

(CuK radiation (=1.5405 Å), 40 kV, 30 mA) and Rigaku RAPID II (CuK, 50 kV, 300 mA). An

imaging plate was used to collect the X-ray intensities as 2-D images. The condition of camera

length of small-angle measurement was calibrated by a reference sample of lead stearate (d = 50.1

Å) as 1.364 m. The amorphous samples of 1b was prepared by irradiation with UV (313 nm) light

to acetonitrile solution of 1a and subsequent evaporation. The dehydrated amorphous 1b was put

between two nearly orthogonal strips of Kapton film (ca. 1 x 1 cm) and located on the X-ray optical

path. The hydration of 1b was achieved by adding a drop of water, kneading by spatula, and drying

gradually in the dark (Figures 9 and S12).

XRD data shown in Figure S13 were obtained using synchrotron at the beam line BL-4A in

KEK photon factory (18 keV, =0.69 Å ) with thin glass plate. The condition of camera length was

calibrated by a reference sample of silver behenate (d = 58.5 Å) as 0.85 m.

7. TEM Microscopy.

TEM microscopy was performed with a JEOL JEM-1400 instrument. The acceleration

voltage was selected from 80, 100, and 120 kV. Samples on carbon-coated copper grids were

prepared by air-drying an aqueous suspension of diarylethene 1. Sodium phosphotungstate was used

for negative staining.

The preparation of samples in Figures 8a-d was carried out as follows: A dilute aqueous

solution of 1 having various conversion ratio was prepared, dropped on a grid (ca. 15 L), and

allowed to stand for 10 min. The grid, onto which the aggregates were adsorbed, was washed by 2

wt % aqueous solution of sodium phosphotungstate. Additional washing was carried out twice by

dropping 1 wt % solution of sodium phosphotungstate on the grid, removing the residual solution

by filter paper after 1 s, and drying at room temperature.

The preparation of samples for the "snapshot" of the photoinduced phase transition in

Figures 8e and 8f was carried out as follows: for Figure 8e, a dilute suspension of 1a was dropped

on the grid and photoirradiation was performed using low-pressure mercury lamp (AS ONE

SLUV-4, 254 nm, 8.8 mW/cm2) and washed by sodium phosphotungstate as described above; for

Figure 8f, an acetonitrile solution of 1 was irradiated with 313 nm light and then pure water was

S13

added to make the starting suspension of fiber 1b. The irradiation with visible light to the fiber

suspension was carried out using the optical microscope as described in the section of

Photochromic Reaction under Microscope. The grid was washed by sodium phosphotungstate as

described above.

8. Molecular Modeling

The supramolecular structure of closed-ring isomer 1b was modeled by a molecular

mechanics/molecular dynamics (MM/MD) approach using Materials Studio v7.0, Accelrys

Software Inc. The Dreiding force field implemented in the Forcite module was used for MM and

MD calculations. The initial geometries were inspired from TEM images. First, the closed-ring

isomers were arranged in a concentric pattern to give the model for one layer of fiber structure.

When the number of molecule was set at 19 molecules, a ring structure, in which both alkyl chains

and diarylethene core moieties are closely packed, was observed. After geometrical optimization,

the diameter of the ring structure was ca. 9 nm. Then, a fibrous geometry (consisting of 16 layers of

the observed ring structure, composed of 304 molecules) was generated as an initial structure. The

initial structure was geometrically optimized. During the optimization, the diameter of fiber

structure slightly decreased to ca. 8 nm. The resulting structure was subsequently subjected to

molecular dynamics calculations. The dynamics simulation was carried out at 298 K with NVE

ensemble. Total simulation time and dynamics time step were set at 100 ps and 1 fs, respectively.

S14

B. ADDITIONAL FIGURES

Figure S5. Temperature dependence of photoinduced macroscopic morphological transformation of supramolecular assemblies of 1. (a-b) At 5 °C and 10 °C, division occurred vigorously and the visible irradiation gave small microspheres. Red color was due to the optical filters. (c) At 15 °C, gentle division of microsphere was observed. (d-e) At 20 °C and 25 °C, division of microsphere was not observed, however the red-purple fibers were generated. The initial states indicate that the colorless microspheres, which were composed of open-ring isomer 1a, did not show transformation by temperature control. Additionally, at the region of high intensity of visible light, the decoloration of microspheres occurred rapidly and the convergence was scarcely observed.

S15

Figure S6. Morphological change of well-colored microsphere, which was mainly composed of closed-ring isomer 1b, by fast cooling from 40 °C to 15 °C. The slight decoloration at 437 s was caused by the illumination of transmitted light.

Figure S7. Morphological change of well-colored microsphere on standing after UV irradiation at room temperature. Initially, some colored microspheres were observed and this time was set to be 0 s. At 17.0 s, the green-arrowed microsphere started to deform like ellipsoid. At the time period between 20.5 s and 21.3 s, the yellow-arrowed small microsphere was squeezed out and appeared like necklace. At 110.0 s, almost spherical microsphere was observed as indicated by green arrow. At 164 s, the green-arrowed microsphere restarted to deform. At the time period between 168.3 s and 169.0 s, the pink-arrowed small microsphere was squeezed out again.

S16

Figure S8. Change of absorbance at 600, 640, and 800 nm and absorption maximum wavelength monitored upon cooling for the sample containing 5% of 1a and 95% of 1b. The values are taken from the absorption spectra shown in Figure 7(a). The clearing point temperature was obtained as 23 °C from the change of absorbance at 800 nm. When 600- and 640-nm light was employed for the monitoring, gradual decrease of absorption band was observed upon cooling from 35 °C to 23 °C and subsequent rapid decrease was observed below 23 °C. The shift of absorption maximum also changed at similar temperatures.

Figure S9. (a) Change of absorption spectra by cooling from 25 °C to 5 °C: 25 °C (black), 15 °C (dotted orange), 14 °C (yellowish green), 13 °C (green), 11 °C (sky blue), and 5 °C (dark blue). The aqueous suspension contained 67% of 1a and 33% of 1b. (b) Change of absorbance at 600, 640, and 800 nm and absorption maximum wavelength monitored upon cooling from 50 °C to 5 °C. The temperature of clearing point was obtained as 15 °C. Gradual decrease and the shift of absorption band were observed upon cooling from 26 °C to 15 °C and subsequent rapid decrease was observed below 15 °C.

S17

(a) 0% (b) 13%

(c) 22% (d) 28%

(e) 45%(g) 83%(f) 58%

Wavelength / nm

Wavelength / nm

Wavelength / nm

Wavelength / nm

Wavelength / nm

Wavelength / nm

Abso

rban

ceAb

sorb

ance

Abso

rban

ce

Abso

rban

ce

Abso

rban

ceAb

sorb

ance

Abso

rban

ce

Figure S10. Change of absorbance at 600, 640, and 800 nm and absorption maximum wavelength monitored upon cooling for the samples with different open/closed ratios of photoisomerization. The content of the closed-ring isomer 1b was (a) 0%, (b) 13%, (c) 22%, (d) 28%, (e) 45%, (f) 58%, and (g) 83%.

S18

Figure S11. The effect of evaporation of solvent water. (a) The pale-blue sphere of the open-ring isomer 1a gathered at the edge and finally the sphere became amorphous by drying. (b) UV-irradiated mixture of 1a and 1b which showed red–purple color in water turned blue by evaporation of water. In 172 and 184 s, the left side kept red-purple color, but the right side turned blue. It means that the ratio of the closed-ring isomer 1b was still high to keep fiber structure in spite of illumination with weak observation light through optical microscope. However, the removal of water caused random orientation of the closed-ring isomers. Similar disassembly of the fiber structure of the closed-ring isomer by removing water was observed in XRD measurement as shown in Figures 9c and d.

0

50

100

150

200

0 10 20 30 40

Inte

nsity

/ a.

u.

Scattering vector q / nm -1

Figure S12. Wide angle XRD patterns of 1a in the paste containing water.

S19

0

1000

2000

3000

4000

5000

0 1 2 3 4 5 6 7

Inte

nsity

/ a.

u.

q / nm-1

0

1000

2000

3000

4000

5000

0 1 2 3 4 5 6 7

Inte

nsity

/ a.

u.

q / nm-1

0

1000

2000

3000

4000

5000

0 1 2 3 4 5 6 7

Inte

nsity

/ a.

u.

q / nm-1

0

1000

2000

3000

4000

5000

0 1 2 3 4 5 6 7

Inte

nsity

/ a.

u.

q / nm-1

0

1000

2000

3000

4000

5000

0 1 2 3 4 5 6 7

Inte

nsity

/ a.

u.

q / nm-1

(a) 0%

(b) 34%

(c) 61%

(d) 88%

(e) blank

d = 3.4 nm

2.2 nm

Figure S13. Synchrotron small angle XRD patterns of the samples with different open/closed ratios of photoisomerization. The content of the closed-ring isomer 1b was (a) 0%, (b) 34%, (c) 61%, and (d) 88% in the paste including small amount of water. (e) Blank cell composed of only thin glass and water. The diffraction images are shown in the right column. The background scattering at the region of small angle was caused by the glass cell and sample paste. The XRD patterns were obtained by subtraction of simulated background scattering from the raw data. The simulation was carried out with an assumption that the size of large domains were approximated as gamma distributed spheres around a few ten nanometers and these parameters were determined by fitting to the base line of raw data.[S3]

S20

Figure S14. Additional TEM images of the diarylethene assembly for the samples with different open/closed ratios of photoisomerization. The content of the closed-ring isomer 1b was (a) and (b) 0%, (c) 17%, (d) and (e) 48%, (f) and (g) 64%. Only (a) was positive-stained by RuO4 and others were negative-stained by sodium phosphotungstate. For (g), brightness was controlled for the visibility.

S21

References

[S1] Vasconcelos, U. B.; Schrader, A.; Vilela, G. D.; Borges, A. C. A.; Merlo, A. A. Tetrahedron

2008, 64, 4619-4626.

[S2] Hagiwara, T.; Murano, Y.; Watanabe, Y.; Hoshi, T.; Sawaguchi, T. Tetrahedron Lett. 2012, 53,

2805-2808.

[S3] Chu, B.; Liu, T. J. Nanopart. Res. 2000, 2, 29–41.