Supporting Information Control of photomechanical crystal twisting by illumination direction Daichi Kitagawa, Hajime Tsujioka, Fei Tong, Xinning Dong, Christopher J. Bardeen,*, and Seiya Kobatake*, Department of Applied Chemistry, Graduate School of Engineering, Osaka City University, 3-3- 138 Sugimoto, Sumiyoshi-ku, Osaka 558-8585, Japan. Department of Chemistry, University of California, 501 Big Springs Road, Riverside, CA 92521, USA. *E-mail: kobatake@a-chem.eng.osaka-cu.ac.jp; christopher.bardeen@ucr.edu S1
General The photoinduced crystal twisting behavior of diarylethene crystals was observed using a Keyence VHX-500 digital microscope. UV irradiation was conducted using a Keyence UV-LED UV-400 with an attached UV-50H head (365 nm light). The light power on the crystal surface was measured using a Neoark PM-335A power meter. Visible light irradiation was conducted using a halogen lamp (100 W). Powder X-ray diffraction profile was recorded on a Rigaku MiniFlex 600 diffractometer employing CuK radiation ( = 1.54184 Å). Material Diarylethene 1a was synthesized according to a procedure described in the literature. S1 S2
Table S1. Crystal data for diarylethene 1a. S1 empirical formula C27H18F6S2 formula weight 520.53 temperature/k 296(2) crystal system monoclinic space group P21/c Z 8 a/å 18.786(2) b/å 11.762(1) c/å 21.675(3) /deg 96.323(2) V/Å 3 4759.9(10) density calcd /g cm 3 1.453 goodness-of-fit on F 2 1.023 R1 [I > 2 (I)] 0.0485 wr2 for all data 0.1516 S3
Figure S1. Magnified image of photomechanical crystal twisting. Scale bar is 300 m. Figure S2. Magnified images of photomechanical twisting in different direction depending on the angle of the incident light. Scale bar is 300 m. UV irradiation was carried out from (a) upper back to front, (b) lower back to front, (c) upper front to back, and (d) lower front to back, respectively. The twisting direction was determined by following the focus of the crystal edge. S4
Figure S3. Magnified images of different twisting motion, a helicoid and a cylindrical helix, depending on the angle of the incident light. Scale bar is 300 m. UV irradiation was carried out from left side on the image. The angle of the incident light respect to the crystal in each panel is (a) 0, (b) 12, (c) 22, (d) 34, (e) 45, (f) 55, (g) 74, and (h) 90, respectively. S5
Mechanism of Helical Deformation Resulting from Different Light Directions Figure S4. Schematic illustration of size change due to 1a 1b photoisomerization. Note that this is a simplified picture in two dimensions. The photoisomerization generates stress along two different molecular axes, one compressive and one expansive. Figure S5. Top view of ribbon illuminated by light incident at a 90 angle to the bc face. Dashed red lines show approximate alignment of the 1a transition dipole moments for the two molecular orientations in the crystal. The polarization of the incident light is perpendicular to the direction of propagation, lying along the b and c crystal axes, and can excite both molecules equally. S6
a b hν Figure S6. Because 90 excitation isomerizes molecules with both orientations, it generates stress vectors that add together and generate net stress in directions both parallel and perpendicular to the ribbon axis (red arrows show compressive and expansive components). If the stress parallel to the long (b) axis is larger, it will generate bending, as observed. If the short axis (a) stress is larger, it would generate curling. The actual motion will be a mixture of both deformations. Figure S7. Top view of ribbon illuminated by light incident at a shallow angle with respect to the bc face. Dashed red lines show approximate alignment of the 1a transition dipole moments. The polarization of the incident light is perpendicular to the direction of propagation, and the second layer of molecules, oriented close to parallel to the direction of propagation and perpendicular to the light polarization, can no longer be efficiently excited. S7
hν a b Figure S8. Because shallow excitation isomerizes molecules with only one orientation, it generates stress vectors that are rotated with respect to the ribbon axis (red arrows show compressive and expansive components). This diagonal stress tensor will generate a twisted helicoid, as opposed to bending or cylindrical helix, as outlined in ref. S2. Figure S9. Illustration of two types of twisting: helicoid and cylindrical helix. S8
Figure S10. Different strain tensors on a ribbon surface lead to different motions. Reprinted with permission from ref. S3, Copyright 2015 American Chemical Society. It is generally assumed that the shape of the stress tensor follows the strain tensor. The twisting in panels d) and e) corresponds to the case of compressive and expansive stresses, rotated at a 45 angle with respect to the crystal axis. As the stress tensor becomes more closely aligned with the ribbon axis (panels f) and g)), the twist is replaced by a cylindrical coiling. In the limit where the stresses are perfectly aligned with the ribbon (the initial square is transformed to a rectangle) only bending or curling would be observed. S9
Preparation of ribbon crystals of 1a by a sublimation method A vial containing several tens of milligrams of powder crystals 1a was heated up for 1 week at 135 C. The ribbon crystals were obtained on the inner wall of the vial as shown below. Caption of videos Video S1. Photomechanical crystal twisting of the ribbon crystal of 1a in a cylindrical helix. UV irradiation was carried out from left side on the image. The movie is fast-forwarded as much as 4 times. Video S2. Photomechanical crystal twisting of the ribbon crystal of 1a in a helicoid. UV irradiation was carried out from left side on the image. The movie is fast-forwarded as much as 4 times. S10
Reference (S1) Irie, M.; Lifka, T.; Kobatake, S.; Kato, N. J. Am. Chem. Soc. 2000, 122, 4871 4876. (S2) Chen, Z.; Majidi, C.; Srolovitz, D. J.; Haataja, M. Appl. Phys. Lett. 2011, 98, 011906. (S3) Naumov, P.; Chizhik, S.; Panda, M. K.; Nath, N. K.; Boldyreva, E. Chem. Rev. 2015, 115, 12440 12490. S11