Research 1..10

Home , Spontaneous recovery

Research Article

www.acsami.org

1 Camera Method for Monitoring a Mechanochromic Luminescent

2 β‑Diketone Dye with Rapid Recovery †,‡ ,† 3 Tristan Butler, Alexander S. Mathew, Michal Sabat, and Cassandra L. Fraser*

† 4 Department of Chemistry, University of Virginia, Charlottesville, Virginia 22904, United States ‡ 5 Department of Materials Science and Engineering, University of Virginia, 395 McCormick Road, Charlottesville, Virginia 22904, 6 United States

7 *S Supporting Information

8 ABSTRACT: Mechanochromic luminescent (ML) materials, which show a change in emission due to an applied mechanical 9 stimulus, are useful components in a variety of applications, including organic light-emitting diodes, force sensors, optical 10 memory storage, and next-generation lighting materials. While there are many different ML active derivatives, few show room 11 temperature self-erasing. Thin films of the methoxy substituted β-diketone, gbmOMe, initially exhibited blue (428 nm) emission; 12 however, green (478 nm) emission was observed after smearing. The mechanically generated smeared state recovered so rapidly 13 that characterization of its emission was difficult at room temperature using traditional luminescence techniques. Thus, a new 14 complementary metal oxide semiconductor camera imaging method was developed and used to calculate the decay time of the fi 15 mechanically generated smeared state (i.e., smeared-state decay; τSM) for gbmOMe thin lms. Additionally, this method was used 16 to evaluate substrate and film thickness effects on ML recovery for glass and weighing paper films. The recovery behavior of 17 gbmOMe was largely substrate-independent for the indicated matrixes; however, thickness effects were observed. Thus, film 18 thickness may be the main factor in determining ML recovery behavior and must be accounted for when comparing the recovery ff ° 19 dynamics of di erent ML materials. Moreover, when heated above the melting point (Tm = 119 C), bulk gbmOMe powders 20 assumed a metastable state that eventually crystallized after a few minutes at room temperature. However, melted thin films 21 remained in an amorphous state indefinitely despite annealing at different temperatures (50−110 °C). The amorphous phase was 22 identified as a supercooled liquid via changing the rate of cooling in differential scanning calorimetry thermograms. 23 KEYWORDS: mechanochromic luminescence, self-healing, camera RGB image analysis, supercooled liquids

24 ■ INTRODUCTION The methoxy-substituted dinaphthoylmethane derivative 40 ﬂ (dnmOMe) showed solvatochromism and AIE in addition to 41 25 Di uoroboron β-diketonate (BF2bdk) materials are noted for 18 high contrast ML with rapid spontaneous recovery. 42 26 their unique optical properties in both solution and the solid 1−3 Furthermore, because β-diketones, including methoxy-substi- 43 27 state. Many BF2bdk derivatives show room temperature tuted derivatives, show large molar absorptivities and photo- 44 28 phosphorescence (RTP) in poly(lactic acid) (PLA) matrixes 4−6 stabilities, they have been screened and used as UV absorbing 45 29 and can be used to quantify oxygen in biological systems. 19 additives in sunscreens. 46 30 Additionally, BF2bdk materials exhibit mechanochromic 7 31 luminescence (ML) with spontaneous recovery as well as Successful integration of bdk dyes into commercial 47 8 9 − 32 aggregation induced emission (AIE) and solvatochromism. applications requires a thorough understanding of structure 48 33 These diverse properties coupled with their straightforward property relationships and material processing parameters (e.g., 49 34 synthesis make this family of boron dyes potentially useful for temperature, thickness, and substrate). Prior reports describe 50 ﬀ 7,20 21 35 numerous applications such as mechanical sensors, next- the e ects of alkyl chain length, halide substitution, arene 51 36 generation light sources, display technologies, security inks, 10−16 37 and biological probes. Similar to their boronated counter- Received: February 10, 2017 38 parts, unique solution and solid-state optical properties have Accepted: April 17, 2017 17 39 also been observed in boron-free β-diketone (bdk) materials.

22 52 size, and even BF2 coordination on the optical and room solubility of various active pharmaceutical ingredients and 108 18 39 53 temperature self-healing properties of ML materials. Sub- excipients, supercooled liquids are relatively rare. 109 54 stitution of stimuli responsive materials with different donor Compared to temperature, relatively little is known about 110 55 and acceptor groups is also an common strategy for modulating how film thickness, dye−substrate interactions, and repeated 111 56 solid-state ML properties such as emission wavelength, smearing affect ML recovery properties. This may be due in 112 23−26 57 intensity, and spontaneous recovery. For example, alkoxy part to the relative scarcity of RT self-healing systems. 113 58 substitution modulates solution and solid-state optical proper- Furthermore, certain fast recovering ML dyes such as diketones 114 fl 59 ties through electronic properties and control of molecular are incompatible with standard room temperature uorimetry, 115 27−30 60 packing. Recently, Sket et al., showed that the ML active, given they recover faster (seconds to minutes) than the time 116 117 61 methoxy-substituted BF2bdk 1-phenyl-3-(3,5-dimethoxyphen- required to record a spectrum. On the other hand, slow self- 62 yl)-propane-1,3-dione can be crystallized as uniquely emissive erasing ML systems are inconvenient to monitor given intensity 118 63 polymorphs whose formation is partially governed by the changes little over long time periods, and subtle substrate and 119 31 fi ff 64 different conformations of the methoxy substituents. Addi- lm thickness e ects could be imperceptible. Additionally, 120 65 tionally, selective methoxy-substitution can induce intra- standard characterization often averages emission intensity over 121 66 molecular charge transfer (ICT) in certain bdk dyes, which relatively large sample regions and does not account for 122 ff 67 has been correlated with the solvatochromic and phosphor- localized di erences in ML recovery. Alternatively, techniques 123 18,32 68 escent properties. Literature precedent combined with the that excite locally do not provide good spatial resolution of ML 124 69 abundance and variety of commercially available starting recovery processes. To address some of these challenges and 125 126 70 materials makes methoxy substitution a good strategy for obtain 2D spatiotemporal information on ML recovery as a 127 71 tuning the optical and ML properties of bdk materials. Thus, in function of substrate, thickness, and smearing, we developed a video camera method to monitor the intensity decay of 128 72 this study, we employed methoxy substitution to modulate the individual pixels over time. Previously, we employed this 129 73 emission and self-recovery properties of a bdk dye. Though the camera imaging technology to track photostability, oxygen- 130 74 tetramethoxy-substituted dibenzoylmethane, 3-hydroxy-3-(4- 40 6 dependent lifetime, and intensity of BF bdkPLA phosphor- 131 75 methoxyphenyl)-1-(3,4,5-trimethoxyphenyl)prop-2-en-1-one 2 escent materials. This method not only allows for investigation 132 76 (gbmOMe) was synthesized and evaluated previously as a 19 of recovery dynamics of gbmOMe thin films averaged as a 133 77 sunscreen additive, the stimuli responsive solid-state proper- whole but also provides a spatially resolved means of evaluating 134 78 ties of gbmOMe have yet to be reported. localized recovery effects associated with the smearing process. 135 Structural and stimuli responsive properties of gbmOMe were 136 also investigated using powder and single crystal X-ray 137 diffraction, differential scanning calorimetry (DSC), and 138 through excitation and emission spectra recorded at room 139 temperature and 77 K. 140 − 79 Material processing parameters such as temperature, dye ■ EXPERIMENTAL DETAILS 141 80 substrate interactions, film thickness, and repeated smearing Materials. Solvents THF and CH2Cl2 were dried over molecular 142 81 fl 41 also in uence ML properties. The thermal properties of sieves activated at 300 °C as previously described. Reactions were 143 82 luminophores are important both for fabrication and because monitored using silica TLC plates. Compounds purchased from 144 83 the mechanically triggered changes in optical properties are Sigma-Aldrich were reagent grade and used without further 145 84 often the result of temperature-dependent crystalline-to- purification. The β-diketone 3-hydroxy-3-(4-methoxyphenyl)-1- 146 21,33 85 amorphous phase transitions. For example, amorphous (3,4,5-trimethoxyphenyl)prop-2-en-1-one (gbmOMe) was synthesized 147 2 as previously described. 148 86 melt quenched thiophene and furan substituted BF2bdk Methods. 1H NMR (600 MHz) spectra were recorded in dilute 149 87 derivatives showed red-shifted emission compared to their 34 CDCl using a Varian VRMS/600. Spectra were referenced to the 150 88 emission in the crystalline state. Thus, it is possible to access 3 signals for residual protio-CDCl3 at 7.26 ppm and coupling constants 151 89 the amorphous phase of these materials via thermal or were recorded in Hz. UV−vis spectra were collected on a Hewlett- 152 90 mechanical stimulation. Similar optical properties were Packard 8452A diode-array spectrophotometer. Steady-state fluores- 153 91 observed for a triphenyl amine (TPA)-based material reported cence emission and excitation spectra were obtained on a Horiba 154 92 by Mizuguchi et al., which showed rapid ML recovery of a Fluorolog-3 Model FL3-22 spectrofluorometer (double-grating 155 93 green-yellow emissive state that was produced via melting and excitation and double-grating emission monochromator). Time- 156 35 fl 94 cooling to room temperature in addition to smearing. The correlated single-photon counting (TCSPC) uorescence lifetime 157 measurements were performed with a NanoLED-370 (λ = 369 nm) 158 95 identity of this emissive melted state was determined to be a ex excitation source and a DataStation Hub as the SPC controller. 159 96 supercooled liquid with red-shifted emission resulting from a Lifetime data were analyzed with DataStation v2.4 software from 160 97 twisted intramolecular charge transfer (TICT) state. Addition- Horiba Jobin Yvon. Solid-state quantum yield measurements were 161 98 ally, Kim et al. observed a high contrast change in emission for acquired using a F-3029 Quanta-Φ Integrating Sphere purchased from 162 99 an alkylated diketopyrrolopyrrole due to shear-triggered Horiba ScientificandanalyzedusingFluorEssencesoftware.163 36 100 crystallization of a thermally stable supercooled liquid state. Fluorescence quantum yields, φF, in CH2Cl2 were calculated versus 164 101 a dilute quinine sulfate solution in 0.1 M H2SO4 as a standard using a 165 The phase transition was attributed to a small Gibbs free energy 42 ff previously described method and the following values: φF quinine 166 102 di erence between the crystalline and the supercooled liquid 43 44 45 sulfate in 0.1 M sulfuric acid = 0.54; n 0.1 M H SO = 1.33; n 167 103 state. Such changes in emission were even activated by small D 2 4 D CH2Cl2 = 1.424. Optically dilute CH2Cl2 solutions of all samples were 168 104 forces associated with live cell attachment. This example prepared in 1 cm path length quartz cuvettes with absorbances <0.1 169 105 demonstrates an application for thermally stable amorphous (au). ML emission was observed by applying a constant, gentle 170 106 solids, namely, as cell force sensors. Yet, despite their potential pressure via a cotton swab or a Kimwipe against glass and weigh paper 171 37,38 107 for optical and optoelectronic uses and for improving the films. Powder XRD patterns were obtained using a Panalytical X’Pert 172

B DOI: 10.1021/acsami.7b01985 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX ACS Applied Materials & Interfaces Research Article

173 Pro MPD diffractometer operating at 40 kV and 40ma using Cu Kα However, qualitative screening of the thermal properties of 240 174 radiation. DSC was performed on the pristine powders using a TA gbmOMe not only revealed differences between bulk solid and 241 175 Instruments DSC 2920 Modulated DSC, and data were analyzed using thin film samples but also showed new phenomena. As a bulk 242 176 the Universal Analysis software v2.3 from TA Instruments. Thermo- Φ powder, gbmOMe glowed light blue (λem = 453 nm, = 5.1%) 243 177 grams were recorded using the standard mode while heating at a under UV irradiation. When the powder was heated above the 244 178 constant rate of 10 °C/min. The cooling rate was changed as indicated ff melting point, a melted (MT) viscous green emissive phase 245 179 for di erent measurements but was held constant throughout a given Φ 180 heating/cooling cycle. formed (λMT = 481 nm, = 10.0%). Crystallization of this 246 181 Crystallographic Information. Crystals for single crystal XRD were melted phase occurred within a few minutes after the heat 247 182 grown by vapor diffusion of hexanes into ethyl acetate. Data collection source was removed to yield a blue-emissive solid. However, 248 183 was performed using a Bruker Kappa Duo CCD diffractometer at similar crystallization was not observed for melted thin films of 249 − ° 184 120 CwithMoKα radiation. Crystal data for gbmOMe: gbmOMe. Instead, the green emissive state persisted upon 250 185 orthorhombic space group Pbca; a = 17.7840(19); b = 7.5472(8); c cooling. Powder XRD patterns further confirmed that bulk 251 186 = 7.5472(8) Å; β = 90°, Z = 8, V = 3393.2(6) Å3. The structure was gbmOMe is crystalline and MT thin films are amorphous, as no 252 187 solved by the charge flipping method of the Bruker SHELXTL ff fi fl peaks were observed in the di raction pattern for the latter 253 188 program and re ned to an R = 0.0475 using 5976 re ections with I > ff σ samples (Figure S2). This relative di erence in thermal stability 254 189 2 (I). fi 190 Preparation of Thin Films. Bulk powder films for ML character- of the MT phase for bulk powder and thin lms may indicate a 255 191 ization were fabricated on weigh paper substrates by smearing ∼2 mg thickness dependence on the crystallization rate of melted 256 192 of sample across the entire surface with a Kimwipe. Drop cast films gbmOMe. 257 193 were prepared by adding 20 drops of gbmOMe/THF solutions (0.090 The thermal properties of gbmOMe were further inves- 258 ″ × ″ 194 and 0.018 M) to 3 1 weigh paper and glass microscope slides. tigated using differential scanning calorimetry. The sample was 259 fi 195 Spin-cast lms were made using a Laurel Technologies WS-65OS spin- heated at a constant rate of 10 °C/min while the rate of cooling 260 196 coater and applying 20 drops to glass microscope slides rotating at was varied for each cycle (Figure 1). Each cycle was performed 261 f1 197 3000 rpm. Both spin-cast and drop cast samples were dried under 198 vacuum for 20 min prior to annealing at 75 °C for 10 min. Green 199 emission remained after annealing, so films were gently smeared with a 200 Kimwipe followed by annealing at 75 °C for another 10 min to 201 produce uniform blue emission. 202 Smeared State Decay Measurements. Spontaneous smeared 203 emission decay measurements were performed at room temperature 204 using a PGR GS3-U3-41C6C-C video camera with a complementary 205 metal oxide semiconductor chip capable of 90 frames per second 206 (FPS) at a maximum resolution of 2048 × 2048 pixels. The camera 207 was also equipped with a Spacecom f/0.95 50 mm lens and an 208 Edmund Optics 425 nm long pass filter to minimize excitation 209 background. The camera was operated with a Lenovo W530 laptop 210 connected via a USB 3.0 cable. Point Gray FlyCap2 software was used 211 to record videos of smeared glass and weigh paper films, and the data 212 were analyzed using a custom MATLAB 2014b program. The camera 213 was placed approximately 0.5 m above the sample, which was 214 illuminated with a 100W Black-Ray B-100AP/R lamp at 365 nm 215 located ∼10 cm above the sample. Videos of recovering samples were 216 recorded using a frame rate of 1 frame per second (fps), and the 217 intensity of the green channel was monitored over time. The initial 218 intensity of the green channel due to sample emission prior to 219 smearing was subtracted from each pixel before fitting data to a double 220 exponential decay, and the pre-exponential weighted lifetime was 221 reported. Photostability measurements were performed using the same 222 experimental setup as recovery measurements and irradiating the 223 sample continuously for 4 h while monitoring the pixel intensity every 224 5 min.

225 ■ RESULTS AND DISCUSSION Figure 1. (a) DSC thermograms of consecutive heating/cooling cycles 226 Optical Properties in Solution. β The -diketone gbmOMe of gbmOMe with varying cooling rates (ramp rate: 10 °C/min). (b) 227 was synthesized via Claisen condensation of ester and ketone ff 2 DSC thermograms of gbmOMe cooled to di erent minimum 228 building blocks as previously described. The optical properties temperatures (ramp rate: 10 °C/min). 229 of gbmOMe were measured in dichloromethane solution 230 (Figure S1, Table S1). UV−vis spectra reveal an absorption −1 231 peak (λabs) at 365 nm with a molar absorptivity of 64 000 M in succession to probe the thermal history. Thermograms show 262 −1 232 cm . Though the solution appeared nonemissive to the eye that gbmOMe melts at 119 °C regardless of cooling rate. When 263 233 due to a very low quantum yield (Φ = 0.01%), a peak was cooled faster than 1 °C/min, crystallization was not observed 264 234 detectable at λem = 427 nm in the steady-state emission for the duration of the cooling cycle; however, crystallization 265 235 spectrum. was observed upon subsequent heating. When the rate was slow 266 236 Solid State Thermal Properties. To achieve high color enough (1 or 0.5 °C/min), crystallization could be seen during 267 237 contrast, mechanochromic luminescent materials are typically the cooling cycle. These results indicate that the difference in 268 238 processed by thermal annealing to produce an ordered emissive thermal stability of the green amorphous state of gbmOMe as a 269 239 state. Smearing then generates a red-shifted amorphous state. bulk melt compared to melted thin films could be due to 270

C DOI: 10.1021/acsami.7b01985 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX ACS Applied Materials & Interfaces Research Article

fi Figure 2. Images of gbmOMe thin lms (λex = 365 nm) (a) on glass in the melted phase (MT) under ambient (left) and UV light (right) at room temperature; (b) on glass in AS, TA, and SM states at room temperature; and (c) on WP films in AP and MT states. Emission spectra of gbmOMe fi lms (λex = 369 nm) on (d) WP and glass at room temperature and (e) WP at 77K.

271 differences in the cooling rates when samples are quenched on glass were crystalline, and AS films were amorphous (Figure 311 fi 272 (i.e., cooled) in air. S2). Smearing of TA lms produces green emission (λSM = 478 312 273 To gain further insight into the identity of the green emissive nm) that self-recovers over the course of a few minutes (Figure 313 274 state, successive DSC scans of gbmOMe were run while varying 2, Table S2). 314 275 the minimum cooling temperature from 25 to −20 °C(Figure Mechanochromic luminescence properties of gbmOMe were 315 276 1b). When a relatively fast constant rate of heating/cooling (10 also investigated on weighing paper for ready comparison with 316 21,22 277 °C/min) was used to prevent crystallization of gbmOMe, all many previous studies. Thin films were prepared by 317 278 scans that were cooled to a sufficiently low temperature showed smearing a small amount of gbmOMe (∼2 mg) across a 318 ° fi 279 a glass transition temperature (Tg =2 C). Combined with the piece of weighing paper. The as prepared (AP) WP lm glowed 319 280 cooling rate-dependent crystallization, the presence of a clearly blue (λAP = 428 nm) under UV exposure, and smearing 320 fi 281 de ned Tg could indicate a supercooled liquid state for rapidly produced green emission that rapidly regained its original blue 321 ∼ 282 cooled gbmOMe above Tg that becomes a glass at lower color ( 30 s). Unlike most other ML samples, thermal 322 283 temperatures. annealing is not required to produce the maximally blue-shifted 323 284 Mechanochromic Luminescence. In addition to thermal state. Recovery happens rapidly and spontaneously. Recovery of 324 285 responsiveness, changes in solid-state emission of gbmOMe smeared WP films was so rapid that an emission spectrum in 325 286 thin films can also be produced mechanically. Thin films of the SM state could not be obtained at room temperature given 326 287 gbmOMe were investigated on both glass (G) and weighing that the green smeared emission disappeared faster than the 327 f2 288 paper (WP) substrates (Figure 2, Table S2). Glass films were time required to perform the measurement. 328 289 fabricated by spin coating dilute (0.018 M) dye/THF solutions The self-healing properties of BF2bdk ML materials have 329 290 on microscope coverslips. In the as spun (AS) state (i.e., after been previously ascribed by Ito et al., to the thermal back 330 46 291 solvent evaporation), gbmOMe films on glass were locked into population from the amorphous phase to the crystalline state. 331 fi 292 a transparent state that glowed green (λAS = 499 nm) under UV In order to slow the rapid recovery of gbmOMe WP lms and 332 293 light and qualitatively resembled the melt phase. This observe them in the SM state, the optical properties of WP 333 294 transparency is atypical given that as spun films are often films were measured at 77K (Figure 2e, Table S3). Films in the 334 7,21 295 opaque. Typical thermal annealing of BF2bdk or bdk SM state were produced by rubbing the sample, followed by 335 296 samples involves heating below the melting temperature immediate submersion in liquid N2. Excitation spectra were also 336 297 followed by cooling to produce a maximally blue-shifted state. recorded to probe the different emissive species that exist in 337 298 However, this process was insufficient to anneal gbmOMe AP, MT, and SM weigh paper films (Figure S3). The peak 338 fi fi 299 lms, which showed no change in emission and remained in the emission of smeared lms (λSM = 454 nm) at 77K falls between 339 fi 300 transparent green emissive state after heating and cooling. the slightly red-shifted peak emission of melt lms (λMT = 463 340 301 Instead, gbmOMe films on glass were heated at 75 °C (below nm) and the blue-shifted peak emission of films in the AP state 341 ° 302 Tm = 119 C) for 10 min, followed by cooling to room (λAP = 429 nm). The relative proximity in peak emission and 342 303 temperature, then gentle smearing with a Kimwipe. After ∼1 h, their similarly broad emission profiles indicate that the smeared 343 304 blue emissive films formed. To speed up the processing, after and melt states of WP films contain similar emissive species. 344 305 smearing, films were again heated at 75 °C for 10 min followed However, a shoulder corresponding with the emission of AP 345 306 by cooling to produce an opaque, blue emissive, thermally films is also visible in the spectrum of the smeared film. This 346 307 annealed (TA) state (λTA = 428 nm) (Figure 2d). It is possible suggests that SM gbmOMe samples are comprised of both 347 308 that perturbation of the amorphous phase via smearing induced green amorphous and blue crystalline emissive species. This is 348 309 gbmOMe nucleation to form the crystalline TA phase. further supported by the excitation spectra monitored at the 349 310 According to their corresponding XRD pattern, TA thin films peak emission of each film. The excitation spectrum of the SM 350

D DOI: 10.1021/acsami.7b01985 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX ACS Applied Materials & Interfaces Research Article

351 sample contains features of AP and SM films. Incomplete interactions (2) are also observed between hydrogen atoms 388 352 conversion of the sample to the amorphous state upon of methoxy groups on the monosubstituted ring and the ketone 389 353 smearing or rapid partial recovery before immersion in liquid moiety of neighboring molecules (distance: 2.667 Å) (Figure 390 354 nitrogen may explain the presence of blue and green emissive 3b). Examination of the unit cell shows that gbmOMe 391 355 states in the smeared WP film. On the basis of the optical molecules are arranged in a herringbone configuration with 392 356 characterization, the mechanically produced color changes no evidence of face-to-face or offset dimers often observed for 393 20,47 357 observed in gbmOMe films on both glass and WP substrates ML active bdk-based materials. The intermolecular distance 394 358 appear to be produced by the same emissive species. was estimated between neighboring dyes using centroids 395 359 Structural Characterization. X-ray diffraction studies were calculated for the trimethoxy substituted rings of each molecule. 396 360 performed to gain insight into crystallinity and molecular As evidenced by the large intermolecular distance (5.334 Å), 397 361 packing. As previously described, powder XRD patterns of bulk π−π interactions were not observed. This may be the result of 398 362 gbmOMe and thermally annealed (TA) glass films indicate that out-of-plane methoxy substituents in gbmOMe preventing π- 399 363 crystalline species produce blue emission and that green stacking and other strong associations linked with dimer 400 48,49 364 emission results from amorphous states (i.e., MT and AS). To formation in similar bdk systems. These groups may also 401 365 further investigate the solid-state emission and intermolecular play a role in more rapid recovery. 402 366 interactions, single crystals were grown by vapor diffusion of Camera Characterization of Mechanochromic Lumi- 403 367 hexanes into a concentrated EtOAc solution of gbmOMe. The nescence Recovery. The ML behavior of glass and weigh 404 368 emission spectrum of the crystals is similar that of other paper films is similar; however, there is a large difference in 405 369 material forms with blue emissive states; however, the peak their recovery times. All samples show blue to green ML, but 406 370 emission of the crystal (λC = 452 nm) is red-shifted relative to the smeared emission of WP films vanishes much more quickly 407 fi 371 emission of WP lms at room temperature (λAP = 429 nm). compared to that of smeared samples on glass. Film thickness, 408 372 The difference in emission profiles could indicate the presence substrate, or other fabrication effects may be involved. In fact, 409 373 of multiple phases in WP films (Figure S4). the rapid recovery of gbmOMe provides a good handle for 410 374 According to the crystal structure, gbmOMe adopts a mostly investigation given the intensity of the SM emission changes on 411 375 planar conformation. However, the methoxy group in the 4- a convenient time scale for measurement. 412 376 position of the trimethoxy substituted phenyl ring is out of A camera method was developed to monitor intensity 413 377 plane, which can be attributed to steric crowding by the changes during mechanochromic luminescence recovery in 414 f3 378 neighboring methoxy substituents (Figure 3). Despite the gbmOMe films. Specifically, a video was recorded immediately 415 after smearing, and the intensity decay of the smearing induced 416 color changes was monitored for each pixel. The green channel 417 of the camera was used, given significant overlap between its 418 quantum efficiency and the emission profile of the smeared 419 state. Furthermore, the green channel specifically captures the 420 smeared state given that emission from the thermally annealed 421 state does not tail into this region (Figure S5). The smeared- 422 state decay for each pixel (i.e., recovery lifetime, τR) was 423 determined using Matlab by fitting the intensity to a double 424 exponential decay and calculating the pre-exponential weighted 425 lifetime (Figures S6 and S7). Using the τR for each pixel, a 426 spatially resolved colormap was generated for each video, 427 depicting the smeared region decay process on glass and WP 428 substrates (Figure 4a). By computing the mean τR across a 429 f4 region of interest in the colormap, calculation of the average 430 smeared-state lifetime (τSM) is possible, and processes on WP 431 and glass samples may be compared. 432 To validate the camera method, it is important to consider 433 other decay pathways that could affect the intensity of emission. 434 For example, continuous UV irradiation for the duration of the 435 Figure 3. (a) Crystal structure of gbmOMe from top (left) and side − − − video recording could result in photobleaching. To ensure that 436 (right) view. (b) View highlighting C O···H Ar and C H···O=C fi interactions in gbmOMe crystals (interaction distances: (1) m-OMe: the photodegradation of gbmOMe does not signi cantly 437 contribute to the intensity decay, a video of gbmOMe WP 438 2.711 Å; p-OMe: 2.416 Å; (2) 2:2.667 Å) and intermolecular distances fi (green line; 5.334 Å). (c) Unit cell of gbmOMe. lms was recorded under constant UV illumination. The 439 intensity change of the green channel was measured for films in 440 AP and MT states given both crystalline and amorphous species 441 379 presence of a hydroxyl group in gbmOMe, no intermolecular are present during the smearing process (Figure 4b). For the 442 380 hydrogen bonding is observed. In fact, the only interactions photostability control experiment, films were irradiated for over 443 381 influencing the crystal packing are C−H···O−C, C−H···arene, 3 h, which is much longer than gbmOMe recovery on both WP 444 382 and weak C−H···H−C van der Waals interactions. For and glass substrates. Films typically recover (i.e., show little 445 383 example, multiple C−O···H−Ar interactions (1) between change in emission) within seconds (WP) or a few minutes 446 384 oxygen atoms of meta-(m-OMe) and para- (p-OMe) methoxy (glass) after smearing. During the course of the 3 h 447 385 groups of the trisubstituted arene ring and aryl hydrogens of an photostability measurement, the intensity of crystalline AP 448 386 adjacent dye are observed (interaction distances: m-OMe: films decreased by ∼20%, whereas amorphous MT films 449 387 2.711 Å, p-OMe: 2.416 Å) Additionally, C−H···O=C showed only a slight decrease in intensity (5−10%) over the 450

E DOI: 10.1021/acsami.7b01985 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX ACS Applied Materials & Interfaces Research Article

was calculated for up to ten smearing/recovery cycles (Figure 483 f5 5). 484 f5

Figure 4. (a) Initial frame from the video of a smeared gbmOMe glass film (left) and the corresponding colormap for the smeared intensity decay process (right). The colormap was generated from the pre- exponential weighted lifetimes of the double exponential fit associated with the intensity decay of the green channel for each pixel (τR). A region of interest (black rectangle) was defined to calculate the average lifetime of the smeared state (τSM). (b) Photostability experiments showing intensity over time of continuously illuminated weigh paper films in the AP and MT states (λ = 365 nm). Figure 5. Smeared-state decay (τSM) of repeatedly smeared gbmOMe ex fi ″ × ″ thin lms on (a) 3 1 weigh paper and (b) glass substrates (τSM = smeared-state decay (τR, per pixel) averaged over a region of interest).

451 same time period. Inspection of photostability data for shorter With the exception of G3, most films show comparable 485 452 time frames typically used to monitor smeared emission decay smeared-state decays (τSM < 100) regardless of substrate and 486 ff fi fl 453 (i.e., minutes) revealed that intensity di erences were smearing cycle. Data for WP lms shows that τSM uctuates 487 454 imperceptible. These results indicate that photodegradation is randomly between 32 and 52 s regardless of the fabrication 488 455 minimal and should not interfere with methods utilizing method (Figure 5a). This fluctuation may simply be due to the 489 456 intensity to calculate smeared emission decay, τSM. noise that can be expected in the fabrication and smearing 490 ff ff 457 Sample thickness e ects were investigated by three di erent process. To assess this, the τSM values for all WP samples were 491 fi ff 458 approaches. Namely, lms were prepared by di erent common averaged (τSM̅ = 41.2 s), and the standard deviation was 492 459 methods (i.e., surface application of the solid, drop casting, and computed to be σSM = 6.0 s. Comparison of τSM data for G1 493 460 spin-casting), by varying the sample loading within a given and G2 reveals similar smeared-state decays though G2 and in 494 461 method (e.g., spin-casting from solutions of different dye particular shows a slight upward trend with each smearing cycle 495 462 concentrations), and by repeated smearing, which removes (G1: cycle 1 = 55 s; cycle 10 = 68 s; G2: cycle 1 = 36 s; cycle 10 496 463 some sample with each iteration. An initial investigation was = 84 s) (Figure 5b). For G3, however, the thinnest glass film, 497 ff ff fi 464 conducted to get a rough idea of how these di erent this e ect is much more dramatic. After the rst cycle, τSM is 498 ff 465 approaches a ect smeared-state decay and to screen for already roughly double that of G1 and G2 (e.g., G3 τSM = 98 s; 499 ff fi 466 di erences. Afterward, a more detailed study was conducted G1 τSM = 55 s) and increases signi cantly with each smearing 500 fi 467 for glass lms, and results are presented below. (i.e., thinning) cycle. After the fourth cycle (τSM = 266 s), the 501 468 Films for initial screening were prepared in the following signal intensity was too low for reliable camera detection, so 502 469 ways. Glass films were prepared by two methods. Thicker glass data were no longer recorded. These results point to a film 503 470 films were made by drop casting from a concentrated thickness effect upon smeared-state decay and a possible 504 471 gbmOMe/THF solution (0.090 M) and evaporating in air threshold, below which larger decay times are detected. While 505 fi 472 (G1). Thinner lms were produced via spin-casting from the τSM represents an average for an entire 2D region of interest, 506 473 same concentrated (G2) and more dilute (0.018 M) solutions greater resolution is achieved with colormaps, which display 507 474 (G3) (note: the fabrication procedure for G3 is the same as smeared-state decays on a pixel-level scale (τR). Colormaps 508 475 that used to make films for ML studies on the glass above). associated with both weigh paper (Figure S8) and glass films 509 476 Thicker weigh paper films were fabricated by spreading bulk (Figure S9) show that regions with longer lifetimes are more 510 477 gbmOMe powder (∼2 mg) across the substrate (WP1; prevalent in films that are thinner (i.e., initially, upon repeated 511 478 standard fabrication procedure for weigh paper ML films smearing or due to localized smearing effects at the center of 512 479 above), and thinner films were made by drop casting from the the sample). 513 480 same concentrated (WP2) and dilute (WP3) solutions that Differences between weigh paper and glass samples may arise 514 481 were used to produce glass films. Then, glass and WP gbmOMe from substrate or thickness effects, initially or upon repeated 515 fi 482 thin lms were repeatedly smeared in the same region, and τSM smearing. Though G1 and WP2 were drop cast from the same 516

F DOI: 10.1021/acsami.7b01985 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX ACS Applied Materials & Interfaces Research Article

517 initial solution, cycle 10 smeared-state lifetimes are slightly thickness. Smeared-state decays increase with decreasing film 564 518 longer on glass (τSM = 68 s) than weigh paper (τSM = 48 s). thickness, and quite dramatically so for the thinnest samples. 565 519 Furthermore, the colormap of G1 after the tenth cycle shows To investigate these effects more systematically using a single 566 fi fi 520 well-de ned regions with large τR that are not present in common lm fabrication method and substrate, gbmOMe thin 567 521 colormaps corresponding to WP2. These findings indicate that films of variable thickness were prepared via spin-casting from 568 522 the dye likely interacts with cellulose paper fibers differently stock solutions of different concentrations (C1 = 0.090 M; C2 569 523 than with SiO2 glass. Furthermore, qualitative observations = 0.072 M; C3 = 0.054 M; C4 = 0.036 M; C5 = 0.018 M) on 570 524 suggest that more sample is removed from hard glass surfaces glass. This range is bracketed by concentrations used to 571 525 than from paper with each smearing cycle; additionally, fabricate G1 and G3 films above. Multiple films were prepared 572 526 smearing on paper may cause the dye to penetrate into the from each solvent, and the relative film thickness was estimated 573 527 fibers and be somewhat more resistant to removal with by measuring the absorption spectrum of each film in the AS 574 ff ’ 528 mechanical perturbation. These substrate di erences may help state and using the peak absorption (λabs = 363 nm) and Beer s 575 529 to explain why films on paper retain their reversible quality for Law to calculate the corresponding path length of each film 576 530 longer despite similar preparation and treatment as films on (Figure S11). Films in the AS state were used to determine 577 531 glass. relative initial thickness due to their transparency; TA films are 578 532 To test whether weigh paper films also reach a point at which opaque and incompatible with this measurement method given 579 fi 533 τSM values begin to increase, WP lms with initial thicknesses scattering effects. Approximate thicknesses were calculated 580 534 smaller than those of WP1−WP3 were fabricated by spin- relative to the thinnest film such that the maximum thickness 581 535 casting from a 0.018 M solution. Though spin-casting is was over eight times greater in the initial AS state. On the basis 582 536 somewhat unconventional for paper substrates, it nonetheless of the absorption spectra, some deviations in thickness were 583 537 addressed challenges associated with drop casting dilute dye observed when spin-casting from stock solutions of the same 584 538 solutions, which still produced relatively thick films after concentration. This was especially evident for films fabricated 585 539 evaporation due to the amount of solvent needed to cover the from solution C1. Samples spin-cast from concentrated 586 540 entire sample. After spin-casting and annealing at 75 °C, solutions produce less reproducible film thicknesses, which 587 541 heterogeneous films were obtained that were thin at the centers account for multiple points per concentration in Figure 7b. 588 f7 542 and thicker at the edges, as confirmed by colormap histograms After annealing, each film was smeared in three separate 589 543 with bimodal distributions (Figure S10). Though a thickness locations with a cotton swab for finer control (Figure 7). Then, 590 fi 544 gradient was observed, these lms allow monitoring of the τSM values for each smeared region were averaged to quantify 591 545 distribution of smeared-state decays on the same surface. After the recovery dynamics for a given film (Figure 7b) (mean τ = 592 fi SM 546 the rst smearing/recovery cycle, the colormap showed regions τSM̅ ). This approach provides insight into the reproducibility of 593 547 with longer recovery lifetimes (τR) where the sample was smeared-state decay values for regions of comparable thickness. 594 548 thinnest compared to the thicker edges that recovered more The representative colormaps for films of different initial 595 f6 549 quickly (Figure 6). The colormaps for subsequent smearing/ thicknesses, T1−T5, produced via spin-casting from different 596 solvents, C1−C5, show a narrow distribution of smeared-state 597 decays within each smeared region (Figure 7). A relatively 598 − narrow range of τSM̅ values (26 41 s) was observed for all but 599 the thinnest samples (max τSM̅ = 120 s), which corresponds to 600 an 87% decrease in initial thickness compared to the thickest 601 sample. Furthermore, most samples showed very little variation 602 in τSM̅ regardless of the region being sampled; however, the 603 thinnest samples showed the most significant deviation (e.g., 604 T5, SM1). Despite the relatively large error bars associated with 605 fi data for the thinnest samples, τSM̅ was still signi cantly longer 606 fi compared to thicker lm decays. The trend toward longer τSM̅ 607 Figure 6. (a) Images of spin-cast gbmOMe films on WP under for thinner films is additional evidence of a thickness effect in 608 ambient light (left) and UV irradiation (right). (b) Spatially resolved gbmOMe glass films, and here, too, it appears that a critical 609 colormaps of repeatedly smeared spin-cast gbmOMe on WP (λex = thickness threshold must be reached before thickness effects are 610 365 nm). detected and smeared-state decays noticeably increase. 611 To gain further insight into a possible thickness threshold, 612 550 recovery cycles show the erosion of the thicker regions of the the regime where smeared-state decays change most dramat- 613 551 sample and a general trend toward longer smeared-state decays. ically, another set of films with varying initial thicknesses, T6− 614 552 Furthermore, the smeared-state decays (i.e., calculated from T10, was prepared using the same method as described above. 615 fi fi 553 recovery lifetimes τR for a region of interest that encompasses These lms were further thinned by repeated smearing for ve 616 554 both thicker and thinner regions) for each of the cycles (τSM = cycles (Figure 8). As expected, regions of longer smeared-state 617 f8 555 66 s for cycle 1; τSM = 79 s for cycle 2; τSM = 83 s for cycle 3) decay (τSM > 100 s) developed and grew as the sample was 618 556 were significantly longer than those observed for bulk and drop continuously smeared. For thinner samples, these areas became 619 − fi 557 cast WP1 WP3 lms (35 < τSM < 55 s). Not only does this pronounced after fewer cycles. This trend is illustrated by 620 558 clearly demonstrate thickness-dependent recovery on weigh comparing the two extremes. For the thickest sample, T6, red 621 559 paper substrates, it also shows that differences in ML recovery colored regions with elevated recovery lifetimes started to 622 560 behavior can be used to evaluate relative film thickness in a develop only after cycle 4. However, for T10, the thinnest 623 561 given heterogeneous sample. sample, longer recovery lifetimes were observed beginning with 624 562 According to the initial studies, it is clear that the recovery the first cycle. Additionally, as T10 was repeatedly smeared, the 625 563 dynamics of gbmOMe thin films are dependent on sample size of the longer recovering red region increased such that it 626

G DOI: 10.1021/acsami.7b01985 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX ACS Applied Materials & Interfaces Research Article

Figure 8. Spatially resolved smeared-state decay colormaps of repeatedly smeared spin-cast gbmOMe thin ﬁlms with varying thicknesses on glass substrates. Samples were smeared with a Kimwipe for 2 s to aim for consistent sample removal for 5 smearing/recovery cycles. T6−T10 represent ﬁlms spin-cast from progressively more dilute stock solutions (T6 = 0.090 M; T7 = 0.072 M; T8 = 0.054 M; T9 = 0.036 M; T10 = 0.014 M).

Figure 7. (a) Representative recovery lifetime (τR) colormaps of spin- cast gbmOMe thin films on glass with different initial thicknesses fabricated via spin-casting from different concentration stock solutions. Relative thicknesses were calculated from absorption spectra of each film. Three different regions of each film (SM1, SM2, and SM3) were smeared with a cotton swab to assess reproducibility of smeared-state decays without changing initial thickness due to sample removal. Thicknesses T1−T5 represent films spin-cast from progressively more dilute stock solutions (C1 = 0.090 M; C2 = 0.072 M; C3 = 0.054 M; C4 = 0.036 M; C5 = 0.018 M). (b) Mean smeared-state decays (τSM̅ ) as a function of relative thickness. Error bars represent one standard Figure 9. Smeared-state decay (τ ) versus relative initial thickness of deviation from the mean τ̅ for each sample and indicate smearing SM SM smeared gbmOMe glass films after one and five smearing/recovery and thickness variations within a given film. Multiple films were made cycles. from each stock solution, which produced films with slight variations in initial thicknesses. certain amount of sample must be removed from thicker films 648 before accessing an interfacial regime where substrate 649 627 encompassed nearly the entire film after the third recovery interactions can be observed. It is also possible that the 650 fi fi 651 628 cycle. After ve smearing cycles, larger τSM values were material morphology changes in thinner lms. Given that ML is 629 observed for all samples regardless of initial film thickness; a solid-state property, a certain particle size and number and 652 fi 653 630 however, the increases in τSM are less signi cant for the thickest arrangement of molecules is likely required in the extended 631 sample. One potential explanation for this deviation is that the array to observe rapid reversibility. The thinnest of samples 654 632 thickness of T6 has not yet reached a critical region even after explored in this study may reach the limits of this regime for 655 633 five smearing/recovery cycles. As observed for the first cycle, a gbmOMe. 656 634 general upward trend was observed with decreasing initial ■ CONCLUSION 657 635 thickness until a maximum smeared-state decay (τSM = 104 s) 636 was measured for T10. Coupled with the results for the Stimuli responsive gbmOMe has many interesting thermal and 658 fi 637 colormaps, these trends in τSM provide further evidence that the mechanochromic properties. As a thin lm, gbmOMe forms a 659 638 smeared-state decay is highly dependent on the film thickness thermally stable supercooled liquid phase when melted and 660 639 of gbmOMe. subsequently cooled at ambient temperature. Yet, when the 661 640 Experiments indicate that τSM is relatively insensitive to bulk dye is cooled at room temperature, a metastable 662 641 thickness effects until a critical threshold is reached. This is supercooled liquid state is formed, which eventually recrystal- 663 642 evidenced by the small deviations in τSM that were observed lizes after a few minutes. Mechanochromic properties include a 664 643 until a requisite amount of gbmOMe was removed, and films blue to green shift in emission upon smearing along with rapid 665 f9 644 with smaller initial thicknesses show longer τSM values (Figure ML recovery at room temperature. This presents a unique 666 f9 645 9). One potential explanation for this apparent thickness handle for evaluating the performance of self-recovering ML 667 646 threshold is that sample−substrate interactions become more materials by a new camera method. The intensity decay of each 668 fi fi 647 signi cant for thinner lms and slow ML recovery. Thus, a pixel (i.e., recovery lifetime, τR) was used to generate a spatially 669

H DOI: 10.1021/acsami.7b01985 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX ACS Applied Materials & Interfaces Research Article

670 resolved colormap of the recovery process. When averaged over ■ ACKNOWLEDGMENTS 729 671 a region of interest, smeared-state decay values τ were SM We thank the National Science Foundation (NSF Grant CHE- 730 672 calculated and provided a means of comparison between 1213915) and the University of Virginia Department of 731 673 different sample preparations, substrates, and film thicknesses. Chemistry for support for this research and Professor James 732 674 Findings with thicker films illustrate the potential of gbmOMe Demas for helpful discussions concerning the camera detection 733 675 as a dynamic “renewable ink” with the ability to recover even method. 734 676 after repeated use. Thin films of gbmOMe, however, showed 677 thickness-dependent smeared-state decays on both glass and ■ REFERENCES 735 678 WP substrates, but significant fluctuations in recovery behavior (1) Morris, W. A.; Kolpaczynska, M.; Fraser, C. L. Effects of α- 736 679 were observed only in the thinnest films investigated, indicating Substitution on Mechanochromic Luminescence and Aggregation- 737 680 that a critical thickness must be reached before gbmOMe films Induced Emission of Difluoroboron β- Diketonate Dyes. J. Phys. Chem. 738 681 lose full reversibility. Kim et al. showed that quantitative force − fi C 2016, 120, 22539 22548. 739 682 measurements yield reliable results for uniform lms but are 740 36 (2) Zhang, G.; Lu, J.; Fraser, C. L. Mechanochromic Luminescence 683 unreliable for molecular solids with heterogeneous coverage. Quenching: Force-Enhanced Singlet-to-Triplet Intersystem Crossing 741 684 These results show, however, that smeared-state decays for Iodide-Substituted Difluoroboron−Dibenzoylmethane−Dodecane 742 685 obtained from camera analysis are quite reproducible even in the Solid State. Inorg. Chem. 2010, 49, 10747−10749. 743 686 with a qualitative smearing method. Particularly for fast (3) Zhang, G.; Lu, J.; Sabat, M.; Fraser, C. L. Polymorphism and 744 687 recovering ML systems, films of similar thickness on identical Reversible Mechanochromic Luminescence for Solid-State Difluor- 745 − 688 substrates must be used to accurately compare and quantify oboron Avobenzone. J. Am. Chem. Soc. 2010, 132, 2160 2162. 746 (4) Zhang, G.; Palmer, G. M.; Dewhirst, M. W.; Fraser, C. L. A Dual- 747 689 recovery in ML materials. Additionally, threshold effects may Emissive-Materials Design Concept Enables Tumour Hypoxia 748 690 well vary for different reversible ML dyes and substrates and Imaging. Nat. Mater. 2009, 8, 747−751. 749 691 fi should be identi ed in each case. Future studies with (5) DeRosa, C. A.; Samonina-Kosicka, J.; Fan, Z.; Hendargo, H. C.; 750 692 functionalized surfaces may provide additional insight into the Weitzel, D. H.; Palmer, G. M.; Fraser, C. L. Oxygen Sensing 751 ff 693 origin of substrate e ects and new methods for modulating ML Difluoroboron Dinaphthoylmethane Polylactide. Macromolecules 2015, 752 694 recovery. For certain applications, it may be desirable to 48, 2967−2977. 753 695 generate films that can fully recover after many smearing/self- (6) DeRosa, C. A.; Seaman, S. A.; Mathew, A. S.; Gorick, C. A.; Fan, 754 696 erasing cycles. Methods that greatly hinder or prevent sample Z.; Demas, J. N.; Peirce, S. M.; Fraser, C. L. Oxygen Sensing 755 − 697 removal and a reduction in sample thickness would be Difluoroboron β Diketonate Polylactide Materials with Tunable 756 − 698 important. Other dye−substrate combinations and protective Dynamic Ranges for Wound Imaging. ACS Sens. 2016, 1, 1366 1373. 757 (7) Nguyen, N. D.; Zhang, G.; Lu, J.; Sherman, A. E.; Fraser, C. L. 758 699 coatings that do not significantly affect ML color change and Alkyl Chain Length Effects on Solid-State Difluoroboron β-Diketonate 759 700 recovery behaviors are avenues worth exploring. Mechanochromic Luminescence. J. Mater. Chem. 2011, 21, 8409− 760 8415. 761 701 ■ ASSOCIATED CONTENT (8) Zhang, X.; Liu, X.; Lu, R.; Zhang, H.; Gong, P. Fast Detection of 762 702 *S Supporting Information Organic Amine Vapors Based on Fluorescent Nanofibrils Fabricated 763 703 The Supporting Information is available free of charge on the from Triphenylamine Functionalized β-Diketone-Boron Difluoride. J. 764 − 704 ACS Publications website at DOI: 10.1021/acsami.7b01985. Mater. Chem. 2012, 22, 1167 1172. 765 fi (9) Zhang, G.; Kim, S. H.; Evans, R. E.; Kim, B. H.; Demas, J. N.; 766 705 Spectra and data tables for gbmOMe glass and WP lms Fraser, C. L. Luminescent Donor-Acceptor β-Diketones: Modulation 767 706 at room temperature and 77 K, X-ray diffraction patterns of Emission by Solvent Polarity and Group II Metal Binding. J. 768 707 for the different phases of gbmOMe, additional Fluoresc. 2009, 19, 881−889. 769 708 experimental details for the camera detection method, (10) Zeng, S.; Zhang, D.; Huang, W.; Wang, Z.; Freire, S. G.; Yu, X.; 770 709 colormaps for repeatedly smeared glass and WP thin Smith, A. T.; Huang, E. Y.; Nguon, H.; Sun, L. Bio-Inspired Sensitive 771 and Reversible Mechanochromisms via Strain-Dependent Cracks and 772 710 films, absorbance spectra of spin-cast glass films from Folds. Nat. Commun. 2016, 7, 11802. 773 711 different concentration gbmOMe/THF stock solutions, (11) Ciardelli, F.; Ruggeri, G.; Pucci, A. Dye-Containing Polymers: 774 712 and colormaps and histograms of smeared-state decays of fi Methods for Preparation of Mechanochromic Materials. Chem. Soc. 775 713 repeatedly smeared WP lms fabricated via spin-casting Rev. 2013, 42, 857−870. 776 714 (PDF) (12) Sagara, Y.; Yamane, S.; Mitani, M.; Weder, C.; Kato, T. 777 715 Mechanochromic luminescence recovery processes for Mechanoresponsive Luminescent Molecular Assemblies: An Emerging 778 − 716 gbmOMe films AP on weigh paper (AVI) Class of Materials. Adv. Mater. 2016, 28, 1073 1095. 779 717 Mechanochromic luminescence recovery processes for (13) Chi, Z.; Zhang, X.; Xu, B.; Zhou, X.; Ma, C.; Zhang, Y.; Liu, S.; 780 Xu, J. Recent Advances in Organic Mechanofluorochromic Materials. 781 718 gbmOMe films spin-cast on weigh paper (AVI) Chem. Soc. Rev. 2012, 41, 3878−3896. 782 719 Mechanochromic luminescence recovery processes for fi (14) Zhang, X.; Chi, Z.; Zhang, Y.; Liu, S.; Xu, J. Recent Advances in 783 720 gbmOMe lms spin-cast on glass (AVI) Mechanochromic Luminescent Metal Complexes. J. Mater. Chem. C 784 721 (CIF) 2013, 1, 3376−3390. 785 (15) Fang, Y.; Ni, Y.; Leo, S.-Y.; Wang, B.; Basile, V.; Taylor, C.; 786 Jiang, P. Direct Writing of Three-Dimensional Macroporous Photonic 787 722 ■ AUTHOR INFORMATION Crystals on Pressure-Responsive Shape Memory Polymers. ACS Appl. 788 723 Corresponding Author Mater. Interfaces 2015, 7, 23650−23659. 789 724 *E-mail: [email protected]. (16) Jiang, Y. An Outlook Review: Mechanochromic Materials and 790 725 ORCID Their Potential for Biological and Healthcare Applications. Mater. Sci. 791 Eng., C 2014, 45, 682−689. 792 726 Cassandra L. Fraser: 0000-0002-8927-4694 (17) Butler, T.; Morris, W. A.; Samonina-Kosicka, J.; Fraser, C. L. 793 727 Notes Mechanochromic Luminescence and Aggregation Induced Emission 794 728 The authors declare no competing financial interest. for a Metal-Free β-Diketone. Chem. Commun. 2015, 51, 3359−3362. 795

I DOI: 10.1021/acsami.7b01985 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX ACS Applied Materials & Interfaces Research Article

796 (18) Butler, T.; Morris, W. A.; Samonina-Kosicka, J.; Fraser, C. L. (35) Mizuguchi, K.; Kageyama, H.; Nakano, H. Mechanochromic 863 797 Mechanochromic Luminescence and Aggregation Induced Emission of Luminescence of 4-[bis(4-Methylphenyl)amino]benzaldehyde. Mater. 864 798 Dinaphthoylmethane β-Diketones and Their Boronated Counterparts. Lett. 2011, 65, 2658−2661. 865 799 ACS Appl. Mater. Interfaces 2016, 8, 1242−1251. (36) Chung, K.; Kwon, M. S.; Leung, B. M.; Wong-Foy, A. G.; Kim, 866 800 (19) Hubaud, J. C.; Bombarda, I.; Decome, L.; Wallet, J. C.; Gaydou, M. S.; Kim, J.; Takayama, S.; Gierschner, J.; Matzger, A. J.; Kim, J. 867 801 E. M. Synthesis and Spectroscopic Examination of Various Substituted Shear-Triggered Crystallization and Light Emission of a Thermally 868 − 802 1,3-Dibenzoylmethane, Active Agents for UVA/UVB Photoprotection. Stable Organic Supercooled Liquid. ACS Cent. Sci. 2015, 1, 94 102. 869 803 J. Photochem. Photobiol., B 2008, 92, 103−109. (37) Shirota, Y.; Kageyama, H. Charge Carrier Transporting 870 871 804 (20) Morris, W. A.; Sabat, M.; Butler, T.; DeRosa, C. A.; Fraser, C. L. Molecular Materials and Their Applications in Devices. Chem. Rev. 2007, 107, 953−1010. 872 805 Modulating Mechanochromic Luminescence Quenching of Alkylated (38) Shirota, Y. Photo- and Electroactive Amorphous Molecular 873 806 Iodo Difluoroboron Dibenzoylmethane Materials. J. Phys. Chem. C Materials-Molecular Design, Syntheses, Reactions, Properties, and 874 807 2016, 120, 14289−14300. Applications. J. Mater. Chem. 2005, 15, 75−93. 875 808 (21) Morris, W. A.; Liu, T.; Fraser, C. L. Mechanochromic (39) Zografi, G.; Newman, A. Interrelationships Between Structure 876 809 Luminescence of Halide-Substituted Difluoroboron β-Diketonate and the Properties of Amorphous Solids of Pharmaceutical Interest. J. 877 810 Dyes. J. Mater. Chem. C 2015, 3, 352−363. Pharm. Sci. 2017, 106,5−27. 878 811 (22) Liu, T.; Chien, A. D.; Lu, J.; Zhang, G.; Fraser, C. L. Arene (40) Mathew, A. S.; De Rosa, C. A.; Demas, J. N.; Fraser, C. L.; 879 812 Effects on Difluoroboron β-Diketonate Mechanochromic Lumines- β − DeRosa, C. A.; Demas, J. N.; Fraser, C. L. Difluoroboron -Diketonate 880 813 cence. J. Mater. Chem. 2011, 21, 8401 8408. Materials with Long-Lived Phosphorescence Enable Lifetime Based 881 814 (23) Shen, X. Y.; Yuan, W. Z.; Liu, Y.; Zhao, Q.; Lu, P.; Ma, Y.; Oxygen Imaging with a Portable Cost Effective Camera. Anal. Methods 882 815 Williams, I. D.; Qin, A.; Sun, J. Z.; Tang, B. Z. Fumaronitrile-Based 2016, 8, 3109−3114. 883 816 Fluorogen: Red to Near-Infrared Fluorescence, Aggregation-Induced (41) Williams, D. B. G.; Lawton, M. Drying of Organic Solvents: 884 817 Emission, Solvatochromism, and Twisted Intramolecular Charge Quantitative Evaluation of the Efficiency of Several Desiccants. J. Org. 885 818 Transfer. J. Phys. Chem. C 2012, 116, 10541−10547. Chem. 2010, 75, 8351−8354. 886 819 (24) Han, T.; Gu, X.; Lam, J. W. Y.; Leung, A. C. S.; Kwok, R. T. K.; (42) Carraway, E. R.; Demas, J. N.; DeGraff, B. A. Luminescence 887 820 Han, T.; Tong, B.; Shi, J.; Dong, Y.; Tang, B. Z. Diaminomaleonitrile- Quenching Mechanism for Microheterogeneous Systems. Anal. Chem. 888 821 Based Schiff Bases: Aggregation-Enhanced Emission, Red Fluores- 1991, 63, 332−336. 889 822 cence, Mechanochromism and Bioimaging Applications. J. Mater. (43) Zhu, H.; Wang, X.; Li, Y.; Wang, Z.; Yang, F.; Yang, X. 890 823 Chem. C 2016, 4, 10430−10434. Microwave Synthesis of Fluorescent Carbon Nanoparticles with 891 824 (25) Poon, C.-T.; Wu, D.; Lam, W. H.; Yam, V. W. A Solution- Electrochemiluminescence Properties. Chem. Commun. 2009, 5118− 892 825 Processable Donor-Acceptor Compound Containing Boron(III) 5120. 893 826 Centers for Smal0l-Molecule-Based High-Performance Ternary (44) Chow, Y. L.; Johansson, C. I.; Zhang, Y.-H.; Gautron, R.; Yang, 894 827 Electronic Memory Devices. Angew. Chem. 2015, 127, 10715−10719. L.; Rassat, A.; Yang, S.-Z. Spectroscopic and Electrochemical 895 828 (26) Gong, Y.; Tan, Y.; Liu, J.; Lu, P.; Feng, C.; Yuan, W. Z.; Lu, Y.; Properties of 1,3-Diketonatoboron Derivatives. J. Phys. Org. Chem. 896 − 897 829 Sun, J. Z.; He, G.; Zhang, Y. Twisted D-π-A Solid Emitters: Efficient 1996, 9,7 16. (45) Heller, C. A.; Henry, R. A.; McLaughlin, B. A.; Bliss, D. E. 898 830 Emission and High Contrast Mechanochromism. Chem. Commun. Fluorescence Spectra and Quantum Yields. Quinine, Uranine, 9,10- 899 831 2013, 49, 4009−4011. Diphenylanthracene, and 9,10-Bis(phenylethynyl)anthracenes. J. 900 832 (27) Mutai, T.; Satou, H.; Araki, K. Reproducible On−off Switching Chem. Eng. Data 1974, 19, 214−219. 901 833 of Solid-State Luminescence by Controlling Molecular Packing (46) Ito, F.; Sagawa, T. Quantitative Evaluation of Thermodynamic 902 834 through Heat-Mode Interconversion. Nat. Mater. 2005, 4, 685−687. Parameters for Thermal Back-Reaction after Mechanically Induced 903 835 (28) Liu, W.; Wang, Y.; Sun, M.; Zhang, D.; Zheng, M.; Yang, W. Fluorescence Change. RSC Adv. 2013, 3, 19785−19788. 904 836 Alkoxy-Position Effects on Piezofluorochromism and Aggregation- (47) Sun, X.; Zhang, X.; Li, X.; Liu, S.; Zhang, G. A Mechanistic 905 837 Induced Emission of 9,10-Bis(alkoxystyryl)anthracenes. Chem. Com- − Investigation of Mechanochromic Luminescent Organoboron Materi- 906 838 mun. 2013, 49, 6042 6044. als. J. Mater. Chem. 2012, 22, 17332−17339. 907 839 (29) Zhang, X.; Ma, Z.; Yang, Y.; Zhang, X.; Jia, X.; Wei, Y. Fine- (48) Morris, W. A.; Sabat, M.; Butler, T.; DeRosa, C. A.; Fraser, C. L. 908 840 Tuning the Mechanofluorochromic Properties of Benzothiadiazole- Modulating Mechanochromic Luminescence Quenching of Alkylated 909 − 841 Cored Cyano-Substituted Diphenylethene Derivatives through D A Iodo Difluoroboron Dibenzoylmethane Materials. J. Phys. Chem. C 910 − 842 Effect. J. Mater. Chem. C 2014, 2, 8932 8938. 2016, 120, 14289−14300. 911 843 (30) Zhang, X.; Chi, Z.; Xu, B.; Jiang, L.; Zhou, X.; Zhang, Y.; Liu, S.; (49) Fedorenko, E. V.; Mirochnik, A. G.; Beloliptsev, A. Y.; Isakov, V. 912 844 Xu, J. Multifunctional Organic Fluorescent Materials Derived from V. (S2→S0) and (S1→S0) Luminescence of Dimethylaminostyryl-β- 913 845 9,10-Distyrylanthracene with Alkoxyl Endgroups of Various Lengths. Diketonates of Boron Difluoride. Dyes Pigm. 2014, 109, 181−188. 914 846 Chem. Commun. 2012, 48, 10895−10897. ̌ 847 (31) Galer, P.; Korosec,̌ R. C.; Vidmar, M.; Sket, B. Crystal Structures 848 and Emission Properties of the BF2 Complex 1-Phenyl-3-(3,5- 849 Dimethoxyphenyl)-Propane-1,3-Dione: Multiple Chromisms, Aggre- 850 gation- or Crystallization-Induced Emission, and the Self-Assembly 851 Effect. J. Am. Chem. Soc. 2014, 136, 7383−7394. 852 (32) Samonina-Kosicka, J.; Derosa, C. A.; Morris, W. A.; Fan, Z.; 853 Fraser, C. L. Dual-Emissive Difluoroboron Naphthyl-Phenyl β- 854 Diketonate Polylactide Materials: Effects of Heavy Atom Placement 855 and Polymer Molecular Weight. Macromolecules 2014, 47, 3736−3746. 856 (33) Hariharan, P. S.; Venkataramanan, N. S.; Moon, D.; Anthony, S. 857 P. Self-Reversible Mechanochromism and Thermochromism of a 858 Triphenylamine-Based Molecule: Tunable Fluorescence and Nano- 859 fabrication Studies. J. Phys. Chem. C 2015, 119, 9460−9469. 860 (34) Morris, W. A.; Butler, T.; Kolpaczynska, M.; Fraser, C. L. 861 Stimuli Responsive Furan and Thiophene Substituted Difluoroboron 862 β-Diketonate Materials. Mater. Chem. Front. 2017, 1, 158−166.

J DOI: 10.1021/acsami.7b01985 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX