Mechanochromic luminescence is a phenomenon in which the luminescence color changes when mechanical stimulus is applied and is recovered by other external stimuli, such as heat (Sagara et al., 2016; Ito, 2021). Such phenomena have received significant interest not only because they visualize otherwise invisible force histories, but also because of their mechanism involved in transducing macroscopic force to the molecular level. Basically, the color change is due either to a change in the molecular environment, the intermolecular arrangement, the molecular conformation, or a combination thereof.

Phosphorescent organic molecules are particularly promising mechanochromic materials because phosphorescence is highly sensitive to changes in the molecular environment (Xue et al., 2016; Huang et al., 2020). Phosphorescence is a spin-forbidden form of luminescence that involves a change in the spin multiplicity. Unlike metal complexes, such as Ir and Pt, which benefit from the heavy atom effect that accelerates spin-forbidden processes, metal-free organic molecules seldom show room-temperature phosphorescence (RTP) (Hirata, 2017; Kenry et al., 2019). However, the crystalline states of some metal-free organic molecules were reported to exhibit RTP as early as 1939 (Clapp, 1939; Bilen et al., 1978). More recently, significant interest has been directed toward “crystallization-induced phosphorescence” (CIP) following the seminal report published in 2010 by Tang et al. (Yuan et al., 2010) A rigid environment is generally regarded to be crucial for observing RTP from a metal-free organic molecule; otherwise, RTP is quenched by molecular motions (Baroncini et al., 2017). Indeed, the conventional RTP of an organic crystal is quenched by applying mechanical stimulation, which amorphizes the crystal. While such mechanoresponsive RTP turn-off behavior has been used to achieve RTP-to-fluorescence mechanochromism, (Gong et al., 2015; Mao et al., 2015; Xu et al., 2017; He et al., 2019; Huang et al., 2019; Ma et al., 2019; Wang et al., 2019a; Liu et al., 2021a; Liu et al., 2021b) the RTP mechanoresponse is largely limited to the turn-off type, with a molecular design that modulates mechanochromic RTP behavior remaining elusive.

To establish design principles for mechanochromism, two types of structure–mechanochromic-property relationship are important: the molecular-structure–property and the crystal-structure–property relationship. However, in many cases, the molecular structure and crystal structure are strongly interdependent, which prevents revealing the essential causal relationship. Polymorphic crystals, in which the same molecule forms different crystal structures, provide practical bases for studying the crystal-structure–property relationship (Wang and Li, 2017). On the other hand, the molecular-structure–property relationship can be elucidated when different molecules form isostructural crystals; i.e., crystals with the same space group, lattice constants, crystal packing, and molecular conformations (Fábián and Kálmán, 1999; Kálmán et al., 1993). The differences in isostructural crystal structures are so small that they usually do not disturb the properties derived from the molecular structure itself.

Previously, we reported the thienyl diketone BrTn (Figure 1A), the first metal-free organic molecule to exhibit RTP-to-RTP mechanochromism (Tani et al., 2019); in contrast to RTP quenching observed for conventional organic phosphors, mechanical stimulation amorphizes the BrTn crystals, resulting in RTP color change. Detailed investigations revealed that mechanical stimulus turns off the initial RTP (from a skew conformer) while providing a small amount of the highly emissive trans-planar (TP) conformer. In addition, we desymmetrized the C ₂-symmetrical BrTn structure. (Tani et al., 2020; Komura et al., 2021) Notably, replacing one of the two Br atoms with a hydrogen atom led to the formation of a crystal that is isostructural with that of BrTn (Tani et al., 2020). Interestingly, the crystal of the unsymmetrical diketone was non-emissive, which was attributed to the presence of voids that result from the volumetric difference between Br and H (Figure 1B top). Hence, the difference in the crystal structures were highly localized, facilitating the stringent crystal-structure–property relationship study.

Herein, we report the first RTP-to-RTP mechanochromism modulated by halogen-atom exchange. We envisaged that halogen-atom exchange in BrTn (e.g., bromine to chlorine) would afford isostructural crystals (Reddy et al., 2006; Saidykhan et al., 2021; He et al., 2017; Wen et al., 2019; Lai et al., 2021) that retained the mechanoresponsive nature of BrTn, as observed for our previous system (Figure 1B top). On the other hand, altering the heavy atoms is expected to affect the molecular phosphorescence properties, thereby modulating the responsiveness to RTP color. Heavy atom effect—the enhancement of the rate of a spin-forbidden process by the presence of an atom of high atomic number—has been widely used to modulate or achieve phosphorescence from metal-free organic molecules. (He et al., 2017; Wen et al., 2019; Lai et al., 2021; Wang et al., 2019b; Liao et al., 2021) With these aims in mind, we designed new diketone ClTn, in which the Br atoms in BrTn are replaced with Cl atoms (Figure 1B bottom). As a result, the fundamental stimulus-responsive nature of BrTn is retained, while the sensitivities of RTP color to mechanical/thermal stimuli are modulated. Our study demonstrates the usefulness of the halogen-exchange strategy for modulating the stimulus-responsive photofunctions of crystals that involve spin-forbidden processes.

Diketone ClTn was synthesized by reacting BrTn with CuCl. The X-ray structure of a single-crystal of ClTn revealed that it is isostructural to that of BrTn (CCDC), with superimposable molecular geometries, the same space group, and very similar lattice constants (Figure 2 and Table 1). The conformation of aromatic 1,2-diketones is well represented by two torsion angles: the vicinal-dicarbonyl torsion angle θ and the thiophene–carbonyl torsion angle ϕ (Mukai et al., 1992; Singh et al., 2002). A comparison of these angles in ClTn and BrTn highlights their almost identical crystal conformations (θ: 114.0(1) vs. 109.5(2)°; ϕ: –19.1(2) vs. –17.2(3)°). In addition, the diketones have three kinds of intermolecular interactions (a total of 12 interactions from one molecule), which are also comparable in both systems due to their identical crystal packing (Figure 2C). Hence, these isostructural crystals provide both conformationally and environmentally consistent systems that are ideal for investigating structure–property relationships.

ClTn was found to exhibit RTP-to-RTP mechanochromism in a qualitatively similar manner to BrTn (Tani et al., 2019); the RTP color changed from green (G-phase) to yellow (Y-phase) upon grinding. To investigate these photophysical properties in detail, we prepared samples of ClTn-G and ClTn-Y; ClTn-G is a crystalline powder reprecipitated from CHCl₃/MeOH, while ClTn-Y was prepared by uniformly grinding ClTn-G using an agate mortar and a pestle (see Experimental Section for details). The photoluminescence (PL) emission maximum of ClTn-Y (λ _PL = 560 nm) was observed to be redshifted by ∼40 nm compared to that of ClTn-G (λ _PL = 522 nm) (Figure 3A). By comparing the behavior of BrTn with that of ClTn (Figure 3B), (Tani et al., 2019) we conclude that the PL spectrum of ClTn-Y consists of an emerging emission from the TP conformer and a remaining small emission from the skew conformer (vide infra). The PL lifetimes of ClTn-G and ClTn-Y were determined to be 66 and 122 μs, respectively, without any nanosecond-order decay component, confirming that these are phosphorescence emissions (Figure 4). Moreover, the yellow RTP of ClTn-Y returned to green upon heating (Figure 5); this color-change cycle was repeated for five-times, thereby demonstrating reversible mechano/thermoresponsiveness. The observed RTP mechanochromism is based on crystal amorphization, because the PXRD pattern of ClTn-G reveals sharp diffraction peaks, while that of ClTn-Y shows peak broadening (Figure 6). Overall, the RTP-to-RTP mechanochromism of BrTn was well reproduced by ClTn; however, close inspection revealed notable differences in stimulus-responsiveness that are compared and mechanistically rationalized below.

Despite its isostructural nature, the RTP color of ClTn responds to mechanical stimulus more slowly than that of BrTn. For comparison, crystalline powder (ClTn-G or BrTn-G) was placed between two quartz plates, ground by rotating the upper plate while being pressed, with PL spectra acquired at rotating angles of 0, 90, 180, and 360° (Figure 7). The RTP color of ClTn changed gradually; the relative intensity of the emission from the TP conformer (λ _PL = 560 nm) continuously increased with the application of the mechanical stimulus. On the other hand, the TP emission of BrTn emerged rapidly; its RTP color changed dramatically during the first quarter turn and remained almost unchanged thereafter (Figure 7C). These results may seem to suggest that BrTn-G is more easily amorphized than ClTn-G; however, we note that what responds faster in BrTn is the RTP color, for which the ease of amorphization is just one of the possible factors. Evaluating/monitoring the rate of amorphization (loss of crystallinity) during mechanical stimulation is difficult because the preferred orientation may also affect the PXRD peak, the full-width at half-maximum (FWHM) of which is a measure of crystallinity, and can be alleviated by mechanical stimulation.

Monitoring the heat-induced recovery from the Y phase to the G phase is expected to provide a better understanding of the crystallinity–RTP color relationship, as heating does not disturb the orientation of the powder. With this in mind, PL spectra and PXRD profiles were acquired after heating the Y phase samples (h0) at the temperatures and times indicated in Figures 8A–C. In contrast to the mechanical response observed for ClTn, its PL spectrum responded faster to temperature than the spectrum of BrTn. More interestingly, ClTn and BrTn exhibited the same crystallinity response; the FWHMs of the PXRD peaks at 2θ = 6° were observed to decrease at the same rate (Figures 8D,E). These results indicate that, while the crystallinity of the powder is important for determining the RTP color, the stimulus-responsiveness of RTP color is not associated with ease of amorphization.

We hypothesize that the difference in the stimulus-responsiveness of the RTP color originates from a conformation-dependent heavy atom effect. To test this hypothesis, we determined the phosphorescence quantum yields Φ _p of each compound/phase and constructed the Φ _p-weighted PL spectrum of ClTn-Y. As expected from the weaker heavy atom effect of Cl compared to Br, the Φ _p of ClTn is smaller than that of BrTn (Table 2). Interestingly, the amorphous Y phase exhibited a larger quantum yield ratio (α = [Φ _p of BrTn]/[Φ _p of ClTn]) than the crystalline G phase, implying that the extent of the heavy atom effect depends on conformation. Next, we subtracted the PL spectrum of ClTn-G, which is purely derived from the skew conformer, from that of ClTn-Y (Figure 9A). The difference spectrum corresponds to the pure PL spectrum of the TP conformer. Indeed, other PL spectra of ClTn with partial amorphization (e.g., 180° in Figure 7A) can be reconstructed from the obtained spectra of the skew and TP conformers (Figure 9B), which indicates that the PL spectra are composed of the two emissions. Finally, we corrected the PL spectrum of ClTn-Y by multiplying the spectra of each conformer by α and then recombining them. The Φ _p-weighted PL spectrum of ClTn-Y (Figure 9C, blue trace) obtained in this manner matches the PL spectrum of BrTn-Y; therefore, we concluded that the conformational composition of amorphous ClTn-Y is similar to that of BrTn-Y, and that the differences in spectral shape are attributable to the difference in PL efficiency. A proposed mechanism for the different stimulus-responsivenesses of RTP color is described in Figure 10. Even though the ease of amorphization is similar as expected for the isostructural crystals, different PL efficiency (represented by color depth) makes the RTP color response different. Upon amorphization, the PL quantum yields are increased in BrTn-G/Y (3.9/10%) while slightly decreased in ClTn-G/Y (1.7/1.4%) (Table 2). Thus, the mechanoresponse in the RTP color is different. We emphasize that, in the present system, isostructural crystals exhibited a similar amorphization behavior, enabling the stimulus-responsiveness of the bulk solid to be modulated by tuning the molecular properties.

The RTP-to-RTP mechanochromism of a thienyl diketone was successfully modulated by Br-to-Cl halogen exchange. Modulating the molecular structure does not disturb the crystal structure, with both diketones forming isostructural crystals. The stimuli-responsiveness of the crystallinity is also retained, as evidenced by PXRD peak widths. In contrast, relative RTP efficiency is affected, which is ascribable to a conformation-dependent heavy-atom effect. Consequently, we were able to successfully modulate the stimulus-responsiveness of RTP color. We are currently investigating the conformation-dependent heavy atom effect, which will be reported in due course.