Efficient optical photoswitching of benzyloxy-substituted TCF-based D–π–A molecules

Abstract

Molecules exhibiting photo-switching behavior are essential for the development of various photonic and optoelectronic devices. In this study, we report the synthesis of five novel donor–π–acceptor organic dyes, in which a stilbene moiety serves as a π-linker between the electron-donating and electron-withdrawing units. To address the challenge of developing photo-switchable molecules responsive to visible light, we performed fundamental optical characterization in both solution and solid state. The photoisomerization capability was confirmed for all compounds through real-time absorption measurements, enabling the calculation of E → Z isomerization kinetics. The presence of both isomeric forms was further validated by ¹H NMR spectroscopy. These experimental findings were supported by quantum chemical calculations, which identified the most stable conformers and accurately predicted their spectral properties. Moreover, pump–probe experiments demonstrated that irradiation with linearly polarized light efficiently triggers photo-induced birefringence in dye-doped polymer systems, with a notable “memory effect” observed for all studied compounds. Remarkably, the dyes are highly sensitive to light, and the birefringence saturation can be achieved at light intensities comparable to natural daylight, highlighting their broad potential application in optoelectronic devices.

---

Introduction

Recently, one of the main research focuses has been the design of photo-switchable molecules, due to their potential applications in various fields such as sensing, optoelectronics, and photon energy storage^1–3. Molecular photoswitching can occur in a broad range of material systems, including liquid crystals,^4,5 metal–organic frameworks,^6,7 and organic dyes.^8,9 Organic compounds, in particular, offer significant advantages in comparison to other systems due to the relative ease of tuning their spectroscopic properties.¹⁰ Additionally, they can be integrated into diverse media, such as liquid solutions or solid films,^11,12 while maintaining their photoresponsive behavior.

Among the various types of organic dyes, those featuring a donor–π–acceptor (D–π–A) architecture can exhibit multiple functionalities, depending on the specific molecular structures and substituents used during synthesis.^10,13 Their modular nature allows the design process to be compared to a “LEGO block” concept, where different properties can be tailored by simply varying the donor (D), π-bridge, or acceptor (A) components within a common synthetic framework. Moreover, the synthesis of these compounds is generally considered cost-effective, which represents a significant advantage for potential large-scale production and commercialization.

D–π–A molecules featuring vinyl π-linkers have attracted significant attention due to their ability to facilitate precise molecular-level control during photoisomerization.^14,15 Materials with reversible optical photoswitching,¹⁶ responsiveness to visible light,¹⁷ tunable optical properties, and controllable photoisomerization processes¹⁸ are currently of great interest to researchers. These molecules are promising candidates for intelligent optical materials,^19,20 sensors,^21,22 and optical data storage,^23,24 as well as applications in optoelectronics²⁵ and super-resolution microscopy.²⁶ However, the alkene bond has the drawback of requiring UV light to induce the photoisomerization transition between its two forms, which limits its applicability.²⁵ To overcome this limitation, substituting D–π–A compounds with a strong electron-acceptor group, such as 2-dicyanomethylidene-3-cyano-4,5,5-trimethyl-2,5-dihydrofuran (tricyanofuran, TCF), can significantly red-shift the absorption spectrum, thereby enhancing the usability of these molecules within the visible range of the electromagnetic spectrum^10,27.

In this work, we present a family of five different D–π–A structured compounds. Each molecule features a TCF acceptor group and a poly-benzyloxybenzyl donor group, connected via a vinylene linker to facilitate intramolecular charge transfer. It is important to note that the presented compounds share a structural similarity with dyes that have already been published²⁷. Motivated by the interesting properties of those previously reported materials – such as solid-state emission and red-shifted luminescence toward the infrared region – our goal was to design and synthesize new compounds with improved solubility, expecting them to exhibit similar optical properties. To achieve this, we retained the same acceptor group but modified the donor part of the molecules. For the new family of dyes, we performed optical investigations like absorption and emission spectral measurements, both in solution and solid state, as well as photoluminescence lifetime and quantum yield estimations. We performed quantum chemical calculations to gain deeper insight into the photophysical properties of the investigated molecules, including the identification of their stable conformers and the nature of the ground-to-excited state transitions. Afterwards, we investigated E ↔ Z photoisomerization through absorption spectroscopy and ¹H NMR measurements in DCM and CDCl₃, respectively. Finally, in a pump–probe experiment, we demonstrated that the synthesized molecules could function as efficient molecular switches for inducing optical birefringence (Δn) in dye-doped polymeric layers, even at very low pump intensities.

Results and discussion

Material synthesis

The synthesis of all five TCF-based derivatives was carried out using the Knoevenagel reaction, as illustrated in Fig. 1. This reaction was performed under microwave irradiation in anhydrous ethanol with a few drops of piperidine serving as a catalyst. The synthesis and characterization of the dyes, including ¹H NMR, ¹³C{¹H} NMR, and high-resolution mass spectrometry (HRMS) analyses, which were performed on the powdered samples, are summarized in Fig. S1–S15.

Fig. 1 Synthesis scheme of molecules 1a–e.

Optical properties

The optical properties were measured in solution as well as in the solid state using powders directly obtained from the synthesis. Detailed experimental procedures can be found in the SI.

Optical properties in solutions

The absorption and emission spectra of compounds 1a–e were measured in dilute tetrahydrofuran (THF) solutions. THF was chosen due to its intermediate polarity. The concentration was approximately 10⁻⁵ M for both absorption and photoluminescence (PL) measurements (see Fig. 2(a) and (b)). All compounds in this series exhibit broad absorption bands ranging from 320 nm to 550 nm, which is characteristic of D–π–A systems.²⁷ The optical properties of the dyes in THF solution are summarized in Table 1. As predicted, compound 1a displays the most blue-shifted absorption maximum. This observation aligns with the general principle of this molecular systems: increasing the electron-donating strength of the donor segment typically leads to a red shift in the absorption spectrum.¹⁰

Fig. 2 (a) Absorption and (b) emission spectra of 1a–e in ∼10⁻⁵ THF solutions. (c) CIE 1931 graph calculated for the emission spectra (d) with the x and y coordinates placed under the squares filled with the corresponding colors. Emission was recorded at an excitation wavelength equal to the maximum of the absorption spectra of each compound.

Table 1 Optical properties of 1a–e in solution and solid state

Dye	Solution				Solid state
	λabs (nm)/ε^a (M^−1 cm^−1)	λemi^a (nm)	Δλ (nm)/(cm^−1)	Φ^a^b (%)	λemi (nm)	Φ^c (%)	τ^d (ns)
a THF solution.b Using 10^−1 M NaOH solution of fluorescein as a reference (QY = 79%).^29c Absolute quantum yield measured using a calibrated integrative sphere in direct excitation.d Values presented in parentheses are relative amplitudes of PL decay fitted with the two-exponential model.
1a	434/27 942	548	114/4794	4.8	599	8.3	τ1 = 0.40 (50%)
							τ2 = 1.20 (50%)
1b	453/22 250	560	107/4218	<1	621	22	τ1 = 2.10 (25%)
							τ2 = 4.90 (75%)
1c	443/21841	556	113/4587	<1	587	8.8	τ1 = 0.42 (86%)
							τ2 = 0.89 (14%)
1d	477/25 410	554	77/2913	<1	637	14	τ1 = 0.87 (46%)
							τ2 = 1.80 (54%)
1e	439/21 417	578	139/5478	<1	641	n.d.	n.d.

---

Di-benzyloxy compound 1b shows a more red-shifted absorption band with respect to that of 1a. Tri-benzyloxy 1d shows an absorption maximum at 477 nm, while 1c and 1e absorption is blue-shifted. This may be caused due to steric hindrance, which reduces charge transfer between the donor and acceptor groups.²⁸ The molar absorption coefficient reaches approximately 20 000 M⁻¹ cm⁻¹ for 1b–e, while for 1a, it is around 28 000 M⁻¹ cm⁻¹.

To obtain photoluminescence spectra, we excited solutions with wavelengths matching absorption maxima (see Fig. 2). Compounds 1a–e were characterized by a large Stokes shift (from 2913 cm⁻¹ obtained for 1d to 5478 cm⁻¹ for 1e). However, such a significant value of the aforementioned shift is typical for molecules with this structural architecture. It is noteworthy that the spontaneous emission of the investigated compound family is placed mostly in the yellow region of light and in the yellow-orange light range for 1e dye (see the CIE coordinates and diagram in Fig. 2(c) and (d)). The quantum yields of fluorescence (QYs) were calculated using a 10⁻¹ M NaOH solution of fluorescein (QY = 79%) from ref. 29. The value of QYs for 1a–e was relatively low, reaching 4.8% only for 1a. Details about measurements are presented in the Materials and methods section in the SI.

Optical properties in the solid state

In all cases, the fluorescence obtained for solids shifts towards the red region of light (Fig. 3(a)), which might be important from an application point of view.^30,31

Fig. 3 (a) Emission spectra of compounds 1a–e in the solid state; (b) CIE 1931 graph calculated for the emission spectra (c) with the x and y coordinates placed under the squares filled with the corresponding colors. Emission was recorded at an excitation wavelength equal to the maximum of the absorption spectra of each compound.

The fluorescence measurements were performed directly on powders (see details in the SI). The emission maxima of 1a–e are located between 587 nm (1c) and 641 nm (1e), which corresponds to the orange-red region (see the CIE 1931 coordinates in Fig. 3(b) and (c)). The most significant shift was observed for 1e, with a value of 202 nm (7180 cm⁻¹). 1b, which differs in one –OBn group from 1e at the meta position, shows a relatively high value of the solid–liquid shift as well, namely, 168 nm (5974 cm⁻¹). On the other hand, 1c shows the lower value of Stokes shift – 144 nm (5546 cm⁻¹). The photoluminescence quantum yields for solids were determined using a direct method utilizing an integrated sphere (see Methods details in the SI). The values of the QYs in solids compared to those estimated in solutions were significantly higher. Dyes 1a and 1d achieve a similar quantum yield of ∼8%. The most noteworthy results are QYs up to 22% for 1b and 14% for 1d. Additionally, we measured the lifetimes (τ) of the PL for the powders (see Fig. S16 and S17). Photoluminescence lifetimes for 1a–d were estimated using a double exponential fit, where amplitudes are expressed as the percentage of each time constant contribution. The longest exponential decay was measured for 1b, for which values of decay times of 4.9 ns (75%) and 2.1 ns (25%) were determined. The rest of the PL lifetimes for the dyes oscillated around 1–2 ns. Additionally, we calculated the radiative and non-radiative rate constants for the studied compounds in the solid state. Since all of the presented photoluminescence decay curves were best fitted with a bi-exponential model, we decided to utilise the weighted-average photoluminescence lifetime. The rate constants were calculated using a simple approach, as presented below:³²

where k_r is the radiative rate constant, k_nr is the non-radiative rate constant, Φss is the fluorescence quantum yield estimated for solid-state emission, and τ_av is the weighted-average photoluminescence lifetime. It is important to note that for all dyes, we clearly observe k_nr > k_r, indicating that non-radiative processes are dominant. However, for compound 1b, which exhibits the highest values of τ_av and Φ_ss, the ratio between k_r and k_nr is the smallest, suggesting a higher fluorescence efficiency compared to the rest of the dyes. For the values of both rate constants and the averaged photoluminescence lifetimes, please refer to Table 2.

Table 2 Average photoluminescence lifetime (τ_av) and radiative (k_r) and non-radiative (k_nr) rate constants for the studied compounds in the solid state

Dye	τav (ns)	kr × 10^8 (s^−1)	knr × 10^8 (s^−1)
1a	0.80	1.04	11.5
1b	4.20	0.52	1.86
1c	0.49	1.80	18.6
1d	1.37	1.02	6.23
1e	n.d.	n.d.	n.d.

---

Quantum chemical calculations

The representative sample of the energetically lowest conformers of compounds 1a–e has been selected based on metadynamic simulations at the GFN2-xTB level.³³ Among several exchange–correlation functionals tested (cf. Table S2), we chose PBE0, which shows good agreement with the experimental spectra. The equilibrium molecular structures were determined using the PBE0-D3BJ functional with dispersion correction and assuming an aug-ccpVDZ basis set and a PCM implicit solvent model, as implemented in the GAUSSIAN 16 package.³⁴ Vertical excitation spectra were then calculated using the TD-DFT method (cf. Table 3). We performed these calculations using a PCM model, focusing on the results obtained for THF, primarily because the basic spectroscopic properties were measured in this solvent.

Table 3 Relative S₀ energies (ΔE_SO), vertical excitation energies (E_exc) of the bright S₁ state, the corresponding wavelengths and oscillator strengths (f_osc) of the selected 1a–e rotamers (depending on the orientation of phenyl rings in the donor part of the dye) calculated using the PBE0-D3BJ/aug-cc-pVDZ/PCM(THF) method at the respective ground state local minima. The experimental absorption spectra are also shown for reference

Compound	ΔES0 (kJ mol^−1)	Eexc (eV)	λabs (nm)		Fosc
			Calc.	Exp.
1a (E′b)	0	2.783	445	434	1.63
1a (E′′b)	2.57	2.733	454		1.39
1a (E′a)	3.39	2.776	447		1.70
1a (E′′a)	5.57	2.729	454		1.48
1b (E′b)	0	2.613	474	453	1.52
1b (E′′b)	2.16	2.579	481		1.24
1c (E′b)	0	2.722	455	443	1.52
1c (E′′b)	2.33	2.669	464		1.24
1d (E′b)	0	2.704	458	477	1.63
1d (E′′b)	2.03	2.662	466		1.33
1e (E′b)	0	2.538	488	439	1.15
1e (E′′b)	0.28	2.510	494		1.07

---

Considering that the studied dyes have several plausible rotamers, we performed a detailed conformational search for the simplest molecule 1a and focused on the most stable rotamers of the remaining compounds. Among the optimized structures of 1a, we found E and Z isomers of the ethylenic bridge (–HC=CH–) and bent- (b) and axial-like (a) conformers of the methoxy bridge (–O–CH₂–) in the –OBn group (cf. SI). The latter differ in the mutual orientation of phenyl rings (the dihedral angle of the R–O–CH₂–R is either ∼180° in the ‘a’ form and ∼75° in the ‘b’ form). The E isomers are found in two nearly isoenergetic conformations, which differ in the orientation of the ethylenic bridge.

According to our calculations in the ground state, only the E′ and E′′ isomers with the axial or bent orientation of –OBn groups are likely to be present in the solution as the located Z isomers that are over 20 kJ mol⁻¹ are less stable (see the SI for details). Thus, in the case of compounds 1b–e, we focused on the E′b and E′′b isomers. The most relevant results are shown in Table 3, with more data reported in the SI.

Considering the size of the studied compounds, we chose 1a for further studies of the photoisomerization mechanism. Fig. 4 shows a schematic representation of the most likely E–Z photoisomerization pathway of the E′b isomer of 1a. To obtain more reliable relative energies, the XMS-CASPT2/cc-pVDZ energies were calculated using the BAGEL package³⁵ at the geometries obtained using the PBE0-D3BJ/def2-SVP/CPCM(THF) approach implemented in the ORCA 6.0.1 package.³⁶ The E′b isomer, after photoexcitation to the S₁ ¹ππ* state having a significant charge-transfer (CT) characteristic (from phenyl to TCF), undergoes structural relaxation (mainly bond-length alteration) to the S₁ (E) state minimum with a similar geometry. Further relaxation leads to an intermediate structure with the ethylenic –HC=CH– bridge rotated by 90°, which is close to the conical intersection with the ground state (the S₁–S₀ energy gap is only 0.2 eV at the PBE0-D3BJ/def2-SVP/PCM(THF) level) and then to the Z′b isomer. To further confirm this mechanism, we performed a simulation of the emission band of the latter using the Adiabatic Hessian approach implemented in the excited state dynamics³⁷ subroutine of an ORCA 6.0.1 package,³⁶ which agrees well with the experimental data (Fig. S18). The calculated fluorescence rate is 5.908 × 10⁸ s⁻¹ (τ = 1.69 ns). The character of the S₁ state is essentially a HOMO–LUMO transition (see Fig. 4), and it remains virtually unchanged throughout the E–Z photoisomerization process, although for the rotated S1 structure, it becomes a dark state. The low wavelength shoulder of the experimental absorption band may be due to the S₂ ¹ππ* locally excited (LE) state, localized mostly at the TCF fragment, which, according to PBE0-D3BJ/aug-cc-pVDZ/PCM(THF) calculations, lies at 3.55 eV (349 nm) for 1a. However, the computed oscillator strengths are usually substantially smaller than that for the bright S₁ state (0.02–0.06) – only for the 1b dye, the computed relative oscillator strengths for S₀ → S₁ and S₀ → S₂ transitions correspond to the experimental relative intensities, and for the E′b isomer, they amount to 1.22 and 0.30, respectively. The energy differences between rotamers are low enough (see Table 3) that molecules at room temperature can freely undergo rotameric transitions. Only the presence of isomeric structural changes (E to Z and Z to E) can be identified in the performed spectroscopic experiments. Thus, in the next sections of the article, we decided to streamline the nomenclature, and we simply categorized all conformers as E and Z, depending only on the rotation around the ethylene bridge (–HC=CH–), without distinguishing between rotameric states.

Fig. 4 A schematic representation of the most likely E–Z photoisomerization pathway of dye 1a. The relative XMS-CASPT2(6,6)/cc-pVDZ energies were calculated at the geometries obtained using a PBE0-D3BJ/def2-SVP/PCM(THF) method. The rotated ¹ππ* (R) intermediate is close to the minimum-energy crossing point (MECP) with the ground state.

Photochromism

The photochromic reaction based on photoisomerization between stable conformers of the studied dyes, predicted in quantum chemical calculations, was further investigated experimentally by UV-VIS absorption measurements and NMR spectroscopy. Finally, we also checked if cyclic photoisomerization can lead to molecular photo-orientation and, thus, induce optical birefringence.

Real-time absorption measurements

The photoisomerization process can be monitored by measuring the intensity of the absorption band of the E and Z isomers (see details in the SI). In our case, the reaction was carried out in dichloromethane (DCM) solution with a concentration of C = ∼10⁻⁵ M. DCM was chosen as the solvent due to its medium polarity and lack of light absorption in the UV region, where the Z-isomers of the dyes are expected to absorb. Although the quantum chemical calculations were performed using THF as the solvent model, the polarity of DCM is comparable to that of THF. Therefore, we considered it reasonable to compare the experimental and theoretical results, assuming only a small margin of error between the two solvents. The absorption of E isomers is located at around ∼430–470 nm, while Z isomers can absorb the light near ∼370 nm. By changing the wavelength of the light irradiation, we can control the relative concentration of both conformers, obtaining the typical photo-switching behavior. Thus, we used the 455 nm light irradiation to force transformation for all of the studied compounds. This transition can be observed in absorption spectra as the E band decreases with the increase of the Z band (see Fig. 5(a) and (b) for the representative compound 1a). When we obtained the photostationary state for the E → Z photoisomerization, we used 365 nm light to obtain a reverse transition. The well-defined isosbestic points clearly show that the isomerization reaction is carried out only between the dyes’ isomers.

Fig. 5 The electronic absorption spectra of compound 1a over the period of irradiation at (a) 455 nm and (b) 365 nm, (c) the dependence of as a function of the time and (d) the changes in the ¹H NMR spectra, with regions characteristic of the E and Z transitions highlighted by dotted red and blue rectangles, respectively.

The photoisomerization degree (Table 4) for the E → Z transition was obtained using eqn (3):

where A₀ is the absorbance just before the irradiation and A_∞ is the absorbance in the photostationary state at the maximum of the E band. The degree of photoisomerization varies between 31 and 50% for different compounds and is shown in Table 4. The kinetics data (Fig. 6(c)) present the determination of the k for 1a, and Fig. S19 shows kinetics data for 1b–1e. However, it is worth noting that, based on ¹H NMR measurements, the irradiation of the samples at 455 nm, which is used to determine the S(%), results in a simultaneous decrease in the signals corresponding to the E isomers and an increase in the signals associated with the Z isomers. Nevertheless, the whole process can be obtained by the calculation of the slope coefficient in the linear function plotted as a logarithmic dependence of the time:³⁸where A(t) is the value of the absorbance as a function of time and k is the kinetic constant, shown in Table 4. It is clear to see that the slowest transition of the E–Z photoreaction is observed for the 1a molecule, which can be determined by the most favorable molecular arrangement obtained in the E form of the isomer. Note that 1c and 1e have the fastest kinetic reaction with respect to the other dyes.

Table 4 The kinetic rate constant (k) of the E → Z isomerization under 455 nm irradiation was determined for the studied compounds in DCM solution (C ≈ 10⁻⁵ M) at room temperature. The values are expressed in s⁻¹ and the corresponding isomerization times are expressed in seconds. R² represents the goodness of fit of the linear regression, and S denotes the photoisomerization degree (and it was calculated for the maximum absorption wavelength for the E isomer)

Compound	E → Z
	k (s^−1)	k (s)	R^2	S (%)
1a	(2.29 ± 0.05) × 10^−2	44 ± 1	0.9947	42
1b	(3.04 ± 0.07) × 10^−2	33 ± 1	0.9954	50
1c	(3.43 ± 0.14) × 10^−2	29 ± 2	0.9861	31
1d	(3.13 ± 0.05) × 10^−2	32 ± 1	0.9963	n.d.
1e	(3.37 ± 0.11) × 10^−2	30 ± 1	0.9898	41

---

Fig. 6 (a) Impact of the external light on the changes in optical birefringence for 1a, (b) time constant for the growth curve measured for all of the studied compounds, (c) saturable refractive index fit, and (d) linear dependence of changes in transmittance (ΔT) as a function of pump intensity (I).

The structural changes of all compounds were also studied by ¹H NMR analysis (the details of the experiment can be found in the Materials and methods section of the SI). For all samples (see Fig. 5(d) and Fig. S20–S23), irradiation at 455 nm resulted in a decrease in the doublet characteristic of the CHs of the trans double bond (i.e., doublet at around 6.85 ppm, J ∼ 16.4 Hz, marked with the red rectangle in Fig. 5(d)) and the appearance of the new doublet characteristic of the CHs of the Z double bond (i.e., doublet at 5.8–5.9 ppm, J – 12.1 Hz). For the exact positions of the E and Z isomer doublets and their relative integration values over time, see Table S5. Please note that 1d was the only dye to exhibit the presence of Z isomers immediately after dissolution, as observed in the ¹H NMR spectra.

Photoinduced birefringence

The photoisomerization process can lead to the development of optical birefringence in the sample. Typically, the optical anisotropy for azo-dyes is obtained by molecular ordering through the Weigert effect.³⁹ Such optically induced birefringence can be used to construct optical valves or to store optical data. As our previous experiments have shown, photoisomerization can be obtained for all of the molecules described here. Thus, we decided to investigate if they can also generate optical anisotropy. The experiment was carried out in the so-called pump–probe optical setup for dye-doped polymeric samples inserted between two crossed polarizers, where the DPSS diode laser of wavelength 473 nm was used to induce optical birefringence while the He–Ne laser was probing the induced Δn over time.

The details of the experiment and materials preparation are provided in the SI. Afterwards, we can calculate the value of photoinduced birefringence measured in a crossed-polarized system, namely, Δn (I_pump, t), using the below equation:⁴⁰

where λ_prob is the wavelength of the probe laser, d is the thickness of the sample, and ΔT (I_pump, t) is the sample transmittance. In all of the cases, the measurements were conducted on the polymeric layers made of the dye-doped poly(methyl methacrylate) (PMMA) polymer with a concentration of 2%. Nevertheless, for the whole family, we noticed a typical exponential growth of intensity, which can correspond to the optical birefringence caused by the multiple changes in the molecular conformations.⁴¹ On the other hand, when the laser pump is off, we cannot obtain the return of the material to the initial value of Δn (see Fig. 6(a) and Fig. S24), confirming the long-time thermal relaxation for the studied materials.

Both growth and decay curves have shown complex dynamics and could only be fitted by the triple exponential functions with considerable accuracy. All time constant values were determined for the used power of the laser pump (∼60.5 ± 6.9 mW cm⁻²) (see eqn (6) for the growth and eqn (7) for the decay, and Fig. 6(a)). The values of the calculated time constants are shown in Table S6,

where Δn^gr_1–3 and Δn^dec_1–3 are amplitudes of growth and decay curves, respectively, while τ^gr_1–3 and τ^dec_1–3 are the corresponding time constants, t₀ is the time when laser light irradiation was turned on or off, and Δn^dec is a constant birefringence observed after decay. It is worth adding that in all cases, we observed the memory effect, and once the placed sample was irradiated by the pumping laser, the birefringence never returned to the 0 value. Exemplary Δn dynamics for the 1a molecule with fitting functions is shown in Fig. 6(a).

Growth time constants are shown in Fig. 6(b). It is clearly seen that in the case of the growth curve, the shortest time constant (τ^gr₁) presents the highest amplitude, based on calculated Δn, which was estimated using eqn (6) (see Table S6). On the other hand, the τ^gr₂ and τ^gr₃ present the lowest amplitude, with a value between 4 and 18 s for τ^gr₂ and with a range of 77 s to 194 s for τ^gr₃. When we consider the decay curve for all of the studied compounds, we see a relatively equal distribution for each of the time constants except for 1b and 1d, where the τ^gr₁ presents a visibly higher contribution than τ^gr₂ and τ^gr₃ (Table S6).

Due to the observed saturation effect in photoinduced birefringence (see Fig. 6(c)) at low power pump intensity, which is due to the saturation of E isomer absorption, we propose to fit obtained data with the following expression (see the derivation in the SI):

where Δn(I) is the maximal birefringence for intensity I, Δn is the maximal birefringence for the given system when I → ∞, and I₀ is the saturation intensity. Note that the I₀ values measured for all of the dyes are exceptionally small. For instance, compound 1c achieves its saturation level at the pump intensity as low as 8 mW cm⁻². On the other hand, 1b presents I₀ ∼ 22 mW cm⁻² (see Table 5). Moreover, the most striking fact is that available birefringence values for the measured samples are on the order of 10⁻⁴–10⁻³, which is a significant value with respect to other systems utilizing small, isomerizable molecules for the generation of optical birefringence.^42–44

Table 5 The values of the Δn, I₀, Δn₀, and ξ for 1a–e

Compound	Δn	SD^a (Δn)	I0 (mW cm^−2)	SD^a (I0)	R^2	Δn0^b	ξ^c (%)
a SD(−) is the standard deviation of value.b The Δn0 was estimated from the extrapolated triple exponential decay fit over 1000 seconds to obtain a plateau.c ξ represents the percentage share of Δn0 in Δn.
1a	2.7 × 10^−3	8.1 × 10^−5	20	2	0.9986	3.97 × 10^−4	14.65
1b	1.8 × 10^−3	1.0 × 10^−4	22	4	0.9937	2.58 × 10^−4	14.02
1c	9.0 × 10^−4	3.4 × 10^−5	8	3	0.9914	2.22 × 10^−4	24.69
1d	2.1 × 10^−3	8.5 × 10^−5	12	3	0.9902	1.72 × 10^−4	8.23
1e	2.0 × 10^−3	8.0 × 10^−5	11	3	0.9939	4.44 × 10^−4	21.98

---

Fast saturation and strong nonlinear character can be seen as disadvantages in various optical modulators. Luckily, for the mentioned application, typically, the value of interest is the transmittance or transmitted light intensity. In Fig. 6(d), we plot the transmittance ΔT obtained using crossed polarizers vs. pumping beam intensity.

Discussion

The analysis of the optical properties of the studied compounds, both in solution and in the solid state, shows that only compound 1a exhibits a photoluminescence quantum yield above 1% in solution. Please note that compound 1d exhibited the lowest Stokes shift among the studied dyes (in solution). This may be caused by the bonding effect of the –OBn groups on the carbon–carbon double bond, which could disturb the planarity of the entire system. As a result, this may interfere with the charge transfer between donor and acceptor groups, disrupt electron delocalization, and thus blue-shift the emission spectrum. (Please refer to Table S7 for the corresponding calculations.)

Additionally, the photosensitivity of the materials was evaluated through real-time monitoring of their absorption spectra under irradiation at two different wavelengths (455 nm and 365 nm). The results demonstrated full reversibility of isomerization processes in liquid environments, with the kinetics of the E → Z transformations indicating a rapid response that can be triggered even under daylight conditions. As previously mentioned, the rate of isomerization depends on the molecular structure, specifically the number and position of –OBn substituents. Compounds 1c and 1e exhibited the fastest E → Z transitions, whereas 1a showed the slowest response.

Structural changes between the isomeric forms were confirmed by ¹H NMR spectroscopy. Note that the ¹H NMR measurements were performed in deuterated chloroform – CDCl₃ (unlike in the case of calculation and real-time absorption measurements). Please note that Z–E photoisomerization can occur in different solvents for all dyes, indicating their versatility and making them ideal for photoswitching applications. Notably, thermal relaxation back to the initial E form was observed only for compounds 1d and 1e, as confirmed by the reappearance of characteristic signals in the ¹H NMR spectra (considering the given time of 72 h). Please note that the relative integration of the respective E/Z doublets presented in Table S5 clearly shows a decrease in the integration of the doublet corresponding to the Z-isomer over time. This observation indicates that thermal relaxation does occur for these dyes, although it proceeds at a significantly slow rate. In contrast, for compounds 1a, 1b, and 1c, the doublet corresponding to the Z isomer remained detectable even after 72 hours in the dark, indicating a significantly slower thermal back-isomerization process. This prolonged stability of the Z isomer suggests that such compounds, particularly 1a–1c, may hold potential for optical data storage applications, where long thermal relaxation times are considered highly advantageous.⁴⁵

The typical nonlinear refractive index parameter (n₂) could not be accurately determined, as the system appears to reach a saturation regime, where the birefringence response no longer follows a linear dependence on pump intensity. Nevertheless, for all of the studied compounds, we estimated the relevant parameters using eqn (8) (see eqn (S11)), which accounts for saturable absorption. This approach yielded excellent fits, with R² values exceeding > 0.99 for all samples. Importantly, based on these estimations, we determined the saturation intensity values, which were remarkably low, below approximately 22 mW cm⁻². We also calculated the maximum achievable birefringence for each compound. In all cases, the change in the refractive index (Δn) was found to be significantly high, in the range of 10⁻⁴ to 10⁻³. This combination of high birefringence and low saturation intensity highlights the strong potential of these materials for use in low-power-consumption optical devices. Furthermore, the evaluation of saturation intensity and maximum birefringence provides a more meaningful assessment of the materials’ utility in photoswitching applications than the standard n₂ parameter alone.

The dynamics of birefringence changes during growth and decay observed in our materials exhibit a complex, triple-exponential behavior, which is relatively uncommon in the literature. In the studied compounds, the birefringence build-up process can be characterized by fast time constants ranging from 0.28 to 1.20 s, depending on the specific compound (see Fig. 6(b) and Table S6). The intermediate components vary from 4.91 to 18.8 s, while the slowest processes extend to tens or even hundreds of seconds. As expected, the fastest components generally contribute the largest amplitude to the overall birefringence, whereas the slowest components have the smallest contributions. These findings confirm the ability of the materials to modulate optical signals effectively; however, the presence of slower processes suggests the involvement of additional mechanisms in birefringence formation. These may include thermal effects, photo-fluidization, or depth-dependent absorption saturation, all of which require further investigation and fall beyond the scope of this work. Regarding birefringence decay, as previously mentioned, the systems do not fully revert to their initial states after the removal of the excitation light (see Fig. 6(a), Fig. S24 and Table 4), indicating a residual birefringence or the so-called “memory effect”. This behavior implies potential applicability in optical data storage. To further quantify this effect, we introduced an additional parameter, ξ (see Table 4), which represents the percentage contribution of the non-reversible birefringence component (Δn₀) to the total induced birefringence (Δn). Notably, all compounds demonstrated a measurable memory effect, with the highest ξ value (∼25%) observed for compound 1c and the lowest (∼8%) for 1d. These results suggest that the materials studied here could serve as multifunctional platforms: depending on the thresholding criteria, they could function either as optical memory elements or as reversible optical modulators. The abovementioned combination of fast transition (and even saturation under the low value of power pump intensities) and relatively low recovery with the material's high and constant birefringence are desired properties in development, for instance, in telecommunication devices⁴⁶ or modulators.⁴¹

Conclusions

A new family of TCF-based D–π–A molecules were synthesized via simple, cost-effective, and efficient synthetic routes. This series of compounds demonstrates a range of advantageous optical properties, including strong solid-state photoluminescence and efficient photoswitching capabilities. The electronic structures and the ability to form distinct, stable conformers were further supported by quantum chemical calculations. All synthesized compounds exhibited photoisomerization, characterized by well-defined isosbestic points and reversible switching behavior. Importantly, by tuning the number and positional arrangement of –OBn substituents in the donor segment of the molecules, we were able to modulate their photosensitivity, particularly in relation to steric hindrance, which plays a significant role in slowing the thermal relaxation to the more stable isomer. Efficient photoisomerization and photoswitching were observed in both liquid and polymeric systems at remarkably low pumping intensities, comparable to natural sunlight (∼100 mW cm⁻²⁴⁷), making these materials highly promising for applications requiring low-power operation and moderate switching speeds. Notably, all compounds demonstrated a measurable “memory effect”, further underscoring their versatility.

Author contributions

K. L.: investigation, conceptualization, methodology, data curation, formal analysis, verification, visualization, writing – original draft, and writing – review and editing; K. E. S.: conceptualization, resources, methodology, validation and writing – review and editing; K. P.: resources and methodology; R. G.: conceptualization, resources, methodology, validation and writing – review and editing; P. F.: resources, methodology, formal analysis and investigation, validation and writing – review and editing; Y. B.: conceptualization, resources, and writing – review and editing; C. A.: conceptualization, resources, and writing – review and editing; L. S.: conceptualization, methodology, formal analysis, data curation, visualization, writing – original draft, supervision, funding acquisition, and project administration.

Conflicts of interest

There are no conflicts to declare.

Data availability

The data supporting this article have been included as part of the SI with more comprehensive information to facilitate a deeper understanding of the study's findings. The SI contains the methodology, characterization of the synthesized compounds, and their general synthesis along with the results mentioned in the main text. See DOI: https://doi.org/10.1039/d5tc02366a.

Acknowledgements

K. L. and L. S. would like to thank the National Science Centre of Poland for financial support within grant no. 2020/39/O/ST5/0185. K. L. would like to thank the Polish National Agency for Academic Exchange for the financial support provided through internship no. BPN/PRE/2022/1/00028. Calculations have been carried out at the Wroclaw Centre for Networking and Supercomputing. The time-resolved spectrofluorometer used in this work has been funded by the National Laboratory for Photonics and Quantum Technologies, project number POIR.04.02.00.00-B003/18.

References

1. A. Abu-Rayyan, I. Ahmad, N. H. Bahtiti, T. Muhmood, S. Bondock and M. Abohashrh, et al., Recent Progress in the Development of Organic Chemosensors for Formaldehyde Detection, ACS Omega, 2023, 8(17), 14859–14872.
2. Y. Watanabe, H. Sasabe, D. Yokoyama, T. Beppu and H. Katagiri, Kido J. Synthesis, properties, and OLED characteristics of 2,2′-bipyridine-based electron-transport materials: The synergistic effect of molecular shape anisotropy and a weak hydrogen-bonding network on molecular orientation, J. Mater. Chem. C, 2016, 4(17), 3699–3704.
3. Q. Qiu, S. Yang, M. A. Gerkman, H. Fu, I. Aprahamian and G. G. D. Han, Photon Energy Storage in Strained Cyclic Hydrazones: Emerging Molecular Solar Thermal Energy Storage Compounds, J. Am. Chem. Soc., 2022, 144(28), 12627–12631.
4. I. Bala, J. T. Plank, B. Balamut, D. Henry, A. R. Lippert and I. Aprahamian, Multi-stage and multi-colour liquid crystal reflections using a chiral triptycene photoswitchable dopant, Nat. Chem., 2024, 16(12), 2084–2090.
5. D. Wang, J. Chen, Y. Wang, X. Hao, H. Peng and Y. Liao, et al., Photoswitching in a Liquid Crystalline Pt(II) Coordination Complex, Chem. – Eur. J., 2024, 30(20), e202304366.
6. A. M. Rice, C. R. Martin, V. A. Galitskiy, A. A. Berseneva, G. A. Leith and N. B. Shustova, Photophysics Modulation in Photoswitchable Metal–Organic Frameworks, Chemical Reviews, American Chemical Society, 2020, pp. 8790–813, vol. 120.
7. M. Mostaghimi, H. Pacheco Hernandez, Y. Jiang, W. Wenzel, L. Heinke and M. Kozlowska, On–off conduction photoswitching in modelled spiropyran-based metal–organic frameworks, Commun. Chem., 2023, 6(1), 275.
8. C. Hillel, S. Rough, C. J. Barrett, W. J. Pietro and O. Mermut, A cautionary tale of basic azo photoswitching in dichloromethane finally explained, Commun. Chem., 2024, 7(1), 250.
9. F. Xu, J. Sheng, C. N. Stindt, S. Crespi, W. Danowski and M. F. Hilbers, et al., All-visible-light-driven stiff-stilbene photoswitches, Chem. Sci., 2024, 15(18), 6763–6769.
10. S. Redon, G. Eucat, M. Ipuy, E. Jeanneau, I. Gautier-Luneau and A. Ibanez, et al., Tuning the solid-state emission of small push-pull dipolar dyes to the far-red through variation of the electron-acceptor group, Dyes Pigm., 2018, 156, 116–132.
11. A. Gonzalez, E. S. Kengmana, M. V. Fonseca and G. G. D. Han, Solid-state photoswitching molecules: structural design for isomerization in condensed phase, Materials Today Advances, Elsevier Ltd, 2020, vol. 6.
12. J. J. B. van der Tol, T. A. P. Engels, R. Cardinaels, G. Vantomme, E. W. Meijer and F. Eisenreich, Photoswitchable Liquid-to-Solid Transition of Azobenzene-Decorated Polysiloxanes, Adv. Funct. Mater., 2023, 33(36), 2301246.
13. M. Rémond, Z. Zheng, E. Jeanneau, C. Andraud, Y. Bretonnière and S. Redon, 4,5,5-Trimethyl-2,5-dihydrofuran-Based Electron-Withdrawing Groups for NIR-Emitting Push–Pull Dipolar Fluorophores, J. Org. Chem., 2019, 84(16), 9965–9974.
14. B. Wüstenberg and N. R. Branda, A photoswitchable donor-π-linker-acceptor system based on a modified hexatriene backbone, Adv. Mater., 2005, 17(17), 2134–2138.
15. J. M. Park, C. Y. Jung, W. D. Jang and J. Y. Jaung, Effect of donor-π-acceptor structure on photochromism of dithienylethene-based dyes, Dyes Pigm., 2020, 177, 108315.
16. L. Fang, Z. Lin, Y. Zhang, B. Ye, J. Li and Q. Ran, et al., Robust, Ultrafast and Reversible Photoswitching in Bulk Polymers Enabled by Octupolar Molecule Design, Angew. Chem., Int. Ed., 2024, 63(16), e202402349.
17. S. L. Walden, P. H. D. Nguyen, H. K. Li, X. Liu, M. T. N. Le and L. Xian Jun, et al., Visible light-induced switching of soft matter materials properties based on thioindigo photoswitches, Nat. Commun., 2023, 14(1), 8298.
18. A. Goulet-Hanssens, F. Eisenreich and S. Hecht, Enlightening Materials with Photoswitches, Adv. Mater., 2020, 32(20), 1905966.
19. G. Liu, J. Leng, Q. Zhou, Z. Deng, L. Shi and C. Fan, et al., Fluorescence photoswitch of stiff-stilbene derivatives for anti-counterfeiting, Dyes Pigm., 2022, 203, 110361,  DOI:10.1016/j.dyepig.2022.110361.
20. V. Trovato, S. Sfameni, G. Rando, G. Rosace, S. Libertino and A. Ferri, et al., A Review of Stimuli-Responsive Smart Materials for Wearable Technology in Healthcare: Retrospective, Perspective, and Prospective, Molecules, 2022, 27, 5709.
21. K. Li, Y. Xiang, X. Wang, J. Li, R. Hu and A. Tong, et al., Reversible photochromic system based on rhodamine B salicylaldehyde hydrazone metal complex, J. Am. Chem. Soc., 2014, 136(4), 1643–1649.
22. L. Hu, Q. Zhang, X. Li and M. J. Serpe, Stimuli-responsive polymers for sensing and actuation, Materials Horizons, Royal Society of Chemistry, 2019, pp. 1774–93, vol. 6.
23. S. Shruthi, M. Smahel, M. Kohout, G. Shanker and G. Hegde, Influence of linking units on the photo responsive studies of azobenzene liquid Crystals: Application in optical storage devices, J. Mol. Liq., 2021, 339, 116744,  DOI:10.1016/j.molliq.2021.116744.
24. J. Hu, T. Han, Y. Liu, X. Zhang, Y. Duan and Z. Li, et al., Strategy for optical data encryption and decryption using a D-A type stimuli-responsive AIE material, Spectrochim. Acta, Part A, 2020, 239.
25. T. Ganguly, P. Pal, A. Paul and S. Baitalik, Synthesis and manifold but controllable emission switching of stilbene-appended polyaromatic terpyridine derivatives via aggregation and trans–cis isomerization, J. Photochem. Photobiol., A, 2022, 430, 113966,  DOI:10.1016/j.jphotochem.2022.113966.
26. U. Endesfelder, S. Malkusch, B. Flottmann, J. Mondry, P. Liguzinski and P. J. Verveer, et al., Chemically induced photoswitching of fluorescent probes-A general concept for super-resolution microscopy, Molecules, 2011, 16(4), 3106–3118.
27. M. Ipuy, Y. Y. Liao, E. Jeanneau, P. L. Baldeck, Y. Bretonnière and C. Andraud, Solid state red biphotonic excited emission from small dipolar fluorophores, J. Mater. Chem. C, 2015, 4(4), 766–779.
28. D. J. Stewart, M. J. Dalton, S. L. Long, R. Kannan, Z. Yu and T. M. Cooper, et al., Steric hindrance inhibits excited-state relaxation and lowers the extent of intramolecular charge transfer in two-photon absorbing dyes, Phys. Chem. Chem. Phys., 2016, 18(7), 5587–5596.
29. Z. C. Yang, M. Wang, A. M. Yong, S. Y. Wong, X. H. Zhang and H. Tan, et al., Intrinsically fluorescent carbon dots with tunable emission derived from hydrothermal treatment of glucose in the presence of monopotassium phosphate, Chem. Commun., 2011, 47(42), 11615–11617.
30. T. Han, X. Gu, J. W. Y. Lam, A. C. S. Leung, R. T. K. Kwok and T. Han, et al., Diaminomaleonitrile-based Schiff bases: aggregation-enhanced emission, red fluorescence, mechanochromism and bioimaging applications, J. Mater. Chem. C, 2016, 4(44), 10430–10434.
31. J. Chen, D. Li, W. Chi, G. Liu, S. H. Liu and X. Liu, et al., A Highly Reversible Mechanochromic Difluorobenzothiadiazole Dye with Near-Infrared Emission, Chem. – Eur. J., 2018, 24(15), 3671–3676.
32. S. Dhar, D. K. Rana, S. Singha Roy, S. Roy, S. Bhattacharya and S. C. Bhattacharya, Effect of solvent environment on the Photophysics of a newly synthesized bioactive 7-oxy(5-selenocyanato-pentyl)-2H-1-benzopyran-2-one, J. Lumin., 2012, 132(4), 957–964.
33. C. Bannwarth, S. Ehlert and S. Grimme, GFN2-xTB – An Accurate and Broadly Parametrized Self-Consistent Tight-Binding Quantum Chemical Method with Multipole Electrostatics and Density-Dependent Dispersion Contributions, J. Chem. Theory Comput., 2019, 15(3), 1652–1671.
34. M. J. Frisch; G. W. Trucks; H. B. Schlegel; G. E. Scuseria; M. A. Robb; J. R. Cheeseman; G. Scalmani; V. Barone; G. A. Petersson; H. Nakatsuji; X. Li; M. Caricato; A. V. Marenich; J. Bloino; B. G. Janesko; R. Gomperts; B. Mennucci and Hratch, Gaussian 16, Revision C.01, 2016.
35. T. Shiozaki, BAGEL: Brilliantly Advanced General Electronic-structure Library, Wiley Interdiscip. Rev.: Comput. Mol. Sci., 2018, 8(1), e1331.
36. F. Neese, F. Wennmohs, U. Becker and C. Riplinger, The ORCA quantum chemistry program package, J. Chem. Phys., 2020, 152(22), 224108.
37. B. De Souza, F. Neese and R. Izsák, On the theoretical prediction of fluorescence rates from first principles using the path integral approach, J. Chem. Phys., 2018, 148(3), 035204.
38. A. Airinei, N. Fifere, M. Homocianu, C. Gaina, V. Gaina and B. C. Simionescu, Optical properties of some new azo photoisomerizable bismaleimide derivatives, Int. J. Mol. Sci., 2011, 12(9), 6176–6193.
39. A. Natansohn and P. Rochon, Photoinduced motions in azo-containing polymers, Chem. Rev., 2002, 102(11), 4139–4175.
40. A. Szukalski, K. Haupa, A. Miniewicz and J. Mysliwiec, Photoinduced birefringence in PMMA polymer doped with photoisomerizable pyrazoline derivative, J. Phys. Chem. C, 2015, 119(18), 10007–10014.
41. A. Szukalski, R. Stottko, P. Krawczyk, B. Sahraoui and B. Jędrzejewska, Application of the pyrazolone derivatives as effective modulators in the opto-electronic networks, J. Photochem. Photobiol., A, 2023, 437, 114482.
42. L. M. Silva, D. L. Silva, M. V. Boas, Y. Bretonniere, C. Andraud and M. G. Vivas, Probing the high performance of photoinduced birefringence in V-shaped azo/PMMA guest–host films, RSC Adv., 2020, 10(67), 40806–40814.
43. F. Fan, Y. Liu, Y. Hong, J. Zang, A. Wu and T. Zhao, et al., Improving the polarization-holography performance of PQ/PMMA photopolymer by doping with THMFA, Opt. Express, 2018, 26(14), 17794.
44. C. R. Mendonça, L. Misoguti, A. A. Andrade, S. B. Yamaki, V. D. Dias and T. D. Z. Atvars, et al., Photoinduced birefringence in di-azo compounds in polystyrene and poly(methyl methacrylate) guest-host systems, Opt. Mater., 2007, 30(2), 216–221.
45. B. N. Sunil, W. S. Yam and G. Hegde, Photoresponsive behavior of hydrophilic/hydrophobic-based novel azobenzene mesogens: Synthesis, characterization and their application in optical storage devices, RSC Adv., 2019, 9(69), 40588–40606.
46. P. Drexler and P. Fial, Optical Fiber Birefringence Effects – Sources, Utilization and Methods of Suppression. Recent Progress in Optical Fiber, Recent Progress in Optical Fiber Research, 2012, pp. 127–150.
47. D. R. Myers, Solar radiation modeling and measurements for renewable energy applications: Data and model quality, Energy, 2005, 30(9), 1517–1531.

---