Improving the Luminescence and Stability of Carbon-Centered Radicals by Kinetic Isotope Effect

The kinetic isotope effect (KIE) is beneficial to improve the performance of luminescent molecules and relevant light-emitting diodes. In this work, the influences of deuteration on the photophysical property and stability of luminescent radicals are investigated for the first time. Four deuterated radicals based on biphenylmethyl, triphenylmethyl, and deuterated carbazole were synthesized and sufficiently characterized. The deuterated radicals exhibited excellent redox stability, as well as improved thermal and photostability. The appropriate deuteration of relevant C-H bonds would effectively suppress the non-radiative process, resulting in the increase in photoluminescence quantum efficiency (PLQE). This research has demonstrated that the introduction of deuterium atoms could be an effective pathway to develop high-performance luminescent radicals.


Introduction
Organic radicals with unpaired electrons exhibit great application prospects for their unique optical, electrical, and magnetic properties [1][2][3][4][5][6][7][8][9][10]. Especially, through the doublet emission process, the theoretical upper limit of the internal quantum efficiency for radicalbased light-emitting devices can reach 100% [11,12]. In order to improve the properties of luminescent radicals, several molecular design strategies were proposed. Guo et al. designed a few triphenylmethyl derivatives with different electron-rich groups, the electron structure of which did not follow the Aufbau principle, which exhibited higher photostability and photoluminescence quantum efficiency (PLQE) [13]. Alim et al. proposed the introduction of specific groups to construct non-alternant hydrocarbon which was beneficial to improve PLQE [14]. Mattiello et al. improved the photostability of PyBTM derivatives with the introduction of four phenyl groups, which improved even further when extra methoxy groups were added [15]. With terminal benzene rings on carbazole, Matsuda et al. improved the photostability of TTM-1Cz radicals efficiently [16]. Moreover, many studies have explored the effects of different substituent groups on the properties of luminescent bi-or triphenylmethyl radicals [17][18][19][20][21][22].
In addition to incorporating different substituted groups, the substitution of isotopes on relevant C-H bonds could also influence the properties of luminescent molecules, especially PLQE and stability [23,24]. Although the steric and electronic configurations are barely influenced, isotopic replacement of the protium atom (H) by the deuterium atom (D) would enormously change the bond stretch bands and decrease the energy of vibration modes, namely the kinetic isotope effect (KIE) of H/D substitution. Usually, after being substituted by D, the non-radiative transitions of molecules would be suppressed and the rates of relevant C-H(D) bond breaking would be decreased, resulting in higher PLQE and better stability under heated or irradiated conditions, even in working lightemitting diodes. Thus, the deuteration of relevant chemical bonds has become an effective way to develop high-performance luminescent materials, as well as light-emitting diodes based on closed-shell molecules with fluorescence, phosphorescence, or thermally activated delayed fluorescence (TADF) [25][26][27][28][29]. However, up until now, the KIE on properties of open-shell luminescent radicals has not been reported. Hence, in this work, molecules based on different luminescent radical systems with different numbers of deuterated carbazole groups were synthesized to investigate the KIE on properties of open-shell radicals, especially PLQE and stability.
Molecules 2023, 28, x FOR PEER REVIEW 2 of 10 being substituted by D, the non-radiative transitions of molecules would be suppressed and the rates of relevant C-H(D) bond breaking would be decreased, resulting in higher PLQE and better stability under heated or irradiated conditions, even in working lightemitting diodes. Thus, the deuteration of relevant chemical bonds has become an effective way to develop high-performance luminescent materials, as well as light-emitting diodes based on closed-shell molecules with fluorescence, phosphorescence, or thermally activated delayed fluorescence (TADF) [25][26][27][28][29]. However, up until now, the KIE on properties of open-shell luminescent radicals has not been reported. Hence, in this work, molecules based on different luminescent radical systems with different numbers of deuterated carbazole groups were synthesized to investigate the KIE on properties of open-shell radicals, especially PLQE and stability.

Photophysical Properties
The ultraviolet-visible (UV-Vis) absorption spectra of four radicals were measured in cyclohexane solvent ( Figure 2). Most of them show three typical absorption bands. The strong absorption bands, peaking at 285 nm for BTM-1DCz and BTM-2DCz, and 290 nm for TTM-1DCz and TTM-2DCz, come from carbazole. The medium absorption bands,

Photophysical Properties
The ultraviolet-visible (UV-Vis) absorption spectra of four radicals were measured in cyclohexane solvent ( Figure 2). Most of them show three typical absorption bands. The strong absorption bands, peaking at 285 nm for BTM-1DCz and BTM-2DCz, and 290 nm for TTM-1DCz and TTM-2DCz, come from carbazole. The medium absorption bands, peaking at 375 and 387 nm for TTM series and BTM-1DCz, and 369 and 403 nm for BTM-2DCz, are the characteristic absorption from carbon-centered radicals. Additionally, the weak absorption bands of four radicals, peaking from 450 to 700 nm which are not identical, mainly come from the intramolecular charge-transfer states compared with previous research [30,31]. Compared to the non-deuterated molecules, there is no significant change for the UV-Vis absorption spectra since there was no obvious transformation for both steric and electronic configurations between deuterated and non-deuterated molecules. In addition, BTM-2DCz, the only one whose non-deuterated molecule has not been reported before, shows some differences in absorption spectrum. There is an obvious absorption band peaking at 470 nm different from BTM-1DCz, and slight absorption enhancement around 600 nm. That is because two carbazole groups in BTM-2DCz are located at different sites. One is similar to the site in BTM-1DCz, and the other is similar to the site in TTM-1DCz. Thus, there would be two possible types of intramolecular charge-transfer in BTM-2DCz, resulting in the changes of long-wavelength absorption compared to the spectra of BTM-1DCz and TTM-1DCz. With increasing carbazole groups, the absorption of carbazole around 280 nm clearly increases. At the same time, the ratios of absorption from charge-transfer state at long wavelengths to absorption from carbon-center radicals around 380 nm are also increased, which indicates more characteristics of the charge-transfer state in more carbazole-substituted radicals. Absorption spectra in different solvents with different polarity were also measured ( Figure S5), and no obvious change could be found.
Molecules 2023, 28, x FOR PEER REVIEW 3 peaking at 375 and 387 nm for TTM series and BTM-1DCz, and 369 and 403 nm for B 2DCz, are the characteristic absorption from carbon-centered radicals. Additionall weak absorption bands of four radicals, peaking from 450 to 700 nm which are not tical, mainly come from the intramolecular charge-transfer states compared with pre research [30,31]. Compared to the non-deuterated molecules, there is no signi change for the UV-Vis absorption spectra since there was no obvious transformatio both steric and electronic configurations between deuterated and non-deuterated cules. In addition, BTM-2DCz, the only one whose non-deuterated molecule has not reported before, shows some differences in absorption spectrum. There is an obviou sorption band peaking at 470 nm different from BTM-1DCz, and slight absorptio hancement around 600 nm. That is because two carbazole groups in BTM-2DCz a cated at different sites. One is similar to the site in BTM-1DCz, and the other is simi the site in TTM-1DCz. Thus, there would be two possible types of intramolecular ch transfer in BTM-2DCz, resulting in the changes of long-wavelength absorption comp to the spectra of BTM-1DCz and TTM-1DCz. With increasing carbazole groups, th sorption of carbazole around 280 nm clearly increases. At the same time, the ratios sorption from charge-transfer state at long wavelengths to absorption from carbon-c radicals around 380 nm are also increased, which indicates more characteristics o charge-transfer state in more carbazole-substituted radicals. Absorption spectra in d ent solvents with different polarity were also measured ( Figure S5), and no ob change could be found. Photoluminescence (PL) spectra of deuterated radicals were measured in cycl ane solvent (Figure 2). BTM-1DCz and BTM-2DCz exhibit weak deep-red emission ing at 711 and 706 nm, respectively. The absolute PLQE values measured by integr sphere are 3.0% for BTM-1DCz and 3.6% for BTM-2DCz. TTM-1DCz and TTMexhibit bright red emission peaking at 638 and 647 nm with considerable absolute P values of 78.4 and 56.7%, respectively. Similar to absorption spectra, besides slight shift, there is no obvious change in PL spectra of all deuterated radicals compared to deuterated molecules ( Figure 2). It is noteworthy that the PLQE values of BTM-1DC TTM-1DCz in cyclohexane are almost 1.5 times as high as the relevant non-deute molecules reported in the literature [30,31].
To better explore the KIE on the luminescent properties of deuterated radical transient PL decays of four radicals in cyclohexane were measured, as well a Photoluminescence (PL) spectra of deuterated radicals were measured in cyclohexane solvent ( Figure 2). BTM-1DCz and BTM-2DCz exhibit weak deep-red emission peaking at 711 and 706 nm, respectively. The absolute PLQE values measured by integrating sphere are 3.0% for BTM-1DCz and 3.6% for BTM-2DCz. TTM-1DCz and TTM-2DCz exhibit bright red emission peaking at 638 and 647 nm with considerable absolute PLQE values of 78.4 and 56.7%, respectively. Similar to absorption spectra, besides slight blue-shift, there is no obvious change in PL spectra of all deuterated radicals compared to non-deuterated molecules ( Figure 2). It is noteworthy that the PLQE values of BTM-1DCz and TTM-1DCz in cyclohexane are almost 1.5 times as high as the relevant non-deuterated molecules reported in the literature [30,31].
To better explore the KIE on the luminescent properties of deuterated radicals, the transient PL decays of four radicals in cyclohexane were measured, as well as the unreported relevant photophysical parameters of BTM-2Cz. According to the measurements, radiative and non-radiative transition rates of luminescent radicals, k r and k nr , were calcu-lated, respectively. All the relevant photophysical parameters are summarized in Table 1. From the results, the k nr values of BTM-1DCz and BTM-2DCz decrease significantly from 245.0 × 10 6 and 241.8 × 10 6 to 215.5 × 10 6 and 214.2 × 10 6 s −1 compared to the relevant non-deuterated molecules. A similar reduction could also be found for the k nr values of TTM-1DCz and TTM-2DCz. However, the k r values of these four deuterated radicals show few variations. These photophysical results indicate that the deuteration of luminescent radicals does not influence the pathway and probability of radiative transitions, but significantly influences the non-radiative transitions. Because of the KIE from deuterium atoms, the non-radiative transitions of radicals were suppressed effectively, resulting in higher PLQE. The longer fluorescence lifetimes of deuterated radicals also indicate the effective suppression of bond breaking, that is to say the excited states of deuterated radicals become more stable. No matter which radical systems are used, bi-or triphenylmethyl radicals, more deuterated carbazole groups do not yield stronger suppression, so the PLQE values of BTM-2DCz and TTM-2DCz show little increases. Meanwhile, if the ratios of k nr to k r of luminescent radicals are as larger as biphenylmethyl radical systems, the influences on PLQE from the KIE of deuteration would be weak. For luminescent radicals with more balanced k nr and k r , appropriate deuteration could significantly improve PLQE values.

Electrochemical Properties
Cyclic voltammetry (CV) analysis was performed to study the redox characteristics of deuterated radicals. Similar to non-deuterated radicals ( Figure S7 (Table S1).
To test the redox stability of deuterated radicals, continuous multicycle (20 cycles) CV measurements were performed. From the relevant CV curves ( Figure S8), neither the potentials nor the signal intensities of the redox peaks changed clearly. These results demonstrate that the unpaired electron features and molecular structures would not change after continuous multicycle CV scan, indicating the excellent redox stability of deuterated radicals. To test the redox stability of deuterated radicals, continuous multicycle (20 cy CV measurements were performed. From the relevant CV curves ( Figure S8), neithe potentials nor the signal intensities of the redox peaks changed clearly. These re demonstrate that the unpaired electron features and molecular structures would change after continuous multicycle CV scan, indicating the excellent redox stabili deuterated radicals.

Theoretical Calculations
Density functional theory (DFT) calculations (B3LYP/6-31G(d,p)) and time-dep ent density functional theory (TD-DFT) calculations (CAM-B3LYP/6-31 G(d,p)) were formed to further investigate the electronic structures and transitions of deuterated cals. Frequency calculations were carried out to ensure that all the optimized struct were minima on the potential energy surface. Figure 4 shows the optimized ground structures and frontier orbitals of deuterated radicals. Compared to non-deuterated cals, the calculation results did not clearly change due to the negligible influence on tronic structures from deuteration. The frontier energy levels calculated from DFT ar most the same as the SOMO and SUMO energy levels calculated from the CV meas ments (Table S2). From the results of optimized TD-DFT calculations, the emission pro of radicals, namely the transitions between the lowest doublet excited states (D1) and doublet ground states (D0), are mainly the transitions between SUMO and β-HOMO the electron cloud changes of these transitions are from the donors, carbazole parts to radical centers, demonstrating the charge-transfer feature of luminescence from t radicals. The relevant calculation results of excited states are summarized in Table S3

Theoretical Calculations
Density functional theory (DFT) calculations (B3LYP/6-31G(d,p)) and time-dependent density functional theory (TD-DFT) calculations (CAM-B3LYP/6-31 G(d,p)) were performed to further investigate the electronic structures and transitions of deuterated radicals. Frequency calculations were carried out to ensure that all the optimized structures were minima on the potential energy surface. Figure 4 shows the optimized ground state structures and frontier orbitals of deuterated radicals. Compared to non-deuterated radicals, the calculation results did not clearly change due to the negligible influence on electronic structures from deuteration. The frontier energy levels calculated from DFT are almost the same as the SOMO and SUMO energy levels calculated from the CV measurements (Table S2). From the results of optimized TD-DFT calculations, the emission process of radicals, namely the transitions between the lowest doublet excited states (D 1 ) and the doublet ground states (D 0 ), are mainly the transitions between SUMO and β-HOMO. All the electron cloud changes of these transitions are from the donors, carbazole parts to the radical centers, demonstrating the charge-transfer feature of luminescence from these radicals. The relevant calculation results of excited states are summarized in Table S3.

Thermal and Photostability
Stability is a vital property in influencing the applications of luminescent radicals in most fields. Besides the redox stability discussed above, the thermal stability was also evaluated using thermogravimetric analysis (TGA). The results show that the thermal decomposition temperatures of the deuterated radicals are raised in different degrees due to the KIE from deuteration ( Figure S9). In particular, the thermal decomposition temperatures of BTM-1DCz and TTM-2DCz are about 20 • C higher than non-deuterated radicals.
Another important stability that needs to be considered for luminescent radicals is photostability. The decays of fluorescence intensity from deuterated radicals were recorded under continuous irradiation from xenon lamp and contrasted with non-deuterated radicals and TTM radicals. The data of these decays were fitted, and the half-life was calculated ( Figure 5). The half-life values of the deuterated radicals were 5.39 × 10 3 (BTM-1DCz), 1.32 × 10 4 (BTM-2DCz), 8.57 × 10 3 (TTM-1DCz), and 1.54 × 10 4 s (TTM-2DCz), respectively, indicating excellent photostability. Even compared with non-deuterated radicals, the photostability increases around 2 to 10 times ( Figure S10). These results indicate that the photostability of luminescent radicals could be notably increased by deuteration, namely the excited states of luminescent radicals would be more stable due to the KIE, which is beneficial for applications.

Thermal and Photostability
Stability is a vital property in influencing the applications of luminescent radicals in most fields. Besides the redox stability discussed above, the thermal stability was also evaluated using thermogravimetric analysis (TGA). The results show that the thermal decomposition temperatures of the deuterated radicals are raised in different degrees due to the KIE from deuteration ( Figure S9). In particular, the thermal decomposition temperatures of BTM-1DCz and TTM-2DCz are about 20 °C higher than non-deuterated radicals.
Another important stability that needs to be considered for luminescent radicals is photostability. The decays of fluorescence intensity from deuterated radicals were recorded under continuous irradiation from xenon lamp and contrasted with non-deuterated radicals and TTM radicals. The data of these decays were fitted, and the half-life was calculated ( Figure 5  respectively, indicating excellent photostability. Even compared with non-deuterated radicals, the photostability increases around 2 to 10 times ( Figure S10). These results indicate that the photostability of luminescent radicals could be notably increased by deuteration, namely the excited states of luminescent radicals would be more stable due to the KIE, which is beneficial for applications.

Materials and Methods
All chemical reagents and starting materials used in this work were purchased from ERNEGI and Xilong Science Co., Ltd. (Shanghai, China) without further purification. The synthetic routes of target radicals followed the previous literature and can be found in

Materials and Methods
All chemical reagents and starting materials used in this work were purchased from ERNEGI and Xilong Science Co., Ltd. (Shanghai, China) without further purification. The synthetic routes of target radicals followed the previous literature and can be found in Supplementary Materials Figure S1.
HRMS were recorded on a Shimadzu LCMS-IT-TOF. Infrared spectra were recorded on a Bruker Tensor 27 spectrometer. Elemental analysis data were recorded on an Elementar Vario microcube spectrometer. EPR data were measured using a Bruker Instruments A320 spectrometer with 10 −5 M concentration in cyclohexane at room temperature. The UV-Vis spectra in various solvents were measured with a Shimadzu UV-1900i UV-Vis spectrometer. The PL spectra in cyclohexane were measured with a Shimadzu RF-6000 spectrometer. The transient PL decays were recorded using an Edinburgh FLS1000 spectrometer, and the absolute PLQEs were recorded on the same instrument using the integrating sphere method. DFT and TD-DFT calculations were performed on Gaussian16 commercial software (Revision C.02) [32]. Thermal stability measurements were performed on a TA INSTRUMENTS Q600 TGA analyzer under air and a ramp rate of 10 • C·min −1 . The CV measurements were performed using a CH Instruments CHI660E electrochemical analyzer with a glassy carbon electrode as the working electrode, a platinum wire as the counter electrode, and Ag/AgCl as the reference electrode. Redox couple ferrocenium/ferrocene was used as an internal standard. The photostability was also recorded on RF-6000 under continuous irradiation from a xenon lamp.

Synthesis of BTM-1DCz and BTM-2DCz
Compound BTM-1DCz and BTM-2DCz were prepared following the literature procedure ( Figure S1) [30]. Purple-black solid BTM-1DCz and dark green solid BTM-2DCz were simultaneously obtained from the final step with yields of 22

Synthesis of TTM-1DCz and TTM-2DCz
Compound TTM-1DCz and TTM-2DCz were prepared following the procedure in the literature ( Figure S2) [31]. Reddish-brown solid TTM-1DCz was obtained with 71% yield of relevant step, and dark green solid TTM-2DCz was obtained with 83% yield of relevant step.

Conclusions
In summary, four deuterated luminescent radicals were synthesized to explore the KIE on luminescent radicals. The non-radiative process could be effectively suppressed by deuteration, which is beneficial to the PLQE of luminescent radicals. In particular, the PLQE value of TTM-1DCz significantly increased from 53.0 to 78.4%, and its k nr deceased from 19 × 10 6 to 5 × 10 6 s −1 . The KIE of deuteration also made the luminescent radicals more stable, including redox stability, thermal stability, and photostability. Especially compared with non-deuterated radicals, the photostability of TTM-2DCz increased almost 10 times. These results from this paper demonstrate that the deuteration of relevant C-H bonds would be an effective pathway to develop high-performance luminescent radicals, especially to the non-deuterated luminescent radicals with balanced k nr and k r . The influence of deuteration on the properties of radical-based light-emitting diodes are also under investigation.
Supplementary Materials: The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/molecules28124805/s1, Figure S1: The synthetic routes of deuterated radicals. Figure S2: Mass spectrometry of deuterated radicals. Figure S3: Infrared absorption (IR) spectrogram of deuterated radicals. Figure S4: EPR of deuterated radicals in cyclohexane solution at room temperature. Figure S5: UV-Vis absorption spectra of deuterated radicals in solutions of different polarities (10−5 M). Figure S6: Transient fluorescence decay of deuterated radicals in cyclohexane solution (10−5 M). Figure S7: Voltammetry (CV) curves of non-deuterated radicals. Figure S8: Voltammetry (CV) curves of deuterated radicals for multiple (20-turn) cycles. Figure S9: TGA curve of deuterated radicals and non-deuterated radicals. Figure S10: Photostability of deuterated radicals and non-deuterated radicals. Table S1: Redox potentials and corresponding orbital energy levels of deuterated radicals and non-deuterated radicals. Table S2: The corresponding SOMO orbital energy levels calculated theoretically and measured experimentally. Table S3: Parameters corresponding to the emission bands in TD-DFT calculations of deuterated radicals.