Dynamics of the Ligand Excited States Relaxation in Novel β-Diketonates of Non-Luminescent Trivalent Metal Ions

Complexes emitting in the blue spectral region are attractive materials for developing white-colored light sources. Here, we report the luminescence properties of novel coordination compounds based on the trivalent group 3, 13 metals, and the 1-phenyl-3-methyl-4-cyclohexylcarbonyl-pyrazol-5-onate (QCH) ligand. [M(QCH)3] (M = Al, Ga, and In), [M(QCH)3(H2O)] (M = Sc, Gd, and Lu), [Lu(QCH)3(DMSO)], and [La(QCH)3(H2O)(EtOH)] complexes were synthesized and structurally characterized by a single-crystal X-ray diffraction study. It has been found that the luminescence quantum yields of the ligand increase by one order of magnitude upon metal coordination. A significant correspondence between the energies of the ligand’s excited states and the luminescence quantum yields to the metal ion’s atomic numbers was found using molecular spectroscopy techniques. The replacement of the central ion with the heavier one leads to a monotonic increase in singlet state energy, while the energy of the triplet state is similar for all the complexes. Time-resolved measurements allowed us to estimate the intersystem crossing (ISC) rate constants. It was shown that replacing the Al3+ ion with the heavier diamagnetic Ga3+ and In3+ ions decreased the ISC rate, while the replacement with the paramagnetic Gd3+ ion increased the ISC rate, which resulted in a remarkably bright and room-temperature phosphorescence of [Gd(QCH)3(H2O)].


Introduction
White organic light-emitting diodes (WOLEDs) are the most economical light sources for street, home, and display lighting and many other special applications [1][2][3]. White emissions in such devices are conditioned by simultaneous emissions of several luminophores in blue, green, and red spectral areas [2], providing an R-G-B scheme of white light [3], or emissions of blue and orange emitters, providing a B-O scheme. One of the most popular classes of materials used in WOLEDs are the platinum-based group materials, especially those based on Ir(III) complexes containing 2-phenylpyridine fluorinated derivatives and imidazole-type carbene compounds [4,5]. The commercial price of production of such materials is relatively high, which makes the technology quite expensive [6].
An attractive class of blue light-emitting materials is formed by small organic molecules, such as triarylboranes and triarylmethanes [7,8]. A great interest has arisen in recent years Corresponding lanthanide derivatives have been widely investigated, but in most of cases the attention has been paid on the luminescence properties of emitting f-elements, mainly terbium and europium [27,28].
Here, we report two series of coordination compounds obtained from the reaction of the trivalent metals, i.e., the 13 group metals (Al 3+ , Ga 3+ , and In 3+ ), 3 group metals (Sc 3+ , La 3+ , Gd 3+ , and Lu 3+ ), and the proligand (4-(cyclohexanecarbonyl)-5-methyl-2-phenyl-2,4dihydro-3H-pyrazol-3-one HQ CH , with the aim of qualitatively and quantitatively correlating the photophysical ligand parameters to the central ion choice. We select this ligand since it allows the synthesis of compounds containing ions of various radii from Al 3+ to Lu 3+ . Earlier, we showed that Q CH lanthanide complexes exhibit excellent luminescent properties [42,43], and that the cyclohexyl substituent prevents the formation of molecular aggregates due to the molecular volume [44].
In order to disclose the influence of different metal ions on the ligand's electronic structure, we also investigated in detail the photophysical properties of all the complexes by exploring the absorption, excitation, emission spectra, quantum yields of luminescence, and the lifetime of the excited states. All of the complexes reported here exhibit strong emissions in the blue-green region of the spectra, which makes them promising for use as blue emitting layer components in WOLEDs.

Synthesis of Complexes
The complexes [M(Q CH )3] 2-4 (M = Al, Ga, and In) of all the three elements can be readily prepared in high yields in aqueous EtOH media, using hydrated salts as the metal precursors and NaOH as the base (Scheme 1): Scheme 1. Preparation of Al 3+ , Ga 3+ , and In 3+ complexes (2)(3)(4).
All three complexes were obtained in an anhydrous form and can be purified by recrystallization from hot EtOH. Gallium (III) nitrate was used as a source of Ga 3+ ion as it is the most soluble salt. The nature of the anion did not affect the yield of the complexes. Scheme 1. Preparation of Al 3+ , Ga 3+ , and In 3+ complexes (2)(3)(4).
All three complexes were obtained in an anhydrous form and can be purified by recrystallization from hot EtOH. Gallium (III) nitrate was used as a source of Ga 3+ ion as it is the most soluble salt. The nature of the anion did not affect the yield of the complexes.
The interaction of HQ CH (1) with Sc(ClO 4 ) 3 and NaOH in the EtOH-H 2 O mixture led to the formation of [Sc(Q CH ) 3 (H 2 O)] n (5), as it was confirmed by an EA and FTIR data. This complex is insoluble in most solvents, including alcohols, which may testify to its polymeric structure with bridge water molecules. Previously, we have shown that for Sc 3+ diketonates with heterocyclic ligands, a coordination number (CN) equal to seven is preferable [15]. This complex is readily soluble in coordinating solvents, such as DMSO, upon gentle heating. A ligand exchange occurs due to dissolution with the formation of a monomeric species (Scheme 2). It is worth noting that the Al 3+ , Ga 3+ , and In 3+ complexes containing acylpyrazolonates have been scarcely studied to date [26]. Several aluminum complexes have been previously obtained from the interaction of Al(i-PrO)3 or anhydrous AlCl3 with three different 4-acylpyrazolones-5, bearing 4-acetyl, 4-benzyl, or 4-propionyl fragments in benzene [45]. Analogous In 3+ complexes were prepared using In(i-PrO)3 as the starting material with a method also employed for the preparation of Al 3+ ion complexes [46]. Al 3+ and In 3+ ion derivatives of 1-phenyl-3-methyl-4-trifluoroacetyl-pyrazolonate [47] and a binuclear In 3+ complex with the multitopic 1,10-bis(1-phenyl-3-methyl-5-hydroxy-4-pyrazolyl)-1,10decanedionate were synthetized and investigated in the extraction of indium and aluminum from the solutions [48]. Only Ga(1-phenyl-3-methyl-4-benzoyl-pyrazolonate) was reported in the literature without a detailed description of its preparation [49].
The interaction of HQ CH (1) with Sc(ClO4) 3 and NaOH in the EtOH-H2O mixture led to the formation of [Sc(Q CH )3(H2O)]n (5), as it was confirmed by an EA and FTIR data. This complex is insoluble in most solvents, including alcohols, which may testify to its polymeric structure with bridge water molecules. Previously, we have shown that for Sc 3+ diketonates with heterocyclic ligands, a coordination number (CN) equal to seven is preferable [15]. This complex is readily soluble in coordinating solvents, such as DMSO, upon gentle heating. A ligand exchange occurs due to dissolution with the formation of a monomeric species (Scheme 2).

Scheme 2. Preparation of Sc 3+ complexes.
Notably, upon the slow diffusion of EtOH vapor into a saturated DMSO solution of [Sc(Q CH )3(H2O)]n (5) fine crystals of [Sc(Q CH )3(DMSO)] (6), a complex formed as a sole product due to the higher coordination ability of DMSO over EtOH. The choice of Sc(ClO4)3 is not critical for the synthesis; other soluble salts, such as chlorides or nitrates, can be used. However, it is more convenient to dissolve Sc2O3 in non-volatile HClO4 rather than in HCl or HNO3.
For scandium (III), only one complex with 1-phenyl-3-methyl-4-benzoyl-pyrazolone-5 (HL1) was reported [50]. It was obtained by an interaction between the free ligand, hydrated Sc(NO3)3 in MeOH without a base, and identified as [Sc(L1)3]·H2O on the basis of an elemental analysis (EA) and FTIR data. Upon crystallization from hot MeOH, this amorphous complex transformed into anhydrous crystalline [Sc(L1)3], but no crystal structure data were provided. Complexes of La 3+ and Lu 3+ ions were obtained by a modified method, which has been described previously in the literature for other lanthanides [42,51,52]  Notably, upon the slow diffusion of EtOH vapor into a saturated DMSO solution of [Sc(Q CH ) 3 (H 2 O)] n (5) fine crystals of [Sc(Q CH ) 3 (DMSO)] (6), a complex formed as a sole product due to the higher coordination ability of DMSO over EtOH. The choice of Sc(ClO 4 ) 3 is not critical for the synthesis; other soluble salts, such as chlorides or nitrates, can be used. However, it is more convenient to dissolve Sc 2 O 3 in non-volatile HClO 4 rather than in HCl or HNO 3 .
For scandium (III), only one complex with 1-phenyl-3-methyl-4-benzoyl-pyrazolone-5 (HL 1 ) was reported [50]. It was obtained by an interaction between the free ligand, hydrated Sc(NO 3 ) 3 in MeOH without a base, and identified as [Sc(L 1 ) 3 ]·H 2 O on the basis of an elemental analysis (EA) and FTIR data. Upon crystallization from hot MeOH, this amorphous complex transformed into anhydrous crystalline [Sc(L 1 ) 3 ], but no crystal structure data were provided. Complexes of La 3+ and Lu 3+ ions were obtained by a modified method, which has been described previously in the literature for other lanthanides [42,51,52] (Scheme 3).
Since the ionic radius of La 3+ is bigger than that of Lu 3+ due to lanthanide contraction, La 3+ demonstrates a higher CN (8) and adopts two additional ligands (EtOH and water molecules) alongside three bulky diketonate anions. For the Gd 3+ ion and especially Lu 3+ , the ion coordination number is 7, and only one additional water molecule can be inserted in the inner sphere of the complex together with three anions of Q CH ligands. From hot EtOH complex [La(Q CH ) 3  Since the ionic radius of La 3+ is bigger than that of Lu 3+ due to lanthanide contraction, La 3+ demonstrates a higher CN (8) and adopts two additional ligands (EtOH and water molecules) alongside three bulky diketonate anions. For the Gd 3+ ion and especially Lu 3+ , the ion coordination number is 7, and only one additional water molecule can be inserted in the inner sphere of the complex together with three anions of Q CH ligands. From hot EtOH complex [La(Q CH )3(H2O)(EtOH)] (7) crystallized as a solvate with one molecule of EtOH, but it can be fully desolvated by heating at 45 °C at a diminished pressure.  Table 1. The isostructurality of single-crystals to polycrystalline bulk samples was confirmed by the powder X-ray diffraction method (PXRD) (Figures S1-S3).  Table 1. The isostructurality of single-crystals to polycrystalline bulk samples was confirmed by the powder X-ray diffraction method (PXRD) (Figures S1-S3). All three structures are mononuclear complexes, where the metal ion is coordinated by three diketonate ligands (Figures 1 and 2), and according to the CCDC analysis, these structures present the first example of such complexes of Al 3+ , Ga 3+ , and In 3+ ions with β-diketones. It must be noted that the asymmetric unit of the [In(Q CH ) 3 ] (3) crystal structure contains two molecules of the complex ( Figure 2). Each central ion (Al 3+ , Ga 3+ , and In 3+ ) bonds with two oxygen atoms of each ligand, leading to the octahedral coordination polyhedron {MO 6 } and neutral charge of the complexes. As for the {InO 6 } polyhedron, moderate distortion of the angle between two vertices can be observed, leading the O1-In1-O6 and O7-In2-O9 angles to be 171. 2 3 ], which is attributed to the increase in the ionic radius of the central metal ion. The analysis of crystal packing revealed no presence of any hydrogen bonds, but the presence of rather weak C-H . . . π was observed. yhedron {MO6} and neutral charge of the complexes. As for the {InO6} polyhedron, moderate distortion of the angle between two vertices can be observed, leading the O1-In1-O6 and O7-In2-O9 angles to be 171. 2 and 173.2° instead of 180° for the ideal octahedron. The elongation of M-O bonds (Table S1) is observed for complexes [Al(Q CH )3], [Ga(Q CH )3], and [In(Q CH )3], which is attributed to the increase in the ionic radius of the central metal ion. The analysis of crystal packing revealed no presence of any hydrogen bonds, but the presence of rather weak C-H…π was observed.   9), which are suitable for single-crystal X-ray diffraction, were obtained by the slow evaporation of solutions in EtOH for La 3+ and Gd 3+ ion complexes or by the slow diffusion of EtOH vapors into a saturated DMSO solutions of complexes for Lu 3+ and Sc 3+ ion complexes at room temperature. The selected crystal data and refinement parameters for complexes  (9), which are suitable for single-crystal X-ray diffraction, were obtained by the slow evaporation of solutions in EtOH for La 3+ and Gd 3+ ion complexes or by the slow diffusion of EtOH vapors into a saturated DMSO solutions of complexes for Lu 3+ and Sc 3+ ion complexes at room temperature. The selected crystal data and refinement parameters for complexes [Sc(Q CH ) 3 Table 2. The isomorphism of the crystal structures of the studied single crystals [La(Q CH ) 3 Figures  S4 and S6). Table 2. Crystal data and refinement parameters for [Sc(Q CH ) 3  It should also be noted that, according to the PXRD data ( Figure S5), the [Sc(Q CH ) 3

(H 2 O)] and [Lu(Q CH ) 3 (H 2 O)
] complexes also turned out to be isostructural; however, due to the depressingly low solubility levels in most of the weak or moderate coordinating solvents, it was not possible to grow the single crystals of these compounds.    9), having one coordinated DMSO molecule instead of a water molecule. This alteration does not affect the coordination polyhedron {MO7}, which is a capped trigonal prism in all cases, but rather leads to a slight change in the relative arrangement of the ligands around the metal ion ( Figure S7).
As for the [La(Q CH ) 3 (H 2 O)(EtOH)]·(EtOH) (7) complex, a relatively larger ionic radius of La 3+ compared to Sc 3+ , Gd 3+ , and Lu 3+ ions causes the formation of an octa-coordinated complex that includes two solvent molecules in the inner coordination sphere, leading to the {MO 8 } La 3+ polyhedron, which is best described as a square antiprism.
The analysis of Ln 3+ -O bond lengths in the corresponding La 3+ , Gd 3+ , and Lu 3+ ion complexes (Table S2) allows one to observe the influence of lanthanide contraction, which results not only in the shortening of bond lengths, but also in changes in the coordination number from eight for [La(Q CH ) 3

(H 2 O)(EtOH)]·(EtOH) to seven for [Gd(Q CH ) 3 (H 2 O)] and [Lu(Q CH ) 3 (DMSO)
]. It is also worth noting that the hepta-coordinated lanthanide ion is observed in Eu 3+ [29] and Dy 3+ [42] complexes that are isostructural to [Gd(Q CH ) 3 9), having one coordinated DMSO molecule instead of a water molecule. This alteration does not affect the coordination polyhedron {MO7}, which is a capped trigonal prism in all cases, but rather leads to a slight change in the relative arrangement of the ligands around the metal ion ( Figure S7).
As for the [La(Q CH )3(H2O)(EtOH)]·(EtOH) (7) complex, a relatively larger ionic radius of La 3+ compared to Sc 3+ , Gd 3+ , and Lu 3+ ions causes the formation of an octa-coordinated complex that includes two solvent molecules in the inner coordination sphere, leading to the {MO8} La 3+ polyhedron, which is best described as a square antiprism.
The analysis of Ln 3+ -O bond lengths in the corresponding La 3+ , Gd 3+ , and Lu 3+ ion complexes (Table S2) allows one to observe the influence of lanthanide contraction, which results not only in the shortening of bond lengths, but also in changes in the coordination number from eight for [La(Q CH ) 3

Spectroscopic Studies
Optical absorption spectra were obtained for all the complexes and HQ CH (1). As can be seen in Figure 5, the absorption spectra exhibit two pronounced absorption bands located in the UV region of spectra. The bands correspond to ligand absorption and we observed no ion absorption for any of the complexes. The spectra recorded for the complexes [La(Q CH ) 3

(H 2 O)(EtOH)], [Lu(Q CH ) 3 (H 2 O)], and [Gd(Q CH ) 3 (H 2 O)]
, designated further as La, Lu, and Gd, respectively, qualitatively resemble the spectrum of HQ CH . However, the molar extinction (ε) increases from 6 × 10 3 for HQ CH to 3.2-5.5 × 10 4 L × mol −1 × cm −1 for complexes. Moreover,ε increases monotonically from 2.2 × 10 4 to 5.5 × 10 4 L × mol −1 × cm −1 with the replacement of the central ion by one with a higher atomic number. On the contrary, the 13 group metal and Sc 3+ ions affect the optical absorption of the complexes [Sc(Q CH ) 3 3 ], M = Al, In, and Ga (designated as Sc, Al, In, and Ga, respectively), by two factors. Firstly, they multiply ligand extinction up to 10 times. Secondly, these complexes have the red-shift of absorption bands in comparison with the complexes of lanthanide ions and HQ CH . The maximum red-shift of an absorption band is observed for the Sc complex. Notably, the absorption bands are better resolved in the spectra recorded for the Al, Sc, In, and Ga complexes than that in the spectra recorded for HQ CH and lanthanide ion complexes. This is caused by the redistribution of the oscillator strengths corresponding to these bands [15].

(H 2 O)] and [M(Q CH )
ondly, these complexes have the red-shift of absorption bands in comparison with the complexes of lanthanide ions and HQ CH . The maximum red-shift of an absorption band is observed for the Sc complex. Notably, the absorption bands are better resolved in the spectra recorded for the Al, Sc, In, and Ga complexes than that in the spectra recorded for HQ CH and lanthanide ion complexes. This is caused by the redistribution of the oscillator strengths corresponding to these bands [15]. Photoluminescence (PL) spectra, measured under excitations at 340 nm wavelengths, are shown on Figure 6. The spectrum of H Q CH reveals a wide spectral band (FWHM = 92 nm) with the emission maximum at 530 nm and with a wide shoulder at longer wavelengths up to 800 nm. Additionally, the additional fluorescence band appears within the region of 380-420 nm. We observed significant influences of the central ion on the PL spectrum. Actually, the emission maximum (λem) shifts to the blue region of the optical spectrum from 530 nm for the HQ CH ligand to 490 nm for Gd, 464 nm for In and Ga, and 458 nm for Al. Finally, the maximum blue-shift of the PL maximum is for Sc, La, and Lu, measuring 430, 436, and 428 nm, respectively. Surprisingly, the spectrum taken for Gd Photoluminescence (PL) spectra, measured under excitations at 340 nm wavelengths, are shown on Figure 6. The spectrum of H → Q CH reveals a wide spectral band (FWHM = 92 nm) with the emission maximum at 530 nm and with a wide shoulder at longer wavelengths up to 800 nm. Additionally, the additional fluorescence band appears within the region of 380-420 nm. We observed significant influences of the central ion on the PL spectrum. Actually, the emission maximum (λ em ) shifts to the blue region of the optical spectrum from 530 nm for the HQ CH ligand to 490 nm for Gd, 464 nm for In and Ga, and 458 nm for Al. Finally, the maximum blue-shift of the PL maximum is for Sc, La, and Lu, measuring 430, 436, and 428 nm, respectively. Surprisingly, the spectrum taken for Gd exhibits a wide spectral band (FWHM = 104 nm) and the maximum is centered at 490 nm. There is a low intensity spectral band located within 380-420 nm, which matches with the PL maxima of the Sc, La, and Lu complexes. We suppose that the redistribution of the emission intensity is determined by the presence of dual emission: ligand phosphorescence located at longer wavelengths and ligand fluorescence at 425 nm. There is a low intensity spectral band located within 380-420 nm, which matches with the PL maxima of the Sc, La, and Lu complexes. We suppose that the redistribution of the emission intensity is determined by the presence of dual emission: ligand phosphorescence located at longer wavelengths and ligand fluorescence at 425 nm.  To check this hypothesis, the fluorescence and phosphorescence spectra at 77 K were measured (see Figure 7). It is clearly noticeable that the spectral band in the short-wave region of the phosphorescence spectrum disappears, while in the spectrum without delay (fluorescence spectrum), the band is still observed, which unequivocally confirms the fluorescent nature of this spectral band. To check this hypothesis, the fluorescence and phosphorescence spectra at 77 K were measured (see Figure 7). It is clearly noticeable that the spectral band in the short-wave region of the phosphorescence spectrum disappears, while in the spectrum without delay (fluorescence spectrum), the band is still observed, which unequivocally confirms the fluorescent nature of this spectral band. A comparison between the spectra recorded for the complexes of lanthanide ions and those for the complexes of the 13 group metals implies that the valence electrons caused the shift of the emission maximum toward 458 nm for Al and 464 nm for In.
Photoluminescence excitation (PL) spectra were obtained for all the compounds and HQ CH with the registration wavelength located at the emission maxima, respectively. All the spectra are qualitatively similar, revealing a broad excitation band at 320-400 nm. An excitation maximum of 361 nm was estimated for the free ligand (HQ CH ) under emission registration at 530 nm, whereas the spectra for complexes revealed a blue-shifted excitation band with the maxima located in a neighborhood of 340 nm, except for the Lu complex (see Figure 8 and Table 3).  A comparison between the spectra recorded for the complexes of lanthanide ions and those for the complexes of the 13 group metals implies that the valence electrons caused the shift of the emission maximum toward 458 nm for Al and 464 nm for In.
Photoluminescence excitation (PL) spectra were obtained for all the compounds and HQ CH with the registration wavelength located at the emission maxima, respectively. All the spectra are qualitatively similar, revealing a broad excitation band at 320-400 nm. An excitation maximum of 361 nm was estimated for the free ligand (HQ CH ) under emission registration at 530 nm, whereas the spectra for complexes revealed a blue-shifted excitation band with the maxima located in a neighborhood of 340 nm, except for the Lu complex (see Figure 8 and Table 3). To gain insight into the electronic excitation relaxation processes in the investigated compounds, luminescence decays were recorded. All the experiments were conducted at room temperature. As seen from Table 3 (see Figure S10), the formation of trivalent ion complexes increases the observable luminescence lifetime (τ) in comparison with HQ CH . The luminescence decays recorded for the p-metal ion complexes Al, Ga, and In reveal single exponential behavior with characteristic lifetimes of τ = 9.9, 6.3, and 4.2 ns, respectively. Therefore, the replacement of the central Al 3+ ion with the heavier one (Ga 3+ and In 3+ ) leads to a decrease in the observed lifetimes of up to two times. On the contrary, the decays obtained for rare earth ion complexes have more complicated behaviors. The multiexponential law can fit the recorded kinetic traces: where τi and Ai are decay times and amplitudes, respectively. The measured luminescence decay is determined by the following equation: where I irf (t ′ ) is the instrument response function (IRF), which can be described as a double Gaussian shape with the characteristic lifetime of τirf = 0.5 ns. Specifically, the decays for the complexes of the La 3+ , Lu 3+ , and Sc 3+ ions fit with the bi-exponential function. The presence of two relaxation components in the fluorescence decays of the Sc, La, and Lu complexes can be attributed to the distinct emitting sites that are responsible for luminescence [15]. Their characteristic lifetimes are listed in Table 3. Unexpectedly, the luminescence of the Gd compound has a significantly longer decay with characteristic lifetimes of τ1 = 22 μs and τ2 = 36 μs (see Figure S11). Therefore, the relatively long lifetime proves the phosphorescence nature of long wavelength bands in the PL spectrum for Gd. We also measured the luminescence decay for the Gd compound at a cryogenic temperature of 77 K. Cooling leads to the suppression of all the rotationalvibrational processes with consequentially higher values of characteristic lifetimes of τ 77 1 = 220 μs and τ 77 2 = 536 μs.  To gain insight into the electronic excitation relaxation processes in the investigated compounds, luminescence decays were recorded. All the experiments were conducted at room temperature. As seen from Table 3 (see Figure S10), the formation of trivalent ion complexes increases the observable luminescence lifetime (τ) in comparison with HQ CH . The luminescence decays recorded for the p-metal ion complexes Al, Ga, and In reveal single exponential behavior with characteristic lifetimes of τ = 9.9, 6.3, and 4.2 ns, respectively. Therefore, the replacement of the central Al 3+ ion with the heavier one (Ga 3+ and In 3+ ) leads to a decrease in the observed lifetimes of up to two times. On the contrary, the decays obtained for rare earth ion complexes have more complicated behaviors. The multiexponential law can fit the recorded kinetic traces: where τ i and A i are decay times and amplitudes, respectively. The measured luminescence decay is determined by the following equation: (2) where I irf (t ) is the instrument response function (IRF), which can be described as a double Gaussian shape with the characteristic lifetime of τ irf = 0.5 ns. Specifically, the decays for the complexes of the La 3+ , Lu 3+ , and Sc 3+ ions fit with the bi-exponential function. The presence of two relaxation components in the fluorescence decays of the Sc, La, and Lu complexes can be attributed to the distinct emitting sites that are responsible for luminescence [15]. Their characteristic lifetimes are listed in Table 3. Unexpectedly, the luminescence of the Gd compound has a significantly longer decay with characteristic lifetimes of τ 1 = 22 µs and τ 2 = 36 µs (see Figure S11). Therefore, the relatively long lifetime proves the phosphorescence nature of long wavelength bands in the PL spectrum for Gd. We also measured the luminescence decay for the Gd compound at a cryogenic temperature of 77 K. Cooling leads to the suppression of all the rotationalvibrational processes with consequentially higher values of characteristic lifetimes of τ 77 1 = 220 µs and τ 77 2 = 536 µs.

Discussion
The energies of the first excited singlet and triplet states were estimated by generally recognized methods [25,53]. Due to the energy reorganization in absorption and emission processes for the non-adiabatic approximation, energies of the S 0 → S 1 and T 1 → S 0 transitions can be determined as the low-energy edges of the absorption spectrum and highenergy edges of the phosphorescence spectrum with the use of the tangent method [25]. To suppress the rotational-vibrational processes during phosphorescence measurements, the complexes were cooled down to 77 K. To remove the fast fluorescence contribution, a 200 µs delay was employed. The estimated energies of the T 1 state are listed in Table 3. We obtained the close values for all the complexes, which lie in the range of 23,500-23,700 cm −1 , except for Sc. An energy of 22,715 cm −1 was obtained for Sc. Therefore, the central ion leads to a T 1 state energy increase from 22,100 cm −1 for HQ CH to 22,715 cm −1 for Sc and approximately 23,600 cm −1 for all other complexes.
The first excited singlet state energy (S 1 ) increases for all the complexes in comparison with HQ CH , which has an S 1 energy of 26,000 cm −1 . The highest energy was obtained for Lu-28,000 cm −1 ; other complexes' energies lay in the range of 27,000-27,700 cm −1 . Thereby, we did not observe significant changes in the energy gap (∆E ST ) between the S 1 and T 1 states due to the influence of ion substitution (see Table 3).
It should be noted that the ∆E ST values of the La and Gd complexes are sufficiently close to the ligand's values. As ∆E ST values of La and Gd and the ligand are close, and the La and Gd absorption spectra qualitatively resemble the ligand's one, we consider that, specifically, La 3+ and Gd 3+ do not distort the potential energy surfaces of the S 1 and T 1 states. Figure 7 demonstrates the fluorescence and phosphorescence spectra obtained for the Gd complex. The emission spectrum, recorded at 77 K, reveals two emission bands located at 380-420 nm and at 420-650 nm. The band located in the region 380-420 nm vanishes in the phosphorescence spectrum, proving the presence of two radiative relaxation processes with different emission states for the Gd complex. Namely, fluorescence appears within 380-420 nm and phosphorescence is observed on a long-wavelength spectral range. It is interesting that, according to the literature, room temperature phosphorescence is quite rare to see [7,54,55]. Therefore, there are two possible explanations for this phenomenon. First, the intersystem crossing process (ISC) of the Gd complex has a higher rate compared with the other complexes. The second explanation is that it has a much lower rate of non-radiative processes due to the different molecule structures and different symmetry groups in particular.
Two observed lifetimes can be assigned to radiative relaxation from two local minimums on the T 1 state's potential energy surface (PES). Since the long time component's lifetime is much higher under cooling than the short time component's lifetime, we conclude that the longest time component (τ 2 ) is associated with radiative relaxation from the deepest minimum of T 1 PES. For all the investigated complexes, we measured the PL quantum yield values Φ under optical excitations at 340 nm, providing excitations in the maximum of the luminescence excitation spectra (see Figure 4). As follows from Table 3, the formation of complexes leads to strong increases in the PL quantum yield by up to 39 times. In particular, the maximum Φ value was recorded for La. In this compound ligand, the fluorescence Φ was enhanced from 0.5% to 19.5%. In the 13 group metal complexes, the replacement of the central ion with a heavier one led to a decrease in the quantum yield value from 16.9% for Al to 6.6% for Ga and 3.3% for In. We see the same dependence for the La (19.5%) and Lu (6.5%) complexes. The probabilities of radiative (k rad ) and non-radiative (k nrad ) processes were evaluated using the following formulas [53]: As can be seen from Tables 3 and 4, the rate of the radiative process monotonically reduces from 1.7 × 10 7 to 0.7 × 10 7 s −1 , and the quantum yield decreases from 16.9 to 3.0 under the replacement of the central Al 3+ ion on Ga 3+ and In 3+ . Notably, the fluorescence quantum yield value recorded for the Gd complex is only 1.8%, since the radiative relaxation of S 1 is a less pronounced pathway than the intersystem crossing process (ISC) followed by phosphorescence. As we noted a significant increase in the observed luminescence lifetime under cooling up to 77 K, a huge enhancement of the quantum yield was expected. While the number of emitted photons equals the integrated intensity of the emission spectrum, the luminescence quantum yield at a temperature of 77 K (Φ 77 ) can be calculated by the following formula: where I 77 integrated luminescence intensity at 77 K, I 300 integrated luminescence intensity at 300 K, and Φ 300 quantum yield at 300 K. According to this procedure, we achieved the quantum yield value of 45.8% for the Gd complex at a temperature of 77 K. To estimate the energy transfer process from a singlet state to a triplet manifold, the intersystem crossing rates (k isc ) were calculated using emission lifetimes and fluorescence quantum yields, both at 300 K and 77 K. Since only the Al, In, Ga, and Gd complexes and HQ CH demonstrate the single exponential fluorescence behavior, calculations were only performed for these compounds. The intersystem crossing rate can be evaluated as follows [25]: After comparing the ISC rates obtained at 300 K and 77 K, we conclude that the rates for the Al, In, and Ga complexes are lower at 77 K. This is caused by significant vibrational relaxation at room temperature. Taking into account the fact that the probability of nonradiative vibrational relaxation processes is negligibly low at 77 K, we assume that the applied method is more beneficial for calculations at 77 K. The rates calculated at 77 K increase with an increase in the atomic number of the 13 group ions. The k isc rate of the HQ CH rate remained unchanged with the decrease in the temperature (13.4 × 10 7 s −1 and 13.3 × 10 7 s −1 at 300 K and 77 K, respectively), suggesting that phosphorescence is predominant in the relaxation channel (see Figure 6). On the contrary, the Gd complex rate slightly increases from 12.5 × 10 7 s −1 (300 K) to 13.8 × 10 7 s −1 (77 K) due to the paramagnetic properties of the Gd 3+ ion (See Table 4). Notably, since the Gd compound has a rigid geometry, vibrational relaxation is reduced in comparison with HQ CH . Therefore, phosphorescence in the Gd complex is more effective relaxation pathway (see Table 3).
Stock 1 M Sc(ClO 4 ) 3 solution was prepared as follows: after being freshly calcinated at 600 • C, Sc 2 O 3 (6.896 g, 50 mmol, 99.999%) was dissolved upon heating in a quartz flask in mixture of 26 mL of perchloric acid (70%, 99.999% trace metals basis, Aldrich) and 20 mL of deionized water. Excess water was slowly evaporated at 90 • C; the residue was quantitatively transferred to a 100 mL volumetric flask and brought to volume by deionized water. Solution was stored in a polypropylene bottle.
Elemental analysis was performed by Elemental Vario MicroCube CHNO(S) analyzer (Elementar Analysensysteme, Langenselbold, Germany). The metal content was determined by complexometric titration with a Trilon B (disodium salt of ethylenediaminetetraacetic acid) solution in the presence of Xylenol Orange as an indicator (for scandium, lanthanum, and lutetium) or by ICP-MS analysis (for aluminum, gallium, and indium). Before the analysis, the complexes were decomposed by heating with concentrated HNO 3 . ICP-MS was performed using an inductively coupled plasma mass spectrometer ELAN mod. 9000, DRC II, DRC-e (PerkinElmer, Waltham, MA, USA). FTIR spectra were recorded in KBr pellets on Perkin Elmer Spectrum One instrument (PerkinElmer, Waltham, MA, USA).
Single-crystal X-ray diffraction analysis of [Al (Q CH ) 3 ] were due to their weak scattering of X-ray caused by disorders of cyclohexyl substituents. All hydrogen atoms were located in calculated positions and refined within riding model. All calculations were performed using the SHELXTL [56,57] and Olex2 [58] software packages. Atomic coordinates, bond lengths, angles, and thermal parameters have been deposited at the Cambridge Crystallographic Data Centre with deposition numbers-CCDC 2208569-2208571, 2215463, 2215494, 2215723, and 2215486, which are all available, free of charge at www.ccdc.cam.ac.uk (accessed on 3 April 2023).