Tuning of bandgaps and emission properties of light-emitting diode materials through homogeneous alloying in molecular crystals

Abstract

Alloy formation is ubiquitous in inorganic materials science, and it strongly depends on the similarity between the alloyed atoms. Since molecules have widely different shapes, sizes and bonding properties, it is highly challenging to make alloyed molecular crystals. Here we report the generation of homogenous molecular alloys of organic light emitting diode materials that leads to tuning in their bandgaps and fluorescence emission. Tris(8-hydroxyquinolinato)aluminium (Alq₃) and its Ga, In and Cr analogues (Gaq₃, Inq₃, and Crq₃) form homogeneous mixed crystal phases thereby resulting in binary, ternary and even quaternary molecular alloys. The M_xM′_(1−x)q₃ alloy crystals are investigated using X-ray diffraction, energy dispersive X-ray spectroscopy and Raman spectroscopy on single crystal samples, and photoluminescence properties are measured on the exact same single crystal specimens. The different series of alloys exhibit distinct trends in their optical bandgaps compared with their parent crystals. In the Al_xGa_(1−x)q₃ alloys the emission wavelengths lie in between those of the parent crystals, while the Al_xIn_(1−x)q₃ and Ga_xIn_(1−x)q₃ alloys have red shifts. Intriguingly, efficient fluorescence quenching is observed for the M_xCr_(1−x)q₃ alloys (M = Al, Ga) revealing the effect of paramagnetic molecular doping, and corroborating the molecular scale phase homogeneity.

---

Introduction

Alloy formation represents one of the cornerstones in inorganic materials science. Ever since the bronze age, humankind has relied on numerous materials that owe their unique properties to the continuous mixing of similar atoms or ions in the crystalline state thereby forming solids with random disorder. In sharp contrast the formation of alloys between molecular species is highly challenging since molecules are unique in size, shape and properties. Crystal engineering approaches directed towards multi-component crystals are largely based on cocrystal design utilizing supramolecular synthons that non-covalently bind different component molecules in crystals.¹ While cocrystals are limited to the fixed stoichiometries of molecular components, organic alloys are characterized by their continuous stoichiometries in crystals. Since the possibilities of different molecular components occupying similar crystal lattices are limited, the reports of molecular alloy crystals are relatively rare.^2–10 Moreover, studies that explore tuning of physical properties of molecular materials by alloy formation are even rarer – despite the promising prospects of alloys in bandgap engineering of organic semiconductors and molecular photovoltaic crystals.^11,12 Although mixed solid state phases formed by melt-cooling of different molecular components, or by doping in polymer matrices, are sometimes referred to as alloy phases, these composite phases need to be clearly distinguished from molecular alloy crystals, which are strictly characterized by homogenous mixing of molecular components in their crystal lattices. Studies of metal organic frameworks and organic–inorganic hybrid perovskite structures that show composition variation with respect to ionic species in the crystal lattices have been reported with tuning in the optical/electrical properties, but they are not strictly classified as organic alloys.^13,14 A well-known class of organic alloy crystals are those formed by donor–acceptor charge transfer complexes. Matzger et al. and Pei et al. explored the variation in bandgaps of molecular solid solutions, but continuous or systematic trends in bandgaps could not be observed in these alloys.^15,16 Recently, we demonstrated systematic tuning in the optical band gaps in molecular alloys formed by a series of diphenyl dichalcogenides.¹⁷ In the context of photophysical properties of molecular alloys, Fu et al. demonstrated the tuning of photoluminescence emission and optical waveguiding performances of charge transfer alloy crystals formed by perylene and 9,10-dicyano-anthracene (DCA)_1−x(Pe)_x.¹⁸ Although the composition range of the alloy components in these examples were limited to <10%, this report points to the prospects of achieving desirable tuning of photophysical properties if high composition ranges could be achieved in molecular alloys. In this article, we report the crystalline molecular alloys formed by organic light-emitting diode (OLED) materials Mq₃, where M = Al, Ga, In, and Cr and q = 8-hydroxyquinoline ligand. Tris(8-hydroxyquinoline)aluminum or Alq₃ is widely used as the light emitting layer material in OLED devices.^19,20 Light emission is generated in such emissive layers by electron–hole recombination. Hence, it is logical to assume that the tuning of bandgaps (such as the trend reported by us in dichalcogenide molecular alloys¹⁷) may also result in the tuning of emission properties of OLED components such as Alq₃. Tuning of emission properties in Mq₃ compounds are primarily attempted by means of chemical derivatives of the ligand molecule by adding functional groups to 8-hydroxyquinoline ligand.^21,22 However, such methods result in significant changes from the original crystal structure of Mq₃ in packing features, intermolecular interactions and thereby in opto-electronic properties. Recently, Han et al. reported alloy crystal forms of Mq₃ methanol solvates and Mq₃ nano/microrods obtained through a solution pathway.²³ However, the structural characterization of the mixed crystals were limited to the solvate crystal Al_0.5Ga_0.5q₃ MeOH, the crystal structure of which was quite different from that of the pure Alq₃ or Gaq₃. To our knowledge there is no systematic structure–property relation study on the multicomponent molecular alloy forms of Mq₃ series of crystalline OLED materials. Here we report the binary, tertiary and quaternary molecular alloys of Alq₃, Gaq₃, Inq₃, and Crq₃ that form homogenously mixed crystal phases. A detailed structural study correlating the crystal structures, composition and domain structures with the trends in optical band gaps and fluorescence emission properties observed for the alloy crystals is described here employing single crystal X-ray diffraction (SCXRD), powder XRD, energy dispersive X-ray (EDX), Raman and fluorescence spectroscopy. Since Mq₃ systems exhibit multiple crystal polymorphs and two different stereoisomers (see Fig. 1a), the effect of such structural preferences in the alloy formation and emission properties are investigated. In addition, we report an hitherto unknown phenomenon of fluorescence quenching in molecular alloys (in the alloy crystals formed by Crq₃) revealing the effect of paramagnetic molecular doping/impurity on the emission properties of these OLED materials.

Fig. 1 (a) The two possible stereoisomers, namely meridional (mer) and facial (fac), of Mq₃ molecules discussed in this study, where M = Al, Ga, In, and Cr and q = 8-hydroxyquinoline ligand. Asymmetric units of (b) the new polymorph of Gaq₃ and (c) the binary alloy Al_0.48Ga_0.52q₃ shown as thermal ellipsoid plots at the 50% probability level (metal atoms of the alloy structure have been highlighted as enlarged balls).

Results and discussion

Polymorphs and stereoisomers of Mq₃ compounds and the generation of their alloy crystals

In this study, we employed a two-zone sublimation–condensation method to obtain crystals of the Mq₃ compounds. The Mq₃ series of compounds exhibit two stereoisomers; a meridional (mer) isomer with C₁ symmetry and facial (fac) isomer with C₃ symmetry, as shown in Fig. 1a. They are known to crystallize in five different crystal forms – resulting from a combination of polymorphs and stereoisomers – referred to as α, β, γ, δ, and ε. While α, β, and ε forms are adopted by the mer isomer, the γ and δ forms are associated with the fac stereoisomer.^24,25 The central metal atom M is in +3 oxidation state in all the Mq₃ compounds discussed here. The preferences for the stereoisomers are known to be different in the series of Mq₃ compounds, depending upon the size and covalent radii of the central metal atom. The preference for fac or mer isomer is a result of the tradeoff between two opposite factors²⁶ (i) the structure destabilizing steric effect between the ligands which is higher in fac form and increases with decreasing covalent radii, (ii) the stabilizing effect from stronger metal–ligand orbital interactions and bonding, which is predominant in the fac isomer. It is also known that the In–N and In–O bonds are more covalent as compared to other M–O and M–N bonds in this series.²⁰ Hence, for Alq₃, Gaq₃, and Crq₃ the mer isomer is found to be more stable, whereas for Inq₃ the fac isomer is preferred (owing to the higher covalent radius of In). The crystal forms obtained in this study are in accordance with this trend in stability of isomers. For Alq₃, Gaq₃ and Crq₃ the crystal forms obtained by two-zone sublimation–condensation method were determined to be mer-ε form, whereas the Inq₃ crystals obtained were found to be predominantly fac-δ form. The crystals showed acicular morphology (Fig. S1†), and their crystal structures were confirmed by single crystal XRD experiments at 100 K, except for Inq₃ for which the poor quality of the single crystals allowed only the determination of unit cell parameters. Table S2 in the ESI† shows the crystallographic cell parameters for the different crystal forms discussed here. The crystal structures of mer-ε form adopted by Alq₃ and Crq₃ have Z = 6, with 3 molecules in the asymmetric unit (Z′ = 3) and crystallographic space group P1̄. The cell parameters of Alq₃ and Crq₃ showed similarity implying their isostructurality. Although only the β form of Gaq₃ is known (CSD code: ETELAQ), in this study we discovered a new polymorph of Gaq₃Z = 6 and Z′ = 3 in triclinic space group P1̄, which is isostructural to the ε form of Alq₃ and Crq₃. Fig. S2 in the ESI† shows the crystal packing diagrams of mer-Gaq₃ in β and ε forms. The discovery of this mer-ε polymorph points to the propensity of Gaq₃ to potentially occupy the lattice of Alq₃ or Crq₃ crystals.

A commonly reported class of solid solution/alloy crystals are formed between planar aromatic molecules as donor–acceptor charge transfer complexes, where two or more electron-acceptor molecules are sandwiched between planar aromatic electron donor molecules.^15,18,27 Hence the individual crystal structures of the electron-acceptor molecules are not a crucial factor in the alloy formation. On the contrary, for nonplanar molecules to form alloys, isostructurality of the crystal structures of individual components is an important structural criterion.

The crystal structures of the Mq₃ compounds (Alq₃, Gaq₃, Crq₃ and Inq₃) are stabilized by intermolecular C–H⋯π, C–H⋯O interactions and π⋯π stacking interactions (with a partial stacking of the aromatic rings). To obtain a more quantitative understanding of the crystal packing and interactions in these structures, we employed Hirshfeld fingerprint analysis²⁸ which revealed percentage contribution of different types of intermolecular contacts. Hirshfeld surface analysis revealed that the isostructurality in crystal packing and unit cell parameters also manifest as the percentage contributions of C⋯C and H⋯C, and H⋯O interatomic contacts (corresponding to the proportions of C–H… π, C–H⋯O interactions and π⋯π stacking interactions) – which are very similar among the series of Mq₃ crystal structures (Fig. S3 and Table S1†). These results are akin to the previously reported examples of isostructural interaction topologies in analogous crystal structures^4,17 reported by us (represented as energy frameworks²⁹), which facilitated alloy formation. Their unit cell volumes are also very similar, and show the following order: V_cell(Alq₃) < V_cell(Gaq₃) < V_cell(Crq₃) (see ESI Table S2†).

Inspired by the isostructurality of Alq₃, Gaq₃ and Crq₃ (as can be seen from Tables S1 and S2†), we further probed the possibility of alloy formation between these compounds. It should be noted that Han et al. recently reported mixed alloy-like crystals of Mq₃ solvates and crystalline nanorods.²³ The powder diffractograms of the nano/microrods obtained by the mixed phases of Alq₃ and Gaq₃ from chloroform solvent indicated a possible α polymorph formation, although these phases were not structurally characterized. Moreover, the composition of the alloy phases were not determined but arbitrarily assigned to the ratio of parent compounds used for crystallization. Here, firstly we set out to synthesize binary molecular alloys using the following combinations: Alq₃–Gaq₃, Alq₃–Crq₃ and Gaq₃–Crq₃ in different stoichiometric ratios. In order to obtain the single crystals of the binary alloys devoid of any solvent molecules (that might potentially co-crystallize and alter the crystal structure), we employed two-zone sublimation–condensation method. Crystals of Alq₃–Gaq₃, Alq₃–Crq₃ and Gaq₃–Crq₃ alloy phases thus obtained were of acicular morphology, very similar to their parent crystals. SCXRD analysis of the single crystals of the alloys revealed a wide range of compositions (see Tables S3–S7†). We have recently demonstrated that single crystals of molecular alloys obtained from the same crystallization experiment can show a range of composition of their components.¹⁷ Here, we find that accurate stoichiometry control in the alloy crystals is not possible for the systems in this study (using the sublimation–condensation method). Nevertheless, the compositions of the alloy single crystals from SCXRD roughly match with the stoichiometry of the compound mixtures used in their crystallization although with some exceptions (see Tables S3–S7†). The deviation of the composition from their starting stoichiometries were particularly dominant for the alloy crystals containing Crq₃, where the percentage of Crq₃ was generally found to be higher than the stoichiometry used in the crystallization mixture.

As opposed to Alq₃, Gaq₃ and Crq₃, the readily obtained form of Inq₃ is the fac isomer crystallizing in δ form (Z = 2). Although the mer-Inq₃ is reported, it crystallizes in β form (CSD code: WADGUZ).³⁰ Hence, the alloys of Inq₃ with other Mq₃ compounds were found to be quite difficult to obtain possibly due to the fact that Inq₃ prefers a fac stereoisomer as opposed to the other three Mq₃ compounds. We could not obtain alloy single crystals of Inq₃ with Alq₃, Gaq₃ or Crq₃ for higher composition of Inq₃ in the alloy form. Since the major focus of this study is to obtain structure–property relations in the molecular alloys of this series of OLED materials at a molecular to crystal domain level, our investigations were focused mainly on single crystals. Because good quality single crystals could not be obtained for 30% and 50% alloys of Inq₃ with Alq₃ and Gaq₃, we used powder-XRD to characterize these phases (see ESI†). Molecular alloy formation inherently leads to crystallographic disorders with respect to both positions and site occupancy values, and we tested two strategies to model the alloy crystal structures. The structural models with constraints on the positions and atomic displacement parameters (ADPs) of the central metal atoms of Mq₃ were especially found to result in satisfactory convergence in the crystallographic refinements. The occupancy values of metal atoms in the binary alloy crystals were refined using an occupancy free variable (see ESI†).

It is interesting to note some structural features of Mq₃ compounds that exhibit changes in the alloy crystal structures as compared to their parent crystal forms. The metal–ligand bond distances (M–O and M–N bonds) in the series of Mq₃ structures and in the alloys are of particular interest. The average M–O distances for the series of Mq₃ crystal structures from the SCXRD analysis are in the following order: Alq₃ (1.86 Å) < Gaq₃ (1.95 Å) < Crq₃ (1.96 Å). Whereas the M–N bond distances shows the order of Alq₃ (2.03 Å) < Crq₃ (2.04 Å) < Gaq₃ (2.08 Å). Interestingly, the M–O and M–N bond distances in the alloy crystal structures are found to be intermediate between the corresponding bond distances in their parent crystal structures. Fig. 2 shows the trends in M–O and M–N bond distances in Alq₃–Gaq₃ and Alq₃–Crq₃ alloy crystals, with the bond distance variations roughly corelating with the composition of the alloys (also, see Table S9†). This ‘bond averaging effect’ in the alloy crystals is likely due to the spatial averaging of the corresponding metal and ligand atom positions in the alloy crystals, rather than a realistic bond elongation/compression of the M–O and M–N bonds. However, this indicates the ability of different component Mq₃ molecules to occupy very similar crystallographic positions in the alloy crystal lattice in a more or less homogenous manner.

Fig. 2 Variations in M–O and M–N bond distances observed in the crystal structures of (a, b) Alq₃–Gaq₃ and (c, d) Alq₃–Crq₃ alloys with respect to the corresponding bond distances in their parent crystal structures.

Ternary and quaternary alloy crystals of Mq₃

Further, we set out to examine if ternary and quaternary molecular alloys could be obtained from this series of Mq₃ compounds. By extending the same crystallization approach with 1 : 1 : 1 stoichiometric ratio of Alq₃, Gaq₃ and Crq₃ as well as 1 : 1 : 1 : 1 mixture of Alq₃, Crq₃, Gaq₃ and Inq₃, we obtained single crystals which were further analyzed using SCXRD and EDX spectroscopy. For binary molecular alloys, the composition obtained from EDX spectra were found to be comparable to the values obtained from SCXRD. However, for ternary and quaternary alloy crystals, the crystallographic occupancy refinement of the closely positioned metal atoms into three or four linearly correlated free variable failed to attain convergence (leading to unreasonable values of occupancies, including negative occupancies). Hence, to obtain realistic crystal structure models for the three component and four component alloys, we measured EDX spectra on the same single crystals used for SCXRD data collection (Fig. 3a and b). The crystal structures of the ternary and quaternary alloys are thus refined using SCXRD data with values of metal atom occupancies fixed to the composition values obtained from the EDX spectra. To our knowledge, these structures represent the rare examples of quaternary molecular alloys (medium entropy alloys) formed by nonplanar molecules. The examples of six-component solid-solutions with three stoichiometry-tunable molecular components reported by Desiraju et al.,^7,8 and the quaternary solid solutions of donor–acceptor type charge transfer complexes reported by Matzger et al. are to be noted in this context.¹⁵

Fig. 3 Scanning electron microscopic images of (a) ternary and (b) quaternary alloy single crystals along with their EDX spectra, and the corresponding single crystal X-ray structural models (thermal ellipsoid plot at 50% probability level) obtained using the composition values from the EDX spectra. EDX maps of alloy crystals showing the homogenous distribution of their components in (c) Alq₃–Crq₃ binary, (d) Alq₃–Gaq₃–Crq₃ ternary and (e) Alq₃–Gaq₃–Inq₃–Crq₃ quaternary alloys.

Spectroscopic studies of the domain structure and homogeneity of the alloy crystals

Although SCXRD provides the structure of the alloy crystals with multiple components in the lattice, it gives only a spatially averaged picture of the molecules in the alloy crystals. Beyond molecular structure and crystallographic parameters, domain structure is an important feature in alloy crystal phases that influences their physical properties. The long-range distribution of molecular components in a homogenous or heterogenous manner can dictate the degree of mixing of electronic states corresponding to individual components in the alloy crystals. In a recent report, we investigated the molecular alloy crystals formed by diphenyl disulfide and diphenyl diselenide analogues using EDX spectroscopy revealing a homogenous distribution in the alloy phases.¹⁷ In this study, we employed EDX spectroscopy to (i) probe the domain structure of the alloy phases, and (ii) obtain their accurate compositions. We performed high-resolution mapping of the elemental composition on the single crystals of the alloys (Fig. 3c–e) as well as on the microparticles obtained by dispersing the single crystals of the alloys in ethanol (Fig. S19†). Fig. 3c–e show that the alloy crystals show homogenous distribution of molecular components within the resolution of the EDX mapping (2 nm). Since the domain structure from the EDX spectra implies the molecular level mixing of different components in the alloy crystal lattices, such a homogenous distribution can lead to new type of intermolecular interactions – different from those present in the parent crystals. For example, the nature of intermolecular interactions present in the Alq₃–Crq₃ alloy crystal could be different from those in Alq₃ and Crq₃ parent crystals. To probe any intermolecular effects such as charge transfer between the different components in the alloy crystals, we performed Raman spectroscopic studies. Fig. 4a–d shows the Raman spectra measured on the single crystals of pure Alq₃, Gaq₃, Crq₃, Inq₃ and their alloy crystals. In this study, we focused on the spectral range of 80–800 cm⁻¹ which contain the characteristic peaks corresponding to the metal–ligand inter-actions. For pure Alq₃ crystals, we observe a peak at 118 cm⁻¹ corresponding to the Al–oxine deformation and Al–N stretching, and a significant peak is observed at 525 cm⁻¹ corresponding to ring deformation and Al–O stretching. The peak observed at 314 cm⁻¹ is attributed to Al–N stretching and C–C–O bending.^31,32 All these peaks are characteristic features of the mer isomer of Alq₃ and it confirms that the crystals obtained are in mer form. All the observed peak positions and their vibrational assignments are tabulated in Tables S10 and S11 in the ESI.† Gaq₃ and Crq₃ crystals also showed similar peaks with slight differences in the peak positions with the change in the central metal atom. However, the spectra obtained for pure Inq₃ crystals are different from those of the Alq₃, Gaq₃ and Crq₃. A comparison of the Raman spectra obtained for Inq₃ with those reported in the literature for fac-Alq₃ indicate that the crystals could be predominantly in fac form³¹ as it was also indicated by SCXRD measurements (since the weakly diffracting single crystals of Inq₃ were too small to obtain complete diffraction data for the structural refinement, only the unit cell parameters were obtained from the SCXRD measurements). Raman spectra of the binary alloy crystals formed by Alq₃, Gaq₃ and Crq₃ were found to be very similar to their parent crystals with no significant shifts in their characteristic peak positions.^31,32 This implies the absence of any considerable charge transfer type interactions between the molecular components in the alloys. An observable effect in the Raman spectra of the alloy crystals is a relative broadening of these characteristic peaks compared to parent crystals. The fact that the characteristic peaks corresponding to the M–O and M–N stretching modes do not show any significant change in the alloy crystals (compared to the parent crystals) further substantiates our argument that the bond distance averaging in the alloy structures is the consequence of the crystallographic spatial averaging rather that a bond relaxation effect. On the other hand, alloys formed by Inq₃ show a different trend from that of the pure Inq₃ crystals. As discussed earlier, the pure Alq₃ Raman spectrum corresponds to the mer isomeric form while that of Inq₃ exhibit spectral features corresponding to fac isomer as it is the preferred isomer for Inq₃. However, the Raman spectra of the alloy phases of Inq₃ reveal an intriguing trend: when the composition of Alq₃ is higher in Alq₃–Inq₃ alloy crystal, the spectra match well with that of the mer isomeric form. On the other hand, as the concentration of Alq₃ decreases, the characteristic peaks of the mer isomer disappears and peaks corresponding to fac isomer emerge in the spectra (Fig. 4c). In crystals of Alq₃, the peak at 525 cm⁻¹ corresponds to mer isomer and the peaks at 535 cm⁻¹ and 548 cm⁻¹ correspond to fac isomer. These peaks at 535 cm⁻¹ and 548 cm⁻¹ in Alq₃–Inq₃ alloy crystals with higher composition of Inq₃, such as Alq₃–Inq₃ (50 : 50), and Alq₃–Inq₃ (30 : 70) are replaced by a new peak at 525 cm⁻¹ for alloy crystals with low composition of Inq₃ (as in 90 : 10 alloy of Alq₃–Inq₃). This implies that the relatively less stable mer-isomer of Inq₃ could be getting entrapped in the Alq₃–Inq₃ alloy crystal lattice, where it is a minor component.

Fig. 4 (a–d) Raman spectra of Alq₃, Gaq₃, Crq₃, Inq₃ and their binary alloys measured on single crystals. (e) Structural overlay of fac-Inq₃ and mer-Inq₃ along with the crystal packing diagrams for fac-Inq₃ and the supramolecular ‘entrapment’ of mer-Inq₃ in the lattice of mer-Alq₃ leading to Alq₃-Inq₃ alloy crystals.

Optical bandgaps of the alloys from UV-Vis diffuse reflectance spectra

A major consequence of the mixing of different molecules that are chemical analogues in the alloy phases could be the possible mixing of their electronic states. This could manifest as changes in the energy band gaps of the alloy crystals. In a recent report, we demonstrated that alloy formation can indeed lead to tuning in optical band gaps of crystals (even in the absence of π–π stacking) in the crystal structure.¹⁷ The bandgap is a fundamental feature of a material that is linked to many of its optical and electronic properties. Especially since the Mq₃ series of compounds are OLED emissive layer materials, the trends in variation of band gaps in the alloy crystals is of particular interest. Hence, we measured UV-visible diffuse reflectance spectra on the solid-state samples of pure Mq₃ crystals and their binary alloys. The diffuse reflectance spectra were transformed using the Kubelka–Munk function³³ and the optical bandgaps were estimated from Tauc's plots.³⁴ The bandgaps observed for the series of pure crystals are in the following order Crq₃ < Gaq₃ < Alq₃ < Inq₃. Fig. S12† shows that Tauc's plot of Crq₃ shows a distinct feature of two separate slopes. It should be noted that the Tauc plot is usually effective when there is nearly zero absorption below the band gap energy, however in case of Crq₃ and its alloy crystals, there is a fair amount of absorption below the band gap energy. These absorption features usually arise from impurity or modified structure, or surface modifications. Hence a modified Tauc plot is used in these cases.³⁵ See ESI Fig. S10–S14† for Tauc's plots of all the Mq₃ crystals and their alloys. The values of the optical bandgaps for the parent crystals and their alloys are given in Table 1. Notably for Alq₃–Gaq₃ alloys the bandgap was found to be intermediate to those of pure Alq₃ and Gaq₃ similar to the trend we reported recently for diphenyl dichalcogenide alloys.¹⁷ In case of Alq₃–Inq₃ alloys the tuning of bandgap to intermediate values (of Alq₃ and Inq₃) is observed for lower composition of Alq₃. Higher concentration of Alq₃ (lower concentration of Inq₃) in the alloy phase leads to a significant lowering of the band gap compared to the values of both pure Inq₃ and Alq₃. A similar trend is also observed for Gaq₃–Inq₃ alloy phases for which bandgap tuning (at high compositions of Inq₃) and bandgap lowering (at low compositions of Inq₃) were observed.

Table 1 Optical bandgaps (eV) for the crystalline Mq₃ samples and their binary alloys obtained from UV-vis diffuse reflectance spectra

Alloy	Composition ratios
	100 : 0	90 : 10	50 : 50	30 : 70	0 : 100
Alq3–Gaq3	2.65	2.63	2.62	2.60	2.59
Alq3–Inq3	2.65	2.61	2.67	2.66	2.71
Gaq3–Inq3	2.59	2.55	2.6		2.71
Alq3–Crq3	2.65	2.54	2.53		2.53
Gaq3–Crq3	2.59	2.53	2.53		2.53

---

A conjoined analysis of the features from Raman spectra alongside SCXRD structures suggest that the possible origin of this distinct trend in bandgaps is in the preferential stabilization of different stereoisomers of Inq₃ with different compositions in its alloys. We find that the supramolecular entrapment of mer-Inq₃ (an otherwise unfavored isomer of Inq₃) as a minor component in the lattice of a mer-Alq₃ or mer-Gaq₃ major component could be the structural origin of this trend. This leads to intriguing effects in the emission properties of such alloy crystals as evident from the photoluminescence emission maxima described in the following section. Interestingly the bandgaps of the Alq₃–Crq₃ and Gaq₃–Crq₃ alloy crystals were found to be very close to the value of pure Crq₃ crystals (2.53 eV). Such a lowering of bandgap close to that of the paramagnetic component in molecular alloy crystals is hitherto unknown.

Effect of alloy formation on the photoluminescence emission

To probe the effect of alloy formation in the emission properties of the alloy crystals, we examined the photoluminescence (PL) on the single crystals of the pure Mq₃ and their binary alloys. The crystals mounted on a quartz plate were excited at a wavelength of 370 nm. The PL properties of pure Mq₃ systems are well studied and their emission is understood to originate from π–π* electronic transition in the ligand (from the HOMO localized around the phenoxide ring to the LUMO localized around the pyridyl ring).³⁶ Their emission maxima are known to vary with the central metal atom, polymorphic forms, and stereoisomers of the Mq₃. It was previously reported that with the increase in the size of the metal atom, there could be an increase in the covalent nature of the metal ligand bond leading to the red shift in the emission peak value.^20,37 Moreover, the mer isomeric forms showed a red shifted emission compared to the fac isomer in all the Mq₃ systems.³⁸ In addition to the stereochemical differences in the mer and fac isomers, variations in their crystal packing and intermolecular interactions might also be resulting in different relative energies for the ground and excited states – leading to shifts in their corresponding emission wavelengths. We observed the following trend in the wavelengths of emission maxima (λ_max) for the series of crystals studied here (Fig. 5): Gaq₃ (526 nm) > Alq₃ (504 nm) > Inq₃ (495 nm). Notably in the case of Crq₃ no photoluminescence emission was observed. This is evident from the optical microscopic images of Crq₃ single crystals (Fig. 5a) as well from the fluorescence spectra (Fig. 5d). Although the PL emission maximum of Inq₃ (at 498 nm) seems to defy the trend with respect to metal atom size, it is to be noted that the crystal form of Inq₃ obtained in this study are found to be fac form (the more stable stereoisomer for Inq₃). The emission maximum we obtained for Inq₃ matches with the previously reported PL spectra of fac-Inq₃ in powder form.³⁸ Further, we examined the emission features in the PL spectra of the binary alloy crystals. In the case of Alq₃–Gaq₃ alloy crystal, we observed a tuning of the emission wavelength between the parent compounds i.e., the emission maxima of the alloy crystals fall between the emission maxima of pure Alq₃ and Gaq₃ (Fig. 5b). With increasing concentration of Gaq₃ in the alloy, the emission showed a red shift. However, in the case of Alq₃–Inq₃ alloy crystals, the trend was quite different. In Alq₃–Inq₃ alloy crystals with higher composition of Alq₃, the emission peak exhibits a red shift compared to both the parent compounds, however, for lower compositions of Alq₃, the peak positions are close to the parent compounds (Fig. 5c and g, Table 2). The significant red shifts in the PL emission maxima of Alq₃–Inq₃ alloys red-shifted from those of the parent crystal forms may be understood in terms of the preferred stereoisomer and the lattice structure of Inq₃ molecules in their alloy phases. Possible explanations for the observed red shifts can be (i) charge transfer interaction between Inq₃–Alq₃ molecules in the crystal lattice leading to changes in the electronic energy levels, or (ii) supramolecular entrapment of the relatively less stable mer isomer of Inq₃ in the crystal lattice of the alloys. The latter is consistent with our structural and spectroscopic observations on the alloys. The crystal packing features of the major component lattice of the mer-ε Alq₃ prefers the incorporation of mer-Inq₃ compared to fac-Inq₃. Hence the possible origin of the red shifts observed in the PL emission should be the occurrence of the otherwise less stable mer-Inq₃ ‘trapped’ in the alloys, which is supported by SCXRD and Raman spectral analysis as well as by the trend observed in the optical bandgaps.

Fig. 5 (a) Optical microscopic images of the single crystals of the parent compounds (top row) and some binary alloys (bottom row), under UV light (λ = 395 nm). (b–e) Photo-luminescence spectra of Alq₃, Gaq₃, Inq₃ and Crq₃ in comparison with those of the binary alloys (background corrected). The absence of PL in Crq₃ and quenching of PL in Alq₃–Crq₃ alloy are evident from (d). The samples are excited at λ = 370 nm. (f) The tuning in λ_max of the PL emission in Alq₃–Gaq₃ alloy crystals, and (g) the red shifts in the λ_max in Alq₃–Inq₃ alloy crystals away from the λ_max values of the parent crystals.

Table 2 Photoluminescence emission maxima (λ_max in nm) of the binary alloy crystals

Alq3–Gaq3		Alq3–Inq3		Gaq3–Inq3
Composition	Em. λmax	Composition	Em. λmax	Composition	Em. λmax
Gaq3 pure	526	Inq3 pure	495	Gaq3 pure	526
Al0.59Ga0.41	523	Al0.27In0.73	496	Ga0.88In0.12	527
Al0.78Ga0.22	517	Al0.49In0.51	502	Ga0.51In0.49	515
Al0.89Ga0.11	515	Al0.86In0.14	524
Alq3 pure	504	Alq3 pure	504	Inq3 pure	495

---

Quenching of photoluminescence emission in the alloy crystals by the paramagnetic Crq₃ molecules

As mentioned above, the crystals of Crq₃ exhibit no significant PL emission (Fig. 5). However, we observed an intriguing phenomenon of fluorescence quenching caused by Crq₃ molecules in its alloy crystals. The binary alloys of Crq₃ with Alq₃ and Gaq₃ shows an effective quenching of PL emission. A similar phenomenon of quenching is also found for the ternary alloy crystals of Al_xGa_yCr_1−x−yq₃ and the quaternary alloy crystals of Al_xGa_yIn_zCr_{1−x−y−z}q₃. Crq₃ is paramagnetic, with the central Cr³⁺ atom having an electronic configuration of 3d³ (t³_2ge⁰_g). It is known that paramagnetic metal centers can interact with the excited states of fluorescing molecules leading to a quenching in fluorescence.^39,40 Fluorescence quenching by paramagnetic metal ions, such as Cr³⁺, Fe³⁺, Co²⁺, Ni²⁺, and Cu²⁺ have been reported,^41–43 and this phenomenon has been utilized in the detection of nitroxyl (HNO) in biological systems.⁴⁰ On the other hand, fluorescence enhancement induced by Cr³⁺ ions upon complexation with the ligand molecules is also reported.⁴³ It should be noted that most of this reported examples of fluorescence quenching or enhancement are associated with the complexation of paramagnetic metal ions with fluorophore molecules in solution state. The mechanism of quenching by paramagnetic metal centers are generally understood to be based on the photoinduced electron transfer (PET) from a singlet excited state of the fluorophore to paramagnetic metal center upon metal–ligand complexation. However, in the present study the fluorescence quenching is found to be originating from sheer intermolecular interactions between the Crq₃ molecules and the fluorescing molecule such as Alq₃ or Gaq₃ in the alloy lattice. Moreover, we find that the presence of Crq₃ leads to highly efficient fluorescence quenching in its molecular alloys with Alq₃ and Gaq₃ – not only in 1 : 1 alloys, but also in alloys with compositions as low as ∼5% of Crq₃. These observations indicate a highly homogenous mixing of the molecular components in the alloy crystals, allowing effective interaction of the paramagnetic Crq₃ molecules with the electronic states of the Alq₃ or Gaq₃ molecules in the neighborhood of the alloy lattices. These results call for in-depth experimental and theoretical investigations, as the exact origin of the quenching phenomenon caused by the intermolecular interactions in the Crq₃ alloy crystals are not clear.

Conclusion

In summary, the molecular alloys formed by the Mq₃ series of compounds reported in this study exhibit intriguing trends with respect to their spectral and emission properties. The detailed analysis combining SCXRD, EDX and Raman spectra reveal correlations with the observed trends in their structure, optical band gaps and photoluminescence emission properties. In addition, we have reported the rare examples of ternary and quaternary molecular alloys formed by nonplanar molecules. The highly efficient fluorescence quenching caused by Crq₃ in its molecular alloy crystals reveals the effect of paramagnetic molecular components in the photoluminescence of alloy phases and provides further evidence for the highly homogenous nature of these alloys. The trends in PL emission peaks observed for different series of binary alloys have been corelated with the trends in their optical band gaps as well as Raman spectroscopic features and the stereoisomeric preference for mer or fac forms. We have shown that the formation of molecular alloys of Inq₃ with Alq₃ leads to a supramolecular entrapment of the less favored mer stereoisomer of Inq₃ in the alloy crystal lattice, leading to red shifts in the emission properties. Alloy crystal engineering using molecular components of chemical analogues as demonstrated in this study could be an efficient design strategy to fine tune the photophysical properties of functional molecular materials.

Data availability

The crystallographic information files have been deposited in the CCDC database with CCDC deposition numbers 2073957–2073962, 2074033–2074036, 2074042–2074044, 2074050–2074051, 2074080–2074081.

Author contribution

RT and SPT performed conceptualisation, data curation, formal analysis, investigation, methodology, and writing of the original draft. HL, AHM, MB and VB performed data curation, formal analysis and methodology. BBI performed conceptualisation, funding acquisition, project administration, supervision, discussion, review and editing.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was supported by the Villum Foundation grant to BBI. VB acknowledges Novo Nordisk foundation (grant number NNF20OC0061417). SPT acknowledges EU funding for Marie Skłodowska-Curie Individual fellowship (grant number 798633). We thank Prof. Brian Julsgaard for useful discussions.

References

1. G. R. Desiraju, J. Am. Chem. Soc., 2013, 135, 9952–9967.
2. D. S. Reddy, D. C. Craig and G. R. Desiraju, Chem. Commun., 1994, 1457–1458,  10.1039/C39940001457.
3. M. K. Mishra, U. Ramamurty and G. R. Desiraju, J. Am. Chem. Soc., 2015, 137, 1794–1797.
4. S. P. Thomas, R. Sathishkumar and T. N. Guru Row, Chem. Commun., 2015, 51, 14255–14258.
5. M. Lusi, CrystEngComm, 2018, 20, 7042–7052.
6. S. Chakraborty, S. Joseph and G. R. Desiraju, Angew. Chem., Int. Ed., 2018, 57, 9279–9283.
7. M. Paul, S. Chakraborty and G. R. Desiraju, J. Am. Chem. Soc., 2018, 140, 2309–2315.
8. S. Chakraborty and G. R. Desiraju, Cryst. Growth Des., 2018, 18, 3607–3615.
9. M. Dabros, P. R. Emery and V. R. Thalladi, Angew. Chem., Int. Ed., 2007, 46, 4132–4135.
10. M. Lusi, Cryst. Growth Des., 2018, 18, 3704–3712.
11. X. Xu, B. Shan, S. Kalytchuk, M. Xie, S. Yang, D. Liu, S. V. Kershaw and Q. Miao, Chem. Commun., 2014, 50, 12828–12831.
12. M. Schwarze, W. Tress, B. Beyer, F. Gao, R. Scholz, C. Poelking, K. Ortstein, A. A. Günther, D. Kasemann, D. Andrienko and K. Leo, Science, 2016, 352, 1446.
13. C. J. Adams, M. F. Haddow, M. Lusi and A. G. Orpen, Proc. Natl. Acad. Sci. U.S.A., 2010, 107, 16033.
14. W.-Q. Liao, D. Zhao, Y.-Y. Tang, Y. Zhang, P.-F. Li, P.-P. Shi, X.-G. Chen, Y.-M. You and R.-G. Xiong, Science, 2019, 363, 1206.
15. R. A. Wiscons, V. Coropceanu and A. J. Matzger, Chem. Mater., 2019, 31, 6598–6604.
16. J.-H. Dou, Z.-A. Yu, J. Zhang, Y.-Q. Zheng, Z.-F. Yao, Z. Tu, X. Wang, S. Huang, C. Liu, J. Sun, Y. Yi, X. Cao, Y. Gao, J.-Y. Wang and J. Pei, J. Am. Chem. Soc., 2019, 141, 6561–6568.
17. S. P. Thomas, R. Thomas, T. B. E. Gronbech, M. Bondesgaard, A. H. Mamakhel, V. Birkedal and B. B. Iversen, J. Phys. Chem. Lett., 2021, 3059–3065,  DOI:10.1021/acs.jpclett.1c00614.
18. X. Han, Y. Lei, Q. Liao and H. Fu, Angew. Chem., Int. Ed. Engl., 2021, 60, 3037–3046.
19. C. W. Tang and S. A. VanSlyke, Appl. Phys. Lett., 1987, 51, 913–915.
20. C. H. Chen and J. Shi, Coord. Chem. Rev., 1998, 171, 161–174.
21. V. A. Montes, R. Pohl, J. Shinar and P. Anzenbacher Jr, Chem.–Eur J., 2006, 12, 4523–4535.
22. J. P. Heiskanen and O. E. O. Hormi, Tetrahedron, 2009, 65, 8244–8249.
23. C. Dai, Z. Wei, Z. Chen, X. Yan, J. Fan, Y. Yang, Z. Pang and S. Han, ACS Appl. Energy Mater., 2018, 1, 4367–4373.
24. M. Rajeswaran, T. N. Blanton, C. W. Tang, W. C. Lenhart, S. C. Switalski, D. J. Giesen, B. J. Antalek, T. D. Pawlik, D. Y. Kondakov, N. Zumbulyadis and R. H. Young, Polyhedron, 2009, 28, 835–843.
25. R. Katakura and Y. Koide, Inorg. Chem., 2006, 45, 5730–5732.
26. C. F. R. A. C. Lima, R. J. S. Taveira, J. C. S. Costa, A. M. Fernandes, A. Melo, A. M. S. Silva and L. M. N. B. F. Santos, Phys. Chem. Chem. Phys., 2016, 18, 16555–16565.
27. Y. Lei, Y. Sun, Y. Zhang, H. Zhang, H. Zhang, Z. Meng, W.-Y. Wong, J. Yao and H. Fu, Nat. Commun., 2018, 9, 4358.
28. M. A. Spackman and D. Jayatilaka, CrystEngComm, 2009, 11, 19–32.
29. M. J. Turner, S. P. Thomas, M. W. Shi, D. Jayatilaka and M. A. Spackman, Chem. Commun., 2015, 51, 3735–3738.
30. M. Brinkmann, B. Fite, S. Pratontep and C. Chaumont, Chem. Mater., 2004, 16, 4627–4633.
31. S.-Y. Hung, R.-L. Kao, K.-Y. Lin, C.-C. Yang, K.-S. Lin, Y.-C. Chao, J.-S. Wang, J.-L. Shen and K.-C. Chiu, Mater. Chem. Phys., 2015, 154, 100–106.
32. M. Halls and R. Aroca, Can. J. Chem., 1998, 76, 1730–1736.
33. J. H. Nobbs, Rev. Prog. Coloration Relat. Top., 1985, 15, 66–75.
34. J. Tauc, Mater. Res. Bull., 1968, 3, 37–46.
35. P. Makuła, M. Pacia and W. Macyk, J. Phys. Chem. Lett., 2018, 9, 6814–6817.
36. J. G. Mahakhode, S. J. Dhoble, C. P. Joshi and S. V. Moharil, Bull. Mater. Sci., 2011, 34, 1649–1651.
37. B. J. Chen, X. W. Sun and Y. K. Li, Appl. Phys. Lett., 2003, 82, 3017–3019.
38. R. I. Avetisov, A. A. Akkuzina, A. G. Cherednichenko, A. V. Khomyakov and I. C. Avetissov, Dokl. Chem., 2014, 454, 6–8.
39. A. R. Freitas, M. Silva, M. L. Ramos, L. L. G. Justino, S. M. Fonseca, M. M. Barsan, C. M. A. Brett, M. R. Silva and H. D. Burrows, Dalton Trans., 2015, 44, 11491–11503.
40. W. Yang, X. Chen, H. Su, W. Fang and Y. Zhang, Chem. Commun., 2015, 51, 9616–9619.
41. A. W. Varnes, R. B. Dodson and E. L. Wehry, J. Am. Chem. Soc., 1972, 94, 946–950.
42. H. S. Jung, P. S. Kwon, J. W. Lee, J. I. Kim, C. S. Hong, J. W. Kim, S. Yan, J. Y. Lee, J. H. Lee, T. Joo and J. S. Kim, J. Am. Chem. Soc., 2009, 131, 2008–2012.
43. S. Pal, N. Chatterjee and P. K. Bharadwaj, RSC Adv., 2014, 4, 26585–26620.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: 10.1039/d1sc03714e

‡ These authors contributed equally.

§ Current address: Department of Chemistry, Indian Institute of Technology Delhi, New Delhi-110016, India.

---