On the nature of the red shift in absorption spectra of β-dicarbonyl chelates

Abstract

The present work offers a theoretical interpretation of the red shift observed in the absorption spectra of mononuclear dicarbonyl chelates. The energy of the transition ultimately depends on two factors: the redistribution of π-electronic density from a homogeneous distribution around the dicarbonyl fragment toward the multiple bonds of the substituted ring, and the decrease in the interatomic distance between the oxygen atoms of the dicarbonyl fragment. The O–O′ distance is dependent on the coupling between the 2pσ_z orbitals of oxygen atoms and the d_z² and d_xz orbitals of the metal ion.

---

In the almost two decades since Sony used 8-quinolinolate complexes as luminophores in the world's first production OLED TV,¹ similar chelate structures have formed the basis of nearly all economically successful materials for OLED matrix displays. The optical π* ← π and π* ← n transitions at the used wavelengths correspond to metal-to-ligand charge transfer.² The oscillator strength and quantum yield of such transitions are largely determined by the structure of the delocalized molecular orbitals of the donor ligand. Conjugated β-dicarbonyl chelates are coordination compounds with significant π-delocalization within a six-membered chelation ring, closely resembling the electronic structure of hydroxyquinolinates. The most stable complexes are formed by perfluorinated β-diketones bearing a heterocyclic moiety. Many of these diketonate complexes act as green and red luminophores^3,4 and are promising candidates for application as OLED materials.^5,6

Some aspects of how the ligand structure influences luminescence parameters remain unexplained. Most solid-state diketonates for which single-crystal structures are known are mixed-ligand complexes. In such complicated systems, the effect of an individual ligand is overlapped by the electronic contributions of other ligands and by the polynuclear nature of the central ions. The study of mononuclear diketonates may therefore shed light on the underlying mechanism.

Although the synthesis of solid-state isostructural mononuclear complexes is a difficult task, the thermodynamic and spectral parameters of such compounds can be effectively estimated by investigating equilibrium complexation using electronic absorption spectroscopy in aqueous solutions (Fig. 1a). Recently, we have compiled a large dataset on equilibrium systems of alkaline-earth metals, rare-earth metals, transition metals, actinides, and heterocyclic β-diketones.^7–13 During the analysis of the obtained dataset, a correlation was identified between the maximum absorption wavelength of mononuclear complexes and the coordination nature of the metal ions.¹⁴ Lanthanide complexes exhibit identical maximum wavelengths, which depend only on the nature of the substituted ring and not on the type of lanthanide ion⁷ (Fig. 1b). Complexes of transition-metal ions (Zn²⁺, Mn²⁺, Ni²⁺, and others) exhibit absorption wavelengths that fall between those of the anionic form of the ligand and the lanthanide complexes.¹² The bands of the anionic ligand forms are shifted toward the blue region and serve as reference points. The absorption wavelengths of alkaline-earth complexes (Mg²⁺ and Ca²⁺) are slightly shifted relative to the anionic form and fall between the anionic and transition-metal complexes. The most red-shifted complexes are the actinide diketonates.⁸ Depending on the ligand, these complexes exhibit red shifts of 11–15 nm relative to rare-earth complexes and 17–23 nm relative to the anionic form. Thus, the observed order of absorption wavelengths is identical for each diketone: anionic form < alkaline-earth metals < divalent d-metals (Zn²⁺, Co²⁺, Mn²⁺, etc.) < lanthanide ions < Sc³⁺ < Th⁴⁺. The exact absorption wavelength values are provided in Table S1 (SI).

Fig. 1 Spectra and structures. (a) Scheme of chelation studied β-diketones and the structures of their substituted rings, along with the corresponding ligand abbreviations. The structure of the ligand is presented in keto-form for the sake of clarity; (b) effect of red shift in spectra of diketone complexes: the comparison of the electronic absorption spectra of anionic forms of ligands and their mononuclear ML complexes; (c) frontier molecular orbitals (HOMOs – left and LUMOs – right) corresponding to electronic transitions in the ultraviolet spectra: example for anionic form S-CF₃ and for [Gd(H₂O)₆(S-CF₃)]²⁺ species. The structure of the LUMOs is identical for all forms but the HOMOs significantly differ: the chelation leads to the localization of π-electronic density near multiple bonds while in anionic form the electronic density is delocalized around the dicarbonyl fragment. cc-pVDZ/PBE0/SMD level of theory.

The similarities in the shapes of the absorption spectra have a simple explanation: the most intense band for both the anionic form of diketones and their chelated forms corresponds to intraligand charge transfer between the molecular orbitals of the deprotonated dicarbonyl fragment and the π-molecular orbitals of the substituted ring (Fig. 1c and Fig. S1, S2 in the SI). However, the origin of the red shift in the absorption bands remains unclear. Initially, we suggested that the red shift was related to the hydration shell.¹⁰ This is partially correct, as the number of water molecules in the first hydration shell has a direct impact on the main absorption band.¹⁵ Nevertheless, this alone does not account for the differences in transition energies. The present work aims to provide a detailed explanation of the observed spectral parameters.

First, we need to explain why the spectra of the complexes are always red-shifted relative to the spectra of the free ligands. The observed spectral bands are associated with the frontier molecular orbitals. Fig. 1c shows the typical shapes of the HOMO and LUMO in the anionic form of diketones and their complexes (example: Gd³⁺). The structures and compositions of the LUMOs are similar for both species, whereas the HOMO orbitals differ. In the free ligand, the π-electrons are more evenly delocalized around the dicarbonyl fragment. In the chelated form, the bond orders change and the π-electron density is localized near the double C=C bond of the dicarbonyl fragment and depleted around the second C=O group (Scheme 1). Analysis of atomic Mulliken population for the dicarbonyl fragment in different forms is given in Table 1. We can see that, across the series of complexes, the atomic populations are reduced and directly correlate with the experimental wavelengths. The populations are identical for the lanthanide complexes, which corresponds to the experimentally observed single wavelength for these chelates. Fig. 2 shows that the reduction in the population of the dicarbonyl fragment leads to an increase in the population of the heterocyclic ring.

Scheme 1 Distribution of π-electron density in the HOMO of the anion (left) and chelation form (right).

Table 1 An atomic Mulliken population in the HOMO and experimental parameters (S-CF₃ system, PBE0/cc-pVDZ)

Form	C-β	C-α	C-γ	O-α	O-β	λ, nm^a	log K^b
a λ-wavelength at maximum absorbance. b Stability constant, defined in the SI.
Anion	0.99	0.08	0.04	0.45	0.21	340	—
MgL^+	0.69	0.13	0.03	0.34	0.04	341	2.04 ± 0.06
ZnL^+	0.64	0.13	0.03	0.32	0.03	344	3.52 ± 0.06
LaL^2+	0.58	0.13	0.03	0.29	0.01	348	3.15 ± 0.06
GdL^2+	0.57	0.12	0.03	0.28	0.01	348	3.38 ± 0.06
ScL^2+	0.47	0.13	0.02	0.24	0.00	353	5.4 ± 0.1
ThL^3+	0.38	0.13	0.01	0.18	0.00	357	6.25 ± 0.12

---

Fig. 2 Dynamics of atomic Mulliken population in the HOMOs (in the β-dicarbonyl fragment (a) and in the substituted ring (c)) for different complexes. Example for the S-CF₃ (top) and Phen-CF₃ (bottom) system; (b) schematic image of the redistribution of π-electronic density from the dicarbonyl fragment to multiple bonds of the substituted ring.

Thus, the intercalation of the metal ion into the diketone structure leads to a redistribution of the π-electronic density in the HOMO from the β-dicarbonyl group to the multiple bonds of the heterocyclic moiety. This reduces delocalization and leads to an increase in orbital energies and a decrease in the energy of the corresponding electronic transition.

To understand the spectral red shift in the complexes, we must explain why the HOMO–LUMO energy gap decreases from light to heavy metal complexes. Although the energy of the electronic transition directly depends on the wave functions of the ground and excited states, the important factor is the dipole moment.¹⁶ In the series of dicarbonyl chelates, the origin of these changes should be sought in the structure of the σ-bonds of the dicarbonyl fragment. Mononuclear dicarbonyl complexes contain a six-membered chelation ring, and the geometry of this ring is related to the radius of the metal ion. The metal ion forms two σ-bonds with sp-hybrid orbitals and two π-bonds with the 2pπ_y orbitals of the oxygen atoms. Additional σ- and π-bonds are formed between the metal ion and water molecules of the hydration shell. Fig. 3a illustrates an approximate MO diagram of the dicarbonyl chelation ring, using Sc³⁺ as an example.^17,18 Direct σ- and π-bonds between metal ions and oxygen atoms slightly change the C=O distances and do not cause deformations of the dicarbonyl fragment. The structure of the –C–C(β)–C– chain also does not change across different complexes and ligands. Only one structural parameter changes significantly—the distance between the chelating O (β-O) and O′ (α-O) oxygen atoms (Fig. 3c).

Fig. 3 Structures of molecular orbitals and complexes. (a) The scheme of MO for the dicarbonyl fragment for the [Sc(H₂O)₆L]²⁺ complex; the orbitals that are formed as a result of overlap between the 2pσ_z orbitals and d_z² and d_xz orbitals are shown in dark blue (and their suggested symmetries). (b) Schematic (left) and calculated (right) structure of the overlap between the 2pσ_z orbitals of the oxygen atoms in the free form of the ligand and the coupling of 2pσ_z orbitals with d_z² and d_xz orbitals of the metal ion. (c) The correlation between the O–O′ distance in complex species and the observed maximum wavelength (example for O-CF₃ complexes; 1 Å = 100 pm); (d) TDDFT maximum wavelength for chalcogen ligand geometry with different O–O′ length. The spectral range at the variation of O–O′ distance is shifted in the red part of the spectrum, as it is observed in experiments for the corresponding complex species: O-CF₃ < S-CF₃ < Se-CF₃ < Te-CF₃. (c) Correlation between the calculated O–O′ distance in complex species and observed wavelength.

Estimation of the O–O′ distance by DFT (cc-pVDZ/PBE0/SMD level of theory) demonstrates that this parameter is decreased from the free ligand to the metal complexes. In addition, the TDDFT calculations of the anionic form of diketones with different O–O′ distances demonstrate that the calculated maximum wavelength shifts to the red side with the decrease of the distance between oxygens (Fig. 3d). Characteristically, the range of spectral changes in the TDDFT calculations is similar to the experimentally observed spectral range of diketone complexes (see Table S1). This distance depends on the coupling between the 2pσ_z orbitals of the oxygen atoms. This coupling arises from the overlap of the 2pσ_z orbitals of the oxygen atoms with the d orbitals of the metal ions (Fig. 3b, top). In the anionic form of the diketones, the exchange interaction between the oxygen atoms is weak, and the overlap of the 2pσ_z orbitals forms a delocalized σ-bond with an order of less than 0.1. In the complex species, the overlap between the oxygen atoms increases due to the participation of the d_z² orbitals in the coupling (Fig. 3b, middle). This leads to an enhancement of the exchange interaction and a reduction in the O–O′ bond length.

In contrast, the coupling between the 2pσ_z orbitals of oxygen and the d_xz metal orbitals leads to the formation of a molecular orbital with different symmetry, in which the sign of the wave function is opposite near the oxygen atoms (Fig. 3b, bottom). Such overlap decreases the exchange interaction and increases the O–O′ distance. The final distance between the two oxygen atoms results from the combined contributions of both types of molecular orbitals: the d_z² + 2pσ_z interaction, which strengthens the O–O′ coupling, and the d_xz + 2pσ_z interaction, which is responsible for oxygen–oxygen repulsion. The numerical estimation of the contribution of each orbital to the O–O′ distance is difficult due to the different energies and compositions of the MOs. This type of coupling explains the abrupt changes in the absorption wavelengths between the diketonates of 3d, 4f, and 5f metals. The coupling involves the interaction between oxygen 2pσ_z-orbitals and 3d_z², 4d_z², and 5d_z² orbitals of transition metals(III), lanthanides(III), and actinides(IV), respectively. Since the size of an orbital depends directly on the principal quantum number, the orbital size and the O–O′ distance in the corresponding molecular orbital increase abruptly upon transition from 3d, 4f, and 5f metals. In addition, the suggested model explains the correlation between the wavelengths and the stability of the complexes (Table 1): an increase in the 2σpz and d-orbital overlap enhances the metal–ligand interaction, which in turn influences the O–O′ distance and the position of the absorption wavelength.

Taking into account the above conclusions, we can explain the observed spectrochemical order of the ligands (Fig. S3a). As shown earlier, the Mulliken population of the dicarbonyl fragment in the HOMO orbital directly correlates with the observed wavelength within a single ligand. However, these populations do not correlate with the absolute wavelength values when comparing different ligands. Although no direct correlation is observed between the wavelengths and the atomic populations, the difference in population between the dicarbonyl fragment and the substituted ring exhibits a clear relationship with the wavelengths (Fig. S3c).

The difference in population between the dicarbonyl fragment and the heterocyclic moiety demonstrates the polarity of the π-electron distribution in the HOMO, which depends on the electron-donating or electron-accepting ability of the substituted rings. In the order phenyl → pyridine → furan → thiophene → selenophene → tellurophene, the electron-donating ability increases, and so does the wavelength of the ligands and their chelated forms. The exceptions are the nitrogen-containing diketones, N-Me-CF₃ and NN-CF₃. These ligands contain methyl-substituted pyrazole and pyrrole rings, whose dipole moments are not in the plane of the rings, which should alter their π-electron distribution.

Conclusions

In summary, the red shift observed in the spectra of dicarbonyl chelates arises from two factors. First, an increase in the ionic potential of the metal ions leads to a decrease in the π-electronic density of the dicarbonyl fragment and its redistribution toward the multiple bonds of the substituted ring. Second, changes in the O–O′ distance result from the coupling between the 2pσ_z orbitals of oxygen and the d_z² and d_xz orbitals of the metal ions. The observed spectrochemical order of the ligands can be explained in terms of the electron-donating ability of the substituted rings.

Author contributions

MAL – writing, investigation, and conceptualization.

Conflicts of interest

There are no conflicts to declare.

Data availability

The data supporting this article have been included as part of the supplementary information (SI). Supplementary information is available. See DOI: https://doi.org/10.1039/d5nj03964a.

Acknowledgements

This work was performed as part of State Task no FWES-2026-0011 for the Institute of Chemistry and Chemical Technology SB RAS, using the equipment of the Krasnoyarsk Regional Research Equipment Centre of SB RAS. Numerical computations were performed on the MVS-1000M cluster of the Institute of computational modeling SB RAS.

Notes and references

1. G. Hong, X. Gan, C. Leonhardt, Z. Zhang, J. Seibert, J. M. Busch and S. Bräse, Adv. Mater., 2021, 33, 2005630.
2. A. Köhler and H. Bässler, Electronic Processes in Organic Semiconductors: An Introduction, John Wiley & Sons, 2015.
3. E. A. Katlenok, A. V. Rozhkov, R. R. Ramazanov, R. R. Valiev, O. V. Levin, D. O. Goryachiy, I. V. Taydakov, M. L. Kuznetsov and V. Y. Kukushkin, Inorg. Chem., 2022, 61, 8670–8684.
4. E. A. Varaksina, I. V. Taydakov, S. A. Ambrozevich, A. S. Selyukov, K. A. Lyssenko, L. T. Jesus and R. O. Freire, J. Lumin., 2018, 196, 161–168.
5. I. V. Taydakov, A. A. Akkuzina, R. I. Avetisov, A. V. Khomyakov, R. R. Saifutyarov and I. C. Avetissov, J. Lumin., 2016, 177, 31–39.
6. V. S. Sizov, D. A. Komissar, D. A. Metlina, D. F. Aminev, S. A. Ambrozevich, S. E. Nefedov, E. A. Varaksina, M. T. Metlin, V. V. Mislavskii and I. V. Taydakov, Spectrochim. Acta, Part A, 2020, 225, 117503.
7. M. A. Lutoshkin, A. I. Petrov, Y. N. Malyar and A. S. Kazachenko, Inorg. Chem., 2021, 60, 3291–3304.
8. M. A. Lutoshkin, I. V. Taydakov, A. A. Patrusheva and P. I. Matveev, Eur. J. Inorg. Chem., 2024, e202400407.
9. M. A. Lutoshkin, A. A. Patrusheva, I. V. Taydakov and A. I. Rubaylo, J. Mol. Liq., 2024, 408, 125302.
10. M. A. Lutoshkin and I. V. Taydakov, Polyhedron, 2021, 207, 115383.
11. M. A. Lutoshkin and I. V. Taydakov, Polyhedron, 2023, 237, 116395.
12. M. A. Lutoshkin and I. V. Taydakov, J. Sol. Chem., 2023, 52, 304–325.
13. M. A. Lutoshkin, J. Phys. Chem. A, 2023, 127, 8383–8391.
14. M. A. Lutoshkin, I. V. Taydakov and P. I. Matveev, New J. Chem., 1989, 75, 173–194.
15. M. A. Lutoshkin, J. Phys. Chem. A, 2025, 129, 5490–5498.
16. I. B. Bersuker, Electronic Structure and Properties of Transition Metal Compounds: Introduction to the Theory, John Wiley & Sons, New York, 1996.
17. A. M. May and J. L. Dempsey, Chem. Sci., 2024, 15, 6661–6678.
18. M. Suta, F. Cimpoesu and W. Urland, Coord. Chem. Rev., 2021, 441, 213981.

---