Homonuclear 3d transition-metal diatomics: A systematic density functional theory study

A Vision for the Future of Multiscale Modeling

The rise of modern computer science enabled physical chemistry to make enormous progresses in understanding and harnessing natural and artificial phenomena. Nevertheless, despite the advances achieved over past decades, computational resources are still insufficient to thoroughly simulate extended systems from first principles. Indeed, countless biological, catalytic and photophysical processes require ab initio treatments to be properly described, but the breadth of length and time scales involved makes it practically unfeasible. A way to address these issues is to couple theories and algorithms working at different scales by dividing the system into domains treated at different levels of approximation, ranging from quantum mechanics to classical molecular dynamics, even including continuum electrodynamics. This approach is known as multiscale modeling and its use over the past 60 years has led to remarkable results. Considering the rapid advances in theory, algorithm design, and computing power, we believe multiscale modeling will massively grow into a dominant research methodology in the forthcoming years. Hereby we describe the main approaches developed within its realm, highlighting their achievements and current drawbacks, eventually proposing a plausible direction for future developments considering also the emergence of new computational techniques such as machine learning and quantum computing. We then discuss how advanced multiscale modeling methods could be exploited to address critical scientific challenges, focusing on the simulation of complex light-harvesting processes, such as natural photosynthesis. While doing so, we suggest a cutting-edge computational paradigm consisting in performing simultaneous multiscale calculations on a system allowing the various domains, treated with appropriate accuracy, to move and extend while they properly interact with each other. Although this vision is very ambitious, we believe the quick development of computer science will lead to both massive improvements and widespread use of these techniques, resulting in enormous progresses in physical chemistry and, eventually, in our society.

Unraveling actinide-actinide bonding in fullerene cages: a DFT versus ab initio methodological study

Actinide-actinide bonding poses a challenge for both experimental and theoretical chemists because of both the scarcity of experimental data and the exotic nature of actinide bonding due to the involvement and mixing of actinide 7s-, 6p-, 6d-, and particularly 5f-orbitals. Only a few experimental examples of An-An bonding have been reported so far. Here, we perform a methodological study of actinide-actinide bonding on experimentally known Th2@C80 and U2@C80 systems. We compared selected GGA, meta-GGA, hybrid-GGA and range-separated hybrid-GGA functionals with the results obtained using a multireference CASPT2 method, which we consider as a reference point. We show that functionals such as BP86, PBE or TPSS perform well for predicting geometries, while range-separated hybrids are superior in the description of the chemical bonding. None of the tested functionals were deemed reliable regarding the correct electronic spin ground state. Based on the results of this methodological study, we re-evaluate selected previously studied diactinide fullerene systems using more reliable protocol.

The performance of CCSD(T) for the calculation of dipole moments in diatomics

This work analyzes the accuracy of the coupled cluster with single, double, and perturbative triple excitation [CCSD(T)] method for predicting dipole moments. In particular, we benchmark CCSD(T) predictions for the equilibrium bond length, vibrational frequency, and dipole moment versus accurate experimental data. As a result, we find that CCSD(T) leads to accurate dipole moments. However, in some cases, it disagrees with the experimental values, and the disagreement can not be satisfactorily explained via relativistic or multi-reference effects. Therefore, our results indicate that benchmark studies for energy and geometry properties do not accurately describe other electron density magnitudes.

Computational Investigation of the Metal and Ligand Substitution Effects on the Structure and Electronic States of the Phosphoranimide Tetramer Complexes of Cu(I), Ni(I), Co(I), and Fe(I)

The structurally unique saddle-shaped paramagnetic tetrametallic clusters of Co(I) and Ni(I) with phosphoranimide ligands have been synthesized and proposed as catalyst precursors. The analogous Cu(I) nanocluster is planar and diamagnetic. These notable variations in geometry and ground electronic states indicate that the effect of metal and ligand substituents on the structure and electronic properties of these complexes requires investigation. We present a computational study of a series of these novel homoleptic complexes containing Co(I), Ni(I), and Cu(I) as well as Fe(I) coordinated to phosphoranimides with electron-donating and withdrawing substituents, conducted at the relativistic density functional theory level using ZORA-PBE/TZP. The optimized structures of the saddle-shaped Co(I) and Ni(I) and planar Cu(I) tetramers with linear N-M-N coordination are validated with respect to X-ray diffraction determinations. The ground-state analysis indicates that Cu(I) complexes are diamagnetic, whereas Ni(I) and Co(I) complexes are in high-spin states, in agreement with magnetic susceptibility measurements. The computational results show that Fe(I) complexes are saddle shaped and high spin. The Co(I) complex is stabilized by a metal macrocycle distortion from square to diamond, as elucidated from its Walsh diagram. The effects of metals and ligand substituents on the ground electronic state, metal center coordination environment, and energy of the complexes are investigated. The bulky tertiary butyl substituent causes the largest saddle-shape distortion of the tetramer marcocycle, which partially offsets its electron-donating effect. Macrocycle distortions with N-M-N site angles ranging from obtuse to alternating obtuse reflex are correlated with the increasing number of unpaired electrons. The phenyl-substituted complexes are expected to have the highest reactivity toward electrophiles. Understanding the interplay between structural and electronic parameters is intended to guide the development of synthetic cooperative systems for multielectron redox reactions, models of biological systems, and molecular magnets.

A Study of NbMo and NbMo- by Anion Photoelectron Spectroscopy

Photoelectron spectra of the niobium-molybdenum diatomic anion, obtained at 488 and 514 nm, display vibrationally resolved transitions from the ground state and one excited electronic state of the anion to the ground state and one excited electronic state of the neutral molecule. The electron affinity of NbMo is measured to be 1.130 ± 0.005 eV. Its ²Δ_3/2 spin-orbit component is observed to lie 870 ± 20 cm^-1 above its previously identified ²Δ_5/2 ground state. For ⁹³Nb⁹⁸Mo, vibrational energies measured for levels up to v = 4 for the ²Δ_5/2 and ²Δ_3/2 states give harmonic frequency and anharmonicity constant values of ω_e = 492 ± 12 cm^-1 and ω_ex_e = 8.0 ± 3.2 cm^-1, the former value corresponding to a force constant of 6.80 ± 0.35 mdyn/Å. These two vibrational parameters suggest a bond dissociation energy that is too low by at least a factor of 3, indicating that the ground state potential energy curve of NbMo deviates markedly from a Morse potential at higher energies. An excited electronic state of NbMo, assigned as a ²Σ⁺ state, is observed at 2900 ± 25 cm^-1 (T₀). Vibrational energies up to v = 8 in this excited state give values of ω_e = 544 ± 8 cm^-1 and ω_ex_e = 1.9 ± 1.2 cm^-1 for ⁹³Nb⁹⁸Mo. The former value corresponds to a high vibrational force constant of 8.30 ± 0.25 mdyn/Å. Both doublet states of the neutral molecule are accessed from the anion ground state, which is assigned as ¹Σ⁺. For the ⁹³Nb⁹⁸Mo^- anion, the fundamental vibrational frequency (ΔG_1/2) is 484 ± 15 cm^-1. Electron affinity data indicate that the bond dissociation energy of NbMo^- is 0.213 ± 0.005 eV greater than that of neutral NbMo, whose previously reported value then gives D₀ = 4.85 ± 0.27 eV for the anion. An excited state of the anion lying 3050 ± 25 cm^-1 (T₀) above its ground state is assigned as ³Δ, and the energies of its spin-orbit components above the ³Δ₃ lowest energy level are measured to be 450 ± 20 cm^-1 (³Δ₂) and 1100 ± 20 cm^-1 (³Δ₁). Their uneven spacing suggests that the energy of the ³Δ₂ level is lowered by interaction with a higher energy Ω = 2 anion state. The vibrational frequency (ΔG_1/2) for the ³Δ₁ and ³Δ₂ states is measured to be 433 ± 20 cm^-1. Bond length differences among the observed states are estimated from Franck-Condon fits to vibrational band intensity profiles. When combined with the previously reported NbMo bond length, these provide bond length estimates for the ground state of the anion (1.940 ± 0.025 Å) and for the observed excited states. These species offer extreme examples of multiple metal-metal bonding, with formal bond orders of 5¹/₂ for the ²Δ ground and ²Σ⁺ excited doublet states of NbMo, 6 for the singlet ground state of the anion, and 5 for its low-lying triplet state. The relationships among their bonding properties and those of related multiply bonded transition metal dimers are discussed.

Taming the Antiferromagnetic Beast: Computational Design of Ultrashort Mn-Mn Bonds Stabilized by N-Heterocyclic Carbenes

The development of complexes featuring low-valent, multiply bonded metal centers is an exciting field with several potential applications. In this work, we describe the design principles and extensive computational investigation of new organometallic platforms featuring the elusive manganese-manganese bond stabilized by experimentally realized N-heterocyclic carbenes (NHCs). By using DFT computations benchmarked against multireference calculations, as well as MO- and VB-based bonding analyses, we could disentangle the various electronic and structural effects contributing to the thermodynamic and kinetic stability, as well as the experimental feasibility, of the systems. In particular, we explored the nature of the metal-carbene interaction and the role of the ancillary η6 coordination to the generation of Mn2 systems featuring ultrashort metal-metal bonds, closed-shell singlet multiplicities, and positive adiabatic singlet-triplet gaps. Our analysis identifies two distinct classes of viable synthetic targets, whose electrostructural properties are thoroughly investigated.

From π Bonds without σ Bonds to the Longest Metal-Metal Bond Ever: A Survey on Actinide-Actinide Bonding in Fullerenes

Actinide-actinide bonds are rare. Only a few experimental systems with An-An bonds have been described so far. Recent experimental characterization of the U₂@I_h(7)-C₈₀ (J. Am. Chem. Soc. 2018, 140, 3907) system with one-electron two-center (OETC) U-U bonds as was predicted by some of us (Phys. Chem. Chem. Phys. 2015, 17, 24182) encourages the search for more examples of actinide-actinide bonding in fullerene cages. Here, we investigate actinide-actinide bonding in An₂@D_5h(1)-C₇₀, An₂@I_h(7)-C₈₀, and An₂@D_5h(1)-C₉₀ (An = Ac-Cm) endohedral metallofullerenes (EMFs). Using different methods of the chemical bonding analysis, we show that most of the studied An₂@C₇₀ and An₂@C₈₀ systems feature one or more one-electron two-center actinide-actinide bonds. Unique bonding patterns are revealed in plutonium EMFs. The Pu₂@I_h(7)-C₈₀ features two OETC Pu-Pu π bonds without any evidence of a corresponding σ bond. In the Pu₂@D_5h(1)-C₉₀ with r_Pu-Pu = 5.9 Å, theory predicts the longest metal-metal bond ever described. Predicted systems are thermodynamically stable and should be, in principle, experimentally accessible, though radioactivity of studied metals may be a serious obstacle.

Analysis of the structures, stabilities and electronic properties of MB16− (M = V, Cr, Mn, Fe, Co, Ni) clusters and assemblies

Stacking of lowest-energy structures of Fe₂B₂₄⁻ and Co₂B₂₄⁻ dimers.

Assessing the Strength of Metal-Metal Interactions

A proper comparison between bond strengths of different atom pairs is relevant only for the same formal bond order (BO) of atomic interactions, e.g., for single bonds, because it is clear that the higher is the BO or the number of the electron pairs responsible of bonding, the stronger is the bond. For the metal-metal interactions, such a comparison of the bond strengths is especially problematic because formal BOs may differ from the effective ones by more than 1 v.u. (valence units). In this paper, we investigated the strength of bonding and its correlation to structural parameters (bond length/BO) for 24 metal-metal pairs. A simple way of the BO and bond-strength analysis was proposed and verified on the transition metal dimers. In contrast to the previous studies, the effective BOs were not calculated from spectroscopic data or related to reference compounds, but determined independently based on the Pauling model and bond valence parameters obtained from recent quantum chemistry data. To characterize the bond strength, we used the force constants based on the available experimental or DFT data of stretching frequencies. A linear correlation between the effective BOs and force constants of the metal-metal bonds, which was confirmed in this work, allows for prediction of the stretching frequencies according to the effective BOs (and/or the bond lengths) and vice versa.

Metal-Metal (MM) Bond Distances and Bond Orders in Binuclear Metal Complexes of the First Row Transition Metals Titanium Through Zinc

This survey of metal-metal (MM) bond distances in binuclear complexes of the first row 3d-block elements reviews experimental and computational research on a wide range of such systems. The metals surveyed are titanium, vanadium, chromium, manganese, iron, cobalt, nickel, copper, and zinc, representing the only comprehensive presentation of such results to date. Factors impacting MM bond lengths that are discussed here include (a) the formal MM bond order, (b) size of the metal ion present in the bimetallic core (M₂) ⁿ⁺, (c) the metal oxidation state, (d) effects of ligand basicity, coordination mode and number, and (e) steric effects of bulky ligands. Correlations between experimental and computational findings are examined wherever possible, often yielding good agreement for MM bond lengths. The formal bond order provides a key basis for assessing experimental and computationally derived MM bond lengths. The effects of change in the metal upon MM bond length ranges in binuclear complexes suggest trends for single, double, triple, and quadruple MM bonds which are related to the available information on metal atomic radii. It emerges that while specific factors for a limited range of complexes are found to have their expected impact in many cases, the assessment of the net effect of these factors is challenging. The combination of experimental and computational results leads us to propose for the first time the ranges and "best" estimates for MM bond distances of all types (Ti-Ti through Zn-Zn, single through quintuple).

Electronic ground state of Ni2. J Chem Phys 2016;145:194302. [PMID: 27875883 DOI: 10.1063/1.4967821]

The Φ9/24 ground state of the Ni₂⁺ diatomic molecular cation is determined experimentally from temperature and magnetic-field-dependent x-ray magnetic circular dichroism spectroscopy in a cryogenic ion trap, where an electronic and rotational temperature of 7.4±0.2 K was reached by buffer gas cooling of the molecular ion. The contribution of the spin dipole operator to the x-ray magnetic circular dichroism spin sum rule amounts to 7T_z=0.17±0.06μ_B per atom, approximately 11% of the spin magnetic moment. We find that, in general, homonuclear diatomic molecular cations of 3d transition metals seem to adopt maximum spin magnetic moments in their electronic ground states.

Low-energy excited states of divanadium: a matrix isolation and MRCI study

The ground and excited electronic states of the vanadium dimer (V2) have been studied using Ne matrix isolation experiments and quantum chemical calculations (multireference configuration interaction based on complete active space self-consistent orbitals). In the near infrared absorption spectrum, two vibrational progressions of a new electronic term with a large number of members have been observed with the origin at 1.08 eV and a fundamental vibrational quantum of 475 cm(-1). With the aid of calculations, it has been assigned to a (3)Πu electronic term. The calculations yield potential energy curves for a large number of singlet, triplet, and quintet electronic terms.

Distinguishing Bonds

The energy change per electron in a chemical or physical transformation, ΔE/n, may be expressed as Δχ̅ + Δ(VNN + ω)/n, where Δχ̅ is the average electron binding energy, a generalized electronegativity, ΔVNN is the change in nuclear repulsions, and Δω is the change in multielectron interactions in the process considered. The last term can be obtained by the difference from experimental or theoretical estimates of the first terms. Previously obtained consequences of this energy partitioning are extended here to a different analysis of bonding in a great variety of diatomics, including more or less polar ones. Arguments are presented for associating the average change in electron binding energy with covalence, and the change in multielectron interactions with electron transfer, either to, out, or within a molecule. A new descriptor Q, essentially the scaled difference between the Δχ̅ and Δ(VNN + ω)/n terms, when plotted versus the bond energy, separates nicely a wide variety of bonding types, covalent, covalent but more correlated, polar and increasingly ionic, metallogenic, electrostatic, charge-shift bonds, and dispersion interactions. Also, Q itself shows a set of interesting relations with the correlation energy of a bond.

A unified set of experimental organometallic data used to evaluate modern theoretical methods

We applied a test set of ligand dissociation enthalpies derived entirely from a unified experimental approach to evaluate the efficacy of various methods for modeling organometallic chemistry.

Multiconfiguration Pair-Density Functional Theory: A Fully Translated Gradient Approximation and Its Performance for Transition Metal Dimers and the Spectroscopy of Re2Cl8(2-)

We extend the on-top density functional of multiconfiguration pair-density functional theory (MC-PDFT) to include the gradient of the on-top density as well as the gradient of the density. We find that the theory is reasonably stable to this extension; furthermore, it provides improved accuracy for molecules containing transition metals. We illustrate the extended on-top density functionals by applying them to Cr2, Cu2, Ag2, Os2, and Re2Cl8(2-) as well as to our previous database of 56 data for bond dissociation energies, barrier heights, reaction energies, proton affinities, and the water dimer. The performance of MC-PDFT is comparable to or better than that of CASPT2.

Ligand-free Ni nanocluster formation at atmospheric pressure via rapid quenching in a microplasma process

The production of metal nanoclusters composed of less than 10(3) atoms is important for applications in energy conversion and medicine, and for fundamental studies of nanomaterial nucleation and growth. Unfortunately, existing synthesis methods do not enable adequate control of cluster formation, particularly at atmospheric pressure wherein formation typically occurs on sub-millisecond timescales. Here, we demonstrate that ligand-free, unagglomerated nickel nanoclusters can be continuously synthesized at atmospheric pressure via the decomposition of bis(cyclopentadienyl)nickel(II) (nickelocene) in a spatially-confined microplasma process that rapidly quenches particle growth and agglomeration. The clusters were measured on line by ion mobility spectrometry (IMS) and further analyzed by atomic force microscopy (AFM). Our results reveal that stable clusters with spherical equivalent mean diameters below 10 Åare produced, and by controlling the nickelocene concentration, the mean diameter can be tuned up to ∼50 Å. Although diameter is often the sole metric used in nanocluster and nanoparticle characterization, to infer the number of atoms in AFM and IMS detected clusters, we compare measured AFM heights and IMS inferred collision cross sections to theoretical predictions based on both bulk matter approximations and density functional theory and Hartree-Fock calculated Ni nanocluster structures (composed of 2-15 atoms for the latter). The calculations suggest that Ni nanoclusters composed of less than 10(2) atoms can be produced repeatably with simple microplasma reactors.

Photoelectron spectroscopy of CoC2H2(-) and density functional study of Co(n)C2H2 (n = 1-3) anion and neutral clusters

The anionic and neutral ConC2H2 (n = 1-3) clusters were investigated using anion photoelectron spectroscopy and density functional calculations. The adiabatic detachment energies and vertical detachment energies of ConC2H2(-) (n = 1-3) were determined. Our results show that the most stable geometries of anionic ConC2H2(-) (n = 1-2) and neutral ConC2H2 (n = 1-3) are composed of ConC2H clusters adsorbing a hydrogen atom on the top or bridge sites of Con, whereas Co3C2H2(-) consists of a five-member ring of Co3C2 carbide adsorbing two hydrogen atoms on two bridge sites of Co3. The reaction mechanisms show that the inserted isomer HCoC2H can convert into the vinylidene complex Co═C═CH2 via a side-on isomer M-η(2)-(C2H2).

Comparative analysis of the performance of commonly available density functionals in the determination of geometrical parameters for copper complexes

In this study, a set of 50 transition-metal complexes of Cu(I) and Cu(II), were used in the evaluation of 18 density functionals in geometry determination. In addition, 14 different basis sets were considered, including four commonly used Pople's all-electron basis sets; four basis sets including popular types of effective-core potentials: Los Alamos, Steven-Basch-Krauss, and Stuttgart-Dresden; and six triple-ζ basis sets. The results illustrate the performance of different methodological alternatives for the treatment of geometrical properties in relevant copper complexes, pointing out Double-Hybrid (DH) and Long-range Correction (LC) Generalized Gradient Approximation (GGA) methods as better descriptors of the geometry of the evaluated systems. These however, are associated with a computational cost several times higher than some of the other methods employed, such as the M06 functional, which has also demonstrated a comparable performance. Regarding the basis sets, 6-31+G(d) and 6-31+G(d,p) were the best performing approaches. In addition, the results show that the use of effective-core potentials has a limited impact, in terms of the accuracy in the determination of metal-ligand bond-lengths and angles in our dataset of copper complexes. Hence, these could become a good alternative for the geometrical description of these systems, particularly CEP-121G and SDD basis sets, if one is considering larger copper complexes where the computational cost could be an issue.

Unprecedented tris-phosphido-bridged triangular clusters with 42 valence electrons. Chemical, electrochemical and computational studies of their formation and stability

This paper presents the synthesis and structural characterization of the unprecedented tris-phosphido-bridged compounds Pt3(μ-PBu(t)2)3X3 (X = Cl, Br, I), having only 42 valence electrons, while up to now analogous clusters typically have 44e(-). The new species were obtained by an apparent bielectronic oxidation of the 44e(-) monohalides Pt3(μ-PBu(t)2)3(CO)2X with the corresponding dihalogen X2. Their X-ray structures are close to the D3h symmetry, similarly to the 44e(-) analogues with three terminal carbonyl ligands. The products were also obtained by electrochemical oxidation of the same monohalides in the presence of the corresponding halide. In a detailed study on the formation of Pt3(μ-PBu(t)2)3I3, the redox potentials indicated that I2 can only perform the first monoelectronic oxidation but is unsuited for the second one. Accordingly, the 43e(-) intermediate [Pt3(μ-PBu(t)2)3(CO)2I](+) was ascertained to play a key role. Another piece of information is that, together with the fully oxidized product Pt3(μ-PBu(t)2)3I3, the transient 44e(-) species [Pt3(μ-PBu(t)2)3(CO)3](+) is formed in the early steps of the reaction. In order to extract detailed information on the formation pathway, involving both terminal ligand substitutions and electron transfer processes, a DFT investigation has been performed and all the possible intermediates have been defined together with their associated energy costs. The profile highlights many important aspects, such as the formation of an appropriate couple of 43e(-) intermediates having different sets of terminal coligands, and suitable redox potentials for the transfer of one electron. Optimizations of 45e(-) associative intermediates in the ligand substitution reactions indicate their possible involvement in the redox process with reduction of the overall energy cost. Finally, according to MO arguments, the unique stability of the 42e(-) phosphido-bridged Pt3 clusters can be attributed to the simultaneous presence of three terminal halides.

Performance of density functional methods. Some difficult cases for small systems containing Cu, Ag, or Au

Twenty-six density functional theory (DFT) methods were tested in conjunction with three different effective core potentials (ECPs) and their corresponding valence basis sets, for studying the behavior of DFT methods in small systems containing Cu, Ag, or Au where it is well-known that some functionals fail. The DFT results were compared with those obtained with post Hartree-Fock methods: second-order many-body perturbation theory (MP2), coupled cluster singles-doubles (CCSD), and coupled cluster singles-doubles with perturbative triples, CCSD(T). Calculations were carried for M3 (M = Cu, Ag, Au); M4(-), (M = Cu, Ag) and [H2O-Cu](+2). The comparison of the DFT calculated values with the Post Hartree-Fock values showed that, in general, all generalized gradient approximation (GGA) type functionals fail to describe these systems. The hybrid GGA functionals (H-GGA) showed a better behavior; however, when the Lee-Yang-Parr (LYP) exchange-correlation functional was used, wrong results were obtained. The results with the hybrid meta (HM-GGA) functionals, as in the case of H-GGAs, showed that, to obtain similar results to MP2 or CCSD(T), it is necessary to have a high Hartree-Fock exchange percentage. Spurious results obtained with the H-GGA or HM-GGA methods can be eliminated increasing the Hartree-Fock exchange percentage in the H-GGA or HM-GGA type functionals. Among the different functionals tested, the BB1K and MPWB1K functionals showed the best agreement with the MP2 and CCSD(T) results.

Defect induced electronic structure of uranofullerene

The interaction between the inner atoms/cluster and the outer fullerene cage is the source of various novel properties of endohedral metallofullerenes. Herein, we introduce an adatom-type spin polarization defect on the surface of a typical endohedral stable U(2)@C(60) to predict the associated structure and electronic properties of U(2)@C(61) based on the density functional theory method. We found that defect induces obvious changes in the electronic structure of this metallofullerene. More interestingly, the ground state of U(2)@C(61) is nonet spin in contrast to the septet of U(2)@C(60). Electronic structure analysis shows that the inner U atoms and the C ad-atom on the surface of the cage contribute together to this spin state, which is brought about by a ferromagnetic coupling between the spin of the unpaired electrons of the U atoms and the C ad-atom. This discovery may provide a possible approach to adapt the electronic structure properties of endohedral metallofullerenes.

Comparative Study of Single and Double Hybrid Density Functionals for the Prediction of 3d Transition Metal Thermochemistry

The performance of 13 density functionals, including hybrid-GGA, hybrid-meta-GGA, and double-hybrid functionals, in combination with the correlation consistent basis sets, has been evaluated for the prediction of gas phase enthalpies of formation for a large set of 3d transition-metal-containing molecules with versatile bonding features. Of the methods studied, the hybrid B97-1 functional and the double hybrid functional mPW2-PLYP exhibit the best overall performance with mean absolute deviations (MAD) from experimental data of 7.2 and 7.3 kcal mol(-1), respectively. For single reference molecules, where dynamic correlation predominates, the results of the hybrid functionals B97-1, B98, and ωB97X and the double hybrid functionals B2-PLYP, B2GP-PLYP, and mPW2-PLYP yield the smallest deviations from the experimental enthalpies of formation. For the prediction of thermodynamic properties of coordination complexes including metal carbonyls, B97-1 and mPW2-PLYP are the most promising functionals of those investigated. When the size of the molecule is considered, B97-1 and B98 outperform mPW2-PLYP for diatomics and triatomics, while mPW2-PLYP yields the lowest MAD for larger molecules.