Relativistic Approximations to Paramagnetic NMR ... - ACS Publications

Jun 21, 2017 - Results obtained with the scalar relativistic (SR) Douglas−Kroll−Hess Hamiltonian used for the g- and ZFS tensor calculations are c...
0 downloads 0 Views 2MB Size
Subscriber access provided by EAST TENNESSEE STATE UNIV

Article

Relativistic Approximations to Paramagnetic NMR Chemical Shift and Shielding Anisotropy in Transition Metal Systems Syed Awais Rouf, Ji#í Mareš, and Juha Vaara J. Chem. Theory Comput., Just Accepted Manuscript • DOI: 10.1021/acs.jctc.7b00168 • Publication Date (Web): 21 Jun 2017 Downloaded from http://pubs.acs.org on June 23, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Journal of Chemical Theory and Computation is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Chemical Theory and Computation

Relativistic Approximations to Paramagnetic NMR Chemical Shift and Shielding Anisotropy in Transition Metal Systems Syed Awais Rouf, Jiˇr´ı Mareˇs, and Juha Vaara∗ NMR Research Unit, P.O. Box 3000, FIN-90014 University of Oulu, Finland E-mail: [email protected]

Abstract We apply approximate relativistic methods to calculate the magnetic property tensors, i.e., the g-tensor, zero-field splitting (ZFS) tensor (D), and hyperfine coupling (HFC) tensors, for the purpose of constructing paramagnetic nuclear magnetic resonance (pNMR) shielding tensors. The chemical shift and shielding anisotropy are calculated by applying a modern implementation of the classic Kurland-McGarvey theory (J. Magn. Reson. 1970, 2, 286), which formulates the shielding tensor in terms of the g- and HFC tensors obtained for the ground multiplet, in the case of higher than doublet multiplicity defined by the ZFS interaction. The g- and ZFS tensors are calculated by ab initio complete active space self-consistent field and N -electron valence-state perturbation theory methods, with spin-orbit (SO) effects treated via quasidegenerate perturbation theory. Results obtained with the scalar relativistic (SR) Douglas-Kroll-Hess Hamiltonian used for the g- and ZFS tensor calculations are compared with nonrelativistically based computations. The HFC tensors computed using the fully relativistic four-component matrix Dirac-Kohn-Sham approach are contrasted against perturbationally SO-corrected nonrelativistic results, in the density-functional

1

ACS Paragon Plus Environment

Journal of Chemical Theory and Computation

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

theory framework. These approximations are applied on paramagnetic metallocenes (MCp2 ) (M = Ni, Cr, V, Mn, Co, Rh, Ir), a Co(II) pyrazolylborate complex, and a Cr(III) complex. SR effects are found to be small for g and D in these systems. The HFCs are found to be more influenced by relativistic effects for the 3d systems. However, for some of the 3d complexes, nonrelativistic calculations give a reasonable agreement with the experimental chemical shift and shielding anisotropy. The influence of scalar relativity is strong for the 5d IrCp2 system. This mixed ab initio/DFT technique, with a fully relativistic method used for the critical HFC tensor, should be useful for the treatment of both electron correlation and relativistic effects at a reasonable computational cost, to compute the pNMR shielding tensors in transition-metal systems.

1

Introduction

Paramagnetic ions play an important role in the determination of nuclear magnetic resonance (NMR) spectra of proteins and other biomolecules, as well as in materials sciences. 1–5 Electronic structure methods for determining the NMR shielding tensor σ and the associated chemical shift δ can aid in the prediction, interpretation, and assignment of NMR spectra. 6 The involvement of relativistic electronic structure theory 7–11 and the application of ab initio wavefunction-based methods 8,9,12,13 have increased the potential of quantum-chemical approaches in the prediction and interpretation of paramagnetic NMR (pNMR) shifts and shielding anisotropies. When parameterizing the pNMR shielding in terms of the electron paramagnetic resonance (EPR) tensors of the ground multiplet, as in the classic Kurland-McGarvey pNMR shielding theory, 14 one needs to reliably compute the g-, hyperfine coupling (HFC, A), and zero-field splitting (ZFS, D) tensors. The treatment of relativistic effects is important for the EPR parameters. The HFC tensor is due, in the nonrelativistic (NR) framework, to the isotropic Fermi contact and the anisotropic spin-dipolar interactions. Relativistic effects 2

ACS Paragon Plus Environment

Page 2 of 42

Page 3 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Chemical Theory and Computation

on HFC are pronounced because such hyperfine interactions depend on the region of the electronic charge distribution close to the nuclei. 7 The g- and ZFS tensors arise from the spin-orbit (SO) interactions, which can be included perturbationally or variationally. The latter means 2- or 4-component calculations, where SO effects are treated to infinite order. An infinite-order treatment is preferred in cases with strong SO interaction. ZFS determines, in systems with more than one unpaired electron, the energy level structure of the ground multiplet. In calculations including treatment of SO coupling, the ZFS can be found from the energies of the 2S + 1 lowest states, where S is the electron spin quantum number. In heavy-element systems, g, A, and D are also affected by scalar relativistic (SR) effects. Most of the available relativistic quantum-chemical implementations of EPR parameters are based on one- or two-component versions of methods based on transformed Hamiltonians, 15,16 e.g., the zeroth-order regular approximation (ZORA 17 ) and the Douglas-Kroll-Hess (DKH 18,19 ) Hamiltonian. However, fully relativistic, four-component calculations of the gtensor at the Hartree-Fock level have been reported in Refs. 20,21 and at the density-functional theory (DFT) level in the matrix Dirac-Kohn-Sham (mDKS) framework, in Ref. 22 In 2006, Komorovsk´ y et al. 23 published a relativistic two-component DKS2-RI approach, based on the DKS Hamiltonian and the resolution-of-identity (RI) technique. The approach was applied to relativistic calculations of g- and A tensors. Malkin et al. 24 studied the effect of a finite-size model for both the nuclear charge and magnetic moment distributions on HFC tensors calculated by the mDKS technique. This method provides an attractive alternative to the approximate two-component methods, as it avoids picture-change effects in property calculations. 25 Recently, the four-component mDKS approach for the calculations of EPR gand HFC tensors, within the restricted kinetic balance framework, was extended 26 to hybrid functionals in the ReSpect code. 27 The development of relativistic methods for the NMR shielding tensor of diamagnetic systems has been reviewed in the literature. 6,28–43 In pNMR, experimental data are typically obtained for light nuclei in heavy-element systems, with the NMR nuclei relatively far from

3

ACS Paragon Plus Environment

Journal of Chemical Theory and Computation

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

the paramagnetic center. This is due to the fact that the nuclei close to the paramagnetic center relax very fast and the associated line broadening renders the NMR signals unobservable in such cases. This somewhat alleviates the demands for a relativistic treatment in the pNMR context, but already for 3d transition element systems, the SR or strong SO effects may be expected to be relevant for comparison with pNMR experiments. Several computational studies on pNMR shieldings and chemical shifts have been reported in recent years. 8–13,22,44–51 In Ref., 49 pNMR shieldings for doublet systems (group-IX metallocenes) were calculated by using a first-principles method with SR effective core potentials, while the SO effects were included perturbationally. In Ref., 50 a DFT-based method for paramagnetic molecules with arbitrary spin multiplicity was implemented, featuring an a posteriori inclusion of the ZFS interaction. Autschbach, Patchkovskii, and Pritchard 10 calculated the hyperfine tensors and paramagnetic chemical shifts for Ru(III) complexes using DFT with the SR effects treated at the ZORA level and SO effects perturbationally by linear response theory. In 2013, Komorovsk´ y et al. 11 introduced a four-component relativistic method for the pNMR shielding tensors of doublet systems, using both restricted kinetically and magnetically balanced basis sets 25 along with gauge-including atomic orbitals to ensure rapid basisset convergence. Application to the 1 H NMR chemical shifts of a series of 3d, 4d, and 5d transition-metal hydrides was reported. 51 For, e.g., lanthanide and actinide systems, still stronger relativistic effects can be expected. Eventually, the Kurland and McGarvey approach based on standard EPR parameters 14 may have to be abandoned due to the strong SO coupling and, instead, generalized EPR tensors have to be employed. 8 Ref. 8 reported pNMR chemical shifts calculated for lanthanide-doped fluorite crystals using the ab initio complete active space self-consistent field (CASSCF) method. 52 One may also refrain from using the EPR parameters altogether and, instead, directly employ the relativistic manyelectron wave functions of the ground and excited electronic states, to calculate pNMR shielding. 9 Gendron, Sharkas, and Autschbach 9 used multiconfigurational methods, includ-

4

ACS Paragon Plus Environment

Page 4 of 42

Page 5 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Chemical Theory and Computation

ing SR effects taken into account with the DKH Hamiltonian, to compute pNMR shieldings in this way. The present authors 12,13 reported pNMR shifts according to the Kurland-McGarvey theory 14,46 for 3d transition metal systems, where SO effects were included for the g- and ZFS tensors by an ab initio two-step, post-SCF treatment, combined with DFT to calculate the HFC tensors. A useful accuracy for such systems was obtained, despite the fact that no SR effects were included. In the present paper, we investigate the 1 H and

13

C pNMR chemical

shifts and shielding anisotropies for both S = 1/2 and S > 1/2 metallocenes, as well as the 1 H pNMR shifts and shielding anisotropies for larger quartet S = 3/2 systems, using the formalism applied in Refs. 12,13 We extend upon our previous methodology by including SR approximations to the g-, ZFS, and HFC tensors, used to parameterize the pNMR shielding tensors. As previously, 12,13 g and D are calculated ab initio, 53–55 based on a NR one-component CASSCF wavefunction followed by quasidegenerate perturbation theory (QDPT) treatment of the SO interaction, 54,55 with dynamic correlation corrections by N electron valence-state perturbation theory (NEVPT2 56,57 ). This level is denoted ”NR+SO” in the following. Alternatively, the relativistic second-order DKH Hamiltonian is used to produce the underlying one-component wavefunction, before the SO step, which involves the appropriately picture-changed SO as well as, for the g-tensor calculation, the Zeeman interaction 55 (DKH+SO). On the other hand, the HFC tensors are calculated by hybrid DFT using the PBE0 58 exchange-correlation functional at both the NR level, supplemented with first-order perturbational SO corrections (the level used in Refs., 12,13 NR+SO) and at the fully relativistic four-component mDKS level. 24 The different approximations used for the three present tensors (g, D, A) can be motivated by their different expected sensitivity to relativistic effects (vide supra). We believe that the influence of SR effects on pNMR shielding tensors has not been subjected to such a systematic investigation before. Corresponding methodology was recently applied to the calculation of pNMR spin-spin coupling enhancement. 59

5

ACS Paragon Plus Environment

Journal of Chemical Theory and Computation

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 42

We demonstrate this combined methodology with calculations on paramagnetic metallocenes that have been selected, on the one hand, with the central metal ion across the 3d period and, on the other hand, down group IX of the periodic table, with 3d, 4d, and 5d doublet metallocenes. These metallocenes have already been investigated in pNMR shielding calculations in many other studies 4,12,13,44,48–50,60 and represent different spin states, i.e., S = 1 (nickelocene and chromocene), 3/2 (vanadocene), and 5/2 (manganocene), all with the central atom in the 3d period, as well as the group-IX S = 1/2 cobaltocene, rhodocene, and iridocene. In addition, larger quartet (S = 3/2) Co(II) and Cr(III) complexes are studied in the present work.

2 2.1

Theory Paramagnetic NMR shielding

We used the formalism of the pNMR nuclear shielding and chemical shift presented in Refs. 12–14,46 for S > 1/2 systems, valid for comparatively weak SO coupling 8 and the ZFS Hamiltonian possessing the form HZFS = S · D · S. It gives us the following expression for the nuclear shielding tensor: µB σ = σorb − g · hSSi · A γkT P Q hn|S|mihm|S|ni nm P mn hSSi = n exp (−En /kT )    exp (−En /kT ) Qmn =   − kT [exp (−Em /kT ) − exp (−En /kT )] Em −En

(1) (2) ;

En = Em

;

En 6= Em

,

(3)

which, for S=1/2 systems, reduces to the doublet expression 45,49

σ = σorb −

µB g· A, γkT

6

ACS Paragon Plus Environment

(4)

Page 7 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Chemical Theory and Computation

where the effects of ZFS are absent. In these formulae, γ, µB , k, and T are the gyromagnetic ratio of the nucleus, the Bohr magneton, the Boltzmann constant, and the absolute temperature, respectively. S is the effective spin operator and hSSi is a dyadic with the components hSǫ Sτ i evaluated in the manifold of electronic states |ni, with energies En , consisting here of the ground-state zero-field split spin multiplet, the eigenfunctions and eigenvalues of HZFS . In Eq. (1), the first term is the orbital shielding, which is approximately independent of temperature. 61–63 The second, both explicitly and implicitly (via hSSi) temperature-dependent hyperfine shielding term consists of the generalized product of the g-tensor, hSSi, and the HFC tensor. The different physical contributions to the hyperfine shielding are listed in Table S1 in the Supporting Information. A detailed discussion of the shielding formalism is available in Refs. 12,13 The isotropic shielding constant σ and the chemical shift δ were obtained from the calculated shielding tensor by using

σ =

σxx + σyy + σzz 3

δ = σref − σ,

(5) (6)

respectively. Here, σref is the isotropic shielding constant in a diamagnetic reference compound, in our case tetramethylsilane, TMS. The shielding anisotropy was calculated for the axially symmetric metallocenes as

∆σ = σk − σ⊥ ,

(7)

the difference of the shielding tensor component along the effective molecular symmetry axis and the perpendicular component.

7

ACS Paragon Plus Environment

Journal of Chemical Theory and Computation

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

2.2

Relativistic Approximations

The g- and D tensor calculations were carried out using the method and implementation of Refs. 54,55 at two levels of treating relativistic effects, i.e., the NR+SO and DKH2+SO levels. In the first, a multiconfigurational one-component NR wavefunction was supplemented by SO interactions treated by a post-SCF QDPT method using Orca. 64 The two-electron SO contributions were treated by the mean-field approximation developed by Hess et al. 65 Secondly, in the DKH2 computations, the SR effects were included in the one-component CASSCF calculation, after which QDPT with picture-changed SO operator was carried out. Dynamic correlation effects were included at the NEVPT2 level. The HFC tensors (A) were calculated first at the NR level using hybrid DFT and SO effects treated via leading-order perturbational (Breit-Pauli) SO-corrections. 66 Secondly, fully relativistic four-component mDKS level was used, with hybrid DFT, including both SR and SO effects variationally.

3

Computational Details

Calculations were performed for computationally optimized, eclipsed (as opposed to staggered 67,68 ) geometries of the metallocenes, illustrated in Figure 1. The geometry optimization was carried out with the Turbomole software 69 using the B3LYP 70 level of theory with the def2-TZVP 71 basis set for all metallocenes except CoCp2 (cobaltocene), RhCp2 (rhodocene), and IrCp2 (iridocene). For these three molecules, the def2-TZVP basis set was only used for the light atoms. For the Co, Rh, and Ir centers, the Stuttgart-type energyconsistent, SR effective core potentials ECP10MDF, 72 ECP28MWB, 73 and ECP60MWB, 73 with 10, 28, and 60 electrons in the core, respectively, were used with the appropriate 8s7p6d2f 1g/6s5p3d2f 1g (in the primitive/contracted notation) valence basis sets. 74 The optimized structures for vanadocene, manganocene, cobaltocene, rhodocene, and iridocene are available in the Supporting Information. The optimized structures for nickelocene, chro-

8

ACS Paragon Plus Environment

Page 8 of 42

Page 9 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Chemical Theory and Computation

Figure 1: Eclipsed structure of metallocenes (MCp2 ); in this work, M = Ni, Cr, V, Mn, Co, Rh, and Ir. The pNMR shielding and shielding anisotropy calculations are reported for the 1 H and 13 C nuclei. mocene, and the larger studied Cr(III) complexes were taken from Ref. 12 and that of the Co(II) system from Ref. 13 The last two complexes are depicted in Figure 2.

Figure 2: Structures of the studied (a) Co(II) pyrazolyborate (HPYBCO) and (b) Cr(III) complexes. The groups of 1 H and 13 C nuclei, for which the pNMR shift calculations are reported, are indicated.

9

ACS Paragon Plus Environment

Journal of Chemical Theory and Computation

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 42

The computations of σorb were carried out with Gaussian 09 (G09) 75 at the DFT level with the PBE0 functional using the all-electron def2-TZVP basis set. σorb are calculated non-relativistically, which should not present a drastic approximation for the light nuclei subjected to pNMR measurements, as well as taken the typical dominance of the hyperfine shielding contributions (vide infra). The HFC tensors were calculated, for comparison purposes, both with Orca and ReSpect, at two different levels of theory. The NR+SO level computations (SO interactions treated perturbationally to first order) were carried out with Orca, whereas the fully relativistic mDKS approach was used with ReSpect. The computations of g and D were performed using the ab initio CASSCF and NEVPT2 levels of theory with Orca. 53–55,76 The SO effects were included in these calculations by a QDPT procedure on top of the state-average CASSCF procedure including, in the relevant spin multiplicities, all the roots allowed by the active space (vide infra). 54 This was done at the NR+SO level as well as basing on the one-component reference wavefunction obtained using the SR DKH2 Hamiltonian. A comparison of treating relativity at both these levels is presented for the resulting pNMR shieldings. Similarly, the shielding results obtained with the A tensors computed at the NR+SO and mDKS levels, are compared. The notation used throughout the article for the electron correlation treatment is CASSCF/PBE0 or NEVPT2/PBE0, denoting that either CASSCF or NEVPT2 was used for g and D, whereas PBE0 was used for A. The corresponding notation used for relativistic approximations is NR+SO/NR+SO, DKH+SO/NR+SO, and DKH+SO/mDKS, where always the first method was used for g and D, and the second one for A. The active space in our state-average CASSCF calculations, which also underlay the application of the NEVPT2 method, was chosen as the five valence d-orbitals of the metal ion. For CoCp2 , RhCp2 , and IrCp2 , all with S = 1/2, the seven metal d-electrons were correlated in the Co 3d, Rh 4d, and Ir 5d orbitals, respectively. All the 30 doublet states enabled by the CAS(7,5) specification were included for these doublet systems, to facilitate the QDPT step. Correspondingly, among the S > 1/2 systems, NiCp2 was treated by

10

ACS Paragon Plus Environment

Page 11 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Chemical Theory and Computation

CAS(8,5), CrCp2 by CAS(4,5), VCp2 by CAS(3,5), and MnCp2 by CAS(5,5) (see Table 1). Table 1: Ground-state spin quantum number of the studied systems along with the specification of the active space used to run state-average CASSCF calculations, as well as the number of computed roots therein.

System NiCp2 CrCp2 VCp2 MnCp2 CoCp2 RhCp2 IrCp2 Co(II)b Cr(III)b

Ground state S 1 1 3/2 5/2 1/2 1/2 1/2 3/2 3/2

Correlated electrons 8 4 3 5 7 7 7 7 3

Computed rootsa S − 1 S S+1 15 10 45 35 5 30 10 21 1 30 30 30 40 10 40 10 -

a

The S+1 roots are computed only for CrCp2 because of the limitations posed by the active space and number of correlated electrons. CoCp2 , RhCp2 , and IrCp2 are doublet (S = 1/2) systems and, hence, do not feature ZFS nor necessitate including the S + 1 roots in the calculation. b See text for specification of these complexes.

For the CASSCF and NEVPT2 calculations at the NR+SO level, the balanced def2TZVP basis was used for all the atoms of the metallocenes, whereas a locally dense basis was used for the larger Co(II) and Cr(III) systems. The locally dense basis consisted of the def2-TZVP set applied for the metal ion and the directly bonded atoms, and the def2-SVP basis 71 for the more distant atoms. The DKH-TZV and DKH-SV basis sets 77 were used, instead of the def2 sets, for the corresponding calculations at the DKH+SO level. These sets are DKH recontractions of the def2 sets, designed for relativistic calculations. E.g., in the def2-TZVP basis of Ni, the 17s11p7d1f set of primitive Gaussian functions is contracted to 6s4p4d1f , whereas in the DKH-TZV basis, the 17s11p7d1f primitives are contracted to 10s6p4d1f . The performance of the locally dense basis approach is reliable in calculations of the present type, as seen previously from the results of Refs. 12,13 For A, DFT PBE0 calculations with both the full def2-TZVP basis (results are shown in parentheses in the Tables) and entirely uncontracted DKH-TZV basis on all atoms, were employed with the 11

ACS Paragon Plus Environment

Journal of Chemical Theory and Computation

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 42

NR+SO approximation, while the entirely uncontracted DKH-TZV basis sets were used on all atoms with the mDKS method. As the A tensors are sensitive to both relativistic effects and basis-set quality, we decided to compare the different means of treating relativistic effects for HFC using the same uncontracted DKH-TZV basis. The PBE0 functional is used in the present work as a typical, well-performing representative of the family of hybrid functionals in computations of magnetic properties. 12,13,29 It should be noted that the comparison with experiment can be significantly affected by this choice. The main topic of the present work, the changes due to SR effects on pNMR shieldings, is not expected to be strongly dependent on the choice of this particular functional, however. The total chemical shifts were calculated with respect to the diamagnetic tetramethylsilane molecule, with the geometry optimized as described in Ref. 13 and with the shielding constants σ(13 C) = 188.5 ppm and σ(1 H) = 31.8 ppm at the NR PBE0/def2-TZVP level. The total 13 C and 1 H chemical shifts and shielding anisotropies of all systems were averaged over all the experimentally equivalent nuclei.

4

Results and Discussion

4.1

Adequacy of the basis set

We studied the basis set effect for NiCp2 chosen as the example system, with the relativistic effects included at the mDKS level for A and at the DKH level for g and D. The basis sets used here are DKH-SV, DKH-TZV, and Dyall-CVQZ⋆ 78 sets, where the



denotes that the

Dyall-CVQZ set was used only for the Ni atom, while the DKH-QZV 71 basis was used for the C and H atoms. Tables 2 and 3 contain the data for the basis-set dependence of the calculated EPR parameters of NiCp2 . The g- and ZFS tensors (Table 2) indicate a rather modest sensitivity to the basis-set quality, whereas the HFC tensors (Table 3) show a more pronounced dependence. Therefore, the behavior of the pNMR shielding tensors with the different basis sets follows closely that 12

ACS Paragon Plus Environment

Page 13 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Chemical Theory and Computation

Table 2: Calculated isotropic g-factor and the eigenvalues of the g-tensor, as well as the D (in units of cm−1 ) parameter of ZFS at the CASSCF and NEVPT2 levels for NiCp2 , using the DKH-SV, DKH-TZV, and Dyall-CQVZ⋆ basis sets at the DKH+SO level.

Parameters Number of basis functionsb D giso g-eigenvalues

CASSCF DKH-SV DKH-TZV Dyall-CQVZ⋆a 231 435 1130 66.451 67.901 68.871 2.246 2.248 2.250 1.990 1.990 1.990 2.373 2.377 2.380 2.373 2.377 2.380

NEVPT2 DKH-SV DKH-TZV Dyall-CQVZ⋆a 231 435 1130 38.688 39.340 39.797 2.125 2.125 2.123 1.997 1.997 1.997 2.189 2.188 2.186 2.189 2.188 2.186

a

The Dyall-CVQZ basis set was used for the Ni atom, while for the C and H atoms, the DKH-QZV basis was used. b Number of contracted basis functions in the different basis sets.

Table 3: Calculated 13 C and 1 H hyperfine coupling constants and coupling anisotropies (in MHz) for NiCp2 , using the uncontracted DKH-SV, DKH-TZV, and Dyall-CQVZ⋆ basis sets at the DFT (PBE0) level in the mDKS approach. Averages over all the 10 experimentally equivalent carbon and hydrogen nuclei are given. Parametersa Number of basis functionsc A(13 C) (isotropic) ∆A(13 C) A(1 H) (isotropic) ∆A(1 H)

DKH-SV DKH-TZV Dyall-CQVZ⋆b 383 632 1310 4.19 5.07 5.18 11.36 11.82 11.97 -3.018 -3.361 -3.511 0.018 0.222 0.274

a

The hyperfine coupling anisotropy ∆A is calculated as ∆A = Azz − (Axx + Ayy )/2, where the z axis is chosen along the axis of the effective cylindrical symmetry of the molecule. b See footnote a in Table 2. c Number of uncontracted basis functions in the different basis sets.

of the A tensors. It can be conjectured that, while the triple zeta-level basis set is not entirely converged, it represents a sufficient level for a qualitative consideration of the influence of SR effects on the pNMR shielding tensor. The deficiencies of the basis set should not dominate over the other systematic errors, such as those due to our choice of the CASSCF active space or the DFT functional.

13

ACS Paragon Plus Environment

Journal of Chemical Theory and Computation

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

4.2

Relativistic effects on g and D

The calculated isotropic g-values, the eigenvalues of the g-tensor, as well as the D (and E) parameters of ZFS are listed in Table 4 for the S > 1/2 metallocenes, as well as the two larger complexes. It is seen that introducing scalar relativity using the DKH Hamiltonian mostly decreases the magnitude of the D-parameter very slightly, to much smaller degree than that of dynamic correlation when changing from CASSCF to NEVPT2. The g-tensor data are also very little affected by adopting the DKH Hamiltonian. Overall, no major improvement can be found in the agreement with experimental ZFS or g-tensor data due to adopting scalar relativity. The fact that the SR influences are small on g and D, suggests that DKH2 may be quite sufficient for them, at least for the 3d systems. The isotropic g-factor and the eigenvalues of the g-tensor for the S = 1/2 cobaltocene, rhodocene, and iridocene, are given in Table 5. CoCp2 shows similar changes due to SR effects as the other 3d metallocenes. The heavier 4d complex RhCp2 does not feature more marked changes due to scalar relativity as compared to CoCp2 . Expectedly, IrCp2 shows the largest dependence on the treatment of SR effects. Also a clear influence of the dynamic electron correlation treatment is observed. For the A tensors in the present metallocenes, the effects of relativity, both SR and SO effects beyond the leading order, are indirectly visible from the present chemical shift and shielding anisotropy Tables 6–9 below, in the changes of the shielding constants and anisotropies corresponding to the change of the method of calculating A from the NR+SO to the mDKS level. Due to the fairly modest dependence of g and D on the treatment of relativity, the results for δ and ∆σ (vide infra) reflect rather faithfully the large variations of A for practically all the presently studied systems (concerning the group-IX metallocenes, see Table S7 in SI).

14

ACS Paragon Plus Environment

Page 14 of 42

Page 15 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Chemical Theory and Computation

Table 4: Calculated D- and E/D-parametersa of the zero-field splitting tensor, as well as the isotropic g-factor and the eigenvalues of the g-tensor, for the present S > 1/2 metallocenes and two larger paramagnetic complexes. CASSCF and NEVPT2 calculations using the def2TZVP and DKH-TZV basis sets, with nonrelativistic (NR+SO) and relativistic (DKH+SO) wavefunctions, respectively.

NiCp2 b Parameter D (cm−1 ) E/D giso g-eigenvalues

Parameter D (cm−1 ) E/D giso g-eigenvalues

Parameter D (cm−1 ) E/D giso g-eigenvalues

CASSCF NEVPT2 NR+SO DKH+SO NR+SO DKH+SO 70.1 67.9 40.1 39.3 0.000 0.000 0.000 0.000 2.253 2.248 2.125 2.125 1.989 1.990 1.997 1.997 2.385 2.377 2.189 2.188 2.385 2.377 2.189 2.188 VCp2 d CASSCF NEVPT2 NR+SO DKH+SO NR+SO DKH+SO 2.4 2.3 2.5 2.4 0.000 0.000 0.000 0.000 1.979 1.980 1.984 1.985 1.968 1.969 1.976 1.977 1.968 1.969 1.976 1.977 2.000 2.000 2.000 2.000 Co(II) pyrazolylborate complex CASSCF NEVPT2 NR+SO DKH+SO NR+SO DKH+SO -111.3 -109.9 -113.4 -111.8 0.201 0.192 0.155 0.148 2.111 2.110 2.113 2.113 1.446 1.450 1.473 1.478 1.685 1.680 1.663 1.659 3.200 3.200 3.204 3.202

CrCp2 c CASSCF NEVPT2 NR+SO DKH+SO NR+SO DKH+SO -24.5 -24.7 -11.3 -10.4 0.125 0.130 0.002 0.002 1.635 1.628 2.070 2.066 1.446 1.444 1.964 1.965 1.686 1.675 1.964 1.966 1.773 1.765 2.280 2.267 MnCp2 e CASSCF NEVPT2 NR+SO DKH+SO NR+SO DKH+SO 0.8 -0.1 -0.2 -0.2 0.087 0.182 0.226 0.212 1.997 2.000 1.995 2.000 1.993 2.000 1.987 2.000 1.998 2.000 1.998 2.000 2.001 2.000 1.999 2.000 Cr(III) cyclopentadienyl complex CASSCF NEVPT2 NR+SO DKH+SO NR+SO DKH+SO 3.0 2.9 2.8 2.8 0.047 0.046 0.032 0.032 1.958 1.959 1.970 1.970 1.944 1.945 1.959 1.959 1.952 1.952 1.965 1.965 1.979 1.980 1.985 1.985

a

For the definitions of the D- and E-parameters see, e.g., Table S5 of Ref. 13 b Experimental D parameters in the range 25.6 ± 3.0 . . . 33.6 ± 0.3 cm−1 . 79–83 Experimental g-tensor data from susceptibility measurements: gk = 2.00, as well as g⊥ = 2.06 ± 0.10 79 and 2.11 ± 0.03. 81 c Experimental D parameter from susceptibility measurement −15.1 cm−1 . 84 Experimental EPR data: gk = 2.012, g⊥ = 1.988. 85 d Experimental D parameter 2.8 cm−1 . 86 Experimental isotropic g-value is 2. 86 e Experimental D parameter −0.24 . . . − 0.2 cm−1 . 87 Experimental isotropic g-value is 2. 87

4.3 4.3.1

3d metallocenes with ZFS Isotropic shifts

The calculated isotropic carbon and proton chemical shifts for the present S > 1/2 metallocenes at 298 K are listed in Table 6. Experimental 15

13

C and 1 H pNMR shift data for these

ACS Paragon Plus Environment

Journal of Chemical Theory and Computation

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 42

Table 5: Calculated isotropic g-factor and the eigenvalues of the g-tensor for the doublet CoCp2 , RhCp2 , and IrCp2 systems. CASSCF and NEVPT2 calculations carried out using the def2-TZVP and DKH-TZV basis, with nonrelativistic (NR+SO) and scalar relativistic (DKH+SO) wavefunctions. CoCp2 a Parameter giso g-eigenvalues

Parameter giso g-eigenvalues

Parameter giso g-eigenvalues

CASSCF NEVPT2 NR+SO DKH+SO NR+SO DKH+SO 1.999 2.010 1.979 1.980 1.850 1.867 1.835 1.829 2.050 2.050 2.037 2.035 2.095 2.112 2.065 2.075 b RhCp2 CASSCF NEVPT2 NR+SO DKH+SO NR+SO DKH+SO 1.842 1.837 1.827 1.829 1.443 1.442 1.461 1.469 1.923 1.919 1.909 1.912 2.161 2.151 2.111 2.107 c IrCp2 CASSCF NEVPT2 NR+SO DKH+SO NR+SO DKH+SO 1.472 1.613 1.532 1.612 0.854 0.965 0.958 1.050 1.415 1.547 1.529 1.661 2.145 2.327 2.109 2.124

a

Experimental EPR data.: 88 g11 = 1.940, g22 = 1.161, g33 = 1.377, giso = 1.493 in NiCp2 host lattice. g11 = 1.585, g22 = 1.140, g33 = 1.219, giso = 1.315 in RuCp2 host lattice. g11 = 1.69, g22 = 1.81, g33 = 1.81, giso = 1.77 in FeCp2 host lattice. b Experimental EPR data: gk = 2.033, g⊥ = 2.003. 89 c Experimental EPR data: gk = 2.033, g⊥ = 2.001. 89

metallocenes have been reported in Refs. 86,87,90–96 At the CASSCF level, the isotropic

13

C

shifts show very modest changes of 1 to 3 ppm and 1 H shifts practically no change (≤ 1 ppm) due to the inclusion of SR effects for g and D. However, with fully relativistic calculations of A, much larger changes of around 10 ppm or more are found in the

13

C shifts of CrCp2 ,

VCp2 , and MnCp2 , as well as changes of up to 10 ppm in the 1 H shifts of NiCp2 , VCp2 , and MnCp2 , when all HFCs are computed with the same uncontracted DKH-TZV basis

16

ACS Paragon Plus Environment

Page 17 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Chemical Theory and Computation

set to facilitate comparison. At the NEVPT2 level, essentially the same findings are made concerning relativistic approximations as at the CASSCF level. Whereas the effects of relativity are similar, somewhat large dynamical correlation effects are seen with NEVPT2. Our best calculations (g and D with DKH+SO; A with mDKS) at the NEVPT2 level lead to a generally rather good agreement of the calculated carbon and proton shifts with experiment, although the situation is not dramatically affected by the improved treatment of relativity, for these 3d metallocenes. In the case of MnCp2 , however, a 10 ppm change in the proton shifts is obtained due to employing the fully relativistic (mDKS) HFC. Some remaining difference can be expected even in the most favorable comparison of experiments performed in the condensed phase (solution or solid state) at a finite temperature, with quantum-chemical calculations in vacuo, without considering rovibrational effects. Table 6: Computed isotropic 13 C and 1 H chemical shifts (in ppm, with respect to TMS) for S > 1/2 metallocenes at 298 K (unless otherwise noted). Notation: Method for g and D/method for A. The CASSCF/PBE0 and NEVPT2/PBE0 levels are used with the def2TZVP basis set for NR+SO calculations and DKH-TZV basis set for DKH+SO and mDKS calculations. The uncontracted DKH-TZV basis set is used for all the HFC calculations, except for the results shown in parentheses, where the def2-TZVP basis set was used.

Nucleus/System 13 C NiCp2 CrCp2 VCp2 MnCp2

NR+SO/NR+SO (1332.2) (-154.1) (-443.5) (3621.1) (2792.7)

1421.3 -159.3 -414.6 1865.9 1452.2

CASSCF/PBE0 DKH+SO/NR+SO (1335.9) (-153.0) (-443.7) (3623.4) (2795.4)

DKH+SO/mDKS

NR+SO/NR+SO

1424.9 -158.2 -414.8 1865.6 1452.2

1424.2 -168.4 -426.0 1866.1 1452.6

(1373.5) (-245.0) (-445.8) (3611.1) (2785.6)

1459.2 -253.9 -416.8 1857.5 1453.1

(-280.0) -269.0 (251.6) 246.3 (335.1) 328.9 (208.2) -2.9 (160.6) -0.7

-262.5 248.4 332.6 7.4 7.1

(-264.1) -253.6 (323.3) 316.6 (335.7) 329.5 (207.3) -3.2 (159.9) -0.6

NEVPT2/PBE0 DKH+SO/NR+SO

DKH+SO/mDKS

Exp.

Earlier computational

1460.6 -252.7 -417.0 1867.2 1453.1

1462.7 -258.4 -428.3 1867.6 1453.5

1514,a 1715b -250,e -229f -510,g -422g 1187h 1206i

1056.8...1503.7,c 1276.26...1401.24d -226.2...-62.2,c -215.88...-102.00d -433.3...-218.9,c -429.17...-287.57d 1734.9...2084.8,c 1728.18...1820.40d

(-264.0) -253.6 (322.6) 315.9 (335.8) 329.6 (208.4) -2.7 (160.7) -0.6

-247.3 318.7 333.3 7.5 7.2

-253bj 320f j 318,g 307g -11h -23.3i

-254.3...-238.5,c -258.24...-242.62d 312.3...336.8,c 311.30...336.34d 339.9...362.8,c 336.33...357.69d -15.4...-4.1,c 5.18...12.12d

(1374.9) (-244.0) (-446.0) (3624.8) (2796.1)

1

H NiCp2 CrCp2 VCp2 MnCp2

(-280.7) -269.6 (252.6) 247.3 (335.0) 328.8 (208.3) -2.6 (160.6) -0.7

a

At room temperature. Solution state. 90 b At room temperature. Solid state. 91 c A range of DFT results obtained using different exchange-correlation functionals: BP86, B3LYP, B3PW91, and BPW91-30HF. 50 d A range of DFT results obtained using the PBE and PBE0 functionals. 48 e At 298 K. Solution state. 92 f At 298 K. Solid state. 93 g At room temperature. Ref. 86 h Ref. 91 Mean shift value of terminal and bridging ligands of the polymeric (non-sandwich) compound. i At 390 K. Solution state. Ref. 87 j At 298 K. Solution state. 94

In the case of nickelocene, the CASSCF and NEVPT2 results show a small relativistic effect on the isotropic carbon and proton chemical shifts. There is a difference of a couple 17

ACS Paragon Plus Environment

Journal of Chemical Theory and Computation

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

of ppm in the

13

Page 18 of 42

C shift when g and D are calculated with the DKH+SO approximation, as

compared to using the NR wavefunction. The 1 H shift indicates an even smaller difference due to the DKH-based treatment of g and D. However, the variation of 1 H shifts is large when we use the relativistic four-component mDKS approximation for A instead of the NR+SO level, where the only relativistic influence is via the perturbational treatment of SO effects. At the NEVPT2 level, the difference in the isotropic 13 C shift is about 3.5 ppm, while the 1 H shift shows a change of about -6.3 ppm, when SR effects are included both for g, D (DKH+SO method), and A (mDKS), as compared to the basic NR+SO level. The isotropic 13

C and 1 H shifts also show the significant 38.5 ppm and -15.2 ppm differences, respectively,

due to improving the correlation treatment from the CASSCF to the NEVPT2 wavefunction level, using our best combination of relativistic methods (g and D with DKH+SO; A with mDKS). The best calculations using NEVPT2 give the isotropic 13 C shift of 1462.7 ppm and 1

H shift of -247.3 ppm, in a fairly good agreement with experimental values in the range

from 1514 to 1715 ppm and -253 ppm, respectively. For chromocene, the

13

C and 1 H shifts show a negligible difference, when g and D are

calculated at the DKH+SO level as compared to using a NR wavefunction (NR+SO). Again, the variation of 13 C and 1 H shifts is larger if we use the mDKS approximation for A instead of NR+SO. At NEVPT2, the difference in the 13 C isotropic shift is about -4.5 ppm, while the 1 H shift shows a change of +2.1 ppm with g and D using DKH and A with mDKS, as compared to the NR+SO/NR+SO level. Much larger, 90 and 70 ppm differences are seen in 1

13

C and

H shifts between the CASSCF and NEVPT2 wavefunctions, with our DKH+SO/mDKS

combination of methods for g and D/A. The relativistically calculated NEVPT2 shifts at the DKH+SO/mDKS level, -258.4 and 318.7 ppm, are in an excellent agreement with the experimental shifts of -250...-229 and 320 ppm for

13

C and 1 H, respectively.

Similar findings concerning the effects of the relativistic approximations are found for vanadocene VCp2 (S=3/2). Negligible differences result in the chemical shifts when g and D are calculated with DKH+SO as compared to the NR+SO level. Adopting the fully

18

ACS Paragon Plus Environment

Page 19 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Chemical Theory and Computation

relativistic mDKS HFCs leads, then, to differences in the isotropic 13 C and 1 H shifts of -11.5 and 3.8 ppm. For this system, the effect of the choice of the wavefunction (either CASSCF or NEVPT2) in the calculation of the g and D tensors, is very small on the chemical shifts. Our best calculation, DKH+SO/mDKS at the NEVPT2 level, gives the isotropic 1

13

C and

H shifts of -428.3 and 333.3 ppm, in a good agreement with the experimental values at

-510...-422 ppm and 307...318 ppm, respectively. In the case of manganocene, MnCp2 (S=5/2), somewhat larger deviations are found from experimental isotropic

13

C shifts at 298 K even at our best level of computation. The data

corresponding to measurements at 390 K are in a much better agreement, however. At the lower temperature (298 K), the larger difference in the shifts is likely to be due to the fact that the experiment 91 was carried out for a polymeric instead of a single-molecular species. The overall effects on chemical shifts due to the incorporation of the present relativistic approximations are the same as seen for previous systems. Our best DKH+SO/mDKS calculation at the NEVPT2 level gives the shifts of 1867.6 and 7.5 ppm at 298 K, while the experimental values are at 1187 ppm and -11 ppm for

13

C and 1 H, respectively. It is

seen that MnCp2 shows a large basis-set dependence, as major differences are seen between the chemical shifts when the A tensors are calculated with the contracted def2-TZVP and the uncontracted DKH-TZV basis set. The smaller def2-TZVP basis gives clearly erroneous HFCs for this system. 4.3.2

Group-IX doublet metallocenes

The isotropic carbon and proton chemical shifts for the present group-IX metallocenes, i.e., CoCp2 , RhCp2 , and IrCp2 are given in Table 7. The overall variation in the calculated chemical shift values can be seen in Fig. 3. In the cases of CoCp2 and RhCp2 , changing the method for g from NR+SO to DKH+SO produces small changes at both the CASSCF and NEVPT2 levels, similarly as seen for the S > 1/2 metallocenes in their isotropic chemical shifts. Qualitatively similar results are found in the

19

13

C and 1 H shifts of the doublet

ACS Paragon Plus Environment

Journal of Chemical Theory and Computation

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 42

metallocenes both using the CASSCF and NEVPT2 wave functions. For cobaltocene, the fully relativistic mDKS HFCs bring the isotropic carbon chemical shift into an excellent agreement with experiment (614 ppm) at both the CASSCF and NEVPT2 levels, with the results 630 and 622 ppm, respectively. A few ppm changes in the proton shift due to the SR calculations appear for CoCp2 and RhCp2 . Table 7: Computed isotropic 13 C and 1 H chemical shifts (in ppm, with respect to TMS) for S = 1/2 metallocenes at 298 K. Notation: Method for g/method for A. The CASSCF/PBE0 and NEVPT2/PBE0 levels used with the def2-TZVP basis set for NR+SO calculations and DKH-TZV basis set for DKH+SO and mDKS calculations. The uncontracted DKH-TZV basis set is used for all the HFC calculations, except for the results shown in parentheses, where the def2-TZVP basis set was used.

Nucleus/System 13 C CoCp2 RhCp2 IrCp2 1 H CoCp2 RhCp2 IrCp2

NR+SO/NR+SO

CASSCF/PBE0 DKH+SO/NR+SO

DKH+SO/mDKS

NEVPT2/PBE0 NR+SO/NR+SO DKH+SO/NR+SO

DKH+SO/mDKS

Exp.

Earlier computational

a

545...681.1,b 541.2...618c 562.9,d 601.4e 685.9e -59...-34.2,b -56.2...-54.7c -87.7,d -84.2e -78.7e

(639.5) 629.9 (641.8) 678.1 (289.7) 329.3

(643.1) 633.5 (641.2) 677.4 (331.2) 374.1

630.4 681.3 359.7

(634.1) 624.6 (643.4) 679.4 (319.5) 360.6

(633.8) 624.3 (645.3) 681.3 (349.2) 392.2

621.6 685.6 379.7

614 -

(-37.5) -34.5 (-90.1) -86.3 (-92.0) -88.0

(-37.8) -34.7 (-89.9) -86.0 (-100.7) -96.4

-35.9 -82.1 -83.2

(-37.1) -34.1 (-89.2) -85.3 (-94.9) -90.8

(-37.1) -34.1 (-89.3) -85.4 (-99.3) -95.0

-35.3 -81.5 -80.8

-51a -

a

At 298 K. Solution State. 95 b A range of DFT results obtained using different exchange-correlation functionals: BP86, B3LYP, B3PW91, and BPW91-30HF. 50 c A range of DFT results obtained using the PBE and PBE0 functionals. 48 d All-electron DFT calculations using the PBE functional in Ref. 49 e Relativistic pseudopotential calculations using the PBE functional in Ref. 49

In the cases of CoCp2 and RhCp2 , the comparison of the nonrelativistically based (NR+SO) calculations with DKH+SO for g shows only small SR effects conveyed to isotropic chemical shifts. Slightly larger, a few ppm relativistic changes follow when the HFCs are calculated with the mDKS method, as compared to the NR+SO level. However, the incorporation of SR effects becomes indispensable when reaching the last (5d) transition element row in the periodic table of elements. The findings point to a genuine, albeit modest need to include SR effects in pNMR shielding calculations already for 3d and 4d systems, 11 for NMR nuclei that reside in the immediate vicinity of the paramagnetic center such as in the case of the metallocenes. The results in parentheses in Tables 6 and 7 obtained with the def2-TZVP basis indicate, however, that it is for 3d and 4d systems even more important to take care 20

ACS Paragon Plus Environment

Page 21 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Chemical Theory and Computation

Figure 3: Calculated isotropic chemical shifts (in ppm, with respect to the TMS reference compound, at 298 K) for the group-IX doublet metallocenes using various approximations of relativistic influences: (a) carbon and (b) proton. Notation: Method for g/method for A. NR denotes nonrelativistic one-component wavefunction and DKH+SO the secondorder Douglas-Kroll-Hess Hamiltonian supplemented with spin-orbit coupling effects. mDKS denotes a fully relativistic calculation of A. NEVPT2/PBE0 level used. See text for details. of sufficient flexibility of the basis set. Furthermore, intermolecular interaction effects are likely to be equally important as SR influences, for such systems. In contrast to the lighter doublet metallocenes studied presently, the inclusion of SR effects for g at the DKH+SO level already produces marked influence on the

13

C and 1 H

shifts in the 5d iridocene system. Furthermore, SR effects on the HFC also change the 13

C and 1 H pNMR shifts significantly in this 5d system. The difference of the relativistic

results at the DKH+SO/mDKS level (for g/A) from the NR-based shifts amounts to 30-40 ppm for carbon shifts and about 10-15 ppm for proton shifts. The calculated

13

C and 1 H

isotropic shifts are also comparable with earlier computations. 48,49 The comparison of our best calculation with the earlier obtained CoCp2 and RhCp2 , but the

13

13

C shifts shows small differences in the cases of

C shifts in IrCp2 shows a large difference of about 300 ppm

with the relativistic effective core potential (ECP) calculation of Ref. 49 This suggests that the relativistic ECP calculations may not be sufficient for the

13

C shifts of IrCp2 . The

comparison of isotropic 1 H shifts of the doublet metallocenes with earlier ECP computations only shows small differences of few ppm, indicating that the ECP method may indeed be a viable one for the proton shifts.

21

ACS Paragon Plus Environment

Journal of Chemical Theory and Computation

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

4.3.3 The

13

Page 22 of 42

Scalar relativistic effects on the shielding anisotropy C and 1 H shielding anisotropies for the S > 1/2 metallocenes are listed in Table 8.

The calculated

13

C shielding anisotropies are negative for these metallocenes except for

CrCp2 . They show variations of up to a few ppm units, when we switch from NR+SO to the DKH+SO method for g and D. The relativistic four-component mDKS HFCs have a clearly larger effect on the 13 C anisotropy for all S > 1/2 metallocenes at both the CASSCF and NEVPT2 levels. Table 8: Calculated 13 C and 1 H shielding anisotropies (in ppm) for S > 1/2 metallocenes at 298 K. Notation: Method for g and D/method for A. The CASSCF/PBE0 and NEVPT2/PBE0 levels used with the def2-TZVP basis set for NR+SO calculations and DKH-TZV basis set for DKH+SO and mDKS calculations. The uncontracted DKH-TZV basis set is used for all the HFC calculations, except for the results shown in parentheses, where the def2-TZVP basis set was used.

Nucleus/System 13 C NiCp2 CrCp2 VCp2 MnCp2 1 H NiCp2 CrCp2 VCp2 MnCp2

NR+SO/NR+SO

CASSCF/PBE0 DKH+SO/NR+SO

DKH+SO/mDKS

NR+SO/NR+SO

NEVPT2/PBE0 DKH+SO/NR+SO

DKH+SO/mDKS

(-2683.0) (161.0) (-659.4) (-3961.8)

-2700.3 184.2 -633.9 -4252.0

(-2699.0) (162.1) (-659.7) (-3952.4)

-2717.2 185.2 -634.2 -4248.0

-2697.8 182.3 -665.9 -4157.0

(-2914.7) (325.0) (-662.9) (-3923.8)

-2946.7 360.1 -637.3 -4224.8

(-2919.3) (320.7) (-663.2) (-3960.7)

-2951.6 355.5 -637.5 -4253.3

-2926.7 342.1 -669.3 -4162.3

(-111.7) (-22.8) (-210.9) (-462.9)

-108.1 -23.1 -210.3 -426.8

(-109.3) (-23.9) (-210.9) (-462.6)

-105.8 -24.2 -210.4 -427.0

-104.7 -23.7 -210.6 -430.4

(-66.6) (-127.7) (-209.6) (-460.0)

-64.8 -126.1 -209.1 -425.4

(-66.1) (-124.5) (-209.7) (-463.1)

-64.3 -123.0 -209.2 -427.2

-63.9 -123.5 -209.3 -430.7

Exp. a

-2855 -831d -

Earlier computational -4030.2...-3116.4,b 119.5...371.4,b -904.7...-567.1,b -5233.2...-4304.4,b

-3901.47...-2509.73c 148.35...299.65c -899.42...-645.05c -5074.53...-4238.57c

a

At room temperature. Solid state. 91 b A range of DFT results obtained using different exchange-correlation functionals: BP86, B3LYP, B3PW91, and BPW91-30HF. 50 c A range of DFT results obtained using the PBE and PBE0 functionals. 48 d At room temperature. Ref. 86

Upon the switch from the CASSCF level to NEVPT2, the absolute value of the anisotropy increases for NiCp2 and CrCp2 , while for VCp2 and MnCp2 , the influence of dynamical correlation is not large. The SR effects on the anisotropic HFCs are equally large, if not larger than on their isotropic counterparts, and the two dominate the isotropic and anisotropic parts of nuclear shielding, respectively. Somewhat larger changes in the carbon shielding anisotropy are found than in the isotropic carbon shift due to incorporation of scalar relativity. For

13

C anisotropies at the NEVPT2 level, the effect of SR g and D is negligible

for VCp2 , while NiCp2 and CrCp2 show less than 5 ppm changes in the absolute value, and 22

ACS Paragon Plus Environment

Page 23 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Chemical Theory and Computation

MnCp2 features a much larger change of about 28 ppm. All 3d metallocenes show large changes of up to 90 ppm in the

13

C shielding anisotropies due to the mDKS HFCs. The

NiCp2 carbon anisotropy is in a very good agreement and VCp2 in a reasonable agreement with experiment, as seen from Table 8. However, reaching that level of agreement is not significantly affected by incorporating scalar relativity in the calculations. For 1 H anisotropies, minor changes up to 3 ppm appear due to adopting SR Hamiltonian for g and D. With mDKS HFCs, at both the CASSCF and NEVPT2 levels, the 1 H anisotropies are generally negligibly affected, except for the 3.5 ppm increase in the absolute value of ∆σ(1 H) in MnCp2 . The

13

C and 1 H shielding anisotropies for doublet (S = 1/2), group-IX metallocenes are

listed in Table 9. The outcome of the different levels of incorporating relativistic effects are shown in Fig. 4. The calculated proton and carbon shielding anisotropies for CoCp2 show changes of a few ppm due to a SR calculation of g at both the CASSCF and NEVPT2 levels. Fully relativistic HFCs at the mDKS level produce a modest 6 ppm change for ∆σ(13 C) in CoCp2 . The proton shielding anisotropies in CoCp2 are hardly affected by the incorporation of SR effects. For RhCp2 , a substantial change of about 17 ppm in carbon anisotropy occurs at both the CASSCF and NEVPT2 levels, due to adopting relativistic mDKS HFCs. The corresponding, non-negligible change in the proton shielding anisotropy amounts to about 5 ppm. The incorporation of scalar relativity does not improve agreement of the CoCp2 shielding anisotropy with experiment, 91 and the remaining difference remains large, about 187 ppm. For IrCp2 , a large variation is seen for the carbon shielding anisotropy already from calculating g using a SR Hamiltonian. Also, the inclusion of scalar relativity for A is more significant and increases the absolute value of carbon shielding anisotropies by about 60 ppm and decreases the absolute value of proton shielding anisotropies by circa 20 ppm at both the CASSCF and NEVPT2 levels. Our best calculation at the NEVPT2 level, with the inclusion of SR effects with the

23

ACS Paragon Plus Environment

Journal of Chemical Theory and Computation

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 24 of 42

Table 9: Calculated 13 C and 1 H shielding anisotropies (in ppm) for S = 1/2, group-IX metallocenes at 298 K. Notation: Method for g/method for A. The CASSCF/PBE0 and NEVPT2/PBE0 levels used with the def2-TZVP basis set for NR+SO calculations and DKHTZV basis set for DKH+SO and mDKS calculations. The uncontracted DKH-TZV basis set is used for all the HFC calculations, except for the results shown in parentheses, where the def2-TZVP basis set was used.

Nucleus/System 13 C CoCp2 RhCp2 IrCp2 1 H CoCp2 RhCp2 IrCp2

a c

NR+SO/NR+SO

CASSCF/PBE0 DKH+SO/NR+SO

DKH+SO/mDKS

NEVPT2/PBE0 NR+SO/NR+SO DKH+SO/NR+SO

(-262.7) -266.5 (-597.2) -591.6 (-466.7) -450.0

(-268.3) -272.0 (-598.0) -592.5 (-531.7) -514.5

-266.4 -609.7 -579.5

(-260.7) -264.5 (-614.1) -609.5 (-542.0) -527.3

(-19.8) -19.4 (-38.4) -37.1 (-72.9) -69.6

(-19.8) -19.4 (-38.0) -36.7 (-76.6) -73.1

-19.1 -31.5 -50.0

(-19.5) -19.2 (-35.6) -34.5 (-68.4) -65.3

DKH+SO/mDKS

Exp.

Earlier computational

(-257.7) -261.4 (-619.9) -615.5 (-603.2) -589.3

-255.9 -632.0 -649.5

-443a -

-659.0b -995.4,b -964.5c -970.6c

(-19.8) -19.4 (-35.2) -34.0 (-67.2) -64.1

-19.0 -28.9 -39.7

-

-11.9b -27.9,b -23.1c -

At room temperature. Solid state. 91 b All-electron DFT calculations using the PBE functional in Ref. 49 Relativistic pseudopotential calculations using DFT with the PBE functional in Ref. 49

Figure 4: Calculated shielding anisotropies (in ppm, at 298 K) for the group-IX doublet metallocenes at the NEVPT2/PBE0 level with the different relativistic approximations for g/A: (a) carbon and (b) proton. See text for details. DKH+SO/mDKS combination of methods, gives the

13

C shielding anisotropies of (-255.9,

-632.0, and -649.5 ppm) for CoCp2 , RhCp2 , and IrCp2 , respectively. The calculated carbon shielding anisotropies are quite different from the earlier computational data, 49 i.e., -659.0, -995.4...-964.5, and -970.6 ppm for CoCp2 , RhCp2 , and IrCp2 , respectively. These differences are mainly due to different exchange-correlation functionals (the earlier work 49 used the non-hybrid PBE functional). Additionally, the earlier 49 SR pseudopotential treatment may not be enough for carbon shielding anisotropies, based on the comparison of results for

24

ACS Paragon Plus Environment

Page 25 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Chemical Theory and Computation

RhCp2 and IrCp2 . The 1 H shielding anisotropies of -19.0, -28.9, and -39.7 ppm, respectively, were obtained for CoCp2 , RhCp2 , and IrCp2 , respectively, at our best level. These results agree qualitatively with the earlier computational (at either NR-based all-electron or pseudopotential levels) shielding anisotropies, suggesting that the pseudopotential approximation may be quite sufficient for the SR effects in proton shielding anisotropies in these doublet metallocenes. 4.3.4

Scalar relativistic effects on the physical contributions to shielding

The different physical contributions to the total isotropic

13

C and 1 H nuclear shielding con-

stants for NiCp2 and IrCp2 , chosen as example systems, are illustrated in Figures 5 and 6, respectively. The corresponding illustrations for the shielding anisotropies are in Figs. S1 and S2 in the Supporting Information. Numerical values for the contributions both to the isotropic shielding constants and shielding anisotropies are listed in Tables S8-S15 in the SI. In the case of NiCp2 , the largest contributions to both

13

C and 1 H shielding constants arise

from the contact term (term 1 in Table S1). Other, somewhat large contributions to the 13 C shielding constant are given by the dipolar term (term 2), the modification of the contact term on account of the g-shift (term 6) and the pseudocontact term (term 9). For NiCp2 , the main relativistic influence arises from terms 1, 6, and 9 to the

13

C, and from term 1 to

the 1 H shielding constant. Similarly for IrCp2 , the largest contributions to both

13

C and 1 H

shielding constants arise from the contact term (term 1). Other large contributions to both the

13

C and 1 H shielding constants come from terms 3, 6, and 9. Major relativistic changes

from the NR+SO to the mDKS treatment of the A arise from the big terms 1, 6, and 9.

4.4

Proton shifts of the larger Co(II) and Cr(III) complexes

The notation of the groups of 1 H nuclei, for which the pNMR shifts of the two studied quartet (S = 3/2) Co(II) and Cr(III) complexes are reported, is illustrated in Fig. 2. The computed isotropic proton chemical shifts for the Co(II) pyrazolylborate complex (HPYBCO), are 25

ACS Paragon Plus Environment

Journal of Chemical Theory and Computation

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 5: Physical contributions to the total calculated isotropic nuclear shielding constants and the total chemical shifts (in ppm, shifts with respect to TMS, at 298 K) for NiCp2 : (a) carbon and (b) proton. The numbering of the hyperfine terms refers to Table S1 in the Supporting Information. The shielding tensors are calculated at the NEVPT2/PBE0 level using the DKH+SO/mDKS approximation (for g and D/A, respectively), except that the orbital part is calculated at the NR level with DFT(PBE0). The uncontracted DKH-TZV basis is used for all the HFCs.

Figure 6: Physical contributions to the total calculated isotropic nuclear shielding constants and the total chemical shifts (in ppm, shifts with respect to TMS, at 298 K) for IrCp2 : (a) carbon and (b) proton. The numbering of the hyperfine terms refers to Table S1 in the Supporting Information. The shielding tensors are calculated at the NEVPT2/PBE0 level using the DKH+SO/mDKS approximation (for g/A, respectively), except that the orbital part is calculated at the NR level with DFT(PBE0). The uncontracted DKH-TZV basis is used for all the HFCs. given in Table 10. Both at the CASSCF and NEVPT2 levels, the DKH+SO approximation for g and D causes negligible changes of up to 0.5 ppm and 0.8 ppm, respectively, in the chemical shift for nucleus 3H. However, the mDKS approximation for A induces appreciable relativistic effects, causing change in the chemical shift of the 3H nucleus by -6 ppm and, for

26

ACS Paragon Plus Environment

Page 26 of 42

Page 27 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Chemical Theory and Computation

the 4H and 5H nuclei, changes of 1 ppm and, for BH, 4.5 ppm, at the CASSCF level. The correlation effects at NEVPT2 level are a couple of ppm in the chemical shift of 3H and BH nuclei. The relativistic effects slightly deteriorate the agreement of the data obtained for 4H, 5H, and BH nuclei with the solution-state experiment. As already stated, this can be due to the fact that the solvent and rovibrational effects influence the experimental data, whereas these factors are not taken into account in our modelling. However, for the 3H nucleus, relativity improves the agreement with experiment. Table 10: Calculated isotropic 1 H chemical shifts (in ppm) for pyrazolylborate Co(II) complex at 298 K. Notation: Method for g and D/method for A. The CASSCF/PBE0 and NEVPT2/PBE0 levels used with the def2-TZVP basis set for NR+SO calculations and DKHTZV basis set for DKH+SO and mDKS calculations. The uncontracted DKH-TZV basis set is used for all the HFC calculations, except for the results shown in parentheses, where the def2-TZVP basis set was used.

Nucleusa 3H 4H 5H BH

NR+SO/NR+SO (-90.4) -91.3 (38.0) 37.4 (94.3) 93.6 (125.8) 125.9

CASSCF/PBE0 DKH+SO/NR+SO (-89.9) -90.8 (38.0) 37.4 (94.1) 93.3 (125.3) 125.4

DKH+SO/mDKS -96.7 36.7 94.7 129.9

NR+SO/NR+SO (-92.4) -93.3 (37.9) 37.3 (95.2) 94.5 (127.8) 127.9

NEVPT2/PBE0 DKH+SO/NR+SO (-91.6) -92.5 (37.9) 37.3 (94.9) 94.2 (127.0) 127.1

DKH+SO/mDKS -98.5 36.7 95.5 131.7

Exp.b -111.0 42.0 94.2 122.0

a

In Ref., 12,13 we reassigned the experimental chemical shifts for 3H and 5H nuclei according to calculated ones, stating that ”The need to reassign can be due to different numbering conventions or a wrong experimental signal assignment”. In fact, our own numbering was wrong for the 3H and 5H nuclei, while the experimental signal was correctly assigned. Therefore, we have listed experimental chemical shifts for the 3H and 5H nuclei in their original order here and there is no need for reassignment, contrary to what was suggested in our previous studies. b At 298 K. Solution-state. 97

The isotropic proton shifts for the Cr(III) complex are given in Table 11. The proton shifts show identical behavior at both the CASSCF and NEVPT2 levels; similar relativistic effects and no large differences (maximally about 1 ppm) in the results when one switches from CASSCF to NEVPT2. There is hardly any effect of the different treatment of relativity on the chemical shifts when using the DKH+SO approximation for g and D, instead of the NR-based wave function. However, significant relativistic effects (change by circa 7 ppm) are seen for the methyl groups CH3-1 and CH3-2: there are changes from -10.6 ppm to -3.4 ppm and from -76.4 ppm to -70.7 ppm, respectively, at the NEVPT2/PBE0 level, when we adopt 27

ACS Paragon Plus Environment

Journal of Chemical Theory and Computation

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 28 of 42

the 4-component mDKS approximation for A. In these two cases, incorporating SR effects improves the agreement with experiment, albeit a relatively large deviation remains. Again, as the experiment was performed in solution state, the possible reason for the deviation may be partially due to neglecting the solvation and rovibrational effects in the calculations. Table 11: Calculated isotropic 1 H chemical shifts (in ppm) for quinolyl-functionalized cyclopentadienyl Cr(III) complex at 298 K. Notation: Method for g and D/method for A. The CASSCF/PBE0 and NEVPT2/PBE0 levels used with the def2-TZVP basis set for NR+SO calculations and DKH-TZV basis set for DKH+SO and mDKS calculations. The uncontracted DKH-TZV basis set is used for all the HFC calculations, except for the results shown in parentheses, where the def2-TZVP basis set was used.

Nucleus CH3-1 CH3-2 H-10 H-11 H-12 H-26 H-27 H-28 a

NR+SO/NR+SO (-12.0) -10.6 (-78.7) -75.9 (-64.8) -63.2 (53.4) 53.1 (-97.4) -96.8 (-28.2) -27.6 (22.0) 21.6 (-7.4) -7.2

CASSCF/PBE0 DKH+SO/NR+SO (-12.0) -10.6 (-78.8) -75.9 (-64.8) -63.3 (53.4) 53.1 (-97.3) -96.8 (-28.3) -27.6 (22.0) 21.6 (-7.4) -7.2

DKH+SO/mDKS -3.4 -70.2 -65.3 54.3 -98.6 -27.9 22.3 -7.9

NR+SO/NR+SO (-12.0) -10.5 (-79.3) -76.4 (-65.1) -63.6 (53.6) 53.3 (-98.7) -98.1 (-28.4) -27.7 (22.1) 21.7 (-7.5) -7.3

NEVPT2/PBE0 DKH+SO/NR+SO (-12.0) -10.6 (-79.3) -76.4 (-65.1) -63.6 (53.6) 53.3 (-98.6) -98.1 (-28.4) -27.7 (22.1) 21.7 (-7.5) -7.3

DKH+SO/mDKS -3.4 -70.7 -65.7 54.5 -99.9 -28.0 22.4 -8.0

Exp.a 27.6 -41.1 -56.0 51.8 -78.0 -15.8 15.3 -

At 295 K. Solution-state. 98

5

Conclusions

First-principles calculations of pNMR chemical shifts and shielding anisotropies are reaching a level at which they can be used for meaningful spectral assignment, interpretation and prediction. In this context, multiconfigurational wave function methods are important for the treatment of electron correlation effects in the calculation of the g- and D tensors. In this contribution we have investigated the influence of various approximations for including relativistic effects in the calculations of the pNMR shielding tensor. For the primarily valence-type properties g and D, we used the second order Douglas-Kroll-Hess approximation to investigate the

13

C and 1 H shielding tensors in a series of 3d and group-IX

metallocenes, as well as the 1 H shielding tensors in two larger 3d complexes. In these cal-

28

ACS Paragon Plus Environment

Page 29 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Chemical Theory and Computation

culations, the spin-orbit interaction effects were included in a two-step, post-SCF treatment based on all CASSCF roots allowed by the chosen active space. Scalar relativistic effects on the pNMR shifts and anisotropies arising from the g- and D tensors were found generally small for the 3d systems, whereas a non-negligible influence is seen for the 5d iridocene. As is well-known from earlier literature, the hyperfine coupling is very sensitive to relativistic effects and our present results obtained with the fully relativistic four-component matrix Dirac-Kohn-Sham approach, at the hybrid DFT level, confirm the influence of SR effects on the pNMR shifts and shielding anisotropies. This is seen already for many of the 3d systems and the contribution becomes entirely indispensable for the 5d IrCp2 system. SR effects are quantitatively important for the 1 H shifts in paramagnetic 3d complexes and even more important in the

13

C case. The presently calculated chemical shifts and

shielding anisotropies show that both a reliable treatment of electron correlation and relativistic effects are important for achieving predictive quality in first-principles pNMR calculations. We are presently investigating similar methodology for f -element systems. Despite the progress in electron correlation treatment and incorporating relativistic effects, particularly the intermolecular interaction and solid matrix effects on pNMR shieldings need to be systematically elaborated in forthcoming work.

Acknowledgement Michal Repisk´ y (Tromsø) and Peter Cherry (Bratislava) are thanked for help and useful advice with the ReSpect code. We acknowledge Michael Odelius and Jozef Kowalewski (Stockholm) for useful discussions concerning ORCA. The research leading to these results has received funding from the People Programme (Marie Curie Actions) of the European Union’s Seventh Framework Programme FP7/2007-2013 under REA grant agreement n0 317127. Computational resources were partially provided by CSC-IT Center for Science (Espoo, Finland) and the Finnish Grid Initiative project. The work of JV and JM was

29

ACS Paragon Plus Environment

Journal of Chemical Theory and Computation

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

additionally supported by the Academy of Finland projects 258565 and 296292.

Supporting Information Available Table of the various physical contributions to the paramagnetic nuclear shielding tensor; tables of the optimized geometries of VCp2 , MnCp2 , CoCp2 , RhCp2 , and IrCp2 ; table of the computed

13

C and 1 H hyperfine coupling constants of CoCp2 , RhCp2 , and IrCp2 with

different treatments of relativistic effects; illustration of the shielding anisotropies of the 3d metallocenes; illustrations of the physical contributions to the shielding anisotropies of NiCp2 and IrCp2 ; tables of the physical contributions to the nuclear shielding constants and shielding anisotropies of NiCp2 and IrCp2 . This material is available free of charge via the Internet at http://pubs.acs.org/.

References (1) Bertini, I.; Luchinat, C.; Parigi, G. Solution NMR of Paramagnetic Molecules. Elsevier: Amsterdam, 2001. (2) Harriman, J. E. Theoretical Foundations of Electron Spin Resonance. Academic Press: New York, 1978. (3) Grey, C. P.; Dupr´e, N. NMR studies of cathode materials for lithium-ion rechargeable batteries. Chem. Rev. 2004, 104, 4493-4512. (4) Kaupp, M.; K¨ohler, F. H. Combining NMR spectroscopy and quantum chemistry as tools to quantify spin density distributions in molecular magnetic compounds. Coord. Chem. Rev. 2009, 253, 2376-2386. (5) Keizers, P. H. J.; Ubbink, M. Paramagnetic tagging for protein structure and dynamics analysis. Progr. NMR Spectrosc. 2011, 58, 88-96. 30

ACS Paragon Plus Environment

Page 30 of 42

Page 31 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Chemical Theory and Computation

(6) Calculation of NMR and EPR Parameters: Theory and Applications. Kaupp, M.; B¨ uhl, M.; Malkin, V. G. (editors), Wiley-VCH: Weinheim, 2004. (7) Filatov, M.; Cremer, D. Relativistically corrected hyperfine structure constants calculated with the regular approximation applied to correlation corrected ab initio theory. J. Chem. Phys. 2004, 121, 5618-5622. (8) Van den Heuvel, W.; Soncini, A. NMR chemical shift in an electronic state with arbitrary degeneracy. Phys. Rev. Lett. 2012, 109, 073001. (9) Gendron, F.; Sharkas, K.; Autschbach, J. Calculating NMR chemical shifts for paramagnetic metal complexes from first-principles. J. Phys. Chem. Lett. 2015, 6, 2183-2188. (10) Autschbach, J.; Patchkovskii, S.; Pritchard, B. Calculation of hyperfine tensors and paramagnetic NMR shifts using the relativistic zeroth-order regular approximation and density functional theory. J. Chem. Theory Comput. 2011, 7, 2175-2188. (11) Komorovsk´ y, S.; Repisk´ y, M.; Ruud, K.; Malkina, O. L.; Malkin, V. G. Four-component relativistic DFT calculations of NMR shielding tensors for paramagnetic systems. J. Phys. Chem. A 2013, 117, 14209-14219. (12) Vaara, J.; Rouf, S. A.; Mareˇs, J. Magnetic couplings in the chemical shift of paramagnetic NMR. J. Chem. Theory Comp. 2015, 11, 4840-4849. (13) Rouf, S. A.; Mareˇs, J.; Vaara, J. 1 H chemical shifts in paramagnetic Co(II) pyrazolylborate complexes: A first-principles study. J. Chem. Theory Comp. 2015, 11, 1683-1691. (14) Kurland, R. J.; McGarvey, B. R. Isotropic NMR shifts in transition metal complexes: The calculation of the Fermi contact and pseudocontact terms. J. Magn. Reson. 1970, 2, 286-301. (15) Samzow, R.; Hess, B. A.; Jansen, G. The two-electron terms of the no-pair Hamiltonian. J. Chem. Phys. 1992, 96, 1227-1231. 31

ACS Paragon Plus Environment

Journal of Chemical Theory and Computation

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

(16) Reiher, M.; Wolf, A. Relativistic Quantum Chemistry: The Fundamental Theory of Molecular Science, Wiley-VCH: Verlag GmbH & Co. KGaA, 2009. (17) Van Lenthe, E.; Baerends, E. J.; Snijders, J. G. Relativistic regular two-component Hamiltonians. J. Chem. Phys. 1993, 99, 4597-4610. (18) Douglas, M.; Kroll, N. M. Quantum electrodynamical corrections to the fine structure of helium. Annals. Phys. 1974, 82, 89-155. (19) Hess, B. A. Relativistic electronic-structure calculations employing a two-component no-pair formalism with external-field projection operators. Phys. Rev. A. 1986, 33, 3742-3748. (20) Quiney, H. M.; Belanzoni, P. Relativistic calculation of hyperfine and electron spin resonance parameters in diatomic molecules. Chem. Phys. Lett. 2002, 353, 253-258. (21) Vad, M. S.; Pedersen, M. N.; Norager, A.; Jensen, H. J. Aa. Correlated four-component EPR g-tensors for doublet molecules. J. Chem. Phys. 2013, 138, 214106. (22) Repisk´ y, M.; Komorovsk´ y, S.; Malkin, E.; Malkina, O. L.; Malkin, V. G. Relativistic four-component calculations of electronic g-tensors in the matrix Dirac-Kohn-Sham framework. Chem. Phys. Lett. 2010, 488, 94-97. (23) Komorovsk´ y, S.; Repisk´ y, M.; Malkina, O. L.; Malkin, V. G.; Malkin, I.; Kaupp, M. Resolution of identity Dirac-Kohn-Sham method using the large component only: Calculations of g-tensor and hyperfine tensor. J. Chem. Phys. 2006, 124, 084108. (24) Malkin, E.; Repisk´ y, M.; Komorovsk´ y, S.; Mach, P.; Malkina, O. L.; Malkin, V. G. Effects of finite size nuclei in relativistic four-component calculations of hyperfine structure. J. Chem. Phys. 2011, 134, 044111. (25) Komorovsk´ y, S.; Repisk´ y, M.; Malkina, O. L.; Malkin, V. G.; Malkin Ondik, I.; Kaupp, M. A fully relativistic method for calculation of nuclear magnetic shielding 32

ACS Paragon Plus Environment

Page 32 of 42

Page 33 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Chemical Theory and Computation

tensors with a restricted magnetically balanced basis in the framework of the matrix Dirac-Kohn-Sham equation. J. Chem. Phys. 2008, 128, 104101. (26) Gohr, S.; Hrob´arik, P.; Repisk´ y, M.; Komorovsk´ y, S.; Ruud, K.; Kaupp, M. FourComponent Relativistic Density Functional Theory Calculations of EPR g- and Hyperfine-Coupling Tensors Using Hybrid Functionals: Validation on Transition-Metal Complexes with Large Tensor Anisotropies and Higher-Order Spin-Orbit Effects. J. Phys. Chem. A 2015, 119, 12892-12905. (27) Repisk´ y, M.; Komorovsk´ y, S.; Malkin, V. G.; Malkina, O. L.; Kaupp, M.; Ruud, K.; Bast, R.; Ekstr¨om, U.; Knecht, S.; Malkin Ondik, I.; Malkin, E. Relativistic Spectroscopy DFT program, version 3.3.0 (beta), 2013. (28) Autschbach, J. Calculating NMR chemical shifts and J-couplings for heavy element compounds. Encyclopedia of Analytical Chemistry, John Wiley & Sons, 2014, p. 1-14. (29) Vaara, J. Theory and computation of nuclear magnetic resonance parameters. Phys. Chem. Chem. Phys. 2007, 9, 5399-5418. (30) Jameson, C. J.; de Dios, A. C. Theoretical and physical aspects of nuclear shielding, Nuclear Magnetic Resonance, The Royal Society of Chemistry: London, 2007, Vol. 36, p. 50-71. (31) Casabianca, L. B.; de Dios, A. C. Ab initio calculations of NMR chemical shifts. J. Chem. Phys. 2008, 128, 052201. (32) Cheng, L.; Xiao, Y.; Liu, W. Four-component relativistic theory for NMR parameters: Unified formulation and numerical assessment of different approaches. J. Chem. Phys. 2009, 130, 144102; Publishers Note: Four-component relativistic theory for NMR parameters: Unified formulation and numerical assessment of different approaches. J. Chem. Phys. 2009, 131, 019902.

33

ACS Paragon Plus Environment

Journal of Chemical Theory and Computation

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

(33) Autschbach, J.; Zheng, S. Relativistic computations of NMR parameters from first principles: Theory and applications. Annual Reports on NMR Spectroscopy, Elsevier, 2009, Vol. 67, p. 1-95. (34) Autschbach, J. Relativistic effects on magnetic resonance parameters and other properties of inorganic molecules and metal complexes. in Relativistic Methods for Chemists, Barysz, M. and Ishikawa, Y. (eds.), Springer Science + Business Media B.V. 2010, Vol. 10, p. 521-598. (35) Olejniczak, M.; Bast, R.; Saue, T.; Pecul, M. A simple scheme for magnetic balance in four-component relativistic KohnSham calculations of nuclear magnetic resonance shielding constants in a Gaussian basis. J. Chem. Phys. 2012, 136, 014108. (36) Autschbach, J. Relativistic effects on NMR parameters. in High Resolution NMR Spectroscopy; Contreras, R. H., Ed.; Elsevier: Amsterdam, 2013; Vol. 3, p. 69-117. (37) Autschbach, J. Relativistic calculations of magnetic resonance parameters: Background and some recent developments. Phil. Trans. R. Soc. A. 2014, 372, 20120489. (38) Autschbach, J. NMR calculations for paramagnetic molecules and metal complexes. Annual Reports in Computational Chemistry, Elsevier B.V., 2015, Vol. 11, p. 3-36. (39) Bolvin, H.; Autschbach, J. Relativistic methods for calculating electron paramagnetic resonance (EPR) parameters in Handbook of Relativistic Quantum Chemistry, Liu, W. (eds.), Springer-Verlag: Berlin, Heidelberg, 2017, p. 725-763. (40) de Dios, A. C.; Jameson, C. J. Recent advances in theoretical and physical aspects of NMR chemical shifts. KIMIKA, 2015; Vol. 26, p. 2-31. (41) Repisk´ y, M.; Komorovsk´ y, S.; Bast, R.; Ruud, K. Relativistic calculations of NMR parameters. New Developments in NMR No.6 Gas Phase NMR, Jackowski, K. and Jaszu´ nski, M. (eds.), The Royal Society of Chemistry, 2016, p. 267-303. 34

ACS Paragon Plus Environment

Page 34 of 42

Page 35 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Chemical Theory and Computation

(42) Y. Xiao, Q. Sun, and W. Liu, Fully relativistic theories and methods for NMR parameters. Theor. Chem. Acc. 2012, 131, 1080: 1-17. (43) Xiao, Y.; Liu, W.; Autschbach, J. Relativistic theories of NMR shielding. in Liu, W. (editor), Handbook of Relativistic Quantum Chemistry, Springer, Berlin, 2017, 657-692. (44) Martin, B.; Autschbach, J. Temperature dependence of contact and dipolar NMR chemical shifts in paramagnetic molecules. J. Chem. Phys. 2015, 142, 054108. (45) Moon, S.; Patchkovskii, S. First-principles calculations of paramagnetic NMR shifts, in Ref. 6 Calculation of NMR and EPR Parameters: Theory and Applications, edited by M. Kaupp, M. B¨ uhl, and V. G. Malkin (Wiley-VCH, Weinheim, 2004). (46) Soncini, A.; Van den Heuvel, W. Paramagnetic NMR chemical shift in a spin state subject to zero-field splitting. J. Chem. Phys. 2013, 138, 021103. (47) Van den Heuvel, W.; Soncini, A. NMR chemical shift as analytical derivative of the Helmholtz free energy. J. Chem. Phys. 2013, 138, 054113. (48) Pennanen, T. O.; Vaara, J. Nuclear magnetic resonance chemical shift in an arbitrary electronic spin state. Phys. Rev. Lett. 2008, 100, 133002. (49) Pennanen, T. O.; Vaara, J. Density-functional calculations of relativistic spin-orbit effects on nuclear magnetic shielding in paramagnetic molecules. J. Chem. Phys. 2005, 123, 174102. (50) Hrob´arik, P.; Reviakine, R.; Arbuznikov, A. V.; Malkina, O. L.; Malkin, V. G.; K¨ohler, F. H.; Kaupp, M. Density functional calculations of NMR shielding tensors for paramagnetic systems with arbitrary spin multiplicity: Validation on 3d metallocenes. J. Chem. Phys. 2007, 126, 024107. (51) Hrob´arik, P.; Hrob´arikov´a, V.; Meier, F.; Repisk´ y, M.; Komorovsk´ y, S.; Kaupp, M. Relativistic four-component DFT calculations of 1 H NMR chemical shifts in transition35

ACS Paragon Plus Environment

Journal of Chemical Theory and Computation

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

metal hydride complexes: Unusual high-field shifts beyond the Buckingham-Stephens model. J. Phys. Chem. A. 2011, 115, 5654-5659. (52) Roos, B. O. A new method for large-scale Cl calculations. Chem. Phys. Lett. 1972, 15, 153-159. (53) Neese, F.; Solomon, E. I. Calculation of zero-field splittings, g-values and the relativistic nephelauxetic effect in transition metal complexes. Application to high spin ferric complexes. Inorg. Chem. 1998, 37, 6568-6582. (54) Ganyushin, D.; Neese, F. First principles calculation of zero-field splittings. J. Chem. Phys. 2006, 125, 024103. (55) Ganyushin, D.; Neese, F.

A fully variational spin-orbit coupled complete active

space self-consistent field approach: Application to electron paramagnetic resonance g-tensors. J. Chem. Phys. 2013, 138, 104113. (56) Angeli, C.; Cimiraglia, R.; Evangelisti, S.; Leininger, T.; Malrieu, J. P. Introduction of n-electron valence states for multireference perturbation theory. J. Chem. Phys. 2001, 114, 10252-10264; Angeli, C.; Cimiraglia, R.; Malrieu, J. P. N -electron valence state perturbation theory: A fast implementation of the strongly contracted variant. Chem. Phys. Lett. 2001, 350, 297-305; Angeli, C.; Cimiraglia, R.; Malrieu, J. P. N electron valence state perturbation theory: A spinless formulation and an efficient implementation of the strongly contracted and of the partially contracted variants. J. Chem. Phys. 2002, 117, 9138-9153; Angeli, C.; Borini, S.; Cimiraglia, R. An application of second-order n-electron valence state perturbation theory to the calculation of excited states. Theor. Chem. Acc. 2004, 111, 352-357. (57) Cestari, M.; Angeli, C.; Borini, S.; Cimiraglia, R. A quasidegenerate formulation of the second order-electron valence state perturbation theory approach. J. Chem. Phys. 2004, 121, 4043-4049. 36

ACS Paragon Plus Environment

Page 36 of 42

Page 37 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Chemical Theory and Computation

(58) Adamo, C.; Barone, V. Toward reliable density functional methods without adjustable parameters: The PBE0 model. J. Chem. Phys. 1999, 110, 6158-6170. (59) Cherry, P. J.; Rouf, S. A.; Vaara, J. Paramagnetic enhancement of nuclear spin-spin Coupling. J. Chem. Theory Comput. 2017, 13, 1275-1283. (60) Aquino, F.; Pritchard, B.; Autschbach, J. Scalar relativistic computations and localized orbital analyses of nuclear hyperfine coupling and paramagnetic NMR chemical shifts. J. Chem. Theory Comput. 2012, 8, 598-609. (61) Rinkevicius, Z.; Vaara, J.; Telyatnyk, L.; Vahtras, O. Calculations of nuclear magnetic shielding in paramagnetic molecules. J. Chem. Phys. 2003, 118, 2550-2561. (62) Ramsey, N. F. Magnetic shielding of nuclei in molecules. Phys. Rev. 1950, 78, 699-703. (63) Ramsey, N. F. Chemical effects in nuclear magnetic resonance and in diamagnetic susceptibility. Phys. Rev. 1952, 86, 243-246. (64) Neese, F. Orca, An ab Initio, Density Functional and Semiempirical Program Package, Version 3.0.1, 2012. (65) Hess, B. A.; Marian, C. M.; Wahlgren, U.; Gropen, O. A mean-field spin-orbit method applicable to correlated wavefunctions. Chem. Phys. Lett. 1996, 251, 365-371. (66) Arbuznikov, A. V.; Vaara, J.; Kaupp, M. Relativistic spin-orbit effects on hyperfine coupling tensors by density-functional theory. J. Chem. Phys. 2004, 120, 2127-2139, and references therein. (67) Wilkinson, G.; Rosenblum, M.; Whiting, M. C.; Woodward, R. B. The structure of iron bis-cyclopentadienyl. J. Am. Chem. Soc. 1952, 74, 2125-2126. (68) Fischer, E. O.; Pfab, W. Cyclopentadien-Metallkomplexe, ein neuer Typ metallorganischer Verbindungen. Z. Naturforsch. B 1952, 7, 377-379. 37

ACS Paragon Plus Environment

Journal of Chemical Theory and Computation

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

(69) Turbomole

V6.5 2013,

a development of University of Karlsruhe and

Forschungszentrum Karlsruhe GmbH, TURBOMOLE GmbH: 1989–2007. Available from http://www.turbomole.com. (70) Lee, C.; Yang, W.; Parr, R. G. Development of the Colle-Salvetti correlation-energy formula into a functional of the electron density. Phys. Rev. B 1988, 37, 785-789; Becke, A. D. Density-functional thermochemistry. III. The role of exact exchange. J. Chem. Phys. 1993, 98, 5648-5652; Stephens, P. J.; Devlin, F. J.; Ashvar, C. S.; Chabalowski, C. F.; Frisch, M. J. Ab initio calculation of vibrational absorption and circular dichroism spectra using density functional force fields. J. Phys. Chem. 1994, 98, 11623-11627. (71) Weigend F.; Ahlrichs, R. Balanced basis sets of split valence, triple zeta valence and quadruple zeta valence quality for H to Rn: Design and assessment of accuracy. Phys. Chem. Chem. Phys. 2005, 7, 3297-3305; Accurate Coulomb-fitting basis sets for H to Rn. Ibid. 2006, 8, 1057-1065. (72) Dolg, M.; Wedig, U.; Stoll, H.; Preuß, H. Energy-adjusted ab initio pseudopotentials for the first row transition elements. J. Chem. Phys. 1987, 86, 866-872. (73) Andrae, D.; H¨außermann, U.; Dolg, M.; Stoll, H.; Preuß, H. Energy-adjusted ab initio pseudopotentials for the second and third row transition elements. Theor. Chim. Acta 1990, 77, 123-141. (74) Martin, J. M. L.; Sundermann, A. Correlation consistent valence basis sets for use with the Stuttgart-Dresden-Bonn relativistic effective core potentials: The atoms Ga-Kr and In-Xe. J. Chem. Phys. 2001, 114, 3408-3420. (75) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H. P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; 38

ACS Paragon Plus Environment

Page 38 of 42

Page 39 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Chemical Theory and Computation

Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, Jr., T.; Montgomery, J. A.; Peralta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin, K. N.; Staroverov, V. N.; Kobayashi, R.; Normand, J.; Raghavachari, K.; Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega, N.; Millam, J. M.; Klene, M.; Knox, J. E.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.; Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, O.; Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. Gaussian 09, Revision D.01. Gaussian, Inc.: Wallingford, CT, 2009. (76) Neese, F. Efficient and accurate approximations to the molecular spin-orbit coupling operator and their use in molecular g-tensor calculations. J. Chem. Phys. 2005, 122, 034107; Calculation of the zero-field splitting tensor on the basis of hybrid density functional and Hartree-Fock theory. Ibid. 2007, 127, 164112. (77) Pantazis, D. A.; Chen, X. Y.; Landis, C. R.; Neese, F. All-electron scalar relativistic basis sets for third-row transition metal atoms. J. Chem. Theory Comput. 2008, 4, 908-919; Pantazis, D. A.; Neese, F. All-electron scalar relativistic basis sets for the lanthanides. J. Chem. Theory Comput. 2009, 5, 2229-2238; Pantazis, D. A.; Neese, F. All-electron scalar relativistic basis sets for the actinides. J. Chem. Theory Comput. 2011, 7, 677-684; Pantazis, D. A.; Neese, F. All-electron scalar relativistic basis sets for the 6p elements. Theor. Chem. Acc. 2012, 131, 1292-1298. (78) Dyall, K. G.; Gomes, A. S. P. available in the ReSpect basis-set library, see Ref.. 27 For methodology, see: Dyall, K. G. Core correlating basis functions for elements 31-118. Theor. Chem. Acc. 2012, 131, 1217-1227. (79) Prins, R.; van Voorst, J. D. W.; Schinkel, C. J. Zero-field splitting in the triplet ground state of nickelocene. Chem. Phys. Lett. 1967, 1, 54-55.

39

ACS Paragon Plus Environment

Journal of Chemical Theory and Computation

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

(80) Zvarykina, A. V.; Karimov, Yu. S.; Leonova, E. V.; Lyubovskii, R. B. Temperatureindependent paramagnetism in nikelecene. Sov. Phys.–Solid State (Engl. Transl.) 1970, 12, 385-387. (81) Baltzer, P.; Furrer, A.; Hulliger, J.; Stebler, A. Magnetic properties of nickelocene. A reinvestigation using inelastic neutron scattering and magnetic susceptibility. Inorg. Chem. 1988, 27, 1543-1548. (82) Li, S.; Hamrick, Y. M.; Van Zee, R. J.; Weltner, Jr., W. Far-infrared magnetic resonance of matrix-isolated nickelocene. J. Am. Chem. Soc. 1992, 114, 4433-4434. (83) Oswald, N. Ph.D. thesis (ETH Z¨ urich, 1977). (84) K¨onig, E.; Schnakig, R.; Kanellakopulos, B.; Klenze, R. Magnetism down to 0.90 K, zero-field splitting, and the ligand field in chromocene. Chem. Phys. Lett. 1977, 50, 439-441. (85) Desai, V. P.; K¨onig, E.; Kanellakopulos, B. Magnetic susceptibility and electron paramagnetic resonance of chromocene and 1,1-dimethyl-chromocene: Evidence for the dynamic Jahn-Teller effect. J. Chem. Phys. 1983, 78, 6299-6306. (86) K¨ohler, F. H.; Xie, X. Vanadocene as a temperature standard for 1 3C and 1 H MAS NMR and for solution-state NMR spectroscopy. Magn. Reson. Chem. 1997, 35 487-492. (87) Hebendanz, N.; K¨ohler, F. H.; M¨ uller, G.; Riede, J. Electron spin adjustment in manganocenes. Preparative, paramagnetic NMR and X-ray study on substituent and solvent effects. J. Am. Chem. Soc. 1986, 108, 3281-3289. (88) Hulliger, J.; Zoller, L.; Ammeter, J. H. Orientation effects in EPR powder samples induced by the static magnetic field. J. Magn. Reson. 1982, 48, 512-518. (89) Keller, H. J.; Wawersik, H. Spectroscopic investigations on complex compounds VI. EPR spectra of (C5 H5 )2 Rh and (C5 H5 )2 Ir. J. Organometal. Chem. 1967, 8, 185-188. 40

ACS Paragon Plus Environment

Page 40 of 42

Page 41 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Chemical Theory and Computation

(90) Rettig, M. F.; Drago, R. S.; Nuclear magnetic resonance contact shifts of substituted paramagnetic metallocenes. Chem. Commun. 1966, 891-892; Electron delocalization in paramagnetic metallocenes. I. Nuclear magnetic resonance contact shifts. J. Am. Chem. Soc. 1969, 91, 1361-1370. (91) Heise, H.; K¨ohler, F. H.; Xie, X. Solid-state NMR spectroscopy of paramagnetic metallocenes. J. Magn. Reson. 2001, 150, 198-206. (92) K¨ohler, F. H.; Pr¨oßdorf, W. Metallocenes: First models for nuclear magnetic resonance isotope shifts in paramagnetic molecules. J. Am. Chem. Soc. 1978, 100, 5970-5972. (93) Bl¨ umel, J.; Herker, M.; Hiller, W.; K¨ohler, F. H. Study of paramagnetic chromocenes by solid-state NMR spectroscopy. Organometallics 1996, 15, 3474-3476. (94) Hebendanz, N.; K¨ohler, F. H.; Scherbaum, F.; Schlesinger, B. NMR spectroscopy of paramagnetic complexes. Magn. Reson. Chem. 1989, 27, 798-802. (95) Eicher, H.; K¨ohler, F. H. Determination of the electronic structure, the spin density distribution, and approach to the geometric structure of substituted cobaltocenes from NMR spectroscopy in solution. Chem. Phys. 1988, 128, 297-309. (96) K¨ohler, F. H.; Doll, K.-H.; Pr¨oßdorf, W. Extreme

13

C-NMR data as information on

organometal radicals. Angew. Chem. Int. Ed. Engl. 1980, 19, 479-480. (97) Dlugopolska, K.; Ruman, T.; Danilczuk, M.; Pogocki, D. Analysis of NMR shifts of high-spin cobalt(II) pyrazolylborate complexes. Appl. Magn. Reson. 2008, 35, 271-283. (98) Fern´andez, P.; Pritzkow, H.; Carb´o, J. J.; Hofmann, P.; Enders, M. 1 H NMR investigation of paramagnetic chromium(III) olefin polymerization catalysts: Experimental results, shift assignment and prediction by quantum chemical calculations. Organometallics 2007, 26, 4402-4412.

41

ACS Paragon Plus Environment

Journal of Chemical Theory and Computation

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Graphical TOC Entry

42

ACS Paragon Plus Environment

Page 42 of 42