DaltonTransactions

PAPER

Cite this: Dalton Trans., 2017, 46,8125

Received 10th March 2017,Accepted 16th May 2017

DOI: 10.1039/c7dt00872d

rsc.li/dalton

trans–cis C–Pd–C rearrangement in hemichelates†

Christophe Werlé, a Sebastian Dohm,b Corinne Bailly,a Lydia Karmazin,a

Louis Ricard,c Nicolas Sieffert,d,e Michel Pfeffer,a Andreas Hansen,b

Stefan Grimme *b and Jean-Pierre Djukic *a

Kinetically unstable heteroleptic trans-bispalladacycles were isolated by using hemichelation. Three struc-

tures of trans isomers and five of cis isomers were characterized by X-ray diffraction analysis. The ready

trans-to-cis isomerization of such hemichelates that was monitored by variable temperature NMR experi-

ments is facilitated dynamically because the Pd(II) center can preserve its square planar coordination in a

rather low lying transition state, which was localized by methods of the density functional theory. This

process is not achievable in the isomerization of conventional homoleptic trans-bispalladacycles since it

involves the preliminary partial chelate decoordination and an unfavorable high-lying planar trigonal co-

ordinated – or Y-shaped-Pd(II) transition state according to DFT investigations.

Introduction

The so-called “antisymbiotic effect” proposed by Pearson1 isthe situation where the coordination of a “soft” ligand to a“soft” metal decreases the affinity of the resulting assemblytowards additional “soft” ligands. Pearson’s proposal isperhaps the very first attempt to rationalize intuitively themutual influence of ligands within a metal’s coordinationsphere and its impact on the stereochemistry of metal com-plexes. In the case of organometallic complexes Pearson’s pro-posal can also be rationalized by the mutual trans influencesoperating between ligands: if one considers a “soft” metalcenter, two metal-bound ligands of similar “softness” willprefer to be positioned cis to each other rather than trans,leaving the latter position preferably to “harder” ligands.2 Asan empirical predictive tool, it makes explicit the drive leadingto the most plausible stereochemistry(ies) of reactions placedunder thermodynamic control.3,4 This issue has been addressedin the past by Vicente et al. for Pd(II) complexes of variousligands leading to the emergence of “transphobia” as anotherformulation of Pearson’s proposal.3,5 This peculiar stereochemi-

cal drive in Pd(II) complexes has been seldom addressed withhomo- and hetero-leptic bispalladacycles, which are known toprivilege mostly the cis C–Pd–C stereochemistry,6,7,8–11 whereasboth trans and cis stereochemistries are reportedly known forPt homologues.7,12 Quite interestingly, a rare case of the trans-isomer of bis-7-membered-palladacycles reported by Arlenet al. was reported to slowly convert into the cis isomer.10 Onemay infer from Chini’s seminal report11 that experimental con-ditions (polarity, solubility of the reactants and intermediates,and temperature) are important and may influence the stereo-chemistry of the final d8-metal complex. Indeed in the case ofPt(II) complexes the synthesis of the trans isomer, which pro-ceeds by the formal displacements of labile L ligands and chlor-ide by reaction of PtL2Cl2 complexes (M = Pt) with ortho-lithiatedN,N-dialkylbenzylamine, is essentially observed at ambient temp-erature. The formation of cis isomers is related to a higher re-action temperature in Chini’s report: the synthesis of N,N-di-alkylbenzylamine-based Pd(II) cis bis-chelates is more sluggish(requiring polar solvent and heating) than for the Pt analogleading to partial decomposition and formation of Pd-black along-side the compound of interest. Other reports issued by differentauthors on the synthesis and isolation of bispalladacycles con-cluded that there was exclusive formation of cis-isomers.8,9

Apart from their possible labile character, it is worth noting thatlittle consideration has been given to the polarity of bischelates,which might be at the root of the practical difficulty to isolate transisomers. Similarly to the data reported11 for Pt analogues, isomertrans-I (Fig. 1) possesses a computed dipole moment lower thanthat of cis-I, which could make the former more difficult to isolatefrom a complex reaction mixture than the latter: μcomp(trans-I) =0.8 D and μcomp(cis-I) = 6.1 D (Fig. 1a, cf. the ESI†).

Although difficult to track, the origin of the almost exclu-sive observation of cis isomers in bispalladacycles could result

†Electronic supplementary information (ESI) available: X-ray diffraction struc-tures. CCDC 1522980–1522986 and 1532762–1532764. For ESI and crystallo-graphic data in CIF or other electronic format see DOI: 10.1039/c7dt00872d

aInstitut de Chimie de Strasbourg, UMR 7177 CNRS, Université de Strasbourg,

4 rue Blaise Pascal, 67000 Strasbourg Cedex 08, France. E-mail: djukic@unistra.frbMulliken Center for Theoretical Chemistry, Institut für Physikalische und

Theoretische Chemie, Universität Bonn, Beringstraße 4, D-53115 Bonn, Germany.

E-mail: grimme@thch.uni-bonn.decDépartement de Chimie, Ecole Polytechnique CNRS, Route de Saclay,

F-91128 Palaiseau Cedex, FrancedUniversité Grenoble Alpes & CNRS, DCM UMR 5250, F-38000 Grenoble, FranceeDCM, CNRS, UMR 5250, F-38000 Grenoble, France

This journal is © The Royal Society of Chemistry 2017 Dalton Trans., 2017, 46, 8125–8137 | 8125

Publ

ishe

d on

17

May

201

7. D

ownl

oade

d by

EC

OL

E P

OL

YT

EC

HN

IQU

E o

n 19

/04/

2018

15:

38:4

8.

View Article OnlineView Journal | View Issue

either from a stereospecific reaction of formation or from theisomerization of a short-lived transient trans-bispalladacycle,which has escaped in a great majority of experimentalists’ trap-ping attempts until now.

Hemichelation14–18 is a mode of metal chelation with aheteroditopic ligand where a covalent coordinative bond issupplemented with a predominantly non-covalent stabilizingdelocalized interaction of the metal with another metalcentred moiety (Fig. 1b). For instance, a large number of14-electron so-called T-shaped Pd and Pt complexes, long con-sidered as elusive key intermediates in the final reductive-elimination steps of various catalyses,19 have been isolatedunder a stable cis C–Pd–C form thanks to hemichelation.14–16

In this report, we show that heteroleptic trans C–Pd–C com-plexes can be isolated under the form of manageable Pd(II)hemichelates.14 Furthermore, we show that those isolable butnonetheless metastable trans isomers slowly isomerize in solu-tion into the cis isomers with a Gibbs barrier of activationlower than that computed for the dissociative isomerization ofelusive trans-I into cis-I with state-of-the-art static DFT-D calcu-lations including COSMO-RS20,21 continuum solvation. Thisstudy shows that the accessibility to trans C–Pd–C hemi-chelates essentially depends on the experimental conditions,and on the solubility of the trans isomer. It counter-intuitivelydemonstrates also that a slightly higher degree of covalence inthe Pd–Cr interaction is not a source of higher molecularcohesion and stability in such heterobimetallic complexes.

Results and discussion

The synthesis of Pd(II) hemichelates14,15 derived from pallada-cycles has been reported recently in a series of articles.16

Hemichelates have seldom displayed17,18 propensity to dispro-portionation into homoleptic complexes like other convention-al bischelates;9,12 they can be considered as a new class of“electron deficient” heteroleptic bispalladacycles. Selective syn-thesis and isolation of trans-3 isomers (Scheme 1) wasachieved by a proper choice of the reaction solvent. Reaction ofthe in situ formed lithium benzylic anion derived from 1,22 i.e.Li-1, with μ-chlorido-bridged palladacycles 2a–d (Scheme 1) attemperatures below −10 °C in Et2O resulted in the precipi-tation of reddish solids corresponding to complexes trans-3a–d.

Isolation of pure samples was hence achieved by collectionof the solid precipitate followed by careful recrystallizationfrom chilled Et2O solutions at −40 °C. It must be noted thatthe same experiments carried out in tetrahydrofuran (THF) ledto more homogeneous solutions and resulted in the exclusiveformation of cis-3a–d. For the reaction of Li-1 with 2e in THF,which even required warming the reaction mixture above roomtemperature owing to the low solubility of the palladacycle,only cis-3e was recovered. The lack of solubility of 2e in Et2Oprecluded all attempts to isolate trans-3e. trans-3a, trans-3b,trans-3c and trans-3d were isolated in 45%, 33%, 59% and 42%yields respectively. The yield was 48% for cis-3e.

Owing to the complex and reactive nature of the reactionmedium, which required prolonged reaction times at varioustemperatures (organolithium derivatives) in situ NMR investi-gations were not attempted. However, all isolated complexesdisplayed reasonable resistance to air and moisture eventhough decomposition would ensue upon prolonged exposureof solutions to air. trans to cis isomerization was found tooperate more rapidly with 3a and 3b than with 3c and 3d. Itwas found that at 298 K a benzene solution of trans-3a wouldconvert quantitatively into cis-3a within 0.5 hours (trans/cisratio: 22/78 after 4 h at 273 K starting from trans-3a) whereastrans-3c would require several hours to convert into the cis-isomer (trans/cis ratio: 6/94 after 19 h at 273 K).

Spectroscopic properties

Attenuated total reflectance Fourier-transform IR (ATR-FT-IR)spectroscopy carried out with solid state (amorphous powder)samples displayed the three major CO ligand stretching modesassigned to the main vibrational modes of the Cr(CO)3 tripod.Within a maximum deviation of 8 cm−1, FT-IR spectra of trans-3 isomers and cis-3 isomers showed no significant differencesin their CO stretching frequencies. Furthermore, the averagefrequency of the three stretching modes of all trans/cis-3compounds (νaver ≈ 1892 cm−1) remained similar to those ofcompound 1 (νaver ≈ 1892 cm−1). This result contrasts slightlywith the significant frequency shifts noted for 2-methyl-

Fig. 1 (a) Homoleptic trans- and cis-bispalladacycles I with their com-puted dipole moments at the PBEh-3c13 level (in D, this study cf. theESI†), (b) heteroleptic bischelate made of a conventional chelatingligand and of a hemichelating ligand.

Scheme 1 Synthesis of trans-3a–d and cis-3a–e from 1.

Paper Dalton Transactions

8126 | Dalton Trans., 2017, 46, 8125–8137 This journal is © The Royal Society of Chemistry 2017

Publ

ishe

d on

17

May

201

7. D

ownl

oade

d by

EC

OL

E P

OL

YT

EC

HN

IQU

E o

n 19

/04/

2018

15:

38:4

8.

View Article Online

indenyl-based Pd(II) hemichelates.14,15 13C NMR spectra ofboth trans-3 in [D]8-toluene (223 K) and cis-3 in [D]6-benzene(298 K) displayed three distinct broad signals of the carbonylligands: the signals at δ 241, 238 and 233 ppm in trans-3a–dare characteristic of the hindered Cr(CO)3 tripod rotation. Inthe cis isomers the same signals appeared only slightlyshielded at δ 238, 236 and 232 ppm. The most significantdifference between the 13C NMR spectra of trans and cisisomers was found to arise from the signal of the benzyliccarbon atom bound to the Pd centre: in trans isomers itappeared at ca. δ 53–55 ppm, whereas in cis isomers it wouldshow up at ca. δ 42–44 ppm.

This significant ca. 10 ppm difference might be related to thedrastic difference of the mutual trans influence operating on thebenzylic carbon ligand and the chelating ligand (N,N-dimethyl-benzylamine abbr. DMBA, or arylpyridine) of the palladacycle; atrans CAr–Pd–Cbenzyl (Fig. 2a) arrangement causing a net down-field shift of the benzylic carbon 13C NMR signal. Like shownbelow, this spectroscopic feature is also accompanied by majorstructural differences between trans-3 and cis-3.

Structural analysis

Fig. 3 displays the molecular structures of trans-3a, trans-3c–dand cis-3a–e (see the ESI† and cif for full crystallographicacquisition and data refinement information). It must benoted that two structures of cis-3a were acquired, one of whichcontains a molecule of palladium complex 2a in the asym-metric unit (not shown in Fig. 3, cf. the ESI,† CCDC 1522982).

The most significant differences between trans and cisisomers are discernible from Cr–Pd and Cbenzyl–Pd distances(Fig. 2) that vary (Δd = dcis–dtrans) in the trans → cis transform-ation of 3a and 3c taken here arbitrarily as reference cases(Fig. 3a and d): Δd(Cr–Pd) +0.17 Å (3a), +0.20 Å (3c),Δd(Cbenzyl–Pd) ≈ −0.16 Å (in both 3a and 3c). The Pd–N andPd–CAr distances are surprisingly much less impacted by thestereochemical change affecting the palladacycle and varywithin the limits of the esds: Δd(Pd–N) = +0.01 Å (3a), +0.03 Å(3c); Δd(Pd–CAr) = −0.01 Å (3a), −0.02 Å (3c). Overall, upontrans → cis isomerization the Cbenzyl–Pd distance shortens,which is consistent with the change of the trans influenceoperated by CAr and N atoms respectively. In all the cases thePd coordination geometry is T-shaped23,24 and the Cr(CO)3moiety virtually “occupies” the fourth vacant coordination site

expected for a square-planar Pd(II) complex. The structuresdepicted here are somewhat similar to those of Rh(I) hemi-chelates prepared from 1.18

In solutio dynamic behaviour

The trans-3a → cis-3a and trans-3b → cis-3b isomerizationsmonitored in [D]8-toluene by 1H NMR spectrometry at fivetemperatures ranging from 275 K to 298 K at a total concen-tration of the analyte of ca. 3–4 × 10−3 M were all found totaland irreversible. The Eyring plots ln(k/T ) = f (T−1) (Fig. 4) weredrawn from the values of first order reaction rates extractedfrom the plots of the concentration of trans-3a and trans-3b vs.time. For the trans-3a → cis-3a system, linear fitting gave aGibbs enthalpy of activation ΔG‡ of +21 ± 3 kcal mol−1 at 298 K(ΔH‡ = +17.5 ± 1.5 kcal mol−1, ΔS‡ = −11 ± 5 cal mol−1 K−1,deviations are deduced exclusively from linear regressionfitting, R2 = 0.97). A lower Gibbs enthalpy of activation of+14 kcal mol−1 at 298 K was obtained for the trans/cis-3b system(ΔH‡ = +27 ± 3 kcal mol−1 and ΔS‡ = 44 ± 8 cal mol−1 K−1,R2 = 0.97). Both Eyring plots display linear behaviour thatsuggests that a single rate-determining step was involved.

Theoretical investigation of the trans-3a → cis-3a isomerization

A DFT analysis of the reaction profile was carried out at the –

B3LYP25-NL26/def2-QZVP//PBEh-3c13+COSMO-RS(v16)20,21 levelsto confirm the experimental values and gain deeper insight intothe mechanism of the isomerization reaction. Our DFT modelindeed predicts that cis-3a is more stable than trans-3a (by4.6 kcal mol−1; Table 1). Reaction path optimization produced alow activation barrier profile passing through transition state TS-3a (see Fig. 5a and 9).‡

The associated computed Gibbs barrier to activation ΔG‡ at298.15 K amounts to ca. +15 kcal mol−1 (relative to trans-3a), inrather good agreement with the experimental value derived bythe Eyring plot (+21 ± 3 kcal mol−1). TS-3a results from a slightshift of the Pd centre towards the Cr and the ipso aromaticcarbon and from the rotation of the palladacycle approximatelyaround a Pd–Cipso axis. The projection of the Cr-atom alongthe arene–Cr axis itself is shifted off the center of the aromaticring while palladium partly binds to the Cortho position as aconsequence of the tendency of Pd(II) to preserve its square-planar coordination geometry. The Cr–Pd distance of 2.878 Åin TS-3a, which is longer than in trans-3a (2.694 Å) and cis-3a(2.832 Å), was first analysed as resulting from repulsive d–dinteractions. In fact, Wiberg bond indexes (w) counter-intui-tively indicate that the Cr–Pd interaction is slightly more co-valently binding in TS-3a (w = 0.23 at the PBEh-3c level) thanin the cis isomer (w = 0.19) but still less binding than the transisomer (w = 0.37). This effect can be explained by the transinfluence exerted between Cr and CAr (Fig. 5c). In TS-3a (cf. the

Fig. 2 Schematic representation of the local atoms proximal to the Pdcentre with the descriptors used throughout this article.

‡ Intrinsic reaction coordinate calculations have been performed to confirm thatTS-3a effectively connects trans-3a and cis-3a on the PBEh-3c surface. We alsochecked that the location of this transition state is reproducible at different DFTlevels. For instance, the latter has also been located on the PBE-D3/def2-SV(P)surface.

Dalton Transactions Paper

This journal is © The Royal Society of Chemistry 2017 Dalton Trans., 2017, 46, 8125–8137 | 8127

Publ

ishe

d on

17

May

201

7. D

ownl

oade

d by

EC

OL

E P

OL

YT

EC

HN

IQU

E o

n 19

/04/

2018

15:

38:4

8.

View Article Online

ESI†), the Pd–N (2.169 Å) distance is slightly shortened as com-pared to reference situations computed in models trans-3a(Pd–N 2.268 Å) and cis-3a (Pd–N 2.240 Å), which suggests thatno full decoordination of the N,N-dimethylamino ligand isrequired for the trans-3a → cis-3a isomerization to be effective.

Fig. 3 (a–g) Thermal ellipsoid type diagrams with partial atom numbering of the structures of trans-3a (a), cis-3a (b) cis-3b (c) trans-3c (d), cis-3c(e), trans-3d (f ) and cis-3d (g) drawn at the 50% probability level with omission of solvent molecules for the sake of clarity (for the sake of concise-ness structural details related to 1 and cis-3e were relegated to the ESI† and cif ). Selected interatomic distances (Å) and angles (°) for trans-3a: Pd1–Cr1 2.7123(8), Pd1–C1 2.045(4), Pd1–C10 2.230(4), Pd1–N1 2.169(4), Pd1–C24 3.023(4), Cr1–C25 1.843(4), Cr1–C24 1.851(4), Cr1–C23 1.898(4),C10–Pd1–Cr1 80.63(11), C24–Cr1–C23 103.42(18). Selected interatomic distances (Å) and angles (°) for cis-3a: Cr1–Pd1 2.9149(14), N1–Pd1 2.181(7),C17–Pd1 2.035(8), C7–Pd1 2.069(8), C14–Cr1 1.851(9), C15–Cr1 1.847(10), C16–Cr1 1.841(10), C14–Pd1 2.524(8), C17–Pd1–C7 90.4(3), C15–Cr1–C14 99.7(4). Selected interatomic distances (Å) and angles (°) for cis-3b: Pd1–Cr1 2.9179(6), Pd1–N1 2.188(3), Pd1–C1 2.018(3), Pd1–C23 2.627(4),Pd1–N1 2.188(3), C23–Pd1–C1 140.91(12), C10–Pd1–Cr1 79.48(9). Selected interatomic distances (Å) and angles (°) for trans-3c: Pd1–Cr1 2.6976(10),Pd1–C1 2.253(5), Pd1–C24 2.041(5), Pd1–N1 2.080(5), Cr1–C30 1.881(5), Cr1–C31 1.866(6), Cr1–C32 1.846(6), Pd1–C30 2.424(6), Pd1–C31 2.627(6),C24–Pd1–C1 172.82(19), C31–Cr1–C30 103.6(2). Selected interatomic distances (Å) and angles (°) for cis-3c: Pd1–Cr1 2.8625(7), Pd1–N1 2.112(4),Pd1–C1 2.097(5), Pd1–C24 2.019(4), Cr1–C32 1.851(5), Cr1–C31 1.847(5), Cr1–C30 1.844(5), Pd1–C32 2.647(4), C24–Pd1–C1 92.92(16), C31–Cr1–C32 98.4(2). Selected interatomic distances (Å) and angles (°) for trans-3d: Cr1–Pd1 2.6777(9), C18–Pd1 2.269(5), N1–Pd1 2.092(4), C1–Pd1 2.056(5),C31–Pd1 2.289(5), C1–Pd1–C18 175.2 (2). Selected interatomic distances (Å) and angles (°) for cis-3d: Cr1–Pd1 2.8576(3), N1–Pd1 2.1115(17), C18–Pd1 2.0982(19), C1–Pd1 2.0124(18), C31–Pd1 2.648(2), C32–Pd1 2.649(2), C1–Pd1–C18 93.01(8).

Table 1 Energy contributions (E in Eh, G in kcal mol−1)

SpeciesE (B3LYP-NL/def2-QZVP)

Gsolv (COSMO-RS v16)a GRRHO

b ΔG

cis-3a −2416.7532 −15.7 212.4 −4.6ctrans-3a −2416.7427 −17.4 212.1 0c

TS-3a −2416.7184 −16.8 210.9 14.7c

cis-I −936.7736 −12.4 208.6 0.1d

trans-I −936.7754 −12.4 209.5 0d

TS-I −936.7181 −13.3 206.9 32.3d

a Corrections for solvation Gibbs free enthalpy computed withCOSMO-RS. b GRRHO is the sum of corrections from energy to Gibbsfree enthalpy in the rigid-rotor-harmonic-oscillator approximation(RRHO) also including zero-point-vibrational energy. c Relative to trans-3a. d Relative to trans-I.

Fig. 4 Eyring plots of ln(k/T ) vs. 1/T (K−1) (left) for the spontaneous iso-merization of (a) trans-3a into cis-3a in [D]8-toluene and (b) trans-3binto cis-3b in [D]8-toluene. Rate constants k were determined at varioustemperatures by monitoring the decay of trans isomers and their con-version into cis isomers over time. A first order kinetic law in trans-3a–bwas assumed for the determination of rate constants k.

Paper Dalton Transactions

8128 | Dalton Trans., 2017, 46, 8125–8137 This journal is © The Royal Society of Chemistry 2017

Publ

ishe

d on

17

May

201

7. D

ownl

oade

d by

EC

OL

E P

OL

YT

EC

HN

IQU

E o

n 19

/04/

2018

15:

38:4

8.

View Article Online

The latter shortening of the Pd–N distance is concomitantwith the increase of the Cortho–Pd interaction (2.22 Å and w =0.43 in TS-3a, 3.33 Å and w ≈ 0 in cis-3a) and the decrease ofthe Cbenzyl–Pd bonding interaction (2.37 Å and w = 0.29 in TS-3a, 2.08 Å and w = 0.61 in cis-3a). The strength of the coordina-tive Cipso–Pd interaction (2.33 Å and w = 0.43 in TS-3a, 2.70 Åand w = 0.13 in cis-3a) also increases. Further studies of thechemical bonding within trans and cis-3a were carried outby complementary methods spanning from electron densitytopology analyses (QTAIM and NCI, vide infra) and byfragment-based interaction energy decomposition analyses(ETS-NOCV and EDA, vide infra).

Electron density-based QTAIM and NCI analyses

Complementary Quantum Theory of Atoms in Molecule(QTAIM)27 investigations (Fig. 6, cf. the ESI† for full details) ofboth trans and cis-3a revealed no bond path nor bond criticalpoints (abbr. BCP) (3,−1) in the Cr–Pd segment. In the case ofTS-3a a BCP (3,−1) was found in the Cr–Pd segment, to whichwas assigned a value of Cr–Pd bond ellipticity ε of 0.59 (∇2ρ =0.0469). The localization of a BCP in one Pd–CCrO bond pathin both isomers is consistent with previous observations ofincipient bridging CO interaction made for a number of othercis C–Pd–C hemichelates.28 The ellipticities ε of the mainbonds around the Pd atom (Pd–CAr, Pd–N, Pd–Cbenzyl) in cis-3aare close to 0 (cylindrically symmetric bonds), whereas intrans-3a Pd–Cbenzyl and Pd–N bonds feature some dissymmetrywith ε values around 0.10.

Electron delocalization indexes δ indicate that the electrondensity around the Cr(CO)3 tripod is more localized withinPd–Cr and Pd–CCO diatomic segments in trans-3a (δ(Cr–Pd) =

0.38, δ(CCO–Pd) = 0.53 cf. the ESI†) than in cis-3a (δ(Cr–Pd) =0.25, δ(CCO–Pd) = 0.39 cf. the ESI†).

Interestingly, the participation of attractive interactions inthe energetic differentiation between trans-3a and cis-3a (aspointed out also by EDA analyses, vide infra) can also beshown on non-covalent interaction plots (NCI plots)29 of thetwo stereoisomers (see Fig. 6). Indeed, blue regions (corres-ponding to such attractive interactions) are found between Pdand the Cr(CO)3 fragment in the two complexes, but exhibitdifferent distributions: in trans-3a (Fig. 6a), non-covalent inter-actions are essentially found between Pd and the carbonylligand labelled “C2”. Between Pd and “C1”, the electronicdensity exceeds the cutoff (0.04 a.u.), giving rise to a bluetorus, normal to the Pd–C1 direction. In cis-3a (Fig. 6b), theblue domain spreads between Pd and the whole Cr(CO)3 frag-ment (including Cr and the two carbonyl ligands labelled “C1”and “C2”). Also, red and yellow domains illustrate the exist-ence of repulsive (steric) interactions between the Cr(CO)3 frag-ment and the rest of the complex, in both stereoisomers.Finally, NCI plots also suggest that van der Waals interactions(green domains) are found between the DMBA ligand and theCr(CO)3 fragment, and between the DMBA and fluorenylligands, in both trans-3a and cis-3a.

Fig. 5 The geometries of (a) TS-3a (40 icm−1) and (b) TS-I (193 icm−1)located at the PBEh-3c level.

Fig. 6 Combined representations of QTAIM and NCI analyses on PBEh-3c optimized structures of trans-3a (a) and cis-3a (b) (see Fig. S30 in theESI† for TS-3a). Bond critical points are shown as purple spheres andbond paths are shown as black lines. NCI plot isosurfaces (s = 0.4 a.u.)are superimposed on the QTAIM structures, in the range of −0.04 <sign(λ2)ρ < +0.04 a.u. The colour code (according to the value ofsign(λ2)ρ) is: red +0.04, yellow +0.02, green 0.00, cyan −0.02 and blue−0.04 a.u. The labeling of selected atoms is shown in black. Dipolemoments were computed at the PBEh-3c level.13

Dalton Transactions Paper

This journal is © The Royal Society of Chemistry 2017 Dalton Trans., 2017, 46, 8125–8137 | 8129

Publ

ishe

d on

17

May

201

7. D

ownl

oade

d by

EC

OL

E P

OL

YT

EC

HN

IQU

E o

n 19

/04/

2018

15:

38:4

8.

View Article Online

Fragment-based analyses (ETS-NOCV and EDA)

Standard Energy Decomposition Analysis30,31 (EDA) wascarried out assuming a fictitious closed-shell fragmentationcorresponding to the interaction of the “prepared” pallada-cyclic cation with the Cr-bound fluorenyl anion, a fragmentationscheme that was used also for an Extended Transition State-Natural Orbitals for Chemical Valence (ETS-NOCV)32 analysis.The EDA indicates that the total interfragment interactionenergy ΔEint is by −7 kcal mol−1 more favorable to cis-3a(Table 2). Whereas ΔEorbital, ΔEdisp, ΔEPauli are roughly similarwithin 1 to 3 kcal mol−1 for both complexes, the main differ-ence arises with the value of ΔEelectrostatic that is by −10kcal mol−1 more favourable to cis-3a. The ETS-NOCV32 analysisoutlines the interactions between the Pd center and the benzylicposition as well as the Cr(CO)3 moiety. The electron densitydeformation plot displayed in Fig. 8 depicts the directionalityof electron density transfer upon interaction of the two con-sidered fragments in the subset of interacting orbitals (Δρ1) ofthe highest orbital interaction energy. Red lobes (mostly loca-lized at the “fluorenyl anion”) materialise electron-densitydonating orbitals whereas blue ones materialise acceptingorbitals (Pd–Cbenzyl, Pd–Cr, Pd–CAr and Pd–N segments)among which some are bonding orbitals located between thetwo considered fragments. Fig. 8 qualitatively suggests that the

orbital interaction of the Pd centre with the Cr-bound fluorenylanion is more delocalized in trans-3a than in cis-3a where thelargest electron density buildup (blue lobes) within the Cr–Pd–Cbenzyl triangle is directed to the Pd–Cbenzyl segment.

In summary, in spite of the slightly higher covalent charac-ter of the Pd–Cr(CO)3 interaction in the trans isomer, thechemical system rather privileges the cis isomer where cova-lence in the Pd–Cr interaction recedes to consolidate thePd–Cbenzyl bond. This conclusion is supported by the shorteningof the Cbenzyl–Pd distance (by −0.16 Å in the X-ray structureand by −0.18 Å at the PBEh-3c level) accompanying the elonga-tion of the Cr–Pd distance (by +0.17 Å in the X-ray structureand by +0.27 Å at the PBEh-3c level), when going from trans-3ato cis-3a. Attractive interactions, at least in the first analysis,participate in the energetic differentiation between trans-3aand cis-3a. As a result, QTAIM analyses exhibit a small increaseof the density at the bond critical point between Pd and Cbenzyl

(by 0.04 u.a., see Table S3†). Also, electron delocalisationindices δ(Pd–Cbenzyl) and δ(Pd–Cipso) increase (by 0.24 and0.06, respectively) while δ(Pd–Cr) decreases (by 0.17) whengoing from trans-3a to cis-3a.

Comparisons with homoleptic trans- and cis-bispalladacycles I

Solvent effect-corrected DFT-D computations (COSMO-RSmodel)33 indicate that the mechanism of the trans-I to cis-I iso-merization reaction is dissociative as already suggested byChini.11 Reaction path optimization showed that the isomeri-

Table 2 EDA analyses of trans- and cis-3a (ZORA-PBE-D3(BJ)/all elec-tron QZ4P) according to the closed-shell fragmentation mode depictedin Fig. 7. Energies are expressed in kcal mol−1

ΔEtrans-3a

cis-3a

ΔEint −170 −177ΔEorbital −124 −124ΔEelectrostat −189 −199ΔEdispersion −13 −13ΔEPauli 156 159

Fig. 8 ETS-NOCV analyses (gas-phase singlet ground stateZORA-PBE-D3(BJ)/all electron QZ4P) of the interfragment orbital inter-actions within trans-3a (upper) and cis-3a (lower) where the arbitraryfragments are the Cr-bound fluorenyl anion and the cationic pallada-cycle in their so-called “prepared” geometry (Fig. 7). The isosurfaces(0.005 e bohr−3) depict electron deformation densities Δρ1 of thehighest orbital interaction energy contribution, the red and blue colorsbeing assigned to electron density donating and electron density receiv-ing orbitals respectively.

Fig. 7 Closed-shell fragmentation mode used in the EDA andETS-NOCV analyses for trans-3a and cis-3a.

Paper Dalton Transactions

8130 | Dalton Trans., 2017, 46, 8125–8137 This journal is © The Royal Society of Chemistry 2017

Publ

ishe

d on

17

May

201

7. D

ownl

oade

d by

EC

OL

E P

OL

YT

EC

HN

IQU

E o

n 19

/04/

2018

15:

38:4

8.

View Article Online

zation occurs via full decoordination of one amino ligand toyield TS-I (Fig. 5d). The latter transition state is Y-shaped,24 i.e.of a rather unfavorable trigonal planar coordination geometryat the Pd atom (sum of the coordination angle in the trigonalgeometry Σαβγ = 359.5°). In this case, according to compu-tation, the Gibbs barrier of activation of the trans-I → cis-I iso-merization lies by around 17 kcal mol−1 at least above that ofthe trans-3a/cis-3a isomerization (Fig. 9).

This result echoes the conclusions drawn by Hoffmannet al.24 to rationalize the favorable energy bias for T-shapedML3 (where L is any ligand and M a transition metal) geome-tries over Y-shaped ones. In addition, it is interesting to notethat still in the isomerization of trans-3a a slight contractionrather than an elongation of the Pd–N bond is necessary tooperate the rotation of the palladacycle around the Cr–Pd axisand conversion to cis-3a.

Conclusion

In conclusion, this study shows that trans heteroleptic bische-lates of Pd(II) can be isolated because of their lower solubilityin Et2O compared to the cis-isomers, and the crystal structuresof both trans- and cis-isomers of a variety of bischelates,differing by the nature of their constitutive {C,N} chelatingligand, are reported. Variable temperature NMR experimentsshow that trans-3a and trans-3b are metastable and spon-taneously convert into their cis-isomers, yielding new hemiche-

lates containing a Pd centre of lower coordination number.This may result from the major drive operated by Pearson’s so-called “antisymbiotic effect”.1 The linear Eyring plots revealthat the isomerisation processes require a rather low activationbarrier and that the two systems follow a first-order kinetic law(between 275 K and 298 K). DFT calculations also show that inspite of the more important contribution of covalence to thePd⋯Cr(CO)3 interaction in trans-3a, the latter is metastableand converts readily into the cis isomer in a single elementaryreaction step. The present report shows that an enhanceddegree of covalence between two adjacent metal centres suchas in trans isomers is not a sufficient condition of persistence.Additionally, hemichelation by an arenetricarbonylchromiumligand dynamically offers in the transition state of the trans →cis isomerization process alternative coordination positions tothe Pd centre, which preserves its square planar coordinationand keeps the energetic barrier payload lower than in aY-shaped system devoid of such delocalized secondary inter-actions, such as in TS-I. According to theory, the main conse-quence is that the trans–cis isomerization occurs by rotation ofthe palladacycle roughly around the Cr–Pd axis without partialdecoordination of the {C,N} chelate.

Experimental sectionGeneral

All experiments were carried out under a dry argon atmosphereusing standard Schlenk techniques or in an argon-filled glove-box when necessary. n-Butyllithium was purchased fromAldrich Chem. Co as a 1.6 M solution in hexanes, hexacarbo-nylchromium was purchased from ABCR and fluorene (98%)was purchased from Aldrich Chem. Co. Celite 545 was pur-chased from VWR Prolabo. 4-(t-Butyl)-2-(p-tolyl)pyridine wasprepared according to the literature procedure.34 The pallada-cycles used in this study were prepared according to literatureprocedures. Anhydrous tetrahydrofuran and diethyl ether weredistilled from purple solutions of Na/benzophenone underargon. All other solvents were distilled over sodium or CaH2

under argon. Deuterated solvents were dried over sodium orCaH2 and purified by trap-to-trap techniques, degassed byfreeze–pump–thaw cycles and stored under argon. 1H and 13CNMR spectra were obtained on Bruker DPX 300, 400, Avance I500 and Avance III 600 spectrometers. Chemical shifts(expressed in parts per million) were referenced against solventpeaks or external references. Full NMR spectral assignments areprovided in the ESI.† Infrared spectra of powdered amorphoussamples were acquired with a Fourier transform-IR Brukeralpha spectrometer using an ATR solid state sample cell.

X-ray diffraction analyses

The crystals were placed in oil, and a single crystal wasselected, mounted on a glass fibre and placed in a low-temp-erature N2 stream. Acquisition and processing parameters aredisplayed in Table S1.† Unless otherwise stated the structureswere solved by direct methods using the program

Fig. 9 The mechanisms of isomerization for trans-I (right hand side)and trans-3a (left hand side) as modeled by DFT.

Dalton Transactions Paper

This journal is © The Royal Society of Chemistry 2017 Dalton Trans., 2017, 46, 8125–8137 | 8131

Publ

ishe

d on

17

May

201

7. D

ownl

oade

d by

EC

OL

E P

OL

YT

EC

HN

IQU

E o

n 19

/04/

2018

15:

38:4

8.

View Article Online

SHELXS-97,35 the refinement and all further calculations werecarried out using SHELXL-97.36 The crystal structures acquiredwith the Nonius Kappa CCD were solved using SIR-97 andrefined with SHELXL-97.36 The refinement and all further cal-culations were carried out using SHELXL-97.36 The H-atomswere included in calculated positions and treated as ridingatoms using SHELXL default parameters. The non-H atomswere refined anisotropically, using weighted full-matrix least-squares on F2. A semi-empirical absorption correction wasapplied using SADABS in APEX2.37

For cis-3a (named cis-3a_CH2Cl2 in the .cif, CCDC 1522983),X-ray diffraction data collection was carried out on a NoniusKappa-CCD38 diffractometer equipped with an OxfordCryosystem liquid N2 device, using Mo-Kα radiation (λ =0.71073 Å). The crystal–detector distance was 36 mm. The cellparameters were determined (Denzo software)38 from reflec-tions taken from one set of 10 frames (1.0° steps in phi angle),each at 20 s exposure. The structure was solved by directmethods using the program SHELXS-2013.35 The refinementand all further calculations were carried out usingSHELXL-2013.36 The H-atoms were included in calculated posi-tions and treated as riding atoms using SHELXL default para-meters. The non-H atoms were refined anisotropically, usingweighted full-matrix least-squares on F2. A semi-empiricalabsorption correction was applied using MULscanABS inPLATON;39 transmission factors: Tmin/Tmax = 0.65602/0.87667.

For cis-3b, X-ray diffraction data collection was carried outon a Bruker APEX II DUO Kappa-CCD diffractometer equippedwith an Oxford Cryosystem liquid N2 device, using Mo-Kα radi-ation (λ = 0.71073 Å). The crystal–detector distance was38 mm. The cell parameters were determined (APEX2 soft-ware)37 from reflections taken from three sets of 12 frames,each at 10 s exposure. The structure was solved by directmethods using the program SHELXS-2013.35 The refinementand all further calculations were carried out usingSHELXL-2013.36 The H-atoms were included in calculatedpositions and treated as riding atoms using SHELXL defaultparameters. The non-H atoms were refined anisotropically,using weighted full-matrix least-squares on F2. A semi-empirical absorption correction was applied using SADABS inAPEX2;37 transmission factors: Tmin/Tmax = 0.6162/0.7456.

For trans-3d, X-ray diffraction data collection was carriedout on a Bruker APEX II DUO Kappa-CCD diffractometerequipped with an Oxford Cryosystem liquid N2 device, usingCu-Kα radiation (λ = 1.54178 Å). The crystal–detector distancewas 40 mm. The cell parameters were determined (APEX2 soft-ware)37 from reflections taken from three sets of 20 frames,each at 10 s exposure. The structure was solved by directmethods using the program SHELXS-2013.35 The refinementand all further calculations were carried out usingSHELXL-2013.36 The H-atoms were included in calculated posi-tions and treated as riding atoms using SHELXL default para-meters. The non-H atoms were refined anisotropically, usingweighted full-matrix least-squares on F2. A semi-empiricalabsorption correction was applied using SADABS in APEX2;37

transmission factors: Tmin/Tmax = 0.5638/0.7528.

For cis-3d, X-ray diffraction data collection was carried outon a Bruker APEX II DUO Kappa-CCD diffractometer equippedwith an Oxford Cryosystem liquid N2 device, using Mo-Kα radi-ation (λ = 0.71073 Å). The crystal–detector distance was38 mm. The cell parameters were determined (APEX2 soft-ware)37 from reflections taken from three sets of 12 frames,each at 10 s exposure. The structure was solved by directmethods using the program SHELXS-2013.35 The refinementand all further calculations were carried out usingSHELXL-2013.36 The H-atoms were included in calculated posi-tions and treated as riding atoms using SHELXL default para-meters. The non-H atoms were refined anisotropically, usingweighted full-matrix least-squares on F2. A semi-empiricalabsorption correction was applied using SADABS in APEX2;37

transmission factors: Tmin/Tmax = 0.6733/0.7458. The SQUEEZEinstruction in PLATON39 was applied. The residual electrondensity was assigned to half a molecule of the dichloro-methane solvent.

For cis-3e, X-ray diffraction data collection was carried outon a Bruker APEX II DUO Kappa-CCD diffractometer equippedwith an Oxford Cryosystem liquid N2 device, using Mo-Kαradiation (λ = 0.71073 Å). The crystal–detector distance was38 mm. The cell parameters were determined (APEX2 soft-ware)37 from reflections taken from three sets of 12 frames,each at 10 s exposure. The structure was solved by directmethods using the program SHELXS-2013.35 The refinementand all further calculations were carried out usingSHELXL-2013.36 The H-atoms were included in calculated posi-tions and treated as riding atoms using SHELXL default para-meters. The non-H atoms were refined anisotropically, usingweighted full-matrix least-squares on F2. A semi-empiricalabsorption correction was applied using SADABS in APEX2;37

transmission factors: Tmin/Tmax = 0.6549/0.7456.

Synthesis of 2d

A mixture containing palladium(II) acetate (1.5 g, 6.68 mmol)and 4-(4-(tert-butyl)pyridin-2-yl)-N,N-dimethylaniline40 (1.7 g,6.68 mmol) in 10 mL of dry degassed dichloromethane wasstirred for 18 h at room temperature. The resulting mixturewas filtered through Celite and solvents were removed fromthe filtrate under reduced pressure. The crude residue was dis-solved in a acetone : H2O (3 : 1, 20 mL) mixture and an excessof LiCl was added. After stirring at room temperature for 1 h,the resulting solution was extracted with three volumes(15 mL) of dichloromethane. The combined organic layerswere dried over anhydrous MgSO4 and the solvent wasremoved under reduced pressure. Finally recrystallisation in adichloromethane/pentane mixture of solvents led to theexpected pale yellow compound 2d (2.419 g, 3.06 mmol, 92%yield). Anal. calcd for C34H42Cl2N4Pd2·1/10 CH2Cl2: C, 51.26;H, 5.32; N, 7.01. Found: C, 51.13; H, 5.37; N, 6.97. HRMS-ESI(m/z): [M + 1Na]+ calcd for C34H42Cl2N4Pd2, 789.0929; found,789.0872. 1H NMR (400 MHz, C6D6 + [D]5-pyridine, 298 K)δ 9.90 (d, J = 6.2 Hz, 1H, H1), 7.43 (d, J = 8.6 Hz, 1H, H7), 7.33(d, J = 2.1 Hz, 1H, H4), 6.44 (dd, J = 8.6, 2.5 Hz, 1H, H8), 6.40(dd, J = 6.3, 2.1 Hz, 1H, H2), 5.66 (d, J = 2.5 Hz, 1H, H10), 2.41

Paper Dalton Transactions

8132 | Dalton Trans., 2017, 46, 8125–8137 This journal is © The Royal Society of Chemistry 2017

Publ

ishe

d on

17

May

201

7. D

ownl

oade

d by

EC

OL

E P

OL

YT

EC

HN

IQU

E o

n 19

/04/

2018

15:

38:4

8.

View Article Online

(s, 6H, H17, H16), 0.96 (s, 9H, H14, H15, H13).13C NMR

(101 MHz, C6D6 + Pyr-d5, 298 K) δ 166.5 (C5), 162.2 (C3), 158.2(C11), 152.3 (C1), 150.7 (C9), 134.8 (C6), 124.4 (C7), 117.5 (C2),116.5 (C10), 113.5 (C4), 108.5 (C8), 39.7 (C17, C16), 34.8 (C12),30.1 (C14, C15, C13) (Fig. 10).

Standard procedure for the synthesis of trans-3 compounds

Tricarbonyl(η6-fluorene)chromium(0), i.e. 1, was dissolved indiethyl ether (5 mL) and treated with n-BuLi at −40 °C underargon. The resulting solution was transferred after 30 min viaa cannula to another Schlenk vessel containing a diethyl ether(3 mL) solution of the corresponding μ-chloro-bridged pallada-cycle 2. The resulting mixture was stirred for 2 h and the temp-erature was slowly raised to −10 °C. During this time producttrans-3 precipitated out of the solution as a reddish solid. Thesolvent was subsequently removed and the residue was washedwith cold diethyl ether (15 mL). Filtration through Celite andrecrystallization from a dichloromethane/pentane mixture ofsolvents afforded the product as a red solid.

Standard procedure for the isomerization of trans-3 into cis-3complexes

trans-3 was dissolved under argon in dry benzene (5 mL) andstirred at 40 °C for 2 hours, which led to the quantitative for-mation of cis-3.

Synthesis of complexes trans-3a and cis-3a

Tricarbonyl(η6-fluorene)chromium(0), i.e. 1, (0.200 g,0.66 mmol), n-BuLi (0.45 mL, 0.73 mmol), bis(μ-chlorido)(N,N-dimethylaminophenylenemethane-κC,κN)palladium(II)41

(0.186 g, 0.34 mmol): trans-3a (0.160 g, 0.30 mmol, 45% yield).Calcd for C25H21CrNO3Pd: C, 55.42; H, 3.91; N, 2.58. Found: C,55.39; H, 4.22; N, 2.79. HRMS-ESI (m/z): [M + 1H]+ calcd forC25H21CrNO3Pd, 542.0034; found, 541.9968. trans-3a: IR(cm−1) ν(CO): 1944 (s), 1876 (s), 1839 (vs). 1H NMR (600 MHz,C7D8, 223 K) δ 7.52 (d, J = 7.7 Hz, 1H, H′6), 7.43 (d, J = 6.8 Hz,1H, H′18), 7.36–7.32 (m, 2H, H′4, H′3), 7.24–7.19 (m, 1H, H′5),7.03–6.97 (m, 2H, H′19, H′20), 6.92 (d, J = 7.2 Hz, 1H, H′21), 5.64(d, J = 6.5 Hz, 1H, H′9), 5.53 (d, J = 6.5 Hz, 1H, H′12), 4.77 (t, J =6.6 Hz, 1H, H′11), 4.32 (t, J = 6.4 Hz, 1H, H′10), 4.15 (s, 1H, H′1),3.93 (d, J = 13.4 Hz, 1H, H′23a), 2.81 (d, J = 11.0 Hz, 1H, H′23b),2.42 (s, 3H, H′24), 2.08 (s, 3H, H′25).

13C NMR (151 MHz, C7D8,

223 K) δ 241.3 (C′16), 238.3 (C′15), 234.1 (C′14), 163.0 (C′17),147.8 (C′22), 146.6 (C′2), 139.0 (C′18), 129.0 (C′7), 128.4 (C′3),124.8 (C′19), 124.2 (C′20), 122.3 (C′6), 121.9 (C′21), 121.0 (C′5),118.0 (C′4), 114.2 (C′8), 102.4 (C′13), 91.34 (C′10), 90.7 (C′12),90.3 (C′9), 89.8 (C′11), 74.5 (C′23), 54.2 (C′1), 52.0 (C′25), 51.6(C′24). cis-3a: IR (cm−1) ν(CO): 1940 (s), 1886 (s), 1843 (vs). 1HNMR (600 MHz, C6D6, 298 K) δ 8.23 (d, J = 7.4 Hz, 1H, H18),7.71–7.67 (m, 1H, H6), 7.42 (td, J = 7.4, 1.5 Hz, 1H, H19),7.26–7.24 (m, 1H, H3), 7.22 (td, J = 7.4, 1.1 Hz, 1H, H20), 7.00(d, J = 7.2 Hz, 1H, H21), 6.97–6.90 (m, 2H, H5, H4), 5.42 (d, J =6.5 Hz, 1H, H9), 5.20 (d, J = 6.4 Hz, 1H, H12), 4.99 (t, J = 6.4 Hz,1H, H10), 4.85 (s, 1H, H1), 4.23–4.14 (m, 1H, H11), 3.58 (d, J =12.9 Hz, 1H, H23a), 2.82 (d, J = 12.9 Hz, 1H, H23b), 2.33 (s, 3H,H24), 2.17 (s, 3H, H25).

13C NMR (151 MHz, C6D6, 301 K)δ 237.3 (bs, C16), 236.2 (bs, C15), 232.7 (bs, C14), 160.9 (C17),151.0 (C7), 147.4 (C22), 133.2 (C18), 130.0 (C2), 129.4 (C5), 125.9(C19), 124.0 (C20), 123.5 (C21), 122.9 (C6), 122.5 (C4), 121.1 (C3),108.8 (C8), 108.5 (C13), 94.9 (C12), 92.8 (C10), 90.2 (C11), 88.4(C9), 73.0 (C23), 50.1 (C25), 48.8 (C24), 41.5 (C1) (Fig. 11).

Synthesis of complexes trans-3b and cis-3b

1 (0.200 g, 0.66 mmol), n-BuLi (0.45 mL, 0.73 mmol), bis(μ-chlorido)(4-fluoro,N,N-dimethylaminophenylenemethane-κC,κN)palladium(II)42 (0.198 g, 0.34 mmol): trans-3b (0.109 g,0.22 mmol, 33% yield). Calcd for C25H20CrFNO3Pd·0.1CH2Cl2:C, 53.04; H, 3.58; N, 2.46. Found: C, 53.07; H, 3.76; N, 2.18.HRMS-ESI (m/z): [M + 1H]+ calcd for C25H20CrFNO3Pd,559.9940; found, 559.9936. trans-3b: IR (cm−1) ν(CO): 1956 (s),1922 (vs), 1829 (s). 1H NMR (600 MHz, Tol, 223 K) δ 7.44 (d, J =7.8 Hz, 1H, H′6), 7.34–7.28 (m, 3H, H′18, H′4, H′3), 7.22–7.16(m, 1H, H′5), 6.75–6.68 (m, 2H, H′20, H′21), 5.53 (d, J = 6.5 Hz,1H, H′9), 5.39 (d, J = 6.5 Hz, 1H, H′12), 4.71 (t, J = 6.3 Hz, 1H,H′11), 4.24 (t, J = 6.6 Hz, 1H, H′10), 4.07 (s, 1H, H′1), 3.83 (d, J =13.3 Hz, 1H, H′23a), 2.73 (d, J = 13.4 Hz, 1H, H′23b), 2.39 (s, 3H,H′24), 2.04 (s, 3H, H′25).

13C NMR (151 MHz, Tol, 223 K) δ 240.4(C′16), 238.1 (C′15), 233.7 (C′14), 166.7 (d, J = 1.5 Hz, C′22), 161.2(d, J = 238.3 Hz, C′19), 147.0 (C′7), 142.9 (d, J = 2.1 Hz, C′17),129.5 (C′2), 128.6 (C′4), 124.5 (d, J = 18.7 Hz, C′18), 122.7 (d, J =6.2 Hz, C′21), 122.4 (C′6), 121.3 (C′5), 118.2 (C′3), 113.8 (C′8),110.3 (d, J = 23.4 Hz, C′20), 102.5 (C′13), 91.5 (C′10), 90.9 (C′12),90.5 (C′9), 90.2 (C′11), 73.8 (C′23), 53.4 (C′1), 51.8 (C′24), 49.7(C′25).

19F NMR (282 MHz, Tol, 296 K) δ −115.8. cis-3b: IR(cm−1) ν(CO): 1945 (vs), 1883 (s) 1843 (vs). 1H NMR (500 MHz,293 K, C6D6) δ 8.09 (dd, J = 9.0, 2.2 Hz, 1H, H18), 7.66 (d, J = 7.3

Fig. 10 Atom numbering schemes for NMR assignments in 2d.

Fig. 11 Atom numbering schemes for NMR assignments in trans (left)and cis (right)-3a.

Dalton Transactions Paper

This journal is © The Royal Society of Chemistry 2017 Dalton Trans., 2017, 46, 8125–8137 | 8133

Publ

ishe

d on

17

May

201

7. D

ownl

oade

d by

EC

OL

E P

OL

YT

EC

HN

IQU

E o

n 19

/04/

2018

15:

38:4

8.

View Article Online

Hz, 1H, H6), 7.19 (d, J = 7.2 Hz, 1H, H3), 6.96–6.88 (m, 3H, H20,H5, H4), 6.79 (dd, J = 7.9, 5.6 Hz, 1H, H21), 5.38–5.33 (m, 1H,H9), 5.10 (d, J = 6.4 Hz, 1H, H12), 4.96 (t, J = 6.4 Hz, 1H, H10),4.62 (s, 1H, H1), 4.16 (t, J = 6.4 Hz, 1H, H11), 3.49–3.43 (m, 1H,H23a), 2.72 (d, J = 12.8 Hz, 1H, H23b), 2.26 (s, 3H, H24), 2.11 (s,3H, H25).

13C NMR (126 MHz, C6D6, 293 K) δ 237.8 (bs, C16),236.1 (bs, C15), 232.3 (bs, C14), 164.2 (d, J = 1.4 Hz, C17), 161.8(d, J = 247.5 Hz, C19), 151.0 (C7), 143.3 (d, J = 2.5 Hz, C22),130.5 (C2), 130.0 (C4), 124.6 (d, J = 7.0 Hz, C21), 123.2 (C6),123.1 (C5), 121.5 (C3), 120.0 (d, J = 17.4 Hz, C18), 110.6 (d, J =21.9 Hz, C20), 109.0 (C8), 108.5 (C13), 95.2 (C12), 93.3 (C10), 90.8(C11), 89.0 (C9), 72.72 (C23), 50.3 (C25), 48.9 (C24), 41.9 (C1).

19FNMR (282 MHz, Tol, 296 K) δ −116.4 (Fig. 12).

Synthesis of complexes trans-3c and cis-3c

1, (0.200 g, 0.66 mmol), n-BuLi (0.45 mL, 0.73 mmol), bis(μ-chlorido)[2-(4-tolylene),4-tert-isobutylpyridine,κC,κN]palla-dium(II)14 (0.247 g, 0.34 mmol): trans-3b (0.231 g, 0.39 mmol,59% yield). Calcd for C32H27CrNO3Pd·0.4CH2Cl2: C, 58.44; H,4.21; N, 2.10. Found: C, 58.37; H, 4.29; N, 2.06. HRMS-ESI(m/z): [M + 1H]+ calcd for C32H27CrNO3Pd, 632.0504; found,632.0572. trans-3c: IR (cm−1) ν(CO): 1945 (s), 1875 (s), 1858(vs). 1H NMR (600 MHz, C7D8, 223 K) δ 8.84 (d, J = 6.1 Hz, 1H,H′27), 7.68 (d, J = 8.0 Hz, 1H, H′6), 7.53 (d, J = 7.7 Hz, 1H, H′3),7.44 (d, J = 1.3 Hz, 1H, H′18), 7.39 (d, J = 1.9 Hz, 1H, H′24), 7.31(d, J = 8.1 Hz, 1H, H′21), 7.20 (t, J = 7.5 Hz, 1H, H′5), 7.16–7.10(m, 1H, H′4), 6.80 (d, J = 7.4 Hz, 1H, H′20), 6.55 (dd, J = 6.0, 2.1Hz, 1H, H′26), 5.55 (d, J = 6.4 Hz, 1H, H′9), 5.50 (d, J = 6.6 Hz,1H, H′12), 4.93 (t, J = 6.4 Hz, 1H, H′11), 4.63 (s, 1H, H′1), 4.43(td, J = 6.4, 1.1 Hz, 1H, H′10), 2.22 (s, 3H, H′32), 1.05 (s, 9H, H′

30, H′31, H′29). 13C NMR (151 MHz, C7D8, 213 K) δ 241.6(C′16), 238.3 (C′15), 233.2 (C′14), 170.3 (C′17), 167.5 (C′23), 161.3(C′25), 147.8 (C′27), 146.9 (C′7), 143.7 (C′22), 140.5 (C′18), 139.4(C′19), 129.6 (C′2), 129.1 (C′5), 124.6 (C′20), 123.8 (C′21), 122.2(C′3), 121.2 (C′4), 119.0 (C′6), 119.0 (C′26), 115.7 (C′24), 114.6(C′8), 103.7 (C′13), 91.9 (C′9), 91.6 (C′12), 91.4 (C′11), 91.1 (C′10),54.8 (C′1), 34.7 (C′28), 29.7 (C′30, C′31, C′29), 22.2 (C′32). cis-3c: IR(cm−1) ν(CO): 1952 (s), 1880 (s), 1857 (vs). 1H NMR (500 MHz,C6D6, 293 K) δ 8.46 (s, 1H, H18), 8.44 (dd, J = 6.1, 0.6 Hz, 1H,H27), 8.07 (d, J = 7.9 Hz, 1H, H6), 7.55 (d, J = 7.9 Hz, 1H, H21),7.35 (d, J = 2.1 Hz, 1H, H24), 7.34–7.29 (m, 1H, H3), 7.11–7.06(m, 2H, H5, H20), 7.03–6.96 (m, 1H, H4), 6.26 (dd, J = 6.1, 2.1Hz, 1H, H26), 5.42 (d, J = 6.5 Hz, 1H, H9), 5.34 (d, J = 6.5 Hz,1H, H12), 5.10 (s, 1H, H1), 5.08 (t, J = 6.4 Hz, 1H, H10), 4.27 (td,

J = 6.4, 1.0 Hz, 1H, H11), 2.55 (s, 3H, H32), 0.83 (s, 9H, H30, H31,H29).

13C NMR (126 MHz, C6D6, 293 K) δ 237.6 (bs, C16), 236.9(bs, C15), 232.0 (bs, C14), 165.5 (C17), 165.4 (C23), 161.4 (C25),151.9 (C27), 150.9 (C7), 145.2 (C22), 139.5 (C19), 135.5 (C18),130.9 (C2), 129.9 (C5), 125.3 (C20), 124.5 (C21), 122.9 (C4), 122.7(C6), 121.7 (C3), 119.5 (C26), 115.3 (C24), 109.4 (C13), 109.2 (C8),95.6 (C12), 93.2 (C10), 90.5 (C11), 88.9 (C9), 43.5 (C1), 34.6 (C28),29.5 (C30, C31, C29), 22.4 (C32) (Fig. 13).

Synthesis of complexes trans-3d and cis-3d

1 (0.200 g, 0.66 mmol), n-BuLi (0.45 mL, 0.73 mmol), bis(μ-chlorido)(4-t-butyl-(4-N,N-dimethylaminophenylene)pyri-dine-κC,κN) palladium(II) (0.267 g, 0.34 mmol): trans-3d(0.184 g, 0.28 mmol, 42% yield). Calcd forC33H30CrN2O3Pd·1.1CH2Cl2: C, 54.29; H, 4.30; N, 3.71. Found:C, 53.95; H, 4.69; N, 3.62. HRMS-ESI (m/z): [M + 1H]+ calcd forC33H30CrN2O3Pd, 661.0769; found, 661.0734. trans-3d. IR(cm−1) ν(CO): 1948 (s), 1885 (s), 1860 (vs). 1H NMR (600 MHz,Tol, 223 K) δ 8.82 (d, J = 6.0 Hz, 1H, H′27), 7.71 (d, J = 8.1 Hz,1H, H′6), 7.56 (d, J = 8.1 Hz, 1H, H′3), 7.42 (d, J = 9.0 Hz, 1H,H′21), 7.38 (d, J = 2.0 Hz, 1H, H′24), 7.22–7.16 (m, 1H, H′5),7.16–7.10 (m, 1H, H′4), 6.78 (d, J = 2.5 Hz, 1H, H′18), 6.49 (dd,J = 6.0, 2.0 Hz, 1H, H′26), 6.27 (dd, J = 8.6, 2.6 Hz, 1H, H′20),5.57 (d, J = 6.9 Hz, 2H, H′12, H′9), 4.93 (t, J = 6.7 Hz, 1H, H′10),4.66 (s, 1H, H′1), 4.47 (t, J = 6.3 Hz, 1H, H′11), 2.63 (s, 6H, H′32,H′33), 1.10 (s, 9H, H′30, H′31, H′29).

13C NMR (126 MHz, C6D6,293 K) δ 242.1 (C′16), 238.9 (C′15), 233.8 (C′14) (fast acquisitionat room temperature was required in order to obtain the carbo-nyl signals that couldn’t be obtained at low temperature due tolow solubility in toluene). 13C NMR (151 MHz, Tol, 223 K) δ170.9 (C′17), 168.1 (C′23), 160.5 (C′25), 149.8 (C′19), 147.5 (C′27),147.2 (C′7), 133.8 (C′22), 129.3 (C′2), 128.2 (C′5), 125.0 (C′21),122.4 (C′18), 121.9 (C′3), 120.9 (C′4), 119.0 (C′6), 116.7 (C′26),114.7 (C′13) 114.4 (C′24), 107.3 (C′20), 103.6 (C′8), 91.4 (C′12, C′9),91.0 (C′11), 90.3 (C′10), 54.8 (C′1), 39.5 (C′32, C′33), 34.8 (C′28),29.8 (C′30, C′31, C′29). cis-3d: IR (cm−1) ν(CO): 1937 (s), 1886 (s),1849 (vs). 1H NMR (500 MHz, C6D6, 293 K) δ 8.51 (d, J = 6.1 Hz,1H, H27), 8.25 (d, J = 7.9 Hz, 1H, H6), 8.11 (d, J = 2.6 Hz, 1H,H18), 7.74 (d, J = 8.6 Hz, 1H, H21), 7.36–7.29 (m, 2H, H3, H24),7.12–7.06 (m, 1H, H5), 7.05–6.98 (m, 1H, H4), 6.73 (dd, J = 8.6,2.5 Hz, 1H, H20), 6.32 (dd, J = 6.1, 2.1 Hz, 1H, H26), 5.63 (d, J =6.4 Hz, 1H, H9), 5.57 (d, J = 6.5 Hz, 1H, H12), 5.12–5.04 (m, 2H,H1, H10), 4.37 (td, J = 6.5, 1.1 Hz, 1H, H11), 3.10 (s, 6H, H32,

Fig. 12 Atom numbering schemes for NMR assignments in trans (left)and cis (right)-3b.

Fig. 13 Atom numbering schemes for NMR assignments in trans (left)and cis (right)-3c.

Paper Dalton Transactions

8134 | Dalton Trans., 2017, 46, 8125–8137 This journal is © The Royal Society of Chemistry 2017

Publ

ishe

d on

17

May

201

7. D

ownl

oade

d by

EC

OL

E P

OL

YT

EC

HN

IQU

E o

n 19

/04/

2018

15:

38:4

8.

View Article Online

H33), 1.08 (s, 9H, H30, H31, H29).13C NMR (126 MHz, C6D6,

293 K) δ 237.9 (bs, C16), 237.1 (bs, C15), 232.3 (bs, C14), 166.5(C19), 166.1 (C25), 160.9 (C23), 151.7 (C27), 151.4 (C17), 151.0(C7), 136.4 (C22), 131.0 (C2), 129.8 (C5), 125.7 (C21), 123.0 (C6),122.7 (C4), 121.7 (C3), 118.1 (C18), 117.9 (C26), 114.2 (C24), 109.7(C13), 109.0 (C8), 108.6 (C20), 95.6 (C12), 93.1 (C10), 90.4 (C11),88.7 (C9), 44.2 (C1), 40.4 (C32, C33), 34.6 (C28), 30.0 (C30, C31,C29) (Fig. 14).

Synthesis of compound cis-3e

1 (0.200 g, 0.66 mmol), n-BuLi (0.45 mL, 0.73 mmol), THF(5 mL), bis(μ-chlorido)(8-methylquinolynyl,κC8,κN)palladium(II)43

(0.199 g, 0.35 mmol): cis-3e (0.175 g, 0.32 mmol, 48% yield).Calcd for C26H17CrNO3Pd·0.1CH2Cl2: C, 56.15; H, 3.11; N,2.51. Found: C, 56.28; H, 3.29; N, 2.75. HRMS-ESI (m/z): [M +1Na]+ calcd for C26H17CrNO3Pd, 571.4541; found, 571.9518. IR(cm−1) ν(CO): 1931 (s), 1873 (s), 1849 (vs). 1H NMR (600 MHz,C6D6, 298 K) δ 8.72 (dd, J = 4.9, 1.6 Hz, 1H, H25), 7.69 (d, J = 7.9Hz, 1H, H6), 7.52 (d, J = 7.1 Hz, 1H, H21), 7.40 (d, J = 7.5 Hz,1H, H3), 7.26–7.19 (m, 2H, H5, H23), 7.11–7.05 (m, 2H, H4,H20), 6.95 (d, J = 8.0 Hz, 1H, H19), 6.35 (dd, J = 8.2, 4.9 Hz, 1H,H24), 5.56 (d, J = 6.6 Hz, 1H, H9), 5.33 (d, J = 6.5 Hz, 1H, H12),5.08 (t, J = 6.4 Hz, 1H, H10), 4.74 (s, 1H, H1), 4.30 (t, J = 6.4 Hz,1H, H11), 4.25 (d, J = 13.9 Hz, 1H, H17a), 3.69 (d, J = 13.9 Hz,1H, H17b).

13C NMR (126 MHz, C6D6, 293 K) δ 238.1 (bs, C16),237.9 (bs, C15), 232.9 (bs, C14), 153.4 (C26), 152.3 (C25), 151.3(C7), 151.2 (C18), 136.9 (C23), 130.3 (C2), 129.6 (C21), 129.6 (C5),129.5 (C22), 127.5 (C20), 123.6 (C19), 122.2 (C4), 121.8 (C3), 121.7(C24), 121.3 (C6), 109.2 (C8), 107.9 (C13), 94.3 (C12), 92.3 (C10),90.1 (C11), 89.4 (C9), 38.9 (C17), 34.6 (C1) (Fig. 15).

Computational details related to transition state search andgeometry optimizations

DFT calculations were performed with the TURBOMOLEprogram package in V7.0.2.44 (see the ESI† for completedetails). Geometry optimizations were carried out using thePBEh-3c density functional composite method.13 Single-pointenergies were obtained with the B3LYP25 method, togetherwith the extended quadruple-zeta basis set def2-QZVP.45 In allDFT calculations the resolution-of-identity (RI) approximationfor the Coulomb integrals46 with matching default auxiliarybasis sets47 was applied. For integration of the exchange corre-lation contribution, the numerical quadrature grid m4 wasemployed for B3LYP. For DFT calculations with the PBEh-3cfunctional, the D3 dispersion-correction scheme,48 applyingthe Becke–Johnson (BJ) damping,49,50 was used. Reaction pathoptimization was performed with the WOELFLING51 programincluded in TURBOMOLE. For evaluation of the single pointenergies the nonlocal (NL) van der Waals functional approachby Vydrov and van Voorhis52 DFT-NL26 was employed non-self-consistently (i.e., calculating the NL energy only once in a post-SCF manner) for B3LYP. Computations of the harmonicvibrational frequencies were performed at the PBEh-3c level.Thermal corrections (Table 1) from energy to Gibbs freeenthalpy were obtained by a coupled rigid-rotor-harmonic-oscillator approximation for each molecule in the gas phase.53

The PBEh-3c vibrational frequencies were scaled by a factor of0.95. The COSMO-RS continuum solvation model,20,54 asimplemented in COSMOtherm 2016,21 was used to obtain allsolvation enthalpies. For this purpose, single-point calcu-lations employing the default BP86 55/def-TZVP56 level oftheory were performed on the optimized gas-phase geometriesfor each molecule and the solvation contribution was added tothe gas-phase enthalpies.

Quantum theory of atoms in molecules (QTAIM) and non-covalent interaction (NCI) plots

For QTAIM analyses,27 the wavefunctions were recalculated byperforming single point calculations at the PBE057/SDD level(i.e. with the def2-TZVPP58 basis set on all elements plus asmall core Effective Core Potential (ECP) on Pd) on PBEh-3cgeometries. Wavefunction files (“.wfx” format) were generated,including additional atomic core electron densities to rep-resent the ECP-modeled electrons. These calculations wereperformed with the Gaussian09 (revision D0159) package.Then, the electron density (ρ), the Laplacian of the electrondensity (∇2ρ) and the bond ellipticity (ε) at bond critical pointswere computed, along with delocalization indexes (δ(A,B))between selected pairs of atoms. These calculations were per-formed with the AIMAll package (version 16.01.09).60

Combined graphical representations of QTAIM and NCI ana-lyses (shown in Fig. 6 and S30 in the ESI†) were performed atthe PBE0/SDD level on PBEh-3c optimized structures, usingthe AIMstudio module of AIMAll, by mapping the magnitudeof the reduced density gradient (so called |RDG|) isosurfacewith the function constructed as the electron density multi-

Fig. 14 Atom numbering schemes for NMR assignments in trans (left)and cis (right)-3d.

Fig. 15 Atom numbering schemes for NMR assignments in cis-3e.

Dalton Transactions Paper

This journal is © The Royal Society of Chemistry 2017 Dalton Trans., 2017, 46, 8125–8137 | 8135

Publ

ishe

d on

17

May

201

7. D

ownl

oade

d by

EC

OL

E P

OL

YT

EC

HN

IQU

E o

n 19

/04/

2018

15:

38:4

8.

View Article Online

plied by the sign of the second Hessian eigenvalue (referred toas sign(λ2)ρ) i.e. following the definition from Johnson and co-workers.§ 29

ETS-NOCV and EDA analyses

Computations were performed with methods of the densityfunctional theory using the SCM-ADF2014.01 package.61 ThePBE62 functional implemented in the Amsterdam DensityFunctional package61 (ADF2014 version) and augmented withGrimme’s DFT-D3(BJ) implementation of dispersion with aBecke–Johnson (BJ) damping function50 was used in theoptimizations of trans and cis-3a geometries. Preliminary geo-metry optimizations by energy gradient minimization werecarried out in all cases with integration grid accuracy com-prised between 4.5 and 7.5, an energy gradient convergencecriterion of 10−3 au and a tight to very tight SCF convergencecriterion. Counterpoise correction for the basis set superposi-tion error (BSSE) was neglected throughout this study. Unlessotherwise stated all computations were carried out usingscalar relativistic corrections within the zeroth order regularapproximation for relativistic effects with ad hoc all-electron(abbr. ae) single and quadruple polarization function tripleand quadruple-ζ Slater type basis sets (ae-TZP and ae-QZ4P).63

Extended Transition State-Natural Orbitals for ChemicalValence (ETS-NOCV) analyses32 as well as calculations ofvibrational modes were performed with optimized geometriesusing ADF2014 subroutines. Energy decomposition analyses30

were carried out using ADF2014 subroutines; it was noted thatthe quality of the basis set was not affecting (energy variationfrom TZP to QZ4P < 2%) the computed interfragment inter-action energies. Vibrational modes were analytically computedto verify that the optimized geometries were related to energyminima: statistical thermodynamic data at T = 298.15 K andP = 1 atm. Representations of molecular structures and isosur-faces were produced with ADFview2014.

Acknowledgements

We are grateful to the Agence Nationale de la Recherche, theDeutsche Forschungsgemeinshaft (COCOORDCHEM jointANR-DFG project), the LABEX “Chimie des SystèmesComplexes” and the Centre National de la RechercheScientifique. C. W. thanks the ANR (projectWEAKINTERMET-2DA) for financial support. S. D. thanks theDFG (project “Kohäsion in der Koordinationschemie”) for thefinancial support. N. S. thanks the HPC-Strasbourg, andCIMENT (project “liqsim”) platforms for allocation of CPU-time.

Notes and references

1 R. G. Pearson, Inorg. Chem., 1973, 12, 712–713.2 D. A. Redfield and J. H. Nelson, J. Am. Chem. Soc., 1974, 96,

6219–6220; J. J. MacDougall, J. H. Nelson and F. Mathey,Inorg. Chem., 1982, 21, 2145–2153.

3 J. Vicente, A. Arcas, M.-D. Galvez-Lopez, F. Julia-Hernandez,D. Bautista and P. G. Jones, Organometallics, 2008, 27,1582–1590; J. Vicente, A. Arcas, M.-D. Galvez-Lopez andP. G. Jones, Organometallics, 2006, 25, 4247–4259;J. Vicente, A. Arcas, D. Bautista and M. C. Ramirez DeArellano, J. Organomet. Chem., 2002, 663, 164–172;J. Vicente, A. Arcas, D. Bautista and P. G. Jones,Organometallics, 1997, 16, 2127–2138.

4 M. Crespo, J. Granell, X. Solans and M. Font-Bardia,J. Organomet. Chem., 2003, 681, 143–149.

5 J. Vicente, J.-A. Abad, A. D. Frankland and M. C. R. DeArellano, Chem. – Eur. J., 1999, 5, 3066–3075; M.-J. Oliva-Madrid, J.-A. Garcia-Lopez, I. Saura-Llamas, D. Bautistaand J. Vicente, Organometallics, 2014, 33, 33–39.

6 P. L. Alsters, H. T. Teunissen, J. Boersma, A. L. Spek andG. van Koten, Organometallics, 1993, 12, 4691–4696;A. Pietrzykowski, Pr. Nauk. – Politech. Warsz., Chem., 1991,52, 3–123; T. Janecki, J. A. D. Jeffreys, P. L. Pauson,A. Pietrzykowski and K. J. McCullough, Organometallics,1987, 6, 1553–1560.

7 A. F. M. J. Van der Ploeg, G. Van Koten and K. Vrieze,J. Organomet. Chem., 1981, 222, 155–174; A. F. M. J. Van derPloeg, G. Van Koten, J. E. J. Schmitz and J. G. M. Van derLinden, Inorg. Chim. Acta, 1982, 58, 53–58.

8 A. Berger, A. de Cian, J.-P. Djukic, J. Fischer and M. Pfeffer,Organometallics, 2001, 20, 3230–3240.

9 A. Berger, J.-P. Djukic, M. Pfeffer, J. Lacour, L. Vial, A. DeCian and N. Kyritsakas-Gruber, Organometallics, 2003, 22,5243–5260.

10 C. Arlen, M. Pfeffer, O. Bars and D. Grandjean, J. Chem.Soc., Dalton Trans., 1983, 1535–1544.

11 G. Longoni, P. Fantucci, P. Chini and F. Canziani,J. Organomet. Chem., 1972, 39, 413–425.

12 P. Jolliet, M. Gianini, A. von Zelewsky, G. Bernardinelli andH. Stoeckli-Evans, Inorg. Chem., 1996, 35, 4883–4888.

13 S. Grimme, J. G. Brandenburg, C. Bannwarth andA. Hansen, J. Chem. Phys., 2015, 143, 054107.

14 C. Werle, C. Bailly, L. Karmazin-Brelot, X.-F. Le Goff,L. Ricard and J.-P. Djukic, J. Am. Chem. Soc., 2013, 135,17839–17852.

15 C. Werle, M. Hamdaoui, C. Bailly, X.-F. Le Goff, L. Brelotand J.-P. Djukic, J. Am. Chem. Soc., 2013, 135, 1715–1718.

16 P. Petrović, J.-P. Djukic, A. Hansen, C. Bannwarth andS. Grimme, in Non-covalent Interactions in Synthesis and Designof New Compounds, ed. A. M. Maharramov, K. T. Mahmudov,M. N. Kopylovich and A. J. L. Pombeiro, John Wiley & Sons,Inc., Hoboken, New Jersey, 2016, pp. 113–143.

17 C. Werlé, L. Karmazin, C. Bailly, L. Ricard and J.-P. Djukic,Organometallics, 2015, 34, 3055–3064.

§We note that single points at the PBE-D3(BJ)/def2-TZVPP level on PBEh-3c opti-mized structures (see Table S3 in the ESI†) show that the contribution of thedamped – D3(BJ) dispersion correction (+0.7 kcal mol−1) to the overall relativeenergies of cis-3a vs. trans-3a (−5.5 kcal mol−1) is very small, and slightly stabil-izes trans-3a over cis-3a

Paper Dalton Transactions

8136 | Dalton Trans., 2017, 46, 8125–8137 This journal is © The Royal Society of Chemistry 2017

Publ

ishe

d on

17

May

201

7. D

ownl

oade

d by

EC

OL

E P

OL

YT

EC

HN

IQU

E o

n 19

/04/

2018

15:

38:4

8.

View Article Online

18 C. Werle, C. Bailly, L. Karmazin-Brelot, X.-F. Le Goff,M. Pfeffer and J.-P. Djukic, Angew. Chem., Int. Ed., 2014, 53,9827–9831.

19 J. P. Stambuli, Z. Weng, C. D. Incarvito and J. F. Hartwig,Angew. Chem., Int. Ed., 2007, 46, 7674–7677; J. P. Stambuli,C. D. Incarvito, M. Bühl and J. F. Hartwig, J. Am. Chem.Soc., 2004, 126, 1184–1194; D. V. Yandulov and N. T. Tran,J. Am. Chem. Soc., 2007, 129, 1342–1358; R. Gerber,O. Blacque and C. M. Frech, Dalton Trans., 2011, 40, 8996–9003; W. Clegg, G. R. Eastham, M. R. J. Elsegood,B. T. Heaton, J. A. Iggo, R. P. Tooze, R. Whyman andS. Zacchini, Organometallics, 2002, 21, 1832–1840;M. J. Ingleson, M. F. Mahon and A. S. Weller, Chem.Commun., 2004, 2398–2399; H. Braunschweig, K. Radacki,D. Rais and D. Scheschkewitz, Angew. Chem., Int. Ed., 2005,44, 5651–5654; C. E. Kefalidis and C. A. Tsipis, J. Comput.Chem., 2012, 33, 1689–1700.

20 F. Eckert and A. Klamt, AIChE J., 2002, 48, 369–385.21 F. Eckert and A. Klamt, in COSMOtherm, Version C3.0,

Release 16.01, COSMOlogic GmbH & Co. KG, Leverkusen,Germany, 2016.

22 T. E. Bitterwolf, R. Herzog and P. D. Rockswold,J. Organomet. Chem., 1987, 320, 197–209.

23 J. A. Casares, P. Espinet and G. Salas, Chem. – Eur. J., 2002,8, 4843–4853; M. A. Ortuño, S. Conejero and A. Lledós,Beilstein J. Org. Chem., 2013, 9, 1352–1382; M. D. Walter,P. S. White and M. Brookhart, New J. Chem., 2013, 37,1128–1133.

24 K. Tatsumi, R. Hoffmann, A. Yamamoto and J. K. Stille,Bull. Chem. Soc. Jpn., 1981, 54, 1857–1867.

25 A. D. Becke, J. Chem. Phys., 1993, 98, 5648–5652; C. Lee,W. Yang and R. G. Parr, Phys. Rev. B: Condens. Matter, 1988,37, 785–789; P. J. Stephens, F. J. Devlin, C. F. Chabalowskiand M. J. Frisch, J. Phys. Chem., 1994, 98, 11623–11627.

26 W. Hujo and S. Grimme, J. Chem. Theory Comput., 2011, 7,3866.

27 R. F. W. Bader, Atoms in Molecules, Oxford University Press,Oxford, 1990.

28 C. Werle, D. M. Anstine, L. Karmazin, C. Bailly, L. Ricardand J.-P. Djukic, Dalton Trans., 2016, 45, 607–617.

29 J. Contreras-Garcia, E. R. Johnson, S. Keinan, R. Chaudret,J. P. Piquemal, D. N. Beratan and W. Yang, J. Chem. Theor.Comput., 2011, 7, 625–632; E. Johnson, S. Keinan, P. Mori-Sanchez, J. Contreras-Garcia, A. J. Cohen and W. Yang,J. Am. Chem. Soc., 2010, 132, 6498–6506.

30 T. Ziegler and A. Rauk, Inorg. Chem., 1979, 18, 1755–1759.31 M. von Hopffgarten and G. Frenking, WIREs Comput. Mol.

Sci., 2012, 2, 43–62; F. M. Bickelhaupt and E. J. Baerends,in Reviews in Computational Chemistry, ed. K. B. Lipkowitzand B. D. Boyd, 2000, pp. 1–86.

32 M. P. Mitoraj, A. Michalak, T. Ziegler, F. M. Bickelhaupt andE. J. Baerends, J. Chem. Theory Comput., 2009, 5, 962–975;M. P. Mitoraj, A. Michalak, T. Ziegler, F. M. Bickelhaupt andE. J. Baerends, Organometallics, 2009, 28, 3727–3733.

33 A. Hansen, C. Bannwarth, S. Grimme, P. Petrović, C. Werléand J.-P. Djukic, ChemistryOpen, 2014, 3, 177–189.

34 J.-P. Djukic, K. Parkhomenko, A. Hijazi, A. Chemmi,L. Allouche, L. Brelot, M. Pfeffer, L. Ricard and X.-F. LeGoff, Dalton Trans., 2009, 2695–2711.

35 G. M. Sheldrick, Acta Crystallogr., Sect. A: Fundam.Crystallogr., 1990, 46, 467–473.

36 G. M. Sheldrick, Acta Crystallogr., Sect. A: Fundam.Crystallogr., 2008, 64, 112–122.

37 M86-E01078 APEX2 User Manual, Bruker AXS Inc., Madison,USA, 2006.

38 Kappa CCD Operation Manual, Nonius B. V., Delft: TheNetherlands, 1997.

39 A. L. Spek, J. Appl. Crystallogr., 2003, 36, 7–13.40 J.-P. Djukic, W. Iali, M. Pfeffer and X.-F. Le Goff, Chem. –

Eur. J., 2012, 18, 6063–6078.41 A. Mentes, R. D. W. Kemmitt, J. Fawcett and D. R. Russell,

J. Mol. Struct., 2004, 693, 241–246.42 J. Vicente, I. Saura-Llamas, M. G. Palin, P. G. Jones and

M. C. Ramirez de Arellano, Organometallics, 1997, 16, 826–833.43 G. E. Hartwell, R. V. Lawrence and M. J. Smas, J. Chem. Soc.

D, 1970, 912.44 R. Ahlrichs, M. Bär, M. Häser, H. Horn and C. Kölmel,

Chem. Phys. Lett., 1989, 162, 165–169.45 F. Weigend and R. Ahlrichs, Phys. Chem. Chem. Phys., 2005,

7, 3297–3305.46 K. Eichkorn, O. Treutler, H. Öhm, M. Häser and

R. Ahlrichs, Chem. Phys. Lett., 1995, 242, 652–660.47 F. Weigend, Phys. Chem. Chem. Phys., 2006, 8, 1057–1065.48 S. Grimme, J. Antony, S. Ehrlich and H. J. Krieg, Chem.

Phys. Lett., 2010, 132, 154104.49 A. D. Becke and E. R. Johnson, J. Chem. Phys., 2005, 123,

154101; E. R. Johnson and A. D. Becke, J. Chem. Phys.,2005, 123, 024101.

50 S. Grimme, S. Ehrlich and L. Goerigk, J. Comput. Chem.,2011, 32, 1456–1465.

51 P. Plessow, J. Chem. Theory Comput., 2013, 9, 1305.52 O. A. Vydrov and T. v. Voorhis, J. Chem. Phys., 2010, 133, 244103.53 S. Grimme, Chem. – Eur. J., 2012, 18, 9955–9964.54 A. Klamt, J. Chem. Phys., 1995, 99, 2224–2235.55 A. D. Becke, Phys. Rev. A, 1988, 38, 3098–3100; J. P. Perdew,

Phys. Rev. B: Condens. Matter, 1986, 33, 8822–8824.56 A. Schäfer, C. Huber and R. Ahlrichs, J. Chem. Phys., 1994,

100, 5829–5835.57 C. Adamo and V. Barone, J. Chem. Phys., 1999, 110, 6158.58 F. Weigend and R. Ahlrichs, Phys. Chem. Chem. Phys., 2005,

7, 3297–3305.59 J. A. Pople, et al., in Gaussian 09, Revision D.01, Gaussian,

Inc., Pittsburgh, PA, 2009 (full reference is given in the ESI†).60 T. A. Keith, in AIMAll (Version 15.09.27), TK Gristmill

Software, Overland Park KS, USA, 2016.61 Amsterdam Density Functional version 2014, SCM,

Theoretical Chemistry, Vrije Universiteit, Amsterdam, TheNetherlands, 2014.

62 J. P. Perdew, K. Burke and M. Ernzerhof, Phys. Rev. Lett.,1996, 77, 3865–3868.

63 E. Van Lenthe and E. J. Baerends, J. Comput. Chem., 2003,24, 1142–1156.

Dalton Transactions Paper

This journal is © The Royal Society of Chemistry 2017 Dalton Trans., 2017, 46, 8125–8137 | 8137

Publ

ishe

d on

17

May

201

7. D

ownl

oade

d by

EC

OL

E P

OL

YT

EC

HN

IQU

E o

n 19

/04/

2018

15:

38:4

8.

View Article Online