- Email: [email protected]
Adventures in divalent early transition metal coordination chemistry: On the way to metal–metal bonded species
Accepted Manuscript Review article Adventures in Divalent Early Transition Metal Coordination Chemistry: On the Way to Metal–Metal Bonded Species Carlos A. Murillo PII: DOI: Reference:
S0020-1693(16)30537-0 http://dx.doi.org/10.1016/j.ica.2016.12.024 ICA 17386
To appear in:
Inorganica Chimica Acta
Received Date: Revised Date: Accepted Date:
6 October 2016 1 December 2016 16 December 2016
Please cite this article as: C.A. Murillo, Adventures in Divalent Early Transition Metal Coordination Chemistry: On the Way to Metal–Metal Bonded Species, Inorganica Chimica Acta (2016), doi: http://dx.doi.org/10.1016/j.ica. 2016.12.024
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
1
Adventures in Divalent Early Transition Metal Coordination Chemistry: On the Way to Metal–Metal Bonded Species Carlos A. Murillo*,† Department of Chemistry, P. O. Box 3012, Texas A&M University, College Station, Texas 77842-3012 and Department of Chemistry, University of Texas at El Paso, El Paso, Texas 79968
Abstract During the search for appropriate early-transition-element starting materials to generate dinuclear paddlewheel species having metal–metal bonded cores of the type M24+, it was necessary to gain more insight into the relatively unexplored coordination chemistries of divalent vanadium, niobium and titanium. Here it is shown how an increased understanding of the chemical behavior eventually led to such species for vanadium and niobium while those of titanium remain elusive. The narrative chronicles some of the strategies behind the preparations and also how careful data analyses provided the blueprint for fully-designed syntheses. Even though the emphasis is on the syntheses, the figures show the structures of the compounds synthesized and structurally characterized in our laboratory in conjunction with Al Cotton.
*
To whom correspondence should be addressed. E-mail: [email protected]; [email protected] In honor of my good friend and spirited scientist Luis Echegoyen, an inspiring example of what can be accomplished by focusing on achieving a goal and excelling in advancing science. †
2
Keywords
metal–metal bonds, low valent, vanadium-, niobium-, tantalum-,
titanium-coordination compounds, triple bonds, tetragonal paddlewheels
1. General aspects It has been over half a century since the discovery of the quadruple bond in the Re2Cl82– anion depicted in Fig 1 [1]. This species has two rhenium atoms separated by ca. 2.21– 2.25 Å, a distance that is about 0.25 Å shorter than that in the rhenium metal. In this anion the four chlorine atoms bonded to each d4 metal atom are in an eclipsed configuration best described by an idealized D4h symmetry. If the metal–metal bond is aligned with the z axis, the eclipsed configuration may be explained as being due to the existence of a delta bond formed by lateral overlap of the corresponding d orbitals with x and y components (either from each of the two ݀௫ మି௬ మ or the two ݀௫௬ orbitals depending on how the Cartesian coordinates are defined [2,3].) The core of this diamagnetic species possesses a σ2π4δ2 electronic configuration, where the σ bond arises from overlap of the ݀௭ మ orbitals from each of the two metal atoms, the π orbitals originate from the overlap of the corresponding d orbitals with a z component (݀௫௭ and ݀௬௭ ). Because of the eclipsed configuration and the D4h symmetry one of the d orbitals with an x and y component must be used for metal–ligand bonding. Therefore, for this configuration, there is a maximum of four metal d-orbital overlaps capable of generating metal–metal bonding. To date a large number of quadruple bonded compounds have been synthesized having metal centers, M, for which M = Cr, Mo, W,
3
Tc and Re [4]. In many of these compounds, bidentate ligands such as carboxylates span the two transition metal atoms. This type of arrangement is commonly referred to as a tetragonal paddlewheel or simply as a paddlewheel. For such complexes the resultant idealized MO interaction diagram is given in Fig. 2. As electrons are added to the quadruple bonded manifold, these occupy antibonding orbitals and thus the bond order decreases from 4 when there are eight metal-base electrons available to a minimum of ½ when there is only one unpaired electron in species containing 15 metal-based electrons, e. g., in M25+ species, M = Ni [5] and Pt [6]. Similar decreases in bond order may be attained by removal of electrons. With the exception of chromium, only a handful of paddlewheel compounds containing first row elements have been synthesized. Similarly, only few such species with early transition elements are known [4]. The main reason for the small number of paddlewheel species with early transition elements is possibly the underdeveloped low-valent coordination chemistry and perhaps the high affinity for oxygen exhibited by the higher valent species. To date, notable exceptions of tetragonal paddlewheel compounds for elements in Groups 5–10 are those of Ta, Fe and Mn. There are, however, trigonal paddlewheels with Fe23+ cores and relatively short Fe–Fe distances of 2.20–2.23 Å [7]. One of them having the formula Fe2(DPhF)3 (DPhF = the anion of N,N′diphenylformamidine) is shown in Fig. 3. Other dinuclear compounds having iron–iron bonds are known but these will not be a focus of this manuscript. Some examples are Fe2(tim)2 (tim = 2,3,9,10-tetramethyl-1,4,8,11-tetraazacyclotetradeca-1,3,8,10-
4
tetraene) that has an unsupported Fe–Fe bond of 2.6869(6) Å [8,9]. Finally, it is appropriate to mention two low-coordinate dinuclear compounds bridged by two bulky guanidinates having very short iron–iron bonds. These have the formulas [Fe2{µ(guanidinate)2}2 where guanidinate represents the ligands denominated as Pipiso {Pipiso = (DipN)2C(cis-2,6-MeC5H8)} [10] or But2C=N)C(NDip) (Dip = 2,6diisopropylphenyl) which has a very short iron–iron bond of 2.1516(5) Å [11]. In this chronicle, it will be described how a sought-after systematic understanding of the coordination chemistry of vanadium allowed the synthesis of the first compound with a vanadium–vanadium triple bond in a species that has divalent metal atoms, six metal-based electrons and a V24+ core. Subsequently an entire family of such compounds was prepared and these were accompanied by related species having V23+ cores and a formal bond order of 3½. It will also be shown how an akin exploration of the divalent chemistry of Nb(II) allowed the syntheses of Nb24+ paddlewheel species. Finally, some attempts to synthesize Ti(II) species will be described. The narrative will highlight how basic knowledge of early transition metal coordination chemistry has allowed rational syntheses of metal–metal containing species. 2. Developing low-valent coordination chemistry of early transition elements 2.1. Vanadium It is often useful to apply lessons learned from the synthesis of previously isolated compounds to attempt preparation of related ones. However, this statement
5
has to be taken with a grain of salt since diverse elements, even those in the same group, often exhibit remarkable differences in chemical behavior. To prepare divanadium species with metal bonds it was necessary to analyze the relatively little explored chemistry of vanadium(II) which tended to produce highly air-sensitive species. It was long known that finely divided vanadium metal would react in acidic aqueous media to generate purple solutions containing the divalent aqueous cation [12]. Alternatively, reduction of higher-valent species in acidic media using amalgamated zinc (Jones reductor) would also produce such species. In this way, VSO4∙7H2O could be isolated [13,14]. This compound has a V(H2O)62+ cation (Fig. 4) and an interstitial water molecule involved in an elaborate and extensive hydrogen bonded network in which every hydrogen atom interacts with an atom acceptor provided by the anion or the interstitial water molecule. Because of this tight network, dry crystals are much more stable to air than the corresponding solutions. Exploration of the aqueous chemistry with anions such as saccharinate [13] (Fig. 5), pyridine-2-carboxylate (picolinate; pic} [15] (Fig. 6) also yielded species with extended hydrogen networks that provided considerable stability towards air to the crystalline materials that allowed handling of the compounds outside a drybox for several minutes without appreciable decomposition [16]. For convenience some of the ligands mentioned in the manuscript are illustrated in Scheme 1. Whenever the water molecules that contribute to the hydrogen bonding networks are removed the stability towards air significantly decreases as shown by
6
trans-V(sacch)2(py)4, V(bipy)2(η2-SO4), [V(en)3]SO4, V(pySO3)2(py)4 [17] (Figs 7, 8, 9 and 10, respectively) as well as related amine derivatives [18–28], all of which are easily oxidized even in the solid state. However, these compounds—all of which possess V–N bonds—are thermodynamically stable and remain unchanged as long as they are protected from oxygen. Additionally, early reactivity studies showed that water replacement by pyridine is quite facile. One of the first structurally characterized species of this type—described by Brauer and Krueger in 1973 and shown in Fig. 11— is trans-VCl2(py)4 [29]. Its preparation was initially cumbersome but eventually reaction of the very insoluble commercially available VCl2 solid with pyridine in a pressure vessel at 250 °C gave good yields of trans-VCl2(py)4 [30]. Interestingly, early attempts to synthesize V(II) compounds by reducing solutions of VCl3(THF)3 with zinc metal produced the dinuclear mixed-metal species [(THF)3V(µ-Cl)3V(THF)3][Cl2Zn(µ-Cl)2ZnCl2] [31,32]. It is important to note that all divalent species described so far are sixcoordinate with an idealized octahedral inner sphere environment. It should also be pointed out that part of the reason VCl2 is so insoluble is the extended crosscoordination to attain six-coordinate vanadium atoms but perhaps more importantly because of the d3 electronic configuration with the three unpaired electrons in the t2g orbitals for each metal atom. Thus there is some resemblance to the slow reaction rates of other octahedral species with half-filled or filled t2g orbitals such as Fe(II) and Cr(III).
7
Because many dinuclear compounds with metal–metal bonds contain carboxylate groups spanning the dimetal unit [4], several attempts were made to use either aqueous V(II) solutions or those in non-aqueous media such as trans-VCl2(py)4 in pyridine solutions with M(O2CR) reagents, M = Li, Na, but for the most part this lead to the isolation of oxo-centered trivanadium species of the type V3(µ3-O)(O2CR)6L3n+, n = 0, 1 and L = H2O, THF, py [33] shown in Fig. 12. It quickly became evident that this type of oxo-centered species represented a thermodynamic sink for such reactions. As mentioned, early work with divalent vanadium species showed that Ncontaining ligands such as pyridine, ethylenediamine, bipyridine, 3-pyridinesulphonate generated stable V(II)–N bonds. In this way, it was decided to explore reactions with ligands containing two donor nitrogen atoms such as N,N′-diarylformamidinates (ArNC(H)NAr– commonly abbreviated as DArF where Ar = an aryl or alkyl group). Initial reactions with starting materials such as [(THF)3V(µ-Cl)3V(THF)3][Cl2Zn(µ-Cl)2ZnCl2] led to intractable species. Therefore the use of VCl3(THF)3 with various reducing agents was pursued. Early reaction attempts showed that whenever the formamidinate ligands were added to the reaction mixture previous to the addition of the reducing agents, the very stable mononuclear compounds having the formula V(formamidinate)3n–, n = 0, 1, were the only isolable products (Fig. 13) [34]. Indeed, the structure of the neutral trischelated compound represents a ubiquitous motif for many trivalent transition metalcontaining species [35]. However, in the presence of excess reducing agents the [V(formamidinate)3]– monoanion is prevalent (Fig. 14) [36]. Importantly, whenever
8
VCl3(THF)3 was first treated with sodium triethylborohydride (NaHBEt3) in THF solution previous to the addition of the formamidinate ligands, isolation of paddlewheel compounds was finally realized. This reaction presumably generates trans-VCl2(THF)4. Upon addition of an appropriate lithium or sodium salt of an DArF ligand this divalent mononuclear species reacts to form a dinuclear species containing a V24+ core with a vanadium–vanadium triple bond having a σ2π4 electronic configuration (Fig. 15) [37,34]. The first isolated compound of this family has the formula V2(DArF)4, where Ar = tolyl, as well as a short V–V distance of 1.975(4) Å. The success of this procedure may be attributed to the cleanliness of the reduction using NaHBEt3 which generates byproducts such as H2 and (THF)BEt3 both of which are volatile and thus easily removed from the reaction mixture when placed under vacuum. Sodium chloride, also produced during the reaction (R1), is insoluble in THF and thus easily removed by filtration. 2VCl3(THF)3 + 2NaHBEt3 + 4NaDArF → V2(DArF)4 + 6NaCl + H2 + 2(THF)BEt3 + 3THF R1 Subsequently, another laboratory used the reaction of trans-VCl2(tmeda) (tmeda = N,N,N',N'-tetramethylethylendiamine) with lithium formamidinates such as Li(CyNC(H)NCy)—abbreviated as DCyF, where Cy is cyclohexyl—as a way of obtaining an analogous paddlewheel complex [38]. In this process the intermediate sixcoordinate mononuclear species, V(DCyF)2(tmeda), initially formed. The dinuclear V2(DCyF)4 paddlewheel forms upon heating.
9
It has been shown that R1 is generally appropriate for the syntheses of other triple bonded V24+ species with ligands such as those having Ar = phenyl (Ph; Fig. 16), anisyl (Ani; Fig. 17), p-chlorophenyl (ClPh; Fig. 18), N,N'-triphenylguanidinate (tpg; Fig. 19) and 2-anilinopyridinate (ap; Fig. 20) [36]. All of these compounds have paddlewheel structures devoid of axial interactions and V–V bond distances of less than 2.0 Å, being 1.979(1), 1.9876(5), 1.974(1), 1.9521(7) and 1.9295(8) Å, respectively. In addition, the bicyclic guanidinate compound V2(hpp)4 (hpp = the anion of 1,3,4,6,7,8-hexahydro-2Hpyrimido[1,2-a]pyrimidine) was prepared in a similar manner and it has a metal–metal bond length of 1.932(1) Å [39,40].The longest metal–metal bond distances in these diamagnetic compounds correspond to the formamidinates while the shorter ones belong to compounds having the more basic guanidinate and aminopyridinate ligands. It is noteworthy that the absence of axial ligation along the V–V bond gives raise to species with five-coordinate vanadium atoms. Interestingly, when an excess of a strong reductant such as KC8 is used instead of NaHBEt3 further reduction of V2(DPhF)4 takes place yielding [K(THF)3]V2(DPhF)4, shown in Fig. 21 [41]. The core of this paramagnetic paddlewheel species has a V23+ unit with a formal bond order of 3½. The increase in bond order relative to those in the V24+ analogues is reflected in a smaller bond distance of 1.9295(8) Å which is shorter by about 0.05 Å relative to that of to the parent V2(DPhF)4. A multiline EPR spectrum shows that the unpaired electron is coupled to each vanadium atoms and has a g of 1.9999 [42].
10
It is noteworthy that in [K(THF)3]V2(DPhF)4, the potassium atom is nestled between the formamidinate ligands and thus determining independently whether the observed shortening in the metal–metal distance was affected by such arrangement was an important consideration. To unequivocally judge whether such arrangement was relevant, a reaction with a crown ether was carried out to trap and extract the nestled potassium atom. In this way a free paddlewheel, shown in Fig. 22, was isolated in [K(18-crown-6)(THF)2]V2(DPhF)4 [36]. The V–V distance of 1.924(2) Å is essentially the same as that in [K(THF)3]V2(DPhF)4. Thus, it was demonstrated that the K interactions with the formamidinate ligands did not alter the V–V bond distances. 2.2. On the way to a Nb24+ paddlewheel Because niobium, like vanadium, is in Group 5, one may naively expect some similarities in reactivity to that of the lighter congener. However, it is also well documented that the heavier elements in any transition metal group also display significant differences, one of which is that mononuclear species are much less prevalent in the heavier congeners. Another significant difference between V and Nb is the absence of aqueous Nb(II) chemistry.12 However, initial work showed that similarly to vanadium, Nb–amine bonds are thermodynamically stable. For example, transNbCl2(py)4 is easily made by reacting trans-NbCl4(THF)2 with intercalated potassium in graphite (KC8) in a mixture of THF/py (Fig. 23) [43]. This reaction mixture produced a royal blue solution containing the mononuclear species (R2).
11
trans-NbCl4(THF)2 + 2KC8 + 4py → trans-NbCl2(py)4 + 2KCl + 16Cgraphite + 2THF R2 Upon filtration of the insoluble KCl salt and graphite byproducts, followed by solvent removal under vacuum, the solid was extracted using a mixture of ethyl ether having a few drops of pyridine. Upon cooling, crystals were obtained. The structure resembles that of the vanadium analogue being six-coordinate with an idealized octahedral environment as well as a trans arrangement of the chlorine atoms. This suggested that similarly to vanadium, there is an inherent stability of the Nb(II)–N bonds. Therefore an analogous reaction to that used to prepare V2(DArF)4 (R1) was an obvious path in attempting to make a Nb2(DArF)4 paddlewheel species. 2.3. A side trip to formamidinate cleavage promoted by tantalum species However, it was rapidly found that reactions of niobium or the heavier congener tantalum in the presence of formamidinates led predominantly to ligand cleavage or other ligand rearrangements (Scheme 2) [44–46]. In such reactions the cleavage generally occurs at a C–N bond. Depending on the reaction conditions, one or both of the arylimido (ArN2–) or arylformidoyl (ArNCH3–) fragments, were observed and these fragments were frequently incorporated into the isolated products. The electrons necessary for the cleavage can be accounted by the metal atom oxidation. An example is provided by R3: Ta2Cl6(SMe2)3 + 4LiDTolF → Cl2(η2-DTolF)TaV(µ-NTol)(η2-η2-HNCTol)TaV(η2-DTolF)2 + 3SMe2 + 4LiCl R3
12
The product Cl2(η2-DTolF)TaV(µ-NTol)(η2-η2-HNCTol)TaV(η2-DTolF)2 (Fig. 24) is diamagnetic. An 1H NMR study revealed the position of the H atom in the unique carbon atom of the tolylformidoyl ligand which supports its formulation. Additionally, since there are no d electrons in the tantalum(V) atoms there is no metal–metal bonding and the long Ta∙∙∙Ta of 3.0005(6) Å supports this assertion. In the presence of reducing agents, the hydrogen atom in the unique carbon atom of the arylformidoyl group can formally be removed to give a dinegative phenylisocyanide species commonly referred to as arylimidoyl—the charge of which was inferred from the bonding parameters. Examples of such species are provided by the reaction of Ta2Cl6(SMe3)3 with DArF (Ar = Ph) in the presence of Zn powder used as a reducing agent. For the phenyl analogue, the reaction may be described by R4: Ta2Cl6(SMe2)3 + 5LiDPhF + ½Zn → (η2-DPhF)2TaIV(µ-PhN)(μ2-η2-PhNC)TaIV(η2-DPhF)2 + 3SMe2 + 5LiCl + ½ZnCl2 + ½H2
R4
In an analogous reaction with the tolyl analogue employing sodium amalgam a somewhat related species having two bridging formamidinates instead of the corresponding chelating ones was observed, as shown in R5: Ta2Cl6(SMe2)3 + 5LiDTolF + Na/Hg → (η2-DTolF)TaIV(µ-DtolF)2(µ-TolN)(μ2-η2PhNC)TaIV(η2-DPhF)2 + 3SMe2 + 5LiCl + NaCl + Hg + ½H2
R5
The structure of the phenyl analogue which has a Ta–Ta bond distance of 2.9411(7) Å is provided in Fig. 25. In the first reaction both the arylimidoyl and
13
arylformidoyl fragments are present. Interestingly, the fashion in which both the arylformidoyl and arylimidoyl (ArNCH3–and ArNC2–) ligand bridge the metal atoms is quite different. Even though they both span the tantalum atoms, for ArNC2– the arrangement is quite unsymmetrical with the ligand bonding in a μ2-η2-fashion while for ArNCH3– it is more symmetrical exhibiting a η2-η2-fashion as shown in Scheme 3. 2.4. Back to niobium Therefore, if a paddlewheel were to be made, it was necessary to find an alternative route that would prevent such cleavage. This is where the bicyclic guanidinate hpp ligand shown in Scheme 1 came into play. Because of the support provided to the C–N bond by other ring bonds, this ligand was projected to be stable towards cleavage. Furthermore, it had been shown that it could effectively form paddlewheel complexes such as those in M2(hpp)4, M = V, Cr, Mo,39 and Ru2(hpp)4Cl2 [47]. Using this strategy, Nb2(hpp)4 was isolated from the reaction in THF of NbCl3(DME) (DME = dimethoxyethane) with hpp and the reducing agent KC8 [48], as shown in R6. 2NbCl3(DME) + 4Lihpp + 2KC8 → Nb2(hpp)4 + 2KCl + 4LiCl + 16Cgraphite + DME R6 This compound has a Nb–Nb bond distance of 2.2035(9) Å (Fig. 26). Remarkably, the structure, as well as the metal–metal distance, had been predicted in the early days of density functional theory before the compound was synthesized [49]. With the knowledge that shielding the C–N–C moiety, other ligands with protected NC bonds were investigated and in this way, 7-azaindole (azin) was used to synthesize Nb2(azin)4
14
(Fig. 27) [50]. This compound crystallizes along with solvated LiCl producing a dinuclear [Li(THF)2Cl]2 moiety whose chlorine atoms occupy axial positions along the Nb–Nb bond at the relatively long distance of 2.849(3) Å from the niobium atom. Nevertheless, this weak interaction is enough to lengthen the Nb–Nb distance by about 0.07 Å to 2.278(2) Å. It is possible that such lengthening may be partly due to the larger increased bite angle of the azin ligand relative to that of hpp. When an excess of the reducing agent NaEt3BH was used instead of KC8 in R6, complexes with composition Nb2(hpp)4∙NaEt3BH and Nb2(hpp)4∙2NaEt3BH were isolated (Figs. 28 and 29, respectively) [50]. In these species the sodium atom is nestled within the hpp ligands and there is no direct interaction of the anion with the dinuclear species and thus the Nb–Nb distance is unaffected being 2.2064(3) Å and 2.2187(7) Å for Nb2(hpp)4∙NaEt3BH and Nb2(hpp)4∙2Et3BH, respectively. This behavior resembles that of [K(THF)3]V2(DPhF)4 (vide supra) [36]. Such reactions employing the guanidinate ligand hpp as well as azin demonstrate that the reaction scheme involving the strategy of protecting the ligand from cleavage is general and that the triple bonded Nb24+ unit is thermodynamically stable. An unusual compound with an Nb26+ core is obtained by oxidizing Nb2(hpp)4 with FeCp2PF6 in DME. This reaction generates [Nb2(η2-hpp)4(µ-η2,η2-hpp]PF6 (Fig. 30) [51]. In this compound there are four chelating hpp groups as well as a symmetrical hpp unit positioned essentially perpendicular to the Nb–Nb bond. In this compound the
15
metal atoms are separated by 2.6014(8) Å; a distance similar to that found in a related Ti26+ species (vide infra). The presence of a fifth hpp ligand suggests other products are also formed but they were not isolated. It is worth mentioning that a double bonded Nb26+ unit with formamidinate groups has been isolated by reacting Nb2(µ-SMe)3Cl6 with NaEt3BH. This reaction generates Nb2(µ-SMe)2(µ-DTolF)2(η2-DTolF)2 (DTolF = the anion of N,N'-di-ptolylformamidine) (Fig. 31) [52]. In this compound the Nb–Nb distance of 2.665(2) Å is significantly longer (by about 0.46 Å) than that in the triple bonded Nb24+ species. Unfortunately, thus far tantalum paddlewheel complexes analogous to those of Nb have not yet been isolated, even though the nononuclear mer-TaCl3(py)3 (Fig. 32) has shown that Ta–N bonds should be stable [53]. This compound has been isolated in three crystalline forms. 2.3. Some adventures pursuing Ti(II) coordination species An exploration in our laboratories, showed that the structure of trans-TiCl2(py)4 is analogous to that of the V and Nb species (Fig. 33) [43]. As with its analogues, the coordination sphere is octahedral with two trans chlorine atoms. This species has recently been the topic of further studies including theoretical computations, EPR analyses and a structural redetermination that confirmed that reported earlier [54].
Reactions with the ligand hpp gave surprising results that again showed that the chemistry of each element is unique and few generalizations are possible. By reducing
16
the pseudo octahedral species Ti(η2-hpp)2Cl2 with excess KC8 in THF the unusual dinuclear compound Cl(η2-hpp)Ti(µ-η2,η2-hpp)2Ti(η2-hpp)Cl species (Fig. 34) was isolated. The arrangement of the µ-η2,η2-hpp unit is reminiscent of that in [Nb2(η2hpp)4(µ-η2,η2-hpp2]PF6 (vide supra) [51]. Similarly to the triple bonded V24+ and Nb24+ hpp-compounds, this species contains a Ti26+ unit but the arrangement of the four hpp ligands is quite unusual and not that of a paddlewheel. The Ti–Ti distance of 2.5944(9) Å is quite long. Nevertheless, DFT calculations suggest that there is a single bond with a σ2 electronic configuration. Solution studies by 1H NMR spectroscopy indicate that the solution structure and that in the solid may be different, with that in solution being much more symmetrical. One could speculate that the more symmetrical solution species may be a paddlewheel. Interestingly, a reaction of the bulky amido-containing compound Ti[NButAr]3 (Ar = 3,5-C6H3Me2 or Ar' = C6H5 and But = tert-butyl) with carbon dioxide at –40 °C produces a diamagnetic Ti26+ paddlewheel with the formula of Ti2[µO2CN(But)Ph]4(NButPh)2 (Fig. 35) [55]. This compound has four carbamate ligands spanning the Ti26+ unit, thus forming a paddlewheel with two axial amido ligands (NButPh). However the Ti∙∙∙Ti separation of over 3.4 Å is rather long and precludes the existence of a metal–metal bond. It is clear that the strongly bound axial ligands (Ti–N = 1.920(3) Å) compete with the metal–metal formation. These assertions are supported by DFT calculations.
17
Nonetheless, a single bonded tri-titanium unit was obtained by reducing titanium tetrachloride or TiCl3(THF)3 in the presence of phosphines and Grignard reagents. It was found that reaction of TiCl4(dmpe) (dmpe = 1,2bis(dimethylphosphino)ethane) in toluene with ButMgCl in diethyl ether (R7) produces a red-brown solution from which the triangular species [Ti(µ-Cl)Cl(dmpe)]3 crystallizes (Fig. 36) [56]. 3TiCl4(dmpe) + 6ButMgCl → [Ti(µ-Cl)Cl(dmpe)]3 + 6MgCl2 + 3But– But
R7
This compound represents the first species with a genuine Ti–Ti single bond and an example of a relatively thermally stable—though pyrophoric in air—divalent titanium compound in the solid state. Nevertheless, solutions did decompose rather quickly at ambient temperature.
An interesting dinuclear species with a short Ti–Ti bond of 2.399(2) Å has a sandwich structure with a bispentalene unit above and below the Ti2 unit and a formula Ti2(µ:η5,η5-Pn)2 (Pn = C8H4{SiPri3-1,4}2) [57]. DFT computations suggest there is a Ti=Ti with σ and π components.
Other dinuclear compounds with trivalent titanium are calix[4]arene derivatives such as K4(THF)8[Ti2(µ-NMe2)2(calix[4]arene)2] and Na2(THF)6[Ti2(calix[4]arene)2] (Figs. 37 and 38, respectively) [58]. However, the long TiIII···TiIII separations of 3.278(2) and 3.133(1) Å, respectively, preclude the existence of a metal–metal bond.
18
3. Concluding remarks While attempting preparation of metal–metal bonded compounds of diverse transition metal species, searching for parallels from those of other metal atoms is sometimes a useful exercise to guide the new synthetic procedures. However, more often than not completely new routes must be developed. What is evident from the narrative is that understanding the complexity of the coordination chemistry of a given element is essential in guiding the development of new processes. Often driving on country roads, identifying dead ends before reaching them and making adjustments is essential to find the highway that eventually leads to a targeted destination. This is similar to what a winding, sinuous river does on its way to the ocean. It does not matter how many turns and twists the stream makes, the water always reaches its target. It is hoped that the insights presented here will help students understand the intricacies of developing new fields and help them learn to appreciate that there is no such thing as a “the reaction did not work”. There are always positive aspects that an inquisitive mind can extract and then use to attain the final goals. Acknowledgments This work was supported by Texas A&M University as well as IR/D support from the National Science Foundation. The author also thanks the multiple co-workers responsible for the advance of this field whose names are provided in the references. With this short review, the author cheers the friendship with Luis Echegoyen as well as his wife Lourdes. With Luis I have had the privilege of knowing him for many years
19
before working together at the National Science Foundation and now we continue to strengthen the friendship by collaborating at the University of Texas at El Paso.
References
[1] F.A. Cotton, N.F. Curtis, C.B. Harris, B.F.G. Johnson, S.J. Lippard, J.T. Mague, W.R. Robinson, J.S. Wood, Science 145 (1964) 1305–1307. [2] For a historical account on the δ bond, see: F.A. Cotton, Pure. Appl. Chem. 64 (1992) 1383–1393. [3] C.A. Murillo, Comments Inorg. Chem. 35 (2015) 39–58. [4] F.A. Cotton, C.A. Murillo, R.A. Walton (Eds.), Multiple Bonds between Metal Atoms, third ed. Springer Science and Business Media, Inc., New York, 2005. [5] J.F. Berry, E. Bothe, F.A. Cotton, S.A. Ibragimov, C.A. Murillo, D. Villagrán, X. Wang, Inorg. Chem. 45 (2006) 4396–4406. [6] F.A. Cotton, J.H. Matonic, C.A. Murillo, Inorg. Chim. Acta 264 (1997) 61–64. [7] F.A. Cotton, L.M. Daniels, L.R. Falvello, J.H. Matonic, C.A. Murillo, Inorg. Chim. Acta 256 (1997) 269– 275. [8] C.R. Hess, T. Weyhermüller, E. Bill, K. Wieghardt, Angew. Chem. 48 (2009) 3703–3706. [9] C.A. Murillo, Angew. Chem. 48 (2009) 5076–5077. [10] L. Fohlmeister, S. Liu, C. Schulten, B. Moubaraki, A. Stasch, J.D. Cashion, K.S. Murray, L. Gagliardi, C. Jones, Angew. Chem. 51 (2012) 8294–8298. [11] A.K. Maity, A.J. Metta-Magaña, S. Fortier, Inorg. Chem 54 (2015) 10030–10041. [12] F.A. Cotton, G. Wilkinson, C.A. Murillo, M. Bochmann, Advanced Inorganic Chemistry, sixth ed., John Wiley and Sons, New York, 1999. [13] F.A. Cotton, L.R. Falvello, R. Llusar, E. Libby, C.A. Murillo, W. Schowtzer, Inorg. Chem. 25 (1986) 3423–3428. [14] F.A. Cotton, L.R. Falvello, C.A. Murillo, I. Pascual, A.J. Schultz, M. Tomás, Inorg. Chem. 33 (1994) 5391–5395. [15] F.A. Cotton, L.M. Daniels, M.L. Montero, C.A. Murillo, Polyhedron 11 (1992) 2767–2774. [16] F.A. Cotton, E. Libby, C.A. Murillo, G. Valle, Inorg. Synth. 27 (1990) 306–310. [17] L.M. Daniels, C.A. Murillo, K.G. Rodríguez, Inorg. Chim. Acta 229 (1995) 27–32. [18] For example, see: A.C. Niedwieski, P.B. Hitchcock, J.D. Da Motta Neto, F. Wypych, G.F. Leigh, F.S. Nunes, J. Brazil. Chem. Soc. 14 (2003) 750–758. [19] A.C. Niedwieski, G.F. Leigh, T. Hasegawa, F.S. Nunes, Acta Crystallogr. E59 (2003) m939–m941. [20] A.C. Raimond, T. Hasegawa, F.S. Nunes, Acta Crystallogr. E60 (2004) m1010–m1012. [21] D.M. Halepoto, L.F. Larkworthy, S.K. Sengupta, Transit. Metal Chem. 28 (2003) 800–802. [22] J.G. Reynolds, S.C. Sendlinger, A.M. Murray, J.C. Huffman, G. Christou, Inorg. Chem. 34 (1995) 5745– 5752. [23] P. Gosh, H. Taube, T. Hasegawa, R. Kuroda, Inorg. Chem. 34 (1995) 5761–5775. [24] J.J.H. Edema, S. Gambarotta, A. Meetsma, A.L. Spek, N. Veldman, Inorg. Chem. 30 (1991) 2062–2066. [25] J.J.H. Edema, W. Stouthamer, F. Van Bothuis, S. Gambarotta, W.J.J. Smeets, A.L. Spek, Inorg. Chem. 29 (1990) 1302–1306. [26] F.A. Cotton, R. Poli, Inorg. Chim. Acta 141 (1988) 91–98. [27] L.F. Larkworthy, B.J. Tucker, Inorg. Chim. Acta 33 (1979) 167–170. [28] M.M. Khamar, L.F. Larkworthy, K.C. Patel, G. Beech, Austr. J. Chem. 27 (1974) 41–51.
20
[29] D.J. Brauer, C. Krueger, Cryst. Struct. Commun. 3 (1973) 421–426. [30] F.A. Cotton, C. A. Murillo, Ing. Cienc. Quim. 9 (1985) 5–6. [31] F.A. Cotton, S.A. Duraj, M.W. Extine, G.E. Lewis, W.J. Roth, C.D. Schmulbach, W. Schwotzer, J. Chem. Soc., Chem. Commun. (1983) 1377–1378. [32] R.J. Bouma, J.H. Teuben, W.R. Beukema, R.L. Bansemer, J.C. Huffman, K.G. Caulton, Inorg. Chem. 23 (1984) 2715–2718. [33] F.A. Cotton, M.W. Extine, L.R. Falvello, D.B. Lewis, G.E. Lewis, C.A. Murillo, W. Schowtzer, M. Tomás, Inorg. Chem. 25 (1986) 3505–3512. [34] F.A. Cotton, L.M. Daniels, C.A. Murillo, Inorg. Chem. 32 (1993) 2881–2885. [35] F.A. Cotton, L.M. Daniels, D.J. Maloney, C.A. Murillo, Inorg. Chim. Acta 242 (1996) 31–42. [36] F.A. Cotton, E.A. Hillard, C.A. Murillo, X. Wang, Inorg. Chem. 42 (2003) 6063–6070. [37] F.A. Cotton, L.M. Daniels, C.A. Murillo, Angew. Chem., Int. Ed. Engl. 31 (1992) 737–738. [38] S. Hao, P. Berno, R.K. Minhas, S. Gambarotta, Inorg. Chim. Acta 244 (1996) 37–49. [39] F.A. Cotton, D.J. Timmons, Polyhedron 17 (1997) 179–184. [40] C. Fernández-Cortabitarte, F. Gacía, J.V. Morey, M. McPartlin, S. Singh, A.E.H. Wheatley, D.S, Wright, Angew. Chem. Int. Ed. 46 (2007) 5425–5427. [41] F.A. Cotton, E.A. Hillard, C.A. Murillo, J. Am. Chem. Soc. 125 (2003) 2026–2027. [42] For a discussion on the usefulness of EPR spectroscopy in dimetal species, see: N.S. Dalal, C.A. Murillo, Dalton Trans. 43 (2014) 8565–8576. [43] M.A. Araya, F.A. Cotton, J.H. Matonic, C.A. Murillo, Inorg. Chem. 34 (1995) 5424–5428. [44] F.A. Cotton, L.M. Daniels, C.A. Murillo, X. Wang, Inorg. Chem. 36 (1997) 896–901. [45] F.A. Cotton, J.H. Matonic, C.A. Murillo, X. Wang, Bull. Soc. Chim. Fr. 133 (1996) 711–720. [46] F.A. Cotton, L.M. Daniels, J.H. Matonic, X. Wang, C.A. Murillo, Polyhedron 16 (1997) 1177–1191. [47] J.L. Bear, Y. Li, B. Han, K.M. Kadish, Inorg. Chem. 35 (1996) 1395–1398. [48] F.A. Cotton, J.H. Matonic, C.A. Murillo, J. Am. Chem. Soc. 119 (1997) 7889–7890. [49] F.A. Cotton, X. Feng, J. Am. Chem. Soc. 119 (1997) 7514–7520. [50] F.A. Cotton, J.H. Matonic, C.A. Murillo, J. Am. Chem. Soc. 120 (1998) 6047–6052. [51] F.A. Cotton, S.A. Ibragimov, C.A. Murillo, P.V. Poplaukhin, Q. Zhao, J. Molec. Struct. 890 (2008) 3–8. [52] F.A. Cotton, J.H. Matonic, C.A. Murillo, J. Clust. Sci. 7 (1996) 655–662. [53] F. A. Cotton, C.A. Murillo, X. Wang, Inorg. Chim. Acta 245 (1996) 115–118. [54] G.B. Wijeratne, E.M. Zolnhofer, S. Fortier, L.N. Grant, P.J. Carroll, C-H. Chen, K. Meyer, J. Krzystek, A. Ozarowski, T.A. Jackson, D.J. Mindiola, J. Telser, Inorg. Chem. 54 (2015) 10380–10397. [55] A. Mendiratta, C.C. Cummins, F.A. Cotton, S.A. Ibragimov, C.A. Murillo, D. Villagrán, Inorg. Chem. 45 (2006) 4328–4330. [56] F.A. Cotton, C.A. Murillo, M. Petrukhina, J. Organomet. Chem. 573 (1999) 78–86. [57] A.F.R. Kilpatrick, J.C.C. Green, F.G.N. Cloke, N. Tsoureas, Chem. Commun. 49 (2013) 9434–9436. [58] F.A. Cotton, E.V. Dikarev, C.A. Murillo, M.A. Petrukhina, Inorg. Chim. Acta 332 (2002) 41–46.
21
Fig. 1. The structure of the quadruple bonded Re2Cl82– anion.
Fig. 2. Idealized molecular orbital diagram for a dimetal species with D4h symmetry. In a species with metal atoms with d4 electronic configurations there are eight metal-based electrons that fill the four bonding orbitals given a σ2π4δ2 electronic configuration and a bond order of four. For simplicity the d orbitals are drawn at arbitrary levels. Note that one of the d orbitals with x and y components is used in metal-to-ligand bonding and thus it is unavailable for metal-tometal bonding.
22
Fig. 3. The core structure of the trigonal paddlewheel compound Fe2(DPhF)3 that has an Fe23+ moiety. DPhF = the anion of N,N'-diphenylformamidine.
Fig. 4. The structure of [V(H2O)6]SO4∙H2O. All H atoms, including those of the interstitial water molecule, are engaged in H-bonding. For simplicity this network is not shown.
23
Fig. 5. The N-bonded saccharinate compound [trans-V(C7H4NO3S)2(H2O)4]∙2H2O. Only one of the symmetry related interstitial water molecules is shown. There is an extensive hydrogen bonding network (not shown) with all water-hydrogen atoms being involved.
Fig. 6. A carboxylate derivative trans-V(pic)2(H2O)2 where pic = pyridine-2-carboxylate (picolinate).
24
Fig. 7. The structure of trans-[V(C7H4NO3S)2(py)4]∙2py. Note that as the water molecules are substituted by pyridine the bonding mode of the saccharinate ligands changes from N-bonded to O-bonded. For comparison, see Fig. 5.
Fig. 8. The mononuclear compound V(bipy)2(η2-SO4).
Fig. 9. The tris(ethylenediamine) compound [V(en)3]SO4.
25
Fig. 10. The compound trans-V(pySO3)2(py)4 where pySO3 = 3-pyridinesulphonate.
Fig. 11. An early structure— trans-[VCl2(py)4]∙THF.
Fig. 12. A few examples of oxo-centered trinuclear compounds: (i) [V3(µ3O)(O2CCHCl)6(H2O)3](CF3SO3) having three trivalent vanadium atoms; (ii) [V3(µ3O)(O2CCH3)6(THF)3]2[Cl2OV(µ-Cl)2VOCl2] with three V(III) atoms and (iii) [V3(µ3O)(O2CCH3)6(py)3]2 having one divalent and two trivalent vanadium atoms.
26
Fig. 13. The mononuclear V(DTolF)3 compound having a trivalent vanadium atom and three chelating formamidinate anions. DTolF = the anion of N,N′-di-p-tolylformamidine.
Fig. 14. [Li(THF)4]V2(DCl2PhF)3 having a divalent vanadium atom. DCl2PhF = the anion of N,N′-di3,5-dichloroformamidine.
27
Fig. 15. The triple bonded V2(DTolF)4 compound where DTolF is the anion of N,N′-di-ptolylformamidine. The V–V bond distance is 1.978(2) Å.
Fig. 16. The paddlewheel compound V2(DPhF)4 where DPhF = N,N′-diphenylformamidinate. The V–V bond distance is 1.979(1) Å.
28
Fig. 17. The structure of V2(DAniF)4 where DAniF = N,N′-di-p-anisylformamidinate. This ligand has an important advantage over analogous formamidinates in that typically its compounds are more soluble than analogues without the methoxy group. The V–V bond distance is 1.9876(5) Å.
Fig. 18. The structure of V2(DClPhF)4 where DClPhF = N,N′-di-p-chloroformamidinate. The V–V bond distance is 1.974(1) Å.
29
Fig. 19. The structure of the guanidinate derivative V2(tpg)4, where tpg = 1,2,3triphenylguanidinate. The V–V bond distance is 1.9521(7) Å.
Fig. 20. The structure of V2(ap)4 where ap = 2-anilinopyridinate. The V–V bond distance is 1.9425(4) Å.
30 Fig. 21. An unusual species having a V23+ core and a formal metal–metal bond order of 3½: [K(THF)3]V2(DPhF)4. Note that the potassium atom is nestled within phenyl rings of the formamidinate ligands and bound to the oxygen atom of the three THF molecules. The V–V bond distance is 1.9295(8) Å.
Fig. 22. The compound [K(18-crown-6)(THF)2]V2(DPhF)4 having a V23+ core with a vanadium– vanadium bond order of 3½. The V–V bond distance is 1.924(2) Å.
Fig. 23. The structure of the THF solvate of trans-[NbCl2(py)4].
31
Fig. 24. The structure of Cl2(η2-DTolF)TaV(µ-NTol)(η2-η2-HNCTol)TaV(η2-DTolF)2. Note that both the p-tolylamido and p-tolylformidoyl fragments bridge the metal atoms.
Fig. 25. The structure of (η2-DPhF)2TaIV(µ-PhN)(μ2-η2-PhNC)TaIV(η2-DPhF)2. Only one orientation of a disordered PhCN2– group related by a 2-fold axis is shown. The structure resembles that of the tolyl analogue.
32
Fig. 26. The triple bonded Nb2(hpp)4 where hpp = the anion of 1,3,4,6,7,8-hexahydro-2Hpyrimido[1,2a]pyrimidine. The Nb–Nb distance is 2.2035(9) Å.
Fig. 27. The structure of Nb2(azin)4∙2LiCl∙4THF in where azin = 7-azaindole. The Nb–Nb distance is 2.278(2) Å. For simplicity the lithium atoms and THF molecules are not shown.
Fig. 28. The structure of Nb2(hpp)4∙NaHBEt3. The Nb–Nb distance is 2.2064(3) Å. Note that the sodium interaction resembles that of potassium in [K(THF)3]V2(DPhF)4 in Figure 21.
33
Fig. 29. The compound Nb2(hpp)4∙2NaHBEt3. The Nb–Nb distance is 2.2187(7) Å. For simplicity only one of the interstitial NaHBEt3 molecules is shown.
Fig. 30. The compound [Nb2(η2-hpp)4(µ-η2-η2-hpp)]PF6 having a Nb26+ core where hpp = the anion of 1,3,4,6,7,8= hexahydro-2H-pyrimido[1,2a]pyrimidine. The Nb–Nb distance is 2.6014(8) Å.
Fig. 31. The structure of [(η2-DTolF)Nb(µ-SMe2)(µ-DTolF)]2 where DTolF = N,N′-di-ptolylformamidinate. The Nb–Nb distance is 2.655(2) Å.
34
Fig. 32. The trivalent, nononuclear mer-TaCl3(py)3 compound.
Fig. 33. The mononuclear THF solvate of trans-[TiCl2(py)4].
Fig. 34. The trivalent titanium species TiIII2(η2-hpp)2(µ-η2-η2-hpp)2Cl2 where hpp = the anion of 1,3,4,6,7,8-hexahydro-2H-pyrimido[1,2a]pyrimidine. The Ti–Ti distance is 2.5944(9) Å.
35
Fig. 35. The tetracarbamate paddlewheel species TiIII2[µ-O2CN(But)Ph]4(NButPh)2∙Et2O which is devoid of metal–metal bonding; the long Ti∙∙∙Ti separation is 3.515(1) Å. The amido NButPh groups occupy axial positions along the Ti∙∙∙Ti axis.
Fig. 36. The trinuclear compound [TiII(µ-Cl)Cl(dmpe)]3 where dmpe = dimethylphosphinoethane. The Ti–Ti single bond length is 2.827(3) Å.
Fig. 37. The structure of K4(THF)8[TiIII2(µ-NMe2)2L2] where H4L = C28H20O4H4; calix[4]arene. The Ti∙∙∙Ti separation of 3.28 Å is too long to result in significant Ti–Ti bonding.
36
Fig. 38. The dinuclear species [Na(THF)3]2[TiIIIL2]2∙2THF where H4L = C28H20O4H4; calix[4]arene. The Ti∙∙∙Ti separation is 3.133(1) Å.
Scheme 1. A diagrammatic representation of some of the ligands mentioned in the text showing the corresponding abbreviations.
37
Scheme 2. How a formamidinate may be fragmented by a transition metal-containing species having Ta or Nb producing arylimidoyl (ArN2–) and arylformidoyl (ArNCH3–) anions. The electrons are provided by concomitant oxidation of the metal atom.
Scheme 3. Schematic representation of the bonding of the ArNC2– fragment to a ditantalum species (left) and that of the HCNAr3– anion (right).
38
Synopsis
39
Highlights • The coordination chemistry of some low valent early metal complexes is presented • A roadmap to the syntheses of paddlewheel compounds with V–V triple bonds is given • The syntheses of paddlewheel compounds with metal–metal bonds is described • The preparation of diniobium compounds is presented • Some adventures in low valent titanium are described • Lessons learned on similarities and differences in their chemistries are offered
40 Carlos A. Murillo studied chemistry at the University of Costa Rica and Texas A&M University where he received his Ph. D. in 1973 with F. Albert Cotton. He then did postdoctoral work with Malcolm H. Chisholm at Princeton University. He went back to Costa Rica where he quickly moved to the rank of Professor. In 1991 he moved to Texas A&M as director of the Laboratory for Molecular Structure and Bonding and in 2007 he took a position as program director in the Division of Chemistry at the US National Science Foundation. He has continued his research as an adjunct professor at Texas A&M University and the University of Texas at El Paso. He is a Fellow of the American Association for the Advancement of Science (AAAS) and a charter member of the Costa Rican Academy of Sciences.
41
S0020-1693(16)30537-0 http://dx.doi.org/10.1016/j.ica.2016.12.024 ICA 17386
To appear in:
Inorganica Chimica Acta
Received Date: Revised Date: Accepted Date:
6 October 2016 1 December 2016 16 December 2016
Please cite this article as: C.A. Murillo, Adventures in Divalent Early Transition Metal Coordination Chemistry: On the Way to Metal–Metal Bonded Species, Inorganica Chimica Acta (2016), doi: http://dx.doi.org/10.1016/j.ica. 2016.12.024
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
1
Adventures in Divalent Early Transition Metal Coordination Chemistry: On the Way to Metal–Metal Bonded Species Carlos A. Murillo*,† Department of Chemistry, P. O. Box 3012, Texas A&M University, College Station, Texas 77842-3012 and Department of Chemistry, University of Texas at El Paso, El Paso, Texas 79968
Abstract During the search for appropriate early-transition-element starting materials to generate dinuclear paddlewheel species having metal–metal bonded cores of the type M24+, it was necessary to gain more insight into the relatively unexplored coordination chemistries of divalent vanadium, niobium and titanium. Here it is shown how an increased understanding of the chemical behavior eventually led to such species for vanadium and niobium while those of titanium remain elusive. The narrative chronicles some of the strategies behind the preparations and also how careful data analyses provided the blueprint for fully-designed syntheses. Even though the emphasis is on the syntheses, the figures show the structures of the compounds synthesized and structurally characterized in our laboratory in conjunction with Al Cotton.
*
To whom correspondence should be addressed. E-mail: [email protected]; [email protected] In honor of my good friend and spirited scientist Luis Echegoyen, an inspiring example of what can be accomplished by focusing on achieving a goal and excelling in advancing science. †
2
Keywords
metal–metal bonds, low valent, vanadium-, niobium-, tantalum-,
titanium-coordination compounds, triple bonds, tetragonal paddlewheels
1. General aspects It has been over half a century since the discovery of the quadruple bond in the Re2Cl82– anion depicted in Fig 1 [1]. This species has two rhenium atoms separated by ca. 2.21– 2.25 Å, a distance that is about 0.25 Å shorter than that in the rhenium metal. In this anion the four chlorine atoms bonded to each d4 metal atom are in an eclipsed configuration best described by an idealized D4h symmetry. If the metal–metal bond is aligned with the z axis, the eclipsed configuration may be explained as being due to the existence of a delta bond formed by lateral overlap of the corresponding d orbitals with x and y components (either from each of the two ݀௫ మି௬ మ or the two ݀௫௬ orbitals depending on how the Cartesian coordinates are defined [2,3].) The core of this diamagnetic species possesses a σ2π4δ2 electronic configuration, where the σ bond arises from overlap of the ݀௭ మ orbitals from each of the two metal atoms, the π orbitals originate from the overlap of the corresponding d orbitals with a z component (݀௫௭ and ݀௬௭ ). Because of the eclipsed configuration and the D4h symmetry one of the d orbitals with an x and y component must be used for metal–ligand bonding. Therefore, for this configuration, there is a maximum of four metal d-orbital overlaps capable of generating metal–metal bonding. To date a large number of quadruple bonded compounds have been synthesized having metal centers, M, for which M = Cr, Mo, W,
3
Tc and Re [4]. In many of these compounds, bidentate ligands such as carboxylates span the two transition metal atoms. This type of arrangement is commonly referred to as a tetragonal paddlewheel or simply as a paddlewheel. For such complexes the resultant idealized MO interaction diagram is given in Fig. 2. As electrons are added to the quadruple bonded manifold, these occupy antibonding orbitals and thus the bond order decreases from 4 when there are eight metal-base electrons available to a minimum of ½ when there is only one unpaired electron in species containing 15 metal-based electrons, e. g., in M25+ species, M = Ni [5] and Pt [6]. Similar decreases in bond order may be attained by removal of electrons. With the exception of chromium, only a handful of paddlewheel compounds containing first row elements have been synthesized. Similarly, only few such species with early transition elements are known [4]. The main reason for the small number of paddlewheel species with early transition elements is possibly the underdeveloped low-valent coordination chemistry and perhaps the high affinity for oxygen exhibited by the higher valent species. To date, notable exceptions of tetragonal paddlewheel compounds for elements in Groups 5–10 are those of Ta, Fe and Mn. There are, however, trigonal paddlewheels with Fe23+ cores and relatively short Fe–Fe distances of 2.20–2.23 Å [7]. One of them having the formula Fe2(DPhF)3 (DPhF = the anion of N,N′diphenylformamidine) is shown in Fig. 3. Other dinuclear compounds having iron–iron bonds are known but these will not be a focus of this manuscript. Some examples are Fe2(tim)2 (tim = 2,3,9,10-tetramethyl-1,4,8,11-tetraazacyclotetradeca-1,3,8,10-
4
tetraene) that has an unsupported Fe–Fe bond of 2.6869(6) Å [8,9]. Finally, it is appropriate to mention two low-coordinate dinuclear compounds bridged by two bulky guanidinates having very short iron–iron bonds. These have the formulas [Fe2{µ(guanidinate)2}2 where guanidinate represents the ligands denominated as Pipiso {Pipiso = (DipN)2C(cis-2,6-MeC5H8)} [10] or But2C=N)C(NDip) (Dip = 2,6diisopropylphenyl) which has a very short iron–iron bond of 2.1516(5) Å [11]. In this chronicle, it will be described how a sought-after systematic understanding of the coordination chemistry of vanadium allowed the synthesis of the first compound with a vanadium–vanadium triple bond in a species that has divalent metal atoms, six metal-based electrons and a V24+ core. Subsequently an entire family of such compounds was prepared and these were accompanied by related species having V23+ cores and a formal bond order of 3½. It will also be shown how an akin exploration of the divalent chemistry of Nb(II) allowed the syntheses of Nb24+ paddlewheel species. Finally, some attempts to synthesize Ti(II) species will be described. The narrative will highlight how basic knowledge of early transition metal coordination chemistry has allowed rational syntheses of metal–metal containing species. 2. Developing low-valent coordination chemistry of early transition elements 2.1. Vanadium It is often useful to apply lessons learned from the synthesis of previously isolated compounds to attempt preparation of related ones. However, this statement
5
has to be taken with a grain of salt since diverse elements, even those in the same group, often exhibit remarkable differences in chemical behavior. To prepare divanadium species with metal bonds it was necessary to analyze the relatively little explored chemistry of vanadium(II) which tended to produce highly air-sensitive species. It was long known that finely divided vanadium metal would react in acidic aqueous media to generate purple solutions containing the divalent aqueous cation [12]. Alternatively, reduction of higher-valent species in acidic media using amalgamated zinc (Jones reductor) would also produce such species. In this way, VSO4∙7H2O could be isolated [13,14]. This compound has a V(H2O)62+ cation (Fig. 4) and an interstitial water molecule involved in an elaborate and extensive hydrogen bonded network in which every hydrogen atom interacts with an atom acceptor provided by the anion or the interstitial water molecule. Because of this tight network, dry crystals are much more stable to air than the corresponding solutions. Exploration of the aqueous chemistry with anions such as saccharinate [13] (Fig. 5), pyridine-2-carboxylate (picolinate; pic} [15] (Fig. 6) also yielded species with extended hydrogen networks that provided considerable stability towards air to the crystalline materials that allowed handling of the compounds outside a drybox for several minutes without appreciable decomposition [16]. For convenience some of the ligands mentioned in the manuscript are illustrated in Scheme 1. Whenever the water molecules that contribute to the hydrogen bonding networks are removed the stability towards air significantly decreases as shown by
6
trans-V(sacch)2(py)4, V(bipy)2(η2-SO4), [V(en)3]SO4, V(pySO3)2(py)4 [17] (Figs 7, 8, 9 and 10, respectively) as well as related amine derivatives [18–28], all of which are easily oxidized even in the solid state. However, these compounds—all of which possess V–N bonds—are thermodynamically stable and remain unchanged as long as they are protected from oxygen. Additionally, early reactivity studies showed that water replacement by pyridine is quite facile. One of the first structurally characterized species of this type—described by Brauer and Krueger in 1973 and shown in Fig. 11— is trans-VCl2(py)4 [29]. Its preparation was initially cumbersome but eventually reaction of the very insoluble commercially available VCl2 solid with pyridine in a pressure vessel at 250 °C gave good yields of trans-VCl2(py)4 [30]. Interestingly, early attempts to synthesize V(II) compounds by reducing solutions of VCl3(THF)3 with zinc metal produced the dinuclear mixed-metal species [(THF)3V(µ-Cl)3V(THF)3][Cl2Zn(µ-Cl)2ZnCl2] [31,32]. It is important to note that all divalent species described so far are sixcoordinate with an idealized octahedral inner sphere environment. It should also be pointed out that part of the reason VCl2 is so insoluble is the extended crosscoordination to attain six-coordinate vanadium atoms but perhaps more importantly because of the d3 electronic configuration with the three unpaired electrons in the t2g orbitals for each metal atom. Thus there is some resemblance to the slow reaction rates of other octahedral species with half-filled or filled t2g orbitals such as Fe(II) and Cr(III).
7
Because many dinuclear compounds with metal–metal bonds contain carboxylate groups spanning the dimetal unit [4], several attempts were made to use either aqueous V(II) solutions or those in non-aqueous media such as trans-VCl2(py)4 in pyridine solutions with M(O2CR) reagents, M = Li, Na, but for the most part this lead to the isolation of oxo-centered trivanadium species of the type V3(µ3-O)(O2CR)6L3n+, n = 0, 1 and L = H2O, THF, py [33] shown in Fig. 12. It quickly became evident that this type of oxo-centered species represented a thermodynamic sink for such reactions. As mentioned, early work with divalent vanadium species showed that Ncontaining ligands such as pyridine, ethylenediamine, bipyridine, 3-pyridinesulphonate generated stable V(II)–N bonds. In this way, it was decided to explore reactions with ligands containing two donor nitrogen atoms such as N,N′-diarylformamidinates (ArNC(H)NAr– commonly abbreviated as DArF where Ar = an aryl or alkyl group). Initial reactions with starting materials such as [(THF)3V(µ-Cl)3V(THF)3][Cl2Zn(µ-Cl)2ZnCl2] led to intractable species. Therefore the use of VCl3(THF)3 with various reducing agents was pursued. Early reaction attempts showed that whenever the formamidinate ligands were added to the reaction mixture previous to the addition of the reducing agents, the very stable mononuclear compounds having the formula V(formamidinate)3n–, n = 0, 1, were the only isolable products (Fig. 13) [34]. Indeed, the structure of the neutral trischelated compound represents a ubiquitous motif for many trivalent transition metalcontaining species [35]. However, in the presence of excess reducing agents the [V(formamidinate)3]– monoanion is prevalent (Fig. 14) [36]. Importantly, whenever
8
VCl3(THF)3 was first treated with sodium triethylborohydride (NaHBEt3) in THF solution previous to the addition of the formamidinate ligands, isolation of paddlewheel compounds was finally realized. This reaction presumably generates trans-VCl2(THF)4. Upon addition of an appropriate lithium or sodium salt of an DArF ligand this divalent mononuclear species reacts to form a dinuclear species containing a V24+ core with a vanadium–vanadium triple bond having a σ2π4 electronic configuration (Fig. 15) [37,34]. The first isolated compound of this family has the formula V2(DArF)4, where Ar = tolyl, as well as a short V–V distance of 1.975(4) Å. The success of this procedure may be attributed to the cleanliness of the reduction using NaHBEt3 which generates byproducts such as H2 and (THF)BEt3 both of which are volatile and thus easily removed from the reaction mixture when placed under vacuum. Sodium chloride, also produced during the reaction (R1), is insoluble in THF and thus easily removed by filtration. 2VCl3(THF)3 + 2NaHBEt3 + 4NaDArF → V2(DArF)4 + 6NaCl + H2 + 2(THF)BEt3 + 3THF R1 Subsequently, another laboratory used the reaction of trans-VCl2(tmeda) (tmeda = N,N,N',N'-tetramethylethylendiamine) with lithium formamidinates such as Li(CyNC(H)NCy)—abbreviated as DCyF, where Cy is cyclohexyl—as a way of obtaining an analogous paddlewheel complex [38]. In this process the intermediate sixcoordinate mononuclear species, V(DCyF)2(tmeda), initially formed. The dinuclear V2(DCyF)4 paddlewheel forms upon heating.
9
It has been shown that R1 is generally appropriate for the syntheses of other triple bonded V24+ species with ligands such as those having Ar = phenyl (Ph; Fig. 16), anisyl (Ani; Fig. 17), p-chlorophenyl (ClPh; Fig. 18), N,N'-triphenylguanidinate (tpg; Fig. 19) and 2-anilinopyridinate (ap; Fig. 20) [36]. All of these compounds have paddlewheel structures devoid of axial interactions and V–V bond distances of less than 2.0 Å, being 1.979(1), 1.9876(5), 1.974(1), 1.9521(7) and 1.9295(8) Å, respectively. In addition, the bicyclic guanidinate compound V2(hpp)4 (hpp = the anion of 1,3,4,6,7,8-hexahydro-2Hpyrimido[1,2-a]pyrimidine) was prepared in a similar manner and it has a metal–metal bond length of 1.932(1) Å [39,40].The longest metal–metal bond distances in these diamagnetic compounds correspond to the formamidinates while the shorter ones belong to compounds having the more basic guanidinate and aminopyridinate ligands. It is noteworthy that the absence of axial ligation along the V–V bond gives raise to species with five-coordinate vanadium atoms. Interestingly, when an excess of a strong reductant such as KC8 is used instead of NaHBEt3 further reduction of V2(DPhF)4 takes place yielding [K(THF)3]V2(DPhF)4, shown in Fig. 21 [41]. The core of this paramagnetic paddlewheel species has a V23+ unit with a formal bond order of 3½. The increase in bond order relative to those in the V24+ analogues is reflected in a smaller bond distance of 1.9295(8) Å which is shorter by about 0.05 Å relative to that of to the parent V2(DPhF)4. A multiline EPR spectrum shows that the unpaired electron is coupled to each vanadium atoms and has a g of 1.9999 [42].
10
It is noteworthy that in [K(THF)3]V2(DPhF)4, the potassium atom is nestled between the formamidinate ligands and thus determining independently whether the observed shortening in the metal–metal distance was affected by such arrangement was an important consideration. To unequivocally judge whether such arrangement was relevant, a reaction with a crown ether was carried out to trap and extract the nestled potassium atom. In this way a free paddlewheel, shown in Fig. 22, was isolated in [K(18-crown-6)(THF)2]V2(DPhF)4 [36]. The V–V distance of 1.924(2) Å is essentially the same as that in [K(THF)3]V2(DPhF)4. Thus, it was demonstrated that the K interactions with the formamidinate ligands did not alter the V–V bond distances. 2.2. On the way to a Nb24+ paddlewheel Because niobium, like vanadium, is in Group 5, one may naively expect some similarities in reactivity to that of the lighter congener. However, it is also well documented that the heavier elements in any transition metal group also display significant differences, one of which is that mononuclear species are much less prevalent in the heavier congeners. Another significant difference between V and Nb is the absence of aqueous Nb(II) chemistry.12 However, initial work showed that similarly to vanadium, Nb–amine bonds are thermodynamically stable. For example, transNbCl2(py)4 is easily made by reacting trans-NbCl4(THF)2 with intercalated potassium in graphite (KC8) in a mixture of THF/py (Fig. 23) [43]. This reaction mixture produced a royal blue solution containing the mononuclear species (R2).
11
trans-NbCl4(THF)2 + 2KC8 + 4py → trans-NbCl2(py)4 + 2KCl + 16Cgraphite + 2THF R2 Upon filtration of the insoluble KCl salt and graphite byproducts, followed by solvent removal under vacuum, the solid was extracted using a mixture of ethyl ether having a few drops of pyridine. Upon cooling, crystals were obtained. The structure resembles that of the vanadium analogue being six-coordinate with an idealized octahedral environment as well as a trans arrangement of the chlorine atoms. This suggested that similarly to vanadium, there is an inherent stability of the Nb(II)–N bonds. Therefore an analogous reaction to that used to prepare V2(DArF)4 (R1) was an obvious path in attempting to make a Nb2(DArF)4 paddlewheel species. 2.3. A side trip to formamidinate cleavage promoted by tantalum species However, it was rapidly found that reactions of niobium or the heavier congener tantalum in the presence of formamidinates led predominantly to ligand cleavage or other ligand rearrangements (Scheme 2) [44–46]. In such reactions the cleavage generally occurs at a C–N bond. Depending on the reaction conditions, one or both of the arylimido (ArN2–) or arylformidoyl (ArNCH3–) fragments, were observed and these fragments were frequently incorporated into the isolated products. The electrons necessary for the cleavage can be accounted by the metal atom oxidation. An example is provided by R3: Ta2Cl6(SMe2)3 + 4LiDTolF → Cl2(η2-DTolF)TaV(µ-NTol)(η2-η2-HNCTol)TaV(η2-DTolF)2 + 3SMe2 + 4LiCl R3
12
The product Cl2(η2-DTolF)TaV(µ-NTol)(η2-η2-HNCTol)TaV(η2-DTolF)2 (Fig. 24) is diamagnetic. An 1H NMR study revealed the position of the H atom in the unique carbon atom of the tolylformidoyl ligand which supports its formulation. Additionally, since there are no d electrons in the tantalum(V) atoms there is no metal–metal bonding and the long Ta∙∙∙Ta of 3.0005(6) Å supports this assertion. In the presence of reducing agents, the hydrogen atom in the unique carbon atom of the arylformidoyl group can formally be removed to give a dinegative phenylisocyanide species commonly referred to as arylimidoyl—the charge of which was inferred from the bonding parameters. Examples of such species are provided by the reaction of Ta2Cl6(SMe3)3 with DArF (Ar = Ph) in the presence of Zn powder used as a reducing agent. For the phenyl analogue, the reaction may be described by R4: Ta2Cl6(SMe2)3 + 5LiDPhF + ½Zn → (η2-DPhF)2TaIV(µ-PhN)(μ2-η2-PhNC)TaIV(η2-DPhF)2 + 3SMe2 + 5LiCl + ½ZnCl2 + ½H2
R4
In an analogous reaction with the tolyl analogue employing sodium amalgam a somewhat related species having two bridging formamidinates instead of the corresponding chelating ones was observed, as shown in R5: Ta2Cl6(SMe2)3 + 5LiDTolF + Na/Hg → (η2-DTolF)TaIV(µ-DtolF)2(µ-TolN)(μ2-η2PhNC)TaIV(η2-DPhF)2 + 3SMe2 + 5LiCl + NaCl + Hg + ½H2
R5
The structure of the phenyl analogue which has a Ta–Ta bond distance of 2.9411(7) Å is provided in Fig. 25. In the first reaction both the arylimidoyl and
13
arylformidoyl fragments are present. Interestingly, the fashion in which both the arylformidoyl and arylimidoyl (ArNCH3–and ArNC2–) ligand bridge the metal atoms is quite different. Even though they both span the tantalum atoms, for ArNC2– the arrangement is quite unsymmetrical with the ligand bonding in a μ2-η2-fashion while for ArNCH3– it is more symmetrical exhibiting a η2-η2-fashion as shown in Scheme 3. 2.4. Back to niobium Therefore, if a paddlewheel were to be made, it was necessary to find an alternative route that would prevent such cleavage. This is where the bicyclic guanidinate hpp ligand shown in Scheme 1 came into play. Because of the support provided to the C–N bond by other ring bonds, this ligand was projected to be stable towards cleavage. Furthermore, it had been shown that it could effectively form paddlewheel complexes such as those in M2(hpp)4, M = V, Cr, Mo,39 and Ru2(hpp)4Cl2 [47]. Using this strategy, Nb2(hpp)4 was isolated from the reaction in THF of NbCl3(DME) (DME = dimethoxyethane) with hpp and the reducing agent KC8 [48], as shown in R6. 2NbCl3(DME) + 4Lihpp + 2KC8 → Nb2(hpp)4 + 2KCl + 4LiCl + 16Cgraphite + DME R6 This compound has a Nb–Nb bond distance of 2.2035(9) Å (Fig. 26). Remarkably, the structure, as well as the metal–metal distance, had been predicted in the early days of density functional theory before the compound was synthesized [49]. With the knowledge that shielding the C–N–C moiety, other ligands with protected NC bonds were investigated and in this way, 7-azaindole (azin) was used to synthesize Nb2(azin)4
14
(Fig. 27) [50]. This compound crystallizes along with solvated LiCl producing a dinuclear [Li(THF)2Cl]2 moiety whose chlorine atoms occupy axial positions along the Nb–Nb bond at the relatively long distance of 2.849(3) Å from the niobium atom. Nevertheless, this weak interaction is enough to lengthen the Nb–Nb distance by about 0.07 Å to 2.278(2) Å. It is possible that such lengthening may be partly due to the larger increased bite angle of the azin ligand relative to that of hpp. When an excess of the reducing agent NaEt3BH was used instead of KC8 in R6, complexes with composition Nb2(hpp)4∙NaEt3BH and Nb2(hpp)4∙2NaEt3BH were isolated (Figs. 28 and 29, respectively) [50]. In these species the sodium atom is nestled within the hpp ligands and there is no direct interaction of the anion with the dinuclear species and thus the Nb–Nb distance is unaffected being 2.2064(3) Å and 2.2187(7) Å for Nb2(hpp)4∙NaEt3BH and Nb2(hpp)4∙2Et3BH, respectively. This behavior resembles that of [K(THF)3]V2(DPhF)4 (vide supra) [36]. Such reactions employing the guanidinate ligand hpp as well as azin demonstrate that the reaction scheme involving the strategy of protecting the ligand from cleavage is general and that the triple bonded Nb24+ unit is thermodynamically stable. An unusual compound with an Nb26+ core is obtained by oxidizing Nb2(hpp)4 with FeCp2PF6 in DME. This reaction generates [Nb2(η2-hpp)4(µ-η2,η2-hpp]PF6 (Fig. 30) [51]. In this compound there are four chelating hpp groups as well as a symmetrical hpp unit positioned essentially perpendicular to the Nb–Nb bond. In this compound the
15
metal atoms are separated by 2.6014(8) Å; a distance similar to that found in a related Ti26+ species (vide infra). The presence of a fifth hpp ligand suggests other products are also formed but they were not isolated. It is worth mentioning that a double bonded Nb26+ unit with formamidinate groups has been isolated by reacting Nb2(µ-SMe)3Cl6 with NaEt3BH. This reaction generates Nb2(µ-SMe)2(µ-DTolF)2(η2-DTolF)2 (DTolF = the anion of N,N'-di-ptolylformamidine) (Fig. 31) [52]. In this compound the Nb–Nb distance of 2.665(2) Å is significantly longer (by about 0.46 Å) than that in the triple bonded Nb24+ species. Unfortunately, thus far tantalum paddlewheel complexes analogous to those of Nb have not yet been isolated, even though the nononuclear mer-TaCl3(py)3 (Fig. 32) has shown that Ta–N bonds should be stable [53]. This compound has been isolated in three crystalline forms. 2.3. Some adventures pursuing Ti(II) coordination species An exploration in our laboratories, showed that the structure of trans-TiCl2(py)4 is analogous to that of the V and Nb species (Fig. 33) [43]. As with its analogues, the coordination sphere is octahedral with two trans chlorine atoms. This species has recently been the topic of further studies including theoretical computations, EPR analyses and a structural redetermination that confirmed that reported earlier [54].
Reactions with the ligand hpp gave surprising results that again showed that the chemistry of each element is unique and few generalizations are possible. By reducing
16
the pseudo octahedral species Ti(η2-hpp)2Cl2 with excess KC8 in THF the unusual dinuclear compound Cl(η2-hpp)Ti(µ-η2,η2-hpp)2Ti(η2-hpp)Cl species (Fig. 34) was isolated. The arrangement of the µ-η2,η2-hpp unit is reminiscent of that in [Nb2(η2hpp)4(µ-η2,η2-hpp2]PF6 (vide supra) [51]. Similarly to the triple bonded V24+ and Nb24+ hpp-compounds, this species contains a Ti26+ unit but the arrangement of the four hpp ligands is quite unusual and not that of a paddlewheel. The Ti–Ti distance of 2.5944(9) Å is quite long. Nevertheless, DFT calculations suggest that there is a single bond with a σ2 electronic configuration. Solution studies by 1H NMR spectroscopy indicate that the solution structure and that in the solid may be different, with that in solution being much more symmetrical. One could speculate that the more symmetrical solution species may be a paddlewheel. Interestingly, a reaction of the bulky amido-containing compound Ti[NButAr]3 (Ar = 3,5-C6H3Me2 or Ar' = C6H5 and But = tert-butyl) with carbon dioxide at –40 °C produces a diamagnetic Ti26+ paddlewheel with the formula of Ti2[µO2CN(But)Ph]4(NButPh)2 (Fig. 35) [55]. This compound has four carbamate ligands spanning the Ti26+ unit, thus forming a paddlewheel with two axial amido ligands (NButPh). However the Ti∙∙∙Ti separation of over 3.4 Å is rather long and precludes the existence of a metal–metal bond. It is clear that the strongly bound axial ligands (Ti–N = 1.920(3) Å) compete with the metal–metal formation. These assertions are supported by DFT calculations.
17
Nonetheless, a single bonded tri-titanium unit was obtained by reducing titanium tetrachloride or TiCl3(THF)3 in the presence of phosphines and Grignard reagents. It was found that reaction of TiCl4(dmpe) (dmpe = 1,2bis(dimethylphosphino)ethane) in toluene with ButMgCl in diethyl ether (R7) produces a red-brown solution from which the triangular species [Ti(µ-Cl)Cl(dmpe)]3 crystallizes (Fig. 36) [56]. 3TiCl4(dmpe) + 6ButMgCl → [Ti(µ-Cl)Cl(dmpe)]3 + 6MgCl2 + 3But– But
R7
This compound represents the first species with a genuine Ti–Ti single bond and an example of a relatively thermally stable—though pyrophoric in air—divalent titanium compound in the solid state. Nevertheless, solutions did decompose rather quickly at ambient temperature.
An interesting dinuclear species with a short Ti–Ti bond of 2.399(2) Å has a sandwich structure with a bispentalene unit above and below the Ti2 unit and a formula Ti2(µ:η5,η5-Pn)2 (Pn = C8H4{SiPri3-1,4}2) [57]. DFT computations suggest there is a Ti=Ti with σ and π components.
Other dinuclear compounds with trivalent titanium are calix[4]arene derivatives such as K4(THF)8[Ti2(µ-NMe2)2(calix[4]arene)2] and Na2(THF)6[Ti2(calix[4]arene)2] (Figs. 37 and 38, respectively) [58]. However, the long TiIII···TiIII separations of 3.278(2) and 3.133(1) Å, respectively, preclude the existence of a metal–metal bond.
18
3. Concluding remarks While attempting preparation of metal–metal bonded compounds of diverse transition metal species, searching for parallels from those of other metal atoms is sometimes a useful exercise to guide the new synthetic procedures. However, more often than not completely new routes must be developed. What is evident from the narrative is that understanding the complexity of the coordination chemistry of a given element is essential in guiding the development of new processes. Often driving on country roads, identifying dead ends before reaching them and making adjustments is essential to find the highway that eventually leads to a targeted destination. This is similar to what a winding, sinuous river does on its way to the ocean. It does not matter how many turns and twists the stream makes, the water always reaches its target. It is hoped that the insights presented here will help students understand the intricacies of developing new fields and help them learn to appreciate that there is no such thing as a “the reaction did not work”. There are always positive aspects that an inquisitive mind can extract and then use to attain the final goals. Acknowledgments This work was supported by Texas A&M University as well as IR/D support from the National Science Foundation. The author also thanks the multiple co-workers responsible for the advance of this field whose names are provided in the references. With this short review, the author cheers the friendship with Luis Echegoyen as well as his wife Lourdes. With Luis I have had the privilege of knowing him for many years
19
before working together at the National Science Foundation and now we continue to strengthen the friendship by collaborating at the University of Texas at El Paso.
References
[1] F.A. Cotton, N.F. Curtis, C.B. Harris, B.F.G. Johnson, S.J. Lippard, J.T. Mague, W.R. Robinson, J.S. Wood, Science 145 (1964) 1305–1307. [2] For a historical account on the δ bond, see: F.A. Cotton, Pure. Appl. Chem. 64 (1992) 1383–1393. [3] C.A. Murillo, Comments Inorg. Chem. 35 (2015) 39–58. [4] F.A. Cotton, C.A. Murillo, R.A. Walton (Eds.), Multiple Bonds between Metal Atoms, third ed. Springer Science and Business Media, Inc., New York, 2005. [5] J.F. Berry, E. Bothe, F.A. Cotton, S.A. Ibragimov, C.A. Murillo, D. Villagrán, X. Wang, Inorg. Chem. 45 (2006) 4396–4406. [6] F.A. Cotton, J.H. Matonic, C.A. Murillo, Inorg. Chim. Acta 264 (1997) 61–64. [7] F.A. Cotton, L.M. Daniels, L.R. Falvello, J.H. Matonic, C.A. Murillo, Inorg. Chim. Acta 256 (1997) 269– 275. [8] C.R. Hess, T. Weyhermüller, E. Bill, K. Wieghardt, Angew. Chem. 48 (2009) 3703–3706. [9] C.A. Murillo, Angew. Chem. 48 (2009) 5076–5077. [10] L. Fohlmeister, S. Liu, C. Schulten, B. Moubaraki, A. Stasch, J.D. Cashion, K.S. Murray, L. Gagliardi, C. Jones, Angew. Chem. 51 (2012) 8294–8298. [11] A.K. Maity, A.J. Metta-Magaña, S. Fortier, Inorg. Chem 54 (2015) 10030–10041. [12] F.A. Cotton, G. Wilkinson, C.A. Murillo, M. Bochmann, Advanced Inorganic Chemistry, sixth ed., John Wiley and Sons, New York, 1999. [13] F.A. Cotton, L.R. Falvello, R. Llusar, E. Libby, C.A. Murillo, W. Schowtzer, Inorg. Chem. 25 (1986) 3423–3428. [14] F.A. Cotton, L.R. Falvello, C.A. Murillo, I. Pascual, A.J. Schultz, M. Tomás, Inorg. Chem. 33 (1994) 5391–5395. [15] F.A. Cotton, L.M. Daniels, M.L. Montero, C.A. Murillo, Polyhedron 11 (1992) 2767–2774. [16] F.A. Cotton, E. Libby, C.A. Murillo, G. Valle, Inorg. Synth. 27 (1990) 306–310. [17] L.M. Daniels, C.A. Murillo, K.G. Rodríguez, Inorg. Chim. Acta 229 (1995) 27–32. [18] For example, see: A.C. Niedwieski, P.B. Hitchcock, J.D. Da Motta Neto, F. Wypych, G.F. Leigh, F.S. Nunes, J. Brazil. Chem. Soc. 14 (2003) 750–758. [19] A.C. Niedwieski, G.F. Leigh, T. Hasegawa, F.S. Nunes, Acta Crystallogr. E59 (2003) m939–m941. [20] A.C. Raimond, T. Hasegawa, F.S. Nunes, Acta Crystallogr. E60 (2004) m1010–m1012. [21] D.M. Halepoto, L.F. Larkworthy, S.K. Sengupta, Transit. Metal Chem. 28 (2003) 800–802. [22] J.G. Reynolds, S.C. Sendlinger, A.M. Murray, J.C. Huffman, G. Christou, Inorg. Chem. 34 (1995) 5745– 5752. [23] P. Gosh, H. Taube, T. Hasegawa, R. Kuroda, Inorg. Chem. 34 (1995) 5761–5775. [24] J.J.H. Edema, S. Gambarotta, A. Meetsma, A.L. Spek, N. Veldman, Inorg. Chem. 30 (1991) 2062–2066. [25] J.J.H. Edema, W. Stouthamer, F. Van Bothuis, S. Gambarotta, W.J.J. Smeets, A.L. Spek, Inorg. Chem. 29 (1990) 1302–1306. [26] F.A. Cotton, R. Poli, Inorg. Chim. Acta 141 (1988) 91–98. [27] L.F. Larkworthy, B.J. Tucker, Inorg. Chim. Acta 33 (1979) 167–170. [28] M.M. Khamar, L.F. Larkworthy, K.C. Patel, G. Beech, Austr. J. Chem. 27 (1974) 41–51.
20
[29] D.J. Brauer, C. Krueger, Cryst. Struct. Commun. 3 (1973) 421–426. [30] F.A. Cotton, C. A. Murillo, Ing. Cienc. Quim. 9 (1985) 5–6. [31] F.A. Cotton, S.A. Duraj, M.W. Extine, G.E. Lewis, W.J. Roth, C.D. Schmulbach, W. Schwotzer, J. Chem. Soc., Chem. Commun. (1983) 1377–1378. [32] R.J. Bouma, J.H. Teuben, W.R. Beukema, R.L. Bansemer, J.C. Huffman, K.G. Caulton, Inorg. Chem. 23 (1984) 2715–2718. [33] F.A. Cotton, M.W. Extine, L.R. Falvello, D.B. Lewis, G.E. Lewis, C.A. Murillo, W. Schowtzer, M. Tomás, Inorg. Chem. 25 (1986) 3505–3512. [34] F.A. Cotton, L.M. Daniels, C.A. Murillo, Inorg. Chem. 32 (1993) 2881–2885. [35] F.A. Cotton, L.M. Daniels, D.J. Maloney, C.A. Murillo, Inorg. Chim. Acta 242 (1996) 31–42. [36] F.A. Cotton, E.A. Hillard, C.A. Murillo, X. Wang, Inorg. Chem. 42 (2003) 6063–6070. [37] F.A. Cotton, L.M. Daniels, C.A. Murillo, Angew. Chem., Int. Ed. Engl. 31 (1992) 737–738. [38] S. Hao, P. Berno, R.K. Minhas, S. Gambarotta, Inorg. Chim. Acta 244 (1996) 37–49. [39] F.A. Cotton, D.J. Timmons, Polyhedron 17 (1997) 179–184. [40] C. Fernández-Cortabitarte, F. Gacía, J.V. Morey, M. McPartlin, S. Singh, A.E.H. Wheatley, D.S, Wright, Angew. Chem. Int. Ed. 46 (2007) 5425–5427. [41] F.A. Cotton, E.A. Hillard, C.A. Murillo, J. Am. Chem. Soc. 125 (2003) 2026–2027. [42] For a discussion on the usefulness of EPR spectroscopy in dimetal species, see: N.S. Dalal, C.A. Murillo, Dalton Trans. 43 (2014) 8565–8576. [43] M.A. Araya, F.A. Cotton, J.H. Matonic, C.A. Murillo, Inorg. Chem. 34 (1995) 5424–5428. [44] F.A. Cotton, L.M. Daniels, C.A. Murillo, X. Wang, Inorg. Chem. 36 (1997) 896–901. [45] F.A. Cotton, J.H. Matonic, C.A. Murillo, X. Wang, Bull. Soc. Chim. Fr. 133 (1996) 711–720. [46] F.A. Cotton, L.M. Daniels, J.H. Matonic, X. Wang, C.A. Murillo, Polyhedron 16 (1997) 1177–1191. [47] J.L. Bear, Y. Li, B. Han, K.M. Kadish, Inorg. Chem. 35 (1996) 1395–1398. [48] F.A. Cotton, J.H. Matonic, C.A. Murillo, J. Am. Chem. Soc. 119 (1997) 7889–7890. [49] F.A. Cotton, X. Feng, J. Am. Chem. Soc. 119 (1997) 7514–7520. [50] F.A. Cotton, J.H. Matonic, C.A. Murillo, J. Am. Chem. Soc. 120 (1998) 6047–6052. [51] F.A. Cotton, S.A. Ibragimov, C.A. Murillo, P.V. Poplaukhin, Q. Zhao, J. Molec. Struct. 890 (2008) 3–8. [52] F.A. Cotton, J.H. Matonic, C.A. Murillo, J. Clust. Sci. 7 (1996) 655–662. [53] F. A. Cotton, C.A. Murillo, X. Wang, Inorg. Chim. Acta 245 (1996) 115–118. [54] G.B. Wijeratne, E.M. Zolnhofer, S. Fortier, L.N. Grant, P.J. Carroll, C-H. Chen, K. Meyer, J. Krzystek, A. Ozarowski, T.A. Jackson, D.J. Mindiola, J. Telser, Inorg. Chem. 54 (2015) 10380–10397. [55] A. Mendiratta, C.C. Cummins, F.A. Cotton, S.A. Ibragimov, C.A. Murillo, D. Villagrán, Inorg. Chem. 45 (2006) 4328–4330. [56] F.A. Cotton, C.A. Murillo, M. Petrukhina, J. Organomet. Chem. 573 (1999) 78–86. [57] A.F.R. Kilpatrick, J.C.C. Green, F.G.N. Cloke, N. Tsoureas, Chem. Commun. 49 (2013) 9434–9436. [58] F.A. Cotton, E.V. Dikarev, C.A. Murillo, M.A. Petrukhina, Inorg. Chim. Acta 332 (2002) 41–46.
21
Fig. 1. The structure of the quadruple bonded Re2Cl82– anion.
Fig. 2. Idealized molecular orbital diagram for a dimetal species with D4h symmetry. In a species with metal atoms with d4 electronic configurations there are eight metal-based electrons that fill the four bonding orbitals given a σ2π4δ2 electronic configuration and a bond order of four. For simplicity the d orbitals are drawn at arbitrary levels. Note that one of the d orbitals with x and y components is used in metal-to-ligand bonding and thus it is unavailable for metal-tometal bonding.
22
Fig. 3. The core structure of the trigonal paddlewheel compound Fe2(DPhF)3 that has an Fe23+ moiety. DPhF = the anion of N,N'-diphenylformamidine.
Fig. 4. The structure of [V(H2O)6]SO4∙H2O. All H atoms, including those of the interstitial water molecule, are engaged in H-bonding. For simplicity this network is not shown.
23
Fig. 5. The N-bonded saccharinate compound [trans-V(C7H4NO3S)2(H2O)4]∙2H2O. Only one of the symmetry related interstitial water molecules is shown. There is an extensive hydrogen bonding network (not shown) with all water-hydrogen atoms being involved.
Fig. 6. A carboxylate derivative trans-V(pic)2(H2O)2 where pic = pyridine-2-carboxylate (picolinate).
24
Fig. 7. The structure of trans-[V(C7H4NO3S)2(py)4]∙2py. Note that as the water molecules are substituted by pyridine the bonding mode of the saccharinate ligands changes from N-bonded to O-bonded. For comparison, see Fig. 5.
Fig. 8. The mononuclear compound V(bipy)2(η2-SO4).
Fig. 9. The tris(ethylenediamine) compound [V(en)3]SO4.
25
Fig. 10. The compound trans-V(pySO3)2(py)4 where pySO3 = 3-pyridinesulphonate.
Fig. 11. An early structure— trans-[VCl2(py)4]∙THF.
Fig. 12. A few examples of oxo-centered trinuclear compounds: (i) [V3(µ3O)(O2CCHCl)6(H2O)3](CF3SO3) having three trivalent vanadium atoms; (ii) [V3(µ3O)(O2CCH3)6(THF)3]2[Cl2OV(µ-Cl)2VOCl2] with three V(III) atoms and (iii) [V3(µ3O)(O2CCH3)6(py)3]2 having one divalent and two trivalent vanadium atoms.
26
Fig. 13. The mononuclear V(DTolF)3 compound having a trivalent vanadium atom and three chelating formamidinate anions. DTolF = the anion of N,N′-di-p-tolylformamidine.
Fig. 14. [Li(THF)4]V2(DCl2PhF)3 having a divalent vanadium atom. DCl2PhF = the anion of N,N′-di3,5-dichloroformamidine.
27
Fig. 15. The triple bonded V2(DTolF)4 compound where DTolF is the anion of N,N′-di-ptolylformamidine. The V–V bond distance is 1.978(2) Å.
Fig. 16. The paddlewheel compound V2(DPhF)4 where DPhF = N,N′-diphenylformamidinate. The V–V bond distance is 1.979(1) Å.
28
Fig. 17. The structure of V2(DAniF)4 where DAniF = N,N′-di-p-anisylformamidinate. This ligand has an important advantage over analogous formamidinates in that typically its compounds are more soluble than analogues without the methoxy group. The V–V bond distance is 1.9876(5) Å.
Fig. 18. The structure of V2(DClPhF)4 where DClPhF = N,N′-di-p-chloroformamidinate. The V–V bond distance is 1.974(1) Å.
29
Fig. 19. The structure of the guanidinate derivative V2(tpg)4, where tpg = 1,2,3triphenylguanidinate. The V–V bond distance is 1.9521(7) Å.
Fig. 20. The structure of V2(ap)4 where ap = 2-anilinopyridinate. The V–V bond distance is 1.9425(4) Å.
30 Fig. 21. An unusual species having a V23+ core and a formal metal–metal bond order of 3½: [K(THF)3]V2(DPhF)4. Note that the potassium atom is nestled within phenyl rings of the formamidinate ligands and bound to the oxygen atom of the three THF molecules. The V–V bond distance is 1.9295(8) Å.
Fig. 22. The compound [K(18-crown-6)(THF)2]V2(DPhF)4 having a V23+ core with a vanadium– vanadium bond order of 3½. The V–V bond distance is 1.924(2) Å.
Fig. 23. The structure of the THF solvate of trans-[NbCl2(py)4].
31
Fig. 24. The structure of Cl2(η2-DTolF)TaV(µ-NTol)(η2-η2-HNCTol)TaV(η2-DTolF)2. Note that both the p-tolylamido and p-tolylformidoyl fragments bridge the metal atoms.
Fig. 25. The structure of (η2-DPhF)2TaIV(µ-PhN)(μ2-η2-PhNC)TaIV(η2-DPhF)2. Only one orientation of a disordered PhCN2– group related by a 2-fold axis is shown. The structure resembles that of the tolyl analogue.
32
Fig. 26. The triple bonded Nb2(hpp)4 where hpp = the anion of 1,3,4,6,7,8-hexahydro-2Hpyrimido[1,2a]pyrimidine. The Nb–Nb distance is 2.2035(9) Å.
Fig. 27. The structure of Nb2(azin)4∙2LiCl∙4THF in where azin = 7-azaindole. The Nb–Nb distance is 2.278(2) Å. For simplicity the lithium atoms and THF molecules are not shown.
Fig. 28. The structure of Nb2(hpp)4∙NaHBEt3. The Nb–Nb distance is 2.2064(3) Å. Note that the sodium interaction resembles that of potassium in [K(THF)3]V2(DPhF)4 in Figure 21.
33
Fig. 29. The compound Nb2(hpp)4∙2NaHBEt3. The Nb–Nb distance is 2.2187(7) Å. For simplicity only one of the interstitial NaHBEt3 molecules is shown.
Fig. 30. The compound [Nb2(η2-hpp)4(µ-η2-η2-hpp)]PF6 having a Nb26+ core where hpp = the anion of 1,3,4,6,7,8= hexahydro-2H-pyrimido[1,2a]pyrimidine. The Nb–Nb distance is 2.6014(8) Å.
Fig. 31. The structure of [(η2-DTolF)Nb(µ-SMe2)(µ-DTolF)]2 where DTolF = N,N′-di-ptolylformamidinate. The Nb–Nb distance is 2.655(2) Å.
34
Fig. 32. The trivalent, nononuclear mer-TaCl3(py)3 compound.
Fig. 33. The mononuclear THF solvate of trans-[TiCl2(py)4].
Fig. 34. The trivalent titanium species TiIII2(η2-hpp)2(µ-η2-η2-hpp)2Cl2 where hpp = the anion of 1,3,4,6,7,8-hexahydro-2H-pyrimido[1,2a]pyrimidine. The Ti–Ti distance is 2.5944(9) Å.
35
Fig. 35. The tetracarbamate paddlewheel species TiIII2[µ-O2CN(But)Ph]4(NButPh)2∙Et2O which is devoid of metal–metal bonding; the long Ti∙∙∙Ti separation is 3.515(1) Å. The amido NButPh groups occupy axial positions along the Ti∙∙∙Ti axis.
Fig. 36. The trinuclear compound [TiII(µ-Cl)Cl(dmpe)]3 where dmpe = dimethylphosphinoethane. The Ti–Ti single bond length is 2.827(3) Å.
Fig. 37. The structure of K4(THF)8[TiIII2(µ-NMe2)2L2] where H4L = C28H20O4H4; calix[4]arene. The Ti∙∙∙Ti separation of 3.28 Å is too long to result in significant Ti–Ti bonding.
36
Fig. 38. The dinuclear species [Na(THF)3]2[TiIIIL2]2∙2THF where H4L = C28H20O4H4; calix[4]arene. The Ti∙∙∙Ti separation is 3.133(1) Å.
Scheme 1. A diagrammatic representation of some of the ligands mentioned in the text showing the corresponding abbreviations.
37
Scheme 2. How a formamidinate may be fragmented by a transition metal-containing species having Ta or Nb producing arylimidoyl (ArN2–) and arylformidoyl (ArNCH3–) anions. The electrons are provided by concomitant oxidation of the metal atom.
Scheme 3. Schematic representation of the bonding of the ArNC2– fragment to a ditantalum species (left) and that of the HCNAr3– anion (right).
38
Synopsis
39
Highlights • The coordination chemistry of some low valent early metal complexes is presented • A roadmap to the syntheses of paddlewheel compounds with V–V triple bonds is given • The syntheses of paddlewheel compounds with metal–metal bonds is described • The preparation of diniobium compounds is presented • Some adventures in low valent titanium are described • Lessons learned on similarities and differences in their chemistries are offered
40 Carlos A. Murillo studied chemistry at the University of Costa Rica and Texas A&M University where he received his Ph. D. in 1973 with F. Albert Cotton. He then did postdoctoral work with Malcolm H. Chisholm at Princeton University. He went back to Costa Rica where he quickly moved to the rank of Professor. In 1991 he moved to Texas A&M as director of the Laboratory for Molecular Structure and Bonding and in 2007 he took a position as program director in the Division of Chemistry at the US National Science Foundation. He has continued his research as an adjunct professor at Texas A&M University and the University of Texas at El Paso. He is a Fellow of the American Association for the Advancement of Science (AAAS) and a charter member of the Costa Rican Academy of Sciences.
41