A Xanthene‐Based Mono‐Anionic PON Ligand: Exploiting a Bulky, Electronically Unsymmetrical Donor in Main Group Chemistry

Abstract The synthesis of a novel mono‐anionic phosphino‐amide ligand based on a xanthene backbone is reported, togetherr with the corresponding GaI complex, (PON)Ga (PON = 4‐(di(2,4,6‐trimethylphenyl)phosphino)‐5‐(2,6‐diisopropylanilido)‐2,7‐di‐tert‐butyl‐9,9‐dimethylxanthene). The solid‐state structure of (PON)Ga (obtained from X‐ray crystallography) reveals very weak O⋅⋅⋅Ga and P⋅⋅⋅Ga interactions, consistent with a R2NGa fragment which closely resembles those found in one‐coordinate amidogallium systems. Strong N‐to‐Ga π donation from the amido substituent is reflected in a very short N−Ga distance (1.961(2) Å), while the P⋅⋅⋅Ga contact (3.076(1) Å) is well outside the sum of the respective covalent radii. While the donor properties of the PON ligand towards GaI are highly unsymmetrical, oxidation to GaIII leads to much stronger coordination of the pendant phosphine as shown by P−Ga distances which are up to 20 % shorter. From a steric perspective, the PON ligand is shown to be significantly bulkier than related β‐diketiminate systems, a finding consistent with reactions of (PON)Ga towards O‐atom sources that proceed without oligomerization. Despite this, the enhanced P‐donor properties brought about by oxidation at gallium are not sufficient to quench the reactivity of the highly polar Ga−O unit. Instead, intramolecular benzylic C−H activation is observed across the Ga−O bond of a transient gallanone intermediate.


Introduction
In recent years, b-diketiminate ('Nacnac')l igandsh aveb een extensively used in coordination chemistry,t os upport aw ide range of metal complexes from across the PeriodicT able. [1][2][3][4] Within main group chemistry,an umber of landmark compounds have been reported incorporating these chelating monoanionic LX ligand systems. These include Group 13 metal complexes in the + 1o xidations tate-systems which are challenging both in terms of their intrinsic tendency to undergo disproportionation and their lability towards externalr eagents. [5] Thus in 2000, the groups of Roesky and Power successfully synthesized the monomeric Al I and Ga I compounds [HC{(Me)C(Dipp)N} 2 ]E (or (Nacnac)E, I,w here E = Al, Ga) using sterically encumbered Dipp-substituted b-diketiminate ligands ( Figure 1; Dipp = 2,6-iPr 2 C 6 H 3 ). [6,7,8] In 2018, we employed ac helating dianionicd iamido (X 2 ) ligand, [ NON] 2À ,t oa ccess anionic Al I and Ga I compounds of the type K 2 [(NON)E] 2 (II,E = Al, Ga;N ON = 4,5-bis(2,6-diiso-propyl-anilido)-2,7-di-tert-butyl-9,9-dimethylxanthene). [9][10][11] These systems show unusualp atterns of reactivity:t he aluminyl anion can act as an ucleophile in the formation of CÀAl and MÀAl bonds, [9,12] and effect the reversible oxidativea ddition of the CÀCb ond in benzene. [13] Given the track record of the Nacnacl igand family in supportingc harge neutral E(I) systems, however, [2][3][4] we targeted relatedm ono-anionic LX donors based on the dimethylxanthenebackbone, in which one of the amido groupso ft he [NON] 2À system is formally replaced by a neutral donor.M oreover,g iven the successful developmentb y Power, and by Jones, of one-coordinate Ga I amides (stabilized to ag reater or lesser extent by interactions with the flanking aryl substituents, III), [14] we targeted xanthenes ystemsf eatur-  ing Ls ubstituents which are known to act as relativelyw eak donors towards Ga I .O ur aim was to develop amido complexes featuring aw eakly interacting (or hemi-labile) Ld onor that might display the high levels of reactivity associated with genuine one-coordinate systems, while retaining some of the ground state stability of Nacnac compounds. Accordingly,w e report in the current manuscript on the development of a mono-anionic [PON] À ligand ( Figure 1) and the use of Ga I and Ga III chemistry to probe its coordination capabilities.

Results and Discussion (i)Synthesis of H(PON) and (PON)Ga
The H(PON) protio-ligand 2 can be synthesized in good yield (ca. 70 %) from the (known)4 -bromo-5-(dimesitylphosphino)xanthenep recursor (1) [15] andD ippNH 2 via aB uchwald-Hartwig coupling reaction. 2 has been characterized by multinuclear NMR, elemental microanalysis and single crystal X-ray diffraction (see SI). Moreover,i ti sc onveniently deprotonated by benzylpotassium, KCH 2 Ph, giving the potassiated ligand K(PON) (3), which can be used for onward reactionc hemistry without furtherp urification, or recrystallized from toluenef or structural and spectroscopic characterization. 3 is thus obtained free of coordinating solvent as dimericK 2 (PON) 2 ,i nw hich K(PON) monomer units are linked in head-to-tail fashion throughc lose contacts between the potassium centre of one unit and the Dipp aromatic p system of the other ( Figure 2). With this new ligand system in hand we set out to probe its coordination chemistry,u sing gallanediyl (Ga I )a nd gallium oxide (Ga III )s ystems as probes of its ability to support highly reactive main group metal centres.A ccordingly,t he Ga I system,( PON)Ga, 4 can be prepared in ca. 60 %i solated yield via as alt-metathesis reactionb etween 3 and "GaI" (Scheme 1), [16] and obtained in crystalline form by recrystallization from benzene. The 1 HNMR spectrum of 4 features as imilar pattern of resonances to those of H(PON) and K(PON), with equivalent phosphorus-bound mesityl substituents, Dipp iPr groups and xanthene backbone methyl groups implyingt he presenceo faplane of symmetry on the NMR timescale. The 31 P{ 1 H} resonanceo f4 (at d P = À34.3 ppm) is very close to those of protio-ligand 2 (d P = À36.9 ppm) andp otassiated system 3 (d P = À34.4 ppm), implying that the interaction of the phosphine donorw ith the gallium centre in 4 is relatively weak.C onsistently,t he (monomeric)s tructure of 4 in the solid state ( Figure 2) features av ery long Ga···P distance (3.076(1) ) which is comfortably outside the sum of the respective covalent radii (1.22(4)+ +1.07(3) ), [17] and much longert han that measured for the corresponding Ga III system (PON)GaI 2 which was synthesized for comparativep urposes (2.612(1) ;s ee SI). The GaÀOs eparation involving the xanthene ether linkagei s also relativelyl ong (2.631(2) )c ompared to the sum of the respectivec ovalent radii( 1.22(3)+ +0.66(2) ), [17] and is broadly comparable to that measured for the potassium gallyl system K 2 [(NON)Ga] 2 ,f eaturing the related [NON] 2À diamidel igand (2.542(2) ). [9] On the other hand, the GaÀNb ondi sv ery short (1.961(2) )i nc omparison with those in K 2 [(NON)Ga] 2 (2.093(2) and 2.106(2) ), [9] presumably reflecting the fact that there is only one Ga-X covalent bond in 4.I ndeed, the GaÀNs eparation in 4 is in line with the bond lengths reported by Power and by Jones for Ga I amides with geometries which approach mono-coordinate (1.954(2) À1.985(4) ), [14] consistent with idea that the coordinationo ft he phosphine and ether donors in 4 is very weak.

(ii)E lectronic and steric properties of (PON)Ga
The electronic structure of (PON)Ga (4)w as probedu sing Density Functional Theorya tt he PBE1PBE/TZVP level. These calculations suggestt hat the LUMO + 2( À0.68 eV) is comprised predominantly of gallium p z character ( Figure 3). The HOMO is ligand-based, with the orbitald isplaying gallium-centred lone pair character being the HOMO-1 (À5.67 eV). The associated energy separation (4.99 eV) is very similart ot hat calculated for (Nacnac)Ga [6] using the same method( 4.86 eV). The essentially identicale nergies of the formally vacant pp orbital in each case (À0.66 eV for (Nacnac)Ga) presumablyr eflect the fact that in 4,the extent of p donation from the (single) amido substituent is markedlye nhanced, consistent with the very short crys- tallographically determined GaÀNb ond length. Both of these systemsh ave wider HOMO-LUMO gaps than that determined for the anionic diamido system [(NON)Ga] À .I nt hat case, the narrow HOMO-LUMO gap (4.21 eV) reflects the factt hat the LUMO is not destabilized to the same degree by N-to-Ga p donation due to the constraints of the xantheneb ackbone (which mean that the amido groups cannota ttain coplanarity with the GaN 2 unit). [9,11] The (PON)Ga system can also be compared with Jones' one-coordinate amidogallium(I) system, [14b] which is similar in that it features p stabilization from only as ingle Nd onor.I nt hat system, the HOMO-LUMO gap (calculated using the same method) is narrower( 4.62 eV) and the LUMO somewhat lower than in 4 (À0.91 eV). This presumably reflects the fact that the LUMO is orthogonal to the amido p system, and (compared to 4)h as less opportunity for interaction with ancillary neutral donors (featuring as it does, only very weak interaction with the flankinga rene p system).
Steric factorsa re known to be as ignificant influence on the patterns of reactivity displayed by low-valent main group species. In both (PON)Ga and( Nacnac)Ga, the Ga I centre sits in a "pocket" between flankinga ryl substituents:i nt he b-diketiminate system these are the N-bound Dipp groups, which are aligneds uch that the distances from the arene centroids to the metal are each ca. 4.0 ,a nd the "pocket" defined by the open centroid-Ga-centroid angle occupies 189.18 in angular terms ( Figure 4). [6] In 4 the gallium centrei sf lanked by NDipp and PMes substituents (the centroids of which are also ca. 4.0 from the metal centre), and the open "pocket" by comparisono ccupies 156.88.A ss uch, we hypothesizedt hat 4 might offer the possibility of am ore sterically shielded pocket in which to carry out chemistrya tt he metal centre.

(iii)Oxidation of (PON)Ga
To probe the idea of enhanced steric bulk and encouraged by the finding that the weakly-bound phosphine donor becomes more strongly stabilizing on oxidation of the metal centre (cf. (PON)Gaa nd (PON)GaI 2 ), we examined the reactivity of 4 towards oxygen atom transfer agents. The isolation of a( hitherto unknown) gallanone complex containing at erminal GaO moiety, would presumably require both very strongly s-donating and sterically demanding ancillary ligands. [18] Power et al. have reported the formation of [(Nacnac)Ga(m-O)] 2 from the reaction of (Nacnac)Ga with N 2 O, [19] with the oxobridged structure presumablyr eflecting as teric profile which permits ready dimerization. In contrast, oxidation of 4 with N 2 Oo rM e 3 NO at room temperature leads to am onomeric product (5,S cheme 2), albeit one which results from intramolecular CÀHa ctivation of one of the ortho-methyl groups of the PMes 2 substituent across the transient gallium oxo unit. [20] Compound 5 has been characterized by spectroscopic methods and its structure in the solid state confirmed crystallographically ( Figure 5). The formation of 5 is signalled by the appearanceo fa 31 PNMR resonance at d p = À41.0 ppm and by increased complexity in the pattern of 1 HNMR signals associated with the ortho methyl substituents. Its structure in the solid state is based around an approximately tetrahedral gallium centre, featuring as ignificantlys hortened GaÀPd istance (2.468(1) cf. 3.076(1) for (PON)Ga), but with aw eak Ga···O contact which is essentially unchangedf rom its precursor (2.578(2) vs. 2.631 (2) ). The GaÀCa nd Ga-OH bond lengths (1.972(4) and 1.828 (2) )f all within the ranges previously reportedf or related (Nacnac)Ga-containing compounds. [21] Nikonov and co-workersh ave recently reported the trapping of the monomeric [(Nacnac)GaO] fragment by reactions of (Nacnac)Ga with N 2 Oi nt he presence of ar ange of relatively acidic CÀHb onds( e.g. those in sulfoxides and ketones, and at the 2-position of pyridine and related heterocycles). [20] In the case of 4/5,t he fact that CÀHa ctivation occurs in the absence   31 PNMR signals at À36.1 and À35.6 ppm, respectively.B oth products could be unambiguously characterizedb yX -ray crystallography,a nd are shown to result from cleavage of aB ÀCb ond in BPh 3 (6)o r B(C 6 F 5 ) 3 (7)a cross the GaÀOu nit formed in the initial reaction of gallylene 4 with N 2 O( see Scheme 2, Figure 5a nd SI). As such, the isolation of compounds 6 and 7 provides further evidence for the formation of at ransient, highly reactive gallanone speciesb yo xygen atom transfer to the galliumc entre of 4.I ti sw orth noting that similar reactivity-leading to the cleavage of one of the BÀCb onds in B(C 6 F 5 ) 3 across at erminal Si=Ob ond-has been reported by Iwamoto and co-workersi n studies of ar eactive (but isolable) silanone. [22] Conclusion In summary,w er eport on the development of am ono-anionic phosphino-amide ligand based on ax antheneb ackbone, togetherr with the synthesis of the corresponding Ga I complex, (PON)Ga( 4), as ap robe of its coordination properties in low valent main group chemistry. Structurals tudies of 4 are consistent with very weak O···Ga and P···Ga interactions, and with an R 2 NGa fragment that closely resembles those found in one-coordinate amidogallium systems. Strong N-to-Ga p donation from the singlea mido group leads to av ery short N-Ga distance (1.961(2) ), while the P···Ga contact (3.076(1) )i sc omfortably outside the sum of the respective covalent radii. While the electronic properties of the PON donor are shown to be highly unsymmetrical towards Ga I ,o xidation to Ga III leads to stronger coordination of the pendant phosphine arm. From a steric perspective, the profile of the PON ligand is shown to be significantly larger than those of related N,N'-chelated b-diketiminates ystems, and underpins reactivity of (PON)Gat owards O-atoms ourcest hat proceeds without oligomerization. However,t he enhanced P-donor properties seen on oxidation are not sufficient to quencht he reactivity of the highly polar GaÀ Ou nit. Instead, intramolecular benzylic CÀHa ctivation is observed across the GaÀOb ond of at ransient gallanone intermediate. Furthers tudies of the coordination chemistry of monoanionic [PON] À (and related) ligandst owards main group E(I) centres are in progress and will be reported in due course.

Generalp rocedures
All manipulations were carried out using standard Schlenk line or dry-box techniques under an atmosphere of argon or dinitrogen. Solvents were degassed by sparging with argon and dried by passing through ac olumn of the appropriate drying agent. NMR spectra were measured in [D 6 ]benzene (which was dried over potassium), with the solvent then being distilled under reduced pressure and stored under argon in Te flon valve ampoules. NMR samples were prepared under argon in 5mmW ilmad 507-PP tubes fitted with J. Young Te flon valves. 1 H, 31 P{ 1 H}, 13 C{ 1 H}, 11 B{ 1 H}, 19 F{ 1 H} NMR spectra were measured on aB ruker AvanceI II HD nanobay 400 MHz or Bruker Avance 500 MHz spectrometer at ambient temperature and referenced internally to residual protio-solvent ( 1 H) or solvent ( 13 C) resonances and are reported relative to tetramethylsilane (d = 0ppm). Assignments were confirmed using two-dimensional 1 H-1 Ha nd 13 C-1 HNMR correlation experiments. Chemical shifts are quoted in d (ppm) and coupling constants in Hz. Elemental analyses were carried out by Elemental Microanalysis Ltd, Okehampton, Devon, UK. Compound 1, [15] B(C 6 F 5 ) 3 [23] and KCH 2 Ph [24] were prepared by literature methods. All other reagents were used as received. The synthetic and characterizing data for H(PON) (2) and (PON)GaI 2 are included in the Supporting Information.

Crystallographic details
Single-crystal X-ray diffraction data for compounds 2-7 and (PON)GaI 2 were collected at 150 Ko na nO xford Diffraction/Agilent SuperNova diffractometer using Cu-Ka radiation (l = 1.54184 ), and equipped with an itrogen gas Oxford Cryosystems cooling unit. [25] Raw frame data were reduced using CrysAlisPro. [26] The structures were solved using SHELXT [27] and refined to convergence on F 2 by full-matrix least-squares using SHELXL [28] in combination with OLEX2. [29] Distances and angles were calculated using the full covariance matrix. Restraints were used to maintain sensible geometries for the disordered groups and approximate the displacement parameters to typical values. Selected crystallographic data are summarized in the Table S1. Deposition Numbers 2036919, 2036920, 2036921, 2036922, 2036923, 2036924, and 2036925 contain the supplementary crystallographic data for this paper.T hese data are provided free of charge by the joint Cambridge Crystallo-

Density functional theory (DFT) calculations
The computational work was performed using DFT within the Gaussian09 (Revision D.01) program package. [30] Geometry optimizations of the monoanionic ligand systems were performed with the PBE1PBE hybrid exchange-correlation functional [31] using a TZVP basis set. [32] Grimme's empirical dispersion correction (DFT-D3) was included in all geometry optimizations. [33] Unless otherwise stated, geometry optimizations were carried out for the full system, and frequency calculations were performed to confirm the nature of the stationary points found (minimum). Graphics were created with the Avogadro program. [34] The natural bond orbital (NBO) analyses were performed using NBO 3.1 as implemented in Gaussian09. [35] Syntheses of novel compounds [K(PON)] 2 ,3 :To luene (30 mL) was added to aS chlenk flask containing am ixture of 2 (1.01 g, 1.50 mmol) and benzylpotassium (0.21 g, 1.60 mmol). The reaction mixture was stirred for 3h at room temperature, filtered and volatiles removed from the filtrate in vacuo to give 3 as an off-white solid (1.02 g, 95 %). Single crystals suitable for X-ray crystallography were obtained by slow evaporation from at oluene solution. 1  (PON)Ga, 4: To am ixture of Ga metal (0.12 g, 1.70 mmol) and iodine (0.22 g, 0.65 mmol) in aS chlenk flask was added toluene (30 mL). The mixture was sonicated until ag reen precipitate appeared, and as olution of 3 (1.00 g, 1.50 mmol) in toluene (20 mL) added dropwise. After stirring at room temperature overnight, the reaction mixture was allowed to settle and filtered by cannula. The resulting filtrate was concentrated (to ca. 20 mL) and crystalline product obtained by slow evaporation at room temperature.