Monofluorophos–Metal Complexes: Ripe for Future Discoveries in Homogeneous Catalysis

The discovery that cyclic (ArO)2PF can support Rh-catalysts for hydroformylation with significant advantages in tuning regioselectivity transformed the study of metal complexes of monofluorophos ligands from one of primarily academic interest to one with potentially important applications in catalysis. In this review, the syntheses of monofluorophosphites, (RO)2PF, and monofluorophosphines, R2PF, are discussed and the factors that control the kinetic stability of these ligands to hydrolysis and disproportionation are set out. A survey of the coordination chemistry of these two classes of monofluorophos ligands with d-block metals is presented, emphasising the bonding of the fluorophos to d-block metals, predominantly in low oxidation states. The application of monofluorophos ligands in homogeneous catalysis (especially hydroformylation and hydrocyanation) is discussed, and it is argued that there is great potential for monofluorophos complexes in future catalytic applications.


Introduction
Phosphorus ligands containing P-C, P-N, and P-O bonds are ubiquitous in homogeneous catalysis.By contrast, fluorophos ligands (those containing a P-F bond) have attracted relatively little attention in catalysis, despite the extensive fluorophos coordination chemistry of late transition metals that has been developed and the industrial interest in the application of monofluorophosphite L1 (Figure 1) in Rh-catalysed hydroformylation dating back to 1998 [1].In other contexts, L1 (commercial name: Ethanox 398) has been employed as an antioxidant [2] and as a flame retardant [3].
The extreme electronegativity of fluorine means that it can withdraw electron density from any atom it is bonded to, contributing to its reputation as the Tyrannosaurus Rex of chemistry [4].It should be noted that the electron-withdrawing power of F is a σ-inductive effect and, in some cases, this is offset by an electron-donating π-resonance effect (see later) [5].This property, combined with the diminutive size of P-F (only P-H is smaller), makes the steric and electronic properties of an F substituent of particular academic interest.The high electronegativity of F would be expected to enhance the π-acceptor capacity of ligands containing P-F bonds compared to analogous ligands containing P-O bonds.Since one of the reasons cited for the success of phosphites such as L2-4 (Figure 1) as ligands in Rh-catalysed hydroformylation is their strong π-acceptor capacity, it is understandable why monofluorophosphite L1 performs well in hydroformylation [6][7][8][9][10][11].
The simplest fluorophos ligand, PF 3 , has a special place in coordination and organometallic chemistry as a ligand that has π-acceptor properties on par with, or surpassing, those of CO [12].The volatility of some PF 3 complexes has made them attractive for applications in chemical vapour deposition [13][14][15] and recently, a PF 3 complex, identified as [Co 2 (µ-CO) 2 (CO) 2 (PF 3 ) 4 ], was reported to be a catalyst precursor for 1-hexene hydroformylation [16].However, progress in the application of PF 3 as an ancillary ligand is hampered by it being an odourless gas with toxicity similar to phosgene [17], and it is not amenable to chemical modification.The simplest fluorophos ligand, PF3, has a special place in coordination and organometallic chemistry as a ligand that has π-acceptor properties on par with, or surpassing, those of CO [12].The volatility of some PF3 complexes has made them attractive for applications in chemical vapour deposition [13][14][15] and recently, a PF3 complex, identified as [Co2(µ-CO)2(CO)2(PF3)4], was reported to be a catalyst precursor for 1-hexene hydroformylation [16].However, progress in the application of PF3 as an ancillary ligand is hampered by it being an odourless gas with toxicity similar to phosgene [17], and it is not amenable to chemical modification.
There are no such disadvantages for the collage of P-F ligands, depicted in Figure 2, which have C-, O-, or N-substituents.These substituted fluorophos ligands have the advantages of being systematically modifiable via R substituents and they are generally straightforward to synthesise.The focus of this review is acyclic and cyclic monofluorophos ligands of the type (RO)2P-F and R2P-F, since these are amongst the simplest achiral P III compounds that Ethanox-398 (L1) and some other landmark phosphite ligands L2-L4 for Rh-catalysed hydroformylation.
There are no such disadvantages for the collage of P-F ligands, depicted in Figure 2, which have C-, O-, or N-substituents.These substituted fluorophos ligands have the advantages of being systematically modifiable via R substituents and they are generally straightforward to synthesise.The simplest fluorophos ligand, PF3, has a special place in coordination and organometallic chemistry as a ligand that has π-acceptor properties on par with, or surpassing, those of CO [12].The volatility of some PF3 complexes has made them attractive for applications in chemical vapour deposition [13][14][15] and recently, a PF3 complex, identified as [Co2(µ-CO)2(CO)2(PF3)4], was reported to be a catalyst precursor for 1-hexene hydroformylation [16].However, progress in the application of PF3 as an ancillary ligand is hampered by it being an odourless gas with toxicity similar to phosgene [17], and it is not amenable to chemical modification.
There are no such disadvantages for the collage of P-F ligands, depicted in Figure 2, which have C-, O-, or N-substituents.These substituted fluorophos ligands have the advantages of being systematically modifiable via R substituents and they are generally straightforward to synthesise.The focus of this review is acyclic and cyclic monofluorophos ligands of the type (RO)2P-F and R2P-F, since these are amongst the simplest achiral P III compounds that contain a P-F bond.Both of these classes of P-ligand have attracted considerable academic and industrial interest since the 1960s, including in the area of homogeneous catalysis.To the best of our knowledge, there has not previously been a review of monofluorophos ligands, although difluorophos ligands have been reviewed [18].The topics covered in this review include (1) the synthetic routes to monofluorophosphites and monofluorophosphines; (2) the factors controlling the stability of monofluorophos ligands that limit their applications; (3) the transition metal coordination chemistry of monofluorophos ligands that may be pertinent to an understanding of their role in homogeneous catalysis; (4) the homogeneous catalysis that has been reported with metal-monofluorophos complexes.This review is not comprehensive and there is a bias to more recent developments The focus of this review is acyclic and cyclic monofluorophos ligands of the type (RO) 2 P-F and R 2 P-F, since these are amongst the simplest achiral P III compounds that contain a P-F bond.Both of these classes of P-ligand have attracted considerable academic and industrial interest since the 1960s, including in the area of homogeneous catalysis.To the best of our knowledge, there has not previously been a review of monofluorophos ligands, although difluorophos ligands have been reviewed [18].The topics covered in this review include (1) the synthetic routes to monofluorophosphites and monofluorophosphines; (2) the factors controlling the stability of monofluorophos ligands that limit their applications; (3) the transition metal coordination chemistry of monofluorophos ligands that may be pertinent to an understanding of their role in homogeneous catalysis; (4) the homogeneous catalysis that has been reported with metal-monofluorophos complexes.This review is not comprehensive and there is a bias to more recent developments that build upon the early foundational work reported by the groups of Schmutzler and Nixon.The main conclusion that is drawn from this review is that the tunability of the steric and electronic effects in monofluorophosphites and monofluorophosphines augurs well for future applications of these and related classes of P-F ligands in homogeneous catalysis.

Synthesis and Hydrolytic Stability of Monofluorophosphites
Cyclic and acyclic monofluorophosphites are most readily prepared from the corresponding chlorophosphite, PCl(OR) 2 , and a source of fluoride, such as CsF or SbF 3 .The precursor chlorophosphites are prepared from PCl 3 and the appropriate phenol/alcohol, or a siloxy derivative (as exemplified in Scheme 1) [19].Monofluorophosphites have also been made from PCl 2 F, but this precursor is not readily accessible [20,21].Nixon.The main conclusion that is drawn from this review is that the tunability of the steric and electronic effects in monofluorophosphites and monofluorophosphines augurs well for future applications of these and related classes of P-F ligands in homogeneous catalysis.

Synthesis and Hydrolytic Stability of Monofluorophosphites
Cyclic and acyclic monofluorophosphites are most readily prepared from the corresponding chlorophosphite, PCl(OR)2, and a source of fluoride, such as CsF or SbF3.The precursor chlorophosphites are prepared from PCl3 and the appropriate phenol/alcohol, or a siloxy derivative (as exemplified in Scheme 1) [19].Monofluorophosphites have also been made from PCl2F, but this precursor is not readily accessible [20,21].At first sight, the prospects for using any halophos ligand of the type Z2P-Hal (Z = alkyl, aryl, OR, OAr, NR2; Hal = F, Cl, Br, I) in catalysis may appear bleak because of the reactivity of P-Hal bonds.For example, chlorophos compounds (Z2P-Cl) are normally viewed as useful intermediates rather than ligands because they react readily with a wide range of C-, O-, or N-nucleophiles [22]; this reactivity makes chlorophos ligands incompatible with many reactive functional groups.Moreover, chlorophos compounds commonly fume in air because of their high susceptibility to hydrolysis, during which HCl is produced (Equation (1), X = Cl). (1) The favourable thermodynamics of P-Cl hydrolysis are largely driven by the P=O bond formation in the P-containing product (Equation (1), X = Cl).However, the thermodynamics of P-OAr hydrolysis (Equation (1), X = OAr) are at least as favourable as those of P-Cl hydrolysis and yet ligands containing P-OAr groups are widely used in coordination chemistry and catalysis.It can therefore be surmised that the high reactivity of chlorophosphites is primarily due to their high kinetic lability.Indeed, chlorophosphites that are remarkably stable to moisture have also been developed and some have been applied in catalysis [23,24] . .It has been shown that phosphite P-O bonds can be stabilised to hydrolysis by integrating them into cyclic structures and/or incorporating bulky hydrophobic groups into the ligand framework, as in aryl phosphite ligands L2-4 (Figure 1).Indeed, diphosphite L3 and its derivatives have been successfully applied in large scale industrial hydroformylation processes [7].It is of no surprise, therefore, that the Eastman monofluorophos At first sight, the prospects for using any halophos ligand of the type Z 2 P-Hal (Z = alkyl, aryl, OR, OAr, NR 2 ; Hal = F, Cl, Br, I) in catalysis may appear bleak because of the reactivity of P-Hal bonds.For example, chlorophos compounds (Z 2 P-Cl) are normally viewed as useful intermediates rather than ligands because they react readily with a wide range of C-, O-, or N-nucleophiles [22]; this reactivity makes chlorophos ligands incompatible with many reactive functional groups.Moreover, chlorophos compounds commonly fume in air because of their high susceptibility to hydrolysis, during which HCl is produced (Equation (1), X = Cl).Nixon.The main conclusion that is drawn from this review is that the tunability of the steric and electronic effects in monofluorophosphites and monofluorophosphines augurs well for future applications of these and related classes of P-F ligands in homogeneous catalysis.

Synthesis and Hydrolytic Stability of Monofluorophosphites
Cyclic and acyclic monofluorophosphites are most readily prepared from the corresponding chlorophosphite, PCl(OR)2, and a source of fluoride, such as CsF or SbF3.The precursor chlorophosphites are prepared from PCl3 and the appropriate phenol/alcohol, or a siloxy derivative (as exemplified in Scheme 1) [19].Monofluorophosphites have also been made from PCl2F, but this precursor is not readily accessible [20,21].At first sight, the prospects for using any halophos ligand of the type Z2P-Hal (Z = alkyl, aryl, OR, OAr, NR2; Hal = F, Cl, Br, I) in catalysis may appear bleak because of the reactivity of P-Hal bonds.For example, chlorophos compounds (Z2P-Cl) are normally viewed as useful intermediates rather than ligands because they react readily with a wide range of C-, O-, or N-nucleophiles [22]; this reactivity makes chlorophos ligands incompatible with many reactive functional groups.Moreover, chlorophos compounds commonly fume in air because of their high susceptibility to hydrolysis, during which HCl is produced (Equation (1), X = Cl). (1) The favourable thermodynamics of P-Cl hydrolysis are largely driven by the P=O bond formation in the P-containing product (Equation (1), X = Cl).However, the thermodynamics of P-OAr hydrolysis (Equation (1), X = OAr) are at least as favourable as those of P-Cl hydrolysis and yet ligands containing P-OAr groups are widely used in coordination chemistry and catalysis.It can therefore be surmised that the high reactivity of chlorophosphites is primarily due to their high kinetic lability.Indeed, chlorophosphites that are remarkably stable to moisture have also been developed and some have been applied in catalysis [23,24] . .It has been shown that phosphite P-O bonds can be stabilised to hydrolysis by integrating them into cyclic structures and/or incorporating bulky hydrophobic groups into the ligand framework, as in aryl phosphite ligands L2-4 (Figure 1).Indeed, diphosphite L3 and its derivatives have been successfully applied in large scale industrial hydroformylation processes [7].It is of no surprise, therefore, that the Eastman monofluorophos The favourable thermodynamics of P-Cl hydrolysis are largely driven by the P=O bond formation in the P-containing product (Equation (1), X = Cl).However, the thermodynamics of P-OAr hydrolysis (Equation (1), X = OAr) are at least as favourable as those of P-Cl hydrolysis and yet ligands containing P-OAr groups are widely used in coordination chemistry and catalysis.It can therefore be surmised that the high reactivity of chlorophosphites is primarily due to their high kinetic lability.Indeed, chlorophosphites that are remarkably stable to moisture have also been developed and some have been applied in catalysis [23,24].
It has been shown that phosphite P-O bonds can be stabilised to hydrolysis by integrating them into cyclic structures and/or incorporating bulky hydrophobic groups into the ligand framework, as in aryl phosphite ligands L2-4 (Figure 1).Indeed, diphosphite L3 and its derivatives have been successfully applied in large scale industrial hydroformylation processes [7].It is of no surprise, therefore, that the Eastman monofluorophos ligand L1 is a phosphadioxacycle which contains bulky t-butyl substituents that shroud the P-F moiety [1].
Molecules 2024, 29, x FOR PEER REVIEW 4 of 28 ligand L1 is a phosphadioxacycle which contains bulky t-butyl substituents that shroud the P-F moiety [1].

Coordination Chemistry of Monofluorophosphites
Metal complexes of monofluorophosphites have been produced by the two routes shown in Scheme 2: (a) by substitution of a labile, neutral ligand (A) by a monofluorophosphite; (b) by methanolysis of a coordinated PF3 or by addition of an equivalent of HF to a coordinated P(OR)3.
The cis-disubstituted Mo complexes 6-9 were prepared by substitution of the norbornadiene ligand in [Mo(nbd)(CO)4] with the cyclic monofluorophosphites L5-L8 and the products were fully characterised, including by X-ray crystallography.The IR data for 6-9 are consistent with the π-acceptor capacities of L5-L8 lying between those of PF3 and P(OPh)3.The νCO values for the highest frequency band increases in the order P(OPh)3 < L8 ~ L7 < L6 < L5 < PF3, which is consistent with the π-acceptor capacity of the cyclic phosphites increasing as the ring size decreases [19].

Coordination Chemistry of Monofluorophosphites
Metal complexes of monofluorophosphites have been produced by the two routes shown in Scheme 2: (a) by substitution of a labile, neutral ligand (A) by a monofluorophosphite; (b) by methanolysis of a coordinated PF 3 or by addition of an equivalent of HF to a coordinated P(OR) 3 .
Molecules 2024, 29, x FOR PEER REVIEW 4 of 28 ligand L1 is a phosphadioxacycle which contains bulky t-butyl substituents that shroud the P-F moiety [1].

Coordination Chemistry of Monofluorophosphites
Metal complexes of monofluorophosphites have been produced by the two routes shown in Scheme 2: (a) by substitution of a labile, neutral ligand (A) by a monofluorophosphite; (b) by methanolysis of a coordinated PF3 or by addition of an equivalent of HF to a coordinated P(OR)3.
The cis-disubstituted Mo complexes 6-9 were prepared by substitution of the norbornadiene ligand in [Mo(nbd)(CO)4] with the cyclic monofluorophosphites L5-L8 and the products were fully characterised, including by X-ray crystallography.The IR data for 6-9 are consistent with the π-acceptor capacities of L5-L8 lying between those of PF3 and P(OPh)3.The νCO values for the highest frequency band increases in the order P(OPh)3 < L8 ~ L7 < L6 < L5 < PF3, which is consistent with the π-acceptor capacity of the cyclic phosphites increasing as the ring size decreases [19].

Group 6 Metal Complexes of Monofluorophosphites
The range of Group 6 metal(0) complexes of monofluorophosphites that have been prepared is summarised in Scheme 3 [20,26,27].UV photolysis of each of the metal hexacarbonyls in the presence of (MeO) 2 PF (L9) gave the homoleptic complexes 1-3 [27].
The cis-disubstituted Mo complexes 6-9 were prepared by substitution of the norbornadiene ligand in [Mo(nbd)(CO) 4 ] with the cyclic monofluorophosphites L5-L8 and the products were fully characterised, including by X-ray crystallography.The IR data for 6-9 are consistent with the π-acceptor capacities of L5-L8 lying between those of PF 3 and P(OPh) 3 .The ν CO values for the highest frequency band increases in the order P(OPh) 3 < L8 ~L7 < L6 < L5 < PF 3 , which is consistent with the π-acceptor capacity of the cyclic phosphites increasing as the ring size decreases [19].
Complex 12 has been identified by IR spectroscopy as a product of the methanolysis of the PF 3 complex 14 in a detailed study of the alcoholysis of [Fe(PF 3 ) x (CO) 5-x ] (x = 1-4) species [29].
Complex 12 has been identified by IR spectroscopy as a product of the methanolysis of the PF3 complex 14 in a detailed study of the alcoholysis of [Fe(PF3)x(CO)5-x] (x = 1-4) species [29].The equilibrium proportions of equatorial (e) and apical (a) isomers of [Fe(CO) 4 L] can be determined by IR spectroscopy; sterically demanding and good π-acceptor ligands prefer to bind at the equatorial sites [29].As shown in Scheme 4 (d), for complex 14, the predominant isomer has the PF 3 equatorial, although the e:a ratio is close to the statistical 60:40 ratio, reflecting the similarity of PF 3 and CO as ligands.For complex 12, only the apical isomer was detected, consistent with L9 being small and a poorer π-acceptor than PF 3 .For complex 11, a higher proportion of equatorial isomer was present than even in the PF 3 complex 14, as expected for the bulky L10.The ν CO values for the complexes 14 and 11 are very similar, showing that PF 3 and L10 have similar π-acceptor properties.This demonstrates that the steric and electronic effects of monofluorophosphite ligands can be controlled via the phosphorus alkoxy substituents.
The ruthenium(II) phosphite complexes trans-[(dppe) 2 Ru(H){P(OR) 3 }] + react with HBF 4 to give the homologous series of monofluorophosphite complexes 15-17 (Scheme 5); the HBF 4 is providing the source of HF in these reactions.The coordinated monofluorophosphite ligands L9, L11, and L12 are readily displaced by a H 2 to give the η 2 -H 2 complex 18 (Scheme 5) [30].The equilibrium proportions of equatorial (e) and apical (a) isomers of [Fe(CO)4L] can be determined by IR spectroscopy; sterically demanding and good π-acceptor ligands prefer to bind at the equatorial sites [29].As shown in Scheme 4 (d), for complex 14, the predominant isomer has the PF3 equatorial, although the e:a ratio is close to the statistical 60:40 ratio, reflecting the similarity of PF3 and CO as ligands.For complex 12, only the apical isomer was detected, consistent with L9 being small and a poorer π-acceptor than PF3.For complex 11, a higher proportion of equatorial isomer was present than even in the PF3 complex 14, as expected for the bulky L10.The νCO values for the complexes 14 and 11 are very similar, showing that PF3 and L10 have similar π-acceptor properties.This demonstrates that the steric and electronic effects of monofluorophosphite ligands can be controlled via the phosphorus alkoxy substituents.

Group 9 Metal Complexes of Monofluorophosphites
The tetrahedral cobalt complexes 19 and 20 containing the coordinated L9 have been separated by preparative GLC from the mixtures obtained by methanolysis of the corresponding PF3 complexes (Scheme 6) [31].The IR spectra of the complexes showed that the νCO and νNO stretching bands are both shifted to significantly lower wavenumber in the

Group 9 Metal Complexes of Monofluorophosphites
The tetrahedral cobalt complexes 19 and 20 containing the coordinated L9 have been separated by preparative GLC from the mixtures obtained by methanolysis of the corresponding PF 3 complexes (Scheme 6) [31].The IR spectra of the complexes showed that the ν CO and ν NO stretching bands are both shifted to significantly lower wavenumber in the monofluorophosphite complexes 19 and 20 with respect to their PF 3 precursors, consistent with L9 being a poorer π-acceptor ligand than PF 3 .Scheme 5.In situ generation of monofluorophosphite ligands on ruthenium(II).

Group 9 Metal Complexes of Monofluorophosphites
The tetrahedral cobalt complexes 19 and 20 containing the coordinated L9 have been separated by preparative GLC from the mixtures obtained by methanolysis of the corresponding PF3 complexes (Scheme 6) [31].The IR spectra of the complexes showed that the νCO and νNO stretching bands are both shifted to significantly lower wavenumber in the monofluorophosphite complexes 19 and 20 with respect to their PF3 precursors, consistent with L9 being a poorer π-acceptor ligand than PF3.There is a consistent trend of increasing proportion of binuclear complex 21 formed with decreasing ring size; indeed, with L5, binuclear 21d is exclusively formed.It is significant that PF 3 is the only other monophos ligand that selectively forms the binuclear product 21e [32,33].The interpretation of these observations is that L5 and PF 3 are sufficiently good π-acceptors to displace the CO from the Rh.
The trend of increasing PF 3 -like behaviour with decreasing size of phosphacycle in relation to the reactions of L8-L5 with [Rh 2 Cl 2 (CO) 4 ] parallels the trend observed in the spectroscopic properties of cis-[Mo(CO) 4 (L) 2 ] (see above) [19].
clear NMR spectroscopy and comparison of the spectra with the products exclusively formed from [Rh2Cl2(diene)2] (diene = 1,5-hexadiene or 1,5-cyclooctadiene) and [Rh(cod)2][BF4].There is a consistent trend of increasing proportion of binuclear complex 21 formed with decreasing ring size; indeed, with L5, binuclear 21d is exclusively formed.It is significant that PF3 is the only other monophos ligand that selectively forms the binuclear product 21e [32,33].The interpretation of these observations is that L5 and PF3 are sufficiently good π-acceptors to displace the CO from the Rh.
The trend of increasing PF3-like behaviour with decreasing size of phosphacycle in relation to the reactions of L8-L5 with [Rh2Cl2(CO)4] parallels the trend observed in the spectroscopic properties of cis-[Mo(CO)4(L)2] (see above) [19].Scheme 7. Cyclic monofluorophosphite chemistry of rhodium(I).

Catalysis with Complexes of Monofluorophosphites 2.3.1. Hydroformylation Catalysis with Rhodium Complexes of Monofluorophosphites
The most notable example of the application of monofluorophos ligands in homogeneous catalysis is the use of cyclic monofluorophosphites such as L1 in the Rh-catalysed hydroformylation reactions, reported by Eastman and shown in Scheme 11 [1,25].Initially, the application of monofluorophosphite ligands in catalysis was approached with scepticism, as it was suspected that monofluorophosphites may be thermally unstable, and be prone to hydrolysis, especially at elevated temperatures, generating hydrogen fluoride (HF), which is a known catalyst poison [25,38,39].However, it was demonstrated that L1 is stable to degradation at temperatures up to 350 • C and stable to hydrolysis even in refluxing aqueous isopropanol, with no free fluoride ions detected [40].While acidic conditions promote the degradation of monofluorophosphites, it has been shown that the catalyst system can be stabilised by the addition of an epoxide or a complex such as [Co(acac) 3 ] [41,42].Scheme 8. Nickel(0) and platinum(0) chemistry of monofluorophosphites.
The trans-palladium(II) and cis-platinum(II) complexes 28 and 29, containing the cyclic monofluorophosphite L5, were prepared by cleavage of the corresponding binuclear complex (Scheme 9) [37].The phosphacycle L14, which can be viewed as a saturated analogue of L5, forms the cis-platinum(II) complex 30; comparison of the 31 P NMR parameters for 29 and 30 shows that they are similar, e.g., JPt,P = 5600 and 5490 Hz, respectively.The platinum(0) complex 31 contains monofluorophosphite L15, a saturated analogue of L6 (Scheme 9) [37].The tetrahedral platinum(0) complexes 32a-d are readily formed by the addition of 4 equiv. of L5-L8 to [Pt(nbe)3] (nbe = norbornene).Complex 32b crystallised from solution even when a sub-stoichiometric amount of L6 was added (Scheme 10).However, the addition of 2 equiv. of L5, L7 or L8 to [Pt(nbe)3] in THF gave mixtures of [Pt(L)4] (32a,c,d) [Pt(L)2(nbe)] (33a,c,d), and [Pt(L)(nbe)2] (34a,c,d), identified from their characteristic 31 P  and 195 Pt NMR signals (Scheme 10) [19].The ratios of complexes observed at equilibrium (Scheme 10) were rationalised to be the result of the competing steric and electronic factors for the nbe and monofluorophosphite ligands; for example, while [Pt(L)4] is more sterically crowded than [Pt(L)2(nbe)], the greater π-acceptor properties of monofluorophosphites makes them better than norbornene at stabilising Pt(0) [19].The striking stability of L1 is attributed to the 8-membered phosphacycle which entropically stabilises the ligand to P-O cleavage and to the t Bu substituents which sterically shield the P atom and provide a hydrophobic environment in the vicinity of the P-F bond.
The hydroformylation of ethylene to produce propionaldehyde (Scheme 11, reaction iii) is a potentially useful transformation but acetylene, typically present in ethylene feedstocks in small quantities, acts as a reversible poison towards Rh-based catalysts [25].The activity of ethylene hydroformylation using a Rh-PPh3 catalyst suffered greatly when subjected to ethylene containing 1000 ppm of acetylene.By contrast, the Rh-L1 catalyst system was shown to be remarkably acetylene-tolerant under the same conditions; the activity of the Rh-L1 catalyst eventually deteriorated upon increasing the concentration of acetylene to 10 000 ppm [48].
The hydroformylation of formaldehyde (in the form of paraformaldehyde) is potentially a valuable route to produce glycolaldehyde (Scheme 11, reaction iv) which can then be hydrogenated to ethylene glycol.It has been shown that a Rh-L1 catalyst is more active and selective than a Rh-PPh3 catalyst under the same conditions [49].

Other Catalytic Reactions with Monofluorophosphite Ligands
The bulky, optically active monofluorophosphite BIFOP-F (L19), derived from fenchol, has been employed in the intramolecular Pd-catalysed cross-coupling reaction shown in Scheme 12 [50].A library of 12 related fenchol-derived BIFOP-X ligands were screened for catalysis and complex 35, derived from L19, was the most enantioselective (64% ee) and gave good yields (88%).Ligand L1 exists as two geometric isomers, labelled cis-L1 and trans-L1 in Figure 4, associated with the relative stereochemistry of the F substituent on P and the Me substituent on the CH of the ligand backbone.The isomers of L1 have been separated, and it was shown by 31 P NMR spectroscopy that, when [Rh(CO) 2 (acac)] was treated with 2 equiv. of cis-L1, a mono-ligated RhL 1 species was produced whereas with 2 equiv. of trans-L1 a bisligated RhL 2 species was the product.Furthermore, trans-L1 readily displaced cis-L1 from its Rh(acac) complex, showing that trans-L1 has a greater affinity for the Rh(I) centre than cis-L1 [43,44].These differences in coordination chemistry are likely due to the 8-membered heterocycle having to adopt a more strained ring conformation in cis-L1 than in trans-L1 in order to accommodate the bulky metal moiety being bound at a pseudo-equatorial site.The observed coordination chemistry differences of the isomers of L1 may be the source of the differences in hydroformylation activity and selectivity that are observed with the various mixtures of isomers of L1 [43,44].
The alkene substrates employed in Rh/L1 catalysed hydroformylations include terminal alkenes (1-propene and 1-octene), and internal alkenes (isomeric nonenes) [1,38].As a consequence of their unsymmetrical nature, alkenes other than ethene give linear (l) and branched (b) aldehydes.For propene, two isomeric aldehydes (one linear and one branched) are formed (reaction i in Scheme 11), while for longer chain alkenes, alkene isomerisation is a competing reaction which can lead to several branched aldehyde products, e.g., for 1-hexene, there are two branched isomers (see reaction ii where R = n Pr in Scheme 11).The l:b ratio of products is affected by a wide array of factors, including temperature, syngas pressure, ligand-metal (L:Rh) ratio, and the nature of the ligands [7][8][9]25,45].With monofluorophosphite ligands, it has been shown that the impact of the L:Rh ratio on the alkene hydroformylation activity is strongly dependent on the structure of the ligand.Increasing the L:Rh ratio (L = P-donor ligand) normally decreases catalytic activity, and this is indeed observed with monofluorophosphite L1.However, with the bulkier cyclic monofluorophosphite L16, increasing the L:Rh ratio increased catalytic activity.The cyclic structure of L16 appears to be critical for this unusual concentration effect on rate, since the conventional decrease in activity with increase in L:Rh is observed with L17, an acyclic analogue of L16 [46].The alkene substrates employed in Rh/L1 catalysed hydroformylations include terminal alkenes (1-propene and 1-octene), and internal alkenes (isomeric nonenes) [1,38].As a consequence of their unsymmetrical nature, alkenes other than ethene give linear (l) and branched (b) aldehydes.For propene, two isomeric aldehydes (one linear and one branched) are formed (reaction i in Scheme 11), while for longer chain alkenes, alkene isomerisation is a competing reaction which can lead to several branched aldehyde products, e.g., for 1-hexene, there are two branched isomers (see reaction ii where R = n Pr in Scheme 11).The l:b ratio of products is affected by a wide array of factors, including temperature, syngas pressure, ligand-metal (L:Rh) ratio, and the nature of the ligands [7][8][9]25,45].With monofluorophosphite ligands, it has been shown that the impact of the L:Rh ratio on the alkene hydroformylation activity is strongly dependent on the structure of the ligand.Increasing the L:Rh ratio (L = P-donor ligand) normally decreases catalytic activity, and this is indeed observed with monofluorophosphite L1.However, with the bulkier cyclic monofluorophosphite L16, increasing the L:Rh ratio increased catalytic activity.The cyclic structure of L16 appears to be critical for this unusual concentration effect on rate, since the conventional decrease in activity with increase in L:Rh is observed with L17, an acyclic analogue of L16 [46].
A thorough study of the alkene hydroformylation catalytic properties of Rh complexes of monofluorophosphite L18 has been reported, which includes in-flow and batch hydroformylation of propene, 1-octene, and 2-octene [47].High activities, with TOF up to 75 000 mol(RCHO) mol(Rh) −1 h -1 , have been observed and outstanding control of the aldehyde l:b ratio can be achieved by modulating the temperature, PCO, PH2, time of reaction, the pre-activation of the catalyst, and Rh:L18 ratio; for example, for 1-octene, the l:b ratio can be 'tuned' from 0.27 to 15 (corresponding to selectivity ranging from 78% branched to 94% linear).The higher the concentration of L18, the more the linear aldehyde is favoured, and this has been rationalised by postulating two mechanisms are operating in parallel: one based on RhL2(CO) species, favouring linear aldehyde formation, and the other based on the less bulky RhL(CO)2 moiety, favouring branched aldehyde formation [47].A thorough study of the alkene hydroformylation catalytic properties of Rh complexes of monofluorophosphite L18 has been reported, which includes in-flow and batch hydroformylation of propene, 1-octene, and 2-octene [47].High activities, with TOF up to 75,000 mol(RCHO) mol(Rh) −1 h −1 , have been observed and outstanding control of the aldehyde l:b ratio can be achieved by modulating the temperature, PCO, PH 2 , time of reaction, the pre-activation of the catalyst, and Rh:L18 ratio; for example, for 1-octene, the l:b ratio can be 'tuned' from 0.27 to 15 (corresponding to selectivity ranging from 78% branched to 94% linear).The higher the concentration of L18, the more the linear aldehyde is favoured, and this has been rationalised by postulating two mechanisms are operating in parallel: one based on RhL 2 (CO) species, favouring linear aldehyde formation, and the other based on the less bulky RhL(CO) 2 moiety, favouring branched aldehyde formation [47].
The hydroformylation of ethylene to produce propionaldehyde (Scheme 11, reaction iii) is a potentially useful transformation but acetylene, typically present in ethylene feedstocks in small quantities, acts as a reversible poison towards Rh-based catalysts [25].The activity of ethylene hydroformylation using a Rh-PPh 3 catalyst suffered greatly when subjected to ethylene containing 1000 ppm of acetylene.By contrast, the Rh-L1 catalyst system was shown to be remarkably acetylene-tolerant under the same conditions; the activity of the Rh-L1 catalyst eventually deteriorated upon increasing the concentration of acetylene to 10,000 ppm [48].
The hydroformylation of formaldehyde (in the form of paraformaldehyde) is potentially a valuable route to produce glycolaldehyde (Scheme 11, reaction iv) which can then be hydrogenated to ethylene glycol.It has been shown that a Rh-L1 catalyst is more active and selective than a Rh-PPh 3 catalyst under the same conditions [49].

Other Catalytic Reactions with Monofluorophosphite Ligands
The bulky, optically active monofluorophosphite BIFOP-F (L19), derived from fenchol, has been employed in the intramolecular Pd-catalysed cross-coupling reaction shown in Scheme 12 [50].A library of 12 related fenchol-derived BIFOP-X ligands were screened for catalysis and complex 35, derived from L19, was the most enantioselective (64% ee) and gave good yields (88%).
An attempt to use the same ligand L19 in a Cu-catalysed 1,4-addition of R 2 Zn or RMgBr (R = Me, Et) to enones was unsuccessful; it was suggested that L19 was unstable under the reaction conditions used [51].An attempt to use the same ligand L19 in a Cu-catalysed 1,4-addition of R2Zn or RMgBr (R = Me, Et) to enones was unsuccessful; it was suggested that L19 was unstable under the reaction conditions used [51].

Synthesis and Stability of Monofluorophosphines
Two general routes to R2PF where R = alkyl or aryl are shown in Scheme 13.The R2PCl route has the advantage of the ready availability of chlorophosphines from PCl3 but the Cl2PF route can provide access to R2PF for which the corresponding R2PCl is unknown, as demonstrated for (PhC≡C)2PF [52].Simple R2PF (which are P III species) are generally unstable with respect to the disproportionation to the P V in R2PF3 and P II in R2P-PR2, as shown in Scheme 14 [53,54].The pathway shown in Scheme 14, involving the intermediates A and B, has been proposed for the disproportionation; examples of P III -P V species A have been isolated and characterised spectroscopically [55,56].This chemistry would militate against the application of monofluorophosphines as ligands in homogeneous catalysis unless, under the catalytic reaction conditions, the equilibrium in Scheme 14 lies in favour of the R2PF, or the equilibrium is rapidly reversible, such that it can be entrained via metal complexation (see below).

Synthesis and Stability of Monofluorophosphines
Two general routes to R 2 PF where R = alkyl or aryl are shown in Scheme 13.The R 2 PCl route has the advantage of the ready availability of chlorophosphines from PCl 3 but the Cl 2 PF route can provide access to R 2 PF for which the corresponding R 2 PCl is unknown, as demonstrated for (PhC≡C) 2 PF [52].An attempt to use the same ligand L19 in a Cu-catalysed 1,4-addition of R2Zn or RMgBr (R = Me, Et) to enones was unsuccessful; it was suggested that L19 was unstable under the reaction conditions used [51].

Synthesis and Stability of Monofluorophosphines
Two general routes to R2PF where R = alkyl or aryl are shown in Scheme 13.The R2PCl route has the advantage of the ready availability of chlorophosphines from PCl3 but the Cl2PF route can provide access to R2PF for which the corresponding R2PCl is unknown, as demonstrated for (PhC≡C)2PF [52].Simple R2PF (which are P III species) are generally unstable with respect to the disproportionation to the P V in R2PF3 and P II in R2P-PR2, as shown in Scheme 14 [53,54].The pathway shown in Scheme 14, involving the intermediates A and B, has been proposed for the disproportionation; examples of P III -P V species A have been isolated and characterised spectroscopically [55,56].This chemistry would militate against the application of monofluorophosphines as ligands in homogeneous catalysis unless, under the catalytic reaction conditions, the equilibrium in Scheme 14 lies in favour of the R2PF, or the equilibrium is rapidly reversible, such that it can be entrained via metal complexation (see below).Simple R 2 PF (which are P III species) are generally unstable with respect to the disproportionation to the P V in R 2 PF 3 and P II in R 2 P-PR 2 , as shown in Scheme 14 [53,54].The pathway shown in Scheme 14, involving the intermediates A and B, has been proposed for the disproportionation; examples of P III -P V species A have been isolated and characterised spectroscopically [55,56].This chemistry would militate against the application of monofluorophosphines as ligands in homogeneous catalysis unless, under the catalytic reaction conditions, the equilibrium in Scheme 14 lies in favour of the R 2 PF, or the equilibrium is rapidly reversible, such that it can be entrained via metal complexation.
The stabilising effects of the P-substituents noted in generalisation (2) accounts for the dominance of t Bu2PF (L20) and (CF3)2PF (L21) in the early literature concerning the coordination chemistry of monofluorophosphines (Figure 5).A simple rationale for the R2PF-stabilising effect of bulky and electron-withdrawing substituents is that these substituents raise the energy of the disproportionation diphosphane product, R2P-PR2 because (a) bulky R groups maximise 1,2-steric repulsions in the relatively crowded diphosphanet Bu2P-P t Bu2 has been calculated to have a weak P-P bond [62]; (b) electron-withdrawing groups destabilise the P-P bond due to electrostatic repulsion between the resulting δ+ charges on each of the P atoms-it has been reported that (CF3)2P-P(CF3)2 has an elongated P-P bond [63].A mechanism for disproportionation involving sterically crowded intermediates A and B, which would also be disfavoured by electron-withdrawing substituents [54][55][56], has been proposed (Scheme 14).The monofluorophosphines CgPF (L22), containing a phospha-adamantane cage and the PhobPF species L23 and L24, containing a phospha-bicycle (Figure 5), are remarkably stable to disproportionation [61].The CgP and PhobP moieties are rigid and bulky and so the stability of L22-L24 may be, at least in part, explained using similar steric con- The following generalisations on the stability of R 2 PF to disproportionation (Scheme 14) have been established from extensive studies: (1) Many common R 2 PF (e.g., R = Ph, Me, n Bu) readily disproportionate [54,57,58]; (2) Bulky substituents and electron-withdrawing substituents stabilise R 2 PF with respect to disproportionation [59,60]; (3) Cyclic monofluorophosphines with constrained C-P-C bonds are more stable with respect to disproportionation than acyclic analogues [61].
The stabilising effects of the P-substituents noted in generalisation (2) accounts for the dominance of t Bu 2 PF (L20) and (CF 3 ) 2 PF (L21) in the early literature concerning the coordination chemistry of monofluorophosphines (Figure 5).A simple rationale for the R 2 PF-stabilising effect of bulky and electron-withdrawing substituents is that these substituents raise the energy of the disproportionation diphosphane product, R 2 P-PR 2 because (a) bulky R groups maximise 1,2-steric repulsions in the relatively crowded diphosphanet Bu 2 P-P t Bu 2 has been calculated to have a weak P-P bond [62]; (b) electron-withdrawing groups destabilise the P-P bond due to electrostatic repulsion between the resulting δ+ charges on each of the P atoms-it has been reported that (CF 3 ) 2 P-P(CF 3 ) 2 has an elongated P-P bond [63].A mechanism for disproportionation involving sterically crowded intermediates A and B, which would also be disfavoured by electron-withdrawing substituents [54][55][56], has been proposed (Scheme 14).The following generalisations on the stability of R2PF to disproportionation (Scheme 14) have been established from extensive studies: (1) Many common R2PF (e.g., R = Ph, Me, n Bu) readily disproportionate [54,57,58]; (2) Bulky substituents and electron-withdrawing substituents stabilise R2PF with respect to disproportionation [59,60]; (3) Cyclic monofluorophosphines with constrained C-P-C bonds are more stable with respect to disproportionation than acyclic analogues [61].
The stabilising effects of the P-substituents noted in generalisation (2) accounts for the dominance of t Bu2PF (L20) and (CF3)2PF (L21) in the early literature concerning the coordination chemistry of monofluorophosphines (Figure 5).A simple rationale for the R2PF-stabilising effect of bulky and electron-withdrawing substituents is that these substituents raise the energy of the disproportionation diphosphane product, R2P-PR2 because (a) bulky R groups maximise 1,2-steric repulsions in the relatively crowded diphosphanet Bu2P-P t Bu2 has been calculated to have a weak P-P bond [62]; (b) electron-withdrawing groups destabilise the P-P bond due to electrostatic repulsion between the resulting δ+ charges on each of the P atoms-it has been reported that (CF3)2P-P(CF3)2 has an elongated P-P bond [63].A mechanism for disproportionation involving sterically crowded intermediates A and B, which would also be disfavoured by electron-withdrawing substituents [54][55][56], has been proposed (Scheme 14).The monofluorophosphines CgPF (L22), containing a phospha-adamantane cage and the PhobPF species L23 and L24, containing a phospha-bicycle (Figure 5), are remarkably stable to disproportionation [61].The CgP and PhobP moieties are rigid and bulky and so the stability of L22-L24 may be, at least in part, explained using similar steric congestion arguments to those used above for the stability of L20 [64][65][66][67].In addition, it has The monofluorophosphines CgPF (L22), containing a phospha-adamantane cage, and the PhobPF species L23 and L24, containing a phospha-bicycle (Figure 5), are remarkably stable to disproportionation [61].The CgP and PhobP moieties are rigid and bulky, and so the stability of L22-L24 may be, at least in part, explained using similar steric congestion arguments to those used above for the stability of L20 [64][65][66][67].In addition, it has been argued that the constrained C-P-C angles in L22-L24 also contribute to their observed stability to disproportionation (generalisation (3) above) using the following reasoning [61].The two geometric isomers of R 2 PF 3 have diapical-equatorial (aae) or apical-diequatorial (aee) F groups, with the high apicophilicity of F leading to the aae isomer being preferred for R 2 PF 3 [68].Therefore, the favoured isomer has the two R substituents occupying two equatorial sites with a 120 • angle between them, as depicted in Scheme 14. X-ray crystallography has shown that the C-P-C angles are close to 90 • in multiple compounds containing either the CgP or PhobP moieties [64][65][66][67].Consequently, the observed stability to disproportionation of L22-L24 can be partly attributed to the high degree of C-P-C ring strain in R 2 PF 3 that would be incurred by the 2 C substituents occupying equatorial sites; if, instead, the eea isomer were adopted, there would be an unfavourable cost in the P-F bond energies associated with two of the F substituents occupying equatorial sites [68].

Coordination Chemistry of Monofluorophosphines
In general, monofluorophosphine (R 2 PF) complexes are made just like many other P-ligand complexes: by the substitution of a labile ligand on a precursor complex.In metal complexes of monofluorophosphines, the coordinated R 2 PF is not susceptible to disproportionation.Consequently, ligated Ph 2 PF (which is unstable as the free ligand) has been generated within a Cr, Mo, or W coordination sphere by fluoride substitution of a labile X group on a precursor R 2 PX complex [69-71].
A notable conclusion drawn on the basis of the IR spectra of cis-[MoL 2 (CO) 4 ] and mer-[MoL 3 (CO) 3 ] is that (CF 3 ) 2 PF and CF 3 PF 2 are stronger π-acceptors than PF 3 , notwithstanding the greater electronegativity of F than that of CF 3 (χ of and 3.3, respectively, on the Pauling Scale).It has been suggested [74] that an explanation for this apparent anomaly lies in the π component present in the P-F bond that involves a HOMO (lone pair) orbital on F and the LUMO (σ *) on P which has π symmetry.This is the same orbital on P that is involved in the π backbonding from the metal.Thus, in a M-P-F fragment, the M competes with F for the π acceptor orbital on P (Figure 6(i)); this competition is not present in a M-P-CF 3 fragment which would explain the greater π acceptor capacity of (CF 3 ) 2 PF than PF 3 [74].This explanation in terms of π interactions between the LUMO (σ *) on P and a HOMO with π symmetry on a P-substituent is reminiscent of the arguments used by Woollins et al. to explain why P t Bu(pyrrolyl) 2 is a stronger σ donor than P(pyrrolyl) 3 (see Figure 6(ii)) [75].

Scheme 15. Monofluorophosphine complexes of Group 6 metals.
A notable conclusion drawn on the basis of the IR spectra of cis-[MoL2(CO)4] and mer-[MoL3(CO)3] is that (CF3)2PF and CF3PF2 are stronger π-acceptors than PF3, notwithstanding the greater electronegativity of F than that of CF3 (χ of 4.0 and 3.3, respectively, on the Pauling Scale).It has been suggested [74] that an explanation for this apparent anomaly lies in the π component present in the P-F bond that involves a HOMO (lone pair) orbital on F and the LUMO (σ *) on P which has π symmetry.This is the same orbital on P that is involved in the π backbonding from the metal.Thus, in a M-P-F fragment, the M competes with F for the π acceptor orbital on P (Figure 6(i)); this competition is not present in a M-P-CF3 fragment which would explain the greater π acceptor capacity of (CF3)2PF than PF3 [74].This explanation in terms of π interactions between the LUMO (σ *) on P and a HOMO with π symmetry on a P-substituent is reminiscent of the arguments used by Woollins et al. to explain why P t Bu(pyrrolyl)2 is a stronger σ donor than P(pyrrolyl)3 (see Figure 6(ii)) [75].A notable conclusion drawn on the basis of the IR spectra of cis-[MoL2(CO)4] and mer-[MoL3(CO)3] is that (CF3)2PF and CF3PF2 are stronger π-acceptors than PF3, notwithstanding the greater electronegativity of F than that of CF3 (χ of 4.0 and 3.3, respectively, on the Pauling Scale).It has been suggested [74] that an explanation for this apparent anomaly lies in the π component present in the P-F bond that involves a HOMO (lone pair) orbital on F and the LUMO (σ *) on P which has π symmetry.This is the same orbital on P that is involved in the π backbonding from the metal.Thus, in a M-P-F fragment, the M competes with F for the π acceptor orbital on P (Figure 6(i)); this competition is not present in a M-P-CF3 fragment which would explain the greater π acceptor capacity of (CF3)2PF than PF3 [74].This explanation in terms of π interactions between the LUMO (σ *) on P and a HOMO with π symmetry on a P-substituent is reminiscent of the arguments used by Woollins et al. to explain why P t Bu(pyrrolyl)2 is a stronger σ donor than P(pyrrolyl)3 (see Figure 6(ii)) [75].

Group 7 Metal Complexes of Monofluorophosphines
The only reported Group 7 metal complexes containing a monofluorophosphine ligand are the isomeric hydridomanganese(I) complexes 45 and 46, formed as a 3:1 mixture by the reaction of [HMn(CO) 5 ] with L21 (Scheme 16) [76].The 2 J(HP) values for 46 (72 Hz) and 47 (4 Hz) are consistent with the assignment of their respective trans and cis geometries.

Group 7 Metal Complexes of Monofluorophosphines
The only reported Group 7 metal complexes containing a monofluorophosphine ligand are the isomeric hydridomanganese(I) complexes 45 and 46, formed as a 3:1 mixture by the reaction of [HMn(CO)5] with L21 (Scheme 16) [76].The 2 J(HP) values for 46 (72 Hz) and 47 (4 Hz) are consistent with the assignment of their respective trans and cis geometries.

Group 8 Metal Complexes of Monofluorophosphines
The tetracarbonyliron complex 47 can be generated in situ by photolysis of a mixture of L21 and [Fe(CO)5] and the IR spectrum suggests that 47 is predominantly the equatorial isomer (Scheme 17).This is consistent with L21 being bulkier than PF3 and of comparable π acceptor capacity to it [29].

Group 8 Metal Complexes of Monofluorophosphines
The tetracarbonyliron complex 47 can be generated in situ by photolysis of a mixture of L21 and [Fe(CO) 5 ] and the IR spectrum suggests that 47 is predominantly the equatorial isomer (Scheme 17).This is consistent with L21 being bulkier than PF 3 and of comparable π acceptor capacity to it [29].

Group 8 Metal Complexes of Monofluorophosphines
The tetracarbonyliron complex 47 can be generated in situ by photolysis of a mixture of L21 and [Fe(CO)5] and the IR spectrum suggests that 47 is predominantly the equatorial isomer (Scheme 17).This is consistent with L21 being bulkier than PF3 and of comparable π acceptor capacity to it [29].

Scheme 17. Monofluorophosphine complexes of iron(0).
The anthracene-derived monofluorophosphine L25 and the naphthalene-derived monofluorophosphine L26 were prepared from Cl2PF (Scheme 13) [52].Ligand L25 was purified by distillation and showed no tendency to undergo disproportionation presumably because it is stabilised by its bulky substituents.Reaction of L25 with [Fe2(CO)9] gave complex 48, whose IR spectrum was consistent with C2v symmetry and was therefore assigned to the equatorial isomer.Ligand L26 was not obtained in pure form but the impure material was reacted with [Fe2(CO)9] to produce the iron complex 49, the IR spectrum of which was consistent with C3v symmetry and was therefore assigned to the apical isomer [52].The different geometries assigned to 48 and 49 may be rationalised by L25 being larger and more electron poor (making it a better π-acceptor) than L26.
The unusual monofluorophosphine 50 has been prepared by treatment of its anionic precursor with N-fluoropyridinium tetrafluoroborate which acts as an electrophilic source of F + (see Scheme 18) [77].The P-F bond in 50 was shown to be covalent in the solid state The anthracene-derived monofluorophosphine L25 and the naphthalene-derived monofluorophosphine L26 were prepared from Cl 2 PF (Scheme 13) [52].Ligand L25 was purified by distillation and showed no tendency to undergo disproportionation presumably because it is stabilised by its bulky substituents.Reaction of L25 with [Fe 2 (CO) 9 ] gave complex 48, whose IR spectrum was consistent with C 2v symmetry and was therefore assigned to the equatorial isomer.Ligand L26 was not obtained in pure form but the impure material was reacted with [Fe 2 (CO) 9 ] to produce the iron complex 49, the IR spectrum of which was consistent with C 3v symmetry and was therefore assigned to the apical isomer [52].The different geometries assigned to 48 and 49 may be rationalised by L25 being larger and more electron poor (making it a better π-acceptor) than L26.
The unusual monofluorophosphine 50 has been prepared by treatment of its anionic precursor with N-fluoropyridinium tetrafluoroborate which acts as an electrophilic source of F + (see Scheme 18) [77].The P-F bond in 50 was shown to be covalent in the solid state by single-crystal X-ray diffraction (d P-F = 1.658(4)Å), and in solution by 31 P and 19 F NMR spectroscopy, which showed that 1 J PF = 918 Hz).The data for 50 are comparable to values for conventional R 2 PF compounds: d P-F = 1.619(7)Å for t Bu 2 PF [78]; 1 J PF = 905 Hz for Ph 2 PF [57].In principle, 50 could act as an monofluorophos ligand, but this has not been reported to date.
The osmium cluster complexes 51 and 52 were readily formed by the addition of an excess of the bulky monofluorophosphine L20 to the corresponding labile MeCN complex precursors (Scheme 19) [79].The osmium cluster complexes 51 and 52 were readily formed by the addition of an excess of the bulky monofluorophosphine L20 to the corresponding labile MeCN complex precursors (Scheme 19) [79].
The osmium cluster complexes 51 and 52 were readily formed by the addition of an excess of the bulky monofluorophosphine L20 to the corresponding labile MeCN complex precursors (Scheme 19) [79].

Group 9 Metal Complexes of Monofluorophosphines
The paramagnetic cobalt complexes 53a and 53b were prepared by stirring a suspension of CoX2 in CH2Cl2 with L20.The highly coloured 53a (blue) and 53b (blue-green) had electronic spectra, IR spectra, and magnetic moments (µ ≈ 4.5 BM) consistent with the tetrahedral geometry depicted in Scheme 20 [80].

Group 9 Metal Complexes of Monofluorophosphines
The paramagnetic cobalt complexes 53a and 53b were prepared by stirring a suspension of CoX 2 in CH 2 Cl 2 with L20.The highly coloured 53a (blue) and 53b (blue-green) had electronic spectra, IR spectra, and magnetic moments (µ ≈ 4.5 BM) consistent with the tetrahedral geometry depicted in Scheme 20 [80].Reaction of 1 equiv. of L21 with [Co(CO)3(NO)] at ambient temperatures over 7 days yielded a mixture of monosubstituted and disubstituted complexes 54 and 55, which were separated by fractional distillation [81].The trisubstituted complex 56 was obtained by heating a mixture of [Co(CO)3(NO)] and an excess of L21 to 120 ˚C (Scheme 20).The position of the νNO band in the IR spectra of 54 (1832 cm −1 ), 55 (1842 cm −1 ), and 56 (1854 cm −1 ) are consistent with L21 being a better π-acceptor than CO.
Mononuclear rhodium complexes 57-61 are formed rapidly upon reaction between [Rh2Cl2(CO)4] and the appropriate monofluorophosphine in CH2Cl2 (Scheme 21).The vCO values given in Scheme 21 show that the cage monofluorophosphine L22 is the strongest π-acceptor followed by the dimethoxynaphthalene ligand L26 and then the sym and asym Reaction of 1 equiv. of L21 with [Co(CO) 3 (NO)] at ambient temperatures over 7 days yielded a mixture of monosubstituted and disubstituted complexes 54 and 55, which were separated by fractional distillation [81].The trisubstituted complex 56 was obtained by heating a mixture of [Co(CO) 3 (NO)] and an excess of L21 to 120 • C (Scheme 20).The position of the ν NO band in the IR spectra of 54 (1832 cm −1 ), 55 (1842 cm −1 ), and 56 (1854 cm −1 ) are consistent with L21 being a better π-acceptor than CO.
Mononuclear rhodium complexes 57-61 are formed rapidly upon reaction between [Rh 2 Cl 2 (CO) 4 ] and the appropriate monofluorophosphine in CH 2 Cl 2 (Scheme 21).The v CO values given in Scheme 21 show that the cage monofluorophosphine L22 is the strongest π-acceptor followed by the dimethoxynaphthalene ligand L26 and then the sym and asym isomers of the bicyclic fluorophobanes L23 and L24 straddle the bulky L20 [52,61].Scheme 20.Monofluorophosphine-cobalt complexes.
Reaction of 1 equiv. of L21 with [Co(CO)3(NO)] at ambient temperatures over 7 days yielded a mixture of monosubstituted and disubstituted complexes 54 and 55, which were separated by fractional distillation [81].The trisubstituted complex 56 was obtained by heating a mixture of [Co(CO)3(NO)] and an excess of L21 to 120 ˚C (Scheme 20).The position of the νNO band in the IR spectra of 54 (1832 cm −1 ), 55 (1842 cm −1 ), and 56 (1854 cm −1 ) are consistent with L21 being a better π-acceptor than CO.
Mononuclear rhodium complexes 57-61 are formed rapidly upon reaction between [Rh2Cl2(CO)4] and the appropriate monofluorophosphine in CH2Cl2 (Scheme 21).The vCO values given in Scheme 21 show that the cage monofluorophosphine L22 is the strongest π-acceptor followed by the dimethoxynaphthalene ligand L26 and then the sym and asym isomers of the bicyclic fluorophobanes L23 and L24 straddle the bulky L20 [52,61].The fluoro analogue of Wilkinson's Catalyst, [RhF(PPh 3 ) 3 ], undergoes the rearrangement shown in Scheme 22 to generate complex 62, which contains a 'trapped' Ph 2 P-F ligated to Rh [82].This remarkable isomerisation occurs under mild conditions and is reversible.Several examples are known where late transition metal fluoro complexes with PR 3 ancillary ligands undergo related P-C/M-F rearrangements to generate coordinated R 2 P-F ligands as products or transient intermediates [82].
Molecules 2024, 29, x FOR PEER REVIEW 20 of 28 PR3 ancillary ligands undergo related P-C/M-F rearrangements to generate coordinated R2P-F ligands as products or transient intermediates [82].
Scheme 22. Rearrangement leading to in situ formation of Ph2PF complex.

Group 10 Metal Complexes of Monofluorophosphines
Treatment of nickel tetracarbonyl with an excess of L21 at 25 °C gives predominantly monocarbonyl 63 with traces of dicarbonyl 64, which can be separated by fractional distillation.The fully substituted complex 65 is produced under more forcing conditions (95 °C, 24 h), but the product is contaminated with traces of 63 (Scheme 22) [83].The volatile, air-stable nickel(0) complex 65 can be more readily prepared by mixing L21 with nickelocene [84] or by reaction of L21 with metallic nickel, generated by thermolysis of nickel oxalate at 60 °C (Scheme 23) [85].The reaction between [Ni(cod)2] and the phospha-cage flurophosphine L22 was reported to give complex 66 (Scheme 23), identified in solution on the basis of the stoichiometry used and the characteristic AA'XX' pattern observed in the 31 P NMR spectrum [61].

Group 10 Metal Complexes of Monofluorophosphines
Treatment of nickel tetracarbonyl with an excess of L21 at 25 • C gives predominantly monocarbonyl 63 with traces of dicarbonyl 64, which can be separated by fractional distillation.The fully substituted complex 65 is produced under more forcing conditions (95 • C, 24 h), but the product is contaminated with traces of 63 (Scheme 22) [83].The volatile, air-stable nickel(0) complex 65 can be more readily prepared by mixing L21 with nickelocene [84] or by reaction of L21 with metallic nickel, generated by thermolysis of nickel oxalate at 60 • C (Scheme 23) [85].The reaction between [Ni(cod) 2 ] and the phosphacage flurophosphine L22 was reported to give complex 66 (Scheme 23), identified in solution on the basis of the stoichiometry used and the characteristic AA'XX' pattern observed in the 31 P NMR spectrum [61].
air-stable nickel(0) complex 65 can be more readily prepared by mixing L21 with nickelocene [84] or by reaction of L21 with metallic nickel, generated by thermolysis of nickel oxalate at 60 °C (Scheme 23) [85].The reaction between [Ni(cod)2] and the phospha-cage flurophosphine L22 was reported to give complex 66 (Scheme 23), identified in solution on the basis of the stoichiometry used and the characteristic AA'XX' pattern observed in the 31 P NMR spectrum [61].Diamagnetic nickel(II) complexes 67a-c are formed when suspensions of NiX2 in acetone or toluene are treated with L20 (Scheme 24) [80].The trans geometry of 57a was established from the large 2 JPP of 425 Hz and the crystal structure of 57b confirms its trans geometry in the solid state [86].Diamagnetic nickel(II) complexes 67a-c are formed when suspensions of NiX 2 in acetone or toluene are treated with L20 (Scheme 24) [80].The trans geometry of 57a was established from the large 2 J PP of 425 Hz and the crystal structure of 57b confirms its trans geometry in the solid state [86].Chiral monofluorophosphine L27 disproportionates (Scheme 14) over a period of 16 h, but the rate of the disproportionation for dilute solutions of L27 in benzene was slow enough to measure its optical purity [87].Reaction of a racemic mixture of L27 with the optically pure dipalladium complex shown in Scheme 25 gave a diastereomeric mixture of complexes 68 and 69.Pure complex 68 was obtained selectively by repeated crystallisation from diethyl ether and the absolute configuration at P was determined by X-ray crystallography.Enantiomerically pure S-L27 was then displaced from complex 68 by addition of a chelating diphosphine.It was shown by polarimetry that S-L27 racemised in benzene over a period of 6 h [87].Chiral monofluorophosphine L27 disproportionates (Scheme 14) over a period of 16 h, but the rate of the disproportionation for dilute solutions of L27 in benzene was slow enough to measure its optical purity [87].Reaction of a racemic mixture of L27 with the optically pure dipalladium complex shown in Scheme 25 gave a diastereomeric mixture of complexes 68 and 69.Pure complex 68 was obtained selectively by repeated crystallisation from diethyl ether and the absolute configuration at P was determined by X-ray crystallography.Enantiomerically pure S-L27 was then displaced from complex 68 by addition of a chelating diphosphine.It was shown by polarimetry that S-L27 racemised in benzene over a period of 6 h [87].
optically pure dipalladium complex shown in Scheme 25 gave a diastereomeric mixture of complexes 68 and 69.Pure complex 68 was obtained selectively by repeated crystallisation from diethyl ether and the absolute configuration at P was determined by X-ray crystallography.Enantiomerically pure S-L27 was then displaced from complex 68 by addition of a chelating diphosphine.It was shown by polarimetry that S-L27 racemised in benzene over a period of 6 h [87].The platinum(0) complex 70 was prepared by heating K2[PtCl4] (or PtCl2) with a large excess of L21 followed by prolonged shaking at ambient temperature (Scheme 26); the P V by-product (CF3)2PFCl2 was identified, consistent with L21 acting as the reducing agent [88].Complex 70 was inert to the addition of MeI, HCl, C2H4, or CS2, even upon prolonged heating, in contrast to the triphenylphosphine analogue [Pt(PPh3)4].This behaviour likely reflects the greater π-acceptor properties of L21 stabilising Pt(0) and reducing its nucleophilicity, coupled with the greater steric bulk of PPh3 promoting the formation of reactive, coordinatively unsaturated PtL3 species [88].
The insoluble platinum(0) complex 71 was prepared by the replacement of PMePh2 by L21 in the reaction shown in Scheme 26, a reaction presumably driven by the greater π-acceptor properties of L21 than PMePh2 [89].The platinum(0) complex 70 was prepared by heating K 2 [PtCl 4 ] (or PtCl 2 ) with a large excess of L21 followed by prolonged shaking at ambient temperature (Scheme 26); the P V by-product (CF 3 ) 2 PFCl 2 was identified, consistent with L21 acting as the reducing agent [88].Complex 70 was inert to the addition of MeI, HCl, C 2 H 4 , or CS 2 , even upon prolonged heating, in contrast to the triphenylphosphine analogue [Pt(PPh 3 ) 4 ].This behaviour likely reflects the greater π-acceptor properties of L21 stabilising Pt(0) and reducing its nucleophilicity, coupled with the greater steric bulk of PPh 3 promoting the formation of reactive, coordinatively unsaturated PtL 3 species [88].The substituted diarylfluorophosphines L28, L29, and L30 form the platinum(II) complexes 72, 73, and 74 by the routes shown in Scheme 27.The cis geometry of 72 and 73 was confirmed by their X-ray crystal structures [52,90], and the trans-configuration of 74 was confirmed by the large value of 2 JP,P = 567 Hz [91].
Scheme 27.Platinum(II) complexes of monofluorophosphines.The insoluble platinum(0) complex 71 was prepared by the replacement of PMePh 2 by L21 in the reaction shown in Scheme 26, a reaction presumably driven by the greater π-acceptor properties of L21 than PMePh 2 [89].

Catalysis with Complexes of Monofluorophosphines
The substituted diarylfluorophosphines L28, L29, and L30 form the platinum(II) complexes 72, 73, and 74 by the routes shown in Scheme 27.The cis geometry of 72 and 73 was confirmed by their X-ray crystal structures [52,90], and the trans-configuration of 74 was confirmed by the large value of 2 J P,P = 567 Hz [91].

Hydroformylation Catalysis with Rhodium Complexes of Monofluorophosphines
The first step in the homologation of 1-heptene to 1-octene is the hydroformylation shown in Scheme 28 [92].Rhodium complexes of monofluorophos ligands L20, L22, L23, and L24 all showed catalytic activity comparable to the commercialised Rh-PPh 3 catalyst.The l:b ratio of 3.9 obtained for the Rh-L22 catalyst compares favourably with the l:b ratio of 2.2 for the Rh-PPh 3 catalyst under the same conditions.The 31 P NMR spectrum of the exit solutions for the Rh-L22 catalysis showed the presence of Rh-monofluorophos complexes, indicating that the coordinated L22 had survived the reaction conditions [61].

Conclusions and Prospective Applications of Monofluorophos Ligands in Coordination Chemistry and Catalysis
The combination of the extreme electronegativity and smallness of F has made ligands containing a P-F bond of academic interest for many years.The strength of the P-F bond at 490 kJ mol −1 dwarfs other P-X single bonds (cf.P-C, 264 kJ mol −1 ; P-O, 335 kJ mol −1 ) and is the source of the thermodynamically stability of P-F compounds.PF 3 is often characterised as the ultimate π-acceptor, outstripping even CO in its capacity to stabilise electron-rich, low oxidation state metal complexes.What has attracted particular attention to substituted monofluorophos ligands is their capacity to be 'tunable' analogues of PF 3 and indeed to make ligands such as (CF 3 ) 2 PF which are more powerful π-acceptors.
The focus of this review has been on the coordination chemistry of monofluorophosphites, (RO) 2 PF, and monofluorophosphines, R 2 PF, and the successful applications of monofluorophos-metal complexes in homogeneous catalysis.At the outset, the prospects for applications of monofluorophos ligands in homogeneous catalysis appeared to be inauspicious because of two fundamental instabilities: (1) notwithstanding the great P-F bond strength, monofluorophos compounds are generally susceptible to hydrolysis, a reaction driven by the formation of the even stronger bonds, H-F (565 kJ mol −1 ) and P=O (544 kJ mol −1 ); (2) the propensity of F to stabilise high oxidation states explains the observation that many P III -F compounds readily decompose by disproportionation into P V -F compounds and P II species containing P-P bonds.
The 1998 report by Puckette and coworkers at Eastmann of the application of the cyclic monofluorophosphite L1 in Rh-catalysed hydroformylation under commercially viable conditions and the impressive advantages of this catalyst (including its tunable regioselectivity) emphatically established that monofluorophos ligands have great potential as ligands for catalysis.It was shown that L1 has structural features that make it resistant to both hydrolysis and disproportionation.These features were borrowed from diphosphites such as L3 which are: the PO 2 heterocycle and the bulky hydrophobic t-butyl groups that protect the P-F group and kinetically stabilise the monofluorophosphite.
Early studies (in the 1970s and 1980s) demonstrated that monofluorophosphines L21 and L22 were stable to disproportionation and this was rationalised in terms of the great steric bulk and strong electron-withdrawing properties of the substituents.It was later shown that constraining the C-P-C angle in bicyclic or tricyclic monofluorophos ligands such as L22 also led to greater stability with respect to disproportionation.Ligands such as L22 have been shown to be effective not only in hydroformylation but also in hydrocyanation under commercially viable conditions.
In view of the observed powerful stabilising effects of P-substituents on monofluorophos ligands, and the demonstrated capacity of monofluorophos ligands to support homogeneous catalysis, it is surprising to us that, to date, the area of monofluorophos chemistry remains so underdeveloped and it is our contention that there are a plethora of opportunities in the areas of ligand design, fundamental coordination chemistry studies, and catalyst discovery based on ligands containing a P-F bond.
It is clear from this review that, firstly, a P-F group confers unusual donor properties on the P III ligand, but there are striking 'holes' in our knowledge due to the paucity of information on monofluorophos coordination chemistry of many d-block metals; for instance, to the best of our knowledge, there are no examples of monofluorophos complexes of Re or Au.Secondly, the few catalytic studies on monofluorophos-metal complexes that have been reported have led to impressive discoveries.Some suggestions for potentially fruitful lines of enquiry that build on the results presented in this review are outlined below.
The monofluorophosphites, denoted {O,O}PF, and monofluorophosphines, denoted {C,C}PF, that are the subject of this review represent only a minor portion of the monofluorophos landscape that is available (Figure 2).There are many related {N,N}PF as well as mixed {C,O}PF, {C,N}PF, and {N,O}PF ligands waiting to be developed.Indeed, a series of acyclic and cyclic {N,O}PF ligands, (see Figure 7) of general structure L31 (R = alkyl) [95] and L32 (R = aryl or alkyl) [96], have been reported.Ligand L32 generates Rh catalysts for alkene hydroformylation with l:b ratios ranging from 0.41 to 12.8 depending on ligand concentration and the nature of R [96].

below.
The monofluorophosphites, denoted {O,O}PF, and monofluorophosphines, denoted {C,C}PF, that are the subject of this review represent only a minor portion of the monofluorophos landscape that is available (Figure 2).There are many related {N,N}PF as well as mixed {C,O}PF, {C,N}PF, and {N,O}PF ligands waiting to be developed.Indeed, a series of acyclic and cyclic {N,O}PF ligands, (see Figure 7) of general structure L31 (R = alkyl) [95] and L32 (R = aryl or alkyl) [96], have been reported.Ligand L32 generates Rh catalysts for alkene hydroformylation with l:b ratios ranging from 0.41 to 12.8 depending on ligand concentration and the nature of R [96].Chelating bis(monofluorophos) ligands would be an exciting avenue to explore and an example of a bis{N,N}PF ligand was recently described: the "Pacman" fluorophos ligand L33 (see Figure 7) [97].
Hydroformylation and hydrocyanation catalysis have been successfully demonstrated with monofluorophos ligands.These observations are consistent with the monofluorophos ligands behaving like other P-donors that are relatively electron-poor, such as phosphites.Monofluorophos-metal catalysts should be capable of catalysing other reactions that are catalysed by metal-phosphites and related ligands such as alkene isomerisation, hydrogenation, and C-C coupling reactions.
It was discovered that the optically active monofluorophosphite L19 was an effective ligand for the enantioselective Pd-catalysed intramolecular C-C coupling reaction.It would certainly be of interest to develop other optically active monofluorophos ligands (including bidentates) and investigate their efficacy in asymmetric catalysis.All of the {X,Y}PF heterocycles shown in Figure 2 have a stereogenic P-centre, and it should be possible to resolve these molecules and investigate the application of their complexes in asymmetric catalysis.
The overarching conclusion is that there is great scope to design new fluorophos ligands containing a PF group and expand the range of steric and electronic effects such ligands can have.There are good reasons to believe that new catalysts will emerge.

Figure 2 .
Figure 2. A selection of P-F containing monophos ligands (R = alkyl or aryl group) including Pheterocycles showing the diversity of ligands that are potentially available.The structures in the red box are the subject of this review.

Figure 2 .
Figure 2. A selection of P-F containing monophos ligands (R = alkyl or aryl group) including Pheterocycles showing the diversity of ligands that are potentially available.The structures in the red box are the subject of this review.

Figure 2 .
Figure 2. A selection of P-F containing monophos ligands (R = alkyl or aryl group) including Pheterocycles showing the diversity of ligands that are potentially available.The structures in the red box are the subject of this review.

Scheme 1 .
Scheme 1.Typical examples of the synthesis of acyclic and cyclic monofluorophosphites.

Scheme 1 .
Scheme 1.Typical examples of the synthesis of acyclic and cyclic monofluorophosphites.

Scheme 1 .
Scheme 1.Typical examples of the synthesis of acyclic and cyclic monofluorophosphites.

Scheme 2 .
Scheme 2. Routes to monofluorophosphite complexes: (a) conventional substitution at the metal of neutral ligand A by monofluorophos ligand; (b) substitution at the phosphorus of a coordinated Pligand.

Scheme 2 .
Scheme 2. Routes to monofluorophosphite complexes: (a) conventional substitution at the metal of neutral ligand A by monofluorophos ligand; (b) substitution at the phosphorus of a coordinated Pligand.

Scheme 2 .
Scheme 2. Routes to monofluorophosphite complexes: (a) conventional substitution at the metal of neutral ligand A by monofluorophos ligand; (b) substitution at the phosphorus of a coordinated P-ligand.

Figure 4 .
Figure 4. Some of the Eastman fluorophophites used as ligands in hydroformylation.

Scheme 21 .Scheme 21 .
Scheme 21.Monofluorophosphine-rhodium complexes.The fluoro analogue of Wilkinson's Catalyst, [RhF(PPh3)3], undergoes the rearrangement shown in Scheme 22 to generate complex 62, which contains a 'trapped' Ph2P-F ligated to Rh [82].This remarkable isomerisation occurs under mild conditions and is reversible.Several examples are known where late transition metal fluoro complexes with

62 Scheme 22 .
Scheme 22. Rearrangement leading to in situ formation of Ph 2 PF complex.

Figure 7 .
Figure 7. Monofluorophos ligands worthy of future study for catalysis.