Novel Fluorinated Phosphorus–Sulfur Heteroatom Compounds: Synthesis and Characterization of Ferrocenyl- and Aryl-Phosphonofluorodithioic Salts, Adducts, and Esters

A series of novel ferrocenyl- and aryl-phosphonofluorodithioic salts, adducts, and esters has been prepared. The reaction of 2,4-diferrocenyl-1,3,2,4-diathiadiphosphetane 2,4-disulfide {[FcP(μ-S)S]2, FcLR} with dry KF or tetrabutylammonium fluoride (TBAF) led to the corresponding potassium and tetrabutylammonium salts of ferrocenyldithiofluorophosphinic acids. Treating potassium ferrocenyldithiofluorophosphinic acid with an equimolar amount of tetraphenylphosphonium chloride readily yielded the corresponding organic adducts, and with mono- and di-halogenated alkanes generated a series of the corresponding esters of ferrocenylphosphonofluoridodithioates. Similarly, using 1,3-epithionaphtho[1,8-cd][1,2,6]oxadiphosphinine 1,3-disulfide or Belleau’s Reagent in place of FcLR resulted in the corresponding novel salts, adducts, and ester derivatives. All new compounds have been characterized by means of multi-NMR (1H, 13C, 31P, 19F) spectroscopy and accurate mass measurement in conjunction with single crystal X-ray crystallography of four structures.


Introduction
Organophosphorus-fluorine heteroatom compounds (OPFHACs) bearing a P-F bond are of interest due to their diverse chemical or biological activities, such as selective phosphorylating agents in synthesis, and efficient inhibition of several classes of enzyme [1][2][3][4][5][6][7][8][9][10][11][12]. This new class of thiophosphates bearing one or more anion FPS − units was first prepared from the reaction of alkali metal fluorides and P4S10 by Roesky's group [13,14]. Since then, the synthesis of thiophosphoryl halides S=PF3 and S=PFCl2, and their derivatives S=PF2NH2, S=PFClNH2, S=PF2N=PF2X (X = Br, NH2 or OH), S=PF2N=PCl3, and S=PF2N=PF2N=C=NSiMe3 was reported successively [15][16][17][18][19]. There are few examples of the synthesis of simple phosphonofluorodithioates ROP(S)(S − )F containing a fluorine atom attached directly to the phosphorus atom [15,20,21]. The nucleoside phosphonofluoridodithioate monoesters were also prepared via oxidation of nucleoside phosphonodithioate with I2 in pyridine in the presence of TMSCl, followed by addition of triethylamine trihydrofluoride (TAF) [22,23]. Similar analogues were obtained from a one-pot sequential reaction of 1,3,2-dithiaphospholane P(III) derivatives, which were converted readily into the corresponding P(V) compounds by addition of elemental sulfur and finally into phosphonofluoridodithioates by further treating with TBAF [24]. The importance of phosphoro-fluorine compounds in pure and applied chemistry invigorated our interest in synthesizing new phosphorodithioates bearing the P-F group. Recently, we have reported the synthesis of a series of phenylphosphonofluorodiselenoic salts, adducts, and esters [25]. Herein, we extend this procedure for the synthesis of potassium and tetrabutylammonium salts of ferrocenyl-and aryl-phosphonofluoridodithioates, and the related organic adducts and esters. To the best of our knowledge, this is the first reported synthesis and characterization of ferrocenyl-phosphonofluorodithioates [FcPS2F] − and their structural analogues, providing a valuable addition to the library of phosphodithioate compounds.

Results and Discussion
The preparation, spectroscopic characterization, and crystal structures of a ferrocene analogue of Lawesson Reagent, 2,4-diferrocenyl-1,3,2,4-diathiadiphosphetane 2,4-disulfide {[FcP(μ-S)S]2, FcLR} has been reported by our group [26,27]. FcLR reacted with two equivalents of fresh dry potassium fluoride in dry acetonitrile at 80 °C under N2 atmosphere for 1 h, giving rise to potassium ferrocenylphosphonofluoridodithioate 1 in 97% yield; or with two equivalents of tetrabutylammonium fluoride (TBAF) in tetrahydrofuran at room temperature for 1 h providing tetrabutylammonium ferrocenylphosphonofluoridodithioate 2 in 99% yield (Scheme 1). Both reactions were fast and very straightforward and must be performed in a moisture and oxygen-free atmosphere. Treatment of FcLR with dry KCl or KBr in acetonitrile or with HCl and HBr in the presence of triethylamine in dry acetonitrile or tetrahydrofuran at room temperature did not result in the similar chloride and bromide products to 1 and 2, indicating that the Cl − and Br − anions are much less reactive nucleophiles than the F − anion. Scheme 1. Salts 1 and 2 derived from FcLR and KI or TBAF.
Using an analogous process to Yilmaz and coworkers [28], salt 1 was obtained as a yellow solid and is insoluble in organic solvents but soluble in oxygen-free water and slowly decomposed when stored at room temperature. Organic salt 2 was prepared as golden sticky oil, is soluble in organic solvents, and shows good air stability at room temperature. Both compounds show the anticipated molecular ion peaks [M − K] − or [M -N(n-Bu)4] − and satisfactory accurate mass measurement. The 31 P-NMR spectra exhibits doublets at δp = 132.1 ppm in compound 1 and 127.3 ppm in compound 2, respectively, attributable to the presence of the P-F single bond, and the values are significantly bigger than that in their selenium counterpart PSe2 ions [25]. In the 19 F-NMR spectra, doublets are observed with 1 J(P,F) coupling constants of 1008 Hz for both 1 and 2 within the known literature values [14,25,28,29]. It should be noted that the 1 H-NMR spectrum for compound 1 is poor quality ( Figure S1); this may be explained by the inherent "shielding" effect of salt ions in the solution leading to sample conductivity [30] or, more likely, the presence of traces of oxidized paramagnetic ferrocenium species. However, the 1 H-NMR spectrum still clearly shows that only ferrocenyl ring protons are present.
We presumed that compounds 1 and 2 are, like phenylphosphonofluorodiselenoic salts, strong nucleophiles [25] and therefore are able to serve as useful precursors for the synthesis of a wide variety of functionalized heteroatom systems and ligands. Compounds 1 and 2 should have the same reactivity toward organic substituents; therefore we chose compound 1 as a target staring material to explore their reactivity. Treating 1 with an equal molar amount of tetraphenylphosphonium chloride in degassed water at room temperature led to the formation of tetraphenylphosphonium ferrocenylphosphonofluoridodithioate 3 in 91% yield, though it should be noted that we have not established quantitative exchange of the cations in this or subsequent reactions. Reacting 1 with half equimolar amount of p-xylylene dibromide in dry tetrahydrofuran in room temperature generated 1,4-phenylenebis(methylene) bis(ferrocenylphosphonofluoridodithioate) 4 in 87% yield. Similarly, compound 1 was allowed to react with an equivalent of mono-halogenated alkanes, giving a series of esters of ferrocenylphosphonofluoridodithioates 5-9 in 64%-86% yields, respectively (Scheme 2). It is worth noting that products 5 and 8, in which the phenyl groups bear the strong electron-withdrawing group NO2 and C≡N group, were obtained in rather lower yields (64% and 66%, respectively); thus, the results indicate that strong electron-withdrawing groups may be unfavorable.
Compounds 3-9 are air-and moisture-stable oils, pastes, or solids and are soluble in common organic solvents such as dichloromethane, chloroform, acetone, and tetrahydrofuran. All new compounds show the anticipated molecular ion peaks [M] + , and were confirmed by satisfactory accurate mass measurements. Not surprisingly, the 31 P-and 19 F-NMR spectra of 3 show similar patterns to 1 and 2 with identical 1 J(P,F) coupling constants apart from another singlet signal in the 31 P-NMR spectrum at δP = 23.2 ppm, assigned to the PPh4 cation ion. The 31 P-NMR spectra of 4-9 display signals ranging from δP = 116.0 to 118.0 ppm with 31 P-19 F coupling constants differing from J(P,F) = 1098-1101 Hz. In the 19 F-NMR spectra of compounds 4-9, two equal signals in the range δF = −42.6-−38.3 ppm being considerately bigger than that in their selenium counterpart esters (δF = −58.5-−55.8 ppm) [25] are observed with the matching 31 P-19 F coupling constants.
By using the same procedure, we have carried out the synthesis of potassium 3-fluoronaphtho [1,8- [1,2,6]oxadiphosphinine-1-thiolate 1,3-disulfide 11 in 98% yield as a pale yellow solid; or with two equivalents of tetrabutylammonium fluoride (TBAF) at room temperature in tetrahydrofuran for 1 h gave tetrabutylammonium 3-fluoronaphtho [1,8-cd] [1,2,6]oxadiphosphinine-1(3H)-thiolate 1,3-disulfide 12 in 85% yield as a brown sticky paste. The P-O-P bond could not be broken even when an excess of potassium fluoride or tetrabutylammonium fluoride was used, indicating that the P-O-P bridge is more robust than the P-S-P bridge in the reaction. Once again, presuming that both compounds 11 and 12 possessing the same ArP2OS3F − ion should have similar reactivity, we selected salt 11 as a target starting material to explore their reactivity toward organic substituents. Complete conversion of 11 to the corresponding organic tetraphenylphosphonium salt 13 was carried on straightforwardly. Similarly, the reactions of 1,3,4-triphenyl-1H-1,2,4-triazol-4-ium tetrafluoroborate and 1,3-dimesityl-4,5-dihydro-1H-imidazol-3-ium chloride with 1 under identical conditions afforded the corresponding organic salts 14 and 15 in excellent yields (Scheme 4). Alkylation product 18 was obtained in 46% yield when 11 or 12 was dissolved in medium dichloromethane and the solution was stirred at room temperature for 48 h. It is believed that the mechanism for this reaction involved the intermediate 16, a product of the alkylation of two molecules of 11 or 12 with one molecule of dichloromethane; afterward, the intermediate, 16, broke and cyclized to give newly formed six-membered P2S2CO heterocyclic compound 18 by loss of a molecule of 1,3-difluoronaphtho [1,8cd] [1,2,6]oxadiphosphinine 1,3-disulfide 17. Unfortunately, we are not able to grasp and identify compound 17 due possibly to its instability on the chromatography column or ready decomposition when exposed to air. Attempts to study the mechanism of the reaction by 31 P-NMR were unsuccessful because complex mixtures were observed. Salt 11 was obtained as a pale yellow solid and is poorly soluble in organic solvents but soluble in oxygen-free water. All of 11-15 are very stable under an inert atmosphere of nitrogen, but they appear to undergo slow decomposition when exposed to the air after days at room temperature. Organic salts 12-15 were synthesized as a sticky paste, foam, or solid, are soluble in organic solvents, and also show good air stability at room temperature. Heterocycle 18 was isolated as yellow foam and was very stable in air and moisture. Surprisingly, in contrast to its starting material 10, compound 18 is soluble in normal organic solvents such as dichloromethane, chloroform, acetone, and so on. All salts show the anticipated molecular cation and anion ion peaks and satisfactory accurate mass measurement. Meanwhile, compound 18 displays the anticipated molecular ion peak and satisfactory accurate mass measurement. The 31 P-NMR spectra of salts 11-15 exhibit a similar pattern of two unequal phosphorus signals ranging from δP = 106.4-107.1 ppm for the non-fluorinated phosphorus and 78.1-78.6 ppm for the fluorinated phosphorus atom with the corresponding 1 J(P,F) and 2 J(P,P) coupling constants. The 1 J(P,F) coupling constants differing from 1099 to 1106 Hz are considerately lower than that in the analogous tetrabutylammonium, 3-fluoronaphtho[1,8-cd][1,2,6]thiadiphosphinine-1-thiolate 1,3-disulfide salt (1144 Hz) [32], and in the selenium counterpart PhPSe2F ions (1141-1146 Hz) [26]. The most significant difference is that the much bigger 2 J(P,P) coupling constant of 54.0 Hz was found in 11-15, comparing to its reported analogy (12.6 Hz) [33], indicating the noticeably differing effects of the P-S-P and P-O-P bridges. The singlet due to the PPh4 cation in 13 is observed at δP = 23.8 ppm. In the 19 F-NMR spectra, doublets at δF = −42.6-− 28.4 ppm are observed along with the matching 1 J(P,F) coupling constants, falling within the literature values [14,25,28,29]. Detailed analysis revealed that the small 3 J(P,F) = 3.2 or 3.3 Hz was observed in compounds 12, 13, and 15. For 18, the 31 P-NMR spectrum shows a singlet at δP = 106.0 ppm with a 2 J(P,P) coupling constant of 54.0 ppm. The 1 H-NMR spectrum displays signals from the aryl and CH2 protons present within compound.
Crystals of 3, 13, 14, and 18 suitable for X-ray analysis were obtained by diffusion of hexane into the dichloromethane solution at room temperature. Crystal data and structure refinement are summarized in Table 1. All structures have a single molecule of the compound in the asymmetric unit, aside from compound 14, in which the asymmetric unit comprises two independent molecules. The X-ray structural analysis of 3, 13, and 14 as shown in Figures 1-3 reveals they crystallize as cation and anion ion-separated species. The X-ray structure of 3 shows that the ferrocenylphosphonofluoridodithioate anionic part contains a distorted tetrahedral phosphorus atom with P-S bond distances of 1.963(3) and 1.953 (3)     The X-ray structure of 13 reveals that the naphthalene part of the molecule and two phosphorus atoms lie very close to the mean plane fitted to these atoms [maximum deviation 0.092 Å for C(1)], with the oxygen atom lying 0.548 Å above the mean plane and its cationic part being identical to that in 3. The structures of the cation in compounds 3 and 13 are noteworthy. The orientations of four phenyl ring are distorted away from each other, as evidenced by the different dihedral angles between three facing anion ion phenyl rings and one away phenyl ring (78.64°, 71.03° and 57.38° in 3 and 68.41°, 75.63° and 39.73° in 13) in the cation Ph4P + ions. This distortion presumably arises as a result of steric interactions of the Ph4P + cation ion with the anionic fragments FPS2 − and C10H6P2S3OF − . The three-dimensional network in Figure 2b shows the weak intramolecular and intermolecular C-H···S interactions (yellow dashed line), intermolecular C-H···F interactions (blue dashed line), intermolecular C-H···O interactions (pink dashed line), and π-π stacking interactions responsible for the stabilization of the crossed-layers supramolecular assembly. The single crystal structure of compound 14 has the same anionic part as that in the structure of 13. However, in contrast to 13, the C10H6P2 part of the anion is planar with the oxygen atom being significantly distorted and lying 0.495 [0.490] Å above the mean plane. The C3P2O ring is buckled with the C3P2 and P2O planes, inclined with respect to each other by 46  Using the Yilmaz methodology [28], we also carried out a similar reaction with Belleau's Reagent. The reaction of Belleau's Reagent with two molar equivalents of dry potassium fluoride at 80 °C gave potassium (4-phenoxyphenyl)phosphonofluoridodithioate 19 in 97% yield as a pale yellow paste; or with two equivalents of tetrabutylammonium fluoride (TBAF) at room temperature in tetrahydrofuran for 1 h it led to tetrabutylammonium (4-phenoxyphenyl)phosphonofluoridodithioate 20 in 99% yield as a slightly greenish yellow oil (Scheme 5). Furthermore, reacting 19 with equivalent of tetraphenylphosphonium chloride [28] at room temperature in degassed water/acetone medium furnished tetraphenylphosphonium (4-phenoxyphenyl)phosphonofluoridodithioate 21 in 86% yield as a white foam. All three reactions took place immediately and must be performed in a moisture-and oxygen-free atmosphere. Similar to salts 1 and 11, compound 19 is poorly soluble in organic solvents but soluble in oxygen-free water and is air and moisture stable. Compounds 20 and 21 are insoluble in oxygen-free water but soluble in normal organic solvents. The 31 P-NMR spectra of salts 19-21 reveal similar pattern to salts 1-3. Doublets at δP = 130. 2

General
Unless otherwise stated, all reactions were carried out under on oxygen-free nitrogen atmosphere using pre-dried solvents and standard Schlenk techniques; subsequent chromatographic and work-up procedures were performed in air. NMR spectra were recorded on Bruker Avance-400 ( 1 H at 400 MHz, 13 C at 100.6 MHz, 31 [40] with SHINE optic using Mo Ka radiation (k = 0.71073 Å). The data were corrected for Lorentz, polarization, and absorption. The data was collected and processed using CrystalClear (Rigaku) [41]. The structures were solved by direct methods [42] and expanded using Fourier techniques [43]. Hydrogen atoms were refined using the riding model. All calculations were performed using the CrystalStructure [

Synthesis of 1,4-Phenylenebis(methylene) Bis(ferrocenylphosphonofluoridodithioate) (4)
A mixture of tetrabutylammonium ferrocenylphosphonofluoridodithioate (0.108 g, 2.0 mmol) in THF (30 mL) was added to p-xylene dibromide (0.264 g, 1.0 mmol). The mixture was stirred at room temperature for 24 h. After filtration to remove the unreacted solid and drying under reduced pressure, the residue was purified by column chromatography on silica gel using dichloromethane as eluent to give the title compound 4 as greenish yellow oil (0.608 g, 87%). Selected IR (KBr, cm

Synthesis of 4-Nitrobenzyl Ferrocenylphosphonofluoridodithioate (5)
A mixture of tetrabutylammonium ferrocenylphosphonofluoridodithioate (0.541 g, 1.0 mmol) in THF (30 mL) was added to 4-nitrobenzyl bromide (0.216 g, 1.0 mmol). The mixture was stirred at room temperature for 24 h. After filtration to remove the unreacted solid and drying under reduced pressure, the residue was purified by column chromatography on silica gel using dichloromethane as eluent to give the title compound 5 as pale yellow paste (0.286 g, 64%). Selected IR (KBr, cm

Synthesis of 4-Bromobenzyl Ferrocenylphosphonofluoridodithioate (6)
A mixture of tetrabutylammonium ferrocenylphosphonofluoridodithioate (0.567 g, 1.05 mmol) in THF (30 mL) was added to 4-bromobenzyl bromide (0.260 g, 1.05 mmol). The mixture was stirred at room temperature for 24 h. After filtered to remove unreacted solid and dried under reduced pressure, the residue was purified by column chromatography on silica gel using dichloromethane as eluent to give the title compound 6 as reddish yellow paste (0.400 g, 86%). Selected IR (KBr, cm   (7) A mixture of tetrabutylammonium ferrocenylphosphonofluoridodithioate (0.541 g, 1.0 mmol) in THF (30 mL) was added to iodomethane (0.142 g, 1.0 mmol). The mixture was stirred at room temperature for 24 h. After filtration to remove the unreacted solid and drying under reduced pressure, the residue was purified by column chromatography on silica gel using dichloromethane as eluent to give the title compound