A water-soluble phosphite derived from sulfonated calix[4]arene. The remarkable stability of its rhodium complexes and two phase hydroformylation studies

Abstract

The reaction of tetrasulfonated calix[4]arene (trioctylammonium salt) with P(NMe₂)₃ followed by treatment with gaseous NMe₂H gave a zwitterionic six-coordinate phosphorus(V) species 1 containing a P(H)(NHMe₂) group which can be stored for months without decomposition. In water, 1 loses NHMe₂ and rearranges to form the water soluble phosphite L_cc. Phosphite L_cc has a half-life in water of ca. 5 h decomposing to H₃PO₃ and free tetrasulfonated calix[4]arene. Compound 1 serves as a convenient precursor to L_cc and complexes of L_cc are formed by dissolving 1 in water in the presence of labile metal complexes. The products have been identified by comparison of their ³¹P NMR data with well established analogues of calix[4]arene derived phosphites. In this way the water soluble complexes [Rh(acac)(CO)(L_cc)₂] (2c) [Rh₂Cl₂(CO)₂(L_cc)₂] (3c), [Pt₂Cl₄(L_cc)₂] (4c) and [PtCl₂(L_cc)₂] have been tentatively identified. The rhodium complex 3c has a half-life in water of 4 months. Two-phase (water/toluene) hydroformylation of 2-methylpentenoate with 2c as a catalyst has been investigated and the results compared with the same reaction in toluene with lipophilic analogues of 2c. Under mild conditions, with 2ccatalyst, branched aldehydes are the only products of hydroformylation and one of the branched aldehydes is selectively hydrogenated to give the lactone derivative.

---

Introduction

Sulfonated arylphosphines have been extensively studied ever since the advent of triphenylphosphine trisulfonate (TPPTS) and its commercial application in two-phase hydroformylation catalysis.^1,2 Triaryl phosphites are excellent ancillary ligands for Rh-catalysed hydroformylation³ but the kinetic instability of P–OR bonds to hydrolysis would appear to preclude making water-soluble sulfonated phosphites with any useful lifetime. Indeed Fell et al.⁴ showed that a mixture of sulfonated triphenylphosphites (TPOPS) (as tri-isooctylammonium salts) completely decompose in aqueous acetone after 24 h and Karakhanov et al.⁵ also reported that a polyether phosphite rapidly decomposed in the presence of water.

Phosphite ligands derived from calix[4]arene have been extensively studied^6–8 and we reported⁷ that phosphites L_aa and L_bb are stable to hydrolysis under conditions in which P(OPh)₃ is rapidly hydrolysed. We now report the synthesis of phosphite L_cc derived from sulfonated calix[4]arene and show that its complexes are sufficiently stable in water to allow their application in two-phase catalysis.

Results and discussion

Ligand synthesis and stability

The hydrophilic phosphite L_cc was made according to the route shown in Scheme 1. Treatment of the rigorously dried trioctylammonium salt of the commercially available calix[4]arene tetrasulfonic acid with P(NMe₂)₃ in CH₂Cl₂ followed by bubbling gaseous NHMe₂ through the solution gave a white precipitate which was assigned to the six-coordinate phosphorus species 1 (³¹P NMR in D₂O, δ −124.5 ¹J(PH) 730 Hz) from the close similarity of its ³¹P NMR data to the analogous non-sulfonated species.⁶ Solid 1 could be stored for weeks without change but in water, 1 converted over a period of 1 hour to phosphite L_cc as shown by its characteristic ³¹P chemical shift of 113.5 ppm (cf. 112.9 ppm for L_aa). Since L_cc is unstable in water (see below), compound 1 is a convenient precursor for the generation of L_ccin situ.

Scheme 1

Phosphite L_cc decomposes in water over 24 h to give H₃PO₃ and free sulfonated calix[4]arene, as shown by ³¹P and ¹H NMR spectroscopy. The kinetics of the decomposition was investigated by following the decline in the mole fraction L_cc (estimated from the ³¹P NMR integrals) as a function of time. A plot of ln[L_cc] against time was linear showing that the decomposition is a pseudo first order process with a t_1/2 of 5 h. In view of our previous observations of the very high hydrolytic stability of L_aa and L_bb and Fell's results which showed that sulfonation of triphenyl phosphite leads to enhanced hydrolytic stability, we had anticipated that L_cc would be more stable in water than is observed. The explanation for the lability of L_cc may lie in its conformation: assuming the conformation of L_cc is analogous to L_aa and L_bb (i.e. the partial cone shown in Scheme 1), the SO₃⁻group would be held in close proximity to the P and may therefore provide anchimeric assistance to the hydrolysis. Consistent with this hypothesis is that the rate of hydrolysis was independent of pH, being the same in buffers of pH 4.0, 7.0, and 9.0.

Coordination chemistry of L_cc

The coordination chemistry of L_cc with Rh, Pt, and Pd has been investigated by generating aqueous solutions of L_ccin situ from 1 in the presence of a labile metal complex as described below. Attempts to isolate the product complexes from their aqueous solutions by removal of the water or by precipitation were unsuccessful, invariably giving intractable oily residues which contained partially decomposed complex and several by-products. The structural assignments of the complexes of L_cc are therefore tentative, being based on comparison of solution ³¹P NMR data with the data for the analogous well-characterised complexes of L_aa or L_bb.

We were interested in making a water-soluble analogue of the previously reported⁷hydroformylation catalysts 2a,b. Thus a solution of the precursor 1 in D₂O was added to an acetone solution of [Rh(CO)₂(acac)] and after 24 h the main rhodium–phosphite complex present (δ 93.2 ppm, J(RhP) 282 Hz) was assigned structure 2c on the basis of the similarity of the ³¹P NMR parameters to those for 2a (δ 109.8 ppm, J(RhP) 309 Hz) and 2b (δ 110.5 ppm, J(RhP) 311 Hz). Addition of precursor 1 to an aqueous solution of Na[RhCl₂(CO)₂] gave a rhodium complex whose ³¹P NMR parameters (δ 107 ppm, J(RhP) 329 Hz) were very similar to those for 3b (δ 103 ppm, J(RhP) 317 Hz) and therefore it was assigned the structure 3c. The stabilisation of the hydrophilic phosphite upon coordination to rhodium(I) is phenomenal: we have monitored the first order decomposition of 2c over six months in aqueous solution and thereby estimated its half-life in water to be 4 months at ambient temperatures.

Treatment of L_cc (generated from 1) in D₂O with an acetone solution of K[PtCl₃(η-C₂H₄)] gave two platinum-containing complexes according to the ³¹P NMR spectrum. One species was assigned the binuclear structure 4c on account of its ³¹P NMR parameters (δ 7.2 ppm, J(PtP) 7320 Hz) being similar to those for the crystallographically characterised⁷ analogue 4b (in CDCl₃, δ 2.2 ppm, J(PtP) 7348 Hz). The other product was assigned the mononuclear structure 5c on account of its ³¹P NMR parameters (δ 32.6 ppm, J(PtP) 6246 Hz) being similar to those for complex 5a (in (CD₃)₂SO, δ 37.9 ppm, J(PtP) 6629 Hz).

Like the rhodium(I) complexes 2c and 3c, the platinum species 4c and 5c were much more stable to hydrolysis than the free phosphite L_cc. However treatment of Na₂[PdCl₄] with L_cc in D₂O gave initially a bright yellow solution but within 30 seconds a black precipitate had formed. The ³¹P NMR spectrum of the colourless supernatant showed the presence of a major peak at −22 ppm which is tentatively assigned to the phosphate L_cc(O) on the basis of the similarity of the δ(P) with that for the previously reported⁶phosphate L_aa(O) of −22 ppm; this is consistent with oxidation of L_cc by Pd(II) and the formation of metallic palladium.

Hydroformylation catalysis

The high stability of aqueous solutions of rhodium(I) complexes of L_cc prompted us to investigate two-phase (aqueous/organic) hydroformylation catalysis with this ligand. The hydroformylation of 2-methylpentenoate (2-MP) was studied, first with 2a and 2b as catalysts in a single toluene phase and then with hydrophilic complex 2c as a catalyst in an aqueous/toluene mixture. The matrix of products that are formed is shown in Scheme 2. The aldehydes A2 and A3 are derived from hydroformylation of the internal alkene 2-MP and hydrogenation of these aldehydes followed by lactonisation with loss of MeOH would give the corresponding lactones Lt2 and Lt3. Isomerisation of 2-MP to 1-MP followed by hydroformylation would give aldehydes A1 and A2. This last reaction is of particular commercial interest⁹ because 2-MP is the product of methoxycarbonylation of butadiene and linear aldehyde A1 is a precursor to adipic acid.

Scheme 2 Products formed in the hydroformylation of 2-methylpentenoate (2-MP).

The results from the catalysis are collected in Table 1 and the details of the conditions are given in the Experimental section. Pre-prepared complexes 2a–c were used and for each run, the conversion of 2-MP was >75%. The results with 2a–c (entries 2–7) can be compared with the control run using [Rh(CO)₂(acac)] (entry 1). The main effect of the catalysts derived from the lipophilic phosphites L_aa and L_bb (entries 2–5) is to reduce the amount of A1 produced as a proportion of the total aldehyde yield from 33% with the control to ca. 5% (entries 4 and 5).

Table 1 Hydroformylation results^a

Run	Ligand	Conversion	Aldehyde	Linearity^b	A1	A2	A3	Lt2	Lt3
a Unless otherwise stated, products obtained after reaction for 3 h in toluene at 160 °C and 60 atm H2/CO with Rh catalyst of 6.8 mM, 1 ∶ 1 ratio of L ∶ Rh, and ca. 150 ∶ 1 ratio of 2-methylpentenoate ∶ Rh. Product amounts are in % as determined by GC using toluene or cumene as internal standard; the balance (1–4%) is 2-methylpentenoate. See Experimental under single phase catalysis for full details. b Mol of A1/total mol of aldehyde. c 10 ∶ 1 ratio of L ∶ Rh. d Two phase catalysis, at 160 °C for Run 6 and 60 °C for Run 7. See Experimental for details.
1	None	87	46	33	15	14	17	23	31
2	Laa	79	40	22	9	17	14	29	30
3	Lbb	78	31	16	5	16	10	28	40
4^c	Laa	87	23	5	1	18	4	34	42
5^c	Lbb	88	24	5	1	20	3	28	46
6^d	Lcc	83	13	0	0	12	1	38	45
7^d	Lcc	94	46	0	0	46	0	4	48

---

The results of the two-phase catalysis with 2c carried out at 160 °C (entry 6) are suspect because at the end of the run, substantial amounts of rhodium metal precipitate were observed in the autoclave. However at 60 °C (entry 7), the catalysis proceeded with no Rh metal deposited and the highest conversion (94%) of all the runs was obtained. Elemental analysis of the aqueous and organic layers at the end of the run showed that ca. 60% of the rhodium resided in the organic layer showing that considerable leaching had occurred. Over 90% of the product was accounted for by A2 and Lt3 in approximately equal amounts. This suggests that the branched aldehydes A2 and A3 are the exclusive products of hydroformylation and subsequently A3 is selectively hydrogenated to the alcohol which, after lactonisation, gives Lt3. The selective hydrogenation of A3 may imply a selective binding of the A3 to the Rh; inspection of the proposed intermediates (Scheme 3) might suggest that the smaller chelate formed by A3 would be favoured over the one formed by A2.

Scheme 3 Proposed chelates formed prior to hydrogenation by (a) A3 and (b) A2.

It is clear that more work is required to further characterise the catalysis but we have shown that water-soluble phosphite complexes are stable in water for long periods and therefore could be used for two-phase catalysis.

Experimental section

General procedures

Unless otherwise stated, all reactions were carried out under a dry nitrogen atmosphere using standard Schlenk line techniques. Commercial reagents were used as supplied unless otherwise stated. The phosphites L_aa and L_bb were made as previously described^6,7 and the starting materials [Rh₂Cl₂(CO)₄]¹⁰ and K[PtCl₃(C₂H₄)]¹¹ were prepared by literature methods. NMR spectra were measured on a Jeol EX 90 or Jeol GX 400. ³¹P{¹H}, ¹³C{¹H}, and ¹H NMR spectra were recorded at ambient temperature of the probe using a deuterated solvent to provide the field/frequency lock.

Synthesis and stability of water soluble phosphite precursor 1

Trioctylamine (9.15 cm³, 21 mmol) was added to a suspension of p-(tetrasulfonic acid)calix[4] arene (3.00 g, 3.52 mmol) in CH₂Cl₂ (30 cm³) and after 1 min a clear yellow solution was obtained. Toluene (50 cm³) was added and then any water was azeotropically removed by concentration of the solution to 10 cm³ by distillation of the solvent. The remaining volatiles were removed under reduced pressure to give a tan oil. The oil was dissolved in CH₂Cl₂ (30 cm³) and P(NMe₂)₃ (0.88 cm³, 0.79 g, 4.84 mmol) added. A white precipitate began to form after 30 s and the mixture was stirred for 25 min after which time gaseous NHMe₂ was bubbled through the mixture at a rate of 3 bubbles per second for 5 min. The white solid product was then filtered off and washed with CH₂Cl₂ (10 cm³). ³¹P NMR (D₂O, 36.4 MHz): δ −124.5, ¹J(PH) 730 Hz. Elemental analysis was consistent with the formulation 1·N(C₈H₁₇)₃ (calc. for C₆₂H₁₀₇N₆O₁₆PS₄ in parentheses): C 53.57 (55.09); H 8.01 (7.98); S 9.08 (9.49); P 2.29 (2.29); N 6.13 (6.22). A sample of compound 1 (65 mg) was dissolved in D₂O (0.6 cm³) and the ³¹P NMR spectrum of the solution was recorded every 10 min for the first 1 h and then hourly for the next 12 h and then finally after 24 h (see Results and Discussion).

Generation of aqueous solutions of rhodium complexes of L_cc

(1) [Rh(CO)₂(acac)] (17 mg, 0.070 mmol) and precursor 1 (65 mg, 0.070 mmol) were added together to a mixture of D₂O (0.5 cm³) and (CD₃)₂CO (0.5 cm³) in an NMR tube. A bright red precipitate forms immediately but then dissolves upon shaking the tube. After 24 h, the ³¹P NMR spectrum showed the presence of one main (75%) Rh-phosphite species in solution (δ 93.2 ppm, J(RhP) 282 Hz) assigned the structure 2c (see Results and Discussion). Attempts to isolate the product by evaporation of the water under reduced pressure led to a viscous brown oil that contained a mixture of several unidentified Rh-containing species and ligand decomposition products.

(2) [Rh₂Cl₂(CO)₄] (12 mg, 0.030 mmol) was dissolved in a solution of NaCl (40 mg, 0.68 mmol) in D₂O (1.0 cm³) in an NMR tube and then precursor 1 (60 mg, 0.065 mmol) was added. After 1 h the ³¹P NMR spectrum showed the presence of a single Rh-phosphite species in solution (δ 107.0 ppm, J(RhP) 329 Hz) assigned the structure 3c. The decomposition of 3c in D₂O to several P-containing species was monitored by ³¹P NMR spectroscopy over six months. Attempts to isolate the product by evaporation of the water under reduced pressure or by precipitation of the product by addition of [NBu₄]Cl led to viscous red-brown oils that contained a mixture of products including unidentified Rh-containing species and ligand decomposition products.

Generation of aqueous solutions of platinum complexes of L_cc

A solution of precursor 1 (30 mg, 0.032 mmol) in D₂O (0.5 cm³) was added to a solution of K[PtCl₃(C₂H₄)] (11 mg, 0.032 mmol) in (CD₃)₂CO (0.5 cm³) in an NMR tube. After 3 h the ³¹P NMR spectrum showed the presence of two Pt-phosphite species in solution in the ratio 10 ∶ 1, the major species with δ 7.2 ppm, J(PtP) 7320 Hz, assigned structure 4c and the minor species with δ 32.6 ppm, J(PtP) 6246 Hz assigned structure 5c (see Results and Discussion). Attempts to isolate the product by evaporation of the water under reduced pressure or by precipitation of the product by addition of [NBu₄]Cl led to viscous yellow-orange oils that contained a mixture of products including unidentified Pt-containing species and ligand decomposition products.

Hydroformylation of 2-methylpentenoate

Single phase catalysis. A solution of 2a or 2b (0.27 mmol [Runs 2 and 3] or 2.7 mmol [Runs 4 and 5]) in toluene (35 cm³) and 2-methylpentenoate (4.65 g, 40.8 mmol) were placed in a 100 cm³ steel autoclave. The autoclave was evacuated and refilled with 1 ∶ 1 mixture of H₂ and CO. The pressure was then increased to 30 atm with H₂/CO and heated to 160 °C. Then the pressure was increased to 60 atm and the reaction mixture stirred for 3 h at 800 rpm. The autoclave was then cooled to ambient temperature and the pressure slowly released and the product analysed by GC.

Two phase catalysis. A solution of 2 (0.27 mmol) in water (30 cm³), toluene (35 cm³), methanol (2 cm³) and 2-methylpentenoate (4.65 g, 40.8 mmol) were placed in a 100 cm³ steel autoclave. The catalysis in Run 6 was then carried out using the same procedure to the single phase experiment above. For Run 7, the reaction temperature was lowered to 60 °C.

We would like to thank Rhône-Poulenc for a studentship (for CJC) and Johnson-Matthey for a loan of rhodium compounds.

References

1. E. G. Kuntz, Fr. Patent, 2314910, 1975, (to Rhone-Poulenc).
2. Aqueous Phase Organometallic Catalysis: Concepts and Applications, ed. B. Cornils and W. A. Herrmann, Wiley, 2004.
3. Rhodium Catalyzed Hydroformylation, ed. P. W. N. M. van Leeuwen and C. Claver, Kluwer Academics Publishers, Dordrecht, 2000.
4. B. Fell, G. Papadogianakis, W. Konkol, J. Weber and H. Bahrmann, J. Prakt. Chem., 1993, 335, 75.
5. E. A. Karakhanov, Y. S. Karasheva, E. A. Runova and V. A. Semernina, J. Mol. Catal. A: Chem., 1999, 142, 339.
6. (a) D. V. Khasnis, J. M. Burton, M. Lattman and H. Zhang, J. Chem. Soc., Chem. Commun., 1991, 562; (b) D. V. Khasnis, J. M. Burton, J. D. McNeil, C. J. Santini, H. Zhang and M. Lattman, Inorg. Chem., 1994, 33, 2657.
7. (a) C. J. Cobley, D. D. Ellis, A. G. Orpen and P. G. Pringle, J. Chem. Soc., Dalton Trans., 2000, 1101; (b) C. J. Cobley, D. D. Ellis, A. G. Orpen and P. G. Pringle, J. Chem. Soc., Dalton Trans., 2000, 1109.
8. (a) F. J. Parlevliet, C. Keiner, J. Fraanje, K. Goubitz, M. Lutz, A. L. Spek, P. C. J. Kamer and P. W. N. M. van Leeuwen, J. Chem. Soc., Dalton Trans., 2000, 1113; (b) C. Dieleman, S. Steyer, C. Jeunesse and D. Matt, J. Chem. Soc., Dalton Trans., 2001, 2508; (c) S. Steyer, C. Jeunesse, J. Harrowfield and D. Matt, Dalton Trans., 2005, 1301; (d) S. Steyer, C. Jeunesse, J. Harrowfield, D. Matt, R. Welter and M. Wesolek, J. Chem. Soc., Dalton Trans., 2002, 4264; (e) C. Kunze, D. Selent, I. Neda, M. Freytag, P. G. Jones, R. Schmutzler, W. Baumann and A. Borner, Z. Anorg. Allg. Chem., 2002, 628, 779; (f) P. Maji, S. S. Krishnamurthy and M. Nethaji, Polyhedron, 2008, 27, 3519; (g) A. Sarkar, M. Nethaji and S. S. Krishnamurthy, J. Organomet. Chem., 2008, 693, 2097; (h) D. Semeril, D. Matt and L. Toupet, Chem.–Eur. J., 2008, 14, 7144; (i) L. Monnereau, D. Semeril, D. Matt and L. Toupet, Adv. Synth. Catal., 2009, 351, 1629; (j) L. Monnereau, D. Semeril and D. Matt, Eur. J. Org. Chem., 2010, 3068; (k) L. Monnereau, D. Semeril and D. Matt, Green Chem., 2010, 12, 1670; (l) S. Liu and C. A. Sandoval, J. Mol. Catal. A: Chem., 2010, 325, 65.
9. (a) E. Drent and W. W. Jager, Eur. Patent, EP 577205, 1992, (to Shell); (b) M. Beller, A. Krotz and W. Baumann, Adv. Synth. Catal., 2002, 344, 517.
10. R. Cramer, Inorg. Synth., 1974, 19, 218.
11. J. Chatt and M. L. Searle, Inorg. Chem., 1950, 5, 210.

---