Milestones in microwave-assisted organophosphorus chemistry

Synthesis of organophosphorus compounds under MW conditions

The beneficial effect of MWs was first experienced by us in the Diels–Alder reaction of 1,2-dihydrophosphinine oxides with dimethyl acetylene dicarboxylate (DMAD) to afford 2-phosphabicyclo[2.2.2]octadiene oxides [8], and in the inverse Wittig-type reaction of 1-(2,4,6-trialkylphenyl) P-heterocycles and DMAD [9]. Regarding the former case, the MW accomplishment goes with a 32-fold rate enhancement (if no solvent is used in the MW-assisted variation), while the latter instance represents an example, where the formation of the corresponding products, the β-oxophosphoranes is rather reluctant on conventional heating. Another reaction that practically does not take place under common conditions is the direct esterification of phosphinic acids [10]. However, we found that on MW irradiation cyclic phosphinic acids, such as 1-hydroxy-3-phospholene oxides (1 and 2), 1-hydroxy-phospholane oxides (3 and 4) and 1,2,3,4,5,6-hexahydrophosphinine oxide 5 underwent reaction with alcohols above 200°C to furnish the corresponding phosphinates (6–10) (Scheme 1) [11], [12].

Scheme 1:

MW-assisted direct esterification of cyclic phosphinic acids.

The above procedure required rather harsh conditions (T/p). It was found that using 10% of [bmim][PF₆] as an additive, the esterifications were possible already at 180°C, the reaction times shortened, and the yields were higher [13]. Other esterifications satisfying green chemical requirements were also elaborated [14]. Alternative routes are the alkylating esterification [15] and the T3P^®–promoted esterification [16], [17] of phosphinic acids. It is worth noting that our theoretical study on the conformation of cyclic phosphinic acids 3–5 revealed unexpected and interesting features. At the same time a few of the racemic phosphinates (e.g. 6) were subjected to optical resolution to provide valuable chiral building blocks or P-ligands [18].

MW irradiation was also useful in the alcoholysis (transesterification) of P-esters [19]. In this way, dialkyl phosphites with different alkyl groups were prepared that served as useful intermediates, e.g. they were used as starting materials in Kabachnik–Fields condensations to afford α-aminophosphonates (see also later) [20]. The hydrolysis of phosphinates and phosphonates is still a challenging field that could be performed advantageously under MW conditions. Earlier, we found that if the solid–liquid phase alkylation of active methylene containing substrates is performed under the effect of MW, there was no need to use a phase transfer catalyst [21]. This was also true for starting materials with P=O-function(s) [22], [23]. Many catalysts were described to promote the Kabachnik–Fields condensation (phospha-Mannich reaction) of amines, aldehydes (ketones) and >P(O)H reagents (mostly dialkyl phosphites) to afford α-aminophosphonates or α-aminophosphine oxides [24], [25], [26], [27], [28]. It was observed by us that in the MW-assisted condensations there was no need for any catalyst [29]. Moreover, bis(Kabachnik–Fields) protocols were also elaborated, and novel bis(phosphinoylmethyl)amines synthesized [26], [27], [30], [31]. A possible intermediate of the Kabachnik–Fields reaction is α-hydroxyphosphonate, a few derivatives of which were also prepared by us under MW conditions from aromatic aldehydes and dialkyl phosphites in the presence of Na₂CO₃ as the catalyst [32]. However, it was found that α-hydroxyphosphonates may also be formed from aldehydes and dialkyl phosphites in the presence of catalysts even without any solvent at room temperature. We were successful in developing the “greenest” variation of the Pudovik reaction [33]. The deoxygenation of phosphine oxides by a variety of silanes is of great importance [34], [35]. On the one hand, P-ligands may be prepared in this way, on the other hand, the phosphine reagents may be regenerated from the phosphine oxide by-products. Although catalytic versions for the reductions were also described, we proved that, especially under MW irradiation, there was no need for any catalyst [36], [37]. The above examples elucidated cases, where it was possible to substitute catalysts by MW irradiation. It is also an option to simplify catalysts or catalytic systems under MW conditions. A good example for this is the Hirao reaction between aryl halides and >P(O)H species, most often dialkyl phosphites to give arylphosphonates [38], [39], as it was found that using Pd(OAc)₂ or NiCl₂ as the catalyst, there was no need to add a P-ligand [40], [41], [42]. In special cases there was no need for any catalyst [43]. Arylphosphonates could also be obtained by the MW-assisted catalytic Arbuzov reaction of aryl bromides and triethyl phosphite [44].

Theoretical background

In order to get insights into the scope and limitations of the MW-assisted reactions, high level quantum chemical calculations were carried out on the energetics of the direct esterification of 1-hydroxy-3-phospholene 1-oxides (1 and 2) (Scheme 2) [11], [45].

Scheme 2:

Kinetic and thermodinamic data for the esterification of 1-hydroxy-3-phospholene 1-oxides obtained by B3LYP/6-31G(d,p)//PCM (MeOH) calculations.

It can be seen that the enthalpy of activation values are above 130 kJ mol⁻¹, and the esterifications are practically thermoneutral. These kinds of reactions may be performed well under MW irradiation. However, the endothermic tioesterifications and amidations of phosphinic acids remained uncomplete even on MW irradiation [46], [47]. The role of MWs is to overcome the relatively high enthalpy of activation that is possible due to the beneficial effect of the statistically occurring local overheatings [11]. This may be efficient for thermoneutral transformations. Our approach is in accord with that of Kappe et al. [48] and others [49] assuming only thermal effects and excluding non-thermal (“magic”) effects. It is important to note that already the effect of small overheatings may be important [50].

Modeling the thermal effect of MWs, and calculating the rate enhancing effect

Figure 1 shows that the total volume (V₀) of the reaction mixture consists of a V_bulk characterized by a temperature of T_bulk, and infinite “micro segments” ( V O H i ) with actual temperatures of T OH i . In the simplified model, T_OH is a constant value belonging to a certain V_OH/V₀ proportion.

Fig. 1:

Definition of the V and T values in a MW irradiated mixture.

T_bulk=temperature in the bulk of the reaction mixture

T OH i =temperature in the individual overheated segments

V_bulk=volume of the bulk of the mixture

( V OH i ) =volume of the individual overheated segments

V 0 = V bulk + ∑ i=1 n V OH i

In the first, simplified approach, it was supposed that the MW irradiation results in a distinct overheating (T_OH) in a percentage segment (V_OH/V₀) of the mixture. Regarding the esterification of 1-hydroxy-3,4-dimethyl-3-phospholene oxide 2 with butanol, and assuming that the temperature of the bulk (T_bulk) is 200°C, while the enthalpy of activation is 138.8 kJ mol⁻¹, and considering T_OH along with V_OH as the running parameters, the Arrhenius equation (1) suggested the rate enhancements (k_rel) defined by eq. (2) and shown in Table 1 and in Fig. 2.

ΔH^#: activation enthalpy (kJ mol⁻¹)

T: reaction temperature (T_bulk or T_OH [ T OH i ]

R: universal gas constant

Fig. 2:

Rate enhancements (k_rel) calculated for the 2→7a transformation in function of T_OH–T_bulk (°C) and V_OH/V₀ (%) at T_bulk=200°C.

Rate enhancements up to 5.5-times were obtained in function of T_OH–T_bulk and the proportion of the overheated segment (V_OH/V₀). A rate enhancement of ca. 3.3 can be expected if an overheating of 30°C occurs in 30% of the reaction mixture, or if the overheating is 35°C in 20% of the volume. The above supposition is obviously an over-simplified, and hence a not acceptable approach, as the extent of the local overheatings is different throughout the reaction mixture.

To approach the reality better, an exponentially decreasing relationship was assumed between T_OH–T_bulk and the proportion of the overheated segment (V_OH/V₀) as suggested by eq. (3). Depending on the proportion of V_OH, six cases were differentiated. The overheated segments were chosen to be 5, 10, 20, 30, 40 and 50%. Within the six cases, overheatings of 5, 10, 20, 30, 40, 50 and 60°C were assumed. The curves for the six cases selected are shown in Fig. 3.

For the A, B and C constants see Fig. 3.

Fig. 3:

The volume percentage (V_OH/V₀) – overheating (T_OH–T_bulk) relationships assumed in the overheated segments of 5–50%.

The relative accelerations (k_rel) were provided by the sum of the weighted partial rate constants obtained for V_bulk and the individual segments ( V OH i ) as suggested by eq. 4.

The partial rate components, and the overall k_rel values obtained for the six models selected are summarized in Table 2. A rate enhancement of 3.37 is predicted by model V assuming an overheated segment of 40%.

The next task was to calculate the experimental rate enhancement. This is possible if the corresponding conversion (x) – time (t) data pairs (Table 3) obtained after a careful experimental work (see Experimental) are substituted in the logarithmic pseudo first order kinetic equation (eq. 5). Doing so, a k_rel of 3.32 was obtained suggesting that the MW-assisted esterification is 3.32 times faster. This is shown in Fig. 4. Consulting Table 2, one can see that model V describes the real situation, where V_OH/V₀ is 40%.

where k′: pseudo first order rate constant

x: conversion

t: time

Fig. 4:

Conversion – time diagram for the thermal and MW-assisted 2→7a esterification at 200°C.

Eventually, the pseudo first order rate constants were also calculated for the thermal and the MW-assisted versions. According to this k′^Δ=1.62·10⁻⁵ s⁻¹, while k′^MW=4.76·10⁻⁵ s⁻¹. This latter constant should be regarded a nominal value.

The esterification of 1-hydroxy-3-methyl-3-phospholene 1-oxide (1) with butanol was investigated by a somewhat different exponential model in our preliminary communication [51]. In that approach, the maximum proportion of the overheated segment (V_OH/V₀) was only 30%, and the maximum extent of the overheating (T_OH–T_bulk) was 50°C. In the lack of a reliable time – conversion data pair, the experimental rate enhancement (k_rel) was underestimated. All these reasons prompted us that the modeling study on the esterification of 1-hydroxy-3-phospholene oxide 1 with BuOH should be revisited. Assuming a T_bulk of 200°C and a ΔH^# of 135.0 kJ mol⁻¹, using the exponential model applied for the 2→7a transformation, and operating with (V_OH/V₀)=50% and (T_OH–T_bulk)=60°C as the maximum values, the rate enhancements predicted are listed in Table 4.

The reliable experimental data pair of t=2h−x=17% for the thermal reaction, and t=3 h−x=58% for the MW-assisted variation suggested a k_rel of 3.10 that is somewhat smaller than that for the previous reaction (3.32). It can be seen that again model V assuming a V_OH/V₀ of 40% describes the real situation. k^Δ was obtained as 2.59·10⁻⁵ s⁻¹, while the nominal k^MW=8.03·10⁻⁵ s⁻¹. The esterification of dimethyl-1-hydroxy-phospholene oxide 2 is indeed somewhat slower than that of the monomethyl derivative 1.

MW-assisted esterification of 1-hydroxy-3,4-dimethyl-3-phospholene 1-oxide (2) with butanol

A mixture of 0.11 g (0.76 mmol) of 1-hydroxy-3,4-dimethyl-3-phospholene oxide 2 and 1.0 mL (11.0 mmol) of butanol was measured in a sealed tube, and irradiated in a CEM Discover microwave reactor equipped with a stirrer and a pressure controller (applying 150–300 W irradiation) at 200°C for 3 h. The pressure developed was 8 bar. Then, the excess of the alcohol was removed under reduced pressure, and the residue so obtained filtered through a 10 mm silica gel layer using ethyl acetate as the eluent to afford a sample analyzed by ³¹P NMR spectroscopy. The proportion of the integrals of the ³¹P NMR signals of hydroxy-phospholene oxide 2 (δ_P (CDCl₃) 69.7) and 1-butoxy-3,4-dimethyl-3-phospholene 1-oxide (7a) (δ_P (CDCl₃) 68.4; δ_P [52] 68.4) suggested a conversion of 44%.

The thermal variation was performed similarly, heating the two components in an oil bath at a constant temperature of 200°C for 2 h. In this case, the ³¹P NMR analysis that followed the filtration via a thin silica gel layer suggested a conversion of 11%. The conversion values were obtained as the average of 3–3 independent experiments.

The time – conversion data pairs for the MW-assisted and thermal esterifications of 1-hydroxy-3-methyl-3-phospholene oxide 1 with BuOH were obtained in similar experiments. The average of 3–3 parallel experiments gave the results shown in Table 5.