Nucleophilic Substitution Reactions of O-Methyl N,N-Diisopropylamino Phosphonochloridothioate with Anilines and Pyridines

Abstract

The kinetic studies on the reactions of O-methyl N,N-diisopropylamino phosphonochloridothioate with X-anilines and X-pyridines have been carried out in acetonitrile. The free energy relationship with X in the anilines exhibits biphasic concave upwards with a break region between X = (H and 4-F), giving unusual negative ${\beta}_X$ and positive ${\rho}_X$ values with weakly basic anilines. The unusual phenomenon is rationalized by isokinetic relationship. A stepwise mechanism with a rate-limiting leaving group departure from the intermediate is proposed based on the selectivity parameter and variation trend of the deuterium kinetic isotope effects with X. The free energy relationship with X in the pyridines exhibits biphasic concave upwards with a break point at X = 3-MeO. A concerted mechanism is proposed based on relatively small ${\beta}_X$ value, and frontside and backside nucleophilic attack are proposed with strongly and weakly basic pyridines, respectively.

Introduction

To extend the kinetic studies on the aminolyses, anilinolysis and pyridinolysis, of the phosphonochloridothioates, the nucleophilic substitution reactions of O-methyl diisopropylamino phosphonochloridothioate (7) with substituted anilines (and deuterated anilines) and pyridines are kinetically investigated in acetonitrile (MeCN) at 65.0 ± 0.1 and 45.0 ± 0.1 ℃, respectively (Schemes 1 and 2). The kinetic results of the present work are compared with those of the anilinolyses1 and pyridinolyses2 of Y-O-aryl methyl [1; Me(YC6H4O)P(=S)Cl],1b,2b O-methyl phenyl [2; Ph(MeO)P(=S)Cl],1e,2d O-ethyl phenyl [3; Ph(EtO) P(=S)Cl],1d,2d O-propyl phenyl [4; Ph(PrO)P(=S)Cl],1e,2e O-𝑖-propyl phenyl [5; Ph(𝑖-PrO)P(=S)Cl],1e,2e Y-O-aryl phenyl [6; Ph(YC6H4O)P(=S)Cl]1c,2a and Y-S-aryl phenyl [8; Ph(YC6H4S) P(=S)Cl]1a,2cphosphonochloridothioates. The aim of this work is to obtain further information on the reactivities, selectivity parameters, substituent effects, free energy correlations, deuterium kinetic isotope effects (DKIEs) and mechanisms of the thiophosphoryl transfer reactions.

Scheme 1.Anilinolysis of O-methyl N,N-diisopropylamino phosphonochloridothioate (7) in MeCN at 65.0 ℃.

Scheme 2.Pyridinolysis of 7 in MeCN at 45.0 ℃.

 

Results and Discussion

Anilinolysis.Table 1 lists the second-order rate constants (kH and kD) in MeCN at 65.0 ℃ together with the selectivity parameters (ρX and βX) and DKIEs (kH/kD). The Brönsted βX (Figure 1) and Hammett ρX (Figure S1) plots show biphasic concave upward free energy relationships with a break region between X = (H and 4-F). The rate is compatible with typical nucleophilic substitution reaction with partial positive charge development at the nucleophilic N atom in the transition state (TS) for X = (4-MeO, 4-Me, 3-Me, H) [electron-donating substituents (EDSs)]. On the other hand, the rate is not consistent with a typical nucleophilic substitution reaction for X = (4-F, 3-MeO, 4-Cl, 3-Cl) [electron-withdrawing substituents (EWSs)], giving unusual negative βX (and positive ρX) values. The magnitudes of the kH/kD values invariably decrease from the primary normal DKIEs (kH/kD = 1.23-1.08 ＞ 1) with X = (4-MeO, 4-Me) to the secondary inverse DKIEs (kH/kD = 0.99-0.86 ＜ 1) with X = (3-Me, H, 4-F, 3- MeO, 4-Cl, 3-Cl) as the aniline becomes less basic.

Table 1.aX = EDSs. bCorrelation coefficient, r = 0.999. cX = EWSs. dr = 0.999. er = 0.990. fr = 0.993. gr = 0.999. hr = 0.999. ir = 0.990. jr = 0.993. kStandard error {= 1/kD[(ΔkH)2 +(kH/kD)2 × (ΔkD)2]1/2}.

Figure 1.Brönsted plots of the reactions of 7 with XC6H4NH2(D2) in MeCN at 65.0 ℃.

Table 2 summarizes the second-order rate constants (kH) with unsubstituted aniline (C6H5NH2) at 55.0 ℃, natural bond order (NBO) charges at the reaction center P atom in the gas phase [B3LYP/6-311+G(d,p) level of theory],3 summations of the Taft steric constants of Ph and Ri [ΣES = ES(Ph) + ES(Ri): ES(Ri) = 0.00(Me); –0.07(Et); –0.36(Pr); –0.47(𝑖-Pr); –2.48(Ph)],4 Brönsted coefficients (βX(H)), crossinteraction constants (CICs, ρXY),5 DKIEs (kH/kD) and variation trends of the kH/kD values with X of the reactions of 1-8 with XC6H4NH2(D2) in MeCN. The variation trends of kH/kD values with X are represented with arrow. The vertical arrows (↑ or ↓) indicate the direction of the consistent increase of kH/kD value with X. For example, the symbol of ↑ indicates that kH/kD value increases as the aniline becomes more basic. The symbol of ↑↓ indicates that kH/kD value increases or decreases with X depending on Y.6

There is no linear correlation between the anilinolysis rates of 1-8 and positive NBO charges at the reaction center P atom, implying that the inductive effects of the two ligands do not play any role to decide the reactivity of anilinolyses of phosphonochloridothioates. The anilinolysis rates of six substrates (1-6) are inversely proportional (roughly) to the size of the two ligands, suggesting that the steric effects of the two ligands are the major factor to determine the anilinolysis rates.7 Taking into account considerably slower rate of 7 (with two ligands of 𝑖-Pr2N and MeO) compared to that of 2 (with two ligands of Ph and MeO), 𝑖.𝑒., kH(3)/kH(2) = 23.3, the steric hindrance of 𝑖-Pr2N group seems to be much greater than that of phenyl group in the TS.

When partial deprotonation of the aniline occurs in a ratelimiting step by hydrogen bonding (𝑒.𝑔., TSf in Scheme 3), the kH/kD values are greater than unity, primary normal (kH/kD ＞ 1.0).8 In contrast, the DKIEs can only be secondary inverse (kH/kD ＜ 1.0) when an increase in the steric congestion occurs in the bond-making process because the N―H(D) vibrational frequencies invariably increase upon going to the TS (𝑒.𝑔., TSb in Scheme 3).9 The real primary normal DKIE due to the hydrogen bond between the hydrogen of the N―H(D) moiety and Cl leaving group should be greater than the observed value because the observed DKIEs would be the sum of: (i) the primary normal DKIE, kH/kD ＞ 1, because of the partial deprotonation of one of the two N―H(D) bonds in the TSf for a frontside attack; (ii) the secondary inverse DKIE, kH/kD＜ 1, because of the steric hindrance that increases the out-of-plane bending vibrational frequencies of the other N―H(D) bond in TSf for a frontside attack; (iii) lowering the kH/kD value because of the nonlinear and unsymmetrical structure of N…H(D)…Cl in TSf; and finally (iv) lowering the kH/kD value because of heavy atom (N in the nucleophile and Cl in the leaving group) contribution to the reaction-coordinate motion. When the reaction proceeds simultaneously through both pathways, backside (TSb) and frontside (TSf) attack, the observed DKIEs are the sum of both effects, secondary inverse and primary normal, and the obtained value of kH/kD can be greater or lesser than unity depending on the fraction of the two pathways.

Table 2.aValue at 55.0 ℃. bValue with Y = H. cEmpirical value. dNote that the value of ΣES is not ‘ES(Ph) + ES(RiO)’ but ‘ES(Ph) + ES(Ri)’ for 2-6 because the data of Taft steric constants of RiO are not available. The ΣES values of 1 and 8 are estimated. eX = (4-MeO, 4-Me, H)/X = (4-Cl, 3-Cl, 3-NO2).fX = EDSs/EWSs.

Scheme 3.Backside attack in-line-type TSb and frontside attack involving hydrogen-bonded, four-center-type TSf (Lig1 and Lig2 represent the two ligands).

The biphasic concave downward free-energy correlations with a break region between X = H and 4-Cl were observed for the anilinolyses of 1:1b herein, (i) the kH/kD values explicitly show a break region between X = H and 4-Cl; (ii) as a results, the DKIEs are distinctly divided into two parts, primary normal (kH/kD = 1.03-1.30) with strongly basic anilines (X = 4-MeO, 4-Me, H) while extraordinary secondary inverse (kH/kD = 0.37-0.57 << 1) with weakly basic anilines (X = 4- Cl, 3-Cl, 3-NO2), rationalized by predominant frontside attack TSf with strongly basic anilines while predominant backside attack TSb with weakly basic anilines; and (iii) the proposed mechanism is a stepwise process with a ratelimiting bond formation for strongly basic anilines, while with a rate-limiting bond breaking for weakly basic anilines, based on the negative and positive ρXY values, respectively.10

In the present work, however, the magnitudes of the kH/kD values consistently decrease as the aniline becomes less basic, from kH/kD = 1.23 (primary normal) with X = 4-MeO to 0.86 (secondary inverse) with X = 3-Cl. The kkH/kD values with X do not show any break region (or point) in spite of unusual negative βX (and positive ρX) values with X = EWSs. This may indicate that: (i) the TS structure is gradually changed from dominant frontside attack involving hydrogenbonded, four-center-type TSf to dominant backside attack inline- type TSb as the aniline becomes less basic; and (ii) the anilinolysis mechanism with X = EDSs is the same as that with X = EDSs, although the attacking direction of the nucleophile is gradually changed with X.

Focus will now shift to the discrete biphasic free energy correlation with X and unusual negative βX value with X = EWSs, and discuss the following three plausible cases: (i) The discrete free energy relationships are reported because of a desolvation step prior to the rate-limiting nucleophilic attack with a smaller value of βX when the nucleophile is anion and the solvent is polar protic, 𝑒.𝑔., water.11 However, in the present work, the negative βX value with X = EWSs is not ascribed to a desolvation step because the aniline nucleophile is neutral and the MeCN solvent is dipolar aprotic; (ii) The negativeβX value can be observed because of TS imbalance phenomenon.12 The authors reported that the sign of βX changes from negative with relatively strong electrondonating Y substituents to positive with more electronwithdrawing Y for the reactions of 1-(Y-phenyl)ethyl chlorides with X-anilines in methanol.13,14 The anilinolyses of Ybenzhydryl chlorides in MeOH-MeCN also exhibited the change of the sign of βX depending on Y substituents.14,15 The negativeβX values were interpreted in terms of a TS structure in which nearly complete bond formation between the nucleophile and cation formed in an ion-pair preequilibrium is coupled with a TS imbalance phenomenon, advocated by Jencks and Bernasconi.12 However, the negative βX value of the present work is inadequate to a TS imbalance phenomenon because the existence of cation formed in an ion-pair pre-equilibrium of 7 can be hardly expected with X = EWSs. Moreover, when partial cleavage of the N―H(D) bond occurs with a TS imbalance phenomenon, the DKIEs should be primary normal,16 opposite to the results of the secondary inverse DKIFs with X = EWSs in the present work; and (iii) The unusual negativeβX (and positive ρX) value with X = EWSs generally implies that the nucleophilic N atom becomes more negative in the TS compared to in the ground state (GS). The authors reported the anilinolysis of Y-aryl phenyl isothiocyanophosphates in MeCN, and the free energy relationships with X exhibit biphasic concave upwards with a break region between X = H and 4-Cl, giving unusual negative βX value with weakly basic anilines (X = 4-Cl and 3-Cl).17 The negativeβX values with weakly basic anilines were substantiated by very tight TS, in which the extent of the bond formation is great and the degree of the bond breaking is considerably small. In other words, the TS structure is similar to that as if the isothiocyante is the nucleophile and aniline is the leaving group. The proposed very tight TS is possible because the isothiocyanate is a poor leaving group. In the present, however, ‘very tight TS’ model is inadequate because the chloride is better leaving group compared to the aniline.

In the present work, it is the suggestion of the authors that unusual negative βX (and positive ρX) value with X = EWSs is attributed to isokinetic relationship. For a large number of reaction series, it is found that δΔH≠ and δΔS≠ are proportional, resulting in isokinetic relationship.18 The secondorder rate constants at 55.0, 65.0 and 75.0 ℃, and enthalpies and entropies of activation with 3-methoxy-, 4-chloro- and 3-chloro-anilines are summarized in Table 3. The enthalpy and entropy of activation remarkably decrease as the aniline becomes less basic, and the ρX value decreases as the reaction temperature becomes higher as seen in Figure 2. As a result, the calculated isokinetic temperature is TISOKINETIC = 471.6 K = 198 ℃, according to Eq. (3) and Figure 3, where the ρX value is null.

δΔG≠ = δΔH≠ − TδΔS≠ = 0; when δΔH≠ = TISOKINETIC δΔS≠ (3)

A concerted mechanism was proposed for the anilinolyses of 2-5 based on the magnitudes of the Brönsted coefficient: βX(H) = 1.26(2), 1.23(3), 1.04(4) and 1.10(5) comparable with βX(H)(6) = 1.22-1.33 and βX(H)(8) = 1.21-1.25, because a concerted mechanism was proposed based on the negative ρXY values for the anilinolyses of both 6 and 8. In the present work, the authors propose a stepwise mechanism with a ratelimiting leaving group departure from the intermediate for both X = EDSs and EWSs based on: (i) the Brönsted coefficient of βX(H) = 1.43(7) with X = EDSs is comparable with βX(H)(1) = 1.31-1.54 with weakly basic anilines where a rate-limiting bond breaking is proposed; and (ii) as mentioned earlier, the gradual DKIEs variation with X indicates that the mechanism is not changed with both X = EDSs and EWSs. Table 4 lists the free energy relationships for substituent X variations in the anilines and proposed mechanisms for the anilinolyses of 1-8. The variation trends of the kH/kD values with X are not correlated with the anilinolysis mechanism as seen in Tables 2 and 4.19

Table 3.Second-Order Rate Constants (kH × 104/M.−1 s.−1) at 55.0, 65.0 and 75.0 ℃, and Activation Parameters for the Reactions of 7 with 4- MeO, 4-Cl and 3-Cl Anilines in MeCN

Figure 2.Hammett plots of log kH ν𝑠 σX for the reactions of 7 with 3-MeO, 4-Cl and 3-Cl anilines in MeCN. The ρX values are 4.21 ± 0.02 (r = 0.999), 4.02 ± 0.01 (r = 0.999) and 3.46 ± 0.02 (r = 0.999) at 55.0, 65.0 and 75.0 ℃, respectively.

Figure 3.Isokinetic relationship for the reactions of 7 with 3-MeO, 4-Cl and 3-Cl anilines in MeCN, giving the slope of isokinetic temperature, slope = TISOKINETIC = 471.6 K = 198 ℃ (r = 0.999).

Table 4.aStrongly/weakly basic anilines. ‘Bond formation/breaking’ indicates a stepwise mechanism with a rate-limiting bond formation and with a ratelimiting leaving group expulsion from the intermediate, respectively.

Pyridinolysis. Table 5 lists the second-order rate constants (kPyr) of the pyridinolysis in MeCN at 45.0 ℃. The substituent effects of the nucleophiles upon the pyridinolysis rates correlate with those for a typical nucleophilic substitution reaction. Both the Brönsted (Figure 4) and Hammett (Figure S2) plots, however, are biphasic concave upwards with a break point at X = 3-MeO. The magnitudes of the ρX (= −3.76 ± 0.03; r = 0.999) and βX (= 0.83 ± 0.02; r = 0.999) values with strongly basic pyridines (X = 4-MeO, 4-Me, 3- Me, H, 3-Ph, 3-MeO) are much greater than those [ ρX = −0.89 ± 0.01 (r = 0.998) and βX = 0.14 ± 0.05 (r = 0.960)] with weakly basic pyridines (X = 3-MeO, 3-Cl, 3-Ac, 4-Ac, 3-CN, 4-CN).

Table 5.Second-Order Rate Constants (kPyr × 104/M−1 s−1) of the Reactions of 7 with XC5H4N in MeCN at 45.0 ℃

Table 6 lists the second-order rate constants (k2 × 103/M−1 s−1) with unsubstituted pyridine at 35.0 ℃, Brönsted coeffi-cients ( βX ), cross-interaction constants (CICs; ρXY)5 and free energy correlations for the substituent X and Y variations in the nucleophiles and substrates, respectively, of the pyridinolyses of 1-8 in MeCN. The substituent effects of X and Y on the pyridinolysis rates are more significant compared to those on the anilinolysis rates, so that the free energy correlations of the pyridinolyses are more complex compared to those of the anilinolyses. The two substrates of 1 and 7 exhibit biphasic free energy correlations, not with Y but with X, in the anilinolyses. On the contrary, only one substrate of 6 exhibits linear free energy correlation with X and Y, and furthermore, the two substrates of 1 and 8 exhibit biphasic free energy correlations with both X and Y in the pyridinolyses. The pyridinolysis rates do not show linear correlation with the positive NBO charge at the reaction center P atom, as observed in the anilinolysis rates. The pyridinolysis rates of 2-5 show linear correlation with the size of the two ligands.20

Figure 4.Brönsted plot of the reactions of 7 with XC5H4N in MeCN at 45.0 ℃.

In view of biphasic concave upward free energy correlations for substituent X variations of 2-5, 7 and 8, the greater βXvalues with strongly basic pyridines compared to those with weakly basic pyridines indicate a frontside attack TSf with strongly basic pyridines and a backside attack TSb with weakly basic pyridines (Scheme 4). It is well known that a weakly basic group has a greater apicophilicity so that apical approach is favored for such nucleophiles.21 The apical nucleophilic attack should lead to a looser P―N bond in the TBP-5C structure because the apical bonds are longer than the equatorial bonds, and hence a smaller magnitude of βX is obtained.

Scheme 4.Backside attack TSb and frontside attack TSf for the reactions of phosphonochloridothioates with X-pyridines.

Table 6.aValue with Y = H. bEmpirical value. cStrongly/weakly basic pyridines. da-block (strong nucleophiles and weak electrophiles)/b-block (weak nucleophiles and weak electrophiles)/c-block (strong nucleophiles and strong electrophiles)/d-block (weak nucleophiles and strong electrophiles). eBiphasic discrete two plots, neither concave upwards nor downwards. fBiphasic concave upwards. gBiphasic concave downwards. hBiphasic concave upwards with a minimum point.

Table 7.aDicrete, V, A, Vmin: see the footnotes in Table 6. ba-/b-/c-/d-block: see the footnotes in Table 6.

Table 8.Activation Parameters for the Reactions of 1-8 with Aniline and Pyridine in MeCN

In the present work, the authors propose a concerted mechanism based on relatively small βX values, βX = 0.83 and 0.14 with strongly and weakly basic pyridines, respectively, comparable with the βX (= 0.87-0.95) values for the pyridinolyses of 6 in which a concerted mechanism was proposed.2a A frontside and backside attack is proposed with strongly and weakly basic pyridines, respectively. The reaction mechanisms of 1-6 and 8 were proposed based on the βX and ρXYvalues.2a-e The proposed mechanisms of the phosphonochloridothiates are listed in Table 7.

Activation Parameters. The enthalpies and entropies of activation, for the anilinolysis and pyridinolysis of 7 are summarized in Table R1,22 and those of 1-8 are summarized in Table 8. The relatively low value of activation enthalpy and large negative value of activation entropy are typical for the aminolyses of P=S (and P=O) systems. It is worthy of note that the clarification of the reaction mechanism by means of the activation parameters is not easy for the aminolyses of P=S (and P=O) systems.

 

Experimental Section

Materials. O-Methyl N,N-diisopropylamino phosphonochloridothioate was prepared by one step synthetic route. N,N-Diisopropylamino methoxy chlorophosphine was stirred about 3.0 h with sulfur in THF at −5.0 ℃. The solvent was removed under reduced pressure and a liquid product was isolated through column chromatography (10% ethyl acetate + n-hexane). Analytical and spectroscopic data of the product gave the following results (supporting information): (𝑖-Pr2N)(MeO)P(=S)Cl: Liquid; 1H NMR (400 MHz, MeCN-d3) δ 1.24-1.32 (d, 12H), 3.79 (s, 3H), 3.83-3.97 (m, 2H); 13C NMR (100 MHz, MeCN-d3) δ 22.3, 49.7, 54.3; 31P NMR (162 MHz, MeCN-d3) δ 83.9 (1P, P=S); GC-MS (EI, m/z) 229 (M+).

Kinetic Procedure. The second-order rate constants and selectivity parameters were obtained as previously described: anilinolysis1 and pyridinolysis.2 Initial concentrations for both anilinolysis and pyridinolysis were as follows; [substrate] = 5 × 10−3 M and [nucleophile] = (0.10-0.30) M.

Product Analysis. In the anilinolysis, O-methyl phenyl phosphonochloridothioate was reacted with excess aniline for more than 15 half-lives at 65.0 ℃ in MeCN. The aniline hydrochloride salt was separated by filtration. Acetonitrile was removed under reduced pressure. The product was isolated through treatment with ether and dilute HCl by a work up process and, then dried over MgSO4. After filtration the product was isolated by evaporating the solvent under reduced pressure. In the pyridinolysis, O-methyl N,N-diisopropylamino phosphonochloridothioate was reacted with excess pyridine, for more than 15 half-lives at 45.0 ℃ in MeCN. Solvent was removed under reduced pressure. The product was isolated after treatment with ether and acetonitrile, and then dried under reduced pressure. The analytical and spectroscopic data of the product gave the following results (supporting information): (𝑖-Pr2N)(MeO)P(=S)NHC6H5: Brown liquid; 1H-NMR (400 MHz, MeCN-d3) δ 1.02-1.34 (d, 12H), 1.36-1.47 (m, 2H), 3.45 (s, 3H), 6.64-6.70 (d, 2H), 7.08-7.21 (t, 2H) 7.32- 7.41 (d, 1H), 9.19 (s, br., 1H); 13C-NMR (100 MHz, MeCN-d3) δ 19.3, 48.2, 49.2, 115.4, 130.1, 138.8, 142.0, 149.1, 158.8; 31P-NMR (162 MHz, MeCN-d3) δ −11.7 (1P, P=S); GC-MS (EI, m/z) 286 (M+). [(𝑖-Pr2N)(MeO)P(=S)NC5H5]+Cl–: Gummy solid crystal; 1H-NMR (400 MHz, D2O) δ 1.21-1.34 (d, 12H), 3.52-3.55 (m, 2H), 4.43 (s, 3H), 8.07-8.09 (d, 2H), 8.56-8.62 (t, 1H), 8.81-8.82 (d, 2H); 13C-NMR (100 MHz, D2O) δ 18.5, 41.3 47.4, 127.3, 128.1, 144.4, 145.2, 153.4; 31P-NMR (162 MHz, D2O) δ −13.7 (1P, s, P=O); LC-MS for C12H22N2OPSCl (m/z), 308 (M+).