J. Chem. Soc., Perkin Trans. 2, 1999, 2551–2556 2551 This journal is © The Royal Society of Chemistry 1999 Alcoholysis of dialkyl tetrazolylphosphonites Erkki J. Nurminen,*a Jorma K. Mattinen a and Harri Lönnberg b a Department of Organic Chemistry, Åbo Akademi University, FIN-20500 Åbo, Finland b Department of Chemistry, Turku University, FIN-20014 Turku, Finland Received (in Cambridge, UK) 2nd June 1999, Accepted 1st September 1999 Kinetics of the reaction of diisopropyl tetrazolylphosphonite with tert-butyl alcohol in dry THF have been studied in the presence of various acids, bases and salts that catalyze the process. Ammonium azolide salts were found to be considerably more e.cient catalysts than the corresponding azole acids or tertiary amine bases.For instance, the relative rates obtained with N,N-diisopropylethylammonium tetrazolide, N,N-diisopropylethylamine and tetrazole were 104, 28 and 1, respectively. The salts of strong protolytes turned out to be better catalysts than those of weak ones.The susceptibility of the reaction rate to the pKBH of the base is fairly strong (Brønsted ß = 0.41) compared to the sensitivity to the pKa of azoles (ß = 0.17). The mechanisms of catalysis are discussed. Introduction Modern automated synthesis of DNA utilizes the reactive phosphoramidites 1, the dialkylamino group of which may be easily displaced by the entering sugar hydroxy function in the presence of an acidic activator.1,2 Amine hydrohalides and azoles are known to catalyze the reaction and the most commonly applied catalyst is tetrazole 3 (TH), which is in principle capable of acting as both an acid and a nucleophile.Intermediary formation of tetrazolylphosphonites † (2) has been detected during tetrazole-promoted alcoholyses of 1 and they are usually considered to be intermediates of the reaction. One of the mechanistic key questions of this reaction is whether it proceeds solely via a tetrazolylphosphonite intermediate or is there possibly a competing route that bypasses involvement of tetrazole as a nucleophile in the reacion.1,4,5 This problem was addressed in our previous study,6 where some evidence for 2 lying on the pathway from 1 to 3 was presented (Scheme 1).Tetrazolylphosphonites can be prepared from 1 and tetrazole, but they have seldom been isolated in pure form, because they are unstable compounds, being especially susceptible to hydrolysis by atmospheric moisture.Compounds 2 are known to react with alcohols to give 3 and with secondary amines to give 1.5,6 Nevertheless, the data on the reactivity of tetrazolylphosphonites and factors in.uencing it are scarce. A thorough understanding of the behaviour of 2 is, however, a prerequisite for solid conclusions concerning the mechanism of phosphoramidite alcoholysis. For this purpose, we now report the reaction of diisopropyl tetrazolylphosphonite (2a) with tertbutyl alcohol (ButOH), a reaction known to be .rst-order in the concentrations of both reactants.Second-order catalysis by tetrazole as well as a marked sensitivity towards added salts were detected in the previous study.6 A more extensive study on catalysis of the reaction by various azole acids, amine bases and their salts has been carried out to learn how the e.ciency of a Scheme 1 † In a previous paper (E. J. Nurminen, J. K. Mattinen and H. Lönnberg, J. Chem. Soc., Perkin Trans. 2, 1998, 1621) these species were incorrectly named by the editor as dialkyl tetrazolylphosphites.given type of catalyst depends on its structure and protolytic strength. Results Structural properties of 2a The 31P NMR spectrum of 2a shows that in dry THF this compound exists as several isomers: one major component (d = 125.9 ppm, ~75%) and two or three (broadened overlapping signals) minor components (d = 128–129 ppm, ~10%). According to saturation transfer experiments these isomers are in equilibrium with each other and cannot therefore be isolated.This is in agreement with the observed parallel disappearance of the signals of all isomers during the alcoholysis. The observed isomers are best explained by the fact that a tetrazole ring can be attached to phosphorus either via its 1- or 2- nitrogen and in both cases with two di.erent rotational orientations (Scheme 2). Similar 1H–2H tautomerism has recently been observed for the tetrazole moiety of Irbesartan, a novel drug compound.7 The interconversion of both the rotational and constitutional isomers is slow on the NMR timescale because rotation around the P–N bond is hindered by pp–dp interaction 8 and bulky isopropyl groups, and cleavage of the bond is required for the constitutional isomerization.The equilibrium between the isomers of 2a can be rationalized by a mechanism similar to that of racemization observed for optically active trivalent phosphorus compounds bearing an appropriate leaving group.9–11 Racemization does not involve the energetically unfavourable inversion of the phosphorus center, but takes place via intermolecular exchange of labile groups at phosphorus.9,10 A cyclic transition state involving two Scheme 22552 J.Chem. Soc., Perkin Trans. 2, 1999, 2551�C2556 (or more) molecules has been proposed11 for the process that is known to be accelerated by traces of ammonium salts 9 and nucleophilic impurities.10 As a result of this, in the presence of salt the 1H NMR signals of the diastereotopic methyl protons of the isopropyl groups in 2a are broadened to two lumps instead of clear doublets.The interconversion processes of isomers are even faster in acetonitrile, where both isopropyl groups give only one broad methyl signal, and the 31P NMR signals of the isomers collapse to one averaged signal. In addition to the above mentioned isomers, there are some impurities visible in the 31P NMR spectrum including the hydrolysis product, diisopropyl hydrogenphosphonate (PriO)2P(O)H (¦Ä = 6.4 ppm), the concentration of which remained at the average level of 7% during the kinetic runs.The phosphorus impurities have not been found to aect the rate of the studied reaction. Pure isolated 2a that is practically free from protolytic impurities is surprisingly stable. One must bear in mind that both the hydrolysis and alcoholysis of tetrazolylphosphonite release tetrazole, known to be a moderate catalyst for these processes, and hence they are autocatalytic: once initiated, the reactions proceed with increasing rate until most of the starting material has been consumed.When protected from atmospheric moisture, puried 2a can be stored practically unchanged for several weeks and its uncatalyzed reaction with 3 equiv. of ButOH has a t1/2 longer than 3 h. However, even traces of a suitable salt added to the reaction medium will speed up the process to completion within a few minutes.The signicant catalytic eect that even traces of protolytes produce gives rise to some practical considerations: 1) to obtain reliable kinetic results 2a must be puried as well as possible from protolytic impurities present in the reaction mixture. 2) The purity of 2a is best controlled by measuring the rate of its alcoholysis in the absence of added catalysts. Hence, the kinetic results obtained with dierent synthesized batches of 2a are only comparable with each other if the rates of their uncatalyzed alcoholyses are equal. 3) The determination of the true rate constant of the uncatalyzed alcoholysis of 2a is a complicated and elaborate task, since it would require either complete absence of all impurities or exact determination of their remaining trace concentrations. In the scope of the present work the problem is not very relevant, and hence the eort required was not considered worthwhile. Accordingly, it remains unclear whether the non-zero value observed results only from the presence of impurities or if it does really reect the true ability of 2a to react with ButOH without any catalytic assistance.Kinetics of alcoholysis of 2a The reaction of 2a with tert-butyl alcohol in THF in the presence of azole acids, tertiary amine bases or ammonium azolide salts was followed. The kinetic runs were performed 20 C using 31P NMR spectroscopy for monitoring. Since the high reactivity of tetrazolylphosphonite prevents measurements under pseudo-rst-order conditions, the method of initial rates was applied: for each catalyst, the initial rate of the appearance of the phosphite product 3a at several catalyst concentrations was determined.In most cases, a good linear relationship (r > 0.99) between v0/{[2a][ButOH]} and catalyst concentration [cat] was obtained, as indicated by eqn. (1). lim t¡ú0 d[3a] dt = v0 = (ku kc[cat]0)[2a]0[ButOH]0 (1) Some of the most weakly catalytic salts that were used at high concentrations, however, showed deviation from the linear dependence, suggesting higher than rst-order dependence on the catalyst concentration (Fig. 1). This may refer to catalysis by ion triplets, quadrupoles etc. resulting from ion aggregation that typically takes place in solvents of low relative permittivity (¦Å(THF) = 7.4). This phenomenon is taken into account in eqn. (2) by inducing a [cat]n term (n > 1). In order to compare only the catalytic strengths of single ion-pairs and to avoid complications that arise in nonlinear cases, where several parameters are involved for each catalyst, we solved the parameters of eqn.(2) for these catalysts and used the rstorder coecients as a measure of catalyst eciency. These values equal the slopes of the dependencies at [cat] = 0, and the results are thus being extrapolated to innite dilution, where aggregation is minimal. v0 = (kn kc[cat]0 kc [cat]0 n)[2a]0[ButOH]0 (2) The relative acidity constants of the catalysts were determined in dry THF at 20 C by examining the protolytic equilibrium between two acids of comparable strength and extending the measurements to the whole set of catalysts in a stepwise manner.Accordingly, a strong base was added to an equimolar mixture of acids and the degree of deprotonation was determined on the basis of 13C NMR shifts as described in the Experimental section. The pKa value of tetrazole was assumed to be 4.9 and the pKBH value of dimethylbenzylamine 8.9.The pKa values of the other protolytes were calculated from these with the aid of the relative protolytic strengths obtained by the stepwise approach. The acidity constants and the catalytic constants are listed in Table 1. Acid catalysis. Various azoles, such as 5-(4-nitrophenyl)-1Htetrazole, 12 5-(ethylthio)-1H-tetrazole 13 and 3-nitro-1,2,4-triazole, have previously been used as activators in oligonucleotide synthesis. The usage of these (we applied 5-(methylthio)-1Htetrazole instead of the ethyl derivative) as catalysts in the studied reaction gave an interesting result: an azole introduced to the solution of 2a very rapidly replaced tetrazole at phosphorus, and a mixture of the new azolylphosphonite and the original 2a is obtained.The exchange reaction was completed in less than 20 seconds, which is the time required for registration of the rst NMR spectrum. The concentration ratio of the azolylphosphonites depended on the amount of azole added, and it remained constant during the alcoholysis, indicating that the two species were rapidly equilibrated.Azoles being more acidic or less acidic than tetrazole behaved similarly in this respect. On using tetrazole as catalyst a similar exchange reaction undoubtedly takes place, although it cannot be observed as the reactant and product in this case are the same compound. Fig. 1 The reaction of 2a with ButOH in the presence of various ammonium tetrazolides in THF at 20 C: the dependence of v0/ {[2a][ButOH]} on salt concentration.Notation: = DBU tetrazolide, = diisopropylethylammonium tetrazolide, = tributylammonium tetrazolide, = dimethylbenzylammonium tetrazolide, = methylmorpholinium tetrazolide, = triallylammonium tetrazolide, = diisopropylanilinium tetrazolide.J. Chem. Soc., Perkin Trans. 2, 1999, 2551–2556 2553 Because of the rapid transazolidation it was impossible to measure a rate constant referring to the catalytic e.ect of a single azole: the introduction of the catalyst yielded a mixture of two dialkyl azolylphosphonites and two azoles, and hence four di.erent catalytic processes may occur simultaneously.Still, it is possible to measure the alcoholysis rate of this mixture at various initial concentrations of the added azole, and calculate the catalytic constants for the mixtures. These constants, although not having a well-de.ned physical signi.- cance, re.ect both the nucleophilic and protolytic in.uence of the azoles, and they can therefore be regarded as the best available indicator of the overall catalytic e.ciency.Stronger acids seem to be better catalysts than the weaker ones. Base catalysis. Time-dependent product distributions of the base-catalyzed reactions of 2a with ButOH show a clear indication of autocatalysis: the reaction releases tetrazole, a relatively strong acid that protonates the amine used as base catalyst.The salt thus formed is a more e.cient catalyst than the original amine, as discussed below in more detail, and the gradual conversion of the base to salt during the reaction hence results in the observed rate enhancement. Consequently, detailed quantitative study of the base catalysis becomes dif- .cult, since the catalytic e.ect of the added base cannot be distinguished from that of the mixture of protolytes present in the solution already at early stages of the reaction.That is why the catalytic constants for only a few amines were determined: in each case a similar autocatalysis was observed. The strongest bases proved to be the best catalysts. Catalysis by salts. This is the most e.ective type of catalysis for tetrazolylphosphonite alcoholysis. The catalytic constants determined for di.erent ammonium azolide salts indicate that the salts derived from strong protolytes are better catalysts than those derived from weak ones.The pKBH of the cation is clearly the most important factor a.ecting the catalytic e.- ciency: Plotting the log kc values obtained with ammonium tetrazolides vs. pKBH gives a Brønsted ß value of 0.41 (± 0.05) with correlation coe.cient r = 0.92 (Fig. 2). The considerable variation of amine structure, wide pKBH range and di.erent degrees of aggregation are likely to be responsible for the observed deviations. The pKa of the azole has a less pronounced Table 1 The reaction of 2a with ButOH in presence of various catalysts in THF at 20 C: third-order rate constants kc (dm6 mol2 s1) and dissociation constants of protolytic additives Catalyst pKa a pKBH a kc 3-Nitro-1,2,4-triazole 1H-Tetrazole 5-Methylthio-1H-tetrazole 5-(4-Nitrophenyl)-1H-tetrazole Diisopropylethylamine Tributylamine 1-Methylmorpholine DIEA 3-nitro-1,2,4-triazolide DIEA 5-methylthiotetrazolide DIEA 5-(4-nitrophenyl)tetrazolide DIEA perchlorate b Tetrabutylammonium tetrazolide b DBU tetrazolide DIEA tetrazolide Tributylammonium tetrazolide N,N-Dimethylbenzylammonium tetrazolide 1-Methylmorpholinium tetrazolide Triallylammonium tetrazolide N,N-Diisopropylanilinium tetrazolide 5.2 4.9 4.2 3.7 ——— 5.2 4.2 3.7 — 4.9 4.9 4.9 4.9 4.9 4.9 4.9 4.9 ———— 10.0 9.7 8.5 10.0 10.0 10.0 10.0 — 13.2 10.0 9.7 8.9 8.5 7.8 6.8 0.004 0.009 0.011 0.21 0.25 0.05 0.01 0.96 1.2 1.7 0.32 3.1 5.7 0.94 0.40 0.078 0.054 0.033 0.024 a pKa and pKBH values in THF measured in current work. For details see Experimental section.b Measurements carried out using a di.erent batch of 2a than for other catalysts. e.ect on kc, the ß value being 0.17 (± 0.04) with r = 0.89 (Fig. 3). In spite of the limited accuracy of the calculated Brønsted ß values, the trend favouring salts of weakly protolytic ions is clear, the emphasis being on the acidity of the ammonium ion. The catalytic constants of tetrabutylammonium tetrazolide and DIEA perchlorate cannot be included in the above mentioned trends, since these salts represent the extreme cases, where the salt cannot act as an acid or base, correspondingly.The observed non-zero values suggest that the capability to act as proton donor or acceptor is not a prerequisite for the catalytic activity of salts in the reaction studied. This is in agreement with the catalysis of tetrabutylammonium bromide observed previously.6 To gain information on the possible co-operativity of bases and their conjugate acids as catalysts the reaction was followed in the presence of both diisopropylethylammonium tetrazole (DIEAT) and tetrazole (TH) or DIEAT and diisopropylethylamine (DIEA).The experiments are analogous to conventional bu.er catalysis measurements in aqueous solution. The kc values for the mixtures were obtained as slopes of the plot v0/{[2a][ButOH]} vs. [DIEAT] (Fig. 4). When an additional protolyte, either an acid or base was present, higher kc values resulted (Table 2).Additional base had a more marked rate-accelerating e.ect than additional acid. The observed catalytic constants of the mixtures were then plotted vs. [TH]0 and [DIEA]0 and catalytic rate constants of TH and DIEA were obtained as the slopes and that of DIEAT as the intercept. These values were used as the initial values for an iteration that optimized the .t of calculated constants with the observed data points yielding kc values 0.92, 0.34 and 1.02 for DIEAT, TH and DIEA, respectively.Fig. 4 shows how the result of the .tting (presented with solid lines) meets Fig. 2 The reaction of 2a with ButOH in the presence of various ammonium tetrazolides in THF at 20 C: Brønsted dependence of log kc on pKBH of ammonium ions. Fig. 3 The reaction of 2a with ButOH in the presence of various diisopropylethylammonium azolides in THF at 20 C: Brønsted dependence of log kc on pKa of azolide ions.2554 J. Chem.Soc., Perkin Trans. 2, 1999, 2551�C2556 the observed data. It is worth noting that the catalytic constants of TH and DIEA are considerably higher in the presence than in the absence of the salt. Discussion The relative permittivity of THF (¦Å = 7.4) is close to that of ethyl acetate (6.0), diethyl ether (4.2) and chloroform (4.7), rather than acetone (20.7), acetonitrile (36.7) or DMSO (46.7).14 Its protolytic contribution can be neglected and it has very limited capability to solvate ions: in such a medium reactions that involve proton transfer, a partially charged transition state, or charged primary products are likely to proceed reluctantly.At low relative permittivity charged particles are unfavoured species that require better stabilization than the solvent is able to oer. This makes the interactions between the polar solvate molecules more important as a means of lowering the energy of charged species and charge formation during reactions.15 These properties provide a reasonable background to the two characteristic features of tetrazolylphosphonite alcoholysis in THF: the surprisingly slow uncatalyzed reaction and high sensitivity of rate to protolytic additives.In solvents of low relative permittivity acids and bases may prefer dimeric forms 16 while the energetically most favoured alternative for ions is association as contact ion-pairs with cation and anion in the same solvent cage.15 Especially at higher concentrations, salts may also form aggregates of several ion-pairs.Considerable association of protolytes leads to altered kinetic orders,17 as seen in the cases of some ammonium salts in this work and tetrazole in our previous work. Catalysis may simultaneously take place by the rst-order and higherorder pathways and therefore distinguishing between these alternatives is neither possible nor meaningful. From the mechanistic point of view, the order of catalysis is actually rather insignicant as it describes more the aggregation degree Fig. 4 The reaction of 2a with ButOH in the presence of mixtures of DIEAT and DIEA or TH in THF at 20 C: the dependence of v0/{[2a][ButOH]} on [DIEAT]. Notation: DIEA:DIEAT = 2:1, DIEA:DIEAT = 1 : 1, DIEA:TH = 1 : 2, ¡Á = DIEA:TH = 1:1, DIEAT only. Dotted lines are calculated from data with linear regression (representing kc values of Table 2) and solid lines are based on the optimized catalytic constants of DIEAT, DIEA and TH.Table 2 The reaction of 2a with ButOH in presence of mixtures of DIEAT and DIEA or TH in THF at 20 C: third order rate constants kc (dm6 mol2 s1) with respect to DIEAT concentration [DIEA] : [DIEAT] : [TH] kc 0:1:2 0:1:1 0:1:0 1:1:0 2:1:0 1.66 1.22 0.92 1.99 3.06 of the catalyst rather than the interaction between the catalyst and the substrate. The characteristic property of tetrazolylphosphonite is the lability of the bond between the electron-withdrawing tetrazole group and trivalent phosphorus.The situation is similar to that proposed for N-protonated phosphoramidite:18 the P�CN bond is lengthened and polarized with partial negative charge on nitrogen and positive charge on phosphorus, the susceptibility of which to nucleophilic attack is thus enhanced. The dipolar starting material is easily associated with nucleophiles present in solution (Scheme 3). Without an added catalyst this preassociated complex is, however, practically unable to cross the energy barrier involved in cleavage of the P�CN bond and formation of the P�CO bond.An acid catalyst can assist the cleavage of the P�CN bond by protonation of the leaving group. An acidic azole, for example, can form a hydrogen bond to the departing tetrazolide. Base, in turn, can promote the formation of a covalent P�CO bond by deprotonating the attacking nucleophile. Both catalysts trigger the rate-limiting substitution step of the reaction that is likely to take place in a more or less concerted manner.The full charges developed on the immediate products are stabilized by the opposite charge still present within the same solvate cage. The remaining protolytic rearrangements take place rapidly after this (Scheme 4). The catalysis by salts is a more complicated matter. The nucleophilic contribution of the azolide can be ruled out, since a) no transazolidation is observed in the presence of salts; the azole can only be changed if the attacking azole is introduced in acidic form, and b) nucleophilic catalysis would not assist the reaction in the case of the tetrazolide anion since the leaving group already is a tetrazolide ion.Neither can the salt catalysis be solely of protolytic origin, because a) salts have larger catalytic constants than stronger acids and bases, b) the reactions in the presence of salt catalysts lack the autocatalytic shape although strongly acidic tetrazole is liberated in the course of the process and c) salts of the most weakly acidic cation and the most weakly basic anion are the best catalysts.Our results are best explained by ion-pair catalysis, a subtype of salt catalysis common in nonpolar solvents where salts exist as ion-pairs. The catalytic eect of the salt rises from its ability to polarize the medium 19 and provide stabilization to the Scheme 3 Scheme 4J. Chem. Soc., Perkin Trans. 2, 1999, 2551–2556 2555 partially charged and ionic species involved in the reaction.15 In solvents such as THF, whose ability to solvate charged species is highly limited, salt e.ects of this type may be enormous17 and salts are commonly more e.ective catalysts than acids or bases.20,21 It is worth noting that the studied salts derived from a tertiary amine and an azole do not .t the conventional concept of salt e.ects because they can, and obviously do, participate in the reaction as protolytes. This is not the origin of the catalytic e.ect, as discussed above, and salts may well catalyze the reaction without being proton donors (tetrabutylammonium tetrazolide) or acceptors (DIEA perchlorate).Nevertheless, observed catalytic coe.cients indicate that salts capable of acting both as acids and bases are superior catalysts due to their ampholytic nature. Ion-pair catalysis is often attributed to intimate interactions between the ion-pairs and substrate molecules, such as those depicted in Scheme 5.These associations are thought to promote the ionization of substrate molecule (A) 22 or stabilize the charges of the transition state (B and C).15,23 Alternatively, the e.ect of ion-pairs has been rationalised on a macromolecular level by an ion-pair atmosphere model that explains the rate enhancement by thed polarization of the medium rather than speci.c events taking place on molecular level.19,22 The e.ect is thought to be analogous to substitution of the solvent with a more polar one, such as acetonitrile, that is known to be a better solvent for the studied reaction.Anyway, the result is a more marked polarization of tetrazolylphosphonite, which leads to the closer association of the attacking alcohol and the phosphorus center before the rate-limiting step. Since the pre-equilibrium thus yields a structure closer to the transition state, the activation energy is lowered and the reaction is able to proceed without protolytic assistance. It is worth noting that the added salt also lowers the energy level of the immediate products and the salt ions are able to act as protolytes promoting the formation of .nal products after the rate-determining step.The dependence of reaction rate on the salt concentration is generally written as eqn. (3),19,20,22–24 consistent with eqn. (1) kobs = k0(1 b[salt]) (3) applied to our results. Winstein et al. have reported curvature that arises from association of the salt at high concentration and leads to catalysis by dimers or other aggregates.17 Scheme 5 Mathematical formulation of this phenomenon would require the insertion of a 1.5th or second order term in the equation as is the case in salts of weakly basic amines used in the present work.Brønsted dependencies qualitatively similar to the one observed for ammonium ions in the present work have previously been reported by others.20 The rationalization for this is that the catalytic e.ectiveness of an ion-pair depends on the charge density on its poles: in the case of a weakly acidic cation and a weakly basic anion the charge is at a maximum and the diminished interaction within the ion-pair enables larger interactions with external charged centers of the substrate.15 The reason for the less pronounced dependence on the azolide ion basicity might be that the di.erences in their protolytic strength arise from delocalization of negative charge in the aromatic heterocycle.In this case pKa is a poor measure of the charge density in the dipole, and the ß value remains small. The acid or base catalysis in the presence of salt has in several cases been found to result in a higher reaction rate than the one expected from additivity of the individual e.ects.21,25 The unexpectedly high acceleration has been attributed to cooperative catalysis involving the protolyte and ion-pair in same transition state. The phenomenon has not been thoroughly investigated, but in cases where data allowed formulation of a rate equation, a term containing the concentration of both catalysts was included (eqn.(4)).25 Our results are qualitatively v0 = (ku kc1[HA]0 2 kc2[HA]0[salt]0)[2a]0[ButOH]0 (4) consistent with this, but we did not detect co-operative secondorder catalysis as seen clearly from the linearity of dependencies in Fig. 4: the catalytic constant of the salt proved to be independent of the presence of acids or bases while the catalytic constants of the latter ones were higher in the presence of salt than in its absence, but were still not dependent on the salt concentration (eqn. (5)).v0 = (ku kc1[HA]0 kc2[salt]0)[2a]0[ButOH]0 (5) Ionization and dissociation of an organic molecule in a weakly dissociating medium occurs in a stepwise manner: heterolytic bond cleavage leads to a contact-ion-pair (CIP), which can dissociate partially to form a solvent-separated ion pair (SSIP) and completely to give two free ions (eqn (6)).A–B AB A||B A B (6) Reactions occurrng via free ions are subject to retardation of the reaction rate by a common ion mass e.ect .rst shown by Ingold et al. SSIPs, in turn, enable the so-called special salt e.ect proposed by Winstein et al.26 Neither of these e.ects is observed in the present study, as expected in a nonpolar solvent such as THF, which means that the charge formation during the reaction takes place within one solvate cage and only CIPs are involved before the rate-limiting step. This explains why azolide anions were not capable of replacing the tetrazole ring of 2a: transazolidation would require a SSIP with a full positive charge on phosphorus and since this is never formed, azolide anion is always more strongly bound to its counterpart in the ion-pair than any charged center outside of it.Hence, the mechanism of the alcoholysis can not be dissociative; fully charged ions can only develop after the attack of nucleophile on the phosphorus center.Unlike the azolide anions, azoles resulted in transazolidation so rapidly that it had already reached equilibrium while alcoholysis had barely started. The reactivity di.erence has two alternative explanations: a) the azole may have a signi.- cantly higher a.nity towards the trivalent phosphorus; competing successfully with the alcohol at the pre-association2556 J. Chem. Soc., Perkin Trans. 2, 1999, 2551–2556 stage, or b) the azole is much more acidic than the alcohol and hence the deprotonation involved in the substitution is faster for the azole. The latter explanation, however, is valid only if the deprotonation takes place during the ratedetermining step, which is unlikely in an acidic medium. Therefore, it appears reasonable to conclude that azole is a better, kinetically favoured, nucleophile for tetrazolylphosphonites than alcohol, which gives the thermodynamically more stable product.Experimental Synthesis of 2a, a description of the kinetic measurements and an analysis of the data have been published previously.6 The amount of hydrolytic decomposition was kept to a minimum by puri.cation of the phosphoramidite used as starting material, usage of only a very dry acetonitrile solution of tetrazole and careful .ltration of the diisopropylammonium tetrazolide salt. Attempts to purify 2a with distillation at reduced pressure (argon atmosphere) were not successful since at the pressure achieved by our vacuum pump (0.1 mbar) the boiling point of the product was too high and it was partially decomposed.Additional hydrolysis during kinetic measurements was kept to a minimum: sample solutions were dried with molecular sieves and stored in vials under airtight Te.on-coated butyl rubber septa, and NMR tubes were dried in an oven prior to use. NMR spectra were recorded at 500 MHz for 1H, 202 MHz for 31P and 125 MHz for 13C. The relative dissociation constants of acids were determined as follows: the 13C NMR shifts were measured for each acid and the corresponding conjugate base, which was obtained by adding strong base (DIEA) in the solution.The strength of two acids were then compared with each other by introducing a small amount of a strong base (DIEA) into a mixture of the two acids. The concentrations of deprotonated species of the acids were calculated from the 13C NMR chemical shift of the mixture using eqn.(7), and the ratio of the dissociation constants was then calculated by eqn. (8). The ratio [A] = dHA dHA A dHA dA [HA]0 (7) KA1 KA2 = [A1 ][HA2] [HA1][A2 ] (8) was determined at several concentrations of added base and the relative pKa value of the acid was calculated using the average of these. Each acid was compared in this manner to its closest neighbours in the aqueous pKa-value scale. Relative dissociation constants of bases were measured in a similar manner using strong acid (TFA) as the proton donor.References 1 É. E. Nifant’ev and M. K. Grachev, Russ. Chem. Rev. (Engl. Transl.), 1994, 63, 575; Usp. Khim., 1994, 63, 602. 2 S. L. Beaucage and R. P. Iyer, Tetrahedron, 1992, 48, 2223. 3 (a) G. I. Koldobskii and V. A. Ostrovskii, Russ. Chem. Rev. (Engl. Transl.), 1994, 63, 797; Usp. Khim., 1994, 63, 847; (b) S. J. Wittenberger, Org. Prep. Proced. Int., 1994, 26, 499. 4 (a) B. H. Dahl, J. Nielsen and O.Dahl, Nucleic Acids Res., 1987, 15, 1729; (b) O. Dahl, Phosphorus Sulfur, 1983, 18, 201. 5 (a) S. Berner, K. Mühlegger and H. Seliger, Nucleic Acids Res., 1989, 17, 853; (b) S. Berner, K. Mühlegger and H. Seliger, Nucleosides Nucleotides, 1988, 7, 763. 6 E. J. Nurminen, J. K. Mattinen and H. Lönnberg, J. Chem. Soc., Perkin Trans., 2, 1998, 1621. 7 M. Bauer, R. K. Harris, R. C. Rao, D. C. Apperley and C. A. Rodger, J. Chem. Soc., Perkin Trans., 2, 1998, 475. 8 D. Imbery and H.Friebolin, Z. Naturforsch., Teil B, 1968, 23, 760. 9 M. Mikolajczyk, Pure Appl. Chem., 1980, 52, 959. 10 J. Nielsen and O. Dahl, J. Chem. Soc., Perkin Trans. 2, 1984, 553. 11 L. Horner and M. Jordan, Phosphorus Sulphur, 1980, 8, 225; 8, 235. 12 (a) B. C. Froehler and M. D. Matteucci, Tetrahedron Lett., 1983, 24, 3171; (b) Y. Hayakawa and M. Kataoka, J. Am. Chem. Soc., 1997, 119, 11758. 13 P. Wright, D. Lloyd, W. Rapp and A. Andrus. Tetrahedron Lett., 1993, 34, 3373. 14 J. E. Gordon, The Organic Chemistry of Electrolyte Solutions, John Wiley & Sons, Inc., 1975, pp. 138 and 156 and references cited therein. 15 J. E. Gordon, The Organic Chemistry of Electrolyte Solutions, John Wiley & Sons, Inc., 1975, pp. 127–129. 16 J. E. Gordon, The Organic Chemistry of Electrolyte Solutions, John Wiley & Sons, Inc., 1975, pp. 234–245. 17 S. Winstein, E. C. Friedrich and S. Smith, J. Am. Chem. Soc., 1964, 86, 305. 18 (a) A. A. Korkin and E. N. Tsvetkov, Russ. J. Gen. Chem. (Engl. Transl.), 1987, 57, 1929; Zh. Obshch. Khim., 1987, 57, 2155; (b) A. A. Korkin, A. M. Mebel and E. N. Tsvetkov, Russ. J. Gen. Chem. (Engl. Transl.), 1988, 58, 900; Zh. Obshch. Khim., 1988, 58, 1015; (c) A. A. Korkin and E. N. Tsvetkov, Bull. Soc. Chim. Fr., 1988, 335; (d) A. A. Korkin and E. N. Tsvetkov, Russ. J. Inorg. Chem. (Engl. Transl.), 1989, 34, 161; Zh. Neorg. Khim., 1989, 34, 295. 19 J. E. Gordon, The Organic Chemistry of Electrolyte Solutions, John Wiley & Sons, Inc., 1975, pp. 137–140. 20 A. M. Eastham, E. L. Blackall and G. A. Latremouille, J. Am. Chem. Soc., 1955, 77, 2182. 21 E. L. Blackall and A. M. Eastham, J. Am. Chem. Soc., 1955, 77, 2148. 22 C. L. Perrin and J. Pressing, J. Am. Chem. Soc., 1971, 93, 5705. 23 A. Loupy and B. Tchoubar, Salt e.ects in organic and organometallic chemistry, VCH Verlagsgesellschaft mbH, Weinheim, 1991, p. 12. 24 A. H. Fainberg and S. Winstein, J. Am. Chem. Soc., 1956, 78, 2763. 25 Y. Pocker and R. F. Buchholtz, J. Am. Chem. Soc., 1971, 93, 2905. 26 A. Loupy and B. Tchoubar, Salt e.ects in organic and organometallic chemistry, VCH Verlagsgesellschaft mbH, Weinheim, 1991, pp. 18–21 and references cited therein. Paper 9/04406J