摘要:
Pre-reactive intermediates in gas-phase chemical reactions: a contribution from rotational spectroscopy A. C. Legon Department of Chemistry, University of Exeter, Stocker Road, Exeter, UK EX4 4QD The detection and detailed characterisation of pre-reactive Introduction intermediates in gas mixtures through their microwave This review describes some recent research aimed at identifying rotational spectra are described. Spectra were observed by intermediates in binary gas mixtures, the components of which, using two special types of nozzle (a fast-mixing nozzle or a if allowed to, react with each other. The term ‘reaction’ is here glow-discharge nozzle) in a pulsed-nozzle, Fourier-transform interpreted in a broad sense to include chemical reactions of the instrument.The pre-reactive intermediates are considered for conventional type, the precipitation of an ionic solid when two one example of each of the following reaction types: a gases are mixed, the formation/precipitation of a molecular ring-opening reaction (l),the addition reaction of a halogen solid composed of complexes involving a dative bond, or a to a 3c bond (2), the abstraction reaction of H from a molecule prototype unimolecular reaction. The technique involved inby fluorine (3), a proton transfer reaction (4), a reaction in detecting and characterising the intermediates is pulsed-nozzle, which a compound having a dative bond is formed (9,and a Fourier-transform microwave spectroscopy which allowsprototype unimolecular reaction (6). ground-state rotational spectra of such species to be observed with both high resolution and high sensitivity.This method has bO+HCl -+ [ O-..HCl] -+ been reviewed elsewhere recently1 and will not be discussed in detail. The only experimental facet to be highlighted in the ClCH2CH20H (1) present article will be two special nozzles that allow the CH2--CH2+Cl2 + [CH2=CH2-.-C12] -+ transient intermediates to be formed and then rapidly isolated. CH2Cl-CH2Cl (2) The rotational spectroscopy is then conventional. Most of the NH3 + F2 -+ [H3N*-*F2] + article will be concerned with the conclusions of chemical NHZF + NHF2 + NF3 (3) interest that result from characterising the various intermediate species. Me3N+ HI + [Me3NH+-.I-] -+solid trimethyl ammonium iodide (4) Some prototype reactions in the gas phase and NH3 + BF3 -+ [H3N-.-BF3] + pre-reactive intermediates solid H3N.BF3 (5) Some reactions of fundamental interest in chemistry are shown Ar+HCl(v = 1) + [(lOOO) Ar-.*HCl] + in Table 1.In each case, two reactant molecules come together Ar + HC1 (v = 0) (6) and form an encounter complex. We shall use the term ‘pre- Table 1 Six gas-phase reactions of interest in chemistry ~-Reaction Reactants Pre-reactive Product Reaction type Intermediate Ring-opening reaction Addition of halogen to 7c bond 3 NH3 + F2 -Violent substitution 4 R3N + HX or -solid ionic Proton transfer ammonium halides before or afterb RsNH+..- X-? clustering? 5 NH3 + BF3 -solid H3N*BF3 Formation of prototype dative bond 6 Ar + HCI(v= 1) -[(lOOO)Ar...HCI] -Ar + HCI (v= 0) Prototype unimolecular step Chem.Commun., 1996 109 reactive intermediates' to describe such complexes because in each case the complex exists at a potential energy minimum but the system is reactive in one way or another, as described below for the particular examples. It is not possible to state with certainty that this potential energy minimum falls on the reaction coordinate in every example, although it does seem likely that the formation of the encounter complex is often a preliminary step. Reaction 1 in Table 1 is the rapid opening of the three- membered ring that follows mixing oxirane and hydrogen chloride in the gas phase. When the components are so mixed, is the hydrogen-bonded species la formed first, followed by proton transfer to give the protonated ring (a species invoked in proposed solution-phase mechanisms) or is there already incipient proton transfer, i.e.does the pre-reactive intermediate already have a substantial extent of ionic character of the type described by lb? Reaction 2 involves the addition of molecular chlorine to ethene. There has been much discussion about the nature of the pre-equilibrium complex.2 Is it a weak outer complex 2a, of the type first classified by Mulliken, or is there a significant extent of charge transfer to give the strong inner complex 2b?3 Reaction 3 is included in the discussion because of the legendary reactivity of fluorine with molecules containing hydrogen.When F2 and NH3 are mixed in the gas phase the result is a spectacular flame. Can an intermediate be detected in such a violently reactive mixture? What is the nature of this intermediate? Proton transfer is a process of considerable interest in chemistry. There has been intense speculation about the nature of the ammonium and methylammonium halides in the gas phase.4 When, say, ammonia and hydrogen chloride gases are mixed (reaction 4), does the formation of complexes H3N...HX 4a precede the precipitation of the ionic solid that constitutes the white smoke or is there already an appreciable extent of proton transfer in the complex, e.g. H3N+H..-X-4b? Can the extent of proton transfer in the gas phase be enhanced by progressive methylation of ammonia? Reaction 5 also involves ammonia.The complex H3N-BF3 was the first coordination compound to be discovered5 and has been taken as the prototype of the dative bond. However, although the molecular solid has been well characterised, attempts to identify the gas-phase molecule H3N...BF3 were fruitless, for when the gases NH3 and BF3 are mixed the solid is rapidly precipitated and reheating of the solid produces H2N=BF2 in the vapour,6 not H3N.-BF3. Clearly, a detailed experimental description of various properties of H3N.BF3 in the vapour phase (including the extent of electric charge redistribution) is of value in describing the dative bond. Reactions 1-5 all involve complexes which presumably must undergo collisions if further reaction is to occur.Reaction 6, on the other hand, is an example of one proceeding by the unimolecular mechanism. Vibrationally excited (v = 1) HCl collides with argon atoms to produce the weakly bound complex (1000) Ar...HCl, in which the excited vibration is that involving predominantly HC1 stretching. But one quantum of the HCl stretching vibration has an energy ca. 25 times that required to dissociate the complex into Ar and HCl (v = 0). The second step in reaction 6 (called vibrational pre-dissociation) is the true unimolecular step in which the molecules having sufficient energy decompose. What is the lifetime of this excited complex before vibrational pre-dissociation? In classical lan- guage, how long does it take before sufficient vibrational energy leaks from the HC1 stretching motion into the mode that leads to dissociation? Does the excited complex live long enough to be observed? To answer the questions raised in the discussion of reactions 1-6 above, a means must be found for producing the pre- reactive intermediates alluded to and then inhibiting further progress along the reaction coordinate for long enough to detect and characterise these transient species.110 Chem. Commun., 1996 How to produce, freeze and characterise pre-reactive intermediates It is well known that rotational spectroscopy is a powerful method for measuring the precise properties of molecules in isolation in the gas phase. Indeed, much of our detailed knowledge of the structure of small molecules has been gained by this means and has been subsumed into textbooks, often without explicit acknowledgement of its source.In more recent times, it has become possible to investigate the rotational spectra of weakly bound complexes through the techniques of molecular beam electric resonance spectroscopy and pulsed- nozzle, Fourier-transform microwave spectroscopy, in which the pioneering work is associated with the names of Klemperer7 and Flygare,s respectively. The central feature in each case is a supersonic beam or jet of gas formed when the pre-mixed components, heavily diluted in, e.g. argon, are expanded from a reservoir at a relatively high pressure through a nozzle (usually a small circular hole) into a vacuum.The gas emerging from the nozzle contains a high proportion of weakly bound complexes which rapidly achieve a state of collisionless expansion and which have very low rotational and vibrational temperatures. While in this collisionless state the complexes can interact with electromagnetic radiation of appropriate frequency (micro- waves for rotational transitions) and their spectra recorded. The results to be discussed here have been obtained by using the pulsed-nozzle, Fourier-transform technique' in which, as the name implies, individual short pulses of gas are expanded through the nozzle via a solenoid valve. We shall not be concerned with the details of rotational spectroscopy, about which there are excellent texts.9 It is sufficient to mention that several of the spectroscopic observables lead to important molecular properties.Those of relevance to the present discussions are summarized in Table 2, which will be referred to from time-to-time. Given that rotational spectra of weakly bound complexes can be observed in the manner outlined, the only remaining problem is that the supersonically expanded gas is pre-mixed. Clearly, to observe pre-reactive intermediates in rapidly reacting systems, a method must be found of mixing the components on a sufficiently short timescale immediately prior to expansion. What is the time between mixing and the collisionless phase of the expansion? Is it fast enough for the purposes outlined above? To answer these questions it is necessary to be familiar with the properties of the supersonic expansion,lO which are dominated by those of argon because it is usually the preponderant (98%) component of the gas mixture.The properties of the expansion are conveniently discussed in terms of the ratio X/d, where X is the distance travelled by the gas downstream from the nozzle exit and d is the nozzle diameter, as defined in Fig. 1. The ratio Xld is proportional to time because within only about one nozzle diameter the gas flow has reached its terminal speed ( = 5 X lo4 cm s-1 for Ar under conditions appropriate to this discussion). Gas dynamics calculations10 show that the translational temperature of the expanded gas achieves a value of ca. 1 K within X/d = 10, i.e.within a time of about 10 ps if the nozzle has typical diameter d =: 0.5 mm. Fig. 1 shows how the number of binary collisions 22 and the number of ternary collisions 2, remaining to each molecule vary with Xld.lo Complexes are formed in three-body collisions which, as Fig. 1 shows, are effectively finished at X/d ;= 2, that is after a microsecond or two. But the nascent complexes can still undergo two-body collisions (which can lead to either dissociation or reaction) until Xld --10, although after that (i.e.after ca. 10 ps) all such collisions cease. Hence, if the complexes survive until Xld = 10,no bimolecular reaction is possible thereafter. Moreover, the cooling of the internal energy modes, which accompanies the translational temperature drop, ensures that the complexes will usually have insufficient energy to allow a truly unimolecular decomposition to occur.In short, at ca. 10 ps after passing through the nozzle, no further reaction, clustering or precipita- Some conclusions of chemical interest tion is possible. Clearly, if the reactive gases can be kept By using the technique of pulsed-nozzle, Fourier-transform separate and away from surfaces until the moment they expand microwave spectroscopy ,l but modified to incorporate either a through the nozzle, reactions requiring times > 10 ps can be fast-mixing nozzle12 or a glow-discharge nozzle,14 it has been precluded. This has been achieved by using the device described possible to observe and analyse the rotational spectra of the pre- below.The so-called fast-mixing nozzle' 1.12 is illustrated in Fig. 2. It reactive intermediates in reactions 1-6. From the spectroscopic is usually bolted onto the bottom plate of a standard solenoid constants so determined, conclusions of chemical interest concerning these species have been derived (see Table 2).valve used in the pulsed-nozzle experiment and consists of a Although the description of reactions 1-6 in Table 1 cited pair of concentric, approximately coterminal tubes of circular particular examples of complexes B.-A, the conclusions of cross section. One of the reactive components (e.g.NH3 in the chemical interest have been generalised in the first four cases by NH3-F2 system) is flowed continuously through the central tube applying the same techniques to a range of carefully chosen (typically of 0.3 mm internal diameter) into the vacuum complexes in which either B or A or both have beenchamber of the spectrometer.The other component (typically as systematically varied. These more general conclusions will be a 1% mixture in argon) is pulsed from a stagnation pressure of referred to in the discussion that follows and references given, 3 bar down the outer tube. Because the tubes are nearly as appropriate. coterminal, the reactive gases meet only in the roughly cylindrical interface between the concentric flows as they Reaction 1simultaneously undergo the expansion. According to the preceding discussion, complexes formed at the interface rapidly The ground-state rotational spectra of several isotopomers of a achieve collisionless expansion, are then effectively frozen and complex of oxirane and hydrogen chloride could be detected can be spectroscopically investigated.The efficacy of the fast- when the components were mixed using the fast-mixing mixing technique is demonstrated by Fig. 3 which shows a rotational transition of H3N--F2 observed when the component gases are pure NH3 (central flow tube) and F2-Ar (pulsed through the outer tube).13 d = nozzle exit The fast-mixing nozzle has been used to investigate the pre- diameter reactive intermediates associated with reactions 1-5, as defined in Table 1. For the prototype unimolecular process (reaction 6), t a different type of device, called a pulsed glow-discharge nozzle,l4 was used to produce the excited species (10OO) 120 Rg.-.HCl, where Rg = Ar or Kr.It consists simply of a standard 18 -solenoid valve with a stainless steel ring held concentrically -16with its outlet orifice but insulated from it. The ring is maintained at a nominal dc potential of -1400V while the case -14 of the solenoid valve is at earth potential. When a pulse of gas 12mixture (e.g. 1% of HCl in Ar or Kr) is expanded from the solenoid valve in the usual way, it travels through the centre of -10 this ring electrode. At some stage in the expansion through the -8 ring into the vacuum chamber, the instantaneous pressure is -6appropriate to allow a dc glow-discharge to be sustained in the gas.A few percent of HC1 molecules are thereby excited to the 411 v = 1 state and these form complexes Ar-..HCl or Kr.-.HCl in -12 .'.53the (1000) state during further expansion. Given that these 1 I1 I0 I I I1 I 1 iJ0vibrationally excited complexes do not have too short a lifetime 0 2 4 6 8 10 12 14 16 18 20 before vibrational pre-dissociation, they can then be detected XI d through their rotational spectrum in the usual way. Fig. 4 shows Fig. 1 The number of binary (Z,) and ternary (Z,) collisions remaining per a region of the rotational spectrum of X4Kr-.H35Cl with the glow molecule plotted against X/d, the number of nozzle diameters d travelled discharge off (a) and on (h).'5 The appearance of rotational downstream X in an axisymmetric expansion of argon.For details, see ref. transitions due to (10°O) 84Kr...H35C1 is evident in (b). 10 (redrawn, with permission of Oxford University Press, from ref. 10). Table 2 Spectroscopic constants from rotational spectroscopy and the molecular properties to which they lead Spectroscopic constant Molecular property Comment Nature of the spectrum Symmetry Special pattern is different for linear, symmetric-top and asymmetric-top molecules. Observation of a particular pattern often allows molecular symmetry to be established. Rotational constants, A", Bo. Co Radial and angular geometry Bo = h/Xn2Ib, where Ib= Elm, (a;?+ c;?)is a principal moment of inertia. Ibdepends on the relative positions (i.e.principal axis coordinates a,, c,) of atoms.Centrifugal distortion constant, DJ or AJ Intermolecular stretching force constant For weakly bound complexes B . . .A, the components k, of a weakly bound complex B and A can in good approximation be assumed rigid. Then, if higher than quadratic force constants can be neglected, DJ 0~ ko-l. Nuclear quadrupole coupling constants Electric field gradient q& at a q;, depends on the detailed electric charge X,,W) = eQx q&/h nucleus X having a nonzero electric distribution within the molecule. If a molecule quadrupole moment Qx X2 is subsumed into a complex, q,& changes and X,,(X) can be used as a probe of the change. Chem. Commun., 1996 111 nozzle. 16 The rotational constants of the various isotopomers demonstrate that this complex has C, symmetry and a geometry of the type shown in projection in the symmetry plane (ac) of the parent species in Fig. 5.Although it is straightforward to obtain the position of the C1 atom relative to that of the oxirane ring, it is notoriously difficult to place the H atom in a hydrogen- bonded complex in this way. However, the availability of the complete C1-nuclear quadrupole coupling tensor, Xuu(Cl), Series 9 solenoid valve Teflon outer tube pulse of F2-Ar from solenoid valve 0.3 mm i.d. glass capillary Fig. 2 The fast-mixing nozzle used to observe the rotational spectra of complexes B-XY (see text for discussion). A dilute mixture of one of the components (either B or XY) in argon is pulsed from a Series 9 (General Valve Cop) solenoid valve down the outer of the two concentric tubes, as indicated.The other component is flowed, undiluted, down the inner glass capillary, the internal diameter of which at its end is cu. 0.3 mm. The example of the NH3-F2 system is illustrated. I I I I I 6222.45 50 55 60 Frequency / MHz Fig. 3 The F = 0 t 1 '4N-nuclear quadrupole component of the J = 1 c0 transition of H3'4N-eF2 at 6222.5235 MHz, as observed with the fast- mixing nozzle/FT microwave spectrometer combination 112 Chem. Commun., 1996 Xbb(C1) and Xuc(Cl), allows an unambiguous conclusion in this case. It can be shown16J7 that the off-diagonal element Xu, leads in good approximation to the angle auzbetween the HCl axis (z) and the a-inertial axis in the equilibrium form of the complex (see Fig.5),that is after allowance has been made for the zero- point angular oscillation of the HC1 subunit. Because the position of C1 is well determined from the rotational constants, auzgives the position of H, as long as it can be assumed that the H atom remains attached to C1 at approximately the same distance as in free HC1. That this assumption is justified follows from a detailed interpretation of the component X,,(Cl) of the C1-nuclear quadrupole coupling tensor. Once auzis known, it is a simple matter to obtain Xzz(Cl), the value of the coupling constant along the equilibrium direction z of the HC1 subunit, from the value X,,(Cl) along the inertial axis direction a. After allowance for the zero-point angular oscilla- tion of the HCl subunit (discussed elsewhere),16 it is found the X,,(Cl) is only a little changed from that, Xo(Cl), of the free HCl Ti II I1 IIIIII 0 100 200 300 400 500 0 100 200 300 400 500 Frequency offset / kHz Frequency offset / kHz Fig.4 (a)The F = 9/2 t9/2 C1-nuclear quadrupole component of the J = 5 t4transition of (WOO) *4Kr-.H3s C1 observed in the absence of a glow discharge. The transition is a doublet as a result of a Doppler effect. (b)The same region of the spectrum observed under the same conditions, except with the glow-discharge nozzle switched on. The new features arise from two AF = 1 C1-nuclear quadrupole components of the J = 5 t4transition of (10OO) 84Kr-.H3s C1. In both (a)and (b),frequencies are offset at a rate of 3.90625 kHz per point from 11922.0207 MHz.(Reproduced, with permission of the American Institute of Physics, from ref. 15). n n I--W * 0a 3.128 A0 0 molecule. What does this reveal about the nature of the HCI subunit when within the complex? If there were a significant extent of transfer of the proton from HCl to (CH2)20 in the detected complex of oxirane and hydrogen chloride, i.e. if the valence-bond description of the complex has a significant contribution from the structure (CH2)20H+.-CI-(lb in Table l), we would expect X,,(CI) in the complex to be substantially decreased in magnitude relative to Xo(C1). This can be readily understood when it is recalled (Table 2) that X,,(Cl) is proportional to qzzCl, the electric field gradient at C1 along the HC1 axis z.In a free C1- ion, the spherically symmetric electronic charge distribution ensures that 4;-= 0. Of course, in (CH2)20H+-.-CI- there would be a nonzero value of 6:-as a result of the presence nearby of the electric charge distribution of (CH&O+H. However, this is expected to be a small effect, as it is in the diatomic ion-pair species Na+-.Cl-, for which 6;-is less than 10%of 6;in the covalent HCI molecule. l8 Hence, the fact that & of the complex is close to that in free HC1 and very much larger than that expected of the ion pair (CH2)20H+...CI- establishes that the HC1 subunit is little perturbed when incorporated into the complex.Accordingly, the geometry is that shown in Fig. 5, with the HC1 bond length effectively unchanged from free HC1. In those circumstances, knowledge of the angle aUzaccurately places the H atom and we conclude (see Fig. 5) that the hydrogen bond is significantly nonlinear.16,17 In fact, the O.-H- C1 nuclei deviate from collinearity by 8 = 16.5". The nonlinearities 8 of several other 0.e.H-X and S..-H-X hydrogen bonds have been dete1mined'6.'7,19-~5 in this way for a series of complexes B-eHX, where B = furan, S02, 2,5-dihydrofuran, oxirane, thiirane or formaldehyde and X = F, C1, Br, CN or C-CH. It has been possible to rationalise the behaviour of the angle 8 when the Lewis base B and the acid HX are varied by introducing a refinement to a simple set of rules that had been proposed earlier.26 The part of the original rules appropriate to the present discussion stated that the observed angular geometry of B..-HX was that resulting if the HX axis were assumed to lie along the axis of a nonbonding electron pair carried by the acceptor atom of B.The pyramidal configuration at 0 in (CH2)20-.HCl and the C, symmetry of the complex can then be readily understood16 in terms of the usual (grossly exaggerated) pictorial representation of nonbonding pairs, as can the shallower pyramid in the complex of 2,5-dihydrofuran -..HCll9 and the steeper pyramid in thiirane...HC1.22 The refinementl7.19.23 to the rules requires that a linear hydrogen bond can be envisaged to be formed initially, as described above, but then, with the position of the H atom fixed, the C1 atom swings round in the plane of the C, molecule until the sum of the force necessary to bend the hydrogen bond by the angle 8, the ring-..Cl repulsive force and the force of the secondary interaction of the CH2 ring protons and the C1 atom is zero.Highly nonlinear primary hydrogen bonds experimen- tally observed in H2CO-.HCCH,24 (CH2)20.-HCCH25 and (CH2)2S--.HCN23 have been rationalised in terms of a weak primary interaction that becomes significantly nonlinear in order to benefit from the stabilising secondary interaction. Reaction 2 The aim of investigating this system is to determine the nature of the so-called pre-equilibrium complex formed by C12 with alkenes.A detailed interpretation of the spectroscopic constants obtained from analysis of the rotational spectrum of ethene ...C12, as observed with the fast-mixing nozzle/FT microwave spectrometer combination, characterises this prototype complex unambigu~usly.~~~~~The magnitudes of the rotational constants Ao, Bo and Co of the species C2H4..-35C12, C2H4--37C135C1 and C2H4.-35C137Cl demonstrate that the geometry is as shown in Fig. 6, i.e. with the C12 internuclear axis lying along the C2axis of ethene that is perpendicular to the plane containing the C2H4 nuclei. This establishes that the complex is formally of the bmao type classified by Mulliken.3 But the question remains: Is the complex of the weak, outer type (2a in Table 1) or the strong, inner type (2b in Table l)? This question can be answered by considering the C1-nuclear quadrupole coupling constants Xgg(Clx) associated with the inner (x = i) and outer (x = 0)C1 nuclei and the intermolecular stretching force constant k,, as determined from the centrifugal distortion constant AJ of the complex (see Table 2).For a complex in which the C12 internuclear axis (z)coincides with the inertial axis a and is also a C, axis (n 3 2) in the equilibrium geometry, it has been shown28 that the electric field gradients &> and &3 at the inner and outer C1 nuclei increase and decrease in magnitude, respectively, by equal and opposite amounts to a good degree of approximation. In view of the proportionality of X,,(Cl) and qzi described in Table 2, the quantity defined by eqn.(1) then provides a measure of the f = [XZACli) -X~u(Clo)l/"%a(Cli)+ XZu(C1o)l (1) mean fractional difference in the electric field gradient at the two nuclei, since in the above approximation $[X;,(Cli) + X&(Cl,)] =: X,(CI), the C1-nuclear quadrupole coupling constant of free C12. Moreover, the equilibrium values in eqn. (1) can be replaced by the observed zero-point values, since the effects of zero-point averaging would then cancel from the expression in good approximation. Hence, f is available from the experi- mental observables. The value obtained is only 0.022 for C2H4---C12, which means that the electric field gradients at Cli and C1, are affected to only a minor extent by complex formation.The quantity fcan be modelled by using the Townes-Dailey method2' of interpreting nuclear quadrupole coupling constants in terms of valence p, d, ...electrons. According to this model in its simplest form, the transfer of a fraction 6 of a valence 3p, electron from Cli and C1, on formation of the complex to give C2H4...ClF+Clg- would lead to coupling constants (1 + 6) Xo(C1) at Cli and (1 -6) Xo(Cl) at CI,, where Xo(C1) is the free C12 value. Hence, the transfer of a single 3p, electron (6 = 1) to give the inner complex 2b, C2H4C1+... C1-, would lead to XaU(C1i) =: 2 Xo(C1) and Xua(Clo) =: 0. In fact, the observed quantities Xuu(C1i) -Xuu(Clo)=: Xo(C1) demonstrate that the Cl;! molecule is essentially unperturbed by complex formation and therefore the complex is of the weak outer type 2a, C2H4...CI-Cl .By substitution of the expressions (1 k 6)Xo(C1) for Xa,(Cli) and Xua(Clo) into eqn. (l), it is immediately clear thatf = 6 = 0.022. Therefore, the electric perturbation of the C12 subunit in the complex is very small and can be described in terms of a polarization of C12 that is envisaged as a transfer of -0.02 e from Cli to C1,. This conclusion is reinforced by the small value k, = 5.9 N m-1 obtained for the restoring force per unit infinitesimal displacement along the C2 axis of the complex. \c,c -H'H 1 Fig. 6 The experimentally determined geometry of the ethene-chlorine complex. The CH2 planes are perpendicular to the plane of the paper. The complex is formally of the bnao type defined by Mulliken (see ref.3) but is of the weak, outer type with no significant extent of electric perturbation of the CI2 subunit (see text for discussion). Chem. Commun., 1996 113 This magnitude of the force constant k, is typical of that of a weakly bound complex. An ion pair, such as C2H4Cl+.-C1-, would have a much larger k,. For example, the force constant for Na+-.Cl- is 108.6 N m-1.30 Similar investigations to that referred to above for C2H4"' C12 have been carried out for the whole series B-.C12, where B = CO, HF, C2H2, C2H4, PH3, H2S, HCN or NH3. It was found31 that the strength of the interaction, as measured by k,, increases in the order CO < HF < HCSH --PH3 --H2C=CH2 < H2S < HCN < NH3.In fact,f = 6 plotted against k, leads to a reasonably straight line. Moreover, even in the case of the most strongly bound of these complexes, namely H3N-.C12,6 is only ca. 0.06, establishing that all members of the series are outer complexes of the Mulliken type.3 Two aspects of the geometry of the series B-..C12 defined above are of interest. First, it is found3' that the angular geometry of a given member of the B..C12 series is identical to that of the corresponding member of the B.-HCl series, even though the angular geometries change considerably from one B to the next and the binding is via H in the B...HCl series but C1 in the B.-C12. For the B..-HCl series, the equilibrium angular geometries can be understood in terms of simple rules26 (alluded to earlier) in which the molecular axis of HC1 lies along the axis of a nonbonding or n-bonding electron pair of the acceptor B, as conventionally envisaged.Evidently, these rules also apply to the B--C12, except that the internuclear axis of C12 now lies along the axis of the appropriate electron pair on B.31 The reason for the parallelism of the angular geometries can be understood, and an explanation of why the rules apply to both series can be obtained, if the angular geometries are controlled predominantly by the electrostatic contribution to the inter- action energy of the two subunits. In particular, a similarity in the electric charge distributions of HCl and C12 means that, in so far as the electrostatic part of the interaction is concerned, these molecules appear very similar to the approaching Lewis base B.32 The second point of interest in the geometries of B-.Cl2 and B.a.HC1 lies in the distance r(B.;Cl), which is found to be systematically shorter by 0.53(7) A in the B.--C12 series.31 This contraction has been identified in the purely repulsive part of the interaction of a helium atom with HCl and C12 through ab initio calculations and has been explained, in part at least, by an anisotropy of the chlorine atom in a chlorine molecule.33 A parallel series of investigations of the complexes B-BrC1 has revealed a pattern similar to that in the B...C12 series as B is systematically varied.34 There is now a congruence of the angular geometries with those of the B-HBr complexes and a systematic contraction of r(B--Br) from B.-HBr to Be-BrCl.Although the extent of electron transfer from Br to Cl when B.-BrCl is formed is larger [G(B--BrCl) = 26 (B-C12)] it is still small, suggesting that the B.-BrCl are also weak, outer complexes in which the interaction is dominated by the electrostatic term. This is reflected in the k, values, which are again small for the B...BrCl series, although in general k,(B.-BrCl) = 2k,(B...C12) . It is also revealing to compare the properties of the complexes in a series Be-XY, where B is fixed and XY = F2, Cl2, ClF, Br2 or BrC1. So far only the case B = NH3 has been investigated for all the defined XY.13335-38 The order of the strength of the binding (as measured by the intermolecular stretching force constant k,) is XY = ClF > BrCl > Br2 > C12 > F2, which is the order expected if electrostatics dominate the interaction in these complexes but not if charge transfer were predominant. Reaction 3 Despite the spectacular nature of the flame produced when NH3 and F2 gases are mixed under normal conditions of temperature and pressure, it has been possible to observe the rotational spectrum of the complex H~N-BF~ by employing the fast-mixing nozzle.13 Such an observation testifies to the power of this 114 Chem.Commun., 1996 device as a tool for observing pre-reactive intermediates. The complex H3N--F2 is found to have C3,, symmetry, with the weak interaction between the N atom of NH3 and the inner F atom of F2.The relative propensity of F2 as an electron acceptor in the formation of B-.XY donor-acceptor complexes has been discussed earlier in connection with reaction 2. Reaction 4 The complexes B.-.C12 discussed earlier are of the weak, outer type in the gas phase. If inner complexes [BCl]+ C1- are involved as intermediates in chlorination reactions in solution, it is likely that progress from B.--C12 to [BCl]+ C1- is solvent assisted. This raises the general question: Do inner complexes, implying significant transfer of charge, occur in the gas phase and can they be detected? If we include proton transfer in the general definition of charge transfer, the answer is yes. To encourage proton transfer in the gas phase and favour an ion pair BH+-.X- relative to the simple hydrogen-bonded species B-m-HX, B should have a high proton affinity while the energy required to dissociate HX into ions H+ and X- should be as small as possible.Ammonia has a high gas-phase proton affinity and this increases monotonically with progressive methylation. Among the hydrogen halides, the order of the appropriate dissociation energy is HF > HC1 > HBr > HI. Hence, of the ammonium and methylammonium halides, H3N.e.HF is the least likely and Me3N...HI is the most likely to exhibit proton transfer in the gas phase.39 By using the fast-mixing nozzle it has been possible to observe the rotational spectrum of the complex that results from mixing Me3N and HI.40 The detected complex is definitely a symmetric top molecule and the nuclei N.-H-..I lie in that order along the C3 axis.But the central question is: where is the proton? This question can be answered by considering the iodine nuclear quadrupole coupling constant X(1) and the intermolecular stretching force constant k,, which is available from the centrifugal distortion constant DJ. Table 3 compares X(1) for free HI, HCN-.HI, H3N-HI, Me3N.-HI, Na+...I- and I-. The complex HCNq-HI provides an example in which the Lewis base is a poor proton acceptor, so that proton transfer to HCN from HI is not expected.4' In the free ion I-the iodine nuclear quadrupole coupling constant would be zero because the electronic distribution is spherically symmetric and therefore the electric field gradient at the nucleus would be zero.It is generally accepted that the diatomic molecule NaI is an ion pair Na+-.I- in the gas phase, as indicated by the small magnitude of X(I)42 which differs from zero only because of the presence nearby of the Na+ ion. Table 3 illustrates that the values of X(1) for H3N...HI and Me3Na-HI are quite different. Clearly, the X(1) value of H3N--HI is close to that of HCN-HI, and both are a little reduced in magnitude from that of free HI. The reductions are readily understood in terms of a relatively large zero-point angular oscillation of the HI subunit in these c0mplexes.~9 On the other hand, Me3N-HI has a value of X(1) close to that of Na+--I-, indicating that a more accurate Table 3 Iodine nuclear quadrupole coupling constants X(1) and inter- molecular stretching force constants k, of some B .. . HI complexes HI -1823.40 -HCI4N...HI -1475.7( 1)' 4.56( 1)b H3'4N*.*HI -1324.891(18)~ 7.18(9)c Me314N...HI -341.204(14)d 66.5(2)dNa+...I--262. 14e 77.w ~ ~~~~~~ a F. C. DeLucia, P. Helminger and W. Gordy, Phys. Rev. A, 1971, 3, 1849. Ref. 41. ' Ref. 43. d Ref. 40. e Ref. 42. f Calculated from o, = (2nc)-l(k/p)+ using me given by J. R. Rusk and W. Gordy, Phys. Rev. A, 1962,127, 817. description of the complex is Me3NH+-.I-. This conclusion is reinforced by consideration of the magnitude of k, for trimethylammonium iodide, which is much larger than those of the simple hydrogen-bonded species HCN---HI and H3N.-HI but is indeed similar to that of the ion pair Na'ee-1- (see Table 3).39 The conclusions about trimethylammonium iodide and ammonium iodide in the vapour phase have implications for proton transfer.In observing the rotational spectrum of the former, the starting materials were Me3N and HI held separately until they met at the point of expansion into a vacuum at the coterminal exits of the concentric tubes that constitute the fast- mixing n0zzle.~0 Yet within ca. 10 ps of the encounter the proton, which was attached to I, was observed to be associated with Me3NH+. Conversely, the experiments with H3N. .-HI began with the ionic solid ammonium iodide which, on heating, led to the hydrogen-bonded species in the vap0ur.~3 In fact, the rotational spectra of several Me3-,HnN--HX complexes in the methylammonium halide series have been observed.This work has been reviewed elsewhere.39 The general conclusion is that as NH3 is progressively methylated the proton, which is attached to C1 in the first member H3N.-HCl of the X = C1 series, gradually shifts towards the N of the base but is only partially transferred by the time Me3N...HCl is reached. In the series Me3N-..HX, where X = F, C1, Br or I, the proton affinity of the base remains constant while the dissociation energy for HX = H+ + X- becomes smaller. It is found that Me3N-HF is a simple hydrogen-bonded complex, the proton is partially transferred in Me3N--H*+..Cl~- but Me3NH+-.Br-and Me3NH+.-I- are both ion pairs.These results are consistent with energetic considerations, as discussed elsewhere.39 By contrast, all Me 3-,H,P...HX (n = 3 and 0) so far investigated show no significant extent of proton transfer from X to P. Ultimately, the difference in behaviour in the N and P series can be attributed to the larger ionic radius of P+ compared with N+, which leads to a smaller Coulombic energy of interaction in P+H...X- than N+H...X- and this reflects unfavourably on the energy of the proton-transfer process P...H-X = PH+ ...X-. 39 Reaction 5 H3N.BF3 can be considered the prototype donor-acceptor complex. It was discovered by Gay-Lussac in 18095 and is the simplest member of the class of compounds R3B.NH3 used by G.N. Lewis to introduce the idea of the dative bond when discussing exceptions to the octet rule.44 Surprisingly, the compound H3N.BF3 was not identified in the gas phase until recently. Evidently, it is not regenerated when the solid precipitated on mixing the component gases is heated.6 In view of the place of H3N-BF3 as the first coordination compound, its r6le as a prototype for the dative bond and the fact that it is composed entirely of first row atoms, this species provides an important link between theory and experiment. A detailed investigation of its experimental properties in isolation in the gas phase therefore seemed desirable. The fast-mixing nozzle/FT microwave spectrometer combi- nation offered an ideal means of detecting the rotational spectrum of H3N.BF3 in mixtures of the components NH3 and BF3 before precipitation of the solid.Indeed, by this means the rotational spectra of the four isotopomers H3 l4N.IOBF3, H314Nv1 lBF3, H315N-IOBF3 and H315N.IIBF3 were detected, albeit with unexpectedly low intensity.45 After a difficult analysis, the 14N, loB and B nuclear quadrupole coupling constants were determined. The magnitudes of these quantities were small, thereby suggesting that the 14N and B nuclei sit in tetrahedral environments since in that case the electric field gradients (and therefore the nuclear quadrupole coupling constants, see Table 2) would be reduced in magnitude relative to the free components. An ab initio calculation46 has confirmed this conclusion and has demonstrated that when the electric field gradients & and gZalong the molecular symmetry axis z at N and B, respectively, are plotted against the separation r(N...B) both show a rapid increase in slope at a certain value of r(N.-B), implying a sudden transfer of an electron pair from N to B at a well defined distance. This is presumably a characteristic of the dative bond.Reaction 6 The interest in complexes (1000) Rg-mHX, where Rg is a rare gas atom and X is a halogen atom, lies in the fact that the state (1000) corresponds to one quantum of excitation in the HX stretching motion. In addition, the excitation energy is greatly in excess (by a factor of ca. 25 for Ara-HCI) of the energy required to dissociate the van der Waals bond.If (10OO) Rg..-HX molecules can be formed and brought to a state of collisionless expansion before collisional deactivation, they provide a model for the simplest possible unimolecular decomposition step in the familiar unimolecular mechanism, namely A* -+ products (2) or to use (1000) Ar...HCl as a specific example (1000) Ar--.HCl+ Ar + HC1 (v = 0) (3) The rotational spectra of ( 1000)Ar...HCl and (1O"0) Kr--.HCl have both been observed14.15 by using the glow-discharge nozzle (described earlier) in a Fourier-transform microwave spectrometer. Interpretation of the observed spectroscopic constants showed that the properties of these excited state complexes (such as the intermolecular stretching force constant k,, the angular oscillation of the HCl subunit and the Rg...CI distance) were only slightly changed by the vibrational excitation, although there was some indication of a minor strengthening of the interaction on vibrational excitation.The widths of individual rotational transitions were not detectably increased by the vibrational excitation (see Fig. 4, for example), so no measurement of the lifetime of (1OOO) Rg...HCI before the vibrational pre-dissociation process (3) was available from line broadening. However, another manifestation of a significant rate of vibrational pre-dissociation for ( 1OOO) Rg...HCl would be the loss of population of this species during the time of its collisionless flight between the point of vibrational excitation (the glow-discharge nozzle) and its detection in the appropriate region of the spectrometer.Loss of population by fluorescence from the (1000) state is negligible. By means of a series of experiments in which this time of flight was varied over a range of 400 ps, it was found that the relative population N,/No of the excited and ground states of both Ar-eHCl and Kr.-.HCl remained c0nstant.~7 The conclusion was that the characteristic decay time for vibrational pre-dissociation T,~exceeds 1 ms for both (1000)Ar-aeHCl and (1000) Kr...HCl. It is of interest to note that in 1 ms the HC1 molecule executes in excess of 10" vibrations. The van der Waals mode and the HC1 stretching mode in these Rg.-HCl are indeed extremely weakly coupled.Acknowledgements I thank my coworkers, whose names appear in the appropriate references, and the EPSRC for their support. A. C. Legon is the Professor of Physical Chemistry in the University of Exeter. He was born at Rookery Farm, near Sudbury in Suffolk but was educated in London: at the Coopers' Company School, Bow and afterwards at University College London. His recent research interests include a systematic investigation of the nature of hydrogen-bonded dimers and other types of complex through the spectroscopy of supersonic jets. He was Tilden Lecturer and Medallist of the Royal Society of Chemistry for 1989-90. Chem. Commun., 1996 115 References 1 A. C. Legon, in Atomic and Molecular Beam Methods, ed. G. Scoles, Oxford University Press, New York, 1993, vol.2, ch. 9. 2 C. K. Ingold, Structure and Mechanism in Organic Chemistry, 2nd edn., G. Bell, London, 1969, ch. XIII,pp. 964-982 and references therein. 3 R. S. Mulliken and W. B. Person, Molecular Complexes, Wiley-Interscience, New York, 1969 and references therein. 4 R. S. Mulliken, Science, 1967, 157, 13 and references therein. 5 J. L. Gay-Lussac and J. L. ThCnard, Mem. de Phys. et de Chim. de la SOC. d’Arcueil, 1809, 2, 210. 6 F. J. Lovas and D. R. Johnson, J. Chem. Phys., 1973, 59,2347. 7 T. R. Dyke, B. J. Howard and W. Klemperer, J. Chem. Phys., 1972,56, 2422. 8 T. J. Balle, E. J. Campbell, M. R. Keenan and W. H. Flygare, J. Chem. Phys., 1979, 71, 2723. 9 See, for example, C. H. Townes and A.L. Schawlow, Microwave Spectroscopy, McGraw-Hill, New York, 1955; W. Gordy and R. L. Cook, Microwave Molecular Spectra, in Techniques of Chemistry, ed. A. Weissberger, 3rd edn., Wiley-Interscience, New York, 1984, vol. XVIII; H. W. Kroto, Molecular Rotation Spectra, Dover Publications, Inc., New York, 1992. 10 D. R. Miller, in Atomic and Molecular Beam Methods, ed. G. Scoles, Oxford University Press, New York, 1988, vol. 1, ch. 2. 11 T. E. Emilsson, T. D. Klots, R. S. Ruoff and H. S. Gutowsky, J. Chem. Phys., 1990, 93, 6971. 12 A. C. Legon and C. A. Rego, J. Chem. Soc. Faraday Trans., 1990,86, 1915. 13 H. I. Bloemink, K. Hinds, J. H. Holloway and A. C. Legon, Chem. Phys. Lett., 1995, 245, 598. 14 J. W. Bevan, A. C. Legon, C.A. Rego and J. Roach, Chem. Phys. Lett., 1992, 198, 347. 15 J. W. Bevan, A. C. Legon and C. A. Rego, J. Chem. Phys., 1993, 98, 2783. 16 A. C. Legon, C. A. Rego and A. L. Wallwork, J. Chem. Phys., 1992,97, 3050. 17 A. C. Legon, Faraday Discuss.Chem. Soc., 1994,97, 19. 18 F. H. de Leeuw, R. van Wachem and A. Dymanus, Symposium on Molecular Structure and Spectroscopy, Ohio, 1969, Abstract R5. 19 A. C. Legon and J. C. Thorn, Chem. Phys. Lett., 1994,227,472. 20 A. C. Legon, A. L. Wallwork and D. J. Millen, Chem. Phys. Lett., 1991, 178, 279. 21 M. J. Atkins, A. C. Legon and H. E. Warner, Chem. Phys. Lett., 1994, 229, 267. 22 C. M. Evans and A. C. Legon, Chem. Phys., 1995, 198, 119. 23 J. CoslCou, D. G. Lister and A. C. Legon, Chem.Phys. Lett., 1994,231, 151. 24 N. W. Howard and A. C. Legon, J. Chem. Phys., 1988,88,6793. 25 A. C. Legon, Chem. Phys. Lett., 1995,247,24. 26 A. C. Legon and D. J. Millen, Faraday Discuss. Chem. Soc., 1982,73, 71; Chem. SOC.Rev., 1987,16,467. 27 H. I. Bloemink, K. Hinds, A. C. Legon and J. C. Thorn, J. Chem. SOC., Chem. Commun., 1994, 1321. 28 H. I. Bloemink, K. Hinds, A. C. Legon and J. C. Thorn, Chem. Eur. J., 1995, 1,17. 29 B. P. Dailey and C. H. Townes, J. Chem. Phys., 1955,23, 118. 30 Calculated using me = (2nc)-l (k/p)i with o,given by P. L. Clouser and W. Gordy, Phys. Rev. A, 1964,134, 863. 31 A. C. Legon, Chem. Phys. Lett., 1995,237,291.This reference cites the individual papers for the complexes in the B..C12 series. 32 H. I. Bloemink, K. Hinds, A. C. Legon and J. C. Thorn, Chem. Phys. Lett., 1994, 223, 162. 33 S. A. Peebles, P. W. Fowler and A. C. Legon, Chem. Phys. Lett., 1995, 240, 130. 34 A. C. Legon, J. Chem. SOC., Faraday Trans., 1995, 91, 1881. This reference cites the individual papers for complexes in the Be-BrC1 series. 35 A. C. Legon, D. G. Lister and J. C. Thorn, J. Chem. SOC., Faraday Trans., 1994, 90, 3205. 36 H. I. Bloemink, C. M. Evans, A. C. Legon and J. H. Holloway, Chem. Phys. Lett., in the press. 37 H. I. Bloemink and A. C. Legon, J. Chem. Phys., 1995,103, 876. 38 H. I. Bloemink, A. C. Legon and J. C. Thorn, J. Chem. Soc., Faraday Trans., 1995, 91, 781. 39 A. C. Legon, Chem. SOC. Rev., 1993,22, 153. This review is a source of references to the rotational spectra of all members of the series Me3-,H,N.-HX and Me3-,H,P.-HX so far investigated and also of references to the proton affinities of the bases and the dissociation energies for HX = H+ + X-. 40 A. C. Legon and C. A. Rego, J. Chem. Phys., 1993,99, 1463. 41 P. W. Fowler, A. C. Legon and S. A. Peebles, Chem. Phys. Lett., 1994, 226, 501. 42 C. E. Miller and J. C. Zorn, J. Chem. Phys., 1969, 50, 3748. 43 A. C. Legon and D. Stephenson, J. Chem. Soc., Faraday Trans., 1992, 88, 761. 44 G. N. Lewis, Valence and the Structures of Atoms and Molecules, Dover Publications, Inc., New York, 1966, ch. VIII, p. 98. 45 A. C. Legon and H. E. Warner, J. Chem. Soc., Chem. Commun., 1991, 1397. 46 D. Fujiang, P. W. Fowler and A. C. Legon, J. Chem. SOC., Chem. Commun., 1995, 113. 47 J. W. Bevan, A. C. Legon and C. A. Rego, Chem. Phys. Lett., 1993,204, 551. Received, 14th September 1995; 5106095H 116 Chem. Commun., 1996
ISSN:1359-7345
DOI:10.1039/CC9960000109
出版商:RSC
年代:1996
数据来源: RSC