摘要:

QUARTERLY REVIEWS MOLECULAR INTERACTIONS IN CEATHRATES A COM- PARISON WITH OTHER CONDENSED PHASES By WILLIAM C. CHILD jun. (EDWARD DAVIES CHEMICAL LABORATORY ABERYSTWYTH) * Introduction General Characteristics of C1athrates.-As implied by their name the principal novelty of clathrates lies in their geometrical structure. The X-ray studies of Powell and his co-workers1 provided the details for a number of typical examples in which “guest” molecules (A) are entrapped within cavities often pseudo-spherical in form which appear in the “host” lattice (B). Two aspects are immediately noteworthy first the lattice structure of B in the clathrate is not its normal crystalline state and the form with unoccupied cavities is usually less stable; secondly the composi- tion of the clathrate A,nB can vary over a range often large in which the clathrate is stable with respect to decomposition into guest and normal host.Accordingly the clathrates are not stoicheiometric compounds minimum values of n and the restrictions on the possible A components are dictated largely by geometric considerations. The best known examples have B = H20 (ice) forming the so-called gas hydrates of approximate comp- ositions n = 6 when A = Cl, SOz CH4 Ar Kr Xe etc. or n = 17 when A = CH31 C3H, etc. ; and B = p-HOC,H,-OH (quinol) A = CO N, H2S CH30H Ar Kr Xe etc. with n = 3.0 ideally (Le. all cavities filled) but usually n - 3.5 to 8 or more depending on the conditions under which the clathrate is formed. Powell’s and von Stackelberg’s researches have brought out very clearly the structural features of these cage-like complexes while J.H. van der Waals has evaluated the thermodynamic aspects leading to their formation. In many respects they provide an interesting model of solution equilibria in that the “solute” molecules (A) are placed in cavities within the “solvent” (B). As van der Waals and other assessors have concluded the energy of interaction (A)-(B) is normally small in the clathrates and the entropy term is similar in magnitude to the entropy of vaporisation of a solute from a solution in which the solvent obeys Raoult’s law. Further the A component may be said to obey Henry’s law in the sense that (A)-(A) interactions are negligible compared with (A)-(B) interactions. Because the (a) D. E. Palin and H. M. Powell J. 1947,208; (b) H. M. Powell ibid. 1948 61; (c) D. E.Palin and H. M. Powell ibid. 1948 571; ( d ) D. E. Palin and H. M. Powell ibid. 1948 815; (e) J. H. Rayner and H. M. Powell ibid. 1952 319; (f) S. C. Wallwork and H. M. Powell ibid. 1956 4855; (8) D. Lawton and H. M. Powell ibid. 1958 471. 321 * Present address Carleton College Northfield Minnesota. U.S.A. 1 322 QUARTERLY REVIEWS environment of the guest molecule can be described fairly precisely calculation of the interaction energy is more straightforward than it is for liquid solutions and the cell model sometimes employed in theoretical treatments of the liquid phase can be applied realistically to clathrates. Not surprisingly therefore both theoretical and practical interest in clathrates appears to be increasing. Diverse uses for clathrates have become apparent and more will certainly be developed the purification of sea water,2 the separation of aromatic mixture^,^ and the introduction of clathrates as carriers4 for radioactive 85Kr and for free radicals5 are a few examples.The major purpose of this Review is to summarise the information revealed by thermodynamic data on the interaction energies of guest molecules with their encaging structures and that on the motion of the guest molecules provided by dielectric and infrared studies. Most of the discussion will relate to the quinol clathrates and the gas hydrates because it is for these that the largest body of physicochemical data has accumu- lated. While a number of clear descriptive surveys of clathrates is avail- able,6-lo a brief reminder of the relevant structural aspects is first presented.Structures of Host Lattices.-As a first approximation the cavities in which the guest molecules are held can be regarded as spherically sym- metric. Refinements to this approximation which are required by a number of experimental results will be discussed later. The structures of the cavi- ties in the quinol clathrates and the gas hydrates revealed by X-ray diffraction studies are the chief basis for the assumption of spherical symmetry. In quinol clathrates the principal forces which hold the lattice together are hydrogen bonds between hydroxyl groups. Groups of six oxygen atoms are linked together to form hexagonal rings which are parallel to one another and constitute the “top” and “bottom” of each cage (see Fig. 1).l0 The sides of the cage consist of portions of six benzene rings which are inclined at angles of 45” with respect to the plane of the hydroxyl hexagons.Since each hexagon participates in two cages and since six quinol molecules are linked together in each hexagon there are on the average three molecules of quinol per cage. The “free diameter” of the cage or the distance from the boundary of an atom on the wall to the A. J. Barduhn H. E. Towlson and Yee-Chien Hu U.S. Dept. Office Tech. Ser. P.B. Rept. 171,031 (1960); Chem. A h . l962,57,4477J P. de Radzitsky and J. Hanotier Ind. Eng. Chem. Process Design Develop. 1962 1 10. G. J. Rotariu E. L. Hoskins and D. M. Hatori U.S. Atomic Energy Comm. TI D-17223. ii P. Goldberg J. Chem. Phys. 1964 40,427. Martinette Hagan J. Chem. Educ. 1963,40 643. ’ Martinette Hagan “Clathrate Inclusion Compounds” Reinhold New York 1962.R. M. Barrer in “Non-Stoichiometric Compounds” ed. L. Mandelcorn Academic L. Mandelcorn Chem. Rev. 1959,59,827 lo H. M. Powell in “Non-Stoichiometric Compounds” ed. L. Mandelcorn Academic Press Inc. New York 1964 chap. 6. Press Inc. New York 1964 ch. 7. CHILD MOLECULAR INTERACTIONS IN CLATHRATES 323 boundary of an atom directly opposite is approximately 4-2 8 when van der Waals radii are used. The geometry of the cage reveals that the force field within it is not precisely spherical. In fact guest molecules such as sulphur dioxide and carbon dioxide have preferred orientations with their long dimension parallel to a line joining the centres of the hexagons.ld ( b ) FIG. 1. (a) The positions of the six quinol molecules that form the immediate surroundings of an argon atom are shown in aperspective drawing.This is deceptive because the centres of the atoms are shown and the space occupied by the atoms cannot be shown at the same time. (6) The spaces occupied by the atoms are indicated in a drawing on the same scale as that above. A few only of the atoms surrounding the central argon atom are shown. They are distinguished from those which have been omitted by small circles drawn around their centres. Broken lines are used to indicate that a part of the structure so represented lies behindsome portion drawn with full lines. (Reproduced with permission from H. M. Powell “Non-Stoichiometric Compounds,” ed. L. Mandelcorn Academic Press Inc. New York 1964 ch. 7.) A closer approximation to spherical symmetry is found in the crystal structures of the gas hydrates.*J1J2 Two common cubic crystal structures termed Structure I and Structure IT have been found for this class of clathrate.Recently several more novel structures have been di~c0vered.l~ The unit cell of Structure I has forty-six water molecules two small cavities and six larger cavities. The small cavities are bounded by twenty water molecules the oxygen atoms of which are nearly equidistant from the centre so that the cavity has a free diameter of about 5.1 I$. Twenty-four M. von Stackelberg and H. R. Muller 2. Electrochem. 1954 58 25. l2 J. H. van der Waals and J. C. Platteeuw Adv. Chem. Phys. 1959,2 1. l3 P. T. Beurskens and G. A. Jeffrey J. Chem. Phys. 1964,40,906 and earlier papers. 324 QUARTERLY REVIEWS water molecules surround the larger cavities which have an average free diameter of 5.8 A with a variation of & 0-3 A.From the ratio of water molecules to cavities it can be seen that if each cavity contains one guest molecule the composition of the hydrate is 5.75 molecules of water per molecule of guest. Structure I1 hydrates contain cavities which are more disparate in size and so these hydrates usually form when the guest molecules are too large to fit into the cavities of Structure I. The unit cell of 136 water molecules contains sixteen small and eight large cavities. Co-ordination numbers and free diameters are 20 and 28 and 5.0 A and 6.7 A respectively. Distances from centre to wall are nearly constant for the two kinds of cavity. If only the large cavities are singly occupied by guest molecules the composition of the hydrate is seventeen water molecules per guest molecule e.g.C,H,,l 7H20. Thermodynamic Analysis of Clathrate Stability Phase Relationships.-When compared with the vapour pressures of the pure guest liquids the dissociation pressures of clathrates are surprisingly small. For example at the boiling point of argon - 186"c the equilibrium dissociation pressure of the argon-quinol clathrate is 2 x atm. The hydrogen sulphide hydrate has a dissociation pressure of 0.9 atm. at O" compared with a vapour pressure of 10 atm. for liquid hydrogen sul- phide at this temperature. Thus it is apparent that the difficulty of escape of the guest molecule from the cage that is a high activation energy for decomposition is not the only factor relating to clathrate stability. The thermodynamic driving force for the formation of many of these clathrates is relatively large.Yet it is also evident that these complexes have neither fixed composi- tions nor chemical bonds between host and guest. They are best regarded as solid solutions of the guest molecule in the host lattice which when empty has a chemical potential slightly larger than that of the stable form of the host lattice ( a - f ~ r m ) . ~ ~ ~ ~ If one were to start with the empty metastable lattice of quinol (p-form) and slowly increase the pressure of a suitable foreign gas in contact with it the cages would gradually become occupied with a consequent stabilisation of the host lattice and lowering of its chemical potential. At a certain pressure called the dissociation pressure the degree of cavity occupation is such that the chemical potential of the clathrate host becomes equal to that of the a-form.Under this condition the three phases clathrate a-form and gas can be in equilibrium. As the pres- sure is raised above the minimum value needed to make the clathrate stable an increasing fraction of the cages is filled and the chemical potential of the host in the clathrate falls below that of the a-form. This is the region of stable solid solutions of varying composition. These considerations can be illustrated by reference to data for the l4 L. A. K. Staveley Adv. in Chem. 1963 No. 39 218. CHILD MOLECULAR INTERACTIONS IN CLATHRATES 325 argon-quinol clathrate at 25".12 The molar free energy of p-quinol is 82 cal./mole greater than that of a-quinol. When a fraction equal to 0.34 of the cavities is filled which occurs at an argon pressure of 3.4 atm.the chemical potential of the P-quinol has been reduced by 82 cal./mole and the chemical potentials of the a- and ,%forms are equal. At higher pressures larger fractions of the cavities are filled the relationship between the two quantities being given by an equation identical in form to the Langmuir adsorption isotherm. The latter fact suggests that there is no interaction between the guest molecules in neighbouring cages i.e. that the "solution" is ideal. This conclusion is confirmed by the observation that the equili- brium clathrates of both argon-quinol and krypton-quinol have 34% of the cages filled at 25". Thus this faction appears to be independent of the identity of the guest as long as the lattice is not distorted by the guest molecules and the solvent (P-quinol) then obeys Raoult's law.A portion of the phase diagram at constant temperature (25") of the argon-quinol system is given in Fig. 2. The solid curve represents equili- brium pressures of stable clatlirate systems while the broken curve corresponds to metastable clathrates. Point I is the invariant point (triple point) clathrate-a-quinol-argon gas. 3 - 4 1 1 OO (4)(0.34) G:)(l-oo) Mole fitio argon:quinol FIG. 2. Phase diagram for argon-quinol at 25". Enthalpies and Entropies of Dissociation.-Theoretical calculations of dissociation pressures and interaction energies of quinol clathrates and/or gas hydrates have been made by several author^.^^,^^-^^ The cell model has been employed in all treatments. It is assumed that the potential energy of interaction between a guest molecule and an atom in the wall of its cage is given by an equation such as the Lennard-Jones 6-12 potential function.l5 R. M. Barrer and W. I. Stuart Proc. Roy. SOC. 1957 A 243,172. l6 R. M. Barrer and D. J. Ruzicka Trans. Faraday SOC. 1962,58,2253. V. McKoy and 0. Sinanoglu J. Chem. Ptiys. 1963,38,2946. 326 QUARTERLY REVIEWS After summation over all atoms in the wall of the cage the potential curve for the interaction between the guest molecule and its cage is found to have a broad minimum and steep sides. Therefore it would not be far wrong to visualise the guest molecule as moving about in its cage in somewhat the same manner as a molecule in the gas phase moves in its container with the very important difference that the volume of the cage is approximately 1/600th of the volume of a gas at S.T.P.divided by the number of molecules in the gas. Rather than speak of the translational motion of the guest molecule in the cage the more realistic term “rattling motion” is used. The theoretical treatments have also employed the assumption that the internal degrees of freedom of the guest molecule are the same as for the molecule in the gas phase. With the use of one or no adjustable parameters depending on the particular treatment the agreement between calculated and observed energies of interaction and dissociation pressures is fairly good for a considerable number of systems. Hence the assumption of spherically symmetric cavities is adequate to allow the prediction of some of the thermodynamic properties of clathrates with non-polar or slightly polar guest molecules.By use of some of the theoretical results in a qualitative or semi- qualitative fashion it is possible to ascertain whether clathrate stability is related primarily to enthalpy or to entropy factors. Before proceeding further we must be precise about the dissociation process whose equili- brium position determines the stability. From the theoretical point of view it is convenient to discuss dissociation of the clathrate into the metastable host lattice and gaseous guest according to the equation A,nB (s) -f nB (s /?-form) + A(g) . . . . (1) Here A and B stand for guest and host respectively and n has the value corresponding to the invariant point on the phase diagram for example point I in Fig. 2. If this change occurs in a system showing no specific interaction energy beyond dispersion and repulsion terms for (A)-(B) then the appropriate concentration and statistical factors lead to an en- tropy change OSoD given by the equationla (2) Aso - R In - - RT ~ R - S .. . . Vf B - where V is the molar volume of the gaseous guest in the standard state Vf is the “free” volume per mole in the clathrate and is a measure of the empty space in the cavities and Sc is the configurational entropy of the clathrate. The first term on the right-hand side of equation (2) is simply the increase in entropy which occurs on expanding a gas from volume Vf to volume V. The second term is related to the departure of the potential function for the interaction between guest and cage from the square-well potential function.There is a contribution of R to the AS because the gas- W. C. Child jun. J. Phys. Chern. 1964 68 1834. CHILD MOLECULAR INTERACTIONS IN CLATHRATES 327 eous guest has communal entropy while the guest molecules in the clath- rate do not. Finally Sc the configurational entropy appears because some of the cavities are vacant and there are a number of ways in which N guest molecules can be arranged in a larger number of cavities. Equation (2) it is important to emphasise is valid provided that the internal degrees of freedom of the guest molecule are the same as in the gas phase and there is no distortion of the host lattice by the guest. The change in internal energy AU", accompanying the process speci- fied by equation (1) also has a simple interpretation it is equal to the average potential energy of interaction between guest and host to a fair approximation ( < 15 % error).Before we consider experimental values of AS" and AH" and their significance it may be informative to examine the way in which these two thermodynamic quantities depend on cavity size. Three cases will be con- sidered (1) cavity diameter much larger than the diameter of the guest molecule (2) cavity slightly larger than guest and (3) cavity slightly smaller than the size of the guest as measured by van der Waals radii. The qualitative features of the potential functions for these three situations are given in Fig. 3. (1) When the cage is considerably larger than the guest molecule for example 20 A versus 48 for the respective diameters then the internal energy of the gaseous guest is not very different from that of the guest in the clathrate because the guest molecule spends a large proportion of the time near the centre of the cavity where the potential energy of the system is zero relative to the separated guest and host.Therefore AU" N 0 and AHoP 21 RT "_ 600 cal./mole at 2 9 8 " ~ . AS"$ + Sc can readily be calculated from equation (2) with the assumption that the second term on the right is zero. To obtain Vf for the calculation the approximate relationship12 can be used. In of the cage and this equation a is the average distance between the centre the nuclei of the atoms in the wall of the cage u is the dist- ance of closest approach between a guest molecule and an atom in the wall (i.e. zero-energy collision diameter) and N is Avogadro's number.Equation (3) is valid in the square-well approximation. Using 10 8 and 3 8 for a and 0 respectively we find Vf to be 860 ~m.~/rnole thirty times smaller than the molar volume of the gaseous guest. Substitution of this value into equation (2) yields when the second term is neglected AS" + Sc = 8.6 e.u. at 298 OK. Combining this quantity with the AH" estimated earlier we obtain AGO - TSc = AHoj - T(AS"B + S,) = 600 - (298) (8.6) - - -2000 cal. at 2 9 8 " ~ 328 QUARTERLY REVIEWS It will be seen later that AGO - TS for the argon-quinol clathrate for which a = 4 A is equal to -1200 cal. at 2 9 8 ” ~ . Thus there is only a slightly greater tendency for a clathrate having the relatively large cavities assumed in this example to decompose into the metastable host lattice plus guest than there is for the argon-quinol clathrate to undergo a similar decomposition provided that in each clathrate the cavities are completely occupied so that Sc = 0.Now the stability usually observed in the laboratory is related not to d Gop but rather to d Go, which corresponds to the process A,nB (s) -+ nB (s a-form) + A (g) . . . . (4) Comparison of equations (1) and (4) shows that d G ’ - d G O = n (GB - GJ where G - G is the difference between the molar free energies of metastable and stable host. Therefore clathrate stability depends in addition to the factors already mentioned upon the value of G - G,. This quantity is only 82 cal./mole for the a-quinol-P-quinol systeiT1 as discussed earlier. It could be expected to be much larger however for a lattice with larger cavities because the sum of the van der Waals attractions in such a lattice would be much less than in a more compact form.If the fraction of the cavities occupied in this host were small when the clathrate was in equilibrium with stable host and gaseous guest additional stability would arise from the configurational entropy. This is unlikely however because a small fraction occupied implies a small value of G - G, contrary to the statement just made. Therefore the relative instability of the host lattice probably accounts for the fact that such clathrates are not commonly found unless there is more than one guest molecule per cavity. It should be noted also that clathrates with large cavities generally have fairly large escape holes.8 An additional characteristic therefore is a small activation energy for decomposition.It is undoubtedly this reason in addition to low thermodynamic stability which accounts for the non- existence of clathrates with the very small guest molecules H2 He H,O HF and Ne. (The last molecule however is found in some double gas hydrates1 g (2) When the “free diameter” of the cavity is say 1 A greater than the van der Waals diameter of the guest molecule the average interaction energy can be appreciable as shown qualitatively by curve (b) of Fig. 3. Since the free volume is of the order of 1 ~rn.~/mole ASo + Sc is much larger than it is for (hypothetical) clathrates with relatively large cages. That is the partial molar entropy of the guest in this second type of clath- rate is relatively small reflecting the reduced freedom of motion of the guest molecule within the cage.Again using as an example the argon-quinol clathrate we have the experimental values dHoB = 6.0 kcal./mole and ASo + Sc = 24~0e.u.~ which whencombined as beforegivedGoB-TSc = - 1 150 cal. at 298 OK. This quantity is related to d Go as follows R. M. Rarrer and D. J. Ruzicka Tram. Frrrnday SOC. 1962 58 2239. CHILD MOLECULAR INTERACTIONS IN CLATHRATES 329 AGOp - TSc = -1150 cal. + TSc = 1130 -n (GB - G,) = -720 AGO = -740 cal at 2 9 8 " ~ AGO is actually less negative than AGO - TSc showing the important contribution of configurational entropy to clathrate stability in this ex- ample. In general clathrates of this type are among the most stable found. Typical values of AHop and AS" + Sc will be compared with heats and entropies of vaporisation of the pure liquid guest substances in a later section.(3) Conning now to those clathrates in which the guest molecule is so large that it can barely be squeezed into the cavity we find that the trend in stability is easily predicted. In the first place the average interaction energy is small compared with that of clathrates in the second group because repulsive forces are more significant (see curve ( c ) of Fig. 3). FIG. 3. Potential functions for three cage sizes showing only the qualitative features. V = potential energy r = distance from the centre of the cage. Curve (a) cage much Iarger than guest molecule curve (b) cage slightly larger than guest; curve (c) cage slightly smaller than guest. Secondly d So + Sc is larger than before because of a smaller free volume. Thirdly AS" + Sc may be larger still because of a restriction on the rotation of the guest molecule in the cavity.Finally GB - G may be anomalously large because of lattice distortion. All four factors operate in the direction of reducing clathrate stability. While quantitative data for clathrates of this sort are sparse methanol-quinol and especially methyl cyanide-quinol are examples of this category. van der Waals and Plat- teeuw12 have found that the fraction of the cavities occupied in the equili- brium methanol-quinol clathrate at 298 OK is 0.474 compared with 0.34 for both the argon- and krypton-quinol clathrates. This difference is to be attributed to a larger value of G - G in the methanol-quinol clath- 330 QUARTERLY REVIEWS rate. d Hop for the methyl cyanide-quinol clathrate is smaller than expected when compared with AHvap of liquid methyl cyanide (see later discussion).Experimental enthalpies and entropies of dissociation of quinol clath- rates gas hydrates and a few others are given in Table 1. Only for the quinol clathrates are dHop and dSop known with any accuracy. These quantities have been estimated for the gas hydrates by the application of several assumptions.18 Comparison of AH" and dHop with dHvap of the Pure Liquid Guest.- Examination of the data in Table 1 reveals that dHop is larger than dHvap of the liquid guest at its boiling point by a factor of 2-3 in most cases. The ratio LlH"p/AHvap appears to be systematically larger for the quinol clathrates than for the gas hydrates but this observation may not be sig- nificant because the values of AH", for the latter clathrates are based on an estimate of H p - Ha which has not been determined experimentally.In any case it is evident that AH" is also considerably larger than AH,,,. Therefore one can conclude that one reason why the dissociation pressures of clathrates are smaller than the vapour pressures of the pure guest liquids is that AH" > AH,,,. An explanation for the large ratios AHo&AHvap can be sought among three factors (1) the removal of a guest molecule from the clathrate leaves an empty cavity whereas the removal of a molecule from the liquid does not leave a hole in the liquid; (2) the energy of interaction between a guest molecule and an element in the wall of its cage may be larger than the energy of interaction between a pair of molecules in the liquid; (3) the co-ordination number of the guest molecule may be larger than the co- ordination number of a molecule in the liquid.(1) The removal of one mole of guest from a clathrate requires an energy d Uop equal to NZE where N is Avogadro's number z is the co-ordination number of a guest molecule and E is the energy of interaction between the guest molecule and a molecule or small group of atoms in the cavity wall. By contrast the energy of vaporisation of a liquid A Uvap should be com- puted as ~ N z E . The factor 4 is present to avoid counting interacting pairs twice. If we ignore for the moment any differences in z or E it becomes apparent that AUop should more properly be compared with 2dUvap which represents the energy required to vaporise one mole from a large amount of liquid in such a way as to leave behind one mole of holes in the remaining liquid.Examination of Table 1 shows that 2AHvap comes fairly close to AHOF although the latter is frequently larger than the former. (2) This observation that AHop often exceeds 2AHvap must in this simple approach be related to differences in z and/or E . The second of these factors the average energy of interaction between a molecule and one of its neighbours depends on the Lennard-Jones constants for the molecule and its neighbour and the free volume of the system. Inspection of typical TABLE 1. Guest Ar Kr N2 0 2 co CH4 HC1 HBr H*C02H CH3-OH so2 CH3CN Experimentally determined and estimated molar enthalpies and entropies of dissociation of clathrates. AH;" AH;^ AH,,,^ AH~/2AHvap AS2 ASP' A s ; + sCd (kcal./mole) (kcal./mole) (kcal./mole) (e.u./mole) (e.u./mole) (e.u./mole) Quinol clathrates A,8.83C6H,(OH), T = 298 OK.4.6e 6-Of 1.56 1.9 17.8' 20.1" 23.9" 18.1k 20.4"" 24.2' 4.9 6-39 2.16 1.5 14.6' 16.9" 20.7" 2O*Ok 22*3aa 26-19 4.4 5.8f 1.33 2.2 18.3' 20.6" 24.4" 1 8 ~ 6 ~ 20.9"" 24*7bb 4.1 5.5f 1 -63 1-7 21-4k 23.7"" 27*Sbb 20-3k 22*6aa 26*4bb 20*2k 22.5aa 26-3cc 5.8 7.2h 1 -96 1.8 21.Ok 23~3"~ 27.1h 7.8 9-2f 3.86 1.2 -1 7.0' -1 9*Om ~ 2 3 . 0 ~ 8.8 10.2' 4.21 1.2 22-5k 24.gaa 28.6" 10-8 12.2f 5.32 -10.0 -1 1.0f 8-43 9-2' 21.7" 14.2' 6.1 3 7.4" 7.9 7.83 0.5 5.6' 6.7' 4.46 0.8 22.3' 4.55 5.6 1 -96 1.4 22.6 5.8 6.9 3.52 1.0 26.2 7.7 8.8 5-96 0.7 26-5 Structure I gas hydrates A,6.0H20 T = 2 7 3 " ~ 20.2t Y 20.5 24- 1 24.4 w w CL w bJ TABLE 1 .-continued w Guest AH:^ AHjb A Hvapc A H;/2A Hvap A p Asjb AS; + SCd (kcal./mole) (kcal./mole) (kcal./mole) (e.u./mole) (e.u./mole) (e.u./mole) Structure I gas hydrates A,7.7H20 T = 2 7 3 " ~ Cl2 6.5O 7.9' 4-88 0.8 23*78 20*5t 'I/ CH,Br 8.1 9.5 5.72 0.8 27.0 24.0 CH31 7 .3 O 10.7" 6-7 0.8 22.15 23*Ot Y C,H,+CI 8.7 12.1 5.9 1.0 29.2 30.1 6.3 9.7 4.49 1.1 25-3 26.2 Structure I1 gas hydrates A,17H20 T = 2 7 3 " ~ 8.Y 11.9 6.4 0.9 27.3P 28.2 0 27.7 s Miscellaneous T = 2 9 8 " ~ 76 C G H f P -JlO.OW 7.35 3 GH8 CHCl2F CBrClF 8.259 11.6 5-34 1.1 26.8 P-CsH4(CH3),,YV 17.8" 10.1 0.9 p-C,H4CI2,YV 21.3" 11.0 1.0 3 %efers to the process A,nB (s) 3 nB (s stable) + A (g) where n has the value given in the various formulae in the Table. bRefers to the process A,nB (s) -+ nB (s metastable) + A (g) where n is the same as in (a).'Enthalpy of vaporisation of the pure liquid guest at its normal boiling point. Most of the values were taken from National Bureau of Standards Circular 500. dSc = configurational entropy eUnless noted otherwise AH; for the quinol clathrates was calculated from AH; and Hp - Ha the latter from D. F. Evans and R. E. Richards Proc. Roy. SOC. 1954 A 223 238. fEvans and Richards ref. e. gRef. 26c. hRef. 26b. iH. G. McAdie Canad. J. Chem. 1963,41 2137. jRef. 26a. kCalc. from AS; and SB - S, the latter from ref. 12. zCalc. from AH; and dissociation pressure the latter from ref .12. MCalc. from AS; and S - S, the latter from ref. 12. flCalc. from AS; and Sc = 3.8 e.u. 'Except where noted otherwise AH for the gas hydrates is taken from 6. N. Glew J. Phys. Chem. 1962,66,605. PW. P. Banks B.0. Heston and I?. F. Blankenship J. Phys. Chem. 1957 58 962. QD. N. Glew Canad. J. Chem. 1960 38 208. "All values of AH; for the gas hydrates are estimates obtained from AH and an estimated value of H8 - Ha ref. 18. Wnless noted otherwise AS; for the gas hydrates is taken from D. N. Glew ref. q. tAll values of AS; for the gas hydrates are estimates obtained from AS and an estimated value of SB - S, ref. 18. uX = Ni(CN)2NH3. vY = Ni(CNS)z (CH,C,H,N). E. Aynsley W. A. Campbell and R. E. Dodd Proc. Chem. SOC. 1957 210. "M. I. Hart jun. and N. 0. Smith J. Amer. Chern. SOC. 1962,84 1816. YWithin experimental error Sc = 0 for the gas hydrates. 'W. F. K. Wynne-Jones and A. R. Jones Compt. rend. de la 2e RCunion de Chimie Physique (2-7 juin 1952 Paris). aaCalc. from AS; + SC and Sc = 3.8 e.u."Calc. from experimental heat capa- cities given in ref. 26d. CCG. L. Stepakoff Diss. Abs. 1963 24 1424. 2 CHILD MOLECULAR INTERACTIONS IN CLATHRATES 333 values12,18,20 of these quantities for both clathrates and the pure guest liquids shows that the pair-wise interaction energies are not appreciably different. (3) In a close-packed liquid the co-ordination number is 12 while in a more random and perhaps more realistic molecular arrangement the most probable co-ordination number21 is 8. On the other hand the number of nearest elements about a guest molecule is about 24. Here “element” refers to a water molecule in the case of the gas hydrates and to a C-H 0-H or C group or atom in the case of the quinol clathrates. Hence an element is not always comparable in size to the guest molecule.Never- theless it is felt that the larger co-ordination numbers in the clathrates probably account for some of the difference between AH”, and 2AHvap at least for the smaller guest molecules. To summarise this discussion of enthalpies we can visualise the vapori- sation of a liquid and the dissociation of a clathrate into stable products as each occurring in two hypothetical steps. Regarding the liquid we imagine that first one mole of vapour is removed from a liquid with the simultaneous creation of one mole of holes a process which is accompanied by an enthalpy increase of 2AHvap. The second step is the collapse of the remaining liquid to produce a normal liquid; the enthalpy change for the process is -AHvap. In an analogous fashion the sequence of steps for the clathrate is first the removal of one mole of guest from the lattice for which the enthalpy change is AHop and second the collapse of the /3-lattice to give to the a-form for which A H = -n(Hg - Ha).By com- parison first of the second steps our discussion has shown that an im- portant enthalpy factor which makes clathrates more stable than the pure guest liquids is that n(Hp - Ha) is often considerably less than AHva,. We return therefore to the notion stressed by other authors that the metastable host lattice must have a structure which although open is particularly stable usually because of hydrogen bonding and that this is most readily attained when the cavities are small. A second factor which is apparent in much of the data is that A Hop > 2AHvap probably because z the co-ordination number is larger in the clathrates.As a final remark on enthalpies it should be noted that for the methyl cyanide-quinol clathrate AHop is no larger than AHvap (Table 1). This unusually small AHog is undoubtedly related to the tight fit of the methyl cyanide molecule in the /3-quinol cavities as mentioned earlier. Comparison of AS” and dSop with d S v a p of the Pure Liquid Guest.- According to Table 1 AS” of the quinol clathrates is close to d&ap of the pure liquid guests (18-20 e.u./mole.) This fact of course means that the extra stabilities of these clathrates relative to the pure liquid guests are not related to entropy considerations. One would expect ASoB to be approxi- 2o D. H. Everett J. 1960,2566. 21 J. D. Bernal paper presented at the 1964 Anniversary Meeting of the Chemical Society at Birmingham.334 QUARTERLY REVIEWS mately equal to A Svap because the theoretical entropy of vaporisation of a liquid with spherical molecules is given by an equation20 similar to equa- tion (2) and the free volumes of the two systems are the same order of magnitude (-1 ~m.~/rnole). It happens that for the quinol clathrates the term [-n(S - S,)] which must be added to to give AS", is small. The values of AS" for the gas hydrates are significantly larger than ASvap of the pure guest liquids. Therefore these clathrates have lower entropies and would be less stable than the liquids if the entropy of the system were the only consideration. Only because AH" > AHvap do the clathrates have the greater stabilities. The reasons for the fairly large values of AS" cannot be discussed with precision because the values of Sb - S for these systems are known only approximately.One factor however is known with certainty and must be significantly responsible for the reduced stabilities of the gas hydrates relative to the quinol clath- rates. Since the fraction of the cavities occupied in the gas hydrates is found experimentally to be nearly loo% Sc = 0 while for the quinol clathrates Sc = 3.8 e.u. This difference is related ultimately to the fact that the empty p-type ice lattices are considerably less stable than ordinary ice (cf. a- and 8-quinol) and hence the degree of occupation of the former structures by guest molecules must be high in order that these structures be stabilised. Thus one finds in the enthalpy and entropy data for these two types of clathrate detailed explanations for the not unexpected conclusion that the clathrate with the least stable host lattice turns out to be the least stable clathrate.-06 -0.4 -0-2 0.0 0.2 0.4 06 log (cmJ/moIe) FIG. 4. AS" v. log V for the quinol clathrates (curve 1) and several structure II gas hydrates (curve 2). The lines have been given slopes of -2*3R valid in the square-well approximation. Within a group of clathrates in which S - S and Sc are constant for the series AS" varies approximately linearly with log Vf as predicted by equation (2) even when the approximate equation (3) is used for the calcu- CHILD MOLECULAR INTERACTIONS IN CLATHRATES 335 lation of the free volumes. This relationship is illustrated in Fig. 4 with data for the quinol clathrates and the structure I1 gas hydrates.Heat Capacities of Quinol Clathrates Deduction of Barriers to Rotation from Heat-capacity Data.-Although the assumption of spherically symmetric cavities does not introduce appreciable error into the theoretical prediction of clathrate stabilities several other kinds of experimental data demonstrate the existence of small potential barriers to the rotation of guest molecules showing that the force-field within the cavities is not strictly isotropic. Three kinds of experiment which bear on this question will be discussed heat capacity dielectric and infrared measurements. These and others have recently been discussed by S t a ~ e l e y . ~ ~ ~ ~ The effect on thermodynamic properties produced by a barrier to free internal rotation in molecules has been known for more than twenty years and in fact a discrepancy between the “Third Law” entropy of ethane and the entropy calculated from statistical thermodynamics with the assumption of free internal rotation first led Kemp and to postulate a potential barrier hindering internal rotation of about 3 kcal./ mole in ethane.Pitzer and Gwinn2j later developed quantitative relation- ships between barrier height and thermodynamic properties such as heat capacity. These relationships apply not only to internal rotations but also to any molecular rotation which experiences a resisting force and in the last few years workers at two l a b o r a t o r i e ~ ~ ~ ~ ~ ~ have calculated barriers to rotation in clathrates from heat capacity measurements over a range of temperatures. When RT V, the latter quantity being the barrier to rotation then the guest molecule undergoes a librational motion which is approximated very closely by the equation of motion for a harmonic oscillator.The contribution to the heat capacity G o t from the two-dimensional libra- tional motion of a diatomic molecule is therefore 2R. On the other hand when RT 9 Yo the molecule undergoes nearly free rotation and Grot becomes approximately R. The corresponding values of Grot for a poly- atomic guest molecule are 3R and 3/2R. In either case Grot starts out at very nearly zero for temperatures close to O’K because the librational motion is not excited rises to a maximum at higher temperatures as the librational motion is excited and then falls off slowly toward the limiting 22 L. A. K. Staveley in “Non-Stoichiometric Compounds” ed.L. Mandelcorn Academic Press Inc. New York 1964 ch. 10. 23 J. D. Kemp and K. S. Pitzer J. Chem. Phys. 1936 4 749. 24 J. D. Kemp and K. S. Pitzer J. Amer. Chem. SOC. 1937 59 276. 25 K. S. Pitzer and W. S. Gwinn J. Chem. Phys. 1942,10,428. 26 (a) N. G. Parsonage and L. A. K. Staveley Mol. Phys. 1959 2 212; (b) ibid. 1960,3,59; (c) N. R. Grey,N. G. Parsonage,and L. A. K. Staveley ibid. 1961 4 153; (d) N. R. Grey and L. A. K. Staveley ibid. 1963 7 83. 27 L. V. Coulter G. L. Stepakoff and C. G. Roper J. Phys. Chem. Solids 1963,24 711. 336 QUARTERLY REVIEWS value for the free rotator at still higher temperatures. From the exact shape of the curve for Crot against T one can evaluate the barrier to rotation. The problem then is to deduce Grot from the total measured heat capacity C of a clathrate and this is discussed in the next Section.Analysis of Heat-capacity Data for C1athrates.-The total heat capacity of the clathrate per three moles of quinol is first subdivided26 into contribu- tions from the P-quinol lattice and the guest molecule as expressed in the eq uatioii where CQ is the molar heat capacity of p-quinol X is the fraction of the cavities occupied and C is the molar heat capacity of the guest in the clathrate. Since there are three molecules of quinol per cavity Xis numeric- ally equal to the number of moles of guest per three moles of quinol. If both CQ and CG are truly independent of the degree of filling of the cavities then C at a given temperature should vary linearly with X as expressed by equation (5). Nearly all the experimental results for six quinol clathrates have verified the linear relationship,2s and C at a number of temperatures has been evaluated from the slopes of plots of C against X .Next Grot is obtained by subtracting Cvib the contribution to the heat capacity from the rattling of the guest molecule in the cage from CG. Cvib can be calculated from the theory of van der Waals and Platteeuw,12 which has been confirmed by the experimental results for argon- and kryp- ton-quinol clathrates,26ayb where the question of rotation does not arise. Finally the barrier to rotation of the guest molecule can be determined by comparison of the experimental Grot with Grot calculated from various barrier heights (see Fig. 5). It is evident that the accumulation of errors in the final result is rather unfavourable because Grot is only a small fraction of c.The barrier heights determined by Staveley for the quinol clathrates of methane oxygen nitrogen and carbon monoxide are 4 200-250 -1 100 and -1200 cal./mole respectively. In no case is the barrier height large enough to cause an appreciable discrepancy between the true ASop for these clathrates and the values predicted by equation (2) which is based on the assumption of free rotation. The virtually free rotation of methane in the clathrate at temperatures above -150"~ is reminiscent of the similar behaviour in solid methane slightly below the melting pointZs and is undoubtedly related to the high symmetry of this molecule. Finite barriers are found in the other clathrates however even though the guest molecules are non-polar or only slightly polar.It has been suggestedZBd that an interaction between the quadrupole moment of the guest molecule and the cavity wall which is electrically more negative in the regions of the two hydrogen-bonded hexagons than in the neighbourhood of the benzene rings could give rise to preferred orientations of the guest molecule in the ( 5 ) c=3cQ+xcG . . . . . . J. G. Aston Record Chem. Progr. 1959 20,41. CHILD MOLECULAR INTERACTIONS IN CLATHRATES 337 -II 100 200 3 0 0 T'K FIG. 5. Heat capacity-temperature plots for the nitrogen-quinol clathrate. Full circles CN2 = total heat capacity contribution per mole of N2 in cal. deg-I mole-I. Open circles calculated values of cUib. Half-shaded circles derived values of Grot. Dotted curves calculated values of Got for hindering energy barriers of 500 900 and 1500 cal.mole1. (Reproduced with permission from N. R. Grey and L. A. K. Staveley Mol. Phys. 1963 7 83). cavity. The observed barrier heights do in fact correlate qualitatively with quadrupole moments of the three guest molecules. One anomaly appears in the heat capacity studies CQ which is obtained as the intercept in a plot of C against X (see equation 5) is not constant for all clathrates at the same temperature. Some of the differences are outside experimental error particularly at the lowest temperatures. Furthermore the entropy of the guest at 298"~ obtained from the equation . . . . . (6) differs by 2-3 e.u. from the entropy calculated from the theory of van der Waals and Platteeuw12 for the argon methane and krypton clathrates. The difference is 1 e.u.larger for krypton than for argon. These discrepan- cies suggest that the low frequency vibrations of the quinol lattice may be perturbed by the guest For example the entropy contributions from one-dimensional lattice vibrations of 120 and 150 cm.-l differ by 0-43 e.u. at 298"~. Several other lines of evidence however argue against perturbations of appreciable magnitude. First the theory which neglects lattice distortion correctly predicts the dissociation pressures of a number of quinol clathrates. Secondly the experimentally determined fraction of the cavities occupied in both the argon and the krypton clathrates is 0.34 when the clathrate is in equilibrium with a-quinol and gaseous guest at 298 OK. If appreciable lattice perturbations occur in these two clathrates they would have to be of similar magnitudes for such a result to be ob- served.Perhaps careful comparison of the infrared spectra of /3-quinol and the clathrates would help to resolve this question. 338 QUARTERLY REVIEWS In a study of heat capacities similar to that described above Coulter Stepakoff and Roperz7 found barriers of 5 11 and 691 cal./mole restricting the rotation of N2 and CO respectively in the quinol clathrates. They also found that the data of Parsonage and Staveley26b for methane-quinol in the range 15-5O0K are consistent with a barrier of 193 cal./mole. These results seem to be in satisfactory agreement with those of Staveley when the unfavourable propagation of error is recalled. The barriers found for a number of guests in the quinol lattice are summarised in Table 2.Results of measurements other than heat capacity are included for completeness. TABLE 2. Barriers hiizdering rotation in quinol clathrates Guest Temp. Range (OK) 0 2 0.25-4 1 - 5 4 . 2 100-300 Nz 1.5-25 120-300 15-100 co 100-300 15-100 a34 150-300 15-50 Barrier height (cal. /mole.) 128 -250 -200 940 -1 100 51 1 -1200 69 1 -0 193 Method A B C D C C C c D C Ref. a b 26d 26d 27 26d 27 26b 27 C A Magnetic susceptibility; By Paramagnetic resonance; C Heat capacity; D Nuclear quadrupole resonance. aH. Meyer M. C. M. O’Brien and J. H. van Vleck Proc. Roy. SOC. 1957 A 243,414. %. Foner H. Meyer and W. H. Kleiner J . Phys. Chem. Solids 1961,18,273. W. Meyer and T. A. Scott ibid. 1959 11,215. Measurements of Dielectric Loss Results for Rotator-phase Solids and Solutions.-Typical dielectric data which show the loss region are given for solid D-camphor in Fig.6. At temperatures between -38” and its melting point 178” D-camphor is similar to a liquid in that its nearly spherical molecules experience very little resistance to rotation. For this reason the dielectric constant or permittivity E ‘ is nearly constant from 0 to -500 Mc./sec. showing that in this frequency range the dipoles are able to reorient in phase with the applied field. At higher frequencies E’ drops off because of the inability of the dipoles to rotate in phasz with the field and at sufficiently high frequencies E’ reaches a small limiting value arising from atomic and electronic induced dipoles all concerted rotation of permanent dipoles having ceased. In the “loss” region or region of decreasing E ’ a conduct- ance which is proportional to E” (the loss factor) appears.This existence of a high-frequency conductance in a material with a high D.C. resistance can be traced to the existence of a component of (A.C.) current in phase with CHILD MOLECULAR INTERACTIONS IN CLATHRATES 339 the applied e.m.f. when the rotation of the molecular dipoles lags behind this e.m.f. Outside the loss region the A.C. current produced by the in- phase orientations of the dipoles with the field turns out to be precisely 90" out-of-phase with the field and there is then no loss of energy or joule heating i.e. the medium is an ideal dielectric. FIG. 6. Dielectric data for solid D-camphor at -2O"c. (Reproduced with permission from C. Clemett and M. Davies Trans. Faraday Sue.1962,58,1718.) The particular frequencies at which an appreciable loss occurs are related to the relaxation time or spectrum of relaxation times for the system. In the simplest case the single relaxation time is defined by the first-order rate equation . . . . . (7) in which p stands for specific polarisation k the first-order rate constant and T = l/k the relaxation time. Equation (7) describes the manner in which the specific polarisation of the sample decays with time after the instantaneous removal of a static electric field. The relaxation time 7 is the time required for p to fall to l/e of its original value. In the more com- plicated physical situation consisting of the application of an alternating field to the sample dielectric theory shows that the maximum value of E" occurs at an angular frequency w = 2nf equal to 1 / ~ = k.Inspection of Fig. 6 shows that for D-camphor at -20" E ' ' ~ ~ ~ . occurs at -3700 Mc./sec. which implies a relaxation time of 4.3 x 10-l1 sec. This macroscopic relaxation time is very nearly equal to the relaxation time for individual dipole reorientations. Since k is a first-order rate constant its temperature dependence (and that of 1 / ~ ) might be expected to follow the Arrhenius or Eyring equations. That this is approximately so has been verified by a variety of experiments and a large number of energies or heats of activation have been evaluated. From the values of T for D-camphor at a number of temperatures AH of the Eyring equation has been found to be 1.8 & 0.7 k ~ a l . / m o l e . ~ ~ ~ Both the small AH and the small T point to the existence of fairly free 340 QUARTERLY REVIEWS rotation in solid D-camphor down to -38"c.For a number of other rotator-phase solids activation enthalpies of 1-2 kcal./mole have also been found. ga 9 b Dielectric studies of polar solutes in non-polar solvents have shown behaviour similar to that of the rotator-phase solids. Typical values of T and AH are 10-12-10-11 sec. at room temperature and 1-2 kcal./mole. When the polar solute molecules are approximately spherical in shape the rotations in solution may be even freer. For example AH is 1.5 and 0.3 kcal./mole for camphor and fluorobenzene respectively in carbon tetra- chloride Apparently the fluorobenzene molecule rotates quite freely about the axis perpendicular to the plane of the molecule. Quinol C1athrates.-In view of their structures and thermodynamic properties one would expect to find for clathrates the very short relaxation times and small enthalpies of activation characteristic of rotator-phase solids and solutions.This expectation is borne out by the available data although there is at least one interesting exception in the methyl cyanide- quinol clathrate. The most significant published dielectric study of quinol clathrates is that of Dryden and Meakin~,~lQ who measured the permittivities and losses of the methanol and methyl cyanide clathrates over a range of frequency and temperature. In addition measurements of permittivity at one low frequency (50 kc./sec.) and room temperature were made for the H2S HCN SOz and CzH2 clathrates. These measurements showed that at 50 kc./sec.the permittivity values are larger than E' of quinol the differences being approximately proportional to p2 of the guest molecule. Hence the loss regions must occur at frequencies higher than 50 kc./sec. and the barriers to rotation in these clathrates are probably small (< 5 kcal ./mole). In the more extensive study31b of the methanol clathrate however an absorption peak which had a shape corresponding to a single relaxation time was revealed at -loll c./sec. at room temperature. The existence of a single relaxation time at a given temperature means that each methanol molecule has the same environment and that co-operative motions involv- ing several methanol molecules are not significant in the reorientations of these molecules in an oscillating electric field. Neither conclusion is surprising in view of the structure of the clathrate.An Arrhenius activation energy of 2.3 kcal./mole was calculated from the temperature dependence of the relaxation time. For comparison with the activation energies quoted in the preceding section this value should be reduced somewhat as the Eyring activation energy is smaller than the Arrhenius value by the factor RT. Thus it is apparent that the barrier hindering the rotation of the (a) C. Clemett and M. Davies Trans Faraday SOC. 1962,58,1705; (b) ibid. p. 1718. Calculated from data given by W. E. Vaughan W. P. Purcell and C. P. Smyth 31 (a) J. S. Dryden and R. J. Meakins Nature 1952 169 324; (b) J. S. Dryden J. Amer. Chem. SOC. 1961 83 571. Trans. Faraday SOC. 1953,49,1333. CHILD hfIOLECULAR INTERACTIONS IN CLATHRATES 34 1 methanol molecules in the quinol clathrate is about the same as the barriers found in rotator-phase solids and in solutions.Measurements with single crystals of the clathrate and a detailed consideration of the values of (EL. - EL) (Fig. 6) at different temperatures indicated a lack of preferred orientations for the methanol molecule in the cage contrary to the finding of Palin and Powell,lc whose X-ray diffraction measurements suggested that on the average the (C-0) axis of the methanol molecule points towards the hexagons of the hydrogen bonded (0-H) groups in the cage. The activation energy deduced from the dielectric absorption must probably be regarded as an average value for the overall rotation of the molecule within its cage. Owing to its small value and the inevitable uncertainties in it unless the hindering potential were distinctly aniso- tropic it is not surprising that the dielectric study failed to show preferred directions for rotation.An appropriately marked contrast is provided by the data for the methyl cyanide-quinol lathr rate.^^^ Dielectric loss occurs at much lower fre- quencies (3 kc./sec. at room temperature) and the activation energy is 18 kcal./mole. Furthermore anisotropic properties are revealed by experiments with single crystals of the clathrate. When the field is parallel to the c axis (the axis perpendicular to the hydroxyl hexagons) a region of strong dielectric loss appears and the permittivity is 27 at 50 c./s. When the field is perpendicular to the c axis there is no absorption peak and the permittivity is only 3.Both observations are consistent with the hypothesis that the equilibrium orientation for the methyl cyanide molecule is along the c axis of the crystal owing to the CSV symmetry of the molecule the dipole moment is precisely along the (C-C-N) line. Introducing a field parallel to this axis causes the molecules to rotate through 180" by passing over a high barrier reversing the whole of the dipole moment while a perpendicular field encounters at most a very small component of the dipole moment. Accordingly to such a field the crystal behaves like a non-polar medium. All these data fall in neatly with the observations of lattice distortion and a relatively small AH" for this clathrate. Gas Hydrates.-The relaxation of the water molecules in the host lattice has been studied by Wilson and DavidsonS2 for acetone hydrate and Brey and LeggS3 for trichlorofluoromethane hydrate.Both systems exhibit loss regions which are broader than would correspond to a single relaxation time in contrast with the single relaxation time found for ordinary ice.34 If the relaxation mechanism in both ice and the clathrates is assumed to be the diffusion of rotational defects through the lattice then the existence of several diffusion paths in the gas hydrate lattices is implied. The activation energy for dipole reorientation was found to be smaller for the gas hydrates than for ice (6-3 and 9-2 versus 13.2 kcal./mole.). This lower hindrance to 32 G. J. Wilson and D. W. Davidson Canad. J. Chem. 1963,41,264. 33 W. S. Brey jun, and J. W. Legg J. Phys. Chem. 1963,67,1737. 34 R.P. Auty and R. H. Cole J. Chem. Phys. 1952,20 1309. 342 QUARTERLY REVIEWS reorientation is consistent with slightly weaker hydrogen bonds in the lattices of the gas hydrates.32 Motions of the guest molecules themselves have recently been examined in ethylene oxide hydrate (Structure I) and tetrahydro furan hydrate (Struc- ture 11) by Davidson Davies and Williams.35 Very short relaxation times ( ~ 1 0 - l ~ sec.) were found even down to liquid-nitrogen temperature and the estimated (Eyring) enthalpies of activation are 0.5 & 0.1 and 0.3 0.1 kcal./mole for ethylene oxide and tetrahydrofuran hydrates respectively. These are extraordinarily small values for polar (oxygenated) molecules in an ice lattice. It appears that departures from spherical symmetry in the cavities are very slight indeed and that atoms in the wall of a cavity need not be pushed out of the way when the guest molecule rotates.Striking confirmation of this remarkable freedom is found in the permittivity values ( E ' ) of these ice-hydrates the dielectric increment [ E' (clathrate) - E' (ice)] measured even at 3 x lo8 c./sec. and -180"~ shows the guest molecules to be rotating as completely as in benzene solutions at room temperatures. The dipole moments deduced from the dielectric increments are practically identical despite the great difference in the media. Infrared Spectra Molecular Rotation revealed by Solution-phase Spectra.-The large widths and peculiar contours of the vibration-rotational bands observed in the spectra of small molecules in the gas phase are familiar aspects of infrared spectroscopy.Less well understood are the often appreciable band widths and occasional multiple peaks found in the vibrational spectra of molecules in solution. If all the solute molecules were unable to rotate and had identical environments which did not vary with time their infrared spectra would consist of sharp bands whose finite widths would be determined mainly by instrumental parameters because rotational transi- tions would be absent and the set of force constants would be the same for all molecules. Obviously this simple situation does not exist in solution and the separation of all the factors influencing band shape has proved to be a difficult problem. Nevertheless a substantial body of evidence recently reviewed by Jones and S h e ~ p a r d ~ ~ shows that when at least one moment of inertia of the solute molecule is small some of the absorption bands have wings attributable to vibration-rotation transitions.For example the perpendicular bands of the methyl halides in a variety of solvents have wings whose shapes approach the contours of the P and R branches of the gas-phase spectra. The half-widths of these bands are 20-50 cm.-l. There seems little doubt that at least some of the solute molecules are able to rotate freely and to undergo rotational transitions although the rotational energy levels are blurred by interactions with the solvent. 35 D. W. Davidson M. Davies and K. Williams J. Chem. Phys. 1964 40 3449. 36 W. J. Jones and N. Sheppard Spectroscopy Report Conf. Organic Hydrocarbon Research Group Inst. Petrol. London 1962 p.181. CHILD MOLECULAR INTERACTIONS IN CLATHRATES 343 Analysis of the shapes of infrared bands to give barrier heights for rota- tion in solution has been attempted in only a few studies and even in these interpretations differ. For example Bulanin and Orlova3' have interpreted the spectra of solutions of carbon monoxide in thirteen solvents at room temperature to indicate potential barriers of 300-1000 cal./mole hindering the rotation of the carbon monoxide molecule. From the spectrum of CO dissolved in liquid oxygen at 9 0 " ~ they concluded that nearly all rotational excitation is inhibited in this solution. On the other hand E ~ i n g ~ ~ deduced from the spectrum of liquid CO at 8 0 " ~ a barrier to rotation of -120 cal./mole. According to the Boltzmann law 55 % of the molecules would be rotating freely in this liquid.A very similar condition was found for CO dissolved in liquid argon or nitrogen. While the spectra of solute molecules in inactive solvents might be expected to have features in common with the spectra of the guest mole- cules in clathrates an even greater similarity can be anticipated between the spectra of clathrates and those of small molecules trapped in solid rare-gas matrices. The infrared spectra of dilute solid solutions of HCl in argon,39 CH and CD in argon krypton and xenon,40 and H,O in argon krypton and xenon41 do in fact have multiple peaks which have been attributed to free or slightly hindered rotation of the trapped mole- cules. In the case of H,O in rare-gas matrices fine structure similar to that found in the spectrum of water vapour has been observed.It is only for the smallest solute molecules however that evidence for rotation has been noted. The infrared spectra of all other systems of this kind consist of purely vibrational bands. It is noteworthy that the free diameter of a cavity in P-quinol is midway between the diameters of the cavities produced by the removal of rare gas atoms from the crystal lattices of krypton and xenon. The cavities in ice of Structures I and I1 are even larger than those in xenon. Thus one can pre- dict the probable occurrence of complex features in the infrared bands of at least some of the clathrates. Quinol C1athrates.-As is true for other physical properties most of the infrared spectra reported for clathrates are for the quinols. The earliest study was that of Hexter and G ~ l d f a r b ~ ~ who investigated the HCl H2S SO, and CO clathrates.The spectra of the first three are dominated by quinol absorptions so that only for the SO clathrate could a few weak bands of the guest be seen. On the other hand the spectrum of the CO clathrate shows peaks at 650 and 2340 cm.-l which are 2% and 0.4% 37 M. 0. Bulanin and N. D. Orlova Optics and Spectroscopy 1963,15 112. 38 G. E. Ewing J. Chem. Phys. 1962,37,2250. 39 L. J. Schoen D. E. Mann C. Knobler and D. White J. Chem. Phys. 1962 37 40 A Cabana G. B. Savitsky and D. F. Hornig J. Chem. Phys. 1963,39,2942. 41 R. L. Redington and D. E. Milligan J. Chem. Phys. 1963 39 1276 and references 42 R. M. Hexter and T. D. Goldfarb J. Znorg. NucIear Chem. 1957,4 171. 1146. therein. 344 QUARTERLY REVIEWS lower in frequency than the corresponding band centres in the spectrum of gaseous COB.Similar “solvent shifts” are found in solution spectra and reflect interactions between solute and solvent. The band at 2340 cm.-l has two maxima each of which has a half-width of about 5 cm.-l. Because of the narrowness of this band it was considered to be a Q branch; therefore the rotation of the CO molecule in the cage appears to be con- siderably hindered a finding consistent with the known distortion of the /3-quinol lattice by the CO molecule.ld It should be noted that only if the barrier hindering rotation is less than -2 kcal./mole will a significant fraction of the molecules have rotational energies lying above the barrier at room temperature. At lower temperatures this figure of course becomes smaller.Therefore the absence of appreciable width or P and R contours in an infrared band only sets a fairly small lower limit to the barrier. A detailed study of the CO-quinol clathrate over a range of tempera- tures was made by Ball and M ~ M e a n . ~ ~ This clathrate is particularly suit- able for the study of rotational motion in the infrared because the CO molecule has a small moment of inertia and its vibrational frequency occurs in a part of the spectrum where quinol absorbs only weakly. The band due to CO is shown in Fig. 7 for the clathrate in an NaBr disc at several cm .-’ FIG. 7. Infrared spectrum of CO-quinol clathrate at 20° -55” and - 130”~. (Repro- duced with permission from D. F. Ball and D. C. McKean Spectrochim. Acta 1962 18 933.) temperatures. An NaBr disc containing the krypton or nitrogen clathrates was used in the reference beam of the spectrometer to cancel out the quinol absorption.The appearance of a strong Q branch at 2133 cm.-l demonstrates that a considerable fraction of the molecules are in rotational energy states lying below the barrier for if there were free rotation the Q branch would be forbidden. Wings of weaker intensity displaced about 40 cm.-l from the Q branch are also evident. Although the entire band is similar in some ways to the spectra of CO in solution the half-widths of the Q branches 43 D. F. Ball and D. C. McKean Spectrochim. Ada 1962,18,933. CHILD MOLECULAR INTERACTIONS IN CLATHRATES 345 are significantly different in the two cases. They are 15-20 and -30 c1n-l for clathrate and hydrocarbon and halogenated hydrocarbon respectively at room temperatures.Similarly at 1 4 0 " ~ the half-width of the clathrate Q branch is -5 cm.-l while at 80-90"~ half- widths are 22 18 26 and -30 cm.-l in the pure liquid liquid argon liquid nitrogen and liquid oxygen r e s p e c t i ~ e l y . ~ ~ ~ ~ ~ Whether these differ- ences imply greater rotational freedom for CO in solution or whether the broadening of the Q branch in solution should be attributed to a diffusion of the rotational process as claimed by Bulanin and Orl~va,~' is a question which cannot be answered at present. A decision regarding the source(s) of the several satellites appearing in the clathrate spectra is made difficult by the existence of a number of possibilities ificluding combination bands with rattling and librational frequencies and P and X branches arising from vibration-rotational transitions between states above the barrier.Ball and McKean ascribe the high-frequency satellites appearing at low temperature to rattling and librational combinations ; the high temperature bands on either side of the Q branch are likely to be P and R branches although the other possibili- ties cannot be ruled out. By use of the P-R separation in a calculation similar to that employed by Bulanin and Orlova the height of the barrier hindering rotation was estiniated to be 720 cal./mole in reasonable agree- ment with other determinations of this barrier height (see Table 2). A solvent shift of 0.5% to lower frequency was observed for the clath- rate. This is larger than most of the solvent shifts found for CO in solution and may be related to a possible quadrupole interaction in the clathrate as discussed by Staveley.*6d Quinol clathrates with HCI DCI SO, CH,-OH CH,.CN HC02H CH3F and CH2C1 as guests and eight clathrates of Dianin's compound have recently been in~estigated.~~ In general very few manifestations of the rotational motions of the guest molecules can be seen in the infrared spectra.Several features of the spectrum of the methanol clathrate are interesting i n light of the dielectric evidence already obtained for this clathrate."lb A band at 3520 ciii.-l is clearly evident and belongs to the 0-H stretching mode of inonomcric CH3.0H. Comparison of the ratio of the heights ofthis peak aid a methanol absorption at 1024 cm.-l with the corresponding ratio for monomeric methanol in solution shows that probably all the CH,.OH molecules in the clathrate are in a non-hydrogen- bonded condition even though the cavity presents twelve oxygen atoms each of which uses only one lone pair in hydrogen bonding.The (0-H) monomer stretching absorption of methanol is shifted downward from vgas by 1-8 % in the clathrate and 1.3 % in carbon tetra- chloride solution; the respective half-widths are 41 and -24 cm.-l. Since 0-H stretching frequencies are known to be particularly sensitive to changes in environment the appreciable solvent shift in the clathrate is not 44 M. Davies and W. C. Child jm. to be published. 346 QUARTERLY REVIEWS surprising and is indicative of interactions between the 0-H dipole and the atoms in the wall of the cavity. At least two causes for the considerably larger half-width in the clathrate band can be found.A greater freedom of rotation of the CH,.OH molecule in the clathrate is consistent with this observation. A second possibility relates to the variation in the force field for the molecule as it undergoes librational and rattling motions. Such motions produce varying perturbations on the 0-H bond and a spread of frequencies in the observed vOpH. This kind of explanation is supported by the findings from dielectric measurements that the activation energy for rotation is over 2 kcal./mole and that there are probably a number of equilibrium positions for the CH,.OH molecule. The spectra of some samples of CH,F-quinol are anomalous in that several weak bands which do not belong either to methyl fluoride or to a- or P-quinol are evident.As the fraction of cavities occupied decreases the intensities of these bands increase. Partially occupied CH,.OH-quinol exhibits a similar but less pronounced behaviour. The most probable explanation is that these samples suffer lattice distortions or have a crystal form different from that of P-quinol. A similar observation was made by Hexter and G01dfarb~~ for the C0,-quinol clathrate. Other C1athrates.-Only a few additional spectra have been reported and in these instances the guest spectra are similar to the spectra of the guest in the liquid phase. Aynsley Campbell and D ~ d d ~ ~ and Drago Kwon and have studied some clathrates of the type Ni(CN),NH,,M where M the guest is benzene aniline pyrrole or thiophen. Hart and Smith4' have noted the spectra of the p-xylene and p-dichlorobenzene clathrates of tetra-(4-methyl pyridine)nickel(II) thiocyanate.Mandelcorn Goldberg and H ~ f f ~ ~ observed one band of sulphur hexafluoride in the spectrum of the clathrate of this compound and Dianin's compound. In conclusion it appears that a complete interpretation of the infrared spectra of clathrates is not yet possible even though the forces acting on the guest molecules are more easily predicted and analysed than are the forces operating in the liquid phase. It is to be hoped that an elucidation of the spectra of clathrates will help to disentangle the various factors affecting spectra in solution. The author is grateful to Dr. Manse1 Davies for his many helpful suggestions 45 E. E. Aynsley W. A. Campbell and R. E. Dodd Proc. Chem. SOC. 1957,210. 46 R. S. Drago J. T. Kwon and R. D. Archer J. Amev. Chem. SOC. 1958 80 2667. 47 M. I. Hart jun. and N. 0. Smith J. Amer. Chem. SOC. 1962 84 1816. 48 L. Mandelcorn N. N. Goldberg and R. E. Hoff J . Atner. Chem. SOC. 1960 82 regarding this Review. 3297.

ISSN:0009-2681    

DOI:10.1039/QR9641800321

出版商:RSC    

年代:1964

数据来源: RSC

摘要:

QUINONE METHIDES By A. B. TURNER (UNIVERSITY OF ABERDEEN) 1. Introduction THE reactive intermediates of organic chemistry have gained credence partly because many of them exist in stable forms having the chemical properties required of the hypothetical species. This situation is well exemplified in the case of quinone methides which on paper are derived from quinones by replacement of one of the carbonyl oxygens by a methylene or substituted methylene group. These compounds are also known as methylenequinones or quinone methines. Increased interest in their chemistry during the last few years has arisen largely from their probable involvement in biochemical processes notably oxidative phos- phorylation. 2. Stable Quinone Methides This section is concerned only with para-quinone methides as no members of the ortho-series have yet been obtained in a pure state.The chemistry of most of the natural products mentioned here has been covered in earlier reviews.l Quinone methides occur in Nature both as fungal metabolites familiar as a rich source of quinones themselves and as wood pigments. Thus various species of Penicilliurn elaborate citrinin ( 1),2 pulvilloric acid (2),3 and purpurogenone (3),4 while fuscin (4) is obtained from cultures of Oidiodendron fuscum R ~ b a k . ~ Structural work on these compounds was facilitated by the ease with which they undergo reversible addition reac- tions giving colourless phenolic products. Pulvilloric acid for example can be crystallised as a colourless ethanol adduct which rapidly reverts to the yellow acid on keeping. It also forms a colourless adduct with sodium bisulphite.This behaviour is typical of quinone methides in general (see Section 4). R. G. Cooke and R. H. Thomson Rev. Pure Appl. Chenz. (Australia) 1958 8 8 5 ; J. P. Brown A. Robertson W. B. Whalley and N. J. Cartwright J . 1949 867. J. F. W. McOmie M. S. Tute A. B. Turner and B. K. Bullimore Chem. andtnd. J. C. Roberts and C . W. H. Warren J. 1955 2992. D. H. R. Barton and J. B. Hendrickson J. 1956 1028. W. B. Whalley “Progress in Organic Chemistry” ed. J. W. Cook 1958 4 72. 1963 1689. 347 348 QUARTERLY REVIEWS Brazilein (5; R = H) and hamatoxylein (5; R = OH) are the dyeing principles of redwood and logwood respectively. The orange wood-pigment pristiineriii ( 6 ) 7 3 8 is the only member of the group which is prone to rearrange under the influence of acid.As might be expected from its structural resemblance to steroidal dienones these skeletal rearrangements are of the dienone-phenol type and involve initial migration of a methyl group to the terminus of the quinone methide system.s In addition there are the anhydro-base forms of the anthocyanins which are responsible for some of the more brilliant colours of living plants. These pigments are all based upon the C, unit (7) having the carbon skele- ton of the flavones and have been reviewed by Se~hadri.~ A feature common to all of these natural products is the isolation of the quinone methide chromophore from labile hydrogen atoms which prevents tautomeric rearrangement of the colouring matter to a phenol.1° This is achieved in most cases by the situation of a cyclic ethereal oxygen atom at the end of the conjugated system remote from the di-unsaturated carbonyl group.In the remaining compounds carbon atoms terminating the chromophoric system do not bear hydrogen. These observations were first made by Grant and Johnson1* in aid of their elucidation of the nature of the chromophore of pristimerin (6). An illustration of the instability of the quinone methide structure in rigid ring systems when not isolated from labile hydrogen atoms has been found in the steroid series. Elimination of R. Robinson Bull. Soc. chim. France 1958 125. R. Harada H. Kakisawa S. Kobayashi M. Musya K. Nakanishi and Y . Taka- A. W. Johnson P. F. Juby T. J. King and S . W. Tam J. 1963,2884. T. R. Seshadri “Festschrift A. Stoll” Birkhauser Bask 1957 p. 318. hashi Tetrahedron Letters 1962 603.l o P. K. Grant and A. W. Johnson J. 1957 4079. TURNER QUINONE METHIDES 349 hydrogen chloride from 10/3-chlorodien-3-0nes (8) by means of calcium carbonate in refluxing dimethylformamide gives A 9(11)-phen~l~ ( 10).l1 The loss of a severe 1,3-diaxial interaction (1 1 P-hydrogen and 13p-methyl) in going to the product (10) no doubt contributes to the ease of aromatisa- tion of the transient intermediate (9). Simple quinone methides in which the terminus of the chromophore is linked to carbon atoms bearing hydrogen have in fact been isolated.12 Oxidation of the phenols (1 1 ; R = Me or Et) gives the crystalline products (12; R = Me) and (12; R = Et). These two compounds show no tendency towards spontaneous aromatisation. They react with hydrogen bromide and alcohols by 1,6-addition to give the corresponding phenols.Me$/ CMe Me3C I I We 6 Q 1% (12) Me R 10-Methylenearithrone (13) is the sole example of a stable quinone methide having an unsubstituted methylene group.13 The low probability of the central ring’s becoming aromatic makes its quinone methide character negligible. The comparative isolation of the ethylenic bond from the carbonyl group in this molecule is shown by its typically olefinic be- haviour in adding a molecule of bromine to give the dibromide (14). Other monomeric quinone methides lacking tautomeric possibilities include fuchsone (15; R = Ph),14 which is attacked by concentrated alkali giving the carbinol(l6; R = Ph X = OH). The related photoadduct (1 5 ; R = COPh) of benzoquinone and diphenylacetylene reacts similarly with hot alkali but the resulting ketol (16; R = COPh X = OH) is degraded under these conditions.15 Catalytic hydrogenation gives the dihydro-derivative (16; R = COPh X = H).l1 J. S. Mills T. Barrera E. Olivares and H. Garcia J. Amer. Chem. SOC. 1960 82 l2 C. D. Cook and B. E. Norcross J. Amer. Chem. SOC. 1956,78,3797. l3 K. H. Meyer Annalen 1920 420 135. l4 A. Bistrzycki and C. Herbst Ber. 1903 36 2335. l5 H. E. Zimmerman and L. Craft Tetrahedron Letters 1964 2131 ; D. Bryce-Smith 5882. G . I. Fray and A. Gilbert Tetrahedron Letters 1964 2137. 350 QUARTERLY REVIEWS cfj B()Br :& / C1 /c / \ (16) (17) (1 8) I%' 'R P h i R PhC=CPh 05) Further examples are the diphenoquinocyclopropene (1 7),16 and the orange steroid (1 S) which results from chlorination followed by de- hydrochlorination of ce~tr0ne.l~ 3.Reactive Quinone Methides (a) Simple Quinone Methides.-Simple members of both the ortho and para series are very unstable molecules which polymerise spontaneously. Some are nevertheless sufficiently stable in dilute solution or at low tem- peratures to allow studies of their properties. They are generally prepared either by oxidation of the corresponding cresols or by elimination of the appropriate elements from ortlzo- and para-hydroxybenzyl derivatives. The parent compound of the ortho series o-benzoquinone methide (20) is obtained1* as a solid at liquid-nitrogen temperature by pyrolysis of o-methoxymethylphenol (19). On warming to -50" the pyrolysate liquefies and appears quire stable. Although spectral data are lacking the chemical properties of the substance accord with its formulation as the monomer (20).Three reactions define the character of the product (Chart I). Reduction by metal hydride yields o-cresol ; addition of methylmagnesium iodide followed by hydrolysis gives o-ethylphenol ; and the colourless solid trimer (21) is formed on warming to room temperature. Analogous trimers are obtained from ring substituted derivatives of o-benzoquinone methide.lg Attempts to prepare p-benzoquinone methide by pyrolysis of p-methoxy- methylphenol were unsuccessful this compound being stable up to 900". A. S. Kende J. Amer. Chem. Soc. 1963 85 1882. l7 E. Schwenk C. G. Castle and E. Joachim J. Org. Chem. 1963,28 136. l8 S. B. Cavitt H. Sarrafizadeh R. and P. D. Gardner J. Org. Chem. 1962,27,1211. l9 A. Merijan B. A.Shoulders and P. D. Gardner J. Org. Chem. 1963,28,2148. TURNER QUINONE METHIDES 35 1 p-Hydroxymethylphenols are likewise more resistant to dehydration than their ortho isomers.2o These differences may be clue to the six-membered transition states available for the fragmentation of the ortho compounds. Although the parent compound of the para series has not been isolated Filar and WinsteinZ1 have studied its 2,6-dimethyl derivative (23 ; R = Me). This can be prepared either from the phenol (22; R = Me) by treatment with a base such as triethylamine or by oxidation of mesitol(25; R = Me) with silver oxide. In the latter reaction it is interesting that very little if any of the ortho isomer is formed. OH R / R Qc Et,N ___c - HCLl MoOH R 6 CH2 (23) LiAlH4 0 0 Me \ (24) Me (25) R The identity of the quinone methide (23; R = Me) is clear from its ultra- violet and infrared spectra.It is stable for days at room temperature in inert solvents at high dilution ( 10-5~) but its disappearance becomes quite rapid as the concentration is increased. The di-t-butyl analogue (23 ; R = But) is rather more stable in solution but again cannot be obtained pure. In both cases the reaction products are primarily dimeric (see p. 353). These para-quinone methides are attacked at the terminal methylene group by a variety of nucleophilic reagents yielding the appropriate benzyl derivative (e.g. 26; R’ = Me or COMe) by 1,6-addition. Attempts to prepare quinone methides of the naphthalene and phenan- threne series have met only with the formation of dimers. Pummerer and C1ierbuliez22 have shown that dehydrogenation of 1-methyl-2-naphthol (29) leads to the dimer (27).This compound when heated in xylene disproportionates to the parent phenol (29) and the quinone methide (28) which in turn gives the stable dimer (30). That the original dimer (27) does indeed give the intermediate (28) is substantiated by the work of Smith and H ~ r n e r ~ ~ who found that the dihydrocoumarin (31) is formed when the dimer (27) is warmed with sodium malonate in dry ethanol (Chart 111). Condensation of phenanthr-9-01 with formaldehyde and dimethylamine under mild conditions24 gives the unstable Mannich base (32; X = NMe,). 2o N. J. L. Megson “Phenolic Resin Chemistry,” Butterworths 1958 p. 165. 21 L. J. Filar and S . Winstein Tetrahedron Letters 1960 25 9. 32 R. Pummerer and E.Cherbuliez Ber. 1919 52 1392. 23 L. I. Smith and J. W. Horner J . Amer. Chem. SOC. 1938,60,676. 24 P. D. Gardner and H. Sarrafizadeh R. J. Org. Chem. 1960 25 641. 352 QUARTERLY REVIEWS (27) - $02Et -CH (C02 Et) Loss of nitrogen occurs during attempts to purify it and the high-melting dimer (34) is obtained. This material is also obtained from the reaction of phenanthr-9-01 with formaldehyde alone indicating similar instability of the 10-hydroxymethyl derivative (32; X = OH). The reactivity of 10- methylene-9-phenanthrone (33) contrasts with the stability of 10-methyl- eneanthrone (1 3). r 1 __t (33) (b) Quinone Methides as Chemical Intermediates.-In addition to the cases in which quinone methides have been isolated there are many reactions in which they are probably involved as transient intermediates.It is well known that ortho- and para-hydroxybenzyl derivatives readily undergo nucleophilic substitution at the benzylic position. Thus displace- ment of the amine moiety of Mannich base methiodides by nucleophils such as methoxide cyanide and hydride ions is facilitated by an anionic oxygen in the ortho position. Gardner and his c o - ~ o r k e r s ~ ~ conclude that although the reaction is bimolecular the transition state closely resembles the quinone methide structure in its charge-separated state 26 P. D. Gardner H. Sarrafizadeh R. and L. Rand J. Amer. Chem. Soc. 1959 81 3364. TURNER QUINONE METHIDES 353 Similarly the intermediate (23) is involved in the hydrolysis of the chloride (22) and in its reaction with sodium acetate in acetic acid to produce the corresponding benzyl acetate (26; R = COMe).A further example is the ready replacement of the substituted amino-group of the benzylamine (35) by active methylene compounds under basic conditions.26 Side-chain C-alkylation of this type of phenol is thereby achieved by nucleophilic attack on the intermediate (28) in contrast to the more familiar mechanism of ring C-alkylation of phenoxide anions. Earlier work23 (described on p. 351) in which the quinone methide (28) generated by other means was shown to react with the malonate anion in the same way substantiates this reaction path. The relatively stable phenoxyl radical (24; R = CH (COPh& &w But) formed by one- electron oxidation of the hindered phenol (25; R = But) disproportion- ates to the parent phenol and the corresponding quinone methide (23; R = B ~ t ) .~ ~ 9 ~ ~ The initial radical decay is bimolecular and the quinone methide reacts further through the formation of free-radical intermediates yielding equal amounts of the biphenylethane (36) and the stilbenequinone n (37). M e C O CMa Y The suggestion that quinones having nuclear alkyl substituents might react in a tautomeric quinone methide form was originally made by Fuson as an extension of the principle of ~ i n y l o g y . ~ ~ On this basis a reasonable mechanism could be written for the mysterious condensation30 of malonic ester with duroquinone (38) to give the coumarin (41; R = C02Et) involving initial attack of malonate ion on the tautomer (39). The enolate 26 H. Hellmann and J. L. W. Pohlmann Annalen 1961 642,28.27 C. D. Cook and B. E. Norcross J. Arner. Chem. SOC. 1959 81 1176. 28 R. H. Bauer and G. M. Coppinger Tetrahedron 1963 19 1201. 28 R. C. Fuson Chem. Rev. 1935 16 1. 30 L. I. Smith and F. J. Dobrovolny J. Arner. Chem. SOC. 1926,48 1693. L 354 QUARTERLY REVIEWS of methyl cyanoacetate adds similarly giving the coumarin (41 ; R = CN) but acetylacetone does not condense.31 Thus cyclisation to the dihydro- coumarin (40) may be necessary to complete the reaction. The final dehydrogenation is brought about by the quinone. The active form (39) is also responsible for the base-catalysed dimerisa- tion of duroquinone (38),32 biduroquinone (42) being formed by Diels-Alder addition of the tautomer (39) to the quinone (38). 2,3-Diniethylnaphtha- quinone gives the same type of dimer under similar condition^.^^ It has recently been found34 that fully alkylated quinones readily under- v 20 amination involving this same tautomeric form at carbon atoms adjacent to the nucleus.In the case of duroquinone reaction with piperi- dine takes place at rooni temperature to give the bis-piperidino-derivative (43). The reaction proceeds by addition of piperidine to the tautomer (39) followed by oxidation of the resulting aminated quinol. Repetition of the process with the new basic centre controlling the direction of the (43) second tautomerisation gives the observed pro- duct which is stabilised by hydrogen bonding. This new type of amination is analogous to the well known nuclear amination of unsubstituted quinones. (c) Quinsne Methides as Biochemical Intermediates.-Quinone methides are implicated as intermediates in a number of biochemical processes 31 L.I. Smith and E. W. Kaiser J. Amer. Client. SOC. 1940 62 138. 32 L. I. Smith R. W. H. Tess and G. E. Ullyot J . Amer. Chem. Suc. 1944 66 1320. 33 K. Chandrasman and I?. H. Thomson unpublished results. 34 D. W. Cameron P. M. Scott and (Lord) Todd J. 1964 42. TURNER QUINONE METHIDES 355 in which they are thought to arise either by tautomeric rearrangement of quinones or by oxidation of phenols. Three mechanisms may be considered for the oxidation of ortho- and para-alkylphenols Path B - U f H Chemical evidence indicating that plienoxy-radical formation is the initial step in the majority of phenol points to one electron oxidation (Path A ) as the most likely pathway. If an ionic process is in- volved then in biological systems it could well be hydride abstraction from carbon (Path C).This has the advantage of yielding the quinone methide directly perhaps by a concerted reaction and the transformation could be brought about enzymically through the agency of pyridine nucleo- tide or quinonoid coenzymes. Recent indicates that enols are sub- ject to two-electron oxidation. Thus in both chemical and biochemical dehydrogenation of steroidal ketones it is the enolic form of the substrate which undergoes hydride abstraction at the ,%carbon atom. In addition suitable enolic compounds are rapidly deh~drogenated~~ by high-potential quinones reagents thought to function by accepting hydride ions.38 In other enzymic dehydrogenations for examples those mediated by pyridine nucleotides the experimental evidence is markedly in favour of a mechan- ism involving direct transfer of a hydride (i) Oxidative phosphorylation.Many lines of evidence point to the participation of quinones in the vital processes of electron transport along the respiratory chain and in the phosphorylations which accompany oxidation. Detailed schemes for the reaction cycles operating during oxida- 36 H. Musso Angew. Chem. Internat. Ed. 1963 2 723. 36 H. J. Ringold M. Hayano and V. Stefanovic J. Biol. Chem. 1963 238 1960. 37 S. K. Pradhan and H. J. Ringold J. Org. Chem. 1964 29 601. 38 L. M. Jackman “Advances in Organic Chemistry” ed. R. A. Raphx! E. C. Taylor 39 B. Pullman and A. Pullman “Quantum Biochemistry” Interscience 1963 p. 522. and H. Wynberg 1960,2 329. 356 QUARTERLY REVIEWS tive phosphorylation have been suggested by Vilkas and Ledereq4O taking advantage of structural features common to the tocopherol Vitamin K and ubiquinone groups.Quinones of these series have a methyl group at position 2 and a 3-substituent capable of participating in chroman ring formation allowing rearrangement to the quinone methide (partial structure 45). Nucleophilic attack by phosphate anion at the terminal methylene group followed by intramolecular migration of the phosphate group to the phenolic oxygen by way of the cyclic phosphate (47) complete the formation of the active quinol phosphates (48). The ease of the transformation (44)+(45) is apparent from the ready formation of ortho-quinone methide-type dimers in both groups of ita am ins.^^-^^ In addition the quinone methides (45) of the Vitamin K and uniquinone series have been trapped by reaction with acetyl chloride and acetic anhydride,4z and The suggestion has also been made42 that 5-phosphomethyl-6-chromanols (46) are active phosphorylating species generating metaphosphate on oxidation.However there is as yet no direct evidence for the participation of quinone methides in oxidative phosphorylation. Ollis and S~therland~~ have suggested the following mechanism for the biosynthesis of the ubiquitous (ii) Biogenesis of 2,2-dimethylchromenes. 40 M. Vilkas and E. Lederer Experientia 1962,18,546. 41 D. McHale and J. Green Chem. and Ind. 1964 366. 42 R. E. Erickson A. F. Wagner and K. Folkers J. Amer. Chem. SOC. 1963,85,1535. 43 P. Mamont R. Azerad P. Cohen M. Vilkas and E. Lederer Cumpt. rend. 1963 44 W.D. Ollis and I. 0. Sutherland “Chemistry of Natural Phenolic Compounds” 257 706. Pergamon Press 1961 p. 84. TURNER QUINONE METHIDES 357 2,2-dimethylchromene system (5 1) from o-dimethylallylphenols (49) (5 1) (50) While support for the formation of the quinone methide (50) is found in the occurrence of derivatives of the parent phenol (49) oxygenated at the benzylic position these being presumed to arise by hydration of the inter- mediate (50) alternative processes may be considered for the oxidation of the phenol (49) as set out at the beginning of this section. Direct formation of the quinone methide (50) by hydride abstraction from the benzylic position appears particularly attractive in this case. The mechanism proposed for the biosynthesis of 2,2-dimethylchro- menes is analogous to that suggested for base-catalysed cyclisations of polyisoprenylated quinones such as the ubiquinones (52).45 MeO1 Pe I M e O v M e MeoQ MeO\ Me OH (iii) Lignin biosynthesis.The part played by quinone methides in the biosynthesis of lignin has been uncovered by the work of Fre~denberg.~~ 45 B. 0. Linn C. H. Shunk E. L. Wong and K. Folkers J. Amer. Chem. SOC. 1963 46 K . Freudenberg Fortschr. Chem. org. Naturstoffe 1962,20,41. 85 239. 358 QUARTERLY REVIEWS Both intermolecular and intramolecular additions of nucleophilic reagents to the quinonoid intermediates are involved. An illustration is the forma- tion of dehydrodiconiferyl alcohol (55) by one-electron oxidation of coniferyl alcohol (53). Dimerisation of the mesomeric free radical pro- duced yields the quinone methide (54) which cyclises as shown.(iH20H EH" OH (53) CH20H k 4. Reactions of Quinone Methides Quinone methides are much more reactive than vinyl ketones owing to the additional driving force of aromatisation which characterises all their reactions. These reactions can be divided into three main types along the lines laid down by Hultzsch.4' Many specific examples of these reaction types have already been mentioned. (a) Addition of Nucleophilic Reagents.-The most characteristic pro- perty of quinone methides both stable and unstable is susceptibility to attack by nucleophils at the terminal methylene group acHrz- co - aCH2* 0- Reagents involved in these processes include alcohols amines carbo- hydrates carboxylic acids alkylmagnesium halides hydrogen cyanide metal hydrides phenols thiourea water and active methylene com- pounds.The net result is 1,4- or 1,6-addition of one molecule of the reagent yielding a phenol. In the case of stable quinone methides the addition is readily reversible and this has been a useful guide in identifying the chro- mophoric system in the natural products. 47 I<. Hultzsch Angew. Chem. 1948 60 179. TURNER QUINONE METHIDES 359 (b) Polymerisation.-Steric factors largely determine the stability of these compounds towards polymerisation and simpler members of the group cannot be isolated in pure form when the terminal methylene group is unsubstituted. The ortho and para isomers give different products on polymerisation the former yielding cyclic ethers and the latter undergoing tail-to-tail dimerisation. Reactive members of the para series under- go a combined disproportionation and dimerisation forming equal amounts of a stilbenequinone and a dihydroxybiphenylethane (see p.353). They have also been to dimerise to dihydroxystilbenes of type (56). (i) para- Quinone methides. (ii) ortho- Quinone methides. Simple compounds of the ortho series give dimers (e.g. 30) or trimers (e.g. 21) by a type of Diels-Alder reaction in which one molecule adds across a double bond of another. etc. 0 The reactions lead specifically to one of a number of possible types of product as do dimerisations of ap-unsaturated ketones and aldehydes.49 The direction of addition can be rationalised on the basis of a stepwise rather than a synchronous mechanism with formation of the most stable radical intermediates,60 as indicated.(c) Addition of 0lefin.s.-Just as one molecule of a reactive quinone methide adds across a double bond of another so will it combine with an olefin to give a chroman derivative. Thus Hultzsch51 has shown that o- 48 H. Euler E. Adler and A. 0. Caspersson Arkiv. Kemi 1943,16A No. 11 1. 4 9 K. Alder H. Offermanns and E. Ruden Ber. 1941 74 926. 6 o J. D. Roberts and M. C. Caserio “Basic Principles of Organic Chemistry” 61 K. Hultzsch Ber. 1941 74 898; J. prakt. Chein. 1941 1§8,275. Benjamin 1964 p. 268. 360 QUARTERLY REVIEWS hydroxymethylphenols react in this way with unsaturated compounds at high temperatures. An example of this is the formation of 2-phenyl- chroman (58) from saligenin (57) and styrene (57 1 PR-CH=CH In the condensation of the phenolic alcohol (59) with oleic acid (60) both possible isomers (61; R = Me R = C02H) and (61; R = C02H R = Me) are ~ b t a i n e d ~ ~ owing to the similar stabilities of the radical intermediates.+ Reactions of this type are important industrially since they are involved in the Combination of drying oils with phenolic resins and perhaps also in vulcanisation. Wakselman and V i l k a ~ ~ ~ have shown that the high temperatures of these reactions are necessary only for the formation of the quinonoid species and not for their subsequent condensation with the olefins. Thus substi- tuted chronians can also be obtained by reaction of styrene or diphenyl- ethylene at room temperature with ortho-quinone methides generated in situ from chloromethylphenols in basic media or from hydroxymethyl- phenols by the action of acid. The author thanks Professor R. H. Thomson for his helpful comments on 52 G. R. Sprengling J. Amer. Chem. Soc.,1952,74,2937. 53 M. Wakselman and M. Vilkas Cumpt. rend. 1964,258 1526. the manuscript.

ISSN:0009-2681    

DOI:10.1039/QR9641800347

出版商:RSC    

年代:1964

数据来源: RSC

摘要:

INORGANIC NITRATES AND NITRATO-COMPOUNDS By B. 0. FIELD* and C . J. HARDY~ (ATOMIC ENERGY RESEARCH ESTABLISHMENT HARWELL BERKS.) I. Introduction MANY simple compounds containing the nitrate group have recently been made for the first time and some exhibit interesting and unexpected properties for example volatility and high reactivity. Nitrates are generally dealt with only briefly in textbooks; these often erroneously state that nitrates of some elements are unknown and fail to refer to important older work. This Review outlines the methods of preparation and the important properties of compounds containing the nitrate group but does not deal in general with the large number of co-ordination complexes of metal nitrates. Recent research has indicated that compounds of elements (M) and the nitrate group can be divided into two classes according to whether the M-NO bond is predominantly ionic or covalent.The main differences upon which this classification is based are summarised in Table 1 and emphasised where appropriate in the text. We propose to apply syste- matically the term nitrate to a compound in which the M-NO3 bond is ionic that is a compound containing the nitrate ion and to use the term nitrato-compound or the prefix nitrato to the element when the NO3 group is covalently bonded through one or more of its oxygen atoms. 2. Preparative Methods The four principal reagents which have been employed in the synthesis of inorganic nitrates and nitrato-compounds are aqueous nitric acid (HNO,) dinitrogen tetroxide (N,O,) dinitrogen pentoxide (N,Q,) and “chlorine nitrate” (CINO,).Which reagent is used and how it is applied in the preparation and isolation of a particular nitrate or nitrato-com- pound varies according to the physical properties and chemical reactivity of the product. Many hydrated metal nitrates can be isolated from aqueous nitric acid solution when hydrolytic or other reactions do not interfere but the preparation of anhydrous nitrates or nitrato-compounds with very few exceptions such as those of the alkali metals silver and barium require anhydrous conditions. The three principal non-aqueous reagents are not of equal applicability or convenience; for example many anhy- drous nitrates and nitrato-compounds can be prepared with liquid dinitrogen pentoxide1s2 or a solution of dinitrogen tetroxide in an ionising organic solvent3s4s5 but Sn(NO,) and Al(NO,) have been isolated only [Present addresses * Chemistry Department Mid-Essex Technical College Chelms- ford Essex.t Chemistry Division Oak Ridge National Laboratory Oak Ridge Tennessee U.S.A.] M. Schmeisser Angew. Chem. 1955 67,493. B. 0. Field and C. J. Hardy J. 1964,4428. C. C. Addison and B. J. Hathaway Proc. Chem. SOC. 1957 19. C. C. Addison B. J. Hathaway and N. Logan Proc. Chem. SOC. 1958,51. C. C. Addison and B. M. Gatehouse J. 1960 613. 361 362 QUARTERLY REVIEWS TABLE 1. Main diferences between nitrates and nitrato-compounds Nitrates M+.NO,- Examples of M in Anhydrous compounds alkali metals; alkaline earth metals (except Be); some heavy metals (Ag Cd Co Pb Tl). Hydrated compounds majority of metals Conditions of preparation (Section 2) Volatility (Section 3.2) Generally by aqueous methods.Low or involatile. Alkali-metal nitrates can be distilled at 300-500"/ 0.005. Majority of hydrated nitrates de- compose on heating. Infrared absorption Three fundamental bands. (6-15p region) (Section 5) (Section 4) (a) With organic Generally do not react compounds for at room temperature. example ethers hydrocarbons. dinitrogen compounds. t etroxi de . Reactions (b) With liquid Do not form addition Nitrato-compounds Anhydrous compounds M-ON02 Be Al; majority of heavy metals; halogens. Hydrated compounds some heavy metals (Zr Th v>. Non-aqueous methods essential except for the few hydrated compounds. Many sublime at low temperature e-g.7 Ti(NO,) at 25"/0.005. Six fundamental bands. Rapid nitration and oxi- dation at room temper- ature.Form addition com- pounds e . ~ . Cu(N0,) 2.N,04. from reactions involving ClN03.6 This is an inconvenient reagent to use because it is an explosive gas at room temperature and the reaction has to be performed at low temperatures. There is also a limitation to the use of a solution of dinitrogen tetroxide in an electron donor solvent because the solvent may co-ordinate so strongly with the metal that its removal dis- rupts the entire molecule; ZT(NO,)~ for instance cannot be obtained with the help of this reagent,' but can be prepared8 with liquid N20,. The anhydrous inorganic nitrates and nitrato-compounds isolated up to early 1964 and their methods of preparation have been summarised in Table 2 based on a periodic classification. M. Schmeisser and K. Brandle Angew.Chem. 1961; 73 388. B. 0 Field and C. J. Hardy unpublished work 1962 and C. C. Addison personal communication. * B. 0. Field and C. J. Hardy Pruc. Chem. SOC. 1962,76. FIELD AND HARDY INORGANIC NITRATES AND NITRATO-COMPOUNDS 363 2.1. Nitric Acid.-Aqueous nitric acid reacts with many metals and their compounds. The rate of reaction and the nature of the products varies according to the metal or compound the strength of acid and the temperature. The presence of nitrous acid has an important influence on the reaction rate of metals such as Bi Cu Ag Hg and Fe which have a negative oxidation-reduction potential relative to that of the standard hydrogen electrode. Metal carbonates and hydroxides dissolve rapidly in nitric acid as also do most oxides although some of these become recalcitrant when they have been heated to a high temperature.Evaporation of the solution produced to the point of crystallisation usually gives a nitrate hydrate [e.g. Cu(NO3),,5H2O] an oxide- or hydroxide-nitrate which is a salt inter- mediate in composition between normal nitrate and oxide or hydroxide [e.g. ZrO(N0,),,2H20] or a hydrogen salt { e.g. H [Au(N0,~,],3H20). Unusual products are obtained from (a) a solution of tin(i1) oxide in nitric acid; this on concentration gives an explosive hydroxide-nitrate Sn3(OH)4(N0,)2,10 and (b) black silver oxide which is oxidised by nitric acid to the black complex salt Ag,O8NO3. Most non-metals and their anhydrous chlorides are oxidised by nitric acid to the corresponding acid essentially the hydrated oxide; thus sulphur phosphorus and arsenic give sulphuric acid phosphoric acid and arsenic acid respectively.The explosive gas “fluorine nitrate” FNO, is formed when fluorine is bubbled through concentrated HNO3,’ and iodine reacts with concentrated HN03 to yield first a yellow powder claimedll to be iodyl nitrate 10(N03) which is further oxidised to iodic acid. 2.2. Dinitrogen tetroxide N,O,.-Liquid N204 by itself’ reacts at low temperatures with only a few metals e.g. Na K Ag and Pb to form nitrates with Zn to form the addition compound Zn(N0,),,2N20,,12 and with Hg to form dinitratomercury(r1). The uranium oxides U03 and U30 both produce U02(NOJ2,N204 from which the anhydrous compound U02(N03)2 can be obtained by heating the addition complex in vacuum at 165” for two hours.13 Alkali-metal salts react with liquid N204 to pro- duce nitrates KCl+ KNO + NOCl and NaC10 + NaNO + NO2 + C102.Provided there is a trace of water to initiate the reaction liquid N204 reacts with lithium carbonate and the carbonate chloride chlorate oxide and hydroxide of magnesium to form the respective metal nitrates.14 Liquid N204 is non-polar and so displays very poor solvent properties for metal salts and its reaction with many metals for instance Cu is P. J. Durrant and B. Durrant “Introduction to Advanced Inorganic Chemistry” Longmans London 1962 p. 682. lo J. D. Donaldson and W. Moser J. 1961 1996. l1 T. Kikindai Compt. rend 1954,238,1229. la P. Gray and A. D. Yoffe Quart. Rev. 1955,9,379. l3 B. Jesowska-Trzebiatowska and B. Kedzia Bull. Acad. Polon. Sci. Sdr. Sci. l4 J. D. Archambault H. H. Sisler and G. E. Ryschkewitsch J.Inorg. ArucZear Chem. chim. 1962 10 No. 5 213. 1961,17 130. TABLE 2. Anhydrous nitrates and nitrato-compounds isolated up to 1964 and their methods of prepaiation. General rnethods- . Special methods- a. Aqueous nitric acid on metal or hydroxide (also on fluorine and iodine). b. Liqujd NaOr or NaO4 in ethyl acetate on metal chloride or carbonyl. c. Liquid N,O on metal anhydrous chloride hydrated nitrate etc. d. Chlorine nitrate on anhydrous chloride. e. Gaseous NOP-N104 at elevated temperature and pressure. f. Dehydration of hydrate followed by vacuum distillation or sublimation. g. Vacuum dehydration of hydrate at 100" (Sc). h. Ozone on BrN03. 1. Iz on I(NO& j . NOa-N204 on carbonyl in gas phase [FeO(NO,)]. Notes. (i) Roman type is used for nitrates Italic type for nitrato-compounds and Bold type for compounds for which there is insufficient information to allow classification.(ii) The absence of a formula indicates that the anhydrous compound is unknown. (iii) Anhydrous double salts e.g. CsU02(NOJ3 or addition complexes with N,04 and N106 are not included unless srmple nitrates are unknown (e.g. for Hf and Po). 0 s E TABLE 2.-continued &(NO 3) 2 b ~ f Zn(NO,)& As BrN03a Kr Br(N0 3)se Se 1 BrOa(NOs)b Au(NO,),c 1 Hg(NO,),a.c I TlNOp 1 Pb(N03)*a 1 Bi(N03),b 1 Po(N03)p,Nz0,b I Rn * Lanthanide series I Generally M(N03)$0 .. iActinide series Th(N03),b.c I Pa I U 0 2 ( N O J p b 1 Simple anhydrous nitrates not reported for higher members. 366 QUARTERLY REVIEWS exceedingly slow. However when N204 is dissolved in an organic electron- donor solvent such as ethyl acetate or acetonitrile the ionisation of N20 is enhanced and since the metal nitrato-compound produced is usually soluble in the organic solvent the rate of reaction is very much increased.Addison and his co-workers have studied the behaviour of this reagent in detail and have used it generally dissolved in dry ethyl acetate for the prcparation of the following anhydrous compounds from the starting material indicated Metal Cu(N03)2; Cd(N03)2 Co(N03), Hg(N03), Mn(N03)25; Ca(N03)2 Mg(N03)$5; Zn(N03)2‘; U02(N03)217; also Bi(N03) by Straub and his co-workers.18 Anhydrous chloride Be(N03)2 Be,0(N03),19; Ni(N03)220; Fe(NO,), FeNO(N0,) t. Garbonyl (in liquid N20J Fe(NO,), FeNO(N0,):; Co(NO,), Ni(N03)220. Carbonyl (in the gas phase) FeO(N0,) [the carbonyls of Co and Ni give nitrites].20 Some metal oxides react at high temperatures and pressures with N204 to produce an N204 addition compound of the metal nitrate or nitrato-compound e.g.MgO ZnO and CuO at 87” and 14.5 atmospheres’ pressure. Other metal oxides require much lower temperatures and pres- sures; HgO Hg20 and Ca(OH) react at 25” and 1-1 atrn.,l The lanthan- ide nitrates can be prepared by heating the sesquioxides (M,O,) with N 204 at 150”c for 24 hours in a sealed tube.22 2.3. Dinitrsgen Pentoxide N ,O,.-Dinitrogen pentoxide dissolved in concentrated nitric acid was used by Guntz and Martin in 1909 to de- hydrate the hydrated nitrates of Mn Cu Ni and Co and to isolate the anhydrous compounds M(N03)2 for the first time.23 Schmeisser and his co-workers allowed liquid N205 to react with metal oxides and anhydrous chlorides and obtained N205 addition compounds.They prepared the anhydrous nitrato-compounds Cr02(NOJ2 VO(NO3), MoO~(NO~)~ WO,(NO3)2 SbO(NO,), and Ti(N03)4 by heating the addition compounds in a vacuum.l The Reviewers2 extended the method to isolate Zr(N03), Ti(N03)4 Hf(N03)4,N205 NbO(N03)3 Au(NO,), Ill(No3)3 Pd(N03)2 Zn(No3)2 Cu(NO3)2 FeNO(NO3)4 Hg(N03)2 l5 C. C. Addison and A. Walker J. 1963 1220. lo C. C. Addison J. Lewis and R. Thompson J. 1951,2829. l7 C. C. Addison H. A. J. Champ N. Hodge and A. H. Norbury J. 1964,2354. l8 D. K. Straub H. H. Sisler and G. E. Ryschkewitsch J. Znorg. Nuclear Chem. 1962 lg C. C. Addison and A. Walker Proc. Chem. SOC. 1961,242. 2o C. C. Addison “Advances in Chemistry Series” Amer Chem. SOC.Special Publ. 21 J. R. Ferraro and G. Gibson J. Amer. Chem. SOC. 1953 75 5747. 22 T. Moeller V. D. Aftandilian and G. W. Cullen “Inorganic Syntheses” McGraw 23 A. Guntz and F. Martin Bull. SOC. chim. France 1909 5 1004. 24 919. No. 14 1962 131. Hill V 1957 37 ff. FIELD AND HARDY INORGANIC NITRATES AND NITRATO-COMPOUNDS 367 Be4O(NO& and RuNO(NO,),(N 204)0. 75 from the metal hydrated nitrate or anhydrous chloride. The nitrato-compound TaO(N O3I3 has also been prepared2* by this method from the pentaciiloride and froin vacuum-dried hydroxide. Liquid N205 has advantages over N204 in the preparative field because its reactions do not require the presence of an electron-donor solvent. It is however sometimes desirable to dissolve N,05 in an inert solvent such as carbon tetrachloride or trichlorofluoromethane in order to carry out the reaction at a low temperature or to moderate its rate.Trinitrato- chromium(m) has been prepared25 by adding N205 to a suspension of chromium cabonyl in CCl at 20”; and Br(NO,) has been prepared26 from BrF and N,05 in trichlorofluoromethane at -30 to -50”. Under the latter conditions C102 and N205 give NQ3Cl. Most non-metal halides react with liquid N2Q to form an oxyhalide or an oxide; SiC14 BCl, and PC15 give Si02 B2Q3 and POCl respectiveIy,l 2.4. Chlorine Nitrate ClNO,.-This is one of the most powerfcl nitrating agents known; at room temperature it reacts explosively with most metals metal chlorides and organic compounds but at between -40” and -70” the reaction can be controlled. Schrneisser and his co- workers have used it to prepare6 the nitrato-compounds Sn(NQ,), Ti(N03)4 and I(N03) from the respective anhydrous chlorides; they also prepared Al(NO,) from aluminium tribromide and CINO in liqaid bromine.Most non-metal chlorides react with ClN03 and the formation of compounds such as B(N03)3 and S(N03)2 have been postulated6 as the initial product but oxidation-reduction reactions then take place and the final products are usually highly polymeric and of indefinite composition. 3. Structures and Physical Propertics 3.1. Structures.-(a) Nitrates. Metal nitrates are quite dift‘erent structurally from salts of the type M,(X03)n formed by other elemcats of Group V. The NO3- ion is geometrically similar to the planar B0,3- and CO2- ions and differs from the pyramidal SQ,2- aiid C103- ions of the elements of the second Short Period.The P4-0 bond length is gener- ally close to 1.22 AZ7 and the structure of the ion can be explained in terms of valence bond theory as a resonance hybrid of three canonical forms (I) (11) and (111) or of molecular orbital theory by assuming that nitrogen (I) a> (3) 24 K. W. Bagnall D. Brown and P. J. Jones J. 1964 2396. 25 C. C. Addison and D. J. Chapman J. 1964 539. 26 M. Schmeisser and L. Taglinger Angew. Chern. 1959 71 523 Chem Bey. 1961 27 A. F. Wells “Structural Inorganic Chemistry” 3rd Edition Oxford Univ. Press 94 1533. 1962; R. L. Sass et al. Acta Cryst. 1957 10 567. 368 QUARTERLY REVIEWS forms three (T bonds using sp2 hybrid orbitals for the purpose and that the p z orbitals of the nitrogen and the three oxygen atoms combine to form a ' i ~ molecular orbital containing two electrons.The length of the N-0 bond is the same as that in nitric acid and not much greater than that of a double bond (1.19 A) in N20 and NO2. Its explanation poses a similar problem to that of the short S-0 bonds in a number of molecules. There is however a notable exception in the higher value (1.267 A) found in a neutron diffraction study of lead nitrate.28 A number of different crystalline forms are displayed by many metal nitrates and these depend upon the temperature and treatment to which they have been subjected. There is however considerable lack of agree- ment on the details as the work on the alkali-metal nitrates discussed in detail in the S~pplements~~ to Mellor's Comprehensive Treatise on Inorganic and Theoretical chemistry shows.(b) Nitrato-compounds. When the NO3 group is covalently bonded through one of the oxygen atoms to another atom (X) the N-0 bond lengths lie between 1.36 and 1.41 A for the N-O(X) bond and between 1.22 and 1.29 A for the other N-0 bonds for instance in nitric acid(~v),~O fluorine nitrate(~),~l and methyl n i t r a t e ( v ~ ) ~ ~ * ~ ~ (all in the vapour phase) The structure of the anhydrous volatile compound dinitratocopper(I1) has been determined33 by electron diffraction in the vapour phase (in which its vapour density corresponds to the monomer34) and it is of considerable interest in that the nitrato-groups are bidentate i.e. bonded through two of the oxygen atoms as in (VII) and not as first reported.35 The bonding of the ,O\ /O\ NOa = 1.30 & 0.04 A 0 0 ("'I) Cu-N = 2.30 0.03 A O-N\-~~CU<I~TN - 0 cu-0 = 2.00 & 0.02 A NO3 groups is different in the crystalline state and X-ray diffraction data have been inter~reted~~ in terms of two different types of NO3 group.Eight oxygen atoms are linked to each copper atom two by bonds of length 1.9 A and six by weaker bonds of length about 2.5 A. The long bonds are regarded as being largely ionic and the short bonds as covalent and 28 W. C. Hamilton Acta Cryst. 1957,10 103. 2s "Mellor's Comprehensive Treatise on Inorganic and Theoretical Chemistry" Vol. 11 Suppl. 11 Parts 1 and 2 The Alkali Metals Longmans London 1961. 30 L. R. Maxwell and V. M. Mosley J. Chem. Phys. 1940 8 738. 31 L. Pauling and L. 0. Brockway J. Amer. Chem. SOC. 1937 59 13. 32 J. C. D. Brand and T. M. Cawthon J. Amer. Chem. SOC.1955 77 319. 33 R. E. LaVilla and S. H. Bauer J. Amer. Chem. SOC. 1953 85 3597. 34 C. C. Addison and B. J. Hathaway J. 1958 3099. s6 H. Bauer and C. C. Addison Proc. Chem. SOC. 1960 251. 38 S. C. Wallwork Proc. Chem. SOC. 1959 311. FIELD AND HARDY INORGANIC NITRATES AND NITRATO-COMPOUNDS 369 the structure has been formulated as [Cu(N03)],n+,n(N03)-; replacement of the nitrate ions by perchlorate ions could account for the existence of the compound CU(NO,)(C~O~).~~ However this structure does not provide an immediate explanation of the volatility and the infrared spectrum3* contains no bands which can definitely be assigned to the nitrate ion*. A further refinement of the structural analyses has indicated39 a more normal distorted (4 + 2) co-ordination for each copper atom instead of the (6 + 2) co-ordination suggested above.The occurrence of nitrato-groups bidentate to one metal atom in the solid state has been suggested on the evidence of X-ray diffraction measure- ments on the crystalline nitrato-complexes U02(N03)2,6H2040 and RbU02(N03)341 containing the uranyl group. The structure [U02(NOJ2- (H20)2] 4H20 for the hexahydrate accords better with recent infrared data42 than the structure [U02(H20),](N03), suggested previ- ously;43 the two planar nitrato-groups appear to be disposed asymmetric- ally about the uranyl group (VIII) (the uranyl oxygens which are not shown above are approximately at right angles to the plane of the paper and the L 0-U-0 = 173”). The very different mean N-0 distances in the two NO3 groups are difficult to understand. 3.2.Melting Points Boiling Points and Volatility.-Although most anhydrous metal nitrates are stable at the melting point there are frequently large discrepancies between published values of this property and the melting point cannot generally be taken as an accurate indication of purity. Melting points for some of these nitrates are given in Table 3 mainly to show that a clear trend is not apparent in the series of nitrates of the alkali and alkaline earth metals. * [Note addedinproofi Logan Simpson and Wallwork (Proc. Chem. SOC. 1964 341) have shown that two forms of Cu(NO,) exist and that these differ in their X-ray powder photographs and infrared spectra. The published crystal is said to relate to the a-form whereas the published infrared spectra5 relate to the 8-form.] s7 B. J. Hathaway Proc.Chem. SOC. 1958 344. 38 C. C. Addison and B. J. Hathaway J. 1960 1468. S . C. Wallwork Internat. Congress of Internat. Union of Crystallog. Rome Sept. 1963. 40 J. E. Fleming and H. Lynton Chem. and Znd. 1960 1416. 41 J. L. Hoard and J. D. Stroupe quoted in “Spectroscopic Properties of Uranium Compounds” by G. H. Dieke and A. B. F. Duncan McGraw Hill New York 1949 p. 13. 42 J. G. Allpress and A. N. Hambly Austral. J. Chem. 1959,12,569. 43 V. N. Volovenko E. V. Stroganov A. P. Sokolov and V. N. Zandin Radiokhimiya 1960,2,24. 370 QUARTERLY REVIEWS The relatively low melting points of the alkali-metal nitrates have been explaineda as being due to the formation of association complexes which contribute a term to the overall entropy of melting without greatly increasing the heat of fusion.Differential thermal analysis (see section 4.4) is particularly useful in the determination of melting points and the temperatures of phase transitions. A number of commonly available anhydrous and hydrated nitrates have been examined in this way.45p4G TABLE 3. Melting points of unhydrozis metal nitrutes. M in MNOs Li Na K Rb Cs Ag TI MinM(NO,) Mg Ca Sr Ba M. p.41 255" 307" 333" 310" 414" 212" 206" M. p.44 561" 605" 595" * * Anhydrous Mg(NOs)a decomposes almost immediately on Melting points and particularly the so-called "boiling-points" are even less reproducible and meaningful for hydrated metal nitrates than for anhydrous ones because these compounds tend to dissolve in the water of crystallisation and also to decompose by the loss of water and nitrate groups simultaneously or consecutively according to the experimental conditions (see also section 4.4 on thermal decomposition).On the other hand stable nitrato- and nitrato-oxide compounds which sublime or distil in a vacuum have been isolated for a number of metals and non- metals (Table 4). However m.p.s have been given for only a few of these partly because of the experimental difficulties of handling them under anhydrous conditions. The vapour of dinitratocopper(~~)*~ and the solid compounds of several other metals2 have been shown by infrared spectro- scopy to contain covalent nitrato-groups. These nitrato-compounds exhibit high volatility compared with the low volatilityg9 of the alkali- metal nitrates (Table 4). The latter are ionic in the solid state and are probably similar in the vapour state to the slightly volatile ionic alkali- metal chlorides which from measurement of electron diffraction dipole moments and photodisso~iation,~ have been postulated to exist as ion pairs in the vapour state.A qualitative comparison of the volatility of the nitrato-compounds with that of the covalent chlorides of the respective metals can be made but a quantitative comparison is not possible with the exception of dinitratocopper(rI) because boiling points or sublimation temperatures have not been measured under identical conditions or as a function of the pressure. 44 W. J. Davis S. E. Rogers and A. R. Ubbelohde Proc. Roy. Suc. 1953 A 220 14. 46 S. Gordon and C. Campbell Analyt. Chem. 1955 27 1102. M. M. Karavaev and I. P. Kirillov Izvest. V. U.Z. M. V.O. S.S.S. R. Khim.i khim. 47 C. C. Addison and J. M. Coldrey J. 1961 468. 48 B. M. Gatehouse Ph.D. Thesis London 1958. 4s C. J. Hardy and B. 0. Field J. 1963 5130. Technol. 1959 2 231. FIELD AND HARDY INORGANIC NITRATES AND NITRATO-COMPOUNDS 371 3.3. Solubility.-Metal nitrates are soluble in water but there is con- siderable variation in the molar solubility and the temperature coefficient of solubility through a series of metal nitrates of similar formula e.g. for alkali-metal nitrates Metal Na K Rb c s Solubility (moles/kg. H20) 0" 8-6 1.32 2.36 (20") 0.47 100" 22-0 25.3 31.8 10.5 No simple correlation is apparent between the solubility of a metal nitrate and the ionic or hydrated radius of the cation. Concentrated aqueous solutions of metal nitrates can be visualised as consisting of small un- hydrated nitrate ions packed between large hydrated cations.The occur- rence of nitrato-complexes is thought50 to be the exception for instance the strong association of NO3 groups with the ions of Th4+ and InS+ in aqueous solution as indicated by Raman spectra. Little quantitative information is available on solubilities of nitrates in organic solvents e.g. alcohols ethers ketones and organo-nitrogen compounds although some of these have been used in the separation of metals and in analytical chemistry. The solubilities of silver nitrate and the dinitrato-compounds of copper(I1) and zinc@) are compared in Table 5 for four different solvents.61 TABLE 5. Solubilities of silver nitrate and dinitrato-compounds of copper and zinc in various solvents at 20" (except where otherwise stated).Solvent Solubility (moles per mole of solvent) CWO,) Zn(NO,) &NO H2O 0.144 (25") 0.122 (25") 0.242 Ethyl acetate 0.71 (25") > 0.6 0-014 Nitromethane 0.016 0.0015 negligible Acetonitrile 0.074 0.15 0.27 The high solubility in ethyl acetate of the covalent copper and zinc compounds ( 5 times higher than in water on a molar basis) and the low solubility of the ionic silver nitrate led to the inference that high solubility in this kind of solvent was mainly a consequence of the covalent bonding between the metal and the nitrato-group. However anhydrous cadmium nitrate is ionic in the solid state and is also very soluble in ethyl acetate hence the bond character is only one of a number of contributory factors. The solubilities of many metal nitrates have been determined in tri-n- butyl phosphate (TBP) largely because the solvent is used in the separation of uranium and plutonium from fission products.A summary of the results obtained up to 1958 has been given by McKay and HeaIys2 who 6o C. K. Jorgenson "Inorganic Complexes" Academic Press London 1963 p. 82. 61 C. C. Addison B. J. Hathaway N. Logan and A. Walker J. 1960,4308. 63 H. A. C. McKay and T. V. Healy Progress in Nuclear Energy Series 111 Process Chemistry Pergamon Press London 1958 Vol. 2 pp. 549. Class TABLE 4. Physical properties of volatile nitrates and nitrato-compounds. Compound Ref. M.p. FeNO(N03) C Nitrato- Hf(N0&N,O5 a compounds Hg(NO3I2? C Ref. Volatility ("C/mm.Hg.) Subl. 80/0-004 b V.P. 150/0-25 to 214/ 3.0; ht. of sublima- tion 15 kcal/mole Subl. 120/vac. 120/0*001 Subl.100/0*01 Subl. in vac. 240/0.07 Subl. 130/0401 Subl. 80/0403 Subl. 40/vac. 20/0.02 e a Subl in vac. Ref. B.p. of chloride a AuCl, subl. 265" CuCl decomp. to CuCl 993 O b c FeCl, 315" a a HfCl, subl. 417" c HgCl, 302" a a InCl, subl. <400" a PdCI2 m.p. 500" (decomp.) d SnCI, 114" e TiCl, 136" a TABLE 4.-continued Class Compound Be,O(NO,) VO(N0,) 3 Nitrato-oxide CrO,(NO,) compounds MNO (where Ionic M=Li Na K nitrates Rb Cs). NH,NO Ref. a f g i k 1 1 0 M.p. 140 (decomp.) - 175 - 107 - 70 - 35 2 253 (Li) to 414 (Cs). 169 Ref. a h j 1 m e Volatility ("C/mm.Hg.) Subl. 275/0*03 Subl. lOO/O*Ol B.P. -42 B.P. -18 Subl. 125/0*05 B.p. 28/0.001 B.p. 45/2 B.p. 68-70/vac. Distil at 3 50-5OO/0~005 Subl. 75/0*001 Ref B.p. of chloride a ZnCl, 732" f ZrCl, subl. 331 " h FCI -101" j Cl, -34" a BeCl, 520" I CrO,Cl, 117" m e VOCI, 127" E.g.NaCl 1413" n n Subl. 335" aB. 0. Field and C. J. Hardy J. 1964,4428. C. C. Addison and B. J. Hathaway Proc. Chem. SOC. 1957 19; J. 1958 3099. C. C. Addison B. J. Hathaway and N. Logan Proc. Chem. SOC. 1958 51. M. Schmeisser and K. Brandle Angew Chern. 1957 69 781. M. 2 Schmeisser Angew. Chem. 1955 67 493. f B. 0. Field and C. J. Hardy Proc. Chem. SOC. 1961 76. G. H. Cady J . Amer. Chem. SOC. 1934,56 2635. A P. J. Durrant and B. Durrant "Introduction to Advanced Inorganic Chemistry" Longmans London 1962 p. 94. H. 8 Martin and Th. Jacobsen Angew. Chem. 1955,67,524. j H. Martin Angew. Chem. 1958,70,97. C. C. Addison and A. Walker Proc. Chem. Soc. 1961,242. M. Schmeisser and D. Lutzow Angew. Chem. 1954,66,230. * W. H. Hartford and M.Darrin Chem. Reviews 1958 58 cd 1-61; these authors claim m.p. in ref (I) too low owing to supercooling. C. J. Hardy and B. 0. Field J. 1963 5130. O Ray and Jana J. 1913 103 1565. z tl v1 w 4 w 374 QUARTERLY REVIEWS showed that for highly soluble salts the composition of the saturated solutions corresponds to the formation of definite solvates ; for instance U02(N03)2(TBP)2 which has a sharp m.p. of -6". Solubilities of a few anhydrous and hydrated metal nitrates and of U02(N03)2,6H20 in TBP are given in Table 6. In general anhydrous ionic metal nitrates have low solubilities in TBP whereas hydrated ionic metal nitrates show appreciable solubility and water is also dissolved with them. TABLE 6. Solubilities of metal salts in TBP at 25". Solid phase Solubility of salt (moles/l.) 1 -32 0.054 0.0025 0.99 0.19 0-01 6 0.00075 1.14 0.83 1.6 Solubility of water (mole@ .) 3.14 - 1 *84 3.19 - 2.28 3.20 0.33 3.4.Other Physical Properties.-Molten salts particularly nitrates and chlorides are being studied extensively for use as high-temperature solvents and reaction media and an increasing amount of information on their viscosity electrical conductivity and surface tension is becoming available. Viscosities of the nitrates of Li Na K NH4 T1 and Ag and electrical conductances of the nitrates of Li Na K Rb Cs T1 and Ag expressed as a function of temperature have been tabulated by Ubbelohde and his co-w~rkers,~~ aqd association into complexes has been postulated to explain the lower melting points of these salts than of the ionic halides of the corresponding metals.The surface tensions of the anhydrous nitrates of the alkali metals and T1 Ag Ca Sr and Ba have been measured over wide temperature ranges by Addison and C01drey.~' They consider that a degree of covalency is evident in the nitrates of metals in Groups I and 11 in the liquid state. The densities and some other physical properties of solid metal nitrates and their aqueous solutions are generally given in compilations of physical properties of inorganic substances but these are usually derived from single measurements and few detailed studies have been made. A notable exception is found in the determination of data on aqueous and organic solutions relevant to solvent extraction recovery systems for nuclear 63 J. P. Frame E. Rhodes and A. R. Ubbelohde Trans. Furaday Suc. 1959,55,2039.64 C. M. Slansky Progress In Nuclear Energy Series 111 Process Chemistry Perga- mon Press London 1958 Vol. 2 p. 535. FIELD AND HARDY INORGANIC NITRATES AND NITRATO-COMPOUNDS 375 4. Reactions The more important reactions of inorganic nitrates and nitrato- compounds are summarised for convenience under the headings of oxida- tion nitration hydrolysis thermal decomposition radiolytic decomposi- tion and addition compounds with N204 and N205. Some of the reactions for instance the reduction of nitrate to ammonia and to nitric acid,which are used in the analytical determination of the nitrate group are discussed at the end of this Section. 4.1. Oxidation Reactions.-The explosive oxidation of sulphur and carbon by solid potassium nitrate mixed in the form of gunpowder has been known in Europe since the thirteenth century and has had a pro- found effect on the development of the modern World.All solid nitrates and nitrato-compounds are oxidising agents to a greater or less extent and will react with organic matter to form carbon monoxide carbon dioxide and oxides of nitrogen. Intimate mixtures of nitrates with certain metals can explode violently; the mixture of zirconium powder and UO2(NO,), 6H20 formed during the selective dissolution of uranium from a zir- conium-uranium alloy by concentrated nitric acid is extremely sensitive to percu~sion.~~ Solid nitrates and liquid nitric acid have been used as a source of oxygen in rocket propellants. 4.2. Nitration Reactions.-Hydrated metal nitrates react with ali- phatic hydrocarbons in sealed tubes at 100-150” to produce in general a mixture of nitro-alkanes and oxidation product^.^^^^^ Aromatic hydro- carbons are not nitrated when heated with the anhydrous nitrates of K Na NH4 Ba Pb or Ag but in the presence of aluminium trichloride they are nitrated at 30-40”.56 These salts are probably not effective as nitrating agents alone because they are ionic nitrates and are unable to produce the necessary reactive species such as NO3.or NO,. radicals or the nitronium ion NO2+. Metal nitrato-compounds with the covalent structure M-0-NO are generally highly reactive owing to dissociations occurring at the M-0 or 0-N bond (or both) to give NO,. radicals. This behaviour is similar to the dissociation of alkyl nitrates R-0-NO2 to give NO2. radicals which is an important step in the reactions (and pyrolysis) of alkyl and is in contrast to the nitration of aromatic compounds by alkyl nitrates in the presence of sulphuric acid59 when the nitronium ion NO,+ is the active species.The stronger the covalent bond between the metal or non-metal and the nitrato-group the more reactive is the nitrato complex and the more likely is the complex to be volatile; the highly volatile titanium 55 F. S. Martin and B. 0. Field A.E.R.E. Report C/R 2692 (1958). 66 A. V. Topchiev “Nitration of Hydrocarbons and Other Organic Compounds” 57 F. Asinger G. Geiseler and W. D. Wirth Chem Ber. 1957 90 1987. 58 P. Gray and A. D. Yoffe Chem. Rev. 1955 55 1069. 59 R. Boschan R. T. Merrow and R. W. van Dolah Chem. Rev. 1955,55,485. Pergamon Press London 1959. 376 QUARTERLY REVIEWS compound Ti(NO,) readily nitrates and oxidises paraffinic hydrocarbons at 25°.60 The reactions of the anhydrous volatile copper compound Cu(N0,) with diethyl ether dimethyl ether and nitromethane have been studied in detail by Addison and his co-workers.20 When (CuNO,) is dissolved in a basic solvent (L) a complex such as (IX) is formed and the strength of the Cu-ONO bonds varies with the nature of L.When ether vapour is L O~NO- cU”- ONO L’ (1x1 passed into a solution of Cu(NO,) in nitrobenzene NO2 and acetaldehyde are evolved and a green solid is precipitated. When ether is added to a solution of Cu(NO,) in ethyl acetate or methyl cyanide it does not react because these solvents co-ordinate so strongly to the copper atom that the Cu-ONO bond is so weakened that dissociation to NO,. radicals does not occur (N0,)Cu-0-NO -+ (N0,)Cuf + NO3- (in basic solvents; no (N0,)Cu-0-NO -+ (N0,)CuO.+ NO2. (in very weakly basic Thus the anhydrous nitrato-copper compound can be regarded as a low- temperature source of NO radicals and the reaction with ether can be formulated as attack on ether) solvents or with Cu(NO,) alone; ether attacked by NO2.) CH3*CH ,*OCH ,*CH3 CO bond fission and addition ofN0 H abstraction by NO2* / \ \ CH3*CH2* + C2H,.0N0 / R*+HONO J explosive decornp- osition 1 Oxidation by 4 NO2 CH,CHO .1 CH3C02H In addition nitrate ions and hydroxyl ions may result from the NO,,CuO. radical formed initially. Nitromethane is a good solvent for Cu(NO,) on account of its high dielectric constant and the solution is stable at room temperature but NO is evolved when the solution is boiled and a green salt containing copper is precipitated.This reaction probably proceeds by a free-radical mechanism in the same way as that with ether and the gas-phase reaction of NO2 with nitromethane at 400”. 6o B. 0.Field and C. J. Hardy J. 1963 5278. FIELD AND HARDY INORGANIC NITRATES AND NITRATO-COMPOUNDS 377 Chlorine nitrate and fluorine nitrate react explosively in contact with many inorganic and organic compounds but under controlled conditions i.e. at low temperature in a suitable organic solvent C1NO3 will both nitrate and chlorinate organic compounds. The many reactions of halogen nitrates have been reviewed recently.6 4.3. Hydrolysis.-Anhydrous and hydrated metal nitrates and nitrato- compounds in general dissolve in water to give nitrate ions and metal ions or hydrolysis products of the metal ion e.g.xFe3+ + yH,O + Fe,(OH),3z-~+ + yH+ Hydrolytic reactions of this kind have been reviewed by Sillkn;61 they depend primarily on the nature of the metal and occur with compounds containing many other anionic groups for example chloride perchlorate and sulphate and will not be discussed further. Dinitratoberyllium(I1) evolves brown fumes on dissolving in water and produces nitrate and nitrite ions in sodium hydroxide solution in contrast to the anhydrous ionic nitrates of Ca Sr and Ba.l5pZ0 It has been suggested that the dissociation of the covalently bonded -Be-0-NO group by the breaking of the 0-N bond can explain the small amount of nitrite formed The NO2 produced gives equal quantities of nitrite and nitrate ions. In non-aqueous media such as ethyl acetate the anhydrous beryllium com- pound dissolves to a pale yellow solution which has an absorption spec- trum corresponding to that of The volatile nitrato-oxide compound of beryllium Be,O(NO,), hydrolyses slowly in water to give only nitrate ions.This compound is thought to have the same structure as the volatile acetato-oxide compound Be,O(Ac),; each nitrato-group is part of a six-membered ring with two of the oxygen atoms bonded to beryllium atoms. It is therefore sug- gested159z0 that the -NO2 group cannot break away as in dinitratoberyl- I i u m( I 1). No other conclusive example has yet been found of a covalent nitrato- compound of a metal which gives nitrite ions on hydrolysis; highly covalent nitrato-compounds of titanitinPo and zirconium,* which might be expected to react in this way only give nitrate ions on hydrolysis.The unusual binuclear nitratonitrosylrutheniiim c0mplex~~9~~ RU 2N601 (for which infrared spectroscopy shows nitrite ions and nitrito- and nitro-groups to be absent) gives nitrite on hydrolysis. However this cannot unequivocally be said to arise from the co-ordinated nitrato-groups because the nitrosyl 61 L. G. SillCn Quart. Rev. 1959 13 No. 2 146. 62 F. S. Martin J. M. Fletcher P. G. M. Brown and B. M. Gatehouse J. 1959 76. 63 D. Scargill and J. M. Fletcher Proc. Chem. Sue. 1961 251. 378 QUARTERLY REVIEWS groups may be more reactive than in the more usual series of complexes of the general formula RUNO(NO~)~(H~O)~-~ which do not give nitrite on hydrolysis. 4.4. Decomposition.-(a) Thermal. The long-established view of the thermal decomposition of metal nitrates is that the first step is probably the loss of oxygen with the formation of metal nitrite.If this is stable at the temperature of the experiment the reaction stops e.g. the alkali-metal nitrates at temperature less than about 750" 2MN03 + 2MN02 + O2 The equilibrium pressure of oxygen is 1 atmosphere at the following temperatures Li 474"; Na 525"; K 533"; Rb 548"; Cs 584".64 If the nitrite is unstable at the given temperature further decomposition into metal oxide and oxides of nitrogen occurs and the amount of nitrite at any given time depends on the relative rates of decomposition of the nitrate and nitrite; thus for Ba(NO,) at 600" Ba(NO3)2 + Ba(N02)2 + 0 2 Ba(NO,) -+ BaO + NO + NO2 If the nitrite is very unstable the only product is the oxide provided it is stable at the given temperature Finally if the oxide is thermally unstable at the temperature at which the nitrate and nitrite decompose the metal is produced The above treatment is over-simplified; other reactions are known to proceed simultaneously for many nitrates and the nature and amounts of the products frequently depend upon the rate of heating and the size of the sample.In addition a hydrated nitrate can behave differently from the anhydrous nitrate or nitrato-complex of the same metal. Much of the early work in this field consisted of qualitative observations on materials of unknown purity and it is not surprising that some of the results are ap- parently contradictory and that there is little detailed knowledge of the basic reactions.The decomposition of the alkali-metal nitrates illustrates some of the simultaneous and consecutive reactions which can occur even for these "simple" ionic salts. Sodium nitrate decomposes rapidly above 800"c to give a substantial amount of as well as nitrite and as the tempera- 64 M. C. Sneed and R. C. Bransted "Comprehensive Inorganic Chemistry" Van hlostrand New York 1957 p. 125. 65 K. Leschewski and G . Zulla Ber. Ges. Freundeit tech. Hochschule Berlin 1942 1 168. FIELD AND HARDY INORGANIC NITRATES AND NITRATO-COMPOUNDS 379 ture is increased sodium peroxide is also formed.se It seems likely that alkali-metal ions nitrate ions and nitrite ions (all probably as ion-pairs) exist in the vapour phase at these temperatures because these metal nitrates and nitrites can be distilled without decomposition at temperatures in the range 450-550°/10-3mm.49 The two main reactions for alkali-metal nitrates are endothermic :66 NaNO -+ NaNO + 0; AH = + 23 kcal.2NaN0 -+ Na,O + 2N + 5 0 ; AH = + 120 kcal. But when the nitrate is heated in the presence of a metal with a high heat of combustion it is possible for an exothermic chain reaction to be initiated which may lead to an explosion 3KN0 + 5A1+ 3KA10 + A1,03 + 3/2N2 When molten mixtures of sodium nitrate and potassium nitrate are used for the heat treatment of metals and alloys it is therefore necessary not to exceed certain maximum temperatures to avoid serious accident^.^' The thermal decomposition of hydrated heavy-metal nitrates and nitrato- compounds is more complicated than that of the alkali-metal nitrates because hydrolysis can occur and complexes (probably polymeric) con- taining hydroxo- and 0x0-groups in addition to nitrato- and nitrito-groups and nitrate and nitrite ions can be formed as intermediates.An empirical formula for such an intermediate with a definite X-ray pattern is Ni(N03)- (OH),J-Hz0.68 Thermogravimetric analysis (T.G.A.) in which loss of weight is measured as a function of temperature and differential thermal analysis (D.T.A.) in which the thermal capacity is measured relative to that of a standard substance have been used in recent years to study these reactions. The results must however be interpreted with caution unless chemical analysis and/or infrared spectra are also available at intermediate stages. T.G.A. of rare-earth nitrates has been claimed6g to show the forma- tion of intermediate oxide-nitrates mixed with the oxides but an infrared study of praseodymium nitrate shows that it decomposes via! an inter- mediate nitrite.’O T.G.A.is particularly useful for determining the tempera- ture at which the final stable product is obtained and results for forty-five metal nitrates have been collected by DuvaP who has discussed their application in quantitative analysis. The thermal decomposition of ten anhydrous45 and several hydrated nitrate^^^,^^ has been studied with the D.T.A. method. A further complication which adds to the difficulty of 66 K. Leschewski and W. Degenhard Ber. 1939 72 1763. 67 H. Remy “Treatise on Inorganic Chemistry” Vol. 1 Elsevier London 1956 68 E. M. Vander Wall U.S.A.E.C. Report IDO-14597 (1962). 69 W. W. Wendlandt and J.L. Bear J. Znorg. Nuclear Chem. 1960 12 276. 70 F. Vratny et al. Trans. Faraday Soc. 1960,56,1051; J. Inorg. Nuclear Chern. 1961 71 C. Duval “Thermogravimetric Analysis of Inorganic Compounds” Elsevier p. 601. 17 281. London 1963. 380 QUARTERLY REVIEWS understanding the reactions is the occurrence of allotropic forms of the final product. There is X-ray diffraction and other e v i d e n ~ e ~ ~ ~ ~ that the thermal decomposition of the dihydrate of dinitratouranyl(r1) (obtained from the more common hexahydrate) at temperatures up to about 450" can give uranium trioxide with at least three different crystallographic forms (p y and amorphous) and that the majur product can be any of these according to the experimental conditions. (b) Radidytic. The overall effect of irradiating solutions of metal nitrates with X-rays,74 fast neutrons or garnrna-ra~s,~~ is to reduce nitrate ion to nitrite ion and to liberate oxygen and hydrogen.A little nitrogen is also produced from concentrated solutions of The large reduc- tion in the hydrogen yield as the nitrate concentration is increased can be qualitatively explained by the reaction of nitrate ions with diffusing hydrogen atoms NO3- + H -+ NO2 + OH- Irradiation of these solutions with fission recoil fragments does not pro- duce nitrite but greatly increases the yield of nitrogen and the stoicheio- metry of the gaseous products from solutions of calcium nitrate75 cor- responds to the overall equations H 2 0 + H 2 + 4 0 2 Nitrogen is thought to be produced by the direct action of highly energised ions or electrons on the nitrate ion and the nature of the cation has little effect on the nitrogen yield.The irradiation of crystalline metal nitrates in a nuclear reactor generally gives equivalent amounts of oxygen gas and nitrite ions mainly by elec- tronic ionisation and excitation rather than by elastic collisions with particle^.^^^^^ The photolysis of solid nitrates of the alkali metals and some heavy metals with light from a high-pressure mercury arc produces nitrite and oxygen with a quantum yield varying from 0.002 to 0 ~ 1 9 ; ~ ~ water of hydration increases the sensitivity to radiation. 4.5. Addition Compounds of Metal Nitrates and Nitrato-compounds with N204 and with N205.-These are important intermediates in the isolation of many metal nitrates and nitrato-compounds from reactions involving N20 or N205; they are generally thermally unstable and dis- 72 R.S. Ondrecjin U.S.A.E.C. document TID 17733 (1963). 73 R. S. Ondrecjin and T. P. Garrett J. Phys. Chem. 1961 65 470. 74 N. A. Bakh Conf. Acad. Sci. U.S.S.R. Peaceful Uses of Atomic Energy 1955 75 R. G. Sowden Trans. Faraday SOC. 1959 55 2084. 76 G. Hening R. Lees and M. S. Matheson J. Chem. Phys. 1953 21 664. 77 D. Hall and G. N. Walton J. Inorg. Nuclear Chem. 1959 10 215. 78 D. Doigan and T. W. Davis J . Pliys. Cliem. 1952 54 764. (referred to in ref. 75). FIELD AND HARDY INORGANIC NITRATES AND NITRATO-COMPOUNDS 38 1 sociate when heated in a vacuum to produce the simple nitrate or nitrato- compound and oxides of nitrogen. The conditions of thermal dissociation must be carefully controlled particularly if the required product is itself thermally unstable because then complex mixtures of oxide-nitrates can easily be formed; for example14 80 O/vac.4 hr. Al(N03)3,0*38N204 - -f Al(N03)3 -k Al2o(No3) Among the more definitely characterised addition compounds are M(N03),,N204 M(N03),,2N204 M(NO3),,2NzO4 M = In79 M(N03)4,N205 M = HP M(N03)4,2N,05 M = Th80 M = Mg,15 Cu,15s7' uo217~79 M = Be,15 C O ~ ~ Zn79 These addition compounds give the simple anhydrous compound on heat- ing in a vacuum for times and at temperatures which vary with their stability; for instance 5 hr. at 160°/10-5 mm. for the thorium addition compound and 52-54"/10-160 mm. for the magnesium compound. The more covalent the bond between a metal and the NO3 group the more likely is the molecule to form an adduct with N204 or N,O,; thus the ionic nitrates of the alkali metals and of Ca Sr Ba and Cd do not form these adducts.Magnesium nitrate is the only ionic nitrate known to form such an addition compound and the bonding of N204 in this adduct is different from that in the addition compounds formed with nitrato- compounds for example in Cu(NO3),,N,O4. This difference in bonding is clearly seen in differences in behaviour on thermal dissociation. The shapes of the vapour pressure-temperature curves for the dissociation of Mg(NO,),,N2O4 and Cu(N03),,N204 are not the same; the magnesium compound shows a sudden pressure increase at 52-54" when an irrevers- ible reaction occurs Mg(N03)2,N204 -+ Mg(N03)2 + N2°4 but a smooth curve is obtained for the reversible reaction Cu(N03)2,N204 Cu(N03)2 + N204 The structure of these addition compounds is open to some doubt; neither Zn(NO,) nor N204 reacts appreciably with dry ether but the addition compound Zn(N03),,2N,0 reacts vigorously.Formulation as nitrosonium salts for example (NO+) [Zn(N03)4]2- is particularly attrac- tive as the solid compounds exhibit a sharp absorption band at 2260-2300 cm.-l characteristic of the NO+ g r o ~ p . ~ ~ ~ ~ ~ The complexes Fe(NO,\, 7 9 C. C. Addison and B. J. Hathaway Chem SOC. Special Publication No. 10 1957 33. 8o 3. R. Ferraro L. I. Katzin and G. Gibson J. Amer. Chem. Soc. 1955 77 327. 382 QUARTERLY REVIEWS N 204 formulated* as (NO+) [Fe(NO,),]- and Hf(N03)4,N 205,2 are stable and can be sublimed as such in a vacuum. In general the co-ordinated oxides of addition can be replaced by other compounds such as pyridine (py) in UO2(N03)2,N204 which gives U02(N03)2,2py.4.6. Reactions of Importance in Chemical Analysis.-The reduction of the nitrate ion to ammonia has long been used for the determination of nitrate and is usually carried out in strongly alkaline solution in the presence of aluminium zinc or Devarda’s alloy (50 Cu 45 Al 5 Zn) 3N03- + 8A1 + 2H20 + 50H- -+ 8 A102- + 3NH3 The ammonia is distilled into a standard solution of an acid and the excess of acid is titrated with a standard solution of alkali. The nitrate ion can also be reduced to ammonia in neutral solution by the use of Arndt’s alloy (60 Mg 40 Cu). The reduction of nitric acid to nitric oxide occurs when a solution of the acid or a nitrate in concentrated sulphuric acid is shaken with mercury 2HN03 + 6Hg + 3HzS04 -+ 2N0 + 3Hg2S04 + 4H20 This reaction is used for the estimation of nitrates or nitrites or oxides of nitrogen in commercial sulphuric acid in the Lunge nitrometer.Ferrous salts reduce nitric acid to nitric oxide which in the cold dissolves in the excess of ferrous salt to give a dark brown solution the “brown ring” in the qualitative test for nitrate. The solution evolves nitric oxide on heating 6FeS0 + 2HN03 + 3H2S0 -+ 3Fe2(SQ& + 2N0 + 4H20 Various other reactions involving the reduction of the nitrate ion both catalytically and by strong reducing agents have been discussed by Szabo and Bartha.81 Other tests for nitric acid and nitrates are (i) the red colour with a solution of brucine in concentrated sulphuric acid (ii) the deep blue colour with a solution of diphenylamine in concentrated sulphuric acid (iii) the evolution of red-brown oxides of nitrogen on heating with concentrated sulphuric acid and copper turnings and (iv) the white crystalline precipitate which develops on adding a solution of the nitron reagent (1 % w/v in 5 % acetic acid) to a solution of the nitrate acidified with acetic or sulphuric Z.G. Szabo and L. G. Bartha “Recent Aspects of the Inorganic Chemistry of Nitrogen” Chemical Society Special Publication No. 10 1957 p. 131. FIELD AND HARDY INORGANIC NITRATES AND NITRATO-COMPOUNDS 383 acids. The nitron reagent is now widely used for the gravimetric determina- tion of nitrate in the presence of many other ions. Nitron (X) is a strong organic base 4,Sdihydro- 1,4-dipheny1-3,5-phenylimino- 1,1,4-triazole (to which a sterically improbable structure is commonly ascribed) and gives a 1 :1 complex with the formula CzoHl6N4,HNO,.5. Spectra The infrared and Raman spectra of simple metal nitrates in the solid state and in aqueous solution have been known for many years as exemplified by the work of Menzies.s2 The publishing of many data since about 1957 on the vibrational spectra of ionic nitrates and covalent nitrato-compounds has been stimulated in part by the preparation and study of anhydrous and volatile nitrato-compounds and also by the widespread interest in the separation of metals by solvent extraction of their nitrates. In most of the recent work the infrared absorption technique has been applied to solid compounds over the easily accessible wavelength range 2-15p and this field will mainly be discussed in this Section.The application of the Raman and infrared techniques to the study of molten metal nitrates has been reviewed recentlyB3 aad will only be mentioned briefly as will the well- known ultraviolet spectrum of the nitrate ion. 5.1. Infrared Spectra of Solid Metal Nitrates and Nitrato-compmds.- The nitrate ion belongs to the group of planar XY3 molecules and has D3h symmetry which gives rise to four fundamental vibrations one of which is infrared forbidden. In compounds in which the NO3 group is covalently bonded through one of the oxygen atoms the symmetry of the group is lowered to CZv and six fundamental vibrations are expected. Arising out of the different symmetries of the nitrate ion and the nitrato- group two different conventions are useds4 in numbering the vibrations and these are given in Table 7 to avoid confusion in subsequent discussion.The frequencies given for the vibrations of the nitrato-group are those found experimentally in the early works4 on nitrato-co-ordination com- plexes of various metals and are included for comparison with those of the nitrate Although the spectra in the 2-15p region can often be used5*l5 to distinguish clearly between the nitrate ion and the covalent nitrato-group it should be noted that the symmetry of the nitrate ion can be lowered if it is subject to the field of a crystal lattice and that in ionic crystals the co- hesive energy can be of the same order of magnitude as the bond energies of covalent compounds. A detailed study by Buijs and Schutte86 of the 82 A. C. Menzies Nature 1929 124 511.83 S. C. Wait and G. J. Janz Quart. Rev. 1963 17 225. 85 G. Herzberg “Infrared and Raman Spectra of Polyatomic Molecules” Van 86 K. Buijs and C. J. H. Schutte Spectuochim. Acta. 1962 18 307. B. M. Gatehouse S. E. Livingstone and R. S. Nyholm J. 1957 4222. Nostrand New York 1945 p. 178. 384 QUARTERLY REVIEWS infrared spectra of the anhydrous nitrates of Li Na K Ca Sr and Ba has shown that the assumptiona7 of partially covalent bonding is not necessary to explain the observed differences from ideal D3h symmetry in these com- pounds. Unfortunately these authorss6 use a different notation from other workers in the field and this can lead to confusion; they refer to the very strong v band for the nitrate ion (Table 7) as v4 and vice versa. TABLE 7. Type v1 (4) v2 (A",) v3 (E') v4 (E? Description of the vibrations of the nitrate ion and nitrato-group.Nitrate ion NO,- Assignment NO stretch out-of-plane NO2 asymm. NO2 bend Active in or Raman (R) Frequency (cm.-l) 1050 83 1 1390 720 bend stretch (planar rock) infrared (I) R I 1 3 IYR Nitrato-group O-N02 Assign- NO out-of- NO2 NO2 NO2 NO bend ment stretch plane symm. asymm. bend asymm. rock stretch stretch symm. (out-of- Type v 2 v6 (B2) v1 (A,) v4 (B1) v3 (A,) v5 (B1) plane rock Active in infrared (I) I,R I,R J,R 1 3 1 3 1 3 or Raman (R) Frequency (cm.-l) 1034-970 800-781 1290-1253 1531-1481 -739 -713 The infrared spectra measured by various workers are generally consist- ent with (i the nitrate ion being present in (a) the anhydrous nitrates of the metals Li Na K Rb Cs Mg Ca Sr Ba Cd Co Pb Ag; (b) the hydrated nitrates of the metals Cd Co Cu FelIr Ni Zn; (ii) the co- ordinated nitrato-group being present in (a) the anhydrous compounds (many of which are volatile see Table 4) of the metals Be Cu Hg" In Mn Pd Ti1" Zn Zr and groups Be40(vr) FeNO(rIr) NbO(rr1) and in (NH,) [Ce(NO,),] ; (b) the hydrated compounds.Th(N03),,4H20 ZrO( NO,) 2,2H 20 RuNO( NO,) ,,2H 20 and 6 - 3 UO 2( N 0 3) 2,2H 20. The nature of the bonding in the nitrato-compounds of the above metals merits closer examination. The strong band usually assigned to the asym- metric stretching frequency (v4) of the nitrato-group in some of these compounds occurs2 at as high a value as 1620-1646 cm.-l and the differ- ence between this band and that usually assigned to the symmetric stretch- ing frequency ( v ~ ) is greater than 400 cm.-l.These values are the highest reported for nitrato-compounds of metals and we have related them2 to the J. R. Ferraro J. Mol. Spectroscopy 1960 4 99. FIELD AND HARDY INORGANIC NITRATES AND NITRATO-COMPOUNDS 385 presence of bridging nitrato-groups from a comparison of the spectra with those of a number of compounds in which it is believed that the nitrato- group must be (a) unidentate to satisfy the known co-ordination number of the metal e.g. nitrato-complexes of Co'I' amines hexanitrato-complexes of U V I and Ce'V; (6) bidentate according to X-ray evidence e.g. Rb[U02(N03\3] U02(N03)2,6H20 ; and (c) bridging two metal atoms according to X-ray evidence e.g. Cu(NO,), or by analogy with similar compounds of known structure e.g. Be40(N03), which is similar to Be40(Ac),.All the compounds believed to contain the unidentate nitrato-group give two main bands (sometimes split) in the 1200-1600 cm.-l region which lie within the limits v4 1560-1454 v1 1346-1254 with a difference (v4-v1) of between 100 and 300 cm.-l. The uranyl compounds which con- tain the bidentate nitrato-group have two strong bands which lie within the above ranges for the unidentate nitrato-group. It would however be unwise to infer that a bidentate nitrato-group cannot be distinguished from a unidentate group on the basis of the bands in this region because the uranyl compounds containing the bidentate group are octaco-ordinated with respect to the metal whereas the others are hexaco-ordinated and the bond angles may be considerably different. We think that it is significant that the anhydrous compounds Be,O- strong bands lying well outside the limits given above for v4 and v1 in unidentate nitrato-compounds.We therefore suggest that the occurrence of a strong band at a frequency greater than 1570 cm.-l together with a strong band at a frequency less than about 1280 crn.-l may indicate a bridging nitrato-group. An attempt to assign these bands is being made* because they are not identical to those named v4 and v1 for the unidentate nitrato-group. Information about the structures of these compounds from X-ray or neutron diffraction measurements will be required to substantiate the above suggestions and the measurement of infrared spectra in the region in which the metal-oxygen bonds are expected to vibrate (300-500 cm.-l) would be particularly useful in determining whether a compound contains unidentate or bidentate nitrato-groups but little information about this region has been published.Great care is required to obtain meaningful spectra of these covalent nitrato-compounds owing to their high reactivity; for instance the compound Ti(N03)4 reacts rapidlyao with the paraffinic hydrocarbons used to prepare mulls. Infrared and Raman' Spectra of Metal Nitrates in Solution and in the Molten State.-The infrared and Raman spectra of metal nitrates have been studied less in aqueous and organic solvents than in the solid state but interest in this field is increasing as it is also in regard to molten salts. Published data up to 1958 on the Raman spectra of metal nitrates in *Unpublished work by the Reviewers 1964; also independent work by Professor C.C. Addison and his co-workers (personal communication). (NO,), NbO(N03)3 In(NO8)3 Pd(NO3)m Ti(N03)4 and Zr(NO3)4 have 5.2 3 386 QUARTERLY REVIEWS solution were discussed by Gatehouse considerable deviations were found from the spectra expected for the Dab symmetry of the nitrate ion. More recent ~ ~ r k ~ ~ ~ ~ ~ has indicated strong association of the nitrate ion with Thw and In*n in aqueous solution. It is expected that further measurements by the Raman method in conjunction with X-ray diffraction studies on the solutions will considerably advance our knowledge of the structure and bonding in solutions of electrolytes. Infrared spectra have been measured for solutions of metal nitrates in organic solvents particularly in tri-n-butyl 91 and solutions of tri-n-octylphosphine oxide in carbon tetrachloride and tri-n-octylamine in benzene.92 In general the u1 and up frequencies of the nitrato-group are independent of the solvent but strongly dependent on the metal ion and a separation of between 150 and 250 cm.-l between these two frequencies is found for many of the complexes; this separation corresponds to that for unidentate bonding in many solid nitrato-complexes.The experimental techniques which have been developed recently for the study of vibrational spectra of molten salts and the qualitative and quantitative interpretation of the spectra have been the subject of many recent reviewss3 and monograph^.^^ The molten nitrates of the alkali metals and silver have been extensively examined partly because they are non-corrosive and low-melting and have a reasonable thermal stability.They are examples of highly ionic salts with simple cations and a highly polarisable anion; the latter acts as an ideal “detector” of changes in electrostatic forces as the composition of the melt is altered. The general conclusions deduced from the vibrational spectra of molten salts are that ionic and molecular species known to exist in aqueous solution and solid phases retain their identity in the molten state. Environmental perturba- tions and ionic and molecular interactions are clearly revealed and can be understood in terms of short-range ordering in the melt and strong cation- anion forces. 5.3. Ultraviolet Spectra of the Nitrate Ion.-Many non-metallic inorganic ions in solutions give intense absorption bands in the ultra- violet region but only the absorption of the monatomic ions such as halide or quasi-monatomic ions such as hydroxide can be ascribed with certainty to anion-solvent charge-transfer transitions.94 The intense absorption of the polyatomic ions is less sensitive to environmental effects and it is accompanied by weak absorption at longer wavelengths of the n -+ u* or n + T* type. The latter bands indicate the presence of low energy u* and T* orbitals which serve as the upper levels of the strongly absorbing s8 R. E. Hester Diss. A h . 1962 23 1510. 8 9 R. E. Hester and R. A. Plane Inorg. Chem. 1964,3 769. 92 J. M. P. J. Verstegen J. Znorg. Nuclear Chem. 1964 26 25. 93 D. W. James “Selected Topics in Molten Salt Chemistry,” ed. M. Blander 94 S. F. Mason Quart. Rev. 1961 15 No.3 355. J. R. Ferraro J. Znorg. Nuclear Chem. 1959 10 319. L. I. Katzin J. Inorg. Nuclear Chem. 1962 24 245. Interscience Publishers New York 1963 ch. “Vibrational Spectra of Molten Salts”. FIELD AND HARDY INORGANIC NITRATES AND NITRATO-COMPOUNDS 387 0 + (T* and 7~ -f 7 ~ * transitions in the polyatomic ions. The absorption of the nitrate ion in aqueous solution may be compared (see Table 8) with TABLE 8. Ultraviolet spectra of various ions. Ion Solvent ~Irla,.(A> Emax. Ref. c1- H2O 1810 4 0 4 94 Br- H2O 1995; 1900 11,Ooo; 12,000 94 1870 5,ooo 94 3570;2980;2110 23; 8; 6000 94 OH- H2O NO2- H2O 3025; 1936 7; 8800 94 17.4; 8270 95 NO3- H2O NO3- molten 2830; 1950 that in molten LiC104 at 140° and with that of some other ions in aqueous solution. In molten LiClO the intense band is consideredg5 to be the electronically allowed Al’ -f E,’ ( 7 ~ -f n*) transition involving a shift of electronic charge-density from the oxygens towards the nitrogen; it is suggestedg5 that experimental difficulties may have led to this band’s being overlooked in other work on crystalline and molten alkali-metal nitrates.LiC104 6. Conclusions Anhydrous nitrates of several metals such as silver and the alkali metals have long been known but Guntz and Martin isolated the first anhydrous compounds of the transition metals Co Cu Mn and Ni with the nitrate group in 1909. The first volatile nitrato-compounds to be isolated were those of Cr and V by Schmeisser and his co-workers in 1954. Since then many metals have been shown to form slightly volatile nitrates or highly volatile nitrato-compounds (Table 4) and volatility cannot now be regarded as an unusual property of these compounds but rather as a consequence either of the intrinsic stability of the nitrate ion in an anhy- drous compound such as sodium nitrate or of the strong covalent bonding of the nitrato-groups to the central metal atom and their screening effect on the cationic charge.This covalent bonding of one or more oxygen atoms of the nitrato-group may be so strong that dissociation of the group in certain organic media occurs not at the metal-oxygen bond but at a nitrogen-oxygen bond to give an NO2. radical with powerful nitrating properties. Nitrates or nitrato-compounds of some of the rarer metals e.g. Nb Ta Hf V in addition to some of the more common ones Al Sn Zr are gener- ally not considered in current textbooks to exist in an anhydrous stoicheio- metric fonn but work over the last ten years has shown that they can be so prepared.The more extensive application of the non-aqueous methods based on N204 N205 and CINO may lead to the preparation of anhy- drous nitrates or nitrato-compounds of some of the elements for which no *6 D. W. James C. R. Boston and G. P. Smith J. Chem. Phys. 1964 40 609. 3 1 388 QUARTERLY REVIEWS such compounds are known at present e.g. Ge Pa Tc (Table 2). Detailed investigation of the infrared spectra and X-ray or neutron diffraction of the presently known anhydrous and hydrated compounds is expected to greatly increase our knowledge of the bonding of the nitrate group in them and their behaviour under many conditions. The Reviewers consider that systematic use of the terms nitrate and nitrato-compound as has been attempted in this Review will help the understanding by non-specialists of the various preparative methods used and properties observed in this field.Whilst it is possible and often con- venient to divide these compounds into two general classes on account of the criteria sumarised in Table 1 there are some properties and reactions e.g. solubility m.p. and hydrolysis which cannot at present be seen to reflect clearly the character of the metal-nitrate bond. The application of the term nitrate or nitrato-compound to a particular compound at present must be carefully decided in the light of experimental evidence from measurement of infrared spectra X-ray diffraction volatility and reactivity and compounds for which insufficient evidence is available should be regarded as nitrates for the present. The Reviewers thank Professors C. C. Addison and P. L. Robinson and Dr. J. M. Fletcher for reading the manuscript and for their many suggestions.

ISSN:0009-2681    

DOI:10.1039/QR9641800361

出版商:RSC    

年代:1964

数据来源: RSC

摘要:

SUBSTITUENT INTERACTIONS IN ortho-SUBSTITUTED NITRO- BENZENES By J. D. LOUDON (UNIVERSITY OF GLASGOW) and G. TENNANT (QUEEN’S COLLEGE DUNDEE) THE interactions to be discussed in this Review extend beyond the influence exerted by two substituents on each other’s reactivity to mutually accom- modating changes in their chemical structures. The purpose of the Review is to assemble the scattered information on one of the less familiar aspects of nitro-group behaviour. The subject embraces redox reactions intra- molecular condensations involving the nitro-group and photochemical transformations. The last topic has recently been reviewed by De Mayo and Reid1 and except for incidental mention it is omitted here. The other topics together comprise a wide range of reactions of varying complexity and whilst no attempt is made to exhaust individual examples the Re- viewers have tried to include representatives of all known and relevant types of interaction.It is common experience in aromatic chemistry that the reactions of a nitro-compound unexpectedly differ from those of its parent. The differ- ence frequently disclosed by an abnormal display of colour and an un- inviting product is often dismissed as merely regrettable evidence of the nitro-group’s ease of reduction. Yet such departures can lead to products which are otherwise inaccessible. The nitro-group’s demand for electrons may be supplied from outside the molecule or from within and for ortho- nitro-compounds the internal supply lines may run either through the molecular framework or across space. The last route is barred to para- substituted nitrobenzenes wherein significantly interactions are neither so numerous nor so varied some of them are mentioned in the sequel.There has been no systematic study of interactions between ortho- situated nitro-groups and side-chains. Information on the reactions which occur has to be gathered almost entirely from the nature of the products isolated and this limitation creates some major difficulties. Two broad paths of interaction may be differentiated according as oxygen or nitrogen of the original nitro-group becomes attached to an appropriate (not necessarily carbon) centre of the side-chain. Thereby the formation of among others oxaza- or aza-heterocycles is implied and subsequent scission of these heterocycles can account for the observed transfer of an oxygen atom from nitrogen to the side-chain (cf.p. 408) or that of a carbon- centred group from the side-chain to nitrogen (cf. p. 409). Such postulates however do not specify the exact oxidation level of the nitrogen atom when the ring-forming step occurs and analogy with para-substituted nitro- benzenes shows that acinitro- and nitroso-intermediates resulting from P. de Mayo and S. T. Reid Quart. Rev. 1961 15 393. 389 390 QUARTERLY REVIEWS adjustments within the conjugated system must be taken into general account. Moreover reduction by an externaZ reagent is admissible as a step in a reaction where it does not preclude or reverse some interaction of the nitro-group and side-chain. The questions to which these general considerations give rise can seldom be given definite answers.It follows that even the choice and arrangement of material for review must be somewhat arbitrary whilst discussion of reaction mechanisms can only be speculative and is here based on the assumption that polar mechanisms operate throughout. Basecatalysed Cyclisations One proposition to be considered is that a nitro-group can provide the electrophilic centre for additive reactions of the type exemplified by the aldol condensation. :C=O + H2CXY + :C(OH).CHXY -+ :C=CXY :C=O + H,NR + :C(OH).NHR -+ :C=NR .N=O + H,CXY + .N(OH).CHXY -+ .N=CXY .N=O + H,NR + *N(OH).NHR + -N=NR O+N=O + HZNR + 0-N(OH).NHR -+ O-N=NR O+N=O + HzCXY + 0-N(OH).CHXY -+ 0-N=CXY This proposition is neither new nor commonplace it has been advanced by various authors sometimes in less specific terms but it has never achieved prominence or wide acceptance.There are several reasons for this. Nitroso-compounds by their intermolecular reactions with primary amines to form azo-compounds or with reactive methylene compounds to form azomethines provide cogent evidence that the nitroso-group can attract and add nucleophils. But in the nitro-group it must be expected that resonance will restrict the reactivity of the nitrosyl component. The restriction is apparently severe for among intermolecular reactions it is difficult to find an example which unequivocally shows an aldol-type of interaction between nitro-group and nucleophil. The best evidence comes from intramolecular reactions where the steric factor is most favourably enlisted and a heterocyclic product is formed. Yet even here the mechanism of cyclisation is seldom clear and the re- current difficulty in interpreting the course of such interactions makes it expedient to consider first the more compelling evidence for the nitro- group’s direct participation in aldol-type reactions.Diphenyl Derivatives.-Muth and his colleagues,2 surprised by the outcome of an attempt to saponify the ester (la) established that this ester in common with some other 2‘-substituted 2-nitrodiphenyls (lb-e) C. W. Muth J. C . Ellers and 0. F. Folmer J. Amer. Chem. Soc. 1957 79 6500; C. W. Muth N. Abraham M. L. Linfield R. B. Wotring and E. A. Pacofsky J. Org. Chem. 1960 25 736. LOUDON AND TENNANT SUBSTITUENT INTERACTIONS 39 1 undergoes cyclodehydration when briefly heated with sodium hydroxide in methanol. It will be observed that among the phenanthridine derivatives so obtained (2a-e) a few are not simple dehydration products but in interesting variety (2; R‘ = CO,H) (24 (2e) are exactly the products to be expected of dehydration followed by hydrolysis.R=C02Me R=CN R=CONHz 8:; R=SOePh R=COPh (0 No OH Me OH p:o \ / R’= COBMe CO2H R’=CONHz R=CN R’=H (+ PhCOZH) R’=OH (+ PhaSOeH) 0 The stability of the diphenyl system the mild conditions of the reaction and the stark simplicity of the result provide in these examples the best available evidence for the aldol type of interaction. A conditioning feature appears to be the provision of a nucleophilic centre (e.g. a carbanion) immediately attached to the un-nitrated nucleus. Thus with diphenyls (1) having the substituents R = H OH Br Ph or C02H or having COCH,CN instead of CH $R comparable cyclisation fails whereas 2’-amino-2-nitrodiphenyl (3) yields the benzocinnoline (4) although it cyclises more slowly.Furthermore methylene reactivity in a quinoxaline derivative allows the similar cyclisation (5) -+ (6).3 o-Nitrobenzoyl Derivatives etc.-o-Nitroacetophenone and especially those of its derivatives wherein methylene reactivity is enhanced by an w-substituent should be capable of this type of cyclisation. Examples however are few and complicated perhaps because the reagents and the isatogens (8) to be expected as products are both highly sensitive. u- Nitrophenacyl chloride (7; R=C1) when heated with aqueous ethanolic potassium hydroxide yields anthranil-3-carboxylic acid ( and a rational course would be the formation and known rearrangement of 1-hydroxyisatin (9) (= 2-hydroxyisatogen).Under similar conditions o-nitrobenzoylacetone (7; R = COMe) yields isatin4 which may well be formed by hydroxide attack on the hydrated form (1 1) of 2-acetylisatogen. o-Nitrophenylpropiolic acid when heated with dilute a1 kali also yields isatin5 whereas its ethyl ester when heated with pyridine yields ethyl R. P. Barnes J. H. Graham and M. A. Salim Qureshi .I. Org. Chem. 1963 28 2890. J. D. Loudon and G. Tennant J. 1963,4268. A. Baeyer Bev. 1880,13,2254; 1881 14 1741. 392 QUARTERLY REVIEWS isatogenate (8; R = C02Et).6 In these two reactions o-nitrobenzoylacetic acid and ester respectively are plausible intermediates but direct proof is lacking. However compounds of this class (12) are converted by cold aqueous sodium hydrogen carbonate into isatogen derivatives :' these are formulated as (1 3) although their relationship to certain co-products and the observed formation of the latter by condensation of ethyl isatogenate with the reagents (12) await elucidation.Indigo is formed when o-nitrobenzoylacetic acid is treated with alkali in presence of glucose as reducing agent5 o-Nitro-a-hydroxybenzyl compounds are already at the required level of reduction and such com- pounds are intermediates in the Baeyer-Drewsen synthesis of indigo.* Thus with aqueous sodium hydroxide the pre-formed aldol (14) or its formative mixture of o-nitrobenzaldehyde and acetone rapidly yields indigo (1 7) ; moreover the aldehyde reacts in similar fashion with' 17-0x0- steroids forming via the aldols (18) products of type (21).9 For the two reactions parallel courses are conceivable leading through 1 -hydroxy- indoxyls (15) and (19) to deoxyisatogens (16) and (20) respectively but whereas the latter of these can achieve the stable isatinoid structure (21) P.Pfeiffer Annalen 1916 411 72. R. T. Coutts M. Hooper and D. G. Wibberley J. 1961 5205. A. Hassner and M. J. Haddadin Tetrahedron Letters 1962 No. 21 975. * A. Baeyer and V. Drewsen Ber. 1882 15 2856; L. E. Hinkel E. E. Ayling and W. H. Morgan J. 1932 985. LOUDON AND TENNANT SUBSTITUENT INTERACTIONS 393 by a simple proton shift the former must combine with a precursor for the ultimate formation of indigo. o-Nitrostyryl ketones do not yield indigo under the conditions of the Baeyer-Drewsen reaction but they are con- vertible by alkali into salt-like products from which the dyestuff can be obtained.* ,1 O &}- A. 0 For another type of cyclisation the relative importance of aldols and their dehydration products appears to be reversed. Krohnke et al. have shownll that aldols of type (22) undergo retroaldol scission when treated with bases. On the other hand their acetates or the derived styrylpyridi- nium salts (23) yield isatogens e.g. (24) in reaction with sodium carbonate or pyridine-diethylamine. The ease of such cyclisations is incompatible with a course which is otherwise practicable running from the styryl compounds (23) to tolanes (25) and hence in presence of pyridine to isatogens (24).6 The latter stage normally requires irradiation (although this is not necessary for all derivatives of o-nitrophenylacetylene) and Huisgen suggests12 that these routes to isatogens may converge upon an inter- mediate pyridiniumbetaine reaction (i) formed by elimination of hydrogen halide from the salt (23) or by photochemical addition of pyridine to the tolane (25).On this basis the precursor of the isatogen is a nitrosophenyl ketone (or its equivalent) whose formation involves transfer of an oxygen atom from the nitro-group to the side-chain a process which detailed mechanism apart is commonly postulated for photochemical inter- actions.lJ3 Nitrosophenyl ketones* are also possible intermediates in the Baeyer-Drewsen type of reaction but here even if real their formation from the aldols could be the result of internal transfer either of oxygen from or of hydrogen to the nitro-group. Equally uncertain is the mechan- * It has recently been shown that 2-nitrodiphenylmethanol reacts with (a) formic acid or toluene-p-sulphonyl chloride in pyridine yielding 2-nitrosobenzophenone; (b) thionyl chloride in hot chloroform yielding 5-chloro-3phenylanthranil (W.B. Dickinson J . Arnev. Chem. SOC. 1964 86 3580). lo Cf. I. Tanasescu and E. Tanasescu Bull. SOC. chim. France 1936 3 865; A. Georgescu J. prakt. Chem. 1934,139 189. l1 F. Krohnke and M. Meyer-Delius Chem. Ber. 1951 84 932 941; F. Krohnke and I. Vogt Ber. 1952 85 376. l2 R. Huisgen Angew. Chem.. Internat. Edn. 1962 2 565. l3 J. S. Splitter and M. Calvin J. Urg. Chem. 1955 20 1086. 394 QUARTERLY REVIEWS ism by which phenylisatogen is formed as a by-product when o-nitro- benzaldehyde sodium phenylacetate and sodium acetate are heated in acetic anhydride .I4 I Anthranil derivatives more commonly found under acidic conditions (p.402) have been isolated from a few base-catalysed interactions (cf. p. 391 and formation of thioanthranil p. 412) and may well be inter- mediates in others (p. 408). During a study of the Kishner-Wolff reaction Seibert15 noted the formation of anthranil (as well as o-nitrotoluene) by the action of dilute alkali on the hydrazone of o-nitrobenzaldehyde [reaction (ii)]. A similar course initiated by extraction of an a-proton can explain the conversion of 3,4-dimethoxy-2-nitrophenylacetic acid by alkali into the corresponding anthranil-3-carboxylic acid (iii).16 Here the hetero- cyclic intermediate (the equivalent of an o-nitrosornandelic acid) may undergo direct dehydration or conceivably can yield the anthranil after isomerisation to an hydroxyaminophenyl ketone [cf.(iii)]. A simple variant of the latter process then accounts for the formation of isatinsl' from certain azlactones as in reaction (iv). While the presence of methoxyl adjacent to the nitro-group appears to assist formation of isatins in this L W J l4 Cf. P. Ruggli E. Caspar and B. Hegedus Helv. Chim. Acta 1937 20 250. l5 W. Seibert Ber. 1947 80,494; 1948 81 266. l8 J. M. Gulland J. 1931 2872. l7 H. Burton and J. L. Stoves J. 1937 402. LOUDON AND TENNANT SUBSTITUENT INTERACTIONS 395 way it is noteworthy that isatin itself is formed by the action of aqueous alkali on o-nitrophenylpyruvic acid.18 The formation of anthranil derivatives by a different type of mechanism is inherent in a re-interpretation of the base-catalysed transformations of compounds of type (26).These compounds prepared by halogenating arylhydrazones of o-nitrobenzaldehyde eliminate hydrogen halide yielding products which were formulated by Chattaway and Walker19 as triazine oxides (28) but are regarded by Gibson20 as anthranil 1-oxides (27) formed through cyclisation of transient nitrilimines. Anthranil 1 -oxides are very elusive compounds but according to Szmant and Harmuth21 3-phenyl- anthranil 1-oxide may be the product isolated from the reaction of o- nitrobenzoic acid with trifluoroacetic anhydride and boron trifluoride. o-Nitrobenzyl Derivatives.- [See reactions (iii) and (iv) above. 3 Proton removal from o-nitrophenylacetone is more likely to occur at the doubly activated methylene than at the terminal methyl group and a condition such as shown in (29; R = Me) may help to explain why there is no record of cyclisation to a quinoline derivative here or with 2,4-dinitrophenyl- acetone or 2,4-dinitrophenylacetoacetic esters (30; R = H).Yet a closely similar compound (30; R = Ph) undergoes cyclodehydration. Zaki and Iskander who noted22 this reaction regarded the product as the naphtha- lene derivative (32) but the quinoline structure (31) is in accord with the ester behaviour and is supported by the infrared spectra which show that the ester and the derived acid are of the salicylic type.23 o-Nitrophenylacetamide likewise does not undergo cyclodehydration but here the nucleophilic potentialities of the amino-group should be released in the anion (29; R = NH,) and indeed o-nitrophenylcyano- acetamide (33) which is better equipped to provide an anion is cyclised to the cinnoline (34) by warm aqueous sodium hydroxide.24 The amide (33) la A.Reissert Ber. 1897 30 1030. l9 F. D. Chattaway and A. J. Walker J. 1927 323; cf. J. G. Ericson in "The Chem- 2o M. S. Gibson Tetrahedron 1962 18 1377. 21 H. H. Szmant and C. M. Harmuth J. Amer. Chem. Soc. 1959 81 962. 22 A. Zaki and Y. Iskander J. 1943 68. 23 J. P. Cairns J. D. Loudon and A. S. Wylie unpublished work 24 J. P. Cairns Ph.D. Thesis Glasgow 1964. istry of Heterocyclic Compounds," Interscience New York 1956 Vol. 10 p. 27. 396 QUARTERLY REVIEWS can be prepared by base-catalysed Smiles rearrangement followed by hydrolysis from the sulphonamide (35; R = CN).25 The same procedure applied to the sulphonamide (35; R = Ph) leads to 3-phenylindazole (36a) possibly through ring-contraction of the cinnoline (36) (for analo- gous contraction of triazines see p.400). In the next set of nitro-compounds each of the a- and P-carbon atoms of the side chain carries a hydrogen atom which is acidic but although the degree of acidity at the respective centres varies widely the compounds are remarkably consistent in their ability to yield derivatives of l-hydroxy- indole. Nevertheless in certain cases and in a strongly alkaline environ- ment this type of product is accompanied by another derived from 1 -hydroxyquinoline. (37) 25 T. Naito and R. Dohmori cf. Chem. Abs. 1954 48 10647. LOUDON AND TENNANT SUBSTITUENT INTERACTIONS 397 o-Nitrobenzylmalonic acid26 and ethyl o-nitrobenzylacetoacetate (37)27 each reacts with aqueous alkali to form 1 -hydroxyindole-2-carboxylic acid the latter reaction occurring so readily that the ester (38) is also isolated.The allied nitrile (39) and the derived amide (39; CONHz for CN) respect- ively yield the 3-cyano- and 3-carbamoyl-derivatives of the indole (38) in smooth reactions effected by sodium carbonate.28 But the nitrile (39) reacts with hot ethanolic potassium hydroxide to form the 1-hydroxyquinoline (40) and this product may be obtained directly by the action of potassium cyanide on diethyl o-nitrobenzylidenemalonate in ethanol. Moreover as exemplified in reactions ( ~ i i ) ~ ~ and (~iii)~O where the first-formed adducts (as 39) have not been isolated other o-nitrobenzylidene compounds react with potassium cyanide affording mixed products of the indole and quinoline types.Two derivatives of a-benzyl-o-nitrobenzyl cyanide have been examined:2s both are methoxylated in the nitrobenzene ring and both yield indoles in reaction [cf. (ix)] with potassium hydroxide despite the feeble activation of the /3-methylene centre of the side-chain. There are too many uncontrolled variables to warrant extensive dis- cussion of these reactions. Broadly a highly reactive a-hydrogen atom in the side-chain of the nitro-compound appears to be a facilitating factor even for indole formation but whether it operates before or after the ring-forming step is uncertain. Reduction presumably by the reaction medium is involved in the formation of the quinolines. Thus the product of reaction (vi) requires a hydroxylamine precursor whereas the nitro- compound (39) can only provide a nitroso-intermediate for instance by base-catalysed transfer of protons from its side-chain.On the other hand direct external reduction of the nitro-compound (39) by zinc and ammo- nium chloride designed28 to yield the corresponding hydroxylamine and 26 A. Reissert Ber, 1896 29 639. 27 S. Gabriel W. Gerhard and R. Wolter Ber. 1923 56 1024. 28 J. D. Loudon and I. Wellings J. 1960 3462. 2 9 J. D. Loudon and G. Tennant J. 1960,3466. 30 J. D. Loudon and A. C. Mackay unpublished work. 4 398 QUARTERLY REVIEWS hence the 3,4-dihydroquinolone (41) yields instead the de-cyanoquinoline (40; H for CN). If on this evidence the dihydride can be rejected as an intermediate in reaction (vi) then the course depicted for this particular reaction becomes a likely one.Among the reactions so far described there are several examples of ring- closure through formation of a nitrogen-nitrogen bond (pp. 391 395). Although 2-nitrobenzylamine is well known it shows no tendency to under- go cyclodehydration. The simple amino-nitrile (43 ; H for Ph) is not known but reactions likely to yield it yield 2-nitrosobenzoic acid instead (p. 408). However the related anilino-nitrile (43) is known and is readily cyclised by sodium carbonate to the indazole 1-oxide (44)31 and shares this be- haviour with a number of its derivative^.^^ A similar course via the anilino- nitrile in all probability accounts for the formation33 of the indazole oxide (44) from o-nitromandelonitrile and aniline in acetic acid. Effectively there- fore 2-nitrobenzylidene-anilines (42 ; R = H) are convertible into inda- zoles as (44) via adducts of type (45)34 wherein again the a-cyano- substituent seems to facilitate cyclisation.2-Nitrobenzylidene-anilines are themselves isomerised to 2-nitrosobenzanilides (45) by light,35 but not by purely chemical means. On the other hand more highly nitrated analogues e.g. (42; R = NO2) are isomerised by sodium carbonate in ethanol and although the products have been given heterocyclic structures such as (46)36 there is an implied close relation to 2-nitrosobenzanilides e.g. (45 ; R = NO 2) through ring-chain tautomerism. Confirmation is desirable for these structures and also for the course depicted for the reaction (x)~’ wherein the alleged product is indicative of an a-anilino-ester as precursor [cf.(43) -+ (44)] whereas the reagents suggest an anil as inter- mediate. CN CN OH 31 A. Reissert and F. Lemmer Ber. 1926 59 351. s* L. C. Behr J. Amer. Chem. SOC. 1954,76,3672; J. Org. Chem. 1962,27,65. *3 G. Heller and G. Spielmeyer Ber. 1925 58,834. 34 Cf. K. Akashi Chem. Abs. 1949 43 7934. ss F. Sachs and R. Kempf Ber. 1902,35,2704. s6 S . Secareanu and I. Lupas Bull. SOC. chim. France 1933 [4] 53 1436; 1934 [51 1,373; 1935 [ 5 ] 2,69. Cf. L. Jacobs in “Heterocyclic Compounds” Wiley New York 1957 Vol. 5 p. 169. 37 I. Tanasescu and E. Tanasescu Brill. SOC. chim. France 1935 2 1016. LOUDON AND TENNANT SUBSTITUENT INTERACTIONS 399 F02Et C02Et (X) ocH2 \ NO2 + C ) N o N M e 2 [ cN'c6H4'] / 0 Derivatives of o-Nitroaniliae.-The prototype of this class of reaction was discovered by Nietzki and Braunschweig who that treatment of o-nitrophenylhydrazine (47) with aqueous alkali did not liberate hydrazine as expected but gave the salt of l-hydroxybenzotriazole (48).The reaction has been extended in various ways. 2,4-Dinitrophenyl- hydrazine is conveniently cyclised to 1 -hydroxy-6-nitrobenzotriazole by hydrazine hydrate39 which may thus be used to prepare hydroxybenzo- triazoles in one operation from o-ha loge no nitro benzene^^^ or from o- dinitr~benzenes.~~ The comparable use of phenylhydrazine or cyclisation of pre-formed 2-nitrohydrazobenzenes yields N-oxides of type (49),42 but in this series reduction to 2-substituted benzotriazoles is often incurred ; moreover cyclisation is variously effected e.g. by alkali or by hot acetic acid and reductive cyclisation occurs in acetic acid in presence of potas- sium iodide.A particular example of this type of cyclisation is given in reaction (xi) although other interpretation of its course has been sug- g e ~ t e d . ~ ~ By contrast methylhydrazine,'which with reactive chloronitro- or dinitro-benzenes affords 1 l-disubstituted h y d r a z i n e ~ ~ ~ ~ leads to triazoles of type (50).41 The internal-aldol mechanism offers an attractively simple interpreta- tion of how these N-oxygenated benzotriazoles are formed but the power- ful reducing properties of hydrazines are also much in evidence. Thus 2,4-dinitrophenylhydrazine yields 46 3,3'-dinitroazoxybenzene m-dinitro- benzene and 1 -hydroxy-6-nitrobenzotriazole in proportions which vary with the pH of the medium and are minimal for the heterocycle at low concentrations of basic condensing agent.While for each of these products an oxidised hydrazine of the type ArN=NH (Ar variously substituted) is a conceivable precursor it cannot in so far as it leads to rn-dinitrobenzene 38 R. Nietzki and E. Braunschweig Ber. 1894 27 3381. 39 T. Curtius and M. Mayer J. prakt. Chem. 1907 76 369. 40 E. Muller and G. Zimmermann J. prakt. Chem. 1925,111,277. 41 B. Vis Rec. Trav. chim. 1939 58 847. 42 Cf. F. R. Benson and W. L. Savell Chem. Rev. 1950 46 1; cf. A. Angeletti 43 R. Grashey Angew Chem. Internat. Edn. 1962 1 158. 44 K. Fries W. Franke and W. Bruns Annalen 1934 511,264; J. J. Blanksma and 45 B. Vis Rec. Trav. chim. 1939,58 387. 46A. K. Macbeth and J. R. Price J. 1934 1637; cf. 0. M. Shemyakina B. M. Gazzetta 1923 53 672. M. L. Wackers Rec.Trav. chim. 1936,55 655. Bogoslovskii and M. M. Shemyakin Chem. Abs. 1957 51 5057. 400 QUARTERLY REVIEWS Br- OHC b / \ 0 be derived by internal oxidation-reduction of the hydrazine yet a com- peting internal process by supplying an o-nitroso-substitueiit in the oxi- dised hydrazine could still be a route to the triazole. 3-Aminobenzotriazole 1-oxide (52; R = H) is formed rapidly and almost quantitatively when o-nitrophenylguanidine (51 ; R = H) is warmed with dilute alkali. Neither ammonia nor aqueous sodium carbonate is effective as condensing agent and cyclisation of the analogous urea derivative (54) -+ (53) requires stronger alkali e.g. 10 % potassium hydroxide. These reactions were discovered by Arndt4' and were later extended48 to include cyclisation of the guanidine (51; R = Ph) of o-nitrophenylthiourea (51; S for NR) and various aryl-substituted guanidines (as 51 ; R = H).49 Since the process often occurs with deepening followed by fading of colour Arndt suggested that a salt of the pseudo-nitro-compound might be an intermediate here again removal of the a-proton may assist cyclisation as in the nitrophenylacetamide (33).On the practical side it should be noted that while these 1-oxides are formed in presence of alkali they are themselves affected by prolonged exposure to hot alkali.50 In this respect the 3-amino- are more susceptible than the 3-hydroxy-compounds but both undergo ring-contraction to benzotriazole derivatives [reaction (xii) 1. o-Nitrophenylbenzamidines are also cyclised by alkali yielding 3- arylbenzotriazine 1 -oxides [reaction (xiii)J51 Other variants include the use 47 F.Arndt Ber. 1913 46 3522. 48 F. Arndt and B. Rosenau Ber. 1917 50 1248. 49 F. J. Wolf K. Pfister R. M. Wilson and C. A. Robinson J. Amer. Chem. Suc. 1954 76 3551; F. J. Wolf R. M. Wilson K. Pfister and M. Tishler ibid. 1954 76 4611; J. Jiu and G. P. Mueller J. Org. Chem. 1959 24 813. 6o J. A. Carbon J. Org. Chem. 1962,27,185. 61 R. F. Robbins and K. Schofield J. 1957 3186; R. Fusco and G. Bianchetti Chem. Abs. 1959,53 9243. LOUDON AND TENNANT SUBSTITUENT INTERACTIONS 401 o-" _Ic %,I- H ty \ (xii) ' I 7' - t 0 co-x of o-nitrobenzenesulphonyl derivatives of g ~ a n i d i n e ~ ~ and urea53 respect- tively to provide in a combined operation [cf. (xiv)] the reagents and con- ditions for cyclisation. Applications to the synthesis of fused heterocyclic systems are exemplified by reactions ( x v ) ~ ~ and ( x v ~ ) ~ ~ and further examples may be implicit in reactions still obscure which occur when 5-chloro-3- methyl-l-(2,4-dinitrophenyl) pyrazole is heated with ammonia or primary a r n i n e ~ .~ ~ Gc", 0 0 0 Two reactions (xvii; R = COPh)4 and (xviii; R = Me or Ph),57 respectively illustrate extension of the Nietzki and Arndt cyclisations to cases wherein the side-chain provides the nucleophil in the form of a carbanion. The potentialities in such extensions are virtually unexplored 52 H. J. Backer and H. D. Moed Rec. Trav. chim. 1947 66 689. 53 H. J. Backer and J. Groot Rec. Trav. chim. 1950 69 1323. 54 J. A. Carbon and S. H. Tabata J. Org. Chem. 1962 27 2504. 56 E. Lieber T. S. Chao and C. N. R.Rao J. Org. Chem. 1957 22 654. 66 C. A. Rojahn and H. Fegeler Ber. 1930,63 2510. 57 G. Tennant J. 1963 2428; R. Fusco and S. Rossi Chimica e Industria 1963 4'5 834; cf. Y. Ahmad M. S. Habib and Ziauddin Tetrahedron 1964 20 1107. 402 QUARTERLY REVIEWS but it is particularly noteworthy that as in reaction (xix) N-methylation of the reagent is found58 to enhance both the yield and the apparent rate of cyclisation. However the recently cyclisation of N-benzyl-o- nitroaniline to 1 -hydroxy-2-phenylbenziminazole reaction (xvii ; R = Ph H for Me) invites comparison with cyclisations of type (ix; p. 397) wherein a mobile a-hydrogen atom appears to compensate for feeble activation at the P-methylene centre of the side-chain. (xviii) (xi x) Acid-catalysed Cyclisations Anthranil derivatives are often products or intermediates in acid- catalysed transformations of o-nitrobenzyl compounds.The parent anthranil is formed by the action of hydrochloric acid on o-nitrobenzyl- idene dimercurichloride it is almost certainly an intermediate in the con- version61 (reaction xx) of o-nitrophenylacetic acid by hot acetic anhydride into the internal anhydride of N-acetylanthranilic acid ; moreover 6-nitro- anthranil is obtained when 2,4-dinitrophenyl-a~etone~~ or -acetic acide3 is heated with concentrated sulphuric acid and amides of the appropriate anthranil-3-carboxylic acids resuW4 from heating certain derivatives of o-nitrophenylacetamide with phosphorus pentachloride in benzene. o-Nitrophenylglycidic acid ( 5 9 possibly reacting via o-nitroplienyl- acetaldehyde yields a mixture of anthranil and its aldehyde (56) when distilled in steam or heated in acetic acid.65 Heating alone sometimes leads to anthranils as in the cyclisation66 of ethyl o-nitrophenyl-malonate 58 G.Tennant J. 1964 2666; cf. R. Fusco and S . Rossi Gazetta 1964 94 3. 59 G. W. Stacey B. V. Ettling and A. J. Papa J. Org. Chem. 1964 29 1537. 6o Kalle and Co. Chem. Zentr. 1908 11 210; A. Reissert Ber. 1907 40,4209. 61 G. N. Walker J. Amer. Chem. SOC. 1955 77 6698. 62 S. S. Joshi and I. R. Gambhir J. Amer. Chem. Soc. 1956,78,2222; J. Org. Chem. 63 H. G. Garg J. Org. Chem. 1962 27 3683. 64 D. H. Hey and A. L. Palluel J. 1956 4123. 65 A. Schillinger and S . Wleugel Ber. 1883 16 2222. 66 C. A. Grob and 0. Weissbach Helv. Chim. Acta 1961,44 1748. 1961 26 3714. LOUDON AND TENNANT 1 SUBSTITUENT INTERACTIONS 403 and -cyanoacetate to 3-ethoxycarbonyl- and 3-cyano-anthranil respectively.Moreover anthranils are probably intermediates of thermal or acid- catalysed reactions which lead to acridones (see below). Thus aluminium chloride catalyses cyclisation of o-nitrodiphenylmethane to a product which is form~lated~~ as acridine 10-oxide (57) whereas thermal cyclisa- tion6* affords acridone (58) possibly via 3-phenylanthranil (see below). Nevertheless the ease of anthranil formation is disconcertingly varied 1 for instance neither o-nitrobenzyl cyanide nor esters of o-nitrophenylacetic acid are cyclised under the conditions of reaction ( x x ) . ~ ~ o-Nitrobenzaldehyde reacts with aromatic compounds affording triaryl- me thane^,^^ 3-arylanthranil~~~~ or ac~idones,~~ 9 7 2 depending on the acidic environment and on the reactivity of the aromatic compound.Although 2-nitrodiarylmethanols are undoubtedly intermediates there is no con- vincing evidence that they have isolated from such reactions:73 on the other hand 2-nitrobenzophenones are not uncommon b y - p r o d ~ c t s ~ ~ ~ ~ and are presumably formed by oxidation of the methanols. For the reaction which occurs at ordinary temperature between o-nitrobenzaldehyde and benzene in presence of concentrated sulphuric acid a plausible course71 is given by the sequence (xxi). In this it will be noted that the step from 2- nitrosobenzophenone to 3-phenylanthranil requires a reducing agent which may well be the methanol since o-nitrobenzophenone is also isolated. In general N-unsubstituted acridones must also be reckoned among the products of such reactions thus N-hydroxyacridones7 is formed as a by- product in reaction (mi) whilst acridone is the principal product when the condensation (mi) is effected by hot polyphosphoric Moreover it is known that the 3-arylanthranils are converted into acridones by traces 67 M.Freund Sitzungsber. Akad. Wiss. Wien. 1896 105 381; A. Kliegl and A. Brosamle Ber. 1936 69 197. A. Kliegl Ber. 1909 42 591. 6 0 J. E. Driver and S. F. Mok J. 1955 3914. 70 A. Kliegl Ber. 1908 41 1845; cf. J. C. E. Simpson and 0. Stephenson J. 1942 71 A. Albert “The Acridines” Arnold and Co. London 1951. 72 I. Tanasescu M. Ionescu I. Goia and H. Mantsch Bull. Soc. chim. France 73 J. D. Loudon and G. Tennant J. 1962 3092. 74 K. Lehmstedt Ber. 1932 65 999. 353. 1960 4 698.404 QUARTERLY REVIEWS of nitrous acid7j whose presence therefore whether in~idental'~ or con- t r i ~ e d ~ ~ greatly affects the proportions of products. Anthranils are almost exclusively the heterocyclic products when o-nitrobenzaldehyde condenses with the more reactive types of aromatic reagent. Thus with aniline the reaction effected by zinc chloride is reported" to give a mixture of the triarylmethane (59; R = R = NH2) and the 3- arylanthranil (60; X = H) whereas in hydrochloric acid-acetic acid it affords the same anthranil together with the 5-chloro-derivative (60; X = Cl).78 Dimethylaniline likewise reacts to give the chlorinated anthranil (60; X = Cl NMe for NH2).79 The allied reactions of phenols are dis- cussed below and are especially interesting because of features held in common with some reactions of reactive rnethylene compounds.The following reactions all occur at low temperatures (0-20") and in acetic acid or ether as solvent. Therein phenol and o-nitrobenzaldehyde are converted by sulphuric acid into the triarylmethane (59; R = R = OH),69 by hydrogen chloride into 5-chloro-3-p-hydroxyphenylanthranil 75 E. Bamberger Ber. 1909 42 1716. 76 F. R. Bradbury and W. H. Linnell J. 1942 377. 77 I. Tanasescu and A. Silberg Bull. SOC. chim. France 1932,51 1357; cf. I. Tanasescu 78 S. Seczreanu and A. Silberg Bull. SOC. chirn. France 1936,3 1777. T 9 T. Zincke and W. Prenntzell Ber. 1905 38 4116. and M. Suciu ibid, 1937 4 245. LOUDON AND TENNANT SUBSTITUENT INTERACTIONS 405 (61),80 and by hydrogen bromide into a mixture of 3-p-hydroxyphenyl- anthranil and its 5-bromo-deri~atives.~~ A more complete contrast in the behaviour of the two hydrogen halides as condensing agents is found73 in the reactions of phenol with 4-bromo- or 4-chloro-2-nitrobenzaldehyde these reagents with hydrogen chloride yield 5,7-dihalogentated anthranils e.g.(62) through entry of chlorine whereas with hydrogen bromide they react forming 5-halogenated anthranils without entry of bromine. It is noteworthy also that when the aromatic component is a reducing phenol namely quinol even hydrogen chloride effects the reaction without entry of chlorine the product from o-nitrobenzaldehyde being 3-(2,5-dihy- dro~ypheny1)anthraniI.~~ Of three pre-formed carbinols (63 ; R = Me) is not appreciably affected by hydrogen chloride under the reaction condi- tions (63; R = Ph) reacts slowly and incompletely giving 5-chloro-3- phenylanthranil whereas (63; R = p-C,H,*OH) readily yields the an- thranil (61).*l (See also footnote p.393.) Hydrogen chloride reacts with o-nitrobenzaldehyde and ethyl aceto- acetate to form the 6-chloro-1-hydroxy-4-quinolone (64).82 Analogous products are formed when the last reagent is replaced by acetylacetone benzoylacetone or diethyl acetonedicarboxylate and 6,8-dichloro- 1 - hydroxyquinolones result from the use of 5-chlor0-2-nitrobenzaldehyde.~~ Nitrobenzylidene derivatives of the methylene reagents behave as inter- mediates but those derived from acetone deoxybenzoin and notably ethyl benzoylacetate fail to cyclise. Hydrogen bromide again provides a contrast by effecting the reaction without inserting a halogen substituent ; and again through the presence of quinol ethyl o-nitrobenzylideneacetoacetate may be cyclised even by hydrogen chloride to the halogen-free 1-hydroxy- quinolone (64; H for Cl).73 HCL C02Et W C O M e - OH The 3phydroxyphenylanthranils and 1-hydroxy-4-quinolones may both be regarded as cyclodehydration products of appropriate o-hydroxyl- aminophenyl ketones.Their formation seems to require at an early stage a carbonium type of intermediate [cf. sequence (xxii)] whereby an oxygen atom of the nitro-group can be linked to the benzylic carbon atom. Then 8o T. Zincke and K. Siebert Ber. 1906 39 1930. *l J. P. Cairns unpublished work. J. D. Loudon and I. Wellings J. 1960 3470. 406 QUARTERLY REVIEWS or after transformation to a nitrosophenyl ketone reduction to the hydro- xylamino-level must occur.The evidence suggests that hydrogen bromide or quinol can effect this reduction whereas hydrogen chloride cannot readily do so but supplies the necessary electrons through entry of chloride ion into the nucleus. For this stepwise process a part-analogy is available in the conversions3 of nitrosobenzene into p-chlorophenylhydroxylamine by hydrogen chloride but a more concerted mechanism e.g. (65) is not excluded. By the action of concentrated sulphuric acid the acetylenes (66; R = C0,Et) and (68) are converted respectively into the isotogens (67; R = COeEt) and (69) and although o-nitrophenylpropiolic acid (66; R = C02H) thereby yields isatin the unstable isatogenic acid (67; R = C02H) is the probable intermediate.s4 Under similar conditions p-nitrophenyl- propiolic acid yields p-nitrobenzoylacetic acids5 and governed by the powerful orienting influence of two nitro-groups 2,4-dinitrotolane [as (66; R = Ph)] affords the deoxybenzoin (70).88 However hydration of the acetylene (66; R = C02Et) does not appear to be a significant step in forming the isatogenic ester since the hydration product ethyl o-nitro- benzoylacetate is hydrolysed rather than cyclised by concentrated .sulphuric acid.87 Presumably therefore cyclisation involves some direct interaction between the nitro-group and the acetylenic side-chain [cf. sequence (xxiii)]. Interaction as shown in sequence (xxiv) is the mechanism proposedss 0 88 E. Bamberger H. Busdorf and B. Szolayski Ber. 1899 32 210. 84 A. Baeyer Rer. 1881 14 1741; 1882 15 50. 85 C . Engler and 0.Zielke Ber. 1889 22 203. 86 P. Pfeiffer Annalen 1916 411 72. E. R. Needham and W. H. Perkin J. 1904,85 148. J. A. MooreandD. H. Ahlstrom J. Urg. Chem. 1961 26 5254; but cf. E. C. Taylor and D. R. Eckroth Tetrahedron 1964,20,2057. LOUDON AND TENNANT SUBSTITUENT INTERACTIONS 407 for the conversionsg of o-nitrobenzoyldiazomethane into 1 -hydroxyisatin by mixed formic and acetic acids. Thus the homologous diazoketone (71) is found to yield the normal ester (72) instead of the analogous 1,3-dihy- droxycarbostyryl (73) and the difference is plausibly explained by the greater difficulty of forming a 7-membered in place of a 6-membered cyclic intermediate it would be less easily explained if the initial step in the mechanism were direct nucleophilic attack on the nitro-group by the w-carbon atom of the side-chain.uxiv) ~t \ NO2 -@o OH Among derivatives of o-nitroaniline a remarkable type of interactiong0 is represented by the reactions (xxv) and (xxvi) which are effected by zinc chloride in hot acetic anhydride. Acid-catalysed cyclisation of 2-nitro- hydrazobenzenes has already been mentioned (p. 399 ; ref. 42) and operates in the formationQ1 of benzotriazole derivatives from 4-alkyloxybutan-2- ones and 2,4-dinitrophenylhydrazine in presence of hydrochloric acid. Uncyclised Products When heated with aqueous alkali o-nitrotoluene yields anthranilic acid (and o-toluidine),02 2,4-dinitrotoluene yields 4-nitroanthranilic acid,93 and 2-nitrotoluene-4-sulphonic acid aided by its solubility smoothly yields 4-sulphoanthranilic acid. 92 However all nuclear substituted 0- P.van Romburgh and H. W. Huyser Rec. Trav. chim. 1930,49 165; cf. “Hetero- cyclic Compounds,” ed. R. C. Elderfield Wiley and Sons Inc. New York 1957 Vol. 6 8s F. Arndt B. Eistert and W. Partale Ber. 1927 60 1364. p. 493. O1 H. J. Shine L. Fang H. E. Mallow N. F. Chamberlain and F. Stehling J. Org. Ba Kalle and Co. Chem. Zentr. 1908 1 1345; L. Preuss and A. Binz 2. angew O3 K. G. Rosdahl Chem. A h . 1950,44,9480. Chem. 1963,28,2326. Chem. 1900 13 385. 408 QUARTERLY REVIEWS nitrotoluenes do not behave in this way for some e.g. the 5-6- and 7,6-methylnitroquinolinesg* more closely resemble the p-nitrotoluenes which commonly undergo reduction to azoxy-compounds or oxidative coupling to dinitrodiaryl-ethanes or -stilbenes. 95 (xxvii) Although claimsQ2 to have isolated anthranil and its precursor 0- nitrosobenzyl alcohol from the reaction with o-nitrotoluene have been questionedQ6 and other courses proposed,97 experiments using l80 labels indicateQ8 that only one of anthranilic acid’s two oxygen atoms comes from the reaction medium.The assumption that the second oxygen atom is internally transferred from the nitro-group then allows a mechanism pro- posed by Sch011,~~ to be elaborated as in (xxvii). Although fairly stable o-nitrobenzyl alcohol is variously convertedg9 by alkali into products which include anthranil o-aminobenzaldehyde anthranilic acid the azo-compounds (74; X = Y = C02H) and (74; X = CH0,Y = C02H) and the presumably derived indazoles of type (75). Some of these products are also obtained from o-nitrobenzaldehyde which can furnish o-nitrobenzyl alcohol as part-product of a normal Cannizzaro reaction with aqueous alkali.loO As mentioned later (p.412) o-nitro- phenylmethanethiol reacts in aqueous alkali to form thioanthranil. It is possible that some of these changes begin with a step similar in its effect to the light-catalysed transformation of o-nitrobenzyl alcohol into o-nitrosobenzaldehyde,lol but this has never been demonstrated as a purely chemical process. On the other hand the comparable conversion of o-nitrobenzaldehyde into o-nitrosobenzoic acid is known both as a photo- chemicaPO2 and chemical lo3 reaction. The latter is best effected by the joint action of ammonium cyanide and hydroxide on the aldehyde in aqueous O4 R. Huisgen Annalen 1948 559 101. 95 0. Fischer and E. Hepp Ber. 1893,26 2231. 96 R.Scholl Monatsh. 1913 34 1011. s7 G. Lock Ber. 1940,73,1377; G. A. Russell and E. G. Janzen J. Amer. Chem. SOC. @* I. I. Kukhtenko Chem. Abs. 1960 54 24619. 99 G. Lock Ber. 1930,63 855; P. CarrC Compt. Reiid. 1905 140,663. loo T. A. Geissman Org. Reactions 1944 11 112. lol E. Bamberger Ber. 1918,51 606. lo2 G. Ciamician and P. Silber Ber. 1901 34 2040. lo3 G. Heller J. prakt. Chem. 1923 106 1. 1962,84,4153. LOUDON AND TENNANT SUBSTITUENT INTERACTIONS 409 ethanol. o-Nitromandelonitrile is a potential intermediate and yields o-nitrosobenzoic acid by reaction with ammonium hydroxide :lo4 moreover by irradiation in liquid hydrogen cyanide it yields the same acid and not the conceivable alternative o-nitrosobenzoyl cyanide.lo5 It is claimedlo3 that 2,4-dinitrobenzaldehyde reacts with ammonium cyanide to form 4-nitro- 2-nitrosobenzoic acid but with potassium cyanide in acetic acid yields the isomeric 2-nitro-4-nitrosobenzoic acid.o-Nitromandelic acids are also susceptible to oxidation-reduction for instance the acid (76) forms the appropriate azobenzoic acid as (74; X = Y = C02H) when heated in nitrobenzene and yields a mixture of this product with the corresponding azoxybenzaldehyde when warmed in aqueous alkali.lo6 (74) a-Phenyl-o-nitrocinnamonitrile reacts exothermally with potassium hydroxide in warm aqueous ethanol forming N-benzoylanthranilic acid.lo7 A plausible course ( ~ x v i i i ) ~ ~ involving formation and hydrolytic scission of an indolinone derivative has some analogy in the hydrolysislo8 of the adduct which is obtainable from methyl isatogenate and methanol [cf.(xxix)]. However methyl a-cyano-o-nitrocinnamate is merely hydro- lysed to the corresponding cyano-acid by potassium hydroxide here for a comparable redox reaction cyanide ion appears to be necessarylo9 and leads to the formation of N-oxalylanthranilic acid [cf. (xxx)]. Reactions which occur under acidic conditions include the formation of 3,5-dibromoanthranilic acid by the action of bromine on o-nitrotolu- ene,ll0 and that of N-o-nitrobenzoylanthranilic acid by heating o-nitro- benzaldehyde with polyphosphoric acid in the presence of anthracene (which furnishes some anthraquinone).lll o-Nitrophenylethylene oxide yields the very sensitive o-nitrosobenzoylmethanol reaction (xxxi) on careful treatment with acid or the acetate of this methanol in reaction with ace tic anhydride.112 An intriguing example of an ortho-interaction is provided by hydrogena- tion of o-nitrobenzonitrile over a platinum or palladium catalyst in methanol [reaction (xwrii)]. The product isolated in high yield is o- aminobenzamide which however cannot arise by reduction with incidental lo4 G. Heller Ber. 1906 39 2334. lo5 F. Sachs and S. Hilpert Ber. 1904 37 3425. lo6 G. M. Robinson J. 1917 111 109. lo' R. Pschorr and 0. Wolfes Ber. 1899 32 3399. lo8 G. Heller and W. Boessneck Ber. 1922 55 474. lo9 J. D. Loudon and G. Tennant unpublished work. ll1 M. Ionescu H. Mantsch and I. Goia Chern. Ber. 1960 93 2063. 112 F. Amdt B. Eistert and W. Partale Ber. 1928 61 1107. P. Greiff Ber. 1880,13,288; W. Gluud Ber 1915 48 432. 410 QUARTERLY REVIEWS (xxix) H+ - MeOH &OMe C02Me OH OH- - hydrolysis since there is no incorporation of lSO when the reaction is conducted in presence of [180] The oxygen of the amide function must therefore come from the nitro-group ultimately but not necessarily directly.It may also be relevant that the same reaction (xxxii) can be effected by hydrazine in presence of Raney nickel,l14 and that hydrogena- tion of the ester (77) is said115 to produce anthranilic acid. Sulphur-containing Compounds Whilst examples of substituent interaction are fairly common among derivatives of o-nitroaniline they are lacking among derivatives of o-nitro- phenol and scarce among those of o-nitrothiophenol. There are indeed a number of reactions which give rise to products wherein a thio-substituent has been oxidised and a nitro-group reduced but the two changes are not always in balance and the reactions are usually complex.o-Nitrothiophenol (78; R = H) when heated with sodium pentyloxide in pentyl alcohol affords the sodium salt of the sulphinic acid (79; R = R = H) but under the same conditions the thio-ether (78; R = Me) is llS H. Moll H. MUSSO and H. Schroder Angew. Chem. Internat. Edn. 1963 2 212; C. W. Jefford Diss. Abs. 1963,834. 114 K. Butler and M. W. Partridge J. 1959 2396. ll8 R. T. Coutts M. Hooper and D. G. Wibberley J. 1961,5058. LOUDON AND TENNANT SUBSTITUENT INTERACTIONS 41 1 simply reduced to the corresponding azo- or azoxy-compound without oxidation of the sulphur atom.lle In alkaline media (o-nitropheny1thio)- acetic acid is relatively stable whereas the acetophenone (78; R = CH,.COPh) yields substantial amounts of 2-benzoylbenzothiazole (8O),ll7 but the reduction here involved could either precede or follow the cyclisa- tion step and is probably effected by o-nitrothiophenol which is always formed in a competing reaction.The isomerisation of o-nitrobenzene- sulphenanilide (81) to phenyl azobenzenesulphinic acid (82) is an example of a balanced oxidation-reduction and the suggested course of the reac- tion is outlined.ll* However in the 2,4-dinitrobenzenesulphenyl series (83; R = NHPh or OMe) the aminosulphonic acid (84) is one of the ob- served products and in a strongly acidic medium can be obtained in 70% yield from the methyl sulphenate (83; R = OMe).l19 o-Nitrobenzenesulphenyl chloride (78; R = Cl) also reacts in hot aqueous methanol to form orthanilic acid (84; H for and in a vigorous reaction with hydrogen fluoride yields the azobenzenesulphonyl fluoride (79; R = R = F) among other products.121 According to Kaluza and Perold122 the reaction (83; R = Cl) -+ (84) in boiling acetic acid re- quires activation by light and proceeds through the hydroxylamine (84; 118 C.Simons and L. G. Ratner J. 1944,421. 118 M. P. Cava and C. E. Blake J. Amer. Chem. SOC. 1956,78,5444. ll9 N. Kharasch W. King and T. C. Bruice J. Amer. Chem. SOC. 1955 77 931. lZo T. Zincke and F. Fur Annalen 1912,391,57. lZ1 D. L. Chamberlain D. Peters and N. Kharasch J. Org. Chem. 1958,23 381. lZz F. Kaluza and G. W. Perold Chem. Abs. 1961,55 11346. K. J. Morgan J. 1959 3502. 412 QUARTERLY REVIEWS NHOH for NH,) which can be isolated as its 0-acetyl derivative. Such compounds as the sulphenyl chloride (83 ; R = Cl but not NHPh) dissolve in concentrated sulphuric acid to give bright red solutions containing the dinitrophenylsulphenium ion to which a resonance-stabilised structure (85) is ascribed.123 In interesting contrast with the slow reaction of o-nitrobenzyl alcohol (p.408) o-nitrophenylmethanethiol (86) reacts vigorously with strong aqueous alkali forming thioanthranil (87) together with the disulphide (88).124 The retention of sulphur in the heterocyclic product is noteworthy but in the strongly reducing environment provided by the thiol the stage at which the sulphur-nitrogen bond is formed is again not clear. Thio- anthranil is also formed in base-catalysed decomposition of the sulphide ( 89).lZ5 Reactions in Neutral Media Some interactions occur under seemingly neutral conditions (pp.402 and 403). For instance when heated with sand at 220° o-nitrodiphenyl- amine (90; R = Ph) yields phenazine (91) and N-benzyl-o-nitroaniline (90; R := CH,Ph) yields 2-phenylbenziminazole (92).126 Benzofuroxans which are capable of the isomerisation shown in (x~xiii),~,~ can be prepared by thermolysis of o-nitrophenyl azides e.g. (93),12* and intermediate formation of a nitrene (94)129 is one of several possible pathways.12 Alternatively benzofuroxans are formed through oxidation of o-nitro- anilines by alkaline hypo~hloritel~~ or in benzene but not in acetic acid by phenyliodoso diacetate.131 U Nitroso-compounds already mentioned in this Review as products or as possible intermediates of various reactions may also rank as initial lZ3 N.Kharasch C. M. Buess and W. King J. Amer. Chem. Soc. 1953,75 6035. 124 S. Gabriel and R. Stelzner Ber. 1896 29 160. 125 Y. Iskander and Y. Riad J. 1951 2054. 126 R. H. Smith and H. Suschitzky Tetrahedron 1961 16 80. 12' Cf. F. B. Mallory and C. S. Wood J. Org. Chem. 1962 27,4109. lZ8 Cf. J. H. Boyer and F. C. Canter Chem. Rev. 1954,54 35. lZ9 Cf. R. A. Abramovitcli and K. A. H. Adams Canad. J. Chem. 1961,39 2516. 130 A. G. Green and F. M. Rowe J. 1912 101 2443; cf. F. M. Rowe and J. S. H. 131 L. K. Dyall and K. H. Pausacker Austral. J. Chem. 1958,11,491. Davies J. 1920,117 1344. LOUDON AND TENNANT SUBSTITUENT INTERACTIONS 41 3 reagents. Thus when reacting in ether nitrosobenzene and o-nitrophenyl- acetylene provide a mixture of products from which the di-isatogen (69) azobenzene and a compound formulated as (95) have been The reaction occurs in absence of light and when conducted in acetic acid affords some 1-hydroxyisatin (9).Similarly from o-nitrophenyl- propiolic acid the products identified are azoxybenzene isatin 1 -hydroxy- isatin the nitrone (95) and (from the ethyl ester) ethyl isatogenate. The process despite its complexity is recommended133 for preparing 2-phenyl- isatogen (8; R = Ph) from o-nitrotolane and here as in allied reactions nitrones of type (96) are the suggested intermediates from which upon cyclisation o-nitrosobenzene is regenerated. In connection with isatogen- formation it is of interest to note that the isatogen (97) is part-product of the reaction at 0" between dinitrogen tetroxide and t01ane.l~~ 132 L. Alessandri Gazzetta 1927 57 195; 1928,58 551; (Chem. Abs. 1929,23,3690.) 133 P. Ruggli E. Caspar and B. Hegedus Helv. Chiin. Acta 1937 20 250. K. N. Campbell J. Shavel and B. K. Campbell J . Amer. Chem. Suc. 1953 75 2400.

ISSN:0009-2681    

DOI:10.1039/QR9641800389

出版商:RSC    

年代:1964

数据来源: RSC

摘要:

ERRATUM 1963 Vol. XVII p. 359 line 14. Omit “and by conversion of desosamine (3) into chalcone. 97a’’

ISSN:0009-2681    

DOI:10.1039/QR9641800414

出版商:RSC    

年代:1964

数据来源: RSC