摘要:

Russian Chemical Reviews 69 (11) 911 ± 923 (2000) The use of objects and methods of colloid chemistry in nanochemistry B D Summ, N I Ivanova Contents I. Introduction II. Ultradisperse systems III. The formation of ultradisperse systems IV. The structures and forms of ultradisperse particles V. The application of theoretical concepts of colloid chemistry in nanochemistry VI. Self-organised colloidal structures VII. Conclusion Abstract. of concepts theoretical and methods Experimental Experimental methods and theoretical concepts of colloid chemistry regarding their possible use in nanochemistry colloid chemistry regarding their possible use in nanochemistry and nanotechnology are considered. The main types of disperse and nanotechnology are considered. The main types of disperse systems which can be regarded as nanosystems are distinguished.systems which can be regarded as nanosystems are distinguished. Some methods for the preparation of colloidal nanosystems are Some methods for the preparation of colloidal nanosystems are described. of extrapolation for Conditions described. Conditions for extrapolation of phenomenological phenomenological laws are objects nanosize to chemistry colloid of laws of colloid chemistry to nanosize objects are considered. considered. Examples given. are structures colloidal self-organised of Examples of self-organised colloidal structures are given. The The bibliography references 205 includes bibliography includes 205 references. I. Introduction According to the classification accepted in colloid chemistry, nanosystems are referred to as ultradisperse colloidal systems with the particle sizes ranging from 1 to 100 nm.1 ±3 This range of particle sizes corresponds to the maximum degree of dispersion at which a colloidal system still preserves one of its basic properties, viz., heterogeneity.According to Rehbinder,4 the minimum size of the disperse phase particles is about 1 nm (3 ± 5 molecular diameters). In order to establish the relationship between colloid chem- istry and nanochemistry, it seems reasonable to recall the main stages in the development of colloid chemistry. After the pioneer- ing studies byMFaraday (1857), who was the first to obtain stable colloidal solutions (sols) of highly dispersed red gold, it was these systems with the particle sizes of 5 ± 100 nm that were the focus of attention of numerous physicists and chemists during the second half of the 19th century and the beginning of the 20th century.Experimental and theoretical studies of different properties of colloidal solutions led to the establishment of the main laws of colloid chemistry, viz. , the law of Brownian motion and diffusion of colloidal particles (A Einstein), the heterogeneous nature of colloidal solutions (R Zsigmondy), sedimentation ± diffusion equilibria of dispersions in the gravitation field (J Perrin) and in B D Summ, N I Ivanova Department of Chemistry, MV Lomonosov Moscow State University, Leninskie Gory. 119899 Moscow, Russian Federation.Fax (7-095) 932 88 46. Tel. (7-095) 939 10 31. E-mail: summ@colloid.chem.msu.ru (B D Summ). Tel. (7-095) 939 53 87. E-mail: ini@colloid.chem.msu.ru (N I Ivanova) Received 27 July 2000 Uspekhi Khimii 69 (11) 995 ± 1008 (2000); translated by V D Gorokhov #2000 Russian Academy of Sciences and Turpion Ltd DOI 10.1070/RC2000v069n11ABEH000616 911 912 912 915 916 918 920 a centrifuge (T Svedberg), light scattering (J Rayleigh), coagula- tion of sols with electrolytes (G Schulze andWHardy). Studies of properties of colloidal solutions of various substances have led to the establishment of the fundamental principle of the universality of the colloidal state (P Weimarn).5 During this period, the leading scientists in the field of colloid chemistry understood well the crucial importance of systematic studies of highly dispersed systems; however, experimental meth- ods of that time did not possess sufficiently high resolution, in particular for the establishment of the structures of ultradisperse particles.Probably it was this circumstance that determined a shift of the peak of studies towards disperse systems with larger particles (foams, emulsions, suspensions, aerosols) and structured systems (gels, coagulation and crystallisation structures). The appearance of high-resolution methods for the structural studies of substances (NMR, electron and atom-force micro- scopy, special equipment for measurements of very low surface tensions down to 1073 mJ m72 , computer simulation, photon- correlation spectroscopy and a number of others) in the second half of the 20th century allowed the initiation of systematic studies of the structures and properties of ultradisperse colloidal systems.Over the last 30 ± 35 years, considerable progress has been achieved in this field; in particular, radically novel classes of ultradisperse systems { were obtained and studied. A tremendous experience gained in colloid chemistry in the investigations of ultradisperse systems can effectively be used for the solution of a number of basic and applied research problems of nanochemistry, since these systems are nanosystems { in their essence. However, it should be emphasised that one should be very careful in the extrapolation of laws of colloid chemistry and concepts of the structures of disperse particles and thin layers to the nanosize range.The closer the size of a disperse particle (d ) to the minimum possible one (i.e., to the nanosize), the stronger become the scale effects, viz. , the dependence of various properties on the particle size.7, 8 Nonetheless, the methods and concepts of colloid chem- istry can also be successfully used in nanochemistry with due respect for appropriate restrictions. This problem is rather com- {A spectacular example of ultradisperse systems may be microemulsions, which are thermodynamically stable systems with the disperse phase composed of droplets several nanometers in diameter.6 { To a certain extent, the problem of what to call the particles with the sizes from 1 to 100 nm is terminological rather than that of meaning: the terms `nanoparticles' and `nanosystems' appeared only after the introduction of the SI system of units, i.e., much later than the term `ultradispersity'.912 prehensive and includes a large set of theoretical and methodo- logical aspects.The present review pays major attention to the classification of ultradisperse systems, the methods of their preparation, as well as to the theoretical description of processes occurring in these systems. II. Ultradisperse systems In the first place, it is useful to discuss what classes (types) of colloidal systems may be regarded as nanosystems. Let us use the classification which is currently accepted in colloid chemistry and relies on certain important characteristics of disperse systems, such as the sizes of the disperse particles d , their dimensionalities (one-, two-, three-dimensional), concentration of particles n (m73), the character of interaction between the dispersed phase and the dispersion medium, aggregate states of the dispersed phase and the dispersion medium.1± 3 As regards their geometrical parameters, nanosystems and ultradisperse systems are virtually equivalent since in the latter the particle size is in the range 1 ± 100 nm.Based on their geometrical characteristics (dimensionalities of disperse particles), colloidal nanosystems may be divided into three groups. 1. Three-dimensional (bulk) nanoparticles in which all the three sizes (d1, d2, d3) are in the nanorange. It should be noted that the bulk particles have rather small radii of curvature. This type includes colloidal solutions (sols), microemulsions, nucleation particles formed upon phase transitions of the first type (crystals, drops, gas vesicles), spherical micelles of surfactants in aqueous and non-aqueous media (direct and reverse micelles).2. Two-dimensional (thin films and layers) nanoparticles in which only one size (thickness) is in the nanorange, while two others (length and width) may be unlimitedly large. These systems include thin liquid films, adsorption mono- and polylayers at the interface (including the Langmuir ± Blodgett films),9 two-dimen- sional lamellar micelles of surfactants.10 Thin liquid films are subdivided into foamy films (between two foam cells), emulsion films (between droplets of direct and reverse emulsions) and wetting films (separating a solid surface and a gas or a liquid).The foamy and emulsion films are referred to as symmetrical films, whereas the wetting films are regarded as non-symmetrical ones. The thickness of symmetrical foamy films stabilised by appropriate surfactants can range from several nanometers (the so-called Newtonian black films) 11 ± 16 to several tens of nano- meters. 3. One-dimensional nanoparticles in which the thickness and width are in the nanorange, while the length may be unlimitedly large. The one-dimensional ultradisperse particles include thin fibres, very thin capillaries and pores, cylindrical micelles of surfactants and nanotubes which much resemble them.17, 18 This group also includes the wetting line (or the line of three-phase contact), which separates three phases, viz., a solid, a liquid and a gas.19, 20 The classification of disperse particles based on their dimen- sionalities is important not only from the formal viewpoint.In conformity with the Ehrenfest principle, the geometry of particles (the dimensionality of space) has a substantial effect on the character of dependences relating physical parameters.21, 22 Thus the power index for the distance in the Newtonian gravita- tion laws and for the electrostatic Coulomb interaction is 72 for three-dimensional space and71 for two-dimensional space.22 Yet another remarkable parameter confirming the Ehrenfest principle is the different dependence of the heat capacity of solids at constant volume CV on temperature T in the range of suffi- ciently low temperatures. Thus Tarasov 23 ± 25 showed experimen- tally that the law of squares (CV*T2) is fulfilled for the lamellar structures (e.g., for graphite and gallium) instead of the Debye law of cubes for isotropic materials whereas the linear dependence (CV*T) is observed for the chain structures (of crystals selenium, B D Summ, N I Ivanova HF, BiO3 and MgSiO3).The theory of heat capacity of thin films and thin rods (needles) developed by Lifshitz 26, 27 states that at low temperatures the CV(T) dependences for two- and one-dimen- sional objects differ substantially from those for three-dimen- sional crystals.Yet another example concerns the theory of capillarity. Clerot who put forward in 1743 the hypothesis of the existence of molecular forces suggested that they depend on the distance r in the same way as the gravitation forces (*r72). However, this hypothesis contradicted the empirical Juren law (1718) describing the upward movement of wetting liquids in capillary tubes. On the contrary, the Laplace hypothesis (1804) 28 of the short-range action of molecular forces (*r71) described correctly this phe- nomenon. It should be noted that the Ehrenfest principle provides the physical substantiation of the Laplace hypothesis. Indeed, the change in the power index from n=72 in the Coulomb law to n=71 in the Laplace hypothesis corresponds, according to Ehrenfest, to the transition from a three-dimensional to a two- dimensional object.The above-cited regularities reflect changes in a number of physicochemical properties of nanoparticles compared to the corresponding properties of the macrophase. For example, a strong decrease in the melting temperature is observed for nano- particles of many metals and semiconductors (Ag, Au, Pb, Sn, In, Bi, Ga, CdS).29, 30 It is also known that the strength of filamentous crystals and fibres can be several times higher than that of macroscopic bodies. Nanosize particles possess an enhanced chemical reactivity manifested in the increase in the rates of chemical reactions involving their particles.This property of nanoparticles is used in the preparation of catalysts. It should also be noted that some phases can appear in nanoparticles which are not revealed in macrosystems. A sufficiently detailed descrip- tion of special properties of nanoparticles can be found in relevant reviews.31 ± 34 Thus, the Ehrenfest principle has a fundamental significance. However, it has not yet found uses in colloid chemistry, including the theory of surface phenomena, although its potentialities seem to be very large here, especially for the description of nanosystems. As has been noted, colloidal systems comprise several phases, at least two, therefore, properties of a system depend on the interaction between these phases.Based on the energy of inter- action of the disperse phase and the dispersion medium, disper- sion systems are divided into two large classes, viz., lyophilic and lyophobic. According to Rehbinder,4 the formation of stable lyophilic systems results from spontaneous dispersion of a macro- phase into particles of colloidal size d. The thermodynamic condition of spontaneous dispersion is described by the expression (1) sd 24bkT , where s is the surface tension at the disperse phase ± dispersion medium interface, b is the coefficient taking into account the particle shape and changes in the entropy of a system due to the involvement of particles in the thermal motion (b=15 ± 30), k is the Boltzmann constant and T is temperature.It follows from the condition (1) that the formation of colloidal systems with the particle size of 1 ± 100 nm requires a strong decrease in the surface tension, approximately down to 1072 ± 1073 mJ m72. This type of nanocolloids includes microemulsion systems, critical emul- sions and micellar systems.4, 6 Thus, the range of ultradisperse systems (nanosystems) is very broad and they are very interesting from the scientific viewpoint for reviews, (see, e.g., Refs 35, 36). III. The formation of ultradisperse systems Numerous methods have been developed in colloid chemistry for the preparation of ultradisperse systems allowing a rather fine control over the sizes of particles, their shapes and structures. The Nobel Prize winner Th Svedberg 37 proposed to divide methods for the preparation of ultradisperse particles into two groups:The use of objects and methods of colloid chemistry in nanochemistry dispersion methods (mechanical, thermal, electrical grinding or spraying of a macroscopic phase) and condensation methods (chemical or physical condensation).It is impossible to consider all these methods in sufficient detail within the frame of this review; therefore, we will discuss only some examples of chemical methods for the preparation of ultradisperse systems. Historically, it was by these methods that the first highly dispersed colloidal solutions were obtained. 1. Preparation of sols Systematic development of methods for the synthesis and purifi- cation of highly dispersed colloidal systems began in the middle of the 19th century when Faraday prepared stable sols of gold (with the particles 2 ± 50 nmin size) by reduction of a dilute solution of a gold salt with yellow phosphorus.38 Au + P(OH)3 + 3 HCl .AuCl3 + 3H2O + P Later, Zsigmondy 39 developed methods (which became clas- sical) for the synthesis of monodisperse gold sols with a predeter- mined degree of dispersion by the reduction of gold(III) compounds with hydrogen peroxide and formaldehyde. 2Au + 8 HCl + 3O2 , 2 HAuCl4 + 3H2O2 2 HAuCl4 + 3HCHO + 11KOH 2Au + 3HCOOK + 8 KCl +8H2O . This process proceeds in two steps. In the first step, nuclei of a new phase are formed and then a slight oversaturation occurs in the sol where no new nuclei are formed and only growth of the existing nuclei takes place.Yellow (d*20 nm), red (d*40 nm) and blue (d*100 nm) sols of gold can be prepared in this way. The invention of an ultramicroscope (1903) which allowed measurement of ultradisperse particles (down to 3 nmin size) gave a strong impetus to the search for new methods for the prepara- tion of sols and studies of their properties. It is worthwhile to note several basic achievements. Svedberg 40 synthesised highly dis- persed sols of alkali metals in organic solvents on strong cooling. Weimarn 41 was the first to establish the conditions under which any substance can be present in the colloidal state: `Any substance may be obtained in the form of colloidal solution if one succeeds in bringing about the reaction of its formation in a dispersion medium, which dissolves this substance negligibly, at a sufficiently low concentration of reactants.' He prepared sols of several hundred different substances under laboratory conditions.In the early 1980s, it was found that the solid-phase nanosize particles have special mechanical, optical, electrical and magnetic properties different from the corresponding properties of macro- systems.34, 42, 43 This fact determined the wide uses of nanopar- ticles in different fields of science and technologies. Numerous methods of synthesis of nanoparticles have been developed to date. Analysis of methods for the preparation of nanoparticles of different nature is reviewed in Refs 31, 34 ± 46.The present review considers only briefly recent methods for the preparation of sols based on condensation reactions. Hydrolysis of inorganic salts of metals or metal alkoxides resulting in the formation of sols of oxides and hydroxides of the corresponding metals can serve as an example.47 ± 51 Thus iron hydroxide sol can be prepared by the reaction Fe(OH)3 + 3HCl . FeCl3 + 3H2O In this approach to sol preparation, it is important to observe thoroughly the reaction conditions; in particular, it is necessary to control strictly pH and the presence of certain organic compounds in the system. Thus Kandori et al.52 showed that the size of Fe2O3 particles resulting from hydrolysis of FeCl3 depends on the concentration of triethanolamine, isopropylamine and piperazine.The presence of a large excess energy in disperse systems related to the highly developed interface favours the processes of aggregation of colloidal particles (see Section V). To prepare sols of required dispersity, it is necessary to arrest growth of particles 913 in due time. To this end, the surface of particles in the disperse phase is deactivated by formation of a protective layer of surfactants 53 ± 55 or by complex formation.56, 57 As regards control of the shape and size of nanoparticles, highly promising proved to be the synthesis of nanoparticles in microemulsion systems which will be considered in more detail in Section III.3. 2. The formation of micellar systems Surfactants, organic substances (synthetic or natural) with limited solubility in water, which can adsorb on the interface thereby reducing the interfacial tension, represent unique aspect of colloid chemistry.These compounds have diphilic structures: a surfactant molecule or ion contains a hydrophobic part and a polar group of some kind. The hydrophobic part is a hydrocarbon radical (CnH2n+1, CnH2n71, CnH2n+1C6H4, etc.) containing from 8 to 18 carbon atoms. Depending on the nature of the hydrophilic group, the surfactants are divided into cationic (these include primary, secondary and tertiary amines and quaternary ammo- nium bases), anionic (molecules of these compounds contain carboxy and sulfonic groups and others) and non-ionogenic (these surfactants exist in solution in the form of molecules).58 The specificity of the behaviour of surfactants in aqueous solutions is related to peculiarities of the interaction between the water molecules and surfactants.According to numerous studies, at room temperature water is a structured liquid and its structure is similar to that of ice but in contrast to ice, water has only near ordering (r<0.8 nm).59, 60 Solubilisation of surfactants induces further structuring of water molecules around non-polar hydro- carbon radicals of surfactants, which leads to a decrease in the entropy of the system. As the system tends to the maximum entropy, at a certain concentration called the critical micelle concentration (CMC), the surfactant molecules or ions begin to form spontaneously associates which are called micelles (as proposed by Mac-Ban who discovered them in 1913).The formation of micelles is accompanied by the release of a portion of structured water which is a thermodynamically advantageous process since it leads to an increase in the entropy of the system. The formation of micelles is usually revealed from changes in some physical property of the surfactant solution (e.g., surface tension, electric conductivity, density, viscosity, light scattering, etc.) depending on the surfactant concentration. The magnitude of CMC depends on a number of factors: the surfactant nature, the length and degree of branching of a hydrocarbon radical, the presence of an electrolyte or other organic compounds and pH of the solution.However, the basic factor is the ratio between hydrophilic and hydrophobic properties of the surfactant. The longer the hydrocarbon radical and the weaker the polar group the smaller the CMC.} At concentrations close to CMC, micelles represent nearly spherical formations in which polar groups contact with water, while hydrophobic radicals are directed inwards to form a non- polar nucleus.4, 61 Molecules or ions constituting micelles are in dynamic equilibrium with the bulk of the solution. This is one of the reasons for the `roughness' of the outer surface of micelles.61, 62 The degree of hydration of polar groups, the structure of the hydrate layer, and the structure of the internal nucleus depend on the nature of the surfactant.63 ± 66 At concentrations higher than CMC, the formation of several types of differently shaped micelles is possible (Fig.1): spherical, cylindrical, hexagonal packed and lamellar ones. Thus, micelles may be regarded as one-, two- and three-dimensional nanoobjects. Depending on the surfactant nature, the aggregation number (n) can change from tens to several hundreds, the sizes of micelles will be changing accordingly.67 } Modern concepts of the thermodynamics of surfactant solutions and micellisation processes are considered in sufficient detail in Rusanov's monograph.61914 1 2 310 ± 35A 6 Figure 1. Structures arising in solutions of surfactants: (1) monomers; (2) micelle; (3) cylindrical micelle; (4) cylindrical micelles as packed hexagons; (5) lamellar micelle; (6) water droplets packed as hexagons in the reverse micellar system. Water-insoluble surfactant molecules with a long hydrocar- bon radical and a weakly polar group can be dissolved in non- polar liquid phases.In this case, at a definite surfactant concen- tration the formation of micelles is also observed and is deter- mined by specific interactions between the polar groups of the surfactant.68 ± 71 These are called reverse micelles. The shapes of the reverse micelles are not necessarily the same, this depends on the surfactant concentration. There are two approaches to the description of micelle formation. According to the first approach (the quasi-chemical model), the micellate formation is considered from the standpoint of the mass action law.The second approach interprets the appearance of micelles as the formation of a new phase. One of the most important properties of the micellar systems is their ability to solubilise, viz. , to increase considerably the solubilities of hydrocarbons in aqueous micellar solutions, or, respectively, of polar liquids in the reverse micellar systems.72 ± 74 Solubilisation results in the appearance of thermodynamically stable equilibrium isotropic systems, which are called microemul- sions (these systems will be considered in more detail in the next section). Due to numerous factors which influence solubilisation (the nature of contacting phases and surfactants, the presence of electrolytes, temperature), the maximum solubilities of substances in surfactant micelles can vary within very broad limits.It should be noted that upon solubilisation the properties of compounds change largely and can also result in changes in the rates of chemical reactions occurring in these systems. This phenomenon known as the micellar catalysis finds wide uses in chemistry, biology, medicine and various technological processes. For exam- ple, the increase in the reactivities of compounds is widely used in the processes of emulsional polymerisation 75 ± 77 and enzymatic catalysis.78 ± 80 3. Formation of microemulsions Microemulsions are thermodynamically stable isotropic disper- sions of two non-miscible liquids. Upon mixing these liquids, the droplets of one of them stabilised by the interfacial film of a surfactant and a co-surfactant (a low-molecular-mass alcohol) are distributed in the other liquid.The term `microemulsion' was proposed by Schulman 81 who was the first to obtain such systems upon addition of surfactants to a macroemulsion. Microemul- sions are regarded as lyophilic disperse systems and may be obtained either by spontaneous dispersion of two non-miscible B D Summ, N I Ivanova 4 Water 10 ± 35A 5 liquids as a result of a strong decrease in the interface tension or in the process of solubilisation, as noted above. The thermodynamic stabilities of microemulsion systems are determined by low inter- face tension which, according to the literature data, may be 1075 mJ m72 for ionic surfactants and 1074 mJ m72 for non- ionogenic surfactants.82 Depending on which phase is dispersed and which is continuous, the microemulsions can be direct [oil in water (o/w)] or reverse [water in oil (w/o)].} In both cases the dispersed phase consists of droplets whose size does not exceed 100 nm.Microemulsions are, as a rule, multicomponent systems comprising different structures (bilayer, cylindrical or spherical micelles).83 ± 85 In the process of micelle formation, optically anisotropic micellar phases, for example, lamellar smectic and hexagonal phases consisting of rod-shaped aggregates of infinite lengths, are formed in addition to the liquid isotropic micellar phases, i.e., microemulsions possess internal microstructures which are currently studied by different methods.86 ± 88 In the case where the contents of water and oil in the system are comparable, bicontinuous systems can be formed.89, 90 The properties of microemulsions are largely determined by the sizes and shapes of the dispersed phase particles as well as by rheological properties of interface adsorption layers formed by surfactants.Since the microemulsions are highly mobile and posses large interfaces, they can serve as versatile media for numerous chemical syntheses, including the preparation of solid nanoparticles. 4. The formation of solid particles in microemulsions In a microemulsion system, the dispersed phase particles contin- uously collide, coalesce and degrade again, which leads to a continuous exchange of their contents.Collision of droplets depends on their diffusion in the oil phase (for the reverse micro- emulsion system), whereas the process of exchange is determined by the interaction of adsorption layers of surfactants and the flexibility of the interface (the latter circumstance is crucial for carrying out chemical reactions in such systems).91, 92 Reverse microemulsion systems are often used for the prepa- ration of solid nanoparticles. To this end, two identical micro- emulsion systems (w/o) containing in their aqueous phases } The term `oil' means a non-polar organic liquid.The use of objects and methods of colloid chemistry in nanochemistry A B B A + + C Figure 2. Scheme of a reaction occurring in the reverse microemulsion system.compounds A and B which form a poorly soluble product C in the chemical reaction are mixed. The new compound C is formed in droplets as a result of coalescence (Fig. 2). The sizes of the new phase particles will be restricted by the size of droplets of the polar phase. Metal nanoparticles can also be obtained by adding a reducing agent (e.g., hydrazine) to a microemulsion containing a metal salt or by bubbling a gas (e.g., H2 or CO) through this emulsion. It was in this way (reduction of a metal salt by hydrazine) that mono- dispersed metal particles of Pt, Pd, Rh and Ir were obtained with a particle sizes of 3 ± 5 nm.93 An analogous method was applied for the synthesis of bimetallic particles of platinum and palladium.94 At present, precipitation reactions in microemulsion systems are widely used for the synthesis of metal nanoparticles,95 ± 97 semiconductors,98 ± 102 barium, calcium and strontium carbonates and sulfates,103 ± 107 monodispersed SiO2 particles 108 ± 110 and high-temperature ceramics.111 ± 113 Despite the fact that the mechanism of formation of nano- particles has not yet been ultimately established, it is possible to identify a number of factors which influence this reaction.These are primarily the ratio of the aqueous phase to the surfactant (W=[H2O]/[surfactant]) in the system, the structure and proper- ties of the solubilised aqueous phase, the dynamic behaviour of microemulsions and the mean concentration of reagents in the aqueous phase.The sizes of droplets in the disperse phase are also influenced by the nature of the surfactants which stabilise the microemulsion system. However, in all cases the sizes of nano- particles formed in the reaction are governed by the sizes of droplets of the initial microemulsion. For instance, Petit and Pileni 99 showed that the sizes of CdS nanoparticles increase virtually linearly as the ratio W increases. On the other hand, the sizes of particles obtained in the reverse microemulsion system stabilised by sodium di(ethylhexyl) sulfosuccinate (Aerosol OT) proved to be smaller than those in the system stabilised by the non- ionogenic surfactant Triton X-100 [polyethylene glycol (n=10) p-iso-octylphenyl ether].Microemulsion systems are also used for conducting hydrol- ysis. Hydrolysis of tetraethoxysilane in the reverse micellar system stabilised by Aerosol OT can serve as an example.114 SiO2 + 4 EtOH . Si(OEt)4+ 2H2O It should also be noted that microemulsion systems are used for the preparation of organic compounds, which is important for the creation of new medicinal forms.115 Most of studies in this field relate to the synthesis of spherical nanoparticles. On the other hand, the preparation of asymmet- rical particles (filaments, discs, ellipsoids) and a rigorous control of their shapes is also of high scientific and practical interest. Syntheses carried out in the microemulsion systems allowed the preparation of BaCO3 nanofibres 116 and asymmetrical nano- particles of various compounds with unusual magnetic proper- ties.117 ± 119 915 Highly interesting is the synthesis of nanocomposites consist- ing of particles of one material (the particle size is 50 ± 100 nm) coated with a thin layer of another meterial.120, 121 5.Preparation of mono- and polymolecular layers Surfactants can form monomolecular films on various interfaces: water ± air, solid ± liquid and liquid ± liquid. Such films may be regarded as two-dimensional nanosystems. The surfactant mono- layers on the water ± air interface were studied by Langmuir who developed an experimental method for the studies of these films (the Langmuir balances).122 For this work Langmuir was awarded the Nobel Prize in 1932.Monomolecular films on the surface of liquids may be in the gaseous, liquid and solid states. These states are characterised by different energies of interaction of the surfactant molecules. Under definite conditions (pH, temperature), the water ± air interface is the site of spontaneous formation of highly ordered structures, in which the surfactant molecules (or ions) are arranged in such a way that the polar group is in solution, while the hydrocarbon radical is oriented into the air at a small angle to the interface. The processes of self-organisation in a film are determined by the diphilicity of the surfactant molecules and may be analysed from the viewpoint of interaction of the polar group with the water substrate and interaction of the hydrocarbon radicals with each other.122 ± 126 Chemical reactions occurring in monomolecular films are of considerable interest.By varying the surface pressure it is possible to control the orientation of molecules in the surface layer and influence thereby specifically the occurrence of these reactions. Thus the Langmuir ± Blodgett films are used for the preparation of solid nanoparticles of different nature directly in the process of chemical reaction or photochemical reduction of metal salts.127 ± 130 Similar processes also occur in biological systems.131 The films deposited onto the solid surface can form both mono- and polylayers. For example, if a glass plate positioned vertically is drawn from water through a monolayer of barium stearate on its surface, this plate becomes coated with a surfactant layer in which the hydrocarbon radicals are oriented outwards.The result is that the surface of this plate becomes hydrophobic. It can then be covered with the next layer. Hydrophilic or hydro- phobic surfaces may be obtained by successive deposition of layers.132 The films composed of layers of the same orientation are called the X-films, whereas those with the oppositely oriented layers are called the Y-films. In this way, it is possible to obtain polylayer coatings with thicknesses in the range of nanometers. IV. The structures and forms of ultradisperse particles The questions concerning the mechanisms of formation and the structures of nanosize particles are regarded as the most important and essential for colloid chemistry.Indeed, ultradisperse particles are some kind of `elementary particles' of colloid chemistry. The transition from a simple qualitative definition of the concept of disperse particles to the determination of their quantitative parameters and relationships requires detailed elucidation of the structures of ultradisperse particles in different colloidal systems, such as sols, micellar solutions, microemulsions, gels, etc. The early concept of the structures of solid ultradisperse particles was based on the suggestion that their structures are analogous to that of the macrophase of the same substance. However, further studies of the process of nucleation and growth of a new phase showed that depending on crystallisation con- ditions (the degree of oversaturation or overcooling, the presence of admixtures and a number of other factors), both amorphous and crystalline ultradisperse particles can be formed from solu- tions.133, 134 Back in 1910, Weimarn 5 found that the shapes of BaSO4 particles formed by crystallisation from solution depended on the degree of its oversaturation.Thus he obtained highly dispersed sols, flocculent structures, well faceted microcrystals916 and needle-shaped crystals. The temperature of the synthesis of nanoparticles also plays an important role. For example, nano- particles of titanium dioxide obtained by the sol ± gel method are rod-shaped at low temperature and have the form of bipyramidal crystals at high temperature.135 Yet another confirmation of the diversity of forms of nanoparticles is the formation of dendrites upon crystallisation from solutions and melts.136 The diversity of forms is due to the fact that the processes of new phase formation (the self-organisation processes) occur under essentially non-equilibrium conditions, and the extent of struc- tural perfection depends on how far the crystallisation conditions deviate from the equilibrium conditions.For example, in the synthesis of diamond from a dense gas phase and plasma, a more perfect structure is formed under more non-equilibrium condi- tions.137 Surfactants present in solution can strongly influence the crystallisation process. Depending on the nature and concentra- tion of the surfactants, they can change the rate of nucleation and growth of nuclei of the new phase, the shape of crystals and distribution of nanoparticles with respect to the size.138 ± 141 All these effects are related to the selective adsorption of surfactant molecules or ions on different facets of the crystals formed and consequently to the retardation of growth of some facets com- pared to others.142 In addition, the nature of the surfactant also influences the polymorphism of the compounds formed.For example, FuÈ redi-Milhofer et al.143 showed that crystallisation of calcium oxalate in solutions of octaethylene glycol monohexa- decyl ether results in the formation of calcium oxalate monohy- drate, whereas in solutions of sodium dodecyl sulfate, dihydrate is formed.An important feature of the crystallisation processes leading to the formation of nanoparticles is that their shapes cannot be described by the methods of usual geometry. These systems are described using the fractal geometry, since in the case of strong deviations from equilibrium and consequently high values of the driving force of the crystallisation process the instability of the interface leads, as a rule, to the formation of fractal structures.144 Of interest are the studies reported by Melikhov et al.145, 146 who showed that highly dispersed primary single crystals are formed first in the mutual crystallisation of ammonium halides and cesium iodide from highly oversaturated vapour.Owing to highly developed interface, the disperse system formed possesses a large energy excess and for this reason it is the site of the occurrence of aggregation processes, which are accompanied by the concrescence of the initial single crystalline particles of approximately the same size. This aggregation results in the formation of pseudo-single crystals. The processes of formation of ultradisperse systems in the case of metal crystallisation are much more complex and diverse. In the essence, investigation of these processes laid the foundation for a new direction known as the chemistry of clusters.147 ± 149 Cluster particles occupy the intermediate position between mononuclear compounds and disperse particles. The metal particles with d<30 nm can be obtained by the methods of gas-phase nuclea- tion, cryogenic synthesis and reactions in polymeric matrices and reverse microemulsion systems.150 ± 153 It should be noted in conclusion that until now there is no sufficiently comprehensive quantitative theory of nucleation and growth of colloidal particles.V. The application of theoretical concepts of colloid chemistry in nanochemistry The concepts of colloid chemistry may also be used for the description of nanosystems (ultradisperse colloids). However, this involves rather complex methodological problems. These are due to the intermediate position occupied by the ultradisperse particles between molecules (atoms, ions) and macroscopic bodies (phases).B D Summ, N I Ivanova Both chemical and physical approaches are used for the theoretical description of colloidal systems. This is related to the boundary position of colloid chemistry between chemistry and physics, which exerts a strong effect on its methodology.154 For example, the chemical approach prevails in studies of the highly dispersed state (including surface layers). Chemical methods are applied for the calculation of the electronic structure of nano- particles (clusters) and its effect on their catalytic properties,149 surface forces and Hamaker constants,155, 156 as well as for the determination of the constants of adsorption equilibrium using the mass action law.61 Based on the analogy of a bimolecular reaction and an elementary act of coagulation, the theory of rapid coagulation of sols under the action of electrolytes was developed (M Smolukhovsky, 1916).The theory of electrokinetic phenom- ena (electrophoresis and electroosmosis) is based on the concepts of electrochemistry on the structure of the double electrical layer and electrocapillary effects. At the same time, the concepts of molecular physics, chemical thermodynamics and mechanics of continuous media are widely used for the description of relatively large particles and colloidal systems formed by them. Of particular conceptual interest are the colloidal systems where the progress in the studies has been achieved upon the parallel use of chemical and physical theories, models and methods.A spectacular example of this approach is the inves- tigation of the formation of direct and reverse micelles in solutions of surfactants. The modern theory of micelle formation has been developed based on two complementary models, viz., the quasi-chemical model (the formation of a micelle resulting from the reversible aggregation reaction) and the model of a phase transition of the second type.61 Such a conceptual dualism is rather characteristic of the methodology of colloid chemistry. Combination of macro- scopic and molecular properties in a single object (colloidal particle, surface layer) was duly appreciated long ago by both chemists and physicists. Thus studies of the capillary and surface phenomena have made it possible to derive first quantitative estimates of the sizes of molecules (B Franklin, T Young) and have led to the discovery of forces of molecular attraction (A Clairaut, P Laplace).Studies of the temperature dependence of the capillary rise of organic liquids made it possible to establish their critical temperatures (D Mendeleev). The Boltzmann con- stant was calculated from the temperature coefficients of surface tension of pure substances. Yet another example of the combination of physical and chemical approaches to studies of colloidal systems is the discov- ery of the effect of adsorption-induced decrease in strength (the Rehbinder effect).157 The modern theory of this effect borrows a number of notions and concepts from the physics of solids and mechanics (dislocation, the concept of tensions in the uppermost part of a fissure, the Griffith theory for the calculation of the critical sizes of nuclei of a crack and others).On the other hand, studies of molecular mechanisms of the Rehbinder effect involve inevitably chemical processes and factors, since the decrease in the strength and plasticity of metals, ionic crystals and other solids was caused by the adsorption of the corresponding compounds in the predestruction zone.157 Phenomenological laws of colloid chemistry established on the basis of molecular physics, thermodynamics and mechanics of continuous media have a fundamental character. However, their direct extrapolation to the ultradisperse field is not, as a rule, sufficiently accurate.This is due to the fact that many properties of substances in the ultradisperse field are substantially dependent on the particle size (d), and changes in properties may be caused by different reasons. For instance, in the range of d<1 mm an increase in the strength of crystals is observed as their diameter d decreases. To an order of magnitude, this `threshold' corresponds to the mean distance between dislocations in non-deformed crystals.158The use of objects and methods of colloid chemistry in nanochemistry Colloidal systems are characterised by a very large specific surface O of the dispersed phase O& Kdr , (2) sr 1 á 2 d à s0 , r where K is a dimensionless coefficient (for the spherical and cubic particles, K=6), r is the disperse phase density.Other most important parameters which characterise colloidal systems are the specific free surface energy s (surface tension), the surface entropy Z and the specific adsorption G. For the systems with the moderate degree of dispersion, the surface tension s is determined only by chemical compositions (the nature) of the dispersed phase and dispersion medium. However, for the nanosystems one has to take into account the dependence of the surface tension on the sizes of dispersed particles. Because of special importance of the surface tension, let us consider this aspect in more detail. Two approaches are widely used for the analysis of the scale (dimensional) dependence of the surface tension.One of them relates the surface tension to the surface curvature and is used, e.g., for the description of surface properties of droplets, gas bubbles and nuclei of the new phase particles. Tolman 159 proposed the following equation for the description of this relationship: where sr is the surface tension corresponding to the curvature radius r, s0 is the surface tension for a flat surface, d is the thickness of the surface layer of a substance characterised (accord- ing to Gibbs) by excess thermodynamic functions. According to a number of estimates, d&1 nm; therefore, for the nanosize particles and gas bubbles a substantial change (increase) in the surface tension is observed. For this reason, the corresponding corrections considering the dependence s=f(r) should be made in all fundamental laws of colloid chemistry, viz., the Laplace law of capillary pressure, the Young law of contact angles of wetting, the Kelvin law of the saturated vapour pressure over a curved surface and others.The dimensional factor plays a rather substantial role in the formation of nucleation particles during phase transitions. Indeed, the classical thermodynamic relations which determine the critical size of the nucleus of a new phase include the surface tension for the surface of small curvature, though it is in the nucleation particle that the surface curvature is the largest. Let us clarify this statement with the following example. The critical size of the new phase nucleus rc in the homogeneous formation resulting from the phase transition of the first type is determined in the Gibbs theory by the equation (3) rc à 2sTm QmDT , where Tm andQm are the temperature and heat of phase transition and DT is the deviation of temperature from Tm.Thus, rather considerable overcoolings DT are required for the crystallisation of very small droplets of liquid metals, starting approximately from 5 nm. For instance, the values of DT for mercury, tin, copper and platinum were found to be 77, 118, 236 and 370 8C, respec- tively.160 The simplest rationale is the increase in the surface tension as the curvature radius decreases [Eqn (2)], which leads to the corresponding increase in the overcooling DT (at a fixed value of the droplet size rc).A different approach to the description of special properties of ultradispersed objects (thin liquid films) is used in the concept of disjoining pressureP introduced by Deryagin.19 The disjoining pressure is the difference between the pressure PH inside the film of thickness H and the pressure in the bulk phase P0 of the same substance at the same values of temperature and chemical potentials as in the thin film. (4) P=PH ±P0 . 917 The most important feature of thin films is that their thicknessHis commensurable with the surface layer thickness d. The over- lapping of the surface layers of two contacting phases (for H<d) makes an additional contribution to the free energy of the film. cosYr=cosY07 K The concept of disjoining pressure has been developed in sufficient detail and is widely used nowadays for the description of many phenomena.Thus the concept of disjoining pressure forms the basis of the theory of stability of colloidal solutions [the Deryagin ± Landau ± Verwey ± Overbeck theory (DLVO)].161 ± 163 Certain problems also arise in the description of wetting of solids with nanosize droplets. In these cases, the equilibrium contact angle of wetting Yr depends on the linear tension K. The corresponding equation is sr , (5) where Y0 is the equilibrium angle of a droplet at a large curvature radius (50.1 mm) and r is the radius of the droplet base. This relation, which is analogous to the Laplace law of capillary pressure Pr=2s/r, was obtained for the first time and confirmed experimentally in measurements of forces on the wetting line in the plane of a bubble base on a solid support.164 The contact angleYr for a droplet or a bubble with a small radius r is determined from the conditions of mechanical equilibrium on the three-phase contact line (TPL).165, 166 The direct confirmation of the dependence of Yr on the linear tension was obtained for microdroplets of organic liquids on asbestos filaments.167 According to different estimates, the mag- nitude of linear tension is *1077 mN.166 Therefore, this can make a noticeable contribution to the magnitude of contact angles (for liquids with the surface tension s&50 ± 100 mN m71) only in the case of nanosize droplets (r*1079 m).The influence of linear tension becomes rather substantial in the heterogeneous formation of the new phase nuclei, e.g., upon vapour condensation on wetted solid substrates. Thermodynamic analysis 166 provides the following equation for the work of a nucleus (droplet) formation as a spherical segment (6) A à PV 2 +K2l , where P is the capillary pressure, V is the droplet volume, l is the length of the TPL. Relation (6) may be regarded as a generalised equation with the aid of which one can calculate the work of formation of the new phase nucleus in the nanosize range (it is for this purpose that the linear tension K is introduced). In the case of a spherical droplet K=0, and relation (6) is transformed into the equation which describes the work of homogeneous formation of the new phase nuclei (A0).In order to calculate the work of heterogeneous formation of the new phase nucleus Ah, use is made of the well known Volmer equation 168 (7) Ah=A0 (273 cosY0 + cos3Y0) . Finally, for V=0 equation (6) describes the work of forma- tion of a two-dimensional nucleus (A=Kl/2). The above-cited examples of extrapolation of some laws of colloid chemistry to nanodispersed objects should be regarded only as a first approximation. The basic reason for the limited action of these laws, even with the consideration of the mentioned corrections, consists in peculiarities of the structure of nano- systems (see Section III). The complexity and non-triviality of the problems arising in this extrapolation will be explained by the example of two nanoobjects which are widely distributed in technological and natural processes: the surface layer of a liquid and LTC.At the interface of a pure liquid and a gas phase the surface layer (SL) has a very small thickness (d) which is of the order of several molecular layers (far from the critical temperature), i.e., it represents formally a two-dimensional nanoparticle with two918 macroscopic sizes (length and width) and one nanosize (thick- ness). Such a definition follows from the classical Young ± Laplace theory; it is also confirmed by calculations based on macroscopic theories. The Gibbs theory considers the SL of a substance as an independent phase, the thermodynamic parameters of which differ from analogous parameters of the bulk phase of the same substance.Evidently, it is for this reason that the structure of SL of a liquid will differ from the structure in the bulk. Until recently, the SL structure could not be studied in detail because of the absence of adequate experimental methods. Only in the last 5 ± 7 years has the improved method of small-angle scattering made it possible to obtain reliable information about the SL structure of different liquids (mercury, gallium, alkanes starting from C16H34 and some others). It was found for all the compounds studied that their SL represents a two-dimensional quasi-crystalline structure at temperatures several degrees higher than the melting temperature Tm.The thickness of SL is estimated to be 1 ± 2 molecular (atomic) diameters. It was established that the parameters of a two-dimensional crystal differ from the parameters of the lattice of a `bulk' crystal of the same sub- stance.169, 170 These experimental data confirm the recent network model of the surface layer of a liquid,171 ± 173 according to which it contains nanoparticles of the solid phase. They form a structure (`network') which determines the appearance of the surface tension of a liquid. The model predicts rather simple correlations between the surface tension and the heat of melting, which are in good agreement with experimental data for liquid metals and liquefied inert gases. These results seem to be of principal significance. It follows from them that the surface nanolayer of a pure liquid has in fact an aggregation state different from the state in the bulk.This confirms one of the initial postulates about the thermodynamics of the surface Gibbs phenomena which states that the SL is an independent phase.171, 174 On this basis, the process of SL formation may be regarded as the phase transition liquid ± two-dimensional solid.162 In the framework of this model, the surface tension sm is proportional to the heat of melting Qm near the melting temperature Tm. It thus follows that (8) sm=Qmrsd=Qsf , where rs is the solid phase density andQsf is the heat of SL melting per surface unit. In the first approximation, the thickness of the SL (d) is equal to a (a is a parameter of the crystal lattice of a given compound).Figure 3 shows that the dependence sm=f(Qm) is very well respected for many liquid metals and liquefied inert gases. The thickness of the SL for liquid alkali metals, other melts and liquefied gases corresponds to four, two and one monolayers, respectively. The model of the solid SL makes it also possible to calculate the temperature coefficient of surface tension 7ds/dT from the a b sm /mN m71 400 Li Xe 300 Kr Ar 200 Ne Na K 100 sm /mN m71 25 20 15 105 Rb Cs 0 0 0 0 40 10 20 Qsf /mJ m72 Figure 3. Correlation of the surface tension sm near the melting point with heat of melting Qsf: (a) liquefied inert gases, (b) alkali metals. B D Summ, N I Ivanova entropy of melting DS=Qm/Tm.Calculations of the coefficients of surface tension for many liquid metals give a rather close correspondence to experimental data.171, 172 Let us consider yet another example of the insufficiency of introduction of additional thermodynamic parameters (the dis- joining pressure P, the linear tension K ) for the development of a comprehensive theory of colloidal nanosystems. In three-phase systems, the most important nanoobject is the three-phase contact line. This appears upon wetting, capillary flow, adhesion of liquids, spreading of drops over the surface of solids, condensa- tion of vapours on supports and other phenomena. In these processes, the TPL is the front boundary of the liquid meniscus.Besides the term `three-phase contact line' (TPL), the term `wetting line' is used in a number of cases. As noted above, the linear tension describes special properties of TPL. Various theo- retical models which consider the influence of the linear tension K on the contact angles Y [see Eqn (5)] include in the implicit form the suggestion that the wetting line is a regular circumference, i.e., it has the same curvature 1/r along the entire length. An analogous assumption is also used, e.g., in the theoretical description of spontaneous spreading of a drop of wetting liquid over a solid or liquid surface,7, 175, 176 flows in capillaries and porous media,177, 178 etc. However, precision studies of the contour of drops of various liquids on solid supports (electron microscopy, laser interferom- etry) led to the conclusion about a totally different form of the wetting line.In the nanoscale, it represents an irregular alternation of protrusions and depressions, which are respectively ahead of or behind the mean position of the `round' wetting line, i.e., it is composed of small sites of positive and negative curvature. This result is rather important for the development of a microscopic theory of wetting. In the first place, it shows that the movement of the wetting line upon the drop spreading should not be regarded as a smooth expansion of the wetted area within the limits of the entire contour (this assumption is used in the hydro- dynamic theory of spreading).15, 176 A different mechanism of motion appears to be more probable, viz., the fluctuational formation of leading meniscuses in the radial direction and subsequent tangential flow of a liquid under the action of the capillary pressure created.This mechanism of the wetting line dislocation requires a much lower activation energy than that based on the simultaneous spreading of a drop over the entire perimeter. Such a movement of TPL has much in common with the sliding of edge dislocations in crystals;137 therefore, one can effectively use the corresponding models (formation of kinks, their expansion, etc) for its description. At the same time, good prospects are open for the use of methods of fractal geometry for the description of these objects.VI. Self-organised colloidal structures A fundamental property of dispersed systems is their capacity for a gradual evolution. This evolution is associated with the nature of the colloidal (dispersed) state and primarily with the thermody- namic non-equilibrium of most colloidal systems. The excess of free energy, which is determined by the presence of a highly developed interface separating the dispersed phase and dispersion medium, stimulates different processes (physical, physicochemi- cal) leading to a decrease in the Helmholtz free energy F. Without consideration of the chemical component, the variation of the Helmholtz free energy dF at a constant temper- ature is (9) dF=7TdS + d(sO) , where O is the interface magnitude. The thermodynamic condition of evolution dF<0 may be created in energy- and entropy-based ways.The former is charac- terised by changes in the second member of Eqn (9). This may be attained by decreasing the interface separating the dispersed phase and dispersion medium (sdO<0) and by decreasing the inter-The use of objects and methods of colloid chemistry in nanochemistry facial free surface energy (the surface tension) (Ods<0). Upon the decrease in the interface, an increase in the mean sizes of particles d takes place, i.e., the distribution of particles over their sizes, which is fixed directly at the moment of their formation, is shifted with time towards the larger particles. The mechanisms of this process may be completely different: coalescence of droplets in emulsions and of gas bubbles in foams upon their direct contact, isothermal distillation of the disperse phase particles in conform- ity with the Kelvin law, the Ostwald ripening of solid particles.As a result of these processes, the solid particles can become so large that they loose their sedimentational stability. Ultimately, these processes can lead to the partition of a dispersed system into two macrophases.179 The decrease in the specific free surface energy ds<0 also occurs upon the interaction of the dispersed particles through the dispersion medium film, as well as upon the loss of aggregation stability by the system in the process of coalescence. The theory of sol stability (the DLVO theory) 163 considers in the first approx- imation the interaction of colloidal particles as the sum of dispersion (attraction) and electrostatic (repulsion) interactions which take place upon overlapping of the ionic atmospheres due to the approach of sol particles to each other.In the case where the radius of the dispersed particles is much larger than the thickness of ionic atmosphere, the energy of interaction of colloidal particles U(h) may be expressed as U(h) = Bc0kTg2 exp(7KH)7 A 12pH2 , (10) K where B is a numerical coefficient, c0 is the electrolyte concen- tration in the system, g=tanh(zef0/4kT) (z is the valence of the counterion and f0 is the thermodynamic potential of the solid surface), K is the magnitude of the reciprocal thickness of ionic atmosphere theory, in the Debye ±HuÈ ckel K=(ee0kT/2z2e2c0)71/2 (e is the dielectric constant of the medium, e0 is the dielectric constant equal to 8.85610712 F m71), H is the thickness of the dispersion medium film between particles, A* is the Hamaker complex constant characterising the interaction of contacting phases.It follows from the DLVO theory that if the energy barrier related to the electrostatic repulsion of particles (Eel) is substan- tially higher than the energy of thermal motion (i.e., Eel 44 kT), then the system is stable to aggregation, and no enlargement of particles occurs. If Eel 55 kT , then each collision will lead to the formation of an aggregate, and the rate of formation of aggregates will be limited by the rate of their diffusion in the dispersion medium.This process is called rapid coagulation (M Smolukhovsky, 1916). The parameter which characterises the process of rapid coagu- lation is the coagulation period (t), viz., the time during which the concentration of particles n0 decreases by a factor of two. According to the Smolukhovsky theory, (11) 3Z t à 8kTn , 0 where Z is the viscosity of the dispersion medium. For the rapid coagulation, the value of t does not usually exceed a few seconds. Upon a decrease in the electrolyte concentration, the effi- ciency of collisions decreases and only a portion of them leads to the formation of aggregates. This is the case of the so-called slow coagulation (N Fuchs, 1934). In conformity with the Fuchs theory, the decrease in the rate of aggregate formation is associ- ated with the presence of a barrier in the dependence of potential energy on the distance between the centres of particles.180 The theoretically predicted dependences of the coagulation rate on concentration, as well as the ratio of rates as a function of the coagulating ion charge are in a satisfactory correlation with experimental results.The coagulation of sols yields aggregates, the fractal dimen- sionality of which depends on the coagulation type. For the aggregates formed under the conditions of slow coagulation, the 919 fractal dimensionality proves to be higher 181 ± 183 than for the aggregates formed under the conditions of rapid coagulation, which points to a higher degree of their organisation.Analysis of Eqn (10) showed that at sufficiently large values of H a minimum can be observed on the curve of the dependence of interaction energy of particles on the distance. The interaction of particles at large distances, the so-called far interaction, leads to the formation of ensembles possessing the phase stability. Such ordered formations are called periodic colloidal structures.184 A spectacular example of the periodic structures are the Schiller layers, viz., the structures formed in sols of iron hydrox- ides, vanadium oxides and some others whose particles have the anisometric forms. At low sol concentration, the particles form horizontal layers upon sedimentation separated from each other by several hundred nanometers.185 The increase in the sol concen- tration leads to the spindle-like aggregates which are called tactoids.In this case, the dispersed particles are arranged nearly in parallel to the spindle axis, while the distance between them is decreased to tens of nanometers. The dispersed particles can also be oriented near a solid support.135, 186 The periodic colloidal structures include also the Liesegang layers and rings,187 which were apparently the first examples of the self-organised structures studied. Let us consider these objects in more detail. The Liesegang rings (1896) can be prepared in the following way. A drop of 15% solution of AgNO3 is placed on a film of the gelatine gel which has been pre-impregnated with 0.4% potassium dichromate solution.The silver salt diffuses gradually into the gel. This results in the precipitation of silver dichromate. The precip- itate does not form a continuous zone around the drop but is in the form of concentric rings separated by transparent clear areas. Usually, the Liesegang rings appear when a concentrated salt solution diffuses through a gel which contains another electrolyte at a lower concentration. The chemical reaction results in the formation of a salt of a poorly soluble compound which is precipitated in some periodic way.187 ± 190 The precipitate can form not only the concentrated rings but also segments and radial formations.191, 192 It should be noted that despite the long time which has passed since the discovery of Liesegang rings, a sufficiently complete theory of their formation has not been developed until now.It should be mentioned that this phenom- enon is of a considerable interest not only for colloid chemistry but also for other sciences, e.g., biophysics. Indeed, the Liesegang rings are similar to ring structures which result from the autowave processes occurring in unidimensional systems (thin layers).193 In such `distributed' systems, the periodic wave structures appear due to diffusion undisturbed by convective fluxes, i.e., owing to the distribution of concentration waves. The same conditions are necessary for the formation of Liesegang rings. In the case of the reaction under discussion, one of the possible reasons for the appearance of Liesegang rings may be formation of colloidal particles of silver dichromate which coagulate because of the excess of one of the reagents. The processes of self-organisation are observed in the disperse systems under external impacts as well.Thus it was shown in a number of studies 194 ± 198 that the aggregation of initial particles with the formation of ordered structures is possible in the shear flow of suspensions. The above-cited examples refer to the ordered colloidal nano- systems with relatively large sizes of particles. Let us consider now another extreme case, viz., the self- organised nanosystems in which the size of dispersed particles is only slightly larger than the molecular size.These systems include, e.g., aqueous solutions of surfactants. The possible cases of self- organisation of the surfactant molecules are shown in Fig. 4; they include adsorption layers of surfactants at the interface, as well as direct and reverse micelles. Indeed, the close-to-saturation layers of surfactants at the interfaces water ± air, water ± oil and polar solid surface ± surfac- tant solution can have a rather high degree of ordering at which all920 6 1 2 4 The nanoparticles considered in colloid chemistry are the ultra- disperse particles. Among the ultradisperse systems a special place is occupied by the systems in which particles have a specific colloidal nature: they include micellar systems, microemulsions, adsorption layers of surfactants at the interface and the Lang- muir ± Blodgett layers.5 3 Figure 4. A scheme of the formation of organised structures in solutions of surfactants: (1) surfactant solution; (2) direct micelle in aqueous solution; (3) solubilisation of a non-polar liquid by the direct micelle; (4) reverse micelle in a non-polar liquid; (5) solubilisation of the polar phase by the reverse micelle; (6) adsorption layer of the surfactant at the aqueous solution ± air interface. the surfactant molecules (ions) are oriented perpendicular to the interface in a way ensuring the contact of the polar groups of the surfactant with the polar phase. It should be specially noted that the formation of self-organised biologically active structures on the water surface plays an extremely important role in biological evolution.Numerous data and calculations show that the surfaces of oceans (seas) are the most favourable regions for the appear- ance of Protozoa.193 Another class of self-organised colloidal nanostructures of minimum size includes the direct and reverse spherical micelles of surfactants. The ordering of micelles is characterised by the same orientation of the surfactant molecules (ions) in the radial direction from the centre of the micelle to its outer surface, on which the functional groups similar in their nature to the dis- persion medium (water or an organic liquid) are positioned. The self-organised nanostructures of this type also include lamellar and cylindrical micelles, bilayers of surfactants and vesicles.The factors favouring the formation of self-organised struc- tures of surfactants depend to a considerable extent on the nature of the liquid phase or the nature of the solid substrate (when considering the processes of adsorption of surfactants from solutions on solid surfaces). In aqueous solutions of surfactants, the processes of self-organisation are determined by the general increase in the entropy of the system dS>0 which is the sum of changes in the entropy of the dispersed phase dSd and in the entropy of the dispersion medium dSm. In this case, the local decrease in the entropy of the surfactant molecules in the processes of adsorption on the interface aqueous solution ± air and micelle formation is compensated by the increase in the entropy of the water molecules so that the total change in the entropy of the system dS>0.61, 199 ± 202 On the contrary, the appearance of micelles in the oil phase, as well as the adsorption of surfactants at the interface water ± oil or on a polar solid surface are determined by changes in the enthalpy of the system due to the interaction of the polar groups of surfactants with each other in the nuclei of reverse micelles or because of the interaction of the polar groups of surfactants with a polar solid surface or a polar solvent.203, 204 It should be noted that the processes of self-organisation of surfactant ions can also be observed in gels.205 B D Summ, N I Ivanova Thus, the most important feature of the evolutionary proc- esses in colloidal systems consists in that they often lead to the appearance of ordered structures of different scale (nano-, meso-, and micro-macrostructures).These structures may be two-dimen- sional and three-dimensional, while their size can change from several nanometers to hundreds of nanometers. Such structures are rather widely distributed. They are of a considerable applied and scientific interest, being convenient models of a number of biophysical and biochemical systems. VII. Conclusion The methods for the preparation and studies of ultradispersed systems developed in colloid chemistry may be used rather efficiently in nanochemistry and nanotechnology. In many cases, the methods of colloid chemistry make it possible to regulate rigorously the sizes, shapes, compositions of ultradispersed par- ticles and properties of the structures formed by them.The theoretical concepts of colloid chemistry concerning the surface phenomena, stability and evolution of disperse systems combined with the models used for studies of the thermodynamics of irreversible processes and synergism may also be applied success- fully to the nanodisperse systems. References 1. P A Rehbinder, G I Fuchs Uspekhi Kolloidnoi Khimii (Advances in Colloid Chemistry) (Moscow: Nauka, 1973) p. 5 2. D H Everett Basic Principles of Colloid Science (London: Royal Society of Chemistry, 1988) 3. J Israelachvili Intermolecular and Surface Forces (London: Academic Press, 1994) 4.P A Rehbinder Izbrannye Trudy. Poverkhnostnye Yavleniya v Dis- persnykh Sistemakh. Kolloidnaya Khimiya (Selected Proceedings. Surface Effects in Disperse Systems. Colloid Chemistry) (Moscow: Nauka, 1979) 5. P P Weimarn, B V Malyshev Zh. Russ. Khim. O-va 42 484 (1910) 6. S E Friberg, P Botorel (Eds) Microemulsions: Structure and Dynamics (Boca Raton, FL: CRC Press, 1988) 7. A Adamson Physical Chemistry of Surfaces (New York: Wiley- Interscience, 1976) 8. Yu I Petrov Fizika Malykh Chastits (The Physics of Small Particles) (Moscow: Nauka, 1982) 9. G G Roberts (Ed.) Langmuir ± Blodgett Films (New York: Plenum, 1990) 10. P Mukerjee, K J Mysels Critical Micelle Concentrations of Aqueous Surfactant Systems (Washington, DC: National Bureau of Standards, 1971) 11.I B Ivanov (Ed.) Thin Liquids Films (Surf. Sci. Ser. Vol. 29) (New York: Marcel Dekker, 1985) 12. V N Izmailova, G P Yampol'skaya, B D Summ Poverkhnostnye Yavleniya v Belkovykh Sistemakh (Surface Effects in Protein Systems) (Moscow: Khimiya, 1988) 13. P M Kruglyakov, D R Ekserova Pena i Pennye Plenki (Foam and Foam Films) (Moscow: Khimiya, 1990) 14. P M Kruglyakov, Yu G Rovin Fizikokhimiya Chernykh Uglevodo- rodnykh Plenok (Physical Chemistry of Black Hydrocarbon Films) (Moscow: Nauka, 1978) 15. B D Summ, Yu V Goryunov Fiziko-Khimicheskie Osnovy Smachivaniya i Rastekaniya (Physicochemical Foundations of Wetting and Spreading) (Moscow: Khimiya, 1976) 16. S S Dukhin, N N Rulev, D S Dimitrov Koagulyatsiya i Dinamika Tonkikh Plenok (Coagulation and Dynamics of Thin Films) (Kiev: Naukova Dumka, 1986) 17.E G Rakov Usp. Khim. 69 41 (2000) [Russ. Chem. Rev. 69 53 (2000)]The use of objects and methods of colloid chemistry in nanochemistry 18. P M Ajayan, T W Ebbesen Rep. Prog. Phys. 60 1025 (1997) 19. B V Deryagin, N V Churaev Smachivayushchie Plenki (Wetting Films) (Moscow: Nauka, 1984) 20. B D Summ, E A Raud, in Uspekhi Kolloidnoi Khimii i Fiziko-Khimi- cheskoi Mekhaniki (Advances in Colloid Chemistry and Physico- chemical Mechanics) (Moscow: Nauka, 1992) p. 31 21. P Ehrenfest Ann. Phys. 61 440 (1920) 22. G E Gorelik Pochemu Prostranstvo Trekhmerno? (Why is Space Three-Dimensional?) (Moscow: Nauka, 1982) 23. V V Tarasov Zh.Fiz. Khim. 24 111 (1950) a 24. V V Tarasov Dokl. Akad. Nauk SSSR 46 22 (1945) b 25. V V Tarasov Dokl. Akad. Nauk SSSR 58 577 (1947) b 26. I M Lifshits Zh. Eksp. Teor. Fiz. 22 471 (1952) 27. I M Lifshits Zh. Eksp. Teor. Fiz. 22 475 (1952) 28. P S Laplace Exposition du Systeme du Monde. (Paris: Imprimeur- Libraire, 1835) 29. T Castro, R Reifenberger, E Choi, R Andres Phys. Rev. B, Condens. Matter 42 8548 (1990) 30. A N Goldstein, C M Echer, A P Alivisatos Science 256 1425 (1992) 31. R A Andrievskii Usp. Khim. 63 431 (1994) [Russ. Chem. Rev. 63 429 (1994)] 32. A I Gusev Usp. Fiz. Nauk 168 53 (1998) c 33. R F Khairutdinov Kolloid. Zh. 59 581 (1997) d 34. A I Gusev Nanokristallicheskie Materialy: Metody Polucheniya i Svoistva (Nanocrystalline Materials: The Methods of Preparation and Properties) (Ekaterinburg: Urals Branch of the Russian Academy of Sciences, 1998) 35.E Matiyevic' Langmuir 10 8 (1994) 36. M P Pileni J. Phys. Chem. 97 6961 (1993) 37. T Svedberg Obrazovanie Kolloidov (Formation of Colloids) (Leningrad: Research Chemical Technological Institute, 1927) 38. M Faraday Philos. Trans. 147 145 (1857) 39. R Zsigmondy Liebigs Ann. 301 29 (1898) 40. T Svedberg Berichte 38 3616 (1905) 41. P P von Veimarn Koll.-Zeitschr. 9 25 (1911) 42. S Mann Nature (London) 365 499 (1993) 43. A Henglein J. Phys. Chem. 97 5457 (1993) 44. R V Siegel Nanostruct. Mater. 3 1 (1993) 45. A D Pomogailo Usp. Khim. 66 750 (1997) [Russ. Chem. Rev. 66 679 (1997)] 46. R F Khairutdinov Usp.Khim. 67 125 (1998) [Russ. Chem. Rev. 67 109 (1998)] 47. T Tsuchida, J Kitayima J. Mater. Sci. 27 2713 (1992) 48. Z-C M Hu, M T Harris, C H Byers, J Sjoblom, S E Friberg, A Amran J. Disp. Sci. Technol. 16 31 (1995) 49. O B Pavlova-Verevkina, V V Nazarov, E D Politova, A V Pertsov Kolloid. Zh. 59 686 (1997) d 50. Z-C M Hu, M T Harris, C H Byers J. Colloid Interface Sci. 198 87 (1998) 51. O B Pavlova-Verevkina, E D Politova, V V Nazarov Kolloid. Zh. 62 515 (2000) d 52. K Kandori, A Yasukawa, T Ishikawa J. Colloid Interface Sci. 180 446 (1996) 53. M E Carlotti, S Morel, R Cavalli J. Disp. Sci. Technol. 14 35 (1993) 54. M C van der Leeden, G Mvan Rosmalen J. Colloid Interface Sci. 171 142 (1995) 55. J H Fendler Chem. Rev. 87 877 (1987) 56.V Peyre, O Spalla, L Belonli, M Nabavi J. Colloid Interface Sci. 187 184 (1997) 57. O Spalla, P Ke'kicheff J. Colloid Interface Sci. 192 43 (1997) 58. A A Abramzon Poverkhnostno-Aktivnye Veshchestva (Surfactants) (Leningrad: Khimiya, 1981) 59. J E Gordon The Organic Chemistry of Electrolyte Solutions (New York: Wiley-Interscience, 1977) 60. O Ya Samoilov Struktura Vodnykh Rastvorov Elektrolitov i Gidra- tatsiya Ionov (The Structure of Aqueous Solutions of Electrolytes and Hydration of Ions) (Moscow: Academy of Sciences of the USSR, 1957) 61. A I Rusanov Mitselloobrazovanie v Vodnykh Rastvorakh Poverkh- nostno-Aktivnykh Veshchestv (Micellisation in Aqueous Solutions of Surfactants) (St. Petersburg: Khimiya, 1992) 62.E E Zaev Kolloid. Zh. 62 175 (2000) d 63. E Szajdzinska-Pietek, R Maldonado, L Kevan, R R M Jones, M J Coleman J. Am. Chem. Soc. 107 784 (1985) 921 64. R R M Jones, R Maldonado E Szajdzinska-Pietek, L Kevan J. Phys. Chem. 90 1126 (1986) 65. L Kevan, P Baglioni Pure Appl. Chem. 62 275 (1990) 66. N Gorski, J Kalus J. Phys. Chem. B 101 4390 (1997) 67. K L Mittel (Ed.) Micellization, Solubilization and Microemulsions (New York: Plenum, 1978) 68. Structure and Reactivity in Reversed Micelles (Amsterdam: Elsevier, 1989) 69. H F Eicke, H Christen J. Colloid Interface Sci. 46 417 (1974) 70. H Christen, H F Eicke J. Phys. Chem. 78 1423 (1974) 71. B Derecskei, A Derecskei-Kovacs, Z A Schelly Langmuir 15 1981 (1999) 72. I Abou-Nemeh, H J Bart Langmuir 14 4451 (1998) 73.D Liu, J Ma, H Cheng, Z Zhao Colloids Surf. A 143 59 (1998) 74. Z A Schelly J. Mol. Liq. 72 3 (1997) 75. M Antonietty, S Lohmann, C van Niel Macromolecules 25 1139 (1992) 76. MAntonietty, WBremser, DMuÈ schenborn, C Rsenauer, B Schupp, M Schmidt Macromolecules 24 6636 (1991) 77. D Donescu, M Teodorescu, L Fusulan, C Petcu J. Disp. Sci. Technol. 20 1085 (1999) 78. C Sonesson, K Holmberg J. Colloid Interface Sci. 141 239 (1991) 79. A V Levashov Pure Appl. Chem. 64 1125 (1992) 80. G Headstrom, J P Slotte,M Baclund, O Molander, J B Rosenholm Biocatalysis 6 281 (1992) 81. T P Hoar, J H Schulman Nature (London) 152 102 (1943) 82. R Aveyard, T Lawless J. Chem. Soc., Faraday Trans. 1 82 1755 (1986) 83. A Pouchelon, J Meunierv, D Langevin Chem.Phys. Lett. 76 277 (1980) 84. B Medhage, M Almgren, Y Alsins J. Phys. Chem. 97 7753 (1993) 85. D Langevin, in Reverse Micelles (Eds P L Luisi, B E Straub) (New York: Plenum, 1984) 86. M-J Schwuger, K Stickdorn, R Schomicker Chem. Rev. 95 849 (1995) 87. J Gu, Z A Schelly Langmuir 13 4256 (1997) 88. P D I Fletcher, A M Hove, B H Robinson J. Chem. Soc., Faraday Trans. 1 83 985 (1987) 89. M A Middleton, R S Schechter, K P Johnson Langmuir 6 920 (1990) 90. L E Scriven Nature (London) 263 123 (1976) 91. T Hellweg, D Langevin Physica A 264 370 (1999) 92. C Tojo,M C Blanco, M A Lo'pez-Quintela Langmuir 13 4527 (1997) 93. M Boutonnet, J Kizling, P Stenius, G Marie Colloids Surf. 5 209 (1982) 94. R Toroude, P Girard, G Maire,M Kizling,M Boutonnet-Kizling, P Stenius Colloids Surf.67 19 (1992) 95. M P Pileni Langmuir 13 3266 (1997) 96. J B Nagy Colloids Surf. 35 201 (1989) 97. A Manna, B D Kulkarni, K Bandyopadhyay, K Vijayamohanan Chem. Mater. 9 3032 (1997) 98. P Lianos, J K Thomas Chem. Phys. Lett. 125 299 (1986) 99. C Petit, M P Pileni J. Phys. Chem. 92 2282 (1988) 100. T F Towey, A K Lodhi, B H Robinson J. Chem. Soc., Faraday Trans. 86 3757 (1990) 101. L Motte, C Petit, L Boulanger, P Lixon, M P Pileni Langmuir 8 1049 (1992) 102. A L Rogach, A Kornowski, M Gao, A EychmuÈ ller, H Weller J. Phys. Chem. B 103 3065 (1999) 103. K Kandori,K Kon-no,A Kitahara J. Colloid Interface Sci. 115 579 (1987) 104. K Kandori, K Kon-no, A Kitahara J.Colloid Interface Sci. 122 78 (1988) 105. LNErofeev,AVRaevskii,OI Kolesova, B S Grishin, T I Pisarenko Khim. Fiz. 15 106 (1996) e 106. J D Hopwood, S Mann Chem. Mater. 9 1819 (1997) 107. L Qi, J Ma, H Cheng, Z Zhao Colloids Surf. A 108 117 (1996) 108. H Yamauchi, T Ishikawa, S Kondo Colloids Surf. 37 71 (1989) 109. K Osseo-Asare, F J Arrigada Colloids Surf. 50 311 (1990) 110. F J Arrigada, K Osseo-Asare Colloids Surf. 69 105 (1992) 111. P A Ayyub, A N Maitra, D O Shah Physica C 168 571 (1990) 112. P Kumar, V Pillai, S R Bates, D O Shah Mater. Lett. 16 68 (1993) 113. P Kumar, V Pillai, D O Shah Appl. Phys. Lett. 62 675 (1993)922 114. F J Arriagada, K Osseo-Asare J. Disp. Sci. Technol. 15 59 (1994) 115. B SjoÈ stroÈ m, B BergenstaÊ hl, M Lindenberg, A Ê C Rasmuson J.Disp. Sci. Technol. 15 89 (1994) 116. L Qi, J Ma, H Cheng, Z Zhao J. Phys. Chem. B 101 3460 (1997) 117. S G Dixit, A R Mahadeshwar, S K Haram Colloids Surf. A 133 69 (1998) 118. D L Naylor J. Mater. Sci. 7 2288 (1992) 119. N Duxin, N Brun, C Colliex, M P Pileni Langmuir 14 1984 (1998) 120. S J Oldenburg Chem. Phys. Lett. 288 243 (1998) 121. V V Hardikar, E Matijevic' J. Colloid Interface Sci. 221 133 (2000) 122. I Langmuir J. Am. Chem. Soc. 39 1848 (1917) 123. A M Blinov Usp. Fiz. Nauk 155 443 (1988) c 124. G P Yampol'skaya, V N Izmailova, G Ts Raznikova, P V Nuss Izv. Acad. Nauk, Ser. Fiz. 59 109 (1995) 125. V V Arslanov Usp. Khim. 63 3 (1994) [Russ. Chem. Rev. 63 1 (1994)] 126. A-F Mingotand, C Mingotand, L K Patterson Handbook of Monolayers (Orlando: Academic Press, 1993) 127.V L Colvin, A N Goldstein, A P Alivisatos J. Am. Chem. Soc. 114 5221 (1992) 128. G Chumanov, K Sokolov, B W Gregory, T M Cotton J. Phys. Chem. 99 9466 (1995) 129. J Jang, F C Meldrum, J H Fendler J. Phys. Chem. 99 5500 (1995) 130. A D Pomogailo Usp. Khim. 69 60 (2000) [Russ. Chem. Rev. 69 35 (2000)] 131. R H Tredgold, R A Allen, P Hodge, E Khoshdel J. Phys. D, Appl. Phys. 20 1385 (1987) 132. K B Blodgett J. Am. Chem. Soc. 57 1007 (1935) 133. Z Ya Berestneva, V A Kargin Usp. Khim. 24 249 (1955) 134. N V Belov (Ed.) Protsessy Real'nogo Kristalloobrazovaniya (Processes of Real Crystal Formation) (Moscow: Nauka, 1977) 135. S D Burnside, V Shklover, C Barbg, P Comte, F Arendse, K Brooks, M GraÈ tzel Chem.Mater. 10 2419 (1998) 136. M M Khvorov, R R Nigmatullin Fiz. Tv. Tela 32 2294 (1990) 137. V I Trefilov, G I Savvakin, V V Skorokhod, Yu M Solonin, A F Khrienko Dokl. Akad. Nauk SSSR 239 838 (1978) b 138. J P Caselier J. Disp. Sci. Technol. 14 625 (1993) 139. L Fernandez-Diaz, A Putnis, T J Cumberbatch Eur. J. Miner. 2 495 (1990) 140. G G Stefanov, D M Dimitrova, S B Borisova, G G Ivanov Polym. J. 28 732 (1996) 141. N I Ivanova, A F Putilin, E D Shchukin Kolloid. Zh. 58 613 (1996) d 142. H FuÈ redi-Milhofer, V B Ivancic, L Brecevic, N Filipovic'- Vincekovic', D Kralj, L Komunjer,M Marcovic, D Skrtic Colloids Surf. 48 219 (1990) 143. H FuÈ redi-Milhofer, R Bloch, D Skrtic, N Filipovic'-Vincekovic', N Garti J.Disp. Sci. Technol. 14 355 (1993) 144. D Anvir (Ed.) The Fractal Approach to Heterogeneous Chemistry. Surfaces, Colloids, Polymers (Chichester; New York; Brisbane; Toronto; Singapore: Wiley, 1997) 145. I V Melikhov, E N Kitova, A N Kamenskaya, E D Kozlovskaya, N B Mikheev, S A Kulyukhin Kolloid. Zh. 59 780 (1997) d 146. I V Melikhov, I E Mikheeva, V N Rudin Kristallografiya 34 1272 (1989) f 147. S P Gubin Khimiya Klasterov (The Chemistry of Clusters) (Moscow: Nauka, 1987) 148. P Jena, B K Pao, S N Khanna Physics and Chemistry of Small Clusters (New York: Plenum, 1987) 149. Yu I Petrov Klastery i Malye Chastitsy (Clusters and Small Particles) (Moscow: Nauka, 1986) 150. S P Gubin, I D Kosobudskii Usp. Khim.52 1350 (1983) [Russ. Chem. Rev. 52 766 (1983)] 151. V V Smirnov, L A Tyurina Usp. Khim. 63 57 (1994) [Russ. Chem. Rev. 63 55 (1994)] 152. S B Brichkin, V F Razumov, M G Spirin Kolloid. Zh. 62 12 (2000) d 153. S B Brichkin, V F Razumov, M G Spirin, M V Alfimov Dokl. Akad. Nauk 358 198 (1998)b 154. B D Summ Ros. Khim. Zh. 40 57 (1996) g 155. L BergstroÈ m Adv. Colloid Interface Sci. 70 125 (1997) 156. B V Deryagin, N V Churaev, V M Muller Poverkhnostnye Sily (Surface Forces) (Moscow: Nauka, 1985) B D Summ, N I Ivanova 157. P A Rehbinder Izbrannye Trudy. Poverkhnostnye Yavleniya v Dis- persnykh Sistemakh. Fiziko-Himicheskaya Mekhanika Dispersnykh Sistem (Selected Proceedings. Surface Effects in Dispersed Systems. Physicochemical Mechanics of Dispersed Systems) (Moscow: Nauka, 1979) 158. J Fridel Dislocations (New York: Pergamon Press, 1964) 159.R C Tolman J. Chem. Phys. 17 333 (1949) 160. B Chalmers Teoriya Zatverdevaniya (The Theory of Hardening) (Moscow: Metallurgiya, 1968) 161. B V Deryagin Teoriya Ustoichivosti Kolloidov i Tonkikh Plenok (The Theory of Stability of Colloids and Thin Films) (Moscow: Nauka, 1986) 162. B V Deryagin, L D Landau Zh. Eksp. Teor. Fiz. 15 662 (1945) 163. E J W Verwey, J Th G Overbeek Theory of Stability of Hydrophobic Colloids (Amsterdam, 1948) 164. V S Veselovskii, V N Pertsov Zh. Fiz. Khim. 8 245 (1936) a 165. A D Sheludko Kolloidnaya Khimiya (Colloid Chemistry) (Moscow: Mir, 1984) 166. A D Sheludko, B V Toshev, D Platikanov, in Sovremennaya Teoriya Kapillyarnosti (Modern Theory of Capillarity) (Eds A I Rusanov, F Ch Gudrich) (Leningrad: Khimiya, 1980) p.274 167. I V Petryanov, N P Rozenblyum Dokl. Akad. Nauk SSSR 61 661 (1948) b 168. M Volmer Kinetic der Phasenbildung. (Ann Arbor: Edwards Brothers, 1945) 169. X Z Wu, B M Ocko, E B Sirota S K Sinha,M Deutsch, B H Cao, M W Kim Science 261 1018 (1993) 170. M J Regan, E H Kawamoto, S Lee, P S Pershan, N Maskil, M Dentsch, O M Magnussen, B M Ocko, L E Berman Phys. Rev. Lett. 75 2498 (1995) 171. B D Summ Vestn. Mosk. Univ., Ser. 2, Khim. 34 59 (1993) h 172. B D Summ Vestn. Mosk. Univ., Ser. 2, Khim. 40 400 (1999) h 173. B D Summ, E Yu Il'ichev Zh. Fiz. Khim. 70 316 (1996) a 174. V K Semenchenko Poverkhnostnye Yavleniya v Metallakh i Spla- vakh (Surface Effects in Metals and Alloys) (Moscow: Gostekhteorizdat, 1957) 175. A I Bykhovskii Rastekanie (Spreading) (Kiev: Naukova Dumka, 1983) 176. P G de Gennes Rev. Mod. Phys. 57 827 (1985) 177. G A Aksel'rud,M A Al'tshuler Vvedenie v Kapillyarno-Khimiche- skuyu Tekhnologiyu (Introduction in Capillary Chemical Technology) (Moscow: Khimiya, 1983) 178. Yu A Chizmadzhev, V S Markin,MR Tarasevich, Yu G Chirkov Makrokinetika Protsessov v Poristykh Sredakh (Macrokinetics of Processes in Porous Media) (Moscow: Nauka, 1971) 179. S A Kukushkin, V V Slezov Dispersnye Sistemy na Poverkhnosti Tverdykh Tel (Evolyutsionnyi Podkhod): Mekhanizmy Obrazova- niya Tonkikh Plenok [Dispersed Systems on the Surface of Solids (Evolution Method) The Mechanisms of Formation of Thin Films] (St Petersburg: Nauka, 1996) 180. N A Fuchs Mekhanika Aerozolei (The Mechanics of Aerosol) (Moscow: Academy of Sciences of the USSR, 1955) 181. T Nakayama, A Nakahara,M Matsushita J. Phys. Soc. Jpn. 64 1114 (1995) 182. M Y Lin,H M Lindsay,D A Weitz, R C Ball,R Klein, P Meakin Phys. Rev. A, Gen. Phys. 41 2005 (1990) 183. D A Weitz, M Y Lin, H M Lindsay Chemometr. Intell. Lab. Syst. 10 133 (1991) 184. I F Efremov Periodicheskie Struktury (Periodical Structures) (Leningrad: Khimiya, 1971) 185. H Zocher Z. Anorg. Allgem. Chem. 147 91 (1925) 186. J D Bernal, I Fankuchen Nature (London) 139 923 (1937) 187. I Das, P Singh, N R Agrawal, R P Rastogi J. Colloid Interface Sci. 192 420 (1997) 188. S Kai, S C MuÈ ller, J Ross J. Phys. Chem. 87 806 (1983) 189. I Das, A Pushkarna, R S Lall J. Cryst. Growth 82 361 (1988) 190. I Das, A Pushkarna, N R Agrawal J. Phys. Chem. 93 7269 (1989) 191. S C MuÈ ller, S Kai, J Ross Science 216 635 (1982) 192. I Das, S Chand, A Pushkarna J. Phys. Chem. 93 7435 (1989) 193. T A Aizatullin, V L Lebedev, K M Khailov Okean. Aktivnye Poverkhnosti i Zhizn' (Ocean Active Surface and Life) (Leningrad: Gidrometeoizdat, 1979)923 The use of objects and methods of colloid chemistry in nanochemistry 194. I V Melikhov, B M Dolgonosov Dokl. Akad. Nauk SSSR 299 1175 (1988) b 195. I V Melikhov, B M Dolgonosov Kolloid. Zh. 51 911 (1989) d 196. T Serra, J Colomer, X Casamitijana J. Colloid Interface Sci. 187 466 (1997) 197. H Tateyama, P J Scales, M Ooi, S B Johnson, K Rees, D V Boger, Th W Healy Langmuir 13 6393 (1997) 198. A K Arora, B V R Tata Adv. Colloid Interface Sci. 78 49 (1998) 199. Z N Markina, in Uspekhi Kolloidnoi Khimii (Advances in Colloid Chemistry) (Moscow: Nauka, 1973) p. 234 200. L De Maeyer, C Trachimov, U Kaatze J. Phys. Chem. B 102 8480 (1998) 201. A I Rusanov, F M Kuni, A K Shchekin Kolloid. Zh. 62 199 (2000) d 202. F M Kuni, A K Shchekin, A P Grinin, A I Rusanov Kolloid. Zh. 62 204 (2000) d 203. V P Tikhonov, in Uspekhi Kolloidnoi Khimii (Advances in Colloid Chemistry) (Tashkent: FAN, 1987) p. 310 204. V A Volkov, in Uspekhi Kolloidnoi Khimii (Advances in Colloid Chemistry) (Leningrad: Khimiya, 1991) p. 185 205. P Hansson Langmuir 14 2269 (1998) a�Russ. J. Phys. Chem. (Engl. Transl.) b�Dokl. Chem. Technol., Dokl. Chem. (Engl. Transl.) c�Physics-Uspekhi (Engl. Transl.) d�Russ. Colloid. J. (Engl. Transl.) e�Chem. Phys. Rep. (Engl. Transl.) f�Crystallogr. Rep. (Engl. Transl.) g�Mendeleev Chem. J. (Engl. Transl.) h�Moscow Univ. Bull.

ISSN:0036-021X    

出版商:RSC    

年代:2000

数据来源: RSC

摘要:

Russian Chemical Reviews 69 (11) 925 ± 945 (2000) Dynamic and stimulated nuclear polarisation in photochemical radical reactions E G Bagryanskaya, R Z Sagdeev Contents I. Introduction II. Experimental implementation of the methods III. Dynamic nuclear polarisation IV. Stimulated nuclear polarisation V. Chemically induced dynamic nuclear polarisation in switched external magnetic fields VI. Conclusion Abstract. tech- resonance magnetic time-resolved sensitive Highly Highly sensitive time-resolved magnetic resonance tech- niques based on the effect of variable radio-frequency and niques based on the effect of variable radio-frequency and switched polarisation nuclear the on fields magnetic constant switched constant magnetic fields on the nuclear polarisation of of diamagnetic considered.are reactions radical of products diamagnetic products of radical reactions are considered. The The applications of investigation the to methods these of applications of these methods to the investigation of short-lived short-lived radical radicals free biradicals, pairs, (radical intermediates radical intermediates (radical pairs, biradicals, free radicals and and radical in reactions photochemical in formed ions) radical ions) formed in photochemical reactions in homogeneous homogeneous and The discussed. are media molecular-organised and molecular-organised media are discussed. The bibliography bibliography includes references 148 includes 148 references. I. Introduction The review is devoted to the use of new time-resolved magnetic resonance techniques based on the detection of nuclear polar- isation of diamagnetic products of radical reactions for the study of spin dynamics and chemical kinetics of short-lived radical species (free radicals, radical pairs and radical ion pairs) in photochemical radical reactions. The demand for investigation of fast photochemical reactions in solutions has stimulated the appearance and extensive development of methods for the detec- tion of short-lived radical species, for example, laser flash photol- ysis, time-resolved EPR, reaction yield detected magnetic reso- nance methods (RYDMR), etc.The progress in the investigations into the mechanisms of photochemical radical reactions is largely associated with the use of magnetic resonance { because these methods provide information on the structure of radicals, the rates of chemical reactions, electron exchange interaction, the parameters determining the mobility of radical species in solu- tions.Methods based on the application of variable resonance magnetic fields have been greatly developed during the last two decades.1± 12 A radio-frequency (RF) resonance field induces nuclear polarisation of in-cage (M1) and escape (M2, M3 ) dia- magnetic products of a photochemical reaction.Polarisation arises either upon RF-induced changes in the singlet ± triplet conversion rate in radical pairs (selectively with respect to the E G Bagryanskaya, R Z Sagdeev International Tomography Centre, Siberian Branch of the Russian Academy of Sciences, ul.Institutskaya 3a, 630090 Novosibirsk, Russian Federation. Fax (7-383) 233 13 99. Tel. (7-383) 233 34 25. E-mail: elena@tomo.nsc.ru (E G Bagryanskaya) Received 25 April 2000 Uspekhi Khimii 69 (11) 1009 ± 1031 (2000); translated by Z P Bobkova #2000 Russian Academy of Sciences and Turpion Ltd DOI 10.1070/RC2000v069n11ABEH000599 925 926 926 930 938 943 orientation of nuclear spins) or upon electron or nuclear tran- sitions in the free radicals R1, R2 (Scheme 1). The magnitude and the sign of nuclear polarisation induced by an RF field depend on the field frequency and on the type of resonance transition (nuclear or electronic) to which this frequency corresponds. Scheme 1 RF field 1 3 [R1 ..R2 ] [R1 . .R2 ] M M* M1 (in-cage products) R1+R2 M2, M3 (escape products) RF field There are several methods for detecting magnetic resonance spectra of short-lived radical species. (i) Stimulation of nuclear transitions in polarised intermediate short-lived radicals with simultaneous recording of the NMR spectra of diamagnetic products.13 ± 18 This method was first proposed and implemented in a study by Sagdeev et al.13 and was developed further by Trifunac et al.14 ± 18 An obvious advant- age of the method is high sensitivity (below 108 spins in the sample) and simplicity of line assignment � the lines of the product detected byNMRcorrespond to a particular nucleus in the radical precursors. The main shortcoming of the method is that its scope is limited to radicals the lifetimes of which are not shorter than several microseconds. (ii) The method of dynamic nuclear polarisation (DNP).Pumping of electron transitions in the intermediate short-lived radicals gives rise to nuclear polarisation, which is transferred to diamagnetic products after radical recombination.12, 19 ± 22 The dependence of the signal intensity in the NMR spectrum of diamagnetic products on the external magnetic field is an EPR spectrum of the intermediate radicals. This method provides information on the mechanism of electron ± nucleus interaction in short-lived radicals (contact or dipole ± dipole); it opens the way for determining the relaxation times, the signs and the magnitudes of hyperfine coupling (HFC) constants and for investigation of the kinetics of these radicals in solution.(iii) The method of stimulated nuclear polarisation (SNP). The effect of an RF field with a specified frequency on an ensemble of { The sensitivity of EPR (109 spins in a sample) does not always suffice for the investigation of fast radical processes.926 radical pairs changes nuclear polarisation in the products formed.12, 23 ± 25 By investigating the dependence of nuclear polar- isation of the diamagnetic products on the frequency of the saturating field, one can gain information on the EPR spectrum of the intermediate radical pair, exchange interaction, pathways to diamagnetic products, etc. II. Experimental implementation of the methods For recording the spectra of short-lived radicals, it was necessary to devise instruments which either produce a variable resonance magnetic field at radio frequency or ensure the possibility of switching an external magnetic field directly during a radical reaction with simultaneous recording of the NMR spectra of diamagnetic products. Coupling of a resonator or an oscillator circuit with the probe of a standard NMR spectrometer was a complicated engineering problem.This was done by Sagdeev et al.13 in the first experiment on recording the NMR spectra of short-lived radicals. In scanning the resonance lines, the frequency sweep of the RF field was accomplished manually; due to the restrictions in the frequency range of the resonance circuit of the NMR spectrometer, only one resonance line was detected.The separate pumping of resonance transitions in radical pairs and recording the NMR spectra of the reaction products proved to be an easier procedure (Fig. 1). A fast-flow system is normally used to transfer the sample into the photolysis cell. A liquid is injected into the cell through a thin Teflon capillary either using a pump 17, 20, 26 or under a gas pressure.12 The flow rate of the liquid should be such that the time it takes to transfer the sample from the reaction vessel to the probe of the NMR spectrometer be shorter than the nuclear relaxation time of diamagnetic products, i.e. this should require fractions of a second. An original setup for detecting SNP in weak magnetic fields of about 10.0 mT has been developed.27 The transfer of a sample from the magnetic field in which the photochemical reaction took place to the probe of the NMRspectrometer was performed using the pneumatic system of the probe.The frequency of pumping of electron transitions (300 MHz) was equal to the resonance frequency of the NMR spectrometer. The limitations of this method include the use of only one resonance frequency of an RF field. The magnetic flux chosen (B0) is largely determined by the goals of the experiment. The resonance frequency employed in existing experimental setups ranges from 50 MHz to 10 GHz. A3 1 hn 4 2 11 (I) 3 8 6 10 9 11 (II) 5 7 Figure 1. Block diagram of the setup for the study of SNP: (1) photolysis cell with a sample; (2) RF resonator or oscillating circuit; (3) electro- magnet; (4) laser (308 nm); (5) superconductive magnet and the probe of a Bruker MSL-300 NMR spectrometer (300 MHz); (6) RF generator; (7) control computer of the NMR spectrometer; (8) pulse generator; (9) an oscilloscope; (10) power supply of the magnet and B0 sweep ; (11) fast-flow system consisting of a vessel (I ) with the initial substances (argon feed under a pressure of*1.5 atm) and a vessel for the collection of the reaction products (II ).E G Bagryanskaya, R Z Sagdeev a b t RF field RF field Time Laser pulse Time Laser pulse Figure 2. Schematic illustration of the two methods for recording the spectra: continuous (a) and pulse (b).In pulse methods 28 (Fig. 2), the dependence of the intensity of NMRspectra on the time delay (t) between the laser pulse and the RF pulse is studied. The main difference between the pulse and steady-state methods is as follows. When the continuous variant is used, radical pairs and the intermediate short-lived radicals exist throughout their lifetimes under the influence of an RF field, which induces the singlet ± triplet conversion in radical pairs or the DNP. The use of pulse modification provides a new, in principle, opportunity to expose separate sub-ensembles of radical pairs (or radicals) with a particular lifetime to an RF field. The main advantages of the pulse variant over the steady-state method include differentiation of geminate and diffusion pairs by measur- ing the SNP for different delays t; separation of the contributions of SNP andDNPto the RFfield-induced nuclear polarisation; the possibility of obtaining information on the pattern of distribution of radical pairs and radicals over lifetimes.In the case of pulse method, time resolution is determined by the rise times of the microwave pulse and the laser pulse and equals 20 ± 30 ns. High time resolution can be achieved by using either cavities with low Q-factors or cavities with high Q-factors but with a higher coupling constant between the resonator and the generator.29 Upon mathematical processing of the data, the time resolution of the method can be improved to 5 ns.27 Asubstantial drawback of the above methods of recording the spectra is the necessity of using large samples. The use of deuterated solvents markedly increases the cost of the experiment; however, this problem can be solved by taking solvents containing no protons (CCl4, C6F6 and so on) or by recording the spectra for other nuclei (13C, 31P, 19F and so on.).In the case where theNMR chemical shifts of the solvent and the compound under study differ by several ppm (at least, by 1 ppm), the signal of the solvent should be suppressed. III. Dynamic nuclear polarisation The DNP effect means that saturation of the electron levels of free radicals in an external magnetic field gives rise to a non-equili- brium nuclear polarisation.30, 31 A radio-frequency resonance magnetic field makes the occupancies of the electron levels equal and thus creates non-equilibrium occupation.Cross-relaxation transitions caused by hyperfine coupling of an electron with the nucleus are directed at restoring the thermal equilibrium between the electron levels. However, since cross-relaxation processes are accompanied by inversion of nuclear spins, polarisation arises on the radical nuclei, p=xsf ge gn, where, the parameter x is determined by the mechanism of the electron ± nucleus coupling, s=(S07<Sz>)/S0 is the saturation factor of the electron spin system, f is the leakage factor determin- ing the fraction of nuclear relaxation occurring without partic- ipation of electron spins, and ge and gn are gyromagnetic ratios of the electron and nuclear spins.For contact interaction [the change in the total spin of the electron and the nucleus (Dm) is equal to zero], x=1, while for dipole interaction (Dm=2), x=71/2. Figure 3 illustrates qualitatively the mechanism of formation of the DNP upon non-equilibrium electron polarisation of a radical with two equivalent magnetic nuclei (I1=I2=1/2) and positiveDynamic and stimulated nuclear polarisation in photochemical radical reactions mSmI mSmI +1 +1 +1/2 +1/2 0 0 mI mI 71 71 710 710 1 1 2 +1 +1 2 71 71 71/2 71/2 0 +1 0 +1Radical Diamagnetic molecule Radical Diamagnetic molecule Figure 3. Schematic illustration for the mechanism of the generation of multiplet DPN of the diamagnetic products formed from short-lived radicals.The arrows show the resonance transitions induced by an RF field (1) and cross-relaxation transitions (2). HFC constants (A1=A2).32 Cross-relaxation occurs by a contact mechanism (Dm=0). Study of the dependences of DNP of stable radicals of para- magnetic ions on the temperature, solvent viscosity, and the magnetic field strength makes it possible to solve diverse prob- lems, for example, to study the effect of the medium on the interaction of various molecules with free radicals, to study the roles of the contact and dipole interactions during relaxation, to determine the mechanisms of interaction between radicals and nuclei, etc.{ 1. Dynamic nuclear polarisation of neutral intermediate short-lived radicals Grishin et al.19 was the first to consider the possibility of using the DNP effects for the investigation of intermediate short-lived radicals.Photolysis of benzaldehyde in CCl4 affords the radical pairs [PhCH(OH) PhCO].19 a-Hydroxybenzyl radicals are con- verted into the starting benzaldehyde upon hydrogen atom trans- fer. In strong magnetic fields (330 mT), DNP can be detected in theNMRsignals of all the diamagnetic reaction products.19, 33 ± 35 The negative sign of DNP corresponds to a dipole mechanism of cross-relaxation in the intermediate a-hydroxybenzyl radical. In weak and medium magnetic fields (3 and 60 mT) in which the cross-relaxation efficiency is much lower, DNP is negligibly small, the predominant contribution to polarisation being made by SNP.As noted above, an RF field induces nuclear polarisation both by the DNP mechanism, due to cross-relaxation processes in free radicals, and by the SNP mechanism, due to the change in the rate of singlet ± triplet transitions in radical pairs. The contributions of these mechanisms to the polarisation of the nuclei of diamagnetic products formed from short-lived radicals can be separated based on the dependence of the width of resonance lines and the SNP and DNP intensities on the RF field amplitude (B1). For example, photolysis of benzaldehyde in CHCl3 has been used to study the influence of the RF-field amplitude on the ratio of the SNP and DNP contributions and on the width of individual components of the spectrum.20 This permitted the researchers to estimate the lifetime of the radical pairs (*20 ns).The possibility of separating the contributions of DNP and SNP to the total polarisation by varying the external magnetic field, the RF-field amplitude and the lifetimes of radicals was demonstrated 12 in relation to the photolysis of benzoquinone in solutions. For semiquinoid radicals, the efficiency of cross-relax- ation in 50.0 mT fields becomes substantial, and the observed DNP is comparable in magnitude with SNP. For durosemiqui- { The DNP studies of stable radicals are surveyed comprehensively in Ref. 31. 927 noid radicals, cross-relaxation is due to the modulation of HFC caused by rotation of the methyl groups (i.e. cross-relaxation follows a contact mechanism); in this case, a positiveDNP effect is observed.20, 21, 31 ± 36 Information on electron relaxation times in short-lived radi- cals and on the rate of chemical reactions can be gained by using pulsed generation of radicals and by investigating the variation of DNP intensity in the magnetic field corresponding to the max- imum of the resonance line vs.the time delay between the laser pulse and the leading edge of the RF pulse. These experiments for the neutral radicals generated upon photolysis of benzaldehyde in chloroform were first performed by Yamakage et al.37 The experimental results were processed by calculations in terms of the kinetic scheme for the occupancies of the radical levels with allowance for cross-relaxation, electron transitions and the chem- ical reaction.It was shown that the DNP kinetics actually reflect the radical kinetics and do not depend on the cross-relaxation rate, while the DNP intensity is determined by the ratio of the cross- relaxation, electron relaxation and the chemical reaction rates. The results provided grounds for concluding that the cross- relaxation rate for the a-proton of the a-hydroxybenzyl radical is one or two orders of magnitude lower than that for aromatic protons. Cross-relaxation processes influence substantially the chemi- cally induced dynamic electron polarisation (CIDEP) of short- lived radicals 38 ± 40 and, in some cases, account for the change in its sign. Thus McLauchlan et al.39, 40 found the change in the CIDEP sign in the photolysis of aliphatic ketones and suggested that this is due to the cross-relaxation in the intermediate radical.Subsequently, in a theoretical study of the cross-relaxation processes, it was shown 41 that CIDNP and CIDEP are related in time and their kinetics should be considered jointly. Thus, the transfer of electron polarisation to the nuclei should substantially affect the CIDNP kinetics. Experimental research on photolysis of aliphatic ketones by the time-resolved CIDNP method gave ambiguous results. Tsentalovich et al.42 ± 46 concluded that, when analysing the kinetics of CIDNP detected during the photolysis of acetone in isopropanol and 2,4-dihydroxy-2,4-dimethylpenta- none, one should take into account the contribution of the cross- relaxation mechanism. However, Fischer et al.47 showed that the CIDNP kinetics can be adequately described without allowance for cross-relaxation; only two-photon processes resulting in CIDNP of some reaction products should be taken into account.In addition, CIDEP with a positive sign is observed in the photolysis of 2,4-dihydroxy-2,4-dimethylpentanone, which refutes the conclusions about generation of CIDNP by the cross- relaxation mechanism made by Salzmann.45 The main difficulty in comparison of the results obtained by the CIDEP and CIDNP methods is that the fields used in CIDEP experiments are about 330 mT, while the operating fields of NMR spectrometers are usually in the range 3 ± 7 T. The electron and cross-relaxation times are known to depend appreciably on the external magnetic field, while the processes of polarisation transfer are, in turn, determined by these times.The use of the DNP method clarified the point concerning the nature of inversion of the CIDEP sign.32 A study of the DNP in reactions involving Me2XC radicals (X=Me, OH, D, COMe) demonstrated unambiguously the presence of effective scalar (Dm=0) electron ± nuclear cross-relaxation in the tert-butyl and 2-hydroxyprop-2-yl radicals and the absence of cross-relaxation in 2-acetyl and prop-2-yl radicals. Figure 4 shows the DNP and EPR spectra recorded in the X and L ranges for the Me2(OH)C radical, resulting from photolysis of 2,4-dihydroxy-2,4-dimethyl- pentanone in benzene.It can be seen that the signs of polarisation coincide and the intensity ratios of the lines in the spectra are closely similar. Comparison of the obtained results with the results of time-resolved CIDNP studies led to the conclusion that the cross-relaxation is due to the modulation of the isotropic HFC with a characteristic correlation time of tc441 ps.928 330 335 Magnetic field /mT Figure 4. Spectra of 2-hydroxyprop-2-yl radical recorded in the X and L ranges by the time-resolved EPR (a, b) and DNP (c, d ) methods. The resonance frequency is 10 (a, c), 1.6 (b) and 1.53 GHz (d ). 2. Dynamic nuclear polarisation in radical ions Radical-ion reactions are accompanied, in the majority of cases, by degenerate electron exchange (DEE) reactions D+D , D+D A+A A+A (D is donor, A is acceptor); therefore, the DNP in radical-ion reactions has several specific features.In a study of the influence of an RF field on the nuclear polarisation of the products of radical-ion reactions in weak magnetic fields, a pronounced DNP effect was found.48 Further studies demonstrated 22, 49 that this effect is observed in the presence of DEE provided that the magnitide of the external magnetic field is comparable with the HFC constant of the radical ion.Two different approaches have been proposed for quantita- tive description of DNP under the conditions of DEE in weak magnetic fields.22, 49 One approach is based on numerical solution of the Liouville equation for the spin density matrix of the radical with allowance for the spin dynamics in a weak magnetic field and the DEE process, while the other approach includes solution of the Bloch phenomenological equations with allowance for the relaxation times Te e 1 Te1 à 2D2t 1 á o2 e 1 Te2 à D2t 1 á 1 á o2 1 where D2 à 13 i X spectrum of the radical ion for nuclei with spins Ii and HFC constants Ai , te is the DEE time, and o0 is the Larmor frequency of electron spins.The former approach is applicable to any DEE rates, while the latter one is suitable only for high exchange rates (D2t 2e 55 1). The main advantages of the latter model include its simplicity and the possibility of using it to study radicals with complex EPR spectra.Comparison of the DNP spectra calculated in terms of these two models showed that at high rates of electron exchange, the spectral patterns are identical; in this case, the use of the phenomenological approach is quite justified. Note that both models take into c d a b account only processes which occur in radical ions, the influence of the RF field on the radical pairs being ignored. Of course, this assumption is rather simplified because the RF field induces singlet ± triplet transitions in the radical ion pairs and, thus, it can, in principle, induce CIDEP.50 The line shape in the DNP spectrum in the absence of CIDEP was compared to that obtained with the assumption that CIDEP does not depend on the fields B1 and B0 . It was shown that the distinctions become significant in the case of DEE and large B1 values, i.e.in those cases where the spectrum width becomes comparable with the field magnitude B0 . 60 55 50 340 345 Magnetic field /mT 1, T e2 1 ; 0t2e , 0t2e ! IiÖIi á 1ÜA2i is the second moment of the EPR E G Bagryanskaya, R Z Sagdeev The DNP effects have also been studied in the photolysis of anthracene with dimethyl- or diethyl-aniline and in the photo- sensitised cis ± trans isomerisation of fumaronitrile in the presence of naphthalenes.22, 24, 49, 51 The researchers studied the influence of the electron exchange rate and the amplitude of theRFfield on the pattern of DNP spectra; the results were in good agreement with theoretical predictions.The dependence of the half-width of the spectrum on the reciprocal concentration of diamagnetic mole- cules participating in the charge exchange was used to determine the rate constants for the electron exchange of the dimethyl- and diethyl-aniline radical cations. The sign of CIDEP is known to be determined by the multi- plicity of the radical pair and by the sign of exchange interaction.52 Upon the change in multiplicity, the sign of CIDEP and, hence, the sign of DNP should be reversed. This was observed,25, 51 for example, in the photosensitised isomerisation of fumaronitrile in 1,4-dimethoxynaphthalene. The triplet state of the excited 1,4- dimethoxynaphthalene molecule is lower in energy than the radical-ion pair; therefore, by changing the concentration of fumaronitrile, one can change the multiplicity of the radical pair.53 The measurement of the DEE rate constants can be markedly simplified by proceeding from the dependence of theDNPkinetics on the concentration of the diamagnetic product.This method has advantages over investigation of the concentration dependence of the half-with of the DNP spectra because the half-width of the spectra depends appreciably on the amplitude of the RF field B1, whereas the DNP kinetics do not. For the first-order reaction of radical decay, an analytical solution was found; for the second- order reaction, a numerical solution was obtained.49 It was shown that the DNP kinetics are determined by the time of electron exchange te , the time of electron relaxation Te1 and the lifetime of radical ions tl.For high concentrations of diamagnetic molecules involved in DEE, the kinetics are described by an exponential function with a decay parameter equal to Te1. Figure 5 shows the experimental curves for the DNP kinetics of the photolysis of fumaronitrile with perdeuterated naphthalene. Any curve for t1>100 ns can be described by the exponential function exp(7kobst1), where kobs is the observed rate constant for the exponential decay. The initial pattern of the curve presented in Fig. 5 b is due to the time for generation of radical ions. At high concentrations of molecules participating in charge exchange, i.e. at low times of DEE, the condition T e155tl holds, T e1 and the experimental dependences of kobs on the concentration of fumar- onitrile being adequately described by the formula kobs ^ 1 à 2D2te .Te1 0t2e 1 á o2 The rate constant for DEE ke is equal to (4.520.23)6109 litre mol71 s71 and is close to the value estimated from the concentration dependence of the spectra. In a series of studies,25, 54, 55 the DNP method together with time-resolved EPR were employed to identify the radical species resulting from photolysis of heteroaromatic azines in the presence of dimethylaniline in polar solvents. In the DNP study of all the compounds, except for phenazine, the recorded spectra of the dimethylaniline radical cation were narrowed down due to the DEE reaction. This provided grounds for conclusion that the radical-ion pathway contributes to the formation of the com- pounds under consideration.Dynamic and stimulated nuclear polarisation in photochemical radical reactions a DNP (rel.u.) 1 0.1 b DNP (rel.u.) 0.50 70.5 71.0 c DNP (rel.u.) 0.50 0.5 1.0 70.5 71.0 Figure 5.Time dependences of DNP obtained for the photolysis of naphthalene-d8 (c=561074 mol litre71) with fumaronitrile in CD3CN at different DEE rates. Fumaronitrile concentration (mol litre71): (a) 561073 (1), 261072 (2), 461072 (3); (b) 561073; (c) 461072. The absolute magnitudes of DNP are plotted on a logarithmic scale (a). The continuous lines (b, c) correspond to the decay of radical ions by an exponential law, dashed lines show the second-order decay of radical ions and describe the contributions of the minor geminate (dashed line) and diffusion polarisations (chain line).3. Electron ± nuclei transitions in short-lived radicals in weak magnetic fields The influence of an RF field on the nuclear polarisation of diamagnetic products of radical reactions in weak magnetic fields, whose magnitudes are comparable with the HFC constants of radicals, possesses several specific features: the intrinsic wave functions of radicals are no longer the products of the electronic and nuclear wave functions, and the resonance transitions are not purely electronic or nuclear transitions. Therefore, all transitions in these fields involve both the nuclear and the electronic spins.Therefore, the use of relatively low RF-field amplitudes results in substantial changes in the occupancy of the electron ± nuclear levels. This effect was called NMR detected electron ± nuclear transitions in short-lived radicals (NMR ENR).56, 57 The diagram of the electron ± nuclear spin levels of a radical with one magnetic nucleus with the spin 1/2 is shown in Fig. 6. In a weak magnetic field, the probabilities (P) of transitions n1>n2 and n3>n4 on exposure to a resonance field are proportional to the expression Pni >nj *j < Cni j ^ B1xÖ ^ Sx á ^IxÜjCnj > j2 à à B1 2âc21 ÖgebeÜ2 á c22 ÖgnbnÜ2ä , where ni , nj correspond in pairs to n1, n2 and n3, n4 ; Cni , j are the wave functions of the ni, j states; the coefficients c1 and c2 are determined by the relationship between the HFC constant and the B0 value.When the magnetic field changes, the coefficients c21 and c22 change from 1 and 0, respectively, in strong magnetic fields (where n1>n2 and n3>n4 are purely nuclear transitions) to 0.5 and 0.5, respectively, in the zero magnetic field in which these transitions involve changes in the projections of both electron and nuclear spins. In a strong magnetic field (c1=1, c2=0), sub- 123 1.5 t1 /ms 929 E(A) n4 aean n3 c1aebn+c2bean A/4 0 n2 bebn 3A/4 n1 c1bean7c2aebn 0.5 0 1.0 B0/A Figure 6. Scheme of the energy levels of a radical with one magnetic nucleus and the spin 1/2 in a magnetic field. The arrows show all the transitions allowed in a weak field.stantial amplitudes of the radio-frequency field are required to saturate these transitions in short-lived radicals B1& 1 gnt , where t1 à T1n1 á t1l á T1cr , T n1 is the time of nuclear relaxation of the radical, tl is the lifetime of the radicals and T cr is the time of electron ± nuclear cross- relaxation. With the real values T n1 &1074 s, tl&1075± 1076 s, Tcr&1075 s, we have B1&1 ± 10 mT. In weak magnetic fields (c 21 &c 22 ), the probability of transitions is high even for much smaller amplitudes of the radio-frequency field B1* 1 c1get , where t1 à T1e1 á t1l á T1cr , T e1 is the time of electron relaxation of the radical.Thus, in weak magnetic fields, the probability of transitions is high when radio- frequency fields of about 1072 mT are used. In radical reactions, the occupancies of electron ±nuclear sublevels of radicals are usually non-equilibrated due to CIDNP. It is evident that saturation of transitions would result in equal- isation of the occupancies and in a substantial change in the nuclear polarisation of the diamagnetic products. In weak mag- netic fields, a radio-frequency field acting parallel to an external magnetic field induces the n1>n3 transition (see Fig. 6), forbid- den in strong magnetic fields. The probability of this transition is proportional to the value (c1c2geB1).2 Theoretical description of the NMR ENR effect was made using the numerical solution of the Liouville equation for the spin density matrix of the radical with allowance for electron and nuclear relaxation.Theoretical analysis of the main features of this effect showed that its magnitude is determined by several factors, namely, by the amplitude of the applied radio-frequency field, the coefficients c1 and c2 , the lifetime of radicals in solution and by the occupancy of the corresponding electron ± nuclear levels (Fig. 7). It can be seen that at relatively small amplitudes of the radio-frequency field, the magnitude of this effect does not depend on the type of the transition excited, while at large amplitudes, it does depend on the type of transition. Thus, by comparing the polarisation intensities corresponding to different transitions in the same magnetic field, one can calculate the occupancies of all the electron ± nuclear spin levels of free radicals; analysis of the line shape provides information concerning the930 n2>n4625 525 n2?n3 n4?n3 500 Figure 7.Theoretical NMR ENR spectra. Initial occupancies of the levels n1=0, n2=1/2, n3=0, n4=1/2. The multiplet CIDNP arises in a 42 mT magnetic field, tl 55 T1,2 , A=70 mT, the amplitude of the applied radio-frequency field B1 (mT): (a) 3, (b) 0.001. relaxation times of radicals in a weak magnetic field and on their lifetimes. Ananchenko et al. 56 ± 58 were the first to carry out experiments on the study of the electron ± nuclear transitions in phosphonyl radicals resulting from photolysis of (2,4,6-trimethylbenzoyl)di- phenylphosphine oxide and dimethyl 2,4,6-trimethylbenzoyl- phosphonic acid.Figures 8 and 9 present the NMR ENR spectra recorded upon photolysis of 2,4,6-(trimethylbenzoyl)di- phenylphosphine oxide (B1 is parallel to B0 , Fig. 8) and dimethyl Polarisation (%) 0 7400 740 780 200 n2>n1 1000 7100 30 20 Figure 8. NMR ENR spectra of the diphenylphosphinyl radical Ph2(O)P . RF-field frequency n (MHz): (a) 431, (b) 750, (c) 1530; RF- field amplitude B1 (mT): (a, c) 1.1, (b) 1.2. Continuous lines are the simulated spectra constructed as the sum of the SNP contributions (chain line) and the theoretical NMR ENR spectra (dashed line). a n2?n1 n4?n1 b 1500 2500 f /MHz a n4>n3b n2>n3 c I 50 60 40 70 B0 /mT a Polarisation (%) 200 720 740 b 200 1000 c 200 150 100 500 750 25 15 Figure 9.NMR ENR spectra of the dimethylphosphonyl radical (MeO)2(O)P . RF-field frequency n (MHz): (a) 315, (b) 1530, (c) 694; RF-field amplitude B1 (mT): (a) 1.2, (b) 0.7, (c) 0.9. For the designation of lines, see Fig. 8. 2,4,6-trimethylbenzoylphosphonate in dioxane (B1 is perpendicu- lar to B0 , Fig. 9). Polarisation was measured in percent of CIDNP for B0=42 mT. It can be seen that the intensity of the n1>n2 and n1>n3 transitions exceeds the intensity of the n4>n3 transitions by a large factor. Based on analysis of the intensities and the signs of resonance lines in the spectrum, it was shown that the initial occupancy of the level n1 is negligibly small compared to that of the levels n2, n3 or n4 .Thus, it can be concluded that in weak magnetic fields, the magnitudes of which are lower than the HFC constants of radicals, intense electron polarisation should be observed. It is noteworthy that the results are in good agreement with the conclusions based on analysis of the field dependences of CIDNP for radical pairs with large HFC constants.59, 60 IV. Stimulated nuclear polarisation Stimulated nuclear polarisation is a phenomenon tightly related to other magnetic and spin effects in radical reactions, in particular, to the chemically induced dynamic nuclear polarisation. The CIDNP method does not bear direct quantitative information about the magnetic-resonance parameters of radical pairs; there- fore, it should be combined with other methods such as EPR, optical spectroscopy, etc.For example, the use of time-resolved CIDNP, first proposed and implemented by Closs et al.,61 provides quantitative information on the kinetics of the inter- mediate radical species, the lifetimes of triplet molecules, etc.62 ± 64 The SNP method is a combination of CIDNP with the RYDMR principles,1 ± 10, 65, 66 based on the selective influence of a variable field on the probability of recombination of separate sub-ensembles of radical pairs. The influence of an RF field on the yield of the reaction products is detected from the change of parameters of the system under study (for example, the resonance photoconductivity of organic semiconductors, the intensity of the recombination fluorescence, the optical absorption of the prod- ucts, the EPR spectra of stable radicals,67 etc.).We shall consider the mechanism of SNP in relation to a radical pair with one magnetic nucleus in strong magnetic fields. Let us assume that the pair has formed in the triplet state and E G Bagryanskaya, R Z Sagdeev n4>n3 n3>n1 n2>n1 35 45 B0 /mTDynamic and stimulated nuclear polarisation in photochemical radical reactions recombination from the triplet state is forbidden. In the case of strong magnetic fields, the S and T0 states are degenerate and, therefore, transitions from the non-reactive T0 state to the reactive S state occur efficiently due to hyperfine coupling. However, transitions from the T7 and T+ states to the S state are supressed due to the great energy gap between these states.The CIDNP effect is known to be based on the fact that the probabilities of the S ¡À T transitions are different for radical pairs with different orientations of the nuclear spin. Let the g-factors of the radicals be equal. In this case, the CIDNP effect for a radical pair with one magnetic nucleus is equal to zero. The resonance RF field mixes the T0 state of the radical pair with the T7 and T+ states, which accelerates the S ¡À T conversion at small amplitudes of theRFfield (o1<jDoj) and decelerates it at great amplitudes (o1>jDoj).[Here, o1 is the amplitude of the RF field, Do= oa7o0=o07ob is the difference between the resonance fre- quencies of the radicals R1 and R2 (partners of the R1 R2 radical pairs), R1 contains the magnetic nucleus (oa, ob) and R2 contains no nuclei with a non-zero HFC constant (o0).] We assume that o1<jDoj. If the resonance-field frequency is equal to the reso- nance frequency oa of electron transitions in the radical R1 with a-orientation of the nuclear spins, the acceleration of S ¡À T transitions would involve only the sub-ensemble of radical pairs with this spin orientation and the products of recombination of radical pairs would mainly contain nuclei with the a-projection of spins. This would result in enhanced absorption in the NMR line of in-cage products.The nuclei of the escape products of the recombination of the radical pair would be mainly characterised by the b-projection of spins and the corresponding line in the NMR spectrum would be an emission line (Fig. 10). If the frequency of the RF field is equal to the ob frequency of the electron transitions in the R1 radical with the b-orientation of nuclear spins, emission would be observed in theNMRlines of the in-cage products and absorption would be observed in the spectra of escape products. The effect of the RF field on the resonance frequency of the R2 radical (o0) will induce acceleration of the S ¡À T conversion with the same efficiency for both sub-ensembles of radical pairs and, hence, will not change the nuclear polar- a 3 R2 R1(H) Exit to the bulk RF field 1 R1(H), R2 R2 R1(H) Escape products In-cage products b o0 o0 o0 ob oa ob oa ob oa c Escape producs In-cage products Figure 10.Scheme for the generation of SNP for a radical pair with one magnetic nucleus (a) and its EPR (b) and SNP (c) spectra. 931 isation of the products. Thus, the positions of the resonance lines in the SNP spectrum correspond to the resonance frequencies in the EPR spectrum of the radical pair, and the signs of the resonance lines are determined by the signs of the HFC constants and by the pathways to diamagnetic products theNMR spectrum of which is used to detect the SNP. In the case where o1>jDoj, i.e. under spin-locking condi- tions, the RF field retards the S ¡À T conversion and, hence, the signs of SNP would be opposite to those observed in the case described above.On the basis of these considerations, Mikhailov et al.68 formulated the rule for determining the sign of polarisation of the products for the case of net SNP G=mecZAk , where G>0 corresponds to the positive (absorption) and G<0, to the negative SNP effect (emission), m=1or71 for a triplet or a singlet radical pair; e=1or71 for the in-cage or escape product; c=1 or 71 for low- or high-field component of the HFC spectrum of the chosen nucleus; Z=1 for o1<jDoj and Z=71 for o1>jDoj, Ak is the isotropic HFC constant at the k-th nucleus polarisation of which is considered. For radical pairs with any number of magnetic nuclei, the SNP spectrum is a superposition of the spectra of separate sub- ensembles of radical pairs with fixed configurations of nuclear spins.In a real situation when the spin dynamics of radical pairs are overlapped by the molecular motion of radicals, the spectral lines have a finite width and, in the general case, for several nuclei with different HFC constants, the SNP spectral pattern is mark- edly complicated. Therefore, to analyse the experimental SNP spectra, one should compare them with the theoretical spectra. 7>KS¡ÀT0, KS¡ÀT+); 0 , KS¡ÀT7). In the ¡¦ state is lower than those of the �¢ or T a0 states. Radio frequency pumps the T a¡¦ level and thus ¡¦?Sb 0>KS¡ÀT7 , KS¡ÀT+) (Fig. 11 b) and Exchange interaction markedly influences the pattern of the SNP spectra.We shall consider this influence qualitatively in relation to a biradical with one magnetic nucleus (Fig. 11 a,b). The relative positions of the energy levels are determined by the ratio of the HFC constants A, the effective exchange interaction J0 and the external maetic field B0 . The effects of nuclear polarisation, both chemically induced and stimulated, depend substantially on the ratio of the intersystem crossing rates (KS¡ÀT case where the average exchange interaction geB0<2j J0 j (Fig. 11 a), the rate of intersystem crossing via the S ¡À T7 channel is higher than that via other channels (KS¡ÀT therefore, the occupancy of the Ta T a increases the number of radical pairs recombining upon Ta transitions. As a consequence, the products of recombination become enriched in the nuclei with the b-projection of nuclear spins.The SNP spectrum of radical pairs is an emission line. Otherwise, geB0>2j J0 j(KS¡ÀT the occupancy of the Ta0 state is lower than that of the T a¡¦ or T a�¢ states. In this case, the number of radical pairs recombining via the S¡ÀT0 channel of intersystem crossing increases under the action of the RF field. The SNP spectral pattern is normalDthe signs of polarisation of the spectral lines depend on the signs of the projection of the nuclear spins, the exchange interaction being manifested as a change in the line splitting in the spectrum. Thus, a b T+ T+ ba ab geB0 geB0 T0 T0 ab abgeB0 2ab jSJ0 j geB0 2 j J0 j T7 T7 ba ba b S a Figure 11.Diagram of the electron levels of a biradical with one magnetic nucleus for strong (a) and weak (b) exchange interaction.932 the magnitude of exchange interaction can be judged both from the pattern of the S ±T0 or S ± T7 type spectrum and from the splitting in the S ± T0 type SNP spectrum. Apart from the exchange interaction, splitting in the SNP spectra is influenced by spin-selective decay of radical pairs, which can diminish splitting even in the absence of exchange interaction.69 It was shown 69, 70 that exchange interaction, electron relaxation and fast decay of the singlet state are factors influencing the splitting of lines in the SNP spectra because they all result in dephasing of the singlet ± triplet transitions.Theoretical description of the SNP effects has been the object of a number of studies 69 ± 81 in which various approaches were employed. Thus Mikhailov et al.68 developed a theory of SNP for reactions of neutral radicals with isotropic reactivity in strong magnetic fields based on the mathematical apparatus used tradi- tionally to study the geminate recombination of radicals with allowance for the S ± T0 transitions in radical pairs. The SNP effects were calculated in this study by summing the contributions to recombination of all the radical re-encounters.82 This method is based on the assumption that the spin dynamics in radical pairs and relative molecular motion of radicals in solution are com- pletely separate and implies the use of numerical calculations.However, numerical analysis of the SNP effects in systems with many magnetic nuclei is highly time-consuming. Analytical expressions have been obtained 72, 83 for the calculation of the probability of recombination of radical pairs with a specified configuration of nuclear spins and of the SNP effects on the assumption of low efficiency of singlet ± triplet transitions in radical pairs. The effects of various parameters on the line intensity and the pattern of the SNP spectrum have been studied theoretically.68 It was shown that the intensity and the sign of SNP as well as the spectral pattern depend substantially on many molecular-kinetic and magnetic-resonance parameters of the radical pair, solvent viscosity, reactivity of radicals, the occurrence of reactions of the radicals with acceptors, the amplitude and the frequency of a microwave field, the HFC constants and g-factors of radicals.Therefore, detailed interpretation of experimental data can only be performed resorting to calculations of the SNP effects. Osintsev et al.84 used the kinematic approach to derive the analytical expressions for calculating the SNP effects of radical pairs with one HFC constant in arbitrary magnetic fields. In another study,73 these reseachers analysed the line profile in the SNP spectra in weak magnetic fields, comparable with the HFC constant, for different RF-field amplitudes. It was shown that in weak fields, additional lines which are missing in strong fields appear in the spectrum.The effect of a powerful RF field during recording of the SNP spectra in weak magnetic fields gives rise to new resonance characteristics in the spin dynamics of radical pairs, which are multi-quantum resonances from the standpoint of quantum theory.74 Experimentally, this shows itself as the appearance of signals in the SNP spectra in the Zeeman fields B0 approximately divisible by the field of the main resonance.25 The kinematic approximation has also been used successfully 73 to design the theory of multi-quantum resonances under the conditions of SNP. Characteristic features of the SNP spectra for spin-correlated radical pairs in strong magnetic fields have been considered taking a two-position model as an example.81 It was found that narrow lines the nature of which is due to the degeneracy of two electron transition frequencies appear in the SNP spectra under definite conditions.From the width of these lines, the lifetime of spin- correlated radical pairs can be determined. Note that similar features are characteristic of the spectra of radical pairs measured using the optically detected electron paramagnetic resonance (OD EPR) technique.85 The SNP effects in systems with non-zero exchange interac- tion depending on the distance between the radical sites (short- lived biradicals,69 radical pairs in micelles 75 ± 78) can be described E G Bagryanskaya, R Z Sagdeev using an approach based on the numerical solution of the stochastic Liouville equation.1. Stimulated nuclear polarisation of radical pairs in homogeneous solutions The SNP effects were first observed experimentally for recombi- nation of the biradicals formed upon photolysis of cyclododeca- none.86 Subsequently, it has been shown in relation to the photolysis of benzoyl peroxide in various solvents 87 that recom- bination of radicals in the liquid phase proceeding via the formation of radical pairs consisting of the phenyl and phenacyl radicals [Ph OC(O)Ph] is accompanied by substantial polarisa- tion. Bagryanskaya et al.87 carried out experiments in weak magnetic fields (B0=*3.0 mT, n=94.5 MHz), the magnitudes of which are comparable with the HFC constants in the phenyl radical.The pattern of the SNP spectrum was complex, which substantially hampered interpretation of the data obtained, espe- cially for B1&A. Subsequently, the main regularities of SNP have been investigated in the photolysis of p-benzoquinone and dur- oquinone in CD3CN, CDCl3 and CD3OD in weak and strong magnetic fields.12, 20, 21, 34, 88 Photolysis of benzoquinone in CD3CN has been used as an example to study 86 the effect of spin locking, which was then employed to estimate the amplitude of the RF field when working in a high-frequency resonator. By comparing the dependences of SNP on the amplitude of the radio- frequency field in weak magnetic fields in the case of low pumping frequency (100 MHz, when measurements of the B1 value are rather simple) and in strong magnetic fields in the case of high pumping frequency (1.6 GHz), one can determine the amplitude B1 in the resonator for the high-frequency case.Multi-quantum resonances were detected experimentally in the high-field region of the SNP spectra measured at great RF field amplitudes (B1 44 A).25 As noted above, the method of SNP permits identification of intermediate short-lived species (radical and radical-ion pairs) and also provides information on the pathways of formation of the reaction products. For example, SNP has been used 89 to study elementary steps in the photolysis of anthraquinone (AQ) with triethylamine in solvents with different polarities.For the reaction carried out in polar solvents (CD3CN, CD3OD), the spectrum of triethylamine radical cation was measured using the NMR line of the triethylamine CH2 group. Based on the sign of the spectrum, it was ccluded that triethylamine is an in-cage product of recom- bination of the radical-ion pair {[AQ]7 +[Et3N]}. For the photolysis of triethylamine and anthraquinone in CD3OD, the SNP spectrum was measured based on the NMR signal of the proton of the hydroxy group in CD3OH. This spectrum differed from the SNP spectrum recorded using the signal of the triethyl- amine CH2 group only by the sign; hence, it corresponds to an escape product of transformation of the triethylamine radical cation. This fact indicates that the hydroxyl proton of the alcohol has come from the CH2 group of the triethylamine escape radical cation.Thus, the authors concluded that the reaction takes place between hydroxyalkyl and aminoalkyl radicals. In non-polar solvents, in particular, in benzene, the SNP spectrum corresponds to the EPR spectrum of the aminoalkyl radical MeC.HNEt. In medium-polarity solvents (cyclohexane and isopropanol), both pathways are effective (the radical-ion and radical pathways). It should be borne in mind that signals in the SNP spectra can differ substantially for continuous and pulse photolysis modes.28, 90 For example, the intensity of signals recorded during pulse laser photolysis of benzoquinone and benzaldehyde decreases by a factor of dozens when a mercury light is used as a continuous source of light.The signal intensity in the SNP spectra does not change if the radicals decay in a first-order reaction, for example, in the photolysis of benzoyl peroxide, cyclic ketones, etc. In the steady-state photolysis, the lifetime of radicals normally exceeds the time of nuclear relaxation of the intermediate radicals (1074 s), the observed polarisation being due to the contribution of the polarisation of the in-cage recombination products. In theDynamic and stimulated nuclear polarisation in photochemical radical reactions case of pulse generation of radicals, the opposite situation is observed, i.e. the lifetime of radicals is usually shorter than the nuclear relaxation time. Hence, the total polarisation of diamag- netic products formed both in the cage and as a result of radical recombination in the bulk markedly decreases or disappears. In a study of laser photolysis of p-benzoquinone in methanol by SNP, it has been found 25, 91 that cyclohexadienyl type radicals with methoxy- or hydroxymethyl groups participate in the reac- tion.The SNP spectra based on the signals present in the NMR spectra of benzoquinone, hydroquinone and the alcohol (the hydroxyl-proton signal) differ appreciably from the SNP spectra of radical pairs of semiquinoid radicals measured for continuous photolysis under the same conditions. The distance between the terminal lines in the spectrum (4.5 mT) indicates that the reaction involves radicals with the high HFC constants of the ring protons.Comparison of experimental SNP spectra with those calculated for several assumed radical species with allowance for HFC constants (calculated by the MNDO method) made it possible to establish the structure of the intermediate radical. Figure 12 a,b shows the NMR and SNP spectra recorded for the photolysis of 2-chloro-2-methylbutan-3-one in C6D12 .32 1) hn 2) C6D12 Me2CCOMe+C6D11 Me2C(Cl)COMe 7DCl 2 C6D11H+H2C CMeCOMe 1 Me2C(C6D11)COMe 3 MeOCCMe2CMe2COMe 4 Stimulated nuclear polarisation is observed for all diamagnetic products 1 ± 4. The experimental spectra of in-cage products are in good agreement with the spectra calculated for the [Me2(CO- Me)C C6D11] radical pair. To attain agreement between the theoretical and experimental SNP spectra of escape products, a 1 2 3 2 1 4 5 4 3 2 ppm b 3 1 4 2 55 55 50 45 mT mT 50 45 Figure 12.NMR spectra recorded for the photolysis of 2-methyl-2- chlorobutan-3-one in C6D12 (a) [(1) spectrum in the dark, (2) CIDNP spectrum] and experimental and simulated SNP spectra for reaction products (b); (1) product 1, (2) product 2, (3) product 4, (4) product 3. The continuous lines are simulated spectra. (1, 2) with allowance for the contributions to SNP of the escape products of the geminate radical pair [Me2(COMe)C C6D11] and in-cage products of the diffusion radical pair [Me2(COMe)C C(COMe)Me2]; (3, 4) calculations of the SNP of the in- cage products from the geminate radical pair [Me2(COMe)C C6D11].contributions of both radical pairs, [Me2(COMe)C C6D11] and [Me2(COMe)C C(COMe)Me2], should be taken into account. The influence of the DEE rate on the pattern of SNP spectra has been studied.48 It was shown that, since DEE results in frequency exchange between the components of the spectrum with opposite polarisation signs, the narrowing of lines is accom- panied by a decrease in their intensities; in the limit, the lines disappear. The influence of DEE on the SNP effect has been studied in detail both experimentally and theoretically.92 The researchers used two approaches to the theoretical description of the SNP spectra with allowance for DEE. In the former case, the spectra were calculated using the theory of uncorrelated frequency migration for a radical pair with two magnetic nuclei.In the latter case, for a weak RF field (B155A), a two-position model was proposed. In the case of low DEE rates {Ate 441, where te is the time of DEE, te=1/ke[A] (or [D]), [A] or [D] is the concentration of the acceptor or the donor participating in the DEE reaction}, this model provides an analytical expression describing the variation of the SNP amplitude vs. the exchange time te . Photo- induced isomerisation of fumaronitrile in the presence of deuter- ated naphthalene in acetonitrile was chosen as the model reaction for experimental investigations. Figure 13 shows the SNP spectra of radical pairs resulting from photolysis of fumaronitrile with naphthalene over broad ranges of DEE rates and RF-field amplitudes.At low fumaronitrile concentrations when the influ- ence of DEE is negligibly small, the SNP spectra resemble the spectra based on calculations for neutral radical pairs without SNP (rel.u.) 10 7110 7110 71 52 54 53 Figure 13. SNP spectra recorded for the photolysis of naphthalene-d8 (261073 mol litre71) with fumaronitrile and detected from the NMR signal of the fumaronitrile protons. Fumaronitrile concentration (mol litre71): (1) 1073, (2) 1072, (3) 561072 . RF-field amplitudes (mT): (a) 0.25, (b) 0.67, (c) 1.2. The dashed lines correspond to the spectra calculated with allowance for all the HFC constants using the theory of SNP for neutral radical pairs without DEE.The continuous lines are the spectra calculated using the theory of uncorrelated frequency migration for a radical pair with two magnetic nuclei. 933 a 123 bc 55 56 57 B0 /mT934 allowance for charge exchange. The calculation was carried out taking into account all the HFC constants in radical pairs in terms of the theory of SNP in strong magnetic fields. It can be seen from Fig. 13 that the pattern of the SNP spectra is extremely sensitive to the RF-field amplitude. As the concentration of fumaronitrile increases, SNP sharply decreases; line splitting in the spectra also decreases. To measure the rate constant for DEE in radical-ion reactions by comparison of the theoretical and experimental SNP spectral patterns, the dependence of the SNP intensity on the exchange time te at low RF-field amplitudes in the absence of fast trapping of radical ions was used.The SNP method was utilised to study the photoinduced electron transfer from quadricyclanes to tetrafluorobenzoqui- none. It was shown 93 that the radical-ion pair in this reaction incorporates the quadricyclane radical cation. Isomerisation of quadricyclane into norbornadiene takes place in the free radical ion rather than in the step of formation of the radical-ion pair. The majority of researchers studied reactions in homogeneous solutions by 1H NMR spectroscopy. However, detection of SNP using other nuclei (13C, 31P, 19F, etc.) appears quite promising. Figure 14 presents the 31P SNP spectrum of the radical pair formed in the photolysis of (2,4,6-trimethylbenzoyl)diphenyl- phosphine oxide.Splitting in the spectrum corresponds to the HFC constant at the phosphorus nucleus. A specific feature of SNP at the 31P nuclei is large HFC constants comparable with the magnitude of the external magnetic field. Therefore, calculations cannot be carried out using the strong magnetic field approxima- tion. In addition, for radical pairs with large HFC constants, contributions of S ± T7 transitions occurring in the area of intersection of terms should be taken into account.59, 60 The potential and advantages of the time-resolved SNP method were verified experimentally 28 taking photolysis of ben- zoyl peroxide (inCD3ODorCD3CN) as an example.It was shown that the delay (t) between the laser pulse and the RF pulse inverts the observed SNP spectrum.Astudy of the dependence of SNP on t in different magnetic fields for different resonance lines demon- strated that the inversion of the spectrum takes place at t >50 ns, and subsequently the spectral pattern changes only slightly. Apparently, the spectrum corresponds to the [Ph Ph] and [Ph CD2OD] pairs. Osintsev et al.84 have considered theoretically the problem of the line shape in the SNP spectra recorded in time-resolved experiments and the dependence of SNP on the character of motion of the partners of radical pairs. From the theoretical standpoint, consideration of these processes is reduced to the Polarisation (%) 60 123 40 200 720 740 760 780 50 40 20 60 30 B0 /mT Figure 14.31P SNP spectra of radical pairs formed in the photolysis of (2,4,6-trimethylbenzoyl)diphenylphosphine oxide. The spectra were detected using the NMR signal of diphenylphosphine oxide. Polarisation was measured in percent of the CIDNP for B0=43 mT. B1 (mT): (1) 1, (2) 0.6, (3) 0.4. The continuous lines are the spectra simulated using the analytical formula for weak magnetic fields.73 E G Bagryanskaya, R Z Sagdeev calculation of the probability of recombination of geminate radical pairs. (The theory of recombination of radical pairs with a Hamiltonian changing abruptly with time has been formu- lated.94, 95) It was shown 84 that the width of the SNP line at small amplitudes of the RF field is determined by the amplitude of the RF field, while the intensity of the SNP spectral line depends on this amplitude according to a law describing diffusion motion.The use of a delay between the laser and RF pulses should not result in a substantial narrowing of SNP lines. This is due to the fact that the re-encounter of partners in radical pairs which takes place after a prolonged diffusion walk contributes significantly to the SNP. As long as the delay t differs from zero, this contribution remains crucial. The variation of the intensity of the SNP line in the magnetic field corresponding to the resonance line maximum vs. the delay t appears the most informative. For great delays, the dependence typical of diffusion motion should be observed.The SNP kinetics calculated using the two-position model is described by an exponential curve with a characteristic exponent parameter equal to the constant for the irreversible decay of radical pair tc . The expected line widths in the SNP spectrum should not depend on the amplitude of the RF field and are determined by the reciprocal time 1/tc . Time-resolved SNP experiments for geminate radical pairs in homogeneous solutions are currently impossible because the characteristic lifetime of radical pairs is about several nano- seconds and, to obtain real kinetics, it would be necessary to initiate the reaction by lasers with nanosecond pulse duration. In addition, it would be necessary to create a rise time of about 1 ns for the RF pulse, which is a difficult engineering problem. However, the lifetimes of short-lived biradicals and micellised radical pairs lie in the range from tens to hundreds of nano- seconds; the kinetics of these systems can be studied experimen- tally.2. Stimulated nuclear polarisation of short-lived biradicals Biradicals are formed as intermediates in many photochemical processes; therefore, it is important to investigate them in order to elucidate reaction mechanisms. As a rule, the lifetimes of flexible biradicals are about 0.1 ± 1 ms. They are determined by the rates of intersystem crossing from the non-reactive triplet state to the singlet state, recombination from the singlet state and the con- formational motion of radical centres.Photolysis of aliphatic cyclic ketones with rupture of the a-bond is a perfect method for the investigation of short-lived biradicals. The a-bond rupture occurs from the excited triplet state almost in all cases. The introduction of methyl or phenyl groups in the a-position decreases the activation energy and increases the rate of a-rupture to 109 s71 or more. The acylalkyl biradical thus formed can be decarbonylated giving rise to alkyl biradicals. The rate of decar- bonylation increases substantially upon a-substitution. For bi- radicals formed in the photolysis of aliphatic cyclic ketones, the rate of molecular motion in non-viscous solutions at room temperature is much higher than the rate of spin dynamics. In the absence of steric hindrance, recombination from the singlet state is not the rate-determining step either; thus, the lifetime of biradicals is determined by the rate of the triplet ± singlet con- version, which depends on the hyperfine electron-exchange and spin-orbital interactions and on the electron spin relaxation. The main features of the generation of SNP in biradicals have been studied experimentally in relation to the photolysis of several aliphatic cyclic ketones O=C(CH2)n (n=10 ± 13).86, 96 ± 99 As noted above, the major pathway of the photolysis of cycloalka- nones is a-rupture in the triplet-excited ketone, resulting in the formation of an acylalkyl biradical.Variation of the length of the polymethylene chain of the ketone, the HFC constants (by using ketones in which the protons in the a-position have been replaced by D or CH3) and the external magnetic field, as well as detection of SNP from 1Hand 13CNMRspectra make it possible to vary the ratio of A, J and B0 over a broad range.Dynamic and stimulated nuclear polarisation in photochemical radical reactions The 13C SNP spectra of acylalkyl biradicals with n=11, in which the 13C isotope occurs in the carbonyl position or in the a-position, differ significantly.This is due to the difference between the HFC constants and, correspondingly, different ratios of the external magnetic field to the HFC constants, resulting in dissimilar efficiencies of the competing intersystem crossing channels.102 As shown by calculations performed in terms of the two-position model in weak and zero magnetic fields, the pattern of the SNP spectra of short-lived biradicals is exceptionally sensitive to the rate of spin-non-selective intersystem crossing; this can be used to measure their rates.For biradicals with n=10 ± 13 in weak magnetic fields (3 ± 12 mT), the S ± T7 channel of intersystem crossing predom- inates. Both emission CIDNP and emission SNP can be observed in the NMR signals of all groups of protons of ketones and aldehydes. Polarisation on the ketone g-protons, for which the HFC constant in the biradical is close to zero, arises due to the transfer of the nuclear polarisation from the a-CH2 groups of the ketone, induced by the transfer of the sample from a weak to a strong magnetic field. For ketones with n=11 ± 13 in a *50 mT field, the main intersystem crossing channel is S ± T0; no signals for the ketone protons are observed in the SNP spectra in this field because the HFC constants of a- and b-protons in biradicals have different signs, and the lines for opposite signs overlap. When the chain length decreases, the effective value of exchange interaction increases; an emission SNP spectrum is obtained in a 50 mT field for biradicals formed in the photolysis of cyclodecanone (n=10).Koptyug et al.96 reported calculations of SNP of short-lived biradicals performed for a realistic model, which had been used previously 100 to describe the field dependences of CIDNP. Presumably, exchange interaction occurs through the spatial overlap of the orbitals of unpaired electrons and is attenuated according to an exponential law as the distance r between the electrons increases J(r)=J0 exp(7ar) , The main regularities of the origination of SNP in the consecutively formed biradicals were studied experimentally and theoretically 103 in relation to the acylalkyl and bisalkyl biradicals formed in the photolysis of tetramethyl-substituted cycloalka- nones with n=11 ± 15.A specific feature of the consecutive biradicals, as well as of the consecutive radical pairs, is the `memory effect', which implies that nuclear polarisation of the diamagnetic products of the secondary radical pairs or biradicals reflects the spin dynamics processes in both the primary and the secondary pair (biradical). Due to the high spectral selectivity, typical of the NMR method, when recording the NMR spectra of various recombination products, the authors were able to obtain simultaneously the SNP spectra of primary and secondary bi- radicals arising during photolysis.A model for the calculation of SNP for the consecutive biradicals was developed; the generation of SNP was studied both experimentally and theoretically. The main characteristics influencing the line intensity and the pattern of the SNP spectra of secondary biradicals are the rate constant for the electron relaxation of both biradicals, the time required for relaxation processes, the rate constant for the spin-orbital inter- action, parameters of the exchange interaction of both primary and secondary biradicals, the rate constant for decarbonylation, the type of the main intersystem crossing channel in the biradicals, and the ratio of the polarisation intensities for the first and the second biradicals.Note that for the symmetrical biradicals, in which the HFC constants for various radical centres are equal, the SNP spectral pattern differs appreciably from the case considered above. This is due to the fact that, in symmetrical biradicals or radical pairs in strong magnetic fields, the SNP effect is equal to zero. This fact can be easily understood from a qualitative consideration. Saturation of the electron transitions in radical pairs with identical orientations of nuclear spins does not accel- erate the intersystem crossing. In the radical pairs with the opposite orientation of nuclear spins, the S ± T conversion is accelerated for both sub-ensembles of radical pairs.Thus, the overall polarisation induced by a microwave field would be equal to zero. When calculating the SNP in symmetrical biradicals in weak magnetic fields, it is necessary to take into account both HFC constants in the secular approximation. If one HFC con- stant is taken into account in the secular approximation, while the second, is in a non-secular approximation, a non-symmetrical radical pair is actually considered and an S ±T0 type spectrum appears in the calculated SNP spectra. Unlike biradicals with different HFC constants, for which the spectral pattern is deter- mined by the major channel of the singlet ± triplet conversion, for symmetrical biradicals, the SNP spectra would always correspond to the S ± T7 type even in those cases where S ± T0 is the major channel of intersystem crossing.the interaction anisotropy being averaged due to the fast rotations of the terminal groups of the polymethylene chain. Two mecha- nisms of electron spin relaxation were taken into account, one due to fluctuation of local magnetic fields and one, due to dipole ± dipole interaction of the spins of the unpaired electrons. The pattern of the calculated SNP spectra is highly sensitive to the exchange interaction constant J0 . Using experimental and theo- retical spectra obtained for different radio frequencies, the opti- mal values for J0 and a were found.Koptyg et al.96 studied theoretically the influence of several parameters, namely, the diffusion coefficient, the rate of trapping of biradicals, the rate of decay of the triplet biradical due to the spin-orbital interaction, and the electron spin relaxation, on the pattern of the SNP spectra. It is significant that all parameters, except the diffusion coeffi- cient, have little influence on the shape of spectral lines but can only cause a decrease in the signal intensity. As noted above, the dependence of the signal intensity in the SNP spectrum on the delay t bears information on the lifetimes of radical pairs. For instance, this dependence has been used 96 to estimate the lifetimes of short-lived biradicals formed upon photolysis of cyclododeca- none.The characteristic times of SNP decay were*80 ns. Due to the lack of a theory of time-resolved SNP for biradicals, the researchers could not use the experimental data to perform a profound analysis or to gain new information. The kinetics of SNP of short-lived acylalkyl biradicals in homogeneous solutions in weak magnetic fields (*10 mT) have been studied.27, 101 Two types of radio-frequency pulses were used, a long pulse (*2 ms) and a short pulse with a duration at half- height of 15 ns. It was shown that the kinetic dependences of SNP pass through a maximum when the delay between the laser pulse and the RF pulse is 40 ns and then decay by an exponential law. The appearance of the maximum is due to the competition between spin-dependent and spin-independent channels of the triplet ± singlet conversion. The kinetics of SNP were described using a simple kinetic model.From comparison of the experimen- tal data with the calculated data, the researchers estimated the rate of spin-orbital interaction, the electron relaxation rate and the lifetimes of the excited triplet molecules of ketones. Upon the addition of a quencher for triplet states (cis-piperylene), the time for attenuation of the SNP signal diminishes to reach a limiting value equal to the lifetime of the biradical. Proceeding from these results, the lifetimes of biradicals (80 ns for cyclododecanone and 50 ns for cycloundecanone) and of triplet molecules (250 ns for cycloundecanone and 150 ns for cyclododecanone) were esti- mated.Study of the SNP of acylalkyl and bisalkyl biradicals in different magnetic fields (Fig. 15) allowed the authors to deter- mine the J0 value, which describes adequately all the experimental data. In the photolysis of ketones with n=11, on the basis of the NMR signals of the products resulting from the bisalkyl birad- icals, the S ± T7 type spectra represented by one emission line (B0=54.6 mT) were observed (Fig. 15 c). For n=12, inverted S±T0 type spectra were recorded (Fig. 15 d ). The lower limit of the J0 range was established by comparing the calculated and exper- imental SNP spectra of biradicals with n=11 (S±T7 type), while 935936 a 52 57 b 57 52 B0 /mT Figure 15.Experimental (dots) and calculated (continuous lines) SNP of acylalkyl [OC CMe2(CH2)n72C Me2 , n=11 (a), 12 (b)] and bisalkyl [MeC (CH2)n72C Me2 , n=11 (c), 12 (d )] biradicals formed in the photolysis of tetramethyl-substituted cycloalkanones. the upper limit was found by comparing the spectra of biradicals with n=12 (S¡ÀT0 type) in a 27 mT field. In the literature, the short lifetime of the acylalkyl biradicals is attributed most often to the presence of spin-orbital interaction between the radical centres of the biradical (see, for example, Ref. 104). This statement is true for biradicals with relatively short chains (n=5, 6). Evidently, the influence of spin-orbital inter- action on the lifetime of the biradical depends on the residence time of radical centres in the reaction volume.For biradicals with long chains, the residence time of the radical centres in the reaction area is short; therefore, the role of relaxation occurring due to spin-orbital interaction should markedly decrease. On the basis of analysis of the SNP experimental kinetics for biradicals with n=12 ¡À 15, 17, it was concluded that in addition to the spin- orbital interaction, spin-rotational relaxation also plays an impor- tant role.103 The influence of temperature and viscosity on the spin and molecular dynamics of short-lived alkylacyl and bisalkyl birad- icals formed successively has been studied both experimentally and theoretically.105 The 1Hand 13CCIDNP and SNP effects were studied in the temperature range from 200 to 360 K for magnetic fields varying from 0 to 0.1 T.The investigators found a unified set of parameters that describe adequately the nuclear polarisation effects for both protons and 13C atoms with different HFC constants in the biradical. The introduction of a temperature dependence of the biradical reactivity proved especially important for attaining quantitative agreement between the calculated and experimental data in the low-temperature range. Comparison of the experimental data with the results of model calculations showed that broadening of the field dependence of CIDNP and slowing down of the decay of SNP signals at temperatures below 240K is due both to the retardation of the molecular mobility of the polymethylene chain and to the decrease in the reactivity of the acylalkyl biradicals.For the bisalkyl biradicals, spin-orbital interaction is less important for the spin dynamics than in the acylalkyl radicals. The effect of molecular dynamics on the field dependences of the CIDNP of the products formed from bisalkyl biradicals with the change in temperature is more pronounced. Since the lifetime of bisalkyl biradicals is much longer than the lifetime of acylalkyl biradicals, the intensity of nuclear polar- isation of the products formed from bisalkyl biradicals is much higher. c 55 45 65 d 57 52 B0 /mT E G Bagryanskaya, R Z Sagdeev 3. Stimulated nuclear polarisation of micellised radical pairs Micellar solutions as a specific microstructured medium have been drawing attention of photochemists during the last decades.Numerous publications 106 ¡À 115 are devoted to the detection and investigation of extremal magnetic effects and magnetic isotope effects, nuclear and electron polarisation and other effects in radical pairs located in micelles. In micelles, radical pairs are isolated; this enables detailed investigation of the geminate recombination of radicals and spin and molecular dynamics of radical pairs. Since in radical pairs occurring in micelles, the partners are located at short distances from each other throughout their whole lifetime, the electron exchange interaction has a considerable influence on the intersystem crossing in these radical pairs.The field dependences of 13C CIDNP of radical pairs formed upon photolysis of dibenzyl ketone in sulfate micelles of different sizes have been studied.108 The polarisation of the 13C nuclei in both strong and weak fields (irrespective of the sign of the HFC constant) was found to be emission polarisation. The `maximum' of the field dependence of CIDNP shifts upfield as the size of micelles decreases. On the basis of experimental data, it was concluded that CIDNP is mainly due to S ¡À T7 transitions. The effective exchange interaction estimated from the field dependence of CIDNP was *30 mT. However, the results obtained by CIDEP for a number of systems are at variance with the conclusion concerning large values for average exchange interactions.In the EPR spectra of micellised radical pairs, each hyperfine-structure line is split into two. This splitting can be described by including a constant exchange interaction j J j&5 ¡À 10 MHz into the Hamiltonian in model calculations.115 Similar estimates were obtained in RYDMR experiments in micelles.114 Later, it was shown that modulation of the exchange interaction due to the mutual motion of radicals should be taken into account in the description of the spectral patterns and the CIDEP kinetics of spin-correlated pairs in micelles.116 The use of the SNP method for the investigation of the effect of exchange interaction on the intersystem crossing in radical pairs confined in micelles proved quite effective.75 ¡À 78, 117 ¡À 123 Stimulated polarisa- tion is detected from the NMR signals of the diamagnetic products of recombination of radical pairs; therefore, it reflects the influence of the exchange interaction on the intersystem crossing in radical pairs localised in micelles to a greater extent than CIDEP, in which the predominant contribution to the spectral signal is made by free radicals.The 13C and 31P SNP techniques were used to study the radical pairs resulting from photolysis of a-methyldeoxyben- zoin,77, 119, 122, 123 dibenzyl ketone,76, 78, 117, 118, 122 benzoin,78 de- oxybenzoin,124 benzophenone in the presence of tert-butylphe- nol 121 and (2,4,6-trimethylbenzoyl)diphenylphosphine oxide 120 in the sulfate micelles Na+SO¡¦3 O(CH2)nMe of various sizes (n=6 ¡À 11).The significant differences in the ratios of the HFC constants and the lifetimes of these radical pairs permitted the researchers to examine the influence of the exchange interaction on the pattern of the SNP spectra. It was discovered that line splitting and line width in the spectra depend on the size of micelles. From comparison of the spectra recorded in strong and weak magnetic fields and the field dependences of CIDNP with the spectra calculated theoretically, parameters of the molecular and spin dynamics of micellised radical pairs were found. The dynamics of radical pairs in micelles were described using a microreactor model D one radical was fixed in the centre of a spherical micelle, whereas the other was moving inside the micelle core.75 ¡À 78 In addition, a diffusional character of the motion was assumed.The electron spin relaxation due to the dipole ¡À dipole interaction and due to modulation of the anisotropic part of HFC was taken into account.77 The allowance for the dipole ¡À dipole relaxation was made by two methods. One approach is similar to that described previously 69 for the calculation of the field depend- ences of CIDNP and SNP in short-lived biradicals. The other approach is more consistent; it was used (in combination with the Monte Carlo method) 125 to calculate the correlation function ofDynamic and stimulated nuclear polarisation in photochemical radical reactions the dipole ¡À dipole interactions in radical pairs for micelles of different sizes.Subsequently, an analytical expression describing the dipole ¡À dipole relaxation was obtained.126 The main param- eters determining the molecular dynamics of radicals are the coefficient of mutual diffusionD, the micelle radius L, the reaction radius and the rate constant for the radical exit to the bulk kex . Splitting in the SNP spectra depends on the ratio of the following values: the HFC constants, the characteristic time between the collisions of radicals inside the micelle t2 �� L3 ¡¦ R3 * L3 3RD 3RD , the exchange interaction and the rate of the spin-selective decay. Upon a decrease in the micelle size, the time the radical pair spends in the reactive state increases relative to the time spent in the non- reactive state.Thus, the time-average exchange interaction and the time-average recombination rate increase. The experimental implementation of the SNP method for the investigation of radical pairs in micelles has several specific features. On the one hand, due to the higher viscosity of micellar media compared to homogeneous solutions, the nuclear relaxa- tion time of diamagnetic products decreases, resulting in a less intense signal in the SNP spectra. On the other hand, the lifetime of radical pairs in micelles (*0.1 ¡À 1 ms) is commensurable with the characteristic time of RF transitions (*10 ¡À 100 ns), which is calculated from the expression o1=geB1 (the B1 value lies in the range 0.1 ¡À 1 mT).This results in a pronounced increase in the intensity of spectral lines and provides the possibility of measuring 13C SNP spectra for samples with natural abundance of 13C. An advantage of the 13CSNP method is that the spectra of radical pairs containing 13C magnetic nuclei are recorded on the basis of the corresponding lines in the 13C NMR spectrum; therefore, the splitting in the SNP spectrum is equal to the HFC constant for this nucleus in the radical. In small micelles (n=7¡À9) for A=2 ¡À 10 mT, the dece in the spectral splitting is due to both the exchange interaction J and the rate of spin-selective decay of the radical pair. In larger micelles (n=11, 12) for A>2.0 mT, splitting in the spectrum is equal to the HFC constant, exchange interaction influencing only the widths of individual lines.The 13C SNP effects involved in the photolysis of a-methyl- deoxybenzoin enriched to 96% in the 13C isotope at the carbonyl position have been studied in micelles of different sizes.51, 77 The splitting in the SNP spectrum for sodium dodecyl sulfate micelles amounts to 12.4 mT; it is equal to the HFC constant for the carbonyl carbon atom in the benzoyl radical (Fig. 16 a). A decrease in the length of the polymethylene chain of the detergent molecule results in a shorter spacing between the lines in the SNP spectrum and in a changed width of individual lines. From calculations of the SNP spectra and field dependences of CIDNP, the optimal parameters of the electron exchange inter- action and spin-selective decay were found (kst): J0=50 mT (the parameter for the decay of exchange interaction a is 2A) and kst=8.It is noteworthy that the high sensitivity of the SNP spectra to the change in the rate of re-encounters of radicals in the micelles and, hence, to the change in the intra-micellar viscosity and micelle size can be used to investigate the dependence of the properties of micelles on the temperature and on salt additives (see Fig. 16).123 Parnachev et al.121 studied experimentally the SNP of radical pairs formed in the photolysis of benzophenone-d10 in sodium dodecyl sulfate micelles in the presence of 2,4,6-tri-tert-butylphe- nol with the natural abundance of the 13C isotope. On exposure to a 0.03 mT radio-frequency field, the NMR spectrum exhibits signals from six radical pairs differing in the position of 13C (denoted by the Roman numerals I ¡À VI) (Fig.17 a). 937 a 123 b 60 55 50 45 B0 /mT Figure 16. Temperature dependence of the SNP spectra of the radical pairs formed in the photolysis of a-methyldeoxybenzoin in sodium dodecyl sulfate (a) and sodium octyl sulfate (b) micelles. The spectra were detected using the NMR signal of the carbonyl carbon atom of a-methyldeoxybenzoin. The continuous lines are the calculated SNP spectra. T (8C): (1) 12, (2) 29, (3) 45 (a) and (1) 10, (2) 17, (3) 45 (b). VI But D II III VI But HO C I O IV V VI III II But VI D The SNP spectra were obtained for all the six radical pairs with different positions of 13C.Figure 17 b shows the SNP spectra for pairs with the 13C isotope in positions I and II. The opposite phases of these spectra are due to the different signs of the HFC constants for the corresponding 13C nuclei. Measurement of the SNP kinetics makes it possible to determine the lifetimes of radical pairs and to draw conclusions about the mechanisms of their relaxation. The value of the contact electron exchange interaction estimated from comparison of the experimental and theoretically calculated spectra is j J0 j>40 mT. The HFC constants in radical pairs with different positions of the 13C isotope were calculated from analysis of the SNP spectra. In the limiting case, where the characteristic time of intersys- tem crossing markedly exceeds the characteristic time period between the re-encounters in the micelle (At2441), splitting of lines in the SNP spectra does not depend on the micelle size, and the effect of the exchange interaction reduces to line broaden- ing.120 For example, this is the case in the radical pairs formed in the photolysis of (2,4,6-trimethylbenzoyl)diphenylphosphine oxide.In accordance with theoretical predictions, the experimen- tal SNP spectra of radical pairs in sodium octyl sulfate micelles and in sodium dodecyl sulfate micelles show the same splitting; however, agreement between experimental and theoretical spectra is attained only when sodium dodecyl sulfate micelles with a radius of 17 ¡À 21Aor sodium octyl sulfate micelles with a radius of 13 ¡À 15A are used.Thus, it was concluded that the effective radius of micelles varies depending on the compounds involved in the reaction. As noted above, in measuring the EPR spectra, the observed polarisation contains contributions of both radical pairs in micelles and free radicals, whereas SNP arises only in the presence of radical pairs. For example, using SNP and time-resolved EPR, the spectra of the radical pairs formed in the photolysis of a-methyldeoxybenzoin and benzoin in micelles of various size938 a III VI V I IV II ppm 60 100 180 140 b SNP (rel.u.) 1.0 0.50 55 54 56 B0 /mT 70.5 71.0 c SNP (rel.u.) 1.0 0.8 123 0.6 0.4 0.2 0 2000 1000 3000 t /1079 s Figure 17.NMR spectrum of radical pairs formed in the photolysis of benzophenone-d10 with 2,4,6-tri-tert-butylphenol in the micelles on expo- sure to a 0.03 mT RF field (a); SNP spectra of radical pairs labelled at position I (1) and II (2) (b) and the kinetic curves of SNP of the radical pairs labelled at positions I ± III (1 ± 3, respectively) (c) in different magnetic fields. B0 (mT): (1) 53.5; (2), (3) 54.1. The calculated relaxation constants krel 1.26106 (dashed line) and 3.06105 s71 (continuous line). were recorded. The SNP spectra measured for micelles with different sizes are closely similar because the HFC constants, electron relaxation parameters, radical size, and the reactivity of radical pairs are very close. Meanwhile, the EPR spectra recorded for the photolysis of a-methyldeoxybenzoin and benzoin are different in principle.Whereas the RP formed in a-methyldeoxy- benzoin is responsible for an antiphase structure of the spectrum depending on the micelle size, in the case of benzoin, no line splitting due to exchange interaction is observed.116 The radical pairs consisting of hydroxybenzyl radicals formed in benzoin photolysis leave the micelles at a markedly higher rate than the radical pairs consisting of benzyl and benzoyl radicals. As shown by Tarasov et al,116 within 200 ns after the photolysis of a-meth- yldeoxybenzoin, 70% of the radicals are located in micelles, while in the case of benzoin, this value is only 5%. The kinetics of micellised radical pairs have been studied by time-resolved SNP.It is obvious that the coherent transitions between singlet and triplet sublevels in radical pairs cannot be described within the framework of a kinetic scheme.124 However, since for most radical pairs, the rate of recombination markedly exceeds the relaxation rate, the experimental kinetic dependences of SNP are well described by an exponential curve and are consistent with the calculation based on the simple kinetic scheme. E G Bagryanskaya, R Z Sagdeev Yet another approach to the theoretical description of the dynamics of radical pairs is numerical solution of the Liouville equation for the density matrix.78 The kinetic curves for time- resolved SNP measured for the photolysis of a-methyldeoxyben- zoin and dibenzyl ketone in sodium dodecyl sulfate micelles are described adequately by an exponential function in accordance with the kinetic scheme.Figure 17 c shows the kinetic curves of SNP obtained for the photolysis of benzophenone-d10 with 2,4,6-tri-tert-butylphenol in sodium dodecyl sulfate micelles. The SNP kinetic curves for radical pairs labelled in positions I ±V bear information on the relaxation in various radical pairs. The decay of SNP for the pairs labelled in positions II ±V is due to the exit of radicals from the micelle and dipole ± dipole relaxation. For a pair labelled in position I, relaxation caused by the modulation of the anisotropic part of HFC plays an important role. In the photolysis of (2,4,6-trimethylbenzoyl)diphenylphos- phine oxide, the rate of the decay of SNP kinetics for sodium dodecyl sulfate micelles is 2.5 times higher than that for sodium octyl sulfate; meanwhile, in the case of photolysis of ketones, this difference was not more than 20% ± 30%.Analysis of the resulting time dependences in terms of the kinetic model made it possible to estimate the relaxation and recombination rates for the radical pair in this reaction. For sodium octyl sulfate micelles, the relaxation rate is only 1.8 times lower than that for sodium dodecyl sulfate micelles. Thus, the SNP kinetics observed in sodium octyl sulfate micelles is determined by both recombination and relaxation. Note that the parameters determining the kinetics of radical pairs and those determined from theoretical and experimental SNP spectra are in good agreement with the data from the time-resolved EPR spectra.127, 128 The use of the SNP methods in investigating micellised radical pairs provides unique information on their spin dynamics and kinetics.This possibility is due to the high sensitivity of the SNP method and to its selectivity with respect to radical pairs. The information obtained in SNP studies of micellised radical pairs was used in experiments on the measurement of the magnetic isotope effect (MIE) induced by a resonance RF field. The reaction investigated in these experiments was photolysis of dibenzyl ketone.129 Saturation of the electron transitions in radical pairs can give rise to enhanced MIE in two different regimes: at small amplitudes of the RF field (B155A), the S ± T conversion in one sub-ensemble of radical pairs containing a magnetic isotope is accelerated, while at large amplitudes of the RF field (B144A), closing of the S ± T conversion for radicals devoid of a magnetic isotope takes place.Tarasov et al.129 used the former regime but the latter appears to be more promising as regards enhancement of MIE.130, 131 The SNP method can provide exact information on the RF-field amplitude that ensures the maximum MIE value. V. Chemically induced dynamic nuclear polarisation in switched external magnetic fields The method of time-resolved CIDNP permits one to obtain quantitative information about the kinetics of radical reactions, lifetimes of triplet molecules and electron ± nuclear cross-relaxa- tion processes.132 ± 135 The procedure involved in the time-resolved CIDNP method is the measurement of the dependence of CIDNP intensity on the variable delay t between the laser pulse and the detecting radio-frequency pulse (90-degree pulse of an NMR spectrometer) (Fig.18 a). CIDNP arises in the strong magnetic field of the NMR spectrometer; as a rule, this is not the optimal field for observing this effect when many photochemical reactions take place. The time resolution of the method is determined by the duration of the radio-frequency pulse and is equal to 40 ns at best (normally, it is of the order of microseconds). It is often necessary to study CIDNP in weak and medium magnetic fields 52 because this method can provide information about the exchange inter- action, exchange reactions and so on.When measuring CIDNP inDynamic and stimulated nuclear polarisation in photochemical radical reactions B a 1 2 0 b B 1 B0 + Bs t1 0 B0 c B B0 + Bs 1 B00 t2 t1 Figure 18. Schematic illustration of time-resolved CIDNP (a) and switched external magnetic field CIDNP (b, c); (1) laser, (2) region of CIDNP generation, (3) detecting radio-frequency pulse. Field switching can occur both with (b) and without (c) change in its direction. weak and medium magnetic fields, photolysis is normally carried out in the field of a separate magnet and then the sample is transferred into the probe of an NMR spectrometer to record the spectrum.Photolysis can be performed directly in the probe of an NMR spectrometer with detection in weak magnetic fields;136 however, in this case, spectral resolution and the sensitivity are very low, which restricts the scope of this method to systems with abnormally high CIDNP intensity. It is evident that time reso- lution by a traditional time scheme cannot be used in the experi- ments with transfer of the sample from the field of an individual magnet to the probe of an NMR spectrometer. This opportunity appears in the case of fast switching of the magnetic field within a variable period after the laser pulse initiating the radical reaction (Fig. 18 b, c). The idea of switching the external magnetic field (SEMF) during a photochemical reaction was first proposed and implemented by Lavrik et al.137 The authors studied the effect of magnetic field switching on the intensity of fluorescence of excimers in radical-ion reactions.On the basis of the results obtained, estimates were made of the lifetimes of radical pairs. The potential of the methods based on the SEMF during the lifetime of radical pairs was studied theoretically.83 The research- ers considered the influence of SEMF on both the probability of recombination of radical pairs and the nuclear polarisation of diamagnetic products in strong magnetic fields (i.e. with allow- ance for only S ±T0 transitions). Three different modes of radical recombination were chosen. In the first mode, the singlet ± triplet transitions occur in the period between two contacts.This is accompanied by quantum oscillations. In the second mode, the S±T0 transitions do not have time to take place in the interval between two contacts but do occur effectively over the lifetime of the pair in the cage. In this case, no quantum oscillations are displayed; however, the recombination kinetics depend on the frequency of S ± T0 transitions. Finally, in the third mode, the singlet ± triplet transitions are relatively ineffective. This mode of recombination is of interest because it permits one to determine the lifetime of the pair in the cage. It should be noted that in experiments, the recombination mode is largely `controllable'. It can be changed, for example, by changing the solution viscosity.Bagryanskaya et al.138 ± 140 carried out experiments based on switching of the external magnetic field within a variable period after a laser pulse followed by recording of the CIDNP effect (the SEMF CIDNP method). It was found that the change of nuclear 3 t t t2 t t 939 polarisation of the products of radical reactions upon switching the external magnetic field can be induced both by the change in the conditions of spin dynamics of radical pairs and intermediate short-lived free radicals and by the change in the electron and nuclear polarisation at the instant of switching. On the one hand, this treatment substantially complicates interpretation of the experimental results but, on the other hand, it opens the following ways for the investigation of the chemical kinetics and spin dynamics of radical species.(i) Study of the dependence of polarisation on the delay t between the laser pulse and the instant of magnetic field switching allows one to explore the kinetics of CIDNP of geminate radical pairs. The time it takes to switch a magnetic field by 1073 T equals 1 ± 3 ns. The time resolution is determined by the accuracy of synchronisation of the laser pulse and the instant of magnetic field switching (20 ± 30 ns). Unlike the traditional time-resolved CIDNP, this method allows investigation of the CIDNP kinetics in weak and medium magnetic fields. This time resolution in NMR spectrometers is, in principle, impossible in weak fields because it is determined by the duration of the radio-frequency pulse, which lies in the millisecond time interval for weak and medium fields. (ii) Variation of the delay t permits one to change selectively the conditions of spin dynamics for geminate and diffusion radical pairs and thus to study the chemical kinetics of radical species.(iii) Switching of the magnetic field changes the electron and nuclear polarisation of free radicals; this makes it possible to study the mechanisms of electron relaxation and the electron ± nuclear cross-relaxation and CIDEP in weak magnetic fields. Since the change in the magnetic field by*1072 T causes only a minor change in the CIDNP in strong magnetic fields, realisa- tion of the SEMF CIDNP technique in strong magnetic fields requires that a pulsed magnetic field of a substantial magnitude be created.This is a complex engineering problem. In weak magnetic fields, CIDNP changes quite substantially as the magnetic field changes by *1073 T. The SEMF CIDNP method in weak and medium magnetic fields was implemented using a setup in which photolysis was carried out in the field of a separate magnet and, subsequently, after polarisation appeared, the polarised diamag- netic product was rapidly (over a period shorter than the charac- teristic time of nuclear relaxation) delivered via a fast-flow system to the probe of an NMR spectrometer in which CIDNP was recorded. The magnetic field is switched by use of Helmholtz coils, which create an additional field in the magnet with a switching amplitude of 7 mT.This markedly restricts the range of fields in which the SEMF CIDNP kinetics can be measured and is due to a requirement of the method according to which the absolute magnitudes of CIDNP in the initial and the final fields should be substantially different. Therefore, the highest sensitivity of the method is attained in those fields, in which the curve for the field dependence of CIDNP has `steep slopes'. 1. Chemically induced nuclear polarisation under conditions of non-adiabatic switching of the magnetic field The time of electron relaxation of free radicals does not actually depend on the switching of the external magnetic field. However, the situation changes if the chemical processes, for example DEE, influence the electron relaxation.For radical-ion reactions with DEE, the relaxation times in weak fields can be markedly depend- ent on the external magnetic field. As has been shown above, CIDNP in radical-ion reactions is influenced substantially by DEE reactions. The dependence of CIDNP kinetics on the concentration of the donor or acceptor molecules can be used to measure the DEE rate and to study the electron polarisation in weak magnetic fields. The utility of the SEMF CIDNP method for measurement of DEE rates has been demonstrated in relation to the photoisomer- isation of fumaronitrile with naphthalene in polar solvents.139 Figure 19 presents the kinetic curves for the SEMF CIDNP of fumaronitrile and maleonitrile recorded for magnetic fields940 CIDNP (rel.u.) a 1.0 0.5 123 0 300 100 200 t /ns 70.5 71.0 b 1.0 0.50 600 400 200 800 t /ns 70.5 71.0 Figure 19.Kinetic curves of the SEMF CIDNP for the photoisomerisa- tion of the fumaronitrile with naphthalene constructed using the NMR signals of maleonitrile (B0=1 mT, B0=71 mT) (a) and fumaronitrile (B0=3 mT, B0=5 mT) (b). Fumaronitrile concentration (mol litre71): (1) 561073, (2) 261072, (3) 461072. switched either with or without inversion of their direction. The kinetics of CIDNP detected based on an in-cage reaction product, maleonitrile, are determined by the time required for the forma- tion of radical pairs. For the escape reaction product, fumaroni- trile, the kinetics are stipulated by the time of electron relaxation of the intermediate radical anion.The characteristic time of the decay of the CIDNP signal (of maleonitrile) is determined by the time of formation of geminate radical pairs; at high fumaronitrile concentrations [(2 ± 4)61072 mol litre71] it amounts to 15 ± 20 ns, while at low concentrations (561073 mol litre71), it is 30 ± 40 ns. On the basis of analysis of the kinetic curves of CIDNP constructed using the signals of maleonitrile in the NMR spectra, it was concluded that the time it takes for CIDNP in geminate radical pairs to appear is either shorter than or com- parable with the duration of a laser pulse. Meanwhile, the kinetic curves of CIDNP detected from the NMR signal of fumaronitrile depend substantially on the rate of DEE.In the case of fast exchange, the obtained values coincide with the times of electron relaxation due to DEE, while for slow exchange, the CIDNP kinetics are determined by both the rate ofDEE and the lifetime of free radicals. It is worth noting that the typical lifetimes of radical pairs in homogeneous solutions are in the nanosecond range. Lasers with a pulse duration of 10 ± 20 ns are used to initiate photochemical reactions; this stipulates the characteristic time of formation of the radical pair. Thus, experimental kinetic curves for CIDNP con- structed for geminate radical pairs in homogeneous solutions are expected to reflect the kinetics of formation rather than the lifetime of the pair.The main advantage of the method discussed here is separation of the contributions of geminate and diffusion radical pairs to CIDNP. The high time resolution and the possibility of investigating the kinetics of CIDNP in weak magnetic fields makes this method attractive for the investigation of spin dynamics of micellised radical pairs and short-lived biradicals. However, an attempt to carry out SEMF CIDNP experiments in micellised radical pairs demonstrated that, to obtain a satisfactory signal-to-noise ratio, the experiments should be carried out in fields whose magnitudes are close to the CIDNP maximum intensity. The field dependence of CIDNP for micellised radical pairs has almost in all cases the characteristic emission appearance, indicating the predominant E G Bagryanskaya, R Z Sagdeev contribution of the S ± T7 transitions to CIDNP.Thus, the kinetics of SEMF CIDNP should also be calculated with allow- ance for S ±T7 transitions. This brought about the necessity of modifying the theory of CIDNP under conditions of SEMF for the case of radical pairs with limited mobility. This problem was successfully solved 140, 141 on the basis of a kinematic approach. The time of switching of the magnetic field was assumed to be shorter than the time of intersystem crossing, electron relaxation and the exit of radicals from the micelles. To describe the motion of a micellised radical pair, the model for motion in the confined space was utilised.The calculation of the CIDNP effect as a function of the instant t0 when the magnetic field is switched using typical parameters of a micellised radical pair showed that no quantum oscillations occur and the effect decays monotonically. The value by which the field changes has no influence on the pattern of time dependence, which is determined by the viscosity of the solution inside the micelle, but influences only the scale of the CIDNP effect. An interesting result is drawn if singlet ± triplet transitions in one sub-ensemble of radical pairs stop after the magnetic field switching. In this case, quantum oscillations are observed on the time dependence of the effect. An expression was derived describing averaging of the S ± T0 transitions in the absence of relaxation and at low reactivity of the radical pair. In this case, conclusions about the character of movement of radicals in the radical pair can be drawn from the shape of the kinetic curve of SEMF CIDNP. The purpose of the work by Fedin et al.140 was to test the SEMF CIDNP method on a series of model reactions involving micellised radical pairs, to study the correspondence between the theoretical and experimental data, and to study the possibility of using this method for determination of the parameters of molec- ular and spin dynamics of micellised pairs.The researchers chose three photochemical reactions corresponding to different limiting cases of the ratio of the rates of radical exit from the micelles (kex), electron relaxation (krel) and S ± T7 conversion (klc) and conform- ing to the approximation at which only one magnetic nucleus is taken into account.In relation to these reactions, it was demon- strated that approximation of the kinetics by numerical calcu- lations can provide information on the exit rate of the radicals from the micelle, the rate of decarbonylation, the rate of dipole ± dipole electron relaxation, the rate of electron relaxation caused by anisotropy of HFC, and the rate of S ± T7 transitions occur- ring in the area of energy level crossing. Estimation of the rates of S±T7 transitions for the systems studied suggested strong aniso- tropy of the exchange interaction. The capacity of the SEMF CIDNP method for the investiga- tion into short-lived biradicals has been evaluated 142 taking photolysis of acylalkyl cyclic ketones as an example.The use of substituted ketones allowed the researchers to compare the CIDNP kinetics for biradicals differing only in HFC constants and to estimate the lifetime of the triplet molecule. The measure- ment of the lifetimes of short-lived biradicals formed upon photolysis of a,a-diphenylcyclododecanone showed that the kinetic curves of CIDNP for this reaction in weak magnetic fields are described with rather high accuracy by an exponential curve and the characteristic time of the curve decay coincides with the data obtained from the optical absorption of biradicals. 2. Chemically induced nuclear polarisation under conditions of adiabatic switching of the external magnetic field Switching of magnetic field can be accompanied by the transfer of electron polarisation to the nuclei and vice versa. The time and conditions of magnetic field switching are important factors determining the type of polarisation transfer (adiabatic or non- adiabatic). It has been shown 143 that adiabatic switching of the magnetic field with the change in its direction results in the transfer of the integral electron polarisation to the nuclei.The effect of the adiabatic change of the external magnetic field on the transfer of integral CIDEP on the radical nuclei and the subse- quent transfer to the diamagnetic products of radical chemicalDynamic and stimulated nuclear polarisation in photochemical radical reactions a b bb ba aa ab bb ba aa ab aa ab bb ba aa ab bb ba Figure 20.Schematic illustration of the effect of adiabatic switching of the external magnetic field on the nuclear polarisation of radicals; (a) the initial integral electron polarisation completely disappears after adiabatic switching of the magnetic field, (b) the initial multiplet electron polar- isation does not change after the magnetic field switching. reactions was studied. The limits of the adiabatic character of magnetic field switching were found from the Landau ± Zener formula 144 to be 45 ± 1100 nm. The lower limit is determined by the elucidation of time evolution of the spin system under adiabatic conditions, while the upper limit results from the requirement that radicals should react to give a diamagnetic product over a period of time shorter than the time of nuclear relaxation.The switching of the field was supposed to take place after the appearance of CIDEP. The adiabatic field switching is selective with respect to the mechanism of the appearance of CIDEP; the triplet and radical-pair mechanisms were considered as examples (Fig. 20). Passage through the zero magnetic field is accompanied by the transfer of the occupancy of levels: bean>aebn and aebn>bean. It was found that the integral CIDEP is completely transferred to the radical nuclei and, after radical recombination, it can be detected in diamagnetic products. In the case of multiplet CIDEP, switching of the field does not influence the polarisation of the diamagnetic products.The optimal conditions for the transfer of electron polarisation were analysed numerically in the study cited as functions of the instant the field switching starts, of the duration of its front and of the velocity of passage through the zero-field region. Evidently, non-zero nuclear polarisation is also transferred on electrons upon switching the magnetic field. This may result in the decay of escape nuclear polarisation. Since most radical reactions in weak magnetic fields are usually accompanied by substantial CIDNP, adiabatic change of the magnetic field can be used to separate the contributions of in-cage and escape polarisations.It has been shown 139 that transfer of electron polarisation to the nuclei takes place in weak magnetic fields even in the case of non- adiabatic switching of the magnetic field. The potential of the SEMF 1H CIDNP method as applied to short-lived radical species in a homogeneous solution has been analysed.145 For this purpose, several photochemical reactions with participation of alkyl radicals PhCH2, Me3 C and PhC(O)CH2 in magnetic fields of 0.5 ± 2.5 mT were studied. From analysis of experimental data, it was concluded that the kinetics of polarisation is determined by the electron ± nuclear relaxation in the intermediate radicals. The measured relaxation times of the PhCH2, Me3C and PhC(O)CH2 radicals in benzene were 2.20.3, 7.10.3 and 1.20.2 ms.Figure 21 shows the kinetic curves of CIDNP for the photolysis of dibenzyl ketone constructed using the NMR signals of dibenzyl ketone. The CIDNP kinetics for dibenzyl ketone is determined by the rate of decarbonylation of the phenacyl radical, whereas for dibenzyl, the decay of the kinetic curve does not depend on the initial concen- 941 CIDNP (rel.u.) 1.0 123 0.8 0.6 0.4 0.20 0 2 4 6 8 10 12 t /ms Figure 21. Experimental kinetic curves of SEMF CIDNP for the photol- ysis of dibenzyl ketone detected using theNMRsignals of dibenzyl ketone (1 ) and dibenzyl: (2) I1 , (3) I2=3I1 . tration of radicals and is equal to the electron relaxation time of the benzyl radical. The decays of the kinetic curves found for the photolyses of other compounds involving the benzyl radical do not depend on the initial concentration of the radicals either (Fig. 22).CIDNP (rel.u.) 1.0 12 0.8 0.6 0.4 0.20 15 10 5 0 t /ms Figure 22. Kinetic curves of SEMF CIDNP for the photolysis of dibenzyl ketone (1) and deoxybenzoin (2). TheSEMF31PCIDNP method was used to determine the spin relaxation times for a series of phosphonyl radicals O=P(O)XY (X, Y are various substituents) in a 5 mT field in acetonitrile and in dioxane. If the anisotropy of HFC is modulated only by the rotation of the radical as a whole with characteristic times of the order of several picoseconds, the relaxation time in acetonitrile should be shorter than that in dioxane. However, an opposite relationship was found in the experiment, which was explained by the occurrence of deformation vibrations with characteristic periods of several nanoseconds.It was found that, if the relaxation time of the radical is shorter than its lifetime, the kinetics of SEMF CIDNP are determined by relaxation. 3. Polarisation under conditions of double switching of the external magnetic field Recently, further updating of the SEMF CIDNP method was proposed, which implies the use of two non-adiabatic magnetic field switchings separated in time during the reaction.146 The researchers considered the variation of the polarisation of free short-lived radicals induced by double switching of the magnetic field and determined the dependence of CIDNP of the diamag- netic products on the delay between the switchings.The time diagram of the experiment is shown in Fig. 23 a. In the first stage when 0<t<t1 , the radicals are formed under the action of a laser pulse in the external magnetic field H01 . It is assumed that by the time instant t1, the primary polarisation of the radicals has arisen and they have moved out of the cage to the bulk. In the second stage when t1<t<t2 , the magnetic field undergoes a942 b a z H Heff ~ Hfin o0 H01~s0 4 2 1 H02~sS 3 2 x A t 0 t2 t1 3 4 ~sSOtU Figure 23. Time diagram (a) and vector model (b) in CIDNP experiments with double switching of the external magnetic field; (1) laser pulse, (2) preparation, (3) evolution; (4) recording.non-adiabatic change to reach the value H02 . The radical polar- isation starts to oscillate, the characteristic frequencies of oscil- lation being dependent on the structure of the spin levels of the radicals in the fieldH02 and on the level occupancies. Finally, in the third stage when t2<t, the field undergoes a non-adiabatic change to reach the value Hfin 44 H02 . The nuclear polarisation of the radicals ceases to change and, upon recombination, passes to the diamagnetic products. Thus, the polarisation of oscillations in the intermediate radicals can be recorded using the dependence of the CIDNP intensity of the recombination products on the delay t=t27t1. The appearance of beats of the CIDNP of radicals and their diamagnetic products can be illustrated using a vector model (Fig.23 b). For the case of radical with one magnetic nucleus with the spin I=1/2, the two-level approximation jabN>,<baNj and the vector model are applicable. According to Adrian's approach,147 we consider the operators of the projections ^ Fz , ^ Fx of the effective spin, which act on the basis functions jabN>,<baNj as the corresponding Pauli matrices F^zjabN >=12 jabN >; F^zjbaN >=¢§12 jbaN >; F^xjabN >=12 jbaN >; F^xjbaN >=12 jabN > . Then the Hamiltonian H a o0 S^z a AO S^ I^U ^ for the radical in question in a weak field determined in the subspace O with the basis set jabN>,<baNj assumes the form H a A ^ Fx a o0 ^ Fz a O~Heff , ^ F) , ^ where ~Heff= (A, 0, o0).Thus, evolution of the state jC> from the subspace O can be represented as precession of the effective spin ~s in an effective magnetic field with the z component o0 , the x component A and the frequency o20 a A2 (see Fig. 23 b). The projection of the qAAAAAAAAAAAAAAAAA vector ~s on the z axis is equal to the difference between the occupancies of the levels jabN>,<baNj. The nuclear polar- isation corresponding to the state jC> is proportional to this value. The permanent generation of effective spins at the prepa- ration stage results in the vectors ~s0 covering uniformly the surface of the cone, so that the average vector ~sS is directed along the cone axis, i.e.along the vector ~Heff . After non-adiabatic switching of the magnetic field to reach, for example, a zero magnitude, the vector ~Heff is directed along the x axis and has the length A.At the initial instant, the vector~sS remains constant then it starts to precess along the x axis with a frequency equal in magnitude to the HFC constant. The projection of the vector ~s on the z axis varies periodically with the same frequency. If the magnetic field switches from the zero value to the valueHfin 44A after a time t, the nuclear polarisation of radicals stops changing and passes to the diamagnetic reaction products after recombina- tion. The vector ~sS would precess about the z axis and its projection on this axis would remain constant.Thus, by applying a second switching of the magnetic field after different delays t, one can record the dependence of the oscillation of the CIDNP of diamagnetic products on t. The paramagnetic relaxation in the radicals would result in a decrease in the length of the vector ~s with an increase in the time t1 and in a decrease in the amplitude of beats with an increase in the delay t. When t144T1 (T1 is the spin ¡¾ lattice relaxation time), all the spin levels of the radical are almost equally occupied, which corresponds to the decrease in the length of the vector~sS to zero. Evidently, in this case, the CIDNP of the diamagnetic products of the reaction does not oscillate. Thus, for high-quality recording of the oscillation, it is requited that T1 be much longer than t1.Data on the oscillation of the CIDNP of the diamagnetic products bear information on the short-lived intermediates from which these products are formed. The amplitude, the frequency and the phase of oscillations depend on the structure and occupancy of the spin levels in the short-lived intermediate species. The method described above was first implemented by Makarov et al.148 in relation to the photolysis of p-benzoquinone with 2-propanol-d8 and to sensitised photoisomerisation of fumaroni- trile. The oscillations of the CIDNP of the reaction products were CIDNP (t)/CIDNP (B01 ) 0.2 0.1 0 70.1 70.2 80 Amplitude 0.79 1.0 1.02 0.8 0.6 0.4 0.20 0 1.0 0.5 Figure 24. Oscillation of the CIDNP of the products of photolysis of p-benzoquinone with isopropanol-d8 in acetonitrile in experiments with double switching of the external magnetic field (a) and Fourier spectrum of the resulting kinetic curve (b).The experimental positions of the resonance frequencies in the spectrum are given (the calculated positions are 0.76 and 1.04). E G Bagryanskaya, R Z SagdeevSS a 160 240 t /ns b 2.0 1.5 2.5 n/ mTDynamic and stimulated nuclear polarisation in photochemical radical reactions recorded as functions of the delay t between the magnetic field switchings (Fig. 24). It was found that the amplitude, phase and the frequency of CIDNP beats are determined by the hyperfine structure and by the occupancies of the spin levels in the intermediate radicals.It was concluded from comparison of the experimental and theoretical data that the electron polarisation of semiquinoid radicals and fumaronitrile radical ions in weak magnetic fields is negative. It appears promising to use this method for investigating electron polarisation and the times of electron ± nuclear relaxation of short-lived radicals in weak mag- netic fields to which conventional time-resolved EPR is inappli- cable. VI. Conclusion A vast body of experimental information has been accumulated on the use of high-sensitivity magnetic-resonance techniques based on the influence of variable resonance magnetic fields and switched constant magnetic fields on the nuclear polarisation of the products of radical reactions.The examples given in this review demonstrate the applicability of these methods to the study of the mechanism of radical reactions, identification of latent radical stages and measurement of DEE rates. The main regu- larities of polarisation of radical species with a broad range of lifetimes and in homogeneous and molecular-organised media have been elucidated. By now, the theory of the influence of RF fields on the nuclear polarisation in radical reactions has been thoroughly developed and employed to gain information on short-lived radical species � their spin dynamics and chemical kinetics. The time-resolved SNP method can be used to separate the contributions to the polarisation of geminate and diffusion radical pairs and to obtain quantitative data on their kinetics.The SEMF CIDNP technique provides the possibility of measuring the rates of DEE, the kinetics of radicals, the lifetimes of short-lived biradicals and micellised radical pairs. This can prove useful in those cases where the use of flash photolysis is difficult due to the overlap of optical absorption bands. An additional advantage of the SEMF CIDNP method is the spectral resolution of NMR, which allows one to follow the transformation of biradicals into diamagnetic products. The DNP method provides the possibility of measuring the electron relaxation times for short-lived radicals and the signs and magnitudes of their HFC constants. Detection of DNP of the final reaction products can be used to investigate the mechanisms of radical processes.Unlike the CIDNP method, in the case of DNP, the formation of radical pairs in solution is not necessary. The doubly switched external magnetic field CIDNP method is supposed to be suitable for measuring the kinetics and the spin relaxation times of radicals, for investigating the CIDEP in weak magnetic fields, and for gaining information on the occupancy of the spin levels in the radicals. Thus, nowadays, there exists a basis for the use of highly sensitive magnetic-resonance methods in the studies of complex photochemical reactions. The highest efficiency in the application of these methods can be attained by combining them with the techniques of laser flash photolysis and time-resolved EPR.This review was supported by the Russian Foundation for Basic Research (Projects Nos 99-03-33488 and 99-03-32459) and by the INTAS (Grant No. 99-03-1764). References 1. V M Grigoryants, O A Anisimov, Yu N Molin Zh. Strukt. Khim. 23 (3) 4 (1982) a 2. Yu N Molin, O A Anisimov, V I Melekhov, S N Smirnov Faraday Discuss. Chem. Soc. 78 1 (1984) 3. E L Frankevich, A I Pristupa Pis'ma Zh. Techn. Fiz. 24 397 (1976) 4. E L Frankevich, S I Kubarev, in Triplet State ODMR Spectroscopy (Ed. R H Clarke) (New York: Wiley, 1982) p. 137 943 6. M K Bowman, D E Budil, G L Closs, A G Kostka, C A Wraight, 5. V I Lesin, V P Sakun, A I Pristupa, E L Frankevich Phys. Status Solidi B 84 513 (1977) J R Norris Proc. Natl. Acad. Sci. USA 78 3305 (1981) 7. J R Norris, M K Bowman, D E Budil, J Tang, C A Wraight, G L Closs Proc. Natl.Acad. Sci. USA 79 5532 (1982) 8. A I Grant,K A McLauchlan, S R Nattrass Mol. Phys. 55 557 (1985) 9. K W Moehl, E J Lous, A J Hoff Chem. Phys. Lett. 121 22 (1985) 10. E L Frankevich Zh. Eksp. Teor. Fiz. 50 1139 (1993) 11. H Murai, A tsuyma, Y Iwasaki, K Enjo, K Maeda, T Azumi Appl. Magn. Reson. 12 411 (1997) 12. E G Bagryanskaya, Yu A Grishin, N I Avdievich, R Z Sagdeev, Yu N Molin Chem. Phys. Lett. 128 162 (1986) 13. R Z Sagdeev, Yu A Grishin, A Z Gogolev, A V Dushkin, A G Semenov, Yu I Molin Zh. Strukt. Khim. 20 1132 (1979) a 14. A D Trifunac, W T Evanochko J. Am. Chem. Soc. 102 4598 (1980) 15. R H D Nuttall, A D Trifunac Chem. Phys. Lett. 81 151 (1981) 16.R H D Nuttall, A D Trifunac J. Phys. Chem. 86 3963 (1982) 17. R H D Nuttall, A D Trifunac J. Magn. Reson. 51 118 (1983) 18. A D Trifunac, R G Lawler, D M Bartels,M C Thurnauer Prog. React. Kinet. 14 43 (1986) 19. Yu A Grishin, A Z Gogolev, E G Bagryanskaya, A V Dushkin, A G Semenov, R Z Sagdeev, Yu N Molin Dokl. Akad. Nauk SSSR 255 1160 (1980) b 20. K Maeda, Q-X Meng,Y Yamakage, T Azumi Z. Phys. Chem. 182 99 (1993) 21. QMeng,YYamakage,KMaeda, T Azumi Phys. Chem. 180 95 (1993) 22. E G Bagryanskaya, N N Lukzen, I V Koptyug, R Z Sagdeev, M N Usacheva Appl. Magn. Reson. 1 431 (1990) 23. R Z Sagdeev, Yu N Molin, K M Salikhov, Yu A Grishin, A V Dushkin Bull. Magn. Reson. 2 66 (1980) 24. R Z Sagdeev, E G Bagryanskaya Pure. Appl.Chem. 62 1547 (1990) 25. E G Bagryanskaya, R Z Sagdeev Prog. React. Kinet. 18 63 (1993) 26. Y Yamakage, Q Meng, K Maeda, T Azumi Chem. Phys. Lett. 204 411 (1993) 27. S V Dvinskikh, A V Yurkovskaya, H-M Vieth J. Phys. Chem. 100 8125 (1996) 28. N I Avdievich, E G Bagryanskaya, Yu A Grishin, R Z Sagdeev Chem. Phys. Lett. 155 141 (1989) 29. V P Ivannikov, Yu D Chernousov, I V Shebolaev Vopr. At. Nauki Tekhn. 6 55 (1989) 30. A L Buchachenko, A M Vasserman Stabil'nye Radikaly (Stable Radicals) (Moscow: Nauka, 1973) 31. K H Hausser, D Stelik Adv. Magn. Reson. 3 1772 (1070) 32. E G Bagryanskaya, G S Ananchenko, T Nagashima, K Maeda, S Milikisyants, H Paul J. Phys. Chem. A 103 1127 (1999) 33. S Nagakura, H Hayashi, T Azumi (Eds) Dynamic Spin Chemistry.Magnetic Controls and Spin Dynamic of Chemical Reactions (Tokyo: Kodansha, 1998) 34. K Maeda, Y Yamakage, K Shindo, T Azumi Appl. Magn. Reson. 5 177 (1993) 35. Y Yamakage, Q Meng, S S Ali, K Maeda, T Azumi Indian Acad. Sci. 105 629 (1993) 36. Q Meng, Y Yamakage, T Aizawa, K Maeda, T Azumi Indian Acad. Sci. 105 619 (1993) 37. Y Yamakage, K Maeda, T Azumi Mol. Phys. 90 3 (1997) 38. K A McLauchlan, in Modern Pulsed and Continuous Wave Electron Spin Resonance (Eds L Kevan,MK Bowman) (New York: Wiley, 1990) p. 285 39. K A McLauchlan, D G Stevens Mol. Phys. 57 223 (1986) 40. F Jent, H Paul, K A McLauchlan, D G Stevens Chem. Phys. Lett. 141 443 (1987) 41. V I Valyaev, Yu N Molin, R Z Sagdeev, P J Hore, K A McLauchlan, N J K Simpson Mol.Phys. 63 891 (1988) 42. A V Yurkovskaya, Yu P Tsentalovich, R Z Sagdeev Chem. Phys. Lett. 171 406 (1990) 43. Yu P Tsentalovich, A A Frantsev, A B Doctorov, A V Yurkovskaya, R Z Sagdeev J. Phys. Chem. 97 8900 (1993) 44. Yu P Tsentalovich, A V Yurkovskaya, A A Frantsev, A B Doctorov, R Z Sagdeev Z. Phys. Chem. 182 119 (1993) 45. M Salzmann, Yu P Tsentalovich, H Fischer J. Chem. Soc., Perkin. Trans. 2 2119 (1994) 46. O Morozova, Yu P Tsentalovich, A V Yurkovskaya, R Z Sagdeev Chem. Phys. Lett. 246 499 (1995)944 47. S N Batchelor, H Fischer J. Phys. Chem. 100 556 (1996) 48. E G Bagryanskaya, N I Avdievich, Yu A Grishin, R Z Sagdeev Chem. Phys. 135 123 (1989) 49. V R Gorelik, E G Bagryanskaya, N N Lukzen, I V Koptyug, V V Perov, R Z Sagdeev J.Phys. Chem. 100 580 (1996) 50. KMSalikhov, C H Boch, D Stehlik Appl. Magn. Reson. 1 195 (1990) 51. E G Bagryanskaya, R Z Sagdeev Z. Phys. Chem. 80 63 (1993) 52. K M Salikhov, Yu N Molin, R Z Sagdeev, A L Buchachenko, in Spin Polarization and Magnetic Effects in Radical Reactions (Ed. Yu N Molin) (Amsterdam: Elsevier, 1984) 53. E Schaffner, H Fischer J. Phys. Chem. 100 1657 (1996) 54. M N Usacheva, E G Bagryanskaya, T I Viktorova, I V Drozdenko, in Mezhdunarodnaya Konferentsiya po Fotokhimii (Tez. Dokl.), Kiev, 1992 [International Conference on Photochemistry (Abstracts of Repots), Kiev, 1992] p. 170 55. E G Bagryanskaya, V R Gorelik, R Z Sagdeev, M N Usacheva, H Paul, in The Xth International Conference `Magnetic Resonance in Chemistry and Biology' (Abstracts of Reports), Suzdal, 1998 p.49 56. G S Ananchenko, E G Bagryanskaya, R Z Sagdeev Chem. Phys. Lett. 282 450 (1998) 57. E G Bagryanskaya, G S Ananchenko, A P Parnachev, P A Purtov, R Z Sagdeev, in Symposium on Magnetic Field and Spin Effects in Chemistry and Related Phenomena (Abstracts of Reports), Ierusalim, 1997 p. 49 58. G S Ananchenko, Candidate Thesis in Chemical Sciences, Institute of Chemical Kinetics and Combustion, Siberian Branch of the Russian Academy of Sciences, Novosibirsk, 1999 59. G S Ananchenko, P A Purtov, E G Bagryanskaya, R Z Sagdeev J. Phys. Chem. A. 101 3848 (1997) 60. G S Ananchenko, P A Purtov, E G Bagryanskaya, R Z Sagdeev J. Phys. Chem. A. 103 3430 (1999) 61. G L Closs, R J Miller, O D Redwein Acc.Chem. Res. 18 196 (1985) 62. J K Vollenweider, H Fischer Chem. Phys. 108 365 (1986) 63. N J Turro,MB Zimmt, I R Gould J. Am. Chem. Soc. 105 6347 (1983) 64. Yu P Tsentalovich, A V Yurkovskaya, R Z Sagdeev, A A Obynochny, P P Purtov Chem. Phys. 139 307 (1989) 65. S I Kubarev, E A Pshenichnov Chem. Phys. Lett. 28 66 (1974) 66. S N Batchelor, K A McLauchlan, I A Shkrob Chem. Phys. Lett. 181 327 (1991) 67. M Okazaki, S Sakata, R Konaka, T Shiga J. Chem Phys. 86 6792 (1987) 68. S A Mikhailov,KMSalikhov, AMPlato Chem. Phys. 117 197 (1987) 69. I V Koptyug, N N Lukzen, E G Bagryanskaya, A B Doktorov Chem. Phys. Lett. 175 467 (1990) 70. A I Shushin Chem. Phys. Lett. 181 274 (1991) 71. S M Mikhailov, Candidate Thesis in Physicomathematical Sciences, Institute of Chemical Kinetics and Combustion, Siberian Branch of the Academy of Sciences of the USSR, Novosibirsk, 1988 72.S M Mikhailov, P P Purtov Teor. Eksp. Khim. 24 519 (1988) c 73. AMOsintsev, P P Purtov, KMSalikhov Chem. Phys. 174 237 (1993) 74. V A Morozov, A B Doktorov, R Z Sagdeev Chem. Phys. 179 287 (1994) 75. I A Shkrob, V F Tarasov, E G Bagryanskaya Chem. Phys. 153 427 (1991) 76. E G Bagryanskaya, V F Tarasov,N I Avdievich, I A Shkrob Chem. Phys. 162 213 (1992) 77. V F Tarasov, E G Bagryanskaya, I A Shkrob, N I Avdievich, N D Ghatlia, N N Lukzen, N J Turro, R Z Sagdeev J. Am. Chem. Soc. 117 110 (1995) 78. A P Parnachev, E G Bagryanskaya, V F Tarasov, N N Lukzen, R Z Sagdeev Chem. Phys.Lett. 244 245 (1995) 79. A I Shushin Chem. Phys. Lett. 282 413 (1998) 80. P A Purtov, Candidate Thesis in Physicomathematical Sciences, Institute of Chemical Kinetics and Combustion, Siberian Branch of the Academy of Sciences of the USSR, Novosibirsk, 1982 81. K M Salikhov, R T Galeev Mol. Phys. 99 1005 (1998) 82. J B Pedersen Chem. Phys. Lett. 52 333 (1977) 83. A B Doktorov, S A Mikhailov, P P Purtov Chem. Phys. 160 223 (1992) 84. A M Osintsev, P A Purtov, A B Doktorov Chem. Phys. 185 281 (1994) 85. K M Salikhov, Y Sakaguchi, H Hayashi Chem. Phys. 220 355 (1997) 86. E G Bagryanskaya, Yu A Grishin, R Z Sagdeev Chem. Phys. Lett. 113 234 (1985) E G Bagryanskaya, R Z Sagdeev 87. E G Bagryanskaya, Yu A Grishin, R Z Sagdeev, Yu A Molin Chem.Phys. Lett. 114 138 (1985) 88. E G Bagryanskaya, Yu A Grishin, R Z Sagdeev, T V Leshina, N E Polyakov, Yu A Molin Chem. Phys. Lett. 117 220 (1985) 89. Yu P Tsentalovich, E G Bagryanskaya, Yu A Grishin, A A Obynochny, R Z Sagdeev, H-K Roth Chem. Phys. 142 75 (1990) 90. I V Koptyug, Candidate Thesis in Physicomathematical Sciences, Institute of Chemical Kinetics and Combustion, Siberian Branch of the Academy of Sciences of the USSR, Novosibirsk, 1990 91. N I Avdievich, Candidate Thesis in Physicomathematical Sciences, Institute of Chemical Kinetics and Combustion, Siberian Branch of the Russian Academy of Sciences, Novosibirsk, 1992 92. V R Gorelik, N N Lukzen, A V Morozov, E G Bagryanskaya, R Z Sagdeev Chem. Phys. 224 229 (1997) 93.Y Yamakage, T Nagashima, K Maeda, H Murai, T Azumi Appl. Magn. Reson. 12 441 (1997) 94. A B Doktorov, S A Mikhailov, P P Purtov Chem. Phys. 160 239 (1992) 95. S A Mikhailov, P P Purtov, A B Doktorov Chem. Phys. 166 35 (1992) 96. I V Koptyug, N N Lukzen, E G Bagryanskaya, A B Doktorov, R Z Sagdeev Chem. Phys. 162 165 (1992) 97. I V Koptyug, E G Bagryanskaya, Yu A Grishin, R Z Sagdeev Chem. Phys. 145 375 (1990) 98. I V Koptyug, E G Bagryanskaya, R Z Sagdeev Chem. Phys. Lett. 163 503 (1989) 99. R Z Sagdeev, E G Bagryanskaya, I V Koptyug, Yu P Tsentalovich, A V Yurkovskaya, in Spin Chemistry (Oji International Conference on Spin Chemistry) (Ed. Y J I'Haya) (Tokyo: University of Electro- 100. F J J De Kanter, R Kaptein, R A von Santen Chem. Phys. Lett. 45 101. A V Yurkovskaya, O B Morozova, R Z Sagdeev, S V Dvinskikh, 102. S V Dvinskikh, G Buntkowsky, K M Salikhov, H-M Vieth Chem. Communications, 1991) p. 138 575 (1977) G Buntkowsky, H M Vieth Chem. Phys. 197 157 (1995) Phys. Lett. 268 401 (1997) 103. N V Lebedeva, E G Bagryanskaya, I V Koptyug, R Z Sagdeev, M D E Forbes Chem. Phys. Lett. 308 295 (1999) 104. J C Scaiano Acc. Chem. Res. 15 252 (1982) 105. A Yurkovskaya, S Grosse, S Dvinskikh, O Morozova, H M Vieth J. Chem. Phys. 103 980 (1999) 106. A L Buchachenko, R Z Sagdeev, K M Salikhov Magnitnye i Spinovye Effekty v Khimicheskikh Reaktsiyakh (Magnetic and Spin Effects in Chemical Reactions) (Novosibirsk: Nauka, 1978) p. 296 107. N J Turro,W R Cherry J. Am. Chem. Soc. 100 7431 (1978) 108. M B Zimmt, C Doubleday Jr , N J Turro J. Am. Chem. Soc. 106 3363 (1984) 109. I R Gould,M B Zimmt, N J Turro, B H Baretz, G F Lehr J. Am. Chem. Soc. 107 4607 (1985) 110. P P Levin, V A Kuzmin Chem. Phys. 162 79 (1992) 111. T Ulrich, U E Steiner Chem. Phys. Lett. 112 365 (1984) 112. E N Step, V F Tarasov, A Z Yankelevich Khim. Fiz. 7 1487 (1988) d 113. I A Shkrob, V F Tarasov, A L Buchachenko Chem. Phys. 153 443 (1991) 114. K A McLauchlan, S R Nattrass Mol. Phys. 65 1483 (1988) 115. G L Closs, M D E Forbes, J R Noris Jr J. Phys. Chem. 91 3592 (1987) 116. V V Tarasov, H Yashiro, K Maeda, T Azumi, I A Shkrob Chem. Phys. 212 353 (1996) 117. E G Bagryanskaya, A P Parnachev, G S Ananchenko, V F Tarasov, N N Lukzen, R Z Sagdeev, in The VIth Chianti Workshop on 120. G S Ananchenko, E G Bagryanskaya, V F Tarasov, R Z Sagdeev, Magnetic Resonance (Abstracts of Reports), San Miniato, 1995 p. 3 118. E G Bagryanskaya, N V Lebedeva, M V Fedin, in The 17th IUPAC Symposium on Photochemistry (Abstracts of Reports), Barcelone, 1998 p. 53 119. E G Bagryanskaya,M V Fedin, N V Lebedeva, in International Conference on Reactive Intermediates and Reaction Mechanisms (Abstracts of Reports), Ascone, 1998 p. 31 H Paul Chem. Phys. Lett. 255 267 (1996) 121. A P Parnachev, E G Bagryanskaya,R Z Sagdeev J. Phys. Chem.A 101 3854 (1997)945 Dynamic and stimulated nuclear polarisation in photochemical radical reactions 122. E G Bagryanskaya, Yu A Grishin, R Z Sagdeev, V F Tarasov, I A Shkrob, A L Buchachenko Khim. Fiz. 10 1365 (1991) d 123. N V Lebedeva, E G Bagryanskaya, V R Gorelik, I V Koptyug, R Z Sagdeev, in The VIth International Symposium on Magnetic Field and Spin Effects in Chemistry and Related Phenomena (Abstracts of Reports), Emmetten, 1999 p. 76 124. N I Avdievich, E G Bagryanskaya, V F Tarasov, R Z Sagdeev Z. Phys. Chem. 182 107 (1993) 125. U E Steiner, J R Wu Chem. Phys. 162 53 (1992) 126. V A Morozov, S V Isakov, R Z Sagdeev Chem. Phys. Rep. 16 559 (1997) 127. V F Tarasov,M D Forbes Spectrochim. Acta, Part A 56 245 (2000) 128. V V Tarasov, H Yashiro, K Maeda, T Azumi, I A Shkrob Chem. Phys. 212 253 (1996) 129. V F Tarasov, E G Bagryanskaya, Yu A Grishin, R Z Sagdeev, A L Buchachenko Mendeleev Commun. 85 (1991) 130. M Okazaki, Y Tamimoto, Y Konishi, K Toriyama J. Phys. Chem. 100 9403 (1996) 131. Y Konishi, M Okazaki, K Toriyama Appl. Magn. Reson. 11 135 (1996) 132. G L Closs, R J Miller J. Am. Chem. Soc. 101 1639 (1979) 133. R J Miller, G L Closs Rev. Sci. Instr. 52 1876 (1981) 134. J-K Vollenweider, H Fischer, J Hennig, R Leuschner Chem. Phys. 97 217 (1985) 135. J-K Vollenweider, H Fischer Chem. Phys. 124 333 (1988) 136. Yu A Grishin, Candidate Thesis in Physicomathematical Sciences, Institute of Chemical Kinetics and Combustion, Siberian Branch of the Academy of Sciences of the USSR, Novosibirsk, 1978 137. N L Lavrik, V E Khmelinsky Chem. Phys. Lett. 140 582 (1987) 138. A P Parnachev, E G Bagryanskaya, V F Tarasov, N N Lukzen, N I Avdievich, R Z Sagdeev, in The IIIrd International Symposium on Magnetic Field and Spin Effects in Chemistry and Related Phenomena (Abstracts of Reports), Chicago, 1994 p. 69 139. E G Bagryanskaya, V R Gorelik, R Z Sagdeev Chem. Phys. Lett. 264 655 (1997) 140. M V Fedin, P A Purtov, E G Bagryanskaya J. Chem. Phys. 111 5491 (1999) 141. A P Parnachev, P A Purtov, E G Bagryanskaya, R Z Sagdeev J. Chem. Phys. 107 9942 (1997) 142. E G Bagryanskaya, V R Gorelik, A P Parnachev,R Z Sagdeev, in The 28th Congress on Magnetic Resonance and Related Phenomena (Abstracts of Reports), Canterbury, 1996 p. 79 143. N N Lukzen, U E Steiner Mol. Phys. 86 1271 (1995) 144. L D Landau, E M Lifshits Teoreticheskaya Fizika (Theoretical Physics) Vol. 3 (Moscow: Nauka, 1998) 145. M V Fedin, E G Bagryanskaya, in Vseros. Shkola-Simpozium po Khimicheskoi Fizike (Tez. Dokl.), Tuapse, 1999 [ All-Russian School-Symposium on Chemical Physics (Abstracts of Reports), Tuapse, 1999] p. 42 146. N N Lukzen, V A Morozov, R Z Sagdeev Chem. Phys. 241 193 (1999) 147. F J Adrian J. Chem. Phys. 54 3918 (1971) 148. T N Makarov, E G Bagryanskaya, S R Shakirov, N N Lukzen, R Z Sagdeev Chem. Phys. Lett. 317 252 (2000) a�Russ. J. Struct. Chem. (Engl. Transl.) b�Dokl. Chem. Technol., Dokl. Chem. (Engl. Transl.) c�Theor. Exp. Chem. (Engl. Transl.) d�Russ. J. Chem. Phys. (Engl.

ISSN:0036-021X    

出版商:RSC    

年代:2000

数据来源: RSC

摘要:

Russian Chemical Reviews 69 (11) 947 ± 955 (2000) Reagents for the preparation and cleavage of 1,3-dithiolanes A K Banerjee,MS Laya Contents I. Introduction II. Reagents for the synthesis of 1,3-dithiolanes based on 1,2-ethanedithiol III. Reagents for the cleavage of 1,3-dithiolanes IV. Conclusion Abstract. group carbonyl the of protection for procedures on Data Data on procedures for protection of the carbonyl group by cleavage on and moiety 1,3-dithiolane the into conversion its by its conversion into the 1,3-dithiolane moiety and on cleavage of of the surveyed. are reagents different of action the under latter the latter under the action of different reagents are surveyed. The The preparation of 1,3-dithiolanes from carbonyl compounds fol- preparation of 1,3-dithiolanes from carbonyl compounds fol- lowed various of syntheses in used be can deprotection by lowed by deprotection can be used in syntheses of various organic organic compounds.The bibliography includes 54 references. compounds. The bibliography includes 54 references. I. Introduction Syntheses of organic compounds, including natural products, often generate a need to use protective groups. Thus if it is necessary that the reaction proceed selectively at one of several reaction centres of a polyfunctional compound, all other centres should be temporarily blocked. A good protective group must be removed selectively and quantitatively under the action of readily accessible reagents which do not react with functional groups to be regenerated. Various protective groups have already been devel- oped and these investigations are being continued.Analysis of the published data demonstrated that the introduction and removal of a protective group is often a complex multistep process. In the synthesis of particular carbonyl-containing products, it is sometimes necessary to prevent the reaction of the carbonyl group with reducing agents of acidic or alkaline nature, catalytic reducing agents and hydrides, some oxidising agents, strong and moderate nucleophiles, organometallic compounds, etc. In the reactions of polyfunctional organic compounds, it is very con- venient to protect the carbonyl group by converting it into the 1,3- dithiolane moiety because sulfur-containing functional groups are inert with respect to many reagents.Various reagents, which are stable under conventional reaction conditions, have been devel- oped for the synthesis of 1,3-dithiolanes.1, 2 Generally, 1,3-dithio- lanes are prepared by condensation of carbonyl compounds with 1,2-dithiols catalysed by protic or Lewis acids.3, 4 In the present review, the data on the reagents for the synthesis and cleavage of 1,3-dithiolanes developed in the last decade are surveyed. The mechanisms of formation and cleavage of 1,3- dithiolanes are beyond the scope of this review. A K Banerjee, MS Laya Centre of Chemistry, Venezuelan Scientific Research Institute, Apartado 21827, Caracas 1020-A, Venezuela. Fax (58-2) 504 13 50. Tel. (58-2) 504 13 24.E-mail: abanerje@quimica.ivic.ve Received 10 April 2000 Uspekhi Khimii 69 (11) 1032 ± 1041 (2000); translated by T N Safonova #2000 Russian Academy of Sciences and Turpion Ltd DOI 10.1070/RC2000v069n11ABEH000583 947 947 951 954 II. Reagents for the synthesis of 1,3-dithiolanes based on 1,2-ethanedithiol 1. Catalysis by Amberlyst-15 Amberlyst-15 is an efficient catalyst of the synthesis of 1,3- dithiolanes from some aldehydes,5 for example, from 1 or 2, under the action of 1,2-ethanedithiol (CH2SH)2. (CH2SH)2 O Me Amberlyst-15 1 S Me S S 3 (93%) S (CH2SH)2 O Amberlyst-15 4 (94%) 2 The reactions proceed at room temperature to form deriva- tives 3 or 4 in high yields. If the reaction mixtures contain ketones along with aldehydes, the former do not react with (CH2SH)2 under these conditions.However, this procedure can also be used for protection of ketones (even sterically hindered ketones), if the reaction is carried out with boiling. Under these conditions, 1,3- dithiane 6 is formed from pinacolone 5. O S S (CH2SH)2 But Me But Amberlyst-15 5 Me 6 (88%) In this case, the workup procedure is substantially simpler than those used in the synthesis of 1,3-dithiolanes catalysed by other reagents, in particular, by thionyl chloride and impregnated silica gel,6 or in the synthesis of 1,3-dithianes from 2,2-dimethyl- 1,3,2-dithiasilane in the presence of boron trifluoride etherate as a catalyst.7 Ketones of the type 5 react with (CH2SH)2 in the presence of Amberlyst-15 to form dithiolane 6.An important characteristic feature of this system is the fact that the reactions with ketones proceed at a slow rate because of which this procedure can be used for separation of aldehydes and ketones. 2. Catalysis by tetrachlorosilane It is known 8 that tetrachlorosilane (Lewis acid) is a soft and efficient catalyst of the synthesis of 1,3-dithiolanes. This com- pound catalyses the formation of dithiolanes from aromatic aldehydes containing both electron-donating and electron-with-948 drawing substituents. Thus the reactions of aldehydes 7± 9 with (CH2SH)2 in the presence of SiCl4 afforded dithiolanes 10 ± 12, respectively. S CHO S (CH2SH)2, SiCl4 7 10 (95%) S CHO S (CH2SH)2, SiCl4 Cl Cl 11 (89%) 8 S CHO S (CH2SH)2, SiCl4O2N O2N 12 9 The SiCl4-catalysed reactions of aromatic aldehydes (for example, 7 and 13) with (CH2SH)2 in the presence of acetophe- none (14) proceed absolutely chemoselectively to form the 1,3- dithiolanes 10 and 15, respectively.Methyl-substituted dithiolane 16 is not formed in this reaction. Good selectivity is observed also in the case of aliphatic aldehydes. COMe CHO (CH2SH)2, SiCl4 + 14 7 S S + S S Me 10 (98%) 16 (0%) S O O CHO (CH2SH)2, SiCl4 +16 +14 S 15 (99%) (0%) 13 Ku and Oh 8 believed that the bulkiness and branching of the carbonyl compound are the major factors governing the chemo- selectivity of formation of 1,3-dithiolanes. 3. Catalysis by bentonite Miranda et al.9 developed an excellent procedure for the synthesis of 1,3-dithiolanes.This procedure involves condensation of ketones with (CH2SH)2 in the presence of bentonite, which has also been used for the oxirane ring-opening 10 and conversion of oximes into ketones.11, 12 Dithiolanes 21 ± 24 were prepared in high yields from ketones 17 ± 20, respectively. The reactions were performed in anhydrous toluene in the presence of bentonite as a catalyst. S (CH2SH)2, bentonite O S 21 (99%) 17 O S S (CH2SH)2, bentonite 18 22 (90%) O S S (CH2SH)2, bentonite Me Me C6H13-n C6H13-n 19 23 (90%) Me MeMe 20 This procedure is very simple and bentonite is inexpensive. It should be noted that the amount of the catalyst required in this case is seven times smaller than the amount of montmorillonite KSF in analogous reactions and the crude reaction mixture gives only one spot on a thin-layer chromatogram.4. Catalysis by anhydrous LaCl3 Anhydrous LaCl3 proved to be a convenient and efficient cata- lyst 13 of conversions of the carbonyl compounds 7, 9, 13 and 17 under the action of (CH2SH)2 into the dithiolanes 10, 12, 15 and 21, respectively. The reactions proceeded smoothly giving the products in very high yields. Owing to mild conditions and the simplicity, this procedure is suitable for multistep syntheses of complex organic compounds. Hydrated rare-earth metal chlor- ides were ineffective for the preparation of 1,3-dithiolanes from some carbonyl compounds.Apparently, anhydrous salts adsorb water much more actively than the hydrates, which facilitates substitution in the intermediate hemithioacetal with elimination of a water molecule and coordination of the lanthanum ion at the oxygen atom of the carbonyl group resulting in an increase in the electrophilicity of the carbon atom. 5. Catalysis by H-Y zeolite The reactions of crotonaldehyde (25) and aromatic and sterically hindered ketones 26, 14 and 27 with (CH2SH)2 in the presence of H-Y zeolite (Si :Al=2.43) afforded 1,3-dithiolanes 28, 29, 16 and 30, respectively.14 Me O Ph 2627 The use of H-ZSM-5 zeolite (Si : Al = 45) and H-mordenite has met with only limited success. The advantages of H-Y zeolite are the simplicity of separation, high yields of the products and the possibility of regeneration of the catalyst.6. Catalysis by montmorillonite KSF Montmorillonite KSF was used 15 for conversions of some alde- hydes, for example, of 7, 9 and 31 ± 33, in the absence of solvents at room temperature to form 1,3-dithiolanes 10, 12 and 34 ± 36, respectively. A K Banerjee,MS Laya Me MeMe O O (CH2SH)2, bentonite S S O 24 (80%) S (CH2SH)2, H-Y zeolite CHO S Me 28 25 S S (CH2SH)2, H-Y zeolite Ph Ph Ph 29 S COMe S (CH2SH)2, H-Y zeolite Me 16 14 O S S (CH2SH)2, H-Y zeolite 30Reagents for the preparation and cleavage of 1,3-dithiolanes CHO (CH2SH)2, montmorrillonite KSF R 31 ± 33 R=C(O)Me (31, 34), OMe (32, 35), NMe2 (33, 36).In these reactions, clay KSF acts as a Brùnsted solid acid. The procedure for the synthesis is simple. Montmorillonite KSF is an inexpensive and selective catalyst, which does not exhibit corro- sive action. The reactions of aromatic and aliphatic aldehydes with (CH2SH)2 in the presence of clay KSF are chemoselective, viz., only aromatic aldehydes are converted into dithiolanes. The reactions with aromatic ketones (for example, 14 and 18) proceed more rapidly than those with aliphatic aldehydes. 7. Catalysis by SnCl2 .2H2O A simple and efficient procedure for conversions of steroid ketones 37 ± 39 into 1,3-dithiolanes 40 ± 42, respectively, involves their treatment with 1,2-ethanedithiol in the presence of soft Lewis acid, viz., SnCl2 .2H2O, in THF.16 Me Me Me O 37 Me S S 40 Me Me Me HO 38 O Me HO S 41 Me Me Me O 39 Me S S 42 S S R 34 ± 36 Me Me (CH2SH)2, SnCl2 . 2H2O Me Me Me Me Me Me (CH2SH)2, SnCl2 . 2H2O Me Me Me Me S Me Me (CH2SH)2, SnCl2 . 2H2O Me Me Me Me The reactions can be performed in a neutral medium. This procedure is suitable for the synthesis of various aliphatic, aromatic and sterically hindered steroid ketones. Under the reaction conditions, the double bond at the 4,5 position of a,b-unsaturated ketone 39 remains intact. This procedure for the preparation of dithiolanes may also be of use in more general cases. Anhydrous SnCl2 can be used as a catalyst as well.8. Catalysis by bis(trimethylsilyl) sulfate adsorbed on silica gel Bis(trimethylsilyl) sulfate adsorbed on silica gel is a soft and efficient catalyst of conversions of some carbonyl compounds (7, 9, 14, 17 and 18) into dithiolanes (10, 12, 16, 21 and 22, respectively) under the action of (CH2SH)2 in anhydrous dichloromethane at room temperature.17 The reaction conditions and the synthetic procedure are simple and convenient. The high activity of this catalyst is supported by the fact that low-reactivity ketones, for example, benzophenone (26), aceto- phenone (14) and 9-fluorenone (18), are smoothly converted into the dithiolanes 29, 16 and 22, respectively. These compounds were also synthesised with the use of another catalyst. It was demon- strated that the reaction in the absence of silicon dioxide requires drastic conditions, viz., boiling in benzene with azeotropic distil- lation of water. Undoubtedly, this reagent will find wider use in organic synthesis.9. Catalysis by MgI2 in diethyl ether Magnesium iodide in diethyl ether is an efficient catalyst of conversions of a series of carbonyl compounds, for example, of 7, 18, 38 and 39, into the dithiolanes 10, 22, 41 and 42, respectively, under the action of (CH2SH)2.18 The reactions proceed at room temperature giving dithiolanes in very high yields. This procedure is chemoselective. Thus dithiolanes are not formed from aromatic ketones, for example, from benzophenone (26) and acetophenone (14), even upon prolonged storage of the reaction mixtures in the presence of this reagent at room temperature.10. Catalysis by Dowex-50W-X8 Bhattacharyya and coworkers 19 proposed that cation exchange resin Dowex-50W-X8 be used for conversions of ketones 43 ± 46 under mild conditions to form dithiolanes 47 ± 50, respectively, in high yields. CHO (CH2SH)2, Dowex-50W-X8 Me 43 CHO O (CH2SH)2, Dowex-50W-X8 O 44 (CH2SH)2, Dowex-50W-X8 O 45 (CH2SH)2, Dowex-50W-X8 O 46 Apparently, resin Dowex-50W-X8 treated with HCl acts as an acidic catalyst and increases the electrophilicity of the carbon atom of the carbonyl group. In the presence of this catalyst, the sterically hindered aromatic ketones 14 and 26 are smoothly converted into the dithiolanes 16 and 29, respectively. This 949 S S Me 47 S O S O 48 S S 49 SS 50950 method can also be used in the case of compounds containing acid-sensitive groups. The reactions are completed in a short time.11. Catalysis by CuSO4 The reactions of some carbonyl compounds, for example, of a-tetralone (27), 5a-cholestan-3-one (37), dihydrocarvone (51), 4-phenylbutan-2-one (52), acetophenone (14) and benzophenone (26), with 1,2-ethanedithiol in dry THF in the presence of anhydrous copper sulfate, which is a gentle and inexpensive catalyst, afforded the thiolanes 30, 40, 53, 54, 16 and 29, respectively.20 S O S CH2 CH2 (CH2SH)2 Me Me CuSO4 Me Me 53 51 S Ph O (CH2SH)2 Ph S CuSO4 Me 52 Me 54 The reactions catalysed by CuSO4 .5H2Oafford dithiolanes in very low yields. The advantage of this method is the fact that it can be applied in the case of various aliphatic, aromatic and steroid ketones. 12. Catalysis by bismuth halides and sulfate Bismuth halides BiX3 (X = Cl, Br or I) and bismuth sulfate Bi2(SO4)3 in acetone are powerful catalysts 21 of conversions of carbonyl compounds, for example, of 8, 14, 55 and 56, under the action of (CH2SH)2 at room temperature giving rise to the dithiolanes 11, 16, 57 and 58, respectively. S CHO (CH2SH)2, BiCl3 S Ph Ph 55 57 O (CH2SH)2, Bi2(SO4)3 S S H H n-C7H15 n-C7H15 58 56 These bismuth salts offer the following advantages over various Lewis and Brùnsted acids: (1) minimum amounts of these salts are required; (2) the salts exhibit catalytic activities under mild conditions; (3) the synthetic procedure is simple; (4) the products are prepared in high yields; (5) the reactions are chemoselective; (6) the catalyst is inexpensive and nontoxic; (7) it is possible to recognise the type of the carbonyl functional group.The high activity of this catalyst is retained in the presence of a small amount of water. An acid-sensitive substrate, such as furfurol (13), is readily converted into dithiolane without self- condensation and ring opening. 13. Catalysis by lithium perchlorate in diethyl ether The reactions of aldehydes 59 ± 61 with (CH2SH)2 yielding dithio- lanes 62 ± 64, respectively, are catalysed by lithium perchlorate in diethyl ether.22 Analogously, acetals 65 and 66 are converted into the dithiolanes 10 and 35, respectively.S CHO S (CH2SH)2 LiClO4±Et2O MeO MeO OMe 62 59 OMe A K Banerjee,MS Laya S CHO S (CH2SH)2 LiClO4±Et2O HC H2C HC H2C 60 OMe OMe 63 S (CH2SH)2 Me Me CHO S LiClO4±Et2O Me 64 Me 61 S OMe (CH2SH)2 S OMe LiClO4±Et2O 65 10 S OEt (CH2SH)2 OEt S LiClO4±Et2O 35 66 MeO MeO These reactions proceed at room temperature to form 1,3- dithiolanes in high yields.{ It appeared that ketones and their acetals react much more slowly than aldehydes and their acetals. Yet another advantage of the above-mentioned system is its ability to perform chemoselective thioacetalisation of oxoalde- hydes 31 and 67 ± 69 in the synthesis of dithiolanes 34 and 70 ± 72, respectively.S (CH2SH)2, LiClO4±Et2O MeCO(CH2)4CHO 67 MeCO(CH2)4 70 S S PhCO(CH2)4CHO (CH2SH)2, LiClO4±Et2O PhCO(CH2)4 71 68 S O O S (CH2SH)2, LiClO4±Et2O CHO S 69 72 In this case, the aldehyde carbonyl group is selectively protected, while the keto group remains intact. The selectivity is absent in the reactions performed with the use of such Lewis acids as BF3 or AlCl3. 14. Catalysis by ammonium cerium nitrate Some aldehydes, for example, 7, 25, 59 and 61, were converted into dithiolanes 10, 28, 62 and 64, respectively, under the action of (CH2SH)2 catalysed by (NH4)2Ce(NO3)6.24 These dithiolanes were prepared in high yields at room temperature. In the presence of ketones, only aldehydes enter into the reactions.The reaction conditions and the synthetic procedure are very simple and convenient. This procedure can be extended to ketones. For example, cyclohexanone (17) gave the dithiolane 21 upon reflux- ing in chloroform for 10 h. It should be noted that aromatic ketones, g-lactones and acyclic ketones do not react under the above-mentioned conditions, and hence, this catalyst is conven- ient for chemoselective thioacetalisation. Ammonium cerium nitrate is comparable in efficiency with the above-described solution of lithium perchlorate in diethyl ether.22 { It should be noted that the reaction of dimethyl acetal 65 with 1,2- ethanedithiol in the presence of TeCl4 in dichloromethane affords the dithiolane 10 in 99% yield.23Reagents for the preparation and cleavage of 1,3-dithiolanes 15.Catalysis by zirconium(IV) chloride adsorbed on silica gel The dithiolanes 10, 16, 21, 22, 29 and 57 are formed in high yields by the reactions of the carbonyl compounds 7, 14, 17, 18, 26 and 55, respectively, with 1,2-ethanedithiol in the presence of ZrCl4 adsorbed on silica gel.25 This method has the following advan- tages: (1) the procedure is simple and convenient; (2) the reactions proceed smoothly; (3) the reactions are rapid and the system can be used in the case of low-reactivity ketones, such as 26 and 14. Thioacetalisation of a,b-unsaturated aldehydes, for example, of cinnamaldehyde (55), can be performed with excellent yield, but attempts to prepare dithiolanes from a,b-unsaturated ketones in good yields were unsuccessful. 16.Catalysis by natural kaolin Recently, it has been demonstrated 26 that kaolin efficiently catalyses conversions of the carbonyl group into the dithiolane moiety, for example, in the reactions of the compounds 7, 13, 26 and 55 with (CH2SH)2 giving rise to the 1,3-dithiolanes 10, 15, 29 and 57, respectively. It was also noted that the reaction of a mixture of benzaldehyde (7) and acetophenone (14) with 1,2- ethanedithiol in the presence of 10% of kaolin afforded only the dithiolane 10, while acetophenone remained unconsumed. Owing to high chemoselectivity, the reactions catalysed by kaolin may be advantageous in the selective protection of alde- hydes in the presence of other carbonyl compounds.17. Catalysis by LiBr Aldehydes, for example, 7 and 55, are converted into the dithio- lanes 10 and 57, respectively, in high yields under the action of (CH2SH)2 without a solvent in the presence of catalytic amounts of lithium bromide.27 Under similar conditions, the aromatic and aliphatic ketones 14 and 61 do not react even over several hours. This procedure is unsuitable for the preparation of dithiolanes in such solvents as THF and dichloromethane. In the latter solvents, aldehydes remain unconsumed. Due to the neutrality of the reaction medium, the procedure under consideration is very useful in the case of substrates highly sensitive to acids. The excellent selectivity and simplicity allow one to use this method successfully for selective conversions of aromatic and a,b-unsaturated alde- hydes containing other carbonyl groups into 1,3-dithiolanes under very mild conditions.18. Catalysis by copper(II) trifluoromethanesulfonate adsorbed on silica gel Alarge number of aldehydes and ketones, for example, 7, 8, 14, 26, 32 and 43, were converted into 1,3-dithiolanes by reactions with (CH2SH)2 in the absence of a solvent and in the presence of catalytic amounts of copper(II) trifluoromethanesulfonate Cu(OTf)2 adsorbed on silica gel.28 Analogous conversions are catalysed by Cu(OTf)2 in dichloromethane. Aldehydes react more rapidly than ketones. The reactions with aldehydes and ketones proceed chemoselectively, viz., primarily aldehydes are involved in the reactions.In most cases, the reactions proceed smoothly. Such a sterically hindered and low-reactivity ketone as benzophenone (26) is inert with respect to the system under consideration without a solvent at room temperature, but the reaction proceeds in toluene on heating to 80 8C. Conversions of some aldehydes and ketones into 1,3-dithiolanes without a solvent were also observed in the presence of catalytic amounts of copper chloride adsorbed on silica gel, but the reactions proceeded slowly with low yields. The procedure for thioacetalisation of carbonyl compounds with- out a solvent may find wide use in organic synthesis in the future. 19. Catalysis by dibutyl(ethylenedithio)stannane in the presence of Bu2Sn(OTf)2 Sato et al.29 reported the conversions of the ketones 17 and 73 ± 75 into the dithiolanes 21 and 76 ± 78, respectively, upon treatment 951 with dibutyl(ethylenedithio)stannane Bu2Sn(SCH2)2 in the pres- ence of Bu2Sn(OTf)2 in dichloromethane.S Bu2Sn(SCH2)2±Bu2Sn(OTf)2 O CH2Cl2 S 73 76 O S S Bu2Sn(SCH2)2±Bu2Sn(OTf)2 Ph Ph Me Me CH2Cl2 77 74 Bu2Sn(SCH2)2±Bu2Sn(OTf)2 S CH2Cl2 O 75 78 S This procedure for the synthesis of dithiolanes is to some extent comparable with that proposed by Evans et al.30 The latter procedure involves thiosilanes and has been used only for car- bonyl compounds. An important advantage of the procedure under consideration is the simplicity of the synthesis of thiostan- nane from various organotin compounds.In addition, thiostan- nane is air- and moisture-stable and can be stored in air. The procedure with the use of thiostannane and organotin triflate may be useful in preparing dithiolanes not only from ketones but also from their ketals. Organotin triflates can be used for separation of ketones (or their ketals). If the reaction mixtures contain cyclopentanone (73), heptan-2-one (79) or 2-methylcyclohexanone (80) along with cyclohexanone (17), the reactions proceed primarily with cyclo- hexanone. The yield of the dithiolane 21 is very high compared to the yields of the compounds 76, 81 and 82. Bu2Sn(CH2S)2±Bu2Sn(OTf)2 O O+ 73 17 S S + S 76 (6%) S 21 (71%) O S S Bu2Sn(CH2S)2±Bu2Sn(OTf)2 17+ Me Me C5H11-n C5H11-n 21+ (89%) 79 81 (2%)S Bu2Sn(CH2S)2±Bu2Sn(OTf)2 O 17+ 21+ (73%) Me 80 S Me 82 (12%) This high selectivity cannot be achieved by conventional methods, for example, with the use of (CH2SH)2 and BF3 .OEt2. This possibility of separation is of great importance in the chemistry of steroids. III. Reagents for the cleavage of 1,3-dithiolanes A large number of procedures were developed for conversions of dithiolanes into the initial carbonyl compounds.1, 31 Some of them require drastic conditions or the use of toxic compounds, for example, mercury salts. In some cases, dithiolanes are decom- posed under mild conditions in several steps. In this connection, a demand arose for the development of new procedures for depro- tection of dithiolanes and construction of new reagents which do not affect other functional groups.952 1.The phenyl phosphorodichloridate ± sodium iodide system in DMF Liu and Wiszniewski 32 used phenyl phosphorodichloridate in combination with sodium iodide in DMF for conversions of dithiolanes, for example, of 57 and 83 ± 85, into the corresponding carbonyl compounds 55 and 86 ± 88. O S S H H PhOPOCl2 ±NaI, DMF 55 57 O PhOPOCl2 ±NaI, DMF S S C9H19-n Me C9H19-n Me 86 83 S PhOPOCl2 ±NaI, DMF O S 87 84 O S S PhOPOCl2 ±NaI, DMF 88 85 Ketones are formed in high yields. The reactions can be performed at room temperature. In the reactions with the use of acetonitrile instead of DMF, dithiolanes are recovered. The reactions carried out with boiling afford mixtures of products.This procedure 32 is applicable to aliphatic, aromatic and allylic dithiolanes. In all cases, the reactions proceed smoothly yielding the only desired product, which was confirmed by TLC data. 2. Phosphorus tetraiodide in acetic anhydride Shigemesa et al.33 developed a procedure for decomposition of 1,3-dithiolanes, for example, of 89 ± 92, under the action of P2I4 in acetic anhydride, chloroform or CCl4 giving rise to ketones 93 ± 96, respectively, in high yields. O S S P2I4, Ac2O R2 R2 93 ± 96 R1 89 ± 92 R1 R2 R1 Ketone Dithiolane 93 94 95 96 89 90 91 92 Ph Ph Ph Bun MeO Me Cl MeO The reactions of these 1,3-dithiolanes with P2I4 in dichloro- methane result primarily in reductive desulfurisation and partial deprotection.In all cases, the amount of the reduced product is substantially higher than the amount of the ketone formed. 3. Amberlyst-15 ± paraformaldehyde in wet acetone The dithiolanes 10, 22, 35, 97 and 98 were converted into the ketones 7, 18, 33, 99 and 100, respectively, under the action of Amberlyst-15 and paraformaldehyde.34 A K Banerjee,MS Laya CO2Me CO2Me Me Me Amberlyst-15, (CH2O)n Me2CO±H2O Me MeS Me O Me S 99 97 Amberlyst-15, (CH2O)n O S S Me2CO±H2O Me Me CO2Me CO2Me 98 100 The dithiolanes were dissolved in aqueous acetone. Then paraformaldehyde and Amberlyst-15 were added and the reaction mixture was heated at 80 8C for 15 ± 20 h.In these reactions, acid- sensitive ethers and esters remain intact. 4. Trimethylsilyl trifluromethanesulfonate ± 4-nitro- benzaldehyde in CH2Cl2 A new procedure for the cleavage of dithiolanes 35 involves the addition of CF3SO3SiMe3 to a mixture of dithiolane and 4-nitro- benzaldehyde in dichloromethane at room temperature. This procedure was used for conversions of the dithiolanes 16, 30, 35 and 57 into the ketones and aldehydes 14, 27, 32 and 55, respectively. The reactions can be performed with the use of CF3SO3SiButMe2 instead of CF3SO3SiMe3.Of the four aldehydes tested in the study,35 4-nitrobenzaldehyde was chosen because this reagent made it possible to accelerate the reaction and to improve the efficiency of migration of the sulfur.This method can also be used for selective deprotection of dithiolanes prepared from ketones in the presence of dithiolanes obtained from aldehydes. Thus 1,3-dithiolane 81 was hydrolysed to give ketone 79 in the presence of compound 101. This is the first example of catalytic transthioacetalisation with simultaneous migration of oxygen and alkanethiol proceeding irreversibly in an anhydrous medium. CF3SO3SiMe3 ± 4-NO2C6H4CHO + S S S S CH2Cl2, N2 Me H n-C5H11 n-C5H11101 81 O + 101 (83%) Me n-C5H11 79 5. Dimethyl sulfoxide, 140 ± 160 8C In the absence of other reagents, DMSO catalyses 36 cleavage of some dithiolanes, for example, of 30, 57, 102 and 103, at high temperature giving rise to the corresponding carbonyl compounds 27, 55, 104 and 105 in high yields.DMSO O S S D Et Et Et Et 102 104DMSO O S S D C11H23-n C11H23-n n-C11H23 n-C11H23 105 103 It is known that DMSO can decompose dithiolanes in the presence of iodine 37 or tert-butyl chloride.38Reagents for the preparation and cleavage of 1,3-dithiolanes 6. Dowex-50W ± paraformaldehyde in acetone Giri and Sankar 39 used a mixture of the acidic catalyst Dowex- 50W and paraformaldehyde in acetone for the cleavage of dithiolanes of nitrogen-containing heterocyclic compounds 106 ± 108 to yield ketones 109 ± 111, respectively. S O Me Me S Dowex-50W, (CH2O)n, Me2CO N N Me Me 109 O 106 O S O Me Me S Dowex-50W, (CH2O)n, Me2CO NH NH 107 110 S O N N Dowex-50W, (CH2O)n, Me2CO S Me Me 111 108 Although the cleavage of such dithiolanes is generally per- formed with the use of mercury salts, the latter sometimes form insoluble complexes.This difficulty was overcome 40 by protection of the nitrogen atom before hydrolysis of dithiolanes with triethyloxonium tetrafluoroborate. Taking into account the sim- plicity of the procedure and the good yield of the product, we believe that this procedure 39 is superior in efficiency to all other available procedures for dethioacetalisation of nitrogen-contain- ing heterocycles. 7. N-Fluoro-2,4,6-trimethylpyridinium triflate in a 1 : 1 CH2Cl2 ±THF mixture and H2O Dithiolanes 12, 112 and 113 were decomposed to aldehydes 9, 114 and 115, respectively, with the use of N-fluoro-2,4,6-trimethylpyr- idinium triflate [C8H11NF]+[CF3SO3]7 in a 1 : 1 CH2Cl2 ±THF mixture and 5 equiv.of H2O.41 High yields of compounds 9, 114 and 115 were achieved by performing the reactions under an inert atmosphere (Ar orN2). The reactions performed in the presence of oxygen or nitrobenzene give rise to a number of unidentified by- products. The reactions in pure dichloromethane, ether, hexane or acetonitrile as the solvent afford aldehydes in low yields. In DMF, dithiolanes do not decompose. O [C8H11NF]+[CF3SO3]7 S S CH2Cl2 ± THF, H2O H H 4-NO2C6H4 4-NO2C6H4 9 12 O [C8H11NF]+[CF3SO3]7 S S CH2Cl2 ± THF, H2O 4-MeCOHNC6H4 H H 4-MeCOHNC6H4112 114 O [C8H11NF]+[CF3SO3]7 S S CH2Cl2 ± THF, H2O H MeSC6H4 MeSC6H4115 H 113 In the reactions of the title reagent with dithiolanes 116a ± c, the latter decompose to form mono- and diketones.The ratio between the monoketones 117a ± c and the diketones 118a ± c 953 prepared from the thioacetals 116a ± c decreases as the length of the carbon chain separating two dithiolane groups increases. S Me Me O O S Me [C8H11NF]+[CF3SO3]7 (CH2)n (CH2)n + (CH2)n CH2Cl2 ±THF Me Me S S O Me S S 116a ± c 117a ± c 118a ± c n Yield (%) Compound 118 117 116 ± 118 0 74 5 1 57 15 2 52 21 abc The cleavage of dithiolanes of aromatic derivatives 119 and 120 with N-fluoro-2,4,6-trimethylpyridinium triflate in the pres- ence of water also afforded predominantly monoketones. The higher selectivity of the formation of para-substituted 121 ketone compared to the meta-isomer 122 is indicative of the important role of electronic effects and agrees with the cationic nature of intermediates. Me Me S S [C8H11NF]+[CF3SO3]7 CH2Cl2 ±THF S S 119 Me Me S O S 121 Me Me S S [C8H11NF]+[CF3SO3]7 CH2Cl2 ±THF S S 120 Me Me S O S 122 Finally, it should be noted that hydrolysis of dithiolanes with a mixture of N-fluoro-2,4,6-trimethylpyridinium triflate and water was not accompanied by migration of the electrophilic fluorine atom.This fact is in contradiction with the mechanism of fluorination of organic substrates with N-fluoro-2,4,6-trimethyl- pyridinium triflate or other N-fluoropyridinium cations proposed previously.42 8.Copper(II) chloride ± silica gel in CH2Cl2 The dithiolanes 16, 40, 53 and 54 were decomposed under the action of CuCl2 . 2H2O adsorbed on silica gel under mild con- ditions to form the corresponding carbonyl compounds 14, 37, 51 and 52.22 The yields of the products obtained with the use of dichloromethane are substantially higher than those obtained with the use of THF. The reagent under consideration can be successfully used for the cleavage of dithiolanes of various aliphatic, aromatic and sterically hindered steroid compounds. The reaction conditions are very simple and convenient. 9. Selenium dioxide in acetic acid A solution of SeO2 in acetic acid proved to be an efficient reagent for the deprotection of 1,3-dithiolanes 123 ± 126 to yield the aldehydes and ketones 127 ± 130, respectively.43954 SeO2, AcOH S S H Bun 123 SeO2, AcOH S S Prn H 124 Me OBn Me S S H 125 Me S Me AcO 126 H The reactions proceed readily at room temperature giving rise to carbonyl compounds in very high yields.Widespread func- tional groups, such as ethers and esters, do not react with selenium dioxide. This procedure is very useful in the case of dithiolane derivatives of C(7)- and C(13)-steroid ketones whose cleavage generally requires drastic conditions. The presence of acetic acid is of importance because the reactions in aprotic solvents, for example, in THF, dichloromethane or toluene, proceed with very low yields. Hydrolysis in aqueous acetonitrile or acetone proceeds slowly (20 ± 30 h) and affords cleavage products in low yields.Heating of the mixture accelerates the reaction, but the yield becomes even lower because a number of undesirable products are formed. The best results of the cleavage of dithio- lanes and the highest yields of ketones regenerated are achieved with the 1,3-dithiolane to SeO2 ratio of 1 : 5. On the whole, the procedure proposed for hydrolysis of 1,3-dithiolanes is rapid and convenient, characterised by high yields and is applicable to various compounds. 10. Oxone ± wet Al2O3 Dithiolanes 30, 131 and 132 were converted into ketones 27, 133 and 134, respectively, in high yields under mild conditions with the use of Oxone (potassium hydrogen persulfate, 2KHSO5 .KHSO4 .K2SO4) and wet alumina.44 S But S131 S S MeO 132 H The carbonyl compounds were obtained in high yields by refluxing 1,3-dithiolanes in chloroform in the presence of Oxone on Al2O3, the functional groups, such as ethers and esters, and double bonds remaining intact. The procedure is simple and convenient, which is an additional advantage of this reagent for dethioacetalisation. This procedure is comparable in efficiency O H Bun 127 O Prn H 128 Me SeO2, AcOH O H S Me SeO2, AcOHAcO H Oxone ±Al2O3, H2O O Oxone ±Al2O3, H2O MeO Me OBn 129 Me O 130But 133 O134 H A K Banerjee,MS Laya with the above-described procedures for hydrolysis of dithiolanes or even superior to these methods.11. Iron(III) nitrate ±montmorillonite K10 ± hexane Hirano et al.45 developed an efficient and simple procedure for oxidative cleavage of 1,3-dithiolanes with the use of iron(III) nitrate and montmorillonite K10 in hexane at 50 8C. This proce- dure was used for conversions of the 1,3-dithiolanes 12, 57 and 89 ± 92 into the carbonyl compounds 9, 55 and 93 ± 96, respec- tively, in very high yields. In the absence of the clay, the reactions proceed slowly and afford ketones in lower yields. The advantages of this system over other supported reagents, such as iron(III) or copper(II) nitrates on montmorillonite K10 in the absence of hexane 46 or copper(II) nitrate on silica gel,47 are the high reaction rate and the fact that there is no need for pretreatment of the reagent.Therefore, this procedure is an improved modification of the above-described methods. Hexane is a better solvent from economic and environmental standpoints compared to chlori- nated hydrocarbons (CH2Cl2, CHCl3 and CCl4), which are often used in standard heterogeneous systems. On the whole, this procedure for the cleavage of 1,3-dithiolanes and regeneration of carbonyl compounds is simple and convenient and satisfies economic requirements. 12. Zirconium sulfophenylphosphonate Mixed zirconium(IV) methylphosphonate sulfophenylphospho- nate, Zr(O3PCH3)1.2(O3PC6H4SO3H)0.8, proved to be a very useful catalyst 48 for mild hydrolysis of 1,3-dithiolanes to form the corresponding carbonyl compounds. Dithiolanes are heated with a catalyst in the presence of glyoxylic acid monohydrate as an exchange reagent.This procedure was used for the preparation of the carbonyl compounds 7, 14, 37 and 133 from the dithiolanes 10, 16, 40 and 131, respectively, in very high yields. Zirconium sulfophenylphosphonate is also a very efficient catalyst of the cleavage of cyclic dithioacetal, which is resistant to a number of other reagents.49 It is noteworthy that zirconium sulfophenyl- phosphonate is comparable in activity with the above-described reagents or even superior to them. The essential advantages of this procedure are the absence of the solvent, mild conditions and the simplicity. 13. Miscellaneous procedures The major reagents and catalysts used for the deprotection of 1,3- dithiolanes were considered in the previous sections. In addition, some dithiolanes were decomposed with the use of systems based on silver nitrate,50 bis(trifluoroacetoxy)iodobenzene,51 nitrogen oxides,52 visible light 53 and tantalum chloride.54 Since these procedures were rarely used, we do not consider them in detail in the present review.IV. Conclusion The problem of protection of the carbonyl groups in aldehydes and ketones, in particular, the protection giving rise to 1,3- dithiolanes, is of great importance in organic synthesis due to which an array of procedures for the introduction of this group were developed. The most frequently used methods were consid- ered above. It should be noted that a large number of reagents for subsequent cleavage of 1,3-dithiolanes and regeneration of car- bonyl-containing compounds have been proposed in recent years.Apparently, new reagents for introduction and removal of the dithiolane group will be developed in the future. In our opinion, the dithiolane moiety should be more extensively used as a protective group. We hope the present review will help to attract the attention of chemists engaged in organic chemistry to this problem.955 Reagents for the preparation and cleavage of 1,3-dithiolanes References 50. K Nishide, K Yokata, D Nakamura, T Sumiya,M Node, M Ueda, K Fuji Tetrahedron Lett. 34 3425 (1993) 51. G Stork, K Zhao Tetrahedron Lett. 30 287 (1989) 52. G Mehta, R Uma Tetrahedron Lett. 37 1897 (1996) 53.G A Epling, Q Wang Synlett 335 (1992) 54. S Chandrasekhar, M Takhi, Y R Reddy, S Mohapatra, C R Rao, K V Reddy Tetrahedron 53 14997 (1997) 1. T W Greene Protective Groups in Organic Synthesis (New York: Wiley, 1981) p. 129 2. H J E Lowenthal, in Protective Groups in Organic Chemistry (Ed. J FWMcOmie) (New York: Plenum, 1973) p. 323 3. B S Ong, T H Chan Synth. Commun. 7 283 (1977) 4. V Kumar, S Dev Tetrahedron Lett. 24 1289 (1983) 5. R B Perni Synth. Commun. 19 2383 (1989) 6. Y Kamitori,M Hojo, R Masuda, T Kimura, T Yoshida J. Org. Chem. 51 1427 (1986) 7. J A Soderquist, E I Miranda Tetrahedron Lett. 27 6305 (1986) 8. B Ku, D Y Oh Synth. Commun. 19 433 (1989) 9. R Miranda, H Cervantes, P Joseph-Nathan Synth. Commun. 20 153 (1990) 10.M Salmon, G Penieres, R Miranda, C Alva rez J. Heterocycl. Chem. 18 1475 (1981) 11. M Salmon, R Miranda, E Angeles Synth. Commun. 16 1827 (1986) 12. A C Cano, F Delgado, A Co rdoba, C Ma rquez, C Alva rez Synth. Commun. 18 2051 (1988) 13. L Gariaschelli, G Vidari Tetrahedron Lett. 31 5815 (1990) 14. P Kumar,R S Reddy,A P Singh, B Pandey Tetrahedron Lett. 33 825 (1992) 15. D Villemin, B Labiad, M Hammadi J. Chem. Soc., Chem. Commun. 1192 (1992) 16. N B Das, A Nayak, R P Sharma J. Chem. Res. (S) 242 (1993) 17. H K Patney Tetrahedron Lett. 34 7127 (1993) 18. P K Chowdhury J. Chem. Res. (S) 124 (1993) 19. A K Maiti, K Basu, P Bhattacharyya J. Chem. Res. (S) 108 (1995) 20. A Nayak, B Nanda, N B Das, R P Sharma J. Chem. Res. (S) 100 (1994) 21. N Komatsu, M Uda, H Suzuki Synlett 984 (1995) 22. V G Saraswathy, S Sankararaman J. Org. Chem. 59 4665 (1994) 23. H Tani, K Masumoto, T Inamasu, H Suzuki Tetrahedron Lett. 32 2039 (1991) 24. P K Mandal, S C Roy Tetrahedron 51 7823 (1995) 25. H K Patney, S Margan Tetrahedron Lett. 37 4621 (1996) 26. D Ponde, H B Borate, A Sudalai, T Ravindranathan, V H Deshpande Tetrahedron Lett. 37 4605 (1996) 27. H Firouzabadi, N Iranpoor, B Karimi Synthesis 58 (1999) 28. R V Anand, P Saravanan, V K Singh Synlett 415 (1999) 29. T Sato, J Otera, H Nozaki J. Org. Chem. 58 4971 (1993) 30. D A Evans, L K Truesdale, K G Grimm, S L Nesbitt J. Am. Chem. Soc. 99 5009 (1977) 31. B-T Grlbel, D Seebach Synthesis 357 (1977) 32. H J Liu, V Wiszniewski Tetrahedron Lett. 29 5471 (1988) 33. Y Shigemesa, M Ogawa, H Sashiwa, H Saimoto Tetrahedron Lett. 30 1277 (1989) 34. R Ballini, M Petrini Synthesis 336 (1990) 35. T Ravindranathan, S P Chavan, R B Tejwani, J P Varghese J. Chem. Soc., Chem. Commun. 1750 (1991) 36. Srinivasa Rao Ch ,M Chandrasekharam, H Ila, H Junjappa Tetrahedron Lett. 33 8163 (1992) 37. J B Chattopadhyaya, A V Rama Rao Tetrahedron Lett. 3735 (1973) 38. G A Olah, A K Mehrotra, S C Narang Synthesis 151 (1982) 39. V S Giri, P J Sankar Synth. Commun. 23 1795 (1993) 40. T Oishi, H Takechi, K Kamemoto, Y Ban Tetrahedron Lett. 11 41. A S Kiselyov, L Strekowski, V V Semenov Tetrahedron 49 2151 (1974) (1993) 42. T Umemoto, S Fukami, G Tomizawa, K Harasawa, K Kawada, K Tomita J. Am. Chem. Soc. 112 8563 (1990) 43. S A Haroutounian Synthesis 39 (1995) 44. P Ceccherelli, M Curini, M C Marcotullio, F Epifano, O Rosati Synlett 767 (1996) 45. M Hirano, K Ukawa, S Yakabe, J H Clark, T Morimoto Synthesis 858 (1997) 46. A Cornelis, P Laszlo Synthesis 909 (1985) 47. J G Lee, J P Hwang Chem. Lett. 507 (1995) 48. M Curini, M C Marcotullio, E Pisani, O Rosati, U Costantino Synlett 769 (1997) 49. A B Smith III, B D Dorsey, M Visnick, T Matda,M S Malamas J. Am. Chem. Soc. 108 3110 (1986)

ISSN:0036-021X    

出版商:RSC    

年代:2000

数据来源: RSC

摘要:

Russian Chemical Reviews 69 (11) 957 ± 984 (2000) Polynuclear complexes of copper(I) halides: coordination chemistry and catalytic transformations of alkynes BMMykhalichko, O N Temkin,MG Mys'kiv Contents I. Introduction II. Complexes of copper(I) halides in solutions and in the crystalline state III. Organocopper compounds IV. Catalytic conversions of acetylenic compounds in solutions of copper(I) chloride complexes V. Conclusion Abstract. of chemistry coordination the of features Characteristic Characteristic features of the coordination chemistry of Cu( the in alkynes of conversions catalytic of mechanisms and I) and mechanisms of catalytic conversions of alkynes in the CuCl or metal alkali is (MCl system 7MCl7H2O7HC:CR CR system (MCl is alkali metal or ammonium chloride or amine hydrochloride; R=H, CH ammonium chloride or amine hydrochloride; R=H, CH2OH, CH=CH compositions the of studies on based analysed are 2, etc.) .) are analysed based on studies of the compositions and alkyne complexes, (bromide) chloride of structures and structures of copper( copper(I) chloride (bromide) complexes, alkyne p compounds polynuclear organometallic ethynyl and -complexes -complexes and ethynyl organometallic polynuclear compounds formed state.crystalline the in and solutions in system this in formed in this system in solutions and in the crystalline state. The The role alkynes of reactions various in complexes polynuclear of role of polynuclear complexes in various reactions of alkynes is is discussed. references 149 includes bibliography The discussed.The bibliography includes 149 references. I. Introduction Copper(I) complexes occupy a significant place in the chemistry of coordination compounds and in the development and evolution of homogeneous metal-complex catalysis. Copper(I) salts in quino- line were the first catalysts used in homogeneous hydrogena- tion.1, 2 Copper(I) chloride was among the first commercial homogeneous metal-complex catalysts (dimerisation of acety- lene).3, 4 Acetylenic complexes of copper(I) were prepared for the first time by Chavastelon as early as 1898 in studies of the reactions of CuCl with C2H2 in aqueous solutions of KCl.5, 6 The compositions 2KCl . 8CuCl .C2H2, KCl . 4CuCl .C2H2, KCl . . 2CuCl .C2H2 and 2CuCl .C2H2 were assigned to the crystalline adducts isolated.5 ±7 However, interest in the reactions of acety- lene with CuCl and in catalysis by copper(I) complexes has particularly increased 3, 4, 8, 9 after brilliant studies on catalysis of dimerisation and hydrochlorination of acetylene by copper(I) chloride complexes performed by Nieuwland 3, 10 and the develop- ment of the commercial synthesis of chloroprene and synthetic rubber neoprene by Carothers (Du Pont de Nemours; see the monograph 3).Thus, Nieuwland and coworkers 10 (in 1929 ± 1931) were the first to perform the catalytic synthesis of linear oligomers of C2H2 by the reaction of a concentrated solution of NH4Cl and BMMykhalichko,MG Mys'kiv Ivan Franko Lvov National University, ul.Kirilla i Mefodia 6, 79005 Lvov, Ukraine. Fax (38-032) 297 16 68. Tel. (38-032) 279 45 06. E-mail: margm@chem.franko.lviv.ua (B MMykhalichko) O N TemkinMV Lomonosov Moscow State Academy of Fine Chemical Technology, prosp. Vernadskogo 86, 117571 Moscow, Russian Federation. Fax (7-095) 434 87 11. Tel. (7-095) 434 86 41. E-mail: lbruk@dol.ru Received 3 July 2000 Uspekhi Khimii 69 (11) 1042 ± 1070 (2000); translated by T N Safonova #2000 Russian Academy of Sciences and Turpion Ltd DOI 10.1070/RC2000v069n11ABEH000609 957 958 963 975 981 CuCl in water (the molar ratio NH4Cl : CuCl :H2O=2:3:7) with acetylene. NH4Cl, CuCl, H2O 2HC:CH H2C=CHC:CH pH*4,*80 8C NH4Cl, CuCl, H2O H2C=CHC:CCH=CH2 3HC:CH *20 8C Conversions of acetylene in these systems are accompanied by a sharp decrease in pH of the medium (from*4 to 1.1 at 80 8C).Alkali metal chlorides 11 or amine hydrochlorides (hereinafter, MCl) 12 can be used instead of NH4Cl as a component of the catalytic Nieuwland system. Concentrated solutions of MCl and CuCl proved to be also very efficient catalysts of other catalytic reactions of acetylene (for example, the addition of H2O, HCl, HCN, RSH and C5H6; see Refs 1 and 13), dienes and dichloro- butenes.14 The major proposals for the mechanisms of various reactions of alkynes catalysed by copper(I) chloride complexes, including oxidation processes with the participation of the Cu(I) ± Cu(II) system, were developed in 1960 ± 1990.13, 14 Certain specific fea- tures of copper(I) halide complexes are of most interest in their coordination chemistry and catalysis, among which are the pronounced tendency to form polynuclear complexes and, con- sequently, the possibility of the occurrence of multielectron processes, the formation of polynuclear p-complexes and s-orga- nometallic compounds with m2- and m3-bridging groups, the formation of heterovalent polynuclear complexes 15 and the ability to activate alkynes, alkenes and dienes in various addition, substitution and oxidation reactions.The distinguishing feature of concentrated solutions of MCl and CuCl in water is the formation of a large number of polynuclear (CumCln)(n7m)7 complexes, which can participate to a variable extent in the same catalytic process and react with acetylene and its derivatives giving rise to intermediate p- and s-complexes with different compositions and structures and of different nature.The understanding of the mechanism of catalytic reactions depends substantially on the available information on the structures of possible intermediates. Until recently, investiga- tions of these intermediates were restricted to isolation of adducts and determination of their chemical compositions. During the last 15 years, a series of studies has been aimed at synthesising copper- containing complexes both of chloride (inorganic) and p- and s-organocopper compounds and establishing their crystal struc- tures.16 ± 19958 In this review, the data on the coordination chemistry, syn- thesis and structures of complex compounds of CuCl (CuBr) with acetylene and its derivatives and on catalysis with the participa- tion of copper(I) chloride (bromide) complexes and alkynes are generalised and systematised and the possible mechanisms of oligomerisation of acetylene and other processes catalysed by copper complexes are discussed.In this connection, particular attention will be given to catalytic reactions of acetylene and its derivatives proceeding in highly concentrated aqueous solutions of MCl and CuCl because it is in these homogeneous systems that copper(I) chloride complexes exhibit the highest catalytic activ- ity.13 Researchers from the former Soviet Union and the Commun- ity of Independent States (primarily, from Russia and Ukraine) have made a significant contribution to these investigations.In the present review, these studies are considered in detail. II. Complexes of copper(I) halides in solutions and in the crystalline state Systems of copper(I) chloride with aqueous solutions of alkali metal chlorides, ammonium chloride or amine hydrochlorides are of such great interest that studies of the state of copper in these solutions and mechanisms of their catalytic reactions, which were started late in the 19th century, are being continued. Investiga- tions of the solubility 8 and other physicochemical properties of concentrated solutions of MCl and CuCl in water and isolation of the products obtained in the reaction of acetylene with CuCl, which incorporate one, two, three or even six CuCl molecules per C2H2 molecule,4, 7, 11, 20 provided evidence that these solutions contained polynuclear complexes with (CumCln)(n7m)7.1. Physicochemical properties of concentrated MCl ± CuCl ±H2O solutions The abundant data on the solubility of CuCl in aqueous solutions of MCl have been reported in a monograph.4 The NH4Cl ± CuCl ±H2O system has been studied extensively.21, 22 The solubility of CuCl increases sharply as the concentration of NH4Cl increases and the mole fraction of water decreases (Figs 1 and 2). The [CuCl] : [NH4Cl] ratio is noticeably larger than unity at 80 and 100 8C, which is indicative of the formation of polynuclear copper(I) chloride complexes. This is also evident from calculations of the Bjerrum formation function n *(or the Sille'n function Z) based on the results of the potentiometric determination of the equilibrium concentration of the ligand ([Cl]7): n *=âCl¡äS ¡ âCl¡ä .âCuCläS [CuCl] / [NH4Cl] 1.5 1.0 2 0.5 1 30 0 10 20 Figure 1. Dependence of the [CuCl]/[NH4Cl] molar ratio on the molal concentration of NH4Cl at different temperatures.21 T (8C): (1) 25; (2) 50; (3) 80; (4) 100. composition 4 3 40 50 [NH4Cl] (m *) BMMykhalichko, O N Temkin,M G Mys'kiv H2O 1 23 CuCl NH4Cl Figure 2. Triangular phase diagram of the solubility of CuCl (mole fractions).21 T (8C): (1) 25; (2) 80; (3) 100. At the overall concentration of NH4Cl and NH4NO3 of 10 m * (m *is the molal concentration), the n *value changes from 2.80 to 1.73 as the concentration of CuCl changes from 0.5 to 10 m *and the concentration of NH4Cl changes from 1.5 to 10 m *(80 8C).The dependence of n *on [CuCl] also indicates the existence of polynuclear complexes.22 Dissolution of CuCl in solutions of NH4Cl is a slightly endothermic process (DH0 changes from 16.7 to 4.18 kJ mol71 as the concentration of NH4Cl changes from 1.5 to 15 m *) and is accompanied by an increase in the entropy (DS 0298 changes from 36 to 45 J mol71 K71).21, 23 Dissolution of CuCl is accompanied by dehydration of chloride ions resulting in an increase in pH2O and the activity of water (aH2O).23 In the range of CuCl concen- trations from 0 to 10 m *([NH4Cl]=8.84 m *), the aH2O value increases from*0.74 to 0.83 (55 8C) and 0.88 (85 8C).mCuClsolid+p Cl(H2O)¡ [CumClm+p(H2O)pq7x] p7+xH2O. q Studies of concentrated solutions of copper(I) chloride in NH4Cl by electronic spectroscopy 24 (lmax were 200, 235 and 273 ± 400 nm) demonstrated that these solutions contain two groups of complexes because the optical density of the solution increases in the region of 250 ± 290 nm and decreases in the region of 340 ± 370 nm as the concentration of NH4Cl increases. The absorption band in the region of 340 ± 370 nm is the sum of absorption bands of different polynuclear complexes. A comparison of the absorption spectra of chloride and bromide complexes in solutions in KX (5 m *, X=Cl or Br) in the regions of 273 and 300 ± 330 nm at the CuX concentrations of 0.01 and 0.4 m *demonstrated 22 that the shifts of the absorption maxima to the long-wavelength region are no larger than 5 ± 8 nm on going from Cl7 to Br7.It was assumed 24, 25 that the longer- wavelength bands (l ^ 320 nm, eBr7 ^ 150) belong to the for- bidden transitions 3d10?3d 94s1 and the bands at 200 and 273 nm belong to the transitions 3d 10?3d 94p1. 2. Complexation equilibria in the MCl ± CuCl ±H2O and NH4Cl ± CuCl ±H2O catalytic systems The equilibria of the formation of copper(I) chloride complexes in rather dilute solutions have been studied in detail.26 The concen- tration stability constants (bm;n) for the complexes (CumCln)(n7m)7 in the NaClO4 ± NaCl ± HCl ± CuCl system with an ionic strength of the solution I=5 mol litre71 and [HCl]=0.1 mol litre71 (25 8C) were determined by two methods.The potentiometric method was used for [NaCl]=0.2 ± 0.88 mol litre71 and [CuCl]40.012 mol litre71 and for [NaCl]=0.2 ± 4.268 mol litre71 and [CuCl]40.047 mol litre71. The solubility method was used for [NaCl]45.0 mol litre71 and [CuCl]41.16 mol litre71. In the latter case, polynuclear com- plexes were revealed and the b2, 4 value for Cu2Cl2¡ 4 was evaluated (Table 1). The constant pKs=6.73 was also determined by Andrews and Keefer.32Polynuclear complexes of copper(I) halides: coordination chemistry and catalytic transformations of alkynes Table 1. Stability constants (bm;n) for selected (CumCln)(n7m)7 complexes. Ref.Table 2. Stability constants (bm,n) for the reaction mCu++nCl7> (CumCln)(n7m)7 at different temperatures and molal concentrations (m *) of NH�¢ pKs log b1,2 log b1,3 log b1,4 log b2,4 4 (see Ref. 31). 25 8C 508C 808C Complex anion Temper- Ionic ature /8C strength of the solution /mol litre71 [NH�¢4 ]=6.5 [NH�¢4 ]=14.0 [NH�¢4 ]=10.0 [NH�¢4 ]=10.0 234 6.50 27, 28 29 <4.7 130.2 7.38 26 5.6 3 5.31 5.30 6.00 6.04 6.30 005.0 6.5 a 14.0 a 25 20 25 25 25 5.70 5.99 5.98 6.08 12.2 6.73 30 31 12.8 5.7 5 Note: Ks is the solubility product for CuClsolid. a The equilibrium constant bm,n=bm ;n/gm,n ; the constant sum of the concentrations (m *) of NH4Cl+NH4NO3. 7567 CuCl¡¦ CuCl2¡¦ CuCl3¡¦ Cu2Cl¡¦ Cu2Cl24 ¡¦ Cu2Cl3¡¦ Cu2Cl46 ¡¦ Cu3Cl¡¦4 Cu3Cl25 ¡¦ Cu3Cl36 ¡¦ Cu3Cl4¡¦ Cu4Cl¡¦ Cu4Cl2¡¦ Cu4Cl3¡¦ Cu4Cl48 ¡¦ Cu5Cl¡¦6 Cu5Cl27 ¡¦ Cu5Cl38 ¡¦ Cu5Cl49 ¡¦ 1.16106 9.66105 4.06105 2.061012 1.561012 1.061012 4.361011 3.661018 4.061018 2.861018 1.861018 2.061025 2.061025 5.061024 5.061024 1.061032 5.061031 3.061031 2.061031 discrepancy increases due to the low accuracy of the graphical method of solution of the Hedstro�� m equation.35 Analysis of the distribution of the complexes (Fig. 3) demon- strated that polynuclear complexes appear in noticeable amounts starting with a CuCl concentration of 3 ¡À 4 m *.The total concen- tration of the (CumCln)(n7m)7 complexes with m=2, 3 and 4 reaches*30% of the sum of the concentrations of all complexes. Based on the bm,n values (Table 2) determined at I=6.5 and with the aCu+ and aCl¡¦ values, which were measured 26 for the max- imum concentrations of NaCl and CuCl, it can be demonstrated that the mole fraction of complexes with m52 is smaller than [(CumCln)(n7m)7] /mol litre71 With the aim of establishing the compositions of copper(I) chloride complexes in solutions concentrated with respect to CuCl andNH4Cl, which were used in catalytic reactions of alkynes,3, 4, 8 the equilibria in theNH4Cl ¡À CuCl ¡ÀH2Osystem were examined at 25, 50 and 80 8C at the concentrations of CuCl varying from 0.03 to 10 m *(at 80 8C) and at different constant concentrations of the supporting electrolyte, viz., NH4Cl+NH4NO3 (6.5, 10 and 14 m *).30, 31, 33 The activities of the Cu+ (aCu+) and Cl7 (aCl¡¦) ions used in the cited studies were calculated based on the potentials of copper and silver-chloride electrodes measured in chains with transference (with respect to a standard hydrogen electrode).For the standard conditions (pH2=1 atm, aH+=1, aCu0=1, aAgCl(sol)=1, aAg0=1), these activities are directly pro- portional to the thermodynamic activities of individual Cu+ and Cl7 ions and are taken as equal to these values at ED=0 (ED is the diffusion potential). Under the conditions of ED=const,30, 31, 33 the proportionality coefficient (which is close to unity) remains constant throughout the concentration ranges of CuCl and NH4Cl, which made it possible to obtain reliable data on the compositions of the complexes.The activity coefficient of Cl7 (gCl¡¦) was found to remain constant at the constant concen- tration ofNH�¢4 ions throughout the concentration ranges of CuCl and NH4Cl. It was assumed that g(CumCln)(n7m)7 also remains constant under these conditions. The compositions of polynuclear complexes were determined and the stability constants were estimated using the Sille'n ¡À Hedstro�� m method.34, 35 When using the aCu+ and aCl¡¦ values, the stability constants bm,n differ from the thermodynamic constants bm;n by the activity coefficient gm,n. 1.2 bm;n �� bm;n . m;n g 1.0 0.8 0.6 0.4 0.20 2 Figure 3.Distribution of the (CumCln)(n7m)7 complexes in the CuCl ¡ÀNH4Cl ¡ÀH2O system depending on the concentration of CuCl (80 8C, [NH4�¢]=10 m *). (1) CuCl¡¦ (7) Cu4Cl2¡¦ 2 ; (2) Cu2Cll23 ¡¦; (4) Cu2Cl24 ¡¦; (5) Cu3Cl25 ¡¦; (6) Cu3Cl¡¦4 ; 6 ; (8) Cu4Cl¡¦5 . The bm,n values for 19 (CumCln)(n7m)7 complexes are given in Table 2. It should be taken into account that the bm,n constants for complexes with large m (4 or 5) and n (7, 8 or 9) values calculated by this method can exhibit an error of up to one order of magnitude. Hence, it can only be said with assurance that single-, double- and triple-charged anions containing from one to four copper atoms exist in solution.The discrepancy between the experimental and calculated concentrations of CuCl at all temper- atures is no higher than 10%.22 It can be seen from Table 2 that the multiply-charged complexes (n7m=4) and complexes with m=5 disappear as the temperature increases, the compositions of the complexes being virtually independent of the increase in [NH�¢4 ]. The possibility of the determination of a large number of complexes and their compositions by the Sille'n ¡À Hedstro�� m method follows primarily from the independent determination of the aCu+ and aCl¡¦ values. Calculations of the activity (and concentration) of Cl7 ions by the Hedstro�� m equation with the use of the aCu+, [CuCl]S and [Cl7]S values demonstrated that the calculated and experimental aCl¡¦ values coincide only up to the CuCl concentrations of 0.5 m *.At higher concentrations, the 959 4.86105 1.06105 2.06104 1.361011 5.061010 2.861010 1.061010 3.061016 3.061016 7.061015 76.061021 5.061021 7 2.06106 1.26106 5.06105 7.561012 6.461012 5.561012 1.561012 1.761019 2.361019 1.761019 1.561019 2.461026 1.261026 8.061025 7.06105 2.86105 7.36104 1.661012 7.061011 4.061011 2.061011 1.561018 1.161018 5.061017 3.061017 3.061024 1.861024 1.061024 7 7 7 7 7 8.061032 8.061032 7 1 23 45 67 8 8 6 4 [CuCl] (m *)960 Table 3. Enthalpy of the reaction mCu++n Cl7>(CumCln)(n7m)7. (CumCln)(n7m)7 (CumCln)(n7m)7 298 298 DH /kJ mol71 DH /kJ mol71 2Cl46 ¡ 2 3Cl¡43 3Cl25 ¡ 4 3Cl36 ¡ 770.64 789.56 793.66 7111.56 717.12 Cu 736.32 Cu 748.52 Cu 755.64 Cu 4 4Cl26 ¡ 7139.610 7147.110 766.94 Cu4 Cl¡5 773.64 Cu CuCl¡ CuCl2¡ CuCl3¡ Cu2Cl¡3 Cu2Cl2¡ Cu2Cl35 ¡ Note.[NH4+]=10 m *. 3 (see Table 1). 5%. Because of this, these complexes were not determined experimentally.26 Interestingly, the different approaches gave close values of the stability constants bm;n and bm,n for CuCl¡2and CuCl2¡ Since the activity coefficients of different chlorides (KCl, NaCl and NH4Cl) in aqueous solutions are virtually independent of the temperature, it was assumed that dgm,n/dT ^ 0. This assumption was used for the estimation of the DH values of the reactions 33 (Table 3) mCu++n Cl7 (CumCln)(n7m)7. 4 ]=10 m *were determined by the linear interpolation on the kinetics of the reactions catalysed by copper(I) complexes.The bm,n values change in parallel with the temperature (the dependence of logbm,n on 1/T is linear).33 The bm,n values at 25 8C and at [NHá of the values given in Table 2. The addition of each successive CuCl molecule to the (CumCln)(n7m)7 complex causes a decrease in the DH value, on the average, by 30 ± 42 kJ mol71. According to the estimates obtained, the formation of polynuclear complexes from smaller complexes is also accompanied by a decrease in the enthalpy (Table 4). However, the equilibria are noticeably shifted to the right at rather low temperatures (below 50 ± 60 8C) due to a large decrease in the entropy.An increase in the bm,n values as the concentration of the cations increases may be associated both with the decrease in gm,n and with the formation of outer-sphere complexes with the NHá4 ion. The effect of the anionic complexes with n ±m52 on the ionic strength of the solution was assessed 13 with the use of the Moiseev ± Flid equation,36, 37 which describes the dependence of the acidity function of HCl (7H0) on the ionic strength of the solution (I). 7H0=log cH3O++LI7log aH2O , where L is a constant dependent on the nature of the salt and the solvent (L=0.051 m *71:0.063 mol71 litre at 40 ± 80 8C and [HCl]=0.18 ± 0.8 m *). The measurements demonstrated 38 that the H0 value remained unchanged over a wide range of CuCl concentrations at [HCl]=0.16 m *(80 8C) (Fig.4).13 Taking into account the change in the aH2O value, the ionic strength increases by no more than 0.5 for [NH4Cl]=12 m *on going from the NH4Cl ± HCl ±H2O system to the NH4Cl ± CuCl ± HCl ±H2O system. Hence it follows that multiply-charged anions have virtually no effect on the ionic strength due to their low concen- trations (calculated from the data in Table 2) as well as due to Table 4. Thermodynamic parameters for the association reactions. Reactions 298 298 DS /J mol71 K71 DH /kJ mol71 CuCl¡2 +CuCl¡2 CuCl¡2 +Cu2Cl¡3 CuCl¡2 +Cu3Cl¡4 7100.3 762.7 7121.2 7112.8 731.8 720.9 740.5 735.9 Cu2Cl24 ¡ Cu3Cl25 ¡ Cu4Cl26 ¡ Cu4Cl26 ¡ Cu2Cl¡3 +Cu2Cl¡3 BMMykhalichko, O N Temkin,M G Mys'kiv H0 0.4 1 2 0.2 3 0 8 6 4 2 0 10 [CuCl] (m *) Figure 4.Effect of the concentration of CuCl on the acidity function H0 at 80 8C and at different concentrations of NH4Cl ([HCl] = 0.16 m *).13 [NH4Cl]: (1) 5 m *; (2) 9 m *; (3) 12 m *. their interactions withNHá4 ions. Therefore, the bm,n value (taking into account the coordination of water) should be represented by the following equation:31 m;n a p bm,n= a q 2O. H NHá4m;n;p;q q p XX b g The presence of various polynuclear complexes and their participation in catalytic reactions were confirmed in subsequent studies of ethynyl complexes of copper in solution, research on the synthesis and isolation of crystalline complexes and investigations 3.Crystal chemistry of (CumXn)(n7m)7 anionic complexes Analysis of the available information on the compositions and crystal structures of complexes isolated from copper halide systems is necessary for an understanding of the mechanisms of catalytic processes and reasons for high catalytic activities of concentrated solutions containing polynuclear copper complexes. In this section, we analyse the structures of the (CumXn)(n7m)7 anions and consider the effect of the outer-sphere M+ cation on the structures of the complexes. All types of (CumXn)(n7m)7 anions found in the structures of copper(I) halide complexes can be divided into four groups, viz., island, chain (including ribbon), layered and framework struc- tures.The structural classification of the (CumXn)(n7m)7 anions is given in Table 5. Crystal-chemical analysis of the complexes demonstrated that the copper atoms in the (CumXn)(n7m)7 anionic fragments (X=Cl or Br) are most often located in a tetrahedral environ- ment (the coordination number is 4) formed by the Cl or Br atoms. The vertices of the tetrahedra are generally located at different distances from the central atom. The X7Cu bond lengths depend on the coordination number of the X atom, i.e., on the degree of association of the Cu(I) coordination polyhedra. On the whole, the X atoms characterised by larger coordination numbers are located at larger distances from the central atom (see Table 5).4 3 Planar-trigonal coordination (the coordination number is 3) about the Cu atoms was also found in copper(I) halide complexes. The angular or linear environments (the coordination number is 2) occur more rarely. Among the structurally characterised copper(I) halide complexes, no isolated tetrahedral CuX3¡ anions were found. Apparently, discrete CuX2¡ anions containing the metal atom in a planar-trigonal environment also do not exist. In the (CumXn)(n7m)7 anions containing two or more Cu atoms, the coordination polyhedra can be linked to each other in different fashions, viz., through vertices, edges or faces. The coordination polyhedra are more often linked in a combined mode simultaneously through vertices and edges, which in some cases gives rise to (CumXn)(n7m)7 anions with zeolite-like struc- tures.The numeies of the latter structures can be occupied by small molecules (HC:CH, H2O, CuC:CH, etc.) and even complexes [for example, the Cu(II) complexes in com-Table 5. Structural characteristics of the (CumXn)(n7m)7 anions in copper(I) halide complexes. Compound (PyH)CuCl2 (1) KCuBr2 . 1/2H2O (2a) b NH4CuBr2 . 1/2H2O (2b) b K2CuCl3 (3a) K2CuBr3 (3b) b (NH4)2CuBr3 (3c) b (NH4)2CuCl3 (4) CsCu2Cl3 (5 2X¡3 chain (II) ) Cu (ImH)Cu2Cl3 (6) c Cs3Cu2Cl5 (7 2X3¡ 5 chain ) Cu (Me4N)3Cu2Br5 (8) b (NH4)2Cu3Cl5 . 1/3H2O. 1/15 CuCl (9a 3X2¡ 5 ) Cu framework (NH4)2Cu3Cl5 . 4/9H2O. 1/9C2H2 (9b) d (NH4)2Cu3Cl5 .4/9H2O. 4/9C2H2 (9c) d Rb9Cu13Cl22 . (H3O+Cl7) (9d 3X2¡ 5 " ) Cu NH4Cu4Cl3(I,Cl)2 (10 4X¡5 chain ) Cu (C6H5NH3)3Cu4Cl7 (11 4X3¡ 7 " ) Cu (PyH)2Cu5Cl7 . 1/5H2O (12 5X2¡ 7 ) Cu framework Rb4Cu5Cl9 (13 5X4¡ 9 chain (ribbon) ) Cu (MeEt3N)3Cu6Br9 (14) b Rb11Cu15Cl16Br6(CuIICl6) .CuC:CH (15a 15 ) C X u 9¡ 24 Rb11Cu15Cl16Br6(CuIICl6) (15b 15X9¡ 24 " ) Cu {Rb11[Cu2(H2O)2]}Cu15Cl24(CuIICl6) (16 15X9 ) C ¡ 24 u Notes. For the Cu7Cl bonds, the first and second figures correspond to the minimum and maximum values, respectively; for the Cu7Cu bonds, the shortest distances are given. The complexes 2a and 2b, the complexes 3a ± c and 4, the complexes 9a ± d and the complexes 15a,b and 16 are isostructural. a The distances from the Cu atom to the nonbridging Cl atom; b the Cu7Br bond lengths are given; cImH is the imidazolium cation; d acetylene is adsorbed and does not occupy particular crystallographic positions.Structural type Anion CuX¡2 CuX¡2 CuX¡2 CuX2¡ 3 CuX2¡ 3 CuX2¡ 3 CuX2¡ 3 chain layered "chain """ Cu2X¡3 chain (I) Cu2X3¡ 5 island Cu3X2¡ 5 " Cu3X2¡ 5 " Cu6X3¡ 9 island layered " Coordination number of Cu 4, 3 44444444, 3 434, 3+1 4, 3+1 4, 3+1 4, 3+1 44, 3 4, 3 44, 2 4, 2 4, 2 4, 2 Interatomic distances /A Cl7Cu a m2-Cl7Cu 2.269(4), 2.269(4) 2.353(4), 2.433(4) 2.44(1), 2.545(8) 2.44(1), 2.545(8) 2.44(2), 2.48(2) 2.34(2), 2.35(2) 2.44(2), 2.48(2) 2.34(2), 2.35(2) 2.44(2), 2.48(2) 2.34(2), 2.35(2) 2.44(2), 2.48(2) 2.34(2), 2.35(2) 2.37(1), 2.37(1) 2.283(4), 2.366(4) 2.375(3), 2.513(4) 2.262(5), 2.262(5) 2.40(1), 2.40(1) 2.38(1), 2.38(1) 2.25(1), 2.30(1) 2.25(1), 2.30(1) 2.25(1), 2.30(1) 2.25(1), 2.30(1) 2.29(2), 2.39(2) 2.549(7), 2.549(7) 2.240(8), 2.41(1) 2.24(2), 2.64(2) 2.31(2), 2.31(2) 2.196(5), 2.578(5) 2.299(6), 2.652(5) 2.386(3), 2.528(2) 2.287(7), 2.340(7) 2.305(7), 2.63(1) 2.287(7), 2.340(7) 2.305(7), 2.63(1) 2.287(7), 2.340(7) 2.305(7), 2.63(1) m3-Cl7Cu 2.30(1), 2.71(1) 2.30(1), 2.71(1) 2.30(1), 2.71(1) 2.30(1), 2.71(1) 2.56(1), 2.631(8) 2.48(2), 2.63(2) Ref.Cu_Cu m4-Cl7Cu 2.803(5) 39 40 18 41 41 2.891(8) 2.891(8) 4.08(1) 4.08(1) 42, 43 4.08(1) 4.08(1) 43 44 45 46 47 48 49 49 48 50 51 18 52 53 54 54 55 2.81(1) 2.45(1), 2.45(1) 2.748(3) 2.911(8) 2.34(1), 2.881(8) 72.98(1) 2.40(1), 2.68(1) 2.98(1) 2.40(1), 2.68(1) 2.98(1) 2.40(1), 2.68(1) 2.98(1) 2.40(1), 2.68(1) 2.48(2) 2.60(2), 2.74(2) 2.801(9) 2.74(2) 3.092(4) 2.650(3) 2.697(3), 2.697(3) 3.006(6) 3.006(6) 3.006(6)962 pounds 15a,b and 16; see Table 5] to form `guest ± host' systems.The fact that the coordination polyhedra about the Cu atoms are linked via edges (faces) results in short Cu . . . Cu contacts typical of cluster compounds (see Table 5). Let us consider the types of the (CumXn)(n7m)7 anions found in the structures of copper(I) halide complexes in detail. Two types of mononuclear anions, viz., CuX¡2 and CuX23 ¡, are known.The CuX¡2 anion serves as an independent unit of the (CuX¡2 )n polymeric anions, which exist as chains or layers in complexes 1 and 2a,b, respectively (see Table 5). In both case, the coordination polyhedra about the Cu atoms are linked to each other both through vertices and edges. In isostructural compounds 3a (or 3b,c) and 4, the (CuX2¡ 3 )n 3 chain fragments of identical structure are formed based on the CuX2¡ anion. In these compounds, the coordination tetrahedra about the Cu(I) atoms are linked only through vertices. 3 The binuclear Cu2X¡ anion can produce two types of (Cu2Cl¡3 )n chain fragments. Fragments of the first type are observed in complex 6 and are formed by the Cu(I) coordination tetrahedra and triangles linked to each other simultaneously through vertices and edges.The second type (complex 5) is formed by the Cu(I) coordination tetrahedra linked to each other exclu- sively through edges. 5 5 5 )n For yet another binuclear Cu2X3¡ anion, both the island and chain modes of arrangement of the anionic fragments are observed. Complex 8 contains the Cu2Br3¡ anion of the island type. In this complex, two coordination triangles about the Cu(I) atoms are linked via a shared vertex. In complex 7, the (Cu2Cl3¡ chain anion consists only of the Cu(I) coordination tetrahedra, which are linked to each other both through vertices and edges. The trinuclear Cu3X2¡ anion was found in isotypic com- 5 pounds 9a ± c with the zeolite-like framework structures formed by the Cu(I) coordination polyhedra linked to each other both through vertices and edges.The (Cu3Cl2¡ 5 )n framework anion was also found in complex 9d whose composition can be represented as Rb2Cu37xCl57x . (2/9H3O+Cl7) (x=0.1) by analogy with the compounds 9a ± c. Crystals of these four compounds were isolated from the catalytic MCl ± CuCl ±H2O system (M=NH4+ or Rb+) in the presence of which oligomerisation of C2H2 was performed. The compounds 9a ± d can be considered as stable stoichiometric points in the region of a phase with the variable composition M2Cu3Cl5 . x . (yC2H2), where x=3/9H2O. . 1/15CuCl and (2/9H3O+Cl7) at y=0 for M=NHá4 and Rb+, respectively, and x=4/9H2O at y=1/9 or 4/9 forM=NHá4 .The tetranuclear Cu4X¡5 anion forms an unusual [from the 7 viewpoint of the mode in which the coordination polyhedra about the Cu(I) atoms are linked] (Cu4X¡5 )n chain fragment. In complex 10, these fragments exist as the Cu(I) coordination tetrahedra linked to each other via edges and faces. The tetranuclear Cu4X3¡ anion is observed in complex 11. These anions form (Cu4Cl3¡ 7 )n chains in which the coordination tetrahedra and triangles about Cu(I) are linked to each other through vertices and edges. In the structurally characterised copper(I) halide complexes, two types of pentanuclear anions, viz., Cu5X2¡ 7 and Cu5X49 ¡, were found. In complex 12, the (Cu5Cl2¡ 7 )n anions are built of the coordination tetrahedra and triangles about the Cu(I) atoms, which are successively linked to each other through vertices and edges to form a zeolite-like framework structure.9 The Cu5X4¡ anions in complex 13 are linked in a ribbon-like 9 )n fragment in which the Cu(I) coordination polyhedra 9 (Cu5X4¡ are linked to each other both through vertices and edges. The island Cu6Br3¡ fragment is an example of a hexanuclear anion. In complex 14, this anion consists of four Cu(I) coordina- tion tetrahedra linked to each other exclusively through edges. Two opposite vertices of this fragment are additionally coordi- nated by two Cu atoms with the coordination number 2. Finally, the polynuclear Cu15X9¡ 24 anion was found in the zeolite-like structure 16. The anions form a [Cu15Cl24(CuIICl6)]n layer in which the coordination polyhedra about Cu(I) are linked BMMykhalichko, O N Temkin,M G Mys'kiv (CumCln)(n7m)7 Cu15Cl924¡ Cu5Cl49 ¡ Cu5Cl27 ¡ Cu4Cl37 ¡ Cu4Cl¡5 Cu3Cl25 ¡ Cu2Cl35 ¡ II-Cu2Cl¡3 I-Cu2Cl¡3 CuCl2¡ 3 II-CuCl¡2 I-CuCl¡2 1 2 3 4 Figure 5.Relationship between the degree of association of the coordi- nation polyhedra about the metal atoms and the number of nuclei (m) for the (CumCln)(n7m)7 anion; (1) terminal Cl; (2), (3) and (4) m2-, m3- and m4-Cl, respectively. to each other through vertices and edges. The `guest ± host'-type structures of 15a,b contain the structurally similar [Cu15Cl16Br6(CuIICl6)]n anionic layers in which some crystallo- graphic positions are partially occupied by Cl atoms.For a range of the above-described structural fragments in copper(I) chloride complexes, a marked trend is observed toward their complication as the number of nuclei in the (CumCln)(n7m)7 anions increases. The relationship between the number of nuclei in the (CumCln)(n7m)7 fragment (m) and the character of the bridg- ing Cl atom (mn, where n=2, 3 or 4) can be represented by a diagram (Fig. 5). It can be readily seen from Fig. 5 that an increase in m is accompanied by an increase in the ability of the chlorine atoms to act as bridging ligands. Thus, the very low degree of association of the Cu(I) coordination polyhedra is characteristic of the anionic fragments at m=1, whereas complex structural forms of anions appear at m=2, 3 or 4.For m54, though the ability of the bridging chlorine atoms to undergo association is somewhat lower, the structural fragments are, on the whole, complex. This regularity agrees well with the results of studies of the equilibria in the CuCl ± MCl system.13, 30, 31, 33 Thus among the (CumCln)(n7m)7 complexes (m=1±5 and n=2 ± 7), which are present in concentrated solutions of NH4Cl and CuCl, the com- plexes with m=2, 3 and 4 prevail. Therefore, the ability of the Cu(I) coordination polyhedra to form complex (CumCln)(n7m)7 associates at m52, 3 or 4 is, apparently, one of the major reasons for the high catalytic activity of concentrated solutions of MCl and CuCl. Complex associates in solution are formed from coordination polyhedra of layered, chain, island or framework types upon cleavage of the bridging mn-Cl bonds by water molecules or Cl7 anions, which occupy the coordination vacancies in (CumCln)(n7m)7 complexes in solutions.Let us consider the effect of the nature of the outer-sphereM+ cation involved in the MCl ± CuCl ±H2O system on the crystal structures of the anionic copper(I) halide complexes. As mentioned above, dissolution of CuCl and formation of anionic complexes in aqueous solutions of MCl proceed according to the reaction mCuClsolid+pCl(H2O)¡ [CumClm+p(H2O)pq7x] p7+xH2O. qPolynuclear complexes of copper(I) halides: coordination chemistry and catalytic transformations of alkynes It is known that copper chloride is a salt, which contains covalent polar Cu7Cl bonds and is characterised by a framework sphalerite-like structure.56 Dissolution of CuCl in aqueous solu- tions of ionic salts MCl involves destruction of the crystal lattice of CuCl under the action of adsorbed Cl7 anions with the simulta- neous cleavage of some Cu7Cl bonds and the appearance of individual (CumCln)(n7m)7 fragments resulting in the establish- ment of the equilibria between these fragments.Both processes are accompanied by dehydration of Cl7 ions present in the solution. Apparently, the structures of the anions (as well as the stoichiom- etry of the equilibria) are determined to a large extent by the nature of theM+ cations. The crystal structures of the complexes isolated from solutions also depend on the type (the size, symme- try, configuration, charge density and other characteristics) of the outer-sphere cation involved in complexation. Analysis of the structural data demonstrated that NHá cations can form crystalline complexes with extremely complex zeolite-like structures, such as the complexes 9a ± c, due to the tetrahedral form of the ammonium cations. The latter interact with Cl7 ions through N(H) .. . Cl hydrogen bonds 57 imparting the directed character to the M+. . . Cl7 ionic bonding, which is responsible for the `code' of the structure construction of the complex as a whole. The skeleton structures 1, 6, 11 and 12 containing planar PyH+, ImH+ (Im is imidazole) and C6H5NHá cations are also formed through (N)H .. . Cl hydrogen bonds. The replacement of NH4Cl by KCl gives rise to the complexes 2a, 3a and 3b in which K+ cations cannot form directed bonds with Cl7 ions, and, consequently, no complex combinations of the Cu(I) coordination polyhedra are observed. Apparently, the ability of NHá cations and protonated amines to generate complex structural forms of (CumCln)(n7m)7 anions through spatially oriented M+. . . Cl7 interactions giving rise to outer-sphere complexes in solution is one of the reasons for the fact that catalytic activities in systems containing NHá4 and amine hydrochlorides are higher than those in the KCl ± CuCl ±H2O systems. Table 6. Mode of coordination of the C:C group and the geometric characteristics of the p-coordinated Cu(I) units in the structures of p-complexes of copper(I) halides with acetylene and its monosubstituted derivatives.Complex Cu2Cl2(HC:CH) (17) Cu3Cl3(HC:CH) (18) Cu6Cl6(HC:CH) (19) b KCu3Cl4(HC:CH) (20) NH4Cu3Cl4(HC:CH) (21a) RbCu3Cl4(HC:CH) (21b) 43 4 Mode of coordination of the C:C /A base of the poly- length /A /A /A group m-Z2,Z2 m-Z2,Z2 m-Z2,Z2 m-Z2,Z2 m-Z2,Z2 m-Z2,Z2 III. Organocopper compounds In the previous section, we considered M(n7m)CumXn complexes, which were formed in concentrated aqueous solutions of MCl and CuCl and isolated from solutions in the crystalline state. The compositions and structures of these complexes give an idea of the state of the initial catalytic system. In the course of the reaction of the MCl ± CuCl ±H2O system with acetylene (for example, under conditions of the oligomerisation according to Nieuwland), acetylene molecules and then products of their conversions (mono- and disubstituted derivatives) form p- and s-organo- copper compounds with (CumCln)(n7m)7 anions.4, 13 In the first stage, p-complexes form.Then these complexes are converted into compounds of the mono- and diacetylenide types depending on pH of the medium.58 In the initially weakly acidic medium, organometallic compounds exist predominantly as ace- tylenide (ethynyl) derivatives, which can be reversibly converted into p-complexes as pH decreases. The latter complexes are the major products of the reactions of alkynes with Cu(I) in acidic solutions.In this section, we will consider p- and s-organocopper complexes, which are formed in the course of the reactions of acetylenic compounds with concentrated aqueous solutions of MCl and CuCl. Analysis of the structural features of polynuclear ethynyl complexes of copper(I) is required for an understanding of their role in catalytic processes (provided that the structures of the major fragments of the crystalline compounds are retained in solution). Selected structural characteristics of the compounds under consideration are given in Tables 6 ± 9. For simplicity, the types (modes) of coordination of the C:C group (Z1 or Z2) are denoted by s or p, respectively (or s,p-mode if both modes of coordination occur). Examination of the data given in Tables 6 ± 9 allows one to judge the efficiency of p-interactions in copper(I) p-complexes, to estimate the degree of activation of the triple bond and the alkyne molecule by the metal ion, to follow the possible pathways of conversions of p-complexes into acetylenide (ethynyl) complexes Sum of LCuL angles in the Cu7C* distance hedron /deg 2.01(8) 1.76(8) 2.12(7) 1.99(9) 2.16(9) 2.29(9) 2.01(3) 2.01(3) 1.97(2) 1.97(2) 356.1 359.4 356.5 350.5 348.9 351.5 352.8 352.8 349.4 349.4 7 7359.2 359.2 359.5 359.5 359.3 359.4 359.0 359.0 359.5 359.9 358.5 358.5 1.940(4) 1.940(4) 1.947(4) 1.947(4) 1.94(1) 1.94(1) 1.93(1) 1.93(1) 2.03(3) 1.93(3) 1.89(2) 1.89(2) 963 Ref.Cu7Xax bond Difference C:C bond length in Cu7C distances 1.18(14) 20 1.26(13) 0.94(24) a 2.75(3) 3.08(4) 3.07(4) 2.67(3) 2.99(5) 2.74(4) 1.12(4) a 20 1.32(5) 0.00(8) 0.00(8) 0.11(7) 0.07(9) 0.00(9) 0.00(9) 2.663(6) 0.08(3) 2.663(6) 0.08(3) 2.596(7) 0.02(2) 2.596(7) 0.02(2) 7 7 7 17 20 1.26(1) 1.28(1) 59 1.32(2) 1.25(1) 60 1.33(3) 1.27(3) 3.177(2) 0.011(4) 3.177(2) 0.011(4) 3.462(2) 0.004(4) 3.462(2) 0.004(4) 3.664(5) 0.00(1) 3.583(5) 0.00(1) 3.256(5) 0.03(1) 3.256(5) 0.03(1) 3.674(9) 0.00(3) 3.596(9) 0.00(3) 3.307(9) 0.04(2) 3.307(9) 0.04(2)964 Table 6 (continued). Complex Cu7C* Sum of LCuL distance angles in the Mode of coordination of the C:C /A base of the poly- length /A /A /A hedron /deg group m-Z2,Z2 KCu8Cl9(HC:CH)4 .1/2HCu2Cl3 .H2O (22) NH4Cu8Cl9(HC:CH)4 . 2/5 {[Cu(H2O)2] . [CuCl2] . m-Z2,Z2 .H2O} (23) CuCl(HC:CPh) (24) CuCl(HC:CCH2Cl) (25) CuCl(HC:CCH2OH) (26) Cu2Cl2(HC:C(CH2)4C:CH) (27) Cu2Cl2(HC:CCH2OH) (28) Z2 Z2 Z2 Z2;Z2 m-Z2,Z2 m-Z2,Z2 (C6H5NH3)Cu2Cl3(HC:CCH2OH) (29) 356.7 356.7 356.9 356.9 358.8 358.6 358.7 359.7 357.5 354.7 358.7 349.1 359.8 341.3 359.8 354.8 1.948(4) 1.948(4) 1.957(5) 1.957(5) 1.941(4) 1.912(4) 1.92(1) 1.946(6) 1.94(1) 1.98(1) 1.95(1) 1.99(1) 1.91(1) 2.10(1) 1.93(1) 1.964(9) Note. Hereinafter, Cu7C* is the distance from the copper atom to the midpoint of the multiple bond.a The positions of the C atoms in the acetylene fragment were not refined because of the low stability of the complexes; b the first stage of X-ray diffraction analysis; cXax is the O atom of the alcoholic group. Table 7. Mode of coordination of the C:C group and the geometric characteristics of the p-coordinated Cu(I) units in the structures of p-complexes of copper(I) halides with alkenynes. Complex Cu7C* distance /A Sum of LCuL angles in the base of the polyhedron /deg /A /A /A Mode of coordina- tion of the C:C group m-Z2,Z2 Cu7Cl7(HC:CCH=CH2)3 (30) CuCl(CH2=CHCH2C:CCH2OH) (31) Cu2Cl2(CH2=CHCH2C:CPh) (32) Z2 Z2 Cu2Br2(CH2=CHCH2C:CPh) (33) Z2 354.2 353.8 359.5 357.1 359.1 357.5 358.4 355.5 1.98(1) 1.96(2) 1.92(2) 1.953(7) b 1.937(7) 1.85(12) b 1.94(9) 1.961(4) b Cu2Cl2(CH2=CHC:CCH=CH2) (34) Z2 Cu3Cl3(CH2=CHC:CCH=CH2) (35) Z2 Cu3Cl3(CH2=CHCH2C:CCH2CH=CH2) (36) Z2 Cu3Br3(CH2=CHCH2C:CCH2CH=CH2) (37a) Z2 (modification A) Cu3Br3(CH2=CHCH2C:CCH2CH=CH2) (37b) Z2 (modification B) 359.8 357.2 359.9 350.5 360.0 358.2 353.1 359.5 359.6 358.6 359.9 1.951(7) b 1.962(7) b 1.910(6) 1.97(1) b 1.86(1) 2.00(4) b 1.99(4) b 2.05(4) 1.99(3) b 1.98(3) b 1.92(3) a The positions of the corresponding Catoms were not refined because of the disorder of the C:Cgroups in the crystal; b the Cu(I) atom is coordinated by the C:C bond.and to analyse the characteristic structural features of polynuclear ethynyl compounds.1. Complexes of CuX with acetylene and its derivatives coordinated simultaneously by two metal atoms (Fig. 6). In this respect, the acetylene ligand is isolobal to the bridging m2-Cl ligand. Some monosubstituted acetylene derivatives in p-com- plexes with CuCl behave analogously (for example, 28 ± 30; see Tables 6 and 7). However, the C:C bond in p-complexes of a series of alkynes and alkenynes, for example, of disubstituted acetylene derivatives (like the alkene bond), is coordinated exclu- sively by one copper atom [compounds 24 ± 27, 38 ± 42 and 31 ± 37 The triple bond consists (in addition to the s-bond) of two mutually orthogonal p-bonds, which are potentially able to interact with Cu(I). In the structures of copper(I) chloride com- plexes, acetylene always acts as a bridging p-ligand and is BMMykhalichko, O N Temkin,M G Mys'kiv Ref.Cu7Xax bond Difference C:C bond length in Cu7C distances 20 1.23(1) 1.233(9) 61 1.20(1) 1.213(5) 62 1.204(5) 62 62 1.196(9) 63 64 1.26(2) 1.24(2) 1.25(2) 65 1.27(1) 2.832(2) 0.000(4) 2.832(2) 0.000(4) 2.846(4) 0.000(5) 2.846(4) 0.000(5) 3.037(1) 0.067(4) 3.102(1) 0.000(4) 3.078(4) 0.012(9) 3.134(2) 0.024(6) 2.467(9) c 0.00(1) 2.725(5) 0.02(1) 3.303(6) 0.03(1) 2.604(5) 0.04(1) 3.730(6) 0.01(1) 2.416(6) 0.08(1) 3.058(7) 0.006(9) 2.830(7) 0.027(9) Ref. Difference in Cu7C distances Cu7Xax bond length C:C bond length 66 1.26(2) 67 68 69 0.03(1) 0.02(2) 0.10(2) 0.063(7) 0.010(7) 0.03(12) 0.03(9) 0.050(4) 2.643(6) 2.647(6) 3.516(7) 2.930(2) 2.924(2) 2.63(3) 3.04(3) 2.705(1) 1.14(3) a 1.35(1) 1.23(1) 1.3(1) 1.2(1) 1.361(6) 70 70 71 71 72 1.34(1) 1.36(1) 1.225(8) 1.35(2) 1.18(2) 1.45(5) 1.36(5) 1.21(5) 1.37(4) 1.36(4) 1.18(3) 0.034(6) 0.008(6) 0.002(5) 0.07(1) 0.00(1) 0.12(4) 0.24(3) 0.07(3) 0.00(3) 0.10(3) 0.03(2) 3.147(2) 2.881(2) 3.380(2) 2.627(3) 3.742(3) 3.028(7) 2.895(6) 3.450(6) 3.36(3) 3.05(2) 3.33(3)Polynuclear complexes of copper(I) halides: coordination chemistry and catalytic transformations of alkynes Table 8. Mode of coordination of the C:C group and the geometric characteristics of the p-coordinated Cu(I) units in the structures of p-complexes of copper(I) halides with disubstituted acetylenes.Complex Sum of LCuL angles in the base of the Cu7Xax bond length Cu7C* distance /A Difference in Cu7C distances polyhedron /deg /A /A /A Mode of coordina- tion of the C:C group 1.923(3) 1.928(3) 73.183(1) 1.923(7) 7 73 40 7 7 7 40 40 7 7 7 40 74 Z2 Z2 Z2 Z2 Z2 Z2; Z2 CuCl(Me3SiC:CSiMe3) (38) CuCl(HOCH2C:CCH2OH) (39a) CuBr(HOCH2C:CCH2OH) (39b) a KCuBr2(HOCH2C:CCH2OH) (40a) KCuCl2(HOCH2C:CCH2OH) (40b) a Cu2Cl2(ButC:CSC:CBut) (41) Z2; Z2 [Cu2Cl2(ButC:CSC:CBut)]n (42) 360.0 359.3 7 7359.9 7 7360.0 359.8 359.4 358.8 359.3 359.3 3.550(3) 3.501(3) 73.01(1) 3.34(1) 7 1.886(4) 1.916(4) 1.88(2) 1.87(2) 1.91(2) 1.88(2) a The first stage of X-ray diffraction analysis.b The positions of the corresponding C atoms were not refined because of the disorder of the C:C groups in the crystal. Table 9. Mode of coordination of the C:C group and the geometric characteristics of the p-coordinated Cu(I) units in the structures of s,p-complexes of copper(I) halides. Complex Difference C:C in Cu7C bond length /A /A /A Cu7C* Sum of LCuL Mode of Cu7Xax coordination distance angles in the bond of the C:C /A base of the poly- length group hedron /deg (ImH)4Cu9Cl11(C:CH)2 (43) (PyH)4Cu9Cl12(C:CH) (44) (NH4)8Cu29Cl29(C:C)4 .7H2O (45) 2.18(3) 2.01(2) 2.01(2) 2.06(2) 2.06(2) 159.1 a 358.4 358.4 359.1 359.1 2.980(9) 2.980(9) 2.79(1) 2.79(1) 770.34(3) 0.01(2) 0.01(2) 0.06(2) 0.06(2) 7 7 71.29(2) 45 m-Z1,Z1 m-Z1,Z1 m3-Z1,Z1,Z1, m-Z2,Z2, m3-Z1,Z1,Z1 m-Z1,Z1, m3-Z2,Z2,Z2, m3-Z1,Z1,Z1 m-Z1,Z1, m-Z2,Z2 (ImH)Cu4Cl4(C:CCH2OH) .H2O (46) 2.773(5) 2.730(5) 2.01(1) 2.01(1) 0.01(1) 0.01(1) 7 7 7 18 1.29(1) 76 m-Z1,Z1, m-Z2,Z2 m-Z1,Z1, m-Z2,Z2 NH4Cu4Cl4(C:CCH2OH) .H2O (47) c (NH4)2Cu4Cl5(C:CCH2OH) .H2O (48) m-Z1,Z1, m-Z2,Z2 (C6H5NH3)2Cu4Cl5(C:CCH2OH) (49) m-Z1,Z1, m-Z2,Z2 Cs2Cu5Cl6(C:CCH2OH) (50) Rb3Cu7Cl9(C:CCH2OH) .H2O (51) 2.00(2) 2.01(2) 1.94(3) 1.98(2) m3-Z1,Z1,Z1, m-Z2,Z2 2.09(2) 1.97(2) (PyH)2Cu8Cl8(C:CCH2OH)2 .H2O (52) m-Z1,Z1, m-Z2,Z2 m-Z1,Z1, m-Z2,Z2 (C6H5NH3)3Cu8Cl10(C:CCH2OH) .2H2O (53) 1.97(1) 1.98(1) Cu25Cl17(C:CCH=CH2)8 .14H2O (55) 353.9 353.9 7 7 1.986(8) 355.9 2.005(8) 358.0 354.8 357.5 359.1 358.8 358.9 359.9 m3-Z1,Z1,Z1, m-Z2,Z2 1.974(5) 356.6 1.994(5) 355.9 1.989(5) 352.8 2.020(5) 358.8 360.0 355.8 HO(CH2)2NH3Cu9Cl8(C:CCH2OH)2 . (H2O)2 (54) m3-Z1,Z1,Z1, m-Z2,Z2 1.988(6) 356.1 1.965(6) 358.7 357.7 357.2 2.850(3) 2.979(3) 2.937(9) 2.944(0) 3.061(7) 3.23(1) 3.558(7) 3.489(6) 2.973(3) 2.754(3) 2.728(3) 2.931(3) 3.392(9) 2.858(9) 2.778(3) 2.979(3) 77 m-Z1,Z1, m-Z2,Z2 1.93(4) 2.00(3) a The coordination number of Cu is 2; b the Z2-mode of coordination of theC:Cgroup to three Cu(I) atoms; c the first stage of X-ray diffraction analysis.C:C bond to be involved in the p-interaction with copper(I) is higher than that of the C=C bond. According to the Dewar ± Chatt ± Duncanson concept,13, 80, 81 (Fig. 7)]. Moreover, the C:C bond of the alkenyne ligand (i.e., the more sterically hindered bond) successfully competes with the peripheral vinyl groups even in conjugated systems (complex 35). In the case of the CuCl complex with allylpropargyl alcohol (31), the C=C bond of the allyl group, unlike the C:C bond, is not coordinated to the copper atom. Therefore, the ability of the the formation of acetylene complexes with Cu(I) compounds (3d104s 04p0) is accompanied by interactions between the frontier molecular orbitals (MO) of acetylene and the metal complex [for 965 Ref.C:C bond length 1.227(5) 1.219(4) 0.003(3) 0.024(3) 1.24(1) 0.000(7) 74 1.232(4) 1.071(5) b 1.25(2) 1.13(2) b 1.24(2) 1.23(2) 0.029(4) 0.36(4) 0.04(2) 0.12(2) 0.02(2) 0.09(2) Ref. distances 1.21(2) 45 1.19(7) 39 75 71.27(5) b 71.23(5) 71.25(4) 77 1.22(3) 78 1.23(3) 52 1.258(6) 39 1.267(6) 1.25(2) 51 1.259(8) 79 1.21(5) 66 0.028(8) 0.028(8) 0.04(2) 0.03(3) 0.08(2) 0.03(2) 0.00(2) 0.02(2) 0.024(5) 0.009(5) 0.127(5) 0.009(5) 0.04(1) 0.02(1) 0.012(7) 0.069(6) 0.03(3) 0.03(3)966 1/2 x Figure 6. Projection of the structure of Cu2Cl2(HC:CH) (17) onto the xy plane. y x Figure 7. Projection of the structure of CuC2(HOCH2C:CH2OH) (39a) onto the xy plane. example, (CumCln)(n7m)7] giving rise to two three-centre MOs.One of them is formed by the overlap of the unoccupied sp3- hybridised orbital of Cu(I) with the bonding p-MO of acetylene (the donor-acceptor component of the M/L bond, where M is metal and L is alkyne), whereas the second MO is formed due to overlapping of one of the symmetry-suitable occupied d orbitals of the metal atom and the antibonding p*-MO of acetylene (the dative component of the M?L bond). Both three-centre bonds contribute to the increase in the strength of the Cu7p-ligand bond; however, they have different effects on the properties of the p-ligand. For example, the occupation of the antibonding p*-MO of acetylene with the electrons of the metal atom (the dative component) in Cu(I) p-complexes leads to a change in the C:C7R angle (cis distortion) and elongation of the C:C bond.In turn, donation of electrons from p-MO of acetylene to one of the sp3-hybridised orbitals of Cu(I) (the donor-acceptor component) results in a decrease in the electron density on the C:C bond due to which the carbon atoms acquire positive charges (d+) and this p-coordinated acetylene group acts as a soft electrophile. The transfer of the electron density from the ligand to Cu(I) leads also to a decrease in the distance from the Cu(I) atom to the midpoint of the C:C bond. In essence, the above-considered model of the bond in the metal (M) p-complex with alkyne is synergistic because an increase in the M/L component leads to an increase in the back M?L transfer.82 However, these components make differ- Cl(1) Cl(6) C(1) C(2) Cu(1) Cu(2) H Cu(5)Cl(4) Cl H Cu C(2) C(3) C(1) H(6) C(4) O(1) H(1) O(2) BMMykhalichko, O N Temkin,M G Mys'kiv y Cl(5) Cl(2) Cu(6) C(4) Cu(4) C(3)Cu(3) Cl(3) ent contributions to p-interactions.Calculations of the energies of the Cu(I)7acetylene bond [or the Cu(I)7ethylene bond] demon- strated 83 that the contribution of the donor-acceptor electron transfer to the bond energy is approximately three times larger than that of the dative interaction. Hence, the specificity of activation of the acetylene bond on the metal centre in Cu(I) ± p-complexes is primarily determined by interactions of theM/L type.High-level quantum-chemical calculations 84, 85 of the Cu(C2H2)+ complex (taking into account the low-lying excited states) demonstrated that the Cu+7C2H2 bond is predominantly ionic in character and the Coulomb Cu(I) ion ± p-ligand interac- tion makes the major contribution (the C:C bond length is 1.242 A, the C:CH angle is 168.7 8). This is consistent with the conclusion 83 that the electron transfer to the metal atom prevails. The ionic character of the bonds was also revealed in bisalkyne complexes Cl2X(C:CH)2 .CuMe (X=C, Si or Ti).86 In these complexes, CuMe even donates electrons to the p-ligand. The Cu(I)7(C:C) [or Cu(I)7(C=C)] bonding in copper(I) chloride p-complexes leads to deformation of the tetrahedral environment about the metal atom characteristic of Cu(I) s-com- pounds toward the trigonal pyramid and even toward the planar triangle.87, 88 Deformation of coordination polyhedra is associ- ated with the appearance of a deficiency of electron density in the dx27y2 orbital { due to its interaction with antibonding MOs of the multiple carbon7carbon bond.In contrast, the dz2 orbital, which is directed toward the axial ligand (for example, Clax), possesses residual electron density. Hence, the interaction of Cu(I) with the multiple carbon7carbon bond is accompanied both by short- ening of the bonds between the copper atom and the ligands located in the base of the pyramid and by elongation of the Cu7Lax bond, i.e., the Cu(I) atom moves toward the base of the polyhedron.The more efficient the dative component of the p-bond, the more substantial the deformation, the distance from the Cu atom to the midpoint of the multiple carbon7carbon bond (Cu7C*) being also shorter due to the synergism of the Cu(I)/L and Cu(I)?L components. Knowing the parameters of the coordination unit, such as the Cu7C* distance, the Cu7Xax bond length (Xax is the axial Cl or Br atom), the sum of the LCuL angles (PLCuL) in the base of the polyhedron, which characterises the deviation of the metal atom from the equatorial plane,{ and { The dx27y2 and dz2 orbitals in the initial tetrahedral ligand field after the removal of degeneracy are characterised by lower energies than the remaining three orbitals of de symmetry.89 { The sums of the LCuL angles equal to 329.5 8 and 360 8 correspond to two limiting cases, viz., to the ideal tetrahedron and planar triangle.Polynuclear complexes of copper(I) halides: coordination chemistry and catalytic transformations of alkynes a b Lax H2C Cu L HC L R cH Lax C Cu Cu CH L L L Figure 8.Modes of coordination in the p-complexes of CuCl and CuBr; (a) the trigonal pyramid typical of the Cu(I) atom p-coordinated by the C=C bond; (b) the planar-trigonal environment of Cu(I) in p-complexes with disubstituted acetylene derivatives; (c) the bridging p-coordination of the C:C bond to the Cu atoms in p-complexes of C2H2 and its monosubstituted derivatives.the length of the p-coordinated multiple carbon7carbon bond, one can assess the state of the C:C (or C=C) group p-activated on the metal centre and the efficiency of the Cu(I)7(C:C) [Cu(I)7(C=C)] interaction. The above-mentioned parameters for a series of molecular and anionic copper(I) chloride p-com- plexes are given in Tables 6 ± 8. To reveal the differences in the behaviour of the C:C and C=C bonds with respect to Cu(I), let us consider the data on the alkenyne p-complexes of CuCl and CuBr (Table 7). In the structures of alkenyne complexes 30 ± 37, the coordination envi- ronment about the Cu(I) atom, which is p-coordinated to the C=C bond, is most often a tetrahedron slightly deformed toward the trigonal pyramid (Fig. 8 a). The complexes 34 and 35 contain conjugated alkenyne ligands.However, the Cu(I) coordination polyhedron in the complex 34 is more similar to the tetrahedron than that in the complex 35. In the structure of 35, the C=C groups interact more efficiently with Cu(I) due to the involvement of the C:C bond in p-coordination of the metal atom. The distances from the copper atom to the midpoint of the C=C (Cu7C*) bond in the complexes 30 ± 37 are in the range of 1.951(7) ± 2.00(4) A, which indicates that the efficiency of the p-bonding is smaller than that of Cu7(C:C). In particular, the Cu7C*(C=C) distances in the complex 35 are 1.951(7) and 1.962(7) A, whereas the Cu7C*(C:C) contact is substantially shorter [1.910(6) A]. More substantial differences are observed in the structure of 36 in which the multiple bonds of the ligand are isolated [the Cu7C*(C=C) and Cu7C*(C:C) bonds are 1.97(1) and 1.86(1) A, respectively].In this complex, the metal atom coordinated by the C:C bond is in a planar-trigonal environment (Fig. 8 b), whereas the Cu atoms coordinated by the C=C bond are in a trigonal-pyramidal environment. In the copper(I) bromide complexes 33 and 37a,b, the corresponding geometric parameters of the coordination p-units differ only slightly from each other due to steric hindrances caused by the bulkier bromine atoms. In the complex 32, the geometric charac- teristics of the Cu(I) polyhedra formed with the participation of the C:C or C=C groups are also only slightly different, which suggests that the p-interactions of these bonds with the metal atom are virtually identical in efficiency.To put it differenlty, the efficiency of p-coordination of the C:C bond decreases due to R C C Cu R L L Lax L 967 the presence of the phenyl substituent giving rise to steric hindrances. On the whole, the Cu(I) atoms in alkenyne p-com- plexes tend to interact predominantly with the C:C bond. The affinity of alkyne ligands for Cu(I) is particularly pro- nounced both in molecular and anionic copper(I) halide p-com- plexes (24 ± 27 and 38 ± 42; Tables 6 and 8). If the C:C group of the alkyne is coordinated only to one Cu atom, the environment of the latter is exclusively planar-trigonal (Fig. 8 b). In this case, very short Cu7C*(C:C) distances [1.886(4) ± 1.92(1) A] are observed, which is indicative of efficient interactions (compounds 25, 26 and 38 ± 42).In the compounds 24 and 27, the Cu7C*(C:C) distances are 1.941(4) and 1.946(6) A, respec- tively. Efficient p-interactions in the complexes 24 and 27 are hindered by the phenyl group at the C:C bond and the bridging character of the octadiyne molecule, respectively. The propargyl alcohol molecules in the molecular (28) and anionic (29) p-complexes and the vinylacetylene molecule in the compound 30 whose terminal C:C group acts as a bridging p-ligand simultaneously coordinated to two metal atoms (Fig. 8 c), exhibit peculiar behaviour with respect to Cu(I). One of two copper atoms of the Cu(I)7(C:C)7Cu(I) unit is always located somewhat closer to the midpoint of the triple bond.If the coordination polyhedra of Cu(I) in the the complexes 28 ± 30 are arranged in order of increasing Cu7Lax distances, it can be readily seen that the deformation of the Cu coordination poly- hedra from the virtually regular tetrahedron [Cu7Lax, 2.416(2) A] to [Cu7Lax, the trigonal pyramid 2.605(5) ± 3.058(7) A] and then to planar-trigonal coordination [Cu7Lax, 3.738(6) A] is accompanied by a corresponding short- ening of the Cu7C*(C:C) distance, which indicates an increase in the efficiency of p-bonding in the series under consideration. The triple bond in the p-complexes of acetylene (17 ± 23; Table 6), as in the above-considered complexes 28 ± 30, is coordi- nated simultaneously to two metal atoms (Fig.8 c). This coordi- nation is retained both in molecular and anionic compounds. In the structures of the compounds 17 ± 23, the Cu7C* distances are noticeably longer than those in alkyne and alkenyne p-complexes and are similar to the Cu(I)7(C=C) distances. The average Cu7C* distance in the molecular p-complexes of acetylene 17 (see Fig. 6) and 18 [1.97(2) ± 2.29(9) A] are noticeably larger that those in the anionic compounds 20 ± 23 [1.89(2) ± 1.957(5) A]. The latter compounds are characterised by virtually planar-trigonal environments about the metal atom. Therefore, the C:C group is more prone to interaction with Cu(I) than the C=C group even if the former occupies a position inside the ligand, which is less favourable for coordination to the metal atom.To account for this fact, let us use the data on the electron-donating and electron-withdrawing properties of the ethylene and acetylene molecules and the C2H2 molecule p-coor- dinated to Cu(I). It is known that acetylene in the ground state exhibits the weaker electron-donating properties (two HOMOs have lower energies) and the weaker electron-withdrawing properties (two LUMOs have higher energies) compared to ethylene.13, 90 How- ever, in the case of cis-distortion of the alkyne molecule in the coordination sphere about the metal atom, the degeneracy of the MOs is removed. As a result, the energies and geometric param- eters of the frontier MOsof the alkyne change and it becomes both a better electron-acceptor and electron-donor than alkene. There- fore, the more efficient Cu(I)7(C:C) p-interaction results from the presence of two p-MOs in the alkyne molecule, which are split into MOs with different energies (pk and p\), as well as from the stronger cis-distortion of the molecule compared to that of alkenes.Due to the presence of two MOs in the alkyne, it can interact with Cu(I) and other metals as a bridging ligand and can coordinate two Cu(I) atoms. In this case, as mentioned above, the Cu7C* distances in the Cu(I)7(C:C) p-units are noticeably shorter than those in the Cu(I)7(C:C)7Cu(I) p-units, which indicates that the efficiency of the p-interaction in the first case is968 higher. This fact is in agreement with the general regularities of coordination chemistry according to which the bond between a bridging m-L ligand and the metal atom is always weaker than that with the participation of the terminal ligand.This also takes place in the cases of m-Cl, m-Me, m-CO and m-C2H2. In particular, this fact can be explained as follows: if each p-MO of alkyne acts independently (as alkene), two p-MOs, which are weaker p-donors than the highest pk-MO of acetylene in the nonbridging p-complex, are involved in bonding. It is also possible that pk-MO as well as pk -MO and p\ -MO are involved in the formation of the coordination bond in the Cu(I)7(C:C) units.91 However, this question calls for special investigation. The efficiency of the interaction between Cu(I) and the multi- ple carbon ± carbon bond affects not only the distance from Cu(I) to the midpoint of the Cu7C* bond, which is determined by the contribution of the donor-acceptor component, but also the length of the coordinated bond determined predominantly by the dative component. The data in Tables 6 ± 8 demonstrate that the smaller the difference in the Cu7Cdistances (d), the longer the C:C or C=C bond.Thus, two crystallographically independent C2H2 ligands in the crystal of the compound 21a (Fig. 9) differ in the C:C bond length [1.25(1) and 1.32(2) A]. This is indicative of the higher and lower efficiency of the correspodning dative components of the Cu(I)7(C:C) p-interactions. In the second case [the coordination unit with the participation of the C(1):C(1)0 bond], the asymmetry of the Cu7C bonds is absent, whereas the less elongated C(2):C(2)0 bond is located somewhat asymmetrically with respect to the Cu atom [d=0.03(1) A], which decreases the efficiency of the dative Cu(I)7C(2):C(2)0 inter- action.This tendency was revealed for a rather large number of examples. However, the C:C bond in the Cu(I)7(C:C) p-unit of the compound 36 is not elongated [1.18(2) A] in spite of the total symmetry of the Cu7C distances (d=0 A), although the x C(20) N(1) C(2) Cl(3) y Cu(3) H(2) Cu(4) Cl(1) H(1) Cl(4) Cl(2) C(1) N(2) Cu(2) C(10) Cu(1) Figure 9. Projection of the structure of NH4Cu3Cl4(HC:CH) (21a) onto the xy plane. BMMykhalichko, O N Temkin,M G Mys'kiv Cu7C* distance [1.86(1) A] is indicative of a substantial charge transferM/L increasing the degree of ionicity of the bond.In most of the structurally studied Cu(I) p-complexes, the terminal carbon atom of the terminal C=C bond is generally located somewhat closer to the Cu atom, whereas the coordinated inner C:C bond in alkyne and alkenyne ligands is most often symmetrical with respect to the Cu atom. Apparently, the inner C:C bond is more stable to the shift with respect to the Cu7C* axis than the C=C bond due to the presence of two p-MOs in the former case. However, the C:C group in CuCl p-complexes with terminal alkynes is characterised by a high tendency to form the asymmetrical Cu(I)7(C:C) p-unit primarily due to the forma- tion of:C7H.. . Cl hydrogen bonds with the participation of the terminal H atom. Acetylene and its monosubstituted derivatives are character- ised by high CH-acidity.90 The dipole moment (m) of the C7H bond in acetylene is 1.0 D. The positive charge of the dipole is localised on the H atom (the same direction of the dipole of the =C7H bond is observed in alkenes; however, m for alkenes is approximately 0.6 D). The acidity of the C7H bond in alkynes increases upon the formation of p-complexes with Cu(I) and some other metals. Due to polarisation of the Cd77Hd+ bond, p-com- plexes become strong acids, which can dissociate to form acetyle- nide compounds depending on pH of the solution. The tendency for dissociation is associated not only with polarisation of the C7H bond, but also with the high stability of coordination products of the resulting RC:C7 anions with Cu(I).4, 13 At the same time, polarisation of the C7H bond in p-complexes is favourable for hydrogen bonding with the chloride ion or the coordinated chloride ligand.Thus, acetylene or its monosubsti- tuted derivatives in strongly acidic media form p-complexes with CuCl. In these complexes, :C7H. . . Cl hydrogen bonds exert a stabilising effect on the structures of the compounds.92, 93 The energies of the hydrogen bonds of this type are comparable with p-interaction that [EC7H. . . Cl*20 kJ mol71; of the ECu(I)7(C:C)*50 kJ mol71]. In structures of p-complexes with acetylene, the :C7H.. . Cl contacts cause additional polarisa- tion of the Cu7Cl bonds, thus being favourable for strengthening of the p-bonding in the Cu(I)7(C:C)7Cu(I) unit,17 which is, in turn, reflected in the geometry of the p-coordinated C:C bond. Moreover, there is a particular relationship between elongation of the coordinated C:C bond and the length of the :C7H. . . Cl hydrogen bond. Thus elongation of the triple bond is accompa- nied by weakening of the corresponding C(H) . . . Cl contact. An increase in the C:C bond length due to weakening of the donor- acceptor interaction and strengthening of the dative interaction leads to a decrease in polarisation of the C7H bond and weakening of the hydrogen bond. On the contrary, a smaller elongation of the triple bond due to the stronger donor-acceptor interaction and an increase in polarisation of the C7H bond results in a stronger hydrogen bond and shortening of the H.. . Cl distance. The electron-withdrawing chlorine atoms, in turn, can substantially affect polarisation of the C7H bonds of the p- coordinated acetylene molecule due to the C(H) . . . Cl interaction. Thus the stronger the C(H) . . . Cl hydrogen bond, the higher the degree of polarisation of the:Cd77Hd+ bond. In other words, the formation of the C7H. . . Cl hydrogen bond enhances the donor-acceptor interaction, which, in turn, strengthens the hydro- gen bond. The scheme of action of the :C7H. . . Cl hydrogen bonds in the structures of copper(I) chloride p-complexes of C2H2 is shown in Fig.10. The characteristics of the :C H. . . Cl hydrogen bonds in the structures of CuCl p-complexes with acetylene and its monosub- stituted derivatives are graphically represented in Fig. 11. Accord- ing to the criteria reported in a study,94 the C7Cl distances indicate the existence of a hydrogen bond. It can be seen from the plot that there is a linear correlation between the elongation of the p-coordinated C:C bond and theH. . . Cl distance. However, this is true only for p-complexes (anionic and molecular) in the structures of which the C:C bond of the ligand molecule acts aPolynuclear complexes of copper(I) halides: coordination chemistry and catalytic transformations of alkynes Cld7 C d+ C d+ Hd+ H H d7Cl Cl d7 Cd7 d+ d+ Cu Cu Cd7 Cl Cl d7 d7 H Hd+ d+ Hd+ C C Cld7 Figure 10.Scheme of :C H. . . Cl hydrogen bonds in the structures of copper(I) chloride p-complexes of acetylene. d(C)H_Cl /A 21b 25 3.0 21a 29 20 20 2.8 26 24 21b 28 22 2.6 30 21a 28 2.4 28 23 1.24 1.20 1.28 1.32 lC:C /A Figure 11. Correlation between the (C)H . . . Cl distance and the C:C bond lengths in CuCl p-complexes with acetylene and its monosubstituted derivatives (the linear correlation coefficient r=0.85). 4 4 bridge to form the Cu(I)7(C:C)7Cu(I) p-unit. The complexes 24 ± 26 are exceptional (see Fig. 11) due to the formation of the Cu(I)7(C:C) p-unit in which the p-interaction characteristic of nonbridging p-complexes prevails. Against its background, the polarising effect of the:CH.. . Cl contact is less significant. Hydrogen bonds play an important role in the structures of the anionic p-complexes 20 and 21a. In the complex 20, the lengths of the hydrogen bonds between the hydrogen atoms of two crystallographically independent C2H2 molecules and the chlorine atoms are virtually equal [2.85(8) and 2.80(9) A]. In the structure of 21a, these bonds are substantially different [2.98(12) and 2.55(12) A] due to which the p-coordinated triple bond of one C2H2 molecule is substantially elongated [1.32(2) A] compared to the second bond [1.25(1) A]. In the structure of 20, the parameters of the p-coordinated triple bond are virtually identical [1.28(1) A] and correspond to the average value for the two limiting C:C bond lengths in the structure of 21a.In these isostoichiometric complexes, the differences are observed not only in the geometric parameters of the crystallographically independent p-coordinated units, but also in the symmetry of their unit cells due to the different effect of the :C7H. . . Cl hydrogen bonds. The struc- tures of the complexes 20 and 21a contain structurally similar (but not identical) (Cu4Cl4)n chain fragments (Fig. 12). Thus, all units of the chain in the structure of 20 have the symmetry 2 (Fig. 12 b), whereas the Cu4Cl4 units in the chains of the structure of 21a are characterised by the symmetry 2 and m (Fig. 12 a). Apparently, this effect caused by the replacement of the K+ ions by the NHá ions results from the NHá .. . Cl7 interaction. Thus the ammo- nium cation's adopting the tetrahedral configuration forms (N)H . . . Cl hydrogen bonds [2.20 ± 2.87(9) A], which is favourable for the formation of the crystal structure of the complex 21a, which differs from the crystal structure of the complex 20. There- fore, the competition of two interactions, viz., M+. . . Cl7 and :C7H. . . Cl, rather than the :C7H. . . Cl hydrogen bond affects the geometry of the p-coordinated C:C bond in the C2H2 molecule in the anionic complexes. In the anionic CuCl p-complexes with acetylene and its monosubstituted derivatives, the efficiency of the Cu(I)7(C:C) p-interaction depends on the a Cl Cu HC 2m2m Figure 12.Chain fragments (Cu4Cl4)n in the structures of the complexes NH4Cu3Cl4(HC:CH) 21a (a) and KCu3Cl4(HC:CH) 20 (b). efficiency of :C7H. . . Cl hydrogen bonding whose strength is determined by the type of the outer-sphereM+ cation (cations of alkali metals, ammonium or protonated amines). In the case of soluble anionic p-complexes, hydrogen bonding can also occur if the bridging or terminal chlorine atoms forming the coordination polyhedra occupy convenient positions. Weak hydrogen bonds between the NHá4 ion and the terminal Cl atoms can also exist. The possibility of hydrogen bonding with the participation of water molecules is also beyond question Cu C:C7H_OH2 . In this connection, the role of theM+ cations as a component of the Nieuwland catalyst (MCl7CuCl7H2O) becomes appa- rent.2. Acetylenide (ethynyl) complexes of copper(I) As mentioned above, the formation of ethynyl compounds (com- pounds possessing at least one M7C: bond) in protophilic media (or in the presence of bases) is a characteristic feature of the reactions of acetylene and monosubstituted alkynes with Cu(I) complexes. The polarisation of the:C7H bond in Cu(I) p-com- plexes is substantial and the strength of the resulting Cu7C: bond is so high that the Cu+ ion reacts with C2H2 to form CuC:CH even in strongly acidic media in the Cu2SO4±H2SO4 system (up to [H2SO4]=20%± 25%). This reaction is accompa- nied by the formation of vinylacetylene in small amounts.95 Cu(C2H2)++H2O Studies of the formation of the so-called `yellow' complexes in the reactions of C2H2 with solutions of (CumCln)(n7m)7 polynu- clear anions by spectrophotometric and potentiometric methods demonstrated that the following equilibria occur in solutions at 25 ± 80 8C depending on the acidity of the medium 58 (Cum)C2H (Cum)(C2H2) 7H+ 969 b Cl Cu C CuC2H+H3O+ (Cup) (Cum+p)C2.7H+970 By increasing the acidity of the medium, one can go from symmetrical acetylenide (Cum+p)(C2¡ 2 ) to unsymmetrical (Cum)(C2H7) and again to the (Cum)(C2H2) p-complex. The formation of solid acetylenide complexes containing Cu2C2 has been thoroughly studied by Vestin.96 In particular, the equilibrium constant of the formation of the `violet' complex 56 (K=a2HCl/pC2H2=4 at 25 8C) was measured K C2Cu2(CuCl)x(solid)+2HCl C2H2+(x+2) CuClsolid 56 and the composition of the `yellow' solid complex 57 was established C2H2+8CuClsolid+2K+ K2C2(CuCl)8(solid)+2H+. 57 It was concluded 96 that the solution of the CuCl7KCl7H2O system contains the [C2(CuCl)n]27 anion.Unsymmetrical ethynyl complexes in solutions 58 {[HCl]=0.05 ± 0.6 m *, the acidity of the medium (h0) is 0.2 ± 5, pH=1.1 ± 0.0} have predominantly the compositions [Cu4Cl4(C:CH)]7 and [Cu4Cl5(C:CH)]27. Recently, ethynyl complexes and acetylenide complexes of copper(I) chloride have been synthesised and their structures have been studied (see Table 9 and references therein). These studies provided convincing evidence that unsymmetrical polynuclear ethynyl complexes [CumCln(C:CR)](n7m+1)7 do exist.New structures, which have not been described previously for non- halide complexes, were revealed.97 ± 116 In the case of acetylene, the (ImH)4Cu9Cl11(C:CH)2 (43) 45 and (PyH)4Cu9Cl12(C:CH) (44) 39 compounds were obtained as insoluble (crystalline) acetylenide complexes. In these complexes, the C:CH group is bridging and is coordinated to the copper atoms only through one a-carbon atom (the m-Z1-, Z1-C:CH- or s,s-modes of coordination) (Fig. 13). The C:C bond lengths in these complexes are 1.21(2) and 1.19(7) A, respectively, and the Cu7C bond lengths are 1.97(1) and 1.92(3) A, respectively. This type of the ligand is rather often observed in ethynyl complexes of the Cu2(C:CR)2L4 (L=PR3) type.97 For simplicity, these Cu(8) C(4) Cu(2) 2.05(3) C(1) C(2) Figure 14.Fragment of the structure of the (NH4)8Cu29Cl29(C:C)4 . 7H2O complex (45). The distances from the Cu(3) and Cu(5) atoms to the midpoint of the C(1)7C(2) bond are 2.18(3) and 2.01(2) A, respectively. BMMykhalichko, O N Temkin,M G Mys'kiv x Cl(2) z C(2) C(1) Cu(4) Cu(3) Cl(1) N Cu(2) Cu(1) 1/2 Cl(3) Cl(4) Cl(5) Figure 13. Projection of the structure of (PyH)4Cu9Cl12(C:CH) (44) onto the xz plane (one layer is shown). 2 ethynyl compounds will be called s-complexes. The complex 43 contains the Cu4Cl5(C2H)27 anion. In the acetylenide complex (NH4)8Cu29Cl29(C:C)4 . 7H2O (45), the C2¡ anion, unlike the C2H7 anion, is coordinated both in the Z1- and Z2-modes (s,p- mode) and serves as a bridge between eight copper atoms (Fig.14). In the structure of 45, two crystallographically inde- pendent C2¡ species are present. Each of these species acts as a 2 Cl Cu(9) C(3) Cu(7) Cu(3) Cu(6) 2.09(3) 1.99(3) 2.43(3) C(2) C(1) 2.10(2) 2.16(3) 2.11(2) Cu(4 0) Cu(5) Cu(2)Polynuclear complexes of copper(I) halides: coordination chemistry and catalytic transformations of alkynes a bCu Cu Cu Cu Cu 1078 1068 Cu Cu C(2) C(1) C(4) C(3) Cu Cu 1258 Cu Cu Cu Cu Cu Cu Cu Figure 15. p,s-Coordination units in the structure of the (NH4)8Cu29Cl29(C:C)4 .7H2O complex (45). 2 bridging s,p,s-ligand, which coordinates simultaneously eight metal atoms. Noteworthy are the modes of coordination of the Cu(I) atoms by two independent C2¡ 2 species.Each of two carbon atoms of the C(3):C(4) group forms s-bonds with three copper atoms (Fig. 15 a). Therefore, the C(3):C(4) group is coordinated in the m3-Z1,Z1,Z1,m2-Z2,Z2,m3-Z1,Z1,Z1-mode, i.e., this group acts as the s,p,s-ligand. The effect of six copper(I) atoms Z1-coordi- nated to the C2¡ ion and two copper atoms Z2-coordinated to the same ion on the C:C bond is so high that the geometry of the Z2-unit changes and the Cu7C*7Cu angle reaches 106 8. In this case, the C(1):C(2) bond can additionally coordinate the third copper atom (Fig. 15 b). Apparently, the unusual mode of coor- dination to the copper(I) atoms is the characteristic feature of the C2¡ 2 anion.Thus the perfectly planar coordination s,p,p,s-unit Cu Cu7C:C7Cu, Cu was found in the structure of the complex [Cu4(Ph2PCH2PPh2)4(C:C)](BF4)2 . 4Me2CO.117 In this unit, the Cu7C*7Cu angle is 180 8. Apparently, the structure of the `yellow' complex studied by Vestin 96 also contains the fragments (a) and (b) (see Fig. 15). The s,p-coordination of several copper atoms by the C2¡ 2 group gives rise to clusters in which the shortest Cu . . . Cu distance is 2.452(6) A. Coordination of the ethynyl ligand to three copper(I) atoms (m3-Z1,Z1,Z1-C:C) to form the trinuclear cluster is also very typical of ethynyl complexes.97 The question about the nature of the Z2-bonds between the C:C group and three metal atoms deserves special analysis.Structural motifs found in ethynyl complexes containing RC:C groups are widely diversified. The efficiency of Z2-coor- dination of the Cu(I) atom by the RC:C group depends on the Z1-mode of coordination. The geometric parameters of the p-coordination units in the structures of the CuCl s,p-complexes with acetylene and its terminal derivatives [which, according to the criteria used in the previous section of this chapter, characterise the efficiency of the Cu(I)7(Z2-C:CR) bonding] are given in Tables 9 and 10. It can be readily seen that two metal atoms in the coordination p-unit (m-Z2,Z2-C:CR) in cluster ethynyl com- plexes of copper(I) are located at somewhat different distances from the midpoint of the C:C bond, as in the CuCl p-complexes with HC:CCH2OH and HC:CCH=CH2 [compounds 28 and 29 (see Table 6) and 30 (see Table 7)].However, shortening of the Cu7C* distance in s,p-complexes, which is determined primarily by the donor-acceptor component of the p-interaction, unlike that in CuCl p-complexes with HC:CCH2OH and HC:CCH=CH2, is not accompanied by an increase in the dative transfer, i.e., it does not cause additional deformation of the trigonal-pyramidal environment about the Cu(I) atom. Moreover, the s,p-complexes are characterised by the following effect: the smaller elongation of the Cu7Xax distance in the coordination p-unit (m-Z2,Z2-C:CR) corresponds to the shorter Cu7C* distance due, apparently, to the synergism of the Cu(I)7(Z1- C:CR) and Cu(I)7(Z2-C:CR) interactions. Such mutual 971 Table 10.The CuC*Cu angles in the p-coordinated Cu(I) units in the structures of p,s-complexes of CuCl with acetylene and its terminal derivatives. CuC*Cu angle /deg Ref. Cluster type Complex 45 75 45 76 77 78 52 39 46 48 49 50 51 52 125(1) 125(1) 107(1) 106(1) 93.5(6) 95(1) 93.6(9) 97(1) 90.4(7) 96.9(2) 89.8(2) 96.9(6) 98.1(3) 98(2) 51 79 66 53 54 55 7777II II II IIV IIV III IV III strengthening of the Z1- and Z2-bonds is manifested to a greater or lesser extent depending on the number of the s-coordinated Cu atoms and their arrangement with respect to the C:C group. In this connection, it is worthwhile to consider the stereo- chemistry of the [Cu4(C:CR)] and [Cu5(C:CR)] clusters found in the structures of copper(I) chloride s,p-complexes with termi- nal derivatives of acetylene.Four major types of cluster were distinguished (see Table 10). The differences in their structures are determined by the difference in the orientation of two mutually perpendicular p-MOs of the ethynyl group of the s,p-coordinated C:CR ligand with respect to the plane P passing through the C:C bond (Fig. 16). In the structures of the complexes contain- ing clusters of type I, the plane of one of the p-MOs of the C:C group (for example, pk) coincides with the plane P, i.e., the p-coordinated C(2) atom and the C:C bond together with two s-bonded Cu(3) and Cu(4) atoms are in a single plane, whereas the normal mn to this plane coincides with the axis of another p\-MO [the Cu(1) ±C* axis] (Fig.16 a). Both p-MOs of the C:C group (pk and p\) in clusters of the types II ± IV, unlike those in clusters of type I, are located at an angle of 45 8 to the normal mn (as well as to the plane P). [In the case of clusters of the types III and IV (Fig. 16 c,d ), the plane P passes perpendicularly to the normal mn to retain the spatial orientation of the p-coordinated copper(I) atoms analogous to that in clusters II (Fig. 16 b).] The clusters II, III and IV differ primarily in the orientation of the s-coordinated Cu atoms. Thus, the s-coordinated Cu(3) and Cu(4) atoms in clusters of the type II lie in the plane P (Fig. 16 b). The rotation of these s-coordinated Cu(3) and Cu(4) atoms about the C:C axis by 90 8 in such a way that they form a new plane (Q) with the normal mn leads to clusters of the type III (Fig.16 c). The replacement of one s-coordinated Cu(4) atom by two atoms [Cu(4) and Cu(5)] in such a way that these atoms are no longer located in the plane Q [the Cu(4) and Cu(5) atoms are located behind and in front of the plane Q, respectively; the Cu(4)C(1)Cu(5) angle &80 8] gives rise to a cluster of the type IV (Fig. 16 d). In addition to the above-considered four cluster types, combined types of clusters are also observed in the structures of Cu(I) s,p-complexes. An example is the [Cu8] cluster in the structure of 52, which consists of clusters of the types I and IV linked to one another through the bridging Cu atom (Fig.17). The [Cu9] cluster in the structure of 54 is composed of two clusters of the type IV (Fig. 18). Clusters of the types I ± IV were found for the first time in ethynyl complexes of copper(I) chloride. In the structures of s,p-complexes containing clusters of the type I, the coplanar arrangement of the p- and s-coordinated metal atoms [for example, the p-coordinated Cu(3) atom and the s-coordinated Cu(1) and Cu(2) atoms in the structure of 50 (Fig. 19)] and the C:C bond leads to the mutual strengthening972 Cu(4) Cu(3) Cu(4) Figure 16. Types of clusters observed in the structures of ethynyl copper chloride complexes: (a) type I {the [Cu4(C:CCH2OH)] cluster}; (b) type II {the [Cu4(C:CCH2OH)] cluster}; (c) type III {the [Cu4(C:CR)] cluster, where R=CH=CH2 or CH2OH}; (d) type IV {the [Cu5(C:CCH2OH)] cluster}; mn is the normal to the plane P.C(6) O(2) C(5) H C(4) Cu(7) Cu(4) Figure 17. Cluster fragment [Cu8(C:CCH2OH)2] in the structure of (PyH)2Cu8Cl8(C:CCH2OH)2 .H2O (52). H C(3) C(2) Cu(4) Cu(5) Figure 18. Cluster [Cu9(C:CCH2OH)2] fragment [HO(CH2)2NH3]Cu9Cl8(C:CCH2OH)2 .2H2O complex (54). a n Cu(1) Cu(2) P Cu(3) C(2) O C(1) m C(3) c n Q Cu(1) Cu(2) P C(2) C(1) m R O(1) Cu(8) C(3) Cu(6) C(2) Cu(1) Cu(5) C(1) Cu(3) Cu(2) O(1) Cu(3) Cu(2) C(1) Cu(1) H Cu(4) Cu(3) Cu(4) Cu(5) x z y Figure 19. Fragment of the structure of Cs2Cu5Cl6(C:CCH2OH) (50).the in of two individual types of interactions, viz., the acetylenide and p-interactions. This process becomes so efficient that the coordi- nation polyhedra about the s- and p-coordinated copper(I) atoms become planar-trigonal (see Tables 9 and 10). Apparently, this synergism occurs due to the angular overlapping simultaneously of three orbitals, viz., the occupied dx27y2 orbitals of the s- [Cu(1)] and p-coordinated [Cu(3)] metal atoms, and the antibonding unoccupiedMO(for example, pk ) of the acetylene bond. It should be noted that there are spatial prerequisites for this `three-centre' overlapping in the structure of 50 [the Cu(2)C(1)Cu(1), Cu(1)C(1)Cu(3) and Cu(2)C(1)Cu(4) angles are 81.3(9) 8, 84.2(9) 8 and 89.9(9) 8, respectively].However, mutual strengthen- ing of the Cu(I)7(Z1-C:CR) and Cu(I)7(Z2-C:CR) interac- BMMykhalichko, O N Temkin,M G Mys'kiv b n Cu(1) Cu(2) Cu(3) C(2) C(1) m C(3) d n Q Cu(1) Cu(2) P C(2) C(1) m C(3) 1/2 Cl(4) Cs(1) Cl(5) Cs(2) Cu(5) Cl(2) Cu(4) Cu(1) Cl(1) Cu(2) Cl(3) C(1) C(2) P Cl(6) Cu(3)Polynuclear complexes of copper(I) halides: coordination chemistry and catalytic transformations of alkynes z C(2) C(3) Cl(1) Cl(2) Cu(1) Cu(3) Cl(3)Cu(4) H Cu(2) Cl(5) C(1) Cl(4) Figure 20. Anion [Cu4Cl5(C:CCH2OH)27]n in the structure of the (NH4)2Cu4Cl5(C:CCH2OH) .H2O complex (48). tions does not cause the expected substantial elongation of the C:C bond in the complex 50 [1.22(3) A] because the above- described angular dative process causes polarisation of the C(1)d7:C(2)d+ bond.Quite a different situation is observed in the case of formation of the cluster of the type II (see Fig. 16 b). In this cluster, strengthening of two types of interactions (s and p) is less efficient for geometric reasons because two p-MOs of the alkyne are located at an angle of 45 8 to the normal mn.As a result, the coordination polyhedra about the s-coordinated Cu(I) atoms in the structures of 46, 48 (Fig. 20) and 49 remain tetrahedral, whereas the p-coordinated Cu(I) atoms are characterised by the trigonal-pyramidal environment. In this case, the C:C bond is substantially elongated [1.29(1) A] due to lower polarisation. The considerable synergism of the s and p interactions is also observed in the complexes 54 (Fig.21), 53 and 55 (Fig. 22), which contain clusters of the types III and IV (see Figs 16 c,d). Their p- and s-coordinated Cu(I) atoms are characterised predominantly by the planar-trigonal environment and the C:C bond is virtually not elongated (see Tables 9 and 10). Cu(1) C(3) Cu(2) C(2) O C(1) Cu(3) H Cu(4) Cu(5) Figure 21. Cluster fragment [Cu5(C:CCH2OH)] in the structure of Rb3Cu7Cl9(C:CCH2OH) .H2O (51). Cu(5) Cu(4) Cu(3) C(1) C(3) C(2) H C(4) Cu(2) Figure 22. Cluster fragment [Cu4(C:CCH=CH2)] in the structure of Cu25Cl17(C:CCH=CH2)8 . 14H2O (55). 3. Structural aspects of formation of some p-complexes and s-ethynyl derivatives Structures of p-complexes and products of their reversible con- versions (s-complexes) are determined both by the structures of the (CumCln)(n7m)7 anions in solution and the stability of the products, i.e., by the favourable formation of a particular struc- 973 ture.However, the structures of compounds isolated from sol- ution may not be directly associated with the mechanism of their formation because conversions of the initial intermediates of any type lead to the formation of more stable structures, but the least soluble compounds are isolated from solution. Thus, the ethynyl compound with composition Cum(C2H) can be formed by the reaction of Cup(C2H) with Cuq (p+q=m) in solution. At the same time, it is of interest to analyse possible `genetic' relations between the structures of the initial copper chloride anions and the products formed in the reactions of the latter with RC:CH.Let us consider the group of crystalline compounds formed in the (C6H5NH3)Cl ± CuCl ±H2O±HC:CCH2OH system. Col- ourless crystals of (C6H5NH3)Cu4Cl7 (11), which precipitated from a concentrated almost neutral solution of CuCl (4 mol) and (C6H5NH3)Cl (3 mol) in H2O (20 mol), rapidly dissolved after addition of propargyl alcohol to give yellow crystals of the ethynyl complex 53. 2 (C6H5NH3)3Cu4Cl7(solid)+HC:CCH2OH+2H2O 11 (C6H5NH3)3Cu8Cl10(C:CCH2OH) .2H2Osolid+ 53 +3(C6H5NH3)Cl+HCl. The reaction was accompanied by a decrease in pH to *1.5. Under analogous conditions [but when CuCl and (C6H5NH3)Cl were taken in the molar ratio of 2 : 1], the complex 49 was obtained.4CuCl+2C6H5NH3Cl+HC:CCH2OH (C6H5NH3)2Cu4Cl5(C:CCH2OH)solid+HCl. 49 Unlike the complex 53, the ethynyl complex 49 was converted into the p-complex 29 upon storage in the mother liquor (pH41). This fact indicates that the process is non-equilibrium in charac- ter. (C6H5NH3)2Cu4Cl5(C:CCH2OH)solid+HCl 49 (C6H5NH3)Cu2Cl3(HC:CCH2OH)solid+(C6H5NH3)Cu2Cl3 . 29 Let us consider the relationship between the stereochemistry of compounds 11 and 53 and the possible mechanism of formation of the latter in the solid-phase process (Fig. 23). The subunit of the (Cu4Cl3¡ 7 )n anion in the structure of 11, which consists of the Cu coordination polyhedra linked to each other, is shown in Fig.23 a. The molecule of propargyl alcohol can replace the m-Cl atom (which is hatched in Fig. 23 a) in 11 and can be incorporated into the initial macropolyhedron to form the p-complex (Z2-RC:CH). In this case, the octahedral cavity in the macro- polyhedron is transformed into a square-pyramidal cavity. Sub- sequent deprotonation of the p-ligand with the H2O molecule affords the expected p-complex with the 7C:CR ligand giving rise to the bridging m-Z2,m-Z2-C:CCH2OH ligand (Fig. 23 b). The formation of the p-complex leads to weakening of some Cu7Cl bonds (indicated by dashed lines in Fig. 23 b) and to their cleavage accompanied by the formation of two new Cu7C:Z1- bonds (Fig. 23 c) in the [Cu4(C:CCH2OH)] cluster, which is a component of the complex 53.In the structure of the latter, the [Cu8Cl10(C:CCH2OH)37]n anions form layers in which the s,p-coordinated C:CR group of each propargylium anion is clamped between four Cu atoms inside the square-pyramidal cavity. Due to steric hindrance to the penetration of the H3O+ ion into the cavity toward the a-carbon atom of the C:CCH2OH ligand and the high strength of bonding of the s- and p-coor- dinated anion [in the structure of 53, the length of one of the Cu7C bonds of the acetylenide type is 1.89(2) A and the Cu7C*974a b c Cl Cl Cu Cu Cl ClCu Cl Cl Figure 23. Conversion of the (C6H5NH3)3Cu4Cl7 complex (11) into the (C6H5NH3)3Cu8Cl10(C:CCH2OH) .2H2O complex (53) in the presence of HC:CCH2OH (the stereochemical aspect).distance is 1.97(1) A], the complex 53 is very resistant to decom- position or conversion into the p-complex under the action of HCl. The formation of the complex 49 differs from the above- considered complex in the structural features. Judging from the stoichiometry of the formation of the complex 29, the Cu2Cl¡¦ anion is involved in the formation of the complex 49. This anion was revealed for the first time in the CsCu2Cl3 complex (5) (Fig. 24) and its structure differs fundamentally from that of the (Cu4Cl3¡¦ 7 )n anion. The Cu2C¡¦3 anion is characterised by the fact that the coordination polyhedra about the Cu atoms are linked to each other exclusively through edges to form infinite (Cu2Cl¡¦ chain fragments within which all metal atoms are virtually in a single plane.In this case, a peculiar sequence of planar (approx- imately square) n[Cu4] matrices with Cu . . . Cu contacts [2.81(1) A] a b Figure 25. Scheme of formation and conversions of the crystals of the (C6H5NH3)2Cu4Cl5(C:CCH2OH) complex (49) into (C6H5NH3)2Cu2Cl3(HC:CCH2OH) (29): (a) the [Cu4Cl5(C2R)27]n anion; (b) the [Cu2Cl3(RC2H)7]n anion. BMMykhalichko, O N Temkin,M G Mys'kiv Cl H Cl C + Cu H O Cl Cl Cu 3 Dm4-Cl. Figure 24. Anion in the structure of the CsCu2Cl3 complex (5). 3 )n 3 appears. In solution, the complex 5 (Fig. 24) produces the Cu2Cl¡¦ anions (A), which interact with alkyne to form the p-complex [Cu2Cl3(RC2H)]7 (B).HCl Cu2Cl¡¦3Polynuclear complexes of copper(I) halides: coordination chemistry and catalytic transformations of alkynes 29C6H5NHá3Cu RC2H 2Cl¡3 (7HCl) [Cu Cu2Cl¡3 3 ) HCl (7Cu2Cl¡ 2Cl3(RC2H)]7 B CsCu2Cl3 5 A C6H5NHá349 [Cu4Cl5(C2R)]27 C R=CH2OH. 33 The formation of the complex 49 is a non-equilibrium process and is, most likely, determined by the kinetics of its crystallisation. As a consequence, the p-complex B rapidly reacts with the Cu2Cl¡ ion in solution to yield the ethynyl complex C, which precipitates as the compound 49. The slow reaction of the complex 49 (through the soluble form C) with HCl followed by elimination of Cu2Cl¡ to form (through the complex B) gives rise to the complex 29, which is thermodynamically more stable under these conditions.In the [Cu4Cl5(C:CCH2OH)27]n anion of the compound 49 3 fragments to form the [Cu2Cl3(HC:CCH2OH)7]n (Fig. 25 a), the a-carbon atom of the 7C:CCH2OH fragment is accessible to attack by the proton. In addition, the bonds between the s-coordinated Cu atoms and the a-carbon atom in the structure 49 are weaker [the Cu7C: bonds are in the range of 1.95(2) ± 2.00(2) A] than those in the complex 53. Hence, the possibility of the conversion of the [Cu4Cl5(C:CCH2OH)27]n anion in crystals of the complex 49 accompanied by elimination of the Cu2Cl¡ anion of the complex 29 must not be ruled out (Fig. 25 b). Crystals of the complex 29 appear in the mother liquor (pH41) immedi- ately after the disappearance of crystals of the complex 49.IV. Catalytic conversions of acetylenic compounds in solutions of copper(I) chloride complexes Studies of the physicochemical properties of the NH4Cl ± CuCl ±H2O system provided the basis for the development of the procedure for investigation of the kinetics of reactions in concentrated aqueous solutions.13, 118 The principal results of these investigations are considered below. 1. Kinetics of catalytic reactions of acetylene and its derivatives in concentrated aqueous solutions ofMCl ± CuCl In studies of the kinetics of various reactions of acetylene (and its derivatives) in concentrated aqueous solutions of MCl and CuCl, kinetic models are generally constructed 13 with the use of an empirical dependence of the following type: r=j (pC2H2, [HCl]S, [CuCl]S, [NH4Cl]S), where r is the reaction rate and p is the partial pressure.If the solution contains polynuclear complexes and does not contain an excess of chloride anions with respect to metal, there is no way to express the concentration of free catalytically active complexes (one or several) in terms of [CuCl]S (i.e., to account for the material balance with respect to the catalyst in the kinetic equation). Hence, the only approach to a theoretically justified adequate kinetic model and to the elucidation of the reaction mechanism in these systems consists in measuring the concen- trations (or activities) of the free reagents and components of the solution (Cu+ and Cl7) under stationary conditions of catalytic reactions and in maintaining constant conditions, the partial pressures and concentrations of the reagents being varied, i.e., in maintaining the [CuCl]S : [Cu+] and [Cl7]S : [Cl7] ratios and, consequently, [(CumCln)(n7m)7].In other words, it is necessary to add the initial catalyst (CuCl and, if required, MCl) as the catalyst is trapped by the reagent. The values, which are directly proportional to the activities (and concentrations) of the Cu+, Cl7 and H3O+ ions (see Section II), can be determined by the potentiometric method (under the conditions of constant cationic content).22, 30, 118 975 Table 11. Catalytic reactions in the CuCl ±NH4Cl ±H2O system. Catalytic reaction Catalytic [CuCl] Ref. system (m *) abc CH2=CHCl (1) MeC(Cl)=CH2 (2) MeCH=CHCl (3) CH2=CHC(Cl)=CH2 (4) de C2H2+HCl C3H4+HCl C3H4+HCl C4H4+HCl C2H2+H2O MeCHO (5) f a C3H4+H2O Me2 CO (6) gh b i b 5 ± 12 jk 5 ± 12 119 0.6 ± 12 120 4 ± 12 120 0.5 ± 8 121 5 ± 12 119 5 ± 12 122 0 ± 12 120 4 ± 12 123 4 ± 12 123 95 5 ± 12 124 5 ± 12 125 5 ± 10 126 MeCH2CHO (7) HC:CCH=CH2 (8) CH2=CHC:CCH=CH2 (9) CH2=CHCH=CHC:CH (10) lm C3H4+H2O 2C2H2 3C2H2 3C2H2 C2H2+HCN CH2=CHCN (11) (12) C2H2+ 13 13 7 ± 12 7 ± 21 no Note.In the reactions (1± 11) [NH4 Cl]=12 m *; in the reaction (12) [NH4Cl]=21 m *. a In addition toNH4Cl, Cu2Ssolid is present; b in addition to NH4Cl, RSH is present. This procedure was used for studying 12 reactions of acety- lenic compounds in 15 catalytic systems (Table 11).The kinetic equations obtained under conditions of the constant content of NHá4 and at the constant aCu+ and aCl¡ values are given in Table 12. It was found that the type of equation is independent of the concentration of CuCl (in the kinetic equations, only the observed kij values are a function of this concentration, where i is the number of the reaction and j is the catalytic system) (see Table 12). In essence, kij is the function of the stationary concen- trations of the (CumCln)(n7m)7 complexes and Cl7 ions or of the stationary activities of Cu+ and Cl7 kij=f (aCu+, aCl7). In the simplest case (the compositions of the initial copper(I) complex and the intermediate of the first irreversible stage are identical in the Cl7 component), the equation is as follows: k (1) kij= effÖm;nÜ[(CumCln)(m7n)7].n m XX Table 12. Kinetic equations for the reactions of acetylenic compounds in the CuCl ±NH4Cl ±H2O system. Ref. Kinetic equation 119 120 120 121 119 122 120 123 123 95 C r1a=k1a pC2H2 [HCl] r2b=k2b pC3H4 [HCl]+kpC3H4 aCl7 r3c=k3c pC3H4 [HCl] r4d=k4d pC4H4 [HCl]+kpC4H4 [HCl] (a2Cl¡ /aH2O) r5e=k5e pC2H2 [H3O+]/(1+k [H3O+] r5f=k5f pC2H2 [H3O+]+ kpC2H2 r6g=k6g pC3H4 r6h=k6h pC3H4 r7i=k7i pC3H4 r8j=k8j p2 2H2 /(kpC2H2+[H3O+]) r9k=k9k pC2H2 pC4H4/(kpC2H2+[H3O+]) r10l=k10l pC2H2 [H3O+]/(kpC2H2+[H3O+]) 124 125 126 r11m=k11m pC2H2 pHCN/[H3O+] r12n(o)=k12n(o) pC2H2 pC5H6 /[H3O+] 13 Note.The subscript in the kinetic equation denotes the number of the reaction and the letter denotes the catalytic system (see Table 11).976 k kij= n m The keff(m,n) values can be determined only if all complexes present in the system are taken into account, i.e., by solving the system of equations of the same type (1). However, taking into account the presence of a large number of complexes in solution, the large error in the determination of kij (10% ¡¾ 15%) and the errors in the determination of the total set of complexes, one cannot expect that the coefficients of this system thus obtained reflect the true contributions of individual complexes to the reaction rate. In this connection, the average composition of the complexes, which make the major contribution to kij both due to the large keff(m,n) values and the high concentrations of (CumCln)(n7m)7, was estimated.From Eqn (1) it follows: XX effOm;nUbm;na Cu m aa nCl¢§ aXXkm;na mCuaa nCl¢§ . (2) m n mb [CuCl]S= m;na mCuaa nCl¢§ , n m The aCl7 value decreases as [CuCl]S increases, because aCu+ and aCl¢§ are interdependent: XX nb [Cl7]S=[Cl7]+ m;na mCuaa nCl¢§ , n m XX due to which the independent determination of m and n is impossible. It was demonstrated 126, 127 that the aCu+ and aCl¢§ values determined by the potentiometric method are related to each other by the empirical dependence throughout the aCu+ range met in the systems under consideration (aCu+=1078 ¡¾ 1076): (3) aCl¢§ a a 0Cl¢§ .ba gCua k k m;nOa 0Cl¢§ Un bn a m¢§ng a Cua ij a n m n m The b and g coefficients were determined experimentally for each case represented in Table 12. From Eqns (2) and (3) it can be shown that XX m;na Cu m aa nCl¢§ aXXk k (4) ka lk Cua . a k X The average observed order (lA) with respect to aCu+ can be determined by approximating Eqn (4) by Eqn (5) , (5) kij=keff(i, j) a l^Cua where lA=mA7nAg is the weighted-mean value, which accounts for the lk values for each complex and the fraction of the reaction rate (yk) determined by the participation of the kth complex a b c kij kij 4 6 6 2 1 3 4 4 2 2 0 0 2 4 6 8 10 12 2 4 6 8 10 12 Figure 26. Dependence of kij on the concentration of CuCl: (1) k1a6103; (2) k2b6105; (3) k3c6107; (4) k4d60.56105; (5) k5e6103; (6) k5f6102; (7) k6g; (8) k6h60.56105; (9) k7i60.56105; (10) k8j6104; (11) k9k6107; (12) k11m; (13) k12n6106. BMMykhalichko, O N Temkin,M G Mys'kiv (6) d lnkij d lnaCua k a l ^aXlkyk .From Eqn (6) it follows that the nAand mAvalues are also weighted-mean. In the general case, the lAvalue should be variable because yk can change as the [CuCl]S changes. The fact that the lA value for all processes is constant throughout the range of CuCl concentrations under study indicates that the m(n) values for the complexes, which make the major contribution to kij, are similar. To estimate the m and n values and the compositions of the initial (CumCly)(y7m)7 complexes (y=n upon elimination or addition of Cl7 ions in the individual stages of the reaction), the following equation was used: l^ (7) q^g ^ m a 1 ¢§ g a 1 ¢§ g a m^min a 1 ¢§ g , q^g where qA=nA7mA.It was assumed that multiply-charged complexes, which are formed in insignificant amounts in concentrated CuCl solu- tions,22, 30, 31, 33 are not involved in the reactions with acetylenic compounds, and hence analysis of the compositions of the (CumCly)(y7m)7 complexes was restricted to y7m=1 or 2. In addition, it was assumed that acetylene replaces the water mole- cule upon the formation of a p-complex, thus retaining the charge of the complex. Then the parameter qAin Eqn (7) can take the values 0, 1, 2 and 3 depending on the presence of stages of elimination of one Cl7 ion (y=n+1, qA=0 or 1) and stages of conversion of intermediates with the participation of one Cl7 ion (y=n71, qA=2 or 3) as well as on the presence of mechanisms in which y=n (qA=1 or 2).In this case, the mAvalues change from mAmin (qA=0) to mAmin+3g/(17g). On these assumptions, mA51 and nA5mAhave a physical meaning. 4 , The dependence of kij on [CuCl]S is shown in Fig. 26. Although the kij value increases as [CuCl] increases in all cases, the character of kij=f[CuCl] is substantially different for different reactions (see Table 11). For dimerisation (8) of acetylene, hydro- cyanation (11) of acetylene, hydration of methylacetylene (6h and 7i) and acetylene (5f) and the reactions with cyclopentadiene (12), the kij value sharply increases only at [CuCl]S55 m *, i.e., in the concentration range in which polynuclear complexes (Cu3Cl¢§ Cu3Cl2¢§ 5 , Cu4Cl¢§5 , Cu4Cl26 ¢§, etc.) appear in noticeable amounts.The observed order with respect to [CuCl]S is smaller than unity only for the reaction with HCl (4d). The observed orders with respect to aCu+ (lA) are given in Table 13. The (CumClnCN)(n7m + 1)7 and [CumCln(C:CH)](n7m + 1)7 complexes are involved in hydrocyanation (11) and diene con- densation (12) of acetylene possessing rate-determining stages. These complexes are formed due to the replacement of the Cl7 ion. kij 9 10 3 8 6 12 2 13 5 1 7 11 0 2 4 6 8 10 [CuCl]S (*m)Polynuclear complexes of copper(I) halides: coordination chemistry and catalytic transformations of alkynes Table 13.The ^mvalues and the probable compositions of the initial (CumCly)(y7m)7 complexes. lA kij k1a k2b k3c k4d k5e k5f k6g k6h k7i k8j k9k k10l k11m k12n 0.73 0.88 1.10 0.71 0.63 1.16 0.67 0.77 0.99 2.60 0.13 0.64 2.21 1.52 2.02 k12o Note. At qA=0, the (CumCly)7 complexes with m=3 and 4 (catalytic systems m and n) and m=5 (catalytic system o) are involved in the reactions (11) and (12), respectively. (CumCly)(y7m)7+HC:CH [CumCln(C:CH)](n7m+1)7+H3O++Cl7. Hence, y=n+1 (at qA=0 or 1) and mAmin=3.6 ¡¾ 5.0 for the reactions (11) and (12). In the reactions (8) ¡¾ (10), which proceed without rate-determining stages, qAreflects the composition of the transition state (y=n).Polynuclear complexes with m=3 ¡¾ 7 are involved in the reactions (8), (9) and (12) (see Table 3). The participation of polynuclear Cu(I) complexes in the dimerisation of acetylene (8) was confirmed by spectrophotomet- ric studies of the dependence of the dimerisation rate on the concentration of the `yellow' complexes [Cu4Cl4(C:CH)7 and Cu4Cl5(C:CH)27] and on their compositions as a function of aCu+, aCl7, pC2H2 and [H3O+] 58 (see Section III) and was addi- tionally supported by isolation of crystalline copper chloride and ethynyl complexes formed in the catalytic Nieuwland system (the structures of these complexes are considered in Sections II and III).2. Kinetics and crystal-chemical aspects of the mechanism of Nieuwland oligomerisation of acetylene Comprehensive investigations of the kinetics of dimerisa- tion 58, 95, 125, 128 ¡¾ 131 and trimerisation 124, 125 of acetylene made it possible to state the major features of the mechanism of these reactions. Let us consider dimerisation of acetylene (8) (see Table 11) giving rise to vinylacetylene (VA) in detail. At constant activities of Cu+ and Cl7, i.e., at constant concentrations of catalytically active complexes, the reaction rate rVA is described by the equation H p2C2H2 k rVA=kpC2H2 a aH3Oaa . A two-stage mechanism without a rate-determining stage corresponds to this equation, where the copper chloride complex is denoted by [Cu] (1) [Cu]+C2H2 (2) [Cu]C2H7+C2H2 min g mA (qA=0) mA (qA=1) 1.49 2.04 1.49 2.34 2.75 4.30 1.03 1.31 1.64 1.06 71.64 1.00 1.45 1.87 3.837 7 7 1.45 4.26 5.30 6.55 0.29 0.33 0.33 0.33 0.32 0.29 0.33 0.47 0.47 0.32 0.33 0.33 0.39 0.60 0.60 73.62 3.80 5.05 (8) [Cu]C2H7+H3O+, k2 X1, 977 (CumCly)(y7m)7 (CumCly)(y7m)7 (qA=2) (qA=1) mA (qA=3) mA (qA=2) 3 CuCl¢§2 , Cu2Cl¢§44 2.25 2.79 3.12 2.54 4 Cu2Cl¢§3 , Cu3Cl¢§ Cu2Cl¢§3 , Cu3Cl¢§ Cu2Cl¢§3 , Cu3Cl¢§ Cu2Cl2¢§ 4 3 CuCl¢§2 , Cu2Cl¢§ Cu2Cl¢§ 5 3 Cu2Cl24 ¢§, Cu3Cl2¢§ Cu2Cl2¢§ 4 3 CuCl¢§2 , Cu2Cl¢§ 6 4 Cu3Cl25 ¢§, Cu4Cl2¢§ Cu2Cl¢§3 , Cu3Cl¢§ 6 4 Cu3Cl25 ¢§, Cu4Cl2¢§ Cu2Cl¢§3 , Cu3Cl¢§ 7 6 Cu4Cl26 ¢§, Cu5Cl2¢§ Cu4Cl¢§5 , Cu5Cl¢§ CuCl2¢§ 3 , Cu2Cl24 ¢§ 1.85 2.30 2.65 2.26 1.96 2.45 1.99 3.23 3.64 4.76 1.18 1.94 Cu2Cl2¢§ 4 2 Cu2Cl¢§3 , CuCl¢§ 7 Cu4Cl26 ¢§, Cu5Cl2¢§ 8 Cu5Cl27 ¢§, Cu6Cl2¢§ Cu7Cl29 ¢§ k3 [Cu]+C4H4. (3)X1+H3O+ The stages 1 and 2, which follow unambiguously from the kinetics, are supplemented by the stage 3, which is required to obtain the final stoichiometric equation for the catalytic reaction (8).It was also found that the k constant in the denominator is independent of [CuCl]S (and aCu+), but depends on aCl7 (k=k0/aCl7). This fact leads to the equation H pC2H2 (9) rVA a k 0pC k 0 2H2 a aH3Oaa aCl¢§a , according to which Cl7 and H3O+ are eliminated in the same stage (the stage 1 or the following quasi-equilibrium stage).Data on the composition of the intermediate X1 (and to some extent on its nature) were obtained in studies of the kinetics of the preparation of vinylacetylene and 2-chlorovinylacetylene (CVA) in systems containing additionally small concentrations (*1073 mol litre71) of CuCl2.132, 133 In the presence of CuCl2, the conversion of the intermediate X1 takes two competitive paths, viz., the stages 3 and 4 k4 CuCl2 X2 (4)X1+CuCl2 CH2=C(Cl)7C:CH+2 CuCl . Under these conditions, the sum of the rates rVA and rCVA is equal to the rate of formation of X1 (i.e., rVA). The rate ratio is described by the equation VA (10) r rCVA 2a , a K aH3Oaa aCuCl where K =k3/k4.As a result, the equation for the rate of formation of vinyl- acetylene includes the concentration of the H3O+ ion, which is involved in the stage 3 after the irreversible stage 2 kH p2C2H2 (11) r kp VA a C : 2H2 a aH3Oaa KaH KaH3Oaa 3Oaa a aCuCl2a . As a consequence, the rate of formation of vinylacetylene (rVA) passes through a maximum as [H3O]+ changes. It was978 suggested that the intermediate X1 with composition [Cu]C4H3 was formed due to the insertion of acetylene at the Cu7C:CH bond, i.e., this intermediate is the b-ethynylvinyl derivative of copper (CuCH=CHC:CH). The reaction of X1 (successively with two CuCl2 molecules) afforded only chlorovinylacetylene. An analogous process took place in the reaction of CuCH=CHCl with two CuCl2 molecules to formCH2=CCl2 (acidolysis yielding CH2=CHCl was the competitive process).134, 135 The formation of chlorovinylacetylene from the intermediate X1 (it is the inter- mediate common to vinyl- and chlorovinylacetylenes) is also confirmed by the fact that none of such compounds as vinyl- acetylene, diacetylene, vinylidene chloride or vinyl chloride in these systems is converted into chlorovinylacetylene under the above-mentioned conditions.132, 133 Investigation of the effect of the addition of HgCl2 on the synthesis of vinylacetylene demon- strated 131 that an increase in the acidity of the CuCl ±NH4Cl ±H2O systems upon the addition of HgCl2 is associated with an exchange reaction giving rise to the organo- mercury derivative ClHgCH=CHC:CH, which is stable to HCl.Vinylacetylene can be isolated from this derivative only under the action of concentrated HCl. Let us consider the stages 1 ± 3 in more detail. Activation of acetylene resulting in its deprotonation (stage 1) occurs, most likely, in the course of several reactions the first of which involves the formation of p-complexes B (CumCln)(n7m)7+C2H2(aqueous) A Kinetic studies demonstrated that replacement of Cl7 in this stage is contradictory to Eqn (9). The formation of the p-complex can occur either in the course of the replacement of the coordi- nated water molecule or as a result of the cleavage of the m-Cl bond without the replacement of Cl7.The replacement of the bridging chlorine atom accompanied by its migration to other coordination vacancies can also occur. Based on the assumption that the short- range order of complexes dissolved in concentrated solutions (which is similar to that in a melt), for example, of Cu3Cl2¡ (see Refs 48 and 49), corresponds to the structure of the complex 9a in the crystalline state (Tm.p.&80 8C), the process of replacement of the water molecules can be represented by the scheme shown in Fig. 27. In this process, the formation of the intermediate B is accompanied by the migration of the bridging chlorine atom in the complex A to the coordination site, which was previously occu- Cl C H H3O+ Cu C C2H2D Figure 27. Structural changes in the course of catalytic oligomerisation of acetylene. [CumCln(C2H2)](n7m)7+H2O.B 5 C2H2 B+ 7 D* pied by the water molecule of crystallisation (this molecule is hatched in Fig. 27). The intermediate B (more precisely, the coordination p-unit) is structurally similar to one of the p-com- plexes (20, 21a,b, 22 or 23). The reaction of the p-complex (B) with a water molecule affords the acetylenide (ethynyl) intermediate C. In this case, the Cl7 ion can be eliminated simultaneously with H3O+, but it can also be eliminated in the following quasi-equilibrium stage, the type of the kinetic equation being the same [CumCln(C2H2)](n7m)7+H2O B [CumCln71(C2H)](n7m)7+H3O++Cl7. Among a wide diversity of the structures of ethynyl com- pounds (see Section III), the complexes 43 and 44 (see Fig.13) correspond to the intermediateCwhose existence was suggested in the study 96 without convincing reasons. As mentioned above, variations in the partial pressure of acetylene (pC2H2) and [H3O+] with the maintenance of constant aCu+ and aCl7, i.e., the concentrations of the active copper chloride complex, allows one to solve the question as to what particular s-organometallic compound, which is present in the solution, serves as an intermediate in the catalytic reaction. In the synthesis of vinylacetylene, an unsymmetrical ethynyl complex serves as such an intermediate. In solution, other `yellow' complexes, for example, the symmetrical acetylenide complex [(Cu2C2)CupClq)](q7p)7 existing in equilibrium with the [(CumCln71)(C2H)](n7m)7 ion, are present in solution depending on pH.The above-mentioned inactive complex enters into the material balance with respect to [CuCl]S. Hence, the reaction rate of the formation of vinylacetylene passes through a maximum as the concentration of HCl changes if the concentration of (CumCln)(n7m)7 is not maintained constant.128 The maximum rate is observed in the range pH=1.5 ± 0.9 ([HCl]= 0.05 ± 0.15 m *). In this case, an increase in the acidity of the solution (pH<1.5) leads to destruction of symmetrical acetyle- nide and to an increase in the concentration of unsymmetrical [Cu]C2H acetylenide. A further increase in the acidity leads to the disappearance of this intermediate and to a decrease in the rate to zero.The formation of a complex containing the Cu2C2 fragment of the type 45 at pH=4 and low temperature is in complete agreement with the above reasoning and observations. H2O A X1 BMMykhalichko, O N Temkin,M G Mys'kiv C H3O+ VA EPolynuclear complexes of copper(I) halides: coordination chemistry and catalytic transformations of alkynes Evidently, the subsequent stage of the Nieuwland process (stage 2) giving rise to the intermediate X1 involves the insertion of acetylene at the Cu7C2H bond to form the CuCH=CHC:CH fragment whose acidolysis affords C4H4. Below are represented the stages of the conversion of the intermediate C into the intermediate X1 (see Fig. 27). [CumCln71(C2H)](n7m)7+C2H2 C[CumCln71(C2H2)(C2H)](n7m)7+H2O, D [CumCln71(C4H3)](n7m)7.[CumCln71(C2H2)(C2H)](n7m)7+H2O D X1 The stage of acidolysis of the intermediate X1 (the electrophilic replacement of the Cu atom by theH+ ion in the Cu7CH=frag- ment) is beyond question. This stage was confirmed by the kinetic data [Eqn (11)]. The reaction proceeds very readily even with water.136, 137 Stages of this type were also observed in hydration, hydrochlorination and hydrocyanation of acetylene.13 [CumCln71(C4H3)](n7m)7+H3O+ X1 [CumCln71(CH2=CHC:CH)](n7m+1). E The intermediate E decomposes to form vinylacetylene (it can be replaced by acetylene) or is involved in trimerisation of acetylene to give hexa-1,5-dien-3-yne [divinylacetylene (DVA)]. The insertion of acetylene giving rise to the intermediate X1 is the most important and intricate stage of the process.The fact of insertion is beyond a doubt because the reactions of cuprates R2Cu .MgX and RCu.MgX2 with alkynes are well known 136, 137 and proceed stereoselectively (100% syn-addition) under mild conditions. However, the question about the mechanism of the process of insertion remains open. In the mononuclear case, the stage of insertion CuCH=CHR HC:CH Cu7R is not an elementary concerted process because it is symmetry- forbidden due to the d 10 configuration and sp3 hybridisation of the atomic orbitals of Cu(I).138 Three alternative paths are formally possible. 1. The formation of the metallocarbenium ion (F) followed by the rapid migration of the ethynyl group giving rise to the intermediate X1 in the first irreversible stage.H +CH C D Cu 7 X1 CH C Cu F 2. Dissociation of the [Cu](m-Z1-C:C) bond to form the contact ionic pair followed by the addition of the carbanion to the Z2-coordinated acetylene. Cu+ 7C CH C C D X1 H H Cu 3. Dissociation of the bridging [Cu](m-Z1-C:C) bond accom- panied by the conversion of the ethynyl group into the Z2-state 979 followed by the addition of the Z2-HC:C7 carbanion to Z2-alkyne. Cu+ H H C C 7 D X1 C Cu CH Path 1 was considered 13 for the reactions of cationic Pt(II) complexes with alkynes. This path was assumed13 for reactions with the participation of Hg2+, but it is less probable for Cu(I) halides.Path 3 is more interesting because a softer nucleophile and a weak base (Z2-HC:C7) are involved in the reaction. In addition, paths 2 and 3 provide a more adequate explanation for the high stereoselectivity of the insertion of alkynes. To discuss these alternatives, let us use the data on the structures of the complexes (PyH)4Cu9Cl12(C:CH) (44) and NH4Cu8. .Cl9(HC:CH)4 . 2/5{[Cu(H2O)2] . [CuCl2] .H2O} (23). Crystal- chemical analysis demonstrated that these compounds are genet- ically related (Fig. 28). The subunit Cu8Cl9 (Fig. 28 b) is common to both complexes. This subunit contains four pairs of [Cu2] `matrices' with the Cu . . . Cu distance of *2.8 A. The Z2,Z2- coordination of one C2H2 molecule at the [Cu2] matrix each affords the structural fragment shown in Fig.28 c. The coordina- tion of the Z1,Z1-C:CHligand at the [Cu2] matrix gives rise to the fragment shown in Fig. 28 a. The relation between the fragments (a) and (c) suggests the possible stereochemistry of the second and third paths of conversions of the intermediate D into X1. Prob- ably, the intermediate D is formed by the replacement of the s-coordinated m4-Cl ligand in the complex with acetylene (m-Z2,Z2-C2H2). The resulting complex D (see Fig. 27) contains simultaneously several ligands coordinated at the [Cu2] matrix, viz., Z1,Z1-C:CH and Z2,Z2-C2H2. The intermediate D can be converted into X1 via two paths. From the viewpoint of the stereochemistry, the path 2 is the most favourable and involves the cleavage of the Cu7C Z1-bond to form the contact ionic pair D* (see Fig.27). The 7C2H ion rapidly attacks one of the Cd+ atoms of the p-bonded C2H2 molecule. In the case of the path 3, the Z1,Z1-C:CH ligand should be converted into the Z1,Z2-C:CH ligand. Then the Cu7C Z1-bond can be cleaved followed by the attack of the Z2-coordinated anion on the Z2-acetylene. Protolysis of the intermediate X1 affords p-coordi- nated vinylacetylene (Fig. 27, the intermediate E). b a c Figure 28. Structural fragment of the (PyH)4Cu9Cl12(C:CH) complex (44) (a), the Cu8Cl9 subunit (b) and the structural fragment of the NH4Cu8Cl9(HC:CH)4 . 2/5{[Cu(H2O)2] . [CuCl2] .H2O}complex(23)(c).980 The [CumCln(C4H4)](n7m)7 complex (E) can be deprotonated to form the intermediate G (of the type 55, see Fig.22). Judging from the kinetic data, the latter is converted into divinylacetylene. C2H2 [CumCln(C:CCH=CH2)](n7m)7 G H3O+ X3 [CumCln(C4H3)C2H2)](n7m)7 H Cu2Cl2(CH2=CHC:CCH=CH2)+(Cum72Cln72)(n7m)7. I Protolysis of the intermediateX3 (an analogue ofX1) yields the intermediate I (complex 34). The intermediate X1 can interact with yet another acetylene molecule (through the p-complex), which can be inserted at the Cu7CH= bond, because the insertion of alkynes at the Cu(I) ± alkenyl bond occurs rather readily.137 As a result, hexa- 1,3-dien-5-yne (the isomer of divinylacetylene) is formed. On certain assumptions, the scheme of its formation agrees with the kinetic data.13 3.Addition of cyclopentadiene and HCN to acetylene The Diels ± Alder reaction (12) of acetylene with cyclopentadiene to form norbornadiene J proceeds efficiently in highly concen- trated solutions of CuCl. HC CH + J The polynuclear complexes Cu5, Cu6 and Cu7 also participate in this reaction. The process involves the rate-determining stage because the reaction H3O+ X4 C7H8+[Cu] [Cu]C:CH+C5H6 proceeds more slowly than the reaction of the ethynyl intermedi- ate with C2H2. Two possible mechanisms of formation and conversion of X4 should be considered. 1. The classical diene synthesis, viz., [4+2]-cycloaddition, with the participation of an active (polarised) dienophile. [Cu] Cu H3O+ [Cu]+J C + CH 2. The insertion of diene at the Cu7C bond to form the Z3-allylic intermediate followed by the repeated intramolecular insertion of C:CH at the Cu7C bond.HC CH C C [Cu] Cu Cu H3O+ J C + CH Cu Hydrocyanation of acetylene (11, see Table 12) to form acrylonitrile proceeded in more acidic solutions of CuCl (the HCl content was up to 3 mass %). The process involves the rate- determining stage. At constant aCu+ and aCl7, the process is described by the equation (12) r11m=k11m pC2H2 pHCN . 3Oáä âH If the concentrations of the active components are not main- tained constant, the reaction rate passes through a maximum both BMMykhalichko, O N Temkin,M G Mys'kiv as pHCN and the concentration of H3O+ change.126 Although the conclusion about the participation of the free CN7 ion or a complex containing the *CuCN fragment cannot be made because of the presence of the rate-determining stage, the exper- imental data give grounds to suggest that the the mechanism is analogous to that of dimerisation of acetylene.The concentration of free cyanide ions in acidic solutions of CuCl cannot be higher than 1079 ± 10710 mol litre71 due to which the attack of the free cyanide ion present in solution on the p-coordinated acetylene ligand is improbable, as in the case of the HC:C7 anion in the dimerisation of acetylene. Catalysis of hydrocyanation of acetylene in concentrated solutions makes it possible to achieve high concentrations of the copper cyanide complex *CuC:N. Thus the concentration of dissolved CuCN in the CuCl ±NH4Cl ±H2O system can reach 15 mass %.If HCN is passed through these systems, different mono- and polycyanide complexes of copper, viz., [CumCln71(CN)](n7m)7 and [CumCln7x(CN)x](n7m)7, can be formed depending on the concentration of HCl. High reactivities of polynuclear cyanide complexes are also associated with their bridging structures analogous to those of the complex NH4Cu2(CN)3 .H2O139 (Fig. 29 a) or KCu2(CN)3 .H2O.140 In addition, due to the bridging structure, CN7 can be held as an anion in the coordination sphere of the (Cu+C7:NCu) complex in the course of insertion of acetylene at the Cu7CN bond. Acrylonitrile formed after acidolysis of the intermediate CuCH=CHC:NCu can be coordinated to form the p-complex of the type 2CuCl .CH2=CHCN.141 In the latter, the acrylonitrile molecule serves as the bridging Z2- (at the C=C bond) and Z1- (at the N atom) ligand. Interestingly, the CN7 ion can be replaced in the structure of the cyanide complex KCu2(CN)3 .H2O by the C2¡ 2 a x Cu(1) N(3)C(3) N(1) C(1) H(2) Cu(2) C(2) N(2) H(1) NH(4) H(3) H2O y b N(2) C(2) z Cu(2) C(4) C(3) K H3O+ Cu(1) N(1) C(1) x Figure 29. The [Cu2(CN)¡3 ]n anions in the structure of the NH4Cu2(CN)3 .H2Ocomplex 139 (a) and the [Cu2(CN)2(C:C)27]n anions in the structure of the [K(H3O)+]Cu2(CN)2(C:C) complex18 (b).Polynuclear complexes of copper(I) halides: coordination chemistry and catalytic transformations of alkynes dianion (Fig.29 b) to form the complex [K(H3O+)] . . [Cu2(CN)2(C:C)].18 4. Participation of ethynyl copper complexes in homogeneous oxidative conversions of alkynes in the CuCl ± MCl ± CuCl2 systems Many oxidative conversions of acetylene and alkynes proceed in concentrated solutions of CuCl in the presence of CuCl2.13, 142 As mentioned above, 2-chlorobut-1-en-3-yne,132, 133 trans-1,2- dichloroethylene and 1,1-dichloroethylene 134, 135 formed from acetylene in these systems. Soluble s-ethynyl derivatives are also involved in oxidative dehydrocondensation of alkynes yielding substituted diacetylenes K13, 143 and in substituting oxidative chlorination giving rise to chloroalkynes L.144 ± 147 RC:CC:CR 2CuCl2 K *CuC:CR ClC:CR L n The CuCl ± CuCl2 catalytic systems (CuCl2 is a catalyst in the presence of O2 or in an electrochemical system) are characterised by the formation of heterovalent polynuclear complexes very typical of copper.148 The structure of the simplest representative of these complexes, viz., [(C2H5)4N]Cu2Cl4, has been described.15 The Cu(I) and Cu(II) centres in the polynuclear chain (Cu2Cl4)n¡ differ in symmetry. Concentrated solutions, which enable one to achieve high concentrations of intermediates (*CuC:CR, *CuCN or *CuCH=CHX), play an important role in the catalysis of addition at the C:C bonds of the HC:CH, HCl, HOH, HCN and C5H6 molecules.In oxidative conversions of alkynes, electron transfer from CuR to CuCl2 in mixed complexes occurs so rapidly (k=0.56108 mol71 litre s71)149 that the reaction rates are high even in solutions dilute with respect to CuCl (for example, at the concentrations of LiCl, CuCl and CuCl2 equal to 5, 0.2 and 0.5 mol litre71, respectively).The kinetic equations for some reactions proceeeding in these systems under conditions of con- stant high concentration of the cations (Li+) {when [CuCl]S! [(CuCln)(n71)7] and [CuCl2]S![(CuCln)(n72)7]} were obtained. Thus, the rate of oxidative dimerisation of methylacetylene (MA) is described by the equation 142 (13) 3Oáä rDDA=kDDA pMAâCuCl2äâCuClä2 , âH where DDA is dimethyldiacetylene. The process occurs through the intermediate CuC:CMe and involves the rate-determining stage whose transition state includes Cu(I) and Cu(II) ions [2Cu(I) .Cu(II)(7C:CR)]#. It should be noted that two CuCl molecules enter into in the reaction even in rather dilute solutions. Apparently, these molecules are required for stabilisation of the assumed intermediate, viz., the ethynyl complex of copper(II) (ClCuC:CMe). The kinetic data 142 dem- onstrate that two molecules of the intermediate decompose to form dialkyne MeC:CC:CMe+6 CuCl. 2 ClCuC:CMe. 2 CuCl M It may be suggested that the ethynyl group in the M complex acts as the m3-Z1,Z2,Z2-ligand. It is also conceivable that two CuCl molecules act as catalysts in the formation of ClCuIIC:CMe ClCuC:CMe+2CuCl, ClCuC:CMe. 2CuCl MeC:CC:CMe+2CuCl. 2ClCuC:CMe The reaction of *CuIC:CMe with two CuCl2 molecules affords 1-chloroprop-1-yne (CP).The latter is formed in two stages the first of which is rate-determining. 981 2 CuC:CMe+2CuCl CuCl2(ClCuIIC:CMe)+CuCl, ClC:CMe+2 CuCl. CuCl2(ClCuIIC:CMe) The reaction rate is described by the equation (14) rCP=kCP pMAâCuCl2ä2âCuClä . âH3Oáä It should be noted that heterovalent copper complexes with organic ligands (of the s- and p-types) are as yet unknown although the complex 15 may be assigned to heterovalent organo- metallic complexes because its structure contains Cu2+ ions occupying octahedral cavities and CuC:CH species as `guests' along with the `host' complexes. V. Conclusion Analysis of the experimental data obtained in studies of the compositions and structures of the complexes M(n7m)CumCln in solutions and in the solid phase, copper(I) p-complexes with acetylene and its derivatives and s-ethynyl compounds, and information gained from investigations of the kinetics of the catalytic conversions of alkynes in the MCl ± CuCl ±H2O and MCl ± CuCl ± CuCl2±H2O systems reveals the specific features of the coordination chemistry of copper(I) halide complexes and the peculiarities of the behaviour of copper(I) as a catalyst of conversions of acetylenic compounds.The formation of coordination polyhedra through bridging chloride (bromide) ligands in which Cl7 acts as the m2-, m3- or m4-ligand gives rise to low-charged polynuclear associates (CumCln)(n7m)7, which can efficiently interact with acetylene and alkynes to form various p- and s-organometallic compounds.The geometric characteristics of the p-coordination polyhedra about the Cu(I) atoms and of the triple C:C bond p-activated at the metal centre in the CuCl p-complexes with acetylene and its derivatives and the higher efficiency of the Cu(I)7(C:C) p-inter- action compared to that of Cu(I)7(C=C) are consistent with the statement that two degenerate p-MOs are split into two MOs with different energies (pk and p\) in the course of p-bonding with Cu(I) due to which the reactivities of alkynes increase. Using the copper(I) halide p-complexes of organic compounds containing the C:C bond as an example, two qualitatively different modes of p-coordination of the C:C groups to the Cu(I) atoms were revealed. These modes differ in the efficiency of the p-interaction.Thus, the first mode of coordination occurs in the case of the probable simultaneous participation of pk-, pk and p\ -MOs to form the Cu(I)7(C:C) p-unit. The second mode of coordination occurs if two individual p-MOs of theC:C bond interact with the Cu(I) atom idependently of one another giving rise to the bridging Cu(I)7(C:C)7Cu(I) p-unit (Z2,Z2-HC:CR). In the latter case, each p-MO of the acetylene group behaves analogously to that of the ethylene group and is independently coordinated to one metal atom. The participation of the higher pk-MO and two p*-MOs is favourable for an increase in the contribution both of the donor- acceptor and dative components of the bond due to which Cu(I)7alkyne bonding in the first case is more efficient.In the crystalline anionic p-complexes of CuCl with acetylene and its monosubstituted derivatives, the efficiency of the Cu(I)7(C:C) p-interaction, which is manifested in the elonga- tion of the p-coordinated C:C bond, is substantially affected by :C H. . . Cl hydrogen bonds (the length of the p-coordinated C:C bond depends linearly on the H. . . Cl distance). The reason is that the Cl atoms whose electron-withdrawing properties depend substantially on the strength of the M+. . . Cl7 ionic interaction can affect the polarisation of the:C7H bond of the p-coordinated molecule of terminal alkyne through the C(H) . . . Cl contact due to which the electron density on the carbon atoms can increase or decrease resulting in the corresponding decrease or increase of the contribution of the dative transfer982 Cu(I)?(HC:CR).In this case, the C:C bond is subjected to an insignificant or substantial elongation, respectively. The high efficiency of p-interactions in copper(I) complexes of the ethynyl type (s,p-complexes) whose formation in the MCl ± CuCl ±H2O systems (according to the kinetic data) plays a decisive role in many conversions of C2H2 and alkynes is associated with the mutual strengthening of individual types of interactions, viz., of Cu(I)7(Z1-C:CR) and Cu(I)7 (Z2-C:CR). This effect of mutual strengthening (synergism) of acetylenide and p-types of interactions, which was revealed in cluster acetylenides and ethynyl complexes of Cu(I), is often accompanied by unusual structural manifestations and by the appearance of qualitatively new properties of the C:C group p-activated at the metal centre [the 7C:C7 dianion in the compound 45 can be p-coordinated simultaneously to three Cu(I) atoms].In addition to the modes of coordination of ethynyl groups (C:CR) to the copper(I) atom (A, B, C and D) known in organometallic chemistry,96 ± 117 new types of clusters (E and F) were found in the case of copper chloride complexes. R R R C C C C C C M M M M M M M m3-Z1,Z1,Z1 (C) m-Z1,Z2 (B) m-Z1,Z1 (A) R R R M C C C M M M C C C M M M M M M M M m5-Z1,Z1,Z1,Z2,Z2 (F) m4-Z1,Z1,Z2,Z2 (E) m3-Z1,Z2,Z2 (D) The polydentate character of coordination of acetylene in p-complexes and of ethynyl ligands in s-organometallic com- pounds (as well as of cyanide groups) is the governing factor responsible for the catalytic activity of polynuclear Cu(I) com- plexes, which is characterised by the following features.1. The high rate of cleavage of the bridging Cu7Cl (m,m3,m4- Cl)-bonds to form the initial p-complexes (in some cases, the m-Cl atom is isolobally replaced by alkyne). 2. The high solubility of such molecules as Cu:CH, CuC:CR and CuCN due to reactions with polynuclear (CumCln)(n7m)7anions and, consequently, the high concentration of intermediates in solutions (the concentration of `yellow' acety- lenide complexes in the Nieuwland reaction reaches 0.1 mol litre71). 3. The bridging nature of the coordinated ethynyl and cyanide groups provides various stereochemical possibilities for the appearance of carbanionic (RC:C7) and cyanide (N:C7) nucleophiles in the complex.4. The tendency for cluster-formation in polynuclear s,p- ethynyl complexes has a particular effect on the course of catalytic processes. 5. All stages of the process, including coordination of the products, can occur on the `matrix' of one complex formed from the [Cun] cluster. 6. The formation of polynuclear heterovalent complexes in oxidation reactions is favourable for the electron transfer in multiple-electron processes. 7. Hydration and hydrochlorination of alkynes in which the s- organometallic intermediate is formed in the course of the attack of the nucleophile (Cl7orH2O), which is present in the solution of the copper(I) p-complex, can also occur without the participation of polynuclear complexes and, correspondingly, need not be catalysed by concentrated solutions.BMMykhalichko, O N Temkin,M G Mys'kiv References 1. M Calvin Trans. Faraday Soc. 34 1181 (1931) 2. B R James Homogeneous Hydrogenation (New York: Wiley, 1973) 3. J A Nieuwland, R Fogt Khimiya Atsetilena (The Chemistry of Acetylene) (Translated into Russian; Moscow: Izd. Inostr. Lit., 1947) 4. O N Temkin, R M Flid Kataliticheskie Prevrashcheniya Atsetilenovykh Soedinenii v Rastvorakh Kompleksov Metallov (Catalytic Conversions of Acetylene Compounds in Solutions of Metal Complexes) (Moscow: Nauka, 1968) 5.M Chavastelon C.R. Acad. Sci. 126 1810 (1898) 6. M Chavastelon C.R. Acad. Sci. 130 1764 (1900) 7. M Herberhold Metal p-Complexes (Amsterdam: Elsevier, 1972) 8. O A Chaltykyan Kuprokataliz (Cuprocatalysis) (Erevan: Aipetrat, 1963) 9. S A Vartanyan Khimiya Vinilatsetilena i Ego Proizvodnykh (The Chemistry of Vinylacetylene and Its Derivatives) (Erevan: Academy of Sciences Arm. SSR, 1966) 10. J A Nieuwland,W S Calcott, F B Downing, A S Carter J. Am. Chem. Soc. 53 4197 (1931) 11. O A Chaltykyan Zh. Obshch. Khim. 18 1626 (1948) 12. I M Dolgopol'skii, A L Klebanskii, Z F Dobler Zh. Obshch. Khim. 33 1074 (1963) a 13. O N Temkin, G K Shestakov, Yu A Treger Atsetilen. Khimiya. Mekhanizmy Reaktsii. Tekhnologiya (Acetylene. Chemistry. Mechanisms of Reactions.Technology) (Moscow: Khimiya, 1991) 14. G W Parshal, S D Ittel Homogeneous Catalysis (New York: Wiley, 1992) 15. R D Willett Inorg. Chem. 26 3423 (1987) 16. M G Mys'kiv, Doctoral Thesis in Chemical Sciences, Institute of Organic and Inorganic Chemistry, Academy of Sciences of the USSR, Moscow, 1991 17. S I Osechkin, Candidate Thesis in Chemical Sciences, Lvov State University, Lvov, 1991 18. B M Mykhalichko, Doctoral Thesis in Chemical Sciences, Lvov National University, Lvov, 1999 19. M G Mys'kiv, B M Mykhalichko Zh. Strukt. Khim. 35 (5) 120 (1994) b 20. S I Osechkin,M G Mys'kiv, P Yu Zavalii, A N Sobolev Metalloorg. Khim. 4 997 (1991) c 21. I S Morozov, G V Ustavshchikova Izv. Acad. Nauk SSSR, Otd. Khim. Nauk 451 (1944) 22.T G Sukhova, Candidate Thesis in Chemical Sciences, Moscow Institute of Fine Chemical Technology, Moscow, 1969 23. O N Temkin, R M Flid, T G Sukhova, E G Chepaikin, G F Tikhonov Zh. Prikl. Khim. 41 633 (1968) d 24. T G Sukhova, O N Temkin, R M Flid Zh. Neorg. Khim. 15 1849 (1970) e 25. D S McClure Solid State Phys. 9 399 (1959) 26. S Ahrland, J Rawsthorne Acta Chem. Scand. 24 157 (1970) 27. A A Noyes, M Chow J. Am. Chem. Soc. 40 739 (1918) 28. W W Latimer The Oxidation States of the Elements and their Potentials in Aqueous Solutions (New York: Prentice-Hall, 1952) 29. T Hurlen Acta Chem. Scand. 15 1231 (1961) 30. T G Sukhova, O N Temkin, R M Flid, T K Kaliya Zh. Neorg. Khim. 13 2073 (1968) e 31. T G Sukhova, N Ya Borshch, O N Temkin, R M Flid Zh.Neorg. Khim. 14 694 (1969) e 32. L J Andrews, R M Keefer J. Am. Chem. Soc. 71 2379 (1949) 33. T G Sukhova, O N Temkin, R M Flid Zh. Neorg. Khim. 14 928 (1969) e 34. F G C Rossotti, H Rossotti The Determination of Stability Constants and Other Equilibrium Constants in Solution. (New York: McGraw-Hill, 1961) 35. B HedstroÈ m Acta Chem. Scand. 9 613 (1955) 36. R M Flid, I I Moiseev Zh. Prikl. Khim. 27 1145 (1954) 37. I I Moiseev, Candidate Thesis in Chemical Sciences, Moscow Institute of Fine Chemical Technology, Moscow, 1958 38. G F Tikhonov, O N Temkin, R M Flid Zh. Fiz. Khim. 40 3075 (1966) f 39. B M Mykhalichko, M G Mys'kiv Koord. Khim. 25 533 (1999) g 40. M G Mys'kiv, S I Osechkin, P Yu Zavalii, V S Fundamenskii Koord.Khim. 14 524 (1988) gPolynuclear complexes of copper(I) halides: coordination chemistry and catalytic transformations of alkynes 41. C Brink, C H McGillavary Acta Crystallogr. 2 158 (1949) 42. C Brink Struct. Rep. 15 171 (1951) 43. C Brink, A E van Arkel Acta Crystallogr. 5 506 (1952) 44. C Brink, N F Binnendijk, J van de Linde Acta Crystallogr. 7 176 (1954) 45. B M Mykhalichko Zh. Neorg. Khim. 43 1840 (1998) e 46. B M Mykhalichko, V N Davydov, L G Aksel'rud Zh. Neorg. Khim. 42 1125 (1997) e 47. M Asplund, S Jagner Acta Chem. Scand., Ser. A 39 47 (1985) 48. B M Mykhalichko, L G Aksel'rud, V N Davydov Zh. Neorg. Khim. 42 410 (1997) e 49. B M Mykhalichko, M G Mys'kiv, P Yu Zavalii,M D Mazus, S I Osechkin Koord. Khim. 17 827 (1991) g 50.S Geller, G R Akridge, S A Wilber Struct. Rep. 46A 186 (1980) 51. B M Mykhalichko,M G Mys'kiv Zh. Strukt. Khim. 40 1160 (1999) b 52. B M Mykhalichko, M G Mys'kiv Koord. Khim. 24 596 (1998) g 53. S Andersson, S Jagner Acta Chem. Scand., Ser. A 43 39 (1989) 54. B M Mykhalichko, T Glovyak, M G Mys'kiv Zh. Neorg. Khim. 40 757 (1995) e 55. B M Mykhalichko, K Gelzak-Kots'vin, M G Mys'kiv Zh. Strukt. Khim. 41 559 (2000) b 56. M P Shaskol'skaya Kristallografiya (Crystallography) (Moscow: Nauka, 1984) 57. L P Shevardina, V M Pinchuk Zh. Strukt. Khim. 30 (4) 17 (1989) b 58. O N Temkin, T G Sukhova, G K Shestakov, G F Tikhonov, R M Flid, E G Chepaikin, G M Novikova, N Ya Borshch Kinet. Katal. 10 1004 (1969) h 59. B M Mykhalichko, M G Mys'kiv, L G Aksel'rud Koord.Khim. 19 722 (1993) g 60. B M Mykhalichko Ukr. Khim. Zh. 64 16 (1998) 61. B M Mykhalichko, M G Mys'kiv, V N Davydov Zh. Neorg. Khim. 44 411 (1999) e 62. K Brantin, M HaÈ kansson, S Jagner J. Organomet. Chem. 474 229 (1994) 63. M HaÈ kansson, K WettstroÈ m, S Jagner J. Organomet. Chem. 421 347 (1991) 64. B M Mykhalichko, M G Mys'kiv, E A Goreshnik Koord. Khim. 25 65 (1999) g 65. B M Mykhalichko, M G Mys'kiv Zh. Neorg. Khim. 44 1661 (1999) e 66. B M Mykhalichko, M G Mys'kiv Koord. Khim. 24 939 (1998) g 67. B M Mykhalichko, M G Mys'kiv, M D Mazus Ukr. Khim. Zh. 59 690 (1993) 68. B M Mykhalichko,M G Mys'kiv, P Yu Zavalii, V S Fundamenskii Koord. Khim. 16 512 (1990) g 69. B M Mykhalichko Ukr. Khim.Zh. 62 39 (1996) 70. M G Mys'kiv, P Yu Zavalii, B M Mykhalichko, V S Fundamenskii Koord. Khim. 14 1619 (1988) g 71. B M Mykhalichko, M G Mys'kiv, L G Aksel'rud, M D Mazus Koord. Khim. 18 985 (1992) g 72. B M Mykhalichko, V K Pecharskii, M G Mys'kiv Koord. Khim. 19 544 (1993) g 73. G G Aleksandrov, I R Gol'ding, S R Sterling, A M Sladkov, Yu T Struchkov, I A Garbuzova, V T Aleksanyan Izv. Akad. Nauk SSSR, Ser. Khim. 2679 (1980) i 74. G Schmidt, N Schittenhelm, U Behrens J. Organomet. Chem. 496 49 (1995) 75. B M Mykhalichko, M G Mys'kiv, V N Davydov Zh. Neorg. Khim. 44 46 (1999) e 76. B M Mykhalichko Zh. Neorg. Khim. 43 1639 (1998) e 77. B M Mykhalichko Koord. Khim. 25 359 (1999) g 78. B M Mykhalichko, M G Mys'kiv Zh. Neorg. Khim.44 1127 (1999) e 79. B M Mykhalichko, M G Mys'kiv Koord. Khim. 25 461 (1999) g 80. M J S Dewar Bull. Soc. Chim. Fr. C71 (1951) 81. A K Barnard Theoretical Basis of Inorganic Chemistry (New York McGraw-Hill, 1965) 82. M J S Dewar,RC Dougherty ThePMOTheory of Organic Chemistry (New York: Plenum Press, 1975) 83. P H Kasai, D McLeod Jr , T Watanabe J. Am. Chem. Soc. 102 179 (1980) 84. M Sodupe, C W Bauschlicher Jr J. Phys. Chem. 95 8640 (1991) 85. M BoÈ hme, T Wagener, G Frenking J. Organomet. Chem. 520 31 (1996) 86. A Kov'acs, G Frenking Organometallics 18 887 (1999) 87. M G Mys'kiv, V V Oliinik Koord. Khim. 21 290 (1995) g 983 88. M G Mys'kiv, V V Oliinik, B M Mykhalichko, E A Goreshnik, in Pratsi NTSh (Proceedings of T Shevchenko Scientific Society) (Lvov: NTSH, 1997) Vol.1, p. 201 89. J E Huheey Inorganic Chemistry. Principles of Structure and Reactivity (New York: Happer and Row, 1983) 90. Khimiya Atsetilenovykh Soedinenii (The Chemistry of Acetylene Compounds) (Moscow: Khimiya, 1973) 91. K Tatsumi, R Hoffman, J L Templeton Inorg. Chem. 21 466 (1982) 92. G R Desiraju Acc. Chem. Res. 24 290 (1991) 93. Z Berkovitch-Yellin, L Leizerovitz Acta Crystallogr., Ser. B 40 159 (1984) 94. R Taylor, O Kennard J. Am. Chem. Soc. 104 5063 (1982) 95. O N Temkin, R M Flid, G K Shestakov, A Ermakova, G F Tikhonov, D I Yarovaya, V G Mykhal'chenko Kinet. Katal. 10 1230 (1969) h 96. R Vestin Svensk. Kem. Tidskr. 66 65 (1954) 97. K Osakada, T Takizawa, T Yamamoto Organometallics 14 3531 (1995) 98. MPilar Gamasa, J Gimeno, E Lastra, A Aguirre, S Garcia-Granda J. Organomet. Chem. 378 C11 (1989) 99. J Diez, M Pilar Gamasa, J Gimeno, E Lastra, A Aguirre, S Garcia-Granda Organometallics 12 2213 (1993) 100. J Diez, M Pilar Gamasa, J Gimeno, A Aguirre, S Garcia-Granda Organometallics 10 380 (1991) 101. V W-W Yam, W-K Lee, K-K Cheung, B Crystall, D Phillips J. Chem. Soc., Dalton Trans. 3283 (1996) 102. D M Knotter, A L Spek, G van Koten J. Chem. Soc., Chem. Commun. 1738 (1989) 103. DMKnotter, A L Spek,DMGrove,Gvan Koten Organometallics 11 4083 (1992) 104. M D Janssen, J G Donkervoort, S B van Berlekom, A L Spek, D M Grove, G van Koten Inorg. Chem. 35 4752 (1996) 105. R W M ten Hoedt, J G Noltes, G van Koten, A L Spek J. Chem. Soc., Dalton Trans. 1800 (1978) 106. MG B Drew, F S Esho, SMNelson J. Chem. Soc., Chem. Commun. 1347 (1982) 107. V W-W Yam, S W-K Choi, C-L Chan, K-K Cheung J. Chem. Soc., Chem. Commun. 2067 (1996) 108. D L Reger, J E Collins, M E Huff, A L Rheingold, G P A Yap Organometallics 14 5475 (1995) 109. V W-W Yam, W-K Lee, K K Cheung, H-K Lee, W-P Leung J. Chem. Soc., Dalton Trans. 2889 (1996) 110. A J Edwards,MA Paver, P R Raithby, M-A Rennie, C A Russell, D S Wright Organometallics 13 4967 (1994) 111. V W-W Yam, W-K Lee, T-F Lai Organometallics 12 2383 (1993) 112. L Naldini, F Demartin,M Manassero, M Sansoni, G Rassu, M A Zoroddu J. Organomet. Chem. 279 C42 (1985) 113. M Pilar Gamasa, J Gimeno, E Lastra, X Solaus J. Organomet. Chem. 346 277 (1988) 114. V W-W Yam, W-K Lee, K-K Cheung J. Chem. Soc., Dalton Trans. 2335 (1996) 115. F Olbrich, J Kopf, E Weiss Angew. Chem., Int. Ed. Engl. 32 1077 (1993) 116. P W R Corfield, H M M Shearer Acta Crystallogr. 21 957 (1966) 117. V W-W Yam,W K-M Fung, K-K Cheung Angew. Chem., Int. Ed. Engl. 35 1100 (1996) 118. O N Temkin, Doctoral Thesis in Chemical Sciences, Moscow Institute of Fine Chemical Technology, Moscow, 1972 119. G K Shestakov, A Ermakova, V G Mykhal'chenko, O N Temkin, R M Flid Zh. Fiz. Khim. 44 406 (1970) f 120. G K Shestakov, A V Massal'skaya, S M Airyan, O N Temkin Kinet. Katal. 18 399 (1977) h 121. G K Shestakov, F I Bel'skii, S M Airyan, O N Temkin Kinet. Katal. 19 334 (1978) h 122. O N Temkin, G K Shestakov, R M Flid Dokl. Akad. Nauk SSSR 184 362 (1969) j 123. N Yu Vsesvyatskaya, G K Shestakov, O N Temkin Kinet. Katal. 27 1330 (1986) h 124. V S Vartanyan, G K Shestakov, O N Temkin Arm. Khim. Zh. 32 259; 264 (1979) 125. G K Shestakov, Doctoral Thesis in Chemical Sciences, Moscow Institute of Fine Chemical Technology, Moscow, 1989 126. A A Khorkin, A Ermakova, O N Temkin, R M Flid, V S Kozina Kinet. Katal. 14 336 (1973) hBMMykhalichko, O N Temkin,M G Mys'kiv 984 127. O N Temkin, G K Shestakov, N Yu Kozlova Kinet. Katal. 31 850 (1990) h 128. G F Tikhonov, O N Temkin, R M Flid Kinet. Katal. 8 520 (1967) h 129. O N Temkin, G F Tikhonov, R M Flid, Yu R Galiev Kinet. 130. G K Shestakov, G F Tikhonov, O N Temkin, R M Flid 131. G K Shestakov, G F Tikhonov, O N Temkin, R M Flid, Katal. 8 1236 (1967) h Kinet. Katal. 11 575 (1970) h I Sh Gershenzon, S M Brailovskii, S I Dolzhnikova Kinet. Katal. 11 875 (1970) h 132. S M Brailovskii, O N Temkin, R M Flid, N G Belova Kinet. Katal. 11 107; 1346 (1970) h 133. S M Brailovskii, O N Temkin, R M Flid Kinet. Katal. 12 1152 (1971) h 134. S M Brailovskii, O N Temkin, A S Kostyushin, K Yu Odintsov Kinet. Katal. 31 1371 (1990) h 135. S M Brailovskii, Huin Man Hoan, O N Temkin Kinet. Katal. 35 734 (1994) h 136. J F Normant Synthesis 63 (1972)] 137. J F Normant, A Alexakis Synthesis 841 (1981) 138. Ya A Dorfman Zhidkofaznyi Kataliz (Orbital'noe Modelirovanie) [Liquid-Phase Catalysis (Orbital Modelling)] (Alma-Ata: Nauka, 1981) 139. B M Mykhalichko, M G Mys'kiv, L G Aksel'rud Zh. Neorg. Khim. 41 75 (1996) e 140. D T Cromer, A C Larson Struct. Rep. 27 416 (1962) 141. M Massaux, M-T Le Bihan, R Chevalier Acta Crystallogr., Ser. B 33 2084 (1977) 142. S M Brailovskii, O N Temkin Zh. Vsesoyuz. Khim. O-va im. D I Mendeleeva 30 331 (1985) k 143. Hyunh Manh Hoan, S M Brailovskii, O N Temkin Kinet. Katal. 35 266; 367 (1994) h 144. L V Shchel'tsin, S M Brailovskii, May Suan Tin, O N Temkin Kinet. Katal. 29 581 (1988) h 145. L V Shchel'tsin, S M Brailovskii, E Yu Murugova, O N Temkin Kinet. Katal. 29 1044 (1988) h 146. L V Shchel'tsin, S M Brailovskii, O N Temkin Kinet. Katal. 31 469 (1990) h 147. L V Shchel'tsin, S M Brailovskii, O N Temkin Kinet. Katal. 31 1361 (1990) h 148. M Dunaj-Iurco, G Ondrejovic,M Melnik Coord. Chem. Rev. 83 1 (1988) 149. H M McConnell, H J Davidson J. Am. Chem. Soc. 72 3168 (1950) a�Russ. J. Gen. Chem. (Engl. Transl.) b�Russ. J. Struct. Chem. (Engl. Transl.) c�Russ. J. Organomet. Chem. (Engl. Transl.) d�Russ. J. Appl. Chem. (Engl. Transl.) e�Russ. J. Inorg. Chem. (Engl. Transl.) f�Russ. J. Phys. Chem. (Engl. Transl.) g�Russ. J. Coord. Chem. (Engl. Transl.) h�Kinet. Catal. (Engl. Transl.) i�Russ. Chem. Bull. (Engl. Transl.) j�Dokl. Chem. Technol., Dokl. Chem. (Engl. Transl.) k�Mendeleev Ch

ISSN:0036-021X    

出版商:RSC    

年代:2000

数据来源: RSC

摘要:

Russian Chemical Reviews 69 (11) 985 ± 999 (2000) Reactions of nitrogen oxides with polymers { G B Pariiskii, I S Gaponova, E Ya Davydov Contents I. Introduction II. Reactions of polymers with nitrogen dioxide III. The effect of nitric oxide on polymers IV. Conclusion Abstract. and oxides nitrogen of reactions the of mechanisms The The mechanisms of the reactions of nitrogen oxides and different considered. are polymers solid of classes different classes of solid polymers are considered. Particular Particular emphasis of mechanisms the of analysis the to given is emphasis is given to the analysis of the mechanisms of the the formation of stable nitroxyl radicals. Double bonds and amide formation of stable nitroxyl radicals. Double bonds and amide groups of macromolecules, as well as hydroperoxides and per- groups of macromolecules, as well as hydroperoxides and per- oxide reactions the in involved be to shown are macroradicals oxide macroradicals are shown to be involved in the reactions with with nitrogen for oxides nitrogen of application The oxides. nitrogen oxides.The application of nitrogen oxides for the the preparation the of use the and polymers spin-labelled of preparation of spin-labelled polymers and the use of the ESR ESR imaging of investigation the for tomography) (ESR technique imaging technique (ESR tomography) for the investigation of the the structure of the reaction front during nitration of solid polymers structure of the reaction front during nitration of solid polymers are references 111 includes bibliography The considered.are considered. The bibliography includes 111 references. I. Introduction Investigations aimed at finding relations between the processes of thermal, thermal-oxidative, photolytic, photo-oxidative and mechanochemical ageing of different classes of polymers and destruction effects, as well as studies on stabilisation of polymers to withstand these effects have been carried out for a long time. The main goal of these investigations is to retain the operating characteristics of polymeric materials. Among numerous publications devoted to generalisation of the results obtained in this area of scientific exploration (see, e.g., Refs 1 ± 20), monographs by Emanuel and Buchachenko 2, 4 can be highlighted.In these studies, the kinetics of the ageing of polymers under the action of different factors have been analysed taking account of specific features of elementary acts of the corresponding solid-phase chemical transformations. In the mid-twentieth century, it was recognised that the increasing amount of industrial waste discharged to the atmos- phere pollutes it with such aggressive compounds as sulfurous anhydride, different nitrogen oxides and ozone. These give rise to acid rains and smogs,21 which cause deterioration of structural materials (concrete, marble, bricks, polymeric materials, lacquers and paints). Reactions of polymers with nitrogen oxides constitute an important factor responsible for deterioration of the functions of polymers. Studies on nitration of organic compounds both in the G B Pariiskii, I S Gaponova, E Ya Davydov NMEmanuel Institute of Biochemical Physics, Russian Academy of Sciences, ul.Kosygina 4, 117977 Moscow, Russian Federation. Fax (7-095) 137 41 01. Tel. (7-095) 939 71 03 (G B Pariiskii, I S Gaponova), (7-095) 939 74 46 (E Ya Davydov) Received 11 July 2000 Uspekhi Khimii 69 (11) 1071 ± 1086 (2000); translated by AMRaevsky #2000 Russian Academy of Sciences and Turpion Ltd DOI 10.1070/RC2000v069n11ABEH000611 985 986 994 997 liquid and gas phases began in the mid-nineteenth century. Both nitric acid and nitrogen oxides were used as nitrating agents. The nitration of aromatic compounds is, as a rule, an ionic proc- ess,22 ± 24 whereas that of the paraffin hydrocarbons is a free- radical process initiated by nitrogen dioxide, NO2.25 ± 27 In this case, nitric acid served as a source of nitrogen oxides. The addition of small amounts of nitrogen dioxide initiates degenerate branched chain oxidation of hydrocarbons in the liquid phase and, on the other hand, results in chain termination of oxidation reactions.28 Systematic studies of the effect of nitrogen oxides on the stability of polymeric materials began in the late 1950s.29 ± 32 The major contribution to the solution of this problem was made by Jellinek who studied degradation of different classes of polymers in an atmosphere of nitrogen dioxide and its dimer at ambient temperatures and under the combined action of nitrogen dioxide, oxygen and near-UV light and visible light.32 Eight oxides of nitrogen are known; however, only three of them, viz., NO, NO2 and N2O4, are thought to be air pollutants.33 The atmospheric concentrations of other nitrogen oxides are too low to have a pronounced effect on the ageing of polymers under natural conditions.The concentrations of nitrogen oxides in the atmosphere are subject to diurnal and seasonal variations over a rather wide range and depend on the position and weather conditions. The max- imum permissible concentrations of nitric oxide in urban atmos- pheres are 261078 mol litre71 (the peak concentration) and 261079 mol litre71 (the daily average concentration). The max- imum permissible concentration of nitrogen dioxide is 1.8561079 mol litre71 in both cases.In the production areas, the permissible concentration of NO2 is about 1.161077 mol litre71 (Refs 34, 35). Nitrogen dioxide always exists in equilibrium with its colour- less linear dimer, N2O4. The heat of formation of N2O4 is 761.4 kJ mol71 at 298 K. 2NO2 O2N7NO2 . The equilibrium constant decreases with the increase in the total pressure of NO2+N2O4 (P) and is 0.1415 atm at P=4.361072 atm and T=298 K.36 Based on experimental and calculated data,37 ± 40 it was assumed that nitrosation of alkenes in non-polar solvents is the result of the reaction of carbon ± carbon double bonds with nitroso nitrate, ON7ONO2, a less stable isomer of linear N2O4. { Dedicated to the memory of Academician NMEmanuel (1915 ± 1984).986 Some additional remarks are to be made before analysing the results of investigations on the reactions of polymers with nitrogen oxides and establishing the possibility of extrapolating them in order to assess the ageing of polymers in a polluted atmosphere.Since the atmospheric concentrations of nitrogen oxides are low, the rates of their reactions are also low. Therefore, the ageing processes due to other factors (e.g., thermal oxidation, photo- initiated oxidation, ozone ageing, biodegradation of polymers etc.) can compete with or be more efficient than the reactions involving nitrogen oxides. Therefore, the results of tests under natural conditions can hardly be used for evaluation of the contribution of the processes initiated by nitrogen oxides.These reactions are usually studied at the concentrations of air pollutant gases which are several orders of magnitude higher than their atmospheric concentrations. The results obtained under these conditions are, as a rule, linearly extrapolated to the atmospheric concentrations. Such a priori extrapolation is not obvious, since the contributions of particular stages to the overall ageing process under the conditions of accelerated tests can vary depending on the experimental conditions. Ageing of polymers exposed to nitrogen oxides is due to the reactions of gases diffusing into the polymer with reactive groups of macromolecules. These processes are characterised by non- uniform propagation of the reaction through the bulk of the sample and require complicated description of their kinetics.`Kinetic inequivalence' of reactants is typical of the chemical reactions in solid polymers.2, 4 This means that the species with the same chemical structure have different nearest environments and therefore react with different rate constants. As a consequence, the most reactive species are consumed rapidly so that the overall rate constants decrease with time. However, relaxation processes occurring in polymers favour the recovery of the initial distribu- tion of reactivities of the species. Therefore, the reaction kinetics appears to be dependent on the ratio of the rates of chemical reactions and relaxation processes.41 This also requires a critical view on simple transfer of the results of accelerated tests to the ageing of polymers in atmospheres polluted with nitrogen oxides.Finally, it should be noted that most of the studies of the effect of nitrogen oxides on polymeric materials involve the determina- tion of changes in the average degree of polymerisation. However, this parameter characterises only one reaction, viz., degradation of macromolecules, though this is of prime importance.Adetailed study of degradation in reactions of various polymers with nitro- gen dioxide was carried out by Jellinek.32 He has divided all polymers into three main groups differing in stability towardNO2. In the present review, the results will also be analysed using this classification.Changes in the mechanical properties can be considered as important operating characteristics of a polymer as a structural material; however, this parameter provides no direct information on the chemical transformations and structure of macromolecules. This information can be obtained using IR spectroscopy, but changes in the absorption spectra can usually be detected only at rather high degrees of conversion. Therefore, the use of other physicochemical methods for the analysis of the transformations occurring in the polymers exposed to nitrogen oxides, as well as the analysis of the distribution of the reaction over the bulk of the sample are of crucial importance for the elucidation of complicated mechanisms of these reactions. Some- times, ageing products can be more reactive towards nitrogen oxides than the initial polymer.In the present review, we considered the results obtained in studies on the ageing of different classes of polymers by traditional methods of physicochemical analysis and in the investigations of these processes by ESR spectroscopy, which allow one to draw conclusions on free-radical stages of the ageing of polymers involving nitrogen oxides from the structure of stable nitrogen- containing macroradicals produced in the reactions. G B Pariiskii, I S Gaponova, E Ya Davydov II. Reactions of polymers with nitrogen dioxide Nitrogen dioxide plays the role of a moderately reactive free radical (the strength of the ONO7H bond is 322 to 326 kJ mol71),42 which can initiate several types of reactions, e.g., hydrogen abstraction from molecules of organic compounds and reversible addition to carbon ± carbon double bonds.Addi- tionally, NO2 and its dimer can enter into specific reactions with functional groups of macromolecules. 1. Carbon-chain polymers containing no carbon ± carbon double bonds Carbon-chain polymers containing no carbon ± carbon double bonds in both the main chain and side groups [e.g., polyethylene (PE), polypropylene (PP), polystyrene (PS), poly(methyl metha- crylate) (PMMA), polyacrylonitrile (PAN), poly(vinyl chloride) (PVC) and poly(vinyl fluoride) (PVF)] are characterised by insig- nificant changes in their properties under the conditions of accelerated tests (Jellinek et al.43, 44 carried out their experiments at aNO2 concentration of 7.861073 mol litre71 and T=308 K; the exposure time was*30 h).Linear extrapolation of the results obtained to the NO2 concentration typical of polluted atmos- pheres predicts a long-term stability of the polymers.45 Let us consider the results of investigations of the effect of NO2 on particular representatives of these polymers in more detail. Pioneering studies of the reaction of nitrogen dioxide with PE and PP have been carried out by Ogihara et al.30, 31 Using IR spectroscopy, they have found that nitrogen dioxide cannot abstract secondary and tertiary hydrogen atoms from the PE and PP molecules at 298 K. It can only add to the vinyl, vinylene and vinylidene units that are formed in the synthesis of, e.g., PE.These reactions resulted in dinitro compounds and nitro nitrites. (1) C C(NO2) C C +NO2 R (2) C(NO2) C(NO2) R+NO2 (3) C(ONO) C(NO2) At T5373 K, nitro, nitrite, nitrate, carbonyl and hydroxy pendent groups are formed in these polymers. The following reaction mechanism was proposed: HNO2+ CH2 CH2 +NO2 CH CH2 R RNO2 R+NO2 RONO RONO RO +NO. Subsequent reactions involving the RO. radicals result in macromolecular nitrates, alcohols and carbonyl compounds. This scheme allows rationalisation of the accumulation of the nitro groups, which proceeds at a constant rate, and autoaccel- erated formation of nitrates, alcohols and carbonyl compounds; however, it provides no explanation for the S-shaped dependence of the accumulation of nitrites.The activation energy (Ea) for the NO2 addition to the double bonds of PE obtained by different methods is 8 to 16 kJ mol71. The activation energy for hydrogen abstraction lies between 56 and 68 kJ mol71 for PE and is 60 kJ mol71 for PP. At room temperature and NO2 concentrations lying between 5.461074 and 5.461073 mol litre71, the characteristics of PE, PP, PAN and PMMA change only slightly even if they are simultaneously exposed to a combination of NO2, O2 and UV radiation.43 The reactions ofNO2 with PVC and PVF resulted in aReactions of nitrogen oxides with polymers slight decrease in the content of chlorine and fluorine atoms, respectively.32, 43 At temperatures between 298 and 328 K, nitrogen dioxide (7.861073 ± 3.461072 mol litre71) can abstract tertiary hydro- gen atoms from PS molecules to form pendent nitro and nitrite groups.32 These reactions proceed at low rates and are accom- panied by chain scissions of the macromolecules.32, 45, 46 Accord- ing to Jellinek, the dependence of the decrease in the average degree of polymerisation of PS exposed to NO2 on the exposure time has three (initial, middle and final) linear regions.A decrease in the apparent degradation rate was observed during the second stage of the reaction (the middle region of the dependence). Presumably, this was related to the association of the macro- molecules in solution, which is due to the effect of polar groups and can affect the results of viscosimetric measurements. Subse- quent increase in the apparent degradation rate was attributed to the consumption of these nitrogen-containing groups and to a decrease in the degree of association of the macromolecules.Polystyrene films were also simultaneously exposed to NO2 (1.161074 mol litre71) and light (l>280 nm).46 No polymer degradation was observed in the initial stage (at exposure times of no longer than 10 h). Then chain scissions occurred at a constant rate. After 40 h, degradation ceased. The mechanism of degrada- tion under these conditions was not discussed. An attempt to carry out a quantitative study of the ageing of PS and poly(tert-butyl methacrylate) (PTBMA) under the con- ditions simulating the ageing of these polymers in an atmosphere polluted with NO2 has been undertaken by Huber.47 Glass supports coated with the samples were exposed to a stream of air containing NO2 (2.561076 to 3.761075 mol litre71) at 300 K and simultaneously irradiated with light (l>290 nm).The num- ber of chain scissions per 10 000 monomer units (a) was deter- mined from the data of viscosimetric measurements. The dependence of a on the exposure time can be described by an empirical equation a=PQÖexp Qt ¡ 1Ü, where P and Q are constants and t is time. This equation describes an autocatalytic process. At Q?0, degradation occurs at a constant rate. As the thickness of the sample increases, both constants, P andQ, decrease, the former decreasing faster than the latter.Autocatalytic process is more pronounced for thin films. Degradation of thin PS films under the same conditions occurs slower than that of the PTBMA films and its autocatalytic nature is more pronounced. Possible reasons for the dependence of the kinetic parameters of degradation on thickness of the samples have not been considered. The autocatalytic type of degradation of PTBMA was asso- ciated 47 with the photo-induced formation of isobutylene, which reacts withNO2, thus initiating free-radical degradation processes of macromolecules. The IR spectrum of PS exposed to NO2 and light under conditions mentioned above exhibits two bands at 1686 and 3400 cm71 corresponding to the carbonyl and hydroxy groups, respectively.The formation of nitrogen-containing products has not been observed in both PTBMA and PS. The following reactions have been assumed 47 to proceed in PS: C(Ph)H CH2 CH2 HNO2+ +NO2 C(Ph) (R1). R1O2. (R1).+O2 ROOH +(R1). R1O2.+RH R1NO2 (R1).+NO2 R1ONO hn R1O.+NO R1ONO 987 R1O.+.OH+NO R1OOH+NO (4) hn R1OOH R1O.+.OH (5) R1O. (R2).+degradation products. It is believed that the decomposition of hydroperoxides exposed to NO and light [reactions (4) and (5)] leads to autocata- lytic degradation of PS. The results of investigations on the mechanisms and kinetics of the reactions of macromolecular hydroperoxides with nitric oxide are considered in more detail in the forthcoming sections.2. Reactions of nitrogen dioxide with carbon-chain polymers containing carbon ± carbon double bonds Among the reactions of NO2 with carbon-chain polymers con- taining carbon ± carbon double bonds, its reactions with rubbers have been studied in most detail. Rubbers are much more susceptible to NO2 than the polymers containing no double bonds in both the main chain and side groups of the macro- molecules. First, this is due to the ability of NO2 to add reversibly to carbon ± carbon double bonds to give nitroalkyl radicals [reaction (1)], thus initiating free-radical transformations of elastomers. Second, it was found 48 that the rate of the formation of products (Prod) of the liquid-phase reactions of NO2 with alkenes is described by the equation dâProdä=k 0[alkene][NO2]+k 00[alkene][NO2]2, dt where k0 is the effective rate constant for the formation of products in the reactions of allyl radicals with NO2 and k00 is the effective rate constant for the formation of products in reactions (1) ± (3).The second term of this equation dominates at high NO2 concentrations; this corresponds to the formation of products in the recombination reactions of nitroalkyl radicals with nitrogen dioxide [reactions (1) ± (3)]. The first term of this equation dominates at low NO2 concentrations; this is associated with the ability of nitrogen dioxide to abstract hydrogen atoms in b-posi- tion to the double bond to give allyl radicals, which then combine with an NO2 molecule. Depending on the structure of the alkene, the reaction resulting in the formation of the allyl radical can be either weakly exothermic or weakly endothermic. For instance, the strengths of the weakest C7H bonds in the structural frag- ment 7C(1)H27[C(2)H3]C=CH7C(3)H27 of synthetic iso- prene rubber are 334, 351 and 316 kJ mol71 for the C(1)7H, C(2)7H and C(3)7H bonds, respectively.49 In preliminary studies it was shown 43 that exposure of polyisoprene and polybutadiene to nitrogen dioxide leads to both degradation and cross-linking of the macromolecules, whereas butyl rubber (a copolymer of isobutylene with 1.75% of isoprene) only undergoes degradation.Jellinek et al.32, 50, 51 car- ried out a detailed study of the ageing of butyl rubber exposed to (i)NO2 (5.261077 to 5.261075 mol litre71) alone, (ii) anNO2 ± O2 mixture and (iii) an NO2±O2 mixture plus UV light (l>280 nm).It was found that in the first case the polymer degradation is described by the formula a=keft , where kef depends on the NO2 pressure. This equation is valid except for the initial portion of the curve where the reaction rate decreases with time. A kinetic scheme of the main-chain scission was proposed. H3C NO2 CH3 C CH +NO2 C CH R chain fragments+NO2 R+NO2 to Exposure to a mixture of O2 (6.861074 7.861073 mol litre71) and NO2 (5.261077 to 5.261075 mol litre71) leads to an increase in the degradation rate of the988 rubber with the increase in the concentrations of both gases. According to Jellinek, there are three factors governing the main- chain scissioning in the polymers exposed to O2±NO2 mixtures.These are (i) degradation induced by NO2 alone, (ii) oxidative degradation of macromolecules and (iii) scissions of macromole- cules owing to the synergistic effect of NO2 and O2, which in turn is due to the NO2-induced decomposition of hydroperoxides formed in the reaction. According to Jellinek, the degradation kinetics of butyl rubber exposed to NO2±O2 mixture and UV irradiation is described by the following equation: a=k0eft2+k 00 ef[NO2][17exp(7k 000t)], where k0 and k00 are the effective rate constants for different groups of reactions and k000 is the rate constant for degradation of the elastomer exposed to NO2 alone.At low NO2 pressures, the degradation rate increases with time. At high NO2 pressures, the reaction slows down. 3. ESR studies of the reactions of nitrogen dioxide with double bonds in solid polymers As was mentioned above, nitration of aromatic compounds usually follows an ionic mechanism, whereas nitration of aliphatic compounds (in particular, the reaction of NO2 or its dimer with carbon ± carbon double bonds in non-polar media) proceeds by free-radical mechanism.37, 48, 52, 53 Liquid-phase reactions of low- molecular-mass alkenes with NO2±N2O4 mixture produce para- magnetic centres.54 It was found that the ESR spectra arising are temperature-dependent and that their parameters differed from those of the spectra of nitroalkyl radicals.These spectra were associated with the p-complexes formed in the reactions of nitro- gen dioxide with the double bonds of alkenes. Nitroalkyl radicals produced in the styrene ±NO2 (or a-methylstyrene ±NO2) systems were identified by Kommandeur et al.55 in the ESR spectra of the spin adducts of these radicals with nitrosobenzene. Paramagnetic centres observed 54 were assigned to the nitroxyl radicals gener- ated in the reactions NO2 C C +NO2 C C R. (6) RNO R.+NO O. RNO+R. (7) R N R It is believed that reaction (6) is due to the trace amounts of NO in nitrogen dioxide. Analysis of this scheme shows that the formation of nitroxyl radicals (7) must be autoaccelerated. This was confirmed in further studies.56 a.Reactions of UV-irradiated poly(methyl methacrylate) with nitrogen dioxide Nitrogen dioxide virtually does not react with PMMA. On the other hand, it is known that irradiation of this polymer with UV light causes elimination of the ester groups leading mainly to methyl formate, HCOOCH3, and macromolecular structures containing double bonds in the main chain.57 The reaction of nitrogen dioxide with the double bonds formed must initiate free- radical ageing reactions in the polymer. The formation of paramagnetic species in PMMA containing 0.3 ± 0.4 mol kg71 of the *(CH3)C(COOCH3) ±CH=C(CH3) ± CH2*units (produced upon photolysis) under the action of NO2 has been studied by ESR spectroscopy.58 The experimental ESR spectrum of the solid polymer is a superposition of two aniso- tropic triplets typical of dialkylnitroxyl and acylalkylnitroxyl radicals in the solid phase with low frequencies of rotational motions.59, 60 Dialkylnitroxyl macroradicals R7N(O.)7R give an anisotropic triplet with ANk =3.20.1 mT and gk= 2.00260.0005, while the triplet corresponding to acylalkylni- G B Pariiskii, I S Gaponova, E Ya Davydov troxyl radicals RC(=O)7N(O.)7R is characterised by the parameters ANk =2.10.1 mT and gk=2.00270.0005. In a benzene solution of PMMA, these radicals exhibit triplet ESR spectra with the parameters aN=1.50.1 mT, g= 2.00640.0005 and aN=0.80.1 mT and g=2.00650.0005, respectively.The absence of an additional hyperfine splitting in the spectra of these radicals indicates that the carbon atoms neighbouring to the NO.group are tertiary. The accumulation of dialkylnitroxyl macroradicals occurs without an induction period and its initial rate is proportional to the concentration of nitrogen dioxide in the gas phase. The kinetics observed shows that the scheme of the formation of dialkylnitroxyl radicals [reactions (1), (6) and (7)] in the liquid phase (see above) is inapplicable to the solid polymer due to the low probability of the encounter of a macroradical with the macromolecular nitroso compound. It was assumed that these radicals are produced as a result of a series of `pseudocage' reactions such as those shown below: CH3 CH3 CH3 CH3 C CH C C CH C +NO2 N O COOCH3 COOCH3 OCH3 CH3 C CH C +COOCH3 CH3 CH3 N O C CH C O CH3 CH3 N O COOCH3 C CH C O N O COOCH3 O O CH3 H3CC CH C N O COOCH3 O O CH3 H3C CH3 H3C C C CH C C N+ HCO N O COOCH3 COOCH3 O CH3 H3C C C N HCO COOCH3 O O O H3C H3CC N COCH3 C N+ COOCH3 HCO HCO According to this scheme, the NO2 group and the free valence are localised on the neighbouring carbon atoms of the primary nitroalkyl radical and can react with each other to produce unstable, four-membered alkoxyalkylnitroxyl radicals, which decompose to give the alkoxyl radical and nitroso group at the adjacent carbon atom.Subsequent decomposition of the alkoxyl macroradical results in an alkyl radical and nitroso compound located in the immediate vicinity of each other.The reaction between them, resulting in the formation of a dialkylnitroxyl macroradical, can be initiated by small-amplitude translational and rotational motions. Acylalkylnitroxyl radicals appear in the reactions of the low-molecular-mass methoxycarbonyl radicals with tertiary nitroso compounds. This reaction scheme ignores concurrent transformations of reactive free radicals, which pro- ceed without formation of nitroxyl radicals. These processes seem to be responsible for the low yield of R7N(O.)7R radicalsReactions of nitrogen oxides with polymers (about one hundredth of the total number of consumed NO2 molecules). The ability of nitro compounds to undergo reduction to nitroso compounds through the stages of formation and decom- position of alkoxyalkylnitroxyl radicals in the liquid phase has been pointed out.61 ± 63 It was also concluded that the reaction between the double bonds of PMMA and NO2 is accompanied by chain scissions in the macromolecules and elimination of the ester side groups.58 In addition, the structure of the nitroxyl macroradicals produced suggests that NO2 is added not only to the less sterically shielded carbon atom of the double bond, but also to the neighbouring atom.This was also observed in the liquid-phase reaction of methyl methacrylate with nitrogen dioxide.64 b. Nitroxyl radicals produced in reactions of nitrogen dioxide with rubbers As was mentioned in the preceding Section, macromolecular nitroxyl radicals produced in the reactions ofNO2 with the double bonds of polymers are spin labels which carry information on macromolecular dynamics.Knowledge of the formation rules and structure of nitroxyl radicals allows elucidation of the mechanisms of macromolecular transformations in the polymers. This approach was used 65 to obtain spin-labelled macro- molecules of 1,4-cis-polyisoprene ( synthetic isoprene rubber, or SKI2) and a copolymer of ethylene, propylene and dicyclopenta- diene (SKEPT-402). These spin labels were used to study the macromolecular dynamics in solid SKI2 and the structure of the reaction front during nitration of the polymer.66 Nitroxyl macroradicals formed in these polymers at 293 K E tc=t0 exp exhibit nearly identical ESR spectra. The ESR spectrum of nitroxyl radicals in SKI2 recorded at 298 K is shown in Fig.1 a. It is an anisotropic triplet with ANk =3.10 mT and gk=2.00280.0005. This indicates that the correlation time, tc, of rotational motions at this temperature is longer than 1079 s. As the temperature increases, the spectra become triplets with aN=1.530.03 mT and g=2.00570.0005 (Fig. 1b), which indicates a substantial shortening of the correlation time. The change in tc in the temperature interval from 293 to 373 K(Fig. 2) is described by the equation , RT where logt0=714.2 and E is the activation energy for rotational diffusion (34.7 kJ mol71). These values are close to the corresponding values determined for SKI2 using a bulky spin probe.67 Studies on the kinetics of the formation of nitroxyl radicals in solid elastomers showed that free-radical transformations of double bonds in the presence of nitrogen dioxide differ funda- mentally from analogous liquid-phase reactions.56 b a 2 mT 1 mT H H Figure 1.ESR spectra of nitroxyl macroradicals produced in the reac- tion of nitrogen dioxide with SKI2-3M at 293 (a) and 373 K (b). logtc (s) 78 79 103T71 /K71 3.0 2.6 Figure 2. Temperature dependence of the correlation time tc of rota- tional diffusion of SKI2 macromolecules measured using spin labels obtained in the reaction of polyisoprene with nitrogen dioxide. The changes in the concentration of nitroxyl radicals as functions of the nitration time of SKI2at different concentrations of nitrogen dioxide in the gas phase are shown in Fig.3 a.As in the case of PMMA, these radicals are formed in an NO2±N2O4 atmosphere without an induction period and the initial rate of formation is proportional to the concentration of nitrogen diox- ide. Initially, the concentration of radicals increases and reaches a maximum value at [NO2]=(7 ± 8)61074 mol litre71 and then decreases. It was assumed that this shape of the kinetic curves is due to the loss of nitroxyl radicals in the trapping reactions and in the redox reactions involving the nitrogen dioxide dimer. Along induction period is observed in the presence ofO2 in the gas mixture (Fig. 3 b, curve 2). The apparent accumulation rate of radicals to the end of this period is 50 to 60 times lower than that in the absence of oxygen at the same NO2 concentration; however, the amount of radicals accumulated is three times as large as that a 2 43 1 21 100 t /min b 20 000 t /min 105 100 t /min Figure 3.Kinetic curves of accumulation of nitroxyl macroradicals in SKI2 exposed to NO2 (a) and an NO2±O2 mixture (b); (a): [NO2] /mol litre71: 2.261073 (1), 761074 (2) and 1075 (3); (b): kinetic curve (1) and the initial portion of curve 1 (2); [NO2]=761074 mol litre71 and [O2 ]=1.261072 mol litre71. 10716 n /spin per sample 10716 n /spin per sample 10716 n /spin per sample 989 0.8 3 0.6 0.4 0.2 200 40 000 1 1.0 0.6 2 0.2 200 10716 n /spin per sample990 accumulated in the former case.It was concluded that in the absence of O2 nitroxyl macroradicals are formed in two ways. The first mechanism is the same as the liquid-phase one, while the second is a `pseudocage' mechanism, which predominates if the reaction is carried out in the absence of oxygen. In the presence of O2, `pseudocage' reactions are inhibited due to the conversion of initially formed nitroalkyl macroradicals into nitroperoxide radicals. From these data it also follows that the recombination of nitroxyl macroradicals with nitroalkyl and allylic macro- radicals in the absence of O2 must lead to gradual cross-linking of the SKI2 macromolecules. Similar kinetics of accumulation of nitroxyl macroradicals was also observed for SKEPT-402, which differs from SKI2 by a substantially lower proportion of double bonds.The mechanism of formation of nitroxyl radicals in these polymers by the `pseudocage reactions' is in essence the same as that in the case of PMMA containing double bonds. Alkoxyl macroradicals produced in the `pseudocage reactions' cause degradation of polymers, while recombination of nitroxyl radicals with nitroalkyl radicals leads to cross-linking of the elastomer. Reactions ofNO2 with solid polymers can be considered as the interaction between the molecules diffusing from the gas phase and reacting in the bulk of the polymer. In this case, a non- uniform distribution of the reaction front in the bulk of the polymer must be present. The method of ESR tomography (or the ESR imaging technique),68 ± 70 which has been rapidly devel- oped recently, allows determination of the spatial distribution of paramagnetic species.In the case of reaction of NO2 with the double bonds of elastomers, the spatial distribution of nitroxyl macroradicals characterises the structure of the reaction front during nitration of the polymers. The results of investigations of the reactions of SKI2 with NO2 and NO2±O2 mixtures by ESR tomography have been reported.66 The spatial distributions of nitroxyl macroradicals in the initial and deep stages of the nitration of SKI2 by a NO2±N2O4 mixture are shown in Fig. 4 a and Fig. 4 b, respec- a b 0.02 0.06 0.1 0.07 0.04 0.02 0.02 0.04 0.02 0.02 0.02 0.02 0.02 0.04 0.02 0.02 0.06 0.06 0.04 0.06 0.06 0.06 0.04 0.02 0.02 c d 0.02 0.07 0.05 0.07 0.03 0.03 0.03 0.05 0.02 0.07 0.05 0.07 0.03 0.05 0.02 0.08 0.03 0.03 0.05 0.07 0.07 0.11 0.02 1 mm 0.03 Figure 4.Bulk distribution of nitroxyl macroradicals produced upon nitration of SKI2 by NO2 with a concentration of 3.861074 mol litre71 (a, b) and a mixture of NO2 (5.561074 mol litre71) with O2 (1.561073 mol litre71) (c, d). Nitration time /h: 2.5 (a), 720 (b ), 168 (c) and 505 (d). It should be remembered that NO2 always is in equilibrium with N2O4. G B Pariiskii, I S Gaponova, E Ya Davydov tively. As can be seen, after the initial 2.5 h the reaction front penetrates the surface of the polymer to a depth of *1 mm.The interior of the sample is inaccessible to nitrogen dioxide. As the reaction time is extended to 740 h, the depth of the reaction zone increases only by 20% to 30%. The extent of the reaction is maximal in the near-surface layer and decreases as the reaction front moves deeper into the sample. The results obtained show that the initial structure of the reaction front is due to the penetration of NO2 into the bulk of the sample along pores, cracks, sites with low local density and that further propagation of the reaction front is very slow. Studies of the rates of diffusion of the low-molecular-mass radical (2,2,6,6-tetramethyl-4-oxopiper- idinoxyl) from the gas phase into the bulk of an SKI2 sample nitrated for different periods of time showed that nitration was accompanied by the efficient formation of a three-dimensional network structure, which precludes the diffusion of low-molec- ular-mass compounds into the sample.When an SKI2 sample is exposed to an NO2±N2O4±O2 mixture, the limitations on diffusion due to the changes in the structure of the polymer must diminish since the concentration of cross-links decreases in the presence of oxygen. Therefore, one could expect a greater depth of the reaction zone and a more uniform distribution of nitroxyl macroradicals in the bulk of the polymer. In fact, the distribution of nitroxyl macroradicals in the bulk of the sample and its change during nitration of the samples with an NO2±O2 mixture (Fig. 4 c,d) are similar to those for the sample exposed to an NO2±N2O4 mixture, but the reaction zone penetrates the sample to a smaller depth.This can likely be rationalised by the fact that the structure of the reaction front is due to the membrane-like type of the process.71 Under these conditions, the propagation velocity of the reaction front is determined by the consumption rate of nitroalkyl macroradicals: the higher the consumption rate, the smaller the depth to which the reaction front penetrates. Introduction of O2 into the system causes an increase in the consumption rate of nitroalkyl radicals, thus decreasing the depth of the reaction front. Thus, the results obtained by ESR tomography showed that if rubbers are exposed to high NO2 concentrations, the degradation and cross-linking of macromolecules occur in the narrow (of the order of 1 mm thick) surface layer and that no modification occurs in the bulk of the polymer sample for long.4. Reactions of nitrogen dioxide with polyamides, polyurethanes and polyamidoimides a. Effect of nitrogen dioxide on aliphatic polyamides Polymers containing amide and urethane groups form a particular class of materials sensitive to NO2. Jellinek et al.72, 73 showed that exposure of nylon-66 films of different morphology to NO2 (2.661075 to 10.461075 mol litre71) causes main-chain scis- sion in the polymers. The degradation of nylon is a diffusion- controlled reaction. Its rate and depth depend essentially on the degree of crystallinity of specimens and on the size of crystallites.The reaction can be inhibited by small amounts of benzaldehyde or benzoic acid. The degradation is accelerated in the presence of air andUVlight in addition toNO2. It is believed that benzoic acid blocks the amide groups and that only a few of them, not involved in hydrogen bonding, enter into the reaction. The following mechanism for the polymer degradation under the action of NO2 was proposed. O +NO2 +HNO2 C N O NO2 +NO2 C N OC O H C N OC N OC CH2 N CH2 N CH2+CH2Reactions of nitrogen oxides with polymers The effect of nitrogen dioxide on nylon fibres has been studied.74 It was found that long-term storage of samples in an atmosphere with low NO2 concentration at room temperature did not lead to their darkening or to a decrease in the tensile strength.These characteristics deteriorated at elevated temperatures. Deg- radation of nylon fibres irradiated with UV light in air was more pronounced in the presence of nitrogen oxides. The results of the UV, IR and ESR studies on the ageing of NO O +HNO3 . +N2O4 polycapramide (PCA) in a nitrogen dioxide atmosphere have been reported.75 The absorption bands in the region of 390 ± 435 nm typical of N-nitroso amides were observed in the UV spectra of PCA films stored in anNO2 atmosphere.76 IR studies showed that the intensity of the band at 3293 cm71 assigned to the stretching vibrations of hydrogen-bonded N7H groups decreases on storage of thin PCA films in an NO2 (461075 to 461074 mol litre71) atmosphere.The intensities of the bands at 1642 and 1563 cm71 also decrease. A band at 1730 cm71 charac- teristic of the stretching vibrations of the C=O groups not involved in hydrogen bonding appears instead of these bands, as well as two bands, at 1504 and 1387 cm71, corresponding to the antisymmetrical and symmetrical stretching vibrations of the 7N=O groups of nitroso amides. It was found that under these conditions the initial rate of the nitrosation of PCA is propor- tional to [NO2]n, where n&2. The formation of the nitroso amide groups can be depicted as follows: OC NH C N CH2 CH2 Nitrosation of PCA leads to disappearance of the system of hydrogen bonds, which determines its mechanical properties.After exposure to NO2, the PCA films degraded at minimum mechanical actions. White 77, 78 carried out detailed studies of the liquid-phase reactions of low-molecular-mass amides with nitrogen dioxide. He suggested that the action ofN2O4 used as the nitrosating agent is due to the equilibrium NO+NO¡ N2O4 3 and that the reaction proceeds by an ionic mechanism. Jellinek et al.72, 73 proposed a free-radical mechanism of the interaction of polyamides with NO2, which implies the formation of an N-nitro amide with a characteristic IR absorption band at 1580 cm71 as the main product. However, no formation of this product was detected, probably, due to the low proportion of the free-radical reactions in the overall process.The formation of amidyl radicals was confirmed by the accumulation of acylalkylnitroxyl macro- radicals (which occurs without induction period, however).75 These macroradicals are generated in the reactions (8) +HNO2 , +NO2 O O +NO. (9) +NO2 OC N O ONO C N O H C N OC N C N Thus, there are two concurrent mechanisms for the trans- formation of polyamides in the presence of nitrogen dioxide, namely, the free-radical mechanism with the rate proportional to the concentration of NO2 and the ionic mechanism, the rate of which is proportional to the square of theNO2 concentration. The ratio of these processes depends on the concentration of nitrogen dioxide in the gas phase. Therefore, linear extrapolation of the results obtained under the conditions of accelerated tests cannot be used for predicting the stability of polyamides.991 b. The effect of nitrogen dioxide on polyvinylpyrrolidone Polyvinylpyrrolidone (PVP) is a carbon-chain polymer containing a tertiary amide group in the pendent ring. CH2CH N O The interaction of PVP with NO2 has been described.75 Two absorption bands, at lmax=413 and 435 nm, were observed in the UVspectra of products of the reaction ofNO2 with PVP. As in the case of PCA, these peaks can be assigned to the absorption of the nitroso amide groups N(NO)7C(=O)7(CH2)3 formed in PVP due to the elimination of the side cyclic fragments. CH2CH (10) N PVP+NO2 O 7HNO2 (R1). N O (11) (R1). CH2CH + NO2 N N O O O2N X.(12) +2NO2 NO ONOX N N O O O2N O2N OX+ X is a free radical. ESR studies of the reaction ofNO2 (1074 to 1073 mol litre71) with PVP revealed two additional mechanisms for the trans- formation of the polymer. The ESR spectrum of paramagnetic species formed in this polymer (Fig. 5) is a superposition of the signals of iminoxyl and acylalkylnitroxyl radicals with the parameters ANk = 4.33 mT, ? =2.44 mT, gk=2.0029 and g\=2.0053 and ANk =1.94 mT AN and gk=2.003, respectively. The iminoxyl radicals are produced as follows. Nitric oxide recombines with (R1). to give nitroso compounds which undergo 2 mT H AN? ANk Figure 5. The ESR spectrum of paramagnetic species formed in the reaction of PVP with nitrogen dioxide. The low-field component of the hyperfine structure of acylalkylnitroxyl macroradicals is shown by an arrow.992 isomerisation into oximes to produce iminoxyl radicals in the reaction with NO2 CH2CH N (R1) +NO O ON (13) CH2CH CH2CH NO2 N N O ON O HON 7HNO2 Nitrous acid is the source of nitric oxide in the system H2O+NO2+NO.2HNO2 The formation of acylalkylnitroxyl radicals can be represented as follows. CH2C (14) N PVP+NO2 O 7HNO2 (R2) CH2C CH2C X. NO (R2) N N (15) NO O O CH2CN N(X)O. O It was found that the kinetics of the accumulation of iminoxyl and acylalkylnitroxyl radicals are different. The concentration of iminoxyl radicals increases initially (the rate of accumulation decreases with time), reaches a maximum value and then decreases.The accumulation of acylalkylnitroxyl radicals occurs initially in an autoaccelerated manner and then the rate of the process decreases and the concentration of the radicals reaches a constant value. This indicates that nitrogen-containing radicals are produced from different precursors following reactions with different rate-limiting stages. It was found that the structure and properties of the matrix affect the relative rates of the formation of nitrogen-containing radicals in PVP samples filled with aerosil.79 Thus, one can conclude that the presence of tertiary amide groups in the monomer units of PVP leads to weakening of the C7H bonds in the CH7 and 7CH27 groups neighbouring to the nitrogen atom. Nitrogen dioxide can abstract hydrogen atoms from these groups with ease to produce macroradicals which decompose with the cleavage of the C7C bonds in the b-position to the free valence.Alkyl macroradicals also recombine with nitrogen oxides to give nitrites, nitrates and nitroso compounds and then nitroxyl macroradicals. Trapping of the alkyl macro- radicals (R1). and (R2). by the nitroso compounds and nitroxyl radicals must be accompanied by cross-linking of the macro- molecules. c. Reactions of NO2 with polyurethanes Jellinek et al.80, 81 studied the effect of NO2 on films of linear polyurethane (PU) synthesised from tetramethylene glycol and hexamethylene diisocyanate. It was found that the degradation of PU is accompanied by cross-linking of macromolecules and that the degree of degradation and the yield (the weight percentage ) of the gel fraction are complex functions of the exposure time.For instance, the yield of the gel fraction initially increases up to 20% and then decreases down to nearly zero at 333 K and a NO2 concentration of 1073 mol litre71. The number of chain scissions in the sol fraction (i.e., the degree of degradation) increases initially, then decreases and eventually increases again; however, the final degradation rate is lower than the initial one. Exposure of G B Pariiskii, I S Gaponova, E Ya Davydov the PU films toNO2 is accompanied by evolution of CO2, which is a monotonic function of the exposure time. The IR spectra of the films allow assessment of the consumption of the NH bonds (n=3300 cm71).The reaction mechanism proposed by Jellinek et al.80, 81 involves the abstraction of hydrogen atoms from two types of structures, namely, a carbamate structure (A) and a tertiary amido structure (B) O O C N CH2 O C NH CH2 A where Z is a side alkyl group, as the initial step. The mechanism involves the following main stages: O NO2 A 7HNO2 O C N CH2 (R1) O NO2 B O C N CH2 7HNO2 Z (R2) R1NO2 (R1) +NO2 O (R1) CH2 N + C O CH2 (R3) (R3) CO2+ CH2 cross-linking products. (R1) +(R2) According to Jellinek, recombination of the (R1). and (R2). radicals leads to cross-linking of the polymer chains, while decomposition of the (R1). radicals results in the degradation of macromolecules and evolution of CO2.Energetically, the decom- position of the (R1). radicals seems to be hardly probable since this reaction results in the terminal macroradical (R3). and a nitrene, which is a very reactive species. On the other hand, more probable decomposition reactions of (R1). involving cleavage of the C7C or C7O bonds produce no alkoxycarbonyl macroradicals (R3)., which can undergo decarboxylation (see Ref. 82). Taking into account the mechanisms of the reaction of nitrogen dioxide with aliphatic polyamides and PVP, the ageing of PU in an NO2 atmosphere can be represented as follows: O NO A O C N CH2 N2O4 7HNO3 O A NO2 7HNO2 O C N CH2 (R1) O B O C N CH2 NO2 7HNO2 Z (R2) O NO2 B O C N CH 7HNO2 (R4) ZH (R4) (R3).+HZ N CH (R3) CO2+ CH2 nitration products (Ri ) +NO2 2 (Ri ) cross-linking products i=1±4.This scheme allows a rationalisation of the degradation accompanied by cross-linking of macromolecules, the consump- OB ZH CH2Reactions of nitrogen oxides with polymers tion of NH groups of the polymer as well as the evolution of carbon dioxide upon degradation. d. The reaction of NO2 with a polyamidoimide The effect of NO2 on the mechanical properties of a polyamidoi- mide (PAI) obtained by the polycondensation of chloroformyl- phthalic anhydride with 4,40-diaminodiphenyl ether has been studied.83 The experiments were carried out at 323 K and an NO2 concentration of 561073 mol litre71 and involved the measurements of the temperature dependence of the storage modulus E0 and the loss modulus E 00 for different exposure times of the PAI samples to NO2.At 473 K, the parameter E 0 revealed a non-monotonic decrease and the strongest temperature dependence. An increase in E0 at the glass transition temperature (*563 K ) was observed for the films exposed to NO2 over a period of eight days. When studying the temperature dependences of the loss moduli E00 of the films exposed to NO2 for different times, an absorption peak was found to appear at 473 K, which indicates degradation of polymers. All these facts were associated with degradation accompanied by cross-linking of macromole- cules. A decrease in the band intensities of the stretching vibrations of the N7H bonds and the appearance of the peaks associated with the addition of NO2 groups to macromolecules was observed in the IR spectra of the PAI films.No reaction mechanisms were discussed. To elucidate possible reaction mechanisms, we carried out ESR studies of the effect of NO2 (1073 mol litre71, at 293 K) on PAI of the same structure as that of the polymer studied by Kambe and Yokota.83 Under these conditions, iminoxyl radicals with the ESR spectral parameters ANk =4.1 mT, AN? =2.1 mT and gk=2.0028, g?=2.0058 are mainly produced in PAI. The proportion of the acylalkylnitroxyl macroradicals which have nearly the same ESR spectral parameters as those of polyamide and are likely produced by the same mechanism [reactions (8) and (9)] is less than 35%.Formation of iminoxyl radicals in this system is quite unexpected.Aconventional mechanism of their formation [reactions of the types (10) and (13) in PVP] involves (i) hydrogen abstraction from the CH2 group of a molecule, (ii) NOaddition to the radical produced and (iii) isomerisation of the newly formed nitroso compound into an oxime. Abstraction of labile hydrogen atoms from the oxime molecules by NO2 results in iminoxyl radicals. Since the PAI molecules contain no CH2 groups, this mechanism is impossible. It was assumed that the spin density in the amidyl radicals produced is partially delocalised over the phenyl ring, which facilitates cleavage of the polymer ether bond.The following mechanism of the reaction of NO2 with PAI was proposed. O H N2O4 C N O 7HNO3 CON CO ON O N C O CON CO O H NO2 C N O 7HNO2 CON CO O N C O CON CO (R1) O (R1) C N O + CON CO(R2) 993 O. N O H NO2 O N C O O N C O H (R3) NO O HN C (R3) O OH NO NO O O O O N C N C HO HOH H (R4) NO2 (cross-linking). (R2) + 7HNO2 R2 Decomposition of the radicals (R1). leads to the scission of the main chain of PAI, while their reactions withNO2, as in the case of PCA, result in acylalkylnitroxyl radicals. The reaction of the radicals (R2). with the phenyl rings of neighbouring macromole- cules causes cross-linking of the polymer and the reaction with NO2 results in the appearance of NO2 groups in the sample.83 Kinetic curves of the accumulation of the iminoxyl radicals correspond to the following overall reactions (R4)., RH+NO2 R4NO2 .(R4).+NO2 Their transformation in the coordinates of the following equation 0t s ln aR4a aOR4U as ¢§ aOR4U a a aOR4U as , W lation of the iminoxyl radicals; Wd where [(R4).] and [(R4).]s, are the current and steady-state concen- trations of the radicals, respectively, and W0 is the rate of formation of the iminoxyl radicals, allows the minimum degrada- tion rate of the polymer, Wd, to be determined from the accumu- was found to be 1078 mol kg71 s71. The mechanism proposed allows a ration- alisation of the degradation accompanied by cross-linking of the PAI macromolecules as well as the introduction of the NO2 groups.83 5.On the possibility of polymer degradation resulting from the reaction of nitrogen dioxide with peroxyl macroradicals As was mentioned above, at room temperature nitrogen dioxide can (i) abstract labile hydrogen atoms from macromolecules to produce alkyl or allyl macroradicals and (ii) add to the double bonds, thus producing the nitroalkyl macroradicals. If the reac- tion is carried out in air, carbon-centred macroradicals are oxidized to give peroxyl radicals. Presumably, the reactions of the latter with NO2 must result in relatively unstable peroxy nitrates, ROONO2, similarly to the low-molecular-mass radicals rO2.84 ¡¾ 86 O k16 (16) CH3COONO2 O k17 Decomposition of peroxyacetyl nitrate, CH3C(=O)OONO2, was studied in detail and two mechanisms were considered.87, 88 O CH3COO +NO2 O CH3CO +NO3 (17) CH3COONO2994 Using the spin trap technique, it was shown that peroxyacetyl nitrate decomposes mainly following reaction (16), which is reverse to the formation reaction of CH3C(=O)OONO2.89 An analogous conclusion was also drawn based on the results of quantum-chemical calculations.90 Stabilities and transformation mechanisms of other low-molecular-mass peroxy nitrates have been much less studied.The thermal stability of products of the reactions of NO2 with internal *CF2CF(OO.)CF2* (RinO2) and terminal *CF27 CF27OO. (RtO2) peroxyl macroradicals of poly(tetrafluoroethy- lene) (PTFE) produced under the action of penetrating radiation or UV light in an O2 atmosphere was studied by ESR spectro- scopy.91 These macroradicals are stable at T4333 K but can enter into reactions with nitrogen dioxide.If the reaction of the type (16) is the only way of the transformation of peroxy nitrates into PTFE, complete regener- ation of the initial peroxyl macroradicals must be observed in vacuo. The alkoxyl macroradicals produced by the reaction (17) can undergo effective decomposition accompanied by scission of the main chain of the polymer.92 For instance, O k18 CF2 CF CF2 The reaction of terminal alkyl macroradicals with NO3 or NO2 results in the formation of stable products and in a decrease in the degree of regeneration of RO It was found 91 that RinOONO2 do decompose with partial regeneration of RinO2 at room and elevated temperatures.At 298 K, the peroxy nitrate molecules which undergo dissociation are kinetically inequivalent. At 313 K, they undergo a unimolec- ular decomposition with a rate constant of the reaction of the type (16), 4.061074 4 k16 4 4.861074 s71 (Fig. 6 a), which is a ¡ln âRO2ä1 ¡ âRO2 ä âRO2ä1 012 2 3 3 2 1 b âRO2ä1 âRO2ä0 0.8 1 0.6 0.4 0.20 4 2 Figure 6. Kinetic curve of regeneration of RinO2 radicals after exposure of PTFE to NO2 (2.861073 mol litre71) for 1.5 h (a) and the limiting degree of regeneration of (RinO2)? radicals as a function of the time of preliminary exposure of sample t* to NO2 (761073 mol litre71) (b).Kinetic curve (1) and semilogarithmic transformation of curve 1 (2). O CF2 CF+ CF2 2. âRO2ä âRO2ä0 0.8 1 0.6 0.4 0.2 t /h 4¡ln âRO2ä1 âRO2ä0 2 4321 t* /h 6 (18) CF2 G B Pariiskii, I S Gaponova, E Ya Davydov independent of either [RinO2]0 or the exposure time (t*) of the samples to nitrogen dioxide. The effective constant k16 obtained is about three orders of magnitude smaller than the rate constant for decomposition of CClF2O2NO2.93 This seems to be due to the low probability of the escape of NO2 molecules from the `cage' in the solid phase. However, the t* value determines the limiting concentration of regenerated peroxyl macroradicals, [RinO2]? (Fig.6 b) k16 [RmO2 .]?=k16 á k17 [RmO2 .]0 exp(7keft*), where kef is the effective rate constant for various conversions of peroxy nitrate exposed to NO2 including those involving alkoxyl macroradicals. It was found that k17<kef<7.661075 s71, i.e., the rate constant of a reaction of the type (17), which can result in degradation of the PTFE macromolecules, does not exceed *20% of k16. Unexpected results were obtained in studies of the thermal stability of the reactions products of RtO2 with nitrogen dioxide. Heating of samples in vacuo at temperatures up to 413 K did not lead to production of free radicals. This can be due to either the fact that no peroxy nitrates are formed in this case or to the fact that the reaction proceeds through a six-membered cyclic tran- sition state and results in the fast formation of stable products.Thus, if a polymer containing rather stable peroxy radicals (e.g., PTFE) is exposed to high concentrations of nitrogen dioxide for long, the second decomposition route of peroxy nitrate [reaction of the type (17 )] can lead to a high degree of degradation of the polymer due to the irreversibility of reaction (18) despite the fact that the corresponding rate constant is five or more times smaller than that typical of the first decomposition route [reaction of the type (16)]. III. The effect of nitric oxide on polymers Nitric oxide is less reactive than nitrogen dioxide. Currently, it is known with certainty that at room temperature nitric oxide cannot abstract tertiary or allylic hydrogen atoms from organic molecules, since the H7NO bond with a strength of 208.62.1 kJ mol71 is weaker than the C7H bonds in organic compounds.Nitric oxide cannot also add to isolated double bonds.94 ± 96 Opinions differ widely as to the ability of NO to react with low-molecular-mass dienes and polyenes. It is known that NO can add to substituted o-quinodimethane, phorone and b-carotene to produce free radicals.97 ± 99 It is also believed that the reactions ofNOwith dienes are actually initiated by admixtures of NO2.96 In any case, it is of interest to study the regularities of the reaction of nitric oxide with the ageing products of polymers, e.g., hydroperoxides.100, 101 In addition, NO readily reacts with free radicals to give nitroso compounds, which are efficient spin traps for the free radicals. Analysis of the kinetic curves of the formation and the structures of nitrogen-containing radicals which are stabilised in this case makes it possible to draw conclusions on the mechanisms of free-radical processes in the polymers.1. Photolysis and radiolysis of poly(methyl methacrylate) and cellulose triacetate in the presence of nitric oxide Photolysis of PMMA exposed to NO leads to the formation of stable nitrogen-containing radicals. The structures of the radicals formed depend on the temperature and the wavelength of the photolysing light.102 For instance, acylalkylnitroxyl radicals, R1N(O.)C(=O)OR2, are produced upon photolysis of PMMA by unfiltered light of a mercury lamp at 298 K.In solid PMMA, they exhibit an anisotropic triplet ESR spectrum with ANk =2.10.1 mT and gk=2.0027 (Fig. 7 a). In a solution of the polymer, the parameters of the ESR spectrum of these radicals are aN=0.80.1 mT and g also dialkylnitroxyl radicals, RN(O.)R, are produced in the k=2.0065. Not only acylalkyl, but samples photolysed at 383 K. The latter exhibit the ESR spectraReactions of nitrogen oxides with polymers b a 2 mT H 2A1k2A1k2A2k Figure 7. ESR spectra of solid PMMAphotolysed in an NO atmosphere at 298 (a) and 383 K (b). The parameters 2A1k and 2A2k correspond to the components of hyperfine structure (HFS ) of the acylalkyl and dialkylni- troxyl radicals, respectively.with the parameters ANk =3.20.1 mT and gk=2.0026 for the solid phase (Fig. 7 b) and aN=1.50.1 mT and g=2.0064 for a solution. Under the action of UV light (260<l<400 nm) alone, iminoxyl macroradicals are also formed, which exhibit triplet ESR spectra with aN=2.8 to 3.1 mT and g=2.0050 (in solution). The formation of acylalkylnitroxyl radicals indicates that methoxycarbonyl radicals, CH3OC.(=O), are split off during photolysis. Dialkylnitroxyl radicals are produced as a result of both decomposition of tertiary *CH2(CH3)C.CH2* macrorad- icals at 383 K and depolymerisation of the chain carrier radicals. The formation of iminoxyl radicals implies generation of (CH3)C(COOCH3)C.H(CH3)C(COOCH3)* radicals during photolysis.In an atmosphere of nitric oxide, the latter are converted into nitroso compounds, which isomerise into oximes [the reactions are similar to the process (13)]. Radiolysis of PMMA at room temperature results in the formation of acylalkylnitroxyl and iminoxyl macroradicals,102 the proportion of the latter being much higher than in the case of photolysis. The results obtained show that radiolysis causes an increase in the contribution of the reaction of hydrogen abstrac- tion from the alkyl groups of the polymer to the overall reaction. Photolysis of cellulose triacetate (CTA) in an atmosphere of NO at 298 K leads to the production of dialkylnitroxyl and acylalkylnitroxyl radicals, which exhibit the ESR spectra with AN the AN k =3.10.1 mT, parameters gk=2.0026 and k =2.00.1 mT and gk=2.0032, respectively.Removal of NO causes a substantial increase in the intensities of the spectral components assigned to acylalkylnitroxyl radicals, which seems to be due to decomposition of the diamagnetic complexes of these radicals with NO. Dialkylnitroxyl radicals are known to form no such complexes.103 The appearance of acylalkylnitroxyl radicals indicates that the acetate groups decompose with the elimination of CH3C.(=O) radicals. Methyl radicals possibly participate in the formation of dialkylnitroxyl radicals. The very weak signal observed during radiolysis of CTA exposed to NO precludes elucidation of the nature of the radicals formed. After long-term evacuation of the samples, an ESR spectrum appears.The main signal in this spectrum is assigned to iminoxyl radicals (ANk =4.50.1 mT, AN? =2.40.1 mT, gk=2.0025 and g?=2.0054). In addition, a weak signal of acylalkylnitroxyl radicals is observed, which has the same param- eters as in the case of photolysis. This composition of radicals indicates that hydrogen abstraction from the CH3 groups of macromolecules plays a significant role during radiolysis. Carrying out free-radical processes in the polymers exposed to nitric oxide is in essence a specific modification of the spin trap technique. This method is, as a rule, used in the studies of the composition of free radicals in the liquid phase where the ratio of the spin adducts is determined by the composition of the radicals 995 that are trapped, the rate constants for their reactions with nitroso compounds and by the stabilities of the nitroxyl radicals formed.In the liquid phase, the encounter of reagents is due to diffusion, which is not the limiting stage of a process. In solid polymers, nitric oxide, low-molecular-mass free radicals and nitroso com- pounds formed from them can diffuse with relative ease. The encounter of macromolecular species can occur only if efficient mechanisms of free valence migration exist. Under the conditions discussed here, such encounters can be the limiting steps of the reactions. In particular, this is responsible for the formation of only acylalkylnitroxyl radicals during photolysis of PMMA exposed to NO at room temperature despite the decomposition of the ester group to give methoxycarbonyl and tertiary alkyl macroradicals, which are converted in the presence of nitric oxide into macromolecular nitroso compounds.Dialkylnitroxyl radi- cals are produced only at elevated temperatures, where the free valence migration occurs due to the depolymerisation reaction. 2. Free-radical processes in poly(perfluoroalkenes) exposed to nitric oxide g-Radiolysis of PTFE and a copolymer of tetrafluoroethylene with hexafluoropropylene exposed toNOrevealed the crucial role of the motion of radical centres. In this case, reactions g +NO CF2CFCF2 CF2CF(NO)CF2 CF2C(NO)CF2 CF2CCF2 NO ON CF2 CF3 g, NO g CF2 CF CF2 CF2 CF CF2 CF ON CF ON CF CF2 CF2 CF CF2 CF2 produce only iminoxyl macroradicals 104 NO FC NO CF and CF2 CF2 C CF2 CF2 No nitroxyl radicals are formed in perfluoroalkenes under these conditions.However, they can be generated in the post-radiation reactions proceeding in an NO atmosphere.105, 106 For instance, terminal and internal peroxyl macroradicals are stabilised in the g-radiolysis of PTFE in air at room temperature. If the samples evacuated after g-radiolysis are exposed to nitric oxide for a rather long time, the ESR spectrum of nitroxyl radicals *CF2N(O.)CF2* is observed after removal of NO. The ESR spectra of oriented films (Fig. 8) are quintets of triplets with the AN parameters AF k =0.46 mT, k =1.11 mT, gk=2.0060; AN? =1.12 mT, AF?=1.61 mT and g?=2.0071.The following mechanism for production of these radicals was proposed.105 In b a 1 mT H Figure 8. ESR spectra of oriented PTFE films g-irradiated in air and exposed to NO for 96 h at 298 K. The angle between the direction of film orientation and the permanent magnetic field is 0 8 (a) and 90 8 (b).996 the case of g-radiolysis of PTFE in oxygen-containing atmos- phere, a certain fraction of internal alkyl macroradicals has no time to be oxidised and undergoes decomposition accompanied by scission of the main chain of the polymer. C.F2+CF2=CF7CF2 CF27CF27C.F7CF2 The terminal alkyl groups are oxidised into terminal peroxide groups so that the neighbouring terminal peroxyl macroradicals and terminal double bonds are stabilised in the polymer.After exposure of such samples to NO, the peroxyl radicals can be converted as follows: CF2CF2O.+NO2 , CF27CF2OO.+NO CF2CF2ONO, CF2CF2O.+NO CF2CF2O.+NO, CF2CF2ONO C.F2+COF2 . CF2CF2O. Terminal alkyl macroradicals recombine with NO to produce terminal nitroso groups neighbouring to the terminal double bonds, which interact with terminal nitrosyl fragments to generate nitroxyl macroradicals CF2NO, C.F2+NO CF2N(O. )CF2C.FCF2 , CF2NO+CF2=CFCF2 CF2N(O. )CF2C.FCF2 +NO CF2N(O. )CF2CF(NO)CF2 . It was found 107 that low-molecular-mass perfluoroalkenes of different structures as well as hydrogen- or fluorine-containing nitroso compounds can enter in these reactions.Similar radicals were also produced in the copolymer of tetrafluoroethylene with hexafluoropropylene.106 This copolymer was used to generate yet another type of radicals, namely, *CF2N(O.)CF3, with an ESR spectrum consisting of eight HFS components split by 0.880.02 mT and g=2.0067. These parameters nearly coincide with the corresponding values for low-molecular-mass perfluoro- nitroxyl radicals of a similar structure in the liquid phase.107 The pattern of the ESR spectrum of the internal nitroxyl macroradical is due to partial averaging of the anisotropy of HFS and g-factor owing to fast rotation of the nitroxyl fragment about the long axis of the macromolecule. The mobility of the radical fragment in terminal nitroxyl radical is much higher, which in fact is respon- sible for a complete averaging of the anisotropy of the hyperfine interaction and g-tensors in its ESR spectrum.Thus, thermally stable spin-labelled macromolecules are gen- erated in the bulk of the samples as a result of post-radiation, free- radical transformations in fluorinated polymers exposed to NO. Their concentration is sufficiently high to allow the ESR studies on molecular dynamics in polymers. The advantage of this method for the introduction of a spin label is that the free-radical (nitroxyl) group can be introduced into various chain fragments, which allows one to obtain optimum information on the macro- molecular dynamics. 3. Interaction of polypropylene hydroperoxides with nitric oxide Polymeric hydroperoxides are reactive intermediates in the oxi- dative degradation of polymers. Investigations of the reactions of nitrogen oxides with these species furnishes insights into the mechanisms of the effect of air pollutant gases on the stabilities of polymeric materials.For instance, autoaccelerated degradation of PS macromolecules subject to the combined action of NO2, air and UV radiation was associated with the NO-induced decom- position of hydroperoxides.47 It is believed (see, e.g., Refs 100 and 108) that the process begins with the reaction (19) ROOH+NO RO.+HNO2 . G B Pariiskii, I S Gaponova, E Ya Davydov 0 âROOHä âROOHä 1.0 1 2 3 0.5 t /h 10 15 5 Figure 9. Kinetic curves of consumption of PP hydroperoxides at NO concentrations (in the gas phase) of 1.661073 (1), 3.261073 (2) and 4.161072 mol litre71 (3).It was also assumed (see Ref. 101) that the reaction proceeds to give the peroxyl radicals. (20) ROO.+HNO. ROOH+NO At high concentrations of hydroperoxides, NO-induced decomposition of their dimers was postulated for the liquid phase.108 (21) (ROOH)2+NO ROO.+HNO2+ROH. The decomposition kinetics of PP-derived hydroperoxides at 298 K and different NO pressures has been studied in detail.109 The kinetic curves for decomposition of hydroperoxides are shown in Fig. 9. As can be seen, the rate of hydroperoxide consumption is initially low and then increases substantially. Reactions (19) ± (21) cannot provide an explanation for this type of reaction kinetics. Using ESR spectroscopy, it was found that the decomposition of hydroperoxides is accompanied by accumu- lation of dialkylnitroxyl macroradicals and that the induction period is long.The duration of the induction periods for the decomposition of hydroperoxides and the formation of nitroxyl macroradicals increases as the concentration of nitric oxide in the gas phase decreases and is dependent on the presence of trace amounts of impurities (in particular, NO2) at a given NO concen- tration. It was concluded that the primary step in the decom- position of hydroperoxide is initiated by N2O3 rather than NO. [ROONO]+HNO2 . ROOH+N2O3 This reaction results in a thermally unstable peroxy nitrite, which can either undergo a fast `intracage' conversion into nitrate RONO2 [ROONO] [RO +NO2] RO +NO2 or decompose into alkoxyl macroradical and nitrogen dioxide, which then pass in the bulk of the sample. In turn, the alkoxyl radicals can either decompose or enter into substitution reaction, thus producing internal (Rin) and terminal (Rt) alkyl macro- radicals and low-molecular-mass alkyl radicals r (see Ref.100). These radicals react with NO to give nitroso compounds. RO. Rin(R.t , r. ), R RinNO(RtNO, rNO). in (Rt. , r. )+NO Acceleration of peroxide decomposition can be associated with the reactions which involve nitroso compounds 110 RO.+rN(O. )OH, ROOH+rNO rN(O. )OH r.+HNO2 . Alkyl radicals formed in the system can initiate decomposition of hydroperoxides.Reactions of nitrogen oxides with polymers rH(RinH,RtH)+RO2.r.(Rin ,Rt. )+ROOH RO +NO2 [ROONO] +NO RO2 RONO2 Yet another process which can be responsible for the increase in the hydroperoxide decomposition rate is disproportionation of NO into N2 and NO3 in the reactions with the nitroso com- pounds.94 RNO+2NO [R N N ONO2] RONO2+N2 [R +N2+ ONO2] R+N2+NO3 The ONO2 radical is a highly reactive species (the O2NO7H binding energy is 423 kJ mol71, see Ref. 111), which can abstract hydrogen atoms from the PP macromolecules, thus increasing the concentration of free radicals in the polymer R.+HNO3 , RH+.ONO2 and also reacts with NO 2NO2 , NO+NO3 which results in an increase in the concentration of nitrogen dioxide. Thus, taking PP as an example, it was found that decom- position of hydroperoxides in an atmosphere of NO is initiated by higher nitrogen oxides and proceeds by an autocatalytic mecha- nism.In an atmosphere polluted with nitrogen oxides, the primary step of the decomposition of hydroperoxides can be their nitro- sation by the dimeric form of nitrogen dioxide, the concentration of which is much higher than that of nitrous anhydride N2O3. On account of the great excess of oxygen compared to nitrogen oxides, the probability of the formation of nitroso compounds, which play the key role in the development of autoaccelerated reaction under the conditions used by Gaponova and Pariiskii,109 must be rather low in a polluted atmosphere due to the oxidation of alkyl radicals into peroxyl radicals.Investigation of the ROOH±NO system showed 109 that the results obtained under conditions of accelerated tests must be used with great caution when assessing the stability of polymers in atmosphere polluted with nitrogen oxides. IV. Conclusion The results summarised in this review show that nitrogen dioxide, which is an inorganic free radical of moderate reactivity, can perform some functions in the ageing of polymers in polluted atmosphere. On the one hand, at ambient temperatures it can abstract labile hydrogen atoms (tertiary hydrogen atoms in PS and polyvinylpyrrolidone and allylic hydrogen atoms in the polymers containing carbon ± carbon double bonds in the main chain) and add to the double bonds of macromolecules, thus generating free radicals and, hence, initiating free-radical ageing processes in the system.At elevated temperatures (T5373 K), NO2 also reacts with stronger C7H bonds, e.g., in the CH2 groups of polymers. On the other hand, it can react with alkyl, alkoxyl and peroxyl macroradicals generated in the reactions, thus inhibiting free-radical transformations of polymers and resulting in valence-saturated products of different thermal stabilities. The dimer of nitrogen dioxide,N2O4, is an efficient nitrosating agent and its reactions are responsible for the ageing of the polymers containing amide groups (polyamides, polyurethanes and polyamidoimides). Free-radical reactions initiated by NO2 are accompanied by production of stable nitrogen-containing radicals.Analysis of their structures and kinetics of formation provides additional information on the free-radical stages of complex processes of nitration of polymers under various con- ditions. 997 Nitric oxide, NO, is also a paramagnetic molecule, which is much less reactive than nitrogen dioxide. At ambient temper- atures, it cannot abstract even the most mobile (labile) hydrogen atoms from the molecules of organic compounds or add to isolated double bonds to produce free radicals. In addition, it is an efficient acceptor of free radicals which can inhibit free-radical chain reactions. Nitroso compounds formed in the reactions with alkyl radicals are efficient spin traps and can in turn inhibit free- radical processes in polymers.However, at high NO pressures their reactions with nitroso compounds lead to disproportiona- tion of NO into N2 and ONO2, which is a reactive radical capable of initiating free-radical processes in polymers. This reaction seems to be the stage responsible for autoaccelerated decomposi- tion of hydroperoxides in an atmosphere of nitric oxide. Free radicals in solid polymers can be generated by photolysis, g-radiolysis, thermal decomposition of initiators etc. Free-radical processes in solid polymers carried out in an atmosphere of NO are accompanied by production of nitrogen-containing macro- radicals of different structures.Analysis of their structures and kinetics of formation allows one to obtain additional information on the mechanisms of the above-mentioned processes. Performing free-radical reactions in chemically inert, insoluble polymers [e.g., poly(perfluoroalkenes)] in an atmosphere of nitric oxide makes it possible to prepare spin-labelled macromolecules with either internal or terminal NO. groups. These spin labels perturb the polymer chain dynamics to the smallest extent and are therefore the most appropriate for using in the ESR studies. References 5. G E Zaikov Usp. Khim. 60 2220 (1991) [Russ. Chem. Rev. 60 1145 1. N M Emanuel' Usp. Khim. 48 2113 (1979) [Russ. Chem. Rev. 48 1139 (1979)] 2. N M Emanuel', A L Buchachenko Khimicheskaya Fizika Stareniya i Stabilizatsii Polimerov (Chemical Physics of Ageing and Stabilisation of Polymers) (Moscow: Nauka, 1982) 3.N M Emanuel' Vysokomol. Soedin., Ser. A 27 1347 (1985) a 4. N M Emanuel', A L Buchachenko Khimicheskaya Fizika Moleku- lyarnogo Razrusheniya i Stabilizatsii Polimerov (Chemical Physics of Molecular Destruction and Stabilisation of Polymers) (Moscow: Nauka, 1988) (1991)] 6. N Grassie, G Scott Polymer Degradation and Stabilization (Cambridge, 1985) 7. E T Denisov Okislenie i Destruktsiya Karbotsepnykh Polimerov (Oxidation and Degradation of Carbon-Chain Polymers) (Leningrad: Khimiya, 1990) 8. A A Popov, N Ya Rapoport, G E Zaikov Okislenie Orientirovan- nykh i Napryazhennykh Polimerov (Oxidation of Oriented and Tense Polymers) (Moscow: Khimiya, 1987) 9.V A Roginskii Fenol'nye Antioksidanty. Reaktsionnaya Sposobnost' i Effektivnost' (Phenol Antioxidants. Reactivity and Efficiency) (Moscow: Nauka, 1988) 10. Yu A Shlyapnikov, S G Kiryushkin, A P Mar'in Antiokislitel'naya Stabilizatsiya Polimerov (Antioxidant Stabilisation of Polymers) (Moscow: Khimiya, 1986) 11. H H G Jellinek (Ed.) Aspects of Degradation and Stabilization of Polymers (New York: Elsevier, 1978) 12. H H G Jellinek (Ed.) Degradation and Stabilization of Polymers Vol. 1 (New York: Elsevier, 1983) 13. K Thinius Stabilisierung und Alterung von Plastwerkstoffen (Berlin: Akademie Verlag, 1969; 1970) Vols I, II 14. V Ya Shlyapintokh Fotokhimicheskie Prevrashcheniya i Stabilizat- siya Polimerov (Photochemical Conversion and Stabilisation of Polymers) (Moscow: Khimiya, 1979) 15.B Raà nby, J F Rabek Photodegradation, Photooxidation and Photo- stabilization of Polymers. Principles and Applications (New York: Wiley, 1975) 16. J Voigt Die Stabilisieeerung der Kunststofegegen Licht und WaÈrme (Berlin: Springer, 1966)998 17. B M Kovarskaya, A B Blyumenfel'd, I I Levantovskaya Termicheskaya Stabil'nost' Geterotsepnykh Polimerov (Thermal Stability of Heterochain Polymers) (Moscow: Khimiya, 1977) 18. A Charlesby Atomic Radiation and Polymers (Oxford: Pergamon Press, 1960) 19. MDole (Ed.) The Radiation Chemistry of Macromolecules Vol. 1 (New York: Academic Press, 1972) 20. J Sohma Prog. Polym. Sci. 14 451 (1989) 21.G E Zaikov, S A Maslov, V L Rubailo Kislotnye Dozhdi i Okruzhayushchaya Sreda (Acid Rains and Environment) (Moscow: Khimiya, 1991) 22. E D Hughes, C K Ingold, R I Reed J. Chem. Soc. 2400 (1950) 23. C A Bunton, E D Hughes, C K Ingold, D I H Jacobs,M H Jones, G J Minkoff, R I Reed J. Chem. Soc. 2628 (1950) 24. L Eberson,M P Hartshorn, F Radner Acta Chem. Scand. 48 937 (1994) 25. A I Titov Usp. Khim. 21 881 (1952) 26. A V Topchiev Nitrovanie Uglevodorodov i Drugikh Organicheskikh Soedinenii (Nitration of Hydrocarbons and Other Organic Compounds) (Moscow: Academy of Sciences of the USSR, 1956) 27. S S Novikov, G A Shveikhgeimer, V V Sevost'yanova, V A Shlyapochnikov Khimiya Alifaticheskikh Alitsiklicheskikh Nitrosoedinenii (The Chemistry of Aliphatic Alicyclic Nitro Compounds) (Moscow: Khimiya, 1974) 28.E T Denisov, N M Emanuel' Zh. Fiz. Khim. 31 1266 (1957) b 29. H C Beacheb, S P Nemphos J. Polym. Sci. 24 113 (1956) 30. T Ogihara Bull. Chem. Soc. Jpn. 31 58 (1963) 31. T Ogihara, S Tsuchiya, K Kuratani Bull. Chem. Soc. Jpn. 38 978 (1965) 32. H H G Jellinek, in Aspects of Degradation and Stabilization of Polymers Ch. IX (Ed. H H G Jellinek) (New York: Elsevier, 1978) 33. V Buffalini Envir. Sci. Technol. 5 685 (1971) 34. D O Gorelik, L A Konopel'ko Monitoring Zagryazneniya Atmosfery i Istochnikov Vybrosov. Aero-Analiticheskie Izmereniya (Monitoring of the Atmospheric Pollution and Waste Sources. Aero-Analytical Measurements) (Moscow: Izd. Standartov, 1992) 35.B G Artem'ev, S M Golubev Spravochnoe Posobie dlya Rabotnikov Meteorologicheskikh Sluzhb (Handbook for Workers of Meteorological Services) (Moscow: Izd. Standartov, 1990) 36. T C Hall, F E Blacet J. Chem. Phys. 70 1745 (1952) 37. P Golding, J L Powell, J H Ridd J. Chem. Soc., Perkin Trans. 2 813 (1996) 38. J H Ridd Acta Chem. Scand. 52 11 (1998) 39. E Bosch, J K Kochi J. Org. Chem. 59 3314 (1994) 40. M L McKee J. Am. Chem. Soc. 117 1629 (1995) 41. O N Karpukhin Usp. Khim. 47 1119 (1978) [Russ. Chem. Rev. 47 587 (1978)] 42. L V Gurevich, G V Karachintsev, V N Kondrat'ev, Yu A Lebedev, V A Medvedev, V K Potapov, Yu S Khodeev Energiya Razryva Khimicheskikh Svyazei. Potentsial Ionizatsii i Srodstvo k Elektronu (Energy of Rupture of Chemical Bonds.Ionisation Potential and Electron Affinity) (Moscow: Nauka, 1974) 43. H H G Jellinek, F Flajsman, F J Kryman J. Appl. Polym. Sci. 13 107 (1969) 44. H H G Jellinek, in The 2nd International Symposium on Degradation and Stabilization of Polymers (Abstracts of Reports), Dubrovnik, 1978 R1 45. H H G Jellinek, Y Toyoshima J. Polym. Sci., A-1 5 3214 (1967) 46. H H G Jellinek, F Flajsman J. Polym. Sci., A-1 7 1153 (1969) 47. A Huber, Diss. Doktor Naturwiss, FakultaÈ t Chemie der UniversitaÈ t Stuttgart, Stuttgart, 1988 48. D H Giamalva,G B Kenion,D F Church,W A Pryor J. Am. Chem. Soc. 109 7059 (1987) 49. Comprehensive Chemical Kinetics Vol. 14 (Amsterdam: Elsevier, 1975) 50. H H G Jellinek, F Flajsman J. Polym. Sci., A-1 8 711 (1970) 51.H H G Jellinek, P Hrdlovic J. Polym. Sci., A-1 9 1219 (1971) 52. H Shechter Rec. Chem. Prog. 25 55 (1964) 53. J L Powell, J H Ridd, J P B Sandall J. Chem. Soc., Chem. Commun. 402 (1990) 54. B H J Bielski, J M Gebicki J. Phys. Chem. 73 1402 (1969) 55. L Jonkman, H Muller, J Kommandeur J. Am. Chem. Soc. 93 5833 (1971) 56. A Rockenbauer,M GyoÈ r, F TuÈ doÈ s Tetrahedron Lett. 27 3425 (1986) G B Pariiskii, I S Gaponova, E Ya Davydov 57. T S Popravko, Yu A Mikheev, D Ya Toptygin Dokl. Akad. Nauk SSSR 232 856 (1977) c 58. T V Pokholok, G B Pariiskii, G O Bragina Vysokomol. Soedin. Ser. A 31 2049 (1989) a 59. A L Buchachenko, A M Vasserman Stabil'nye Radikaly (Stable Radicals) (Moscow: Khimiya, 1973) 60. A M Vasserman, A L Kovarskii Spinovye Metki i Zondy v Fiziko- khimii Polimerov (Spin Labels and Probes in Physical Chemistry of Polymers) (Moscow: Nauka, 1986) 61.R A Jackson, W A Waters J. Chem. Soc. 1653 (1960) 62. G R Chalfont, D H Hee, K S Y Liang,M J Perkins J. Chem. Soc., B 233 (1971) 63. W Ahrens, A Berndt Tetrahedron Lett. 4281 (1973) 64. W M Fox, J A McRae, M C R Symons J. Chem. Soc., A 1773 (1967) 65. T V Pokholok, G B Pariiskii Vysokomol. Soedin., Ser. A 39 1152 (1997) a 66. E N Degtyarev, T V Pokholok, G B Pariiskii, O E Yakimchenko Zh. Fiz. Khim. 68 461 (1994) b 67. V A Livshits, V A Kuznetsov, I I Barashkova, A M Vasserman Vysokomol. Soedin., Ser. A 24 1085 (1982) a 68. A I Smirnov, E N Degtyarev, O E Yakimchenko, Ya S Lebedev Pribory Tekhn. Eksperim. 195 (1991) 69. O E Yakimchenko, A I Smirnov, Ya S Lebedev Appl. Magn. Res. 1 1 (1991) 70. A I Smirnov, O E Yakimchenko, H A Golovina, Sh Kh Bekova, Ya S Lebedev J. Magn. Res. 91 386 (1991) 71. Ya S Lebedev Kinet. Katal. 6 522 (1965) d 72. H H G Jellinek, A Chandhuri J. Polym. Sci., A-1 10 1773 (1972) 73. H H G Jellinek, R Yokata, Y Itoh Polym. J. 4 601 (1973) 74. H Herlinger, B KuÈ ster, H Essig Text. Prax. Int. 44 655 (1989) 75. I S Gaponova, E Ya Davydov, G G Makarov, G B Pariiskii, V P Pustoshnyi Vysokomol. Soedin., Ser. A 40 551 (1998) a 76. C Djerassi, E Lund, E Bunnenberg, B Sjlberg J. Am. Chem. Soc. 83 2307 (1961) 77. E H White J. Am. Chem. Soc. 77 6008 (1955) 78. E H White J. Am. Chem. Soc. 77 6011 (1955) 79. G E Zaikov, G B Pariyskii, I S Gaponova, E Ya Davydov Oxidation Commun. 23 196 (2000) 80. H H G Jellinek, T J W Wang J. Polym. Sci., Polym. Chem. Ed. 11 3227 (1973) 81. H H G Jellinek, F H Martin, H Wegener J. Appl. Polym. Sci. 18 1773 (1974) 82. D Griller, B P Roberts J. Chem. Soc., Perkin Trans. 2 747 (1972) 83. H Kambe, R Yokota, in The 2nd International Symposium on Degradation and Stabilization of Polymers (Abstracts of Reports), Dubrownik, 1978 p. 39 84. C W Spicer, A Villa, H A Wiebe, J Heicklen J. Am. Chem. Soc. 95 13 (1973) 85. N Basco, S S Parmar Int. J. Chem. Kinet. 19 115 (1987) 86. F Zabel, A Reimer, K H Becker, E H Fink J. Phys. Chem. 93 5500 (1989) 87. E F J Duynstee, J L J P Hennekns, J G H M Housmans, W van Raayen, W Voskuil Recl. Trav. Chim. Pays-Bas 92 1272 (1973) 88. D G Hendry, R A Kenley J. Am. Chem. Soc. 99 3198 (1972) 89. J Maeda, H Masaoka, T Suetaka, M Munemori Chem. Express 3 591 (1988) 90. V L Antonovskii, K V Bozhenko Dokl. Akad. Nauk 343 337 (1995) c 91. I S Gaponova, G B Pariiskii J. Chem. Biochem. Kinet. 1 349 (1991) 92. V P Mel'nikov, L S Gulyaeva, A M Markevich Khim. Fiz. (5) 637 (1983) e 93. D KoÈ ppenkastrop, F Zabel Inter. J. Chem. Kinet. 23 1 (1991) 94. J F Brown Jr J. Am. Chem. Soc. 79 2480 (1957) 95. A Rockenbauer, L Korecz J. Chem. Soc., Chem. Commun. 145 (1994) 96. J S B Park, J C Walton J. Chem. Soc., Perkin Trans. 2 2579 (1997) 97. H-G Korth, R Sustmann, P Lommes, T Paul, A Ernst, H de Groot, L Hughes, K U Ingold J. Am. Chem. Soc. 116 2767 (1994) 98. I Gabr,M C R Symons J. Chem. Soc., Faraday Trans. 92 1767 (1996)999 Reactions of nitrogen oxides with polymers 99. I Gabr, R P Patel, M C R Symons, M T Wilson J. Chem. Soc., Chem. Commun. 915 (1995) 100. L A Tatarenko, V S Pudov Vysokomol. Soedin. 22 671 (1980) a 101. D J Carlsson, R Brousseau, Zhang Can, D M Wiles Polym. Degrad. Stab. 17 303 (1987) 102. I S Gaponova, G B Pariiskii, D Ya Toptygin Vysokomol. Soedin., Ser. A 30 262 (1988) a 103. B G Gowenlock, G Kresze, J Pfab Tetrahedron Lett. 593 (1972) 104. I S Gaponova, G B Pariiskii, D Ya Toptygin Vysokomol. Soedin., Ser. A 31 1238 (1989) a 105. I S Gaponova, G B Pariiskii Khim. Fiz. 16 (10) 49 (1997) e 106. I S Gaponova, G B Pariiskii Vysokomol. Soedin. 40 1906 (1998) a 107. L H Sutcliffe, A Zylnik J. Chem. Soc., Faraday Trans. 1 679 (1985) 108. J R Shelton, R F Kopczewski J. Org. Chem. 32 2908 (1967) 109. I S Gaponova, G B Pariiskii Vysokomol. Soedin. 37 1855 (1995) a 110. S Terabe, R Konaka J. Chem. Soc., Perkin Trans. 2 2163 (1972) 111. R Atkinson J. Phys. Chem., Ref. Data 20 459 (1991) a�Polym. Sci. (Engl. Transl.) b�Russ. J. Phys. Chem. (Engl. Transl.) c�Dokl. Chem. Technol., Dokl. Chem. (Engl. Transl.) d�Kinet. Catal. (Engl. Transl.) e�Chem. Phys. Rep. (Engl

ISSN:0036-021X    

出版商:RSC    

年代:2000

数据来源: RSC