Development of new methods in modern selective organic synthesis: preparation of functionalized molecules with atomic precision

A popular method for the introduction of a trifluoromethyl group is nucleophilic trifluoromethylation.⁴⁷ Unlike classical organomagnesium and -lithium reagents, the trifluoromethylated analogues (for example, ${\text{F}}_{3}{\text{CLi}}, {\text{F}}_{3}{\text{CMgBr}}$) are unstable even at reduced temperatures $(-78\ ^{\circ}{\text{C}})$. The most convenient synthetic equivalent of the ${\text{CF}}_{3}$ carbanion is trimethyl(trifluoromethyl)silane (1) (Ruppert – Prakash reagent). Treatment of silane 1 with basic activating reagents (fluoride, carboxylate and alkoxide anions) results in the generation of a pentacoordinate intermediate, which serves as the source of the trifluoromethyl anion (Scheme 1).

Zoom In Zoom Out Reset image size

Using silane 1, carbonyl compounds ^{47 – 49} and imines ⁵⁰ (and many other compounds with the ${\text{C}}{=\!=}{\text{N}}$ bond) ⁵¹ can be easily converted to the corresponding ${\text{CF}}_{3}$-substituted alcohols and amines — useful building blocks for medicinal chemistry (Scheme 2). (Some schemes in the review are presented in the general form reflecting the possibility of the processes; excessive detailing was considered inexpedient.) Examples of the asymmetric addition of the trifluoromethyl group to aldehydes using the fluoride anion for activation in combination with the chiral ammonium counter-ion based on cinchona alkaloids were described.⁵²

Zoom In Zoom Out Reset image size

Nucleophilic addition of the trifluoromethyl anion to Michael acceptors was also studied using several examples. High yields of conjugate addition products were obtained only for substrates with a highly electrophilic double bond (Scheme 3).^{53 – 55} However, the addition to classical $\unicode{945}, \unicode{946}$-unsaturated aldehydes, ketones, esters and nitro compounds has not yet been performed. Using the acylated Baylis – Hillman adducts 2 in combination with a chiral catalyst, the asymmetric nucleophilic substitution of the acyloxy group was carried out. This process includes a double allylic substitution: first, the chiral nitrogen-based nucleophile (L) attacks substrate 2 at the double bond thus generating chiral electrophilic intermediate 3, which reacts with silane 1 under activation by the released carboxylate anion.^{56, 57}

Zoom In Zoom Out Reset image size

A lot of attention was paid to the cross-coupling of aryl halides with nucleophilic trifluoromethylating reagents catalyzed by transition metal complexes.⁴¹ In the case of catalysis by palladium complexes, there is a problem of slow reductive elimination from ${\text{RPd(L)CF}}_{3}$. This problem was solved only in 2010 by using monophosphine biphenyl ligands in the reaction of aryl chlorides with a silicon reagent (Scheme 4).⁵⁸ Later, this cross-coupling reaction was extended to vinyl triflates and vinyl nonaflates.⁵⁹

Zoom In Zoom Out Reset image size

Trifluoromethylation reactions promoted by copper(I) salts were intensively studied.⁴¹ Trifluoromethylcopper can be prepared from silane 1 (see Ref. 60) or from fluoroform on treatment with potassium di(tert-butoxy)cuprate generated in situ from potassium tert-butoxide and copper(I) chloride (Scheme 5).⁶¹ Note that the copper complex with 1,10-phenanthroline (phen), ${\text{CF}}_{3}{\text{Cu(phen)}}$, has become commercially available. It was shown that trifluoromethylcopper can act as an efficient trifluoromethylating reagent for aromatic, benzyl, allyl and propargyl substrates.⁶² Conditions for conducting the trifluoromethylation in the presence of catalytic amounts of copper salts were also found.⁶³

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The conduction of electrophilic trifluoromethylation is a very difficult task. The presence of three fluorine atoms at the reaction centre hampers classical ${\text{S}}_{{\text{N}}}1$ and ${\text{S}}_{{\text{N}}}2$ reaction mechanisms. In 2006, reagents 4 and 5 based on trivalent iodine — Togni reagents — were proposed (Scheme 6).^{64, 65}

Zoom In Zoom Out Reset image size

These reagents proved to be efficient for electrophilic trifluoromethylation of a very broad range of substrates. Besides C-nucleophiles (1,3-dicarbonyl compounds, $\unicode{945}$-nitroesters, aromatic compounds), this reaction may involve thiols, phosphines, alcohols and azoles. Scheme 6 depicts an example of asymmetric trifluoromethylation of aldehydes with compound 5 in the presence of the MacMillan catalyst.⁶⁶

A new direction is related to the ability of some electrophilic trifluoromethylating reagents to transfer a cationic ${\text{CF}}_{3}$ group to a transition metal. This increases the oxidation state of the metal, which considerably promotes the reductive elimination. Presumably, the trifluoromethylation of boronic acids with 5 in the presence of copper salts occurs by this mechanism (Scheme 7).^{67, 68} In the palladium chemistry, a cycle based on the ${\text{Pd}}^{\text{II}}/{\text{Pd}}^{\text{IV}}$ pair was also proposed for CH-trifluoromethylation of aromatic compounds.⁶⁹ In the latter case, the Umemoto reagent was used as the source of the ${\text{CF}}_{3}$ cation (6) (see Scheme 7).

Zoom In Zoom Out Reset image size

Starting from 2010, studies of reactions involving cross-coupling of a nucleophilic component with a nucleophilic trifluoromethylation reagent in the presence of an oxidant (additional reagent or air oxygen) have been rapidly developing (Scheme 8). For example, a combination of simple terminal acetylenes with silane 1 in the presence of KF and oxygen affords, in one step, ${\text{CF}}_{3}$-substituted acetylenes 7, which are very difficult to prepare by any other method.⁷⁰ Similarly, heterocyclic compounds containing a relatively acidic hydrogen atom can be trifluoromethylated in the presence of a silver salt as an oxidant (see Scheme 8).⁷¹

Zoom In Zoom Out Reset image size

In recent years, the interest in free radical trifluoromethylation processes has substantially increased.⁴² Indeed, unlike the carbocation or carbanion, the trifluoromethyl radical is generated rather easily. This can be done under either oxidative or reductive conditions (Scheme 9). The trifluoromethyl radical can react with various $\unicode{960}$-nucleophiles (alkenes, aromatic compounds) to give products of ${\text{CF}}_{3}$-group substitution for hydrogen. Thus, ${\text{CF}}_{3}$-containing compounds can be prepared without preliminary functionalization of the substrate. However, the problem of low regioselectivity of the radical trifluoromethylation of substituted aromatic compounds remains unsolved in most cases.

Zoom In Zoom Out Reset image size

A special but still important case of synthesis of ${\text{CF}}_{3}$-containing compounds is the introduction of the 2,2,2-trifluoroethyl substituent, which can be performed by a nucleophilic substitution reaction (Scheme 10). Methods of Pd-catalyzed cross-coupling involving boronic acids or pinacolyl boronates and trifluoroethyl iodide were proposed; for this reaction to occur, the presence of a sterically hindered phosphine ligand, either bidentate Xanthphos or monodentate SPhos, is required.^{72, 73} However, reactions involving 2,2,2-trifluoroethyl organometallic reagents 8, which are prone to $\unicode{946}$-elimination, are still unknown. Just recently, stable organoboron compounds of this type [8, ${\text{M}} = {\text{B(OR)}}_{2}$] were obtained but cross-coupling reactions with these reagents have not been studied as yet.⁷⁴

Zoom In Zoom Out Reset image size

The most general approach to the formation of a ${\text{CF}}_{2}$ moiety is direct transformation of a carbonyl compound (aldehyde or ketone) according to the deoxofluorination reaction (Scheme 11).⁷⁵ However, a considerable drawback of this process is difficulty of handling of fluorine reagents (toxicity of sulfur tetrafluoride; detonation susceptibility of sulfotrifluorides 9). Recently, less hazardous reagents 10 ⁷⁶ and 11 ⁷⁷ were proposed. Nevertheless, deoxofluorination implies, by its nature, fairly drastic reaction conditions, which narrows down the range of functional groups that may be present in the substrate.

Zoom In Zoom Out Reset image size

Often, the target compounds are prepared by functionalization of accessible compounds that already have a ${\text{CF}}_{2}$ group.^{78, 79} As simple examples, of note are halodifluoroacetic acids $({\text{XCF}}_{2}{\text{CO}}_{2}{\text{H}}; {\text{X}} = {\text{Cl, Br}})$ and their derivatives and some difluoroalkanes $({\text{CF}}_{2}{\text{Br}}_{2}, {\text{ClCHF}}_{2})$, although the availability of the latter compounds can be considerably restricted by the Montreal Protocol due to ozone layer depletion.

Methods based on difluorocarbene reactions are rather efficient.⁴⁴ In particular, an approach to the synthesis of difluoromethylene compounds from three components — a nucleophile, difluorocarbene and an electrophile — has been proposed recently (Scheme 12).⁸⁰ The possibility of independent variation of the nucleophile and the electrophile makes this method quite versatile. This approach was demonstrated by a number of examples where organozinc compounds served as nucleophiles and bromine, iodine, proton and allyl halide were the electrophiles. Note that with an allylic electrophile, three-component coupling furnished two C–C bonds.⁸¹ In the reactions with organozinc reagents, (bromodifluoromethyl)trimethylsilane (12) in the presence of the acetate anion was used to generate difluorocarbene.

Zoom In Zoom Out Reset image size

gem-Difluoro-substituted cyclopropanes and cyclopropenes represent a separate class of compounds that are usually synthesized by 1,2-cycloaddition of difluorocarbene (Scheme 13).⁸² The suitable sources of carbene include salts of halodifluoroacetic acids, perfluoropropylene oxide, difluorohaloalkanes, difluoro(fluorosulfonyl)acetic acid derivatives ⁸³ or fluorosilicon and fluoromercury reagents.^{84, 85} Difluorocarbene is an electrophilic carbene; therefore, it readily reacts with electron-enriched double bonds, whereas the addition to electron-deficient alkenes is less efficient. 1,1-Difluoroalkenes 13 are prepared most often by the Wittig reaction (Scheme 14). However, the corresponding phosphonium ylide 14 is unstable and is generated in situ from dibromodifluoromethane ⁸⁶ or from zwitter-ionic reagent 15, which is easily formed from bromodifluoroacetic acid.⁸⁷

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

For the preparation of functionalized compounds with a ${\text{CF}}_{2}$ group, a series of heteroorganic (including organometallic) reagents containing both standard carbon functional groups (ester,⁸⁸ nitrile ⁸⁹) and substituents based on chalcogenides,⁹⁰ phosphorus ⁹¹ and even silicon ⁹² have been devised (Fig. 3). The presence of these substituents opens up prospects for the subsequent activation of the carbon–heteroatom bond (C–S, C–P, C–Si) by heterolytic and homolytic mechanisms, which provides a broad range of difluoro-substituted products.

Zoom In Zoom Out Reset image size

A classical method for the introduction of a fluorine atom into the aromatic ring is thermolysis of diazonium tetrafluoroborates (Schiemann reaction).⁹³ However, this method is potentially hazardous especially on an industrial scale due to low stability of diazonium salts. Meanwhile, the direct nucleophilic substitution of fluoride ion for the chloride ion, so-called Halex process is applicable only to a narrow range of substrates.⁹⁴

Replacement of the phenolic hydroxy group by a fluorine atom by deoxofluorination is difficult to implement, and standard fluorine – sulfur reagents used for aliphatic alcohols prove to be inefficient. The most reactive reagent for the deoxofluorination of phenols is difluoroimidazoline 16, which is suitable only for phenols containing electron-donating substituents in the ring (Scheme 15).⁹⁵

Zoom In Zoom Out Reset image size

In the last five years, methods for C–F bond formation based on the transition metal chemistry have been rapidly developed.⁹⁶ The fluorine atom substitution for the triflate leaving group was accomplished under palladium catalysis in the presence of sterically crowded monodentate phosphines (Scheme 16).⁹⁷ An alternative method includes replacement of the iodine atom by fluorine on treatment with an excess of copper(I) triflate and silver(I) fluoride at high temperature.⁹⁸

Zoom In Zoom Out Reset image size

A new, in principle, method for the preparation of aromatic fluorides comprises electrophilic fluorination of the nucleophilic carbon – element bond (Scheme 17). Originally, quite expensive xenon difluoride was used. A substantial progress along this line is related to the appearance of reagents with a nitrogen–fluorine bond,⁹⁹ although they are prepared using elemental fluorine. The Selectfluor and NFSI reagents are stable crystalline compounds convenient for handling and are sold by many companies at reasonable prices. The reaction of the organometallic substrate with an N-fluorinating reagent can occur as a direct nucleophilic attack on the fluorine atom (for ${\text{M}} = {\text{Li, MgX}}$). However, in some cases (for boron and tin compounds), the reaction is catalyzed by transition metal complexes (palladium, nickel, copper and silver) and, probably, includes the initial formation of the bond between the fluorine atom and the transition metal followed by reductive elimination.

Zoom In Zoom Out Reset image size

Electrophilic fluorinating reagents have started to be used in enantioselective processes catalyzed by transition metals or under organocatalysis. Scheme 18 shows the electrophilic fluorination of substrates 17 with chiral phosphoric acid 18 functioning as the asymmetric inductor. Presumably, salt 19 containing an achiral N–F group and two chiral counter-ions serves as the enantioselective reagent.¹⁰⁰ In this reaction, high enantiomeric excess values can be achieved, although the chiral inductor has no covalent bond with either substrate 17 or the fluorinating reagent.

Zoom In Zoom Out Reset image size

Academician O M Nefedov and co-workers ^{101 – 104} proposed an interesting approach to the synthesis of fluoroaromatic compounds, fluoroalkenes and fluorodienes, comprising the cycloaddition of fluorine-substituted carbenes to unsaturated substrates and the subsequent skeletal rearrangement of cyclopropanes involving three-membered ring opening, resulting in new structures with the fluorine – carbon bond being retained. The process is conducted in the flow mode. This method can be efficiently used to prepare fluoroaromatic compounds (aromatization is attained upon elimination of hydrogen halide) (Scheme 19).¹⁰¹ Indeed, various mono- and difluoroarene products (including 2,3-difluoronaphthalenes inaccessible by other methods) can be obtained from readily available starting compounds.

Zoom In Zoom Out Reset image size

Carbene syntheses of fluoroarenes can also be arranged for liquid-phase processes with generation of fluorochlorocarbene by treatment with aqueous alkali under phase transfer catalysis (at ${\sim} 0\ ^{\circ}{\text{C}}$). In this reaction, relatively labile spiro derivatives of cyclopentadiene and fulvenes can be used as substrates, apart from alkylcyclopentadienes and indenes.¹⁰¹

This approach is also rather efficient for the synthesis of fluoro-substituted dienes and alkenes. For example, cyclopropanation can be carried out under mild conditions of carbene decomposition of fluorodichloromethane giving rise to fluorochlorocyclopropanes in good yields (Scheme 20). The subsequent rearrangement of the fluorochlorocyclopropane moiety can be induced not only by thermolysis $(300 {-} 500\ ^{\circ}{\text{C}})$ but also on moderate heating $(80\ ^{\circ}{\text{C}})$ in the presence of a catalytic amount of copper(I) chloride or a CuCl and LiCl mixture in acetonitrile.^{103, 104}

Zoom In Zoom Out Reset image size

Skeletal rearrangements of methoxy-substituted gem-fluorochlorocyclopropanes can also be used to prepare fluorine-substituted $\unicode{945}, \unicode{946}$-unsaturated aldehydes, ketones and methoxy-substituted fluorodienes.^{105, 106}

The fluorine chemistry has started to play an important role in the positron emission tomography (PET) based on the use of isotopes that decay to give off positrons.^{36, 37} Among these isotopes, ${}^{18}{\text{F}}$ with the half-life of $109.7\ {\text{min}}$ is most convenient. After injection of a substance containing ${}^{18}{\text{F}}$ into the blood, it is possible to determine the exact distribution pattern of this substance in the body by means of the detection equipment. Fast decay of the ${}^{18}{\text{F}}$ isotope imposes substantial restrictions on the method used: the synthesis, isolation and purification of the desired compound should not require long time. In addition, the ${}^{18}{\text{F}}$ isotope is formed in the cyclotron as either elemental fluorine or a solution of hydrofluoric acid in water (i.e., as the fluoride anion). The latter alternative is, on the one hand, more practical and, on the other hand, more complicated as regards the introduction of fluorine into an organic molecule. This is why effective methods for the introduction of fluorine as the fluoride anion (especially into the aromatic ring) are highly demanded. Scheme 21 shows an example ¹⁰⁷ in which ${}^{18}{\text{F}}$-containing boronic acid 20 is attached to a protein containing a iodophenyl substituent via the Suzuki reaction. Reagent 20 is prepared in two steps using nucleophilic aromatic substitution to introduce the fluorine atom by the reaction of the iodonium salt with the fluoride anion in the presence of cryptand K-222 (the cryptand strongly binds the potassium ion thus transferring the fluoride anion into the organic solvent).

Zoom In Zoom Out Reset image size

Thus, unlike the ${\text{CF}}_{3}$ group, which can be introduced into the molecule by a variety of reported methods, there only a few methods for the precise and highly selective introduction of fluoro- and difluoro-substituents. Conduction of nucleophilic substitution by the ${\text{S}}_{{\text{N}}}2$ mechanism using fluorine-containing nucleophilic reagents remains an unsolved problem.

A separate topical direction is synthesis of compounds with a ${\text{CF}}_{2}$ group. For example, it is difficult to prepare compounds containing a halodifluoromethyl group $({\text{CF}}_{2}{\text{X, X}} = {\text{Cl, Br, I}})$ using existing reactions, and the corresponding carbanions are quite unstable.

The metathesis of linear 1,7-dienes containing an $\unicode{945}$-amino acid skeleton was first employed by Grubbs ¹²² in 1996 to prepare dehydropipecolic acid ($n = 1$, Scheme 22). It was found that the corresponding five-membered derivatives — dehydroprolines — cannot be synthesized with catalysis by the carbene complex G-I (see Fig. 4): due to the acidic nature of the $\unicode{945}$-proton in the initial 1,6-dienes ($n =0$, see Scheme 22), the reactions afford only linear $\unicode{945}, \unicode{946}$-unsaturated oligomers. More recently, this strategy was successfully used by other research teams to prepare six-, seven- and eight-membered amino acid derivatives ($n =1 {-} 3$, see Scheme 22).^{123, 124}

Zoom In Zoom Out Reset image size

Ring closing metathesis was first used in the synthesis of fluorine-containing cyclic $\unicode{945}$-amino acids in 1998.¹²⁵ The starting di- and trifluoromethyl dienes were prepared under mild conditions by amidoalkylation of the corresponding C-nucleophiles with highly electrophilic fluorinated methyl pyruvate imines. The subsequent cyclization was easily accomplished at room temperature in the presence of $5\ {\text{mol}}.\hbox{%}$ catalyst G-I to give metathesis products in high yields. The absence of the $\unicode{945}$-proton ensured the formation of previously inaccessible five-membered proline derivatives ($n =0, m=1$, Scheme 23).¹²⁶

Zoom In Zoom Out Reset image size

2-Trifluoromethyl-substituted 4,5-dehydropipecolic acids containing additionally an alkenyl substituent in the 4-position can be prepared by combined ring opening — ring closing metathesis (ROM – RCM).¹²⁷ The synthesis of the starting aza-1,7-dienes containing one of the double bonds in the ring comprises the ene reaction of methyl trifluoropyruvate imine with methylidenecycloalkanes and subsequent N-allyllation with sodium hydride-induced deprotonation. The aza-1,7-dienes obtained in this way readily undergo carbene complex G-I-catalyzed rearrangement of the carbon skeleton, which includes opening of cycloalkene and closure of a new heterocycle, resulting in the desired dehydropipecolinates with an alkenyl substituent the length of which can be controlled by varying the ring size in the starting methylidenecycloalkane (Scheme 24).

Zoom In Zoom Out Reset image size

The catalytic cycle, resulting in the rearrangement of the carbon skeleton occurs, apparently, in the following way: four-membered metal ring A formed initially is fragmented with evolution of a styrene molecule to give carbene complex B, which undergoes intramolecular cyclization to give unstable $[2+2]$-cycloadduct C. The latter rearranges into a new ruthenium complex D, which in turn undergoes cross-metathesis with a styrene molecule occurring in the reaction area to give stable reaction products and a catalyst molecule for a new catalytic cycle (Scheme 25).

Zoom In Zoom Out Reset image size

Ring closing enyne metathesis (RCEYM), like diene metathesis, has been rapidly developed in recent years and has acquired importance in the synthesis of carbo-, hetero- and macrocyclic compounds and other biologically important derivatives.^{128, 129} A characteristic feature of intramolecular ring closing enyne metathesis is that the formation of a new C–C bond occurs without the loss of the carbon skeleton and results in cyclic 1,3-dienes with one double bond in the ring, which are widely used to form polycyclic systems.

However, the RCEYM synthesis of cyclic 1,3-diene-containing $\unicode{945}$-amino acids often gives unsatisfactory results. On the one hand, this may be due to catalyst inhibition upon coordination of the nucleophilic nitrogen atom to the metal centre. On the other hand, metathesis of terminal enynes often gives low product yields due to poisoning of the active catalytic species in the secondary metathesis processes of diene intermediates.¹³⁰

Taking into account the above-indicated features, we studied metathesis of 1,6- and 1,7-enynes containing $\unicode{945}\hbox{-}{\text{XCF}}_{2}\hbox{-}\unicode{945}$-amino acid moieties. The starting enynes can be obtained either from the corresponding imines ¹³¹ similarly to dienes or from fluorinated diazocarbonyl compounds by [2,3]-sigmatropic rearrangement of the allyl group in nitrogen ${\text{CF}}_{3}$ ylides (Scheme 26).¹³²

Zoom In Zoom Out Reset image size

As a result, it was found that aza-1,6-enynes cyclize only in the presence of $5\ {\text{mol}}.\hbox{%} {-} 8\ {\text{mol}}.\hbox{%}$ of the allenylidene ruthenium complex $[{\text{Ru}}{=\!=}{\text{C}}{=\!=}{\text{C}}{=\!=}{\text{CPh}}_{2}({\text{Cl}})$ $({\text{PCy}}_{3})(p\hbox{-}{\text{cymene}})]^{+}{\text{TfO}}^{-}$, which was deliberately synthesized from the readily accessible precursor $[{\text{RuCl(PCy}}_{3})(p\hbox{-}{\text{cymene}})]^{+}{\text{TfO}}^{-}$ and the alcohol ${\text{HC}}{\equiv} {\text{CCPh}}_{2}{\text{OH}}$. The metathesis products obtained in moderate and good yields were then introduced into the Diels – Alder reaction followed by oxidation with 2,3-dichloro-4,5-dicyanobenzoquinone (DDQ), which resulted in the corresponding functionally substituted benzoprolines (Scheme 27).¹³¹

Zoom In Zoom Out Reset image size

As expected, in the case of terminal aza-1,7-enyne, the desired metathesis product is formed in a moderate yield when the reaction is catalyzed by commercially available Grubbs (G-II) or Hoveyda – Grubbs (H-II) catalysts. Even an increased catalyst loading and long-term heating in toluene produced mainly the product of homo-cross-metathesis of cyclic 1,3-diene (Scheme 28).¹³³

Zoom In Zoom Out Reset image size

Since cross-metathesis of 1,1-disubstituted alkenes is known to be a challenging task requiring either drastic conditions or more active catalytic systems,¹¹⁷ the decision was made to introduce an additional substituent to the aza-1,7-enyne triple bond; then the RCEYM step will give 1,1-disubstituted alkene, which will suppress the undesirable side homo-cross-metathesis reaction. A number of new aza-1,7-enynes containing an internal triple bond were prepared by the Pd-catalyzed Sonogashira reaction and subjected to ring closing metathesis. In all cases, aza-1,7-enynes smoothly cyclized on heating in toluene, which resulted in the selective formation of cyclic 1,3-dienes in high yields (Scheme 29).¹³³

Zoom In Zoom Out Reset image size

In order to demonstrate the synthetic potential of the obtained conjugated dienes in the design of functionally substituted polycyclic systems, tricyclic $\unicode{945}$-amino acid derivatives were synthesized by a procedure based on $[4+2]$-cycloaddition of cyclic 1,3-dienes to $N$-phenylmaleimide (Scheme 30).

Zoom In Zoom Out Reset image size

In addition, a new metathesis reaction involving aza-1,6- and 1,7-enynes was discovered. In the presence of $5\ {\text{mol}}.\hbox{%}$ ruthenium complex ${\text{Cp}}^{*}{\text{Ru}}({\text{cod}}){\text{Cl}}$ (${\text{Cp}}^{*}= \eta^5 \hbox{-}{\text{C}}_{5}{\text{Me}}_{5}$; cod is cycloocta-1,5-diene) and an equimolar amount of substituted diazoalkane, the starting enynes undergo combined cyclization – cyclopropanation reaction to give the corresponding bicyclic products (Scheme 31). The products formed in reactions with trimethylsilyldiazomethane have the alkenyl group exclusively in the $Z$-configuration, whereas reactions with ethyl diazoacetate give products with the $E$-configured alkenyl group.¹³⁴

Zoom In Zoom Out Reset image size

The presumed mechanism of this unusual transformation comprises the primary reaction of a diazo compound with the catalyst to give carbene complex E, which reacts with the substrate by the classical enyne metathesis mechanism up to the formation of intermediate ${\bf{H}} ({\bf{E}} \rightarrow {\bf{F}} \rightarrow {\bf{G}} \rightarrow {\bf{H}})$, and after that, the pathway becomes different. In particular, intermediate H does not give off the corresponding carbene complex (as in the enyne metathesis) but undergoes instead reductive elimination to give unusual bicyclic products. The observed stereochemistry of the alkenyl moiety may be related to the sterically more favourable arrangement of the trimethylsilyl (TMS) and pentamethylcyclopentadiene groups in intermediate F, namely, in anti-positions relative to each other. Meanwhile, it cannot be ruled out that the interaction between the neighbouring chlorine atom and the TMS group can also be responsible for the reaction stereochemistry (Scheme 32).

Zoom In Zoom Out Reset image size

The intra- and intermolecular cyclotrimerizations of acetylenes catalyzed by transition metal complexes have been widely spread during the last two decades as a versatile method for the synthesis of substituted aromatic and heteroaromatic compounds. Catalysts based on transition metal complexes, most often, cobalt and rhodium, were successfully used in the alkyne cyclotrimerization serving as the key step for the synthesis of a series of biologically important structures, including natural compounds.¹³⁵ However, only a few examples of application of this strategy to prepare $\unicode{945}$-amino acid derivatives were reported, and data on the synthesis of fluorine-containing amino acids and their phosphorus analogues are totally absent.

It is known that the tetrahydroisoquinoline moiety is encountered in many organic compounds that exhibit various types of biological activity including antihypertensive, antimalarial and antitumour activities. Among these, 1,2,3,4-tetrahydroisoquinoline-3-carboxylic acid (TIC), being a conformationally rigid analogue of phenylalanine, attracts particular attention as an important component of highly selective enzyme inhibitors and integrin and $\delta$-opioid receptor antagonists with a broad spectrum of pharmacological activity.^{136, 137} The most vivid example of using the TIC pharmacophore for the drug design is the discovery of Accupril®,¹³⁸ which is currently used to treat hypertension and congestive heart failure (Fig. 5). Therefore, the development of efficient catalytic processes for the synthesis of new TIC derivatives is of obvious fundamental and applied interest.

Zoom In Zoom Out Reset image size

An original approach was proposed to the synthesis of ${\text{CF}}_{3}\hbox{-}{\text{TIC}}$ derivatives ¹³⁹ and their phosphorus analogues ¹⁴⁰ comprising regioselective co-trimerization of aza-1,7-diynes with terminal acetylenes catalyzed by the Grubbs ruthenium complex (G-II). The starting trifluoromethyl-aza-1,7-diynes were prepared from the imines ${\text{CF}}_{3}({\text{X}}){\text{C}}{=\!=}{\text{N}}{-}{\text{PG}}$ $[{\text{X}}={\text{CO}}_{2}{\text{Me}},\ {\text{P(O)(OEt)}}_{2})]$ ¹⁴¹ using the following synthetic sequence: (i) the reaction of the imine with propargylmagnesium bromide; (ii) the introduction of the aryl substituent at the terminal triple bond by means of Pd-catalyzed cross-coupling with aryl iodides; (iii) N-propargylation of the Sonogashira reaction products during deprotonation of sodium hydride (Scheme 33).

Zoom In Zoom Out Reset image size

A study of the Ru-catalyzed cyclotrimerization of aza-1,7-diynes with various alkynes (acetylene, hex-1-yne, oct-1-yne and phenylacetylene) ascertained that reactions with terminal acetylenes occur on moderate heating in dichloroethane (DCE) in the presence of $5\ {\text{mol}}.\hbox{%}$ of the second-generation Grubbs catalyst (G-II) and lead to the corresponding bicyclic derivatives with high meta-selectivity and in good yields (Scheme 34). In all cases, the content of the meta-isomer was above 92%, and in the case of the 2-MeO-phenyl substituent, it was $> 97\hbox{%}$.

Zoom In Zoom Out Reset image size

The presumptive mechanism of cyclotrimerization includes the preferred addition of the ruthenium carbene complex to the terminal triple bond of aza-1,7-diyne to afford intermediate I, this step being facilitated by the coordination of ruthenium to the remaining triple bond. This is followed by a cascade of well-known intra- and intermolecular metathesis steps ${\bf{J}}\to{\bf{K}} \to{\bf{L}}$, which ends in the release of the ruthenium benzylidene active species and predominant formation of the meta-isomer. The selective $[2+2]$-addition of the terminal acetylene to the ${\text{Ru}}{=\!=}{\text{C}}$ bond of intermediate J, resulting in the formation of ruthenacyclobutene K, is responsible for the observed chemo- and regioselectivity of the process (Scheme 35).

Zoom In Zoom Out Reset image size

Note that new fluorine-containing $\unicode{945}$-amino acids and their derivatives prepared by ring closing metathesis hold good promise for the use in bioorganic and medicinal chemistry for modification of biologically active peptides and are of interest by themselves for biological activity assays.

Generally, despite the substantial achievements in the metathesis, some problems of intra- and intermolecular metathesis transformations remain unsolved. First of all, this concerns the regio- and stereoselectivity of reactions, especially in the metathesis involving polyfunctional unsaturated compounds containing internal double (RCM process) and terminal triple (RCEYM process) bonds. Furthermore, due to the thermodynamic nature of metathesis, the $Z$-selective alkene metathesis is still a highly problematic reaction. The use of most of commercially available catalysts usually furnishes a difficult-to-separate $Z$- and $E$-alkene mixture in which the more thermodynamically favourable $E$-isomer predominates.¹⁴² This fundamental problem of the alkene metathesis markedly restricts its industrial implementation. Solution of this problem can lead to creation of industrially important processes for utilization of by-products of petroleum cracking,^{143, 144} alternative energy sources based on bioresources ^{145, 146} and new stereoregular polymers with unique properties.^{147, 148} Therefore, the search for new, in principle, ruthenium-based catalytic systems with enhanced characteristics,^{149, 150} first of all the activity and selectivity, is a highly topical line of research of modern chemical science.

The first example of catalytic alkyne hydrophosphorylation has been known since 1996 and implies the use of palladium phosphine complexes.¹⁹⁴ A considerable part of synthetic procedures published later also used various Pd phosphine complexes as catalysts and provided branched alkenylphosphonates ($\unicode{945}$-isomers).^{195 – 197}

Successful use of rhodium complexes for catalytic hydrophosphorylation of alkynes giving selectively only the linear $\unicode{946}\hbox{-}E$-isomer was demonstrated for some examples. This difference of the selectivity of addition in the presence of complexes of different metals (Pd and Rh) was attributed to differences at the step of alkyne insertion, which involves the Pd–P bond in one case and the Rh–H bond in the other case. However, in both cases, the reaction proceeds as syn-addition. Thus, Pd- and Rh-catalyzed hydrophosphorylations of alkynes supplement each other, thus enabling the selective synthesis of isomeric $\unicode{945}$- and $\unicode{946}\hbox{-}E$-alkenylphosphonates, respectively.^{198, 199}

A recently developed procedure allows controlling the selectivity of Pd-catalyzed reaction, which can be used to synthesize both branched $(\unicode{945}\hbox{-})$ and linear $(\unicode{946}\hbox{-}E\hbox{-})$ products.²⁰⁰ The control of the selectivity was achieved by deliberate modification of the coordination sphere of palladium intermediates by choosing the appropriate ligand out of the series ${\text{P}}\{{\text{(MeO)}}_{n}{\text{C}}_{6}{\text{H}}_{5{-}n}\}_{3}$.²⁰⁰ This approach to the regioselectivity control has no analogues among reactions of this type.

In 2004, a study of alkyne hydrophosphorylation catalyzed by nickel complexes was published,²⁰¹ being concerned with the effect of diphenylphosphinic acid ${\text{[Ph}}_{2}{\text{P(O)OH}}]$ on the reaction regioselectivity with the use of Ni systems with different phosphorus ligands (Scheme 56). Unfortunately, in this case, it is impossible to identify the key factor responsible for the regioselectivity of addition, because the reactions were carried out in different solvents (THF, ethanol) with addition of the acid ${\text{[Ph}}_{2}\ {\text{P(O)OH}}]$ and without the acid using various catalyst precursors and ligands of various nature (${\text{PPh}}_{2}{\text{Me}}$ and ${\text{PPhMe}}_{2}$) (see Scheme 56).²⁰¹ The authors also did not present the interpretation of the regioselectivity change from the standpoint of the reaction mechanism.

Zoom In Zoom Out Reset image size

Note that a similar effect of the ${\text{Ph}}_{2}{\text{P(O)OH}}$ additive was also observed for the Ni-catalyzed addition to alkynes of other ${\text{P}}^{\text{V}}{-}{\text{H}}$ substrates — Ph(EtO)P(O)H and ${\text{Ph}}_{2}{\text{P(O)H}}$.²⁰¹ However, in these cases, too, the factor responsible for the control of regioselectivity remained unidentified for the above-indicated reasons. Moreover, commercially available diphenylphosphine oxide $[{\text{Ph}}_{2}{\text{P(O)H}}]$ used by the researchers ²⁰¹ contains, as a rule, a noticeable amount of the acid ${\text{Ph}}_{2}{\text{P(O)OH}}$ formed upon oxidation. Therefore, study of the effect of the acid on alkyne hydrophosphinylation with diphenylphosphine oxide requires thorough purification of this ${\text{P}}^{\text{V}}{-}{\text{H}}$ substrate. Unfortunately, the researchers did not describe the purification of ${\text{Ph}}_{2}{\text{P(O)H}}$ before the reaction.

Subsequently, the $[{\text{Ni}}]{-}{\text{Ph}}_{2}{\text{P(O)OH}}$ catalytic system was used in the reaction of propargyl alcohols with various ${\text{P}}^{\text{V}}{-}{\text{H}}$ substrates [H-phosphonates, H-phosphinates and ${\text{Ph}}_{2}{\text{P(O)H}}$], which made it possible to obtain phosphinoyl-1,3-dienes in high yields (Schemes 57 and 58).²⁰²

Zoom In Zoom Out Reset image size

A more detailed comparative analysis of published data (see Schemes 56 and 58) ^{201, 202} discloses some contradictions in the presented results and brings about some questions. Indeed, earlier the same authors demonstrated ²⁰¹ the possibility of hydrophosphorylation of oct-1-yne with dimethylphosphite on a nickel catalyst (see Scheme 56); however, only $1\ {\text{mol}}.\hbox{%}$ of the nickel-containing catalyst precursor was used, whereas in the reaction of dimethyl phosphite with 2-methylbut-3-yn-2- ol (see Scheme 58), the amount of the catalyst precursor was $5\ {\text{mol}}.\hbox{%}$.

Zoom In Zoom Out Reset image size

In addition, in the latter case, the reaction was carried out for $16\ {\text{h}}$ rather than for $2\ {\text{h}}$ as the reaction with oct-1-yne. It can hardly be conceived that the presence of the hydroxy group in the aliphatic moiety at the triple bond can have so adverse effect on the addition reactions and can require a $5\ {\text{times}}$ greater amount of the catalyst for a $8\ {\text{times}}$ longer reaction time. One more fact is noteworthy, in particular, in this synthetic procedure, the authors use $10\ {\text{mol}}.\hbox{%}$ ${\text{Ph}}_{2}{\text{P(O)OH}}$ (see Scheme 58) instead of $2\ {\text{mol}}.\hbox{%}$ used earlier (see Scheme 56) but the product yield was much lower in this case: 68% and 91%, respectively. Note that in the presence of $10\ {\text{mol}}.\hbox{%}$ ${\text{Ph}}_{2}{\text{P(O)OH}}$, the regioselectivity is lower: the content of the $\unicode{945}$-isomer is only 77%, while in the presence of $2\ {\text{mol}}.\hbox{%}$ ${\text{Ph}}_{2}{\text{P(O)OH}}$, it is 92% (see Schemes 56 and 58).^{201, 202} The use of different amounts of the catalyst precursor and contradictory data on the effect of acid addition on the regioselectivity with the alkyne structure being varied preclude drawing unambiguous conclusions about the general applicability of the $[{\text{Ni}}]{-}{\text{PPhMe}}_{2}{-}{\text{Ph}}_{2}{\text{P(O)OH}}$ system proposed by the authors to hydrophosphorylation of terminal alkynes.

A procedure of alkyne hydrophosphorylation using inexpensive and oxygen- and air moisture-stable ${\text{Ni(acac)}}_{2}$ as the catalyst precursor was developed in 2010.²⁰³ With this nickel complex, the catalytic reaction can be carried out without addition of the acid with no decrease in the yield or selectivity. The addition of H-phosphonates with various substituents to terminal and internal alkynes was accomplished with high regio- and stereoselectivity (Scheme 59).

Zoom In Zoom Out Reset image size

The mechanism of catalytic hydrophosphorylation of alkynes was examined using quantum chemical calculations.²⁰⁴ Comparative analysis was done for the two alternative pathways of alkyne insertion either at the metal–phosphorus bond or at the metal–hydrogen bond. It was demonstrated for the first time that alkyne can be inserted much more readily into the M–H bond than into the M–P bond. The same publication ²⁰⁴ demonstrated that the catalytic hydrophosphorylation of internal alkynes can, in principle, occur without a phosphine ligand.

Yet another vivid example of nickel-catalyzed hydrophosphorylation is the addition of H-phosphonates to ynamides.²⁰⁵ The best result was achieved with ${\text{NiBr}}_{2}\ (10\ {\text{mol}}.\hbox{%})$ (Scheme 60).

Zoom In Zoom Out Reset image size

The current trend of going from palladium to nickel catalysts is attributable, first of all, to the economic factor, i.e., high cost of palladium compounds. Also, some organic nickel derivatives [for example, ${\text{Ni(acac)}}_{2}$] are attractive from the practical standpoint, as they are air stable, can be stored for long without the loss of activity, whereas some palladium compounds have low stability even in the solid state [for example, ${\text{Pd}}_{2}({\text{dba}})_{3}$, dba is dibenzylideneacetone].²⁰⁶

Recently, high catalytic activity of the system ${\text{Ni(acac)}}_{2}{-}{\text{DIBAL}}\ ({\text{DIBAL}}={\text{Bu}}_2^{{\text{i}}}{\text{AlH}})$ in the hydrophosphorylation of terminal and internal alkynes was found and explored.²⁰⁷ The hydrophosphorylation proceeds without a solvent and does not require the use of phosphine ligands or acid additives. The ${\text{Ni(acac)}}_{2}{-}{\text{DIBAL}}$ system provides a unique possibility for controlling the selectivity by changing the amount of ${\text{Ni(acac)}}_{2}$ ($1\ {\text{mol}}.\hbox{%}$ or $9\ {\text{mol}}.\hbox{%}$): either mono- or 1,2-bis-phosphonate can be obtained in a quantitative yield (Scheme 61).

Zoom In Zoom Out Reset image size

To date, only a few publications have been devoted to the catalytic hydrophosphorylation of diynes.^{194, 198, 208} In the presence of palladium ^{194, 208} and rhodium ¹⁹⁸ catalysts, addition of simple aliphatic H-phosphonates [${\text{(MeO)}}_{2}{\text{P(O)H}}$,^{194, 208} $({\text{cyclo}}\hbox{-}{\text{OCMe}}_{2}{\text{CMe}}_{2}{\text{OP}})({\text{O}}){\text{H}}$ ^{198, 208} and $({\text{EtO}})_{2}{\text{P(O)H}}$ ²⁰⁸] was performed to give $\unicode{945}$- and $\unicode{946}$-adducts and cyclic vinylphosphonate derivatives.

The moderate number of publications is due to low reactivity of diynes in hydrophosphorylation. The ${\text{Ni(acac)}}_{2}$–dppe-based catalyst, which showed high efficiency in the hydrophosphorylation of hept-1-yne,²⁰³ proved to be inefficient for hepta-1,6-diyne.²⁰⁹ Evidently, the substantially lower reactivity of diynes compared to alkynes is due to the presence of the second triple bond in the molecule; however, the cause of this unexpected phenomenon is still unclear.

In 2013, hydrophosphorylation of a number of diynes was accomplished in the presence of a nickel catalyst.²⁰⁹ The ${\text{Ni(acac)}}_{2}{-}{\text{DIBAL}}$ catalytic system described above ²⁰⁷ provided the formation of alkyltetraphosphonates 21 upon the addition of H-phosphonates of different nature to diynes. The products were obtained in good (up to 91%) yields. This work was the first example of the synthesis of alkyltetraphosphonates by the catalytic reaction of H-phosphonates with diynes.²⁰⁹

Zoom In Zoom Out Reset image size

As in the case of hydrophosphorylation, studies of the catalytic hydrophosphinylation of acetylene hydrocarbons were started in the second half of the 1990s. The development of both reactions was pursued in parallel, both similar and distinctive features being revealed in the course of studies.

Commercially available H-phosphonates are represented by a series of inexpensive compounds, while the marketed secondary phosphine oxides are limited to only a few expensive ${\text{P}}^{\text{V}}{-}{\text{H}}$ substrates of this class. Among these, diphenylphosphine oxide has the lowest cost, but still it is tens and sometimes hundreds times more expensive than H-phosphonates. Furthermore, as has already been noted, commercial ${\text{Ph}}_{2}{\text{P(O)H}}$ contains, most often, noticeable quantities of the acid ${\text{Ph}}_{2}{\text{P(O}}){\text{OH}}$; therefore, thorough preliminary purification of this reagent is required, which increases the duration and the cost of the whole synthetic procedure. Nevertheless, there are several examples of reported catalytic addition reactions of commercial ${\text{Ph}}_{2}{\text{P(O)H}}$ to acetylene hydrocarbons.^{201, 202, 210 – 215} Complexes of various metals (Pd, Rh, Ni, Cu) were used as catalysts. Analysis of published data shows that many characteristic features of hydrophosphorylation reactions are also observed for hydrophosphinylation. The use of nickel catalysts for this class of reactions was discussed above.^{201, 202}

It was shown that the Rh complex-catalyzed addition of ${\text{Ph}}_{2}{\text{P(O)H}}$ to alkynes can proceed without a solvent under microwave irradiation (MW),²¹³ the reaction time in this case being decreased to several minutes. This reaction, like all the processes involving Rh complexes discussed above, gave, with high selectivity, linear $\unicode{946}\hbox{-}(E)$-alkenylphosphine oxides (for terminal alkynes) and syn-addition products (for internal alkynes). The possibility of microwave-assisted bis-hydrophosphinylation with 2 – 3 equivalents of ${\text{Ph}}_{2}{\text{P(O)H}}$ was demonstrated ²¹³ (Scheme 62).

Zoom In Zoom Out Reset image size

The same study reports the possibility of using polymer-supported catalysts, which provides regeneration and reuse of the metal complex.²¹³ This approach to the design of a catalytic system, i.e., the use of microwave radiation, catalyst recycling and the absence of solvent, is consistent with the green chemistry requirements and makes it possible to decrease the expenditures for the synthesis.

Yet another important transformation catalyzed by Rh complexes is the addition of ${\text{Ph}}_{2}{\text{P(O)H}}$ to the triple bond of ethynylsteroids — microwave-assisted hydrophosphinylation insensitive to air oxygen and moisture.²¹⁶ The reaction occurs in water as the solvent and affords a linear ($\unicode{946}\hbox{-}E$-isomer) addition product in good yield (Scheme 63). The use of water is highly untypical for the addition of molecules with a ${\text{P}}^{\text{V}}{-}{\text{H}}$ bond to alkynes; as a rule, these transformations are adversely sensitive to even traces of water.

Zoom In Zoom Out Reset image size

Among examples of bis-hydrophosphinylation, mention should be made of the reaction of terminal alkynes with diphenylphosphine oxide in the presence of ${\text{Pd}}({\text{PPh}}_{3})_{4}$ ²¹⁷ or binuclear or trinuclear Pd–M catalysts $({\text{M}}={\text{Ti, Zr, Hf}})$; the process occurs under mild conditions $(40\ ^{\circ}{\text{C}}, 1\ {\text{h}})$ and gives products in high yields $(> 95\hbox{%})$.²¹⁸

A recent paper ²¹⁵ describes the procedure of addition of chiral enantiomerically pure $1r\hbox{-}{\text{oxo}}\hbox{-}2{\text{c}},5t$-diphenylphospholane to terminal alkynes catalyzed by Pd and Rh complexes. The selective formation of $\unicode{945}$- and $\unicode{946}\hbox{-}E$-isomers in Pd- and Rh-catalyzed reactions, respectively, was demonstrated (Scheme 64). Unfortunately, although the product yield calculated from NMR data was often higher than 95%, most products were isolated in yields of about 25% – 40%, which was attributed by the authors ²¹⁵ to destruction of the alkenylphosphine oxides on silica gel during chromatography.

Zoom In Zoom Out Reset image size

The ratio between two alternative processes, namely, regioselective addition to give the $\unicode{945}$-isomer and bis-hydrophosphinylation, in the reaction of terminal alkynes with ${\text{Ph}}_{2}{\text{P(O)H}}$ catalyzed by Pd complexes was studied.²¹⁹ In the presence of the bidentate dppe ligand, branched $\unicode{945}$-adducts were formed in good yields (64% – 88%) and with high selectivity (95% – 99%). When a monodentate ligand was used [${\text{P}}(o\hbox{-}{\text{Tol}})_{2}{\text{Ph}}, o\hbox{-}{\text{Tol}}$ is $o$-tolyl], bis-hydrophosphinylation occurred predominantly to give a reasonable yield (48%) of the corresponding product.

In 2007, alkyne hydrophosphinylation was implemented in the presence of the CuI–1,2-diaminoethane system (Scheme 65).²²⁰ The use of ${\text{Ph}}_{2}{\text{P(O)H}}$ and ${\text{Bn}}_{2}{\text{P(O)H}}$ as the ${\text{P}}^{\text{V}}{-}{\text{H}}$ substrates resulted in the formation of $\unicode{946}$-alkenylphosphine oxides in good yields and with high stereoselectivity.²²⁰

Zoom In Zoom Out Reset image size

In 2011, the first studies were published devoted to the addition of secondary dialkylphosphine oxides to alkynes.^{221, 222} Palladium complex 22, in the presence of bidentate phosphine ligands, catalyzed the addition of the ${\text{P}}^{\text{V}}{-}{\text{H}}$ substrates (H-phosphonates, H-phosphinates and secondary phosphine oxides) to terminal alkynes.²²¹ This resulted in the synthesis of a series of $\unicode{945}$-adducts in quantitative yields with high regioselectivity (Scheme 66).

Zoom In Zoom Out Reset image size

The use of the catalytic system ${\text{CpPd}}(\eta^3\hbox{-}{\text{All}})$ – dppben – $[3,5\hbox{-}({\text{CF}}_{3})_{2}{\text{C}}_{6}{\text{H}}_{3}]_{2}{\text{P(O)OH}}\ [{\text{Cp}}={\text{Z}}^{5}$$\hbox{-}{\text{C}}_{5}{\text{H}}_{5},$ ${\text{dppben}}=1,2\hbox{-}({\text{PPh}}_{2})_{2}{\text{C}}_{6}{\text{H}}_{4}]$ resulted in the addition of dibutylphosphine oxide to some terminal alkynes and diphenylacetylene.²²²

The number of reactions of this class reported in the literature is much less compared to those for hydrophosphorylation and hydrophosphinylation. Effective catalytic systems based on palladium ([Pd] is phosphine ligand) for selective addition of hypophosphites to acetylenic hydrocarbons are known.^{223 – 227} The use of catalytic amounts of ${\text{NiCl}}_{2}$ provided the addition of alkylphosphinates to terminal and internal alkynes without a phosphine ligand.²²⁸ Unfortunately, in the case of ${\text{NiCl}}_{2}$, the reaction regioselectivity was lower than in a similar process involving Pd complexes, and hydrophosphonylation of terminal alkynes led to a mixture of $\unicode{945}$- and $\unicode{946}$-isomers.

In 2007, a two-step procedure for the synthesis of various phosphonic acids was developed.²²⁹ The first step was the addition of hypophosphorous acid $[{\text{HOP(O)H}}_{2}]$ to the alkyne triple bond to give alkenyl-H-phosphinate (Scheme 67). In the second step, the product was oxidized with air oxygen in dimethylformamide, which gave rise to the corresponding phosphonic acid.

Zoom In Zoom Out Reset image size

In the small number of examples of H-phosphinate addition to alkynes, phenyl-H-phosphinates were used in all cases as the ${\text{P}}^{\text{V}}{-}{\text{H}}$ substrates.^{201, 202, 220, 221, 230} It was found that an acceptable selectivity of $\unicode{945}$-isomer formation in the Pd-catalyzed reaction can be achieved without the addition of acid.²³¹ Good results were attained by using the ${\text{Pd(OAc)}}_{2}$–dppe system as the catalyst (Scheme 68).²³¹

Zoom In Zoom Out Reset image size

In 2013, Montchamp and co-workers ²³² published a study devoted to Pd-catalyzed reactions of phosphorus–carbon bond formation. By using the ${\text{Pd(OAc)}}_{2}$–dppf system as the catalyst [dppf is 1,1'-bis(diphenylphosphino)ferrocene], the addition of various H-phosphinates to internal oct-4-yne was accomplished.

Thus, the addition of ${\text{P}}^{\text{V}}{-}{\text{H}}$ substrates to the triple carbon – carbon bond is of obvious interest as a tool of hydroheterofunctionalization of organic compounds and opens up broad prospects for the development of effective procedures for the synthesis of valuable ${\text{C}}{-}{\text{P}}^{\text{V}}$ derivatives. The variation of substituents in one or both reactants (${\text{P}}^{\text{V}}{-}{\text{H}}$-substrate and alkyne) may furnish adducts with a specified structure and properties (polarity, solubility, melting and boiling points). The ${\text{C}}{-}{\text{P}}^{\text{V}}$ derivatives of various organic compounds find use in important areas of human activity.

Despite the achievements (see above), many problems in this field of chemistry still remain unsolved. These include: (i) the absence of microwave-assisted synthetic procedures and catalyst recycling methods for catalytic hydrophosphorylations; (ii) rather limited number of substrates for some reactions; (iii) unexpectedly low reactivity of diynes and, as a consequence, few examples of catalytic reactions involving them; (iv) contradictory data obtained for reaction mechanisms; and (v) the absence of procedures for conducting these reactions in 'green' solvents such as polyethylene glycols, ionic liquids or supercritical media.

Suzuki reaction is the cross-coupling of aryl halides with arylboronic acids (Scheme 71). For the discovery of this reaction, A Suzuki, together with E Negishi and R Heck, were awarded the Nobel Prize in chemistry for 2010.³⁰⁷ This reaction is most efficiently catalyzed by palladium complexes containing highly electron – donating, sterically crowded ²³⁹ phosphine ³⁰⁸ or carbene ligands.³⁰⁹ The use of expanded ring carbenes as ligands holds good prospects, because they surpass both phosphines and five-membered ring carbenes in the electron donor and steric properties.²⁵²

Zoom In Zoom Out Reset image size

The first example of using palladium complexes with expanded ring carbenes in the Suzuki reaction was published in 2006 by Herrmann ad co-workers.³¹⁰ Owing to the high donor capacity of the carbene ligand, complex 53 is very active in this reaction. The authors demonstrated that in the presence of this catalyst, the reaction can involve both deactivated aryl bromides and aryl chlorides. In the case of bromides, the process occurs efficiently in the presence of $0.005\ {\text{mol}}.\hbox{%}$ palladium complex, the turnover number being as high as $1\times 10^6$ in $14\ {\text{h}}$. For chlorides, these characteristics are $0.01\ {\text{mol}}.\hbox{%}$ and $6\times 10^3$, respectively.

Subsequently, ligand 54 was employed to generate the palladium complex in situ.³¹¹ High activity of the complex was demonstrated (turnover number of up to $1.87\times 10^5$, amount of the catalyst $0.005\ {\text{mol}}.\hbox{%}$) but the results were obtained only for highly active aryl iodides.

It was shown ³¹² that on treatment with a base, salts 55 and 56 form in situ palladium complexes, which catalyze the coupling of $p$-acetylchlorobenzene and phenylboronic acid, giving products in almost quantitative yields within $2\ {\text{h}}$ at $50\ ^{\circ}{\text{C}}$. However, it is noteworthy that large loads of palladium salts and ligand precursors were used ($1.5\ {\text{mol}}.\hbox{%}$ and $3.0\ {\text{mol}}.\hbox{%}$, respectively).

The synthesis of a series of palladium complexes based on ferrocenyl diaminocarbene 57 was reported.²⁷⁰ The catalytic activity of the complexes was tested in the reactions of para-substituted aryl bromides with phenylboronic acid. The complexes were found to give products in high yields with a catalyst amount of down to $1\times 10^{-4}\ {\text{mol}}.\hbox{%}$.

Zoom In Zoom Out Reset image size

A publication of 2013 presents the results of a comparative study of the catalytic activity of palladium complexes 58 in the Suzuki reaction of hetaryl halides in water.²⁸⁸ Complexes based on carbenes containing five-, six- and seven-membered rings and bulky aromatic substituents $[{\text{Mes}}, 2,6\hbox{-}{\text{Pr}}_2^{{\text{i}}}{\text{C}}_{6}{\text{H}}_{3}\ ({\text{Dipp}})]$ were used. The complex comprising a six-membered ring carbene and the Dipp substituent proved to be the most active catalyst. Using this catalyst, hetaryl chlorides of various nature can be coupled with boronic acids containing electron-donating, electron-withdrawing or bulky substituents. The developed procedure complies with the green chemistry principles. The following advantages of this procedure should be noted: (a) the reaction is carried out in water without adding organic solvents; (b) the reaction is carried out in air without solvent degassing; (c) a mild and cheap base, ${\text{NaHCO}}_{3}$, is used; (d) small amounts of ${\text{Bu}}_4^{\text{n}}{\text{NBr}}$ are needed; (e) a small excess of boronic acid is taken; (f) high yields are obtained after $1\ {\text{h}}$ of refluxing; (g) high purity of the products containing no side homocoupling products is achieved. High activity of the catalyst is due to the appropriate selection of the donor and steric properties of the ligand. This resulted in the complex that exhibited high activity and was stable in the activated state in the reaction medium. It was also shown that these complexes efficiently catalyze the Suzuki reaction without a solvent.³¹³

The Heck reaction is the coupling of alkyl and aryl halides with alkenes catalyzed by palladium complexes (Scheme 72).³¹⁴

Zoom In Zoom Out Reset image size

The first example of using expanded ring carbenes in the Heck reaction was published in 2010.²⁹⁵ The reaction was catalyzed by a palladium complex of a six-membered ring carbene chemically grafted to a polymeric resin. This catalyst showed a high activity in the coupling of styrene or n-butyl acrylate with aromatic bromides (turnover number of up to $1\times 10^5$). Homogeneous catalyst 59 demonstrated a three times higher activity under similar conditions.

Zoom In Zoom Out Reset image size

Complexes formed by benzyl-substituted carbenes 55 and 56 catalyze the reactions of styrene with para-substituted phenyl bromides containing either electron-withdrawing or electron-donating substituents to give products in $> 70\hbox{%}$ yields in $2\ {\text{h}}$.³¹² It was shown in 2011 that ${\text{Pd}}^0$ complexes with six- and seven-membered ring carbenes (compound 60) catalyze the reaction of $p$-acetylbromobenzene with n-butyl acrylate.²⁸⁶ The product yields varied from 50% to 100% for the reaction time of $3\ {\text{h}}$ depending on the carbene ring size and the nature of aromatic substituents at nitrogen atoms. The amounts of catalysts used were low $(0.1\ {\text{mol}}.\hbox{%})$. The complex containing ferrocenyl ligand 57 demonstrated moderate activity in the coupling of para-substituted aryl bromides with n-butyl acrylate.²⁷⁰

One example of using complexes of expanded ring diaminocarbenes as catalysts in the Kumada reaction was reported in the literature.²⁸⁵ A series of ${\text{Ni}}^{\text{I}}$ complexes 61 have been prepared and characterized by X-ray diffraction. The coupling reactions with aryl chlorides gave 40% – 50% yields (Scheme 73); in the case of aryl fluorides, the product yields did not exceed 30%.

Zoom In Zoom Out Reset image size

Cavell and co-workers ^{315, 316} reported the synthesis of rhodium (62) and iridium (63) complexes containing functionalized aromatic and/or aliphatic substituents. These complexes show high activity and stability in alkene hydrogenation with molecular hydrogen. At room temperature, a pressure of $1\ {\text{atm}}$ and $24\ {\text{h}}$ reaction time, hydrogenation products of various cycloalkanes, styrenes and allylic derivatives were obtained in quantitative yields. Similar rhodium complexes were used in hydroformylation of oct-1-ene.²⁹³ It was shown that the catalyst turnover frequency is $1500\ {\text{h}}^{-1}$; however, the reaction selectivity is low: both linear nonanal and a mixture of isomeric octa-, hepta- and hexanals are formed.

Zoom In Zoom Out Reset image size

In 2010, carbene complexes of rhodium, iridium and palladium chemically grafted to a polymeric support were synthesized.²⁹⁵ The iridium catalyst was active in the reduction of benzaldehyde with isopropyl alcohol. In the presence of $0.1\ {\text{mol}}.\hbox{%}$ catalyst, the reaction proceeded quantitatively over a period of $4\ {\text{h}}$; the product contamination by the metal was $< 50\ {\text{ppm}}$.

It was shown ³¹⁷ that iridium complexes 63 $({\text{X}}={\text{Br}}; {\text{R}}^1=2\hbox{-}{\text{MeOC}}_{6}{\text{H}}_{4};$ ${\text{R}}^2={\text{Mes}},\ {\text{Dipp}};\ n =1, 2)$ are highly active as catalysts in the reduction of ketones with isopropyl alcohol. The reduction of acetophenones and cyclohexanone occurred quantitatively at $80\ ^{\circ}{\text{C}}$ over a period of $24\ {\text{h}}$.

Palladium complexes 60 $({\text{Ar}}={\text{Mes}},\ {\text{Dipp}};\ n =1, 2)$ catalyze the reduction of alkynes with formic acid (Scheme 74).³¹⁸ It was found that at short reaction times, cis-alkenes are predominantly formed. During $24\ {\text{h}}$, the reactions give alkanes almost quantitatively.

Zoom In Zoom Out Reset image size

In 2005, copper complexes 64 and 65 were synthesized ³¹⁹ and found to efficiently catalyze the addition of triethylsilane to carbonyl compounds of various nature. The catalyst amounts were $0.001\ {\text{mol}}.\hbox{%}{-} 0.002\ {\text{mol}}.\hbox{%}$, complete conversion being observed in $1 {-} 3\ {\text{h}}$. Ionic rhodium complex 66a exhibits high activity in the hydrosilylation of alkynes, styrenes and cyclic ketones.²⁹⁴ High yields were obtained in the presence of $0.05\ {\text{mol}}.\hbox{%}$ catalyst in $12\ {\text{h}}$.

Platinum(0) complexes 67 structurally similar to Pd complexes 60 serve as active catalysts for hydrosilylation of ketones, alkenes and alkynes.³²⁰ For a catalyst amount of only $0.005\ {\text{mol}}.\hbox{%}$, high conversion is achieved. It is noteworthy that the selectivity of reactions is also high in most cases.

It was shown that the zwitter-ionic iron complexes 68 can be successfully used in the catalytic hydrosilylation of aromatic aldehydes.³²¹ For the catalyst load of $1\ {\text{mol}}.\hbox{%}$, high conversion is achieved at room temperature in $1 {-} 3\ {\text{h}}$. Analogous zwitter-ionic rhodium complexes 69 are also active in catalytic hydroboration of styrene; however, the reaction rate and selectivity are low.²⁶⁷

Zoom In Zoom Out Reset image size

In 2010, complex 70a was synthesized,²⁶⁸ and the next year, complex 70b was reported (Scheme 75).³²² These complexes proved to be very efficient catalysts for asymmetric borylation of unsaturated compounds. The addition products to Michael acceptors were obtained in high yields [see Scheme 75, reaction (1)].²⁶⁸ The cleavage of allyl ethers to give borylated allyl compounds [reaction (2)] ^{322, 323} followed by their oxidation to diols [reaction (3)] ³²⁴ occur in ${\sim} 80\hbox{%}$ yields. Noteworthy is the formation of chiral bimetallic boron and silicon derivatives [reaction (4)].³²⁵ In 2013, complex 70a was used for catalytic silylation of allyl phosphates with silylated pinacolborane ${\text{Me}}_{2}{\text{PhSi}}{-}{\text{BPin}}$ [reaction (5)].³²⁶

Zoom In Zoom Out Reset image size

In 2004, the catalytic activity of rhodium 62 and iridium 63 complexes in the arylation of carbonyl compounds with arylboronic acids was studied (Scheme 76).³²⁷ The complexes were found to be highly active (yields of up to 99%). Also, complexes containing isopropyl substituents are more active than methyl-substituted derivatives. A decrease in the nucleophilicity of the anion X increases the electrophilicity of the metal atom and, hence, increases the catalytic activity. The iridium complexes are less active than rhodium complexes but they increase the reaction selectivity by reducing the trend of oxidation of alcohols to ketones. The same authors showed that a similar ionic rhodium complex containing the ${\text{BF}}_4^-$ anion is more active than neutral complexes. In this case, the selectivity of the catalytic reaction shifts towards the formation of ketones.²⁹⁴ In a similar reaction, rhodium complexes 62 with benzyl substituents (see Scheme 76) exhibit high activity and selectivity in the formation of alcohols, the yields being 70% – 95%.³²⁸

Zoom In Zoom Out Reset image size

A study of the catalytic activity of rhodium complexes 62a – d [${\text{R}}^1={\text{R}}^2={\text{Mes}}, n =1,\ {\text{X}}={\text{Cl}}$ (a), OTf (b), ${\text{OBu}}^{\text{t}}$ (c), ${\text{OC(O)CF}}_{3}$ (d)], 66a,b and iridium complexes 63 $[{\text{R}}^1={\text{R}}^2={\text{Mes}}, n =1: {\text{X}}={\text{Cl}}, {\text{OC(O)CF}}_{3}]$ in the polymerization of phenylacetylene was reported.³⁰⁰ The polymer thus formed (polyphenylacetylene, PPA) can have different microstructures depending on the type of the catalyst used (Scheme 77). All rhodium complexes showed a high catalytic activity with predominant formation of the cis-isomer of PPA. In the case of catalysts 66a,b having ionic structure, the selectivity towards cis-PPA was 100%. The use of iridium complexes 63 resulted in the selective formation of trans-PPA.

Zoom In Zoom Out Reset image size

Rhodium carbene complexes chemically grafted to a polymeric support were also used to prepare PPA.²⁹⁵ These catalysts performed quantitative polymerization of phenylacetylene $({\text{substrate : catalyst}}=100 : 1)$ at room temperature over a period of $2\ {\text{h}}$ to give cis-PPA. The metal washing out from the support was only 4.5% per catalytic cycle.

Rhodium complexes containing carbenes with a malonate moiety (complexes 69) and carbenes modified by an electrophile (complexes 71a – c) were used for phenylacetylene polymerization.²⁶⁵ Zwitter-ionic complexes 69 were found to be more active than cationic complexes 71. However, complexes of both types show much lower catalytic activity and lower specificity towards cis-PPA than complexes 62a – d, 66a,b containing the tetrahydropyrimidine ligand with more pronounced donor properties.

The cationic metal complex-catalyzed polymerization of ethylene is among the most important processes in the polymer industry. In 2012, Al Thagfi and Lavoie ²⁷² synthesized a series of iron, cobalt and chromium complexes (72 – 74, respectively) containing a potentially tridentate ligand similar to bis(imino)pyridine ligands, which are widely used to design post-metallocene catalysts for the synthesis of polyolefins.³²⁹ An X-ray diffraction study, which confirmed the tridentate coordination of the ligand, was performed only for Cr complex 74. Upon activation by methylalumoxane, iron (72) and cobalt (73) complexes do not exhibit catalytic activity. Conversely, chromium complex 74 shows a substantial catalytic activity. However, it should be noted that catalyst 74 is 3 – 4 orders of magnitude less active than analogous catalysts containing bis(imino)pyridine ligands.³³⁰

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

It was demonstrated ³³¹ that complex 75 efficiently catalyzes the addition of hydrazine to alkynes (Scheme 78). Depending on the substrate nature, the product yields of $> 80\hbox{%}$ can be attained over periods of time from 3 to $6\ {\text{h}}$ either at room temperature or on heating.

Zoom In Zoom Out Reset image size

Recently, the intramolecular hydroamination reaction to give indoles has been studied in detail (Scheme 79). A comparative study of the catalytic activities of gold complexes with carbenes 76 containing five-, six- and seven-membered rings has been performed. The highest activity was observed in the case of the complex with seven-membered ring carbene containing bulky Dipp substituents at nitrogen atoms. This is due to the fact that the ring expansion and an increase in the substituent bulk results in efficient steric and electronic stabilization of the cationic form of the complex, which is the catalytically active species. Quantitative yields of products are attained with $2\ {\text{mol}}.\hbox{%}$ of the catalyst over a period of $15\ {\text{min}}$. The developed catalytic system is most active among those described in the literature to date.

Zoom In Zoom Out Reset image size

Stable expanded ring carbenes have been actively studied in the last decade. These compounds are promising for the design of metal complex catalysts for homogeneous processes. This is due to the large structural diversity, synthetic accessibility and easy functionalization of expanded ring carbenes and the possibility of varying the steric and electronic properties over broad limits. It should be noted, however, that only few catalytic systems able to compete with effective catalysts based on phosphines and other ligands have been developed to date. These systems include catalysts for the cross-coupling of hetaryl halides in water based on palladium complexes, catalysts for asymmetric borylation based on copper complexes, ruthenium catalysts for alkene metathesis, and hydroamination catalysts based on ionic gold complexes. It is evident that the advantages of expanded ring carbenes would be embodied in the near future in new high-performance catalytic processes to run chemical transformations under mild conditions, to activate substrates having low reactivity and to prepare compounds inaccessible by other methods.

Reduction is among the best studied reactions of cyclic ketimines. Recent investigations have paid considerable attention to diastereo- and enantioselective hydrogenation. For instance, 2,5-disubstituted 1-pyrrolines and 2,6-disubstituted piperideines 79 are reduced, most often, by catalytic hydrogenation in ethanol. The reaction is catalyzed by Adams's catalyst $({\text{PtO}}_{2})$,³⁵⁴ palladium on carbon,^{355 – 357} platinum on carbon,³⁵⁸ palladium hydroxide.³⁵⁹ Under the described conditions, syn-addition products 88 were formed, usually in high yields. Complex hydrides (diisobutylaluminium hydride,^{360, 361} lithium aluminium hydride ³⁶²) can also be used to prepare syn-addition products 88 (Scheme 87).

Trans-2,6-disubstituted piperidines 89 are prepared by lithium aluminium hydride reduction in the presence of triethylaluminium. As a result of ${\text{Et}}_{3}{\text{Al}}$ coordination to the imine nitrogen atom, this reaction furnishes anti-addition products 89 (see Scheme 87).^{362, 363}

Zoom In Zoom Out Reset image size

When a substituent is present in position 3 of the cyclic imine, the reduction affords the cis-diastereomer. For example, 2,3-diphenylpyrroline 79a is reduced by sodium borohydride in methanol in the presence of a catalytic amount of acetic acid to give the target product 90 in a yield of 71% (Scheme 88).³⁶⁴

Zoom In Zoom Out Reset image size

When there are no additional substituents in positions adjacent to the imino group, the same reducing agent can produce both syn- and anti-addition products. Indeed, the reduction of optically pure piperideine 79b containing a 4-substituent with sodium cyanoborohydride affords cis-diastereomer 91 in a good yield (Scheme 89).³⁶⁵

Zoom In Zoom Out Reset image size

However, the reduction of 5-aryl-2-substituted piperideines 79c with sodium cyanoborohydride gives, conversely, trans-diastereomers 92 also in good yields (Scheme 90).³⁶⁶

Zoom In Zoom Out Reset image size

Modern medicinal chemistry permanently needs optically active compounds, in particular, amines; therefore, methods for enantioselective reduction of imines are being actively developed.³⁶⁷ Several recent studies have been devoted to the reduction of cyclic ketimines.

$(R)$-1-Phenylpyrrolidine (87a) can be obtained from 2-phenylpyrroline (79d) in almost quantitative yield (96%) and with high enantioselectivity (98% ee). As the hydrogenation catalyst, the titanium ansa-complex 93 with a biphenyl bridge was used in the presence of $2\ {\text{equiv}}.$ of n-butyllithium (Scheme 91).³⁶⁸

Zoom In Zoom Out Reset image size

Recently, ruthenium catalyst 94 was developed affording the preparation of $\unicode{945}$-substituted piperidines 87b from piperideines 79e in high yields (83% – 94%) with moderate enantioselectivity (50% – 61% ee) (Scheme 92).³⁶⁹

Zoom In Zoom Out Reset image size

A higher enantioselectivity was achieved by using a similar ruthenium complex 95. A series of cyclic imines 79 were effectively reduced in the presence of complex 95 giving the products in high yields (up to 98%) and most often with high enantioselectivity (up to 98% ee) (Scheme 93).³⁷⁰

Zoom In Zoom Out Reset image size

The catalytic complex prepared from di-$\unicode{956}$-chlorobis(cycloocta-1,5-diene)diiridium(I) and $(S, S)$-f-BINAPHANE (96) is an efficient catalyst for enantioselective hydrogenation of cyclic imines 79 suitable for preparing pyrrolidine and piperidine derivatives 87 in high yields and with good enantioselectivity (Scheme 94).³⁷¹

Zoom In Zoom Out Reset image size

Further, the method of biocatalyzed reduction of cyclic imines 79 has recently started to be actively developed (Scheme 95). The conversion can be as high as 99%, while the enantiomeric purity of the products is $> 99\hbox{%}$.^{372 – 376}

Zoom In Zoom Out Reset image size

Of particular interest are reactions of imines with C-nucleophiles because they result in the formation of a new C–C bond. For example, the addition of hydrogen cyanide or its synthetic equivalents to imines results in the synthesis of aminonitriles, which are readily converted to $\unicode{945}$-amino acids widely used in pharmacology and medicinal chemistry. Recently, the synthesis of the alkaloid hasubanonine belonging to the hasubanan family was performed.³⁷⁷ The key step in this synthetic route is the Strecker reaction involving imine 79f (Scheme 96). The target compound 97 was isolated as the only diastereomer in a high yield (87%).

Zoom In Zoom Out Reset image size

Alkylation of $\unicode{960}$-donor aromatic systems with imines is classified in different papers as either Mannich reactions or Friedel – Crafts alkylation. The reaction between indoles 98 and cyclic ketimines 79 has recently been studied.³⁷⁸ It was shown that only perfluoroalkyl-substituted cyclic ketimines can participate in this reaction, while non-fluorinated analogues proved to be insufficiently electrophilic. The reaction is catalyzed by Lewis acids, the best results can be achieved by using boron trifluoride etherate. Target compounds 99 were mainly formed in high yields (Scheme 97).

Zoom In Zoom Out Reset image size

Perfluoroalkyl-substituted cyclic imines 79 can also alkylate pyrroles 100 in the presence of boron trifluoride etherate (Scheme 98). This reaction is characterized by unusual $\unicode{946}$-selectivity. Pyrrole reacts with five- and seven-membered ring ketimines to give $\unicode{945}$- (101a) and $\unicode{946}$-substitution (101b) products in ${\sim} 1 : 1\ {\text{ratio}}$. ${\text{CF}}_{3}$-Piperideine alkylates pyrrole only into the $\unicode{945}$-position. In the case of N-substituted pyrroles, electrophilic substitution at the $\unicode{946}$-position unusual for pyrroles is observed, which is due to thermodynamic control of the reaction.^{379, 380} It was shown that other $\unicode{960}$-donor aromatic compounds cannot be alkylated by cyclic ketimines 79.

Zoom In Zoom Out Reset image size

The addition of pentafluoroethyllithium to cyclic imines 79 affords pentafluoroethyl-substituted pyrrolidines and piperidines 102. The oxidative cleavage of 2-pentafluoroethyl-2-furylpyrrolidine proved to be an effective route to $\unicode{945}$-pentafluoroethylproline (103) (Scheme 99).^{381, 382}

Zoom In Zoom Out Reset image size

The addition of methylmagnesium bromide to imine 79g was carried out in the presence of boron trifluoride etherate.³⁸³ The desired compound 104 was isolated as a single diastereomer in high yields (Scheme 100).

Zoom In Zoom Out Reset image size

Recently, a method for the synthesis of chiral amines was developed based on the reaction of imines with chiral Lewis acid 105 to give iminium salts, which are then treated with an organometallic reagent (Scheme 101). For example, upon the reaction with allylmagnesium bromide, complex 106 obtained from imine 79h was converted to optically active pyrrolidine 107 in a high yield and with high enantiomeric excess.³⁸⁴

Zoom In Zoom Out Reset image size

In various publications, the reactions of imines with any of ketones, $\unicode{960}$-donor aromatic derivatives, nitroalkanes or other compounds are classified as the Mannich reactions. For example, on treatment with a methanol solution of HCl, bicyclic imine 79i is converted to tricyclic product 108 in a 83% yield (diastereomer ratio of 4 : 1, the predominant isomer is shown). Then compound 108 is converted to the alkaloid clavolonine (Scheme 102).³⁸⁵

Zoom In Zoom Out Reset image size

In an alternative synthesis of clavolonine,³⁸⁶ bicyclic imine 79j was converted to compound 109 by treatment with HBr generated in situ. The Mannich reaction is accompanied by acid-induced cleavage of methyl ether, formation of the enol ether and cyclization (Scheme 103).

Zoom In Zoom Out Reset image size

One more similar route to clavolonine has been reported.³⁸⁷ Imine 79k used as the precursor was treated with a methanol solution of HCl, which induced decarboxylation accompanied by epimerization of one stereocentre, the subsequent Mannich reaction, acid-induced cleavage of methyl ether, the formation of enol ether and cyclization (Scheme 104).

Zoom In Zoom Out Reset image size

The Mannich reaction was also used in the synthesis of natural alkaloids: lycopodine ^{388, 389} and paniculine ³⁹⁰ (Scheme 105). Instead of ketone, ketone silyl ether was used. Bicyclic imines 79l were treated with zinc triflate to afford tricyclic amines 110 in moderate or high yields.

Zoom In Zoom Out Reset image size

The Ugi reaction is a multicomponent reaction involving isonitrile, an acid, an amine and a carbonyl compound (imine). The mechanism of this reaction comprises several successive steps: imine formation and protonation, isonitrile and carboxylate addition and intramolecular transfer of the acyl group. Upon this transformation, $\unicode{945}$-amino acid amides of complex structure can be obtained from simple molecules (Scheme 106).

Zoom In Zoom Out Reset image size

The three-component Ugi reaction with 2-substituted cyclic imines 79 opened up an effective route to proline and pipecolic acid derivatives with an additional substituent in the $\unicode{945}$-position.^{391, 392} By means of the Ugi reaction, semisynthetic dipeptides 111 containing a natural amino acid moiety and $\unicode{945}$-substituted proline or pipecolic acid have been prepared (Scheme 107).

Zoom In Zoom Out Reset image size

The modified Ugi reaction with cyclic imines 79 and hydrazoic acid (formed from ${\text{TMSN}}_{3}$ and MeOH) being used in place of carboxylic acid opens up the way to 1,5-disubstituted tetrazoles 112, which are most often formed in high yields (Scheme 108).^{393, 394} Using benzylisonitrile as the isocyanide component, upon hydrogenolysis of N-benzyl-substituted Ugi reaction products, it was possible to obtain $1H$-5-substituted tetrazoles 113, which are of considerable interest as organocatalysts (see Scheme 108).

Zoom In Zoom Out Reset image size

Cyclic imines 79 react with the Ruppert – Prakash reagent [trimethyl(trifluoromethyl)silane] in the presence of potassium hydrogen fluoride to give the corresponding $\unicode{945}$-trifluoromethyl pyrrolidine and piperidine derivatives 114 mainly in good yields (Scheme 109). New analogues of nicotine, anabasine and homoanabasine were synthesized in this way.³⁹⁵

Zoom In Zoom Out Reset image size

Pyrroline 79m was used in the synthesis of the immunosuppressive agent FR901483 ³⁹⁶ (Scheme 110). The key step of this scheme was the reaction of the iminium ion, formed from pyrroline 79m in the presence of trimethylsilyl triflate (TMSOTf), with allylsilane 115. This was followed by carbamoylation of the resulting amine and isolation of the Cbz derivative 116.

Zoom In Zoom Out Reset image size

It was demonstrated in a number of papers ^{397 – 399} that $\unicode{945}$-substituted cyclic imines 79 add dialkyl phosphites in the presence of boron trifluoride etherate as the catalyst, thus enabling the synthesis of alkyl aminophosphonates 117 (Scheme 111). Hydrolysis of these products yields potentially biologically active aminophosphonic acids.

Zoom In Zoom Out Reset image size

A similar result was obtained upon the addition of diethyl phosphite to 3-substituted cyclic imine 79n also in the presence of ${\text{BF}}_{3}\cdot\ {\text{Et}}_{2}{\text{O}}$. The reaction gave the target compound as a single diastereomer 118 in 54% yield (Scheme 112).³⁸³

Zoom In Zoom Out Reset image size

This approach was utilized to synthesize a new DEPMPO analogue widely used in EPR as a trap for detection of short-lived radicals. In this case, too, the key step was the addition of dialkyl phosphite 119 to 2-methylpyrroline 79h giving rise to amino-phosphonate 120 as a single diastereomer in 66% yield (Scheme 113).⁴⁰⁰ This product was converted to DEPMPO analogue 121 containing a cholesterol moiety.

Zoom In Zoom Out Reset image size

Cyclic imines 79 react with peroxy acids [$m$-chloroperoxybenzoic acid (mCPBA) or magnesium monoperoxyphthalate (MMPP)] to give the corresponding diastereomeric oxaziridines 122 and 123 in yields ranging from moderate to nearly quantitative.^{401 – 403} On treatment of trisubstituted pyrrolines 79 $({\text{R}}^2={\text{Alk, R}}^3={\text{CO}}_{2}{\text{Et}})$ with magnesium monoperoxyphthalate, two diastereomers with predominance of cis-diastereomer 122 are formed (Scheme 114). Oxaziridines possess a broad synthetic potential. They tend to react with nucleophiles, undergo radical reactions and some other types of chemical transformations.

Zoom In Zoom Out Reset image size

The oxidation of imine 79o to desired oxaziridine 124 in a good yield (76%) was reported.⁴⁰⁴ Product 124 was converted to new efficient reagent 125 for enantioselective oxidation of sulfides to sulfoxides (Scheme 115).

Zoom In Zoom Out Reset image size

Treatment of imines with lithium diisopropylamide (LDA) affords aza-enolates, which are highly reactive nucleophiles. Treatment of deprotonated ketimines with ethyl trifluoroacetate gave acylation products — enaminoketones 126. The by-products formed in this reaction, $N$-(trifluoroacetyl)enamines 127, i.e., aza-enolates, behave as typical ambident nucleophiles. In the case of 2-methylpyrroline (79h), deprotonation occurs at the methyl group, and, after acylation, exocyclic aminoenone 128 is formed.⁴⁰⁵ It is of interest that trifluoroacetylation of imine 79p also involved the methyl group of thiophene, resulting in the formation of compound 129 in 39% yield (Scheme 116).

Zoom In Zoom Out Reset image size

In the case of 2-phenylpiperidine 79e, the reaction gave a considerable amount of dimeric by-product 130. This is caused by addition of the anion derived from the starting imine to trifluoroacylated imine (Scheme 117).

Zoom In Zoom Out Reset image size

An unusual transformation has been reported.^{406, 407} The reaction of spiro compound 131 with 3-aroyl-$1H$-pyrrolo[2,1-$c$][1,4]benzoxazine-1,2,4-thiones 132 or with 5-arylfuran-2,3-diones 133 occurs on heating. In both cases, imine 131 reacts as the enamine tautomer. In the case of compound 132, the addition of imine is followed by heterocyclization yielding polycyclic spiro-coupled product 134 (Scheme 118).

Zoom In Zoom Out Reset image size

Allenyl trimethylsilyl thioketenes undergo $[4+2]$-cycloaddition to imines 79 to give $\delta$-thiolactams 135.⁴⁰⁸ The starting thioketenes are generated in situ from propargyl sulfides 136 upon [3,3]-sigmatropic rearrangement (Scheme 119).

Zoom In Zoom Out Reset image size

New synthetic ketene equivalent 137 ($\unicode{945}$-halovinyl acetate) was studied as both a nucleophilic and electrophilic reagent in a tandem reaction with imines. Diethylaluminium ethoxide was used as the catalyst (Scheme 120). The reaction between 2-methylpyrroline 79h and $\unicode{945}$-halovinyl acetate 137 gives intermediately $\unicode{946}$-lactam 138, which was detected by NMR spectroscopy. However, on attempted isolation by passing through silica gel, lactam 138 decomposed to be converted to amide 139. The other product formed in the reaction was tricyclic compound 140, which was isolated as a by-product in 25% yield (see Scheme 120).⁴⁰⁹

Zoom In Zoom Out Reset image size

Yet another type of cycloaddition reaction based on imines is the reaction of cyclic aza-allyllithium derivatives generated in situ from stannanes 79q with alkenes 141 and polyenes 142a,b (Scheme 121). In most cases, $[3+2]$-cycloaddition products 143 – 146 were isolated in acceptable yields and with 1 : 1 ratio of regioisomers in the case of alkenes 141 and 2 : 3 ratio in the case of cyclohexadiene (142a). Cycloheptatriene (142b) reacts with aza-allyllithium to give cycloaddition product 147 and new cyclic imine 79r with 1 : 1 product ratio and a total yield of 67% (see Scheme 121).⁴¹⁰

Zoom In Zoom Out Reset image size

The same publication describes the reaction of azamethine ylide generated in situ from pyrroline derivative 79q with $N$-methylmaleimide (148). The $[3+2]$-cycloaddition products 149 were formed in moderate yields on refluxing in toluene (Scheme 122).

Zoom In Zoom Out Reset image size

The presented information clearly demonstrates the great synthetic potential of cyclic imines, which can be easily prepared from commercially available chemicals and used in the synthesis of pyrrolidine and piperidine derivatives. The reactions involving these compounds have already found use in the synthesis of biologically active molecules or precursors of natural products. Certainly, the scope of synthetic applicability of five- and six-membered cyclic imines is not exhausted by reactions covered in this review, which were mainly published in the last $10 {-} 15\ {\text{years}}$. The approach based on the use of the open form of cyclic imines, aminoketones, is also of great interest. These bifunctional reagents can be used in the synthesis of heterocycles having an aminoalkyl moiety, some of the latter compounds, for example, tryptamines, exhibit clear-cut physiological activities.^{411 – 416}

The interest of researchers in single crystals as model metal catalysts, which predominated in the 1970 – 1990s, was due to the quest for elucidating the effect of the surface structure on the nature of species formed upon adsorption of small molecules and transformation steps of these species into the intermediates and products of the catalytic reaction. The preparation of single crystals with a definite surface structure for experiments starts with growing perfect single crystalline rods of diameter ${\sim} 1\ {\text{cm}}$, which are then cut along one of the crystallographic directions to form a $1 {-} 2\ {\text{mm}}\hbox{-}{\text{thick}}$ parallel-plate pellet, which thus restricts the contribution of the side surfaces with uncertain structure $(< 10\hbox{%})$. The disorder of the principal faces of the pellet should not exceed $1\ {\text{deg}}$ (better, fraction of a degree) from the chosen orientation. Cutting at a large angle provides allows for the formation of stepped single crystal surfaces with a strictly definite terrace size and monoatomic step structure. The subsequent treatment of the single crystal surface is performed in high vacuum chambers of spectrometers and, as a rule, involves the standard cleaning cycle repeated many times, comprising surface etching by argon ions, heating in an oxygen atmosphere under specified temperature conditions and the final UHV annealing at premelting temperatures.

The most popular method for the reproducible preparation of model supported catalysts is the ultrahigh-vacuum deposition of disperse metal particles on a planar support.^{466 – 468} When this preparation of metal nanoparticles is carried out inside spectral equipment, it is possible, first, to pretreat and prepare of the support and the sputtering system and, second, to transfer the prepared sample into the analysis zone without contact with air. Thus, one can avoid surface contamination and quickly prepare a series of samples with variable size of metal clusters. This circumstance accounts for the large number of studies that used this technique to model real catalysts in order to study the electronic properties and morphology of metal nanoparticles depending on their size. Attempts were also made to study adsorption on disperse metal particles.^{469, 470}

Despite the obvious advantages, for effective use of this preparation technique in the catalytic studies, special precautions in the support preparation should be taken. On the one hand, the support should provide heating of the sample up to the reaction temperature (as a rule, several hundred degrees) in the presence of the reaction medium, while the chemical composition of the support should remain invariable, and ensure the stability of metal particles against sintering, on the other hand, the substrate should have a good conductivity, which is required for investigations by scanning tunnelling microscopy (STM, measures the tunnel current between the tip and the conducting sample), X-ray photoelectron spectroscopy (XPS, removing the electron photoemission-induced charge leading to spectral line broadening) and so on. These problems can be solved without deterioration of the atomic smoothness of the support (which is necessary for STM determination of the size of nanoscale metal particles) by using thin oxide films [e.g., ${\text{Al}}_{2}{\text{O}}_{3}$ films grown on the surface of NiAl(110) metallic single crystals ⁴⁶⁶] and highly oriented pyrolytic graphite (HOPG) the surface of which is modified either chemically or by formation of induced defects.⁴⁶⁸

The procedural details of using physical methods for in situ investigations of the surface and adsorbed species during the catalytic reactions also deserve special discussion, which is beyond the scope of this Section (see, for example, Ref. 471).

Despite the immense interest in the study of silver-catalyzed ethylene epoxidation using physical techniques of surface science,^{472 – 477} the nature of oxygen species active in ethylene epoxidation has long remained a debated issue.^{472 – 475} This situation was related to impossibility of simultaneous investigation of adsorption layers and testing of catalytic properties, while only in this case, it would be possible to elucidate the correlations between the concentration of various adsorbed oxygen species and the yield of the reaction products — ethylene oxide and ${\text{CO}}_{2}$ (resulting from total oxidation of ethylene). The situation changed only in the last decade owing to the advent of the in situ XPS method, which can be used to measure the spectra in the millibar pressure range ⁴⁷¹ in which ethylene oxide starts to be detected in the reaction products by, for example, mass spectrometry.⁴⁷⁸ Figure 13 presents the mass spectra of ethylene oxide measured by the proton transfer reaction mass spectrometry (PTR-MS) for two different pressures; the presented data confirm the existence of the pressure gap problem for this reaction. Indeed, no ethylene oxide signal is present in the PTR-MS spectra at $P({\text{C}}_{2}{\text{H}}_{4}) = 0.0715\ {\text{mbar}}$, and only a pressure increase to $1\ {\text{mbar}}$ results in the appearance of this signal starting from the temperature of $420\ {\text{K}}$.

Zoom In Zoom Out Reset image size

The O1s core level spectra recorded under similar conditions indicate that only nucleophilic oxygen $({\text{O}}_{{\text{nucl}}})$ characterized by the O1s binding energy of $528.2\ {\text{eV}}$ occurs on the surface at low pressure. This state was observed in most studies that used post-reaction analysis; it was shown that this oxygen is active only towards the total oxidation of ethylene to ${\text{CO}}_{2}$ and ${\text{H}}_{2}{\text{O}}$. The pressure increase to $1\ {\text{mbar}}$ results in a second component with a higher binding energy appearing in the spectrum. Results of numerous experiments of this sort at different reactant ratios, temperatures and pressures made it possible to plot linear correlations between the concentration of this oxygen species and the ethylene oxide yield (Fig. 14). This substantiated the involvement of this oxygen, which was called electrophilic $({\text{O}}_{{\text{el}}})$, in the ethylene epoxidation step.⁴⁷⁸

Zoom In Zoom Out Reset image size

An attempt to plot a similar correlation for nucleophilic oxygen showed, as expected, the opposite trend: the yield of ethylene oxide decreased upon the increase in the ${\text{O}}_{{\text{nucl}}}$ concentration (see Fig. 14). On the basis of the obtained data, a reaction mechanism was proposed (Scheme 123).

Zoom In Zoom Out Reset image size

More recently, it was shown that the formation of electrophilic oxygen is related to considerable restructuring of the initial silver surface, which is also caused by interactions with components of the reaction medium at high pressure.⁴⁷⁹

Yet another issue that would have not been resolved without the use of model silver catalysts is interpretation of the size effect in the epoxidation, which is manifested as an increase in the reaction rate by more than an order of magnitude upon the increase in the silver particle size to $> 50\ {\text{nm}}$.⁴⁸⁰ Indeed, the intense O1s signal from alumina in 'real' ${\text{Ag}}/\unicode{945}\hbox{-}{\text{Al}}_{2}{\text{O}}_{3}$ catalysts overlaps the signal from the surface oxygen species, which made the XPS method non-informative for identification of ethylene epoxidating oxygen (${\text{O}}_{{\text{el}}}$, see the previous Section). It was proposed to replace alumina by a carbon substrate, namely by HOPG. Owing to the atomically smooth surface of HOPG, this would enable using STM and SEM methods to determine the silver particle size. This brings about two problems that are to be solved for successful in situ studies. The first problem is related to the weak interaction of silver particles with the defect-free surface of the annealed pyrolytic graphite and, as a consequence, high mobility of silver particles on the atomically smooth HOPG surface with the diffusion coefficient reaching $10^{-8}\ {\text{cm}}^2\ {\text{s}}^{-1}$ (Ref. 481), which may result in fast agglomeration of the silver particles. The second problem is possible degradation of HOPG due to burning of the carbon support under the oxidative atmosphere of the reaction.

For the solution of the first problem, a procedure for forming a defective HOPG surface by soft ion etching was proposed. It was assumed that these defects would serve as sites for crystallization and stabilization of silver particle.⁴⁶⁸ Figure 15 shows the STM and SEM images of two model Ag/HOPG samples with equal atomic ratio ${\text{Ag : C}} \approx 0.5$ prepared by UHV thermal sputtering of silver onto atomically smooth (sm) (see Fig. 15 a) and defective (see Fig. 15 b) graphite surfaces followed by annealing of samples at $250\ ^{\circ}{\text{C}}$. It can be seen that silver particles deposited on a smooth HOPG surface sinter at high temperatures to form agglomerates located near steps, that is, borders of atomically smooth surface terraces. In the case of the Ag/HOPG(Ar) sample with the initially defective surface, the situation is quite different: silver nanoparticles are uniformly distributed over the surface and have a rather narrow size distribution. Analogous conclusions can be drawn from the high-resolution SEM data (see Fig. 15 c, d). These results served for the development of a preparation procedure for model Ag/HOPG catalysts having high stability of Ag particles against sintering at elevated temperature, comprising the following stages: (i) etching of the HOPG surface with argon ions in order to create stabilization sites for the silver particles being sputtered; (ii) silver sputtering (it was shown that the amount of introduced metal determines the average particle size); 3) heating of the pretreated surface at $T \geqslant 250\ ^{\circ}{\text{C}}$ in a vacuum in order to anneal defects and stabilize the surface of the model catalyst.

Zoom In Zoom Out Reset image size

The stability of pyrolytic graphite against the oxidative atmosphere was also verified by SEM. Figure 16 presents the SEM images of the Ag/HOPG(Ar) sample after being used in ethylene epoxidation experiments $({\text{O}}_{2} : {\text{C}}_{2}{\text{H}}_{4}=5 : 1, P=0.25\ {\text{mbar}})$ at various temperatures for several hours. It can be seen that below $230\ ^{\circ}{\text{C}}$, the sample surface is stable (see Fig. 16 a), whereas the temperature rise to $250\ ^{\circ}{\text{C}}$ results in burning-out of graphite layers under diffusing and agglomerating silver particles (see Fig. 16 b). The detected temperature limit of stability of pyrolytic graphite in the reaction medium in the presence of silver particles restricts the application of model Ag/HOPG catalysts for in situ investigation of ethylene epoxidation to a temperature of $230\ ^{\circ}{\text{C}}$, which, however, is high enough for ethylene oxide to be detected among the reaction products.⁴⁷⁸

Zoom In Zoom Out Reset image size

To study the nature of the size effect in ethylene oxidation, two Ag/HOPG(Ar) samples with an average metal particle size of 8 and $40\ {\text{nm}}$, were prepared by the developed procedure. Figure 17 presents the changes in the PTR-MS signals of ethylene oxide and XPS spectra for the Ag/HOPG $(8\ {\text{nm}})$ and Ag/HOPG $(40\ {\text{nm}})$ samples recorded in the $170 {-} 210\ ^{\circ}{\text{C}}$ temperature range.⁴⁸² It can be easily seen that, in full conformity with published data,⁴⁸⁰ ethylene oxide cannot be detected among the reaction products when a sample with fine silver particles is used, whereas the Ag/HOPG sample with coarse particles exhibits catalytic activity in ethylene epoxidation (see Fig. 17 c, d). This behaviour can be interpreted based on analysis of the XPS spectra. The key distinctive feature of the O1s spectrum of the coarse-particle sample is the presence of a peak with the binding energy of $529.2\ {\text{eV}}$ due to nucleophilic oxygen on supported silver particles.⁴⁸² The peak with $E_{{\text{b}}}({\text{O1s}}) = 530.8\ {\text{eV}}$ is a superposition of two oxygen species — oxygen dissolved in the surface layers of silver particles and ${\text{O}}_{{\text{el}}}$ providing the formation of ethylene oxide (see Fig. 17 a, b). Peaks at higher binding energies correspond to oxygen-containing functional groups on the support surface.⁴⁸² Thus, it is the appearance of nucleophilic oxygen that triggers the activity of silver in the ethylene epoxidation, which accounts for the size effect.⁴⁸⁰

Zoom In Zoom Out Reset image size

The question of why the presence of nucleophilic oxygen, which is active in the total oxidation of ethylene, increases the activity in the epoxidation, can be answered by taking into account the transfer of electron density from silver to oxygen and formation of ${\text{Ag}}^{+}$ ions.⁴⁷⁹ These silver ions serve as sites of ethylene adsorption as $\unicode{960}$-complexes without activation of the C–H bond. Only after that, is the adsorbed ethylene able to react with electrophilic oxygen to give ethylene oxide either via the formation of oxymetallacycle ⁴⁷⁷ or directly.⁴⁸⁰ The reaction mechanism proposed relying on model investigations can not only account for specific features of the process such as the need to promote the silver catalyst used in industry by chlorine and caesium compounds and the higher selectivity of bimetallic Ag–Cu catalysts in this reaction but can also help to develop approaches to the design of the optimal silver catalyst for propylene epoxidation.⁴⁸³

The ethylene adsorption as a $\unicode{960}$-complex is also the key factor ensuring the high selectivity of oxidative acetoxylation of ethylene. Along with this desired reaction, other side reactions can occur in the system and thus decrease the selectivity towards vinyl acetate (Scheme 124).

Zoom In Zoom Out Reset image size

The contribution of these reactions has been estimated.⁴⁸⁴ It was shown that at the temperature of vinyl acetate synthesis $(413\ {\text{K}})$,⁴⁸⁵ the rate of palladium-catalyzed oxidation of acetic acid is relatively low, and acetic acid addition does not affect the kinetics of ethylene oxidation. Similar experiments dealing with the effect of vinyl acetate on the ${\text{CO}}_{2}$ formation rate demonstrated that introduction of $3.5\ {\text{kPa}}\ {\text{CH}}_{3}{\text{CO}}_{2}{\text{CH}}{=\!=}{\text{CH}}_{2}$ into the reaction mixture comprising ethylene, acetic acid and oxygen resulted in an increase in the total oxidation rate by only 5%.⁴⁸⁴ Thus, it was concluded that the major contribution to the decrease in the selectivity towards vinyl acetate is made by the total oxidation of ethylene, which, in turn, depends on the type of ethylene adsorption.

The ethylene adsorption and decomposition have been studied in detail by temperature-programmed desorption (TPD) on model ${\text{Pd/SiO}}_{2}$ and ${\text{Au}}{-}{\text{Pd}}/{\text{SiO}}_{2}$ catalysts.⁴⁸⁶ Uniform heating of the sample, without which TPD experiments are impossible, has been provided by a special design of the model catalyst based on a refractory molybdenum single crystal. A ${\text{SiO}}_{2}$ film grown epitaxially on the Mo(110) surface is complementary to the metal structure.⁴⁸⁷ One monolayer (ML) of palladium and then gold in different amounts (from 0.1 to $1\ {\text{ML}}$) were sputtered on the prepared ${\text{SiO}}_{2}$ film and then the samples were annealed in a vacuum at $800\ {\text{K}}$. A study of adsorption of deuterated ethylene on the samples prepared in this way showed that in the case of a monometallic palladium sample, adsorption at $90\ {\text{K}}$ gives rise to a broad desorption peak with $T_{\max}=250\ {\text{K}}$. Upon introduction and increase in the content of gold, this peak narrows down and the desorption maximum shifts towards lower temperatures (down to $215\ {\text{K}}$) (Fig. 18 a). In view of the fact that in the case of bimetallic samples, ${\text{D}}_{2}$ formed upon ${\text{C}}_{2}{\text{D}}_{4}$ decomposition disappears from the TPD spectra (see Fig. 18 b), the low-temperature peak in the TPD spectrum of ${\text{C}}_{2}{\text{D}}_{4}$ was assigned to ethylene desorption from the $\unicode{960}$-complex. Considering the intense ${\text{D}}_{2}$ desorption peaks recorded after adsorption on the monometallic palladium sample or on the samples with low gold contents (see Fig. 18 b ), the high-temperature shoulder in the TPD spectrum of ethylene was identified as desorption from the ${\text{di}}\hbox{-}\unicode{963}\hbox{-}{\text{state}}$, which is the intermediate for the total ethylene oxidation pathway.⁴⁸⁶ These data demonstrate that the introduction of gold changes the surface composition of particles.

Zoom In Zoom Out Reset image size

The surface composition of the prepared bimetallic catalysts was studied by the ion scattering spectroscopy (ISS),⁴⁸⁸ while the atomic structure of the surface sites was investigated by IR spectroscopy.⁴⁸⁹ These experiments were carried out using model samples of a different type in which gold and palladium were successively sputtered onto a cleaned surface of the Mo(110) face. Owing to the high melting point of molybdenum, the samples could be heated up to the gold $(1400\ {\text{K}})$ and palladium $(1420\ {\text{K}})$ desorption temperatures,⁴⁸⁶ which was used for calibration and for estimation of the amount of introduced metal depending on the sputtering time. During the deposition of metal onto metal, layer-by-layer coverage by the Frank – van der Merwe mechanism ⁴⁹⁰ takes place to give a continuous film, which is necessary for efficient use of ISS and IR spectroscopy. If the metal is deposited on silica, as in the previous example, the stronger metal – metal interaction (cohesion) compared with the metal – oxide interaction (adhesion) provides the formation of 3D particles by the Volmer – Weber mechanism.^{486, 490}

Two samples with different metal sputtering sequences, (i) Pd/Au/Mo(110) and (ii) Au/Pd/Mo(110), were prepared for these experiments. Figure 19 a presents the initial ion scattering spectra, which represent the dependence of the number of elastically reflected low-energy ${\text{He}}^{+}$ ions on the ion energy, while Fig. 19 b shows the temperature dependence of the palladium and gold concentration in the surface layer. The ${\text{He}}^{+}$ ions monochromatic in energy are formed in a special source and reflected from palladium and gold atoms and, hence, they are manifested in the ion scattering spectra at different energies (see Fig. 19 a ), which enables chemical analysis of the surface. It can be seen that successive deposition of gold and palladium gives rise to the corresponding peaks in the ion scattering spectra; the introduction of palladium hides almost completely the gold signal, which proves that the films grow by the Frank – van der Merwe mechanism. An increase in the temperature leads to blending of the gold and palladium films to give an alloy, which is manifested as a gradual decrease in the palladium signal and an increase in the gold signal. Quantitative data (see Fig. 19 b ) demonstrate that the system achieves at equilibrium at $700\ {\text{K}}$, and then up to the annealing temperature of $900\ {\text{K}}$, the surface composition does not change. The same state of the surface is also achieved for the second sample in which gold was deposited after palladium (see Fig. 19 b). Further heating of the samples results in the desorpion of gold and then palladium (see Fig. 19 a). In full conformity with surface tension data for gold and palladium, the equilibrium surface of the alloy is enriched in gold (${\sim} 80\hbox{%}$ Au and ${\sim} 20\hbox{%}$ Pd). The surface segregation of gold was detected previously for massive Pd–Au alloys.^{491, 492}

Zoom In Zoom Out Reset image size

The data shown in Fig. 19 also indicate that by varying the annealing temperature, it is possible to control the composition of the alloy surface. Meanwhile, it is of interest to study the structure of the palladium surface sites as it is diluted with gold. This was done using the IR spectroscopy for adsorbed CO the molecular vibration frequency of which depends on the surface site geometry.⁴⁸⁹ Three types of CO adsorption are distinguished in the literature: (i) three-point adsorption where the CO molecule is located above a triangle of palladium atoms, (ii) bridging adsorption where the CO molecule is bonded to two neighbouring palladium atoms, (iii) terminal adsorption where the CO molecule is located above one palladium atom.^{493 – 496} Figure 20 shows the IR spectra of adsorbed CO for samples prepared by successive sputtering of gold and then palladium onto the surface of the Mo(110) single crystal followed by annealing at 600 and $800\ {\text{K}}$. These temperatures were selected in view of the data presented in Fig. 19; in the first case, the alloy surface contains $1.5\ {\text{times}}$ more palladium than in the second case. From comparison of the IR spectra, it can be seen that the major difference between there two model catalysts is the appearance of the bridging CO form adsorbed on palladium sites in the sample annealed at lower temperature. This is indicated by the signal at ${\sim} 2030\ {\text{cm}}^{-1}$ (Ref. 495), which appears in the spectrum in addition to the intense signal at $2050 {-} 2080\ {\text{cm}}^{-1}$ due to the terminal species ${\text{CO}}_{{\text{ads}}}/{\text{Pd}}$.⁴⁹⁴ Because of weak interaction of CO with gold $(8 {-} 10\ {\text{kcal mol}}^{-1})$ (Ref. 497), the contribution of ${\text{CO}}_{{\text{ads}}}/{\text{Au}}$ to the IR spectra recorded at room temperature (see Fig. 20) should be excluded. This result indicates that in the equilibrium state, the surface of the Pd–Au alloy contains only single palladium atoms surrounded by only gold atoms. As a consequence, ethylene can be adsorbed on this surface only as $\unicode{960}$-complexes. Annealing at lower temperature retains some neighbouring palladium atoms. The possible formation of ${\text{di}}\hbox{-}\unicode{963}\hbox{-}{\text{complexes}}$ by adsorbed ${\text{C}}_{2}{\text{H}}_{4}$ would increase the probability of the total oxidation pathway for ethylene, thus decreasing the process selectivity with respect to vinyl acetate. A practically important recommendation that follows from the performed fundamental research of the oxidative acetoxylation of ethylene on model Pd–Au catalysts is to use high annealing temperatures.

Zoom In Zoom Out Reset image size

Taking into account these recommendations, the ${\text{Pd}}(1\hbox{%})-{\text{Au}}(0.5\hbox{%})/{\text{SiO}}_{2}$ catalyst was prepared by incipient wetness method and annealed at $500\ ^{\circ}{\text{C}}$.⁴⁹⁸ Comparison of its catalytic properties with those of the monometallic ${\text{Pd}}(1\hbox{%})/{\text{SiO}}_{2}$ catalyst prepared in a similar way showed ⁴⁹⁹ that the addition of gold increases by more than an order of magnitude the catalyst activity expressed as the number of vinyl acetate molecules formed in $1\ {\text{s}}$ on one surface palladium atom; the selectivity in this case reaches 96% (for comparison, the best monometallic catalyst provides only 90% selectivity).⁴⁹⁸ High annealing temperatures $(500 {-} 600\ ^{\circ}{\text{C}})$ are also indicated in modern patents devoted to improvement of palladium – gold supported catalysts for vinyl acetate synthesis.⁵⁰⁰

Thus, the understanding of the mechanisms of catalytic reactions and determination of the active site structure may serve for pronounced improvement of the reaction selectivity towards the target product. The use of this approach will be extended in the near future; however, there are some obvious limitations to applying it to complex organic reactions in the liquid phase. The first step in solving this problem may be the post-reaction analysis of heterogeneous catalysts by surface-sensitive physical methods for detection of changes in the structure and chemical state of the surface upon the catalytic reaction. The atmosphere-free loading technique (dry chamber) becomes especially significant for transfer of samples from the reaction medium inside spectrometers or microscopes. Yet another direction for the development of this area is to choose simple compounds that model some organic reaction (hydrogenation, oxidation, functionalization, etc.).

The catalytic hydrogenation of alkynes has a fundamental value both for laboratory practice and for chemical industry,⁵²¹ because it can create trans- and cis-alkene moieties serving as building blocks in fine organic synthesis. The hydrogenation is typically catalyzed by supported metallic catalysts. Therefore, numerous research teams all over the world are engaged in the research of the main regularities of the catalytic action of these materials.^{522, 523} The key goals are, first, to minimize the conversion of alkenes formed in the reaction to alkanes and, second, to select process conditions that would result in the desired stereoselectivity.

For cis-alkene synthesis from alkynes, the heterogeneous Lindlar catalyst (Pd–Pb) is often used. A considerable drawback of this catalyst is the toxicity of lead compounds, which are needed for partial catalyst deactivation for preventing the reduction of the target products (alkenes) to alkanes. In addition, this catalyst is inapplicable to some substrates. For example, terminal alkynes cannot be selectively converted to alkenes as they are rapidly hydrogenated to alkanes. In some cases, the use of the Lindlar catalysts is accompanied by instability and irreproducibility of the results under experimental conditions.⁵²⁴

A promising way for attaining selective hydrogenation of the ${\text{C}}{\equiv} {\text{C}}$ bond is the use of bimetallic catalysts. The studies along this line carried out to date demonstrated that the highest selectivity results from the use of bimetallic catalysts containing a Group 8 metal (usually Pd) as one of the active components.^{525, 526} The key role is then played by two factors: the nature of the second metal and the degree of homogeneity of the bimetallic nanoparticles that are formed during the catalyst preparation.

For example, the introduction of Zn or Ag into a Pd catalyst considerably increases the selectivity in alkyne hydrogenation to alkenes, although the activity somewhat decreases.⁵²⁷ The observed increase in the selectivity to alkenes is usually attributed to electronic and/or ligand effects and to the disappearance of the palladium hydride $\unicode{946}$-PdH phase. The two first-mentioned factors are tightly interrelated. The presence of the second metal on the surface and in the bulk of a bimetallic alloy particle changes the electronic properties, resulting in a decrease in the desorption energy of the alkene being formed, which thus becomes lower than the activation energy of the second $({\text{alkene}} \to {\text{alkane}})$ hydrogenation step. Hence, alkene desorption rather than further hydrogenation becomes the predominant reaction pathway.⁵²⁸ The second, geometric effect of the introduction of the second metal is 'dilution' of the Pd surface layer by inactive metal atoms, which reduces the probability of formation of multiatomic sites in which strong multicentre adsorption of alkylidene intermediates is possible. These intermediates are believed to be responsible for the direct hydrogenation of alkynes to alkanes.⁵²⁹

A considerable role is also played by the degree of homogeneity of the prepared bimetallic nanoparticles. In the ideal case, the composition of every nanoparticle should correspond to the component ratio of the whole catalyst; however, obtaining this high degree of homogeneity is a challenging task. An effective process for the preparation of highly homogeneous bimetallic catalysts is the use of heterobimetallic palladium acetate complexes with the second metal ${\text{PdM(OAc)}}_{4}({\text{OH}}_{2})\ ({\text{M}}={\text{Zn, Ce, Co, Ni}})$.^{527, 530} These catalysts are more active in the hydrogenation of alkynes to alkenes than the catalyst samples prepared by the traditional deposition of the metal precursors from solutions of individual salts.⁵²⁷

An ingenious method for increasing the selectivity of bimetallic catalysts for hydrogenation of alkynes to alkenes is the use of intermetallic compounds in which the electronic effect of the second component on the adsorption and catalytic properties of the active metal is much more pronounced than in disordered alloys. Indeed, it was ascertained ⁵³¹ that the ${\text{Ni}}_{3}{\text{Ge}}/{\text{MCM}}\hbox{-}41$ catalyst performs highly selective hydrogenation of acetylene in an excess of ethylene up to almost complete conversion of the alkyne. Similar results have been obtained ^{532, 533} for the intermetallic compounds PdGa and ${\text{Pd}}_{3}{\text{Ga}}_{7}$.

It should be noted, however, that the results of reactions of fine organic synthesis in the liquid phase catalyzed by supported bimetallic catalysts are not always consistent with the results obtained for the gas-phase heterogeneous catalysis. For example, liquid-phase hydrogenation of phenylacetylene on the above-mentioned PdGa intermetallic ^{532, 533} showed only a minor increase in the selectivity with respect to that on monometallic Pd catalyst.⁵³⁴ Furthermore, it was found that the same is true for the Lindlar catalyst. The authors attributed these significant discrepancies of the results obtained for the gas-phase hydrogenation of acetylene and liquid-phase hydrogenation of phenylacetylene to specific effect of the liquid phase on the surface chemistry of the PdGa intermetallic compound, resulting in partial oxidation and pronounced surface segregation of gallium.

Nevertheless, in some cases, the results of the gas-phase and liquid-phase hydrogenation are well correlated with each other. Thus for catalysts based on the bimetallic complexes ${\text{PdM(OAc)}}_{4}({\text{OH}}_{2})\ ({\text{M}}={\text{Zn, Ce, Co, Ni}})$, it was found that introduction of the second metal leads to higher selectivity of diphenylacetylene hydrogenation, as in the case of gas-phase hydrogenation, and results in higher percentage of the target cis-isomer in the reaction products. It was shown that the activity of catalysts deposited from bimetallic complexes considerably depends on the nature of the second metal and increases in the order ${\text{PdZn}} < {\text{PdCo}} < {\text{PdNi}}$.⁵³⁵ Therefore, the results confirm good perspectives for the catalysts based on bimetallic palladium acetate complexes in the selective hydrogenation of ${\text{C}}{\equiv} {\text{C}}$ bonds.

Homogeneous bimetallic particles can also be prepared in solution. Spee et al.⁵³⁶ reported a detailed investigation of the catalytic properties of the ${\text{Pd}}{-}{\text{Cu/SiO}}_{2}$ bimetallic systems in the selective hydrogenation of substituted alkenes and propargyl alcohols. The Pd–Cu nanoparticles were obtained by the reaction of lithium di(4-tolyl)cuprate with palladium acetate, resulting in deposition of Pd–Cu particles on the silica gel surface. The formation of the Pd–Cu alloy was verified by high-resolution electron microscopy and EXAFS. The ${\text{Pd}}{-}{\text{Cu/SiO}}_{2}$ catalysts demonstrated much higher selectivity than the Lindlar catalyst towards the formation of cis-alkenes in the hydrogenation of mono- and disubstituted alkynes.

Interesting results of selective hydrogenation of terminal alkynes were obtained in the already mentioned study by Lin et al.⁵²⁰ It was found that bimetallic Rh–Co nanoparticles provide high yields of terminal alkenes and diphenylethylene (Scheme 128). Hydrazine hydrate served as the hydrogenating agent and, hence, the reaction could be easily conducted using common laboratory ware for organic synthesis.

Zoom In Zoom Out Reset image size

A new method for the preparation of bimetallic Rh–Ag nanoparticles has been proposed,⁵³⁷ namely, entrapment of the catalytically active Rh within a silver metal matrix. This was done using an original procedure based on reduction of ${\text{AgNO}}_{3}$ with Zn metal to give finely dispersed Ag in solution. This is accompanied by entrapment of the Rh complex, giving rise to a bimetallic particle. This Rh–Ag complex showed high selectivity towards the formation of cis-stilbene. This result is even more notable because monometallic Rh nanoparticles provide the formation of only exhaustive hydrogenation products.

The synthesis of trans-alkenes from alkynes is a more complicated task than the synthesis of cis-alkenes by the hydrogenation reaction. An ingenious method for the synthesis of trans-isomers upon diphenylacetylene hydrogenation was proposed by Komatsu et al.⁵³⁸ Using the ${\text{Pd}}_{3}{\text{Bi/SiO}}_{2}$ catalyst containing a palladium intermetallic compound as the active component, the authors were able to achieve high selectivity in the transformation of diphenylacetylene to cis-stilbene up to a $> 90\hbox{%}$ conversion. The catalyst shows much higher selectivity than the classical Lindlar catalyst. By using ${\text{Pd}}_{3}{\text{Bi/SiO}}_{2}$ as the hydrogenating component and by mechanically mixing it with the H-USY zeolite, the authors achieved a ${\sim} 74\hbox{%}$ yield of trans-stilbene. The ${\text{Pd}}_{3}{\text{Bi/SiO}}_{2}$ catalyst performs the selective hydrogenation of diphenylacetylene to cis-stilbene, whereas on the zeolite component, the resulting cis-stilbene isomerizes to the thermodynamically more stable trans-product (Scheme 129).

Zoom In Zoom Out Reset image size