Pyrethroid insecticides. Chapter III: Synthesis of 3-phenoxymandelonitrile

Racemic and enantiopure 3-phenoxymandelonitrile are valuable building blocks for the synthesis of pyrethrin insecticides cypermethrin and deltamethrin. Their synthesis involves two crucial steps: the synthesis of 3-phenoxybenzaldehyde which is produced through ether coupling, and its hydrocyanation.

Several other conditions and catalysts have been successfully used to perform the cyanation of benzaldehyde. For example, it has been reported that zirconium(IV) t-butoxide favors the transhydroxycyanation of benzaldehyde (11) from acetone cyanohydrin (13a) (Scheme 8, entry c). 30 Otherwise acetoxy nitriles can be synthesized directly from the benzaldehyde and acetyl cyanide 13b in the presence of potassium carbonate (Scheme 8, entry d). 31 Scheme 8. Synthesis of mandelonitrile and O-protected forms from benzaldehyde. 26,28,[30][31][32][33][34]36,38,39,[41][42][43] Similarly, t-butyl cyanoformate (13c) reacts with benzaldehyde (11) to provide t-butyloxocarbonyl mandelonitrile 12c through triethylamine catalysis (Scheme 8, entry e), 28 and diethyl cyanophosphonate (13d) reacts with (11) extremely rapidly at room temperature (0.1h) in the presence of triethylamine (10%, neat) to produce the diethyl phosphonate 12d in almost quantitative yield (Scheme 8, entry f). 32 Trimethylsilylcyanide (TMSCN) (13e) is by far the most efficient reagent for the addition of a cyano group to aldehydes and ketones as well. The reaction was originally carried out by heating the two reactants in neat form for several hours (Scheme 8, entry g) and related conditions are listed in Scheme 9. 33,34 It has been suggested 33 that replacing a proton in HCN for silicon in Me3SiCN (13e) favors the cyanation of benzaldehyde (11) by ca. 20 kcal mol -1 . Furthermore, the stability of the TMS adduct 12e, conferred by an energetically favorable Si-O bond, (i) prevents the reverse cyanation reaction from occurring and hence removes a potential pathway for racemization for enantiopure compounds (ii) offers an easy synthesis and recovery of the related cyanohydrin 12a by desilylation on hydrolysis (1 N HCl, rt, 6 h) that occurs with complete stereocontrol. 35

Synthesis of enantiopure (S)-3-phenoxymandelonitrile and protected species from 3-phenoxybenzaldehyde
Enantiopure cyanohydrins were known in nature long before 3-phenoxymandelonitrile (S)-(3a) was found to be a crucial partner for the synthesis of deltamethrin (1b). Unfortunately, most of the hydroxynitrile lyases that perform such synthesis on analogous benzaldehydes proved to produce the (R)-enantiomers. 5,23,24 Thus, intensive research had to be carried out to achieve the desired transformation as it will be reported below (Section 2.4.3.2.). In fact, many approaches to enantiopure cyanohydrins have been proposed since 1990 and almost all of them have been successfully applied to 3-phenoxymandelonitrile (S)-(3a). They are shown in the following sections and are summarized below. (i) the synthesis of the racemate as reported above (Section 2.3.) and after derivatization leading to a diastereoisomeric mixture, (ii) an adequate method of separation of the diastereoisomers, and (iii) recovery of the 3-phenoxymandelonitrile (S)-(3a) without epimerization, (iv) recycling the unwanted stereoisomer/enantiomer that amount at least 50% that is otherwise lost. This recycling can either involve epimerization leading back to a mixture of diastereoisomers/enantiomers and repeating the separation/isolation processes or better inverting selectively their (R) to the (S) benzylic carbon either directly on the unwanted separated species or after adequate treatment.
As already pointed out, epimerization of the cyanohydrin (R)-3 easily takes place in basic media by equilibration involving the intermediate formation of the 3-phenoxybenzaldehyde-(4a) and hydrogen cyanide or through epimerization at the benzylic carbon of a suitably "protected" compound.
Thus, enantiopure biocartol (16b), from the ozonolysis of enantiopure (1R)-cis-chrysanthemic acid (2e), 66 has proven to be an extremely valuable reactant for the resolution of racemic 3-phenoxymandelonitrile (3a) ( Scheme 19). 63 It offers the advantage to produce the conjugate diastereoisomeric mixture 10 in an acid medium, under conditions under which both the lactol 16b and the cyanohydrin 3a are stable. The reaction takes advantage of the easy substitution, after protonation, of its hydroxyl group by the alkoxy group of 3phenoxymandelonitrile (3a) at its anomeric center on which a formal cation can be stabilized by the remaining carboxy oxygen.
This exceptional structural arrangement also allows the reverse reaction leading to biocartol and recovery of the alcohol (cyanohydrin) after resolution keeping untouched all chiral centers. 63 The "acetalization" takes place by removal of water under reduced pressure whereas the recovery of the cyanohydrin is carried out in the presence of an excess of water. Under these conditions, the enantiopure cyanohydrin (S)-3a is obtained from the enantiopure "acetal" (S)- (10) as the result of the resolution of the racemate (R,S)-10 by crystallization of the later in iso-propanol (Scheme 19). 63 Biocartol (16b) 66 is also extremely valuable for resolution of racemic mixtures of a wide variety of alcohols including allethrolone (Chapter 2) 64,65,66 and trans-chrysanthemol. 64,65 Interestingly, if the crystallization is performed under suitable conditions in the presence of triethylamine, epimerization takes place at the benzylic carbon of 10 and crystallization of (S)-10 shifts the equilibrium towards its formation (Scheme 19, entry b). 63 The resulting acetate (S)-1c can be directly used (Scheme 17) 57 or deprotected to the cyanohydrin then acylated with a permethrinic acid chloride (Scheme 14, entry a). 50 The lipase selected and the conditions used are very important for the success of the process especially for industrial purposes. The key factors are: the accessibility of the lipase, the selectivity of the process, the nature of the enantiomers formed, the pH of the medium, the ease of recovery of the cyanohydrin and/or its acetate, the efficiency of the recycling of the compounds possessing the unwanted (R)-stereochemistry and the reuse of the lipase.

Chirality transfer in asymmetric synthesis of 3-phenoxymandelonitrile.
Another approach to of 3phenoxymandelonitrile (S)-(3a) has been carried out by Johnson on 3-phenoxybenzaldehyde (4a) and enantiopure 1,3-propane diols (17) possessing at least one asymmetric center as reported in Scheme 24 and Scheme 25. 70 The successful strategy 70,71 involves (i) the synthesis of the acetal 9 derived from 2,4-pentanediol (17) and 3-phenoxybenzaldehyde (4a) using pyridinium tosylate as catalyst, 72 (ii) the Lewis acid-assisted ring opening of acetal 9 by trimethylsilyl cyanide 70,71 leading to the β-benzyloxy alcohol 18a (iii) its transformation to 3phenoxymandelonitrile (3a) in good yield in a two-step process that involves the synthesis of β-benzyloxy methyl ketone 19a on oxidation with pyridinium chlorochromate (PCC), 73 and its decomposition through a β-elimination reaction catalyzed by an acidic medium (Scheme 24). 70 The whole process initiated by the prochiral (2S,4S)-pentane diol (17a) provides the (S)-mandelonitrile (S)-(3a) in almost quantitative yield and high stereocontrol (ee: 91%) but the chiral center is generated at the expense of the two chiral centers of the pentane diol (17a) that are both destroyed in the process! Scheme 24. Chirality transfer in asymmetric synthesis of 3-phenoxymandelonitrile (S)-3a from 3-phenoxybenzaldehyde (4a) through its chiral dioxolane from enantiopure (S,S)-2,4-pentanediol. 70 The key step of the process involves the titanium tetrachloride assisted ring opening of acetal 9a using excess of trimethylsilyl cyanide. It is temperature-dependent with a de of 95% at-78 °C and only 84% at 0 °C. 70 The reaction is believed to occur through an SN2-like transition state A (Scheme 24, route a), stabilized by lengthening of the 2,3-bond of the ground state conformer A with consequential relief of the large 2,4-diaxial H/Me interaction. 70 As can be seen, the same is not operative if route b would have been instead involved (Scheme 24, route b). 70 The reaction involving (S)-2,4-butanediol (17b), offers advantages of (i) its ready access from the related βketo ester by yeast reduction, 69 and (ii) a 1/1 asymmetric center transfer, as compared to the one more being lost in the previous approach (Scheme 25, compare to Scheme 24). 70 Differences between the related approaches disclosed in Scheme 24 and Scheme 25 are listed below: (i) the benzylic carbon of acetal 9b 69 is chiral whereas that of 9a is prochiral. 70 Therefore, 9b can exist in two diastereoisomeric forms. Fortunately however, the cis-stereoisomer cis-9b, the more stable of all possible forms, is produced on equilibrating conditions involving for example pyridinium tosylate (97% yield). 72 (ii) Titanium tetrachloride catalyzed ring opening can take place by cleavage of one the C-O bonds leading either to an aldehyde after oxidation of the primary alcohol (S)-18b (Scheme 25, Route a) or to a methyl ketone through the secondary alcohol 18b' (Scheme 25, Route b). 69 It was indeed found that the cleavage leads to the primary alcohol (S)-18b (de: 90%) and to the secondary alcohol 18b' (de: 0%) in identical amounts if the reaction is performed at -78 °C. 69 This ratio varies from 1/1 to 99/1 depending on the conditions. It has been however found that slow addition at 0 °C of TiCl4 to a mixture of cis-9b and 3 equivalents of TMSCN (13e) provides, after "deprotection", 3-phenoxymandelonitrile (S)-(3a) in good yields (97%) and high enantiomeric excess (ee 97%). 70 The preferential cleavage of the C-O bond leading to the primary alcohol can be rationalized by preferential TiCl4 complexation at the least hindered site on the acetal 9b. 70 Scheme 25. Chirality transfer in asymmetric synthesis of 3-phenoxymandelonitrile (S)-3a from 3-phenoxy benzaldehyde (4a) through its chiral dioxolane from enantiopure (S)-2,4-butanediol. Chemically catalyzed asymmetric cyanohydrin synthesis has been the subject of several publications including comprehensive reviews. [6][7][8][9] Over the past two decades, significant advances have been made towards developing asymmetric cyanohydrin synthesis of aldehydes possessing different structures including aliphatic aldehydes, [6][7][8]113,114 and ketones, 6-9 but relatively few have been performed on 3phenoxybenzaldehyde (4a). The reviews, especially that of Holmes 7 and later North 9 efficiently cover the field, and we direct the reader's attention to them. We have restricted our contributions to two starting materials: (i)-3-phenoxybenzaldehyde (4a) (Scheme 28, Table 3, Scheme 29, table 4) the substance requested for the synthesis of cypermethrin and deltamethrin and (ii) benzaldehyde (11) (Scheme 27, table 2), 36,74-100 the parent aromatic compound that has been used as a model in almost all the researches published. We have gathered in Table 3 35,36,75,77-80,85-90,97-99,101 and Table 4 102,103,104,105 the reaction that leads directly to the chiral cyanohydrins 3a and 12a, respectively, or to their trimethylsilyl precursors 3e and 12e. We have gathered in Scheme 29 the reactions of 4a that lead instead to the related chiral phosphonate 3g, acetate 1c or methoxy carbonate 3f. Those are quite difficult to deprotect to the related cyanohydrins 3a and 12a and are moreover poorly enantioselective, that preclude their use for an industrial synthesis of 3a.
The trimethylsilylated cyanohydrins occupy a special place since as already pointed out trimethysilyl cyanide their precursor is easily prepared and commercially available, they are produced easily in high yield, they are stable and the enantiopure derivatives are stereoselectively desilylated in acidic media.
The synthesis of enantiopure cyanohydrins or their silyl-protected analogs have been thus carried out using a metal as pre-catalyst and a chiral-ligand that combines to produce the chiral catalyst that ideally should be reused. The different metals used are displayed on a periodic table of elements (Table 1): they include those belonging to the s-block (Li, Mg); p-block (B, Al, Sn, Bi); d-block (Sc, Ti, V, Mn, Co, Y, Zr, Re) and f-block (La, Sm, Eu, Yb, Lu) -metals. [6][7][8][9] The nature of the metal (entry D), the structure of the related metallic compounds (entry E), the structure of the related ligands (entry B) and the postulated structure of the catalyst (entry F) are listed in Tables 2-4. We have also gathered succinct experimental details about each cyanation method (entry E). The yield (entry G) and the enantiomeric excess (entry H) in cyanohydrin or its protected form are shown in the Tables 2-4. The following sections will discuss sequentially some of these parameters.  Table 2, entries 8f, 14a, 17f, 18b, 18c, 19a-c; Table 3, entry 5;  Table 4, entries 3, 4; [Bi]: Table 2, entry 16; Table 3, entry 11),  • Titanium has been by far the most used metal providing the highest yields and enantiomeric excesses. It has been introduced as titanium tetra-ethoxide, -isopropoxide or -chloride, • Vanadium introduced as oxovanadium sulfate has proven in rare cases to be superior ( Table 2, compare entry 8d to entries 7, 8a, 8b), • Aluminum as aluminum trichloride, dimethyl aluminum chloride, aluminum tetraisopropoxide and triethyl aluminum proved reasonably efficient, in particular when used in conjunction with lithium (Table 2, entries  14a, 17f; Table 3, entry 5; Table 4, entries 3, 4), • Samarium as samarium trichloride (         Krief, A.

Scheme 28.
Reactions that allow the enantioselective synthesis of 3-phenoxymandelonitrile-phosphonate 3g, acetate 1c and carbonate 3f from 3-phenoxybenzaldehyde (4a).   Table 4, entries 1, 2). They are ligands among others for Ti(IV), Li(I), Al(III), It has been reported that salen-Ti(IV) 23b is only operative in anhydrous medium. Otherwise it is transformed to 23bTiO that possesses a different catalytic profile. 103 In the case of Schiff bases, it has been noticed that the presence of a hindered group such as a t-butyl group at the 6-position of the phenols dramatically favor the enantioselectivity ( Table 2, compare entry 4 to entry 3). Many other ligands of the series possess this structural feature (Table 2, entries 5,8a,8b,8c,8d,8e,8f; Table 3, entries 1, 2, 3, 5; Table 4, entries 1, 2), It has however been found 81 that the ligand 23a lacking the t-butyl group on the aromatic ring leads to a higher enantioselectivity as compared to the analogue possessing the t-butyl group in this position.
In some cases, it has been noticed that the ligand has been easily separated from the other compounds after the reaction, recovered in up to 95% yield and successfully recycled. 100 2.4.3.1.1.3. The catalyst. The catalyst produced by reaction between the "metal salt" and the ligand has been in some case prepared then introduced in the reaction mixture whereas in other case it has been prepared in situ. Although at the early time, stoichiometric amount of catalyst has been used ( Table 2, entry a), catalysts have been later designed that leads to the cyanohydrins in high enantiomeric excess when used in only 10-20% (  Table 3, entry, 1, 2, 4, 6, 7, 8, 9, 10, 11; Table 4, entry 3) and even 1-5% (Table 2, entry 5, 8a, 8b, 8c, 8d, 8e, 8f, 17b, 20, 21; Table 3, 5, 12, 13; Table  4, entry 1, 2, 4). Tables 2 and 3 have been carried out with trimethylsilyl cyanide. Other reagents that have been only scarcely used are acetone cyanohydrin (13a) and potassium cyanide in the presence of a trapping agent such as acetic anhydride or ethoxycarbonyl cyanide or cyanodiethylphosphonate (Scheme 28, Table 4).

The cyanide donor. Most of the reactions presented in
Reactions have been usually performed in chlorinated solvents such as dichloromethane or chloroform. In some cases, dramatic differences have been observed between those closely related solvents (Table 4, compare entries 4a and 4b). 105 However, toluene, ether, THF, ethanol, acetonitrile, have been from time to time used (Tables 2, 3, 4). Some striking differences have been in some cases noticed ( Table 2, compare entry 17c to 17d,e; entry 19a to 19b; Table 4, entry 4a to entries 4b-d). See also ref. 103.
The reactions have often been carried out in the presence of additives, that either increase the yield or the enantioselectivity but the mechanistic implications of which have rarely been elucidated. We can cite among others (i) tetrabutyl ammonium fluoride ( In the last case, for example, several additives have been tested with little success in the reaction of potassium cyanide, acetic anhydride and 1% catalyst 23bTiO with benzaldehyde: 103 For example, the addition of acids generally and acetic acid in particular decreases the rate of the reaction, and addition of hydrogen cyanide leads to dramatic loss of enantioselectivity. 103 Water, t-butanol, mineral acids, N-bases, thiols, thiourea, CS2, Ph4BNa, mineral carbonates, surfactants, phosphines, and phosphine oxides have improved greatly the reaction rate, but the most successful additives proved to be either 1H-imidazole (10 mol-% rel. to benzaldehyde) or the mixture of H2O/t-BuOH (10 and 100 mol-% rel. to benzaldehyde, resp.) added to the reaction mixture (see Table  4, entry 2 for application of those conditions to 3-phenoxybenzaldehyde (4a). 103 The role of water ( Table 2, compare entries 8a to 8b) 103 and triphenylphosphine 100 (Table 2, entries 8f,15a,17f,22; Table 3, entry 10), has been however identified in some cases. As already discussed above, it has been found that salen-Ti(IV) 23bTiO rapidly reacts with water, even with trace of moisture, to produce, in situ, the dimer 23bTiO that possesses a different catalytic profile and changes the course of the reaction. 103 Since (23bTiO) is a better catalyst than 23b, the later has been voluntarily reacted with an equivalent amount of water prior introduction of the reactants in the medium for the next step (Table 2,compare entries 8; Table 4, entry 2). 82,83,103 It was originally noticed by  and Najera 93,94 that increased enantioselectivity is observed in reactions carried out in the presence of Bu3P=O, 113,114 MeP(=O)Ph2 113,114 or Ph3P=O. 93,94,100 Those results have triggered the incorporation of the phosphine oxide moiety into the ligands 113,114 (Table 2, entry 23/F) 115 and has been at the root of the concept of "two-center catalysis".
Corey also found that addition of triphenylphosphine (20 mol%) to the reaction containing the boron catalyst 32 (10 mol%), improved dramatically the enantioselectivity of the reaction. 100 Basic NMR and infrared studies suggested that triphenylphosphine oxide reacts with trimethylsilyl cyanide to produce the new and more reactive species, whose structure are shown on Scheme 29, that exhibits an isocyanato structure and is suspected to be responsible of the observed increased selectivity. 100 Scheme 29. Cyanide-to-isocyanide transformation. 100

Organocatalysis in asymmetric synthesis of 3-phenoxymandelonitrile.
Catalysis implying metals presented above although it allows the synthesis of cyanohydrins in good yields and with high stereocontrol, suffers from (i) the loss of the metal involved in the catalyst, (ii) low temperature (<-30°C) often required, and long reaction times (>24h) that are not compatible with industrial requirements.
On the contrary, organocatalysis proved to be efficient for the enantioselective synthesis of mandelonitriles especially that of 3-phenoxymandelonitrile (3a). This is in particular the case when the cyclic dipeptides: cyclo[(R)-phenylalanyl-(R)-histidyl]diketopiperazide (R,R)- (33) or its enantiomer cyclo[(S)-phenylalanyl-(S)histidyl] diketo piperazide (S,S)-(33) 4,116-122 that to a certain extend mimic the nitrile lyase enzymes, are used as catalyst. 4 It was found that (S)-3a is produced in about 4h in high yield (94%) and high enantiomeric excess (ee > 92%, Table 5, entry 2d) on reaction of 4a with two equivalents of hydrogen cyanide (from NaCN and H2SO4) 121 and only 0.2% equivalents of (R,R)-33 as a catalyst in a suitable solvent at 5°C (Scheme 30, Table 5, entry 2d). 120 Lower yields are obtained if the aldehyde 4a is subjected to a transcyanation using instead acetone cyanohydrin (13a) in the presence of the same catalyst that requires to be carried out at higher temperature (Table 5, compare entry 1a to 2a). 116 Poor results are observed when DMSO is used as the solvent 119,120 whereas very good results have been observed with 33 in toluene, 120 that exhibits thixotropic behavior. Increased enantioselectivity is observed upon increasing the stirring rate that concomitantly decreases the viscosity. 119,121 Otherwise, good results are obtained when the reactions are performed in a gel or when the catalyst is present in the medium as an amorphous solid (by precipitation from methanol, spray drying, supercritical CO2 drying, lyophilization). 121 The presence of an alcohol at early stage of the process besides the aprotic inert solvent is especially useful in reducing the induction period, 122 increasing the reaction rate, and the enantiomeric excess of the cyanohydrin. 122 Higher enantioselectivity is observed when the solution is extracted with acid (aq. H2SO4) prior to the addition of the catalyst since this treatment allows the removal of trace of amines, if any, that could compete with the catalyst to produce racemic 3a. 117 It was also found 120 that the enantioselectivity of the transformation of 3-phenoxybenzaldehyde (4a) to 3phenoxymandelonitrile (3a) increases by increasing the reaction time (Scheme 30, Table 5, entries 2g,2h; Compare entry 2d to entries 2a,2b,2c). This implies that the product 3a formed interacts with the original catalyst 33 to form a more enantioselective one! This behavior named "enantioselective autoinduction" has been observed and defined previously by Alberts and Wynberg. 123 It has been thus reported that addition, at the early stage of the reaction, of small amount of: (i) (S)-3-phenoxymandelonitrile (S)-(3a) (0.09 eq.), the enantiomer expected to be formed, 121 generates 121 (S)-3-phenoxymandelonitrile (S)-(3a) more rapidly and with higher enantioselectivity compared to the original reaction (Table 5, compare entry 2h to entries,2e,2f,2g; Compare entry 2e to entry 2a especially entry H), (ii) (R)-3-phenoxymandelonitrile (R)-(3a) (0.09 eq.), the mismatched enantiomer, generates 121 (S)-3phenoxymandelonitrile (S)-(3a) more rapidly but with lower enantioselectivity compared to the original reaction (Table 5, compare entries 2l, 2h and 2d), (iii) (S)-mandelonitrile (S)-(12a) (0.04 eq.) leads 116 to (S)-3-phenoxymandelonitrile (S)-(3a) with a slightly higher enantioselectivity compared to the original reaction (Table 5, entry 3b compare to entry 3a). This enantioselectivity proved to be lower than the one generated by addition of the same seed amount (0.04 eq.) of (S)-3-phenoxymandelonitrile (S)-(3a) ( Table 5, entry 3c compare to 3a, b). 116 Scheme 30. Enantioselective synthesis of a cyanohydrin from an aromatic aldehyde using enantiopure manmade catalysts.

Asymmetric hydroxycyanation of 3-phenoxybenzaldehyde using hydroxynitrile lyase.
Enantiopure cyanohydrins 4,23,24,124 are known in nature where they play the role of chiral building block or alternatively as a stock of ammunition by producing on request the deadly hydrogen cyanide for leaving system that possess a respiratory system such as mammals. This field has been investigated in plants more than a century ago by Rosenthaler. 125 Hydroxynitrile lyases are responsible of the transformation of carbonyl compounds to the corresponding cyanohydrins. 23,24,[126][127][128][129][130][131][132][133][134][139][140][141][142][143] Applying successfully this process to the enantioselective organic synthesis request that the enzymatic process is favored over the non-enzymatic reaction that could compete leading to racemic cyanohydrins.
It has been calculated 126 that to achieve an enantiomeric excess of 99% a ratio of 100:1 catalyzed / uncatalyzed process is required whereas to reach an ee of 99.9% a ratio of 1,000:1 is obligatory and therefore, the non-enzymatic must be significantly reduced. 131 This has been usually accomplished: 126 (a) by decreasing the pH-value in the aqueous phase and the reaction temperature. For example, at 5 °C and pH 5.5, the non-enzymatic reaction involving benzaldehyde is drastically inhibited over that of hydroxynitrile lyase from Prunus amygdalus (for which the rate determining step seems to be the conversion of the ternary complex into the free enzyme and mandelonitrile) although its rate is also lowered, 127 (b) by introducing a chemical shunt in the transformation of carbonyl compounds to their cyanohydrins that could divert the flux of reactants away from the nonenzymatic direction that favors the enzymic pathway. 132 This strategy has been originally successfully applied by Kyler 132 on phenylacetaldehyde using acetone cyanohydrin (13a) as a cyanide donor in water-immiscible solvent [1.3 eq. 13a, acetate buffer (pH 5.0), Roxynitrilase, ether, 23 °C; 83% yield, ee 88%; in absence of the oxynitrilase; 8% yield, ee 0%]. 132 It has been successfully applied to 3-phenoxybenzaldehyde (3a) using (R)-hydroxynitrile lyase from Prunus mume ( Table 6, compare entry 2 to entry 1). 133 Although crude enzymes are available and often used, genes encoding HNL's enable the heterologous production of HNL's in industrially relevant expression system such as Escherichia coli so sufficient quantities of proteins can be produced with constant quality and batch-to -batch reproducibility at low cost. 124 It has been also described by Effenberger 128 that immobilized hydroxynitrile lyase (e.g. on nitrocellulose) in an organic solvent offers a definite advantage over biocatalysis in aqueous media. 129 The use of organic solvents greatly enhances the solubility of the substrate and suppresses the non-enzymatic reaction. Recovery of product is easier and substrate and/or product inhibition is reduced. However, limitations arise from limited stability and lower activity of the enzymes.
Immobilization has been successfully applied to the synthesis of 3-phenoxymandelonitrile (3a) out of a solution of 3-phenoxybenzaldehyde (4a) and hydrogen cyanide is passed through a porous membrane comprising a polymeric resinous binder having finely divided filer particles dispersed through-out the binder to which the (S)-hydroxynitrile lyase from sorghum shoots enzyme has been chemically bound (Table 6). 142 It can be observed that di-n-butyl ether was the best solvent, the higher yield being achieved at 20 °C (Table 6, entry 6a) and the higher ee obtained when the reaction is performed at lower temperature (6 °C, Table 6, entry 6b). 142 Reactions have even been also successfully carried out in "dry" organic solvents. 141 Homogeneous pure organic systems wherein the substrates and products are dissolved afford high degrees of conversion even at low hydrogen cyanide-substrate ratios. 126,130 Hydrogen cyanide 131 and more efficiently acetone cyanohydrin (13a) 132 in an ether-aqueous-buffered biphasic solvent system, proved to be valuable although in the latter case the solubility properties of the substrate have a pronounced effect on the enantiomeric purity of the mandelonitrile and although exchanging ether by ethanol still generates mandelonitrile in high yield it takes place with poor enantioselectivity. 132 Several hydroxynitrile lyases belong to the(R)-selective series. 4 This is among others the case of those related to the following species: Prunus amygdalus (almond, nuts), 126 Prunus laurocerasus (cherry laurel, seeds), Prunus lyonii (California cherry, seeds), Prunus serotina (black cherry, seeds), Linum usitatissimum (linseed, seedling), 126 Phlebodeum aureum (fern, leaves) and Arabidopsis thaliana. They have been the first to be discovered. (See Table 6, entries 1-3 for results involving 3-phenoxybenzaldehyde (4a) Others belong to (S)-selective series. 4 This is among others the case of those related to the following species: Manihot esculenta (manioc, leaves; Table 6, entry 5), Hevea brasiliensis (rubber tree, leaves; Table 6, entry 4), Sorghum bicolor (millet, seedlings), Sorghum vulgare (millet, seedlings), Ximenia americana (sandalweed, leaves). The hydroxynitrile lyase from Manihot esculenta (MeHNL), transforms a broad spectrum of aldehydes and ketones and some variants such as MeHNL-W128A expands the substrate range to more bulky substrates (Table 6, entry 5a, compare with entry 5b). 131 The synthesis of the (S)-enantiomers is hampered by the limited availability of the required biocatalyst (S-oxynitrilase from Sorghum) and its narrow substrate specificity. (See Table 6, entries 4-6 for results involving 3-phenoxybenzaldehyde (4a). Table 6.

Scheme 32. Hydroxynitrile lyases: cyanohydrin syntheses under conditions shown in
Most of the hydroxynitrile lyases known earlier were related to that of bitter almond (Prunus amygdalus) 24 promoting the addition of KCN in buffered solution (pH: 4.5) to produce the (R)-mandelonitrile (12a) from benzaldehyde (11). This is also the case of the seeds of the Japanese apricot (Prunus mume, Table 6, entry 1) 133 or that of the white apricot of the Indian Himalaya (Prunus armeniaca, Table 6, entry 2). 134 They allow, among others, the enantioselective synthesis of (R)-3-phenoxymandelonitrile (R)-(3a) from 3-phenoxybenzaldehyde (4a). Its (S)-enantiomer required for the synthesis of deltamethrin for example, needs therefore an inversion of configuration that could be achieved using a subsequent Mitsonobu reaction, a substitution reaction of the hydroxyl group using ethyl azodicarboxylate, triphenylphosphine and a carboxylate nucleophile (Scheme 32). 13 This reaction leads usually to an ester whose configuration is inverted as compared to that of the starting alcohols. 136,137 In the case of the cyanohydrin derived from benzaldehyde relatively poor results have been obtained when acetic acid was used (ee: 92%). 135 Better results have been observed using instead benzoic acids or arylacetic acids. 135 4-nitrophenyl acetic acid has been finally selected because (i) it provides the corresponding ester in good yield and high enantioselection (ee: 99%) (ii) the resulting ester can be easily purified by recrystallization and (ii) transformed in high yield and excellent stereocontrol to (S)-mandelonitrile in acidic media (MeSO3H/MeOH, 20 °C, 48 h, Scheme 33). 135 Use of at least one equivalent of methane or paratoluenesulfonic acid and methanol as the solvent was crucial since dilute hydrochloric or sulfuric acid caused no reaction at all and stronger aqueous acidic conditions, such as concentrated hydrochloric acid or 20% sulfuric acid, resulted in partial hydrolysis of the cyano group. 135 The whole transformation of benzaldehyde sequentially to (R)-then (S)-mandelonitrile is presented in Scheme 33. 133,135 It should be noticed that many other esters require a basic media for saponification that is incompatible with the stability of cyanohydrins. Scheme 35. Retrosynthetic approach to 3-phenoxybenzaldehyde (4a) from related 3-substitued benzaldehydes.
The transformation of 3-phenoxytoluene (4b) into 3-phenoxybenzaldehyde (4a) can be achieved (i) directly by substitution of two hydrogens by oxygen or (ii) in two steps that involve as the first step a halogenation reaction. Both routes are far from regioselective. Oxidation of 3-phenoxytoluene (4b) to 3phenoxybenzaldehyde (4a) is not an easy task due to competing formation of the 3-phenoxybenzyl alcohol (4c) and 3-phenoxybenzoic acid (4eH). Lateral halogenation is not an asy task since not only mono and polyhalogenation coexist but also ring halogenation compete.
Anyhow the synthesis of 4b, 4c, or 4e the potential precursors of 4a (Scheme 36) uses the same approaches as the ones proposed for the direct synthesis of 4a (Scheme 35) that involves the formation of an oxygen aryl bond but involves one different partner in each case 5b, 5c, (5d) instead of 5a; 6b, 6c, 6d instead of (6a).
We have also briefly included the case of 3-phenoxybenzonitrile (4f) whose transformation to 4a has been partially achieved and that of 3-phenoxybromo-benzene (4g) that involves a strategy different to that of used to synthesize 4b, 4c, or 4e and has been used for the synthesis of the 14 C labelled 4a required for biochemical tracing (Scheme 37).
We have gathered before discussing the related specific examples, some recent knowledge about (i) the Ulmann reaction that was not available at the time the industrial transformations have been carried out and (ii) Buchwald-Hartwig condensation that use the more expensive palladium catalyst and was unknown at that time. We have purposefully selected the more judicious examples that do not rely on the specific topic of interest and we have included the more recent advances about their mechanisms. 3.1.1. The copper-catalyzed Ullmann condensation reaction. The copper-catalyzed Ullmann condensation reaction involves the coupling of aryl halides and phenols and delivers diaryl ethers by copper catalysis. 145 It originally took place with excess of phenol in pyridine or polar solvents, at high temperature (120-250°C), in the presence of strong bases and stoichiometric amounts of copper powder, salts or oxides as the catalyst, and delivered, after long reaction time (12-48h), the diaryl ether in low to moderate yields that furthermore is difficult to isolate from the inorganic brownish sludge. [10][11][12][13][14]146,147 In the context of pyrethrin synthesis that require to be carried out at large industrial scale, it offers the advantage of using a cheap pre-catalyst but suffers from the drastic conditions required, the low yields obtained, the better reactivity of more expensive aryl iodides and the poorer reactivity of those reactants that possess electron-withdrawing group (such as the formyl group). Such behavior has triggered the synthesis of the 4b or 4c that are easier to produce but require an extra step to produce 4a.
As a general trend the reaction: (i) Proceeds faster with the aryl bromides or iodides than with the corresponding chlorides (the cheaper reagent) or fluorides (F<Cl<Br= I) but still requires high temperatures. 163 It has been however found that the arylation of phenols, including 3-hydroxybenzaldehyde (5a) can be conveniently carried out at room temperature without the need of copper catalyst if performed, in THF, with diphenyl iodonium triflates or tetrafuoroborates 6dI+ in the presence of stoichiometric amount of potassium t-butoxide (see below, Scheme 84, Scheme 85), 149,151 (ii) Does not generally work with aryl halides possessing strong electron-donating groups and phenols with electron-withdrawing. In such cases the reduction of aryl halides to arenes (Ar-Br to Ar-H), usually the major side reaction in Ullmann ether synthesis, takes place to a large extent. 164 It has been also found that bulky substituents at ortho-position of the reactants lower the reaction rate, in particular that of phenols with ortho methoxy or acetoxy groups, whereas the presence of an additional halogen on the aryl halide increases the reaction rate. 163 More recently however, some of those restrictions seem to have been overcome using an original process performed in toluene as the solvent at 110 °C for 12-26h with aryl bromides and iodides, that uses an excess of phenol (1.4-2 eq.) as a reactant, cesium carbonate as the base, copper triflate (CuOTf, 0.25-2.5% eq.) as precatalyst, small amount of ethyl acetate (5% eq.) and eventually naphthoic acid (1 eq.) as an additive. 165 (iii) Tolerates a variety of functional groups 10-14,146,147 either on (a) the aryl halide, which include alkyl, alkoxy, hydroxymethyl, amino, halogeno including fluoro, nitro, carbonyl, amido, sulfonamido and cyano group or on (b) the phenol: [10][11][12][13][14]146,147 such as halogeno including fluoro, alkoxy, hydroxymethyl, carbonyl, carboxy, nitro, trifluoromethyl groups.
(iv) Has been carried out with a large variety of bases such as Et3N, DIPEA, DBU, 1,2,2,6,6-pentamethyl piperidine, dicyclohexylamine, as well as K2CO3, K3PO4, Li2CO3, Na2CO3, BaCO3), but cesium carbonate (2 eq. Cs2CO3) has often been preferred. 165 (v) Use copper 145 or copper compounds as the catalyst/pre-catalyst in the presence of additives or ligands in stoichiometric or catalytic amounts (Scheme 38, entry a, Table 7). [10][11][12][13][14]146,147 It includes (Table 7) copper chlorides (CuCl, CuCl2), copper iodide (CuI), 165 copper bromide (CuBr, 165 CuBr2), copper triflate (CuOTf), 165 copper acetate (Cu(OAc2)) 165 and copper sulfate (CuSO4) as well as copper oxide (CuO, Cu2O) that includes reusable Cu2Onanocubes 195 and copper fluoroapatite 166 or Cu/CNFs nanofiber composite. 167 Although the choice of the copper salt did not appear to be critical, 165 the use of the more soluble copper triflate (CuOTf) or its complex with benzene has been suggested. 165 (vi) Is improved: • Under sonication 168 or photo-assistance 169 that allows a copper assisted coupling of phenols and aryl iodides at room temperature; • With Microwave assistance that offers the advantage 170 of about 15% increased yields over the same unirradiated process. Microwave irradiation has also been used in the synthesis of diaryl ethers in the absence of copper catalyst, 171 but seems almost exclusively limited to halogeno-benzenes substituted in 2and 4-position by an electron withdrawing group that usually involves a two steps mechanism (aromatic substitution by addition/elimination).
The structure of some of the ligands as well as the related copper compound and the solvent used in the modified Ullmann reactions are depicted in Table 7. In fact, the copper compounds act as pre-catalyst since except in rare cases where the effective catalyst is produced in situ on interaction with a ligand also named additive (Scheme 38, Table 7) 147 and in some cases with the solvent that plays a similar role. The proper use of ligand/additive and the copper pre-catalyst increases the rate of the reaction and allow to perform the reaction at lower temperature probably by increasing in conjunction with the solvent the solubility of the copper catalyst and that of the copper phenolate responsible of the substitution leading to the diaryl ethers.
The crucial role of the ligand/ and the solvent on the course of the Ullmann condensation has been further investigated. 15,16 A ligand-directed selectivity in N-versus O-arylation reactions of ambident nucleophiles has been observed and rationalized. 15 It has been suggested 15 that it does not derive from initial "Cu(I) (nucleophile)" complex formation but from the subsequent steps involving aryl halide activation leading to O-arylation via an iodine atom transfer (IAT) process whereas N-arylation takes place via a single-electron transfer (SET) process that depends on the electron-donating abilities of the ligand and the nucleophile (Scheme 39).

Scheme 39. IAT versus SET mechanism potentially involved in the Ullmann condensation reaction. 15
Detailed study on the intimate mechanism of Ullmann condensation, has been carried out by Hartwig who investigated the nature of the reactive phenolate species. 16 For that purpose neutral 38Cu (Scheme 40, entry a), 16 and ionic 54 copper alkoxides (Scheme 40, entry c), 16 have been synthesized unambiguously characterized even by X-ray crystallography in the case of 38dCu (Scheme 40, entry b) and reacted with a few aryl halides. 16 The reaction of reagents 38dCu with different ligands, in which the embedded phenoxy group bears electron donating-or electron withdrawing groups in para-position, with 4-fluoro iodobenzene (57), provides interesting information about the relative reactivity of related phenolates (Scheme 42): 16 the reaction is faster when the reactive ligand is more electron rich, most likely because it helps make the metal more electron rich and thereby accelerates oxidative addition of the aryl halide. 16 Scheme 42. Comparative rates of reaction of a series of aryloxy copper reagents toward 4-fluoro iodobenzene. 16

The palladium catalyzed Buchwald-Hartwig condensation reaction
It has been demonstrated that the cross-coupling of electron-deficient, electron-neutral, and electron-rich aryl halides and sulfonates with a variety of phenols could be conducted in high yield under relatively mild conditions using palladium as a pre-catalyst (Scheme 38, entry b) and electron-rich, bulky phosphines ligands ((59)-(66), Table 8). 153,172,186,189,190 The coupling has been also successfully achieved using supported palladium catalyst (67Pd, Table 8) 148 crafted onto cellulose whose ligand (79,  The reaction is very dependent upon the nature of the catalyst, especially the ligand. The Buchwald-Hartwig reaction has been continuously elaborated since its discovery in 1999 172,186 to expand its scope and to get some insight about its intimate mechanism. 153,[189][190][191] As general trends the limitations are often orthogonal to that of the copper mediated Ullmann reaction discussed above since for example phenols missing a substituent in ortho position pose problems, and except one case (Scheme 43, entry l, compare to entries j,k), 148 aryl bromides and eventually aryl chlorides are far better substrates than aryl iodides (Scheme 43). 153,172,186,189,191 Limitations have been observed with little or no product from reactions of: (i) Phenols with aryl halides that possess either strongly electron-withdrawing or electron-donating ortho substituents. 2-bromo-acetophenone for example is reluctant to couple, (ii) electron-deficient phenols and phenols that did not possess an ortho substituent at least as large as a methyl group and especially the first member of the series, the phenol as already pointed out.
The coupling proceeds equally well with various source of Pd such as palladium diacetate [Pd(OAc)2], or palladium(0) bis(dibenzylideneacetone) [Pd(dba)2]. The latter has been nevertheless in some instances preferred, 172 and in the case of reaction carried out at room temperature, the liganded (cinnamyl)PdCl2 proved by far the best pre-catalyst, 190 accounting for its exceptional aptitude to generate the active Pd(0) species at low temperatures as compared to the other commonly used precursors. 190 It has been also once noticed 186 that the amount of ligand compared to the palladium pre-catalyst (1/1 to 5/1) does not affect the course of the reaction.

Scheme 43.
Selected examples of the Buchwald-Hartwig condensation reaction. 148,153,172,189,190,192 In fact, over the years, following the discovery of new ligands ( Table 8), many of the original limitations have been overcome.
The first reports involve (i) phosphines 59-62 whose phosphorus atom is surrounded by at least two bulky groups such as t-butyl, adamantyl groups, and a ferrocenyl or biphenylyl group as the third substituent, (ii) quite large amount of non-recoverable palladium species (2-5% molar equivalent), (iii) various bases such as NaOH as K3PO4 (Scheme 43 entries a,b,e,h). 172,186 , (iv) toluene as solvent and (v) heating at 80-110°C for several hours (14-48 h).
The next improvement involves increasing substitution on each of the two aryl groups of the biphenylyl moiety of the ligand (63, 64). 153 It dramatically increases the type of phenols and aryl halides that can couple (Scheme 43 entry c). 153 The latter generation of ligands (65,66, Table 8, Scheme 43, entries d,f,g,i) includes 65 that instead bears an 1-arylindole moiety and 66 that still belongs to the biphenyl series but possesses larger substituants on the biphenyl group. It extends the scope of the reaction to o-unsubstituted phenols, lowers the amount of palladium and ligand used and offers exceptional TON. These ligands allow in some specific case to carry out the reaction at room temperature (66, Table 8, Scheme 43, entry f). 189,190 The final improvement uses palladium grafted on cellulose and liganded by two nitrogen atoms. It still allows low loading of the catalyst, but also allows its reuse, couples a large spectrum of aryl halides including aryl iodides that usually pose problems, involves benign bases such as potassium carbonate, short reaction times and offers remarkable TON and TOF performances (Table 8, Scheme 43, entries j,k,l). 148 The ligand has been prepared stepwise from filter paper (i) acidic hydrolysis (2.5M HBr) of filter paper under ultrasonication (100 °C, 3h) (ii) isolation of the residue after washing with deionized and centrifugation (12.000 RPM) (iii) reaction with tosyl chloride in pyridine, leading to tosylated cellulose nanocrystals "CNC-Tos" (iv) grafting "CNC-Tos" with 2-(1H-benzo[d]imidazol-2-yl) aniline "BIA" (DMF, 100 °C, 24 h) leading to "CNC-BIA" (67) (v) reaction of 67 with palladium dichloride in the same solvent (60°C, 24h). The resulting "CNC-BIA-Pd" (67Pd) ( Table 8) accounts 0.047g Pd/ 1g 67Pd (4.7%) and is used in 0.4 g amount for 1 mmol of aryl halide. 148 K2CO3 proved a better base than Na2CO3, NaOH or KOH; DMSO proved to be better than MeCN, DMF, dioxane or ethanol. 148 The coupling is efficient even with phenols with withdrawing group and aryl halides with electron donor groups. The yields were in the order of para> ortho> meta, the catalytic performance of the substrates is I > Br > Cl and the CNC-BIA-Pd 67Pd catalyst could be reused even eight times without losing its effectiveness with very low Pd leaching (5%). 148 Successful coupling has been also achieved using recyclable nano-particles implying graphene oxide grafted with a Pd compound. 192 Little is known about the molecular basis of the catalysis and although X-ray crystallographic data of some postulated intermediates are available, the rational about the request of bulky ligands for successful catalysis is still missing. It has been for example proposed 153 that the better performance of ligand 64 compared to that of 63 (Table 8) could be due to restricted rotation around the C-aryl, P bond in catalyst 67Pd (compared to its destetra-methyl analogue 64Pd) that fixes, for steric reasons, the location of the palladium-containing entity in 64' as shown in Scheme 44.

Scheme 44.
Modelling the reactivity of related postulated catalyst in the Buchwald-Hartwig condensation. 153 The mechanism of the reaction has been at several occasions discussed 172,186 and is admitted to consist of three distinct steps depicted in Scheme 45: (1) oxidative addition of aryl halide I to the ligand-palladium complex LnPd(0) to give J; (2) formation of the Pd-aryloxide complex L from the Pd-halide adduct via transmetallation of metal phenolate K; and (3) reductive elimination leading to the diaryl ether product M with concomitant regeneration of the active L n Pd(0) species H.
While the oxidative addition and transmetallation may be expected to be relatively facile, the reductive elimination to form the C-O bond is disfavored due to the Pd,C (LUMO) and Pd,O (HOMO) energy gap. at lower concentrations (82% at 0.2 M aryl halide vs 23% at 1 M aryl halide), lead to the suggestion that monomeric complexes are almost certainly the real intermediates. 172 The authors have also proposed that "coordination of ligand P(t-Bu)3 (59) to arylpalladium phenoxides makes reductive elimination of ethers much faster than it is from complexes containing more conventional arylphosphine ligands ". 172 Thermolysis of aryl alkoxide complex 74 (Scheme 47, entry a) carried out by raising the temperature from -50 to 23 °C in THF, leads to the C-O reductive elimination producing the p-neopentoxybenzaldehyde (75a) in excellent yield. 191 Scheme 47. Modelling the reactivity of palladium species as an insight about the mechanism of the Buchwald-Hartwig condensation reaction. 191 Decomposition of 74a at 23 °C obeys first-order kinetics and the rate of C-O reductive elimination from (1b) is nearly an order of magnitude faster than decomposition of p-cyanophenyl derivatives 74 (E= CN). It has also been found that the rate of reductive elimination decreases in the order E= NO2 > CHO > COPh (74, Scheme 47). 191 It was found that C-O reductive elimination from aryl(neopentoxide) complexes occurs far more readily than that of aryl(aryloxide) complexes and is facilitated by the presence of substituents on the palladium-bound aryl group capable of delocalizing negative charge.
These observations are consistent with the buildup of negative charge in the palladium-bound aryl group in the transition state for C-O reductive elimination.
This charge accumulation can be accounted for by a mechanism initiated by inner-sphere nucleophilic attack of the alkoxide ligand at the ipso-carbon atom of the palladium-bound aryl group to form a zwitterionic Meisenheimer intermediate or transition state (Scheme 47, entry b). However, the proposed mechanism does not appear to be the only mechanism available for C-O reductive elimination.
In an effort to further probe the intimate mechanism of palladium-mediated C-O reductive elimination, the rate and efficiency of the thermal decomposition of palladium (aryl) neopentoxide complexes has been investigated as a function of the electronic nature of the palladium-bound aryl group. Kinetics are consistent with rapid and reversible generation of intermediate O followed by rate-limiting Pd,C heterolysis to form the liganded palladium 74 and the aryl ethers 75 via a charge-separated transition state such as P (Scheme 51, entry c). Alternatively, kinetic results are consistent with rate-limiting alkoxide migration via a transition state such as N followed by rapid Pd,C bond cleavage (Scheme 47, entry b). Using these informations, it could be suggested that: 186 (1) For electron-deficient aryl halides, a mechanism involving transfer of the phenolate from the palladium to the ipso-carbon of the aryl halide to form a zwitterionic intermediate such as O then P is suggested which then converts to the diaryl ether and a palladium(0) complex, (2) For electronically neutral and electron-rich aryl halides, a different mechanism is proposed for reductive elimination to form the C-O bond that most likely involves a three-centered transition state such as N. In these cases, the bulkier ligands are necessary to destabilize the ground state of the LnPd(OAr)Ar' complex, forcing the palladium-bound aryl and aryloxy groups to be closer, favoring a distorted complex and a three-centered transition state. 186.

Syntheses of 3-phenoxybenzaldehyde from 3-halogenobenzaldehydes and 3-hydroxybenzaldehyde
This approach takes advantage of the fact that one of the two partners, the phenol (5d), is a basic compound and the fact that the coupling proceeds efficiently from the 3-bromobenzaldehyde (6aBr) catalyzed by copper chloride under the classical Ullmann reaction (Scheme 48, entries a,b). 173

Syntheses of 3-phenoxybenzaldehyde from 3-halogenobenzaldehydes.
The approaches to 3phenoxybenzaldehyde (4a) from 3-halogenobenzaldehydes 6a and 3-hydroxybenzaldehyde (5a) are apparently the most straightforward ones. They have been achieved in the first case by coupling 3-halogenobenzaldehydes 6a with the parent phenol in the presence of a base or with pre-generated sodium or potassium phenoxide in the presence of a transition metal (Section 3.2.1) or 3-hydroxybenzaldehyde (5a) and a diphenyliodonium salt or triphenylbismuth diacetate (Section 3.2.2.).

Scheme 48.
Syntheses of 3-phenoxybenzaldehyde from 3-halogenobenzaldehydes. 148,173,187 These results do not seem to align with the admitted knowledge that (i) aromatic halides react poorly when they bear electron withdrawing groups, the fact that (ii) contrary to what is believed, the reaction carried out in pyridine does not perform better than that performed in xylene (Scheme 48; compare entries a,b) and that (iii) the fluoro derivative 6aF, which is believed to be the least reactive perform so well and even does not require the presence of copper (Scheme 48, entry c). 187 This suggests a process that involves an addition-elimination reaction through a Meisenheimer intermediate 188 that does not usually proceed with fluoro derivatives that bears an electron attractive group in meta-position but rather in ortho/para. The reaction catalyzed by palladium supported on cellulose nanocrystals (CNC-Bia-Pd (67Pd), Table 8) according to a modified Buchwald-Hartwig process, provides an exceptionally high yield of 3-phenoxybenzaldehyde under mild conditions (DMSO, 80 °C, <1h), profiling reasonably good catalytic performances (TON 440, TOF 543 h -1 , Scheme 52, entry d) and efficient reuse of the catalyst. 148 This reaction however proceeds on the aryl iodide that is the most expensive and the heaviest of the series.

Syntheses of 3-phenoxybenzaldehyde from 3-hydroxybenzaldehyde.
Arylation of 3hydroxybenzaldehyde does not seems to have been carried out using the Ullmann or Buchwald-Hartwig reactions. It has been however successfully achieved at room temperature and in high yield using the particularly reactive diphenyliodonium tetrafluoroborate (Scheme 49). 151 Furthermore it does not require the presence of a metal catalyst (Scheme 49, entry a). 151 Scheme 49. Syntheses of 3-phenoxybenzaldehyde from 3-hydroxybenzaldehyde. 151,158 The other successful synthesis involves arylation of 3-hydroxybenzaldehyde (5a) using triphenylbismuth diacetate. 193 This reaction also takes place at room temperature but requires longer time (24 h) and uses copper powder as catalyst (Scheme 49, entry b). 158

Synthesis of 3-phenoxytoluene from 3-halogenotoluene or 3-hydroxytoluene. 3.3.1.1. Synthesis of 3phenoxytoluene from 3-halogenotoluenes and phenol.
The results concerning the synthesis of 3phenoxytoluene (4b) from 3-halogeno-toluenes 6b and phenol (5d) are gathered in Scheme 50. They all involve the Ullmann reaction using copper salts as the pre-catalysts, cesium carbonate as the base or already prepared potassium phenate, are usually carried out in aprotic polar solvents such as N,N-dimethyl formamide (DMF), N,N-dimethyl acetamide (DMAC), N-methyl-2-pyrolidone (NMP) or acetonitrile and provides 3-phenoxytoluene (4b) in medium to high yields. 166,168,169,181,194,195 Scheme 50. Synthesis of 3-phenoxytoluene from 3-halogenotoluenes. [166][167][168][169][170][171][172][173][174][175][176][177][178][179][180][181]195 They are however striking differences in conditions and yields depending on the catalytic system used. The reaction has usually been carried out with 3-bromotoluene (6bBr) (Scheme 50, entries a-d) and its iodo analogue 6bI (Scheme 50, entries e-g) usually at temperature of 120-150 °C for 3 to 24 h. In one case, the transformation has been performed on 6bBr without solvent under sonication at 120 °C with extremely low copper iodide loading (0.05%, Scheme 50, entry d) 168 or even at room temperature under light irradiation (254 nm, Scheme 50, entry e). 169 The reaction performed using a more elaborated copper chloride catalyst (Cu/CNFs) deposited on carbon nanofibers, resulting from a carbonization process, delivers 4b in similar modest yield (Scheme 54, entry b, compare to entry a). 167 It nevertheless offers the advantage of easy recovery of the catalyst that is usually cumbersome to achieve, and its easy recycling. 167 It was noticed 167 that regioisomeric bromotoluenes behave similarly, generating the isomeric phenoxytoluenes in similar modest yield although under similar conditions diphenyl ether is produced in almost quantitative yield. 167 Scheme 51. Synthesis of copper deposited on fluorapatite (CuFAp) and complexed copper catalysts. 166,177 Higher yields have been observed using copper deposited on fluorapatite (CuFAp) solid support that requires a high loading and no ligand (100 mg/mmol, Scheme 50, entry c), 166 or Cu2O-nanocubes (Scheme 50, entry g) 195 that similarly allow the easy isolation of the product from the inorganic phase. Note that the former reaction that use (CuFAp) has been successfully extended to the whole series of halogeno-benzenes, except the  166 Excellent yield of 4a have been obtained from the reactions carried out neat under sonication (Scheme 54, entry d), 168 or induced by light (Scheme 50, entry e). 169 Scheme 52. Possible pathway for photoinduced copper catalyzed cross-coupling between 3-halogeno-toluenes and phenol. 169 The latter reaction merits further comments. It has been established 169 that: (i) Use of longer-wavelength light result in less efficient C-O bond formation (300 nm: 64%), (ii) Combination of DBU and t-BuOK provides a better yield (80%), than each one alone (DBU: 70% ; t-BuOK: 74%), (iii) Excess t-BuOK is detrimental, (iv) the alkali-metal cation associated with the t-butoxide base has a significant impact on the coupling efficiency (t-BuOLi: 18%), (v) the use of a lower quantity of CuI (5%: 42% yield instead of 80%) or replacement of CuI with CuCl (78% yield), CuCl 2 (65% yield) or Cu (22% yield) leads to diminished yields in the diaryl ether 4b.
A mechanism that involves the intermediate formation of the aryl radical Q has been postulated to account for this process (Scheme 52, entry a) 169 and interestingly 3-phenoxytoluene (4b) has been successfully synthesized on reacting on 3-iodo-toluene (6bI) with 76, a postulated intermediate in the process, even in the absence of copper iodide (Scheme 52, entry b). 169

Synthesis of 3-phenoxytoluene from 3-hydroxytoluene and halogenobenzenes.
3-phenoxytoluene (4b) has also been synthesized from m-cresol (5b) and halogenobenzenes 6d using the Ullman procedure as shown in Scheme 53,164,170,166,178,197 the reaction employing the stannyl derivative 6dSn instead is reported in Scheme 54. 155 The reaction has been carried on almost the whole family of halogenobenzenes with the exclusion of fluorobenzene and involves in most of the cases in situ formation of a cresolate on reaction with potassium hydroxide or cesium carbonate (Scheme 52, entries a-f). 178,164,170 In other cases, the phenolate is prepared separately and introduced in the reaction mixture (Scheme 57, entries h,i). 166,197 It has been observed, as expected, that bromobenzene (6dBr) in PEG 4000 as the solvent and ligand (Table  7) is far more reactive than the related chloride 6dCl (Scheme 53, compare entries a,d) and that in the latter case performing the reaction under pressure has a favorable impact on the yield of 4b (Scheme 57, compare entries a-c).

Scheme 53.
Copper-catalyzed synthesis of 3-phenoxytoluene from metal cresolates and aryl halides. 166,178,170,195,197 Successful coupling has been performed on iodo-benzene (6dI) in the presence of N,N-dimethylglycine (Table 7) as the ligand (Scheme 53, entry e), 164 under microwave irradiation in NMP, although it proceed at very high temperature (Scheme 53, entry f), 170 or using copper deposited on fluorapatite (CuFAp) in NMP (Scheme 51, entry b, Scheme 53, entry h). 166 Apparently, the process implying Cu2O-nanocubes in THF was very attractive since it proceeds 195 rapidly (3 h) with chlorobenzene (6dCl) (Scheme 53, entry g). It does not seem to be impacted by the halophilicity that follows the bond reactivity order of C-I > C-Br > C-Cl. It requires an extremely low loading (0.1 mol%), allows easy isolation of the product and provides the product in high yield and the catalyst can be recycled at least 3 times without losing of its activity and furthermore these Cu(I)] particles are easily produced at 240 °C by a onepot process from Cu(II)(acac) 2, in the presence of poly-(vinyl pyrrolidone) (PVP) as a surfactant and 1,5-pentanediol (PD) as both reductant and solvent. 195 Nevertheless this process requires 2 equivalents of Cs2CO3 as base (Scheme 57, entry g) that precludes its industrialization due to its high cost and to the fact that cesium is genotoxic. This topics is discussed below (Scheme 54).
According to a patent, the most convenient industrial route to 3-phenoxytoluene (4a) should use the cheapest chlorobenzene (6dCl), sodium cresolate (5bNa) and the smallest possible amount of copper chloride catalyst. Although the reaction takes place at 160 °C with an amount of copper chloride as low as 0.05 eq. with the more expensive potassium cresolate, it does not work with its sodium analogue. 197 In order to perform the reaction at the lowest cost, It has been found that the reaction proceeds under similar conditions (160 °C, 5 mol% CuCl), neat under anhydrous conditions when sodium cresolate is mixed with at least 40% of the related potassium salt or if chlorobenzene (6dCl) is reacted neat with a mixture of sodium cresolate / potassium cresolate / cresol (1/1/2) (Scheme 54). 197 It has also been reported that: (i) the catalytic activity of copper chloride is limited to about four hours and (ii) the red-brown solid resulting from the coupling dissolves in aqueous acidic media allowing extraction of the coupling product (4b) with chlorobenzene that can be reused as starting material for another run. 197 Scheme 54. Industrial synthesis of 3-phenoxytoluene from chlorobenzene. 197 The synthesis of 3-phenoxybenzaldehyde (4b) has also been achieved from m-cresol and triphenylstannyl chloride (6dSn) and 3-hydroxytoluene (5b) in the presence of copper acetate. The reaction occurs at room temperature and proceeds with 0.5 equivalents of reagent suggesting that the triphenyltin chloride is able to transfer more than one phenyl group in the process (Scheme 55, entry a). 155 Since tetraphenyltin does not provide 4b, it has been suggested that the presence of the Sn-Cl bond in triphenyl chlorostannane (6dSn) provides the site of insertion for copper to produce the postulated intermediate R that on reaction with cresol (5b) leads, in the presence of triethylamine, to S by phenyl migration, then to intermediates T and U (Scheme 58, entry b). 155 interestingly, U has still the capability to transfer one more phenyl group to (5b) to produce 4b. 155 Scheme 55. Copper-catalyzed synthesis of 3-phenoxytoluene from meta-cresol and arylstannanes and a proposed mechanism. 155 3.3.2. Syntheses of 3-phenoxybenzaldehyde from 3-phenoxytoluene. The strategies involved for the transformation of 3-phenoxytoluene (4b) to 3-phenoxybenzaldehyde (4a) are summarized in Scheme 56.
Oxidation of 3-phenoxytoluene (4b) to produce 3-phenoxybenzaldehyde (4a) (Scheme 56, entry b) is not an easy task since the removal of the first benzylic hydrogen leading formally to the benzyl alcohol (4c) is the most difficult task (Scheme 59, entry a) and over-oxidation competes (Scheme 56, entry c).
Oxidation by oxygen is without context the cheapest but the synthesis of 3-phenoxybenzaldehyde (4a) request fine tuning since the formation of the intermediate 3-phenoxybenzyl alcohol (4c) is difficult (Scheme 56, entry b then a) and overoxidation leading to 3-phenoxybenzoic acid (4e) particularly easy (Scheme 56, entry c). Scheme 56. Alternative approaches to 3-phenoxybenzaldehyde from 3-phenoxytoluene.
Oxidation using metal oxides offers the advantage of tuning their reactivity by selection of the metal. Oxidation has been also carried out with halogens (Scheme 56, entries c,d) and again the first exchange leading to 3-phenoxybenzyl halides 4h is usually the most difficult and polyhalogenation that can either occur laterally leading to 4i (Scheme 56, entry d) or on the ring. Furthermore, use of halogens is not only more expensive than use of oxygen but it also requires an extra step to deliver the required aldehyde 4a. The

Transformation of 3-phenoxytoluene to 3-phenoxybenzaldehyde.
Oxidation of benzylic hydrogen of 3-phenoxytoluene (4b) by dioxygen in the presence metal catalyst does not go to completion and is not usually chemoselective. 178,[198][199][200][201][202][203][204]205, It provides besides the required aldehyde (4a) a mixture of 3-phenoxybenzyl alcohol (4c) and in some case the carboxylic acid 4e. These have been then transformed to 3phenoxybenzaldehyde (4a) on chemoselective oxidation of the benzyl alcohol 4c (Chapter 3.3.) or by reduction of the carboxylic acid 4e that has been achieved among others through the corresponding benzoyl chloride 4j using usually the Rosenmund reaction. 206 The direct oxidation of 3-phenoxytoluene (4b) using dioxygen as oxidant and cobalt diacetate as catalyst [Co(OAc)2/O2] has been reported in the patent literature (see references quoted in refs 178, 207 ) to provide 3phenoxybenzaldehyde (4a) (29% yield), the corresponding benzyl alcohol (4c) or its acetate 4cAc (21% yield) beside a large amount of by-products (21-50%). 207 Scheme 57. Oxidation of 3-phenoxytoluene by air catalyzed by cobalt acetate. 178 More recent results using the same catalyst are shown in Scheme 57. 178 The reaction is best achieved by dioxygen at quite high temperature (110 °C) and under high pressure (12 bar), in the presence of cobalt acetate (Co(OAc)4) as a catalyst and an additive such as paraldehyde, sodium bromide alone or mixed with copper acetate. These reactions deliver in all the cases 3-phenoxybenzaldehyde (4a), the corresponding carboxylic acid 4e and the corresponding benzyl acetate 4cAc in low overall yield (18-33%). 178 Oxidation of 3-phenoxytoluene has been also performed by stoichiometric amounts of potassium permanganate (KMnO4) and added potassium or copper halides. It delivers a small number of compounds resulting from side-chain oxidation, besides larger amounts of compounds 78x resulting from ring halogenation (Scheme 58, entries a-c). 207 These poor results have been ascribed to the detrimental role of the phenoxy group attached in meta position to the methyl group. 207 Scheme 58. Oxidation of 3-phenoxytoluene with manganese compounds. 205,207 Oxidation has also been achieved by hydrogen peroxide in the presence of N,N′,N′′-trimethyl-1,4,7triazacyclononane (Mn-tmtacn) sulfate and oxalic acid in the presence of a buffer. It leads to a mixture of the Scheme 59. Oxidation of 3-phenoxytoluene by "solid" ceric methanesulfonate. 204 Oxidation of 3-phenoxytoluene takes place with solid ceric methanesulfonate [Ce(MeSO3)2(OH)2·H2O] in methanesulfonic acid (110°C, 1.10 h) and produces the aldehyde 4a in around 40% (Scheme 59). 204 Performing the same reaction with fully dissolved ceric reagent does not produce the aldehyde 4a but leads to substantial loss of the 3-phenoxytoluene (4b), the latter being recovered in less than 27 % after the same reaction time. 204 It has been found 208 that ceric trifluoroacetate in aqueous trifluoroacetic acid is particularly effective for the oxidation of activated toluenes including (4b) to the corresponding aldehydes such as (4a), but unfortunately aromatic ring oxidation compete with benzylic oxidation (Scheme 60). 208 These competitive processes have been rationalized by the involvement of the ring centered radical cation X as shown in Scheme 63. 208 Finally, since the ceric ion is consumed in stoichiometric amounts in these process, its electrochemical regeneration at high current efficiencies (95%) allows to achieve the oxidation using catalytic amounts of cerium. 208 Scheme 60. Oxidation of 3-phenoxytoluene by cerium(IV) and rationalization. 208 As already pointed out, only few oxidants have been found selectively to transform 3-phenoxytoluene (4b) into 3-phenoxybenzaldehyde (4a). At best, this oxidation delivers besides 4a the benzyl alcohol 4c or/and the carboxylic acid 4e. Separation and recycling 4c by selective oxidation and 4e by selective reduction allows the production of the desired aldehyde 4a in fair yield.
Otherwise, 178 the mixture can either be reduced by for example by lithium aluminum hydride 209 to 3phenoxybenzyl alcohol (4c) or oxidized to the carboxylic acid 4e by the Jones reagent for example, 210 then selectively oxidized or reduced to the corresponding 3-phenoxybenzaldehyde (4a) (Scheme 61).

Scheme 61.
Synthesis of 4-phenoxybenzaldehydes. 178,211 Thus, the carboxylic acid 4e has been successfully transformed to 3-phenoxybenzaldehyde 4a using the Rosenmund reduction 206 that involves the transformation to the corresponding acid chloride 4j on reaction of thionyl chloride in DMF at reflux and its reduction with hydrogen in the presence of palladium on charcoal as catalyst (Scheme 61, entry a). 178 The transformation of 3-phenoxybenzyl alcohol (4c) to 3-phenoxybenzaldehyde (4a) discussed in Section 3.4.2.

Transformations of 3-phenoxytoluene to 3-phenoxybenzaldehyde involving its side-chain halogenation.
Transformation of 3-phenoxytoluene (4b) to 3-phenoxybenzaldehyde (4a) has been achieved by side chain halogenation of the starting material and subsequent substitution. Halogenation usually involves the intermediate formation of a radical and provides benzyl halides 4h and subsequently benzal halides 4i. Although both are good precursors of 3-phenoxybenzaldehyde (4a), there is an economical advantage to produce the former since a single halogen is consumed and although both 3-phenoxybenzyl bromide (4hBr) and the 3phenoxybenzyl chloride (4hCl) are able to generate 4a the latter has to be preferred due to lower cost.  202 substantially improves the conversion and the amount of chlorination. AIBN proved to be more selective than BP providing the benzyl derivative 4hCl in higher yield (Scheme 65, entry c, compare to entries a,e). 202 It has been found that the amount of benzal chloride is highly dependent on the reaction time and increases by increasing it since the amount of gas is dependent on the reaction time (Scheme 62, compare entries b-d). 202 Under sunlamp irradiation that also favor the radical process (Scheme 65, entries f-h) chorination is more efficient by increasing the amount of chlorine gas involved but this at the same time substantially favors the formation of benzal chloride (4iCl) produced (Scheme 62, entries f,g,h: 1/1.27/2.1 eq.; 4hCl: 15/48/74% yield; 4iCl/4hCl/4iCl: 3/6/23 ratio).  212 and mixture of compounds that include 3-phenoxybenzyl chloride( 4hCl) is formed when the reaction is performed with AIBN in different solvents (CCl4, 1,2-dichloroethane, chlorobenzene) 212 or in the presence of different radical initiator such as lauryl or benzoyl peroxide (low conversion). 212 As a general rule and in order to avoid the formation of ring chlorinated compounds, the chlorination reaction must be performed, whatever is the reagent (i) at low concentration (5-50%), 202 in a solvent and therefore not neat 212 (ii) in non-polar solvent, 202 and (iii) should be prevented from going to completion and should be stopped at 95-98% conversion. 202 Scheme 63. 212 Oxidation of 3-phenoxytoluene to 3-phenoxybenzylchloride using sulfuryl chloride. 212

Side-chain bromination of 3-phenoxytoluene.
Bromination of 3-phenoxytoluene (4b) has been carried out in a stream of excess of bromine (1.25 eq.) under ultraviolet irradiation close to the place the bromine is introduced in the reactant and vigorously circulated (Scheme 64). 202 After cooling for more than 12 h and flushing with nitrogen to remove the unreacted bromine, the resulting mixture (4b/4hBr/4iBr:2.1/61.5/36.4%) is not purified but instead directly engaged for the next step leading to 4a as will be discussed in the forthcoming paragraph. 202 Scheme 64. Oxidation of 3-phenoxytoluene to 3-phenoxybenzylbromide and transformation to 3-phenoxy benzaldehyde. 202
aldehyde without overoxidation to the corresponding acid 4e. It offers therefore a net advantage over the transformation that instead involve the intermediate formation of 3-phenoxytoluene (4b) discussed above since selective oxidation of benzylic primary alcohol is much easier and selective compared to that of the methyl group of 3-phenoxytoluene (4b) (See above).

Synthesis of 3-phenoxybenzyl alcohol from 3-halogenobenzyl alcohols and 3-hydroxytoluene.
In the two approaches that involves the 3-halogeno-benzyl alcohols 6c and phenol (5d) (Section 3.4.1.1.), or the 3hydroxybenzyl alcohol (5c) and halogenobenzenes 5d (Section 3.4.1.2.), remains the possibility that the alcohol moiety competes with the phenol in the coupling process forming competitively the unwanted phenyl benzyl ether instead of the diaryl ether and potentially a polymer from 6c. It has been interestingly found that such possibility does not occur under the conditions used (Scheme 69a, Scheme 70). 175,179,180,217 3.4.1.1. Synthesis of 3-phenoxybenzyl alcohol from 3-halogenobenzyl alcohols and phenol. The former coupling that is not so easy to achieve when using copper catalyst 176 that has been successfully carried out with a nickel catalyst. 217 This method offers the advantage to occur under mild conditions (80 °C), in the presence of a mild base (K2CO3) in aqueous medium along with a surfactant such as sodium dodecyl sulfate (SDS), catalyzed by alumina-supported nickel nanoparticles as a stable recyclable heterogeneous catalyst (Scheme 69, entry a).

Scheme 69.
Synthesis of 3-phenoxybenzyl alcohol from 3-bromo-benzyl alcohol and phenol. Aryl versus alkyl coupling. 217 A control experiment carried out using a 1/1/1 mixture of p-cresol, iodobenzene and benzyl alcohol showed explicitly that the diaryl ether is produced selectively and the benzyl alcohol fully recovered at the end of the process (Scheme 72, entry b). 217

Synthesis of 3-phenoxybenzyl alcohol from 3-hydroxybenzyl alcohol and halogenobenzenes.
The synthesis of 3-phenoxybenzyl alcohol (4c) from 3-hydroxybenzyl alcohol (6c) and chlorobenzene (5cCl) has been successfully achieved in high yield (> 80%) in 3-dimethyl-2-imidazolidinone (DMI) using copper chloride complexed by 8-hydroxyquinoline (37) ( Table 7) as catalyst. The reaction is carried out at high temperature (150-180 °C) in chlorobenzene, as one of the reagent and as the solvent, so the excess of chloro-benzene can be distilled of during the process to allow, by azeotropic distillation, the removal of the water produced on reaction of the phenol with potassium carbonate (Scheme 70, entry a). 175 Interestingly, the coupling chemoselectively occurs to produce the diaryl ether and not to the alkyl aryl ether. 175 Scheme 70. Synthesis of 3-phenoxybenzyl alcohol from 3-hydroxybenzyl alcohol and chloro-and bromobenzenes. 175,179,180 Even better results (91 % yield) have been obtained from 1.5 eq. of chlorobenzene (half the amount used in the previous process) with less than 2% of copper iodide and N-aryl,N'-alkyl oxalamines (PMPBO) 42c as ligand (Table 7) in DMSO at 120°C for 24 h (Scheme 70, entry b). 179,180 The reaction can be achieved at lower temperature when performed 179 on bromobenzene (90 °C, instead of 120 °C, Scheme 70, entry c, compare to entry b) but requires a slightly different N-aryl,N'-alkyl oxalamine ligand (BPPO 42b instead of PMPBO 42c, Table 7). 179 Apparently, the proper selection of the ligand and the solvent proved to be crucial for the success of the reaction.
Those conditions allow the efficient coupling of ortho-and meta-substituted aryl bromides using a 1 mol% loading of the catalyst-ligand mixture and proceed well with electron-rich phenols but requires a higher loading (2 mol%) with the less reactive electron-poor phenols. 179

Synthesis of 3-phenoxybenzaldehyde by chemoselective oxidation of 3-phenoxybenzyl alcohol.
Selective oxidation of alcohols to the corresponding aldehydes is a ubiquitous transformation in organic chemistry 196, that include several named reactions such as Swern oxidation, Moffat oxidation, Jones oxidation, Corey-Suggs oxidation, Dess-Martin oxidation (references cited). 211 It has been traditionally achieved by numerous stoichiometric oxidants, often involving transition metals, that lead to the release of quantities of toxic by-products into the environment. Molecular oxygen as a pure substance or diluted in air and hydrogen peroxide offer several advantages since they are environmentally friendly and exhibit a high efficiency per weight of oxidant. Molecular oxygen is inexpensive but difficult to handle whereas hydrogen peroxide is easy to handle as a water miscible liquid. 196,218 The direct oxidation of organic substrates by either O2 or H2O2 is rare as the energy barrier for electron transfer from the organic substrate to the oxidant is usually high. For molecular oxygen, which has a triplet ground state, this high energy barrier is nature's way of protecting organic compounds from destructive oxidation". 218 For a catalytic oxidation reaction, the substrate-selective catalyst, which may often be a transition metal (M n+2 /M n ), oxidizes the substrate to the desired product. The reduced form of the catalyst is subsequently reoxidized by the stoichiometric oxidant that could be ideally oxygen or hydrogen peroxide. 218 In several cases however this process fails because the electron transfer between M n and O2 or H2O2 is too slow compared to the decomposition of the reduced metal; that can be circumvented by use of a (i) ligand to stabilize the species (Scheme 71, entry a) or (ii) an extra electron-transfer mediator (ETM) (Scheme 71, entry b). 218

Scheme 71.
Oxidation with a substrate-selective redox catalyst b. Electron transfer facilitated by an electrontransfer mediator (ETM).
We have selected a few examples of oxidation that cover the field of oxidation of 3-phenoxybenzyl alcohol (4c) to 3-phenoxybenzaldehyde (4a). It has been for example reported 223 that choline peroxydisulfate (ChPS) (81), easily generated from the commercially available related potassium salt, performs efficiently neat at 70 °C and in a short time the reagent playing also the role of an ionic liquid (Scheme 72).

Heterogeneous oxidations of 3-phenoxybenzyl alcohol.
Oxidation in a heterogeneous medium offers the advantages of an easy separation of the catalyst once the reaction is over and allow the easy recycling of the catalyst but are usually less efficient due to difficult contact between the starting material and the catalyst. In some cases the oxidant has been inserted in a nickel-alumina matrix 82a to produce the reagent 82b (Scheme 73, entry a; Scheme 74, entry a), 224 or complexed through an organic linker to a magnetic end of a silicon coated the catalyst 83c is easily extracted from the medium using a magnet and has been recycled for at least five time without affecting its catalytic activity. 230

Platinum/palladium-promoted oxidation of 3-phenoxybenzyl alcohol.
Platinum and palladium (1 wt-% and 5 wt-% respectively) deposited on pulverised carbon proved extremely valuable for promoting the oxidation of 3-phenoxybenzyl alcohol (4c) to 3-phenoxybenzaldehyde (4a). 200 Best conditions involve bubbling dioxygen on a heated mixture (80 °C) of the alcohol 4c in the presence of 0.1 equivalents of a 1 wt-% Pt on carbon, 0.5% of lead nitrate in an aqueous solution of sodium hydroxide leading to 4a in up to 90% yield besides 7% of the carboxylic acid 4e resulting from an overoxidation that is easily removed by alkali (Scheme 73, entry c). Each ingredient is essential for the success of the reaction although interchanging lead-by bismuth-nitrate does not affect too much the process. 200 Similar results are obtained by replacing platinum by palladium (Scheme 73, entries f,g). 200 However, the percentage of 4e arising from overoxidation of 4a is higher (12% instead of 7%; Scheme 73, entry f compare to entry d). 200 The palladium catalyzed reaction does not strictly require the use of an additive as it is the case in the platinum catalyzed reaction (Scheme 73, compare entry g to entry e) but in its absence the reaction time increases substantially as well as the overoxidation. 200

Homogeneous oxidation of 3-phenoxybenzyl alcohol.
Specific examples of oxidation of 3phenoxybenzyl alcohol (4c) in a homogeneous solution are gathered in Scheme 75.
Scheme 75. Aerobic oxidation of 3-phenoxybenzyl alcohol (4c) to 3-phenoxybenzaldehyde (4a), involving metal catalysis and iodosyl benzene , TEMPO or TCQ as oxidant. 211,213,[225][226][227][228][229] 3.4.2.2.1. Palladium-promoted oxidation of 3-phenoxybenzyl alcohol. The oxidation of 3-phenoxybenzyl alcohol (4c) to 3-phenoxybenzaldehyde (4a) by dioxygen or even by air, has been achieved also in homogeneous solution, in toluene at 90 °C 225 using Pd(OAc)2 (5 mol%), pyridine (1 eq.) and molecular sieve (MS 3 Å). The role of the latter proved to be essential for the success of the reaction since it catalyzes the decomposition of the hydrogen peroxide, formed in the process, to water and oxygen (Scheme 75, entry a). 225 The process can be advantageously compared to the heterogeneous process that uses the same transition metal catalyst (Scheme 73, entries f,g). 200,222,225 The oxidation takes place with lower loading of catalyst (1 mol%) for benzylic alcohols but the rate of the reaction is slower. Air can be replaced by oxygen as already disclosed but the reaction rate is much slower. The reaction does not take place 225 with (i) other palladium compounds such as PdCl2, Pd(dba)3 or Pd(PPh3)4 (ii) 2,6dimethylpyridine whose nitrogen atom is unable properly to complex the metal, as pyridine does, (iii) If the reaction is carried out in toluene but at reflux of the solvent instead of 80 °C since palladium black is released rapidly and its catalytic activity lost. It has been also mentioned that the yield of 4a greatly diminishes when methylene dichloride, THF, ether or 1,4-dioxane are replacing toluene as the solvent. 225 A tentative mechanism in which Pd(II) species work all along the process is shown in Scheme 76. 225 The reaction is expected to take place via the formation of the Pd(II) alcoholate AP, from the Pd(II)pyridine complex AO, to provide by β-elimination the aldehyde and the Pd(II)-hydride species AQ that reacts with dioxygen to provide the Pd(II)-hydroperoxide species AR. The latter exchanges its ligand with the alkyl group to produce hydrogen peroxide that is decomposed to water and dioxygen by the MS 3 Å molecular sieves at 80 °C (Scheme 76). 225 Scheme 76. Postulated mechanism of the Pd(II) catalyzed aerobic oxidation of alcohols to aldehydes. 225

Ruthenium-promoted oxidation of 3-phenoxybenzyl alcohol.
Complexed ruthenium compounds have been successfully used for the same purpose with iodosylbenzene as oxidant (Scheme 75, entries be). 213,229 The required complexes (Phen-Ru-Phen, Pyr-Ru-Pyr, Quin-Ru-Quin, (Scheme 75) have been synthesized, in a straightforward way, by mixing, at 20 °C for 4 h, stoichiometric amounts of ruthenium trichloride hydrates (1-3) and 1,10-phenanthroline (Phen), 8-hydroxyquinoline (Quin) or 2,2 ′ -bipyridine (Pyr) or their mixture in acetonitrile. Although all exhibit catalytic activities, those containing 8-hydroxyquinoline (Quin) possesses the highest aptitude to oxidize alcohols and offers the advantage of possessing the lower molecular weight and to do not perform overoxidation (Scheme 75, entry d). 229 Related oxidation of benzyl alcohol delivers benzaldehyde in lower yields when using lower amounts of iodosylbenzene or when acetonitrile (100%) is replaced by toluene (18%), THF (31%), or dichloromethane (58%). 229 A related polymer supported catalyst (PS) has also been used for the same purpose (Scheme 75, entry e). 213 The polymeric ligand has been prepared from chloromethyl polystyrene (Merrifield resin) and 5-amino-1,10phenanthroline then reacted with ruthenium trichloride in THF (20 °C, 12 h). The best conditions and solvent proved to be similar to those selected in the homogeneous version (Scheme 78, compare entry e to entries b,c) and similarly no overoxidation is taking place. 213 This supported ruthenium complex can also be reused for at least three time with benzyl alcohol as substrate before its activity decreases (40 mg/mmol benzyl alcohol: 1-3rd recycling: 100%; 4, 5th recycling 71, 55% yield) but using a larger amount of supported catalyst (60 mg instead of 40 mg/mmol) allows an overuse of at least 7 runs. 213

TEMPO-promoted oxidation of 3-phenoxybenzyl alcohol.
Oxidation of 3-phenoxybenzyl alcohol (4c) to 3-phenoxybenzaldehyde (4a) has been advantageously carried out using 2,2,6,6-tetramethylpiperidine Noxide (TEMPO) (83a) or its 4hydroxy-TEMPO (83b) analog (easier to separate by chromatography, cheap, highly reactive) under a large number of conditions (Scheme 75, entries f-i). Most of the conditions use catalytic amount of TEMPO and therefore require the in situ oxidation of the resulting TEMPOH (84a) by-product. This has been effectively achieved by dioxygen (Scheme 75, entry f) 226 or air (Scheme 78, entries g-i) 211,227,228 in some cases in the presence of transition metal catalysts (iron or copper, Scheme 75, entries h,i). 211,227 Originally the oxidation of TEMPO required the presence of a transition metal catalyst and several conditions have been tested in order to perform the aerobic oxidation reaction under mild metal-free conditions. Among the various combinations tested, 226,228 two high yielding conditions have been selected for the oxidation of 3-phenoxybenzyl alcohol (4c).
The first one is carried out neat, without solvent and uses gaseous dioxygen (1 atmosphere), in the presence of a small amount of 4-hydroxyTEMPO (83b) (1%), as well as tetrachlorobenzophenone (TCQ, 2%), tert-butyl nitrite (TBN, 5%) and hydrochloric acid (HCl) generates 4a in extremely high yield at room temperature (98%, Scheme 75, entry f). 226 The other involves a slightly higher amount of TEMPO (5%) but uses air in place of dioxygen, sodium nitrite (NaNO2) instead of TBN and hydrochloric acid to promote the catalytic oxidation that takes place for a little longer time (overnight) at room temperature (Scheme 78, entry g) but minimizes the waste produced. 228 The choice of sodium nitrite and hydrochloric acid takes advantage of the intimate mechanism of the TEMPO oxidation that is presented in Scheme 77, entry a. 228 The processes shown above could be rationalized in terms of the positive halogen species being able to oxidize TEMPO (83a)/TEMPOH (84a) to the oxoammonium cation TEMPO + AR (Scheme 77). 231 It was found that, HCl for example, in combination with TEMPO and NaNO2, is effective in promoting the catalytic aerobic oxidation of the alcohol to the corresponding aldehyde under mild conditions. The role of the acid was assumed to be the donation of a proton (H + ) to NaNO2 to generate NO/NO 2 . The effect of the halide anion on the catalysis may stem from its reaction with NO2 to generate oxidizing species such as NOCl, which is known to oxidize TEMPO (83a)/TEMPOH (84a) to TEMPO (83a). These preliminary studies revealed that acidic conditions and chloride or bromide anions have beneficial impacts on the catalytic oxidation. 228 Importantly, the results clearly demonstrated that oxidizing halogen-containing compounds are not essential to drive the TEMPO/NaNO2-based catalytic system. The TEMPO/HCl/NaNO2 system, efficiently achieves in a wide range of solvents, such as ClCH2CH2Cl, EtOAc, CH3CN, HOAc, and PhF, the conversions of benzyl alcohols at atmospheric pressure and ambient temperature (without dioxygen or air bubbling). 228

Scheme 77.
Mechanism for the TEMPO mediated process that imply any oxidant such as dioxygen as the cooxidant (Scheme 77, entry a) and a related one that includes copper acetate (Scheme 77, entry b). 228,231 It was also found that silica gel adsorbed 4-hydroxyTEMPO, smoothly catalyzed the aerobic oxidation of 3phenoxybenzyl alcohol (4c) in the presence of Fe(NO3)3·9H2O and NaCl (1/1) in a nonpolar solvent (Scheme 75, entry h). 211 The reaction proceeds at room temperature in six hours and the resulting 3-phenoxybenzaldehyde (4a) is obtained in excellent yield just after filtration and moreover the recovered the catalyst can be reused at least six times without loss of catalytic activity. 211 It has been reported that 4-hydroxyTEMPO possess a much higher propensity (x8) than TEMPO to be absorbed on silica gel and that the mixture of sodium chloride and iron nitrate cannot be replaced by ferric chloride (FeCl3). Although ferric chloride proved in some cases to be more efficient than the sodium chloride, iron nitrate combination, its high Lewis acid acidity often promotes unwanted competing reactions. 211 The requirement of chlorine for successful result could be related to the comment discussed in the previous transformation. 228 Finally, 3-phenoxybenzyl alcohol (4c) has been oxidized by "ligand-and additive-free" process involving the couple Cu(OAc)2/TEMPO as catalyst that enables efficient and selective aerobic oxidation, at low catalyst loading, of a broad range of primary and secondary benzylic and aliphatic alcohols to the corresponding aldehydes and ketones. 227 This ambient temperature oxidation protocol is of practical features like aqueous acetonitrile as solvent, ambient air as the terminal oxidant, and low catalyst loading, presenting a potential value in terms of both economical and environmental considerations. 227 Based on the experimental observations, a plausible reaction mechanism was proposed that originates from the original work of Semmelhack on TEMPO (Scheme 77, entry b). 231 Accordingly cupric ion effects one-electron oxidation of TEMPO (83a) to the nitrosonium ion AR that in turn is able to oxidize the alcohol to the aldehyde and return the hydroxyl amine TEMPOH (84a) (Scheme 77, entry b). Rapid symproportionation of 84 with AR regenerates TEMPO (83a). Finally, Cu(I) is regenerated by dioxygen in a process that consume protons and gives Cu(II) and water in a well know process. The net reaction is the oxidation of the alcohol to the aldehyde and water (Scheme 77, entry b)" with (i) no formation of the Scheme 81. Synthesis of 3-phenoxybromo-benzene from bromobenzene substituted at C-3 by a hydroxy, an iodonio-or a bromo-group. 155,232,233 Better results have been obtained on coupling phenol (5d) with the 3-bromo-iodoniums 6eI+ that possess two leaving groups exhibiting different reactivity (Scheme 81, entry b) in a reaction that takes place, 234 as already discussed, under milder conditions and does not require a copper catalyst (Scheme 53, entry a). 151 The choice of the mesityl iodidonium among different other iodonium is documented. 234 The selected one offers the advantage (i) to react selectively on the phenyl instead of the 2,4,6-methyl phenyl group and (ii) to be recyclable. 234 Even better results have been obtained from 3-bromo-phenol (5g) as starting material and triphenyl tin chloride 6eSn that requires a copper catalyst (Scheme 81, entry a) 155 in a reaction that has been also used for the synthesis of 3-phenoxytoluene (4b) (Scheme 58, entry a), 155 and for which a mechanism has been proposed (shown in Scheme 55, entry b). 155 3.5.2. Synthesis of 3-phenoxybenzaldehyde from 3-phenoxybenzoic acid/ esters. The transformation of 3phenoxybenzoic acid/ esters 4e to 3-phenoxybenzaldehyde (4a) is shown in two different approaches (Scheme 61, entry a; Scheme 81, entry b). The Rosenmund reduction, that involves the transformation of the carboxylic acid 4e to its acid chloride 4j using thionyl chloride and the catalytic hydrogenation of the latter to the aldehyde 4a, is probably the most efficient large-scale transformation especially in the industrial context (Scheme 61, entry a). The strategy used to produce the tiny amount of 14 C-radiolabelled 3-phenoxybenzaldehyde 14 C (4a) has been achieved in two steps that involves the reduction of the aldehyde using lithium aluminum hydride 209 leading to 14 C-radiolabelled 3-phenoxybenzyl alcohol 14 C (4c) and selective oxidation to the aldehyde 14-C 4a using chromium oxidant 210 (Scheme 81, entry b). This transformation is unsuitable for industrial purpose.

Multistep syntheses of 3-phenoxybenzaldehyde from 3-phenoxybenzonitrile
This approach has not been documented; nevertheless the synthesis of 3-phenoxybenzonitrile (4f) according to the two strategies traditionally used to produce the ether linkage from phenol or from a 3-functionalized phenol have been reported as presented below. Furthermore, the transformation of aromatic nitriles to the corresponding aldehydes directly or after transformation to the corresponding carboxylic acids is a wellestablished route. 3.6.1. Syntheses of 3-phenoxybenzonitrile. 3.6.1.1. Synthesis of 3-phenoxybenzonitrile from 3hydroxybenzonitrile. The synthesis of 3-phenoxybenzonitrile (4f) from 3-hydroxybenzonitrile (5f) has been achieved on reaction of diphenyliodonium tetrafluoroborate 6eI+ in the presence of potassium t-butoxide (Scheme 82), 151,149 in a process similar to the one involved in the synthesis of 3-phenoxybenzaldehyde (4a) from 3-hydroxybenzaldehyde 5a (Scheme 57), 151 or the synthesis of 3-phenoxybromobenzene (4g) from phenol (5e) (Scheme 81, entry b). 234 Scheme 82. Transformation of 3-hydroxybenzonitrile to 3-phenoxybenzonitrile. 149,151 3.6.1.2. Synthesis of 3-phenoxybenzonitrile from 3-halogenobenzonitrile. The second approach that involves 3-halogenobenzonitriles 6f, phenol (5d) and a base or a preformed phenate, is better documented (Scheme 83). 169,176,184,235 All the reactions involve copper as the pre-catalyst and belong to the category of coupling with a poorly reactive aryl halide since they bear an electron withdrawing group. Therefore, the ligand plays a key role to favor the coupling.
Thus, in 3-phenoxybenzonitrile (4f) is produced in no more than 15% yield on reaction at 155 °C of 3-chlorobenzonitrile (6fCl), potassium phenate in the presence of copper chloride in anisole as the solvent (Scheme 83, entry a), 184 and a dramatic increase in the yield, up to 85%, is observed by performing the reaction under the same conditions but in the presence of no more than 1 mol% of the tridentate ligand 52 (Table 7, Scheme 83; compare entries a,b). 184 Similarly: performing the reaction in the presence of no more than 1.25 mol% of "Nano-Cu" (from sodium citrate-assisted generated CuI nanoparticles) in DMF not only increases the reaction yield (+8%) but at the same time allow to perform coupling at lower temperature (by -45 °C) and reaction time (-2 h) compared to use liganded copper catalyst (Scheme 83, entry c). 235 It also offers the possibility to recycle, after ethyl acetate washing and drying at 65 °C, the catalyst isolated by centrifugation. 235 Finally, using the air stable Cu(I)-bipyridyl complex 38M offers the advantage of an extremely low loading of the catalyst and lowering the reaction temperature (Scheme 83, entry d). 176 It however suffers from a much longer reaction time (24 h), use of the more expensive aryl bromide and of lower yield in 6f (Scheme 83, entry d). 176 Similarly, the photochemically promoted Ullmann reaction, requires the even more expensive 3-iodobenzonitrile but offers the advantage to carry the coupling at room temperature. 169

Conclusions
We have discussed in this chapter several syntheses of 3-phenoxymandelonitrile involved in the synthesis of cypermethrin and deltamethrin, one of the most active man-made but nature-inspired pyrethroid insecticides.
We have presented the strategy and methods specifically devoted to the synthesis of diaryl ethers bearing a formyl group and one of its precursors, especially the original Ullmann coupling that uses copper catalysis and its modifications that use copper ligands and the more recent Buchwald-Hartwig coupling that instead use palladium catalysts. Alternative methods such as Barton-Gagnon are also presented.
We have also extensively reviewed the different methods to synthesize cyanohydrins from aldehydes especially that derived from 3-phenoxybenzaldehyde. Methods for separation of the enantiomers, recycling the unwanted enantiomer as well as their enantioselective synthesis, including enzymic ones, are reported.
We have only reported the methods and conditions that specifically refer to the synthesis of the selected products. Several have not been invented to specifically allow such synthesis but have been used as model to generalize those methods. Other methods could apply to the synthesis of such compounds and there are rooms to test them.