Syntheses and catalysis of transition metal complexes with hemilabile ligands

Chiral diaryl methanols are important intermediates for the synthesis of medically active compounds Figure 5.1.135 Generally, they can be obtained either by reduction of the correspondin

Chapter 5 – Palladium Catalyzed C-C bond Formation Involving

5.1.1 Palladium catalyzed Suzuki-Miyaura cross coupling reactions

Palladium-catalyzed Suzuki-Miyaura cross-coupling between organoboranes and electrophiles has become one of the most important C-C bond formation methodologies (Equation 5.1).128 It enables facile syntheses of biaryls and alkylaromatics that are

organo-intermediates of pharmaceuticals, natural products and stereoselective reactions etc.130

For example, tetra-ortho-substituted biaryls can be prepared from this method with the

use of bulky phosphines,131 hybrid ligands,132 and carbene complexes133 (Scheme 5.1 and 5.2) etc There are however some drawbacks, such as the cost of boron reagents, need for high temperatures and high Pd loadings (2-12 mol%) etc These have restricted the use of

asymmetric Suzuki cross-coupling reactions in producing biologically active biaryls134and prompted the development of more efficient catalytic systems especially on ligand innovations Therefore, part of our research focus is to develop an efficient ligand/Pd catalyst system for coupling sterically hindered substrates and preferably, to extend further for asymmetric Suzuki-Miyaura cross-coupling reactions.

Ar Ar

Ar = Ph, 10 mol% Pd used, 25h, Product Yield = 25%

Scheme 5.1 Reported Pd systems used to catalyze Suzuki – Miyaura cross couplings of

hindered Suzuki naphthalene substrates.131a-b,132b

100 - 110oC12-18h

X = Br,Cat loading = 2mol%,Product Yield = 88%

5.1.2 Palladium catalyzed 1,2- addition of organoboronic acids to aldehydes

Chiral diaryl methanols are important intermediates for the synthesis of medically active compounds (Figure 5.1).135 Generally, they can be obtained either by reduction of the corresponding unsymmetrical diaryl ketones or enantioselective aryl transfer reactions to aldehydes (1,2-addition reactions) (Scheme 5.3).136 However, both methods have severe limitations and only work well for a limited range of substrates It is accordingly pertinent

a need to develop alternative pathways that can be applied to wider variety of carbinol derivatives

OH

HO

Figure 5.1 Some examples of diaryl methanols as intermediates for the synthesis of medically active compounds.135

OHO

OH

catalyst[H]

catalystR"2Zn

Scheme 5.3 General Synthetic routes to obtain biaryl methanols.136

The 1,2-addition to aldehydes has been investigated with a variety of protocols employing organoboron,129c,137 organozinc,138 organotin139 and organosilane140 compounds Among the catalytic methodologies used, the rhodium catalyzed 1,2-addition of arylboronic acids

to aldehydes is especially notable. 137a-b,f, 141 It would also be desirable to use cheaper metals Promising palladium catalysts such those of phosphapalladacycle,137c anionic palladacycles137i and thioether-imidazolinium137k that are active towards the 1,2-addition

reactions have accordingly emerged (Figure 5.2) On the other hand, use of palladium complexes with hybrid ligands towards catalytic additions of arylboronic acids to aldehydes is not known It is therefore of our interest to investigate the efficacy of our Pd/[P, N] ligands system towards 1,2-addition of arylboronic acids to aldehydes

5.1.3 Palladium catalyzed 1,4- conjugate additions of organoboronic acids to α,β unsaturated ketones

-Some of the enantiomerically enriched compounds possessing a chiral diarylmethane unit are known to be biologically active in antimuscarinics,142 antidepressants,143 and endothelin antagonists,144 and have received growing attention for their syntheses One of the straightforward methods is via conjugate addition of an aryl nucleophile to electron-

deficient olefins substituted with another aryl groups at the β position The 1,4-addition of arylboronic acids to enones and related substrates is a versatile method reported.137f,141c

Although the catalysts reported are dominated by rhodium-based complexes,129c,140c,145palladium(0)-SbCl3 (cat A),146 palladium-bipyridine (cat B),147 cationic palladium(II)

complex (cat C)137g-h and palladacycles (cat D and E)148,137i have been reported as active

catalysts for 1,4-addition reactions of arylboronic acids with α,β-unsaturated ketones and

α-ketoesters (Scheme 5.4) Despite the ready availability of the reagents, the conjugate addition of arylboronic acids to activated ketones has been elusive until now Therefore, it

is of interest to explore the potential application of the mixed donor hybrid ligands in the palladium-catalyzed 1,4 addition reaction of arylboronic acids to these ketones as well

Pd PPh2OAc

O

OAr

Scheme 5.4 Reported active catalysts for 1,4-addition reactions of arylboronic acids with

α,β-unsaturated ketones and α-ketoesters.137g-i, 146-148

5.2 Results and discussion

5.2.1 Catalytic Studies

5.2.1.1 Suzuki-Miyaura cross coupling reactions of aryl halides and organoboronic acids

The ligand system (t-Bu)2PFcC=NCH(CH3)R (Fc = ferrocenyl (C5H4)2Fe) is chosen as a model for this study because it contains both strong (phosphine) and weak (imine) donor sites and that they are separated by a conformationally flexible ferrocenyl skeleton which would in principle enable the weak donor to undergo facile reversible coordination.8a It is also important that both donating sites contain a substituent group that can be changed This allows the introduction of substituents (R or R’) to systematically and independently alter the electronic and steric properties of both sites.149

Within this ligand system, there are two types (Figure 5.3) The former (ligand L1a) has

a fixed phenyl on the imine but the R on the phosphine varies, whereas the latter (ligand

L1k-o) has a variable group on the imine with a constant tBu on the phosphine Ligand

L1a is known to support Pd-catalyzed Suzuki coupling of aryl chlorides and aryl boronic acids8a-band Ni-catalyzed ethylene oligomerization.8c

We first compared the catalytic ability of L1a and L1k–o towards extremely hindered

substrate combination involving 1-bromo-2-methylnaphthalene and boronic acid in the presence of Pd2(dba)3 under typical Suzuki cross-coupling reactions (Table 5.1) Change of -Ph to -CH(CH3)(Ph) could be manifested in chelate dissociation and halide-bridge formation, thus reducing the metal sphere congestion promoting metal-substrate interaction Indeed, a significant increase in the cross coupling product is observed (Table 5.1, entry 1 and 2) It is evident that the product yields are sensitively dependent on R (Table 5.1, entry 3 – 6) One possible explanation is the different oxidative addition products that are formed as a result of the hemilability of the ligands

2-methylnaphthyl-1-Table 5.1: Ligand effect on the Suzuki cross-coupling reactions of methylnaphthalene and 2-methylnaphthyl-1-boronic acid.a

1-bromo-2-Entry Ligand Isolated Yieldb

a3 equiv base b Not optimized cIsolated Yield is 100% if CsF is used.

Since L1o gives the highest yield in the above coupling, it is used as a model to examine

the efficiency of sp2-sp2 Ar-Ar’ couplings (Table 5.2) The yields are satisfactory with many near-quantitative conversions Consistent with published results128a,e-f, sterically

B(OH)2Br

1.2 mol% ligand 0.5mol% Pd2(dba)3THF, Cs2CO3reflux, 24hr

+

demanding substrates require higher catalytic loads to achieve satisfactory yields (Table

5.2, entries 6-8).Cross-coupling using sterically bulky substrates to form

tetra-ortho-substituted products were obtained in remarkably high isolated yields at low catalytic loads of 1.0 mol% Pd (Table 2, Entries 6-8) These catalyst loadings are comparable or lower than the reported catalyst systems.131-132 For example, only 0.1 mol% of Pd2(dba)3

(with 0.23 mol% of L1k) is sufficient to promote quantitative coupling of

1-bromonaphthalene and naphthyl-1-boronic acid at r.t (Table 5.2, Entry 1) This is more favourable than the coupling of 1-bromonaphthalene and naphthyl-1-boronic acid at 1 mol% of Pd(dba)2 and 2 mol% of 134 at r.t over 2 days.131b Although electron-rich aryl halides tend to be more sluggish in Suzuki couplings,128 electron-rich 1-bromo-2-methoxynaphthalene can perform better than the electron-poor 1-bromonaphthyl-2-aldehyde in this system when sterically demanding substrates are used

This system is also effective towards a range of aryl chlorides at low Pd loadings of 0.05–1.0 mol% The coupling of activated para-substituted aryl chlorides and phenylboronic acids can be effectively carried out under ambient conditions (Table 5.3, Entry 1-7), which is comparable to many efficient systems using bulky phosphines131,150 and N-heterocyclic carbenes (NHC).133,151 The more sterically hindered substrates can also be effectively consumed under higher temperatures and high catalytic loadings (Table 5.3, Entry 3 – 7)

Table 5.3 sp2-sp2 Suzuki cross-coupling of aryl chlorides and aryl boronic acids catalyzed

by Pd2(dba)3 with ligand L1o.a

Isolated

% Yield

4d 4-chlorobenzonitrile B(OH)2 0.1 reflux, 24 h 100

6e 2,5-(MeO)2C6H3Cl PhB(OH)2 1 reflux, 12 h 86

a Pd:ligand = 1:1.2 b Not optimized. c At 50 °C, 12 h and Pd loading = 0.01 mol%, isolated yield = 100% d1,4-dioxane was used instead of THF

Sp2-sp3 couplings using hexylboronic acid could be achieved with good yields (Table 5.4) that are comparable with the few known systems.150a This subset of Suzuki cross coupling reaction is rare According to the survey conducted by Suzuki and Miyaura, the alkyl boronic acids are the least reactive coupling partners (Chart 5.1).152 This system does not require the normal use of Tl(I) or Ag(I) bases152 or air-sensitive trialkylboranes such as 9-BBN derivatives,131c trifluoroalkylborates153 and dialkylpinacolborates.154 It is however ineffective towards sp3-sp3 couplings

Table 5.4. sp2-sp3 Suzuki cross-coupling of aryl halides and alkyl boronic acids catalyzed

by Pd2(dba)3 with ligand L1o.a

Cs2CO3, THF,reflux

Ooctyloctyl

octyl

Chart 5.1 Reactivity of alkyl boranes surveyed by Suzuki and Miyaura et al.152

5.2.1.2 1,2- and 1,4- additions of organoboronic acids to carbonyl compounds

Initial experiments were carried out to explore the effect of additives towards our

Pd/ligand L1k system (Table 5.5) Prelimary catalytic testings showed that the Pd/ligand L1k system catalyzed 1,2-addition of phenylboronic acid to benzaldehyde in the presence

of CHCl3 as additive to give the corresponding alcohols at quantitative yields of 82% (Table 5.5, Entry 2) Recently, Ito and Ohta also reported that the presence of chloroform

is essential in this type of palladium-catalyzed reaction of arylboronic acids with aldehydes.137d Low product yield (20%) was obtained when an organozinc compound like diethylzinc was added to the catalytic reaction No conversion was observed when salts like ZnCl2, InCl3, AlCl3 were added It suggested that the presence of an organo group in the additives might be needed to trigger the addition reaction 1,2-addition of phenylboronic acid to benzaldehyde

Table 5.5 1,2-Addition of arylboronic acids to aldehydes catalyzed by Pd2(dba)3 with

Pd2(dba)3/ L1k

AdditiveDioxane

Among the metal precursors tested, only Pd2(dba)3 and [Rh(COD)Cl]2 could show satisfactory yields (82% and 98% respectively, Table 5.6, entries 3 and 11) Use of Pd(II) compounds such as PdX2 (X = Cl, OAc) or Pd (COD)Cl2 are less effective (4-39%, Table 5.6, entries 4 – 7) Other d10 compounds (such as CuX (X = Cl, I) and Ni(COD)2 are largely inactive The reported active rhodium and palladium catalysts reported are invariably supported by diphosphines (dppf, dppp, dppe)137f and monodentate phosphines [e.g PPh3137d and P(1-Nap)3 (1-Nap = 1-Naphthyl)]116e respectively The activity shown

by a potentially hemilabile ligand such as L1k suggests that it is more compatible with

the metals as it could perform dual function, viz monodentate and chelating

Table 5.6: Effect of metal precursor towards 1,2 –addition of phenylboronic acid to benzaldehyde.a

Entry Metal Precursor %Yield

aReaction conditions (not optimized): benzaldehyde (1.0 equiv), phenylboronic acid (2.0

equiv), L1k (1.0 mol%), metal precursor loading (1.0 mol%), Cs2CO3 (2.0 equiv), CHCl3

(0.1 mL) 1,4-dioxane (2 mL), 70°C, 16 h bIsolated Yields (average of 2 runs)

Further catalytic investigations were then performed using the palladium solvent adduct,

Pd2(dba)3·CHCl3 as the palladium(0) source instead This reaction is strongly solvent dependent, showing good yields for 1,4-dioxane and CHCl3 (98% and 82% respectively, Table 5.7, entries 1 and 6) but negligible for the likes of PhMe, MeOH and MeCN The carbonates of K+ or Cs+ are the best supporting bases (89% and 98% respectively, Table

5.8, entry 1-2) Increasing the L1k:Pd ratio from 1.2 to 2 would also suppress the product

aReaction conditions (not optimized): benzaldehyde(1.0 equiv), phenylboronic acid (2.0

equiv), L1k (1.0 mol%), Pd loading (1.0 mol%), Cs2CO3 (2.0 equiv), solvent (2 mL), 70°C, 16 h bIsolated Yields (average of 2 runs)

Table 5.8: Effect of base towards 1,2-addition of phenylboronic acid to benzaldehyde.a

Entry Base %Yieldb

aReaction conditions (not optimized): benzaldehyde (1.0 equiv), phenylboronic acid (2.0

equiv), L1k (1.2 mol%), base (2.0 equiv), Pd loading (1.0 mol%), 1,4-dioxane (2 mL), 70

°C, 16 h bIsolated Yields (average of 2 runs)

Based on the conditions optimized above, a list of ligands at 1 mol% was screened towards the 1,2-addition of phenylboronic acid to benzaldehyde (Chart 5.2) The ligands cover monophosphines[(C5H5)Fe(C5H4PPh2)], diphosphines (dppf, dppp, dppb) and a range of monophosphine with a secondary donor of varying strengths (Chart 5.2)

Chart 5.2 Ligand screening towards Pd-catalyzed 1,2-addition of phenyl boronic acid to benzaldehyde.a

O

H

B(OH)2

OH+

Ligand

Pd2dba3.CHCl31,4-dioxane

Cs2CO3, 70oCovernightLigand:

PR2

N

PR2N

PR2N

R = Cy, L1d: Product Yield = 85%

= Ph, L1f: Product Yield = 89%

=tBu, L1k: Product Yield = 98%

R = Cy, L1e: Product Yield = 23%

= Ph, L1j: Product Yield = 56%

=tBu, L1o: Product Yield = 90%

R = Ph, L1a: Product Yield = 86%

R = tBu, L1c: Product Yield = 50%

From Chart 5.2, the hemilabile iminophosphine ligands (L1a, L1c-f, L1j, L1k & L1o)

generally show the highest activities, surpassing the diphosphines (9-23%) This is in agreement with literature results which also reported mediocre yields for diphosphines (e.g dppp, dppb, Yields: <52%)137d-e, monodentate phosphines/Pd and the palladacycles which require at least 5mol% of Pd loadings to achieve satisfactory yields (73 – 98%).137d-e,i The yields are dependent on the substituents on both donors Bulky

phosphine substituents tend to perform the best The aminophosphine 59, in stark contrast

to the iminophosphine counterpart, is effectively inactive This could be attributed to the much stronger sp3 amine as a donor, thus giving a chelate as stable as that of a diphosphine The donor strength is not the only parameter that influences the catalytic efficacy Even the ligand skeleton has a strong effect on the yields This is exemplified

when replacement of the ferrocenyl skeleton (in L1f & L1j) by an aryl (in L3) would

inactivate the system This reflects the advantage offered by a skeletally flexible framework which could support the engagement and disengagement of the second donor The ability for the ferrocenyl iminophosphine to switch facilely between the monodentate and chelating coordination modes is hence believed to be a key parameter that influences

the catalytic efficiency This is further evident in 13 & L2 when the yields are mediocre

(47 – 50%) as the basicity of the second donor is weakened to the extent that the ligand resembles more a monodentate than a hemilabile chelating system

The catalytic addition is effective towards a range of aldehydes (Table 5.9), giving excellent yields (92 – 100%) except when electron-rich aldehydes or arylboronic acids were involved (33 – 68%) are used A slower transmetallation in these arylboronic acids makes the system more prone to homocoupling and protodeboronation side reactions.155

Table 5.9 1,2-Addition of arylboronic acids to aldehydes catalyzed by Pd2(dba)3·CHCl3

8 Benzaldehyde 2,6-(Me)2C6H3B(OH)2 66

9 4-methylbenzaldehyde 4-(MeO)C6H4B(OH)2 68

aThe reaction was carried out with aldehyde (1.0 mmol), arylboronic acid (2.0 mmol), Cs2CO3

(1.0 mmol), Pd2(dba)3·CHCl3 (0.5 mol%) and ligand L1k (1.0 mol%) in 2 mL of 1,4-dioxane, at

70 ºC, 16 h b 96% when 2.5 mol% Pd2(dba)3·CHCl3 was used

For comparisons, the activities of these hybrid ligands towards the 1,4-addition of

arylboronic acids to α,β-unsaturated ketones, represented by trans-chalcone

(1,3-diphenyl-2-propen-1-one), under similar reaction conditions have also been examined

(Table 5.10) Similar to the 1,2-addition described above, the tBu-substituted

iminophosphine L1k is also the best performer Electron-rich or sterically hindered

arylboronic acids are also less effective (Yield: 43 – 78%, Table 5.10, entries 5 -7) Other

α,β-unsaturated ketones examined (such as 2-cyclohexenone, benzylideneacetone, methyl-3-hexen-2-one…) are inactive

5-Table 5.10 1,4-Addition of Arylboronic Acids to trans-chalcone catalyzed by

Pd2(dba)3·CHCl3 with selected mixed hybrid donor ligand.a

O Ph

Pd2(dba)3.CHCl3Ligand

1,4-Dioxane,Cs2CO3

+ R'-B(OH)2Ph

O Ph Ph

a The reaction was carried out with trans-chalcone (1.0 mmol), arylboronic acid (2.0 mmol),

Cs2CO3 (1.0 mmol), Pd2(dba)3·CHCl3 (0.5 mol%) and ligand (2.2 mol%) in 2 mL of 1,4-dioxane,

at 70 ºC, 16 h

5.2.2 Coordination studies of potential catalytic active palladium [P, N] intermediate

formed between [P, N] ferrocenyl ligands and Pd(0) precursors

5.2.2.1 [P, N] palladium complexes formed oxidative addition in Suzuki-Miyaura

cross-coupling reaction

Extensive synthetic and mechanistic studies have been carried out over the past decades.128 The commonly accepted mechanism of Palladium catalyzed Suzuki – Miyaura cross coupling reaction involves an initial oxidation addition of the aryl halide to

a palladium (0) species, followed by transfer of the aryl group from the boronic acid to the palladium(II) centre (transmetallation) and eventually by the reductive elimination to give the biaryl product and regenerate the catalytically active palladium(0) species

(Scheme 5.5).128 However, when hindered substrates were involved, the process was complicated by protonolysis or deboronation.155 Therefore, fast reductive elimination to compete with hydrolysis and deboronation was the key to the success of the reaction A

more comprehensive mechanism had been proposed by Cammidge et al (Scheme 5.6).144c

Observations that hindered aryl iodides gave faster and better reactions during the cross coupling of hindered aryl halides and boronic acids suggest that reductive elimination might be triggered by oxidative addition step in the mechanism.155

Transmetallation

Scheme 5.5 A general mechanism for Suzuki-Miyaura cross coupling reaction.128

LPdL

ArX

LPdX

ArL

Ar'L

Ar'-B(OH)2Nu

Ar'-H

Deboronation

Ar'B(OH)2Base

n-Oxidative Addition

L1o with Pd2dba3 and aryl halides, ArX such as pentafluoroiodobenzene, iodonaphthalene or 1-bromonaphthalene (Scheme 5.7)

C12a:R = C6F5

C12b: R = Nap

P Pd

II

R

PdRP

Scheme 5.7 Reaction of ligand L1o and Pd2(dba)3 with aryl halides

Reaction of L1o with Pd2(dba)3 in situ to give proposed intermediate

{η-C5H4CH=N[CH3(CH)C10H7]}Fe[η-C5H4P(t-Bu)2]Pd(dba), C11, which undergoes oxidative addition with pentafluoroiodobenzene or 1-iodonaphthalene to give the d8 Pd(II) complex {η-C5H4CH=N[CH3(CH)C10H7]}Fe[η-C5H4P(t-Bu)2]Pd(I)(R), C12a–b with aryl

trans to imine nitrogen.8b Complex C12a (ESI m/z 811; δP 75.2 ppm) is thermally unstable, which is exemplified for its decomposition to C11 (major, ESI m/z 633.2; δP

56.2 ppm) when subject to heating in THF at 60°C for 1 h (Figure 5.4) in the presence of

dba The ability of dba to keep the metal (in form of C11) in solution is apparent.158 Upon

prolonged stirring in the absence of dba, C12a converts to

{{η-C5H4CH=N[CH3(CH)C10H7]}Fe[η-C5H4P(t-Bu)2]Pd(I)(C6F5)}2, C13a (major species, δP

86.7 ppm) instead However, unlike C12a, complex C12b converts to

{{η-C5H4CH=N[CH3(CH)C10H7]}Fe[η-C5H4P(t-Bu)2]Pd(I)(C10H7)}2 C13b (major) when

heated at 60°C for 24 h The formation of C13 from C12 highlights the ability for the

hemilabile iminophosphine to adapt to dimerization

Figure 5.4. The 31P-{1H} NMR spectra of complex {{η-C5H4CH=N[CH3(CH)C10H7]} Fe[η-C5H4P(t-Bu)2]Pd(I)(C6F5)}2, C13a at room temperature (a) and the decomposition

of complex {{η-C5H4CH=N[CH3(CH)C10H7]}Fe[η-C5H4P(t-Bu)2]Pd(I)(C6F5)}2, C13a at 60°C (b) in the presence of dba

Reaction of 1-bromonaphthalene with L1o/Pd2(dba)3 or C11 gives a mononuclear Pd(II)

species, {{η-C5H4CH=N[CH3(CH)C10H7]}Fe[η-C5H4P(t-Bu)2]}2Pd(Br)(C10H7), C14, with two iminophosphine ligands with dangling imines (Scheme 5.7) The contrast

between C14 and C13 suggests that the oxidative addition product is dependent on the

5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95

5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95

(ppm)

File 1 : C:\WIN1D\SPC\ASPX32\SH0827~1\004001.1R File 2 : C:\WIN1D\SPC\ASPX32\SH0924~1\011001.1R

(a)

(b)

aryl halide, as reported.46b,156b NMR analysis of C14 suggests that it exists as a mixture of

cis- (δP = 60.74) and trans- (δP = 60.01) isomers The 31P NMR resonances of C14

broadened at 60°C (Figure 5.5) Complex C14 decomposes to Pd black and L1o under

elevated temperatures

5 10 15 20 25 30 35 40 45 50 55 60 65 70

75

5 10 15 20 25 30 35 40 45 50 55 60 65 70

75

5 10 15 20 25 30 35 40 45 50 55 60 65 70

75

(ppm)

Figure 5.5. The variable temperature 31P-{1H} NMR study of C5H4CH=N[CH3(CH)C10H7]}Fe[η-C5H4P(t-Bu)2]}2Pd(Br)(C10H7), C14, at room temperature (a), 60 °C (b) and back to room temperature (c)

{{η-(a)

(b)

(c)

5.2.2.2 Contribution of chloroform in promoting palladium catalyzed 1,2 addition of

phenylboronic acids to aldehydes

The dynamic coordinative abilities of 1,1’-iminophosphino ferrocenyl ligands L1a-c offers a possible reason how a ligand such as L1k could best support the catalytically

palladium Based on the catalytic cycles proposed by Miyaura and Ohta et al., the

beneficial effect of chloroform in promoting the 1,2-addition reactions137d,159 can be explained by the formation of Pd(II) chloro(dichloromethyl) complex {η-C5H4CH=N[CH3(CH)C6H5]}Fe[η-C5H4P(t-Bu)2]Pd(Cl)(CHCl2), C15 as an oxidative addition product (Scheme 5.10) The mixture of Pd2(dba)3, ligand and chloroform were made Preliminary analyses were performed on a one-pot reaction of Pd2(dba)3, ligand

L1k in chloroform stirred at temperature of 60°C for 3 h has provided spectroscopic

support for C15 This complex provides an entry to catalytic cycle through

transmetallation Keto or enone-coordination is facilitated by de-chelating of the hemilabile iminophosphine, which is followed by insertion to the Pd-Ph moiety to give alkoxy or enolate, leading finally to the product through hydrolysis (Scheme 5.11)

Ph

PdP

PdNP

N

OH

PdN

2

Ph

CHCl2CHCl2

Pd2(dba)3+ P N

Heat

PhB(OH)2Base

Pd

PhR'

R

O

PdROPhR'

5.3 Conclusion

In conclusion, we have demonstrated that the steric and electronic properties of a ferrocene ligand can be tuned to raise the application value of the catalyst system with a wider scope This results in higher performance in Suzuki-Miyaura cross coupling of challenging substrates These ligands could also be applied to palladium catalyzed addition of arylboronic acids to carbonyl compounds in the presence of chloroform Further attempts to understand the contribution of the [P, N] ligands towards the Pd intermediates in the Suzuki-Miyaura cross coupling reaction and 1,2-addition of arylboronic acids to alehydes were made Important structural data of the Pd intermediates were obtained These observations would allow us to design better catalysts that could switch easily among different forms of active forms that could meet the demands of a range of catalytic reactions or the same reaction but carried out under different conditions

Chapter 6 – Potential Applications of Chiral 1,1’-Iminophosphine

ferrocenyl ligands in Asymmetric Rhodium–Catalyzed Hydrosilylation

of Ketones to Alcohols

6.1 Introduction

6.1.1 Importance of optically active secondary alcohols

Innovation in asymmetric catalysis development is being driven by the increasing demand for chiral compounds, especially in the production of chiral drugs in pure stereoisomer form as active ingredients in pharmaceuticals.160 There is a growing demand for chiral alcohols because of the vast application of these compounds in chemistry (Figure 6.1).161Optically active secondary alcohols incorporate many structures with biological activity They are structural units in pharmaceuticals, perfumes, herbicides, pesticides, and important components of liquid crystals.161a The hydroxyl group is also an excellent chiral building block or precursor for the synthesis of a wide variety of chiral compounds (Figure 6.2).162 Thus,numerous methods have been explored for the synthesis of optically active secondary alcohols Many of these synthetic procedures involve the transformation

of prochiral ketones, namely through enantioselective reduction such as asymmetric hydrosilylation and asymmetric hydrogenation.32a

OH HO

β-Agonist

Figure 6.1 Some useful optically active secondary alcohols.161

Figure 6.2. Some applications of chiral alcohols as building block or precursors for syntheses of chiral compounds

6.1.2 Asymmetric rhodium-catalyzed hydrosilylation of ketones to alcohols

Asymmetric metal-catalyzed hydrosilylation of prochiral ketones (Equation 6.1) has been

a versatile method in the production of optically active secondary alcohols which is of great importance in preparative organic chemistry.163 It leads to optically active silyl ethers; subsequent hydrolysis of these ethers gives the corresponding carbinols Owing to the mild reaction conditions involved in the absence of water- and air-sensitive metal hydrides and pressurized hydrogen gas, it is also an attractive alternative to current catalytic hydrogenation method.164 Since the past decades, intensive studies have been focused on the development of efficient hydrosilylation chiral catalysts and impressive chemical yields and enantioselectivities have been attained.165 Chiral rhodium complexes have been widely used as catalysts in asymmetric hydrosilylation of ketones owing to their high catalytic activity.166

Equation 6.1 Metal-catalyzed asymmetric hydrosilylation of ketones.163

The rhodium-catalyzed hydrosilylation of acetophenone is believed to occur according to the catalytic cycle shown in Scheme 6.1.167 The first step is the oxidative addition of the

silane to form Rh(III) complex, 142 In the second step, the ketone coordinates to form species 143 The third step is insertion of the carbonyl bond into the rhodium – silicon bond to give 144 Subsequent reductive elimination completes the cycle, releasing the

silyl alkyl ether and reforming the Rh(I) catalyst To obtain the secondary alcohol, the silyl group is removed by hydrolysis Depending on the ligand used, the silyl enol ether

145 (upon hydrolysis reverts back to the starting ketone) is a common by-product The quantitative analysis of ligand effects concerning the stereochemistry of the reaction reveals a non-linear relationship between ligands’ steric effect and observed enantioselectivity.168 There is also no simple correlation between the size of the silane and the enantioselectivity of the reaction.168

L

L

Rh SiHPh2Cl

H O Ph

OSiHPh 2

Ph +

O Ph

Oxidative addition

Association Insertion

Reductive elimination

142

143 144

6.1.3 Reported ligands in rhodium-catalyzed asymmetric hydrosilylation of ketones to alcohols

Three major ligand families have been developed ensuring high performance in rhodium catalyzed enantioselective reduction of ketones to secondary alcohols They are nitrogen-based chelating ligands, diphosphines and heterobidentate [P, N] ligands

6.1.3.1 Nitrogen-based ligands

A number of highly efficient nitrogen-based chelating ligands were synthesized since 1980s (Figure 6.3).165c Efficient systems (ee up 85%) include chiral pyridine-thiazolidene

(PYTHIA, 146),168c,169 pyridine-bisoxazoline (PYBOX, 147),170 bipyridine-bisoxazoline

(BIPYMOX, 148)171 and pyridine-oxazoline (PYMOX, 149).172

Figure 6.3. Reported N-based ligands for Rh-catalyzed asymmetric hydrosilylation.165c

Neutral or cationic rhodium complexes combined with these ligands catalyze the reduction of aryl alkyl ketones under mild conditions with enantioselectivities from modest to high enantioselectivities.165c In most cases, these ligands have to be used in large excess relative to rhodium to attain high enantioselectivity. 165c The reaction can be carried out under solvent free conditions or performed in solution and proceed below room temperature (down to -78°C). 165c

6.1.3.2 Diphosphine ligands

In the last decade, a number of new phosphorus-based ligands have been synthesized and tested in rhodium catalyzed hydrosilylation (Figure 6.4).165c For most of these ligands developed, high catalytic activities but modest ee (20-58%) was observed.165c However, exceptional high ee (up to 99%) was achieved when rhodium cationic complex combined

with diphosphine ligands like MiniPHOS (159) 173 C2-symmetric P-chiral bis(phosphine)

ferrocene ligand (164),174 and (R)-BINAP (165)175 was used

MeMe

159: (t-Bu)-MiniPHOS

PP

PhPh

PP

O

OO

165: ( R)-BINAP

PPh2PPh2

Figure 6.4 Some reported diphosphines used in Rh-catalyzed asymmetric hydrosilylation

of ketones to alcohols.165c

6.1.3.3 Heterobidentate [P, N] ligands

In 1996, Helmchen163b and Williams176 independently reported enantioselective reduction

of simple ketones in the presence of rhodium(I) complexes of chiral phosphine-oxazoline

ligands (166, Figure 6.5) In the optimized conditions, the hydrosilylation of

acetophenone led to (R)-1-phenylethanol with up to 86% ee A successful modification of

this type of ligand was BPOI (167, Figure 6.5) which was found to induce ee up to 94%

in the reduction of aryl methyl ketones.177

PPh2

167 : BPOI

Figure 6.5. Reported phosphine-oxazoline ligands for enantioselective reduction of simple ketones.163b, 176-177

Subsequently, many active [P, N] ligands were synthesized and tested for their efficiency

in asymmetric hydrosilylation of carbonyl derivatives (Figure 6.6).178 These hybrid [P, N] ligands combine a relatively high enantioselectivity characteristic of nitrogen-based ligands and high activity characteristic of diphosphine ligands, thus allowing selectivities above 90%.165c

Ph2PN

MeNR

67a: R = Ph67b: R = 3-C6H4CF367c: R = 4-C6H4CF3

Fe PPh2

NOR

Ph

168a: R = H168b: R = Ph, ( S,S,S)-DIPOF

NO

of 1,1’-disubstituted iminophosphino ferrocenyl ligands towards asymmetric catalyses are not known One possible explanation is that only a limited number of such catalytic complexes at our disposal are fully characterized There are advantages of the 1,1’-disubstituted unsymmetrical iminophosphino ferrocenyl ligands over their 1,2-counterparts which have been overlooked One example is the simplicity of the synthetic pathway to obtain the chiral ferrocenyl ligands A synthetic pathway of 1,2-disubstituted

ferrocenyl ligand, 67 is shown in Scheme 6.2.41f,180 Compared to 67, the introduction of a

chiral imino group onto the ligand structure for 1,1’-disubstituted ferrocenyl ligand is

easier (Scheme 6.3) The rhodium-catalyzed hydrosilylation reaction has attracted us mainly due to the mild reaction conditions and manipulative simplicity.181 Therefore, we were interested to find out whether the synthesized 1,1’-disubstituted unsymmetrical

iminophosphino ferrocenyl ligands, L1f-j could be applied to the asymmetric

rhodium-catalyzed hydrosilylation reaction

Scheme 6.2 Synthesis of chiral ligand 67 from ferrocene

Scheme 6.3 Synthesis of chiral 1,1’-disubstituted iminophosphino ligands L1f-j

6.2 Objective of this work

We are interested to extend the use of 1,1’-disubstituted iminophosphino ferrocenyl ligands towards asymmetric hydrosilylation of ketones to alcohols Synthesis and crystallographic elucidation of the rhodium complexes chelating to these ligands will be attempted Examination on the synthesized rhodium complexes’ effect on asymmetric hydrosilylation of ketones to alcohols will also be performed

6.3 Results and Discussion

6.3.1 Catalytic evaluation of [P, N] ligands with Rh and Ir precursors towards asymmetric hydrosilylation of ketones to alcohols

Rhodium-catalyzed hydrosilylation of acetophenone was performed employing diphenylsilane together with [RhCl(COD)]2 and ligands L1f-j in the presence of AgBF4,

summarized in Table 6.1 The best catalytic performance was observed when ligand

(R)-L1j was used, giving the corresponding (R)-alcohol product of 83% yield and 60 %ee

(Table 6.1, entry 5) The product yield and ee are better to those obtained with nitrogen

based ligands such as 151 (Yield = 70%, %ee = 42),182 153 (Yield = 76%, %ee = 50),183

154 (Yield = 75%, %ee = 37),184 and 156 (Yield = 32%, %ee = 33).185 The results are also comparable (in terms of yield) and better (in terms of ee) with the reported diphosphine

ligands such as 160 (Yield = 84%, %ee = 37),186 161 (Yield = 87%, %ee = 25)187 and 163

(Yield = 41%, %ee = 20).188 This is however less efficient than the best reported systems

on 1,2-disubstituted ferrocenyl ligands like 47 (Yield = 94%, %ee = 98)35c and 67 (Yield

OH1.1 mol% ligand

0.5 mol% [RhCl(COD)]2AgBF4, THF, -20oC

From Table 6.1, it was observed that the conversion and enantioselectivity are sensitive towards the imine aryl substituent ligand A bulkier substituent would lead to an increase

in the enantioselectivity of the alcohol product obtained (Table 6.1, entries 5 and 6) Lower enantioselectivity and opposite alcohol product configuration were observed when

(S)-L1j was used instead of (R)-L1j although they were enantiomers Lower enantiomeric

excess suggested that the configuration of the imine substituent on the Rh-complexes had some influence on the enantioselectivity The presence of any electron withdrawing or donating properties of the aryl group on the imine substituent resulted in a decrease in the product yield (Table 6.1, entries 2 and 4)

The performance of rhodium catalysts is reported to be dependent on the nature of the silane.160c Usually, bulky diarylsilanes such as diphenylsilane and phenyl(1-naphthyl)silane furnished the best results Similarly, from Table 6.2, the yield and enantiomeric excess also increase generally with the increase in the steric demand of the silane Diaryl silane results in best yield and enantiomeric excess observed compared to the dialkyl silanes (Table 6.2, entry 2) The tertiary silanes however, led to low alcohol products (Table 6.2, entry 4)

Table 6.2: Effect of silanes on asymmetric hydrosilylation of acetophenone catalyzed by

RhI/(R)-L1j

O

Silane+

OH

1.1 mol% (R)-L1j

0.5 mol% [RhCl(COD)]2AgBF4, THF, -20oC

aAverage of 2 runs, after column chromatography bDetermined by HPLC using ChiralCel OD-H

To investigate the influence of the metal, ligand (R)-L1j was screened with various

rhodium and iridium precursors on asymmetric hydrosilylation of acetophenone, as outlined in Table 6.3 Similar to those reported,189 the addition of AgBF4 to the catalytic

(R)-L1j/[RhCl(COD)]2 system raised the yield from 50% to 83%, probably by creating a

vacant site for substrate coordination (through oxygen) Use of alternative precursors such

as [Rh(OH)(COD)]2, [Rh(OMe)(COD)]2 and [IrCl(COD)]2 in the absence of silver salt did not improve on the catalytic system (Yield 33-78%, ee: 17-32%)