Initial Mechanistic Investigation

List of Abbreviations

1.3 Initial Mechanistic Investigation

Chapter 1—The History and Previous Investigations of Palladium-Catalyzed Decarboxylative Asymmetric Allylic

Alkylation of Ketone Enolates Using the PHOX Ligand Architecture

10

or steps related to enantioinduction and bond forming in the mechanism involve only a single palladium species.

Figure 1.1. Nonlinear Effect Study of Palladium-Catalyzed Decarboxylative Allylic Alkylation

Nonlinear Data for Tsuji Allylation _R² _{= 0.988}

0 10 20 30 40 50 60 70

-5 5 15 25 35 45 55 65 75 85 95 105

% ee of ligand

% ee of product

O O BnO

[Pd2(dba)3] THF, 25 °C

O Bn

XX XX

PPh2 N O

i-Pr

The ee of the i-PrPHOX ligand (X-axis) was varied by mixing freshly prepared stock solutions of enantiopure (S) and (R) i-PrPHOX ligand prior to each experiment and the mixture ratio confirmed by chiral HPLC. The product of each reaction was isolated and purified before obtaining ee (Y-axis) via HPLC.

The nonlinear effects studies were supported with traditional reaction kinetics studies. Kinetics studies of the decarboxylative allylic alkylation of allyl enol carbonate XX and allyl !-ketoester XX to form tetralone 23 determined that both the allyl enol

carbonate and !-ketoester reactions were first order in catalyst and zero order in substrate (Figure 1.2 on page 13).

²³

Notably all three substrate classes give allylic alkylation products in similar yields and practically identical ee (Table 1.2 on page 12).

¹⁰

This is

17 16

Chapter 1—The History and Previous Investigations of Palladium-Catalyzed Decarboxylative Asymmetric Allylic

Alkylation of Ketone Enolates Using the PHOX Ligand Architecture

12

strongly suggestive of a single underlying mechanism that must converge at or before the formation of the ketone enolate intermediate (Scheme 1.12 on page 14). Together these results favor a universal mechanism involving a single monomeric PHOX palladium species for each step of both the productive catalytic cycle as well as any unproductive catalyst resting states that might exist.

Table 1.2. Consistent Results Across All Three Substrate Classes¹⁰

O O

O

O O

O

85 88 85 87 95 87

87 92 89 92 79 91

94 85 87 86 99 81

83 87 81 87 94 86

Product yield^a ee^b yield^a ee^b yield^a ee^b

R² O R¹

O O

n

R² OTMS

R¹

n

R² O

R¹

n O

O

OR OR

R² O R¹

n [Pd2(dba)3] (2.5 mol%)

(S)-t-BuPHOX (6.25 mol%)

allyl !-ketoester allyl enol carbonate silyl enol ether THF, 25 °C R³

R³

R³ bis(2-R³)allyl carbonate

+

[a] Isolated yield (%) from reactions with 1 mmol of substrate. [b] Measured by chiral GC or HPLC

9

18

19

20

Figure 1.2. Kinetics Studies Show Zero-Order Dependence on Substrate

y = -0.0046x + 63.656 R² = 0.9941

0 10 20 30 40 50 60 70

0 2000 4000 6000 8000 10000 12000 14000 time (s)

allyl enol carbonates

y = -0.0335x + 698.56 R² = 0.9979

0 100 200 300 400 500 600 700 800

0 5000 10000 15000 20000 25000

time (s) Allyl _-ketoester

[Pd₂(dba)₃] (S)-t-BuPHOX

THF-d8, 0 °C 1,4-dimethoxybenzene

(internal standard)

O O

O O

O O

O [Pd2(dba)3]

(S)-t-BuPHOX THF-d₈, 0 °C 1,4-dimethoxybenzene

(internal standard) O

XX XX

XX XX

!

Reactions were performed in an NMR tube on an 0.05 mm scale and monitored by ¹H NMR. Substrate concentration (Y-axis) is in arbitrary integration units relative to an internal standard consisting of 0.0175 mmol (35 mol%) 1,4-dimethoxybenzene.

23 21

23 22

Chapter 1—The History and Previous Investigations of Palladium-Catalyzed Decarboxylative Asymmetric Allylic

Alkylation of Ketone Enolates Using the PHOX Ligand Architecture

14

With this knowledge in hand we turned to DFT to simulate the reaction of free enolate 24 with a single PHOX palladium "-allyl species 25.

^23,24

Via DFT a traditional outer-sphere allylic alkylation path was identified favoring nucleophilic attack at the "- allyl terminus trans to phosphorous. This is in perfect accordance with previous mechanistic studies for palladium-catalyzed asymmetric allylic alkylation using the PHOX ligand architecture (Figure 1.3 on page 15).

¹⁷

However, DFT simulation also predicted that this outer-sphere attack has practically no energy difference between the two facial approaches of the enolate nucleophile. If so, such an outer-sphere mechanism should result in near racemic allylic alkylation product.

²⁴

The inconsistency of this simulated mechanism versus the experimentally observed results was highly unsatisfactory.

Scheme 1.12. Consistent Product Yield and Enantioinduction Implies a Common Mechanism.

O O O

OTMS

O O

O

+ diallyl carbonate TBAT

O^– +

O

87% ee common asymmetric

discrimination and enantioselective bond

fomration steps Ph3SiF

Me3SiF CO2

CO2 CO2 [Pd2(dba)3]

(S)-t-BuPHOX

[Pd2(dba)3] (S)-t-BuPHOX

[Pd2(dba)3] (S)-t-BuPHOX XX

XX

XX

Ph2P PdN

O +

t-Bu XX

XX

XX

24 25

9 8

4

6

Figure 1.3. Results of DFT Simulation for Outer-Sphere Allylic Alkylation Starting with 24 and 25

!E + E_Solv(THF) / kcal/mol

–30.0 3.4

0.0

PPdN O

O

Ph2P PdN

O

t-Bu O

O^– + Ph2P

PdN O

+

t-Bu XX

XX

–26.6 O

low ee

Ph2P PdN

O

t-Bu +

XX

Literature precedent for the palladium-catalyzed allylic alkylation of hard nucleophiles suggests that they can proceed via an inner-sphere mechanism, whereby nucleophilic attack occurs at the metal center and subsequent bond forming occurs by a reductive elimination process (Scheme 1.13 on page 16).

^4,15

Noting this, we sought to use DFT to investigate an alternate inner-sphere alkylation pathway. DFT placed the energy of the ion-paired free enolate 24 and palladium "-allyl cation 25 as roughly isoenergetic to palladium "-allyl complex 26 with the enolate apically bound (Figure 1.4 on page 17). From complex 26 It was determined that an internal rearrangement involving the isomerization of the allyl ligand from an #-3 to an #-1 binding mode in conjunction with the collapse of the apical enolate ligand into the square plane could form palladium allyl enolate 27. This internal rearrangement was computed to have a kinetic barrier 1.9 kcal/mole smaller then the outer-sphere allylic alkylation process

24 25

9

Chapter 1—The History and Previous Investigations of Palladium-Catalyzed Decarboxylative Asymmetric Allylic

Alkylation of Ketone Enolates Using the PHOX Ligand Architecture

16

making the internal rearrangement the more kinetically favorable of the two processes.

^24,25

Scheme 1.13. Generalized Inner-Sphere Mechanism for Palladium-Catalyzed Allylic Alkylation

X

Nu X^–

Pd(II)

Nu

Pd(II) Pd(II) Nu or Pd(0)

Nu

O- to C-bond rearrangement of the enolate ligand from palladium allyl enolate 27