Transition Metal-Free Direct Hydrogenation of Esters via a Frustrated Lewis Pair

Joshua S. Sapsford, D á niel Cs ó k á s, Roland C. Turnell-Ritson, Liam A. Parkin, Andrew D. Crawford,

Research Article

Transition Metal-Free Direct Hydrogenation of Esters via a

Frustrated Lewis Pair

Imre P á pai,* and Andrew E. Ashley*

Cite This: ACS Catal. 2021, 11, 9143 −9150

Read Online

ACCESS

Metrics & More

Article Recommendations

sı Supporting Information

ABSTRACT: “Frustrated Lewis pairs” (FLPs) continue to exhibit

unique reactivity for the reduction of organic substrates, yet to date, the

catalytic hydrogenation of an ester functionality has not been

demonstrated. Here, we report that iPr SnNTf (1-NTf ; Tf =

SO CF ) is a more potent Lewis acid than the previously studied

iPr SnOTf; in an FLP with 2,4,6-collidine/2,6-lutidine (col/lut), this

translates to faster H activation and the catalytic hydrogenolysis of an

ester bond by a main-group compound, furnishing alcohol and ether

(

minor) products. The reaction outcome is sensitive to the steric and

electronic properties of the substrate; CF CO Et and simple formates

(

HCO Me and HCO Et) are catalytically reduced, whereas related esters

2 2

CF CO nBu and CH CO Et show only stoichiometric reactivity. A

computational case study on the hydrogenation of CF CO Et and

CH CO Et reveals that both share a common mechanistic pathway;

however, key diﬀerences in the energies of a Sn-acetal intermediate and transition states emerge, favoring CF CO Et reduction. The

−

alcohol products reversibly inhibit 1-NTf /lut via formation of resting-state species 1-OR/[1·(1-OR)] [NTf ] ; however, the extra

energy required to regenerate 1-NTf /lut exacerbates the unfavorable reduction energy proﬁle for CH CO Et, ultimately preventing

turnover. These ﬁndings will assist the design of future main-group catalysts for ester hydrogenation, with improved performance.

KEYWORDS: hydrogenation, frustrated Lewis pair, main group catalysis, tin, ester

ydrogenations of unsaturated bonds in organic mole-

cules are typically catalyzed by precious metals, but their

increasing cost and scarcity have motivated the design of

products, which sequester Ar B in preference to the ester

carbonyl. FLP-type BAr -catalyzed reductions of esters (to

21,22

23,24

aldehydes

or alkanes)

have been reported, though

catalysts incorporating more abundant elements. “Frustrated

these require more reactive hydrosilanes as reductants. In and

Ga halides can also promote ester hydrosilylations, proceeding

Lewis pairs” (FLPs), which comprise hindered Lewis acids

25,26

via a radical mechanism

(Scheme 1).

(

LAs) and bases (LBs), can cooperatively activate H to H

−

We previously reported that iPr SnOTf (1-OTf; Tf =

and H equivalents that are subsequently delivered across an

−6

O SCF ), a surrogate for [1] , catalyzes reductive aminations

unsaturated bond.

Ar B-catalyzed reductions of functional

7,28

−10

and the hydrogenation of imines and carbonyls.

^H2

groups such as imines,

ketones

alkenes,

alkynes,

and

For successful

3−15

16,17

cleavage was found to be the rate-determining step and that

are now well-established.

−

changing the anion partner of the [1] fragment from TfO to

ketone hydrogenation, the product alcohols typically associate

−

weakly coordinating {Al[OC(CF ) ] } aﬀects the kinetics and

to Ar B to form strongly Brønsted acidic adducts [e.g., pK

3 3 4

thermodynamics of H cleavage. We proposed that using ₋a

H O·B(C F ) = 8.4 in MeCN]; this issue can be

6 5 3

larger anion with similar coordinating properties to TfO

circumvented by employing ether solvents as weak LB

3−15

might increase the reactivity of the [1] fragment, through

partners

and/or tuning the steric and electronic proper-

internal Sn-anion bond frustration. Here, we report the novel

ties of Ar B species to disfavor strong association with

19,20

hydroxylic species.

Despite the prevalence of Ar B species

Received: April 29, 2021

Revised: June 15, 2021

Published: July 9, 2021

that catalyze ketone hydrogenations, FLP-catalyzed direct

hydrogenation of more inert carbonyl functionalities such as

esters have not been reported previously. Resonance within the

CO R group diminishes the electrophilicity of the central C

−

atom from H attack, even when Ar B coordination polarizes

the CO bond. Furthermore, reduction would yield alcohol

2021 American Chemical Society

143

ACS Catal. 2021, 11, 9143−9150

ACS Catalysis

(AN: 80.9 vs 68.0). In addition, the calculated

Research Article

Scheme 1. Overview of the Stepwise Reduction of Esters by E−H (E = H and R Si) Reductants

−

LA 1-NTf which is a superior catalyst for imine and carbonyl

hydrogenation and also promotes the ﬁrst examples of main

group-catalyzed ester hydrogenation.

hydride aﬃnities [ΔG(H ) for the reaction 1-X + H → 1-H +

−

X ] are −43.9 and −34.8 kcal/mol, for X = NTf and OTf,

respectively. Collectively, these data reveal a substantial

increase in Lewis acidity and hydridophilicity for 1-NTf versus

-NTf₂was prepared on a multigram scale by proto-

dealkylation of iPr Sn with HNTf in 96% yield (Scheme 2).

1-OTf.

Admission of H (4 bar) to a solution of 1-NTf and 2,4,6-

collidine (col) in 1,2-dichlorobenzene (DCB) resulted in the

Scheme 2. Synthesis of 1-NTf and H Cleavage by 1-NTf +

col/lut (2,4,6-Collidine/2,6-Lutidine) Lewis Pairs

appearance of resonances consistent with 1-H (Sn−H: δ =

.09 ppm; δ

= −47 ppm) and [col-H] (N−H: δ =

19Sn

3.48 ppm) in the H NMR spectra after 15 h at RT.

Notably, H activation by 1-OTf/col requires elevated pressure

of H (10 bar) to react at a comparable rate, demonstrating a

signiﬁcantly enhanced reactivity of 1-NTf toward H₂

heterolysis.

Density functional theory (DFT) calculations (see Figure

) reveal that the ground-state speciation in DCB solution

‑

favors the separated ion pair [1·(col)] [NTf ] versus the 1-

NTf /col FLP (by 2.6 kcal/mol), whereas the situation is

reversed in the case of 1-OTf/col (the FLP being more stable

by 6.8 kcal/mol), reﬂecting the stronger Lewis acidity of the

^t^rⁱ^ﬂⁱ^mⁱ^d^e^c^o^m^p^o^uⁿ^d^.^F^u^r^t^h^e^r^m^o^r^e^,^c^o^m^p^u^t^a^tⁱ^oⁿ^s^s^u^p^p^o^r^t^H2

heterolysis observations, predicting the reaction to be more

Single-crystal X-ray diﬀraction studies reveal that 1-NTf₂

aggregates in the solid state (Figure 1); NTf moieties bridge

[

1] units via Sn−O linkers exclusively, which is consistent

with the structures of related R Sn[N(SO Me) ] (R = Cy and

thermodynamically favorable for 1-NTf + col cf. 1-OTf + col

from their respective ground states (−4.7 vs 1.7 kcal/mol) and

Ph) compounds.

with a slightly lower activation barrier for the former (19.8 vs

In CDCl solution, 1-NTf exhibits a more positive ¹¹⁹Sn-

20.2 kcal/mol at RT).

{¹

respectively) and a higher Gutmann−Beckett acceptor

H} chemical shift than 1-OTf (δ/ppm = 214 vs 168,

We subsequently canvassed the activity of 1-NTf versus 1-

OTf toward CN and CO hydrogenation, using archetypal

2,33

number

substrates in FLP chemistry (Table 1). 1-NTf was found to be

Figure 1. Oak Ridge thermal ellipsoid plot (ORTEP) of a single iPr Sn-NTf (1-NTf ) fragment (top) and a linear chain of four 1-NTf units

(

bottom). Ellipsoids are shown at 50% probability. Atoms are colored, C = black, F = green, N = blue, O = red, S = yellow, and Sn = pink. H atoms

are omitted for clarity.

144

ACS Catal. 2021, 11, 9143−9150

ACS Catalysis

Research Article

conversion, in marked contrast with the previously reported

27,28

performance of 1-OTf/col.

All our previous attempts to hydrogenate ester function-

alities using 1-OTf/LB LPs were unsuccessful, which we

attribute to an insuﬃcient Lewis acidity at the Sn center;

accordingly, we hypothesized that the enhanced reactivity of 1-

NTf might prove to be more fruitful, especially in conjunction

with more electrophilic ester substrates. Gratifyingly, admis-

sion of H (10 bar) to CF CO Et and 1-NTf /lut (10 mol %,

DCB) resulted in the appearance of resonances attributable to

alcohol products CF CH OH (δ CH = 3.81 ppm; δ

19F

−

76.8 ppm) and EtOH (δ₁_HCH₂= 3.58 ppm) in the

respective NMR spectra, after heating to 150 °C. A small

amount of Et O (∼6%) was also formed, as conﬁrmed by

comparison with an authentic sample in DCB (δ CH = 3.31

ppm); a correspondingly small amount of Et O was also

produced when EtOH and 10 mol % 1-NTf were heated

(

DCB, 150 °C), demonstrating that 1-NTf promotes the

observed condensation of EtOH to Et O (and H O) during

ester hydrogenations. After 72 h, a respectable 80%

conversion of the ester was achieved which represents a

turnover number of 8 and 16 with respect to ester conversion

Figure 2. H activation with 1-NTf /col. Relative stabilities (in kcal/

mol; at RT) are shown in parentheses. For comparison, the barrier

and the reaction free energy for the analogous process with 1-OTf/col

are 20.2 and 1.7 kcal/mol, respectively. Selected bond distances for

the computed transition state TS_H_Hare given in Å.

and H molecules cleaved, respectively (Table 2; entry 1).

Table 2. Direct Hydrogenation of Esters Mediated by the 1-

Table 1. Conditions and Conversions for the

Hydrogenations of Imines and Acetone Catalyzed by 1-NTf₂

NTf + lut Lewis Pair

entry

R′

PH bar T °C

t h conv. %

entry

R′

nBu

t, h

conv. %

TON H₂

NtBu

NPh

120

180

120

>99

CF₃

2.2

7.4

7.6

0.8

4-Br-Ph

iPr

b d

Conversions measured using in situ 1_H_N_M_R_s_p_e_c_t_r_o_s_c_o_p_y_,

calculated as the ratio of peak integrals assigned to alcohol and

ether products versus reactant ester. CF CH O ethers were not

formed. Conversions measured by H NMR spectroscopy in situ by

the ratio of the integrals of ester and product peaks versus PhMe

internal standard, due to the volatility of Me O. TON de ﬁ ned as H

0 mol % lut added. 10 mol % col added. Conversions measured

using in situ H NMR spectroscopy, calculated as the ratio of the

integrals of peaks assigned to the product and reactant (see the

Supporting Information for details). Reaction prepared on the open

bench using undried commercial grade solvent before pressurization

(see the Supporting Information for details).

molecules cleaved per 1-NTf . See the Supporting Information for full

details.

a superior catalyst than 1-OTf for the hydrogenation of imines,

with lower H pressures (4 vs 10 bar) required for reactions to

Curiously, using the bulkier substrate CF CO nBu resulted

reach >90% conversion, under otherwise analogous conditions

in only stoichiometric conversion to CF CH OH and nBuOH

(

Table 1, entries 1−4).

As seen previously for 1-OTf, N-aryl imines were

insuﬃciently basic to cleave H with 1-NTf and required

(Table 2; entry 2). Simple formates HCO R (R = Me and

Et) were more successful with turnover numbers 7.6 and 7.4

with respect to H molecules cleaved, respectively (Table 2;

,6-lutidine (lut) as an auxiliary LB; nevertheless, the former

entries 3 and 4). MeCO Et, however, was hydrogenated in

catalyst required longer reaction times for these substrates

sub-stoichiometric amounts, whereas iPrCO Et was not

reduced at all, presumably due to steric hindrance.

(

≥24 h; compare entries 3 and 4). Additionally, 1-NTf /col

was able to hydrogenate Me CO (entry 5) almost three times

To gain mechanistic information on the reduction and

account for reactivity diﬀerences, either CF CO Et or

faster than 1-OTf/col (78%, 96 h, 10 bar) and without the

need for periodic repressurization of the reaction vessel;

MeCO Et was combined with authentic 1-H (2.2 equiv, 10

similar conversion at 4 bar H could also be achieved (entry 6),

albeit at a higher temperature and longer reaction time.

mol % excess) in DCB at RT; in neither case was a reaction

observed. Upon addition of 10 mol % 1-NTf ; however,

Conversely, while Me CO hydrogenations using 1-NTf /col in

undried solvent were still catalytic (entry 7), the presence of

H O clearly has a deleterious eﬀect on the reaction rate and

hydrostannation of CF CO Et to 1-OCH CF and 1-OEt

proceeded smoothly (44%, 12 h), gentle heating of which

promoted increased conversion (65%, 60 °C, 12 h).

145

ACS Catal. 2021, 11, 9143−9150

ACS Catalysis

Research Article

Figure 3. Computed Gibbs free-energy proﬁle (ΔG in kcal/mol) of the proposed mechanism of 1-NTf -catalyzed reduction of RCO Et (R = CH ,

red; R = CF , blue). The reference state is normalized to 3 1-NTf + 2 lut + 2 H + RCO Et.

Scheme 3. Stoichiometric Hydrostannations and Proposed Reduction Steps of RCO Et (R = CF and CH ) Mediated by 1-H

and 1-NTf₂

Conversely, MeCO Et required heating to 80 °C for reduction

The reaction commences with H activation by the lut/1-

initiation (21%, 12 h), progressing to only 31% conversion at

NTf

50 °C (TS , Figure 3) to produce 1-H + [Lut-H] +

NTf‑Lut

FLP, proceeding with a barrier height of 23.8 kcal/mol at

20 °C after 12 h. Curiously in these reactions, only the

−

[NTf ] . For the ensuing hydride transfers, the reductant was

Sn{ H} NMR resonance for 1-OCH CF could be

−

modeled as [1-H·NTf

] (ΔG: 1-H + [NTf

] = 11.3 kcal/

observed, as a sharp peak at δ

= 44 ppm.

Sn{ H}

mol) since an anion needs to be in close proximity to 1-H in

order to sequester the incipient [1]⁺fragment upon H⁻

departure, thereby avoiding formation of a prohibitively high-

In order to garner insights into the stark reactivity

diﬀerences between the reduction of CF CO Et and

CH CO Et, DFT calculations (see the Supporting Information

energy “free” stannylium cation. It was discovered that for

for full details) were performed to predict the most feasible

pathways mediated by 1-NTf , which is clearly necessary to

both ester substrates, the lowest energy reduction scheme

involves essentially the same mechanistic steps (Scheme 3): (i)

initiate reduction by 1-H.

pre-activation of the ester through complexation of 1-NTf to

−

Gibbs free energy changes are shown in parentheses (kcal/

form [1·(RCO Et)] [NTf ] ; (ii) subsequent hydride attack

−

mol; T = 150 °C; R = CF /CH , respectively).

on CO by [1-H·NTf ] to form a Sn-acetal, via transition

146

ACS Catal. 2021, 11, 9143−9150

ACS Catalysis

Research Article

Scheme 4. Interconversion of Species Involved in Liberation of Alcohol Products and Regeneration of 1-NTf Catalyst during

Direct Hydrogenations of RCO Et (R = CF and CH )

state TS ; (iii) elimination of 1-OEt via TS_a_l_dto produce

subsequent dissociation of the latter into 1-NTf + RCH OH

2 2

(see Scheme 4) thereby regenerates the LA catalyst for

red‑1

aldehyde RCHO; (iv) activation of RCHO by 1-NTf to form

−

[

(

1·(RCHO)] [NTf ] ; and (v) reduction of [1·

RCHO)] [NTf ] by [1-H·NTf₂] yields the Sn-alkoxides

turnover in ensuing H cleavage/CO activation steps.

−

DFT calculations (at 25 °C) reveal that CH CH OH forms

-OCH R and 1-OEt, via TS

a weak dative Sn−O bond with 1-NTf (ΔG = 1.7 kcal/mol),

red‑2

Despite the similarity of the elementary steps for both esters,

key energetic diﬀerences emerge. For the more electron-rich

CH CO Et, both TS and TS are noticeably lower in

but this adduct preferentially dissociates into [1·

−

(CH CH OH)] + [NTf₂] (ΔG = −1.0 kcal/mol),

suggesting all Sn-containing species are in equilibrium; in

contrast, analogous states for CF CH OH + 1-NTf are

= 4.2 kcal/mol; ΔG = 8.1 kcal/

red‑1

red‑2

energy than that for CF CO Et, mirroring the more favorable

free energies of formation of activated adducts [1·(RCO Et/

signiﬁcantly less stable (ΔG

−

RCHO)] [NTf ] when R = CH versus CF (Scheme 3), due

mol).

to the greater Lewis basicity of CH CO Et/CH CHO.

Similarly, the reaction barrier for TS_a_l_dis also lower when R

Gibbs free-energy changes are shown in parentheses (kcal/

mol; T = 150 °C; R = CF /CH , respectively). Underlined

values show the largest energy input for liberation of alcohol

product and active catalyst 1-NTf , from the most stable

species in the catalytic cycle of the respective ester substrate.

In the presence of lut, however, the coordinated RCH OH is

CH ; however, a crucial divergence between the stability of

the preceding Sn-acetal intermediates is evident, whereas the

formation of ₃Sn-acetal is favorable (−3.9 kcal/mol), and

_t_h_a_t_o_fC Sn-acetal is considerably endergonic (+5.9 kcal/

deprotonated in both cases, resulting in 1-OCH R + [Lut-H]

mol). This impacts upon the reaction barrier for the collapse of

−

+ [NTf ] species, which overall are predicted to be highly

Sn-acetal with respect to the preceding reactant state (TS_a_l_d

favored thermodynamically over the RCH OH + 1-NTf + lut

21.4 kcal/mol), eﬀectively increasing it to 27.3 kcal/mol

states for both alcohols (ΔG = −8.6 and −8.7 kcal/mol for R

(

since the CH CO Et/1-H reactant state is more stable) such

= CH and CF , respectively). Initially, it might appear

that it becomes the highest overall barrier on the CH CO Et

surprising that two alcohols with such diﬀering donor

properties could exert almost identical inhibitory eﬀects on

^r^e^d^u^c^tⁱ^oⁿ^p^a^t^h^w^a^y^,^e^x^c^e^e^dⁱⁿ^g^t^h^a^t^f^o^r^H2 activation.

Accordingly, it may be envisaged that faster H activation

catalysis, via sequestration of the active catalyst 1-NTf . It

would result in the complete conversion of 1-NTf into 1-H,

should be noted, however, that two factors should be taken

and the reaction would be stranded in the state of 1-H + [Lut-

into account here: ﬁrst, the acidity of the alcohol, and second,

−

H] + [NTf ] , thus depleting 1-NTf available for the critical

the strength of interaction between RCH OH and [1] .

activation of the ester.

Notably, 1-H is experimentally observed by H NMR after

limited conversion of CH CO Et (δSn-H = 5.10 ppm; see the

Clearly, the electron-withdrawing group in CF CH OH

weakens alcohol binding to 1-NTf , yet in this case, this is

practically counterbalanced by the lower pK of the adduct,

Supporting Information ), which corroborates this mechanistic

scenario.

rendering deprotonation more favorable. These results are in

line with the experiment whereby 1-OCH R are observed as

Conversely, although the precise rate-determining step for

CF CO Et reduction cannot be identiﬁed due to the

sharp resonances in the Sn{ H} NMR spectra when lut/1-

NTf and RCH OH are combined (1:1:1, DCB, RT; δ: 34 and

uncertainty of DFT methods (TS

, TS , and TS_a_l_d

44 ppm for R = CH and CF , respectively); upon admission

NTf‑Lut

red‑1

3 3

within a 0.9 kcal/mol range), it may be proposed that these

elementary steps proceed in parallel which is pivotal for the

reduction reaction to progress successfully, as experimentally

ascertained.

of H to these solutions (see the Supporting Information), no

H cleavage was apparent at RT (even after 7 days), although

heating to 150 °C did produce 1-H. This demonstrates that

the inhibitory eﬀect these alcohols exert on the reactivity of 1-

It has previously been observed that O donors such as

alcohols can act as inhibitors of FLP-type reactions using BAr₃

and 1-OTf LA catalysts, via strong adduct formation and/or

subsequent deprotonation by the LB partner, which forms

NTf , is reversible at the temperature of ester reduction. It was

also recognized that another route to 1-NTf sequestration

might arise from coordination of the catalyst to the Lewis basic

Sn-alkoxides 1-OCH R, and computational results reveal that

kinetically incompetent alkoxide species. For completion of

the catalytic cycle in these ester hydrogenations, it is essential

to liberate the free alcohol products from 1-OCH R (R = CF /

such species exist in their dissociated form [1·(1-OCH R)] +

−

[NTf ] (Scheme 4; 150 °C). Adduct formation for R = CF is

predicted to be thermodynamically unfavorable, whereas the

CH ) species. Conceptually, this would occur via protonation

opposite is true for the more basic R = CH derivative;

−

with [Lut-H] [NTf ] to form lut + [1-O(H)CH R] [NTf ] ,

accordingly, 1-OCH CF and [1·(1-OCH CH )] [NTf ]

2 3 2 3 2

147

ACS Catal. 2021, 11, 9143−9150

ACS Catalysis

https://pubs.acs.org/doi/10.1021/acscatal.1c01940.

Research Article

represent the lowest energy Sn-containing species in their

in the reduction steps, ultimately leading to poor conversion

for CH CO Et. Overall, the computational trends compare

respective catalytic cycles for RCO Et hydrogenation. This has

important implications for the reduction of CH CO Et; as 1-

very well with the experimental observations. The success of a

given substrate is proposed to depend on the electrophilicity of

the ester and the basicity/volatility of the product alcohols.

−

NTf is converted into [1·(1-OCH CH )] [NTf ] in the

early-phase of the reaction, the adduct becomes a reservoir for

the most Lewis acidic [1] fragment, in active 1-NTf . As

While 1-NTf is more Lewis acidic than 1-OTf (as evidenced

shown in Scheme 4, full closure of the catalytic ester

by the superior activity of the former for catalytic hydro-

genations of imines and acetone under milder conditions),

hydrogenation loop, to produce RCH OH and regenerate 1-

NTf from the most stable Sn compounds, requires an energy

which is amenable toward faster H activation and more potent

−

input of 7.4 and 11.9 kcal/mol for R = CH and CF ,

substrate activation to promote H transfer from 1-H, this

respectively. It follows that in ensuing cycles, the reaction

barriers increase by these values as the elementary steps of the

reduction remain unchanged, which requires the involvement

comes at the expense of being more susceptible to product

inhibition by hydroxylic products. We anticipate that further

steric or electronic modulation of [1] and/or its anion is likely

of 1-NTf LA. This explains the necessity of the relatively high

to uncover further, novel reactivity.

temperatures and prolonged reaction times needed for the

successful reduction of the CF CO Et substrate, and why for

ASSOCIATED CONTENT

* Supporting Information

The Supporting Information is available free of charge at

■

CH CO Et the reaction stalls prior to a full turnover of ester

sı

(

due to a prohibitively high eﬀective reaction barrier of 39.2

kcal/mol for the rate-determining step), despite otherwise

identical reaction conditions. Finally, for the more successful

hydrogenation of HCO Me and HCO Et, it is plausible that

Crystallographic data of 1-NTf , CCDC 2072866 (CIF)

Full synthetic and computational details and NMR

spectra of the reported reactions (PDF )

the more volatile product CH OH has an increased tendency

to partition in the vapor phase in preference to solution,

thereby reducing its concentration and shifting the equilibrium

of alcohol coordination to 1-NTf toward the free species

required for ester reduction.

AUTHOR INFORMATION

Corresponding Authors

■

From the DFT results, it may be envisaged that the use of

slightly weaker LBs than lut (2,6-dimethylpyridine) might

improve the conversion of CH CO Et, both by rendering the

Imre Pápai − Research Center for Natural Sciences, Institute of

Organic Chemistry, Budapest H-1117, Hungary;

orcid.org/0000-0002-4978-0365; Email: papai.imre@

ttk.mta.hu

Andrew E. Ashley − Molecular Sciences Research Hub,

Imperial College, London W12 0BZ, U.K.; orcid.org/

protonative release of EtOH from 1-OEt more facile and by

making H heterolysis more reversible, thus increasing the

concentration of free 1-NTf . To this end, we employed the

structurally related 2-methyl-6-chloropyridine or 2,6-dichlor-

opyridine which retain a similar steric proﬁle to lut (both

0000-0002-5575-7866 ; Email: a.ashley@imperial.ac.uk

formed FLPs with 1-NTf ), yet with diminished basicity due to

the electron-withdrawing chloro substituents; unfortunately, in

Authors

Joshua S. Sapsford − Molecular Sciences Research Hub,

Imperial College, London W12 0BZ, U.K.

each case, negligible CH CO Et hydrogenation occurred.

Computational calculations reveal that H cleavage becomes

Dániel Csókás − Research Center for Natural Sciences,

Institute of Organic Chemistry, Budapest H-1117, Hungary;

orcid.org/0000-0002-4150-477X

Roland C. Turnell-Ritson − Molecular Sciences Research

Hub, Imperial College, London W12 0BZ, U.K.

Liam A. Parkin − Molecular Sciences Research Hub, Imperial

College, London W12 0BZ, U.K.

notably endergonic in these cases (+5.3 and +16.2 kcal/mol at

50 °C, respectively) and attempts to locate TS structures for

this process were unsuccessful; instead spontaneous H

elimination from the proposed 1-H + [LB-H] [NTf ]

−

product structures was observed, indicating that H splitting

is also kinetically disfavored.

In summary, we have reported the ﬁrst examples of FLP-

catalyzed direct hydrogenations of esters. Although the scope

of compatible substrates is limited, detailed mechanistic

analysis shows that variation of ester acyl substituents can

dramatically alter the reactivity and overall thermodynamic

proﬁle of the reduction. A reaction pathway accounting for the

origin of the poor activity of CH CO Et was found, wherein

Andrew D. Crawford − Molecular Sciences Research Hub,

Imperial College, London W12 0BZ, U.K.

Complete contact information is available at:

https://pubs.acs.org/10.1021/acscatal.1c01940

Author Contributions

The manuscript was written through contributions of all

authors.

the collapse of Sn-acetal intermediate has a decisive role in the

entire catalytic cycle. The reaction barriers involved in

CF CO Et reduction, however, are comparable and suﬃciently

Funding

We wish to thank the EPSRC for PhD (J.S.S.) and grant

funding (EP/N026004) and the Royal Society for a University

Research Fellowship (A.E.A.; UF/160395 and RGF\EA

180179). This work was also supported by the NKFIH (K-

15660). Computational resources provided by the NIIF

Supercomputer Center (Project 85708 kataproc) are acknowl-

edged.

low for Sn-acetal collapse to be competent, thus preventing

catalyst 1-NTf from complete consumption in the early phase

of the reaction and permitting subsequent ester activation. The

activity of the lut/1-NTf system has also been shown to be

inhibited by alcohol products due to the formation of quite

stable 1-alkoxide species, which is exacerbated by the

−

sequestration of 1-NTf in [1·(1-alkoxide)] [NTf ] adducts

for more Lewis basic alcohols. Although this is reversible, the

additional energy input augments unfavorable reaction barriers

Notes

The authors declare no competing ﬁnancial interest.

148

ACS Catal. 2021, 11, 9143−9150

ACS Catalysis

Research Article

19) (a) Fasano, V.; Ingleson, M. Recent Advances in Water-

ACKNOWLEDGMENTS

■

Tolerance in Frustrated Lewis Pair Chemistry. Synthesis 2018, 50,

We would like to thank Pete Haycock for providing support

and assistance during the collection of NMR data.

1783−1795. (b) Gyömöre, A.; Bakos, M.; Földes, T.; Pápai, I.;

Domján, A.; Soós, T. Moisture-Tolerant Frustrated Lewis Pair

Catalyst for Hydrogenation of Aldehydes and Ketones. ACS Catal.

2015, 5, 5366−5372.

REFERENCES

■

(20) Carden, J. L.; Dasgupta, A.; Melen, R. L. Halogenated

Triarylboranes: Synthesis, Properties and Applications in Catalysis.

1) Welch, G. C.; Juan, R. R. S.; Masuda, J. D.; Stephan, D. W.

Reversible, Metal-Free Hydrogen Activation. Science 2006, 314,

Chem. Soc. Rev. 2020, 49, 1706−1725.

(

124−1126.

(21) Parks, D. J.; Piers, W. E. Tris(pentafluorophenyl)boron-

2) Chen, D.; Klankermayer, J. Frustrated Lewis Pairs: From

Catalyzed Hydrosilation of Aromatic Aldehydes, Ketones, and Esters.

Dihydrogen Activation to Asymmetric Catalysis. In Frustrated Lewis

Pairs II: Expanding the Scope; Erker, G., Stephan, D. W., Eds.; Topics

in Current Chemistry; Springer-Verlag: Berlin, 2013; Vol. 334, pp 1−

J. Am. Chem. Soc. 1996, 118, 9440−9441.

(22) Parks, D. J.; Blackwell, J. M.; Piers, W. E. Studies on the

Mechanism of B(C F ) -Catalyzed Hydrosilation of Carbonyl

26.

6 5 3

Functions. J. Org. Chem. 2000, 65, 3090−3098.

3) Stephan, D. W. Frustrated Lewis pairs: From Concept to

Catalysis. Acc. Chem. Res. 2015, 48 , 306 −316.

4) (a) Stephan, D. W.; Erker, G. Frustrated Lewis Pair Chemistry:

Development and Perspectives. Angew. Chem., Int. Ed. 2015 , 54 ,

400 −6441. (b) Jupp, A. R.; Stephan, D. W. New Directions for

Frustrated Lewis Pair Chemistry. Trends Chem. 2019 , 1 , 35 −48.

c) Stephan, D. W. The Broadening Reach of Frustrated Lewis Pair

Chemistry. Science 2016 , 354 , aaf7229.

5) Stephan, D. W. Frustrated Lewis Pairs. J. Am. Chem. Soc. 2015,

37, 10018−10032.

6) Scott, D. J.; Fuchter, M. J.; Ashley, A. E. Designing Effective

Frustrated Lewis Pair ” Hydrogenation Catalysts. Chem. Soc. Rev.

017, 46, 5689−5700.

7) Chase, P. A.; Welch, G. C.; Jurca, T.; Stephan, D. W. Metal-Free

Catalytic Hydrogenation. Angew. Chem., Int. Ed. 2007, 46, 8050−

053.

8) Spies, P.; Schwendemann, S.; Lange, S.; Kehr, G.; Fröhlich, R.;

(23) Gevorgyan, V.; Rubin, M.; Liu, J.-X.; Yamamoto, Y. A Direct

Reduction of Aliphatic Aldehyde, Acyl Chloride, Ester, and

Carboxylic Functions into a Methyl Group. J. Org. Chem. 2001, 66,

672−1675.

24) Bajracharya, G. B.; Nogami, T.; Jin, T.; Matsuda, K.;

(25) Sakai, N.; Moriya, T.; Konakahara, T. An Efficient One-Pot

Gevorgyan, V.; Yamamoto, Y. Reduction of Carbonyl Function to a

Methyl Group. Synthesis 2004, 308−311.

Synthesis of Unsymmetrical Ethers: A Directly Reductive Deoxyge-

(

nation of Esters Using an InBr /Et SiH Catalytic System. J. Org.

3 3

“

Chem. 2007, 72, 5920−5922.

(26) Biermann, U.; Metzger, J. O. Synthesis of Ethers by GaBr ₃ -

Catalyzed Reduction of Carboxylic Acid Esters and Lactones by

Siloxanes. ChemSusChem 2014, 7, 644−649.

(

(27) Scott, D. J.; Phillips, N. A.; Sapsford, J. S.; Deacy, A. C.;

Fuchter, M. J.; Ashley, A. E. Versatile Catalytic Hydrogenation Using

A Simple Tin(IV) Lewis Acid. Angew. Chem., Int. Ed. 2016, 55,

Erker, G. Metal-Free Catalytic Hydrogenation of Enamines, Imines,

and Conjugated Phosphinoalkenylboranes. Angew. Chem., Int. Ed.

(

4738−14742.

9) Chase, P. A.; Jurca, T.; Stephan, D. W. Lewis Acid-Catalyzed

008, 47, 7543−7546.

Hydrogenation: B(C F ) -Mediated Reduction of Imines and Nitriles

28) Sapsford, J. S.; Scott, D. J.; Allcock, N. J.; Fuchter, M. J.; Tighe,

C. J.; Ashley, A. E. Direct Reductive Amination of Carbonyl

Compounds Catalyzed by a Moisture Tolerant Tin(IV) Lewis Acid.

Adv. Synth. Catal. 2018, 360, 1066−1071.

5 3

with H . Chem. Commun. 2008, 1701 −1703.

10) Chen, D.; Klankermayer, J. Metal-Free Catalytic Hydro-

(29) Sapsford, J. S.; Csókás, D.; Scott, D. J.; Turnell-Ritson, R. C.;

genation of Imines with Tris(perfluorophenyl)borane. Chem.

Commun. 2008, 2130−2131.

Piascik, A. D.; Pápai, I.; Ashley, A. E. Establishing the Role of Triflate

Anions in H Activation by a Cationic Triorganotin(IV) Lewis Acid.

(

11) Greb, L.; Ona-Burgos, P.; Schirmer, B.; Grimme, S.; Stephan,

ACS Catal. 2020, 10, 7573−7583.

D. W.; Paradies, J. Metal-Free Catalytic Olefin Hydrogenation: Low-

(30) 1-NTf is only sparingly soluble in common non-donor

Temperature H Activation by Frustrated Lewis Pairs. Angew. Chem.,

solvents, although suﬃciently so in CDCl to obtain NMR spectral

information. Even in reactivity studies, which 1,2-dichlorobenzene

Int. Ed. 2012, 51, 10164−10168.

12) Chernichenko, K.; Madarász, A.; Pápai, I.; Nieger, M.; Leskelä,

(

DCB) is used it only appreciably dissolves when suitably strong

donors (e.g. lut, col) are present in solution.

31) Blaschette, A.; Lange, I.; Krahl, J.; Koch, D.; Jones, P. G.

Polysulfonylamine. Organomet. Chem. 1994 , 467 , 169 −180.

32) Mayer, U.; Gutmann, V.; Gerger, W. The Acceptor Number ? A

M.; Repo, T. A Frustrated-Lewis-pair Approach to Catalytic

Reduction of Alkynes to Cis-alkenes. Nat. Chem. 2013 , 5 , 718 −723.

13) Scott, D. J.; Fuchter, M. J.; Ashley, A. E. Nonmetal Catalyzed

Hydrogenation of Carbonyl Compounds. J. Am. Chem. Soc. 2014, 136,

5813−15816.

14) Mahdi, T.; Stephan, D. W. Enabling Catalytic Ketone

Hydrogenation by Frustrated Lewis Pairs. J. Am. Chem. Soc. 2014,

36, 15809−15812.

15) Mahdi, T.; Stephan, D. W. Facile Protocol for Catalytic

Quantitative Empirical Parameter for the Electrophilic Properties of

Solvents. Monatsh. Chem. 1975, 106, 1235−1257.

(33) Beckett, M. A.; Strickland, G. C.; Holland, J. R.; Sukumar

Varma, K. A convenient n.m.r. method for the measurement of Lewis

acidity at boron centres: correlation of reaction rates of Lewis acid

initiated epoxide polymerizations with Lewis acidity. Polymer 1996,

Frustrated Lewis Pair Hydrogenation and Reductive Deoxygenation

of Ketones and Aldehydes. Angew. Chem., Int. Ed. 2015, 54, 8511−

(

7, 4629−4631.

16) Sitte, N. A.; Bursch, M.; Grimme, S.; Paradies, J. Frustrated

514.

34) Gibbs free energies were computed by combining ωB97X-D/

Def2TZVPP electronic energies with thermal and solvation

contributions obtained at the ωB97X-D/Def2SVP level of DFT.

For further computational details, see the Supporting Information.

Lewis Pair Catalyzed Hydrogenation of Amides: Halides as Active

Lewis Base in the Metal-Free Hydrogen Activation. J. Am. Chem. Soc.

019, 141, 159−162.

17) Lam, J.; Szkop, K. M.; Mosaferi, E.; Stephan, D. W. FLP

Catalysis: Main Group Hydrogenations of Organic Unsaturated

Substrates. Chem. Soc. Rev. 2019, 48, 3592−3612.

18) Bergquist, C.; Bridgewater, B. M.; Harlan, C. J.; Norton, J. R.;

Friesner, R. A.; Parkin, G. Aqua, Alcohol, and Acetonitrile Adducts of

Tris(perfluorophenyl)borane: Evaluation of Brønsted Acidity and

Ligand Lability with Experimental and Computational Methods. J.

Am. Chem. Soc. 2000, 122, 10581−10590.

NTf versus Me Si-OTf: see Mathieu, B.; Ghosez, L. Trimethylsilyl

−

(

35) Presumably this arises from the larger size of NTf cf. OTf ,

−

which creates greater “internal frustration” between [1] and NTf ,

resulting in more stannylium (e.g. [1] ) character. A similar argument

was invoked to rationalize the superior catalytic activity of Me

bis(trifluoromethanesulfonyl)imide as a Tolerant and Environ-

mentally Benign Lewis Acid Catalyst of the Diels −Alder Reaction.

Tetrahedron 2002, 58, 8219−8226.

(

Si-

149

ACS Catal. 2021, 11, 9143−9150

ACS Catalysis

silylium adducts e.g. [Et Si ·H-SiEt ] : see Nava, M.; Reed, C. A.

Research Article

(36) The resonances for both 1-H and 1-NTf₂are broadened

signiﬁcantly indicative of an interaction between the two species in

17/119

these reactions, which we assume prevents resolution of the

J_S_n_Hsatellites. Analogous observations have been seen for silane-

Triethylsilyl Perfluoro-Tetraphenylborate, [Et Si ][F -BPh ], a

−

Widely Used Nonexistent Compound. Organometallics 2011, 30,

798−4800.

37) 2,6-Lutidine (lut) showed comparable H activation kinetics to

(

2,4,6-collidine, as anticipated by their close structural similarity, yet it

was found that lut was an optimal base for catalytic imine

hydrogenations.

(

38) No evidence of debromination was observed (e.g. by iPr Sn·

radicals), which is consistent with the polar hydrogenation of the

same imine catalyzed by 1-OTf (see ref 27). Hydrodebromination of

aryl bromides is mediated by n Bu SnH via a radical mechanism; see

Neumann, W. P. Tri-n-butyltin Hydride as Reagent in Organic

Synthesis. Synthesis 1987, 6655−6683 For more recent reports of

“frustrated radical pairs ” see. (a) Liu, L.; Cao, L. L.; Shao, Y.; Ménard,

G.; Stephan, D. W. A Radical Mechanism for Frustrated Lewis Pair

Reactivity. Chem 2017, 3, 259−267. (b) Holtrop, F.; Jupp, A. R.;

Kooij, B. J.; Leest, N. P.; Bruin, B.; Slootweg, J. C. Single-electron

Transfer in Frustrated Lewis Pair Chemistry. Angew. Chem., Int. Ed.

020, 59, 22210−22216. (c) Soltani, Y.; Dasgupta, A.; Gazis, T. A.;

Ould, D. M. C.; Richards, E.; Slater, B.; Stefkova, K.; Vladimirov, V.

Y.; Wilkins, L. C.; Willcox, D.; Melen, R. L. Radical Reactivity of

Frustrated Lewis Pairs with Diaryl Esters. Cell Rep. Phys. Sci. 2020, 1,

00016 None of these reaction pathways was observed in this study,

thus a radical mechanism was excluded.

39) Fulmer, G. R.; Miller, A. J. M.; Sherden, N. H.; Gottlieb, H. E.;

(

Nudelman, A.; Stoltz, B. M.; Bercaw, J. E.; Goldberg, K. I. NMR

Chemical Shifts of Trace Impurities: Common Laboratory Solvents,

Organics, and Gases in Deuterated Solvents Relevant to the

Organometallic Chemist. Organometallics 2010, 29, 2176−2179.

(

40) The low conversion rates of EtOH to Et O (∼6%) could also

arise due to inhibition of 1-NTf by the H O product.

(41) Other triﬂuoroacetate esters were tested, yet none were

hydrogenated at super-stoichiometric conversion. See Supporting

Information for details.

(42) Further heating of the solution at higher temperatures

continued to increase the conversion. The ester hydrostannations

−

can also be eﬀected by [Lut-H] [NTf ] (at T > 80°C), however this

generated 1-NTf in situ by reaction with 1-H at these temperatures

(

as evidenced by the formation of H ). Accordingly these experiments

could not unambiguously probe an alternative reduction pathway by

which the ester is activated by [Lut-H]⁺instead of 1-NTf . See

Supporting Information for further details.

(

43) Unfortunately, it was not possible to discern any reaction

119

intermediates by either H, F or Sn{ H} NMR spectroscopy for

these reduction reactions, even at low temperature.

(

44) Alternative mechanistic pathways for RCO Et reduction

involving either direct reaction of the CO bond with 1-H, or

pre-activation of CO bond through H-bonding to [Lut-H] prior to

hydride transfer from 1-H, were also examined yet the associated

energy barriers (>42 kcal/mol) are considered unfeasible to be

operative under the experimental conditions.

(45) (a) Das, S.; Mondal, S.; Pati, S. K. Mechanistic Insights into

Hydrogen Activation by Frustrated N/Sn Lewis Pairs. Chem.Eur. J.

018 , 24 , 2575 −2579. (b) Das, S.; Pati, S. K. Unravelling the

Mechanism of Tin-based Frustrated Lewis Pair Catalyzed Hydro-

genation of Carbonyl Compounds. Catal. Sci. Technol. 2018, 8, 5178−

(

189.

46) A variety of other Lewis bases were also surveyed for MeCO Et

hydrogenation, albeit unsuccessfully. See Supporting Information for

more details.

150