Scholarly article on 1<i>H</i>-heptafluoro-1-methoxy-butan-1-ol 377-52-6 from Organic Process Research and Development p. 415

pubs.acs.org/OPRD

Article

Engineering Catalysts for Selective Ester Hydrogenation

Pavel A. Dub,* Rami J. Batrice, John C. Gordon, Brian L. Scott, Yury Minko, Jurgen G. Schmidt,

and Robert F. Williams

Cite This: Org. Process Res. Dev. 2020, 24, 415 −442

Read Online

sı

*

Metrics & More

Article Recommendations

Supporting Information

ACCESS

ABSTRACT: The development of eﬃcient catalysts and processes for synthesizing functionalized (oleﬁnic and/or chiral) primary

alcohols and ﬂuoral hemiacetals is currently needed. These are valuable building blocks for pharmaceuticals, agrochemicals,

perfumes, and so forth. From an economic standpoint, bench-stable Takasago Int. Corp.’s Ru-PNP, more commonly known as Ru-

MACHO, and Gusev’s Ru-SNS complexes are arguably the most appealing molecular catalysts to access primary alcohols from esters

and H₂(Waser, M. et al. Org. Proc. Res. Dev. 2018, 22, 862). This work introduces economically competitive Ru-SNP(O)_zcomplexes

(z = 0, 1), which combine key structural elements of both of these catalysts. In particular, the incorporation of SNP heteroatoms into

the ligand skeleton was found to be crucial for the design of a more product-selective catalyst in the synthesis of ﬂuoral hemiacetals

under kinetically controlled conditions. Based on experimental observations and computational analysis, this paper further extends

the current state-of-the-art understanding of the accelerative role of KO-t-C₄H₉in ester hydrogenation. It attempts to explain why a

maximum turnover is seen to occur starting at ∼25 mol % base, in contrast to only ∼10 mol % with ketones as substrates.

KEYWORDS: homogeneous catalysis, ester hydrogenation, ﬂuoral hemiacetal, base eﬀect, mechanism

The resultant functionalized esters can rather be more

conveniently and selectively hydrogenated into the correspond-

ing primary alcohols with homogeneous and even some

heterogeneous catalysts.^13,14At least two examples of catalytic

hydrogenation of chiral esters with heterogeneous catalysts

without signiﬁcant racemization are reported.¹⁵The drawback

of these systems is either of the following: limited substrate

scope, high catalyst loading, and/or potential poisoning of the

catalyst with H₂. Homogeneous catalysts have emerged as

eﬃcient catalysts for mild hydrogenation of esters.^9,13,14We

note some particularly eﬃcient catalysts [in terms of high

substrate-to-catalyst (S/C) ratios] developed by the ﬁne

chemical industry¹⁶and academia.¹⁷Several of the-

se^{16c,17a,d,e,h,18}as well as other¹⁹catalysts were probed in the

hydrogenation of oleﬁnic or chiral esters in a C(O)O/CC

chemoselective manner or without aﬀecting optical purities,

respectively. Fewer catalysts have been studied in the hydro-

genation of α-ﬂuorinated esters into ﬂuoral hemiacetals under

kinetically controlled conditions.^18b,20From an economic

standpoint (synthetic protocols, stability, toxicity, turnover

eﬃciencies),^18abench-stable Takasago Int. Corp.’s^16cRu-

MACHO and Gusev’s Ru-SNS complexes^17hare arguably the

most notable catalysts, Figure 2.^18a,20d,21

1. INTRODUCTION

The U.S. Department of Energy (DOE) has set the goal of

replacing 25% of industrial petroleum-derived organic chemicals

with biomass-derived chemicals by 2025.^1,2According to the

initial DOE report of 2004,¹and its 2010 update,³the majority

of top value-added platform chemicals that have become

accessible from biomass are carboxylic acids (CAs). The

combined list includes enantio-enriched (S)-lactic, (S)-aspartic,

(S)-glutamic, and ^D-glucaric acids, as well as oleﬁnic itaconic

acid, Figure 1A.

The 2015 industrial production of these CAs from biomass

feedstocks through microbial fermentation and biotransforma-

tion reached ∼3 × 10⁶tons, Figure 1A.⁴Other functionalized

CAs that are already produced or have the potential to be

produced from biomass include three isomers of muconic acid

(MA),⁵^L-lysine,⁶fumaric acid,⁷and acrylic acid,⁸Figure 1B.

Most of the CAs in Figure 1A,B are envisioned as platform

chemicals for bio-based chemical and polymer production (e.g.,

synthetic resins, industrial plastics, polyesters, polyols, food

ingredients, biodegradable and engineering polymers). We, in

addition, envision these bio-derived (as well as other fossil-

derived) CAs as synthons for highly functionalized primary

alcohols. Indeed, the latter are valuable building blocks for

medicines, agrochemicals, perfumes, and others.⁹

In this work, we present derivatization of these two molecular

The catalytic hydrogenation of CAs is an ideal method for the

bulk production of alcohols from a practical/industrial point of

view because water is the only byproduct, and salt waste is not

formed as is always the case in stoichiometric hydrogenations

(e.g., with LiAlH₄), Figure 1C. Unfortunately, selective

hydrogenation of CAs is very challenging or diﬃcult even if

homogeneous¹⁰or enzymatic¹¹catalysts are employed. On the

other hand, CAs can be easily converted into esters without loss

in optical purities and/or aﬀecting other functional groups.¹²

complexes, speciﬁcally, Ru-SNP(O)_zcatalysts, as shown in

Received: December 31, 2019

Published: January 30, 2020

https://dx.doi.org/10.1021/acs.oprd.9b00559

Org. Process Res. Dev. 2020, 24, 415−442

415

Organic Process Research & Development

pubs.acs.org/OPRD

Article

Figure 1. (A) 2015 industrial production of functionalized CAs from biomass feedstocks through microbial fermentation and biotransformation,

^aincluding ∼100,000 t/year of racemic lactic acid via chemical synthesis and a pilot-scale production of the (R)-enantiomer.²⁵(B) Other functionalized

CAs already produced from biomass, to be produced, or potentially to be produced from biomass. ^bThree isomers of MA designated as cis,cis-MA,

cis,trans-MA, and trans,trans-MA.⁵Only cis,cis-MA is shown. ^cDL- and enantio-enriched D-lysine can also be assessed.^6a^dValue for the 2007market.^6b

^eBased on petrochemical routes, but biotechnological production is under consideration.⁷^fCurrently produced from crude oil-derived propylene by

selective oxidation, but biomass-derived path from glycerol is discussed.⁸(C) Limitations in catalytic hydrogenation of CAs into oleﬁnic and/or

enantio-enriched primary alcohols, valuable building blocks.

Figure 2. Takasago Int. Corp.’s Ru-MACHO Ru-PNP^16cand Gusev’s Ru-SNS precatalysts,^17has well as Ru-SNP(O)_zprecatalysts developed in this

work for selective ester hydrogenation. Abbreviations: Me = CH₃, Bn = CH₂C₆H₅, ^tBu = C(CH₃)₃, Ph = C₆H_6.

Figure 2, where z = 0, 1, and R is selected from Me, Bn, ^tBu, or

Ph. The Ru-SNP catalysts (z = 0) and corresponding ligands

were independently developed by us at LANL,²²as well as

researchers from Givaudan SA²³and Takasago Int. Corp.²⁴

Herein, we report the LANL synthetic protocols to SNP(O)_z

ligands and their Ru complexes, as well as provide a comparison

with competing methodologies. The important goal of this work

was, however, to compare the catalytic eﬃciency of Ru-

SNP(O)_zwith Ru-MACHO and Ru-SNS complexes, and

report an unexpected “nonlinear” behavior that was identiﬁed.

Furthermore, based on the example of the complex Ru-2a (vide

infra), the advantage of Ru-SNP catalysts was demonstrated in

the hydrogenation of α-ﬂuorinated esters into the corresponding

ﬂuoral hemiacetals under kinetically controlled conditions. In

addition to the introduction of new powerful catalysts for

selective ester hydrogenation, this work provides further insights

into the accelerative role of excess KO-t-C₄H₉in catalytic ester

hydrogenations. In particular, based on hybrid dispersion-

corrected density functional theory (DFT) calculations, we

demonstrate the eﬀect of a “solvated” KO-t-C₄H₉molecule on

the catalytic potential energy surface (PES) for the ﬁrst time.

2. RESULTS AND DISCUSSION

2.1. Synthesis and Characterization of SNP(O)_z

Ligands (z = 0, 1) and Their Ru Complexes. The targeted

SNP ligands of the formula RSCH₂CH₂NHCH₂CH₂PPh₂,

t

where R is selected from Me (2a), Bn (2b), Bu (2c), and Ph

(2e), were synthesized by using an optimized three-step

methodology shown at the top of Scheme 1 (LANL method).

Speciﬁcally, the ﬁrst step of the process corresponds to the

isolation of known^26,272-aminoethyl alkyl sulﬁdes

RSCH₂CH₂NH₂by reacting the inexpensive starting materials

416

https://dx.doi.org/10.1021/acs.oprd.9b00559

Org. Process Res. Dev. 2020, 24, 415−442

Organic Process Research & Development

pubs.acs.org/OPRD

Article

Scheme 1. Comparison of Synthetic Methodologies Leading to the Synthesis of SNP(O)_zLigands Described by LANL (This

Work),²²Givaudan SA,²³and Takasago Int. Corp.²⁴

a

b

Column chromatography in air. Column chromatography under an inert atmosphere. Trt =C(C₆H₅)₃.

(bromoethylamine hydrobromide and RSX, where X = Na for R

= Me, X = H for R = Ph and Trt; chloroethylamine

hydrochloride and BnSH²⁸). Alternatively, all RSCH₂CH₂NH₂

are commercially available from several vendors; for example,

^tBuSCH₂CH₂NH₂used in this study was purchased from

Enamine, LLC. The second step of the process corresponds to

an aza-Michael addition of RSCH₂CH₂NH₂to diphenylvinyl-

phosphine oxide, easily obtainable from commercial diphenylvi-

nylphosphine. This step of the process was carefully optimized

based on ³¹P{¹H} NMR and/or thin-layer chromatography

(TLC). In particular, for R = Me, the reaction can be carried out

in water as a solvent, and the desired SNPO ligand 1a can be

further isolated as a yellowish oil in 78% yield after ﬂash

chromatography. In contrast, for R = Bn, ^tBu, and Ph, the same

procedure aﬀorded signiﬁcant amounts of the corresponding

tertiary amine byproduct as a result of double addition.

Switching from water to a binary water−EtOH mixture

increased the yield of the desired product and allowed for the

isolation of 1b, 1c, and 1e in 80−85% yield after ﬂash column

chromatography. The same methodology for R = Me gave 1a in

a 69% isolated yield. Finally, the third step of the process

corresponds to the reduction of 1a−c,e by HSiCl₃/NEt₃under

an argon atmosphere to aﬀord the desired 2a−c,e in 79−93%

isolated yield. In the course of this work, the SNPO ligand 1d

was also synthesized in 82% isolated yield (R = trityl ≡ Trt; 20.8

g from 5.7 g of diphenylvinylphosphine oxide), but no further

417

https://dx.doi.org/10.1021/acs.oprd.9b00559

Org. Process Res. Dev. 2020, 24, 415−442

Organic Process Research & Development

pubs.acs.org/OPRD

Article

Scheme 2. Synthesis of Ru Complexes Supported by SNP(O)_zLigands (z = 0, 1)

a

b

c

d

Contains 1−4% of PPh₃. Isolated yield after additional recrystallization. 1.1 equiv. of 1c was used. Contains solvent as minor impurity (up to

e

∼0.5 equiv). Purity ~93%.

³¹P{¹H} NMR. The same result was obtained if the reaction time

was increased to 2 h. After several trials, we found that the

reaction works well if 1.2 equiv of ligand 2a is used and reﬂux is

applied. Following this methodology, complexes Ru-2a−c and

Ru-2e supported by SNP ligands 2a−c and 2e were prepared

and isolated in 53−88% yield, depending on the conditions

used, Scheme 2. The solvent used was either 1,2-dichloroethane

or toluene. Higher yields were isolated in toluene, but the

product retained some residual solvent impurity which required

more extensive drying. We note here that Takasago Int. Corp.²⁴

and Givaudan SA²³reported similar methodologies and isolated

Ru-2a in 98 and 93% yield, respectively. In the former case, the

product was claimed to be 99.8% pure, and the major impurity

was residual toluene.

The Ru complexes shown in Scheme 2 were characterized by

elemental analysis, X-ray crystallography whenever possible, and

NMR spectroscopy. The X-ray structures of complexes Ru-1a

and Ru-2a are compared in Figure 3, whereas those of Ru-1b−c,

Ru-1e, and Ru-2c are available in the Experimental Section

(Figures 14 and 15). For all the complexes analyzed by X-ray

crystallography, mer-coordination of the tridentate ligands 1 or 2

is observed. The N and S atoms of the SNP ligands are both

chiral centers, leading to four sets of stereoisomers for each

metal complex. Because all structures reported herein were

found to be centrosymmetric, the individual enantiomers were

not resolved as expected under nonchiral conditions. Thus, the

selected crystals for diﬀraction contained either enantiomeric

pairs R, S↔S, R or R, R ↔ S, S, Figure 4c.

attempts were made to reduce 1d to 2d. All ligands 1a−e and

2a−c,e were characterized by elemental analysis, NMR

spectroscopy, and/or ESI-MS. Most of the ligands appear as

pale yellow oils, except 1d which forms a waxy solid at room

temperature and 1b that exists as a white solid. We found that 1a

can further crystallize into an oﬀ-white solid upon standing at

−35 °C.

The LANL methodology can be further compared to those

independently described by Givaudan SA²³and Takasago Int.

Corp.,²⁴middle and bottom of Scheme 1, respectively.

Speciﬁcally, the key building block for Givaudan’s strategy²³is

n

alkylthioethanol RSCH₂C(O)H (R = Me, Bu, Oct), the

synthesis of which is well described starting from chloro- and/or

alkylthioacetaldehyde dialkyl acetals.²⁹Reaction of RSCH₂C-

(O)H with commercial Ph₂PCH₂CH₂NH₂aﬀorded 2a, 2f, and

2g in 19−31% isolated yield after ﬂash chromatography under

an inert atmosphere. The key building block used in Takasago’s

strategy²⁴is 3-(2-chloroethyl)-2-oxazolidinone, which is

obtained from N,N-bis(chloroethyl)amine hydrochloride/

CO₂/NEt₃in 99% yield. Further steps involve its reaction

t

with Ph₂PLi or RSNa (R = Me, Et, Bu) to generate 3-[2-

(diphenylphosphino)ethyl]-2-oxazolidinone and 3-[2-

(alkylthio)ethyl]-2-oxazolidinone, respectively. Finally, each

reagent is reacted with RSX (X = H or Na) and Ph₂PLi,

respectively, to generate the ﬁnal products in 78−90% isolated

yield after column chromatography under an inert atmosphere.

2.1.1. Synthesis and Characterization of Ru Complexes

Supported by SNP(O)_zLigands (z = 0, 1). Ru complexes Ru-

1a−c and Ru-1e supported by SNPO ligands 1a−c and 1e were

isolated in 76−99% yield by reacting [RuCl₂(PPh₃)₃] with 1

equiv of the corresponding ligand in dichloromethane at 25 °C

(2 h), as shown in Scheme 2.³⁰In some cases (complexes Ru-1b

and Ru-1e), additional recrystallization was beneﬁcial to liberate

from traces of “free” PPh₃, present at 1−4% according to NMR.

It was found that [RuCl₂(PPh₃)₃] reacted with 1 equiv of SNP

ligand 2a in dichloromethane to aﬀord Ru-2a and “free” PPh₃as

the sole products within 10 min based on in situ ³¹P{¹H} NMR.

Leaving this mixture unaﬀected (no stirring) spontaneously (∼2

h) produced crystals of Ru-2a·CH₂Cl₂based on X-ray

diﬀraction and NMR spectroscopy, with a moderate (∼57%)

isolated yield. Attempts to increase the yield of Ru-2a in a

separate attempt by layering with diethyl ether were undertaken,

but the isolated product contained minor impurities manifested

as some broad peaks around the desired product based on

In the ¹H, ³¹P{¹H}, and ¹³C{¹H} solution-state NMR spectra

of all Ru complexes in Scheme 2, two isomers are observed.³¹

Figure 4a,b shows the ³¹P{¹H} NMR spectra of crystals of Ru-1a

and Ru-2a after dissolving in CD₂Cl₂. Observation of two

isomers is attributed to equilibration of diastereomers, as shown

in Figure 4c.^20d

2.1.2. Benchmark Studies: Catalytic Hydrogenation of

Methyl Triﬂuoroacetate FE1 by Ru Complexes Supported by

SNP(O)_zLigands (z = 0, 1), Ru-MACHO, and Ru-SNS. In order

to initially limit the precatalyst scope, identify the inﬂuence of

PO versus P heteroatom, SR parameter (R = Me, Bn, ^tBu, Ph),

and compare the catalytic activity with commercially available

Ru-MACHO and Ru-SNS precatalysts, we studied the catalytic

hydrogenation of methyl triﬂuoroacetate FE1 into triﬂuoroace-

taldehyde methyl hemiacetal Hem1 under kinetically controlled

conditions,^18b,20Table 1.

418

https://dx.doi.org/10.1021/acs.oprd.9b00559

Org. Process Res. Dev. 2020, 24, 415−442

Organic Process Research & Development

pubs.acs.org/OPRD

Article

Figure 4. (a) ³¹P{¹H} NMR (CD₂Cl₂, 162 MHz, r.t.) spectrum of

complex Ru-1a, z = 1; (b) ³¹P{¹H} NMR spectrum of complex Ru-2a, z

= 0; (c) relationship between enantiomeric pairs in the centrosym-

metric crystals reported herein. Any given crystal contains either the red

or blue set of enantiomers.

The conditions employed were based on previous optimiza-

tion studies with Ru-MACHO and similar catalysts,^18b,20d

speciﬁcally, S/C ratios of 20,000, 40 °C, methanol (MeOH), 25

mol % of MeONa, and 10 h.

Homogeneous hydrogenation of FE1 that bears the electron-

withdrawing CF₃group aﬀords Hem1 as the kinetic product and

2,2,2-triﬂuoroethanol Alc1 as the thermodynamic product.^18b

Of great interest is the selective kinetically controlled synthesis

and further isolation of Hem1 via catalytic hydrogenation as

demonstrated in several patents^20a−cof Central Glass Co. based

on the Ru-MACHO complex.^18bIndeed, Hem1 is an important

synthon for the introduction of CF₃functionality directly or via

intermediate ﬂuoral formation.³²The following conclusions can

be made based on results presented in Table 1: (1) SNP ligands

are more eﬃcient than their SNPO analogues; nevertheless, Ru-

SNPO catalysts 1a−1e may ﬁnd applications in other reactions

with Noyori-type catalysts;^9,18a,33(2) there is a negligible

inﬂuence within experimental accuracy of the SR parameter on

both catalytic activity and selectivity within the SNP(O)_zfamily,

with slight preference for R = Me; (3) Ru-2a precatalyst (R =

Me) is comparable with Ru-MACHO for activity but over-

whelms the latter in terms of selectivity (∼98 vs ∼86%, runs 6

and 11); (4) Gusev’s Ru-SNS precatalyst performs better than

Ru complexes supported by SNPO ligands, but the activity is

unsatisfactory when compared with Ru-2a−c, Ru-2e, and Ru-

MACHO precatalysts. We have previously noted that Ru-

MACHO overwhelms Ru-SNS under similar conditions for this

reaction.^20dThis observation is attributed to a lower stability of

Ru-SNS in methanol^16c,20das discussed below.

Figure 3. X-ray structure of complexes Ru-1a and Ru-2a·CH₂Cl₂(35%

level of thermal ellipsoids, CH₂Cl₂is omitted for clarity). Selected bond

lengths and angles for Ru-1a: Ru1−O1 = 2.1304(17) Å, Ru1−N1 =

2.201(2) Å, Ru1−S1 = 2.2651(7) Å, Ru1−P2 = 2.2712(8) Å, Ru1−Cl1

= 2.4074(8) Å, Ru1−Cl2 = 2.4069(8) Å, Cl1−Ru1−Cl2 = 165.92(2)°,

and Ru-2a: Ru1−P1 = 2.3112(4) Å, Ru1−N1 = 2.1766(13) Å, Ru1−S1

= 2.4225(4) Å, Ru1−P2 = 2.3213(4) Å, Ru1−Cl1 = 2.4115(4) Å,

Ru1−Cl2 = 2.4151(4), Cl1−Ru1−Cl2 = 167.820(14)°.

2.1.3. Base Eﬀect on the Catalytic Hydrogenation of Neat

Natural Methyl Hexanoate with the Ru-2a Complex. Similar

419

https://dx.doi.org/10.1021/acs.oprd.9b00559

Org. Process Res. Dev. 2020, 24, 415−442

Organic Process Research & Development

pubs.acs.org/OPRD

Article

selectivity of the Ru-2a complex with commercial Ru-MACHO

and Ru-SNS precatalysts in the hydrogenation of commercial

E1, functionalized esters E2−E4, and chiral esters E5−E6 under

the same conditions and typically in the presence of 25 mol % of

KO-t-C₄H₉, Table 2. Our preference was given to the

hydrogenation of these esters under neat conditions whenever

possible. For those substrates that require a solvent for certain

reasons, preference was given to MeOH. Indeed, because most

of the substrates are methyl esters, their reduction will

necessarily produce methanol; thus, its direct use as the reaction

medium might greatly simplify solvent recycle in a possible

scale-up (no solvent separation steps are required, an important

consideration in pharmaceutical and large-scale industrial

processes).

Table 1. Comparative Hydrogenation of Methyl

Triﬂuoroacetate FE1 Catalyzed by Ru Complexes Supported

by SNP(O)_zLigands (z = 0, 1), Ru-MACHO, and Ru-SNS

In agreement with previous studies of Gusev et al.,^17hE1 is

hydrogenated more eﬀectively with Ru-SNS than Ru-MACHO

under neat conditions (runs 1−2). Interestingly, under the same

conditions, Ru-2a performs equally well as Ru-SNS, as

evidenced by the same rate of H₂pressure drop in the autoclave

and the level of conversion/yield (run 3). The same behavior

was noted for functionalized E2, which was quantitatively

(>97% conversion) and CO/CC chemoselectively

(∼99%)³⁷converted into 1-undecen-11-ol (∼94% yield) with

both Ru-SNS and Ru-2a, whereas Ru-MACHO proved to be

much less eﬃcient (runs 4−6).

a

b

entry

cat.

Ru-1a

Ru-1b

Ru-1c

Ru-1e

Ru-2a

Ru-2b

Ru-2c

conv., %

sel., %

1

2

3

4

5

6

7

8

∼12

∼8

∼98

∼91

∼98

∼99

∼98

∼86

∼11

c

>97

∼9

∼92

∼82

∼79

∼90

∼43

∼96

In the next step, we attempted to hydrogenate α,β-

unsaturated esters E3 and E4 into the corresponding allyl

alcohols, cinnamyl alcohol, and prenol (Table 2). These

chemicals are important intermediates to pharmaceuticals and

aroma compounds and industrial demand of cinnamyl alcohol is

usually fulﬁlled by selective enzymatic hydrogenation of

cinnamaldehyde.³⁸Prenol is produced industrially (e.g., by

BASF and Kuraray) by the reaction of formaldehyde with

isobutene, followed by the isomerization of the resulting

isoprenol.³⁹The CO/CC chemoselective hydrogenation

of E3 and E4 could aﬀord us with alternative routes to these

chemicals. Whereas we are unaware of attempts to hydrogenate

E4 into prenol, hydrogenation of E3 into cinnamyl alcohol has

been attempted by Firmenich SA researchers,^16eas well as by

Pidko^17dand de Vries^19ain academia, Scheme 3. The Pidko

complex^17dseems to be the most eﬃcient to selectively

hydrogenate E3 into cinnamyl alcohol in terms of the turnover

number (TON = 1800), but this catalyst suﬀers from a

complicated and expensive synthetic methodology.^17d,40

Methyl cinnamate E3 is a white solid at room temperature and

melts above 40 °C. Attempts to hydrogenate this substrate

under neat conditions at 80 °C were practically unsuccessful for

all three catalysts (runs 7−9). Some small (∼10%) conversion

into unidentiﬁed products was noted for Ru-MACHO and Ru-

2a (runs 7, 9). Therefore, methanol was further used as the

solvent (runs 10−12). Similar to observations with E1−E3,

performance with Ru-MACHO was poor (run 10), whereas Ru-

SNS and Ru-2a delivered the desired product in 33 and 48%

yields, respectively (runs 11−12). The conversion was higher

with Ru-2a (98 vs 59%), and according to the GC−MS and

NMR spectroscopic analysis of the mixture, this is due to further

reduction of the product. Nevertheless, a TON of 2400 (e.g.,

slightly higher than Pidko’s complex) leading to cinnamyl

alcohol is noted, and this number can be compared with TONs

with other catalysts, Scheme 3. When methanol was replaced

with toluene, the same reactions proceeded much faster with

both Ru-SNS and Ru-2a, as evidenced by a faster drop in H₂

pressure (runs 13−16).

9

10

11

Ru-2e

Ru-SNS

Ru-MACHO

a

Total conversion into triﬂuoroacetaldehyde methyl hemiacetal

Hem1 and 1,2,3-triﬂuoroethanol Alc1, average of three runs,

b

determined by ¹⁹F NMR. Percentage of Hem1 from converted

c

FE1. S/C = 1000, 2 h.

to ketone hydrogenations, ester hydrogenations with Noyori-

type catalysts are also further accelerated by the addition of

inorganic bases containing alkali metal cations, such as KOH,

NaOH, NaOCH₃, KO-i-C₃H₇, or KO-t-C₄H₉, or even sources

generating alkali metal cations such as NaBAr^F, NaBH₄, and so

forth,⁹Whereas the maximum rate is typically observed starting

at the ∼5−10 mol % level for ketones with Noyori-type

catalysts,³⁴esters typically require up to ∼25 mol% according to

titration curves in a given solvent.^16d,18b,35The eﬀect of the base

has not been studied under neat conditions to the best of our

knowledge. Figure 5 shows the eﬀect of base on the catalytic

hydrogenation of neat natural (from biomass)³⁶commercial

methyl hexanoate (E1) with the Ru-2a precatalyst.

The following conclusions can be made: (1) as expected, no

reaction occurs in the absence of the base, consistent with the

fact that at least 2 equiv of base is needed with respect to the

trans-Ru dichloride precatalyst to transform it into the

catalytically active trans-Ru dihydride complex;^18b(2) no

reaction occurs in the presence of LiOMe, possibly because of

the base’s limited solubility; (3) the reaction slightly depends on

the identity of the anion (MeO⁻vs ^tBuO⁻) in the 10−35 mol %

range and is accelerated in moving from Na to K; (4) the

maximum yield of the 1-hexanol product is observed at ∼25 mol

% for K⁺, although a satisfactory conversion is already observed

at ∼5 mol % of the inorganic base.

2.1.4. Comparative Hydrogenation of Methyl Hexanoate,

Functionalized, and Chiral Esters with Ru-2a, Ru-MACHO,

and Ru-SNS. In the next step, we compared the activity and

420

https://dx.doi.org/10.1021/acs.oprd.9b00559

Org. Process Res. Dev. 2020, 24, 415−442

Organic Process Research & Development

pubs.acs.org/OPRD

Article

Figure 5. Eﬀect of added base on the catalytic hydrogenation of neat methyl hexanoate (natural, ≥99%, W270814 Sigma-Aldrich) with the Ru-2a

precatalyst, 50 mL Parr autoclave. Each point is an average of two runs (the largest error margin is shown), except for LiOMe (one run). Yield of 1-

hexanol is determined by ¹H NMR spectroscopy. The balance of the material present was unreacted methyl hexanoate and hexyl hexanoate byproduct.

This fact could point to faster decomposition in methanol of

both catalysts, and it seems that either Ru-SNS decomposes

faster or Ru-2a performs faster hydrogenation. Under the

conditions of runs 14−15, ∼91% conversion of E3 and ∼45%

yield of cinnamyl alcohol were observed with Ru-SNS, whereas

100% conversion of E3 and only ∼4% yield of cinnamyl alcohol

were observed with Ru-2a, suggesting that Ru-2a is too active

under the conditions employed. We also note here that fully

saturated benzenepropanol can be additionally produced from

cinnamyl alcohol via fast parallel catalytic cinnamyl alcohol →

benzenepropanal isomerization and following fast hydro-

genation.⁴¹

In the following step, we attempted to hydrogenate E4 under

neat conditions and with the Ru-2a complex. The reaction did

proceed according to a drop in H₂pressure, but in contrast to E1

and E2 as substrates, the overall mixture represented white

suspension at the end. Attempts to homogenize the system using

a smaller loading of base (5 mol % of KO-t-C₄H₉) indeed

aﬀorded a clean solution at the end, but the mixture was

complicated according to NMR analysis (desired product

present at ∼24% level according to integration with respect to

added pyridine as the internal standard). We further used

methanol as a media and the standard 25 mol % of KO-t-C₄H₉

(runs 17, 19−20). Under these conditions, Ru-MACHO

delivered prenol with only 4% yield (run 17). The overall

conversion of E4 reached >85%, however. A separate experi-

ment shows that in the absence of a catalyst and H₂pressure, E4

is converted into methyl 3,3-dimethylacrylate ester (30%) as

well as β-methoxy-isovaleric acid methyl ester (60%), as a result

of a competitive (background) KO-t-C₄H₉-catalyzed oxa-

Michael reaction (run 18).⁴²In contrast to Ru-MACHO, Ru-

SNS and Ru-2a delivered the desired product with somewhat

appreciable yields of 45 and 40%, respectively (runs 19−20).

Note that similarly to E3, the overall conversion with Ru-2a was

higher (88 vs 54%). According to the in situ analysis of the

mixture, this is due to further reduction of the product, that is,

Ru-2a was slightly over reactive under the conditions employed.

Chiral substrates (S)-E5 and (S)-E6 were tested with an

enantiomeric excess (ee) of 97 and 98%, respectively (runs 21−

30). To recall, Takasago Int. Corp. originally developed Ru-

MACHO for the hydrogenation of chiral esters into

corresponding alcohols, ideally without loss in optical purity.^16c

For example, 2200 kg of methyl (R)-lactate (99.6% ee) can be

hydrogenated with 6.4 kg of Ru-MACHO (S/C = 2000) in

methanol into 1477 kg of isolated (R)-1,2-propanediol with a

99.2% ee in a single batch process. The optical purity of (R)-1,2-

propanediol obtained using this process was higher than that via

the asymmetric hydrogenation of hydroxyacetone (99.2 vs

98.5% ee), conducted at Takasago plants since 1992 with the

Ru-SEGPHOS complex.^16cOur studies performed on methyl

(S)-lactate E5 indicate that complexes Ru-2a and Ru-SNS

perform equally well in terms of ee retention but approximately

twice as eﬃcient as Ru-MACHO (runs 21−23). In the presence

of even 5 mol % of KO-t-C₄H₉, the ﬁnal product is obtained with

∼97% ee based on chiral GC−MS analysis, that is, there is less

than 1% drop in the ee based on accuracy of the method used.

Note, as expected, increasing the reaction time or amount of the

base results in a better yield and a faster reaction, respectively,

but there is a signiﬁcant drop in optical purity to ∼80% ee (runs

24−25). Nevertheless, it seems that both Ru-2a and Ru-SNS

421

https://dx.doi.org/10.1021/acs.oprd.9b00559

Org. Process Res. Dev. 2020, 24, 415−442

Organic Process Research & Development

pubs.acs.org/OPRD

Article

Table 2. Comparative Hydrogenation of Methyl Hexanoate E1 and Functionalized Esters E2−E6 with Ru-2a, Ru-MACHO, and

Ru-SNS Complexes

a

entry

subs.

cat.

solvent

time

conv., %

yield, % (ee, %)

1

2

3

4

5

6

7

8

E1

E2

E3

E4

E5

E6

Ru-MACHO

Ru-SNS

Ru-2a

Ru-MACHO

Ru-SNS

Ru-2a

Ru-MACHO

Ru-SNS

Ru-2a

Ru-MACHO

Ru-SNS

Ru-2a

Ru-MACHO

Ru-SNS

Ru-2a

Ru-MACHO

6

∼33

∼99

∼22

∼97

∼10

∼0

∼10

∼18

∼59

∼98

<1%

∼91

100

>85

91

54

88

30

74

69

>94

91

99

82

∼14

∼98

∼0

∼94

0

∼33

∼48

∼0

∼51

∼4

b

6

b

6

c

6

24

6

c

d

9

24

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

MeOH

toluene

MeOH

24

e

4

f

4

g

4

1

h

m

4

i

j

48

∼4

0

45

40

ib

Ru-SNS

Ru-2a

Ru-MACHO

Ru-SNS

Ru-2a

Ru-MACHO

Ru-SNS

Ru-2a

48 ^,

ib

48 ^,

k

ib

24 ^,

MeOH

28 (∼97)

72 (∼97)

67 (∼97)

92 (∼78)

86 (∼80)

99 (∼1)

80 (∼3)

94 (∼1)

74 (∼4)

92 (∼5)

k

ib

24 ^,

ib

24 ^,

i

72

24

il

18 ^,

il

18 ^,

il

18 ^,

95

>74

92

io

72 ^,

Ru-2a

no

,

48

a

Total conversion into unsymmetrical ester, alcohol, and for some substrates, other products. Determined by NMR spectroscopy and/or (chiral)

b

c

d

GC−MS. Average of two runs. Average of three runs. At 80 °C, conversion predominantly into C₆H₅CH₂CH₂CO₂CH₃.

e

f

C₆H₅CH₂CH₂CO₂CH₃(8%), C₆H₅CH₂CH₂CH₂OH (1%), and transesteriﬁcation products (17%). C₆H₅CH₂CH₂CO₂CH₃(17%),

g

C₆H₅CH₂CH₂CH₂OH (28%) and transesteriﬁcation products (2%). C₆H₅CH₂CH₂CO₂CH₃(16%), C₆H₅CH₂CH₂CH₂OH (13%), and

transesteriﬁcation products (11%). Major product is C₆H₅CH₂CH₂CH₂OH (82%). S/C = 2500. Total conversion into methyl 3,3-

dimethylacrylate (30%) and β-methoxy-isovaleric acid methyl ester (60%). Base = 5 mol %. Base = 15 mol %. At 25 °C, C₆H₅CH₂CH₂CO₂CH₃

(29%), C₆H₅CH₂CH₂CH₂OH (28%), and transesteriﬁcation products (39%). S/C = 1000. Base = 2.5 mol %.

h

i

j

k

l

m

n

o

422

https://dx.doi.org/10.1021/acs.oprd.9b00559

Org. Process Res. Dev. 2020, 24, 415−442

Organic Process Research & Development

pubs.acs.org/OPRD

Article

a

Scheme 3. Catalytic Hydrogenation Pattern of Methyl Cinnamate E3 with Homogeneous Catalysts

a

TON refers to the desired product formation, cinnamyl alcohol

could potentially replace Ru-MACHO in large-scale methyl (R)-

lactate hydrogenation described above.

(THF) or toluene gave unsatisfactory results.^18bIndeed,

regardless of the substrate, catalyst, or even the base used,

methanol was found to be an excellent solvent, producing

extremely transparent reaction solutions, Table 3.

Methyl mandelate (S)-E6 was found to be hydrogenated

equally well with both Ru-MACHO and Ru-2a (>95%

conversion; runs 26, 28), and slightly lower activity is noted

for Ru-SNS (82% conversation, run 27). However, the reaction

product was completely racemized consistent with the higher

acidity of the α-CH-proton in E6 versus E5. Attempts to

decrease the amount of base from 15 to 2.5 mol % resulted in an

approximate four-fold increase in reaction time when the Ru-2a

precatalyst was used (runs 28−29), but the racemization

problem was not resolved. Further decrease of S/C to 1000

resulted in a faster conversion as expected (runs 29−30), but

again, a racemate was obtained. No further attempts to optimize

the reaction were made.

The conversion of other substrates (other functionalized

esters) for the hydrogenation with Ru-2a and similar complexes

is available in independent patents by Givaudan SA²³and

Takasago Int. Corp.²⁴We note here excellent scalability of Ru-

SNP catalysts, for example, Givaudan SA scientists reported

selective hydrogenation of highly functionalized sclareolide on a

250 g scale and S/C = 40,000.²³

2.1.5. Comparative Hydrogenation of α-Fluorinated Esters

with Ru-2a, Ru-MACHO, and Ru-SNS. In the next step, we

compared the activity and selectivity of the Ru-2a complex with

commercial Ru-MACHO and Ru-SNS precatalysts in the

hydrogenation of α-ﬂuorinated esters FE1−FE6 into corre-

sponding hemiacetals Hem1−Hem6 under kinetically con-

trolled conditions,^18b,20Table 3. Because all of the ester

substrates FE1−FE6 are liquids under the conditions employed,

attempts to hydrogenate several of them with the Ru-2a

complex were undertaken under neat conditions. However,

upon exposure of the base additive to ﬂuorinated ester,

signiﬁcant heat was produced, which resulted in the

complication of the experimental procedure and, more

importantly, small conversions were noticed. We therefore

used methanol as a media, following previous optimization

studies with Ru-MACHO where methanol was found to be the

solvent of choice for these reactions because tetrahydrofuran

The following conclusions can be made: (1) substrates FE1−

FE5 can be hydrogenated with Ru-2a, Ru-MACHO, and Ru-

SNS under kinetically controlled conditions producing hemi-

acetals Hem1−5 with appreciable selectivity and TONs (entries

2−4, 6−8, 10−18); (2) addition of 25 mol % KO-t-C₄H₉to FE6

resulted in immediate precipitate formation, and no reaction was

further observed even after exposing this mixture to H₂(no

further eﬀorts to optimize the reaction were attempted); (3) as

expected, no reaction occurs in the absence of any of these

catalysts based on the example of esters FE1 and FE2 (entries 1

and 5); (4) in agreement with previous studies,^18bunder lower

S/C, prolonged reaction times, and/or higher H₂pressure, the

thermodynamic product alcohols Alc can be produced

quantitatively (e.g., entry 9); (5) the Ru-2a precatalyst produces

hemiacetals Hem1−5 with better selectivities and TONs than

Ru-MACHO and Ru-SNS (entries 2−4, 6−8, 10−18).⁴³For

example, a TON of 69000 is noted in the hydrogenation of

methyl diﬂuoroacetate FE2 into Hem2, which can be compared

to ∼53,000 turnovers with Ru-MACHO and Ru-SNS (entries

6−8). Such a turnover eﬃciency with Ru-2a is arguably a non-

negligible advance in the homogeneous hydrogenation of FE2

and other α-ﬂuorinated esters. For instance, Central Glass Co.

reported three (3 × 40 kg = 120 kg) batches of the FE2

hydrogenation process with Ru-MACHO (3 × 47 g = 141 g, S/C

= 5000) that aﬀorded stable Hem2 in the equivalent form of

ethyl hemiacetal in 70 kg (46%) isolated yield after a series of

distillations from a combined residue, Scheme 4.^20cIt is obvious,

therefore, that the use of the Ru-2a precatalyst for the same or

similar reactions could lower the cost of the process via

minimizing the amount of the catalyst used and an increased

yield of the product.⁴⁴We note here, however, that while Ru-

SNS is less eﬃcient compared to Ru-2a, its SNS ligand is

synthesized in one-step under aerobic conditions, followed by

simple puriﬁcation using vacuum distillation.^17h

423

https://dx.doi.org/10.1021/acs.oprd.9b00559

Org. Process Res. Dev. 2020, 24, 415−442

Organic Process Research & Development

pubs.acs.org/OPRD

Article

Table 3. Comparative Hydrogenation of α-Fluorinated Esters FE1−FE6 into Corresponding Fluoral Hemiacetals Hem1−Hem6

under Kinetically Controlled Conditions with Ru-2a, Ru-MACHO, and Ru-SNS Complexes

a

e

entry

subs.

cat.

S/C

time

conv., %

Hem, %

Alc, %

TON

1

2

3

4

5

6

7

8

FE1

FE2

FE3

FE4

FE5

FE6

24

8

10

6

0

92

40

91

0

72

54

0

84

38

89

0

53

52

69

0

8

2

0

19

2

11

100

43

14

10

9

∼0.3

∼1.4

35

0

42,000

19,000

44,500

0

53,000

52,000

69,000

0

2850

2150

3750

31,000

20,000

36,500

30,000

18,500

31,500

0

b

Ru-MACHO

Ru-SNS

Ru-2a

50,000

b

Ru-MACHO

Ru-SNS

Ru-2a

Ru-MACHO

Ru-SNS

Ru-2a

Ru-MACHO

Ru-SNS

Ru-2a

Ru-MACHO

Ru-SNS

100,000

5000

b

80

9

100

57

85

75

43

76

>95

>37

c

10

11

12

13

14

15

16

17

18

19

5000

57

43

75

62

40

73

60

37

63

c

6

b

50,000

5000

24

b

traces

7

0

b

Ru-2a

>70

d

0

a

b

c

d

Determined by ¹⁹F and/or ¹H NMR. Average of two runs. MeONa was used as a base. Crashed out (solid in the beginning and at the end of

e

the reaction). TON leading to Hem.

2.1.6. Comparative Hydrogenation of N,N-Diethyl-2,2,2-

triﬂuoroacetamide CA1 with Ru-MACHO, Ru-SNS, Ru-2a,

and Ru-1e. In the course of this work, we have also undertaken

attempts to hydrogenate N,N-diethyl-2,2,2-triﬂuoroacetamide

CA1 under kinetically controlled conditions, Scheme 5.

Catalytic hydrogenation of α-ﬂuorinated carboxamides could

be of tremendous interest for the same reasons as the catalytic

hydrogenation of α-ﬂuorinated esters. Catalytic synthesis of

ﬂuoral hemiaminals has not yet been reported, but similar to the

ﬂuoral hemiacetals described above, they can be envisioned as

valuable ﬂuoroalkylating agents.^32a,45In the example of N,N-

diethyl-2,2,2-triﬂuoroacetamide CA1, Scheme 5 shows that

kinetically controlled hydrogenation of α-ﬂuorinated carbox-

amides is indeed possible; however, under the conditions

employed, the reaction product is not the expected hemiaminal

1-(diethylamino)-2,2,2-triﬂuoroethan-1-ol^32abut rather hemi-

acetal Hem1; the latter is likely obtained from hemiaminal and

methanol via a metal-catalyzed process.⁴⁶Other aspects to note

are as follows: (1) as expected,¹³hydrogenation of CA1 is more

diﬃcult (lower S/C ratios) with respect to α-ﬂuorinated esters

described above; (2) Ru-MACHO is very active (>93%

conversion), but the selectivity is almost negligible (5%); (3)

Ru-SNS and Ru-2a aﬀord Hem1 with ∼80−90% selectivity but

moderate 36−54% conversions; (5) Ru-1e complex aﬀords

424

https://dx.doi.org/10.1021/acs.oprd.9b00559

Org. Process Res. Dev. 2020, 24, 415−442

Organic Process Research & Development

pubs.acs.org/OPRD

Article

use.^17d,eIndeed, CA and water impurities can lead to the

formation of ruthenium-carboxylate and -hydroxo species,

possible inactive catalytic species. Based on experimental

observation of such species under stoichiometric conditions,

Gauvin⁴⁹suggested that excess of inorganic base is needed to

regenerate the active catalytic species, that is, trans-RuH₂

complexes, Figure 6a.⁵⁰One can reasonably add here that,

prior to the formation of such carboxylate and hydroxo species,

the base can also destroy excess of corresponding parent acid

and H₂O in the very beginning of the reaction, Figure 6b.

Here, we note important experimental observations: ester

hydrogenation,^17eand somewhat related transfer hydrogena-

tion,⁵¹can, in principle, proceed without any added bases

providing that the precatalyst is used in the form of the dihydride

complex and that the substrate is at least distilled or otherwise

treated with basic alumina. It is, however, not clear whether a

base further accelerates the reaction rate under these conditions

just as it does for related ketones³⁴because most of the titration

curves in the wide range (up to 75 mol % base), including this

work, were obtained with commercial esters.^35a,52Our

experimental studies suggest that such acceleration is seen for

neat methyl hexanoate (cf. Figure 5) because no diﬀerence in

conversion/product yield is observed if the substrate is used in

the commercial or “preactivated” form (treated with basic

alumina/molecular sieves)⁵³under the condition of 2.5 mol %

base, Table 4, entries 1−2. We further noted that decreasing the

base amount to 1.0 mol % in this reaction resulted in a slightly

slower conversion (8 vs 6 h), and “preactivated”E1 is either

hydrogenated slightly better than the commercial ester (89 vs

86% conversion) or similarly taking into account the equipment

and instrumental accuracy (entries 3−4). On the other hand,

when 0.5 mol % of the base was used, there was only ∼33%

conversion of commercial E1, cf. almost ∼92% for “preacti-

vated” E1 after 12 h (entries 5−6). This observation is

Scheme 4. Three-Batches Large-Scale Hydrogenation of

Methyl Diﬂuoroacetate FE2 with the Ru-MACHO Complex

Reported by Central Glass Co.^20c

a

b

¹⁹F NMR yield. Isolated yield of the ethyl hemiacetal after series of

distillations from a combined three batches residue, purity > 87%.

similar activity as Ru-SNS and Ru-2a but under a lower S/C of

1000.

2.1.7. Experimental and Computational Investigation of

the Accelerative Role of KO-t-C₄H₉in Methyl Hexanoate E1

Hydrogenation. The mechanism of ester hydrogenation^18b,47is

understood to some degree with Noyori-type M/NH catalysts,⁹

but the accelerative role or set of roles that KO-t-C₄H₉(or a

similar base) plays in the reaction is less clear. Commercially

available esters, even though claimed to be >99% pure, often

contain a signiﬁcant amount of the corresponding parent acid

and H₂O impurities.⁴⁸This fact represents an important

consideration as large-scale commercial processes are typically

based on “as-received” esters (liquids are sometimes additionally

degassed by sparging with argon). In contrast, those in academia

performing small-scale reactions, particularly those employing

liquid esters, usually treat such materials with basic alumina and/

or molecular sieves (or even distillation) under argon prior to

Scheme 5. Hydrogenation of N,N-Diethyl-2,2,2-triﬂuoroacetamide CA1 Catalyzed by Ru-MACHO, Ru-SNS, Ru-2a, and Ru-1e

a

Complexes Under Kinetically Controlled Conditions, ¹⁹F NMR (rd = 10s)

a

Selectivity of Hem1 is with respect to major Alc1 and minor product, tentatively hemiaminal 1-(diethylamino)-2,2,2-triﬂuoroethan-1-ol (<1%;

∼3% for Ru-1e).⁵⁴

425

https://dx.doi.org/10.1021/acs.oprd.9b00559

Org. Process Res. Dev. 2020, 24, 415−442

Organic Process Research & Development

pubs.acs.org/OPRD

Article

Fully homogeneous solutions in runs 1−4 and 6 (Table 4) are

consistent with alcohol production upon turnover. Relevant to

this is another experimental observation: 25 mol% of KO-t-C₄H₉

in neat commercial E1 begins as a suspension, and does not

appear to undergo any appreciable dissolution upon stirring.

However, when the precatalyst is added, H₂is applied, and the

reaction is allowed to stir, and a clean, homogeneous solution is

observed at the end of the reaction (runs 2−3, Table 2).

Although this fact seems to be partially due to formation of the

alcohol product which further solubilizes the remaining base, it

can also point to more profound involvement of KO-t-C₄H₉in

the catalytic reaction. Therefore, although the transformations

shown in Figure 6a,b indeed most likely take place in the general

case, they are not relevant to the accelerative eﬀect. Previously

established for ketones,⁵⁵the formation of alkali metal amidate

complexes as shown in Figure 6c in the example of KO-t-C₄H₉

only partially explains what is occurring. Calculations of Pidko

shows that barriers of both outer-sphere hydride transfer and

H−H bond cleavage are lowered when the N−H functionality is

replaced by N−K,⁵⁶but similar observations were made for

ketones.^55cHowever, the formation of potassium amidate

complexes does not explain why a maximum turnover begins to

be observed typically using ∼25 mol % of base or more (cf.

Figure 5). This question becomes particularly interesting and

relevant by taking into account that only ∼5 to 10 mol% is often

needed for optimal conversion of ketones, for example, the

Noyori-catalyst.³⁴Figure 7 further supports this statement with

respect to the Ru-2a precatalyst and acetophenone as a

substrate. A maximum turnover begins to be observed at ∼10

mol % KO-t-C₄H₉.

Figure 6. Current state-of-the-art explanation of the accelerative role of

inorganic bases in ester hydrogenation relies in chemical modiﬁcation

of the catalyst or CA and water impurities: (a) the base is needed to

regenerate the active trans-RuH₂complex from presumably inactive

carboxylate and hydroxo species generated within the catalytic reaction

mixture from impurities present in ester; (b) destruction of CA and

water, discussion in this paper; (c) formation of potassium amidate

complexes from KO-t-C₄H₉or similar that more eﬃciently (lower

barrier) transfers the hydride atom (outer sphere) or regenerates the

catalyst via H−H bond cleavage. E, L = 2e⁻donor.

Table 4. Comparative Hydrogenation of Commercial and

“Pre-activated” (by Basic Alumina and Molecular Sieves)⁵⁶

Methyl Hexanoate E1 with the Ru-2a Precatalyst and Small

Excess of KO-t-C₄H₉

a

entry

substrate

base, mol %

time

conv., %

sel., %

1

2

3

4

5

6

commercial

“preactivated”

commercial

“preactivated”

commercial

“preactivated”

2.5

1.0

0.5

6

8

12

∼92

∼86

∼89

∼33

∼92

∼94

∼90

∼92

∼33

∼93

b

a

Conversion and 1-hexanol selectivity is determined by ¹H NMR (rd

= 10 s). The balance of the material present was unreacted methyl

b

hexanoate and byproduct hexyl hexanoate. Average of two runs.

Figure 7. Base eﬀect on the catalytic hydrogenation of acetophenone

(Acros Organics, 99%)⁵⁷with Ru-2a precatalyst, 50 mL Parr autoclave,

15 min (blue) vs 2 h (purple). Conversion and yield of the product is

determined by ¹H NMR (rd = 10 s). The balance of the material present

was the unreacted starting material.

consistent with the reactions in Figure 6a,b. In fact, a white,

translucent solution was seen for run 5 at the end of the reaction.

426

https://dx.doi.org/10.1021/acs.oprd.9b00559

Org. Process Res. Dev. 2020, 24, 415−442

Organic Process Research & Development

pubs.acs.org/OPRD

Article

Figure 8. Computed physical and chemical “activation” of E1 by KO-t-C₄H₉at the ωB97X-D/def2-QZVP/SMD (methylbutanoate, EPS = 4.70) level

of the theory. Noncritical H-atoms are omitted for the optimized structures. q_Cis NBO charge on the carbonyl carbon atom. KOtBu ≡ KO-t-C₄H₉.

Figure 9. Optimized geometries of two diastereomeric trans-Ru dihydride complexes int0a and int0b as well as potassium amidate complexes derived

from them at the ωB97X-D/def2-SVP//def2-QZVP/SMD(methylbutanoate, EPS = 4.70) level. All energies are calibrated relative to int0a. Dotted

lines represent cation−π and cation−lone pair noncovalent interactions. Noncritical H-atoms are omitted for clarity.

Therefore, when an ester is employed as a substrate, either its

physical⁵⁸and/or even chemical activation using an inorganic

base, and eventually the catalyst−substrate complex (CSC), is

most likely to take place.

2.1.7.1. Physical and Chemical Activation of Ester E1 using

KO-t-C₄H₉. Experimentally, esters possessing α-CH-function-

ality can react with inorganic bases to form ester enolates.⁵⁹In

computational modeling of such ester hydrogenations with

Noyori-type M/NH catalysts,^{17h,18b,35a,47a,56,58}the substrate is

usually used as is. We posited if this indeed was the case under

high base concentrations. Figure 8 shows DFT-computed

physical and chemical “activation” of methyl hexanoate E1

using KO-t-C₄H₉at the ωB97X-D/def2-QZVP/SMD-

(methylbutanoate, EPS = 4.70) level.⁶⁰

427

https://dx.doi.org/10.1021/acs.oprd.9b00559

Org. Process Res. Dev. 2020, 24, 415−442

Organic Process Research & Development

pubs.acs.org/OPRD

Article

Figure 10. Intrinsic reaction coordinates (IRCs) for the outer-sphere hydride transfer from int0a and int1a to physically or chemically activated E1.

Computed at the ωB97X-D/def2-SVP//def2-QZVP/SMD (methylbutanoate, EPS = 4.70) level. For the optimized geometries of the transition states,

noncritical H-atoms are omitted for clarity. Dotted lines represent noncovalent interactions. All energies are calibrated with respect to int0a.

Physical Activation. Starting ester E1 exists in the form of anti-

E1 and syn-E1 conformers (blue bars, Figure 8), of which the

former is expected^58,61and computed to be energetically more

favorable, ΔG_298K° = −4.5 kcal·mol⁻¹. Both conformers react

kcal·mol⁻¹), which is actually stabilized by dipole−dipole

C^(δ+)O^(δ−)···O^(δ−)−K^(δ+)interactions, green dotted lines in

Figure 8.⁶²Formation of this adduct is accompanied with the

increase of natural bond orbital (NBO) charge of the carbonyl

carbon atom from 0.79 (in anti-E1) to 0.85, resulting in a

physical activation of the CO group of E1(enhanced

electrophilicity).

with KO-t-C₄H₉to aﬀord Lewis acid adducts LA1−3 (ΔG_298K

°

= −0.5 to 3.1 kcal·mol⁻¹with respect to anti-E1), which are

stabilized primarily because of cation K⁺−oxygen lone pair(s)

electrostatic interactions (violet bars, Figure 8). Of note is

thermodynamically favorable complex LA1 (ΔG_298K° = −0.5

Chemical Activation. KO-t-C₄H₉can further react with E1

chemically to aﬀord tert-butyl hexanoate esters anti-^tBuH and

428

https://dx.doi.org/10.1021/acs.oprd.9b00559

Org. Process Res. Dev. 2020, 24, 415−442

Organic Process Research & Development

pubs.acs.org/OPRD

Article

syn-^tBuH (orange bars, Figure 8). Their formation, however, is

thermodynamically unfavorable (ΔG_298K° = 3.2−10.3 kcal·

mol⁻¹with respect to anti-E1, respectively).⁶³Moreover, these

esters are signiﬁcantly bulkier than E1, and therefore are unlikely

to participate in the actual catalytic reaction. Ester E1 bears an α-

C−H proton, and KO-t-C₄H₉can produce enolates en1−2/

KOMe (four isomers located, gray bars). Their formation is

thermodynamically unfavorable (ΔG_298K° = 6.2−12.8 kcal·

mol⁻¹) but kinetically accessible. Enolate E-syn-en1 could also

similar and are shown in Figure 10A, ts1, and ts2. In each case,

the substrate approaches the hydride atom within the Re-face

position and in the form of the anti-E1 conformer, in agreement

with the unfavorable thermodynamics of anti-E1 → syn-E1

described above; there appears to be an N−X···O noncovalent

interaction in each case, and S−Me is located anti with respect to

N−X functionality, obviously to avoid steric repulsion with the

approaching substrate. As expected,^55cthe barrier for the outer-

sphere hydride transfer is lower on the potassium amidate

complex int1a (ts1 vs ts2), computationally by 1.7 kcal·mol⁻¹.

Chemically Activated Ester in the Form of Enolate E-syn-en1 and

Keteneket. Taking into account that enolate E-syn-en1 is placed

only 6.2 kcal·mol⁻¹above anti-E1 (Figure 8), we have

performed various constrained PES scans to model outer-sphere

hydride transfer to the former from both complexes int0a and

int1a, Figure 10B,C. Two similar pathways leading to the

hexanaldehyde enolate and KOMe, both bound to each other

and the remaining cationic Ru residue via several noncovalent

interactions, were located for each of int0a and int1a complexes,

as shown in Figure 10B. Similarly, in the case of anti-E1 as

described in the previous subsection, the barrier for the outer-

sphere hydride transfer is lower on the potassium amidate

complex int1a (ts3 vs ts4), and in this case, by 8.6 kcal·mol⁻¹.

However, the overall path via enolate E-syn-en1 is kinetically

inaccessible: ts4 is uphill by 26.5 kcal·mol⁻¹relative to ts2.

Therefore, we may conclude that localized reaction channels are

not realized in the actual catalytic reaction. Interestingly, an

exclusive channel was also located for int0a, Figure 10C. In this

channel, enolate E-syn-en1 ﬁrst deprotonates N−H function-

ality via ts5, leading to int2. The latter represents noncovalently

bound complex int1c and syn-E1 (cf. Figures 8 and 9). Outer-

sphere hydride transfer via ts6 aﬀords the corresponding salt of

potassium hemiacetal bonded to the 16e⁻Ru amido complex

int3′ (cf. int1c). This path, however, is not realized because of

kinetic reasons: ts6 is uphill by 7.7 kcal·mol⁻¹relative ts2.

Finally, we considered ket as a possible equivalent entry to E1,

Figure 10D. As expected, this path is too high in energy with

respect to the path involving inactivated E1 (cf. Figure 10A).

The outer-sphere hydride transfer to ket produces the same

intermediate as E-syn-en1: hexanaldehyde enolate bonded to the

catalyst residue (cf. Figure 10B). Interestingly, although ket is

placed signiﬁcantly higher in energy with respect to E-syn-en1

(Figure 8), outer-sphere hydride transfer to ket is more

favorable. Nevertheless, all these paths seem to be too high in

energy, and we have concluded that ester E1 enters the cycle in

the chemically inactivated form (Figure 10A). In addition, based

on the results reported in the previous subsections, we will

assume that under high KO-t-C₄H₉concentrations, all trans-

formations take place at potassium amidate complex int1a

(Figures 9 and 10), which represents a chemically activated form

of int0a.

produce ketene ket (red bar).⁶⁴This process seems to be too

⧧

unfavorable, both kinetically and thermodynamically (ΔG_298K

°

= 29.7 and ΔG_298K° = 27.8 kcal·mol⁻¹, respectively).

2.1.7.2. Feasibility of Potassium Amidate Complex

Formation. In order to evaluate possible chemical activation

of the trans-Ru dihydride complex (active form of the Ru-2a

precatalyst)^18bint0 by KO-t-C₄H₉, we have optimized geo-

metries and compared the relative energies of various isomers of

potassium amidate complexes int1a−1f, Figure 9.^55b,cFor int0

and each of the located isomeric complexes int1a−1f, both a/b

diastereomers were considered in our calculations. They diﬀer

by the absolute conﬁguration on the sulphur atom, whereas the

absolute conﬁguration on the N(H) atom is S and will be ﬁxed

from now on. The computed diﬀerences in stabilities of various

amidate complexes are typically within −1.9 to 9.3 kcal·mol⁻¹

relative to int0a and KO-t-C₄H₉/HO-t-C₄H₉. The computed

energy range is similar to the Noyori catalyst,^55cfor which such

complexes were detected experimentally.^55b

Although an N−K bond is inherently ionic, and one could

expect some degree of its dissociation into various ion-pairs,

calculations shown in Figure 9 suggest that K⁺prefers to stay

bound to N⁻within the diastereomeric complexes int1a and

int1b. One should understand that in real solution, dynamics is

operative, as well as various speciﬁc and nonspeciﬁc solvations

which could be present from both the solvent and/or excess of

KO-t-C₄H₉, and the N−K dissociation could be subject to an

equilibrium process. Nevertheless, it is reasonable to assume

that K⁺remains in close proximity to N⁻(tight ion-pair N⁻···K⁺)

in a productive catalytic cycle, at least based on results in Figure

9. Such an approach was used by us^55c,58and others^56,65in the

static DFT computational modeling.

2.1.7.3. Outer-Sphere Hydride Transfer to E1: Does the

Substrate Enter the Cycle in a Chemically or Physically

Activated Form? The possibility of hydride transfer from the

metal to carbon atom of the carbonyl moiety of a substrate

without its preliminary coordination is one of the central

mechanistic components of Noyori M/NH catalysts.⁹Previous

studies demonstrated that the role of N−X functionality (X =

H,^18b,47aK⁵⁶) is to stabilize the corresponding transition states

via N−X···O noncovalent interactions, and, in contrast to

stoichiometric processes, the N−X functionality seems to

remain intact during the catalytic processes.

Inactivated Ester. There are eight possibilities for outer-sphere

hydride transfer from each of the complexes int0a and int1a to

the carbon atom of the carbonyl functionality in E1. These

originate from two conformers of E1 (anti and syn), two

possibilities for substrate approach (Re- and Si-face), and two

diastereomeric forms of the catalyst.

There are multiple transition states for each possible hydride

transfer scenario originating from numerous possible geo-

metrical combinations within the CSC along the Ru−H−C

vector. We performed numerous constrained PES scans for each

of the eight possibilities and each of the int0a and int1a

complexes. The most stable transition states were found to be

Physical Activation of Ester by 1 equiv of KO-t-C₄H₉. Dipole−

dipole complex LA1 is an interesting possible catalytic cycle

entry for E1 to consider in modeling. Indeed, its formation from

anti-E1 and KO-t-C₄H₉is not only slightly thermodynamically

favorable (ΔG_298K° = −0.5 kcal·mol⁻¹) but also results in

enhanced electrophilicity of the carboxyl carbon atom as

discussed above. Similar to E1, there are multiple transition

states for a hydride transfer scenario originating from numerous

possible geometrical combinations within the catalyst-LA1

complex along the Ru−H−C vector. A random scan resulted in

the location of the transition state ts2·base, Figure 10E. This

structure is stabilized by several noncovalent interactions, and

429

https://dx.doi.org/10.1021/acs.oprd.9b00559

Org. Process Res. Dev. 2020, 24, 415−442

Organic Process Research & Development

pubs.acs.org/OPRD

Article

overall is placed 1.9 kcal·mol⁻¹lower than ts2! It is therefore

obvious that under higher KO-t-C₄H₉concentration, equili-

brium anti-E1 ↔ LA1 is shifted to the right, and LA1 could be

the actual point of entry of E1.

2.1.7.4. Eﬀect of Explicit “Solvation” of KO-t-C₄H₉on the

Catalytic PES for Ester → Aldehyde. After learning that E1

likely enters the cycle in the physically “activated” form of LA1 as

described in the previous subsection, we aimed to understand

the eﬀect of the remaining KO-t-C₄H₉molecules on the PES for

the ester → aldehyde catalytic cycle. We recall here that the

hydrogenation of esters lacking electron-withdrawing groups

into primary alcohols with M/NH Noyori-type catalysts can be

partitioned into two distinct catalytic cycles: ester → aldehyde

and aldehyde → alcohol.⁹The accelerative role of the base in the

latter cycle is well understood and seems to rely solely on the

formation of amidate complexes.^55cHere, we would like to

understand the eﬀect of explicit solvation of one equivalent of

KO-t-C₄H₉on the former cycle, the details of which were

elaborated in the work of Gusev under base-free conditions.^47a

We note here that with the exception of Gusev’s work, there

were other computational investigations of the mechanism of

ester hydrogenation with M/NH Noyori-type catalysts.^35a,56,66

However, all of these reports implicitly assumed involvement of

the N−H functionality in the catalytic reaction via a reversible

proton transfer following Noyori’s initial hypothesis. In addition,

some of those works incorporated an artifact of the model that

did not include solvent eﬀects, leading to an irreversible mental

“trap”, further enhancing this implicit assumption.⁶⁷We would

like to re-emphasize here that the mechanism of ketone, ester,

and so forth hydrogenation with M/NH Noyori-type catalysts in

which the N−H functionality remains intact is not only a

possible alternative but is energetically more feasible.^9,68

The “free” KO-t-C₄H₉could be placed anywhere in the

model, which might introduce artifacts in the computations

because each orientation has a diﬀerent energy. In order to

partially address this issue, we performed further constrained

PES scans from stationary point int3·base (Figure 10), the

geometry of which resulted from forward intrinsic reaction path

calculations of the ts2·base, obtained in its turn from int1a and

dipole−dipole complex LA1 as discussed above. This approach

is not perfect also because of the fact that only 1 equiv of KO-t-

C₄H₉is used to understand its accelerative role, but a better

solution does not seem to be available for static DFT modeling

based on computing resources used and/or parallelization issues

of the employed Gaussian code. Figure 11 compares the eﬀect of

one KO-t-C₄H₉equivalent “solvation” of the CSC along the

computed catalytic cycle, leading from anti-E1 to hexanaldehyde

and catalyst regeneration with H₂.

Figure 11. Eﬀect of explicit “solvation” of 1 equiv of KO-t-C₄H₉on the

catalytic cycle RCO₂CH₃→ RC(O)H with the Ru-2a precatalyst, R =

C₅H₁₂. Computed at the ωB97X-D/def2-SVP//def2-QZVP/SMD-

(methylbutanoate, EPS = 4.70) level. For the optimized geometry of

int5·base, noncritical H-atoms are omitted for clarity. All energies are

calibrated with respect to int0a/anti-E1/KO-t-C₄H₉.

For the case where E1 enters the cycle in the form of dipole−

dipole complex LA1, the reaction starts with the outer-sphere

hydride transfer from int1a to LA1 via transition state ts2·base

⧧

to aﬀord ion-pair int3·base, ΔΔG_298K

°

= 14.6 kcal·mol⁻¹,

ΔΔG_298K° = 8.7 kcal·mol⁻¹.

shown in Figure 11, is therefore the bona ﬁde catalytic

intermediate for ester hydrogenation, in contrast to ketone

hydrogenation with Noyori-type catalysts where such species

can be only oﬀ-loop species.^55c

Its rearrangement via rotation along the C−O⁻vector

through transition state ts3·base aﬀords isomeric ion-pair

⧧

int4·base, ΔΔG_298K

°

= 1.0 kcal·mol⁻¹, ΔΔG_298K° = −1.2

kcal·mol⁻¹. Hexanaldehyde is further obtained from int4·base

via transition state ts4·base, in which simultaneous cleavage of

C−O and another Ru−O bond formation takes place to aﬀord

int5·base. This process takes place with the relative activation

barrier of 4.9 kcal·mol⁻¹and is extremely thermodynamically

favorable, ΔΔG_298K° = −24.6 kcal·mol⁻¹. “Neutral” alkoxide

ruthenium complex int5·base, the optimized ﬁgure of which is

The following step consists of Ru−O heterolytic bond

cleavage in int5·base via transition state ts5·base to aﬀord ion

pair int6·base.⁶⁹This process takes place with the relative

activation barrier of 6.0 kcal·mol⁻¹and is thermodynamically

unfavorable, ΔΔG_298K° = −4.1 kcal·mol⁻¹. Ion pair int6·base

further coordinates H₂via transition state ts6·base to aﬀord

⧧

dihydrogen-complex int7·base, ΔΔG_298K

°

= 7.9 kcal·mol⁻¹

430

https://dx.doi.org/10.1021/acs.oprd.9b00559

Org. Process Res. Dev. 2020, 24, 415−442

Organic Process Research & Development

pubs.acs.org/OPRD

Article

Figure 12. Optimized geometries and activation barriers for the outer-sphere hydride transfer from various potassium amidate complexes to methyl

hexanoate E1. Computed at the ωB97X-D/def2-SVP//def2-QZVP/SMD(methylbutanoate, EPS = 4.70) level, noncritical H-atoms are omitted for

clarity. All free energies are calibrated with respect to parent N−H complexes, 1 M.

and ΔΔG_298K° = −3.5 kcal·mol⁻¹. Cleavage of the H−H bond in

int7·base via deprotonation by the methoxide anion takes place

with an extremely small barrier of 2.2 kcal·mol⁻¹(transition state

ts7·base) to aﬀord int8·base, which represents the regenerated

catalyst bound to methanol and KO-t-C₄H₉(not shown in the

ﬁgure for simplicity, see the Supporting Information). The

overall catalytic PES curve is shown in orange color in Figure 11.

As expected, the hydrogenation of E1 into hexanaldehyde and

methanol is slightly thermodynamically unfavorable (ΔG_298K° =

3.9 kcal·mol⁻¹), but the former should be further rapidly

hydrogenated into 1-hexanol via a catalytic cycle established for

the Noyori catalyst.^55c

For the case when E1 enters the cycle as it is, similar stationary

points were located on the catalytic PES, blue curve in Figure 11.

The diﬀerence being their relative position, they are located

signiﬁcantly higher on the PES, especially so for the second part

of the reaction after liberation of hexanaldehyde (int5·base →

int1a). When one KO-t-C₄H₉molecule is present in

calculations, it modiﬁes the geometries of stationary points,

which results in better stabilization. In particular, complex int5·

base is signiﬁcantly more stable than int5 (ΔΔG_298K° = −22.6

kcal·mol⁻¹), and this gain in energy is largely due to multiple

ionic interactions present in int5·base, that is, we note a perfect

tetrahedral environment around the oxygen atom of the −OCH₃

ligand, Figure 11. In reality, more than 1 equiv of KO-t-C₄H₉

could be involved in the speciﬁc solvation of CSC, and one could

expect even slightly deeper lying stationary points on the PES,

but the eﬀect of this 1 equiv seems to be the most profound. We

therefore conclude that base excess accelerates the reaction both

via chemical modiﬁcation of the catalyst (int0a → int1a) and via

shifting equilibria toward LA1 and/or speciﬁcally solvating

CSC. Our results are in line with our previous computationally

unveriﬁed assumption⁵⁸as well as the recent computational

studies of Pidko who concluded that “the base not only acts as a

reagent in the conversions...but it also (at least at high

concentrations) acts as a solvent, aﬀecting thus nonlinearly the

energetics of the elementary steps”.⁵⁶Our studies suggest that

base is particularly important in the second half of the ester →

aldehyde catalytic cycle.

K bond remains intact but plays an important role in stabilizing

transition states.

2.1.7.5. Toward Eﬃcient Molecular Catalyst Design: the

Origin of Nonlinear Activity for All Three Catalysts. Is there

further interest to replace the CO ligand in the Ru-MACHO

complex with PPh₃? The answer to this question has already

been given in an independent patent of Takasago Int. Corp.²⁴

The catalytic activity increases in the order PNP ≪< SNS ≤

SNP under the conditions employed (S/C = 1000, 6 h, 80 °C,

THF), Table S10. This trend on average mimics most of our

results with Ru-MACHO but not all. Based on results of our

work, catalytic eﬃciency (total conversion) depends largely on

two factors: nature of the ester and presence/absence of the

solvent (Tables 2 and 3). For example, the catalytic activity

increases in the order Ru-MACHO ≪< Ru-SNS ≈ Ru-2a in the

hydrogenation of fatty and/or functionalized esters E1−E4

under neat conditions, but the trend becomes Ru-MACHO ≪<

Ru-SNS < Ru-2a in the presence of a solvent. Is it possible to

design a catalyst computationally? For example, the activation

barriers for the outer-sphere hydride transfer⁷¹from potassium

amidate complexes of Ru-SNS, Ru-2a, and Ru-MACHO to

methyl hexanoate E1 are indistinguishable taking into account

the standard 3 kcal·mol⁻¹accuracy of DFT,⁷²Figure 12. To

recall, experimentally (Table 2, runs 1−3) E1 is hydrogenated

by following the order Ru-MACHO ≪< Ru-SNS ≈ Ru-2a, but

the rate diﬀers only by a factor of ∼3 in favor of Ru-SNS or Ru-

2a. Computations only suggest that replacement of the CO

ligand with PPh₃in Ru-MACHO complex could be prohibitive

(Figure 12), and this is what is observed experimentally.²⁴What

makes Ru-MACHO less eﬃcient in these reactions? One can

imagine that KO-t-C₄H₉can solvate the corresponding CSC via

cation−π interactions that are caused by the numerous phenyl

rings of the catalyst. Indeed, based on the example of E1 (Figure

13 vs Figure 12), the overall solvation increases the relative

barrier from 14.3 to 18.7 kcal·mol⁻¹, consistent with the

destabilization of the resultant organometallic cation.

However, hydrogenation of substrate E6 in methanol follows

Ru-SNS ≈ Ru-2a < Ru-MACHO. Similarly, thermodynamic

hydrogenation of α-ﬂuorinated esters FE1, FE3, and FE5, as

well as carboxamide CA1, follows the order Ru-SNS < Ru-2a <

Ru-MACHO, whereas for FE2 and FE4, it becomes Ru-SNS <

Ru-2a ≈ Ru-MACHO in the presence of 25 mol % KO-t-C₄H₉

or NaOCH₃. It is therefore not possible to computationally

In summary, the catalytic cycle with or without KO-t-C₄H₉in

Figure 11 is qualitatively similar to the one established in the

work of Gusev based on N−H functionality under base-free

conditions.^47aThe diﬀerence is that the N−H covalent bond is

replaced with N−K ionic one. Note that similar to N−H, the N−

431

https://dx.doi.org/10.1021/acs.oprd.9b00559

Org. Process Res. Dev. 2020, 24, 415−442

Organic Process Research & Development

pubs.acs.org/OPRD

Article

physical site-speciﬁc activation of the CSC through nonspeciﬁc

“solvation”. In particular, the eﬀect of this “solvation” is

particularly manifested for the second half of the catalytic half-

reaction ester → aldehyde, in which the dihydride intermediate

is regenerated from the neutral alkoxide complex through its

reaction with H₂. This computational observation may provide

insight as to why a maximum turnover is observed for esters

under signiﬁcantly larger base loadings than what is required for

ketones: the neutral alkoxide complex is not an intermediate of

ketone hydrogenation but rather an oﬀ-loop species. The

existence of a neutral alkoxide complex of the type int5 on the

catalytic reaction PES for ester hydrogenation as a bona ﬁde

intermediate requires additional KO-t-C₄H₉for the faster

regeneration of the dihydride-catalyst complex.

Ligands 1a−e and 2a−c,e and catalyst-complexes Ru-1a−c,e

and Ru-2a described in this work are soon to be commercially

available from Strem Chemicals, Inc (https://www.strem.com/

catalog/pl/).

Figure 13. Optimized geometry and activation barrier for the outer-

sphere hydride transfer for the potassium amidate complex of Ru-

MACHO to methyl hexanoate E1 in the presence of three KO-t-C₄H₉

“molecules”. The dotted gray lines correspond to cation−π

interactions. Computed at the ωB97X-D/def2-SVP//def2-QZVP/

SMD(methylbutanoate, EPS = 4.70) level, noncritical H-atoms are

omitted for clarity. All energies are calibrated with respect to parent N−

H complexes.⁷⁰

4. EXPERIMENTAL SECTION

4.1. Materials and Analytical Techniques. Diphenylvi-

nylphosphine (95%), H₂O₂(30 wt % in H₂O), 2-bromoethyl-

amine hydrobromide (99%), 2-chloroethylamine hydrochloride

(99%), sodium ethoxide (95%), sodium thiomethoxide (95%),

thiophenol (97%), benzyl mercaptan (99%), and triphenylme-

thanethiol (97%) were purchased from Sigma-Aldrich and used

as received. MeSCH₂CH₂NH₂, PhSCH₂CH₂NH₂, and

^tBuSCH₂CH₂NH₂were purchased from Enamine or prepared

using modiﬁcations of literature procedures (vide infra).

design a universally eﬃcient catalyst, as catalytic activity depends

on the speciﬁc reaction conditions used.

3. CONCLUSIONS

28

BnSCH₂CH₂NH₂was prepared as described elsewhere. All

We report the synthesis, characterization, solution behavior, and

the hydrogenation of oleﬁnic, chiral, and α-ﬂuorinated esters by

newly engineered Ru-SNP(O)_zmolecular catalysts. Compara-

tive experiments suggest that economically viable catalyst Ru-2a

may overwhelm parent Ru-MACHO and Gusev’s Ru-SNS

complexes in the hydrogenation of several esters in terms of

activity and/or selectivity. In particular, the advantage of the Ru-

2a catalyst was demonstrated in the hydrogenation of α-

ﬂuorinated esters into the corresponding ﬂuoral hemiacetals

under kinetically controlled conditions. This work also shows,

for the ﬁrst time, the potential to hydrogenate α-ﬂuorinated

carboxamides into, at a minimum, the corresponding ﬂuoral

hemiacetals, valuable synthons for organic syntheses. The work

provides further insights into the accelerative role of KO-t-C₄H₉

excess in catalytic ester hydrogenations. In the beginning or later

stages of the reaction, the base might indeed destroy impurities

of H₂O and CA present in commercial ester and/or regenerate

the active catalytic species from organometallic products caused

by these impurities. Additionally, in the beginning of the

reaction, base does indeed chemically react with the ester

possessing α-CH-functionality to aﬀord an (experimentally

observed) equilibrated amount of ester enolates. None of these

reactions is the origin for acceleration, however. Ester enolates

are further removed from the system via reversible equilibrium

with the starting material upon turnover. Consistent with this

statement is the fact that products of possible ester enolate

reactions are not observed, even if large base excess is employed.

New experimental observations and computational analysis of

the catalytic cycle suggests that a large KO-t-C₄H₉excess (up to

25 mol % experimentally from titration) is needed for two main

reasons. Our studies suggest that the base accelerates the

reaction via (1) chemical modiﬁcation of the catalyst as

established for ketones (N−H → N⁺K⁻ N−K); and (2)

solvents for aerobic organic syntheses were purchased from

Sigma-Aldrich and used as received in air in a fume hood. All

syntheses of organometallic complexes were performed in an

MBraun MB 200B glovebox under argon (<0.1 ppm O₂/H₂O).

Dichloromethane (anhydrous, ≥99.8%, Sigma-Aldrich), 1,2-

dichloroethane (anhydrous, ≥99.8%, Sigma-Aldrich), toluene

(anhydrous, 99.8%, Sigma-Aldrich), THF (anhydrous, ≥99.9%

%, Sigma-Aldrich), diethyl ether (anhydrous, ≥99.7%, Sigma-

Aldrich), pentane (anhydrous, ≥99%, Sigma-Aldrich), meth-

anol (anhydrous, 99.8%, Sigma-Aldrich), [RuCl₂(PPh₃)₃] (97%,

Sigma-Aldrich), Ru-MACHO (739103 Aldrich), and Gusev’s

Ru-SNS (97%, 746339 Aldrich) were used as received.

Unless otherwise speciﬁed, elemental analysis was performed

by Midwest Microlab, LLC (Indianapolis, IN 46250) alone or in

the presence of V₂O₅(nitrogen atmosphere). Selected elemental

analyses of air-stable compounds were performed in-house using

a Thermo-Finnigan Flash EA 1112 Elemental Analyzer. Samples

sizes of 1.5−2.5 mg were used and rolled in tin cups, which were

then dropped in a combustion column set to 950 °C with the aid

of an autosampler. All NMR experiments were carried out on a

Bruker AV400 MHz spectrometer. ¹H and ¹³C{¹H} and

³¹P{¹H} NMR spectra were calibrated relative to tetramethylsi-

lane and H₃PO₄, respectively, in ppm (δ). ¹⁹F NMR spectra were

measured without lock but properly shimmed in methanol

(relative to CFCl₃). GC−MS analysis was performed using a

Shimadzu GCMS-QP2010 Series spectrometer. For achiral

compounds, a Shimadzu SH-Rxi-5Sil MS column was used (30

m × 0.25 mm ID × 0.25 μm df); for chiral 1,2-propanediol, an

Agilent CP-Chirasil Dex CB column was utilized (25 m × 0.25

mm ID × 0.25 μm df). Chiral HPLC was performed using an

Agilent 1200 Series instrument equipped with a diode array

detector set to 215 nm. Separation was performed using 2%

432

https://dx.doi.org/10.1021/acs.oprd.9b00559

Org. Process Res. Dev. 2020, 24, 415−442

Organic Process Research & Development

pubs.acs.org/OPRD

Article

isopropanol in hexanes as the mobile phase on a Daicel IB N-3

cellulose-tris(3,5-dimethylphenylcarbamate) column (4.6 mm

× 150 mm) using a ﬂow rate of 1 mL/min and a sample

concentration of 1 mg/mL.

4.2.1.4. 2-Aminoethyl Trityl Sulﬁde.²⁷Bromoethylamine

hydrobromide (10.245 g, 50 mmol) was added to a freshly

prepared solution of sodium ethoxide (0.4 M, 125 mL, 100

mmol, 2 equiv). The reaction was stirred for 5 min and the

triphenylmethanethiol added (13.82 g, 50 mmol). The reaction

was stirred for 1 h at room temperature and then heated to reﬂux

overnight. The salts were removed by ﬁltration, and the ﬁltrate

was evaporated to dryness. The crude material was suspended in

water and extracted with three portions of ethyl ether. The

combined organics were washed with 1 N NaOH, dried over

Na₂SO₄, ﬁltered, and evaporated. The residual solid was

recrystallized from ethyl acetate/hexanes to yield the tritylth-

iolethylamine as a white solid (11.99 g, 39 mmol, 78%).

4.2. Synthesis. 4.2.1. Synthesis and Characterization of

Building Blocks. 4.2.1.1. Diphenylvinylphosphine Oxide.⁷³

Diphenylvinylphosphine (25 g, 118 mmol) was dissolved in 500

mL of dichloromethane and cooled to 0 °C. Hydrogen peroxide

(13.5 g, >30% w/w, ∼12.2 mL, 120 mmol) was added dropwise

over 30 min, and the colorless solution warmed to room

temperature. After stirring overnight, 100 mL of water was

added to the yellow solution. The organic phase was separated,

the aqueous phase extracted twice with dichloromethane, and

the combined organics dried over Na₂SO₄, ﬁltered, and

evaporated. The residual solid was recrystallized from hexanes

to yield phosphine oxide (24.572 g; 107.7 mmol; 91%) as a

white solid.

4.2.2. Synthesis and Characterization of SNPO Ligands

1a−e (Aerobic Conditions): Aza-Michael Addition of 2-

Aminoethyl Alkyl Sulﬁdes to Diphenylvinylphosphine Oxide.

The reactions were monitored using TLC and optimized to

reduce double addition to the amine. For 2-aminoethyl methyl

sulﬁde, the reaction can be carried out in pure H₂O; however, a

mixture of ethanol in water is needed to maximize the yield of

other 2-aminoethyl alkyl sulﬁdes.

4.2.1.2. 2-Aminoethyl Methyl Sulﬁde.^26b,dBromoethyl-

amine hydrobromide (10.245 g, 50 mmol) was added to a

freshly prepared solution of sodium ethoxide (0.4 M, 62.5 mL,

50 mmol, 1 equiv). The reaction was stirred for 5 min and then

sodium thiomethoxide added (3.505 g, 50 mmol). The reaction

was stirred for 1 h at room temperature and then heated to reﬂux

overnight. The salts were removed by ﬁltration and the solvent

removed under vacuum. The residue was suspended in water

and extracted with three portions ethyl ether. The combined

organics were washed with 1 N NaOH, then dried over Na₂SO₄,

ﬁltered, and evaporated. The residual oil is distilled using a

Kugelrohr (100 mbar/110 °C) to yield the methylthioethyl-

amine (3.291 g, 36.15 mol, 72%) as a pale yellow oil.

4.2.2.1. (2-((2-(Methylthio)ethyl)amino)ethyl)-

diphenylphosphine Oxide, 1a. Method A. A mixture of

diphenylvinylphosphine oxide (1 g, 4.38 mmol) and 2-

(methylthio)ethylamine (1.1 equiv, 448 μL, 4.82 mmol, 97%

Aldrich) was stirred in 10 mL of water at 100 °C for 6 h. The

reaction was cooled and stirring ceased, resulting in phase

separation of the reaction mixture to an orange organic phase

and transparent water phase. The aqueous phase was extracted

with dichloromethane (2 × 15 mL), and the combined organics

were dried over MgSO₄, ﬁltered, and evaporated on a rotary

evaporator (1 h at 60 °C). The yellow-orange viscous oil

comprised product 1a (>88%), tertiary amine (>9%), and traces

of the starting material (δ 28.0 ppm) according to ³¹P{¹H} NMR

(1250 mg). The crude product was puriﬁed by ﬂash column

chromatography (8.5 × 5.0 cm, SiO₂230−400 mesh, 40−63 μ,

av. pore diameter 60 Å, Sigma, ca. 120 g, CHCl₃/methanol,

100:13, ca. 500 mL, R_f= 0.31 for the product, R_f= 0.63 for the

tertiary amine, TLC Baker-Flex silicagel IB-F). First fraction:

tertiary amine N(CH₂CH₂SMe)(CH₂CH₂P(O)Ph₂)₂, yield

after pentane trituration (2 × 10 mL): 111 mg, oﬀ-white solid

Elem. Anal. Calcd for C₃₁H₃₅NO₂P₂S (547.63): C, 67.99; H,

6.44; N, 2.56%. Found: C, 67.87; H, 6.32; N, 2.49%. ³¹P{¹H}

(162 MHz, CDCl₃, rt): δ 30.7 (s). ¹³C{¹H} (100.5 MHz,

CDCl₃, 25 °C): δ 15.9 (s, 1C), 27.0 (d, J_C−P= 70 Hz, 2C), 31.8

4.2.1.3. 2-Aminoethyl Phenyl Sulﬁde.^26a,cBromoethylamine

hydrobromide (10.245 g, 50 mmol) was added to a freshly

prepared solution of sodium ethoxide (0.4 M, 125 mL, 100

mmol, 2 equiv). The reaction was stirred for 5 min, followed by

addition of thiophenol (5.5 g, 50 mmol). The reaction was

stirred for 1 h at room temperature, and then it was heated to

reﬂux overnight. The salts were removed by ﬁltration, and the

ﬁltrate was evaporated to dryness using a rotary evaporator. The

crude material is suspended in water and extracted with three

portions of ethyl ether. The combined organics were washed

with 1 N NaOH, dried over Na₂SO₄, ﬁltered, and evaporated.

The residual oil is puriﬁed by ﬂash liquid chromatography

(FLC) on a Biotage Isolera (SNAP 25 Silica, 2−10% MeOH in

dichloromethane using linear gradient, product elutes at 12CV)

to yield the target product as pale yellow oil (6.199 g, 40.05

mmol, 81%).

(s, 1C), 45.6 (s, 2C), 56.9 (s, 1C), 128.7 (d, J_C−P= 12 Hz, 8C_meta

,

Ph), 130.7 (d, J_C−P= 9 Hz, 8C_ortho, PPh₃), 131.8 (d, J_C−P= 3 Hz,

4C_para, Ph), 133.0 (d, J_C−P= 99 Hz, 4C_ipso). Second fraction:

product 1a, yield after solvent evaporation and drying (2 h, 60

°C): 1082 mg (78%), transparent yellow oil. Elem. Anal. Calcd

for C₁₇H₂₂NOPS (319.40): C, 63.93; H, 6.94; N, 4.39%. Found

(air, V₂O₅): C, 63.86; H, 6.84; N, 4.37%. ESI-MS, m/z: 319.9

(calcd 320.4 for 1aH⁺). ¹H NMR (400 MHz, CDCl₃, r.t.): δ 1.86

(brs, 1H, NH), 2.05 (s, 3H), 2.52 (m, 2H), 2.58 (vt, ³J_H−H≈ 7

Hz, 2H), 2.77 (vt, ³J_H−H≈7 Hz, 2H), 2.97 (m, 2H), 7.42−7.56

(overlapped, 6H), 7.69−7.81 (overlapped, 4H). ¹³C{¹H} (100.5

433

https://dx.doi.org/10.1021/acs.oprd.9b00559

Org. Process Res. Dev. 2020, 24, 415−442

Organic Process Research & Development

pubs.acs.org/OPRD

Article

MHz, CDCl₃, r.t.): δ 15.3 (s, 1C), 30.5 (d, ¹J_C−P≈ 71 Hz, 1C),

34.2 (s, 1C), 42.6 (s, 1C), 47.8 (s, 1C), 128.7 (d, J_C−P≈ 11 Hz,

4C), 130.7 (d, J_C−P≈ 10 Hz, 4C), 131.8 (d, J_C−P≈2 Hz, 2C),

133.0 (d, ¹J_C−P≈ 99 Hz, 2C). ³¹P{¹H} (162 MHz, CDCl₃, r.t.):

δ 30.9 (s).

Method B. Diphenylvinylphosphine oxide (5.7 g, 25 mmol)

was dissolved in 250 mL of ethanol/water (80%), methyl-

thioethylamine (25 mmol, 2.275 g) was added, and the reaction

heated to reﬂux for 12 h. The solvent was then removed, and the

residual oil was dissolved in 100 mL of dichloromethane. The

organic phase was washed once with water, dried over Na₂SO₄,

ﬁltered, and evaporated. The crude product was puriﬁed by FLC

on a Biotage Isolera (SNAP 25 Silica, 2−10% MeOH in

dichloromethane using linear gradient, product elutes at 15CV)

to yield 1a as a colorless oil (5.49 g, 17.7 mmol, 69%). The

product solidiﬁes into a waxy solid over the course of a week

when refrigerated.

≈ 71 Hz, 1C), 31.0 (s, 3C), 42.1 (s, 1C), 42.6 (s, 1C), 49.2 (s,

1C), 128.7 (d, J_C−P≈ 11 Hz, 4C), 130.7 (d, J_C−P≈ 10 Hz, 4C),

1

131.8 (d, J_C−P≈2 Hz, 2C), 133.0 (d, J_C−P≈ 99 Hz, 2C).

³¹P{¹H} (162 MHz, CDCl₃, r.t.): δ 30.9 (s).

4.2.2.4. (2-((2-(Tritylthio)ethyl)amino)ethyl)-

diphenylphosphine Oxide, 1d. Diphenylvinylphosphine oxide

(5.7 g, 25 mmol) was dissolved in 250 mL of ethanol/water

(80%) followed by the addition of tritylthioethylamine (25

mmol, 8.0 g) and heating to reﬂux for 8 h. The solvent was

removed under vacuum and the residual oil dissolved in 100 mL

dichloromethane. The organic mixture was washed with water,

dried over Na₂SO₄, ﬁltered, and evaporated. The crude product

was puriﬁed by FLC on a Biotage Isolera (SNAP 25 Silica, 0−

10% MeOH in dichloromethane using linear gradient over

20CV, product elutes at 10CV) to yield the target product as a

viscid solid (20.77 g, 82%). The product solidiﬁes upon cooling

to −35 °C. No satisfactory elemental analysis was obtained after

several attempts. ESI-MS, m/z: 548.9, 570.6 (calcd 548.7 for

4.2.2.2. (2-((2-(Benzylthio)ethyl)amino)ethyl)-

diphenylphosphine Oxide, 1b. Diphenylvinylphosphine oxide

(3.4 g, 15 mmol) was dissolved in 250 mL of ethanol/water

(80%), and then benzylthioethylamine (15 mmol, 2.5 g) was

added. The reaction was heated to reﬂux for 12 h, cooled to

room temperature, and evaporated. The residual oil was

dissolved in 100 mL of dichloromethane and washed with

water. After drying of the organic phases over Na₂SO₄, the

solution was ﬁltered and reduced using a rotary evaporator. The

crude material was puriﬁed by FLC on a Biotage Isolera (SNAP

25 Silica, 0−10% MeOH in dichloromethane using linear

gradient for 20CV, product elutes at 12−13CV) to yield the

diphenylphosphine oxide product as an oﬀ-white solid (4.529 g,

11.46 mmol, 67.4%). Elem. Anal. Calcd for C₂₃H₂₆NOPS

(395.50): C, 69.85; H, 6.63; N, 3.54%. Found (air, V₂O₅): C,

69.82; H, 6.42; N, 3.55%. ESI-MS, m/z: 396.7, 418.6 (calcd

1

1dH⁺, 570.7 for 1dNa⁺). H NMR (400 MHz, CDCl₃, r.t.): δ

3

1.60 (br s, 1H, NH), 2.33 (vt, J_H−H≈6 Hz, 2H), 2.39−2.51

(overlapped, 4H), 2.82 (m, 2H), 7.17−7.31 (overlapped, 9H),

7.38−7.58 (overlapped, 12H), 7.70−7.78 (overlapped, 4H).

1

¹³C{¹H} (100.5 MHz, CDCl₃, r.t.): δ 30.4 (d, J_C−P≈ 71 Hz,

1C), 32.1 (s, 1C), 42.4 (s, 1C), 48.0 (s, 1C), 66.6 (s, 1C), 126.6

(s, 3C), 127.9 (s, 6C), 128.7 (d, J_C−P≈ 11 Hz, 4C), 129.6 (s,

6C), 130.7 (d, J_C−P≈ 10 Hz, 4C), 131.8 (d, J_C−P≈2 Hz, 2C),

133.0 (d, ¹J_C−P≈ 99 Hz, 2C), 144.8 (s, 3C). ³¹P{¹H} (162 MHz,

CDCl₃, r.t.): δ 30.9 (s).

4.2.2.5. (2-((2-(Phenylthio)ethyl)amino)ethyl)-

diphenylphosphine Oxide, 1e. Diphenylvinylphosphine oxide

(4.5 g, 20 mmol) was dissolved in 250 mL of ethanol/water

(80%), phenylthioethylamine (20 mmol, 3.06 g) was added, and

the reaction heated to reﬂux for 12 h. The solvent was removed

under vacuum, and the residual oil dissolved in 100 mL of

dichloromethane. The organic mixture was washed with water,

dried over Na₂SO₄, ﬁltered, and evaporated. The residual⁷⁴is

puriﬁed by FLC on a Biotage Isolera (SNAP 25 Silica, 0−10%

MeOH in dichloromethane using linear gradient over 20CV,

product elutes at 16CV) to yield the diphenylphosphine oxide

product as a pale yellow solid (5.95 g, 15.48 mmol, 77%). Elem.

Anal.Calcd for C₂₂H₂₄NOPS (381.47): C, 69.27; H, 6.34; N,

3.67%; Found (air, V₂O₅): C, 69.19; H, 6.20; N, 3.66%. ESI-MS,

m/z: 382.0, 403.9 (calcd 382.5 for 1eH⁺, 404.5 for 1eNa⁺). ¹H

NMR (400 MHz, CDCl₃, r.t.): δ 1.87 (br s, 1H, NH), 2.49 (m,

1

396.5 for 1bH⁺, 418.5 1bNa⁺). H NMR (400 MHz, CDCl₃,

r.t.): δ 2.44−2.65 (overlapped, 5H, NH + 4CH), 2.70 (vt, ³J_H−H

≈6 Hz, 2H), 2.93 (m, 2H), 3.67 (s, 2H), 7.16−7.33 (overlapped,

5H), 7.41−7.58 (overlapped, 6H), 7.70−7.81 (overlapped, 4H).

1

¹³C{¹H} (100.5 MHz, CDCl₃, r.t.): δ 30.6 (d, J_C−P≈ 71 Hz,

1C), 36.2 (s, 1C), 42.5 (s, 1C), 48.0 (s, 1C), 127.0 (s, 1C), 128.5

(s, 2C), 128.7 (d, J_C−P≈ 11 Hz, 4C), 128.8 (s, 2C), 130.7 (d,

J_C−P≈ 10 Hz, 4C), 131.8 (d, J_C−P≈2 Hz, 2C), 133.0 (d, ¹J_C−P

≈

99 Hz, 2C), 138.4 (s, 1C). ³¹P{¹H} (162 MHz, CDCl₃, r.t.): δ

31.1 (s).

4.2.2.3. (2-((2-(tert-Butylthio)ethyl)amino)ethyl)-

diphenylphosphine Oxide, 1c. Diphenylvinylphosphine oxide

(5.7 g, 25 mmol) was dissolved in 250 mL of ethanol/water

(80%) followed by addition of tert-butylthioethylamine (30

mmol, 4.0 g, from Enamine) and heating to reﬂux for 8 h. The

solvent was removed under vacuum, and the residual oil was

dissolved in 100 mL of dichloromethane. The organic mixture

was washed with water, dried over Na₂SO₄, ﬁltered, and

evaporated. The crude product was puriﬁed by FLC on a Biotage

Isolera (SNAP 25 Silica, 0−10% MeOH in dichloromethane

using linear gradient over 15CV, product elutes at 10CV) to

yield the target product as a pale yellow oil (7.67 g, 85%). ESI-

MS, m/z: 362.4, 384.3, 400.3 (calcd 362.5 for 1cH⁺, 384.5 for

1cNa⁺, 400.6 for 1cK⁺). No satisfactory elemental analysis was

3

2H), 2.78 (vt, J_H−H≈6 Hz, 2H), 2.89−3.05 (overlapped, m,

4H), 7.12−7.37 (overlapped, 5H), 7.40−7.55 (overlapped, 6H),

7.68−7.81 (overlapped, 4H). ¹³C{¹H} (100.5 MHz, CDCl₃,

r.t.): δ 30.4 (d, ¹J_C−P≈ 71 Hz, 1C), 34.0 (s, 1C), 42.5 (d, ²J_C−P

≈2 Hz, 1C), 48.0 (s, 1C), 126.2 (s, 1C), 128.7 (d, J_C−P≈ 11 Hz,

4C), 128.9 (s, 1C), 129.7 (s, 1C), 130.7 (d, J_C−P≈ 10 Hz, 4C),

131.8 (d, J_C−P≈2 Hz, 2C), 133.0 (d, ¹J_C−P≈ 99 Hz, 2C), 135.7

(s, 1C). ³¹P{¹H} (162 MHz, CDCl₃, r.t.): δ 31.0 (s).

4.2.3. Synthesis and Characterization of SNP Ligands 2a−

c and 2e (Anaerobic Conditions). 4.2.3.1. General Procedure.

In the glovebox, to an oven-dried 100 mL Kontes Schlenk

vacuum tube equipped with a magnetic stirring bar was added

the SNPO ligand 1 (3.15 mmol), 20 mL MeCN, and degassed

NEt₃(31.5 mmol, ∼4.4 mL). The mixture was stirred until full

dissolution occurred and then removed from the glovebox (if

necessary, sonication was applied). The solution was cooled to 0

°C, and then, HSiCl₃(26.8 mmol, ∼2.7 mL) was added

dropwise under argon. The mixture was heated to reﬂux for 2 h,

and then it was cooled to room temperature. The ³¹P NMR

1

obtained after several attempts. H NMR (400 MHz, CDCl₃,

r.t.): δ 1.30 (s, 9H), 1.66 (brs, 1H, NH), 2.52 (m, 2H), 2.63 (vt,

³J_H−H≈7 Hz, 2H), 2.78 (vt, ³J_H−H≈7 Hz, 2H), 2.97 (m, 2H),

7.38−7.58 (overlapped, 4H), 7.67−7.83 (overlapped, 6H).

¹³C{¹H} (100.5 MHz, CDCl₃, r.t.): δ 28.6 (s, 1C), 30.6 (d, ¹J_C−P

434

https://dx.doi.org/10.1021/acs.oprd.9b00559

Org. Process Res. Dev. 2020, 24, 415−442

Organic Process Research & Development

pubs.acs.org/OPRD

Article

reveals full conversion of the starting material. The reaction was

cooled to 0 °C and quenched by dropwise addition of NaOH in

degassed water (280 mmol, ∼11 g in 30 mL of distilled H₂O;

36% w/v). Toluene (30 mL) was added and the layers stirred

and separated using a cannula. The aqueous layer was again

extracted with toluene (2 × 15 mL), and the combined organics

was washed with degassed brine (60 mL). Toluene (2 × 15 mL)

was added to brine and the layers stirred and separated. All

organics were combined, dried over MgSO₄, ﬁltered in the

glovebox, and the solvent evaporated to yield the ﬁnal product.

4.2.3.2. (2-((2-(Methylthio)ethyl)amino)ethyl)-

diphenylphosphine, 2a. The product was obtained in 92%

yield (0.88 g), yellow oil. Elem. Anal. Calcd for C₁₇H₂₂NPS

(303.40): C, 67.30; H, 7.31; N, 4.62%. Found (nitrogen, V₂O₅):

C, 67.29; H, 7.17; N, 4.55%. ¹H NMR (400 MHz, CDCl₃, r.t.): δ

1.84 (brs, 1H, NH), 2.10 (s, 3H), 2.32 (vt, ³J_H−H≈7 Hz, 2H),

138.2 (d, J_C−P≈ 12 Hz, 2C). ³¹P{¹H} (162 MHz, CDCl₃, r.t.):

δ−20.7 (s).

4.2.4. Synthesis and Characterization of Ru-SNPO Com-

plexes Ru-1a−1c and Ru-1e (Glovebox, Argon Atmosphere).

4.2.4.1. General Procedure. A solution of ligand 1 (1.5 mmol or

1.65 mmol for 1c) in 15 mL of dichloromethane was added to

1440 mg (1.5 mmol) of RuCl₂(PPh₃)₃in a scintillation vial

under stirring. The obtained solution was then stirred for 2 h at

room temperature. Depending on the ligand identity (vide

infra), the solution was evaporated to ∼0.7 of volume. The

solution was then transferred to a 120 mL Ace-tube and layered

with 100 mL of diethyl ether. In one week, the mother liquor was

decanted and the crystalline or powder material worked up as

described below. The products are further characterized by X-

ray diﬀraction studies, elemental analysis, and solution-state

NMR. In the NMR spectra, diastereomers or diastereomers sets

(complex Ru-1c) were observed.

3

2.64 (vt, J_H−H≈7 Hz, 2H), 2.81 (overlapped, m, 4H), 7.25−

7.60 (overlapped, 10H). ¹³C{¹H} (100.5 MHz, CDCl₃, r.t.): δ

15.3 (s, 1C), 28.8 (d, ²J_C−P≈ 10 Hz, 1C), 34.1 (s, 1C), 46.2 (d,

¹J_C−P≈ 20 Hz, 1C), 47.6 (s, 1C), 128.5 (d, J_C−P≈6 Hz, 4C),

128.7 (s, 2C), 132.7 (d, J_C−P≈ 20 Hz, 4C), 138.5 (d, J_C−P≈ 12

Hz, 2C). ³¹P{¹H} (162 MHz, CDCl₃, r.t.): δ −20.6 (s). Stability

test: solution of the ligand in CDCl₃(NMR tube, ca. 10 μL in

300 μL) prepared under argon then exposed to air. The NMR

was removed and replaced after 30 s, and the tube left for 43 h.

³¹P{¹H} NMR reveals ∼1% of the phosphine oxidized to 1a after

this time.

4.2.4.2. Ru-1a. A solution of ligand 1a (480 mg, 1.5 mmol) in

15 mL of CH₂Cl₂was added to 1440 mg (1.5 mmol) of

RuCl₂(PPh₃)₃in a scintillation vial with stirring. The obtained

solution was stirred for 2 h at room temperature and evaporated

to ∼0.7 volume. The solution was transferred to a 120 mL Ace-

tube, where it was layered with 100 mL of diethyl ether. In one

week, the mother liquor was decanted and the crystalline

material triturated with diethyl ether (100 mL). The obtained

bright pink powder was collected on a frit, washed with diethyl

ether, and dried under high-vacuum overnight (40 °C) to aﬀord

1009 mg (89%) of the title product. The product is air-stable in

the solid state but decomposes over 24 h in solution under

aerobic conditions. Elem. Anal. Calcd for C₃₅H₃₇Cl₂NOP₂RuS

(753.66): C, 55.78; H, 4.95; N, 1.86%. Found (nitrogen): C,

55.46; H, 4.82; N, 1.55%. ³¹P{¹H} NMR (162 MHz, CD₂Cl₂,

r.t.): δ 44.2 (s, 0.47P), 45.4 (s, 0.53P), 52.3 (s, 0.47P), 52.4 (s,

0.53P). For ¹H and ¹³C{¹H} NMR spectra, see Figure S15.

4.2.4.3. Ru-1b. According to the general procedure (ligand

1b: 592 mg, 1.5 mmol; 1440 mg (1.5 mmol) of RuCl₂(PPh₃)₃in

15 mL of CH₂Cl₂), 1232 mg (99%) of the title product was

isolated as a salmon-colored powder containing ∼4% of PPh₃

according to ³¹P{¹H} NMR. Additional recrystallization from

dichloromethane/diethyl ether aﬀorded a high-purity product

after washing with ethyl ether and drying (1063 mg, 85%). Elem.

Anal. Calcd for C₄₁H₄₁Cl₂NOP₂RuS (829.76): C, 59.35; H,

4.98; N, 1.69%. Found (nitrogen, V₂O₅): C, 59.08; H, 5.02; N,

1.95%.³¹P{¹H} NMR (162 MHz, CDCl₃, r.t.): δ 43.9 (s, 0.45P),

4.2.3.3. (2-((2-(Benzylthio)ethyl)amino)ethyl)-

diphenylphosphine, 2b. The product was obtained in 93%

yield (1.11 g), yellow oil. Elem. Anal. Calcd for C₂₃H₂₆NPS

(379.50): C, 72.79; H, 6.91; N, 3.69%. Found (nitrogen, V₂O₅):

C, 72.45; H, 6.85; N, 3.83%. ¹H NMR (400 MHz, CDCl₃, r.t.): δ

1.50 (br s, 1H, NH), 2.27 (vt, ³J_H−H≈7 Hz, 2H), 2.57 (vt, ³J_H−H

≈7 Hz, 2H), 2.74 (overlapped, m, 4H), 3.72 (s, 2H), 7.11−8.00

(overlapped, 15H). ¹³C{¹H} (100.5 MHz, CDCl₃, r.t.): δ 28.0

(d, ²J_C−P≈ 11 Hz, 1C), 31.6 (s, 1C), 36.2 (s, 1C), 46.2 (d, ¹J_C−P

≈ 21 Hz, 1C), 48.0 (s, 1C), 127.0 (s, 1C), 128.5 (d, J_C−P≈6 Hz,

4C), 128.6 (s, 2C), 128.7 (s, 2C), 128.8 (s, 2C), 132.7 (d, J_C−P

≈

20 Hz, 4C), 138.5 (d, J_C−P≈ 12 Hz, 2C), 138.4 (s, 1C). ³¹P{¹H}

(162 MHz, CDCl₃, r.t.): δ −20.6 (s).

4.2.3.4. (2-((2-(tert-Butylthio)ethyl)amino)ethyl)-

diphenylphosphine, 2c. The product was obtained in 79%

yield (0.86 g), pale yellow oil. Elem. Anal. Calcd for C₂₀H₂₈NPS

(345.48): C, 69.53; H, 8.17; N, 4.05%. Found (nitrogen, V₂O₅):

C, 68.29; H, 7.93; N, 3.77%. ¹H NMR (400 MHz, CDCl₃, r.t.): δ

1.32 (s, 9H), 2.39 (m, 2H), 2.57 (vt, ³J_H−H≈7 Hz, 2H), 2.71−

2.89 (overlapped, m, 6H), 5.00 (br s, 1H, NH), 7.32−7.52

(overlapped, 10H). ¹³C{¹H} (100.5 MHz, CDCl₃, r.t., Csp³-

region): δ 27.4 (s, 1C), 27.6 (d, ²J_C−P≈ 11 Hz, 1C), 30.8 (s, 3C),

42.2 (s, 1C), 45.8 (d, ¹J_C−P≈ 21 Hz, 1C), 48.8 (s, 1C). ³¹P{¹H}

(162 MHz, CDCl₃, r.t.): δ −20.8 (s).

1

45.9 (s, 0.55P), 51.1 (s, 0.55P), 51.3 (s, 0.45P). For H and

¹³C{¹H} NMR spectra, see Figure S16.

4.2.4.4. Ru-1c. A similar procedure was employed, however,

1.1 equiv of the ligand was used (1c: 598 mg, 1.65 mmol; 1440

mg (1.5 mmol) of RuCl₂(PPh₃)₃in 15 mL of CH₂Cl₂), aﬀording

902 mg (76%) of the product as dark-brown crystals. Elem. Anal.

Calcd for C₃₈H₄₃Cl₂NOP₂RuS (795.75): C, 57.36; H, 5.45; N,

1.76%. Found (air): C, 56.91; H, 5.17; N, 1.64%. ³¹P{¹H} NMR

(162 MHz, CD₂Cl₂, r.t.): δ 43.7 (s, 0.05P), 43.9 (s, 0.43P), 44.6

(s, 0.49P), 44.9 (s, 0.05P), 49.9 (s, 0.49P), 52.0 (s, 0.43P), 52.7

(s, 0.05P), 53.2 (s, 0.03P). For ¹H and ³¹P{¹H} NMR spectrum,

see Figures S17 and S18.

4.2.4.5. Ru-1e. A similar procedure (ligand 1e: 573 mg, 1.5

mmol; 1440 mg (1.5 mmol) of RuCl₂(PPh₃)₃in 15 mL of

CH₂Cl₂) aﬀorded 1134 mg (93%) of the product as a crimson

powder containing ∼1% of PPh₃according to ³¹P{¹H} NMR.

Additional recrystallization from dichloromethane/diethyl

ether, followed by subsequent washing with diethyl ether and

drying, yielded analytically pure Ru-1e (781 mg, 64%). Poor

4.2.3.5. (2-((2-(Phenylthio)ethyl)amino)ethyl)-

diphenylphosphine, 2e. The product was obtained in 87%

yield (1.02 g) as a pale yellow solid which melts at room

temperature. Elem. Anal. Calcd for C₂₂H₂₄NPS (365.47): C,

72.30; H, 6.62; N, 3.83%. Found (nitrogen, V₂O₅): C, 72.29; H,

6.50; N, 3.60%. ¹H NMR (400 MHz, CDCl₃, r.t.): δ 1.54 (brs,

3

1H, NH), 2.28 (t, J_H−H≈7 Hz, 2H), 2.77 (vq, J ≈8 Hz, 2H),

2.84 (t, ³J_H−H≈7 Hz, 2H), 3.05 (t, ³J_H−H≈7 Hz, 2H), 7.17−7.50

(15H). ¹³C{¹H} (100.5 MHz, CDCl₃, r.t.): δ 28.8 (d, ²J_C−P≈ 12

Hz, 1C), 33.8 (s, 1C), 46.2 (d, ¹J_C−P≈ 20 Hz, 1C), 47.9 (s, 1C),

126.3 (s, 1C), 128.5 (d, J_C−P≈6 Hz, 4C), 128.7 (s, 2C), 129.0 (s,

2C), 129.8 (s, 2C), 132.8 (d, J_C−P≈ 20 Hz, 4C), 135.6 (s, 1C),

435

https://dx.doi.org/10.1021/acs.oprd.9b00559

Org. Process Res. Dev. 2020, 24, 415−442

Organic Process Research & Development

pubs.acs.org/OPRD

Article

solubility in CD₂Cl₂is noted. Elem. Anal. Calcd for

C₄₀H₃₉Cl₂NOP₂RuS (815.74): C, 58.90; H, 4.82; N, 1.72%.

Found (nitrogen, V₂O₅): C, 58.56; H, 4.81; N, 1.89%. ³¹P{¹H}

NMR (162 MHz, CD₂Cl₂, r.t.): δ 44.4 (s, 0.8P), 44.7 (s, 0.2P),

52.1 (s, 0.8P), 52.3 (s, 0.2P). For ¹H NMR spectrum, see Figure

S19.

entity (11 mg from 96 mg of RuCl₂(PPh₃)₃). Elem. Anal. Calcd

for C₃₈H₄₃Cl₂NP₂RuS (779.75): C, 58.53; H, 5.56; N, 1.80%.

Found (nitrogen, V₂O₅): C, 56.52; H, 5.75; N, 1.57%. ³¹P{¹H}

NMR (162 MHz, CD₂Cl₂, r.t.): δ 40.9 (br s), 41.1 (br s), 44.2

(br s), 44.4 (br s). For ³¹P{¹H} NMR spectrum, see Figure S22.

4.2.5.4. Ru-2e. The product was obtained as yellow powder

(extremely poor solubility in CD₂Cl₂is noted). Prepared

following method B in 80% yield. The product contains some

residual 1,2-dichloroethane solvent. Elem. Anal. Calcd for

C₄₀H₃₉Cl₂NP₂RuS (799.74): C, 60.07; H, 4.92; N, 1.75%.

Found (nitrogen, V₂O₅): C, 59.85; H, 5.00; N, 2.01%. NMR

spectra were not collected due to very poor solubility. Some

signals were observed in the ³¹P{¹H} NMR (162 MHz, CD₂Cl₂,

r.t.): δ 44.0 (s, 0.58P), 44.2 (s, 0.42P), 45.8 (br s, 1P).

4.3. X-ray Structural Analysis. Data were collected at 100

K on either a Bruker Apex II (Ru-1a, Ru-1c, and Ru-1e) or a

Bruker Quest diﬀractometer (Ru-1b, Ru-2a and Ru-2c). Both

instruments were equipped with a graphite monochromatized

Mo Kα X-ray source (l = 0.71073 Å). The Apex II was equipped

with a monocapillary and CCD detector, and the Quest

employed a Triumph curved graphite monochoromator and

Photon CMOS detector. Structure solution, reﬁnement,

graphics, and creation of publication materials were performed

using SHELXTL software. Additional details can be found in

Tables S1 −S7 or in the crystallographic information ﬁle.

4.4. Catalytic Hydrogenation of Esters and Carbox-

amides Mediated by Ruthenium Complexes. All sub-

strates were purchased from Sigma-Aldrich, except (S)-

(+)-methyl mandelate (Combi-Blocks), methyl heptaﬂuorobu-

tyrate (Matrix Scientiﬁc), methyl chlorodiﬂuoroacetate (Oak-

wood Chemical), and methyl pentaﬂuoropropionate (Oakwood

Chemical). CA1 was synthesized according to the previous

literature.⁷⁵Bottles containing solid substrates were opened and

immediately cycled through the antechamber of an argon-ﬁlled

glovebox to remove residual oxygen. Liquid substrates were

degassed either by three freeze−pump−thaw cycles or by

sparging with argon. All reactions mixtures were prepared in an

inert atmosphere glovebox under an argon atmosphere in 50 mL

Parr autoclave reactors containing a borosilicate glass liner and a

7 mm stir bar. In a typical reaction, the catalyst was ﬁrst added

either directly as a solid (≥1.0 mg) or as a dichloromethane

stock solution (<1.0 mg), followed by the addition of the desired

base; when the catalyst was added as a stock solution, the solvent

was removed by gentle heating at 45 °C prior to the addition of

the base. If a solid substrate was used, 10 mmol was added to the

aforementioned mixture, followed by dissolution in the desired

solvent (5 mL). When liquid substrates were employed in these

reactions, the catalyst/base mixture was ﬁrst dissolved in 5 mL of

the desired solvent (where indicated), followed by the addition

of 10 mmol of the substrate with the aid of a calibrated pipette.

The glass insert was then placed in the reactor, followed by

sealing the autoclave and removing from the glovebox. The head

of the reactor was tightened to 25 ft·lbs, and then the system was

connected to a hydrogen gas line. The line was then purged three

times with hydrogen, followed by slowly ﬁlling the Parr reactor

to the desired pressure (25−50 bar). The reactor was placed in

the heating element which was placed atop a magnetic stirring

plate. The reaction was then set to stir at 630 revolutions per

minute, and the reaction slowly heated to the target temperature

using a model 4838 Parr Temperature Controller. At the end of

the reaction, the heating was switched oﬀ, the reactor was cooled

with the aid of an ice water bath (0 °C), and the excess hydrogen

was vented. An aliquot of the reaction mixture was withdrawn

4.2.5. Synthesis and Characterization of Ru-SNP Com-

plexes Ru-2a−2c and Ru-2e (Glovebox, Argon Atmosphere).

4.2.5.1. Ru-2a. Yellow powder (poor solubility in CD₂Cl₂is

noted). ³¹P{¹H} NMR (162 MHz, CD₂Cl₂, r.t.; 2^dorder

spectrum): δ 44.4 (d, ²J_PP= 30 Hz, 1P), 44.7 (d, ²J_PP= 30 Hz,

1.34P), 46.1 (d, ²J_PP= 30 Hz, 1P), 46.6 (d, ²J_PP= 30 Hz, 1.34P).

¹H NMR (400 MHz, 25 °C, CD₂Cl₂): d 1.52 + 1.57 (s, 1H, S−

CH₃), 2.35−2.74 (m, 2H, CH₂), 2.86−3.38 (m, 6H, CH₂× 3),

4.47 + 4.67 (br s, 1H, N−H), 6.80−7.52 (m, 25H, P-Ph).

Method A. To a scintillation vial containing ligand 2a (57 mg,

0.188 mmol) and RuCl₂(PPh₃)₃(180 mg, 0.188 mmol) was

added 2.5 mL of dichloromethane. A homogenous solution

formed after approximately 10 min of stirring; ³¹P{¹H} spectrum

revealed quantitative formation of the product. The solution was

allowed to stir for a few more hours until a microcrystalline

product began to deposit, followed by cooling to −35 °C. The

following day, crystals were collected and washed with pentane

to aﬀord the product as a 1:1 solvate of CH₂Cl₂according to

NMR and X-ray diﬀraction studies. Yield: 88 mg (57%).

Method B. A mixture containing ligand 2a (366 mg, 1.2 mmol,

1.2 equiv), RuCl₂(PPh₃)₃(963 mg, 1 mmol), and 15 mL of 1,2-

dichloroethane was stirred at 90 °C for 2 h (Ace tube). The

yellow precipitate was collected on a frit, washed with cold 1,2-

dichloroethane (20 mL) and pentane, and dried under vacuum

overnight at 40 °C. The product was obtained in 53% yield (391

g). Elem. Anal. Calcd for C₃₅H₃₇Cl₂NP₂RuS (737.67): C, 56.99;

H, 5.06; N, 1.90%. Found (nitrogen, V₂O₅): C, 57.05; H, 5.05;

N, 2.35%.

Method C. To a 120 mL pressure tube in an argon-ﬁlled

glovebox was added 5.01 g (5.22 mmol) RuCl₂(PPh₃)₃followed

by 1.74 g (5.74 mmol) 2a in 65 mL of toluene. The tube was

sealed with a Teﬂon stopper and heated to 120 °C with rapid

stirring for 12 h. After this time, the reaction was cooled to room

temperature and the resulting orange suspension was ﬁltered

over a ﬁne porosity frit. The orange precipitate was rinsed with

three 50 mL portions of toluene and then three 25 mL portions

of n-hexane. The product was then collected and dried under

vacuum overnight at 65 °C, yielding 3.41 g of the desired

product as an orange powder (88% isolated yield). Anal. Calcd

for C₃₅H₃₇Cl₂NP₂RuS (737.67): C, 56.99; H, 5.06; N, 1.90.

Found: C, 56.60; H, 5.20; N, 1.45.

4.2.5.2. Ru-2b. Orange powder (very poor solubility in

CD₂Cl₂is noted). Prepared following method C in 84% yield.

The product contains some residual toluene (∼0.5 equiv)

1

according to H NMR (suspension). Elem. Anal. Calcd for

C₄₁H₄₁Cl₂NP₂RuS (813.76): C, 60.52; H, 5.08; N, 1.72%.

Found (nitrogen, V₂O₅): C, 59.82; H, 5.11; N, 1.58%. ³¹P{¹H}

NMR (162 MHz, CD₂Cl₂, r.t.; 2^dorder spectrum, suspension):

δ 44.5 (d, ²J_PP= 30 Hz, 1P), 44.6 (d, ²J_PP= 30 Hz, 1.33P), 46.0

(d, ²J_PP= 30 Hz, 1.33P), 46.5 (d, ²J_PP= 30 Hz, 1P). For ³¹P{¹H}

NMR spectrum, see Figure S21.

4.2.5.3. Ru-2c. Orange powder. Prepared following method B

in 59% yield. The product is ∼93% pure. Byproduct present was

found to be ∼7% according to the ³¹P{¹H} NMR and

characterized by two 1:1 doublets: 31.0 (d, J_P−P≈ 35 Hz),

51.5 (d, J_P−P≈ 35 Hz). If the reaction is performed in

dichloromethane at 25 °C, the byproduct is the only isolable

436

https://dx.doi.org/10.1021/acs.oprd.9b00559

Org. Process Res. Dev. 2020, 24, 415−442

Organic Process Research & Development

pubs.acs.org/OPRD

Article

and analyzed by NMR and/or GC−MS, alone or in the presence

of an internal standard. When possible, residual KO-t-C₄H₉

served as the internal standard for NMR studies. Chiral GC−MS

or HPLC was used to determine the percentage enantiomeric

excess (% ee). Select representative examples are provided

below.

solution was evaporated and redissolved in a 98:2 mixture of

hexanes/isopropanol (1 mg/mL) and then ﬁltered to remove

any suspended particles. The obtained solution was analyzed by

chiral HPLC using a mobile phase of 98:2 hexanes/isopropanol.

Example 2. Catalytic Hydrogenation of Methyl Diﬂuoroace-

tate by Ru-2a, S/C = 100,000. A stock solution of 0.147534 mg/

mL Ru-2a was prepared, and 0.5 mL of this solution was

transferred to a 50 mL glass liner with a stir bar. The liner was

heated to 45 °C without stirring for approximately 20 min after

all solvent had evaporated. This ensured complete removal of

dichloromethane. To the liner was then added 279 mg KO-t-

C₄H₉(2.49 mmol, 25 mol %) followed by 5 mL of MeOH. The

mixture was stirred until the solution was homogenous, and

then, methyl diﬂuoroacetate was added using a calibrated

pipette (873 μL, 1.10 g, 10 mmol). The glass liner was placed in

the stainless steel autoclave and sealed as described above. After

purging the gas line with dihydrogen, the reactor was pressurized

to 25 bar (∼50 mmol) and then slowly heated to 40 °C. The

reaction mixture was heated with stirring for 8 h, ceased by

cooling in an ice bath, andhydrogen gas was vented. The

solution was analyzed by ¹⁹F NMR spectroscopy to determine

reaction yield and selectivity, Figure 16.

Example 1. Catalytic Hydrogenation of Methyl (S)-

(+)-Mandelate Mediated by Ru-1a, S/C = 2500. To a 50 mL

Example 3. Catalytic Hydrogenation of N,N-diethyl-2,2,2-

triﬂuoroacetamide by Ru-2a, S/C = 5000. To a 50 mL glass liner

with a stir bar was carefully added 1.5 mg of Ru-2a, followed by

136 mg NaOMe (2.51 mmol, 25 mol %). The solids were

dissolved in 5 mL of MeOH and then stirred until complete

dissolution was achieved. To this solution was added 1.410 mL

(1.69 g, 10 mmol) of N,N-diethyl-2,2,2-triﬂuoroacetamide. The

glass liner was placed in the stainless steel autoclave and sealed as

described above. After purging the gas line with dihydrogen, the

reactor was pressurized to 25 bar (∼50 mmol) and then slowly

heated to 40 °C. The reaction mixture was heated with stirring

for 10 h, ceased by cooling in an ice bath, and hydrogen gas was

vented. The solution was analyzed by ¹⁹F NMR spectroscopy to

determine reaction yield and selectivity, Figure 17.

Figure 14. X-ray structures (35% level of thermal ellipsoids) of Ru-1b

and Ru-1e. Noncritical H-atoms are omitted for clarity.

4.5. Computational Analysis. Computations were per-

formed using unabridged models with code Gaussian 09 (rev.

E01),⁷⁶DFT⁷⁷by using hybrid ωB97X-D⁷⁸functional

incorporating Grimme’s D2 dispersion model⁷⁹and the SMD

polarizable continuum model.⁸⁰Because methyl hexanoate is

absent in the list of solvents, it has been approximated using the

following keyword in the route section SCRF=(SMD, solvent =

methylbutanoate, read), which implies default parameters of

available methyl butanoate adjusted by a new custom value of

the dielectric constant (EPS = 4.70) in a separate PCM input

section. Organic reactions (geometry optimization and

frequency calculations) were modeled with the def2-QZVP

basis set⁸¹and an increased integral accuracy, integral-

(ultraﬁnegrid,Acc2E = 12). Geometry optimizations and

frequency calculations for organometallic complexes were

performed by using a def2-SVP basis set⁸¹and an ultraﬁne

grid, integral = ultraﬁnegrid. The standard reaction Gibbs

energies were calculated by combining the single-point def2-

SVP//def2-QZVP energies [integral(ultraﬁnegrid,Acc2E =

12)] with the thermal corrections from frequency calculations

under the def2-SVP level. Frequency calculations were carried

out for all optimized geometries in order to verify their nature as

local minima, under the harmonic approximations, and for the

identiﬁcation of all transition states (one imaginary frequency in

the Hessian Matrix). The Gibbs free energies, G, were calculated

under standard-state conditions of 1 atm (as default for the

continuum model) and then corrected to 1 M (standard-state in

Figure 15. X-ray structures (35% level of thermal ellipsoids) of Ru-1c

and Ru-2c. Noncritical H-atoms are omitted for clarity.

glass liner with a stir bar was added 3.0 mg of Ru-1a (4.07 μmol),

168.9 mg of KO-t-C₄H₉(1.51 mmol, 15 mol %), and 1.662 g of

methyl (S)-(+)-mandelate (10.00 mmol). To this mixture was

added 5 mL of methanol followed by stirring until a

homogenous solution was obtained. The glass liner was placed

in the stainless steel autoclave and sealed as described above.

After purging the gas line with dihydrogen, the reactor was

pressurized to 50 bar (∼100 mmol) and then slowly heated to 40

°C. The reaction mixture was heated with stirring for 18 h and

then ceased by cooling to room temperature and venting the

hydrogen gas. The solution was analyzed by ¹H NMR

spectroscopy (conversion and product yield were determined

with respect to residual KO-t-C₄H₉; rd = 10 s). To determine the

enantiomeric excess of the formed product, an aliquot of the

437

https://dx.doi.org/10.1021/acs.oprd.9b00559

Org. Process Res. Dev. 2020, 24, 415−442

Organic Process Research & Development

pubs.acs.org/OPRD

Article

Figure 16. ¹⁹F NMR (376.5 MHz, r.t.) analysis of run 8 in Table 3: (a) starting methyl diﬂuoroacetate in MeOH; (b) starting methyl diﬂuoroacetate in

the presence of 25 mol % KO-t-C₄H₉in MeOH; (c) catalytic reaction mixture of run 8 in Table 3.

Figure 17. ¹⁹F NMR (376.5 MHz, r.t.) analysis of catalytic hydrogenation of CA1 with Ru-2a (Scheme 5): (a) starting material in MeOH in the

presence of 25 mol % MeONa; (b) reaction performed in the absence of the catalyst; (c) reaction performed in the presence of the ruthenium catalyst

(pressure drops from 26 to 23 bar). Selected ¹⁹F NMR for 1-(diethylamino)-2,2,2-triﬂuoroethan-1-ol: δ −75.9 ppm, d, ³J_F−H= 6 Hz; for Hem1: δ

−83.3 ppm, d, ³J_F−H= 4 Hz.

solution) by adding 0.00301 Hartree. The IRC⁸²calculations

NMR and ESI-MS spectra, X-ray crystallographic data,

were carried out in both directions starting from the located

transition states. Molecular graphic images were produced using

the UCSF Chimera package⁸³or Chemcraft graphical

program.⁸⁴

and further computational details (PDF)

X-ray crystallographic data for Ru-1a, Ru-1b, Ru-1c, Ru-

1e, Ru-2a·CH2Cl2, and Ru-2C (CIF)

AUTHOR INFORMATION

Corresponding Author

■

ASSOCIATED CONTENT

■

sı

* Supporting Information

The Supporting Information is available free of charge at

https://pubs.acs.org/doi/10.1021/acs.oprd.9b00559.

Pavel A. Dub − Chemistry Division, Los Alamos National

Laboratory, Los Alamos, New Mexico 87545, United States;

orcid.org/0000-0001-9750-6603; Email: pdub@lanl.gov

438

https://dx.doi.org/10.1021/acs.oprd.9b00559

Org. Process Res. Dev. 2020, 24, 415−442

Organic Process Research & Development

pubs.acs.org/OPRD

Article

Effect of Morphology of Rhizopus arrhizus NRRL 2582. Fermentation

2017, 3, 33.

Authors

Rami J. Batrice − Chemistry Division, Los Alamos National

Laboratory, Los Alamos, New Mexico 87545, United States

John C. Gordon − Chemistry Division, Los Alamos National

Laboratory, Los Alamos, New Mexico 87545, United States

Brian L. Scott − Materials and Physics Applications Division, Los

Alamos National Laboratory, Los Alamos, New Mexico 87545,

United States

Yury Minko − Biochemistry Division, Los Alamos National

Laboratory, Los Alamos, New Mexico 87545, United States

Jurgen G. Schmidt − Biochemistry Division, Los Alamos National

Laboratory, Los Alamos, New Mexico 87545, United States

Robert F. Williams − Biochemistry Division, Los Alamos

National Laboratory, Los Alamos, New Mexico 87545, United

States; orcid.org/0000-0002-4310-6249

(8) (a) Sun, D.; Yamada, Y.; Sato, S.; Ueda, W. Glycerol as a potential

renewable raw material for acrylic acid production. Green Chem. 2017,

19, 3186−3213. (b) Yang, S.; Kim, M.; Yang, S.; Kim, D. S.; Lee, W. J.;

Lee, H. Production of acrylic acid from biomass-derived allyl alcohol by

selective oxidation using Au/ceria catalysts. Catal. Sci. Technol. 2016, 6,

3616−3622.

(9) Dub, P. A.; Gordon, J. C. The role of the metal-bound N-H

functionality in Noyori-type molecular catalysts. Nat. Rev. Chem. 2018,

2, 396−408.

(10) Yoshioka, S.; Saito, S. Catalytic hydrogenation of carboxylic acids

using low-valent and high-valent metal complexes. Chem. Commun.

2018, 54, 13319−13330.

(11) Ni, Y.; Hagedoorn, P.-L.; Xu, J.-H.; Arends, I. W. C. E.;

Hollmann, F. A biocatalytic hydrogenation of carboxylic acids. Chem.

Commun. 2012, 48, 12056−12058.

Complete contact information is available at:

https://pubs.acs.org/10.1021/acs.oprd.9b00559

(12) (a) Matsumoto, K.; Yanagi, R.; Oe, Y. Recent Advances in the

Synthesis of Carboxylic Acid Esters. Carboxylic Acid: Key Role in Life

Sciences; Badea, G. I., Radu, G. L., Ed.; IntechOpen, 2018. Available

from: https://www.intechopen.com/books/carboxylic-acid-key-role-

in-life-sciences/recent-advances-in-the-synthesis-of-carboxylic-acid-

esters; (b) Li, J.; Sha, Y. A Convenient Synthesis of Amino Acid Methyl

Esters. Molecules 2008, 13, 1111−1119. (c) Nakao, R.; Oka, K.;

Fukumoto, T. A Simple Method for the Esterification of Carboxylic

Acids Using Chlorosilanes. Bull. Chem. Soc. Jpn. 1981, 54, 1267−1268.

(13) Dub, P. A.; Ikariya, T. Catalytic Reductive Transformations of

Carboxylic and Carbonic Acid Derivatives Using Molecular Hydrogen.

ACS Catal. 2012, 2, 1718−1741.

Notes

The authors declare no competing ﬁnancial interest.

ACKNOWLEDGMENTS

■

Research presented in this article was supported by the

Laboratory Directed Research and Development (LDRD)

program of Los Alamos National Laboratory (LANL) under

project numbers 20140672PRD2, 20160666ER reserve project,

and 20170048DR. RB was supported by a LANL Director’s

Postdoctoral Fellowship. Computations were performed by

using the LANL Darwin Computational Cluster. LANL is

operated by Triad National Security, LLC, for the National

Nuclear Security Administration of U.S. Department of Energy

(contract no. 89233218CNA000001).

(14) Pritchard, J.; Filonenko, G. A.; van Putten, R.; Hensen, E. J. M.;

Pidko, E. A. Heterogeneous and homogeneous catalysis for the

hydrogenation of carboxylic acid derivatives: history, advances and

future directions. Chem. Soc. Rev. 2015, 44, 3808−3833.

(15) (a) Zhang, S.; Yu, J.; Li, H.; Mao, D.; Lu, G. High-effective

approach from amino acid esters to chiral amino alcohols over Cu/

ZnO/Al₂O₃catalyst and its catalytic reaction mechanism. Sci. Rep.

2016, 6, 33196. (b) Studer, M.; Burkhardt, S.; Blaser, H.-U. Catalytic

Hydrogenation of Chiral α -Amino and α -Hydroxy Esters at Room

Temperature with Nishimura Catalyst without Racemization. Adv.

Synth. Catal. 2001, 343, 802−808.

REFERENCES

■

(1) Werpy, T. A.; Holladay, J. E.; White, J. F. Top Value Added

Chemicals from Biomass: I. Results of Screening for Potential Candidates

from Sugars and Synthesis Gas; Paciﬁc Northwest National Laboratory

(PNNL): Richland, WA, 2004.

(16) (a) Ogata, O.; Nakayama, Y.; Nara, H.; Fujiwhara, M.; Kayaki, Y.

Atmospheric Hydrogenation of Esters Catalyzed by PNP-Ruthenium

Complexes with an N-Heterocyclic Carbene Ligand. Org. Lett. 2016,

18, 3894−3897. (b) Murata, Y.; Torihara, M. Metal complex including

tridentate aminodicarbene ligand and hydrogenation reduction method

using same. WO 2014203963 A1, 2014; (c) Kuriyama, W.; Matsumoto,

T.; Ogata, O.; Ino, Y.; Aoki, K.; Tanaka, S.; Ishida, K.; Kobayashi, T.;

Sayo, N.; Saito, T. Catalytic hydrogenation of esters. development of an

efficient catalyst and processes for synthesising (R)-1,2-propanediol

and 2-(l-menthoxy)ethanol. Org. Process Res. Dev. 2012, 16, 166−171.

(d) Touge, T.; Hakamata, T.; Nara, H.; Kobayashi, T.; Sayo, N.; Saito,

T.; Kayaki, Y.; Ikariya, T. Oxo-Tethered Ruthenium(II) Complex as a

Bifunctional Catalyst for Asymmetric Transfer Hydrogenation and H₂

Hydrogenation. J. Am. Chem. Soc. 2011, 133, 14960−14963.

(e) Saudan, L. A.; Saudan, C. M.; Debieux, C.; Wyss, P. Dihydrogen

Reduction of Carboxylic Esters to Alcohols under the Catalysis of

Homogeneous Ruthenium Complexes: High Efficiency and Unprece-

dented Chemoselectivity. Angew. Chem., Int. Ed. 2007, 46, 7473−7476.

(f) Abdur-Rashid, K. Transfer hydrogenation processes and catalysts.

WO 2004096735 A2, 2004.

(17) (a) He, T.; Buttner, J. C.; Reynolds, E. F.; Pham, J.; Malek, J. C.;

Keith, J. M.; Chianese, A. R. Dehydroalkylative Activation of CNN- and

PNN-Pincer Ruthenium Catalysts for Ester Hydrogenation. J. Am.

Chem. Soc. 2019, 141, 17404−17413. (b) Wang, F.; Tan, X.; Lv, H.;

Zhang, X. New Ruthenium Complexes Based on Tetradentate

Bipyridine Ligands for Catalytic Hydrogenation of Esters. Chem.

Asian J. 2016, 11, 2103−2106. (c) Tan, X.; Wang, Y.; Liu, Y.; Wang, F.;

Shi, L.; Lee, K.-H.; Lin, Z.; Lv, H.; Zhang, X. Highly Efficient

Tetradentate Ruthenium Catalyst for Ester Reduction: Especially for

(2) Ragauskas, A. J.; Williams, C. K.; Davison, B. H.; Britovsek, G.;

Cairney, J.; Eckert, C. A.; Frederick, W. J.; Hallett, J. P.; Leak, D. J.;

Liotta, C. L.; Mielenz, J. R.; Murphy, R.; Templer, R.; Tschaplinski, T.

The Path Forward for Biofuels and Biomaterials. Science 2006, 311,

484−489.

(3) Bozell, J. J.; Petersen, G. R. Technology development for the

production of biobased products from biorefinery carbohydrates-the

US Department of Energy ’s “Top 10 ” revisited. Green Chem. 2010, 12,

539−554.

(4) (a) Becker, J.; Lange, A.; Fabarius, J.; Wittmann, C. Top value

platform chemicals: bio-based production of organic acids. Curr. Opin.

Biotechnol. 2015, 36, 168−175. (b) Choi, S.; Song, C. W.; Shin, J. H.;

Lee, S. Y. Biorefineries for the production of top building block

chemicals and their derivatives. Metab. Eng. 2015, 28, 223−239.

(5) Eudes, A.; Berthomieu, R.; Hao, Z.; Zhao, N.; Benites, V. T.;

Baidoo, E. E. K.; Loque, D. Production of muconic acid in plants.

Metab. Eng. 2018, 46, 13−19.

(6) (a) Takahashi, E.; Furui, M.; Shibatani, T., Scale-up of D-lysine

production from L-lysine by successive chemical racemization and

microbial asymmetric degradation. 1997, 19, 245-250; DOI: 10.1023/

a:1018357607976 (b) Anastassiadis, S. L-Lysine Fermentation. Recent

Pat. Biotechnol. 2007, 1, 11−24.

́

(7) (a) Martin-Dominguez, V.; Estevez, J.; Ojembarrena, F.; Santos,

V.; Ladero, M. Fumaric Acid Production: A Biorefinery Perspective.

Fermentation 2018, 4, 33. (b) Papadaki, A.; Androutsopoulos, N.;

Patsalou, M.; Koutinas, M.; Kopsahelis, N.; Castro, A.; Papanikolaou,

S.; Koutinas, A. Biotechnological Production of Fumaric Acid: The

439

https://dx.doi.org/10.1021/acs.oprd.9b00559

Org. Process Res. Dev. 2020, 24, 415−442

Organic Process Research & Development

pubs.acs.org/OPRD

Article

Hydrogenation of Fatty Acid Esters. Org. Lett. 2015, 17, 454−457.

species: Sinharines, penimides, and illukumbins. Tetrahedron 1994, 50,

6279−6286.

(d) Filonenko, G. A.; Aguila, M. J. B.; Schulpen, E. N.; van Putten, R.;

(27) (a) Ward, S. M.; Skinner, M.; Saha, B.; Emrick, T. Polymer-

Temozolomide Conjugates as Therapeutics for Treating Glioblastoma.

Mol. Pharm. 2018, 15, 5263−5276. (b) Yu, X.; Zhang, B.; Shan, G.; Wu,

Y.; Yang, F.-L.; Lei, X. Synthesis of the molecular hybrid inspired by

Largazole and Psammaplin A. Tetrahedron 2018, 74, 549−555.

(c) Yang, Q.; He, C.; Zhang, Z.; Tan, L.; Liu, B.; Zhu, Z.; Shao, Z.;

Gong, B.; Shen, Y.-M. Redox-responsive flower-like micelles of poly(l-

lactic acid)-b-poly(ethylene glycol)-b-poly(l-lactic acid) for intra-

cellular drug delivery. Polymer 2016, 90, 351−362.

Wiecko, J.; Muller, C.; Lefort, L.; Hensen, E. J. M.; Pidko, E. A. Bis-N -

̈

heterocyclic Carbene Aminopincer Ligands Enable High Activity in Ru-

Catalyzed Ester Hydrogenation. J. Am. Chem. Soc. 2015, 137, 7620−

7623. (e) Spasyuk, D.; Vicent, C.; Gusev, D. G. Chemoselective

Hydrogenation of Carbonyl Compounds and Acceptorless Dehydro-

genative Coupling of Alcohols. J. Am. Chem. Soc. 2015, 137, 3743−

3746. (f) Tan, X.; Wang, Q.; Liu, Y.; Wang, F.; Lv, H.; Zhang, X. A new

designed hydrazine group-containing ruthenium complex used for

catalytic hydrogenation of esters. Chem. Commun. 2015, 51, 12193−

12196. (g) Li, W.; Xie, J.-H.; Yuan, M.-L.; Zhou, Q.-L. Ruthenium

complexes of tetradentate bipyridine ligands: highly efficient catalysts

for the hydrogenation of carboxylic esters and lactones. Green Chem.

2014, 16, 4081−4085. (h) Spasyuk, D.; Smith, S.; Gusev, D. G.

Replacing Phosphorus with Sulfur for the Efficient Hydrogenation of

Esters. Angew. Chem., Int. Ed. 2013, 52, 2538−2542.

̈

(28) Blomenkemper, M.; Schroder, H.; Pape, T.; Hahn, F. E.

Copper(I) complexes of N-centered aliphatic tripodal trithioether

ligands − Adjustment of complex geometry by variation of spacer

lengths. Inorg. Chim. Acta 2011, 366, 76−80.

(29) (a) Keiko, N. A.; Stepanova, L. G.; Verochkina, E. A.; Chuvashev,

Y. A. Synthesis and properties of alkylthioethanals. ARKIVOC 2011,

2011, 127−138. (b) Mikołajczyk, M.; Bałczewski, P. Methylenomycin

B: A New Synthesis from a β -Ketophosphonate. Synthesis 1987, 659−

661.

̈

(18) (a) Schorgenhumer, J.; Zimmermann, A.; Waser, M. SNS-

Ligands for Ru-Catalyzed Homogeneous Hydrogenation and Dehy-

drogenation Reactions. Org. Process Res. Dev. 2018, 22, 862−870.

(b) Otsuka, T.; Ishii, A.; Dub, P. A.; Ikariya, T. Practical Selective

Hydrogenation of α -Fluorinated Esters with Bifunctional Pincer-Type

Ruthenium(II) Catalysts Leading to Fluorinated Alcohols or Fluoral

Hemiacetals. J. Am. Chem. Soc. 2013, 135, 9600−9603.

(30) Similar reaction with ligand 1d produced a deep blue color,

possibly due to generation of trityl radicals and triggered by this

transformations. No attempts to isolate and identify the product were

undertaken.

(31) For the complex Ru-1c, a second minor (3−5%) set of isomers is

observed in the NMR spectra, which can be attributed to fac-isomerism.

That minor set of isomers can also be present in solution of the related

complex Ru-2c.

(19) (a) Stadler, B. M.; Puylaert, P.; Diekamp, J.; van Heck, R.; Fan,

Y.; Spannenberg, A.; Hinze, S.; de Vries, J. G. Inexpensive Ruthenium

NNS-Complexes as Efficient Ester Hydrogenation Catalysts with High

C=O vs. C=C Selectivities. Adv. Synth. Catal. 2018, 360, 1151−1158.

(b) Henrion, M.; Roisnel, T.; Couturier, J.-L.; Dubois, J.-L.; Sortais, J.-

B.; Darcel, C.; Carpentier, J.-F. Ruthenium complexes bearing amino-

bis(phosphinite) or amino-bis(aminophosphine) ligands: Application

in catalytic ester hydrogenation. Mol. Catal. 2017, 432, 15−22.

(20) (a) Ishii, A.; Ootsuka, T.; Imamura, M.; Nishimiya, T.; Kimura,

K. Process for producing α-ﬂuoroaldehydes. WO 2013018573 A1,

2013; (b) Ishii, A.; Ootsuka, T.; Ishimaru, T.; Imamura, M. Method For

Producing β-Fluoroalcohol. WO 2012105431 A1, 2012; (c) Ootsuka,

T.; Imamura, M.; Ishii, A.; Ueda, K.; Mimura, S. α,α-Diﬂuoroacetalde-

hyde Production Method. WO 2014115801 A1, 2014; (d) Dub, P. A.;

Scott, B. L.; Gordon, J. C. Air-Stable NNS (ENENES) Ligands and

Their Well-Defined Ruthenium and Iridium Complexes for Molecular

Catalysis. Organometallics 2015, 34, 4464−4479.

(21) Ogata, O.; Nara, H.; Fujiwhara, M.; Matsumura, K.; Kayaki, Y. N-

Monomethylation of Aromatic Amines with Methanol via PNHP-

Pincer Ru Catalysts. Org. Lett. 2018, 20, 3866−3870.

(22) Dub, P. A.; Gordon, J. C. Polydentate ligands and their

complexes for molecular catalysis. WO 2015191505 A1, 2015.

(23) Geisser, R. W.; Oetiker, J. D.; Schroeder, F. Improvements in or

relating to organic compounds. WO 2015110515 A1, 2015.

̈

(32) (a) Landge, S. M.; Borkin, D. A.; Torok, B. Microwave-assisted

preparation of trifluoroacetaldehyde (fluoral): isolation and applica-

tions. Tetrahedron Lett. 2007, 48, 6372−6376. (b) Peerannawar, S.;

̈

Sood, A.; Brown, A.; Schafer, C.; Alonzo, J.; Sutton, S.; Christianson,

̈

M.; Stocking, R.; Naclerio, N.; Torok, B.; Landge, S. M. Effect of solvent

polarity on the regioselective hydroxyalkylation of indole with

trifluoroacetaldehyde hemiacetals. Struct. Chem. 2019, 30, 1941−

1956. (c) Ramanjaneyulu, B. T.; Mahesh, S.; Vijaya Anand, R. N-

Heterocyclic Carbene Catalyzed Highly Chemoselective Intermolec-

ular Crossed Acyloin Condensation of Aromatic Aldehydes with

Trifluoroacetaldehyde Ethyl Hemiacetal. Org. Lett. 2015, 17, 6−9.

̈

(d) Borkin, D. A.; Landge, S. M.; Torok, B. Enantioselective Friedel −

Crafts reaction of indoles with trifluoroacetaldehyde catalyzed by

Cinchona alkaloids. Chirality 2011, 23, 612−616.

(33) (a) Sordakis, K.; Tang, C.; Vogt, L. K.; Junge, H.; Dyson, P. J.;

Beller, M.; Laurenczy, G. Homogeneous Catalysis for Sustainable

Hydrogen Storage in Formic Acid and Alcohols. Chem. Rev. 2018, 118,

372−433. (b) Werkmeister, S.; Junge, K.; Beller, M. Catalytic

Hydrogenation of Carboxylic Acid Esters, Amides, and Nitriles with

Homogeneous Catalysts. Org. Process Res. Dev. 2014, 18, 289−302.

(c) Zhao, B.; Han, Z.; Ding, K. The N −H Functional Group in

Organometallic Catalysis. Angew. Chem., Int. Ed. 2013, 52, 4744−4788.

(24) Nakayama, Y.; Ogata, O. N-(Phosphinoalkyl)-N-(thioalkyl)-

amine derivative, method for producing same, and metal complex

thereof. WO 2016031874 A1, 2016.

(34) Sandoval, C. A.; Ohkuma, T.; Muniz, K.; Noyori, R. Mechanism

̃

of Asymmetric Hydrogenation of Ketones Catalyzed by BINAP/1,2-

Diamine-Ruthenium(II) Complexes. J. Am. Chem. Soc. 2003, 125,

13490−13503.

(25) (a) Alves de Oliveira, R.; Komesu, A.; Vaz Rossell, C. E.; Maciel

Filho, R. Challenges and opportunities in lactic acid bioprocess design-

From economic to production aspects. Biochem. Eng. J. 2018, 133, 219−

(35) (a) van Putten, R.; Uslamin, E. A.; Garbe, M.; Liu, C.; Gonzalez-

de-Castro, A.; Lutz, M.; Junge, K.; Hensen, E. J. M.; Beller, M.; Lefort,

L.; Pidko, E. A. Non-Pincer-Type Manganese Complexes as Efficient

Catalysts for the Hydrogenation of Esters. Angew. Chem., Int. Ed. 2017,

56, 7531−7534. (b) Ito, M.; Ootsuka, T.; Watari, R.; Shiibashi, A.;

Himizu, A.; Ikariya, T. Catalytic Hydrogenation of Carboxamides and

Esters by Well-Defined Cp *Ru Complexes Bearing a Protic Amine

Ligand. J. Am. Chem. Soc. 2011, 133, 4240−4242.

239. (b) Klotz, S.; Kaufmann, N.; Kuenz, A.; Pruße, U. Biotechnological

̈

production of enantiomerically pure d-lactic acid. Appl. Microbiol.

Biotechnol. 2016, 100, 9423−9437.

́

(26) (a) Lenstra, D. C.; Lenting, P. E.; Mecinovic, J. Sustainable

organophosphorus-catalysed Staudinger reduction. Green Chem. 2018,

́

20, 4418−4422. (b) Mirzahosseini, A.; Noszal, B. The species- and site-

specific acid-base properties of biological thiols and their homo-

disulfides. J. Pharm. Biomed. Sci. 2014, 95, 184−192. (c) Kim, S.;

Ginsbach, J. W.; Billah, A. I.; Siegler, M. A.; Moore, C. D.; Solomon, E.

I.; Karlin, K. D. Tuning of the Copper-Thioether Bond in Tetradentate

(36) Organic Chemicals from Biomass; Goldstein, I. S., Ed.; CRC Press,

Inc., 2017; p 310.

(37) The presence of trace amounts of methyl undec-9-enoate and

undecanoate in the E2 substrate somewhat complicates the

interpretation of the composition of the catalytic products by ¹H

NMR; nevertheless, the observed selectivity to undec-10-enol appears

to be at least 99%, or practically quantitative (runs 5−6).

N₃S(thioether) Ligands; O-O Bond Reductive Cleavage via a [Cu^II(μ -

2

1,2-peroxo)]²⁺/[Cu^III₂(μ -oxo)₂]²⁺Equilibrium. J. Am. Chem. Soc. 2014,

136, 8063−8071. (d) Hinterberger, S.; Hofer, O.; Greger, H. Synthesis

and corrected structures of sulphur-containing amides from Glycosmis

440

https://dx.doi.org/10.1021/acs.oprd.9b00559

Org. Process Res. Dev. 2020, 24, 415−442

Organic Process Research & Development

pubs.acs.org/OPRD

Article

(38) Zucca, P.; Littarru, M.; Rescigno, A.; Sanjust, E. Cofactor

Recycling for Selective Enzymatic Biotransformation of Cinnamalde-

hyde to Cinnamyl Alcohol. Biosci., Biotechnol., Biochem. 2009, 73,

1224−1226.

4−8 mesh beads, 208590) which had been previously activated by

evacuating several times immediately after heating at full power for

approximately 10 min in a commercial microwave. After 3 days over the

molecular sieves, the ester was used in further catalytic transformations.

(54) Selected NMR for 1-(diethylamino)-2,2,2-triﬂuoroethan-1-ol:

¹⁹F NMR: δ −75.9 ppm, d, ³J_F−H= 6 Hz; ¹³C{¹H} NMR: δ 90.6 ppm, q,

²J_C−F= 32 Hz, 1C, CF₃CH. For comparison, the same data for Hem1:

¹⁹F NMR: δ −83.3 ppm, d, ³J_F−H= 4 Hz; ¹³C{¹H} NMR: δ 96.6 ppm, q,

²J_C−F= 34 Hz, 1C, CF₃CH.

(55) (a) Hartmann, R.; Chen, P. Noyori ’s hydrogenation catalyst

needs a Lewis acid cocatalyst for high activity. Angew. Chem., Int. Ed.

2001, 40, 3581−3585. (b) John, J. M.; Takebayashi, S.; Dabral, N.;

Miskolzie, M.; Bergens, S. H. Base-Catalyzed Bifunctional Addition to

Amides and Imides at Low Temperature. A New Pathway for Carbonyl

Hydrogenation. J. Am. Chem. Soc. 2013, 135, 8578−8584. (c) Dub, P.

A.; Henson, N. J.; Martin, R. L.; Gordon, J. C. Unravelling the

Mechanism of the Asymmetric Hydrogenation of Acetophenone by

[RuX₂(diphosphine)(1,2-diamine)] Catalysts. J. Am. Chem. Soc. 2014,

136, 3505−3521.

(56) Liu, C.; van Putten, R.; Kulyaev, P. O.; Filonenko, G. A.; Pidko, E.

A. Computational insights into the catalytic role of the base promoters

in ester hydrogenation with homogeneous non-pincer-based Mn-P,N

catalyst. J. Catal. 2018, 363, 136−143.

(57) Acetophenone was used as received for the 2 h run but was

additionally puriﬁed (by stirring over calcium sulfate overnight,

ﬁltering, then distilling from calcium hydride under vacuum) for the

15 min run.

(58) Dub, P. A.; Scott, B. L.; Gordon, J. C. Why Does Alkylation of the

N-H Functionality within M/NH Bifunctional Noyori-Type Catalysts

Lead to Turnover? J. Am. Chem. Soc. 2017, 139, 1245−1260.

(59) Seebach, D. Structure and Reactivity of Lithium Enolates. From

Pinacolone to SelectiveC-Alkylations of Peptides. Difficulties and

Opportunities Afforded by Complex Structures. Angew. Chem., Int. Ed.

Engl. 1988, 27, 1624−1654.

(60) No aggregation was considered for ester enolates in our

modelling, see representative examples: (a) Seebach, D.; Amstutz, R.;

Laube, T.; Schweizer, W. B.; Dunitz, J. D. Structures of three lithium

ester enolates by x-ray diffraction: derivation of reaction path for

cleavage into ketene and alcoholate. J. Am. Chem. Soc. 1985, 107, 5403−

5409. (b) Williard, P. G.; Carpenter, G. B. X-ray crystal structures of

lithium, sodium, and potassium enolates of pinacolone. J. Am. Chem.

Soc. 1986, 108, 462−468. (c) Williard, P. G.; Carpenter, G. B. X-ray

crystal structure of an unsolvated lithium enolate anion. J. Am. Chem.

Soc. 1985, 107, 3345−3346.

(61) Kuze, N.; Ishikawa, A.; Kono, M.; Kobayashi, T.; Fuchisawa, N.;

Tsuji, T.; Takeuchi, H. Molecular Structure and Internal Rotation of

CF₃Group of Methyl Trifluoroacetate: Gas Electron Diffraction,

Microwave Spectroscopy, and Quantum Chemical Calculation Studies.

J. Phys. Chem. A 2015, 119, 1774−1786.

(62) Nocera, G.; Young, A.; Palumbo, F.; Emery, K. J.; Coulthard, G.;

McGuire, T.; Tuttle, T.; Murphy, J. A. Electron Transfer Reactions:

KO^tBu (but not NaO^tBu) Photoreduces Benzophenone under

Activation by Visible Light. J. Am. Chem. Soc. 2018, 140, 9751−9757.

(63) We observed a minor amount of these products by GC−MS in

performing methyl hexanoate hydrogenation under the conditions of

Figure 5 and evident by a small drop in H₂pressure.

(39) Kogan, S.; Kaliya, M.; Froumin, N. Liquid phase isomerization of

isoprenol into prenol in hydrogen environment. Appl. Catal., A 2006,

297, 231−236.

(40) see ref 16b.

(41) Xia, T.; Wei, Z.; Spiegelberg, B.; Jiao, H.; Hinze, S.; de Vries, J. G.

Isomerization of Allylic Alcohols to Ketones Catalyzed by Well-Defined

Iron PNP Pincer Catalysts. Chem.Eur. J. 2018, 24, 4043−4049.

(42) (a) Hu, J.; Bian, M.; Ding, H. Recent application of oxa-Michael

reaction in complex natural product synthesis. Tetrahedron Lett. 2016,

̈

57, 5519−5539. (b) Nising, C. F.; Brase, S. Recent developments in the

field of oxa-Michael reactions. Chem. Soc. Rev. 2012, 41, 988−999.

̈

(c) Nising, C. F.; Brase, S. The oxa-Michael reaction: from recent

developments to applications in natural product synthesis. Chem. Soc.

Rev. 2008, 37, 1218−1228.

(43) Inazu, A.; Nishizawa, E.; Akiba, S.; Nadano, R. Method for

Producing of α, α-diﬂuoroacetaldehyde. WO 2017126233 A1, 2017.

(44) Dub, P. A.; Batrice, R. J.; Gordon, J. C. Synthesis of Fluoro

Hemiacetals and Hemiaminals via Transition Metal-Catalyzed Ester

Hydrogenation. U.S. Provisional Patent Application 62/827,627, April

1, 2019.

(45) (a) Kondoh, A.; Ota, Y.; Komuro, T.; Egawa, F.; Kanomata, K.;

Terada, M. Chiral Brønsted acid-catalyzed enantioselective Friedel-

Crafts reaction of 2-methoxyfuran with aliphatic ketimines generated in

situ. Chem. Sci. 2016, 7, 1057−1062. (b) Mispelaere, C.; Roques, N.

Hemiaminals of trifluoroacetaldehyde, as trifluoromethylating agents.

Tetrahedron Lett. 1999, 40, 6411−6414.

(46) No reaction occurs in the absence of catalyst under the

conditions of Scheme 5, which further supports our statement that

hemiacetal is obtained from hemiaminal and/or its anion.

(47) (a) Gusev, D. G. Dehydrogenative Coupling of Ethanol and Ester

Hydrogenation Catalyzed by Pincer-Type YNP Complexes. ACS Catal.

2016, 6, 6967−6981. (b) Vicent, C.; Gusev, D. G. ESI-MS Insights into

Acceptorless Dehydrogenative Coupling of Alcohols. ACS Catal. 2016,

6, 3301−3309.

(48) One can also imagine glass- or surface-catalyzed hydration of

ester by H₂O, i.e. presence of water in the system could result in the

production of acid.

́

(49) Nguyen, D. H.; Trivelli, X.; Capet, F.; Swesi, Y.; Favre-Reguillon,

A.; Vanoye, L.; Dumeignil, F.; Gauvin, R. M. Deeper Mechanistic

Insight into Ru Pincer-Mediated Acceptorless Dehydrogenative

Coupling of Alcohols: Exchanges, Intermediates, and Deactivation

Species. ACS Catal. 2018, 8, 4719−4734.

(50) A similar idea was earlier proposed by Bergens to explain the

accelerative role of inorganic bases in related ketone hydrogenation

with Noyori catalyst, based on stoichiometric observation that alkoxo

complexes were not reacting with H₂in the absence of base, see

Hamilton, R. J.; Bergens, S. H. An Unexpected Possible Role of Base in

Asymmetric Catalytic Hydrogenations of Ketones. Synthesis and

Characterization of Several Key Catalytic Intermediates. J. Am. Chem.

Soc. 2006, 128, 13700−13701.

(51) Farrar-Tobar, R. A.; Wozniak, B.; Savini, A.; Hinze, S.; Tin, S.; de

Vries, J. G. Base-Free Iron Catalyzed Transfer Hydrogenation of Esters

Using EtOH as Hydrogen Source. Angew. Chem., Int. Ed. 2019, 58,

1129.

(52) Wang, Z.; Chen, X.; Liu, B.; Liu, Q.-b.; Solan, G. A.; Yang, X.;

Sun, W.-H. Cooperative interplay between a flexible PNN-Ru(ii)

complex and a NaBH₄additive in the efficient catalytic hydrogenation

of esters. Catal. Sci. Technol. 2017, 7, 1297−1304.

(53) Basic aluminum oxide (Brockman I, Sigma-Aldrich, 199443, 58 Å

pore size, ∼150 mesh) was activated by heating to 180°C under high

vacuum overnight, then taken into the glovebox. Methyl hexanoate

(Sigma-Aldrich, 99%, 259942) which was degassed by three freeze−

pump−thaw cycles was stirred over the basic alumina powder for 3 h,

then ﬁltered through a 0.2 μm Whatman syringe ﬁlter. The methyl

hexanoate was then stored over 4 Å molecular sieves (Sigma-Aldrich,

(64) Sullivan, D. F.; Woodbury, R. P.; Rathke, M. W. The self-

condensation reaction of lithium ester enolates. Isolation of a ketene

intermediate. J. Org. Chem. 1977, 42, 2038−2039.

(65) (a) Gu, G.; Yang, T.; Yu, O.; Qian, H.; Wang, J.; Wen, J.; Dang,

L.; Zhang, X. Enantioselective Iridium-Catalyzed Hydrogenation of α -

Keto Amides to α -Hydroxy Amides. Org. Lett. 2017, 19, 5920−5923.

(b) Liang, Z.; Yang, T.; Gu, G.; Dang, L.; Zhang, X. Scope and

Mechanism on Iridium-f-Amphamide Catalyzed Asymmetric Hydro-

genation of Ketones. Chin. J. Chem. 2018, 36, 851−856.

(66) (a) Chen, X.; Jing, Y.; Yang, X. Unexpected Direct Hydride

Transfer Mechanism for the Hydrogenation of Ethyl Acetate to Ethanol

Catalyzed by SNS Pincer Ruthenium Complexes. Chem.Eur. J. 2016,

22, 1950−1957. (b) Qu, S.; Dai, H.; Dang, Y.; Song, C.; Wang, Z.-X.;

441

https://dx.doi.org/10.1021/acs.oprd.9b00559

Org. Process Res. Dev. 2020, 24, 415−442

Organic Process Research & Development

pubs.acs.org/OPRD

Article

Guan, H. Computational Mechanistic Study of Fe-Catalyzed Hydro-

genation of Esters to Alcohols: Improving Catalysis by Accelerating

Precatalyst Activation with a Lewis Base. ACS Catal. 2014, 4, 4377−

4388. (c) Chen, T.; Li, H.; Qu, S.; Zheng, B.; He, L.; Lai, Z.; Wang, Z.-

X.; Huang, K.-W. Hydrogenation of Esters Catalyzed by Ruthenium

PN₃-Pincer Complexes Containing an Aminophosphine Arm. Organo-

metallics 2014, 33, 4152−4155.

(80) Marenich, A. V.; Cramer, C. J.; Truhlar, D. G. Universal Solvation

Model Based on Solute Electron Density and on a Continuum Model of

the Solvent Defined by the Bulk Dielectric Constant and Atomic

Surface Tensions. J. Phys. Chem. B 2009, 113, 6378−6396.

(81) (a) Weigend, F.; Ahlrichs, R. Balanced basis sets of split valence,

triple zeta valence and quadruple zeta valence quality for H to Rn:

Design and assessment of accuracy. Phys. Chem. Chem. Phys. 2005, 7,

3297−3305. (b) Weigend, F. Accurate Coulomb-fitting basis sets for H

to Rn. Phys. Chem. Chem. Phys. 2006, 8, 1057−1065.

(67) Dub, P. A.; Gordon, J. C. The mechanism of enantioselective

ketone reduction with Noyori and Noyori-Ikariya bifunctional catalysts.

Dalton Trans. 2016, 45, 6756−6781.

(82) Fukui, K. The path of chemical reactions - the IRC approach. Acc.

Chem. Res. 1981, 14, 363−368.

(68) Dub, P. A.; Gordon, J. C. Metal −Ligand Bifunctional Catalysis:

The “Accepted ” Mechanism, the Issue of Concertedness, and the

Function of the Ligand in Catalytic Cycles Involving Hydrogen Atoms.

ACS Catal. 2017, 7, 6635−6655.

(69) We have also performed various constrained PES scans

modelling transfer of each separate K “atom” to neighboring oxygen

atom in the hopes that this would trigger a generation of coordination

vacancy at the metal via any Ru-O bond cleavage and/or open existence

of alternative pathways to int7·base (e.g. via coordinated to metal

CH₃OK obtained via a transfer of “K” atom from explicit KO-t-C₄H₉).

However, no minima were located. We therefore conclude that the

methoxide ligand in int5·base generated the coordination vacancy via

direct Ru−O heterolytic bond cleavage, ts5·base.

(83) Pettersen, E. F.; Goddard, T. D.; Huang, C. C.; Couch, G. S.;

Greenblatt, D. M.; Meng, E. C.; Ferrin, T. E. UCSF Chimera A

visualization system for exploratory research and analysis. J. Comput.

Chem. 2004, 25, 1605−1612.

(84) http://www.chemcraftprog.com/index.html (accessed 2020-02-

14).

(70) For the potassium amidate complex of Ru-MACHO in the

presence of three KO-t-C₄H₉“molecules” the barrier is calculated by

summation of thermodynamics of NH/NK exchange and activation

barrier with respect to optimized geometry of the potassium amidate

complex of Ru-MACHO in the presence of three KO-t-C₄H₉

“molecules” and E1.

(71) Note, the relationship between these barriers and overall catalytic

reaction rates is not necessarily linear.

(72) Laury, M. L.; Wilson, A. K. Performance of Density Functional

Theory for Second Row (4d) Transition Metal Thermochemistry. J.

Chem. Theory Comput. 2013, 9, 3939−3946.

(73) Barbaro, P.; Bianchini, C.; Giambastiani, G.; Togni, A. The first

tridentate phosphine ligand combining planar, phosphorus and carbon

chirality. Chem. Commun. 2002, 2672−2673.

(74) We also found that dissolving this residue in minimum amount of

dichloromethane results in following crystallization of the product.

(75) Smith, C. C.; Hollenstein, M.; Leumann, C. J. The synthesis and

application of a diazirine-modified uridine analogue for investigating

RNA-protein interactions. RSC Adv. 2014, 4, 48228−48235.

(76) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.;

Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.;

Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H. P.;

Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.;

Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima,

T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Montgomery, J. A., Jr.;

Peralta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin,

K. N.; Staroverov, V. N.; Keith, T.; Kobayashi, R.; Normand, J.;

Raghavachari, K.; Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.;

Cossi, M.; Rega, N.; Millam, J. M.; Klene, M.; Knox, J. E.; Cross, J. B.;

Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.;

Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.;

Martin, R. L.; Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador,

P.; Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, O.;

Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. Gaussian 09,

Revision E.01; Gaussian, Inc.: Wallingford CT, 2013.

(77) (a) Hohenberg, P.; Kohn, W. Inhomogeneous Electron Gas.

Phys. Rev. 1964, 136, B864−B871. (b) Kohn, W.; Sham, L. J. Self-

Consistent Equations Including Exchange and Correlation Effects.

Phys. Rev. 1965, 140, A1133−A1138.

(78) Chai, J.-D.; Head-Gordon, M. Long-range corrected hybrid

density functionals with damped atom-atom dispersion corrections.

Phys. Chem. Chem. Phys. 2008, 10, 6615−6620.

(79) Grimme, S. Semiempirical GGA-type density functional

constructed with a long-range dispersion correction. J. Comput. Chem.

2006, 27, 1787−1799.

442

https://dx.doi.org/10.1021/acs.oprd.9b00559

Org. Process Res. Dev. 2020, 24, 415−442

Article Doi

Engineering Catalysts for Selective Ester Hydrogenation

DOI: 10.1021/acs.oprd.9b00559

Source and publish data:

Authors:

Article abstract of DOI:10.1021/acs.oprd.9b00559

Full text of DOI:10.1021/acs.oprd.9b00559

Products guided by the article

R&D Labs maybe for 377-52-6

Relevant to this article

Competitive insertion of isocyanide and carbon dioxide into niobium- and silicon-amido bonds

New stable form of nitroacetonitrile