Hydrophosphonylation of aldehydes catalyzed by cyclopentadienyl ruthenium(II) complexes

Article abstract of DOI:10.1016/j.mcat.2018.03.001

This work reports the first method for the synthesis of α-hydroxyphosphonates from aldehydes catalyzed by cyclopentadienyl ruthenium(II) complexes. The best results were obtained using the system HP(O)(OEt)₂/[RuClCp(PPh₃)₂

Full text of DOI:10.1016/j.mcat.2018.03.001

MolecularCatalysis450(2018)77–86
Contents lists available at ScienceDirect
Molecular Catalysis
journal homepage: www.elsevier.com/locate/mcat
Hydrophosphonylation of aldehydes catalyzed by cyclopentadienyl
ruthenium(II) complexes
Ivânia R. Cabritaa, Pedro R. Florindoa, Paulo J. Costab, M. Conceição Oliveiraa,
Ana C. Fernandesa,⁎
a
Centro de Química Estrutural, Instituto Superior Técnico, Universidade de Lisboa, Av. Rovisco Pais, 1049-001 Lisboa, Portugal
Centro de Química e Bioquímica, DQB and BioISI – Biosystems & Integrative Sciences Institute, Faculdade de Ciências da Universidade de Lisboa, Campo Grande, 1749-
b
016 Lisboa, Portugal
A R T I C L E I N F O
A B S T R A C T
Keywords:
This work reports the first method for the synthesis of α-hydroxyphosphonates from aldehydes
catalyzed by cyclopentadienyl ruthenium(II) complexes. The best results were obtained using the system
HP(O)(OEt)2/[RuClCp(PPh3)2] (5 mol%), affording the α-hydroxyphosphonates in good to excellent yields with
high chemoselectivity. The catalyst [RuClCp(PPh3)2] can be used for at least 12 catalytic cycles with excellent
activity and the reactions were carried out under solvent free conditions. DFT calculations were performed to
rationalize the mechanism showing that the barriers associated with the H-phosphonate tautomer, HP(O)(OR)2
are unrealistically high. This led us to propose that the catalyst promotes the tautomerization towards phosphite,
P(OH)(OR)2, via P-atom coordination, which accounts for the observed reactivity.
Hydrophosphonylation
α-hydroxyphosphonates
cyclopentadienylruthenium(II) complexes
Aldehydes
DFT calculations
1. Introduction
using oxo-molybdenum complexes as efficient catalysts for the hydro-
phosphonylation of aldehydes [14], herein we report the first method
The synthesis of α-hydroxyphosphonates is one of the rapidly
growing research fields in organic chemistry [1] due to the intriguing
biological properties of these compounds and their derivatives as an-
tibiotics, antivirals, anticancer drugs, herbicides, insecticides, fungi-
cides, and enzyme inhibitors [2]. α-Hydroxyphosphonates are also
useful precursors for the synthesis of other biologically important α-
functionalized phosphonates such as amino-, keto-, and acetoxypho-
sphonates [3].
The addition of H-phosphonates to carbonyl compounds (Pudovik
or Abramov reaction) is a crucial and highly valuable method for the
synthesis of α-hydroxyphosphonates and for C-P bond formation. Many
efforts have been made to develop highly active catalysts for this re-
action in recent years. As consequence, several catalysts based on bis-
muth [4], barium [5], tungsten [6], vanadium [7], zinc [8], alkaline
earth [9], lanthanide [10], titanium [11], ytterbium [12], or aluminum
[13] have been studied. However, some of these systems require rela-
tively harsh reaction conditions, such as high temperature (above
100 °C) and high catalyst loading.
for the synthesis of α-hydroxyphosphonates catalyzed by ruthenium
complexes.
2. Experimental
Carbonyl compounds and phosphites were obtained from
commercial suppliers and were used without further purification.
The catalysts [RuClCp(PPh3)2] [15], [RuClCp(dppe)] [15],
[RuClCp(2,2′-bipy)(PPh3)][PF6] [16], [RuCp(TFru)(PPh3)2][PF6] [17],
and [RuCp(NCXylAc)(PPh3)2][PF6] [18], were prepared according to
literature procedures. Flash chromatography was performed on MN
Kieselgel 60 M 230–400 mesh. 1H NMR, 13C NMR and 31P NMR spectra
were measured on a Bruker Avance II+ 400 MHz and 300 MHz spec-
trometers. Chemical shifts are reported in parts per million (ppm)
downfield from an internal standard.
ESI mass experiments were carried out on a LCQ Fleet mass spec-
trometer operated in the ESI positive/negative ion mode (Thermo
Scientific). The optimized parameters were as following: ion spray
voltage, 4.5 kV; capillary voltage, 16 (-11) V; tube lens offset, −63 V;
sheath gas (N2), 80 arbitrary units; auxiliary gas, 5 arbitrary units;
capillary temperature, 250°C. The spectra were recorded in the range
100–1500 Da. Spectrum typically corresponds to the average of 20–35
Due to the fast development of the chemistry and biology of α-hy-
droxyphosphonates over the past decade, the introduction of novel
catalysts that show wide applicability for the hydrophosphonylation of
aldehydes is strongly required. In continuation of our previous work
⁎
Corresponding author.
E-mail address: anacristinafernandes@tecnico.ulisboa.pt (A.C. Fernandes).
https://doi.org/10.1016/j.mcat.2018.03.001
Received 6 November 2017; Received in revised form 30 January 2018; Accepted 1 March 2018
2468-8231/©2018ElsevierB.V.Allrightsreserved.

I.R. Cabrita et al.
MolecularCatalysis450(2018)77–86
scans. CID experiments were obtained with an isolation window of
8 Da, a 20% relative collision energy and with an activation energy of
30 msec.
The Density Functional Theory calculations [19] were performed
with Gaussian09 package [20] using the PBE1PBE functional, also
known as PBE0. This functional was obtained combining the general-
ized gradient functional of Perdew, Burke and Ernzerhof (PBE) [21]
with a predefined amount of exact exchange functional defined by
Adamo and Barone [22] (25% exchange and 75% correlation
weighting).
For Ru and P, the triple-ζ basis set with one polarization function
LANL2TZ(f) and LANL08(d), respectively, were used with the asso-
ciated effective core potential (ECP) [23]. Both were downloaded from
the EMSL Basis Set Library [24]. The remaining elements were de-
scribed with the standard 6–31G* basis set. All geometries were opti-
mized without symmetry constrains and frequency calculations were
performed for all species in order to characterize the nature of the
stationary points. The transition-state structures, which yielded one
imaginary frequency, were relaxed following the reaction path by in-
tegrating the intrinsic reaction coordinate (IRC calculation). All calcu-
lations were performed using a Polarizable Continuum Model (PCM) to
describe solvent effect via the integral equation formalism variant
(IEFPCM) [25], as implemented in the software. Throughout this
manuscript, toluene (ε = 2.3741) was chosen, given that it was used in
the ESI–MS study.
The reaction mixture was cooled and the yield was determined by
1H NMR spectroscopy using mesitylene (0.5 mmol) as the internal standard.
In the next catalytic cycles, hydrocinnamaldehyde (0.5 mmol), HP(O)(OEt)2
(2.4 mmol) and mesitylene (0.5 mmol) was added to the reaction mixture
and stirred at 80 °C. The yields were determined by 1H NMR spectroscopy.
3. Results and discussion
Initially, we studied the hydrophosphonylation of the test substrate 4-fluoro
benzaldehyde catalyzed by the cyclopentadienyl ruthenium (II) complexes
[RuClCp(PPh3)2] 1, [RuClCp(dppe)] 2, [RuCp(NCXylAc)(PPh3)2][PF6] 3,
[RuClCp(2,2′-bipy)(PPh3)][PF6] 4 and [RuCp(TFru)(PPh3)2][PF6] 5 (Fig. 1)
with different H-phosphonates in order to assess the best reaction conditions
(Tables 1–3). The progress of the reactions was monitored by
thin layer chromatography and by 1H NMR spectroscopy. Complex
[RuClCp(PPh3)2] 1 (5 mol%) was the best catalyst for this reaction, affording
the corresponding α-hydroxyphosphonate in 89% yield after 3 h under
solvent free conditions at 80 °C in air atmosphere (Table 1, entry 1). Using
only 2.5 mol% of [RuClCp(PPh3)2], the α-hydroxyphosphonate was
obtained in 85% yield, but the reaction required 24 h (Table 1, entry 2). At
room temperature, the reaction did not occur in presence of
5 mol% of [RuClCp(PPh3)2] (Table 1, entry 3). The ruthenium complex
[RuClCp(dppe)] 2, containing the bidentate ethylenebis(diphenylphosphine)
(dppe), required 24 h to produce the α-hydroxyphosphonate in 63% yield
(Table 1, entry 4).
In order to study the effect of the solvent on the reaction barriers,
tributylphosphate (ε = 8.1781) aiming at representing the H-phos-
phonate, and benzaldehyde (ε = 18.220) were also tested. The triple-ζ
basis set 6–311G** in replacement of 6–31G*was employed in some
calculations to test the basis set effect (vide infra).
The reaction catalyzed by complex [RuCp(NCXylAc)(PPh3)2][PF6] 3,
containing
a carbohydrate moiety prepared from xylose, afforded
the α-hydroxyphosphonate in 77% yield (Table 1, entry 5). In
contrast, low yields were obtained in the reaction catalyzed by
[RuClCp(2,2′-bipy)(PPh3)][PF6] 4 (Table 1, entry 6). Finally, no reaction
was observed using the catalyst [RuCp(TFru)(PPh3)2][PF6] 5 or in the
absence of catalyst (Table 1, entries 7 and 8).
After the screening of catalysts, different H-phosphonates were
evaluated in the hydrophosphonylation of 4-fluorobenzaldehyde
(Table 2). The reaction carried out with HP(O)(OEt)2 produced the best
yield (89%) of the corresponding α-hydroxyphosphonate after 3 h
under solvent free conditions (Table 2, entry 1). Good yields of
α-hydroxyphosphonates were also obtained using HP(O)(OBu)2 and HP
(O)(OMe)2, but these reactions required 24 h (Table 2, entries 2–3).
Finally, in the reaction performed with HP(O)(OPh)2, the α-hydro-
xyphosphonate was obtained in 58% yield, along with several sec-
ondary products (Table 2, entry 4).
To evaluate the scope and limitations of the system
HP(O)(OEt)2/[RuClCp(PPh3)2] (5 mol%), we decided to study the
hydrophosphonylation of several aldehydes at 80 °C under solvent free
conditions in air atmosphere. Generally, good to excellent yields of
α-hydroxyphoshonates were obtained under the optimized reaction
conditions, including aldehydes bearing electron-withdrawing or elec-
tron-donating groups (Table 3). Indeed, the hydrophosphonylation of
aldehydes containing several functional groups such as eOCH3, eSCH3,
eF, eCN, eCF3, eNO2, eBr, eCO2R was successfully accomplished
2.1. General procedure for the hydrophosphonylation of aldehydes with HP
(O)(OEt)2 catalyzed by [RuClCp(PPh3)2]
To a mixture of aldehyde (2 mmol) and [RuClCp(PPh3)2] (5 mol%)
was added HP(O)(OEt)2 (2.4 mmol). The reaction mixture was stirred at
80 °C under inert atmosphere (the reaction times are indicated in the
Table 3) and the reaction progress was monitored by TLC and 1H NMR
spectroscopy. When the reaction was complete, water (3 mL) was added
and the mixture was stirred at 80 °C. After 1 h, the reaction mixture was
cooled to ambient temperature and extracted with ethyl acetate
(2 × 10 mL). The combined organic layers were dried over Na2SO4 and
filtered, and the solvent was removed under reduced pressure. The
residue was purified by flash chromatography with appropriate mix-
tures of n-hexane and ethyl acetate.
2.2. Use of the catalyst [RuClCp(PPh3)2] in several catalytic cycles
To a mixture of hydrocinnamaldehyde (0.5 mmol) and [RuClCp(PPh3)2]
(5 mol%) was added HP(O)(OEt)2 (2.4 mmol). The reaction mixture
was stirred at 80 °C under an air atmosphere during 1 h 30 min.
Fig. 1. Structures of Ru(II)Cp complexes.
78

I.R. Cabrita et al.
MolecularCatalysis450(2018)77–86
Table 1
Hydrophosphonylation of 4-fluorobenzaldehyde catalyzed by different Ru(II)Cp complexes.a
Entry
Catalyst
Catalyst [mol%]
Temp. [°C]
Time [h]
Yield [%]b
1
2
3
4
5
6
7
8
[RuClCp(PPh3)2] 1
[RuClCp(PPh3)2] 1
[RuClCp(PPh3)2] 1
[RuClCp(dppe)] 2
[RuCp(NCXylAc)(PPh3)2][PF6] 3
[RuClCp(2,2′-bipy)(PPh3)][PF6] 4
[RuCp(TFru)(PPh3)2][PF6] 5
Without catalyst
5
2.5
5
5
5
5
5
–
80
80
r. t.
80
80
80
80
80
3
89
85
24
24
24
24
24
24
24
No reaction
63
77
10
No reaction
No reaction
a
The reactions were carried out with 1.0 mmol of 4-fluorobenzaldehyde and 2.4 mmol of HP(O)(OEt)2.
Isolated yields.
b
Table 2
Hydrophosphonylation of 4-fluorobenzaldehyde catalyzed by [RuClCp(PPh3)2] using different HP(O)(OR)2.a
Entry
1
HP(O)(OR)2
HP(O)(OEt)2
Product
Time[h]
3
Yield [%]b
89
2
3
4
HP(O)(OBu)2
HP(O)(OMe)2
HP(O)(OPh)2
24
24
4.5
84
82
58
a
The reactions were carried out with 2.0 mmol of 4-fluorobenzaldehyde and 2.4 mmol of HP(O)(OR)2.
Isolated yields.
b
with the system HP(O)(OEt)2/[RuClCp(PPh3)2], showing the high
and chloride-containing species that agrees well with the one calculated
for [RuClCp(PO)(OEt)2(PPh3)]−, as shown in Fig. 2a (insert). The (+)
ESI mass spectrum (not shown) only presented peaks that can be at-
tributed to the catalyst. For further characterization of the [M]− spe-
cies, its isotopic cluster was isolated, and collision induced dissociation
experiments (CID) were performed. The MS2 spectrum showed only a
fragmentation channel due to the loss of 262 u (a triphenylphosphine
unit), leading to a group of peaks at m/z 339/341 attributed to
[RuClCp(PO)(OEt)2]−, according to the calculated isotopic distribution
(Fig. 2b).
The reaction between the catalyst [RuClCp(PPh3)2] and
4-(methylthio)benzaldehyde at room temperature (Eq. (1)) was also
analyzed by ESI–MS. However, the mass spectra acquired in both ESI
ion modes present only peaks related to the catalyst. This result sug-
gests that a direct interaction between the catalyst and the aldehyde
does not occur.
chemoselectivity of this novel method. The reaction of 2-ni-
trobenzaldehyde gave the corresponding α-hydroxyphosphonate in
lower yield (62%) than the reaction of 4-nitrobenzaldehyde (85%),
which may be due to steric effects.
The substrate versatility of this method was further demonstrated by
the hydrophosphonylation of hydrocinnamaldehyde (Table 3, entry 8),
furfural (Table 3, entry 11) and ferrocenecarboxaldehyde (Table 3,
entry 12) in good yields. In contrast, the reaction of heteroaromatic
aldehyde 2-pyridinecarboxaldehyde gave the corresponding α-hydro-
xyphosphonate in low yield (Table 3, entry 13) and the aliphatic al-
dehyde iPrCHO did not react.
We have also evaluated the catalytic activity of [RuClCp(PPh3)2] in
more than one cycle using the hydrocinnamaldehyde as test substrate,
by sequential addition of fresh substrate and HP(O)(OEt)2 to the reac-
tion mixture. The reactions were followed by 1H NMR spectroscopy and
we observed that [RuClCp(PPh3)2] can be used for at least 12 catalytic
cycles with excellent activity (Table 4).
To study the activation of the phosphite by the catalyst, the reaction
between the complex [RuClCp(PPh3)2] with HP(O)(OEt)2 in toluene, at
room temperature, was performed and investigated by ESI–MS meth-
odologies.
The (−)ESI mass spectrum displayed a group of peaks (m/z 601,
base peak) with a characteristic isotopic distribution for a ruthenium
(1)
ESI–MS results clearly indicate that the catalytic cycle starts with
the formation of the unsaturated complex [RuClCp(PPh3)], by
de-coordination of a PPh3 ligand and subsequent coordination of
79

I.R. Cabrita et al.
MolecularCatalysis450(2018)77–86
Table 3
Hydrophosphonylation of aldehydes with the system HP(O)(OEt)2/[RuClCp(PPh3)2].a
Entry
1
Substrate
Product
Time[h]
3
Yield [%]b
94
2
3
92
3
4
6
3
91
89
5
6
3
6
88
86
7
3
85
8
3.5
1.5
4
62
82
79
9
10
11
3
74
12
13
3
70
30
18
14
18
70
a
The reactions were carried out with 2.0 mmol of aldehyde and 2.4 mmol of HP(O)(OEt)2.
Isolated yields.
b
HP(O)(OMe)2. This is further corroborated by DFT calculations (see
below). Assuming this as a common feature for catalysts 2-5, the lower
catalytic activity of these complexes in relation to complex 1 can be
rationalized based on these findings. In the case of catalyst 2, the de-
coordination of one phosphorus atom of 1,2-bis(diphenylphosphino)
ethane (dppe) is entropically disfavored (chelate effect). In catalysts
3-5, the cationic nature of the metal centers makes them less prone for
PPh3 de-coordination and thus less active than catalyst 1; the
differences in activity within these three catalysts may be rationalized
based on the lower electronic density of the metal center in 4 when
compared with 3, due to the different ligand sphere, while the in-
activity of complex 5 may be related with the acidic nature of the tet-
razole hydrogen, with the possible existence of an equilibrium between
the tetrazole and the tetrazolate neutral complex [26].
Density Functional Theory (DFT) calculations [19] were performed
to rationalize the possible mechanism for hydrophosphonylation of
80

I.R. Cabrita et al.
MolecularCatalysis450(2018)77–86
Table 4
Use of the catalyst [RuClCp(PPh3)2] in multiple cycles.a
Catalytic cycle
Time [h]
Yield [%]b
TON
1
2
3
4
5
6
7
8
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
2
92
92
91
92
91
92
90
91
91
90
89
88
20
40
60
80
100
120
140
160
180
200
220
240
9
10
11
12
2.5
a
The reactions were carried out with 0.5 mmol of hydrocinammaldehyde and 2.4 mmol of HP(O)(OEt)2.
Yields determined by1H NMR spectroscopy.
b
Fig. 2. Negative ion ESI–MS spectra of Ru complex with HP(O)(OEt)2 in toluene. a) Full scan mass spectrum. The insert shows the theoretical isotope pattern for [M]−; b) Tandem mass
spectrum of precursor ion [M]−, m/z 601/603.
Fig. 3. Comparison between the DFT optimized structure and the X-ray experimental structures of complex 1, [RuClCp(PPh3)2]. Hydrogens were omitted for clarity. Distances are
presented in Å.
Ru-Cl distance 2.474 Å (DFT) vs. 2.453/2.462 Å (experimental) as seen
in Fig. 3.
aldehydes by CpRu(II) complexes. The mechanism was modeled
using complex [RuClCp(PPh3)2] 1, HP(O)(OMe)2 and benzaldehyde.
Complex 1 is coordinately saturated, thus being reasonable to as-
sume that the catalytic cycle starts with the formation of an unsaturated
species. Experimentally, MS analysis of the reaction of 1 with HP(O)
Complex
1 crystallizes as two polymorphs in the solid state,
CPDRUA [27] and CPDRUA01 [28] (CSD refcodes [29]) and the
DFT optimized structure compares nicely with both, e.g.,
81

I.R. Cabrita et al.
MolecularCatalysis450(2018)77–86
Fig. 4. Mechanism for the reaction of A with benzaldehyde via intermediate B in toluene (ΔH top, bold and ΔG bottom, italics in kcal mol−1). The energy scale corresponds to the
calculated enthalpies.
Fig. 5. Relevant distances in the reagents, products, intermediates, and transition states involved in the reaction of A with benzaldehyde via intermediate B. In some structures, the CeH
hydrogen atoms were omitted for clarity.
(OEt)2 in toluene at room temperature revealed a peak at m/z = 601,
consistent with the [RuClCpP(O)(OEt)2(PPh3)]− species. This clearly
indicates that the formation of an active coordinately unsaturated Ru-
species occurs via de-coordination of a PPh3 ligand, possibly through
the formation of intermediate A, [RuClCp(PPh3)], followed by
HP(O)(OMe)2 coordination, ultimately yielding complex B (Fig. 4), in
agreement with the MS peak at m/z = 601. Complex B has a higher
interaction between the catalyst and aldehyde was experimentally ex-
cluded, therefore, initial coordination of HP(O)(OMe)2 seems likely.
Moreover, PPh3 de-coordination and replacement is a common process
observed in catalysis promoted by this type of Ru-complexes [30–32].
Fig. 5 presents the optimized geometries and relevant calculated
distances of reagents, products, intermediates, and transition states
involved in the reaction of A with benzaldehyde via intermediate B. In
this intermediate, HP(O)(OMe)2 is coordinated by the P = O oxygen
energy (ΔH in solution = 13.8 kcal mol−1
) than 1, however, an
82

I.R. Cabrita et al.
MolecularCatalysis450(2018)77–86
Fig. 6. Mechanism for the reaction of A with benzaldehyde via intermediate E1 (ΔH top, bold and ΔG bottom, italics in kcal mol−1) in toluene. The energy scale corresponds to the
calculated enthalpies.
atom (Ru-O distance = 2.276 Å) leading to an increase of that distance
from 1.499 Å to 1.518 Å when compared with the free HP(O)(OMe)2
molecule while the P-H distance remains practically constant (1.41 Å).
The reaction of complex B with benzaldehyde yields a hydrogen
bonded complex C (PeH⋯O distance = 2.345 Å), dropping the energy
to 9.9 kcal mol−1 if ΔH values are considered (Fig. 4). The free energy
ΔG of C rises to 14.3 kcal mol−1 given the entropy penalty of joining the
isolated molecules (benzaldehyde and B) to form C. The activation of
the aldehyde proceeds through transition state TSCD which lies
98.9 kcal mol−1(ΔH) or 110.2 (ΔG) kcal mol−1 above the starting re-
agents (1 + HP(O)(OMe)2+ benzaldehyde). In this transition-state, the
benzaldehyde rearranges and the C]O bond, which is stretched from
1.214 Å to 1.371 Å, is aligned parallel to the PeH bond. New P⋯C and
O⋯H bonds start to form (1.877 and 1.835 Å, respectively) although
the O⋯H distance is still quite long. Attempts to locate non-concerted
transition-states, i.e., a stepwise formation of the P⋯C and O⋯H bonds,
were not successful. The concerted TSCD leads to complex D with a
coordination of P(OH)(OMe)2 to A affords complex E1 which is stable,
i.e., ∼24 kcal mol−1 (ΔH) lower in energy than the starting products
(Fig. 6). More importantly, E1 is also compatible with the MS peak at
m/z = 601.
Fig. 7 presents the optimized structures of E1 together with all the
relevant species discussed in the reactivity of this complex (see below).
In E1 the coordinated phosphite presents a Ru-P bond of 2.246 Å,
shorter than the analogous Ru-PPh3 (2.334 Å). The P-OH moiety is
available for hydrogen bonding and indeed, in the presence of benzal-
dehyde, species F is formed, the aldehyde being hydrogen bonded to the
coordinated phosphite with a OH⋯O distance of 1.728 Å. The reaction
proceeds through TSFG with ΔH = 35.9 kcal mol−1 relative to the
starting reagents (1 + phosphite + benzaldehyde). In this transition
state, the original OeH bond of the phosphite is stretched from 0.989 Å
to 1.002 Å whereas a new OeH bond starts to form, although the dis-
tance is quite long (1.664 Å). Simultaneously, the C]O bond starts to
lose the double bond character, increasing the distance from 1.225 Å in
F to 1.350 Å in TSFG as a new P-C bond starts to form (2.230 Å). As the
reaction proceeds, the α-hydroxyphosphonate is formed and released.
In G, the product lies close to the complex A, whereas re-coordination
of the phosphine PPh3 regenerates the catalyst 1, the process being
favorable in terms of the calculated enthalpies or free energies when
compared with the starting reagents.
The comparison of the paths presented in Figs. 4 and 6, and more
specifically TSCD and TSFG, indicates that formation of the α-hydro-
xyphosphonate via the phosphite tautomer coordinated to Ru(II), E1,
seems indeed more likely. The fact that the reaction occurs at 80° but
not at room temperature (see Table 1) accounts for the relatively high
calculated barrier (∼48 kcal mol−1). However, we are aware that this
value is not totally in agreement with the experimental reaction times
and conversions reported in the tables above. We will come back to this
issue below. Nonetheless, the ability of Ru(II) to shift the tautomeric
equilibrium towards the phosphite, given the stabilization of this form
through 4dπ(Ru)···3dπ(P) back-bonding with the coordination of the P
atom [35], is likely to be the key for the catalytic activity. The im-
portance of the Ru(II)-promoted tautomeric shift is further evidenced
by analysis of the calculated non-catalyzed reaction profiles (Fig. 8). In
solution, the H-phosphonate tautomer is present but this form presents
a higher barrier for the formation of the α-hydroxyphosphonate. On the
other hand, the phosphite tautomer, although never experimentally
observed [33], is highly reactive presenting very low barriers.
As mentioned earlier, the calculated barrier of 48 kcal mol−1
coordinated α-hydroxyphosphonate containing
a long Ru-O bond
(2.268 Å) and a short P = O bond of 1.527 Å, comparable to the one
obtained for B (1.519 Å) or in the final α-hydroxyphosphonate product
(1.510 Å). De-coordination of the final product from D regenerates A.
The calculated mechanism presented in Fig. 4 seems unlikely, given
the high barrier calculated for the conversion of C to D via TSCD. Dialkyl
H-phosphonates possess low P–H acidity and a positive charge at the
phosphorus atom [33], hindering the reaction via this tautomer. Indeed,
our calculated Mulliken charges at the P atom in free HP(O)(OMe)2 or
coordinated in B were 1.023 and 1.068, respectively.
Alternative pathways, consistent with the experimental conditions
and the existence of a species with an m/z = 601 peak in the MS spectra
were thus considered. The coordination of HP(O)(OMe)2 offers other
possibilities owing to the existence of the tautomerization process be-
tween dimethyl H-phosphonate, HP(O)(OMe)2, and dimethyl phos-
phite, P(OH)(OMe)2. However, the non-catalyzed tautomerization has a
barrier of 62.2 kcal mol−1, calculated at the MP2 level of theory [34],
or 59.1 kcal mol−1 when calculated at the B3LYP level in solution [14].
Additionally, the phosphite tautomer has higher energy and therefore,
the H-phosphonate tautomer was considered in the mechanism pre-
sented in Fig. 4, as in the hydrophosphonylation of aldehydes promoted
by MoO2Cl2 [14]. Nonetheless, phosphite tautomerisation is often re-
ferred to explain the reactivity of these compounds [33]. More strik-
ingly, Ru(II) is able to shift the tautomeric equilibrium towards the
phosphite form through coordination by the P atom [35]. Indeed,
83

I.R. Cabrita et al.
MolecularCatalysis450(2018)77–86
Fig. 7. Relevant distances in the reagents, products, intermediates, and transition states involved in the reactions between E1 and benzaldehyde. In some structures, the CeH hydrogen
atoms were omitted for clarity.
(Fig. 6) is not in full quantitative agreement with the reaction times and
conversions reported in Tables 2 and 3. To check the consistency of our
calculations, we recalculated all reagents, products, intermediates, and
transition states involved in the reactions reported in Figs. 4 and 6 using
different solvents or an extended basis set (vide infra). The newly cal-
culated energies are collected in Table 5 together with those previously
calculated in toluene.
We first investigated the solvent effect on the reaction barriers by
using benzaldehyde and tributylphosphate. The first emulates the
dielectric of the aldehyde reagents whereas the latter aims at emulating
the H-phosphonate (which unfortunately is not available in Gaussian09).
When analyzing the calculated values (columns 2, 3, and 5), especially
those pertaining to the critical transitions states (TSCD and TSFG),
one can observe that the variation is minimal. For TSCD, the ΔH values
barely change going from 99 kcal mol−1 (toluene) to 100 kcal mol−1
(benzaldehyde and tributylphosphate). The ΔG values are also
practically unchanged. More importantly, for TSFG the ΔH value
increases c.a. 2 kcal mol−1 from toluene to benzaldehyde or
tributylphosphate, the same being observed for the ΔG values (from
∼48 kcal mol−1to ∼51 kcal mol−1). Therefore, the reaction mechanism
appears to be relatively solvent independent and the above described
qualitative picture holds: the reaction proceeds via TSFG
.
We also checked the basis set dependence of our values by replacing
the 6–31G* basis set for the larger 6–311G** (see Experimental) using
tributylphosphate as solvent. For TSCD
, ΔH decreases only c.a.
1 kcal mol−1, the same occurring for ΔG. For TSFG, a similar variation
of ∼1 kcal mol−1 is observed for both ΔH and ΔG. This is an indication
that the calculated barriers are also relatively insensitive to the basis set
size. Therefore, in spite the values do not agree per se with the ex-
perimental reaction times and conversions, being higher than expected,
the robustness our calculations provide an approximate quantitative
view, allowing to discriminate between the possible paths
(TSCD vs TSFG). Finally, we should add that is not uncommon for DFT
calculated barriers to be slightly higher than those expected based on
the experimental data, as reported in several recent examples for Ru
catalysts [36].
Fig. 8. Comparison between the non-catalyzed mechanisms involving the H-phosphonate tautomer (left) and the phosphite tautomer (right). ΔH top, bold and ΔG bottom, italics in kcal
mol−1
.
84

I.R. Cabrita et al.
MolecularCatalysis450(2018)77–86
Table 5
Calculated energies (ΔH, bold and ΔG, italics in kcal mol−1) for the various reagents, products, intermediates, and transition states involved in the reactions depicted in Figs. 4 and 6
calculated in toluene, tributylphosphate, and benzaldehyde. A calculation was also performed expanding the original 6–31G* basis set to an 6–311G** (denoted tributylphosphate −6-
311G**).
Toluene
Tributylphosphate
Tributylphosphate
-6-311G**
Benzaldehyde
Fig. 4
1 + aldehyde + HP(O)(OMe)2
0
0
0
0
A + aldehyde + HP(O)(OMe)2
B
C
TSCD
23.7 (5.9)
13.8 (8.6)
9.9 (14.3)
98.9 (110.2)
2.7 (12.1)
12.6 (8.0)
−11.5 (2.1)
22.4 (4.1)
13.1 (8.7)
9.7 (14.3)
100.0 (110.3)
3.3 (12.8)
12.3 (6.9)
−10.1 (2.8)
20.9 (3.1)
16.6 (7.3)
9.8 (14.2)
98.6 (109.5)
0.9 (10.5)
8.9 (3.7)
−12.1 (0.6)
22.1 (4.7)
12.8 (8.8)
9.5 (14.0)
100.3 (111.7)
3.4 (13.1)
12.3 (7.7)
−9.8 (3.0)
D
A + product
1 + product
Fig. 6
1 + aldehyde + P(OH)(OMe)2
0
0
0
0
A + aldehyde + P(OH)(OMe)2
E1
F
TSFG
23.7 (5.9)
−23.8 (−24.9)
−22.8 (−14.2)
35.9 (47.7)
3.4 (8.8)
22.4 (4.1)
−22.3 (-23.5)
−22.5 (-14.2)
37.8 (50.5)
4.2 (10.0)
6.2 (1.4)
20.9 (3.1)
−22.0 (-23.2)
-21.2 (−13.8)
39.1 (51.8)
4.6 (10.3)
6.3 (1.7)
22.1 (4.7)
−21.5 (-22.2)
−22.4 (-13.8)
38.3 (50.9)
4.5 (10.3)
6.3 (2.4)
G
A + product
1 + product
6.2 (2.0)
−17.5 (−3.8)
−16.2 (-2.7)
−14.6 (1.3)
−15.7 (-2.3)
4. Conclusion
[7] S.-S. Weng, G.-Y. Lin, H.-C. Li, K.-C. Yang, T.-M. Yang, H.-C. Liu, S.-H. Sie, Appl.
Organometal. Chem. 26 (2012) 455–460.
[8] S.K.T. Basha, S. Rasheed, K.C. Sekhar, C.N. Raju, M.S. Ali, C.A. Rao, Sulfur Phosphorus
Silicon 189 (2014) 1546–1556.
[9] B. Liu, J.-F. Carpentier, Y. Sarazin, Chem. Eur. J. 18 (2012) 13259–13264.
[10] Q. Wu, J. Zhou, Z. Yao, F. Xu, Q. Shen, J. Org. Chem. 75 (2010) 7498–7501.
[11] (a) L. Chen, N. Zhao, Q. Wang, G. Hou, H. Zi, G. Song, Inorg. Chim. Acta 402 (2013)
140–155;
(b) T. Yokomatsu, T. Yamagishi, S. Shibuya, J. Chem. Soc. Perkin Trans. 1 (1997)
1527–1534.
[12] W. Chen, Y. Hui, X. Zhou, J. Jiang, Y. Cai, X. Liu, L. Lin, X. Feng, Tetrahedron Lett. 51
(2010) 4175–4178.
In conclusion, we have developed the first method for the synthesis
of a large variety of α-hydroxyphosphonates catalyzed by ruthenium(II)
complexes in good to excellent yields, tolerating a large variety of
functional groups. This work demonstrates that Ru(II)Cp complexes are
efficient catalysts for the synthesis of α-hydroxyphosphonates and for
C-P bond formation. This novel method has other advantages such as
the use of the catalyst [RuClCp(PPh3)2] in at least 12 catalytic cycles
with good activity and the reactions were carried out under solvent free
conditions.
[13] (a) B. Saito, T. Katsuki, Angew. Chem. Int. Ed 44 (2005) 4600–4602.
[14] R.G. Noronha, P.J. Costa, C.C. Romão, M.J. Calhorda, A.C. Fernandes, Organometallics 28
(2009) 6206–6212.
[15] G.S. Ashby, M.I. Bruce, I.B. Tomkins, R.C. Walli, Aust. J. Chem. 32 (1979) 1003–1016.
[16] V. Moreno, M. Font-Bardia, T. Teresa Calvet, J. Lorenzo, F.X. Avilés, M.H. Garcia,
T.S. Morais, A. Valente, M.P. Robalo, J. Inorg. Biochem 105 (2011) 241–249.
[17] P.R. Florindo, I.J. Marques, C.D. Nunes, A.C. Fernandes, J. Organomet. Chem. 760 (2014)
240–247.
DFT calculations showed that the Ru(II) catalyst promotes the shift
of the tautomeric equilibrium towards the phosphite, forming a P-atom
coordinated active species (E1) which is the key for the observed cat-
alytic activity.
[18] P.R. Florindo, D.M. Pereira, P.M. Borralho, C.M.P. Rodrigues, M.F.M. Piedade,
A.C. Fernandes, J. Med. Chem. 58 (2015) 4339–4347.
[19] R.G. Parr, W. Yang, Density Functional Theory of Atoms and Molecules, Oxford
University Press, New York, 1989.
Acknowledgements
[20] M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb, J.R. Cheeseman,
G. Scalmani, V. Barone, B. Mennucci, G.A. Petersson, H. Nakatsuji, M.X. Li, H.P. Caricato,
A.F. Hratchian, J. Izmaylov, G. Bloino, J.L. Zheng, M. Sonnenberg, M. Hada, K. Ehara,
R. Toyota, J. Fukuda, M. Hasegawa, T. Ishida, Y. Nakajima, O. Honda, H. Kitao, T. Nakai,
J.A. Vreven, J.E. Montgomery, F. Peralta, M. Ogliaro, J.J. Bearpark, E. Heyd,
K.N. Brothers, V.N. Kudin, R. Staroverov, J. Kobayashi, K. Normand, A. Raghavachari,
J.C. Rendell, S.S. Burant, J. Iyengar, M. Tomasi, N. Cossi, J.M. Rega, M. Millam,
J.E. Klene, J.B. Knox, V. Cross, C. Bakken, J. Adamo, R. Jaramillo, R.E. Gomperts,
O. Stratmann, A.J. Yazyev, R. Austin, C. Cammi, J.W. Pomelli, R.L. Ochterski, K. Martin,
V.G. Morokuma, G.A. Zakrzewski, P. Voth, J.J. Salvador, S. Dannenberg, A.D. Dapprich,
J.B. Daniels, J.V. Ö. Farkas, J. Foresman, D.J. Ortiz, Cioslowski Fox, Gaussian 09,
Gaussian, Inc., Wallingford CT, 2009.
[21] J.P. Perdew, K. Burke, M. Ernzerhof, Phys. Rev. Lett. 77 (1996) 3865–3868.
[22] C. Adamo, V. Barone, J. Chem. Phys. 110 (1999) 6158–6169.
[23] L.E. Roy, P.J. Hay, R.L. Martin, J. Chem. Theory Comput. 4 (2008) 1029–1031.
[24] (a) D. Feller, J. Comp. Chem 17 (1996) 1571–1586.
[25] J. Tomasi, B. Mennucci, R. Cammi, Chem. Rev. 105 (2005) 2999–3093.
[26] P.R. Florindo, P.J. Costa, M.F.M. Piedade, M.P. Robalo, Inorg. Chem. 56 (2017)
6849–6863.
This research was supported by Fundação para a Ciência e a Tecnologia
through projects UID/QUI/00100/2013 and UID/MULTI/00612/2013.
Ivânia Cabrita (SFRH/BD/74280/2010) and Pedro Florindo thank FCT for
grants. Ana Fernandes (IF/00849/2012) and Paulo Costa (IF/00069/2014)
acknowledge FCT for the “Investigador FCT” Program. Authors also thank
the Portuguese MS and NMR Networks (Node IST and IST-UTL Center) for
providing access to the MS and NMR facilities.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in the
online version, at http://dx.doi.org/10.1016/j.mcat.2018.03.001.
References
[27] M.I. Bruce, F.S. Wong, B.W. Skelton, A.H. White, J. Chem. Soc. Dalton Trans. (1981)
1398–1405.
[28] E.V. Gorbachuk, E.K. Badeeva, R.G. Zinnatullina, P.O. Pavlova, A.B. Dobrynin,
A.T. Gubaidullin, M.A. Ziganshin, A.V. Gerasimov, O.G. Sinyashin, D.G. Yakhvarov, J.
Organomet. Chem. 805 (2016) 49–53.
[29] C.R. Groom, I.J. Bruno, M.P. Lightfoot, S.C. Ward, Acta Cryst. B 72 (2016) 171–179. The
Cambridge Structural Database.
[30] A. Tenaglia, S.J. Marc, Org. Chem. 71 (2006) 3569–3575.
[31] F. Boeck, T. Kribber, L. Xiao, L. Hintermann, J. Am. Chem. Soc. 133 (2011) 8138–8141.
[32] W.-H. Mu, D.-C. Fang, S.-Y. Xia, R.-J. Cheng, G.A. Chass, J. Chem Eur, 22 (2016)
15396–15403.
[1] P. Merino, E. Marques-Lopez, R.P. Herrera, Adv. Synth. Catal. 350 (2008) 1195–1208.
[2] (a) A. Szymanska, M. Szymczak, J. Boryski, J. Stawinski, A. Kraszewski, G. Collu,
G. Sanna, G. Giliberti, R. Loddo, P. La Colla, Bioorg. Med. Chem. 14 (2006) 1924–1934.
[3] (a) B. Kaboudin, Tetrahedron Lett. 44 (2003) 1051–1053.
[4] A. Kumar, S. Jamwal, S. Khan, N. Singh, V.K. Rai, Phosphorus Sulfur Silicon Relat. Elem.
192 (2017) 381–385.
[5] M. Pandi, P.K. Chanani, S. Govindasamy, Appl. Catal. A: General 441–442 (2012)
119–123.
[6] S. Santhisudha, P. Sreelakshmi, S.H. Jayaprakash, B.V. Kumar, C.S. Reddy, Phosphorus
Sulfur Silicon Relat. Elem. 190 (2015) 1479–1488.
[33] K.D. Troev, 3 – Reactivity of H-Phosphonates, In Chemistry and Application of H-
85

I.R. Cabrita et al.
MolecularCatalysis450(2018)77–86
Rosal, C. Faradji, G. Alcaraz, L. Vendier, S. Sabo-Etienne, ACS Catal. 8 (2018) 939–948. b)
ΔGcalc = 34.1 kcal mol−1, reaction time = 20 h, conversion = 98%, T = 80 ºC, in X. Tan,
W. Zeng, X. Zhang, L. W. Chung, X. Zhang, Chem. Commun. 54 (2018) 535–538.
c)ΔGcalc = 28 kcal mol−1, reaction time = 3 h, conversion = 97%, T = 90 ºC, in L. Piola,
J. A. Fernández-Salas, F. Nahra, A. Poater, L. Cavallo, S. P. Nolanbe, Molecular Catalysis
440 (2017) 184–189.
Phosphonates, Elsevier Science Ltd., Amsterdam, 2006, pp. 23–105.
[34] E.M. Georgiev, J. Kaneti, K. Troev, D.M. Roundhill, J. Am. Chem. Soc. 115 (1993)
10964–10973.
[35] R.L. Sernaglia, D.W. Franco, Inorg. Chem. 28 (1989) 3485–3489.
[36] Recent examples in the literature on Ru catalysts: a) ΔGcalc = 25.1 kcal mol−1, reaction
time = 5 min, conversion = 51% , T = room temperature, in M. Beguerie, C. Dinoi, I. del
86