Methyltriphenylphosphonium Methylcarbonate, an All-In-One Wittig Vinylation Reagent

Article abstract of DOI:10.1002/cssc.201500935

The methyltriphenylphosphonium methylcarbonate salt [Ph₃PCH₃][CH₃OCO₂], obtained directly by quaternarization of triphenylphosphine with dimethylcarbonate, is a latent ylide that promotes Wittig vinylation of aldehydes and ketones. Alkenes are obtained simply by mixing [Ph₃PCH₃][CH₃OCO₂] and the carbonyl and heating in a solvent (no base, no halides, and no inorganic byproducts). Deuterium exchange experiments and the particularly short anion-cation distance measured by XRD in [Ph₃PCH₃][CH₃OCO₂] allowed to explain the nature and reactivity of this species. Green chemistry metrics (atom economy, mass index, environmental factor) indicate that this vinylation procedure is more efficient than comparable ones. Deuterated [Ph₃PCD₃][CH₃OCO₂] promoted the synthesis of deuterated olefins. Add carbonyl and go! The Wittig reaction is widely applied, despite having shortcomings such as low atom economy (AE), the necessity to use solvents, and stoichiometric amounts of waste from generating the ylide. [Ph₃PCH₃][CH₃OCO₂] is a latent phosphorous ylide that promotes Wittig vinylation of aldehydes and ketones by simply heating with the carbonyl. No external base and no halides involved, and the protocol offers good AE, environmental factor, and mass index.

Full text of DOI:10.1002/cssc.201500935

DOI: 10.1002/cssc.201500935
Communications
Methyltriphenylphosphonium Methylcarbonate, an All-In-
One Wittig Vinylation Reagent
Lisa Cattelan,[a] Marco No›,[a] Maurizio Selva,[a] Nicola Demitri,[b] and Alvise Perosa*[a]
The methyltriphenylphosphonium methylcarbonate salt
[Ph3PCH3][CH3OCO2], obtained directly by quaternarization of
triphenylphosphine with dimethylcarbonate, is a latent ylide
that promotes Wittig vinylation of aldehydes and ketones. Al-
kenes are obtained simply by mixing [Ph3PCH3][CH3OCO2] and
the carbonyl and heating in a solvent (no base, no halides, and
no inorganic byproducts). Deuterium exchange experiments
and the particularly short anion-cation distance measured by
XRD in [Ph3PCH3][CH3OCO2] allowed to explain the nature and
reactivity of this species. Green chemistry metrics (atom econo-
my, mass index, environmental factor) indicate that this vinyla-
tion procedure is more efficient than comparable ones. Deuter-
ated [Ph3PCD3][CH3OCO2] promoted the synthesis of deuterat-
ed olefins.
oxide.[9] A catalytic alternative, based on in situ reduction of
the phosphine oxide with a silane,[10] still suffers from the co-
production of stoichiometric amounts of waste, although life-
cycle analysis (LCA) of the process seem to indicate that it is
more efficient than traditional procedures.[11] A second issue is
related to the need for organic solvents; this has met with
only partial solutions through the use of water,[12] biphasic sys-
tems,[13] ionic liquids,[14] and solvent-free conditions.[15] The
third drawback of the Wittig reaction is the need for alkyl hal-
ides in the preparation of the phosphonium,[16] along with the
necessity of an external base to generate the ylide. This leads
to stoichiometric amounts of halide salts that require separa-
tion and disposal. To the best of our knowledge, there are no
reports on halide- and base-free Wittig reactions. Likewise,
other direct procedures to convert a carbonyl into a methylene
group involve complex and expensive reagents such as
Tebbe’s complex, which is water- and air-sensitive.[17]
While its mechanism is still an open topic,[1] the Wittig reac-
tion[2] is widely applied in both the academic[3] and industrial[4]
sectors for the olefination of carbonyl substrates, including
formaldehyde,[5] aldehydes,[6] and ketones.[7] Even so, the Wittig
reaction suffers from a number of shortcomings, summarized
in Figure 1.
We have previously established the halide-free synthesis of
phosphonium and ammonium ionic liquids using dimethylcar-
bonate (CH3OCO2CH3, DMC) as a green methylating agent.[18]
The resulting compounds have the structure [R3QCH3]
[CH3OCO2] (Q=N, P) and behave as very efficient basic organo-
catalysts.[19] This catalytic behavior was explained by a coopera-
tive double activation mechanism of the substrate.[20]
In particular, a primary issue is the reaction’s low atom econ-
omy (AE),[8] due to the unavoidable formation of phosphine
Considering the structures of these methyl-phosphonium
compounds from a different perspective (Scheme 1, left box)[18]
Scheme 1. Phosphonium–ylide equilibrium.
Figure 1. Major shortcomings of the traditional Wittig reaction protocol.
suggests that they incorporate both the basic and phosphoni-
um moieties indispensable for the preparation of phosphorous
ylides. We therefore imagined that the methylcarbonate anion
might be able to generate the ylide in situ, as shown in
Scheme 1, and thus give access to a new Wittig vinylation pro-
tocol.
[a] L. Cattelan, Dr. M. No›, Prof. M. Selva, Prof. A. Perosa
Dipartimento di Scienze Molecolari e Nanosistemi
Università Ca’ Foscari Venezia
Dorsoduro 2137, 30123 Venezia (Italy)
E-mail: alvise@unive.it
Because methyltriphenylphoshponium is one of the most
commonly reported Wittig ylide precursors, we started by de-
veloping its synthesis by methylation of triphenylphosphine
(TPP) with DMC as a generalization of our previous results with
trialkylphosphines (Scheme 2).[18] Optimization of the reaction
[b] N. Demitri
Elettra—Sincrotrone Trieste, S.S. 14 Km 163.5 in Area
Science Park, 34149 Basovizza—Trieste (Italy)
Supporting Information and ORCIDs from the authors of this article are
available on the WWW under http://dx.doi.org/10.1002/cssc.201500935.
ChemSusChem 2015, 8, 3963 – 3966
3963
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communications
Scheme 2. Synthesis of 1 from TPP and DMC.
Figure 3. 1H NMR spectra of [Ph3PCH3][CH3OCO2] 1 in the P-CH3 region ac-
quired in CDCl3 at 10 min time intervals.
Figure 2. X-ray crystal structure (50% probability ellipsoids) of 1 (100 K, C:
grey, P: yellow, O: red, H: white).
conditions by 1H NMR monitoring (TPP: 89.6 mmol, DMC:
617 mmol, CH3OH: 52 mL, N2 atmosphere, 1408C for 24 h) led
to an 89% yield of 1 as a white powder. Recrystallization yield-
ed 1 suitable for crystal structure determination by X-ray dif-
fraction (XRD) (Figure 2).
Scheme 3. Hydrogen–deuterium exchange mechanism.
The crystal structure of 1 (Figure 2) highlighted a short dis-
tance between the methylcarbonate oxygen and the PÀCH3
methyl group (3.174 Š at 100 K), suggesting significant hydro-
gen-bonding in the solid state.[21] Comparison with the distan-
ces between the PÀCH3 methyl group and the anion in other
[Ph3PCH3][X] salts (Supporting Information, Table S1 and refer-
ences therein) further supports this peculiar proximity. This
offers strong evidence for the possibility of “counterion medi-
ated” deprotonation, as hypothesized in Scheme 1, and there-
fore in favor of the existence of a latent ylide able to promote
Wittig olefination.
1H NMR analysis of 1 provided additional evidence of the
latent ylide intermediate postulated in Scheme 1. Specifically,
1H NMR of 1 in the resonance region of the PÀCH3 protons
(3.29–3.39 ppm) collected in [D6]DMSO showed a doublet (J=
14.79 Hz) as expected, due to coupling of the methyl protons
with 31P (Supporting Information, Figure S7). Instead, the same
spectral region collected in CDCl3 as solvent showed a time-de-
pendent disappearance of the PÀCH3 protons signal (Figure 3).
This disappearance is attributed to hydrogen–deuterium ex-
change between the PÀCH3 protons (pKa >20) in 1 and deute-
rium in CDCl3 (pKa =25), in analogy to other studies[22] and
with the methylcarbonate anion acting as the base.[23]
phosphoranes.[22,24] Others have explained a similar carbonate-
induced deprotonation of a phosphonium by invoking a pre-
ceding decarboxylation of the alkylcarbonate.[23b] In our opin-
ion, the available data points towards deprotonation occurring
first. In the reverse process, 1-d3 was dissolved in CHCl3 to
yield 1 back again.
Next, benzaldehyde 2a was chosen as the model carbonyl
substrate to test the Wittig olefination with 1.[25] Prescreening
of different solvents (Supporting Information, Figure S3) indi-
cated that in biobased 2Me-THF[26] the reaction was extremely
effective both in terms of conversion and selectivity, notwith-
standing that 1 is only sparingly soluble in all the tested sol-
vents. By simply mixing 1 and 2a at the reflux temperature of
2Me-THF (808C), total conversion of benzaldehyde 2a to yield
styrene was observed after 20 min, with 100% selectivity
(Scheme 4).
The scope of this procedure was extended by reacting
a series of aldehydes (4-nitro-benzaldehyde, 2b; 4-chloro-ben-
zaldehyde, 2c; 4-methoxy-benzaldehyde, 2d; piperonal, 2e;
furfural 2 f, 1-decanal 2g; 3-phenylbutanal 2h) with 1 in 2Me-
THF to obtain the corresponding vinyl derivatives 3b–3h
(Scheme 5).
H/D exchange was further supported by treating the phos-
phonium 1 with a 15-fold molar excess of CDCl3 that yielded
quantitatively (1.34 g) the deuterated analog [Ph3PCD3]
[CH3OCO2] 1-d3. The equilibrium postulated in Scheme 1 trig-
gers complete H/D exchange by successive steps (Scheme 3),
in analogy to what was previously described for hindered
Scheme 4. Wittig reaction between benzaldehyde and [Ph3PCH3] [CH3OCO2]
1.
ChemSusChem 2015, 8, 3963 – 3966
www.chemsuschem.org
3964
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communications
Table 1. Calculated green metrics for the conversion of 2e into 3e.[a]
Reference
Atom
economy (AE)
Environmental
factor (EF)
Mass
index (MI)
[31]
[27]
[32]
This work
22.9%
25.9%
25.9%
29.5%
4.34
2.99
3.90
3.82
15.85
15.70
19.31
9.60
[a] Metrics were calculated by using the method developed by An-
draos.[33] For details see Figures S21–S23 (Supporting Information).
atom efficient (29.5%) due to the absence of halides and of
added base. Sheldon’s EF is favored in Aslam’s procedure due
to a staggering 97% yield using purely stoichiometric amounts
(1.0 equiv) of phosphonium salt and BuLi. On the contrary, the
MI of our protocol was significantly more advantageous (9.60)
than the other three (15.70–19.31). This was due to the very
high conversion and selectivity achieved with a low amount of
solvent (2e/2Me-THF=0.150 gmLÀ1) that also favored the re-
action rate. In all cases, chromatography solvents were not in-
cluded in the calculation.
Scheme 5. Wittig vinylation products obtained from different aldehydes
(%yields and reaction time are indicated), conditions: 808C, 1 atm. N2, 5 mL
2Me-THF, 1: substrate ratio=1.1. [a] Yield as determined by GC at 100% se-
lectivity. [b] Isolated yield.
Reactivity followed the order 2d !2g<2h<2e<2 f<2a<
2c<2b. Aromatic aldehydes with electron-withdrawing
groups (EWGs) reacted faster than ones with electron-donating
groups (EDGs) and aliphatic aldehydes required longer reaction
times. Notably, aliphatic aldehydes 2g and 2h did not show
any competing aldol-type condensation reactions despite the
basicity of 1. To confirm the mass balance a selection of prod-
ucts including 4-NO2-styrene 3b, methylene-dioxystyrene 3e,
1-undecene 3g and 3-phenyl-pent-1-ene 3h were isolated, pu-
rified, fully characterized, and obtained in 94%, >98%, 78%
and 87% yields, respectively.
A complete evaluation in terms of efficiency and environ-
mental impact would require a full life-cycle analysis (LCA), as
was shown elsewhere.[11] However, the complexity of such an
investigation is meaningful for specific industrial processes,
and is therefore rather beyond the scope of the present work.
With deuterated [Ph3PCD3][CH3OCO2] 1-d3 in hand, we ap-
plied our protocol to the synthesis of 2,2-dideutero-methylene-
dioxystyrene 3e-d2, which was obtained in 75% yield and 90%
isotopic purity (by NMR) (Scheme 7).
Methylene-dioxystyrene 3e, a pharmaceutical intermedi-
ate,[26] was selected to compare the efficiency of the present
protocol with classic Wittig ones. The reported synthesis of 3e
involves the Wittig reaction of [Ph3PCH3][X] with BuLi at 08C
followed by the addition of piperonal 2e. In our case, by
simply mixing 1 and 2e in 2Me-THF and heating at reflux,
>98% isolated yield of methylene-dioxystyrene 3e was ob-
tained on a multigram scale (~3 g, 20 mmol), without forma-
tion of halide salts (Scheme 6).
Scheme 7. Preparation of deuterated 3e-d2 by reaction of 2e with 1-d3.
In order to take a snapshot of the potential environmental
impact, we then compared the efficiency of our protocol with
three existing methods for the synthesis of 3e by using simple
green chemistry metrics:[28] AE,[29] environmental factor (EF),[30]
and mass index (MI). The results are listed in Table 1.
The scope of the new Wittig vinylation reaction was further
expanded by testing the reactivity of 1 with two representative
ketones: acetophenone 2i and benzophenone 2j. The results
are summarized in Scheme 8. The vinylation of ketones re-
quired to increase the molar ratio 1/ketone up to 3 in order to
achieve 100% conversion and 100% selectivity toward the de-
sired products (Scheme 8). A third representative ketone, cyclo-
hexanone, did not undergo Wittig olefination, affording only
unidentifiable acyclic products.
AE is always low due to the formation of triphenylphosphine
oxide (TPPO), nonetheless our procedure is still slightly more
In summary, the phosphonium salt 1—prepared by a halide-
free procedure—is a latent ylide able to perform Wittig-type vi-
nylation reactions simply by heating it at reflux with a variety
of carbonyls. The solvent of choice was biobased 2Me-THF, but
Scheme 6. Synthesis of 3e from piperonal 2e in the presence of 1.
ChemSusChem 2015, 8, 3963 – 3966
www.chemsuschem.org
3965
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communications
[5] a) A. Streitwieser, R. S. McDowell, THEOCHEM 1986, 138, 89–95; b) F.
Volatron, O. Eisenstein, J. Am. Chem. Soc. 1987, 109, 1–14.
[6] a) A. El-Batta, C. Jiang, W. Zhao, R. Anness, A. L. Cooksy, M. Bergdahl, J.
Org. Chem. 2007, 72, 5244–5259; b) M. Schlosser, B. Schaub, J. Am.
Chem. Soc. 1982, 104, 5821–5823.
[7] a) Z. Huang, R. Sun, J. Chem. Res. 2003, 40–41; b) D. Enders, M. Bartsch,
D. Backhaus, J. Runsink, G. Raabe, Synthesis 1996, 1438–1442; c) S.
Park, D. Yang, K. T. Kim, H. B. Jeon, Tetrahedron Lett. 2011, 52, 6578–
6580.
[8] M. Eissen, R. Mazur, H. G. Quebbemann, K. H. Pennemann, Helv. Chim.
Acta 2004, 87, 524–535.
[9] M. B. Smith, March’s Advanced Organic Chemistry: Reactions, Mecha-
nisms, and Structure, Wiley, 2013.
[10] C. J. O’Brien, J. L. Tellez, Z. S. Nixon, L. J. Kang, A. L. Carter, S. R. Kunkel,
K. C. Przeworski, Angew. Chem. Int. Ed. 2009, 48, 6836–6839; Angew.
Chem. 2009, 121, 6968–6971.
Scheme 8. Wittig reaction between [Ph3PCH3][CH3OCO2] 1 and ketones 2i–j
[a] molar ratio 1/ketone=1.1 [b] molar ratio 1/ketone=3. Conditions: 808C,
1 atm. N2, 5 mL 2Me-THF; conversion as determined by GC.
[11] H. A. van Kalkeren, A. L. Blom, F. P. J. T. Rutjesa, M. A. J. Huijbregts, Green
Chem. 2013, 15, 1255–1263.
the reaction was shown to proceed also in dimethylcarbonate
(DMC), EtOAc, cyclopentylmethyl ether, toluene and in a 2Me-
THF/EtOH mixture. The absence of halides throughout the
whole process guaranteed the lack of formation of inorganic
halide salts. The vinylation products were isolated from the
crude reaction mixture by a simple filtration over silica to
remove triphenylphosphine oxide (TPPO) and residual 1. Syn-
thesis of deuterated 1-d3 confirmed formation of the ylide in-
termediate through “counterion-mediated” deprotonation and
allowed us to prepare deuterated olefins using CDCl3 as an in-
expensive and readily available deuterium source. The straight-
forward synthesis of 1 and its simple handling should enable
the use of this new vinylation reagent in academic and indus-
trial research and development laboratories, as well as in ad-
vanced teaching laboratories.
[12] G. Molander, R. Oliveira, Tetrahedron Lett. 2008, 49, 1266–1268.
[13] Y. Moussaoui, K. Sa¨ıd, R. B. Salem, Arkivoc 2006, 12, 1–22.
[14] V. Le Boulaire, R. GrØe, Chem. Commun. 2000, 2195–2196.
[15] V. Balema, J. Wiench, M. Pruski, J. Am. Chem. Soc. 2002, 124, 6244–
6245.
[16] V. K. Aggarwal, J. N. Harvey, R. Robiette, Angew. Chem. Int. Ed. 2005, 44,
5468–5471; Angew. Chem. 2005, 117, 5604–5607.
[17] S. H. Pine, G. S. Shen, H. Hoang, Synthesis 1991, 165–167.
[18] M. Fabris, V. Lucchini, M. No›, A. Perosa, M. Selva, Chem. Eur. J. 2009,
15, 12273–12282.
[19] M. Selva, A. Caretto, M. No›, A. Perosa, Org. Biomol. Chem. 2014, 12,
4143–4155.
[20] V. Lucchini, M. No›, M. Selva, M. Fabris, A. Perosa, Chem. Commun.
2012, 48, 5178–5180.
[21] G. Adamovµ, R. L. Gardas, M. Nieuwenhuyzen, A. V. Puga, L. P. N. Rebelo,
A. J. Robertson, K. R. Seddon, Dalton Trans. 2012, 41, 8316–8332.
[22] I. Szele, S. J. Kubisen, F. H. Westheimer, J. Am. Chem. Soc. 1976, 98,
3533–3536.
[23] a) L. Jedinak, L. T. Rush, M. Lee, D. Hesek, J. F. Fisher, B. Boggess, B. C.
Noll, S. Mobashery, J. Org. Chem. 2013, 78, 12224–12228; b) E. E. Coyle,
B. J. Doonan, A. J. Holohan, K. A. Walsh, F. Lavigne, E. H. Krenske, C. J.
O’Brien, Angew. Chem. Int. Ed. 2014, 53, 12907–12911; Angew. Chem.
2014, 126, 13121–13125.
Acknowledgements
Dr. Marco Bortoluzzi is gratefully acknowledged for help with
2H NMR spectra.
[24] D. Phillips, I. Szele, F. H. Westheimer, J. Am. Chem. Soc. 1976, 98, 184–
189.
[25] H. Yamataka, K. Nagareda, K. Ando, T. Hanafusa, J. Org. Chem. 1992, 57,
2865–2869.
Keywords: green chemistry · halides · phosphorus · Wittig
reactions · ylides
[26] V. Pace, P. Hoyos, L. Castoldi, P. D. de Marìa, A. R. Alcàntara, ChemSus-
Chem 2012, 5, 1369–1379.
[27] S. N. Aslam, P. C. Stevenson, S. J. Phythian, N. C. Veitch, D. R. Hall, Tetra-
hedron 2006, 62, 4214–4226.
[28] Green Chemistry Metrics, R. K. Henderson, D. J. C. Constable, C. JimØ-
nez-Gonzµlez in Green Chemistry in the Pharmaceutical Industry, Wiley-
VCH, Weinheim, 2010.
[29] Atom economy is defined as: AE=[MW(gmolÀ1)product]/[S
MW(gmolÀ1) of all reagents]”100. See also B. M. Trost, Science 1991,
254, 1471–1477.
[30] R. A. Sheldon, Green Chem. 2007, 9, 1273–1283.
[31] Y. Le Bigot, M. Delmas, A. Gaset, Synth. Commun. 1982, 12, 107–112.
[32] C. A. Faler, M. M. JoulliØ, Org. Lett. 2007, 9, 1987–1990.
[33] a) J. Andraos, M. Sayed, J. Chem. Educ. 2007, 84, 1004–1010; b) J. An-
draos, Org. Process Res. Dev. 2009, 13, 161–185.
[1] D. G. Gilheany, Chem. Rev. 1994, 94, 1339–1374.
[2] G. Wittig, G. Geissler, Justus Liebigs Ann. Chem. 1953, 580, 44–57.
[3] a) Z. Chen, Y. Nieves-Quinones, J. R. Waas, D. A. Singleton, J. Am. Chem.
Soc. 2014, 136, 13122–13125; b) P. A. Byrne, D. G. Gilheany, Chem. Soc.
Rev. 2013, 42, 6670–6696; c) G. W. Wong, C. R. Landis, Angew. Chem. Int.
Ed. 2013, 52, 1564–1567; Angew. Chem. 2013, 125, 1604–1607; d) P. A.
Byrne, Investigation of Reactions Involving Pentacoordinate Intermediates,
Springer, 2012, pp. 1–92; e) P. A. Byrne, D. G. Gilheany, J. Am. Chem.
Soc. 2012, 134, 9225–9239; f) J. McNulty, P. Das, D. McLeod, Chem. Eur.
J. 2010, 16, 6756–6760; g) J. J. Cao, F. Zhou, J. Zhou, Angew. Chem. Int.
Ed. 2010, 49, 4976–4980; Angew. Chem. 2010, 122, 5096–5100; h) A.
Maercker, The Wittig reaction, Wiley, 1965.
[4] a) H. Pommer, Angew. Chem. Int. Ed. Engl. 1977, 16, 423–429; Angew.
Chem. 1977, 89, 437–443; b) H. J. Bestmann, L. Kiesilowmski, W. Distler,
Angew. Chem. Int. Ed. Engl. 1976, 15, 298–299; Angew. Chem. 1976, 88,
297–298; c) H. Yokoyama, P. Calif, Carotene derivative and use thereof.
3321315, 1967.
Received: July 10, 2015
Published online on September 18, 2015
ChemSusChem 2015, 8, 3963 – 3966
www.chemsuschem.org
3966
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim