Lanthanide triflate-catalyzed arene acylation. Relation to classical Friedel-Crafts acylation

Article abstract of DOI:10.1021/jo800158k

(Chemical Equation Presented) Lanthanide trifluoromethanesulfonates, Ln(OTf)₃ (OTf^- = trifluoromethanesulfonate), serve as effective precatalysts for the rapid, regioselective, intermolecular acylation of activated arenes. This contribution probes mechanism and metal ionic radius effects in the catalytic lanthanide triflate-mediated acylation of anisole with acetic anhydride. Kinetic studies of Ln(OTf)₃ (Ln = La, Eu, Yb, Lu)-mediated anisole acylation with acetic anhydride in nitromethane reveal the rate law ν ～ k₃ [Ln³⁺]¹[acetic anhydride]¹[anisole]¹. Eyring and Arrhenius analyses yield ΔH? = 12.9 (4) kcal·mol^-1, ΔS? = -44.8 (1.3) e.u., and E_a = 13.1 (4) kcal·mol^-1 for Ln = Yb, with the negative ΔS? implying a highly organized transition state. The observed primary kinetic isotope effect of k _H/k_D = 2.6 ± 0.15 is consistent with arene C-H bond scission in the turnover-limiting step. The proposed catalytic pathway involves precatalyst formation via interaction of Ln(OTf)₃ with acetic anhydride, followed by Ln³⁺-anisole π-complexation, substrate-electrophile σ-complex formation, and turnover-limiting C-H bond scission. Lanthanide size effects on turnover frequencies are consistent with a transition state lacking significant ionic radius-dependent steric constraints. Substrate-Ln³⁺ interactions using paramagnetic Gd³⁺ and Yb³⁺ NMR probes and factors affecting reaction rates such as arene substituent and added LiClO₄ cocatalyst are also explored.

Full text of DOI:10.1021/jo800158k

Lanthanide Triflate-Catalyzed Arene Acylation. Relation to Classical
Friedel–Crafts Acylation
Alma Dzudza and Tobin J. Marks*
Department of Chemistry, Northwestern UniVersity, 2145 Sheridan Road, EVanston, Illinois 60208-3113
t-marks@northwestern.edu
ReceiVed January 22, 2008
Lanthanide trifluoromethanesulfonates, Ln(OTf)3 (OTf- ) trifluoromethanesulfonate), serve as effective
precatalysts for the rapid, regioselective, intermolecular acylation of activated arenes. This contribution probes
mechanism and metal ionic radius effects in the catalytic lanthanide triflate-mediated acylation of anisole
with acetic anhydride. Kinetic studies of Ln(OTf)3 (Ln ) La, Eu, Yb, Lu)-mediated anisole acylation with
acetic anhydride in nitromethane reveal the rate law ν ∼ k3 [Ln3+]1[acetic anhydride]1[anisole]1. Eyring and
Arrhenius analyses yield ∆H‡ ) 12.9 (4) kcal ·mol-1, ∆S‡ ) -44.8 (1.3) e.u., and Ea ) 13.1 (4) kcal ·mol-1
for Ln ) Yb, with the negative ∆S‡ implying a highly organized transition state. The observed primary kinetic
isotope effect of kH/kD ) 2.6 ( 0.15 is consistent with arene C-H bond scission in the turnover-limiting step.
The proposed catalytic pathway involves precatalyst formation via interaction of Ln(OTf)3 with acetic anhydride,
followed by Ln3+-anisole π-complexation, substrate-electrophile σ-complex formation, and turnover-
limiting C-H bond scission. Lanthanide size effects on turnover frequencies are consistent with a transition
state lacking significant ionic radius-dependent steric constraints. Substrate-Ln3+ interactions using
paramagnetic Gd3+ and Yb3+ NMR probes and factors affecting reaction rates such as arene substituent
and added LiClO4 cocatalyst are also explored.
Introduction
solvents.3a,b,4 Acylations must be carried out under anhydrous
conditions and require an acylating agent, e.g., an acyl chloride,
acyl anhydride, or carboxylic acid, an aromatic substrate, and
a stoichiometric excess of Lewis acid.3a,b,4 Aluminum trichloride
is often referred to as a “catalyst” even though it is known to
be required in greater than stoichiometric amounts due to
formation of stable Lewis acid-aryl ketone product complexes,
thus necessitating a final hydrolysis step for product isolation.3,4
This reaction is catalytic only if the Lewis acid-aryl ketone
complex is partially dissociable, which is both Lewis acid- and
temperature-dependent.3a,b While it is possible to carry out FC
acylations in the presence of some Lewis acids at higher
temperatures, this usually provokes side reactions.3,4 The
standard FC workup procedure leads to catalyst destruction and
The most important synthetic organic transformations are
arguably those that generate carbon-carbon bonds, such as
reactions which facilitate assembly of new molecular frame-
works.1 In this context, one of the most significant recent
synthetic advances has been the development of broad classes
of Lewis acid-catalyzed reactions, of great interest due to their
unique reactivity and selectivity patterns.1,2
The Friedel–Crafts (FC) acylation of arenes is a fundamental
and useful carbon-carbon bond-forming transformation for the
preparation of aromatic ketones.3 Classical intermolecular FC
arene acylations are typically mediated by strong, moisture-
sensitive Lewis acids such as AlCl3, TiCl4, and BF3 in organic
(1) (a) Lewis Acids in Organic Synthesis; Yamamoto, H., Ed.; Wiley-VCH:
Weinheim, 2000. (b) Lewis Acid Reagents: A Practical Approach; Yamamoto,
H., Ed.; Oxford University Press: Oxford, UK, 1999.
(2) (a) SelectiVities in Lewis Acid Promoted Reactions; Schinzer, D., Ed.;
Kluwer Academic Publishers: Dordrecht, Holland, 1989; pp 73–125. (b) Santelli,
M.; Pons, J.-M. Lewis Acids and SelectiVity in Organic Synthesis; CRC: Boca
Raton, 1995.
(3) (a) Olah, G. A. Friedel-Crafts and Related Reactions, 4th ed.; Interscience:
New York, 1964; Vol 2, Part 2, Chapter 2. (b) Olah, G. A. Friedel-Crafts and
Related Reactions, 4th ed.; Interscience: New York, 1964; Vol. 3, Part. 2, Chapter
36. (c) Nightingale, D. V. Chem. ReV. 1939, 25, 329–376.
(4) Taylor, R. Electrophilic Aromatic Substitution; John Wiley & Sons: New
York, 1990; Chapter 6.
4004 J. Org. Chem. 2008, 73, 4004–4016
10.1021/jo800158k CCC: $40.75 2008 American Chemical Society
Published on Web 04/30/2008

Lanthanide Triflate-Catalyzed Arene Acylation
to the formation of large amounts of acid and metal waste.3,4
In addition, sluggish reaction conditions and the excessive
amounts of organic solvents required in workup procedures and
in product isolation are environmentally problematic for FC
reactions with classical Lewis acids.3,4
preparing precursors to many natural products such as aromatic
ketones and related structural variants such as indanones,
tetralones, hydrindones, etc.13a,d These structural motifs play
major roles in the medicinal, pharmaceutical, and polymer
industries.13
As a promising means to address the aforementioned issues
with conventional Lewis acids, notable progress has been made
in the past few decades.5,6 Among the newly emerged FC
acylation catalyst such as binary Lewis acid mixtures,5a–c air-
stable late transition-metal catalysts,5f new types of environ-
mentally friendly f-element Lewis acids, lanthanide trifluo-
romethanesulfonates, Ln(OTf)3, stand out.6,7b–d,8,9 Lanthanide
ions have far larger ionic radii and formal coordination numbers
than typical transition metal ions.7a–c They are expected to act
as very strong Lewis acids because of their hard, electrophilic
character and yet are relatively stable to hydrolysis.7a–c,8,9
Moreover, the strongly electron-withdrawing properties of the
triflate counterion suggest it will act as a good leaving group,
as is typical of its role in numerous organic transformations.6,8b,10
Furthermore, lanthanide triflates remain catalytically active in
the presence of many oxygen-, nitrogen-, phosphorus-, and
sulfur-containing Lewis bases.7b–d In most cases, only substo-
ichiometric (catalytic) quantities of triflate reagents are sufficient
to effect useful catalytic transformations.6,7b–d,9 Importantly, rare
earth metal triflates can be easily recovered after reaction
completion and recycled without significant loss of activity.6,7b–d,9
Lanthanide triflates are employed effectively in a wide range
of carbon-carbon bond-forming transformations in aqueous,
organic, and biphasic reaction media in which their oxophilicity
and regioselectivity have been exploited extensively.6,7d Koba-
yashi and co-workers have studied the application of lanthanide
triflates as catalysts in organic transformations such as Aldol
reactions, Diels–Alder, retro and aza Diels–Alder reactions,
allylation, etc.6,7,9 While the above reactions proceed readily
in water or in biphasic media, Friedel–Crafts acylation mediated
by lanthanide triflates proceeds only in organic solvents and
ionic liquids.6,11,12 Our attention was drawn to the application
of lanthanide triflates in intermolecular FC acylation, a funda-
mental reaction in organic synthesis and one of great utility in
In comparison to classical Lewis acids, lanthanide triflate-
mediated acylation processes do not require rigorously anhy-
drous reaction conditions since these are water-tolerant Lewis
acids.6,9,11b Furthermore, only substoichiometric amounts of
catalyst are sufficient for complete conversions, and catalysts
can be readily recovered from the aqueous layer after workup,
without catalytic activity loss.6,9,11b Ln(OTf)3 complexes ef-
fectively and selectively catalyze FC acylations in standard
organic media and most efficiently in polar nitromethane.5f,11b
The scope of these catalytic systems has been extensively
explored by Kobayashi and co-workers, generally using acid
anhydrides as acylation agents and electron-rich aromatic
substrates such as anisole.11b The final product is a single
regioisomer in excellent yield, with efficiency alterable either
by increasing the catalyst loading to stoichiometric levels or
by introducing a cocatalyst such as LiClO4, which promotes
acylation of less reactive aromatics, such as xylenes and toluene,
but in certain cases alters the regioselectivity.14 The catalytic
activities of metal triflates can also be enhanced by the addition
of triflic acid and the use of acid chlorides as acylating
reagents.15 Since triflic acid is a known catalyst for these
reactions, the role of the lanthanide center remains unclear in
this case.
To date, little is known about the actual mechanism of
Ln(OTf)3-mediated Friedel–Crafts acylation, and there have been
no kinetic or mechanistic studies to probe the role of the
Ln(OTf)3 catalyst and the interactions/transformations leading
to C-C bond formation. It is known that arenes form π-com-
16
17
plexes with electrophilic ions such as Sm3+ and Ag+ and
with neutral acceptors such as iodine,18 so a goal here is to
define the nature of the interaction between the Ln3+ center and
(11) Ln(OTf)3-mediated FC acylation in organic solvents: (a) Kawada, A.;
Mitamura, S.; Matsuo, J.; Kobayashi, S. Bull. Chem. Soc. Jpn. 2000, 73, 2325–
2333. (b) Kawada, A.; Mitamura, S.; Kobayashi, S. J. Chem. Soc., Chem.
Commun. 1993, 1157–1158.
(5) (a) Izumi, J.; Mukaiyama, T. Chem. Lett. 1996, 739–740. (b) Hachiya,
I.; Moriwaki, M.; Kobayashi, S. Bull. Chem. Soc. Jpn. 1995, 68, 2053–2060.
(c) Suzuki, K.; Kitagawa, H.; Mukaiyama, T. Bull. Chem. Soc. Jpn. 1993, 66,
3729–3734. (d) Mukaiyama, T.; Ohno, T.; Nishimura, T.; Suda, S.; Kobayashi,
S. Chem. Lett. 1991, 1059–1062. (e) Mukaiyama, T.; Nagaoka, H.; Ohshima,
M.; Murakami, M. Chem. Lett. 1986, 165–169. (f) Fürstner, A.; Voigtländer,
D.; Schraider, W.; Giebel, D.; Reetz, M. T. Org. Lett. 2001, 3, 417–420.
(6) Reviews of Ln(OTf)3-mediated reactions: (a) Li, C.-J.; Chen, L. Chem.
Soc. ReV. 2006, 35, 68–82. (b) Luo, S.; Zhu, L.; Ralukdar, A.; Zhang, G.; Mi,
X.; Chen, J.-P.; Wang, P. G. Mini-ReV. Org. Chem. 2005, 2, 177–202. (c) Li, C.
Chem. ReV. 2005, 105, 3095–3165. (d) Kobayashi, S. Chem. ReV. 2002, 102,
2227–2302. (e) Kobayashi, S.; Manabe, K. Acc. Chem. Res. 2002, 35, 209–217.
(f) Topics in Organometallic Chemistry. Lanthanides: Chemistry and Use in
Organic Synthesis; Kobayashi, S., Ed.; Springer-Verlag: Berlin, 1999; Vol.
2, Chapter 2.
(7) Lanthanide chemistry: (a) Imamoto, T. Lanthanides in Organic Synthesis;
Academic Press: San Diego, 1994. (b) Cotton, S. Lanthanide and Actinide
Chemistry; Wiley: Chichester, England, 2006. (c) Aspinall, H. C. Chemistry of
the f-Block Elements; Gordon and Breach Publishers: Amsterdam, 2001. (d)
Aqueous-Phase Organometallic Catalysis; Cornils, B., Herrmann, W. A., Eds.;
Wiley-VCH: New York, 1998.
(12) Ln(OTf)3-mediated FC acylation in ionic liquids (ILs): (a) Hardacre,
C.; Katdare, S. P.; Milroy, D.; Nancarrow, P.; Rooney, D. W.; Thompson, J. M.
J. Catal. 2004, 227, 44–52. (b) Xiao, J.; Ross, J. Green Chem. 2002, 4, 129–
133.
(13) Examples of FC acylation used in natural product synthesis: (a) Jackson,
W. J. Macromolecules 1983, 16, 1027–1033. (b) Goudie, A. C.; Gaster, L. M.;
Lake, A. W.; Rose, C. J.; Freeman, P. C.; Hughes, B. O.; Miller, D. J. Med.
Chem. 1978, 21, 1260–1264. (c) Crenshaw, R. R.; Luke, G. M.; Bialy, G. J. Med.
Chem. 1972, 15, 1179–1180. (d) Harrison, I. T.; Lewis, B.; Nelson, P.; Rooks,
W.; Roszkowski, A.; Tomolonis, A.; Fried, J. H. J. Med. Chem. 1970, 13, 203–
205.
(14) Properties of LiClO4: (a) Malona, J. A.; Colbourne, J. M.; Frontier, A. J.
Org. Lett. 2006, 8, 5661–5664. (b) Komoto, I.; Kobayashi, S. Tetrahedron Lett.
2000, 56, 6463–6465. (c) Nesakumar, E. J.; Sankararaman, S. Eur. J. Org. Chem.
2000, 11, 2003–2001. (d) Kawada, A.; Mitamura, S.; Kobayashi, S. Chem.
Commun. 1996, 183–184. (e) Hachiya, I.; Moriwaki, K.; Kobayashi, S.
Tetrahedron Lett. 1995, 36, 409–412. (f) Ayerbe, M.; Cossío, F. P. Tetrahedron
Lett. 1995, 36, 4447–4450. (g) Mukaiyama, T.; Suzuki, K.; Han, J.-S.; Kobayashi,
S. Chem. Lett. 1992, 3, 435–438.
(15) TfOH-mediated FC acylation: Dumeunier, R. D.; Markó, I. Tetrahedron
Lett. 2004, 45, 825–829.
(8) (a) Molander, G. A. Chem. ReV. 1992, 92, 29–68. (b) Lawrence, G. Chem.
(16) Ln3+-arene complexes: (a) Liang, H.; Shen, Q.; Guan, J.; Lin, Y. J.
Organomet. Chem. 1994, 474, 113–116. (b) Deacon, G. B. Aust. J. Chem. 1990,
43, 1245–1257. (c) Fan, B.; Shen, Q.; Lin, Y. J. Organomet. Chem. 1989, 376,
61–66. (d) Fan, B.; Shen, Q.; Lin, Y. J. Organomet. Chem. 1989, 377, 51–58.
(e) Cotton, A.; Schwotzer, W. Organometallics 1987, 6, 1275–1279. (f) Cotton,
A.; Schwotzer, W. J. Am. Chem. Soc. 1986, 108, 4657–4568.
(17) Ag+-arene complexes: Keffer, R. M.; Andrews, J. J. Am. Chem. Soc.
1952, 74, 640–643.
ReV. 1986, 86, 17–33.
(9) Ln(OTf)3-mediated C-C bond-forming reactions: (a) Kobayashi, S.;
Hachiya, I. J. Org. Chem. 1994, 59, 3590–3596. (b) Engberts, J. B. F. N.; Wijnen,
J. W. J. Org. Chem. 1997, 62, 2039–2044. (c) Bellucci, C.; Cozzi, P. G.; Umani-
Ronchi, A. Tetrahedron Lett. 1995, 36, 7289–7292.
(10) (a) Parker, D.; Dickins, R.; Puschmann, H.; Crossland, C.; Howard, J.
Chem. ReV. 2002, 102, 1977–2010. (b) Kobayashi, S.; Nagayama, S.; Busujima,
T. J. Am. Chem. Soc. 1998, 120, 8287–8288. (c) Baes, C. F.; Mesmer, R. E.
The Hydrolysis of Cations; Wiley: New York, 1976.
(18) I2-arene complexes: Keefer, R. M.; Andrews, L. J. Am. Chem. Soc.
1955, 77, 2164–2168.
J. Org. Chem. Vol. 73, No. 11, 2008 4005

Dzudza and Marks
substrate and to better define the sequence of events leading to
these facile, lanthanide-mediated arene functionalizations. For
the present study, we chose anisole acylation as the representa-
tive reaction under the same reaction conditions as employed
by Kobayashi et al. (eq 1).11b Herein, we present a detailed
kinetic/mechanistic investigation of this reaction with the aim
of elucidating the role of the Ln3+ center in Friedel–Crafts
anisole acylation with acetic anhydride. Substrate-Ln3+ interac-
tions are investigated using long-T1e and dipolar shift paramag-
netic probes. Factors affecting reaction rates such as lanthanide
ionic radius, arene substituents, kinetic isotope effects, and the
role of LiClO4 cocatalyst are investigated. Finally, these
mechanistic observations are compared and contrasted with what
is known from classical Friedel–Crafts acylation.
FIGURE 1. Representative 500 MHz 1H NMR spectrum of the reaction
for kinetic studies of the Yb(OTf)3-catalyzed acylation of anisole with
acetic anhydride with 1,1,2,2-tetrachloroethane as the internal standard
in nitromethane-d3 at 50 °C.
anisole:anhydride ratio held to greater than 1:10 (pseudofirst-
order in anhydride). The catalyst:anisole ratio was then held at
1:4, and the anisole concentration monitored until ∼90%
consumption. The disappearance of the 1H NMR anisole OCH3
resonance (δ ∼ 3.84 ppm) and appearance of the p-methoxy-
acetophenone OCH3 resonance (δ ∼ 4.1 ppm) were normalized
to an internal (1,1,2,2-tetrachloroethane) standard (δ ∼ 6.2 ppm;
Results
This section begins with a discussion of the reaction kinetics
and rate law for Ln(OTf)3-mediated anisole acylation with acetic
anhydride. This includes analysis of molecularity, activation
parameters, and the CH/CD kinetic isotope effect. We then
explore solvent effects, the effects of Ln(OTf)3, substrate
variation and LiClO4 cocatalyst effects, and Ln3+-substrate
interactions using long-T1e 4f7 and dipolar shift paramagnetic
probes.
General Reaction Procedure. Preliminary studies of
Ln(OTf)3-mediated anisole acylations by acetic anhydride were
carried out in the presence of air and moisture to ensure
agreement with the literature results.11b The stoichiometry used
in these initial reactions was Ln(OTf)3:anisole:acetic anhydride
) 1:5:10, respectively, at 50 °C for 18 h in nitromethane. Upon
completion, the catalyst was recovered from the aqueous phase
on workup. The only product detected is the para regioisomer,
4-methoxyacetophenone, and it was characterized by 1H NMR
spectroscopy.
1
Figure 1). In situ H NMR analysis of all Yb(OTf)3-mediated
acylations indicates that, within the detection limits, only
substrate and product are present in detectable quantities at all
stages of conversion. Monitoring the reaction for at least 3
turnovers gives initial rate constants for each concentration. For
example, a van’t Hoff plot of ln kobs versus the ln[anisole]
reveals a linear relationship consistent with a first-order
dependence (0.936 ( 0.036) in anisole concentration (Figure
2a). The same trend is observed for the dependence of kobs on
the concentration of acetic anhydride and Yb(OTf)3 (Figure 2
b,c). Combining all of these results, the overall experimental
kinetic order for Yb(OTf)3-mediated acylation of anisole with
acetic anhydride is third-order, first-kinetic order in each of the
reactants and the catalyst: ν ∼ k3[Ln3+]1[Ac2O]1[anisole]1.
Activation Parameters. To obtain activation parameters, the
temperature dependence of the acylation rates was studied,
within the practical temperature limits imposed by the system.
The rate of Yb(OTf)3-catalyzed acylation was monitored by 1H
NMR spectroscopy over a 30–80 °C temperature range, and
results are summarized in Table 1. Standard Eyring and
Arrhenius kinetic analyses20 (Figure 3) afford the following
activation parameters: ∆H‡ ) 12.9 (4) kcal ·mol-1, ∆S‡ ) -44.8
Kinetic Studies of Anisole Acylation. Most kinetic data
acquired in this study were obtained using Yb(OTf)3 as the
precatalyst. The rationale for this choice is that Yb3+ exhibits
the largest turnover frequencies among the lanthanide triflates
3+
examined, and the relatively short Yb
T allows convenient
1e
and informative in situ NMR monitoring of the reaction without
the excessive line-broadening induced by some other lanthanides
(e.g., Figure 1).19a The former also allows greater efficiency in
kinetic data acquisition and greater accuracy in integration. To
minimize potential artifacts, the solvent and substrates used were
rigorously purified by distillation under reduced pressure, dried
over an appropriate drying agent, and stored under argon (see
Experimental Section for more details). For each kinetic data
set, a standard solution was prepared. All reaction components
were transferred to J-Young NMR tubes under an argon
atmosphere on a high-vacuum line to minimize exposure to air
and moisture. Quantitative kinetic studies were carried out using
in situ 1H NMR spectroscopy at 50 °C, by monitoring reactant
consumption as a function of time. For example, catalyst and
anisole concentrations were initially held constant, with the
(1.3) e.u., and Ea ) 13.1 (4) kcal ·mol-1
.
Kinetic Isotope Effects. Three kinetic runs were performed
with normal anisole and three runs, under identical conditions
with [4-2H]anisole (see Experimental section). Analysis of
kinetic data (Table 2) obtained in these experiments indicates a
substantial primary KIE, kH/kD ) 2.6 ( 0.15 (see more in
Discussion section below).
Solvent Effects. The effect of reaction medium on the
catalytic activity of Yb(OTf)3-mediated anisole acylation with
acetic anhydride was also explored (Table 3). When nitro-
methane or acetic anhydride are used as a solvent, the reaction
proceeds smoothly. However, in water, no product formation
is observed over 48 h at 50 °C. The product percent yield in
neat acetic anhydride is lower, versus that in nitromethane,
owing possibly to competition of acetic anhydride with anisole
(19) Ln3+-induced line-broadening: (a) Bertini, I.; Luchinat, C. Coordination
Chemistry ReViews: NMR of Paramagnetic Substances; Lever, A. B. P., Ed.;
Elsevier: Netherlands, 1996; Vol. 150, Chapters 1–9. (b) Nolan, S.; Marks, T. J.
J. Am. Chem. Soc. 1989, 111, 8538–8540. (c) Glavincevski, B.; Brownstein, S.
J. Org. Chem. 1982, 47, 1005–1007.
(20) More, P. M.; Spencer, M.
Organometallics 1994, 13, 1646–1655.D.;
Wilson, S. R.; Girolami, G. S.
4006 J. Org. Chem. Vol. 73, No. 11, 2008

Lanthanide Triflate-Catalyzed Arene Acylation
TABLE 1. Variable-Temperature Kinetic Experiments Rate Constantsa
temp (K)
k1,obs (min-1
)
k1,obs (s-1
)
k3,obs
ln(k3/T)
ln(k3)
1/T
303.15 ( 1
313.15 ( 1
323.15 ( 1
333.15 ( 1
353.15 ( 1
0.0082
0.0143
0.0254
0.0531
0.1575
1.3667 × 10-4
2.3833 × 10-4
4.2333 × 10-4
8.8501 × 10-4
2.6250 × 10-3
5.5449 × 10-7
9.6695 × 10-7
1.7175 × 10-6
3.5907 × 10-6
1.0650 × 10-5
-20.1194
-19.5958
-19.0528
-18.3458
-17.3168
-14.4052
-13.8491
-13.2746
-12.5371
-11.4499
3.2987 × 10-3
3.1933 × 10-3
3.0945 × 10-3
3.0016 × 10-3
2.8317 × 10-3
a Reaction conditions: [acetic anhydride] ) 2.2 M, [anisole] ) 0.138 M, [Yb(OTf)3] ) 0.0294 M, [1,1,2,2-tetrachloroethane] ) 0.3245 M in
nitromethane-d3.
FIGURE 2. (a) van’t Hoff plot of ln[kobs] vs ln[anisole] concentration
for catalytic production of 4-methoxyacetophenone. Reaction conditions:
[acetic anhydride] ) 0.14 M, [Yb(OTf)3] ) 0.069 M, [1,1,2,2-
tetrachloroethane] ) 0.26 M in nitromethane-d3 at 50 °C, R2 ) 0.9969.
(b) van’t Hoff plot of ln[kobs] vs ln[acetic anhydride] for catalytic
production of 4-methoxyacetophenone. Reaction conditions: [anisole]
) 0.14 M, [Yb(OTf)3] ) 0.069 M, [1,1,2,2-tetrachloroethane] ) 0.25
M in nitromethane-d3 at 50 °C, R2 ) 0.9883. (c) van’t Hoff plot of
ln[kobs] vs ln[Yb(OTf)3] concentration for catalytic production of
4-methoxyacetophenone. Reaction conditions: [acetic anhydride] ) 2.5
M, [anisole] ) 0.24 M, [1,1,2,2-tetrachloroethane] ) 0.26 M in
nitromethane-d3 at 50 °C, R2 ) 0.9956.
FIGURE 3. (a) Plot of acylation rate constants as a function of
temperature. (b) Eyring plot for Yb(OTf)3-mediated anisole acylation
with acetic anhydride in nitromethane-d3. (c) Arrhenius plot for
Yb(OTf)3-mediated anisole acylation with acetic anhydride in ni-
tromethane-d3. The lines are least-squares fits to the data points.
anaerobic conditions.11b The formation of the ortho product
isomer is not observed in the 1H NMR spectrum, indicative of
exclusive para-selectivity. Moving across the lanthanide series,
for fixed reaction time, turnover frequency increases with
increasing Lewis acidity of the lanthanide ion7a–c (Tables 4 and
5, Figure 4; see more in Discussion section below).
Substrate Substituent and Cocatalyst Effects. Arene sub-
stituent effects on acylation rates and regioselection were
investigated under the present anhydrous/anaerobic Yb(OTf)3-
mediated reaction conditions (eq 2). Several substituted benzenes
were subjected to Yb(OTf)3-catalyzed acylation with acetic
for coordination at the metal center (see more in Discussion
section below).
Effect of Ln3+ Ionic Radius on Reaction Efficiency and
Turnover Frequencies. A catalytic rate dependence on metal
ion size was previously observed for classical Friedel–Crafts
acylation processes, with AlCl3 being the smallest and most
effective catalyst. The effect of lanthanide variation on Ln(OTf)3-
mediated anisole acylation by acetic anhydride on reaction
efficiency was probed in nitromethane at 50 °C under rigorously
anhydrous, anaerobic conditions. In the present study, the most
active catalyst is found to be Yb(OTf)3 (Table 4), with efficiency
scaling inversely with Ln3+ ionic radius, while the regioselec-
tivity remains constant. The results obtained are similar to those
reported in the literature conducted under less rigorous anhydrous/
anhydride (Table 6). Reactions occurring with non-negligible
rates produced a single acylated product, and the formation of
1
other isomers was not observed by H NMR analysis. It is
known that this catalytic system generally turns over only with
J. Org. Chem. Vol. 73, No. 11, 2008 4007

Dzudza and Marks
TABLE 2. Rate Constants Obtained from the Least-Squares Fits
to Experimental Data Used in Determination of the Kinetic Isotope
trial
kH (s-1
)
σa kH × 10-3
((0.021)
((0.037)
((0.014)
((0.031)
kD (s-1
)
σakD × 10-4
((0.023)
((0.161)
((0.139)
((0.151)
1
2
3
1.228 × 10-3
1.286 × 10-3
1.235 × 10-3
4.683 × 10-4
4.867 × 10-4
4.567 × 10-4
4.706 × 10-4
mean 1.249 × 10-3
a Standard deviation associated with each measurement and the
FIGURE 4. Effect of Ln3+ ionic radius on kobs for anisole acylation
with acetic anhydride in nitromethane-d3 at 50 °C.
average.
TABLE 3. Reaction Medium Effects on Yb(OTf)3-Mediated
Anisole Acylation with Acetic Anhydride
TABLE 6. Arene Substituent Effects on Yb(OTf)3-Mediated Arene
Acylation with Acetic Anhydride with and without the Presence of
LiClO4 as a Cocatalyst
product
εd
reaction conditions
reaction time
% yielde
entry
R1
R2
catalyst
cocatalyst
kobs (s-1
)
solvent
1a
2b
3a
4b
5a
6b
7a
8b
OCH3
OCH3
SCH3
SCH3
N(CH3)2
N(CH3)2
CH3
H
H
H
H
H
H
CH3
CH3
Yb(OTf)3
Yb(OTf)3
Yb(OTf)3
Yb(OTf)3
Yb(OTf)3
Yb(OTf)3
Yb(OTf)3
Yb(OTf)3
-
1.57 × 10-3
3.80 × 10-3
1.03 × 10-4
CD3NO2
H2Oa
36.4
80.4
21.0
50 °C
50 °C
50 °C
24 h
48 h
24 h
98
b
LiClO4
-
LiClO4
-
LiClO4
-
LiClO4
Ac2Oc
57
1.84 × 10-4
a Entry 2 uses neat water. b Product formation was not observed.
c Entry 3 uses neat acetic anhydride. d Solvent dielectric constant at 25
°C. e Product percent yield. Reaction conditions: [acetic anhydride] )
0.8 M, [anisole] ) 0.4 M, [Yb(OTf)3] ) 0.08 M at 50 °C, 24 h.
c
d
c
CH3
1.39 × 10-5
a Reaction conditions: [acetic anhydride] ) 1.35 M, [Yb(OTf)3] ) 6.9
TABLE 4. Effect of Lanthanide(III) Ionic Radius on Anisole
Acylation with Acetic Anhydride
×
10-2 M, and [1,1,2,2-tetrachloroethane]
) 0.222 M in nitro-
methane-d3, 50 °C. b Reaction conditions: [acetic anhydride] ) 1.35 M,
[Yb(OTf)3] ) 6.9 × 10-2 M, [LiClO4] ) 0.41 M, and [1,1,2,2-
tetrachloroethane] ) 0.222 M in nitromethane-d3, 50 °C. c No product
formation observed by 1H NMR. d Desired product formation was not
observed. Instead, a gelatinous sludge formed at the beginning of the
ionic radius
3+, Å
product
yield (%)a
Nt
(%) catalyst
recovereda
Ln(OTf)3
M
(h-1
)
La
1.172
1.098
1.087
1.008
37
75
82
95
1.73
3.45
3.74
4.99
93
90
95
93
Sm
Eu
Yb
1
reaction with an unresolved H NMR spectrum.
a Catalyst recovered from the aqueous layer in product isolation
workup and dried on a Schlenk line under vacuum at 100 °C for 24 h.
Reaction conditions: [Ln(OTf)3] ) 0.070 M, [anisole] ) 0.38 M, [acetic
anhydride] ) 0.72 M, [1,1,2,2-tetrachloroethane] ) 0.49 M in nitro-
methane-d3 at 50 °C, 18 h reaction time.
sion section below). The effects of LiClO4 on the acylation rates
of arenes with acetic anhydride using the Yb(OTf)3-LiClO4
catalytic system are summarized in Table 6.
Coordinative Probing. While kinetic studies provide the
overall reaction order and suggest the turnover-limiting step, NMR
line-broadening experiments give qualitative insight into which
substrate and product species enter the Ln3+ coordination sphere.19
The first set of experiments consisted of adding small increments
of paramagnetic Gd(OTf)3 to a 6.36 M solution of acetic anhydride
in nitromethane-d3 (Figure 5). Resonance line width at half-height
ratios obtained from individual peak measurements indicate
significantly larger line-broadening of the acetic anhydride reso-
nance versus that of the nitromethane peak, suggesting preferential
Ln3+-acetic anhydride complexation (for acetic anhydride/ni-
tromethane line width at half-height ratios, see Figure 5). In the
second set of experiments, small increments of paramagnetic
Gd(OTf)3 are added to a solution containing both anisole and acetic
TABLE 5. Correlation between Ln(OTf)3-Mediated Anisole
Acylation Rate Constants and Ln3+ Ionic Radius
ionic radius
a
catalyst
M
3+, Å
kobs (s-1
)
La(OTf)3
Eu(OTf)3
Yb(OTf)3
Lu(OTf)3
1.172
1.087
1.008
1.001
2.20 × 10-4
3.58 × 10-4
4.92 × 10-4
8.98 × 10-4
a Reaction conditions: [acetic anhydride] ) 1.35 M, [anisole] ) 0.14
M, [Ln(OTf)3] ) 2.94 × 10-2 M, and [1,1,2,2,-tetrachloroethane] )
0.222 M in nitromethane-d3, 50 °C.
1
anhydride, and then differences in paramagnetically induced H
strongly electron-donating arene substituents for which the rate
constants were obtained in the present study. No rate data were
obtained in the acylation of N,N-dimethylaniline. The ordering
of substituent effects on the rate of Yb(OTf)3-promoted acylation
is OCH3 > SCH3 >> (CH3)2 ≈ N(CH3)2. The results are
summarized in Table 6.
It has been reported that the presence of a LiClO4 cocatalyst
further enhances acylation rates of even less reactive aromatics,
and the role of LiClO4 appears to be limited to enhancing
Ln(OTf)3 efficiency since LiClO4 alone does not promote the
reaction.14 Therefore, potential cocatalyst effects were investi-
gated as well. Indeed, the presence of 6.0 equiv of LiClO4
promotes arene acylation with less electron-donating groups such
as m-xylene. However, acylation of N,N-dimethylaniline was
immeasurably sluggish even under these conditions (see Discus-
NMR line-broadening observed (Figure 6; see Discussion section
below on Ln3+-Substrate Interactions and the Nature of the Active
Species). Resonance line width at half-height ratios derived from
individual measurements of acetic anhydride and arene peaks
indicate some preferential line-broadening of the acetic anhydride
versus arene aromatic resonances (for ratios, see Figure 6).
In a third set of coordinative probe experiments, small
increments of Gd(OTf)3 were first loaded into J-Young NMR
tubes containing fixed amounts of LiClO4 to which was then
added a standard solution containing acetic anhydride, 1,1,2,2-
tetrachloroethane, and internal NMR standard, in nitromethane-
d3, and then differences in paramagnetically induced line-
broadening in the 1H NMR spectrum were observed. The
NMR spectra are indistinguishable from those in the other
4008 J. Org. Chem. Vol. 73, No. 11, 2008

Lanthanide Triflate-Catalyzed Arene Acylation
1
FIGURE 5. Overlaid 500 MHz H NMR spectra of solutions having different Gd(OTf)3 concentrations demonstrating the effect of increasing
[Gd(OTf)3] on the line width at half-height ratio of acetic anhydride: nitromethane-d3. All spectra were recorded at t ) 0 min. Reaction solution:
1
[acetic anhydride] ) 6.36 M, [1,1,2,2-tetrachloroethane] ) 0.05 M in nitromethane-d3 at room temperature. 500 MHz H NMR CD3NO2: δ 2.219
ppm (a, 3H, s, (CH3CO)2O), 4.33 ppm (b, CD3NO2), 6.269 ppm (c, 2H, s, Cl2CHCHCl2). * (CH3CO)2O/CD3NO2 line width at half-height ratio.
1
FIGURE 6. Overlaid 500 MHz H NMR spectra of the reaction solution having different Gd(OTf)3 concentrations demonstrating the effect of
increasing [Gd(OTf)3] on the line width half-height ratio of acetic anhydride versus anisole. Standard solution: [acetic anhydride] ) 0.917 M,
[anisole] ) 0.131 M, [1,1,2,2-tetrachloroethane] ) 0.263 M, in nitromethane-d3 at room temperature. 500 MHz 1H NMR CD3NO2 (g): δ 2.203 ppm
(a, 6H, s, (CH3CO)2O), 3.796 ppm (b, 3H, s, OCH3), 6.264 ppm (f, 2Η, s, Cl2CHCHCl2), 6.938 ppm (c, 3H, m, ArH), 7.303 ppm (d, 2H, dd, ArH).
* Calculated (CH3CO)2O/ArH line width at half-height ratio.
Gd(OTf)3-acetic anhydride coordination probing experiments
shown in Figure 5 (see Discussion section below).
of reaction (Figure 7). Observation of significant displacements
in the acetic anhydride methyl resonance is precluded by the
large stoichiometric excess of this reagent (see Figure 7).
Yb3+-Arene Interaction. To complement the NMR line-
broadening studies, additional arrayed experiments were carried
out in which in situ monitoring of the course of Yb(OTf)3-
mediated acylation with acetic anhydride probed Yb3+-arene
interactions via chemical shift changes of the aromatic 1H
resonances (Figure 7) (for details, see Experimental section).
Yb3+ was chosen since it induces reasonably small line-
broadening, allowing more accurate peak integration. In the
Yb(OTf)3-mediated acylation of anisole and thioanisole with
acetic anhydride, substantial changes in the aromatic proton
resonance chemical shift dispersion are observed on initiation
Discussion
Molecularity and Rate Law. Early studies of classical FC
acylation processes afforded no examples of simple kinetics.21
The available kinetic data provide ambiguous/conflicting con-
clusions about the identity of the active electrophile. Rate
equations of the form ν ) k[RCOCl·AlCl3]1 [ArH]1 have been
reported for reaction of benzene and toluene with both acetyl
and benzoyl chloride.21a–e These results are consistent with either
J. Org. Chem. Vol. 73, No. 11, 2008 4009

Dzudza and Marks
1
1
FIGURE 7. (A) 500 MHz H NMR spectrum of the thioanisole aromatic region in nitromethane-d3. (B) 500 MHz H NMR spectrum of the
thioanisole aromatic region in the reaction mixture at t ) 0 min of the acylation reaction. (C) 500 MHz 1H NMR spectrum of thioanisole aromatic
region at t ) 5 min of the acylation reaction. Reaction mixture: [acetic anhydride] ) 1.45 M, [thioanisole] ) 0.138 M, [Yb(OTf)3] ) 0.069 M,
[1,1,2,2-tetrachloroethane] ) 0.3245 M in nitromethane-d3 at 50 °C.
TABLE 7. Summary of Activation Parameters Obtained in
of two catalytic reaction pathways: via a free acylium ion pair
(eqs 3 and 4) or by nucleophilic attack (eqs 5 and 6), with
Various Acylation Reactions
∆H‡
∆S‡
(e.u.)
Ea
preference for a particular pathway being solvent- and
temperature-dependent.3a,b,4 For reactions that proceed via an
acylium ion pair pathway, the overall reaction order is second.
The last step, decomposition of the σ-complex is thought to be
rate-determining (see Discussion section below on Temperature
and Kinetic Isotope Effects), which corresponds to observed
rate law (eqs 3 and 4).3,21e The acylium ion pair mechanism is
preferred by activated aromatics, sterically hindered acylating
agents, in polar solvents, and at higher temperatures.3a,b
catalyst solvent (kcal·mol-1
)
(kcal·mol-1
)
ref
a
Yb(OTf)3 CH3NO2
12.9 (4)
12.7
13.3
12.8
10.3
- 44.8 (1.3)
- 27.6
13.1 (4)
13.3
13.9
13.3
10.9
this work
22a
22b
21e
21e, f
b
AlCl3
AlCl3
AlCl3
EDCb
EDC
BCc
c
c
d
- 28.1
- 26.7
- 48.0
SnCl4
BC
a Yb(OTf)3-mediated anisole acylation with acetic anhydride in
nitromethane. b AlCl3-mediated benzene and toluene acylation with
acylchloride in EDC (ethylenedichloride). c AlCl3-mediated benzene and
toluene acylation with benzoylchloride (BC) neat and in EDC.
d SnCl4-mediated benzene and toluene neat benzoylation. Note: This work’s
activation parameters comparison was made to AlCl3-mediated reactions
that exhibit second-order kinetics: ν ) k2[AlCl3 ·RCOCl]1[ArH]1, first-order
with respect to the AlCl3-acylating agent complex and arene.
consistent with observed substituent effects (see Discussion
section below).
Temperature and Kinetic Isotope Effects. Variable-tem-
perature kinetic characterization of Yb(OTf)3-catalyzed anisole
acylation with acetic anhydride yields the following activation
parameters: ∆H‡ ) 12.9 (4) kcal ·mol-1, ∆S‡ ) -44.8 (1.3)
e.u., and Ea ) 13.1 (4) kcal ·mol-1. These values are comparable
to those found in classical AlCl3-catalyzed FC acylation
reactions that exhibit overall second-order kinetics of type ν )
k2[AlCl3 · RCOCl]1[ArH]1.21c,22 As shown in Table 7, AlCl3-
catalyzed benzene acetylation in ethylenedichloride proceeds
with ∆H‡ ) 12.7 kcal ·mol-1, ∆S‡ ) -27.6 e.u., and Ea ) 13.3
22a
kcal·mol-1
.
Similar parameters were reported for both
The present kinetic results for Yb(OTf)3-catalyzed anisole
acylation with acetic anhydride (see Results section above)
clearly indicate the overall empirical rate law of eq 7, first-
order kinetics in each of the reactants, and the catalyst.
(21) Classical FC acylation rate laws: (a) DeHaan, F. P.; Covey, W. D.;
Delker, G. L.; Baker, N. J.; Feigon, J. F.; Ono, D.; Miller, K. D.; Stelter, A. D.
J. Org. Chem. 1984, 49, 3959–3963. (b) Kirkman, S.; Wolfenden, R. J. Am.
Chem. Soc. 1978, 100, 5944–5955. (c) Kobayashi, S.; Olah, G. A. J. Am. Chem.
Soc. 1971, 93, 6964–6967. (d) Brown, H. C.; Marino, G. J. Am. Chem. Soc.
1959, 81, 5611–5615. (e) Brown, H. C.; Jensen, F. R. J. Am. Chem. Soc. 1958,
80, 2291–2296. (f) Brown, H. C.; Jensen, F. R. J. Am. Chem. Soc. 1958, 80,
3039–3040.
ν
k3[Ln3+]1[Ac2O]1[anisole]1
(7)
The significance of the first-order dependence of the rate upon
the aromatic hydrocarbon concentration for this reaction suggests
that the turnover-limiting step is not electrophile formation, but
the subsequent attack on the aromatic hydrocarbon, which is
(22) (a) Brown, H. C.; Marino, G. J. Am. Chem. Soc. 1959, 81, 3310–33114.
(b) Jensen, F. R.; Marino, G.; Brown, H. C. J. Am. Chem. Soc. 1959, 81, 3303–
3306.
4010 J. Org. Chem. Vol. 73, No. 11, 2008

Lanthanide Triflate-Catalyzed Arene Acylation
pected.3,4,21c,24–27 Analysis of the present kinetic data yields a
primary KIE, kH/kD ) 2.6 ( 0.15, suggesting that the turnover-
limiting step in the Yb(OTf)3-mediated anisole acylation with
acetic anhydride is C-H bond cleavage, in general accord with
previous KIE observations for other Lewis acid-mediated
acetylations.21c,27
The magnitude of the present KIE suggests C-H scission is
either partial or complete in the transition state.24,27 These results
are consistent with the overall rate law and the multistep
mechanism of eqs 8, 9, and 10. In classical FC AlCl3-catalyzed
benzene and toluene acetylations in polar nitro solvents,
comparable values are: kH/kD ) 2.45 and 2.68, respectively.21c,22,28
FIGURE 8. Electrophile (E+) interacts with an aromatic substrate
forming a weak reagent-substrate complex. As E+ approaches the
bonding distance, the aromatic HOMO π-orbital overlaps with an empty
electrophile orbital, forming a two-electron, three-center π-complex
(step i), which leads to σ-complex formation (step ii), followed by final
C-H cleavage.3,4,24
acetylation and benzoylation of benzene in either ethylene
dichloride22b or benzoyl chloride as the solvent.21e However,
SnCl4-catalyzed toluene benzoylation in benzoyl chloride
proceeds with a far larger negative activation entropy ∆S‡ )
-48 e.u. versus ∆S‡ ) -27 e.u. for toluene benzoylation using
AlCl3 as the catalyst.21e,f These discrepancies were explained
by differences in the strength of benzoyl halide complexation
to AlCl3 versus SnCl4.21e,f Thus, SnCl4 is not significantly
complexed to benzoyl chloride in solution,21e,f and the large
difference in ∆S‡ values reflects a loss in degrees of freedom
accompanying SnCl4-benzoyl chloride complex formation prior
to the turnover-limiting step.
Reasoning similar to that for SnCl4 can rationalize the large
negative ∆S‡ observed in Ln(OTf)3-mediated acylation, also
considering nitromethane solvation effects. Literature reports
indicate far greater Yb(OTf)3-mediated anisole acylation rates
in polar, moderately coordinating aprotic solvents such as
CH3NO2 and CH3CN, rather than in less polar, less coordinating
aprotic solvents, such as CH2Cl2.6f,11b For neutral substrates,
such as anisole and acetic anhydride, polar aprotic solvents
should accelerate the reaction, due to stabilization of transition
state charge separation.4,23 Therefore, the large negative ∆S‡ is
also indicative of a highly organized transition state, plausibly
reflecting some combination of loss of degrees of freedom in
Ln3+-acetic anhydride complexation, solvent reorganization in
stabilizing the charged transition state, and substrate reorganiza-
tion optimizing Ln3+-arene proximity.
Ln3 + Ac2O yk1z Ln(Ac2O)3+
(8)
k-1
Ln(Ac2O)3+ + anisole yk2z Ln(Ac2O)(anisole)3+ (9)
k-2
k3
Ln(Ac2O)(anisole)3+ 98 Ln3+ + product + AcOH (10)
Solvent Effects. Comparative kinetic studies of Yb(OTf)3-
mediated anisole acylation with acetic anhydride in different
solvents reveal that nitromethane is by far the most effective,
followed by acetic anhydride, with water being the least effective
(Table 3). While both nitromethane and water have high
dielectric constants and large dipole moments (Table 3), the
difference lies in the cation-coordinating strength.21 17O and 19
F
NMR studies show that in aqueous solution, Ln(OTf)3 OTf-
anions are displaced by water molecules, which reside in the
inner coordination sphere of the hydrated complex.28 Hydrogen-
bonding among the coordinated water molecules further impedes
their displacement. The energetic cost of disrupting the H-
bonding increases the activation energy of acylation, and
moreover, Mackenzie and Winter showed that for HClO4-
mediated FC acylation, water rapidly converts acetic anhydride
into acetic acid, thus depleting the acylating agent.29 Therefore,
in aqueous solution, formation of the hydrated Ln3+ complexes
as well as depletion of the acylating agent explain why negligible
turnover is observed.
Kinetic hydrogen isotope effects in electrophilic substitutions
and their interpretation regarding transition state position along
the reaction coordinate have been extensively reviewed.3,4,24 As
noted above in the discussion of classical FC acylation reaction
mechanisms, if the step i (Figure 8), electrophilic attack on the
arene, is turnover-limiting, the transition state would resemble
a π-complex (Figure 8), and a small secondary isotope effect
is expected since no C-H bonds are broken or formed.3,4,24
Examples of electrophilic aromatic substitutions proceeding with
In contrast to water, nitromethane is considered to be a
moderately cation-coordinating solvent as reflected by its Ln3+
coordination number, 2.7.23 As noted above, Gd(OTf)3-induced
line-broadening effects on acetic anhydride + nitromethane
resonances suggest preferential coordination of acetic anhydride
versus nitromethane. Therefore, in moderately cation-coordinat-
ing nitromethane, substrate competition for Ln3+ coordination
is not severe, thus allowing precatalyst activation and catalytic
turnover. Modest product yields in neat acetic anhydride versus
nitromethane possibly reflect competition of acetic anhydride
with anisole for coordination at the Ln3+ center (see Results,
Table 3).
secondary isotope effects include nitration where kH/kD
)
1.17.21c,24–27 However, if the proton elimination is turnover-
limiting, the transition state would evolve from a σ-complex
(step ii, Figure 8), and a primary isotope effect is ex-
(23) Reichardt, S. SolVents and SolVent Effects in Organic Chemistry; Wiley-
VCH: Weinheim, 1990.
(24) Melander, L.; Sanders, W. H. Reaction Rates of Isotopic Molecules;
John Wiley & Sons: New York, 1980.
Lewis Acidity Effects. In the cases of AlCl3, BF3, and TiCl4-
catalyzed acylations, the kinetic order in catalyst depends on
(25) Covalent AlCl3 complexes: (a) Bigi, F.; Casnati, G.; Sartori, G.; Predieri,
G. J. Chem. Soc., Perkin Trans. 2 1991, 1319–1321. (b) Chevrier, B.; Weiss, R.
Angew. Chem., Int. Ed. Engl. 1974, 13, 1–10. (c) Chevrier, B.; Le Carpentier,
J.-M.; Weiss, R. J. Am. Chem. Soc. 1972, 94, 5718–5723.
(28) (a) Campbell, R.; Maraschin, V.; Silber, H. B. J. Solid State Chem.
2003, 171, 225–229. (b) Zhang, S.; Wu, K.; Biewer, M. C.; Sherry, D. A. Inorg.
Chem. 2001, 40, 4284–4290. (c) van Loon, A. M.; van Bekkum, H.; Peters,
J. A. Inorg. Chem. 1999, 38, 3080–3084. (d) Merbach, A.; Helm, L.; Cossy, C.
Inorg. Chem. 1998, 27, 1973–1979. (e) Swift, T. J.; Connick, R. E. J. Chem.
Phys. 1962, 37, 307–320.
(26) Ionic AlCl3 complexes: (a) Xu, T.; Barich, D. H.; Torres, P. D.; Nicholas,
J. B.; Haw, J. F. J. Am. Chem. Soc. 1997, 119, 396–405. (b) Boer, F. P. J. Am.
Chem. Soc. 1968, 90, 6706–6709. (c) Olah, G. A.; Kuhn, S. J.; Toglyesi, W. S.;
Baker, E. B. J. Am. Chem. Soc. 1962, 84, 2733–2740. (d) Olah,G. A.; Germain,
A.; White, A. M. Carbonium Ions; Olah, G. A., Schleyer, P. v. R., Eds.; Wiley:
New York, 1976; Vol. 5, pp 2049-2133.
(29) (a) Mackenzie, H. A.; Winter, E. R. S. Trans. Faraday Soc. 1948, 44,
243–253. (b) Mackenzie, H. A.; Winter, E. R. S. Trans. Faraday Soc. 1948,
44, 171–181. (c) Mackenzie, H. A.; Winter, E. R. S. Trans. Faraday Soc. 1948,
44, 159–171.
(27) Olah, G. A.; Kuhn, S. J.; Flood, S. J.; Hardie, B. A. J. Am. Chem. Soc.
1964, 86, 2203–2209.
J. Org. Chem. Vol. 73, No. 11, 2008 4011

Dzudza and Marks
FIGURE 9. Complex formed in polarization of an acyl halide by a
Lewis acid such as TiCl4.
the Lewis acid strength. The order in weaker Lewis acids such
as SbCl5 or TiCl4 is second, consistent with a requirement for
the two molecules to polarize the acyl halide for acylation to
occur.21e,f,30 If a catalyst has greater chloride than oxygen affinity
(Figure 9), C-Cl bond weakening is thought to occur, followed
by ionization, thus increasing the acylium ion concentration and
favoring an acylium ion mechanism (eqs 3 and 4).3,21e,f,30 The
extent of C-Cl bond polarization depends on the Lewis acid
strength.3,30 As shown here, anisole acylations with acetic
anhydride mediated by lanthanide triflates (Ln ) La, Eu, Yb,
Lu) proceed at rates dependent on Ln3+ size, with larger ions
proceeding more sluggishly (see Results section, Table 5).
Moreover, we find that the kinetic order in [Ln3+] across the
series is invariant. Weaker Lewis acids such as TiCl4 pro-
duce much lower product percent yields than AlCl3.21e,f,30 The
same trend is observed in the present Ln(OTf)3-mediated anisole
acylation with acetic anhydride. Moving across the lanthanide
series, Nt and product yield increase with increasing Ln3+ Lewis
acidity (see Results section on Lanthanide Effects on Efficiency
and Turnover Frequency; Tables 4 and 5, Figure 4).
FIGURE 10. Plausible role of LiClO4 as a cocatalyst in Yb(OTf)3-
LiClO4 mediated arene acylation with acetic anhydride.
Yb(OTf)3-mediated anisole and thioanisole acylation of are 1.57
× 10-3 s-1 and 1.03 × 10-3 s-1, respectively (Table 6), while
methyl substituents generally have little effect on FC arene
acylation rates.34c,f In this study, m-xylene acylation proceeds
in the presence of a LiClO4 cocatalyst with a rate constant of
1.3 × 10-5 s-1 (Table 6; entry 7). N,N-Dimethylaniline may
undergo para acylation with AlCl3, but only traces of ketone
are obtained.34b The sluggish reactivity of N,N-dimethylaniline
is attributed to strongly deactivating coordination to the catalyst
of the basic amine (Table 6, entries 5 and 6).5f,34b
Cocatalyst Effects. The results detailed in Table 6 clearly
indicate that LiClO4 accelerates the acylation of all substrates
examined. A major factor responsible for the acylation rate
enhancement is likely the established Lewis acidity of Li+ in
organic solvents, dependent on the solvent basicity and coor-
dinating characteristics, thus facilitating the formation of polar
and/or ionic intermediates.14e,35 An example of solvent–Li+
Lewis acidity enhancement is found in enhanced rates of
aldehyde and cyclic ketone aldol condensation with silyl enol
ethers in 5 M LiClO4-CH3NO2 versus 5 M LiClO4-Et2O.36
Diethyl ether is a strongly polar cation-coordinating solvent,
and Li+ coordination to diethyl ether moderates the Li+ Lewis
acidity.35,36 Versus diethyl ether, nitromethane is moderately
cation-coordinating as reflected in its donor number (19.2 versus
2.7 for diethylether and nitromethane, respectively).23 Due to
its higher dielectric constant and dipole moment, nitromethane
enhances Li+ Lewis acidity, and thus carbonyl group activation
is more effective.34–36
Turnover frequency variations as a function of Ln3+ ionic
radius for lanthanide triflate-mediated anisole acylation are rather
modest (see Results section above on Lanthanide Effects on
Efficiency and Turnover Frequencies; Table 5, Figure 4). In
contrast, for other lanthanide-catalyzed reactions such as ami-
noalkene hydroamination/cyclization, increasing the Ln3+ ionic
radius increases the turnover frequency (Nt) markedly (100–1000-
fold times Nt increases are not unusual), presumably reflecting
significant steric demands in the transition state.31,32 Combining
all of the above observations suggests minimal ionic radius-
dependent steric impediments in the acylation transition state.
Arene Substituent Effects. Arene substituent trends observed
in the present Yb(OTf)3-mediated acylations parallel classical
FC acylations.3,4,33,34 The present kinetic ordering of aromatic
substituents is: OMe > SMe >> (Me)2 > NMe2. Arenes with
methoxy and methylthio are known to accelerate acylation and
have been successfully acylated using stoichiometric amounts
of many classical Lewis acids such as AlCl3, TiCl4, and
SnCl4.3,4,33,34 In the present catalytic system, the rates of
We find in this work that acylations in the presence of LiClO4
are accompanied by more rapid acetic acid formation, the
concentration of which always equals that of the benzophenone
formed. That the presence of LiClO4 may increase Ln3+ Lewis
acidity was probed by conducting the aforementioned
Gd(OTf)3-acetic anhydride line-broadening experiments in the
presence of added LiClO4 (see Figure 5). The NMR spectra
are indistinguishable from these in the other Gd(OTf)3-acetic
anhydride coordination probing experiments, suggesting the
effects on ground-state substrate coordination equilibria are
small. While the exact role of LiClO4 in Ln(OTf)3-mediated
FC acylation is not completely clear at this stage, it is
conceivable that LiClO4 + Yb(OTf)3 anion metathesis14e,g,25a
may facilitate highly reactive oxocarbenium perchlorate forma-
(30) Classical FC Lewis acidity effects: Brown, H. C.; Jensen, F. R. J. Am.
Chem. Soc. 1958, 80, 2296–2300.
(31) (a) Hong, S.; Kawaoka, A.; Marks, T. J. J. Am. Chem. Soc. 2003, 125,
15878–15892. (b) Ryu, J. S.; Li, G. Y.; Marks, T. J. J. Am. Chem. Soc. 2003,
125, 12584–12605. (c) Hong, S.; Tian, S.; Metz, M. V.; Marks, T. J. J. Am.
Chem. Soc. 2003, 125, 14768–14783.
(32) (a) Giardello, M. A.; Conticello, V. P.; Brard, L.; Gagné, M. R.; Stern,
C. L.; Marks, T. J. J. Am. Chem. Soc. 1994, 116, 10241–10254. (b) Gagné,
M. R.; Stern, C. L.; Marks, T. J. J. Am. Chem. Soc. 1992, 114, 275–294. (c)
Jeske, G.; Schock, L. E.; Mauermann, H.; Swepston, P. N.; Schumann, H.; Marks,
T. J. J. Am. Chem. Soc. 1985, 107, 8103–8110.
(33) (a) Acetic anhydride in classic FC acylation: Jeffery, E. A.; Satchell,
D. P. N. J. Chem. Soc. 1962, 1877–1894. (b) Satchell, D. P. N. J. Chem. Soc.
1960, 1752–1768. (c) Van Dyke, R. E.; Crawford, H. E. J. Am. Chem. Soc.
1951, 73, 2018–2021. (d) Van Dyke, R. E. J. Am. Chem. Soc. 1951, 73, 398–
402.
-
tion, CH3CO+ClO4 (Figure 10, I). A similar mechanism was
proposed for SbCl5-LiClO4- mediated FC acylation.25a
(34) Arene substituent effects in classical FC acylations: (a) Tashiro, M.;
Kobayashi, S.; Olah, G. A. J. Am. Chem. Soc. 1970, 92, 6369–6371. (b) Tuzun,
C.; Ogliaruso, M.; Becker, E. I. Org. Synth. 1961, 41, 1–4. (c) Marino, G.; Brown,
H. C. J. Am. Chem. Soc. 1959, 81, 5475–5481. (d) Cullinane, N. M.; Chard,
S. J.; Leyshon, D. M. J. Chem. Soc. 1952, 376–380. (e) Cutler, R. A.; Stenger,
R. J.; Suter, C. M. J. Am. Chem. Soc. 1952, 74, 5475–5481. (f) Shah, R. C.;
Chaubal, J. S. J. Chem. Soc. 1932, 650–652.
(35) (a) Springer, G.; Elam, C.; Edwards, A.; Bowe, C.; Boyles, D.; Bartmess,
J.; Chandler, M.; West, K.; Williams, J.; Green, J.; Pagni, R. M.; Kabalka, G. W.
J. Org. Chem. 1999, 64, 2202–2210. (b) Pagni, R. M.; Kabalka, W.; Bains, S.;
Plesco, M.; Wilson, J.; Bartmess, J. J. Org. Chem. 1993, 58, 3130–3133.
(36) (a) Sudha, R.; Sankararaman, S. J. Chem. Soc., Perkin Trans. 1 1999,
383–386. (b) RajanBabu, T. V. J. Org. Chem. 1984, 49, 2083–2088.
4012 J. Org. Chem. Vol. 73, No. 11, 2008

Lanthanide Triflate-Catalyzed Arene Acylation
FIGURE 11. Electrophilic species generated from acyl halides and
aluminum halides. II, donor-aceptor complex; III, acylium aluminate
ion pair.
FIGURE 13. Ab initio results on geometry-optimized complexes IV
(C6H6-AlCl3) and V (C6H6-BCl3).40 In the monomeric state, AlCl3
and BCl3 possess D3h geometries. However, in the complexed state,
charge transfer from the benzene HOMO to Al3+ or B3+ leads to
geometric distortion and pyramidalized complexes IV and V. This
entails partial loss of benzene aromaticity, compensated to an extent
by back-donation of Cl 3p electron density to the benzene HOMO. In
complex IV, the greater charge on Al leads to a closer approach to the
aromatic carbon (Al-C ≈ 2.35 Å), thus to more effective charge
transfer than in complex V, with a longer computed B-C distance
(≈ 3.22 Å).40
FIGURE 12. Perspective drawing of the methyloxocarbonium hexaflu-
oroantimonate salt, [CH3CO]+[SbF6]-. Bond lengths: Sb-F ) 1.894
Å (weighted average); C-C ) 1.385 (16) Å; C-O ) 1.108 (15) Å.26
Ln3+-Substrate Interactions and the Nature of the
Active Species. The activation process for acylation reactions
involving Lewis acids depends on the identity of the acylating
agent.3,4 In classical Fridel-Crafts acylations with acid chlo-
rides, it is known that their complexes with Lewis acids involve
electrophilic intermediates, which exist in solution as either
covalent (Figure 11, II)25 or ion paired forms (Figure 11, III),26
with the equilibrium position depending both on the nature of
the acylating agent, the aromatic, solvent, and temperature.3,30
Evidence for formation of 1:1 or 1:2 donor–acceptor covalent
complexes (Figure 11, II), such as acetyl chloride-aluminum
halide complexes in SO2 solution, is provided by NMR
spectroscopy.25 Moreover, a number of stable crystalline acylium
salts, such as methyloxocarbonium hexafluoroantimonate (Figure
12), have been characterized by X-ray diffraction, as well as
solid state NMR and IR spectroscopy.26
In classical FC acylations, kinetic and mechanistic investiga-
tions along with conductivity measurements suggest that in
reaction with aluminum halides, acetic anhydride behaves as
an acyl halide-aluminum halide complex which effects the
actual acylation (eqs 11 and 12).33 Acetic anhydride reacts with
2 equiv of AlCl3 to afford 2 equiv of acyl halide acylating
agent.33
FIGURE 14. (A) Perspective drawing of the crystal structure of the
arene complex Sm(C6Me6)(AlCl4)3, VI. (B) Perspective drawing of the
crystal structure of the aryloxide complex [Yb(O-2,6-Ph2C6H3)3], VII.
binding energies. However, Ln3+-ketone complexes have been
detected in solution by 13C and 17O NMR.39
In all proposed conventional FC acylation reaction mecha-
nisms, the role of the Lewis acid, such as AlCl3, is limited to
generation of the electrophile, which subsequently attacks the
arene to form either a π- or σ-complex (see Discussion section).
Ab initio calculations have probed an additional role of the
Lewis acid in C6H6-AlCl3 and C6H6-BCl3 model structures
and indicate the marked tendency of only one arene carbon atom
to become highly nucleophilic, thereby facilitating an attack by
an incipient electrophile (Figure 13).40
In conjuction with the aforementioned ab initio studies,
experimental work has shown that arenes form isolable π-com-
plexes with electrophiles such as Sm3+ and Ag+.17 Schwotzer
16
and Cotton demonstrated that neutral arenes form isolable
π-complexes with Sm3+ in a sterically congested mononuclear
η6-complex (Figure 14, VI).16e,f The Nd3+, Gd3+, and Yb3+
derivatives have been characterized and described as well.16a
Later, it was shown that less-substituted arenes, such as
m-xylene16c or even benzene, also yield isolable analogues of
the type Ln(C6H6)(AlCl4)3 (Ln ) La, Nd, Sm).16d A homoleptic
Yb(III) aryloxide complex [Yb(O-2,6-Ph2C6H3)3] was also
shown to exhibit an intramolecular Yb3+-π-arene interaction
(Figure 14, VII).16b The existence of thermally stable
Describing Ln3+-acetic anhydride interactions relies on
understanding Ln3+ coordination chemistry.7a–d,37c,d While
numerous lanthanide ꢀ-diketonates have been isolated and
characterized crystallographically,38 Ln3+-ꢀ-diketone com-
plexes have not been isolated, presumably due to their low
(37) (a) Gamp, E.; Shimomoto, R.; Edelstein, N.; McGarvey, B. R. Inorg.
Chem. 1987, 26, 2177–2782. (b) Bertini, I.; Luchinat, C. NMR of Paramagnetic
Molecules in Biological Systems; Benjamin: Menlo Park, 1986; Chapter 10. (c)
Fischer, R. D. Fundamental and Technological Aspects of Organo-f-Element
Chemistry; Marks, T. J., Fragalla, I. L., Eds.; Reidel Publishing: Dordrecht, 1985;
Chapter 8. (d) McGarvey, B. R. Organometallics of the f-Elements; Marks, T. J.,
Fischer, R. D., Eds.; Reidel Publishing: Dordrecht, 1979; Chapter 10. (e)
McGarvey, B. R. Can. J. Chem. 1984, 62, 1349–1355. (f) McGarvey, B. R.
J. Chem. Phys. 1976, 65, 955–961. (g) Kolb, J. R.; Marks, T. J. J. Am. Chem.
Soc. 1975, 97, 27–35. (j) Horrocks, W. D., Jr. NMR of Paramagnetic Molecules;
La Mar, G. N., Horrocks, W. D., Holm, R. H., Eds.; Academic Press: New York,
1973; Chapter 4.
(38) Ln3+-ꢀ-diketonates complexes: (a) Melby, L. R.; Rose, N. J.; Abramson,
E.; Caris, J. C. J. Am. Chem. Soc. 1964, 86, 5117–5125. (b) Halverson, F.; Brinen,
J. S.; Leto, J. R. J. Chem. Phys. 1964, 40, 2790–2792. (c) Whan, R. E.; Crosby,
G. A. J. Mol. Spectrosc. 1962, 3, 315–327. (d) Pope, G. W.; Steinbach, J. F.;
Wagner, W. F. J. Inorg. Nucl. Chem. 1961, 20, 304–313. (e) Biltz, W. Justus
Lieb. Ann. Chem. 1904, 331, 334–358.
(39) (a) Raber, D. J. J. Chem. Soc., Perkin Trans. 2 1986, 853-859. (b)
Okuyama, T.; Fueno, T. J. Am. Chem. Soc. 1980, 102, 6591–6594. (c) Chadwick,
D. J.; Williams, D. H. J. Chem. Soc., Perkin Trans. 2 1974, 1202-1206.
(40) AlCl3-and BF3-arene ab initio calculations: Tarakeshwar, P.; Lee, J-Y.;
Kim, K. S. J. Phys. Chem. A 1998, 102, 2253–2255.
J. Org. Chem. Vol. 73, No. 11, 2008 4013

Dzudza and Marks
(Gd3+, Yb3+) on both individual and competitive levels. The
former involved incremental Gd(OTf)3 addition to a ni-
tromethane solution of anisole at constant [anisole] (see Results
above on Coordination Probing). The results are inconclusive
due to extensive line-broadening of all the anisole peaks and
merging of the methoxy and nitromethane resonances, even with
Gd(OTf)3:anisole ) 0.1:1. However, competition studies reveal
more information (Figure 6). Incremental addition of Gd(OTf)3
to the reaction mixture of anisole and acetic anhydride in
nitromethane solution induces line-broadening of both arene
aromatic and acetic anhydride resonances, with slightly greater
broadening of the latter (for line width at half-height ratios, see
Figure 6), thus providing evidence for the latter type of
Ln3+-subtrate interaction proposed in structure IX, Figure 15.
For Yb(OTf)3, unlike Gd(OTf)3, strong spin–orbit coupling
results in a rapid electron spin–lattice relaxation, resulting in a
reasonable NMR sharp line widths and expectation of substantial
dipolar shifts.19,37 Analysis of Yb(OTf)3-mediated anisole
acylation with acetic anhydride in nitromethane-d3 solutions
reveals a slight downfield displacement of the methoxy peak
(∆ ∼ 0.06 ppm) vs the 1,1,2,2-tetrachloroethane internal stan-
dard, suggesting weak coordinative interactions between the
methoxy oxygen and Yb3+, supporting proposed structure VIII
in Figure 15. Furthermore, the aromatic proton chemical shift
dispersion in Yb(OTf)3-mediated thioanisole and anisole acy-
lation (Figure 7) increases markedly on Yb(OTf)3 addition,
providing evidence for Yb3+-arene complexation as well (e.g.,
IX in Figure 15).
FIGURE 15. (A) Anisole coordination to Ln3+ through the methoxy
oxygen only, VIII. (B) Anisole coordination to Ln3+ through the
methoxy oxygen and arene π-complexation, IX.
Ln3+-arene complexes provides precedent for Ln3+-arene
complexation in the present catalytic arene acylations. In
principle, anisole coordination to Ln3+ could occur through the
methoxy oxygen only (Figure 15, VIII) or accompanied by
additional arene π-cloud Ln3+ polarization (Figure 15, IX).
As discussed above, acetic anhydride-Ln3+ complexation
was probed by 1H NMR spectroscopy using the coordinatively
unsaturated, 4f7 paramagnetic NMR broadening probe, Gd-
(OTf)3. Gd3+ complexes, by virtue of their ground state (8S7/2
,
<SZ>J ≈ -31.5),37 are magnetically isotropic and cannot give
rise to dipolar (or pseudocontact) shifts but can give rise to
contact shifts.37 Because of the slow electronic spin–lattice
relaxation in Gd3+, NMR spectra of these complexes are usually
very broad, as also observed in this study (Figure 5). Incremental
addition of Gd(OTf)3 to acetic anhydride solutions in ni-
tromethane causes slight downfield displacement (versus internal
1,1,2,2,-tetrachloroethane) of the acetic anhydride methyl
resonance from δ 2.094 ppm to δ 2.243 ppm (see Results
Section above on Coordination Probing, Figure 5) and far greater
line-broadening of the acetic anhydride resonance than that of
the nitromethane peak. This behavior can be interpreted in terms
of equilibria as in eq 13, where ligand exchange is rapid on the
NMR time scale at room temperature. Also, the small observed
downfield paramagnetic shift is consistent with minor ligand
Mechanism of Catalytic Ln(OTf)3-Mediated Acylation. As
noted above, the present kinetic results for Ln(OTf)3-catalyzed
anisole acetylation with acetic anhydride indicate first-order rate
dependence on both substrate and catalyst concentrations. Earlier
kinetic and mechanistic studies of classical FC acylation
suggested that acetic anhydride behaves as an acyl halide and
that in polar solvents such as nitromethane, heterolysis favors
an acylium ion (CH3CO+) pathway, because of greater transition
state stabilization/solvation in polar solvents.21,26 However, the
acylium is a relatively weak electrophile,3 and acylation of even
slightly deactivated aromatics is sluggish.3,4
electron density donation (π-electrons from carbonyl) into (or
polarization by) empty or partially filled Gd3+ orbitals (4f or
6s, 6p, 5d in the present case) with exchange interactions
favoring transfer of spin density parallel to the spin already on
Gd(III), leaving unpaired spin density of the opposite sign
remaining on the ligand.6,25,26,32 The effect should attenuate as
the number of bonds between metal and the interacting nucleus
is increased, hence the small change in the methyl 1H chemical
shift.
The proposed Yb(OTf)3-catalyzed reaction mechanism (Fig-
ure 16) has several steps: (i) activation of acetic anhydride via
Ln(OTf)3 coordination, (ii) formation of an anisole complex.
The arene aromaticity is not yet lost in the substrate-reagent
complex. As the reagents move closer to bonding distance, (iii)
anhydride C-O bond heterolysis occurs, possibly facilitated by
arene π-complex formation, to generate a reactive electrophile
CH3CO+, along with (iv) aromatic deprotonation with ben-
zophenone and acetic acid generation. Kinetic isotope data
suggest the deprotonation step to be the turnover-limiting step.
That the results of this anaerobic/anhydrous study (e.g., Ln3+
effect on percent yield) parallel those reported by Kobayashi11b
suggests that under less rigorous reaction conditions water
ligands surrounding the Ln3+ center rapidly exchange with
substrate, allowing their coordination to the Ln3+ center and
the reaction to proceed. Furthermore, neutral Ln3+-η6-arene
complexes synthesized under aerobic/hydrous Friedel–Crafts
reaction conditions are isolable.16 Thus for all aforementioned
reasons, and under less anaerobic/anhydrous reaction conditions,
it is likely that the proposed catalytic pathway in Figure 16 will
most likely be unchanged.
Both structures proposed in eq 13 depict acetic anhydride
binding to Ln3+. The NMR data suggest that, in nitromethane,
formation of a Lewis acid–base complex between acetic
anhydride and the Ln3+ center occurs. Note that while lanthanide
triflates are poorly soluble in nitromethane, addition of acetic
anhydride immediately induces dissolution. These observations
along with the aforementioned NMR shifting/line-broadening
results support the suggestion that initial activation of acetic
anhydride by Ln(OTf)3 precatalyst leads to the formation of
the active catalyst (Figure 16).
The viability and the type of Ln3+-arene complexation
suggested in Figure 15 was also investigated spectroscopically
utilizing coordinatively unsaturated paramagnetic NMR probes
4014 J. Org. Chem. Vol. 73, No. 11, 2008

Lanthanide Triflate-Catalyzed Arene Acylation
FIGURE 16. Mechanistic scenario for Ln(OTf)3-mediated acylation of anisole with acetic anhydride in nitromethane.
Conclusions
catalysts is limited to polar aprotic, moderately coordinating
solvents such as nitromethane. Given these differences, acyla-
tions using Ln(OTf)3 offer many attractions over classical Lewis
acids: the acylations are catalytic, and the air- and moisture-
stable catalysts can be recycled many times.
Lanthanide triflates are clearly more efficient and cleaner
arene acylation catalysts than conventional Lewis acids. They
offer distinctive advantages such as the catalytic stoichiometry
of Ln(OTf)3 required to mediate FC acylations. Moreover, their
air- and moisture-stability and efficient catalyst recyclability add
an environmentally friendly aspect to Ln(OTf)3 catalysis.
The mechanistic results presented here, which focus on the
catalyst role, demonstrate that the present catalytic reaction
pathway has similarities to, and differences from, more classical
Lewis acid acylation. The first step in the classical FC acylation
pathway consists of initial Lewis acid activation of the acylating
agent. In classical FC acylation, acetic anhydride reacts with 2
equiv of Lewis acid to afford 2 equiv of acyl halide, which
through Lewis acid-coordination/abstraction generates the
reactive electrophile, CH3CO+, which effects the actual acyla-
tion. The sequence of events leading to reactive electrophile
generation using Ln(OTf)3 appears to be markedly different.
Acetic anhydride activation via Ln(OTf)3 coordination is
followed by Ln3+-arene π-complex formation, which possibly
facilitates anhydride C-O bond heterolysis to generate the
reactive electrophile. The last step, common in both pathways,
is aromatic C-H bond scission with concurrent benzophenone
generation. Both classical FC and the kinetic isotope data
presented herein suggest C-H scission to be turnover-limiting.
In addition to mechanistic pathway differences, other notable
differences/similarities for classical FC versus Ln(OTf)3-medi-
ated acylation include Lewis acidity effects on the overall
reaction order, arene substituent effects, and solvent effects. In
classical FC acylation, the overall reaction order is Lewis
acidity-dependent, being higher order with weaker Lewis acids.
However, in Ln(OTf)3-mediated acylation, the reaction order
for different Ln3+ ions is unchanged, suggesting minimal ionic
radius-dependent steric constraints in the transition state. While
classical Lewis acids effectively catalyze acylation of activated
and unactivated arenes, Ln(OTf)3 efficiency is limited to
activated ones. Also, classical Lewis acids function in both polar
and nonpolar solvents while the effectiveness of Ln(OTf)3
Experimental Section
Materials and Methods. All manipulations of reagents were
carried out with rigorous exclusion of oxygen and moisture in flame-
or oven-dried Schlenk-type glassware on a dual-manifold Schlenk-
line, or interfaced to a high-vacuum line (10-6 Torr), or in a
nitrogen-filled MBraun glovebox with a high-capacity recirculator
(<1 ppm O2). Argon (prepurified) was purified by passage through
a MnO oxygen-removal column and a Davison 4-Å molecular sieve
column. Nitromethane-d3, used for NMR-scale reactions, was
freshly predistilled under reduced pressure, dried over CaCl2, and
stored over activated Davison 4-Å molecular sieves in a vacuum-
tight storage flask under an Ar atmosphere. All aromatic substrates,
unless stated otherwise, were purchased from a chemical supplier,
distilled under reduced pressure, and dried as appropriate prior to
use. Acetic anhydride was predistilled and dried over BaO in a
vacuum-tight storage flask under an Ar atmosphere prior to use.
Physical and Analytical Measurments. Solution NMR spectra
were recorded either on a Varian Inova-500 (FT, 500 MHz, 1H) or
Varian Inova-400 (FT, 400 MHz, 1H) NMR spectrometer equipped
1
with a variable-temperature unit. All H NMR chemical shifts are
reported relative to 1,1,2,2-tetrachloroethane, an internal reference
standard, unless stated otherwise. Thermogravimetric (TGA) and
differential thermal analysis (DTA) data were collected on a TA
Instruments model SDT 2960 at atmospheric pressure, with an N2
flow rate of 100 mL/min and a temperature ramp of 5 °C/min. X-ray
diffraction (XRD) data were acquired with a computer-interfaced
Rigaku DMAX-A powder diffractometer using Ni-filtered Cu KR
radiation.
Preparation of Ln(OTf)3 Complexes. All lanthanide triflate salts
(Ln ) La, Sm, Eu, Gd, Yb, Lu) were prepared by adding an excess
of respective lanthanide(III) oxide (99.9% purity) to an aqueous
solution of trifluoromethanesulfonic acid (50% v/v) and heating at
boiling for approximately 1 h.41a The mixture was then filtered to
remove the unreacted oxide, and the water was then removed from
the filtrate under reduced pressure. The resulting hydrate was further
J. Org. Chem. Vol. 73, No. 11, 2008 4015

Dzudza and Marks
dried by heating under a vacuum at 180-200 °C for 48 h. The
resulting anhydrous Ln(OTf)3 complexes were stored in a glovebox
under a nitrogen atmosphere. Formation of the desired products
was confirmed by TGA, DTA, and XRD.41
Preparation of Standard Solutions for Kinetic Studies. For
each data set, a separate standard solution was prepared under inert
atmosphere in a three-neck round-bottom flask interfaced to a high-
vacuum line. The solution consisted of nitromethane-d3, an internal
NMR standard, and either the aromatic substrate or acylating agent,
whichever was to be held in pseudofirst-order excess, depending
on the particular kinetic experiment.42
Typical NMR-Scale Catalytic Reaction. In the glovebox, a
J-Young NMR tube equipped with a Teflon valve was loaded with
a precatalyst Ln(OTf)3 (Ln ) La, Eu, Yb, Lu), except for the studies
in which LiClO4 cocatalyst effects were explored, in which the
cocatalyst loading with respect to the catalyst was 6-fold. On the
high-vacuum line, the tube was evacuated (10-6 Torr), then frozen
at -78 °C, followed by addition of 700 µL of the standard solu-
tion and the second substrate via gastight syringe to the NMR tube
under an argon flush. The tube was then evacuated and backfilled
with Ar while frozen at -78 °C and then sealed with the Teflon
valve. The frozen reaction mixture was maintained in a dry ice/
acetone bath until the beginning of the kinetic run.
Absolute Error in ∆H‡:
R2T2maxT2min
2
2
σT
T
∆L
∆T
(σ∆H*)2 )
1 + Tmin
+
{(
)
[
(
(
)
∆T2
2
2
∆L
∆T
σk
k
1 + Tmax
+ 2
(15)
)
]
(
) }
Absolute Error in ∆S‡:
R2 σT
∆L
∆T
σk
2
2
(σ∆S*)2 )
T2max 1 + Tmin
+
{(
)
[
(
)
∆T2
T
2
∆L
∆T
T2min 1 + Tmax
+
2 T2max + T2min (16)
(
)
(
)
]
(
)
}
k
Isotopic Labeling Studies. The substrate [4-2H]anisole was
synthesized by treat 4-bromoanisole with Mg turnings to form the
corresponding Grignard reagent, (p-methoxyphenyl)magnesium
bromide, which was quenched with D2O. The final product was
first extracted with ether, then fractionally distilled under reduced
pressure, and finally stored in a closed storage vessel under
nitrogen.43 Characterization by H NMR 500 MHz CD3NO2: δ
1
3.796 (s, 3H, OMe), 6.935 (d, J ) 7.9 Hz, 2H), 7.304 (d, J ) 7.9
Hz, 2H).
Kinetic Studies of Yb(OTf)3-Mediated Acylation. In a typical
NMR experiment, an NMR sample was prepared as described above
(see Typical NMR-Scale Catalytic Reaction section) but maintained
at -78 °C until kinetic measurements were begun. The sample tube
was then inserted into the probe of the NMR spectrometer, which
had been previously set to the appropriate reaction temperature (T
( 0.2 °C, calibrated with the ethylene glycol temperature standard).
A long pulse delay was used during data acquisition to avoid
saturation. All data for a particular set of kinetic runs were collected
using a single preacquisition delay, which was adjusted to fit the
overall reaction times. The kinetics were usually monitored from
intensity changes in the substrate alkoxy or acetic anhydride methyl
resonances over three or more half-lives.
The substrate concentration C as a function of time was
calculated with respect to the known concentration of a 1,1,2,2-
tetrachloroethane standard in the solution. The turnover frequency
(h-1; eq 14) was calculated from the least-squares-determined ×
intercept (-C0/m, min) where C0 is the initial substrate concentration
and E is the catalyst to substrate ratio.31,42
Paramagnetic Line-Broadening Experiments on Individual
Substrates. In the glovebox, a J-Young NMR tube equipped with
a Teflon valve was loaded with various quantities of the paramag-
netic (4f7) Gd(OTf)3 precatalyst. On the high-vacuum line, each
tube was evacuated to 10-6 Torr, then frozen at -78 °C. Under an
argon flush of 600 µL of the nitromethane-d3 standard solution,
consisting of 6.36 M acetic anhydride or 6.36 M acetic anhydride
with 0.41 M LiClO4 or 6.36 M anisole, 0.05 M 1,1,2,2-tetrachlo-
roethane was added via gastight syringe into each NMR tube, which
was then sealed with the Teflon valve. The NMR tubes were then
evacuated and backfilled with argon. 1H NMR spectra were recorded
at room temperature, and resonance line width at half-height ratios
were obtained from individual peak measurements. Gd3+-induced
line broadening of the anisole spectrum causes partial merging of
the methoxy and nitromethane resonances as well as of the two
aromatic resonances.
Paramagnetic Line-Broadening Experiments on Reaction
Mixtures. In the glovebox, J-Young NMR tubes equipped with a
Teflon valve were loaded with varying amounts of the paramagnetic
(4f7) Gd(OTf)3 precatalyst. On the high-vacuum line, each tube was
evacuated to 10-6 Torr, then frozen at -78 °C. Under an argon
flush, of 600 µL of the nitromethane-d3 standard solution, consist-
ing of 0.917 M acetic anhydride and 0.132 M anisole, 0.263 M
1,1,2,2,tetrachloroethane was added via gastight syringe into each
NMR tube, which were then sealed with the Teflon valve. The NMR
Nt(h-1) ) (60 min h-1 (-C /m))E
(14)
⁄
0
Activation Parameters and Associated Errors. Eyring and
Arrhenius kinetic analyses20 of data obtained in variable-temperature
experiments were used to extract ∆H‡, ∆S‡, and Ea parameters.
The errors in these activation parameters were computed from the
published error propagation formulas (eqs 15 and 16)20a which were
derived from the Eyring equation20b using the OriginPro 7.5
program.
1
tubes were then evacuated and backfilled with argon. H NMR
spectra were recorded at room temperature, and resonance line width
athalf-heightratioswereobtainedfromindividualpeakmeasurements.
(41) Ln(OTf)3 preparation and characterization: (a) Pascal, J. L.; Hamidi,
M. M. Polyhedron 1994, 11, 1787–1792. (b) Abbasi, A.; Lindqvist-Reis, P.;
Eriksson, L.; Sandstrom, D.; Lidin, S.; Persson, I.; Sandstrom, M. Chem.-Eur.
J. 2005, 11, 4065–4077. (c) Egashira, K.; Yoshimura, Y.; Kanno, H.; Suzuki,
Y. J. Therm. Anal. Calorim. 2003, 71, 501–508. (d) Zineddine, H.; Hnach, M.;
Hamidi, M. J. Fluorine Chem. 1998, 88, 139–141. (e) Nakayama, M.; Nakamura,
S.; Yanagihara, N. Polyhedron 1998, 17, 3625–3631.
Acknowledgment. We gratefully acknowledge financial
support by NSF (CHE-04157407). We also thank Dr. X. Yu
for helpful suggestions on the kinetic studies.
JO800158K
(42) (a) InVestigation of Rates and Mechanism of Reactions; Bernasconi, C.,
Ed.; 4th ed.; Interscience: New York, 1986. (b) BensonS. W. Thermochemical
Kinetics, 2nd ed.; Wiley & Sons: New York, 1986; pp 8-10.
(43) Abdul-Malik, N. F.; Jenkins, L. A.; Davis, F. A. J. Org. Chem. 1983,
48, 5128–5130.
4016 J. Org. Chem. Vol. 73, No. 11, 2008