Scandium(III) catalysis of transimination reactions. Independent and constitutionally coupled reversible processes

Article abstract of DOI:10.1021/ja042469q

Sc(OTf)₃ efficiently catalyzes the self-sufficient transimination reaction between various types of C=N bonds in organic solvents, with turnover frequencies up to 3600 h^-1 and rate accelerations up to 6 × 10⁵. The mechanism of the crossover reaction in mixtures of amines and imines is studied, comparing parallel individual reactions with coupled equilibria. The intrinsic kinetic parameters for isolated reactions cannot simply be added up when several components are mixed, and the behavior of the system agrees with the presence of a unique mediator that constitutes the core of a network of competing reactions. In mixed systems, every single amine or imine competes for the same central hub, in accordance with their binding affinity for the catalyst metal ion center. More generally, the study extends the basic principles of constitutional dynamic chemistry to interconnected chemical transformations and provides a step toward dynamic systems of increasing complexity.

Full text of DOI:10.1021/ja042469q

Published on Web 03/22/2005
Scandium(III) Catalysis of Transimination Reactions.
Independent and Constitutionally Coupled Reversible
Processes
Nicolas Giuseppone,* Jean-Louis Schmitt, Evan Schwartz, and Jean-Marie Lehn*
Contribution from the Laboratoire de Chimie Supramole´culaire, Institut de Science et
d’Inge´nierie Supramole´culaires, UniVersite´ Louis Pasteur, 8 Alle´e Gaspard Monge, BP 70028,
67083 Strasbourg Cedex, France
Received December 15, 2004; E-mail: giuseppone@isis.u-strasbg.fr; lehn@isis.u-strasbg.fr
Abstract: Sc(OTf)3 efficiently catalyzes the self-sufficient transimination reaction between various types of
CdN bonds in organic solvents, with turnover frequencies up to 3600 h-1 and rate accelerations up to 6
× 105. The mechanism of the crossover reaction in mixtures of amines and imines is studied, comparing
parallel individual reactions with coupled equilibria. The intrinsic kinetic parameters for isolated reactions
cannot simply be added up when several components are mixed, and the behavior of the system agrees
with the presence of a unique mediator that constitutes the core of a network of competing reactions. In
mixed systems, every single amine or imine competes for the same central hub, in accordance with their
binding affinity for the catalyst metal ion center. More generally, the study extends the basic principles of
constitutional dynamic chemistry to interconnected chemical transformations and provides a step toward
dynamic systems of increasing complexity.
Introduction
reorganization on both the molecular (covalent) and the su-
pramolecular (noncovalent) levels.
Over the last century, numerous studies have been devoted
to CdN-type bonds,1 firmly establishing their relevance in the
fields of chemistry,2 biology,3 and material science.4 Despite
such a venerable age, the literature reveals that research activities
related to these functional groups are more active than ever.5
The versatility of this chemical linkage is mainly due to its fast
and highly chemoselective formation from amino groups and
carbonyl derivatives. A characteristic feature that has become
of special interest is the reversibility of the CdN linkage and
the resulting ability of imino-type compounds to undergo
exchange of the two components against other functionally
compatible partners. It is of particular value for the recently
developed dynamic combinatorial (covalent) chemistry (DCC),
belonging to the molecular domain of constitutional dynamic
chemistry (CDC),6,7 which covers reversible constitutional
DCC depends crucially on the availability of fast reversible
reactions, allowing the generation of highly diverse libraries of
interconnecting compounds, thus creating dynamic constitutional
diversity for both biological and materials science purposes.
Provided fast and efficient reversibility in water or organic
solvents is achieved, the ubiquitous CdN unit, present, for
example, in imines, hydrazones, and oximes, offers highly
attractive potentialities for molecular CDC. Three different types
of equilibrated reactions may allow the exchange of the amino
fragment of the CdN units (Scheme 1). In the first case (a),
the formation/hydrolysis mechanism,1b exchange of the com-
ponents occurs via back-reaction with water to convert the imine
bond into free amine and carbonyl, which can subsequently
undergo reaction with another partner present in the reaction
mixture. In the other two cases, transimination8 (b) and imine
metathesis9 (c), the process is self-sufficient, not involving the
mediation of a third constituent (i.e., water), and the exchange
occurs without disruption of the CdN unit into free amine and
carbonyl.10 This feature is of special interest in CDC for at least
three reasons: (a) numerous reactions have to be performed in
anhydrous conditions; (b) a highly shifted equilibrium is often
(1) (a) Schiff, H. Ann. 1864, 131, 118. (b) Layer, R. W. Chem. ReV. 1963, 63,
489-510.
(2) Katritzky, A. R.; Meth-Cohn, O.; Rees, C. W. ComprehensiVe Organic
Functional Group Transformations; Pergamon: Elmsford, NY, 1995; Vol.
3, pp 403-423.
(3) Stryer, L. Biochemistry; W. H. Freeman and Co.: San Francisco, CA, 1975;
Chapter 18, pp 432-455.
(4) (a) Gray, G. W. J. Chem. Soc. 1958, 552-555. (b) Simon, J.; Bassoul, P.
Design of Molecular Materials, Supramolecular Engineering; Wiley: New
York, 2000; Chapter 8, pp 405-427.
(5) A bibliographic research using SciFinder on more than 19 000 papers related
to the term “imines” indicates that, since 1905, one-half of the publications
appeared in the last 10 years.
(8) For imines transimination, see: (a) Reddelien, G. Ber. 1920, 53B, 355-
358. For oximes transimination, see: (b) Hauser, C. R.; Hoffenberg, D. S.
J. Am. Chem. Soc. 1955, 77, 4885-4887. For pyruvamides transimination,
see: (c) Lopatin, W.; Young, P. R., Jr.; Owen, T. C. J. Am. Chem. Soc.
1979, 101, 960-969.
(9) (a) To´th, G.; Pinter, I.; Messmer, A. Tetrahedron Lett. 1974, 9, 735-738.
(b) Zuckerman, R. L.; Krska, S. W.; Bergman, R. G. J. Am. Chem. Soc.
2000, 122, 751-761.
(10) For the “atom economy” synthetic approach, see: Trost, B. M. Science
1991, 254, 1471.
(6) (a) Lehn, J.-M. Chem.-Eur. J. 1999, 5, 2455-2463. (b) Cousins, R. L.;
Poulsen, S. A.; Sanders, J. K. M. Curr. Opin. Chem. Biol. 2000, 4, 270-
279. (c) Rowan, S. J.; Cantrill, S. J.; Cousins, G. R. L.; Sanders, J. K. M.;
Stoddart, J. F. Angew. Chem. 2002, 114, 938-993; Angew. Chem., Int.
Ed. 2002, 41, 898-952.
(7) (a) Lehn, J.-M. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 4763-4768. (b)
Lehn, J.-M. Science 2002, 295, 2400.
9
5528
J. AM. CHEM. SOC. 2005, 127, 5528-5539
10.1021/ja042469q CCC: $30.25 © 2005 American Chemical Society

Scandium(III) Catalysis of Transimination Reactions
A R T I C L E S
Scheme 2. Exchange Reaction between Imine 1 and
Cyclopentylamine (Equimolar Ratio), Catalyzed by Sc(OTf)3, and
Leading to a Thermodynamic Equilibrium with Imine 2 and
Naphthylamine
Scheme 1. Types of Amine Exchange Processes Involving the
Imine Bond: (a) Condensation and Hydrolysis; (b) Transimination;
(c) Imine Metathesis
processes (turnover frequency, TOF up to 3.6 × 103 h-1 at 25
°C), with respect to both the noncatalyzed (accelerations up to
6 × 105) and the Bro¨nsted acid-catalyzed ones. Moreover, we
broaden the applicability of this reaction by using microencap-
sulated scandium triflate, which has emerged as a promising
supported catalyst, especially for further dynamic constitutional
approaches. Finally, we show that the results obtained are
consistent with a predominantly concerted transimination path-
way, where both the imine and the free amine occupy the
coordination sphere of the scandium in a ternary intermediate
(“promnastic” effect).15a As a consequence, when several
different components are brought together in constitutionally
dynamic combinatorial systems, as compared to individual
reactions, coupled equilibria are established through the forma-
tion of mediator species functioning as a hub where reagents
and products compete for the same catalytic site.
required for the design of dynamic materials, in particular,
dynamers;11-13 and (c) the decrease in number of constituents
is advantageous for the determination of the composition of a
dynamic library; indeed, the presence of significant amounts
of starting materials (e.g., carbonyl and amino compounds) as
well as of reaction intermediates (e.g. hemiaminal) may
complicate the analyses of mixtures.
Although many systems have taken advantage of the revers-
ibility of the imine-type bond in water, according to the
condensation/hydrolysis process,14 fast exchange of these key
units in organic solvents remains difficult to achieve. This is
particularly apparent for the more stable hydrazones and oximes,
as judged from the small number of publications related to the
catalysis of this reaction.15 We have recently studied the rate
of the transimination reaction in the presence of lanthanide ions
as catalyst. The exchange rates increased linearly with a decrease
in ionic radius, the most efficient catalyst being scandium(III).16
In particular, exchange reactions involving molecular helical
strands based on polyheterocyclic units linked by hydrazone-
type connections could be efficiently catalyzed by scandium
triflate in chloroform,16,17 whereas protons did not afford more
than traces of the exchanged products. We here report a more
detailed study of this scandium-catalyzed exchange in Schiff
bases (such as compounds 1 and 2) and extend it to a broader
range of CdN units (i.e., imines and oximes). We have
investigated structural effects on the thermodynamic distribution
of products (selectivity) as well as on the exchange rate. We
also highlight the efficiency of the Lewis acid-catalyzed
Results and Discussion
I. Catalytic Aspects. A. Catalytic Efficiency of Sc(OTf)3
in Exchange Processes. To evaluate the catalytic efficiency of
scandium triflate in exchange processes, we first studied the
exchange between imine 1 and cyclopentylamine, leading to
the equilibrium with the imine 2 and R-naphthylamine (Scheme
2), in deuterated chloroform in the presence of different amounts
of Sc(OTf)3 as catalyst. The evolution of the equilibrium was
followed using 1H NMR at 25 and 60 °C for different
percentages of metallic salt, and the data obtained are plotted
in Figure 1.
The results indicate that the exchange process strongly
depends on the percentage (with respect to amine) of scandium
triflate in solution, as well as on the temperature. At 25 °C, the
uncatalyzed reaction is characterized by an initial rate (V0)18 of
1.9 × 10-2 mM h-1 for an initial concentration 1i ) 22.6 mM,
which is roughly 100 times slower than the rate when 20% of
ScIII is added (1.96 mM h-1). Raising the temperature to 60 °C
(11) The term “dynamers” designates dynamic polymers connected by either
reversible covalent or noncovalent bonds; see ref 7. For dynamic materials,
see: (a) Lehn, J.-M. In Supramolecular Science: Where It Is and Where It
Is Going; Ungaro, R., Dalcanale, E., Eds.; Kluwer: Dordrecht, The
Netherlands, 1999; pp 287-304. (b) Lehn, J.-M. Polym. Int. 2002, 51, 825-
839. (c) Polyacylhydrazones: Skene, W. G.; Lehn, J.-M. Proc. Natl. Acad.
Sci. U.S.A. 2004, 8270-8275. (d) Lehn, J.-M.; Giuseppone, N. Dynamic
Polymers: Fluorescent Polymeric Materials Exhibiting ReVersible Com-
ponent Exchange. Provisory patent application EV337617182US, 2003.
(12) For the constitutional control of dynamers using external stimuli, see:
Giuseppone, N.; Lehn, J.-M. J. Am. Chem. Soc. 2004, 126, 11448-11449.
(13) Nishinaga, T.; Tanatani, A.; Oh, K.; Moore, J. S. J. Am. Chem. Soc. 2002,
124, 5934-5935.
(14) (a) Huc, I.; Lehn, J.-M. Proc. Natl. Acad. Sci. U.S.A. 1997, 94, 2106-
2110. (b) Hochgu¨rtel, M.; Kroth, H.; Piecha, D.; Hofmann, M. W.; Nicolau,
C.; Krause, S.; Schaaf, O.; Sonnenmoser, G.; Eliseev, A. V. Proc. Natl.
Acad. Sci. U.S.A. 2002, 99, 3382-3387.
(15) For zinc-catalyzed transimination in water, see: (a) Leach, B. E.; Leussing,
D. L. J. Am. Chem. Soc. 1971, 93, 3377-3384. For acid-catalyzed
transimination in water, see: (b) Polyakov, V. A.; Nelen, M. I.; Nazarpack-
Kandlousy, N.; Ryabov, A. D.; Eliseev, A. V. J. Phys. Org. Chem. 1999,
12, 357-363. For amine exchange in Schiff-base complexes, see: (c)
Lindoy, L. F. Q. ReV. 1971, 25, 379-391. (d) Olszewski, E. J.; Martin, D.
F. J. Inorg. Nucl. Chem. 1965, 27, 345-351. (e) Verter, H. S.; Frost, A. E.
J. Am. Chem. Soc. 1960, 82, 85-89. For amino exchange of hydrazones
in metal complexes, see: (f) Dong, G.; Chun-ying, D.; Ke-liang, P.; Qing-
jin, M. J. Chem. Soc., Chem. Commun. 2002, 1096-1097.
increases the rate of the uncatalyzed exchange to 0.35 mM h-1
,
whereas in the presence of 4 mol % of catalyst in the same
conditions the rate increases to 10 mM h-1 (i.e., an acceleration
of a factor 30). In terms of selectivity, as long as low relative
concentrations of catalyst are used (no more than 4 mol %),
the ratios at equilibrium at both temperatures are equal to those
obtained without scandium ions. For higher quantities (20 mol
% of metal salt), the catalyst interferes with the thermodynamic
equilibrium of the system and tends to favor compound 1 over
2. Indeed, the ratio at equilibrium between imines 1 and 2 varies
from 20/80 (4 mol % of ScIII) to 60/40 (20 mol % of ScIII) at
60 °C, and from 10/90 (4 mol % of ScIII) to 40/60 (20 mol %
of ScIII) at 25 °C. This effect may be explained by the fact that
ScIII is expected to coordinate more strongly the most basic
(16) Giuseppone, N.; Schmitt, J.-L.; Lehn, J.-M. Angew. Chem. 2004, 116,
5010-5014; Angew. Chem., Int. Ed. 2004, 37, 4902-4906.
(17) For an overview of the use of rare-earth metal triflates in organic synthesis,
see: Kobayashi, S.; Sugiura, M.; Kitagawa, H.; Lam, W. W.-L. Chem. ReV.
2002, 102, 2227-2302.
(18) The V0 values were determined graphically from the corresponding plots
of the evolution of the equilibria versus time.
9
J. AM. CHEM. SOC. VOL. 127, NO. 15, 2005 5529

A R T I C L E S
Giuseppone et al.
Figure 1. Evolution with time of the equilibration between imines 1 and 2 described in Scheme 2 (1i ) 22.6 mM), using various amounts of Sc(OTf)3
catalyst (added as a deuterated acetonitrile solution) and followed by the integration of the 1H NMR signals of the imine proton at 60 °C (a) and 25 °C (b).
The different amounts of catalyst used, relative to cyclopentylamine, are indicated as follows: O 0%; b 0.5%; 2 2%; 4 4%; 9 20%.
amine, here the cyclopentylamine,19 thus shifting the equilibrium
in favor of 2, as was observed in the related case of a bis-imine.12
Scheme 3. Exchange Reactions between Imine 1 and
Benzylamine, 4-Methoxyaniline, Aniline, 3-Nitroaniline, or
2,4-Dinitroaniline (Equimolar Ratios) Leading to Thermodynamic
As a consequence, the selectivity of the exchange depends on
the relative affinity of scandium for the amines involved.
Subsequently, the use of 4 mol % of ScIII appeared to be best
suited for our purposes: it both maximizes the rate of the
exchange process and does not interfere with the equilibrium,
as obtained without catalyst, a crucial point for the formation
of isoenergetic libraries of compounds.6a
Equilibria with Imines 3, 4, 5, 6, and 7, Respectively; Undetectable
Reaction in the Case of 2,4-Dinitroaniline
B. Exchange Equilibria and Kinetics; Structural Effects.
The influence of the nature of the amine on the kinetic and
thermodynamic parameters of the exchange process was studied
by comparing the uncatalyzed and Sc(OTf)3-catalyzed reactions
of the imine 1 with equimolar amounts of benzylamine,
4-methoxyaniline, aniline, 3-nitroaniline, or 2,4-dinitroaniline
leading to imines 3, 4, 5, 6, and 7, respectively (Scheme 3).
The corresponding equilibria, given by the selectivities (S), the
initial rates (V0), and the turnover frequencies (TOF), are listed
in Table 1.
In terms of selectivity (y/x ratio between the imines at
equilibrium) and for both the catalyzed and the uncatalyzed
pathways, the results correlate with the nucleophilicities of the
amines which compete for imine formation, the most reactive/
basic amine leading to the thermodynamically most stable imine
(2 > 3 > 4 > 5 > 1 > 6 > 7), consistent with the literature
data on the addition of amines to aldehydes.20 The same general
behavior was observed starting from the more basic imine 2
that undergoes exchange reactions with R-naphthylamine, ben-
zylamine, allylamine, and isopentylamine to yield imines 1, 3,
8, and 9, respectively (Scheme 4 and Table 2).
C
U
The initial rates of the catalyzed (V0 ) and uncatalyzed (V0
)
amine exchange reactions of a number of imines are listed in
C
U
Tables 1 and 2, together with the acceleration factors (V0 /V0
)
and the turnover frequencies (TOF) of the catalyzed process.
The initial exchange rate from 1 for the uncatalyzed pathway
showed weak correlation with the structure of the free amine,
the most basic exchanging the fastest (entry 1, Table 1).21 In
the case of the scandium triflate-catalyzed reaction, the trend
was opposite. Indeed, the initial rates and TOF (up to 1.8 ×
103 h-1) as well as the rate increase (up to 5 orders of
magnitude) as compared to the control (uncatalyzed) experi-
ments (entry 4, Table 1) displayed the sequence: aniline >
4-methoxyaniline . 3-nitroaniline . benzylamine . cyclo-
pentylamine. The results suggested that the exchange rates were
highest for amines of comparable nucleophilicities, regardless
of the direction of the reaction. Indeed, the initial rates were
similar in both directions as was verified by starting from
compound 2 in the presence of R-naphthylamine and different
amounts of ScIII catalyst (Figure 2).
(19) Dash, A. C.; Nanda, R. K. J. Am. Chem. Soc. 1969, 91, 6944-6947.
(20) (a) Conant, J. B.; Bartlett, P. D. J. Am. Chem. Soc. 1932, 54, 2881-2899.
(b) Sander, E. G.; Jencks, W. P. J. Am. Chem. Soc. 1968, 90, 6154-6162.
(21) Porai-Koshits, B. A.; Remizov, A. L. Probl. Mekhanizma Org. Reaktsii,
Akad. Nauk Ukr. SSR, Otdel Fiz-Mat. I Khim. Nauk 1953, 238-246; Chem.
Abstr. 1956, 50, 16686.
Similarly, the exchange reactions of compound 2 (containing
the basic cyclopentylamine moiety) with benzylamine, allyl-
amine, and isopentylamine leading to imines 3, 8, and 9,
9
5530 J. AM. CHEM. SOC. VOL. 127, NO. 15, 2005

Scandium(III) Catalysis of Transimination Reactions
A R T I C L E S
Table 1. Values of the Selectivity Ratios (S), the Initial Catalyzed (V0C) and Uncatalyzed (V0U) Rates, and the Turnover Frequencies (TOF)
for the Exchange Reactions Starting from Imine 1 and Leading to the Equilibria with the Imines 2, 3, 4, 5, 6, and 7, at 25 °C in CDCl3 and in
the Presence of 4 mol % of Sc(OTf)3 Catalysta (Schemes 2 and 4)
1
1
1
equilibrium
S
V0C/mM h-
V0U/mM h-
acceleration
TOF/h-
ScIII
U
entry
x
h
y
(y/x)
4% Sc(OTf)3
control
V0C/V0
1
2
3
4
5
6
1 h 2
1 h 3
1 h 4
1 h 5
1 h 6
1 h 7
9/1
7/3
2/1
1/1
1/3
f0a
0.84
18.6
19 × 10-3
6.0 × 10-3
5.6 × 10-3
5.4 × 10-3
45
0.93
20.6
1200
1800
540
nd
3.1 × 104
1.9 × 105
3.0 × 105
1.10 × 103
1.62 × 103
486
nd
nd
nd
1
a Added as 45 mM acetonitrile solution; final solution contains 2 vol % acetonitrile in CDCl3. b No detectable amount of 7 by H NMR.
(entry 8) (H+ and ScIII having similar selectivities). In the cases
of the most basic amines, catalysis by H+ is less efficient as
protonation decreases the reactivity of the amines and, by
scavenging the protons, prevents activation of the imines. For
the less basic amines (entries 5 and 6, Table 3), these effects
are less pronounced and proton catalysis becomes comparable
Scheme 4. Exchange Reactions between Imine 2 and
R-Naphthylamine, Benzylamine, Allylamine, or Isopentylamine
(Equimolar Ratios) in CDCl3 and in the Presence of Catalytic
Amounts of Sc(OTf)3, Leading to Thermodynamic Equilibria with
Imines 1, 3, 8, and 9, Respectively
C
to ScIII catalysis. Furthermore, in all cases where V0 /V0H > 1,
attempts to increase V0H by lowering the pH led to degradation
of the imines.
Examining reactivities in different solvents showed that ScIII
catalysis was most pronounced in solvents of low dielectric
constant (Table 4). On the other hand, the use of coordinating
solvents greatly decreased the effect of ScIII, whereas it affected
less the proton-catalyzed process.
D. Effect of Ionic Radius on Metal Ion Catalysis. Crossover
experiments were conducted with salts of different trivalent
metal ions. The initial rates and turnover frequencies in the
exchanges involving 2 and allylamine catalyzed by different
lanthanide triflates were found to be linearly correlated with
the ionic radii of the trivalent ions (Figure 3). This behavior
agrees with that already reported for exchange of hydrazine
derivatives, which furthermore were much more reluctant to
undergo proton catalysis.16
E. Five-Component Competitive Experiment in a Dynamic
Combinatorial Library Approach. To extend the scope of
such a catalytic system to DCC, we finally performed a direct
competitive experiment involving the reaction of 1 with a
mixture of cyclopentylamine, benzylamine, aniline, and dini-
troaniline in equimolar ratios at 25 °C leading to a potential
mixture of imines 1, 2, 3, 5, and 7. The time dependence of the
equilibration is shown in Figure 4.
respectively (Scheme 4, Table 2), showed a global behavior in
which the more closely related the structures of the competing
C
U
amines, the faster were the accelerations Vo /Vo as compared
to the controls without catalyst. Comparing the ratio of the
reaction rates with and without catalyst minimizes the interfer-
ences of other factors, the intrinsic steric or electronic param-
eters. In terms of the absolute reaction rate, the best results were
obtained for allylamine with a TOF of 2800 h-1 (entry 3).
C. Comparison of ScIII and H+ as Exchange Catalysts;
Solvent Effects. To compare the relative catalytic efficiency
of ScIII and protons, we examined the effect of hydrated Sc-
(OTf)3 and of CF3CO2D on different exchange reactions in
deuterated chloroform starting from compounds 1 or 2 (Table
3).
In most of the cases studied, the use of identical percentages
of ScIII or H+ revealed the higher activity of the metal salt (in
terms of initial rate) even though the exchange also proceeded
efficiently using H+. For example, the exchange between 1 and
3 at 25 °C was 15 times faster under scandium triflate catalysis
than under H+ catalysis (entry 4), while the initial rate of the
exchange between 2 and 8 was 100 times greater with ScIII
After 4 days, the final composition of the mixture followed
the order of nucleophilicities (55% of 2; 41% of 3; 2.3% of 5;
1.7% of 1; 0% of 7), as expected from a superposition of the
previous two-imine competitive experiments (see selectivities
in Table 1). The initial reaction rate for the whole mixture was
1
measured by observing via H NMR the disappearance of the
C
CHdN signal of imine 1. The results gave a V0 of 2.72 mM
h-1 and a TOF of 0.75 h-1, which are comparable to the V0
C
and TOF values obtained for the reaction involving 1 and
cyclopentylamine alone (Table 1, entry 1). These data are far
from what one would predict if the reaction occurred indepen-
dently (e.g., formation of the kinetic products at first, followed
by their progressive equilibration to give the thermodynamic
ones), based on the kinetic parameters determined (Table 1).
Thus, for instance, one would have expected initial formation
C
of imine 5 by exchange between 1 and aniline (V0 ) 1620
mM h-1; TOF ) 1800 h-1; Table 1, entry 4).
9
J. AM. CHEM. SOC. VOL. 127, NO. 15, 2005 5531

A R T I C L E S
Giuseppone et al.
Table 2. Values of the Selectivity Ratios (S), the Initial Catalyzed (V0C) and Uncatalyzed (V0U) Rates, and the Turnover Frequencies (TOF)
for the Exchange Reactions Starting from Imine 2 and Leading to Equilibria with the Imines 1, 3, 8, and 9, at 25 °C in the Presence of 4 mol
% of Sc(OTf)3 Catalysta (Scheme 5)
1
1
1
equilibrium
S
V0C/mM h-
V0U/mM h-
acceleration
TOF/h-
ScIII
U
entry
x
h
y
(y/x)
4% Sc(OTf)3
control
V0C/V0
1
2
3
4
2 h 1
2 h 3
2 h 8
2 h 9
1/9
2/3
1/2
5/4
0.84
12.7 × 10-3
8.54
3.29
66
128
760
913
0.93
1220
2830
900
1.1 × 103
2.5 × 103
813
0.89
a Added as 45 mM acetonitrile solution; final solution contains 2 vol % acetonitrile in CDCl3.
Figure 2. Evolution with time of the formation and the disappearance of compound 2 in the equilibration between 1 and 2 (1i ) 2i ) 22.6 mM) in CDCl3
at 60 °C, starting from 1 and cyclopentylamine (1:1) (direction I) or 2 and R-naphthylamine (1:1) (direction II). The equilibria were followed by integration
of the 1H NMR imine signals, in the presence of either 2 or 4 mol % Sc(OTf)3, relative to free amines, labeled as follows: O 2% (I); 4 4% (I); b 2% (II);
2 4% (II).
Table 3. Comparative Values, Determined by 1H NMR, of the Selectivity Constants (S) and of the Initial ScIII-Catalyzed (V0C) and
Proton-Catalyzed (V0H) Rates for the Exchange Reactions between Imines 1, 2, 3, 4, 5, 8, and 9 (Schemes 4 and 5) in the Presence of
Catalytic Amounts of Sc(OTf)3 and CF3CO2D
1
1
equilibrium
temp
S (y/x)
(ScIII
S (y/x)
(H+)
V0C/mM h-
V0H/mM h-
(% H+)
(% ScIII
)
V0C/V0
H
entry
x
h
y
/
°
C
)
1
2
3
4
5
6
7
8
9
1 h 2
60
60
60
20
20
20
20
20
20
4/1
4/1
2/3
7/3
2/1
1/1
3/4
1/2
1/2
7/3
7/3
2/3
7/3
2/1
3/2
3/4
1/2
1/2
10 (4%)
1.2 (4%)
0.9 (2%)
5.5 (40%)a
1.2 (4%)
8.3
4.4
1.8
15.5
1
0.46
14.5
100
12
4.1 (2%)
10 (20%)
18.6 (4%)
1 h 3
1 h 4
1 h 5
2 h 3
2 h 8
2 h 9
1.1 × 103 (4%)
1.6 × 103 (4%)
1.1 × 103 (4%)
2.5 × 103 (4%)
1.1 × 103 (4%)
3.5 × 103 (4%)
77 (4%)
24.8 (4%)
69 (4%)
813 (4%)
a Decomposition into free 9-anthracencarboxaldehyde (20 mol % relative to the sum of imines) was observed.
Table 4. Comparative Values, Determined Using 1H NMR, of the Initial Rates (V0 and V0H) in Various Deuterated Solvents at 60 °C for the
Exchange Process between Imines 1 and 2 (Scheme 2) Using Catalytic Amounts of Sc(OTf)3 and CF3CO2D
C
1
1
1
H
V0C/mM h-
V0H/mM h-
(% H+)
VoU/mM h-
VoC/Vo
entry
solvent
(% ScIII
)
control
(VoC/Vo
)
U
1
2
3
4
5
6
CDCl3
C2D2Cl4
toluene-d8
CDCl3/MeOD (95/5)
CDCl3/MeOD (50/50)
DMSO-d6
10 (4%)
1.2 (4%)
0.35
1.25
8.3 (28)
(34.4)
11.7 (451)
1 (-)
0.46 (-)
0.64 (5.5)
43 (4%)
3.1 (4%)
1 (4%)
0.52 (4%)
1 (4%)
1.35 × 10-2
0.2 (4%)
0.43 (4%)
5.8 × 10-2 (4%)
9 × 10-2 (4%)
1.06 × 10-2
F. Extension of ScIII-Catalyzed Exchange to a Broader
Range of CdN Units. To gain further insight into the scope of
the catalyzed reaction, we investigated different types of
substrates such as those depicted in Scheme 5 with the
corresponding thermodynamic and kinetic data appearing in
Table 5. The reaction is compatible with the presence of electron
rich and/or chelating groups such as phenol in the Schiff bases
(entries 1, 2), even if a decrease in the accelerations values is
observed. As it has been already described in the cases of
heterocyclic helical molecular strands,16 the efficiency of the
exchange is also demonstrated for hydrazone derivatives, even
those that do not contain pyrimidine or pyridine moieties (entry
3). Crossover experiments mixing different CdN-type bonds
may also be performed, such as between hydrazones and oximes
(entry 4). Nevertheless, attempts to exchange hydrazine, and
even more hydroxylamine moieties, against amines failed due
to a highly displaced equilibrium in favor of the most nucleo-
philic nitrogen group. The best result in terms of catalytic
activity (accelerations up to 6 × 105 and TOF of more than
3600 h-1) was obtained in exchange between the oxime 17 of
cyclohexanone and benzylhydroxylamine (entry 5). This result
is particularly interesting in view of the great reluctance of
9
5532 J. AM. CHEM. SOC. VOL. 127, NO. 15, 2005

Scandium(III) Catalysis of Transimination Reactions
A R T I C L E S
catalyst (Figure 5, left) display a plateau that is reached when
more than 4 mol % of catalyst is used. These data fit those
reported for the general acid-catalyzed imine formation,26 where
the catalyst accelerates the rate-determining addition step by
polarizing the imine bond and facilitating the nucleophilic attack
of the free amine, here through cation coordination to its nitrogen
site (Scheme 7). At higher percentages of catalyst, the rate of
the addition step forming the geminal diamino intermediate
reaches a plateau due to deactivation of the amine by the Lewis
acid center.
For low quantities of ScIII (0.05-4 mol %), the plot of log-
(V0) against log(ScIII) shows a linear dependence of rate on
catalyst concentration, demonstrating that the rate of the
uncatalyzed reaction is comparatively negligible (Figure 5,
right).27 Moreover, the initial rate V0 was found to be indepen-
dent of both the imine and the amine concentrations,28 so that
the overall reaction is first order in this domain.
The behavior of the system revealed by the linear relationship
between the ionic radius of the metal ion and the initial rates/
turnover frequencies (Figure 3) is in favor of Lewis acid-type
catalysis. This conclusion is also in agreement with the
competition of coordinating solvents that can displace amines
and imines from the metal ion and thus decrease the reaction
rate (Table 4); the same holds when comparing proton catalysis
with the more efficient ScIII catalysis (Table 3).
Figure 3. Top: Equilibration between imines 2 and 8. Bottom: Initial
rates (V0C) (9) and turnover frequencies (TOF) (b) for amine exchange
between imine 2 and allylamine under catalysis by different lanthanide
triflates as a function of their ionic radius; equimolar ratio with allylamine;
1i ) 22.6 mM, in the presence of 4% Ln(OTf)3 (added as a deuterated
acetonitrile solution).
oximes to undergo exchange.15b Last, but not least, the pos-
sibility of using highly stable five-membered ring imines as well
as amidines with hydroxylamines might be of interest for future
endeavors.
G. ScIII-Supported Catalyst. As an extension of the ScIII
catalysis in homogeneous solutions, we investigated different
types of supported scandium triflate in organic solvents. Such
a heterogeneous process is of special interest in view of its
potential implementation in CDC. Indeed, the use of a supported
catalyst to equilibrate a library offers the possibility to freeze
the reaction by simply removing the supported metal ions from
the solution. We chose two different techniques to support the
scandium triflate: microencapsulation in polystyrene22 and
immobilization on ion-exchange resins.23
In the case of ion-exchange resins, no activity was preserved.
With the polystyrene microencapsulated scandium (MC Sc-
(OTf)3) using acetonitrile as solvent (Scheme 6), the exchange
proceeded in mild conditions for different CdN-type bonds
investigated above, good catalytic activities being conserved as
compared to the control experiments with free ScIII (Table 6).
II. Mechanistic Aspects. To obtain a first indication on the
mechanistic pathway, the reaction of 1 with benzaldehyde (1:1
ratio) was performed at 25 °C in the presence of 4 mol % of
Sc(OTf)3.24 After 3 days, no cross-product could be detected.
Also, no direct imine metathesis between 1 and another Schiff
base (formed from p-anisidine and benzaldehyde) was observed
under the same conditions (25 °C and 4 mol % of catalyst).
Indeed, even though traces of exchange products were detected
via 1H NMR, they appeared to mostly derive from the presence
of residual water.25
B. Effects of Amine Basicities. The results in Tables 1 and
C
U
2 show that the acceleration factors V0 /V0 are largest for
amines of basicity comparable to that of the amine engaged in
the imine. Furthermore, for similar basicities, the accelerations
are higher for less basic partners than for more basic ones.
Indeed, the accelerations are greater (up to a factor of 300) when
two aromatic amines are engaged in crossover experiments, as
compared to what is obtained between aliphatic amines and
aliphatic imines (Table 1, entry 4 and Table 2, entry 4),
indicating that the activity of the scandium catalyst itself is
affected by the global basicity of both amines and imines. Based
on the data obtained for the effect of the nature of the
components (Tables 1 and 2), one may propose that the rate of
the exchange reaction is dependent upon three main factors:
activation of the imine by ScIII coordination; deactivation of
the amine by binding to ScIII ions; and deactivation of the ScIII
by both the bound amine(s) and the imine(s). All three factors
(25) Hydrolysis of the imine bonds has been detected by the observation of the
1H NMR aldehyde signal. This hydrolysis, catalyzed by ScIII, is due to
traces of water in the organic solvents (chloroform and acetonitrile) as well
as to the water coordinated to the scandium ion (4.3 molecules of H2O
were found by elemental analysis of Sc(OTf)3). When the reaction was
performed under the transimination protocol, hydrolysis did not occur as
the aldehyde is not detected. Indeed, had this aldehyde formed, it should
have been detectable, as in the presence of scandium, the transimination
reaction was found to be faster than the condensation of aldehyde and amine.
For example, the exchange of 1 and 5 exhibits a V0 of 1.62 × 103 mM
C
h-1 with 4% ScIII (see, Table 1, entry 4), while the formation of compound
5, from 9-anthracenecarboxaldehyde and aniline in the presence of 4% of
catalyst, displays a rate of only 49.5 mM h-1 (33 times slower). The control
without catalyst has a value of 0.18 mM h-1 (9 × 103 times slower).
(26) For general acid catalysis, see: (a) Jencks, W. P. J. Am. Chem. Soc. 1959,
81, 475-481. (b) Sayer, J. M.; Jencks, W. P. J. Am. Chem. Soc. 1973, 95,
4277-4287.
These results indicate that the exchange described in Scheme
2 occurs via the transimination pathway (Scheme 1b).
A. Effects of Catalyst Concentration. At both 25 and 60
°C, the plots of the initial rates V0 against the percentage of
(27) To avoid the in situ liberation of protons, as slowly occurs in CDCl3,
reactions intended to study the linear correlation between the initial rate
and the catalyst concentration were performed in deuterated tetrachloro-
ethane.
(28) Experiments were set up for 10 different values of imine concentration
from 0.05 equiv to 0.9 equiv (1.13, 4.52, 6.78, 9.04, 11.3, 13.6, 15.8, 18.1,
and 20.3 mM) relative to the amine (22.6 mM; 4 mol % catalyst); and for
eight different values of amine concentration from 0.1 equiv to 1.8 equiv
(2.26, 4.52, 9.04, 18.4, 27.1, 36.2, and 40.7 mM) relative to the imine (22.6
mM; 4 mol % catalyst).
(22) Kobayashi, S.; Nagayama, S. J. Am. Chem. Soc. 1998, 120, 2985-2986.
(23) (a) Kobayashi, S.; Nagayama, S. J. Org. Chem. 1996, 61, 2256-2257. (b)
Yu, L.; Chen, D.; Li, J.; Wang, P. G. J. Org. Chem. 1997, 62, 3575-
3581.
(24) For mechanistic studies of imine formation reactions (from amines and
aldehydes) catalyzed by zinc(II) ions in organic solvents, see: Boate, A.
R.; Eaton, D. R. Can. J. Chem. 1977, 55, 2432-2441.
9
J. AM. CHEM. SOC. VOL. 127, NO. 15, 2005 5533

A R T I C L E S
Giuseppone et al.
Figure 4. Left: Evolution over time of the equilibration of the imines 1, 2, 3, 5, and 7 in CDCl3 at 25 °C starting from imine 1 (1i 22.6 mM) mixed with
equimolar amounts of the amines, cyclopentylamine, benzylamine, aniline, and 2,4-dinitroaniline, corresponding to the imines 2, 3, 5, and 7 and using 16
mol % of Sc(OTf)3 (relative to 1) added as a CD3CN solution. The products are labeled as follows: 9 (1); [ (2); 1 (3); 2 (5); b (7). Right: Selected 400
1
MHz H NMR spectra at different times corresponding to the graph on the left.
effect).15a The behavior of the individual parallel reactions as
Scheme 5. Exchange Reactions between Various Imines,
Hydrazones, Oximes, and an Imide in Deuterated Chloroform and
compared to that of the mixture leads to a set of reactions
in the Presence of Catalytic Amounts of Sc(OTf)3, Added from a
coupled through a common catalyst that may be schematically
represented according to the change in their constitutional
connectivity, as depicted in Figure 6.
CD3CN Solution
In this diagram, the reactions starting from imine 1 and
various amines are represented using nodes and links having
different sizes and thicknesses corresponding, respectively, to
the relative thermodynamic (S) and kinetic (V0) parameters that
both depend on the relative basicity of the amines. In both cases,
individual (left) and coupled (right) equilibria, the basicity of
the amine determines the distribution S at the equilibrium.29 On
the contrary, the initial rates V0 follow an opposite sequence as
a function of basicity on going from the individual to the
competing system. Such a behavior agrees with the presence
of a unique intermediate that constitutes the core of a network
of competing reactions. In the mixed system, every single amine
or imine competes for the same central hub, in accordance with
its binding activity for the catalytic metal ion center. Thus, the
most basic amine (cyclopentylamine) monopolizes the hub and
determines the rate of the overall system because it acts as the
most efficient competitor in the transimination process.30
The competitive mixed system (Figure 6, right) may be
illustrated by the crossover reactions occurring between 1,
cyclopentylamine, and aniline, depicted in Scheme 8. We
suggest that the rate-determining step, for each competitive
catalytic cycle, is the addition reaction of the amine to the imine
1 through the ternary intermediates A and A′, on the following
grounds: (1) the steps involving proton transfer display a rapid
exchange on the NMR time-scale and present a negligible
isotope effect;31 (2) the elimination steps from the gem-diamino
compounds (B and B′) are not rate determining as these
intermediates are not observable in the NMR spectra (conse-
quently, k-1, k-1′ . k1, k1′, and k-2, k-2′ . k2, k2′); and (3) the
initial rate V0 is independent of imine/amine concentrations in
increase with the basicity of the amine, whether free or engaged
in the imine.
C. Effects of Component Competition. Bringing together
the results discussed above, and comparing the separate
equilibrating reactions (Tables 1 and 2) with the behavior
observed for the five-membered competitive experiment (Figure
4), allows the formulation of a mechanism for the catalyzed
transimination involving a ternary intermediate with simulta-
neous coordination of at least one amine and one imine (plus
solvent molecules) to the mediator metal ion (“promnastic”
(29) Spencer, J. N. J. Chem. Educ. 1992, 69, 281-284.
(30) A related behavior is found for carrier-mediated transport. See: Jacquez, J.
A. Biochim. Biophys. Acta 1967, 79, 318-328. Also, in the transport of
alkali ions by cryptates, the strongly bound potassium ions saturate the
carrier and strongly decrease the transport rates, see: Kirch, M.; Lehn, J.-
M. Angew. Chem. 1975, 87, 542-543; Angew. Chem., Int. Ed. Engl. 1975,
14, 555-556. Behr, J.-P.; Kirch, M.; Lehn, J.-M. J. Am. Chem. Soc. 1985,
107, 241-246.
(31) The reaction described in Scheme 2 was performed at 25 °C using deuterated
cyclopentylamine and led to an identical initial rate as compared to the
undeuterated amine.
9
5534 J. AM. CHEM. SOC. VOL. 127, NO. 15, 2005

Scandium(III) Catalysis of Transimination Reactions
A R T I C L E S
Table 5. Values of the Selectivity Constants (S), the Initial Catalyzed (V0C) and Uncatalyzed (V0U) Rates, and the Turnover Frequencies
(TOF) for the Exchanges between Imines Hydrazones and Oximes Depicted in Scheme 5
VoC/mM h-
4% ScIII
VoU/mM h-
control
acceleration
TOF/h-
ScIII
1
1
1
equilibrium
S
U
entry
x
h
y
(y/x)
VoC/Vo
1
2
3
4
5
6
7
8
10 h 11
12 h 13
14 h 15
14 h 16
17 h 18
19 h 20a
19 h 21
22 h 23
50
7.7
1.5
4.55
0.92
27.4
97
21
133
3250
0
0.96
6.8
28.5
14.5
412
30.4
108
23.3
147
5.1 × 10-2
3.13
43
5.4 × 10-3
0
6 × 105
3616
1.5
1
813
0.84
1.27
640
18
903
0.93
4.84 × 10-2
a The reaction was performed at both 25 and 60 °C.
Scheme 6. Exchange Reactions Using Polystyrene
Microencapsulated Scandium (MC Sc(OTf)3) as Supported
Catalyst (4 mol %) for Various Crossover-Type Reactions in
CD3CN
One may suppose that the nucleophile amine occupies the
second coordination sphere of the catalyst, which could explain
a zero-order dependence on reactants because a constant
concentration of ions electrostatically localizes in the immediate
vicinity. Such a rapid reversible amine coordination would result
in intermediate states where no amine occupies the empty
coordination site to deactivate the imine and may facilitate the
addition of the outgoing (or incoming) amine to the activated
imine.
When several amines are brought together in crossover
experiments, the imine formation results from competition
between the two pathways of Scheme 8. The results above
(Table 1, entries 1 and 4) indicate that the formation of imine
5 (bottom pathway of Scheme 8) is about 2000 times faster
than the formation of imine 2 (top pathway of Scheme 8). In
the competitive experiment (both pathways operating simulta-
neously), the observed amount of 5 was never higher than 2.5%
over the course of the exchange (Figure 4), and 2 is formed
faster than 5. This result is a clear illustration of the interplay
of processes occurring in competitive experiments with mixtures.
It may be rationalized by considering that the intermediate A
is significantly more stable than A′ because the lone pair of
cyclopentylamine is more electron-rich than that of aniline,
resulting in a stronger binding to ScIII.19 Consequently, the
catalyst is preferentially surrounded in the coordination sphere
by cyclopentylamine, which circumvents the incorporation of
aniline and decreases the catalytic activity. One may generalize
this interpretation by considering that the most strongly
coordinated amine of any competitive mixture will dominate
the denominator of the rate equation. This analysis holds for
the more complex system described in Figures 4 and 6 (right)
involving altogether five amines and five imines.32 The present
mechanistic study reveals that isoenergetic libraries,6a here sets
of amines of similar basicities, are particularly well-adapted for
the ScIII-catalyzed transimination process, as they yield the
fastest exchanges, as well as equi-distributions of products at
equilibrium, a crucial aspect for further instruction/amplification
of the dynamic system.
Table 6. Values of the Turnover Frequencies (TOF) for the
Exchanges between Imines Hydrazones and Oximes Depicted in
Scheme 6 (17i ) 19i ) 10i ) 22.6 mM), Using 4 mol % of Free
Sc(OTf)3 Catalysta in CDCl3, in CD3CN, and Polystyrene
Microencapsulated ScIII in CD3CNb at 25 °C
1
1
1
TOF/h-
TOF/h-
TOF/h-
MC ScIII
equilibrium
free ScIII
free ScIII
entry
x
h
y
(CDCl3)
(CD3CN)
(CD3CN)
1
2
3
17 h 18
19 h 21
10 h 11
3616
903
108
230
451
33
160c
185
23
a Added as a 45 mM acetonitrile solution; final solution contains 2 vol
% acetonitrile solution. b Prepared according to ref 23. c Without ScIII in
CD3CN, no crossover product is detectable by NMR after 3 days.
max
the range studied (saturation effect),28 corresponding to a V0
and suggesting a rapid preequilibration (fast associations and
exchanges between scandium and amine/imine ligands) with
Michaelis-Menton-type kinetics. One may propose that the
mediators A and A′, involved in the rate-determining step,
assemble several amines or imines into their coordination
max
spheres and that the maximum reaction rate is given by V0
II
) kA. Finally, the results of Figure 2 (V0I ) V0 ) indicate that
Conclusions
the proposed transimination reaction is symmetrical for the
forward and the backward reactions (k1 close to k1′ and k2 close
to k2′), as a result of the interplay between relative amine and
imine basicities, that is, relative binding to the catalytic ion. In
terms of microscopic steps for the conversion of A to B in
Scheme 8, the addition is preceded by rapid preequilibria
reestablishing the distribution of ligands in the first and second
coordination spheres of the metal complex, followed by further
rapid equilibria reestablishing this distribution with the products.
We have shown that transimination reactions can be ef-
ficiently catalyzed by scandium(III) acting as a Lewis acid and
produce turnovers up to 3600 h-1 and rate accelerations up to
6 × 105. A wide variety of CdN-type bonds are exploitable
(32) Moreover, the reaction carried out under conditions that are typically used
for constructing Hammett plots, that is, using the same constitutional mixture
with a 4:1 ratio between 1 and each of the free amines, leads to a similar
general behavior, with the product of the most strongly bound amine coming
first, and the others following afterward.
9
J. AM. CHEM. SOC. VOL. 127, NO. 15, 2005 5535

A R T I C L E S
Giuseppone et al.
Figure 5. Left: Dependence of the initial rate V0 for the forward reaction of the equilibrium described in Scheme 2 (1 ) 22.6 mM) in CDCl3, and determined
using 1H NMR at 25 °C (9) and 60 °C (b), on the amount of catalyst used. Right: Plot of log(V0) against log(ScIII) for the equilibrium described in Scheme
2 at 60 °C in C2D2Cl4, and for 0.05, 0.1, 0.5, 1, 2, and 4 mol % of catalyst.
1
Scheme 7. General Representation of the Addition/Elimination
the residual proton solvent signal. In the H NMR assignments, the
chemical shifts are given in ppm. The coupling constants J are listed
in Hz. The following notation is used for the 1H NMR spectral splitting
patterns: singlet (s), doublet (d), triplet (t), multiplet (m), large (l).
The temperatures that are given for the kinetic and thermodynamic
data were directly measured and regulated in the NMR probe using a
thermocouple. Electrospray (ESI-TOF) studies were performed on a
Bruker MicroTOF mass spectrometer (sample solutions were introduced
into the mass spectrometer source with a syringe pump with a flow
rate of 160 µL min-1). Melting Points (Mp) were recorded on a Kofler
Heizblock and on a Bu¨chi Melting Point B-540 apparatus and are
uncorrected. Microanalyses were performed by the Service de Mi-
croanalyze, Institut de Chimie, Universite´ Louis Pasteur.
Steps in the Transimination Process
using the optimal procedures that require only low amounts of
catalyst and soft conditions. Microencapsulated catalyst repre-
sents a potentially useful tool, as it would be of major interest
to realize the reshuffling of compound libraries on solid support,
with or without an immobilized target, to both freeze and isolate
the equilibrated amplified mixtures by simple filtration.
An analysis of the dependence of reaction rates and product
distributions at equilibrium led to the formulation of a mech-
anism involving a ternary intermediate in which ScIII is
simultaneously coordinated to amine and imine (as well as
solvent) molecules. The application to mixtures of amines
reveals the role played by coupled processes between competing
components as compared to independent reactions. It was found
that the closer are the basicities of the component amines, the
faster are the rates of equilibration, thus stressing the advantage
of isoenergetic (isobasic in the present case) libraries. Further-
more, the intrinsic kinetic parameters for isolated reactions
cannot simply be added up when several components are mixed
due to the formation of a hub in a networked system, where
every single component competes for a unique catalyst. This
competitive inter-relation results in a behavior of the mixture
that is markedly different from what might be expected from
consideration of the separate individual reactions. It bears
ressemblance to processes such as enzymatic reactions or carrier-
mediated transport through membranes in the presence of
mixtures of competing substrates. The results obtained are of
particular significance with respect to the implementation of
dynamic combinatorial processes in chemical reactivity and
catalysis. More generally, they extend the basic principles of
constitutional dynamic chemistry to interconnected chemical
transformations and provide a step toward dynamic systems of
increasing complexity.
General Procedures for the Synthesis of Isolated Imines.
Method a. Equimolar amounts of amine (oxime or hydrazine) and
aldehyde (or ketone) were solubilized or suspended in ethanol at a
concentration of 0.1 M, and the reaction mixtures were refluxed
overnight under vigorous stirring. The solutions were then either
evaporated to dryness and the crude residues were purified by flash
chromatography or, after cooling of the solutions to 0 °C, the
precipitates formed from the reaction mixture were filtered and washed
with cold ethanol and the compounds crystallized or precipitated from
AcOEt/hexane mixtures.
Method b. One equivalent of aldehyde (or ketone) and 1.2 equiv of
amine (or oxime) were heated neat, up at 80 °C, for 5 min. The excess
of amine (or oxime) was then removed under vacuum, and the crude
residues were crystallized or precipitated from AcOEt/hexane mixtures.
General Procedure for Crossover Experiments and Determina-
tions of Kinetic Data. A typical protocol is realized by the preparation
of a fresh CDCl3 solution containing the compounds to exchange in a
NMR tube (0.6 mL, 22.6 mM in imine) or a vial (2 mL, 22.6 mM in
imine). The desired amounts of catalyst were immediately added to
this mixture from a CD3CN solution (5-50 µL, 45 mM, 0.2-8.0 vol
% of CD3CN in CDCl3), prior to NMR measurements. The first
spectrum was usually recorded 30 s after addition without spinning in
the probe. The NMR tubes were topped with Teflon caps to keep a
constant concentration.
(E)-N-(Anthracen-10-ylmethylene)naphthalen-1-amine (1). The
product was synthesized using the general method a and was obtained
as a bright yellow powder (yield 85%). Mp 151-152 °C. Anal. Calcd
for C20H19N: C, 90.60; H, 5.17; N, 4.23. Found: C, 90.23; H, 5.23;
N, 4.05. 1H NMR (CDCl3): δ ) 7.32 (d, 3J ) 7.3 Hz, 1H), 7.55-7.66
(m, 8H), 7.85 (d, 3J ) 7.8 Hz, 1H), 7.95 (d, 3J ) 8.8 Hz, 1H), 8.12 (d,
Experimental Section
3
3
3J ) 8.3 Hz, 2H), 8.47 (d, J ) 7.3 Hz, 1H), 8.65 (s, 1H), 8.97 (d, J
) 8.8 Hz, 2H), 9.86 (s, 1H). 13C NMR (CDCl3): 112.95, 124.25,
124.85, 125.45, 126.00, 126.15, 126.20, 126.50, 127.15, 127.40, 127.70,
128.80, 129.10, 130.90, 131.00, 131.35, 134.05, 150.30, 159.95. MS
(electrospray): m/z 332.1 (100) M + H+.
Experimental Data for Compounds and Procedures. General
Aspects. All reagents and solvents were purchased at the highest
commercial quality and were used without further purification unless
otherwise noted. Deuterated chloroform used for the kinetic measure-
ments was flashed through alumina immediately prior to use, to remove
any trace of acid. Yields refer to purified spectroscopically (1H NMR)
(E)-N-(Anthracen-10-ylmethylene)cyclopentylamine (2). The prod-
uct was synthesized using the general method a and was obtained as a
yellow, fluffy powder (yield 92%). Mp 95-96 °C. Anal. Calcd for
1
homogeneous materials. H NMR spectra were recorded on a Bruker
Avance 400 spectrometer. The spectra were internally referenced to
9
5536 J. AM. CHEM. SOC. VOL. 127, NO. 15, 2005

Scandium(III) Catalysis of Transimination Reactions
A R T I C L E S
Figure 6. Schematic representation of the constitutional connectivities and their related thermodynamic (S) and kinetic (V0) parameters for: (left) the
individual transimination reactions starting from imine 1 and various amines and proceeding via a ternary species involving catalyst, amine, and imine (see
Scheme 3 and Table 1); (right) system interconnected through a common entity, the ternary species, hub (see, Figure 4). The relative size of the dots
represents the relative abundance of products after reaching the equilibrium; the relative thickness of the bonds represents the relative initial rates of the
reactions.
Scheme 8. Simplified Schematic Representation of Two Competing Pathways for the Transimination Reaction That Involves a Simple
Model Exchange between 1, Cyclopentylamine, and Anilinea
a For clarity, the steps representing proton transfers are not depicted as they are not rate-determining, and ScIII represents coordinated ScIII(L)x ions, with
L being amine, imine, and solvent molecules.
C20H19N: C, 87.87; H, 7.01; N, 5.12. Found: C, 87.28; H, 7.07; N,
(E)-N-(Anthracen-10-ylmethylene)(phenyl)methylamine (3). The
product was synthesized using the general method a and was obtained
as an orange solid (yield 72%). Mp 90-91°C. Anal. Calcd for
C22H17N: C, 89.46; H, 5.80; N, 4.74. Found: C, 89.69; H, 5.88; N,
1
5.18. H NMR (CDCl3): δ ) 1.80 (l, 2H), 2.00 (l, 4H), 2.13 (l, 2H),
3
3
4.15 (qt, J ) 5.9 Hz, 1H), 7.53 (m, 4H), 8.05 (d, J ) 8.3 Hz, 2H),
8.49 (d, 3J ) 8.3 Hz, 2H), 8.50 (s, 1H), 9.43 (s, 1H). 13C NMR
(CDCl3): 24.95, 34.90, 73.35, 124.85, 125.20, 126.45, 128.70, 128.85,
129.90, 131.35, 157.55, 162.00. MS (electrospray): m/z 274.15 (100)
M + H+.
1
4.57. H NMR (CDCl3): δ ) 5.19 (s, 2H), 7.39 (m, 1H), 7.46 (m,
2H), 7.54 (m, 6H), 8.06 (d, 3J ) 7.6 Hz, 2H), 8.54 (s, 1H), 8.56 (d, 3J
) 8.5 Hz, 2H), 9.6 (s, 1H). 13C NMR (CDCl3): 67.2, 125.6, 126.0,
9
J. AM. CHEM. SOC. VOL. 127, NO. 15, 2005 5537

A R T I C L E S
Giuseppone et al.
3
127.5, 128.0, 128.8, 129.0, 129.5, 129.7, 130.3, 130.9, 132.1, 139.2,
162.2. MS (electrospray): m/ z 296.146 (100) M + H+.
(E)-N-(Anthracen-10-ylmethylene)-4-methoxybenzenamine (4).
The product was synthesized using the general method a and was
obtained as a yellow powder (yield 84%). Mp 147-148 °C. Anal. Calcd
for C14H13NO: C, 84.86; H, 5.50; N, 4.50. Found: C, 84.66; H, 5.56;
8.2 Hz, 1H), 7.23 (d, J ) 7.0 Hz, 1H), 7.45-7.60 (m, 5H), 7.83 (d,
3J ) 7.0 Hz, 1H), 7.93 (dd, J ) 6.3 Hz and J ) 3.1 Hz, 1H), 7.93
(dd, 3J ) 5.8 Hz and 3J ) 3.1 Hz, 1H), 8.75 (s, 1H). 13C NMR
(CDCl3): 114.70, 118.05, 119.95, 120.25, 124.00, 126.70, 127.25,
127.45, 127.70, 128.70, 129.00, 133.20, 134.20, 134.80, 147.15, 162.20,
164.65. MS (electrospray): m/z 248.099 (100) M + H+.
3
3
1
3
N, 4.27. H NMR (CDCl3): δ ) 9.70 (s, 1H), 8.72 (d, J ) 8.8 Hz,
(E)-2-((Cyclopentylimino)methyl)phenol (11). The product was
synthesized using the general method b and was obtained as an orange
solid (yield 65%). Mp 34-35 °C. Anal. Calcd for C12H15NO: C, 76.16;
H, 7.99; N, 7.40. Found: C, 76.35; H, 7.99; N, 7.52. 1H NMR
(CDCl3): δ ) 1.73 (l, 4H), 1.88 (l, 2H), 1.96 (l, 2H), 3.80 (d, 3J ) 5.5
Hz, 1H), 6.89 (t, 3J ) 7.4 Hz, 1H), 6.98 (d, 3J ) 8.2 Hz, 1H), 7.26 (d,
2H), 8.55 (s, 1H), 8.06 (d, 3J ) 8.3 Hz, 2H), 7.54 (m, 4H), 7.45 (d, 3J
) 8.8 Hz, 2H), 7.04 (d, J ) 8.8 Hz, 2H). 13C NMR (CDCl3): 158.6,
3
157.7, 145.6, 131.4, 130.6, 130.3, 129.0, 127.7, 127.1, 125.4, 124.9,
122.3, 114.5, 55.6. MS (electrospray): m/z 312.134 (100) M + H+.
(E)-N-(Anthracen-10-ylmethylene)benzenamine (5). The product
was synthesized using the general method a and was obtained as yellow
needle-shaped crystals (yield 84%). Mp 111-112 °C. Anal. Calcd for
C21H15N: C, 89.65; H, 5.37; N, 4.98. Found: C, 89.71; H, 5.44; N,
3J ) 7.4 Hz, 1H), 6.89 (t, J ) 8.2 Hz, 1H), 8.37 (s, 1H). 13C NMR
3
(CDCl3): 24.50, 35.00, 70.45, 117.65, 119.15, 119.60, 131.80, 132.70,
162.20, 163.20. MS (electrospray): m/z 190.115 (100) M + H+.
(E)-4-((Naphthalen-1-ylimino)methyl)phenol (12). The product
was synthesized using the general method b and was obtained as a
beige, sandlike solid (yield 78%). Mp 190-191 °C. Anal. Calcd for
C17H13NO: C, 76.16; H, 7.99; N, 7.40. Found: C, 76.35; H, 7.99; N,
7.52. 1H NMR (CDCl3): δ ) 7.00 (d, 3J ) 8.3 Hz, 2H), 7.15 (d, 3J )
7.3 Hz, 1H), 7.55 (m, 3H), 7.77 (d, 3J ) 8.3 Hz, 1H), 7.92 (d, 3J ) 7.3
Hz, 1H), 7.95 (d, J ) 8.8 Hz, 2H), 8.36 (d, J ) 7.3 Hz, 1 H), 8.56
(s, 1H). 13C NMR (CD2Cl2): 112.60, 115.65, 123.85, 125.30, 125.50,
126.10, 126.80, 127.50, 129.75, 130.85, 133.95, 149.40, 157.20, 159.60,
159.65. MS (electrospray): m/z 262.118 (100) M + H+.
1
4.84. H NMR (CDCl3): δ ) 5.19 (s, 2H), 7.39 (m, 1H), 7.46 (m,
2H), 7.54 (m, 6H), 8.06 (d, 3J ) 7.6 Hz, 2H), 8.54 (s, 1H), 8.56 (d, 3J
) 8.5 Hz, 2H), 9.6 (s, 1H). 13C NMR (CDCl3): 121.80, 125.50, 126.20,
127.10, 128.05, 129.85, 130.15, 130.20, 131.45, 131.50, 132.15, 153.55,
160.85. MS (electrospray): m/z 282.12 (100) M + H+.
(E)-N-(Anthracen-10-ylmethylene)-3-nitrobenzenamine (6). The
product was synthesized using the general method a and was obtained
as an orange powder (yield 65%). Mp 181-182 °C. Anal. Calcd for
C21H15N: C, 77.29; H, 4.32; N, 8.58. Found: C, 76.78; H, 4.42; N,
8.21. 1H NMR (CDCl3): δ ) 9.77 (s, 1H), 8.81 (d, 3J ) 8.8 Hz, 2H),
8.63 (s, 1H), 8.25 (s, 1H), 8.18 (d, 3J ) 7.8 Hz, 1H), 8.08 (d, 3J ) 8.3
Hz, 2H), 7.73 (d, 3J ) 8.3 Hz, 2H), 7.61 (m, 3H), 7.63 (t, 3J ) 6.8 Hz,
2H). 13C NMR (CDCl3): 162.0, 153.7, 135.2, 131.3, 131.1, 130.9,
130.1, 129.8, 129.3, 129.1, 127.9, 127.7, 128.8, 125.7, 125.5, 124.4,
123.5, 120.7, 120.5, 115.5. MS (electrospray): m/z 327.1 (100) M +
H+.
3
3
(E)-4-((Cyclopentylimino)methyl)phenol (13). The product was
synthesized using the general method b and was obtained as a reddish-
orange powder (yield 41%). Mp 174-175 °C. Anal. Calcd for C17H13-
NO: C, 76.16; H, 7.99; N, 7.40. Found: C, 76.35; H, 7.99; N, 7.52.
1H NMR (CDCl3): δ ) 1.64-1.93 (m, 8H), 3.75 (qt, J ) 6.8 Hz,
3
3
3
1H), 6.70 (d, J ) 8.8 Hz, 2H), 7.52 (d, J ) 8.8 Hz, 2H), 8.22 (s,
1H). 13C NMR (CDCl3): 24.60, 34.15, 71.65, 115.75, 121.70, 130.10,
159.45, 160.00. MS (electrospray): m/z 190.12 (100) M + H+.
(E)-1-Phenyl-2-((E)-3-phenylallylidene)hydrazine (14). The prod-
uct was synthesized using the general method b and was obtained as a
yellow powder (yield 92%). Mp 175-176 °C. Anal. Calcd for
C15H14N2: C, 81.05; H, 6.35; N, 12.60. Found: C, 81.12; H, 6.39; N,
12.38. 1H NMR (CDCl3): δ ) 6.71 (d, 3J ) 16.1 Hz, 1H), 6.90 (t, 3J
) 7.4 Hz, 1H), 7.01-7.09 (m, 3H), 7.52 (d, 3J ) 8.8 Hz, 2H), 7.27-
(E)-N-(Anthracen-10-ylmethylene)-3,5-dinitrobenzenamine (7).
The product was synthesized using the general method a and was
obtained as an orange powder (yield 63%). Mp 263-264 °C. Anal.
Calcd for C21H15N: C, 67.92; H, 3.53; N, 11.32. Found: C, 67.92; H,
3.64; N, 11.32. 1H NMR (CDCl3): δ ) 7.61 (t, 3J ) 8.1 Hz, 2H), 7.71
3
3
(t, J ) 8.1 Hz, 2H), 8.14 (d, J ) 8.6 Hz, 2H), 8.55 (s, 2H), 8.72 (s,
1H), 8.93 (d, J ) 8.6 Hz, 2H), 9.00 (s, 1H), 9.88 (s, 1H). 13C NMR
3
(CDCl3): 115.35, 121.35, 124.05, 125.75, 128.55, 129.55, 131.25,
131.55, 133.60, 155.00, 163.75. MS (electrospray): m/z 372.1 (100)
M + H+.
3
3
7.31 (m, 2H), 7.37 (d, J ) 7.3 Hz, 2H), 7.49 (d, J ) 7.3 Hz, 2H),
7.58 (d, J ) 9.2 Hz, 2H), 7.59 (s, 1H). 13C NMR (CDCl3): 112.70,
3
120.20, 125.85, 126.55, 128.00, 128.75, 129.35, 134.10, 136.75, 139.80,
(E)-N-(Anthracen-10-ylmethylene)prop-2-en-1-amine (8). The
product was synthesized using the general method b and was obtained
as a yellow solid (yield 75%). Mp 67-68 °C. Anal. Calcd for
C18H15N: C, 88.13; H, 6.16; N, 5.71. Found: C, 88.09; H, 6.24; N,
5.63. 1H NMR (CDCl3): δ ) 4.63 (d, 3J ) 4.7 Hz, 2H), 5.32 (d, 3J )
144.25. MS (electrospray): m/z 223.119 (100) M + H+.
(E)-1-Methyl-1-phenyl-2-((E)-3-phenylallylidene)hydrazine (15).
The product was synthesized using the general method a and was
obtained as a yellow powder (yield 85%). Mp 112-113 °C. Anal. Calcd
for C16H16N2: C, 81.32; H, 6.82; N, 11.85. Found: C, 79.18; H, 6.71;
3
9.9 Hz, 1H), 5.44 (d, J ) 17.2 Hz, 1H), 6.32 (m, 1H), 7.55 (m, 4H),
1
8.06 (d, 3J ) 8.2 Hz, 2H), 8.51 (s, 1H), 8.52 (d, 3J ) 8.5 Hz, 2H), 9.50
(s, 1H). 13C NMR (CDCl3): 65.60, 117.25, 125.55, 126.00, 127.45,
129.65, 129.70, 130.15, 130.80, 132.10, 136.80, 162.15. MS (electro-
spray): m/z 246.132 (100) M + H+.
N, 11.32. H NMR (CDCl3): δ )7.40 (m, 3H), 7.32 (m, 6H), 7.12
3
3
(dd, J ) 8.8 and 15.6 Hz, 1H), 6.93 (m, 1H), 6.73 (d, J ) 15.6 Hz,
2H), 3.38 (s, 3H). 13C NMR (CDCl3): 147.6, 137.2, 132.8, 129.0, 128.7,
127.7, 127.6, 126.3. MS (electrospray): m/z 236.3 (100) M + H+.
(1E,2E)-Cinnamaldehyde O-Benzyl Oxime (16). The product was
synthesized using the general method a and was obtained as a reddish-
brown solid (yield 71%). Anal. Calcd for C16H15NO: C, 80.98; H, 6.37;
N, 5.90. Found: C, 79.89; H, 6.37; N, 3.34. 1H NMR (CDCl3) (mixture
of the s-trans and s-cis isomers with a 1:2 ratio): δ ) 5.20 and 5.24
(E)-N-(Anthracen-10-ylmethylene)-3-methylbutan-1-amine (9).
The product was synthesized using the general method b and was
obtained as an orange waxy solid (yield 70%). Mp 56.5-57.5 °C. Anal.
Calcd for C20H21N: C, 87.23; H, 7.69; N, 5.09. Found: C, 86.82; H,
7.66; N, 5.09. 1H NMR (CDCl3): δ ) 1.05 (d, 3J ) 6.3 Hz, 6H), 1.86
(m, 3H), 4.00 (t, 3J ) 7.0 Hz, 2H), 7.55 (m, 4H), 8.06 (d, 3J ) 8.2 Hz,
3
3
(2s, 2H), 6.77 (dd, AB system, J ) 15.6 Hz and J ) 7.5 Hz, 1H),
3
3
2H), 8.51 (s, 1H), 8.52 (d, J ) 8.5 Hz, 2H), 9.4 (s, 1H). 13C NMR
6.88 (m, AB system, 1H), 7.20-7.60 (m, 9H), 7.99 (d, J ) 8.6 Hz,
3
(CDCl3): 22.80, 26.30, 40.45, 61.80, 125.60, 126.00, 127.35, 129.35,
129.65, 129.85, 130.75, 132.10, 160.75. MS (electrospray): m/z 276.177
(100) M + H+.
1H), 8.60 (d, J ) 7.5 Hz, 1H). 13C NMR (CDCl3): 193.70, 152.80,
151.05, 138.60, 131.30, 129.10, 128.80, 128.55, 128.50, 126.90, 125.20,
121.90. MS (electrospray): m/z 238.12 (100) M + H+.
(E)-2-((Naphthalen-1-ylimino)methyl)phenol (10). The product
was synthesized using the general method b and was obtained as an
orange-brown powder (yield 77%). Mp 46-47 °C. Anal. Calcd for
C17H13NO: C, 82.57; H, 5.30; N, 5.66. Found: C, 82.66; H, 5.36; N,
5.88. 1H NMR (CD3CN): δ ) 7.02 (t, 3J ) 7.4 Hz, 1H), 7.15 (d, 3J )
Cyclohexanone O-Methyl Oxime (17). The product was synthesized
using the general method a and was obtained as a colorless liquid (yield
65%). Anal. Calcd for C7H13NO: C, 66.10; H, 10.30; N, 11.01.
Found: C, 66.20; H, 10.25; N, 11.04. 1H NMR (CDCl3): δ ) 1.62 (l,
4H), 1.69 (l, 2H), 2.21 (t, 3J ) 5.8 Hz, 2H), 2.46 (t, 3J ) 5.8 Hz, 2H),
9
5538 J. AM. CHEM. SOC. VOL. 127, NO. 15, 2005

Scandium(III) Catalysis of Transimination Reactions
A R T I C L E S
3.83 (s, 3H). 13C NMR (CDCl3): 25.15, 25.70, 25.80, 27.00, 32.15,
60.90, 160.20. MS (electrospray): m/z 128.1 (100) M + H+.
) 7.0 Hz, 2H), 1.88 and 1.90 (2s, 3H), 2.24 and 2.43 (2t, 3J ) 7.3 Hz,
2H), 2.69 (t, 3J ) 7.0 Hz, 2H), 5.08 and 5.09 (2s, 2H), 7.30-7.40 (m,
5H). 13C NMR (CDCl3): 14.32, 20.00, 26.80, 29.85, 30.30, 33.30,
41.75, 41.95, 75.70, 75.75, 128.35, 128.40, 128.65, 128.70, 129.05,
129.10, 139.20, 139.25, 158.80, 159.60. MS (electrospray): m/z 221.15
(100) M + H+.
Cyclohexanone O-Benzyl Oxime (18). The product was synthesized
using the general method a and was obtained as a slightly orange viscous
liquid (yield 73%). Anal. Calcd for C13H17NO: C, 76.81; H, 8.43; N,
6.89. Found: C, 76.54; H, 8.54; N, 6.35. 1H NMR (CDCl3): δ ) 1.62-
1.70 (l, 6H), 2.24 (m, 2H), 2.54 (m, 2H), 5.09 (s, 2H), 7.29-7.32 (m,
5H). 13C NMR (CDCl3): 55.85, 115.05, 122.95, 129.35, 129.50, 131.85,
137.30, 144.5, 159.40. MS (electrospray): m/z 204.13 (100) M + H+.
(E+Z)-5-Aminopentan-2-one O-Hydroxy(phenyl)methyl Oxime
(21). The product was synthesized using the general method a and was
obtained as an orange viscous oil (yield 75%). Anal. Calcd for
C13H20N2O: C, 69.87; H, 8.80; N, 13.58. Found: C, 70.10; H, 8.91;
Acknowledgment. This work was supported by a doctoral
fellowship from the Ministe`re de la Jeunesse, de la Recherche
et de la Technologie (J.-L.S.), and a Research Experience for
Undergraduate Students fellowship from the National Science
Foundation (E.S.).
1
N, 13.70. H NMR (CDCl3) (mixture of the E and Z isomers with a
1:2 ratio, respectively): δ ) 1.38 (l, 2H), 1.66 (tt, 3J ) 7.3 Hz and 3J
JA042469Q
9
J. AM. CHEM. SOC. VOL. 127, NO. 15, 2005 5539