The search for efficient intramolecular proton transfer from carbon: The kinetically silent intramolecular general base-catalysed elimination reaction of O-phenyl 8-dimethylamino-1-naphthaldoximes

Article abstract of DOI:10.1002/poc.858

The ready elimination of phenol/phenoxide from the O-phenyl oxime 10E derived from 8-dimethylamino-1-naphthaldehyde, necessarily involving proton transfer from carbon, is catalysed by the neighbouring NMe₂ group at pH > 9. However, reaction is

Full text of DOI:10.1002/poc.858

JOURNAL OF PHYSICAL ORGANIC CHEMISTRY
J. Phys. Org. Chem. 2005; 18: 101–109
Published online 21 September 2004 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/poc.858
The search for efficient intramolecular proton transfer
from carbon: the kinetically silent intramolecular general
base-catalysed elimination reaction of O-phenyl
8-dimethylamino-1-naphthaldoximesy
Nabil Asaad, John E. Davies, David R. W. Hodgson,z Anthony J. Kirby,* Liisa van Vliet and Laura Ottavi
University Chemical Laboratory, Cambridge CB2 1EW, UK
Received 28 November 2003; revised 19 February 2004; accepted 6 March 2004
ABSTRACT: The ready elimination of phenol/phenoxide from the O-phenyl oxime 10E derived from 8-dimethy-
lamino-1-naphthaldehyde, necessarily involving proton transfer from carbon, is catalysed by the neighbouring NMe2
group at pH > 9. However, reaction is faster, rather than slower, at lower pH. It is shown that the step involving proton
transfer is not cleanly rate determining at any pH: the preferred route involves syn/anti isomerization to form the more
reactive Z-isomer. The rate constant for the anti elimination cannot be extracted from the available data, so no reliable
estimate of effective molarity (EM) is possible. Copyright # 2004 John Wiley & Sons, Ltd.
KEYWORDS: proton transfer; intramolecular general base catalysis; elimination; oxime
INTRODUCTION
in rapid exchange with the local medium. However,
proton transfers can be slow, and therefore in need of
catalysis in biological systems, when concerted with the
making or breaking of bonds between heavy atom centres.
Examples are the general acid-catalysed hydrolysis of
acetals, e.g. 1,3 a simple model for the reaction catalysed
by the glycohydrolases such as lysozyme, and general
base-catalysed elimination reactions 2 (Scheme 1). A
special case of the latter process is enolization, 3, and
enzymes such as mandelate racemase, which catalyse the
most difficult (intrinsically slow) enolizations have been
the focus of much recent attention.5
Intramolecular reactions are typically faster (and cataly-
tic processes more efficient) than their intermolecular
counterparts,1 so they are of interest as simple models for
the reactions which take place when the same functional
groups are brought together in enzyme active sites.
Intramolecular nucleophilic additions—cyclizations—
can be very efficient indeed, with effective molarities
(EM) of the order of 109 M in unstrained systems, and as
high as 1013 M in systems in which ground strain is
relieved in the transition state for cyclization.1
Intramolecular proton transfers are different, typically
showing EMs < 10 M, with the important exception of
systems in which the product, and hence the transition
state leading to it, is stabilized by a strong intramolecular
hydrogen bond.2 These systems are therefore of special
interest.2–4
Proton transfer is the commonest reaction taking place
in enzymes, which generally catalyse reactions in aqueous
solution at more or less constant pH. Simple transfers
between electronegative centres are mostly diffusion
controlled, and unlikely to need catalysis in active sites
We have shown that the ketonization of the enol ether 4
is catalysed very efficiently by the neighbouring dimethy-
lammonium group (EM > 60 000).4 So efficient is the
proton transfer to and from carbon that the rate-determin-
ing step is not the proton transfer but the opening of the
intramolecular hydrogen bond of the oxocarbocation
intermediate 5 (the geometry of this hydrogen bond is
patterned on that of proton sponge6).
The logical next step was to examine the enolization of
the acids 6 (R ¼ H) and their anions, but we could find no
incorporation of deuterium even into the ester 6 (ꢀ-CH2,
R ¼ Et) under forcing conditions.7 The geometry is the
same as in the efficient system 4, so we concluded that
*Correspondence to: A. J. Kirby, University Chemical Laboratory,
Cambridge CB2 1EW, UK.
E-mail: ajk1@cam.ac.uk
ySelected article presented at the Seventh Latin American Conference
on Physical Organic Chemistry (CLAFQO-7), 21–26 September 2003,
´
Florianopolis, Brazil.
zPresent address: Department of Chemistry, University of Durham,
South Road, Durham DH1 3LE, UK.
Contract/grant sponsor: Engineering and Physical Sciences Research
Council of Great Britain.
Copyright # 2004 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2005; 18: 101–109

102
N. ASAAD ET AL.
Scheme 2
RESULTS AND DISCUSSION
The elimination of phenol (or phenoxide) from 10E to
form the nitrile 11 (Scheme 2) was conveniently studied
[in 50% (v/v) acetonitrile–water, for solubility reasons] at
60 ꢀC and ionic strength 0.1 M (KCl). Since system 10
was designed to support an efficient intramolecular gen-
eral base-catalysed proton transfer from carbon, it was
expected to react in the free base form, thus at pHs above
the pKa of the dimethylammonium group. In practice, the
pH–rate profile for the elimination (Fig. 1) shows a
reaction that is slower at high pH. Reaction is slow also
at low pH, as expected if rapid elimination depends on the
presence of the dimethylamino group in its free base
form. However, the dimethylamino group is also (pre-
sumably) protonated in the pH-independent region be-
tween pH 2 and 8: only the higher of the two apparent
pKas (1.26 and 8.61) derived from the pH–rate profile can
reasonably be assigned to the dimethylammonium group
(see below). Hence, the mechanism appears not to be a
simple elimination process.
Scheme 1
exchange is not observed because proton return from the
enolate intermediate 7, to regenerate the starting material,
is much faster than the opening of the intramolecular
hydrogen bond needed for exchange of the NHþ with
solvent.7
In a further attempt to observe efficient intramolecular
proton transfer using this basic system, we prepared the
vinyl bromide 8 (D ¼ H, and the labelled form shown),8
designed to support the efficient elimination of HBr.
However, reaction gave not the alkyne but the heterocycle
9. The mechanism probably involves an initial nucleo-
philic addition of the dimethylamino group to form a five-
membered ring: in any event, the target hydron is still
present in the product, so the desired proton transfer has
not occurred.8
We report our first successful attempt to follow an
elimination reaction in a system based on proton sponge.
8-Dimethylamino-1-naphthaldehyde is the common in-
termediate in the synthesis of a number of the compounds
described above, so it was a relatively simple matter to
prepare oxime derivatives (general formula 10*, geome-
try may be Z or E). Of three compounds prepared
(obtained as the E-isomers, no doubt for steric reasons),
the acetate was inconveniently reactive and the methoxy
derivative was hydrolysed faster than it eliminated
methanol. However, the O-phenyl derivative 10E was
converted quantitatively to the nitrile 11 within hours in
aqueous solution at 30 ꢀC (Scheme 2).
Figure 1. pH–rate profiles for the elimination reactions
of 10E, (filled circles) and d-10E (open circles), to form 11:
in 50% (v/v) acetonitrile–H2O at 60ꢀC and ionic strength
0.1M (KCl). The curves represent calculated fits based
on the mechanism discussed below
Copyright # 2004 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2005; 18: 101–109

EFFICIENT INTRAMOLECULAR PROTON TRANSFER FROM CARBON
103
Reaction mechanism
the eliminations of pivalate9 and benzyloxide10 from the
corresponding (E)-O-acyl (alkyl) oximes of benzalde-
hyde catalysed by DBU, show kH/kD ¼ 2.7 and 3.3,
respectively. For the corresponding Z-isomers, which
react 20 000 (36 000) times faster, kH/kD ¼ 7.8 and 7.3,
respectively.9,10
In the plateau region between pH 2 and 8, we observe
a substantial deuterium isotope effect (kH/kD ¼ 3.08 ꢁ
0.11). This is consistent with rate-determining elimina-
tion from 10E; however, this is unlikely to be the correct
explanation, because reaction is faster, rather than slower,
than that of the free base form above pH 9. The elimina-
tion step catalysed by the dimethylamino group of the
free base 10E is expected to be faster than that of its con-
jugate acid, but overall elimination is slower at pH > 9,
because the proton transfer step is not rate determining in
this pH region.
Proton transfers to and from carbon typically show
substantial and informative primary deuterium isotope
effects, so we also measured the pH–rate profile for
elimination from the deuterio-compound d-10E (see
Fig. 1 and Table 1). The observed kinetic deuterium
isotope effects tell us that proton transfer is not rate
determining for the elimination reaction of 10E, and
probably not exclusively rate determining for the faster
reaction of the conjugate acid, 10EHþ.
The plateau reaction
In the high pH region (pH > 9), there is no significant
primary deuterium isotope effect (kH/kD ¼ 1.13 ꢁ 0.04).
For comparison, the most closely analogous reactions,
The O-phenyl oxime 10E has three possible conjugate
acids. Apart from the predominant dimethylammonium
Table 1. Rate data for the conversion of oxime 10E to the nitrile 11, in 50% (v/v) acetonitrile–water at 60 ꢀC and ionic strength
1.0 M (data in sꢂ1
)
Reaction of 10E
Reaction of d-10E
pH
Log kobs
pH
Log kobs
pD
Log kobs
1.2100
1.2800
1.2800
1.3600
1.9500
2.2000
2.2000
2.2000
2.3100
2.6000
2.9800
2.9800
2.9800
3.8200
3.8200
3.8200
3.9700
3.9700
4.4500
4.4500
4.4500
4.4600
4.9600
5.2500
5.2500
5.2500
5.6100
5.6100
5.6100
ꢂ3.5705
ꢂ3.4189
ꢂ3.4265
ꢂ3.3978
ꢂ3.1833
ꢂ3.1703
ꢂ3.1737
ꢂ3.1571
ꢂ3.1438
ꢂ3.1168
ꢂ3.1416
ꢂ3.1455
ꢂ3.1233
ꢂ3.1523
ꢂ3.1352
ꢂ3.1337
ꢂ3.1734
ꢂ3.1591
ꢂ3.1451
ꢂ3.1466
ꢂ3.1417
ꢂ3.1650
ꢂ3.1950
ꢂ3.1653
ꢂ3.1576
ꢂ3.1559
ꢂ3.1639
ꢂ3.1751
ꢂ3.1659
5.8200
5.8200
7.4200
7.4200
7.4200
7.9600
7.9600
7.9600
10.730
10.980
10.980
10.980
11.090
11.110
11.110
11.780
11.950
11.950
11.950
12.260
12.260
12.260
12.920
12.920
12.920
13.460
13.760
13.760
ꢂ3.1651
ꢂ3.2075
ꢂ3.2576
ꢂ3.2740
ꢂ3.2548
ꢂ3.3021
ꢂ3.2892
ꢂ3.2828
ꢂ4.0150
ꢂ4.0230
ꢂ4.0118
ꢂ4.0148
ꢂ4.0189
ꢂ4.0112
ꢂ4.0179
ꢂ4.0310
ꢂ4.0198
ꢂ4.0113
ꢂ4.0249
ꢂ4.0243
ꢂ4.0272
ꢂ4.0197
ꢂ4.0046
ꢂ4.0105
ꢂ4.0020
ꢂ4.0101
ꢂ3.9966
ꢂ4.0064
1.2800
1.2800
2.2000
2.2000
4.4500
4.4500
3.8200
3.8200
3.8200
10.920
7.9600
7.9600
7.9600
12.260
13.770
13.770
13.770
13.770
10.400
10.400
10.400
13.770
13.770
13.770
13.770
12.750
12.750
12.750
11.530
11.530
5.2300
5.2300
5.2300
ꢂ3.9593
ꢂ3.9494
ꢂ3.6926
ꢂ3.6854
ꢂ3.6598
ꢂ3.6708
ꢂ3.6861
ꢂ3.6702
ꢂ3.6609
ꢂ4.0462
ꢂ3.6636
ꢂ3.6766
ꢂ3.6655
ꢂ4.0622
ꢂ4.0814
ꢂ4.0866
ꢂ4.0927
ꢂ4.0952
ꢂ4.0735
ꢂ4.0665
ꢂ4.0736
ꢂ4.0814
ꢂ4.0866
ꢂ4.0927
ꢂ4.0952
ꢂ4.0647
ꢂ4.0625
ꢂ4.0519
ꢂ4.0526
ꢂ4.0503
ꢂ3.6638
ꢂ3.6626
ꢂ3.6687
Copyright # 2004 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2005; 18: 101–109

104
N. ASAAD ET AL.
Scheme 3
form 10EHþ, the oxime group can be protonated, on
either nitrogen orþoxygen (Scheme 3). The N-protonated
—
form 10Eꢃ NH might be expected to be a relatively
—
strong acid (the pKa of the oxime derived from 9-
formylfluorene is ꢂ1.6211) and the O-protonated form
10E ꢃ OHþ a very strong acid, with pKa < ꢂ7.12 If the
pKa of the major species 10EHþ is 8.61, the equilibrium
constant for the formation of 10E ꢃ OHþ will be <10ꢂ15
,
requiring a first-order rate constant for the elimination of
phenol of the order of 1012 sꢂ1 from a species existing for
only a few vibrations. If not formally impossible, such
a rate constant is highly unlikely (particularly for a
syn-elimination, which might be expected to be signifi-
cantly slower than the corresponding trans-elimination
reaction).
A much more likely initial reaction for the conjugate
acid 10EHþ, consistent with the known behaviour of
peri-disubstituted systems,13 and explaining why it
reacts faster than the basic form, is cyclisation of the N-
Scheme 4
þ
—
protonated oxime 10Eꢃ NH . This will generate 12
—
(Scheme 4), making possible free rotation about what
—
was the C N bond of the oxime. Compound 12 can
—
Two parameters are of particular interest, the apparent
pKas of 1.26 and 8.61, and the primary kinetic isotope
effects in the two pH-independent regions.
þ
—
open either to regenerate 10Eꢃ NH orþto give the Z-
—
—
isomer 10Z of the oxime (via 10Zꢃ NH ). This opens
—
the way to a new, low-energy pathway because trans-
elimination from the Z-isomers of oximes derived from
aromatic aldehydes is known to be much faster than syn-
elimination from the E-isomers, as discussed above.
Hence we consider that the reaction between pH 2 and
8 goes through 12 (Scheme 4). The reaction of the neutral
species 10E is presumed to go by a similar mechanism
via the zwitterionic intermediate 12 ꢁ . The reaction of
pKas
We have not been able to measure pKas convincingly for
10EHþ: it is rapidly degraded when titrated with base,
and the changes in the UV spectrum with pH are not
simple. (There is no 1H NMR evidence for the formation
of significant amounts of a third species under the
reaction conditions, either for 10E of for the more stable
O-methyl oxime.) The apparent pKas (1.26 and 8.61)
are very different from those expected for the indivi-
dual dimethylamino and oxime groups of 10E, so they
must represent either kinetic pKas or values shifted by
intramolecular interactions between the groups. Such
intermolecular interactions are well known in 1,8-disub-
stituted naphthalenes,14 and anomalous pKas generally
result from the formation of strong hydrogen bonds
between electronegative peri-substituents. In the case of
10EHþ the strongest intramolecular hydrogen bond pos-
sible is between the Me2NHþ group and the oxime N, and
this is observed (not unexpectedly) in the crystal structure
the conjugate acidþ10EHþ is faster because the cyclisa-
—
—
tion of 10Eꢃ NH is faster than that of 10E.
Scheme 4 provides a qualitative framework for the
more quantitative discussion of the mechanism. {Note: A
referee suggests that the path to the nitrile (Scheme 4)
could involve a Lossen-type rearrangement of 12 ꢁ , with
hydride migration concerted with the departþure of phen-
—
oxide (generating initially the [HN C—N Me ] inter-
—
2
mediate). We considered8 the corresponding mechanism
for the rearrangement of 8, which involved the departure
of Brꢂ, a much better leaving group, and required the
NMe2 group rather than hydride to migrate. Similar
considerations apply to the rearrangement of 12 ꢁ , but
the alternative mechanism cannot formally be ruled out.}
Copyright # 2004 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2005; 18: 101–109

EFFICIENT INTRAMOLECULAR PROTON TRANSFER FROM CARBON
105
reaction involves rate-determining anti-syn isomerization
of 10E to 10Z, via 12 ꢁ , followed by the rapid elimina-
tion of phenoxide from the less abundant but far more
reactive neutral isomer 10Z. No significant primary
isotope effect is observed because the C—H bond is
not broken in the rate-determining step. At lower pHs (on
the plateau between pH 2 and 8) cyclization, leading to
the (reversible) isomerization of 10EHþ, is faster, and
rapid elimination from 10Z (k2) competes with its recy-
clization, via 10ZHþ. (Apart from undergoing elimina-
tion much more rapidly 10Z is also expected to be
substantially less basic than 10E because it is unable,
for geometric reasons, to form the intramolecular hydro-
gen bond that we consider responsible for the unusually
high pKa of 10EHþ. We have no direct evidence on this
point.) We suggest that the observed primary kinetic
isotope effect (3.08 ꢁ 0.11) reflects a larger figure (of
around 7) for the anti elimination of 10Z, reduced by the
negligible isotope effect expected for its (partially rate-
determining) recyclization to 12, which must go at a
comparable rate (via 10ZHþ).
Figure 2. (a) Molecular structure of 13EHþ (crystallized as
the tetrafluoroborate salt) and (b) the parent oxime 14 (free
base form). ORTEP: ellipsoids are drawn at the 30% prob-
ability level and hydrogen atoms are represented as spheres
˚
of radius 0.12 A. For details, see the Experimental section
of the tetrafluoroborate salt of the O-methyl oxime
13EHþ [Fig. 2(a)].
The N–Hꢃ ꢃ ꢃN angle is 164.8ꢀ and the Nꢃ ꢃ ꢃN distance
ꢀ
˚
˚
is 2.628 A (at 150 K), close to those [160(3) and 2.571 A
at 120 K] of the very strong hydrogen bond of the
conjugate acid of proton sponge [1,8-bis(dimethylami-
no)naphthalene].15 However, this is achieved in 13EHþ
only at the expense of significant distortion of the
naphthalene framework: the torsion angle C(1)—
C(9)—C(8)—N(2) is over 20ꢀ (the conjugate acid of
proton sponge is almost planar) and the bond angles
C(9)—C(1)—N(1) and C(9)—C(8)—N(2) are ex-
panded to 128.9ꢀ and 124.5ꢀ, respectively, to accommo-
date the proton. {The Nꢃ ꢃ ꢃN distance of closest approach
in the relevant conformer of the oxime [Fig. 2(b)] would
1
We see no evidence, from H NMR measurements in
D2O (at room temperature) under the reaction conditions,
that 12 is formed in significant concentrations in the
plateau region. The parent aldehyde 14 is largely (2:1
1
mixture observed by H NMR) converted to the cyclic
aminal 15 in CDCl3 in the presence of 10% CF3COOD
(Scheme 5);13 as shown by the reduced intensity of the
—
oxime N CH singlet, the appearance of a new
—
(methine) proton signal at ꢁ 7.0, and the doubling and
downfield shift of the N-methyl proton signal.
None of these NMR indicators is observed for 10E or
for the O-methyl oxime in the plateau region. The oxime
singlet at ꢁ 9.46 does disappear, essentially completely,
in 0.1 M DCl, to be replaced by a broad signal at ꢁ 8.13
(and weaker signals at ꢁ 5.1–5.3, initially obscured by the
HDO peak). This intriguing result is most likely only
indirectly relevant to the mechanism at higher pH,
although we know that the species present is relatively
unreactive.
The calculated curves shown in Fig. 1 represent the fits
derived on the basis of Scheme 4. The data in the plateau
region (pH 1–9) were fitted to the simplified Eqn (1)
based on Scheme 6, using the steady-state assumption.
We assume that all proton transfers between O and N
centres are fast, and that k1, the opening of the inter-
mediate 12 (Scheme 4) to give the thermodynamically
less stable isomer of the oxime, is the rate-determining
˚
be an implausible 1.89 A.} As a result, the thermody-
namic stability of the H-bond, and hence the effect on the
pKa of 10EHþ, is expected to be reduced.
The O-methyl oxime 13E is considerably more stable
than the O-phenyl derivative 10E, and so offers increas-
ed scope for physical measurements. Unfortunately, it
behaves differently under the conditions: it is not suffi-
ciently soluble in 50% aqueous acetonitrile, so measure-
ments were made in 10% dioxane. However, in this
solvent it did not undergo the elimination of interest:
hydrolysis of the oxime group to the parent aldehyde is
faster. Titration at low pH reveals a pKa of 3.22 ꢁ 0.07 (in
10% dioxane at room temperature), falling to 2.34 ꢁ 0.16
in 50% acetonitrile (adding 0.5 equiv. of HCl to a solution
of 13EHþ gave a solution of pH ca 2.6). This is consistent
with a pKa of 1.26 at 60 ꢀC for the O-phenyl derivative.
This evidence suggests that the lower of the two apparent
pKas of 10EHþ is a true pKa; we have no conclusive
evidence that this is the case for the higher pKa at 8.61.
Mechanism of the elimination reaction of 10
A suggested mechanism is outlined in Scheme 4. At high
pH, above the pKa of the dimethylammonium group,
Scheme 5
Copyright # 2004 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2005; 18: 101–109

106
N. ASAAD ET AL.
Scheme 6
step for the isomerization process. Then 10E is in rapid
pre-equilibrium with 12 (K12 is the equilibrium constant
for the formation of 12 from 10EHþ), and we can write
(for the plateau reaction):
are observed), (iii) that the proton transfer step is not
cleanly rate determining at any pH and (iv) that the
preferred route involves isomerisation to the more reac-
tive Z-isomer. The rate constant for the anti elimination
cannot be extracted from the available data, so no reliable
estimate of effective molarity is possible.
kobs ¼ k1K12½10EHþꢄ
ð1Þ
þ
ꢂꢂ
ꢂꢂ
ꢅ k2½10Zꢄ=ðkꢂ1½10Z ꢃ
NH ꢄ þ k ½10ZꢄÞ
2
EXPERIMENTAL
Syntheses
The pH-independent reaction above pH 9 is allowed for
by the addition of a separate term.
The same procedure gives a good fit for the reaction of
d-10E. The rate and equilibrium constants derived from
the fits of the data for 10E and its deuterated form d-10E
All solvents were dried by standard procedures and
distilled and all glassware was oven-dried before use.
Potentially moisture-sensitive reactions were carried out
under an atmosphere of argon. Analytical thin-layer
chromatography was performed on precoated 0.25 mm
Merck Kieselgel 60F254 plates. Flash column chromato-
graphy was carried out using Merck Kieselgel 60 (230–
400 mesh). Melting-points were measured on a Stuart
Scientific SMP1 melting-point apparatus and are uncor-
rected.
are given in Table 2. The derived rate constants k1, k
ꢂ1
and k2 give errors larger than the figures quoted and do
not represent unique solutions, although relative values
should be meaningful. The derived ‘pKas’ match closely
the apparent pKas, except that the values for 10E and 10Z
are reversed. The resultant deuterium isotope effects fall
within ꢁ 20% of unity for k1, for k
and for the rate
ꢂ 1
constant k0 at high pH, but for k2, the elimination reaction
of the reactive isomer 10Z kH/kD ¼ 7.9. This is the value
expected for this process according to the closest pre-
cedent available (see above).10
NMR spectra were recorded on Bruker DPX400,
DRX400 and DRX500 FT spectrometers at probe tem-
peratures of 27 ꢀC unless stated otherwise. The solvent
signal was used as internal deuterium lock. Chemical
shifts are quoted downfield from ꢁ (tetramethylsila-
1
ne) ¼ 0 ppm for H and 13C NMR. The multiplicity of
CONCLUSIONS
first-order signals is indicated by standard notation.
Broadband proton decoupling was used for 13C NMR.
Infrared spectra were recorded on Perkin-Elmer 1600 FT-
IR and Spectrum One FT-IR spectrometers. Mass spectra
were recorded on Kratos MS890 (EI), Kratos FAB
MS890 (FAB), MSI Concept IH (EI and LSIMS), Bruker
Bio-apex II FT-ICR (ESI, LSIMS and EI) and ESI
Micromass Q-TOF (þES) mass spectrometers.
The mechanism of the elimination of phenol from oxime
10E is not characterized down to the last detail, but we
have established (i) that reaction is subject to the ex-
pected catalysis by the neighbouring dimethylamino
group, (ii) that syn elimination from the E-isomer is not
observed (and is therefore slower than the reactions that
Table 2. Derived rate and equilibrium constants for the
elimination reactions of 10E and its deuterated form in
50% (v/v) acetonitrile–water at 60 ꢀC
[10-D]-8-N,N-Dimethylamino-1-naphthaldehyde. N,N- Di
methylamino-1-naphthylamine (14, 3.0 g, 17.5 mmol)
was dissolved in diethyl ether (180 cm3) and cooled to
ꢂ78 ꢀC. n-Butyllithium (15 cm3, 2.3 M solution in hex-
ane, 2 equiv.) was added dropwise and the mixture stirred
for 30 h, warming to room temperature. A white precipi-
tate formed from the yellow solution. The mixture was
again cooled to ꢂ78 ꢀC and a solution of [1-D]-dimethyl-
formamide (1.5 cm3, 1.1 equiv.), was cannulated in at
ꢂ78 ꢀC. Stirring was continued for a further 3 h. Metha-
nol (20 cm3) was added and the mixture allowed to warm
Parameter
10E
d-10E
kH/kD
k1K12
k2
6540
4440
6900
562
0.95
7.9
k
0.0546
1.26 ꢁ 0.04
8.61
0.0556
1.25 ꢁ 0.04
8.25
0.79
0.98
0.44
ꢂ 1
pKa (EHþ)
pKa (ZHþ)
k0 (pH > 9) (9.66 ꢁ 0.15) ꢅ 10ꢂ5 (8.43 ꢁ 0.10) ꢅ 10ꢂ5 1.14
R
0.997
0.996
Copyright # 2004 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2005; 18: 101–109

EFFICIENT INTRAMOLECULAR PROTON TRANSFER FROM CARBON
107
to room temperature. Distilled water (80 cm3) was added
and the solution basified (KOH). The layers were shaken
and separated, then the aqueous layer extracted with
hexane (2 ꢅ 50 cm3). The combined organic layers were
washed (brine), dried (MgSO4) and evaporated to near
dryness. A small volume of toluene was added and
the crude product was stored in a freezer to yield the
aldehyde as transparent pale-yellow plates (1.21 g). The
mother liquor was chromatographed (SiO2, Et2O–hex-
ane, 1:3) to yield further aldehyde as a yellow powder
(0.98 g, 62% overall), m.p. 80–82 ꢀC; RF (Et2O–hexane,
1:3) 0.13; NMR, ꢁ H (CDCl3) 8.22 (1 H, d, J 7, ArH), 7.89
(1 H, dd, J 8 and 1, ArH), 7.63 (1 H, dd, J 8 and 1, ArH),
7.54 (1 H, dd, J 7 and 1, ArH), 7.51–7.45 (1H, m, ArH),
—
(C C N), 131.0, 129.7 (C —H), 128.9, 128.7,
—
Ar Ph
127.9 (CH), 126.6 (CH), 126.0 (CH), 124.8 (CH), 122.2
(CH), 117.6 (CH), 115.0 (CPh—H), 45.6 (CH3—N).
O-Methyl-8-N,N-dimethylamino-1-naphthaldoxime (13E).
N,N-Dimethylamino-1-naphthaldehyde (292 mg, 1.46 mmol),
O-methylhydroxylamine hydrochloride (240 mg, 2 equiv.)
and sodium acetate (480 mg, 4 equiv.) were dissolved in
methanol (20 cm3) and the mixture was stirred at room
temperature for 1 h. Brine (5 cm3) and distilled water
(10 cm3) were added and the mixture was extracted with
hexane (3 ꢅ 30 cm3). The combined organic extracts were
washed with brine (30 cm3), dried (MgSO4) and evaporated
under vacuum. The crude product was chromatographed
(SiO2, Et2O–hexane, 1:3) and freeze-drying yielded the
oxime ether as pale-yellow needles (290 mg, 87%), RF
7.31 (1 H, dd, J 7 and 1, ArH) and 2.68 (6H, s, NMe2);
—
ꢁ C (CDCl ) 191.4 (t, J 29, C O), 150.3, 137.2, 135.0,
—
130.9 (CH), 129.3, 126.7 (CH), 125.9 (CH), 125.0 (CH),
3
(Et2O–hexane, 1:3) 0.29, which decomposed on heating.
—
124.7 (CH), 118.2 (CH), 44.9 (NMe2); IR (Nujol),
—
NMR, ꢁ H (CDCl ) 9.21 (1 H, s, CH N), 7.81(1 H, dd,
—
3
ꢂmax 1649 (C O); MS, m/z (þES) 201.11310
—
J 7 and 1, ArH), 7.64 (1 H, dd, J 7 and 1, ArH), 7.56 (1 H,
dd, J 8 and 1, ArH), 7.43 (1 H, t, J 8, ArH), 7.40 (1 H, t,
J 8, ArH), 7.20 (1 H, dd, J 8 and 1, ArH), 4.00 (3 H, s,
(C13H12DNO ꢃ Hþ requires 201.11302).
O-Phenyl-8-N,N-dimethylamino-1-naphthaldoxime (10E).
N,N-Dimethylamino-1-naphthaldehyde (48.8 mg, 0.24
mmol), O-phenylhydroxylamine hydrochloride (64 mg,
2 equiv.) and sodium acetate (75 mg, 4 equiv.) were
dissolved in methanol (15 cm3) and the mixture was
stirred at room temperature for 15 min. The reaction
mixture was extracted with hexane (3 ꢅ 30 cm3) and the
combined hexane layers were washed with brine
(30 cm3), dried (MgSO4) and evaporated under vacuum,
maintaining the temperature below 40 ꢀC. The crude
product was chromatographed (SiO2, Et2O-hexane, 1:3)
and freeze-drying yielded the oxime ether as pale-yellow
needles (54.1 mg, 76%), RF (Et2O–hexane, 1:3) 0.35;
which decomposed on heating. NMR, ꢁ H (CDCl3) 9.56
—
—
OMe) and 2.68 (6H, s, NMe2); ꢁ C (CDCl ) 150.8 (C
3
—
—
N), 150.1 (C —N), 134.6 (C—C N), 128.9 (CH),
Ar
127.8, 127.0, 125.7 (CH), 124.8 (CH), 124.3 (CH), 123.0
(CH), 115.6 (CH), 60.4 (CH3—O) and 43.8 (CH3—N);
IR (Nujol), ꢂmax 1597, 1577; MS, m/z (þEI) 228.1264
(C14H16N2O requires 228.1263).
Tetrafluoroborate salt of O-methyl-8-N,N-dimethylamino-
1-naphthaldoxime (13EHþ). O-Methyl-8-N,N-dimethyla-
mino-1-naphthaldoxime (20 mg) was dissolved in diethyl
ether (1 cm3). Hydrogen tetrafluoroborate [ꢆ0.1 cm3,
54% (w/w) in diethyl ether] was added dropwise until
precipitation ceased. The liquid was decanted off and the
white precipitate dissolved in methanol (ꢆ0.2 cm3). A
layer of diethyl ether (ꢆ1 cm3) was carefully deposited
on top of the methanol layer and the solution allowed to
crystallize as colourless plates. The compound decom-
posed on heating.
—
(1 H, s, CH N), 7.88 (1 H, dd, J 8.0 and 1, ArH), 7.73
—
(1 H, dt, J 7 and 1, ArH), 7.58 (1 H, dd, J 7 and 1, ArH),
7.46 (1 H, t, J 7, ArH), 7.42 (1 H, t, J 7, ArH), 7.37–7.28
(4 H, m, ArH), 7.25 (1 H, m, ArH), 7.04 (1 H, tt, J 7 and 1,
ArH), 2.72 (6H, s, NMe2); ꢁ C (CDCl3) 160.3 (CPh—O),
—
—
155.4 (C N), 151.6 (C —N), 136.1 (C C N), 130.9
Ar
—
—
Ar
(CH), 129.7 (CPh—H), 128.9, 128.7, 127.9 (CH), 126.6
(CH), 126.0 (CH), 124.8 (CH), 122.2 (CH), 117.6 (CH),
115.0 (CPh—H) and 45.6 (CH3—N); IR (Nujol), ꢂmax
1605, 1583; MS, m/z (þEI) 290.1416 (C19H18N2O re-
quires 290.1419).
8-Cyano-N,N-dimethylamino-1-naphthylamine (11). 8-N,N-
Dimethylamino-1-naphthylaldoxime (60 mg, 0.28 mmol)
was dissolved in dichloromethane (2 cm3). Pyridine
(0.1 cm3, 5 equiv.) and freshly distilled acetic anhydride
(0.5 cm3, 5 equiv.) were added and the mixture was
heated at 50 ꢀC for 2 h. The reaction mixture was washed
with dilute hydrochloric acid (3 ꢅ 5 cm3) and saturated
sodium hydrogencarbonate solution (3 ꢅ 5 cm3) and
dried (MgSO4). The solvent was removed under vacuum
to yield the nitrile as a yellow solid (47 mg, 85%), m.p.
75–76 ꢀC.16 NMR, ꢁ H (CDCl3) 8.01 (1 H, d, J 8, ArH),
7.96 (1 H, d, J 7, ArH), 7.60 (1 H, d, J 8, ArH), 7.52–7.45
(2 H, m, ArH), 7.33 (1 H, d, J 7, ArH) and 2.82 (6H, s,
NMe2); ꢁ C (CDCl3) 151.4, 137.0 (CH), 136.1, 134.3
(CH), 130.1, 128.3 (CH), 126.0 (CH), 125.4 (CH),
121.5, 119.8 (CH), 109.1 (C—N) and 46.6 (NMe2); IR,
[10-D]-O-Phenyl-8-N,N-dimethylamino-1-naphthaldoxime
(d-10E). This was prepared by the above method using
[10-D]-N,N-dimethylamino-1-naphthaldehyde (49.1 mg,
24.4 mmol) to yield the oxime ether as pale-yellow
needles (61.7 mg, 87%); NMR, ꢁ H (CDCl3) 7.87 (1 H,
dd, J 8.0 and 1, ArH), 7.75 (1 H, dt, J 7 and 1, ArH), 7.58
(1 H, dd, J 7 and 1, ArH), 7.48 (1 H, t, J 7, ArH), 7.43 (1
H, t, J 7, ArH), 7.38–7.29 (4 H, m, ArH), 7.25 (1 H, m,
ArH), 7.04 (1 H, m, ArH), 2.72 (6H, s, NMe2); ꢁ C
(CDCl3) 160.3 (CPh—O), 151.6 (CAr—N), 136.8
Copyright # 2004 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2005; 18: 101–109

108
N. ASAAD ET AL.
ꢂmax (CH2Cl2) 1612, 1590, 1574; MS, m/z (EI) 196.0999
(C13H12N2 requires 196.1000).
cuvette (1.0 cm pathlength) for each kinetic run. Repeti-
tive wavelength scans were carried out for each kinetic
substrate in a series of buffer solutions to determine
whether the reaction exhibited one or more isosbestic
points and to allow the selection of an appropriate
wavelength at which to record absorbance–time data.
Experimental data were fitted to the first-order rate law
equation using Kaleidagraph v. 3.08 (Synergy Software),
assuming that the observed change in absorbance was the
result of conversion of reactant to product; accuracies
quoted are ꢁ 1 standard deviation. No buffer catalysis
was observed in this work. Multi-parameter curve fitting
for the plateau region data was based on Eqn (1) and used
the Kaleidagraph implementation of the Levenberg–
Marquardt algorithm.
Kinetic measurements
All buffer reagents were of analytical-reagent grade.
Deionized water was doubly distilled in all-glass appa-
ratus and degassed with argon. KOH and HCl stock
solutions (2 mol dmꢂ3) were made by dilution of BDH
Convol concentrates. Buffer solutions were made by
appropriate dilution in grade A volumetric flasks: the
appropriate quantity of KCl was added to adjust the ionic
strength to 0.10 M. Dioxane was freshly distilled from
NaBH4 prior to use; the cosolvent content is quoted as
percentage (v/v). The reaction was followed at 60 ꢀC in
50% aqueous MeCN using HCl, KOH and formate,
acetate, phosphate, Tris and carbonate buffers. The pH
of each buffer solution was recorded at the temperature
used in the kinetic investigation, using a Radiometer
PHM82 pH meter fitted with a Russell CTWL electrode
calibrated to standard buffer solutions. Rate data are
given in Table 1.
Crystal structure determinations
Selected bond lengths and angles, relevant to the inter-
acting peri-substituent groups, are given in Tables 3 and 4.
Crystal data for 13Hþ. C14H17BF4N2O, M ¼ 316.11,
triclinic, space group P-1 (No. 2), a ¼ 7.5686(6),
˚
UV–visible spectroscopic data were recorded using a
Varian Cary 3 spectrometer fitted with a thermostated cell
holder maintained at the temperature stated. Stock solu-
tions (ꢆ5 ꢅ 10ꢂ3 mol dmꢂ3) of kinetic substrates were
prepared in acetonitrile. The stock solution (10 ml) was
added to preheated buffer solution (2 cm3) in a quartz
b ¼ 8.1133(6), c ¼ 12.2631(6) A, ꢀ ¼ 98.757(5), ꢃ ¼
3
ꢀ
˚
92.343(4), ꢄ ¼ 99.341(3) , U ¼ 732.72(9) A , Z ¼ 2,
ꢅ(Mo Kꢀ) ¼ 0.124 mmꢂ1, 6082 reflections measured at
180(2) K using an Oxford Cryosystems Cryostream
cooling apparatus, 2509 unique (Rint ¼ 0.043); R1 ¼
0.059, wR2 ¼ 0.151 [I > 2_(I)]; goodness-of-fit on F2,
Table 3. Selected bond lengths and angles for 13Hþ [Fig. 2(a)]
˚
Bond lengths (A)
O(1)—N(1)
O(1)—C(12)
N(1)—C(11)
1.401(3)
1.433(3)
1.274(3)
1.475(3)
N(2)—C(13)
N(2)—C(14)
C(1)—C(2)
C(1)—C(9)
1.497(3)
1.500(3)
1.389(3)
1.450(3)
C(1)—C(11)
C(7)—C(8)
C(8)—C(9)
C(9)—C(10)
1.469(3)
1.371(4)
1.432(3)
1.435(3)
N(2)—C(8)
Bond angles (ꢀ)
N(1)—O(1)—C(12)
C(11)—N(1)—O(1)
C(8)—N(2)—C(13)
C(8)—N(2)—C(14)
C(13)—N(2)—C(14)
108.3(2)
C(2)—C(1)—C(9)
C(2)—C(1)—C(11)
C(9)—C(1)—C(11)
C(7)—C(8)—C(9)
C(7)—C(8)—N(2)
118.7(2)
C(9)—C(8)—N(2)
C(8)—C(9)—C(10)
C(8)—C(9)—C(1)
C(10)—C(9)—C(1)
N(1)—C(11)—C(1)
119.7(2)
112.5(2)
115.7(2)
111.2(2)
111.0(2)
111.6(2)
128.9(2)
122.3(2)
117.6(2)
115.1(2)
128.3(2)
116.6(2)
124.5(2)
Table 4. Selected bond lengths and angles for 14 [Fig. 2(b)]
˚
Bond lengths (A)
O(1)—N(1)
N(1)—C(1)
N(2)—C(10)
1.422(3)
1.275(4)
1.419(4)
1.450(4)
N(2)—C(12)
C(1)—C(2)
C(2)—C(3)
C(2)—C(11)
1.464(4)
C(6)—C(11)
C(9)—C(10)
C(10)—C(11)
1.428(4)
1.373(5)
1.443(5)
1.477(4)
1.374(4)
1.439(4)
N(2)—C(13)
Bond angles (ꢀ)
C(1)—N(1)—O(1)
C(10)—N(2)—C(13)
C(10)—N(2)—C(12)
C(13)—N(2)—C(12)
N(1)—C(1)—C(2)
110.8(2)
C(3)—C(2)—C(11)
C(3)—C(2)—C(1)
C(11)—C(2)—C(1)
C(9)—C(10)—N(2)
C(9)—C(10)—C(11)
119.1(3)
N(2)—C(10)—C(11)
C(6)—C(11)—C(2)
C(6)—C(11)—C(10)
C(2)—C(11)—C(10)
118.5(3)
115.7(3)
113.1(3)
111.9(3)
117.4(3)
116.2(3)
124.5(3)
121.8(3)
119.7(3)
117.5(3)
117.6(3)
124.9(3)
Copyright # 2004 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2005; 18: 101–109

EFFICIENT INTRAMOLECULAR PROTON TRANSFER FROM CARBON
109
S ¼ 1.039. The structure was solved with SHELXS-9717
REFERENCES
and refined with SHELXL-9717 (CCDB 226253).
1. Kirby AJ. Adv. Phys. Org. Chem. 1980; 17: 183–278.
2. Kirby AJ. Acc. Chem. Res. 1997; 30: 290–296; Hartwell E,
Hodgson DRW, Kirby AJ. J. Am. Chem. Soc. 2000; 122: 9326–
9327.
3. Kirby AJ, Percy JM. J. Chem. Soc., Perkin Trans. 2 1989; 907–
912.
4. Kirby AJ, O’Carroll F. J. Chem. Soc., Perkin Trans. 2 1994; 649–
655.
5. Gulick AM, Hubbard BK, Gerlt JA, Rayment I. Biochemistry
2000; 39: 4590–4602; St. Maurice M, Bearne SL. Biochemistry
2000; 39: 26.
6. Alder RW, Bowman PS, Steele WRS, Winterman DR. J. Chem.
Soc., Chem. Commun. 1968; 723–724.
7. Hollfelder F. PhD Thesis, University of Cambridge, 1997.
8. Hodgson DRW, Kirby AJ, Feeder N. J. Chem. Soc., Perkin Trans.
1 1999; 949–954.
9. Cho BR, Cho NS, Lee SK. J. Org. Chem. 1997; 62: 2230–2233.
10. Cho BR, Chung HS, Cho NS. J. Org. Chem. 1998; 63: 4685–4690.
11. More O’Ferrall, RA, O’Brien DM, Murphy DG. Can. J. Chem.
2000; 78: 1594–1612.
12. Stewart R. The Proton: Applications to Organic Chemistry.
Academic Press: New York, 1985.
Crystal data for 14. C13H14N2O, M ¼ 214.26, tetragonal,
space group P41 (No. 76), a ¼ b ¼ 12.305(2), c ¼
0.643 mmꢂ1, 1558 reflections measured at 180(2) K using
an Oxford Cryosystems Cryostream cooling apparatus,
1422 unique (Rint ¼ 0.046); R1 ¼ 0.046, wR2 ¼ 0.116
[I > 2_(I)]; goodness-of-fit on F2, S ¼ 1.059. The struc-
ture was solved with SHELXS17 and refined with
SHELXL17 (CCDB 226236).
3
˚
7.506(6), U ¼ 1136.5(9) A , Z ¼ 4, ꢅ(Cu Kꢀ) ¼
Acknowledgements
We are grateful to the Engineering and Physical Sciences
Research Council of Great Britain for the award of a
Studentship (to N.A.) and for financial assistance towards
the purchase of the Nonius Kappa CCD diffractometer.
L.v.V. is a Benefactor’s Scholar of St. John’s College,
Cambridge and is grateful for support from the Cambridge
European Trust in collaboration with the Isaac Newton
Trust. L.O. was a summer (2003) project student from the
University of Perugia.
13. Kirby AJ, Percy JM. Tetrahedron 1988; 44: 6903–6910.
14. Dunitz JD. X-Ray Analysis and the Structure of Organic Mole-
cules. Cornell University Press: Ithaca, NY, 1979; 373–375.
15. Grech E, Malarski Z, SawkaDobrowolska W, Sobczyk L. J. Phys.
Org. Chem. 1999; 12: 313–318.
16. Kiefl C. Eur. J. Org. Chem. 2000; 3279–3286.
17. Sheldrick GM. SHELXS-97/SHELXL-97. University of Gottingen:
¨
Gottingen, 1997.
¨
Copyright # 2004 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2005; 18: 101–109