Thermally latent reaction of hemiacetal ester with epoxide controlled by Schiff-base-zinc chloride complexes with tunable catalytic activity

Article abstract of DOI:10.1016/j.molcata.2007.04.016

Schiff-base-zinc chloride complexes (ZnCl₂/1_R) thermal-latently catalyze the reaction of glycidyl phenyl ether (2) and 1-propoxyethyl-2-ethylhexanate (3) that proceeds at appropriate temperatures for latent curing. This reaction proceeds via the nucleophilic addition of carboxylic acid generated from the thermal dissociation of 3 to 2, which takes place faster than the reaction without ZnCl₂/1_R. Catalytic activities of ZnCl₂/1_R, depending on the basicities of the α-diimine ligands controllable by the substituents on the aromatic rings, were evaluated by kinetic parameters; namely the reaction rate constants (k), the activation energies (E_a), and the frequency factors (A). ZnCl₂/1_Cl bearing the electron-withdrawing chlorine group initiates the reaction above 80 °C, whereas ZnCl₂/1_OMe bearing the electron-donating methoxy group initiates the reaction above 100 °C. The E_a values in the reactions with ZnCl₂/1_Cl and ZnCl₂/1_OMe were estimated to be 52.2 and 177 kJ mol^-1, respectively, which agree with the latencies at ambient temperatures. The A values also differ with the catalysts (6.46 × 10² and 2.04 × 10¹⁹ L mol^-1 s^-1 for ZnCl₂/1_Cl and ZnCl₂/1_OMe, respectively). The very high A values for the catalysts with electron-donating groups manifest the very good latencies under ambient conditions, in spite of the high activities at elevated temperatures. The coordination behavior of ZnCl₂/1_R was evaluated by ¹H NMR, ¹³C NMR, ¹⁵N NMR, and IR spectroscopies to understand the substituent effects.

Full text of DOI:10.1016/j.molcata.2007.04.016

Journal of Molecular Catalysis A: Chemical 273 (2007) 289–297
Thermally latent reaction of hemiacetal ester with epoxide controlled by
Schiff-base–zinc chloride complexes with tunable catalytic activity
Hiroyuki Komatsua, Bungo Ochiaia, Tetsuo Hinob, Takeshi Endoa,∗
a Department of Polymer Science and Engineering, Faculty of Engineering, Yamagata University, 4-3-16 Jonan, Yonezawa, Yamagata 992-8510, Japan
b Graduate School of Science and Engineering, Yamagata University, 4-3-16 Jonan, Yonezawa, Yamagata 992-8510, Japan
Received 6 November 2006; received in revised form 10 April 2007; accepted 15 April 2007
Available online 19 April 2007
Abstract
Schiff-base–zinc chloride complexes (ZnCl2/1R) thermal-latently catalyze the reaction of glycidyl phenyl ether (2) and 1-propoxyethyl-2-
ethylhexanate (3) that proceeds at appropriate temperatures for latent curing. This reaction proceeds via the nucleophilic addition of carboxylic
acid generated from the thermal dissociation of 3 to 2, which takes place faster than the reaction without ZnCl2/1R. Catalytic activities of ZnCl2/1R,
depending on the basicities of the ␣-diimine ligands controllable by the substituents on the aromatic rings, were evaluated by kinetic parameters;
namely the reaction rate constants (k), the activation energies (Ea), and the frequency factors (A). ZnCl2/1Cl bearing the electron-withdrawing
chlorine group initiates the reaction above 80 ◦C, whereas ZnCl2/1OMe bearing the electron-donating methoxy group initiates the reaction above
100 ◦C. The Ea values in the reactions with ZnCl2/1Cl and ZnCl2/1OMe were estimated to be 52.2 and 177 kJ mol−1, respectively, which agree
with the latencies at ambient temperatures. The A values also differ with the catalysts (6.46 × 102 and 2.04 × 1019 L mol−1 s−1 for ZnCl2/1Cl and
ZnCl2/1OMe, respectively). The very high A values for the catalysts with electron-donating groups manifest the very good latencies under ambient
conditions, in spite of the high activities at elevated temperatures. The coordination behavior of ZnCl2/1R was evaluated by 1H NMR, 13C NMR,
15N NMR, and IR spectroscopies to understand the substituent effects.
© 2007 Elsevier B.V. All rights reserved.
Keywords: Schiff-base–zinc chloride complex; Lewis acid; Thermally latent catalyst; Hemiacetal ester; Epoxide
1. Introduction
should increase vigorously with temperatures. Temperature
dependence is evaluated with the well-known Arrhenius equa-
tion, by which the activation energies (Ea) and the frequency
factors (A) are calculated. Although a large Ea value typically
decreases a reaction rate constant (k) as shown in the Arrhenius
equation, a large A value can compensate to achieve a practi-
cal reaction rate at high temperatures [1,3b]. As an example,
Pappas and Hill et al. reported that sulfonium salts are promis-
ing catalysts for latent curing [1], in which polymerization of
epoxides shows large Ea and A value (e.g., 208 kJ mol−1 and
1.8 × 109 s−1, respectively). Endo et al. have reported phos-
phonium ylides as thermally latent catalysts for the reaction
of epoxides with phenol-resins [3]. The large Ea values of
phosphonium ylides bearing bulky substituents are ascribed to
the thermal stability of the linkages between the carbon and
phosphorous atoms that directly affects the catalytic activity
[3b].
Typical catalysts are demanded to decrease activation ener-
gies of specific reactions for acceleration and increase in
selectivity. However, some reactions, for example, latent reac-
tions, do not require simple acceleration. Latent reactions, which
are inactive under ambient conditions but active only with spe-
cific external stimulation such as heating and photoirradiation,
have been widely applied to adhesives, coatings, paintings, and
photo-resists [1,2]. Such reactions allow simple processing,
because mixtures of precursors ready to use may be treated with
simple procedures.
Desirable thermally latent reactions should have sharp tem-
perature dependence, and it means that the rates of the reactions
∗
Corresponding author. Present address: Molecular Engineering Institute,
Kinki University, 11-6 Kayanomori, Iizuka, Fukuoka 820-8555, Japan.
Tel.: +81 948 22 7210; fax: +81 948 22 7210.
As substrates for latent reactions, we have been focusing
on hemiacetal ester derivatives of useful thermally latent car-
E-mail address: tendo@me-henkel.fuk.kindai.ac.jp (T. Endo).
1381-1169/$ – see front matter © 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcata.2007.04.016

290
H. Komatsu et al. / Journal of Molecular Catalysis A: Chemical 273 (2007) 289–297
Scheme 1.
boxylic acids, which thermally dissociate into carboxylic acids
and vinyl ethers (Scheme 1) [4]. Catalysts are required for the
reactions of hemiacetal esters so that carboxylic acids thermal-
latently generated from hemiacetal esters can nucleophilically
react with epoxides under appropriate temperature for curing
(e.g., 120–160 ◦C) [4c,5,6], because spontaneous dissociation
of hemiacetal esters typically takes place only at very high tem-
perature (>160 ◦C) [5,6].
2.2. Measurements
1H and 13C nuclear magnetic resonance (NMR) spectra were
recorded with a Varian INOVA 500 spectrometer using tetram-
ethylsilane as an internal standard in CDCl3 at an ambient
temperature (500 and 125 MHz for H and 13C NMR, respec-
1
tively). 15N NMR spectra (40.5 MHz) were recorded with a
JEOL ECX 400 spectrometer using nitromethane as an inter-
nal standard in CDCl3 at an ambient temperature. The δ values
are given in ppm. Fourier-transform infrared (FT-IR) spectra
were recorded with a JASCO FT/IR-470 plus spectrometer. Gas
chromatography mass spectra (GC–MS) were measured with a
Shimadzu GCGS-QP5050A instrument. Elemental analysis was
carried out with a Perkin-Elmer PE2400-2 CHNS/O instrument.
Melting points (Tm) were determined by differential scanning
calorimetry (DSC) measurements on a SEIKO Instruments DSC
6200 and EXSTAR 6000 system at a scanning rate of 10 ◦C/min
under a nitrogen atmosphere.
We have recently reported that Schiff-base–zinc halide
complexes (ZnX2/1H) thermal-latently catalyze addition of
a hemiacetal ester with an epoxide [6]. Unexpectedly, the
catalytic activity of the Schiff-base–zinc halide complexes
(ZnI2/1H > ZnBr2/1H ≥ ZnCl2/1H) does not correlate with the
electronegativities of the halides. This result supposes that
the Schiff-base ligand does not coordinate with ZnI2 suf-
ficiently owing to the steric hindrance of iodide. In order
to design Lewis acid complexes with more tunabilities, we
have focused on the basicities of the Schiff-base ligands that
may be tuned by para-substituents on the aromatic rings.
The introduction of para-substituents on the aromatic rings
is expected to control both the activity and the latency.
For example, we presume that the introduction of electron-
donating groups increases the latency by decreasing the Lewis
acidity of the complex. Therefore, we have investigated ther-
mally latent reaction of a hemiacetal ester and an epoxide
using zinc chloride complexes with various Schiff-base lig-
ands (ZnCl2/1R). The catalytic activity of the Schiff-base–zinc
chloride complexes (ZnCl2/1R) was also evaluated from kinetic
analysis.
2.3. Syntheses of N,N’-bis(p-substituted
benzylidene)-1,2-diiminoethane (1R)
N,N’-Bis(benzylidene)-1,2-diiminoethane (1H) was prepared
as previously reported [6]. The other ligands were also prepared
in similar manners, and their structures were confirmed by 1H
NMR, 13C NMR, 15N NMR, IR, and GC–mass spectroscopic
and elemental analyses. The ligands were obtained as white
needle crystals.
2.3.1. N,N’-Bis(p-chlorobenzylidene)-1,2-diiminoethane
(1Cl); yield: 91%
2. Experimental
1H NMR: 8.22 (s, 2H, CH N ), 7.63 (d, J = 11.0 Hz, 4H,
C6H4 ), 7.36 (d, J = 11.0 Hz, 4H, C6H4 ), 3.95 (s, 4H,
CH2 ). 13C NMR: 161.2 ( CH N ), 136.5, 134.4, 129.1,
128.7 ( C6H4 ), 61.3 ( CH2 ). 15N NMR: 331.02 ( CH N ).
IR (KBr): 3030, 2915 (C C, CH2 ), 1644 ( CH N ), 1592,
1486( C6H4 ). GC–MS(m/z):303. mp:148.0–148.5 ◦C. Elem.
Anal. for C16H14Cl2N2—Calcd.: C, 62.97; H, 4.62; N, 9.18.
Found: C, 62.93; H, 4.52; N, 9.10.
2.1. Materials
Ethylene diamine, p-chlorobenzaldehyde, benzaldehyde,
p-methylbenzaldehyde, p-tert-butylbenzaldehyde, p-methoxy-
benzaldehyde, p-ethoxybenzaldehyde, and p-N,N-dimethyl-
aminobenzaldehyde (Kanto Kagaku, Tokyo, Japan) were
used as received. Zinc chloride in diethyl ether solution
[(1 M), Sigma Aldrich, Missouri, MO] was used as received.
Glycidyl phenyl ether (2) (Kanto Kagaku, Tokyo, Japan)
was dried over CaH2, and distilled under reduced pres-
sure. 1-Propoxyethyl-2-ethyl hexanoate (3) was prepared
as reported [7]. 2-Ethylhexanoic acid (4) (Kanto Kagaku,
Tokyo, Japan) was used as received. Other chemicals were
commercial products with extra-pure grade, and used as
received.
2.3.2. N,N’-Bis(benzylidene)-1,2-diiminoethane (1H);
yield: 92%
1H NMR: 8.30 (s, 2H, CH N ), 7.70 (m, 4H, C6H5),
7.40 (m, 6H, C6H5), 3.98 (s, 4H, CH2 ). 13C NMR:
162.8 ( CH N ), 136.1, 130.5, 128.5, 128.0, ( C6H5), 61.5
( CH2 ). 15N NMR: 328.36 ( CH N ). IR (KBr): 3060,
3032, 2934 (C C, CH2 ), 1642 ( CH N ) 1597, 1489

H. Komatsu et al. / Journal of Molecular Catalysis A: Chemical 273 (2007) 289–297
291
( C6H5). GC–MS (m/z): 236. mp: 55.1–56.4 ◦C. Elem. Anal.
for C16H16N2—Calcd.: C, 81.32; H, 6.82; N, 11.85. Found: C,
81.24; H, 6.79; N, 11.82.
4H, CH2 ), 2.99 [s, 6H, N(CH3)2]. 13C NMR: 162.4
( CH N ), 152.0, 129.4, 124.5, 111.5 ( C6H4 ), 61.9
( CH2 ), 40.0 [ N(CH3)3]. 15N NMR: 328.88 ( CH N ),
47.36 [ N(CH3)2]. IR (KBr): 2829 (C C, CH2 ), 1638
( CH N ), 1604 ( C6H4 ). GC–MS (m/z): 322. mp:
179.0–179.8 ◦C. Elem. Anal. for C20H26N4—Calcd.: C,
74.50; H, 8.13; N, 17.38. Found: C, 74.56; H, 8.15; N,
17.34.
2.3.3. N,N’-Bis(p-methylbenzylidene)-1,2-diiminoethane
(1Me); yield: 90%
1H NMR: 8.24 (s, 2H, CH N ), 7.59 (d, J = 8.0 Hz, 4H,
C6H4 ), 7.19 (d, J = 8.0 Hz, 4H, C6H4 ), 3.94 (s, 4H,
CH2 ), 2.36 (s, 6H, CH3). 13C NMR: 162.4 ( CH N ),
140.7, 133.5, 129.1, 127.9 ( C6H4 ), 61.6 ( CH2 ), 21.4
( CH3). 15N NMR: 324.67 ( CH N ). IR (KBr): 3046, 2923,
2911 (C C, CH2 ), 1637 ( CH N ), 1605, 1508 ( C6H4 ).
GC–MS (m/z): 264. mp: 161.7–162.6 ◦C. Elem. Anal. for
C18H20N2—Calcd.: C, 81.78; H, 7.63; N, 10.60. Found: C,
81.54; H, 7.57; N, 10.53.
2.4. Preparation of N,N’-bis(p-substituted
benzylidene)-1,2-diiminoethane/zinc chloride complex
(ZnCl2/1R)
ZnCl2/1R [N,N’-bis(benzylidene)-1,2-diiminoethane/zinc
chloride complex] was prepared as previously reported [6].
The other complexes (ZnCl2/1H) were also prepared in
2.3.4. N,N’-Bis(p-tert-butylbenzylidene)-1,2-diiminoethane
(1tBu); yield: 92%
1
similar manners, and their structures were confirmed by H
NMR, 13C NMR, and 15N NMR and IR spectroscopic and
elemental analyses. All the complexes were obtained as a white
powder.
1H NMR: 8.27 (s, 2H, CH N ), 7.64 (d, J = 9.5 Hz,
4H, C6H4 ), 7.41 (d, J = 9.5 Hz, 4H, C6H4 ), 3.95 (s,
4H, CH2 ), 1.32 [s, 18H, C(CH3)3]. 13C NMR: 162.3
( CH N ), 153.8, 133.4, 127.8, 125.4 ( C6H4 ), 61.7
( CH2 ), 34.7 [ C(CH3)3], 31.1 [ C(CH3)3]. 15N NMR:
324.76 ( CH N ). IR (KBr): 3020, 2901 (C C, CH2 ), 1644
( CH N ), 1605, 1503 ( C6H4 ). GC–MS (m/z): 348. mp:
115.6–116.8 ◦C. Elem. Anal. for C24H32N2—Calcd.: C, 82.71;
H, 9.25; N, 8.04. Found: C, 82.93; H, 9.46; N, 8.04.
2.4.1. N,N’-Bis(benzylidene)-1,2-diiminoethane/zinc
chloride complex (ZnCl2/1H); yield: 92%
1H NMR: 8.60 (s, 2H, CH N ), 8.11 (d, J = 9.0 Hz,
4H, C6H5), 7.59 (d, J = 9.0 Hz, 6H, C6H5), 4.10 (s, 4H,
CH2 ). 13C NMR: 170.5 ( CH N ), 146.5, 134.2, 130.7,
129.2, ( C6H5), 59.7 ( CH2 ). IR (KBr): 3060, 3032, 2934
(C C, CH2 ), 1644 ( CH N ), 1597 ( C6H4 ). Elem. Anal.
for C16H16N2Cl2Zn—Calcd.: C, 51.57; H, 4.33; N, 7.52. Found:
C, 51.33; H, 4.44; N, 7.42.
2.3.5. N,N’-Bis(p-methoxybenzylidene)-1,2-diiminoethane
(1OMe); yield: 90%
1H NMR: 8.21 (s, 2H, CH N ), 7.64 (d, J = 7.5 Hz, 4H,
C6H4 ), 6.90 (d, J = 7.5 Hz, 4H, C6H4 ), 3.91 (s, 6H,
OCH3), 3.83 (s, 4H, CH2 ). 13C NMR: 162.0 ( CH N ),
161.7, 129.7, 129.3, 114.0 ( C6H4 ), 61.9 ( CH2 ), 55.5
( OCH3). 15N NMR: 320.46 ( CH N ). IR (KBr): 3046, 2971,
2920 (C C, CH2 ), 1639 ( CH N ). GC–MS (m/z): 296.
mp: 112.7–113.4 ◦C. Elem. Anal. for C18H20N2O2—Calcd.: C,
72.95; H, 6.80; N, 9.45. Found: C, 72.76; H, 6.76; N, 9.40.
2.4.2.
N,N’-Bis(p-methylbenzylidene)-1,2-diiminoethane/zinc
chloride complex (ZnCl2/1Me); yield: 90%
1H NMR: 8.34 (s, 2H, CH N ), 8.00 (d, J = 8.0 Hz, 4H,
C6H4 ), 7.35 (d, J = 8.0 Hz, 4H, C6H4 ), 4.03 (s, 4H,
CH2 ), 2.42 (s, 6H, CH3). 13C NMR: 169.8 ( CH N ),
145.1, 130.8, 129.8, 128.5, ( C6H4 ), 59.7 ( CH2 ), 21.8
( CH3). 15N NMR: 257.84 ( CH N ). IR (KBr): 3046, 2937
(C C, CH2 ), 1637 ( CH N ), 1607, 1512 ( C6H4 ). Elem.
Anal. for C18H20Cl2N2Zn—Calcd.: C, 53.96; H, 5.03; N, 6.99.
Found: C, 53.36; H, 4.89; N, 6.88.
2.3.6. N,N’-Bis(p-ethoxybenzylidene)-1,2-diiminoethane
(1OEt); yield: 90%
1H NMR: 8.20 (s, 2H, CH N ), 7.62 (d, J = 7.5 Hz,
4H, C6H4 ), 6.89 (d, J = 7.5 Hz, 4H, C6H4 ), 4.05 (q,
J = 7.0 Hz, 4H, OCH2 ), 3.91 (s, 4H, CH2 ), 1.42 (t,
J = 7.0 Hz, 6H, CH3). 13C NMR: 161.8 ( CH N ), 160.8,
129.5, 128.9, 114.2 ( C6H4 ), 63.4 ( OCH2 ), 61.6 ( CH2 ),
55.5 ( CH3). 15N NMR: 320.21 ( CH N ). IR (KBr): 2984,
2929 (C C, CH2 ), 1641 ( CH N ), 1603, 1509 ( C6H4 ).
GC–MS (m/z): 324. mp; 115.7–116.1 ◦C. Elem. Anal. for
C20H24N2O2—Calcd.: C, 74.04; H, 7.46; N, 8.64. Found: C,
74.13; H, 7.41; N, 8.65.
2.4.3.
N,N’-Bis(p-tert-butylbenzylidene)-1,2-diiminoethane/zinc
chloride complex (ZnCl2/1tBu); yield: 92%
1H NMR: 8.50 (s, 2H, CH N ), 8.07 (d, J = 9.5 Hz,
4H, C6H4 ), 7.58 (d, J = 9.5 Hz, 4H, C6H4 ), 4.03
(s, 4H, CH2 ), 1.34 [s, 18H, C(CH3)3]. 13C NMR
(CDCl3, 125 MHz): 169.7 ( CH N ), 158.0, 130.8, 128.3,
126.2 ( C6H4 ), 59.7 ( CH2 ), 35.2 [ C(CH3)3], 30.9
[ C(CH3)3]. 15N NMR: 257.36 ( CH N ). IR (KBr):
3056, 2963 (C C, CH2 ), 1641 ( CH N ), 1606, 1510
( C6H4 ). Elem. Anal. for C24H32Cl2N2Zn—Calcd.: C,
59.46; H, 6.65; N, 5.78. Found: C, 59.32; H, 6.57; N,
5.77.
2.3.7.
N,N’-Bis(p-dimethylaminobenzylidene)-1,2-diiminoethane
(1NMe ); yield: 84%
2
1H NMR: 8.16 (s, 2H, CH N ), 7.59 (d, J = 8.0 Hz,
4H, C6H4 ), 6.69 (d, J = 8.0 Hz, 4H, C6H4 ), 3.87 (s,

292
H. Komatsu et al. / Journal of Molecular Catalysis A: Chemical 273 (2007) 289–297
2.4.4.
N,N’-Bis(p-methoxybenzylidene)-1,2-diiminoethane/zinc
chloride complex (ZnCl2/1OMe); yield: 90%
1H NMR: 8.44 (s, 2H, CH N ), 8.09 (d, J = 9.0 Hz,
4H, C6H4 ), 7.05 (d, J = 9.0 Hz, 4H, C6H4 ), 4.00 (s,
6H, OCH3). 13C NMR: 168.9 ( CH N ), 164.1, 133.0,
123.9, 114.6 ( C6H4 ), 59.6 ( CH2 ), 55.5 ( OCH3). 15N
NMR: 252.02 ( CH N ). IR (KBr): 2935 (C C, CH2 ),
1632 ( CH N ), 1600, 1514 ( C6H4 ). Elem. Anal. for
C18H20Cl2N2O2Zn—Calcd.: C, 49.97; H, 4.66; N, 6.47. Found:
C, 50.00; H, 4.54; N, 6.46.
Scheme 2.
1
by H NMR, 13C NMR, 15N NMR, and IR spectroscopic and
2.4.5.
elemental analysis (Scheme 2).
N,N’-Bis(p-ethoxybenzylidene)-1,2-diiminoethane/zinc
chloride complex (ZnCl2/1OEt); yield: 90%
The 1H NMR, 13C NMR, and 15N NMR chemical shifts and
the IR absorption bands of the ZnCl2/1R correlate with the basic-
ities of the original Schiff-bases depending on the electronic
characters of the substituents. Table 1 summarizes the 1H NMR,
13C NMR, and 15N NMR chemical shifts and IR absorption band
of the Schiff-base–zinc chloride complexes. Complexes with
electron-donating substituents show the NMR signals of imine
(CH N) at higher magnetic field, demonstrating the electron
densities of the imine groups depending on the electron charac-
terofthesubstituents. Weestimatedthedegreeofthedonationby
the ligand in ZnCl2/1R from the spectroscopic data. Specifically,
the degree of donation of ␲-electrons in imine groups was evalu-
ated from the differences between the 15N NMR chemical shifts
and IR absorption band of the imine groups in the complexes and
the free ligands based on the fact that donation of ␲-electrons in
the imine groups to acceptors shifts 15N NMR chemical shifts
and the IR absorption bands toward higher magnetic field and
lower wavenumber, respectively [8,9]. The ꢀppm values from
the 15N NMR spectroscopic data increased as the increase in the
electron-donating ability (i.e., the Hammett’s σp values) of the
substituents [10], although we could not evaluate ZnCl2/1H and
1H NMR: 8.43 (s, 2H, CH N ), 8.06 (d, J = 9.0 Hz, 4H,
C6H4 ), 7.00 (d, J = 9.0 Hz, 4H, C6H4 ), 4.13 (q, J = 7.0 Hz,
4H, OCH2 ), 3.99 (s, 4H, CH2 ), 1.43 (t, J = 8.0 Hz, 6H,
CH3). 13C NMR: 168.9 ( CH N ), 163.6, 133.0, 123.6,
114.9 ( C6H4 ), 63.8 ( OCH2 ), 59.6 ( CH2 ), 14.5 ( CH3).
15N NMR: 251.60 ( CH N ). IR (KBr): 3051, 2925 (C C,
CH2 ), 1635 ( CH N ), 1601, 1512 ( C6H4 ). Elem. Anal.
for C20H24Cl2N2O2Zn—Calcd.: C, 52.14; H, 5.25; N, 6.08.
Found: C, 52.07; H, 5.14; N, 6.09.
2.4.6. N,N’-Bis(p-dimethylaminobenzylidene)-1,2-
diiminoethane/zinc chloride complex (ZnCl2/1NMe ); yield:
88%
2
1H NMR: 8.30 (s, 2H, CH N ), 7.99 (d, J = 9.0 Hz, 4H,
C6H4 ), 6.77 (d, J = 9.0 Hz, 4H, C6H4 ), 3.92 (s, 4H,
CH2 ), 3.07 [s, 6H, N(CH3)2]. 13C NMR: 168.5 ( CH N ),
154.1, 133.2, 118.8, 111.5 ( C6H4 ), 59.7 ( CH2 ), 39.6
[ N(CH3)3]. IR (KBr): 2909 (C C, CH2 ), 1628 ( CH N ),
1593 ( C6H4 ). Elem. Anal. for C20H26Cl2N4Zn—Calcd.: C,
52.36; H, 5.71; N, 12.21. Found: C, 52.00; H, 5.64; N, 12.03.
ZnCl2/1NMe , because of their low solubility. The shifts toward
2
higher magnetic field agree with the cases of the coordination
of the nitrogen ligand to metals [11,12] and the protonation of
amines [8] and imines [13], by which the electron densities of
the nitrogen atoms decrease [8]. The larger shifts toward higher
magnetic field are attributable to the stronger donation of the
␲-electron on the nitrogen atoms included in the conjugated
␲ systems consisting of the aromatic rings and the conjugated
diimines to zinc atom (Scheme 3). The electronic characters of
the substituents also affect the IR absorption bands of the ligands
and the complexes. The IR absorption bands for the imines in the
Schiff-base ligands (CH N stretching vibration) were observed
at approximately 1640 cm−1, and the wavenumbers decrease as
the electron-donating ability of the substituents increases. The
wavenumbers were shifted by the coordination, and the ꢀcm−1
values (i.e., the shifts in the wavenumbers of the imine group
from those in the free ligands to those in the complexes) also
increased with the increase in the electron-donating ability of
the para-substituents on the aromatic rings. Lower wavenumber
shifts are attributable to the decreases in the sp2 characters of the
imine group by stronger donation of the ␲-electron. These obser-
vations support that the strength of the coordination increases
2.5. The reaction of 2 and 3 with ZnCl2/1H as a thermal
latent catalyst (typical procedure)
A mixture of 2 (0.150 g, 1.00 mmol), 3 (0.230 g, 1.00 mmol),
and ZnCl2/1H (22.3 mg, 60.0 ␮mol, 3.00 mol% to 2 and 3) was
heated to 30–140 ◦C for 3 h under a nitrogen atmosphere. The
thermal dissociation of hemiacetal ester 3, the consumption of
2, and the reaction of 2 with the carboxylic acid from 3 were
quantified by 1H NMR spectroscopy.
3. Results and discussion
3.1. Preparation of Schiff-base–zinc chloride complexes
Schiff-base–zinc chloride complexes (ZnCl2/1R) bearing
substituents on the para-position of the aromatic rings were pre-
pared via the coordination of Schiff-base ligands (1R) to zinc
chloride. The complexes except ZnCl2/1NMe were obtained as
2
white solids, while ZnCl2/1NMe was yielded as a pale yellow
2
solid. The structures of the obtained products were confirmed

H. Komatsu et al. / Journal of Molecular Catalysis A: Chemical 273 (2007) 289–297
293
Table 1
NMR chemical shift and IR absorption band of CH N in Schiff-base–zinc chloride complexes (ZnCl2/1R) and their differential values to those of Schiff-base ligands
(1R) at 25 ◦C
R (σp)a
15N NMR chemical shiftb of
CH N (ꢀppm)
IR absorption bandc of
13C NMR chemical shiftb of
CH N (ꢀppm)
1H NMR chemical shiftb of
CH N (ꢀppm)
CH N (ꢀcm−1
)
d
d
d
d
Cl (0.23)
H (0.00)
–
–
–
–
e
–
1644.1 (+1.9)
1636.3 (−3.9)
1641.1 (−3.9)
1635.3 (−5.8)
1632.4 (−6.8)
1628.4 (−9.8)
170.52 (+7.76)
169.89 (+7.41)
169.52 (+7.24)
168.85 (+7.01)
168.86 (+6.83)
168.47 (+6.04)
8.602 (+0.306)
8.510 (+0.268)
8.502 (+0.246)
8.444 (+0.236)
8.427 (+0.226)
8.269 (+0.168)
Me (−0.17)
But (−0.20)
OEt (−0.25)
OMe (−0.27)
NMe2 (−0.83)
257.84 (−66.8)
257.36 (−67.4)
251.60 (−68.6)
252.02 (−70.4)
e
–
a
σp values of the substituents on 1R in Ref. [14].
Measured in CDCl3. The difference in the NMR chemical shift (ꢀppm) of imine groups in the complexes and free ligand. Positive and negative values indicate
b
shift toward lower and higher magnetic field, respectively.
c
Measured as KBr disks. The difference in the IR absorption bands (cm−1) of imine groups in the complexes and free ligand. Positive and negative values indicate
shifts toward higher and lower wavenumber regions, respectively.
d
Not isolable from the mixture of ZnCl2/1Cl and unreacted 1Cl
.
e
Resolved 15N NMR spectrum could not be obtained due to the poor solubility of the metal complex in CDCl3.
Scheme 3.
with increasing basicity of the ligand, and hence, ZnCl2/1R with
weakly basic ligands is assumed to consist of weak coordina-
tion. This assumption agrees with the fact that 1Cl did not make
significant coordination with ZnCl2, which results in affording
the mixture containing a partial amount of ZnCl2 and 1Cl.
ZnCl2/1R [Cl, H, Me, tBu, OMe, OEt, NMe2] (3.0 mol%) at
various temperatures for 3 h (Scheme 4). Fig. 1 shows the
relationship between temperatures versus conversion of 2. No
addition reaction of 2 and 3 proceeded with ZnCl2/1R within
24 h under ambient conditions (<70 ◦C), and this indicates that
ZnCl2/1R complexes show activity only at elevated tempera-
tures. It is advantageous over bare ZnCl2 that shows activity
even at room temperature (e.g., approximately 10% amounts of
2 was converted at 25 ◦C) [6]. Among these catalysts, ZnCl2/1Cl
initiated the addition at the lowest temperature (80 ◦C). The
complexes with electron-donating groups, ZnCl2/1OMe and
3.2. The reaction of hemiacetal ester (3) and epoxide (2)
with Schiff-base–zinc chloride complex (ZnCl2/1R)
In order to evaluate the electronic effect of the para-
substituents on the Schiff-base ligands (1R) on the catalytic
activity, 2 and 3 were allowed to react in the presence of
ZnCl2/1NMe initiated the reaction at 100 ◦C and 110 ◦C,
2
Scheme 4.

294
H. Komatsu et al. / Journal of Molecular Catalysis A: Chemical 273 (2007) 289–297
respectively, and the other complexes did at 90 ◦C. The conver-
sions became quantitative at 140 ◦C with ZnCl2/1Cl, ZnCl2/1H,
ZnCl2/1Me, and ZnCl2/1OMe, but stayed approximately 80%
and 70% at 140 ◦C with ZnCl2/1tBu and ZnCl2/1OEt, respec-
tively. The low conversions in the reactions using ZnCl2/1tBu
and ZnCl2/1OEt may be ascribed to their lower catalytic activ-
ities due to the electron-donating abilities of the substituents,
although the reason for the high conversion with ZnCl2/1OMe is
unclear. Very low conversion was observed for the reaction with
ZnCl2/1NMe . This is assumed to originate from the suppressed
2
dissociation of the hemiacetal ester due to the basicity of the
dimethylamino group on the Schiff-base ligand (1NMe ) [15] in
2
addition to the very low Lewis acidities due to the high electron-
donating ability of the dimethylamino group. Fig. 2 shows the
1H NMR spectra of the reaction mixtures before (A) and after
(B) the reaction of 2 and 3 with ZnCl2/1 (3.0 mol%) at 140 ◦C.
The products composed of three compounds, i.e., an ester with
a hydroxyl group (5) and esters with acetal groups (6 and 7) [6].
Fig. 3(A) shows the relationship between the time and the
conversion of 2 in the reaction of 2 and 3 with 3.0 mol% amounts
of ZnCl2/1R at 120 ◦C. As the electron-donating ability of the
Fig. 1. Time vs. conversion of 2 relationship in the addition of 2 and 3 with
3.0 mol% amounts of ZnCl2/1R (1Cl, ♦; 1H; ᭹; 1Me, ꢀ; 1tBu, ꢀ; 1OEt, ꢁ; 1OMe
,
ꢁ; 1NMe2 , ꢂ) at various temperatures.
Fig. 2. 1H NMR spectra of the reaction mixtures before (A) and after (B) the reaction of 2 and 3 with ZnCl2/1 (3.0 mol%) at 140 ◦C for 3 h, recorded in CDCl3.

H. Komatsu et al. / Journal of Molecular Catalysis A: Chemical 273 (2007) 289–297
295
Fig. 4. Hammett plot for the ln(k1R/k1H) values in the reaction of 2 and 3 with
ZnCl2/1R at 120 ◦C.
Fig. 3(B) shows the time versus 1/[2]–1/[2]0 relation-
ships in the reaction of 2 and 3 with 3.0 mol% amounts
of ZnCl2/1R at 120 ◦C. The second-order kinetic plots
show linear relationships between the reaction time and
the 1/[2]–1/[2]0 values, which indicate that the reactions
with these complexes also obey the apparent second-order
kinetics. The slopes of the lines reveal the apparent reac-
tion rate constants (kapp): k = 8.387 × 10−5 L mol−1 s−1
,
k
k
k
k
k
k
= 6.256 × 10−5 L mol−1 s−11Cl
,
1H
= 5.153 × 10−5 L mol−1 s−1
,
,
,
1Me
= 3.958 × 10−5 L mol−1 s−1
1tBu
1OEt
1OMe
1NMe
= 3.003 × 10−5 L mol−1 s−1
Fig. 3. (A) Relationship between the time vs. conversion of 2, and (B) rela-
tionship between the time vs. 1/[2]–1/[2]0 values in the reaction of 2 and 3 with
= 2.603 × 10−5 L mol−1 s−1
,
and
= 5.833 × 10−6 L mol−1 s−1
.
3.0 mol% amounts of ZnCl2/1R, (1Cl, ♦; 1H, ᭹; 1Me, ꢀ; 1tBu, ꢀ; 1OEt, ꢁ; 1OMe
,
2
ꢁ; 1NMe2 , ꢂ) at 120 ◦C.
As the electron-donating ability of the para-substituents on
the aromatic rings increased, the rate constants in the reaction
with ZnCl2/1R decreased. In order to demonstrate the substituent
effect clearly, we plotted the ln(k1R/k1H) values against the Ham-
mett’s σp values (Fig. 4). The ln(k1R/k1H) values correlated very
well with σp values. This correlation clearly demonstrates that
the electronic character of the para-substitutents on the Schiff-
base ligands controls the kinetics of this latent reaction.
The kinetic constants were determined for the reactions at
other temperatures (110, 130, and 140 ◦C), whose kinetics also
obey the apparent second-order kinetics. The rate constants at
the varied temperatures allowed us to calculate the activation
energies (Ea) and the frequency factors (A) expressed by the
Arrhenius’s equation [Eq. (3)]:
para-substituent increased, the conversions of 2 decreased. We
attribute this behavior to the harder coordination abilities of
the imines by the electron-donating groups that decrease the
catalytic activity of ZnCl2/1R by lowering the Lewis acidities.
We expressed the reaction of 2 and 3 in the presence of
ZnCl2/1R bytheapparentsecond-orderkineticsinasimilarman-
ner with the reaction with ZnCl2/1H, whose rate depends on both
of the concentrations of 2 and 3 owing to the slower recombi-
nation of the carboxylic acid and the vinyl ether [6]. Based on
the assumption that the concentration of unreacted 2 and 3 are
equal in the reaction mixtures, the reactant rate constant (k) can
be expressed by the equation below.
k = Ae−E /RT
(3)
a
d[C]
−
= k[C]2 ([C] : concentration of 2)
(1)
Eq. (3) can be logarithmically transformed to Eq. (4):
dt
Ea
Integrating Eq. (1) gives Eq. (2):
ln k = −
+ ln A
(4)
RT
1
1
Ea and A values can be calculated from slopes of lines and
extrapolated intercepts, if logarithms of rate constants (ln k) cor-
−
= kt ([C]0 : initial concentration of 2)
(2)
[C] [C]0

296
H. Komatsu et al. / Journal of Molecular Catalysis A: Chemical 273 (2007) 289–297
Table 2
Activation energy (Ea) and frequency factor (A) in the addition of 2 and 3 in the
presence of ZnCl2/1R (3.0 mol%) at 110–140 ◦C
Ea (kJ mol−1
b
)
Ab (s−1
)
a
Additive
σp
ZnCl2/1Cl
ZnCl2/1H
0.23
0
52.2
74.7
126
129
130
6.46 × 102
4.41 × 105
3.20 × 1012
3.31 × 1012
8.82 × 1012
2.04 × 1019
1.38 × 105
ZnCl2/1Me
ZnCl2/1tBu
ZnCl2/1OEt
ZnCl2/1OMe
ZnCl2/1NMe
−0.17
−0.20
−0.25
−0.27
−0.83
177
78.3
2
a
␴ values of the substituents on 1R in Ref. [14].
Calculated by the Arrhenius equation based on the reaction rate constants
p
b
determined by the first-order plots at 110–140 ◦C.
Fig. 5. Arrhenius plots in the reaction of 2 and 3 with 3.0 mol% amounts of
ZnCl2/1R (1Cl, ♦; 1H, ᭹; 1Me, ꢀ; 1tBu, ꢀ; 1OEt, ꢁ; 1OMe, ꢁ; 1NMe2 , ꢂ) at
110–140 ◦C.
with bare ZnCl2 proceeded even at room temperature, and was
accompanied by observable side reactions due to the polymer-
ization of 2 above 110 ◦C. The Ea and A value of this reaction
with bare ZnCl2 involving side reactions was estimated to be
approximately 45 kJ mol−1 and 4.20 × 104 s−1, which are not
high enough for a latent reaction. The reactions with ZnCl2/1R
relatelinearlywithreciprocalsoftemperatures(T−1)asshownin
Fig. 5. Theplotsshowlinearrelationshipsbetweentheln kvalues
and the reciprocals of temperatures, and allow us to determine
the activation energies (Ea) and the frequency factors (A). Table 2
summarizes the Ea and A values in the reaction of 2 and 3 with
3.0 mol% amounts of ZnCl2/1R (R: Cl, H, Me, tBu, OEt, OMe,
NMe2). Ea values are 50–100 kJ mol−1 in smooth reactions,
which proceed even at room temperature (e.g., a typical Ea value
in an alcoholysis reaction of an isocyanate giving an urethane
in the absence of catalysts is ca. 75 kJ mol−1 [16]). On the other
hand, chemicalreactionswithEa valueslargerthan100 kJ mol−1
typically require additional energies such as heating. Thermally
latent reaction must have a large Ea value at ambient conditions
to attain good latency, and should have a large A value to attain
high reactivity at higher temperature. For example, polymer-
ization of epoxides using sulfonium salts1 shows large Ea and
exhibitlargerEa andAvaluesrangingfrom52.2to170 kJ mol−1
,
6.46 × 102 to 2.04 × 1019
s
−1, respectively, except the reaction
withZnCl2/1NMe . Inparticular, thecomplexesbearingelectron-
2
donating groups tend to exhibit the large Ea values and A factors.
TheAfactorsarecomparabletothoseinthermallylatentreaction
of 2,6-dimethylphenol and glycidyl phenyl ether with phospho-
nium ylides (i.e.; 2.04 × 10−3 to 1.61 × 10 s−1) [3b]. The com-
plex bearing methoxy group (ZnCl2/1OMe) exhibits the largest
Ea and A values among the examined complexes. The large A
value leads to an excellent latency, which was demonstrated by
the fact that ZnCl2/1OMe accomplishes the addition quantita-
tively within 3 h at 140 ◦C in spite of the inertness below 90 ◦C.
These results almost correlate with the aforementioned changes
in the 15N NMR chemical shifts (ꢀppm) and the IR absorption
bands (ꢀcm−1) by ligation, and suggest that the electronic char-
acters of the Schiff-base ligands control both the latency and the
activity in the reaction of hemiacetal esters with epoxides.
A values (e.g., 208–316 kJ mol−1, 1.8 × 109 to 3.2 × 1018 s−1
,
respectively), which result in good latency at ambient condi-
tions and high activity at elevated temperatures. The reaction
Fig. 6. Energy profile in the reaction of hemiacetal ester (3) and epoxide (2) with Schiff-base–zinc chloride complex (ZnCl2/1OMe) at 140 ◦C.

H. Komatsu et al. / Journal of Molecular Catalysis A: Chemical 273 (2007) 289–297
297
Thereactionof2and3proceedsviatwo-stepreactions, which
[2] For reviews:
(a) T. Endo, F. Sanda, Macromol. Symp. 107 (1996) 237;
(b) M. Shirai, M. Tsunooka, Prog. Polym. Sci. 21 (1996) 1;
(c) Y. Yaguchi, T. Endo, Adv. Polym. Sci. 127 (1997) 59;
(d) J.V. Crivello, J. Polym. Sci., Part A: Polym. Chem. 37 (1999)
4241;
is assumed to involve an activation energy for each primary reac-
tion (indicated as Ea1 and Ea2 for the first and second reactions,
respectively, below and in Fig. 6), although the calculated Ea val-
ues are the apparent activation energies of the whole reactions.
The Ea1 value expresses the activation energy of the thermal dis-
sociation of 3, and the Ea2 value expresses the activation energy
of the reaction of 2 and 2-ethylhexanoic acid (4). Accordingly,
we determined the Ea2 value in the reaction of 2 and 4 with
ZnCl2/1OMe in a similar manner to the addition of 2 and 3. The
Ea2 value, 69.3 kJ mol−1 [17], is significantly lower than the
Ea value (177 kJ mol−1). This observation suggests that the Ea
value is almost identical to the Ea1 value and that the primary
dissociation is the rate-determining step. Thus, we concluded
that the control over the primary dissociation with ZnCl2/1OMe
is a necessary condition for the high latency.
(e) J.V. Crivello, J. Ma, F. Jiang, H. Hac, J. Ahn, R.A. Ortiz, Macromol.
Symp. 215 (2004) 165.
[3] (a) M. Kobayashi, F. Sanda, T. Endo, Macromolecules 32 (1999)
4751;
(b) M. Kobayashi, F. Sanda, T. Endo, Macromolecules 33 (2000)
5384;
(c) M. Kobayashi, F. Sanda, T. Endo, Macromolecules 34 (2001)
1134;
(d) M. Kobayashi, F. Sanda, T. Endo, Macromolecules 35 (2002)
346.
[4] (a) H. Otsuka, T. Endo, Macromolecules 32 (1999) 9059;
(b) Y. Nakane, M. Ishidoya, T. Endo, J. Polym. Sci., Part A: Polym. Chem.
37 (1999) 609;
(c) H. Otsuka, H. Fujiwara, T. Endo, React. Funct. Polym. 46 (2001) 293;
(d) H. Komatsu, T. Hino, T. Endo, J. Polym. Sci., Part A: Polym. Chem. 44
(2006) 3966.
4. Summary
[5] (a) Y. Nakane, M. Ishidoya, Prog. Org. Coat. 31 (1997) 113;
(b) T. Yamamoto, M. Ishidoya, Prog. Org. Coat. 40 (2000) 267.
[6] H. Komatsu, B. Ochiai, T. Hino, T. Endo, J. Polym. Sci., Part A: Polym.
Chem., submitted for publication.
[7] H. Komatsu, T. Hino, T. Endo, J. Polym. Sci., Part A: Polym. Chem. 43
(2005) 4260.
[8] G.C. Levy, R.L. Lichter, Nitrogen-15 Nuclear Magnetic Resonance Spec-
troscopy, Wiley, New York, 1979.
[9] M. Amirnasr, A.H. Mahmoudkhani, A. Gorji, S. Dehghanpour, H.R. Bijan-
zadeh, Polyhedron 21 (2002) 2733.
[10] The ꢀppm values from the 1H NMR and 13C NMR spectroscopic data of
the obtained complexes exhibit the contrastive relationships to those from
15N NMR spectroscopy, and namely the shifts toward lower magnetic field
through coordination. Similar behaviors were reported in with 1H NMR and
13C NMR studies for cobalt(III) peptide complexes. The ꢀppm values from
1H and 13C NMR spectra decrease as the increase of the electron-donating
ability of the para-substituents on the aromatic rings, although the ꢀppm
values from 15N NMR spectrum increase as the stronger electron-donating
of substituent (see Table 1).
[11] (a) Y. Wu, H.B. Daryle, J. Am. Chem. Soc. 94 (1972) 4115;
(b) E.J. Evans, J.E. Grice, C.J. Hawkins, M. Heard, Inorg. Chem. 19 (1980)
3496.
[12] (a) N. Juranic, R.L. Lichter, J. Am. Chem. Soc. 105 (1983) 406;
(b) N. Juranic, S. Vukpavlovic, A.T. Nikolic, J. Inorg. Biochem. 62 (1996)
117;
Schiff-base–zinc chloride complexes (ZnCl2/1R) proved the
effective Lewis acids meeting requirements for both latency
and activity in the thermally latent reaction of a hemiac-
etal ester and an epoxide. Kinetic studies revealed that the
activation energies and the frequency factors depend on the
Lewis acidity of the catalysts controlled by the electronic
characters of the substituents. For instance, ZnCl2/1Cl bear-
ing an electron-withdrawing chlorine group promotes the
reaction approximately three times as fast as ZnCl2/1OMe
bearing an electron-donating methoxy group at 120 ◦C (i.e.,
8.39 × 10−5 L mol−1 s−1 and 2.60 × 10−5 L mol−1 s−1, respec-
tively), although the rate for ZnCl2/1OMe is satisfactory for
curing. The differing rates originate from the Ea values. The
Ea value for ZnCl2/1Cl is one-third of that of ZnCl2/1OMe (i.e.,
52.2 kJ mol−1 and 177 kJ mol−1, respectively). As well as the
activity, the latency also depends on the electronic characters
of the substituents. The complexes with basic ligands tend to
have higher A values, and the highest A value was observed for
ZnCl2/1OMe (2.04 × 1019
s
−1) that attains very good latency.
TheelectroniccharactersofthesubstituentscontroltheLewis
acidities of the Schiff-base–zinc chloride complexes that regu-
late both the activity and latency. The controllable activity and
latency are effective in conducting curing under desired condi-
tions. The fine-tuning of the Lewis acidities suggests the wide
applicability of the Schiff-base–zinc chloride complexes not
only for the curing system based on the reaction of hemiacetal
esters and epoxides but also for catalysis of reactions involving
activation of acid-sensitive substrates like epoxides.
(c) N. Juranic, S. Macura, Inorg. Chim. Acta 62 (1996) 213.
[13] E.B. Robert, D.R. John, J. Org. Chem. 44 (1979) 140.
[14] L.P. Hammett, Chem. Rev. 17 (1935) 125.
[15] Hemiacetal esters are stabilized by catalytic amounts of basic reagents such
as tertiary amines and tertiary phosphines, whereas susceptible to cleavage
in the presence of acidic compounds (see Ref. [5]). A plausible reason
for the very low conversion by ZnCl2/1NMe is competitive activation and
2
deactivation by the Lewis acidic zinc-center and the basic dimethylamino
group, respectively.
[16] (a) J.W. Backer, M.M. Davies, J. Gaunt, J. Chem. Soc. (1949) 24;
(b) J.W. Backer, J.B. Holdworth, J. Chem. Soc. (1947) 713;
Acknowledgement
(c) J.W. Backer, J. Gaunt, J. Chem. Soc. (1949) 27.
[17] The Ea and A values in the addition of 2 and 4 with ZnCl2/1OMe
were determined in similar manners with those in the addition of
3 and 2. The Ea (69.3 kJ moL−1) and A (4.11 × 10−4 s−1) were cal-
culated from the slope and the extrapolated intercept of the linear
plots on the relationship between the logarithm of the rate con-
stants and the reciprocals of the temperatures (110–140 ◦C). The rate
constants are follows; 110 ◦C, kobs = 1.44 × 10−5 L mol−1 s−1, 120 ◦C,
The authors acknowledge the kind support of Dr. Takeyoshi
Takahashi at Faculty of Engineering of Yamagata University for
conducting elemental analysis.
References
kobs = 2.64 × 10−5 L mol−1 s−1; 130 ◦C, kobs = 4.47 × 10−5 L mol−1 s−1
;
and 140 ◦C, kobs = 8.17 × 10−5 L mol−1 s−1
.
[1] S.P. Pappas, L.W. Hill, J. Coat. Technol. 53 (1981) 43.