Quinoline-containing networks from enone and aldehyde triglyceride derivatives

Article abstract of DOI:10.1002/pola.23838

The crosslinking reaction of a triglyceride derivative containing α,β-unsaturated ketones with diaminodiphenyl- methane via aza-Michael addition has been extensively studied. First, a model study with monofunctional compounds showed that the conjugated addition product undergoes a series of transformations leading to formation of a substituted quinoline. The proposed reaction pathway is presented as a variation of the Skraup-Doebner-Von Miller quinoline synthesis. The presence of quinolines as crosslinking points in the cured materials has been proved by means of different characterization techniques, and the properties derived from this aromatization process have been described. This new crosslinking approach has been successfully applied to an aldehyde-containing triglyceride to obtain quinoline-containing thermosets.

Full text of DOI:10.1002/pola.23838

Quinoline-Containing Networks from Enone and Aldehyde Triglyceride
Derivatives
`
´
LUCAS MONTERO DE ESPINOSA, JUAN C. RONDA, MARINA GALIA, VIRGINIA CADIZ
Departament de Qu´ımica Anal´ıtica i Qu´ımica Orga`nica, Universitat Rovira i Virgili, Campus Sescelades, Marcel.l´ı Domingo s/n,
43007 Tarragona, Spain
Received 13 October 2009; accepted 18 November 2009
DOI: 10.1002/pola.23838
Published online in Wiley InterScience (www.interscience.wiley.com).
ABSTRACT: The crosslinking reaction of a triglyceride derivative
containing a,b-unsaturated ketones with diaminodiphenyl-
methane via aza-Michael addition has been extensively studied.
First, a model study with monofunctional compounds showed
that the conjugated addition product undergoes a series of
transformations leading to formation of a substituted quinoline.
The proposed reaction pathway is presented as a variation of
the Skraup-Doebner-Von Miller quinoline synthesis. The pres-
ence of quinolines as crosslinking points in the cured materials
has been proved by means of different characterization techni-
ques, and the properties derived from this aromatization process
have been described. This new crosslinking approach has been
successfully applied to an aldehyde-containing triglyceride to
C
V
obtain quinoline-containing thermosets. 2010 Wiley Periodicals,
Inc. J Polym Sci Part A: Polym Chem 48: 869–878, 2010
KEYWORDS: aromatization; aza-Michael; biopolymers; crosslink-
ing; renewable resources; triglyceride
INTRODUCTION The utilization of fossil fuels for the manu-
facture of plastics accounts for about 7% of the worldwide
use of oil and gas, which will arguably be depleted within
the next 100 years.1 In these next decades of increasing oil
prices, global warming, and other environmental concerns, a
change from fossil feedstocks to renewable resources is im-
portant for sustainable development into the future.2 Among
the renewable raw materials, natural oils are the most
widely used renewable resource for the chemical and poly-
mer industries.3 The main components of the triglyceride
vegetable oils are saturated and unsaturated fatty acids.
Although they have double bonds, which can be used as re-
active sites in coatings, for obtaining high-performance poly-
meric materials, the introduction of more reactive functional
groups, such as hydroxyl, epoxy, or carboxyl groups, is much
more suitable.4,5 Various chemical pathways for functionaliz-
ing triglycerides and fatty acids have been studied.6 Epoxida-
tion is one of the most important functionalization reactions
of the CAC double bonds, which can be achieved by environ-
mentally friendly procedures such as catalyzed chemical oxi-
dation with hydrogen peroxide7,8 or by enzymatic oxidation.9
The opening of the epoxide ring is a versatile reaction that
leads to numerous products.4,10,11 Thus, epoxidized vegetable
oils are extremely promising as inexpensive renewable mate-
rials for industrial applications12 because they share many of
the characteristics of conventional epoxy thermosets. In this
way, naturally occurring epoxy oil, as vernonia seed oil or
epoxidized vegetable oils from soybean, linseed, or castor
oils have been cured cationically13 or with conventional
hardeners as diamines or dianhydrides.14–18
Recently, in a previous communication, we reported the syn-
thesis of a new highly reactive triglyceride derivative with
a,b-unsaturated carbonyl groups.19 This enone-containing tri-
glyceride obtained by an environmentally friendly chemical
procedure from high-oleic sunflower oil resulted an interest-
ing alternative to epoxidized vegetable oils to produce ther-
mosets by crosslinking with conventional aromatic diamines.
The Michael addition reaction came out as a valuable tool in
the synthesis of polymeric networks.20 The aza-Michael reac-
tion, a variation in which an amine acts as the nucleophile,
had been used in the synthesis of improved bismaleimide
networks,21 but this reaction had not been applied to the
synthesis of crosslinked polymers derived from vegetable
oils. In the aza-Michael reaction with the enone-containing
triglyceride derivative, we observed the existence of second-
ary reactions during the crosslinking process at high temper-
atures, reactions that can take place not only in this case but
also in any other curing process in which an aza-Michael
addition is involved. Thus, this work is focused on the study
of the nature of these secondary reactions and their extent,
as they can have a great influence in the properties of the
obtained thermosets. The reactions were followed by NMR,
UV, and fluorescence spectroscopy, showing that at high tem-
peratures aromatic moieties are formed in the network. This
finding is relevant to enhance the material properties. Usu-
ally, materials from oils are incapable of displaying the
Correspondence to: V. Ca´diz (E-mail: virginia.cadiz@urv.cat)
C
V
Journal of Polymer Science: Part A: Polymer Chemistry, Vol. 48, 869–878 (2010) 2010 Wiley Periodicals, Inc.
CROSSLINKING REACTION OF TRIGLYCERIDE DERIVATIVE, MONTERO DE ESPINOSA ET AL.
869

JOURNAL OF POLYMER SCIENCE: PART A: POLYMER CHEMISTRY DOI 10.1002/POLA
necessary rigidity and strength required for structural appli-
cations, and so modification or copolymerization with aro-
matic components is required to overcome this drawback.22
Finally, the evaluation of thermal properties of the final ma-
terial could be related to the presence of the aromatic
structures.
nary), 125.93 (CAr5), 36.12 (Ar-CH2), 32.60 (Ar-CH2), 32.10
(CH2ACH2ACH3), 32.07 (CH2ACH2ACH3), 30.77 (CH2), 30.17
(CH2), 30.07 (CH2), 29.80 (CH2), 29.78 (CH2), 29.68 (CH2),
29.54 (CH2), 29.47 (CH2), 22.88 (CH2ACH3), 21.69 (Ar-CH3),
14.32 (CH3).
Synthesis of vic-Diol-Containing Triglyceride (16)
Epoxidized high-oleic sunflower oil (5g, 5.3 mmol) was dis-
solved in 400 mL of THF in a 1-L round-bottomed flask,
which was placed in a water bath at 20 ꢁC. A mixture of
water (80 mL) and 60% HClO4 (2.4 mL) was added drop-
wise under vigorous stirring. The reaction was kept for 12 h,
and then the reaction mixture was extracted with dichloro-
methane and washed with water. The organic layer was
dried over MgSO4, and the solvent was removed at reduced
pressure. Compound 16 was obtained with 96% yield after
crystallization from hexane. The product has an average of
2.5 diol groups per triglyceride (by 1H NMR spectroscopy).
FTIR (cmꢂ1): 3540 (OAH), 1742 (C¼¼O), 1142 (CAO). 1H
NMR (CDCl3, TMS, d in ppm): 5.28–5.20 (m, CHAOCO), 4.28
(dd, J ¼ 12.0, 4.0 Hz, CH2AOCO), 4.12 (dd, J ¼ 12.0, 6.0 Hz,
CH2AOCO), 3.38–3.32 (m, CHAOH), 2.45–2.80 (broad, OH),
2.29 (t, J ¼ 7.4 Hz, CH2ACO), 1.65–1.52 (m, CH2ACH2ACO),
1.52–1.20 (m, CH2ACHOH and CH2), 0.85 (t, J ¼ 6.8 Hz,
CH3). 13C NMR (CDCl3, TMS, d in ppm): 173.55 (COOR),
173.12 (COOR), 74.70 (CHAOH), 74.63 (CHAOH), 69.07
(CHAOCO), 62.31 (CH2AOCO), 34.40 (CH2ACO), 34.23
(CH2ACO), 33.80 (CH2ACHOH), 32.07 (CH2ACH2ACH3),
29.91–29.14 (CH2), 25.91 (CH2ACH2ACHOH), 25.01
(CH2ACH2ACO), 22.87 (CH2ACH3), 14.31 (CH3).
EXPERIMENTAL
Materials
High-oleic sunflower oil (minimum 80% oleic acid) was
R
kindly supplied by BorgesV, BF3ꢀMEA (Aldrich), decanal
(Aldrich), HClO4 (60%, Probus), NaIO4 (Fluka), NaHCO3
(Scharlab), and MgSO4 (Scharlab) were used as received. p-
Toluidine (Aldrich) was recrystallized from heptane. Hexane,
ethyl acetate, tetrahydrofuran, dichloromethane, and 1,4-
dioxane were purchased to Scharlab and used directly. Silica
for column chromatography was purchased to SDS (60 A. C.
C. 40–63 lm) and silica gel TLC aluminum sheets to Merck
(60, F254). The mixture of methyl-9-oxo-10-octadecenoate
and methyl-10-oxo-8-octadecenoate (4),23 glyceryl tris(9-oxo-
10-octadecenoate), and glyceryl tris(10-oxo-8-octadecenoate)
mixture (1)19 and epoxidized high-oleic sunflower oil (15)19
were synthesized as previously reported. TLC plates were
developed with an UV lamp (254 nm) or by spraying with
sulfuric acid/anisaldehyde ethanol solution and heating at
ꢁ
200 C.
Reaction between p-Toluidine and Methyl-9-oxo-10-
octadecenoate/Methyl-10-oxo-8-octadecenoate Mixture
The a,b-unsaturated ketone (300.0 mg, 0.96 mmol), p-tolui-
dine (103.5 mg, 0.96 mmol), and BF3ꢀMEA (3.3 mg, 0.03
mmol) were mixed in a 10-mL round-bottomed flask under
argon. The stirred reaction mixture was heated at 90 ꢁC for
Synthesis of Aldehyde-Containing Triglyceride (17)
Compound 16 (5.2 g, 5.3 mmol) and NaIO4 (3.11 g, 14.5
mmol) were placed in a 100-mL round-bottomed flask, and
50 mL of a 9/1 mixture of 1,4-dioxane/water was added.
The reaction mixture was stirred vigorously at room temper-
ature for 1 h, and then it was diluted with 20 mL of
dichloromethane and washed twice with NaHCO3 and water.
The pale brown oily product was connected to a high vac-
uum pump equipped with a liquid nitrogen trap to remove
nonanal. The product was obtained quantitatively with an
ꢁ
ꢁ
30 min, at 110 C for 2 h, and finally at 140 C for 2 h more.
One hundred milligrams of samples were taken at the end of
1
each reaction step and analyzed by H NMR spectroscopy.
Quinoline (11) Synthesis from Decanal and p-Toluidine
Decanal (200 mg, 1.28 mmol), p-toluidine (137 mg, 1.28
mmol), and BF3ꢀMEA (4.3 mg, 0.038 mmol) were mixed in a
10-mL round-bottomed flask under argon. The stirred reac-
tion mixture was heated at 50 C for 4 h, and then the tem-
perature was raised to 140 C and maintained for 12 h. The
ꢁ
1
average of 2.5 aldehyde groups per triglyceride (by H NMR
ꢁ
spectroscopy). FTIR (cmꢂ1): 2719 (CAH, aldehyde), 1735
(C¼¼O, ester), 1721 (C¼¼O, aldehyde), 1162 (CAO), 1096
final mixture was analyzed by 1H NMR spectroscopy. Quino-
line 11 was isolated by column chromatography using hex-
ane/ethyl acetate 150/1 with 38% yield. FTIR (cmꢂ1): 3060
(CAH, Ar), 3016 (CAH, Ar), 2921 (CAH), 2850 (CAH), 1602
(C¼¼C), 1562 (C¼¼C), 1493 (C¼¼C), 1460 (CH2 and CH3), 823
(CAH, Ar). 1H NMR (CDCl3, TMS, d in ppm) (assignations
of quinoline ring according to IUPAC nomenclature): 7.90 (d,
J ¼ 8.50 Hz, 1H, C8AH), 7.74 (s, 1H, C4AH), 7.46 (s, 1H,
C5AH), 7.43 (dd, J ¼ 8.58, 1.92 Hz, 1H, C7AH), 2.96–2.92 (m,
2H, Ar-CH2), 2.77–2.73 (m, 2H, Ar-CH2), 2.49 (s, 3H, Ar-CH3),
1.81–1.74 (m, 2H, Ar-CH2ACH2), 1.71–1.63 (m, 2H, Ar-
CH2ACH2), 1.50–1.22 (m, 24H), 0.93–0.84 (m, 6H, CH3). 13C
NMR (CDCl3, TMS, d in ppm) (assignations of quinoline ring
according to IUPAC nomenclature): 161.50 (CAr2), 145.25
(quaternary), 135.31 (quaternary), 134.41 (CAr4), 134.22
(quaternary), 130.70 (CAr7), 128.31 (CAr8), 127.41 (quater-
1
(CAO). H NMR (CDCl3, TMS, d in ppm): 9.71 (t, J ¼ 1.77 Hz,
CHO), 5.24–5.18 (m, CHAOCO), 4.25 (dd, J ¼ 11.90, 4.28 Hz,
CH2AOCO), 4.09 (dd, J ¼ 11.91, 5.96 Hz, CH2AOCO), 2.38 (t,
J ¼ 7.36 Hz, CH2ACHO), 2.27 (dt, J ¼ 7.61 Hz, CH2ACOOR),
1.64–1.50 (m, CH2ACH2COOR and CH2ACH2CHO), 1.34–1.16
(m, CH2), 0.83 (t, J ¼ 6.83 Hz, CH3). 13C NMR (CDCl3, TMS, d
in ppm): 202.96 (CHO), 173.25 (COOR), 172.85 (COOR),
68.99 (CHAOCO), 62.19 (CH2AOCO), 43.92 (CH2ACHO),
34.18 (CH2ACOOR), 34.02 (CH2ACOOR), 32.01 (CH2ACH2
ACH3), 29.79–28.89 (CH2), 24.80 (CH2ACH2ACOOR), 22.79
(CH2ACH3), 22.05 (CH2ACH2ACHO), 14.24 (CH3).
Curing Reactions and Extraction of Soluble Parts
For the 1/DDM curing system (samples I and II), 1 and DDM
were dissolved in dichloromethane and placed inside a Petri
870
INTERSCIENCE.WILEY.COM/JOURNAL/JPOLA

ARTICLE
TABLE 1 Soluble Fractions, Tgs, and Thermogravimetric Data of the Cured Samples I–VIII
TGA (N2)
Soluble Part (%)a
Tg (ꢁC)b
T5% loss (ꢁC)
Tmax (ꢁC)
Char800 (%)
c
ꢁ
Sample
Curing System (molar ratio)
Catalyst
C
I
1/DDM (1/1.25)
1/DDM (1/1.25)
1/DDM (1/1.25)
1/DDM (1/1.25)
1/DDM (1/1.25)
17/DDM (1/0.75)
17/DDM (1/1)
–
25.6
20.3
20.5
18.2
13.0
7.0
ꢂ8
16
27
34
48
57
73
84
283
320
275
304
324
339
331
321
466
466
9.4
11.7
9.7
II
–
III
IV
V
BF3ꢀMEA
BF3ꢀMEA
BF3ꢀMEA
BF3ꢀMEA
BF3ꢀMEA
BF3ꢀMEA
466
472
11.1
13.9
16.1
16.3
16.1
468
VI
VII
VIII
a
414/467
413/468
413/468
6.4
17/DDM (1/1.25)
9.9
12 h soxhlet extraction of 0.5 g with 125 mL of DCM.
Maximum of the tan d peak.
Temperature of the maximum weight loss rate.
b
c
sions 10 ꢃ 5 ꢃ 0.2 mm3 were used. Thermal stability stud-
ies were carried out on a Mettler TGA/SDTA851e/LF/1100
with N2 as purge gas. The studies were performed in the
30–800 ꢁC temperature range at a scan rate of 10 ꢁC/min.
The UV–vis data were acquired and monitored by a Hewlett-
Packard 8452A spectrophotometer using the HP89531A soft-
ware. The spectra were recorded from 190 to 800 nm in
4 nm steps. The spectrofluorimetric data were acquired on
an Aminco-Bowman Series 2 Luminescence spectrometer
(SLM Aminco, Rochester, NY) equipped with a 150 W contin-
uous xenon lamp and a PMT detector. The instrument was
interfaced by a GPIB card and driver with a PC Pentium
microcomputer provided with the AB2 software version 1.40
for spectral acquisition. For UV–vis and fluorescence mea-
surements, 20-lm-thick films supported on a quartz plate
were used. For the preparation of the films, the dichloro-
methane or THF solutions used for the curing reactions were
cast over the quartz plate.
ꢁ
dish. The mixture was heated at 50 C for 30 min to remove
the solvent, and then sample I was heated for 4 h at 90 ꢁC
and sample II, 4 h at 90 ꢁC and 12 h at 120 ꢁC. For the 1/
DDM/BF3ꢀMEA curing system (samples III–V), 1, DDM, and
BF3ꢀMEA (3 mol % related to ketone groups) were dissolved
in dichloromethane, and the solution was put inside a Petri
dish and heated at 50 ꢁC for 30 min to remove the solvent.
For each sample, the corresponding curing program was
then applied. Sample III (90 ꢁC/4 h), sample IV (90 ꢁC/4 h
and 120 C/12 h), and sample V (90 ꢁC/4 h and 140 C/12
h). For the 17/DDM/BF3ꢀMEA curing system (samples VI–
VIII), 17 and DDM (see Table 1 for proportions in each sam-
ple) were dissolved in THF separately. BF3ꢀMEA (3 mol %
related to aldehyde groups) was added to the DDM solution
and then both solutions were mixed. The mixture was
ꢁ
ꢁ
ꢁ
quickly placed into a Petri dish and maintained at 40 C for
2 h to eliminate all the solvent. The same cꢁuring program
was applied for the three samples, 4 h at 90 C and 12 h at
ꢁ
140 C.
RESULTS AND DISCUSSION
All samples were subjected to soxhlet extraction with previ-
ously distilled dichloromethane to determine their soluble
fractions. A total of 0.5 g of each sample was extracted with
125 mL of dichloromethane. Before extractions, the samples
were grinded to maximize the extraction efficiency.
As mentioned in the Introduction, we synthesized a new
enone-containing triglyceride derivative (1) by an environ-
mentally friendly chemical procedure from high-oleic sun-
flower oil using the singlet oxygen ‘‘ene’’ reaction. A mixture
of allylic hydroperoxides was obtained and further trans-
formed in a regioisomeric mixture of enones in the presence
of acetic anhydride and pyridine or tertiary amines.19 We
proved the high efficiency of the aza-Michael crosslinking
Characterization
1H NMR 400 MHz and 13C NMR 100.6 MHz NMR spectra
were obtained using a Varian Gemini 400 spectrometer with
Fourier transform. CDCl3 was used as a solvent and TMS as
an internal reference. The IR analyses were performed on a
reaction between
1
and diaminodiphenylmethane (2)
(Scheme 1) in a 1H NMR kinetic experiment using model
compounds. However, although conversion of the a,b-unsatu-
rated ketone was uniform throughout the experiment until
total consumption, it could be observed that formation of
the aza-Michael adduct reached a point from which no fur-
ther growth was observed.
FTIR-680PLUS spectrophotometer with
a resolution of
2 cmꢂ1 in the transmittance mode. An attenuated total
reflection accessory with thermal control and a diamond
crystal was used to determine FTIR/ATR spectra. Dynamic
mechanical thermal analysis (DMTA) was performed using a
TA DMA 2928 in the controlled force-tension film mode with
a preload force of 0.1 N, an amplitude of 10 lm, and at a
This fact, together with a slight decrease of the aza-Michael
product concentration at long reaction times, caused us to
consider the possibility of a secondary reaction taking place
in which the aza-Michael product was involved. Moreover,
ꢁ
fixed frequency of 1 Hz, in the ꢂ100 to 200 C range and at
ꢁ
a heating rate of 3 C/min. Rectangular samples with dimen-
CROSSLINKING REACTION OF TRIGLYCERIDE DERIVATIVE, MONTERO DE ESPINOSA ET AL.
871

JOURNAL OF POLYMER SCIENCE: PART A: POLYMER CHEMISTRY DOI 10.1002/POLA
SCHEME 1 Aza-Michael crosslinking reac-
tion between 1 and DDM (2).
the dynamomechanical analysis of the crosslinked material
gave a very broad tan d peak, indicating low structural ho-
mogeneity and therefore supporting the theory of the exis-
tence of minor secondary reactions during the crosslinking
process.19 We could also observed that at higher tempera-
tures the extent of these secondary reactions increases, lead-
ing to an unexpectedly harder material.
the forming network and a deterioration of the mechanical
properties. However, although the noncatalyzed crosslinking
reaction at 90 ꢁC (sample I, Table 1) gave a material with a
ꢁ
Tg of ꢂ8 C (maximum of the tan d peak) and a soluble frac-
tion of 25.6%, the material obtained at 120 ꢁC (sample II,
ꢁ
Table 1) gave a Tg of 16 C, a soluble fraction of 20.3%, and
a higher crosslinking density (by DMTA). From this data, it
can be inferred that not only the retro-Mannich fragmenta-
tion is taking place but also further reactions leading to
crosslinked networks. As these chemical processes could be
generalized to any other curing processes in which an aza-
Michael addition is involved, it is important to determine
their nature and the effect of the reaction conditions.
Our first aim is to understand and establish the reactions
taking place in the crosslinking reaction after the aza-Mi-
chael adduct is formed. As the reaction of 1 and 2 leads to
an insoluble crosslinked material (Scheme 1), the accurate
analysis of the chemical reactions involved in the process
cannot be performed. For this reason, p-toluidine (3) and a
methyl oleate-derived enone (4) were used as model com-
For this purpose, 3 and 4 were reacted again in the presence
of 3 mol % of BF3ꢀMEA. In this model reaction, no solvent
was added to better reproduce the crosslinking reaction con-
ditions. To determine the effect of the temperature on the
formed products, the mixture was first heated at ꢁ90 ꢁC for
1
pounds, and a H NMR analysis was carried out.
As commented, high temperatures and long reaction times
are important factors in the development of side reactions;
for this reason, we tested the effect of adding a Lewis acid
catalyst (BF3ꢀMEA) to the reaction between 3 and 4 in an
attempt to reduce the reaction time. The conversion of the
reactants was accelerated, and the 1H NMR signals corre-
sponding to the secondary products were even increased.
The chromatographic isolation of the side products revealed
the presence, among other complex structures, of a methyl
ketone. This finding can only be explained by a Lewis acid-
catalyzed retro-Mannich type fragmentation of the aza-Mi-
chael adduct24 that would lead to the formation of a methyl
ketone (8) and an aldimine (6) (Scheme 2).
ꢁ
30 min, then at 110 C for 2 h, and finally at 140 C for 2 h
1
more. The progress of the reaction was followed by H NMR
spectroscopy (Fig. 1).
After heating the mixture at 90 ꢁC for 30 min, the signals c
and d of the aza-Michael product (5) can be observed to-
gether with a decreasing in the intensity of the conjugated
ketone (4) double bond signals (a and b) [Fig. 1(A)]. How-
ever, a singlet (e) at 2.07 ppm appears, which belongs to a
methyl ketone, presumably formed in the expected retro-
Mannich type fragmentation (Scheme 2). This fragmentation,
which proceeds through a BF3ꢀMEA-catalyzed condensation
of a molecule of p-toluidine and the aza-Michael adduct,
During the crosslinking reaction of 1 and 2, the fragmenta-
tion of the aza-Michael adduct would mean the breakage of
SCHEME 2 Lewis acid-catalyzed retro-Mannich type fragmentation of the aza-Michael adduct and formation of a methyl ketone (8)
and an aldimine (6).
872
INTERSCIENCE.WILEY.COM/JOURNAL/JPOLA

ARTICLE
FIGURE
1
1H NMR spectra of
reaction 3 þ 4 in the presence of
3 mol % of BF3ꢀMEA after heat-
ing. (A) At 90 ꢁC for 30 min. (B)
After raising the temperature to
110 ꢁC for 2 h. (C) When the
reaction mixture was heated for
2 h at 140 ꢁC.
produces two new species, an aldimine (6) and a ketimine
(7). Although the ketimine can be hydrolyzed under the
reaction conditions to the observed methyl ketone (8) and p-
toluidine, the aldimine 6 is more reactive and prone to
undergo self-aldol condensation25 [Scheme 3(A)].
omatic signals were clearly observed [Fig. 1(C)]. As mentioned ear-
lier, the aldimine 6 can undergo a BF3ꢀMEA self-aldol condensation
to yield 9, which, in the reaction conditions, can undergo a cycliza-
tion reaction to give the dihydroquinoline 10. The subsequent
elimination of a molecule of p-toluidine,25 followed by aromatiza-
tion through hydrogen transfer to Schiff bases present in the mix-
ture, can give the quinoline 11 [Scheme 3(A)].
ꢁ
When the temperature was raised to 110 C and kept for 2 h,
the starting enone and the intermediate aza-Michael adduct
signals disappeared almost completely [Fig. 1(B)]. The singlet
of the methyl ketone (e) remained unaltered ruling out the
self-aldol condensation in the reaction conditions and new sig-
nals appeared in the aromatic region, which could be attrib-
uted to further reactions involving the aldimine 6.
Thus summarizing, the reaction pathway proposed involves
several reactions: aza-Michael addition, retro-Mannich type
fragmentation, self-condensation of the aldimine fragment
and its cyclization followed by deamination and aromatiza-
tion. The fact that the methyl ketone remains unaltered
throughout all the reaction steps suggests that the retro-
Mannich fragmentation is irreversible and that the methyl
ketone is not involved in further steps of the reaction.
Finally, the reaction mixture was heated for 2 h at 140 ꢁC. The
methyl ketone still remained unaltered, but in this case, the new ar-
SCHEME 3 (A) Proposed self-aldol condensation of the aldimine 6 and further cyclization and aromatization to give quinoline 11.
(B) Reaction of decanal with p-toluidine in the presence of BF3ꢀMEA (3 mol %) to give aldimine 13 and further cyclization and aro-
matization to quinoline 14.
CROSSLINKING REACTION OF TRIGLYCERIDE DERIVATIVE, MONTERO DE ESPINOSA ET AL.
873

JOURNAL OF POLYMER SCIENCE: PART A: POLYMER CHEMISTRY DOI 10.1002/POLA
FIGURE 2 1H NMR spectra of (A)
final reaction mixture of 12 þ 3
in the presence of BF3ꢀMEA and
(B) pure quinoline 14.
Studies on the Skraup-Doebner-Von Miller reaction by Den-
mark and Venkatraman26 allowed them to establish a com-
plex mechanism for the condensation of aniline derivatives
with 3-disubstituted a,b-unsaturated ketones to form quino-
lines. It consists of a conjugated addition followed by a frag-
mentation to the corresponding imine and ketone and a
recombination of both fragments to form an anil that leads
to the final quinoline product. The initial conjugated addition
and subsequent fragmentation are in accordance to our
results. However, the presence of the ketone 8 at the end of
the reaction rules out condensation with the aldimine 6.
Instead of this, the aldimine 6 undergoes self-aldol conden-
sation leading to the final quinoline product as described
earlier. The difference may relay on the Michael acceptor
used. In our study, the a,b-unsaturated ketone is 3-monosub-
stituted and, thus, fragmentation of the aza-Michael adduct
yields an aldimine and a ketimine instead of two ketimines.
The higher reactivity of the aldimine compared with the keti-
mine justifies the self-aldol condensation of the former
instead of the cross condensation.
of 1 and DDM in presence of 3 mol % of BF3ꢀMEA. For this
purpose, equal amounts of the three-component mixture
were put on three different Petri dishes. The first dish (sam-
ꢁ
ple III) was heated at 90 C for 4 h, the second diꢁsh (sample
ꢁ
IV) was heated for 4 h at 90 C and 12 h at 120 C, and the
ꢁ
third dish (sample V) was heated for 4 h at 90 C and 12 h
ꢁ
at 140 C.
The synthesis of quinolines through reaction of in situ-gener-
ated alkylimines was described by Tanaka et al.27 Therefore,
to further confirm the propensity of the intermediate aldi-
mine to give self-aldol condensation and to produce the final
quinoline, decanal (12) was taken as a model compound.
Compound 12 was reacted with p-toluidine in the presence
of BF3ꢀMEA (3 mol %) at 50 ꢁC to allow formation of the
aldimine 13 and then at 140 ꢁC to favor the cyclization and
aromatization reactions [Scheme 3(B)]. The 1H NMR spec-
trum of the final reaction mixture [Fig. 2(A)] revealed the
formation of the expected quinoline 14 [Fig. 2(B)], which
was obtained with 38% yield after column chromatography.
These results confirm the proposed mechanism and can help
to explain the crosslinking temperature-dependent proper-
ties found in the materials obtained when triglyceride deriv-
ative 1 reacts with DDM. Moreover, as most plant oil-derived
materials can be considered as elastomers because of their
high content in long alkyl chains, the formation of quinolines
as crosslinking points would lead to a marked improvement
in their thermal and mechanical properties. Thus, the study
of the properties derived from this aromatization process is
of great interest.
FIGURE 3 Normalized UV–vis spectra as a function of cure
temperature of (A) 1/2/BF3ꢀMEA system (a: initial mixture, b:
90 ꢁC, c: 110 ꢁC, d: 120 ꢁC, and e: 140 ꢁC) and (B) 17/2/BF3ꢀMEA
system (a: initial mixture, b: 90 ꢁC, c: 110 ꢁC, and d: 140 ꢁC).
We started by studying the effect of the crosslinking temper-
ature on the properties of the materials obtained by reaction
874
INTERSCIENCE.WILEY.COM/JOURNAL/JPOLA

ARTICLE
FIGURE 4 Normalized fluorescence emission spectra as a function of the cure temperature of 1/2/BF3ꢀMEA system with excitation
wavelengths at (A) 255 nm and (B) 325 nm (a: initial mixture, b: 90 ꢁC, c: 120 ꢁC, and d: 140 ꢁC) and 17/2/BF3ꢀMEA system with ex-
citation wavelengths at (C) 255 nm and (D) 325 nm (a: initial mixture, b: 60 ꢁC, c: 90 ꢁC, and d: 140 ꢁC).
The expected chemical changes in this curing system include
the aza-Michael reaction of an aromatic amine with an enone
system and the appearance of a new aromatic structure.
Thus, UV–vis and fluorescence spectroscopies can be useful
techniques to monitor the crosslinking process. For this pur-
pose, the measurements were performed with 20-lm-thick
films supported on a quartz plate. Figure 3(A) shows the
normalized UV–vis spectra as a function of cure temperature
for 1/2/BF3ꢀMEA system. Although the spectral regions of
several species overlap, there is a clear change in the spectra
of cure temperatures over 90 ꢁC with the appearance of a
new absorption peak at longer wavelengths (ꢄ370 nm). As
previously mentioned, the aromatization process proceeds at
high temperatures. As the quinoline structure has a higher
conjugation degree than the initial species and possesses a
donor heteroatom, the observed reꢁd shift confirms their
presence, which is maximized at 140 C.
excitation wavelengths at 255 and 325 nm. The excitation
wavelengths were chosen as the ones giving the maxima of
emission in a 2D excitation–emission spectrum of the initial
mixture. By using a constant excitation wavelength through-
out the curing process, the changes in the emission spectra
can be related to the species involved in the process. In Fig-
ure 4(A), as the temperature is raised to 90 ꢁC, part of the
fluorescence of the initial mixture occurring at short wave-
lengths disappears as a result of the reaction between 1 and
2 and the subsequent fragmentation. When the sample is
heated at 120 ꢁC and finally at 140 ꢁC, the fluorescence in-
tensity at longer wavelengths is increased because of the
developing aromatization process. In Figure 4(B), a clear
shift of the fluorescence spectra toward longer wavelengths
is observed as the cure temperature is increased. These find-
ings are in accordance with the development of an aromatic
system with a higher conjugation degree.
The crosslinking reaction could also be monitored following
the changes in the fluorescence spectra. Figure 4(A,B)
depicts the fluorescence emission spectra of the 1/2/
BF3ꢀMEA system as a function of the cure temperature with
As already explained, an increase of the crosslinking temper-
ature should promote the formation of quinolines through
the proposed mechanism. In this way, although sample III
was expected to only suffer retro-Mannich type
CROSSLINKING REACTION OF TRIGLYCERIDE DERIVATIVE, MONTERO DE ESPINOSA ET AL.
875

JOURNAL OF POLYMER SCIENCE: PART A: POLYMER CHEMISTRY DOI 10.1002/POLA
of samples III–V (Fig. 6). As aromatic moieties are known to
promote char formation during the combustion process, the
analysis of the char yield at 800 ꢁC (Table 1) can provide
valuable information. Sample III has a 9.7% char yield
because of the presence of DDM as one of the main compo-
nents. The char yield increases to 11.1% for sample IV as
the quinolines start forming, and finally, for sample V, the
char yield obtained is 13.9%. This increase in the char yield
can be attributed to the aromatization process. In these TGA
plots can also be observed that the temperature of 5% loss
increases from sample III to sample V according to the
higher content in aromatic structures and high crosslinking
density of the sample V.
Based on the information obtained with the model reactions
and the study and interpretation of the properties of these
materials, it seems that quinoline structures are formed as
crosslinking points and that the temperature plays a key role
on the aromatization process. However, the retro-Mannich
fragmentation of the former intermediate aza-Michael adduct
causes the presence of long aliphatic methyl ketones that
can move freely acting as plasticizers. As starting triglyceride
derivative 1 is a statistical mixture of positional isomers, the
retro-Mannich fragmentation also leads to aliphatic methyl
ketones, which are not linked to the polymeric network. To
overcome this problem and improve the thermal and me-
chanical properties of the final materials, a new strategy was
used. Taking into account the high reactivity of aldimine 13
toward quinoline formation, we decided to prepare a triglyc-
eride derivative containing aldehyde groups. The reaction of
this new derivative with DDM in the presence of BF3ꢀMEA
should give an intermediate aldimine, which, at high temper-
ature, would lead to quinoline formation [see Scheme 3(B)].
In this way, we should be able to obtain a crosslinked poly-
mer with higher content of quinolines as crosslinking points.
FIGURE 5 Tan d plots as a function of temperature for the
cured samples III–VIII.
fragmentation, the crosslinking temperature used for sample
V should allow quinoline formation. An increase in the aro-
matic content of the polymeric network would increase the
rigidity, and thus the glass transition temperature (Tg)
should increase from sample III to sample V. The dynamome-
chanical analysis of samples III, IV, and V is shown in Figure
5, and the data are collected in Table 1.
As expected, the Tg (maximum of the tan d peak) increases
from sample III (27 ꢁC) to sample V (48 ꢁC). Interestingly,
the height of the tan d peak decreases from sample III to
sample V indicating an increase in the crosslinking density
despite the initial retro-Mannich fragmentation. However, the
tan d peak width at half-height increases, what is associated
with an increasing number of branching modes and a wider
distribution of structures, showing a lower structural homo-
geneity that should be caused by the mentioned fragmenta-
tion. It was explained earlier that aromatization of the dihy-
droquinoline 10 can be accomplished through hydrogenation
of Schiff bases present in the reaction media. The amines
formed in this way are still crosslinking points, and, thus, no
cleavage of the network is produced. Moreover, Tanaka
et al.27 demonstrated that aromatization of the intermediate
hydroquinoline is promoted in air atmosphere. For this rea-
son, despite the initial fragmentation, the crosslinking den-
sity increases together with the structural diversity.
For the preparation of the aldehyde-containing triglyceride,
epoxidized high-oleic sunflower oil 15 was used as starting
material (Scheme 4).
Soxhlet extraction with dichloromethane (Table 1) gave the
higher soluble fraction for sample III (20.5%) and the lower
for sample V (13.0%). The fact that the soluble part is lower
for the higher crosslinking temperature also supports the pro-
posed crosslinking mechanism taking place after the initial
fragmentation. However, the samples undergo some weight
loss during the crosslinking reaction because of evaporation of
the free aliphatic methyl ketones released in the retro-Mannich
fragmentation. Samples III, IV, and V lost 11, 15, and 15% of
its total weight, respectively, but anyway, these slight differen-
ces do not justify the differences in soluble parts.
Another evidence of the increase in the aromatic content can
be found in the thermogravimetric analysis (TGA) under N2
FIGURE 6 TGA plots under nitrogen for the cured samples III–
VIII.
876
INTERSCIENCE.WILEY.COM/JOURNAL/JPOLA

SCHEME 4 Synthesis of the aldehyde-containing triglyceride (17) and BF3ꢀMEA-catalyzed crosslinking with DDM.
The epoxide groups were hydrolyzed in the presence of
proportion depending on the properties of the final materials. As
the reaction between 2 and 17 proceeds very fast, to properly ho-
mogenize the curing mixture, both components were dissolved in
THF separately. BF3ꢀMEA was added to the DDM solution, and then
both solutions were mixed. The mixture was placed in a Petri dish
and maintained at 40 ꢁC for 2 ꢁh to eliminate all the solvent. The
temperature was raised to 90 C for 4 h, and finally the sample
was heated at 140 ꢁC for 12 h to promote the cyclization and aro-
matization reactions. In this way, three samples were prepared
with 17/2 ratios of 1/0.75 (sample VI), 1/1 (sample VII), and 1/
1.25 (sample VIII) with 3 mol % of BF3ꢀMEA (related to aldehyde
groups).
perchloric acid to give the triglyceride 16 with 2.5 diol
1
groups (determined by H NMR spectroscopy). The oxidative
cleavage of the 1,2-diols was performed with sodium period-
ate to obtain a mixture of aldehyde-containing triglyceride
17 and nonanal. The latter was removed by applying high
ꢁ
vacuum to the stirred mixture at 80 C, and 17 was obtained
as slightly yellow oil with 2.5 aldehyde groups per triglycer-
1
ide (determined by H NMR spectroscopy).
In the model reaction of p-toluidine and decanal [see Scheme
3(B)], the formation of the quinoline structure requires two mole-
cules of aldehyde for one of amine. However, a second amine mole-
cule is needed as catalyst. Normally, a little excess of amine over
the 2/1 ratio could be enough, but in the case of crosslinking reac-
tions, where reactants diffusion is limited, a higher amount of
amine may be needed. For this reason, three different 17/2 ratios
were used for the crosslinking reactions to determine the optimum
The crosslinking of 17/2/BF3ꢀMEA system was monitored by
UV–vis and fluorescence spectroscopy. The measurements
were performed with a 20-lm-thick film supported on a
quartz plate, and to minimize DDM absorption interferences,
the 17/2 ratio used was 1/0.75. Figure 3(B) shows the UV–

JOURNAL OF POLYMER SCIENCE: PART A: POLYMER CHEMISTRY DOI 10.1002/POLA
REFERENCES AND NOTES
vis spectra as a function of cure temperature in which the
development of the absorption bands is similar to that of the
1/2/BF3ꢀMEA system. The appearance of an absorption band
around 370 nm at high temperatures can be related with the
presence of quinolines. The fluorescence emission spectra as
a function of the cure temperature with excitation wave-
lengths at 255 and 325 nm are depicted in Figure 4(C,D),
respectively. The spectra are similar to those of the 1/2/
BF3ꢀMEA system and therefore the same conclusions can be
extracted. The bathochromic shift observed in both spectra
supports the formation of quinolines.
1 Williams, C. K.; Hillmyer, M. A. Polym Rev 2008, 48, 1–10.
2 Feedstocks for the Future: Renewable for the Production of
Chemicals and Materials, ACS Symposium Series 921; Bozell, J. J.;
Patel, M., Eds.; American Chemical Society: Washington, DC, 2006.
3 Biermann, U.; Friedt, W.; Lang, S.; Luhs, W.; Machmuller, G.;
¨
¨
Metzger, J. O.; Klaas, M. R.; Schafer, H. J.; Schneider, M. P.
¨
Angew Chem Int Ed 2000, 39, 2206–2224.
4 Guner, F. S.; Yagci, Y.; Erciyes, A. T. Prog Polym Sci 2006,
¨
31, 633–670.
5 Meier, M. A. R.; Metzger, J. O.; Schubert, U. S. Chem Soc
The analysis of the tan d peaks of samples VI–VIII shows as
Rev 2007, 36, 1788–1802.
the main difference with respect to sample V a shoulder
ꢁ
6 Khot, S. N.; LaScala, J. J.; Can, E. S.; Morye, S. G.; Williams,
around 10 C due to the fragmentation products, which act as
I.; Palmese, G. R. S.; Kusefoglu, H.; Wool, R. P. J Appl Polym
¨
plasticizers in sample V. On the other hand, the tan d peaks of
samples VI–VIII are narrower, indicating a higher structural
homogeneity. As the initial content of 2 is increased, the max-
ima of the tan d peak shift to higher temperaturꢁes because of
the higher aromatic content, reaching a Tg of 84 C for sample
VIII. However, the peak height is similar for the three samples,
indicating that a similar crosslinking density is achieved even
when the amount of amine groups is increased. The diffusion
of amines in the curing mixture is a limiting factor, but adding
an excess does not seem to solve the problem provided that
more amine groups will remain unreacted though quinoline
formation might be favored. To further confirm this fact, the
Sci 2001, 82, 703–723.
7 Lane, B. S.; Burgess, K. Chem Rev 2003, 103, 2457–2474.
8 Grigoropoulou, G.; Clark, J. H.; Elings, J. A. Green Chem
2003, 5, 1–7.
9 Uyama, H.; Kuwabara, M.; Tsujimoto, T.; Kobayashi, S. Bio-
macromolecules 2003, 4, 211–215.
10 Esen, H.; Kusefoglu, S. H. J Appl Polym Sci 2003, 89,
¨
3882–3888.
11 Guo, A.; Javni, I.; Petrovic, Z. S. J Appl Polym Sci 2000, 77,
467–473.
1
soluble fractions of samples VI–VIII were analyzed by H NMR.
12 Kaplan, D. L. Biopolymers from Renewable Resources;
Although DDM was found to be one of the main components
of the soluble fraction of sample V, it was not found in the
soluble part of samples VI and VII. It is also worth to mention
the low soluble fraction weight percentages found in samples
VI–VIII when compared with those of samples III–V.
Springer: Berlin, 1998; p 267.
13 Park, S.-J.; Jin, F.-L.; Lee, J.-R. Macromol Rapid Commun
2004, 25, 724–727.
14 Earls, J. D.; White, J. E.; Lo´ pez, L. C.; Lysenko, Z.; Dettloff,
M. L.; Null, M. J. Polymer 2007, 48, 712–719.
Figure 6 depicts the thermal stability of samples VI–VIII
showing similar behavior to that of samples III–V. Char
yields are found around 16%, slightly over sample V (14%)
because of the higher aromatic content.
15 Miyagawa, H.; Misra, M.; Drzal, L. T.; Mohanty, A. K. Polym
Eng Sci 2005, 45, 487–495.
16 Miyagawa, H.; Mohanty, A. K.; Misra, M.; Drzal, L. T. Macro-
mol Mater Eng 2004, 289, 636–641.
CONCLUSIONS
17 Ahmad, S.; Naqvi, F.; Sharmin, E.; Verma, K. L. Prog Org
Coat 2006, 55, 268–275.
The crosslinking reaction of 1 and DDM is a complex process,
which depends on the reaction conditions. First the aza-Mi-
chael adduct is formed. However, high temperatures and the
presence of BF3ꢀMEA promote a retro-Mannich type fragmen-
tation followed by self-aldol condensation, cyclization, and aro-
matization reactions. As a result, quinoline structures are
formed as crosslinking points into the polymeric network.
Based on these findings, a triglyceride functionalized with
aldehyde groups has been reacted with DDM as an improved
approach to quinoline-containing materials. The properties
observed in the cured materials support the proposed cross-
linking process. Although common plant oil-derived materials
do not have the necessary rigidity and strength required for
structural applications, the obtained materials present
improved properties reaching Tg values up to 84 ꢁC.
18 Lligadas, G.; Ronda, J. C.; Galia`, M.; Ca´diz, V. J Polym Sci
Part A: Polym Chem 2006, 44, 6717–6727.
19 Montero de Espinosa, L.; Ronda, J. C.; Galia`, M.; Ca´diz, V. J
Polym Sci Part A: Polym Chem 2008, 46, 6843–6850.
20 Mather, B. D.; Viswanathan, K.; Miller, K. M.; Long, T. E.
Prog Polym Sci 2006, 31, 487–531.
21 Wu, C. S.; Liu, Y. L.; Chiu, Y. S. Polymer 2002, 43, 1773–1779.
22 Lu, Y.; Larock, R. C. ChemSusChem 2009, 2, 136–147.
23 Mihelich, E. D.; Eickhoff, D. J. J Org Chem 1983, 48, 4135–4137.
24 Lim, S.-G.; Jun, C.-H. Bull Korean Chem Soc 2004, 25, 1623–1624.
25 Layer, R. W. Chem Rev 1963, 63, 489–510.
26 Denmark, S. E.; Venkatraman, S. J Org Chem 2006, 71,
1668–1676.
The authors express their thanks to MICINN (Ministerio de
Ciencia e Innovacio´n) (MAT2008-01412) for financial support
for this work.
27 Tanaka, S.-Y.; Yasuda, M.; Baba, A. J Org Chem 2006, 71,
800–803.
878
INTERSCIENCE.WILEY.COM/JOURNAL/JPOLA