Magnetic-based picolinaldehyde–melamine copper complex for the one-pot synthesis of hexahydroquinolines via Hantzsch four-component reactions

Article abstract of DOI:10.1002/aoc.4151

A picolinaldehyde–melamine copper complex was loaded on a magnetic Fe₃O₄ core, so that it contained 0.33?mmol of Cu per gram, and was used as an efficient catalyst. The as-synthesized catalyst was characterized using various techniqu

Full text of DOI:10.1002/aoc.4151

Received: 21 August 2017
DOI: 10.1002/aoc.4151
Revised: 25 September 2017
Accepted: 26 September 2017
F U L L P A P E R
Magnetic‐based picolinaldehyde–melamine copper complex
for the one‐pot synthesis of hexahydroquinolines via
Hantzsch four‐component reactions
Ardeshir Khazaei
| Maryam Mahmoudiani Gilan | Negin Sarmasti
Faculty of Chemistry, Bu‐Ali Sina
University, Hamedan 6517838683, Iran
A picolinaldehyde–melamine copper complex was loaded on a magnetic Fe O
3 4
core, so that it contained 0.33 mmol of Cu per gram, and was used as an
efficient catalyst. The as‐synthesized catalyst was characterized using various
techniques, including Fourier transform infrared spectroscopy, X‐ray diffrac-
tion, energy‐dispersive X‐ray spectroscopy, field emission scanning electron
microscopy, transmission electron microscopy, vibrating sample magnetometry
and thermogravimetric analysis. The catalyst was used to activate the raw
materials in the synthesis of hexahydroquinoline derivatives in one‐pot
four‐component reactions. Low reaction time (minutes versus half an hour),
solvent‐free condition and magnetically separable catalyst are some salient
features of the developed catalyst. Also, the optimum amount of catalyst and
temperature were determined as 0.07 g and 87.6 °C, respectively, which were
obtained using response surface methodology and optimization techniques.
Correspondence
Ardeshir Khazaei, Faculty of Chemistry,
Bu‐Ali Sina University, Hamedan
6517838683, Iran.
Email: khazaei_1326@yahoo.com
Funding information
development of chemical methods,
reagents and molecules, Grant/Award
Number: 32‐1716; Center of Excellence in
Development of Chemical Method
(
CEDCM), Hamedan, Iran; Research
Affairs Office of Bu‐Ali Sina University,
Grant/Award Number: 32‐1716
KEYWORDS
design of experiment, hexahydroquinolines, one‐pot reaction, picolinaldehyde–melamine copper
complex
1
| INTRODUCTION
followed by further research, so that now it is a rapidly
advancing area, as apparent from several recently pub-
Given the number of substances involved in a reaction,
reactions in organic chemistry fall into three main catego-
ries, namely one‐component reactions, two‐component
reactions and one‐pot multicomponent reactions
lished works. Some prototypic examples of MCRs are
[
4]
Mannich three‐component reaction (3‐CR), Strecker
[
5,6]
[7,8]
[9]
3‐CR,
Passerini 3‐CR,
Ugi 3‐CR, Pauson–Khand
[
10]
[11]
[12]
3‐CR, Van Leusen 3‐CR, Gewald 3‐CR, Hantzsch
[1]
[13]
[14]
(MCRs). In MCRs, three or more starting materials are
3‐CR
and Staudinger 3‐CR.
In 1881, Hantzsch
involved in the reaction, so that the majority of the atoms
of the raw materials are incorporated in the final prod-
synthesized dihydropyridine (DHP) which is the one of
[
13]
the most important and conventional MCRs.
Due to
[2]
uct. MCRs are also called domino reactions or tandem
reactions. The first research in this area is attributed to
Gerhard and Laurent who synthesized cyanohydrin
the diversity and stability of the products, the Hantzsch
approach has remained the most widely accepted
methodology for the synthesis of DHPs. The various
biological properties of DHPs, for instance nifedipine,
[
2]
imines from bitter almond oil and ammonia,
and
seminal work was carried out by Strecker (synthesis of
nicardipine, and amlodipine for the treatment of
[3]
[15,16]
α‐amino acid from NH , aldehyde and HCN). In 1890,
hypertension,
have motivated researchers to
3
Hantzsch synthesized the major constituents of
haemoglobin and chlorophyll using MCRs. This was
synthesize novel DHP derivatives which are more
[2]
[17,18]
effective and with novel modes of action.
Regarding
Appl Organometal Chem. 2017;e4151.
wileyonlinelibrary.com/journal/aoc
Copyright © 2017 John Wiley & Sons, Ltd.
1 of 12
https://doi.org/10.1002/aoc.4151

2
of 12
KHAZAEI ET AL.
this issue, Safak's group, by incorporating a fused ring on
the DHP structure, synthesized hexahydroquinolines
electron microscopy (FESEM) and transmission electron
microscopy (TEM) images were recorded with a SIGMA
VP‐500 (Zeiss) and Zeiss‐EM10C‐100 KV, respectively.
Elemental analysis of the as‐synthesized catalyst was
conducted using energy‐dispersive X‐ray analysis (EDX;
Oxford Instruments). A vibrating sample magnetometer
(model MDKB) was used to determine the saturation
magnetization of the catalyst. To monitor reaction
progress, TLC (using silica gel SIL G/UV 254 plates)
was used.
[
19]
(
HHQs) as novel derivatives.
spread pharmacological properties such as vasodilatory,
bronchodilatory, anti‐atherosclerotic, anti‐tumour,
geroprotective, hepatoprotective and anti‐diabetic
The HHQs have wide-
[20–22]
activities.
Analysing the literature shows that the
published articles in this area are increasing, with 212
papers hitherto. Synthesis of novel HHQ derivatives via
efficient approaches and evaluation of the pharmacologi-
cal properties are the main aims of the research.
Considering the above discussion, various catalysts such
[
23]
[24]
[25]
as γ‐Fe O ,
SBA‐15/SO H nanoreactor,
KH PO ,
2.2 | Synthesis of catalyst
2
3
3
2
4
[
26]
sulfamic acid‐functionalized nano‐titanium dioxide,
2
.2.1 | Synthesis of Fe O nanoparticles
[27]
[28]
[29]
3 4
nano‐CoAl O ,
nano‐ZrO ‐SO H,
ZrOCl ⋅8H O,
2
4
2
3
2
2
[
30]
silica‐bonded imidazolium–sulfonic acid chloride,
Magnetic Fe O4 nanoparticles were synthesized using
3
[31]
[39]
nano‐Fe O –TiO –SO H,
1,3‐disulfonic
amino alcohol,
acid
the co‐precipitation method.
First, 11.3
g
of
3
4
2
3
[32]
[33]
imidazolium hydrogen sulfate,
basic
FeCl ⋅6H O and 5.6 g of FeCl ⋅4H O were dissolved in
3
2
2
2
[
34]
isoreticular metal–organic framework (IRMOF‐3),
deionized water at 80 °C with mixing (600 rpm). After
[
35]
[36]
SBA‐SO H,
[MPIm][HSO ]@SBA‐15,
SBA‐15/
and ionic liquids grafted onto graphene
have been used in the synthesis of HHQs and
complete dissolution of the salts, 25 ml of ammonia
3
4
[
37]
[40]
NHSO H
(28 wt%) was added rapidly to the solution.
The
3
38]
[
oxide
resulting black solution was vigorously stirred for 2 h
at 80 °C under nitrogen atmosphere. Then the precipi-
tated magnetic nanoparticles were separated from the
solution using a magnet. They were washed several
times with water and then with acetone. Finally a stable
black magnetic dispersion was obtained and dried in
an oven.
DHPs. Although impressive successes have been
achieved, most of the developed catalysts suffer from long
reaction time (half an hour to several hours), and not so
simple procedures for separation of the catalyst. So, there
is still a great desire for the synthesis of novel catalysts
which are efficient, feasible, high‐yielding and with low
reaction times.
In the light the above discussion, in the work pre-
sented here, we extended the synthetic applicability of
an as‐prepared catalyst (magnetic‐based picolinaldehyde–
melamine copper complex) in the synthesis of HHQs in
one‐pot reactions and under solvent‐free condition with
low reaction times.
2.2.2 | Synthesis of magnetic Fe
nanoparticles
O @SiO
3 4 2
An amount of 1 g of the Fe O4 nanoparticles was
3
dispersed in water and ethanol mixture (80:20 v/v) using
an ultrasound bath (15 min). Then 2 ml of ammonia
(
28 wt%) and 2 ml of tetramethyl orthosilicate were
added dropwise to the mixture and stirred at 50 °C for
h. The obtained Fe O @SiO nanoparticles were
2
2
| EXPERIMENTAL
2
3
4
2
.1 | Materials and apparatus
separated using a magnet and then washed with water
and ethanol successively, and dried in an oven.
All chemicals were supplied from Merck and Fluka. The
1
13
products were characterized using H NMR and
NMR spectroscopy, melting points as well as infrared
IR) spectroscopy. The relevant spectral data are given
C
2
.2.3 | Synthesis of Fe O @SiO @Si‐
3
4
2
(
(
CH ) Cl nanoparticles
2 3
1
13
in the supporting information. The H NMR and
C
NMR (500 MHz) spectra were recorded with a Bruker
Avance DPX‐400 FT‐NMR spectrometer (δ in ppm).
Melting points were measured using a Büchi B‐545
apparatus in open capillary tubes. A PerkinElmer PE‐
An amount of 1 g of the previously synthesized
Fe O @SiO nanoparticles was dispersed in 50 ml of
3
4
2
toluene (15 min). Then 2 ml of (3‐chloropropyl)
triethoxysilane was gently added to the mixture and
refluxed at 110 °C for 24 h. The synthesized nanoparticles
were separated using a magnet and washed several times
with toluene and water. Finally the product was dried in
an oven.
1
600‐FTIR spectrometer was used to record the IR spec-
tra of products. To measure the mass reduction of the
catalyst, thermogravimetric analysis (TGA) under a
nitrogen atmosphere was used. Field emission scanning

KHAZAEI ET AL.
3 of 12
2
.2.4 | Synthesis of Fe O @SiO @Si‐
3
4
2
(CH ) @melamine nanoparticles
2 3
Amounts of 12 mmol (1.51 g) of melamine and 12 mmol
of K CO3 as a base were dissolved in 70 ml of
2
dimethylsulfoxide (DMSO), and then 1 g of the
Fe O @SiO @Si‐(CH ) Cl nanoparticles was added to
3
4
2
2 3
the mixture and dispersed (15 min). After refluxing the
mixture for 24 h, the nanoparticles were separated using
magnet, washed with 50 ml of DMSO and dried in
an oven.
2
.2.5 | Synthesis of Fe O @SiO @Si‐
3
4
2
(CH ) @melamine@picolinaldehyde
2 3
nanoparticles
First, 1 g of the Fe O @SiO @Si‐(CH ) @melamine
3
4
2
2 3
nanoparticles was dispersed in 50 ml of methanol
(15 min), and refluxed. An amount of 24 mmol of
picolinaldehyde (2.57 g) was dissolved in 30 ml of metha-
nol, and then was added dropwise to the nanoparticles
using a decanter funnel. After refluxing for 24 h, the
product was separated using a magnet, washed with
5
0 ml of methanol and dried in an oven.
2
.2.6 | Synthesis of Fe O @SiO @Si‐
3
4
2
FIGURE 1 Total procedure for synthesis of catalyst
(CH ) @melamine@picolinaldehyde@Cu
2 3
nanoparticles
mixture, so that all of the components were dissolved,
and because of the insolubility of the catalyst, it was
recovered with a super magnet. Finally the product
was purified by crystallization from the added ethyl
acetate.
A solution of Cu(OAc)2 (5% w/v) was prepared in
methanol (solution A), and then 1 g of the nanoparticles
synthesized in the previous step was dispersed in 50 ml
of methanol (15 min). Solution A was added dropwise to
the dispersed nanoparticles, and the resulting mixture
refluxed for 48 h. After the completion of the reaction,
the nanoparticles were separated using a super magnet,
and washed several times with methanol. The final
product was dried in an oven. The overall scheme for
the synthesis of the nanoparticles and catalyst is shown in
Figure 1.
2.4 | Retrieval of catalyst
To investigate the catalytic activity versus the number of
reaction cycles, the benchmark reaction was carried out.
After the complete progress of the reaction, 20 ml of hot
ethyl acetate was added to the reaction mixture to
dissolve all the raw material and final product. To can-
cel out the reaction time, all of the cycles were ended
after 6 min. Then, the catalyst was separated with a
super magnet. The resulting solution without any
work‐up was used for crystallization of the product.
The catalyst was dried in an oven and utilized in the
next cycle. Figure 2 shows the reaction yield versus
number of cycles. As can be seen, the reaction yield
does not change significantly, which implies that the
catalytic activity remains almost unchanged. It should
be noted that all reaction cycles were performed at opti-
mum conditions.
2.3 | General procedure for synthesis of
HHQs
Using the optimum benchmark reaction conditions,
namely 0.06 g of as‐prepared catalyst and at 50 °C,
dimedone (0.28 g, 2 mmol), aryl aldehyde (2 mmol), β‐
ketoester (2 mmol) and ammonium acetate (0.185 g,
2
.4 mmol) were added to a test tube, and then the mix-
ture was stirred magnetically. TLC was used to probe the
reaction progress, and after completion of the reaction,
the mixture was cooled to room temperature. Then,
2
0 ml of hot ethyl acetate was added to the reaction

4
of 12
KHAZAEI ET AL.
−
1
1
00
and then it shifts from 1699 to 1568 cm . These bands
imply the formation of the complex between
picolinaldehyde, melamine and copper.
8
6
4
2
0
0
0
0
0
Vibrating sample magnetometry was used to further
verify the loading of various layers on the nanoparticles.
The saturation magnetization curves of Fe O ,
and Fe O @SiO @Si(CH ) @melamine
3 4 2 2 3
picolinaldehyde@Cu nanoparticles are shown in
3
4
Fe O @SiO
3
4
2
@
Figure 4. As one can see, the saturation magnetization
decreases (54, 50 and 43 emu g for Fe O , Fe O @SiO
2
−
1
3
4
3
4
and as‐prepared catalyst, respectively) with an increasing
number of loaded layers on nano‐Fe O . It has been
3
4
shown that non‐magnetic materials reduce the saturation
magnetization because of the quenching of surface
1
2
3
4
[
44]
Number of cycles
moments,
and this behaviour is seen clearly in
Figure 4. It should be noted that, although the magnetiza-
tion of the as‐prepared catalyst is reduced, it is enough
FIGURE 2 Benchmark reaction yield versus number of cycles
3
3
| RESULTS AND DISCUSSION
.1 | Characterization of catalyst
60
40
20
0
The IR spectra of the intermediate nanoparticles and final
catalyst are shown in Figure 3. The band at 1083.24 cm
belongs to SiO of the SiO group in the Fe O @SiO nano-
−
1
2
3
4
2
−
1
particles, and that at 2946.4 cm is related to CH of
2
propyl chloride in Fe O @SiO @Si(CH ) Cl nanoparti-
cles. The band at 3282.3 cm is an indication of the pres-
3
4
2
2 3
-20
−
1
ence of NH group of melamine, confirming the loading
-40
2
of melamine on the surface of Fe O @SiO @Si(CH ) Cl.
It has been shown that melamine with three nitrogen
atoms can well coordinate with metal ions such as cop-
3
4
2
2 3
-
60
-
1.5
-1
-0.5
0
0.5
1
1.5
[
41–43]
−1
per.
The band at 1428 cm corresponds to the
−
−1
CH COO group, and that at 497 cm is attributed to
FIGURE
4
3 4
Saturation magnetization curves for Fe O ,
3
[41]
−1
O―Cu
and C¼N. The band at 1568 cm can arise
Fe
Fe
3
O
O
4
@SiO
@SiO
2
2
and
2 3
@Si(CH ) @melamine@picolinaldehyde@Cu.
from the complexation of copper ions to the imine group
3
4
3 4 2 2 3
FIGURE 3 IR spectra of intermediate nanoparticles and as‐prepared catalyst (Fe O @SiO @Si‐(CH ) @melamine@picolinaldehyde@Cu)

KHAZAEI ET AL.
5 of 12
such that it can be easily retrieved from the reaction
mixture.
FESEM and TEM images of the as‐prepared
nanomagnetic catalyst are shown in Figure 5. These
techniques were used to identify the size, shape and mor-
phology of micro‐ and nanoparticles. As can be seen from
the FESEM images (Figure 5a,b), the geometric shape of
the nanoparticles is spherical. Besides the spherical shape,
the TEM images (Figure 5c,d) show the catalyst has a
core–shell structure. The shell is attributed to the various
layers loaded on the Fe O surface. Figure 6 shows the
3
4
normal probability and cumulative particle size distribu-
tion (experimental and fitted) of the as‐prepared catalyst.
The experimentally determined curve in the normal prob-
ability plot (Figure 6a) is fairly straight, so the normality
of the particle sizes could be a reasonable assumption.
The cumulative distribution function (Figure 6b) shows
that 60 and 90% of the nanoparticles have sizes of 30
and 40 nm, respectively. The data collected from FESEM
images show that the mean and the standard deviation
of sizes of as‐synthesized catalyst nanoparticles are 29.45
and 7.25 nm, respectively.
The size, purity and crystalline nature of the particles
can be identified with the X‐ray diffraction (XRD) tech-
nique. The XRD patterns of Fe O , Fe O @SiO ,
3
4
3
4
2
Fe O @SiO @Si(CH ) Cl and Fe O @SiO @Si(CH )
2 3
3
4
2
2 3
3
4
2
@melamine@picolinaldehyde@Cu nanoparticles are
shown in Figure 7. To calculate the as‐synthesized cata-
lyst particle size, the Scherrer equation (D =Kλ/βcos θ)
was used, where K, λ, β and θ correspond to dimension-
less shape factor, X‐ray wavelength, full width at half
maximum (FWHM) of the diffraction peak and Bragg dif-
fraction angle in radians, respectively. It was supposed
that the nanoparticles have a spherical geometric shape,
so that the K value is set to 0.9 in the Scherrer equation.
For the X‐ray wavelength (λ), 0.154184 nm was used
which is the averaged value (Kα λ = 0.1540598 nm and
1
Kα λ = 0.154433 nm) of copper Kα. The values of β and
2
θ are determined from XRD patterns. The values of θ,
FWHM and calculated particle size are given in Table 1.
Using the Scherrer equation, 9.86 nm was obtained as
the mean value of particle size. There is a discrepancy
between the mean particle size from FESEM (29.45 nm)
and XRD (9.86 nm) techniques. This difference is attrib-
uted to the inability to measure small particles in FESEM
and the K value in the Scherrer equation.
In addition, the Bragg equation (dhkl =λ/2 sin θ) was
used to identify the distance between the layers (d‐spacing
or interplaner spacing) in the crystalline structure of
as‐prepared catalyst. The d‐spacing values at different
diffraction angles are given in Table 1.
The EDX technique was used to determine the contri-
bution of each element of the sample. The elemental
FIGURE 5 (a, b) FESEM and (c, d) TEM images of as‐synthesized
catalyst

6
of 12
KHAZAEI ET AL.
0
.997
24.6, 17.7, 5.4, 3.5 and 2.1%, respectively. The contribution
of Cu shows that 0.33 mmol of Cu per gram of catalyst
0
.99
0
.98
(a)
−
1
(
0.33 mmol g ) has been coordinated with the melamine
0
0
.95
.90
and picolinaldehyde. The Au peak in Figure 8 is due to
the coating of the sample with Au in the procedure of
sample preparation. There are 5 mmol of N per gram of
catalyst, which is approximately 15 times that of Cu. So
it can be inferred that half (52.8%) of the picolinaldehyde
and melamine have been coordinated with Cu.
0
0
0
.75
.50
.25
0
0
0
.10
.05
The thermal stability of the catalyst has a great effect
on its catalytic activity, because the optimum reaction
temperature is approximately 88 °C. The TGA curve of
the as‐prepared catalyst is shown in Figure 9. The first
mass loss up to 150 °C (region A) is due to the release of
adsorbed water; the second from 200 to 550 °C is related
to the decomposition of organic matter on the
Fe O @SiO core (region B). Region C corresponds to
.02
0
.01
0.003
2
0
30
40
50
1
0
0
0
0
0
0
0
0
0
.9
.8
.7
.6
.5
.4
.3
.2
.1
0
(
b)
3
4
2
Fe O @SiO . The TGA curve of the catalyst system dem-
3
4
2
onstrates high thermal stability, with decomposition
starting at around 200 °C under a nitrogen atmosphere.
3.2 | Effect of solvent on reaction
Exp. particle size
Fitted particle size
To determine the effect of solvent type on the reaction
yield and time, the benchmark reaction (reaction of benz-
aldehyde, dimedone, ethyl acetoacetate and ammonium
acetate) was carried out (under the optimized reaction
conditions) in various solvents, namely ethyl acetate,
dimethylformamide (DMF), water, n‐hexane and ethanol,
and under solvent‐free condition. Results are given in
Table 2. The reaction under solvent‐free condition, with
20
30
40
50
FIGURE 6 (a) Normal probability plot and (b) cumulative
probability plot of as‐synthesized nanoparticle size
12 min and 85% for reaction time and yield, respectively,
is more efficient than reaction with solvent. This can be
attributed to the presence of ethyl acetoacetate, which is
liquid at room temperature and in the reaction, so that
it can play the role of solvent.
3.3 | Experimental design
The yield and time of an organic reaction are affected by
numerous factors such as temperature, type and amount
of catalyst, type of solvent, and so on. Because of complex
interactions between factors, the effect of the different fac-
tors cannot be modelled using rigorous mathematical
relations. In this situation, statistics can be useful, so that
one can perturb the effective variable to measure the
response. The conventional method which is used by
many chemists is ‘change one variable at a time’ (OVAT)
or interchangeably ‘one factor at a time’. If the is a
considerable interaction between parameters the OVAT
procedure is not efficient, so that it gives an unrealistic
FIGURE 7 XRD patterns of intermediate nanoparticles and as‐prepared
catalyst (Fe O @SiO @Si‐(CH ) @melamine@picolinaldehyde@Cu)
3 4 2 2 3
analysis of the as‐synthesized catalyst together with the
contribution of each element is shown in Figure 8. The
mass percentages of Fe, O, C, Si, N and Cu are 46.7,
[
29,45,46]
optimal point.
The large number of experiments
is another relevant feature of this method, so it is not

KHAZAEI ET AL.
7 of 12
3 4 2 2 3
TABLE 1 XRD data for as‐synthesized catalyst (Fe O @SiO @Si(CH ) @melamine@picolinaldehyde@Cu)
Entry
2θ (°)
Peak width, FWHM (rad)
Size (nm)
Interplanar spacing of crystal (nm)
1
9.2
1.111
7.16
0.96
2
30.3
35.7
43.5
57.3
63.0
—
0.789
0.732
0.657
0.852
0.825
—
10.8
10.9
12.1
9.33
9.64
9.86
0.29
0.25
0.21
0.16
0.15
0.34
3
4
5
6
Mean
TABLE 2 Effect of solvents, and no solvent, on the reaction time
and yield
a
Entry
Solvent
Time (min)
Yield at 60° (%)
1
—
12
85
2
3
4
5
6
EtOAc
DMF
>100
>100
>100
>100
90
52
55
43
35
50
2
H O
n‐Hexane
EtOH
a
Isolated yield.
FIGURE
8
Elemental analysis of as‐prepared catalyst
(
Fe
3
O
4
@SiO
2
@Si(CH @melamine@picolinaldehyde@Cu) using
2 3
)
methodology, the response surface method (RSM) is the
most used one. The most well‐known RSM is central com-
posite design which is used in the present work to find the
optimal conditions in the synthesis of HHQs.
EDX
Briefly, the amount of catalyst (X ) and temperature
1
(X2) were considered as the effective variables on the
reaction yield (Y) at a specified reaction time. The factors
X and X were coded to three levels (−1, 0, 1) in face‐
1
2
centred composite design. Three replicates at centre point
were used to probe the curvature in response (yield, Y), if
there is one. The levels of the variables are given in
Table 3. The benchmark reaction (reaction of benzalde-
hyde, dimedone, ethyl acetoacetate and ammonium ace-
tate) was carried out in these designed experimental
runs, with the reaction being stopped after 4 min to cancel
out the effect of time on the reaction yield. The reaction
yields are given in Table 3. It should be noted that the
given yields ate the crystallization yields in ethyl acetate
up to ambient temperature as the minimum temperature.
FIGURE
Fe @SiO
9
TGA curve
of
as‐prepared
catalyst
(
3
O
4
2
2 3
@Si(CH ) @melamine@picolinaldehyde@Cu)
efficient from material, time and economic points of
3
.4 | Optimization of reaction conditions
[47,48]
view.
Unfortunately, many researchers probes their
and validation of model
optimal conditions through OVAT methodology.
Recently, multivariate statistical approaches, namely
design of experiment, have been used to overcome the
drawbacks of the OVAT method. Although various
approaches are used in the design of experiment
From a mathematical point of view, the relation between
effective variables and dependent variable (response) can
be expressed using a Taylor expansion series.
[
49]
In the
present work, a second‐order Taylor expansion series

8
of 12
KHAZAEI ET AL.
TABLE 3 Levels of effective variables and corresponding response values (yield) using face‐centred composite design method
Effective variable
X
1
(amount of catalyst)
2
X (temperature)
Response
(yield)
Run
Coded level
Actual level
Coded level
Actual level
1
0
0.0355
−1
30.00
50
70
38
57
85
83
71
72
71
56
65
2
3
4
5
6
7
8
9
0
−1
−1
+1
+1
0
0.0355
0.0010
0.0010
0.0700
0.0700
0.0355
0.0355
0.0355
0.0010
0.0700
+1
−1
+1
0
110.00
30.00
110.00
70.00
110.00
70.00
70.00
70.00
70.00
30.00
+1
0
0
0
0
0
1
0
1
−1
+1
0
1
−1
was used:
Y ¼ β þ β x1 þ β x2 þ β x1x2 þ β x þ β x þ ε (1)
TABLE 4 Analysis of variance (ANOVA) based on quadratic
polynomial model for yield of reaction
2
1
2
0
1
2
12
11
22
2
Source
Probability value (p‐value)
where β ,…, β stand for the coefficients of the model and
0
22
Model (yield)
<0.0001
x and Y explain the effective variables and response. The
matrix notation of the model is given as:
X
1
2
<0.0001
<0.0001
0.6410
1.0000
<0.0001
0.9974
0.9787
0.1901
X
Y ¼ βX þ ε
(2)
X1X2
2
X
1
To find the best value of the coefficient, the method of
2
[50]
X2
least squares was used:
2
R
À
Á
−
1
T
T
2
β ¼ X X X Y
(3)
Predicted R
Lack of fit
The final polynomial model for the reaction yield
based on the experimental runs and method of least
squares is as follows:
Table 4. The value of 0.9787 shows that the model could
well predict the reaction yield, with new values of X and
1
Y ¼ 7:70 þ 408:82X1
X2 which are not present in Table 3 (temperature and the
amount of catalyst which didn't used in the model devel-
opment procedure).
(
4)
−
3
2
2
þ 1:16X −0:18X X −6:56×10 X
2
1
2
Analysis of variance (ANOVA) showed that based on
the sum of squares and lack of fit, the quadratic model
could well correlate with the experimental data. The
summary of ANOVA is given in Table 4, so that the
probability value (p‐value) is less than 0.0001, which
calls for the rejection of the null hypothesis. Although
the p‐value of the model parameters shows that the
Figure 10 shows graphically the power of the model in
correlation and prediction of the reaction yield. Any devi-
ation from the diagonal line indicates the inability of
model to correlate or predict the experimental data or
the error of the experimental yield data.
Figure 11 shows the three‐dimensional surface and the
contours of reaction yield versus amount of catalyst and tem-
perature. As one can see, an increased yield, Y, is achieved
2
X X and X terms should be cancelled out from equa-
1
2
1
tion 4, the model is significant and can be used for fur-
ther analysis. Although the value of R (0.9974) shows
through an increase in the amount of catalyst, X
, with a pos-
1
2
itive value of β in equation 4. As the temperature increases,
1
that the provided model could well correlate the experi-
mental reaction yields, the predictive power of the model
must be evaluated, which is given as predicted R in
the reaction yield first increases to a maximum value, and
then decreases. An increase in the yield is likely to be attrib-
uted to the temperature dependence of the reaction kinetics,
2

KHAZAEI ET AL.
9 of 12
and a decrease of the yield is probably related to the loss of
catalytic activity at higher temperatures.
To determine the optimum point (using analytical or
numerical methods) of the yield, equation 4 was used,
with 0.07 g and 87.6 °C being obtained for the amount
of catalyst and temperature, respectively. Although the
predictive power of the model was evaluated using pre-
2
dicted R , it is better evaluate the model at a new point.
So, the yield of the optimum point was obtained in the
laboratory and compared with the results of the model,
which are given in Table 5. The deviation between the
experimental and predicted yield is 4.46%, which is low
enough for a predictive model. Accordingly, the obtained
results are compared with those of experimental and a
good agreement between them is noticed.
To further investigate the effect of catalyst on reaction
time, the obtained optimum point was repeated in the lab-
oratory without catalyst. The results are given in Table 6.
As one can see, in the absence of catalyst, even with a
time greater than 30 min, the yield is low (56%). So, the
TABLE 5 Predicted and experimental values of reaction yield at
optimum point
Optimum variable Optimum response
X
1
X
2
Time (min) Yield (%)
Predicted
0.07
87.6
88.0
4
4
86.7
83.0
Experimental 0.07
TABLE 6 Checking optimum condition with and without catalyst
Variable
Response
Condition
X
1
X
2
Time (min)
Yield (%)
2
2
FIGURE 10
obtained model
R and predicted R of reaction yield from the
With catalyst
0.07
88.0
88.0
6
91.0
Without catalyst
—
>30
56.0
FIGURE 11 Surface and contour of reaction yield versus amount of catalyst and temperature

10 of 12
KHAZAEI ET AL.

KHAZAEI ET AL.
11 of 12
4
| CONCLUSIONS
In summary,
a
magnetic‐based picolinaldehyde–
melamine copper complex was synthesized and used in
the synthesis of HHQ derivatives. The synthesized catalyst
contains 0.33 mmol of Cu per gram which can activate the
reaction between starting materials. The as‐synthesized
catalyst has salient features which include: magnetically
separation of the catalyst, solvent‐free condition, green,
benign and considerable short reaction time. Face‐centred
composite design as a response surface methodology was
used to statistically design the experimental runs. Using
the obtained quadratic polynomial model for the reaction
yield, and with the optimization techniques, the true
maximum was achieved. So, the RSM method in this case
is more efficient than OVAT.
FIGURE 12 Plausible mechanism for synthesis of HHQ using as‐
synthesized catalyst
ACKNOWLEDGEMENTS
presence of the catalyst is vital for the progress of the
reaction.
The authors gratefully acknowledge partial support of this
work by the Research Affairs Office of Bu‐Ali Sina
University (grant number 32‐1716, entitled development
of chemical methods, reagents and molecules), Center of
Excellence in Development of Chemical Method
(CEDCM), Hamedan, Iran.
3.5 | Synthesis of HHQs at optimal
conditions
Although the optimum point was achieved through the
benchmark reaction, it was assumed that the optimum
point is fixed for all the benzaldehyde derivatives. Hence,
reactions between various aryl aldehydes, dimedone,
ethyl acetoacetate and ammonium acetate were carried
out at the optimum condition (X1 = 0.07 g and
X2 = 87 °C). Various aryl aldehyde derivatives with elec-
tron‐withdrawing and electron‐donating groups on the
benzene ring were successfully examined. The products
of the reactions together with the experimental melting
points are given in Table 7. Although the reaction yields
are almost identical to those reported in the literature,
for some derivatives such as products 1, 9 and
ORCID
Ardeshir Khazaei http://orcid.org/0000-0001-7990-2266
REFERENCES
[1] T. Zarganes‐Tzitzikas, A. L. Chandgude, A. Dömling, Chem.
Rec. 2015, 15, 981.
[
[
2] I. Ugi, A. Dömling, W. Hörl, Endeavour 1994, 18, 115.
3] Z. Wang, Strecker Synthesis: Comprehensive Organic Name
Reactions and Reagents, John Wiley, New York 2010.
[4] C. Mannich, W. Krösche, Arch. Pharm. 1912, 250, 647.
[26,32,33,51–53]
1
1,
the reaction time is short (minutes versus
[
[
5] A. Strecker, Justus Liebigs Ann. Chem. 1850, 75, 27.
half an hour) and for some products, including 3, 6 and
6] J. T. Kuethe, D. R. Gauthier, G. L. Beutner, N. Yasuda, J. Org.
Chem. 2007, 72, 7469.
[
52,53]
1
2, the reaction time is longer than those reported.
[
[
[
7] M. Passerini, L. Simone, Gazz. Chim. Ital. 1921, 51, 126.
8] S. Rostamnia, RSC Adv. 2015, 5, 97044.
3.6 | Proposed mechanism
9] I. Ugi, Angew. Chem. Int. Ed. Engl. 1962, 1, 8.
A proposed and plausible mechanism based on IR data is
shown in Figure 12. As shown, and based on IR investiga-
tions, the as‐prepared catalyst interacts only with benzal-
dehyde and reduces the carbonyl band wavenumber from
[
10] P. L. Pauson, I. U. Khand, Ann. N. Y. Acad. Sci. 1977, 295, 2.
11] A. M. Van Leusen, J. Wildeman, O. H. Oldenziel, J. Org. Chem.
[
1977, 42, 1153.
[
12] K. Gewald, E. Schinke, H. Böttcher, Chem. Ber. 1966, 99, 94.
13] A. Hantzsch, Ber. Deutsch. Chem. Gesellsch. 1881, 14, 1637.
−
1
1
703.92 to 1687.49 cm . The details of the mechanism,
[
along with the associated intermediates, are presented in
[29,54]
the literature.
[14] H. Staudinger, Justus Liebigs Ann. Chem. 1907, 356, 51.

12 of 12
KHAZAEI ET AL.
[
[
[
15] N. Edraki, A. R. Mehdipour, M. Khoshneviszadeh, R. Miri,
[39] X. Liu, Z. Ma, J. Xing, H. Liu, J. Magn. Magn. Mater. 2004,
270, 1.
Drug Discov. Today 2009, 14, 1058.
16] J. L. Reid, P. A. Meredith, F. Pasanisi, J. Cardiovasc. Pharmacol.
[40] S. Rostamnia, B. Zeynizadeh, E. Doustkhah, A. Baghban, K. O.
1
985, 7, S18.
Aghbash, Catal. Commun. 2015, 68, 77.
17] D. Längle, V. Marquardt, E. Heider, B. Vigante, G. Duburs, I.
Luntena, D. Flötgen, C. Golz, C. Strohmann, O. Koch, D.
Schade, Eur. J. Med. Chem. 2015, 95, 249.
[41] Y.‐P. Wang, Y. Luo, R.‐M. Wang, React. Funct. Polym. 1997,
33, 81.
[
[
[
42] A. B. Wiles, D. Bozzuto, C. L. Cahill, R. D. Pike, Polyhedron
[
18] J. Marín‐Prida, G. L. Pardo Andreu, C. P. Rossignoli, M. G.
Durruthy, E. O. Rodríguez, Y. V. Reyes, R. F. Acosta, S. A.
Uyemura, L. C. Alberici, Toxicol. In Vitro 2017, 42(Suppl. C), 21.
2006, 25, 776.
43] D. M. L. Goodgame, I. Hussain, A. J. P. White, D. J. Williams, J.
Chem. Soc. Dalton Trans. 1999, 16, 2899.
[
[
[
19] C. Safak, R. Simsek, Mini‐Rev. Med. Chem. 2006, 6, 747.
44] H. Xu, N. Tong, L. Cui, Y. Lu, H. Gu, J. Magn. Magn. Mater.
2007, 311, 125.
20] T. Godfraind, R. Miller, M. Wibo, Pharmacol. Rev. 1986, 38, 321.
21] R. Mannhold, B. Jablonka, W. Voigt, K. Schönafinger, E.
[45] M. A. Zolfigol, A. Khazaei, N. Sarmasti, J. Y. Seyf, V.
Schraven, Eur. J. Med. Chem. 1992, 27, 229.
Khakyzadeh, A. R. Moosavi‐Zare, J. Mol. Catal. A 2014,
3
93, 142.
[
22] E. Baydar, M. G. Gündüz, V. S. Krishna, R. Şimşek, D. Sriram,
S. Ö. Yıldırım, R. J. Butcher, C. Şafak, Res. Chem. Intermed.
[46] D. D. Frey, F. Engelhardt, E. M. Greitzer, Res. Eng. Des. 2003,
14, 65.
2
017, https://doi.org/10.1007/s11164‐017‐3087‐0.
[
[
[
[
[
23] S. Rostamnia, A. Nuri, H. Xin, A. Pourjavadi, S. H. Hosseini,
[47] R. Carlson, J. E. Carlson, Design and Optimization in Organic
Synthesis, Elsevier Science, Oxford 2005.
Tetrahedron Lett. 2013, 54, 3344.
24] S. Rostamnia, H. Xin, X. Liu, K. Lamei, J. Mol. Catal. A 2013,
[48] F. Paulo, L. Santos, Mater. Sci. Eng. C 2017, 77, 1327.
3
74(Suppl. C), 85.
25] S. J. Yü, S. Wu, X. M. Zhao, C. W. Lü, Res. Chem. Intermed.
017, 43, 3121.
26] E. Tabrizian, A. Amoozadeh, J. Chinese Chem. Soc. 2017,
4, 331.
[49] D. Baş, İ. H. Boyacı, J. Food Eng. 2007, 78, 836.
[50] T. F. Edgar, D. M. Himmelblau, L. S. Lasdon, Optimization of
Chemical Processes, McGraw‐Hill, Boston, MA 2001.
2
[
[
[
[
51] S. Ko, M. N. V. Sastry, C. Lin, C.‐F. Yao, Tetrahedron Lett. 2005,
6
46, 5771.
27] A. Khazaei, L. Jafari‐Ghalebabakhani, E. Ghaderi, M. Tavasoli,
A. R. Moosavi‐Zare, Appl. Organometal. Chem. 2017, https://
doi.org/10.1002/aoc.3815.
52] A. Khazaei, A. R. Moosavi‐Zare, H. Afshar‐Hezarkhani, V.
Khakyzadeh, RSC Adv. 2014, 4, 32142.
53] M. Tajbakhsh, H. Alinezhad, M. Norouzi, S. Baghery, M.
[
[
[
28] A. Amoozadeh, S. Rahmani, M. Bitaraf, F. B. Abadi, E.
Akbari, J. Mol. Liq. 2013, 177, 44.
Tabrizian, New J. Chem. 2016, 40, 770.
54] M. Tajbakhsh, H. Alinezhad, M. Norouzi, S. Baghery, M.
29] A. Khazaei, N. Sarmasti, J. Y. Seyf, M. Tavasoli, RSC Adv. 2015,
Akbari, J. Mol. Liq. 2013, 177(Suppl. C), 44.
5, 101268.
30] A. R. Moosavi‐Zare, M. A. Zolfigol, M. Zarei, A. Zare, J. Afsar,
Appl. Catal. A 2015, 505, 224.
SUPPORTING INFORMATION
[
[
31] A. Amoozadeh, S. Golian, S. Rahmani, RSC Adv. 2015, 5, 45974.
Additional Supporting Information may be found online
in the supporting information tab for this article.
32] A. Zare, F. Abi, A. R. Moosavi‐Zare, M. H. Beyzavi, M. A.
Zolfigol, J. Mol. Liq. 2013, 178, 113.
[
[
[
33] S.‐J. Song, Z.‐X. Shan, Y. Jin, Synth. Commun. 2010, 40, 3067.
34] S. Rostamnia, H. Xin, Appl. Organometal. Chem. 2014, 28, 359.
35] E. Doustkhah, S. Rostamnia, A. Hassankhani, J. Porous Mater.
How to cite this article: Khazaei A,
Mahmoudiani Gilan M, Sarmasti N. Magnetic‐
based picolinaldehyde–melamine copper complex
for the one‐pot synthesis of hexahydroquinolines
via Hantzsch four‐component reactions. Appl
Organometal Chem. 2017;e4151. https://doi.org/
2
016, 23, 549.
[
36] S. Rostamnia, A. Hassankhani, H. G. Hossieni, B. Gholipour, H.
Xin, J. Mol. Catal. A 2014, 395(Suppl. C), 463.
[
[
37] S. Rostamnia, A. Hassankhani, Synlett 2014, 25, 2753.
10.1002/aoc.4151
38] S. Sobhani, F. Zarifi, J. Skibsted, New J. Chem. 2017, 41, 6219.