Process modeling of reduction and acetylation reactions by spectrophotometric and chemometrics methods

Article abstract of DOI:10.1007/s13738-013-0284-7

Achieving high levels of chemoselectivity is a pivotal feature of chemical synthesis. Although significant progress has been made in this regards, lots of challenges lie ahead. Herein, multivariate curve resolution-alternating least square (MCR-ALS) was employed to overcome the spectral overlapping of the reactants and products in the model reactions of methyl-(4-formyl) benzoate by NaBH₄ and acetylation of p-aminophenol by acetyl chloride. The first-order ultraviolet/visible absorbance spectra were recorded during the addition of different quantities of reagent to the substrate. The absorbing species, coexisting in the system, were detected through employing factor analysis. The soft-modeling analysis of the evolutionary absorbance data by MCR-ALS revealed that both model reactions proceed through a two-step consecutive manner. Using the obtained concentration profiles, one can obtain an idea about the suitable molar ratio of reagent/substrate for selective production of one of the reaction products. Moreover, the pure spectrum of the intermediate species could be estimated.

Full text of DOI:10.1007/s13738-013-0284-7

J IRAN CHEM SOC
DOI 10.1007/s13738-013-0284-7
ORIGINAL PAPER
Process modeling of reduction and acetylation reactions
by spectrophotometric and chemometrics methods
•
Bahram Hemmateenejad Parisa Shadabipour
•
Moahamad Reza Mohamadizadeh
Received: 17 April 2013 / Accepted: 27 May 2013
Ó Iranian Chemical Society 2013
Abstract Achieving high levels of chemoselectivity is a
pivotal feature of chemical synthesis. Although significant
progress has been made in this regards, lots of challenges
lie ahead. Herein, multivariate curve resolution-alternating
least square (MCR-ALS) was employed to overcome the
spectral overlapping of the reactants and products in the
model reactions of methyl-(4-formyl) benzoate by NaBH4
and acetylation of p-aminophenol by acetyl chloride. The
first-order ultraviolet/visible absorbance spectra were
recorded during the addition of different quantities of
reagent to the substrate. The absorbing species, coexisting
in the system, were detected through employing factor
analysis. The soft-modeling analysis of the evolutionary
absorbance data by MCR-ALS revealed that both model
reactions proceed through a two-step consecutive manner.
Using the obtained concentration profiles, one can obtain
an idea about the suitable molar ratio of reagent/substrate
for selective production of one of the reaction products.
Moreover, the pure spectrum of the intermediate species
could be estimated.
Introduction
Analyzing the behavior of reactants and products, in
addition to investigating the mechanism of reactions, is an
important stage in the study of organic reactions, especially
for reduction and acetylation reactions, which are impor-
tant subjects in the world of chemistry. To date, vast
varieties of reducing agents have been introduced for
reduction processes [1–4]. Achieving the highest yield and
purity, and minimizing the amount of by product is con-
sidered as priority for synthetic chemists.
Of the powerful and mild reducing agents, which have
been developed for this purpose, LiAlH4 and NaBH4 are
commonly used reagents [2, 3]. However, some disad-
vantages are associated with their applications. For exam-
ple, because of their reducing power all organic functional
groups in a molecule can be reduced, and it is rather dif-
ficult to utilize lithium aluminum hydride for the selective
reduction of multifunctional molecules [5].
Sodium borohydride is a relatively mild reducing agent,
which primarily has been used for the reduction of reactive
functional groups in protic solvents [6]. Its mild reducing
property allows for considerable selectivity in reduction of
organic compounds. Aldehydes and ketones are, in general,
readily converted into their corresponding alcohols in the
presence of a variety of functional groups including nitro,
carboxylic acid, ester, amide, carbonate, carbamate, lac-
tone, epoxide, nitrile, halide, alkene, and alkyne [7].
However, the rate of reaction is sometimes slow; accom-
panying with relatively low chemoselectivity in some cases
[8].
Keywords Process modeling ꢀ Chemometrics ꢀ
Chemoselectivity ꢀ Spectrophotometry ꢀ Acetylation ꢀ
Reduction
Electronic supplementary material The online version of this
article (doi:10.1007/s13738-013-0284-7) contains supplementary
material, which is available to authorized users.
B. Hemmateenejad (&) ꢀ P. Shadabipour
Department of Chemistry, Shiraz University, 71454 Shiraz, Iran
e-mail: hemmatb@sums.ac.ir
The acylation of amines, which frequently is used as a
fundamental protecting procedure in organic synthesis,
provides an efficient and inexpensive means for protecting
an amino functionality in a multi-step synthetic process [9].
M. R. Mohamadizadeh
Department of Chemistry, Faculty of Sciences,
Persian Gulf University, 75169 Bushehr, Iran
123

J IRAN CHEM SOC
Owing to displaying both nucleophilic and basic charac-
teristics, amines must often be blocked with a protecting
group during a multi-step synthesis, e.g., in the synthesis of
a diverse array of biological molecules such as amino
acids, peptides, glycopeptides, aminoglycosides, b-lac-
tams, nucleosides, sphingosines and alkaloids. N-Acyl
derivatives are well-known protecting groups for the
amines [10]. Acylation of amines is sometimes carried out
using acyl transfer reagents [10] and at other times with
acetic acid [11]; but acylating reagents such as acyl halides
and acid anhydrides are usually employed in the presence
of either acidic or basic catalysts [12]. Recently, the
advantages and drawbacks of these reactions have been
described extensively by Katritzky [13].
[18]. It should be noted that MCR-ALS can be used to
analyze the first-order spectroscopic data (where for each
sample a vector of variables is available) as well as the
second-order data (where for each sample a bi-dimensional
array or a matrix of variables is available) [19–22]. Here,
the first-order spectroscopic data were analyzed by MCR-
ALS.
In the present contribution, a simple chemometrics
based spectrophotometric method has been developed for
investigating the reduction of methyl-(4-formyl) benzoate
(MFB) and acetylation of p-aminophenol (PAP). A soft-
modeling method has been applied to gain insight into the
mechanism of reduction and acetylation.
Monitoring the progress of organic reactions is usually
performed with the help of chromatographic methods.
Whilst thin-layer chromatography is usually used for
qualitative monitoring of organic reactions, gas chroma-
tography and high performance liquid chromatography can
be used to gain quantitative information concerning the
progress of organic reactions. However, these methods are
associated with sophisticated instrumentation, cumbersome
sample preparation and trained operators. Thus, the
development of a new simple practical method for moni-
toring the progress of organic reactions is still in great
demand.
Experimental
Apparatus and software
The UV–vis measurements were performed on an Ultro-
spec 3000 UV/visible spectrometer equipped with a 10-mm
quartz cell. All calculations were performed in MATLAB
environment (Mathwork Inc., version 7) utilizing a Pen-
tium 4 (CPU at 2.8 GHtz) personal computer with Win-
dows XP operating system.
Spectrophotometric methods are in general simple,
sensitive and very suitable for studying chemical reactions
in solutions. When a reaction process is analyzed by
ultraviolet/visible (UV/vis) absorption spectroscopy, a
series of spectra are recorded during the course of the
reaction (e.g., at different molar ratios of substrate to
reagent or at different times). The collected data contain
the information about all light-absorbing species involved
in the chemical process. Spectral band overlapping is one
significant limiting factor, which hinders the application of
spectroscopic methods in monitoring the chemical reac-
tions. However, spectral overlapping could be addressed by
doing different chemometric methods [14, 15].
Reagents
Methyl-(4-formyl) benzoate and p-aminophenol were pur-
chased from Sigma-Aldrich. Sodium borohydride and
acetyl chloride were obtained from Fluka. The analytical
reagent grade solvents ethanol and acetonitrile were pur-
chased from Merck (Darmstadt, Germany).
Stock standard solutions of MFB and NaBH4 were
prepared in ethanol with concentration of 5.0 9 10-3 and
1.0 9 10-2 M, respectively. Solutions of lower concen-
trations of MFB (4.0 9 10-4, 4.0 9 10-5 M) and NaBH4
(1.0 9 10-3 M) were prepared by appropriate dilution of
the stock solutions with ethanol.
Multivariate curve resolution-alternating least square
(MCR-ALS) is emerging as a popular chemometrics
method used for the resolution of multiple component
responses in unknown and unresolved mixtures. MCR-ALS
is a soft-modeling approach for describing and modeling
processes without explicitly using the underlying chemical
model linked to them [16]. MCR-ALS decomposes the raw
measurements (in our case, a series of spectra as a function
of mol ratio) into a bilinear model that contains the con-
centration profiles and pure spectra of each absorbing
species contributing to the collected measurements.
Recently, MCR-ALS method has been used by our group
for the monitoring of drug photodegradation [17], and
studying the behavior of iodine in binary solvent mixtures
Two stock standard solutions of PAP and acetyl chloride
with a respective concentration of 5.0 9 10-3 and
1.0 9 10-2 M were prepared in acetonitrile. Solutions of
lower concentrations of PAP (5.0 9 10-5 M) and acetyl
chloride (3.0 9 10-3 M) were prepared by appropriate
dilution of the stock solutions with acetonitrile.
Experimental procedure
In a typical experiment, a 3.0-mL aliquot of the ethanolic
solution of 4.0 9 10-5 M MFB was transferred into a
quartz cell and afterward it was titrated by addition of
10.0 lL portions of 1.0 9 10-3 M NaBH4, which was
added to both the blank (ethanol) and the main solution
123

J IRAN CHEM SOC
until the reaction was completed. The same procedure was
applied on 1.0 9 10-2 M of NaBH4 and 4.0 9 10-4 M of
MFB. After each addition, the solutions were stirred for
3 min and the absorbance spectra were recorded over the
reagent and blank (NaBH4 solution was added to pure
solvent in the same concentrations as blank).
respectively. Herein, C and S are containing, respectively,
the concentration profiles and pure spectra of the species
coexisting in the studied reactions. Equation (1) proposes
the matrix notation valid for the general MCR procedures:
D ¼ CST þ R
ð1Þ
where R (m 9 n) is being the matrix of residual variation
of the data matrix not modeled by the employed MCR
method [20].
For monitoring of the acetylation reaction, a 3.0-ml
portion of 5.0 9 10-5 M of PAP in acetonitrile was
transferred to the quartz cell and certain amounts (i.e.,
10.0 lL) of the acetyl chloride solution (3.0 9 10-3 M) in
acetonitrile was added to the sample solution. After each
addition, the resulting solutions were stirred for 3 min
following with UV–vis measurement in the same manner
as explained for reduction of MFB.
Since its introduction, MCR-ALS method has received
wide applications among the scientific community. The
first step in MCR-ALS is determination of r (i.e., the
number of systematic contributions in the data matrix D).
Here, the absorbance data matrix was subjected to factor
analysis (FA) [21] to evaluate the number of significant
components. The next step involves in an initiate estima-
tion for either C or S. In this regard, we used evolving
factor analysis (EFA) [22] to suggest an estimate of con-
centration profile (C).
Each recorded absorbance spectrum was digitized in
1 nm intervals and saved as a row vector of length of n,
where n is the number of absorbance readings per spec-
trum. Thus, the digitized absorbance spectra obtained in a
course of a reaction were collected below each other to
form an absorbance data matrix (D) of size of (m 9 n),
where m is number of recorded spectra per experiment or
the number of molar ratios of reagent to substrate, at which
the absorbance spectra were recorded.
It should be noted that the reduction of MFB was
monitored at two different initial concentrations of sub-
strate, while the molar ratios of reagent to substrate at two
experiments were taken the same, whereas wavelength
ranges were different. In this case, row augmentation of the
absorbance data matrices was adapted and MCR-ALS was
run on individual matrices as well as the augmented data
matrix.
Chemometrics procedure
Multivariate curve resolution techniques, as types of soft-
modeling methods, have been proposed for the recovery of
the response profiles (spectra, pH profiles, time profiles,
elution profiles) that comprise more than one component in
an unresolved mixture (obtained from evolutionary pro-
cesses). These methods have been when little or no back-
ground is available about the nature and composition of the
mixture [19]. MCR methods attempt to resolve mixed
evolutionary profiles without presupposing a well-defined
shape function. Instead, concerning the general shape of
the profile and spectra, they impose known constraints such
as: (1) unimodality constraint (each component profiles has
only one maximum); (2) non-negativity constraint (all
points in the concentration and component spectra profiles
must be greater than or equal to zero); and (3) closure
constraint (the sum of species concentration is a known
constant value, i.e., in reaction-based systems, it is known
as mass balance equation). MCR produce a mathematical
decomposition of a multivariate multicomponent experi-
mental data matrix D (m 9 n), which represents the total
instrumental response of a system, into the product of two
simpler matrices C (m 9 r) and ST (r 9 n), which are the
pure response profiles of the r individual contributions
making up the system. The scalar r can be considered as
the number of light-absorbing species evolved or decayed
in the course of reaction. The matrices C and ST are related
to the data variation in the row and column directions of D,
The MCR-ALS subroutine written by Tauler and
coworkers was applied to resolve the components pure
spectra and their corresponding concentration profiles. The
software is freely available by http://www.ub.edu/mcr/
web_mcr/welcome.htm.
Results and discussion
Reduction of MFB by sodium borohydride
Herein, we aimed to represent a simple analytical meth-
odology to monitor the progress of organic reactions, by
which we can judge the selectivity at each reaction stage.
MFB was used as a model substrate (Fig. 1) due to having
both aldehyde and ester groups. According to the literature
[23], NaBH4 as a mild reducing agent, first attacks the
aldehyde functional group and then attack that of ester for
both electronic and steric reasons [23, 24]. We used a
spectroscopic method, as an alternative to separation
methods, to confirm the above statement.
The changes in the absorbance spectra of the ethanolic
solution of MFB (4.0 9 10-5 M) upon addition of differ-
ent quantity of reductant are represented in Fig. 2a. We
used this concentration to have more precise absorbance
data (1.0 \ absorbance \ 0.5). As it is observed, the MFB
123

J IRAN CHEM SOC
Fig. 1 Reduction pathway of
MFB reduction by NaBH4
spectrum possesses two absorbance maxima at around 251
and 293 nm. Upon the addition of NaBH4, the absorbance
intensity at both peak maxima decreases, along with
showing a blue shift. However, in the wavelength region of
275–315 nm, the absorbance changes are very small.
Therefore, to have spectral information about the peak at
around 293 nm, another experiment was established with a
higher concentrations of both reactants (i.e., MFB
4.0 9 10-4 M, NaBH4 1.0 9 10-2 M) while the mole
ratios of NaBH4/MFB were kept at the same as those used
in the previous experiment. The obtained absorbance
spectra are depicted in Fig. 2b. Since the absorbencies of
lower wavelengths were overloaded, we used the absor-
bance data of wavelengths higher than 272 nm. The
observed changes in the absorbance spectra of MFB can be
attributed to conversion of MFB to its reduction product(s).
It should be noted that we have two absorbance data
matrices: D1 for MFB concentration of 4 9 10-5 M in the
wavelength interval of 214–320 nm and D2 for MFB
concentration of 4.0 9 10-4 M in the wavelength interval
of 272–320 nm. Factor analysis was employed to deter-
mine the number of significant principal components (PCs)
in the reaction systems. The results are given in Supple-
mentary section. For system D1, a large separation was
observed between the second and the third Eigen-values
(EV), however, a small separation was also observed
between the third and the fourth ones. In the loading plots
two smooth loadings were found whereas the third and
fourth ones were noisy. Therefore, it is difficult to
accurately estimate the number of significant PCs for D1.
The same procedure applies for D2, where the large sepa-
ration was observed between the first and second Eigen-
values whereas two smooth loadings were observed.
However, according to the reaction Scheme shown in
Fig. 1, reduction of MFB is associated with three species.
The probable reasons for obtaining number of factor lower
than the number of species in D2 are (1) rank deficiency in
spectral profile or (2) one species does not have significant
absorbencies in the used spectral region. To have spectral
information about both data matrices and two break rank
deficiency if existed, matrix augmentation can be a simple,
versatile and smart way [25].
Since the data matrices benefit from same mole ratios in
row wise, but having different region of wavelength in
column wise, they can be augmented in a row-wise man-
ner. The new augmented matrix is easily achieved by set-
ting individual data matrices one beside of another. The
way of augmentation and the information gained by MCR-
ALS analysis of the augmented matrix are shown in Fig. 3.
The number of columns in the new augmented data matrix
equals to the total of the acquired spectra in different
concentration of MFB and the number of rows remains the
same in all the data matrices. In this case, MCR-ALS
results of both experiments with unique concentration
profile and different spectral profiles are accessible.
To investigate the number of independent chemical
absorbing species created through reduction of MFB, the
augmented data matrix of the recorded absorbencies at
Fig. 2 Changes in the
absorbance spectra of MFB
upon successive addition of
NaBH4 a MFB 4.0 9 10-5
and NaBH4 1.0 9 10-3 M,
b MFB 4.0 9 10-4 M and
NaBH4 1.0 9 10-2 M. (to
M
record each spectrum, 10.0 lL
portions of NaBH4 were added
successively to the
corresponding solution of
MFB). The first spectra are
those of MFB in the absence of
NaBH4. All spectra were
recorded versus the reagent
blank
123

J IRAN CHEM SOC
Fig. 3 Schematic
representation of the augmented
data by row-wise augmentation
different mol ratio was subjected to factor analysis [26].
The variation in the Eigen-values of the covariance data
matrix as a function of number of factor is represented in
Fig. 4a for the first 10 factors. As can be seen, there is a
relatively large separation between the third and fourth
Eigen-values. The three first factors, which could explain
about 99 % of total variances in the absorbance data,
contained the systematic variances and the rest of factors
can be considered as noisy factors.
Using the results of EFA as initial estimate of concen-
tration profile, MCR-ALS was run implementing the non-
negativity constraint to both concentration and spectral
profiles and unimodality and closure to concentration
profile [27]. The program was converged after few itera-
tions with lack fit errors of 0.788 and 0.926 for PCA and
model, respectively. The optimized concentration and
spectral profiles after applying MCR-ALS on aug-
mented data matrix are shown in Fig. 5, where A is MFB
and B and C are the intermediate and final products,
respectively.
The number of significant factors was also investigated
by loading plot [17]. As represented in Fig. 4b, the first
three loading represented the systematic variation and the
fourth one represented a noisy distribution. This result was
confirmed by EFA [22]. As is shown in Fig. 4c the forward
and backward EFA plots represent a large separation
between the first three Eigen-values and the reminder Ei-
gen-values. This is further confirmation of the presence of
three chemical species in reduction reaction of MFB.
The detected chemical species by FA results can be
attributed to the pure MFB and its reduction products
(mono and dialcohol products). The calculated initial
estimate of the concentration profiles of the reactant and
products by EFA (see supplementary section) suggest that
the reduction of MFB in the experimental conditions
employed in this work may be preceded in a two-step
manner. The plotted concentration profiles are more likely
to be a consecutive two-step reaction of the form of
A ? B ? C, where A is pure MFB and B and C are the
intermediate and final products, respectively.
The shown concentration profile reveals that the first
product, which is assumed to be that producing from the
reduction of aldehyde functional group, is evolved by first
addition of NaBH4 and its concentration is increased up to
molar ratio of about 0.5. It is observed that in the molar
ratio region of 0.0–0.5, only substrate and the first reduc-
tion product are coexisted and the third component is
absent. The final dialcohol product is evolved after molar
ratio of 0.5. This suggests that the used reagent for
reduction of MFB is highly selective in the reduction of the
first functional group (aldehyde to monoalcoholic product).
However, at molar ratios higher than 0.5, both products are
coexisted up to the molar ratio of about 1.8. At molar ratios
[1.8 the final product is available with high degree of
selectivity. Finally, the shown concentration profiles give
the information on the yield of reaction. As shown, the
maximum amount of B was achieved at molar ratio of
*0.5 with about 90 % conversion yield whereas the
Fig. 4 Factor analysis results for determination of the number of factors in reduction of MFB using augmented data matrix. a Logarithm of the
singular values, b the normalized loadings plots for the first four factors and c EFA plot
123

J IRAN CHEM SOC
Fig. 5 The optimized
concentration profile a and pure
spectra b of the species evolved
through reduction of MFB by
NaBH4 as determined by MCR-
ALS analysis of the augmented
data matrix: a MFB,
b intermediate product and
c final product
product C reach to its maximum value at molar ratio 2.0 or
higher with about 100 % conversion yield.
The changes in the absorbance spectra of the
5.0 9 10-5 M solution of p-aminophenol in acetonitrile
upon adding different quantities of acetyl chloride are
given in Fig. 7. As it is observed, the p-aminophenol
spectrum possess two absorbance maxima at 240 and
310 nm whereas upon adding acetyl chloride, the absor-
bance intensities at around 240 nm are increased along
with decreasing the band at around 310 nm accompanying
with Blue shift in both bands.
The resolved pure spectra of MFB and its reduction
products are also demonstrated in Fig. 5. First of all, there
was an excellent agreement between the calculated pure
spectra of MFB and its experimental spectra (the spectra A
of Fig. 5b and the first spectra of Fig. 2, respectively).
Secondly, in the resolved pure spectra of D2 one can
observe that those of products B and C are very similar.
This confirms the rank deficiency of D2 as the results of
rank deficiency in spectral profiles.
To investigate the number of independent chemical
absorbing species created through acetylation of p-amino-
phenol, the data matrix of the recorded absorbance at dif-
ferent mole ratios were subjected to factor analysis. The
logarithmic plot of the singular values against the number
of factors (shown in Fig. 8a) represents a large and clear
separation between the third and fourth eigenvectors,
indicating the presence of three significant PCs in the
system. The number of significant factors was also inves-
tigated by loading plot as is represented in Fig. 8b. Clearly,
the first three loading represented the systematic variation
and the fourth one represented a noisy distribution. The
results were confirmed by EFA plot (Fig. 8c). It should be
noted since the absorbance intensities of the studied solu-
tions of PAP were comparable at two peaks (240 and
310 nm) and all three methods of principal component
determinations arrived to a unique results, we did not
repeat the experiment with another concentration of PAP
and acetyl chloride. The calculated initial estimate of the
Acetylation of PAP by acetyl chloride
In order to investigate chemoselectivity in the acetyla-
tion reaction, the acetylation of p-aminophenol by acetyl
chloride [12] was chosen. As shown in Fig. 6, both amine
and alcohol groups could be acetylated by acetyl chlo-
ride. However, acetyl chloride prefers amino group
because of the much higher nucleophilicity of the amino
group compared to the hydroxyl group, and the higher
thermodynamic stability of amides compared to esters
[28]. To show the selectivity of this reaction as a function
of the molar ratio of reagent to substrate, UV/vis
absorption spectroscopy was employed as a simple ana-
lytical technique and coupled it with chemometric
methods to overcome the problem of spectral bands
overlapping.
Fig. 6 Acetylation pathway of
p-aminophenol by acetyl
chloride
123

J IRAN CHEM SOC
*0.8 and only the amino group is being acetylated.
However, after molar ratio of 0.8 both products are coex-
isted and the first product is even existed at molar ratio of
2.0. The final product could be selectively obtained at
molar ratio of higher than 2.2. Finally, the shown con-
centration profiles give information on the yield of reac-
tion. As shown, the amount of B and C at respective molar
ratios of about *0.8 and 2.2 are close to the initial amount
of PAP, indicating a high yield of conversion (close to
100 %).
The resolve pure spectra of the different species shown
in Fig. 9b explain that acetylation of the amino group
changes significantly the peak maximum of longer wave-
lengths (to lower wavelengths and reduced molar absorp-
tivity) whereas the second acetylation process does not
affect significantly the peak maximum of monoacetylated
product. However, both the first and second acetylation
processes affect significantly the shorter wavelength peak
of PAP. It is not worthy to note that there is an excellent
agreement between the experimental and resolved spectra
of PAP.
Fig. 7 Absorbance spectra of 5.0 9 10-5 M p-aminophenol with
adding 3.0 9 10-3 M acetyl chloride (to record each spectrum, 10 lL
of acetyl chloride was added to the corresponded solution. Time
interval between spectra is 3 min)
concentration profile of the reactant and products by EFA
(shown in supplementary section) was used to start the
MCR-ALS optimization.
The employed constraints were the same as those used
for reduction of MFB. After convergence of MCR-ALS at
few iterations, the lack of fit of errors was 1.13 and 1.22 %
for PCA and model, respectively. The resolve concentra-
tion profiles and pure spectra are given in Fig. 9.
Conclusion
The shown concentration profiles reveal that the first
product, which is assumed to be the acetylation of amino
group, is evolved by first addition of acetyl chloride and is
grow up at mole ratio of *0.8. In this mole ratio region
(0.0–0.8) the third component (final product, where both
groups are acetylated) is absent. However, the second
acetylated product is evolved after molar ratio of *0.8.
This suggests that the used method for acetylation of
p-aminophenol is highly selective before molar ratio of
In summary, a method based on soft-modeling multivariate
curve resolution alternative least square was developed to
study the chemoselectivity of a reduction and an acetyla-
tion reaction using methyl-(4-formyl) benzoate and
p-aminophenol as substrate, respectively. Augmentation of
the data was performed on the absorbance data matrices of
MFB reduction to break rank deficiency in spectra of
longer wavelengths. The obtained results from MCR-ALS
provide a concentration profile, permitting one to choose
Fig. 8 Factor analysis results for determination of the number of factors in acetylation of PAP. a Logarithm of the singular values, b the
normalized loadings plots for the first four factors and c EFA plot
123

J IRAN CHEM SOC
Fig. 9 The optimized
concentration profile a and pure
spectra b of the species evolved
through acetylation of PAP by
acetyl chloride as determined by
MCR-ALS analysis of the
absorbance data matrix: a PAP,
b intermediate product and
c final product
9. T.W. Greene, P.G.M. Wuts, Protective Groups in Organic Syn-
thesis, 3rd edn. (Wiley, New York, 1999)
10. Y. Murakami, K. Kondo, K. Miki, Y. Akiyama, T. Watanabe, Y.
Yokoyama, Tetrahedron Lett 38, 3751 (1997)
11. N. Narender, P. Srinivasu, S.J. Kulkarni, K.V. Raghavan, Green
Chem. 2, 104 (2000)
12. S. Naik, G. Bhattacharjya, V.R. Kavala, B.K. Patel, Arkivoc
2004, 55 (2004)
13. A.R. Katritzky, H.-Y. He, K. Suzuki, J. Org. Chem. 65, 8210
(2000)
the appropriate mole ratio to achieve the higher chemose-
lective product as well as higher yield of conversion
without need to perform several experiments. Finally, it
should be noted that it is possible to use the proposed
chemometrics procedure for more complex reactions (e.g.,
more than 3/4 reactants). However, by increasing the
complexity of the system, running MCR-ALS requires
close attention to handle rank deficiency and rotational
ambiguity.
14. R.G. Brereton, Data Analysis for the Laboratory and Chemical
Plant (Wiley, New York, 2003)
15. B. Hemmateenejad, G. Absalan, M. Hasanpour, J. Iran. Chem.
Soc. 8, 166 (2011)
16. A.D. Juan, J. Coello, S. Maspoch, Anal. Chim. Acta 567, 245 (2006)
17. B. Hemmateenejad, K. Javidnia, M. Saeidi-boroujeni, J. Pharm.
Biomed. Anal. 47, 625 (2008)
Acknowledgments Financial support of this work by Shiraz Uni-
versity Research Council is highly appreciated.
18. M. Shamsipur, B. Hemmateenejad, J. Soln. Chem. 32, 819 (2003)
19. R. Tauler, J. Chemometr. 15, 627 (2001)
20. M. Esteban, C. Arino, J.M. Diaz-Cruz, M.S. Diaz-Cruz, R. Ta-
uler, Trends Anal. Chem. 19, 49 (2000)
21. M. Meader, Y.M. Neuhold, Practical Data Analysis in Chemis-
try, 1st edn. (Elsevier, New York, 2007)
References
1. S.D. Burke, R.L. Danheiser, Handbook of Reagents for Organic
Synthesis, Oxidizing and Reducing Agents (Wiley-VCH, New
York, 1999)
22. H. Gampp, M. Maeder, C.J. Meyer, A.D. Zuberbuhler, Talanta
32, 1133 (1985)
2. B.T. Cho, S.K. Kang, S. Kim, R. Ryu, D. Keun, Tetrahedron 62,
8164 (2006)
23. D.E. Ward, C.K. Rhee, Can. J. Chem. 67, 1206 (1989)
24. B. Zeynizadeh, F. Faraji, Bull. Korean Chem. Soc. 24, 453 (2003)
25. G. Hanrahan, F.A. Gomez, in Chemometric Methods in Capillary
Electrophoresis, ed. by G. Hanrahan, F.A. Gomez (Wiley, New
jersey, 2009)
3. J.C.S. Costa, K.C. Pais, E.L. Fernandes, P.S.M.D. Oliveira, J.S.
Mendonc¸a, M.V.N.D. Souza, M.A. Peralta, T.R.A. Vasconcelos,
Arkivoc 2006, 128 (2006)
4. B. Zeynizadeh, S. Yahyaei, Bull. Korean Chem. Soc. 24, 1664
(2003)
26. E.R. Malinowski, Factor Analysis in Chemistry, 3rd edn. (Wiley,
New York, 2002)
5. N.G. Gaylord, Reduction with Complex Hydrides (Interscience
Publishers, New York, 1956)
27. M. Garrido, F.X. Rius, M.S. Larrechi, Anal. Bioanal. Chem. 390,
2059 (2008)
6. E. Section, G.A. Olah, J.M. Bollingerlb, J. Am. Chem. Soc. 89,
4744 (1967)
28. http://www.organomet.chem.es.osaka-u.ac.jp. Accessed 28 Nov
2011
7. D.E. Ward, C.K. Rhee, W.M. Zoghaib, Tetrahedron Lett. 29, 517
(1988)
8. H.C. Brown, S. Krishnamurthy, Tetrahedron 35, 567 (1979)
123