Journal of CO₂ Utilization 26 (2018) 212–220

Contents lists available at ScienceDirect

Journal of CO2 Utilization

journal homepage: www.elsevier.com/locate/jcou

Application of Box–Behnken design for processing of mefenamic T

acid–paracetamol cocrystals using gas anti-solvent (GAS) process
⁎
Napada Wichianphonga,b,c, Manop Charoenchaitrakoola,b,c,
a
Department of Chemical Engineering, Faculty of Engineering, Kasetsart University, Bangkok 10900, Thailand
b
Center for Advanced Studies in Nanotechnology for Chemical, Food and Agricultural Industries, KU Institute for Advanced Studies, Kasetsart University, Bangkok 10900,
Thailand
c
NANOTEC-KU-Center of Excellence on Nanoscale Materials Design for Green Nanotechnology, Kasetsart University, Bangkok 10900, Thailand

A R T I C LE I N FO A B S T R A C T

Keywords: The present work deals with the preparation of drug–drug cocrystals of mefenamic acid (MEF) with paracetamol
Drug–drug cocrystal (PAR) using gas antisolvent (GAS) process with an aim to improve the dissolution rate of MEF. Box–Behnken
Mefenamic acid experimental design was used to optimize the GAS process variables for minimal dissolution time of MEF. A
Paracetamol mathematical model was developed to study the effects of operating temperature, PAR-to-MEF molar ratio and
GAS process
%MEF saturation in the ranges of 25 °C–45 °C, 3:1–5:1 and 70–90%, respectively. PAR-to-MEF ratio and %MEF
Box–Behnken
saturation were found to be the main parameters affecting the dissolution time. The fastest dissolution time
(t63.2), 11.38 min, was obtained at the optimal conditions of a temperature of 34.9 °C, PAR-to-MEF ratio of 4.3:1
and MEF saturation of 86.8%. Optimized cocrystals by GAS showed that the dissolution rate of MEF improved by
6.0, 5.3 and 2.3-fold when compared to pure MEF, sieved cocrystals prepared by a traditional slow evaporation
technique and sieved marketed combination drugs, respectively. GAS cocrystallization thus offers an efficient
way to enhance the dissolution rate of the poorly water-soluble drug.

1. Introduction from traditional methods used in preparation of cocrystals. Solvent

Poor water-solubility of drugs is a major problem in the pharma-
ceutical industry. A drug with low solubility also has poor dissolution
properties, leading to poor bioavailability and resulting in a low ther-
apeutic efficacy [1]. Hence, an important challenge now is to find a way
to improve drug dissolution rate. The synthesis of cocrystals, i.e. co-
crystallization, is very attractive because it offers opportunities to im-
prove the dissolution performance of poorly water-soluble drugs,
without changing their molecular structure [2,3]. Pharmaceutical co-
crystals are normally formed from an active pharmaceutical ingredient
(API) and a cocrystal former in the crystal lattice through noncovalent
interactions, often including hydrogen bonding. The cocrystal former or
coformer may be an excipient or another active drug. When the pre-
paration technique uses two drugs, the resulting cocrystals are termed
drug–drug or multidrug cocrystals [4]. This class of cocrystal offers
potential advantages in terms of enhanced solubility and dissolution
rate of API, greater therapeutic effect (i.e., synergistic and/or additive
effects), reduction of the number of prescriptions and administrative
costs and an increase in patient compliance [4,5].
In general, the preparation of drug–drug cocrystals does not differ
evaporation, melting cocrystallization, solid grinding or liquid-assisted
grinding have been applied for the preparation of drug–drug cocrystals
in the literature [6]. Recently, the application of dense CO2 in cocrystal
formation has been widely studied. For this, one useful technique is the
gas antisolvent (GAS) process. This technique offers a rapid single-step
process, has low environmental impact and provides an opportunity to
control of important product characteristics, such as particle size and
morphology, in a more efficient manner. In addition, the use of dense
CO2 can also reduce thermal and mechanical stress acting on APIs
compared to grinding methods, as well as reduce organic solvent use
and reduce residual solvent in cocrystal products compared to a tradi-
tional solvent-based cocrystallization [7–9]. It is important to note that
the GAS technique has been used to form several drug–coformer co-
crystals [8,10–14], however, up until now the use of the technique to
form drug–drug cocrystal has not been studied.
Mefenamic acid (MEF, Fig. 1), a derivative of fenamic acid, is a non-
steroidal anti-inflammatory and analgesic drug used to treat period
pain and other pains. According to the Biopharmaceutical Classification
System (BCS), MEF is classified as class II, having low water-solubility,
and consequently its absorption appears to be limited by solubility after

⁎
Corresponding author at: Department of Chemical Engineering, Faculty of Engineering, Kasetsart University, Bangkok 10900, Thailand.
E-mail address: manop.c@ku.ac.th (M. Charoenchaitrakool).

https://doi.org/10.1016/j.jcou.2018.05.011
Received 11 February 2018; Received in revised form 9 May 2018; Accepted 10 May 2018
Available online 25 May 2018
2212-9820/ © 2018 Elsevier Ltd. All rights reserved.
N. Wichianphong, M. Charoenchaitrakool
Fig. 1. Structural formulas of mefenamic acid (MEF) and paracetamol (PAR).
administration [15]. Therefore, it is necessary to enhance the dissolu-
tion rate of this drug. MEF cocrystallization with more soluble co-
formers (or APIs) is one solution to this problem. Successful MEF co-
crystal formation has been obtained using some coformers, such as
nicotinamide (a form of vitamin B3) [16–18] and 4,4′-bipyridyl [19].
Recently, our work [20] demonstrated that mefenamic acid–-
nicotinamide (MEF–NIC) cocrystal could be produced successfully
using gas anti-solvent (GAS) process. Box–Behnken design was used to
investigate the effects of operating parameters: temperature, coformer-
to-drug molar ratio and %drug saturation in the ranges of 25–45 °C, 3–5
and 70–90%, respectively. At a temperature of 45 °C, a coformer-to-
drug ratio of 5 and a %drug saturation of 70% were found to be the
optimal conditions for achieving the fastest dissolution time. The sieved
cocrystals by GAS showed an enhanced dissolution rate 38 times
greater than that of pure MEF.
Although there is some research on the production of cocrystal of
MEF and coformer, fewer drug–drug cocrystals with MEF have been
reported to date. In early studies, Pathak et al. [21] proposed solvent
evaporation as a method for preparing drug–drug cocrystals of mefe-
namic acid with paracetamol, but there have been no reports on MEF
dissolution improvement for this cocrystal. Therefore, there remains a
challenge to investigate the drug–drug cocrystal.
Paracetamol, acetaminophen or 4-hydroxyacetanilide (PAR, Fig. 1),
is a widely used analgesic and antipyretic drug employed in the treat-
ment of fever, pains, headaches and other minor aches [22]. PAR has
effective functional groups, such as an amide and a hydroxyl group, to
form the intermolecular hydrogen bonds required for cocrystal forma-
tion [23]. In a prospective observational clinical study by Rao and
Sailaja [24], MEF and PAR combination was found to be more effective
than paracetamol as an antipyretic (fever reducer) with high sig-
nificance (p < 0.05). In addition, MEF in combination with PAR is also
very useful for the treatment of moderate pains connected with rheu-
matoid arthritis [25]. From a combination drug perspective, this is an
interesting drug–drug cocrystal because MEF–PAR cocrystal may not
only have benefits for MEF dissolution rate enhancement, but also ad-
vantages in terms of better antipyresis. Currently, a combination of
paracetamol and mefenamic acid is also marketed through various
formulations–for example, MEF 50 mg + PAR 125 mg, MEF
500 mg + PAR 325 mg and MEF 500 mg + PAR 500 mg.
To date, the preparation of MEF–PAR cocrystals using the GAS tech-
nique and the dissolution improvement of MEF through multidrug co-
crystallization have not been investigated in the literature. In this work,
GAS technique was used to produce cocrystals of MEF with PAR in order
to enhance dissolution rate of MEF. The GAS process was employed using
acetone as a solvent and carbon dioxide as an anti-solvent. Box–Behnken
response surface design was chosen to study the influence of process
parameters–namely, operating temperature, PAR-to-MEF molar ratio and
%MEF saturation–on the time at which 63.2% of the drug (i.e., MEF) is
dissolved (t63.2), and the process variables were also optimized so as to
obtain the fastest dissolution time (i.e., the minimal value of t63.2).
Comparative dissolution and morphology studies of optimized cocrystals
obtained by GAS, a marketed combination drug and cocrystals prepared
by a traditional slow evaporation method were also performed. Solid-state
characterization techniques, such as differential scanning calorimetry
(DSC), Fourier transform infrared (FT-IR) spectroscopy and powder X-ray
diffraction (PXRD) were used for confirming cocrystal formation.
213
Journal of CO₂ Utilization 26 (2018) 212–220
2. Materials and methods
2.1. Materials
Mefenamic acid (MEF) (C15H15NO2, 98%, Sigma) and paracetamol
(PAR) (C8H9NO2, 98%, Tokyo Chemical Industry) were used as re-
ceived. Commercial dual-drug-tablets containing mefenamic acid
(500 mg) and paracetamol (325 mg) were purchased from a local
market (India) and used as marketed combination drugs. Acetone
(99.5%, Quality Reagent Chemical) was used as an organic solvent.
Carbon dioxide (high-purity grade, TIG) was used as an anti-solvent.
2.2. Physical mixture preparation
Physical mixtures were obtained by gentle mixing of MEF and PAR
at predetermined molar ratios with a spatula for 1 min prior to analysis.
2.3. Marketed combination drug preparation
Since the marketed combination drugs were originally in the form of
tablets, they were gently ground into powder form and then sieved
through a 100-mesh sieve with openings of 150 microns prior to ana-
lysis.
2.4. Cocrystallization by slow evaporation method
MEF and PAR at predetermined molar ratios and concentrations
were dissolved in 5 mL of acetone. The solution was sonicated in an
ultrasonic bath at 40 °C for 10 min. The clear solution was then allowed
to slowly evaporate at room temperature for 48 h. The resulting solids
were in the form of large dense plates, therefore they were gently
ground and then sieved through a 100-mesh sieve with openings of 150
microns prior to analysis.
2.5. Cocrystallization by GAS process
A schematic flow diagram of GAS precipitation can be found else-
where [13]. Solutions of MEF and PAR in acetone were prepared at
predetermined concentrations and molar ratios. The saturated solubi-
lity values of MEF at various temperatures were obtained from a pre-
vious report [26]. In each experiment, 5 mL of prepared solution was
loaded into the precipitation chamber (Jerguson sight gauge series no.
32), which was immersed in a temperature-controlled water bath
having as a component a recirculation heater (Thermoline Unistat 130).
After the system temperature was stable at the target point (25 °C, 35 °C
or 45 °C), carbon dioxide was then introduced by a syringe pump (ISCO
model 260D) with a constant flow rate of 10 mL/min through a pre-
heating coil and a 0.5 μm filter from the bottom of the precipitation
chamber. The initial pressure prior to feeding CO2 to the system was set
to be 63 bar for each experiment (as in Ref. [20]). During the pressur-
ization, volumetric expansion of the liquid occurred, causing the re-
duction of the solvating power of acetone and thus the precipitation of
the dissolved solute. Precipitated products were trapped onto the filter
located at the bottom of the chamber. The system was further pres-
surized up to 90 bar in order to ensure a complete precipitation (as in
Ref. [20]). Then the solvent in the chamber was removed at a constant
pressure of 90 bar by passing 80 mL of CO2 from the top of the chamber
and flushing the solvent through the filter. After depressurization, the
products were collected for further analyses. Note that cocrystals by
GAS were used as received, without grinding or sieving.
2.6. Experimental design for the GAS cocrystallization
Box–Behnken experimental design with response surface metho-
dology is a systematic approach for determination of the importance of
input variables, their interactions and searching for the optimal
N. Wichianphong, M. Charoenchaitrakool
response [27]. In this study, a three-factor and three-level Box–Behnken
design was used to investigate the dissolution times for different process
conditions. Independent variables and their value range were taken to
be operating temperature (25–45 °C), PAR-to-MEF molar ratio (3:1–5:1)
and %MEF saturation (70–90%). Note that from our preliminary stu-
dies, the product obtained from low PAR-to-MEF molar ratio (1:1–2:1)
had a similar dissolution rate as those of physical mixture, while using
high PAR-to-MEF molar ratio (3:1–5:1) resulted in a significantly en-
hanced dissolution rate of MEF. Moreover, %MEF saturation of at least
70% was considered in this work since lower %MEF saturation can
result in lower and insufficient amount of product collected. A tem-
perature in the range of 25–45 °C was used since the system tempera-
ture was controlled using a temperature-controlled Perspex water bath.
Time required to dissolve 63.2% of MEF (t63.2) was considered as an
output variable (response). Note that t63.2 (i.e., mean dissolution time or
MDT), a characteristic parameter of the Weibull function, is commonly
employed to describe an in vitro dissolution curve and compare dis-
solution performances [28–30]. The ranges and levels of the in-
dependent variables employed in the experiments are given in Table 1.
There were 15 runs with different settings of parameters generated
from the software Minitab 17. Regression and graphical analysis of
experimental data were also performed using Minitab 17. Experimental
data were fitted to a mathematical equation in order to express the
relationship between independent variables and the corresponding
dependent variable. The quality of the fit of the model was evaluated
using the coefficient of determination (R2), adjusted R2 (R2 (adj)), and
Analysis of Variance (ANOVA). Contour plots were generated using the
fitted quadratic model.
The generalized form of the second-order quadratic polynomial
equation commonly used in the response surface [31,32] is presented in
the following equation:
Y= β0 + Σ βi xi + Σ βii x 2i + Σ βij xixj
where Y is the process response or output (t63.2); xi and xj are the un-
coded independent variables; β0 is a constant called intercept term; and
βi , βii and βij are the coefficients of the linear, quadratic (i.e., squared)
and interaction terms, respectively.
2.7. Characterization of cocrystal
Microstructural analysis was carried out using a JEOL JSM-5410LV
scanning electron microscopy (SEM) operated at an accelerating vol-
tage of 10 kV. Samples were mounted on an aluminum stub with
double-sided adhesive tape. A thin coating of palladium was applied to
the sample using a Polaron SC7640 sputter coater prior to microscopy.
Particle size distributions (PSDs) of the powder were measured by
laser diffraction using a Malvern Mastersizer 3000 particle size ana-
lyzer. Silicone oil was used as the dispersion medium in the measure-
ment.
DSC measurements were performed on a Metler Toledo DSC 1 Star
System. Approximately 7–9 mg of sample were placed onto an alu-
minum pan which was then covered with a cap with a pin hole in it. The
samples were heated from 25 °C to 300 °C at a scan rate of 5 °C/min
under a nitrogen purge (50 cm3/min).
FT-IR spectra were obtained on a Shimadzu IR Prestige-21 using KBr
Table 1
Factors and their levels in the GAS crystallization.
Factors Symbols Uncoded levels
Low Medium High
Temperature (°C) T 25 35 45
PAR-to-MEF ratio R 3:1 4:1 5:1
MEF saturation (%) S 70 80 90
Journal of CO₂ Utilization 26 (2018) 212–220
disc method. Samples were scanned with a resolution of 1 cm−1 be-
tween 400 to 4000 cm−1.
Powder X-ray diffraction (PXRD) patterns were collected using a
Shimadzu XRD-6000 diffractometer (40 kV and 30 mA) using Cu Kα
radiation. Data were collected from 5° to 60° 2θ at step size of 0.01° 2θ.
MEF content in the sample was measured with a UV–visible spec-
trophotometer (Genesys 10S, Thermo Scientific). Approximately 3 mg
of dried solid sample was dissolved in 100 mL of phosphate buffer of pH
7.6. The amount of MEF in the solution was then determined by mea-
suring the absorbance at 335 nm. %MEF content was calculated using
the formula shown in Eq. (2):
mass of the MEF in particles
%MEF content = × 100
total mass of particles (2)
In vitro drug dissolution tests were conducted to evaluate dissolution
behavior of MEF–PAR cocrystals in comparison with those of un-
processed MEF, physical mixture, cocrystals by SE and marketed com-
bination drugs. Dissolution studies were performed using 900 mL of
sodium phosphate buffer of pH 7.6 at 37 °C and 200 rpm. Samples
equivalent to 5 mg of MEF were introduced into the dissolution
medium. A volume of 4 mL of aliquot was withdrawn and then filtered
through a syringe filter (0.45 μm) at 5, 10, 15, 20, 25, 30, 60, 90,120,
150, 180, 210, 240, 300 and 360 min. The amount of dissolved MEF
was evaluated by UV spectrophotometric method at 335 nm. The dis-
solution rate profiles were plotted as %MEF release from pure MEF,
physical mixture, marketed combination drugs and cocrystals produced
by SE and GAS process versus time in minute. %MEF released was
defined as the percentage of MEF dissolved into the dissolution medium
with respect to the amount of MEF loaded. For the purposes of com-
parison, the parameter used to evaluate drug dissolution properties of
the samples was the time at which 63.2% of MEF was dissolved (t63.2).
The t63.2 was obtained from the dissolution profile using linear inter-
(1) polation between two known dissolution points [33]. Dissolution tests
were conducted in triplicate for each GAS run, and average values of
t63.2 were reported.
3. Results and discussion
3.1. GAS cocrystallization feasibility
Feasibility studies were performed in order to check whether the
GAS technique is able to produce MEF–PAR cocrystals. MEF–PAR
sample produced by the GAS process at an operating temperature of
34.9 °C, a PAR-to-MEF molar ratio of 4.3 and a %MEF saturation of
86.8% was selected as a representative sample for verifying formation
of MEF–PAR cocrystals by DSC, FT-IR and XRD analysis. In addition,
the sample produced via GAS was also compared to those produced by
slow evaporation (SE) technique and a physical mixture of MEF and
PAR containing the same %MEF content as in the GAS product (22%
MEF).
3.1.1. Differential scanning calorimetry (DSC)
DSC analysis was employed to detect the cocrystal formation from
the difference in melting point of cocrystal in comparison with its own
constituents by endotherm phase [34]. DSC thermograms of MEF, PAR,
their respective physical mixture and their products obtained by SE and
GAS process are shown in Fig. 2. Melting point comparisons of starting
materials and cocrystals are presented in Table 2. The melting point of
pure MEF and that of PAR were determined to be 229.46 °C and
168.90 °C, respectively. Meanwhile, the physical mixture (containing
22.3% MEF) showed a sharp endothermic peak at 168.88 °C, corre-
sponding to the melting point of predominant PAR, and a slight en-
dothermic peak at 170.76 °C, resulting from the presence of MEF in the
mixture. The results indicated that there was no interaction between
MEF and PAR after simple mixing. Products obtained by GAS and SE
showed new melting points at 165.96 °C and 166.17 °C, respectively.
214
N. Wichianphong, M. Charoenchaitrakool
Fig. 2. DSC heating curves of (a) MEF, (b) PAR, (c) PM: physical mixture of
MEF and PAR, (d) MEF–PAR cocrystals by SE, and (e) MEF–PAR cocrystals by
GAS.
Table 2
DSC parameters of MEF, PAR, physical mixture and MEF–PAR cocrystals by
SE and GAS.
Samples Endothermic peak (°C)
MEF 229.46
PAR 168.90
Physical mixture 168.88, 170.76
MEF–PAR cocrystals by SE 166.17, 168.58
MEF–PAR cocrystals by GAS 165.96, 167.22
The new melting points of these products are different from those of the
respective pure components and indeed correspond to the melting point
of MEF–PAR cocrystals (160–165 °C) prepared by solvent evaporation
method using ethanol as solvent [21]. However, there was a second
endothermic peak for each of cocrystal by GAS and cocrystal by SE,
found around 167–168 °C. These peaks may result from eutectic melting
of MEF–PAR cocrystals and uncocrystallized PAR or may be due to the
melting point of excess uncocrystallized PAR in the final products.
3.1.2. Fourier transform infrared (FT-IR) spectroscopy
FT-IR spectroscopy was used to investigate molecular interactions
between MEF and PAR molecules in the formation of the cocrystal. The
formation of cocrystal can be confirmed by shifting of wavenumber or
changing in peak intensities of functional groups in comparison with
the cocrystal constituents [35]. The FT-IR spectra and lists of relevant
IR bands of MEF, PAR, their physical mixture and products obtained by
SE and GAS are presented in Fig. 3 and Table 3, respectively. The IR
bands of the physical mixture matched with those of MEF and PAR,
indicating there was no interaction between MEF and PAR after the
simple mixing process. However, the IR bands of MEF–PAR cocrystals
prepared by SE and GAS processes were different from those of the
individual components. Peaks at 1657 cm−1 (corresponding to C]O
stretching in PAR), 3311 cm−1 (corresponding to NeH stretching in
MEF) and 1576 cm−1 (corresponding to NeH bending in MEF) were
not observed for the cocrystals. The disappearance of these IR peaks
indicated the intermolecular interaction between MEF and PAR (i.e.,
the formation of MEF–PAR cocrystals).
3.1.3. Powder X-ray diffractometry (PXRD)
PXRD is a useful technique for identification of a new crystalline
phase. PXRD spectra of pure MEF, pure PAR, their physical mixture and
products obtained by GAS and SE are shown in Fig. 4. It was observed
that the PXRD pattern of the physical mixture showed a combination of
the diffractograms of both pure compounds presented in the mixture,
although the intensity of the characteristic peaks of MEF was lower due
to the high PAR-to-MEF ratio employed in the formulation. The
215
Journal of CO₂ Utilization 26 (2018) 212–220
diffractogram of cocrystals by SE showed several unique peaks at 2θ
values of 16.38, 25.69, 28.99, 29.45, 36.93 and 37.32 (as indicated by
star marks on Fig. 4), suggesting the formation of a new solid phase and
the cocrystals. However, in the case of cocrystals by the GAS process, no
unique peaks were observed. This could be due to errors associated with
X-ray diffraction determinations of NeH⋯O]C hydrogen bond geo-
metries as previously documented in the literature [36]. It was also
reported that the use of X-ray diffraction may not be able to detect the
cocrystal formation in some cases because suitable X-ray quality co-
crystals cannot always be produced and the exact location of the hy-
drogen atom may be ambiguous [37]. Using a variety of solid-state
characterization techniques, such as DSC and FT-IR combined tech-
nique, also has been recommended [37,38]. In addition, a decrease in
the peak intensities of cocrystals produced by GAS compared to their
respective starting components indicated a loss of crystallinity of co-
crystals by GAS. The results also suggested that cocrystals produced by
the GAS process had a lower crystallinity than those processed by SE.
Loss of drug crystallinity induced by GAS processing has been pre-
viously reported in the literature. It may occur to the rapid nucleation
of particles during GAS precipitation preventing rearrangement of the
drug molecules into the crystalline structure [8,39].
Based on the DSC and FT-IR results mentioned above, the MEF–PAR
products produced by SE and GAS were evidenced to be cocrystals.
Therefore, the effects of GAS process conditions were further in-
vestigated to find optimal conditions for the highest dissolution rate of
MEF in cocrystal in the next section.
3.2. Experimental design and statistical analysis
In this study, the first step in application of the Box–Behnken design
with response surface methodology is statistically designing experi-
ments and collecting data. The next step is to generate a response
surface model (i.e., mathematical relationship between the response
variable and independent variables) by response surface methodology.
The objectives of a response surface model are to describe the effects of
independent variables and determine the optimal conditions so as to
achieve the desired response [40].
3.2.1. Evaluation of fitted model
A design matrix consisting of 15 experimental runs with different
levels of process parameters along with the respective experimental and
predicted values of t63.2 is summarized in Table 4. All experimental runs
were conducted in triplicate, and the average values of t63.2 are re-
ported. The t63.2 values and respective %MEF contents of cocrystals by
GAS were found to be in the range of 12–24 min and of 19–32%, re-
spectively. By applying regression analysis on the experimental data, a
fully quadratic model for t63.2 value was developed in terms of uncoded
factors–namely, T (operating temperature), R (PAR-to-MEF molar ratio)
and S (%MEF saturation)–as shown in Eq. (3). It was observed that this
model was found to fit well, with a coefficient of determination (R2) of
97.01% and R2 (adj) of 91.63%.
t63.2 = 172.3 − 2.577 T − 28.18 R − 1.261 S + 0.0491 T*T + 2.867 R*R-
+ 0.0086nS*S + 0.0165 T*R − 0.01067 T*S + 0.033 R*S (3)
The R2 and R2 (adj) values were used to check the degree of fit (i.e., the
goodness of fit) of the proposed model. A model with larger values of R2
and R2 (adj) will give better correlation between input parameters and
response [41,42]. Joglekar and May [43] recommended that R2 should
be at least 80% for a preferable fit of a model. Consequently, it can be
assumed that this proposed model has high goodness of fit and can give
good prediction of experimental results.
Table 5 shows the ANOVA of the coefficients of the quadratic model
in Eq. (3). The larger the F-values and the smaller the associated p-
values, the higher is the significance of the corresponding term [44]. A
p-value less than 0.05 indicates that the model terms is statistically
N. Wichianphong, M. Charoenchaitrakool Journal of CO₂ Utilization 26 (2018) 212–220

Fig. 3. FT-IR spectra of (a) MEF, (b) PAR, (c) PM: physical mixture of MEF and PAR, (d) MEF–PAR cocrystals by SE, and (e) MEF–PAR cocrystals by GAS.

Table 3 Table 4
Wavenumber of functional groups of MEF, PAR, physical mixture and cocrystals Box–Behnken design matrix of three variables and the obtained %MEF content
by SE and GAS. and t63.2 values.
Group MEF PAR physical cocrystals cocrystals by Run T R S MEF content (%) t63.2 (minutes)
mixture by SE GAS
Theoretical actual Experimental Predicted from
NeH stretching 3311 3313 disappeared disappeared Eq. (3)
3325 3324 3325 3325
C]O stretching 1651 1651 1652 1653 1 25 3 80 34.73 31.77 + 1.10 22.56 + 1.43 21.80
1657 1658 disappeared disappeared 2 45 3 80 34.73 30.05 + 3.65 23.40 + 2.01 22.92
NeH bending 1576 1573 disappeared disappeared 3 25 5 80 24.20 22.00 + 0.48 16.97 + 1.82 17.42
1566 1566 1566 1566 4 45 5 80 24.20 19.49 + 2.16 18.47 + 1.50 19.20
5 25 4 70 28.52 25.89 + 1.83 18.19 + 0.69 17.80
6 45 4 70 28.52 23.27 + 1.65 22.04 + 3.00 21.38
7 25 4 90 28.52 26.73 + 1.17 16.78 + 1.51 17.41
8 45 4 90 28.52 24.99 + 4.43 16.36 + 1.06 16.72
9 35 3 70 34.73 24.01 + 1.00 18.82 + 2.19 19.91
10 35 5 70 24.20 21.80 + 1.75 15.33 + 0.62 15.19
11 35 3 90 34.73 31.46 + 1.28 16.61 + 0.32 16.72
12 35 5 90 24.20 23.96 + 1.51 14.44 + 0.63 13.32
13 35 4 80 28.52 21.84 + 0.50 12.77 + 0.51 12.56
14 35 4 80 28.52 21.86 + 1.63 12.43 + 0.33 12.56
15 35 4 80 28.52 22.41 + 0.31 12.52 + 0.61 12.56

Table 5
ANOVA for the response surface model for t63.2 of the cocrystals by the GAS
process.
Source DF Adj SS Adj MS F-value p-value

Fig. 4. PXRD patterns of (a) MEF, (b) PAR, (c) PM: physical mixture of MEF and Model 9 167.331 18.5923 18.04 0.003
T 1 4.162 4.1616 4.04 0.101
PAR, (d) MEF–PAR cocrystals by SE, and (e) MEF–PAR cocrystals by GAS.
R 1 32.724 32.7240 31.75 0.002
S 1 12.980 12.9795 12.59 0.016
T*T 1 88.999 88.9994 86.36 0.000
significant at the 95% confidence level [42,45,46]. The ANOVA of the R*R 1 30.351 30.3514 29.45 0.003
S*S 1 2.728 2.7282 2.65 0.165
quadratic model demonstrates that this model is statistically significant,
T*R 1 0.109 0.1089 0.11 0.758
as is evident from the model F-value of 18.04 and low p-value of 0.003 T*S 1 4.558 4.5582 4.42 0.089
(< 0.05). Two independent variables studied (R and S) and two R*S 1 0.436 0.4356 0.42 0.544
quadratic terms (T*T and R*R) significantly affect the t63.2 values, with Residual Error 5 5.153 1.0306
Lack-of-Fit 3 5.091 1.6970
small p-values (p < 0.05). The other terms are not significant because
Pure Error 2 0.062 0.0310
the p-values are greater than 0.05. Overall, the results show that PAR- Total 14 172.484
to-MEF molar ratio (R) and %MEF saturation (S) are the most sig-
nificant process parameters affecting t63.2 (p < 0.05). The significance Note: S = 1.01519, R2 = 97.01%, R2 (adj) = 91.63%.
for these effects followed the following order: PAR-to-MEF ratio
(R) > %MEF saturation (S). the straight line (R2 = 0.9701), showing that the deviation between the
The model accuracy was also checked through a parity plot showing predicted and experimental values was small and the proposed model
the relationship between the predicted and experimental t63.2 values, as was adequate in predicting t63.2.
presented in Fig. 5. It can be seen that the data points are spread near to

216
N. Wichianphong, M. Charoenchaitrakool
Fig. 5. Plot of predicted versus experimental t63.2 values.
3.2.2. Effects of independent variables on response
The visualization of the model equation can be accomplished by the
contour plots as presented in Fig. 6. The plots, generated from the
Box–Behnken design, were used to describe changes in t63.2 values of
cocrystals by GAS in terms of changes in two independent variables,
while the other independent variable was kept constant. Fig. 6(a) is the
contour plot showing the effects of PAR-to-MEF ratio and temperature
on the t63.2 at a fixed %MEF saturation of 80%. It can be observed that
middle-range temperatures provided the smallest t63.2 (i.e., less than
14 min) over a range of PAR-to-MEF ratio from 3.6 to 5.0. However, at
low PAR-to-MEF ratios (< 3.6) with decreasing PAR-to-MEF ratio, t63.2
begins to increase. This is most likely because a higher PAR-to-MEF
ratio in the starting solution resulted in a higher PAR content (i.e.,
lower %MEF content) in the final product (see Table 4). An increase in
PAR content in the starting solution may also cause an increase in co-
crystal formation and hence faster dissolution of cocrystals. PAR, a
more soluble compound when compared to MEF, is classified as class I
according to the BSC system, as it has high solubility in water [47], and
it can thus enhance the solubility of MEF; therefore the better wett-
ability of cocrystals in the higher PAR content could be linked to faster
dissolution of the cocrystals. In addition, use of lower or higher tem-
peratures than the middle-range temperatures will lead to an increased
t63.2. The value of t63.2 was found to increase as the temperature was
decreased from the middle-range temperatures down to 25 °C. This was
probably due to the effect of lower PAR precipitation (i.e., higher %MEF
content) at lower temperature. The effect of temperature on the amount
of PAR precipitation was shown in Table 4. A decrease in temperature
from 45 °C to 25 °C resulted in a slight increase of %MEF content from
30.05% to 31.77% (run nos. 2 and 1) and from 19.49% to 22.00% (run
nos. 4 and 3). Moreover, t63.2 was also found to increase when in-
creasing the temperature from 35 °C to 45 °C. This can be explained by
the effect of temperature on the resulting crystal size. Fusaro et al. [48]
Fig. 6. Contour plots of response t63.2 values: (a) PAR-to-MEF ratio versus temperature, (b) %MEF saturation versus temperature, and (c) %MEF saturation versus
PAR-to-MEF ratio.
217
Journal of CO₂ Utilization 26 (2018) 212–220
studied the precipitation of PAR using GAS and reported that an in-
crease in GAS crystallization temperature led to an increase of crystal
size of PAR. Since all the GAS products from this study contained more
than 68% PAR by wt., an increase in temperature could result in a
larger particle size of the product. A larger particle size in turn provides
a lower specific surface area exposed to the dissolution medium, re-
sulting in a lower dissolution time (i.e., higher t63.2).
Fig. 6(b) is the contour plot showing the effects of %MEF saturation
and temperature on the t63.2 at a fixed PAR-to-MEF molar ratio of 4:1.
This contour plot shows a similar trend to the previous plot (Fig. 6(a)).
For middle-range temperatures, the smallest t63.2 (i.e., less than 14 min)
can be obtained using %MEF saturations from 72 to 90%. A possible
reason for this is that a high %MEF saturation (72–90%) promoted
formation of cocrystal, which in turn favored the dissolution rate of
MEF. However, reducing %MEF saturation below 72% causes t63.2 to
increase. Moreover, t63.2 also increases with a decrease in temperature
from 35 °C to 25 °C due to the effect of lower %PAR content in the
product at low temperatures (as previously discussed). Also, t63.2 in-
creases when increasing the temperature from 35 °C to 45 °C due to the
effect of the larger cocrystal size that occurs at higher temperatures.
Fig. 6(c) is the contour plot showing the effects of %MEF saturation
and PAR-to-MEF ratio on t63.2 at a fixed temperature of 35 °C. At PAR-
to-MEF ratios lower than 4.5:1, a higher PAR-to-MEF ratio together
with high MEF saturation resulted in smaller t63.2. This is most likely
because higher %solid concentration promoted more MEF–PAR co-
crystal formation, resulting in higher dissolution time of MEF. How-
ever, increases in PAR-to-MEF ratio beyond a certain value (> 4.5) did
not lead to any further decrease in t63.2, and even caused a small in-
crease (i.e., an increase of t63.2 from less than 12 min to 12–13.5 min).
This is probably because the highest solid concentration in the starting
solution may result in fast precipitation of solutes and low cocrystal
formation and hence low dissolution rate.
3.2.3. Optimization of process and verification of results
The optimum value of t63.2 for the studied independent variables
was obtained using the response surface methodology. The optimal
conditions of the GAS process for the lowest t63.2 value (i.e., fastest
dissolution time) were found to be as follows: operating temperature of
34.9 °C, PAR-to-MEF molar ratio of 4.3:1 and 86.8% MEF saturation.
Using these conditions, the predicted t63.2 value was 11.82 min, with a
desirability value of 1.0000. In order to validate this prediction, ex-
periments using the predicted optimal conditions were carried out in
triplicate. The average t63.2value was found to be 11.38 + 0.99 min.
This experimental value of t63.2 (11.38 min) is close to the predicted
value (11.82 min). Such good agreement justifies the validity of the
model presented in Eq. (3).
N. Wichianphong, M. Charoenchaitrakool Journal of CO₂ Utilization 26 (2018) 212–220

Fig. 7. SEM images at ×100 magnification of (a) MEF, (b) PAR, (c) physical mixture of MEF and PAR, (d) sieved marketed combination drugs, (e) sieved MEF–PAR
cocrystals by SE, and (f) MEF–PAR cocrystals by GAS.

3.3. Characterization of cocrystal

The optimized cocrystals produced by GAS were further in-

vestigated for their microstructural and dissolution behaviors. The ob-
tained results were then compared to those of the physical mixture,
sieved cocrystals produced by SE technique and the marketed combi-
nation drugs.

3.3.1. Product morphology

Cocrystal size and morphology are important parameters when
evaluating the properties of cocrystals [49]. SEM analysis was per-
formed for pure MEF, pure PAR, physical mixture, sieved marketed
combination drugs and cocrystals by SE and GAS as shown in Fig. 7. It
was observed that pure MEF has multi-shaped particles with various
sizes, while pure PAR showed cube-like shapes with a large particle size
of approximately 300–500 μm. For the physical mixture, the char-
acteristic appearances of both MEF and PAR could be found, and some
of the MEF particles were well-dispersed over the PAR particle surface.
The sieved market combination drugs have irregularly shaped particles
with rough surfaces. PAR–MEF cocrystals by SE and GAS exhibited a
morphology of small plate-like crystals covered with some smaller
flakes, which is different from those of individual components and the
marketed drugs. One difference between the GAS and the SE samples
was that the cocrystals by GAS had a smaller particle size, as shown in
Fig. 8.
Fig. 9 shows the PSDs of pure MEF, cocrystals by GAS and sieved
cocrystals by SE analyzed using a Malvern Mastersizer. It was observed
that the cocrystals by GAS had a smaller particle size with a narrower
Fig. 9. PSD graphs of pure MEF, sieved cocrystals by SE and cocrystals by GAS.
particle size distribution and a higher specific surface area (average
particle size of 71.57 + 1.37 μm and specific surface area of
249.67 + 3.33 m2/kg) when compared to sieved cocrystal produced by
SE method (average particle size of 110.67 + 1.67 μm and specific
surface area of 171.33 + 2.23 m2/kg) and pure MEF (average particle
size of 93.23 + 2.63 μm and specific surface area of 277.17 + 1.77 m2/
kg). The presence of smaller particles produced by the GAS process is in
accordance to SEM images shown in Figs. 7 and 8. Due to the unique
properties of dense CO2 such as gas-like diffusivity and viscosity, low
surface tension, high compressibility combined with liquid-like density,
the expanded solution in the GAS process can reach a degree of

Fig. 8. SEM images at ×500 magnification of (a) sieved marketed combination drugs, (b) sieved MEF–PAR cocrystals by SE, and (c) MEF–PAR cocrystals by GAS.

218
N. Wichianphong, M. Charoenchaitrakool
Fig. 10. Dissolution profiles of pure MEF, physical mixture of MEF and PAR,
sieved MEF–PAR cocrystals by SE, sieved marketed combination drugs, and
MEF–PAR cocrystals by GAS.
Table 6
Time required to achieve 63.2% of MEF dissolution for pure MEF, physical
mixture, sieved marketed combination drugs, sieved cocrystals by SE and co-
crystals by GAS.
Samples Sample form MEF content (%) t63.2 (min)
Pure MEF Unprocessed 100 68.75
Physical mixture Unprocessed 22.30 31.90
Marketed drugs Sieved powder 53.32 + 1.16 26.37
Cocrystals by SE Sieved powder 22.16 + 2.61 60.62
Cocrystals by GAS Without sieving 22.38 + 0.57 11.38
supersaturation faster than the SE method, causing a high nucleation
rate. As a result, cocrystal precipitated by GAS was smaller than by the
SE method.
3.3.2. Dissolution studies
In vitro dissolution tests were conducted in order to verify the effect
of GAS cocrystallization technique on the enhancement of MEF dis-
solution rate. The dissolution curves of pure MEF, the physical mixture,
marketed combination drugs and cocrystals by SE and GAS are shown
in Fig. 10. Associated dissolution times t63.2 are given in Table 6. The
obtained results indicated that the cocrystal by GAS showed an im-
provement in dissolution rate of MEF and had the best dissolution
properties with a t63.2 value of 11.38 min. Based on the values of t63.2,
the dissolution rate of MEF in GAS product was found to be 6.0, 2.8, 2.3
and 5.3 times greater than pure MEF, the physical mixture, sieved
marketed combination drugs and sieved cocrystals produced by SE,
respectively.
The optimized cocrystals by GAS showed a dissolution rate mark-
edly higher than that of the marketed combination drug and also that of
the pure drug. In fact, MEF in cocrystals by GAS was completely dis-
solved within 60 min, while the sieved marketed combined drugs was
completely dissolved within 150 min, and the pure MEF was in-
completely dissolved even after 6 h. These results indicate that GAS
cocrystallization is an effective approach for improving the dissolution
rate of MEF. Greater dissolution of MEF from cocrystals can be attrib-
uted to the change in crystallinity pattern, morphology (i.e., crystal
habit), shape and size of cocrystals that lead to enhance solubility of
cocrystal in the dissolution medium [50,51].
Dissolution properties of cocrystals prepared by GAS were found to
be superior compared to those prepared by conventional slow eva-
poration. One possible reason for this is that the slow evaporation
method tends to produce larger primary particles with a wider size
distribution and a lower specific surface area, consequently giving a
lower dissolution rate of MEF. Another possible reason could be the
different packing and arrangement of MEF and PAR associated in the
cocrystal lattice, which favored the dissolution of MEF [50]. The im-
proved dissolution properties of cocrystals by GAS indicate that the
219
Journal of CO₂ Utilization 26 (2018) 212–220
dissolution performance of cocrystal was strongly dependent on the
cocrystallization technique applied and also the particle size of the
cocrystal tested. However, cocrystals by SE also showed an enhance-
ment of MEF dissolution, as they released 100% MEF within 5 h,
whereas the dissolution of pure MEF was incomplete even after 6 h.
In addition, it was observed that the dissolution rate of MEF from
the physical mixture was almost 2.2 times higher than that of the pure
drug. This may be due to a better dispersion of MEF particles in the
form of the physical mixture with PAR. The presence of MEF dispersed
over PAR particles was also confirmed by SEM image in Fig. 7(c). This
leads to reduction of the agglomeration of MEF particles during the
dissolution testing and consequently an increase in their specific surface
area exposed to the dissolution medium, aiding in the increase of dis-
solution rate. However, the release of MEF in the physical mixture was
incomplete even after 6 h.
4. Conclusion
Dual-drug cocrystals of MEF–PAR were successfully produced using
the GAS process to obtain an improved dissolution rate of MEF. The
study showed that Box–Behnken design was suitable to optimize GAS
process conditions for minimizing the dissolution time of MEF in the
resulting cocrystals. Statistical analysis showed that the dissolution
time was significantly influenced by PAR-to-MEF molar ratio and %
MEF saturation. The PAR-to-MEF ratio can be linked to the obtained
PAR content in the cocrystal product, which leads to promotion of the
dissolution rate of MEF. Higher %MEF saturation may promote more
cocrystal formation, but the highest %MEF saturation resulted in a
rapid precipitation and hence low cocrystal formation. The optimized
cocrystals by GAS showed comparatively higher dissolution rate than
the pure drug and also the marketed combination drug by 6.0 and 2.3-
fold, respectively. The obtained results indicate that the GAS cocrys-
tallization approach is an effective means to improve the physical
properties of the product, such as dissolution performance and crystal
habit of a drug substance, without chemical modification. In compar-
ison to the use of the physical mixture and the traditional slow eva-
poration method, the GAS process was proved to be the best method for
improving MEF dissolution rate through this dual-drug cocrystalliza-
tion, as it provided a smaller primary particle size and a larger specific
surface area of cocrystal. The dissolution properties of cocrystal were
found to be dependent on the cocrystallization technique used.
Conflict of interest
The authors have declared no conflict of interest.
Acknowledgements
The authors gratefully acknowledge the financial support of the
Kasetsart University Research and Development Institute (KURDI), the
National Research University Project of Thailand (NRU), the
Nanotechnology Center (NANOTEC) and the Thailand Research Fund
(Institutional Research Grant IRG5980004).
References
[1] T.T.L. Nguyen, N. Anton, T.F. Vandamme, Oral pellets loaded with nanoemulsions,
in: E. Andronescu, A.M. Grumezescu (Eds.), Nanostructures for Oral Medicine,
Elsevier, Amsterdam, 2017pp. 203–230.
[2] S.A. Ross, D.A. Lamprou, D. Douroumis, Engineering and manufacturing of phar-
maceutical co-crystals: a review of solvent-free manufacturing technologies, Chem.
Commun. 52 (2016) 8772–8786.
[3] D.D. Bavishi, C.H. Borkhataria, Spring and parachute: how cocrystals enhance so-
lubility, Prog. Cryst. Growth Charact. Mater. 62 (2016) 1–8.
[4] S. Aitipamula, R.B.H. Tan, Pharmaceutical co-crystals: crystal engineering and ap-
plication, in: E.R.T. Tiekink, J. Zukerman-Schpector (Eds.), Multi-Component
Crystals: Synthesis, Concepts, Function, Walter de Gruyter GmbH, Berlin, 2018pp.
1-31.
N. Wichianphong, M. Charoenchaitrakool
[5] K.V. Drozd, A.N. Manin, A.V. Churakov, G.L. Perlovich, Drug–drug cocrystals of
antituberculous 4-aminosalicylic acid: screening, crystal structures, thermo-
chemical and solubility studies, Eur. J. Pharm. Sci. 99 (2017) 228–239.
[6] R. Thipparaboina, D. Kumar, R.B. Chavan, N.R. Shastri, Multidrug co-crystals: to-
wards the development of effective therapeutic hybrids, Drug Discov. Today 21
(2016) 481–490.
[7] A. Erriguible, C. Neurohr, A.-L. Revelli, S. Laugier, G. Fevotte, P. Subra-Paternault,
Cocrystallization induced by compressed CO2 as antisolvent: simulation of a batch
process for the estimation of nucleation and growth parameters, J. Supercrit. Fluids
98 (2015) 194–203.
[8] C.A. Ober, S.E. Montgomery, R.B. Gupta, Formation of itraconazole/L-malic acid
cocrystals by gas antisolvent cocrystallization, Powder Technol. 236 (2013)
122–131.
[9] N. Ahmed, C.E. Mora-Huertas, C. Jaafar-Maalej, H. Fessi, A. Elaissari, Polymeric
drug delivery systems for encapsulating hydrophobic drugs, in: D. Douroumis,
A. Fahr (Eds.), Drug Delivery Strategies for Poorly Water-Soluble Drugs, John Wiley
& Sons, Ltd., United Kingdom, 2013pp. 151–197.
[10] L. Padrela, M.A. Rodrigues, S.P. Velaga, H.A. Matos, E.G. de Azevedo, Formation of
indomethacin–saccharin cocrystals using supercritical fluid technology, Eur. J.
Pharm. Sci. 38 (2009) 9–17.
[11] C.A. Ober, R.B. Gupta, Formation of itraconazole–succinic acid cocrystals by gas
antisolvent cocrystallization, AAPS PharmSciTech 13 (2012) 1396–1406.
[12] C. Neurohr, A.-L. Revelli, P. Billot, M. Marchivie, S. Lecomte, S. Laugier, S. Massip,
P. Subra-Paternault, Naproxen–nicotinamide cocrystals produced by CO2 anti-
solvent, J. Supercrit. Fluids 83 (2013) 78–85.
[13] R. Imchalee, M. Charoenchaitrakool, Gas anti-solvent processing of a new
sulfamethoxazole–L-malic acid cocrystal, J. Ind. Eng. Chem. 25 (2015) 12–15.
[14] G. Kotbantao, M. Charoenchaitrakool, Processing of ketoconazole–4-aminobenzoic
acid cocrystals using dense CO2 as an antisolvent, J. CO2 Util. 17 (2017) 213–219.
[15] K.R.S.S. Rao, M.V. Nagabhushanam, K.P.R. Chowdary, In vitro dissolution studies on
solid dispersions of mefenamic acid, Indian J. Pharm. Sci. 73 (2011) 243–247.
[16] D. Utami, I. Nugrahani, S. Ibrahim, Mefenamic acid–nicotinamide co-crystal syn-
thesized by using melt crystallization method and its solubility study, Asian J.
Pharm. Clin. Res. 10 (2017) 135–139.
[17] L. Fábián, N. Hamill, K.S. Eccles, H.A. Moynihan, A.R. Maguire, L. McCausland,
S.E. Lawrence, Cocrystals of fenamic acids with nicotinamide, Cryst. Growth Des.
11 (2011) 3522–3528.
[18] D. Utami, I. Nugrahani, S. Ibrahim, Formation and characterization of mefenamic
acid–nicotinamide cocrystal during co-milling based on X-ray powder diffraction
analysis, J. Appl. Pharm. Sci. 6 (2016) 75–81.
[19] K.E. Wittering, L.R. Agnew, A.R. Klapwijk, K. Robertson, A.J.P. Cousen,
D.L. Cruickshank, C.C. Wilson, Crystallisation and physicochemical property char-
acterisation of conformationally-locked co-crystals of fenamic acid derivatives,
CrystEngComm 17 (2015) 3610–3618.
[20] N. Wichianphong, M. Charoenchaitrakool, Statistical optimization for production of
mefenamic acid–nicotinamide cocrystals using gas anti-solvent (GAS) process, J.
Ind. Eng. Chem. 62 (2018) 375–382, http://dx.doi.org/10.1016/j.jiec.2018.01.017.
[21] C.D. Pathak, K.T. Savjani, A.K. Gajjar, J.K. Savjani, Cocrystal formation of para-
cetamol with indomethacin and mefenamic acid: an efficient approach to enhance
solubility, Int. J. Pharm. Pharm. Sci. 5 (2013) 414–419.
[22] V.K. Srirambhatla, A. Kraft, S. Watt, A.V. Powell, Crystal design approaches for the
synthesis of paracetamol co-crystals, Cryst. Growth Des. 12 (2012) 4870–4879.
[23] S. Hiendrawan, B. Veriansyah, E. Widjojokusumo, S.N. Soewandhi, S. Wikarsa,
R.R. Tjandrawinata, Simultaneous cocrystallization and micronization of para-
cetamol-dipicolinic acid cocrystal by supercritical antisolvent (SAS), Int. J. Pharm.
Pharm. Sci. 8 (2016) 89–98.
[24] M.S. Rao, G. Sailaja, Evaluation of efficacy and tolerability of acetaminophen
(paracetamol) and mefenamic acid and paracetamol combination as antipyretic in
pediatric patients with febrile illness: a comparative study, IOSR J. Dent. Med. Sci.
14 (2015) 5–9.
[25] S. Das, S.C. Sharma, S.K. Talwar, P.D. Sethi, Simultaneous spectrophotometric de-
termination of mefenamic acid and paracetamol in pharmaceutical preparations,
Analyst 114 (1989) 101–103.
[26] S.K. Abdul Mudalip, M.R. Abu Bakar, P. Jamal, F. Adam, Solubility and dissolution
thermodynamic data of mefenamic acid crystals in different classes of organic
solvents, J. Chem. Eng. Data 58 (2013) 3447–3452.
[27] N. Tafreshi, S. Sharifnia, S.M. Dehaghi, Box–Behnken experimental design for op-
timization of ammonia photocatalytic degradation by ZnO/oak charcoal composite,
Process Saf. Environ. Prot. 106 (2017) 203–210.
220
Journal of CO₂ Utilization 26 (2018) 212–220
[28] P.S. Collier, The interpretation of in vivo mean dissolution time data, in: L.Z. Benet,
G. Levy, B.L. Ferraiolo (Eds.), Pharmacokinetics, Plenum Press, New York, 1984pp.
403–404.
[29] J.R. Medina, J. Hernandez, M. Hurtado, In vitro release studies of carbamazepine
tablets and benzoyl metronidazole suspensions using the flow-through cell appa-
ratus and simulated gastrointestinal fluids, Int. J. App. Pharm. 9 (2017) 54–60.
[30] F. Langenbucher, Interpretation of in vitro–in vivo time profiles in terms of extent,
rate, and shape, in: J. Dressman, J. Krämer (Eds.), Pharmaceutical Dissolution
Testing, Taylor & Francis Group, New York, 2005pp. 251–279.
[31] K. Zhao, Y. Cheng, H. Liu, C. Yang, L. Qiu, G. Zeng, H. He, Extractive desulfurization
of dibenzothiophene by a mixed extractant of N,N-dimethylacetamide, N,N-diem-
thylformamide and tetramethylene sulfone: optimization by Box–Behnken design,
RSC Adv. 5 (2015) 66013–66023.
[32] M. Mourabet, A. El Rhilassi, H. El Boujaady, M. Bennani-Ziatni, A. Taitai, Use of
response surface methodology for optimization of fluoride adsorption in an aqueous
solution by Brushite, Arabian J. Chem. 10 (2017) S3292–S3302.
[33] J.-M. Cardot, B.M. Davit, In vitro–in vivo correlations: tricks and traps, AAPS J. 14
(2012) 491–499.
[34] I. Sopyan, A. Fudholi, M. Muchtaridi, I.P. Sari, Co-crystallization: a tool to enhance
solubility and dissolution rate of simvastatin, J. Young Pharm. 9 (2017) 183–186.
[35] H. Samsodien, M. Bapoo, T.I. Doms, Z. Harneker, A.S. Louw, I.C. Scheepers,
A.B. Sonday, B. Geldenhuys, FTIR, dissolution and anti-viral activity of nevirapine
co-crystals, Pharm. Anal. Acta 8 (2017) 1–10.
[36] R. Taylor, O. Kennard, Comparison of X-ray and neutron diffraction results for the
NeH⋯O]C hydrogen bond, Acta Cryst. 39 (1983) 133–138.
[37] N. Schultheiss, A. Newman, Pharmaceutical cocrystals and their physicochemical
properties, Cryst. Growth Des. 9 (2009) 2950–2967.
[38] S.-Y. Lin, Simultaneous screening and detection of pharmaceutical co-crystals by
the one-step DSC-FTIR microspectroscopic technique, Drug Discov. Today 22
(2017) 718–728.
[39] A.M. Barrett, F. Dehghani, N.R. Foster, Increasing the dissolution rate of itracona-
zole processed by gas antisolvent techniques using polyethylene glycol as a carrier,
Pharm. Res. 25 (2008) 1274–1289.
[40] D. Granato, V.M. de Araújo, Calado, The use and importance of design of experi-
ments (DOE) in process modelling in food science and technology, in: D. Granato,
G. Ares (Eds.), Mathematical and Statistical Methods in Food Science and
Technology, John Wiley & Sons, Ltd., New York, 2014pp. 1–18.
[41] S.-N. Nam, H. Cho, J. Han, N. Her, J. Yoon, Photocatalytic degradation of ace-
sulfame K: optimization using the Box–Behnken design (BBD), Process Saf. Environ.
Prot. 113 (2018) 10–21.
[42] H.K. Kim, J.G. Kim, J.D. Cho, J.W. Hong, Optimization and characterization of UV-
curable adhesives for optical communications by response surface methodology,
Polym. Test. 22 (2003) 899–906.
[43] A.M. Joglekar, A.T. May, Product excellence through design of experiments, Cereal
Foods World 32 (1987) 857–868.
[44] N. Anuar, A.F.M. Adnan, N. Saat, N. Aziz, R.M. Taha, Optimization of extraction
parameters by using response surface methodology, purification and identification
of anthocyanin pigments in Melastoma malabathricum fruit, Sci. World J. 2013
(2013) 1–10.
[45] M. Mondal, A. Ghosh, K. Gayen, G. Halder, O.N. Tiwari, Carbon dioxide bio-fixation
by Chlorella sp. BTA 9031 towards biomass and lipid production: optimization using
Central Composite Design approach, J. CO2 Util. 22 (2017) 317–329.
[46] F.J. Cui, Y. Li, Z.H. Xu, H.Y. Xu, K. Sun, W.Y. Tao, Optimization of the medium
composition for the production of mycelial biomass and exo-polymer by Grifola
frondosa GF9801 using response surface methodology, Bioresour. Technol. 97
(2006) 1209–1216.
[47] H.V. Chavda, C.N. Patel, I.S. Anand, Biopharmaceutics classification system, Syst.
Rev. Pharm. 1 (2010) 62–69.
[48] F. Fusaro, M. Mazzotti, G. Muhrer, Gas antisolvent recrystallization of paracetamol
from acetone using compressed carbon dioxide as antisolvent, Cryst. Growth Des. 4
(2004) 881–889.
[49] L. Padrela, M.A. Rodrigues, J. Tiago, S.P. Velaga, H.A. Matos, E.G. de Azevedo,
Tuning physicochemical properties of theophylline by cocrystallization using the
supercritical fluid enhanced atomization technique, J. Supercrit. Fluids 86 (2014)
129–136.
[50] P. Panzade, G. Shendarkar, S. Shaikh, P.B. Rathi, Pharmaceutical cocrystal of pir-
oxicam: design, formulation and evaluation, Adv. Pharm. Bull. 7 (2017) 399–408.
[51] G. Liu, Q. Lin, Y. Huang, G. Guan, Y. Jiang, Tailoring the particle microstructures of
gefitinib by supercritical CO2 anti-solvent process, J. CO2 Util. 20 (2017) 43–51.