International Journal of Pharmaceutics 485 (2015) 277287

Contents lists available at ScienceDirect

International Journal of Pharmaceutics

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

Inuence of excipients on solubility and dissolution of pharmaceuticals

Raphael Paus, Anke Prudic, Yuanhui Ji *
TU Dortmund, Department of Biochemical and Chemical Engineering, Laboratory of Thermodynamics, Emil-Figge Str. 70, D-44227 Dortmund, Germany

A R T I C L E I N F O

A B S T R A C T

Article history:
Received 26 January 2015
Received in revised form 27 February 2015
Accepted 2 March 2015
Available online 5 March 2015

In this work, solubilities and dissolution proles of the active pharmaceutical ingredients (APIs)
indomethacin and naproxen were measured in water in the presence of one excipient out of polyethylene
glycol (PEG) 2000, 6000 and 12000, polyvinylpyrrolidone (PVP) K 25 and mannitol. It was found that the
solubility of indomethacin and naproxen was increased with an addition of the selected excipients, which
was also predicted by the perturbed-chain statistical associating uid theory (PC-SAFT). The two-step
chemical-potential-gradientty model was applied to investigate the dissolution mechanism of
indomethacin and naproxen in water in the presence of the excipient. It was found that the dissolution
mechanisms of indomethacin and naproxen were changed by the presence of excipients. Although the
solubility of the API was increased by the addition of excipients, the dissolution rate of the API was
decreased in some cases. This was mainly due to the combination of the molecular interactions between
the API and the polymer with the inuence of the excipients on the kinetic part (rate constant of the
surface reaction or diffusion of the API or both) of API dissolution as function of PEG molar mass as well as
of the API type. Based upon the determined rate constants, the dissolution proles were modeled with a
high accuracy compared with the experimental data.
2015 Elsevier B.V. All rights reserved.

Chemical compounds studied in this article:

Indomethacin (PubChem CID: 3715)
Naproxen (PubChem CID: 156391)
Polyethylene glycol (PubChem CID: 174)
Polyvinylpyrrolidone (PubChem CID: 6917)
Mannitol (PubChem CID 6251)
Keywords:
Dissolution
Excipients
Thermodynamics
Solubility
PC-SAFT
Chemical potential gradient
Poorly soluble APIs

1. Introduction
Many active pharmaceutical ingredients (APIs) show a low
solubility and slow dissolution in aqueous solutions. As a result, the
oral bioavailability of these poorly soluble APIs is often quite low.
Therefore, several formulation strategies were developed to
enhance the aqueous solubility and dissolution rate of these APIs.
Within the formulation process, excipients are commonly used.
Different polymers, e.g., polyethylene glycol (PEG) and polyvinylpyrrolidone (PVP), were reported to be suitable formulation
materials for many poorly soluble APIs (Caron et al., 2010; Joshi
et al., 2004; Kochling et al., 2007; Lakshman et al., 2008;
Papadimitriou et al., 2012; Windbergs et al., 2009). Moreover,
sugars, e.g., mannitol, are often used for pharmaceutical formulations (Liao et al., 2005; Littringer et al., 2013). However, so far,
the detailed inuencing mechanism of these excipients on API
solubility and dissolution is far from being well explained. As the
dissolution process of many poorly soluble APIs often shows
complex dynamic characteristics, an accurate description of their

* Corresponding author. Tel.: +49 231 755 3199; fax: +49 231 755 2572.
E-mail addresses: Yuanhui.Ji@bci.tu-dortmund.de, yuanhuiji@aliyun.com (Y. Ji).
http://dx.doi.org/10.1016/j.ijpharm.2015.03.004
0378-5173/ 2015 Elsevier B.V. All rights reserved.

dissolution proles and an appropriate analysis of their dissolution

mechanism turn out to be difcult (DeAlmeida et al., 1997; Lu et al.,
2011). Additionally, the presence of excipients, which might
inuence the dissolution mechanism of APIs, complicates the
analysis and prediction of the specic API dissolution prole. Based
on the approach proposed by Noyes and Whitney, (1897), several
models, in most cases, empirical approaches (Costa et al., 2001;
DeAlmeida et al., 1997; Gibaldi and Feldman, 1967; Higuchi, 1961,
1963; Higuchi et al., 1958; Hopfenberg, 1976; Korsmeyer et al.,
1983; Mooney et al., 1981a,b) were developed to describe the
dissolution proles of APIs. Mooney et al. (1981a,b); Mooney et al.
(1981a,b) investigated the dissolution of several weak acid APIs
under unbuffered and buffered conditions from a rotating disk die
and proposed a model for describing the initial steady-state
dissolution rate of those APIs based on the Ficks second law of
diffusion. In this model, a diffusion-controlled mass transport of
the investigated APIs was assumed, in which simple, instantaneously established reaction equilibria were considered across a
postulated diffusion layer (Mooney et al., 1981a,b). However, for
the investigation on the inuence of excipients on API dissolution,
these models need to be combined with an appropriate
thermodynamic model to take into account the molecular
interactions between API and water, between API and the excipient

278

R. Paus et al. / International Journal of Pharmaceutics 485 (2015) 277287

as well as between the excipient and water. In previous works,

based on the work of Ji et al. (2010) and Lu et al. (2011), a two-step
chemical-potential-gradient model in which the molecular interactions of the systems were accounted for was developed to
describe the dissolution proles and to analyze the dissolution
mechanism of poorly soluble crystalline APIs and their formulations under various conditions (Ji et al., 2015; Paus et al., 2015).
Within this model, two consecutive steps, namely the surface
reaction and diffusion, were considered and expressed in terms of
the chemical potential gradient of the API (the thermodynamic
driving force) and the corresponding rate constants (from kinetic
aspect). In this paper, the two-step chemical-potential-gradient
model was used to analyze the inuencing mechanism of
excipients on the dissolution of APIs in water.
In this work, indomethacin and naproxen were selected as
model APIs. PEG with different molar masses (PEG 2000, PEG
6000 and PEG 12000), PVP with a molecular weight of 25,000 g/
mol (PVP K 25) and mannitol were selected as model excipients.
The chemical structures of the APIs and excipients are shown in
Fig. 1. In literature, PVP and PEG were reported to show a high
inuence on the aqueous solubility of poorly-soluble APIs
(Afrasiabi Garekani et al., 2003; Bettinetti and Mura, 1994;
Cadwallader and Madan, 1981; Mura et al., 1996). Mura et al.
(1996) investigated the inuence of PEGs with different molar
masses on the aqueous solubility of naproxen. They found that by
addition of 2 wt% of PEG 4000 in the solution, the solubility of
naproxen was increased by more than two times. As during API
dissolution, the amount of excipients is commonly quite small
compared to the amount of solvent (intestinal uid) and as a high
inuence of 2 wt% of PEG on API solubility was already reported in
literature, in this work, the solubility and dissolution proles of
indomethacin and naproxen in water in the presence of 2 wt% of
the model excipients were measured. As the perturbed-chain
statistical associating uid theory (PC-SAFT (Gross and Sadowski,
2001)) was already successfully applied to calculate the
thermodynamic properties of APIs in polymers (Prudic et al.,
2014a,b, 2015b), solvents and solvent mixtures (Ruether and
Sadowski, 2009, 2011) under various conditions, it was applied to
describe the inuence of excipients on API solubility (thermodynamic aspect). The two-step chemical-potential-gradient model
was applied to analyze the inuencing mechanism of excipients
on API dissolution (kinetic aspect) by accounting for the
interactions between the API and the excipient, the API and
water as well as the interactions between the excipient and water
via PC-SAFT (Gross and Sadowski, 2001). Finally the dissolution

proles of the APIs were modeled and compared with the

experimental ndings.
2. Theory
2.1. Two-step chemical-potential-gradient model
As introduced in previous works (Ji et al., 2010, 2015;
Paus et al., 2015), two main consecutive steps are involved in
the API dissolution process. In the rst step the disintegration of
the API crystals and the hydration of the API molecules take place.
This step is called surface reaction. For this step, the chemical
potential gradient of the API between the solid phase mSAPI and the
solidliquid interface mIAPI is the thermodynamic driving force of
the surface reaction. For the second step, the diffusion of the
hydrated API molecules from the solidliquid interface into the
bulk phase of the medium takes place. Here the chemical potential
gradient of the API between the solidliquid interface mIAPI and the
bulk phase mBAPI , is the thermodynamic driving force of diffusion.
The rates of surface reaction and diffusion are described in
Eqs. (1) and (2).
!
mSAPI mIAPI
dcBAPI 1
(1)
 ks

JAPI V 
A
dt
RT
RT

JAPI kd

mIAPI
RT

mBAPI

RT

(2)

In Eq. (1), JAPI is the dissolution rate of the API in mol/(m2 s), V is the
volume of the dissolution medium in m3; A is the surface area of
the dissolving API in contact with the dissolution medium in m2,
cBAPI is the concentration of the API in the bulk phase of the medium
in mol/m3 and t is the time in s. In Eqs. (1) and (2), T is the
temperature in K and R is the ideal gas constant in J/(mol K). mSAPI ,

mIAPI and mBAPI are the chemical potentials of the API in the solid
phase, at the solidliquid interface and in the bulk phase in J/mol,
respectively. ks and kd are the rate constants of the surface reaction
and diffusion in mol/(m2 s), respectively.
Based on the values of ks and kd within Eqs. (1) and (2), the rate
controlling step of API dissolution can be determined. The API
dissolution is controlled by the surface reaction in the case of
ks < kd and it is controlled by diffusion if kd < ks. In the case of ks
equal to or similar to kd both steps are important for API
dissolution.

Fig. 1. Chemical structures of indomethacin (a), naproxen (b), PEG (c), PVP (d) and mannitol (e).

R. Paus et al. / International Journal of Pharmaceutics 485 (2015) 277287

The chemical potential of the API at the solidliquid interface is

determined by the Statistical Rate Theory developed by Dejmek
and Ward (Dejmek and Ward, 1998; Ward et al., 1982). This theory
describes the instantaneous API transport rate across the
solidliquid interface according to Eq. (3).
(
!
!)
mSAPI  mIAPI
mIAPI  mSAPI
JAPI ke exp
 exp
(3)
RT
RT
In Eq. (3), ke is the equilibrium exchange rate of the API molecules
between the solid and the liquid phase in mol/(m2 s). As a detailed
derivation of the equilibrium exchange rate can be found in
literature (Dejmek and Ward, 1998; Ji et al., 2015; Paus et al., 2015;
Ward et al., 1982), it is only briey introduced in this paper. The
equilibrium exchange rate of the API between the solid and liquid
phase per unit area of the interface can be described according to
Eq. (4).
r
p
v3 2
ke xLAPI a1 af T xLAPI a2 ;af 0:51
d
(4)

In Eq. (4), xLAPI is the API solubility in the media in mole fraction, a1
and a2 are the proportionality constants, ; is a constant fraction of
molecules that strike the solid surface, af describes the fraction of
the area of the solidliquid interface available for the transport of
molecules from the medium, T is the temperature of the system in
K, v is the stirring speed in round/s, n is the kinematic viscosity
of the medium in m2/s and d is the thickness of the diffusion layer
in m.
The chemical potential of the API in the solid phase mSAPI , as used
in Eqs. (1) and (3) can be calculated by the solidliquid equilibrium
(SLE). Here, the chemical potential of the API in the solid phase is
equal to that in its saturated solution mIAPI (Eq. (5)).

mSAPI mLAPI mL0API  RT lnaLAPI

(5)

In Eq. (5), aLAPI is the API activity in the saturated solution, mL0API is
the chemical potential of the API in the standard state (here the
pure liquid API).
The chemical potentials of the API at the solidliquid interface
mIAPI and in the bulk phase mBAPI required in Eqs. (1)(3) can be
expressed as shown in Eqs. (6) and (7).

mIAPI mL0API RTlnaIAPI

(6)

mBAPI mL0API RTlnaBAPI

(7)

In

Eq. (6), aIAPI is the activity of the API at the solidliquid interface.
Eq. (7), aBAPI is the activity of the API in the bulk phase of the

In
medium. In this work, the required API activities (see Eqs. (5)(7))
are calculated using the thermodynamic model PC-SAFT (Gross
and Sadowski, 2002b) according to Eq. (8).
aAPI xAPI g API

(8)

In Eq. (8), xAPI is the mole fraction of the API in the solution and g API
the corresponding API activity coefcient. Based on Eqs. (4)(7),
Eqs. (1)(3) can be described in terms of the API activity (see Eqs.
(9)(11)).


(9)
JAPI ks ln aLAPI  ln aIAPI


JAPI kd ln aIAPI  ln aBAPI

JAPI

xLAPI

p
a1 af T xLAPI a2 ;af 0:51

(10)
r

v3 2
d
n

aLAPI aIAPI

aIAPI aLAPI

In

Eqs.

(9)(11),

the

279

parameters

K1

(K1 = a1af)

and

K2

(K 2 a2 ;af d ), as well as kd and ks were tted to the experimental

API dissolution proles based on Eqs. (9)(11).
Finally, based on Eqs. (9) and (10), the dissolution rate of the API
can be described in terms of the rate constant of the whole API
dissolution kT and the chemical potential gradient of the API
between the solid phase and the bulk phase according to Eq. (12).
!


mSAPI  mBAPI
kT ln aLAPI  ln aBAPI
JAPI kT
(12)
RT
In Eq. (12), kT can be expressed in terms of the surface-reaction rate
constant ks and the diffusion rate constant kd (see Eq. (13)).
kT

1
1=kd 1=ks

(13)

2.2. Solidliquid equilibrium

Based on the SLE (Prausnitz et al., 1998), the temperaturedependent solubility of an API in a solution was calculated
according to Eq. (14).
(
!
1
DhSL
T
L
0API
1  SL
xAPI L exp 
RT
g API
T 0API
!
!)
SL
SL
SL
Dc0pAPI
T
T
(14)
ln 0API  0API 1

R
T
T
In Eq. (14), xLAPI is the solubility of the API in the solution in mole
fraction, g LAPI is the activity coefcient of the API in the saturated
solution, which is used to account for the inuence of the solvent
and excipient on API solubility in the solution. T SL
0API , Dh0API and
DcSL
are
the
caloric
properties
of
the
API,
namely
the melting
p0API
SL

temperature in K, melting enthalpy in J/mol and the difference in

the solid and liquid API heat capacities in J/(mol K), respectively.
T is the temperature in K and R is the universal ideal gas constant.
In this work, the caloric properties of the selected APIs were taken
from previous works, as summarized in Table 1.
As the excipients, the API and water show different physicochemical properties due to their different molecular interactions
and shapes, the activity coefcient of the API is often quite different
from one and has to be considered within the solubility
calculations (see Eq. (14)). In this work, the API activity coefcient
was calculated as function of temperature and medium composition using PC-SAFT (Gross and Sadowski, 2001).
2.3. PC-SAFT
Within PC-SAFT, the residual Helmholtz energy ares of the
system is calculated as a sum of various contributions (Gross and
Sadowski, 2001, 2002a) resulting from the repulsion of the hard
chain (ahc), van der Waals attractions (adisp, disp stands for
dispersion) and hydrogen bondings (aassoc, where assoc stands for
association), according to Eq. (15).

Table 1
Melting temperature, melting enthalpy and the difference in the solid and liquid
heat capacities of indomethacin and naproxen used within this work.
API

(11)
Indomethacin
Naproxen

T SL
0API
[K]

DhSL
0API

DcSL
pAPI

[kJ/mol]

[J/(mol K)]

433.25
429.47

39.3
31.5

116.95
87.44

Ref.

(Paus et al., 2015)

(Paus et al., 2015)

280

R. Paus et al. / International Journal of Pharmaceutics 485 (2015) 277287

ares ahc adisp aassoc

(15)

2.5. PC-SAFT binary interaction parameters

In this work, the binary interaction parameters between the
API and water, the API and PVP K 25 as well as those between
mannitol and water were taken from literature (Held et al., 2013;
Paus et al., 2015,b; Prudic et al., 2014a,b). For PVP K 25/water
system, a temperature-independent binary interaction parameter
was estimated by Prudic et al. (2015a) to describe the water
sorption of PVP K 25 (vaporliquid equilibrium) at 298.15 K. In
this work, in order to improve the accuracy of predicted API
solubility in water/PVP K 25 solution, temperature-dependent
binary interaction parameters between PVP K 25 and water were
estimated by tting them to the solubility data of indomethacin in
water containing 2 wt% of PVP K 25 and these parameters were
further used for predicting the solubility of naproxen in water
containing 2 wt% of PVP K 25. As proposed by Prudic et al. (2014a)
the binary interaction parameter between indomethacin and PEG
with different molar masses can be set to zero. As the API
solubility in water in the presence of PEG was found to be strongly
dependent on the PEG molar mass, a binary interaction parameter
between water and PEG as function of PEG molar mass was
considered according to Eq. (21) and it was tted to indomethacin
solubility in PEG/water solutions.

Within PC-SAFT, each molecule is described as a chain of

and a segment diameter
segments with a segment number of mseg
i
of s i. The van der Waals interactions are described by the
dispersion energy parameter ui/kB. Here, kB is the Boltzmann
constant in J/K. For molecules, which can form hydrogen bonds,
two additional parameters, the association-energy parameter
eAi Bi =kB and the association-volume parameter kAi Bi , are required.
Additionally, the number of association sites Nassoc
within a
i
molecule has to be determined.
To describe thermodynamic properties of binary mixtures, e.g.,
mixtures between API, excipient and water, the BerthelotLorentz
mixing rules are applied for the segment diameter and the
dispersion energy of the mixture of different molecules. The
segment diameter and the dispersion energy for a mixture of two
components i and j are given in Eqs. (16) and (17).

s ij


1
s sj
2 i

uij 1  kij

(16)

p
ui uj

(17)

kPEG; water k  MPEG 0:086

A correction for the deviation of the interaction energy of unlike

segments from the geometric mean is made by the binary
interaction parameter kij, as introduced in Eq. (17). In some cases,
this parameter is determined as function of temperature
(see Eq. (18)).
kij T kij; slope  T=K  298:15 kij; 298:15 K

(18)

For systems in which cross association occurs, the crossassociation interactions between two associating components can
be described by the combination rules proposed by Wolbach and
Sandler (Eqs. (19) and (20)) (Wolbach and Sandler, 1998).

e A i Bj

1 Ai Bi Aj Bj 
e e
2

(19)

kA i Bj

)3
(
p 2  p
s i s j

kAi Bi kAj Bj 
s i s j

(20)

(21)

In Eq. (21), k is a parameter which was estimated as 0.23 

104 mol/g and MPEG is the molar mass of the polymer PEG in
g/mol. Additionally, a temperature-dependent binary interaction
parameter between naproxen and PEG was tted to the solubility
data of naproxen in the PEG 12000/water solution and it was
further used for the prediction of naproxen solubility in the PEG
2000/water and PEG 6000/water solutions. The binary interaction
parameter between the selected APIs and mannitol was set to zero.
All the binary interaction parameters between API and water and
between API and the excipient as well as between the excipient
and water are summarized in Table 3.
3. Materials
Crystalline indomethacin (purity 99%) was purchased from
SigmaAldrich Co., LLC (Germany). Naproxen (purity >99%) was
purchased from TCI Deutschland Co., LLC (Germany). PEG 2000
was purchased from Merck (Germany), PEG 6000 from Prolabo
VWR (Germany), and PEG 12000 from SigmaAldrich Co., LLC
(Germany). PVP K 25 (Kollidon1 K 25) was purchased from BASF
(Ludwigshafen, Germany). Mannitol (purity >98%) was purchased
from SigmaAldrich Co., LLC (Germany). All substances were used
as obtained without any further purication. Water was ltered,
deionized and distilled using a Millipore purication system and
used for all experiments.

2.4. PC-SAFT pure-component parameters

In this work, the PC-SAFT pure-component parameters of
indomethacin, naproxen, water and the selected excipients were
taken from literature (Fuchs et al., 2006; Held et al., 2013; Prudic
et al., 2015a, 2014a,b; Stoychev et al., 2009) and are summarized
in Table 2.

Table 2
Molar mass and PC-SAFT pure-component parameters of naproxen, indomethacin, PEG 2000, 6000 and 12000, PVP K 25, mannitol and water used within this work.
Substance

Naproxen
Indomethacin
PEG 2000
PEG 6000
PEG 12000
PVP K 25
Mannitol
Water
a

mseg
i
()

si

ui/kB

eAi Bi =kB

kAi Bi

()

(K)

(K)

()

Nassoc
i
()

Ref.

(g/mol)
230.26
357.79
2000
6000
12000
25700
182.17
18.015

8.110
14.283
101.2
303.6
607.2
1045.21
7.230
1.2047

2.939
3.535
2.899
2.899
2.899
2.710
2.935
2.793a

229.45
262.79
204.6
204.6
204.6
205.59
227.03
353.945

934.19
886.44
1799.8
1799.8
1799.8
0
5000
2425.671

0.02
0.02
0.02
0.02
0.02
0.0451
0.1
0.0451

2/2
3/3
2/2
2/2
2/2
231/231
6/6
1/1

(Prudic et al., 2014b)

(Prudic et al., 2014a)
(Stoychev et al., 2009)
(Stoychev et al., 2009)
(Stoychev et al., 2009)
(Prudic et al., 2015a; Prudic et al., 2014a)
(Held et al., 2013)
(Fuchs et al., 2006)

The expression swater = 2.7927 + 10.11  exp(0.01775  T)  1.417  exp(0.01146  T) was used (Fuchs et al., 2006).

R. Paus et al. / International Journal of Pharmaceutics 485 (2015) 277287

4. Methods

281

Table 4
Measured kinematic viscosities (including standard deviations) of the selected
media at 310.15 K.

4.1. API solubility measurement

Solubility measurements of indomethacin and naproxen in
water in the presence of one excipient out of PEG 2000, PEG 6000,
PEG 12000, PVP K 25 and mannitol were carried out in 100 ml
glass vessels with a heating jacket in a temperature range of
288.15313.15 K. For this purpose, an excess of the crystalline API
was added into each medium (water containing 2 wt% of each
excipient) and mixed with a magnetic stirrer at 600 rpm. After 96 h
(thermodynamic equilibrium was ensured by a constant API
concentration), the solubility was measured at each temperature.
For sampling, syringes (10 ml) and needles were used. To prevent
API nucleation (if necessary), both, the syringe and the needle was
preheated prior to sampling. The sample (23 ml) was ltered by a
PES lter with a mesh size of 0.45 mm and then immediately
analyzed by the UVvis spectroscopy (Analytic Jena Specord 210
Plus spectrophotometer, Jena, Germany) at a wavelength of 331 nm
for naproxen and 320 nm for indomethacin according to the
procedure described in previous works (Paus et al., 2015). Each
measurement was performed in a triplicate and the mean value
was reported.
4.2. Kinematic viscosity measurement of solutions
According to Eq. (11) the viscosities of the solutions in the
presence of the excipients were required for dissolution modeling.
Therefore, the viscosities of the aqueous media containing 2 wt% of
each excipient were measured using a MGW Lauda viscosimeter
(Lauda-Knigshofen, Germany) at 310.15 K (accuracy 0.01 K)
according to the approach proposed by Ubbelohde (1937). Each
measurement was performed in triplicate and the mean value is
reported. The temperature-dependent viscosity data of pure water
were taken from literature (Kestin et al., 1978) and correlated by
the following function: n/(m2/s) = 2.248  1010 (T/ C)2 3.195 
108 (T/ C) + 1.572  106.
4.3. API dissolution measurement
The intrinsic dissolution proles of indomethacin and naproxen
in water in the presence of 2 wt% of each excipient (PEG 2000,
6000, 12000, PVP K 25 and mannitol) were measured using a
rotating disk system (USP II). For this purpose a cylindrical tablet

System

n [mm2/s]

Deviation

Water/mannitol
Water/PEG 2000
Water/PEG 6000
Water/PEG 12000
Water/PVP K 25

0.493
0.843
0.995
1.178
0.967

0.012
0.008
0.032
0.007
0.003

( = 8 mm) of pure API (200 mg  1 mg) was pressed with a

compressive force of 2 kN into the disk using a manual tablet
press device (Brtsch/Regger AG, Zurich, Switzerland). Within an
experiment the disk was rotated at a stirring speed of 50 rpm in the
prepared media at 310.15 K (accuracy 0.3 K). The temperature and
the stirring speed were monitored during each experiment. Each
media was prepared by dissolving a dened mass of each excipient
in water to generate a solution containing 2 wt% of each excipient.
Sample analysis was conducted in spectrometer ow cells which
were circulated with the dissolution media at 25 ml/min using a
piston pump (SOTAX CY 7-50, Allschwil, Switzerland). API
concentrations were measured at the same wavelengths as used
in the solubility measurements. Data analysis was performed by
the use of the WinSOTAX plus dissolution software (SOTAX Co.,
LLC, Allschwil, Switzerland). Each measurement was performed in
a triplicate and the mean value was reported.
5. Results and discussions
5.1. Kinematic viscosity of the selected media
The kinematic viscosities of the selected media at 310.15 K were
measured according to the procedures described in Section 4.2.
The measured results including the standard deviations are
summarized in Table 4.
For a better comparison of the measured kinematic viscosities
of different media, the measured results and the kinematic
viscosity of pure water at 310.15 K are also shown in Fig. 2. As
shown in Fig. 2, the kinematic viscosities of the aqueous solutions
in the presence of 2 wt% PVP K 25, PEG 2000, PEG 6000 and PEG
12000 are obviously higher than that of pure water. For the media
containing selected PEGs with different molar masses, the
viscosity increases with an increase in the molar mass of PEG.
Here, the viscosity of water containing 2 wt% PEG 12000 is around

Table 3
Binary interaction parameters kijs between API and water and between API and the excipient as well as between the excipient and water used within this work.
Component i

Component j

kij(T) = kij  (T/K  298.15)  kij,

Indomethacin

Mannitol
PEG 2000
PEG 6000
PEG 12000
PVP K 25
Water

0
0
0
0
0.000633  (T/K  298.15)  0.09653
0.0001691  (T/K  298.15)  0.05948

this work
(Prudic et al., 2014a)
(Prudic et al., 2014a)
(Prudic et al., 2014a)
(Prudic et al., 2014b)
(Paus et al., 2015)

Naproxen

Mannitol
PEG 2000
PEG 6000
PEG 12000
PVP K 25

this work
this work
this work
this work
(Prudic et al., 2014b)

Water

0
0.043  (T/K  298.15) + 0.07955
0.043  (T/K  298.15) + 0.07955
0.043  (T/K  298.15) + 0.07955
0.000128  (T/K  298.15)  0.09154
8
0.0002269  (T/K  298.15) + 0.006475

Mannitol
PEG 2000
PEG 6000
PEG 12000
PVP K 25

0.00022  (T/K  298.15)  0.0492

0.04
0.052
0.154
0.002667  (T/K  298.15)  0.148

(Held et al., 2013)

this work
this work
this work
(Prudic et al., 2015a)/this work

Water

298.15 K

Ref.

(Paus et al., 2015)

282

R. Paus et al. / International Journal of Pharmaceutics 485 (2015) 277287

Fig. 2. Kinematic viscosities of the selected media at 310.15 K.

1.7 times as high as that of pure water at 310.15 K. The aqueous

media containing 2 wt% PVP K 25 and that containing 2 wt% PEG
6000 show nearly the same kinematic viscosity. In the case of the
media in the presence of 2 wt% mannitol, the kinematic viscosity is
lower than that of pure water.
The measured viscosities of the selected media at 310.15 K
are further used for the dissolution mechanism analysis based on
Eqs. (9)(11).

5.2. Measured API solubility in water in the presence of excipients

The measured aqueous solubility of indomethacin and
naproxen in water in the presence of 2 wt% of each excipient is
shown in Fig. 3(a)(d). As shown in this gure, the solubility of
both APIs is obviously inuenced by the presence of excipients.
In all cases, the API solubility is a function of temperature as
well as of excipient type in water. For a comparison the

temperature-dependent solubility of indomethacin and naproxen

in pure water (taken from a previous work (Paus et al., 2015)) is
also included in Fig. 3. As observed in Fig. 3(a), the presence of
2 wt% of PVP K 25 increases the solubility of indomethacin in
water. At a temperature of 313.15 K, indomethacin solubility is
1.4 times as high as that in pure water. The observed increase in
the aqueous solubility of indomethacin by adding PVP was also
reported in literatures for several other APIs (Afrasiabi Garekani
et al., 2003; Bettinetti and Mura, 1994; Cadwallader and Madan,
1981; Tros de Llarduya et al., 1998). This shows that our
experimental observations agree with the results in the
literatures. Generally, an increase in API solubility in water was
observed with an increase in PVP concentration in the aqueous
media (Afrasiabi Garekani et al., 2003; Tros de Llarduya et al.,
1998). This phenomenon might be mainly attributed to the high
ability of PVP to form hydrogen bonds with the APIs (Afrasiabi
Garekani et al., 2003). The presence of 2 wt% of mannitol in water
has no obvious inuence on the solubility of indomethacin. Only a
slight increase in the solubility of indomethacin is observed at
lower temperatures. As shown in Fig. 3(b), the presence of 2 wt%
of PEG in water has obvious inuence on the solubility of
indomethacin as function of PEG molar mass and it follows the
order: PEG 2000 > PEG 6000 > PEG 12000. The solubility of
indomethacin in water in the presence of 2 wt% of PEG 2000 is
around 3 times at 313.15 K as high as that in pure water. In case of
the presence of 2 wt% of PEG 6000 in water the indomethacin
solubility is 1.5 times as high as that in pure water at 313.15 K and
it remains nearly the same in case of the presence of 2 wt% of
PEG 12000 as that in pure water.
As shown in Fig. 3(c) and (d) almost the same trend is observed
for the solubility of naproxen in water in the presence of 2 wt% of
each excipient. Here again, the solubility of naproxen in water in

Fig. 3. (a) Solubility of indomethacin in water (Paus et al., 2015) (gray hollow stars), in water with 2 wt% of mannitol (light hollow circles), and in water with 2 wt% of PVP K 25
(dark gray triangles). (b) Solubility of indomethacin in water (Paus et al., 2015) (gray stars), in water with 2 wt% of PEG 2000 (gray circles), in water with 2 wt% of PEG 6000
(light gray pentagons) and in water with 2 wt% of PEG 12000 (light gray hollow squares). (c) Solubility of naproxen in water (Paus et al., 2015) (gray stars), in water with 2 wt%
of mannitol (hollow circles), in water with 2 wt% of PVP K 25 (dark gray triangles). (d) Solubility of naproxen in water (Paus et al., 2015) (gray stars), in water with 2 wt% of PEG
2000 (gray circles), in water with 2 wt% of PEG 6000 (light gray pentagons) and in water with 2 wt% of PEG 12000 (light gray squares). The lines represent the modeling results
of API solubility in water and in water with 2 wt% of PEG 2000, 6000 and 12000 (full lines), in water with 2 wt% of mannitol (dotted lines), as well as in water with 2 wt% of PVP
K 25 (dashed lines) using PC-SAFT.

R. Paus et al. / International Journal of Pharmaceutics 485 (2015) 277287

the presence of 2 wt% of PVP K 25 is around 2 times at 313.15 K as

high as that in pure water. Bettinetti and Mura (1994) also
observed an increase in the aqueous API solubility in the presence
of PVP. This shows again that our observations agree with the
results in the literature. The enhancement of naproxen solubility
by adding 2 wt% of PVP K 25 in water might be also attributed to the
high ability of PVP to form hydrogen bonds with naproxen. For the
media in the presence of 2 wt% of mannitol the solubility of
naproxen is only slightly inuenced. Here, the solubility of
naproxen is around 1.17 times as high as that in pure water at
310.15 K. As shown in Fig. 3(d), the solubility of naproxen in water
with 2 wt% of PEG is highly increased and is more increased with
increasing temperature compared to that in pure water. Here again,
the solubility of naproxen in water with different PEGs is a function
of PEG molar mass. With an increase in the molar mass of the PEG,
the solubility of naproxen is decreased (PEG 2000 > PEG
6000 > PEG > 12000). At 313.15 K, the solubility of naproxen in
water in the presence of 2 wt% of PEG 2000 is around 3.4 times as
high as that in pure water. For the media with 2 wt% of PEG 6000
and with 2 wt% of PEG 12000, the naproxen solubility is around 2.8
times and 2.4 times, respectively as high as that in pure water.
Mura et al. (1996) investigated the effect of PEGs with different
molar masses of 4000, 6000 and 20000 on the aqueous solubility
of naproxen. They also observed an increase in the API solubility
with a decrease in the molar mass of the PEG which was mainly
attributed to the formation of hydrogen bonds between the API and
the polymer resulting in an improved local API solubilization.
Additionally, the API-polymer interactions were found to be
decreased with an increase in the polymer molar mass (Mura et al.,
1996). This also helps to explain why the aqueous solubility of
indomethacin and naproxen was decreased with increasing PEG
molar mass.
Compared to indomethacin solubility (see Fig. 3(a) and (b)), the
presence of PEG with a higher molar mass shows a signicantly
higher inuence on the aqueous naproxen solubility.
5.3. Modeling results of API solubility in water in the presence of
excipients
The solubility of indomethacin and naproxen in all systems was
modeled by PC-SAFT (see Sections 2.22.5) and the modeling
results are also included in Fig. 3(a)(d). As presented in Fig. 3, the
modeling results are in good accordance to the experimental data.
To evaluate the accuracy of the modeling results, the maximum
relative deviation (MRD) as well as the average relative deviation
(ARD) between the calculated and experimental API solubility
were calculated according to Eqs. (22) and (23).
MRD 100  maxi1; nexp j

xcalc; i  xexp; i
j
xexp; i

exp
xcalc; i  xexp; i
1 X
j
j
nexp i1
xexp; i

283

PVP K 25 a higher deviation (13.48%) between the modeling results

and the experimental data is observed. However, one has to keep in
mind that the modeling results for the naproxen/water/PVP K25
system were fully predicted as all the pure-component parameters
and binary interaction parameters were determined from binary
systems (Prudic et al., 2014b, 2015a). Additionally it has to be
mentioned that the aqueous solubility of these APIs is very low for
all of the selected systems (xLAPI < 1.5  105 mol/mol in the
selected temperature range).
5.4. Measured API dissolution proles in water in the presence of
excipients
The experimental dissolution proles of indomethacin and
naproxen were measured within a time range of 240 min at
310.15 K according to the procedures described in Section 4.3. The
measured dissolution proles of indomethacin and naproxen in
water in the presence of 2 wt% of PVP K 25, 2 wt% of mannitol and
in the presence of 2 wt% of PEGs are shown in Fig. 4(a). For a
comparison, the dissolution proles of indomethacin and
naproxen in pure water (taken from a previous work
(Paus et al., 2015)) are also included in this gure. As presented
in Fig. 4(a), the dissolution proles of both APIs, indomethacin
and naproxen are inuenced by the presence of the excipients.
From Fig. 4(a) it becomes obvious, that the dissolution rate
of indomethacin is increased by the presence of 2 wt% of
PVP K 25 and mannitol compared to that in pure water. Here,
indomethacin shows the highest dissolution rate in water in the
presence of 2 wt% of mannitol. As also shown in Fig. 4(a), the
dissolution rate of indomethacin in water containing different
PEGs decreases with an increase in the molar mass of PEG. The
same trend is observed for the inuence of the different PEGs on
naproxen dissolution.
As shown in Fig. 4(a), naproxen dissolution rate increases with
addition of both excipients PVP K 25 and mannitol compared to
that in pure water, and the highest dissolution rate of naproxen is
observed in PVP K 25 solution. In the presence of 2 wt% of PEG
6000, naproxen dissolution rate is similar to that in pure water
while naproxen dissolution rate decreases with an addition of 2 wt
% of PEG 12000 in water. In the presence of 2 wt% of PEG 2000 the
observed naproxen dissolution rate is higher than that in pure
water. The experimental investigations on indomethacin and
naproxen dissolution in the media containing different excipients
show that the presence of excipients plays an important role on the
dissolution rate of the selected APIs. In the next section, the
inuence mechanism of the excipients on indomethacin and
naproxen dissolution will be discussed in detail.

(22)
Table 5
The maximum relative deviation (MRD) as well as the average relative deviation
(ARD) of the calculated and experimental API solubilities in the selected media.

ARD 100 

(23)

In Eqs. (22) and (23), xexp, i is the experimental API solubility in

mole fraction, xcalc, i is the calculated API solubility in mole fraction
and nexp is the number of experimental solubility data points. The
results are summarized in Table 5.
As presented in Table 5, the MRD and ARD analysis shows that
the modeling results are in good accordance to the experimental
data. In all cases except for the naproxen/water/PVP K 25 system
the ARD is below 12%, which indicates the good performance of
PC-SAFT for the solubility calculation of indomethacin and
naproxen in water in the presence of 2 wt% of selected excipients.
Only in the case of naproxen in water in the presence of 2 wt% of

API

System

ARD (%)

MRD

Indomethacin

Water
Water/mannitol
Water/PVP K 25
Water/PEG 2000
Water/PEG 6000
Water/PEG 12000

7.55a
11.66
4.74
8.45
10.69
7.91

14.52
30.12
10.44
14.58
19.97
21.79

Naproxen

Water
Water/mannitol
Water/PVP K 25
Water/PEG 2000
Water/PEG 6000
Water/PEG 12000

2.38a
10.41
13.48
5.98
7.20
4.98

6.72
16.91
32.86
13.38
13.45
9.28

taken from (Paus et al., 2015).

284

R. Paus et al. / International Journal of Pharmaceutics 485 (2015) 277287

Fig. 4. (a) Dissolution proles of selected APIs at 310.15 K and 50 rpm: indomethacin in water (dark gray stars) (Paus et al., 2015), in water with 2 wt% of PVP K 25 (hollow
triangles), in water with 2 wt% of mannitol (light gray pentagons), in water with 2 wt% of PEG 2000 (light gray cycles), in water with 2 wt% of PEG 6000 (hollow hexagons) and
in water with 2 wt% of PEG 12000 (light gray squares); naproxen in water (black stars) (Paus et al., 2015), in water with 2 wt% of PVP K 25 (gray triangles), in water with 2 wt% of
mannitol (gray pentagons), in water with 2 wt% of PEG 2000 (gray cycles), in water with 2 wt% of PEG 6000 (gray hexagons) and in water with 2 wt% of PEG 12000 (gray
squares); (b) surface-reaction rate constant ks and diffusion rate constant kd of indomethacin and naproxen in water (Paus et al., 2015), in water with 2 wt% of mannitol, in
water with 2 wt% of PVP K 25, in water with 2 wt% of PEG 2000, in water with 2 wt% of PEG 6000 and in water with 2 wt% of PEG 12000 at 310.15 K and 50 rpm.

5.5. Dissolution mechanism of indomethacin and naproxen in water in

the presence of excipients
The dissolution mechanism of indomethacin and naproxen in
water in the presence of 2 wt% of each excipient was analyzed
using the two-step chemical-potential-gradient model as introduced in Section 2.1. As shown in Eq. (12), the dissolution rate of
indomethacin and naproxen JAPI depends on two main factors.
Firstly the thermodynamic driving force, which is described by the
chemical potential gradient of the API mSAPI  mBAPI and secondly the
kinetic aspect, which is described in terms of the rate constant kT.
The latter is a function of the surface-reaction rate constant ks and
the diffusion rate constant kd of the API. The aim of this section is to
discuss the inuence of the excipients on the dissolution of the API
in water from both aspects, the thermodynamic driving force and
the kinetic aspect.
The surface-reaction rate constant ks and the diffusion rate
constant kd of indomethacin and naproxen were tted to their
experimental dissolution proles in the media containing 2 wt% of
each excipient (Section 5.4). For the data analysis, it was assumed
that the volume of the dissolution medium, the viscosity of the
medium and the surface area of the tablet remain constant during
each measurement.

The corresponding values of ks and kd for indomethacin and

naproxen dissolution are shown in Fig. 4(b). The rate constants of
indomethacin and naproxen dissolution in water were taken from
a previous work (Paus et al., 2015). The detailed values of the
parameters ks, kd and K1 and K2 (as introduced in Section 2.1) and
the ARD (%) between the calculated and experimental dissolution
data are also listed in Table 6.
Although the thermodynamic driving force mSAPI  mBAPI of
indomethacin in water and in water with 2 wt% of mannitol is
nearly the same (see Fig. 3(a)) at 310.15 K, the dissolution rate of
indomethacin in water in the presence of mannitol is obviously
higher than that in pure water, as observed in Fig. 4(a). The main
reason for this is the higher rate constants of both surface reaction
and diffusion for indomethacin dissolution in water in the
presence of 2 wt% of mannitol compared to those in pure water
as shown in Fig. 4(b). In mannitol solution, the surface-reaction
rate constant of indomethacin is around 1.8 times and the diffusion
rate constant is around 1.2 times as high as those in pure water. The
enhancement of the diffusion rate constant for indomethacin
dissolution in mannitol solution might be related to the decrease in
the viscosity of mannitol solution compared to that of pure water
(see Fig. 2). The increase of the surface-reaction rate constant of
indomethacin dissolution in mannitol solution might be due to

Table 6
Fitted parameters ks,kd, K1, K2 and calculated rate constant of the whole API dissolution kT and the average relative deviation (ARD) between the calculated and the
experimental dissolution proles.
System

kd

k1 = a1af

K 2 a2 ;af d

Indomethacin/water (Paus et al., 2015)

Indomethacin/water/PEG 2000
Indomethacin/water/PEG 6000
Indomethacin/water/PEG 12000
Indomethacin/water/PVP K 25
Indomethacin/water/mannitol
Naproxen/water (Paus et al., 2015)
Naproxen/water/PEG 2000
Naproxen/water/PEG 6000
Naproxen/water/PEG 12000
Naproxen/water/PVP K 25
Naproxen/water/mannitol

9.69  1007
1.01 1006
8.13  1007
5.86  1007
4.86  1007
1.20  1006
3.70  1006
1.40  1006
1.26  1006
1.09  1006
2.17  1006
4.09  1006

1.35  1003
9.61 1006
8.13  1007
8.85  1004
9.74  1004
5.90  1003
8.99  1004
3.37  1004
9.08  1004
1.68  1004
3.83  1004
2.88  1005

1.82  1004
4.39  1005
4.64  1005
8.63  1005
1.36  1003
2.15  1004
1.39  1004
3.39  1004
3.15  1004
3.89  1004
1.75  1003
1.05  1004

ks

kT

ARD%

2.80  1007
4.67  1007
2.73  1007
1.35  1007
1.05  1006
5.08  1007
1.73  1006
3.24  1006
3.01 1006
2.77  1006
4.49  1006
1.82  1006

2.17  1007
3.19  1007
2.05  1007
1.10  1007
3.32  1007
3.57  1007
1.18  1006
9.78  1007
8.88  1007
7.82  1007
1.46  1006
1.26  1006

17.52
11.87
5.21
12.76
16.61
6.59
16.35
12.25
8.15
8.84
14.35
14.43

R. Paus et al. / International Journal of Pharmaceutics 485 (2015) 277287

the improvement of hydration of indomethacin by addition of

mannitol. Nevertheless the dissolution of indomethacin in
both pure water and mannitol solution is controlled by
surface reaction which indicates a slow disintegration of
indomethacin crystals from the crystal lattice and a slow hydration
of the indomethacin molecules. In the case of indomethacin
dissolution in water in the presence of 2 wt% of PVP K 25, the
observed increase in indomethacin dissolution rate compared
to that in pure water can be explained from two different
aspects. Firstly the increase in the thermodynamic driving
force mSIND  mBIND due to the higher solubility of indomethacin
in PVP K 25 solution (see Fig. 3(a)) and secondly the increase in
he surface-reaction rate constant. As shown in Fig. 4(b), the
surface-reaction rate constant of indomethacin in PVP K 25
solution is around 4 times as high as that in pure water. However,
the diffusion rate constant of indomethacin molecules into the
bulk phase of the PVP K 25 solution is decreased by the presence of
2 wt% of PVP K 25 in water. This can be related to the increase in
the viscosity of the media (see Fig. 2) and the interactions
between PVP K 25 and indomethacin. This leads to a change
in the dissolution mechanism of indomethacin in PVP K 25 solution
from surface-reaction controlled in pure water to diffusion
controlled.
As observed in Fig. 4(a), the dissolution rate of indomethacin
in water in the presence of 2 wt% of PEG 2000 is slightly increased
compared to that in pure water. This again can be related to two
main aspects: rstly the increase in the thermodynamic driving
force due to the higher solubility of indomethacin in PEG 2000
solution than that in pure water (see Fig. 3(b)) and secondly the
increase in the surface-reaction rate constant from the kinetic
aspect (1.6 times as high as that in pure water as shown in
Fig. 4(b)) due to an improvement of the disintegration of
indomethacin crystals from the crystal lattice and of the
hydration of the indomethacin molecules. It was found in
Fig. 4(b) that the diffusion rate constant of indomethacin in
PEG 2000 solution is nearly the same as that in pure water. For the
dissolution of indomethacin in water in the presence of 2 wt% of
PEG 6000 and PEG 12000, the dissolution rate of indomethacin is
slower than that in water in the presence of 2 wt% of PEG 2000.
For the dissolution of indomethacin in PEG 6000 solution, the
dissolution rate of indomethacin is nearly the same as that in pure
water. Although the thermodynamic driving force of indomethacin dissolution in PEG 6000 and PEG 12000 solutions is increased
due to the higher solubility of indomethacin in these solutions,
the rate constants from the kinetic aspect are decreased due to
the addition of the two excipients. As shown in Fig. 4(b), both the
surface-reaction rate constant and the diffusion rate constant of
indomethacin dissolution are decreased with an addition of PEG
12000. For the dissolution of indomethacin in PEG 12000
solution, the surface-reaction rate constant of indomethacin
dissolution is nearly half of that in pure water, which might be
due to the inuence of PEG 12000 on the mass transport of
indomethacin molecules at the solidliquid interface. Additionally, the diffusion rate constant of indomethacin dissolution is
decreased due to the combination of the molecular interactions
between the API and the polymer with the increase in the
viscosity of the medium with addition of PEG 12000. The decrease
in both rate constants nally leads to a slower indomethacin
dissolution in PEG 12000 solution compared to that in pure water.
For the dissolution of indomethacin in all PEG solutions, it is
controlled by the surface reaction, which shows that the
dissolution rate of indomethacin can be further enhanced mainly
by the improvement of the disintegration of indomethacin
crystals from the crystal lattice and of the hydration of the
indomethacin molecules.

285

As observed in Fig. 4(a), the dissolution rate of naproxen in

PVP K 25 solution and in mannitol solution is obviously higher
than that in pure water. From the dissolutionmechanism
analysis this can be explained by two main aspects. Firstly the
higher thermodynamic driving force due to the increase in the
naproxen solubility in water with addition of both excipients
(see Fig. 3(c)) and secondly the increase in the surface-reaction
rate constant of naproxen in PVP K 25 solution and in both
rate constants of naproxen in mannitol solution from kinetic
aspect. In the presence of 2 wt% of PVP K 25 the corresponding
surface-reaction rate constant of naproxen is nearly 2.6 times
as high as that in pure water. Although the diffusion-rate
constant of naproxen is decreased with an addition of PVP K
25, which leads to a diffusion-controlled dissolution process of
naproxen, the increase in the surface-reaction rate constant as
well as in thermodynamic driving force of naproxen dissolution
leads to the higher naproxen dissolution rates than that in
pure water.
For naproxen dissolution in mannitol solution, the surfacereaction rate constant of naproxen dissolution is nearly the same as
that in pure water. However, the improved diffusion of naproxen
molecules into the bulk phase of the media due to the lower
viscosity of the medium (see Fig. 2) leads to a higher dissolution
rate compared to that in pure water.
As observed in Fig. 4(a), the same trend for the inuence of
PEGs on naproxen dissolution in water is observed as that on
indomethacin dissolution. With an increase in the molar mass of
the PEG, the dissolution rate of naproxen decreases. In PEG 2000
solution, the dissolution rate of naproxen is slightly higher
than that in pure water. While in PEG 6000 solution, the
observed naproxen dissolution rate is nearly identical to that in
pure water. For naproxen dissolution in PEG 12000 solution, the
dissolution rate of naproxen is decreased compared to that in
pure water although the solubility of naproxen is increased
with an addition of PEG 12000 (see Fig. 3(d)). For the dissolution
of naproxen in water in the presence of PEGs, the rate constants
from kinetic aspect play an important role on naproxen
dissolution. Although the thermodynamic driving force is
increased due to the higher solubility of naproxen in water
in the presence of PEGs (see Fig. 3(d)), the diffusion rate
constant is highly decreased with the addition of PEGs,
which leads to a change in the dissolution mechanism from
surface-reaction controlled to diffusion controlled. Although the
surface-reaction rate constant of naproxen in PEG 2000 solution
is nearly 2 times as high as that in pure water, the diffusion rate
constant is decreased. For naproxen dissolution in PEG 6000
solution, the diffusion rate constant is only one third of that
in water. However, there is an increase in the corresponding
surface-reaction rate constant, which explains why the observed
naproxen dissolution rate in PEG 6000 solution is similar to that
in pure water. For naproxen dissolution in PEG 12000 solution,
the decrease in both the surface-reaction rate constant and the
diffusion rate constant (below one third of that in pure water)
of naproxen leads to the lower dissolution rate of naproxen
(see Fig. 4(b)).
The dissolution-mechanism analysis shows that the dissolution
rate of poorly soluble indomethacin and naproxen can be improved
from both thermodynamic aspect (solubility) and kinetic aspect
(rate constants of surface reaction or diffusion or both) with
addition of mannitol and PVP K 25.
For API dissolution in PEG solutions, although the solubility of
the APIs may be obviously increased, their dissolution rates can be
decreased due to the change in the surface-reaction rate constant
and diffusion rate constant of the APIs as function of PEG molar
mass as well as of API type.

286

R. Paus et al. / International Journal of Pharmaceutics 485 (2015) 277287

Fig. 5. (a) Dissolution proles of indomethacin (dark gray stars) (Paus et al., 2015), in water with 2 wt% of PVP K 25 (hollow triangles), in water with 2 wt% of mannitol (light
gray pentagons), in water with 2 wt% of PEG 2000 (light gray cycles), in water with 2 wt% of PEG 6000 (hollow hexagons) and in water with 2 wt% of PEG 12000 (light gray
squares) from a rotating disk at 310.15 K and 50 rpm; (b) dissolution proles of naproxen in water (black stars) (Paus et al., 2015), in water with 2 wt% of PVP K 25 (gray
triangles), in water with 2 wt% of mannitol (gray pentagons), in water with 2 wt% of PEG 2000 (gray cycles), in water with 2 wt% of PEG 6000 (gray hexagons) and in water with
2 wt% of PEG 12000 (gray squares) from a rotating disk at 310.15 K and 50 rpm; the dotted black line, the dashed black line, the full black line, the dashed gray line, the full gray
line and the dotted gray line represent the calculated dissolution proles of the APIs in water, in water with 2 wt% of mannitol, in water with 2 wt% of PVP K25, in water with
2 wt% of PEG 2000, in water with 2 wt% of PEG 6000, in water with 2 wt% PEG 12000 using the two-step chemical-potential-gradient model, respectively.

The dissolution proles of indomethacin and naproxen in water

in the presence of each excipient were calculated according to
Eq. (12). The determined rate constants of the surface reaction ks
and diffusion kd were used for the calculation of the rate constant
of the whole API dissolution kT according to Eq. (13). The detailed
values of ks, kd, and kT are summarized in Table 6. The calculated
dissolution proles of indomethacin and naproxen in water in the
presence of each excipient are shown in Fig. 5(a) and (b). For
comparison, the dissolution proles of indomethacin and
naproxen in pure water are also included in Fig. 5. As presented
in Fig. 5(a) and (b), the calculated dissolution proles are in good
accordance to the experimental data, which is also veried by the
calculated low ARDs as summarized in Table 6. This shows that the
two-step chemical-potential-gradient model can be used to well
represent the dissolution proles of the selected APIs in water in
the presence of excipients.

but decreases the diffusion rate constant of the APIs, which might
be due to the combination of the molecular interactions between
the API and the polymer with the higher viscosity of PVP K25
solution than that of pure water. For API dissolution in PEG
solutions, although the solubility of the APIs were observed to be
obviously increased, their dissolution rates were decreased due to
the change in the surface-reaction rate constant and diffusion rate
constant of the APIs as function of PEG molar mass as well as of
API type.
Based on the obtained rate-constants of surface reaction and
diffusion, the dissolution proles of indomethacin and naproxen
in all of the selected systems were calculated in good accordance
to the experimental ndings. This work shows that the two-step
chemical-potential-gradient model combined with PC-SAFT can
be used to analyze the inuencing mechanism of different
excipients on the dissolution of poorly-soluble APIs and to
calculate the API dissolution proles in water in the presence of
excipients with a high accuracy compared to the experimental
data.

6. Conclusions

Acknowledgments

In this work, the temperature-dependent solubility of

indomethacin and naproxen and the dissolution proles of these
two APIs were measured in water in the presence of 2 wt% of each
excipient out of PVP K 25, PEG 2000, PEG 6000, PEG 12000 and
mannitol. It was found that the solubility of indomethacin and
naproxen was increased with addition of the selected excipients,
which was also predicted by the PC-SAFT. It was also found that
the presence of 2 wt% of both, mannitol and PVP K 25 increases
the dissolution rate of both APIs. However, in the presence of
different PEGs, a decrease in the dissolution rates of both APIs was
observed with an increase in the molar mass of the PEG. With the
analysis of the dissolution mechanism of indomethacin and
naproxen in water in the presence of each excipient using the
two-step chemical-potential-gradient model, it was found that
the dissolution mechanisms of indomethacin and naproxen can
be changed by the addition of the different excipients. Although
the presence of mannitol showed a slight inuence on API
solubility (the thermodynamic driving force of API dissolution),
the API dissolution rate was obviously increased due to the
increase of both rate constants of surface reaction and diffusion.
The presence of PVP K 25 mainly increases the thermodynamic
driving force and the surface-reaction rate constant of the APIs

We would like to acknowledge the nancial support from the

Alexander von Humboldt Foundation (Dr. Yuanhui Ji) as well as
that from the CLIB Graduate Cluster Industrial Biotechnology
(Anke Prudic). We would like to thank Lisa Vahle for assisting with
some solubility measurements. We would also like to thank the
reviewers for their helpful advice.

5.6. Calculation of API dissolution proles in water in the presence of

excipients

References
Afrasiabi Garekani, H., Sadeghi, F., Ghazi, A., 2003. Increasing the aqueous solubility
of acetaminophen in the presence of polyvinylpyrrolidone and investigation of
the mechanisms involved. Drug Dev. Ind. Pharm. 29, 173179.
Bettinetti, G., Mura, P., 1994. Dissolution properties of naproxen in combinations
with polyvinylpyrrolidone. Drug Dev. Ind. Pharm. 20, 13531366.
Cadwallader, D.E., Madan, D.K., 1981. Effect of macromolecules on aqueous
solubility of cholesterol and hormone drugs. J. Pharm. Sci. 70, 442446.
Caron, V., Bhugra, C., Pikal, M.J., 2010. Prediction of onset of crystallization in
amorphous pharmaceutical systems: phenobarbital, nifedipine/PVP, and
phenobarbital/PVP. J. Pharm. Sci. 99, 38873900.
Costa, P., Manuel, J., Lobo, S., 2001. Modeling and comparison of dissolution proles.
Eur. J. Pharm. Sci. 13, 123133.
DeAlmeida, L.P., Simoes, S., Brito, P., Portugal, A., Figueiredo, M., 1997. Modeling
dissolution of sparingly soluble multisized powders. J. Pharm. Sci. 86, 726732.
Dejmek, M., Ward, C.A., 1998. A statistical rate theory study of interface
concentration during crystal growth or dissolution. J. Chem. Phys. 108,
86988704.

R. Paus et al. / International Journal of Pharmaceutics 485 (2015) 277287

Fuchs, D., Fischer, J., Tumakaka, F., Sadowski, G., 2006. Solubility of amino acids:
inuence of the pH value and the addition of alcoholic cosolvents on aqueous
solubility. Ind. Eng. Chem. Res. 45, 65786584.
Gibaldi, M., Feldman, S., 1967. Establishment of sink conditions in dissolution rate
determinations theoretical considerations and application to
nondisintegrating dosage forms. J. Pharm. Sci. 56, 1238.
Gross, J., Sadowski, G., 2001. Perturbed-chain SAFT: an equation of state based
on a perturbation theory for chain molecules. Ind. Eng. Chem. Res. 40,
12441260.
Gross, J., Sadowski, G., 2002a. Application of the perturbed-chain SAFT equation of
state to associating systems. Ind. Eng. Chem. Res. 41, 55105515.
Gross, J., Sadowski, G., 2002b. Modeling polymer systems using the perturbed-chain
statistical associating uid theory equation of state. Ind. Eng. Chem. Res. 41,
10841093.
Held, C., Sadowski, G., Carneiro, A., Rodrguez, O., Macedo, E.A., 2013. Modeling
thermodynamic properties of aqueous single-solute and multi-solute sugar
solutions with PC-SAFT. AIChE J. 59, 47944805.
Higuchi, T., 1961. Rate of release of medicaments from ointment bases containing
drugs in suspension. J. Pharm. Sci. 50, 874.
Higuchi, T., 1963. Mechanism of sustained-action medication theoretical analysis
of rate of release of solid drugs dispersed in solid matrices. J. Pharm. Sci. 52,
1145.
Higuchi, W., Parrott, E.L., Wurster, D.E., Higuchi, T., 1958. Investigation of drug
release from solids II. Theoretical and experimental study of inuences of bases
and buffers on rates of dissolution of acidic solids. J. Am. Pharm. Assoc. 47,
376383.
Hopfenberg, H.B., 1976. Controlled release from erodible slabs, cylinders, and
spheres. Abstr. Pap. Am. Chem. Soc. 62.
Ji, Y.H., Ji, X.Y., Liu, C., Feng, X., Lu, X.H., 2010. Modelling of mass transfer coupling
with crystallization kinetics in microscale. Chem. Eng. Sci. 65, 26492655.
Ji, Y.H., Paus, R., Prudic, A., Lbbert, C., Sadowski, G., 2015. A novel approach for
analyzing the dissolution mechanism of solid dispersions. Pharm. Res. doi:
http://dx.doi.org/10.1007/s11095-015-1644-z.
Joshi, H.N., Tejwani, R.W., Davidovich, M., Sahasrabudhe, V.P., Jemal, M., Bathala, M.
S., Varia, S.A., Serajuddin, A.T.M., 2004. Bioavailability enhancement of a poorly
water-soluble drug by solid dispersion in polyethylene glycolpolysorbate 80
mixture. Int. J. Pharm. 269, 251258.
Kestin, J., Sokolov, M., Wakeham, W.A., 1978. Viscosity of liquid water in range of 8  C
to 150  C. J. Phys. Chem. Ref. Data 7, 941948.
Kochling, J.D., Miao, H., Young, C.R., Looker, A.R., Shannon, M., Montgomery, E.R.,
2007. Understanding the degradation pathway of a poorly water-soluble drug
formulated in PEG-400. J. Pharm. Biomed. Anal. 43, 16381646.
Korsmeyer, R.W., Gurny, R., Doelker, E., Buri, P., Peppas, N.A., 1983. Mechanisms of
solute release from porous hydrophilic polymers. Int. J. Pharm. 15, 2535.
Lakshman, J.P., Cao, Y., Kowalski, J., Serajuddin, A.T., 2008. Application of melt
extrusion in the development of a physically and chemically stable high-energy
amorphous solid dispersion of a poorly water-soluble drug. Mol. Pharm. 5,
9941002.
Liao, X., Krishnamurthy, R., Suryanarayanan, R., 2005. Inuence of the active
pharmaceutical ingredient concentration on the physical state of mannitol
implications in freeze-drying. Pharm. Res. 22, 19781985.
Littringer, E.M., Noisternig, M.F., Mescher, A., Schroettner, H., Walzel, P., Griesser, U.J.,
Urbanetz, N.A., 2013. The morphology and various densities of spray dried
mannitol. Powder Technol. 246, 193200.

287

Lu, X.H., Ji, Y.H., Liu, H.L., 2011. Non-equilibrium thermodynamics analysis and its
application in interfacial mass transfer. Sci. China Chem. 54, 16591666.
Mooney, K., Mintun, M., Himmelstein, K., Stella, V., 1981a. Dissolution kinetics of
carboxylic acids I: effect of pH under unbuffered conditions. J. Pharm. Sci. 70,
1322.
Mooney, K., Mintun, M., Himmelstein, K., Stella, V., 1981b. Dissolution kinetics of
carboxylic acids II: effect of buffers. J. Pharm. Sci. 70, 2232.
Mura, P., Manderioli, A., Bramanti, G., Ceccarelli, L., 1996. Properties of solid
dispersions of naproxen in various polyethylene glycols. Drug Dev. Ind. Pharm.
22, 909916.
Noyes, A.A., Whitney, W.R., 1897. The rate of solution of solid substances in their own
solutions. J. Am. Chem. Soc. 19, 930934.
Papadimitriou, S.A., Barmpalexis, P., Karavas, E., Bikiaris, D.N., 2012. Optimizing the
ability of PVP/PEG mixtures to be used as appropriate carriers for the
preparation of drug solid dispersions by melt mixing technique using articial
neural networks: I. Eur. J. Pharm. Biopharm. 82, 175186.
Paus, R., Ji, Y.H., Braak, F., Sadowski, G., 2015. Dissolution of crystalline
pharmaceuticalsexperimental investigation and thermodynamic modeling.
Ind. Eng. Chem. Res. 54, 731741.
Prausnitz, J.M., Lichtenthaler, R.N., de Azevedo, E.G., 1998. Molecular
Thermodynamics of Fluid-phase Equilibria. Pearson Education.
Prudic, A., Ji, Y.H., Sadowski, G., 2014a. Thermodynamic phase behavior of API/
polymer solid dispersions. Mol. Pharm. 11, 22942304.
Prudic, A., Kleetz, T., Korf, M., Ji, Y.H., Sadowski, G., 2014b. Inuence of copolymer
composition on the phase behavior of solid dispersions. Mol. Pharm. 11,
41894198.
Prudic, A., Ji, Y.H., Lbbert, C., Sadowski, G., 2015. Phase behavior of API/polymer
formulations as function of relative humidity. in preparation.
Prudic, A., Lesniak, A.K., Ji, Y.H., Sadowski, G., 2015b. Thermodynamic phase
behaviour of indomethacin/PLGA formulations. Eur. J. Pharm. Biopharm. doi:
http://dx.doi.org/10.1016/j.ejpb.2015.01.029.
Ruether, F., Sadowski, G., 2009. Modeling the solubility of pharmaceuticals in pure
solvents and solvent mixtures for drug process design. J. Pharm. Sci. 98,
42054215.
Ruether, F., Sadowski, G., 2011. Thermodynamic modeling of solubility. Chem. Ing.
Tech. 83, 496502.
Stoychev, I., Galy, J., Fournel, B., Lacroix-Desmazes, P., Kleiner, M., Sadowski, G., 2009.
Modeling the phase behavior of PEO-PPO-PEO surfactants in carbon dioxide
using the PC-SAFT equation of state: application to dry decontamination of solid
substrates. J. Chem. Eng. Data 54, 15511559.
Tros de Llarduya, M., Martin, C., Goni, M., Martinez-Oharriz, M., 1998. Solubilization
and interaction of sulindac with polyvinylpyrrolidone K30 in the solid state and
in aqueous solution. Drug Dev. Ind. Pharm. 24, 295300.
Ubbelohde, L., 1937. The principle of the suspended level: applications to the
measurement of viscosity and other properties of liquids. Ind. Eng. Chem. Anal.
Ed. 9, 8590.
Ward, C.A., Findlay, R.D., Rizk, M., 1982. Statistical rate theory of interfacial
transport: 1. Theoretical development. J. Chem. Phys. 76, 55995605.
Windbergs, M., Strachan, C.J., Kleinebudde, P., 2009. Tailor-made dissolution proles
by extruded matrices based on lipid polyethylene glycol mixtures. J. Control.
Release 137, 211216.
Wolbach, J.P., Sandler, S.I., 1998. Using molecular orbital calculations to describe
the phase behavior of cross-associating mixtures. Ind. Eng. Chem. Res. 37,
29172928.