You are on page 1of 9

b i o s y s t e m s e n g i n e e r i n g 1 2 9 ( 2 0 1 5 ) 7 8 e8 6

Available online at


journal homepage:

Research Paper

Parametric evaluation using mechanistic model for

release rate of phosphate ions from chitosan-
coated phosphorus fertiliser pellets

N.N.R. Ahmad a, W.J.N. Fernando b, M.H. Uzir a,*

School of Chemical Engineering, Universiti Sains Malaysia, Nibong Tebal, 14300 Pulau Pinang, Malaysia
Faculty of Engineering, University of Peradeniya, Peradeniya, Sri Lanka

article info
Evaluation of nutrient release from a controlled-release fertiliser (CRF) using mechanistic
Article history: model is important in order to increase the understanding towards the mechanism that
Received 29 January 2013 might involve in a release process. In this study, a mechanistic model was used to evaluate
Received in revised form the parameters that govern the phosphate release rate through chitosan-coated phos-
18 September 2014 phorus fertiliser. The model considers the phenomena of boundary layer formation on the
Accepted 24 September 2014 external surface of the coat and the coating thickness changes due to erosion. Static release
Published online experiments were conducted in order to study the effect of the number of coatings and pH
conditions to the release rate of phosphate ions. The respective parameters of the phos-
Keywords: phate release such as diffusion coefficient were evaluated using the model in both cases.
Diffusion The diffusivity of the phosphate was found to increase as the number of coatings and pH
Modelling values of the medium decreased. Regeneration of the released data for the phosphate ions
Coated fertiliser using the model indicated a very satisfactory fit with the experimental data.
Phosphate © 2014 IAgrE. Published by Elsevier Ltd. All rights reserved.
Release rate

coating material. Various coating materials have been used to

1. Introduction coat different types of fertilisers, but the current trend shows
that polymer coated fertilisers are the most popular types of
The application of controlled-release fertilisers (CRF) in agri- CRF and having great advantages over conventional types of
cultural sector is one of the methods that can be used to coatings (Du, Tang, Zhou, Wang, & Shaviv, 2008). However, the
enhance the nutrient consumption efficiency as well as min- use of conventional polymer coating usually deals with the
imising the environmental problems due to leaching of un- high cost of material and accumulation of undesired synthetic
controlled nutrient released (Al-Zahrani, 2000). Among these residues in soil (Golman, 2011; Trenkel, 2010). Recently, a
CRF types, coated fertilisers are more commonly used in number of studies have been focused on the use of biode-
practice (Shaviv, 2000). The nutrients release rate is controlled gradable polymers as the promising coating materials for
by encapsulating the fertiliser granules or pellets into a type of fertilisers (Chen, Xie, Zhuang, Chen, & Jing, 2008; Han, Chen, &

* Corresponding author. Tel.: þ60 4599 6464; fax: þ60 4594 1013.
E-mail address: (M.H. Uzir).
1537-5110/© 2014 IAgrE. Published by Elsevier Ltd. All rights reserved.
b i o s y s t e m s e n g i n e e r i n g 1 2 9 ( 2 0 1 5 ) 7 8 e8 6 79

Nomenclature Cm nutrient concentration within the membrane,

g m3
R rate of phosphate release, g min1
x distance from the membrane surface, m
Z thickness of the pellet, m
Co original concentration of nutrient within the
A area of the nutrient mass transfer, m2
pellet, g m3
Cav average nutrient concentration released at time t,
D diffusion coefficient of nutrient within the
g m3
membrane or coating layer, m2 min1
kc mass transfer coefficient for external surface of
Df diffusion coefficient of nutrient within the bulk,
the coat, m min1
m2 min1
C2 nutrient concentration at the outer surface of the
l coating thickness at time t, m
chitosan coat at a time t, g m3
r K(D/Df)0.5
Ca nutrient concentration outside the pellet at a
K solute partition coefficient
distance ∞ from the inner boundary of the
lo initial thickness of the chitosan coat, m
chitosan coat, g m3
m dissolution rate constant of the coating film,
C1 nutrient concentration within the pellet, g m3
m min1
C2(x,t) concentration of nutrient outside the pellet at a
t time, min
distance x from the inner boundary of the coat at a
time t, g m3

Hu, 2009; Ni, Liu, Lu, Xie, & Wang, 2011; Ni et al., 2010; Wu, reported on the controlled-release systems, however, most of
2008). them have not taken into account the effect of boundary layer
Chitosan is one of the biodegradable polymers, which is formation occurred on the surface of the external coat (Du,
available in abundant in nature. This natural polysaccharide Zhou, Shaviv, & Wang, 2004; Ko, Cho, & Rhee, 1996; Koizumi,
has been widely used in biomedical, pharmaceutical and Ritthidej, & Phaechamud, 2001; Langer & Peppas, 1983;
agricultural fields due to its biocompatibility, non-toxicity and Shaviv, Raban, & Zaidel, 2003). In fact, in the real application
biodegradability (Aranaz et al., 2009; Sionkowska, 2011). The of controlled-release of nutrients into a static release medium,
application of chitosan can be found in the form of gels, the presence of such a boundary layer could impart additional
membranes, coatings and beads (Krajewska, 2005). However, resistance for the solute fluxes and can affect the transport
chitosan-based materials are hydrophilic in nature and make parameters across the membrane layer (Dworecki, Slezak,
it less stable to be used in its pure form. Cross-linking reaction Ornal-Wasik, & Wasik, 2005).
is one of the modification methods, which can be used to The purpose of this study is to investigate the phosphate
improve its stability, including the water vapour barrier ions release rate from the chitosan-coated phosphorus fertil-
properties (Rinaudo, 2006; Tsai & Wang, 2008). Several studies isers based on the variation of pH conditions and number of
have been reported on the chitosan-coated fertilisers, which coatings. The parameters of release kinetics were evaluated
mainly focused on encapsulating urea and nitro- using a carefully derived mechanistic model. The model was
genephosphorusepotassium (NPK) compound fertilisers developed by considering the formation of boundary layer on
(Chen, Lu, Zhang, Wu, & Deng, 2005; Teixeira et al., 1990; Wu & the external surface of the coat. It also considered the effect of
Liu, 2008; Xia, Zhang, Zhao, & Wang, 2009). Nevertheless, none coating thickness to the release rate. Coated fertilisers could
of them reported on the applicability of coating phosphorus be custom-designed if the factors affecting the release of nu-
fertiliser using the chitosan film. trients could be parameterized so that the release rate could
The application of mathematical model in the controlled- be estimated, prior to manufacturing of such coatings.
release fertiliser systems is very important in order to opti-
mize the design of CRF and later to produce fertilisers that can
meet the nutrients requirement by plants (Shaviv, 2000;
Zheng, Liang, Ye, & He, 2009). Mechanistic model is a type of 2. Materials and method
model derived based on a real physical phenomenon, which
takes place during a release process (Siepmann & Siepmann, 2.1. Materials
2008). Unfortunately, the evaluation of nutrient release ki-
netics from the chitosan-coated fertilisers using mechanistic Calcium dihydrogen phosphate was purchased from Acros
model equations is yet to be extensively explored. Some of the Organics (Geed, Belgium) and of analytical grade. Calcium
previous studies only assessed the parameters of nutrient oxide powder was supplied by Hamburg Chemical GmbH
release using the empirical model, which does not provide a (Germany). Chitosan (degree of deacetylation > 85%, molar
better insight on the actual mechanisms (Wu & Liu, 2008). For mass is 100e1000 kDa) was purchased from Chito-Chem,
the case of chitosan membrane system, capillary pore model Malaysia. Sodium carbonate, glacial acetic acid, and citric
and free volume model have been used to investigate the so- acid were purchased from Fisher Scientific (Malaysia).
lute diffusional properties across the membrane (Krajewska, Aqueous solutions of different pH values were prepared using
2001a, 2001b). There have been several mechanistic models citric acid and sodium carbonate.
80 b i o s y s t e m s e n g i n e e r i n g 1 2 9 ( 2 0 1 5 ) 7 8 e8 6

2.2. Preparation and characterisation of chitosan-coated coating. The coated pellets were then left to dry under the
phosphorus fertiliser pellets continuous flow of air. After 15 min, it was observed that the
coat was completely dried without any visible surface mois-
2.2.1. Preparation of the coated fertiliser pellets ture on the surface of the pellet. Double and triple-coated
The powder of calcium dihydrogen phosphate (as a source of layers were obtained by dipping the pellets in the coating so-
phosphate ions) and calcium oxide (as filler) were mixed ac- lution for two and three times respectively with a gap of
cording to the weight ratio of 14.31:8.00 in order to prepare the 15 min of air drying. The second and third coatings were
powdered fertiliser mixture. Preparation of the pellet required carried out without disturbing the previously formed layer of
about 1.9 g of the mixture in order to fill up a stainless steel coat. The dried pellets obtained were kept in a closed
cylindrical mould with a diameter of 2 cm. The mixture in the container in order to avoid moisture and water contamina-
mould was compressed with a pressure of 500 psi using a tions in between the dipping, such that there would be no
hand-held hydraulic press. The formed pellets with the modification of the original coatings. The coated pellets were
thickness of approximately 4 mm were left to dry in an oven left to dry in the oven at 40  C for 12 h.
(Carbolite, Parsons Lane) at 100  C for 3 h.
In order to prepare the coating solution, about 4% w/v of 2.2.2. Measurement of the initial coating thickness
chitosan powder was dissolved in 50 mL of 4% acetic acid Before undergoing the release rate experiment, the initial
solution. The solution was continuously stirred for 30 min at thickness of the single, double and triple-coated films were
room temperature. Citric acid (0.80% /w chitosan solution) measured using a micrometer screw gauge (Mitutoyo, Japan).
was then added as an agent to cross-link the chitosan solu- The measurements were conducted on the dry coating films
tion. The mixture was left for 2 h with constant stirring. In the after they were peeled off from the upper surface of the pel-
case of 4% acetic acid solution (pH 2.4), the added citric acid lets. For each coating, the measurements were taken 10 times
would be slightly dissociated, however preliminary estima- at different locations and the average thickness was calcu-
tion based on dissociation constant indicated that citric acid lated. The same procedure was carried out for double and
dissociation is only minimal. The purpose of chitosan cross- triple-coated films.
linking is to improve its water barrier properties and in-
crease its stability for the application in the acidic medium. 2.2.3. Fourier Transform Infrared (FTIR) spectra analysis
The mechanism of cross-linking action is illustrated in Fig. 1. The non-cross-linked and cross-linked chitosan coating films
The fertiliser pellets were coated by dipping them into the were characterised using an FTIR spectrophotometer
cross-linked chitosan solution in order to form a single layer (IRPrestige-21, Shimadzu). The samples were pelletised with
KBr powder and analysed under the scanning range from 400
to 4000 cm1.

2.2.4. Coating film morphology

The changes of surface morphology for the double-coated film
before and after the nutrient release rate tests under different
pH values were observed using a scanning electron micro-
scopy (SEM) (Zeiss, BP 35). Platinum was used to coat the films
using an ion-sputtering device before carrying out any

2.3. Release rate study

The phosphate release rate through the chitosan-coated

phosphorus fertiliser pellets were studied based on the vari-
ation of number of coatings and pH conditions of aqueous
solution. All the coated pellets (single, double and triple
layers) as well as the one without coating were immersed in
different cylinders containing 70 ml of pH 5.7 aqueous solu-
tions and left for 10 days. Measurement of phosphate release
from the uncoated pellet was used as a control experiment.
The samples were taken for day 1, 3, 5, 8 and 10. For each
sampling, about 20 ml of sample was withdrawn and added
with another 20 mL of fresh aqueous solution with pH 5.7. In
order to study the effect of pH conditions, the same procedure
was repeated for double-coated pellet with the pH values of
5.7, 7.1 and 8.5. Each sample was analysed for phosphate ions
concentration using a phosphate low range meter (Hanna
Fig. 1 e Formation of amide linkage via cross-linking Instruments, HI 93713) with appropriate dilutions. The phos-
reaction between the amino groups of chitosan and phate concentrations in the sample were measured based on
carboxylic groups of citric acid. the application of ascorbic acid method.
b i o s y s t e m s e n g i n e e r i n g 1 2 9 ( 2 0 1 5 ) 7 8 e8 6 81

2.4. Theoretical modelling of the phosphate release nutrient from the coat's boundary is dependent on the thick-
ness of the coat.
The mathematical model was developed based on the The model uses the following initial and boundary condi-
following assumptions; (1) the nutrient released process took tions, first proposed by Hoogervorst et al. (1978);
place on the upper surface of the pellet and in a one-
ðaÞ C1 ðx; t ¼ 0Þ ¼ Co for  ∞ < x < 0
dimensional coordinate system, (2) the nutrient released
ðbÞ Cðx; t ¼ 0Þ ¼ Cm ðxÞ ¼ KC0 f  x=l þ 1Þ for 0 < x < l (5)
was diffusion dependent, (3) the thickness of the pellet un- ðcÞ C2 ðx; t ¼ 0Þ ¼ 0 for l < x < ∞
dergoing nutrient release was fixed at a constant value and
time independent during the release process and (4) the where C1 (x, t ¼ 0) is the initial solute concentrations within
thickness of the coating layer was uniform. The rate of the pellet, Cm is the solute concentration within the mem-
nutrient release, R from the fertiliser pellet at a given time, t brane and C2 (x, t ¼ 0) is the initial solute concentration in the
could be expressed as; bulk fluid and x is referred to the transport region. It was
assumed that for conditions (a) and (c), the thickness of the
dCav pellet (Z) is significantly thicker than the thickness of the
R ¼ ZA (1)
dt chitosan coat (l). Thus, the ratio of both thicknesses could be
where Z is the thickness of the pellet, A is the area of the taken as Z/l / ∞. Meanwhile, condition (a) assumes that the
nutrient mass transfer, and Cav is the average nutrient con- nutrient distribution is uniform across the pellet at the
centration released at time, t. The nutrient was assumed to beginning of the nutrient release process.
diffuse from the core of the pellet to the outer surface through For the condition near the wall of the membrane where
the coating layer. The expression for the nutrient transfer rate (x ¼ l), the concentration C2 (l, t) in Eq. (4) could therefore be
from the upper surface of the pellet to the surrounding me- expressed as Eq. (7) using the correlation given by Eq. (6).
dium could be described using Eq. (2);
ierfc y ¼ pffiffiffiffi$exp  y2  y$erfc y (6)
R ¼ kc AðC2  Ca Þ (2) p

where kc is the mass transfer coefficient for external surface of " pffiffiffiffiffiffi pffiffiffiffiffiffi ∞  n  #
2r Dt 1 4r2 Dt X 1r nl
the coat, C2 is the nutrient concentration at the outer surface C2 ðl; tÞ ¼ Co $pffiffiffiffi þ 2 $ $ierfc pffiffiffiffiffiffi
ðr þ 1Þl p ðr  1Þl n¼1 1 þ r 2 Dt
of the chitosan coat at time, t and Ca is the nutrient concen-
tration outside the pellet at a distance ∞ from the inner (7)
boundary of the chitosan coat. Combination of Eqs. (1) and (2) It can be proven that for all values of r > 0, the term
resulted in Eq. (3); pffiffiffiffiffiffi
ð1  r=1 þ rÞn $ierfcðnl=2 DtÞ converges rapidly with n. Thus,
Eq. (7) could be simplified into:
Z ¼ kc ðC2  Ca Þ (3) "
dt pffiffiffiffiffiffi pffiffiffiffiffiffi    #
2r Dt 1 4r2 Dt 1  r l
During the nutrient release process, the chitosan coat was C2 ðl; tÞ ¼ Co $pffiffiffiffi þ $ierfc pffiffiffiffiffiffi (8)
ðr þ 1Þl p ðr2  1Þl 1 þ r 2 Dt
assumed to behave as a membrane. As the time of release
For a distance far away from the surface of the pellet, it was
getting longer, a boundary layer slowly formed on the surface
assumed that the concentration of phosphate ions is negli-
of the coated pellet. The formation of the layer has a signifi-
gible (i.e. Ca / 0 when x / ∞). Then, substitution of Eq. (8) into
cance role in governing the nutrients transport rate through
Eq. (3) resulted in the following equation;
the coat (Dworecki et al., 2005). The presence of additional
" pffiffiffiffiffiffi pffiffiffiffiffiffi    #
resistance due to this boundary layer causes the release pro- dCav 2r Dt 1 4r2 Dt 1  r l
cess to be a rate-limiting diffusion (Pedley & Fischbarg, 1980; Z ¼ kc Co $pffiffiffiffi þ 2 $ierfc pffiffiffiffiffiffi (9)
dt ðr þ 1Þl p ðr  1Þl 1 þ r 2 Dt
Pohl, Saparov, & Antonenko, 1998). According to
Re-arrangement of Eq. (9) into a more convenient form
Hoogervorst, de Goede, Versluijs, and Smit (1978), the math-
leads into;
ematical descriptions of the solute concentration profile in the
aqueous medium with significant effect of the liquid bound- pffiffiffiffiffiffi "  #
dCav Dt 2r 1 4r2 l
ary layer formation can be expressed using Eq. (4) below; Z ¼ kc Co $pffiffiffiffi  $ierfc pffiffiffiffiffiffi (10)
dt l ðr þ 1Þ p ð1 þ rÞ2 2 Dt
" pffiffiffiffiffiffi ! pffiffiffiffiffiffi ∞  n
2r Dt x1 4r2 Dt X 1r Eq. (10) is only valid for the nutrient release process for the
C2 ðx;tÞ ¼ Co $ierfc pffiffiffiffiffiffiffi þ 2 $
ðr þ 1Þl 2 Df t ðr  1Þl n¼1 1 þ r coated pellets with a constant coating thickness. However, for
!# (4)
nl xl the nutrient release in an acidic medium, the chitosan coating
$ierfc pffiffiffiffiffiffi þ pffiffiffiffiffiffiffi
2 Dt 2 Df t film could be exposed to erosion and later dissolved in the
aqueous solution (El-Sherbiny, Abdel-Bary, & Harding, 2006;
where C2 (x,t) is the concentration of nutrient outside the Wu, Su, & Ma, 2006). Thus, the coat thickness, l could also be
pellet at a distance x from the inner boundary of the coat at assumed to decrease with time as expressed by the following
time, t, Co is the original concentration of nutrient within the equation;
pellet, D is the diffusion coefficient of nutrient within the
l ¼ lo expð  mtÞ (11)
membrane or coating layer, Df is the diffusion coefficient of
nutrient within the bulk, l is the coating thickness at time, t, r where lo is the initial thickness of the chitosan coat and m is
is represented by K(D/Df)0.5 and K is the solute partition coef- the dissolution rate constant of the coating layer. Incorpora-
ficient. It can be seen from Eq. (4) that the concentration of tion of Eq. (11) into Eq. (10) resulted into Eq. (12);
82 b i o s y s t e m s e n g i n e e r i n g 1 2 9 ( 2 0 1 5 ) 7 8 e8 6

pffiffiffiffiffiffi "
dCav Dt 2r 1 4r2
Z ¼ kc Co $pffiffiffiffi 
dt ðlo expð  mtÞÞ ðr þ 1Þ p ð1 þ rÞ2
 # (12)
ðlo expð  mtÞÞ
$ierfc pffiffiffiffiffiffi
2 Dt

The diffusion model provided by Eq. (12) can be used to

describe the nutrient released process from the coated pellets,
undergoing film thickness changes with respect to time.

3. Results and discussion

3.1. Initial thickness of the coating film

Table 1 shows the results of the measurement of the initial

thickness of the coating film. Obviously, the average thickness
of the coating film increases with an increase of the number of
coatings. The thicknesses of the double and triple-coated Fig. 2 e FTIR spectra for (a) non-cross-linked and (b) cross-
films are approximately 1.5 and 2 times higher compared to linked chitosan with citric acid solution.
the single layer of coating film respectively. However, the
increment of the thickness is not directly proportional to the
number of coatings. During the addition of the subsequent
3.3. Morphology of the coating film
coating, the first layer of polymer coat could be slightly dis-
solved into the second layer, which possibly explains such a
The SEM images of double coating films before and after 10
condition. Such an observation is also supported by similar
days of release tests were closely compared. It can be seen
findings reported by Tomaszewska and Jarosiewicz (2002).
from Fig. 3(A) that the initial coating film has a compact sur-
face without any pores and cracks. The figure also shows that
the chitosan film has a dense structure, which could be due to
3.2. FTIR analysis the cross-linking reaction between the citric acid and chito-
san, as revealed by the FTIR analysis in Section 3.2. After 10
The FTIR spectra of non-cross-linked and cross-linked chito- days of immersion in the aqueous solution of different pH
san samples are shown in Fig. 2. For the non-cross-linked values, the coating surface structure has changed accordingly.
chitosan sample, the presence of peak at the frequency Coating film immersed in pH 5.7 (Fig. 3(B)) showed a spongy
around 3444.87 corresponds to the presence of eOH stretch in structure with pores clearly present on its surface. The hy-
chitosan. A peak observed at 1643.35 cm1 represents the drolysis of amide bond in acidic condition is one of the
amide (I) band in chitosan (Salam, Venditti, Pawlak, & El- possible reactions, which can take place on the chitosan layer
Tahlawy, 2011; Yin, Luo, Chen, & Khutoryanskiy, 2006). The (Mujika, Formoso, Mercero, & Lopez, 2006) and therefore, leads
CeO stretch vibration in chitosan was indicated by the peak at to the erosion of chitosan as observed in Fig. 3(B). The pore
1068.56 cm1. It was also observed that the cross-linked chi- structures were not observed from the images of chitosan
tosan sample exhibited the analogous FTIR spectrum films after the release tests at pH 7.1 and 8.5 (Fig. 3C and D).
compared to the non-cross-linked chitosan. However, the This explains that the chitosan films are more stable under
cross-linked chitosan spectrum indicates a characteristic the neutral and basic conditions with no erosion detected,
peak which was found at the frequency of 1712.79 cm1. This which observation was also reported by Wu et al. (2006).
peak could be attributed to the stretching of carbonyl groups
in the secondary amide linkage of cross-linked chitosan,
3.4. Phosphates release rate
whose appearance has been previously observed at about
1716 cm1 (Salam et al., 2011). The presence of this peak cor-
The trend of phosphate ions released under different number
responds to the cross-linking reaction that might occur be-
of coatings could be observed from Fig. 4. It can be apparently
tween the amine groups from the chitosan and carboxylic
seen that the triple-coated pellets show the lowest phosphate
groups from the acid (citric acid) (Fig. 1).
ions concentration in the aqueous solution, followed by the
double and single coated pellets. An increase in the number of
coatings corresponds to the increase in the thickness as
Table 1 e Initial thickness of different coating films, lo.
shown previously in Table 1. Thus, the thicker the coating
Coating film Average thickness (mm)
layer will result in a lower phosphate release rate. This is due
Single 63 to the fact that a longer resistance path will be formed that
Double 96 hinders the nutrient to diffuse out from the centre of the
Triple 128
b i o s y s t e m s e n g i n e e r i n g 1 2 9 ( 2 0 1 5 ) 7 8 e8 6 83

Fig. 3 e SEM images of the double coating film before (A) and after 10 days of release test under pH 5.7 (B), 7.1 (C) and 8.5 (D).

The effect of pH conditions to the phosphate release 3.5. Transport parameters of phosphate release and
through chitosan film is illustrated in Fig. 5. The figure shows model validation
that the phosphate ions release increases as the pH values are
decreased. The highest release of phosphate ions at pH 5.7 The experimental data of phosphate ions concentration ob-
could be attributed to the decrease of coating thickness due to tained at different number of coatings and pH conditions were
the erosion of chitosan under the acidic condition. This is used to calculate the rate of released of the phosphate ions at
supported by the formation of pores on the surface of coating different period of times. These data were examined for
film after the release test at pH 5.7 as shown in Fig. 3(A). Low applicability of the diffusion model as developed in Section
release rate of phosphate ions observed at pH 7.1 and 8.5 could 2.4. The diffusion model in Eq. (12) was used to evaluate the
be due to the increase in the hydrophobicity of chitosan under transport parameters in an acidic condition by taking into
a weakly basic medium (El-Sherbiny et al., 2006). account the decrease of coating thickness due to erosion.
Meanwhile, the experimental data for neutral and basic con-
ditions was evaluated using Eq. (10). The parameters of kcCo, r,

Fig. 4 e The profile of phosphate release under pH 5.7 for

uncoated and coated pellets with different number of Fig. 5 e The profile of phosphate release for double-coated
coatings. pellets under different pH values.
84 b i o s y s t e m s e n g i n e e r i n g 1 2 9 ( 2 0 1 5 ) 7 8 e8 6

Table 2 e Values of parameters evaluated for phosphate ions release at pH 5.7 for pellets with different number of coatings.
Number of coatings Parameters
kcCo (g min1 m2) r m (m min1) D (m2 min1)
Single 3.85  106 0.8286 3.1  105 2.75  108
Double 3.83  106 0.7561 3.0  105 2.53  108
Triple 3.81  106 0.6906 3.0  105 2.11  108

conditions, the degree of swelling for the chitosan film is low

Table 3 e Values of parameters evaluated for phosphate
and causes the mobility of solute across the coating film be-
ions release from double-coated pellets at different pH
conditions. comes slower, which in turn reduce the solute diffusivity (El-
Sherbiny et al., 2006; Miller-Chou & Koenig, 2003). The
pH values Parameters
decrease of the film thickness with respect to time due to
kcCo (g min1 m2) r D (m2 min1) erosion at pH 5.7 also reduces the resistance for the phosphate
5.7 3.83  10 0.7561 2.53  108 diffusion, which explains why the diffusivity value, D was
7.1 2.81  106 0.5955 2.00  108 high at this particular condition.
8.5 2.40  106 0.5911 1.82  108 Figures 6 and 7 provide a plot of the predicted release rate
against the experimental values of the phosphate release rate
m and D were obtained from the parameter fitting procedure based on the model developed for different number of coat-
embedded in Matlab® (The MathWorks, Inc). ings and pH conditions respectively. Both plots seem to pro-
The results of transport parameters for the case of different vide approximate values to that of the experimental data. The
number of coatings and pH conditions are shown in Tables 2 difference between the predicted and the experimental
and 3 respectively. For the effect of different number of release rate seems to be statistically acceptable, which proves
coatings, the diffusion coefficient, D evaluated from the model that the model is able to predict the release rate of phosphate
was found to be the highest for a single coated pellet. This is in ions from the chitosan-coated pellets. This, in turn confirms
agreement with the high concentration of phosphate released the estimation of the diffusivity values obtained from the
observed for a single coated pellet (Fig. 4), which was experimental data.
described previously in Section 3.4. High value of diffusivity, D
for a single coated pellet could be due to lesser effect of
resistance for the nutrient diffusion across a thinner coating
4. Conclusion
film. Table 2 also shows that there are no appreciable differ-
ences between the values for coating film dissolution rate
From the investigation, it can be summarised that the phos-
constant, m under pH 5.7 for different number of coatings.
phate release rate and the resulting diffusion coefficient of the
This result agrees with the report by Sumner, Krause, Sabot,
phosphate ions from phosphorus-based fertiliser pellets to
Turner, and McNeil (2001) who observed that the degrada-
the surrounding are affected by the number of coatings and
tion rate is rather constant and independent on the polymer
the pH conditions. The diffusion coefficient of phosphate ions
film thickness.
seems to increase as the coating numbers and pH of the me-
The results tabulated in Table 3 apparently show that the
dium gradually decreased. The mechanistic model developed
diffusion coefficient, D decreases as the pH values increased.
in this work has successfully described the release rate of
This could be due to the fact that under neutral and basic
phosphate ions and at the same time evaluated the

Fig. 6 e Comparison between experimental and predicted Fig. 7 e Comparison between experimental and predicted
phosphate release rate for pellets with different number of phosphate release rate for double-coated pellets under
coatings at pH 5.7 condition. variation of pH values.
b i o s y s t e m s e n g i n e e r i n g 1 2 9 ( 2 0 1 5 ) 7 8 e8 6 85

parameters values of kcCo, r, m and D. These parameter values Krajewska, B. (2001a). Diffusion of metal ions through gel
are important and need to be estimated prior to designing the chitosan membranes. Reactive & Functional Polymers, 47, 37e47.
fertiliser pellets. Pellets can be designed in such a way that the Krajewska, B. (2001b). Diffusional properties of chitosan hydrogel
membrane. Journal of Chemical Technology and Biotechnology, 76,
thickness of the coat follows the predetermined or customer-
desired rate and therefore leading to a reduced waste as well Krajewska, B. (2005). Membrane-based processes performed with
as minimizing problems associated with nutrient leaching. use of chitin/chitosan materials. Separation and Purification
Technology, 41, 305e312.
Langer, R., & Peppas, N. (1983). Chemical and physical structure of
polymers as carriers for controlled release of bioactive agents:
Acknowledgements a review. Journal of Macromolecular Science, Part C: Polymer
Review, 23(1), 61e126.
Miller-Chou, B. A., & Koenig, J. L. (2003). A review of polymer
The authors wish to acknowledge the financial support dissolution. Progress in Polymer Science, 28, 1223e1270.
received from the USM Incentive Grant (8022033). USM Mujika, J. I., Formoso, E., Mercero, J. M., & Lopez, X. (2006).
Fellowship scheme that funds N.N.R. Ahmad is also greatly Reaction mechanism of the acidic hydrolysis of highly
acknowledged. twisted amides: rate acceleration caused by the twist of the
amide bond. The Journal of Physical Chemistry B, 110,
Ni, B., Liu, M., Lu, S., Xie, L., & Wang, Y. (2011). Environmentally
references friendly slow-release nitrogen fertilizer. Journal of Agricultural
and Food Chemistry, 59, 10169e10175.
Ni, B., Liu, M., Lu, S., Xie, L., Zhang, X., & Wang, Y. (2010).
Al-Zahrani, S. M. (2000). Utilization of polyethylene and paraffin Novel slow-release multielement compound fertilizer with
waxes as controlled delivery systems for different fertilizers. hydroscopicity and moisture preservation. Industrial and
Industrial Engineering and Chemistry Research, 39, 367e371. Engineering Chemistry Research, 49, 4546e4552.
Aranaz, I., Mengibar, M., Harris, R., Panos, I., Miralles, B., Pedley, T. J., & Fischbarg, J. (1980). Unstirred layer effects on
Acosta, N., et al. (2009). Functional characterization of osmotic water flow across gallbladder epithelium. Journal of
chitosan. Current Chemical Biology, 3, 203e230. Membrane Biology, 54(2), 89e102.
Chen, Q., Lu, W. J., Zhang, W. Q., Wu, J. H., & Deng, X. (2005). Pohl, P., Saparov, S. M., & Antonenko, Y. N. (1998). The size of the
Nutrient release characteristic of chitosan coated urea and unstirred layer as a function of the solute diffusion coefficient.
change of penetrability of its coating layer. Journal of East China Biophysical Journal, 75, 1403e1409.
University of Science and Technology, 31(6), 764e767. Rinaudo, M. (2006). Chitin and chitosan: properties and
Chen, L., Xie, Z., Zhuang, X., Chen, X., & Jing, X. (2008). Controlled applications. Progress in Polymer Science, 31, 603e632.
release of urea encapsulated by starch-g-poly(L-lactide). Salam, A., Venditti, R. A., Pawlak, J. J., & El-Tahlawy, K. (2011).
Carbohydrate Polymers, 72, 342e348. Crosslinked hemicellulose citrateechitosan aerogel foams.
Du, C., Tang, D., Zhou, J., Wang, H., & Shaviv, A. (2008). Prediction of Carbohydrate Polymers, 84, 1221e1229.
nitrate release from polymer coated fertilizers using an artificial Shaviv, A. (2000). Advances in controlled release of fertilizer.
neural network model. Biosystems Engineering, 99, 478e486. Advance in Agronomy, 71, 1e49.
Du, C., Zhou, J., Shaviv, A., & Wang, H. (2004). Mathematical Shaviv, A., Raban, S., & Zaidel, E. (2003). Modeling controlled
model for potassium release from polymer-coated fertilizer. nutrient release from polymer coated fertilizers: diffusion
Biosystems Engineering, 88(3), 395e400. release from single granules. Environmental Science and
Dworecki, K., Slezak, A., Ornal-Wasik, B., & Wasik, S. (2005). Technology, 37, 2251e2256.
Evolution of concentration field in a membrane system. Siepmann, J., & Siepmann, F. (2008). Mathematical modeling of
Journal of Biochemical and Biophysical Methods, 62, 153e162. drug delivery. International Journal of Pharmaceutics, 364,
El-Sherbiny, I. M., Abdel-Bary, E. M., & Harding, D. R. K. (2006). 328e343.
Swelling characteristics and in vitro drug release study with Sionkowska, A. (2011). Current research on the blends of natural
pH- and thermally sensitive hydrogels based on modified and synthetic polymers as new biomaterial: review. Progress in
chitosan. Journal of Applied Polymer Science, 102, 977e985. Polymer Science, 36, 1254e1276.
Golman, B. (2011). Kinetics of release from polydisperse core Sumner, C., Krause, S., Sabot, A., Turner, K., & McNeil, C. J. (2001).
particles coated with layers of fine soluble particles. Biosensor based on enzyme-catalysed degradation of thin
International Journal of Chemical Engineering and Applications, 2(4), polymer films. Biosensors & Bioelectronics, 16, 709e714.
283e287. Teixeira, M. A., Paterson, W. J., Dunn, E. J., Li, Q., Hunter, B. K., &
Han, X., Chen, S., & Hu, X. (2009). Controlled-release fertilizer Goosen, M. F. A. (1990). Assessment of chitosan gels for the
encapsulated by starch/polyvinyl alcohol coating. Desalination, controlled release of agrochemicals. Industrial and Engineering
240, 21e26. Chemistry Research, 29, 1205e1209.
Hoogervorst, C. J. P., de Goede, J., Versluijs, C. W., & Smit, J. A. M. Tomaszewska, M., & Jarosiewicz, A. (2002). Use of polysulfone in
(1978). Nonstationary diffusion through membranes. 2. controlled-release NPK fertilizer formulations. Journal of
Transient diffusion through a membrane separating two Agricultural and Food Chemistry, 50, 4634e4639.
semiinfinite volumes of unstirred solutions. Journal of Physical Trenkel, M. E. (2010). Slow- and controlled-release and stabilized
Chemistry, 82(11), 1318e1324. fertilizers. Paris, France: IFA.
Ko, B., Cho, Y., & Rhee, H. (1996). Controlled release of urea from Tsai, H. S., & Wang, Y. Z. (2008). Properties of hydrophilic chitosan
rosin-coated fertilizer particles. Industrial and Engineering network membranes by introducing binary crosslink agents.
Chemistry Research, 35, 250e257. Polymer Bulletin, 60, 103e113.
Koizumi, T., Ritthidej, G. C., & Phaechamud, T. (2001). Mechanistic Wu, C. (2008). Controlled release evaluation of bacterial fertilizer
modeling of drug release from chitosan coated tablets. Journal using polymer composites as matrix. Journal of Controlled
of Controlled Release, 70, 277e284. Release, 132, 42e48.
86 b i o s y s t e m s e n g i n e e r i n g 1 2 9 ( 2 0 1 5 ) 7 8 e8 6

Wu, L., & Liu, M. (2008). Preparation and properties of chitosan- chitosan coated slow release fertilizers. Acta Ecologica Sinica,
coated NPK compound fertilizer with controlled-release and 29(8), 4560e4564.
water-retention. Carbohydrate Polymers, 72, 240e247. Yin, J., Luo, K., Chen, X., & Khutoryanskiy, V. V. (2006). Miscibility
Wu, J., Su, Z. G., & Ma, G. H. (2006). A thermo- and pH-sensitive studies of the blends of chitosan with some cellulose ethers.
hydrogel composed of quaternized chitosan/ Carbohydrate Polymers, 63, 238e244.
glycerophosphate. International Journal of Pharmaceutics, 315, Zheng, T., Liang, Y., Ye, S., & He, Z. (2009). Superabsorbent
1e11. hydrogels as carriers for the controlled release of urea:
Xia, W., Zhang, W. Q., Zhao, X. F., & Wang, W. Q. (2009). Effects of experiments and a mathematical model describing the release
environmental actors on nutrient release characteristic of rate. Biosystems Engineering, 102, 44e50.