international

journal of
ELSEVIER International Journal of Pharmaceutics 125 (1995) 289-296
pharmaceutics

The morphology of blends of ethylcellulose with hydroxypropyl

methylcellulose as used in film coating
P. S a k e l l a r i o u a,l,., R . C. R o w e b
a ICI Paints, Wexham Road, Slough SL5 2DS, UK
b Zeneca Pharmaceuticals, Pharmaceutical Department, Macclesfield SKIO 2TG, UK

Received 10 February 1995; accepted 3 April 1995

Abstract

Near-equilibrium morphologies of thin films (< 0.2 ~m) of ethylcellulose-rich incompatible blends of ethylcellu-
lose (EC) with hydroxypropyl methylcellulose (HPMC) consist of HPMC-rich macroscopic polydisperse domains
(3-approx. 20/~m) dispersed in an EC-rich matrix implying complete leaching of the water soluble phase. However,
for thick films (> 100/~m) complete leaching was observed only for systems with above 60% w/w HPMC, whereas
all other systems showed some HPMC retention. The observed HPMC leaching behaviour is attributed to deviations
of the phase decomposition from equilibrium during film formation arising from the rate of viscosity increase and
changes in the solvent system composition.

Keywords: Hydroxypropyl methylcellulose; Ethyl cellulose; Blend; Interaction; Leaching; Drug release; Morphology;
Tablet coating; Film formation

1. I n t r o d u c t i o n ypropyl methylcellulose as the minor component.

The drug is released through pores created in the
Blends of ethylcellulose with hydroxypropyl film coating by the dissolution of the dispersed
methylcellulose are widely used in the formula- phase. The interactions and morphology in the
tion of delayed or sustained drug release film blend are of prime importance to the drug re-
coatings (Shah and Sheth, 1972; Rowe, 1985). lease profile. Ethylcellulose/hydroxypropyl meth-
The water insoluble ethylcellulose is usually cho- ylcellulose blends have been shown to phase sep-
sen as the matrix and the water soluble hydrox- arate (Sakellariou et al., 1986a) with a very lim-
ited degree of mixing between the two polymers.
In an earlier publication (Sakellariou et al., 1988),
data were reported on the retention/leaching of
* Corresponding author.
i Present address: Courtaulds Coatings Ltd, New Technol- water-soluble and pH-dependent-soluble cellu-
ogy Group, Stoneygate Lane, Felling, Gateshead, Tyne and lose derivatives and PEG6000 from blends with
Wear NEI0 0JY, UK. ethylcellulose containing 20% and 80% w / w eth-

0378-5173/95/$09.50 © 1995 Elsevier Science B.V. All rights reserved

SSDI 0378-5173(95)00147-6
290 P. Sakellariou, R.C. Rowe/International Journal of Pharmaceutics 125 (1995) 289-296

ylcellulose. A small amount of hydroxypropyl solution at 1 cm min-1 and kept suspended in the
methylcellulose is retained in blends with high solvent vapours to dry over 2-3 h, ensuring near
(80% w / w ) ethylcellulose content (Sakellariou et equilibrium solution casting of the blend. The
al., 1988). movements of the microscope slide were con-
This paper reports on the morphology of blends trolled by a gear box equipped with a motor and
of ethylcellulose with hydroxypropyl methylcellu- power supply. The thickness of the film was con-
lose with specific reference to the effect of film trolled by the concentration of the polymer solu-
preparation. tion and the speed of slide withdrawal from the
solution. Under the conditions of our experi-
ments, films of 1000-1500.4, thickness were ob-
tained eliminating the need for microsectioning.
2. Materials and methods The coated films were collected on a clean micro-
scope slide after floating in a water trough. Al-
The hydroxypropyl methylcellulose (HPMC, though swelling of the HPMC phases with water
Pharmacoat 606) was supplied by Shin-Etsu would enhance phase contrast (at room tempera-
Chemical Co., Tokyo, Japan, and ethylcellulose ture), they did not give rise to morphological
(EC, Grade N50) from Hercules Powder Co., artefacts. This was confirmed by the fact that
Wilmington, DE, USA. Molecular weights and morphology remained unaltered as the tempera-
degrees of substitution of the two polymers are ture increased above the boiling point of water,
listed in Table 1. Films were prepared by solution and even at temperatures above the glass transi-
casting from 50:50 v / v methanol/methylene tion temperature (Tg) of both polymers.
chloride. The dynamic mechanical spectra of the blends
To overcome difficulties encountered in study- were obtained with a torsional braid analyser
ing the morphology of thick films of ethylcellu- (TBA) described in detail elsewhere (Sakellariou
lose/hydroxypropyl methylcellulose blends ( > et al., 1985). Leached samples were prepared by
100 ~m), a thin film applicator was built. It immersing the TBA specimens in distilled water
consists of a thermostatted oil bath where the for a period of 16 h before being dried at 85°C to
dilute polymer solution (2% w / v ) is kept in a constant weight. The drying procedure was suffi-
Pyrex glass tube. A cleaned microscope slide was cient to remove all free moisture. The weight loss
suspended over the solution and in solvent vapour after leaching was determined by difference. Re-
for 20 min, and subsequently lowered slowly into peated measurements on selected samples con-
the solution where it remained for 20 min. The firmed the reproducibility of weight loss with a
immersed slide was withdrawn from the polymer low coefficient of variation ( < 2%).

Table 1
Degrees of substitution (% w/w), number average molecularweight (Mn), polydispersityindex (Mw/M,), solubilityparameter (3)
and Tg of EC and HPMC
Degrees of substitution (% w/w) M. ( × 10- 3) Mw/M" ~ (MPal/2) a Tg (o C) b
Ethoxy M e t h o x y Hydroxypropoxyl (g/mol)
EC c 48-49.5 - - 53 2.96 20.6 131.5
HPMC J - 28-30 7-12 58 3.02 24.4 153.5

a Sakellariou et al. (1985).

b Sakellariou et al. (1986b).
c Grade N50.
o Pharmacoat 606.
 P. Sakellariou, R.C. Rowe / International Journal of Pharmaceutics 125 (1995) 289-296 291

3. Results and discussion

Optical micrographs of blends with 60 and

80% EC are shown in Figs. 1 and 2. Both systems
exhibited morphologies typical to nucleation and
growth mechanism of phase separation, with
HPMC-rich domains dispersed in an EC-rich ma-
trix. The size of the disperse phase was in the
region from 3 to 5 / z m for the blend with 80%
w / w EC and 3 to approx. 20/zm for the system
with 60% w / w EC. It is interesting to note that
the essential morphological features of the blend
films prepared under ambient conditions re-
mained unaltered on raising the temperature from
ambient to 200°C, well above the Tg of both
polymers (see Table 1). This confirms that the 20oc
morphologies obtained were near-equilibrium
ones and that the observed polymer incompatibil-
ity was not due to the effect of the solvent system
but due to thermodynamic incompatibility of the
two polymers. The phase separation observed is
consistent with the solubility parameters of the
two polymers (see Table 1). The Flory-Huggins
interaction parameter for the blend (X = 1.04)
was significantly higher than the critical value for
phase separation (X = 0.04) (Sakellariou and
Rowe, 1995). The cause of this incompatibility
has been attributed to absence of intermolecular
interactions in view of the negligible contribution
of the change in entropy on mixing to the free
energy of mixing. Earlier work has elucidated the
absence of intermolecular interactions in terms of 131oc
the solubility parameter maps of the two poly-
mers (Rowe, 1986). Only limited overlap of the
solubility parameter maps of EC and HPMC was
reported indicating that no significant dipole-di-
pole/induced dipole and hydrogen bonding in-
teractions take place between EC and HPMC.
The phase separation in blends of EC with
HPMC was also confirmed by the presence of
two transitions in the TBA spectra corresponding
to the glass transition temperatures of the two
polymers (Sakellariou et al., 1986b). For compar-
ative purposes Fig. 3 shows the dynamic mechani-

Fig. 1. Morphology of thin films of 60:40 w/w EC/HPMC 181o C

blends with increasing temperature (bar 30/~m).
292 P. Sakellariou, R.C. Rowe/International Journal of Pharmaceutics 125 (1995)289-296

cal behaviour of these blends in terms of the

logarithmic decrement (log dec) and relative
rigidity (1/p 2, p = period of sinusoidal damping).
Leaching of the water soluble HPMC from the
blend would be expected to result in lowering the
height of the glass transition peak assigned to the
HPMC-rich phase; the peak would disappear
completely if the whole of the HPMC-rich phase
were leached out. Glass transition peaks in the
dynamic mechanical spectra due to minor blend
components at contents as low as 10% w / w can
be detected (MacKnight et al., 1978). The dy-
namic mechanical spectra of E C / H P M C blends
leached in distilled water are shown in Fig. 4. A
single main transition at 131-133°C was recorded
25oc for all blend compositions. The temperature at
which this transition occurred, together with the
fact that its maximum logarithmic decrement de-
creased in line with EC content in the blend,
suggest strongly that it is due to EC-rich phases
in the blend. The height of the transition peak
due to the HPMC-rich phase of the blend was
reduced markedly, indicating a significant extent
of HPMC leaching from the films. In blends with
20, 40 and 60% w / w HPMC it became a shoul-
der superimposed on the peak due to the EC-rich
phase, and in the system with 80% HPMC it
disappeared completely. Comparison of the dy-
namic mechanical spectra of the blends before
and after leaching demonstrates clearly that sig-
nificant fractions of HPMC have been leached
lOlOC away from the film, and for systems containing
more than 60% w / w HPMC leaching was almost
complete.
Trends observed in the dynamic mechanical
spectra were corroborated by weight loss data
(Fig. 5). Blends with EC as the major phase
exhibited some retention of HPMC in the film.
This was not the case for systems with HPMC as
the major phase. It is interesting to note that the
latter systems showed leaching greater than the
theoretical one implying that perhaps some EC
has also been leached away. From the TBA and
gravimetric data it would appear that leaching of

Fig. 2. Morphologyof thin films of 80:20 w/w EC/HPMC

140oc blends with increasingtemperature (bar 30/~m).
 P. Sakellariot~ R.C. Rowe/International Journal o f Pharmaceutics 125 (1995) 289-296 293

HPMC from the film is related not only to the mechanical analysis and morphological studies.
water solubility of HPMC but also to the phase T B A samples were thick films (> 100 /zm) cast
morphology of the blend. under ambient conditions leading to faster sol-
The morphologies of thin E C / H P M C blend vent evaporation (non-equilibrium conditions). In
films would suggest complete leaching of the contrast, the films used in the morphological
HPMC-rich dispersed phase. On the other hand, studies were thin ( < 0.2/zm) and cast under near
the dynamic mechanical data and gravimetric equilibrium conditions in an atmosphere of the
analysis showed that, under the leaching condi- solvent system vapours. Fig. 6a illustrates the
tions of our experiments, only for the systems phase separation process during solvent evapora-
containing more than 60% w / w HPMC was com- tion. Once the system concentration has crossed
plete leaching achieved. HPMC retention was the binodal and exceeded a critical value, nucle-
indicated for all systems with EC as the major ation and growth leads to formation of two phases.
component. The apparent contradiction is be- On further solvent evaporation under equilibrium
lieved to stem from differences in the film prepa- conditions, phase ripening takes place across the
ration processes for the films used for dynamic binodal leading to phases richer in the corre-

2~
e 8o 20 ~,, EC/HI~C
• ~:~0
76
x 20:~

22

O BO 20'dlw EC/HPMC
• 60 t.O 20
o ~.0 bO
x~O 80

~6-[i,
1~-

t2. 2

b
I I I l I l I 1 I t I f
20 ~ 60 8g 100 120 1LO 160 t~ ?O0 ?n $0 "~OT/.C 120 1~.0 lt, O ~a}
TI°C

Fig. 3. Logarithmic decrement (LOG DEC) (a) and relative rigidity (b) for EC/HPMC blends before leaching.
 294 P, Sakellariou, R.C Rowe/International Journal of Pharmaceutics 125 (1995) 289-296

o 80/20 ECIHPH[ 1 I BOI20EUHR

• 60/~0 [ ongma[ 601&0 ,,
~Q/60 original tomposlhon compos¢flOnI cz ~160 "
x20~ j,
20/80 '

16

1/' 1-
IIp2
12 - 1/p2
.06"

10 •05 -

8~ .04-

61 .03-

41 .02-

.(7t--

a 001 b
T I I l I 110 r I
~,0 60 80 100 120 IM) 160 180 200 20 ~0 60 B0 100 120 160 180
T/% T/%

Fig. 4. Logarithmic decrement (a) and relative rigidity (b) for E C / H P M C blends after leaching in distilled water.

100,

90 .

80,

70.

o 60

"~ so

~ 4o

~ 3o

2o

10

0 I I I I I
20 40 60 80 100
% w/w HPMC
Fig. 5. Experimental weight loss against original blend composition for E C / H P M C blends. The line represents the theoretical
weight loss if all HPMC were leached out.
 P. Sakellariou, R.C Rowe / International Journal of Pharmaceutics 125 (1995) 289-296 295

1
sponding major component. This process is com- that as ~0HPMC and ~p2c increase in the corre-
plicated by the rate of solvent removal and sponding phases (increased mixing) the difference
changes in the composition of the solvent system in the surface energies of the two phases also
during film formation. increases leading to higher interfacial tension (cf.
The effect of the rate of solvent removal can Eq. 1). The final film morphology will, therefore,
now be considered. In the case of TBA films, a reflect a frozen-in morphology of phases with
point is reached when the time scale for molecu- higher interfacial tension and, consequently,
lar diffusion necessary for phase rearrangement larger dispersions of the minor phase. The
becomes comparable to or slower than that for HPMC-rich domains of the TBA films will, there-
thermodynamically driven concentration changes fore, exceed in size those prepared under equilib-
in the system. This will lead to deviations from rium conditions. Although direct evidence of the
equilibrium binodal phase decomposition (indi- pore size in fully formulated E C / H M P C blended
cated by the arrows in Fig. 6a) and give rise to film coated drug release systems is not available,
relatively increased mixing in the two phases. Lippold et al. (1980) revealed the presence of
The interfacial tension between the two phases large pores (10-100 p~m) in coated depots of
can be approximated by (Van Krevelen and Carbutamide with EC/PEG1500 blended film.
Hoftyzer, 1976): The effect of changes in the solvent system
composition can now be examined. Enrichment
3",2 = (1) of the solvent system in the less volatile methanol
where 3' denotes the surface energies of phases 1 will have important implications for polymer sol-
and 2 which are given by (Sakellariou, 1993): vency, system viscosity, chain conformation and
diffusion. The blend of equal volumes of methanol
1
3"1 = ~Ic3'EC -F (~HPMC3"HPMC (2) and methylene chloride is a non-solvent/solvent
system for EC and a cosolvent system for HPMC
3"2 = ~°2c3"Ec + ~ 2HPMC3"HPMC (3) (Sakellariou and Rowe, 1991). Solvent enrich-
where ~p is the volume fraction of EC and HPMC ment in methanol during film formation will con-
homopolymers; superscripts 1 and 2 denote the tribute to deviations from the original binodal by
corresponding phases. If it is assumed that altering the solvency of both polymers and phase
~0HPMC > ~0EC (justified in view of the higher po- rearrangement. HPMC will come out of solution
larity of the former polymer), Eq. 2 and 3 show first leading to purer HPMC-rich phases (Fig.

Solvent System
Solvent System

EC HPMC EC HPMC
Fig. 6. (a) Schematic phase diagram of EC/HPMC/solvent system (50:50 v/v methanol/methylene chloride). (b) Schematic phase
diagram of EC/HPMC/solvent system (methanol/methylene chloride) enriched in methanol (see text for explanation).
296 P. Sakellariou, R. C. Rowe~International Journal of Pharmaceutics 125 (1995) 289-296

6b). A similar process will take place with EC-rich man, S. (Eds), Polymer Blends, Vol. 1, Academic Press,
phases at later stages as the solvent system is London, 1978, pp. 185-242.
Rowe, R.C., Film coating -the ideal process for the produc-
enriched further in methanol. tion of modified release oral dosage forms. Pharm. Int., 6
Both processes have significant ramifications (1985) 14-17.
for drug release from spheroids or tablets coated Rowe, R.C., The prediction of compatibility/incompatibility
in coating drums at 60°C. The non-equilibrium in blends of ethylcellulose with hydroxypropyl methylcellu-
morphologies resulting from the film coating pro- lose or hydroxypropyl cellulose using 2-dimensional solu-
bility parameter maps. J. Pharm. Pharmacol., 38 (1986)
cedures will be characterised by larger HPMC- 214-215.
rich domains leading to larger than expected holes Sakellariou, P., Effect of polymer compatibility on the surface
and relatively reduced polymer demixing reduc- enrichment in polymer blends. Polymer, 34 (1993) 3408-
ing the leaching of HPMC form the film coating. 3415.
At the same time changes in the solvent system Sakellariou, P. and Rowe, R.C., Phase separation and mor-
phology in ethylcellulose/cellulose acetate phthalate
composition would tend to produce purer phases. blends. J. Appl. Polym. Sci., 43 (1991) 845-855.
The final film morphology will depend on the Sakellariou, P. and Rowe, R.C, Interactions in cellulose
relative rate for the two processes. The kinetically derivative films for oral drug delivery. Prog. Polym. ScL,
controlled phase morphology of the final film (1995) in press.
coating will inadvertently lead to complications in Sakellariou, P., Rowe, R.C. and White, E.F.T., The thermo-
mechanical properties and glass transition temperatures of
the reproducibility of the drug release profiles some cellulose derivatives used in film coating. Int. J.
from solvent cast systems. The final product per- Pharm., 27 (1985) 267-277.
formance will depend strongly on film thickness, Sakellariou, P., Rowe, R.C. and White, E.F.T., Polymer/poly-
rate of wet film application, solvent system, tem- mer interactions in blends of ethylcelluiose with cellulose
perature profile and air flow conditions of the derivatives and polyethylene glycol 6000. Int. J. Pharm., 34
(1986a) 93-103.
coating process. Sakellariou, P., Rowe, R.C. and White, E.F.T., The solubility
parameters of some cellulose derivatives and polyethylene
glycols used in tablet film coating. Int. J. Pharrn., 31
(1986b) 175-177.
Sakellariou, P., Rowe, R.C. and White, E.F.T., A study of the
References leaching/retention of water soluble polymers in blends
with ethyicellulose using torsional braid analysis. J. Con-
trolled Release, 7 (1988) 147-157.
Lippold, C., Lippold, B.H. and Sgoll, G.B., Steuerung der Shah, N.B. and Sheth, B.B., A method for study of timed-re-
Arnzeistoff-reisetzung aus Mikrokapsein. Pharm. Ind., 42 lease films. J. Pharm. Sci., 61 (1972) 412-416.
(1980) 745-751. Van Krevelen, D.W. and Hoftyzer, P.J., Properties of Poly-
MacKnight, W.J., Karasz F.E. and Fried, J.R., Solid state mers: Their Estimation and Correlation with Chemical
transition behaviour of blends. In Paul, D.R. and New- Structure, 2nd Edn, Elsevier, Amsterdam, 1976.