observation of swelling process and diffusion front ... · the behavior of gel layer thickness in...

Journal of Controlled Release 61 (1999) 83–91

Observation of swelling process and diffusion front positionduring swelling in hydroxypropyl methyl cellulose (HPMC)

matrices containing a soluble druga , a b*Paolo Colombo , Ruggero Bettini , Nikolaos A. Peppas

aDepartment of Pharmacy, University of Parma, 43100 Parma, ItalybBiomaterials and Drug Delivery Laboratories, School of Chemical Engineering, Purdue University, West Lafayette, IN 47907-1283,

USA

Received 18 September 1998; accepted 14 April 1999

Abstract

The behavior of gel layer thickness in swellable hydroxypropyl methyl cellulose matrices loaded with increasing amountsof soluble and colored drug and exhibiting swelling, diffusion and erosion fronts, was studied using a colorimetric technique.The effect of the drug loading on the front position in the gel layer, in particular, on the presence of a diffusion front and itsmovement, was investigated. In addition, the swelling, diffusion and erosion front positions at different releasing times weremeasured and a theoretical analysis of the overall process was provided. It was found that the diffusion front was visible insystems with more than 30% drug, due to the presence of an undissolved drug layer. The physical analysis of such systemsclearly showed the importance of drug solubility and loading in the observation of the diffusion front. 1999 ElsevierScience B.V. All rights reserved.

Keywords: Swellable matrices; Drug loading; Diffusion front

1. Introduction studies on the swelling and subsequent dissolution ofhydroxypropyl methyl cellulose (HPMC) tablets or

Swelling of hydrophilic polymeric matrices has matrices. For example, Melia et al. [7,8] and Rajabi-been the subject of significant research in the last Siahboomi et al. [9] have investigated the swellingfew years. In particular, there have been several process using scanning electron microscopy andstudies trying to relate molecular characteristics of NMR spectroscopy. Gao and associates [5,10] pre-the swelling process (macromolecular chain exten- sented studies on HPMC swelling using an opticalsion, solvent accommodation) to macroscopic imaging technique. Konrad et al. [11] measured thecharacteristics [1–6]. eroding and swelling front advancement during drug

Of particular interest are recent experimental release using ultrasound.In recent years, our group has also examined

various aspects of the swelling behavior of such*Corresponding author. Tel.: 139-0521-905086; fax: 139-tablets in terms of front positions [3,12–14]. Indeed,0521-905085.

E-mail address: [email protected] (P. Colombo) during contact between a dry HPMC matrix and a

0168-3659/99/$ – see front matter 1999 Elsevier Science B.V. All rights reserved.PI I : S0168-3659( 99 )00104-2

84 P. Colombo et al. / Journal of Controlled Release 61 (1999) 83 –91

dissolution medium, such as water or a biological presence of a diffusion front and its movement. Wefluid, water penetrates the HPMC matrix and swells measured the swelling, diffusion and erosion frontthe macromolecular chains. Molecularly, individual positions at different swelling times and provided achains, originally found in their unperturbed state theoretical analysis of the overall process.[15], absorb water so that their end-to-end distanceand radius of gyration expand to the new solvatedstate. This expansion (swelling) is observed macro- 2. Analysis of swelling behaviorscopically by the formation of distinct fronts thatseparate unswollen and swollen regions. As dis- 2.1. Swelling processcussed previously [3], during macroscopic observa-tion of the swelling process, we have identified a We considered the dynamic swelling behavior of a‘swelling front’ that clearly separates the rubbery hydrophilic, glassy, polymeric matrix that can swellregion (region of swollen HPMC with enough water (and dissolve) in water. Starting with a slab thicknessto have its T below the experimental temperature)g of 2d at time t50; when the slab is placed in water,0and the glassy region (region where HPMC has a Tg there is swelling and dissolution, and an erosion andabove the experimental temperature). A second front a swelling front appear (E and S, respectively) (Fig.is the ‘erosion front’, which separates the matrix 1).from the solvent (water).

The gel layer formed on the glassy core of a2.2. Swelling frontswellable matrix is considered to be the controlling

element of drug release kinetics. The gel layerThe water concentration at the glassy–rubberystructure and composition change during matrix

interface, c*, is established by the glass transitionswelling, due to the molecular extension of thetemperature of the polymer, T and the experimentalgsolvated polymeric chains. The gel layer thicknesstemperature, T [18,19]. Thus, the water concen-expbehavior as a function of time is determined by thetration at position S, can be calculated asrelative position of the moving swelling and erosion

fronts. In addition, a drug diffusion front located T 2 Tg exp]]]c* 5 (1)between the swelling and erosion fronts and con- b /afstituting the boundary separating solid from dis-

solved drug was identified previously [16]. The Here, c* is the threshold concentration, expressed indiffusion front position in the gel phase during drug g of water /g of dry polymer, a is the linear thermalf

release was dependent on the drug’s solubility and expansion coefficient of the polymer, and b is theloading. In fact, the diffusion front movement was contribution of the water to the expansion coefficientrelated to the drug’s dissolution rate [3]. Finally, the of the polymer.diffusion and erosion front positions identify the In problems of polymer dissolution, it is advan-drug’s dissolved gel layer, where the concentration tageous to use volume fractions of components, y ,iprofile relevant to the flux of drug is established. rather than concentrations, c . This is because of thei

The movement of the swelling and dissolution volume change at any point in a slab. Thus, in order(erosion) fronts in HPMC matrices containing to convert c* to an equivalent volume fraction, thefluorescein sodium at different drug loadings and value of c* can also be expressed as a volume of

3polymer viscosities were measured using a non- water per volume of polymer as c*r /r (cm ofp w3destructive mode of operating [17]. water /cm of polymer).

In this work, the behavior of gel layer thickness in At the interface, S, in drug-containing swellableswellable matrices loaded with increasing amounts of systems, some undissolved drug also exists and issoluble and colored drug was studied using a equal to the loading concentration of c (g of drug/gd

3 3colorimetric technique. The main goal of this work of polymer), or c r /r (cm of drug/cm of poly-d p d

was to determine how the drug loading affected the mer), where r is the drug density.d

front position in the gel layer, in particular, the Then, at the interface, the equivalent threshold

P. Colombo et al. / Journal of Controlled Release 61 (1999) 83 –91 85

Fig. 1. Schematic representation of the front positions, water, polymer and drug gradient associated with the dynamic swelling process in aslab-like matrix.

* *volume fractions of water, y , polymer, y , and agitation and the composition of the surroundingw p

*drug, y , are as follows (Fig. 1): fluid (pH and ionic strength).d

When dynamic swelling /dissolution is established,rp a volume fraction (concentration) gradient is estab-]c*rw lished in the region between the erosion and swelling

]]]]]*y 5 (2)w r rp p fronts, as shown in Fig. 1. This figure shows the] ]c* 1 c 1 1dr r shape of the concentration profiles.w d

1 2.4. Diffusion front]]]]]*y 5 (3)p r rp p] ]c* 1 c 1 1dr rw d In this section, we consider incorporation of a

3drug whose solubility in water is c (g of drug/cmsandof water). Drug concentration at any point in the gel

rp is obtained by multiplying c by the correspondings]c*r water volume fraction at that point, y . Then, thed w

]]]]]*y 5 (4)d r r drug concentration in the gel (between S and E) isp p3] ]c* 1 c 1 1d c y (g of drug/cm of gel).r r s ww d

To further express this value as a local drugHere, volume fraction, y , we divide by the drug density tods

obtain:* * *y 1 y 1 y 5 1 (5)w p d c ys w 3 3]]y 5 (cm of drug/cm of gel) (6)ds rd2.3. Erosion frontIt must be noted that, between S and E, the terms

At the erosion front, E, the relevant volume c and r are constant (at constant T ) and only ys d exp w1fractions of the components are for pure water, y , can change as a function of position. As c has aw s

1the disentangled macromolecular chains, y , and limiting value, y will also have a limiting value,p,dis ds1 1the drug, y . Although y is known for each i.e., a single local value at which the drug will not bed p,dis

1 1polymer [4], y and y are not known a priori but completely dissolved. This value identifies the posi-w d

are dependent on experimental conditions, such as tion of the diffusion front, D. From the diffusion


front up to the swelling front, the drug concentration to 10 parts HPMC (Methocel K4M, Colorcon,should change, as indicated by drug profile in Fig. Orpington, UK; particle size, 90%,150 mm). As a1c. binder, four parts of cellulose acetate phthalate

Using Eq. (6), it is also possible to calculate the dissolved in acetone–ethanol (1:1,v /v) was used.*value of the drug volume fraction, y , that corre- After drying, the granules were lubricated with twods

*sponds to the threshold water volume fraction y as: parts of magnesium stearate (USP grade) and fourw

parts of talc (USP grade). Finally, the granules were*c ys w tabletted using a reciprocating tabletting machine]]*y 5 (7)ds rd (EKO Korsch, Berlin, Germany) equipped with flatface cylindrical punches of 7 mm diameter, in orderThis parameter is of importance when trying toto prepare matrices weighing 12560.8 mg, having adetermine the dissolved and undissolved drug as itcrushing strength of between 12 and 13 (Monsantowill be shown in our experimental studies below.scale) and an initial porosity of 9.260.9%.The drug loading below which the diffusion front

Swelling studies of these HPMC matrices werewill not appear can be calculated by making theperformed by clamping each matrix between twofollowing assumption. First, we define the value oftransparent Plexiglas disks, thus forming an assem-* *y , as in Eq. (4), and we set it equal to y , as in Eq.d ds

bled device that was introduced into the vessel of a(7), to obtain:USP 23 Apparatus 2 (Esadissolver, Advanced Prod-

rp ucts, Milan, Italy) containing distilled water at 378C]cd *r c yd s w (agitation speed of 200 rpm) [21].]]]]] ]]* *y 5 5 y 5 (8)d dsr r rp p At fixed times during swelling, the device wasd] ]c* 1 c 1 1dr r taken out of the vessel and pictures of the disc matrixw d

base were video-recorded. Using an Image 1.49Then, we solve for the value of c , which becomesd program (NIH, Bethesda, MD, USA), the frontr rp distance was measured in pixels and these wered] ]c* 1 1S DS Dr r converted into length units.w p

]]]]]]c 5 (9)d The densities of drug and polymer were measuredrd]] 2 1S D using a helium pycnometer (Multivolume Pycnome-*c ys w

ter 1305, Micromeritics, Norcross, GA, USA).*In addition, by solving Eq. (8) for y , we can obtain:w

cd]rp 4. Results and discussioncs

]]]]]*y 5 (10)w r rp p] ]c* 1 c 1 1 The swellable matrices were prepared using mix-dr rw d tures of BPP and HPMC of different ratios. Drug and

This equation shows that the threshold fraction of polymer constituted 90 wt.% of the formulation,water at the diffusion front is directly related to the while the rest were adjuvants for tabletting. Matricesratio of c /c , a similar conclusion and relation as were produced by compression under the samed s

that obtained by Lee [20]. conditions.Swelling studies were conducted in a device

consisting of two transparent Plexiglas discsclamped onto the bases of the cylindrical matrix.3. ExperimentalThus, water penetration and matrix expansionoccurred through the lateral side of the matrix [21].Swellable matrices having a disc shape wereThe transparent discs of the device allowed for directprepared by granulating 10 to 80 parts of buflomedilobservation of the undergoing swelling and forpyridoxal phosphate, (BPP, mol. wt. 553.5; solubility

3 identification of the positions where polymer transi-in water, c 50.65 g cm ; Lisapharma S.p.A., Erba,s

tion and drug dissolution took place.Italy; particle size range, 20–90 mm) mixed with 80


The drug was light yellow in color in the dry state,whereas in solution, it ranged from yellow to orangeas the concentration increased. In fact, the matrixfronts during drug release were visible on the matrixbase as concentric circles corresponding to a sharpchange in drug color. Fig. 2 shows the three frontsformed during the swelling /dissolution process.From the center to the periphery of the matrix, theswelling front (polymer glassy–rubbery transitionboundary), the diffusion front (solid drug–drugsolution boundary) and the erosion front (swollenmatrix–solvent boundary) could be identified.

Clearly, the intensity of the yellow color indicatedFig. 3. Gel layer thickness as a function of time for matricesthe dissolved drug concentration. The fronts movedcontaining different percentages of BPP: (s) 10%, (h) 20%, (n)in relation to the phenomena that were active at their30%, (d) 40%, (j) 60% and (m) 80%. The bars represent the

respective locations. For example, the erosion front standard error of the mean (n53).moved outwards due to the swelling of the matrix orinwards due to matrix dissolution, whereas the distance between the erosion and swelling fronts.swelling front moved inwards as a result of water Fig. 3 shows that the thickness of the gel layer as apenetration. function of time was not very different in the

The gel layer thickness was measured as the systems prepared, despite the varying amounts of

Fig. 2. Image of the matrix containing 60% BPP (w/w) taken after 240 min of swelling.


drug. It was also found that the individual movement of swelling. For example, the color exhibited by theof erosion and swelling fronts (data not shown) was matrix containing 60% of drug ranged from the deepnot significantly different in the various systems orange color of saturated or highly concentrated drugstudied, except for the matrix containing the highest solutions near the diffusion front to the very lightamount of drug (80%), which showed the slowest yellow color near the erosion front.outward movement of the erosion front and the The diffusion front was easy to identify becausefastest movement of the swelling front (i.e., high the deep orange saturated drug solution in the gelledwater penetration rate). In this case, the low amount portion of the matrix was preceded by a distinctof polymer in the matrix rendered the matrix more circular yellow layer. This yellow layer becamesensitive to erosion and, therefore, to water penetra- thicker and denser at high loadings (60–80%). Ittion. appeared to be uniform in color, and corresponded to

In the matrices used in this study, the diffusion the undissolved drug gel layer thickness. Therefore,front moved inwards and was near the swelling front. with an increasing BPP loading, the solid drugThe distance between the diffusion front and the tended to persist after the swelling front and gaveerosion front represented the thickness of the dis- rise to the diffusion front. The sudden increase insolved drug gel layer [3], that is, the diffusive layer color at the diffusion front is an indication that theinvolved in controlling the drug’s release process. local value of the water volume fraction was highFig. 4 shows the variation in the thickness of the enough to dissolve the amount of solid drug presentdrug-dissolved gel layer versus time. As shown in close to the diffusion front. This is also clearlythis figure, samples with 10 and 20% drug exhibited indicated by Eqs. (8) and (9), which indicate that theconsistently higher dissolved drug gel layer thick- diffusion front and the associated finite value of cd

nesses than the other samples, due to the absence of will appear only when the water volume fraction is at*the diffusion front. least y .w

The series of pictures of the matrix base, taken at A detailed analysis of the threshold volume frac-sequential times, showed a gradient of color in the tions (concentrations) of water, drug and polymer ingel layer due to the dissolved drug. The color the specific HPMC matrices investigated here isgradient in the gel is related to the concentration shown in Table 1. For these values, the followinggradient of dissolved BPP. Typical pictures of the calculations were performed.variation of the color in matrices containing increas- First, the threshold concentration, c*, was calcu-ing amount of drug are shown in Fig. 5 after 120 min lated from Eq. (1), where T 5378C, T 51778C forexp g

HPMC, as reported by Doelker [1], the expansion24 21coefficient was a 53.7?10 K , as defined byf

Ferry [18] and the expansion factor b 50.2, asreported by Fujita and Kishimoto [22]. Thus, for ourHPMC system, the value of c* was calculated as0.259 g of water /g of polymer, a value that is closeto those reported by Doelker [1], of 0.23, andHancock and Zografi [23], of 0.20.

To determine the threshold volume fractions, weused Eqs. (2–4), with a polymer density of r 5p

3 31.326 g/cm , a drug density r 51.394 g/cm , asd

determined by helium pycnometry, and a water3density at 378C equal to r 50.993 g/cm [24].w

Finally, as noted before, the water BPP solubility3was c 50.65 g/cm .sFig. 4. Dissolved drug gel layer thickness as a function of time

As indicated in Table 1, the nominal loading offor matrices containing different percentages of BPP: (s) 10%,the studied systems was reported as a percentage of(h) 20%, (n) 30%, (d) 40%, (j) 60% and (m) 80%. The bars

represent the standard error of the mean (n53). the total tablet weight. These values were recalcu-


Fig. 5. Images of the matrices containing different percentages of BPP (w/w), taken after 120 min of swelling.


Table 1Threshold volume fractions for water, drug, polymer and dissolved drug at different nominal loadings of drug, calculated according to Eqs.(2), (3), (4) and (7), respectively

Nominal Loading/concentration Threshold volume fractionsloading cd Water Drug Polymer Dissolved drug(% w/w) * * * *y y y yw d p ds

10 0.125 0.236 0.081 0.683 0.11020 0.286 0.214 0.168 0.618 0.09930 0.500 0.190 0.260 0.549 0.08840 0.800 0.164 0.360 0.475 0.07660 2.000 0.106 0.585 0.308 0.04980 8.000 0.038 0.849 0.112 0.018

lated as values of c required in this work (g of was extremely difficult to visually identify thed

drug/g of polymer) and are shown in the second diffusion front position in matrices containing lesscolumn of Table 1. than 30% drug. According to the theoretical analysis

The threshold volume fractions were calculated presented here, this was due to the high waterunder these conditions and are reported in columns solubility of the drug as well as its loading level. In

* * *3, 4 and 5 as y , y and y for water, drug and fact, while analyzing all matrices that contained BPPw d p

polymer, respectively. It is evident that the relative at concentrations ranging from 10 to 80%, wevolume fractions of each component change accord- noticed that the diffusion front started to be visible ining to the system’s nominal drug loading. Most matrices at above 30% drug, due to the appearancenotably, the water threshold volume fraction varies of the thick circular layer with a uniform yellowfrom 0.0328 to 0.236, with the lowest value being color, which created a sharp variation in color.observed at the highest loading, where only a small In the region beyond the circular yellow layer, theamount of polymer was present. Although we did not matrix was gel-like and the color changed to antake the initial porosity of the matrices (which is of orange color, identifying the boundary separatingthe order of 10%) and the other components used for solid drug from dissolved drug, i.e., the diffusionmatrix production (10% w/w) into consideration in front. It is interesting to note that, as drug-loadingthese calculations, these values are still deemed increased, the presence of this yellow layer appearedreasonable [23]. at earlier swelling times, thus confirming that the

We were also able to calculate the drug volume observed layer was due to incomplete dissolution of*fraction, y , from Eq. (7) using the corresponding the drug at the swelling front. This situation could beds

*threshold water volume fraction, y . The results are explained by considering that the increase in thew

shown in the last column of Table 1 and are quite amount of drug in the matrix augmented the localrevealing as they indicate the amount of undissolved drug’s volume fraction, as shown by the data indrug in the gel layer at the swelling front. For column 4 of Table 1. Therefore, a higher waterexample, for the matrix containing just 10% BPP, the volume fraction was required to completely dissolvetotal drug present at the swelling front was 8.1% the drug present at the diffusion front, as shown in

* *(y 50.081), with 11% (y 50.110) that could Eq. (10).d ds

dissolve. This is the reason why, in Fig. 5, it wasimpossible to visibly detect the diffusion front. Onthe other hand, in the matrix with 80% BPP, the total 5. Conclusionsdrug present at the swelling front was 84.9%, withonly 1.8% being dissolved. From the results obtained in this work, it is

Our results indicate that there was a substantial concluded that a diffusion front can be observed withdifference among the matrices that were loaded with BPP-containing HPMC matrices. This front is visibledifferent drug concentrations. More specifically, it in systems with more than 30% drug, since there is


mobility and diffusion coefficients in the gel layer of HPMCan undissolved drug layer in such systems. A newtablets using NMR imaging, Pharm. Res. 13 (1996) 376–mathematical and physical analysis of such systems380.

shows the importance of drug solubility and loading[10] P. Gao, R.H. Meury, Swelling of hydroxypropyl methyl-

in the observation of the diffusion front. cellulose matrix tablets. 1. Characterisation of swelling usinga novel optical imaging method, J. Pharm. Sci. 85 (1996)725–731.

[11] R. Konrad, A. Christ, G. Zessin, U. Cobet, The use ofAcknowledgementsultrasound and penentrometer to characterize the advance-ment of swelling and eroding fronts in HPMC matrices, Int.

This work was partially supported by a grant from J. Pharm. 163 (1998) 123–131.CNR Italy 96.01066.CT03 and from the Showalter [12] R. Bettini, P. Colombo, G. Massimo, P. Santi, P.L. Catellani,Foundation. N.A. Peppas, Moving fronts and drug release from hydrogel

matrices, Proc. 21st Int. Symp. Control. Rel. Bioact. Mater.21 (1994) 19–20.

[13] R. Bettini, P. Colombo, N.A. Peppas, Drug concentrationReferencesprofiles during drug release from hydrophylic matrices, Proc.24th Int. Symp. Control. Rel. Bioact. Mater. 24 (1997)

[1] E. Doelker, Cellulose derivatives, in: N.A. Peppas, R.S. 635–636.Langer (Eds.), Advances in Polymer Science. Biopolymers I, [14] R. Bettini, N.A. Peppas, P. Colombo, Polymer relaxation inSpringer-Verlag, Berlin, 1993, pp. 199–265. swellable matrices contributes to drug release, Proc. 25th Int.

[2] N.A. Peppas, R.W. Korsmeyer, Dynamically swelling hydro- Symp. Control. Rel. Bioact. Mater. 25 (1998) 36–37.gels in controlled release applications, in: N.A. Peppas (Ed.), [15] P.J. Flory, Principles of Polymer Chemistry, Cornell Uni-Hydrogels in Medicine and Pharmacy, Properties and Appli- versity Press, Ithaca, NY, 1953.cations, Vol. III, CRC, Boca Raton, FL, 1987, pp. 109–135. [16] P.I. Lee, C.-J. Kim, Probing the mechanisms of drug release

[3] P. Colombo, R. Bettini, G. Massimo, P.L. Catellani, P. Santi, from hydrogels, J. Control. Release 16 (1991) 229–236.N.A. Peppas, Drug diffusion front movement is important in

[17] A.T. Pham, P.I. Lee, Probing the mechanisms of drug releasedrug release control from swellable matrix tablets, J. Pharm.

from HPMC matrices, Pharm. Res. 11 (1994) 1379–1384.Sci. 84 (1995) 991–997.

[18] J.D. Ferry, Viscoelastic Properties of Polymers, 3rd ed, John[4] B. Narasimhan, N.A. Peppas, Molecular analysis of drug

Wiley, New York, 1980.delivery systems controlled by dissolution of the polymer

[19] N.A. Peppas, N.M. Fransons, The Swelling interface numbercarrier, J. Pharm. Sci. 86 (1997) 297–304.as a criterion for prediction of diffusional solute release[5] P. Gao, J.W. Skoug, R.P. Nixon, T.R. Ju, N.L. Stemm, K.-C.mechanisms in swellable polymers, J. Polym. Sci. Polym.Sung, Swelling of hydroxypropyl methylcellulose matrixPhys. 21 (1983) 983–997.tablets. 2. Mechanistic study of the influence of formulation

[20] P.I. Lee, Controlled drug release from polymeric matricesvariables on matrix performance and drug release, J. Pharm.involving moving boundaries, in: D.H. Levis (Ed.), Con-Sci. 85 (1996) 732–740.trolled Release of Pesticides and Pharmaceuticals, Plenum[6] I.S. Moussa, L.H. Cartilier, Characterisation of movingPress, New York, 1981, pp. 30–48.fronts in cross-linked amylose matrices by image analysis, J.

[21] R. Bettini, P. Colombo, G. Massimo, P.L. Catellani, T. Vitali,Control. Release 42 (1996) 47–55.Swelling and drug release in hydrogel matrices: polymer[7] C.D. Melia, A.R. Rajabi-Siahboomi, M.C. Davies, Theviscosity and matrix porosity effects, Eur. J. Pharm. Sci. 2development of structural features in the surface gel layer of(1994) 13–219.hydrating HPMC hydrophilic matrices, Proc. 19th Int. Symp.

[22] H. Fujita, A. Kishimoto, Diffusion in polymer–diluentControl. Rel. Bioact. Mater. 19 (1992) 28–29.systems, J. Polym. Sci. 28 (1958) 547–567.[8] C.D. Melia, A.C. Hodsdon, M.C. Davies, J.R. Mitchell,

[23] B.C. Hancock, G. Zografi, The relationship between thePolymer concentration profiles across the surface gel layer ofglass transition temperature and the water content of amor-xanthan, alginate and HPMC matrix systems, Proc. 21st Int.phous pharmaceutical solids, Pharm. Res. 11 (1994) 471–Symp. Control. Rel. Bioact. Mater. 21 (1994) 724–725.477.[9] A.R. Rajabi-Siahboomi, R.W. Bowtell, P. Mansfield, M.C.

Davies, C.D. Melia, Structure and behaviour in hydrophylic [24] K. Diem (Ed.), Documenta Geigy, Tables Scientifiques, 6thmatrix sustained release dosage forms: 4. Studies of water ed., J.R. Geigy S.A., Dept. Pharmaceutique, Basel, 1963.

observation of swelling process and diffusion front ... · the behavior of gel layer thickness in...

Documents