Chemical Engineering and Processing 37 (1998) 257–263
`
`Near infrared drying of pharmaceutical thin films: experimental
`analysis of internal mass transport
`
`S. Le Person a,b, J.R. Puiggali a,*, M. Baron b, M. Roques a
`a Laboratoire de Ge´nie des Proce´de´s de Pau (EAD 1932), ENSGTI, rue Jules Ferry, 64000 Pau, France
`b Centre Poudres et Proce´de´s, ENSTIMAC, 81013 Albi Cedex 09, France
`
`Received 8 December 1997; received in revised form 11 March 1998; accepted 11 March 1998
`
`Abstract
`
`Thin polymer films are produced for pharmaceutical purposes from a solution in water and polar solvent. These coatings
`contain also a heavy solvent with the active substance as solute. A global analysis is presented here of the drying rate in terms
`of average content for each component, i.e. water, heavy solvent and active substance. The results are used to better the global
`drying process. The quality of the final product imposes to know the local distribution of the key products. The use of Laser
`Scanning Confocal Microscopy (LSCM) on two molecules is proposed here; the active substance and the heavy solvent.
`Combining the two set of results, one is able to determine the cases of maldistribution of the active substance and of eventual
`crystallisation which in turn will have consequences on storage and successful transdermal delivery of the drug. © 1998 Elsevier
`Science S.A. All rights reserved.
`
`Keywords: Drying kinetics; Coating; Selectivity; Laser Scanning Confocal Microscopy
`
`1. Introduction
`
`Modern industries produces more and more thin
`polymer films and coatings for many purposes, e.g.
`adhesives, varnishes and paints. In the pharmaceutical
`industry some films are used in patches for transdermal
`drug delivery. Drying is the essential unit operation
`necessary to form the final product. In all cases, master-
`ing of process variable and microscopic aspects of the
`product quality entails chemical and process engineer-
`ing and transport phenomena as basic sciences.
`After preparation, the coating mixture is spread on a
`web and submitted to drying in a tunnel or an oven.
`Frequently, impinging jets and Infra Red Radiation [1]
`accomplish the work in a short drying time (100 s as an
`order of magnitude). In the end, one must be sure that
`the selected process and its conditions is able to ensure
`the right product quality; a limited remanence of the
`process solvent (generally a mixture of volatile solvents)
`and a given quality product [2], i.e. physical and chem-
`
`* Corresponding author. Tel.: (cid:27)33 597 22080; fax: (cid:27)33 559
`722081; e-mail: jean-rodolphe.puiggali@univ-pau.fr
`
`0255-2701:98:$19.00 © 1998 Elsevier Science S.A. All rights reserved.
`PII S0255-2701(98)00032-4
`
`ical homogeneity and an appropriate distribution of
`active substance.
`The tools to design the correct process are pilot plant
`experiments, bench scale experiments and modelisation
`of transfers. In this paper, small scale experiments were
`opted for and an experimental approach of internal
`transfers. Evidently, the diffusional approach of com-
`plex systems containing two immiscible solvents, a
`shrinking polymeric macromolecule network and an
`active substance, cannot be tracked from the basic
`text-book equations. What is modelisable is already
`intuitively and:or experimentally known. It would take
`a lot of basic investigation on simpler systems to make
`a substantial progress on the only problem of cross
`diffusivities [3]. For all those reasons it has been sug-
`gested, in this paper, to track the internal distribution
`of some key products by Laser Scanning Confocal
`Microscopy (LSCM). Adding an integral chemical
`analysis of the film, one is then able to quantify the
`absolute distribution for films produced under variable
`conditions.
`After few other investigators [4,5], the use of LSCM
`for films was advocated because this technique elimi-
`nates all informations in the depth but the one from a
`specific plane. By moving vertically the sample or the
`
`RBP_TEVA05024349
`
`TEVA EXHIBIT 1026
`TEVA PHARMACEUTICALS USA, INC. V. MONOSOL RX, LLC
`
`

`
`258
`
`S. Le Person et al. :Chemical Engineering and Processing 37 (1998) 257–263
`
`Table 1
`Typical wet composition of the coating
`
`Acrylic polymer (Pol)
`
`Light solvents
`
`Active phase
`
`Sl1
`
`Sl2
`
`Sl3
`
`Excipient heavy solvent (E)
`
`Active substance (PA)
`
`Molecular weight
`Typical composition (100 g)
`
`52
`
`18 (water)
`38
`
`46
`3
`
`76
`4
`
`190
`2
`
`340
`1
`
`objective, one gets a resolution of ca. 1 mm. Not all the
`molecules are susceptible to be excited by laser beam
`but we were fortunate to have two in our sample and
`could then reconstruct all interesting distribution.
`The release kinetics of a drug by application onto the
`skin depends on the appropriate distribution of the
`drug which in turn depends on the drying process.
`Hence the selection of a process is based upon the
`thermosensitivity of the active substance and a homoge-
`neous treatment of the film [6]. It appeared that after a
`rapid screening that SIR plus convection was the best
`solution. It is proposed in this paper to validate this
`first guess by an analysis in depth of the product
`quality.
`
`2. Materials and experiences
`
`2.1. Product composition
`
`The coating mixture is constituted of: (1) an acrylic
`adhesive polymer (Pol); (2) three light solvents (Sli); (3)
`one heavy solvent named hereafter excipient (E); (4)
`one pharmaceutical active substance (AS) reputedly
`soluble in E. As excipient is essential to the administra-
`tion of the drug, the drying process must evacuate the
`light solvent and preserve a maximum amount of E,
`hence the selectivity requirement. A typical wet coating
`composition is given in Table 1 together with the
`molecular weight of the essential products.
`The final composition of the coating for each of the
`solvents, excipient and active substance is also part of
`the industrial target, the mass distribution within the
`thickness of the coating was added to this ordinary
`constraint; these two pieces of information have to be
`analysed as selective drying. Moreover during drying
`process the thermal degradation of the active substance
`must be avoided.
`
`2.2. Experimental set-up
`
`In order to substantiate the choice of the best heat
`supply mode with respect to drying time and quality,
`the coating drying were studied by convection, conduc-
`tion, medium and short infra-red respectively (MIR and
`SIR), each of the last drying modes is always associated
`
`with convection. Experiments were carried out in a
`laboratory set-up mainly composed of two parts, the
`drying cell and the wind tunnel (see Fig. 1):
`(cid:147) the wind tunnel is a conventional drying rig where
`temperature
`(U(cid:12)),and
`humidity
`(T(cid:12)db),velocity
`(Y(cid:12)), of air are controlled,
`(cid:147) in the drying cell the sample weight is continuously
`recorded; depending on the combined heat supply, a
`heated slab (Tc), non heated in the case of pure
`convection, can support the sample (conduction) or
`an infrared emitter (MIR or SIR) can irradiate (PIR)
`from a given height (h) the sample.
`Operating conditions for the different experiments to
`qualify quality with respect to heat supply are sum-
`marised in Table 2.
`
`2.3. Drying kinetics
`
`During those experiments the mass of the sample is
`either continuously recorded or acquired at regular
`intervals. Moreover the average composition is deter-
`mined on partially dried coatings by means of chro-
`matographic analysis and coulometry. Then, each point
`of a curve corresponds to a separate drying experiment
`carried out at least twice. The initial coating thickness
`for all the tests was 200 mm.
`Drying kinetics reported on Fig. 2 show that the
`initial drying velocity with infrared is about twice the
`velocity with other heating modes. Source thermal iner-
`tia affects greatly the first two points of the MIR drying
`curve bringing an abnormal delay. Between conduction
`and SIR the latter one is far more preferable.
`
`Fig. 1. Schematic view of the experimental set-up.
`
`RBP_TEVA05024350
`
`TEVA EXHIBIT 1026
`TEVA PHARMACEUTICALS USA, INC. V. MONOSOL RX, LLC
`
`

`
`S. Le Person et al. :Chemical Engineering and Processing 37 (1998) 257–263
`
`259
`
`Table 2
`Typical wet composition of the coating
`
`Table 3
`Final composition indicators for different heating modes
`
`Heat supply
`
`Operating conditions
`
`Indicators
`
`SIR
`
`MIR
`
`Conduction
`
`Pure convec-
`tion
`
`IRC
`
`IRM
`
`Conduction
`
`Pure convection
`
`PIR(cid:30)170 W; h(cid:30)5 cm; U(cid:12)(cid:30)2 m:s; T(cid:12)db(cid:30)
`30°C; T(cid:12)wb(cid:30)18°C
`PIR(cid:30)140 W; h(cid:30)6 cm; U(cid:12)(cid:30)2 m:s; T(cid:12)db(cid:30)
`30°C; T(cid:12)wb(cid:30)18°C
`Tc(cid:30)60°C; U(cid:12)(cid:30)4 m:s; T(cid:12)db(cid:30)65°C; T(cid:12)wb(cid:30)
`44°C
`U(cid:12)(cid:30)4 m:s; T(cid:12)db(cid:30)75°C; T(cid:12)wb(cid:30)36°C
`
`X
`
`Mass ratio
`(Sli(cid:27)E):Pol
`Mass ratio E: Y
`Sl1
`Drying time
`range
`
`:X
`
`0.8X
`
`:0.8X
`
`0.9Y
`
`0.5Y
`
`0.6Y
`
`Analytical results on films show that the various
`heating processes can be classified by the remaining
`ratio of solvents to dry matter (Sli(cid:27)E):Pol, by the
`ratio of heavy solvent to water E:Sl1 and by the drying
`time. In Table 3 it is shown that SIR is vastly preferable
`in order to keep the heavy solvent inside the product in
`the minimum drying time. The ratio (Sli(cid:27)E):Pol is not
`discriminant.
`
`2.4. Temperature during SIR drying
`
`Although rapid, and quality respectful, the SIR dry-
`ing might harm the product because of local over
`heating. In order to evaluate this danger, It was realised
`that a thermal study of the coating for different in-
`frared powers,
`for different
`thickness and coating
`supports.
`Three thermocouples were embedded in a thick coat-
`ing of 2–15 mm, an extrapolation of the temperature
`profile towards the irradiated surface permits to evalu-
`ate the highest temperature obtained by irradiation for
`various drying times and constant infrared power (see
`Fig. 3).
`From Fig. 3, one can infer that 65°C is the maximum
`temperature reached at the end of the drying process
`and hence nor alteration nor destruction is to be feared
`
`according to the stability diagram of the active sub-
`stance. Similar results have been obtained by optic
`pyrometry on similar coatings (acrylic polymers) by
`Navarri and Andrieu [7].
`It was concluded that SIR drying is the appropriate
`answer to the drying time and quality constraints.
`
`2.5. Thermodynamic of the acti6e phase
`
`During the drying process the thin layer temperature
`increases as the solvent contents decrease. At the end of
`the process the coating is cooled at a given AS mass
`fraction. All these changes in temperature and composi-
`tion mass and thermal steps are susceptible to produce
`active substance crystallisation in the thin film and
`therefore the final quality of the product might be out
`of bonds. A good knowledge of solubility behaviour of
`the active substance in the heavy solvent is necessary to
`avoid this arowback. An analysis of the solvent mass
`loss coupled with the coating temperature increase is a
`good tool to propose drying schedules in relation with
`dried product quality.
`
`Fig. 2. Drying kinetics of a 200 mm thick coating under four drying
`different processes: pure convection (
`), combined process: convec-
`tion plus MIR (x), SIR ((cid:147)), conduction ((cid:15)).
`
`Fig. 3. Averaged temperature of a 100 mm thick coating under Short
`Infra Red drying (Table 2) obtained by extrapolation of experimental
`results on 2–15 mm thick coatings.
`
`RBP_TEVA05024351
`
`TEVA EXHIBIT 1026
`TEVA PHARMACEUTICALS USA, INC. V. MONOSOL RX, LLC
`
`

`
`260
`
`S. Le Person et al. :Chemical Engineering and Processing 37 (1998) 257–263
`
`Fig. 4. Active substance solubility curve in the active phase (heavy
`solvent plus active substance) obtained by cooling of a solution
`initially at 70°C by air at 15°C.
`
`Solubility of the active substance in the heavy solvent
`is obtained by cooling (ambient air flow) of the active
`phase solution at a given mass fraction (MAS:MAS(cid:27)
`ME) initially unsaturated and homogenised at 75°C
`with ultra sounds. A thermocouple placed in the solu-
`tion allows to follow the temperature evolution and,
`linked with an observation of active substance crystals
`through a microscope, the saturation temperature is
`obtained and the solubility curve plotted (Fig. 4).
`On the same graph, the AS mass fraction is plotted
`versus temperature as the SIR drying process of the
`heavy phase E proceeds. One can notice that no crys-
`tallisation problem occurs if the mass fraction is consid-
`ered in terms of an average value.
`In order to show some light on the selective character
`of the drying operation, purpose of which is not just to
`remove the moisture, a more accurate study is neces-
`sary. Therefore, in relation with the quality and the
`final product requirements, one has to investigate, more
`precisely than an average kinetics study, the repartition
`of the active phase components (heavy solvent and
`active substance) inside the thin film.
`
`3. Results
`
`3.1. Time e6olution of components
`
`In order to follow the evolution in mass of each
`component during drying, there is only one way; repeat
`a given experiment as many times as you need an
`analysis. This regular dosage is carried out by liquid
`and gas chromatography and also by coulometry,
`(complete analysis every 30 s); results are given on Fig.
`5.
`
`Using the selectivity coefficient Si of a component i in
`a solution as defined by Riede and Schlu¨nder [8] and
`used by Wolf [9] for hygroscopic materials:Ssolvent i(cid:30)
`
`Fig. 5. Mass evolution of each components during drying kinetics of
`a 100 mm thick coating under Short Infra Red drying process.
`nsolvent i
`(cid:28)xsolvent i, one can evaluates the drying effi-
`nall solvents
`ciency with respect to water and heavy solvent. In this
`last relation, ni is the evaporative mass flux of solvent i,
`nall solvents the total evaporative mass flux and ixi the
`mass fraction of the solvent i. As indicated in Fig. 6,
`where Swater and Sheavy solvent are plotted versus drying
`time, water evaporates preferentially, regarding to the
`heavy solvent, at the beginning of the drying process.
`The thin film thickness is also plotted on this figure
`showing that shrinkage occurs essentially when water
`(major solvent) intensely evaporates [10,11].
`In Fig. 5 the selective character of drying is demon-
`strated since 99% of the initial water amount is evapo-
`rated in 10 min. In the same time only 2.05% of heavy
`solvent evaporates; this last evaporation is due to the
`absence of heavy solvent vapour in the drying air. An
`important practical consequence of this heavy solvent
`evaporation is that one must shorten the drying time to
`maintain most of this solvent inside the product.
`
`Fig. 6. Selectivity factors for the two major solvents, i.e. water (Sl1
`
`) and heavy solvent (E (cid:5)) during drying of a 100 mm thick coating
`under Short Infra Red drying process. Indication of the shrinkage in
`terms of reduced thickness.
`
`RBP_TEVA05024352
`
`TEVA EXHIBIT 1026
`TEVA PHARMACEUTICALS USA, INC. V. MONOSOL RX, LLC
`
`

`
`S. Le Person et al. :Chemical Engineering and Processing 37 (1998) 257–263
`
`261
`
`So far there are only global results that is to say
`drying curves for each component of the coating. The
`results are valuable for a good management of the
`drying operation but failed to give detailed informa-
`tions on the quality of the product. Again, precise
`knowledge of the spatial repartition component in the
`volume is completely necessary to assess the functional-
`ity of the final product.
`A new non destructive exploration method has been
`used, tested and validated in order to follow the spatial
`distribution of the components of the active phase.
`Laser Scanning Confocal Microscopy (LSCM) can
`provide this type of informations in as much as the
`probe molecule
`is fluorescent under
`laser beam
`radiation.
`
`3.2. Spatial e6olution of the components
`
`The LSCM allows, without sample preparation, the
`observation of a given surface called focal surface in-
`side the thickness of the sample, avoiding the destruc-
`tive slicing, with a much better resolution. Precisely,
`with a micronic resolution one is able to obtain accu-
`rate data on very thin sub-millimetric films.
`The principle feature of confocal
`imaging is that
`detection is strictly limited to what is in focus; LSCM
`avoids out-of-focus informations from neighbouring
`planes.
`The substance is irradiated by a laser beam. An
`intensity of fluorescence, proportional to the quantity
`of excited substance in the focus plane, is the informa-
`tion in return. Only aromatic molecules and:or ele-
`ments with double conjugated links are excited by laser
`beam UV (364) or Visible (633, 543, 488). A map of the
`average fluorescence relative to the studied component
`can be obtained on the focus plane. By displacement of
`the focus plane, one can follow the fluorescence inten-
`sity of the component in the thickness of the sample
`[12].
`Studies were carried out, with the help of Dr Cather-
`ine Garnier, at the Macromolecule Physico-Chemistry
`Laboratory (LPCM-INRA) at Nantes (F). All fluores-
`cence measurements were performed using a micro-
`scope Axiovert 135M-Zeiss, an UV 364-Microsysteme
`laser as the excitation source and a SLM-Zeiss as
`spectrofluorimeter.
`A miniature drying ring obviously had to be brought
`consisting of the infrared lamp and a blower in the
`vicinity of the LSCM microscope. Each sample was
`dried, on a glass plate, for a given length of time (5, 10,
`15 min) then sealed over by a plastic film, brought to
`the microscope and analysed for 120 s during which the
`relaxation processes are unavailable but limited.
`In parallel a chemical dosage allows the determina-
`tion of the total mass of the fluorescent element in the
`studied substance. Therefore it is easy to connect the
`
`Fig. 7. Volumic contents of the water, active phase and polymer
`during drying of a 100 mm thick coating under Short Infra Red
`drying process. Indication of the shrinkage in terms of coating
`thickness before drying and after 5 min of drying.
`
`fluorescence informations to quantitative data.
`Amongst the constitutive elements of the enduction,
`only the active substance and the heavy solvent can be
`excited by a UV364 laser beam.
`After the disappearance of water from the thin layer,
`no further shrinkage occurs and the film consists of
`polymer and active phase (heavy solvent and active
`component) the volume of which are simply additive.
`It is obvious then to calculate the spatial distribution
`of the polymer, knowing that of the others components.
`
`4. Discussion
`
`Because of the fluorescence of the two constituents of
`the active phase, an analysis of many preparations, with
`different concentrations,
`is necessary before studying
`the real enduction in order to differentiate the respec-
`tive impact both of the active substance and of the
`heavy solvent.
`Observations were carried out on enductions after 5,
`10 and 15 min of drying. Before drying an observation
`is made in order to assess the degree of homogeneity of
`the constituents inside the thin layer.
`After a 5 min drying analytical dosages (Fig. 5) show
`that only three components remain in the layer: (1)
`acrylic polymer; (2) active substance; and (3) heavy
`solvent. As the thin layer shrinkage is known (Fig. 6),
`one can transpose the LCSM data in terms of volumic
`content fields (Fig. 7). The intense moisture removal,
`through the exposed surface of the layer to the radia-
`tion, during the first 3 min of drying (Fig. 7) produces
`a stress on the polymer skeleton. This stress increases in
`the water flow direction and as a result the acrylic
`polymer becomes more and more dense in the upper
`part of the layer (exposed surface). This intense shrink-
`age coupled with the polymer compaction causes a
`displacement of the active phase towards the bottom of
`the layer (Fig. 7).
`
`RBP_TEVA05024353
`
`TEVA EXHIBIT 1026
`TEVA PHARMACEUTICALS USA, INC. V. MONOSOL RX, LLC
`
`

`
`262
`
`S. Le Person et al. :Chemical Engineering and Processing 37 (1998) 257–263
`
`Fig. 8. Volumic content fields of the heavy solvent at three drying
`times (300 2, 600 , 900 s x) during drying of a 100 mm thick
`coating under Short Infra Red drying process.
`
`This instantaneous view of the volumic content fields
`for the thin layer constituents show a very unstable
`state. One can easily imagine that the compressive wave
`must propagates from the top to the bottom of the
`enduction. Therefore at the end of the drying process
`the global skeleton structure will be equilibrated.
`During the thin layer drying, two main stages are
`identified on Figs. 8 and 9 where the spatial evolution
`of the heavy solvent and the active substance are re-
`ported after 5, 10, 15 min of drying:
`a first stage (t55 min for the drying condition),
`where water is intensively removed producing a
`severe shrinkage of the layer and a compaction of the
`polymeric skeleton at the top (in fact the exposed
`surface); as a result, the active phase in solution in
`the heavy solvent is displaced towards the bottom,
`a second stage where the volumic content fields of
`each component equilibrate themselves in the layer
`due to the combined displacement of the solvent and
`the relaxation of the polymer skeleton; the discharge
`of the polymer towards the bottom giving a stable
`
`Fig. 9. Volumic content fields of the active substance at three drying
`times (300 2, 600 , 900 s x) during drying of a 100 mm thick
`coating under Short Infra Red drying process.
`
`Fig. 10. Mass fraction fields of the active substance at three drying
`times (300 2, 600 , 900 s x) during drying of a 100 mm thick
`coating under Short Infra Red drying process. Indications of the
`possible zones of crystallisation.
`
`mechanical state is joined to a movement from bot-
`tom to top of the active phase with a partial removal
`of the heavy solvent.
`Initially,
`in the thin layer the constituents of the
`active phase are homogeneously distributed. But the
`pharmaceutical active substance has a larger steric bulk
`than the heavy solvent and therefore might react differ-
`ently to the stresses imposed by the skeleton of the
`acrylic polymer during the drying process. Then a
`detailed study of the spatial evolution of the active
`substance constituents must be carried out in order to
`prevent risks of separation of the active substance from
`the active phase.
`Combining measurements made by LCSM and solu-
`bility data (Fig. 4) of the active substance in the heavy
`solvent lead one to suppose the existence of the active
`substance in its crystalline form. This is true for the
`residual quantity of the active molecules trapped in the
`retracted skeleton of the adhesive polymer 5 mm under
`the exposed surface (Fig. 10).
`After 5 min of drying the system tends to reequili-
`brate the mechanical stresses. However, because water
`is no more present in the system, the polymeric network
`is not turgescent and the meshes are densely packed.
`The polymer skeleton acts as a filter for the active
`substance when the system reequilibrates. By their re-
`spective mobility and bulk, the heavy solvent will slip
`easier than the active substance through the net during
`this phase of mechanical reequilibrium just as in thin
`layer chromatography.
`Consequently, after 10 min of drying, one can notice
`(Figs. 8 and 9) an inversion of the heavy solvent and
`active substance concentration profiles inside the enduc-
`tion. Between the 5th and the 10th min of drying the
`heavy solvent migrates towards the exposed surface and
`is removed whereas the active substance, slowed down
`in its migration, stays in the bottom of the layer.
`
`RBP_TEVA05024354
`
`TEVA EXHIBIT 1026
`TEVA PHARMACEUTICALS USA, INC. V. MONOSOL RX, LLC
`
`

`
`S. Le Person et al. :Chemical Engineering and Processing 37 (1998) 257–263
`
`263
`
`Because of a quicker migration effect than the evapora-
`tion one the heavy solvent is stored in the top of the
`coating. The lack of heavy solvent with respect to the
`active substance in the bottom of the layer should lead
`to find, in this location, an important crystallisation.
`This is shown in Fig. 10 where crystallisation is indi-
`cated on a 15 mm thickness in the bottom of the layer.
`As the drying process proceeds, the active substance,
`from its strong affinity with the heavy solvent migrates
`and homogenises in the enduction thickness. After 15
`min of drying, a quasi equilibrium is obtained for the
`components of the active phase, taking into account the
`evaporation of the heavy solvent.
`
`5. Conclusion
`
`The coupling between studies performed, on the one
`hand on a temporal basis (chromatographic and coulo-
`metric analysis), and on the other hand on a spatial
`basis (LSCM) allows to propose a model of the con-
`stituent
`transports inside the film whose thickness
`shrinks from 100 to 50 mm during drying.
`The schematic description of the behaviour of the
`acrylic polymer and the solubility curve of the active
`substance in the heavy solvent allow to explain and to
`prevent mechanisms of separation of the active phase in
`the layer thickness. Such separation effects can produce
`harmful crystallisation phenomena.
`These studies on overall and local behaviour of each
`components of the pharmaceutical thin film are fully
`sufficient in order to propose drying schedules. Follow-
`ing these lines and the guide suggested by Nadeau and
`Puiggali J.R. [13], a pilot plant for thin pharmaceutical
`layer drying was designed [14].
`
`References
`
`[1] E.D. Cohen, E.J. Lightfoot, E.B. Gutoff, A primer on forming
`coatings, Chem. Eng. Prog. September, (1990) 30–36.
`[2] Z.W. Wicks, F.N. Jones, S.P. Pappas, Organic coatings: science
`and technology, vol. 1: Film Formation, Components and Ap-
`pearance, Wiley, New York, 1992, pp. 343.
`[3] R. Taylor, R. Krishna, Multicomponent Mass Transfer, Wiley,
`New York, 1993, p. 579.
`[4] H. Verghoot, J. van Dam, P. de Boer, A. Draaijer, M. Houpt,
`Confocal laser scanning microscopy: a new method for determi-
`nation of the morphology of polymer blends, Polymer 34 (6)
`(1993) 1325–1329.
`[5] L. Li, S. Sosnowski, C.E. Chaffey, S.T. Balke, M.A. Winnik,
`Surface morphology of a polymer blend examined by laser
`confocal fluorescence microscopy, Langmuir 30 (8) (1994) 2495–
`2497.
`[6] R.O. Potts, G.W. Cleary, Transdermal drug delivery: useful
`paradigms, J. Drug Targeting 3 (1995) 247–251.
`[7] P. Navarri, J. Andrieu, High intensity infrared drying study. Part
`I. Case of thin coated films, Chem. Eng. Process. 32 (1993)
`319–325.
`[8] T.H. Riede, E.U. Schlu¨nder, Selective evaporation of a ternary
`mixture containing one nonvolatile component with regard to
`drying processes, Chem. Eng. Process. 28 (1990) 151–163.
`[9] H. Wolf, Selective drying of hygroscopic materials wetted with
`binary mixtures, Drying Technol. 12 (6) (1994) 1445–1470.
`[10] W. Aregba, W. Jomaa, J.R. Puiggali, Une approche du se´chage
`de mate´riaux fortement de´formables, Rev. Phys. Appl. 25 (1990)
`1161–1175.
`[11] M. Moyne, N. Kechaou, P. Do Amaral Sobral, M. Roques, A.
`Cairault, H. Bizot, Mechanism of water transport in drying gels,
`Int. Chem. Eng. 34 (3) (1994) 360–369.
`[12] H. Jinnai, Y. Nishikawa, T. Koga, T. Hashimoto, Direct obser-
`vation of three-dimensional bicontinuous structure developed via
`spinodal decomposition, Macromolecules 28 (13) (1995) 4782–
`4784.
`[13] J.P. Nadeau, J.R. Puiggali, De l’analyse d’un produit a`
`conception d’un se´choir, Entropie 49 (1982) 41–47.
`[14] S. Le Person, Se´chage Se´lectif d’Enductions Pharmaceutiques,
`PhD thesis, Universite´ de Pau et des Pays de l’Adour (F), 339
`(1996) 187.
`
`la
`
`.
`
`RBP_TEVA05024355
`
`TEVA EXHIBIT 1026
`TEVA PHARMACEUTICALS USA, INC. V. MONOSOL RX, LLC