MMICROENCAPSULATION

EDITH MATHIOWITZMARK R. KREITZBrown UniversityProvidence, Rhode IslandLISA BRANNON-PEPPASBiogel TechnologyIndianapolis, Indiana

KEY WORDS

CharacterizationCoacervationEmulsionFormulationGelationHot meltInterfacial polymerizationMembraneMicrocapsuleMicroencapsulationMicrospherePhase separationPolymerSolvent evaporationSolvent removalSpray-drying

OUTLINE

IntroductionBackground

Classification of Microencapsulation TechniquesClassification of Microspheres and MicrocapsulesClassification of Polymeric MembranesRelease Characteristics of Microspheres andMicrocapsulesEmulsion FormationEvaluation of EmulsionsPhase Properties of Emulsion: Hydrophilic-Lipophile BalancePolymer CharacterizationPhase Separation of PolymersInterfacial Phenomenon

Solvent EvaporationComplex Coacervation: Gelatin and Gum ArabicOrganic Phase Separation: CoacervationInterfacial Polymerization

Interfacial Coacervation

Thermal DenaturationGelation: AlginateGelation AgaroseHot Melt MicroencapsulationSolvent RemovalSpray-DryingOne-Step Formation of Double-Walled MicrospheresNovel Encapsulation TechniquesApplications of Microencapsulation

Encapsulation in the Food, Consumer Products,and Cosmetics IndustriesApplications of Encapsulation for AgriculturalProductsApplications and Opportunities for Encapsulationto Enhance Human Health

Bibliography

INTRODUCTION

Microencapsulation is one of the most intriguing fields inthe area of drug delivery systems. It is an interdisciplinaryfield that requires knowledge of the field of pure polymerscience, familiarity with emulsion technology (1), and anin-depth understanding of drug and protein stabilization(2). In the early 1970s this area was considered more of anart than a science because most of the research was de-veloped in pharmaceutical companies and very little infor-mation was discussed in scientific meetings. Today, thetopic of microencapsulation is extensively studied insidemajor pharmaceutical companies and universities as wellas research institutes. Some journals are now solely dedi-cated to the area of microencapsulation (e.g., Journal ofMicroencapsulation). Although scientists at the beginningof the 1970s were primarily concerned with the encapsu-lation of dyes to produce carbonless paper, scientists todayhave mastered the technology to such a level that cells aswell as delicate proteins and genes can be encapsulated.Because two entries in this encyclopedia are dedicated toliposomes and nanoparticles, they are not going to be dis-cussed in this article. In addition, traditional methods suchas coating are discussed in a separate article related tocoatings. This article is divided into three main sections.In the first section we discuss general aspects of polymertopics related to microencapsulation. Then we describe inmore detail the main methods of preparing microcapsulesand microspheres, and we end with a general summary ofthe different fields of application of microencapsulation.

BACKGROUND

Classification of Microencapsulation Techniques

Microencapsulation is a technology devoted to entrappingsolids, liquids, or gases inside one or more polymeric coat-

ings (3-43). Two major classes of encapsulation methodshave evolved, chemical and physical. The first class of en-capsulation involves polymerization during the process ofpreparing the microcapsules. Examples of this class areusually known by the name of interfacial polymerizationor in situ polymerization (44-49). The second type involvesthe controlled precipitation of a polymeric solution whereinphysical changes usually occur. Several books examine thedifferent processes of microencapsulation, and readers arestrongly recommended to consult them (2,5,19,30,39,42,49-58). Perhaps one of the best courses on the fabrica-tion of microcapsules is given at the University of Wash-ington by Dr. Thies.

There are a huge and increasing number of encapsula-tion processes. In addition, many new patents evolve solelyon the basis of novel ways to produce microspheres. Thus,many scientists try to develop a systemic nomenclature forencapsulation classification. In Table 1 we classified thedifferent types of methods according to published litera-ture. It is sometimes difficult to classify encapsulationmethods because specific techniques can be hybrids of twoor more methods or can use different mechanisms simul-taneously. Also, many names have changed throughout theyears (e.g., solvent evaporation has been called water dry-ing and double emulsion), and this can create confusion.

Interfacial polymerization involves the condensation oftwo monomers at the interface of the organic and aqueousphases. Polyamide capsules are a great example of thissystem and are discussed later in this article (44-48).

Complex coacervation was the process used to make themicrocapsules in the first successful encapsulated product,carbonless copy paper (11,14,22,27,29,32,59-66). Complexcoacervation encapsulation processes use the interactionof two oppositely charged poly electrolytes in water to forma polymer-rich coating solution called a coacervate (68).This solution (or coacervate) engulfs the liquid or solid be-ing encapsulated, thereby forming an embryo capsule.Cooling the system causes the coacervate (or coating so-lution) to gel via network formation. Gelatin is a primarycomponent of most complex coacervation systems.

Coacervation uses the common phenomenon of poly-mer-polymer incompatibility to form microcapsules. Thepolymer that is to become the capsule wall material is dis-solved in a solvent and to this solution a second polymer(called the phase inducer) is introduced. Because the twopolymers are incompatible, two polymer-rich phases form.

If drug particles are then introduced, one phase, rich in thedesired coating polymer, engulfs the drug being encapsu-lated thereby forming embryo capsules. In principle, therange of polymers that can be used in this process is es-sentially infinite. In practice, the number of polymers thathave been used successfully is relatively small, for reasonsthat will be discussed later.

The precipitation and/or gelation processes listed in Ta-ble 1 cover many techniques. One example is the precipi-tation of water-soluble polymers such as gelatin withwater-miscible solvents such as isopropanol. Other exam-ples include the precipitation of ethyl cellulose from cyclo-hexane by cooling, the gelation of sodium alginate withaqueous calcium salt solutions (41,68-74), and the ther-mally induced precipitation of proteins to form micro-spheres. In all cases, the objective is to precipitate apreformed polymer around the core (sometimes a multi-particulate core) to cause encapsulation.

Salting-out also listed in Table 1, involves the additionof salt to an aqueous polymer solution ultimately causingthe polymer to phase separate from solution. One potentialproblem with this process is the possibility of incorporatinga relatively high concentration of salt in the final capsulewall, as these salts may have an adverse effect on capsulerelease behavior.

Solvent evaporation (18,36,75-80) is the most popularway to accomplish encapsulation. A core material and cap-sule wall material are briefly dissolved in a water-immiscible, volatile organic solvent and the resulting so-lution is emulsified in an aqueous solution. The solvent isallowed to evaporate, thereby producing solid microcap-sules or microparticles. Another version involves forminga double emulsion where an aqueous core material solutionis emulsified in a polymer-volatile organic solvent solu-tion. The resulting emulsion is emulsified in water givinga double emulsion. Evaporation of the volatile solventyields a solid microcapsule with an aqueous core. Much ofthe effort on encapsulation in the pharmaceutical field hasconcentrated on using biodegradable polymers. A recentreview by Brannon-Peppas (43) describes some of these ad-vances with both microparticles and nanoparticles.

Hot melt encapsulation was developed to avoid the useof solvents throughout the process (81). Solvent removal(82,83) was developed as a modification of the solventevaporation technique, using organic solvents as the ex-tracting medium. In spray-drying, the evaporation of the

Table 1. Classification of Microencapsulation Methods

Process

Interfacial polymerizationComplex coacervationCoacervationThermal denaturationSalting-outSolvent evaporationHot meltSolvent removalSpray-dryingPhase separation

Coating material

Water-soluble and insoluble monomersWater-soluble polyelectrolyteHydrophobic polymersProteinsWater-soluble polymerHydrophilic or hydrophobic polymersHydrophilic or hydrophobic polymersHydrophilic or hydrophobic polymersHydrophilic or hydrophobic polymersHydrophilic or hydrophobic polymers

Suspended medium

Aqueous/organic solventWaterOrganic solventOrganicWaterOrganic or waterAqueous/organic solventOrganic solventsAir, nitrogenAqueous/organic

solvent is achieved in a special, temperature-controlled cy-clone. And finally, phase separation is a new method inwhich a one-step precipitation of two polymers or more pro-duces double-walled microspheres (84).

When someone is faced with the challenge of encapsu-lating a substance, the first question should be, "What isthe final application of the product?" As examples, if thefinal target is pesticide encapsulation, one must firstchoose polymers that are stable and nonerodible, or if oneis to encapsulate therapeutic drugs such as proteins, thechoice of polymers becomes more restricted to bioerodibleones. Once the polymer system is chosen, the next step isto select an appropriate encapsulation method. This canundertaken by either reviewing the relevant literature ordeveloping new methods of encapsulation. Once the correctmethod of encapsulation has been determined the nextstep is to prepare the microcapsules, ensuring reproduci-bility, high encapsulation efficiency, and preservation ofthe activity of the encapsulated substance. This is achievedby thorough characterization, both before and after micro-sphere fabrication. It is an integral component, necessaryfor replication and optimization, and is discussed later inthis article.

Classification of Microspheres and MicrocapsuiesTwo general structures exist: microcapsules and micro-spheres. A microcapsule is a system that contains a well-defined core and a well-defined envelope: the core can besolid, liquid, or gas; the envelope is made of a continuous,porous or nonporous, polymeric phase. Figure Ia shows thedifferent microcapsule configurations; the drug can be dis-persed inside the microcapsule as solid particulates withregular or irregular shapes. Other forms may consist of apure or dissolved solution, suspension, emulsion, or a com-bination of suspension and emulsion. Specific applicationssometimes require modifications, e.g., when proteins areencapsulated they may contain stabilizers as well as theactive ingredient (2). Also, the core can be something otherthan a chemical meant for release. An interesting appli-cation is the encapsulation of gases for the use of ultrasonicimaging (85-87). Alternatively, a microsphere is a struc-ture made of a continuous phase of one or more misciblepolymers in which particulate drug is dispersed, at eitherthe macroscopic (particulates) or molecular (dissolution)level (Fig. Ib). However, the difference between the twosystems is the nature of the microsphere matrix, in whichno well-defined wall or envelope exists. Different methodsof encapsulation result, in most cases, in either a micro-capsule or a microsphere. For example, interfacial poly-merization almost always produces a microcapsule,whereas solvent evaporation may result in a microsphereor a microcapsule, depending on the amount of loading.Yet, one can use solvent evaporation twice to create asphere within a sphere, or as we discuss later, novel meth-ods have been designed to use solvent evaporation to cre-ate in one step a double-walled microsphere, which is es-sentially a microcapsule (84).

Classification of Polymeric MembranesPolymeric films and membranes can be classified in vari-ous ways. One such classification is based on porosity, withthe following categories:

1. Macroporous membranes, which have large pores(0. 1-1 jum)

2. Microporous membranes, in which the pores are ap-preciably smaller (100-500 A)

3. Nonporous (gel, solution-diffusion) membranes. Inthe last category the "pores" are of the order of mo-lecular dimensions. They are formed by entangle-ment and/or cross-linking of the molecular chainsand mesh size refers to the space between thesechains.

Release Characteristics of Microspheres and Microcapsules

Release of core material from a nonerodible micro-capsule can occur in several ways. Figure 2 containsfour theoretical curves (A, B, C, and D) that describefour types of release behavior. All curves are plotted aspercent drug released versus time. The mathematicaldescriptions of these release processes have been given(88-90).

Curve A represents the release behavior of a perfect,nonerodible, spherical microcapsule which releases the en-capsulated material by steady-state diffusion through acoating of uniform thickness. The rate of release remainsconstant as long as the internal and external concentra-tions of core material and the concentration gradientthrough the membrane are constant. If a finite time isneeded to establish the initial, constant concentration gra-dient in the capsule wall membrane then there is a timelag in core material release. Curve A in Figure 2 displaysa system with no time lag. If some of the encapsulatedmaterial migrates through the microcapsule membraneduring storage, a burst effect occurs, as represented byCurve B. If the microcapsule acts as an inert matrix par-ticle in which core material is dispersed (a microsphere),the Higuchi model is valid up to 60% release (91). In thiscase, a plot of percent drug released versus square-roottime is linear, as shown by Curve C in Figure 2. First-orderrelease is represented by Curve D. The curve is linear iflog percent core material left in the capsule is plotted ver-sus time.

Actual capsule release data have been plotted for all ofthe described ways and analyzed in terms of the appropri-ate model. Good agreement between actual results and as-sorted model release curves may exist for some portion ofthe release curve, but significant deviations occur often,either during initial states of release or after most of thecore material carried by a capsule sample has been re-leased. Many capsule samples experience an unusuallyrapid rate of release when first immersed in an in vitrorelease medium, i.e., they have a large burst effect. First-order release plots tend to fit a broader range of capsulerelease data than other plots. It is relevant to note thatproper adjustment of the release constant in the first-orderrelease equation can make a first-order release curve ap-proximate, within a few percent, the release curve calcu-lated from the Higuchi equation up to 60% release. If mi-crocapsule release data are linear with square root of timeand also fit a first-order release plot, the expression mostaccurately describing the release behavior can be deter-mined by plotting the total amount of drug released (Q')

Figure 1. Various configurations of(a) microcapsules and (b) micro-spheres.

and 1/Q' versus release rate. For the first-order release,Q' is directly proportional to release rate whereas forsquare root of time release, 1/Q' is directly proportional torelease rate.

The desire to fit capsule release data to various modelsis worthwhile. However, it must be recognized that the ac-tual release kinetics of microcapsules and microspherescan vary greatly from those developed to describe releasefrom macroscopic controlled release delivery devices. Mi-crocapsules are small particles and even small doses con-tain many microcapsules. The release behavior of a micro-capsule formulation is the sum of the release of apopulation of microcapsules (38,39). This population con-sists of individual microcapsules that differ from eachother in the quality of their walls. Some capsule walls aremore permeable than others because of irregularities,which may appear as pits or craters under a scanning elec-

tron microscope. Dappert and Thies (38,39,47) discussedthe release behavior of microcapsule populations. Onecapsule population considered contained microcapsulesthat individually released their contents at a constantrate, but the individual release rate constants fit a log-normal distribution. For such a capsule population, thepredicted release curve will approximate first-order re-lease kinetics.

All microcapsule release behaviors reported in the lit-erature represent microcapsule populations and not indi-vidual microcapsules. Therefore, the successful replicationof microcapsule preparations and confirmation of repro-ducible release rates is an optimal result. The LupronDepo®, a product already in the marketplace, demon-strates that these milestones can be achieved. In the fol-lowing section, we discuss the specific releases as relevantto the specific encapsulation methods.

Gaseous core Solid core Liquid core

Spherical Spherical Irregular Pure or dissolveddrug

Suspension

Irregular Matrix Multi-corn partmental

Emulsion Emulsion-suspension

Gaseous Solid Liquid

Spherical Spherical Solid solution Pure or dissolveddrug

Suspension

Irregular Irregular Emulsion Emulsion-suspension

Figure 2. Theoretical release curves expected for different typesof nonerodible delivery systems. A, Membrane reservoir-type freeof lag time and burst effects; B, same as A, with burst effects;C, matrix or monolithic sphere with square root time-release;D, system with first-order release.

Emulsion Formation

The first step in almost any encapsulation technique (de-scribed in Table 1) involves the formation of an emulsion,usually of a polymeric solution inside a continuous phase(1). Similarly, in order to disperse the drug inside this poly-meric solution (assuming a nonsoluble drug), emulsionsmust be created. Thus, an understanding of the propertiesof emulsions is extremely important. The emulsion for-mation determines the resulting particle size in the finalprocess of encapsulation. An emulsion is achieved by ap-plying mechanical energy which deforms the interface be-tween the two phases to such an extent that droplets form.These droplets are typically large and are subsequentlydisrupted or broken up into smaller ones. The ability todisrupt the larger droplets is a critical step in emulsifica-tion and in encapsulation where an emulsion is prepared.When considering dyes as the main encapsulate, the prob-lems involved at this stage can be minor because most dyesare quite stable. However, drugs (e.g., labile proteins) orcells may be destroyed by the application of mechanicalshear and, if necessary, preventive measures should betaken to stabilize these, even at this stage of the process.

A suitable surfactant is needed to produce a stableemulsion, a result achieved by lowering the surface tension(y, usually from 40 to 5 mN/m"1). For pharmaceutical ap-plications surfactants must acceptable for therapeutic use(92).

Many devices have been designed to produce emulsions.Depending on the desired particle size, a range of devicesfrom simple overhead stirrers or impellers to more sophis-ticated devices such as homogenizers, ultrasonic powergenerators, or ball and roller mills are available. The fol-lowing is a list of techniques and equipment used to formemulsions. Also, an excellent reference is Chapter 2 inBecher (1).

1. Shaking. Shaking is self-evident but in many en-capsulation techniques countertop shakers or agi-tators are used.

Time

2. Pipe flow: laminar and turbulent. Pipe flow may in-clude constriction on the flow baffles on which theliquid can infringe to increase the velocity gradientor turbulence.

3. Injection. The dispersed phase is injected into a con-tinuous phase as a cylindrical jet, where it is brokenup into fairly large droplets.

4. Stirring: simple stirrert rotor-stator, scraper, and vi-brator. Many types of stirring exist, e.g., for pumpsrevolution rates up to 300/s. Rotor-stator machinesexist in great variety and are described under thename of homogenizers. In the laboratory the name"Ultra Turrax" is often used.

5. Colloid mill. This is a rotor-stator device as well,with the exception of having a very narrow slit (e.g.,0.1 mm) and is designed to achieve very high (sim-ple) shear up to 1O7S"1.

6. Ball and roller mills. These are more suited for dis-ruption of solid particles or very viscous emulsiondroplets.

7. High-pressure homogenizer. In the homogenizer theliquid is brought under a high pressure ph (e.g., 10to 40 MPa) by a positive pump and is forced througha narrow (e.g., 0.1 mm) valve slit; owing to the pres-sure, the valve opens against a spring. The poten-tial energy is converted into kinetic energy as theliquid obtains a high velocity u (ph = _u2/2; u is,e.g., 200 m s"1). The kinetic energy is dissipatedinto heat during passage through the valve. Thistakes a very short time (0.1 ms); so energy densityis very high (up to 1012 Wm"3).

8. Ultrasonic: vibrating knife and magneto-striction.Ultrasonic waves can be generated in many ways.A common one is the "liquid whistle" or Pohlmanngenerator. The liquid stream impinges on a knife orblade, which is then brought to vibrate at high fre-quency (6-40 kHz); the liquid is forced through anarrow slit at speeds over 50 m s~ *, which requiresa pressure of some 1 MPa or more. Magneto-striction devices often work at a frequency of 20kHz.

9. Aerosol to liquid: mechanical and electrical. Onemay atomize the dispersed phase in air and let thedroplets be taken up by the continuous phase.

10. Foaming or boiling. Some oils spread over a water-gas interface. If air is beaten in or the water boiled,the thinly spread oil layer is disrupted, and verysmall droplets may result. This may happen as anadditional mechanism during various stirringoperations. Also by steam injection, droplets can bedisrupted to very small ones, though at high cost ofsteam.

Some of the methods are exclusively used in laboratoriessuch as shaking, vibrator stirring, magneto-striction ultra-sonic systems, and aerosol to liquid systems. For large-scale production of emulsions, rotor-stator stirring, col-loidal mills, and high-pressure homogenizers are mostoften used, with some newer processes also using ultra-sonic techniques.

Per

cent

dru

g re

leas

ed

Evaluation of Emulsions

Three types of characteristics are critical in evaluating anemulsion procedure, each with its own methods of evalu-ation (1). They can easily be adapted to encapsulation tech-niques such as solvent evaporation or interfacial polymer-ization. The parameters that should be studied are:

1. Emulsion capacity. The maximum amount of dis-persed phase that can be emulsified under specificconditions without causing aggregation.

2. Emulsion stability. The amount of phase separationtaking place, mostly by the sedimentation rate,which is estimated either under by gravity or by cen-trifugation. The governing theory is that sedimen-tation rate depends on droplet size. This relation,however, may not be so simple because sedimenta-tion rate depends on fluctuation of the droplets andtheir apparent viscosity at very low velocity gradi-ents of the continuous phase. This condition mayalso be affected by the variables studied, leading tomisinterpretation. In microcapsules this first step isimportant because once the emulsion is stabilizedthe next step of hardening the capsules follows.

3. Droplet size. This may be some kind of average orfull distribution. Sometimes the number of drops isdetermined, or some characteristic, depending onthe drop size, such as turbidity under specific con-ditions. Droplet size distribution may change afteremulsification. This can be due to coalescence but italso may be due to isothermal distillation, unless thedispersed phase is completely insoluble in the con-tinuous one. Again, the initial size of the emulsionmay differ from the final size of the microspheres ormicrocapsules.

Determination of droplet size distribution as function ofthe process and product variables is by far the best methodto study emulsification, particularly in microencapsulationwhere the process does not stop at the point of making theemulsion, but continues to the further steps of formationof hard particles. Particle comparison between the originalsize of the emulsion and the final size of the microsphereis beneficial.

Phase Properties of Emulsion: Hydrophilic-Lipophile Balance

One of the important parameters in studying emulsions isthe selection of a surfactant which satisfactorily emulsifiesthe different phases. For this, the hydrophilic-liphophilebalance (HLB) is a useful index. The concept of HLB in itsearly stage was qualitative; however, schemes designed toput this concept on a quantitative basis have been ad-vanced and a new method, which takes into considerationthe temperature and the kind of oil used, has been intro-duced. Around 1950 Griffin [see also Schick (93)] foundthat it was possible to define the polarity for nonionicagents in terms of an empirical quantity which he calledthe HLB. This is represented by an arbitrary scale, inwhich the least hydrophilic materials have low HLB num-

bers and increasing HLB corresponds to increasingly hy-drophilic character. Broadly, these HLB numbers can beused to characterize the applicability of particular surfac-tant. In a given homologous series of surfactants there isa range in which the HLB is optimal for a particular ap-plication. Table 2 lists the HLB ranges suitable for variousapplications.

In many cases the HLB number may be calculated fromcomposition data. For example, for fatty acid esters ofmany polyhydric alcohols,

HLB = 2o(l - | j (1)

where S is the saponification number of the ester and A isthe acid number of the fatty acid. In another example, forglyceryl monostearate, S = 161 and A = 198; hence, itsHLB is 3.8.

For some fatty acid esters it is not practical to obtaingood saponification-number data, e.g., esters of tall oil androsin, beeswax, and lanolin fatty acids. However, for these,the HLB may be calculated from the relation

E + PHLB = = - ^ - (2)

owhere E is the weight percent of oxyethylene content andP is the weight percent of polyol content.

For materials where only ethylene oxide is used to pro-duce the hydrophilic moiety, e.g., fatty-alcohol ethylene ox-ide adducts, this equation reduces to

HLB = E/5 (3)

where E has the same meaning as above.The use of weight percent in these equations is quite

significant, because a better dependence of micellar prop-erties on chain length is found when this quantity is usedrather than the ethylene oxide mole ratio.

Davies devised a method for calculating HLB numbersfor surfactants directly from their formulae using empiri-cally derived group numbers. Thus, a group number is as-signed to various component groups in emulsifiers, e.g.,CH3-CH2-COO-, -CH2CH2O, etc., and the HLB is then cal-culated from the following relation (94):

HLB = 7 + 27(hydrophilic group numbers)- 27(lipophilic group numbers) (4)

For a number of cases, Davies (94) has shown that theHLB numbers calculated from the previous equation and

Table 2. HLB Ranges and Applications

Range Application

3-6 W/O emulsifier7-9 Wetting agent8-15 OAV emulsifier

13-15 Detergent15-18 Solubilizer

Source: From Ref. 93.

those experimentally determined are in satisfactory agree-ment. However, this equation contains the implicitassumption that all ethylene oxide groups make the samehydrophilic contribution, which is manifestly incorrect(94), and the equation fails with increasing poly-(oxyethylene) content.

Additional techniques can be used to determine HLB.These include water titration, algebraic addition of contri-bution from various groups making up the chemical struc-ture from spreading, and by measurements of such prop-erties as dielectric constant, and behavior as a substratein gas-liquid chromatography [see Schick (95), pp. 607-613].

A rough estimate of HLB can frequently be made on thebasis of water solubility or dispersibility. Table 3 shows theranges of HLB numbers indicated by various types of dis-persibility. An up-to-date listing of HLB numbers for alarge range of emulsifying agents (including a few ionicspecies) is given in Schick (95).

A valuable attribute of the HLB scale is that the HLBof mixtures of surfactants can be calculated (to a good firstapproximation, at least) by algebraic addition. For exam-ple, a blend containing 4 parts of Span® 20 and 6 parts ofTween® 60 would have an effective HLB of:

0.4 X 8.6 + 0.6 X 14.9 = 12.3 (5)

It has recently been shown that this algebraic additivityis not strictly obeyed, but the deviation is usually suffi-ciently small so that the system is insensitive to the dis-crepancy. In a few cases, the requirements may be morestringent, but this merely requires final adjustment of thecomposition of the emulsifier.

As a corollary to the previous discussion, each nonaque-ous phase to be emulsified has a required HLB, which is ofcourse different depending on whether water or oil is to bethe continuous phase. The required HLB for a particularoil is usually determined by preparing emulsions over arange of HLB values (obtained, for example, by blendingSpan® and Tween® emulsifiers in various proportions) andobserving the HLB at which maximum stability occurs. Alist of required HLB values for a variety of oil phases isgiven in Table 4.

Using a particular emulsifier pair, the HLB correspond-ing to maximum stability may result, however, in an emul-sion insufficiently stable for the purpose desired. In thiscase, it may be necessary to examine the effect of the chem-ical type of the emulsifying agent, i.e., substituting an ole-ate for a stearate, or an ether-type nonionic for an ester.

Table 3. HLB by Dispersibility

HLB range

No dispersibility in water 1-4Poor dispersion 3-6Milky dispersion after vigorous agitation 6-8Stable milky dispersion 8-10Translucent to clear dispersion 10-13Clear solution 13 +

Source: From Ref. 94.

Table 4. HLB Values Required to Emulsify Various 0:1Phases

In each case, however, the required HLB is equal to (ornearly equal to) the required HLB determined in the initialexperiment.

Geiger et al. (96) studied the behavior of a WfOiW mul-tiple emulsion formulation as potential controlled deliverysystem. The authors were interested in understanding therelease due to a swelling-breakdown phenomenon. Thebreakdown was caused by the water flow from the externalaqueous phase to the internal aqueous phase. Various ex-perimental analyses, such as granulometry, rheology, andconductimetry, as well as a micropipette aspirationmethod, were used to study the stability of W/OAV emul-sions. The predominant role of the lipophilic surfactantduring the swelling phase was confirmed. Two differentmechanisms were proposed and both imply that the mi-gration of the lipophilic surfactant from one interface toanother takes place successively. The lipophilic surfactantcould diffuse from the first to the second interface, thusrigidifying the membrane, or from the oily phase to thefirst interface, resulting in delayed coalescence of the aque-ous droplets during swelling. Thus, the more lipophilic thesurfactant, the more the oil globule capacity increases andthe more the release is delayed.

Polymer Characterization

A broad range of polymers can be used to form microcap-sules. However the list of polymers approved for used in

Oil phase

AcetophenoneAcid, dimerAcid, lauricAcid, linoleicAcid, oleicAcid, ricinoleicAcid, stearicAlcohol, cetylAlcohol, decylAlcohol, laurylAlcohol, tridecylBenzeneCarbon tetrachlorideCastor oilChlorinated paraffinKeroseneLanolin, anhydrousOils

Mineral, aromaticMineral, paraffinic

Mineral spiritsPetrolatumPine oilWaxes

BeeswaxCandelillaCarnubaMicrocrystallineParaffin

Source: From Ref. 93.

W/O emulsion

8

44

4

5

4

OAV emulsion

14141616171617151414141516148

1412

1210147-816

914-15

121010

oral and/or parenteral drug formulations is limited. Thislist includes proteins, polysaccharides, cellulose deriva-tives, synthetic polyesters developed as synthetic suturematerials, poly anhydrides, and polyphosphazene. Char-acterization data for candidate polymer wall materials canplay an important role in the development of a successfuland reproducible encapsulation procedure. Thus, consid-erable attention should be paid to obtaining meaningfulcharacterization data for polymer samples being used toform microcapsules. All the physical properties allow bet-ter fit of individual polymers to the correct encapsulationtechniques. For example, it is not recommended to usepolyanhydrides in solvent evaporation because the poly-mer may degrade during the process of encapsulation.Once the microcapsules are made, a new series of charac-terizations are required because morphological changes oc-cur during encapsulation.

Phase Separation of Polymers

A wide range of polymer phase separation can be used toencapsulate materials: interfacial coacervation, binarypolymer/solvent systems, polymer/solvent/nonsolvent sys-tems, polymer/polymer/sol vent systems, complex coacer-vation, and salting-out.

All these processes provide a way to surround each dis-crete internal phase droplet or particle of drug with ho-mogeneous and reliably concentrated polymeric layerswhich can be later solidified in some manner to form astable capsule. Thus, the next section addresses the gen-eral concepts of phase separation relevant to microencap-sulation.

Interfacial Coacervation. The first example of a suitablephase separation process is the case of polymer adsorption,which Chang (49) has used to prepare small capsules. Inthis process, polymer dissolved in an organic phase mi-grates to the water-organic liquid interface where it pre-cipitates spontaneously from solution to form a membrane.Chang calls this interfacial coacervation. Phase separationoccurs spontaneously. An example of this process is theemulsification of an aqueous protein solution in diethylether using an oil-soluble nonionic surfactant. Span® 85,sorbitan trioleate, is preferred because it does not denaturethe protein. Chang used collodian nitrocellulose as the wallstructure.

The emulsification process gives microcapsules that canbe separated from much of the ether suspending mediumand resuspended in rc-butyl benzoate. ra-Butybenzoate wasselected because it is a poor solvent for collodian, a water-immiscible solvent, and a solvent with a density of 1.0. Thesuspension is left uncovered in order to allow evaporationof ether.

Chang often uses collodian (i.e., nitrocellulose) as thewall material, but stresses that microcapsules (or mem-branes) also can be formed in interfacial coacervation orprecipitation of other polymers using different solvents(e.g., polystyrene in benzene). The success of this proce-dure depends on having a sufficiently high initial proteinconcentration in the aqueous phase. Microencapsulationusually is not possible with a very dilute solution of pro-

tein. Nothing more was noted about the mechanismwhereby the protein (and probably the emulsifying agentin the system) influences the oil/water interfacial tension(}Vw) o r n o w changes in y0/w influences precipitation of theoil phases' polymer.

Phase Separation in Binary Polymer/Solvent Systems. Themost conventional phase equilibria involving nonionicpolymers dissolved in organic solvents have been discussedextensively by Flory (97). Flory's book focused on systemswhere all phases formed are completely liquid in natureand several systems of this nature play a key role in mi-croencapsulation. The simplest case is of a binary systemcontaining polymer and solvent. Figure 3 is a phase dia-gram of this system. Temperature changes alone are usedto induce phase separation. Assuming a monodispersepolymer, phase separation is characterized by formation oftwo distinct phases, one rich in polymer and the other con-taining little polymer. Under equilibrium conditions,chemical potentials of the solvent (Ju1) and polymer (//2) inPhase I equals that in Phase II (i.e., ju\ — //?; /4 = /41)-Furthermore, at some critical temperature (T"c), incipientphase separation will take place:

№) = 0 ; ( ^ ) =0 (6)\ 80% /T,P \dt>2/T,P

and from these relations it can be shown that:

»2C = Ep x i c = o + 1? ( 7 )

1 + yX ^ JX

where t>2 = volume fraction of polymer, v2c = critical vol-ume fraction of polymer, X = number of segments in agiven polymer molecule, and X1C = critical value ofpolymer-solvent interaction free energy. For this simplestcase, theoretical binodials can be calculated and comparedwith experimental values. Such binodials have some im-portant features. First, v2C is predicted to be small, andthis is confirmed by experimentation. In addition, at tem-peratures well below Tc, the dilute phase may retain anunappreciable amount of solute (especially true at higher

Figure 3. Phase separation in a two-phase system (polymer/solvent) induced by temperature.

Polymer concentration

Mw values). However, even the so-called concentratedphase contains a large amount of solvent which means thatthe system is still liquid in nature and thus could flow andengulf the core.

Thus, lowering the temperature (T) of a polymer-solvent mixture at constant D2 and Mw often is a simplemethod of inducing phase separation and thereby obtain-ing a "polymer-rich" phase which can form a capsule wall.This could, in principle, be accomplished at constant T andD2 by increasing Mw (i.e., continuing polymerization of thepolymer). Alternately, at constant T and Mw, D2 could bevaried. Of course, if encapsulation is to occur, the polymer-rich phase must engulf at the internal phase. That is, for-mation of a concentrated polymer-rich phase having a lowtotal volume does not necessarily mean adsorption and en-capsulation occurs. Adsorption of this polymer-rich phaseis a prerequisite for wall deposition (see next section) andmay not be achieved in many cases.

Phase Separation in Polymer/Solvent/Nonsolvent Systems.A second method to produce phase separation involves ad-dition of nonsolvent to a polymer-solvent system (Fig. 4).In such a case one moves from a two- to a three-componentsystem (polymer/sol vent /nonsolvent), thereby grossly com-plicating theoretical calculations. Nevertheless, it still istrue that at equilibrium, chemical potentials of each com-ponent must be equal in the two phases:

A = /i?; A = № = $ (8)

where subscripts 1, 2, and 3 refer to nonsolvent, solvent,and polymer, respectively. If it is assumed:

V1 = V2; X23 = O; X12 = X13 = 1.5 (9)

(V1 = molar volume solvent; V2 = molar volume nonsol-vent; X23 = polymer— solvent interaction free energy; X12

= solvent-nonsolvent interaction free energy; and X13 =polymer-nonsolvent interaction free energy), a specialcase which can be solved is obtained and theoretical bino-

dials can be calculated. Binodials calculated in this man-ner have critical points which occur at low polymer con-centrations. This concentration decreases as Mw increases.As the binodial merges with the solvent-nonsolvent axis,at point D in Figure 4, the polymer concentration in thedilute phase becomes negligible (i.e., approaches zero) ifthe nonsolvent just slightly exceeds that required at thecritical point. Thus, a two-phase system is formed with onephase relatively rich in polymer and the other containingrelatively little polymer and mostly solvent. Provided thepolymer molecules in the polymer-rich phase adsorb at theinternal phase-solution interface, encapsulation occurs. Apossible problem with systems of this type is that the non-solvent could be of such a nature that it is preferentiallyabsorbed by the internal phase and thereby prevents en-capsulation even though phase separation occurs. An ad-ditional important point is that a mixed solvent systemusually cannot be treated like a single solvent system.Generally speaking, the nonsolvent-solvent ratios differmarkedly for the two phases in equilibrium whose com-positions are indicated by the ends of the tie line. Only ifP 12 — /*23 will the solvent compositions in each phase besimilar (they are equal if //13 = //23). A large difference be-tween JUi3

a nd// 2 3 favors absorption of the better solvent inthe polymer-rich phase and this undoubtedly contributesto serious aggregation problems when efforts are made toprepare microcapsules by encapsulation processes involv-ing polymer solvent-nonsolvent mixtures.

Phase Separation in Polymer/Polymer/Solvent System. Athird type of phase separation occurs in systems contain-ing mixtures of two chemically different polymers, as seenin Figure 5. In such cases, formation of true polymer-in-polymer solutions is rarely observed. This is particularlytrue for polymer mixtures free of solvent. In the presenceof solvent, two phases are generally formed under all butdilute solution conditions. The inability of two polymers tomix has led to the general rule of polymer-polymer incom-patibility. This incompatibility has proved to be a problemin many cases where homogeneous mixtures of two poly-

Solvent100%

Solvent100%

100%Nonsolvent

100%Polymer

100%Polymer I

100%Polymer Il

Figure 4. Phase separation diagram for a polymer/solvent/nonsolvent system.

Figure 5. Phase separation diagram for a polymer/polymer/solvent system.

mers are desired (e.g., polymer blends, cases of plastici-zation), but it is a very important phenomena for use asan encapsulation technique; particularly, our group hasused it to make double-walled microspheres in one step(84).

Polymer-polymer incompatibility is a widely encoun-tered phenomenon because the free energy change occur-ring when the two polymers are mixed is generally >0. Thefree energy of mixing (JGm) can be described by:

AGm = ARm - TASm ( 1 Q )

For two polymers: zfSm is small.

JHm generally is >0 (i.e., endothermic mixing). ASm isso small that even if the heat of mixing (per segment) isjust greater than zero, polymer-polymer incompatibilitywould occur. Thus, the widespread nature of incompatibil-ity is attributable to the small contribution of the JS m termand the tendency for polymer-polymer mixing processes tobe endothermic in nature (i.e., ^Hm > O).

It was originally predicted that polymer-polymer in-compatibility is not influenced by the solvent used. Thatis, polymer pairs incompatible in one solvent are incom-patible in all solvents. This often has been found to be thecase, provided that the same molecular weight of polymeris used, but exceptions do exist. Theoretical phase dia-grams developed for specific monodisperse model systemsalso revealed that polymer/polymer/solvent phase sepa-rations are characterized by formation of two phases, eachof which contains essentially only one of the two polymersinvolved. Figure 5 is a phase diagram of such a system.The higher the Mw, the sharper the separation of the twopolymers. These features have been confirmed experimen-tally by numerous workers and by our group. The actualphase separation process is influenced by such factors assolvent, temperature, nature of the two polymers involved,and their initial concentrations and the molecular weights.Complete separation of two polymers is not alwaysachieved by polymer-polymer incompatibility. In mostcases of encapsulation by phase separation, the low molec-ular weight polymers commonly used to induce phase sep-aration become entrapped in the microcapsule wall,thereby affecting properties of the wall. Our experiencewith polymers such as polylactide and polystyrene withMw

in the range of 20,000 to 50,000 always showed completesolubility of 15% WAV, where used with polymer ratiosof 1:1.

The importance of polymer-polymer phase separationin encapsulation is its capability to consistently form a dis-tinct, relatively concentrated polymer phase which, if thepolymer is adsorbed by the internal drug phase, readilywraps it completely thereby forming a capsule.

These methods are theoretically applicable to encap-sulation of both water-soluble and water-insoluble sub-stances. However, when water-insoluble solids or liquidsmust be encapsulated, an additional useful form of poly-mer phase separation phenomena is encountered, complexcoacervation.

Complex Coacervation. Complex coacervation is thespontaneous liquid-liquid phase separation that fre-

quently occurs when solutions of oppositely charged poly-electrolytes are mixed in the same solvent (usually water).It is encountered in biological systems and differs funda-mentally from the polymer-polymer incompatibility men-tioned earlier in that one phase contains most of the twopolymers whereas the second phase is a dilute polymersolution. Figure 6, a phase diagram for the gelatin-gumarabic-water complex coacervation system, illustrates thispoint. Polymer-polymer incompatibility yields two phaseswith each containing predominately one of the two poly-mers.

Significantly, complex coacervation is truly a complexphenomenon. Polyelectrolytes useful for coacervation comein all shapes and forms with varying numbers and typesof charged sites distributed along the polymer chains. Thepolyelectrolytes may have charged sites that are fully ion-ized at all pH values (e.g., SO3) or sites where the group isweakly ionized (e.g., COOH), and the degree of ionizationvaries strongly with pH. Both positively and negativelycharged sites can be on the same molecule (polyampho-lytes). Furthermore, many useful polyelectrolytes are nat-ural gums, biopolymers, etc., and they often (usually) havecomplex structures which are not fully resolved. These fac-tors grossly complicate matters, especially when rigorousanalyses of such processes are desired.

The best known example of complex coacervation is thegelatin-gum arabic (GGA) system studied by Bungenbergde Jong and co-workers (57). Additional studies by Veis(98) have shown that the interaction of two gelatins of dif-fering ionic points (pi = 5.0 and pi = 9.0) also providesan example of complex coacervation. Each polymer pair(under proper pH conditions) interacts to form a complexwhich appears as a concentrated, separated phase (i.e., co-acervate). Temperatures of 4O0C are used in order to obtainthe so-called coacervate in a liquid state. These complexesare not the precipitated complexes formed by interactingoppositely charged polyelectrolytes of high charge density.For proper encapsulation, it is preferred to have liquid co-acervates which can flow around or completely wrap theinternal phase particles or droplets. Of course, the coac-ervate must wet the internal phase (i.e., be adsorbed by it)if it is to form the capsule wall. As mentioned before, a key

Figure 6. Phase separation diagram for a polyelectrolyte system.

100% Pe-100% Pe+

100% H2O

feature of complex coacervation is that the supernatant orequilibrium liquid formed is a dilute polymer solution,whereas the coacervate is a polymer-rich phase (11—15%solids) containing both polymers.

Two factors which affect complex coacervation are pHand neutral salt ions. Because the ionic charge on gelatinand gum arabic varies with pH, coacervation of these spe-cies is sensitive to pH. Bungenberg de Jong (67) found thatcoacervation of alkali-precursor gelatin (pi = 4.8) and gumarabic occurred only between pH 2.0 and 4.8. At pHs out-side this range no coacervate formed due to an imbalancein electrostatic charges. Neutral salt ions decrease the co-acervation tendency of two polymers. This is attributed tothe ability of such ions to screen the charged sites on thepolyions involved and thereby reduce their mutual attrac-tion. For example, Bungenberg de Jong (67) reported that0.02M CaCl2 or 0.035M KCl completely reverses coacer-vation of gelatin and gum arabic at 4O0C. Suppression bysalts is stronger the higher the valence of the added ion.Coacervation tendencies of gelatin and gum arabic also de-crease with increasing initial concentrations of these poly-ions. Such self-suppression of coacervation has been at-tributed to an increase in ionic strength of the system dueto increased diffusible microion concentration. Veis (98) re-ports that the volume of coacervate and fraction of gelatinin the coacervate decreases uniformly with increasing con-centration of gelatin for coacervate systems consisting oftwo isoionic gelatins. The process of coacervation was firsttreated theoretically by Overbeek and Voorn (99) and laterby Veis (98).

lnterfacial Phenomenon

Adsorption of polymers on the core phase is a key step inencapsulation processes based on polymer phase separa-tion phenomena. Thus we discuss were the surface freeenergy (y, dynes/cm or ergs/cm2) effects that occur in suchprocesses.

Consider a case in which a droplet of liquid core mate-rial (representing the active drug) is engulfed by a liquidphase rich in polymer (i.e., the coacervate phase). Figure7 is a schematic diagram of the process. Phase 1 is the dropof core material being engulfed, Phase 3 is the coacervatephase that is engulfing the core material, and Phase 2 isthe continuous (or supernatant) phase in which Phases 1and 3 are dispersed. If Phase 1 is designated such that y12

> y23, the free energy change (AG3) that occurs when Phase3 spreads over Phase 1 is:

AG3 = G a f t e r — ^before Q-J\

AG3 = y13 + y32 - 7i2

Because the spreading coefficient S3 is defined as -AG3:

(12)

Because the core material, coacervate, and supernatantform a three-phase system, the spreading of Phase 1 overPhase 2 and the spreading of Phase 2 over Phase 1 mustalso be considered. The surface free energy change thatoccurs on spreading of Phase 1 over Phase 2 is given by:

AG1 = G a f t e r - Gbefore = }>12 + V13 ~ 723 Q g )

51 = -A1 = J23 - (y12 + y13)

The surface free energy change that occurs on spreading,of Phase 2 over Phase 1 is given by:

AG2 = Gafter ~ ^before = ?12 + ?23 ~ 713 Q 4 )

52 = -AG2 = y13 - (y12 + y23)

Therefore, three spreading coefficients (S1, S2, and S3) de-scribe the equilibrium state of a three-phase system. Thethree S values can have three sets of values (100):

S1 < O, S2 < O, S3 > O; S1 < O, S2 < O, S3

< O; S1 < O, S2 > O, S3 < O (15)

Figure 8 shows the equilibrium configurations that re-sults for each of the above three sets of S values. Of great-est significance to capsule-makers is that complete engulf-ing of core material by the coacervate occurs only when S1

< O, S2 < O, S3 > O, that is, one must have:

51 < O = 723 - (y12 - y13)

52 < O = y13 - (712 + 723) (16)

£3 > O = y12 - (yia + 732)

In encapsulation systems involving polymer phase sep-aration (coacervation), y23 usually is >0.2 dyne/cm. Be-cause y23 is so small, it is reasonable to assume that ^23 «*O. If this is done, and yl2 > y13, S3 is > O, S1 is <0, and S2

is <0. Thus, microcapsules form spontaneously. This is anassumption, and this may be part of the reason why meth-ods such as coacervation may fail to produce microspheres.

Completeengulfing

Partialengulfing

No engulfing

Figure 7. Schematic presentation of a three-phase system.

Figure 8. Schematic diagram showing the spreading coefficient,S, necessary for (a) complete engulfing of an internal phase (1) bya coacervate phase (3) in a continuous phase (2); (b) partial en-gulfing; (c) no engulfing.

But the new method to make double-walled microspheresare more useful.

SOLVENT EVAPORATION

One of the oldest and most widely used methods of micro-sphere preparation is the solvent evaporation technique(18,36,75-80,101). This method yields drug-loaded poly-mer particles that could be called microspheres when drugloading is low or microcapsules when the drug loading ishigh.

The solvent evaporation encapsulation process, de-picted in Figure 9, is a way of precipitating small polymerparticles from an oil-in-water emulsion. The polymer isdissolved in a volatile organic solvent that is immisciblewith water. Methylene chloride is a preferred solvent be-cause of its high volatility (boiling point [bp] 410C) and itscapacity for dissolving a broad range of polymers. Table 5lists a number of solvents that can be used and, it shouldbe noted that, many solvents suitable for this process havea finite degree of water solubility, even though they arenormally classified as water-insoluble solvents. Mixed sol-vents can also be used. The mixtures used so far tend tocontain a water-immiscible solvent (e.g., CH2CI2) and awater-miscible solvent (e.g., acetone). The water-immiscible solvent is the predominant component of themixture.

Once the desired coating polymer is dissolved in the or-ganic solvent, the drug to be encapsulated is added to thissolution. The drug agent may be a solid (crystalline oramorphous) or a nonvolatile liquid. The added drug maycompletely dissolve in the polymer solution or it may becompletely insoluble and simply form a dispersion, sus-pension, or suspension-emulsion. In the latter case, thesolid particles must be micronized so that their mean di-ameter is much less than the desired mean microspheresize. This is true for any encapsulation technique, and gen-erally, a particle size /microsphere size ratio of 1:10 or lessis preferred. The solubility of the drug in the organic sol-vent is also a major factor in determining the morphologyof microspheres produced by the solvent evaporation pro-cess and the final state of the polymer itself (crystalline oramorphous).

The drug/polymer/solvent mixture (i.e., the oil phase)is emulsified in water to form an oil-in-water emulsion.The size of the oil phase droplets obtained is determinedby how rapidly the system is agitated when the oil phaseis added to the aqueous phase, and determines the size ofthe microspheres produced. Emulsification is carried outin a blender if small microspheres are desired (<20//m) orwith a suspended agitator for larger microspheres. In orderto aid emulsification, a surfactant is normally dissolved inthe water phase before the oil-in-water emulsion is formed.A good example is partially hydrolyzed (88%) poly( vinylalcohol) (PVA).

Once the desired oil phase droplet size and emulsionstability have been obtained the system is stirred at a con-stant rate and the solvent evaporates. This is the basis ofthe name, because most of the solvent disappears by evap-oration. Evaporation can occur in an open system at re-duced pressure and range of evaporation temperatures canbe used. Once solvent evaporation appears to be complete,the capsules are separated from the suspending mediumby filtration, washed, and dried. The maximum dryingtemperature must remain below the glass-transition tem-perature of the polymer encapsulant or the microspheresfuse together. Although most of the process depends onevaporation, some of the solvent may diffuse into the aque-ous solution and then evaporate. The amount that diffusesinto the aqueous solution directly depends on the solubilityof the organic solvent in water.

The solvent evaporation process is conceptually simple,but a large number of process variables exist which canprofoundly affect the nature of the product obtained.Table 6 lists a number of such variables. How each of thesevariables influences a given system must be determinedexperimentally, although some general trends are known.For example, semicrystalline polymers often give porousstructures with spherulites on the surface of the micro-spheres. Uniform, pore-free spheres are most readily ob-tained with amorphous polymers. If a polymer is not sol-uble in a single solvent, mixed solvents can be used (e.g.,CH2Cl2/ethanol or CH2Cl2/acetone mixtures).

One requirement of the solvent evaporation process isthat the active agent (i.e., drug) partition favorably intothe oil phase. This partitioning is favored by using active

Add to aqueous solutioncontaining surfactant while

stirring

Polymer dissolvedin organic solvent

Drugparticles Solvent

evaporation

Figure 9. Microencapsulation by solvent evapo-ration. Note that the dark dots represent drug par-ticles; however when the drug is water-soluble, anadditional step is required to prepare a water-in-oil emulsion, which is later encapsulated as shownhere.

Separate and drymicrospheres

Next Page