`PHARMACEUTICAL
`CODEX
`
`--- ·
`
`' 1-.
`
`~ ~
`
`,·Ii
`
`-
`
`•, f
`
`,
`
`. _,• . .
`
`Twelfth Edition
`
`" '
`Principles and Practice of Pharmaceutics
`
`( l ~~·' <
`
`• ;...~; - ~ -
`
`Editor: Walter Lund
`
`~ _-.
`
`-
`
`.
`
`-_·
`
`Published by direction of the Council oftthe Royal Pharmaceutical Society of
`Great Britain and prepared in the Pha/fnaceutics Division of the Society's
`Department of Pharmaceutical Sciences
`
`London
`THE PHARMACEUTICAL PRESS
`1994
`
`<-'-"'-'
`
`-:•
`<i
`
`i.:'<i.-.
`
`1
`
`EX 1011
`IPR of U.S. Pat. No. 7,829,595
`
`
`
`•
`
`ipJE
`1
`
`~we
`or at
`·SU bj
`tees
`:ear
`1elo1
`
`tice
`;t cle
`on tc
`to bE
`:s an
`ibilit
`1aJ p
`; thi.<:
`·and
`
`cove
`tics,
`in es
`ctior
`the>
`nacy
`cl SUJ
`.ates,
`tents
`enta
`·ecol
`mor
`1e Re
`·xper
`
`-::euti
`llics ;
`any
`ndpi
`rage<
`Jre-r,
`harn1
`s-bc
`
`----
`
`AD Aller
`Dr GE A
`MSc, F
`JE Balmf
`Dr A Ble
`MRPh
`Prof TG
`FRPha
`MJS Bur.
`DMS,
`JMW Ca
`JH Carr.
`MIPha
`DL Cole1
`PLM Da
`Dr NJB I
`FRCP.
`
`Copyright I; 1994 hy the Royal Plwrnwcel/ticul Society of Great Britain.
`Copies of this book may be obtained through any good bookseller or, in any case of difficulty,
`direct from the publisher or the publisher's agents:
`
`The Pharmaceutical Press
`(publications division of the Royal Pharmaceutical Societ)
`of Great Britain)
`1 Lambeth High Street. London SE 1 7JN
`
`Al/St ra Ii Cl
`The Australian Pharmaceutical Publishing Co. Ltd.
`40 Burwood Road. Hawthorn, Victoria 3122
`um/
`The Pharmaceutical Society of Australia
`Pharmacy House, PO Box 21. Curtin. ACT 2605
`
`Germany. Austria, Sll'it:erlcmd
`Deutscher Apotheker Verlag
`Postfach 10 10 61. Birkenwaldstrasse 44. D-7000 Stuttgart
`
`Jupun
`Maruzen Co. Ltd.
`3-10 Nihonbashi 2-chome. Chuo-ku, Tokyo 103
`or
`PO Box 5050. Tokyo International, 100-31
`
`Ne11· Zealand
`The Pharmaceutical Society of New Zealand, Pharrracy House,
`124 Dixon Street, PO Box l 1640, Wellington
`
`USA.
`Rittenhouse Book Distributors. Inc.
`51 l Feheley Drive, King of Prussia. Pennsylvania l 9406
`
`A catalogue record for this book is available from the British Library
`
`lnternational Standard Book Number (ISBN): 0 85369 290 4.
`All rights reserved. No part of this publication may be reproduced, stored in a retrieval system,
`or transmitted in any, form or by any means-electronic, mechanical, recording, photocopying,
`or otherwise--without prior written permission from the copyright owner.
`
`Typeset in Great Britain by Alden Multimedia. Northampton
`Printed and bound by The Bath Press, Bath, Avon
`
`* Decease(
`t Deceasec
`
`2
`
`
`
`H
`
`1ci
`
`IlE
`rp
`;h
`en
`·nc
`le'
`
`ac
`la:
`lp
`:d
`m
`stc
`ci
`le
`LC'
`
`t I
`
`aJ
`a
`Ld
`e:
`JI
`
`LS
`I:
`>f
`n
`
`'.ll
`
`n
`e
`r
`,f
`
`Pref ormulation
`
`formulation, with the aim being to highlight th
`Following the identification of a new chemical
`tests that are available and to draw attention t~
`entity that is suitable for developn1ent. the forn1u-
`relevant texts for further reading.
`lator will be called upon
`to produce dosage
`forms. Initially, this may involve the production The order of the sections .below does not represent
`any chronolog1cal order 111 which the experi1nents
`of an injectable fonn suitable for early efficacy
`and toxicity testing and subsequently there will be
`are performed. For example, solubility experi(cid:173)
`n1ents have a higher priority than investigations
`a need to develop the final dosage form, which gen-
`erally will not be an injection. The challenge for the
`of some aspects of crystal form; however, it ;,
`formulator is to develop the initial and final dosage
`easier to describe the solubility of polymorphs
`forms to the highest quality in the shortest time.
`after introducing the concepts of polymorphism
`This process is best achieved when certain physico-
`and that is the basis on which the chapter is
`chemical properties of the drug substance are inves-
`arranged.
`ligated, understood, and effectively utilised: this is
`prefonnulation.
`Prefonnulation studies include investigations of
`chemical form (for example, salts). crystal form
`(for example, polymorphism and habit), solubi(cid:173)
`lity, dissociation (pKa), partitioning, 1nechanical
`properties, stability, and excipient compatibility.
`The number of experiments that could be per(cid:173)
`fonned is extre1nely large and it is necessary to bal(cid:173)
`ance the value of the results against the ti1ne taken
`to obtain thc1n. There is also a need to obtain as
`n1uch infor1nation as possible at the earliest stage.
`Ho\vever. the quantities of drug available will be
`extremely limited until the synthesis has been
`scaled up, and also the changes in the che111ical pro(cid:173)
`duction process n1ay result in changes in the physi(cid:173)
`coche1nical properties of the drug, for cxan1ple,
`different crystal [orms (which is discussed below).
`Thus, there must be careful thought about the
`experiments that will be performed and the stage
`at which they should be considered. The develop(cid:173)
`ment scientist should, if possible. be involved at
`an early stage in the discussion on choice of syn(cid:173)
`thetic route for the bulk production of drug sub(cid:173)
`stance. especially the final crystallisation step.
`Furthennore, the range of tests to be perfonned
`will vary depending upon the desired route of
`adn1inistration and dosage fon11 selected. Thus,
`only for tableted products would one consider
`n1atcrial co1npression properlies. Such co1npres(cid:173)
`sion tests would usually be left to the later stages,
`si1nply on the basis of balancing material supply
`with the requirements of the test.
`ANALYSIS
`As prefonnulation covers such a large range of sub- An essential part of preformulation is lo be able to
`ject areas, the scope of' this chapter has been
`assay the drug. Analysis is a subject that is too large
`restricted to an inlroduction to the concepts of pre-
`to be adequately addressed here: ho\.vever, assays
`178
`
`ORGANOLEPTIC PROPERTIES
`With the Control of Substances Hazardous to
`Health (COSHH) safety regulations it would be
`unusual for preformulation scientists to routinely
`taste new chemical entities. In the following sec(cid:173)
`tion, the need for accurate analysis is discussed·
`however, there are occasions when organolepti~
`aspects provide useful inforn1ation. When an oxi(cid:173)
`dation reaction produces a coloured degradation
`product it will often be detected by the human
`eye before the breakdown product has reached a
`sufficiently high concentration to be detected by
`chemical analysis. Equally, problematic crystal
`transitions can often be detected by simple light
`microscopy, where even qualitative observations
`on particle shape and approxin1ate size distribu(cid:173)
`tion can also be a valuable guide to potential pro(cid:173)
`blen1s
`(for exa111plc, a
`\vide size distribution
`would alert the worker to the possibility of Ost(cid:173)
`vvald ripening in suspensions: acicular shaped crys(cid:173)
`tals lo potential problems with flow). Smell can be
`an efficient 1nethod by vvhich chen1ical and 1nicro(cid:173)
`biological instabilities can be detected. With the
`wide range of erudite techniques available to the
`scientist it is easy to forget that careful organolep·
`tic observations and thorough note taking can be
`valuable tools by which to 111onitor changes in a
`drug and to tnake decisions about subsequent prc(cid:173)
`fonnulation tests, as well as fonnulation and pro·
`cessing factors.
`
`an
`liq
`
`(G
`pfl
`st1i
`a
`ab
`be
`of
`be
`pl<
`th·
`ler
`(fc
`b)
`
`Cl
`Tl
`er
`ar
`va
`m
`ge
`dr
`er
`pl
`
`Pi
`c
`oJ
`w
`se
`pl
`lU
`IV
`o:
`al
`te
`n:
`sl
`Si
`" r;
`fij
`g)
`1
`el
`SI
`p
`a:
`d
`sl
`
`3
`
`
`
`are con1111only- underlaken by high perforn1ance
`liquid chromatography (HPLC). thin layer chro(cid:173)
`(TLC). or gas chromatography
`matography
`(GC), which can allow the drug and degradation
`products or related substances to be monitored (a
`stability-indicating assay). It is also useful to have
`a simple assay (based on an ultraviolet (UV)(cid:173)
`absorbing chromaphore that obeys a Beer-Lam(cid:173)
`bert plot) to allow easy and rapid quantification
`of the preformulation experiments. Users should
`be aware of the litnitations of an assay, for exan1-
`p\e, whether it is stability-indicating or whether
`the breakdown product absorbs at the san1e v..1ave(cid:173)
`length, or if other ingredients in the formulation
`(for example. excipients) interfere with the assay
`by absorbing at the same wavelength.
`
`I
`I
`i
`
`i
`I
`
`CRYSTAL PROPERTIES
`The majority of drug substances are regarded as
`crystalline 111aterials: the n1olecules are packed in
`an ordered and reproducible rnanner. Excipients
`vary fron1 crystalline Inateria\s to ai11orphous poly(cid:173)
`n1ers. However, few san1ples will be entirely ho1no(cid:173)
`geneous, polymers will be partially crystalline and
`drugs at icast partially amorphous. The extent of
`crystallinity of compounds will greatly affect their
`physical properties.
`
`I
`
`Polymorphism
`Changes in the crystallisation process can affect not
`only the degree of crystallinity, but also the way in
`wh!ch the molecules are packed. When a particular
`sohd has been shovvn to exist in n1ore than one
`packing arrangcn1cnt. it is said to exhibit poly(cid:173)
`morphism. There arc two types of polymorpl1ism.
`Monotropic polymorphs are those for which only
`is. stable (irrespective of te1nperature
`one forn1
`and pressure) and the metastable form will revert
`to the stable fonn \Vith tin1c. Enantiotropic poly(cid:173)
`morphs are those for which different forms are
`stable under ditTerenl experin1ental conditions,
`such that a change in pressure or ten1perature
`~~y alter lhc form that is stable.
`iffe1ent n1011otrop1c polyn1orphs often have dif(cid:173)
`ferent n1clting points, vvith the rnost stable forn1
`tenerally having the highest melting point (see
`hermal Methods below). They also exhibit differ-
`ent X ray d'ff
`· f
`·
`d
`1 racl1on patterns and 111 rare
`(IR)
`-
`specl!"\ At
`.
`any one par11cular te111perature and
`'·
`pressure, there will be only one stable polymorph
`alfl other forms that exist for any detectable period
`o tu11e (·1 d I
`'n t 1ere n1av be several) are tenned rncta-
`t
`s able. Metastable p~lymorphs will have a faster
`
`I
`
`l highlight the
`w attention to
`
`~snot represent
`he experin1ents
`ubility experi(cid:173)
`t investigations
`however, it is
`of polymorphs
`polymorphism
`the chapter is
`
`Hazardous to
`ns it would be
`sts to routinely
`' following sec(cid:173)
`;is is discussed;
`:n organoleptic
`. When an oxi·
`·ed degradation
`by the human
`l has reached a
`be detected by
`>len1atic crystal
`by simple light
`ve observations
`te size distribu(cid:173)
`o paten tial pro(cid:173)
`ize distribution
`ssibility of Ost-
`1\.:1r shaped crys(cid:173)
`"1). Smell can be
`1ical and 1nicro(cid:173)
`ected. With the
`available to the
`rcf'ul organolep(cid:173)
`te taking can be
`or changes in a
`subsequent pre(cid:173)
`ulation and pro-
`
`n is to be able to
`t that is too large
`hovJever, assays
`
`Crystal Habit
`
`179
`
`dissolution rate than the stable forn1, and appar(cid:173)
`ently have a greater equilibrium solubility. thus
`the bioavailability from a metastable form can be
`considerably greater than from the stable form of
`that drug. This type of behaviour is due to the
`fact that the melting point is an indication of the
`lattice energy of the crystal, so the most stable crys(cid:173)
`tals will have the largest lattice energy, the highest
`melting point, and the lowest rate of solution. and
`vice \'l"TSa. However, by definition, the n1etastable
`form is not stable and will tend to revert to the
`stable polymorph. Transitions in polymorphic
`forn1 can occur gradually as a function of ti1ne,
`and can be accelerated by changes in storage condi(cid:173)
`tions (such as increases in ten1perature and hu1nid~
`ity) or energetic treatment (processing) of the
`powder. Thus, unit processing such as 1nixing,
`milling, and tableting can cause changes in crystal
`type and consequently change the physical, and
`potentially the biopharmaceutical. properties of a
`drug. It follows that great care must be taken to
`detennine \:vhich poly111orph is present, and under
`what conditions and for how long it will be stable.
`A useful stress test for a drug substance is to ball
`mill it for a defined time and then to eheck for
`any change in polymorphic form. perhaps by use
`of differential scanning calorimetry (sec Thermal
`Methods below).
`
`Pseudopolymorphism
`Changes in crystallisation processes can also result
`in inclusion of inolecules of the solvent in the crys(cid:173)
`tal. producing solvates (or in the unique case vvhere
`water is included, hydrates). These crystals have
`different properties from the non-solvated sample,
`in a si111ilar 1nanncr lo different polyn1orphic
`forms. and are thus often termed 'pseudopoly(cid:173)
`morpbs'. It has been shown that different solvates
`of the same drug can produce different blood con(cid:173)
`centrations following adininistration of a solid oral
`dosage form. However, whereas with polymorphs
`it is the form with the lowest melting point that
`will produce the highest blood concentrations, for
`solvates it can son1ctin1es be the hydrate and for
`other drugs lhe anhydrous form that produces
`the highest concentrations. 1
`
`Crystal Habit
`1--Iabit is the ten11 given to the outward appearance
`of a crystal. It is possible to change polymorphic
`form without altering habit and equally to change
`habit ,.,:vhile 1naintaining the sa1ne poly111orphic
`forn1: the two paran1etcrs arc independent. l--Iabit
`
`4
`
`
`
`180
`
`PrelOrn1ulalion
`
`can be described by variations on the then1e of
`seven syste1ns (cubic. tetragonal, orthorhornbic,
`n1onoclinic. triclinic. trigonal. and hexagonal, see
`I). The phannaceutica\ significance of
`Table
`changes in habit can be an alteration in dissolution
`rate, powder flow, and co1npressibility; thus it can
`influence processing (for exa1nple, flow and com(cid:173)
`pression during tableting2) and use of dosage
`forms. Dissolution rates nre a!Tected by the surface
`to volume ratio, while tenns such as 'needle' (to
`describe the shape of acicular crystals) will indi(cid:173)
`cate those that will have poor flo\V properties.
`
`Table l
`A.11gfes cmd fengrlrs r~f' axes rlrut descri/)e cr.r.11nl lwhi1s
`
`C1y.1·1ttl r s.rnmr_\ 1111 Ang In r>/ /!.\:('.'
`l.englil 11(' uxes 1::.1:c1mplc
`J. = /i = ;· "'90 "= )- = l
`J.=;/i=;·=90 x ""' 'y
`/_
`i=-
`1.=fl=·;=90
`:\-j.y=Fz
`'l. ·""fl= 90 'f. ;· " f: :- * z
`'J_ d:. fi '/=- ;· d:. (j()
`:\ t- )- "'" z
`
`sodium chloride
`nickel sulrbctle
`pO(\\SSilltll
`pcrnmnga11ate
`sucrosi;:
`copper sulphntc
`
`C\1bic (rcguluf)
`Tetragonal
`Orthorhombic
`
`Mo11oclinic
`Tl'iclinic
`(a~ymrnetrkl
`Trig,inal
`(rhombohedral)
`Hc:;.agona!
`
`7. =fl=:·:"' 9[)
`
`x = 'J = 7
`
`sodium nitrate
`
`z al 90
`
`ll' ba~e
`
`silver nitrate
`
`Habit can be altered by changes in the crystallisa(cid:173)
`tion process. The habit is determined by the rate
`of growth of the different faces of the crystal. The
`fastest gro\ving faces will tend to groV'I' out of exis(cid:173)
`tence, and will, therefore. be the smallest foces on
`the final crystal. whereas the slow growing faces
`\Vi\1 do1ninate the final structure. As different faces
`can exhibit different proportions of the functional
`groups that inake up the drug n1olccule. changes
`in
`the crystallising solvent
`tnay preferentially
`favour the interaction with different faces and con(cid:173)
`sequently alter habit. The presence of impurities
`can result in adsorption at certain faces. which in
`turn can prevent (or slow) drug deposition to these
`faces. thus altering the growth rates of the exposed
`areas. Such adaptations can be accidental, due to
`impurities, breakdown products, or synthetic pre(cid:173)
`cursors in the crystallisation 1nixture, or deliberate
`due to the specific addition of i1npurities (such as
`surfactants). A well cited example of this is Lhe
`1nodification of adipic acid crystals by Fairbrother
`and Grant. 3· 4 The preforn1ulation scientist should
`consider the optimum form of habit and, if possi(cid:173)
`ble, influence crystallisation procedures to ensure
`optin1u1n properties arc not due to serendipity,
`but rather a consequence of crystal engineering.
`Crystal habit and n1orphology are best investigated
`by microscopy. Standard light microscopes fitted
`
`with polarising filters and phase contrast facilities
`can allow crystals to be visualised with ease. and
`habit and size to be quantified (for furlher detail
`on size. see Physico1nechanical Properties below).
`
`Crystal Defects
`Bulk crystallisation of drug substances will be
`prone to produce imperfect crystals. The imperfec(cid:173)
`tions will be due lo point defects and dislocations
`during the packing of the latllce. The addition (or
`accidental presence) of low concentrations of
`impurities will increase the disruption in the lat~
`tice. Disruptions in the crystal lattice can result in
`n1ajor changes in the ease of processing. chen1ica\
`reactivity, and dissolution rate, and hence bioavail(cid:173)
`ability.5 York and Grant5· 6 have described a
`method by which it is possible to define a disruption
`index lo quantify the disorder induced by additives
`and/or in1purities in crystals. The disruption index
`can be calculated fro1n differential scanning calori(cid:173)
`n1etry measuretnents or from calorin1etric measure~
`ments of the enthalpy of solution (see Thermal
`Methods for further details of these instruments).
`
`Optical Isomers
`l'
`Historically, atte1npts \Vere not generally n1ade to
`separate optical isomers for tnost compounds.
`However, drugs with one or n1ore chiral centres 1
`i
`are now given special consideration, as the pre~
`j
`ferred ison1cr n1ust be identified and any other iso-
`,~
`n1ers regarded as i1npurities. It is i1nportant to
`identify the correct isomer and to eli1ninate all
`others. Considerable effort is being invested in the
`the
`development of column-based systems for
`separation of iso1ners.
`
`=1
`
`· I
`
`Summary of Tes ts Relating to Crystal Properties
`Standard tests should include melting point (to
`provide an indication of purity and crystal fonn)
`and thennal analysis (see separate section). It is
`usual to obtain a photo1nicrograph or for very
`s1nall particles a scanning electron inicrograph,
`frorr1 which co1nn1ent about habit can be n1ade.
`For the photon1icrograph, ilnage enhancement
`may be utilised (for example, polarised light). The
`tests described under Physicomechanical Proper'
`ties arc also relevant to crystal forn1. The ·aim~
`;ire to identify the polymorph, solvate, habit, and
`optical iso1ner and to tnonitor the changes in prop·
`erties against subsequent batches of drug.
`
`Conclusions
`The crystallisation process will influence the
`vioL1r of the drug, in terms of its physical
`
`i l
`l
`
`che
`phl
`tall
`for
`Th•
`ate
`pre
`tio
`po
`Ph
`en•
`sin
`
`Pl
`Tl
`dr
`fa1
`co
`ar
`ce
`(s
`til
`p
`
`I
`I
`I
`I
`
`I
`
`.I
`
`5
`
`
`
`·asl facilities
`th ease, and
`1rther detail
`ties below).
`
`ices will be
`he imperfec(cid:173)
`dislocations
`addition (or
`1trations of
`1 in the lat(cid:173)
`;an result in
`1g. chemical
`ice bioavail(cid:173)
`:lescribed a
`a disruption
`by additives
`1ption index
`1ning calori(cid:173)
`l"ic n1easure(cid:173)
`ee Thennal
`:;trun1ents).
`
`lly made lo
`:0111pounds.
`iral centres
`as the pre(cid:173)
`y other iso-
`1portant to
`:in1inate all
`ested in the
`ns for the
`
`)roperties
`5 point (to
`ystal rorm)
`;lion). Jt is
`>r for very
`1icro gra phi
`i be 111ade.
`hancen1ent
`light). The
`;al Proper-
`The ain1s
`hc1biL and
`~es in prop-
`1g.
`
`' the bcha(cid:173)
`ysical and
`
`chemical stability, its processability. and its bi6-
`pharn1aceutical perforn1ance. Inappropriate crys(cid:173)
`tallisation procedures can n1ake a successful
`forn1ulation unnecessarily difficult
`to achieve.
`There are precedents for the selection of appropri(cid:173)
`ate crystallisation procedures to yield optimum
`properties for the drug. 7 Subsequent to crystallisa(cid:173)
`tion, the processing of the drug substance will also
`potentially alter
`its physical properties
`(see
`Physicomechanical Properties below). The influ(cid:173)
`ence of physical changes in 1naterials upon proces(cid:173)
`sing and bioavailability cannot be overe111phasised.
`
`PHYSICOMECI-IANICAL PROPERTIES
`The in1portant physicon1echanical properties of a
`drug are its particle size distribution, density, sur(cid:173)
`face area. wettability, hygroscopicity, flow. and
`con1pression properties. Many of these properties
`are influenced or controlled by crystallisation pro(cid:173)
`cedures, however, it is equally true that processing
`(such as 111illing) can affect n1any or all of these cri~
`tical paran1eters.
`
`Particle Size Analysis
`The size of particles can affect processability and
`dissolution rate. For high potency drugs, it can
`be advantageous to have sn1all particle sizes to
`inaxin1ise the nu111ber of particles in order to allow
`adequate 111ixing and dose uniforn1ity. I-lowever,
`static charges can be of greater significance with
`sn1all particles and this can result in aggregation
`and, ironically, poor 111ixing: thus it is in1portant
`to balance the desired properties. Obtaining a par(cid:173)
`ticle size distribution for a povvder is easy, a range
`of instrun1ents being available for the task. How~
`ever, the result lhat is given n1ay require careful
`interpretation and n1ay not be a true reflection of
`the particle size of the pov..1der. For 111any sys(cid:173)
`ten1s, sizing is not as straightforward as it 1nay
`seen1. 'fbe first approach to sizing is to use a n1icro(cid:173)
`scope vvitb a calibrated graticule in order to eSti-
`1nate the size with which to co111pare the result of
`the auto1nated syste1n. Care should be taken with
`the presentation of particle size data, which 111ay,
`for exan1ple, be 'by nu1nbcr' or 'by 111ass'. Micro(cid:173)
`scopy will give a 'by nun1ber' result in which each
`particle is 111casured and the nu111bers in each pre(cid:173)
`d~fined size band arc counted. On a 'by nun1ber'
`distribution, if a san1ple of 500 particles contained
`495 .small particles and 5 large particles, the large
`par1Jcles \vould be 1°/o of the distribution by nu111-
`ber. Other techniques often have data presentation
`'by weight', whereby a sample with many small
`Particles and only a few large particles 1nay, in
`
`Particle Size Analysis
`
`181
`
`fact, have a high percentage of large particles by
`weight (for example. 40%), as one large particle
`contains the equivalent vveight of 111aterial of
`many small particles. The 'by weight' distributions
`are 111ore useful in describing the particle size distri(cid:173)
`bution of phannaceutical systen1s.
`The automated sizing techniques. which have the
`advantage of giving distributions by weight, can
`present problems as they assume spherical parti(cid:173)
`cles and niake no allowance for aggregation.
`Many pharn1aceutical crystals are acicular and
`few are spherical. thus the error can be signifi(cid:173)
`cant, depending upon the aspect ratio of the parti(cid:173)
`in suspension,
`cle. Drugs
`tend
`to aggregate
`especially small particles (with a high electrostatic
`charge density). Aggregation can be worse if the
`particles are not adequately dispersed in the sus(cid:173)
`pending solvent (for example. partially hydropho(cid:173)
`bic solids in \Vater). Surfactants are often used to
`aid dispersion, in which case care niust be taken
`to ensure that the smaller (more rapidly soluble)
`particles are not dissolved in preference to the
`large particles in a solubilising 111ediu111. The rela(cid:173)
`tionship between particle size and the apparent
`solubility of a drug (S,) is obtained from the Gibbs
`Kelvin relation, which for sub1nicron particles
`demonstrates that the true equilibrium solubility
`(that of an infinitely large particle, Sx) of the
`drug can be exceeded due to the high interfacial
`energy that exists between the solid and the liquid
`bsd for very small particles, of radius r:
`
`(s,.)
`
`log Sx
`
`21s1.M
`2.303 RTpr
`
`where Mis the 111olecular weight, R is.the gas con(cid:173)
`stant, Tis the absolute ten1perature, and p is the
`density of the solid.
`There are two popular types of automated sizing
`instrun1ents, the Coulter counter and those that
`use laser light diffraction (for example. Malvern).
`The Coulter counter method involves the disper(cid:173)
`sion of particles in an electrolyte (it can be difficult
`to find a suitable electrolyte for samples that can(cid:173)
`not be dispersed in aqueous liquids). 'fhe instru-
`through a sn1a[J
`1nent draws
`the electrolyte
`orifice, and detects the conductance through the
`liquid. As particles pass through the orifice the con(cid:173)
`ductance is changed and the response is propor(cid:173)
`that passes
`the size of the particle
`tional
`to
`!hrough. 'The approxi111ate powder size 111ust be
`known in order to select an appropriate size of
`type of instru111ent
`tube orifice. The Malvern
`detects diffraction of a laser light. the dispersion
`
`6
`
`
`
`l 82
`
`Preforn1ulation
`
`of which is inversely proportional to the size of the
`particle that passes through the beam.
`A value for a 1nedian size is not an adequate
`description of a sa1nple; so1ne indication of the dis(cid:173)
`tribution around the median is needed. This is
`often obtained by quoting an interquartile range.
`I-laving obtained the size of the 1naterial, the data
`is presented as a distribution. This is convention(cid:173)
`ally done by plotting a cun1ulative percentage over(cid:173)
`size (or undersize) as a function of log of size, which
`will yield a sig1noidal graph that can be linearised
`(to aid the calculation of interquartile range) by
`plotting it on log probability paper.
`
`Surface Area
`One of the reasons for controlling particle size is
`that changes will alter the available surface area
`and consequently affect dissolution and poten(cid:173)
`tially bioavailability. In this respect, size is a crude
`indication of surface area. Most auton1ated sizing
`instru1nents will give a value for the surface area
`of the sa1nple. This is calculated assun1ing that
`the particles are spheres and as this is ahnost
`always untrue the result will be a crude estin1ation
`of the true surface area. A n1ore appropriate test
`for surface area is by a BET (Brunauer, En1n1ett,
`and Teller) gas adsorption experitnent. 8 It is usual
`to adsorb nitrogen gas onto a packed coltunn of
`powder. This is done by passing nitrogen over the
`sa1nple and detecting the output, an<l then quench(cid:173)
`ing the sample in liquid nitrogen, such that the
`nitrogen gas inside the colun1n will adsorb onto
`the available surface; fro1n the reduced output of
`nitrogen gas the a1nount adsorbed is assessed and
`fron1 the cross-sectional area of a nitrogen n1ole(cid:173)
`cule the surface area of the solid is estirnated using
`the BE1~ isotherm equation. Sin1ilar experin1ents
`can be undertaken using krypton (giving i1nprovecl
`accuracy) and son1e workers utilise water sorption
`to provide an indication of the effective surface
`area available to water tnolecules.
`
`Wetting
`The degree of wetting of a solid by a liquid can con(cid:173)
`trol its effective surface area. A finely divided pow(cid:173)
`der (high surface area) that is poorly wetted will
`have a li1nited interface with the liquid, because
`the powder will tend to aggrega te9 and air will be
`trapped at the powder surface preventing contact
`with the liquid. Wells 10 states that wetting is an
`in1portant factor in controlling solubility and can
`be influenced by changes in salt form and by recrys(cid:173)
`tallising to change habit. Wetting is indeed influ(cid:173)
`enced by chcn1ical variation of the n1olecule and
`
`Wl
`for
`val
`tW•
`un
`int
`al~
`(s<
`111'
`pe
`th
`la
`fc
`Cl
`0
`
`I
`l
`
`as dif-Terent faces of a crystal can express different
`proportions of the functional groups that 1nake
`up the molecule, wetting will also be influenced
`(lo some extent) by crystallisation. However, a sig(cid:173)
`nificant factor that affects wetting is that of physi(cid:173)
`cal processing. Buckton et afl I have den1onstrated
`that significantly different wettabilities can be
`observed by changes in the processes used to mill
`a drug. These changes are undoubtedly a function
`of disordering the outer molecules of the crystal,
`possibly causing changes in the degree of crystalli(cid:173)
`nity of the surface. Thus, the duration of changes in
`wettability will be linked to the tendency for the
`surface to recrystallise (for example, humidity
`and storage ten1perature) and the crystal forn1
`that is produced upon surface recrystallisation.
`There is no doubt that changes in the physical pro(cid:173)
`cessing of n1aterials can cause sufficiently large
`changes in properties to completely alter the subse(cid:173)
`quent processing or the functional performance of
`the product.
`The use of wettability data to predict prodnct per(cid:173)
`forn1ance has been a central issue of a number of
`publications. 9· 12- 15 The concept is that during pre(cid:173)
`formulation studies, the wettability of the drug is
`assessed and this is utilised to estitnate the surface
`energy and polarity. These values are then co1n(cid:173)
`parecl with a library of data for con1111only used
`excipients and the n1ost logical choice of adjuvantS
`is then 1nade, rather than resorting to a trial and
`error approach. This concept forn1s the starting
`point for
`'expert systems', whereby computer(cid:173)
`aided formulation should allow the development
`of an initial fonnulation on the basis of physico-
`. chemical characteristics of the drug (preforrnulalion).
`
`Deter11li11ing Values j'or Contact Angle and Surj'ace
`Energies
`While it is possible to assess wettability by use of
`calorimetric methods, 16 the most utilised method
`1
`is to cletenninc the value of a contact angle. For
`·\
`powdered systems this is not straightforward and
`· l
`Lhe available techniques are all fla\ved. 16 Most 1
`workers use an approach whereby a drop of liquid
`
`.1
`
`is placed upon the surface of a prcsaturated powder l
`
`con1pact and the angle is either tneasured directly
`:
`(difficult due to dynamic effects), photographed <
`.ii
`and measured, or the maximum height of liquid
`drop 1neasurecl fron1 \.vhich the contact angle can
`-.
`be calculated. 17 A contact angle (e) is formed due
`,
`J
`to a balance of the interfacial forces between the
`J
`solid/liquid (-ysL), the liquid/vapour (-yLy), which
`is the surface tension of the liquid, and the solid/
`'
`vapour (-rsv ), and is expressed by Young's equation:
`
`7
`
`
`
`'SS different
`that lnake
`: influenced
`~ever, a sigR
`.at of physi(cid:173)
`m1onstrated
`ies can be
`used to mill
`ya function
`the crystal,
`of crystalli(cid:173)
`if changes in
`ency for the
`e, humidity
`:rystal form
`ystallisation.
`ohysical pro(cid:173)
`ciently large
`:er the su bse(cid:173)
`rf or1nance of
`
`product per(cid:173)
`a nun1ber of
`lt during prc(cid:173)
`>f the drug is
`te the surface
`re then con1-
`11111only used
`: of adjuvants
`:o a trial and
`; the starting
`>Y con1puter(cid:173)
`development
`is of physico(cid:173)
`~fonnulation).
`
`1e and Su1:face
`
`ility by use of
`:ilised method
`ict angle. For
`1tforward and
`awed.16 Most
`drop of liquid
`urated powder
`isured directly
`photographed
`eight of liquid
`.tact angle can
`is formed due
`cs between the
`r (~ ), which
`·ct/
`1LV
`and the soh
`ung's equation:
`
`'ILV cos 8 = 1sv -1sL
`While it is impossible to measure directly the terms
`for solid interfacial energy, it is possible to estimate
`values for then1 by n1easuring contact angles for
`two liquids of known surface tension on the
`unknown solid. If the surface tension is divided
`into two con1ponents (polar and dispersion), it is
`also possible to calculate the polarity of the solid
`(see Zografi and Tam 18 for detail of calculation
`method). Having determined the polar (p) and dis(cid:173)
`persion (d) components for the surface energy of
`the solid, it is possible to compare these with simi(cid:173)
`lat values for any other potential con1ponent of the
`formulation by calculation of a spreading coeffi(cid:173)
`cient (.\), which is shown here for the general case
`of phase 1 spreading over phase 2:
`
`~]
`
`A positive value for the spreading coefficient of
`phase 1 over phase 2 will indicate that phase 1
`will indeed spread over phase 2; such a situation
`may be desirable in the example of a binder for
`use in wet granulation. Thus, having detern1ined
`the surface energy and polarity of the drug, the cor(cid:173)
`rect binder can be selected before attetnpting to
`make the formulation. The value of spreading coef