International Journal of Pharmaceutics 148 (I99?) I
`
`2]
`
`The use of solubility parameters in pharmaceutical dosage form
`design
`
`Invited review
`
`Bruno C. Hancock “'*_, Peter York 5, Raymond C. Rowe ”
`
`"Merck Frosst Cctrtada Incorporated, Pharmaceutical R&D Department,
`Quebec HQH 31. l, Camrda
`b School qf fharmacy. University of Bradford. Bradford. UK
`°Zetteca Pharmaceuticals. Maccl'esfiet'd, Cheshire. UK
`
`t'I57ll Trattscanada Highi-m_v. Kirktamt.
`
`Received 15 October 1996: received in revised fonn 2] November 1996: accepted 28 November 1996
`
`Abstract
`
`The use and potential of solubility parameters for pharmaceutical dosage form design are reviewed in this paper.
`Specific reference is given to the development of the approach, its previous usage and likely future applications. The
`advantages, assumptions and limitations of this type of approach are also described. (3 199'? Elsevier Science B.V.
`
`KE_}’1v(JFd5'I Cohesive energy density: Solubility parameter; Interaction; Polarity; Formulation
`
`1. Introduction
`
`The rational design of pharmaceutical dosage
`
`forms results from a clear understanding of: (i)
`the chemical and physical properties of the dosage
`form components and (ii) their potential to inter-
`act with each other and the environments to
`
`which they are exposed. Such material properties
`
`and
`
`subsequent
`
`interactions
`
`can
`
`be
`
`readily
`
`* Corresponding author. Tel; + I 514 4283342; fax: + I
`514 4282677; e—mail: bruno_ hancoclc@merck.com
`
`estimated from a knowledge of the solubility
`parameters (or cohesive energy densities (CED))
`of the formulation components.
`
`2. Background
`
`The cohesive energy of a material is the energy
`which holds that substance together.
`It
`is the
`amount of energy required to separate the con-
`stituent atoms or molecules of the material to an
`
`infinite distance. and hence it is a direct measure
`
`0378-5|'t'3;97_t$lTr'.00 (0 I99? Elsevier Science B,V. All rights reserved.
`PH S03't'3—5l't'3[96)04828-4
`
`MYLAN PHARMS. INC. EXHIBIT 1049 PAGE 1
`
`MYLAN PHARMS. INC. EXHIBIT 1049 PAGE 1
`
`

`
`2
`
`B. C. Hanr.-cad: at at’. flrirerriatiorial Journal of P!iormat'em‘it’s 143 (I997) I -2.’
`
`of the attraction that its atoms or molecules have
`
`for one another. Cohesive energy is the net effect
`of all the inter atomic/molecular interactions in-
`cluding Van der Waals
`interactions, covalent
`
`bonds, ionic bonds, hydrogen bonds, electrostatic
`interactions, induced dipole and permanent dipole
`interactions. An understanding of cohesive ener-
`gies is important to the materials scientist because
`
`they can be used to explain or predict how sub-
`stances will behave when they are subjected to
`external stresses, such as heat, light or mechanical
`forces. Cohesive energies are especially important
`to the pharmaceutical materials scientist because
`
`they determine many of the critical physico-chem-
`ical properties (e.g. solubility, melting point) of
`drugs and excipients. A thorough understanding
`of cohesive energies can increase our awareness of
`
`how pharmaceutical materials will behave when
`processed or when dosed into the human body.
`The cohesive energy of a material can be
`quantified in a number of ways. The most com-
`mon approach is to use the so—called solubility
`parameter
`(:5)
`(Hildebrand and Scott,
`1950;
`Hansen, 1969; Barton,
`1983;
`1985). Solubility
`parameter theory was developed by Hildebrand
`and co-workers
`(Hildebrand and Scott, 1950)
`based on regular solution theory. According to
`their approach when two materials are mixed
`together the heat of mixing (AH) is given by:
`
`AH = VTi(ZlEv|.fVmil0'5 — (/lEv2/ Vm2)0'5i2‘¢"I ‘ $2
`(1)
`
`where VT is the total volume, AEV is the energy of
`vapourisation, V,“
`the molar volume,
`96
`is the
`volume fraction, and l and 2 refer to the solvent
`
`and solute components, respectively. The solubil-
`ity parameter of each component is defined as the
`square root of its CED, measured as the energy of
`vapourisation per unit volume:
`
`5 = (CED)°'5 = (AEvfVm)”"‘
`
`(2)
`
`When the solubility parameters of two materials
`are similar Eq. (1) predicts they will be mutually
`and athermally soluble. The units of the solubility
`parameter are (J/m3)“, MPa°‘5 or
`(cal/cm3_}”-5,
`and one (cal/cm3)°-5 is equivalent to 2.0421 MP3”
`or (.Ifm3)"'5’.
`
`The concept of solubility parameters was origi-
`nally developed for simple liquid mixtures and in
`order to extend the principles to consider more
`complex situations several approximations and
`assumptions are required. Typically gases are
`treated as hypothetical
`liquids whilst solids are
`treated as supercooled liquids. With these as-
`sumptions
`it
`is possible
`to apply solubility
`parameter theories to ideal gases, and to organic
`solids with a low level of crystallinity. Regular
`solution theory, upon which the concept of solu-
`bility parameters is based, also applies best
`to
`non-polar molecules which interact through weak
`dispersion forces. Several methods have been pro-
`posed to extend solubility parameter concepts to
`the more polar strongly interacting species which
`are typical of pharmaceutical materials. Various
`authors (Hansen, 1967a,b. 1969; Karger et al.,
`1978)
`have
`sub-divided
`the
`total
`solubility
`parameter (5,)
`(also known as the Hildebrand
`solubility parameter) into components which ex-
`press the contributions from the different types of
`interatomicfintermolecular forces (e.g. hydrogen
`bonds (6,), dispersion forces (rid), ‘polar’ interac-
`tions Mp):
`
`6§=a'§+aIf,+d§
`
`{3}
`
`This approach allows a more detailed characteri-
`sation of the system of interest. It also permits the
`calculation of the polarity of a material
`(Xp)
`(Zografi and Tam, 1976):
`
`X, = af,,n5§
`
`(4)
`
`This parameter provides insight into the balance
`of polar and non-polar forces operating between
`adjacent atoms/molecules and between material
`surfaces. An
`alternative
`‘extended
`solubility
`parameter’ theory has been developed by Martin
`and co-workers (Adjei et al., 1980; Martin et al.,
`l980, 1981) in order to describe the solubility of
`crystalline solids in both polar and non-polar
`liquids. These authors used an interaction pa-
`rameter to account for specific solute—solvent in-
`teractions.
`In the case of a perfectly regular
`-solution this interaction parameter equals one.
`When there is attraction between the solute and
`
`the parameter is greater than unity and
`solvent
`when there is self association by either component
`
`MYLAN PHARMS. INC. EXHIBIT 1049 PAGE 2
`
`MYLAN PHARMS. INC. EXHIBIT 1049 PAGE 2
`
`

`
`B.C. Hancock et al. /International Journal of Pharmaceutics M8 {I997} l—2l
`
`Table l
`
`Solubility parameters and fractional polarities of some drugs
`
`Material
`
`Sol. param.
`(MPaI}.fi)
`
`Polarity
`
`Method
`
`Reference
`
`Aspirin
`
`24.1-24.9
`
`0.29
`
`Calculated
`
`Barbital
`Benzoeaine
`Benzoic acid
`
`22.6
`31.7
`23.5-24.3
`
`-
`—
`-—---
`
`—
`0.60
`
`0.49
`
`—
`
`Solubility
`Solubility
`Solubility, calculated
`
`Calculated
`Inverse gas chromatography
`
`Calculated
`
`Solubility
`
`Samaha and Naggar. 1990;
`Roberts et al..
`[991
`Khalil and Martin. 1967
`Most. I972
`Chertkoff and Martin, 1960;
`Samaha and Naggar. 1990
`Samaha and Naggar. 1988
`l-luu-Phuoc et al.. 1987: Rowe, 198921
`
`Ticehursl. I994
`
`Adjci ct al.. I980
`
`23.3--28.?‘
`
`0.16-0.59
`
`31.2 33.2
`22.4-- 22.6
`37.4
`
`0.6] -0.65
`0.23-0.25
`0.72
`
`Partition, solubility, inverse gas
`chromatography. calorimetry
`Inverse gas chromatography
`Calculated
`Inverse gas chromotography
`
`Rey-Mermet et al..
`
`l99|
`
`Ticehurst, I994
`Tiochurst. 1994
`Egawa et al.. [992
`
`Betamethasone
`Caffeine
`(anhydrous)
`Caffeine
`(anhydrous)
`Caffeine
`{anhydrous}
`
`Caffeine
`(anhydrous)
`Carbamezapine
`Carbarnezapine
`Ccphalexin
`(20.8% crys-
`tallinel
`
`Ccphalexin
`(36.7% crys-
`talline)
`
`Cephalexin
`(88.6% Crys-
`talline)
`
`Ccphalexin
`(freeze dried}
`
`Cephalexin
`Ethinamale
`Griseofulvin
`Hydrocortisone
`Hydrocortisone
`acetate
`
`Ibuprofen
`Indomethacin
`Norethindronc
`derivatives
`
`Paracetamol
`Phenacetin
`Phenobarbital
`Phcnylbutazone
`
`Propanolol
`hydrochloride
`
`Propanolol
`hydrochloride
`Salicylamide
`Salicylic acid
`Steroids
`
`24.5
`26.6
`
`28.0
`
`28.2
`
`38.0
`
`27.0
`
`31.4
`
`22.4
`28.2
`21.3
`25.3
`23.?
`
`20.4
`25.2
`19.3 -22.2
`
`26.2
`23.6
`25.6
`22.9 -22.3
`
`24.4
`
`35.5
`
`31.3
`22.]
`I’.-‘.2 25.3
`
`0.72
`
`Inverse gas chromotography
`
`Egawa et al., [992
`
`0.60
`
`Inverse gas chrorriotography
`
`Egavva ct al.. I992
`
`Inverse gas chromotography
`
`Egawa et 21]., I992
`
`Calculated
`Calculated
`Calculated
`Calculated
`Calculated
`
`Calculated
`Calculated
`Solubility
`
`0.41
`—-
`0.32
`0.19-0.50
`
`Calculated
`Calculated
`Calculated
`Solubility. calorimetry
`
`Calculated
`
`Ticehursl, I994
`Samaha and Naggar. I990
`Samaha and Nagar, I990
`Samaha and Naggar, I990
`Samaha and Naggar, I990
`
`Roberts et al.. 1994
`Samaha and Naggar. I990
`Lewis and Enever. I979
`
`Ticehurst. I994
`Samaha and Naggar, 1990
`Rowe.
`l989b
`Sarnaha and Naggar, I983:
`Rey-Mermet et al.. I991
`Ticehurst, I994
`
`Inverse gas chromatography
`
`Ticehurst. 1994
`
`Calculated
`Solubility
`Calculated
`
`Roberts el al.. I994
`Khalil and Martin, 1967
`Michaels et al.. 1975'.
`Samaha and Naggar, I988
`
`0.6l
`
`0.30
`—
`-—
`—
`—
`
`l].i4
`—-
`_
`
`0.22
`
`0.68
`
`—
`—
`—
`
`MYLAN PHARMS. INC. EXHIBIT 1049 PAGE 3
`
`MYLAN PHARMS. INC. EXHIBIT 1049 PAGE 3
`
`

`
`4
`
`B.C. Hancock er of. /International Journal of Pharmaceutics I48 (.1997) l -21
`
`Table I (continued)
`
`Material
`
`Sol. param.
`(MPaD.5)
`
`Sulphonarnides
`
`20 28
`
`Polarity
`
`Method
`
`Reference
`
`-—
`
`0.41
`
`-—
`
`0.45
`
`Solubility
`
`Solubility
`
`Solubility
`
`Samaha and Naggar, I988:
`Bustamante ct al., 1993a
`James et al., 1976; Rowe, 1989a
`
`Martin et al., 1980
`
`Inverse gas chromotography
`
`1-luu-Phuoc el al.. 198'.-'; Rowe. 1989a
`
`19.4
`
`28.5
`
`28.6
`
`Testosterone
`proprionate
`Theophylline
`(anhydrous)
`Theophyllinc
`(anhydrous)
`Theophyllinc
`(anhydrous)
`Theophylline
`(anhydrous)
`Tolbutarnide
`
`29.8,24.4
`
`0.36. 0.53
`
`Solubility, calorimetry
`
`Rey—Mern1et et al., 199!
`
`27.4
`
`22.0
`
`0.50
`
`—
`
`Calculated
`
`Calculated
`
`Tieehurst, 1994
`
`Samaha and Nagar, I988
`
`then the parameter is less than one. This ap-
`proach can be used to describe almost any so-
`lute—solvent
`system but
`it has very limited
`predictive capabilities.
`There have been many detailed reviews of the
`development of solubility parameters over the
`past 40 years and the reader is referred to these
`for
`further background information (Hansen,
`1969; Barton, 1983, 1985). In the remainder of
`this paper the use of solubility parameters specifi-
`cally for the design of pharmaceutical dosage
`forms is described. The methods suitable for de-
`
`termining the solubility parameters of pharma-
`ceutical materials
`are
`first
`reviewed,
`then
`
`examples of the properties and interactions that
`can be predicted from solubility parameters are
`given. Finally the advantages and limitations of
`using a solubility parameter approach for phar-
`maceutical dosage form design are outlined.
`
`3. Determination of solubility parameters of
`pharmaceutical materials
`
`Of all the direct and indirect methods available
`
`for determining solubility parameters many are
`
`suitable for use with pharmaceutical materials
`(Tables 1 and 2). Different methods give slightly
`different results (Barton, 1983; Rey-Merrnet et
`al., 1991) and the best methods to choose are
`those which most closely represent the in-use situ-
`
`ation of the material(s) under consideration. The
`level of variation seen between different methods
`
`is illustrated for three typical pharmaceutical ma-
`teriais in Tables 3 and 4. Variations in both the
`
`total solubility parameter and the fractional po-
`larity of pharmaceutical materials are common.
`By definition the solubility parameter ((5) of a
`material
`is linked to its heat of vapourisation
`(-'1lHv)1
`
`:5 = <cED)°-5 = (AK./V..)“—5 = «AH. - Rm V...)‘”
`(5)
`
`For materials which are stable above their boiiing
`points the heat of vapourisation can be directly
`determined. However, this method only provides
`the total solubility parameter, and it
`is often
`unsuitable for drugs and excipients because of
`thermal instabilities. The heat of vapourisation of
`pharmaceutical
`liquids can be indirectly deter-
`mined from their vapour pressure using the Clau-
`sius—Clapeyron equation (Sunwoo and Eisen,
`1971) or from their boiling points using an empir-
`ical equation (Vaughan, 1985; Lin, 1992).
`Several group contribution methods have been
`developed for calculating solubility parameters
`(Van Krevelen and Hoftyzer, 1976). This ap-
`proach requires a knowledge of the chemical
`structure of the material, and this is normally
`available for pharmaceutical substances (Table 5).
`Such an approach is especially useful at the start
`of the pharmaceutical development process as it
`
`MYLAN PHARMS. INC. EXHIBIT 1049 PAGE 4
`
`MYLAN PHARMS. INC. EXHIBIT 1049 PAGE 4
`
`

`
`B.C. Hancock er al. _.»"l'mernan'onal Journal ofPharma::ear1'rs 148 (1997) 1-21
`
`Table 2
`
`Solubility parameters and fractional polarities of some pharmaceutical solvents, cxcipients and packaging materials
`
`Material
`
`Sol. param.
`(MPaE).S}
`
`Polarity
`
`Method
`
`Reference
`
`21.3-21.5
`19.8-20.3
`23.9-24.3
`24.5
`25.3
`23.5
`25.]
`25.3
`21.6
`27.8
`18.2-18.4
`25.7
`36.2
`56.2
`
`30.2
`
`39.3
`
`0.55
`0.41
`0.61
`—
`—
`—
`—
`—
`"r
`0.42
`—
`0.76
`0.69
`0.96
`
`0.73
`
`0.76
`
`19.6 47.9
`
`0.25-0.93
`
`21.7-212
`
`—
`
`——
`0.12
`—
`0.00
`
`0.23
`
`0.26
`
`0.29
`
`0.17
`
`0.52
`
`-
`—
`—
`Calculated
`Viscosity
`—
`Calculated
`Calculated
`—
`—
`—.
`—
`Calculated
`Viscosity,
`swelling
`Calculated,
`modulus
`
`lnversc gas
`chromatogra-
`phy
`Viscosity,
`solubility
`Calculated
`
`Calculated
`
`Calculated
`—
`
`Vaughan, I985: Bochek and Petropavlovslty, I993
`Grulke, I975; Vaughan, I985: Suga and Takahama, I996
`Grulke, 1975; Vaughan, I985
`Cowie. 1965
`Cowie, 1965
`Vaughan, [985
`Vaughan, 1985
`Vaughan, 1985
`Vaughan, 1985
`Hansen, 1967b
`Vaughan, 1985: King. 1995
`Grulke.
`l9'r'5
`Bochel-: and Petropavlovsky. I993
`Bochel-: and Petropavlovsky. I993
`
`Roberts and Rowe, 1993
`
`Huu-Phuoc ct al.. 1987
`
`Archer, 1992: Bochck and Petropavlovsky, I993
`
`Sakellariou et al., 1986
`
`Vaughan, I985
`Grulkc, I975
`Vaughan, I985
`Grulke, 1975
`Vaughan, 1985
`
`—
`
`—
`
`_
`
`Grulke. 1975; Vaughan, I985;
`Rasmussen and Walmstrom, I994
`Kent and Rowe. I978: Grulke, I925
`
`Vaughan, [985
`Grulke. 1975: Kent and Rowe, I978
`
`Grulke, I975; Vaughan, I935
`
`Grulke, I975; Vaughan, I985
`
`Acetic acid
`Acetone
`Acetonitrile
`Amylase
`Amylase
`Benzoic acid
`Benzyl alcohol
`B1-IA
`Butylparabcn
`Carbon black
`Castor oil
`Cellulose
`Cellulose
`Cellulose
`
`Cellulose
`(mic:-ocrys—
`talline}
`Cellulose
`lmicr0crys-
`lallinc}
`Cellulose
`acetate
`Cellulose
`acetate
`
`phthalate
`Cetyl alcohol
`Chloroform
`Cholesterol
`Cyclohcxane
`D and C Red
`No. 22
`[Eosinl
`Dibutyl
`phthalatc
`Diethyl
`phthalale
`Dimethicone
`Dimcthyl
`phlhalate
`Dioctyl
`phthalate
`Dimcthy|-
`sulloxide
`
`Ethanol
`Ethyl acetate
`lithylcellulose
`
`18.3
`19.0
`19.5
`16.8
`22.8
`
`19.0-20.2
`
`20.5
`
`12.]
`21,9-22.1
`
`|8.2
`
`24.6-27.4
`
`25.6 26.5
`18.6-18.8
`20.6
`
`Ethylene glycol
`Freon 12
`Gelatin
`Glycerol
`
`29.6
`||.3
`24.5
`33.2-47.]
`
`0.64
`0.25
`0.34
`
`—
`
`—
`Viscosity,
`solubility
`—
`—
`Swelling
`0.77-0.86 —
`
`Vaughan, 1985; Bochek and Petropavlovsky, 1993
`Grulkc, 1975; Vaughan. I983
`Kent and Rowe. I928: Archer. I992
`
`Vaughan, I985
`Grulke, 1975
`Bajpai, I996
`Grullte, 19".-'5; Lewis and Enever, 1979; Vaughan, I985:
`Bustamante ct al.,
`l993b
`
`MYLAN PHARMS. INC. EXHIBIT 1049 PAGE 5
`
`MYLAN PHARMS. INC. EXHIBIT 1049 PAGE 5
`
`

`
`B.C'. Hancock at al. /International‘ Journal of Pharmaceutics 148 (1997) 1-27
`
`Polarity
`
`Method
`
`Reference
`
`0.8—0.8l
`
`Calculated
`
`Choi et al., 1994
`
`Table 2 {continued}
`
`Material
`
`Sol. param.
`[MPau.S)
`
`Hydroxyethylce1— 25,5-19.8
`lulose
`
`HydroxyethylceI— 31.0 29.2
`lulose
`I-1ydroxypropyl— 22.1-20.8
`cellulose
`
`23.7- 22.1
`
`Hydroxypropyl-
`cellulose
`1-lydroxypropyl 25.5-22.1
`cellulose
`
`Hydroxypropyl- 22.8-30.6
`methylcellulose
`HPMCP
`
`26.4-- 17.2
`
`Iron oxide (red) 28.0
`Isopropanol
`23.0
`Lactic acid
`30.2
`Lactose
`33.2
`(anhydrous)
`Lactose
`(anhydrous)
`Lactose (mono-
`hydrate)
`Lauric acid
`Lauryl alcohol
`Magnesium
`stearate
`
`36.3
`
`17.]
`19.4
`18.2
`
`36.0-39.9
`
`29.3-29.7
`Methanol
`Methyl cellulose 21.3
`Methylene
`19.5
`chloride
`
`Methyl paraben 24.5
`Mineral oil
`14.5
`n—octanol
`17.8 21.3
`
`rt-propanol
`N-methyl
`pyrrolidone
`Nylon 6,6
`Oleic acid
`Palmitit: acid
`Petrolatum
`Polyethylene
`Polyethylene
`oxide
`
`24.3
`23.1
`
`22.9-27.8
`15.8-16.1
`161
`15.0
`17.6
`34.7
`
`18.0 26.1
`
`Polyethylene
`glycol
`Polyoxyethylated 24.9-29.6
`ethers
`
`Polyoxyethylatecl 25.1-27.0
`nonyl phenols
`
`Choi et al., 1994
`
`Choi et al., 1994
`
`Choi et al.. 1994
`
`Roberts and Thomas. 1978; Choi et al., 1994
`
`Rowe. 1983b;
`
`l989a,b; Archer. 1992
`
`Sakcllariou ct al.. 1986
`
`Hansen, 1967!:
`Vaughan, 1985
`Vaughan, 1985
`Ticehursl. 1994
`
`0.70 0.43 Molecular
`modelling
`Calculated
`
`0.76-0.72
`
`0.71-0.58 Molecular
`modelling
`0.71-0.48 —
`
`Solubility.
`calculated
`Calculated
`
`—
`—
`Calculated
`Calculated
`
`0.60-0.66
`
`—
`
`0.45
`—
`—
`0.74
`
`0.74-0.82
`
`0.75
`
`-
`-
`0.26
`
`0.74
`(1.56
`—
`
`I-luu-Phuoc et al., 1987; Ticehurst. 1994;
`Inverse gas
`chromatography Maeda ct al.. 1992. I995
`Inverse gas
`Nakai et al., I989
`chromatography
`Calculated
`Calculated
`—
`
`Vaughan. 1985
`Vaughan, 1985
`Little. 1966: Rowe,
`
`l988c,d; 1989a
`
`—
`—
`‘
`
`Grulke. 1975; Vaughan, 1985
`Rowe. 1988a; l989a,b
`Vaughan. [985
`
`Calculated
`-
`—
`—
`0.34-0.37 —
`
`Vaughan. 1985
`Vaughan, I985
`Grulke, 1975; Bustamante et 31.. 1993b
`
`0.58
`0.38
`
`0.37
`0.18
`—
`—
`0.00
`—
`
`—
`
`Grullce, 1975
`Grullte. 1975
`
`_
`
`Calculated
`—
`—
`—
`
`Tobolsky, 1960: Grulke. 1975
`Grullte. 1975; Vaughan. 1935
`Vaughan. 1985
`Vaughan. 1985
`Barton, 1983: Rowe. 1988a
`Lee. 1968
`
`Calculated
`
`Vaughan, 1985; Sakellariou ct al., 1986
`
`0.52-0.60
`
`Calculated
`
`Sarnaha and Nagar, 1988
`
`0.50-0.55
`
`Calculated
`
`Samaha and Naggar, 1938
`
`MYLAN PHARMS. INC. EXHIBIT 1049 PAGE 6
`
`MYLAN PHARMS. INC. EXHIBIT 1049 PAGE 6
`
`

`
`B.C. Hunr'ot'k er at’. H international Journal of Pharmct'eutn:'s 148 (1997) i'—2l
`
`7
`
`Table 2 (continued)
`
`Material
`
`Polyoxycthylated octyl
`phenols
`Polytctrafiuoro-ethylene
`Polyvinylacetzite
`Polyvinyl
`alcohol
`
`Polyvinyl
`chloride
`
`Polyvinyl
`pyrrolidone
`Propylene
`glycol
`Sorbic acid
`Sorbitan laurate
`Soybean nil
`Stearic acid
`Sucrose
`I-butane-I
`Titanium
`dioxide
`Triacctin
`Twecns
`Water
`
`Sol. pararn.
`{MPafl.5}
`
`Polarity
`
`Method
`
`Reference
`
`23f'—28.4
`
`0.45-0.54
`
`Calculated
`
`Samaha and N-aggar, I988
`
`l2.7
`25.6
`l9.9 34.4
`
`2|.4
`
`21.2
`
`——
`0.33
`—
`
`0.28
`
`0.47
`
`—
`
`Calculated
`
`Tobolsky. I960
`Barton. I983; Rowe. 1988a
`Sakellariou and Rowe. i996
`
`—
`
`Barton, I983; Rowe, l988a; 1989b
`
`Calculated
`
`Rowe. 1988b
`
`25.8 -39.3
`
`0.69
`
`—-
`
`Grulke. I975: Lewis and Enever. 1979: Vaughan. I985
`
`24.4
`8.6]
`l8.2
`17.2
`l5.8 I16
`32.8
`21.0
`34.4
`
`22.0
`24.4-29.4
`47.9
`
`—
`
`—
`—
`U.l3
`—
`—
`0.5]
`
`——
`0.-‘-l5—0.57
`0.90-0.93
`
`Calculated
`Calculated
`—
`Calculated
`Calculated
`—
`—
`
`Calculated
`Calculated
`—
`
`i985
`Vaughan,
`Vaughan. 1985
`King, 1995
`Hansen. 1967b: Vaughan. 1985: Rowe,
`Roberts et 211.. 1994
`Vaughan, 1985
`I-iansen,
`|96'lb
`
`l938c
`
`Vaughan, 1985
`Samaha and Naggar, 1988
`Lewis and Enever. 1979; Grulke, I975:
`Bustamantc et al.. 1993b
`
`allows characterisation of a material when there
`
`may not be sufficient available for experimental
`determinations. Rowe and co-workers (Rowe,
`l988b, 1989b; Roberts et al., 1994) have used
`
`contribution methods to determine the
`group
`partial and total solubility parameters of a wide
`range of pharmaceutical drugs and excipients (Ta-
`bles l and 2). This method has also been used to
`
`estimate the solubility parameters of biological
`systems such as the human skin (Groning and
`Braun. 1996). Solubility parameters can be esti-
`mated from molecular structure using molecular
`modeling and molecular dynamics calculations.
`The solubility parameters of several organic sol-
`
`vents and one common pharmaceutical polymer
`(hydroxypropyl cellulose) have been determined
`in this way and the results compare favourably
`with those determined experimentally (Choi et al.,
`1994; Saga and Takahama, I996).
`
`is possible to determine partial and total
`It
`solubility parameters by measuring the solubility]
`miscibility of a material
`in liquids with known
`cohesive energies (Rcuteier-Faoro et al.. 1988}.
`The solubility parameter of the test substance is
`assumed to be the same as that of the liquid in
`which it most completely and athermally dissolves
`(Fig. 1). This method of determining solubility
`parameters is very popular because of its practical
`simplicity and its applicability to solids,
`liquids
`and gases. Archer (1992) has recently used this
`technique to determine the solubility parameters
`of some pharmaceutical
`film-coating polymers.
`For highly accurate measurements the approach
`can be combined with solution calorimetry (Rey-
`Mermet et al., 1989, 1991).
`is directly
`The cohesive energy of a material
`proportional to its surface free energy (Garden,
`197?; Samaha and Naggar, 1990) and it can be
`shown that:
`
`MYLAN PHARMS. INC. EXHIBIT 1049 PAGE 7
`
`MYLAN PHARMS. INC. EXHIBIT 1049 PAGE 7
`
`

`
`8
`
`Table 3
`
`RC. Hancock 2: ai’. /International Journal of Pharmaceutics I43 (I997) I -2}
`
`Total solubility parameters of caffeine. theophylline and phenylbutazone obtained by various methods (MPa°")
`{Rey—Mermet et a1., 1991]
`
`Method
`
`Calculated
`Sublimation
`
`Vapourisation
`lnverse gas Chromatography
`Solubility
`Partition
`Calorimetry
`Surface tension
`
`Caffeine
`
`25.6
`27.1-28.1
`
`22.4
`26.6
`27.2-28.2
`24.3
`29.9
`26.1}
`
`Theophylline
`
`Phenylbutazone
`
`28.3
`34.4
`
`31.9
`28.7
`28.2-28.7
`—
`24.0
`29.1
`
`23.9
`—-
`
`-
`28.1
`25.2
`—
`26.6
`—
`
`:52 = {y/V:,;'-l)"
`
`(6)
`
`is the surface free energy of a material
`where :2
`and n is a constant related to the arrangement of
`atoms or molecules in space. Thus, it is possible
`to calculate solubility parameters directly from
`surface free energies and molar volumes (Keen-
`hen and Smolders,
`I925; Van Krevelen and
`
`1977; Roberts and
`1976: Qiardon,
`Hoftyzer,
`Thomas, 1978). This method has been compared
`with other methods of determining solubility
`parameters for a wide range of pharmaceutical
`materials and found to correlate very well (Fig. 2)
`(Samaha and Naggar, 1990). As a consequence of
`this relationship it
`is possible to use any of the
`methods used to evaluate surface energetic prop-
`erties to determine solubility parameters. These
`methods (e.g. contact angle analysis) and their
`application to pharmaceutical systems have been
`extensively described in the literature (Stamm et
`al.. 1984; Buckton, 1990, 1992) and will not be
`described any further here.
`The solubility parameters of pharmaceutical
`solids and liquids can also be determined using
`inverse gas chromatographic (IGC) experiments,
`from the retention times of gases of known cohe-
`sive energies. This method has been used for a
`wide range of pharmaceutical
`excipients and
`drugs (‘Tables 1-4) (Huu-Phuoc et al., 1986, 198?;
`Nakai et al., 1989; Egawa et al., 1992; Ticehurst,
`1994; Ticehurst et al., 1994; King, 1995). The
`method gives precise and reproducible solubility
`parameters, but
`it
`is not
`a
`true equilibrium
`method and the results obtained may be affected
`by a heterogeneous distribution of active sites on
`
`the stationary phase. It has been argued that the
`necessary manipulation of the stationary phase
`and its prolonged exposure to the carrier gas(es)
`during the experiment may alter the measured
`cohesive energy density (e.g. by drying) (Tice—
`hurst, 1994; Ticehurst et al.. 1994). The occasional
`differences
`in
`results
`reported between this
`
`method and other simpler approaches (Tables 3
`and 4) probably result from its greater sensitivity
`to surface heterogeneities compared to bulk cohe-
`sive energy determination methods.
`The cohesive energy of a solid plays a major
`role in determining its fundamental mechanical
`properties and, thus, solubility parameters can be
`estimated from the results of mechanical property
`measurements (Roberts and Rowe, 1993; Roberts
`et al., 1994). Willbourn (1926) reported that the
`CED of various polymers is
`related to their
`Young‘s modulus, and Garden has shown a sim-
`ple relationship between the tensile strength of a
`range of inorganic materials and their solubility
`parameters (Garden, 1977). Roberts and Rowe
`(1993) have shown the validity of this type of
`approach for pharmaceutical materials. These au-
`thors measured the Young's modulus and critical
`stress intensity factor of compressed specimens of
`microcrystalline cellulose in solvents of differing
`solubility parameters. They found that the partial
`solubility parameters calculated for the microcrys-
`talline cellulose using this technique were very
`similar to those determined by methods such as
`inverse gas chromatography and contact angle
`analysis. More
`recently
`these
`authors have
`demonstrated correlations between the Young’s
`
`MYLAN PHARMS. INC. EXHIBIT 1049 PAGE 8
`
`MYLAN PHARMS. INC. EXHIBIT 1049 PAGE 8
`
`

`
`B.C. Hancock er of. _/ luternariortaf Journal of Pharmaceutics I45‘ (I99?) I -2.‘
`
`9
`
`Table 4
`
`Partial solubility parameters and fractional polarity of caffeine, theophylline and phcnylbutazone obtained by various methods
`(MPa°'5) (Rey-Mennet ct al.. I991; Ticehurst, 1994)
`
`Material
`
`Caffeine
`
`Theophylline
`
`Phenylbutazone
`
`Method
`
`IGC
`
`Solubility
`Partition
`
`Calorimetry
`Calculated
`
`Solubility
`Calorimetry
`Calculated
`
`Solubility
`Calorimetry
`
`6,
`
`26.7
`
`«id
`
`17.0
`
`6,,
`
`11.7
`
`0",,
`
`l7.0
`
`X,
`
`0.59
`
`26.6-28.7
`23.3
`
`20.1-20.6
`21.4
`
`?.2- 13.9
`3.8
`
`10.5 18.6
`8.3
`
`0.43 0.48
`0.16
`
`25.8
`23.0
`
`29.8
`24.4
`27.4
`
`17.3
`20.0
`
`23.8
`I 6.8
`I9.-4
`
`13.7
`14.3
`
`13.4
`11.3
`14.2
`
`13.4
`13.3
`
`11.9
`13.7
`13.2
`
`0.55
`0.49
`
`0.36
`0.53
`0.5()
`
`23.9-22.3
`24.0
`
`16.9-24.5
`17.5
`
`9.2-- 13.0
`12.5
`
`2.6 -10.?
`10.?
`
`0.19 -0.50
`0.47
`
`modulus, tensile strength and critical stress inten-
`sity factors of a wide range of pharmaceutical
`materials and their solubility parameters (Roberts
`et al., 1996).
`Solubility parameters of pharmaceutical poly-
`mers can be estimated from the upper limiting
`intrinsic viscosities of their solutions in solvents of
`
`from their Flory—Huggins
`varying quality, or
`polymer—solvent
`interaction parameters (Cowie,
`1965; Roberts and Thomas, 1978; LaPack et al.,
`1994; Paik and Writer, 1995). Kent and Rowe
`(1978) used intrinsic viscosity measurements to
`determine the solubility parameter of ethy1ee1lu-
`lose utilised in pharmaceutical film coatings and
`achieved results which were identical
`to several
`
`other methods ( Fig. 3). Flory- Huggins polymer-
`solvent interactions parameters can be determined
`from measurements of solvent sorption, solution
`vapour pressure, osmotic pressure or light scatter-
`ing measurements. Thus,
`this method has great
`potential for determining the solubility parame-
`ters of pharmaceutical polymers.
`The solubility parameters of pharmaceutical
`materials can be estimated from a range of other
`fundamental material properties (e.g.
`refractive
`index (Koenhen and Smolders, 1975; James et al.,
`1926; Vaughan, I985), coefficients of thermal ex-
`
`1926;
`pansion {Van Krevelen and Hoftyzer,
`Vaughan, 1985)). Such methods are not routinely
`used and they have an unknown degree of uncer-
`tainty associated with their results (Van Krevelen
`and Hoftyzer, 1976). For more details the reader
`
`referred to the book by Van Krevelen and
`is
`Hoftyzer (1926)
`
`4. Use of solubility parameters in pharmaceutical
`dosage form design
`
`A knowledge of the cohesive energy density of
`a material is invaluable in determining how it will
`behave when exposed to different external condi-
`tions (e.g. during processing, under physiological
`conditions). As a consequence of this solubility
`parameters have found widespread application in
`all aspects of pharmaceutical dosage form design.
`It is possible to divide the reported applications of
`solubility parameters into three main groups.
`These are:
`(i) prediction of unknown material
`properties; (ii) assessment of processing effects on
`material properties; and (iii)
`the prediction of
`interactions and incompatibilities between materi-
`als.
`
`4.1. Prediction of unknown material properties
`
`are
`Many fundamental material properties
`linked to the cohesive energy holding the atoms or
`molecules of that material
`together.
`It
`is thus
`possible to estimate unknown material properties
`from a knowledge of their solubility parameters.
`For example, the thermal properties of materials
`are connected to their interatomic/molecular co-
`hesive forces, and fundamental relationships be-
`
`MYLAN PHARMS. INC. EXHIBIT 1049 PAGE 9
`
`MYLAN PHARMS. INC. EXHIBIT 1049 PAGE 9
`
`

`
`ll}
`
`Table 5
`
`B.C. Hancock er al. /International Journal of Pharmaceutics l43 (I997) l—2l
`
`Partial solubility parameters of ibuprofen calculated using group contribution parameters (MPa°'5) [Roberts ct al., 1994)
`
`Group
`
`Frequency Partial molar volume
`(cm-‘.n1o|" '}
`
`‘Fa
`(J"'5,cm‘-5.m0l' ‘)
`
`‘F’: [J cm’ moi”)
`
`‘Uh {J mol ‘)
`
`CEH4
`(aromatic ring)
`CH
`
`CH2
`CH3
`COOH
`
`Total
`
`1
`
`3
`
`l
`3
`I
`
`0': = (.s§+6§,+a‘§1"-5
`= 20.36
`
`52.4
`
`-2.0
`
`l6.l
`l00.5
`20.8
`
`l8:-'.8
`
`1270
`
`2 X 80
`
`270
`3 X 420
`530
`
`3490
`
`12 100
`
`U
`
`U
`0
`I76 400
`
`l33 500
`
`0
`
`0
`
`0
`0
`10 000
`
`I0 000
`
`x_, = (¢$§,+<>'i)1o‘,-" = 0.14 ad = 349031352
`= 13.34 MPa”"
`
`5,, =11 885 000)“-‘-7135.2
`= 2.34 MP-a°-5
`
`(5,, = (10 00{lll85.2}"-5
`= 7.35 MPa"-5
`
`Molecular weight = 206.3 gmol“. true der1.sity= l.ll g cm“, total molar volume = |85.2 cm3 moi“.
`
`tween the melting point and glass transition tem-
`perature of pharmaceutical solids and their solu-
`bility parameters have been reported (Tobolsky,
`1960; Lee, 1968; Michaels et al., 1975). Paruta et
`
`al. (1962) have also demonstrated that the dielec-
`tric constant of pharmaceutical solvents can be
`related to their total solubility parameter (Fig. 4).
`The mechanical properties of solids are likewise
`related to their interatomic/intermolecular forces.
`Willbourn (1976) showed that the Young’s mod-
`ulus of various polymers is related to their CED
`
`in a linear fashion, and Garden (l9?'7) showed a
`
`similar relationship between the tensile strength of
`inorganic materials and their solubility parame-
`ters. Roberts et al. (1991, 1994, 1996) and York
`
`(1992) have recently demonstrated that similar
`relationships exist for a wide range of drugs and
`excipients (Figs. 5 and 6). Yamamoto and Fu-
`rukawa (1995) have used cohesive energy densities
`in their model to predict the shear yield stress of
`a series of amorphous polymers. With the recent
`
`
`
`3
`
`l
`
`I0
`
`'
`
`1|
`
`'
`
`I2
`
`'
`
`13
`
`'
`
`l4
`
`0.0250
`
`0.0200
`
`‘E
`.9
`D
`2 00150
`9
`
`)5
`§ 0.0100
`.0
`EU
`U.)
`
`0.0050
`
`0.0000
`
`F"\
`E
`-
`
`15
`
`25
`
`35
`
`45
`
`Solubility parameter of solvent mixture [MPa)"0.5
`
`l5
`
`Solubility parameter (MPa0'5)
`
`1. Solubility of sulfanilamide as a function of solvent
`Fig.
`solubility parameter for ethanol-water mixtures and ethanol-
`cthyl acetate mixtures (data from Bustamante et al.. 1994).
`
`Fig. 2. Correlation between the surface free energies and
`solubility parameters of some pharmaceutical
`solids (data
`from Samaha and Nagar. I990).
`
`MYLAN PHARMS. INC. EXHIBIT 1049 PAGE 10
`
`MYLAN PHARMS. INC. EXHIBIT 1049 PAGE 10
`
`

`
`sol
`Intrinsic
`viscosity I
`
`l
`
`B.C. Home-ck er of. Hnrernationa! Journal’ of Pharmaceutics 148 (I997) I-2!
`
`ll
`
`25
`
`MG
`
`_. (J1
`
`
`
`
`
`Young'smodulus(MPa) u.S
`
`L__
`15
`
`.4
`20
`
`25
`
`30
`
`0
`
`500
`
`1000
`
`1500
`
`2000
`
`Solubtl ity parameter (MPa"0. 5)
`
`Cohesive energy density {MP3}
`
`Fig. 3. The viscosity of ethyl cellulose in three classes of
`solvents as a function of their solubility parameter (data from
`Kent and Rowe. I978}.
`
`Fig. 5. Correlation between the Young's modulus and cohesive
`energy density (CED) of a series of drugs and excipients ( data
`from Roberts et al.. l99l).
`
`with varying degrees of substitution (Boehek and
`
`Petropavlovsky, I993) (Fig. 7"). In a similar way
`the solubility parameters of several alcohols have
`been calculated from those of a related ho-
`
`mologous series of alcohols (Paruta et al., 1962;
`Carre and Vial, 1994) (Fig. 8). Samaha and Nag-
`gar (1988) used a correlative approach to study
`the surface active properties of a series of non-
`ionic surfactants and showed that critical mieelle
`
`concentration (CMC) varied in a systematic way
`
`40
`
`CA)01
`
`Ca)S
`
`development of reliable models for predicting
`macroscopic material properties from molecular
`structure information it may soon be possible to
`use solubility parameters to predict many more
`fundamental material properties.
`The solubility parameters of well characterised
`materials can often be used to calculate those of
`
`less well studied but structurally similar com-
`pounds. For
`example,
`the partial
`solubility
`parameters of pure cellulose have been deter-
`mined by extrapolating those of cellulose acetate
`
`80
`
`
`
`Dielectricconstant
`
`TU(J1
`
`
`
`
`
`Tensilestrength(MP3) 8Zn‘8
`
`U1
`
`I 5
`
`45
`35
`25
`Solubility parameter (MI>’a)"0.5
`
`55
`
`0
`
`500
`
`t 000
`
`1500
`
`2000
`
`Cohesive energy density (MPa)
`
`Fig. 4. Correlation between the dielectric constants and solu-
`bility parameters of some common pharmaceutical solvents
`(data from Table 3).
`
`Fig. 6. Correlation between the tensile strength and cohesive
`energy density (CED) of a series of drugs and excipients (data
`from Roberts et al.. 1996).
`
`MYLAN PHARMS. INC. EXHIBIT 1049 PAGE 11
`
`MYLAN PHARMS. INC. EXHIBIT 1049 PAGE 11
`
`

`
`12
`
`RC. Hancock er a1. /lnternarionai Joumaf of Pharmaceutics I48 (1997) I-21
`
`
`
`
`
`Solubilityparameter(MPaO‘5)
`
`20-
`
`‘Al0 1
`
`Degree of substitution (D8)
`
`Fig. 7. Partial solubiiity parameters for cellulose acetates of
`varying degrees of substitution (DS) I = JP, A = (id, V = Eh,
`I = 6, {data from Bochelc and Petropavlovsky