Particle Engineering in Pharmaceutical Solids Processing: Surface Energy Considerations

Transcription

1 en Orers for Reprints to Current Pharmaceutical Design, 2015, 21, Particle Engineering in Pharmaceutical olis Processing: urface Energy Consierations Daryl R.Williams* Department of Chemical Engineering, Imperial College Lonon, Prince Consort Roa, Kensington Lonon W7 2AZ, Unite Kingom Abstract: During the past 10 years particle engineering in the pharmaceutical inustry has become a topic of increasing importance. Engineers an pharmacists nee to unerstan an control a range of key unit manufacturing operations such as milling, granulation, crystallisation, power mixing an ry power inhale rugs which can be very challenging. It has now become very clear that in many of these particle processing operations, the surface energy of the starting, intermeiate or final proucts is a key factor in unerstaning the processing operation an or the final prouct performance. This review will consier the surface energy an surface energy heterogeneity of crystalline solis, methos for the measurement of surface energy, effects of milling on power surface energy, ahesion an cohesion on power mixtures, crystal habits an surface energy, surface energy an power granulation processes, performance of DPI systems an finally crystallisation conitions an surface energy. This review will conclue that the importance of surface energy as a significant factor in unerstaning the performance of many particulate pharmaceutical proucts an processes has now been clearly establishe. It is still nevertheless, work in progress both in terms of evelopment of methos an establishing the limits for when surface energy is the key variable of relevance. Keywors: urface energy, contact angle, inverse gas chromatography, pharmaceutical powers, surface properties, power processing. INTRODUCTION The formulation an manufacture of moern pharmaceutical particulate proucts is a complex an multi-facete process. Inee our funamental unerstaning of these processes oes not meet the nees of inustry, especially as more challenging types of powers such as hyrophobic rugs an complex amorphous formulations become more common place. The ominance of soli state osage forms which involve particulate materials necessitates that interfacial an surface phenomena have an important role to play in etermining the quality an performance of both processes an final proucts. During the past 10 years particle engineering in the pharmaceutical inustry has become a topic of increasing importance especially as engineers an pharmacists seek to unerstan an control a range of key unit manufacturing operations such as milling, granulation, crystallisation, power mixing an ry power inhale rugs. It has now become clear that in many of these particle processing operations, the surface energy of the starting, intermeiate or final proucts can be a key factor in unerstaning the processing operation an or the final prouct performance. This review will consier the role playe by surface energy in pharmaceutical particle processing operations, an will be precee by a iscussion of the surface energy of pure pharmaceutical solis an experimental methos for its measurement. URFACE FREE ENERGY AND THERMODYNAMIC Our funamental escription of surface free energy follows from the thermoynamic free energy per unit area, ij for two phases i an j in contact [1, 2]. From the first law of thermoynamics, the internal energy of a close system is represente by: However, commonly the liqui roplet oes not fully sprea across the entire surface of a soli substrate to form a liqui film. Inee for many systems such as organic solis, the liqui roplet which exists in the presence of its own vapour, will eventually achieve an equilibrium thermoynamic state whereby a contact angle can be efine between the liqui roplet an the soli sub- U = T pv + N + A C i=1 i i *Aress corresponence to this author at the Department of Chemical Engineering, Imperial College Lonon, Prince Consort Roa, Kensington Lonon W7 2AZ, Unite Kingom; Tel: ; .r.williams@imperial.ac.uk ij (1) where U is the internal energy, the entropy, p an T the conitions of pressure an temperature, i the chemical potential of each component, N i the amount of component i, A the area of the interface an ij the interfacial free energy. The interfacial free energy, ij is efine as the increase in the internal energy of the entire system per unit increase in interface area at constant volume an entropy of the system uner close conitions, as given in (2): G ij = A T, P, N i The interfacial free energy ij may be expresse in terms of the Gibbs free energy (3): U ij = A (3), V, N i o for example, the spreaing of a liqui i over a soli surface j is etermine by their interfacial free energies. At the interface, with a new liqui surface area, A i will result in the iminishing of the soli surface area, A j. This change will also cause the creation of the interface surface area, A ij. o for the spontaneous spreaing of a liqui i over a soli substrate j, i/j is given by solving the exact ifferential shown in Eqn 4. Positive values of i/j woul inicate a spontaneous spreaing of liqui i over soli j. i/j is also known as the spreaing coefficient. G i j A = / i T, P = j i ij (2) (4) /15 $ Bentham cience Publishers

2 2678 Current Pharmaceutical Design, 2015, Vol. 21, No. 19 Daryl R.Williams strate. LV, V an L represent respectively the liqui surface free energy (commonly calle the surface tension for liquis), the solivapor surface energy an the soli-liqui surface energy (Fig. 1). The units for surface free energy are mj m -2, though sometime an equivalent unit mn m -1 is use, especially in the case of surface tension measurements. The measurement of the soli surface free energy V for an unknown material is often etermine by reference to the liqui wetting characteristics of known liquis as obtaine via contact angle measurements. Asorbe Vapour Molecules LV Liqui Droplet angle of a series of liquis with ifferent surface tensions, an thus iffering 's with the soli of interest, linke with a simple extrapolation, typically graphically to =0. Fig. 2 below shows a Zisman plot obtaine for morphine sulfate powers by Prestige an Tsatouhas using a capillary wicking metho [6]. The primary limitation of the metho is that the iffering chemical nature of the interfacial interactions of the contact angle liquis (for example hyrocarbons versus hyrogen boning liquis like water) with the soli being stuie effecte the estimates for c obtaine. Furthermore, the analysis provies no insight chemically to the interfacial phenomena associate with the wetting event. e V L urface Fig. (1). essile rop contact angle schematic for a liqui roplet on a soli substrate with asorbe vapor film. Young s equation (5) escribes the relationship between the interfacial tensions or free energies of the soli an liqui to the contact angle [3]: V L = LV cos( ) (5) The efinition for work of ahesion, WA = G, follows irectly from the interfacial free energy : A W A = i + i j (6) WC imilarly the work of cohesion, W C, is efine: = 2 (7) i Equations (6) an (7) then lea to a key relationship between the work of ahesion, the contact angle an the surface free energy of the contact angle liqui- the Young-Dupre equation (8) [1]: W [ 1+ cos ] (8) A = LV ometimes a more etaile equation is provie (9) which takes into account the change in soli-vapour surface energy cause by the presence of asorbe vapor which exerts a film pressure of e. The significant experimental ifficulty in measuring e means that Eqn (8) is the normally use form. W [ 1+ cos ] + e (9) A = LV The use of a contact angle liqui with known surface tension properties to etermine the surface energetics of an unknown soli state substrate seems a simple enough experimental concept. However, the equations which link these concepts together have only slowly evolve over the past 50 years. The reaer is invite to rea the excellent an comprehensive review on this topic by Eztler [4]. In this review, we will summarise the major theoretical issues core to this topic, in broaly a historical orer. The etermination of the surface energy of solis was pioneere by Zisman who stuie the surface energy of polymeric surfaces in the early 1960's at the Naval Research Laboratories [5]. He introuce the essentially empirical concept of the critical wetting tensions c for a soli surface, which was the surface tension of a liqui which just exhibite a contact angle of zero with the soli surface uner stuy. This analysis was base on measuring the contact Fig. (2). Morphine sulfate wetting rates as a function of the surface tension of wetting liquis [6]. The Zisman approach prove popular until the 1970's when a number of improve but relate methos for surface energy analysis were evelope which coul be broaly escribe as semiempirical in their nature. These evelopments were riven by the concept evelope by Fowkes [7] who avance the important assumption that the surface tension for many organic materials coul be consiere to be compose of two inepenent components; one essentially relating to long range physical forces an another term relating to shorter range chemical forces. Eqn (10) below shows that simple equation where the surface energy or surface tension for a liqui or soli respectively can be consiere to be compose of a long range force component ue to Lonon van e Waals forces which is commonly referre to as, as well as a secon component which escribe shorter range forces, originally escribe as polar forces p. In his original paper, Fowkes consiere a wier range of sources for these interactions other than an p incluing metallic boning, inuction forces etc for a general material system. p = + (10) The evelopment of Eqn (10) catalyse a number of semiempirical moels for preicting the work of ahesion between a pair of organic material phases. These approaches were base on (i) an p being treate as inepenent quantities an (ii) geometric mean approximations base on Berthelot s principle being use to estimate interfacial interactions. Fowkes then allowe thermoynamic terms such as the work of ahesion W A, to be approximate as a simple sum of these inepenent terms, each relating to a specific type of intermolecular interactions. Owens an Went extene Fowkes equation by grouping the non-boning Lonon, Debye an Keesom interactions into a similar term which they calle ispersive (also recognise as the Lifshitz-van er Waals interactions)

3 Particle Engineering in Pharmaceutical olis Processing Current Pharmaceutical Design, 2015, Vol. 21, No while the remaining terms were groupe as the polar contributions [8]. The Owens-Went relationship is also known as the Kaelble Eqn [9] an is given by (11) for a soli-liqui system: p p L = V + LV 2 V LV 2 V LV (11) This specific approach has been popular because of its simplicity an robustness, with 100 s of papers reporting ata using this analysis. Practically using two contact angle liquis, commonly water an iiomethane for whom an p L are known for both, L contact angles for these two liquis on an unknown surface allows an to be etermine for this soli. p The next major theoretical evelopment came from Fowkes an Mostafa [10] who shifte the nature of the formalism by which intermolecular forces were split into two inepenent classes of intermolecular interactions. They consiere two specific classes of intermolecular interactions; long range ispersive interactions an short range aci-base interactions; the later term replacing the previous use term. In the case of the aci-base interactions they p propose the use of the well-establishe linear free energy relationship evelope by Drago [11]; the 4 parameter E&C moel. Incorporation of this moel into a wetting equation theoretically allowe preictions of the interfacial aci-base free energies of interaction. It is fair to say that espite the soun theoretical basis of this approach, this metho i not prove to be popular or useful, mainly because the E&C constants for many contact angle liquis were not known, as well as the lack of reliable estimations for the number of aci-base pair interactions occurring at the interface of interest, a necessary number for using this new theory. A simpler an more useable relationship for aci-base interactions uses the 2 parameter Gutmann onor (DN) an acceptor (AN) number linear free energy relationship which was use by Rile an Fowkes [12] to generate the aci an base constants for the surface, K A an K B. These constants are sometimes reporte as acceptor an onor constants for a surface K A an K D. More recently, van Oss et al. [13, 14] introuce a refine semi-empirical analysis of aci base interactions, which is given in Eqn (12) below, which consiers three inepenent constants for escribing the aci-base an long range interactions; LW, + an. These parameters escribe the long range ispersive, aci an basic surface energies respectively. Unlike previous moels this approach introuce an empirical specific aci an basic escriptor for each surface present. However, the van Oss moel utilise water as reference materials with the prouct + : - equal to unity. This assumption often resulte in an over-basic surface energy parameter estimates. Della Volpe [15] propose ifferent ratios suggesting that water was not amphoteric but preominantly aciic. LW LW + + (12) L = V LV V LV V LV V LV In concluing this section it is worthwhile commenting on the limite take up an use of the aci-base moels evelope in the past 30 years for characterising the surface energy of soli surfaces. The major problems really come from two iffering issues. Firstly, an maybe most critically, it is not currently possible to valiate using an inepenent experimental methoology, the theoretical correctness of any the aci-base moels summarise here for analysing the interfacial thermoynamics of wetting. It is also not clear whether contact angle an surface tension measurements by themselves give accurate enough estimates for an. Therefore + a + lack of high fielity an estimates for the contact angle test fluis as further uncertainties in estimating + an values for surfaces of interest. econly, the assumption implicit in all of the above aci-base moels is that the surface of interest is basic an/or aciic an has only one type of aci or base site present. This assumption will be seen in this review to be both unrealistic an overly simplistic, especially for pharmaceutical soli state materials. MEAUREMENT OF URFACE ENERGY- CONTACT ANGLE METHOD Historically the scientific literature on the surface energy of organic materials has been most ominate by stuies on polymeric materials, often in the form of films an fibres. In the case of films an other flat sample forms, contact angle techniques have long been the mainstay methos, an these approaches have in been in routine use since the work of Zisman an co-workers at NRL in the 1960 s [5]. essile rop techniques represent the most commonly reporte experimental approach for the surface energy etermination of soli state films an monolithic solis. Here a roplet of a known liqui which has a surface tension higher than the surface energy of the soli surface of interest is place on the planar substrate of interest. By imaging the roplet shape profile the contact angle can be estimate. Either by estimating visually the tangent at the three phase contact line or more commonly these ays, the Young-Laplace equation is fitte to a igital image of the roplet profile, which in turn allows the contact angle to be measure. Most commonly the avancing contact angle for the liqui is then reporte as being representative of the equilibrium contact angle. Though the measurement of contact angles is an intrinsically simple metho, there are a number of complicating factors, both in terms of experimental esign as well as in ata interpretation which researchers nee to be cognisant of. An ieal flat substrate for contact angle stuy shoul possess no surface roughness or porosity. The observe experience is that contact angles are not uniquely efine, an that commonly there exist for a specific liqui-soli system a range of stable measurable contact angles, which are characterise by maximum an minimum measurable angles. These are known as the avancing an the receing contact angles respectively, an the ifference between these two angles is referre to as contact angle hysteresis. Hysteresis is a function of a number of experimental systems variables incluing surface roughness an the extent of surface chemical heterogeneity of the soli. The effects of surface roughness an chemical heterogeneity on contact angles are given by the classic papers by Wenzel [16] as well as Cassie an Baxter [17]. The other primary metho for contact angle etermination is base on wetting force measurements. In this case the substrate of interest, assuming it has a regular geometric shape such as a fibre or a film, can be attache to a sensitive microbalance an the wetting force irectly measure when the soli is immerse in the liqui of interest, an the soli pulle through the liqui surfaces. This metho forms the basis of the Wilhelmy approach an forces are escribe by the Eqn (13): F = LV Lcos (13) where F is the wetting force, LV is the surface tension of the test liqui, q is the contact angle an L the contact length of the liqui with the soli, usually the sample perimeter length. By avancing or receing the three phase contact line of the liqui, avancing an receing contact angles can be measure. This approach is the metho of choice for measuring the wetting behaviour of thin fibres. The technique can also be use for thin films an has been aapte for use also with power compacts. It shoul however be note that the formation of power compacts for either sessile rop or Washburn type experiments carries the risk that the power

4 2680 Current Pharmaceutical Design, 2015, Vol. 21, No. 19 Daryl R.Williams compaction process may result in particle fracture, thus creating new surfaces not representative of the sample at large. A key conition for measuring an equilibrium wetting property, which is an equilibrium thermoynamic property, is that there shoul be no significant chemical interactions between the liqui an soli substrate uner investigation that might change the properties of either the liqui or the soli substrate uring the experiment. In the case of pharmaceutical solis any one of a number of phenomena can seriously invaliate or compromise the equilibrium contact angle etermination. welling of the substrate as well as issolution of the substrate can effect both as well as LV, an may even change the topography of the substrate. Chemical reactions, incluing hyrate or solvate formation can funamentally compromise the equilibrium nature of the measurement. ome materials will result in liqui absorption into porous substrates, changing the surface to be characterise by the presence of these liqui species. A further subtle, but not insignificant, constraint is that the probe liqui must have a surface tension greater than the surface energy of the soli to be teste, so as to give a contact angle greater than zero. In general organic materials have surface energies in the range of 35 to 55 mjm -2, so the list of available test liquis tens to be very limite; water, iioo methane, ethylene glycol, glycerol an formamie are the ones most commonly use. Probably 95% plus of all contact angle ata reporte can be attribute to the use of water an or iioo methane. The wetting characterisation of porous samples, incluing power compacts, is a more complex matter. Most methos in use are base on the classic Washburn Eqn (14) for the capillary rise of a liqui in a network of uniform cylinrical capillaries: r LV cos (14) v = 2l Where v escribes the rate of penetration of liqui into cylinrical capillaries, LV is the surface tension of the test liqui, is the liqui viscosity an l the epth of penetration, r is the capillary raius an is the contact angle. One major limitation of this approach is that few porous solis can be consiere to be represente as ensembles of uniform capillaries. essile rop techniques can be use for both power compacts as well as flat substrates which powers have been ahere to an have also both been use to varying levels of success. Both of these methos will be iscusse later in this review. As pharmaceutical proucts are commonly in the form of particulate materials, the characterisation of their surface energy is challenging as the methos available have complications, an their reactivity towars the contact angle fluis can result in a range of potential non-equilibrium effects. Therefore not surprisingly, alternate methos for etermining the surface energy of pharmaceutical powers have been evelope incluing Inverse Gas Chromatography (IGC) an Atomic Force Microscopy (AFM). Of these IGC has attracte the most interest. URFACE ENERGY MEAUREMENT UING ATOMIC FORCE MICROCOPY This review has thus far focusse on the more commonly use chemical methos for surface energy etermination base on liqui an gas phase probing of a soli surface. During the past 10 years, the use of atomic force microscopy (AFM) to measure surface energy of pharmaceutical soli surfaces has attracte research interest. Louey et al. [18] were one of the first groups to report on the surface ahesive forces of pharmaceutical powers using AFM. Dry power inhaler (DPI) formulations were the focus of this stuy an the authors use a stanar spherical AFM probe which enable reproucible ahesional characteristics of the particle surface to be etermine. The subsequently estimate surface energies were very low, < 2mJm -2. Though the specific reasons for such low surface energies were not establishe, reasons such as particle size, chemical composition of the etaching particle, surface roughness an contact geometry were consiere as potential explanations. A subsequent stuy also on DPI powers was reporte by Berar et al. [19]. They irectly measure using an AFM the ahesional forces between carrier particles (lactose) an rug surfaces, incluing both crystalline an amorphous forms of zanamivir. Aitionally, they reporte that the ahesion forces graually increase with the increasing RH. The authors i not report the surface energy of any of the surfaces stuie here. To estimate the surface energy of a soli using AFM force ata, the contact areas nee to be estimate as a precursor to estimating the work of ahesion. uch a stuy was complete by Hooton et al. [20] who were the first workers to report works of ahesion for a series of salbutamol sulphate particles using AFM ata. A stuy of AFM contact geometry was also reporte by Hooton et al. [21] who showe that when consiering iniviual particles, ifferences in surface chemistry an asperity geometry can lea to rastic changes in ahesion with ifferent humiity conitions. Thus, they argue that the interpretation of the AFM measurement of particle ahesion force nees to take account of these factors if sensible conclusions are to be rawn for pharmaceutical power systems. Begat et al. [22] have also looke at DPI powers an specifically stuie how to minimize the variations in contact area between the AFM probe an substrates, by using nanometer smooth crystal surfaces of the rugs an the excipient. This improve uniformity in contact area allowe accurate an reproucible force measurements to be achieve. Cohesive-ahesive force balance graphs were then evelope allowing irect comparison of the interaction forces occurring in moel carrier-base formulations of salbutamol sulphate-lactose an buesonie-lactose. The use of AFM base force ata for pharmaceutical powers has been briefly reviewe by Roberts [23]. Traini et al. [24] has use the works of ahesion an cohesion for preicting the suspension stability for pressurise metere ose inhalers powers using AFM, IGC an contact angle ata. AFM works of ahesion ata i not correlate with base works of ahesion as etermine from IGC or contact angle experiments. However, works of ahesion which inclue polar/aci-base interactions gave more promising correlations with AFM force ata. Detaile surface energy ata using IGC an AFM have been reporte by Davies et al. [25] for buesonie particles as well as other moel substrates. An unmille buesonie material isplaye surface energy, as etermine by AFM of the (002) crystal face of mjm -2. The surface energy of the micronise material as etermine by IGC was 68 mjm -2. The variability in surface energy from AFM, especially apparent for the micronise buesonie was attribute to two factors, intrinsic material variations within a single particle an assumptions present within the contact mechanics moel use. -lactose monohyrate continues to be one of the most stuie materials using IGC. Zhang et al. [26] have use AFM to etermine of mille an crystalline -lactose monohyrate. AFM is potentially a very attractive metho as it allows the surface energy to be irectly relate to local surface topology an features. It s own sie is the uncertainly in the measurement in terms of spring constant calibration, area of contact, tip raius an the choice of the most appropriate DMT or JKR equation for ata analysis, as well as the evelopment of a suitable experimental metho for measuring tip-surface interaction forces. When the surface of interest is a particle, these experiments can be especially challenging. In this stuy the surface energy of crystalline face (0-1 -1) an a cast amorphous

5 Particle Engineering in Pharmaceutical olis Processing Current Pharmaceutical Design, 2015, Vol. 21, No lactose film was etermine to be 23.3 mjm -2 an 57.4 mjm -2 respectively. This paper emonstrates that AFM can provie localise information from iniviual faces or within components of heterogeneous powere samples. MEAUREMENT OF URFACE ENERGY- INVERE GA CHROMATOGRAPHY METHOD The recent popularity in the use of inverse gas chromatography (IGC) methos for the surface energy characterisation of powers reflects new scientific evelopments in the experimental technique as well as a growing appreciation that gas (strictly speaking vapour) asorption methos can provie a rich an unique unerstaning of the surface thermoynamic properties of powers not possible using any other metho. The history of IGC can be trace back to a number of key research publications in the 1960 s an The reaer will fin the seminal books by Kiselev an Yashin [27] as well as Coner an Young [28] a very useful backgroun rea for this topic an reveal the enormous potential of the IGC metho; not just surface energy measurements which is the current topic. Etzler has provie a very etaile an extensive review of surface energy incluing much etail on both theory an experiment [4]. Grimsey et al. have also reviewe some papers on IGC an surface energy effects in 2002 [29]. It is also appropriate to comment on the IGC instrumentation use to measure surface properties. Papers oler than about 15 years were all base on home built IGC instruments, with commercial instrument systems appearing in about The first commercial IGC instrument was the IGC 2000 manufacture by urface Measurement ystems an was able to measure a wie range of surface properties incluing the with an automate gas injection systems capable of elivering up to 12 ifferent vapours. In 2012 this system was supersee by the IGC-EA which allowe IGC to be performe for a series of user selectable fixe surface coverages for all asorbates. This feature in turn allowe to be etermine for specific surface coverages, ie. an isostere metho, which in turn allowe surface energy heterogeneity to be quantifie. This key evelopment allows more robust an more sensible comparison of IGC ata obtaine by ifferent groups or researchers as the values can now be irectly linke to a specific surface coverage. IGC experimental methos are vapour asorption methos, an are conceptually relate to the BET surface areas techniques which are in common usage base on gas (nitrogen) asorption. IGC experiments are however typically conucte at vapor concentrations below that at which monolayer surface coverage occurs, such that the asorbing molecules behave inepenently an retention behaviour is thus in the Henry s Law region. Methane Gas n Octane Vapour Fig. (3). eries of chromatograms obtaine for alkane vapour species interacting with a GC column packe with crystalline fibres. Fig. 3 above shows four overlai chromatograms obtaine for alkane vapour species interacting with a GC column packe with crystalline fibres at 40 o C uner conitions of infinite ilution. The quantity t M is the retention time for an inert non-interacting gas species to sweep through the packe column. This time is known as the experimental ea-time an is typically measure using methane or nitrogen. This retention time t M when multiplie by the carrier gas flow rate F accurately approximates the ea volume V M within the system. This ea volume consists of the internal volume of the instrumentation plumbing as well as ea-space within the sample column an its packing. As the hyrocarbon chain lengths increases of the vapour solute injecte, so oes the propensity of the solute species to interact with the sample surface via asorption. This results in increasing retention times for the solute molecules with increase molecular mass an the tren is clearly shown in the peaks shown below for hexane through octane. They irect reflect the stronger molecular interactions of octane vapor for the surface compare say to hexane. The increase resience time in the GC column for the larger alkanes irectly results in broaer an less intense solute peaks ue to increase longituinal iffusive broaening. The peaks nevertheless maintain their Gaussian shape. The retention time t R per unit of sample mass for an asorbing solute vapour allows the nett retention volume V N to be etermine using Eqn (15): V N = j.t R.F.(T/273.15) - j.t M.F.(T/273.15) (15) where F is the carrier gas flow rate, T is the column temperature, t M is the ea time an j is a correction term allowing for the pressure changes along the column. The retention process for the solute with the stationary phase is etermine by the solute partitioning between the stationary an mobile phases at the relevant temperature, pressure an concentration. For the case in which the retention process is ue to soli-vapor asorption as in the case of surface energy measurements, solute partitioning between the mobile an stationary phases is given by the appropriate asorption isotherm. At low solute concentrations (<<0.01 P/Po) the asorption isotherm is typically linear an this region is commonly escribe as the Henry s Law region. In this region solute molecules asorbing onto a surface are inepenent an nearest neighbor interactions are not significant. Gaussian shape chromatograms as shown in Fig. 3 result in the case of asorption in the Henry s Law region. In this linear region of the asorption isotherm, the nett retention volume V N may be relate irectly to the surface area of the sample A an the partitioning coefficient K in the case of surface retention of the solute: V N = A. K (16) K is also the slope of the asorption isotherm at infinite ilution an is efine as the ratio of the solute concentration q in the stationary phase to the solute concentration c in the mobile phase: K = q /c (17) It is thus apparent that the nett retention volume V N is irectly proportional to the slope of the asorption isotherm an thus the equilibrium constant for the asorption process. Consequently stanar thermoynamic analysis may be applie to the ata. For example, from the temperature epenent partitioning coefficient K the heat of asorption o H A using Equation (18) may be etermine. Choice of appropriate stanar states for the asorbe species allows both free energies of asorption G A o an entropies of asorption o A to be etermine as well. o (ln K ) (18) q = H A = R (1/ T )

6 2682 Current Pharmaceutical Design, 2015, Vol. 21, No. 19 Daryl R.Williams G o A Ks. p = RT ln s s, g (19) o q + G o A A = (20) T A stuy of V N as a function of temperature thus allows a etaile stuy of surface asorption thermoynamics to be unertaken using the above equations. Measuring the retention behavior of a series of alkane probes allows the ispersive (long range) van er Waals component of the surface energy of the surface, to be estimate. By measuring the change in retention volume as one increases the molecular size of the alkane probe, the ifferential free energy of asorption for an imaginary -CH 2 - species can be etermine. Using the assumption that an infinite surface of -CH 2 - groups is equivalent to a polyethylene surface, an with the use of the Fowkes geometric mean work of ahesion analysis, be estimate using Eqn 21 : 4 = V N. a CH2 G A CH2 2 where N is Avogaro s constant, a CH 2 may (21) is the area of a -CH 2 - group an is the surface energy of polyethylene. This simple ratio V measurement was evelope by Gray an Dorris [30] has become very popular, not least because there is no nee to know the exact surface area for the sample or the cross-sectional area of the asorbing molecule. Another approach for estimating was reporte by chultz an co-workers [31]. This approach results in a ifferent epenent variable be plotte in the x axis, am, where a is the L m cross-section asorption area of the alkane species an is the L surface tension of the same alkane liqui. Both are soun methos for ata analyses an espite the aitional nee for accurate estimates for a, the chultz metho is more commonly reporte. m Both methos typically give the same estimate for to within 1 mjm -2 though it has been reporte that the Dorris-Gray metho oes yiel more accurate estimates for [32]. In a more extensive evaluation, hi et al. [33] have looke in etail at the relationship between the Dorris-Gray an chultz methos. They conclue that the Dorris-Gray estimates will always be larger than the chultz estimates, an that the Dorris-Gray estimates have a more accurate basis. o though we have iscusse etails of the analysis of the long range van er Waals forces using alkane vapours, what about the short range forces? These are etermine using a number of aci an basic vapour injections, often measure on the samples after the alkane experiments are complete. The soli-vapour retention by aci an base probes by the surface of interest is partially ue to the ispersive interactions of the aci/base probes with the surface, an partially ue to the aci/base interactions with the surface. ince the ispersive properties of the aci/base probes are well recore, the retention volume ue to aci-base asorption interactions can be etermine, an this is commonly converte into a free energy of sp specific asorption; G A. Data is typically reporte in one of two sp main formats. Values of G A for a particular aci or base probe molecule, such as ichloromethane, ethanol or ethyl acetate for example may be irectly reporte, or this ata may be analyse using the Gutmann approach, in which case specific aci an base numbers (K A, K B ) for the surface maybe reporte, or more commonly a ratio of them K A / K B is reporte. More etails on all of these IGC analyses are provie in the excellent recent review by Mohammai-Jam an Waters [34]. Until relatively recently, IGC ata for estimating was essentially base on a series of injections using alkane probe molecules at infinite ilution injection conitions. This analysis was base on the use of a Henry s Law equation, an therefore assume that the asorption isotherm is linear for low surface coverages an implicitly that the surface must therefore be energetically homogeneous. A key implication of this assumption is that retention times, an thus surface energies, are inepenent of solute injection size. Inee for virtually all of the research work reporte on the characterisation of organic materials such a polymers, an more recently research work on pharmaceutical solis, the assumption that such surfaces must be energetically homogeneous ha not been challenge or valiate. The avent of a number of wetting papers since 2000 [6, 35-37] has clearly confirme the surface energy heterogeneity of crystalline pharmaceutical solis. As a result of this work IGC researchers have aopte new experimental methos so as to establish, or not, the surface energy heterogeneity of these soli state materials. The key reference to this new approach was reporte by Ylä- Mäihäniemi et al. [32] who introuce the concept of surface energy istributions using an isostere base analysis of IGC ata. In this new approach was etermine at a number of specific surface coverages (e.g. 0.2% or 2% surface coverage) by a series of vapour injections with the usual alkane probe molecules. This approach takes into account the specific cross-section area of the asorbing species so that coul be estimate for each of these surface coverages. The premise was that asorption at low surface coverages such as 0.2% surface coverage woul sample the higher energy an more active surface sites, on average, than woul be sense by an experiment conucte at 2.0% surface coverage. Fig. 4 below shows schematically this approach. A refinement of this metho for analysing IGC ata was propose by Jefferson et al. [38]. Here the authors improve the analysis of surface energy istributions using a numerical moel with incorporates a Boltzmann factor into probabilities for the asorption of iffering hyrocarbon species. WETTABILITY OF PHARMACEUTICAL POWDER AND CRYTAL Early stuies on pharmaceutical power wetting focusse on the capillary rise base liqui approaches. Buckton an Newton [39] etermine the critical surface tensions of amylobartbitone powers using a range of liquis an liqui mixtures. Duncan- Hewitt an Nisman [40] comprehensively stuie the wettability of a acetaminophen an aipic aci using a range of methos incluing seimentation, thin coate films, capillary rise as well as sessile rop measurements on power compacts an large single crystals. There was a surprising excellent agreement across the measurement approaches, allowing to be estimate using an equation of state approach. A stuy on pharmaceutical powers by Dove et al. [41] examine the wetting behaviour of caffeine an theophylline powers using two ifferent contact angle techniques as well as infinite ilution IGC. The authors fabricate thin power tablets an use a Wilhelmy plate wetting metho, as well as etermining sessile rop contact angles for very thin layers of powers ahere to glass cover slies using a spray on ahesive. In the case of the Wilhelmy plate ata, some significant inconsistencies in the ata were reporte aroun very high values of. These anomalies are not surpris- ing consiering the complex nature of the topography an porosity

7 Particle Engineering in Pharmaceutical olis Processing Current Pharmaceutical Design, 2015, Vol. 21, No Fig. (4). chematic iagram of the experimental metho for etermining the istribution profile for [32]. of the power plate being stuie. The contact angle an surface energy ata from the power coate glass slies seeme to be more reliable, with values of 44.5 an 43.4 mjm -2 being reporte for caffeine an theophylline surfaces respectively. The real concern about this coate glass slie metho is the risk that the contact angle ata will partially reflect the contact angle of the unerlying glue. The authors were able to argue that this i not seem to be a significant effect in their stuy. In the case of the IGC ata, values of 39.9 an 50.8 mjm -2 being reporte for caffeine an theophylline respectively. o though both the IGC an power coate glass slie contact ata gave sensible numbers for, there were still some significant ifferences in the values reporte of between 4 an 8 mjm -2 between these two methos. An extensive stuy of the surface energy of microcrystalline celluloses using capillary wetting an IGC was reporte by teele et al. [42]. Despite the potential problems in interpreting capillary wetting ata for these water swellable solis, the authors reporte surface energies for a range of inustrial microcrystalline cellulose powers. Wettability stuies of morphine sulfate powers were reporte by Prestige an Tsatouhas [6] using a capillary rise approach. Reasonably reproucible wetting rates were reporte for number of liquis, though sessile rop contact angle experiments on power compacts were not successful. The Zisman wetting tension was then reporte for morphine sulfate an the authors surmise that the relative exposure of ifferent crystal faces plays an important role in controlling the wettability of morphine sulfate powers. Muster an Prestige [35] have use a series of crystal facet specific measurements to etermine the face-specific structure, chemistry, an wettability of two moel pharmaceutical crystals; N,n-octyl-D-gluconamie (OGA) as well as forms I an III of sulfathiazole (TZ). Direct sessile rop contact angles of water were measure on a number of crystal faces, where sufficiently large enough crystals coul be prepare an these experiments were augmente with AFM an ToF-IM ata. ToF-IM molecular fragmentation ata confirme surface compositional ifferences for ifferent facets of TZ form III an OGA crystals. Contact angle ata showe some ifferences in the facet specific results which were interprete in terms of face specific surface chemistries. The surface properties of two crystal habits of celecoxib (CEL), an acicular crystal habit (CEL-A) an a plate-shape crystal habit (CEL-P) have been reporte [43]. Power wetting ata confirme ifferences in the wettability an surface energy of these two habits. Enhance issolution an pharmaceutical performance of CEL-P was attribute to the presence of more hyrophilic surfaces in this habit compare to CEL-A. Heng et al. have reporte a stuy of the surface energy of a number of API systems using large single crystal contact angle eterminations, sessile rops measurements on power compacts as well as capillary rise experiments using a wetting tensiometer [44]. Contact angle eterminations on large macroscopic crystals were foun to be the most effective measurement approach an ata obtaine for aspirin, racemic ibuprofen, -(+)-Ibuprofen as well as paracetamol forms I an II were reporte. They confirme facet specific contact angle an surface energies for all systems stuie. uch rich an reliable ata sets on the surface energy of crystal facets coul not be obtaine with any of the power base contact angle approaches. The most etaile stuy of facet specific surface energies as well as facet surface chemistry for an API was reporte by Heng et al. [36]. In this stuy the anisotropic surface energetics an wettability of macroscopic form I crystals of paracetamol were exam-

8 2684 Current Pharmaceutical Design, 2015, Vol. 21, No. 19 Daryl R.Williams Table I. Paracetamol form I crystals- surface energy, unit crystal cell an OH Group Density [36]. ine in etail, as well as the surface chemical composition of the crystal facets as etermine by XP. Contact angles were etermine using a sessile rop metho on large macroscopic crystals. Experiments were also conucte using paracetamol saturate water solutions to exclue any issolution effects on the water contact angle ata obtaine. Table I above summarises the surface energy ata as well as OH group concentrations for each crystal facet. The relative surface polarity for the facets was in ecreasing orer was (001) > (011) > (201) > (110) > (010) agreeing with the fraction of expose polar hyroxyl groups as etermine using XP, an correlate with the number of non-hyrogen-bone hyroxyl groups per unit area present for each crystal facet using known crystal structures. Cleaving form I crystals expose a more apolar (010) surface with very ifferent surface properties, incluing a of 45 mjm -2 which was also much more hyrophobic surface than the other external crystal surfaces. A subsequent paper on forms I an II of paracetamol conclue that the iffering polymorphic forms exhibite significant variations in their wetting behaviour for the same Miller inexe faces, though anisotropic surface energetics was reporte for both form I an form II crystal surfaces [45]. The wettability of the (001), (100), an (011) crystallographic facets of macroscopic aspirin crystals has been experimentally investigate using a sessile rop contact angle metho [37]. The surface energetics were foun to be anisotropic an facet epenant, being irectly relate to the presence of surface carboxylic groups. A key question which is raise by surface energy characterisation of solis by wetting experiments an by IGC experiments is can this ata be compare, an oes one obtain the same values for if the work is performe rigorously? This crucial question was answere in 2010 by Ho an co-workers at Imperial College [46]. In their stuy they use a moel hyrophilic excipient, -mannitol an from it create a moel hyrophobic excipient, methyl silanise -mannitol. These two materials were then prouce in both a powere form as well as in a large single crystal form, allowing both IGC an contact angle analysis to be eploye on the respective sample forms (Fig. 5). For completeness they also looke at the face specific surface chemistry of the primary crystal facets of - mannitol; (010), (120) an (011). Fig. (5). Macroscopic crystals of -mannitol grown from aqueous solution [46]. Table II. urface energy of untreate an silanise -mannitol as etermine with Owen-Went analysis [46]. 0 Fig. (6). an G AB (ethanol) istributions for untreate an silanise - mannitol [46]. The IGC ata in Fig. 6 for the hyrophobic -mannitol shows to be very constant as a function of surface coverage at 34.5 mjm -2, whereas the contact angle ata summarise in Table II gives to be in the range 34.5 to 35.0 mjm -2. For the normal hyrophilic -mannitol, varie between the ifferent facets, ranging from 39 to 44 mjm -2. The IGC heterogeneity graphs in Fig. 6 shows a maximum value of about 47 mjm -2, ecreasing to a plateaux value of about 40 mjm -2 for high surface coverages. Therefore ata for both hyrophilic an hyrophobic -mannitol gave IGC ata which was entirely consistent with the ata obtaine with contact angles on single crystal facets. XP confirme that the (011) facet on normal -mannitol, which was the most hyrophilic surface, ha the highest surface concentration of OH groups. Ho et al. [47] have examine the surface energy of ifferent size fractions of the form of -mannitol. The essential premise of the work is that ifferent size fractions of -mannitol have ifferent crystal aspect ratios, an that resultant exposure of ifferent crystal facets woul be reflecte by the surface energy istributions measure. Fig. 7 below shows the surface energy istributions as function of size fraction, as well as lines showing the

9 Particle Engineering in Pharmaceutical olis Processing Current Pharmaceutical Design, 2015, Vol. 21, No the specific relationship between surface property escriptors of powers an their physical stability is a much more complex topic which is currently beyon our current unerstaning scientifically. The effect of aing alkylpolyglycosie surfactants to spray rie salbutamol sulphate particles was reporte by Columbano et al. [51]. Using IGC they etermine that values were very similar for all formulations teste, though some small ifferences in K D / K A ratios for some formulations were note. Fig. (7). [47]. istributions of -mannitol obtaine with ifferent sieve sizes surface energy for specific known facets of -mannitol etermine by contact angle stuies on large -mannitol crystals. The work confirme that surface energy istributions are sensitive to the shape an thus population of crystal facets foun in a crystalline power system. TRIBOELECTRIC CHARGING OF POWDER The triboelectric charging of a power is a measure of a powers ability to accept or onate electrons to, or from, another surface with which it is typically in ynamic contact. This phenomena s presence can manifest itself on processing equipment surfaces or even containers use to store or ispense powers. It is of great practical importance in the pharmaceutical inustry an many power processing operations can suffer from such charging effects. In increasing interest in hyrophobic API s means that these effects are likely to be more prevalent in the future. Triboelectric charging of powers has been specifically reviewe recently [48]. Ahfat et al. [49] reporte an early stuy on power tribocharging an surface energy for pharmaceutical powers, both excipients an API s. The surface energies were measure using IGC, as well as by a sessile rop technique where the pharmaceutical power was ahere to glass slies (as previously reporte by Dove et al. p [41]) an values surface energies were reporte from wetting experiments as well as an Gutmann K A an K D numbers from the IGC experiment. First orer power tribocharging ata was obtaine using a Faraay well technique mae from stainless steel. It was anticipate that the tribocharging woul be ascribable to surface chemical groups present on the particle surfaces an there was some promise that the IGC ata using K A an K D might prove incitefull. A positive correlation was establishe between the K A an K D an the electrostatic charge measure. PRAY DRYING OF OLID Ohta an Buckton [50] have examine the stability of two amorphous spraye rie formulations of cefitoren pivoxil both of which exhibite ientical T g s as etermine by DC. However, the formulations were known to exhibit significant ifferences in their physical stabilities. Using IGC at infinite ilution, small ifferences in were observe between the formulations, as well as more significant ifferences in K D an K A for these two formulations which coul be ascribe to subtle ifferences in the aci-base surface chemistry of these two formulations. It was conclue that surface energy escriptors as etermine by IGC coul be use for stuying batch to batch variations in these formulations. However, WET GRANULATION Wet granulation is a complex process for particle enlargement which is known to epen on a wie range of properties incluing particle shape an size, granule packing/porosity, shear forces, power flow ynamics as well as the wetting a spreaing of the biner solution over the power particles. Despite these complexities, various workers have attempte to ientify the specific contribution that power wetting makes to this process. The importance of particle surface energetics ata in optimising wet granulation systems was pioneere by Rowe in the late 1980's [52-54]. urface thermoynamics for cohesion an ahesion, an specifically the spreaing coefficients between substrates an biners solutions i / j were use to inform the selection of biners base on an experimental knowlege of the surface energies of substrates an biners solutions. Yokoi et al. [55] stuie the physicochemical properties of wet granulate an spray rie cefitoren pivoxil with various aitives. They suggeste that ifference in aci/base surface chemistry in the ifferent formulations coul be measure using IGC, an propose that ifferent surface chemical groups were present on the particle surface which relate to the formulation composition. The importance of the surface free energy in the selection of excipients in the course of a wet-granulation of metroniazole was reporte by Tuske et al. [56]. Corn starch, lactose, microcrystalline cellulose, hyroxyl-propylcellulose were all consiere as potential excipients for formulation with metroniazole. Contact angle ata for the excipients allowe the spreaing coefficients to be estimate for biner solution spreaing. These results were correlate with the friability of the final wet granulate proucts. When the spreaing coefficient of a biner over the substrate is positive, the formation of ense, non-friable pellets can be expecte. However, the spreaing coefficient results alone i not preict completely the granule properties of complex formulations. In another stuy of biner liqui spreaing, the wettability of a number of pharmaceutical powers was etermine from contact angle ata, an the subsequent values for surface energies were then use to etermine spreaing coefficients. This etermination allowe the selection of the most appropriate granulating solvent for a wet granulation process [57]. Preicte granulating solvent performance using solvent rug spreaing coefficients was in goo agreement with resulting granules properties, specifically in terms of ensity, porosity an friability. A etaile stuy on the specific effects of particle surface energy an surface chemistry on granule size an granule mechanical properties for high shear wet granulation have been reporte by Ho et al. [58]. Using stanar -mannitol power as well as a silanise hyrophobic version of -mannitol, they prouce a series of power mixtures with ientical power ensity, particle shape an particle size istributions. IGC establishe that the silanise an nonsilanise powers ha very ifferent surface energetics (see Fig. 8). Measurement of the particle size istributions for the final granules prouce as well as stuy of the compressive mechanical properties of the final granules showe significant ifferences which were ascribe to the wetting phase of the granulation processes (see Fig. 9).

10 2686 Current Pharmaceutical Design, 2015, Vol. 21, No. 19 Daryl R.Williams Table III. an acceptor an onor inexes (K A an K B ) of the ifferent mannitol powers at 303 K [60]. Process (mjm -2 ) K A K B Jet mille 47.9 ± ± ± 0.12 pray rie 60.3 ± ± ± 0.23 CLIJ 85.3 ± ± ± 0.29 Fig. (8). surface energy heterogeneity istributions measure for -mannitol [58]. POWDER PREPARATION METHOD A relatively early stuy of API milling an crystal surface energy was reporte by Heng et al. [59]. The effects of milling an particle size on surface energies of form I paracetamol crystals were reporte for both particulate as well as macroscopic crystal materials. Milling of form 1 crystals resulte in fracture along the crystal s lowest attachment energy plane (010), exposing facets of ifferent surface chemistry to that of the native external facets. Milling resulte in the exposure of a more hyrophobic surface for paracetamol form I crystals which becomes increasingly more ominant in its presence with ecreasing particle size; inee for mille samples increase by 20% with ecreasing particle size. A comparison of mannitol particle prepare by iffering methos was reporte by Tang et al. [60] for aerosol elivery applications. Methos use inclue spray rying, air jet milling as well as a confine liqui impinging jet approach (CLIJ). Table III above shows the significant ifferences on both an aci/base surface energies reporte for these iffering types of mannitol. This stuy, in the context of the current review, highlights how sensitive surface properties are to particle manufacturing metho. Moi et al. [61] has reporte the effect of crystal habit an milling on the intrinsic issolution behavior of celecoxib. Celecoxib power compacts forme from planar crystals exhibite higher wettability than the acicular, mille acicular or mille planar power compacts. This enhance wettability manifeste itself in much higher levels of aciic surface energy for the planar crystals which was confirme with higher surface concentrations of nitrogen containing species as etermine by XP. These same crystals ha an intrinsic issolution rate which was 20% higher than the 3 other celecoxib samples stuie. Another stuy of milling an particle shape was reporte by Ho et al. [62]. They examine both ball mille an unmille versions of the form of -mannitol, an consiere iffering facture pathways for this material-see Fig. 10. Milling was reporte to change crystal aspect ratio which was quantifie using ynamic image analysis. Analysis of the surface energy istributions of these powers showe some very clear trens. From Fig. 11, it is clear that is relate to the crystal shape of mille -mannitol. As the BR crystal aspect ratio increases, the profiles isplay a ownwar tren towar lower values of. The mean values, as measure from the istributions of surface energy epicte in Fig. 11, ecreases from 44.7 to 42.1 mjm -2 (6% ifference) as the aspect ratio increases from 0.39 to This reuction in energy was ascribe to the exposure of crystal plane (011) of -mannitol neeles upon milling. This small reuction in the mean values of is rather significant ue to the fact that the absolute ifference between (010) an (011) is less than 5 mjm -2. Comparing mille an unmille -mannitol powers at the ientical aspect ratio, it is clear that the mille materials exhibit consistently smaller particle breaths, implying that mille -mannitol neeles are shorter in length than unmille -mannitol. INHALABILITY OF AEROOL PARTICLE There are a number of particle properties which coul be expecte to control the performance of <5m iameter particles use as active therapeutics for eep lung eposition. Aerosolisation of such powers epens upon range of properties incluing particle size, topography, shape, ensity, surface chemistry an surface energy, an this topic with a specific focus on lactose particle has Fig. (9). Cumulative granule size istribution of -mannitol formulations an the Young s compressive mouli for single 1mm iameter granules for increasing egrees of power mixture hyrophobicity [58].

11 Particle Engineering in Pharmaceutical olis Processing Current Pharmaceutical Design, 2015, Vol. 21, No Fig. (10). Diagrams representing the fracture scenarios of -mannitol with a neele morphology: (a) fracture irection across (011) an (b) fracture irection across (010) [62]. been reviewe by Pilcer et al. [63]. Raula et al. [64] has examine the aerosol performance for l-leucine coate salbutamol sulphate particles. They conclue that particle surface roughness an particle morphology (buckle or spherical) was more important than, though they i note performance correlations with particle surface aciicity. In a stuy esigne to look at how inhalable particle surface properties coul be change, Davies et al. [65] prepare salbutamol sulphate particles via a number of crystallisation routes. This inclue a traitional cooling crystallisation, an anti-solvent crystallisation as well as crystallisation in the presence of moel pulmonary surfactants uner carefully controlle conitions. A number of these approaches ha a significant impact on the istributions, incluing most interestingly the anti-solvent crystallisation. This process resulte in a very heterogeneous surface energy istribution, though the surfactant coate particles also ha higher istributions compare to most other materials. MILLING A large number of stuies have reporte significant ifferences in the surface energy of mille pharmaceutical powers. A set of seminal references were publishe in the mi 1990 s by Ticehurst an co-workers on this topic [66-68]. Using IGC they establishe the changes in for mille of salbutamol-sulphate, l-propranolol hyrochlorie an -lactose monohyrate materials [66]. In the case of inustrial salbutamol-sulphate samples they showe that two materials subjecte to iffering milling regimes iffere by a factor of 10 approximately in surface area, with the high surface area material having a of 83 mjm -2 whilst the low surface area soli ha a value less than of this. These workers have sensibly packe both IGC samples with the same sample surface area, so that the IGC ata represent vapour asorption measurements at the same surface coverage for both samples. They also establishe that the aci-base asorption behaviour of both samples also iffere significantly. Using the same experimental approach these workers also stuies a series of nominally equivalent batches of -lactose monohyrate [67]. They conclue that all values were in the range of mjm -2. They i note small changes in the values for sp G A for a series of aci-base probes. However, it is unclear if sp G A these ifferences in are experimentally significant. Finally, York et al. [68] examine 5 sets of l-propranolol hyrochlorie powers, incluing 4 jet mille samples. As particle size ecrease from 75 to 8.3 m, increase from 45 to 61 mjm -2. Though for 6.6 m, slightly ecrease to 57 mjm -2 ; error bars an uncertainties were not reporte, so this ifference may or may not be significant. The authors also reporte some fairly clear linear trens sp in G A for ichloromethane an THF asorption as a function of log particle size. The surface properties etermine were iscusse in terms of the ifferent crystal facets present on l-propranolol hyrochlorie an their attachment energies as preicte by computational moels. An early stuy on the milling of -lactose monohyrate an surface energy as etermine using IGC was reporte by Newell et al. [69]. These workers examine ifferences in ue to processing inuce isorer so as to unerstan the effects of isorer on lactose particle surfaces. The milling process create 1% of the amorphous lactose, an was 31.2, 37.1 an 41.6 mjm -2 for crystalline, spray rie an mille lactose, respectively. However, a physical mixture of crystalline (99%) an amorphous (1%) material ha a ispersive surface energy of 31.5 mjm -2. Newell et al. [70] then stuie changes in the surface energy of amorphous -lactose monohyrate as a function of relative humiity. ecrease with increasing %RH for amorphous spray rie -lactose monohyrate, whereas a mille -lactose monohyrate exhibite the same trens, though the values were about 5 mjm -2 higher. This stuy is significant as it establishes that the percentage of amorphous material present implies no information about the istribution of amorphous material in a power. In this stuy a mixture of 1% of particles which were purely amorphous was compare to a power which ha the 1% total amorphous content probably sprea across the surface of many particles via milling. These two 1% amorphous powers ha iffering surface values with the 1% physical mixture having effectively the of the unmille material. This work also serves to remin us that IGC ata is most sensitive to the sur-

12 2688 Current Pharmaceutical Design, 2015, Vol. 21, No. 19 Daryl R.Williams Fig. (11). profiles of mille -mannitol sieve fractions as a function of bouning rectangular aspect ratios an, obtaine by area uner curve metho, for mille -mannitol sieve fractions [62]. face area of the sites available for vapour asorption, not their bulk mass present. That is, IGC is generally a surface base phenomena first an foremost. The milling of cefitoren pivoxil has been stuie by Ohta an Buckton [71]. Material mille for 1 minute ha a of 56.2 mjm -2 whereas ecrease to 45.8 mjm -2 after 30 minutes of milling. for the crystalline starting material was 52.3 mjm -2. The effect of milling also showe a clear increase in K D /K A values with increasing milling time, inicating that the power surface was becoming more basic uring milling. Changes in K D /K A with increasing milling time were irectly correlate to sample crystallinity- see Fig. 12 Fig. (12). The relationship between crystallinity an K D /K A ratios for cefitoren pivoxil [71]. Das an tewart have examine the changes in as well as aci-base surface energy istributions for both mille an unmille inomethacin [72]. They also observe that the surface energy istributions for the mille materials gave higher curves than the as receive unmille material. ignificant ifferences in acibase surface properties were also reporte by the authors for the ifferent material forms. A stuy on the surface energy for a series of ifferent milling methos have been reporte by Luner et al. [73] for succinic aci an sucrose. The specific focus of the stuy was to ientify the impact of high shear wet milling in the presence of ifferent solvents, compare to ry impact milling. Fig. 13 below summaries the major finings of this work. Wet milling create the high surface energy solis followe by the ry mille solis an finally the crystalline starting materials. ucrose wet-mille in ethanol showe the highest, 90 mjm -2, which is almost 3-fol higher than that of the unmille particles. uch significant changes were not expecte an as of yet cannot be fully rationalise, though the authors have consiere a number of potential explanations. The surface energy an solubility of ursoeoxycholic aci powers prouce following vibratory milling has been reporte by Chung et al. [74]. Contact angle ata inicate that materials which were vibration mille ha much higher polar surface energies, an thus surface polarities, compare to both jet mille an as receive powers. In general, a correlation between increase rug solubility, ecrease crystallinity with increase surface polarity was reporte. Gamble et al. [75] have investigate the impact of micronisation on the measure surface energy characteristics of Ibipinabant. Micronisation le to an increase in the measure of the rug substance with increasing micronisation energy. The increase in was conclue to be ue to the generation of new, higher energy surface sites. The blening of micronise salbutamol sulphate (B) with ifferent concentrations of magnesium stearate (Mgst) an glycerol monostearate (GM) was followe by co-milling with an air jet mill was recently reporte [76]. This metho coul be use for the surface moification of inhalable rugs particles so as to change inter-particulate forces balance between active ingreient an carrier particles. However, it shoul be note that the reliable interpretation of IGC ata for such complex multicomponent power/chemical systems is clearly more problematic ue to the multiple potential surface species being sense. tank an teckel [76] conclue that of B is lowere using such a process an that the istributions are more homogenous for the co-mille samples (see Fig. 14): In a recent stuy Olusanmi et al. have examine the etaile power properties for a set of mille APIs [77]. They showe that API surface energies increase irectly with the extent of power milling. An examination of the power blening process reveale

13 Particle Engineering in Pharmaceutical olis Processing Current Pharmaceutical Design, 2015, Vol. 21, No Fig. (13). urface energy of wet mille succinic aci an sucrose [73]. that the power wall friction angle was irectly correlate with. It was conclue that the ata presente in this stuy inicates that surface energetics may not be the main factors governing the blening of particulate materials. POWDER ENGINEERING The surface properties of magnesium stearate coate -lactose monohyrate powers as prepare by a mechanical ry processing metho was reporte by Zhou et al. [78]. The untreate an the mechano-fuse lactose particles visually appeare very ifferent uner EM, with the mechano-fuse materials having a 30% lower BET surface area. surface energy istributions showe some heterogeneity for the mechano-fuse lactose particles, with the high coverage being 37 mjm -2, very close to of pure magnesium stearate; 36 mjm -2. The untreate materials exhibite a higher but very constant of 45 mjm -2. Data was also reporte for aci-base surface energies, though these estimates seem to be very high. This may be ue to using IGC ata from low surface overages to estimate high surface coverage properties. For energetically homogeneous surfaces such an analysis may be acceptable, but for energetically heterogeneous surfaces this coul easily lea to overestimates of aci-base properties which are typically very surface coverage epenent. A similar stuy has been reporte by waminathan et al. [79] for magnesium stearate coate -lactose monohyrate powers an microcrystalline cellulose (MCC). These authors reporte the more conventional single values for of their powers at nominal infinite ilution IGC conitions of 66.8, 41.1 an 40.6 mjm -2 for MCC, lactose an magnesium stearate respectively. For the MCCmagnesium stearate power mixtures, change with blen op- erating conitions, trening to lower values of similar to plain magnesium stearate with increase mixing times. The authors conclue that coul be use a measure of a successful coating process. A wor of caution shoul be ae here. For these power mixtures both excipient an magnesium stearate particles are present an therefore changes in measure reflect changes the numbers of the iffering surface sites present on which vapour asorption can occur. In this stuy this coul be because of goo blening (when the lubricant was homogenously isperse) but coul also occur by simply increasing in the surface area of lubricant particles present. Das an tewart examine the changes in surface energy istributions for mechano-fuse magnesium stearate with -lactose monohyrate as well as Turbula mixture of the same powers [72, 80]. In both cases istributions for -lactose monohyrate- magnesium stearate mixtures iffere from the original starting materials. Interesting the mechano fuse -lactose monohyrate ha a lower surface energy istribution than any other power incluing the -lactose monohyrate an magnesium stearate. This ifference was argue to be possibly ue to surface group reorientation in the magnesium stearate film which occurre uring particle compression/processing. Data was also reporte for aci-base surface energies, though these estimates seem to be very high, possible ue to the use of low surface coverage ata being use. Fig. 15 shows the significant variations in istributions which can result from mechanofusion. The surface energy istributions were foun to be consistent with the observe flow behaviour of these surface treate powers urface energy istributions have been recently reporte for an API power system coate with two ifferent high surface area io 2 powers [81]. This stuy specifically consiers the use of IGC ata for binary power systems an the authors showe that iffering istributions were obtaine for the uncoate power, as well as iffering API-iO 2 power systems. This very etaile stuy consiers the important question as to whether the surface of the API particles is passivate by the presence of the io 2 powers. Due to the competing of effects of surface energy, number of surface sites present an affinity of the vapour asorption probes for these iffering sites on the istributions obtaine, precise inter-

14 2690 Current Pharmaceutical Design, 2015, Vol. 21, No. 19 Daryl R.Williams Fig. (14). surface energy istributions, B-Mgst co-mille [76]. pretation of such complex phenomena can be challenging. The authors conclue in this case that the API surface was not being passivate by the io 2 powers present. taile surface energy measurements show that an API such as ibuprofen, being crystalline, has higher surface energy as well as increase surface energy heterogeneity as it was mae into finer particles via milling. In contrast, ry coating with nanosilica uniformly reuce the surface energy to a point that the surface energy of 5 m ibuprofen particles is almost the same as the original 58 m ibuprofen particles using hyrophobic silica. This silica appears to quench high-energy sites an reuce surface energy, which is one of the two major roles playe by ry-coate silica. This stuy highlights the importance of surface energy heterogeneity information as key ata in unerstaning the behavior of complex power mixtures. Though this conclusion is in irect contrast to that of Gamble et al. [79], it highlights the sensitivity of IGC for stuying these types of complex power phenomena. The flowability of a series of surface treate Al powers have been reporte by Jallo et al. [83]. In their extensive investigation they looke at surface energy, power flow, interparticle ahesive forces. Fig. 16 shows the correlation between power ahesion an aci-base surface energetics they reporte. Fig. (15). istributions for Lactose (Ptos) before an after Ptos mechanofuse with 0.1%, 1%, 2%, 5% an 8% magnesium stearate (Mgt) [80]. A key question was recently aresse via a systematic investigation of the influence of ry coating on surface energy of mille pharmaceutical crystals [82]. The observe very high cohesion of mille powers contrast significantly to the improvements observe after ry coating of such materials with nanoparticles. Uncoate micronise ibuprofen showe increase istributions with increasing milling intensity, in contrast, ry-coate mille powers showe a significant reuction in the values. The authors reporte that physical mixtures of uncoate micronise ibuprofen an silica exhibite no reuction in surface energy, inicating that ry coating is necessary for the passivation of high-energy sites of ibuprofen create uring micronisation. pecifically, ry coatings lea to ecrease cohesion an improve flowability because of reuce levels of an the creation of nano scale surface roughness. De- Fig. (16). The relationship between the Angle of Repose an surfaces basicity/aciity. The more basic the aluminum powers, lower the angle of repose which correlate to better flowability [83]. Burnett et al. [84] have investigate the effects of quench cooling an ball milling on the surface energy heterogeneity an surface chemistry of crystalline inomethacin (IMC). The surface energy istributions of crystalline, quench coole an mille IMC samples are shown below in Fig. 16. The quench coole material was fully

15 Particle Engineering in Pharmaceutical olis Processing Current Pharmaceutical Design, 2015, Vol. 21, No amorphous, whereas the mille an crystalline materials processe no amorphous character. ata shows that all IMC samples are energetically fairly heterogeneous, meaning the surface energy changes as a function of surface coverage. Mille IMC is eviently the most active surface with the highest surface energy values an highest surface basicity, having higher values across all surface coverages measure. The quench coole IMC an crystalline IMC ha lower values. Quench coole an crystalline IMC samples possesse wier variations of, highlighting their more heterogeneous surface chemistry (Fig. 17). a shift in the energy istribution of ispersive as well as specific component to higher values, compare to the other two samples. These changes are ascribe to amorphous omains present on the mille crystal surfaces. Co-crystals are an interesting class of materials which are potential caniates for inhalable rugs. Theophylline co-crystals were prepare with urea (THF-URE), saccharin (THF-AC) an nicotinamie (THF-NIC) by spray rying by Alhalaweh et al. [86]. These powers, as well as some mille materials, were then characterise in terms of the physicochemical properties as well as their inhalation performance when aerosolise from a ry power inhaler (see Table. IV). Compare to mille theophylline co-crystals, the novel spray rie co-crystals emonstrate consierably lower values of. Table IV. urface Energy Properties of pray Drie (D) Theophylline Co-crystals an Mille [86]. Fig. (17). profiles of three ifferent IMC samples; crystalline, ball mille an melt quenche [84]. Thielmann et al. [85] have reporte on istributions for mille, recrystallize an as receive lactose materials with very similar particle size istributions. Fig. 18 below shows the istributions obtaine for the three materials: Fig. (18). istributions for crystalline, mille an recrystallise lactose powers [85]. The as receive an recrystallize lactose showe a similar values of 's at infinite ilution, though the istributions for the two samples are quite ifferent. For the mille material there is Facet specific surface energetics of large single crystals of mefenamic aci have been relate to the surface energetic profiles of the corresponing power samples recently by hah et al. [87, 88]. Mefenamic aci crystals with ifferent crystal habits ranging from elongate neeles to hexagonal cubois were obtaine by varying the crystallisation conitions. profiles of mefenamic aci powers obtaine by IGC for varying crystal habits an were correlate to the relative exposure of ifferent crystal facets. Elongate neele shape mefenamic aci crystals were foun to be about eight times more cohesive as powers relative to hexagonal cube shape crystals. This power cohesivity was attribute to the combine effect of ifferences in the API s surface energy an the surface area Contributions from crystal shape on power cohesion of mefenamic aci was ecouple from that of surface energy an surface area using a normalisation proceure. Hexagonal cuboi or elongate plate shape crystals was foun to be less cohesive compare to the neele shape crystals. OTHER PROPERTIE TRACKED BY URFACE ENERGY Hasegara et al. [89] have use surface energy an nett retention volume obtaine using IGC to follow structural relaxations at the surface of amorphous soli ispersions of inomethacin an PVP. Nett retention volumes ecrease by ~15% over a perio of 30 hours reflecting changes in the free energy of the system. Traitionally these types of relaxation processes have been successfully evaluate using DC an isothermal microcalorimetry. These later two approaches measure bulk thermoynamic properties, an specifically enthalpically observable events. By contrast IGC has both a surface focus as well being a technique which measures free energy properties as its primary ata set. tructural relaxations clearly pertain to the system moving to a lower free energy state, so the sensitivity of IGC to the surface relaxations is consistent with expectations. IGC showe faster relaxation kinetics to that etermine by DC, inicating that surface relaxations occurre more quickly than bulk relaxations. A novel use of IGC has been reporte by Miyanishi et al. [90]. They have followe the surface crystallisation of melt-quenche mixture of crystalline nifeipine an polyvinylpyrrolione using the