Heterogeneous Crystallization of Fenofibrate onto ... - ACS Publications


Heterogeneous Crystallization of Fenofibrate onto...

0 downloads 82 Views 4MB Size

Subscriber access provided by - Access paid by the | UCSB Libraries

Article

Heterogeneous Crystallization of Fenofibrate onto Pharmaceutical Excipients Raquel Arribas Bueno, Clare M. Crowley, Peter Davern, B. Kieran Hodnett, and Sarah Hudson Cryst. Growth Des., Just Accepted Manuscript • DOI: 10.1021/acs.cgd.7b01598 • Publication Date (Web): 21 Feb 2018 Downloaded from http://pubs.acs.org on February 26, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Crystal Growth & Design is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 38 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Crystal Growth & Design

Heterogeneous Crystallization of Fenofibrate onto Pharmaceutical Excipients Raquel Arribas Bueno, Clare M. Crowley, Peter Davern, B. Kieran Hodnett, and Sarah Hudson Synthesis and Solid State Pharmaceutical Centre, Department of Chemical Sciences, Bernal Institute, University of Limerick, Limerick, Ireland [email protected] ABSTRACT The crystallization of fenofibrate (FF) from methanol (MeOH) was carried out in the presence of the following dispersed excipients: α/β-Lactose (α/β-Lac), D-Mannitol (D-Man), microcrystalline cellulose, carboxymethyl cellulose (CMC), silica (SiO2) and polycaprolactone (PCL). More control was achieved over the nucleation and crystal growth of the FF particles in the presence of excipients relative to its conventional crystallization using FF seed. Each of the excipients was found to strongly reduce the FF induction time during its crystallization from supersaturated MeOH solutions relative to the rate observed in the absence of the excipients; there was a pronounced reduction in the induction time for FF from > 22 hours in the absence of excipients to ca. 15 minutes in their presence at optimum conditions. Additionally, the FF particle size can be optimized by adjusting the FF loading (% w/w) and the crystallization temperature. The dissolution rate of the small FF particles generated via crystallization in the presence of excipients was comparable to the dissolution rate of the ground commercial FF (Lipantil® Supra) and was faster compared to that of the FF crystallized in the presence of seed. Thus,

1 Environment ACS Paragon Plus

Crystal Growth & Design 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 38

the process parameters of heterogeneous crystallisation in the presence of pharmaceutical excipients can reduce induction times and control API particle size. KEYWORDS: Heterogeneous crystallization, Fenofibrate, Excipient, Dissolution, Induction time, Particle size. INTRODUCTION It is estimated that at least 40% of active pharmaceutical ingredients (APIs) are poorly water soluble, leading to problems such as poor and highly variable bioavailability 1. Common solid formulation approaches to overcome these difficulties and improve the dissolution rates of these drugs include increasing specific surface area of API crystals 2, co-crystal formation 3, polymorphism, saltformation, reduction in crystallinity, and the addition of excipients 2. Among these techniques, increasing surface area by reducing the crystal particle size and/or modifying the crystal habit are very common approaches 4, 5. There are two general approaches to particle size reduction, namely the ‘bottom-up’ or the ‘topdown’ approach. ‘Top-down’ processes involve disintegration methods, such as milling and homogenization 2 and are the most commonly used in the industrial context. However, these techniques are energetically costly 6. In the literature, there are several ‘bottom-up’ approaches to produce small crystals including: controlled crystallization and precipitation after solvent removal via evaporation or freeze-drying 7, 8, the ‘hydrosol’ method 9, supercritical fluid methods 10, 11, cryogenic spray processes 12, and anti-solvent precipitation 8. However, ‘bottom up’ approaches can produce an amorphous material or an undesired polymorphic form which can create problems for long term stability 4. In addition, they can be difficult to scale up. Thus, the ‘bottom-up’ approach has not yet been established as a successful commercial technology 13. That said, the last two decades have seen a shift from empirical formulation efforts to an engineering approach based on a better understanding of particle formation in these processes which may enable future commercial development 14.

2 Environment ACS Paragon Plus

Page 3 of 38 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Crystal Growth & Design

In pharmaceutical formulation, several processing steps can be involved, such as feeding, blending, milling, wet or dry granulation, drying, tableting and coating. In an attempt to improve manufacturing methods and minimize the number of processing steps required a ‘bottom up’ approach is developed here15. API particles of a poorly water-soluble drug are produced in the low micron range (< 15 µm) via the cooling heterogeneous crystallization of the API in the presence of dispersed excipient particles. If amenable to scale-up, this proposed method offers the potential to eliminate the need to (a) mill bulk APIs, and (b) blend milled APIs with excipients.

Figure 1: Chemical structure of FF Table 1: Different polymorphs and characteristics of FF crystals 16, 17

Polymorph Space group

FI

FII

FIII17

Amorphous

FI

FII

FIII

-

P

-

P

P21/n

a (Å)

8.133

13.619

9.4803

-

b (Å)

8.239

7.554

9.7605

-

c (Å)

14.399

17.880

10.9327

-

beta

105.75

92.35

90.352

-

2.0

4.0

2.0

-

Triclinic

Monoclinic

Triclinic

Amorphous

TADLIU01

TADLIU02

No data

-

Z Symmetry cell setting CCDC Ref code

Fenofibrate (FF), an oral medication used to treat high cholesterol levels, Figure 1, was selected as the model API for this present study. It is known to have three crystalline low melting point polymorphs in addition to existing in an amorphous form (Table 1)17. Form I is the most stable polymorph, and is used as an API in tablet and capsule formulations. It has high intestinal permeability but poor solubility 3 Environment ACS Paragon Plus

Crystal Growth & Design 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 38

in water. It therefore exhibits irregular absorption in the body and is classified as a Class II drug according to the Biopharmaceutics Classification System (BCS). Specifically, it is poorly soluble in water ( 99.9

Sigma-Aldrich

-

-

-

Hydrochloric acid

HCl

37

Sigma-Aldrich

-

-

-

Fenofibrate

FF

≥ 98

Sigma-Aldrich

Crystalline

TADLIU01

Form I

α/β-Lac

≥ 99

Sigma-Aldrich

Crystalline

BLACTO

α/β

MCC

-

VWR/MERCK

Amorphous

-

-

D-Man

≥ 98

Sigma-Aldrich

Crystalline

DMANTL07

β

CMC

-

MERCK Millipore

Amorphous

-

-

PCL

> 99.9

Sigma-Aldrich

Amorphous

-

-

SiO2

-

Aeroperl 300 Pharma

Amorphous

-

-

Commercial FF

35

BGP Products Ltd.

Nanocrystalline

TADLIU01

Form I

α/β-Lactose (≤ 30% α-anomer) (MW: 342 g/mol) Microcrystalline cellulose (MW: 36000 g/mol) β-D-Mannitol (MW: 182 g/mol) Carboxymethyl cellulose, Sodium salt (MW: (1−2) × 105 g/mol) Polycaprolactone (MW: 14000 g/mol) Silica (MW: 60 g/mol; specific surface area = 294 m2/g; pore size = 20 to 40 nm) Commercial Fenofibrate Tablet ® (Lipantil Supra)

2. METHODS 2.1. Determination of the solubility of FF and the six excipients in MeOH The solubility of FF in MeOH was measured between 10.0 and 30.0 °C, and that of each of the six excipients was measured between 5.0 and 35.0 °C. Each solubility measurement was performed in triplicate according to the following procedure. Excess solids were added to approximately 20 mL of MeOH, placed in a temperature-controlled water bath (± 0.1 °C) at the required saturation temperature (Tsat) and agitated with a magnetic stirrer at 500 rpm for 24 hours. Agitation was then stopped and the suspensions allowed to settle for more than 1 hour. The concentration of the dissolved FF or excipient was then determined using the dry mass method

32

. The amount of solute present in the supernatant,

expressed in terms of concentration as g solute/kg MeOH, was then calculated according to the method of Granberg et al.33. 6 Environment ACS Paragon Plus

Page 7 of 38 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Crystal Growth & Design

2.2. Determination of the induction time and the metastable zone width (MSZW) for the crystallization FF from MeOH Initially, stock solutions of FF in MeOH (86 g/kg MeOH), saturated at 27.0 °C, were placed in a water bath at Tsat + 3.0 °C (30.0 ± 0.1 °C) and agitated at 400 rpm with a PTFE-coated magnetic stirrer for ≥ 12 hours. 20 mL aliquots of saturated solution were then transferred to 25 mL vials, sealed with PTFE-lined lids, using pre-heated syringes and filters (PTFE, 0.2 µm). Thereafter, these filtered aliquots were used in the following induction time and MSZW experiments. 2.2.1. Induction time Seven vials were equilibrated at 30.0 ± 0.1 °C, 200 rpm for ≥ 12 hours prior to quench-cooling to the crystallization temperature (Tcry). Agitation was maintained via a PTFE-coated magnetic stirrer at 200 rpm throughout the isothermal treatment. The induction time (defined as the time when the first crystals of FF are observed to crystallize in the first vial) was measured using a webcam (Microsoft life cam, wide angle f/2.2, HD Lens 720 p HD, 30 FPS, Autofocus widescreen) at Tcry = 21.9 °C corresponding to a FF supersaturation level (S) of 1.5 (S =

, where c = supersaturated

concentration of FF in MeOH in g FF/kg MeOH, and c* = equilibrium concentration of FF in MeOH at Tcry in g FF/kg MeOH). 2.2.2. Metastable zone width The metastable zone width (MSZW) was specifically determined for a FF-MeOH solution saturated at 27.0 ˚C because most of the crystallization experiments in this study were performed at this saturation level. As such, seven vials were equilibrated at 30.0 ± 0.1 °C and 200 rpm for at least 12 hours prior to cooling at a rate of 0.4 °C/min. Agitation was maintained via a PTFE-coated magnetic stirrer at 200 rpm throughout the treatment. The time and temperature at which the contents of each vial crystallized were recorded using a webcam as in Section 2.2.1 and the average temperature for the onset of crystallization was used to determine the MSZW.

7 Environment ACS Paragon Plus

Crystal Growth & Design 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 38

2.3. Nucleation experiments in the presence and in the absence of ‘hetereosurfaces’ Vials containing 20 mL aliquots of the saturated FF-MeOH solutions were initially prepared at Tsat as described in Section 2.2. Each vial served as a discrete sample for a particular time, FF loading (%) and S point, and was discarded after characterization. Following equilibration at Tsat + 5.0 °C, the vials were transferred to a water bath at Tcry. Thereafter, these cooled filtered aliquots were treated as outlined below in order to crystallize FF at different processing conditions in the absence and presence of excipients, and in the presence of FF seed. 2.3.1. In the absence of excipients FF was crystallized from metastable MeOH solutions at S = 1.5 (21.9 ± 0.1 °C, 200 rpm). The concentration of FF in solution was measured, as explained in Section 2.1, prior to the visual onset of nucleation and thereafter at three time points following the observation of nucleation. 2.3.2. In the presence FF seed FF ‘as received’ (Form I; D50 = 360 µm, span = 847 µm) was used as seed, and was added to the supersaturated solution at an amount equivalent to 1 % w/w with respect to the amount of FF available to crystallize at S = 1.5, as defined in Equation 1. Following the addition of seed, the solution was agitated at 200 rpm with a PTFE-coated magnetic stirrer and held isothermally at Tcry.

Equation 1 where: mseed

= mass of seed (g)

c

= initial concentration of FF prior to addition to the seed (g FF / kg MeOH)

c*

= equilibrium concentration of FF at Tcry obtained from solubility data (g FF / kg MeOH)

mmethanol

= mass of MeOH (kg)

w/w % = mass of seed / (mass of available FF + mass of seed)

8 Environment ACS Paragon Plus

100

Page 9 of 38 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Crystal Growth & Design

2.3.3. In the presence of excipients The required quantity of excipient (either CMC, D-Man, α/β-Lac, MCC, PCL or SiO2) was added to the supersaturated FF solution. The resultant suspensions were agitated at 700 rpm with a PTFE-coated magnetic stirrer and held isothermally at Tcry. The following factors, deemed likely to influence the crystallization of FF in the presence of the excipients, were examined: (a) the FF-excipient contact time, (b) the FF supersaturation level, (c) Tcry, and (d) the amount of FF available to crystallize in the presence of an excipient where all supersaturation is consumed via either heterogeneous or homogeneous nucleation, i.e. the FF loading (% w/w), as defined in Equation 2: Equation 2 where: mexcipient

= mass of excipient (g)

w/w % = mass of available FF / (mass of available FF + mass of excipient)

100

Table 3 summarizes the various parameter combinations examined during the crystallization of FF from metastable MeOH solutions in the presence of FF seed or each of the six excipients. Table 3: The various combinations of Tsat, Tcry, amount of excipient or seed added, amount of FF available to crystallize, % FF loading and S examined for the crystallization of FF from supersaturated MeOH solutions in the presence of different added ‘hetereosurfaces’. The D50 of the excipients and seed is included (n = 15 to 57) D50 of the Amount of FF Amount of Added excipient Tsat Tcry Supersaturation FF loading excipient or seed available to excipient or or seed (°C) (°C) (S) (% w/w) crystallize (g) seed added (g) (µ µm)

FF seed*

330 ± 250

27.0

21.9

1.5

0.45

0.005

CMC D-Man

100 ± 30 250 ± 100

α/β-Lac

100 ± 80

MCC

100 ± 20

PCL SiO2

400 ± 100 20 ± 10

16.0 27.0 16.0 27.0 27.0 27.0 16.0 22.0 16.0 27.0 27.0

10.0 21.9 10.0 21.9 21.9 21.9 10.0 17.5 10.0 21.9 21.9

1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.1 1.4 1.5 1.5

0.23 0.45 0.23 0.45 0.45 0.45 0.23 0.08 0.23 0.45 0.45

3.1 3.3 3.1 0.86 0.86 4.1 3.1 1.1 0.56 6.0 2.8

*: added at 1 % w/w with respect to the amount of FF available to crystallize at S = 1.5 9 Environment ACS Paragon Plus

not applicable 7 12 7 35 35 10 7 7 29 7 14

Crystal Growth & Design 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 38

2.3.4. Characterization of the supernatant and the isolated FF-excipient composite solids Following the required contact time, agitation was stopped and the suspended solids were allowed to settle. The supernatant and solid fraction were separated by vacuum filtration (using a Büchner funnel and 2.5 µm cellulose filter paper), and the solid fraction was dried in an oven at 50 °C and atmospheric pressure to a constant weight (> 24 hours). a. Quantification of the percentage of FF desupersaturated from solution The concentration of FF remaining in the supernatant, expressed as g FF/kg MeOH, was determined as described in Section 2.1. From this, the percentage FF desupersaturation was calculated using Equation 3:

Equation 3 where: csupernatant = concentration of FF remaining in the supernatant (g FF/kg MeOH) c

= supersaturated concentration of FF (g FF/kg MeOH)

c*

= equilibrium concentration of FF at Tcry (g FF/kg MeOH)

b. Analysis of the isolated FF-excipient composite solids i. Powder X-ray diffraction Powder X-ray diffractograms (PXRD) were recorded on a Phillips PANanalytical X'Pert MPD PRO diffractometer using a Cu radiation source (λ=1.541 Å) at 40 mA and 40 kV. Scans were performed between 5 - 40° 2θ at a scan rate of 2.13° 2θ/min. ii. SEM The habits of the MeOH-washed excipients, the FF seed and the isolated FF-excipient composite solids were examined by SEM (JCM-5700 and JSM-6510LV (JEOL)). Samples were gold-coated (SI50B, Edwards) and the surface appearance of MeOH-washed excipients, recrystallized FF, FF crystallized in the presence of FF seed, and isolated FF-excipient composite solids were compared. 10 Environment ACS Paragon Plus

Page 11 of 38 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Crystal Growth & Design

The mean particle sizes of the MeOH-washed excipients, the FF seed and FF crystallized in the presence of the ‘heterosurfaces’ were also measured from the micrographs using image analysis (Adobe Measurement Tool) with 15 to 57 particles being assessed per composite depending on the number of FF particles observed per image. All SEM images shown are representative of the majority of particle sizes and habits observed per sample. iii. In situ SEM-Raman Spectroscopy Micro-Raman measurements were performed on an InVIA Reflex spectrometer (Renishaw) coupled to an optical microscope (DM2500, Leica) and an SEM (JSM-6510LV, JEOL) (the latter being referred to as the SEM-SCA (SEM-Structure & Chemical Analyser)). Instrument calibration was performed using the Si (100) peak (520.5 ± 1 cm−1). Spectra were acquired using the 785 nm laser, variable laser power (0.1 – 10 mW), acquisition times (10 – 500 s) and accumulations (1 – 20) over the spectral range of interest. Spectra collection and processing were performed with the WIRE™ 4.1 software (Renishaw). Samples were gold-coated (SI50B, Edwards) after acquisition of the Raman spectra and prior to obtaining the SEM images. iv. Solid state NMR Carbon-13 solid-state nuclear magnetic resonance (SSNMR) spectra were acquired on a Bruker Avance III HD NMR spectrometer operating at B0 = 9.4 T, with corresponding 1H and

13

C

resonance frequencies of ν0(1H) = 400.1 MHz and ν0(13C) = 100.6 MHz. Fenofibrate samples were packed in 4 mm o.d. zirconia rotors with Kel-F caps under ambient atmosphere, and experimental 13

C NMR spectra were acquired at natural abundance using a 4 mm triple channel (H/X/Y) Bruker

MAS probe operating in double resonance mode. The magic angle was optimised using a rotor packed with KBr and spun at 5 kHz. NMR spectra were referenced to TMS at δiso = 0 ppm by setting the high frequency

13

C resonance in adamantane to 38.48 ppm34.

The

13

C Cross

Polarization Magic Angle Spinning (CP/MAS) NMR spectra were acquired in a single spectral window using the cross-polarization pulse sequence, with a magic angle spinning (MAS) rotor 11 Environment ACS Paragon Plus

Crystal Growth & Design 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 38

frequency of 10 kHz, a 1H 90° pulse width of 2.5 µs, and 50 kHz 1H decoupling during acquisition. Proton decoupling was carried out with the SPINAL6435 decoupling sequence at 100 %. For each sample the 1H T1 relaxation time(s) were checked using the saturation recovery pulse sequence to ensure that the recycle delay allowed for adequate relaxation between the collection of subsequent transients. 13C CP/MAS spectra were collected using optimised contact times (4 ms and 5 ms for FF and SiO2-FF composite samples, respectively) and relaxation delays (2 s and 5 s for for FF and SiO2-FF composite samples, respectively - at least 1.4 x T1 values). 15,000 scans were collected for the SiO2-FF composite sample and 124 scans for pure ‘as received’ FF (Form I). v. Dissolution rate studies The dissolution rates of the various isolated FF-excipient composite solids in 0.1 M HCl with 0.4% w/v Tween-80 were determined at sink conditions of approximately one fifth of the equilibrium FF solubility concentration; these were compared with the corresponding dissolution rates for a ground tablet of a commercial formulation of FF, namely Lipantil® Supra, 145 mg tablet, BGP Products Ltd. The tablet was ground via mortar and pestle. The dissolution rate of a physical mixture of FF (D50 = 7 µm, span = 26 µm, produced by grinding FF via mortar and pestle) and MCC was also determined as a control sample. A solution of 450 mL of 0.1 M HCl with 0.4% w/v Tween-80 was placed in a water bath at 42.0 °C for 24 hours and used as the dissolution medium. These conditions were used to allow for comparison of the determined dissolution rates with those previously reported 8. Following equilibration, the appropriate amount of sample containing 12.5 mg of FF in each case was added to the dissolution medium (450 mL). 1 mL aliquots were withdrawn every 5 minutes during the first 30 minutes and then every 15 minutes thereafter. A 1 mL aliquot was also taken after 24 hours (i.e. complete dissolution) to determine the actual concentration of FF in the original sample. The concentration of dissolved FF was measured with reference to the UV absorbance for FF at λ = 289 nm (Cary 300 Bio). 12 Environment ACS Paragon Plus

Page 13 of 38 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Crystal Growth & Design

3. RESULTS AND DISCUSSION 3.1. Solubility of FF and the six excipients in MeOH and determination of the MSZW The experimentally determined solubility of FF, Figure 2, compares favorably with that previously reported by Watterson and Rasmuson 36. The solubility of FF in MeOH at 25 °C was at least 53 times greater than that of the most soluble excipient (D-Man), Figure 2. As such, the solubilities of all six excipients in MeOH (Table 4) were deemed negligible compared with that of FF at the crystallization conditions used in this study. The MSZW, measured at a cooling rate of 0.4 °C/min, for FF-MeOH solutions saturated at 27.0 °C was 12.5 ± 5.1 °C, Figure 2. This corresponds to a maximum attainable supersaturation of 2.22, after which FF crystallizes spontaneously. Promotion of heterogeneous nucleation and suppression of homogeneous nucleation requires that the supersaturated FF-MeOH solutions be held within this MSZ, in other words, at a supersaturation of less than 2.22.

S = 1.5; 21.9° C

S = 1.1; 17.5° C

S = 1.5; 10° C

Figure 2: (i) Experimental solubility values in MeOH for FF between 10.0 and 30.0 °C (♦) and for the excipients between 5.0 and 30.0 °C (■, ∆, ●,▲,+,-), (ii) the estimated MSZ limit (dashed red line) based on an extrapolation through the

×

experimentally determined MSZW data point ( ) for FF in MeOH for a saturation temperature (Tsat) of 27.0 °C and at cooling rate of 0.4 °C/min. The points ▲, ♦, and ■ represent the different processing conditions (in terms of S and Tcry) used during the experiments to crystallize FF from MeOH

13 Environment ACS Paragon Plus

Crystal Growth & Design 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 38

Table 4: Solubility of the six excipients in MeOH at 25.0 °C Excipient

Solubility (g/kg MeOH)

D-Man

1.42 ± 0.02

α/β-Lac

0.68 ± 0.003

MCC

0.06 ± 0.007

CMC

0.16 ± 0.018

PCL

0.0007 ± 0.00

SiO2

0.0001 ± 0.00

3.2. Induction time and extent of desupersaturation of FF from MeOH solution in the absence of excipients During the induction time experiments (20 mL scale, S=1.5), FF was not observed to crystallize in less than 22 hours; indeed, just two out of seven vials nucleated in 40 hours. The % desupersaturation in the first vial to nucleate was quantified following visual observation of the first nucleation event. At 15 minutes post-nucleation, 79 % desupersaturation was observed, and full desupersaturation was achieved 45 minutes later, i.e. 1 hour post-nucleation. The rate of desupersaturation was seen to decrease considerably over time as ever-less supersaturation remained to ‘drive’ the crystallization to completion (i.e. to equilibrium saturation).

14 Environment ACS Paragon Plus

Page 15 of 38 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Crystal Growth & Design

3.3. Crystallization of FF in the presence of excipients

See d; 1 % w/w ; 21.9 α/β °C Lac ; 35 %

MC DC; Ma 35 n; %; 12 21.9 %; °C 21.9 °C CM C; 7 %; 10.0 °C

SiO 2; 14 %; 21.9 °C PCL; 7 %; 21.9 °C

No FF seed; 21.9 °C



Figure 3: Desupersaturation of FF-MeOH solutions in the absence of seed from induction time experiments ( ), and in the presence of the following ‘hetereosurfaces’ during 2 hours of contact: (■) suspended FF seed (1% w/w); (♦) α/β-Lac, (■) MCC and (▲) D-Man at S = 1.5, Tcry = 21.9 °C and 12 % w/w loading; (●) CMC at S = 1.5, 7% w/w loading, Tcry = 10.0 °C, (-) SiO2 at S = 1.5, Tcry = 21.9 °C and 14 % w/w loading, and (×) PCL at S = 1.5, Tcry = 21.9 °C and 7 % w/w loading. RTlnS, the thermodynamic driving force, was in a range of 954 to 994 J.mol-1 for all the experiments.

(a) α

β

(b)

(c)

(d)

(e)

(f)

15 Environment ACS Paragon Plus

Crystal Growth & Design 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 38

Figure 4: Chemical structure of the excipients: (a) α/β-Lac, (b) MCC, (c) D-Man, (d) CMC, (e) PCL and (f) SiO2

The addition of α/β-Lac, D-Man, MCC, CMC or SiO2 to a supersaturated solution of FF in MeOH reduced the induction time of FF in the 20 mL scale from greater than 22 hours (observed in the absence of additives) to less than 15 minutes (Figure 3). This contrasts sharply with our previous observations for the crystallization of AAP and carbamazepine (CBMZ) in the presence of α/β−Lac at 20 mL scale and at supersaturations lying within the metastable zone (at S = 1.25 and S = 1.34, respectively)

27, 32

. In these cases, the reduction in induction times in the presence of

excipients was more modest. For AAP in MeOH at S = 1.25 the induction time of the first vial (out of 20 vials) was 2 hours in the absence of excipients and less than 30 minutes in the presence of α/β-Lac. Whereas, for CBMZ in MeOH at S = 1.34 the induction time of the first vial (out of 20 vials) was 3 hours in the absence of excipients and less than 2 hours in the presence of α/β-Lac. Each molecule of FF has 3 hydrogen bond acceptors37 and no hydrogen bond donors (Figure 1), whereas α/β-Lac, D-Man, MCC, CMC and SiO2 each have many hydrogen bonds donors and acceptors (Figure 4). By contrast the AAP molecules feature both hydrogen bond donors and acceptors. For FF the hydroxyl group hydrogens of the excipients offer an additional type of interaction capable of hydrogen bonding with the oxygens of FF’s ketone, ester and ether functional groups, arguably enhancing the prospect for FF’s nucleation onto the excipients’ surfaces from supersaturated solutions. This functional group complementarity, in particular the addition of hydrogen bonding capability, could reduce the free energy barrier to nucleation and as a consequence lower the induction time or increase the number and lifetimes of pre-critical size clusters. In addition, α/β-Lac presents a rough surface with a high density of grooves, while SiO2 (specific surface area = 294 m2/g) is highly porous 33; both of these attributes may further promote the heterogeneous nucleation of FF. α/β-Lac, D-Man, MCC, CMC and SiO2 thus have the capacity to promote the nucleation of FF through functional group matching and/or surface roughness. Conversely, PCL (which only possesses hydrogen bond acceptors) has no such functional group 16 Environment ACS Paragon Plus

Page 17 of 38 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Crystal Growth & Design

compatibility with FF and is therefore not expected to promote nucleation via this mechanism. It may still promote heterogeneous nucleation of FF via its rough surface (as seen below in the SEM image in Figure 12a) . The above observations suggest that α/β-Lac, D-Man, MCC, CMC or SiO2 possess ample hydrogen bond donor ability to overcome the MeOH solvent molecules’ ability to solvate FF molecules (also via hydrogen bond donation). In so doing, over time these excipients are therefore able to sequester FF molecules to their surfaces as single molecules or in clusters large enough to ultimately facilitate nucleation. Literature studies have concluded there is a longer lifetime of hydrogen bonding interactions between a molecule in solution and a solid surface than between two molecules in solution

38-40

. Thus, we propose that the interaction of a cluster with a polar or

hydrogen bonding group on a macromolecular solid surface should increase the average lifetime of the interaction compared to dispersive forces, allowing more stable packing of hydrogen bonded molecules and thus increasing the cluster lifetime 39, 40. In water for example, the average lifetime of the interaction between liquid water molecules and solid ice molecules is some value between water molecules in an ice crystal (10-4 s) and that of water liquid molecules (10-11 s)

41

. This can be

extrapolated to our system which consists of FF molecules in solution, an excipient surface, and MeOH molecules. Unlike water crystallization, FF molecules, since they do not have hydrogen bond donors, can only interact with each other via Van der Waals interactions. These interactions are weaker than hydrogen bonds, resulting in lifetimes in the order of 0.8 to 3 ps42, 43. In addition, FF molecules can also interact with MeOH molecules by hydrogen bonding, resulting in lifetimes in the order of 1 to 70 ps41, 43. Finally, the lifetime of interactions between CO2 hydrates and Silica surfaces calculated by Bai et al.

44

using molecular dynamics simulations were taken as an

estimation of a hydrogen bond lifetime between an excipient surface and a FF molecule. The lifetime of those interactions ranged between 1 to 160 ns.Thus, FF molecules in solution can strongly interact with OH groups on the excipient surface for a longer period of time than with 17 Environment ACS Paragon Plus

Crystal Growth & Design 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 38

molecules of MeOH or FF in solution. As the formation of a nucleus generally appears to take in the order of 10-6 to 10-7 s 45-47, the interaction between FF molecules in solution would not be long enough to facilitate the formation of a FF nucleus. However, the interactions between an excipient solid surface and the FF molecules in solution could endure long enough for the formation of an FF nucleus. As a result, this apparent arrangement of molecules on the excipient surfaces may be the factor responsible for the decrease in the induction time compared to that observed in the absence of excipients. This hypothesis is depicted in Scheme 1.

Scheme 1: Schematic representation of the possible interactions during heterogeneous crystallization of FF from MeOH in the presence of α/β−Lac. The scheme shows a single hydrogen bond between a FF molecule and a hydrogen bond donor in the excipient having a much longer lifetime than hydrogen-bonds or Van der Waals interactions in solution. Moreover, the lifetime of the excipient – FF hydrogen bond is such as to allow the generation of a cluster composed of FF molecules which can ultimately grow into a fully mature FF crystal.

The addition of FF seed produced the most dramatic reduction in induction time, with almost complete FF desupersaturation (98.5%) being achieved after 15 minutes (Figure 3). As such, when compared with all of the other heterosurfaces examined, FF seed likely provides the most suitable template for FF nucleation in terms of lattice matching. In the cases of α/β-Lac, D-Man, MCC and CMC, nucleation of FF was also observed 15 minutes after their addition. However, the extent of 18 Environment ACS Paragon Plus

Page 19 of 38 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Crystal Growth & Design

FF desupersaturation was lower at this time compared with the seed: α/β-Lac (70 %), D-Man (63 %), MCC (49 %) and CMC (28 %). Following nucleation, the initial rate of FF desupersaturation in the presence of α/β-Lac, D-Man and MCC was broadly comparable with that observed in the presence of FF seed. Thereafter, even though the corresponding rates of desupersaturation decreased somewhat, ca. 90% FF desupersaturation was still achieved after 1 hour of total contact time with each of these three excipients. By contrast, the crystallisation of FF in the presence of CMC was accompanied by a comparatively lower initial rate of FF desupersaturation, and a total contact time of 2 hours was required to achieve ca. 90% desupersaturation; this may be a consequence of the gel that CMC forms when suspended in MeOH, which may have hindered the diffusion of FF through the solution. The desupersaturation profile for FF in the presence of SiO2 differs from those of the above four excipients in so far as it has a slightly slower nucleation rate and a lower rate of desupersaturation, and requires a slightly longer contact time to achieve ca. 90% FF desupersaturation. In terms of rationalizing these observations, the results obtained following SEM and 13C SSNMR analysis of the related composite solids (reported later in this paper) support the view that the crystallization of FF occurred predominantly within the pores of the SiO2, which pore sizes range 20 to 40 nm. As such, the integration of FF into the SiO2 will depend on the absorption properties of the SiO2. In this regard, the mass transfer of FF will depend on the SiO2 mass resistance, which in turn will depend on film mass transfer, porous diffusion and finally fixation 48. Thus, the desupersaturation profile for FF in the presence of SiO2 may be explained by poor porous diffusion. Finally, the crystallization of FF in the presence of PCL was accompanied by the poorest desaturation profile of all the excipients examined over the 2 hour contact time. This is likely due to functional group mismatch in so far as PCL and FF only possess hydrogen bond acceptors; as such, PCL possesses no hydrogen bond donors to complement FF’s hydrogen bond acceptors, and viceversa. Despite this, ca. 20 – 30% FF desupersaturation was observed after 2 hours of contact time. This may be due to the roughened surface topography that PCL exhibits, as

19 Environment ACS Paragon Plus

Crystal Growth & Design 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 38

evidenced from the results obtained following SEM analysis of the related composite solids (reported later in this paper, Figure 12a).

Figure 5: Desupersaturation of FF-MeOH solutions at S = 1.5, V=20 mL in the presence of MCC during 2 hours of contact: (♦) Tcry = 10.0 °C and 7 % w/w loading, and (■) Tcry = 21.9 °C and 35 % w/w loading. RTlnS, the thermodynamic driving force, was in a range of 954 to 994 J.mol-1.

Additionally, in the specific case of MCC, no pronounced change in the rate of desupersaturation was observed when performing the crystallization at a combination of high FF loading and Tcry (35% and 21.9 °C, respectively) versus at a combination of low FF loading and low Tcry (7% and 10 °C, respectively) (Figure 5). This may be rationalized on the basis that even though different FF loadings were used in each experiment, the processing conditions were such that both experiments were performed at the same FF supersaturation (S = 1.5). As such, the difference in crystallization temperature for the two experiments meant that the thermodymanic driving force for nucleation (defined as RTlnS) decreased only marginally from 994 J/mol (for Tcry = 21.9 °C) to 954 J/mol (for Tcry = 10 °C). Therefore, if heterogeneous nucleation behaves as expected from previous reported results

32

, a degree of control may be exercised over the nucleation and growth rates during FF

crystallization and thus over the final FF particle size, without any change on the desupersaturation rates. 20 Environment ACS Paragon Plus

Page 21 of 38 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Crystal Growth & Design

3.4. Control over FF particle growth 3.4.1. Examination of FF particle sizes for a range of excipients The habit and particle size of the FF crystallized in the presence of the various ‘hetereosurfaces’ were compared with those obtained for the ‘as received’ FF seed and the MeOH-washed excipient particles (Figure 6-12). Copious amounts of FF crystals were formed shortly after seed addition at S=1.5 (Figure 3 and Figure 6b). The cause of the formation of smaller and better formed particles than the original seed particles was a consequence of secondary nucleation that was promoted by the seed at this supersaturation

49

. SEM micrographs of the isolated composite solids after the FF

crystallization in the presence of α/β-Lac, MCC, D-Man, CMC and PCL show the formation of FF particles clearly attached to the surface of the excipient particles (Figure 7b, Figure 8b, Figure 9b, Figure 10b and Figure 12b), with few particles of FF found independently of the excipient. However, no distinct particles of anything other than SiO2 could be detected in the presence of this excipient, even at the highest magnification (Figure 11b); this suggests that the FF may have crystallized inside the pores, which have a size range of 20 to 40 nm50. The presence of FF Form I in all the composite solids at all FF loadings and Tcry examined, was confirmed by PXRD, except in the presence of SiO2, where no FF-specific peaks were detected at normal running conditions. FF (a) see d

(b)

Figure 6: SEM micrographs of: (a) the ‘as received’ FF seed, and (b) the FF crystals produced following crystallization in the presence of FF seed (1% w/w) , S = 1.5, Tcry = 21.9 °C, t = 15 min.

21 Environment ACS Paragon Plus

Crystal Growth & Design 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

α/ βLa c

(a)

Page 22 of 38

(b)

Figure 7: SEM micrographs of (a) MeOH-washed α/β-Lac, and (b) α/β-Lac after the crystallization of FF at S=1.5 in the presence of α/β-Lac at 35% w/w loading, 21.9 °C, t = 2h 30. × = FF particles

M CC

(a)

(b)

Figure 8: SEM micrographs of (a) MeOH-washed MCC, and (b) MCC after the crystallization of FF at S=1.5 in its presence at 35% w/w loading, 21.9°C, t = 30 min. × = FF particles

DMa n

(a)

(b)

Figure 9: SEM micrographs of (a) MeOH-washed D-Man, and (b) D-Man after the crystallization of FF at S=1.5 in its presence at 12% w/w loading, 21.9 °C, t = 2h. × = FF particles

C M C

(a)

(b)

Figure 10: SEM micrographs of (a) MeOH-washed CMC, and (b) CMC after the crystallization of FF at S = 1.5 in its presence at 7% w/w loading, 21.9 °C, t = 2h 30. × = FF particles

22 Environment ACS Paragon Plus

Page 23 of 38 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Crystal Growth & Design

Si (a) O2

(b)

Figure 11: SEM micrographs of (a) MeOH-washed SiO2, (b) after the crystallization of FF at S=1.5 in its presence at 14% w/w loading, 21.9 °C, t = 1h 30

PC L

(a)

(b)

Figure 12: SEM micrographs of: (a) MeOH-washed PCL, (b) after the crystallization of FF at S=1.5 in its presence at 7 % w/w loading, 21.9 °C, t = 1h 30. × = FF particles

Further characterization of the FF-SiO2 composite solids was conducted via PXRD at a longer exposure time and over a shorter range (0.093° 2θ/min; 13.5 – 17.5° 2θ), in order to improve the signal to noise ratio. The XRD pattern exhibited a peak that corresponded to FF Form I. Solid state NMR analysis indicates that the chemical environment of the carbon atoms in the FF-SiO2 composite solid corresponds to that of FF Form I, the same as that found in pure crystalline FF (Figure 14). The absence of peak broadening supports the view that the amount of FF present as an amorphous form is below the limit of detection. Similarly, in the study performed by Dwyer et al.

20

SiO2 with the same pore size, was found to be crystalline as determined by XRD.

23 Environment ACS Paragon Plus

the FF loaded onto

Crystal Growth & Design 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 24 of 38

Figure 13: Comparison of the PXRD diffractograms of (i) FF from CCDC database (TADLIU) (ii) Physical mixture of FF and SiO2, 14 % FF loading (ii) the isolated composite solids following the addition of SiO2 to a supersaturated FFMeOH solution after 1 h 30, 14 % FF loading (iv)‘as received’ SiO2.

(b)

(a)

Figure 14: CP/MAS 13C Solid State NMR spectra of (a) FF as received (Form I) (b) the isolated composite solids following the addition of SiO2 to a supersaturated FF-MeOH solution (FF loading (%) = 14 %)

Further analysis of the SEM micrographs via particle size measurements (Figure 15) revealed the rapid growth of the FF particles over time in the presence of all the excipients, except SiO2. The 24 Environment ACS Paragon Plus

Page 25 of 38 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Crystal Growth & Design

growth of the FF particles occurred during the first 30 minutes of excipient contact time, coinciding with the most pronounced rates of desupersaturation (Figure 3). Thereafter, the FF mean particle size either remained constant or decreased slightly, likely due to the shearing effect during mixing 49. The final mean particle size among the different excipients varied between 40 to 87 µm. Plausible explanations for this varability in the FF particle size are the differences in the excipients surface area and in the FF loading. In comparision, the FF particles formed in the absence and in the presence of FF seed were 80 ± 92 and 73 ± 50 µm, respectively, 30 minutes after the onset of nucleation.

α/β-Lac

35 % loading

35 % loading 12% loading 7 % loading 7 % loading

Figure 15: Mean FF particle size at different time points from 15 minutes to 2 hours after the crystallization of FF from MeOH solutions at 21.9 °C and S = 1.5, and in the presence of the following excipients (and % FF loadings): ♦ MCC (35% w/w), ■ α/β-Lac (35% w/w), ▲CMC (7% w/w), × D-Man (12% w/w) and ● PCL (7% w/w).

MCC and α/β-Lac were selected as the model excipients for further studies as both induced a strong reduction in the induction time of FF during its crystallization from MeOH solutions. 3.4.2. The combined influence of FF loading and crystallization temperature on the particle size of FF The combined effect of FF loading and crystallization temperature had a significant influence on the FF particle size obtained during its crystallization from MeOH solutions in the presence of MCC or α/β-Lac (Figure 16). Firstly, the combination of a low FF loading and a low crystallization 25 Environment ACS Paragon Plus

Crystal Growth & Design 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 26 of 38

temperature, i.e. 7 % FF loading and 10.0 °C, led to the formation of small particles of FF with a D50 of ca. 3 µm on the surface of both MCC and α/β-Lac (Figure 16a, 17e). As determined by SEMRaman, these particles occasionally presented as immature crystals that appeared to coat the excipient particles (Figure 17). This coating was only observed at this combination of processing conditions (i.e. low FF loading and low Tcry). Performing the crystallization at a higher FF loadings of 29 % yet while maintaining a crystallization temperature of 10.0 °C (Figure 16c) generated larger, more wellformed FF particles, with a D50 of ca. 50 µm. FF did not coat the MCC or the α/β-Lac in the composite solids prepared under these conditions. When the crystallization temperature was increased to 21.9 °C while maintaining a low FF loading of 10 % (Figure 16b), well-formed FF particles (with a D50 of ca. 35 µm and a span of 54 µm) were again observed after the crystallization in the presence of MCC. Again, no ‘FF coating’ was observed here. Finally, when the crystallization was performed at a combination of high FF loading (35 %) and at a high crystallization temperature (21.9 °C), the FF particles became significantly bigger, with a D50 of ca. 75 µm and a span of 107 µm (Figure 16d). In the presence of α/β-Lac (Figure 17f), large FF particles with a D50 and a span of 81 µm and 72 µm respectively were also observed at high loading (35 %) and high temperature (21.9 °C). Figure 17 and Figure 18 compare the SEM-Raman analysis of the FF-MCC composite solid isolated using the high FF loading / high crystallization temperature combination with that from a FFMCC composite solid generated using the combination of low FF loading and low crystallization temperature. The presence of the 1647 cm-1 peak, in the carbonyl stretching region, characteristic of FF Form I, indicates that the particles formed on the MCC particles surface are FF Form I. In addition, the nature of the coating formed around the MCC particles observed in the SEM images of the isolated composite solids after the crystallization of FF in the presence of MCC at 7% FF loading and 10.0 °C was also determined to be Form I. No carbonyl stretching vibrations shift at 1656 cm-1, which are characteristic of amorphous FF 51, were observed.

26 Environment ACS Paragon Plus

Page 27 of 38 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Crystal Growth & Design

The above observations may be explained in terms of the interplay between the nucleation and growth of FF crystals on the surface of the excipients, and on the assumption that the lower Tcry of 10 °C is still sufficient to overcome the activation energy barrier for nucleation of FF crystals. This assumption, combined with the comparable nature of the thermodynamic driving forces for FF’s nucleation at both crystallization temperatures, means that nucleation will likely occur as readily at the lower Tcry as it will at the higher Tcry. The small FF particles and immature FF crystal ‘coatings’ observed only at low loading / low Tcry may therefore be a consequence of the slower growth rate of FF nuclei on the surface of the excipients due to (i) the slower diffusion of FF molecules to the developing nuclei at the lower Tcry, since diffusion is proportional to temperature52, and (ii) the relatively lower concentration, and thus reduced availability, of FF molecules at this low loading to feed this slower growth rate. In comparison with seeded crystallization which showed the formation of large FF particles (D50 = 80 µm) (Figure 6), heterogeneous crystallization can form smaller FF particles. By selecting specific process conditions, i.e. 7 % loading and 10.0 °C, a 25-fold reduction in the FF particle size in the best-case scenario has been achieved. It can be concluded that the combined effect of FF loading and crystallization temperature can have a significant impact on the particle size of FF crystals generated during its crystallization from MeOH solutions in the presence of excipients, with the smallest FF particles forming at a combination of low loading and low crystallization temperature.

27 Environment ACS Paragon Plus

Crystal Growth & Design 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 28 of 38

1. (a) Low loadings

(b)

(c) High loadings

(d)

Low temperatures

High temperatures

2. Low loadings and low temperatures

(e)

High loadings and high temperatures

(f)

Figure 16: SEM micrographs after the crystallization of FF at S = 1.5. 1. in the presence of MCC at 2 h (a) % FF loading = 7%, Tcry =10.0 °C (b) % FF loading = 10 %, Tcry =21.9 °C (c) % FF loading = 29 %, Tcry =10.0 °C, and (d) % FF loading = 35 %, Tcry =21.9 °C. 2. in the presence of α/β-Lac at 2 h (e) % FF loading = 7%, Tcry =10.0 °C (f) % FF loading = 35 %, Tcry =21.9 °C (× = FF particle).

28 Environment ACS Paragon Plus

Page 29 of 38 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Crystal Growth & Design

(a) Spot 1

Spot 2

Spot 3

Spot 4

Spot 5

Spot 6

(b) 20000 Spot 6 0 20000 Spot 5 0 20000 Spot 4 0 20000 Spot 3 0 20000 Spot 2 0 25000 Spot 1 0 120000

FF 'as received'

0 1525 1550 1575 1600 1625 1650 1675 1700 -1

Raman shift (cm )

Figure 17: (a) SEM micrographs and (b) Raman spectra at different spots (× = spot) of an isolated composite solid particle obtained following the crystallization of FF from MeoH in the presence of MCC at S = 1.5, % FF loading = 7%, Tcry =10.0 °C after 30 min.

29 Environment ACS Paragon Plus

Crystal Growth & Design 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 30 of 38

(a) Spot 2

Spot 1

1 2

Spot 3

Spot 4 2

1

(b) 40000 Spot 4 15000

Spot 3 (2)

15000 Spot 3 (1) 15000

Spot 2 (2)

15000 Spot 2 (1) 15000 Spot 1 15000 MCC 'as received' 0 125000

FF 'as received'

0 1525 1550 1575 1600 1625 1650 1675 1700

Raman shift (cm-1) Figure 18: (a) SEM micrographs and (b) Raman spectra at different spots (× = spot) of an isolated composite solid particle obtained following the crystallization of FF from MeOH in the presence of MCC at S = 1.5, % FF loading = 35%, Tcry = 21.9 °C after 30 min

3.5. Dissolution rate studies During dissolution studies of selected samples, the quantity of FF dissolved after 24 hours did not differ from the value based on the expected loading (Figure 19). The initial dissolution rate of FF after 30 Environment ACS Paragon Plus

Page 31 of 38 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Crystal Growth & Design

its crystallization from MeOH in the presence of MCC or D-Man at 21.9 °C and at loadings of 35 and 14 %, respectively, where FF particle sizes formed were bigger (D50 = 75 µm; span = 107 µm and 35 µm; span = 105 µm, respectively) was slow (0.24 and 0.38 mg/L.min, respectively) compared with the ground tablet of commercial FF (11.75 mg/L.min) (Table 5). Similar initial dissolution rates were observed after the crystallization of FF by seeding and in the absence of excipients (0.23 and 0.32 mg/L.min, respectively), as the FF particle size was also big (D50 = 80 µm; span = 92 µm and 73 µm; span = 50 µm, respectively). However, FF crystallized in the presence of MCC or α/β-Lac at 7% of FF loading and at 10.0 °C, corresponding with the smallest FF particle sizes formed (D50 = 3 µm; span = 23 µm and 3 µm; span = 5 µm, respectively), present faster initial dissolution rates (0.87 and 0.92 mg/L.min, respectively). The dissolution testing of the physical mixture of milled FF (D50 = 7 µm; span = 26 µm) and MCC has a similar initial dissolution rate (0.72 mg/mL.min) to that of the FF after the crystallization in the presence of MCC and α/β-Lac at 7% FF loadings and 10.0 °C, as the FF particle size was similar. This confirms that the enhancement in the initial dissolution rate was mainly due to the reduction in the FF particle size and that the excipients did not give any additional improvement. The initial dissolution rate was found to be inversely proportional to particle size (Figure 20). The formation of smaller FF particles, with larger surface areas, explains this enhancement in the initial dissolution rate 2

. However, for all composites of FF with MCC, α/β-Lac or D-Man, the initial dissolution rate is slower

than for the commercial FF. Interestingly, a significant improvement in the dissolution rate was observed for the SiO2-FF composites, having an initial dissolution rate of 6.92 mg/mL.min with 79 % of the FF dissolved in 10 min. This SiO2-FF composites presents the best dissolution result of all the tested samples almost matching with the commercial nanomilled FF, which initial dissolution rate was 11.75 mg/L.min having 90.8 % of the FF dissolved in 10 min. The major advantage of mesoporous SiO2 used as excipient for the heterogeneous crystallization of FF lies in its pore size, pore morphology, and in the surface functional groups, which result in optimized interactions between the FF and the mesoporous

31 Environment ACS Paragon Plus

Crystal Growth & Design 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 32 of 38

SiO2 surface53. Furthermore, FF can be loaded into these pores in a nanocrystalline state, increasing the drug dissolution properties dramatically. Table 5: Summary of the composites used for the dissolution rate studies including the conditions of their crystallization and the initial dissolution rates. The D50 of the crystallized FF is included (n = 15 to 57) FF loading Initial dissolution rate T t Composite S D50 (wt-%) (mg. L-1.min-1) from 0-5 min (°C) (min) ● Ground tablet of 34 Nanomilled 11.75 commercial FF ■ 1.2 14 21.9 1.5 6.92 FF - SiO2 ♦ 1.5 7 10.0 15 2±5 0.92 FF - α/β-Lac ■

FF - MCC

×

Physical mixture (milled FF-MCC) FF - D-Man FF recrystallized in the absence of excipients FF - MCC FF recrystallized in the presence of FF Seed

● + ▲ –

1.5

7

10.0

30

3 ± 23

0.87

1.5

-

-

-

7 ± 26

0.72

1.5

14

21.9

15

35 ± 105

0.38

1.5

100

21.9

1440

73 ± 50

0.32

1.5

35

21.9

30

75 ± 107

0.24

1.5

100

21.9

30

80 ± 92

0.23

Figure 19: The dissolution profiles of the FF present on the composites shown in Table 5. Dissolution medium: 450 mL of 0.1 M HCl-0.4% w/v Tween 80 at 42.0 °C. Sink conditions: 12.5 mg FF.

32 Environment ACS Paragon Plus

Page 33 of 38 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Crystal Growth & Design

Figure 20: Initial dissolution rate of FF, from 0 to 5 minutes at different D50. Dissolution medium: 450 mL of 0.1 M HCl-0.4% w/v Tween 80 at 42.0 °C. Sink conditions: 12.5 mg FF.

4. CONCLUSIONS It can be concluded that the incorporation of FF seed or suitable excipient surfaces with functional group complementarity with FF into a supersaturated solution forces the FF to nucleate at a more rapid rate than in their absence. This faster rate could be due to a longer lifetime for hydrogen –bond interactions between FF molecules adsorbed onto a solid excipient surface than for all solution phase hydrogen –bonding or van der Walls interactions in solution. The fastest rate of desupersaturation was found in the presence of FF seed. However, crystallization via seeding leads to the formation of large particles of FF as growth is the dominant mechanism. By contrast, when crystallized in the presence of the excipients, the combination of FF loading (%) and crystallization temperature can be adjusted to make nucleation the dominant mechanism and thus promote the generation of smaller FF particles. Due to their reduced particle size, these small FF particles displayed an enhanced initial dissolution rate relative to that observed for FF particles generated in the presence and absence of seed. SiO2-FF composites obtained by this approach presents the best dissolution result of all the tested samples almost matching with the commercial nanomilled FF. 33 Environment ACS Paragon Plus

Crystal Growth & Design 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 34 of 38

5. ACKNOWLEDGMENTS This publication has emanated from research conducted with the financial support of the Synthesis and Solid State Pharmaceutical Centre, funded by Science Foundation Ireland (SFI) under Grant Numbers 12/RC/2275, as well as support from Bernal institute and Department of Chemical Sciences at the University of Limerick.

34 Environment ACS Paragon Plus

Page 35 of 38 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Crystal Growth & Design

6. REFERENCES (1) Merisko-Liversidge, E.; Liversidge, G. G.; Cooper, E. R., Nanosizing: a formulation approach for poorly-water-soluble compounds. Eur. J. Pharm. Sci. 2003, 18, (2), 113-120. (2) Khadka, P.; Ro, J.; Kim, H.; Kim, I.; Kim, J. T.; Kim, H.; Cho, J. M.; Yun, G.; Lee, J., Pharmaceutical particle technologies: An approach to improve drug solubility, dissolution and bioavailability. Asian J. Pharm. 2014, 9, (6), 304-316. (3) Janjikhel, R. K.; Adeyeye, C. M., Dissolution of Ibuprofen Enantiomers from Coprecipitates and Suspensions Containing Chiral Excipients. Pharm. Dev. Technol. 1999, 4, (1), 9-17. (4) Savjani, K. T.; Gajjar, A. K.; Savjani, J. K., Drug Solubility: Importance and Enhancement Techniques. ISRN Pharm. 2012. (5) Leleux, J.; Williams, R. O., Recent advancements in mechanical reduction methods: particulate systems. Drug Dev. Ind. Pharm. 2014, 40, (3), 289-300. (6) Parikh, D. M., Handbook of Pharmaceutical Granulation Technology. ed.; Taylor & Francis: 1997. (7) de Waard, H.; Hinrichs, W. L. J.; Frijlink, H. W., A novel bottom–up process to produce drug nanocrystals: Controlled crystallization during freeze-drying. J. Controlled Release 2008, 128, (2), 179183. (8) Tierney, T. B.; Guo, Y.; Beloshapkin, S.; Rasmuson, Å. C.; Hudson, S. P., Investigation of the Particle Growth of Fenofibrate following Antisolvent Precipitation and Freeze–Drying. Cryst. Growth Des. 2015, 15, (11), 5213-5222. (9) Junghanns, J. U. A. H.; Müller, R. H., Nanocrystal technology, drug delivery and clinical applications. Int. J. Nanomedicine 2008, 3, (3), 295-309. (10) Bleich, J.; Müller, B. W., Production of drug loaded microparticles by the use of supercritical gases with the Aerosol Solvent Extraction System (ASES) process. J. Microencapsulation 1996, 13, (2), 131-139. (11) Padrela, L.; Rodrigues, M. A.; Velaga, S. P.; Fernandes, A. C.; Matos, H. A.; de Azevedo, E. G., Screening for pharmaceutical cocrystals using the supercritical fluid enhanced atomization process. J. Supercrit. Fluids 2010, 53, (1–3), 156-164. (12) Rogers, T. L.; Johnston, K. P.; Williams, R. O., Solution-Based Particle Formation of Pharmaceutical Powders by Supercritical or Compressed Fluid Co2 and Cryogenic Spray-Freezing Technologies. Drug Dev. Ind. Pharm. 2001, 27, (10), 1003-1015. (13) Sinha, B.; Müller, R. H.; Möschwitzer, J. P., Bottom-up approaches for preparing drug nanocrystals: Formulations and factors affecting particle size. Int. J. Pharm. 2013, 453, (1), 126-141. (14) Vehring, R., Pharmaceutical Particle Engineering via Spray Drying. Pharm. Res. 2008, 25, (5), 999-1022. (15) Yazdanpanah, N.; Testa, C. J.; Perala, S. R. K.; Jensen, K. D.; Braatz, R. D.; Myerson, A. S.; Trout, B. L., Continuous Heterogeneous Crystallization on Excipient Surfaces. Cryst. Growth Des. 2017. (16) Groom, C. R.; Bruno, I. J.; Lightfoot, M. P.; Ward, S. C., The Cambridge Structural Database. Acta Crystallogr., Sect. B: Struct. Sci., Cryst. Eng. Mater. 2016, 72, (2), 171-179. (17) Tipduangta, P.; Takieddin, K.; Fábián, L.; Belton, P.; Qi, S., A New Low Melting-Point Polymorph of Fenofibrate Prepared via Talc Induced Heterogeneous Nucleation. Cryst. Growth Des. 2015, 15, (10), 5011-5020. (18) Granero, G. E.; Ramachandran, C.; Amidon, G. L., Dissolution and Solubility Behavior of Fenofibrate in Sodium Lauryl Sulfate Solutions. Drug Dev. Ind. Pharm. 2005, 31, (9), 917-922. (19) Bouledjouidja, A.; Masmoudi, Y.; Van Speybroeck, M.; Schueller, L.; Badens, E., Impregnation of Fenofibrate on mesoporous silica using supercritical carbon dioxide. Int. J. Pharm. 2016, 499, (1–2), 1-9. 35 Environment ACS Paragon Plus

Crystal Growth & Design 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 36 of 38

(20) Dwyer, L. M.; Michaelis, V. K.; O'Mahony, M.; Griffin, R. G.; Myerson, A. S., Confined crystallization of fenofibrate in nanoporous silica. CrystEngComm 2015, 17, (41), 7922-7929. (21) Chadha, R.; Bhandari, S., Drug-excipient compatibility screening-Role of thermoanalytical and spectroscopic techniques. J. Pharm. Biomed. Anal. 2014, 87, 82-97. (22) Chadwick, K.; Myerson, A.; Trout, B., Polymorphic control by heterogeneous nucleation - A new method for selecting crystalline substrates. CrystEngComm 2011, 13, (22), 6625-6627. (23) Chadwick, K.; Chen, J.; Myerson, A. S.; Trout, B. L., Toward the Rational Design of Crystalline Surfaces for Heteroepitaxy: Role of Molecular Functionality. Cryst. Growth Des. 2012, 12, (3), 11591166. (24) Liu, X. Y., Heterogeneous nucleation or homogeneous nucleation? J. Chem. Phys. 2000, 112, (22), 9949-9955. (25) Quon, J. L.; Chadwick, K.; Wood, G. P. F.; Sheu, I.; Brettmann, B. K.; Myerson, A. S.; Trout, B. L., Templated Nucleation of Acetaminophen on Spherical Excipient Agglomerates. Langmuir 2013, 29, (10), 3292-3300. (26) Zimmermann, A.; Millqvist-Fureby, A.; Elema, M. R.; Hansen, T.; Mullertz, A.; Hovgaard, L., Adsorption of pharmaceutical excipients onto microcrystals of siramesine hydrochloride: Effects on physicochemical properties. Eur. J. Pharm. Biopharm. 2009, 71, (1), 109-116. (27) Verma, V.; Peddapatla, R. V. G.; Crowley, C. M.; Crean, A. M.; Davern, P.; Hudson, S.; Hodnett, B. K., Experimental Study on the Influence of Excipients on the Heterogeneous Crystallization and Dissolution Properties of an Active Pharmaceutical Ingredient. Cryst. Growth Des. 2018, 18, (1), 338-350. (28) Yang, H.; Song, C. L.; Lim, Y. X. S.; Chen, W.; Heng, J. Y. Y., Selective crystallisation of carbamazepine polymorphs on surfaces with differing properties. CrystEngComm 2017, 19, (44), 65736578. (29) Curcio, E.; López-Mejías, V.; Di Profio, G.; Fontananova, E.; Drioli, E.; Trout, B. L.; Myerson, A. S., Regulating Nucleation Kinetics through Molecular Interactions at the Polymer–Solute Interface. Cryst. Growth Des. 2014, 14, (2), 678-686. (30) Yazdanpanah, N.; Ferguson, S. T.; Myerson, A. S.; Trout, B. L., Novel Technique for Filtration Avoidance in Continuous Crystallization. Cryst. Growth Des. 2016, 16, (1), 285-296. (31) Ebrahimi, A.; Saffari, M.; Dehghani, F.; Langrish, T., Incorporation of acetaminophen as an active pharmaceutical ingredient into porous lactose. Int. J. Pharm. 2016, 499, (1–2), 217-227. (32) Arribas Bueno, R.; Crowley, C. M.; Hodnett, B. K.; Hudson, S.; Davern, P., Influence of Process Parameters on the Heterogeneous Nucleation of Active Pharmaceutical Ingredients onto Excipients. Org. Process Res. Dev. 2017. (33) Granberg, R. A.; Rasmuson, A. C., Solubility of paracetamol in pure solvents. J. Chem. Eng. Data 1999, 44, (6), 1391-1395. (34) Morcombe, C. R.; Zilm, K. W., Chemical shift referencing in MAS solid state NMR. J. Magn. Reson. 2003, 162, (2), 479-486. (35) Fung, B. M.; Khitrin, A. K.; Ermolaev, K., An Improved Broadband Decoupling Sequence for Liquid Crystals and Solids. J. Magn. Reson. 2000, 142, (1), 97-101. (36) Watterson, S.; Hudson, S.; Svard, M.; Rasmuson, A. C., Thermodynamics of fenofibrate and solubility in pure organic solvents. Fluid Phase Equilib. 2014, 367, 143-150. (37) Southan, C.; Sharman, J. L.; Benson, H. E.; Faccenda, E.; Pawson, A. J.; Alexander, Stephen P. H.; Buneman, O. P.; Davenport, A. P.; McGrath, J. C.; Peters, J. A.; Spedding, M.; Catterall, W. A.; Fabbro, D.; Davies, J. A., The IUPHAR/BPS Guide to Pharmacology in 2016: towards curated quantitative interactions between 1300 protein targets and 6000 ligands. Nucleic Acids Res. 2016, 44, (D1), D1054-D1068. (38) Danielli, J. F.; Riddiford, A. C.; Rosenberg, M. D., Recent Progress in Surface Science. ed.; Elsevier Science: 2013.

36 Environment ACS Paragon Plus

Page 37 of 38 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Crystal Growth & Design

(39) Sun, Q.; Harvey, J. A.; Greco, K. V.; Auerbach, S. M., Molecular Simulations of Hydrogen Bond Cluster Size and Reorientation Dynamics in Liquid and Glassy Azole Systems. J. Phys. Chem. B 2016, 120, (39), 10411-10419. (40) Janick, J., Horticultural Reviews. ed.; Wiley: 2010. (41) Astley, T.; Birch, G. G.; Drew, M. G. B.; Rodger, P. M., Lifetime of a Hydrogen Bond in Aqueous Solutions of Carbohydrates. J. Phys. Chem. A 1999, 103, (26), 5080-5090. (42) Smirnov, B. M., Cluster Ions and Van Der Waals Molecules. ed.; Taylor & Francis: 1992. (43) Zhang, X.; Zhang, Q.; Zhao, D.-X., Hydrogen Bond Lifetime Definitions and the Relaxation Mechanism in Water Solutions. Acta Phys.-Chim. Sin. 2011, 27, (11), 2547-2552. (44) Bai, D.; Chen, G.; Zhang, X.; Sum, A. K.; Wang, W., How Properties of Solid Surfaces Modulate the Nucleation of Gas Hydrate. Sci. Rep. 2015, 5, 12747. (45) Mullin, J. W., Crystallization. ed.; Butterworth-Heinemann: 2001. (46) Tiller, W. A., The Science of Crystallization: Microscopic Interfacial Phenomena. ed.; Cambridge University Press: 1991. (47) Institution of Chemical, E., 14th International Symposium on Industrial Crystallization. ed.; Inst of Chemical Engineers UK. (48) Fulazzaky, M. A.; Khamidun, M. H.; Omar, R., Understanding of mass transfer resistance for the adsorption of solute onto porous material from the modified mass transfer factor models. Chem. Eng. J. 2013, 228, 1023-1029. (49) de Souza, B.; Cogoni, G.; Tyrrell, R.; Frawley, P. J., Evidence of Crystal Nuclei Breeding in Laboratory Scale Seeded Batch Isothermal Crystallization Experiments. Cryst. Growth Des. 2016, 16, (6), 3443-3453. (50) Aeroperl ® 300 Pharma Improving the dissolution of poorly soluble APIs. In ed.; Industries, E., Ed. p 12. (51) Heinz, A.; Gordon, K. C.; McGoverin, C. M.; Rades, T.; Strachan, C. J., Understanding the solid-state forms of fenofibrate – A spectroscopic and computational study. Eur. J. Pharm. Biopharm. 2009, 71, (1), 100-108. (52) Genck, W. J., Temperature Effects on Growth and Nucleation Rates in Mixed Suspension Crystallization. ed.; Iowa State University: 1969. (53) Maleki, A.; Kettiger, H.; Schoubben, A.; Rosenholm, J. M.; Ambrogi, V.; Hamidi, M., Mesoporous silica materials: From physico-chemical properties to enhanced dissolution of poorly water-soluble drugs. J. Controlled Release 2017, 262, 329-347.

37 Environment ACS Paragon Plus

Crystal Growth & Design 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 38 of 38

For Table of contents use only

Heterogeneous Crystallization of Fenofibrate onto Pharmaceutical Excipients Raquel Arribas Bueno, Clare M. Crowley, Peter Davern Benjamin, K. Hodnett and Sarah Hudson Synthesis and Solid State Pharmaceutical Centre, Department of Chemical Sciences, Bernal Institute, University of Limerick, Limerick, Ireland

FF molecule MeOH molecule

α/β-Lac crystal

FF crystal

The lifetime of the excipient–fenofibrate surface-solution hydrogen-bond is much longer than the lifetime of hydrogen-bonds or van der Waals interactions in solution and can generate clusters composed of fenofibrate molecules on the surface of the excipient which grow into fully mature fenofibrate crystals. Thus, hydrogen bond donating excipients strongly reduce induction times for supersaturated fenofibrate solutions.

38 Environment ACS Paragon Plus