British Association For Crystal Growth Annual Conference 2017
Effect of Dispersed Excipients on the Nucleation of Active
Pharmaceutical Ingredients
V. Verma, R. Arribas-Bueno, C. M. Crowley, B. K. Hodnett, S. P. Hudson, P. Davern
University of Limerick, Bernal Institute, Limerick, Ireland
[email protected]
Crystallisation is an attractive purification and isolation step in the production of active pharmaceutical
ingredients (APIs) with required reproducible physicochemical properties. Nucleation, the first step of
crystallisation, is an activated process leading to the assembly of molecules held together by intermolecular
forces such as van der Waals, hydrogen bonds and π-π interaction [1]. The addition of crystal seeds or
heterosurfaces such as excipients to saturated solutions potentially impacts the rate of nucleation and can
be specifically used to manipulate crystal size and polymorph selectivity. The heterosurface or quasi seeds
with similar molecular functionality and lattice matching with APIs can facilitate the nucleation of a particular
polymorph and has been examined previously [2]. However, the actual mechanism by which heterogeneous
nucleation occurs is not very apparent and may be a consequence of epitaxial interactions, nonspecific
adsorption, surface topography and/or intermolecular bonding. For example, Chadwick et al. [3] reported that
the presence of recrystallised α-Lactose monohydrate (α-LMH) and β-D-Mannitol particles in unstirred
paracetamol (AAP)-ethanol solutions reduced the induction time of Form I AAP by 5-fold while a 2-fold
reduction was observed in the presence of dispersed graphite. This reduction in induction time suggested
that functional group complementarity has a significant influence on the ability of organic heterosurfaces to
sequester an incipient nucleus.
The goal of this work is to increase the nucleation rate of APIs via nucleation onto a heterosurface, thus
reducing the crystal size and enhancing the APIs in-vitro dissolution rate. Selecting excipients as the
heterosurfaces will potentially streamline the manufacturing process by avoiding the need to mill/granulate
API batches later on. This study probes the nature of the nucleation (homogeneous or heterogeneous) by
counting the number of particles in the presence and absence of excipient during nucleation. It examines
the degree of API attachment to the excipient particles under the shear of the Labmax impeller and the
influence of this shear and of the excipient on crystal size distribution of APIs.
Fenofibrate (FF, reduces cholesterol levels), paracetamol (AAP, pain reliever and fever reducer) and
carbamazepine (CBMZ, anti-epileptic drug) were selected as model APIs while the excipient microcrystalline
cellulose (MCC) was selected as a heterosurface.All the APIs have solubility of > 50g/L in methanol while
MCC is nearly insoluble, thus methanol was the selected solvent for the crystallisation experiment. A
LabMax reactor system from Mettler-Toledo was used to estimate the nucleation kinetics of all the APIs in
methanol solution in the absence and presence of MCC. A Mettler-Toledo Focused Beam Reflectance
Measurement (FBRM) probe was utilized to monitor the crystal size distribution (CSD) during
desupersaturation events. An in-situ ATR-FTIR probe was used to track solution concentration during
desupersaturation event.
Figure1 illustrates the change in induction time (tind) and nucleation rate assessed by monitoring the total
counts and solution concentration for FF-MeOH solution (S = 1.5) in the absence and presence of dispersed
MCC. In the absence of MCC tind of the FF-MeOH solution was observed to be more than 5 hrs, while
desupersaturation of the solution occurred within minutes in the presence of dispersed MCC, demonstrating
the effect of heterogeneous nucleation. A similar trend was observed with the other two APIs (CBMZ and
AAP), but the influence of heterogeneous nucleation varied for each API. Table 1 summarizes the induction
time for all three APIs in the absence and presence of MCC. The tind for primary homogeneous nucleation of
CBMZ (S = 1.34) was 110 mins and was reduced to 80 mins in the presence of MCC, whilst for AAP (S =
1.09) in the absence and presence of MCC, the tind was 96 and 59 mins, respectively.
British Association For Crystal Growth Annual Conference 2017
Table 1: Induction time of the APIs in the
absence and presence of MCC at a ≥ 500 mL scale
APIs
Excipient
Induction
time (min)
CBMZ
(S=1.34)
AAP
(S=1.09)
FF
(S=1.5)
No Exp
MCC
No Exp
MCC
No Exp
MCC
110 ± 30
80 ± 20
60 ± 16
25 ± 5.5
360 ± 60
8±3
-1
Fig. 1: FBRM counts and IR peak area (1600 cm ) of FF in the
absence and presence of MCC; Tsat = 25 ̊C, Tcry = 21.9 ̊C, S = 1.5,
35%-Loading
SEM micrographs of the solids isolated following 100% desupersaturation of the API solutions in the
presence of MCC demonstrates strong interfacial attachment of AAP and CBMZ with MCC particles but not
between the FF crystals and MCC. The crystal size of all three APIs crystallised in the presence of MCC
remained unaffected by the presence of MCC and varied between 20-100 μm. Large crystal size is known
to adversely affect the dissolution rate of APIs. By varying agitation time, supersaturation, API loading and
changing impeller design, the size of AAP and CBMZ crystal was reduced to 5-60 μm without influencing the
final solid form. In summary, the presence of dispersed MCC in metastable API-methanol solutions
reduced the tind of the APIs examined with a relative decrease in tind of FF >>> AAP > CBMZ.
Although the exact mechanism by which MCC influences the nucleation rate is unclear, this study
suggests that functional group complementarity and a reduction in the surface tension at the MCC
solution interface may be responsible. Tableting of the isolated solids is currently being examined.
References:
[1] Davey, R. J., et al. (2002). "Crystal engineering - nucleation, the key step." CrystEngComm 4(47): 257-264.
[2] Diao, Y., et al. (2011). "Surface design for controlled crystallization: the role of surface chemistry and nanoscale pores
in heterogeneous nucleation." Langmuir 27(9): 5324-5334.
[3]Chadwick, K., et al. (2012). "Toward the Rational Design of Crystalline Surfaces for Heteroepitaxy: Role of Molecular
Functionality." Crystal Growth & Design 12(3): 1159-1166.