(651e) Enhanced Nucleation at Liquid-Liquid Interfaces Due to Local Supersaturation

Authors: 
McKechnie, D., University of Strathclyde
Zahid, S., University of Strathclyde
Mulheran, P. A., University of Strathclyde
Sefcik, J., University of Strathclyde
Johnston, K., University of Strathclyde

Enhanced Nucleation at Liquid-Liquid
Interfaces due to Local Supersaturation

David McKechnie 1,2, Saraf Zahid 1, Paul A.
Mulheran 1, Jan Sefcik 1,3, Karen Johnston 1

1 Department of
Chemical and Process Engineering, University of Strathclyde, 75 Montrose
Street, Glasgow, UK

2 Doctoral Training
Centre in Continuous Manufacturing and Advanced Crystallisation, University of
Strathclyde, Glasgow, UK

3 EPSRC Future
Manufacturing Hub in Continuous Manufacturing and Advanced Crystallisation,
University of Strathclyde, Glasgow, UK

 

Crystallisation is an important
separation process used across a wide range of industries including pharmaceuticals,
fine chemicals, and food. Crystallisation studies are often performed in
small-scale setups that allow for large numbers of experiments to be performed
simultaneously. However, these differ from the typical macroscale
crystallisations carried out in industrial settings in a number of ways. One
key difference is the presence of interfaces at high surface area to volume
ratios. Microfluidic [1] and microwell [2] set-ups are commonly used
for such experiments, investigating small volumes of solution that are
partially or fully surrounded by immiscible liquids. Heterogeneous nucleation
is known to occur at a much greater rate than homogeneous nucleation, and so
the presence of these interfaces at such high surface area to volume ratios
could allow heterogeneous nucleation to dominate the behaviour of the system. The
aim of this work is to study heterogeneous nucleation using a model system of
aqueous glycine solution in contact with tridecane.

Glycine is commonly selected for
nucleation studies as it is the simplest amino acid, is easy to handle, is polymorphic
[3], and has been observed to form mesoscale clusters [4].  Tridecane was used
in previous microwell studies [2] due to its immiscibility with water and low
vapour pressure. We will use a combined experimental and modelling approach to
understand the impact of a liquid-liquid interface on the crystallisation of
glycine compared to crystallisation from aqueous solution.

Experiment

A large number of vials of aqueous
glycine solutions were prepared at concentrations ranging between
275 – 450 g/kgsolvent and monitored by webcam to
determine the isothermal induction times. The nucleation rate in unagitated
samples was found to be negligible at the range of concentrations investigated,
with no crystallisation occurring at any concentration below 450 g/kgsolvent.At 450 g/kgsolvent only 5% of vials crystallised within a
two week period (see Table 1).

Vials with a layer of tridecane
covering the solution were prepared and observed alongside experiments at
concentrations ranging from 275 – 333 g/kgsolvent. A
diagram of the vial setups is shown in Figure 1. The presence of the
liquid-liquid interface resulted in a major increase in the nucleation rate of
glycine. At the lowest concentration investigated, 63% of vials with an oil
interface nucleated within three days, and as the concentration increased the
nucleation rate increased, as expected. This significant increase in nucleation
rate demonstrates the profound effect the interface has on nucleation
behaviour. The percentage of vials that crystallised for each experiment is
presented in Table 1.

 

Figure 1: Diagram of prepared vials with and without the oil - solution interface.

 

" helvetica> 

 

 

 

 

 

Table 1: Percentage of vials that crystallised at each concentration with and without an oil interface.

Concentration

(g/kgsolvent)

Vials Crystallised

without Oil (%)

Vials Crystallised

with Oil (%)

275

0

63

307

0

83

333

0

95

400

0

-

450

5

-

Simulation

Classical molecular dynamics (MD)
simulations were performed to provide insight to the behaviour of the system at
a molecular level that is not currently accessible by experimental techniques.
Simulations of a glycine solution in contact with a layer of tridecane were
performed (see Figure 2 top). Simulations were performed in the NPT ensemble at
298 K and 1 atm. The simulation box was of the dimensions 3.45 x 3.45 x 20 nm
and allowed to adjust in the z direction. 240 glycine, 4000 water and 256
tridecane molecules were simulated, giving a concentration of 250 g/kgsolvent
corresponding to the experimental saturation point for 298 K.

Figure 2: Top: Snapshot of the simulated system: glycine, water and tridecane molecules are blue, red, and green, respectively. Middle: Solution and tridecane density (primary axis) and glycine concentration (secondary axis). The blue dashed line shows the average concentration of the glycine solution (250 g/kgsolvent). Bottom: bond orientation profile of the glycine molecules.

Structural properties were analysed
with respect to distance from the liquid-liquid interface, where dz = z/Lz. Concentration
profiles show an increased concentration of glycine near the interface as shown
in Figure 2 (middle). A concentration of 367 g/kgsolvent  is
obtained at the interface which is approximately 1.5 times higher than the
average concentration of 250g/kgsolvent. This gives a local supersaturation
at the interface that would drive crystal nucleation.

The bond orientation, P2,
of the carbon-carbon bond in glycine molecules with the interface was analysed,
where P2 is defined as:

where θ is the angle of the
bond vector and the z-axis, where z is normal to the interface. The values of P2
range from -0.5 to 1. A P2 of -0.5 indicates the bonds are
oriented parallel to the interface, a value of 1 corresponds to perpendicular
orientation, and a value of 0 represents random orientation of the bonds. The
bond orientation profile is presented in Figure 2 (bottom). The orientation of
the bonds is random in the centre of the film (dz = 0.35 – 0.65), and as the
glycine molecules approach the liquid-liquid interface they order with the
carbon-carbon bond parallel to the interface.

Conclusions

Our MD simulations clearly
demonstrate that oil-solution interfaces have an increased concentration and
ordering of glycine molecules at the interface. This high concentration at the
interface provides the driving force for the increased crystallisation rate observed
in experiments. Future work will explore the effect of the interface on the
solution dynamics. This work highlights the need to account for, or control,
interface effects during the design of experiments.

Acknowledgements

The
authors would like to thank EPSRC and the Future Continuous Manufacturing and
Advanced Crystallisation Research Hub (Grant Ref: EP/P006965/1) for funding
this work. We would also like to acknowledge that this work was carried out in
the CMAC National Facility supported by UKRPIF (UK Research Partnership Fund)
award from the Higher Education Funding Council for England (HEFCE) (Grant ref
HH13054). Computational results were obtained using the ARCHIE-WeSt High
Performance Computer (www.archie-west.ac.uk) based at the University of
Strathclyde.

 

References

-18.0pt">[1]   " helvetica>Ildefonso, M. et al., Organic Process Research &
Development 16, 556-560 (2012); Teychene, S and Biscans, B, Chemical
Engineering Science 77, 242-248 (2012)

-18.0pt">[2]  Little, L. J., et al. Crystal Growth & Design 15,
5345-5354 (2015); Little, L. J., et al. Journal of Chemical Physics 147, 144505
(2017)

-18.0pt">[3]  Boldyreva, E. V., et al.  Journal of thermal analysis and
calorimetry, 73, 409-418
(2003)

-18.0pt">[4]   " helvetica>Jawor-Baczynska, A., et al. Crystal Growth &
Design 13, 470-478 (2013)