Environmentally Benign Deagglomeration and Mixing of Nanoparticles in Supercritical CO2

1
Environmentally Benign Deagglomeration and
Mixing of Nanoparticles in Supercritical CO2
Yangyang Shen4, Aleksey Vishnyakov4, M. Silvina
Tomassone4 Program NIRT Award DMI 0506722
PI Dr. Rajesh Davé1Co-PIs R. Gupta2, R.
Pfeffer, S. Sundaresan3, M. S. Tomassone4
1New Jersey Institute of Technology, Newark, NJ,
2Auburn University, Auburn, AL, 3Princeton
University, Princeton, NJ, 4Rutgers University,
New Brunswick, NJ
Motivation
Interparticle Forces
Disjoining Pressure
Nanoparticle Models (SiO2)

• T 77.4K, pbulk 1atm

• FCC structured model
• (MD)
• 276 SiO2 units
• D 2.2 nm

• Spherical shell model
• (GCMC)

derived form the Derjaguen approximation

• Nanoparticles (NP) and nanocomposites have great
  potential to improve performance of drugs,
  biomaterials, catalysts and other
  high-value-added materials. They offer unique
  properties that arise from their small size and
  large surface area.
• A major problem in utilizing nanoparticles is
  that they often lose their high surface area due
  to grain growth or unavailability of the high
  surface area where it matters. It is difficult to
  produce forces required to deagglomerate the
  nanoparticles at a sufficiently small length
  scale.
• The addition of nanoparticles to polymer
  composites has been shown to significantly
  influence the mechanical, optical, and electrical
  properties. However when nanoparticles
  aggregate, they lose their nanoscale size and
  corresponding properties.
• The breakup of nanoagglomerates, driven by the
  tensile stresses generated by depressurization,
  has not been studied previously for nanoparticles
  and there is a paucity of published analysis on
  this subject.

particles at contact
i. dependence on bulk pressure
Disjoining pressure for LJ model at T318K and
different pbulk. Long-range repulsion at pbulk
102 and 200 atm, not observed at 68 atm
large separation
Experimental Sorption Isotherms
ii. dependence on fluid model
Peaks correspond to the pore width when a new
layer is formed and the separation distance is
small

• T 318K, pbulk 68atm

Background

• Supercritical fluids
• Liquid-like density and solubility
• Gas-like diffusivity and viscosity
• An ideal medium for the purpose of
  deagglomerating nanoparticles, because it can
  penetrate the pores within the nano-agglomerates,
  and upon rapid depressurization, can cause
  separation of the nanoparticles
• Supercritical CO2
• Tc 31.1 C
• Pc 7.38 MPa
• Low toxicity
• High stability

Minima correspond to pore width with large
distance between adjacent layers
The disjoining force is repulsive when
nanoparticles are close, then becomes attractive,
and finally diminishes to zero when the
separation is sufficiently large.
Sorption isotherms at 273 K on amorphous silicas
that differ only by hydroxylation level
Sorption isotherms at 195K at different amorphous
silicas
Disjoining pressure for LJ and dumbbell models
with smooth walls (10-4-3 potential only) at
pbulk 68atm Nature of force oscillations on the
width differs in narrow pores. Pronounced
oscillation periodicity when LJ model is used.
However, in general the results are consistent
The attraction is substantially weaker for
dehydroxylated particles
Interactions strongly depend on surface
hydroxylation Surface hydroxylation increases gt
more energetic adsorption sites gt adsorption is
intensified
iii. dependence on surface roughness
The attraction is most prominent for strongly
hydroxylated particles
Solid-Fluid Interaction

• Interaction potential (with spherical shell
  nanoparticle)

Disjoining pressure for dumbbell model with
smooth walls (10-4-3 potential) and inhomogeneous
walls at pbulk 102atm Here, the influence of
inhomogeneities is negligible

• RESS Rapid Expansion in Supercritical Solvents
• The drug is dissolved in the supercritical fluid
• The drug containing supercritical fluid is passed
  through an expansion valve
• The nanoparticles are collected when the
  particles settle on a collection plate

Larger Particles and Surface Inhomogeneity
This potential reduces to the 10-4 form of Steele
potential when R approaching infinity, which
represents a flat surface.

• Solid-fluid potential
• Steeles potential point inhomogeneities

Conclusions
Reduction of pressure causes evaporation of the
supercritical solvent --gt supersaturation of drug
and subsequent precipitation

• Both Monte Carlo and Molecular Dynamics
  approaches were successfully applied to study the
  forces between two spherical silica nanoparticles
  in a supercritical carbon dioxide environment at
  realistic pressures.
• The two models considered (dumbbell and
  one-center LJ) with validated parameters,
  accurately reproduce experimental data on bulk
  CO2 and CO2 sorption on silica.
• Particles effectively attract at the lower
  pressures ranging 68-100 atm and they experience
  repulsive forces for pressures above 100 atm.
• These conclusions do not depend on the molecular
  model considered.
• Energetic inhomogeneities do not significantly
  affect the value of the force between the
  particles.

• Interaction potential (with FCC structured
  nanoparticle)

Point inhomogeneity distance between the solute
molecule and inhomogeneity vs. extra energy added
to the base 10-4-3
Fluid Models (CO2)
Attractive blue repulsive red
Here ?sf and ?sf parameters are chosen to get the
best fit of GCMC and MD isotherms with the Steele
potential to experiment

• Dumbbell with point quadruple
• (Moller Fischer, 1994 and 1997)

• Solid-fluid parameters fitting ?sf and ?sf
  for a surface in the presence of inhomogeneities

? 3.033 Å ?/kB 125.57 K l 0.699 Å Q2/??5
3.0255

• One-center effective Lennard-Jones particle
• ? 3.68 Å
• ?/kB 286.2 K

References
Experiment vs GCMC 195K
303K Critical temperature
Bakaev, V. A., W. A. Steele, et al. (1999).
Journal of Chem. Phys. 111(21) 9813-9821.
Katoh, M., K. Sakamoto, et al. (2000). PCCP
2(19) 4471-4474. Morishige, K., H. Fujii, et
al. (1997). Langmuir 13(13) 3494-3498. Möller,
D. and J. Fischer (1994). Fluid Phase Equilibria
100 35-61. Span, R. and W. Wagner (1996). J. of
Phys. and Chem. Refer. Data 25(6) 1509-1596.
LJ model one set of parameters. LJ model does
not fit experimental isotherm at low temperature.
Dumbbell model different ?sf to account for
hydroxylation
GCMC and experimental isotherms of CO2 (Dumbbell
model) at 195K
GCMC and experimental isotherms of CO2 (Dumbbell
and LJ models) at 303K
The fluid model must exactly reproduce the phase
diagram and thermodynamic properties of the bulk
fluid in the given range of temperature and
pressure.
Contact ? Tel. [phone] ? Fax
1-732-445-2421 ? Email silvina_at_soemail.rutgers.
edu
Funding from NSF NIRT (Award 0506722) and
IGERT (Award 050497) is acknowledged.