Available online at www.sciencedirect.

com

Advanced Drug Delivery Reviews 60 (2008) 373 – 387

www.elsevier.com/locate/addr

Applications of supercritical CO2 in the fabrication of polymer systems for

drug delivery and tissue engineering ☆
Owen R. Davies a,1 , Andrew L. Lewis a,1 , Martin J. Whitaker a , Hongyun Tai b,c ,
Kevin M. Shakesheff b , Steven M. Howdle a, c ,⁎
a
Critical Pharmaceuticals Ltd BioCity Pennyfoot St., Nottingham, NG1 1GF, United Kingdom
b
School of Pharmacy, University of Nottingham, University Park, Nottingham, NG7 2RD, United Kingdom
c
School of Chemistry, University of Nottingham, University Park, Nottingham, NG7 2RD, United Kingdom
Received 13 November 2006; accepted 14 December 2006
Available online 5 October 2007

Abstract

Supercritical CO2 has the potential to be an excellent environment within which controlled release polymers and dry composites may be
formed. The low temperature and dry conditions within the fluid offer obvious advantages in the processing of water, solvent or heat labile
molecules. The low viscosity and high diffusivity of scCO2 offer the possibility of novel processing routes for polymer drug composites, but there
are still technical challenges to overcome. Moreover, the low solubility of most drug molecules in scCO2 presents both challenges and advantages.
This review explores the current methods that use high pressure and scCO2 for the production of drug delivery systems and the more specialized
application of the fluid in the formation of highly porous tissue engineering scaffolds.
© 2007 Elsevier B.V. All rights reserved.

Keywords: Supercritical CO2; Drug delivery; Tissue engineering; Polymer plasticization

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 374
2. Solvent properties of scCO2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 374
2.1. Solubility of polymers in scCO2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 374
2.2. Solubility of CO2 in polymers. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 374
3. Drug delivery applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 375
3.1. Rapid expansion of scCO2 (RESS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 375
3.2. Anti-solvent applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 376
3.2.1. Biodegradable polymer microparticles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 376
3.3. Particles from gas saturated solutions (PGSS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 378
3.4. Other drug delivery applications. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 379
3.5. Extraction of solvents with carbon dioxide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 379

☆
This review is part of the Advanced Drug Delivery Reviews theme issue on “Drug Delivery Applications of Supercritical Fluid Technology”.
⁎ Corresponding author.
E-mail address: steve.howdle@nottingham.ac.uk (S.M. Howdle).
1
Joint first author.

0169-409X/$ - see front matter © 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.addr.2006.12.001
374 O.R.
O.R.
Davies
Davies
et al.
et al.
/ Advanced
/ Advanced
Drug
Drug
Delivery
Delivery
Reviews
Reviews
60 60
(2008)
(2008)
373–387
373–387 374

4. Manufacture for tissue engineering scaffolds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 380

4.1. Production of porous polymer scaffolds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 380
4.1.1. Gas foaming . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 380
4.1.2. Supercritical fluid emulsion templating . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 382
4.2. Encapsulation and release of growth factors and cells using supercritical fluids . . . . . . . . . . . . . . . . . . . . . . . . 383
5. Conclusions and future challenges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 384
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 384

1. Introduction Si\O\Si— moieties in their backbones are responsible for the

A very wide range of clinically approved pharmaceutical
products take advantage of biodegradable polymers to control the
rate of drug release within the body [1,2]. The preparation of
most systems relies on either melt processing or organic solvent
based methods to incorporate into the polymer phase. Liquefac-
tion is required to mix the drug and polymer phases and to form
the final delivery system morphology. The use of organic
solvents, the formation of interfaces between organic and
aqueous phases and the requirement for high temperatures are
all factors that are not desirable to retain activity of the drug.
Protein therapeutics are one example of a drug class where new
controlled release devices will be required to overcome the issue
of their short half-lives and long treatment times [3]. This review
considers a range of uses of high pressure and supercritical
carbon dioxide (scCO2) for the preparation of devices in which
the drug is dispersed throughout the polymer phase.
2. Solvent properties of scCO2
CO2 is inexpensive, non-toxic, non-flammable and readily
available in high purity from a variety of sources. scCO2
(Tc = 31.1 °C, Pc = 73.8 bar) has the unique combination of gas-
like diffusivity and liquid-like density, which can be tuned
easily by changes in pressure. In recent years, it has been widely
used as a solvent, anti-solvent and plasticizer for synthesis,
modification and purification of both synthetic and natural
polymers [4–9]. The solvent properties of scCO2, i.e. the
solubility of polymers in scCO2 and the solubility of CO2 in the
polymers, are two key fundamental subjects in this field. Most
pharmaceutical compounds and polymers have demonstrated
low solubility in scCO2. However, it is the high solubility [10]
of CO2 in many polymers that can be particularly advantageous
and can lead to dramatic decreases in the glass transition
temperature (Tg) or melting temperature Tm, thus reducing
polymer melt viscosity [6,11–15].
2.1. Solubility of polymers in scCO2
While CO2 is a good solvent for most non-polar and some
polar low molecular substances, it is a poor solvent for most
high molecular weight polymers under mild conditions (T
b 100 °C, P b 350 bar), although there are some exceptions
such as amorphous fluoropolymers and silicones [4,5]. The
significant solubility of fluorinated polymers is due to the
interaction of CO2 with the C\F bond, while the flexible —
solubility of siloxane polymers [16]. The polarity of polymers
can also be a reason for their limited solubility because CO2 has
a quadrupole moment and no permanent dipole [17]. In
addition, CO2 has the potential to act as both a weak Lewis acid
and Lewis base, and theoretical and experimental evidence
indicates that CO2 can participate in hydrogen-bonding
interactions. This suggests that CO2 might solubilise dipolar
and non-dipolar molecular systems through site-specific solute–
solvent interactions. Beckman et al. [9,18] have shown that by
careful molecular design and choice of co-monomers to balance
the solvent–solute and solute–solute interactions it is possible
to prepare hydrocarbon based copolymers that are CO2-soluble.
2.2. Solubility of CO2 in polymers
CO2 solubility and diffusivity in polymers are influenced by
both the molecular structure (the interaction between CO2 and
molecular chains) and the morphology (crystalline or amor-
phous, related with free volume) of polymers. The solubility of
CO2 in many polymers, e.g. poly(methyl methacrylate) PMMA
and poly(styrene) PS, has been studied by evaluating CO2
sorption and polymer swelling [19–21]. There was a general
perception that sorption and swelling is a purely physical
phenomenon until Fourier transform-infrared (FTIR) and ATR-
IR spectroscopy were used to study the specific intermolecular
interactions between CO2 and polymers [22–24]. While the
chain flexibility of polymers can aid dissolution of CO2,
carbonyl or ether groups that are accessible in the backbone or
on side chains can specifically interact with CO2. Polymers with
ether groups in poly(ethylene glycol) (PEG) displayed stronger
interaction than polyesters due to weak Lewis acid–base
interaction [24]. Although poly(lactic acid) (PLA) and poly
(lactic acid-co-glycolic acid) (PLGA) have the same chemical
structure in the main chains, the steric hindrance close to the
carbonyl group and accessible free volume caused by methyl
pendant groups can lead to their different behaviour of solubility
and interaction. In the case of PLGAs, some methyl groups are
substituted with hydrogen which causes much less hindrance
for the interaction with CO2. However, the accessible free
volume in the polymers was found to have a greater effect on
solubility than the interaction between CO2 and the polymers
[25]. Therefore, the solubility of CO2 in PLGA copolymers
decreases with increasing the glycolic acid content. Moreover,
Oliveira et al. [26] studied the solubility of CO2 in PLA with two
different L:D contents, and found that CO2 is more soluble in
PLA 80:20 (amorphous) than in PLA 98:2 (20% crystallinity).
375 O.R.
O.R.
Davies
Davies
et al.
et al.
/ Advanced
/ Advanced
Drug
Drug
Delivery
Delivery
PGA, PEG and PCL are highly crystalline polymers, and CO2
does not easily diffuse into these at the temperatures below their
melting points, leading to low solubility. In contrast, amorphous
poly(DL-lactic acid) (PDLLA) and PLGAs have larger free
volume, which allows CO2 more easily to diffuse into them.
3. Drug delivery applications
3.1. Rapid expansion of scCO2 (RESS)
RESS has been used to produce a wide range of different
drug delivery systems including fibres, microparticles and
films. In the RESS process the solute is solubilised in the
supercritical fluid [27] and then expanded through a capillary
nozzle, typically b 150 μm in diameter, into a precipitation
chamber. Rapid decompression of the solution leads to super-
saturation, nucleation and particle formation. This process can
result in the formation of very small uniform particles due to the
high supersaturation ratios which can be achieved. Alterna-
tively, larger particles can be produced by controlling factors
such as temperature, pressure and nozzle geometry.
Tom et al. were the first to use the RESS process to produce
drug encapsulated microparticles [28]. PDLLA and lovastatin
were dissolved in scCO2 and co-precipitated by RESS to form
drug entrapped polymeric microparticles and microspheres (10–
100 μm). Varying the concentration of lovastatin resulted in a
change in product morphology from drug–polymer networks to
microspheres. The presence of crystalline needles of lovastatin
embedded in the PDLLA microspheres suggested that the drug
and polymer precipitated independently.
More recently Kim et al. encapsulated the non-steroidal anti-
inflammatory drug naproxen into low molecular weight
(2000 Da) PLLA microparticles using a RESS process [29].
Polymer and drug were co-dissolved in scCO2 and expanded
through a capillary at a pressure of 170–200 bar and
temperature of 90–115 °C into a precipitation chamber. This
process is therefore different to earlier studies where drug and
polymer were dissolved separately in scCO2 and then co-
precipitated [28,30]. Polymeric microparticles (10–90 μm) with
embedded naproxen particles (2–5 μm) were successfully
prepared. Naproxen was found to dominate the core while the L-
PLA coated the surface due to slower precipitation.
The highlighted studies demonstrate the feasibility of the
RESS process for preparing polymeric drug delivery systems.
However, the main limiting factor is the low solubility of most
pharmaceutical compounds and regulatory approved polymers
in scCO2. To overcome this obstacle, co-solvents have been
used to modify the polarity of the extracting phase and thus
enhance solute solubility in the supercritical phase. Tom and
Debenedetti demonstrated that when 1% (w/w) acetone was
used as a co-solvent the solubility of PLLA increased by
approximately 500% [31]. Chlorodifluoromethane has also
been utilised as a co-solvent for the preparation of PLLA
microparticles and microspheres using RESS [30]. The
fluorescent solute pyrene was uniformly incorporated into the
microparticles by co-precipitation with the L-PLA. The nozzle
geometry (orifice or capillary) was demonstrated to be an
Reviews
Reviews
60 60
(2008)
(2008)
373–387
373–387 375
important determinant of product morphology. When an orifice
was used, dendrites were the most common morphology,
whereas microspheres were prepared when capillary nozzles
with low length-to-diameter ratios were used.
In many circumstances, even in the presence of co-solvents,
the solubility of most pharmaceutical compounds is still far too
low to make RESS a viable process. Furthermore, the higher
temperatures which are often required for RESS make it
unsuitable for labile polymers and drugs. To circumvent this
problem, scCO2 can be used as an alternative to conventional
solvents to coat drug particles with polymer [32–34]. This process
takes advantage of the differing solubility of drug and polymer in
CO2–solvent mix. The solubility of the coating material is much
higher than that of the drug so the polymer precipitates onto
preformed drug particles on decompression. An example of this is
the rapid expansion from a supercritical solution with a non-
solvent (RESS-N) process. This has been used to encapsulate both
low molecular weight drugs and proteins into polymeric
microparticles [33,34]. RESS-N utilises polar organic solvents,
such as low molecular weight alcohols (methanol, ethanol and 1-
propanol), as co-solvents. These solvents greatly increase the
solubility of many pharmaceutical polymers (including PEG,
PMMA, PEG-PPG-PEG, and ethylcellulose) in scCO2 due to
their hydrogen-bonding capacity. Furthermore, as they are non-
solvents for the polymer, they do not induce swelling or
microparticle aggregation. Proteins (lysozyme and lipase) con-
taining microparticles were prepared by spraying a suspension of
protein in scCO2 and dissolved polymer at a temperature of 35 °C
and a pressure of 200 bar through a nozzle to atmospheric pressure
[34]. PEG, PMMA, PLLA, PDLLGA, and PEG–PPG–PEG
triblock polymers were used as coating materials. Particles were
fabricated in the range from 8 to 62 μm. More recently the RESS-
N process has been used to encapsulate the low molecular weight
drugs ρ-acetamidophenol, acetylsalicylic acid, 1-3-dimethyl-
xanthine, flavone and 3-hydroxyflavone into PEG, PMMA,
ethylcellulose and PLLA microparticles with an average size of
14.6 μm [33]. Although RESS-N is a promising technique, drug
release from coated microparticles has not been reported to date so
it is unclear if coating using this technique can modify dissolution
kinetics.
A similar RESS process has been successfully used to coat
bovine serum albumin (BSA) microparticles with the lipidic
compounds Dynasan® n 114 or Gelucire® 50-02 [32]. BSA and
the coating material were placed in the autoclave and scCO2 (35
or 45 °C and 200 Bar) introduced to dissolve the coating
material whilst stirring. Cooling induced a phase change from
supercritical to liquid CO2 and reduction in solvent strength
which resulted in the coating material being precipitated onto
the preformed drug particles. Whilst coating with Gelucire® 50-
02 produced a homogeneous coat and protein was released
continuously over 24 h, coating with Dynasan® 114 did not
give a continuous coat which led to rapid in vitro release of
BSA (70% in 30 min).
Another application of the RESS process in the field of drug
delivery is impregnation of polymeric materials with pharma-
ceutical agents (see Kikic and Vecchione for a review) [35].
Impregnation utilises the high diffusivity and low surface
376 O.R.
O.R.
Davies
Davies
et al.
et al.
/ Advanced
/ Advanced
Drug
Drug
Delivery
tension of supercritical fluids. Polymer impregnation can be
achieved when the active substance is soluble in the super-
critical fluid (SCF) and the polymer is swollen by the SCF. If the
drug has high affinity for the polymer matrix (high partition
coefficient between the polymer and SCF phase), or specific
interaction with the polymer, the drug can be molecularly
dispersed within the polymeric matrix. Drugs with low affinity
for the polymer can be precipitated and deposited into the
matrix on decompression. In this scenario the drug will
recrystallise in the matrix and will not be molecularly dispersed.
Gueny and Akgerman impregnated PDLLGA with 5-fluorour-
acil (5FU) and β-estradiol for the purpose of controlled release
[36]. Macroporous foams or microporous particles could be
prepared depending on if the system was depressurised rapidly or
slowly. Drug loadings of 26.8 mg/g and 0.49 mg/g could be
achieved at 55 °C 207 bar for β-estradiol and 5FU respectively.
The majority of the drug was released rapidly from the matrix
suggesting the bulk of the drug was adsorbed to the surface rather
than encapsulated.
Kazarian and Martirosyan used ATR-IR spectroscopy to
monitor ibuprofen impregnation of poly(vinyl pyrrolidone)
(PVP) films [37]. It was demonstrated that ibuprofen was
molecularly dispersed in the polymer matrix, with the drug
molecules interacting with carbonyl group of PVP. In a similar
study the RESS process was used to impregnate indomethacin
into chitosan thermosets for controlled release applications [38].
Chitosan and indomethacin powders were physically mixed and
then exposed to scCO2 in an autoclave. SCF impregnation
resulted in indomethacin being amorphously dispersed through-
out the chitosan matrix. FTIR data suggested that the aliphatic
carbonyl group of indomethacin interacted with NH2 groups of
chitosan. Following supercritical fluid processing, the dissolu-
tion rate of indomethacin was markedly increased. Van Hees et
al. used the RESS process to prepare drug cyclodextran
inclusion complexes to improve the solubility of the poorly
soluble drugs [39,40]. Piroxicam-β-cyclodextran inclusion
complex and miconazole cyclodextran complexes were pre-
pared by physically mixing the compounds prior to exposure to
scCO2. The scCO2 carries the drug into cyclodextran cavity,
substituting water molecules with the less polar guest.
3.2. Anti-solvent applications
scCO2 is a relatively poor solvent for most polymers and
pharmaceutical compounds. Although co-solvents can be used
to increase solubility in CO2, or the process adapted to coat
preformed drug particles (as discussed above), in the vast
majority of circumstances solubility is still too low to make
RESS a practical process. To overcome these problems the anti-
solvent or non-solvent properties of carbon dioxide can be
exploited. Many different anti-solvent techniques have been
used for preparation of polymeric drug delivery systems. These
include: GAS, SAS, PCA, ASES and SEDS. These techniques
use a dense gas as an anti-solvent to precipitate a solute which is
dissolved in an organic solvent. Precipitation occurs as the gas
is absorbed by the organic solvent thus expanding the liquid
phase and reducing the solvent power until nucleation and
Delivery
Reviews
Reviews
60 60
(2008)
(2008)
373–387
373–387 376
particle formation occur. The supercritical anti-solvent techni-
ques offer greater flexibility than the RESS process and
consequently have been studied more thoroughly for prepara-
tion of drug delivery systems.
3.2.1. Biodegradable polymer microparticles
The majority of anti-solvent studies have focussed on
encapsulation of pharmaceuticals into biodegradable poly
(esters) such as PLA and PLGA. Bleich and co-workers were
amongst the first to demonstrate the potential of anti-solvent
techniques for microencapsulation [41]. Hyoscine butylbromide
was encapsulated into PLLA microparticles by co-precipitation
of drug and polymer from a solution of methanol and methylene
chloride using the ASES process. A maximum drug loading of
19.8% (w/w) was achieved. Similarly Sze et al. used ASES to
microencapsulate para-hydroxybenzoic acid (ρ-HBA) into
PLLA [42]. Drug encapsulation efficiency was relatively poor,
but could be modestly improved from 7.0% to 9.2% by spraying
the drug solution and polymer solution through separate chan-
nels using co-axial nozzle assembly, rather than co-dissolving
the drug and polymer and spraying the mixture through a single
nozzle. Particles consisted of fibrous ρ-HBA embedded within
polymeric microparticles. Temperature was found to be an
important parameter in determining particle morphology.
Whereas discrete particles were formed at 25 °C, at temperatures
above the critical point the number of deformed or agglomerated
particles increased due to the plasticization effect of CO2.
Bodmeier and co-workers assessed the suitability of PCA for
the preparation of drug entrapped polymeric microparticles
[43]. Initially the swelling behaviour of a range of polymers in
carbon dioxide was investigated. It was found that amorphous
polymers with low glass transition temperatures agglomerated
even at low temperatures. The more crystalline PLLA was
therefore selected to study further. Microparticles were favoured
at moderate temperatures, low polymer concentrations (1%),
high pressure and high CO2 flow rates, whereas fibres formed at
high polymer concentrations and low flow rates. Chlorphenir-
amine and indomethacin were encapsulated into the particles
but encapsulation efficiency was low. More recently Falk et al.
have used the PCA technique to encapsulate gentamycin,
naloxone and naltrexone into PLLA microparticles [44]. To
improve solubility in methylene chloride, gentamycin and
naltrexone were solubilised in methylene chloride by hydro-
phobic ion pairing (counter ions replaced with an anionic
detergent sodium bis-2-ethylhexyl sulfosuccinate). Drug encap-
sulated PLLA microparticles (0.2–1 μm) were prepared by
expanding the drug /polymer solution through a vibrating
nozzle (120 kHz). Ion paired drugs were encapsulated with high
efficiency whereas the encapsulation efficiency of the hydro-
phobic drug rifampicin, which was not ion paired, was much
lower due to its higher solubility in CO2. In a further publication
Falk and Randolph demonstrated that increasing post-precipita-
tion CO2 flow rate and volume decreased residual organic
solvent levels [45]. Interestingly, increasing CO2 co-flow rate
during precipitation increased residual solvent content as it led
to a less crystalline product and thus reduced diffusivity of
methylene chloride through the polymer matrix.
377 O.R.
O.R.
Davies
Davies
et al.
et al.
/ Advanced
/ Advanced
Drug
Drug
Delivery
Delivery
Reviews
Reviews
60 60
(2008)
(2008)
373–387
373–387 377

The SEDS process has also been exploited for the preparation
of drug loaded polymeric microparticles [46,47]. Ghaderi et al.
used the SEDS process to encapsulate hydrocortisone into
discrete sub 10 μm particles PDLLGA, PLLA and PDLLA [47].
Chen et al encapsulated 5FU into L-PLA microparticles [46].
Sub-micron particles (980 nm) were prepared with low residual
methylene chloride content (46 ppm). A drug load of 3.05% and
encapsulation efficiency of 17.8% were achieved. Drug was
released rapidly (89.5% in first 36 h followed by 5.5% in next
36 h) by a diffusion mechanism.
In addition to small drug molecules, anti-solvent techniques
have also been exploited to encapsulate peptide, protein and
even plasmid DNA into polymeric microparticles. Engwicht
et al. used ASES to encapsulate BSA or estriol in to PDLLGA
50:50 or a triblock co-polymer of b-poly-L-lactide-co-DL-
lactide-co-glycolide (62.5:12.5:25) microparticles [48]. Poly-
mers were dissolved in a mixture of 2,2,2-trifluoroethanol and
methylene chloride and the drug dissolved in methanol prior to
mixing with the polymer solution. PLGA formed small ag-
glomerated particles due to the plasticization effect of CO2 and
residual solvent, whereas the triblock polymer formed larger
individual particles (sub 10 μm). High loading efficiency of
BSA and estriol (over 100% due to extraction of LMW polymer
species) was observed for both polymers. Approximately 50%
burst release occurred in first 24 h due to drug at or near the
microparticle surface. Fig. 1. Insulin release profiles from nanoparticles containing 1900 Da PEG: 23
Elvassore et al. prepared insulin and insulin/PEG loaded (⁎), 41 (○), 50 (Δ), 60 (□), 67 (●) and 75 (■).Reproduced with kind permission
from [50].
PLLA (102,000 MW) nanoparticles by GAS anti-solvent [49].
PLLA was dissolved in methylene chloride and then mixed with
a solution of PEG and insulin co-dissolved in DMSO (50:50 3.7% (w/w) was achieved under optimised conditions, but the
solvent ratio). Nanoparticles (400–600 nm) were prepared in encapsulation efficiency was only 14.6%.
high yield and high drug encapsulation efficiency (up to 93.4%) More recently chitosan–plasmid DNA complexes have been
with very low residual methylene chloride (8 ppm) and DMSO prepared using a supercritical anti-solvent technique [51].
content (300 ppm). Insulin extracted from the particles The complexes were prepared by spraying a chitosan–DNA
maintained N 80% of its hypoglycaemic activity in vivo. When complex solution through a V shaped nozzle, with the anti-
PEGs N 1900 MW were incorporated insulin was released in a solvent flowing through one side and the drug polymer solution
burst over 80 h, whereas when low MW PEGs were used no through the other. A mixture of ethanol and scCO2 were used as
burst occurred and insulin was released slowly over 1500 hour. the anti-solvent. The ethanol functioned as a co-solvent to allow
As the L-PLA used in these studies had a very high MW the water to be miscible with carbon dioxide. Inclusion of
(102,000) and degraded extremely slowly with b 3% of chitosan reduced DNA degradation during processing,
encapsulated insulin being released over the 1500 hour study increased yield and enhanced luciferase expression in the
period. In a more recent report it was demonstrated that insulin murine lung when compared to powder sprayed without
dissolution profile could be improved using high concentrations chitosan or plasmid DNA solutions.
of low MW PEG (350, 750 or 1900 Da) [50]. Using a semi- The majority of drug microencapsulation studies have
continuous gas anti-solvent precipitation technique, under focussed on the semi-crystalline polymer PLLA, whereas as far
optimised conditions, insulin loaded PLA/PEG nanoparticles fewer studies have successfully produced drug entrapped
were prepared in high yield (N 70%) and with high encapsula- microparticles from amorphous polymers such as PDLLGA.
tion efficiency (over 90%). Encapsulation efficiency was PLLA has very limited applications in the field of drug delivery
reduced as PEG 1900 content increased or as PEG MW due to extremely long in vivo degradation times (months to years).
increased at 67% PEG loading. Drug release rate increased as It is only suitable for delivery of extremely potent drugs over
PEG content increased with 100% of insulin released over 3– prolonged periods of time, applications that require infrequent
4 months when particles were produced with a PEG 1900 administration, for example vaccines, or drug eluting medical
content of 67 or 75%, but only 10% at a 23% PEG load (Fig. 1). devices. Most controlled release drug delivery applications
Sze et al. encapsulated lysozyme into PLLA microparticles require dosing frequencies between once weekly and once
using an ASES technique [42]. Lysozyme (dissolved in DMSO) every 6 months. Dose frequency is determined by the stability
and PLLA (dissolved in methylene chloride) were sprayed of the drug in the polymeric system and the potency of the
though separate channels using a co-axial nozzle. A loading of therapeutic. For drug delivery systems faster degrading
378 O.R.
O.R.
Davies
Davies
et al.
et al.
/ Advanced
/ Advanced
Drug
Drug
Delivery
Delivery
Reviews
Reviews
60 60
(2008)
(2008)
373–387
373–387 378

amorphous polymers such as PDLLGA and PDLLA are therefore using a PCA technique. Rather than dissolving lysozyme in a
required. However, carbon dioxide has a higher solubility in solvent, lysozyme particles were suspended in a solution of
amorphous polymers resulting in slower mass transfer and rate of polymer dissolved in methylene chloride. Suspension of protein
solidification. Consequently, microparticles tend to aggregate and in organic solvents generally causes much less degradation than
flocculate due to the plasticizing effect of residual carbon dioxide dissolution. Particles were formed by spraying the protein-
[43,52–54]. The plasticizing effect can be further enhanced by the polymer suspension through a capillary nozzle into a carbon
presence of residual organic solvents in the final product. dioxide vapour phase above a liquid CO2 phase (Fig. 2). Lyso-
To overcome the plasticization effect caused by residual CO2 zyme was successfully encapsulated at a 10:1 by weight of
and solvent, Ghaderi et al. used combinations of scCO2 and polymer to protein into PDLLGA particles, but drug dissolution
supercritical nitrogen [47]. The nitrogen decreased the interac- was not investigated.
tion between carbon dioxide and the polymer inside nozzle
thereby increasing the dispersive effect of the nozzle. Discrete 3.3. Particles from gas saturated solutions (PGSS)
particles b 10 μm were produced from PDLLGA, over 10 times
smaller those prepared in the absence of nitrogen [54]. The PGSS process is an alternative supercritical fluid
Decreased polymer solubility in the CO2–N2 mixture resulted processing technology that can be used to form particles from
in faster particle solidification rates, less aggregation and hence polymers or other materials that are plasticized by a supercritical
smaller particles. Hydrocortisone was entrapped into PDLLGA fluid. First developed for coatings [58] the technique has been
particles with an encapsulation efficiency of 22%. There was no applied also to the preparation of particles of the small molecule
change in particle size between drug loaded and non-loaded drug nifedipine [59,60] which plasticizes well in scCO2. More
particles. recently the technique has been developed to produce drug
Young et al. successfully produced PLGA particles using a particles entrapped within polymeric microparticles, fibres or
modified PCA technique in which PLGA was dissolved in implants or scaffolds. The PGSS process utilises the ability of
methylene chloride and then sprayed into a vapour CO2 phase scCO2 to plasticize amorphous polymers such as PDLLGA and
over a liquid CO2 phase [55]. Temperature was found to be an PDLLA. The addition of CO2 markedly lowers the Tg of the
extremely important determinant of final product morphology. At polymer causing it to liquefy at near ambient temperatures.
higher temperatures PLGA remained highly plasticized (Tg Furthermore, the presence of CO2 significantly lowers the
decreased from 45 °C to approximately − 40 °C at 276 K in the viscosity of the polymer melt which enables drug particles to be
presence of carbon dioxide) resulting in agglomerated particles.
While at temperatures of below − 30 °C dissolution rate of
methylene chloride with CO2 decreased also resulting in large
agglomerates (N 1000 μm). At − 20 °C uniform spherical particles
were prepared. Polymer concentration was also an extremely
important determinant of particle size. As polymer concen-
tration increased, particle size increased (from 0.5–5 μm at 1% to
5–70 μm at 5%) due to increased viscosity and slower pre-
cipitation rates. At concentrations N 10% large agglomerates were
formed.
A draw back of the anti-solvent techniques is that they
require the pharmaceutical agent to be soluble in a solvent
which is miscible with scCO2. This limits the use to compounds
that are soluble in organic solvents. While this does not create a
problem in the case of low molecular weight hydrophobic
compounds, more complex hydrophilic agents, such as proteins
and peptides, are insoluble in most organic solvents. Conse-
quently, CO2 miscible solvents such as DMSO must be used to
solubilise the biological molecule. When dissolved in such
solvents proteins can refold and irreversibly change their
secondary structure resulting in loss of functional activity and
risk of immunogenicity. Two approaches have been used to
overcome this limitation. The first is to use ethanol as a co-
solvent to allow the water to be miscible with carbon dioxide
[51,56]. Although proteins have been sprayed to produce stable
micron-sized powders using such approach [56,57], most
commonly used biodegradable polymers are not water soluble
Fig. 2. SEM micrograph (bottom) and optical micrograph (top) of PLGA
so this technique is unlikely to be suitable. microspheres formed by spraying at 0.5 mL/min a 5.0/0.5% PLGA/lysozyme
A second approach was taken by Young et al. [55] where suspension in CH2 Cl2 through a 100 μm capillary nozzle into CO2 flowing at
lysozyme was encapsulated into PLLA or PDLLGA microspheres 25 mL/min at − 20 °C. Reproduced with kind permission from [55].
379 O.R.
O.R.
Davies
Davies
et al.
et al.
/ Advanced
/ Advanced
Drug
Drug
Delivery
Delivery
homogeneously mixed in by means of a stirring device. In these
processes, the drug itself is usually protein based and is not
plasticized or modified in any way by the scCO2. If the liquefied
mixture is sprayed into a collection chamber fibres or particles
can be produced [61–63], whereas if the process is conducted in
a mould, the polymer will foam on depressurisation leading to
the production of a porous scaffold or monolith [64,65].
The PGSS process has many advantages over other super-
critical fluid and emulsion technologies. These include: the
absence of solvents at any stage during processing, the fact that
neither the polymer nor drug need be soluble in the scCO2 and
the mild processing conditions (typically b 40 °C and 150 bar).
Furthermore, as there are no solvent removal steps, the process
is extremely rapid. The PGSS process is particularly suitable for
entrapment of delicate biopharmaceuticals such as proteins
[63–65]. A significant challenge with the PGSS process is to
produce drug encapsulated polymeric microparticles from a
highly viscous liquefied polymer and with a rapid polymer
solidification rate on depressurisation. Furthermore, residual
CO2 may cause agglomeration of the final particulate product.
To enable microparticle formation the solidification rate must be
tightly controlled so that it is slow enough to allow for droplet
formation and polymer chain rearrangement, but fast enough to
prevent aggregation of newly formed particles when they arrive
at the bottom of the collection vessel. Hao et al. used the PGSS
process to produce microparticles from the crystalline polymer
PEG [61]. Spherical particle production was favoured by
reducing the pressure, increasing the temperature and reducing
the nozzle size.
Production of particles from amorphous polymers is even
more challenging than from crystalline polymers due to the
increased. The increased viscosity of the polymer melt makes
atomisation more difficult, and the greater solubility of CO2 in
the polymer phase can result in plasticization and agglomeration
of the final product. To overcome these problems Hao et al.
produced PDLLA microparticles (6100 Da) by spraying the
liquefied polymer/CO2 mixture into a collection chamber with a
nitrogen back pressure [62]. The nitrogen suppressed the loss of
CO2 from the liquefied CO2/polymer mixture thereby slowing
down rate of polymer solidification allowing particle formation
to occur. Cooling the collection chamber with liquid nitrogen
prevented aggregation of the newly formed microparticles. A
back pressure of N 68 bar of nitrogen was required to produce
fine particles. In the absence of back pressure a very fibrous
product was formed. Whitaker et al. extended this work to
produce drug entrapped PDLLA microparticles. Lysozyme,
ribonuclease, insulin and calcitonin were successfully entrapped
into the microparticles with minimal loss of structural integrity
or functional activity.
3.4. Other drug delivery applications
More recently, scCO2 has been exploited to process a wider
range of polymeric drug delivery systems. Moneghini et al. used
GAS anti-solvent to co-precipitate the poorly soluble drug
carbamazepine with the hydrophilic carrier PEG 4000 [66]. Co-
precipitation with the PEG reduced drug particle size and
Reviews
Reviews
60 60
(2008)
(2008)
373–387
373–387 379
improved crystal shape which markedly enhanced in vitro
dissolution rate.
Duarte et al. produced microparticles from ethylcellulose/
methylcellulose blends using either a solvent evaporation
technique or by a SAS process [67]. Microparticles were then
impregnated with the NSAID naproxen by solubilising the drug
in supercritical carbon dioxide using a RESS technique. Drug
release was slower from particles impregnated using scCO2
when compared to those where drug was encapsulated by
solvent evaporation. Following impregnation with naproxen
using supercritical CO2, particles produced by SAS had faster
and more controlled release than those prepared by solvent
evaporation due to the smaller size of the particles and hence
shorter diffusion path.
Reverchon et al. prepared amoxicillin loaded PMMA porous
structures using a SCF phase inversion process in which CO2
acted as a non-solvent [68]. Amoxicillin was loaded into the
structure by co-dissolving it in the solvent (DMSO) with the
polymer, or suspending it in the polymer phase (PMMA dis-
solved in acetone). Pore size and morphology could be con-
trolled by choice of solvent, temperature and polymer
concentration. When amoxicillin was co-dissolved with the
polymer, amoxicillin was uniformly distributed throughout
matrix and the dissolution rate was much slower than when the
drug was suspended in the polymer (3 h and 20 h respectively).
Kunastitchai et al. encapsulated miconazole into liposomes
prepared with various compositions of phosphatidylcholine,
cholesterol and poloxamer 407 [69]. At optimised conditions,
partially crystalline, spherical and nonporous microparticle
aggregates were prepared with size ca. 40 μm with residual
content of methylene chloride and methanol of 80 ppm and
36 ppm respectively, well below regulatory limits (600 and
3000 ppm respectively). Increasing solute concentration
increased yield but lead to larger less spherical particles and
fibres due to increased viscosity and incomplete atomisation.
Another interesting application of supercritical carbon
dioxide which has been investigated recently is the sterilisation
of biodegradable polymers [70]. Sterilisation occurred without
detectable chemical changes to the polymer, possibly as a result
of the formation of carbonic acid, which lowers the intracellular
pH.
3.5. Extraction of solvents with carbon dioxide
Organic solvents such as methylene chloride are generally
used to solubilise biodegradable polymers during production of
drug delivery systems. The final product will consequently
contain residual amounts of solvent. The U.S.P limit for
methylene chloride is 600 ppm. As methylene chloride is miscible
with carbon dioxide, products produced by anti-solvent techni-
ques generally have very low residual levels. However, products
prepared using conventional emulsion techniques can have very
high residual solvent levels, well in excess of the U.S.P limit.
Recently carbon dioxide has been used to extract residual
solvent from microparticles and rods produced by Three
Dimensional Printing™ (3DP™) or emulsion techniques [71–
73]. Solvent extraction utilises the plasticisation effect of the
380 O.R.
O.R.
Davies
Davies
et al.
et al.
/ Advanced
/ Advanced
Drug
Drug
Delivery
carbon dioxide which increases the solvent diffusion coefficient
and diffusion rate. Herberger et al. used CO2 as an extraction
solvent to reduce the residual levels of methylene chloride in
spray dried PDLLGA-darbepoetin alfa microparticles [71].
Whereas liquid CO2 was found to cause some protein aggre-
gation and microparticle agglomeration, extraction with CO2
gas reduced residual solvent concentration and limited particle
aggregation. Solvent extraction rate and degree of particle
agglomeration were found to increase with CO2 pressure.
Carbon dioxide extractions at constant pressure did not reduce
residual solvent concentration to below target levels. Using
extraction cycles, whereby CO2 pressure was increased as
solvent was removed, a residual methylene chloride content of
b 600 ppm was achieved. In a similar study liquid CO2 was used
to reduce the residual solvent in dense PDLLGA devices to
b 50 ppm [72]. Exposure to liquid CO2 plasticised the rods
making them flexible and easily deformed for several hours
after exposure.
Koushik and Kompella utilised scCO2 to not only reduce
residual methylene chloride content in desorelin loaded
PDLLGA (50:50 23.2 kDa) microparticles, but also to increase
particle porosity [73]. Microparticles prepared by an o/w sol-
vent evaporation technique were exposed to scCO2 (1200 psi,
33 °C, 30 min). Upon depressurisation, pore formation was
induced by rapid expansion of entrapped CO2. Following
scCO2 treatment, particle size increased from 2.2 to 13.8 μm,
porosity increased from 39 to 92.4% and bulk density was
reduced from 0.7 to 0.082 g/cc. Residual solvent content
decreased from 4500 ppm to b 25 ppm. Peptide stability was
maintained following processing. Uptake of scCO2 treated
particles was reduced in A549 cells and rat alveolar macro-
phages when compared to non-treated particles. Use of a large
particle size, to reduce clearance by alveolar macrophages, and
of a low mass density, to increase respirable fraction [74], may
make the particles suitable candidates for sustained delivery to
the respiratory tract. However, the toxicity of PDLLGA when
administered via the inhaled route of delivery has yet to be fully
elucidated.
More recently, Chattopadhyay et al. have evaluated a novel
technique called supercritical fluid extraction of emulsions
(SFEE) [75]. Indomethacin and ketoprofen were encapsulated
into PDLLGA and PMMA microparticles using an oil-in-water
emulsification technique. scCO2 was then used to extract the
solvent from the emulsion. Drug entrapped particles (0.1 and
2 μm) with low residual solvent content (b 50 ppm) and high
loading efficiency (98%) were prepared. Particles were
produced in either a batch fashion, by bubbling scCO2 through
the emulsion, or in a continuous fashion by spraying the
emulsion into an extraction column with supercritical CO2
flowing in the opposite direction. Coalescence was prevented
by the presence of a stabilising surfactant in the aqueous phase.
4. Manufacture for tissue engineering scaffolds
Driven by the massive shortage of donor organs, and
increasing frequency of age-related diseases with unsatisfactory
treatments, the closely allied fields of tissue engineering and
Delivery
Reviews
Reviews
60 60
(2008)
(2008)
373–387
373–387 380
regenerative medicine aim to repair, restore or replace damaged
organs. Bone marrow transplantation demonstrates that cell
based therapies can have outstanding success in treating pre-
viously fatal and untreatable diseases. However, for most
conditions the administration of cells alone does not result in
integration into and repair of the host tissue.
To overcome this, a central strategy of tissue engineering is
to try to provide the correct microenvironment for the cells (a
niche) using biomaterials as a scaffold for tissue repair, regen-
eration or reconstitution in vivo or ex vivo. Typically, a thera-
peutic cell population is seeded onto a biodegradable scaffold
which has been modified to provide the appropriate cues to
promote cell proliferation, differentiation and organization into
the desired tissue. Alternatively biomaterials loaded with the
appropriate growth factors, or signalling domains may be ad-
ministered to the site of injury to encourage infiltration of the
patients' own cells to promote healing.
To be effective such scaffolds must provide a large surface
area for cell attachment and efficient gas and nutrient exchange
(and are therefore usually porous) and release the appropriate
growth factor(s) with the appropriate kinetics. Most conven-
tional techniques to produce such polymer scaffolds require
heat or solvents which can be deleterious to the cells, tissues and
encapsulated biological molecules. Since it is solvent free and
can operate at relatively low temperatures, supercritical fluid
processing offers an attractive alternative.
In the field of regenerative medicine, supercritical fluid
processing (principally scCO2) has been applied to:
1. The production of porous polymer scaffolds and composites.
2. Encapsulation and release of growth factors and cells for
tissue engineering.
3. Extraction of residual solvents or other deleterious com-
pounds from pre-fabricated scaffolds.
4. Sterilisation of implantable materials.
4.1. Production of porous polymer scaffolds
4.1.1. Gas foaming
Supercritical fluid processing lends itself to the production
of porous matrices since alternative methods tend to be solvent
and energy intensive and can lead to pore collapse in certain
materials. Furthermore, their compressed state and ability to
plasticize a wide range of polymers lends them to the
formation of polymer foams [76]. Polymer foams are formed
when a polymer, plasticized by saturation in the supercritical
fluid is rapidly depressurized at a constant temperature. As
pressure is released pockets of gas nucleate and grow in the
polymer. As the supercritical fluid leaves the polymer, the Tg
increases. At the point where the Tg for the polymer is higher
than the foaming temperature, the porous structure is set
[77,78].
This can even operate at sub-critical temperatures and pres-
sures. Mooney et al. exposed compression moulded polymer
pellets and solvent cast discs of the poly(α-hydroxyl acids) to
CO2 at a pressure of 5.5 MPa at room temperature. During an
equilibration period of 72 h the polymers saturated with CO2
381 O.R.
O.R.
Davies
Davies
et al.
et al.
/ Advanced
/ Advanced
Drug
Drug
Delivery
Delivery
before rapid depressurization over 15 s. Using this method,
highly porous sponges could be formed (up to 97% for PLGA).
Unlike the amorphous PLGA, PGA, which is crystalline, does
not to foam due to the low amounts of CO2 it absorbs. This fact
was exploited to control porosity by mixing the crystalline and
amorphous materials [79].
However a non-porous film was formed on the surface of the
scaffolds, making them unsuitable for cell ingress, and
presenting a barrier for cell nutrients and the waste products
of cell metabolism. This surface skin is thought to be formed by
the rapid diffusion of the CO2 from the surface of the scaffold as
the pressure is released [78,79]. Furthermore, it was found that
the scaffolds produced had a relatively closed pore-network. To
overcome this issue, an additional porogen (sodium chloride
crystals of a defined size range) was incorporated into com-
pression moulded PLGA discs before gas foaming. Upon pro-
duction of the polymer foam, the porogen was leached out by
incubation in water for 48 h. This created a scaffold with an
interconnected pore network open to the surface, which was
shown to be accessible to cultured smooth muscle cells [80].
Interconnectivity of the pores could be improved even further
by partially fusing the salt porogen by exposure to 95%
humidity [81].
The leaching of the porogen is a major disadvantage of the
gas foaming/particulate leaching process since it results in loss
of the majority of any incorporated growth factor. This can be
overcome somewhat by encapsulating the active factors in
alginate beads which are then incorporated into the compressed
polymer together with the porogen before processing. This was
able to increase the encapsulation efficiency to 55% ± 1% from
28% ± 18%. In this system the released VEGF was shown to
retain N 90% of its activity post-processing [82]. This study also
investigated the use of alternative gases to foam the polymer —
CO2, N2 and He. Carbon dioxide was found to be the most
effective at producing porous scaffolds. The exact mechanism
of this is not known, but was thought could be a result of an
interaction between CO2 and the carbonyl groups of PLGA
[82].
In addition to the issues surrounding salt leaching, the gas
foaming process suffers from relatively long manufacturing time.
Whilst a 1 hour soak in high pressure CO2 has been reported to be
effective in foaming PDLLGA, it produces particularly fragile
scaffolds. Most studies using the technique report a gas equili-
bration period of 12–24 h or more, followed by the leaching step
of at least 18 h to ensure complete removal of the porogen
[79,80,82,83].
Under supercritical conditions, CO2 foams polymers to create
porous scaffolds suitable for tissue engineering applications.
Since increasing pressure, increases the rate of gas diffusion into
polymer systems, equilibration time is reduced compared to sub-
critical pressures [84]. Gas saturation equilibration times as low
as 10 min have been reported [85,86]. The effectiveness of this
technique in foaming PLA and PLGA and use of the scaffolds for
tissue engineering has been demonstrated both in vitro and in
vivo [87–89].
Barry et al. used scCO2 to foam the non-degradable polymer
poly(ethyl methacrylate)/tetrahydrofurfuryl methacrylate (PEMA/
Reviews
Reviews
60 60
(2008)
(2008)
373–387
373–387 381
THFMA), which has been shown to be an excellent scaffold
material for cartilage repair. An 8 hour equilibration time was
required in order to ensure complete permeation of the polymer.
Like the gas foaming technique, a surface skin was formed,
however a greater degree of pore interconnectivity (up to 57%)
was found compared to that reported for the gas foaming
technique applied to PLGA. In vitro cell culture revealed
enhanced phenotypic maintenance of chondrocytes cultured on
the foamed material (as measured by extracellular matrix
secretion), demonstrating the utility of a three-dimensional
environment for promoting tissue regeneration [90].
Further optimisation of the processing conditions of PEMA/
THFMA demonstrated that the degree of porosity and their
interconnectivity could be tightly controlled simply by control-
ling the venting rate. Venting over 30 s reduced porosity, and the
interconnectivity, whilst venting over 60 min allowed the pores
time to grow, and produced scaffolds with a porosity of around
89% with high connectivity (74% open pores). In this process,
an additional porogen/leaching process is not therefore
necessary. As might be expected, as porosity increases,
mechanical strength decreases, an important factor in many
biomedical applications e.g. weight bearing tissues such as
bone. There is therefore an upper limit, above which pore size
and low mechanical strength precludes their use in certain
applications [91].
In attempting to improve the mechanical properties of the
THFMA for soft prosthetic applications, the foaming of blends
of THFMA with styrene–isoprene–styrene copolymer elastomer
(SIS) has been investigated. These blends produce materials
with properties predominantly plastic or elastomeric (depending
upon their composition) whilst maintaining the hydrophilicity of
the THFMA component. Exposure of the blends to scCO2 at
40 °C/100 bar for 8 h and venting over 15 min foamed the
polymers, which contracted upon cooling. There appeared to be
some elastic forces present in the scaffolds which limited the
degree of foaming. Indeed, mechanical testing revealed the
blends to be elastomeric, and these systems may therefore lend
themselves to use as scaffolds for particular tissues, such as
cartilage. The degree of porosity and interconnectivity of the
pores could be tuned by modifying the blend composition and
processing temperature, with a reduction in the degree of
foaming as temperature increased [92].
Similarly, Matheiu et al. have shown that the morphology of
the foams can be controlled to mimic the structure of bone.
Bone from different sites around the body is anisotropic, both
morphologically and mechanically, to varying degrees. They
found that similar anisotropic structures can be formed by
controlling the cooling rate of the foamed polymer — in this
case PLLA — and the density of the gas nucleation. Rapid
cooling locked in large numbers of spherical pores, whilst
slower cooling enabled pore elongation (provided no coales-
cence occurred). Incorporation of ceramic modifiers also influ-
enced the process, reducing porosity and surface area, it was
thought due to modification of the polymer viscosity [85]. Fig. 3
demonstrates the similarity of the polymer foams produced to
cancellous bone from various sites around the body, simply by
varying the processing parameters.
382 O.R.
O.R.
Davies
Davies
et al.
et al.
/ Advanced
/ Advanced
Drug
Drug
Delivery
Delivery
Fig. 3. Showing the similarity of the macrostructures of foamed polymer scaffolds and cancellous bone. Reproduced with kind permission from [85].
Alternatively, the interconnectivity of the pores can be
improved by post-processing the scaffolds generated by gas
foaming using ultrasound. Wang et al. produced polymer foams
of PLA (using sub-critical pressures of CO2) and exposed them
to pulsed ultrasound at a frequency of 20 kHz and average
power input of 100 W. This not only slightly increased pore
size, but also improved their interconnectivity as a result of pore
wall rupture [93].
4.1.2. Supercritical fluid emulsion templating
An alternative to the use of supercritical fluid in the gas
foaming of polymers to create porous scaffolds is emulsion
templating. Here, a variety of porous hydrophilic materials can be
produced by reaction-induced phase separation of concentrated
oil-in-water emulsions e.g. using sol-gel chemistry, or free radical
polymerisation. Removal of the internal phase (i.e. the emulsion
droplets) yields the porous product. Conventional methods
require large amounts of water immiscible organic phases as the
internal phase (usually N 75%), which can be difficult to remove
after the reaction. For example, for inorganic materials firing the
products at temperatures N 600 °C is often used, which would
clearly not be appropriate for thermolabile materials or those
containing growth factors for regenerative medicine.
Butler et al. demonstrated that the emulsion templating
principle can be applied to supercritical CO2-in-water(C/W)
Reviews
Reviews
60 60
(2008)
(2008)
373–387
373–387 382
emulsions. Using perfluoropolyether (PFPE) surfactants and
poly(vinyl alcohol) to stabilize the C/W emulsions of
acrylamide polymers, they vented the CO2 following poly-
merization, leaving polymer scaffolds with interconnected
pores (see Fig. 4). Increasing the volume fraction of the CO2
internal phase increased porosity, and increasing the surfactant
concentration led to more open interconnected structures [94].
A disadvantage of this protocol for bioengineering applica-
tions is the use of a non-degradable surfactant, and the mean
pore sizes (~ 50 μm) produced may not be suitable for all
applications — it has been reported that scaffolds for bone
regeneration for example require pore sizes between 200 and
400 μm [95].
A variation on the emulsion templating technique has re-
cently been reported by Partap et al., who produced emulsion
templated alginate hydrogels in which the scCO2 not only
served as the templating “oil” phase, but also as a source of
acidity used to release Ca2+ from its chelated form, freeing it to
cross-link the alginate and form the porous hydrogel. Calling
the technique “reactive emulsion templating (RET)”, they were
able to produce 3D hydrogels with an open and interconnected
porous network, with a mean pore size similar to the emulsion
templating technique reported by Butler et al., and with me-
chanical stability demonstrated in cell culture media at 37 °C
over at least 4 weeks [96].
383 O.R.
O.R.
Davies
Davies
et al.
et al.
/ Advanced
/ Advanced
Drug
Drug
Delivery
Delivery
Fig. 4. Schematic of the preparation of porous materials using supercritical fluid
emulsion templating. The external phase is an aqueous solution and the internal
droplet phase is scCO2. Reproduced with permission from reference [94].
4.2. Encapsulation and release of growth factors and cells
using supercritical fluids
Supercritical fluid polymer processing has been shown to be
promising technique for the fabrication of porous scaffolds
suitable for tissue regeneration. However whilst a number of
studies have demonstrated the biocompatibility of unmodified
materials in vitro and in vivo, implantation of unmodified
scaffolds is unlikely to be sufficient to promote cell infiltration,
proliferation and differentiation into the desired tissue. To
provide the necessary signals for this to occur, research has
focused upon the incorporation of growth factors, genes or cell
signalling moieties into the scaffold materials, and surface
modification of the polymers in order to promote cell infiltration
and attachment.
Taking advantage of the plasticisation of biocompatible and
biodegradable polymers such as PLA, PLGA and polycapro-
lactone (PCL) when exposed scCO2, Howdle et al. have
demonstrated that substances can be encapsulated within the
liquified polymers. Since the polymers can be processed under
relatively low temperatures and modest pressures, and in the
total absence of organic solvents, the activity of labile molecules
such as proteins can be maintained.
Reviews
Reviews
60 60
(2008)
(2008)
373–387
373–387 383
In this process the solid bioactive material is homogeneously
mixed into the supercritical fluid plasticized polymer using an
impeller [64]. Alternatively, to promote mixing of low loadings
of active material, solutions can be freeze-dried onto the poly-
mer before processing [65]. Following release of the pressure,
porous scaffolds, microparticles or fibres containing the
bioactive material can be produced.
Yang et al. demonstrated the utility of this technique in
encapsulating bone morphogenetic protein 2 (BMP-2) into PLA
scaffolds for bone tissue engineering in an ex vivo model.
Growth factor released from the scaffolds promoted adhesion,
migration, expansion and differentiation of human osteopro-
genitor cells on the scaffolds [97].
Alternatively, the bioactive factor can be homogeneously
mixed together with the polymer powder or solution before
exposure to scCO2, and fabrication of the porous scaffold [82,98–
101]. Recently, Mathieu compared three different mixing
strategies prior to supercritical fluid foaming with the aim of
efficiently incorporating ceramic fillers for us in orthopaedic tissue
engineering. They assessed pre-mixing the ceramic powder with
polymer pellets before a compression moulding step, dispersion of
the ceramic materials into a polymer solution before evaporation
of the solvent, and melt extrusion. They found that in their system
the solvent and melt processing conditions provided the most
homogeneous distribution of the filler, with the melt extrusion
technique producing scaffolds with the best mechanical proper-
ties. Since the ceramic fillers are stable to high temperatures, and
was solvent free they favoured the latter approach [101].
Kim et al. used the gas foaming particulate leaching method to
encapsulate hydroxyapatite (commonly used in bone tissue
engineering) into porous PLGA scaffolds. The hydroxyapatite
(HA) was mixed with the polymer and salt particles prior to
foaming under high pressure CO2, forming scaffolds with an
interconnected porous network, devoid of the surface skin often
seen using this technique. In vitro and in vivo studies revealed
that the gas foamed scaffolds showed significantly higher cell
growth, alkaline phosphatase activity and mineralization com-
pared to similar scaffolds produced by the solvent casting/
particulate leaching method, possibly due to higher exposure of
HA at the scaffold surface [99].
A variation of these techniques is to process pre-encapsulated
materials with supercritical fluids to generate porous scaffolds
impregnated with bioactive factors [83,102]. Hile et al. encap-
sulated basic fibroblast growth factor (bFGF) in a water-in-oil
emulsion with the protein in the aqueous phase and the polymer
(PLGA) in the oil phase, followed by saturation with supercritical
CO2. The CO2 extracts the organic solvent, and upon pressure
release porous polymer foam is produced. This successfully
produced polymer foams which released bFGF in a controlled
fashion, although residual solvent levels remained higher than
accepted pharmacopoieal limits [102]. The ability of supercritical
CO2 to extract organic solvents has been exploited by others to
remove solvents from PLGA scaffolds fabricated using a three-
dimensional printing process. It was found that chloroform could
be efficiently removed from such scaffolds to within pharmaco-
poieal limits relatively quickly (over a matter of hours) using both
a batch and continuous process [72].
384 O.R.
O.R.
Davies
Davies
et al.
et al.
/ Advanced
/ Advanced
Drug
Drug
Delivery
Richardson et al. took pre-encapsulation a stage further by
demonstrating dual growth factor delivery each with different
release kinetics. Platelet derived growth factor (PDGF) was pre-
encapsulated into PLGA microspheres using a double emulsion
process. These microspheres were then mixed with a different
PLGA together with vascular endothelial growth factor (VEGF)
and alginate (previously shown to improve protein encapsula-
tion efficiency) before gas foaming. Following foaming the
particulate and microsphere polymers had fused, creating a
continuous homogeneous matrix.
The PDGF was released more slowly, and could be
controlled by altering the polymer molecular weight, whilst
the VEGF was release more rapidly, it was thought because it
was more closely associated with the surface of the polymer.
Dual delivery of these growth factors with distinct release
kinetics was shown to promote angiogenesis in vivo to a greater
extent than either growth factor given alone [103]. Such
systems, employing the delivery of numerous biologically
active factors, are likely to be essential to control many tissue
regeneration processes which tend to be the result of a complex
temporal interplay between a multitude of growth factors and
cell signals such as interactions with the extracellular matrix. An
alternative approach to particulate fusing involves the use of
laser sintering to create precision and custom shaped biode-
gradable materials containing bioactive materials [104].
Recently it has been shown that a variety of cell types can
actually survive supercritical fluid mixing into polymer
scaffolds. Ginty et al. mixed the myoblastic C2C12 cell line,
3T3 fibroblasts, and primary hepatocytes and chondrocytes into
PDLLA plasticized with scCO2. Not only were the cells
homogeneously distributed throughout the porous scaffolds
produced, but by controlling the processing conditions they
were able to maintain cell viability, functionality and in the case
of the C2C12 cells, their osteogenic capacity. Gene microarray
analysis revealed that the expression of over 4000 genes was
unaffected, and eight genes down-regulated [105].
Indeed, encouraging cells to colonise the entire scaffold,
throughout the porous matrix, is a particular challenge. By using
plasma gas Barry et al. were able to modify the surface
throughout porous PDLLA scaffolds generated using super-
critical fluid. The plasma treated surfaces were shown to
promote cell attachment onto the scaffolds, however the cells
tended to preferentially attach to the periphery of the scaffolds
due to the greater probability of cell contact during culture[106].
By using plasma-polymerized hexane (which is relatively
resistant to cell attachment) to modify the periphery of the
scaffolds, together with the plasma-polymerized allyl amine to
modify the centres of the scaffolds it was possible to encourage
uniform distribution of the seeded cells throughout the scaffolds
[107].
5. Conclusions and future challenges
Supercritical fluids have shown promise in a number of areas.
In anti-solvent applications, there has been considerable com-
mercialisation activity around the controlled preparation of
controlled drug particles, particularly with respect to control of
Delivery
Reviews
Reviews
60 60
(2008)
(2008)
373–387
373–387 384
particle size and morphology. Some work has been carried out
with respect to encapsulation of drugs into polymeric hosts, but
the requirement of large volumes of both conventional solvent
and of scCO2 have proved limiting. Similarly, the RESS process
showed initial promise for drug delivery devices. However, a
significant limitation is the requirement for high solubility of both
the drug and the polymer in scCO2; a relatively rare combination.
A key challenge in this area will be the development of new
highly scCO2 soluble polymers, and promising work in this area is
beginning to be reported [108,109].
The PGSS technique does not require additional solvent, and
in general the drug material should be insoluble in scCO2. Thus
there are significant attractions to the technique. The particle sizes
of the polymer drug composites are limited by the spray process
and the viscosity of the scCO2 polymer melt. A key challenge in
this area will be to demonstrate good product performance with
high quality controlled release data, and perhaps even to drive
down to submicron particle sizes, opening up potential routes to
pulmonary delivery. Finally we have reviewed the applications to
tissue engineering, and in particular the encapsulation of growth
hormones into polymeric scaffolds. The technique has recently
been extended to the encapsulation of mammalian cells, and the
key challenge in this area is now likely to be the development of
implantable tissues, and the introduction of mechanical strength
for load bearing tissue applications.
Overall, it is clear that SCFs have moved beyond the stage of
scientific curiosity, but the key challenge now is to ensure that
real commercial opportunities are developed in drug delivery
and tissue engineering.
References
[1] M.A. Tracy, Development and scale-up of a microsphere protein delivery
system, Biotechnology Progress 14 (1) (1998) 108–115.
[2] H. Okada, One- and three-month release injectable microspheres of the
LH–RH superagonist leuprorelin acetate, Advanced Drug Delivery
Reviews 28 (1) (1997) 43–70.
[3] G. Walsh, Biopharmaceutical benchmarks 2006, Nature Biotechnology
24 (7) (2006) 769–U5.
[4] A.I. Cooper, Polymer synthesis and processing using supercritical carbon
dioxide, Journal of Materials Chemistry 10 (2) (2000) 207–234.
[5] J.L. Kendall, et al., Polymerizations in supercritical carbon dioxide,
Chemical Reviews 99 (2) (1999) 543–563.
[6] D.L. Tomasko, et al., A review of CO2 applications in the processing of
polymers, Industrial & Engineering Chemistry Research 42 (25) (2003)
6431–6456.
[7] H.M. Woods, et al., Materials processing in supercritical carbon dioxide:
surfactants, polymers and biomaterials, Journal of Materials Chemistry
14 (11) (2004) 1663–1678.
[8] S.D. Yeo, E. Kiran, Formation of polymer particles with supercritical
fluids: a review, Journal of Supercritical Fluids 34 (3) (2005) 287–308.
[9] E.J. Beckman, A challenge for green chemistry: designing molecules that
readily dissolve in carbon dioxide, Chemical Communications (17) (2004)
1885–1888.
[10] S. Areerat, et al., Measurement and prediction of diffusion coefficients of
supercritical CO2 in molten polymers, Polymer Engineering and Science
44 (10) (2004) 1915–1924.
[11] S.P. Nalawade, F. Picchioni, L. Janssen, Supercritical carbon dioxide as a
green solvent for processing polymer melts: processing aspects and
applications, Progress in Polymer Science 31 (1) (2006) 19–43.
[12] J.R. Royer, J.M. DeSimone, S.A. Khan, High-pressure rheology and
viscoelastic scaling predictions of polymer melts containing liquid and
385 O.R.
O.R.
Davies
Davies
et al.
et al.
/ Advanced
/ Advanced
Drug
Drug
Delivery
Delivery
supercritical carbon dioxide, Journal of Polymer Science. Part B, Polymer
Physics 39 (23) (2001) 3055–3066.
[13] J.R. Royer, et al., Polymer melt rheology with high-pressure CO2 using a
novel magnetically levitated sphere rheometer, Polymer 43 (8) (2002)
2375–2383.
[14] J.R. Royer, et al., High-pressure rheology of polystyrene melts plasticized
with CO2: experimental measurement and predictive scaling relation-
ships, Journal of Polymer Science. Part B, Polymer Physics 38 (23)
(2000) 3168–3180.
[15] S.M. Lee, et al., High-pressure rheology of polymer melts containing
supercritical carbon dioxide, Korea–Australia Rheology Journal 18 (2)
(2006) 83–90.
[16] M.A. McHugh, V.J. Krukonis, Supercritical fluid extraction: principles
and practice, Butterworth–Heinemann series in chemical engineering,
2nd ed., Butterworth–Heinemann, Boston, 1996, vii, 512.
[17] P. Raveendran, Y. Ikushima, S.L. Wallen, Polar attributes of super-
critical carbon dioxide, Accounts of Chemical Research 38 (6) (2005)
478–485.
[18] T. Sarbu, T. Styranec, E.J. Beckman, Non-fluorous polymers with very
high solubility in supercritical CO2 down to low pressures, Nature 405
(6783) (2000) 165–168.
[19] M. Pantoula, C. Panayiotou, Sorption and swelling in glassy polymer/
carbon dioxide systems — Part I. Sorption, Journal of Supercritical
Fluids 37 (2) (2006) 254–262.
[20] A. Rajendran, et al., Simultaneous measurement of swelling and sorption
in a supercritical CO2-poly(methyl methacrylate) system, Industrial &
Engineering Chemistry Research 44 (8) (2005) 2549–2560.
[21] Y.T. Shieh, K.H. Liu, Solubility of CO2 in glassy PMMA and PS over a
wide pressure range: the effect of carbonyl groups, Journal of Polymer
Research—Taiwan 9 (2) (2002) 107–113.
[22] A.R.C. Duarte, et al., A comparison between gravimetric and in situ
spectroscopic methods to measure the sorption of CO2 in a biocompatible
polymer, Journal of Supercritical Fluids 36 (2) (2005) 160–165.
[23] S.G. Kazarian, et al., Specific intermolecular interaction of carbon
dioxide with polymers, Journal of the American Chemical Society 118 (7)
(1996) 1729–1736.
[24] S.P. Nalawade, et al., The FT-IR studies of the interactions of CO2 and
polymers having different chain groups, Journal of Supercritical Fluids 36
(3) (2006) 236–244.
[25] V.M. Shah, B.J. Hardy, S.A. Stern, Solubility of carbon-dioxide, methane,
and propane in silicone polymers — effect of polymer backbone chains,
Journal of Polymer Science. Part B, Polymer Physics 31 (3) (1993) 313–317.
[26] N.S. Oliveira, et al., Gas solubility of carbon dioxide in poly(lactic acid)
at high pressures, Journal of Polymer Science. Part B, Polymer Physics 44
(6) (2006) 1010–1019.
[27] J.A. Hyatt, Liquid and supercritical carbon-dioxide as organic-solvents,
Journal of Organic Chemistry 49 (26) (1984) 5097–5101.
[28] J.W. Tom, et al., Applications of supercritical fluids in the controlled
release of drugs, ACS Symposium Series 514 (1993) 238–257.
[29] J.H. Kim, T.E. Paxton, D.L. Tomasko, Microencapsulation of naproxen
using rapid expansion of supercritical solutions, Biotechnology Progress
12 (5) (1996) 650–661.
[30] J.W. Tom, P.G. Debenedetti, R. Jerome, Precipitation of poly(l-lactic
acid) and composite poly(l-lactic acid) — pyrene particles by rapid
expansion of supercritical solutions, Journal of Supercritical Fluids 7 (1)
(1994) 9–29.
[31] J.W. Tom, P.G. Debenedetti, Formation of bioerodible polymeric
microspheres and microparticles by rapid expansion of supercritical
solutions, Biotechnology Progress 7 (5) (1991) 403–411.
[32] I.R. Dos Santos, et al., Microencapsulation of protein particles within
lipids using a novel supercritical fluid process, International Journal of
Pharmaceutics 242 (1–2) (2002) 69–78.
[33] K. Matsuyama, et al., Formation of microcapsules of medicines by the
rapid expansion of a supercritical solution with a nonsolvent, Journal of
Applied Polymer Science 89 (3) (2003) 742–752.
[34] K. Mishima, et al., Microencapsulation of proteins by rapid expansion of
supercritical solution with a nonsolvent, AIChE Journal 46 (4) (2000)
857–865.
Reviews
Reviews
60 60
(2008)
(2008)
373–387
373–387 385
[35] I. Kikic, F. Vecchione, Supercritical impregnation of polymers, Current
Opinion in Solid State & Materials Science 7 (4–5) (2003) 399–405.
[36] O. Guney, A. Akgerman, Synthesis of controlled-release products in
supercritical medium, AIChE Journal 48 (4) (2002) 856–866.
[37] S.G. Kazarian, G.G. Martirosyan, Spectroscopy of polymer/drug
formulations processed with supercritical fluids: in situ ATR-IR and
Raman study of impregnation of ibuprofen into PVP, International
Journal of Pharmaceutics 232 (1–2) (2002) 81–90.
[38] K. Gong, J.A. Darr, I.U. Rehman, Supercritical fluid assisted impregnation
of indomethacin into chitosan thermosets for controlled release applications,
International Journal of Pharmaceutics 315 (1–2) (2006) 93–98.
[39] T. Van Hees, et al., Application of Supercritical carbon dioxide for the
preparation of drug-cyclodextrin inclusion compounds, Journal of
Inclusion Phenomena and Macrocyclic Chemistry 44 (1–4) (2002)
271–274.
[40] T. Van Hees, et al., Application of supercritical carbon dioxide for the
preparation of a piroxicam-beta-cyclodextrin inclusion compound,
Pharmaceutical Research 16 (12) (1999) 1864–1870.
[41] J. Bleich, P. Kleinebudde, B.W. Muller, Influence of gas-density and
pressure on microparticles produced with the ASES process, International
Journal of Pharmaceutics 106 (1) (1994) 77–84.
[42] L. Sze Tu, F. Dehghani, N.R. Foster, Micronisation and microencapsula-
tion of pharmaceuticals using a carbon dioxide antisolvent, Powder
Technology 126 (2002) 132–149.
[43] R. Bodmeier, et al., Polymeric microspheres prepared by spraying into
compressed carbon-dioxide, Pharmaceutical Research 12 (8) (1995)
1211–1217.
[44] R. Falk, et al., Controlled release of ionic compounds from poly (L-
lactide) microspheres produced by precipitation with a compressed
antisolvent, Journal of Controlled Release 44 (1) (1997) 77–85.
[45] R.F. Falk, T.W. Randolph, Process variable implications for residual
solvent removal and polymer morphology in the formation of
gentamycin-loaded poly(L-lactide) microparticles, Pharmaceutical
Research 15 (8) (1998) 1233–1237.
[46] A.Z. Chen, et al., Preparation of 5-fluorouracil-poly(L-lactide) micro-
particles using solution-enhanced dispersion by supercritical CO2,
Macromolecular Rapid Communications 27 (15) (2006) 1254–1259.
[47] R. Ghaderi, P. Artursson, J. Carlfors, A new method for preparing
biodegradable microparticles and entrapment of hydrocortisone in DL-
PLG microparticles using supercritical fluids, European Journal of
Pharmaceutical Sciences 10 (1) (2000) 1–9.
[48] A. Engwicht, G.U., B.W. Muller, Critical properties of lactide-co-
glycolide polymers for the use in microparticle preparation by the aerosol
solvent extraction system, International Journal of Pharmaceutics 185
(1999) 61–72.
[49] N. Elvassore, A. Bertucco, P. Caliceti, Production of insulin-loaded poly
(ethylene glycol)/poly(l-lactide) (PEG/PLA) nanoparticles by gas anti-
solvent techniques, Journal of Pharmaceutical Sciences 90 (10) (2001)
1628–1636.
[50] P. Caliceti, et al., Effective protein release from PEG/PLA nano-particles
produced by compressed gas anti-solvent precipitation techniques,
Journal of Controlled Release 94 (1) (2004) 195–205.
[51] H. Okamoto, et al., Pulmonary gene delivery by chitosan–pDNA
complex powder prepared by a supercritical carbon dioxide process,
Journal of Pharmaceutical Sciences 92 (2) (2003) 371–380.
[52] J. Bleich, B.W. Muller, W. Wassmus, Aerosol solvent-extraction
system — a new microparticle production technique, International
Journal of Pharmaceutics 97 (1–3) (1993) 111–117.
[53] A. Breitenbach, D. Mohr, T. Kissel, Biodegradable semi-crystalline comb
polyesters influence the microsphere production by means of a super-
critical fluid extraction technique (ASES), Journal of Controlled Release
63 (1–2) (2000) 53–68.
[54] R. Ghaderi, A.P., J. Carlfors, Preparation of biodegradable microparticles
using solution- enhanced dispersion by supercritical fluids (SEDS),
Pharmaceutical Research 16 (1999) 676–681.
[55] T.J. Young, et al., Encapsulation of lysozyme in a biodegradable polymer
by precipitation with a vapor-over-liquid antisolvent, Journal of
Pharmaceutical Sciences 88 (6) (1999) 640–650.
386 O.R.
O.R.
Davies
Davies
et al.
et al.
/ Advanced
/ Advanced
Drug
Drug
Delivery
Delivery
[56] S. Moshashaee, B.M., R.T. Forbes, E.A. Quinn, H. Nyqvist, P. York,
Supercritical fluid processing of proteins: lysozyme precipitation from
aqueous solution, Journal of Pharmacy and Pharmacology 55 (2003)
185–192.
[57] N. Jovanovic, et al., Stabilization of proteins in dry powder formulations
using supercritical fluid technology, Pharmaceutical Research 21 (11) (2004)
1955–1969.
[58] E. Weidner, Z. Knez, Z. Novak, PGSS (Particles from Gas Saturated
Solutions) — a new process for powder generation, Proc. 3rd Meeting on
Supercritical Fluids, France: ISASF, Strasbourg, 1994.
[59] J. Kerc, et al., Micronization of drugs using supercritical carbon dioxide,
International Journal of Pharmaceutics 182 (1) (1999) 33–39.
[60] P. SencarBozic, et al., Improvement of nifedipine dissolution character-
istics using supercritical CO2, International Journal of Pharmaceutics 148
(2) (1997) 123–130.
[61] J.Y. Hao, et al., Supercritical fluid assisted melting of poly(ethylene
glycol): a new solvent-free route to microparticles, Journal of Materials
Chemistry 15 (11) (2005) 1148–1153.
[62] J.Y. Hao, et al., Plasticization and spraying of poly (DL-lactic acid) using
supercritical carbon dioxide: Control of particle size, Journal of Pharma-
ceutical Sciences 93 (4) (2004) 1083–1090.
[63] M.J. Whitaker, et al., The production of protein-loaded microparticles by
supercritical fluid enhanced mixing and spraying, Journal of Controlled
Release 101 (1–3) (2005) 85–92.
[64] S.M. Howdle, et al., Supercritical fluid mixing: preparation of thermally
sensitive polymer composites containing bioactive materials, Chemical
Communications (01) (2001) 109–110.
[65] M.S. Watson, et al., Incorporation of proteins into polymer materials by
a novel supercritical fluid processing method, Advanced Materials 14
(24) (2002) 1802–1804.
[66] M. Moneghini, et al., Processing of carbamazepine PEG 4000 solid
dispersions with supercritical carbon dioxide: preparation, characterisa-
tion, and in vitro dissolution, International Journal of Pharmaceutics 222
(1) (2001) 129–138.
[67] A.R.C. Duarte, et al., Preparation of controlled release microspheres
using supercritical fluid technology for delivery of anti-inflammatory
drugs, International Journal of Pharmaceutics 308 (1–2) (2006)
168–174.
[68] E. Reverchon, S. Cardea, E.S. Rappo, Production of loaded PMMA
structures using the supercritical CO2 phase inversion process, Journal of
Membrane Science 273 (1–2) (2006) 97–105.
[69] S. Kunastitchai, et al., Application of aerosol solvent extraction system
(ASES) process for preparation of liposomes in a dry and reconstitu-
table form, International Journal of Pharmaceutics 316 (1–2) (2006)
93–101.
[70] A.K. Dillow, et al., Bacterial inactivation by using near- and supercritical
carbon dioxide, Proceedings of the National Academy of Sciences of the
United States of America 96 (18) (1999) 10344–10348.
[71] J. Herberger, et al., Carbon dioxide extraction of residual solvents in poly
(lactide-co-glycolide) microparticles, Journal of Controlled Release 90
(2) (2003) 181–195.
[72] W.S. Koegler, et al., Carbon dioxide extraction of residual chloroform
from biodegradable polymers, Journal of Biomedical Materials Research
63 (5) (2002) 567–576.
[73] K. Koushik, U.B. Kompella, Preparation of large porous deslorelin-
PLGA microparticles with reduced residual solvent and cellular uptake
using a supercritical carbon dioxide process, Pharmaceutical Research 21
(3) (2004) 524–535.
[74] D.A. Edwards, H.J., G. Caponetti, J. Hrkach, A. BenJebria, M.L. Eskew,
J. Mintzes, D. Deaver, N. Lotan, R. Langer, Large porous particles for
pulmonary drug delivery, Science 276 (1997) 1868–1871.
[75] P. Chattopadhyay, R. Huff, B.Y. Shekunov, Drug encapsulation using
supercritical fluid extraction of emulsions, Journal of Pharmaceutical
Sciences 95 (3) (2006) 667–679.
[76] A.I. Cooper, Porous materials and supercritical fluids, Advanced
Materials 15 (13) (2003) 1049–1059.
[77] S.K. Goel, E.J. Beckman, Generation of microcellular polymeric foams
using supercritical carbon-dioxide.1. Effect of pressure and tempera-
Reviews
Reviews
60 60
(2008)
(2008)
373–387
373–387 386
ture on nucleation, Polymer Engineering and Science 34 (14) (1994)
1137–1147.
[78] S.K. Goel, E.J. Beckman, Generation of microcellular polymeric foams
using supercritical carbon-dioxide.2. Cell-growth and skin formation,
Polymer Engineering and Science 34 (14) (1994) 1148–1156.
[79] D.J. Mooney, et al., Novel approach to fabricate porous sponges of poly
(D,L-lactic-co-glycolic acid) without the use of organic solvents, Bioma-
terials 17 (14) (1996) 1417–1422.
[80] L.D. Harris, B.S. Kim, D.J. Mooney, Open pore biodegradable matrices
formed with gas foaming, Journal of Biomedical Materials Research 42
(3) (1998) 396–402.
[81] W.L. Murphy, et al., Salt fusion: an approach to improve pore inter-
connectivity within tissue engineering scaffolds, Tissue Engineering 8
(1) (2002) 43–52.
[82] M.H. Sheridan, et al., Bioadsorbable polymer scaffolds for tissue
engineering capable of sustained growth factor delivery, Journal of
Controlled Release 64 (1–3) (2000) 91–102.
[83] M. Nof, L.D. Shea, Drug-releasing scaffolds fabricated from drug-loaded
microspheres, Journal of Biomedical Materials Research 59 (2) (2002)
349–356.
[84] L. Singh, V. Kumar, B.D. Ratner, Generation of porous microcellular 85/
15 poly ((DL)-lactide-co-glycolide) foams for biomedical applications,
Biomaterials 25 (13) (2004) 2611–2617.
[85] L.M. Mathieu, et al., Architecture and properties of anisotropic polymer
composite scaffolds for bone tissue engineering, Biomaterials 27 (6) (2006)
905–916.
[86] L.M. Mathieu, et al., Bioresorbable composites prepared by super-
critical fluid foaming, Journal of Biomedical Materials Research Part A
75A (1) (2005) 89–97.
[87] K. Partridge, et al., Adenoviral BMP-2 gene transfer in mesenchymal
stem cells: In vitro and in vivo bone formation on biodegradable polymer
scaffolds, Biochemical and Biophysical Research Communications 292
(1) (2002) 144–152.
[88] D. Howard, et al., Immunoselection and adenoviral genetic modulation of
human osteoprogenitors: in vivo bone formation on PLA scaffold,
Biochemical and Biophysical Research Communications 299 (2) (2002)
208–215.
[89] X.B. Yang, et al., Induction of human osteoprogenitor chemotaxis,
proliferation, differentiation, and bone formation by osteoblast stimulat-
ing factor-1/pleiotrophin: osteoconductive biomimetic scaffolds for tissue
engineering, Journal of Bone and Mineral Research 18 (1) (2003) 47–57.
[90] J.J.A. Barry, et al., Porous methacrylate scaffolds: supercritical fluid
fabrication and in vitro chondrocyte responses, Biomaterials 25 (17) (2004)
3559–3568.
[91] J.J.A. Barry, et al., Porous methacrylate tissue engineering scaffolds:
using carbon dioxide to control porosity and interconnectivity, Journal of
Materials Science 41 (13) (2006) 4197–4204.
[92] J.J.A. Barry, et al., Supercritical carbon dioxide foaming of elastomer/
heterocyclic methacrylate blends as scaffolds for tissue engineering,
Journal of Materials Chemistry 15 (46) (2005) 4881–4888.
[93] X.X. Wang, W. Li, V. Kumar, A method for solvent-free fabrication of
porous polymer using solid-state foaming and ultrasound for tissue
engineering applications, Biomaterials 27 (9) (2006) 1924–1929.
[94] R. Butler, C.M. Davies, A.I. Cooper, Emulsion templating using high
internal phase supercritical fluid emulsions, Advanced Materials 13
(19) (2001) 1459–+.
[95] B.D. Boyan, et al., Role of material surfaces in regulating bone and
cartilage cell response, Biomaterials 17 (2) (1996) 137–146.
[96] S. Partap, et al., “Supercritical carbon dioxide in water” emulsion-templated
synthesis of porous calcium alginate hydrogels, Advanced Materials 18
(4) (2006) 501–+.
[97] X.B.B. Yang, et al., Human osteoprogenitor bone formation using
encapsulated bone morphogenetic protein 2 in porous polymer scaffolds,
Tissue Engineering 10 (7–8) (2004) 1037–1045.
[98] H.M. Aydin, E. Piskin, A. Calimli, Microporous scaffolds from poly
(lactide-co-epsilon-caprolactone) composites with hydroxyapatite and
tricalcium phosphates using supercritical CO2 for bone tissue engineering,
Journal of Bioactive and Compatible Polymers 19 (5) (2004) 383–394.
387 O.R.
O.R.
Davies
Davies
et al.
et al.
/ Advanced
/ Advanced
Drug
Drug
Delivery
Delivery
Reviews
Reviews
60 60
(2008)
(2008)
373–387
373–387 387

[99] S.S. Kim, et al., Poly(lactide-co-glycolide)/hydroxyapatite composite [109] B. Tan, et al., Synthesis of CO2-soluble poly(ether carbonate)s and poly
scaffolds for bone tissue engineering, Biomaterials 27 (8) (2006) 1399–1409. (ether ester)s by step growth polymerization, Macromolecules 38 (2005)
[100] Q.H. Sun, et al., Sustained vascular endothelial growth factor delivery 1691–1698.
enhances angiogenesis and perfusion in ischemic hind limb, Pharmaceu-
tical Research 22 (7) (2005) 1110–1116. Key Papers
[101] L.M. Mathieu, P.E. Bourban, J.A.E. Manson, Processing of homogeneous
ceramic/polymer blends for bioresorbable composites, Composites P.J. Ginty, et al., Mammalian cell survival and processing in supercritical CO2,
Science and Technology 66 (11–12) (2006) 1606–1614.
Proceedings of the National Academy of Sciences of the United States of
[102] D.D. Hile, et al., Active growth factor delivery from poly(D,L-lactide-co- America 103 (19) (2006) 7426–7431.
glycolide) foams prepared in supercritical CO2, Journal of Controlled
Release 66 (2–3) (2000) 177–185. S.M. Howdle, et al., Supercritical fluid mixing: preparation of thermally
[103] T.P. Richardson, et al., Polymeric system for dual growth factor delivery, sensitive polymer composites containing bioactive materials, Chemical Com-
Nature Biotechnology 19 (11) (2001) 1029–1034. munications (01) (2001) 109–110.
[104] E.N. Antonov, et al., Three-dimensional bioactive and biodegradable
scaffolds fabricated by surface-selective laser sintering, Advanced
J. Kerc, et al., Micronization of drugs using supercritical carbon dioxide,
Materials 17 (3) (2005) 327–+.
International Journal of Pharmaceutics 182 (1) (1999) 33–39.
[105] P.J. Ginty, et al., Mammalian cell survival and processing in supercritical
CO2, Proceedings of the National Academy of Sciences of the United
D.J. Mooney, et al., Novel approach to fabricate porous sponges of poly(D,L-lactic-
States of America 103 (19) (2006) 7426–7431. co-glycolic acid) without the use of organic solvents, Biomaterials 17 (14) (1996)
[106] J.J.A. Barry, et al., Using plasma deposits to promote cell population of 1417–1422.
the porous interior of three-dimensional poly(D,L-lactic acid) tissue-
engineering scaffolds, Advanced Functional Materials 15 (7) (2005) J.W. Tom, P.G. Debenedetti, Formation of bioerodible polymeric microspheres and
1134–1140. microparticles by rapid expansion of supercritical solutions, Biotechnology Progress
[107] J.J.A. Barry, et al., Using a core-sheath distribution of surface chemistry 7 (5) (1991) 403–411.
through 3D tissue engineering scaffolds to control cell ingress, Advanced
Materials 18 (11) (2006) 1406–+.
[108] T. Sarbu, T.J. Styranec, E.J. Beckman, Design and synthesis of low cost,
sustainable CO2-philes, Industrial & Engineering Chemistry Research 39
(12) (2000) 4678–4683.