Thermochimica
Acta
522 (2011) 110–
117
Contents
lists
available
at
ScienceDirect
Thermochimica
Acta
j o u r n a
l
h
o
m e
p a g e :
w w w . e l s e v i e r . c o m / l o c a t e / t c a
Crystal
polymorphism
of
poly(
l-lactic
acid)
and
its
influence
on
thermal
properties
Maria
Laura
Di
Lorenzo
∗
,
Mariacristina
Cocca,
Mario
Malinconico
Istituto
di
Chimica
e
Tecnologia
dei
Polimeri
(CNR),
c/o
Comprensorio
Olivetti,
Via
Campi
Flegrei,
34,
80078
Pozzuoli
(NA),Italy
a
r
t
i
c
l
e
i
n
f
o
Article
history:
Received
15
October
2010
Received
in
revised
form
16
December
2010
Accepted
22
December
2010
Available online 12 January 2011
Keywords:
Poly(
l-lactic
acid)
Polymorphism
Cold
crystallization
Rigid
amorphous
fraction
Thermal
analysis
a
b
s
t
r
a
c
t
The
influence
of
crystal
polymorphism
on
the
thermal
properties
of
poly(
l-lactic
acid)
(PLLA)
is
discussed
in
this
contribution.
Crystallization
of
PLLA
at
high
temperatures
yields
the
stable
␣
form,
whereas
at
low
temperatures
the
metastable
␣
modification
develops,
which
is
characterized
by
slightly
larger
lattice
dimensions
compared
to
the
␣
counterpart,
and
by
some
degree
of
conformational
disorder.
Quantitative
analysis
with
conventional
and
temperature-modulated
calorimetry
revealed
a
three-phase
structure
of
PLLA
composed
of
a
crystal
phase
and
two
amorphous
fractions
with
different
mobility,
for
all
the
analyzed
thermal
histories.
A
higher
coupling
of
the
amorphous
chain
segments
with
the
crystal
phase
was
found
in
the
presence
of
␣
crystals,
probably
due
to
the
slightly
larger
lattice
dimensions
and
the
looser
arrangements
of
PLLA
chains
in
the
␣
structure.
Some
peculiarities
in
the
thermal
behavior
were
rationalized,
like
an
unusual
frequency-dependence
of
the
reversing
apparent
heat
capacity
upon
the
solid–solid
transition
from
the
␣
to
the
␣
crystals.
Devitrification
of
the
rigid
amorphous
segments
seems
also
to
be
differently
affected
by
the
coupled
crystal
structure
for
the
two
analyzed
crystal
modifications
of
PLLA.
© 2011 Elsevier B.V. All rights reserved.
1.
Introduction
Poly(
l-lactic
acid)
(PLLA)
is
a
biodegradable
and
biocompati-
ble
polyester
that
can
be
produced
by
renewable
resources,
like
corn.
Being
non-toxic
to
human
body,
PLLA
is
used
in
biomedical
applications,
like
surgical
sutures,
bone
fixation
devices,
or
con-
trolled
drug
delivery.
Moreover,
the
good
mechanical
properties
and
easy
of
processability
make
PLLA
a
good
candidate
to
substitute
petroleum-based
polymers
in
selected
and
commodity
application,
with
the
added
value
of
biodegradability.
Similar
to
other
biodegradable
polyesters,
PLLA
displays
crystal
polymorphism,
as
three
main
different
crystal
modifications
can
develop,
named
␣,
,
and
␥
forms,
depending
on
preparation
con-
ditions.
The
␣
form
of
PLLA
grows
upon
melt
or
cold
crystallization,
as
well
as
from
solution.
The
␣
form
has
two
antiparallel
chains
in
a
left-handed
10
3
helical
conformation
(or
distorted
10
3
helix)
packed
in
an
orthorhombic
(or
pseudo-orthorhombic)
unit
cell
with
a
=
1.066
nm,
b
=
0.616
nm,
c
=
2.888
[1–3]
.
Hot-drawing
melt-spun
or
solution-spun
PLLA
fibers
to
a
high-draw
ratio
leads
to
the

form.
An
orthorhombic
unit
cell
with
six
chains
in
the
3
1
helical
conformation,
with
axes
a
=
1.031
nm,
b
=
1.821
nm
and
c
=
0.900
nm
was
first
proposed
for
the

modification
[4]
.
Similar
to
␣
crystals,
∗ Corresponding
author.
Tel.:
+39
081
867
5059;
fax:
+39
081
867
5230.
E-mail
address:
dilorenzo@ictp.cnr.it
(M.L.
Di
Lorenzo).
the
molecular
chains
of
the

form
crystals
have
a
nearly
hexago-
nal
packing,
as
the
b/a
is
very
close
to
√
3
[3]
.
Puiggalí
et
al.
later
suggested
that
the
-form
rests
on
a
frustrated
packing
of
three
3
1
helix
chains
in
a
trigonal
unit
cell
with
parameters
a
=
b
=
1.052
nm,
c
=
0.880
nm,
˛
=
ˇ
=
90
◦
,
=
120
◦
,
with
a
space
group
P3
2
[5]
.
This
frustrated
structure
seems
to
be
formed
to
accommodate
the
ran-
dom
up-down
orientation
of
neighbor
chains
associated
with
the
rapid
crystallization
under
stretching
[5]
.
The
␥
form
is
obtained
via
epitaxial
crystallization
on
hex-
amethylbenzene
substrate.
It
is
characterized
by
two
antiparallel
helices
with
3
1
conformation
packed
in
an
orthorhombic
unit
cell
with
a
=
0.995
nm,
b
=
0.625
nm,
c
=
0.880
nm
[6]
.
The
a
(0.892
nm)
and
b
(0.886
nm)
axes
of
hexamethylbenzene
crystals
are
close
to
the
chain
axis
repeat
distance
of
the
␥
form
of
PLLA
in
the
3
1
heli-
cal
conformation
(0.880
nm).
This
matching
favors
the
epitaxial
growth
of
␥
form
crystals
of
PLLA
on
hexamethylbenzene
crystal
surface.
Besides
these
three
main
crystal
polymorphs,
a
disordered
mod-
ification
of
the
␣
form,
named
␣
form,
was
recently
proposed
for
PLLA.
The
WAXD
patterns
of
the
␣
and
␣
forms
of
PLLA
are
very
similar,
with
small
differences
seen
in
the
shift
to
higher
2
val-
ues
of
the
two
strongest
reflections,
assigned
to
the
(1
1
0)/(2
0
0)
and
(2
0
3)
planes,
and
in
the
appearance
of
a
weak
reflection
at
2
=
≈24.5
◦
in
the
␣
modification.
This
corresponds
to
a
similar
packing
of
the
two
polymorphs,
as,
analogous
to
the
␣
form,
the
PLLA
chains
in
the
␣
modification
have
a
10
3
helix
conformation
and
orthorhombic
(or
pseudo-orthorhombic)
unit
cell
[7–9]
.
The
0040-6031/$
–
see
front
matter ©
2011 Elsevier B.V. All rights reserved.
doi:
10.1016/j.tca.2010.12.027
M.L.
Di
Lorenzo
et
al.
/
Thermochimica
Acta
522 (2011) 110–
117
113
150
100
50
1.5
2.0
2.5
c
p
[J/(K g)]
Temperature (°C)
St-DSC 20°C/min
TMDSC p=60s
TMDSC p=90s
TMDSC p=120s
TMDSC Q-Iso
150
100
50
2
4
6
8
c
p
[J/(K g)]
Temperature (°C)
St-DSC 20°C/min
TMDSC p=60s
TMDSC p=90s
TMDSC p=120s
TMDSC Q-Iso
Tc=85°C 18 h
a
b
Fig.
3.
(a)
Specific
heat
capacity
of
PLLA
after
cold
crystallization
at
85
◦
C
for
66
h.
The
thick
black
solid
line
is
the
total
heat
capacity
by
St-DSC,
the
red,
green
and
blue
lines
are
the
reversing
specific
heat
capacity
measured
by
TMDSC
at
modulation
periods
p
=
60,
90,
120
s,
respectively,
the
yellow
circles
represent
the
reversing
heat
capacity
measured
in
quasi-isothermal
mode
of
modulation,
the
dashed
black
lines
are
the
solid
and
liquid
specific
heat
capacities,
as
taken
from
Ref.
[22]
.
The
St-DSC
data
of
PLLA
after
cold
crystallization
at
85
◦
C
for
18
h
are
also
shown
as
thin
black
solid
line.
(b)
Enlargement
of
the
plot
shown
in
(a)
in
the
area
of
changing
baseline
c
p
.
(For
interpretation
of
the
references
to
color
in
this
figure
legend,
the
reader
is
referred
to
the
web
version
of
this
article.)
of
the
mobile
amorphous
fraction.
In
the
temperature
region
of
the
glass
transition,
a
minor
frequency-dependence
of
the
reversing
heat
capacity
can
be
observed.
The
dynamic
T
g
,
i.e.
the
glass
transi-
tion
originating
from
temperature
modulation
and
obtainable
from
the
reversing
c
p
curve,
is
observed
at
temperatures
slightly
higher
than
the
devitrification
process
deriving
from
linear
heating
(ther-
mal
glass
transition).
This
can
be
explained
considering
that
the
frequencies
related
to
the
ordinary
linear
heating
rates
are
different
from
those
used
in
TMDSC
measurements,
the
latter
being
gener-
ally
higher
[30–32]
.
The
experimental
data
of
Fig.
3
were
used
to
determine
the
three-phase
composition
of
PLLA.
The
heat
capac-
ity
step
at
T
g
accounts
for
a
mobile
amorphous
phase
content
(
w
A
)
of
0.43.
The
crystal
fraction,
measured
by
WAXS,
is
w
C
=
0
.33.
The
rigid
amorphous
fraction
is
quantified
by
difference
using
Eq.
(2)
:
w
C
+
w
A
+
w
RA
=
1
(2)
which
yields
a
value
of
w
RA
=
0
.24
for
PLLA
after
cold
crystallization
at
85
◦
C
for
66
h.
A
notable
thermal
event
appears
in
Fig.
3
b
a
few
degrees
above
completion
of
the
glass
transition.
On
the
basis
of
St-DSC
data
only,
10
5
1.84
1.86
1.88
c
p
[J/(K g)]
Time (min)
Fig.
4.
Time
dependence
of
the
reversing
specific
heat
capacity
of
PLLA
during
quasi-
isothermal
TMDSC
analysis
at
100
◦
C.
this
thermal
event
may
be
interpreted
as
either
a
second
glass
tran-
sition,
followed
by
a
weak
and
broad
exotherm
that
extends
from
about
100
to
130–135
◦
C,
or
as
a
weak
and
broad
endotherm
cen-
tered
around
100
◦
C.
The
appearance
of
a
double
glass
transition
in
PLLA
was
reported
in
a
number
of
papers,
on
the
basis
of
St-
DSC
or
dynamical–mechanical
analyses
[33–35]
.
In
some
cases,
this
second
relaxation
was
ascribed
to
mobilization
of
the
rigid
amor-
phous
fraction.
From
above
the
glass
transition
temperature
up
to
about
100
◦
C
the
apparent
c
p
curve
measured
by
St-DSC
increases
beyond
the
c
p
level
that
corresponds
to
vitrified
rigid
amorphous
fraction.
This
may
be
connected
to
a
partial
devitrification
of
the
RAF,
as
seen
by
comparison
of
the
St-DSC
trace
with
the
base-
line
heat
capacities
drawn
in
Fig.
3
b
on
the
basis
of
the
two-phase
model,
that
accounts
for
the
crystal
phase
and
a
single
amorphous
fraction,
and
of
the
three-phase
model,
that
takes
into
account
the
crystal
and
two
amorphous
fractions
with
different
mobility,
respectively.
Quantitative
analysis
by
TMDSC
in
Fig.
3
b
shows
a
frequency-dependence
of
the
reversing
heat
capacity,
starting
from
80
◦
C,
which
may
indicate
some
reversing
exchange
of
latent
heat
in
this
temperature
range.
This
interpretation
may
be
not
unique,
since
during
devitrification
the
reversing
c
p
is
also
affected
by
the
periodicity
of
temperature
oscillation
[16]
,
as
also
seen
in
the
tem-
perature
range
of
the
glass
transition
of
the
MAF.
However,
the
two
processes
(fusion
and
devitrification)
have
different
response
to
small
oscillations
of
the
temperatures,
and
may
be
distinguished
by
quasi-isothermal
TMDSC
analysis,
which
usually
provides
differ-
ent
outputs
in
the
time
domain
when
a
polymer
is
analyzed
in
the
glass
transition
or
in
the
melting
range.
In
the
first
case
the
revers-
ing
c
p
remains
practically
constant
with
time,
whereas
a
slow
decay
is
observed
upon
reversing
melting
[16]
.
The
time-dependence
of
the
reversing
c
p
of
the
quasi-isothermal
TMDSC
analysis
of
Fig.
3
at
100
◦
C,
i.e.
at
the
peak
temperature
of
the
apparent
small
endotherm,
or
at
the
end
of
the
apparent
T
g
in
the
St-DSC
plot
of
Fig.
3
b,
is
exhibited
in
Fig.
4
.
The
slight
decrease
of
the
reversing
c
p
with
time
reveals
the
occurrence
of
some
reversing
melting,
and
that
the
frequency-dependence
of
the
TMDSC
curves
measured
at
the
underlying
heating
rate
of
2
◦
C/min
is
to
be
linked
to
latent
heat
exchanges
that
cause
an
increase
of
the
computed
reversing
c
p
beyond
the
reversible
c
p
values
[36–38]
.
The
thermal
event
under
analysis
can
therefore
be
linked
to
fusion
of
smaller
and/or
more
defective
crystals,
probably
grown
under
secondary
crystallization,
followed
by
crystallization
of
additional
chain
segments
above
100
◦
C.
No
quantitative
information
on
devitrification
of
the
RAF
cou-
pled
with
␣
crystals
can
be
derived
from
the
data
of
Fig.
3
,
due
to
the
overlapping
of
partial
melting
of
␣
crystals
and
transformation
114
M.L.
Di
Lorenzo
et
al.
/
Thermochimica
Acta
522 (2011) 110–
117
of
the
metastable
␣
structure
into
the
more
stable
␣
crystals.
The
quasi-isothermal
TMDSC
data
intersect
the
two-phase
baseline
at
135
◦
C,
but
the
temperature
at
which
the
reversing
heat
capacities
reaches
the
value
expected
for
full
devitrification
of
the
RAF
is
prob-
ably
affected
by
other
simultaneous
thermal
events,
which
may
increase
the
level
of
the
measured
reversing
c
p
.
Therefore,
it
is
likely
that
devitrification
of
the
RAF
reaches
completion
in
temperature
range
of
the
main
melting
endotherm
[39]
.
Another
noteworthy
feature
of
the
plots
shown
in
Fig.
3
a
is
the
unusual
frequency-dependence
of
the
reversing
heat
capacity
in
the
temperature
range
around
150
◦
C,
in
correspondence
of
the
exothermic
peak
visible
in
the
St-DSC
traces.
This
uncommon
trend
was
observed
for
all
analyzed
crystallization
temperatures,
where
some
amounts
of
␣
crystals
are
present
(85
≤
T
c
≤
125
◦
C).
As
men-
tioned
above,
in
correspondence
of
polymer
melting,
a
decrease
of
modulation
frequency
(or
an
increase
of
amplitude
of
temperature
oscillation),
usually
leads
to
a
higher
apparent
reversing
heat
capac-
ity,
because
a
decrease
in
the
frequency
of
modulation
permits
a
larger
percentage
of
crystalline
material
to
follow
the
modula-
tion
within
a
single
temperature
cycle
[11,35,40–46]
.
Similarly,
an
increase
in
modulation
amplitude
implies
that
a
higher
fraction
of
the
crystallites
that
is
involved
in
the
melting
process
is
added
to
the
reversing
signal.
This
kind
of
dependency
of
the
reversing
c
p
on
the
frequency
of
modulation
is
seen
in
the
data
of
Fig.
3
a,
except
around
150
◦
C,
where
the
data
gained
at
lower
modulation
period
display
a
higher
apparent
reversing
c
p
.
In
order
to
clarify
the
ori-
gin
of
this
unusual
trend,
the
raw
modulated
heat
flow
data
were
analyzed.
Fig.
5
reports
the
modulated
heat
flow
rate
of
PLLA
isothermally
crystallized
at
85
◦
C
for
66
h,
analyzed
by
TMDSC
at
the
underlying
heating
rate
of
2
◦
C/min
and
at
modulation
periods
of
60
and
120
s.
These
data
are
compared
in
Fig.
5
with
the
modulated
heat
flow
rate
with
the
same
modulation
parameters,
without
distortions
caused
by
the
occurrence
of
thermal
processes.
The
latter
curves
were
obtained
by
computer-simulation
from
the
experimental
raw
data
taken
above
completion
of
melting,
i.e.
in
absence
of
thermal
events,
using
the
procedure
detailed
in
Refs.
[38,47]
.
Comparison
of
experimental
and
simulated
heat
flow
rate
data
allows
to
deter-
mine
the
latent
heat
exchanged
during
each
oscillation
period.
Above
160
◦
C,
in
the
region
of
the
main
melting
peak,
large
dis-
tortions
in
the
experimental
curves
can
be
observed
in
both
their
endothermic
and
exothermic
parts,
and
the
effect
is
much
larger
at
higher
modulation
period.
As
a
result,
the
modulated
heat-flow-
rate
amplitude,
and
in
turn
the
reversing
c
p
,
increases
with
the
period
of
temperature
oscillation.
At
lower
temperatures,
around
150
◦
C,
the
modulated
heat-flow-rate
curves
are
deformed
to
a
lower
extent,
but
still
both
endothermic
and
exothermic
events
can
be
detected,
which
are
linked
to
partial
melting
and
to
the
ongoing
phase
transformation
from
the
metastable
␣
structure
to
the
stable
␣
form.
In
the
area
of
interest,
highlighted
by
the
arrow
in
Fig.
5
b,
in
the
experimental
curve
gained
at
p
=
120
s
endother-
mic
events
take
place
during
the
heating
segment
around
150
◦
C.
The
initial
increase
in
the
heat-flow-rate,
caused
by
the
switch
to
a
different
scanning
rate
overlapping
partial
melting,
is
followed
by
exothermal
effects,
as
revealed
by
comparison
with
the
sim-
ulated
data.
The
initial
increase
of
amplitude
of
modulated
heat
flow
rate
linked
to
latent
heat
release
is
followed
by
a
decrease
of
the
oscillation
amplitude,
as
the
experimental
modulated
heat
flow
rate
curve
falls
below
the
simulated
plot
before
the
switch
to
the
next
oscillation
segment.
Such
a
decrease
of
the
experi-
mental
data
below
the
level
corresponding
to
the
simulated
curve
is
not
seen
in
the
curve
gained
at
p
=
60
s,
shown
in
Fig.
5
a,
due
to
the
short
modulation
period.
Similarly,
in
the
preceding
half-
cycle
at
p
=
120
s,
the
exotherm
overlaps
endothermic
latent
heat
exchange,
and
again,
crosses
the
simulated
curve
before
the
end
of
the
modulation
half-period.
The
overall
result
is
that,
in
the
case
160
150
140
-5
0
5
10
Modulated
Φ
(W/g)
Temperature (°C)
p=60s (exp)
p=60s (calc)
160
150
140
-5
0
5
10
Modulated
Φ
(W/g)
Temperature (°C)
p=120s (exp)
p=120s (calc)
a
b
Fig.
5.
Experimental
and
simulated
heat-flow
rates
of
PLLA,
obtained
after
isother-
mal
crystallization
at
85
◦
C
for
66
h:
(a)
p
=
60
s
and
(b)
p
=
120
s.
of
long
period
of
oscillation,
p
=
120
s,
despite
the
larger
percentage
of
crystalline
material
that
follows
the
modulation
within
a
single
temperature
cycle,
the
neat
latent
heat
that
is
exchanged
in
each
modulation
cycle
(endothermic
minus
exothermic
heat)
is
lower
than
when
lower
periods
of
temperature
oscillation
are
used.
This
results
in
a
lower
amplitude
of
modulated
heat
flow
rate,
when
the
experimental
data
are
approximated
with
a
Fourier
series
in
each
modulation
cycle,
and
in
turn
in
a
lower
apparent
reversing
c
p
,
as
seen
in
Fig.
3
a
around
150
◦
C.
The
thermal
analysis
of
PLLA
after
isothermal
cold
crystalliza-
tion
at
145
◦
C
for
18
h
is
presented
in
Fig.
6
a,
with
an
enlargement
of
the
c
p
data
shown
in
Fig.
6
b.
As
revealed
by
the
WAXS
plots
of
Fig.
2
,
this
thermal
history
leads
to
development
of
the
␣ crystal
modifica-
tion
only,
and
a
single
major
melting
endotherm
appears
in
the
DSC
plots
of
Fig.
6
.
The
glass
transition
of
the
MAF
is
centered
at
64
◦
C,
a
few
degrees
below
the
T
g
of
the
polymer
crystallized
at
85
◦
C
for
66
h
(T
g
=
66
◦
C).
This
slight
decrease
of
the
T
g
of
PLLA
at
increas-
ing
crystallization
temperatures,
very
close
to
the
experimental
uncertainty,
is
in
agreement
with
literature
data
[48]
.
From
the
heat
capacity
step
at
T
g
a
mobile
amorphous
fraction
w
A
=
0
.31
is
measured,
which,
compared
to
the
w
A
=
0
.43
computed
after
crys-
tallization
at
85
◦
C
for
66
h,
indicates
that
crystallization
at
higher
temperatures
leads
to
a
reduction
of
the
MAF
content.
The
crys-
tal
fraction
measured
by
WAXS
after
crystallization
at
145
◦
C
is
w
C
=
0
.45,
which
leads
to
a
rigid
amorphous
content
w
RA
=
0
.24.
Above
completion
of
the
glass
transition,
the
St-DSC
and
the
TMDSC
data
of
Fig.
6
,
including
the
quasi-isothermal
analysis,
over-
M.L.
Di
Lorenzo
et
al.
/
Thermochimica
Acta
522 (2011) 110–
117
115
150
100
50
2
4
6
8
c
p
[J/(K g)]
Temperature (°C
)
St-DSC 20°C/min
TMDSC p=60s
TMDSC p=90s
TMDSC p=120s
TMDSC Q-Iso
150
100
50
1.5
2.0
2.5
c
p
[J/(K g)]
Temperature (°C
)
St-DSC 20°C/min
TMDSC p=60s
TMDSC p=90s
TMDSC p=120s
TMDSC Q-Iso
a
b
Fig.
6.
(a)
Specific
heat
capacity
of
PLLA
after
cold
crystallization
at
145
◦
C
for
18
h.
The
black
line
is
the
total
heat
capacity
by
St-DSC,
the
red,
green
and
blue
lines
are
the
reversing
specific
heat
capacity
measured
by
TMDSC
at
modulation
periods
p
=
60,
90,
120
s,
respectively,
the
yellow
circles
represent
the
reversing
heat
capacity
measured
in
quasi-isothermal
mode
of
modulation,
the
dashed
lines
are
the
solid
and
liquid
specific
heat
capacities,
as
taken
from
Ref.
[22]
.
(b)
Enlargement
of
the
plot
shown
in
(a)
in
the
area
of
changing
baseline
c
p
.
(For
interpretation
of
the
references
to
color
in
this
figure
legend,
the
reader
is
referred
to
the
web
version
of
this
article.)
lap
up
to
130
◦
C,
in
correspondence
of
the
onset
of
the
melting
endotherm,
which
starts
at
temperatures
slightly
lower
than
T
c
,
probably
due
to
some
residual
crystallization
during
cooling
to
room
temperature.
The
overlapping
of
reversing
and
total
c
p
data
reveals
that
negligible
latent
heat
exchanges
take
place
up
to
the
onset
of
melting
and
that
the
increase
of
the
experimental
c
p
values
up
to
the
beginning
of
melting
is
to
be
linked
to
devitrification
of
the
rigid
amorphous
segments
of
PLLA.
Unfortunately,
the
overlap-
ping
of
reversing
melting
to
the
c
p
increase
due
to
devitrification,
before
the
intersection
of
the
experimental
reversing
c
p
data
with
the
two-phase
baseline,
does
not
allow
to
estimate
the
exact
point
of
full
devitrification
of
the
RAF
of
PLLA
coupled
with
the
␣
crystals,
which
however
seems
to
attain
full
mobility
at
temperatures
close
to
the
onset
of
crystal
melting.
The
three-phase
composition
of
PLLA
after
isothermal
cold
crystallization
at
various
temperatures
is
illustrated
in
Fig.
7
.
Crys-
tallinity
increases
with
the
crystallization
temperature
in
the
whole
analyzed
range,
with
a
discontinuity
around
110–120
◦
C,
as
often
reported
in
the
literature
[49–53]
.
This
irregular
trend
is
to
be
linked
to
growth
of
PLLA
crystals
in
the
two
different
polymorphs
and
the
corresponding
varied
crystallization
kinetics.
The
mobile
amor-
phous
fraction
decreases
with
T
c
for
all
the
analyzed
crystallization
temperatures,
with
the
only
exception
of
a
slightly
lower
w
A
value
160
140
120
100
80
0.2
0.4
0.6
w
C
, w
A
, w
RA
T
c
(°C)
w
C
w
A
w
RA
Fig.
7.
Crystalline
(
w
C
),
mobile
amorphous
(
w
A
),
and
rigid
amorphous
(
w
RA
)
frac-
tions
of
PLLA
after
isothermal
cold
crystallization
at
various
T
c
.
measured
after
crystallization
at
85
◦
C
compared
to
crystallization
at
95
◦
C.
This
decrease
is
probably
related
to
the
extended
crystal-
lization
time
at
85
◦
C
(66
h)
compared
to
the
other
T
c
(18
h),
which
may
induce
a
larger
extent
of
secondary
crystallization,
with
possi-
ble
insertion
of
thin
lamellae
in
the
interlamellar
amorphous
layer
or
interlamellar
stacks
[54]
.
The
broad
and
weak
endotherm
in
the
DSC
data
of
Fig.
3
around
100
◦
C,
discussed
above,
confirms
the
exis-
tence
of
a
small
population
of
thin
and
defective
lamellae
with
very
poor
thermal
stability
in
PLLA
crystallized
at
85
◦
C
for
66
h.
It
is
interesting
to
note
that
a
decrease
of
the
crystallization
tem-
perature
from
95
to
85
◦
C
corresponds
to
a
considerable
increase
of
the
rigid
amorphous
fraction.
Upon
melt
crystallization,
it
has
been
often
reported
that
a
higher
rate
of
crystal
formation,
like
upon
crystallization
at
T
c
=
95
◦
C
compared
to
T
c
=
85
◦
C
[50]
,
leads
to
short
times
for
the
adjustment
of
the
crystals
into
the
locally
ener-
getically
most
favorable
states.
Internal
stresses
are
not
released
during
crystal
growth,
and
concentrate
at
the
interface
between
the
crystal
and
amorphous
phases,
resulting
in
a
large
rigid
amorphous
fraction
[16]
.
The
formation
of
secondary
crystals
upon
prolonged
crystallization
at
85
◦
C
may
also
be
linked
to
an
increased
coupling
between
the
amorphous
and
crystalline
areas:
rearrangements
of
the
amorphous
regions
localized
in
proximity
of
the
growing
sec-
ondary
lamellae
are
subjected
to
geometrical
restrictions,
in
which
the
melt
undergoes
larger
constraints,
with
consequent
forma-
tion
of
rigid
amorphous
phase.
This
may
confirm
the
hypothesis
often
appeared
in
the
literature,
of
a
connection
between
secondary
crystallization
and
vitrification
of
the
rigid
amorphous
segments
[40–46]
.
A
reduction
in
crystallization
temperature
corresponds
also
to
lower
chain
mobility,
especially
in
the
temperature
range
under
consideration,
that
is
very
close
to
the
glass
transition
of
the
mobile
amorphous
fraction,
which
may
complicate
rearrange-
ments
of
the
chains
at
the
crystal-amorphous
boundary,
leading
to
an
increased
fraction
of
amorphous
material
under
local
stress
at
the
crystal
surfaces.
The
w
RA
vs.
T
c
plot
of
Fig.
7
displays
a
maximum
at
125
◦
C.
At
high
temperatures
w
RA
decreases
with
T
c
,
as
commonly
reported
in
the
literature
for
a
number
of
semicrystalline
polymers
[16]
.
At
95
◦
C
≤
T
c
≤
125
◦
C
the
opposite
trend
can
be
observed.
An
increase
in
crystallization
temperature
in
this
range
corresponds
to
a
larger
fraction
of
␣ to
␣
crystals,
as
shown
in
Fig.
2
[55]
.
It
is
likely
that
crystallization
of
PLLA
into
the
ordered
␣
modification
leads
to
a
larger
coupling
of
the
amorphous
and
crystalline
chain
segments,
compared
to
the
conformationally
disordered
␣
arrangement.
A
similar
influence
of
the
varied
order
in
the
crystal
structure
was
