A Real-Time Study of Homogeneous Nucleation, Growth, and Phase
Transformations in Nanodroplets of Low Molecular Weight Isotactic
Polypropylene Using AFM
Lekshmi Kailas,†Cvetelin Vasilev,†Jean-Nicolas Audinot,‡Henri-Noe 1l Migeon,‡and
Jamie K. Hobbs*,†
Departments of “Physics & Astronomy” and Chemistry, UniVersity of Sheffield, Sheffield S3 7RH, UK,
and De ´partement Science et Analyse des Mate ´riaux CRPGL, 41 rue du Brill,
L-4422 BelVaux, Luxembourg
ReceiVed April 12, 2007; ReVised Manuscript ReceiVed July 20, 2007
ABSTRACT: Nanodroplets of isotactic polypropylene (iPP) were observed using temperature-controlled AFM
in order to study the polymer’s crystallization behavior. The nucleation, growth, and transformation of iPP crystals
on heating have been directly imaged. The strong confinement of the polymer into nanoscale droplets has allowed
the controlled observation of polymer nucleation as well as access to crystal growth at exceptionally high
supercooling in iPP. Different modes of crystal growth were observed depending on the film thickness, including
the formation of multiple independent homogeneous nuclei within single droplets. The temperature at which the
onset of nucleation was observed in individual droplets was found to be dependent on the thickness as well as
the volume of the droplets. For droplets with smaller thicknesses (<5 nm), the thickness of the droplets was
found to be the dominating factor influencing the nucleation temperature. This is the first real-space, real-time
observation of homogeneous nucleation in iPP, almost 125 °C below its melting point, which could signify crystal
growth in the smectic form, and of the subsequent reorganization into the R-form.
Despite its fundamental importance both to the theoretical
understanding of polymer crystallization and to the industrial
application of semicrystalline polymers, the process of crystal
nucleation in polymeric systems remains infrequently studied
and poorly understood due to the difficulty in observing the
onset of a nucleation event. Crystallization is thought to occur
via a two-step process of nucleation and crystal growth, although
recent hypotheses suggest the involvement of many multistep
processes during the initial stages of crystallization from a
polymer melt, the possibility of an intermediate ordering process,
and transient liquid phases preceding the final crystalline
state.1-3The intermediate, partially ordered states that lie
between the melt and final crystal phase are thought to play a
crucial role in crystal growth.4Therefore, gaining new insights
at the molecular level into the formation of the primary nucleus,
its growth, and phase transformations is imperative in order to
understand fully the mechanisms that govern crystallization.
Here, we use the capability of atomic force microscopy (AFM)
to obtain real-space, real-time images to shed further light on
the processes of nucleation and crystal growth.
Isotactic polypropylene (iPP) is known to exhibit several
crystalline polymorphs (monoclinic R-form, hexagonal ?-form,
orthorhombic γ-form), depending on the crystallization condi-
tions and nucleating agents used; all having packing schemes
based on a 3-fold helical conformation. Slow cooling from the
melt is known to produce predominantly the R-modification;
the usage of specific nucleating agents yields the ?-form while
elevated pressure and low degrees of supercooling results in
the γ-phase.5-15The R-phase is found to be the more stable
phase at atmospheric pressures.16On rapid quenching from the
melt, it is also reported to reveal a mesomorphic or “smectic”
phase, the exact nature and structure of which are still not fully
known. It has been suggested that this metastable phase can be
equated with the onset of homogeneous nucleation.5
There have been previous attempts to follow nucleation using
AFM.17-24However, what is observed remains ambiguous
because of the surface specific nature of the AFM techniques
as the polymer film is relatively thick, it is not possible to say
whether what is seen is a nucleation event or the growth of a
crystal from a (subsurface) nucleus that crosses the sample
surface. Secondary nucleation has been observed in several
cases, both from lamellae25-27and from oriented “shish”
structures.18In the case of rapidly crystallizing polymers such
as iPP (and most other commercial polymers) the probability
of observing the nucleation event in a bulk sample or thin film
is vanishingly small. In most circumstances crystallization in a
polymer melt is initiated by foreign particles. The aim of the
current study is to follow homogeneous nucleation and subse-
quent crystal growth. To achieve this, we made use of the
classical droplet method28in order to allow the formation of
small volumes of highly supercooled melt, which are free from
nucleating impurities. The idea is to break the sample up into
a large number of spatially isolated droplets such that the number
of droplets is greater than the number of active heterogeneities,
so that homogeneous nucleation can be observed in some of
these impurity-free droplets. AFM is utilized to directly image
the nucleation process at the molecular scale.
Previous studies of polymer nucleation using the “droplet
method”29-32have found that, once a sufficiently small droplet
size is obtained, the rate of nucleation depends only on the
supercooling and the droplet volume, in agreement with classical
theories.33This is usually interpreted as strong evidence for
homogeneous nucleation. In ref 31 it is argued convincingly
that the observed behavior discounts the possibility of external
interfaces of any kind (including a droplet’s surface), causing
or enhancing the nucleation rate over a range of length scales
* Corresponding author: Tel +44 114 222 9316; e-mail jamie.hobbs@
†University of Sheffield.
10.1021/ma070861t CCC: $37.00 © xxxx American Chemical Society
Published on Web 09/07/2007PAGE EST: 7.4
from microns to just 10 nm under the experimental situation
studied in those experiments (i.e., a dewetting surface). How-
ever, the question still remains as to what happens at smaller
length scale, as the size scale of the droplet approaches or
becomes smaller than the critical nucleus size in at least one
dimension. Indeed, it is precisely in such confined geometries
where nucleation has an enhanced effect, as the eventual crystal
orientation will be strongly controlled by that of the nuclei as
there is no opportunity, within the small confined volume, for
the subsequent growth to introduce its usual dominance.
In this article we present AFM images of the onset of
nucleation in iPP nanodroplets at high supercooling, the growth
and ensuing morphology of the highly supercooled crystalline
phase, and the phase transformation processes undergone by
the iPP crystals on heating. Different modes of crystal growth
were observed in these droplets depending on the film thickness
as well as the volume of the droplets. In order to ascertain
whether the nanodroplets were truly isolated, we made use of
the depth profiling, chemical mapping technique of NanoSIMS.
Low molecular weight iPP (Mw: 12 000; Mn: 5000) from Sigma-
Aldrich was used as received. It was dissolved at a concentration
of 1 mg/mL in p-xylene at 90 °C. The hot solution was dropped
onto a silicon wafer (plasma treated for 30 min before being placed
on the spin-coater), spinning at the rate of 5000 rpm for a time
period of 30 s. The ultrathin film obtained (thickness <10 nm)
was annealed for 10 min at 180 °C in a nitrogen atmosphere on a
Linkam hot stage to allow dewetting into isolated droplets and then
cooled to room temperature. This relatively low molecular weight
sample was selected as it allowed the dewetting process to be
adopted in order to create the nanodropletsshigher molecular
weight material does not dewet sufficiently fast to avoid sample
degradation. NanoSIMS was carried out on the sample to see
whether the isolation was complete or whether there was a thin
layer of the polymer connecting the various droplets. The dewetted
iPP droplets obtained on annealing were examined at various
temperatures after remelting at 180 °C in a nitrogen atmosphere,
quenching the sample by transferring it to a hot stage preheated to
60 °C kept under the AFM scanner18and then slow cooling at about
0.3 °C/min from the melt while imaging with the AFM. A Peltier
device kept in contact with the underside of the Linkam hot stage
allows the lowering of temperature down to 0 °C.34The temper-
atures quoted are those displayed on the Linkam heater and are
found to deviate from the actual sample surface temperature by
less than a degree over the temperature range 20-40 °C.
Atomic force microscopy observations were carried out in air
using a Digital Instruments D3100 AFM operated in Tapping Mode.
Measurements were carried out using standard silicon cantilevers
with a nominal spring constant of 50 N/m and resonant frequency
of ∼300 kHz. Height and phase data were collected simultaneously.
NanoSIMS experiments were performed using the Cameca
NanoSIMS 50 in the raster imaging mode. The ionic images were
obtained layer by layer using a beam of Cs+primary ions with an
energy impact of 16 keV and a beam current of 0.8 pA on the
sample (sputter rate ∼ 10 nm/s). Negative secondary ion signals
for12C-,12C1H-,16O-, and28Si-were collected simultaneously.
Results and Discussion
We focus our attention on isolated droplets formed on the
silicon wafer by a dewetting mechanism in order to observe
homogeneous nucleation and crystalline growth in small con-
fined volumes. Figure 1a shows the AFM phase image of the
iPP film which has undergone dewetting, in this case a nucleated
dewetting process which helps to leave a final sample that is
free from heterogeneities.31Larger impurities present in the
sample tend to initiate nucleation and facilitate the dewetting
of the film into flower-shaped holes. Inside the boundary of
these holes, the dewetting film gives rise to hundreds of tiny
droplets, many of which will be completely devoid of even tiny
impurity particles, and these droplets are ideal sites for observing
homogeneous nucleation. Figure 1b gives a closer view of the
isolated iPP droplets on the surface of the plasma-treated silicon
wafer, after annealing to 180 °C and then cooling to room
A range of polymer droplets of different sizes and shapes is
apparent. As the dewetting process was not left to reach
equilibrium, the droplet shape is highly flattened, with typical
droplets measuring up to microns in diameter but still only a
few nanometers in height above the substrate. It is evident from
Figure 2 (NanoSIMS image) that these droplets are essentially
isolated. The SiOxlayer on the substrate is responsible for the
high-intensity16O signal coming from the medium surrounding
the droplets, indicating that there is no interconnecting layer of
polymer between the droplets.
In Figure 1b the presence of a population of very small spots
evenly distributed throughout the sample surface is apparent
(arrowed), the nature of which is unclear. As the size of these
spots are lower than the 50 nm lateral resolution of NanoSIMS,
no conclusive evidence can be drawn from the NanoSIMS data
Figure 1. AFM phase images showing the dewetted iPP film on plasma-treated silicon substrate. (a) A classic dewetting pattern; scale bar represents
5 µm. (b) A closer view of the isolated iPP droplets; scale bar represents 1 µm. The arrows indicate small impurity spots. Black to white represents
a change in phase of 70°.
Kailas et al.
regarding the chemical composition of these spots. However,
as they do not change throughout our experiments, and do not
appear to interfere with or initiate crystal growth, we assume
them to be either lower molecular weight fragments of poor
tacticity or, more probably, non-nucleating impurities possibly
inherent in the commercial sample.
The melting temperature of this polymer is 157 °C as
measured from DSC data. The polymer was remelted at 180 °C
so that the sample would be devoid of any crystal nuclei that
could initiate crystallization. After quenching it and cooling
slowly from 60 °C, in-situ AFM studies were carried out to
track the changes undergone by the various droplets on the
sample surface. A series of experiments were conducted on
similar samples prepared under similar conditions.
From these experiments, we were able to observe three
different types of nucleation behavior and crystal growth patterns
in the dewetted droplets. The most common behavior observed
was that of instantaneous nucleation and crystal growth in
isolated droplets, where the rate of growth was too rapid for
the conventional AFM to follow, such that we obtained images
of a molten droplet in one frame which had crystallized entirely
by the time the next frame was imaged. Apart from this rapid
crystal growth behavior, we were also able to observe slow
crystal growth initiated by two distinguishable nucleation
methodssone by the formation of a single nucleus in an isolated
droplet and the other by the formation of multiple nuclei inside
one droplet. Detailed interpretations of AFM images and
possible explanations for these different behavior patterns are
Figure 3 gives a series of AFM phase images which show
the “classical” instantaneous nucleation and rapid crystal growth
behavior while the sample was cooled from 60 °C. In these
phase images, the amorphous areas appear as darker and the
crystallized ones as brighter areas, with the silicon having a
similar phase contrast to the crystalline polymer. Artifacts like
the white rims around the droplets in Figure 3 appear in phase
images as these images are influenced by the surface topography
at sharp edges, steps, etc., where the error in the feedback loop
is large. The white arrows in Figure 3a show the position of
the droplets still in the molten state. The white arrows in the
subsequent images show the position of these droplets which
have subsequently undergone crystallization. The droplets are
found to have thicknesses in the range 5-10 nm.
Figure 4 shows the AFM phase images of the nucleation from
a single nucleus and crystalline growth within one of the droplets
(diameter of ∼700 nm) on cooling at the rate of 0.3 °C/min.
On reaching a temperature of 34.8 °C, nucleation is detected
(Figure 4b). At 33.6 °C, the nucleus is seen to have grown to
have a crystal morphology that is neither cross-hatched nor
granular but probably best described as axialitic (Figure 4c).
As AFM is a serial technique, with images being collected line
by line through a continuous raster scanning of the surface, it
is possible to misinterpret the sequence of events shown in a
series of images. In this case all images were collected while
scanning from the top to the bottom. So, the lower, amorphous,
section of the droplet in Figure 4b was imaged by the AFM tip
after the upper section of the droplet containing the nucleus
had been scanned. Therefore, the growth of the crystal shown
in Figure 4c was not instantaneous but did in fact grow from
the nucleus in Figure 4b at a rate that was slower than the “slow-
scan” velocity of the AFM tip. The droplet is very thin (∼2 nm
as measured by AFM), so it is highly unlikely that there are
any heterogeneities within the droplet to cause nucleation as
these would be imaged by the AFM. The only possible
nucleating “agent” is the substrate itself. Also, there is no
possibility of significant growth having occurred below the
surface imaged, as any hard object larger than a few nanometers
would be seen. We therefore suggest that the crystalline object
seen in Figure 4b is the primary nucleus having dimensions of
120 × 50 nm, captured soon after reaching critical size, and
that this nucleus has formed homogeneously. Studies on natural
rubber imply that the homogeneous nucleus can be no longer
than the lamellar thickness.35Therefore, the critical size is most
probably substantially smaller than the dimensions mentioned
abovesour measurement gives an upper bound. Further cooling
shows no change in the crystal size or shape, as seen from Figure
We can estimate a maximum possible growth rate of the iPP
at this temperature from these images, as the distance from the
nuclei to the lowermost point of the structure (Figure 4c) divided
by the time for the 20 scan lines of the image to be taken when
this point was first past in Figure 4b. This estimation is
complicated by the fact that AFM is a serial technique, the image
being collected line by line. Although the crystal growth might
have taken longer to reach this point, it cannot have got there
faster than this rate, as in that case the growing crystal would
have overtaken the scanning tip and we would have seen it in
the image. This calculation gives 40 nm s-1as the maximum
possible growth rate at this temperature in a ∼2 nm thick film,
with 1.7 nm s-1as the minimum possible rate (i.e., the distance
Figure 2. NanoSIMS images of the outermost surface layer. The image on the left shows12C signal while that on the right shows16O signal. The
color bar from black to red signifies increasing signal intensity.
Nanodroplets of Isotactic Polypropylene
grown/the time interval between two consecutive scans). The
relatively slow growth rate as opposed to the rapid growth rate
implied in ref 36 is most probably due to the high level of
confinement experienced by the very thin droplet. At these high
supercoolings, close to the glass transition temperature, the rate
of material transport in such a thin film must considerably reduce
the crystal growth rate.
As seen in these images, even on cooling the sample, some
of the iPP in the droplets is left uncrystallized (the darker
background that retains the original droplet shape); it is probable
that there is some lower molecular weight and low tacticity
component of the sample that is effectively noncrystallizable.
These could be excluded from the growing crystal, and this
exclusion process would be expected to reduce the growth rate.
Figure 3. A series of AFM phase images showing instantaneous nucleation and crystal growth. The color scale black to white represents a change
in phase of 50°, and the scale bar represents 1 µm. The white rims to the droplets are an imaging artifact (see text). The images were taken at (a)
40, (b) 38, (c) 37.5, and (d) 37 °C.
Figure 4. A series of AFM phase images showing the crystallization on cooling (top row) and reorganization on heating (bottom row) of an iPP
droplet. Black to white represents a change in phase of 50° (a-e) and 80° (f-j), and the scale bar represents 500 nm. (a) Taken at 40, (b) 34.8, (c)
33.6, (d) 23.2, (e) 15, (f) 60, (g) 80, (h) 90, (i) 97, and (j) 102 °C.
Kailas et al.
Most probably, the slow growth rate seen is due to a combina-
tion of sample purity and film thickness effects. Considering
the behavior of the other (thicker) droplets such as those in
Figure 3, we suggest film thickness to be the dominant factor
in this case.
Section profiles of the droplet shown in Figure 4 were taken
and are given in Figure 5. The height of the molten droplet
shown in Figure 4a was found to be ∼2 nm. The initial nucleus
in Figure 4b was found to have a thickness of ∼2.8 nm. The
section profile shows that by the time the crystalline growth
has developed to the shape shown in Figure 4c, it has a height
of ∼5.8 nm. The initial, as nucleated, crystal thickness is
surprisingly thin and rapidly thickens between images to the
final thickness that it then maintains. Such thickening growth
at these high supercoolings is indicative of substantial mobility
within the crystalline phase, supporting the suggestion, discussed
in the next section, that crystal growth is in the smectic phase.
Previous studies at comparable supercoolings have found that
iPP crystallizes in the smectic phase.8,37,38Here, we have used
the in-situ capabilities of the AFM to image the same area on
heating. Figure 4f-j show a series of phase images taken on
heating the sample from 60 to 105 °C, tracking the crystalline
structure inside the droplet. The slight distortion in the droplet
size could be attributed to the mechanical changes associated
with scanning at higher temperatures. The fact that the
background appears darker on reheating most probably occurs
as there are changes in the physical properties of the polymer
with an increase in temperature (i.e., it becomes softer and more
sticky). From the AFM images, it appears that the crystal,
initially formed at 33.6 °C (Figure 4c), remains unaltered even
on heating to 80 °C (Figure 4g). But the image taken at 90 °C
(Figure 4h) shows that the structure has disintegrated and at
97 °C (Figure 4i) there is the emergence of a rearranged
crystalline form. Wang et al.8using WAXD have shown that
the phase transformation from the mesomorphic into a crystalline
phase occurs at temperatures above 80 °C. Hsu et al.39using
DSC also showed a smectic to monoclinic transformation at
ca. 80 °C. This is in accord with our AFM measurements which
show the crystalline structure formed at a lower temperature of
34.8 °C, disintegrating between 80 and 90 °C and rearranging
itself into a different conformation. This is not unequivocal
evidence that the structure is in the smectic phasesR-phase
crystals formed at such low temperatures would most probably
reorganize on heating. However, as the change in morphology
is very significant, with a complete change in the morphology
during the rearrangement, and as the onset of homogeneous
nucleation is also inferred to correspond with the formation of
the mesomorphic phase,5it is highly probable that the primary
nucleus formed at 34.8 °C and which completed its growth at
lower temperatures is the “smectic” form of iPP. Previous
images of smectic iPP8,39,40have shown a course, granular
texture, while in our images we see a clear crystal form. It is
possible that the difference in morphology is due to the
constrained nature of the growth in the system studied here,
preventing in-filling that usually hides the original form.
The third form of nucleationssporadic nucleation in droplets
that does not then lead to substantial growth, with further
nucleation occurring later in the same dropletsis the most
unexpected. Figure 6 shows a series of images showing this
mode of growth. In order to interpret the images more easily,
white circles are used to indicate the droplets under investiga-
tion; black arrows indicate the onset of nucleation.
In Figure 6b,c the commencement of crystalline growth is
observed in two of the droplets (indicated by the black arrows).
In Figure 6e,f the appearance of crystal nuclei wholly uncon-
nected from the previously crystallized areas in the same droplets
can be seen. These crystal nuclei are formed at very high
supercooling, even higher than those which resulted in crystal-
Figure 5. Comparison between the line profiles of the iPP droplet shown in Figure 4 taken at 40, 34.8, and 33.6 °C.
Nanodroplets of Isotactic Polypropylene
lization from a single nucleus (shown in Figure 4). The images
clearly show that instead of one single nucleus initiating growth
inside the droplet, there is more than one nucleus, appearing at
different times which slowly grow and fill the entire droplet.
The growth of these crystals is at such a very slow rate that we
were able to track the formation of the various nuclei as well
as the lamellar growth. The fact that these nanodroplets support
multiple sites of initial nucleation and the extremely slow growth
rate make this type of crystal growth behavior very peculiar.
There are several possible explanations. It may be that, because
of the very thin film thickness, growth following the initial
nucleation event is severely hindered by the depletion of
crystallizable material that occurs during nucleation, with
subsequent transport over the surface being too slow at this
temperature relatively close to the glass transition temperature.
So, nucleation has an opportunity to be initiated in other sites
within the droplet, before it has been completely crystallized
by the growth that has started from one nucleus. Alternatively,
if we assume growth is in the mesomorphic phase at these high
supercoolings, it is possible that this highly defective phase
grows very slowlysalthough this seems unlikely as it supplies
no explanation for the slow growth seen in these thin films
compared to the almost instantaneous growth observed in
macroscopic samples. It may be that the polymer chains adjacent
to the surface are pinned in some way and hence have reduced
mobility, leading to slow growth. However, this seems rather
unlikely, considering the fact that this surface dewets at high
temperatures and leaves no polymer behind, implying minimal
attractive interaction between the polymer and the surface.
Another possibility is that there is an orientational effectsif
the initial nucleus is oriented such that the fast growth direction
is perpendicular to the surface, it cannot grow very rapidly (or
at all for certain possible orientations), so the geometric effect
of confinement heavily slows growth and allows time for
multiple nucleation. Indeed, as the films are so thin it seems
unlikely (though not impossible) that the chains will orient
perpendicular to the substrate in all cases, the orientation that
would be most efficient for subsequent thin film growth. From
the data obtained it is impossible to make a definitive selection
from these (or other) possibilities, although the last seems the
most likely explanation.
AFM not only allows the crystallization process to be imaged
but also provides an opportunity to measure directly the volume
of crystallizable material and to provide some test of nucleation
theory. Data were collected from droplets having thickness
ranging from 0.5 to 6 nm and volumes from 4800 to 700 000
nm3(taken from six different samples analyzed on different
days, but all made using the same sample preparation technique).
We utilized Veeco Nanoscope software in order to measure the
thickness and Image SXM41to compute the volume of each
individual droplet. On analyzing the data, we found that the
rapid growth occurred in droplets having a thickness ∼5 nm
and higher as well as volume greater than 700 000 nm3. Single
nucleus formation and growth was found in highly flattened
droplets of around 2-3 nm height and volumes less than those
which exhibited the rapid growth. Droplets with thicknesses
between 5 and 3.5 nm and volume less than 100 000 nm3
exhibited the formation of multiple nuclei inside the same
Figure 7a shows the graph of the crystallization temperature
vs the volume of the droplets while Figure 7b gives the plot of
the temperature against the thickness of the droplets. The
temperature plotted is the one at which the onset of nucleation
is detected (as in some droplets which produced more than one
nucleation site, nucleation occurred over several degrees).
Ideally, data would have been collected at a series of fixed
temperatures and the rate dependence of nucleation measured
at each temperature for different droplet populations. Unfortu-
nately, the relatively limited “field of view” of the AFM while
maintaining high resolution, as well as the difficulty of imaging
stably for such extended periods, made such an experiment
impossible, despite numerous attempts. However, data of the
form obtained still provide useful information for comparison
with theory. If nucleation is truly random in space (i.e.,
independent of interfaces), the rate (i.e., number of nuclei per
unit volume per unit time) would vary very strongly with volume
Figure 6. A series of AFM phase images showing the slow crystallization in iPP droplets. Black to white represents a change in phase of 30°, and
the scale bar represents 200 nm. (a) Taken at 35.8, (b) 33, (c) 32, (d) 31.5, (e) 30, and (f) 29.5 °C. The white circles indicate the droplets under
investigation, and the black arrows indicate when a new nucleus initiates growth.
Kailas et al.
and little with thickness. Alternatively, a stronger variation with
thickness implies either a surface effect or some other confine-
From the graphs, it is clear that the crystallization temperature
increased with the thickness as well as the volume of the
droplets. Previous experiments have also shown that larger
droplets nucleate faster than smaller ones, owing to a higher
probability of nucleation in a bigger volume.31From Figure 7a
there is a notable influence of volume on the nucleation
temperature for droplets with larger volumes. For smaller
volumes, the curve is essentially vertical, indicating that the
nucleation temperature is independent of the volume in this
range. From Figure 7b, it is seen that the nucleation temperature
vs thickness curve is linear for the entire data set, although very
scattered. So it can be concluded that for the smallest thicknesses
(<5 nm) the thickness of the droplets becomes the dominating
factor which influences the nucleation temperature, with es-
sentially no effect of volume. At larger thicknesses, volume also
starts to have an influence. To define which is the dominant
factor in the thicker film case would require considerably more
data, going to larger volumes and thicknesses (and hence
temperatures). However, we can define this thickness as the
point of a transition in behavior, in that volume is now having
an effectswhether thickness also continues to have a strong
influence at greater thicknesses or whether its apparent continu-
ing effect is due to the coupling between thickness and volume
in the population of (pancake-shaped) droplets measured is
unclear. In ref 32, the authors, on investigating small volumes,
found that the nucleation still had a dependence on the droplet
volume in systems down to ∼10 polymer chains and was not
influenced by the interface. In that work the thinnest droplets
studied were greater than 5 nm in thickness. From the examina-
tion of our data (Figure 7) it appears that the “transition” in
behavior occurs at a thickness slightly smaller than that
encountered in ref 32. It is possible that this change occurs
because the dimension of the critical nucleus is now comparable
to the film thickness. The R-phase unit cell parameters have
been found to be a ) 6.65 ( 0.05 Å, b ) 20.96 ( 0.15 Å, c
) 6.50 ( 0.05 Å, R ) γ ) 90°, and ? ) 99.6°.42,43The chain
axis repeat is 6.5 Å. In these very thin films the number of unit
cell repeat distances available within the initial film thickness
is very limited. Once the critical nucleus size in at least one
dimension is comparable to the film thickness the rate of
nucleation is likely to drop, as nucleation requires a fluctuation
which includes an increase in height of the droplet. We therefore
suggest that this is a thickness, rather than a surface, effect.
We have imaged in-situ, in real space, for the first time, the
nucleation and growth of iPP crystals at a supercooling of more
than 125 °C using AFM. Homogeneous nucleation has been
unambiguously observed. By following the subsequent reorga-
nization of the structure on heating, we suggest that the crystals
formed are in the mesomorphic form. Direct, real-space
observation of crystallization in the mesomorphic form shows
the development of a clear crystal morphology. Real-time
imaging of these crystals during heating shows a complete
rearrangement of the structure as expected, given the consider-
able amount of chain reorganization that is necessary between
the statistical up-down chain alignment in the mesomorphic
and the antiparallel chain packing in the R-phases.
The thicknesses of the droplets as well as their volume have
been found to have an influence on the temperature at which
nucleation is initiated as well as the mode of crystal growth.
For smaller thicknesses (<5 nm), we have found that the
nucleation temperature depends on the thickness of the droplets.
The formation of isolated droplets with nanometric dimen-
sions provides a powerful tool for in-situ observation of
nucleation processes and of crystallization at very high super-
coolings in materials that, in bulk samples, cannot be deeply
quenched. Combination with temperature controlled AFM
allows these nonequilibrium processes to be followed in-situ,
in real time, with nanometer resolution.
Acknowledgment. L. Kailas, C. Vasilev, and J. K. Hobbs
thank the EPSRC, UK, for funding. We also thank Mr. O.
Farrance (Dept. of Physics & Astronomy, University of Shef-
field) for the development of the low-temperature system and
temperature calibration. The NanoSIMS analyses have been
performed with the financial support from Fond National de la
Recherche of Luxembourg.
References and Notes
(1) Strobl, G. Eur. Phys. J. E 2000, 3, 165.
Figure 7. (a) Plot of the crystallization temperature vs the volume of
the droplets and (b) plot of the crystallization temperature vs the
thickness of the droplets obtained from six samples (shown in the graph
using six different symbols).
Nanodroplets of Isotactic Polypropylene
(2) Olmsted, P. D.; Poon, W. C. K.; McLeish, T. C. B.; Terrill, N. J.;
Ryan, A. J. Phys. ReV. Lett. 1998, 81, 373.
(3) Heeley, E. L.; Maidens, A. V.; Olmsted, P. D.; Bras, W.; Dolbnya, I.
P.; Fairclough, J. P. A.; Terrill, N. J.; Ryan, A. J. Macromolecules
2003, 36, 3656.
(4) Keller, A.; Hikosaka, M.; Rastogi, S.; Toda, A.; Barham, P. J.;
Goldbeck-wood, G. J. Mater. Sci. 1994, 29, 2579.
(5) Lotz, B. Acad. Nazionale Lincei 2004, 107.
(6) Lotz, B.; Wittmann, J. C.; Lovinger, A. J. Polymer 1996, 37, 4979.
(7) Meille, S. V.; Bruckner, S.; Porzio, W. Macromolecules 1990, 23,
(8) Wang, Z.-G.; Hsiao, B. S.; Srinivas, S.; Brown, G. M.; Tsou, A. H.;
Chang, S. Z. D.; Stein, R. S. Polymer 2001, 42, 7561.
(9) Nedkov, E.; Dobreva, T. e-Polym. 2002, 042, 1.
(10) Merajver, S. D.; Wunder, S. L.; Wallace, W. J. Polym. Sci., Polym.
Phys. 1985, 23, 2043.
(11) Tordjeman, Ph.; Robert, C.; Marin, G.; Gerard, P. Eur. Phys. J. E
2001, 4, 459.
(12) Fillon, B.; Lotz, B.; Thierry, A.; Wittmann, J. C. J. Polym. Sci., Part
B: Polym. Phys. 1993, 31, 1395.
(13) Stocker, W.; Schumacher, M.; Graff, S.; Thierry, A.; Wittmann, J.
C.; Lotz, B. Macromolecules 1998, 31, 807.
(14) Mezghani, K.; Phillips, P. J. Polymer 1997, 38, 5725.
(15) Mezghani, K.; Phillips, P. J. Polymer 1998, 39, 3735.
(16) Dimeska, A.; Phillips, P. J. Polymer 2006, 47, 5445.
(17) Massa, M. V.; Dalnoki-Veress, K.; Forrest, J. A. Eur. Phys. J. E 2003,
(18) Hobbs, J. K.; Humphris, A. D. L.; Miles, M. J. Macromolecules 2001,
(19) Hobbs, J. K. Chin. J. Polym. Sci. 2003, 21, 135.
(20) Lei, Y.-G.; Chan, C.-M.; Li, J.-X.; Ng, K.-M.; Wang, Y.; Jiang, Y.;
Li, L. Macromolecules 2002, 35, 6751.
(21) Pearce, R.; Vansco, G. J. Macromolecules 1997, 30, 5843.
(22) Li, L.; Chan, C.-M.; Li, J.-X.; Ng, K.-M.; Yeung, K.-L.; Weng, L.-T.
Macromolecules 1999, 32, 8240.
(23) Li, L.; Chan, C.-M.; Yeung, K.-L.; Li, J.-X.; Ng, K.-M.; Lei, Y.
Macromolecules 2001, 34, 316.
(24) Beekmans, L. G. M.; Vancso, G. J. Polymer 2000, 41, 8975.
(25) Chen, E.-Q.; Jing, A. J.; Weng, X.; Huang, P.; Lee, S.-W.; Cheng, S.
Z. D.; Hsiao, B. S.; Yeh, F. Polymer 2003, 44, 6051.
(26) Lei, Y.-G.; Chan, C.-M.; Wang, Y.; Ng, K.-M.; Jiang, Y.; Lin, L.
Polymer 2003, 44, 4673.
(27) Hobbs, J. K.; McMaster, T. J.; Miles, M. J.; Barham, P. J. Polymer
1998, 39, 2437.
(28) Vonnegut, B. J. Colloid Sci. 1948, 3, 563.
(29) Taden, A.; Landfester, K. Macromolecules 2003, 36, 4037.
(30) Tol, R. T.; Mathot, V. B. F.; Groeninckx, G. Polymer 2005, 46, 2955.
(31) Massa, M. V.; Dalnoki-Veress, K.; Phys. ReV. Lett. 2004, 92, 255509.
(32) Massa, M. V.; Carvalho, J. L.; Dalnoki-Veress, K. Phys. ReV. Lett.
2006, 97, 247802.
(33) Strobl, G. The Physics of Polymers; 2nd ed.; Springer-Verlag: Berlin,
(34) Hobbs, J. K.; Humphris, A. D. L.; Miles, M. J. ACS Symp. Ser. 2005,
(35) Edwards, B. C.; Phillips, P. J. J. Polym. Sci., Polym. Phys. Ed. 1975,
(36) Gradys, A.; Sajkiewicz, P.; Minakov, A. A.; Adamovsky, S.; Schick,
C.; Hashimoto, T.; Saijo, K. Mater. Sci. Eng., A 2005, 413-414, 442.
(37) Jin, Y.; Hiltner, A.; Baer, E.; Masirek, R.; Piorkowska, E.; Galeski,
A. J. Polym. Sci., Part B: Polym. Phys. 2006, 44, 1795.
(38) Uriguen, J. I.; Bremer, L.; Mathot, V.; Groeninckx, G. Polymer 2004,
(39) Hsu, C. C.; Geil, P. H.; Miyaji, H.; Asai, K. J. Polym. Sci., Part B:
Polym. Phys. 1986, 24, 2379.
(40) Gezovich, D. M.; Geil, P. H. Polym. Eng. Sci. 1968, 8, 202.
(41) Barrett, S. D. Image SXM 2007, http://www.ImageSXM.org.uk.
(42) Natta, G.; Corradini, P. NuoVo Cimento Suppl. 1960, 15, 40.
(43) Bruckner, S.; Meille, S. V.; Petraccone, V.; Pirozzi, B. Prog. Polym.
Sci. 1991, 16, 361.
Kailas et al.PAGE EST: 7.4