Effects of nanoscale confinement and interfaces on the glass transition temperatures of a series of poly(n-methacrylate) films

Article abstract of DOI:10.1071/CH07234

We use fluorescence from dye-labelled polymer to measure the glass transition temperatures (T_gs) across single-layer films and near surfaces and silica interfaces in bilayer films for a series of poly(n-methacrylate)s. With nanoscale confinemen

Full text of DOI:10.1071/CH07234

Full Paper
CSIRO PUBLISHING
www.publish.csiro.au/journals/ajc
Aust. J. Chem. 2007, 60, 765–771
Effects of Nanoscale Confinement and Interfaces
on the Glass Transition Temperatures of a Series
of Poly(n-methacrylate) Films
Rodney D. Priestley,A Manish K. Mundra,B Nina J. Barnett,B Linda J. Broadbelt,A
and John M. TorkelsonA,B,C
ADepartment of Chemical and Biological Engineering, Northwestern University, Evanston,
Illinois 60208, USA.
BDepartment of Materials Science and Engineering, Northwestern University, Evanston,
Illinois 60208, USA.
CCorresponding author. Email: j-torkelson@northwestern.edu
We use fluorescence from dye-labelled polymer to measure the glass transition temperatures (Tgs) across single-layer films
and near surfaces and silica interfaces in bilayer films for a series of poly(n-methacrylate)s. With nanoscale confinement,
the averageTg across a film supported on silica increases for poly(methyl methacrylate) (PMMA), decreases for poly(ethyl
methacrylate) (PEMA) and poly(propyl methacrylate), and is nearly invariant for poly(iso-butyl methacrylate) (PIBMA).
These trends are consistent with the relative strengths of local perturbations to Tg caused by surfaces and substrates as
measured in bilayer films.The substrate effect, which increasesTg via hydrogen-bonding interactions between the polymer
and hydroxyl groups on the silica surface, is stronger than the free-surface effect in PMMA. The free-surface effect, which
reducesTg via a reduction in the required cooperativity of the glass transition dynamics, is stronger than the substrate effect
in PEMA. The substrate and free-surface effects have similar strengths in perturbing the local Tg in PIBMA, resulting in
a net cancellation of effects when measurements are made across single-layer films.
Manuscript received: 9 July 2007.
Final version: 24 August 2007.
Introduction
This behaviour has implications for applications ranging from
photoresists to asymmetric membranes.
As the confining dimension is decreased below a certain
critical value in thin polymer films (with thickness as the
confining dimension) and polymer nanocomposites (with inter-
nanoparticle distance as the confining dimension), an increas-
ingly large fraction of polymeric material is directly in contact
with interfaces or surfaces. Properties can be strongly per-
turbed by the interfacial interactions and surfaces in such
confined polymers, and these perturbations to average prop-
erties or the distributions of properties within such systems
are commonly referred to as the confinement effect. The study
of the effect of nanoscale confinement on the glass transition
temperature (Tg) of amorphous polymers began more than a
dozen years ago.[1,2] In 1994, Keddie et al. showed that ultra-
thin polystyrene (PS) films exhibit a reduction in Tg with
decreasing thickness.[1] In the same year, they also found that
ultrathin poly(methyl methacrylate) (PMMA) films exhibit a
Tg-confinement effect that depends on the substrate supporting
the films, with Tg decreasing with confinement when PMMA
is supported on gold and increasing with confinement when
PMMA is supported on silica.[2] Since then, many studies have
demonstrated that nanoscale confinement of amorphous poly-
mers can lead to changes in Tg and related dynamics from their
bulk values, whether in supported polymer films,[1–30] freely
standing polymer films,[31–34] or polymer nanocomposites.[19,20]
In systems exhibiting
a
decrease in Tg with
confinement,[1,2,4–14,18,21,22,24–28] it is now reasonably well
accepted that the effect originates at the free surface or
polymer–air interface.[4,11,15] Taking advantage of the fact
that fluorescence has proved to be very useful in studies of
dynamics and thermodynamics associated with polymer-based
materials,[35–46] our group developed a novel multilayer fluores-
cence technique that demonstrated thatTg reductions originating
at the free surface can propagate some tens of nanometres into
the film interior.[11,15,16,17,24] This behaviour can be understood
qualitatively by recognizing that the glass transition in poly-
mers is associated with cooperative segmental mobility of tens
to hundreds of repeat units and that the presence of the free sur-
face locally reduces the requirement for cooperativity, thereby
reducing Tg.[11] In contrast, the presence of attractive polymer–
substrate or polymer–nanoparticle interactions, e.g. hydrogen
bonds, increases the requirement for cooperativity in the dynam-
ics associated with the glass transition.[3,5,12,13,15,17–20] This
leads to an increase in Tg with confinement[2,3,5,8,12,13,15,17–20]
when the effects of attractive interactions are strong and domi-
nate the free-surface effects. Although most research in this area
has been experimental, simulations[47–50] have reached similar
conclusions regarding the roles of free surfaces and attractive
© CSIRO 2007
10.1071/CH07234
0004-9425/07/100765

766
R. D. Priestley et al.
Table 1. Molecular characterization and bulk Tgs of polymers used in the present study
DSC, differential scanning calorimetry; PMMA, poly(methyl methacrylate); PEMA, poly(ethyl methacrylate); PPMA,
poly(propyl methacrylate); PIBMA, poly(iso-butyl methacrylate)
Material
Mn [g mol−1
]
Mw/Mn
Bulk Tg [K]
(onset, DSC)
Bulk Tg [K]
(fluorescence)
Label content
[mol-%]
PMMA
PEMA
PIBMA
TC1-labelled PMMA
Pyrene-labelled PEMA
Pyrene-labelled PPMA
TC1-labelled PIBMA
355 000A
460 000B
300 000B
509 000A
202 000B
188 000B
181 000B
1.54A
1.70B
2.00B
1.67A
1.80B
1.78B
1.96B
394
348
337
394
348
320
337
–
–
–
395
347
320
339
–
–
–
1.4
0.6
0.3
0.6
ADetermined via gel permeation chromatography using the universal calibration method and appropriate Mark–Houwink
parameters.
BDetermined via gel permeation chromatography relative to polystyrene standards using tetrahydrofuran as eluent.
polymer–substrate interactions in defining the Tg-confinement
effect. Recent reviews summarize the results, progress, and
challenges in this field.[51–53]
synthesized by reaction with tetracyanoethylene (TCI Amer-
ica) and 2-(N-ethylaniline)ethanol (TCI America) dissolved in
dimethyl formamide (Fisher) at 55◦C for 15 min and then recrys-
tallized from glacial acetic acid. The 1-pyrenyl butanol-labelled
(Aldrich) and TC1-labelled methacrylate monomers were syn-
thesized through an esterification of 1-pyrenyl butanol or TC1,
respectively, with methacryloyl chloride (Aldrich) in the pres-
ence of triethylamine (Aldrich) and dichloromethane (Aldrich)
at 0◦C for 2 h.[55]
Dye-labelled versions of PMMA, PEMA, PPMA, and
PIBMA were synthesized by bulk free radical polymerization
of methyl methacrylate (Aldrich), ethyl methacrylate (Aldrich),
propyl methacrylate (Aldrich), and iso-butyl methacrylate
(Aldrich), respectively, in the presence of trace amounts of
pyrene-labelled or TC1-labelled methacrylate monomer. All
dye-labelled polymers were washed by dissolving in toluene
(Fisher) and precipitating in methanol (Fisher) at least five times
to remove residual unreacted dye-labelled monomer, and then
dried in a vacuum oven at Tg + 15 K for 24 h. Table 1 lists the
number-average molecular weight (Mn) values, polydispersity
indexes, bulk Tgs, and label content of the poly(n-methacrylate)
samples, and Fig. 1 shows the repeat unit structures.
Recently, our group[10] investigated the impact of slight mod-
ifications to the repeat unit structure of PS on theTg-confinement
effect observed in films supported on silica. In addition to PS,
we studied two other systems, including poly(4-methylstyrene)
(P4MS) and poly(tert-butylstyrene) (PTBS), none of which
exhibits any significant attractive interaction with hydroxyl
groups on the surface of the silica substrate; this means that
all Tg-confinement effects originate at the free surfaces of PS,
P4MS, and PTBS films. We observed that the Tg reductions in
the ultrathin films increased dramatically with the presence of
large, rigid side groups on the phenyl ring. In the case of 25-
nm-thick films supported on silica, we obtained the following
reductions in Tg relative to bulk Tg: ∼12 K for PS, ∼18 K for
P4MS, and ∼47 K for PTBS.[10] In addition, the onset thick-
ness for theTg reduction in PTBS was 300–400 nm, much larger
than in PS or P4MS, where the onset thickness for the Tg reduc-
tions is ∼60 nm.[10] These results indicate that small changes to
the repeat unit structure of the polymer, presumably leading to
changes in the packing efficiency of the polymer repeat units
and thereby the effective chain stiffness, can lead to substantial
tunability in the Tg-confinement effect.
Differential Scanning Calorimetry
The bulk Tg of each polymer was determined using a Mettler-
Toledo 822 differential scanning calorimeter. Sample masses of
5–10 mg were placed in sealed aluminium pans with a pinhole
in the top to allow measurements to be conducted in a nitrogen
environment.TheTgs were determined on heating at 10 K min−1
afterannealingthesamplesat423 Ktoerasethepreviousthermal
history. The reported Tgs are onset values.
Here, we continue this line of investigation by studying how
the Tg-confinement effect is impacted by slight variations to
the side groups of poly(n-methacrylate) films supported on
silica, which exhibit both free-surface effects and attractive
polymer–substrate interactions. We use a fluorescence method
to determine Tg values in single-layer films and in bilayer films
in which only one layer has polymer labelled with a fluores-
cencedye.Thelatterexperimentsallowustomeasuredirectlythe
localTgs in ultrathin free-surface layers and substrate layers.The
polymers used in the present study include PMMA, poly(ethyl
methacrylate) (PEMA), poly(propyl methacrylate) (PPMA), and
poly(iso-butyl methacrylate) (PIBMA).
Gel Permeation Chromatography
The Mn value and polydispersity index of each polymer were
determined using a Waters Breeze gel permeation chromato-
graph equipped with refractive index and fluorescence detec-
tors. Molecular weight values were determined relative to PS
standards or universal calibration using the appropriate Mark–
Houwink constants. Tetrahydrofuran was used as the eluent at
30◦C.
Experimental
Three unlabelled polymers, PMMA, PEMA, and PIBMA, were
synthesized at 70◦C by bulk free radical polymerization using
benzoyl peroxide (Aldrich) as initiator. Following a procedure
originally described by McCusick and coworkers,[54] the dye
4-tricyanovinyl-(N-(2-hydroxyethyl)-N-ethyl)aniline (TC1) was
Determination of Label Content
The label content of each dye-labelled polymer was deter-
mined using a Perkin–Elmer Lambda 35 UV-Vis absorbance

Effects of Nanoscale Confinement and Interfaces
767
CH3
C
CH3
C
CH3
C
H2
C
H2
C
H2
C
Poly(x)
Poly(x-r-x؅)
C
C
C
O
R
O
O
O
O
O
R
Rꢀ
2؅
R ؍
1
2
3
4
R؅ ؍
1؅
H2
C
N
CH3
CH3 CH2 CH2
CH3 CH2
CH3
H3C
CH
CH3
NC
CN
CN
Labelled polymers:
Poly(1-r-1ꢀ) ꢁ TC1-labelled PMMA
Poly(2-r-2ꢀ) ꢁ Pyrene-labelled PEMA
Poly(3-r-2ꢀ) ꢁ Pyrene-labelled PPMA
Poly(4-r-1ꢀ) ꢁ TC1-labelled PIBMA
Neat polymers:
Poly(1) ꢁ PMMA
Poly(2) ꢁ PEMA
Poly(4) ꢁ PIBMA
Fig. 1. Structures of polymers used in the present study.
spectrophotometer. Dye-labelled polymer was dissolved in
chloroform at a known concentration, and the absorbance spec-
trum of the solution was measured. Using wavelengths at which
only the dye contributes to the absorbance spectrum, the con-
centration of the dye in the solution, and thereby the label
content in the dye-labelled polymer, can be determined using
the Beer–Lambert law with measured extinction coefficients for
the TC1-labelled monomer and pyrene-labelled monomer. (The
technique assumes that the dye attached to the polymer has the
same extinction coefficient at its absorbance maximum as the
dye-labelled monomer even though the absorbance spectrum
of the dye attached to the polymer is shifted by a few nano-
metres compared with the dye-labelled monomer.The extinction
coefficient for each dye-labelled monomer was determined from
absorbance measurements taken from solutions containing vary-
ing amounts of dye-labelled monomer.The extinction coefficient
is the slope of a plot of absorbance versus concentration).
The label contents were as follows: 1.4 and 0.6 mol-% TC1-
labelled monomer in PMMA and PIBMA, respectively, and 0.6
and 0.3 mol-% pyrene-labelled monomer in PEMA and PPMA,
respectively.
annealed at Tg + 25 K for 10 min to ensure completely consoli-
dated films. Film thicknesses were measured with a Tencor P10
profilmeter using an average of at least ten measurements.
All polymers were of sufficiently high molecular weight to
ensure that interlayer diffusion occurred over at most several
nanometres during the experimental measurement time, which
includes the time to create a consolidated film. Estimation of
interlayer diffusion of similar bilayer films has been described
elsewhere.[15,24]
Fluorescence Measurements
Steady-state fluorescence emission spectra were measured as
a function of temperature (on cooling) using a Photon Tech-
nology International fluorimeter or an SPEX Fluorog-2 DM1B
fluorimeter with 5-nm bandpass excitation and emission slits
used for pyrene-labelled polymer and 12-nm bandpass excita-
tion and emission slits used forTC1-labelled polymer. Excitation
wavelengths of 254 and 480 nm were used for pyrene-labelled
and TC1-labelled polymer, respectively, where there were max-
ima in absorbance. The emission spectra of pyrene-labelled and
TC1-labelled polymer were measured at 360–460 nm and 540–
690 nm, respectively.The Tgs of films were determined by fitting
the temperature dependence of the integrated fluorescence inten-
sity to linear correlations in both the rubbery and glassy states.
In fitting the data to linear correlations, only data well away from
Tg were used in the fitting procedure.
In all cases, Tg data were obtained from fluorescence mea-
surement obtained on cooling from the equilibrium rubbery or
liquid state. Polymer films were heated to the maximum mea-
surement temperature and held for a minimum of 10 min before
measuring the fluorescence emission spectrum. Then the tem-
perature was decreased by 5 K where the sample was held for
5 min to ensure thermal equilibrium before again measuring the
fluorescenceemissionspectrum.This‘cooling, holding, measur-
ing’protocol was repeated well into the glassy state of each film
Film Preparation
Single-layer films were prepared by spin casting[56] polymer/
toluene solutions onto silica slides using solution concentra-
tions ranging from 0.5 to 4.0 wt-% and spin speeds ranging from
1000 to 4000 rpm. The films were allowed to dry in vacuum at
Tg + 5 K for 8 h. Bilayer films were prepared by spin coating
polymer/toluene solutions onto either a silica slide or an NaCl
salt disc. Films spun-cast onto NaCl salt disks were placed in
water, allowing the salt to dissolve, and leaving the polymer film
floating on the water surface. Films spun-cast on silica were then
submersed in the water and used to pick up the films floating on
the water surface, resulting in bilayer films. These bilayer films
were dried in vacuum at room temperature for 12 h and then

768
R. D. Priestley et al.
can yield a precise and accurate determination of polymerTg.[11]
(Here we employ both pyrene-labelled and TC1-labelled poly-
mers because we already had in our possession several of the
polymers, some labelled with pyrene and others labelled with
TC1, needed for the current study). The intersection of linear
correlations fitted to data points deep in the rubbery state and
glassy state provides a measure of Tg. There is an interesting
difference in fluorescence of pyrene-labelled and TC1-labelled
polymers, with pyrene fluorescence yielding a greater tem-
perature dependence in the rubbery state above Tg and TC1
fluorescence yielding a greater temperature dependence in the
glassy state below Tg. This difference is related to the different
mechanisms by which each dye undergoes non-radiative decay
from its excited state and is explained in more detail in refs
[11], [13], and [19]. Nevertheless, both dyes report Tgs for bulk
polymers that are in good agreement with Tg values obtained by
classic techniques such as DSC.
Fig. 3a illustrates that nanoscale confinement leads to a
decrease in the Tg of PEMA. The Tg values of the 500- and
20-nm-thick pyrene-labelled PEMA films are 347 and 336 K,
respectively. In contrast, Fig. 3b shows that nanoscale confine-
ment leads to no change within experimental error ( 1 K) in the
Tg of PIBMA. The Tg values of the 500- and 20-nm-thick TC1-
labelled PIBMA films are 339 and 341 K, respectively. Related
measurements of Tg were also obtained in thin and ultrathin
PMMA and PPMA films.
Fig. 4 shows the thickness dependence of the deviation of Tg
from bulk Tg for single-layer films of PMMA, PIBMA, PEMA,
and PPMA supported on silica. This figure reveals that slight
modifications to the repeat unit structure of a series of poly(n-
methacrylate)s can have a profound impact on how confinement
to thicknesses less than 100 nm affects Tg. (Little change, in the
case of PPMA, or no change, in the case of PMMA, PIBMA, and
PEMA, inTg is observed as a function of thickness for films with
thickness exceeding 100 nm. This is because the perturbations
toTg response caused by the polymer–air and polymer–substrate
interfaces are too small to modify the averageTg response across
films that are thicker than 100 nm). For TC1-labelled PMMA,
the increase in Tg with confinement is consistent with previous
studies of single-layer PMMA films supported on silica.[2,5,19]
The increase inTg with confinement has been related to attractive
hydrogen-bonding interactions between the ester side groups of
PMMA and the hydroxyl groups on silica that reduce segmental
mobility at the substrate interface.[2,5,15] Changing the side unit
off the ester group of PMMA from a methyl unit to an ethyl
(PEMA) or propyl (PPMA) unit leads to an opposite effect of
confinement on Tg, with Tg decreasing from its bulk value by
11 K in a 20-nm-thick PEMA film and by 17 K in a 30-nm-thick
PPMA film. Changing the methyl unit off the ester group of
PMMA to an iso-butyl unit (PIBMA) leads to aTg that is, within
error, independent of confinement down to a film thickness of
14 nm.
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
360
380
400
420
440
460
Wavelength [nm]
Fig. 2. Temperature dependence of the fluorescence emission spectrum
of a 500-nm-thick pyrene-labelled poly(ethyl methacrylate) (PEMA) film
supported on silica: 383 K (solid curve), 348 K (dashed curve), and 308 K
(dotted curve). Data have been normalized to one at the maximum peak value
at 383 K.
being measured and has previously been shown to yield Tg mea-
surements in bulk films that are in good agreement with onset
Tgs measured by differential scanning calorimetry (DSC) using
a 10 K min−1 heating rate.[10,22] (The good agreement is likely
due to the temperature ramping by DSC being within an order
of magnitude of the effective average cooling rate of 5 K/5 min
(or 1 K min−1) used in our fluorescence measurements of Tg).
Additional information on the use of fluorescence to monitor Tg
is found in refs [10–13] and [15].
Results and Discussion
Fig. 2 shows the fluorescence emission spectrum of a pyrene-
labelled PEMA film supported on silica as a function of tem-
perature, which consists solely of fluorescence from isolated
excited-state pyrenyl units, which is called monomer fluor-
escence. (When two pyrenyl dyes are separated from each other
by a severalAngstrom distance, it is possible to observe excimer
fluorescence or fluorescence from an excited-state dimer, which
is a broad, structureless emission centred with a maximum inten-
sity at 480–490 nm.[57] Such fluorescence is absent in the film
measurements done in the current study, indicating that, within
error, all pyrenyl dye fluorescence is from single pyrene dyes iso-
lated in the polymer matrix). A decrease in temperature yields
an increase in fluorescence intensity. The increase in inten-
sity results because a reduction in temperature leads to reduced
thermal energy and densification of the nanoscale medium sur-
rounding the dye, both of which reduce the rate of non-radiative
decay from the excited state of the pyrene dye. We also observed
increases in intensity with decreasing temperature in the other
pyrene-labelled and TC1-labelled polymers used in the present
study.
Fig. 3 shows the temperature dependence of the normalized
integrated fluorescence intensity of thin and ultrathin, single-
layer pyrene-labelled PEMA and TC1-labelled PIBMA films.
We have previously shown that such data originating from appro-
priately chosen dyes can be used to determine Tg in bulk and
nanoconfined polymers[10–13,15–22] and have provided a rationale
for why the temperature dependence of dye fluorescence, which
is sensitive to the temperature dependence of polymer density,
Why do these small modifications to the repeat unit struc-
ture have such an enormous impact on how confinement affects
Tg in this series of poly(n-methacrylates)? From our previous
studies on the distributions of Tgs and physical aging rates in
PS, PMMA, and other films, we know that the free surface and
substrate interface affect Tg and glassy-state dynamics differ-
ently in different polymers.[11,15–17,58] Here we use a bilayer
fluorescence method to investigate directly the local Tgs in the
free-surface and substrate-interface layers of the series of poly(n-
methacrylate)s and show that the modifications to the repeat

Effects of Nanoscale Confinement and Interfaces
769
(a)
(b)
1.3
1.5
1.4
1.3
1.2
1.1
1.0
0.9
0.8
0.7
N
CH
3
HO
NC
1.2
1.1
1.0
0.9
0.8
CN
CN
290
310
330
350
370
390
310 320 330 340 350 360 370
Temperature [K]
Temperature [K]
Fig. 3. (a) Temperature dependence of integrated fluorescence intensity of 500-nm-thick (ꢀ) and 20-nm-thick (ꢀ) pyrene-
labelled poly(ethyl methacrylate) (PEMA) films. Data have been normalized to Tg and arbitrarily shifted vertically for clarity.
Inset shows the structure of pyrene. (b) Temperature dependence of integrated fluorescence intensity of 500-nm-thick (ꢀ) and
20-nm-thick (ꢀ) TC1-labelled poly(iso-butyl methacrylate) (PIBMA) films. Data have been normalized to Tg and arbitrarily
shifted vertically for clarity. Inset shows the structure of TC1.
12
6
for polymers that can undergo hydrogen bonding with hydroxyl
groups naturally on the surface of the silica substrates. However,
the magnitude of the deviations in Tg at the interfaces depends
strongly on chemical structure. For PMMA, the 12-nm-thick
substrate layer has a Tg enhanced by 10 K compared with the
bulk Tg, whereas for PEMA and PIBMA, the 14-nm-thick sub-
strate layers have Tgs enhanced by only 5 K compared with the
bulk Tgs. The 14-nm-thick free-surface layer of PEMA has a
Tg reduced by 14 K compared with its bulk Tg. In contrast, the
Tg reductions are only 7 and 6 K in 12-nm-thick PMMA and
14-nm-thick PIBMA free-surface layers, respectively.
A comparison of Fig. 4 and Fig. 5 reveals a correlation
between the observed deviations in Tg with confinement for
the single-layer films (Fig. 4) and the relative strength of the
deviations in Tg in the free-surface and substrate-interface lay-
ers (Fig. 5) of the poly(n-methacrylate) films. When substrate
effects are stronger than free-surface effects (as quantified by a
greaterTg deviation in the substrate layer than in the free-surface
layer), the Tgs of single-layer films increase with confinement
as observed with PMMA films. When free-surface effects are
stronger than substrate effects, as in the case for PEMA, the
Tgs of single-layer films decrease with confinement. When free-
surface and substrate effects are of nearly equal strength, as in
the case of PIBMA, the Tgs of single-layer films are invariant
with confinement.
Regarding the effects of the substrate and free surface on
the Tgs of the poly(n-methacrylate) layers shown in Fig. 5, the
increase in Tg in the layer next to the silica substrate is under-
stood qualitatively to arise from hydrogen-bonding interactions
between the ester groups of the poly(n-methacrylate)s and the
hydroxyl groups on the silica surface, whereas the reduction inTg
in the free-surface layer is related to how the free surface reduces
the requirement for cooperative segmental mobility. Poly(methyl
methacrylate), which has the smallest alkyl side group of the
seriesofpoly(n-methacrylate)sstudied, exhibitsthegreatestsub-
strate effect. This is likely because the small methyl units off the
ester side groups in PMMA are relatively ineffective relative to
thelargerethylandiso-butylunitsofPEMAandPIBMA, respec-
tively, in impeding hydrogen bonding between the oxygen atoms
on the ester side groups and hydroxyl units on the surface of the
silica. The detailed cause of the substantially larger effect in the
PMMA
PIBMA
0
ꢂ6
ꢂ12
ꢂ18
PEMA
PPMA
10
100
Film thickness [nm]
1000
Fig. 4. Deviation of Tg from bulk Tg as a function of film thickness
for TC1-labelled poly(methyl methacrylate) (PMMA) (ꢁ), TC1-labelled
poly(iso-butyl methacrylate) (PIBMA) (ꢀ), pyrene-labelled poly(ethyl
methacrylate) (PEMA) (ꢀ), and pyrene-labelled poly(propyl methacrylate)
(PPMA) (♦) films.
unit structure affect how the free surface and substrate interface
affect the local Tgs and thereby the average Tgs measured across
ultrathin single-layer films.
Our bilayer films consist of a dye-labelled poly(n-
methacrylate) layer and an unlabelled poly(n-methacrylate)
layer. We construct bilayer films in such a manner that an ultra-
thin labelled layer can be placed at either the free surface or
substrate interface. Heating the bilayer films for a short time
above Tg produces a consolidated film with only the labelled
layer contributing to the fluorescence signal.
Fig. 5 shows the results of the bilayer film measurements for
PMMA, PEMA, and PIBMA where the films are sufficiently
thick to yield bulk Tg values if measured as single-layer films.
In all cases, there is a reduction in Tg at the 12- or 14-nm-
thick free-surface layers, qualitatively consistent with previous
studies[1,4,6,10,11,13,19,21,22,25,31,33] of how the presence of a free
surfaceperturbsthelocalTg. Inallcases, thereisanincreaseinTg
at the polymer–substrate interface, consistent with expectations

770
R. D. Priestley et al.
(a)
(b)
(c)
PMMA
PEMA
PIBMA
Tg(film) ꢂ Tg(bulk) ꢁ ꢂ7 [K]
Bulk underlayer
Tg(film) ꢂ Tg(bulk) ꢁ ꢂ14 [K]
Bulk underlayer
Tg(film) ꢂ Tg(bulk) ꢁ ꢂ6 [K]
Bulk underlayer
Substrate
Substrate
Substrate
Bulk overlayer
Bulk overlayer
Bulk overlayer
Tg(film) ꢂ Tg(bulk) ꢁ 10 [K]
Tg(film) ꢂ Tg(bulk) ꢁ 5 [K]
Tg(film) ꢂ Tg(bulk) ꢁ 5 [K]
Substrate
Substrate
Substrate
Fig. 5. Deviation of Tg from bulk Tg in ultrathin free-surface and substrate layers in bilayer films. (a) 12-nm-thick TC1-
labelled poly(methyl methacrylate) (PMMA) layer placed at the free surface and substrate interface of PMMA bilayer films.
Film thicknesses of bulk overlayer and bulk underlayer are 240 nm. (b) 14-nm-thick pyrene-labelled poly(ethyl methacrylate)
(PEMA) layer placed at the free surface and substrate interface of PEMA bilayer films. Film thicknesses of bulk overlayer and
bulk underlayer are 500 nm. (c) 14-nm-thick TC1-labelled poly(iso-butyl methacrylate) (PIBMA) layer placed at the free surface
and substrate interface of PIBMA bilayer films. Film thicknesses of bulk overlayer and bulk underlayer are 500 nm.
free-surface layer of PEMA relative to the effects observed in
PMMA and PIBMA is not yet known. However, it likely relates
to how the chemical structure of PEMA better supports a strong
free-surface effect, with its perturbation of Tg dynamics via a
reduction of the requirement for cooperativity at the polymer–
air interface. Further experimental, theoretical, and simulation
studies are warranted to provide an understanding regarding why
the perturbations to Tg at free surfaces are strongly dependent
on chemical structure.
References
[1] J. L. Keddie, R. A. L. Jones, R. A. Cory, Europhys. Lett. 1994, 27, 59.
doi:10.1209/0295-5075/27/1/011
[2] J. L. Keddie, R. A. L. Jones, R. A. Cory, Faraday Discuss. 1994, 98,
219. doi:10.1039/FD9949800219
[3] J. H. van Zanten, W. E. Wallace, W. L. Wu, Phys. Rev. E Stat. Phys.
Plasmas Fluids Relat. Interdiscip. Topics 1996, 53, R2053.
[4] J. S. Sharp, J. A. Forrest, Phys. Rev. Lett. 2003, 91, 235701.
doi:10.1103/PHYSREVLETT.91.235701
[5] D. S. Fryer, R. D. Peters, E. J. Kim, J. E. Tomaszewski, J. J. de Pablo,
P. F. Nealey, C. C. White, W. L. Wu, Macromolecules 2001, 34, 5627.
doi:10.1021/MA001932Q
[6] R. M. Papaleo, R. Leal, W. H. Carreira, L. G. Barbosa, I. Bello,
A. Bulla, Phys. Rev. B2006, 74, 094203. doi:10.1103/PHYSREVB.74.
094203
[7] K. Fukao,Y. Miyamoto, Phys. Rev. E Stat. Phys. Plasmas Fluids Relat.
Interdiscip. Topics 2000, 61, 1743.
[8] Y. Grohens, L. Hamon, G. Reiter, A. Soldera, Y. Holl, Eur. Phys. J. E
2002, 8, 217. doi:10.1140/EPJE/I2001-10088-4
[9] R. Seemann, K. Jacobs, K. Landfester, S. Herminghaus, J. Polym. Sci.
Part B Polym. Phys. 2006, 44, 2968. doi:10.1002/POLB.20920
[10] C. J. Ellison, M. K. Mundra, J. M. Torkelson, Macromolecules 2005,
38, 1767. doi:10.1021/MA047846Y
[11] C. J. Ellison, J. M. Torkelson, Nat. Mater. 2003, 2, 695. doi:
10.1038/NMAT980
[12] M. K. Mundra, C. J. Ellison, R. E. Behling, J. M. Torkelson, Polymer
2006, 47, 7747. doi:10.1016/J.POLYMER.2006.08.064
[13] M. K. Mundra, S. K. Donthu, V. P. Dravid, J. M. Torkelson, Nano Lett.
2007, 7, 713. doi:10.1021/NL062894C
[14] R. D. Priestley, L. J. Broadbelt, J. M. Torkelson, K. Fukao, Phys. Rev.
E Stat. Nonlin. Soft Matter Phys. 2007, 75, 061806.
[15] R. D. Priestley, C. J. Ellison, L. J. Broadbelt, J. M. Torkelson, Science
2005, 309, 456. doi:10.1126/SCIENCE.1112217
[16] C. B. Roth, J. M. Torkelson, Macromolecules 2007, 40, 3328.
doi:10.1021/MA070162G
[17] P. Rittigstein, R. D. Priestley, L. J. Broadbelt, J. M. Torkelson, Nat.
Mater. 2007, 6, 278. doi:10.1038/NMAT1870
[18] C. B. Roth, K. L. McNerny, W. F. Jager, J. M. Torkelson, Macro-
molecules 2007, 40, 2568. doi:10.1021/MA062864W
[19] M. K. Mundra, C. J. Ellison, P. Rittigstein, J. M. Torkelson, Eur.
Phys. J. Special Topics 2007, 141, 143. doi:10.1140/EPJST/E2007-
00032-0
Conclusion
Fluorescence spectroscopy has been used to determine the
effects of confinement and interfaces on the Tgs and their distri-
butions in PMMA, PEMA, PPMA, and PIBMA films supported
on silica. The average Tg across thin and ultrathin films of
PIBMA is invariant with confinement, whereas that of PMMA
increases with confinement and those of PEMA and PPMA
decrease with confinement. The relative roles of perturbations
to the Tg by free surfaces and polymer–substrate interfaces in
the presence of hydrogen-bonding interactions were determined
by a series of bilayer fluorescence experiments. In all cases,
reductions in Tg relative to bulk Tg were observed in ultrathin
free-surface layers, whereas increases in Tg relative to bulk Tg
were observed in ultrathin substrate-interface layers. The effects
of confinement of the average Tg across single-layer films are
fully consistent with observed relative strengths of the perturba-
tions caused by free surfaces and polymer–substrate interfaces.
When the free-surface effect is greater, as in the case of PEMA,
the average Tg across a film decreases with nanoscale confine-
ment. When the substrate effect is greater, as in the case of
PMMA, the average Tg across a film increases with nanoscale
confinement. When the effects of the substrate and free surface
in perturbingTg are nearly equal, as in the case of PIBMA, within
error the average Tg across a film is independent of nanoscale
confinement.
Acknowledgements
[20] P. Rittigstein, J. M.Torkelson, J. Polym. Sci. Part B Polym. Phys. 2006,
44, 2935. doi:10.1002/POLB.20925
[21] C. J. Ellison, R. L. Ruszkowski, N. J. Fredin, J. M.Torkelson, Phys. Rev.
Lett. 2004, 92, 095702. doi:10.1103/PHYSREVLETT.92.095702
This work was supported by the NSF-MRSEC program at Northwest-
ern University (grants DMR-0076097 and DMR-0520513), Northwestern
University, and a DFI fellowship (R.D.P.).

Effects of Nanoscale Confinement and Interfaces
771
[22] C. J. Ellison, S. D. Kim, D. B. Hall, J. M. Torkelson, Eur. Phys. J. E
2002, 8, 155. doi:10.1140/EPJE/I2001-10057-Y
[41] M. D. Major, J. M. Torkelson, A. M. Brearley, Macromolecules 1990,
23, 1700. doi:10.1021/MA00208A025
[23] D. B. Hall, J. C. Hooker, J. M. Torkelson, Macromolecules 1997, 30,
667. doi:10.1021/MA961345Y
[42] B. Frank, A. P. Gast, T. P. Russell, H. R. Brown, C. Hawker,
Macromolecules 1996, 29, 6531. doi:10.1021/MA960749N
[43] R. Gelles, C. W. Frank, Macromolecules 1982, 15, 1486.
doi:10.1021/MA00234A007
[44] R. O. Loutfy, Pure Appl. Chem. 1986, 58, 1239.
[45] M. L. Gee, T. A. Smith, Aust. J. Chem. 2003, 56, 1005.
doi:10.1071/CH03093
[46] M. Wilhelm, C. L. Zhao, Y. C. Wang, R. L. Xu, M. A. Winnik,
J. L. Mura, G. Riess, M. D. Croucher, Macromolecules 1991, 24,
1033. doi:10.1021/MA00005A010
[24] J. M. Torkelson, R. D. Priestley, M. K. Mundra, P. Rittigstein,
L. J. Broadbelt, C. J. Ellison, Polym. Preprints 2007, 48, 745.
[25] C. L. Soles, J. F. Douglas, R. L. Jones, W. L. Wu, Macromolecules
2004, 37, 2901. doi:10.1021/MA035580+
[26] S. Kawana, R.A. L. Jones, Phys. Rev. E Stat. Nonlin. Soft Matter Phys.
2001, 63, 021501.
[27] J. Q. Pham, P. F. Green, J. Chem. Phys. 2002, 116, 5801.
doi:10.1063/1.1456035
[28] L. Hartmann, W. Gorbatschow, J. Hauwede, F. Kremer, Eur. Phys. J. E
2002, 8, 145. doi:10.1140/EPJE/I2001-10073-Y
[47] F. Varnik, J. Baschnagel, K. Binder, J. Non-Cryst. Solids 2002, 307–
310, 524. doi:10.1016/S0022-3093(02)01481-3
[29] V. Lupascu, S. J. Picken, M. Wubbenhorst, J. Non-Cryst. Solids 2006,
352, 5594. doi:10.1016/J.JNONCRYSOL.2006.09.004
[30] S. Napolitano, D. Prevosto, M. Lucchesi, P. Pingue, M. D’Acunto,
P. Rolla, Langmuir 2007, 23, 2103. doi:10.1021/LA062229J
[31] J. A. Forrest, K. Dalnoki-Veress, J. R. Stevens, J. R. Dutcher, Phys.
Rev. Lett. 1996, 77, 2002. doi:10.1103/PHYSREVLETT.77.2002
[32] C. B. Roth, J. R. Dutcher, Eur. Phys. J. E 2003, 12, 103.
doi:10.1140/EPJED/E2003-01-024-2
[48] A. R. C. Baljon, J. Billen, R. Khare, Phys. Rev. Lett. 2004, 93, 255701.
doi:10.1103/PHYSREVLETT.93.255701
[49] R.A. Riggleman, K.Yoshimoto, J. F. Douglas, J. J. de Pablo, Phys. Rev.
Lett. 2006, 97, 045502. doi:10.1103/PHYSREVLETT.97.045502
[50] S. Peter, H. Meyer, J. Baschnagel, J. Polym. Sci. Part B Polym. Phys.
2006, 44, 2951. doi:10.1002/POLB.20924
[51] K. L. Ngai, J. Polym. Sci. Part B Polym. Phys. 2006, 44, 2980.
doi:10.1002/POLB.20923
[33] C. Svanberg, Macromolecules 2007, 40, 312. doi:10.1021/
MA061967B
[52] C. B. Roth, J. R. Dutcher, J. Electroanal. Chem. 2005, 584, 13.
doi:10.1016/J.JELECHEM.2004.03.003
[34] T. Miyazaki, R. Inoue, K. Nishida, T. Kanaya, Eur. Phys. J. Special
Topics 2007, 141, 203. doi:10.1140/EPJST/E2007-00041-Y
[35] D. B. Hall, J. M. Torkelson, Macromolecules 1998, 31, 8817.
doi:10.1021/MA9812128
[36] A. Dhinojwala, J. M. Torkelson, Macromolecules 1994, 27, 4817.
doi:10.1021/MA00095A024
[37] J. C. Hooker, J. M. Torkelson, Macromolecules 1995, 28, 7683.
doi:10.1021/MA00127A014
[38] C. J. Ellison, J. M. Torkelson, J. Polym. Sci. Part B Polym. Phys. 2002,
40, 2745. doi:10.1002/POLB.10343
[53] M. Alcoutlabi, G. B. McKenna, J. Phys. Condens. Matter 2005, 17,
R461. doi:10.1088/0953-8984/17/15/R01
[54] B. C. McKusick, R. E. Heckert, T. L. Cairns, D. D. Coffman,
H. F. Mower, J. Am. Chem. Soc. 1958, 80, 2806. doi:10.1021/
JA01544A056
[55] A. Dhinojwala, J. C. Hooker, J. M. Torkelson, J. Non-Cryst. Solids
1994, 172–174, 286. doi:10.1016/0022-3093(94)90447-2
[56] D. B. Hall, P. Underhill, J. M. Torkelson, Polym. Eng. Sci. 1998, 38,
2039. doi:10.1002/PEN.10373
[57] S. D. Kim, J. M. Torkelson, Macromolecules 2002, 35, 5943.
doi:10.1021/MA0200322
[39] J. S. Royal, J. M. Torkelson, Macromolecules 1992, 25, 1705.
doi:10.1021/MA00032A013
[40] C. L. H. Wong, J. Kim, C. B. Roth, J. M. Torkelson, Macromolecules
2007, 40, 5631. doi:10.1021/MA071187I
[58] R. D. Priestley, L. J. Broadbelt, J. M.Torkelson, Macromolecules 2005,
38, 654. doi:10.1021/MA047994O