In Situ ATR-FTIR Studies on MgCl

2
-Diisobutyl Phthalate Interactions
in Thin Film ZieglerNatta Catalysts
Ajin V. Cheruvathur,
,
Ernie H. G. Langner,

J. W. (Hans) Niemantsverdriet,
,
and Peter C. Thune*
,,

Schuit Institute of Catalysis, Eindhoven University of Technology, P.O. Box 513, 5600 MB Eindhoven, The Netherlands

Dutch Polymer Institute (DPI), P.O. Box 902, 5600 AX Eindhoven, The Netherlands

Department of Chemistry, University of the Free State, P.O. Box 339, Bloemfontein 9300, South Africa
*S Supporting Information
ABSTRACT: To study the surface structure of MgCl
2
support and its interaction with other active components in
ZieglerNatta catalyst, such as electron donors, we prepared a
thin film analogue for ZieglerNatta ethylene polymerization
catalyst support by spin-coating a solution of MgCl
2
in ethanol,
optionally containing a diester internal donor (diisobutyl-
ortho-phthalate, DIBP) on a flat Si crystal surface. The donor
content of these films was quantified by applying attenuated
total internal reflectionFourier transform infrared spectros-
copy (ATR-FTIR) and X-ray photoelectron spectroscopy (XPS). Changes in the interaction of DIBP with MgCl
2
at various
temperatures were monitored by in situ ATR-FTIR. Upon increasing the temperature, a shift in the (CO) band toward lower
wavenumbers was observed together with the depletion of (OH) stretching band due to the desorption of residual ethanol. We
assign this shift to gradual redistribution of adsorbed DIBP from adsorption sites on the MgCl
2
(104) surface toward the more
acidic MgCl
2
(110) surface. The morphologies of MgCl
2
and MgCl
2
/DIBP films were studied by transmission electron
microscopy (TEM) revealing a preferential orientation of ClMgCl layers (001) parallel to the lateral film dimensions. This
orientation becomes more pronounced upon annealing. In the absence of donor, the MgCl
2
grow in to large crystals aligned in
large domains upon annealing. Both crystal growth and alignment is impeded by the presence of donor.

INTRODUCTION
Most of the heterogeneous MgCl
2
-supported ZieglerNatta
catalysts for stereospecific olefin polymerization contain MgCl
2
as a support, TiCl
4
as catalyst, a trialkylaluminium as cocatalyst,
and two Lewis bases acting as internal and external donors.
1
This catalyst is generally described as particles of activated
MgCl
2
, composed of irregularly stacked ClMgCl sandwich-
like monolayers, with the microstructures terminated by the
(110) and (104) lateral cuts. The reaction of MgCl
2
support
with TiCl
4
leads to the adsorption of TiCl
4
on the lateral
cleavage surfaces of MgCl
2
, such as the (110) and (104)
cuts.
24
The Mg atoms at the surface are coordinated with four
chlorine atoms in the case of the (110) and five chlorine atoms
in the case of the (104) surfaces, as opposed to six chlorine
atoms in the bulk. On reaction with the cocatalyst (aluminum
alkyls), Ti
4+
is reduced to Ti
3+
, forming a TiC bond, which is
essential for monomer insertion. Both an internal and an
external donor are required for an active and stereospecific
high-yield MgCl
2
-supported ZieglerNatta catalyst for propy-
lene polymerization.
5,6
It is commonly accepted that the
internal donor appears to be bound to the (110) lateral cuts of
MgCl
2
and thereby provides [i] stability to the more acidic,
four coordinate (110) surfaces and [ii] stereoregularity to active
titanium catalysts adjacently bound to the same (110) MgCl
2
surface.
711
The catalysts of the fourth generation contain alkyl
esters of aromatic ortho diacids as internal donors (compo-
nents of the solid catalysts) and alkoxysilanes as external donors
(components of cocatalyst mixtures). Nowadays, a phthalate
ester such as diisobutyl phthalate (DIBP) is widely used as an
internal donor in such catalyst systems, where one of its
important functions is to control the amount and distribution
of TiCl
4
.
12
Infrared spectroscopy has long been a useful tool for studying
the chemical structure of heterogeneous catalysts. It is
particularly attractive in the field of ZieglerNatta catalysts
for olefin polymerization, where organic complexes are formed
within the catalyst matrix.
13
The IR spectrum of the solid
DIBP/MgCl
2
/TiCl
4
catalyst is in the 19001550 cm
1
spectral
range and contains a complex envelope of (CO) bands at
ca. 1687 cm
1
(shifted from the 1729 cm
1
of free DIBP) as
well as two narrow bands of the ortho disubstituted benzene
ring of DIBP at 1595 and 1582 cm
1
(aromatic ring modes).
The envelope of the carbonyl group vibrations in the DIBP
coordinated catalyst IR spectra can be resolved into
components.
1416
The (CO) bands are assigned to
complexes of DIBP with MgCl
2
and TiCl
4
, as well as complexes
of MgCl
2
with two derivatives of DIBP (these derivatives are
formed due to the reaction between TiCl
4
and DIBP).
1623
Received: October 11, 2011
Revised: January 4, 2012
Published: January 4, 2012
Article
pubs.acs.org/Langmuir
2012 American Chemical Society 2643 dx.doi.org/10.1021/la203972k | Langmuir 2012, 28, 26432651
Fourier transform infrared (FTIR) spectroscopy has been an
important qualitative tool for studying infrared observable
adsorbates
2428
and, under certain circumstances, can be used
to determine surface concentration.
29,30
The integrated Beer
Lambert relation, a = Alc, is often used to relate surface
concentration to IR peak areas, where a is the IR absorbance
peak area, l is the sample thickness, and c is the surface
concentration. A is the integrated absorption intensity which
can also be thought of as an extinction coefficient.
3133
A
knowledge of such extinction coefficients allows the quantita-
tive determination of the concentrations of surface adsorbed
species using FTIR. When using the attenuated total internal
reflection (ATR) technique, infrared radiation penetrates only a
few micrometers into the bulk of a material, making it an
excellent tool for surface studies of thin films.
3438
For thin
films (thickness less than 200 nm) the measured integrated
absorbance can be approximated to the scale linear with the
concentration of the deposited material.
39
So far, X-ray photoelectron spectroscopy (XPS) has been
used to study the effect of the internal donor in flat-model
ZieglerNatta catalysts, by following the influence of the
internal donor on the electron density around titanium.
4045
This is an indirect method. In direct XPS analysis of organic
adsorbents (here, electron donors), shifts in the C 1s signal
could be interpreted qualitatively, but poor deconvolution of
spectra makes quantitative analysis unreliable.
46,47
This
prompts the use of fluorine as an excellent analytical marker
or doping agent for XPS, as fluorine exhibits a clear quantifiable
peak around 692 eV.
31,4851
It should be noted that XPS data
are influenced by surface topology, lateral heterogeneity and
depth from the sample surface. So, angle resolved XPS analysis
is essential for accurate quantification.
52
In the past, transmission electron microscopy (TEM) was
used for macroscopic observation of polyolefins produced on
ZieglerNatta catalysts.
53,54
Direct TEM imaging of -TiCl
3
was also reported in literature.
55,56
A microscopic analysis of the
initial stage of the polyolefin formation on a TiCl
4
/MgCl
2
catalyst was reported by Barbe et al., who used a gold sputtering
technique for the TEM samples.
57
Using high resolution TEM,
the (110) lateral cut of MgCl
2
in ZieglerNatta catalyst
(without electron donor) was identified.
58
Another high
resolution TEM study on industrial catalysts (with ester or
diester as electron donor) revealed that changes in catalyst
preparation methods can affect the atomic structure of MgCl
2
crystals.
59
In situ real-time environmental TEM studies on
donor-free TiCl
4
/MgCl
2
catalyst identified different lateral
terminations of MgCl
2
.
60
However, the complex nature of real catalyst prevents the
acquisition of the atomic level knowledge about the active
centers for polymerization and their local geometry. The active
centers are hidden in the pore structure of the support, thereby
limiting the exposed surface area for characterization. One
solution to this problem is the development of model catalysts.
There are two different approaches to model the ZieglerNatta
catalysts. One is the computational modeling of MgCl
2
/TiCl
4
/
donor bulk and surface structures by means of DFT.
3,4,8,11,6171
The other one is the experimental investigation of model
catalysts using different characterization techniques.
41,7278
The
flat-model approach facilitates the characterization of the
catalyst by surface spectroscopy and microscopy techni-
ques.
41,78
However, when dealing with model catalyst
comprising extremely small amounts of highly reactive surface
species, special care has to be devoted to creating a sufficiently
clean environment that prevents the irreversible deactivation of
the catalyst. In our case, all components of the ZieglerNatta
catalyst are extremely sensitive to even minute amounts of
moisture in the reactor setup. In situ ATR-FTIR is a good
analytical tool to study the surface chemistry of flat-model
ZieglerNatta catalyst system. It is possible to observe the
changes in the infrared features of internal donors during
thermal and chemical treatments on MgCl
2
, adsorption of
TiCl
4
, catalyst activation in the presence of Aluminum alkyls
and external donors. ATR-FTIR can also be used to study the
ethylene and propylene polymerization as well as the
crystallinity of polymer that is produced. However, only a few
in situ infrared studies are reported in literature.
23,79
The present study focuses on the in situ ATR-FTIR analysis
of MgCl
2
/ethanol/DIBP films on a flat surface. The (CO)
band of the MgCl
2
/DIBP films producing an asymmetric
envelope between 1710 and 1692 cm
1
is the prime interest for
this study. Since the TiCl
4
is not present in this film, the
existence of donorTiCl
4
, donor derivativeMgCl
2
complexes
can be excluded. The DIBP content of the films was quantified
by combining ATR-FTIR and XPS data. Changes in the (C
O) band of the DIBP/MgCl
2
films upon heating in a flowing
argon gas stream were monitored in situ, revealing the changes
in the adsorption mode of DIBP on MgCl
2
surface. The
morphology of MgCl
2
and MgCl
2
/DIBP films, supported on
window etched silicon based grids, were studied by TEM.

EXPERIMENTAL SECTION
Materials. Anhydrous magnesium chloride (ball milled, 99.9%),
absolute ethanol (99.9%), titanium tetrachloride (99.9%), diisobu-
tylphthalate (89%), 4-fluorophthalic anhydride (97%), isobutanol
(99%), anhydrous calcium chloride, concentrated sulphuric acid
(96.7%), dichloromethane, sodium bicarbonate, sodium chloride,
and sodium sulfate were purchased from Aldrich Chemicals and
used as received. Argon (grade 6.0) was purchased from Linde and
used after passing through a Dririte/molecular sieve column (4 ).
Synthesis of 4-Fluoro-diisobutyl Phthalate. 4-Fluoro-diisobu-
tyl phthalate (FDIBP) was synthesized based on a procedure found in
the literature.
80
FDIBP was obtained in high yield by the full
esterification of 4-fluorophthalic anhydride with isobutanol (Support-
ing Information, Scheme S1). Obtained 4-fluoro diisobutylphthalate
(FDIBP) was characterized by
1
HNMR (Supporting Information,
Figure S1) and IR (see Supporting Information, Figure S2).
Preparation of MgCl
2
/Donor Thin Films. From a bulk solution
of MgCl
2
(105 mmol dm
3
) in ethanol, a series of solutions with
donor/MgCl
2
molar ratios of 0, 0.1, 0.15, 0.2, 0.25, and 0.50 was
prepared. All of these preparations were carried out in N
2
atmosphere.
In a typical procedure, a solution was spin-coated (2800 r.p.m.) under
glovebox conditions onto Si crystal (for ATR-FTIR studies), Si wafer
(for XPS studies), and SiO
2
/SiN
x
TEM wafers (for TEM studies).
ATR-FTIR Analysis. All FTIR spectra were collected using a
Nicolet Prote ge 460 Fourier transform infrared spectrometer equipped
with a heated HATR flow cell for Spectra-Tech ARK with a Si 45
crystal (cutoff at 1500 cm
1
). The FTIR spectra of uncoordinated
DIBP and FDIBP were recorded using Nicolet Smart Golden Gate
equipment with a diamond crystal (cutoff at 800 cm
1
). The coated Si
crystal was mounted and sealed under a nitrogen atmosphere in an
ATR-cell. In order to mimic the high vacuum conditions during XPS
measurement and simultaneously prevent moisture and oxygen from
atmosphere entering in to the system, Ar gas was flowed through the
cell for 30 min at 2 bar and 30 C. FTIR spectra were measured in
absorbance mode using a Silicon background (taken at 30 C) with a
resolution of 4 cm
1
and 32 scans per measurement. For temperature
programmed in situ studies, the ATR set up was gradually heated
under argon flow with isothermal steps (10 C difference) from 30 C
up to 150 C. FTIR spectra were recorded at the end of each
Langmuir Article
dx.doi.org/10.1021/la203972k | Langmuir 2012, 28, 26432651 2644
isothermal step (5 min), against Si background (taken at that particular
temperature).
XPS Analysis. XPS measurements were performed with a VG
Escalab 200 using a standard aluminum anode (Al K 1486.3 eV)
operating at 300 W. The coated Si wafer was transferred (using XPS
transfer vessel) under nitrogen atmosphere into the XPS-antechamber
and then under high vacuum conditions into the measurement
chamber. Spectra were recorded at a background pressure of 1 10
9
mbar. Binding energies were calibrated to a Cl 2s peak at 199.5 eV.
Measurements were carried out at a 0 and 60 takeoff angle relative to
the surface normal.
TEM Analysis. TEM studies were carried out on an FEI Tecnai 20
(Sphera), operated at 200 kV. The custom-made TEM wafers have
dimensions of 20 20 mm
2
to allow homogeneous distribution of
solution during spin coating. These wafers consist of 36 individual
TEM grids. The grids are arranged in a square pattern and stabilized
by a silicon frame. The central part of each grid is etched away to
create a 1520 nm thick silicon nitride (SiN
x
) membrane. The size of
this membrane (TEM window) is around 100 100 m
2
, which
allows the electron beam to pass, thus allows imaging of supported
particles in a sub nanometer scale. Before spin coating, the wafers were
calcined in dry air. This improves the strength of the SiN
x
membrane
due to the formation of a 3 nm thick SiO
2
surface layer. After
deposition of MgCl
2
/DIBP solution via spin coating, the TEM wafer
was carefully broken in to individual TEM grids. A single TEM grid
was mounted on the sample holder. Then it was transferred to the
transmission electron microscope, under an inert atmosphere (using
TEM transfer tube). This methodology was already applied for flat-
model systems with iron oxide nanoparticles in morphology studies
related to the carbon nanotube growth and FischerTropsch
catalysis
81,82
and tin oxide nanoparticles produced by atmospheric
pressure chemical vapor deposition.
83

RESULTS AND DISCUSSION

Quantification of Coordinated Donor. A series of
MgCl
2
DIBP mixtures in ethanol with different DIBP to Mg
molar ratio were prepared and spin coated onto the Si ATR
crystal. FTIR spectra of these MgCl
2
/DIBP films were recorded
after 30 min of Ar flow at 30 C. All spectra measured at 30 C
contained a (CO) band around 1700 cm
1
, which can be
assigned to the (CO) band of coordinated DIBP. We
observed a linear increase in peak intensity when increasing the
DIBP to Mg ratio up to 0.25.
A fluorine tagged DIBP homologue (4-fluoro-diisobutyl
phthalate or FDIBP) was synthesized for XPS studies. The
ATR-FTIR spectra of uncoordinated DIBP and FDIBP were
compared (see Supporting Information, Figure S2 and Table
S1). The (CO) bands of uncoordinated DIBP and FDIBP,
were separated by only 3 cm
1
. Apparent similarity of the
(CO) bands is highly favored when considering FDIBP as a
tagged analogue for DIBP, since the CO bond plays a vital
role in coordination with metal centers. ATR FTIR spectra of
the MgCl
2
/ethanol/DIBP and MgCl
2
/ethanol/FDIBP films
were also compared (Supporting Information, Figure S3 and
Table S1). Upon coordination, the
max
(CO) (the carbonyl
band peak maxima) shifts from 1725 to 1705 cm
1
(DIBP) and
1728 to 1708 cm
1
(FDIBP). The [difference between the

max
(CO) of uncoordinated and coordinated donor] is
exactly 20 cm
1
in both cases indicating that, the influence of F
atom in the aromatic ring on the coordination behavior of
phthalate ester is negligible. This makes FDIBP an ideal
substitute for DIBP in XPS studies. MgCl
2
/FDIBP films were
prepared by following the same preparation method used for
MgCl
2
/DIBP films. The same ATR-FTIR experimental
procedure used for FDIBP/MgCl
2
films was followed for
FDIBP/MgCl
2
complexes too. ATR-FTIR spectra of MgCl
2
/
DIBP films and MgCl
2
/FDIBP films with different donor
loadings are shown in panels a and b of Figure 1, respectively.
The peak position of uncoordinated donors (in liquid phase) is
also given for comparison (dotted line). Using TQ Analyst
Software, peak maxima (center of gravity peak location at 10%
threshold) and integrated peak intensities for carbonyl
stretching band of DIBP and FDIBP was determined. Both
DIBP and FDIBP show BeerLambert type behavior when
Figure 1. (a) Carbonyl and aromatic IR bands of DIBP in MgCl
2
/
DIBP films. (a, inset) Plot of the integrated peak intensities of the
carbonyl band and the aromatic ring band (intensity multiplied by 10)
versus the DIBP/Mg ratios in ethanol solution prior to spin coating.
(b) Carbonyl and aromatic IR bands of FDIBP in MgCl
2
/FDIBP films.
(b, inset) Plot of the integrated peak intensities of the carbonyl band
and the aromatic ring band (intensity multiplied by 10) versus the
FDIBP/Mg ratios in ethanol solution prior to spin coating. (c) F 1s
XPS spectra of the MgCl
2
/ethanol/FDIBP films. (c, inset) Plot of the
F/Mg ratios measured by XPS (at a photo electron takeoff angle of 0
and 60, relative to the surface normal) versus the F/Mg ratios in
ethanol solution prior to spin coating.
Langmuir Article
dx.doi.org/10.1021/la203972k | Langmuir 2012, 28, 26432651 2645
plotting the integrated intensity (absorbance mode) of the C
O stretching band and aromatic ring bands versus Donor to Mg
ratio (shown in the insets of panels a and b of Figure 1). This
indicates that donor/MgCl
2
ratio in films reproduce the donor
to MgCl
2
ratio in the spin coating solution.
XPS is used to confirm the ATR-FTIR quantification of the
donor concentration in the film. A fluorinated derivative of
DIBP (FDIBP) was used as the internal donor in the films for
XPS studies. Using XPS, the constituent atoms in the FDIBP/
MgCl
2
films were quantified and the lateral heterogeneity of the
films was established. The combined XPS spectrum (F 1s
region) of MgCl
2
/Ethanol/FDIBP films, with donor/Mg molar
ratios ranging from 0.05 to 0.25 is shown in Figure 1c. Angle
resolved XPS experiments showed that the Mg to F ratio
remained unchanged, irrespective of the photoelectron takeoff
angle 0 or 60 (relative to the surface normal). This is an
indication of homogeneous DIBP distribution throughout the
MgCl
2
film. In a previous study, we determined the thickness of
MgCl
2
film around 11 2 nm.
41
Figure 1c inset represents the
plot of actual FDIBP to Mg ratio calculated using XPS versus
initial FDIBP to Mg ratio in the spin coated solution. XPS
proves the one-to-one correlation between concentration of
donor in solution and film. Therefore, ATR-FTIR can be use
for quantitative estimation of the donor in MgCl
2
/donor films.
In Situ ATR-FTIR Studies during Annealing of the
DIBP/MgCl
2
Films under Argon Flow. The spectra recorded
before starting the Ar flow, represents the state of MgCl
2
/film
as spin coated. The (CO) band of DIBP at that state was
super imposable to uncoordinated DIBP (around 1725 cm
1
),
which lead to the conclusion that the donor was not
chemisorbed on Mg. The IR signal of ethanol (remnants
from the impregnation solution) was also observed. Within the
30 min of Ar flow at 30 C, the (CO) band broadened and
shifted to lower wavenumber (around 1700 cm
1
) indicating
the coordination of DIBP to Mg surface. Simultaneously, we
observed the partial removal of residual ethanol from the film.
The
max
(CO) at room temperature was changed from 1704
to 1715 cm
1
, when the DIBP loading was increased from 5%
to 50% of MgCl
2
. This is probably due to the incomplete
conversion of uncoordinated donor to coordinated donor.
When the temperature is increased from 30 to 150 C, all of the
DIBP/MgCl
2
films showed a clear and progressive shift in

max
(CO) of DIBP to a lower wavenumber (around 1690
cm
1
). Further removal of residual ethanol from the film was
observed. In the films with DIBP concentration above 20 mol
% with respect to MgCl
2
, partial desorption of DIBP was also
observed. As an example, the ATR-FTIR spectra recorded in
between the in situ experiment on MgCl
2
/DIBP film with
donor to Mg ratio of 0.15 is shown in Figure 2. The band
(OH) represents the OH stretching mode of ethanol. This
band disappeared at 150 C due to ethanol removal. At 30 C,
the band (CC) represents the CC stretching modes of
ethanol and donor. The ethanol leaves the surface at high
temperature and shape of the band changes. At 150 C, the
band (CC) solely represents CC stretching modes of
DIBP. XPS studies on MgCl
2
/ethanol films showed that the
amount of residual ethanol in the film at 30 C is around 13%
with respect to MgCl
2
(see the Supporting Information, Figure
S4). Based on this, the amount of residual ethanol with respect
to MgCl
2
in the MgCl
2
/ethanol/DIBP film was estimated,
which is around 18.5 mol % before Ar flow, 9 mol % after 30
min of Ar flow, and 1 mol % after annealing at 150 C. The
peak maxima of (CO) shifted from 1705 to 1688 cm
1
.
After cooling the ATR system to room temperature, the
carbonyl band at 1688 cm
1
remains stationary, indicating an
irreversible shift. In situ ATR-FTIR spectra of DIBP/MgCl
2
films at different temperatures and different loadings were
compared. The surface concentration of DIBP at each
temperature was calculated using the quantification method
(based on BeerLambert relation) discussed in the previous
section. Donor coverage as the function of temperature is
shown in Figure 3. The maximum amount of DIBP in the
MgCl
2
/DIBP film declines from 38% at 30 C to 19% at 150
C. The average peak maxima of the spectra at this stage were
around 1692 cm
1
. In the literature, the broad peak around
1687 cm
1
was attributed as a superposition of bands due to
DIBP coordinated on different surface sites of MgCl
2
.
14
The
shift in
max
(CO) occurs at elevated temperature can be
explained as the changes in the coordination sites on MgCl
2
or
changes in the coordination mode of DIBP. The assignments
for the phthalate ester peak found in literature are summarized
in Table 1. It is believed that the four coordinated MgCl
2
(110)
site is more acidic than the five coordinated MgCl
2
(104) site.
Figure 2. Comparative ATR FTIR spectra of the DIBP/MgCl
2
film
(with a donor to Mg ratio of 0.15) at different stages of in situ
experiment. (OH) is the OH stretching band of ethanol, (CC)
is the CC stretching band of ethanol and donor, and (CO) is the
carbonyl stretching of DIBP.
Figure 3. Thermal desorption pattern of DIBP (calculated from
integrated peak intensities of DIBP carbonyl band).
Langmuir Article
dx.doi.org/10.1021/la203972k | Langmuir 2012, 28, 26432651 2646
The interaction of electron donor with the MgCl
2
(110) site
will be stronger than that of the (104) site, which results in a
lower wavenumber infrared band for electron donor
coordinates on the MgCl
2
(110) site.
The experimental ATR-FTIR spectra (recorded at different
donor concentrations and temperatures) of the MgCl
2
/DIBP
films were simulated as a summation of GaussianLorentzian
function. The best fit for the entire spectra was achieved when
three components with peak maxima around 1725, 1705, and
1685 were used. For the details of simulations (peak fitting), we
refer to the Supporting Information, Table S2 and Figure S5.
Based on the information from Table 1, the observed (CO)
band can be explained as a combination of three overlapping
components corresponding to the DIBP at three different
chemical environments/phases: [i] uncoordinated DIBP, [ii]
DIBP coordinated on MgCl
2
(104) site, and [iii] DIBP
coordinated on MgCl
2
(110) site. The relative amounts of the
DIBP at the above-mentioned environments were calculated
using the quantification method (discussed above). XPS
measurements confirmed that the molar extinction coefficient
of the (CO) envelope does not change upon annealing
(Supporting Information, Figures S6 and S7). The relative
amounts of the three DIBP components in the MgCl
2
films as a
function of temperature are plotted in Figure 4. DIBP remains
uncoordinated in the film after spin coating but prior to argon
flow. A large excess of ethanol in the film indicates that the
layered MgCl
2
crystallites may not yet have formed. Non-
layered MgCl
2
ethanol adduct formation from a saturated
solution of MgCl
2
in ethanol is reported in the literature.
84
As
the ethanol desorbs in the Ar flow, crystalline MgCl
2
forms and
DIBP starts to coordinate on MgCl
2
(104) and (110) sites. The
existence of crystalline MgCl
2
in nonannealed films has been
confirmed by TEM (see below). Within the 30 min of Ar flow,
the majority of the DIBP coordinates on MgCl
2
(104) and
(110) sites. At low donor loading (DIBP to Mg mol ratio
<0.15), DIBP prefers to coordinate on the MgCl
2
(104) site. At
high donor loading (DIBP to Mg mol ratio >0.15), DIBP does
not have any preference between the MgCl
2
(104) or (110)
site. The amount of residual ethanol remaining in the film (after
30 min of Ar flow) is decreases with increasing amounts of
DIBP in the film. The above two observations lead to the point
that ethanol preferably coordinates on MgCl
2
(110) site and
DIBP coordinates on the remaining surface sites. On thermal
annealing, ethanol desorbs from the MgCl
2
(110) surface.
Simultaneously, DIBP migrates from MgCl
2
(104) sites to
MgCl
2
(110) sites. In the case of high DIBP loadings, the
relative abundance of DIBP on MgCl
2
(110) sites also increase
with increasing temperature, up to a point at which the total
intensity of the carbonyl band starts decrease. This decrease in
intensity is only observed in films with donor to Mg ratio above
than 0.20. The maximum donor loading observed on MgCl
2
decreases with increasing temperature (assuming that the
maximum donor loading corresponds to fully covered MgCl
2
surface). This implies that the crystals sinter at high
temperature, even in the presence of internal donor. The
maximum amount of donor coordinated to the MgCl
2
surface
sites was around 18 mol %.
The outcome of this peak fitting suggests the conversion of
uncoordinated DIBP into DIBP coordinated on MgCl
2
(104)
and (110) sites at 30 C, followed by the migration of DIBP
from MgCl
2
(104) sites to MgCl
2
(110) sites at elevated
Table 1. Peak Assignments for Phthalate Esters, Reported in
the Literature
(CO) band
assigned
max
(CO) in
cm
1
uncoordinated phthalate ester (condensed state)/
physisorbed to 6-fold Mg ion at (001) plane
1732,
18
1730,
17
1728
15,16,19,20
1733,
14
1722
21
phthalate ester bound to 5-fold Mg ion at (104)
edge
1705,
22
1704,
20
1700,
16
1699
15,21
phthalate ester bound to 4-fold Mg ion at (110)
edge
1692,
16
1684,
20
1672,
15,21
1662
22
phthalate ester bound to 3-fold Mg ion at corner 1656,
20
1650
15,21
Figure 4. Changes in distribution of DIBP among different components with respect to DIBP loading and temperature (the range of error due to the
covariance of different components is shown along with each data point). The approximate amount of ethanol derived from the intensity of the (O
H) stretching band is included. The ethanol amounts are only a rough estimate due to the uncertainty of the XPS quantification of ethanol in MgCl
2
films.
Langmuir Article
dx.doi.org/10.1021/la203972k | Langmuir 2012, 28, 26432651 2647
temperatures. On spin coating, a thin film of MgCl
2
, ethanol,
and DIBP is formed. The chemical state of MgCl
2
at this stage
is either as nonlayered MgCl
2
-ethanol adducts or MgCl
2
crystals (with lateral surface sites coordinated by ethanol)
formed from this adduct. At this stage, DIBP remains
uncoordinated in the film. During Ar gas flow, a large amount
of ethanol removes from the film, and DIBP coordinates to the
vacancies created by ethanol on MgCl
2
(104) and (110) sites.
Upon annealing, DIBP transfers from MgCl
2
(104) sites to
MgCl
2
(110). The transfer of DIBP occurs in the temperature
range of 5090 C, which is the same temperature range at
which most of the ethanol removes from the surface. It can be
assumed that MgCl
2
(110) sites become vacant on ethanol
removal, and DIBP migrates from less acidic MgCl
2
(104) site
to more acidic MgCl
2
(110) site. It is already known that
MgCl
2
(110) sites are unstable in the absence of adsorbent
molecule, because of the high surface energy of the empty
surface.
3
SEM-EDX studies on Ostwald ripened crystals of
MgCl
2
prepared from DIBP/MgCl
2
/ethanol solution (DIBP to
Mg ratio of 0.1), pointed out the formation of crystallites with
120 and 90 edge angles, which indicates the presence of
(104) and (110) edge surfaces of MgCl
2
with five and 4-fold
Mg ions respectively.
85
The coexistence of MgCl
2
(104) and
(110) sites in the MgCl
2
crystal (in the presence of DIBP) is in
line with the results of peak fitting.
MgCl
2
Film Morphology. Transmission electron micros-
copy was used to study the morphology of MgCl
2
and MgCl
2
/
DIBP films. The TEM images of annealed (at 150 C, for 1 h)
and nonannealed (as spin coated) MgCl
2
and MgCl
2
/DIBP
films were compared (Figure 5). The concentration of DIBP in
the precursor solution for MgCl
2
/DIBP film preparation was 15
mol % with respect to Mg. Note that the nonannealed films are
already went through the vacuum conditions of TEM
measurements. Therefore we expect these samples to be
comparable to the ATR films after a 30 min treatment in
flowing argon. In all cases, MgCl
2
platelets were observed. In
contrast to films prepared by chemical vapor deposition in ultra
high vacuum on metallic surfaces,
74,75,77
these platelets were
arranged parallel to the surface normal (edge-on orientation).
This may be due to the chemical affinity of coordinatively
unsaturated crystal side edges with the underlying SiO
2
layer of
TEM grid. The diffraction patterns of these MgCl
2
and MgCl
2
/
DIBP films (annealed and nonannealed) were compared
(Supporting Information, Figure S8). The most intense signal
in the diffraction pattern of all these films corresponds to a d
spacing of 5.9 , which represent the basal (001) plane of
MgCl
2
.
4,60
The TEM images reveals that, MgCl
2
crystals are
already present in the film at 30 C itself. In the nonannealed
films, the crystalline regions are small, both in the lateral
dimensions of the MgCl
2
platelets as well as in the degree of
stacking along the (001) direction. After annealing, we observe
Figure 5. TEM images of MgCl
2
and MgCl
2
/DIBP films: (a) MgCl
2
film at 30 C. (b) MgCl
2
film after annealing at 150 C for 1 h; FFT view of the
same as inset. (c) MgCl
2
/DIBP film at 30 C. (d) MgCl
2
/DIBP film after annealing at 150 C for 1 h; FFT view of the same as inset.
Langmuir Article
dx.doi.org/10.1021/la203972k | Langmuir 2012, 28, 26432651 2648
a more extensive stacking of the MgCl
2
crystals in both films. In
the annealed, donor free MgCl
2
film (Figure 5b), the platelets
also grow in their lateral dimensions leading to large domains of
slightly bent and twisted MgCl
2
platelets in parallel alignment.
These observations demonstrate that the nanocrystalline MgCl
2
platelets reorganize and sinter upon heating. The annealed
MgCl
2
/DIBP film features comparatively smaller crystals, which
leads to smaller degree of platelet stacking as compared to the
annealed MgCl
2
film. We attribute this effect to DIBP
stabilizing the coordinatively unsaturated magnesium sites at
the perimeter of MgCl
2
platelets. The lateral dimensions of the
MgCl
2
crystals in the annealed DIBP/MgCl
2
film was around
48 nm. This corresponds to approximately 2515% surface
Mg ions available for donor coordination, which is in good
qualitative agreement with the concentration of DIBP on
MgCl
2
surface sites in the film, as shown by ATR experiments.

CONCLUSIONS
A sensitive method was developed for the investigation of
coordinated donor on flat-model ZieglerNatta catalysts by
means of in situ ATR-FTIR. The method is based on the linear
relationship between the carbonyl stretching band integrated
peak intensity of DIBP and concentration of DIBP (concen-
tration of DIBP in the MgCl
2
/DIBP adduct prepared) in
MgCl
2
films. The results were calibrated by XPS studies on a F
labeled homologue of DIBP, which is the 4-fluoro-diisobutylph-
thalate (FDIBP). The newly synthesized FDIBP displayed a
well-quantifiable fluorine 1s peak during XPS-studies of spin-
coated MgCl
2
/ethanol/donor surfaces, as opposed to the
hitherto used, poorly resolved carbon signals of DIBP. ATR-
FTIR spectra of the flat surfaces showed sufficient similarity in
coordination behavior between FDIBP and DIBP to regard 4-
fluoro-diisobutylphthalate as a possible substitute for DIBP in
future XPS studies on flat-model catalyst surfaces. Changes in
the coordination behavior of DIBP on MgCl
2
were followed by
in situ ATR-FTIR spectroscopy. Directly after impregnation,
DIBP shows no interaction with MgCl
2
. However, as some
ethanol desorbs at 30 C in flowing argon, DIBP binds to the
MgCl
2
. We tentatively explain the observed changes of the
(CO) adsorption bands as a combination of three
overlapping components assigned to uncoordinated DIBP,
DIBP adsorbed to MgCl
2
(104), and DIBP adsorbed to MgCl
2
(110). The molecular level structure of the coordination sites
(surface site structure, DIBP coordination mode and DIBP
coverage on surface sites) and the corresponding wavenumbers
for the carbonyl bands of adsorbed DIBP can only be found by
DFT calculations, which is in progress.
Based on our tentative assignment, we derive the following
conclusions:
Before annealing, we observe a slight preference of the
DIBP donor for the MgCl
2
(104) site. This preference is
probably induced by the presence of coadsorbed ethanol.
Upon annealing, DIBP transfers from MgCl
2
(104) sites
to MgCl
2
(110) sites as these adsorption sites become
vacant due to desorption of ethanol from MgCl
2
(110)
sites.
The saturation loading of DIBP after annealing at 150 C
is 18 mol % with respect to Mg. The maximum donor
loading observed on MgCl
2
decreases with increasing
temperature (assuming that the maximum donor loading
corresponds to fully covered MgCl
2
surface). This
implies that the crystals sinter at high temperature,
even in the presence of internal donor. Excess donor is
removed from the film under Ar flow.
TEM studies show that sintering of MgCl
2
films occurs
during annealing. In the absence of donor this sintering leads to
very large MgCl
2
platelets that offer only a very small number
of coordination sites. In the presence of donor, MgCl
2
platelets
remain much smaller due to the stabilization of the surface sites
by adsorbed DIBP. The observed lateral dimensions of the
MgCl
2
platelets (48 nm) agree well with the size require-
ments for DIBP adsorbed on MgCl
2
assuming a full coverage
on both MgCl
2
(104) and (110).
The results show the feasibility of preparing ZieglerNatta
catalysts with different donor concentration on different MgCl
2
surface sites. Since the activity and selectivity of the catalyst is
heavily depends on the coordination behavior of internal donor,
it is interesting to compare the activity and selectivity of these
catalysts. The quantification method discussed in this article
can be applied to different types of donors used in Ziegler
Natta catalysis. Moreover, this quantitative information will be
useful for the in situ studies on the further steps of catalyst
preparation, such as treatment with TiCl
4
and aluminum alkyls.

ASSOCIATED CONTENT
*S Supporting Information
Synthesis of 4-fluoro-diisobutyl phthalate (FDIBP);
1
HNMR
spectra of FDIBP; comparative studies on DIBP and FDIBP
coordination; quantification of ethanol in the MgCl
2
/ethanol
and MgCl
2
/ethanol/donor films; peak fitting procedure; molar
extinction coefficient calibration; electron diffraction pattern of
MgCl
2
, DIBP/MgCl
2
films. This material is available free of
charge via the Internet at http://pubs.acs.org.

AUTHOR INFORMATION
Corresponding Author
*E-mail: p.c.thuene@tue.nl. Telephone: +31 40 247 4997.

ACKNOWLEDGMENTS
This research forms part of the research program of the Dutch
Polymer Institute (DPI) under Project 712. The authors are
grateful to Prof. Vincenzo Busico of University of Naples for
the valuable discussions and comments.

REFERENCES
(1) Moore, E. P., Jr. Polypropylene Handbook: Polymerization; Hanser
Publishers: New-York, 1996; Chapter 2.
(2) Giannini, U. Makromol. Chem. Suppl. 1981, 5, 216.
(3) Busico, V.; Causa, M.; Cipullo., R.; Credendino, R.; Cutillo, F.;
Friederichs, N.; Lamanna, R.; Segre, A.; Castelli, V. V. A. J. Phys. Chem
C 2008, 112, 10811089.
(4) Correa, A.; Piemontesi, F.; Morini, G.; Cavallo, L. Macromolecules
2007, 40, 91819189.
(5) Sacchi, M. C.; Tritto, I.; Locatelli, P. Prog. Polym. Sci. 1991, 16
(23), 331360.
(6) Sacchi, M. C.; Forlini, F.; Tritto, I.; Locatelli, P.; Morini, G.;
Noristi, L.; Albizzati, E. Macromolecules 1996, 29 (10), 33413345.
(7) Taniike, T.; Terano, M. Macromol. Rapid Commun. 2008, 29,
14721474.
(8) Boero, M.; Parrinello, M.; Huffer, S.; Weiss, H. J. Am. Chem. Soc.
2000, 122, 501509.
(9) Brambilla, L.; Zerbi, G.; Piementosi, F.; Nascetti, S.; Morini, G. J.
Mol. Catal. A: Chem. 2007, 263, 103111.
(10) Ribour, D.; Monteil, V.; Spitz, R. J. Polym. Sci. A: Polym. Chem.
2008, 46, 54615470.
Langmuir Article
dx.doi.org/10.1021/la203972k | Langmuir 2012, 28, 26432651 2649
(11) Vanka, K.; Singh, G.; Iyer, D.; Gupta, V. K. J. Phys. Chem. C
2010, 114 (37), 1577115781.
(12) Chadwick, J. C. Macromol. Symp. 2001, 173, 2135.
(13) Terano, M.; Kataoka, T. J. Polym. Sci. A: Polym. Chem. 1990, 28,
20352048.
(14) Zohuri, G.; Ahmadjo, S.; Jamjah, R.; Nekoomanesh, M. Iran.
Polym. J. 2001, 10 (3), 149155.
(15) Potapov, A. G.; Bukatov, G. D.; Zakharov, V. A. J. Mol. Catal. A:
Chem. 2006, 246 (12), 248254.
(16) Kissin, Y. V.; Liu, X. S.; Pollick, D. J.; Brungard, N. L.; Chang,
M. J. Mol. Catal. A: Chem. 2008, 287 (12), 4552.
(17) Arzoumanidis, G. G.; Karayannis, N. M. Appl. Catal. 1991, 76,
221231.
(18) Yang, C. B.; Hsu, C. C.; Park, Y. S.; Shurvell, H. F. Eur. Polym. J.
1994, 30 (2), 205214.
(19) Makwana, U.; Naik, D. G.; Singh, G.; Patel, V.; Patil, H. R.;
Gupta, V. K. Catal. Lett. 2009, 131 (34), 624631.
(20) Singh, G.; Kaur, S.; Makwana, U.; Patankar, R. B.; Gupta, V. K.
Macromol. Chem. Phys. 2009, 210 (1), 6976.
(21) Potapov, A. G.; Bukatov., G. D.; Zakharov., V. A. J. Mol. Catal.
A: Chem. 2010, 316 (34), 9599.
(22) Stukalov, D. V.; Zakharov, V. A.; Potapov, A. G.; Bukatov, G. D.
J. Catal. 2009, 266 (1), 3949.
(23) Ystenes, M.; Bache,

.; Bade, O. M.; Blom, R.; Eliertsen, J. L.;

Nirisen,

.; Ott, M.; Svendsen, K.; Rytter, E. In Future technology for

Polyolefin and Olefin Polymerization catalysis; Terano, M., Shiono, T.,
Eds.; Technology and education Publishers: Tokyo, 2002; pp 184
191.
(24) Yang, A. C.; Garland, C. W. J. Phys. Chem. 1957, 61 (11), 1504
1512.
(25) Arai, H.; Tominaga, H. J. Catal. 1976, 43 (13), 131142.
(26) Srinivas, G.; Chuang, S. S. C.; Debnath, S. J. Catal. 1994, 148
(2), 748758.
(27) Chuang, S. S. C.; Brundage, M. A.; Balakos, M. W.; Srinivas, G.
Appl. Spectrosc. 1995, 49 (8), 11511163.
(28) Balakos, M. W.; Chuang, S. S. C.; Srinivas, G.; Brundage, M. A.
J. Catal. 1995, 157 (1), 5165.
(29) Rasband, P. B.; Hecker, W. C. J. Catal. 1993, 139 (2), 551560.
(30) Srinivas, G.; Chuang, S. S. C. J. Catal. 1993, 144 (1), 131147.
(31) Nakahama, K.; Kawaguchi, H.; Fujimoto, K. Langmuir 2000, 16
(21), 78827886.
(32) Cavanagh, R. R.; Yates, J. T. J. Chem. Phys. 1981, 74 (7), 4150
4155.
(33) Tan, C. D.; Ni, J. F. J. Chem. Eng. Data 1997, 42 (2), 342345.
(34) Bugay, D. E. Adv. Drug Delivery Rev. 2001, 48 (1), 4365.
(35) Planinsek, O.; Planinsek, D.; Zega, A.; Breznik, M.; Srcic, S. Int.
J. Pharm. 2006, 319 (12), 1319.
(36) Tickanen, L. D.; Tejedor-Tejedor, M. I.; Anderson, M. A.
Langmuir 1991, 7, 451456.
(37) Leewis, C. M.; Kessels, W. M. M.; van de Sanden, M. C. M.;
Niemantsverdriet, J. W. (Hans) J. Vac. Sci. Technol. 2006, 24 (2), 296
304.
(38) Harrick, N. J. J. Opt. Soc. Am. 1965, 55, 851857.
(39) Muller, M.; Keler, B. Langmuir 2011, 27 (20), 1249912505.
(40) Kaushik, V. K.; Gupta, V. K.; Naik, D. G. Appl. Surf. Sci. 2006,
253 (2), 753756.
(41) Andoni, A.; Chadwick, J. C.; Milani, S.; Niemantsverdriet, J. W.
(Hans); Thune, P. C. J. Catal. 2007, 247 (2), 129136.
(42) Andoni, A.; Chadwick, J. C.; Niemantsverdriet, J. W. (Hans);
Thune, P. C. Macromol. Rapid Commun. 2007, 28 (14), 14661471.
(43) Da Silva, A. A.; Alves, M. D. M.; dos Santos, J. H. Z. J. Appl.
Polym. Sci. 2008, 109 (3), 16751683.
(44) Zhao, X.; Zhang, Y.; Song, Y.; Wei, G. Surf. Rev. Lett. 2007, 14
(5), 951955.
(45) Mori, H.; Hasebe, K.; Terano, M. J. Mol. Catal. A: Chem. 1999,
140, 165172.
(46) Xiao-hong, G.; Guang-hao, C.; Chii, S. J. Environ. Sci. 2007, 19,
438443.
(47) Andoni, A.; Chadwick, J. C.; Niemantsverdriet, J. W. (Hans);
Thune, P. C. Macromol. Symp. 2007, 260, 140146.
(48) Strother, T.; Knickerbocker, T.; Russell, J. N.; Butler, J. E.;
Smith, L. M.; Hamers, R. J. Langmuir 2002, 18 (4), 968971.
(49) De Giglio, E.; Cometa, S.; Cioffi, N.; Torsi, L.; Sabbatini, L.
Anal. Bioanal. Chem. 2007, 389, 20552063.
(50) Ahmed, S. F.; Mitra, M. K.; Chattopadhyay, K. K. Appl. Surf. Sci.
2007, 253 (12), 54805484.
(51) Snow, A. W.; Foos, E. E.; Coble, M. M.; Jernigan, G. G.;
Ancona, M. G. Analyst 2009, 134 (9), 17901801.
(52) Thomas, H. R.; OMalley, J. J. Macromolecules 1979, 12 (2),
323329.
(53) Kakugo, M.; Sadathoshi, H.; Yokoyama, K.; Kojima, K.
Macromolecules 1989, 22, 547551.
(54) Kakugo, M.; Sadathoshi, H.; Sakai, J.; Yokoyama, M.
Macromolecules 1989, 22, 31723177.
(55) Murry, R. T.; Pearce, R.; Platt, D. J. Polym. Sci., Polym. Lett. Ed.
1978, 16, 303308.
(56) Higuchi, T.; Liu, B.; Nakatani, H.; Otsuka, N.; Terano, M. Appl.
Surf. Sci. 2003, 214, 272277.
(57) Barbe, P. C.; Noristi, L.; Baruzzi, G.; Marchetti, E. Makromol.
Chem. Rapid. Commun. 1983, 4, 249252.
(58) Mori, H.; Sawada, M.; Higuchi, T.; Hasebe, K.; Otsuka, N.;
Terano, M. Macromol. Rapid Commun. 1999, 20, 245250.
(59) Mori, H.; Higuchi, T.; Otsuka, N.; Terano, M. Macromol. Chem.
Phys. 2000, 201, 27892798.
(60) Oleshko, V. P.; Crozier, P. A.; Cantrell, R. D.; Westwood, A. D.
J. Electron. Microsc. 2002, 51 (Supplement), S27S39.
(61) Boero, M.; Parrinello, M.; Terakura, K. J. Am. Chem. Soc. 1998,
120, 27462752.
(62) Boero, M.; Parrinello, M.; Terakura, K. Surf. Sci. 1999, 438, 18.
(63) Boero, M.; Parrinello, M.; Weiss, H.; Huffer, S. J. Phys. Chem. A
2001, 105, 50965105.
(64) Puhakka, E.; Pakkanen, T. T.; Pakkanen, T. A. J. Phys. Chem. A
1997, 101, 60636068.
(65) Seth, M.; Margl, P. M.; Ziegler, T. Macromolecules 2002, 35,
78157829.
(66) Taniike, T.; Terano, M. Macromol. Rapid Commun. 2007, 28,
19181922.
(67) Mukopadhyay, S.; Kulkarni, S. A.; Bhaduri, S. J. Mol. Struct.
THEOCHEM 2004, 673, 6577.
(68) Mukopadhyay, S.; Kulkarni, S. A.; Bhaduri, S. J. Organomet.
Chem. 2005, 690, 13561365.
(69) Stukalov, D. V.; Zakharov, V. A. J. Phys. Chem. C. 2009, 113,
2137621382.
(70) Stukalov, D. V.; Zilberberg, I. L.; Zakharov, V. A. Macromolecules
2009, 42, 81658171.
(71) Stukalov, D. V.; Zakharov, V. A.; Zilberberg, I. L. J. Phys. Chem.
C 2010, 114, 429435.
(72) Siokou, A.; Ntais, S. Surf. Sci. 2003, 540, 379388.
(73) Magni, E.; Somorjai, G. A. Surf. Sci. 1997, 377379, 824827.
(74) Magni, E.; Somorjai, G. A. J. Phys. Chem. B. 1998, 102, 8788
8795.
(75) Kim, S. H.; Somorjai, G. A. Surf. Interface Anal. 2001, 31, 701
710.
(76) Kim, S. H.; Somorjai, G. A. J. Phys. Chem. B 2002, 106, 1386
1391.
(77) Kim, S. H.; Tewell, C. R.; Somorjai, G. A. Kor. J. Chem. Eng.
2002, 19 (1), 110.
(78) Thune, P. C.; Niemantsverdriet, J. W. (Hans) Surf. Sci. 2009,
603, 17561762.
(79) Andoni, A.; Chadwick, J. C.; Niemantsverdriet, J. W. (Hans);
Thune, P. C. Catal. Lett. 2009, 130 (34), 278285.
(80) Skrzypek, J.; Sadlowski, J. Z.; Lachowska, M.; Turzanski, M.
Chem. Eng. Process. 1994, 33 (6), 413418.
(81) Moodley, P.; Loos, J.; Niemantsverdriet, J. W. (Hans); Thune,
P. C. Carbon. 2009, 47, 20022013.
(82) Moodley, P.; Scheijen, F. J. E.; Niemantsverdriet, J. W. (Hans);
Thune, P. C. Catal. Today. 2009, 154 (12), 142148.
Langmuir Article
dx.doi.org/10.1021/la203972k | Langmuir 2012, 28, 26432651 2650
(83) Mannie, G. J. A.; van Deelen, J.; Niemantsverdriet, J. W. (Hans);
Thune, P. C. Appl. Phys. Lett. 2011, 98, 051907.
(84) Sozzani, P.; Bracco, S.; Comotti, A.; Simonutti, R.; Camurati, I.
J. Am. Chem. Soc. 2003, 125, 1288112893.
(85) Andoni, A.; Chadwick, J. C.; Niemantsverdriet, J. W. (Hans);
Thune, P. C. J. Catal. 2008, 257, 8186.
Langmuir Article
dx.doi.org/10.1021/la203972k | Langmuir 2012, 28, 26432651 2651