Vacuum vapour deposition of phenylphosphonic acid on amorphous alumina

Article abstract of DOI:10.1016/j.susc.2007.05.007

A study of the growth of phenylphosphonic acid (PPOA) on amorphous alumina thin films on top of polycrystalline aluminium foil is presented. The self-assembled monolayer (SAM) of the acid was grown in ultra high vacuum via vapour phase deposition, which allows us to investigate the molecular growth process in situ by photoelectron spectroscopy. The acid adlayer was deposited on the alumina surface at 300 K from an evaporator consisting of a ceramic tube heated by a tungsten wire. On adsorption of the PPOA acid on the alumina surface the anchoring properties of the phosphonic group lead to the formation of an ordered layer with an outward phenyl ring group and a phosphonate interface. The C 1s, P 2s, P 2p, Al 2p and O 1s core level spectra were analysed as a function of the deposition time. The acid adsorption occurs in two ranges: formation of an initial monolayer phase, followed by saturation coverage where the P 2s peak intensity is constant with deposition time. It was confirmed that the phenylphosphonic acid molecules react with the alumina surface forming P-O-Al bonds. It was shown that the SAM formation is driven by the evolution of acid molecular coverage on the alumina surface. The results on the vapour deposited PPOA SAM monolayer on the amorphous alumina surface are consistent with previous findings for a number of phosphonic acids SAMs prepared through immersion of the alumina surfaces in an acid solution.

Full text of DOI:10.1016/j.susc.2007.05.007

Surface Science 601 (2007) 3060–3066
www.elsevier.com/locate/susc
Vacuum vapour deposition of phenylphosphonic acid
on amorphous alumina
*
N. Tsud , M. Yoshitake
National Institute for Materials Science, 3-13, Sakura, Tsukuba 305-0003, Japan
Received 21 December 2006; accepted for publication 11 May 2007
Available online 17 May 2007
Abstract
A study of the growth of phenylphosphonic acid (PPOA) on amorphous alumina thin films on top of polycrystalline aluminium foil is
presented. The self-assembled monolayer (SAM) of the acid was grown in ultra high vacuum via vapour phase deposition, which allows
us to investigate the molecular growth process in situ by photoelectron spectroscopy. The acid adlayer was deposited on the alumina
surface at 300 K from an evaporator consisting of a ceramic tube heated by a tungsten wire. On adsorption of the PPOA acid on the
alumina surface the anchoring properties of the phosphonic group lead to the formation of an ordered layer with an outward phenyl
ring group and a phosphonate interface. The C 1s, P 2s, P 2p, Al 2p and O 1s core level spectra were analysed as a function of the depo-
sition time. The acid adsorption occurs in two ranges: formation of an initial monolayer phase, followed by saturation coverage where
the P 2s peak intensity is constant with deposition time. It was confirmed that the phenylphosphonic acid molecules react with the alu-
mina surface forming P–O–Al bonds. It was shown that the SAM formation is driven by the evolution of acid molecular coverage on the
alumina surface. The results on the vapour deposited PPOA SAM monolayer on the amorphous alumina surface are consistent with
previous findings for a number of phosphonic acids SAMs prepared through immersion of the alumina surfaces in an acid solution.
ꢀ
2007 Elsevier B.V. All rights reserved.
Keywords: Self-assembled monolayer; Phenylphosphonic acid; Growth; X-ray photoelectron spectroscopy; Aluminium oxide
1
. Introduction
subtle interplay of the intermolecular interaction and the
interaction with the oxide surface determines the order
and the packing density of the overlayer at the saturation
coverage. Self-assembled monolayers based on the phos-
phonic acid anchor group (see Fig. 1) are good candidates
for the stabilisation of the aluminium oxide surface. The
phosphonic acid anchor group reacts with the hydroxyl
groups in the aluminium oxide surface to form chemi-
sorbed phosphonate complexes that bond the molecule to
the surface. A tail group may be chosen to be compatible
with an adlayer enabling the molecule to act as a bridge
between the aluminium oxide and organic materials, as
shown in Ref. [6] for photochemical attachment of poly-
mers on the aluminium surfaces.
Spontaneously adsorbed organic monolayer films have
received a great deal of attention for their enormous poten-
tial in tailoring surface properties of materials for catalysts,
sensors, nonlinear optics and high density memory devices
[
1,2]. Self-assembly offers one of the most general strategies
for synthesis of new materials for organic nanoelectronics.
For example, organic semiconductors and dielectrics are
very promising in the development of nanosize field-effect
transistors [3–5]. Often a self-assembled system consists
of molecules that interact with one another through a bal-
ance of attractive and repulsive forces and spontaneously
form an ordered aggregate on a certain substrate. The
Generally in organic/inorganic hybrid materials the or-
ganic part is introduced through a coupling molecule, such
as thiol for gold [2] or trichlorosilane for silicon [1] sur-
faces. The coupling molecule on the base of the phosphonic
*
Corresponding author. Fax: +81 29 863 5571.
E-mail address: Nataliya.Tsud@nims.go.jp (N. Tsud).
0
039-6028/$ - see front matter ꢀ 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.susc.2007.05.007

N. Tsud, M. Yoshitake / Surface Science 601 (2007) 3060–3066
3061
self-assembled monolayer and the surface of amorphous
alumina on top of polycrystalline aluminium.
C
C
C
C
Tail
2. Experimental
C
C
O
O
P
Experiments were performed in a commercial X-ray
O
Anchor
photoelectron spectroscopy (XPS) system, with a base
À9
pressure of 2 · 10 Torr. The UHV chamber was
H
equipped with a hemispherical electron energy analyzer, a
monochromatized Al Ka X-ray source, and an ion gun.
The core level spectra were recorded at 45ꢁ emission of
the photoelectrons with respect to the surface normal.
The XPS spectra were taken at the pass energy of
58.7 eV. Since the X-ray source was monochromatised,
the spectral resolution was mainly limited by the analyser
resolution, which was 1 eV for used pass energy. The Al
metal 2p core level at 72.9 eV was used as binding energy
reference in the course of the experiment.
Fig. 1. Structure of the phenylphosphonic acid molecule.
acid anchor group shows great potential for aluminium
and other metals [7,8]. In comparison with the large
amount of literature on the self-assembly of alkanethiols
on gold surfaces [2], relatively little has been published
on phosphonic acids SAMs on the aluminium surface. In
almost all experimental works, polycrystalline aluminium
covered by an oxide layer was used as the substrate for
the phosphonic acid SAMs prepared ex-situ, through
immersion in a solution [9–15]. It was shown that the phos-
phonic acid treatment improves the corrosion inhibition
and coating adhesion properties of the aluminium [16].
Moreover the alkane phosphonic acids SAMs were found
to be effective protection layer of the Al O /Al films in pat-
The thin amorphous alumina film was used as a support
for the PPOA molecules deposition. First the mechanically
+
and chemically polished Al foil was cleaned by Ar sputter-
ing. Then it was oxidized by exposing to oxygen (at
À2
1 · 10 Torr for 10 min) in the fast entry chamber. The
vacuum oxidation of the clean Al foil was chosen in order
to avoid the surface carbon impurities of a natural oxide
film formed by air exposure of aluminium. The oxide thick-
ness before deposition (about 0.5 nm) was estimated from
the Al 2p oxide to metal peaks ratio. The cleanliness of
the surface was checked by XPS. No charging was ob-
served during the spectroscopic studies.
2
3
tern transfer by wet chemical etching [17]. It is known that
phosphonic acids are highly reactive to aluminium oxide
surfaces and can form stable and well ordered monolayers
depending on the reaction conditions. Tridentate coordina-
tion of the phosphonic acid molecule on the alumina sur-
face was found, independent of the adsorption site and
the surface treatment [18–20]. First principles total energy
calculations for vinyl phosphonic acid interaction with
the hydroxylated a-Al O (0001) surface led to the conclu-
Core levels of O 1s were fitted by a convolution of a
Gaussian and a Doniach–Sunjic line shape, with subtrac-
tion of a Shirley type background, to separate different
peak components, their positions and widths.
The PPOA powder was bought from TCI Organic
Chemistry supplier with a purity of 98%. Some PPOA acid
powder pressed into an In foil was used as a reference sam-
ple to yield XPS data for the molecule before evaporation.
The PPOA evaporation source was mounted in the fast
2
3
sion that the acid molecules are site sensitive and the most
favourable coordination is tridentate [21].
In Ref. [22] the chemisorption of organophosphorus
compounds on a clean and preoxidised surface of an
Al(111) single crystal was studied by photoelectron spec-
troscopy. To our knowledge that was the only attempt at
vapour phase deposition of methyl phosphonic acid. The
authors found it to be unsuccessful because of acid decom-
position on the walls of the vacuum chamber. To have a
reference for comparison with other organophosphorus
compounds, the methyl phosphonic acid was chemisorbed
on the Al(111) substrate from solution in a helium atmo-
sphere [22].
The present work is the first effort to grow phosphonic
acid SAM in vacuum via vapour phase deposition, since
it allows the study of organic adlayer formation in situ,
eliminating interference due to a solvent. We have chosen
the phenylphosphonic acid (see Fig. 1) because of its rela-
tively high melting temperature of about 440 K expected
to give efficient vacuum evaporation. The aim of the study
was the formation, and characterisation by photoelectron
spectroscopy, of the interface between the PPOA molecules
À8
entry chamber with a base pressure of 3 · 10 Torr. The
acid was evaporated from a Knudsen cell equipped with
a thermocouple attached to the back side of the crucible.
The temperature of the crucible was carefully stabilised be-
fore evaporation. The PPOA acid evaporation temperature
was 380 K; the alumina substrate was held at 300 K during
deposition. The pressure during deposition was 6 ·
À8
10 Torr. The ion gauge used to measure the pressure
was estimated to be a factor of 7 more sensitive to PPOA
vapours than to nitrogen. The estimation is made by com-
parison with the known sensitivity factors for the molecules
such as benzene (6), benzoic acid (5.5), chlorobenzene (7),
nitrobenzene (7.2) given in the manual to the ionization
vacuum gauge. The deposition rate was estimated from
the XPS data. We checked the amount of acid evaporated
on different days and estimate that the variation in deposi-
tion rate due to loss of acid powder from the cell was no
more than a few percent over the course of a day.

3062
N. Tsud, M. Yoshitake / Surface Science 601 (2007) 3060–3066
3
. Results
1000
500
0
ratio
P 2s
2
1
1
0
0
.0
.5
.0
.5
.0
We start with the construction and tests of the PPOA
evaporator. To find suitable evaporation conditions, we
first prepared the PPOA thin film on the amorphous alu-
mina surface. By comparison of the photoemission spectra
from the thin acid film (not shown here) and the acid pow-
der reference sample we found that neither the hot filament
of the ion gauge nor the vacuum chamber walls affects the
stability and the structure of the acid molecules. No change
in the PPOA molecule adlayer with increasing X-ray
exposure time was observed. The kinetics of molecular
adsorption was found to depend on the PPOA molecule
impingement rate. For deposition rates higher than 0.05
ML/min (ML = monolayer), the adsorbed PPOA mole-
cules form a multilayer thin film on the alumina surface,
whereas for lower rates the photoemission signals reach
saturation and remain constant (see Fig. 2).
0
20
40
60
Deposition time [min]
Fig. 3. Ratio of C 1s to Al 2p core level intensity and P 2s core level
intensity plots versus PPOA deposition time.
Fig. 4 illustrates the C 1s core level as a function of the
PPOA evaporation time compared to the signal from the
PPOA powder pressed into the In foil. The spectra from
To provide an overview of the PPOA molecule overlayer
formation, we followed the dependence of the C 1s, P 2p, P
4
0 to 70 min deposition time are omitted for clarity. For
each 5 min deposition step the C 1s core level is a single
symmetric well defined peak. We were unable to distinguish
different components in the C 1s core level. The peak shape
and width remain constant at all adsorption steps. We as-
sign the peak to the carbon atoms of the phenyl ring of the
PPOA molecule. The C 1s peak width of the reference acid
powder was found to be 1.1 eV that is close to the value for
PPOA SAM on the amorphous alumina surface (1.4 eV).
On the PPOA monolayer growth the binding energy of
the C 1s core level shifts to lower binding energies by about
2
s, Al 2p and O 1s core level spectra on the deposition time
(
for 0.05 ML/min deposition rate). The low ionisation
cross section together with the low surface concentration
of phosphorus atoms prevent detailed analysis of the P 2s
and P 2p core levels. Moreover for the P 2p core level only
the binding energy was analysed because the peak overlaps
the Al 2s plasmon loss.
C 1s intensity growth accompanied by the reduction of
the Al 2p signal with increasing coverage of the adsorbed
molecules was observed. The ratio of C 1s to Al 2p core
level intensities versus the PPOA deposition time is plotted
in Fig. 3. The intensity ratio plot displays the changes in
the surface coverage and ordering excluding experimental
uncertainties. For the first 20 min of deposition, it grows
gradually and then reaches the condition where the ratios
fluctuate. Thus, it appears that 20 min deposition time cor-
responds to the completion of the monolayer coverage, i.e.
to the formation of the PPOA molecule SAM on the amor-
phous alumina surface. The small changes after 20 min
deposition are driven by subtle evolution in coverage
and/or film structure. We address this behaviour in greater
detail in the discussion section below.
0.7 eV (see Fig. 5) and for PPOA SAM on the amorphous
alumina surface is still 0.3 eV lower than the reference
value for the acid powder. The C 1s binding energy varia-
tion at the constant peak width was explained by the
changes of the final state screening of the core hole by
the increasing number of the neighbouring molecules.
As it was mentioned above, oscillation in the photoelec-
tron intensity of the C 1s and P 2s core levels was observed
after the formation of the PPOA SAM on the amorphous
alumina surface. There is a systematic change in the P 2s
intensity plot (see Fig. 3) which seems to be related to the
coverage change rather than to the instrumental effects.
C 1s
0.3 eV
7
6
PPOA powder
0
0
.15 ML/min
.08 ML/min
5
4
3
2
1
0
0.05 ML/min
3
5 min
0
min
0
20
40
60
80
290
289
288
287
286
285
284
283
Deposition time [min]
Binding energy [eV]
Fig. 2. Ratio of C 1s to Al 2p core level intensity as a function of PPOA
deposition time and deposition rate.
Fig. 4. C 1s core level as a function of acid evaporation time compared to
the signal from acid powder pressed into In foil.

N. Tsud, M. Yoshitake / Surface Science 601 (2007) 3060–3066
3063
2
2
2
86.5
86.0
85.5
P 2p
0.3 eV
P 2p
oxide
PPOA SAM/Al
2 3
O /Al
1
1
1
34.5
34.0
33.5
PPOA powder
C 1s
1
40
138
136
134
132
130
2
85.0
Binding energy [eV]
0
20
40
60
Deposition time [min]
Fig. 6. P 2p core level spectra for PPOA SAM on alumina and acid
powder pressed into In foil.
Fig. 5. Binding energy change of the C 1s and P 2p core levels versus
PPOA deposition time.
Al 2p
2
1
1
000
500
000
After 20 min deposition time the acid coverage is still
increasing, then at 25 min it starts to decrease and at
Al
2 3
O /Al
PPOA SAM/Al
2
3
O /Al
4
0 min reaches the value of 20 min deposition time. The
same behaviour could be seen for the C 1s to Al 2p inten-
sity ratio and C 1s binding energy plots (see Figs. 3 and 5).
So we can eliminate an error in the X-ray intensity change
or other instrumental instabilities during measurements.
The P 2p binding energy plot versus the PPOA deposi-
tion time is shown in Fig. 5. With increasing acid deposi-
tion time the Al plasmon loss vanishes and the P 2p peak
gradually shifts from the binding energy of 134.2 eV to
0.2 eV
5
00
0
7
8
76
74
72
Binding energy [eV]
Fig. 7. Al 2p core level spectra for PPOA SAM on alumina compared
with the signal for the clean oxide surface.
1
34.6 eV at saturation coverage. The peak shift indicates
either coverage changes or conformation reorganisation
in the adsorbed monolayer. It can be explained by the fact
that PPOA adsorption can start with physisorption or
arrangement of molecules in mono- or bidentate geometry.
In other words, the tridentate bonding of the PPOA mole-
cules does not occur immediately. Only small oscillations
were observed in the P 2p binding energy dependence after
PPOA monolayer formation. The P 2p spectra for the
PPOA SAM on alumina and the reference PPOA powder
in comparison with the signal from the clean amorphous
alumina surface, dominated by the Al 2s plasmon loss
peak, are presented in Fig. 6. The 0.3 eV binding energy
difference with respect to the value for the acid powder
was assigned to the formation of P–O–Al bonds on the alu-
mina surface.
Fig. 7 shows the Al 2p core levels for the clean alumina
and the SAM of the PPOA molecules on the alumina sur-
face. Besides a decrease in intensity only small changes
were observed for the Al 2p core level on the PPOA
adsorption. For the oxidised component of the Al 2p core
level a binding energy difference of 0.2 eV was found. We
were unable to separate a component corresponding to
the P–O–Al bonds formation by fitting of the Al 2p core
level. A small difference in the Al oxide 2p binding energy
indicates a change in the bonding surroundings of the Al
atoms, which is indicative of PPOA SAM formation.
Further information was obtained from an examination
of the O 1s core level. Similar to Al 2p for the O 1s core
level only small changes were detected. Fig. 8 shows O 1s
peak for clean alumina surface, the PPOA SAM on the
amorphous alumina and the PPOA powder reference sam-
ple. Due to the large O 1s signal from the thin alumina
layer it was difficult to track changes due to the acid
adsorption. The analysis of the spectra revealed that the
O 1s peak could be fitted by two components. The thin alu-
mina film gives rise to a component at 532.2 eV, which is in
agreement with previously published values [23,24]. For the
clean substrate another component was separated with
maximum at 533.8 eV that was contributed to OH groups
on the oxide surface [24]. For the PPOA SAM on the
amorphous alumina surface the component with similar
binding energy of 533.4 eV was found and assigned to
the P–O–Al functional groups. The position of this peak re-
mains constant with increasing PPOA molecule coverage.
The P–O–Al peak shifts by 0.4 eV to the lower binding en-
ergy with respect to the OH group component of the clean
surface, while the oxide component remains at the same
position. It confirms that the PPOA molecule interacts with
the OH groups on the alumina surface making three P–O–
Al bonds.
The O 1s peak for the PPOA reference sample also con-
sists of two clear observable components. We did not ob-
serve indium core levels in the photoelectron spectrum of
the PPOA powder pressed into In foil. Taking in the ac-
count the absence of the indium signal, while the O 1s
and In 3d core levels have the comparable ionisation cross
section, we conclude that both O 1s components belong to

3064
N. Tsud, M. Yoshitake / Surface Science 601 (2007) 3060–3066
Table 1
O 1s
Al
Al-O
Binding energy of the O 1s core level components for different surfaces
1
0
0
.0
.5
.0
2
O
3
/Al
Substrate
Assignment
(eV)
Assignment
b
E (eV)
E
b
Al-OH
Amorphous Al
PPOA SAM on Al
PPOA powder
Polyacrylic acid [25,26]
Methylphosphonic acid
on Al(111) [22]
Octadecylphosphonic acid on
2
SiO /Si [27]
Octadecylphosphoric acid [28]
Alkylphosphonic acid [12]
2
O
3
/Al
O /Al
2 3
Al–O 532.2
Al–O 532.2
P@O 532.3
C@O 532
Al–O–H 533.8
Al–O–P 533.4
P–O–H 533.7
C–O 533.7
P@O 535
P–O–Al 533.5
5
36
534
532
530
530
530
P@O 534.7
P–O–Si 531.1
O 1s
PPOA SAM/Al
Al-O
P@O 532.1
P@O 532.7
P–O–H, P–O–C 533.6
P–O–H 533.2
2
3
O /Al
1
0
0
.0
.5
.0
Al-O-P
groups in any of XPS handbooks. The most similar avail-
able reference data is for the polyacrylic acid and other
polymers containing C@O and C–O functional groups.
The reported O 1s binding energies lie in the range of
532.2–532.0 eV for the C@O bond and at 533.8 eV for
5
36
534
532
the C–O bond [25,26] that supports our assignment.
Davies and Newton [22] and Gouzman et al. [27] have
attempted to identify P@O and P–O functional groups
in the O 1s photoelectron spectra of the methylphosphonic
acid on the Al(111) surface and octadecylphosphonic acid
À
O 1s
PPOA powder
P-OH
1
0
0
.0
.5
.0
P=O
on SiO /Si surface, respectively. The P@O functional group
2
was associated with O 1s peak components at 535 eV [22]
and 534.7 eV [27], which disagree with the present data
for the PPOA powder. In Ref. [27] they referred to the
work of Davies and Newton where P@O bond assignment
was done by comparison with C@O group in acetone ad-
sorbed on aluminium surfaces. The binding energy values
5
36
534
532
Binding energy [eV]
À
for the P–O peak in the O 1s spectra are in much better
Fig. 8. Spectra of the O 1s core level of the clean alumina surface, PPOA
SAM on the alumina surface and PPOA powder pressed onto In foil.
agreement: 533.5 eV for P–O–Al bond [22] and 533.1 eV
for P–O–Si bond [27] compared to the 533.7 eV for P–O–
H group of the PPOA powder and 533.4 eV for P–O–Al
group of the PPOA SAM on amorphous alumina surface.
Our assignment is in agreement with the definition of the O
the PPOA molecules. We attribute them to two types of
oxygen present in the PPOA molecular structure: first –
in the double bond with phosphorous atom P@O, and sec-
ond in P–O–H bonds. The O 1s components at 532.3 eV
and 533.7 eV were assigned to the P@O and P–O–H
groups, respectively. The P@O component coincides in en-
ergy with the O1s peak of alumina, thus only minor
changes can be expected there on PPOA acid adsorption.
The P–O–Al component appears at 533.4 eV for the PPOA
SAM on alumina, which is 0.3 eV lower than the P–O–H
component of the reference powder. This evidence, com-
bined with the observed shift of the OH peak in the O 1s
clean alumina spectra, is another indication that the phos-
phonic acid undergoes a condensation reaction with sur-
face Al–OH species forming aluminophosphonates.
1s components for the octadecylphosphoric [28] and alkyl-
phosphonic [12] acids, for details see Table 1.
4. Discussion
PPOA molecules react with the surface hydroxyl groups
on the alumina surface giving covalently bound surface
phosphonate species as a result of the condensation reac-
tion. The photoemission data proved that the PPOA
anchor group is attached to the substrate through the
Al–O–P bonds with the phenyl ring oriented away from
the surface.
In the ideal case, for the PPOA molecule the P@O to P–
O–H groups intensity ratio of the O 1s core level compo-
nents has to be 1:2. The experimental value was found to
be 1:1.5. Taking into account the possible water or OH
groups adsorption from air, which we cannot separate in
the O 1s signal, we conclude that the relative peak compo-
nents ratio supports our assignment. We did not find the
binding energy values of the P@O to P–O–H functional
The intensity behaviour of the photoemission peaks
demonstrates that the PPOA acid adsorption is well de-
scribed by a two stage process in which an initial mono-
layer phase is rapidly formed followed by a saturation
plateau corresponding to the monolayer completion. The
binding energy changes up to 20 min deposition are
0.7 eV for C 1s, 0.4 eV for P 2p, 0.2 eV for the oxidised
component of Al 2p, and the Al–O–P component of the

N. Tsud, M. Yoshitake / Surface Science 601 (2007) 3060–3066
3065
O 1s core level remains constant. It accounts for the first
stage in the PPOA molecule adsorption that includes pos-
sible physisorption and/or the presence of mono- or biden-
tate coordinated phosphonate anchor groups on the
alumina surface.
between the molecular axis and the alumina surface normal
is zero. Thus the film thickness is given by the dimension of
the PPOA molecule while it is connected to the alumina
surface by the phosphonate anchor group, which is
0.5 nm [30]. The calculated ratio of Al 2p oxidised compo-
nent for PPOA SAM on alumina to one from clean surface
was found to be 0.85. The measured value gave 0.65 at
20 min deposition time. Taking into account the uncer-
tainty in the mean free path value for monolayer of the
PPOA molecule we conclude that difference of 25% can
be considered good agreement between the estimated and
measured Al 2p core level signal reduction as the result
of PPOA SAM formation on the amorphous alumina
surface.
We calculated that the top layer of the oxide film con-
tributes about 80% of the total Al 2p signal for a 0.5 nm
thick amorphous alumina film. Provided that one PPOA
molecule binds to three Al atoms on the surface making
three Al–O–P bonds we can calculate the surface stoichi-
ometry from the intensities of photoemission spectra nor-
malised to the ionisation cross sections of corresponding
atoms. In the ideal case the ratio of P:C:Al atoms on the
surface are calculated to be 1:6:3. The experimental value
was found to be 1:6:2.7 for the PPOA SAM on the amor-
phous alumina surface that confirms the fact that 20 min
deposition time resulted in the formation of a saturated
layer of PPOA molecules. This is in agreement with litera-
ture data, as the presence of the tridentate coordinated P
atoms on the substrate surface is commonly observed upon
phosphonic acid adsorption from acid solution.
The PPOA molecule adsorption then proceeds by a
stage in which the photoemission signals have a fluctuating
behaviour. The fact that increasing adsorption time leads
to an even smaller photoemission signals may indicate
either the desorption of the weakly bound molecules from
a second layer or the long term reorganisation of the al-
ready chemisorbed molecules. Taking into account the
imperfect structure of the substrate surface (the oxide is
amorphous on the polycrystalline aluminium) we expect a
domain like structure of the SAM molecules without long
range order. Thus the existence of the different phases of
the PPOA molecular adlayer at slightly different coverage
has low probability. We conclude that the PPOA adsorp-
tion process on the alumina surface is driven by the evolu-
tion of the coverage through a continuous variation in the
local monolayer structure and is not a result of changes in
the molecular orientation or conformation.
The timescale of the PPOA molecule adsorption on the
alumina surface was also examined. There has been discus-
sion in the literature concerning long term reorganisation
processes, in which it is believed that the time scale for
molecular organisation is significantly longer than the time
scale for adsorption. Allara and Nuzzo for alkanoic acid
SAMs on an oxidized Al surface, prepared by immersion
in the solution, concluded that the organisation process
of the molecules takes several days [29]. We did not ob-
served any changes in the photoemission signals between
data taken right after deposition and at the end of measure-
ments, which took about an hour. On the other hand we do
see an intensity decrease of the adsorbed layer signals for
the same surface from one day to the next morning at high
PPOA coverages or for a few hours break during the
course of the experiment. For instance, in Fig. 2 two points
at 20 min deposition time were taken with 4 h break. We
expect that the adsorption energy of weakly bound mole-
cules in the second adlayer corresponds to very slow
desorption. We proved that the desorption process could
be activated by the next deposition, i.e., by the impinge-
ment of the PPOA molecules on the substrate which hit
the surface, transfer their kinetic energy to the second ad-
layer molecules and cause their desorption. The deposition
rate of 0.05 ML with 5 min step was shown to be good
evaporation condition to grow the PPOA SAM monolayer
on the amorphous alumina surface. Apparently for higher
deposition rates the slow desorption has minor impact on
the fast film growth, i.e., the equilibrium conditions for
the self-assembly is not achieved for deposition rates higher
than 0.05 ML/min.
Finally it should be mentioned that the preliminary
desorption experiments yield a desorption temperature to
be higher than 500 K for PPOA SAM on amorphous alu-
mina. The above arguments support the concept of the for-
mation of SAM on the surface of the amorphous alumina.
5. Conclusions
We have reported a study of the formation of the phos-
phonic acid – aluminium oxide interface. This type of inter-
face is very important because the phosphorus anchor
group could be used as a universal coupling unit to form
chemical bonds between aluminium and organic materials
for organic field effect transistors and/or sensors. The
PPOA molecule was found to be stable against X-ray expo-
sure and the presence of the hot filaments in the vacuum
chamber. The molecular adsorption was found to depend
on the PPOA molecule impingement rate. For deposition
rates higher than 0.05 ML/min the adsorbed PPOA mole-
cules form the multilayer thin film on the alumina surface,
whereas for lower rates the photoemission signals reach the
saturation and remain constant. The photoemission spec-
tra were analysed as a function of the deposition time. It
was confirmed that the PPOA molecules react with the alu-
mina surface forming a phosphonate interface with an out-
ward phenyl ring tail group. The results are consistent with
previous findings for a number of phosphonic acid SAMs
We have estimated theoretically the reduction of the Al
2
p signal covered by the PPOA SAM with respect to the
signal from the clean alumina surface. First we estimate
the thickness of the PPOA SAM provided that the tilt angle

3066
N. Tsud, M. Yoshitake / Surface Science 601 (2007) 3060–3066
prepared through immersion of the alumina surfaces in an
acid solution.
[13] E. Hoque, J.A. DeRose, P. Hoffman, H.J. Mathieu, B. Bhushan, M.
Cichomski, J. Chem. Phys. 124 (2006) 174710.
[14] I.L. Liakos, R.C. Newman, E. McAlpine, M.R. Alexander, Surf.
Interface Anal. 36 (2004) 347.
Acknowledgements
[15] M.J. Brukman, G.O. Marco, T.D. Dunbar, L.D. Boardman, R.W.
Carpick, Langmuir 22 (2006) 3988.
N. Tsud acknowledges support of the Japan Society for
the Promotion of Science. This work was also supported by
the Hayashi Research Foundation for Women Research-
ers. We thank K.C. Prince for useful discussions.
[16] I. Maege, E. Jaehne, A. Henke, H.J.P. Adler, C. Bram, C. Jung, M.
Stratmann, Prog. Org. Coat. 34 (1998) 1.
[
[
[
17] L.B. Goetting, T. Deng, G.M. Whitesides, Langmuir 15 (1999) 1182.
18] R.D. Ramsier, P.N. Henriksen, A.N. Gent, Surf. Sci. 203 (1988) 72.
19] G. Steiner, V. Sablinskas, O. Savchuk, R. Bariseviciute, E. Jahne,
H.J. Adler, R. Salzer, J. Mol. Struct. 661–662 (2003) 429.
References
[20] J.S. McNatt, J.M. Morgan, N. Farkas, R.D. Ramsier, T.L. Young, J.
Rapp-Cross, M.P. Espe, T.R. Robinson, L.Y. Nelson, Langmuir 19
(
2003) 1148.
[
1] D.K. Aswal, S. Lenfant, D. Guerin, J.V. Yakhmi, D. Vuillaume,
Anal. Chim. Acta 568 (2006) 84.
[
21] L.G. Hector Jr., S.M. Opalka, G.A. Nitowski, L. Wieserman, D.J.
Siegel, H. Yu, J.B. Adams, Surf. Sci. 494 (2001) 1.
[
[
2] F. Schreiber, Prog. Surf. Sci. 65 (2000) 151.
3] J. Collet, O. Tharaud, A. Chapoton, D. Vuillaume, Appl. Phys. Lett.
[
[
22] P.R. Davies, N.G. Newton, Appl. Surf. Sci. 181 (2001) 296.
23] B.V. Crist, XPS Handbook of the Elements and Native oxides, XPS
International, Inc., 1999.
24] E. Tan, P.G. Mather, A.C. Perrella, J.C. Read, R.A. Buhrman, Phys.
Rev. B 71 (2005) 161401.
25] N. Ikeo, Y. Iijima, N. Nimura, M. Sigematsu, T. Tazawa, S.
Matsumoto, K. Kojima, Y. Nagasawa, Handbook of X-ray Photo-
electron Spectroscopy, JEOL, Tokyo, 1991.
26] B.V. Crist, Handbook of Monochromatic XPS Data, Wide and
Narrow Scan Spectra of Polymers, vol. 3, Hakuto Co, Ltd., 1991.
27] I. Gouzman, M. Dubey, M.D. Carolus, J. Schwartz, S.L. Bernasek,
Surf. Sci. 600 (2006) 773.
28] M. Textor, L. Ruiz, R. Hofer, A. Rossi, K. Feldman, G. Hahner,
N.D. Spencer, Langmuir 16 (2000) 3257.
29] D.L. Allara, R.G. Nuzzo, Langmuir 1 (1985) 45.
30] C. Gervais, M. Profeta, V. Lafond, C. Bonhomme, T. Azais, H.
Mutin, C.J. Pickard, F. Mauri, F. Babonneau, Magn. Reson. Chem.
76 (2000) 1941.
[
[
4] J. Collet, D. Vuillaume, Appl. Phys. Lett. 73 (1998) 2681.
5] T.W. Kelley, L.D. Boardman, T.D. Dunbar, D.V. Muyres, M.J.
Pellerite, T.P. Smith, J. Phys. Chem. B 107 (2003) 5877.
[
[
[6] J. Pahnke, J. R u¨ he, Macromol. Rapid Commun. 25 (2004) 1396.
[7] P.H. Mutin, G. Guerrero, A. Vioux, J. Mater. Chem. 15 (2005) 3761.
[8] A. Vioux, J. Le Bideau, P.H. Mutin, D. Leclercq, New Aspects in
Phosphorus Chemistry IV Topics in Current Chemistry 232 (2004)
[
[
[
145.
[
9] M.J. Pellerite, T.D. Dunbar, L.D. Boardman, E.J. Wood, J. Phys.
Chem. B 107 (2003) 11726.
[
[
[
10] L. Forget, F. Wilwers, J. Delhalle, Z. Mekhalif, Appl. Surf. Sci. 205
2003) 44.
11] T.A. Lewington, M.R. Alexander, G.E. Thompson, E. McAlpine,
Surf. Eng. 18 (2002) 228.
(
[
[
12] B. Adolphi, E. Jahne, G. Busch, X. Cai, Anal. Bioanal. Chem. 379
4
2 (2004) 445.
(2004) 646.