Surface-enhanced infrared absorption of DMIP on gold-germanium substrates coated by self-assembled monolayers

Article abstract of DOI:10.1016/j.molstruc.2003.07.022

Gold islands coated with self-assembled monolayers of three enough different thiols - thiophenol, propanethiol and octadecanethiol - are shown to enable the detection of molecules without polar groups, such as dimehylisophthalate (DMIP), using Surface enhanced infrared absorption (SEIRA). Binding of thiols was studied using SERS and SEIRA. The careful comparison of the transmission and p-polarized external reflection SEIRA spectra of DMIP taken using the modified substrates, with the corresponding spectra taken on the uncoated gold-Ge and bare Ge and CaF₂ surfaces, has allowed to elucidate the role played by the underlayer substrate and the different thiols in the total observed enhancement effect. Furthermore, the DMIP orientation with respect to the gold surface for each thiol-coating has also been deduced, confirming that the surface selection rule in our case applies both to reflection and transmission SEIRA spectra. No SERS spectra of DMIP, either on the modified or on the uncoated gold substrates, were revealed.

Full text of DOI:10.1016/j.molstruc.2003.07.022

Journal of Molecular Structure 661-662 (2003) 419–427
www.elsevier.com/locate/molstruc
Surface-enhanced infrared absorption of DMIP on
gold–germanium substrates coated by self-assembled
monolayers
a,
C. Domingo , J.V. Garcıa-Ramos , S. Sanchez-Cortes , J.A. Aznarez
a
a
b
*
´
´
´
aInstituto de Estructura de la Materia, CSIC, Serrano 123, 28006 Madrid, Spain
b
´
Instituto de Fısica Aplicada, CSIC, Serrano 144, 28006 Madrid, Spain
Received 2 May 2003; accepted 24 July 2003
Abstract
Gold islands coated with self-assembled monolayers of three enough different thiols—thiophenol, propanethiol and
octadecanethiol—are shown to enable the detection of molecules without polar groups, such as dimehylisophthalate (DMIP),
using Surface enhanced infrared absorption (SEIRA). Binding of thiols was studied using SERS and SEIRA. The careful
comparison of the transmission and p-polarized external reflection SEIRA spectra of DMIP taken using the modified substrates,
with the corresponding spectra taken on the uncoated gold–Ge and bare Ge and CaF2 surfaces, has allowed to elucidate the role
played by the underlayer substrate and the different thiols in the total observed enhancement effect. Furthermore, the DMIP
orientation with respect to the gold surface for each thiol-coating has also been deduced, confirming that the surface selection
rule in our case applies both to reflection and transmission SEIRA spectra. No SERS spectra of DMIP, either on the modified or
on the uncoated gold substrates, were revealed.
q 2003 Elsevier B.V. All rights reserved.
Keywords: Surface enhanced infrared absorption; Self-assembled monolayers; Surface-enhanced vibrational spectroscopy; Dimethylisophtha-
late; Gold substrates
1. Introduction
molecules have functional groups capable of
binding them to the metallic surfaces employed as
SERS/SEIRA substrates. This fact promotes
these spectroscopies as ‘promising sensor technol-
ogies’ [1,2]. Nevertheless there are many important
compounds, like some polyaromatic hydrocarbons
(PAHs) in which we are presently specially interested,
that approach weakly or not at all to the metallic
surfaces and, consequently, no enhancement effect is
reached. In order to extend the practical applications
of SEVS to such kind of compounds, the rough metal
Surface Enhanced Vibrational Spectroscopies
(SEVS)–Raman (SERS) and Infrared Absorption
(SEIRA)—have proved to be very useful techniques
for the detection and study of very low quantities of
chemical compounds, mainly when the corresponding
*
Corresponding author. Tel.: þ34-91-561-6800; fax: þ34-91-
5645557.
E-mail address: cdomingo@iem.cfmac.csic.es (C. Domingo).
0022-2860/$ - see front matter q 2003 Elsevier B.V. All rights reserved.
doi:10.1016/j.molstruc.2003.07.022

420
C. Domingo et al. / Journal of Molecular Structure 661-662 (2003) 419–427
surfaces can be coated with organic monolayers to
facilitate the approaching of the chemical to be
detected and/or studied. Sulfur and selenium com-
pounds have a strong affinity to transition metal
surfaces and, specifically, self-assembled monolayers
(SAMs) of thiols on gold and silver-which are exactly
the most extensively metals used in SEVS—have
been the subject of considerable research because
such molecules spontaneously chemisorb to form
densely packed and structurally ordered thin films [3].
Thus, SAMs may be applied or patterned at a surface
to control the properties and reactivities of the
monolayer/surface assembly. Hence, modified SEVS
substrates based on such SAMs have been proposed
and evaluated by other authors, in particular, coating
gold–silicon plates for SEIRA in external
reflection configuration [4] and coating of etched
silver foils [5] and silver film over nanospheres
(AgFON) [6] for SERS.
approach to metallic surfaces, like the PAHs whose
detection and study with SEVS we are interested in.
Besides, there are published data about the SEIRA
spectra of the same molecule in external reflection
configuration over a TP coated gold–Si substrate [4],
therefore, we will be able of comparing our results
with the reported ones. (It must be told here that a
second paper of the same authors [10] reports the
testing of SAMs of thiols as SEIRA substrates for
external reflection geometry of molecules having the
nitro group. Since the presence of this strongly
polarizable functional group increases the affinity to
the metal through a chemisorption [11], we do not
consider that their results can be extrapolated, without
a severe analysis, to molecules with no chemisorption
at all. Then we consider more reasonable to proceed
with DMIP in our comparative work.) Ge is here
preferred to CaF2 as underlayer infrared transparent
substrate, because of its broader spectral range
towards low wavenumbers, thus allowing to register
the very important out-of-plane modes in aromatic
systems, under 900 cm21, which supply relevant
information about the molecular orientation. More-
over, Ge is a commonly substrate in ATR configur-
ation and the conclusions from the present work can
be of utility for ATR measurements. At the same time
it is also interesting to test how this non-dielectric
substrate works for getting SERS spectra, since
usually only dielectric substrates like glass or quartz
are employed. Gold was chosen instead of silver
because it is not oxidized, thus increasing the active
life of the solid substrate. Finally the thiols selected
for this work are thought to have enough different
chemical character (aromatic TP, non-aromatic PPT
and ODT), geometry and chain length (C3 in PPT, C17
in ODT), as for allowing to expect from them different
behavior both in the enhancement, if any, and in
DMIP molecular orientation with respect to the
metallic surface. The spectra obtained over the
modified SAMs substrates have been compared with
the ones taken on uncoated gold–Ge, bare Ge and
bare CaF2 surfaces.
It is nowadays clear that the enhancement and
appearance of the SEV spectra depend greatly on the
morphology of the metal film (shape, size and density
of islands) which is mainly determined by the metal,
the substrate on which the metal is deposited, the
deposition rate and the thickness of the island film [2].
Additionally, the configuration selected for the
infrared measurements, either ATR, external reflec-
tion or transmission, affects also the final spectra. In
view of this casuistic, one can assert that it is
practically impossible to transfer in advance the
good properties of one particular SEVS substrate
tested with one spectroscopic technique and configur-
ation, to a different technique or configuration. In this
sense, studies evaluating the SERS and
SEIRA efficiencies of the same ‘active substrate’,
including different infrared configurations, would be
welcome [7].
In our previous studies, SERS and SEIRA of
a given molecule on Au islands have been confirmed
[8,9]. Here, using the short range property of the
electromagnetic enhancement, the enhanced intensi-
ties and molecular orientation of dimethylisophthalate
(DMIP) on coated Au islands with thiophenol (TP),
propanethiol (PPT) and octadecanethiol (ODT), are
reported. Au island films were evaporated over
germanium substrates.
2. Experimental
These are the reasons for the choices made. DMIP
has no reactive functional groups and does not
The evaporation of gold onto 25 £ 10 £ 2 mm3 Ge
windows supplied by Sorem, was made on a vacuum

C. Domingo et al. / Journal of Molecular Structure 661-662 (2003) 419–427
421
chamber held at a pressure of 1026 Torr. The gold
evaporation rate was maintained at 0.65 nm/min and
the film thickness was monitored using a quartz
crystal oscillator. Ten nanometers (mass thickness)
gold films were used, as similar thickness of silver
[12] and platinum [13] films have been reported to
give the greater enhancements and the less distorted
band shapes due to the morphology of the films where
metallic island were close spaced but not touching
each other. Our TEM micrographs showed gold
particles 10–40 nm width and 40–100 nm length.
Thiophenol (Fluka), Propanethiol and Octadeca-
nethiol (Aldrich) were diluted to 1 mg/ml in methanol,
5 mg/ml in ethanol and 0.5 mg/ml in ethanol, respect-
ively. The corresponding SAMs coatings were pre-
pared by soaking the gold–Ge plates in the mentioned
solutions for 5, 3 and 1 day, respectively,
and then rinsed with the solvent and allowed to air
dry before use. All the organic solvents were
purchased from Merck.
Reflectance accessory Specac SC19650 with a KRS5
polarizer was employed.
The Raman spectra were recorded on a Raman
microscope, Renishaw RM1000, with 785 nm laser
excitation.
It is interesting to mention here that even for the
high vacuum employed in the gold evaporation
process, traces of the vacuum pump oil are deposited
over the gold films, as we have confirmed by the
appearance of the corresponding bands in the infrared
‘enhanced’ spectra of the gold–Ge substrates. Taking
into account that the organic solvents employed for
the SAMs formation can eliminate some part of this
oil, special attention must be paid to the ‘background
sample’ in the SEIRA measurements of the SAMs.
Thus, a gold–Ge plate which had been soaked in the
same solvent and for the same period of time that
needed for the formation of the corresponding SAM,
was taken as background sample, so that the
possibility of misinterpreting oil bands as due to
thiols is ruled out. Despite the care we have taken in
the manipulation of the samples, we have realized that
our coated SAMs surfaces present traces of hydro-
carbons contamination, which manifest in their
corresponding ‘enhanced’ spectra through strong
characteristic aliphatic CH stretching bands in the
2800–3000 cm21 region, and less intense CH bend-
Dimethylisophthalate (Fig. 1) was synthesized at
´
´
Instituto de Ciencia y Tecnologıa de Polımeros of the
CSIC, by reaction of isophthaloylchloride with
methanol, and afterwards dissolved in cyclohexane
(0.5 mg/ml). An aliquot, 10 ml, of the solution was
deposited onto the bare Ge, gold–Ge and thiols-
coated gold–Ge plates, allowing the solvent to
evaporate under cover. A cast film of DMIP on a
bare CaF2 substrate was also formed in order to
consider its infrared transmission spectrum as refer-
ence for determining the enhancement factors of the
other surfaces.
ings bands around 1470 and 1380 cm21
.
3. Results and discussion
3.1. SEV spectroscopy of SAMs
The SEIRA spectra were measured on a FTIR
Bruker IFS66 spectrometer in transmission and
polarized external reflection configuration, using a
DTGS detector, 8 cm21 resolution and coadding 100
scans. For the polarized reflection measurements
made at 768 (Brewster angle of Ge), a variable angle
The binding of thiols to the gold–Ge surface was
clearly evidenced for TP in infrared and for PPT in
Raman (Fig. 2). On its turn, ODT gave a very poor
SERS spectrum. This fact is consistent with the
reported SERS of alkanethiols on Ag, where PPT is
told to provide very intense signals in comparison to
the other thiols [14].
Fig. 2 (top) shows the transmission SEIRA
spectrum of thiophenol SAM on gold–Ge substrate
compared with the infrared spectrum of neat TP. The
absence in the SEIRA spectrum of the bands
associated to the S-H bond (stretching mode at 2565
and bending at around 915 cm21) clearly indicate that
TP is adsorbed on the gold surface as an Au
Fig. 1. DMIP molecule.

422
C. Domingo et al. / Journal of Molecular Structure 661-662 (2003) 419–427
the molecular system from the surface normal Z is
found to be 40 ^ 108, from the intensities ratio of the
a1 modes at 692, 1472 and 1574 cm21 to the
mentioned b1 mode and considering the x value
deduced above. Our molecular orientation differs
from the reported on crystalline gold [15] in the twist
angle (x < 08 there, as indicated by the negligible
intensity of the b2 modes) while the tilt angle agrees
ðu < 308Þ: The different intensity pattern of the b1 and
b2 modes in our spectra and in Ref. [15] led to foresee
such disagreement in the twist angle.
Lower part of Fig. 2 shows the SERS spectrum of
the SAM of propanethiol. The binding of PPT to the
gold surface through the S atom is manifested by
the absence of the S–H stretching band (not shown in
the figure) and the C–S stretching bands of the gauche
and trans conformers at 633 and 707 cm21, respect-
ively. This spectrum matches the SERS of PPT on
electrochemically slightly roughened Au surfaces
[17], both wavenumbers and intensities, and the
corresponding SERS spectrum on an acid etched
silver surface [18], with the C–S bands on our gold
surface shift þ9 and þ5 cm21 respect to the
corresponding ones on Ag. The relative intensities of
these two bands and the large intensities of the C–C
stretching band of the trans conformer at 1027 cm21
allow to assert that the conformation of the C–C bond
adjacent to the C–S bond of the adsorbed molecule is
mostly trans [19].
Fig. 2. Upper part: (a) transmission spectrum of neat thiophenol; (b)
transmission SEIRA spectrum of a SAM of Thiophenol on a gold–
Ge substrate. Lower part: SERS spectrum of a SAM of propanethiol
on a gold–Ge substrate, excited at 785 nm.
For both, PPT and ODT SAMs, it is assumed that
the tilt angle of the chain axis from the surface normal
is on the order of 25–408, as reported for alkanethiol
monolayers on polycristalline gold [20]. This means
that all our SAMs are approximately tilted a similar
angle from the gold surface, a fact that is interesting to
consider when establishing the DMIP molecular
orientation onto the different SAM-gold–Ge
substrates.
benzenethiolate. The molecular orientation of TP in
our gold surface can be determined from the relative
intensities of the respective a1ðzÞ; b1ðxÞ and b2ðyÞ
modes, where ðxyzÞ is the molecular system, in KBr
pellet and in the SEIRA spectra. For such determi-
nation, it is necessary to apply the surface selection
rule formulated and employed for TP adsorbed on
crystalline gold (111) using ATR configuration (see
Eqs. 5 and 6 of Ref. [15]), and to consider that such
rule is also applicable to the transmission SEIRA
spectrum [16].
3.2. SEV spectroscopy of DMIP
3.2.1. SEIRA spectroscopy
The reference transmission FT-IR spectrum of
DMIP dispersed in a KBr pellet is shown in Fig. 3a
Spectra b–d of the same Figure show the transmission
infrared spectrum of 6 mg/cm2 of DMIP deposited on
bare Ge, and the corresponding transmission SEIRA
spectra using Au–Ge and thiophenol-coated Au–Ge
In our case, the twist angle x of the molecular plane
around the z molecular axis is found to be 45 ^ 108, as
deduced from the intensities of the out-of plane b1
mode at 735 cm21 and the b2 mode at 1441 cm21
.
Furthermore, the tilt angle u of the z axis of

C. Domingo et al. / Journal of Molecular Structure 661-662 (2003) 419–427
423
1300 cm21, involving also nCC and ns(OCO), and at
725 cm21 due to the out-of-plane gCH wagging. Low
intensity bands appear in the 1000–1200 cm21
region, characteristic of aromatic ring modes, and at
984 and 864 cm21 which must have participation of
gCH too [21]. Extremely weak bands in the not shown
3000 cm21 region, can be assigned to the aromatic
CH stretchings (3014 and 3080 cm21) and to aliphatic
CH3 stretchings (2960 and 2850 cm21).
Within experimental accuracy, the peak positions
in all the recorded spectra were similar to those
observed in the spectrum of the solid in KBr pellet,
except for some bands sensitive to the molecular
orientation that will be discussed below and whose
frequency shifts were in any case lower than 10 cm21
.
This can be taken as evidence that no chemisorption
occurs which would give rise to greater frequency
shifts.
The first thing to comment is that bare Ge seems to
give an ‘enhanced’ transmission spectrum when
compared with the spectrum on CaF2. It must be
recall here that infrared transmission spectra have a
contribution of the refractive index of the supporting
layer [22]. In this sense, CaF2 has a low refractive
index, n ¼ 1:5; while germanium is a highly refrac-
tive, n ¼ 4; and therefore, highly reflecting layer that
promotes multi-beam interference and in this way the
effective length of interaction between infrared
radiation and adsorbate increases. This must be the
reason of the ‘enhancement’ in bare Ge, as it is the
reason of employing such material in ATR
accessories.
Fig. 3. Upper part: FTIR transmission spectrum of DMIP on: (a)
KBr pellet; transmission SEIRA spectra of DMIP on (b) bare Ge, (c)
gold–Ge and (d) TP-gold–Ge. Lower part: reflection SEIRA
spectra of DMIP at 768: (e) p-polarized on TP-gold–Ge, (f) p-
polarized on gold–Ge and (g) s-polarized on TP-gold–Ge. Spectra
(b) to (g) correspond to 6 mg/cm2 of DMIP.
On going to the transmission spectrum on gold–
Ge, an additional enhancement factor of 3 is found.
Probably this factor must be mostly attributed to the
increase of refractive index of the gold island film
with respect to the n value of Ge. We base our
conclusion in the published data about optical
constants of thin gold islands films deposited on
dielectric substrates [22,23]. Such effective values
have been determined by Infrared Ellipsometry and
simulated trough Effective-Medium Theories. The
reported results show that the metallic island films are
characterized by a unique combination of high
refractive index, n < 7; and low absorption index,
k < 1; for a wide infrared spectral range. We assume
that our gold film deposited on a semiconductor
substrates, respectively. At the same Figure, spectra
(e)–(g) present the corresponding external reflection
SEIRA spectra measured at an incident angle of 768,
for p- and s-polarized incident beam in the TP-coated
case and p-polarized in the uncoated one. It must be
told that the infrared transmission spectrum of the
same aliquot deposited on a bare CaF2 plate (not
shown in the figure) is 5 times weaker than the
corresponding one when deposited on bare Ge plate.
The strongest bands in the spectra of the solid
DMIP correspond to the carbonyl stretching CyO at
1728 cm21 and to the CO stretch at 1248 cm21. Other
intense bands are centered at 1440 cm21, due to
aromatic CC stretchings and CH3 deformations, at

424
C. Domingo et al. / Journal of Molecular Structure 661-662 (2003) 419–427
substrate, Ge, has similar characteristics than the
mentioned above.
the gold-Si surface, according to the surface selection
rules [16]. As mentioned in Section 1, SEV spectra are
mainly dependent on the morphology of the islands
films, which changes with the underlayer substrate,
the metal deposition conditions and the film thickness.
The same thickness was employed in Ref. [4] and
here, but the other metal deposition conditions are
different: (1) Si substrate there and Ge here and (2)
sputtering gold deposition there and vacuum gold
evaporation here, with deposition rates 2:1, respect-
ively. The combined effect of these differences
probably originates distinct gold particle size and
shape in both experiments, which reflects in the
different enhancement factors and molecular orien-
tations. Unfortunately there is no information in Ref.
[4] neither about the morphology of the gold film nor
about the orientation of the TP monolayer.
No additional enhancement factor is found in the
SEIRA transmission spectrum when the gold–Ge film
is coated with Thiophenol. In fact, the total enhance-
ment factor with respect to CaF2 substrate is,
generally speaking, a 10% lower than the resulting
15 (3 £ 5) for gold–Ge film.
On the contrary, the expected enhancement effect
due to the thiol over the gold–Ge surface, is clearly
manifested in the external reflection SEIRA spectrum
of DMIP, where a 3-fold enhancement is found for the
p-polarized component. Besides, this spectrum is
around three times more intense than the transmission
one, as can be seen in Fig. 3, except for the out-of-
plane gCH mode around 730 cm21. This is consistent
with the reported fact that the infrared reflection-
absorption spectroscopy has a sensitivity up to 10
times higher than transmission spectroscopy [24].
From the comparison of the p-polarized SEIRA
spectra of DMIP on gold–Ge surfaces coated with the
three thiols employed in this work, it is concluded that
PPT provides similar enhancement than TP while the
enhancement due to ODT coating is 50% lower than
for TP and PPT coatings. This fact will be taken into
account below, when deducing the DMIP orientation.
It is worthy to mention here the results obtained for
gold–Si surfaces coated with TP, PPT and thiohexane
(TH), and referred to SEIRA spectra of molecules
with the nitro group [10]. There, it was reported that:
(i) TP gave the largest enhancement, (ii) TH gave no
additional enhancement over uncoated gold surface
and (iii) PPT was only slightly better ‘enhancer’ than
TH. In our case, for a molecule like DMIP without
polar groups, the results are a little bit different, as it
has been explained above.
In addition to report the ‘enhancer’ capabilities of
every layer and thiol in the modified substrates and to
verify the superior sensitivity of the reflection
configuration over the transmission one, we have
also investigated the different orientation adopted
by the DMIP molecule on the diverse thiol-coated
gold–Ge substrates. For this purpose, Fig. 4 displays
the corresponding transmission (b–d) and p-polarized
reflection SEIRA spectra at 768 (e–g), all of them
normalized to the intensity of the aromatic ring modes
(1000–1100 cm21 region) which maintain their
positions and intensities as measured in the spectrum
of the solid dispersed in KBr designed as (a).
Once we know how thiols are oriented relative to
the gold surface, (Section 3.1) the question is how
DMIP is actually placed ‘in close proximity’ to the
metallic surface when the SAMs are present: (i) on top
of the thiol monolayer, or (ii) intercalated between the
assembled layers [6a].
On the other side, when our p-polarized SEIRA
spectrum of DMIP on TP-coated gold–Ge surface is
compared with the previously reported on TP-coated
gold–silicon substrate [4] and after taking into
consideration the DMIP quantities in both cases
(6 mg/cm2 here and 75 mg/cm2 there), an increase of
three times in sensitivity is detected for our spectrum.
Besides, the intensity patterns are different, thus the
orientation effects of both TP-coated surfaces are
dissimilar. Specifically, in the work yet published [4]
the out-of-plane modes are the most enhanced
ones, hence the DMIP molecule must be parallel to
It is well known that for ordered films, like the
Langmuir–Blodgett (LB) films [25], normal inci-
dence IR transmission spectra display bands due to
vibrational transition moments projected into direc-
tions within the surface plane, while reflection spectra
on metallic substrates exhibit only bands due to
transition moments projected perpendicular to the
surface, according to the surface selection rule. Our
cast DMIP films have a short range order, then the
above assertion would be fulfilled partially. This
means that we would expect enhancement of bands in
transmission (reflection) SEIRA spectra as far as

C. Domingo et al. / Journal of Molecular Structure 661-662 (2003) 419–427
425
spectra of ODT coated substrates, while for the other
substrates this band decreases (enhances) correspond-
ingly. This fact lead us to conclude that the aromatic
ring of DMIP lies rather parallel to the gold surface
when the metallic surface is ODT coated. For the TP
coating, the situation is the inverse; this means that the
most tilted DMIP orientation takes place when the
gold surface is coated with TP. For PPT coating, the
situation is in between the other two thiols. The
analysis of these data, together with the observed
enhancements reported above, allow to presume that:
(I) DMIP is intercalated between the thiophenols
layers; (II) DMIP lies on top of the ODT SAMs and
(III) a probable mixture of both situations occurs in
the PPT case. Due to the different length of the
aliphatic chains in PPT and ODT, the DMIP
molecules lying on top are closer to the gold surface
in the case of PPT coating and, consequently, the
corresponding enhancement factor is larger than for
ODT coating. Apparently, the reflection spectrum of
DMIP on gold–Ge substrate keeps the relative
intensities of the solid, for the strongest bands,
although the transmission one reflects an special
enhancement of the gCH band. This would lead to
think that, without thiol coating, the aromatic ring
adopts an orientation more perpendicular to the
metallic surface. Other bands involving gCH modes
in the 1000–800 cm21 region, behave like the one at
Fig. 4. Upper part: FTIR transmission spectrum of DMIP on (a) KBr
pellet; transmission SEIRA spectra of DMIP on gold–Ge surfaces
coated with (b) TP, (c) PPT and (d) ODT SAMs. Lower part: p-
polarized Reflection (768) SEIRA spectra of DMIP on gold–Ge
surfaces coated with (e) TP, (f) PPT and (g) ODT SAMs. All the
spectra are normalized to the intensity of the 1000–1100 cm21
region. (þ) bad compensate water band. ( p ) impurity of aliphatic
hydrocarbons.
725 cm21
.
The C–O stretch, at 1248 cm21 in the solid
spectrum, shifts towards low wavenumbers in all the
cases, except on bare Ge, where it also experiments
the largest enhancement in transmission mode. Apart
from this case, this band usually enhances except in
reflection spectra on PPT and ODT coatings and
transmission spectrum on ODT. The main conclusion
is that C–O is more perpendicular to the metallic
surface when DMIP is deposited on bare Ge, and that
in other substrates the C–O must adopt a tilted
configuration.
the corresponding vibrational transition modes have a
predominant parallel (perpendicular) component to
the gold surface. In this context we will discuss the
spectra given in Fig. 4, with reference to the spectra of
the solid dispersed in KBr where random orientation
is assumed.
The most clearly sensitive band, both in intensity
and frequency shift, has proved to be the one due to
the out-of-plane gCH mode, at 725 cm21 in the solid.
It maintains its frequency in the cast films on Ge
(Fig. 3b), gold–Ge (Fig. 3c) and TP coated substrates
(Fig. 4b and e), while shifts towards higher wave-
numbers with PPT and ODT coatings, both in
transmission (Fig. 4c and d) and reflection configur-
ations (Fig. 4f and g). At the same time, it is specially
enhanced (decreased) in the reflection (transmission)
On its turn, the CyO stretching, at 1728 cm21 in
the solid spectrum, decreases its intensity in all cases.
Slight shifts downwards (TP coating) and upwards
(PPT and ODT coatings) are observed.
It is interesting to note that the bands involving
ns(OCO), around 1300 cm21 undergo always a down-
wards frequency shift, like the C–O stretch, with
a change of the intensity pattern relative to

426
C. Domingo et al. / Journal of Molecular Structure 661-662 (2003) 419–427
the spectrum of the solid when PPT and ODT coatings
are used.
thiols-coated substrates have been recorded and
investigated. The p-polarized SEIRA spectrum of
DMIP on Thiophenol-coated gold–Ge, taken in the
reflection mode, presents an enhancement factor of
3 with respect to the p-polarized one on gold–Ge,
a 3-fold enhancement with respect to the corre-
sponding SEIRA spectrum taken in the trans-
mission mode and a total 40-fold enhancement
with respect to the transmission spectrum taken on
bare CaF2. Thiol-coatings provide no enhancement
over uncoated surface for SEIRA spectrum taken in
transmission mode.
Finally, the 1440 cm21 bands, due to aromatic CC
stretchings and CH3 deformations, change also the
pattern of intensities, with enhancement of the lower
wavenumber band at 1426 cm21. This result can be
again attributed to the different orientation adopted by
these groups on the SAMs films.
3.2.2. SERS spectroscopy
We tried to get SERS spectra of DMIP on our solid
gold substrates, both SAMs coated and uncoated,
without any success. Taking into account that it has
been reported hat other phthalic derivatives give
SERS spectrum on silver sols but not on gold sols [26]
we also tried to get SERS spectra of DMIP on silver
colloids, in an attempt to discard that the metal (gold
instead of silver) was not adequate for our case. The
results were unsuccessful too. On the contrary, we got
very good SEIRA spectra of DMIP, as it has been
presented and discussed above. From our SEIRA data
it is deduced that DMIP does not chemisorb neither on
the coated nor on the uncoated gold surfaces. On the
other side, DMIP approach slightly to the gold (and
probably silver) surfaces. As a result, the molecule
does not experiment sufficiently the ‘enhanced
electromagnetic field on the metallic surface’ as for
giving Raman enhanced spectra but the same
electromagnetic field is enough for getting infrared
enhanced spectra. This result indicates once more [8]
that, despite being the SEIRA enhancement only 1–2
orders of magnitude while the SERS one can reach 6
orders of magnitude or more, SEIRA spectroscopy is
found to be a surface enhanced technique with higher
sensitivity than SERS spectroscopy.
For the aliphatic thiols tested with DMIP, it was
found that PPT provides similar enhancement than
TP on gold–Ge substrates, while the enhancement
due to ODT coating is 50% lower than for TP and
PPT coatings. The most remarkable effect of the
different thiols, refers to the different DMIP
location, as follows: (I) intercalated between the
monolayers of TP-aromatic ring of DMIP tilted
with respect to the gold surface-favored by P–P
stacking; (II) adsorbed on top of the monolayers of
ODT-aromatic ring of DMIP parallel to the gold
surface—as prompted by the hydrophobic repulsion
between the aliphatic chains and the DMIP
molecule, and (III) a mixture of both situations in
the PPT case with probable majority of on top
adsorbed molecules. Due to the different length of
the aliphatic chains in PPT and ODT, DMIP is
closer to the gold surface in the case of PPT
coating and, consequently, the corresponding
enhancement factor is larger than for ODT coating.
These orientations are mainly deduced from the
intensities of the out-of-plane ring modes gCH at
around 725 and 984 cm21 in the reflection SEIRA
spectra, according to the surface selection rule [16],
but noticing also that similar conclusions can be
taken from the frequency shifts and changes in the
relative intensities of the same vibrational modes as
measured in the transmission SEIRA ones.
These facts show that for our modified SAMs
gold–Ge substrates, the surface selection rule is
also applicable to the transmission SEIRA spectra,
as it was recently confirmed for platinum island
films [13].
4. Conclusions
Self-assembled monolayers of thiophenol, propa-
nethiol and octadecanethiol have been formed on
gold–Ge SEV substrates. Binding of TP and PPT to
the metallic surface are evidenced by the correspond-
ing SEIRA and SERS spectra, respectively. Their
molecular orientation with respect to the metallic
surface have been deduced.
The location of the DMIP molecules deduced
from SEIRA spectra gives an explanation to the fact
that no SERS spectra were obtained: first, there is
The SEIRA transmission and reflection spectra
at 768 of dimehylisophthalate on the uncoated and

C. Domingo et al. / Journal of Molecular Structure 661-662 (2003) 419–427
427
[6] (a) P. Freunscht, R.P. Van Duyne, S. Schneider, Chem. Phys.
Lett. 281 (1998) 382. (b) K.E. Shafer-Peltier, C.L. Haynes,
M.R. Glucksberg, R.P. Van Duyne, J. Am. Chem.Soc. 125
(2003) 588.
no chemisorption at the gold surface and second, the
DMIP molecules lay at a distance from the metallic
surface where the enhanced electromagnetic field
has decayed so much as for producing enhanced
Raman signals. On the other side, our results have
proved that the non-dielectric Ge material is a good
underlayer substrate for SERS too, as far as the
molecule under study present some chemisorption to
the metallic overlayer surface, like the SAMs of
propanethiol do.
[7] E. Johnson, R. Aroca, J. Phys. Chem. 99 (1995) 9325.
´ ´ ´
[8] S. Sanchez-Cortes, C. Domingo, J.V. Garcıa-Ramos, J.A.
´
Aznarez, Langmuir 18 (2001) 1158.
´ ´ ´
[9] C. Domingo, J.V. Garcıa-Ramos, S. Sanchez-Cortes, J.I.
´ ´
Millan, J.A. Aznarez, Proceedings of the XVIII ICORS,
Wiley, New York, 2002, p. 295.
[10] J.A. Seelenbinder, C.W. Brown, Appl. Spectrosc. 56 (2002)
295.
[11] G.T. Merklin, P.R. Griffiths, Langmuir 13 (1997) 6159.
[12] M. Osawa, M. Ikeda, J. Phys. Chem. 95 (1991) 9914.
[13] A.E. Bjerke, P.R. Griffiths, Appl. Spectrosc. 56 (2002) 1285.
[14] B.J. Kennedy, S. Spaeth, M. Dickey, K.T. Carron, J. Phys.
Chem. B 103 (1999) 3640.
Acknowledgements
´
The Ministerio de Ciencia y Tecnologıa, MCYT,
[15] L.-J. Wan, M. Terashima, H. Noda, M. Osawa, J. Phys. Chem.
B 104 (2000) 3563.
of Spain is gratefully acknowledged for financial
support under Projects BFM2001-2265 and
REN2000-1557-CLIM. J.G. de la Campa and A.E.
[16] M. Osawa, K. Ataka, K. Yoshii, Y. Nishikawa, Appl.
Spectrosc. 48 (1993) 1498.
[17] M.H. Schoenfisch, J.E. Pemberton, J. Am. Chem. Soc. 120
(1998) 4502.
´
Lozano, of the Instituto de Ciencia y Tecnologıa de
´
Polımeros, CSIC, are also acknowledged for synthe-
sizing the DMIP compound.
[18] B.J. Kennedy, R. Milofsky, K.T. Carron, Anal. Chem. 69
(1998) 4808.
[19] M. Bryant, J.E. Pemberton, J. Am. Chem. Soc. 113 (1991)
3629.
[20] M.D. Porter, T.B. Bright, D.L. Allara, C.E.D. Chidsey, J. Am.
Chem. Soc. 109 (1987) 3559.
References
[21] J.F. Arenas, J.I. Marcos, Spectrochim. Acta 35 (1989) 355.
[1] W.E. Smith, C. Rodger, in: J.M. Chalmers, P.R. Griffiths
(Eds.), Handbook of Vibrational Spectroscopy, vol. 1, Wiley,
New York, 2002, p. 785.
¨
[22] A. Roseler, E.-H. Korte, Thin Solid Films 313–314 (1998)
832.
¨
[23] A. Roseler, E.-H. Korte, Appl. Spectrosc. 51 (1998) 902.
[2] M. Osawa, in: J.M. Chalmers, P.R. Griffiths (Eds.), Handbook
of Vibrational Spectroscopy, vol. 1, Wiley, New York, 2002,
p. 885.
[24] F.M. Hoffmann, Surf. Sci. Rep. 3 (1983) 107.
¨
[25] D. Tsankov, K. Hinrichs, E.-H. Korte, R. Dietel, A. Roseler,
Langmuir 18 (2002) 6559.
[3] A. Ulman, Chem. Rev. 96 (1996) 1533.
[4] J.A. Seelenbinder, C.W. Brown, D.W. Urish, Appl. Spectrosc.
54 (2000) 366.
[26] O. Klug, Gy. Parlagh, W. Forsling, in: Z. Horvoelgyi, Zs.
Nemeth, I. Paszli (Eds.), Proceedings of the 8th Conference
Colloid Chemistry, Hungarian Chemical Society, Budapest,
1998, pp. 302–305.
[5] C.E. Taylor, M.H. Schoenfisch, J.E. Pemberton, Langmuir 16
(2000) 2902.