www.elsevier.com/locate/apsusc

Applied Surface Science 252 (2005) 1941–1946

Characterisation and stability of hydrophobic surfaces in water

M. Maccarini, M. Himmelhaus *, S. Stoycheva, M. Grunze

Angewandte Physikalische Chemie, Ruprecht Karls Universitat Heidelberg, Im Neunheimer Feld 253, 69120 Heidelberg, Germany

Received 1 February 2005; received in revised form 20 March 2005; accepted 20 March 2005

Available online 18 April 2005

Abstract

The stability of four different hydrophobic surfaces in contact with water is assessed and discussed: H-terminated silicon,

hexamethyldisilazane (HMDS) coated silicon, silicon surfaces covered with self-assembled monolayers (SAMs) of octadecyl-

trichlorosilane (OTS) and gold surfaces modified with SAMs of alkanethiols. Changes in hydrophobicity and surface oxidation

were determined by contact angle measurements, X-ray photoelectron spectroscopy and AFM.

# 2005 Elsevier B.V. All rights reserved.

Keywords: Hydrophobic surfaces; Stability; Surface modification; Self-assembling monolayers

1. Introduction

The preparation of homogeneous and molecularly

flat stable hydrophobic surfaces plays an important

role in technological applications, but is also of

interest in studies of the dynamical and structural

aspects of the solid/liquid interface. The interaction

between solid surfaces separated by a thin layer of

water is a central issue in many technological fields

such as the study of friction and wear, and lubrication

in an humid environment. Many experimental efforts

have been made to measure the forces between

hydrophobic surfaces in water by surface force

apparatus and atomic force microscopes [1]. One

* Corresponding author. Tel.: +49 6221 545065;

fax: +49 6221 545060.

E-mail address: michael.himmelhaus@urz.uni-heidelberg.de

(M. Himmelhaus).

0169-4332/$ – see front matter # 2005 Elsevier B.V. All rights reserved

doi:10.1016/j.apsusc.2005.03.157

important conclusion derived in review by Christen-

son and Claesson [1] is that a number of controversial

results have been generated due to changes in the

surfaces properties during the experiments and due to

a lack of characterisation before, during, and after the

experiments. Indeed, it was found that surface forces

between more robust and stable hydrophobic surfaces

had a significant shorter range with respect to less

stable surfaces [2,3].

Interfaces between water and hydrophobic surfaces

have been widely studied in the last decades also in

view of the fundamental role in molecular self-

assembly processes governing the behaviour of

biological membranes, micellar formation, and the

stability of supramolecular structures. The properties

of a liquid in the bulk differs quite markedly from the

properties of the same liquid in the vicinity of a solid

surface. The attractive/repulsive interactions between

water molecules and the wall as well as the

.

M. Maccarini et al. / Applied Surface Science 252 (2005) 1941–19461942

geometrical constraints induce structural rearrange-

ments that extend over several molecular layers in

the bulk liquid. These structural changes are

reflected in the hydration forces between the two

surfaces immersed in the fluid. Among the few

methods available to probe experimentally such a

thin interfacial liquid layer are neutron and X-ray

reflectometry. However, both methods require long

acquisition times, up to several hours in the case of

neutron reflectometry. During this time, the surface

must be stable, otherwise the results are question-

able. Furthermore, surface roughness tends to

increase the noise in the reflectivity curves, thereby

limiting the observable range of momentum transfer

[4] and consequently the resolution of the experi-

ment [5]. It is therefore advisable to perform

reflectivity measurements with surfaces as flat as

possible. In addition, theoretical studies showed that

hydrophobic hydration depends significantly on the

radius of curvature of the hydrophobic body [6,7].

While water molecules can rearrange around small

apolar molecules in order to maintain the network of

hydrogen bonds, large objects tend to disrupt this

network, thus affecting the density of water

molecules in their vicinity. For flat surfaces, this

means that the character of surface roughness in

terms of corrugation and waviness will affect the

hydration at the water/solid interface and care must

be taken when results of different experiments are

discussed and compared with each other.

In this work, four commonly used hydrophobic

surfaces have been characterised in terms of their

hydrophobicity, stability in water and surface rough-

ness by means of contact angle measurements, X-ray

photoelectron spectroscopy (XPS), and atomic force

microscopy, thus exploring their potential use for the

study of interfacial water layers.

2. Experimental

2.1. Chemicals and materials

Dodecanethiol (DDT, 98%+), hexamethyldisila-

zane (HMDS), octadecyltrichlorosilane (OTS, 90%+)

were purchased from Sigma–Aldrich and used without

further purification. Also ethanol, methanol, chloro-

form, dimethylformamide (DMF), hydrogen peroxide

(H2O2), concentrated sulfuric acid (all analytical

grade) were used as received. Toluene absolute (H2O

< 0.005%) was purchased from Fluka. The water

used was Baker HPLC analysed purchased from J.T.

Baker. (1 0 0)-Silicon substrates were purchased from

Crystec GmbH, Germany.

2.2. Surface modification

2.2.1. Silicon oxide hydrogenised by HF etching

New (1 0 0)-silicon wafers were first immersed in

piranha solution (1/3 hydrogen peroxide + 2/3

sulfuric acid) for 10 min, and then immersed in

0.5% hydrofluoric acid for 10 min in order to remove

the oxidation layer and replace it with a hydrogen

termination. The HF etching was performed in pre-

cleaned plastic containers (sonication in a 3%

aqueous solution of Decon 90 detergent, and rinsed

with copious amount of Baker HPLC analysed

water) since glass-ware is attacked by HF and could

not be used. The samples were blown dry and stored

under nitrogen for a maximum time of few hours

before exposure to water. The measurement of

contact angle and XPS were performed immediately

after the samples were removed from water and

blown dry with nitrogen.

2.2.2. Silicon oxide modified with

hexamethyldisilazane

New (1 0 0)-silicon wafers were precleaned in

piranha solution for 10 min, and sonicated with

methanol, mixtures of methanol and chloroform and

pure chloroform. Subsequently, they were placed

under nitrogen and HMDS vapour overnight and

cleaned in an ultrasonic bath in mixtures of chloro-

form and methanol.

2.2.3. Self-assembled monolayer of

octadecyltrichlorosilane on silicon

New Si wafers (1 0 0) were immersed for 30 min

in piranha solution, then rinsed with millipore water

and ethanol. Subsequently, they were placed for

30 min in vacuum. Then the incubating solution,

prepared with OTS in toluene at a concentration of

25 mM, was introduced under nitrogen overpressure,

and left for 5 days. The samples were then placed in an

ultrasonic bath of toluene and ethanol and blown dry

with and stored under nitrogen.

M. Maccarini et al. / Applied Surface Science 252 (2005) 1941–1946 1943

Fig. 1. The advancing contact angles of Si(1 0 0)–H (full circles);

SiO2–HMDS (empty circles); OTS terminated SiO2 (empty squares)

and SAM of dodecanethiole on gold (triangles) measured as a

function of the immersion time in water. The initial values of the

dry samples are denoted with the dashed line for the dodecanethiol

SAM, dotted line for the OTS, and with the solid line for the

Si(1 0 0)–H and SiO2–HMDS.

2.2.4. Self-assembled monolayers of dodecanethiol

on gold

Thin films of polycrystalline gold were prepared by

thermal evaporation of 2 nm of titanium as adhesion

promoter and subsequent deposition of 50 nm of gold

of 99.99% purity onto polished single-crystal silicon

wafers (Crystec). Evaporation was performed at a

pressure of 2� 10�7 Torr and a deposition rate of

0.5 nm/s. The freshly evaporated gold substrates were

sonicated in dimethylformamide for 5 min prior to

UV cleaning with a UV lamp for 1 h. Immediately

after the UV irradiation, the substrates was immersed

in ethanol for at least 10 min in order to reduce the

oxidation layer which might result in the formation of

isolated multilayer islands as reported by Woodward

et al. [8]. The gold substrates were then immersed in a

1 mM solution of dodecanethiol in ethanol overnight.

2.3. Measurements

The advancing contact angles were measured with

a Kruss G1 goniometer using purified deionized water

(Milli-Q plus system, Millipore, Eschborn, Germany).

The AFM measurements were performed with a

Park Autoprobe (Veeco Instruments) operated in

contact mode and equipped with a 5mm� 5mm

scanner. Roughness values were calculated as rms

deviations from the mean using built-in software.

X-ray photoelectron spectra were obtained with a

Leybold Max 200 Analysis System using a Al K a

source operated at 200 W and a hemispherical energy

analyzer. The pass energy of the analyzer was set to

96 eV for survey scans and to 48 eV for recording

specific core levels. The electron binding energies

were referenced to the Si 2p peak at 99.15 eV for the

silicon substrates and to the Au 4f7/2 peak at 84.0 eV in

case of the gold-coated substrates.

3. Results and discussion

The advancing contact angles of the four hydro-

phobic surfaces as a function of the immersion time in

water are presented in Fig. 1. The contact angles of the

substrates modified with self-assembled monolayers

are 107� for dodecanethiol and 110� for OTS,

respectively. They prove to be very stable since they

are substantially unaffected by the exposure to water

even for long periods of time. The surfaces coated with

HMDS have a contact angle of 90�. This value remains

stable in contact with water for about 100 min, and

then slightly decreases to 80�. Hydrogen terminated

silicon on the contrary is not stable in contact with

water. The contact angle of the freshly prepared

surface is around 90�, but decreases significantly evenafter few minutes of exposure to water.

In order to track the changes in the properties of our

films, we performed XPS measurements on the

samples as a function of the contact time with water.

In Fig. 2a, we report the XPS survey spectra of

Si(1 0 0)–H before immersion in water, after 5 h in

water and, for comparison, also an XPS spectrum of

the native oxide on Si(1 0 0). In all spectra, the Si 2p

and 2s peaks are clearly discernible. The presence of

an oxygen peak can be observed in the SiO2 sample as

well as in the Si(1 0 0)–H sample which had been

immersed in water.

For the latter, XPS detail spectra of the Si 2p region

are depicted in Fig. 3 for different immersion times of

the samples in water. The energy scale was calibrated

by using the literature value of the Si 2p bulk peak at

99.15 eV. A second peak corresponding to SiO2 is

present at 102.9 eV. As expected, the SiO2 peak is

absent in the spectrum of the freshly prepared sample

due to the HF-etching of the native oxide layer.

However, an oxygen XPS peak is observed for the

M. Maccarini et al. / Applied Surface Science 252 (2005) 1941–19461944

Fig. 2. (a) XPS survey spectra of H–Si(1 0 0), of H–Si(1 0 0) after

5 h immersed in water. For comparison the spectra of native oxide

on Si(1 0 0) are shown. (b) XPS survey spectra of HMDS–SiO2 as

prepared, and after 24 h immersed in water. The spectra are verti-

cally displaced for clarity.

Fig. 3. Si 2p narrow XPS spectra of H–Si(1 0 0) at different

immersion times in water.

Si(1 0 0)–H sample before exposure to water, together

with a carbon peak, yet the intensity of the SiO2

emission is zero. We cannot exclude a residual

oxidation layer, since the sensitivity factor of the O1s

peak is almost four times larger than the one of the Si

2p peak. Hence, the presence of a small amount of

silicon oxide could generate a peak in the O1s region

without a corresponding feature in the Si 2p region.

However, due to the presence of a visible C1s peak

(with a sensitivity factor very close to that of silicon),

it is plausible that unavoidable contamination of the

plastic containers used in the HF etching stage, which

cannot be cleaned with the aggressive methods used

for glass-ware (e.g. piranha solution), are the most

likely origin of the O1s and C1s peaks. The area of the

SiO2 band increases with immersion time in water, and

correspondingly the area of the Si 2p bulk peak

decreases due to the oxide overlayer growth.

Obviously, the Si(1 0 0)–H surfaces are oxidized in

Fig. 4. The normalized area of the SiO2 and O XPS peaks on the

Si(1 0 0)–H and on the SiO2–HMDS, vs. contact time with water.

M. Maccarini et al. / Applied Surface Science 252 (2005) 1941–1946 1945

Table 1

XPS peaks of OTS on SiO2 and DDT on Au

OTS on SiO2 DDT on Au

Binding energy (eV) Peak assignation Binding energy (eV) Peak assignation

99.15 Si 2p 84.0 Au 4f7/2

150.3 Si 2s 87.6 Au 4f5/2

285 C1 164.2 S2 p3/2

532.4 O1s 284.8 C

water. The evolution of the area of the SiO2 and of the

oxygen peaks normalised to the area of the bulk Si 2p

peak is shown in Fig. 4a and b (full symbols). The area

of the SiO2 peak remains substantially zero for 10 min,

and then it growths until it saturates after about

1000 min. The induction period indicates an oxide

nucleation process. The characteristic time scale of this

process te, i.e. the time atwhich the area of the peaks are

1/e their saturation value, is of the order of 120 min.

The survey XPS spectra of SiO2–HMDS are

displayed in Fig. 2b besides the Si 2p and 2s peaks,

a significant O1s peak can be observed due to the

native oxide layer. The C1s peak originates from the

carbon in hexamethyldisilazane. Contact with water

affects the chemical properties of the surface only

marginally, as can be seen from the evolution of the

SiO2 and O peak areas in Fig. 4 (open circles).

However, for long immersion times in water, this

surface shows a decline of hydophobicity (Fig. 1),

which corresponds to a slight increase of the area of

the corresponding Si 2p and O1s peaks.

The XPS peaks of silicon coated with a monolayer

of OTS are summarised in Table 1. Again, the Si 2s

and Si 2p, and the O1s peaks are the dominating

features of the spectra. In addition, the C1s peak at

285 eV is is present due to the presence of the OTS.

The spectra of the dry sample and of the sample after

immersion in water do not show any appreciable

difference: for example, the difference in the carbon

peak area before and after immersion is less than 2%.

Furthermore, no peak other than the components of the

substrate and those of the adsorbed organic layer are

present.

The XPS peaks of gold coated with a monolayer of

DDT are also listed in Table 1. The gold peaks are

discernable over the entire binding energy range with

the 4f5/2 and 4f7/2 at 87.7 and 84.0 eV the most

prominent ones. The peak at 284.8 eV corresponds to

the carbon peak of the aliphatic chains of the

alkanethiols. No other components but those of the

substrate and the organic overlayer are present in the

spectra indicating a film free of contamination. The

spectrum obtained from the sample immersed in water

for 24 h does not show any appreciable change as

compared to the original state.

It follows that only surfaces hydrophobised with

long chain aliphaticmoieties satisfy the requirements of

stability inwater necessary to performmeasurements at

the water/hydrophobic interface. The roughness of

these surfaces, which is assumed to play an important

role in the hydration of hydrophobic surfaces [7], was

studied by atomic force microscopy in contact mode.

The root mean square roughness of the unloaded Au

substrate is 2.3 A, and 3.5 A for theAu substrate coated

with the SAM. This means that the dodecanethiols

adsorb on the gold surface following its irregular profile

only slightly affecting its topography. Also after the

adsorption of the organic monolayer, the roughness

remains small. Also in the case of silicon coated with

the OTS SAM, we found a very small roughness of

about 1 A, slightly smaller than the roughness of bare

silicon which was around 1.6 A.

4. Conclusions

The stability of four hydrophobic surfaces in contact

withwaterwas assessedby contact anglemeasurements

and XPS. Hydrogen-terminated silicon surfaces turned

out to have insufficient stability with significant

changes in the contact angle already after one minute

of exposure to water due to oxidation. HDMS-

terminated silicon exhibits better performance with a

drop in the contact angle from90� to 80� aftermore than

two hours of immersion in water. Only the surfaces

coated with long-chain aliphatic self-assembled mono-

layers, such asOTS on silicon andDDTon gold, did not

exhibit any appreciable changes even after 24 h of

M. Maccarini et al. / Applied Surface Science 252 (2005) 1941–19461946

exposure to water. In view of their stability, these latter

surfaces are proper candidates for studies on water/

hydrophobic surface effects bymeans of neutron,X-ray

reflectometry, surface force apparatus and atomic force

microscopy.

Acknowledgments

We gratefully acknowledge Georg Albert for

preparation of the gold-coated silicon substrates,

Roland Steitz and Jorg Fick for helpful discussions.

This work was funded by the Bundesministerium fur

Bildung und Forschung under grant no. GRE1HD and

in parts by the Deutsche Forschungsgemeinschaft

under grant no. Hi 693/2-1.

References

1 H.K. Christenson, P.M. Claesson, Adv. Colloid Interface Sci. 91

(2001) 391.

2 J. Wood, R. Sharma, Langmuir 11 (1995) 4797.

3 S. Ohnishi, H.K. Christenson, V.V. Yaminsky, Langmuir 16

(2000) 8360.

4 R. Steitz, T. Gutberlet, T. Hauss, B. Klosgen, R. Krastev, S.

Schemmel, A.C. Simonsen, G.H. Findenegg, Langmuir 19 (6)

(2003) 2409–2418.

5 J. Daillant, A. Gibaud, X-Ray and Neutron Reflectivity: Princi-

ples and Applications, Springer-Verlag, Berlin, Heidelberg, 1999.

6 K. Lum, D. Chandler, J.D. Weeks, J. Phys. Chem. B 103 (1999)

4570.

7 S.I. Mamatkulov, P.K. Khabibullaev, R.R. Netz, Langmuir 20

(2004) 4750.

8 J.T. Woodward, M.L. Walker, C.W. Meuse, D.J. Vanderah, G.E.

Poirier, A.L. Plant, Langmuir 16 (2000) 5347–5353.