Covalently attached monolayers on crystalline hydrogen-terminated silicon: Extremely mild attachment by visible light

Article abstract of DOI:10.1021/ja045359s

A very mild method was developed for the attachment of high-quality organic monolayers on crystalline silicon surfaces. By using visible light sources, from 447 to 658 nm, a variety of 1-alkenes and 1-alkynes were attached to hydrogen-terminated Si(IOO) and Si(111) surfaces at room temperature. The presence and the quality of the monolayers were evaluated by static water contact angles, X-ray photoelectron spectroscopy, and IR spectroscopy. Monolayers prepared by thermal, UV light, or visible light initiation were compared. Additionally, the ability of infrared reflection-absorption spectroscopy to study organic monolayers on silicon was explored. A reaction mechanism is discussed on the basis of investigations of the reaction behavior of 1-alkenes with silicon wafers with varying types and levels of doping. Finally, a series of mixed monolayers derived from the mixed solutions of a 1-alkene and an ω-fluoro-1-alkene were investigated to reveal that the composition of the mixed monolayers was directly proportional to the molar ratio of the two compounds in the solutions.

Full text of DOI:10.1021/ja045359s

Published on Web 02/04/2005
Covalently Attached Monolayers on Crystalline
Hydrogen-Terminated Silicon: Extremely Mild Attachment by
Visible Light
Qiao-Yu Sun,† Louis C. P. M. de Smet,† Barend van Lagen,† Marcel Giesbers,†
Peter C. Thu¨ne,‡ Johan van Engelenburg,† Frits A. de Wolf,§ Han Zuilhof,*,† and
Ernst J. R. Sudho¨lter†
Contribution from the Laboratory of Organic Chemistry, Wageningen UniVersity and Research
Center, Dreijenplein 8, 6703 HB Wageningen, The Netherlands, Schuit Institute of Catalysis,
EindhoVen UniVersity of Technology, P.O. Box 513, 5600 MB EindhoVen, The Netherlands, and
A&F InnoVations, Wageningen UniVersity and Research Center, Bornsesteeg 59,
6708 PD Wageningen, The Netherlands
Received August 2, 2004; E-mail: Han.Zuilhof@wur.nl
Abstract: A very mild method was developed for the attachment of high-quality organic monolayers on
crystalline silicon surfaces. By using visible light sources, from 447 to 658 nm, a variety of 1-alkenes and
1-alkynes were attached to hydrogen-terminated Si(100) and Si(111) surfaces at room temperature. The
presence and the quality of the monolayers were evaluated by static water contact angles, X-ray
photoelectron spectroscopy, and IR spectroscopy. Monolayers prepared by thermal, UV light, or visible
light initiation were compared. Additionally, the ability of infrared reflection-absorption spectroscopy to
study organic monolayers on silicon was explored. A reaction mechanism is discussed on the basis of
investigations of the reaction behavior of 1-alkenes with silicon wafers with varying types and levels of
doping. Finally, a series of mixed monolayers derived from the mixed solutions of a 1-alkene and an ω-fluoro-
1-alkene were investigated to reveal that the composition of the mixed monolayers was directly proportional
to the molar ratio of the two compounds in the solutions.
drosilylation catalysts,16-19 and chemomechanical scribing.20-22
Introduction
These studies have allowed a proper characterization of such
monolayers, and now allow a focus on the integration of
functional groups or biomolecules in monolayers on the silicon
surface. Although thermal conditions and UV irradiation yield
stable5,6 and densely packed monolayers,7,23 they are too harsh
to allow the use of labile bioactive materials. This situation has
two ways out (Scheme 1). In the first route (route 1) one makes
use of a protected precursor that can stand such harsh conditions
(1a). Subsequently, this precursor is deprotected (1b) and
Over the past eight to ten years several methods have been
developed to prepare organic monolayers on various morphol-
ogies of hydrogen-terminated silicon.1-3 Si-C bond attached
monolayers have been prepared by making use of thermal
conditions,4-8 UV irradiation,9-12 electrochemistry,13-15 hy-
† Laboratory of Organic Chemistry, Wageningen University and Research
Center.
‡ Eindhoven University of Technology.
§ A&F Innovations, Wageningen University and Research Center.
(1) Wayner, D. D. M.; Wolkow, R. A. J. Chem. Soc., Perkin Trans. 2 2002,
23-34.
(13) de Villeneuve, C. H.; Pinson, J.; Bernard, M. C.; Allongue, P. J. Phys.
Chem. B 1997, 101, 2415-2420.
(2) Buriak, J. M. Chem. ReV. 2002, 102, 1271-1308.
(3) Sieval, A. B.; Linke, R.; Zuilhof, H.; Sudho¨lter, E. J. R. AdV. Mater. 2000,
12, 1457-1460.
(14) Gurtner, C.; Wun, A. W.; Sailor, M. J. Angew. Chem., Intl. Ed. 1999, 38,
1966-1968.
(4) Linford, M. R.; Chidsey, C. E. D. J. Am. Chem. Soc. 1993, 115, 12631-
12632.
(15) Robins, E. G.; Stewart, M. P.; Buriak, J. M. Chem. Commun. 1999, 2479-
2480.
(5) Linford, M. R.; Fenter, P.; Eisenberger, P. M.; Chidsey, C. E. D. J. Am.
Chem. Soc. 1995, 117, 3145-3155.
(16) Boukherroub, R.; Morin, S.; Bensebaa, F.; Wayner, D. D. M. Langmuir
1999, 15, 3831-3835.
(6) Sieval, A. B.; Demirel, A. L.; Nissink, J. W. M.; Linford, M. R.; van der
Maas, J. H.; de Jeu, W. H.; Zuilhof, H.; Sudho¨lter, E. J. R. Langmuir 1998,
14, 1759-1768.
(17) Buriak, J. M.; Allen, M. J. J. Am. Chem. Soc. 1998, 120, 1339-1340.
(18) Buriak, J. M.; Stewart, M. P.; Geders, T. W.; Allen, M. J.; Choi, H. C.;
Smith, J.; Raftery, D.; Canham, L. T. J. Am. Chem. Soc. 1999, 121, 11491-
11502.
(7) Sieval, A. B.; van den Hout, B.; Zuilhof, H.; Sudho¨lter, E. J. R. Langmuir
2001, 17, 2172-2181.
(19) Holland, J. M.; Stewart, M. P.; Allen, M. J.; Buriak, J. M. J. Solid State
Chem. 1999, 147, 251-258.
(20) Niederhauser, T. L.; Jiang, G.; Lua, Y.-Y.; Dorff, M. J.; Woolley, A. T.;
Asplund, M. C.; Berges, D. A.; Linford, M. R. Langmuir 2001, 17, 5889-
5900.
(21) Niederhauser, T. L.; Lua, Y.-Y.; Sun, Y.; Jiang, G.; Strossman, G. S.;
Pianetta, P.; Linford, M. R. Chem. Mater. 2002, 14, 27-29.
(22) Niederhauser, T. L.; Lua, Y.-Y.; Jiang, G.; Davis, S. D.; Matheson, R.;
Hess, D. A.; Mowat, I. A.; Linford, M. R. Angew. Chem., Int. Ed. 2002,
41, 2353-2356.
(23) Sieval, A. B.; van den Hout, B.; Zuilhof, H.; Sudho¨lter, E. J. R. Langmuir
2000, 16, 2987-2990.
(8) Sieval, A. B.; Opitz, R.; Maas, H. P. A.; Schoeman, M. G.; Meijer, G.;
Vergeldt, F. J.; Zuilhof, H.; Sudho¨lter, E. J. R. Langmuir 2000, 16, 10359-
10368.
(9) Cicero, R. L.; Linford, M. R.; Chidsey, C. E. D. Langmuir 2000, 16, 5688-
5695.
(10) Effenberger, F.; Gotz, G.; Bidlingmaier, B.; Wezstein, M. Angew. Chem.,
Int. Ed. 1998, 37, 2462-2464.
(11) Strother, T.; Hamers, R. J.; Smith, L. M. Nucleic Acid Res. 2000, 28, 3535-
3541.
(12) Strother, T.; Cai, W.; Zhao, X.; Hamers, R. J.; Smith, L. M. J. Am. Chem.
Soc. 2000, 122, 1205-1209.
9
2514
J. AM. CHEM. SOC. 2005, 127, 2514-2523
10.1021/ja045359s CCC: $30.25 © 2005 American Chemical Society

Covalently Attached Monolayers on Crystalline Silicon
A R T I C L E S
Scheme 1. Preparation of Mixed Monolayers (Box) Containing Biomolecules (Oval Symbols) via Indirect (1) or Direct (2) Methodsa
a See the text.
Scheme 2. Representation of the Proposed (a) Radical Chain Mechanism5 under Thermal and UV Conditions and (b) Electron/Hole Pair
Mechanism28 for Visible-Light-Initiated Hydrosilylation on Porous Silicon
transformed (1c) into the bioactive monolayer. Such an approach
has been taken by, e.g., the groups of Hamers,12,24 Horrocks
and Houlton,25,26 and Wei et al.27
modification thereof.30 This reaction works well, but requires
an extra step, and involves a somewhat poorly defined substrate
due to the fact that the iodination is only partial.
The second route (2) would make use of mild reaction
conditions that are compatible with bioactive moieties. In the
case of porous silicon such an approach might be possible on
the basis of the white-light-promoted reaction discovered by
Stewart and Buriak.28 It was demonstrated that excitons gener-
ated through illumination of the surface of photoluminescent
porous silicon with white light (g400 nm) are capable of driving
a surface hydrosilylation reaction on hydride-terminated porous
silicon, producing Si-C bonds, also in line with recent
computational studies.29 On the basis of the bond strength of
the Si-H bond, a minimum energy of 3.5 eV (λ < 350 nm) is
required to perform Si-H bond cleavage. Since Stewart and
Buriak used wavelengths g400 nm, they concluded that the
mechanism of hydrosilylation on porous silicon does not start
with a homolytic cleavage of the Si-H bonds. For this reason
a radical chain mechanism5 (Scheme 2, top), as operative in
the case of thermal and UV hydrosilylation, was rejected, and
a new mechanism was proposed that involves attack by a
1-alkene or 1-alkyne on a surface-localized positive charge (the
hole) in a nucleophilic fashion, resulting in Si-C bond formation
(Scheme 2, bottom).28 It was thought that quantum confinement
effects were crucial for this attachment reaction, which was thus
thought to be limited to porous silicon. This was in line with
observations by Effenberger and co-workers,10 who used 380
nm light for the modification of flat hydrogen-terminated Si-
(111) with 1-octadecene, but the reported water contact angles
(only ∼95°) indicated a poor monolayer quality. To circumvent
this limitation, Hamers studied partially iodinated Si(111) and
Si(100) surfaces, and reported visible light (514 nm) initiated
Therefore, we were motivated to find an easy and extremely
mild route to modify planar Si surfaces, and recently, we
introduced this via the use of visible light (447 nm) to attach
organic monolayers on flat hydrogen-terminated silicon sur-
faces.31 Angle-resolved X-ray photoelectron spectroscopy
(ARXPS), infrared spectroscopy, and contact angle measure-
ments were used to show the formation of densely packed
monolayers. In another paper, we demonstrated the potential
of this reaction, by the attachment of a fragile monosaccharide.32
However, no systematic study has as yet been reported on the
influence of the nature and degree of doping, the variation of
the irradiation wavelength, and the type of silicon surface [Si-
(100) vs Si(111)]. In addition, no detailed surface studies by
scanning force techniques and X-ray reflectivity measurements
have yet been made, nor a fully unambiguous comparison
between the thermal and this new photochemical method
regarding the obtainable density of the resulting monolayers.
Finally, no systematic evaluation has yet been undertaken on
the applicability of IRRAS (infrared reflection-absorption
spectroscopy) measurements in the analysis of organic mono-
layers. Such an evaluation would be highly useful, given the
ease of use of this technique in the analysis of organic
monolayers, as recently shown by us in a study of organic
monolayers covalently attached to silicon nitride surfaces.33 In
the present paper we aim to address these issues and in
combination with our earlier work try to develop an overall
picture of the scope of this mild attachment reaction.
Experimental Section
Materials. (a) Chemicals. PE40/60, MeOH, EtOH, and CH2Cl2
were distilled prior to use. Distilled CH2Cl2sused as a solvent in
chemical reactionssand acetonitrile were distilled over CaH2 prior to
(24) Lin, Z.; Strother, T.; Cai, W.; Cao, X.; Smith, L. M.; Hamers, R. J. Langmuir
2002, 18, 788-796.
(25) Pike, A. R.; Patole, S. N.; Murray, N. C.; Ilyas, T.; Connolly, B. A.;
Horrocks, B. R.; Houlton, A. AdV. Mater. 2003, 15, 254-257.
(26) Pike, A. R.; Lie, L. H.; Eagling, R. A.; Ryder, L. C.; Patole, S. N.; Connolly,
B. A.; Horrocks, B. R.; Houlton, A. Angew. Chem., Int. Ed. 2002, 41, 615-
617.
(27) Wei, F.; Sun, B.; Guo, Y.; Zhao, X. S. Biosens. Bioelectron. 2003, 18,
1157-1163.
(28) Stewart, M. P.; Buriak, J. M. J. Am. Chem. Soc. 2001, 123, 7821-7830.
(29) Reboredo, F. A.; Schwegler, E.; Galli, G. J. Am. Chem. Soc. 2003, 125,
15243-15249.
(30) Cai, W.; Lin, Z.; Strother, T.; Smith, L. M.; Hamers, R. J. J. Phys. Chem.
B 2002, 106, 2656-2664.
(31) Sun, Q.-Y.; de Smet, L. C. P. M.; van Lagen, B.; Wright, A.; Zuilhof, H.;
Sudho¨lter, E. J. R. Angew. Chem., Int. Ed. 2004, 43, 1352-1355.
(32) de Smet, L. C. P. M.; Stork, G. A.; Hurenkamp, G. H. F.; Sun, Q.-Y.;
Topal, H.; Vronen, P. J. E.; Sieval, A. B.; Wright, A.; Visser, G. M.; Zuilhof,
H.; Sudho¨lter, E. J. R. J. Am. Chem. Soc. 2003, 125, 13916-13917.
(33) Arafat, A.; Schro¨en, K.; de Smet, L. C. P. M.; Zuilhof, H.; Sudho¨lter, E.
J. R. J. Am. Chem. Soc. 2004, 126, 8600-8601.
9
J. AM. CHEM. SOC. VOL. 127, NO. 8, 2005 2515

A R T I C L E S
Sun et al.
use. Mesitylene (Fluka, 99%) was distilled and stored on CaCl2. The
dried mesitylene was filtered to remove traces of CaCl2. 1-Decene
(Fluka, 97%) and 1-hexadecene (Sigma, ∼99%) were distilled at least
twice at reduced pressure. Triethylamine (TEA; Acros, 99%) was dried
over KOH. 1-Docosene (TCI America, 99+%), 10-undecen-1-ol
(Aldrich, 98%), acetone (Acros, 99+%), methanesulfonyl chloride
(Janssen Chimica, 99%), KF (Jansen Chimica, p.a.), and 18-crown-6
(Acros, 99%) were used as received. HF (Fluka, 50% p.a.-plus) was
diluted with demineralized H2O. 10-Undecenoic acid methyl ester was
prepared and purified as described earlier.6 All reactions were performed
under a nitrogen atmosphere. Straight-chain alkanes for the IRRAS
calibration were obtained from Aldrich (99+% purity).
(b) Wafers. Single-polished Si(100): n-type, 500-550 µm thick,
resistivity 1-2 Ω cm (Seltec Silicon, Mitsubishi Silicon America);
p-type, 375 µm thick, resistivity 1-2 Ω cm (Bayer Solar Freiberg,
Germany); n-type, 500-550 µm thick, resistivity 0.008-0.02 Ω cm;
p-type, 500-550 µm thick, 0.01-0.02 Ω cm (Addison Engineering,
San Jose). Single-polished Si(111): n-type, 475-550 µm thick,
resistivity 1-5 Ω cm (Addison Engineering, CA); p-type, 500-550
µm thick, resistivity 0.009-0.012 Ω cm (Addison Engineering). In this
paper silicon with resistivities of 1-5 and 0.008-0.02 Ω cm is referred
to as lowly doped and highly doped, respectively. Lowly doped n-type
Si(100) was used throughout this paper, unless specifically noted
otherwise.
filtration, evaporation yielded 7.228 g of crude product. Column
chromatography (PE40/60 as eluent) yielded 4.173 g (24 mmol, 48%)
of 11-fluoro-1-undecene: TLC Rf(PE40/60:CH2Cl2 ) 1:4) ) 0.81 and
1
Rf(PE40/60) ) 0.44; GC Rt(115 °C) ) 6.8 min; H NMR δ ) 5.88-
5.78 (m, 1H), 5.04-4.94 (m, 2H), 4.51 and 4.39 (dt, 2H, J ) 47.2 Hz,
J ) ∼6.2 Hz), 2.10-2.05 (q, 2H, J ) ∼6.8 Hz), 1.77-1.65 (m, 2H),
and 1.42-1.37 (m) and 1.33 (br s, 12H); 13C NMR δ ) 139.45, 114.48,
85.22, and 83.59 (J ) 163.5 Hz), 34.21, 30.93, and 30.74 (J ) 19.3
Hz), and 29.88, 29.80, 29.64, 29.51, 28.33, 25.59, and 25.54 (J ) 5.4
Hz); MS m/z 172.1622 (calcd for C11H21F, 172.1627).
Sample Cleaning and Etching. Si samples were first wiped with a
tissue that was saturated with chemically pure acetone. After that, the
samples were sonicated for at least 10 min in acetone. Then the samples
were placed in an oxygen plasma cleaner (Harrick PDC-32G) for 10
min. Subsequently, the Si(100) samples were etched in 2.5% aqueous
HF for 2 min, while Si(111) samples were etched in an argon-saturated
40% aqueous NH4F solution for 15 min under an argon atmosphere.
To recycle the relatively expensive attenuated total reflection (ATR)
crystals, they were cleaned by oxidative removal of a previously formed
monolayer in “piranha solution” (30% H2O2:H2SO4 ) 1:2 (v/v)) at 85
°C for 1 h (caution: piranha solutions should be handled with great
care),8 and subsequently etched with HF as described above.
Monolayer Preparation. (a) Photochemical Method. A solution
(1.5-3 mL, 0.1-0.5 M) of 1-alkene(s) in mesitylene was flushed with
argon for at least 30 min before the freshly etched hydrogen-terminated
Si sample was added. After that, the solution was flushed for another
30 min. Then, the argon inlet was moved to a position just above the
solution to change from a bubbling solution to a decent gas flow. After
that, the lampsin all cases fixed at a distance of 0.5 cm from the
reaction vesselswas turned on. For the modification of ATR crystals
two lamps were used. The setup was covered with aluminum foil.
Depending on the wavelength, different vessels with a rectangular
bottom part were used (quartz vessel for 254 nm and conventional
glassware for all other wavelengths). After illumination for the desired
time, the sample was removed from the solution, and the surface was
excessively rinsed with PE40/60, EtOH, and CH2Cl2, respectively. The
following models of double-bore lamps were obtained from Jelight Co.
Inc. (Irvine, CA): 82-3309-2 (254 nm), 84-2051-2 (371 ( 19 nm )
width at half of the maximum intensity), 84-247-2 (447 ( 32 nm),
84-213-2 (504 ( 30 nm), and 84-236-1 (658 ( 14 nm).
(b) Thermal Method. A solution (8.5 mL, 0.2 M) of 1-alkene(s) in
mesitylene was placed in a small three-necked flask fitted with a
nitrogen inlet, a reflux condenser with a CaCl2 tube, and a stopper.
The solution was refluxed for at least 45 min under a flow of nitrogen.
Subsequently, a cleaned and freshly etched sample was added to the
refluxing solution, while a slow nitrogen flow was maintained. After 2
h the modified sample was removed from the solution and excessively
rinsed with PE40/60, EtOH, and CH2Cl2, respectively.
Purification and Analysis of Synthesized Compounds. Thin-layer
chromatography (TLC) was performed on Merck silica gel 60F254
plastic sheets, and detection was realized by charring with an aqueous
solution of KMnO4. Column chromatography was conducted by elution
of a column of Merck Kieselgel silica (230-400 mesh) using eluents
as specified below. NMR spectra were recorded on a Bruker AC-E
400 spectrometer in CDCl3 (dried over Al2O3) at room temperature.
Gas chromatography (GC) measurements were performed on a Hewlett-
Packard 5890 series II chromatograph that was equipped with a DB-
17 column and an FID detector. GC samples were dissolved in ethyl
acetate or diethyl ether.
Synthesis of 11-Fluoro-1-undecene. 10-Undecen-1-ol (8.501 g,
49.92 mmol) was dissolved in 65 mL of dry CH2Cl2, and 10.4 mL
(7.59 g, 75 mmol) of triethylamine was added. The stirred solution
was cooled to -10 °C, and 8.6 g (75 mmol) of mesyl chloride in 20
mL of dry CH2Cl2 was added dropwise at such a rate that the
temperature of the solution was not higher than 0 °C. Thirty minutes
after the addition was accomplished, TLC (PE40/60:CH2Cl2 ) 1:4)
showed that hardly any starting material (Rf ) 0.17) was left, and a
product with Rf ) ∼0.38 was formed. The solution was stored at -35
°C until further workup. The yellow suspension was washed with 200
mL of 0.5 M HCl and the product extracted with CH2Cl2 (5 × 75 mL).
The combined organic layers were washed with saturated NaHCO3 (3×)
and dried over Na2SO4. After filtration the solvent was removed under
1
a reduced pressure, yielding 12.786 g of crude product. TLC and H
Analysis of the Monolayers. (a) Contact Angle Measurements.
In all cases where modified samples were analyzed by different
techniques a small piece (∼5 mm × 10 mm) was cut from the sample
directly after cleaning. Static water contact angles were obtained using
an Erma G-1 contact angle meter (volume of the drop of ultrapure water,
3.5 µL). Contact angles of two or three drops were measured. The error
of the contact angles is (1°.
(b) IRRAS. FT infrared reflection-absorption spectra were recorded
on a Bruker Tensor 27 instrument equipped with a variable-angle
reflection Auto Seagull accessory. A Harrick grid polarizer was installed
in front of the detector for measuring spectra with p-polarized (parallel)
radiation with respect to the plane of incidence at the sample surface.
Single-channel transmittance spectra (4096 scans) were collected using
a spectral resolution of 4 cm-1. All (data derived from) spectra shown
in this paper are a result of spectral subtraction of modified samples
with a cleaned native oxide-covered silicon sample without any further
data manipulation.
NMR showed that 10-undecenyl mesylate was pure enough for further
use.
KF (5.8 g, 100 mmol) and 18-crown-6 (26.4 g, 100 mmol) were
dissolved in acetonitrile (50 mL). The solution was heated to 60 °C,
and to the resulting white suspension was added 12.394 g (49.9 mmol)
of 10-undecenyl mesylate dissolved in 50 mL of acetonitrile over a
period of 45 min. Subsequently, the solution was refluxed overnight.
The course of the reaction was followed by TLC and 1H NMR via the
workup of small analytical samples. After 20 h, 1H NMR showed that
about 60% of the 10-undecenyl mesylate was still present, and therefore,
1 equiv of KF and 18-crown-6 were added to the reaction mixture.
After another 20 h, TLC and 1H NMR showed that most of the starting
material had disappeared. The solvent was removed under a reduced
pressure. PE40/60 (200 mL) and water (100 mL) were added to the
mixture. The organic layer was separated, and the water layer was
washed twice with PE40/60 (100 mL). The combined organic layers
were washed with brine (300 mL) and dried over Na2SO4. After
9
2516 J. AM. CHEM. SOC. VOL. 127, NO. 8, 2005

Covalently Attached Monolayers on Crystalline Silicon
A R T I C L E S
Figure 1. C1s and Si2p regions of an XPS spectrum of an n-hexadecyl
monolayer on Si(100) prepared via the visible light method (447 nm,
15 h).
Table 1. Static Water Contact Angles (deg) of Monolayers on
Silicon Surfaces Prepared by 447 nm Irradiation and Thermal
Methodsa
Figure 2. X-ray reflectivity profiles for n-docosyl monolayers on lowly
doped n-Si(100) prepared via the thermal (red) and visible light (447 nm,
15 h, blue) methods.
447 nm,
Si(100)
447 nm,
Si(111)
thermal,
Si(100)
reactant
CHtCC10H21
CHtCC12H25
CHtCC14H29
CH2dCHC10H21
CH2dCHC12H25
CH2dCHC14H29
CH2dCHC8H16COOCH3 (I)
CH2dCHC8H16COOCH2CF3 (II)
108
110
110
109
108
109
78
109
110
110
109
110
110
75
108b
110b
110b
108c
108c
109c
77c
with those of high-quality monolayers prepared under thermal
conditions.8 The contact angles of monolayers of compounds I
and II are lower due to the presence of the polar functionality,
and in both cases comparable to those of monolayers prepared
via thermal methods. Finally, the differences between the contact
angles of monolayers of I and II on Si(100) and Si(111) are
small, which indicates that after monolayer formation the top
of the organic monolayer does not obviously “display” the
character of the silicon substrate anymore. These initial char-
acterizations drove us to more detailed investigations using XPS,
X-ray reflectivity, and AFM.
85
86
88d
a Conditions: irradiation method, 15 h, room temperature; thermal
method, heating in a neat reactant6 or refluxing for 2 h in a 0.2 M solution
in mesitylene.8 All experiments were performed at least twice, experimental
error (1°. b From ref 8. c From ref 6. d Newly obtained using conditions
as described in ref 8.
Visible-light-prepared (447 nm) hexadecyl monolayers on Si-
(100) were analyzed by XPS, a very useful tool for the study
of covalently attached monolayers on silicon.9,12,16,24,30,34-41
Figure 1 shows the C1s and Si2p regions of the XPS spectrum
of a monolayer derived from 1-hexadecene on hydrogen-
terminated Si(100). The C1s region spectrum shows a clear peak
at 288 eV that is absent on samples without a monolayer,
supporting the formation of covalently attached monolayers. In
addition, the Si2p spectra display hardly any contribution of
Si-O bonds (103-104 eV), indicating there is only a small
amount of surface oxidation observable.
(c) XPS. XPS measurements were performed on a VG Ionex system
equipped with a Clam II analyzer and a standard Al KR X-ray source.
Spectra were taken in normal emission at 10-9 mbar within 10 min.
All C1s peaks corresponding to hydrocarbons were calibrated to a
binding energy of 285.0 eV to correct for the energy shift caused by
charging. The XPS measurements for Figure 1 were performed on a
Quantera SXM from Physical Electronics, equipped with a monochro-
mator and an Al KR X-ray source of 1486.6 eV. The spot was 100 µm
in diameter.
(d) X-ray Reflectivity. X-ray reflectivity measurements were
performed on a Panalytical X’Pert Pro diffractometer using Cu KR
radiation (tube settings 40 kV and 40 mA). The data were collected
using a fixed divergence slit of 1/32° and a parallel plate collimator on
the diffracted beam side. The layer thickness is calculated from the
interference fringes. The error in the measurement is depicted as the
standard deviation of five measurements on the same sample.
(e) Atomic Force Microscopy (AFM). Surface topography was
imaged using a Nanoscope III atomic force microscope (AFM Digital
Instruments, Santa Barbara, CA). The contact mode (CM-AFM) with
silicon nitride cantilevers (Digital Instruments) with a spring constant
of about 0.58 N/m was used.
X-ray reflectivity measurements were performed to compare
n-docosyl (n-C22H45) monolayers on Si(100) prepared via either
visible light or thermal attachment methods. 1-Docosene was
chosen as the alkene, since this yielded a monolayer with
sufficient thickness to obtain quantitatively reliable X-ray
reflectivity data. Figure 2 shows the X-ray reflectivity profiles
for two monolayers obtained by 447 nm irradiation or thermal
treatment, respectively. Just like the static water contact angles
of these layers (∼107°), the X-ray reflectivity profiles are almost
Results and Discussion
(34) Bansal, A.; Li, X.; Yi, S. I.; Weinberg, W. H.; Lewis, N. S. J. Phys. Chem.
B 2001, 105, 10266-10277.
Formation of Organic Monolayers on Si(100) and Si(111)
Using 447 nm Light. Table 1 lists the static water contact angles
of the Si(100) and Si(111) surfaces modified by a variety of
1-alkenes and 1-alkynes and two esterified-1-alkenes (com-
pounds I and II) using 447 nm light (15 h). The work on
monolayers on Si(100) was communicated previously,31 where
it was shown that the monolayer formation was complete after
10 h. The contact angles of the nonfunctionalized monolayers
indicate a high hydrophobicity, and the values are comparable
(35) Zhang, J.; Cui, C. Q.; Lim, T. B.; Kang, E.-T.; Neoh, K. G.; Lim, S. L.;
Tan, K. L. Chem. Mater. 1999, 11, 1061-1068.
(36) Bergerson, W. F.; Mulder, J. A.; Hsung, R. P.; Zhu, X.-Y. J. Am. Chem.
Soc. 1999, 121, 454-455.
(37) Kim, N. Y.; Laibinis, P. E. J. Am. Chem. Soc. 1998, 120, 4516-4517.
(38) Lee, E. J.; Bitner, T. W.; Ha, J. S.; Shane, M. J.; Sailor, M. J. J. Am.
Chem. Soc. 1996, 118, 5375-5382.
(39) Hong, L.; Sugimura, H.; Furukawa, T.; Takai, O. Langmuir 2003, 19, 1966-
1969.
(40) Linford, M. R.; Chidsey, C. E. D. Langmuir 2002, 18, 6217-6221.
(41) Sieval, A. B.; Linke, R.; Heij, G.; Meijer, G.; Zuilhof, H.; Sudho¨lter, E. J.
R. Langmuir 2001, 17, 7554-7559.
9
J. AM. CHEM. SOC. VOL. 127, NO. 8, 2005 2517

A R T I C L E S
Sun et al.
Figure 3. Section analysis obtained via contact mode AFM of photochemically prepared n-hexadecyl-modified Si(100) (left) and Si(111) (right) wafers
[447 nm, 15 and 5 h, respectively].
Table 2. Static Water Contact Angles (deg) of
1-Hexadecene-Derived Monolayers on Lowly Doped n-Type
possibility to obtain densely packed monolayers under a wide
Si(100) (15 h of Irradiation) and Lowly Doped n-Type Si(111)
thermal initiation reaction, ∼108-110°. This indicates the
range of mild conditions and the resulting compatibility with
biomaterials, one example of which can be found in the direct
attachment of labile sialic acid derivatives that have limited
stability at high temperature.32
(5 h of Irradiation) Prepared by Different Wavelengths
λ
n-type
n-type
λ
n-type
n-type
(nm)
Si(100)
Si(111)
(nm)
Si(100)
Si(111)
254
371
447
109
109
109
110
110
109
504
658
108
108
110
110
The easiest way to obtain fully covered modified Si wafers
is to use cheap wide-spectrum phosphorescent “high-efficiency”
lamps, possibly in combination with appropriate wavelength
filters. Broad-spectrum (15 W, 400-730 nm) lamps, in com-
bination with light filters that remove all light below 500, 550,
or 600 nm, yielded n-hexadecyl monolayers with a contact angle
of g108° upon irradiation for 14 h. The quality of the monolayer
seems to be dependent more on the humidity of the surroundings
and the completeness of oxygen removal than on the precise
light source used.
identical. The monolayer thickness obtained from the X-ray data
is 26 ( 1 Å for both monolayers. Using the calculated length
of an all-trans-docosyl tail (27.9 Å),42 the tilt angle with respect
to the normal of the surface can be calculated. This results in a
tilt angle of 15-26°, of which the top limit is close to the tilt
angles of thinner alkyl monolayers.6
Monolayers of 1-hexadecene on Si(100) and Si(111)s
prepared via the thermal and visible-light methods (447 nm)s
were evaluated using CM-AFM. In Figure 3 the corresponding
section analyses are presented for the photochemically prepared
samples. CM-AFM shows that after wet etching and monolayer
formation flat surfaces can be maintained, and no significant
differences were observed between thermally and photochemi-
cally prepared samples. In addition, the modified Si(111) surface
(Figure 3, right) remains a bit more smooth than the Si(100)
surface (Figure 3, left). The observed flatness is of the same
order of magnitude as reported recently by Cattaruzza et al. for
the thermal functionalization of H-terminated Si(100) with
6-heptynoic acid methyl ester.43 In addition, Wayner and co-
workers used AFM to study decyl-modified Si(111) surfaces
prepared via different techniques, and reported atomically flat
terraces.16
Effect of Irradiation Wavelength. The possibility to achieve
monolayer attachment even under milder conditions was tested
via the study of the photochemical reaction over a wide range
of wavelengths (254-658 nm, Table 2) using phosphorescent
lamps with various emission maxima (see the Experimental
Section), and via the use of a wide-spectrum fluorescent lamp
in combination with the appropriate wavelength filters. The
maximum contact angles of an n-hexadecyl monolayer that can
be achieved using light of 371 nm to up to at least 658 nm on
both H-terminated Si(100) and Si(111) are comparable with that
of high-quality monolayers prepared by UV light (254 nm) or
Upon irradiation with UV light (254 nm) the color of a
1-alkene in mesitylene solution becomes yellow, indicating the
photochemical formation of side products.44 The use of visible
light of 371-658 nm did not result in a color change. Gas
chromatography experiments indeed showed the formation of
additional products after irradiation with UV light, while the
composition of the solution did not change detectably upon
irradiation with 447 nm light. The (likely almost) unchanged
nature of alkene/mesitylene solutions after 371-658 nm ir-
radiation allows the use of recycled alkene/mesitylene mixtures
after minimal purification efforts (washing the complete solution
through a small silica plug to remove trace amounts of any
polymers), which is specifically advantageous in light of the
synthetic efforts that are sometimes required to obtain func-
tionalized alkenes.
One of the most versatile functional groups to have available
on the surface would be the carboxylic acid moiety. In the
thermal reaction this can only be obtained via the use of, e.g.,
esters,6,12,24,45,46 as the direct attachment of compounds with free
hydroxyl moieties yields undesirable reactivity of those func-
tional groups, leading to Si-O(CO)R bond formation. Hydroly-
sis of the ester, e.g., by treatment with potassium tert-butoxide
in DMSO, then yields carboxylic acid-modified monolayers with
contact angles of ∼52°.12 However, recently, Liu et al. showed
that this deprotection method diminishes the amount of attached
alkyl tails in the monolayer.45 Interestingly, Wayner and co-
workers recently reported ATR spectra of monolayers of
undecenylic acid on hydrogen-terminated Si(111), which were
(42) This value was obtained via a semiempirical calculation (MOPAC, PM3,
closed-shell-restricted) in Chem3D Pro.
(43) Cattaruzza, F.; Cricenti, A.; Flamini, A.; Girasole, M.; Longo, G.; Mezzi,
A.; Prosperi, T. J. Mater. Chem. 2004, 14, 1461-1468.
(44) Cornelisse, J. Chem. ReV. 1993, 93, 615-669.
(45) Liu, Y.-J.; Navasero, N. M.; Yu, H.-Z. Langmuir 2004, 20, 4039-4050.
9
2518 J. AM. CHEM. SOC. VOL. 127, NO. 8, 2005

Covalently Attached Monolayers on Crystalline Silicon
A R T I C L E S
prepared by direct irradiation (λ ≈ 300 nm) of a solution of
1-undecenylic acid, but no contact angle data were given.46
To further investigate this, we studied the direct attachment
of 1-undecenylic acid on hydrogen-terminated Si(100) by UV
(254 nm) and visible (447 and 658 nm) light. The contact angles
of the resulting layers prepared by visible light were 66-68°
in all three cases, while in the case of the thermal reaction a
contact angle of 79° was observed. The result of 66-68° in
combination with ATR-IR data that display a strong CdO signal
implies the presence of a significant amount of free carboxylic
acid groups in the monolayer, but also suggests that no densely
packed monolayer of acid-terminated alkyl chains is formed,
as significantly lower contact angles are to be expected in that
case (e.g., reported contact angles of acid-terminated monolayers
on gold are in the range of 0 to ∼15°).47,48 The thermally
obtained value of 79° is indicative of a mixture of Si-C-bound,
acid-terminated alkyl chains and Si-O-bound undecenoyl
chains. The presence of the latter linkage is shown upon
treatment with dilute HF solutions: short treatment of the
photochemically prepared monolayers does not affect their
contact angles, while the contact angle of the thermally obtained
monolayers decreases to 73°, caused by the replacement of
highly apolar Si-O undecenoyl chains by Si-H sites. In
addition, it can be concluded that the photochemical method
works as well for all three wavelengths (254, 447, and 658 nm)
that were used.
Figure 4. C-H stretching vibration region of the IRRA spectrum of a
photochemically (447 nm) obtained n-hexadecyl monolayer on hydrogen-
terminated Si(100) (IR reflection angle 68°, p-polarized light).
Table 3. IRRAS Data of the Antisymmetric and Symmetric
C-H Stretching Vibrations of n-Hexadecyl Monolayers on
Hydrogen-Terminated Si(100) and Si(111) Surfaces Obtained by
Different Attachment Methodsa
Si(100)
Si(111)
method
ν
a(C−H)
ν
s(C−H)
ν
a(C−H)
νs(C−H)
thermal
irradiation
254 nm
371 nm
447 nm
504 nm
658 nm
2923
2854
2923
2854
2926
2925
2925
2924
2926
2856
2853
2856
2855
2855
2925
2925
2924
2924
2924
2856
2856
2855
2854
2856
a IRRAS data: p-polarized light, IR reflection angle 68°, in inverse
centimeters. Irradiation time: 254 nm, 3 h; 371, 447, 504, and 658 nm,
15 h.
Table 4. IRRAS Data of the Antisymmetric and Symmetric
C-H Stretching Vibrations of Organic Monolayers on
Hydrogen-Terminated Si(100) and Si(111) Surfaces Obtained for
Different 1-Alkenes and 1-Alkynes upon Irradiation with 447 nm
Light for 15 ha
Infrared Reflection-Absorption Spectroscopy. Detailed
information about the structure of organic monolayers can
frequently be obtained by infrared spectroscopy. Features
derived from C-H stretching vibrations or functional groups
such as esters or amides can reveal the presence4-6,33,41,45,49 as
well as conversion6,33,41,45 of attached molecules. Furthermore,
the precise positions of the antisymmetric and symmetric C-H
stretching vibration bands in an IR spectrum of a monolayer
may provide information about the quality of the monolayers,
since they can be compared with those of C-H bands of
isotropic and crystalline materials.5 Due to the low absorbance
of an organic monolayer and the concomitantly low signal-to-
noise ratio, one usually has to rely on ATR spectroscopy, a
method in which the signal-to-noise ratio is enhanced by
multiple internal reflections of the IR beam in a modified
double-polished ATR crystal. Alternatively, infrared spectra can
also be obtained by external reflection techniques such as
IRRAS.50 IRRA spectra can be obtained using a commercially
available variable-angle reflection accessory in combination with
a standard IR spectrophotometer. This technique would poten-
tially be advantageous as it allows easy measurements of
vibrational features and subsequent other characterizations (X-
ray reflectivity, XPS, water contact angles) all on the same
sample, which is harder to do using ATR crystals (both
technically and pricewise). Figure 4 gives a typical example of
the C-H stretching region of an n-hexadecyl monolayer on
hydrogen-terminated Si(100) obtained by IRRAS.
Si(100)
H) s(C
Si(111)
H) s(C−H)
reactant
ν
a(C−
ν
−H)
νa(C
−
ν
CHtCC10H21
CHtCC12H25
CHtCC14H29
CH2dCHC10H21
CH2dCHC12H25
CH2dCHC14H29
CH2dCHC8H16COOCH3 (I)
CH2dCHC8H16COOCH2CF3 (II)
2927
2925
2925
2926
2925
2925
2930
2931
2854
2854
2855
2855
2855
2856
2855
2855
2925
2922
2922
2924
2925
2924
2925
2925
2854
2852
2852
2857
2856
2855
2856
2857
a IRRAS data: p-polarized light, IR reflection angle 68°, in inverse
centimeters.
As can be seen from the spectrum in Figure 4, IRRAS clearly
shows the presence of a covalently attached organic monolayer.
Table 3 summarizes the IRRAS results of n-hexadecyl mono-
layers on Si(100) and Si(111). For one wavelength (447 nm,
15 h) the study was extended by the preparation of monolayers
with different types of 1-alkenes/1-alkynes on Si(100) and Si-
(111). The results of that variation are presented in Table 4.
On average, IRRAS on the monolayers yields peak positions
corresponding to antisymmetric and symmetric C-H stretching
vibrations at 2925 and 2856 cm-1, respectively (Tables 3 and
4). When compared with ATR data obtained on such mono-
layers, it is evident that these IRRAS values are all higher. The
obtained values were slightly different for different types of
monolayers, butsas expected on the basis of IRRAS theory50s
also depended on IRRAS parameters, such as the incidence
angle. In addition, the C-H stretching vibrations of samples
that were prepared by our standard thermal procedure51 were
observed in IRRAS at consistently higher wavenumbers than
in ATR (Table 3, 2923 and 2854 cm-1 versus 2921 and 2852
(46) Boukherroub, R.; Wayner, D. D. M.; Wojtyk, J. U.S. Patent 6677163,
Jan 13, 2004.
(47) Jennings, G. K.; Laibinis, P. E. J. Am. Chem. Soc. 1997, 119, 5208-5214.
(48) Laibinis, P. E.; Whitesides, G. M. J. Am. Chem. Soc. 1992, 114, 1990-
1995.
(49) Boukherroub, R.; Wayner, D. D. M. J. Am. Chem. Soc. 1999, 121, 11513-
11515.
(50) Brunner, H.; Mayer, U.; Hoffmann, H. Appl. Spectrosc. 1997, 51, 209-
217.
9
J. AM. CHEM. SOC. VOL. 127, NO. 8, 2005 2519

A R T I C L E S
Sun et al.
Figure 6. C-H (left) and CdO (right) stretching vibration regions of the
IRRA spectrum of an n-type Si(100) modified with compound I using
irradiation with 447 nm light (IR reflection angle 68°, p-polarized light).
ments, and the fact that it allows multiple techniques on the
same surface, IRRAS is a potentially useful and informative
infrared technique.
Figure 5. IRRAS band maxima for the antisymmetric (solid symbols) and
symmetric (open symbols) C-H stretching vibrations of deposited alkane
Analysis of the IRRAS data in Tables 3 and 4 shows that
both the thermally and photochemically prepared covalently
attached monolayers on silicon seem to fall in the regime of
disordered materials. This is in contrast with ATR data on the
same samples, which seem to indicate a substantial degree of
anisotropy, i.e., some degree of ordering, perhaps stemming from
the CH2 moieties bound most closely to the surface. A more
detailed analysis of this difference would be desirable, given
the ease of use of IRRAS, but is outside the scope of the present
study. At the moment it suffices to say that IRRAS is a viable
technique to display the presence or absence of functional groups
on the surface, and that the relation between ATR-obtained
values for the C-H stretching vibrations in organic monolayers
on silicon and the ordering of such monolayers deserves more
detailed attention.
The usefulness of IRRAS can, e.g., be shown with IRRAS
measurements of monolayers of compounds I and II (two
bottom entries in Table 4). Not only can the antisymmetric and
symmetric C-H stretching vibrations be observed, but also a
clear CdO stretching vibration indicative of an ester can be
observed (Figure 6). Hydrolysis of such esters can easily be
followed by IRRAS as shown by our recent work on silicon
nitride.33 All in all, we believe IRRAS measurements to be a
novel and convenient tool for the study of covalently attached
organic monolayers.
Effect of the Type of Silicon: Doping Type, Doping
Concentration, and Orientation. Hamers and co-workers
studied the photoattachment of alkenes on partly I-terminated
Si(111) and Si(100),30 and reported significant effects of the
doping of the Si wafers. Following this approach, we selected
four different types of Si(100) wafers, lowly and highly doped
n-type and lowly and highly doped p-type Si(100), and two types
of Si(111), to perform a more systematic study. The surfaces
were etched and modified with 1-hexadecene using 447 nm
light. The contact angles of the resulting monolayers are plotted
as a function of the irradiation time (Figure 7a). For any given
irradiation time, the contact angles of monolayers on n-type Si-
(100) are higher as compared to those of monolayers on p-type
Si(100). Since the preparation conditions (including light
intensity) were constant, it can be concluded that the efficiency
of the photochemical monolayer formation on n-type Si(100)
is higher than on p-type Si(100). From Figure 7a it can also be
seen that covalent attachment follows the rate order: highly
doped n g lowly doped n > lowly doped p > highly doped p.
Figure 7b presents the time-dependent contact angles of
1-hexadecene-derived monolayers on lowly doped n-type and
highly doped p-type Si(111). According to the contact angle
data, monolayer formation on Si(111) is complete after about
layers on Si(100) as a function of the chain length n of alkanes (CnH2n+2
)
(IRRAS data: p-polarized light, IR reflection angle 68°, in inverse
centimeters).
cm-1, respectively).6 Therefore, a detailed investigation was
desirable to obtain a clear picture of the potential of IRRAS in
the study of such monolayers.
First, it was investigated whether IRRAS could also provide
information about the degree of disorder (liquidlike or solidlike)
in such thin layers, via measurement of thin layers of a
homologous series of liquid and solid alkanes (n-C12H26 to
n-C24H50) deposited on Si wafers with their native oxide layer
still present. All spectra were obtained at an incidence angle of
68° with p-polarized light, since it was found for an n-hexadecyl
monolayer on Si(100) that angles in the range of ∼60° to ∼70°
gave the best signal-to-noise ratio (Supporting Information).
Figure 5 shows the results for the antisymmetric and symmetric
stretching vibrations for the different chain lengths.
It can be seen that the measured wavenumbers for the
antisymmetric (∼2928 cm-1) and symmetric (∼2856 cm-1
)
C-H stretching vibrations of the liquid alkanes (n < 18) are
comparable with the values found with conventional IR.52 In
contrast, the IRRAS values for the antisymmetric and symmetric
C-H stretching vibration bands of the solid alkanes (n > 17)
are found to be lower than the conventional IR values: 2914
cm-1 (vs 2920 cm-1) and 2848 cm-1 (vs 2850 cm-1),
respectively. The results show that for the conditions used (68°,
p-polarized light) application of IRRAS yields in principle a
larger difference between crystalline/nonisotropic media and
isotropic media than conventional infrared measurements: 14
cm-1 (vs 8 cm-1) for the antisymmetric C-H stretching
vibration and 8 cm-1 (vs 6 cm-1) for the symmetric C-H
stretching vibration. The situation is best seen at the data points
obtained for n-C18H38 (mp ) 29-30 °C). At temperatures <15
°C solidlike data points are obtained, while at temperatures >35
°C the peaks belonging to a liquid state are observed. In fact,
during the melting of solid octadecane two sharp peaks for both
the symmetric and antisymmetric C-H stretching vibrations are
observed, indicating the simultaneous presence of two distinct
phases, i.e., degrees of organization. As IRRAS only uses
external reflections, in contrast to ATR that relies on internal
reflections, IRRAS is more broadly applicable, e.g., to IR-
absorbing substrates (e.g., Si3Nx).33 In combination with the
rather straightforward and fast implementation of the measure-
(51) Sieval, A. B.; Vleeming, V.; Zuilhof, H.; Sudho¨lter, E. J. R. Langmuir
1999, 15, 8288-8291.
(52) Snyder, R. G.; Strauss, H. L.; Elliger, C. A. J. Phys. Chem. 1982, 86, 5145-
5150.
9
2520 J. AM. CHEM. SOC. VOL. 127, NO. 8, 2005

Covalently Attached Monolayers on Crystalline Silicon
A R T I C L E S
While the present work clearly displays the scope of the
visible light attachment, the mechanism of this reaction is much
less clear. In a recent paper31 we hypothesized that an electron/
hole pair dissociation mechanism (Scheme 2, route b) would
be feasible. However, two arguments cause us to refine our
hypothesis. First, using scanning tunneling microscopy (STM)
on partially formed monolayers of 1-decene on hydrogen-
terminated Si(111), we showed, in collaboration with Eves and
Lopinski, that the reaction likely proceeds via a radical chain
mechanism.53 This suggests that after some initiation process
reactive sites/dangling bonds are formed at the surface that can
initiate a chain reaction. Second, and more complex, is the role
of supposed band bending in these materials,30,31 since the
surface is in contact with a nonconducting solution. Therefore,
it is presently difficult to ascertain the degree of band bending
that does actually occur in these wafers in contact with a
mesitylene solution, while additionally other factorsse.g., faster
oxidation of p-type hydrogen-terminated silicon than n-type
hydrogen-terminated siliconsmight also play a role. This
complication also arises for the case of UV light initiation, as
the top left graph in Figure 8 shows a dopant dependence for
that wavelength as well, which cannot be fit into the mechanism
(Scheme 2, route a) proposed for that reaction9 without involving
a silicon substrate-dependent parameter.
Figure 7. Water contact angles of n-hexadecyl monolayers as a function
The discussion of the mechanism of the visible-light-induced
reaction therefore needs to be limited to a rather broad
delineation. The cleavage of a Si-H bond requires light of <350
nm, whereas <540 nm is needed for the cleavage of a Si-Si
bond. Since attachment and the formation of hydrophobic
monolayers can be established with visible light of wavelength
even larger than 650 nm, an initiation step involving a
(homolytic) cleavage of the Si-H or Si-Si bonds is not
possible. Therefore, a different initiation is required. In addition,
route b in Scheme 2 cannot explain the observation of island
formation in the attachment of 1-decene to hydrogen-terminated
Si(111) as shown by STM.53
We therefore tentatively propose a new mechanism in Scheme
3. Due to the formation of delocalized radical cations at the
silicon surface (electrons temporarily being moved to the bulk)
upon excitation thereof, this surface is susceptible to nucleophilic
attack. Such a reaction may result in a Si-Si bond cleavage in
a concerted manner, as in Si-Si-containing radical cations the
Si-Si bond is not very strong, while the resulting Si-centered
cation at the surface is highly stabilized by the neighboring Si
atoms. As a result we obtain the structure at the bottom left in
Scheme 3, with a â-CH radical site. This radical can pick up a
H atom, and leave a Si radical at the surface that is available
for the attachment of a second alkene.
of irradiation time for different types of (a) Si(100) and (b) Si(111).
5 h, while it takes using the same light intensity about 10 h on
Si(100). Despite the significant difference for the initially
measured water contact angles for n- and p-type Si(111), the
time at which the plateau value is reached depends only slightly
on the doping type (4 vs 5 h for the n- and p types, respectively),
and the final contact angle (∼109°) is identical in both cases.
The differences between the reaction efficiency of lowly
doped n-type silicon and heavily doped p-type Si(100) were
also studied for 254, 371, 504, and 658 nm light (Figure 8).
For all presented wavelengths, hydrogen-terminated n-type
silicon surfaces are more reactive toward unsaturated hydro-
carbons than p-type silicon surfaces. Note that different lamps
with different intensities were used in this experiment. As a
result comparison of efficiencies is only allowed for the different
types of silicon that were modified with the same wavelength.
The different reaction times required to obtain the plateau
contact angles for the different wavelengths are in fact the result
of differences in light intensity. However, it is evident that, using
254, 504, and 658 nm light, the high contact angles obtained at
n-type Si(100) (∼109°) could not be obtained at p-type Si(100).
For practical purposes the use of n-type Si(100) or n-type Si-
(111) is therefore strongly advocated.
A chain mechanism can also rationalize the differences in
efficiency between Si(100) and Si(111). Since the Si(111) is
more flat as compared to Si(100) (see, e.g., Figure 3), the chain
formation on the latter surface will be hampered due to larger
steps between the terraces. As a result the monolayer formation
on Si(111) is more efficient.
Mixed Monolayers. An important issue in the preparation
of well-defined surfaces with highly functionalized materials,
e.g., for application in sensor technology, is the control over
the amount and depth distribution of sensing functionalities in
The reactivity of H-terminated silicon toward 1-alkenes in
the absence of both heat and light was also studied. When a
freshly etched hydrogen-terminated Si(100) sample in an argon-
saturated 1-hexadecene/mesitylene solution was kept in the dark
for 15 h at room temperature, the contact angle of the sample
did not change (∼78°). Since an increase of the contact angle
would indicate monolayer formation and a decrease would imply
oxidation, it can be concluded that the hydrogen-terminated
surface was stable during this experiment. IRRAS showed low
amounts of C-H stretching vibrations. Most likely there is some
attachment due to surface defects/dangling bonds. However, to
obtain a hydrophobic monolayer, light or heat is required.
(53) Eves, B. J.; Sun, Q.-Y.; Lopinski, G. P.; Zuilhof, H. J. Am. Chem. Soc.
2004, 126, 14318-14319.
9
J. AM. CHEM. SOC. VOL. 127, NO. 8, 2005 2521

A R T I C L E S
Sun et al.
Figure 8. Contact angles of n-hexadecyl monolayers as a function of irradiation time for lowly doped n-type Si(100) and highly doped p-type Si(100) using
254, 371, 504, and 658 nm light, respectively.
Scheme 3. Proposed Mechanism for the Initiation of the Light-Induced Reaction
mixed monolayers. In previous work, we revealed the relation
between the ratio of two esterified 1-alkenes in solution and
their relative amounts in the resulting mixed monolayer as
prepared with the visible light method.31 XPS showed a linear
relation between the amount of fluorine in the monolayer and
the mole fraction of the fluorine-containing ester in the solution
used in the preparation of the monolayer, but we couldsin
hindsightsnot fully exclude trace amounts of carbonyl group-
related photochemistry. To allow a completely unambiguous
study of mixed monolayers obtained via both thermal attachment
and visible light attachment, we aimed to keep the system as
stable and as simple as possible by choosing 1-decene and 11-
fluoro-1-undecene as a binary system. The latter compound was
synthesized in two steps from the commercially available alcohol
derivative. For each of the six different ratios that was used a
solution of the two 1-alkenes in mesitylene was divided over
the photochemical and thermal setups, to ensure the comparabil-
ity of the two methods. After the modification of hydrogen-
terminated Si(100), the resulting monolayers were studied with
water contact angles (Figure 9) and XPS (Figure 10).
monolayers prepared with 100% 1-decene (mole fraction of
F-compound, 0) are the same for both preparation techniques.
However, for every mole fraction >0 the contact angle of a
thermally prepared monolayer is slightly higher than that of the
corresponding photochemically prepared monolayer.
According to Cassie,54 the contact angle θ of a drop of water
on a heterogeneous surface made up of patches of two different
types of polarities is related to the contact angle θi of each of
the patches via the following equation:
cos θ ) f1 cos θ1 + f2 cos θ2
(1)
The cosine of the water contact angles of the mixed monolayers
was plotted against the mole fraction (Figure 9, right). The linear
trend lines have high R2 values, indicating that Cassie’s relation
is applicable to this binary system. Two other binary systems
on silicon41,45 that were studied with Cassie’s relation showed
a lower correlation between solution composition and wetting
behavior than the present data. Most likely this is related to the
nature of the functional groups, which differ relatively little
regarding shape, size, and polarity in the present case, and thus
provides a more ideal application for the Cassie relation.
The water contact angles of the mixed monolayers were
plotted against the molecular fraction of 11-fluoro-1-undecene
in the mesitylene solution (Figure 9). The contact angles of
(54) Cassie, A. B. D. Discuss. Faraday Soc. 1948, 3, 11-16.
9
2522 J. AM. CHEM. SOC. VOL. 127, NO. 8, 2005

Covalently Attached Monolayers on Crystalline Silicon
A R T I C L E S
Figure 9. Water contact angle (left) and cosine of the water contact angle (right, linear least-squares fit) of mixed monolayers derived from 11-fluoro-1-
undecene and 1-decene as a function of the mole fraction of 11-fluoro-1-undecene in the mesitylene solution used for the thermal (solid symbols) and visible
light (open symbols) modification of H-terminated Si(100).
relation with the mole fraction of the ester, while the correlation
between surface and solution composition based on contact
angles was clearly different. By stating that the Cassie relation
is debatable, particularly when the surface contains two totally
dispersed species, the authors believe that the spectroscopic data
provide more reliable composition information than the wetting
studies. Our data display a good applicability of Cassie’s law,
but we agree that quantitative information is better obtained
from other sources, such as XPS in the present study.
Conclusions
Well-defined and highly stable organic monolayers on silicon
Figure 10. XPS F1s/Si2p ratio of mixed monolayers of 11-fluoro-1-undecene
surfaces can easily be prepared by irradiating a hydrogen-
terminated silicon wafer with visible light at room temperature
in the presence of a 1-alkene or 1-alkyne. Using a combination
of water contact angles, X-ray reflectivity, X-ray photoelectron
spectroscopy, atomic force microscopy, and infrared reflection-
absorption spectroscopy, it was shown that high-quality mono-
layers can be formed in this manner. The reaction is rather
flexible with regard to the wavelength of irradiation (371 to
larger than 650 nm) and light source, and thus can avoid any
light absorption by the agent that needs to be attached. These
extremely mild conditions are compatible with a very large
variety of biologically active moieties that can be covalently
linked to the reactive alkene or alkyne functionality. This method
thus allows the development of patterned, (bio)active mono-
layers,32 via the combination of (bio)organic chemistry, surface
science, and (nano)lithographic techniques.
and 1-decene as a function of the percentage of 11-fluoro-1-undecene in
the mesitylene solution used for the thermal (solid symbols) and visible
light (open symbols) modification of H-terminated Si(100).
Figure 10 presents the results of XPS measurements (F1s/
Si2p ratio) of the mixed monolayer, and clearly shows a linear
relation between the amount of F on the surface and in solution
for both the photochemical and thermal methods. The presence
of a small amount of F for the 1-decene monolayer (0% 11-
fluoro-1-undecene in solution) is the result of the etching
procedure, as described earlier.32 The data show that the
distribution of two molecules on the surface can be controlled
via their ratio in the solution that is used for the preparation.
However, the slope of the given trend lines is different for the
two preparation techniques, and the visible light method results
according to these XPS data in a grafting density that is about
25% lower than obtained via the thermal method. It is currently
unclear to which degree this is a general feature of the
photochemical method, or whether this is typical for the systems
in the present study, given the observation (Figure 9) that for
this system the contact angles of the photochemically prepared
samples are systematically somewhat lower than for the
thermally prepared samples, which is not the case for other
systems in the present study (e.g., Table 1). Photoreactivity of
specifically the 11-fluoro-1-undecene can be excluded as its
cause, not only because 11-fluoro-1-undecene does not display
any significant light absorption at 447 nm, but also because
gas chromatography analysis of samples before and after
irradiation does not display any photoproducts, and yields a
constant amount of 11-fluoro-1-undecene (in comparison to an
internal standard). More detailed studies are therefore required
to clarify this observation. Very recently, Liu et al. also used
two techniquessIR and water contact angle measurementssto
study mixed ω-carboxyalkyl/alkyl monolayers on Si(111).45 The
absorbance ratio ν(CdO)/νantisymm(CH2) showed a nearly linear
Finally, a mechanism for this reaction was tentatively
proposed, and the use of IRRAS for the study of organic
monolayers was evaluated and proposed as an easy alternative
for attenuated total reflectance IR spectroscopy.
Acknowledgment. We thank the Netherlands Technology
Foundation (CW/STW), ASML, and the Human Frontier
Science Program (HFSP) for financial support, Dr. Maarten
Posthumus for mass spectrometry measurements, and Dr. Greg
Lopinski (NRC, Ottawa, Canada), Dr. Alex Sieval (Nano-C,
Groningen, The Netherlands), and Dr. Dan Wayner (NRC) for
stimulating discussions.
Supporting Information Available: Angle-dependent (0-
80°) IRRAS data for a Si(111) sample modified with tetradecyne
(447 nm, 15 h) and the synthesis of 10-undecenoic acid 2,2,2-
trifluoroethanol ester (compound II) (PDF). This material is
available free of charge via the Internet at http://pubs.acs.org.
JA045359S
9
J. AM. CHEM. SOC. VOL. 127, NO. 8, 2005 2523