Electrochemical Characterization of Si(111) Modified with Linear and Branched AlkylChains

Xiaomin Bin,† Trevor K. Mischki, ‡ Chaoyang Fan,† Gregory P. Lopinski,† andDanial D. M. Wayner* ,†

Steacie Institute for Molecular Sciences, National Research Council, 100 Sussex DriVe, K1A 0R6 Canada, andDepartment of Chemistry, Carleton UniVersity, Ottawa, Ontario, Canada

ReceiVed: January 16, 2007; In Final Form: June 11, 2007

A simple chemical strategy is described to produce branched alkyl chains on Si(111) from the reaction of anester-terminated silicon surface. The stability of the silicon surfaces with linear and branched monolayers ischaracterized by electrochemical impedance, Kelvin probe, and high-resolution electron energy lossspectroscopy (HREELS). The direct observation of surface states in capacitance-voltage plots can be usedto monitor the growth of electrically active defects associated with oxidation of the silicon substrate. We findthat the total surface state density of the freshly made surfaces increases in the following order: Si-B <Si-UDE < Si-C10 (where B is the branched structure, UDE is ethyl undecanoate, and C10 is decyl) inaqueous and organic solvent/electrolyte systems. After 24 h in the electrolyte solution, the surface state densitiesincrease but the order remains the same. The branched structure is significantly more resistant to oxidation.These observations are consistent with the results of other characterization techniques including HREELSand surface photovoltage and indicate that the branched alkyl chain-modified surfaces are considerably morestable, especially in aqueous electrolytes, making them suitable for future use in biological sensor applications.

Introduction

The covalent attachment of organic films on oxide-freesemiconductor surfaces such as silicon is of growing interestfor potential applications from surface passivation to theincorporation of chemical/biochemical functionality at interfacesfor use in biosensors or biosensor arrays.1-7 On the basis ofthe seminal work of Chidsey in which close-packed monolayerswere formed by the reaction of alkenes with the hydrogen-terminated Si(111) surface,8,9 much of the work reported in theliterature has focused on elaborating the organic chemistry fromfilms terminated by unreactive methyl groups4,8,10-13 to othermuch more reactive terminal groups such as acids or esters,13-17

amines,14,18,19alcohols,13,14,20and aromatic rings21-23 that canbe used for further attachment of more complex organic orbioorganic structures. However, because of the mismatchbetween the distance between Si atoms on Si(111) (0.385 nm)and the diameter of an alkyl chain (0.42 nm), it has been obviousthat the alkyl chains cannot pack 1:1 per silicon. It has beensuggested that the (theoretical) maximum coverage of thesemonolayers is 0.5 ML24 (i.e., 50% substitution of the surfacehydrides), although the maximum coverage of the monolayeron silicon has been the subject of some debate.10 Whatever themaximum number of Si-C bonds, it is clear that a significantnumber of unreacted Si-H sites still remain on the surface.Consequently, the surfaces are vulnerable to oxidation by waterand oxygen, which can penetrate through the monolayer. Thisinherent instability is inconsistent with the development ofstable, robust sensors, the application cited by most of thepublished work in this area. The problem is that oxidation of

silicon can result in the formation of electrically active surfacestates that change the electrical properties of the underlyingsilicon. Any strategy that aims toward hybrid silicon-organicdevices should ideally limit the number of oxide-based surfacesstates and, at minimum, stabilize them so they do not changeover time. We report, herein, an approach to increase the extentof alkylation of hydrogen-terminated silicon resulting is a moreelectrically stable Si-C interface.

In a previous publication,14 we demonstrated that an ester-terminated monolayer can react with alkyl Grignards, resultingin an effective doubling of the number of alkyl chains on thesurface. In this paper, we report comparisons of such a branchedsurface (Si-B) with two other unbranched systems, the decyl(Si-C10) and the ethyl undecanoate (Si-UDE) surfaces, onSi(111) surfaces. The surfaces are characterized using electro-chemical impedance and other techniques such as Kelvin probe,ellipsometry, and high-resolution electron energy loss spectros-copy (HREELS). Differential capacity measurements are per-formed in different electrolytes to give the comparison ofdielectric properties and stability of the modified surfaces.Kelvin probe, ellipsometry and high-resolution electron energyloss spectroscopy (HREELS) provide consistent complementaryresults. All the results indicated that the branched alkyl chain-modified surfaces are considerably more resistant to oxidiation,making them suitable for future development of biologicalsensor applications.

Experimental Section

Preparation of Modified Si (111) Surfaces.Decyl (Si-C10), ethyl undecanoate (Si-UDE), and branched alkyl chain(Si-B)-modified Si(111) surfaces were prepared on hydrogen-terminated silicon (Si-H) surfaces. Single side polished silicon(111) shard (1-5Ω cm, n-type, phosphorus doped, thickness

* To whom correspondence should be addressed. E-mail: dan.wayner@nrc.ca.

† National Research Council.‡ Carleton University.

13547J. Phys. Chem. C2007,111,13547-13553

10.1021/jp070354d CCC: $37.00 © 2007 American Chemical SocietyPublished on Web 08/21/2007

) 250( 25µm, Virginia Semiconductor Inc.) was cleaned withpiranha solution (3:1 H2SO4, 96%: H2O2, 30%) at 120°C for30 min, then was rinsed with Milli-Q water. (Warning: Piranhasolutions should be handled with care and kept isolated fromorganic materials). The surface was hydrogen terminated byetching in argon-sparged ammonium fluoride for 15 minfollowed by a brief rinse in degassed Milli-Q water. It was thendried under a nitrogen stream and was transferred into a PyrexSchlenk tube containing ca. 15 mL of deoxygenated decene orethyl undecylenate solution. Either decene (Aldrich, 94%) orethyl undecylenate (Aldrich, 97%) was distilled under vacuumprior to use. The solution was thoroughly purged with argonfor 15 min, then was irradiated for 1.5 h in a Rayonetphotochemical reactor (300 nm). The decyl (Si-C10) or ester-modified (Si-UDE) surface was then rinsed copiously with1,1,1-trichloroethane and was dried under a nitrogen stream.

The preparation of branched alkyl chain-modified Si(111)surfaces (Si-B) is depicted in Scheme 1. The terminal estergroups of Si-UDE surface were converted into di(1-pentyl)carbinol moieties via reaction with pentylmanesium bromide(C5H11MgBr, 2.0M solution in diethyl ether, Aldrich) at roomtemperature for 2 h. The sample was then sonicated in 1% v/vCF3COOH/TCE for 5 min, rinsed with Milli-Q water, and driedunder a nitrogen stream.

Surface Characterization.HREELS.HREELS was carriedout under ultrahigh vacuum conditions with an LK3000 (LKTechnologies) spectrometer. The nominal system resolution wasset to 36 cm-1. This resulted in elastic peak intensities rangingfrom ∼103-104 counts per second with a full width at half-maximum of the elastic peak of∼100 cm-1. HREELS measure-ments of freshly made Si-C10, Si-UDE, and Si-B surfaceswere performed first; the samples were then dipped into Milli-Qwater for 16 h and PBS solution for another 16 h with HREELSmeasurements performed after each stage.

Thickness Measurements.The thickness of organic mono-layers and silicon oxide were estimated using a Gaertner modelL116S ellipsometer with a HeNe laser, and an angle of incidenceof 70°, n ) 1.46 was used as the refractive index of themonolayer and silicon oxide;n ) 3.85 andk ) 0.02 were usedfor the silicon substrate.

Surface PhotoVoltage Measurements (SPV).A Kelvin probe(KP Technology Ltd, Wick, Scotland) was used to measure theSPV, which provides a direct measure of band-bending (thedifference in the surface potential in the dark and underillumination). A white light fiber optic source was used toilluminate the modified silicon surface, ensuring that the lightwas of sufficient intensity to saturate the photovoltage.

Electrochemical Characterization.The differential capaci-tance versus voltage curves were acquired at a frequency of 5kHz with the amplitude of the ac modulation set to 10 mV.The working electrodes, hydrogen-terminated Si(111) surfaceor other modified Si(111) surfaces, were pressed against anopening in the cell bottom (the area is 0.46 cm2). An O-ringwas used to seal the cell. A small modification was made tothe electrochemical cell according to ref 25. We point out thata very thin layer of electrolyte can infiltrate between the O-ringand the silicon sample as described in that reference. Theformation of this thin layer of electrolyte results in a perturbationof the measured potential, which can reach 100mV, as well asthe irreproducibility of effective working electrode area. Simply,to achieve a much improved liquid-tight seal, the O-ring wasreplaced with a machined protrusion at the bottom of the cellthat was finely polished with a SiC sand paper (#4000, Struers).This protrusion was pressed onto the silicon shard that was heldbetween the Teflon cell and a stainless steel plate. Before use,the Teflon parts of the cell were heated in a piranha solution at80 °C for 30 min, then were rinsed with Milli-Q water and driedin the oven before use. To minimize series resistance effectson the contact of the silicon and steel plate, the rear side ofsilicon sample was scratched and etched by a drop of 2% HF,and then was coated with a drop of In-Ga eutectic (Alfa-Aesar,99.99%). The counter electrode was a platinum wire. Thereference electrode was a Pd wire charged in 0.1M H2SO4 at-2.0V for 3 h (∼ -0.25V versus SCE) when the electrolytewas 0.1M H2SO4 or SCE in the other electrolytes. As aprecaution, the SCE was introduced just prior to the measure-ment. It was also possible to use a pseudoreference electrode(silver wire in the solvent/electrolyte) that was calibrated usingthe SCE. The results using either electrode were the same.

Aqueous electrolytes were 0.1M H2SO4 or phosphate-bufferedsaline (PBS, pH) 7.4, a commonly used biological buffer),while the electrolyte for experiments in acetonitrile and dim-ethylformamide (DMF) was 0.1 M tetra-n-butylammoniumperchlorate, Bu4NClO4 (Fluka, electrochemical grade) (TBAP).In general, the electrolyte solution was purged with argon for30 min in the cell before the measurements. A slight positivepressure of argon over the electrolyte was maintained for theduration of the measurement. All measurements were performedin the dark.

Results and Discussion

Film Thickness and Surface Photovoltage of the ModifiedSi(111) Surfaces.The thicknesses of the freshly prepared Si-C10, Si-UDE, and Si-B surfaces measured by ellipsometry

SCHEME 1: Derivatization of Si(111) Surface with Branched Alkyl Monolayers

13548 J. Phys. Chem. C, Vol. 111, No. 36, 2007 Bin et al.

are 10, 13, and 17 Å, respectively (Table 1). The uncertainly is( 1 Å, which comes from at least 5 independent samples. Si-C10 has 10 carbons on its chain, Si-UDE has 13 carbonsincluding the tail ethyl group, and Si-B has 11 carbons in itsinitial chain and 5 carbons in each branched tail. Ellipsometricresults showed that the thickness increases approximately by 1Å per methylene unit. This corresponds to a calculated valueof 35 degrees of the average tilt angle of the axis of thehydrocarbon chain with respect to the surface normal,2,24,32

provided that two short chains maintain the same orientationof the initial chain. IR data also demonstrated that approximately80% of initial chains have been branched by comparing the peakintensities of C-H stretches.33 The reactivity of the ester groupsis ultimately limited by steric factors.

SPV measurements were performed using a Kelvin probe tomeasure the difference in the surface potential in the dark andunder illumination. On a semiconductor surface, surface chargesgive rise to long-range electric fields that penetrate a substantialdistance into the material giving rise to band-bending. Surfacestates, electronic states within the bulk band gap and thereforespatially localized to the surface, can act as charge traps forelectrons (or holes) giving rise to band-bending. Under illumina-tion, photogenerated electron hole pairs act to screen the surfacecharge responsible for any band-bending, flattening the bands(provided that the light intensity is sufficient). The SPV istherefore a measure of the surface charge or density of occupiedsurface states.

Freshly etched H-terminated Si(111) surfaces typically showa very low SPV (<30mV), as expected for surfaces with a lowdensity of surface states. Alkyl-terminated Si(111) is expectedto maintain the minimal band-bending observed on the H/Si-(111) surfaces due to the nonpolar nature of the Si-C bond.The SPVs of freshly made Si-C10, Si-UDE, and Si-B are80, 80, and 44 mV as shown in Table 1. The uncertainty is(20 mV. For the doping density of the silicon used in thepresent experiments (3.5× 1015 cm-2), this degree of band-bending corresponds to a density of trapped charge of<5 ×1010 cm-2. Somewhat surprisingly, fewer charges appear to betrapped on Si-B surfaces than on Si-C10 and Si-UDEsurfaces, although this difference is barely outside of theexperimental uncertainty. As these SPVs are seen to decreaseupon dipping in HF, most of these surface states are attributedto extrinsic oxidation of the interface during the modificationreactions.34 The values in parentheses in Table 1 are the SPVsof the three different modified surfaces after 17 h in water. Thephotovoltage of Si-B surface remains constant (40 mV) whilethose of Si-C10 and Si-UDE surfaces increased significantlyfrom 80 to 160 and 140 mV, respectively. These results indicatethat water or oxygen molecules are capable of penetratingthrough the shorter nonbranched monolayers and oxidizing theSi surface, while the Si-B surface is much more stable, andthe branched alkyl monolayer provides much better blockingproperties.

However, as mentioned above the SPV measures only thequantity of trapped charges (occupied surface states if one state

only traps one charge) on the surface and are typically performedin a dry environment. On the other hand, electrochemicalmeasurements can provide the information about the totalsurface state density at the interface, and they can be performedin solution under conditions that one might expect a sensor toexperience. These results will be discussed in detail in a latersection.

HREELS. HREELS measurements of Si-C10, Si-UDE,and Si-branched modified surfaces are shown in Figure 1a-c,respectively. Black curves are HREELS of freshly madesamples; gray curves represent HREELS of samples after 16 hin water. The vibration signals are nearly the same for the threemodified surfaces, apart from the CdO modes observed on theSi-UDE surface. The peaks in Figure 1 are assigned as thefollowing: 2920 cm-1 ) C-H asymmetric and symmetricstretching (due to the resolution limitation, these two modesare merged), 2080 cm-1 ) Si-H stretching, 1700 cm-1 ) CdO stretching, 1400 cm-1 ) symmetric and asymmetric C-Hbending overlapping, and 1070 cm-1 ) Si-O-Si stretching.For the purpose of this work, we focus on changes in the Si-O-Si stretches as a monitor of the stability of the three modifiedsurfaces. As shown in Figure 1, the intensity of the Si-O-Sipeak on the Si-C10 surface increased significantly after beingin water for 16 h. This indicates that water and/or oxygenmolecules can penetrate the monolayer and insert into siliconbackbonds. In the case of the Si-UDE surface, although theC-O stretch mode in this region overlaps with the Si-O peak,it is reasonable to assume peak intensity of the C-O stretch

TABLE 1: Thickness and Photovoltage of Modified Si(111)Surfaces

surface thickness/Åa photovoltageb/mV

Si-C10 10 80 (160)Si-UDE 13 80 (140)Si-B 17 44 (41)

a Determined by elliposmetry, uncertainty(1 Å. bDetermined byKelvin probe, uncertainty(20 mV; the numbers in brackets aremeasured SPVs after the samples are in water for 17 h.

Figure 1. HREELS of (a) Si-C10, (b) Si-UDE, and (c) Si-Bsurfaces. (black) Freshly made surface; (gray) after 16 h in water.

Si(111) Modified with Linear and Branched Alkyl Chains J. Phys. Chem. C, Vol. 111, No. 36, 200713549

does not change during the whole process. As a result, the netincrease of peak intensity can be attributed to increased oxidationof the surface. Compared to the Si-C10 surface, the oxidationof Si-UDE is slower in water. For the Si-B surface, the Si-Opeak grows considerably slower in water compared to other twomodified surfaces. HREELS measurements of these threesurfaces after 16 h in PBS solution (the samples have alreadybeen in water for 16 h) were performed and showed nosignificant changes compared with those after 16 h in water.The above results show the branched alkyl surface is the mostoxidation resistant surface, which is consistent with results ofthe above Kelvin probe measurements and later electrochemicalmeasurements (vide infra).

Electrochemical Characterization. Electrochemical ca-pacitance was used to characterize the electrical properties ofthe hydrogen-terminated Si(111) surface and modifiedSi(111) surfaces. It is well known that the interface between asemiconductor and an electrolyte behaves like a capa-citor.23,26-28 The electrical nature of the semiconductor/electrolyte interface region, described by an equivalent circuitof an array of resistors and capacitors that correspond to thephysical structure of the interface, has been publishedelsewhere.21-23

The applied potentialE drops across the interface within thespace charge layer (Vsc), the organic (Vm), and the Helmholtz(VH) layers. The variations of the different componentsreferenced with respect to the flat band potentialEfb, asfollowing:

where

∆Vm,H represents the potential drop across the organic layer+Helmholtz layer,Cm,H is the total capacitance of the organiclayer and Helmholtz layer,CH is the capacitance of theHelmholtz layer,Cm is the capacitance of the monolayer, andQ is the charge at the interface. By charge neutrality,

The overall capacity,C, is related to each of the capacitiesaccording to

Using eqs 2 and 3, we obtain

Thus, the parallel combination ofCsc andCss is in series withcapacitors of the organic and Helmholtz layers. Under appropri-ate conditions of frequency and potential, the overall capacitancecan be simplified and determined by the smallest capacitor. Thecapacitance of the Helmholtz layerCH can be obtained bymeasuring the capacitance of the Si-H/electrolyte interfacewithout the presence of the organic monolayer. In the cathodicrange, where the n-type semiconductor is in accumulation, thespace charge capacitance exponentially increases as the electrode

potential shifts in a negative direction

Therefore, in the case whereE is sufficiently negative ofEfb,Csc

-1 can be neglected, andCH dominates the overall capaci-tance. Furthermore, if it is assumed that the Helmholtz doublelayer capacitanceCH determined from the Si-H/electrolyteinterface is independent of the organic monolayer, the capaci-tance of the organic monolayer (Cm) can be evaluated by eq 5in the presence of minimal surface states.

However, as mentioned above, the purpose of modifying thesilicon surface is to effectively protect the surface from oxidationthat causes surface states. Thus, the change in the density ofsurface states as a function of time is an indicator of how wellthe surfaces are protected. In a capacitance-voltage plot, thepresence of surface states is indicated by the appearance of anextra capacitance peak at electrode potentials near the flat bandpotential. Searson and co-workers29-31 performed the quantita-tive analysis of the energetics and kinetics of surface states atn-type silicon surfaces in aqueous solutions. If no redox coupleis present in the solution (so that transfer of electrons betweenthe surface states and an electron acceptor in the solution canbe neglected), the surface states can only exhange electrons withwith the conduction band (CB)

where SS0 is the empty state andSS- the filled state. Theoccupancy of the surface states is determined by the Fermi level,and the surface state capacitance can be expressed as

whereCp(ω) is the parallel capacitance, the sum of the spacecharge capacitance,Csc, and the pseudocapacitance due to thesurface states.ns is the surface electron concentration,ω is thefrequency of the potential or current modulation, andk1 andk2

are the rate constants for the forward and reverse reactionsdescribed by eq 7.Stot is the total density of the surfaces states.Experimentally, the component of the capacitance due to thesurface states can be obtained by subtraction of the underlyingspace charge capacitance. At low frequency whereω2 , (k1ns

+ k2)2, Cp(ω) exhibits a maximum as a function of potentialfrom which the density of surface states can be determined

Differential capacitance as a function of potential (C-Ecurves) for the hydrogen-terminated (Si-H), as well as the C10,UDE, and branched surfaces as measured in four differentelectrolytes (0.1M H2SO4, 0.1 M PBS, 0.1M TBAP in CH3-CN, and 0.1M TBAP in DMF), are shown in Figure 2a-d,respectively. The applied potential was scanned in the cathodicdirection, and it was limited by the onset of hydrogen evolutionat negative potentials and oxidation of silicon substrate atpositive potentials, ensuring that no Faraday current flows duringthe measurement. All the potentials in theC-E curves werenormalized to the flat band potential. A typicalC-E curvecontains three regions: the depletion region, the flat bandpotential region, and the accumulation region. From each region,we can obtain critical information about the electrical properties,

∆V ) E - Efb ) ∆Vsc + ∆Vm,H (1)

∆Vm,H )Qsoln

Cm,Hand Cm,H )

CHCm

CH + Cm(2)

Qsol + Qsc + Qss) 0 (3)

1C

) -∆V(∆Qsc + ∆Qss)

)-(∆Vsc + ∆Vm,H)

(∆Qsc + ∆Qss))

-∆Vsc

(∆Qsc + ∆Qss)+

∆Vm,H

∆Qsol(4)

C-1 ) (Csc + Css)-1 + Cm

-1 + CH-1 (5)

Csc ) (q2εε0Nd

2kT ) exp(-q(E - Efb)

2kT ) (6)

SS0 + e(CB) y\zk1

k2SS- (7)

Cp(ω) ) e2/kTk1k2ns/[ω2 + (k1ns + k2)

2]Stot (8)

Cp(max)) 14

e2

kTStot (9)

13550 J. Phys. Chem. C, Vol. 111, No. 36, 2007 Bin et al.

such as the flat band potential (Efb), the density of surface states(Stot), and the capacitance of the organic monolayer on thesurface. The comparison of these properties on the differentmodified silicon surfaces in different electrolyte solutions andtheir relationships are discussed in detail as below.

A: Flat Band Potentials(Efb). The flat band potentials wereobtained from the Mott-Schottky relations when the spacecharge layer is in depletion

A linear Csc-2-E plot is obtained in the depletion region, the

intercept of which enables the flat band potential. The dopingdensity,Nd ) (3.5 ( 0.7) × 1015 cm-3 can also be calculatedfrom the slope of this linear relationship. It was found that thevalue of the doping density does not depend on the modificationof the silicon surface or the electrolyte used in the measurements(Figure 3). It is also consistent with the doping density derivedfrom four-probe resistivity measurements of the same sample.The flat band potentials of different modified silicon surfacesin four differential solutions are shown in Table 2. With theexception of the results in PBS for which surface oxidation isexpected to be a problem for unprotected surfaces, the flat bandpotentials do not depend on the structure of the monolayer. InPBS, the Si-C10 surface is significantly shifted from the othertwo, presumably as a result of oxidation (Figure 1). A moredetailed analysis of the flat band potentials in comparison tothe Kelvin probe measurements in air will be reported later. Inthe context of this paper, the flat band potentials allow theenergies of the surface states in the gap to be compared. It shouldbe noted that the data for the capacitance plot and the Mott-Schottky are obtained simultaneously. The Mott-Schottky anddifferential capacitance plots are reproducible over a period ofat least an hour provided that the applied potentials are lessthan about 0.5 V positive of the flat band potential (i.e., belowthe known electrochemical oxidation threshold of the surface).

B: Differential Capacitance. The differential capacity-electrode potential curves of Si-C10, Si-UDE, and Si-B

surfaces in different electrolyte solutions show a clear trend asdisplayed in Figure 2. The total capacitance in the accumulationregion is seen to become gradually depressed as the monolayerbecame thicker. Equation 5 can be used to evaluate thecapacitance of the monolayer (Cm) on the Si electrode surfaces.Under accumulation, the capacitance-voltage curves show asmall saturated plateau that is determined only by the capaci-tance of the Helmholtz layer (CH) for Si-H surface, or the totalcapacitance ofCH and Cm for the modified Si surfaces.

Figure 2. Differential capacitance of (-0-) Si-C10, (-O-) Si-UDE, and (-4-) Si-B surfaces in (a) 0.1 M H2SO4, (b) PBS, (c) 0.1 M TBAP inCH3CN, and (d) 0.1 M TBAP in DMF.

1

Csc2

) 2

qεε0NdA2 (E - Efb - kT

q ) (10)

Figure 3. Mott-Shottky curves of Si-UDE surfaces in 0.1 M H2SO4

(0), PBS(O), and 0.1 M TBAP in CH3CN (4).

TABLE 2: Flat Band Potentials of Different Modified SiSurfaces at Different Electrolytes (V vs SCE)a

Si-C10 Si-UDE Si-B

0.1 M H2SO4 -0.53 -0.59 -0.55PBS -0.80 -0.64 -0.640.1 M TBAP in CH3CN -0.46 -0.48 -0.500.1 M TBAP in DMF -0.47 -0.46 -0.49

a Fitted by Mott-Schottky curves fromC-U measurements, uncer-tainty (0.03 V.

Si(111) Modified with Linear and Branched Alkyl Chains J. Phys. Chem. C, Vol. 111, No. 36, 200713551

Extraction of the monolayer capacitance (Cm) from this plateauvalue requires interfaces with a low density of surface statesand the assumption that the values ofCH are independent ofthe presence or composition of the monolayer.23 Table 3 liststhe values of the total capacitance and the derivedCm of threemodified silicon surfaces in 0.1 M H2SO4 and PBS solution.The capacitance of the Helmholtz layer, 4.0µF/cm2 in 0.1MH2SO4 and 3.0µF/cm2 in PBS solution, was used to calculatetheCm of the organic monolayer-modified silicon surfaces. Thevalues ofCm of Si-C10 surfaces in this study are somewhathigher than a previously reported value in the literature. Forexample, Yu35 obtained a value ofCm of 3.3 ( 0.6 µF/cm2 fora decyl monolayer made via the reaction of an alkyl Grignardwith H/Si(111). This difference may be due to the differentpreparation methods producing surfaces with different totalcoverage of the organic.

The capacitive behavior of an organic film on silicon hasbeen described above as that of an ideal capacitor for whichthe reciprocal capacitance of the unit area is give by

whered, εm, andε0 correspond to the thickness and the dielectricconstant of the organic monolayer and the permittivity of thevacuum, respectively. The values ofεm listed in Table 3 indicatethat there are vacancies in the film through which the solvent/electrolyte can penetrate. However, compared to the decyl- andUDE-terminated surfaces, the branched alky surface shows amuch better blocking property and dielectric constants muchcloser to the expected value for polyethylene-like film.35

We have noted that the capacitance properties of modifiedsilicon surfaces in organic solvents (CH3CN and DMF) showedthe same trend as those in aqueous electrolytes. For the sametype of modified silicon surface, the total capacitance underaccumulation deceases as the electrolyte changes: 0.1 M H2-SO4 > PBS> 0.1M TBAP in CH3CN > 0.1M TBAP in DMF.CH3CN and DMF have almost the same dielectric constants(37.5); however, it is difficult to remove traces of water fromthe CH3CN. The values ofCm in organic solvents cannot becalculated because theC-E curves of Si-H surfaces in CH3-CN and DMF showed a very high surface state peak (not shownin the figure) while the modified surfaces showed relatively lowsurface states so it was not possible to estimateCH (presumablydue to a specific interaction between the solvent and the Si-Hsurface). However, the similar trend shown inC-E curves inorganic solvents indicates the branched alkyl monolayer on thesilicon electrode has a better quality than the decyl or UDEmonolayer.

C: Surface State Densities.A surface state pseudocapacitancepeak appears inC-E curves near the flat band potential. Thenegative end of this peak includes the increase of capacitanceof the space charge layer. To separate parallel capacitance dueto surface states from the underlying space charge layercapacitance, a baseline subtraction was made for eachC-Ecurve; an example is shown in Figure 4. From the maximumof the obtained surface state peak curves, the total density of

the surface states can be obtained according to eq 9. Table 4shows the total density of surface states of Si-C10, Si-UDE,and Si-B surfaces in different electrolyte solutions. The valuesin parentheses represent samples stored in the electrolyte for24 h. The density of the surface states is in the range of 1×1010 cm-2 to 2 × 1011 cm-2.

As shown in Table 4, the total density of surface states fordifferent modified surfaces has the same order in the fourdifferent electrolyte solutions: Si-C10 > Si-UDE > Si-B.For the same modified surface, the sequence of the total densityof surface states in different electrolytes is: PBS> 0.1 M TBAPin CH3CN > 0.1 M H2SO4 > 0.1 M TBAP in DMF. Thedifference between 0.1 M H2SO4 and PBS is consistent withthe fact that oxidation of silicon is inhibited at low pH. Thedifference between acetonitrile and DMF may be related to thedifficulty in obtaining anhydrous acetonitrile. Unless extremeefforts are made, acetonitrile distilled from calcium hydride(typical procedure) contains traces of water that may penetratethe monolayer to oxidize the surface. The stability of themodified silicon surfaces was evaluated by determining thesurface state density after 24 h in the electrolyte (values inparentheses). These results are consistent with the Kelvin probedata (Table 1), that is, the order of stability is Si-C10 < Si-UDE < Si-B.

It is worthwhile briefly discussing the comparison of the SPV,HREELS, and electrochemical measurements. These measure-

TABLE 3: Differential Capacitance-Potential Properties of Si-C10, Si-UDE, and Si-B Surfaces in 0.1 M H2SO4 and PBS

0.1 M H2SO4 PBS

Ctot/µF/cm2 Cm/µF/cm2 εma Ctot/µF/cm2 Cm/µF/cm2 εm

a

Si-C10 2.5( 0.5 5.7( 1.0 6.5( 1.1 1.8( 0.4 4.4( 0.8 5.0( 0.9Si-UDE 2.0( 0.4 3.5( 0.8 5.2( 1.1 1.5( 0.3 3.1( 0.6 4.5( 0.9Si-B 1.1( 0.2 1.5( 0.4 2.8( 0.7 1.1( 0.2 1.8( 0.4 3.4( 0.7

a It is assumed that the thickness of the film, determined by air by ellipsometry, does not change in the different solvent/electrolyte systems.

Cm-1 ) d/(εmε0) (11)

Figure 4. Capacitance plot of Si-C10 in PBS, pH 7.4, showing thecapacitance peak after background subtraction.

TABLE 4: The Total Surface States (Stot, 1010 cm-2) ofModified Surfacesa

solvent/electrolyte Si-C10 Si-UDE Si-B

0.1 M H2SO4 7.2 (11.3) 1.6 (4.9) 1.4 (2.3)PBS 9.3 (16.8) 7.7 (16.7) 5.9 (8.6)0.1 M TBAP in CH3CN 8.2 (14.5) 6.5 (8.5) 3.8 (5.6)0.1 M TBAP in DMF 6.7 (10.2) 0.7 (4.4) 0.4 (1.5)

a The values in the parentheses areStot values measured on the secondday (24 h).

13552 J. Phys. Chem. C, Vol. 111, No. 36, 2007 Bin et al.

ments are all made under very different conditions: in ambientair, ultrahigh vacuum, and electrolyte solutions, respectively.There is no reason, a priori, to expect that the origin of themeasured effects are all a result of the formation of oxides or,if they are, that the origin of the oxides are the same. However,the correlation of the data from the three measurements is atleast a compelling argument that oxide formation (whatever themechanism of formation) is responsible for the increase insurface state density and an increase in the SPV. Furtherevidence in support of this can be obtained by dipping thesurfaces in 2% HF following exposures to water or electrolyte.For example, a decyl-terminated silicon surface with an initialSPV of 80 mV increases to 160 mV on exposure to water (Table1). Upon dipping the surface in 2% HF, the SPV recovers to80 mV without significant loss of alkyl chains on the surface.The SPV of the branched surface does not change upon dippingin HF, again confirming that these surfaces are protected fromthe solvent/electrolyte.

It is interesting at this time to compare these results with theresults recently published by Gorostiza et al.36 In that study,alkyl monolayers were formed by Lewis acid catalysis with Et2-AlCl. It was reported that this leads to films with void defectsless than 2.82 Å in diameter. The effective dielectric constantof the films formed in that process are similar to those of thebranched chains reported in this work. Quantitative comparisonsare difficult because the stability studies were carried out underdifferent conditions (storage in water in our study versus storagein air in the Gorostiza work). While both approaches providehigh quality surfaces, each have limitations with respect tofunctional group compatibility. An advantage of branching usingGrignard reagents and terminal esters is that reactive end-groupscan be introduced to allow simple attachment of biomolecules,for example, as orthoacids (protected acids), acetals (protectedaldehydes), or alkenes (suitable for epoxidation). Perhaps acombination of both approaches will provide the ultimateinterface for biosensor applications.

Conclusions

In this paper, we have presented a detailed characterizationof decyl, ethyl undecanoate, and branched alkyl monolayer-modified silicon surfaces using electrochemical measurements.Flat band potentials, dielectric properties, and surface statesanalysis from the differential capacitance-potential curvesshowed that the branched alkyl monolayer with higher alkyldensity provides excellent chemical and electrical passivationto the silicon electrode surfaces. Kelvin probe and HREELSmeasurements provide the complementary consistent results. Themuch better stability of the branched alkyl chain monolayer-modified surfaces provides an alternative to the formation oforganic monolayers suitable for incorporation into silicon-organic hybrid sensors.

Acknowledgment. This work was partially supported by theNRC Genomics and Health Initiative (GHI). We thank DolfLandheer (NRC Institute for Microstructural Sciences) forhelpful comments on the manuscript.

References and Notes

(1) Yates, J. T., Jr.Science1998, 279, 335.(2) Wayner, D. D. M.; Wolkow, R. A.J. Chem. Soc., Perkin Trans.

2002, 2, 23.(3) Buriak, J. M.Chem. ReV. 2002, 102, 1271.(4) Boukherroub, R.; Morin, S.; Wayner, D. D. M.; Bensebaa, F. G.;

Sproule, I. J.; Baribeau, M.; Lockwood, D. J.Chem. Mater.2001, 13, 2002.(5) Wojtyk, J. T. C.; Tomietto, M.; Boukherroub, R.; Wayner, D. D.

M. J. Am. Chem. Soc.2001, 123, 1535.(6) Bent, S. F.J. Phys. Chem. B2002, 106, 2830.(7) Lasseter, T. L.; Clare, B. H.; Abbott, N. L.; Hamers, R. J.J. Am.

Chem. Soc.2004, 126, 10220.(8) Linford, M. R.; Chidsey, C. E. D.J. Am. Chem. Soc.1993, 115,

12631.(9) Linford, M. R.; Fenter, P.; Eisenberger, M.; Chidsey, C. E. D.J.

Am. Chem. Soc.1995, 117, 3145.(10) Cicero, R. L.; Linford, M. R.; Chidsey, C. E. D.Langmuir2000,

16, 5688.(11) Wagner, P.; Nock, S.; Spudich, J. A.; Volkmuth, W. D.; Chu, S.;

Cicero, R. L.; Wade, C. P.; Chidsey, C. E. D.J. Struct. Biol.1997, 119,189.

(12) Boukherroub, R.; Morin, S.; Bensebaa, F.; Wayner, D. D. M.Langmuir1999, 15, 3831.

(13) Sieval, A. B.; Demirel, A. L.; Nissink, J. W. M.; Linford, M. R.;van der Mass, J. H.; de Jeu, W. H.; Zuilhof, H.; Sudholter, E. J. R. Langmuir1998, 14, 1759.

(14) Boukherroub, R.; Wayner, D. D. M.J. Am. Chem. Soc. 1999, 121,11513.

(15) Mitchell, S. A.; Ward, T. R.; Wayner, D. D. M.; Lopinski, G. P.J.Phys. Chem. B2002, 106, 9873.

(16) Wojtyk, J. T. C.; Morin, K. A.; Boukherroub, R.; Wayner, D. D.M. Langmuir2002, 18, 6081.

(17) Zhang, L.; Li, L.; Chen, S.; Jiang, S.Langmuir, 2002, 18, 5448.(18) Sieval, A. B.; Linke, R.; Heij, G.; Meijer, G.; Zuilhof, H.; Sudholter,

E. J. R.Langmuir2001, 17, 7554.(19) Hart, B. R.; Letant, S. E.; Kane, S. R.; Hadi, M. Z.; Shields, S. J.;

Reynolds, J. G.Chem. Commun.2003, 322.(20) Sieval, A. B.; Vleeming, V.; Zuilhof, H.; Sudholter, E. J. R.

Langmuir1999, 15, 8288.(21) Henry de Villeneuve, C.; Pinson, J.; Bernard, M. C.; Allongue, P.

J. Phys, Chem. B.1997, 101, 2415.(22) Allongue, P.; Henry de Villeneuve, C.; Pinson, J.; Ozanam, F.;

Chazalviel, J. N.; Wallart, X.Electrochim. Acta.1998, 43, 2791.(23) Allongue, P.; Henry de Villeneuve, C.; Pinson, J.Electrochim. Acta.

2000, 45, 3241.(24) Sieval, A. B.; van den Hout, B.; Zuilhof, H.; Sudholdter, E. J. R.

Langmuir2000, 17, 2172.(25) Bertagna, V.; Rouelle, F.; Chemla, M.J. Appl. Electrochem.1997,

27, 1179.(26) Zhang, X. G.Electrochemistry of Silicon and its Oxide; Kluwer

Academic, New York, 1992; Chapter 1.(27) Chazalviel, J. N.Electrochim. Acta. 1990, 35, 1545.(28) Allongue, P.; Henry de Villeneuve, C.; Cherouvrier, G.; Cortes,

R.; Bernard, M. C.J. Electroanal. Chem.2003, 161, 550.(29) Oskam, G.; Hoffmann, P. M.; Schmidt, J. C.; Searson, P. C.J.

Phys. Chem.1996, 100, 1801.(30) Oskam, G.; Hoffmann, P. M.; Searson, P. C.Phys. ReV. Lett. 1996,

76, 1521.(31) Natarajan, A.; Searson, P. C.J. Phys. Chem. B1998, 102, 7793.(32) Sieval, A. B.; van den Hout, B.; Zuilhof, H.; Sudholdter, E. J. R.

Langmuir2000, 16, 2987.(33) Mischki, T. K., Carleton University, unpublished data.(34) Lopinski, G. P.; Wayner, D. D. M. InProperties of Single Organic

Molecules on Crystal Surfaces; Grutter, P., Hofer, W., Rosei, F., Eds.;Imperial College Press: London, U.K., 2006, 287-331.

(35) Yu, H. -Z.; Morin, S.; Wayner, D. D. M.; Allongue, P.; Henry deVilleneuve, C.J. Phys. Chem. B2000, 104, 11157.

(36) Gorostiza, P.; Henry de Villeneuve, C.; Sun, Q. Y.; Sanz, F.;Wallart, X.; Boukherroub, R.; Allongue, P.J. Phys. Chem. B2006, 110,5576.

Si(111) Modified with Linear and Branched Alkyl Chains J. Phys. Chem. C, Vol. 111, No. 36, 200713553