Cysteamine-Based Functionalization of InAs Surfaces: Revealing the

ARTICLE pubs.acs.org/Langmuir

Cysteamine-Based Functionalization of InAs Surfaces: Revealing the Critical Role of Oxide Interactions in Biasing Attachment Maria Losurdo,*,†,‡ Pae C. Wu,† Tong-Ho Kim,† Giovanni Bruno,‡ and April S. Brown*,† † ‡

Electrical and Computer Engineering Department, Duke University, Durham, North Carolina 27708, United States Institute of Inorganic Methodologies and of Plasmas, IMIP-CNR, via Orabona 4, 70126 Bari, Italy ABSTRACT: Attaching functional molecules such as thiols and proteins to semiconductor surfaces is increasingly exploited in functional devices such as sensors. Despite extensive research to understand this interface and demonstrate a robust protocol for attachment, the bonding chemistry of thiolates to III
V surfaces has been under great debate in the literature. This study provides a comprehensive chemical model for the attachment of thiols to InAs, an increasingly device-relevant III
V semiconductor, using cysteamine as a model molecule. We examine the attachment of cysteamine to InAs via the thiol group using X-ray photoelectron spectroscopy and spectroscopic ellipsometry and confirm that thiolate bonding to the substrate occurs preferentially to As sites over In sites as a limit. These experiments explore the interplay of the native oxide chemical properties, the cysteamine concentration, and the evolving InAs surface chemistry with functionalization. The thiol
InAs interaction can be framed as a general acid
base reaction, where the nucleophilic and/or electrophilic attack of the surface (i.e., binding to In sites and/or As sites) depends on the acidity of the thiol. The roles of the initial oxide composition, the solvent of the functionalizing solution, and the cysteamine as a limiting reagent in fully displacing the oxide and creating In
S and As
S bonds are highlighted.

1. INTRODUCTION Functional semiconductor-based devices such as sensors typically rely on surface-attached functional molecules to transduce selective interactions through the inorganic
organic, or hybrid, interface. Si-based hybrid structures have been extensively explored in order to leverage the maturity of Si-based microelectronics.1
4 The III
V compound semiconductors, however, possess numerous favorable electronic and optical properties for sensors, such as their direct band gap (e.g., GaAs and GaN) and high (∼ 10,000 cm2/V s) electron mobility (e.g., InGaAs, InAs, and InSb). Compared to Si, a key barrier to developing hybrid III
V devices is the relative lack of chemical and electronic control of their surface properties due to the complex multiphase native oxide that evolves quickly under ambient conditions on III
V surfaces.5 Consequently, the approach most frequently reported in the literature adopts chemistries based on the attachment to oxide-free III
V surfaces to form a self-assembled monolayer (SAM) that chemically and electronically passivates the surface to prevent oxide regrowth. These passivation layers typically comprise sulfur-based molecules6
11 and thiol tethers.6,12
20 Despite extensive research to understand this interface and demonstrate a robust protocol for attachment, significant debate within the literature describing the bonding chemistry of thiolates to GaAs surfaces persists.21 Two key issues are the extent to which the native oxide is completely removed prior to functionalization and that the functionalization r 2011 American Chemical Society

chemistry conditions do not allow for oxide regrowth. Allara12,21 articulated the need for stringent deoxygenation of the functionalization solution such that oxide formation does not compete with thiol attachment. Another key issue relates to the control of site attachment: does the thiol attach to both the III-cationic and the V-anionic sites and what properties of the interactions drive attachment differences? In postfunctionalization, the electronic properties of the interface are critical. For example, a question remains as to which interface phase, As
S versus As
O and/or Ga
S versus Ga
O, reduces the carrier recombination rate on etched GaAs surfaces.6,22 In addition to the oxide removal process, numerous semiconductor surface parameters, such as the surface stoichiometry and reconstruction, surface roughness, and doping, can impact S binding. The functionalizing chemistry conditions, such as the choice of solvent, the pH of the solution, and its temperature, are also critical. We have summarized the results from the literature in Table 1 to highlight the complex conditions that have already been studied for functionalization. From the table, it can be inferred that Na2S 3 9H2O(aq) solutions (over a broad range of pH from 7 to 14) and a large variety of thiols in nonaqueous solutions result in effective As
S binding, independent of the Received: September 1, 2011 Revised: November 29, 2011 Published: December 01, 2011 1235

dx.doi.org/10.1021/la203436r | Langmuir 2012, 28, 1235–1245

1-dodecanethiol (C12H25SH),

1,9-nonanedithiol

GaAs (100) As/Ga surface

1236

stoichiometric

GaAs(100) GaAs(100)nearly

GaAs(100)

heterodimers

(4  6) GaAs Ga
As


As-enriched


As-depleted
stoichiometric

stoichiometric GaAs(100)


Na2S(aq)*

4-Cl-thiophenol in CCl4


1,2-dithiolethane and

thiophenol(anarobic conditions)

(HS(CH2)3
Si(OCH3)3

3-mercaptopropyl)trimethoxysilane

vapors (no solution)

CH3SH, C2H5SH, C4H9SH, CH3SSCH3

1-octadecanethiol (solution) 1-octadecanethiol (solution)

1-octadecanethiol (vapor phase)

in ethanol

CH3
(C6H4)2
SH), HO
(C6H4)2
SH)

octadecylthiol (CH3(CH2)17SH] in ethanol

GaAs(100)

ratio of 0.85

octadecylthiol

As-rich GaAs

the addition of NH4OH

1-octanethiol (C8H17SH), 1-butanethiol (C4H9SH) in ethanol with

1-octadecanethiol (C18H37SH),

GaAs(100)

hexadecanethiol (CH3[CH2]15SH)

octadecane thiol (CH3[CH2]17SH),

S source

stoichiometric

GaAs(100)

GaAs

III
V

As
-rich GaAs surfaces Aso-rich surface, no oxide, oxidized

oxide-free after etching




Ga2O3

Ga2O3, zAs2O3(main), As2O5

no residual oxide


HCl/EtOH (1/10) for 1 min


30% H2O2


Br2-CH3OH/KOH

1:l:100 H2SO4/H2O2/H2O

1.0 M HSC6H4CI

HCl for 1 min

immersed in concentrated

(8:1:1) for 5 s and then

etched in H2O/H2O2/NH4OH

MBE-prepared


HF etching


H2-cleaned


in HCl for 1 min

oxide removed by etching

30 s in NH4OH

and then dipping for

etching for 5 s in 2% HF,

UV/ozone oxidation for 10 min,


photochemical etching


H2O2/H2SO4/H2O (1/1/100)


HCl for 1 min


P etching yields an As
-rich layer

HCl or photochemical etching

etched in 35% HCl for 1 min

with NH4OH

the native oxide was removed

removal of oxide


stochiometric

As
-rich




As2O3 > As2O5, Ga2O3

oxide information

Ga2S3 Ga
S


Ga
S

Ga
S

Ga
S Ga
S

Ga
S









As
S*

As
S

As
S

As
S As
S




As
S

As
S

As
S As
S




As
S

As
S










(or In
S)

Ga
S As
S

Table 1. Literature Reporting Ga
S, In
S, and As
S Bond Formation Depending on the S-Source Molecule, Oxide Present on the Surface, and Etching Procedure

22 21

20

19

18

17

16

15

14

13

12

ref

Langmuir ARTICLE

dx.doi.org/10.1021/la203436r |Langmuir 2012, 28, 1235–1245

(NH4)2S

(NH4)2S

hot (NH4)2S solution

(NH4)2S at 60
C

GaAs (100)

GaAs(100)

GaAs(100)

1237

undoped InAs(100)

InAs(100)

peptides solutions in PBS

mixture of NH4OH

cysteamine (HSC2H4NH2) in a 1:9

in NH4OH

octadecanethiol in HCl/isopropanol octadecanethiol CH3
(CH2)17
SH

NH4OH (basic solution)

(no initial oxide information)

oxide-free

oxide-free oxide-free

the indium and arsenic oxide layers

shows a significant amount of As
but completely removed

Br2- or HCl-based etching

deprotonates the thiol group

oxide and simultaneously

solution strips the native

the alkaline component of the








In
S

In
S In
S

oxide-free

InAs(100) In-terminated InAs(001)

In
S In
S





oxide-free

As
S

As
S











at 200
C

thioacetamide (CH3CSNH2) and

As
S broken at 500
C

disordered As
S

oxide after H2O rinsing

As
S converts to

Ga
S broken

ordered bridge Ga
S in the [011] azimuth

after rinsing in H2O

Ga
S remains only

As
S As
S stableat 200
C

not present

As
S phase, As/S > 0.66

As
S

Ga
S stable at 400
C

thioacetamide (CH3CSNH2)







H2SO4
H2O2 solution

Ga
S (or In
S)

not present

InAs(001)








etched in a mild




removed by Na2S*9H2O(aq)

As0, As oxides, and Ga oxides

removal of oxide




As0, As2O3, As2O5, Ga2O3

oxide information

InAs epigrown by MBE

InAs

Na2S 3 9H2O(aq)

GaAs(100)

Na2S*9H2O(aq)(anarobic conditions)

S source

GaAs(100)

GaAs GaAs(100)

III
V

Table 1. Continued

31

30

28 29

27

26

11

10

9

8

7

6

ref

Langmuir ARTICLE

dx.doi.org/10.1021/la203436r |Langmuir 2012, 28, 1235–1245

Langmuir initial stoichiometry of the surface and oxide removal procedure. Conversely, (NH4)2S solutions and short-chain thiols result in mixed Ga
S/As
S bonding. Finally, more thiols possessing acidic character such as thiophenol result primarily in Ga
S attachment. Thiol-based GaAs functionalization can therefore be framed in terms of the relative hardness (softness) of the molecules governing the surface reactions and the involved electronic structure modifications.23 Specifically, hard acids (e.g., thiophenol) tend to react with ionic and polarizable sites (e.g., Ga) whereas soft bases (as determined by the increase in the thiol species pKa with increasing numbers of C atoms in the molecule) tend to form covalent bonds with less-polarizable sites such as As. However, a different framework appears to govern S attachment to InAs (100) surfaces because both inorganic24,25 and thiolbased SAMs26
30 have been shown to attach to InAs primarily through In
S bonds. Although GaAs is widely considered to be the prototypical III
V semiconductor, InAs is emerging to be of special interest for numerous devices because of its unique electronic properties. Since the seminal work of Tsui31 in which InAs “intrinsic” surface-confined 2D electron gas (2DEG) was observed and characterized, its presence and its high electron mobility have been extensively exploited for electronic devices and increasingly for sensors and spintronic applications. Surface-confined 2DEG is crucial to enabling highly sensitive sensors because of the intimate coupling between the 2DEG concentration, which is in the mid-1012 cm
2 range, and surface events, including surface adsorption and molecule-based affinity interactions. The efficacy of a hybrid sensor therefore depends upon 2DEG and the degree to which its concentration can be modulated. We refer to 2DEG as intrinsic because it is known to originate from surface states with a donor nature yet its defect-based origin remains unknown and in fact may relate to defects present in or at the interface with its oxide.32 Therefore, the functionalization processes and the fate of the InAs oxide may also be crucial in terms of electronic device performance as well as its chemical passivation and stability. Understanding the impact of III
V oxides on the surface electronic properties remains a very active area of research. Recent studies reveal complex relationships that are not always deleterious to device performance as is typically assumed. For example, two recent studies33,34 show that an InAs low-temperature (300
C) thermal oxide provides an excellent lowdefect-density interface that is useful for electronic devices. In addition, the surface electronic properties of InAs and GaAs are significantly different. The GaAs(100) surface Fermi energy is pinned in the energy gap and is associated with a surface depletion layer, and the surface Fermi energy of InAs(100) is located above the conduction band determining the charge balance between the ionized surface donors and 2DEG. Although S-based passivation schemes have been used to unpin the Fermi energy in GaAs, a modification of the InAs surface Fermi energy is not necessarily desired because surface-based affinity interaction sensors can effectively modulate the conduction electrons. The smaller body of literature on thiol-based S attachment to InAs surfaces has focused on oxide removal and conditions deterring its regrowth.35 As an example, Petrovykh demonstrated that by using basic (NH4OH-based) functionalization solutions the native oxide can be removed during functionalization.25,27,28,30 In this study, we have approached the problem of understanding attachment as the result of a complex, dynamic functionalization process that can be mediated by the presence and composition of the InAs native oxide, which alone might activate

ARTICLE

different surface chemical interactions and therefore bias thiol attachment. Here, we present a systematic study of the attachment chemistry of cysteamine molecules to InAs(100) surfaces with the aim of gaining fundamental insight into the competitive role of sulfur and oxide binding on the resultant molecular overlayer characteristics and its binding. The ultimate objective is to achieve control of the chemical and electronic properties of the InAs surfaces through its interactions with organic molecules. Specifically, this work shows that As2O5 oxide coordination plays a crucial role in establishing (i) how the native oxide and its etching impacts the functionalization chemistry and (ii) how As
S versus In
S bonding results from different chemical interactions and reaction rates during functionalization. Given that the primary motivation for this study is to explore the differential attachment of S to In and As on InAs surfaces, we have chosen to benchmark our work against that of Petrovykh28 in which cysteamine was used as a potential linking molecule for the subsequent attachment of functional molecules to InAs surfaces. The authors point out in this work that the shorter thiol-terminated molecule, cysteamine, was chosen instead of an alkanethiol (explored in prior work) in order to render the XPS analysis highly reliable. In addition, cysteamine avoids oxidizing interactions between the hydroxyl and carboxyl groups in III
V alkanethiols. Furthermore, our choice of cysteamine is also motivated by the desire to provide a basis for investigating InAs interactions with more complex molecules, such as thiolterminated DNA, possessing multiple potential binding groups such as thiols (
SH) and amines (
NH2) because it is known that both sulfur and nitrogen36,37 can bind to III
V groups.

2. EXPERIMENTAL SECTION 2.1. Surface Functionalization. Undoped (WaferTech) InAs(100) wafers (resistivity range = 9.4  10
3
1.2  10
2 Ω cm) were used in the present study. Ethanol solutions of cysteamine (HSC2H4NH2) (hereafter abbreviated as C
SH) with 1, 10, and 100 mM concentration functionalized InAs surfaces were used. The InAs samples were dipped into these solutions for 1 h and rinsed with ethanol. We verified, by performing experiment for longer solution dipping times, that in the case of short-chain thiols such as cysteamine the saturation coverage was achieved within 1 h. Therefore, the data presented are at their steady-state coverage value. The 1 h required to achieve the steady state is consistent with a recent kinetic study on short-chained thiol SAM formation38 and is partially explained by the fact that short-chain thiols do not undergo the lying-down to standing-up conformational change that characterizes long-chain alkanethiol SAM formation.39 Cysteamine, with an isoelectronic point in the range of pKa= 8.6
10.5,40 may interact with the InAs surface via a nondeprotonated thiol. Prior to functionalization, samples were degreased with C2HCl3, acetone, and methanol and, when appropriate, etched for 2 min in 1:20 HF/MeOH acidic solutions to remove the native oxide according to the procedure reported in detail in ref 41. Another set of samples was functionalized without removing the native oxide to compare the resulting surface chemistry and illuminate the role of the InAs oxide in thiol interactions. This protocol differs from and complements the Petrovych approach,25,28,30 which used cysteamine in NH4OH (pKa= 9.25), wherein the alkaline nature of the solution simultaneously removes the native oxide and deprotonates the thiol group during the functionalization process. After functionalization, the samples were rinsed in deionized water and dried in flowing nitrogen. They were then immediately transferred (within 2 to 3 min) to the vacuum chamber for XPS analysis, minimizing air exposure. A corroborating study of the reoxidation of InAs by air 1238

dx.doi.org/10.1021/la203436r |Langmuir 2012, 28, 1235–1245

Langmuir

ARTICLE

Figure 1. O/(In + As) ratio as determined from XPS atomic percentages for InAs with native oxide and for the sample functionalized with 10 mM cysteamine after a few minutes, 1 h, and 1, 3, and 4 days of air exposure. exposure after etching revealed that an ∼3 to 4 Å-thick-oxide is found after 1 day of air exposure, an ∼15 Å oxide is found after a week, and an ∼20 Å oxide layer reforms on the etched InAs after 12 days. A similar reoxidation study has been performed on the functionalized InAs. Because the cysteamine molecular overlayer is not densely packed (see discussion below), InAs reoxidation by air exposure occurs. Using ellipsometry, we determined that ∼3 Å of oxide reforms after 1 day of air exposure. Furthermore, XPS measurements were performed on the 10 mM thiolated samples after 1 h and after 1, 3, 4, and 7 days of air exposure, and the variation of the ratio O/(In + As) is shown in Figure 1, where it can be seen that for thiolated surfaces significant reoxidation takes place after only 1 day of air exposure for our time increments, supporting our interpretation of XPS data taken after functionalization chemistry as insignificantly impacted by reoxidation is carried out. 2.2. X-ray Photoelectron Spectroscopy (XPS). XPS analysis was run using a Kratos instrument with a monochromatized Al Kα source. High-resolution spectra of the As 2p, As 3d, In 3d, S 2p, C 1s, O 1s, and N 1s phototelectron core levels were acquired at takeoff angles of 30 and 90
to profile the depth of the oxide characteristics. The spectra taken at a grazing takeoff angle of 30
characterizes the near surface area resulting from modifications due to the ultrathin (