1 
 
 
1. INTRODUCTION 
1.1 Inspiration for this Project 
Surface modification with self-assembled monolayers. A self-assembled 
monolayer (SAM) constitutes organic molecules that align on the surface of a 
substrate in regular arrays.1-3 These molecules are distinguished by their capability 
of attaching to a solid surface via energetically favorable interactions. They 
adsorb to the surface of interest from a solution or vapor phase to form a densely 
packed molecular barrier on a substrate.3 The interactions between these 
molecules and the surface utilize a variety of intermolecular forces such as 
hydrogen bonding, electrostatic attractions, and covalent bonding. As a result, the 
design and preparation of SAMs are extremely flexible in terms of materials and 
reaction conditions. 
The field of SAMs has experienced enormous development since Zisman 
et al. published a paper on adsorbing an n-alkyl amine-based surfactant to a metal 
surface in 1946,4 initiating the concept of self-assembly. The recent boom in 
nanotechnology has further accelerated the progress of this field. Indeed, SAMs 
are nanostructure in nature.3 They are frequently applied as model substrates in 
research for electrode modification,5 corrosion protection of metals,6 synthesis of 
fullerene and chromophoric films,7, 8 and preparation of biocompatible materials.9 
The great versatility of these surface coatings is primarily attributed to the 
patterned structures of the molecular assemblies, which introduce novel chemical 
2 
 
 
and physical properties to their substrates, activating them for further 
modifications.  
Although self-assembly of molecules has been exploited by a wide range 
of applications, more efforts must be focused on improving the control of the 
growth process to tailor the orientation and dynamics of these molecules for a 
specific purpose. This project was initiated as an attempt to investigate the 
fundamental issues in the preparation of SAMs by reacting aminosilanes with the 
native oxide layer (silica) on smooth silicon wafers. 
Amine-terminated films. Amine-terminated organic molecules are 
frequently used to construct SAMs with their primary amine moieties exposed to 
the exterior. These organic films are anchored to solid surfaces via specific and 
stable interactions without involving their amine moieties. For instance, 3-
aminopropylalkoxysilane attaches to a silica surface via siloxane bonds, Si-O-Si, 
to form an amine-terminated surface that can be used to immobilize enzymes and 
other materials.10, 11 In general, an amine-terminated organic film with a 
horizontal array of primary amine moieties available for further derivatizations is 
ideal for many applications.  
Alkoxy-based silane molecules with the general formula NH2-(CH2)k-
Si(OR)3 are silane coupling agents commonly used to activate silica surfaces 
through a reaction, known as silanization. One of the most typical modification 
processes involves depositing 3-aminopropyltriethoxysilane (APTES) on silicon 
wafers,12-17 which can be then used to graft different organic molecules such as 
3 
 
 
poly(ethylene glycol).16 The multi-functionality of APTES allows it to be attached 
to an inorganic silica surface in a closely packed array via stable covalent siloxane 
bonds through its three alkoxy groups while maintaining the free amine group for 
further modifications.  
In this project, the behaviors of APTES in both solution-phase and vapor-
phase silanization on silicon wafers were analyzed and used as the basis for 
evaluating the performances of four other commercially available silane coupling 
agents.   
Surface morphology and hydrolytic stability of SAMs. The homogeneity in 
surface morphology of SAMs is critical for most research and applications. For 
instance, DNA immobilized on flat APTES-coated mica surface can only be 
imaged properly with atomic force microscopy (AFM) when the amine-
terminated background has a roughness lower than that of the DNA.18 Smooth 
SAMs are also essential for adhesion studies since the homogeneity of functional 
layers affect the specificity and sensitivity of the substrates to their analytes, 
which can be either chemicals or microorganisms.19-21   
The hydrolytic stability of SAMs is another important consideration.  
When the aminosilane-coated substrates are used for grafting hydrophilic 
polymers such as dextrans, they are exposed to an aqueous environment for up to 
96 h.22 A hydrolytically stable SAM is therefore crucial for effective grafting of 
biologically and environmentally relevant molecules onto inorganic substrates. 
4 
 
 
In this project, the surface morphology of aminosilane-derived layers on 
silicon wafers was analyzed with AFM and contact angle goniometry. The 
hydrolytic stability of these organic layers prepared under different conditions 
was studied as a function of their thickness against time.      
1.2 Background Information 
Mechanism of silanization. The covalent attachment of an aminosilane 
molecule to a silica surface proceeds in two steps (Figure 1a). The first step 
begins with the hydrolysis of an aminosilane molecule via an SN2 exchange of an 
alkoxy group for a hydroxyl group. The second step involves the condensation of 
a water molecule, resulting in the formation of a siloxane bond, covalent in nature. 
The amine groups can catalyze the siloxane bond formation both intra-
molecularly and inter-molecularly.23 Intra-molecular catalysis (Figure 1b) is 
initiated by the amine group coordinating to the silicon atom of the same 
molecule via the formation of a penta-coordinated cyclic intermediate, facilitating 
the covalent attachment of an aminosilane molecule to a silica surface. Inter-
molecular catalysis (Figure 1c) involves the amine group of an aminosilane 
molecule coordinating to the silicon atom of another aminosilane molecule in 
close proximity.  
 
 
 
 
 
 
 
 
5 
 
 
(a) 
O O O
SiO2
O
H H
H H H
O
H H
- CH3CH2OH
O O O
SiO2
H H H
H2N
Si
OCH2CH3H3CH2CO
HO
H2N
Si
OCH2CH3H3CH2CO
OCH2CH3 -H2O
O O O
SiO2
H
H2N
Si
OCH2CH3H3CH2CO H
 
(b) 
O O O
SiO2
H H H
Si
H3CH2CO
OCH2CH3
OCH2CH3
NH2
O O O
SiO2
H
H2N
Si
OCH2CH3H3CH2CO H
- CH3CH2OH
 
(c) 
O O O
SiO2
H
H2N
Si
H3CH2CO
OCH2CH3
H3CH2CO
- CH3CH2OH
O O O
SiO2
H
H2N
Si
OCH2CH3H3CH2CO H
NH2
Si
H3CH2CO
OCH2CH3
H3CH2CO
H2N
Si
OCH2CH3
H3CH2CO
H3CH2CO
H H
 
Figure 1. Covalent attachment of 3-aminopropyltriethoxysilane to a silica surface: 
(a) basic mechanism of silanization, (b) intra-molecular catalysis by amine, and 
(c) inter-molecular catalysis by amine.  
 
6 
 
 
Conformations of aminosilane. The conformations of trialkoxy 
aminosilane on a silica surface depend on the density of aminosilane molecules 
and the type of bonding.23 When an aminosilane molecule is covalently attached 
to a silica surface via a siloxane bond, it can have an upright conformation (Figure 
2a) or a tilted one (Figure 2b) due to the attraction between its terminal amine and 
the proton on the silanol oxygen.  
The aminosilane molecules are also capable of cross-linking with one 
another both vertically and horizontally, giving rise to multilayers (Figure 3a, 
vertical polymerization) in addition to dense monolayers (Figure 3b, horizontal 
polymerization).24 The type and extent of cross-linking is particularly sensitive to 
water content in a reaction and thus reduces the reproducibility of this type of 
modification process.  
In addition, the possibility of aminosilane molecules to be physisorbed to a 
silica surface via hydrogen bonding limits the stability of silane derived layers 
since physisorbed molecules are easily removed from the surface by rinsing in 
water (Figure 2c-e).25 In this study, the modified substrates were rinsed rigorously 
in toluene, ethanol, and water to remove loosely attached molecules before 
characterization. The variable surface conformations of aminosilanes make it a 
challenging task to produce high-density and stable SAMs on substrates. 
7 
 
 
Si
H2N
OCH2CH3
O
H3CH2CO
Si
O
H3CH2CO
O
Si
H2N
OCH2CH3
H3CH2CO
H3CH2CO
O
H
Si
H2N
OCH2CH3
HO
H3CH2CO
H2N
Si
H3CH2CO
OCH2CH3
H3CH2CO
O
H
O
HH
H2N
OCH2CH3
 
            (a)                (b)                                           (c)                 (d)                (e) 
 
Figure 2. Possible conformations of 3-aminopropyltriethoxysilane on a silica 
surface. 
 
 
O
Si
H2N
O
Si
OCH2CH3
H2N
O
H3CH2CO
Si
H2N
H3CH2CO
H3CH2CO
O
Si
H2N
O
Si
H2N
O
O
Si
H2N
H3CH2CO OCH2CH3O
 
                                      (a)                                            (b)                                                      
Figure 3. Polymerization of 3-aminopropyltriethoxysilane on a silica surface: (a) 
vertical and (b) horizontal.  
 
Hydrolytic stability of aminosilane-derived layers. Degradations of 
aminosilane-derived layers were observed when modified substrates were 
exposed to an aqueous medium for an extensive period of time.23 The loss of 
thickness is proposed to mainly occur through the hydrolysis of siloxane bonds 
between aminosilane molecules and silica surfaces.23 Shown in Figure 4, water 
attacks the silicon atom via an SN2 reaction and this process can also be catalyzed 
by the amine groups intra- and inter-molecularly similar to the catalysis of a 
8 
 
 
siloxane bond formation. Amine groups can catalyze the formation and the 
degradation of siloxane bonds between aminosilane-derived layers and silicon 
wafer substrates. The main goal of this study is to look for aminosilanes and 
reaction conditions that promote the formation but discourage the degradation of 
siloxane bonds. 
O O O
Si Si
H2N
Si
H2
N
NH2
OH O OH
Si
NH2
O
H
O
H
HH
Si
OH
H2N
Si
OH
H2N
 
 
Figure 4. Intra- and inter-molecular catalysis of the hydrolysis of siloxane bonds 
between 3-aminopropyltriethoxysilane and a silica surface. 
 
Choice of reagents. Tri-alkoxy aminosilanes were chosen over mono- and 
di-alkoxy aminosilanes for their capability of forming multiple siloxane bonds 
with a substrate and neighboring silane molecules as shown in Figure 5, giving 
rise to enhanced hydrolytic stability. 
 A total of five silane coupling agents – APTES, 3-
aminopropyltrimethoxysilane (APTMS), N-(6-
aminohexyl)aminomethyltriethoxysilane (AHAMTES), N-(2-aminoethyl)-3-
aminopropyltriethoxysilane (AEAPTES), and N-(2-aminoethyl)-3-
aminopropyltrimethoxysilane (AEAPTMS) – were analyzed in both solution-
phase and vapor-phase silanization reactions for their hydrolytic stability on 
silicon wafers.  
9 
 
 
Si Si
H2N H2N
O
Si
H2N
O O
O O O
SiO2
O
Si Si
H2N H2N
Si
H2N
O
O O O
SiO2
O
Si Si
H2N H2N
Si
H2N
O O O
SiO2  
                   (a)                                   (b)                                                (c) 
 
Figure 5. SAMs derived with (a) tri-alkoxy aminosilane, (b) di-alkoxy 
aminosilane, and (c) mono-alkoxy aminosilane. 
 
1.3 Foundation Work 
General challenges. The complex binding mechanisms between silane 
coupling agents and silicon wafers have prompted many studies to investigate the 
impact of varying reaction conditions on the covalent attachment of aminosilanes 
and their conformations on substrates. Generally, critical reaction parameters 
include reaction time, temperature, silane concentration, rinsing protocol, drying 
and curing methods. Furthermore, silanization’s acute sensitivity to water 
complicates the experimental design. Although most studies reported in the 
literature focused on analyzing only one aspect of the reaction and their results 
showed poor consistencies with one another, some consensuses have been 
recognized. For instance, submerging silicon wafers in piranha solution (7 parts 
concentrated sulfuric acid and 3 parts 30% hydrogen peroxide) and subsequent 
rinsing in water are known to maximize silanol groups on silicon wafers,26, 27 
which are essential for the formation of siloxane bonds. Reaction parameters that 
are critical to the preparation and stability of SAMs of aminosilanes are discussed 
in the subsequent sections. 
10 
 
 
Solution-phase silanization. A solution-phase silanization reaction is 
usually carried out by submerging silicon wafers in an aminosilane solution for a 
desired amount of time. This method is often used by biologists for its ease of 
preparing relatively thick aminosilane-derived layers, however, homogeneous 
surface morphologies and hydrolytic stability of the silane layers are 
compromised due to the lack of control over the growth process. Specifically, the 
following issues need to be considered.  
First, the nature of solvent affects the solubility of a silane coupling agent 
and the amount of water present in the reaction. Non-polar organic solvents are 
generally suitable for silanization since they readily dissolve aminosilanes and 
have low capacity for dissolving water. In the complete absence of water, alkoxy 
groups cannot be hydrolyzed and silanization solely relies on the amine moieties 
for catalysis, resulting in disordered SAMs with low primary amine contents.26 
Excess water, however, leads to uncontrollable polymerization in both horizontal 
and vertical directions, giving rise to extremely thick and rough multilayers. As a 
result, only a trace amount of water is required for silanization.28, 29 Anhydrous 
toluene that dissolves aminosilanes and contains a very small amount of water 
meets these requirements.  
Secondly, the purity of a silane coupling agent and its concentration in the 
reaction medium affect the surface morphology and the hydrolytic stability of 
aminosilane-derived layers. Impurities in an aminosilane reagent may exist in the 
form of oligomers and polymers of aminosilane molecules, which may be 
11 
 
 
introduced during the manufacturing and storage process. A higher silane 
concentration in the reaction medium increases the chances of aminosilane 
molecules reacting with each other and forming aggregates in solution. The 
aggregates existing as impurities in a silane coupling agent and those formed 
during silanization are variable in size. Their depositions on a substrate disrupt the 
regular arrays of aminosilane molecules, creating structural defects susceptible to 
water penetration. Additionally, these aggregates could be attached to a substrate 
via very few siloxane bonds and hence easily removed by hydrolysis. As a result, 
high silane purity and low solution concentration promote the formation of silane 
layers with uniform morphologies. Based on previous studies,23, 26 0.5 mL and 1.0 
mL of a silane coupling agent in 25 mL of anhydrous toluene were chosen for 
solution-phase silanization. The silane layers prepared at these two concentrations 
presented little differences and therefore the lower concentration was adopted 
throughout this study. 
Thirdly, reaction time and temperature also affect the quality of 
aminosilane-derived layers. Extending reaction time allows the formation of thick 
multilayers at the expense of structural regularity because horizontal and vertical 
polymerizations compete with each other. Raising reaction temperature increases 
the reaction rate and results in greater chances of aggregation as aminosilane 
molecules collide with one another more frequently in the reaction medium. 
These aggregates introduce rough surface features as they are deposited on a 
substrate. In addition, high costs associated with extensive reaction time and high 
12 
 
 
reaction temperatures are not desirable. Thus, moderate reaction time and 
temperature that generate considerably thick and stable aminosilane-derived 
layers are targeted. Previous research has shown that silanization in anhydrous 
toluene at ~70 °C for 24 h produces aminosilane-derived layers with appreciable 
thickness and homogeneous surface morphologies.23, 30  
Furthermore, curing samples at a temperature slightly higher than 100 °C 
after silanization is reported to increase the primary amine contents by promoting 
condensation of water and thermally inducing breakage of hydrogen bonds that 
liberate amine groups from negatively charged silica surfaces.23, 26, 31 However, an 
excessively high curing temperature should be avoided since amine groups may 
be consumed by reacting with atmospheric carbon dioxide to form carbamate.26  
Vapor-phase silanization. Vapor-phase reactions have been developed to 
improve the control of the SAM growth process. Compared to solution-phase 
reactions, vapor-phase reactions greatly simplify the experimental design. First of 
all, they eliminate the impact of silane purity since impurities in the form of 
oligomers and polymers do not vaporize easily and hence do not react with the 
substrate. Secondly, vapor-phase reactions are usually in the absence of a solvent 
and therefore confer a better control of the amount of water present in a reaction, 
minimizing the occurrence of aggregation. Thirdly, they simplify post-silanization 
rinsing of substrates.  
The major drawback of this approach is the unknown nature of bonding 
between the silane and the silica surface. Consequently, evaluating the hydrolytic 
13 
 
 
stability of aminosilane-derived layers is particularly important for vapor-phase 
reactions. 
Multiple studies were reported in the literature to demonstrate the 
feasibility of vapor-phase silanization.33-37 More efforts, however, need to be 
focused on optimizing the reaction conditions, which were rigorously pursued in 
this study. 
Other considerations. Although the capability of an aminosilane molecule 
to covalently attach to a silica surface has been generally accepted, the nature of 
their interactions is not clear. Recent studies attempted to analyze the nature of 
bonding via Fourier transform-infrared spectroscopy (FT-IR).32, 38 It was reported 
that solution-phase reactions at room temperature give rise to siloxane bond 
formation at the interface between a silane coupling agent and a silica surface. 
However, vapor-phase reactions at elevated temperatures do not directly lead to 
siloxane bond formation, which can be induced subsequently by submerging 
modified substrates in water. Our post-silanization treatments such as the rinsing 
and drying of substrates were aimed to promote siloxane bond formation.  
1.4 Objectives 
Hypothesis. We aimed to enhance the hydrolytic stability of aminosilane-
derived layers by optimizing reaction conditions to increase the density of 
aminosilane molecules on silicon wafers. Densely packed aminosilane molecules 
would minimize amine-catalyzed hydrolysis of siloxane bonds. We also evaluated 
the performances of five different silane coupling agents with variable chain 
14 
 
 
lengths in both solution- and vapor-phase silanization. Longer chains would 
increase the steric hindrance to the bending of aminosilane molecules and 
therefore reduce the catalysis of siloxane bond breakage by amine groups. 
Based on their molecular structures, we categorized these five silane 
coupling agents into three groups (Table 1). Group I reagents, including APTES 
and APTMS, contain a primary amine group at propyl or butyl position. They are 
capable of catalyzing the formation and hydrolysis of siloxane bonds both intra- 
and inter-molecularly. Intra-molecular catalysis is achieved via the formation of 
stable five- or six-membered cyclic intermediates. Group II reagents, including 
AEAPTES and AEAPTMS, contain a secondary amine group at propyl or butyl 
position. Similar to Group I reagents, they are capable of intra- and inter-
molecular catalysis, however, due to the steric hindrance imposed by their longer 
chains, amine-catalyzed hydrolysis of siloxane bonds would be reduced. Group III 
reagents, including AHAMTES, do not have any amine group at propyl or butyl 
position. As a result, they are only capable of inter-molecular catalysis.  
Experimental design. Solution-phase silanization was carried out in 
anhydrous toluene at 70 °C for either 24 h or 48 h and vapor-phase silanization 
was carried out mostly at 90 °C for either 24 h or 48 h. Silane coupling agents 
were used as received from Gelest without further purification. All the samples 
were cured at 110 °C post silanization. The hydrolytic stability of aminosilane-
derived layers was studied as a function of time by submerging the samples in 
Milli-Q water at 40 °C for different amounts of time up to 48 h. The temperature, 
15 
 
 
40 °C, was chose to simulate biological media in a slightly accelerated manner. 
Modified substrates were characterized to acquire information about their 
thickness, wettability, surface composition, and surface topography. 
Table 1. Three groups of silane coupling agents. 
Group Reagent(s) 
Si
NH2
RO
OR
OR
 
I. Primary amine can 
coordinate to Si 
Si
OCH2CH3
H3CH2CO OCH2CH3
H2N
         
Si
OCH3
H3CO OCH3
H2N
      
               APTES                     APTMS 
Si
NH
RO
OR
OR
NH2
  
II. Secondary amine can 
coordinate to Si               
Si
OCH3
H3CO OCH3
HN
NH2
 
           AEAPTES                 AEAPTMS 
HN
Si
H2
N
H3CH2CO
OCH2CH3
OCH2CH3
X
 
III. No amine can coordinate 
to Si 
Si
OCH2CH3
H3CH2CO OCH2CH3
HN
NH2
 
 AHAMTES 
 
 
Si
OCH2CH3
H3CH2CO OCH2CH3
HN
NH2
16 
 
 
2. EXPERIMENTAL 
2.1 Materials and Apparati 
General. Silicon wafers (100 orientation, P/B doped, resistivity 1–10 Ω-
cm, thickness 475–575 µm) were purchased from International Wafer Service, Inc. 
(USA). Silane coupling agents, including 3-aminopropyltriethoxysilane (APTES, 
99.7% pure), 3-aminopropyltrimethoxysilane (APTMS, 98.5% pure), N-(6-
aminohexyl)aminomethyltriethoxysilane (AHAMTES, 99.66% pure), N-(2-
aminoethyl)-3-aminopropyltriethoxysilane (AEAPTES, 97.4% pure), and N-(2-
aminoethyl)-3-aminopropyltrimethoxysilane (AEAPTMS, 98.3% pure), were 
purchased from Gelest, Inc. (USA). They were stored under nitrogen and used 
without further distillation. House-purified water (reverse osmosis) was purified 
in a Millipore Milli-Q Biocell System (Millipore Corp., USA) that involved 
reverse osmosis, ion exchange, and filtration (18.2 MΩ-cm). Other reagents were 
used as received from Thermo Fisher Scientific, Inc. (USA). Solution-phase 
silanization was carried out in anhydrous toluene that was dried and deoxygenated 
through a Pure Solv 400-6 solvent purification system (Innovative Technology, 
Inc., USA). All glassware was cleaned in a base bath (potassium hydroxide and 
isopropyl alcohol), rinsed with distilled water (3
 
×), and stored in a clean oven at 
110 °C until use.  
Instrumentation. Thicknesses of aminosilane-derived layers were 
measured with a LSE Stokes Ellipsometer (Gaertner Scientific Corp., USA). 
17 
 
 
Contact angles were measured with a NRL C.A. 100-00 goniometer (ramé-hart 
Instrument Co., USA) with a Gilmont syringe (Gilmont Instrument Co., USA) 
and a 24-gauge flat-tipped needle. Atomic force microscope images were 
obtained with a Veeco Metrology Dimension 3100 Atomic Force Microscope 
(Veeco Instruments, Inc., USA) under tapping mode with a Veeco silicon tip 
(resistivity 1–10 Ω-cm, P doped). Roughness of surface features was determined 
using the Nanoscope software (Veeco Instruments, Inc., USA). Atomic 
compositions (C, O, Si, and N contents) of the surfaces were analyzed by X-ray 
photoelectron spectroscopy using a Quantum 2000 Scanning ESCA Microprobe 
(Physical Electronics, Inc., USA). Thermo NESLAB heated baths (Thermo Fisher 
Scientific, Inc., USA) were used to maintain the temperatures for silanization 
reactions and tests for hydrolytic stability. Precision 51221126 Gravity 
Convection Lab Oven (Thermo Fisher Scientific, Inc., USA) was used for storing 
clean glassware and drying clean and silanized silicon wafers. 
2.2 Methods 
Preparation of silicon wafers. Silicon wafers were cut into rectangular 
pieces (1.2 × 1.5 cm), rinsed with house-purified water, and dried with 
compressed air. They were then placed in a custom-designed glass holder and 
cleaned by submerging in a freshly prepared piranha solution containing 7 parts 
concentrated sulfuric acid and 3 parts 30% hydrogen peroxide for 45 min. After 
18 
 
 
being removed from the solution, wafers were rinsed with copious amounts of 
water and dried in a clean oven at 110 °C for 30 min.  
Solution-phase silanization. Clean wafers were transferred to a clean glass 
holder, placed in a custom-designed Schlenk tube (inner diameter 2.5 cm) fitted 
with an o-ring joint and a Teflon stopcock (ACE Glass, Inc., USA). Using the 
Pure Solv system, the tube was flushed with nitrogen and evacuated (3
 
×) before 
25 mL of anhydrous toluene was dispensed. Under nitrogen, 0.5 mL of an 
aminosilane reagent was cannulated into the reaction tube while the solution was 
being stirred. Reactions were carried out at 70 °C in an oil bath for desired 
amounts of time.  After silanization, the wafers were individually rinsed in 
toluene (2
 
×), ethanol (2
 
×) and water (2
 
×), and dried in an oven at 110 °C for 15 
min.  Thickness and contact angle measurements were carried out immediately 
upon cooling for ~5 min. 
Vapor-phase silanization. Clean wafers were transferred to a clean glass 
holder, placed in a custom-designed Schlenk tube fitted with an o-ring joint and a 
Teflon stopcock. After purging the system with nitrogen for ~30 min, 0.5 mL of 
an aminosilane reagent was cannulated into the reaction tube under nitrogen. 
Reactions were carried out at 70 °C, 90 °C, 100 °C, or 105 °C in an oil bath for 
desired amounts of time. After silanization, the wafers were rinsed individually 
with toluene (2
 
×), ethanol (2
 
×) and water (2
 
×), and dried in an oven at 110 °C 
19 
 
 
for 15 min. Thickness and contact angle measurements were carried out 
immediately upon cooling for ~5 min. 
Test method for hydrolytic stability. Freshly silanized samples were placed 
in a clean glass holder and submerged in Milli-Q water in a tightly closed Schlenk 
tube at 40 °C for 1 h, 3 h, 12 h, 24 h, or 48 h. Samples were then rinsed in 
distilled water and dried in an oven at 110 °C for 15 min before characterization.  
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
20 
 
 
3. SURFACE CHARACTERIZATION TECHNIQUES 
3.1 Ellipsometry 
General. Ellipsometry measures the changes in polarization of elliptically 
polarized light upon its reflection from a surface. It yields information about the 
thickness and optical properties of the characterized surface. In this study, 
ellipsometry was adopted to measure the thickness of aminosilane-derived layers, 
ranging from ~2 Å to ~500 Å. The thickness of a clean silicon wafer substrate and 
that of the modified substrate were respectively measured before and after 
silanization. Hence, the thickness of the aminosilane-derived layer was estimated 
to be the difference between these two measurements.     
Theory. Light is an electromagnetic transverse wave composing of an 
electric field and a magnetic field that are perpendicular to each other and to the 
propagation of the wave. Usually, light is unpolarized with electric fields of 
random orientations and phases. Polarization of light restricts its electric field to a 
well-defined shape and path. Linearly polarized light consists of two orthogonal 
waves that are in phase, tracing out a straight line by its electric field vector on a 
fixed plane as it propagates (Figure 6a). When these two waves are equal in 
amplitude but are out of phase by 90°, they constitute a circularly polarized light 
(Figure 6b). Ellipsometry uses elliptically polarized light by combining two 
orthogonal waves of random amplitude and phase difference (≠ 0° or 90°, Figure 
6c).39  
21 
 
 
(a)  
(b)  
(c)  
 
Figure 6. Examples of light polarization: (a) linearly, (b) circularly, and (c) 
elliptically polarized light.40 
 
When a light beam reaches an interface between two materials of different 
optical densities, reflection and/or refraction occur. The law of reflection (Figure 
7a) states that the angle of incidence is equal to the angle of reflection, i.e., 
𝜃𝑖 = 𝜃𝑟. (1) 
Refraction is governed by the Snell’s law (Figure 7b), i.e., 
𝑛1 sin𝜃1 = 𝑛2 sin𝜃2, (2) 
where 𝑛1 and 𝑛2 denote the refractive indices of the materials, which increase as 
the optical densities increase. 𝜃1 and 𝜃2 denote the incident and refracted angles 
22 
 
 
respectively. Total internal reflection occurs when light travels from a material 
into an optically less dense material at an incident angle greater than the critical 
angle. In this condition, all the light is reflected and refraction does not occur.  
 
 
Figure 7(a). Law of reflection.                                 Figure 7(b). Snell’s law. 
 
When an elliptically polarized light beam reaches multiple interfaces 
(Figure 8), both reflection and refraction occur. The wave that is parallel to the 
plane of incidence is referred as the p wave and the one that is perpendicular to 
the plane of incidence is referred as the s wave. There are four physical 
parameters crucial to the operation of an ellipsometer. First, the Fresnel reflection 
coefficient, r, describes the ratio of the amplitude of the reflected wave to that of 
the incident wave in the p and s directions,41 i.e., 
𝑟12
𝑝 =
𝑁2cos𝜃1−𝑁1cos𝜃2
𝑁2cos𝜃1+𝑁1cos𝜃2
 and 𝑟12
𝑠 =
𝑁1cos𝜃1−𝑁2cos𝜃2
𝑁1cos𝜃1+𝑁2cos𝜃2
. (3) 
𝑁1 and 𝑁2 denote the complex refractive indices of the materials, taking into 
consideration the extinction coefficients.  𝑟23
𝑝  and 𝑟23
𝑠  are defined in the same 
manner. 41 
23 
 
 
 
 
Figure 8. Reflections and refractions at multiple interfaces. 
 
The reflection coefficient r is related to the total reflection coefficient, R, 
by 
𝑅𝑝 =
𝑟12
𝑝 +𝑟23
𝑝 𝑒−𝑖2𝛽
1+𝑟12
𝑝 𝑟23
𝑝 𝑒−𝑖2𝛽
  and 𝑅𝑠 =
𝑟12
𝑠 +𝑟23
𝑠 𝑒−𝑖2𝛽
1+𝑟12
𝑠 𝑟23
𝑠 𝑒−𝑖2𝛽
. (4) 
where β, the film phase thickness, is given by  
𝛽 = 2𝜋 �
𝑑
𝜆
�𝑁2 cos 𝜃2, (5) 
 and d is the thickness of the interested surface layer and λ is the wavelength. 41 
Thirdly, reflectance 𝑹� is defined as the ratio of the reflected intensity to 
the incident intensity,41 i.e.,  
𝑹�𝒑 = |𝑅𝑝|2 and 𝑹�𝒔 = |𝑅𝑠|2. (6) 
An ellipsometer directly measures two physical quantities, ∆ and Ψ.41 The 
parameter ∆ is the change in phase difference between p- and s-polarized light 
upon reflection and tan Ψ is the corresponding amplitude ratio upon reflection. ∆ 
and tan Ψ are related to the previous quantities (𝑟,𝑅,𝛽,𝑎𝑛𝑑 𝑹�) through the 
fundamental equation, 
24 
 
 
tanΨ 𝑒𝑖Δ =
𝑹�𝒑
𝑹�𝒔
. (7) 
When the refractive index of the surface layer, N2, is known, the thickness of the 
surface layer, d, can be calculated by a software program. 
Instrumentation & data analysis. A basic ellipsometer consists of the 
following components: (1) a monochromatic light source, (2) polarizers to 
elliptically polarize the light, (3) a sample stage, (4) an analyzer to determine the 
state of polarization of the reflected light, and (5) a detector to convert light to 
electronic signals (Figure 9). The output state of polarization is compared to the 
input state to yield the thickness or refractive index of the surface layer.  
 
 
Figure 9. Ellipsometer components and operation schematic. 42 
 
The ellipsometer used in this study has a He-Ne laser light source with a 
wavelength of 632.8 nm and a 70° angle of incidence (from the surface normal). 
Thickness calculations were based on the following parameters: air, N1 = 1; 
silicon oxide and silane-derived layers, N2 = 1.46; silicon substrate, N3 = 3.85 and 
k3 = -0.02. The extinction coefficient, k3, indicates the amount of absorption loss 
when light propagates through the silicon substrate. Measurement error is within 1 
25 
 
 
Å as specified by the manufacturer. The reported thicknesses were averages of 
four measurements made on different areas of each sample surface.    
3.2 Contact Angle Goniometry 
General. Contact angle goniometry measures the wettability of a surface, 
which is its ability to maintain contact with a liquid. It is largely affected by the 
outermost few Å of the material. This technique measures energetics, physical 
roughness, and chemical heterogeneities of a surface. Contact angle theories have 
been introduced in physical chemistry over the last couple of centuries and their 
applications in surface characterization are usually for qualitative purposes. 
Theory. A free liquid drop takes the shape that minimizes the free energy 
of the system when it rests on a solid surface (Figure 10). The surface free energy 
is the amount of work required to increase the surface area of a substance by a 
unit area.  A change in the shape of the liquid drop is governed by a change in the 
free energy of the system, i.e. the Gibbs free energy, and it is demonstrated by the 
following equation: 
𝑑𝐺 = 𝛾𝑆𝐿𝑑𝐴𝑆𝐿 + 𝛾
𝑆𝑉𝑑𝐴𝑆𝑉 + 𝛾
𝐿𝑉𝑑𝐴𝐿𝑉, (8) 
where γ denotes the surface free energy of the specific interface and it is also 
referred as the surface tension. dA denotes the change in the surface area of the 
corresponding interface and by geometry, it fulfills that 𝑑𝐴𝑆𝑉 = −𝑑𝐴𝑆𝐿 and 
𝑑𝐴𝐿𝑉 = cos 𝜃𝐶 𝑑𝐴𝑆𝐿. Hence, Equation 8 becomes 
𝑑𝐺 = (𝛾𝑆𝐿 − 𝛾𝑆𝑉 + 𝛾𝐿𝑉 cos𝜃𝐶)𝑑𝐴𝑆𝐿. (9) 
26 
 
 
At equilibrium where dG = 0, 𝛾𝑆𝐿,𝛾𝑆𝑉 and 𝛾𝐿𝑉 are related by the Young’s 
equation,43  
𝛾𝐿𝑉 cos 𝜃𝐶 = 𝛾
𝑆𝑉 − 𝛾𝑆𝐿. (10) 
 
 
Figure 10. Static contact angle. 44 
 
The Young’s equation suggests that the angle θC is the outcome of the 
interactions between the three phases. It is termed static, equilibrium, or intrinsic 
contact angle and its utilization is based on a few important assumptions. First, 
the solid surface is smooth and chemically homogenous. Secondly, the solid 
surface maintains its shape and is not penetrated by the liquid. Thirdly, the 
functional groups on the solid surface do not react with the liquid. These 
conditions are rarely met in practical situations. As a result, dynamic contact 
angles are measured to overcome these limitations. The advancing angle, θA, is 
measured when a liquid drop is added to a solid surface (Figure 11a) and the 
receding angle, θR, is measured when the drop is withdrawn from the surface 
(Figure 11b). Generally, θA > θC > θR but there are exceptions. 
 
27 
 
 
                            
                                (a)                                                          (b) 
 
Figure 11. Measurement of dynamic contact angles: (a) an advancing angle and 
(b) a receding angle.45 
 
The difference between θA  and θR is known as the contact angle hysteresis. 
There are two major categories of hysteresis: thermodynamic hysteresis and 
kinetic hysteresis. Thermodynamic hysteresis is reproducible over time when the 
same liquid is applied to the same surface.  It is contributed by surface roughness 
and chemical heterogeneity.46 Kinetic hysteresis is time dependent. It is critical 
when the liquid is able to penetrate the surface with increased contact time and/or 
when the functional groups on the solid surface are able to reorient after 
contacting the liquid.46 Since silica surfaces used in this study were very stable 
and did not vary with time, kinetic hysteresis was not significant.  
For thermodynamic hysteresis, the Wenzel’s equation,47  
cos 𝜃𝑚 = 𝑟 cos 𝜃, (11) 
describes the effect of surface roughness. 𝜃𝑚 is the measured contact angle, θ is 
the intrinsic contact angle and r, the Wenzel’s roughness ratio, is equal to the ratio 
of the true surface area taking into account of surface roughness to the projected 
surface area. This relationship suggests that roughening a hydrophilic surface with 
an intrinsic angle of less than 90° (θ < 90°) decreases the observed angle 𝜃𝑚. 
28 
 
 
However, when a hydrophobic surface with an intrinsic angle of greater than 90° 
is examined, the opposite is observed.  
The effect of chemical heterogeneity is described by the Cassie’s 
equation,48 
cos 𝜃𝑚 = 𝑄1 cos 𝜃1 + 𝑄2 cos 𝜃2. (12) 
𝜃𝑚 is the measured contact angle. 𝑄1 and 𝑄2 are the fractions of the surface 
covered by respective regions with contact angles of 𝜃1 and 𝜃2 in each phase. 
This relationship is established upon the assumption that the surface consists of 
distinctively defined regions that are large relative to molecular dimensions and 
therefore the observed angle is a weighted average. In general, this model predicts 
that the advancing angle is associated with the low-energy regions (hydrophobic) 
of a heterogeneous surface where the advancing edge of the liquid drop is 
hindered and the receding angle is associated with the high-energy regions 
(hydrophilic) where the receding liquid edge is pinned down.   
Contact angle provides rich information about both the physical and 
chemical characteristics of a surface. The complex interactions between the 
surface and the liquid introduce a variety of important parameters for its 
measurement. In this study, contact angle measurements were used to assess the 
extent of hydrolytic degradation of aminosilane-derived layers. 
Instrumentation & data analysis. Dynamic contact angles were measured 
on silicon wafer substrates with Milli-Q water as the probe fluid. Advancing and 
receding angles were recorded respectively while the probe fluid was released to 
29 
 
 
and withdrawn from the surface. The reported values were averages of three 
measurements made on different areas of each sample surface.  Within the same 
batch of samples, the standard deviation of reported values was less than or equal 
to 2°. 
3.3 Atomic Force Microscopy 
General. Atomic force microscopy (AFM) investigates and images surface 
features at nanoscopic resolution. It consists of a sharp tip microfabricated to a 
soft cantilever. When the tip is in close proximity of a surface, they interact with 
each other and deflect the cantilever. The deflections of the cantilever correlate 
with the surface features and the signals are processed to reflect the topography of 
the surface. 
Theory. AFM images are acquired by measuring the deflections of the 
cantilever due to the interactions between the tip and the surface. The tip-surface 
interaction encompasses a variety of forces that can be van der Waal’s, 
electrostatic, magnetic and so on. As a result, this technique can be applied to 
both conducting and insulating materials.  
The deflection of the cantilever is governed by Hooke’s law, 
𝐹 = −𝑘𝑥, (13) 
where F is the restoring force exerted by the cantilever that corresponds to the tip-
surface force,  k is the spring constant that depends on the elasticity of the 
cantilever material, and x is the displacement of the cantilever from its position of 
equilibrium. In theory, there are two major modes of operation: constant-height 
30 
 
 
mode and constant-force mode. In the constant-height mode, the tip is maintained 
at a fixed height and the tip-surface forces are measured continuously as the tip 
scans across the surface. The output signals are converted to height information of 
the surface and plotted against their positions on the sample. This operation mode 
may cause undesired damages to the sample when the tip touches the surface. In 
the constant-force mode, the height of the tip is varied to maintain a constant 
force between the tip and the surface. Again, the output signals are converted to 
height information of the surface and plotted against their positions. This mode 
helps to preserve the surface features of a sample.  
In practice, the constant-height mode is exemplified by the contact mode 
AFM. The tip maintains contact with the surface usually through an adsorbed 
fluid layer on the sample surface. The tip-surface interaction is usually dominated 
by short-range interatomic forces, which are measured by measuring static 
cantilever deflections.49 
The constant-force mode is represented by the tapping mode AFM, which 
was adopted in this study. The tip-surface interaction is dominated by long-range 
forces, such as van der Waal’s attraction. The tip is attached to a cantilever that 
oscillates at or near its resonance frequency driven by a piezoelectric element. A 
change in the tip-surface interaction is detected by changes in the frequency and 
amplitude of the cantilever. For instance, if the tip is attracted to the surface, the 
cantilever will oscillate at a lowered frequency and amplitude. In principle, this 
operation mode is sensitive to single electrons.50  
31 
 
 
 
 
Figure 12.  Atomic force microscope components and operation schematic. 51 
 
Instrumentation & data analysis. Tapping mode AFM operates by 
scanning a tip across a sample surface (Figure 12).  The tip with ~10 nm of 
curvature is microfabricated to a cantilever. It gently taps the sample, touching the 
surface at the bottom of its swing. 51 A feedback loop with a piezoelectric element, 
which converts electrical energy to mechanical motions, drives the cantilever to 
oscillate at a constant resonance frequency with an amplitude between 20 nm and 
100 nm.51 Changes in the oscillation of the cantilever are recorded by a 
photodiode that detects the deflections of a laser beam, focused on the tip at its 
static position. Larger deflections of the laser beam correspond to higher and 
lower surface features. 
32 
 
 
AFM images were acquired and analyzed with the Nanoscope software 
program. Three-dimensional topography, two-dimensional height profile, and 
root-mean-square roughness of an image can be determined.  
3.4 X-ray Photoelectron Spectroscopy 
General. X-ray photoelectron spectroscopy (XPS) detects the relative 
amounts of elements on a sample surface. It operates based on the photoelectric 
effect by exciting core electrons on a sample surface and recording both the 
number and energy of the ejected electrons. Since the energy of the ejected 
electrons is specific to an element, the identity of the element can therefore be 
determined. This technique detects all elements except hydrogen and helium. 
Theory. When a sample is irradiated with an energetic electromagnetic 
wave with certain frequency, ν, some core electrons capture the energy and escape 
from the material. The kinetic energy of the electrons, K.E., is given by, 
𝐾.𝐸. = ℎ𝜈 − 𝐵.𝐸.− 𝜙, (14) 
where h is the Planck’s constant, B.E. denotes the binding energy of the electrons, 
and 𝜙 denotes the work function of the spectrometer, which is the minimum 
energy required to remove an electron from a solid to a point immediately outside 
the solid surface. The binding energy of core electrons in a specific element is 
known and it is usually in the range of 0 to 1100 eV depending on the electronic 
structure of the atom. Each element has a unique set of core electrons and their 
energy states are characteristic of that element.52 As a result, the number of 
ejected electrons is recorded as a function of their binding energy in order to 
33 
 
 
identify the element and its abundance. This technique is sensitive to the 
outermost few nanometers of a sample since only the electrons near the surface 
can overcome inelastic collisions with neighboring atoms to reach the detector. 
XPS can also be used for detecting elements at different depths of a 
sample (Figure 13). This is particularly relevant to this study since the samples 
were surface-modified and did not have uniform chemical compositions at all 
depths. This function is achieved by varying the take-off angle between the 
sample surface and the detector. When the take-off angle is increased, a larger 
number of electrons are able to escape from the bulk by traveling a shortened 
distance within the sample, and thus the sample is examined at a greater depth.53 
The number of detected electrons, N, is related to the depth via 
𝑁
𝑁𝑜
= 𝑒
−𝑡
𝜆sin𝜃  (15)  
where No is the number of electrons produced at depth, t, λ is the mean free path 
of the electron, and θ is the take-off angle.  
 
 
Figure 13. X-ray photoelectron spectroscopy at two take-off angles: 15° and 45°. 
 
34 
 
 
Instrumentation & data analysis. A basic X-ray photoelectron 
spectrometer consists of the following components: (1) an X-ray tube with a 
magnesium or aluminum anode to produce a monochromatic laser beam, (2) a 
vacuum chamber to enclose the X-ray and the sample, (3) a neutralizer such as an 
electron gun to prevent the build-up of surface charges, (4) an electron multiplier 
to detect the ejected electrons and amplify the signals, (5) an analyzer to record 
and process the signals (Figure 14). 
In this study, an X-ray photoelectron spectrometer equipped with an 
aluminum anode-laser source was used for detecting C, O, Si, and N contents at 
two take-off angles, 15° and 45°.  The number of scans for each element was 
adjusted to optimize the signal-to-noise ratio. 
 
 
Figure 14. X-ray photoelectron spectrometer components and operation 
schematic.54 
 
 
 
35 
 
 
4. RESULTS AND DISCUSSION 
4.1 Solution-phase Silanization 
Overview. Solution-phase silanization was carried out using silane 
coupling agents from all three groups in anhydrous toluene at 70 °C. The pre-
determined conditions were observed to be the most ideal for preparing 
reproducible and relatively thick aminosilane-derived layers.23 The hydrolytic 
stability of aminosilane-derived layers was analyzed as a function of time by 
submerging freshly silanized samples in Milli-Q water at 40 °C for different 
amounts of time up to 48 h. Thickness information, contact angle data, and AFM 
images were acquired before and after hydrolysis and analyzed to assess the 
hydrolytic stability of respective samples.  
Based on their molecular structures, Group I reagents (APTES and 
APTMS) have a chain length of ~5 Å, Group II reagents (AEAPTES and 
AEAPTMS) have a chain length of ~8 Å, and Group III reagents (AHAMTES) 
have a chain length of ~10 Å. As a result, the thicknesses of the corresponding 
monolayers consisting of closely packed aminosilane molecules should increase 
in the order of Group I < Group II < Group III. Through experiments, five major 
factors were identified to have critical impact on the morphology and hydrolytic 
stability of aminosilane-derived layers: purity of silane coupling agent, ambient 
humidity, reaction time, nature of silane coupling agent (Group I, II or III), and 
type of alkoxy group (methoxy group, –OCH3, or ethoxy group, –OCH2CH3). The 
influence of each factor is elaborated in the following sections. 
36 
 
 
Silane purity. When 24-h silanization was performed with the Group I 
reagents, initial thicknesses of ~25 Å were obtained with APTES and APTMS as 
shown in Table 2 and Figure 15. The dynamic water advancing and receding 
contact angles are much higher than those of a clean silicon wafer, which are 
~0°/~0° (θA/θR), as the wafer is extremely hydrophilic with abundant silanol 
groups on the surface. The increase in dynamic contact angles indicates an 
increase in hydrophobicity of the surfaces and confirms the fact that aminosilane-
derived layers have been attached to the substrates. The thicknesses of the silane 
layers also suggest that multilayers instead of monolayers have been formed.  
After 24-h exposure to water, the multilayers degraded to thicknesses 
close to those of monolayers. The dynamic contact angles also decreased 
accordingly.  In addition, the reductions in contact angle hysteresis (θA – θR) 
suggest that the physical smoothness and chemical homogeneity of the silane-
modified surfaces improved. When the silanized substrates were soaked in water, 
loosely attached aminosilane molecules were removed from the surfaces, leaving 
the densely packed monolayers relatively intact, hence improving the physical 
smoothness of the modified surfaces. The remaining aminosilane molecules on 
the substrates might also reorient to assume vertical conformations, exposing their 
primary amine groups to the exterior (Figure 2a). As a result, the hydrolyzed 
surfaces might have become more hydrophilic and chemically homogeneous, 
causing dynamic contact angles and contact angle hysteresis to decrease.   
37 
 
 
Similar trends were observed with silane layers prepared with the Group II 
reagent AEAPTMS and the Group III reagent AHAMTES.  The initial thickness 
of the AEAPTMS-derived layer was close to those of Group I reagents but its 
remaining thickness after 24-h hydrolysis in water was higher due to its longer 
chain. The AHAMTES-derived layer had the most reproducible initial and 
remaining thicknesses. The decrease in dynamic contact angles was also minimal, 
indicating a stable silane layer structure.  
The striking differences observed between AEAPTES and AEAPTMS 
were most likely due to their different purities. AEAPTES has the lowest purity, 
97.4%, among all the five reagents, giving rise to extremely thick and 
irreproducible multilayers that degraded extensively during hydrolysis. Although 
silanization with AEAPTES was carried out under exactly the same conditions, an 
initial thickness of ~280 Å was obtained and its standard deviation (268 Å) 
among samples prepared in the same batch and different batches was extremely 
large. After hydrolysis in water, the thickness was reduced drastically to a value 
that is lower than that of a monolayer, suggesting the exposure of bare silicon 
wafer substrates.  
These observations could arise from the impurities in the silane coupling 
agent, which may exist in the form of oligomers and polymers of aminosilane 
molecules introduced during the manufacturing process. These aggregates could 
be anchored to the substrate surface by very few siloxane bonds and therefore 
easily removed during hydrolysis. Additionally, the irregular conformations of the 
38 
 
 
attached aggregates could disrupt the regular arrays of aminosilane molecules, 
creating structural defects susceptible to water attack. Consequently, these 
multilayers are not stable in water.  
The differences observed between the Group I reagents, APTES and 
APTMS, could also partly caused by their different purities. APTMS, with a 
lower purity (98.5%) than that of APTES (99.7%), produced layers with more 
variable thicknesses, evident by a larger standard deviation.      
Table 2. Thickness and water contact angle (θA/θR) data of some aminosilane-
derived layers before and after 24-h exposure to water. 
 
Group 
Purity 
(%) 
Initial After 24-h hydrolysis 
Thickness 
(Å) 
Contact angle 
(deg) 
Thickness 
(Å) 
Contact angle 
(deg) 
I 
APTES 
APTMS 
99.7 
98.5 
23 ± 6 
26 ± 15 
(32 ± 2)/(16 ± 2) 
(44 ± 2)/(18 ± 2) 
9 ± 4 
11 ± 4 
(24 ± 2)/(14 ± 2) 
(27 ± 2)/(15 ± 2) 
II 
AEAPTES 
AEAPTMS 
97.4 
98.3 
282 ± 268 
24 ± 7 
(36 ± 2)/(13 ± 2) 
(53 ± 2)/(18 ± 2) 
4 ± 2 
16 ± 7 
(31 ± 2)/(14 ± 2) 
(42 ± 2)/(14 ± 2) 
III AHAMTES 99.66 21 ± 2 (48 ± 2)/(18 ± 2) 14 ± 4 (49 ± 2)/(18 ± 2) 
 
39 
 
 
 
Figure 15. Thickness of aminosilane-derived layers before and after 24-h 
exposure to water. Solution-phase silanization was carried out in anhydrous 
toluene at 70 °C for 24 h. 
 
To further confirm the impact of silane purity on the reproducibility of 
aminosilane-derived layers, silanization reactions were performed in anhydrous 
toluene at 70 °C for 1 h, 2 h, and 24 h using AEAPTES (97.4% pure), AEAPTMS 
(98.3% pure), and AHAMTES (99.66% pure).  Hydrolysis tests were continued 
for 48 h to monitor the behaviors of aminosilane molecules in water for an 
extended period of time. Table 3 illustrates that as the silane purity increased, the 
standard deviation in the initial thickness of the derived silane layer decreased, 
indicating an improvement in the reproducibility of the data.  
 
 
0
100
200
300
400
500
600
Thickness (Å)
Initial thickness Thickness after 24-h hydrolysis
APTES APTMS AEAPTES AEAPTMS AHAMTES
40 
 
 
Table 3. Thickness and water contact angle (θA/θR) data of some aminosilane-
derived layers before and after 48-h exposure to water. 
 
 
Reaction 
time (h) 
Initial After 48-h hydrolysis 
Thickness 
(Å) 
Contact angle 
(deg) 
Thickness 
(Å) 
Contact angle 
(deg) 
AEAPTES 
(97.4%) 
1 
2 
24 
94 ± 35 
81 ± 54 
282 ± 268 
(46 ± 2)/(15 ± 2) 
(42 ± 2)/(16 ± 2) 
(36 ± 2)/(13 ± 2) 
14 ± 1 
4 ± 4 
5 ± 2 
(34 ± 2)/(14 ± 2) 
(33 ± 2)/(17 ± 2) 
(33 ± 2)/(15 ± 2) 
AEAPTMS 
(98.3%) 
1 
2 
24 
14 ± 4 
14 ± 1 
24 ± 7 
(52 ± 2)/(18 ± 2) 
(53 ± 2)/(18 ± 2) 
(53 ± 2)/(18 ± 2) 
7 ± 2 
6 ± 2 
17 ± 6 
(41 ± 2)/(15 ± 2) 
(38 ± 2)/(13 ± 2) 
(42 ± 2)/(14 ± 2) 
AHAMTES 
(99.66%) 
1 
2 
24 
7 ± 3 
13 ± 1 
21 ± 2 
(47 ± 2)/(22 ± 2) 
(45 ± 2)/(17 ± 2) 
(48 ± 2)/(18 ± 2) 
5 ± 1 
10 ± 2 
12 ± 1 
(44 ± 2)/(22 ± 2) 
(43 ± 2)/(20 ± 2) 
(46 ± 2)/(20 ± 2) 
 
Moreover, an interesting trend was observed with AEAPTES. When the 
reaction time was increased, the initial thickness of the silane layer increased but 
its remaining thickness after 48-h hydrolysis decreased. This implies that the 
thicker layers of AEAPTES probably contained larger amounts of aggregates that 
reduced the regularity and hence the hydrolytic stability of the derived silane 
layers.  AFM images in Figure 16 confirm this rationale by showing that samples 
prepared at longer reaction times have more and bigger bright features, which 
correspond to a larger number of bigger silane aggregates. After hydrolysis, some 
of the bright features remained on the 1-h silanized sample but disappeared 
completely on the 24-h silanized sample.  
 
 
41 
 
 
 
 1-h silanization 2-h silanization 24-h silanization 
Initial 
 
 
 
  
After 48-h 
hydrolysis 
 
 
 
 
 
 
 
Figure 16. AFM images (1 µm×1 µm, height scale at 20 nm) of AEAPTES-
derived layers.  
 
In summary, the purity of a silane coupling agent has a significant impact 
on the thickness, reproducibility, hydrolytic stability, and morphology of the 
aminosilane-derived layers. When the silane coupling agents are used as received 
from their manufacturers without purification, they contain variable amounts of 
impurities that may be in the form of oligomers and polymers of the aminosilane 
molecules. They may also form during storage while water diffuses into the 
reagents, inducing polymerization/aggregation of monomeric aminosilane 
molecules. These aggregates can be attached to a substrate surface by very few 
siloxane bonds, giving rise to irregular silane structures. Water therefore can 
easily penetrate irregularly structured multilayers, cleaves the few siloxane bonds, 
and removes aggregates from the surface. In order to enhance the hydrolytic 
stability and reproducibility of aminosilane-derived layers, silane coupling agents 
42 
 
 
therefore must be purified prior to solution-phase silanization, for instance, by 
distillation.  In the subsequent discussion of other factors, AEAPTES is not 
included due to the overpowering effect of its low purity. 
Ambient humidity. In addition to silane purity, ambient humidity also 
affects the thickness of aminosilane-derived layers. Illustrated in Figure 17, 
thicker silane layers were obtained at higher ambient humidity while maintaining 
the same reaction conditions.  More water molecules are present in the reaction 
system under higher humidity, promoting siloxane bond formation. However, 
excess water results in the aggregation of aminosilane molecules and hence 
irregularities in silane layer structures. This drastically decreases the 
reproducibility of the experiments. As a result, in order to generate reproducible 
silane layers, only a trace amount of water is required in the reaction medium and 
the reaction system should be kept as dry as possible.    
 
 
Figure 17. Thickness of APTMS- and AEAPTES-derived layers. Solution-phase 
silanization was carried out in anhydrous toluene at 70 °C for 24 h under 
different ambient humidity. 
0
100
200
300
400
500
Thickness (Å)
Lower humidity Higher humidity
(Humidity: 19% and 37%) (Humidity: 39% and 59%)
APTMS
AEAPTES
43 
 
 
Reaction time. It was reported that the thickness of APTES-derived layer 
increased as the reaction time was increased.28 The same trend was observed with 
the Group II reagent AEAPTMS and the Group III reagent AHAMTES when the 
reaction time was increased from 1 h to 24 h as shown in Figure 18. 
 
(a)  
 
(b)  
 
Figure 18. Thickness of aminosilane-derived layers against reaction time: (a) 
AEAPTMS and (b) AHAMTES. Solution-phase silanization was carried out in 
anhydrous toluene at 70 °C for 1 h, 2 h, and 24 h, respectively.  
0
5
10
15
20
25
30
35
Thickness (Å)
Initial thickness Thickness after 48-h hydrolysis
1-h silanization 2-h silanization 24-h silanization
0
5
10
15
20
25
30
35
Thickness (Å)
Initial thickness Thickness after 48-h hydrolysis
1-h silanization 2-h silanization 24-h silanization
44 
 
 
As the reaction progresses, more aminosilane molecules are allowed to 
deposit on the substrate surface. Shown in Figure 18, the remaining thicknesses 
after soaking in water scale with the initial thicknesses of the silane layers, 
suggesting that hydrolytic stability could be enhanced by increasing reaction time. 
However, it is speculated that the multilayers would eventually degrade to 
monolayers given sufficient hydrolysis time.  
Nevertheless, extending reaction time does not necessarily improve the 
surface morphology of the silane layers. Surface features observed on AFM 
images of AHAMTES-derived layers (Figure 19) were similar in size and shape 
for samples prepared at different reaction times. The mean roughness of the 
surface determined by the Nanoscope software program was also independent of 
reaction time. 
In general, the reaction time for solution-phase silanization can be 
adjusted accordingly for acquiring the desired thickness of silane layers because 
their remaining thickness after exposure to water scales with their initial thickness. 
Silane layers also tend to have comparable surface morphologies that are 
independent of their thicknesses within 24-h reaction time.  
 
 
 
 
 
45 
 
 
 1-h silanization 2-h silanization 24-h silanization 
Initial 
 
Roughness: 7.00 Å 
 
Roughness: 5.30 Å 
 
Roughness: 5.30 Å 
After 24-h 
hydrolysis 
 
Roughness: 6.24 Å 
 
Roughness: 4.64 Å 
 
Roughness: 5.71 Å 
 
Figure 19. AFM images (1 µm×1 µm, height scale at 20 nm) of AHAMTES-
derived layers. 
 
Nature of silane coupling agent. The performance of a silane coupling 
agent was evaluated based on the thickness, reproducibility, hydrolytic stability, 
and surface morphology of its silane layers. APTES, the most commonly used 
silane coupling agent for functionalizing silicon wafers, was used as the standard 
for comparison. Again, AEAPTES is excluded due to its low purity. 
The effect of silane structure on silanization. As shown in Figure 20, the 
initial thicknesses of aminosilane-derived layers obtained by 24-h silanization 
were all ~25 Å. The thickness of silane layers thus appears to be not affected by 
the structure of the silane coupling agent. All the aminosilanes examined are 
capable of inter-molecular catalysis in solution, which seems sufficient to produce 
silane layers of comparable thicknesses. 
However, the standard deviation in the thickness of the silane layers 
increased steadily from the Group III, the Group II, to the Group I reagents. 
46 
 
 
Variability in the thickness of silane layers could be attributed to the impact of 
silane purity and the intra-molecularly catalyzed silanization in the Group I and II 
reagents, which leads to the formation of oligomers of aminosilane molecules in 
solution and their deposition on substrates. Particularly, the primary amine groups 
in the Group I reagents have higher tendency to form cyclic intermediates that 
facilitate the occurrence of intra-molecular catalysis. In addition, tri-methoxy 
aminosilanes are more sensitive to the ambient humidity than tri-ethoxy 
aminosilanes and this factor is elaborated further in the subsequent sections. 
Overall, the reactions with AHAMTES generated the most reproducible silane 
layers.  
The effect of silane structure on hydrolytic stability. Although the initial 
thicknesses of silane layers were comparable across all three groups of 
aminosilanes, the remaining thicknesses after 24-h hydrolysis were apparently 
larger for the Group II and III reagents probably due to their structural features 
and enhanced hydrolytic stability. Both groups of aminosilanes have longer 
chains, which hinder the formation of cyclic intermediates, reducing intra-
molecularly catalyzed hydrolysis of siloxane bonds.  
Kinetic studies for hydrolysis in water were performed with APTES and 
AHAMTES as shown in Table 4 and Figure 21. The absence of intra-molecular 
catalysis in AHAMTES is not only manifested by its higher stability but also 
higher reproducibility of the silane layer thickness after hydrolysis. 
47 
 
 
The fluctuating thicknesses of silane layers observed during hydrolysis 
tests could be explained by two reasons. First, the loss of aminosilane molecules 
from the substrate surface to the surrounding water is governed by a chemical 
equilibrium. Secondly, the substrate surface might be contaminated.   
 
 
Figure 20. Thickness of aminosilane-derived layers before and after 24-h 
exposure to water and the percent of remaining thickness. Solution-phase 
silanization was carried out in anhydrous toluene at 70 °C for 24 h.  
 
Table 4. Thickness of APTES- and AHAMTES-derived layers before and after 
exposure to water. 
 
Hydrolysis time (h) 
Thickness (Å) 
APTES AHAMTES 
0 
1 
3 
12 
24 
22 ± 5 
11 ± 1 
12 ± 3 
9 ± 1 
10 ± 3 
22 ± 1 
14 ± 2 
15 ± 1 
20 ± 4 
16 ± 1 
0
10
20
30
40
50
60
70
80
90
100
0
5
10
15
20
25
30
35
40
45
Percentage (%)Thickness (Å) Initial thickness
Thickness after 24-h hydrolysis
Percentage of remaining thickness
APTES APTMS AEAPTMS AHAMTES
48 
 
 
 
 
Figure 21. Thickness of APTES- and AHAMTES-derived layers against 
hydrolysis time (1 h, 3 h, 12 h, and 24 h). Solution-phase silanization was carried 
out in anhydrous toluene at 70 °C for 24 h. 
 
Number of alkoxy groups. To confirm the superior efficiencies of 
constructing silane layers with tri-alkoxy aminosilane reagents, solution-phase 
silanization was carried out with 3-aminopropyldimethylethoxysilane (APDMES, 
97% pure) under the same reaction conditions. As shown in Table 5 and Figure 22, 
the initial thicknesses of its silane layers were close to that of a monolayer 
independent of the reaction time. After 24-h soaking in water, almost all the silane 
layers degraded. This is most likely due to the presence of a single ethoxy group 
per molecule that limits the number of anchoring sites on a substrate surface and 
eliminates the possibility of cross-linking with neighboring molecules. 
Additionally, the methyl groups on APDMES prevent the formation of silane 
0
5
10
15
20
25
30
0 5 10 15 20 25
Thickness (Å)
Hydrolysis time (h)
APTES AHAMTES
49 
 
 
monolayers of high density, and hence the ease of water penetration in between 
the attached aminosilane molecules renders them hydrolytically unstable. As a 
result, tri-alkoxy aminosilanes are more efficient functionalizing reagents.  
Table 5. Thickness and water contact angle (θA/θR) data of APDMES-derived 
layers before and after 24-h exposure to water. 
 
 
Reaction 
time (h) 
Initial After 24-h hydrolysis 
Thickness 
(Å) 
Contact angle 
(deg) 
Thickness 
(Å) 
Contact angle 
(deg) 
APDMES 
(97%) 
1 
2 
24 
5 ± 1 
6 ± 1 
5 ± 1 
(52 ± 2)/(40 ± 2) 
(52 ± 2)/(35 ± 2) 
(52 ± 2)/(40 ± 2) 
0 ± 1 
1 ± 1 
1 ± 1 
(21 ± 2)/(17 ± 2) 
(43 ± 2)/(17 ± 2) 
(25 ± 2)/(16 ± 2) 
 
 
Figure 22. Thickness of APDMES-derived layers before and after 24-h exposure 
to water against reaction time. Solution-phase silanization was carried out in 
anhydrous toluene at 70 °C for 1 h, 2 h, and 24 h, respectively. 
 
Type of alkoxy group. In theory, tri-methoxy aminosilanes are more 
reactive than tri-ethoxy aminosilanes since methoxy is a better leaving group. Our 
-2
-1
0
1
2
3
4
5
6
7
8
0 5 10 15 20 25
Thickness (Å)
Reaction time (h)
Initial thickness Thickness after 24-h hydrolysis
50 
 
 
results corroborate this prediction. Based on the comparison between APTES and 
APTMS of comparable purities (Table 2), the thickness of APTES-derived layer 
(23 ± 6 Å) was more reproducible than that of the APTMS-derived layer (26 ± 15 
Å). Similar comparisons between the reagents in Group II and III are not available 
due to the low purity of AEAPTES and the unavailability of methoxy-based 
Group III reagents. 
Faster silanization with tri-methoxy aminosilanes is less controllable and 
could render the silane layers less regular in structure. At the solid-solution 
interface, the tri-methoxy aminosilane molecules are deposited on the substrate 
more rapidly. Before the molecules could orient themselves in favorable 
conformations, more molecules are deposited on top. Polymerization of tri-
methoxy aminosilanes in solution is also expected to occur at faster rates. 
Consequently, tri-methoxy aminosilane-derived layers are more irregular and less 
reproducible. 
In general, there is a trade-off between rate and reproducibility in 
silanization reactions. However, since both trimethoxy and triethoxy aminosilane 
reagents were observed to give rise to silane layers with similar thicknesses in the 
same of amount reaction time, tri-ethoxy aminosilanes that confer higher 
reproducibility are more desirable.  
Summary. Generally, silane layers produced in solution-phase reactions 
are typically multilayer in nature with some surface roughness. The 
reproducibility of solution-phase silanization is highly dependent on the purity of 
51 
 
 
a silane coupling agent and the ambient humidity. As a result, purification of the 
silane coupling agent and maintenance of a dry reaction system are essential for 
generating reproducible silane layers. The hydrolytic stability of aminosilane-
derived layers can also be enhanced by extending reaction time, which does not 
necessarily alter their surface morphologies. Furthermore, two major 
considerations are involved in choosing a suitable silane coupling agent, which 
are the nature (methoxy or ethoxy) and the number (mono-, di-, or tri-alkoxy) of 
the alkoxy group and the molecular structure of the aminosilane molecule (Group 
I, II, or III). Based on this study, AHAMTES, a Group III reagent with the highest 
purity of 99.66%, was shown to generate the most reproducible and hydrolytically 
stable silane layers via solution-phase silanization.  
4.2 Vapor-phase Silanization 
Overview. The significant dependence of solution-phase silanization on 
silane purity and ambient humidity led us to vapor-phase silanization. In a closed 
environment, a silane coupling agent reaches an equilibrium vapor pressure at 
moderate temperatures. The silane molecules in the vapor phase can react with 
silanol groups on silicon wafers at the solid-vapor interface. Impurities in the 
form of aggregates do not readily vaporize due to their higher boiling points and 
therefore they will not deposit on a substrate surface. The impact of ambient 
humidity should also be minimal since the reaction system can be kept 
consistently drier in the absence of a solvent. Developing a reliable protocol for 
52 
 
 
vapor-phase silanization hence greatly simplifies substrate modification with 
aminosilane reagents.    
Vapor-phase silanization was attempted using silane coupling agents from 
all three groups. The reaction conditions such as temperature, time, and drying 
conditions were optimized to maximize the reproducibility and hydrolytic 
stability of aminosilane-derived layers. The hydrolytic stability was analyzed as a 
function of time by submerging freshly silanized samples in Milli-Q water at 40 
°C for various amounts of time up to 24 h or 48 h. Thickness information, contact 
angle data, AFM images, and XPS analyses were acquired before and after 
hydrolysis and analyzed to assess the hydrolytic stability of respective samples.  
Initial survey. Initial studies were carried out with the Group II reagents 
AEAPTES and AEAPTMS to survey the factors that could play critical roles in 
vapor-phase reactions.  
First of all, vapor-phase silanization is found to be indeed independent of 
silane purity. Shown in Table 6, the AEAPTES of a lower purity (97.4%) 
generated more uniform and reproducible silane layers than the AEAPTMS of a 
higher purity (98.3%). As mentioned earlier, the impurities in the silane coupling 
agent do not vaporize easily and thus do not react with the substrate surface. 
Therefore, purification of a silane coupling agent is not essential for vapor-phase 
reactions.  
Secondly, the initial thickness of aminosilane-derived layers can be 
increased by raising the reaction temperature. When the vapor-phase reactions 
53 
 
 
were performed for 48 h without purging the reaction tubes, the reactions at 90 °C 
generated thicker silane layers than those at 70 °C, as shown in Table 6. Raising 
the reaction temperature increases the vapor pressure and the reaction rate, 
allowing more aminosilane molecules to deposit on the substrate surfaces. After 
48-h soaking in water, the remaining thicknesses of AEAPTES-derived layers 
were close to that of a monolayer, suggesting that the two chosen temperatures 
were sufficiently high for obtaining stable silane structures.  
The third factor is the amount of water in the reaction system. It was 
observed that without purging, the reproducibility of the initial thickness was 
lower at the higher reaction temperature, which was improved by purging the 
reaction tube under N2 for 4 min (Table 6). However, this led to the formation of 
thinner multilayers.  Reducing the amount of water in the reaction system 
decreases the chances of vertical polymerization of aminosilane molecules and 
hence enhances the regularity of aminosilane-derived layers.  
Furthermore, the nature of alkoxy group also affects the reproducibility 
and hydrolytic stability of aminosilane-derived layers. Under all three sets of 
reaction conditions, the thicknesses of AEAPTMS-derived layers were much less 
reproducible than those of AEAPTES-derived layers. The lower reproducibility of 
AEAPTMS could be explained by the higher reactivity of tri-methoxy 
aminosilanes. Consistent with our observations in the solution-phase silanization, 
tri-ethoxy aminosilanes gave rise to more reproducible silane layers in the vapor-
phase reactions and therefore are the more preferred tri-alkoxy aminosilanes.  
54 
 
 
Table 6. Thickness of AEAPTES- and AEAPTMS-derived layers before and after 
48-h exposure to water. 
 
 
 Thickness (Å)  
48 h, 70 °C, 
no purging under N2 
48 h, 90 °C, 
no purging under N2 
48 h, 90 °C, 4 min- 
purging under N2 
Initial After Initial After Initial After 
AEAPTES 
(97.4%) 
14 ± 2 11 ± 1 27 ± 22 9 ± 3 15 ± 4 10 ± 3 
AEAPTMS 
(98.3%) 
157 ± 109 4 ± 2 521 ± 366 13 ± 9 34 ± 35 7 ± 1 
  
Fine-tuning reaction conditions – minimizing the impact of ambient 
humidity. In the initial survey, it was observed that purging reaction tubes under 
N2 for 4 min improved the reproducibility of the experiments. To further remove 
excess water and minimize the impact of ambient humidity, purging time was 
extended to 30 min. In addition, the seal of the glassware was improved by 
tightening the joints. 
Adjusting reaction time and temperature. In the attempt to enhance the 
reproducibility of the experiments, 24-h vapor-phase reactions at 70 °C with 30-
min purging under N2 were carried out with the Group I reagents (APTES & 
APTMS) and Group II reagents (AEAPTES & AEAPTMS). The initial 
thicknesses of the AEAPTMS- and AEAPTES-derived layers decreased to 19 ± 
17 Å and 5 ± 1 Å respectively (Table 7 and Figure 23). Compared to the 48-h 
reactions at 90 °C with 4-min purging (Table 6), these adjustments improved the 
reproducibility of the data at the expense of the thickness of the silane layers.  
55 
 
 
Shown in Table 7 and Figure 23, the initial thickness of the APTMS-
derived layer was less reproducible than that of the APTES-derived layer, which 
again is probably because tri-methoxy aminosilanes have higher reactivity, 
resulting in more sample variations. Therefore, only tri-ethoxy aminosilanes were 
focused on in the vapor-phase study.  
In addition, it is apparent that the APTES- and AEAPTES-derived layers 
generated in the extensively purged reaction systems at 70 °C were too thin to be 
hydrolytically stable. Reaction temperature was therefore set at 90 °C for the 
subsequent experiments with all tri-ethoxy aminosilanes. 
Table 7. Thickness of the Group I and II aminosilane-derived layers before and 
after exposure to water. 
 
24 h, 70 °C, 30 min- 
purging under N2 
Purity 
(%) 
Thickness (Å) 
Initial 
After 1-h 
hydrolysis 
After 24-h 
hydrolysis 
Group I 
APTES 
APTMS 
99.7 
98.5 
6 ± 1 
13 ± 9 
2 ± 1 
9 ± 7 
3 ± 1 
9 ± 5 
Group II 
AEAPTES 
AEAPTMS 
97.4 
98.3 
5 ± 1 
19 ± 17 
3 ± 1 
9 ± 5 
2 ± 1 
4 ± 1 
 
 
56 
 
 
 
Figure 23. Thickness of aminosilane-derived layers before and after exposure to 
water. Vapor-phase silanization was carried out at 70 °C for 24 h. 
 
Vapor-phase silanization with AHAMTES. When the 48-h vapor-phase 
reactions were carried out with the Group III reagent AHAMTES at 90 °C with 
30-min purging under N2, the initial thickness was extremely low as shown in 
Table 8. Attempts to increase the silane layer thickness were made by eliminating 
the purging step to retain water in the reaction system and increasing the reaction 
temperature to 100 °C and 110 °C. Under adjusted conditions, the silane layers 
were either still too thin or unstable in water. 
 As mentioned in the introduction section, the amine group in AHMATES 
is only capable of catalyzing silanization inter-molecularly.  The silane molecules 
are much further apart in the vapor phase than in the solution phase, and hence 
intermolecular interactions are likely to be absent. Intra-molecular amine-catalysis 
0
5
10
15
20
25
30
35
40
Thickness (Å)
Initial After 1-h hydrolysis After 24-h hydrolysis
APTES APTMS AEAPTES AEAPTMS
57 
 
 
thus appears to be important for obtaining silane layers with appreciable 
thicknesses in vapor-phase silanization at moderate temperatures. AHAMTES is 
the best choice for solution-phase silanization due to its high purity and molecular 
structure as discussed earlier, however, it is shown here that AHAMTES or any 
Group III aminosilane reagent is not suitable for vapor-phase silanization due to 
the lack of intra-molecular amine-catalysis.    
Table 8. Thickness of AHAMTES-derived layers before and after 48-h exposure 
to water. 
 
 
Thickness (Å) 
48 h, 90 °C,  
purging under N2 
48 h, 90 °C,  
no purging  
48 h, 100 °C,  
no purging  
48 h, 110 °C,  
no purging  
Initial After Initial After Initial After Initial After 
AHAMTES 
(99.66%) 
2 ± 1 1 ± 1 2 ± 1 1 ± 1 23 ± 5 5 ± 1 140 ± 97 8 ± 5 
 
Vapor-phase silanization with APTES and AEAPTES. Based on all the 
initial studies, we decided to focus on investigating vapor-phase silanization with 
the Group I reagent APTES and the Group II reagent AEAPTES. The reaction 
temperature, 90 °C, was found to be sufficiently high for producing aminosilane-
derived layers with acceptable thicknesses and homogeneous morphologies 
(Table 9). The impact of reaction time was analyzed by characterizing the 
hydrolytic stability and surface morphology of the silane layers prepared by either 
24-h or 48-h reactions.  
Hydrolytic stability. The hydrolytic stability of aminosilane-derived layers 
was analyzed by examining their thickness and contact angle after different 
58 
 
 
durations of soaking in water. For the APTES-derived layers, under purged 
conditions, increasing the reaction time from 24 h to 48 h did not increase the 
thickness or the hydrophobicity of the surfaces, indicated by their contact angle 
measurements (Table 9a). The percent loss of thickness over time (Figure 24) did 
not differ significantly between the samples prepared under these conditions. 
However, when the reactions were allowed to take place without purging, the 
initial thicknesses increased for both sets of conditions. The samples prepared in 
the unpurged 48-h reactions had the greatest initial thickness, which degraded to a 
value close to that of a monolayer. The dynamic contact angles were also higher 
at each time point, indicating the more hydrophobic nature of the surfaces. 
Additionally, the percent loss of thickness over time was more gradual than the 
samples prepared in the 24-h unpurged silanization, indicating a slower rate of 
loss. These observations suggest that increasing the reaction time coupled with a 
slight increment of the amount of water in the reaction system could enhance the 
hydrolytic stability of the aminosilane-derived layers by increasing their initial 
thickness. However, the major drawback of eliminating the purging step is to 
expose the reaction to fluctuating ambient humidity that decreases the 
reproducibility of the experiments. 
For the AEAPTES-derived layers, increasing the reaction time increased 
their initial thicknesses under both purged and unpurged conditions. Particularly 
under the purged conditions, the samples prepared by 48-h silanization exhibited 
the slowest rate of loss (Figure 24). It is worth noting that these vapor-phase 
59 
 
 
reactions under purged conditions generated silane layers corresponding to 
monolayer thickness:  ~5 Å for APTES and ~8 Å for AEAPTES. 
In summary, the Group I and II reagents with either a primary or a second 
amine group at propyl position, which can catalyze silanization intra-molecularly, 
yield silane layers with appreciable thicknesses on silicon wafers at moderate 
reaction temperatures in the vapor phase. The thickness and contact angle data 
suggest that AEAPTES-derived layers are more hydrolytically stable than 
APTES-derived layers, similar to our findings in the solution-phase silanization; 
the most stable amine-functionalized surface was prepared by silanization with 
AEAPTES at 90 °C for 48 h under purged condition. This observation can be 
explained by the molecular structures of Group II reagents. Their longer chains 
are less likely to bend toward the silicon atom due to steric hindrance and 
therefore intra-molecularly catalyzed hydrolysis of siloxane bonds occurs to a 
much less extent.     
Table 9. Thickness and contact angle (θA/θR) data of aminosilane-derived layers 
before and after exposure to water: (a) APTES and (b) AEAPTES. 
 
(a) APTES 
 
Reaction 
time  
(h) 
Initial 
After 1-h 
hydrolysis 
After 3-h 
hydrolysis 
After 24-h 
hydrolysis 
T.* (Å) 
C.A.* 
(deg) 
T. (Å) 
C.A. 
(deg) 
T. (Å) 
C.A. 
(deg) 
T. (Å) 
C.A. 
(deg) 
 
Purged 
 
24 
48 
5 ± 1 
5 ± 1 
50/23 
51/23 
3 ± 1 
2 ± 1 
40/17 
36/14 
3 ± 1 
2 ± 1 
36/14 
36/16 
3 ± 1 
3 ± 1 
34/15 
37/15 
Not 
purged 
24 
48 
8 ± 1 
10 ± 2 
56/32 
55/26 
4 ± 1 
6 ± 2 
40/19 
46/21 
3 ± 1 
6 ± 3 
35/19 
42/20 
3 ± 1 
6 ± 3 
31/17 
36/19 
 
 
60 
 
 
(b) AEAPTES 
 
Reaction 
time  
(h) 
Initial 
After 1-h 
hydrolysis 
After 3-h 
hydrolysis 
After 24-h 
hydrolysis 
T. (Å) 
C.A. 
(deg) 
T. (Å) 
C.A. 
(deg) 
T. (Å) 
C.A. 
(deg) 
T. (Å) 
C.A. 
(deg) 
 
Purged 
 
24 
48 
6 ± 1 
11 ± 2 
50/24 
51/22 
4 ± 1 
8 ± 1 
42/19 
41/18 
3 ± 1 
8 ± 1 
40/16 
40/17 
3 ± 1 
8 ± 1 
35/17 
38/16 
Not 
purged 
24 
48 
10 ± 1 
17 ± 1 
52/26 
58/34 
8 ± 2 
12 ± 1 
46/21 
46/21 
7 ± 3 
12 ± 2 
42/21 
45/20 
6 ± 1 
8 ± 1 
36/19 
39/20 
*T. stands for thickness and C.A. stands for contact angle. The standard deviation 
in the contact angle measurement is smaller or equal to 2°. 
 
 
Figure 24. Percent loss of thickness against hydrolysis time: (a) APTES-derived 
layers and (b) AEAPTES-derived layers. Vapor-phase silanization was carried 
out at 90 °C for 24 h or 48 h.  
 
Surface morphology. The surface morphologies of the APTES- and 
AEAPTES-derived layers were examined by AFM. Shown in Figure 25, the 
surfaces prepared by the 24-h and 48-h reactions under purged conditions were 
extremely smooth with a mean roughness of ~1 Å. After hydrolysis, the surfaces 
0
10
20
30
40
50
60
70
80
90
100
Loss percentage 
(%)
Initial After 1 h After 3 h After 24 h
Initial After 1 h After 3 h After 24 h
24 h & purged 48 h & purged 48 h & not purged24 h & not purged
APTES
AEAPTES
61 
 
 
still appeared smooth and the mean roughness of each sample remained at ~1 Å. 
This confirms the preparation of monolayers using the vapor-phase approach. 
(a) Clean wafer 
Hydrolysis 
time (h) 
(b) 
APTES 
24-h 
silanization 
(c) 
APTES 
48-h 
silanization 
(d) 
AEAPTES  
24-h 
silanization 
(e) 
AEAPTES  
48-h 
silanization 
0 
    
1 
    
3 
    
24 
    
 
Figure 25. AFM images (1 μm×1 μm, height scale at 5 nm) of aminosilane-
derived layers: (a) clean wafer, (b) APTES-derived layers prepared by 24-h 
silanization, (c) APTES-derived layers prepared by 48-h silanization, (d) 
AEAPTES-derived layers prepared by 24-h silanization, and (e) AEAPTES-
derived layers prepared by 48-h silanization. The mean roughness for each 
condition is ~1 Å. All the reactions were carried out under purged conditions at 
90 °C. 
62 
 
 
XPS analysis. XPS analysis was carried out at two take-off angles, 15° and 
45°, to examine the atomic compositions of the substrate surface, or O, Si, C, and 
N contents (Figure 26). Oxygen and silicon come mostly from the native silicon 
oxide layer on the silicon wafers. Carbon and nitrogen are mainly from the 
attached silane layers. In general, hydrolysis or the loss of the aminosilane layers 
should result in the exposure of the underlying silicon wafer substrates, which 
would cause oxygen and silicon contents to increase and carbon and nitrogen 
contents to decrease. The samples were prepared by 48-h silanization at 90 °C 
with 30-min purging under N2. As shown in Table 10, the XPS data obtained at 
45° take-off angle consistently show higher silicon signals (from the underlying 
silicon wafer substrates) and lower oxygen, carbon, and nitrogen signals (from the 
thin native oxide and silane layers) than those obtained at 15° take-off angle.  This 
is because core electrons ejected at 45° were detected at a greater depth.  Since 
our focus is the outermost silane layers, only the data obtained at 15° will be 
examined. 
As shown in Table 10, for the AEAPTES-derived layers, extending 
hydrolysis time did not significantly increase the oxygen or the silicon contents, 
suggesting that the attached silane layers were relatively stable on the substrate 
surfaces. The carbon and the nitrogen contents decreased slightly after 1-h 
hydrolysis but then stayed relatively constant, again indicating stable silane 
structures. These observations were consistent with the thickness and contact 
angle measurements, which became stable after 1-h hydrolysis (Figure 27b). 
63 
 
 
The atomic compositions of the APTES-derived layers largely followed a 
similar pattern (Table 10) but experienced a more rapid loss of its nitrogen 
contents than the AEAPTES-derived layers, as shown in Figure 28. Combined 
with the more rapid loss of thickness as shown in Figure 27a, the XPS analysis 
again suggests that the APTES-derived layers were less hydrolytically stable than 
the AEAPTES-derived layers. As discussed earlier, molecular structures of the 
Group II reagents reduce the occurrence of amine catalyzed hydrolysis of siloxane 
bonds. 
                        
(a) 
 
 
 
 
 
 
 
 
 
 
64 
 
 
(b) 
 
 
Figure 26. XPS of AEAPTES-derived layers: (a) initial and (b) after 24-h 
exposure to water. Red features are associated with take-off angle of 15° and blue 
features are associated with take-off angle of 45°.  
 
Table 10. Atomic compositions of APTES- and AEAPTES-derived layers based 
on XPS analysis. 
 
 Hydrolysis time (h) 0 1 3 24 
Take-off angle (deg) 15 45 15 45 15 45 15 45 
Relative 
abundance 
(%) 
APTES 
O 
Si 
C 
N 
42.7 
28.4 
25.3 
3.8 
41.0 
43.8 
13.3 
2.0 
49.0 
33.4 
14.9 
2.8 
42.7 
47.9 
8.3 
1.2 
42.8 
31.3 
23.8 
2.1 
40.5 
45.1 
13.2 
1.3 
46.5 
33.4 
18.1 
2.1 
41.9 
47.9 
8.9 
1.4 
AEAPTES 
O 
Si 
C 
N 
36.3 
25.2 
31.2 
7.4 
37.8 
40.9 
17.2 
4.1 
38.6 
27.8 
27.7 
6.0 
39.1 
41.3 
15.7 
3.9 
39.7 
26.0 
28.3 
6.1 
39.8 
40.4 
16.3 
3.7 
41.0 
25.7 
27.7 
5.7 
39.0 
41.2 
16.1 
3.7 
 
 
65 
 
 
(a)  
(b)  
Figure 27. Thicknesses, contact angles and atomic compositions of aminosilane-
derived layers: (a) APTES and (b) AEAPTES.  Vapor-phase silanization was 
carried out at 90 °C for 48 h under purged conditions. The standard deviations in 
the thickness and contact angle are 1 Å and 2°, respectively. The take-off angle 
for XPS is 15°. 
 
0
1
2
3
4
5
0
10
20
30
40
50
60
70
80
90
100
Thickness (Å)Percentage (%)
O Si C N Thickness
Initial After 1 h After 3 h After 24 h
(51°/23°)
(36°/14°)
(36°/16°) (37°/15°)
0
2
4
6
8
10
12
0
10
20
30
40
50
60
70
80
90
100
Thickness (Å)Percentage (%)
O Si C N Thickness
Initial After 1 h After 3 h After 24 h
(51°/22°)
(41°/18°) (40°/17°) (38°/16°)
66 
 
 
 
Figure 28. Percent loss of nitrogen content of aminosilane-derived layers based 
on XPS analysis at the take-off angle of 15°. 
 
Summary. Reproducibility of vapor-phase silanization is independent of 
silane purity and not affected by ambient humidity under purged conditions. 
Consistent with solution-phase silanization, reactions with tri-ethoxy aminosilanes 
are generally more reproducible than those with tri-methoxy aminosilanes since 
reagents containing ethoxy groups have lower reactivity that makes them more 
tolerant of fluctuating reaction conditions. In addition, the Group II tri-ethoxy 
aminosilanes generate the most hydrolytically stable layers. The silane layers 
derived from the Group I reagents are more susceptible to hydrolysis and Group 
III reagents are not suitable for vapor-phase reactions due to the lack of intra-
molecular catalysis by amine functionality.   
 
 
 
0
10
20
30
40
50
60
70
80
90
100
Percentage loss 
(%)
Initial
After 1 h
After 3 h
After 24 h
APTESAEAPTES
67 
 
 
5. CONCLUSIONS 
This was the first comprehensive study in which all commercially 
available aminosilane reagents were examined and categorized into three groups 
based on their molecular structures.  A total of five representative silanes from 
these three groups were used to establish necessary structural features and 
reaction conditions for obtaining hydrolytically stable aminosilane-derived layers 
on silicon wafer substrates. Categorizing the silane coupling agents into three 
different groups based on their structural features was shown to be a convenient 
method to generalize the common characteristics of each group of reagents. In 
addition, successful attempts were made to optimize reaction conditions for both 
solution- and vapor-phase silanization to enhance the reproducibility and 
hydrolytic stability as well as surface morphology of aminosilane-derived layers. 
Reliable experimental protocols have been developed for both types of 
silanization reactions.  
Solution-phase silanization. Solution-phase silanization was found to be 
highly dependent on silane purity and ambient humidity. The impurities in a 
silane coupling agent may exist in aggregates of aminosilane molecules that 
attach to a substrate surface via few siloxane bonds, disrupting the regular 
patterns of silane layers and hence increasing their susceptibility to hydrolysis. 
Fluctuating ambient humidity could have introduced varied amounts of water to 
the reaction medium that affected the rate and extent of silanization. Both factors 
were observed to decrease the reproducibility and hydrolytic stability of 
68 
 
 
aminosilane-derived layers. The impact of these two factors could be minimized 
by purifying silane coupling agents prior to solution-phase reactions and using 
anhydrous non-polar solvents. In addition, silane layers with multilayer 
characteristics appeared to be more hydrolytically stable and could be obtained by 
extending reaction time. In general, 24-h solution-phase silanization in anhydrous 
toluene at 70 oC was found to be sufficient for generating reproducible and 
hydrolytically stable silane layers with the Group II and III reagents of high purity.     
Vapor-phase silanization. Vapor-phase silanization is independent of 
silane purity since aggregates of aminosilane molecules do not vaporize easily at 
moderate temperatures and thus they do not react with the substrate surface. The 
impact of ambient humidity was also minimal due to a consistently drier reaction 
system in the absence of a solvent. Purging reaction tubes extensively with N2 
was shown to further improve the reproducibility of the experiments. As a result, 
reproducible aminosilane-derived monolayers are generally easier to obtain in 
vapor-phase silanization.  
As mentioned in the introduction section, recent FT-IR studies challenged 
the possibility of siloxane bond formation on silicon wafer substrates in vapor-
phase silanization. This study demonstrated otherwise. With abundant supporting 
data, hydrolytically stable aminosilane-derived layers were prepared through 
vapor-phase reactions, which is consistent with Jonsson et al’s report 33 on 
generating smooth, stable, and reproducible silane layers that exhibited monolayer 
characteristics by the vapor deposition of the Group II reagent AEAPTMS. The 
69 
 
 
lack of siloxane bond formation reported in the FT-IR studies may be caused by 
the use of undecanal triethoxysilane (Figure 29).  The absence of amine functional 
groups does not allow the catalysis of siloxane bond formation, similar to our 
observations with the Group III reagent AHAMTES. 
In general, 48-h vapor-phase silanization at 90 oC was shown to be 
sufficient for generating reproducible and hydrolytically stable silane layers with 
the Group II reagents.     
 
Si
O
O O
O
 
 
Figure 29. Molecular structure of triethoxysilyl undecanal. 
 
Choice of silane coupling agents. As shown in Table 11, Group I reagents, 
which are the most widely used for current research and applications, tend to 
produce silane layers with low hydrolytic stability due to the ready occurrence of 
both intra- and inter-molecularly catalyzed hydrolysis of siloxane bonds. Group II 
reagents are the most ideal for vapor-phase silanization. Amine catalyzed 
hydrolysis of siloxane bonds is reduced by increased steric hindrance due to their 
longer chains. Group III reagents are only suitable for solution-phase silanization 
since it is incapable of intra-molecular catalysis.  
Moreover, tri-ethoxy aminosilanes produce silane layers with higher 
hydrolytic stability and reproducibility than tri-methoxy aminosilanes since the 
70 
 
 
reagents with ethoxy groups have lower reactivity and are more forgiving of 
fluctuating reaction conditions. 
Table 11. Hydrolysis schematics of attached silane layers from three different 
groups. 
 
Group Hydrolysis 
Si
NH2
RO
OR
OR
 
I. Primary amine can 
coordinate to Si 
        
O O O
Si Si
H2N
Si
H2
N
NH2
    
OH O OH
Si
NH2
  
Si
NH
RO
OR
OR
NH2
II. Secondary amine 
can coordinate to Si 
O O O
Si Si Si
HN
NH2
NH
HN
NH2
O
Si
HN
NH2
NH2
             
O OH O
Si Si
HN
NH2
HN
NH2
O
Si
HN
NH2
 
HN
Si
H2
N
H3CH2CO
OCH2CH3
OCH2CH3
X
 
III. No amine can 
coordinate to Si 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
H2O
O O O
Si Si Si
HN HN NH
MH2 NH2 H2N
O O
Si
HN
NH2
O
H2O
H2O
O O O
Si Si Si
HN HN NH
MH2 NH2 H2N
O O
Si
HN
NH2
O
71 
 
 
REFERENCES CITED 
1. Ulman, A. Chem. Rev. 1996, 96, 1533. 
2. Swalen, J. D.; Allara, D. L.; Andrade, J. D.; Chandross, E. A.; Garoff, S.; 
Israelachvili, J.;  McCarthy, T. J.; Murray, R.; Pease, R. F.; Rabolt, J. F.; 
Wynne, K. J.; Yu, H. Langmuir 1987, 3, 932. 
3. Love, J. C.; Estroff, L. A.; Kriebel, J. K.; Nuzzo, R. G.; Whitesides, G. W. 
Chem. Rev. 2005, 105, 1103. 
4. Bigelow, W. C.; Pickett, D. L.; Zisman, W. A. J. Colloid Interface Sci. 
1946, 1, 513. 
5. Vuillaume, D.; Boulas, C.; Collet, J. Appl. Phys. Lett. 1996, 69, 1646. 
6. Li, Y. S.; Wang, Y.; Tran, T.; Perkins, A. Spectrochim. Acta Part A 2005, 
61, 3032. 
7. Chen, K.; Caldwell, W. B.; Mirkin, C. A. J. Am. Chem. Soc. 1993, 115, 
1193. 
8. Choi, J. U.; Lim, C. B.; Kim, J. H.; Chung, T. Y.; Moon, J. H.;  Hahn, J. 
H.;  Kim, S. B.; Park, J. W. Synth. Met. 1995, 71, 1729. 
9. Kapur, R.; Rudolph, A. S. Exp. Cell Res. 1998, 244, 275. 
10. Almeida, A. T.; Salvadori, M. C.; Petri, D. F. S. Langmuir 2002, 18, 6914. 
11. Sugimura, H.; Nakagiri, N. J. Am. Chem. Soc. 1997, 119, 9226. 
12. Chang, Y. C.; Frank, C. W. Langmuir 1998, 14, 326. 
13. Emoto, K.; Van Alstine, J. M.; Harris, J. M. Langmuir 1998, 14, 2722. 
14. Kohler, N.; Fryxell, G. E.; Zhang, M. J. Am. Chem. Soc. 2004, 126, 7206. 
15. Emoto, K.; Harris, J. M.; Van Alstine, J. M. Anal. Chem. 1996, 68, 3751. 
16. Malsten, M.; Emoto, K.; Van Alstine, J. M. J. Colloid Interface Sci. 1998, 
202, 507. 
17. Miksa, D.; Irish, E. R.; Chen, D.; Composto, R. J.; Eckmann, D. M. 
Biomacromolecules 2006, 7, 557. 
18. Hu, J.; Wang, M.; Weier, H. U. G.; Frantz, P.; Kolbe, W.; Ogletree, D. F.; 
Salmeron, M. Langmuir 1996, 12, 1697. 
19. Callow, M. E.; Callow, J. A.; Ista, L. K.; Coleman, S. E.; Nolasco, A. C.; 
Lopez, G. P. Appl. Environ. Microbiol. 2000, 66, 3249. 
20. Eisenberg, J. L.; Piper, J. L.; Mrksich, M. Langmuir 2009, 25, 13942. 
21. Guo, L. Y.; Zhao, Y. P. J. Adhes. Sci. Technol. 2006, 20, 1281. 
22. Martwiset, S. ; Koh, A. E. ; Chen, W. Langmuir 2006, 22, 8192. 
23. Smith, E. A.; Chen, W. Langmuir 2008, 24, 12405. 
24. Fadeev, A. Y.; McCarthy, T. J. Langmuir 2000, 16, 7268. 
25. Vandenberg, E. T.; Bertilsson, L.; Liedberg, B.; Uvdal, K.; Erlandsson, R.; 
Elwing, H.; Lundstron, I. J. Colloid Interface Sci. 1991, 147, 103. 
26. Zhang, F.; Srinivasan, M. P. Langmuir 2004, 20, 2309. 
27. Zhdanov, S. P.; Kosheleva, L. S.; Titova, T. I. Langmuir 1987, 3, 960. 
28. Simon, A.; Cohen-Bouhacina, T.; Porte, M. C.; Aime, J. P.; Baquey, C. J. 
Colloid Interface Sci. 2002, 251, 278. 
72 
 
 
29. McGovern, M. E.; Kallury, K. M. R.; Thompson, M. Langmuir 1994, 10, 
3607. 
30. Howarter, J. A.; Youngblood, J. P. Langmuir 2006, 22, 11147. 
31. Waddell, T. G.; Leyden, D. E.; DeBello, M. T. J. Am. Chem. Soc. 1981, 
103, 5303. 
32. Pasternack, R. M.; Amy, S. R.; Chabal, Y. J. Langmuir 2008, 24, 12963. 
33. Jonsson, U.; Olofsson, G.; Malmqvist, M.; Ronnberg, I. Thin Solid Films 
1985, 124, 117. 
34. Fiorilli, S.; Rivolo, P.; Descrovi, E.; Ricciardi, C.; Pasquardini, L.; Lunelli, 
L.; Vanzetti, L.; Pederzolli, C.; Onida, B.; Garrone, E. J. Colloid Interface 
Sci. 2007, 12, 41. 
35. Ek, S.; Iiskola, E. I.; Niinisto, L.; Vaittinen, J.; Pakkanen, T. T.; Keranen, 
J.; Auroux, A. Langmuir 2003, 19, 10601. 
36. Ek, S.; Iiskola, E. I.; Niinisto, L. Langmuir 2003, 19, 3461. 
37. Ek, S.; Iiskola, E. I.; Niinisto, L. J. Phys. Chem. B 2004, 108, 9650. 
38. Tian, R.; Seitz, O., Li, M.; Hu, W.; Chabal, Y. J.; Gao, J. Langmuir 2010, 
26, 4563. 
39. Jenkins, Francis A. and Harvey E. White. In Fundamentals of Optics; 
McGraw Hill Inc., New York, 1976. 
40. J.A. Woollam Co., Inc. Ellipsometry Solutions. 
http://www.jawoollam.com (accessed April 5, 2011). 
41. Theoretical Aspects and Instrumentation of Ellipsometry handout. 
42. Wikipedia page. http://en.wikipedia.org/wiki/File:Ellipsometry_setup.jpg 
(accessed April 5, 2011). 
43. Young, T. Philos. Trans. R. Soc. London 1805, 95, 65. 
44. Wikipedia page. http://en.wikipedia.org/wiki/File:Contact_angle.svg 
(accessed April 5, 2011). 
45. Smith, E.A. The Hydrolytic Stability of Aminosilane-functionalized 
Silicon Surfaces. Thesis, Mount Holyoke College, South Hadley, MA, 
2008. 
46. Thermo Fisher Scientific page. 
http://www.thermo.com/eThermo/CMA/PDFs/Various/File_9223.pdf 
(accessed April 5, 2011). 
47. Wenzel, R. N. Ind. Eng. Chem. 1936, 28, 988. 
48. Cassie, A. B. D.; Baxter, S. Trans. Faraday Soc. 1944, 40, 546. 
49. Rugar, D.; Hansma, P. Phys. Today 1990. 
50. Binnig, G.; Quate, C. F. Phys. Rev. Lett. 1986, 56, 930. 
51. Veeco Metrology Group. Scanning Probe Microscopy Training Notebook 
2000. 
52. Hollander, J. M.; Jolly, W. L. Acc. Chem. Res. 1970, 3, 193. 
53. Fadley, C. S. Prog. Surf. Sci. 1984, 16, 275. 
54. Belgian Nuclear Research Center Page. 
http://www.sckcen.be/microstructure/Infrastructure/XPS/Infrastructure_X
PS.htm (accessed April 5, 2011).