S
1
Supporting Information
for
Synthesis, Characterization, and Reactivity of Ethynyl
-
and Propynyl
-
Terminated Si(111)
Surfaces
Noah T. Plymale,
†
Youn
-
Geun Kim,
‡
Manuel P. Soriaga,
‡
,
⊥
Bruce S. Brunschwig,
‡
,
§
and Nathan
S. Lewis
*
,
†
,
‡
,
§
,||
†
Division of Chemistry and Chemical Engineering,
‡
Joint Center for Artificial Photosynthesis
,
§
Beckman Institute
, and
||
Kavli Nanoscience Institute
,
Cali
fornia Institute of Technology,
Pasadena, California 91125, United States
⊥
Department of Chemistry, Texas A&M University, College Station, Texas 77843, United States
*
Author to whom correspondence should be addressed,
e
-
mail:
nslewis@caltech.edu
A
.
Supporting
Experimental Details.
1. Preparation of Atomically Flat H
–
Si(111)
Surface
s.
Wafers were cut with a diamond
-
tipped scribe to the desired size and then rinsed sequentially with water, methanol (
≥
99.8%,
EMD), acetone (
≥
99.5%, EMD), methanol, and water. Organic contaminants we
re removed and
the surfaces were oxidized by immersing the wafers in a freshly prepared piranha solution
(1:3
v/v of 30% H
2
O
2
(aq) (EMD): 18 M H
2
SO
4
(EMD)) at 90
–
95 °C for 10
–
15 min. The wafers
were rinsed with copious amounts of water and immersed in buffe
red HF(aq) (semiconductor
grade, Transene Co., Inc.) for 18 s followed by another water rinse. Atomically flat H
–
Si(111)
surfaces were prepared by immersing the wafers in an Ar(g)
-
purged solution of NH
4
F(aq).
1
-
2
Wafers with a miscut angle of 0.5° were etched for 5.5 min, while wafers with a
miscut angle of
0.1° were etched for 9.0 min to obtain optimal terrace size. The solution was purged throughout
the etching process and the wafers were agitated after each minute of etching to remove bubbles
S
2
that formed on the surface. After etching, the w
afers were rinsed with water and dried under a
stream of Ar(
g
).
2. Preparation of Cl
–
Si(111) Surfaces.
Cl
–
Si(111) surfaces were prepared inside a N
2
(
g
)
-
purged glove
box with <10 ppm O
2
. A saturated solution of PCl
5
(
≥
99.998% metal basis, Alfa
Aesar) in
chlorobenzene (anhydrous,
≥
99.8%, Sigma
-
Aldrich) was preheated with an initiating
amount (<1 mg mL
–
1
) of benzoyl peroxide (
≥
98%, Sigma
-
Aldrich) for 1
–
2 min. The H
–
Si(111)
wafers were rinsed with chlorobenzene and
then
reacted in the PCl
5
solution at 90 ±
2 °C for 45
min.
1, 3
Upon completion of the reaction, the solution was drained and the wafers were rinsed
with copious amounts of chlorobenzene, followed by tetrahydrofuran (THF, anhydrous,
inhibitor
-
free,
≥
99.9%, Sigma
-
Aldrich).
3. Preparation of of Br
–
Si(111) Surfaces.
Br
–
Si(111) surfaces were prepared by reaction
under ambient light at 22 °C of H
–
Si(111) with Br
2
(g) in a drying chamber connected to a
vacuum line as well as to a reservoir of Br
2
(l). Immediately after anisotropic etching, H
–
Si(111)
samples were placed ins
ide the drying chamber, which was then evacuated to <20 mTorr. The
sample was sealed under vacuum and the Br
2
(l) reservoir was quickly opened and closed to
allow a visible amount of Br
2
(g) into the evacuated drying chamber. The reaction was allowed to
proc
eed for 10 s, after which the Br
2
(g) was removed by vacuum to a pressure of <20 mTorr.
2, 4
The sample was sealed under vacuum and transferred to a N
2
(g)
-
purged gl
ove
box.
B. Supporting Instr
umentation Details.
1. Transmission Infrared Spectroscopy.
A custom attachment allowed Si samples (1.3
×
3.2
cm) to be mounted such that the incident IR beam was either 74° or 30° with respect to the
sample surface normal. At 74° (Brewster’s angle for Si), IR modes parallel and perpendicular to
the surface are ob
served, and at 30°, parallel modes remain visible, while perpendicular modes
S
3
are greatly diminished in intensity.
5
Reported spectra are averages of 1500 consecutive scans
collected at a resolution of 4 cm
–
1
. The baseline was flattened and peaks resulting from water
absorption were subtracted in the reported spectra. Background SiO
x
and H
–
Si(111) spectra were
recorded separately for each sample prior to subsequent functional
ization.
2.
X
-
ray Photoelectron
Spectroscopy
.
Photoelectrons were collected at 90° with respect to the
surface plane of the sample, with the lens aperture set to sample a 700
×
300
μ
m spot. The
instrument was operated by Vision Manager software v. 2.
2.10 revision 5. Survey and high
-
resolution scans were collected with analyzer pass energies of 80 eV and 10 eV, respectively. No
signals from Cl, Br, Mg, Na, or Li impurities were detected on alkylated samples prepared as
described. When HCC
–
Si(111) surfa
ces were prepared using DMA as the solvent, however,
residual Br and N were often observed by XPS.
Thermal stability in vacuum was studied by collecting XP spectra as a function of annealing
temperature. Samples were mounted on a resistive heating stage that consisted of a molybdenum
puck heated with a tungsten wire. Stainless
-
steel clips affixed t
he sample to the molybdenum
stage. The temperature was monitored by a type E thermocouple gauge affixed on the
molybdenum stage immediately below the sample. Samples were heated to the desired
temperature at a ramp rate of 10 °C min
–
1
and were held at the indicated temperature for 30 min.
The samples were allowed to cool to 22
–
30 °C prior to collection of XPS data.
3. Surface Recombination Velocity Measurements.
Electron
-
hole pairs were formed by a 20
ns, 905 nm laser pulse from a
n OSRAM diode laser with an ETX
-
10A
-
93 driver. For each laser
pulse, the decay in reflected microwave intensity was monitored by a PIN diode connected to an
oscilloscope. All recorded decay curves were averages of 64 consecutive decays. Between
measurement
s, samples were stored in air
-
filled centrifuge tubes in the dark.
S
4
C. Data Analysis.
1. Fitting and Quantification of XPS Data.
High
-
resolution XP spectra were analyzed using
CasaXPS software v. 2.3.16. The peak positions for XP spectra were calibrate
d using the Si 2p
3/2
peak, which was set to be centered at 99.68 eV.
6
For bulk Si
0
and Si
1+
doublets, the ratio of the
peak area of the Si 2p
1/2
:2p
3/2
was set to 0.51 and the width of the two peaks was set equal.
6
Shirley backgrounds were used for all high
-
resolution data except when analyzing small amounts
of SiO
x
in the 102
–
104 eV range, for which a linear background was applied. C
1s and F 1s high
-
resolution spectra were fitted using the Voigt function GL(30), which consists of 70% Gaussian
and 30% Lorentzian character. Si 2p photoemission signals for bulk Si
0
an
d Si
1+
species were
fitted using asymmetric Lorentzian line shapes convoluted with a Gaussian of the form
LA(
a
,
b
,
n
), where
a
and
b
determine the asymmetry of the line shape and
n
specifies the
Gaussian width of the function. LA(1.2, 1.4, 200) was found t
o fit consistently. Contributions
from high
-
order SiO
x
in the range of 102
–
104 eV were fit to a single peak using the GL(30)
function.
The thickness (
d
A
)
of the overlayer species
A
was estimated
by XPS
for CH
3
–
Si(111),
HCC
–
Si(111), and CH
3
CC
–
Si(111)
surfaces
using the substrate
-
overlayer model
7
-
8
I
A
I
S
i
!
"
#
$
%
&
S
F
S
i
S
F
A
!
"
#
$
%
&
ρ
S
i
ρ
A
!
"
#
$
%
&
=
1
−
e
−
d
A
λ
A
s
i
n
θ
e
−
d
A
λ
S
i
s
i
n
θ
!
"
#
#
#
$
%
&
&
&
(
S
1)
where
I
A
is the area under the photoemission peak arising from the overlayer species
A
,
I
Si
is the
area under the Si 2p photoemission signal,
SF
Si
is the instrument sensitivity factor for Si 2p
(0.328), and
SF
A
is the instrument sensitivity factor for the overlayer species
A
, w
hich is
0.278
for C 1s photoelectrons in hydrocarbon overlayers
.
For the hydrocarbon overlayers,
I
A
is
the total
area under the C 1s photoemission signal corresponding to all C atoms in th
e overlayer, which
is
S
5
the signal at 284.3 eV for CH
3
–
Si(111) surfaces, the signal at 284.5 eV for HCC
–
Si(111)
surfaces, and the signals at 284.3 and 285.3 eV for CH
3
CC
–
Si(111) surfaces.
For Si
–
OH,
I
A
is
the sum of the area under the Si 2p photoemission sig
nal at 100.5
eV
and 101.1 eV. For SiO
x
,
I
A
is the area under the Si 2p photoemission signal appearing from 102
–
104 eV.
The density of Si
(
ρ
Si
)
is
2.3
g
cm
–
3
, and the density of the overlayer species
A
(
ρ
A
)
is
3.0 g cm
–
3
for
hydrocarbon
overlayers.
9
HCC
–
Si(111) surfaces
exhibited a
fractional monolayer (ML)
coverage of
~0.63
ML,
so
the assumed density of the overlayer was adjusted to model an overlayer with
63
%
of the density of a
full
monolayer
(1.9
g cm
–
3
)
.
When estimating the thickness of Si
–
OH or SiO
x
overlayers, the
quantity (
SF
Si
/
SF
A
)(
ρ
Si
/
ρ
A
)
reduces
to a normalizing constant of 1.3 to account for
the difference in Si 2p photoelectron signal intensity for Si
–
OH or SiO
x
relative to bulk Si.
8
The
attenuation length for the overlayer species (
λ
A
)
has been
estimated to be 3.6
nm
for
C 1s
photoelectrons moving through
hydrocarbon overlayers
10
-
11
or 3.4
nm for
Si 2p photoelectrons
moving through
Si
–
OH or SiO
x
overlayers
.
10
-
11
The attenuation length for Si
2p photoelectrons
(
λ
Si
)
moving through hydrocarbon overlayers
has been
estimated to be 4.0
nm.
10
-
11
For Si
–
OH or
SiO
x
overlayers, the value of
λ
A
=
λ
Si
= 3.4 nm.
The angle between the surface plane and the
photoelectron ejection vector (
θ
) is
90°.
The thickness of the overlayer species A was calculated
using an iterative process.
The fractional monolayer coverage for the overlayer species A (
Φ
A
) was estimated by
dividing the
measured
thickness,
d
A
,
by the calculated thickness of
1 ML of overlayer species A
,
depicted in Scheme S1
.
The thickness of
1 ML of each hydrocarbon overl
ayer
was estimated by
summing the bond lengths for the species containing C, but excluding Si and H
.
For Si
–
OH
overlayers, the thickness of 1 ML was estimated to be the distance from the bottom of the atop Si
atom to the top of the O atom.
The thickness of
1 ML of SiO
x
was estimated to be
0.35
nm.
8, 12
S
6
Assuming uniform overlayers,
the
value of
Φ
A
represents the fraction of surface Si(111) sites
that were modified with the overlayer species of interest.
Scheme S1.
The fractional monolayer coverage for 4
-
fluorobenzyl
-
modified HCC
–
Si
(111) and SiO
x
surfaces was estimated using a
three
-
layer model
13
-
14
I
A
I
S
i
!
"
#
$
%
&
S
F
S
i
S
F
A
!
"
#
$
%
&
ρ
S
i
ρ
A
!
"
#
$
%
&
=
1
−
e
−
d
A
λ
A
s
i
n
θ
e
−
d
A
+
d
B
(
)
λ
S
i
s
i
n
θ
!
"
#
#
#
$
%
&
&
&
(S2)
where
d
A
is the thickness of the bound F atom and
d
B
is the thickness of the hydrocarbon layer
between the Si
crystal and the F atom.
The value of
SF
A
for F 1s photoelectrons is 1.00 and the
density of the overlayer was assumed to be the same as for HCC
–
Si(111) surfaces, 1.9
g cm
–
3
.
For F 1s photoelectrons, the value of
λ
A
is
1.6 nm.
8
Scheme S2 shows the two proposed
structure
s for 4
-
fluorobenzyl
-
modified HCC
–
Si(111) and SiO
x
surfaces along with the calculated
thickness for
d
A
and
d
B
. Since the ratio
d
A
/
d
B
is known from Scheme
S2, eq
S2 can be expressed
in terms of
d
A
and solved using an iterative process. The measured thicknes
s
d
A
was divided by
the calculated thickness of 1 ML of F atoms, 0.13 nm, to give the fractional
monolayer coverage
of 4
-
fluorobenzyl groups.
Si
C
Si
H
H
H
H
0.09 nm
0.06 nm
Si
C
Si
O
H
0.09 nm
0.12 nm
0.06 nm
d
A
= 0.15 nm
d
A
= 0.27 nm
0.09 nm
0.12 nm
H
H
H
0.06 nm
0.15 nm
d
A
= 0.42 nm
0.12 nm
0.16 nm
0.05 nm
d
A
= 0.33 nm
S
7
Scheme S2.
2. Calculation of Surface Recombination Velocity and Surface Trap
-
State Density.
The
minority
-
carrier lifetime (
τ
) was estimated by fitting the microwave conductivity decay versus
time curve to an exponential decay, as described previously.
15
-
16
The calculated values of
τ
were converted to surface recombination velocities (
S
) for wafers of thickness
d
using
15, 17
-
18
(
S
3)
The surface recombination velocity was converted to an effective trap
-
state density,
N
t
, using
1, 18
(
S
4)
where the trap
-
state capture cross section,
σ
, was 10
-
15
cm
2
and the thermal velocity of charge
carriers,
ν
th
, was 10
7
cm s
–
1
.
18
N
t
can be used to estimate the absolute number of electrically
active defects per surface Si(111) sites by use of the number density of atop Si sites for an
unre
constructed Si(111) surface,
Γ
Si(111)
, which is 7.83
×
10
14
atoms cm
–
2
. Thus, a wafer with
surface recombination velocity
S
has 1 electrically active defect for every
Γ
Si(111)
/
N
t
surface
sites.
Si
F
0.05 nm
0.10 nm
0.07 nm
0.12 nm
0.09 nm
0.15 nm
0.13 nm
d
A
d
B
Si
O
d
A
d
B
F
d
A
= 0.13 nm
d
B
= 0.58 nm
0.08 nm
0.05 nm
0.15 nm
0.29 nm
0.07 nm
0.13 nm
d
A
= 0.13 nm
d
B
= 0.64 nm
S
=
d
2
τ
N
t
=
S
σ
ν
t
h
S
8
D. Supporting Figures
and Table
.
Figure S1.
TIRS data for CH
3
–
Si(111) surfaces
,
referenced to the H
–
Si(111) surface
,
collected
at 74° (bottom)
and 30° (top)
from the surface normal
.
Panel
a shows high
-
energy region, and
panel
b
shows the low
-
energy region.
The n
egative peaks in panel b
result
ed
from the
H
–
Si(111)
background.
A sharp
negative peak observed in panel b
at 30° incidence
marked with
∗
at
667
cm
–
1
result
ed
from CO
2
in the atmosphere.
The subscript
s
“CH
3
” and “CH
2
” indicate C
–
H
stretching
signals arising from the
–
CH
3
and
–
CH
2
–
groups, respectively.
The peak
positions
and
assignments
are indicated. The 30° spectrum
is
offset vertically for clarity.
3500
3250
3000
2750
2500
-‐0.5
0.0
0.5
1.0
1.5
2.0
S
i
C
H
3
ν
s
(
C
−
H
)
CH
2
ν
s
(
C
−
H
)
CH
3
ν
a
(
C
−
H
)
CH
2
ν
a
(
C
−
H
)
CH
3
74°
30°
a
2856
2910
2926
2961
Absorbance
Wavenumber
(
cm
−
1
)
x 10
−
3
2000
1750
1500
1250
1000
750
500
0.0
1.0
2.0
3.0
S
i
C
H
3
627
δ
(
Si
−
H
)
ν
(
Si
−
H
)
2083
74°
30°
*
ν
(
Si
−
C
)
ρ
(
CH
3
)
δ
s
(
C
−
H
)
CH
3
b
678
753
1257
Absorbance
Wavenumber
(
cm
−
1
)
x 10
−
3
S
9
2000
1750
1500
1250
1000
750
500
-‐0.5
-‐0.3
0.0
0.3
0.5
x 10
−
3
S
i
H
2083
ν
(
Si
−
H
)
74
°
ν
(
C
=
O
)
*
δ
(
C
−
H
)
*
ν
(
Si
−
O
−
Si
)
TO
ν
(
Si
−
OH
)
ν
(
C
≡
C
)
b
828
2023
Absorbance
Wavenumber
(
cm
−
1
)
δ
(
Si
−
H
)
627
Figure S2.
TIRS data for HCC
–
Si(111) surfaces prepared using DMA
,
referenced to the
H
–
Si(111)
s
urface
and
collected at 74°
incidence
.
Pa
nel
a
shows
the
high
-
energy region, and
pa
nel
b
shows the low
-
energy region.
The n
egative peaks in panel b
result
ed
from the
H
–
Si(111)
background.
The subscript “sat”
is used to denote
C
–
H stretching signals arising from saturated
hydrocarbons.
The peak positions
and assignments (
∗
denote
s
tentative)
are indicated
in the
figure
.
Figure S3.
TIRS data for
(a)
CH
3
–
Si(111),
(b)
HCC
–
Si(111), and
(c)
CH
3
CC
–
Si(111) surfaces
referenced to
the
SiO
x
surface
. The position of the Si
–
H stretching peak is indicated by the
dotted line.
3500
3250
3000
2750
2500
-‐0.5
-‐0.3
0.0
0.3
0.5
S
i
H
ν
a
(
≡
C
−
H
)
ν
s
(
≡
C
−
H
)
ν
(
C
−
H
)
sat
ν
(
C
−
H
)
sat
a
2847
2947
3292
3307
Absorbance
Wavenumber
(
cm
−
1
)
74
°
x 10
−
3
2200
2175
2150
2125
2100
2075
2050
0.0
0.5
1.0
1.5
S
i
C
H
3
S
i
H
S
i
O
H
S
i
C
H
3
c
b
Absorbance
Wavenumber
(
cm
−
1
)
a
ν
(
Si
−
H
)
x 10
−
4
S
10
Figure S
4
.
HREELS data for CH
3
–
Si(111)
surfaces. The data were collected
in the specular
geometry
using
an incident beam energy of 5.0 eV
,
and the
fwhm
of the elastic peak was
13.3
meV.
The raw spectrum (bottom) is shown with the magnified spectrum (top) superimposed
for clarity
.
The peak positions
and assignments
are indicated
in the figure
.
290
288
286
284
282
8
9
10
11
12
13
S
i
C
H
3
C
O
C
C
a
Counts Per Second
Binding Energy (eV)
C
Si
x 10
2
106
104
102
100
98
0
1
2
3
4
5
6
S
i
C
H
3
x 20
Si
0
Counts Per Second
Binding Energy (eV)
b
x 10
3
Figure
S5
.
High
-
resolution XP spectra of the
(a)
C 1s
and
(b)
Si 2p
regions for CH
3
–
Si(111)
surfaces. The low binding
-
energy C photoemission signal at 284.3 eV has been ascribed to C
bound to Si
(blue, C
Si
)
, with the C
1s
signals
at 285.2 and 286.4 eV
arising from
C bound to C
(red, C
C
) and C bound to O (green, C
O
), respectively
. The region from 102
–
105 eV in the Si 2p
spectrum is magnified to show the absence of detectable high
-
order SiO
x
.
0
500
1000
1500
2000
2500
3000
3500
0
1
2
3
4
5
6
S
i
C
H
3
δ
(
Si
−
C
)
ρ
(
CH
3
)
ν
(
Si
−
C
)
ν
(
Si
−
O
−
Si
)
TO
δ
s
(
C
−
H
)
CH
3
δ
a
(
C
−
H
)
CH
3
477
665
747
1066
1267
1399
2927
x 20
Counts
Energy Loss
(
cm
−
1
)
ν
(
C
−
H
)
CH
3
x 10
3
S
11
106
104
102
100
98
0
4
8
12
16
Si
H
Si
O
H
Si
1+
x 20
Counts Per Second
Binding Energy (eV)
Si
0
x 10
3
Figure S6.
High
-
resolution XP
spectrum of the Si 2
p
region for HCC
–
Si(111) surfaces.
Contributions from the bulk Si (blue, Si
0
) and Si
1+
(red) species are indicated. The region from
102
–
105 eV in the Si 2
p
spectrum is magnified to show the absence of detectable high
-
order
SiO
x
.
106
104
102
100
98
0
4
8
12
16
Si
CH
3
x 20
Counts Per Second
Binding Energy (eV)
Si
0
x 10
3
Figure S7.
High
-
resolution XP spectrum of the Si 2
p
region for CH
3
CC
–
Si(111) surfaces. The
Si
2
p
spectrum showed only a contribution from the bulk Si (blue, Si
0
). The region from
102
–
105 eV in the Si 2
p
spectrum is magnified to sh
ow the absence of detectable high
-
order
SiO
x
.
S
12
Figure S
8
.
Thermal stability in vacuum of HCC
–
Si(111) surfaces.
The annealing temperature is
indicated above each spectrum
,
and
the
spectra are offset vertically for clarity.
The survey
spectra (a)
show
ed
the presence of
only the
Si
2p, Si
2s, C 1s, and O 1s core
-
level peaks along
with the O
A
uger signal and Si plasmon
-
loss features. The
high
-
resolution C 1s spectra (b)
exhibit
ed the peaks arising from C in the ethynyl group (C
CCH
) and adventitious C (C
C
and C
O
).
M
inimal change
in the C 1s spectra was observed
upon an
nealing to 200 °C. B
roadening
was
observed
as the
C 1s
peak
at ~285.1 eV
(C
C
)
greatly increased in intensity
upon heating to
300
–
500 °C. Heating to 600
–
700
°C resulted in the appearance of
a new C 1s peak at
~283.5 eV
1200
1000
800
600
400
200
0
0
4
8
12
16
20
24
S
i
H
S
i
O
H
O Auger
O 1s
C 1s
Si 2s
700 °C
600 °C
500 °C
400 °C
300 °C
200 °C
100 °C
22 °C
Counts Per Second
Binding Energy (eV)
Si 2p
a
x 10
5
290
288
286
284
282
12
16
20
24
28
S
i
H
S
i
O
H
C
O
C
C
C
CCH
700 °C
600 °C
500 °C
400 °C
300 °C
200 °C
100 °C
22 °C
Counts Per Second
Binding Energy (eV)
SiC
b
x 10
2
106
104
102
100
98
0
5
10
15
20
25
30
35
S
i
H
S
i
O
H
Si
1+
700 °C
600 °C
500 °C
400 °C
300 °C
200 °C
100 °C
22 °C
Counts Per Second
Binding Energy (eV)
Si
0
c
x 10
3
S
13
(SiC)
.
Si 2p
spectra
(c)
showed
gradual
smoothing of the shoulder
indicated as Si
1+
with
increased annealing temperature, indicating the loss of surficial Si
–
OH and formation of
Si
–
O
–
Si
.
Figure S
9
.
Thermal stability in vacuum of CH
3
CC
–
Si(111) surfaces. The annealing temperature
is indicated above each spectrum
,
and
the
spectra are offset vertically for clarity. The survey
spectra (a) show the presence of only the Si 2p, Si 2s, C 1s, and O 1s core level peaks along with
the O
A
uger signal and Si plasmon
-
loss features.
Annealing to 600 and 700 °C resulted in the
observati
on of a
small amount of Cu and Cl
,
which
was likely transferred from the sample holder
1200
1000
800
600
400
200
0
0
4
8
12
16
20
S
i
C
H
3
Si 2p
Si 2s
C 1s
O 1s
700 °C
600 °C
500 °C
400 °C
300 °C
200 °C
100 °C
22 °C
Counts Per Second
Binding Energy (eV)
O Auger
a
x 10
5
290
288
286
284
282
1
2
3
4
5
S
i
C
H
3
C
C
C
Si
700 °C
600 °C
500 °C
400 °C
300 °C
200 °C
100 °C
22 °C
Counts Per Second
Binding Energy (eV)
SiC
b
x 10
3
106
104
102
100
98
0
5
10
15
20
25
30
S
i
C
H
3
700 °C
600 °C
500 °C
400 °C
300 °C
200 °C
100 °C
22 °C
Counts Per Second
Binding Energy (eV)
Si
0
c
x 10
3