of 17
Ultrafast Formation of Charge Transfer Trions at Molecular
-
Functionalized 2D MoS
2
Interfaces
Yuancheng Jing
,
[
a
]
Kangkai Liang
,
[b]
Nicole S. Muir
,
[
a
]
Hao Zhou
,
[b]
Zhehao Li
,
[b]
Joseph M.
Palasz
,
[
a
]
Jonathan Sorbie
,
[
a
]
Chenglai Wang
,
[
a
]
Scott K Cushing
,
[
c
]
Clifford P. Kubiak
,
[
a
]
Zdeněk
Sofer
,
[
d
]
Shaowei Li
,
[
a
]
,
[b]
Wei Xiong
*
[a]
,[b]
[a]
Y. Jing, N. S. Muir, J. M. Palasz, J. Sorbie, C. Wang, C. P. Kubiak, S. Li, W. Xiong
Department of Chemistry and Biochemistry
University of California, San Diego
9500 Gilman Drive, MC 0358, La Jo
lla, California 92093
-
0358, United States
E
-
mail:
w2xiong@ucsd.edu
[b]
K. Liang, H. Zhou, Z. Li, S. Li, W. Xiong
Material Science and Engineering Program
University of California, San Diego
9500 Gilman Drive, MC 0418, La Jolla, California 92093
-
0418
, United States
[
c
]
S. K. Cushing
Division of Chemistry and Chemical Engineering
California Institute of Technology
1200 E California Blvd, MC 127
-
72, Pasadena, California 91125, United States
[
d
]
Z. Sofer
Department of Inorganic Chemistry
University of Chemistry and Technology
,
Prague
Technická 5, 166 28 Prague 6, Czech Republic
Table of Contents
S1. Sample Preparation
................................
................................
................................
................................
.
3
S2. Transient E
-
SFG spectrometer
................................
................................
................................
................
4
S3. Additional photoluminescence measurement
................................
................................
........................
5
S4. Cyclic voltammetry for Ru(DPPZ)
2
L
-
Pro
................................
................................
................................
..
6
S5. UV
-
vis absorption
................................
................................
................................
................................
....
7
S
6
.
Atomic
force microscopy (AFM)
................................
................................
................................
..............
8
S
7
.
Energy Dispersive X
-
ray (EDX) elemental mapping
................................
................................
.................
9
S
8
. Photo Generated Carrier Density Estimation
................................
................................
........................
10
S
9
. Regular vs.
Defect
-
rich Ru
-
MoS
2
................................
................................
................................
...........
10
S
10
. Static and transient ESFG of Ru(DPPZ)
2
L
-
pro
................................
................................
......................
11
S
11
. Feynman Diagram of tr
-
E
SFG.
................................
................................
................................
.............
12
S
12.
Kinetic model equation
................................
................................
................................
.......................
13
A. pristine monolayer
-
MoS
2
................................
................................
................................
....................
13
B. Ru
-
MoS
2
................................
................................
................................
................................
..............
14
S
13
. Globe analysis result on Tr
-
ESFG
................................
................................
................................
.........
15
Reference
................................
................................
................................
................................
....................
17
S
1. Sample Preparation
Monolayer MoS
2
is prepared following the method from Liu
[1]
: gold tape is used for
exfoliating
monolayer
MoS
2
from bulk sample
(natural MoS
2
, collected in Krupka, North of Czech Republic)
.
This gold tape, crafted from thermal release tape (TRT) that picked up a polyvinylpyrrolidone
(PVP)
-
coated 150 nm gold layer from an ultra
-
flat silicon wafer, was gently pressed onto a freshly
cleaved MoS
2
bulk sample. The gold layer, along with the attached MoS
2
monolayer, was then
transferred onto a SiO
2
substrate usi
ng heat. The following cleaning process involved sequential
applications of deionized water, acetone, and O
2
plasma (PE
-
100) to remove PVP from the
surface of gold and residual polymers from the TRT. Gold etching was performed using a
potassium iodide solution, which was then cleaned by isopropanol and deionized water, ensuring
a pristine
monolayer MoS
2
.
The MoS
2
monolayer samples can have a size of up to a few
millimeters. FigureS1 (a) shows a homogeneous monolayer flake used for Tr
-
ESFG under a 10x
optical microscope. This flake is roughly 1000μm by 300μm and can be located by naked eyes
during experiments (Figu
re S1 (b)).
Figure
S1
. Optical image of MoS
2
(a) monolayer flake under 10x optical microscope; (b) optical image of monolayer flake on the substrate
The complex [Ru(
D
PPZ
)
2
L
-
Pro]PF
6
{DPPZ = dipyrido[3,2
-
a:2′,3′
-
c]phenazine), L
-
Pro = L
-
proline}
was prepared through the following literature protocols. Ru(DPPZ)
2
Cl
2
was prepared following the
procedures of Sun
et al.
[2]
The chiral
L
-
proline complex was prepared and enantiomerically
enriched following the chiral auxiliary route detai
led by Meggers and coworkers.
[3]
Briefly,
Ru(DPPZ)
2
Cl
2
was combined with 2 equivalents of
L
-
proline and 0.5 equivalents of K
2
CO
3
in
ethylene glycol and heated to 190
C
for 10 minutes. The mixture turned from a dark purple to a
deep blood red color signaling the disappearance of Ru(DPPZ)
2
Cl
2
.
Deviating from the previously
published protocol, the reaction mixture was diluted with methanol, and the product precipitated
by introduction of a saturated aqueous solution of KPF
6
. The solids were collected by filt
ration
and purified by flash chromatography (silica stationary phase, DCM
-
>15% MeOH in DCM). The
major fraction was the Λ
-
[Ru(DPPZ)
2
(
L
-
Pro)]PF
6
, corroborating the reported characterization of
the Λ stereoisomer. 1H NMR (500. MHz, CD3CN): δ (ppm) 9.77
-
7.38 (m, 20H) 5.38 (br s, 1H),
4.17 (q, J
1,2
= 7.85, 7.82 Hz, 1H), 2.21 (m, 1H), 2.08 (m, 1H), 1.85 (m, 1H), 1.47 (m, 3H) ESI
-
MS
m/z C
41
H
28
N
9
O
2
Ru (M)
+
780.1 (calc) 780.13 (found)
[
Ru(
DPPZ
)
2
L
-
Pro]P
F
6
is dissolved in ethanol
at
a concentration of 1mmol/L.
Monola
yer
MoS
2
on
SiO
2
substrate is immersed in
[
Ru(
DPPZ
)
2
L
-
Pro]P
F6 solution for
12
hr
to facilitate the occurrence
of the molecule self
-
assembly process
. The Ru
-
TMD sample was then rinsed with ethanol for
washing off extra
Ru[(
DPPZ
)
2
L
-
Pro)]P
F
6
.
S
2. Transient E
-
SFG spectrometer
A high
-
repetition
-
rate femtosecond laser system (Pharos, 100kHz, Light Conversion) was used
for the
home
-
built
transient ESFG measurement, employing a pump
-
probe geometry. The Pharos
laser generates pulses with a center wavelength of 1030
nm and a pulse width of 160 fs, delivering
a pulse energy of 100 μJ. 50% of the 1030nm pulses are directed to pump a high
-
power optical
parametric amplifier (Orpheus
-
HP, Light Conversion), generating tunable IR pulses. Specifically,
in this work, we
scanned
the center wavelength of the mid
-
IR from 2.08 μm to 2.43μm to match
the photon energy of ESFG with the trion energy of
monolayer
MoS
2
(1.80eV
-
1.72eV). The
remaining 50% of the 1030nm pulse, post
-
OPA, is filtered through an etalon (SLS optics) to create
a
n up
-
conversion pulse with a narrow bandwidth of 4 cm
-
1
FWHM.
Meanwhile, the reflection of
1030nm from etalon was directed through
a
beta barium borate (BBO) crystal to generate 515nm,
serving as the pump beam. The pump beam passes through a variable
-
length delay line before
reaching the sample.
The mid
-
IR beam and the 1030 nm beam
are recombined
colinearly
by a
customized dielectric
optic and
are focused by an f=10
cm parabolic mirror onto the sample
surface with spatially and temporally overlapping to gen
erate the ESFG signal. The incident
angles set at 60° for both beams. All SFG measurements are conducted under PPP polarization
(p
-
SFG, P
-
visible, p
-
IR).
The beams size
for tr
-
ESFG
are: 1030nm (70μm as diameter), IR (97μm
as diameter), for pump: 515nm (153μm as diameter), all determined by the knife
-
edge method.
The SFG signal is dispersed by a spectrograph (Shamrock, Andor) and detected by a charge
-
coupled device (CCD) (Newton idus, Andor).
[4]
A repetition rate of 10 kHz was used to avoid sample burning
damage.
Thereby,
a 100Hz chopper
in the pump beam line modulates the pump beam on and off (100 pulses on and 100 pulses
blocked).
The CCD acquisition
(using cropping mode)
is synchronized to the
trigger of the chopper
to acquire ESFG spectra every 0.1 s.
All measurements take place in a nitrogen
gas purged
environment to eliminate issues related to sample burning and water absorption. Each scan
includes at least 5 averages, and between each average, the sample is repositioned to a different
spot.
The setup
is illustrated in Figure
S
2
.
F
igure S
2
.
Tr
-
ESFG spectrometer
Non
-
resonance SFG signal (see
F
igure
S
3
) from gold is applied as reference to normalize the
static ESFG signal and Tr
-
ESFG signal.
Figure
S
3
.
Gold non
-
resonance SFG signal serves as reference.
For the
Tr
-
ESFG result,
the
ps
eu
do
color
plot (Figure 3)
represents the intensity of pump probe
data at certain time. The
pump probe intensity (
푝푢푚푝
푝푟표푏푒
) is determined by
:
푝푢푚푝
푝푟표푏푒
=
표푛
표푓푓
표푓푓
where
표푛
is
the
intensity of ESFG signal when the sample is pumped by 515nm
laser
beam, while
표푓푓
stands for the intensity of ESFG signal when the 515nm pump beam is blocked (
표푛
and
표푓푓
here has the same meaning in Figure 5b).
S3.
Additional p
hotoluminescence measurement
PL quenching effect is examined under different excitation
wavelengths
. In the manuscript, we
applied 532n
m(2.33eV) laser as the light source for PL. We further applied 633nm(1.95eV) laser
to re
-
test the PL quenching effect after molecular functionalization. Optical microscope of the
monolayer MoS
2
flake is shown in Figure S
4
(a), which was used for this experiment. The mapping
result of pristine MoS
2
and Ru
-
MoS
2
are shown in Figure
S
4
(b) and Figure
S
4
(c), suggesting
strong quenching effect still exist under the incident beam with photon energy lower than the
HOMO/LUMO gap of Ru(DPPZ)
2
L
-
pro, thus charge trans
fers occur at 1.95 eV too.
Figure
S
4
. PL result under excitation wavelength of 633nm to confirm interfacial charge transfer. (a) 20μm by 20μm optical image of mo
nolayer
MoS
2
flake; (b) PL microscope mapping on same pristine monolayer MoS
2
; (c) PL microsco
pe mapping on Ru
-
MoS
2
(d) PL spectrum
comparison on a typical position of monolayer MoS
2
, 55% quenching effect after Ru(DPPZ)
2
L
-
Pro doped; (e) fitting result for PL spectrum of
pristine monolayer MoS2; (f) fitting result of Ru
-
MoS
2
Interestingly, the quenching effect is not as strong as 532nm beam, dropping from 95% to 55%(See
Figure
S
4
(d)), which could either be that the 1.95eV beam is unable to excite exciton B (2.0eV) of the
MoS
2
leading to a smaller density of photo generated carrier, or additional overhead energy is needed
for an effective charge transfer to occur. In Figure
S
4
(e) and Figure
S
4
(f), fitting results show similar
quenching effect for both trion peak, exciton A peak and a relative higher spectrum weight of trion.
Only two gaussian peaks can be fitted also due to low photon energy cannot excite exciton B.
S
4
.
Cyclic voltammetry for Ru(
DPPZ
)
2
L
-
Pro
The HOMO and LUMO values of Ru[(
DPPZ
)
2
L
-
Pro
)] were
estimated
by doing a fitting of the onset
positions of the first notable reduction and oxidation peaks
(
See FigureS5
),
respectively.
[5]
Three
lines were used: one fitted against the curve of the peak, one fitted to the slope of the CV curve,
and one line through the intersection point of the two previous lines. A voltage value is found by
locating a data point along the final line. The LU
MO value
relative to a vacuum (0.00 eV)
was
found using the followin
g equation, where
푟푒푑
,
표푛푠푒푡
is the voltage value determined by fitting the
onset of the reduction peak:
퐿푈푀푂
(
푒푉
)
=
(
푟푒푑
,
표푛푠푒푡
+
퐹푒
/
퐹푒
1
2
+
퐹푒
+
)
(
푒푉
)
Eq S1
Where
퐹푒
/
퐹푒
1
2
+
is the E
1/2
of ferrocene against an Ag/AgCl reference electrode. The
experimentally determined value using a platinum disc working electrode and platinum wire
counter electrode was 0.4253 eV.
퐹푒
+
is the internal standard redox of ferrocene relative to a
vacuum
with a value of 4.80 eV.
The HOMO value was
calculated
using a similar equation, where
표푥
,
표푛푠푒푡
is the voltage value
determined by fitting the onset of the oxidation peak:
퐻푂푀푂
(
푒푉
)
=
(
표푥
,
표푛푠푒푡
+
퐹푒
/
퐹푒
1
2
+
퐹푒
+
)
(
푒푉
)
Eq S2
Cyclic voltammetry
was conducted with an electrochemical workstation (WaveDriver Model AFP1,
Pine Research). A
5.0 mM Ruthenium molecule in 0.1 M TBAPF
6
in
dry acetonitrile at scan rate
of 100
0
mV/s, with a glassy car
bon disc WE, Pt wire CE, and Ag/AgCl RE.
The result is shown in
Figure S
5
.
As a reference, 5.0 mM ferrocene in 0.1 M TBAPF6/dry acetonitrile is included at 25
mV/s.
F
igure S
5
. CV fitting of Ru molecule and Ferrocene
reference.
The calculated LUMO and HOMO values would be
3.54 eV and
-
5.66 eV, respectively. The
HOMO and LUMO equations above were used to estimate the values of 34 scans with similar
features. The average and standard deviation of the entire dataset for LUMO and HO
MO were
3.41±0.06
eV
and
−5.64±0.05
eV
, respectively.
S
5
. UV
-
vis absorption
To further investigate the potential mechanism behind the PL quenching effect, we also ran a UV
-
vis spectroscopy on the pristine
monolayer
MoS
2
,
Ru
-
MoS
2
and pure
Ru(D
PPZ
)
2
L
-
Pro
.
Figure
S
6
(
a
)
orange curve shows t
he pure Ru
(DPPZ)
2
L
-
Pro
shows no absorption at the interested energy
range
.
In Figure S
6 (
b
)
and Figure S
6 (
c
)
, c
ompared to the pristine
monolayer
MoS
2
, Ru
-
MoS
2
has a much lower spectra weight of trion peak. This result indicates a population decrease of a
trion state in
monolayer
MoS
2.
Figure
S
6.
(a) UV
-
vis of
pristine
monolayer
MoS
2
,
Ru
(DPPZ)2L
-
Pro
,
Ru
-
MoS
2
; (b) fitting of
pristine monolayer
MoS
2
(c) fitting of
Ru
-
MoS
2
S
6
.
Atomic
force microscopy
(AFM)
We conducted AFM to determine the
thickness of Ru(DPPZ)
2
L
-
pro
layer and characterize the
monolayer MoS
2
sample:
we synthesized a
50/50
Ru
-
MoS
2
sample with only half of the substrate
immersed
into
Ru(DPPZ)
2
L
-
pro
solution for molecular functionalization. Therefore, we can measure
the
pristine
MoS
2
and Ru
-
MoS
2
on the same substrate to avoid height error. Based on the result of
a
tomic force microscopy
(AFM), the thickness of organic layer is 0.44nm (See Figure
S
7
), suggesting
a single layer of molecular func
tionalization.
Figure
S
7
.
AFM image and line scan of a Ru
-
MoS2 sample.
S
7
.
Energy Dispersive
X
-
ray (EDX)
elemental mapping
We conducted EDX mapping for
additional
characterization. The result is shown as Figure
S
8
. The
sample is Ru
-
MoS
2
on the SiO
2
substrate.
Scanning electron microscope (SEM) of Ru
-
MoS
2
is shown
in Figure
S
8
(a). The distribution of Mo (Figure
S
8
(
b
)
) and S (Figure
S
8
(
c
)
) shows a clear difference
between sample area and substrate. Figure
S
8
(d) shows that the Ru atoms are enriched on the MoS
2
flake. Weight and atomic percentage for the elemental mapping is collected in Table
S
1.
Figure S
8
.
EDX elemental mapping of Ru
-
MoS
2
sam
ple (a) SEM image of a Ru
-
MoS
2
flake, (b) Mo distribution, (c) S distribution, (d) Ru
distribution, (e) O distribution, (f) Si distribution
Table
S
1
. EDX elemental mapping result: weight and atomic percentage is shown.
The density of the Ru(DPPZ)
2
L
-
pro can be estimated by the EDX signal for Ru atom. The atomic
percentage of Ru atoms and Mo atoms can be extracted from Table
S
1. Therefo
re, the density ratio
can be calculated as:
푅푢
푀표
=
푅푢
푀표
=
0
.
353
%
8
.
601%
=
0
.
0
410
Where
푅푢
푀표
is the atomic ratio provided by the EDX result. Then, we can use this atomic ratio with the
Mo density to calculate the density for the molecule. The density of Mo atom can be determined as: