The electron transfer (ET) process of near-infrared (NIR) quantum dots (QDs) is crucial to photonic system efficiency. Typically, chemical configurations are changed to tune the ultrafast ET rate. However, direct tuning of the ET rate in QD-molecular acceptor systems while maintaining the chemical configurations remains a challenge. To this end, high pressure can be used as a powerful external control knob. Herein, pressure tuning of the ultrafast ET rate in NIR lead sulfide (PbS)–anthraquinone (AQ) complexes was experimentally performed using in situ high-pressure ultrafast transient absorption spectroscopy. As pressure changes, ET lifetimes shorten. The results show that the promotional ET in the NIR range was assigned to the shortened distance between PbS and AQ under compression. This study thus indicates that pressure can effectively tune the ET rate and provides brand routes toward efficient NIR QD-based photonic applications.

Near-infrared (NIR) quantum dots (QDs) have recently attracted research attention because of their promising applications in photonic and optoelectronic devices that partially rely on the utilization of NIR photons, facile solution processing, and easy bandgap tuning.1,2 In particular, lead sulfide (PbS) nanocrystal QDs are promising candidates for future advances in PbS-based devices owing to their excellent high molar extinction coefficients, broad-bandwidth gain media, and tunable and broad spectral responses.3–5 PbS QDs are used as potential charge separation centers in optoelectronic devices. The efficiency of such devices primarily depends on the electron transfer (ET) process from photoinduced PbS QDs to electron acceptors through a nanoscale interface in photoexcited QDs.6,7 Therefore, studying the ET process in a QD-based system is significant as it can improve further designing and realize optimal NIR-responsive photonic devices.

In QD-molecular acceptor systems, the ET rate strongly depends on the donor-acceptor distance according to the Marcus ET theory.8,9 The ET rate exponentially increases with the decreasing shell thickness in core/shell QDs,10,11 which is regarded as one approach to change the donor–acceptor distance. Another common approach is to vary the length of the linker molecule (bridge).12–14 Recently, the ET rate was shown to depend on the bridge length between the QDs and the acceptor, suggesting the exponential decay of the ET rate with the bridge length.12,13 In addition, a monotonous increase in the ET rate with decreasing QD size from QDs to molecular acceptors has been proposed in previous studies.15,16 However, traditional studies have focused on changing the chemical configuration to tune the ET rate, and the direct tuning of the ET rate of independent QD-based systems while maintaining the chemical configurations is a challenge.

Herein, a detection method involving NIR transient absorption (TA) spectroscopy combined with in situ high-pressure measurements is proposed to study the influence of pressure on the ET process from photoexcited PbS QDs to anthraquinone-2-carboxylic acid (AQ). Fluorescence quenching confirms that the transfer of photoinduced electrons from excited PbS QDs to the AQ acceptor can occur. Analyzing the NIR femtosecond TA spectra reveals that the ET process and bandgap strongly depend on pressure. Furthermore, we demonstrate that the ET lifetimes are shortened, owing to the shortened donor–acceptor distance between PbS and AQ induced by pressure. Therefore, our results are useful and significant for NIR QD-based materials and nanodevice designs.

The NIR femtosecond TA measurements were performed with PbS QDs and their PbS–AQ complexes in ambient pressure in toluene, as shown in Fig. 1. The NIR TA spectra reveal bleaching of the 1S exciton band (∼1090 nm), which is attributed to the state filling of the 1S electron and hole levels in the photoexcited PbS QDs.17 Broad positive absorption has been attributed to the Stark-effect-induced red shift of high energy bands in the presence of the 1S exciton.18Figures 1(a) and 1(b) obviously show that the spectral amplitude of free PbS QDs was greater than that of the PbS–AQ complexes at the same delay time, which can be attributed to the ET process between PbS and AQ.17 

FIG. 1.

Spectroscopic behavior of the electron transfer (ET) process at 1 atm. (a) TA spectra of PbS QDs in toluene. (b) Transient absorption (TA) spectra of lead sulfide (PbS)–anthraquinone (AQ) complexes in toluene. (c) Transient absorption kinetics of PbS quantum dots (QDs) and PbS–AQ complexes under identical conditions. Solid lines represent the fittings. (d) Fluorescence spectra of the PbS QDs and PbS–AQ complexes (15×) measured at the same conditions in toluene. The negative features at 1142.6 and 1191.9 nm were due to absorption of the solvent.

FIG. 1.

Spectroscopic behavior of the electron transfer (ET) process at 1 atm. (a) TA spectra of PbS QDs in toluene. (b) Transient absorption (TA) spectra of lead sulfide (PbS)–anthraquinone (AQ) complexes in toluene. (c) Transient absorption kinetics of PbS quantum dots (QDs) and PbS–AQ complexes under identical conditions. Solid lines represent the fittings. (d) Fluorescence spectra of the PbS QDs and PbS–AQ complexes (15×) measured at the same conditions in toluene. The negative features at 1142.6 and 1191.9 nm were due to absorption of the solvent.

Close modal

To probe charge transfer with temporal resolution, the TA kinetic traces of PbS QDs and PbS QD–AQ with global analysis are shown in Fig. 1(c). Double-exponential decays were applied for free PbS QDs, obtaining two temporal constants (τ1 ∼ 0.40 ± 0.108 ps and τ2 ∼ 38.54 ± 2.577 ps). Similarly, the three temporal constants (τ1 ∼0.44 ± 0.070 ps, τ2 ∼ 9.25 ± 0.557 ps, and τ3 ∼ 62.46 ± 3.706 ps) for PbS–AQ complexes were recorded with three-exponential decays. The same time scale of hundreds of femtoseconds and picoseconds was detected, which is attributed to the reportorial carrier relaxation19 and auger recombination (AR) lifetime in correlative PbS QDs,16,20 respectively. The AR lifetime for PbS–AQ complexes was found to be longer than that for free PbS QDs in the presence of AQ owing to the reduced exciton lifetime caused by adsorbates in QDs.21 The picosecond domain could represent the ET process time for PbS–AQ complexes in the obtained component.8,22 Hole transfer is impossible here because the highest occupied molecular orbital of the AQ molecule is lower than the valence band of the PbS QDs. The lack of spectral overlap of PbS QD emission with the absorption of AQ molecules renders fluorescence resonance energy transfer to be unlikely in this system. Finally, the ET process could be caused by the reduction and oxidation potentials of PbS and AQ shown in Table S1 and Fig. S1, respectively. Figure 1(d) shows that the PbS QD fluorescence is drastically quenched in the presence of the AQ molecule, implying that the ET process occurred in PbS–AQ complexes.23 Notably, toluene has some absorption features that slightly influence the measured photoluminescence (PL) spectra of PbS–AQ complexes.24 We conclude that the picosecond domain was proposed photoinduced ET from PbS QDs to AQ. This observation is consistent with previous results obtained for ET in PbS–TiO2.22 

To understand the influence of pressure on the relaxation dynamics process in PbS–AQ complexes, the detected NIR TA spectra and relevant kinetic traces at different pressures in toluene are presented in Fig. 2. There is a negative feature centered at 1090 nm in the TA spectra at atmospheric pressure presented in Fig. 2(a), which has been attributed to the bleach of the ground-state band edge absorption of the QDs. Figures 2(a)–2(c) clearly show that the pressure-induced red-shift of the excitonic peak from 1090 to 1210 nm occurs during the compression process in liquid–solid phase toluene (cognate peak value shown in Table S2), which indicates that the pressure-induced PbS QD bandgap could change.25 Moreover, the pressure-dependent TA kinetic traces of PbS QD–AQ with global analysis are shown in Figs. 2(e) and 2(f). The kinetic curves at different pressures for the QD-AQ complexes are well fitted by three exponential and pressure-dependence fitted lifetime constants detailed below.

FIG. 2.

Near-infrared (NIR) TA spectra and relevant decay curves of PbS–AQ complexes under different pressures in toluene. (a)–(c) TA spectra of PbS–AQ complexes at 1 atm, 0.60 GPa (liquid phase), and 1.34 GPa (solid phase). (d)–(f) Kinetics of TA spectra of PbS–AQ complexes measured in toluene at atmospheric pressure, 0.60 GPa (liquid phase), and 1.34 GPa (solid phase). Solid lines represent the fittings.

FIG. 2.

Near-infrared (NIR) TA spectra and relevant decay curves of PbS–AQ complexes under different pressures in toluene. (a)–(c) TA spectra of PbS–AQ complexes at 1 atm, 0.60 GPa (liquid phase), and 1.34 GPa (solid phase). (d)–(f) Kinetics of TA spectra of PbS–AQ complexes measured in toluene at atmospheric pressure, 0.60 GPa (liquid phase), and 1.34 GPa (solid phase). Solid lines represent the fittings.

Close modal

The stress-induced energy (E) representing the 1S bleach signal peak is plotted in Fig. 3(a). For the detected experimental data, the fitted solid line was obtained by applying the following quadratic equation:26,27

EP=E0+αP+βP2.
(1)

Here, α and β are pressure coefficients. The relationship between E and the selected pressure was fitted according to Eq. (1). We obtained the values −88.35 and 9.75 meV GPa−1 for α and β, respectively. Simultaneously, the fitted E0 is 1.136 eV, which is in accordance with the 1S exciton absorption peak shown in Fig. S2. However, a previous work demonstrated that the freezing pressure of toluene is 0.84 GPa (295 K).28 Here, toluene is labeled as the liquid phase and solid phase, with the liquid–solid transition pressure for toluene being 0.8 GPa at room temperature. Hence, considering two parts divided by the liquid–solid transition pressure is more reasonable. The pressure dependence of E can be characterized by a linear fit using the equation E(P)=E0+αP in every phase of toluene. The fitting values of α are −84.43 and −73.47 meV GPa−1 [Fig. 3(a)], revealing that the compressibility is different in the liquid and solid phases of toluene, respectively.29 

FIG. 3.

Pressure dependence of the PbS QD bandgap and diameter. (a) Stress-induced energy in the PbS–AQ complex. Solid lines represent linear fittings in the liquid and solid phases of toluene. The inset shows the pressure-dependent energy with least squares fit, and the solid line reveals the fittings. (b) Evolution of the PbS QD diameter with pressure.

FIG. 3.

Pressure dependence of the PbS QD bandgap and diameter. (a) Stress-induced energy in the PbS–AQ complex. Solid lines represent linear fittings in the liquid and solid phases of toluene. The inset shows the pressure-dependent energy with least squares fit, and the solid line reveals the fittings. (b) Evolution of the PbS QD diameter with pressure.

Close modal

Variation in the PbS QD bandgap was revealed according to the red shift of the 1S bleach peak in Fig. 2 under compression. The nanoparticle bandgap E(R) closely depends on the nanoparticle radius, which can be characterized by the following expression:30,31

ER=Eg+221me*+1mh*π2R21.786e2ϵ2R+e2Rn=1αn(SR)2n¯,
(2)

where αn=[(ε1)(n+1)/ε2(εn+n+1)], and Eg and R are the bulk bandgap and radius of the nanoparticle, respectively. Evidently, the nanoparticle radius affects each term in Eq. (2). Further, Murnaghan's equation can describe the pressure-related QD radius32 

D=D0PB/B+11/3B.
(3)

Here, B, B′, and P represent the bulk modulus, its pressure derivation, and the pressure. Figure 3(b) shows the pressure-related PbS QD diameter used herein, and the diameter of the PbS QD is observed to decrease under elevated pressures; this is in accordance with previous theoretical results.29,33

Consequently, the red-shift of the bleach peak reveals that bandgap narrows under elevated pressures. According to previous studies, researchers have demonstrated that pressure can increase strong lattice contraction, which induces an increase in confinement energy.25,32,34,35 Thus, we attribute the red-shift of the bleach peak to the quantum confinement energy.

To understand the dynamic behaviors of excited-state carriers for PbS–AQ complexes under pressure, the dependence of carrier relaxation (τ1), ET (τ2), and AR (τ3) lifetime constants on pressure (the schematics reported in Fig. S3) was investigated in this Letter. Three-exponential decay can be applied to fit the kinetics of PbS–AQ complexes (the corresponding fitting lifetime values are listed in Table S3); the variations in τ1, τ2, and τ3 under compression are presented in Fig. 4. τ1, τ2, and τ3 can be significantly modulated by compression. In detail, the lifetime τ1 value sharply increases as the pressure changes from 1 atm to 0.8 GPa [shown in Fig. 4(a)]. Nevertheless, in 0.8–1.5 GPa, the τ1 value decreases during the compression process. Obviously, a saltation point exists at 0.8 GPa. Furthermore, Figs. 4(b) and 4(c) clearly demonstrate that lifetimes τ2 and τ3 markedly decrease under elevated pressures. However, the τ2 and τ3 values gradually increase under pressures of 0.8–1.5 GPa. In addition, to demonstrate that the adsorbates affect the lifetime constant, the carrier relaxation and AR lifetime constants for the free PbS QDs are also shown in Fig. S4 (the identical variation of carrier relaxation and AR could also be observed; the corresponding fitting lifetime constants are shown in Table S4).

FIG. 4.

Kinetic lifetimes for (a) carrier relaxation, (b) ET, and (c) Auger recombination as a function of pressure. Magenta represents the liquid phase of toluene solvent, and yellow represents the solid phase. Images under atmospheric pressure (d), liquid phase (e), solid phase (f). The spots in the pictures are the rubies used to calibrate the pressure.

FIG. 4.

Kinetic lifetimes for (a) carrier relaxation, (b) ET, and (c) Auger recombination as a function of pressure. Magenta represents the liquid phase of toluene solvent, and yellow represents the solid phase. Images under atmospheric pressure (d), liquid phase (e), solid phase (f). The spots in the pictures are the rubies used to calibrate the pressure.

Close modal

Figure 4(a) shows that the time constant τ1 decelerated during the compression process in liquid-phase toluene. The distribution of localized trap states near the band edge may broaden when the radius of the PbS QDs reduces under compression,36,37 which eventually suppresses the carrier relaxation rate. Upon further compression, the pressure-restrained carrier relaxation lifetime undergoes an apparent decrease. This observed acceleration of carrier relaxation can be explained by the increased confinement energy as the radius of PbS QDs decreases15,38 during the compression process.

According to the Marcus ET theory, the nonadiabatic ET rate strongly depends on the donor–acceptor distance (rDA) in QD-molecular complexes, described as describing as kETexpβrDA, where β is the decay constant.8 In this study, the pressure dependence of solvent toluene, PbS QDs, or AQ could change rDA.

In this Letter, pressure acts on the PbS–AQ complexes through toluene acting as the pressure transmission medium under gradually elevated pressures. In accordance with Marciniak's work,28 the toluene molecular volume decreases with pressure, thereby shortening rDA and thus facilitating the ET process.8 Hence, the pressure dependence of time constant τ2 exhibits reduction of liquid-phase toluene, as shown in Fig. 4(b). The reduced rDA is generated from the larger compression degree of toluene molecular volume from 1 atm to phase transition pressure point (0.8 GPa),28 causing a marked decrease in τ2 under elevated pressures. In solid-phase toluene under further increased pressure, the time constant τ2 is suppressed, which can be ascribed to the changes in toluene volume and the radius of PbS QDs. On the one hand, the different lattice structures in two phases could have different effects on the direction of pressure transmission, as shown in Figs. 4(d)–4(f) of toluene optical images. Figures 4(d)–4(f) indicate that the morphology of toluene is uniformly dispersed in the liquid phase, whereas that in the solid phase appears to be crystallized, resulting in a slight decrease in the molecular volume of toluene as compared with the toluene volume in the liquid phase.8 These differences are the exactly macroscopic behaviors of transition of the lattice structure in different phases of toluene. On the other hand, under further increased pressure, in addition to the decrease in toluene molecular volume, the radius of PbS QDs also needs to be considered as well. The radius of PbS QD decreases with increasing pressure, which can strongly enhance confinement energy as Eq. (2),15,38 resulting in the suppression of the ET process for PbS–AQ complexes. Obviously, the influence of radius decreasing induced by pressure is the dominant factor for the ET process of PbS–AQ complexes in toluene phases. The time constant τ2 increases as the quantum confinement effect is enhanced in the solid phase of toluene.

Considering independent PbS QDs and separate AQ molecules, the pressure-induced phase transition of PbS QDs from rock salt to orthorhombic structure occurs at about 7 GPa,25,39 and the pressure-related phase transition of AQ occurs at 2.3 GPa.40 We believe that the phase transition of independent PbS and separate AQ molecules could not occur in the implemented pressure range because of the absence of a mutation point in our experimental results preceding 1.5 GPa. Consequently, the pressure-dependent structural parameters and lattice of the AQ molecule have no influence on the dynamic process of PbS–AQ complexes in the present study.

Figure 4(c) reports the evolution of the AR process with pressure and shows that lifetime τ3 decreases as a function of pressure. The decreased AR lifetime was interpreted according to the depressed trap state in the PbS QDs under compression.37 Upon further compression, the AR lifetime of the PbS QDs gradually increases in solid-phase toluene. The longer AR lifetime may be predominantly attributed to the significant surface nonradiative recombination suppressed by the effective stress-tuned coupling of electronic states38 under increasing pressure.

In summary, the unexpected ET process of the PbS–AQ complexes can be regulated with increasing pressure, as detected via NIR TA spectroscopy combined with in situ high-pressure measurements herein. Across pressure-dependent liquid–solid phase transition of the toluene solvent, ET lifetimes shorten. In detail, the ET process in the NIR was enhanced owing to the shortened distance between PbS and AQ under externally applied pressure. As the pressure varied, the ET lifetime gradually increased, which was attributed to the enhanced quantum confinement effect. These experimental findings provide insights into the possible ways of ET rate tuning by pressure and underlying compression, highlighting potential applications in developing efficient solar-to-fuel conversion.

See the supplementary material for the experimental section, the stress-induced 1S bleach signal peak positions, the steady-state spectral absorption of the PbS QDs, and the tables of fitting lifetimes for PbS-AQ complexes and free PbS QDs.

We are grateful to the Center for High Pressure Science and Technology Advanced Research (Changchun branch). This work was supported by the National Basic Research Program of China (Grant No. 2019YFA0307701), the National Natural Science Foundation of China (No. 11874180), the Young and Middle-aged Scientific and Technological Innovation leaders and Team Projects in Jilin Province (20200301020RQ), the Science and technology development project of Jilin Province of China (Grant No. 20190103101JH), the Natural Science Foundation of Jilin Province of China (Grant No. 20190201138JC).

The data that support the findings of this study are available from the corresponding author upon reasonable request.

1.
R. D.
Schaller
and
V. I.
Klimov
,
Phys. Rev. Lett.
92
,
186601
(
2004
).
2.
P.
Santanu
,
D. S.
Francesco
, and
B.
Yu
,
Nat. Nanotechnol.
14
,
72
(
2019
).
3.
Z.
Wang
,
C.
Schliehe
,
T.
Wang
,
Y.
Nagaoka
,
Y. C.
Cao
, and
H.
Weller
,
J. Am. Chem. Soc.
133
,
14484
(
2011
).
4.
B.
Yang
and
K. L.
Han
,
Acc. Chem. Res.
52
,
3188
(
2019
).
5.
M. I.
Nugraha
,
H.
Matsui
,
S.
Watanabe
,
T.
Kubo
,
R.
Häusermann
,
S. Z.
Bisri
,
M.
Sytnyk
,
W.
Heiss
,
N. A.
Loi
, and
J.
Takeya
,
Adv. Electron. Mater.
3
,
1600360
(
2017
).
6.
J.
Tang
,
K. W.
Kemp
,
S.
Hoogland
,
K. S.
Jeong
,
H.
Liu
, and
E. H.
Sargent
,
Nat. Mater.
10
,
765
(
2011
).
7.
T.
Nishihara
,
H.
Tahara
, and
M.
Okano
,
J. Phys. Chem. Lett.
6
,
1327
(
2015
).
8.
H.
Zhu
,
N.
Song
, and
T.
Lian
,
J. Am. Chem. Soc.
132
,
15038
(
2010
).
9.
R. A.
Marcus
,
J. Chem. Phys.
43
,
3477
(
1965
).
10.
H.
Zhao
,
H.
Liang
,
B.
Gonfa
,
M.
Chaker
,
T.
Ozaki
,
P.
Tijssen
,
F.
Vidal
, and
D. L.
Ma
,
Nanoscale
6
,
215
(
2014
).
11.
Y. S.
Park
,
W. K.
Bae
,
L. A.
Padilha
,
J. M.
Pietryga
, and
V. I.
Klimov
,
Nano Lett.
14
,
396
(
2014
).
12.
H.
Wang
,
E. R.
Mcnellis
,
S.
Kinge
,
M.
Bonn
, and
E.
Cánovas
,
Nano Lett.
13
,
5311
(
2013
).
13.
Z. W.
Wang
,
K.
Finkelstein
,
C.
Ma
, and
Z. L.
Wang
,
Appl. Phys. Lett.
90
,
113115
(
2007
).
14.
B. R.
Hyun
,
A. C.
Bartnik
,
L.
Sun
,
T.
Hanrath
, and
F. W.
Wise
,
Nano Lett.
11
,
2126
(
2011
).
15.
A.
Gocalinska
,
M.
Saba
, and
F.
Quochi
,
J. Phys. Chem. Lett.
1
,
1149
(
2010
).
16.
V. I.
Klimov
,
A. A.
Mikhailovsky
,
D. W.
McBranch
,
C. A.
Leatherdale
, and
M. G.
Bawendi
,
Science
287
,
1011
(
2000
).
17.
B. L.
Wehrenberg
and
P.
Guyot-Sionnest
,
J. Am. Chem. Soc.
125
,
7806
(
2003
).
18.
S. E.
Kohn
,
P. Y.
Yu
,
Y.
Petroff
,
Y. R.
Shen
,
Y.
Tsang
, and
M. L.
Cohen
,
Phys. Rev. B
8
,
1477
(
1973
).
19.
G.
Nootz
,
L.
Padilha
,
L.
Levina
,
V.
Sukhovatkin
,
S.
Webster
, and
L.
Brzozowski
,
Phys. Rev. B
83
,
155302
(
2011
).
20.
Y.
Gao
,
C. S. S.
Sandeep
,
J. M.
Schins
,
A. J.
Houtepen
, and
L. D. A.
Siebbeles
,
Nat. Commun.
4
,
2329
(
2013
).
21.
Y.
Yang
,
W.
Rodríguez-Córdoba
, and
T.
Lian
,
Nano Lett.
12
,
4235
(
2012
).
22.
B. R.
Hyun
,
Y. W.
Zhong
, and
A. C.
Bartnik
,
ACS Nano
2
,
2206
(
2008
).
23.
24.
M.
Masteri-Farahani
and
K.
Khademabbasi
,
J. Lumin.
204
,
130
(
2018
).
25.
K.
Bian
,
B. T.
Richards
,
H.
Yang
,
W.
Bassett
,
F. W.
Wise
,
Z. W.
Wange
, and
T.
Hanrath
,
Phys. Chem. Chem. Phys.
16
,
8515
(
2014
).
26.
W.
Shan
,
W.
Walukiewicz
,
J. W.
Ager
,
K. M.
Yu
,
J.
Wu
, and
E. E.
Haller
,
Appl. Phys. Lett.
84
,
67
(
2004
).
27.
C. T.
Yuan
,
Y. C.
Lin
,
Y. N.
Chen
,
Q. L.
Chiu
,
W. C.
Chou
,
D. S.
Chuu
,
W. H.
Chang
,
H. S.
Lin
,
R. C.
Ruaan
, and
C. M.
Lin
,
Nanotechnology
18
,
185402
(
2007
).
28.
J.
Marciniak
,
J.
Bąkowicz
,
M. A.
Dobrowolski
,
K. F.
Dziubek
, and
A.
Katrusiak
,
Cryst. Growth. Des.
16
,
1435
(
2016
).
29.
P.
Podsiadlo
,
B.
Lee
, and
E. V.
Shevchenko
,
Nano. Lett.
11
,
579
588
(
2011
).
30.
L. E.
Brus
,
J. Chem. Phys.
80
,
4403
(
1984
).
31.
S.
Sapra
and
D.
Sarma
,
Phys. Rev. B
69
,
1253041
(
2004
).
32.
K. K.
Zhuravlev
,
J. M.
Pietryga
,
R. K.
Sander
, and
R. D.
Schaller
,
Appl. Phys. Lett.
90
,
043110
(
2007
).
33.
S.
Wang
,
J.
Zhang
, and
Y. S.
Zhao
,
Inorg. Chem.
52
,
8638
(
2013
).
34.
S. H.
Tolbert
,
A. B.
Herhold
,
C. S.
Johnson
, and
A. P.
Alivisatos
,
Phys. Rev. Lett.
73
,
3266
(
1994
).
35.
E.
Pedrueza
,
A.
Segura
,
R.
Abargues
,
J. B.
Bailach
,
J. C.
Chervin
, and
J. P.
Martinez-Pastor
,
Astron. Asitrophys.
24
,
205701
(
2013
).
36.
P. T.
Erslev
,
H. Y.
Chen
,
J.
Gao
,
M. C.
Beard
,
A. J.
Frank
,
J. V. D.
Lagemaat
,
J. C.
Johnson
, and
J. M.
Luther
,
Phys. Rev. B
86
,
155313
(
2012
).
37.
G. J.
Xiao
,
Y. N.
Wang
,
D.
Han
,
K. X.
Li
,
X. F.
Feng
,
P. F.
Lv
,
K.
Wang
,
L.
Liu
,
S. A. T.
Redfern
, and
B.
Zou
,
J. Am. Chem. Soc.
140
,
13970
(
2018
).
38.
E. V.
Ushakova
,
A. P.
Litvin
, and
A. V.
Baranov
,
ACS Nano
6
,
8913
(
2012
).
39.
F.
Bai
,
K.
Bian
,
X.
Huang
,
Z.
Wang
, and
H.
Fan
,
Chem. Rev.
119
,
7673
(
2019
).
40.
A.
Brillante
,
R. G.
Della Valle
,
R.
Farina
, and
E.
Venuti
,
Chem. Phys.
191
,
177
(
1995
).

Supplementary Material