Chinese Journal of Chemical Physics  2018, Vol. 31 Issue (5): 613-618

#### The article information

Qi-chao Shang, Xin-zuo Fang, Hai-long Jiang, Qun Zhang

Location Effect in a Photocatalytic Hybrid System of Metal-Organic Framework Interfaced with Semiconductor Nanoparticles

Chinese Journal of Chemical Physics, 2018, 31(5): 613-618

http://dx.doi.org/10.1063/1674-0068/31/cjcp1803044

### Article history

Received on: March 20, 2018
Accepted on: April 9, 2018
Location Effect in a Photocatalytic Hybrid System of Metal-Organic Framework Interfaced with Semiconductor Nanoparticles
Qi-chao Shanga,b, Xin-zuo Fanga,c, Hai-long Jianga,c, Qun Zhanga,b
Dated: Received on March 20, 2018; Accepted on April 9, 2018
a. Hefei National Laboratory for Physical Sciences at the Microscale, School of Chemistry and Materials Science, University of Science and Technology of China, Hefei 230026, China;
b. Department of Chemical Physics, Synergetic Innovation Center of Quantum Information & Quantum Physics, University of Science and Technology of China, Hefei 230026, China;
c. Department of Chemistry, CAS Key Laboratory of Soft Matter Chemistry, Collaborative Innovation Center of Suzhou Nano Science and Technology, University of Science and Technology of China, Hefei 230026, China
*Author to whom correspondence should be addressed. Qun Zhang, E-mail: qunzh@ustc.edu.cn
Abstract: We report an ultrafast spectroscopy investigation that addresses the subtle location effect in a prototypical semiconductor-MOF hybrid system with TiO$_2$ nanoparticles being incorporated inside or supported onto Cu$_3$(BTC)$_2$, denoted as TiO$_2$@Cu$_3$(BTC)$_2$ and TiO$_2$/Cu$_3$(BTC)$_2$, respectively. By tracking in real time the interface electron dynamics in the hybrid system, we find that the interface states formed between TiO$_2$ and Cu$_3$(BTC)$_2$ can act as an effective relay for electron transfer, whose efficiency rests on the relative location of the two components. It is such a subtle location effect that brings on difference in photocatalytic CO$_2$ reduction using the two semiconductor-MOF hybrids. The mechanistic understanding of the involved interface electron-transfer behavior and effect opens a helpful perspective for rational design of MOF-based hybrid systems for photoelectrochemical applications.
Key words: Metal-organic framework    Photocatalysis    Ultrafast spectroscopy    Transient absorption
Ⅰ. INTRODUCTION

Metal-organic frameworks (MOFs), a class of porous coordination polymers formed by metal ions and organic ligands, have shown great promise in photoelectrochemical applications due to their unique properties such as high specific surface areas, outstanding porosity, and structural diversity [1-3]. MOF-based heterogeneous composites, in which metal or semiconductor nanoparticles (NPs) are interfaced with a MOF material, have been exploited to optimize the desired performance in photocatalysis [4-10]. Recently, in a prototypical metal-MOF hybrid system, i.e., platinum (Pt) NPs interfaced with UiO-66-NH$_2$, we revealed that the Pt location (i.e., either incorporated inside or supported onto the MOF) can significantly influence the photocatalytic efficiency for hydrogen production [6]. In this case the MOF functions as an electron donor while the Pt NPs serve as a co-catalyst, facilitating electron transfer and promoting charge separation therein. During the carrier-transfer process within heterostructures, MOFs could act as either electron donors [4-6] or electron acceptors [7-10]. Notably, it would be more beneficial when the MOF behaves as an electron acceptor in the hybrid system given that the MOF offers extensive active sites for photocatalysis [11]. Recent efforts have been made to develop semiconductor-MOF hybrid systems, in which the MOF receives electrons donated by the interfaced semiconductor component [7-10].

It is interesting to see if the location effect observed in metal-MOF system also manifests in semiconductor-MOF system. A useful way to examine this is to track in real time the interface electron dynamics in the system by ultrafast optical spectroscopy [12-16]. To this end, we have performed a set of femtosecond (fs) time-resolved transient absorption (TA) measurements on a typical semiconductor-MOF system with TiO$_2$ NPs being incorporated inside or supported onto Cu$_3$(BTC)$_2$ (BTC=benzene-1, 3, 5-tricarboxylate), forming the two composites denoted TiO$_2$@Cu$_3$(BTC)$_2$ and TiO$_2$/Cu$_3$(BTC)$_2$, respectively. Analysis of our fs-TA results revealed that the interface states formed between TiO$_2$ and Cu$_3$(BTC)$_2$ can serve as an effective relay for electron transfer, whose efficiency is highly dependent on the relative location of the two components. It is such a subtle location effect that leads to their different photocatalytic performances in CO$_2$ reduction.

Ⅱ. EXPERIMENTS A. Materials

Polyvinylpyrrolidone (PVP, $M_\textrm{W}$=29000) and 1, 3, 5-benzenetricarboxylic acid (H$_3$BTC) were purchased from Sigma-Aldrich. Cuprous oxide (Cu$_2$O) was purchased from Alfa Aesar. Tetrabutyl titanate, $N$, $N$-dimethylformamide (DMF), and ethanol were purchased from Sinopharm Chemical Reagent Co., Ltd. All of the chemicals were used as received without further purification.

B. Synthesis of the samples 1. Synthesis of TiO$_2$

1 mL tetrabutyl titanate was added into a Pyrex vial. The vial was then placed into an oven and kept at 85 ℃. The precursor, tetrabutyl titanate, was hydrolyzed smoothly and the resulting white powders were dehydrated to yield the amorphous sample of TiO$_2$ NPs.

2. Synthesis of Cu$_3$(BTC)$_2$

1.48 g PVP was dissolved in 10 mL pure water and heated at 70 ℃ under magnetic stirring. 8 mL of DMF containing 21 mg H$_3$BTC was added into the PVP solution, which was heated for 10 min. Then, 2 mL of DMF containing 5.6 mg Cu$_2$O was added and further heated at 70 ℃ for 30 min. The product was washed with ethanol three times to remove PVP. The resulting product was dispersed in ethanol again. After 2 h the supernatant was taken out and centrifuged to obtain the Cu$_3$(BTC)$_2$ microcrystals.

3. Synthesis of TiO$_2$@Cu$_3$(BTC)$_2$

5 mg Cu$_3$(BTC)$_2$ was activated by evacuation at 130 ℃ for 1 day in a Pyrex reaction tube. After being cooled to room temperature, Cu$_3$(BTC)$_2$ was immersed with 1 mL tetrabutyl titanate under evacuation for another day. Subsequently, the sample was removed from the Schlenk flask and washed with ethanol to remove tetrabutyl titanate on the surface. Finally, the sample was dried to yield the TiO$_2$@Cu$_3$(BTC)$_2$ composites.

4. Synthesis of TiO$_2$/Cu$_3$(BTC)$_2$

5 mg Cu$_3$(BTC)$_2$ was dispersed in 1 mL ethanol, and then 50 μL tetrabutyl titanate was added into ethanol under stirring. The mixture was dehydrated quickly to yield the TiO$_2$/Cu$_3$(BTC)$_2$ composites.

C. Characterization 1. Routine characterizations

The concentrations of Ti and Cu were measured with a Thermo Scientific Plasma Quad Ⅲ inductively coupled plasma mass spectrometry (ICP-MS) with the samples being dissolved in HCl/HNO$_3$ (3:1, $V$:$V$). The sample morphology of TiO$_2$, Cu$_3$(BTC)$_2$, TiO$_2$@Cu$_3$(BTC)$_2$, and TiO$_2$/Cu$_3$(BTC)$_2$ was examined by transmission electron microscopy (TEM, JEOL-2010). Prior to electron microscopy characterizations, a drop of the aqueous suspension of particles was placed onto a piece of carbon-coated copper grid or silicon wafer and dried under ambient conditions. Powder X-ray diffraction (XRD) was carried out on a Japan Rigaku Smart Lab rotation anode X-ray diffractometer equipped with graphite monochromatized Cu K$\alpha$ radiation ($\lambda$=1.54 Å). The X-ray photoelectron spectroscopy (XPS) characterization was performed on a Thermo Scientific ESCALAB 250 X-ray photoelectron spectrometer, using nonmonochromatized Al K$\alpha$ X-ray as the excitation source. The expected charging of samples was corrected by setting the C 1s binding energy of the adventitious carbon to 284.5 eV. The steady-state UV-Vis diffuse reflectance spectra (270-510 nm) were recorded on a Shimadzu Solid Spec-3700 spectrophotometer (reference: BaSO$_4$).

2. Ultrafast spectroscopy

The fs-TA measurements were performed under ambient conditions in an integrated Helios spectrometer (Ultrafast Systems LLC) with pump and probe beams derived from an amplified femtosecond laser system (Coherent). The 320-nm pump pulses ($\sim$50 nJ/pulse at the sample cell) were delivered by a travelling-wave optical parametric amplifier system (TOPAS-800-fs), which was excited by a Ti:sapphire regenerative amplifier (Legend Elite-1K-HE; 800 nm, 35 fs, 3 mJ/pulse, and 1 kHz repetition rate) seeded with a mode-locked Ti:sapphire laser system (Micra 5) and then pumped with an Nd:YLF laser (Evolution 30). The pump power was adjusted through a series of neutral-density filter wheels. The stable white-light continuum (WLC) probe pulses (450-650 nm) were generated by attenuating and focusing the 800-nm beam (split from the regenerative amplifier, $\sim$400 nJ/pulse) onto a sapphire crystal plate. To correct the pulse-to-pulse fluctuation of the WLC, the WLC was split into a probe beam and a reference beam. The time delays between the pump and probe pulses were controlled by a motorized optical delay line. The instrument response function (IRF) was determined to be $\sim$100 fs by measuring solvent responses under the same experimental conditions (with the exception of a higher excitation power). A mechanical synchronized chopper operating at 500 Hz was used to modulate the pump pulses such that the fs-TA spectra with and without the pump pulses can be recorded alternately. The temporal and spectral profiles (chirp corrected) of the pump-induced differential transmission of the WLC probe light (i.e., absorbance change) were visualized by an optical fiber-coupled multichannel spectrometer (with a complementary metal-oxide-semiconductor sensor) and further processed by the Surface Xplorer software equipped with the Helios system. The samples were well dispersed in a mixture of spectrum-pure acetonitrile (CH$_3$CN) and ethylene glycol. The solution was contained in a 0.7-mL sealed quartz cuvette under a continuous magnetic stirring condition, ensuring that the photoexcited volume of the sample was kept fresh during the fs-TA measurements.

3. Photocatalytic tests

The photocatalytic CO$_2$ reduction tests were carried out in a 100-mL optical reaction vessel under stirring at ambient temperature using a 300-W Xe lamp as the light source. In addition, a 400-nm short-wave-pass cut-off filter was used. The catalyst (10 mg) was dispersed in CH$_3$CN (19 mL) with 1 mL triethanolamine (TEOA) being added as a sacrificial reagent. The mixture was bubbled with CO$_2$ under stirring for 40 min. The product was measured by a gas chromatograph (Shimadzu GC-2014, nitrogen as carrier gas), during which 200 μL of the product was injected and quantified by a calibration plot against the internal CO standard.

Ⅲ. RESULTS AND DISCUSSION

The two composites with an identical mass ratio of TiO$_2$ over Cu$_3$(BTC)$_2$ (i.e., $\sim$0.2) were prepared based on the pre-synthesized Cu$_3$(BTC)$_2$ product. TEM images (FIG. 1) revealed that the MOF morphology (i.e., octahedron-shaped Cu$_3$(BTC)$_2$ microcrystal with a lateral size of $\sim$800 nm) was well maintained upon being integrated with TiO$_2$ NPs. As for TiO$_2$@Cu$_3$(BTC)$_2$, one can see that the surfaces of Cu$_3$(BTC)$_2$ were very smooth (FIG. 1(a)) and TiO$_2$ NPs were well confined within the Cu$_3$(BTC)$_2$ (FIG. 1(b), which is an enlarged view of the selected region in FIG. 1(a)). In contrast, as for TiO$_2$/Cu$_3$(BTC)$_2$, the surfaces of Cu$_3$(BTC)$_2$ were rough and covered with TiO$_2$ NPs (FIG. 1 (c) and (d)). The excellent encapsulation of TiO$_2$ NPs inside MOF for TiO$_2$@Cu$_3$(BTC)$_2$ and nearly uniform coverage of TiO$_2$ with MOF for TiO$_2$/Cu$_3$(BTC)$_2$ clearly indicated the different locations of TiO$_2$ NPs for the two composites. FIG. 2 shows the comparison of the size distribution of TiO$_2$ NPs in TiO$_2$@Cu$_3$(BTC)$_2$ and TiO$_2$/Cu$_3$(BTC)$_2$, which shows no significant difference, precluding the possible impact of size distribution of TiO$_2$ NPs on the interface electron dynamics in the hybrid systems. Moreover, the observation of nearly identical powder XRD patterns (FIG. 3) suggested that the crystalline structure of Cu$_3$(BTC)$_2$ was well maintained for the two composites.

 FIG. 1 Typical TEM images for the two hybrid systems under investigation: (a, b) TiO$_2$@Cu$_3$(BTC)$_2$ and (c, d) TiO$_2$/Cu$_3$(BTC)$_2$. Panels (b) and (d) are the enlarged views of the selected regions (red rectangles) in panels (a) and (c), respectively.
 FIG. 2 Comparison of the size distribution of TiO$_2$ NPs in TiO$_2$@Cu$_3$(BTC)$_2$ and TiO$_2$/Cu$_3$(BTC)$_2$.
 FIG. 3 Powder XRD patterns recorded on TiO$_2$@Cu$_3$(BTC)$_2$ and TiO$_2$/Cu$_3$(BTC)$_2$, showing no obvious difference.

FIG. 4 exhibits the valence-band (VB) XPS spectra recorded on the two composites and the two bare samples of TiO$_2$ and Cu$_3$(BTC)$_2$. The VB edges for TiO$_2$ and Cu$_3$(BTC)$_2$ were determined to be at 2.74 and 2.25 eV, respectively, below the Fermi level. Notably, both composites turned out to hold an identical VB edge at 1.92 eV below the Fermi level, suggesting that the different loading locations of TiO$_2$ NPs in the composites would not bring on significant modification for the electronic structures and properties of Cu$_3$(BTC)$_2$. The survey and high resolution XPS spectra recorded on the two composites can be found in FIG. S1 in the supplementary materials.

 FIG. 4 Valence-band XPS spectra recorded on the two composite samples of TiO$_2$@Cu$_3$(BTC)$_2$ and TiO$_2$/Cu$_3$(BTC)$_2$ as well as the two bare samples of TiO$_2$ and Cu$_3$(BTC)$_2$.

To track the real-time photoexcited electron dynamics in the two hybrid systems, we performed fs-TA spectroscopy characterization. A routine pump-probe configuration with an ultraviolet (UV) pump (320 nm) and a white-light continuum probe (450-650 nm) was adopted. The 320-nm photoexcitation ($\sim$3.88 eV) is suited to induce interband transition in TiO$_2$ (bandgap $\sim$3.2 eV). FIG. 5(a) displays the fs-TA spectra taken at several representative probe delays for the two composite samples (lower panels) and two reference samples (upper panels). Bare Cu$_3$(BTC)$_2$ exhibits nearly indiscernible fs-TA signal as compared to bare TiO$_2$, echoing to their steady-state UV-Vis absorption spectra (see FIG. S2 in the supplementary materials). Markedly, the negative probe bleach (PB) signal of bare TiO$_2$ is completely reversed to the positive photoinduced absorption (PA) signal upon being interfaced with Cu$_3$(BTC)$_2$. As for bare TiO$_2$, the PB evolution probed in 450-650 nm, a red-shifted region relative to the 320-nm pump, tracks via stimulated emission the population of photoexcited electrons in its conduction band (CB) [13, 15]. As for the two hybrid systems, however, the PA evolution observed in the same region should not correspond to the excited-state absorption (i.e., upward probing) from the CB electrons on the TiO$_2$ side, but rather reflects that of the electrons transferred rapidly from the CB of TiO$_2$ to the interface states formed between TiO$_2$ and Cu$_3$(BTC)$_2$. There should occur a sort of efficient TiO$_2$-to-Cu$_3$(BTC)$_2$ interface electron-transfer process, given that the photoexcited component is dominantly TiO$_2$ instead of Cu$_3$(BTC)$_2$. The mechanistic picture is illustrated in FIG. 5(b).

 FIG. 5 (a) Representative fs-TA spectra (pump at 320 nm) for the two hybrid systems (lower panels) and for the two reference systems (upper panels). (b) Schematic illustration of the interface electro-transfer processes in the semiconductor-MOF hybrid system.

To gain more insights into the interface electron transfer and the subsequent relaxation processes in the hybrid systems, we turn our attention to the involved PA kinetics. FIG. 6 exhibits a set of representative fs-TA kinetics data taken at 560 nm. As for bare TiO$_2$, the negative PB signal builds up within the instrument response function (IRF) of $\sim$100 fs and then recovers with an average lifetime of $\sim$110 ps. The fast build-up ($<$100 fs) reflects the initial generation and cooling of hot excitons, while the following recovery ($\sim$110 ps) accounts for the subsequent exciton trapping and recombination [12, 17]. Also, the early-time build-up for the two hybrid systems occurs within the IRF, suggesting that the interface electron-transfer process therein is extremely fast (indicated by a grey, wiggly arrow in FIG. 5(b)). Strikingly, the subsequent relaxation, which should be linked to the event that the promptly accumulated interface-state electrons are redistributed into the electron acceptor states of Cu$_3$(BTC)$_2$ (indicated by a red, wiggly arrow in FIG. 5(b)), exhibits a pronounced difference: $\tau_1$=(182$\pm$16) ps (69.1%) and $\tau_2$=(1006$\pm$323) ps (30.9%) (or (769$\pm$292) ps on average) for TiO$_2$@Cu$_3$(BTC)$_2$, while $\tau_1$=10.1$\pm$1.1 ps (51.2%) and $\tau_2$=(157$\pm$7) ps (48.8%) (or (148$\pm$6) ps on average) for TiO$_2$/Cu$_3$(BTC)$_2$. A roughly 5-fold acceleration (in terms of average lifetime) was detected for TiO$_2$/Cu$_3$(BTC)$_2$ relative to TiO$_2$@Cu$_3$(BTC)$_2$. This evidences that the interface states formed between semiconductor and MOF can act as an effective relay for semiconductor-to-MOF electron transfer, whose efficiency is highly dependent on the relative location of the two components. In the current case, such a relay effect in the system with TiO$_2$ NPs being supported onto Cu$_3$(BTC)$_2$ (i.e., TiO$_2$/Cu$_3$(BTC)$_2$) was proven more efficient than in the system with TiO$_2$ NPs being incorporated inside Cu$_3$(BTC)$_2$ (i.e., TiO$_2$@Cu$_3$(BTC)$_2$), suggesting that the coupling between the interface states and the electron acceptor states of MOF is stronger in TiO$_2$/Cu$_3$(BTC)$_2$ than in TiO$_2$@Cu$_3$(BTC)$_2$.

 FIG. 6 Representative fs-TA kinetics (pump at 320 nm, probe at 560 nm) for the two hybrid systems as well as for the two reference systems.

Considering that the higher efficiency in interface states-mediated electron transfer would promote electron-hole separation (on the photoexcited semiconductor side) and electron redistribution (on the electron-accepting MOF side), we anticipated that TiO$_2$/Cu$_3$(BTC)$_2$ could achieve a better photocatalytic performance than TiO$_2$@Cu$_3$(BTC)$_2$. This expectation was verified by a set of photocatalytic CO$_2$ reduction tests, in which TiO$_2$/Cu$_3$(BTC)$_2$ was found to stand out among others (FIG. 7). Notably, the two composites were found to possess an identical mass ratio of TiO$_2$ NPs over Cu$_3$(BTC)$_2$ (i.e., $\sim$0.2) and a similar size distribution of TiO$_2$ NPs (FIG. 2), as discussed above. On the basis of these observations, it is safe to conclude that the difference in photocatalytic performance for the two composite samples should not simply be due to the slight differences in the loading distribution and density of TiO$_2$ NPs, but rather related to the different interface electron-transfer efficiencies for the two samples.

 FIG. 7 CO conversion rates for TiO$_2$, TiO$_2$@Cu$_3$(BTC)$_2$, and TiO$_2$/Cu$_3$(BTC)$_2$ in the recycling tests of photocatalytic CO$_2$ reduction. The product mole yields were calibrated against the 15-mg photoactive TiO$_2$. Signals for by-products were indiscernible.
Ⅳ. CONCLUSION

We have presented an ultrafast spectroscopy investigation addressing the subtle location effect in a prototypical semiconductor-MOF hybrid system with TiO$_2$ nanoparticles being incorporated inside or supported onto Cu$_3$(BTC)$_2$, denoted TiO$_2$@Cu$_3$(BTC)$_2$ and TiO$_2$/Cu$_3$(BTC)$_2$, respectively. By tracking in real time the interface electron dynamics in the hybrid system, we find that the interface states formed between TiO$_2$ and Cu$_3$(BTC)$_2$ can act as an effective relay for electron transfer, whose efficiency rests on the relative location of the two components. Such a subtle location effect is responsible for the performance difference in photocatalytic CO$_2$ reduction using the two hybrid systems. The mechanistic insights into the involved interface electron-transfer effect provide a helpful perspective for rational design of MOF-based hybrid systems for photoelectrochemical applications.

Supplementary materials: XPS spectra and steady-state UV-Vis absorption spectra of the samples are available in the supplementary materials.

Ⅴ. ACKNOWLEDGMENTS

This work was supported by the Ministry of Science and Technology of China (No.2016YFA0200602), the National Natural Science Foundation of China (No.21573211 and No.21633007), and the Fundamental Research Funds for the Central Universities of China (No.WK2340000063).

Reference
 [1] H. C. Zhou, J. R. Long, and O. M. Yaghi, Chem. Rev. 112 , 673 (2012). DOI:10.1021/cr300014x [2] P. Falcaro, R. Ricco, A. Yazdi, I. Imaz, S. Furukawa, D. Maspoch, R. Ameloot, J. D. Evans, and C. Doonan, Coord. Chem. Rev. 307 , 237 (2016). DOI:10.1016/j.ccr.2015.08.002 [3] Q. H. Yang, Q. Xu, and H. L. Jiang, Chem. Soc. Rev. 46 , 4774 (2017). DOI:10.1039/C6CS00724D [4] C. Wang, K. E. deKrafft, and W. B. Lin, J. Am. Chem. Soc. 134 , 7211 (2012). DOI:10.1021/ja300539p [5] L. Y. Chen, Y. Peng, H. Wang, Z. Z. Gu, and C. Y. Duan, Chem. Commun. 50 , 8651 (2014). DOI:10.1039/C4CC02818J [6] J. D. Xiao, Q. C. Shang, Y. J. Xiong, Q. Zhang, Y. Luo, S. H. Yu, and H. L. Jiang, Angew. Chem. Int. Ed. 55 , 9389 (2016). [7] R. Li, J. H. Hu, M. S. Deng, H. L. Wang, X. J. Wang, Y. L. Hu, H. L. Jiang, J. Jiang, Q. Zhang, Y. Xie, and Y. J. Xiong, Adv. Mater. 26 , 4783 (2014). DOI:10.1002/adma.v26.28 [8] S. B. Wang, and X. C. Wang, Appl. Catal. B 162 , 494 (2015). DOI:10.1016/j.apcatb.2014.07.026 [9] L. Shi, T. Wang, H. B. Zhang, K. Chang, and J. H. Ye, Adv. Funct. Mater. 25 , 5360 (2015). DOI:10.1002/adfm.201502253 [10] D. A. Giannakoudakis, N. A. Travlou, J. Secor, and T. J. Bandosz, Small 13 , 1601758 (2017). DOI:10.1002/smll.201601758 [11] J. Y. Lee, O. K. Farha, J. Roberts, K. A. Scheidt, S. B. T. Nguyen, and J. T. Hupp, Chem. Soc. Rev. 38 , 1450 (2009). DOI:10.1039/b807080f [12] D. A. Wheeler, and J. Z. Zhang, Adv. Mater. 25 , 2878 (2013). DOI:10.1002/adma.v25.21 [13] S. Bai, J. Jiang, Q. Zhang, and Y. J. Xiong, Chem. Soc. Rev. 44 , 2893 (2015). DOI:10.1039/C5CS00064E [14] B. Pattengale, S. Z. Yang, J. Ludwig, Z. Q. Huang, X. Y. Zhang, and J. E. Huang, J. Am. Chem. Soc. 138 , 8072 (2016). DOI:10.1021/jacs.6b04615 [15] Q. Zhang, and Y. Luo, High Power Laser Sci. Eng. 4 , e22 (2016). DOI:10.1017/hpl.2016.23 [16] Q. Y. Li, and T. Q. Lian, Nano Lett. 17 , 3152 (2017). DOI:10.1021/acs.nanolett.7b00587 [17] W. A. Tisdale, K. J. Williams, B. A. Timp, D. J. Norris, E. S. Aydil, and X. Y. Zhu, Science 328 , 1543 (2010). DOI:10.1126/science.1185509

a. 中国科学技术大学化学与材料科学学院，合肥微尺度物质科学国家研究中心，合肥 230026;
b. 中国科学技术大学化学物理系，量子信息与量子物理协同创新中心，合肥 230026;
c. 中国科学技术大学化学系，中国科学院软物质化学重点实验室，苏州纳米科技协同创新中心，合肥 230026