Tuning the excitonic and plasmonic properties of copper chalcogenide
Supporting Information for
Tuning the excitonic and plasmonic properties of copper chalcogenide nanocrystals Ilka Kriegel1&, Chengyang Jiang2&, Jessica Rodríguez-Fernández1, Richard D. Schaller3,4, Dmitri V. Talapin2,3*, Enrico da Como1* and Jochen Feldmann1
1
Photonics and Optoelectronics Group, Department of Physics and CeNS, Ludwig-Maximilians-Universität München, Munich, Germany 2 Department of Chemistry, University of Chicago, Chicago, IL 60637, USA 3 Argonne National Laboratory, Center for Nanoscale Materials, Argonne, IL 60439, USA 4 Department of Chemistry, Northwestern University, Evanston, IL 60208, USA
&
These authors contributed equally to this work. 1
*Corresponding author: [email protected], [email protected] Chemicals. All syntheses were carried out using standard Schlenk line techniques and nitrogen-filled gloveboxes. Copper(II) acetylacetonate (Cu(acac)2, 98%) and trioctylphosphine (TOP, 97%) were purchased from Strem. Ammonium diethyldithiocarbamate, oleic acid (OA, 90%), 1-dodecanethiol (DDT, 98%), copper(I) chloride (99.995%), selenium powder (99.99%), 1-octadecene (ODE, 90%), tellurium shot (99.999%), diisobutylaluminium hydride (1 M in toluene), ammonium sulfide (40-48 wt% in water), didodecyldimethylammonium bromide (DDAB, 98%), hexane (99%, anhydrous), toluene (99.8%, anhydrous), dimethylsulfoxide (DMSO, 99.9%, anhydrous), formamide (FA, 99.5%) and acetonitrile (99.8%, anhydrous) were purchased from Sigma-Aldrich. Oleylamine (OAm, 80-90% C-18 content) was purchased from Acros and arsenic sulfide (99.9%), from Alfa Aesar.
TEM images of Cu2-xS/Se after oxidation TEM images of Cu2-xS/Se NCs confirm that the NCs preserve their original shape and size during oxidation.
Figure S 1 TEM images of Cu2-xS NCs (left image) and Cu2-xSe NCs (right image) after oxidation
2
Size dependence of the LSP band
Figure S 2 Size dependence of the LSP band in the NIR for Cu2-xS NCs of sizes ranging from 4.5 to below 2 nm.
Reduction of Cu2-xTe NCs: addition of the reducing agent [Cu(CH3CN)4][PF6] We have performed additional experiments adding the reducing agent [Cu(CH3CN)4][PF6] (which is milder than DIBAH) to a colloidal solution of Cu2-xTe nanocrystals. Upon addition of [Cu(CH3CN)4][PF6] we have observed the suppression of the plasmon band. Moreover, the band can
Intensity (a.u.)
be retrieved by additional oxidation.
JCPDS #06-0661 30
40
50
2Θ (degree)
3
60
70
Figure S 3 Powder X-ray diffraction patterns measured from four different batches of Cu2-xTe NCs.
Absorbance (a.u.)
as-synthesized Cu2Te NCs Cu2Te NCs reduced with DIBAH
A
B
500 750 1000 1250 1500 1750
Wavelength (nm)
Figure S 4 Reduction of Cu2-xTe NCs with (a) DIBAH and (b) [Cu(CH3CN)4][PF6] followed by subsequent oxidation.
Effect of the solvent refractive index on the LSP resonances of Cu2-xS/Se (x>0) NCs: The sensitivity of the LSP band of non-stoichiometric Cu2-xS (x=0.03) and Cu2-xSe NCs (x=0.2) to the dielectric properties of the surrounding medium was evaluated by first drying 20 µL of a 5 mg/mL NC dispersion in a glovebox and subsequently redispersing the NCs in 2 mL of the respective anhydrous solvent (hexane, toluene or carbon disulfide, with n=1.38, 1.50 and 1.63, respectively). To prevent any further oxidation and hence an oxidation-induced change of the LSP resonance wavelength of the NCs, the NC dispersions were transferred in a glovebox into screwcap- and septum- sealed cuvettes for optical extinction measurements. The position of the LSP band in the copper chalcogenide was actively tuned by changing the dielectric environment of the NCs. In refractive index sensing experiments with Cu2-xSe (x>0) and Cu2-xS (x>0) NCs with well-defined NIR LSPs, both types of NCs were redispersed in organic solvents of different refractive indexes, namely hexane (n=1.38), toluene (n=1.50) and carbon disulfide (CS2, n=1.63). Note that due to the high sensitivity of our NCs to oxygen, the redispersion of the NCs in the different solvents was performed under well-controlled, air-free, conditions in a glovebox and using anhydrous solvents. The results summarized in Figure S demonstrate that both LSPs red-shift with increasing solvent refractive index in agreement with the behavior of LSP resonances in metal nanoparticles1 and with recent work from Alivisatos and co-workers.2 Figure S1c summarizes the LSP peak position of the same Cu2-xS and Cu2-xSe (x>0) NCs plotted versus the refractive index of the solvent clearly showing a red-shift in the maximum of the NIR plasmon band with increasing refractive index of the solvent. 4
a) Extinction (a.u.)
Cu2-xSe
Extinction (a.u.)
b)
CS2 Toluene Hexane Cu2-xS
Toluene 1.496 Hexane 1.375
0.6
c)
CS2 1.62774
0.8 1.0 Energy (eV)
1.2
Energy (eV)
1.0 Cu2-xSe
0.9 0.8 0.7 0.6
Cu2-xS 1.4 1.5 Refractive index
1.6
Figure S 5 (a) and (b) Extinction spectra of Cu2-xSe (x>0) as well as Cu2-xS (x>0) NCs in hexane, toluene, and CS2 with the solvent refractive indices 1.38, 1.50, and 1.63, respectively. (c) Red-shift of NIR LSP with increasing refractive index of the solvent.
Effect of ligands on the LSP in Cu2-xSe (x>0) NCs: Ligand exchange. To cap Cu2-xS (x>0) and Cu2-xSe (x>0) NCs with inorganic-organic hybrid ligands, a two-step procedure was carried out by applying a method developed by Kovalenko et al..3 In a typical experiment, Cu2-xS/Cu2-xSe (x>0) NCs dispersed in 2 mL hexane (~2.5 mg/mL) were mixed with a 1.5 mL DMSO solution of (NH4)3AsS3 (~1.3 mg/mL) under vigorous stirring. Once the NCs were transferred to the DMSO phase, the hexane phase was discarded and the DMSO phase was washed three times with hexane. Next, 3 mL acetonitrile were added to precipitate the NCs. The precipitate was redispersed in FA, yielding a colloidally stable NC dispersion. To transfer the NCs 5
back to toluene, a 2 mL 1 mM solution of DDAB in toluene was mixed with the previous NC dispersion in FA and the two-phase mixture was vigorously stirred for hours before the NCs completely transferred back to the organic phase. This toluene phase was then washed with water and filtered with a 0.2 µm PTFE membrane. The effect of ligands on the NIR plasmon band was investigated. The organic ligand shell was exchanged to inorganic-organic hybrid ligands. Note that here we compare the optical properties of the LSP for NCs in the same solvent but with a different ligand shell. To cap Cu2-xSe (x>0) NCs with inorganic-organic hybrid ligands, a two-step procedure was carried out. First the NCs were transferred to a DMSO solution by exchanging the ligands to charged AsS33-. The back transfer to toluene was carried out by the addition of a 1 mM solution of DDBA to the NCs dispersion. The effect on the LSP is shown in Figure S d. While the NCs do not change shape and size (Figure S a-c), the LSP in the NIR is blue shifting by more than 100 nm and decreasing in intensity. Upon the addition of the hybrid inorganic-organic ligands we introduce charges at the NCs surface. While expecting an increase of the dielectric constant going from organic to hybrid inorganic-organic ligands, entailing a red shift of the LSP band, we observe an intense blue shift instead. This result suggests that the change in the NCs surface charge from neutral (with the organic ligands) to negative (with the inorganic ones), may have a stronger influence on the LSP. By introducing the negative charge of the hybrid organic-inorganic ligand the positive carriers exhibiting the LSP are damped heavily leading to the observed shifts in the spectrum.
6
Figure S 6 TEM images of (A) oleylamine, (B) (NH4)+ - [AsS3]3-, and (C) [DDA]+ - [AsS3]3- capped Cu2-xSe (x>0) NCs. (D) vis-NIR extinction spectra of [DDA]+ - [AsS3]3- capped Cu2-xSe (x>0) NCs (solid curve) and oleylamine-capped Cu2-xSe (x>0) NCs (dashed curve), both dispersed in toluene.
Effect of close-packing on the LSP in Cu2-xS (x>0) NCs: Preparation of Cu2-xS (x=0.03) superlattices. Non-stoichiometric copper sulfide NCS (Cu2-xS, x=0.03) of 12 nm were synthesized as described elsewhere from the thermolysis of a Cu-oleate complex in a mixture of OAm and DDT.4 Sample washing was not performed in the glovebox and carried out by the addition of 10 mL of toluene and 10 mL of ethanol, followed by centrifugation. The resulting deep brown precipitate was afterwards redispersed in 10 mL of toluene and typically consisted of micron-sized superlattices consisting of close-packed (Cu2-xS, x=0.03) NCs. Superlattice disassembly. The careful addition of OAm ligands in excess yielded typically in a colloidal dispersion of disassembled, non-close packed Cu2-xS (x>0) NCs.5 Typically, 40 µL of OAm were added to 2 mL of the superlattices in toluene, followed by bath sonication and mild heating (ca. 40 °C) for several minutes. The effect of NC close packing on the LSP band in Cu2-xS (x>0) NCs of 12 nm size (Figure S b, right panel inset) has been investigated. The synthesis of Cu2-xS NCs, as described elsewhere,4 results in a deep brown precipitate typically consisting of close-packed superlattices consisting of 12 nm sized Cu2-xS (x>0) NCs (Figure S , left TEM image). The disassembly process was monitored via extinction 7
spectroscopy and triggered by the addition of an OAm excess. Figure S shows the evolution of the LSP during NC disassembly. The LSP blue-shifts and decreases in intensity upon superlattice disassembly as recently reported by us.5 LSP coupling is a well studied phenomenon in metallic nanoparticles. It occurs when the particles are close enough to interact with each other. This results in a red-shift of the LSP band with respect to the non-interacting particles. These results further proof on the tunability of the LSP band in the NIR and suggest interesting opportunities for plasmon coupling
Absorbance (a.u.)
phenomena and structures active at telecommunication wavelengths.
0.5 0.4 0.3 0.2 0.1 1200 1600 Wavelength (nm)
2000
Absorbance (a.u.)
Figure S 7 (a) Evolution of the optical properties in the NIR of 12 nm sized Cu2-xS NCs upon controlled disassembly (from the red to the blue curve).
600
800 Wavelength (nm)
1000
Figure S 8 Size-dependent excitonic absorption of Cu2S NCs of < 2 nm to 6 nm (from the black to the green curve). The fluorescence quantum yield Φi has been determined by optical measurements. It has been derived from the equation given below according to which the fluorescence quantum yield of a compound i is determined relative to that of a standard s with known Φ.
8
with fx(λex) = 1 – 10–Ax(λex) and Ax(λex) the absorbance. S describes the reference and i the sample. The absorbance of Cu2S NCs in toluene and the reference IR140 in ethanol was adjusted to be equal at the excitation wavelength (614 nm). Then, fluorescence spectra of both solutions were measured. The fluorescence quantum yield has been corrected for the different solvent refractive indices (n). Extinction IR140 Extinction Cu2S PL IR140 PL Cu2S
PL
Extinction
Excitation
300 350 400 450 500 550 600 650 700 750 800 850 900 950 1000
Wavelength (nm)
Figure S 9 Extinction spectra of Cu2S NCs in toluene (red line) together with the reference spectrum IR140 (black line; not normalized) and PL spectra (dotted lines, respectively), normalized to the maximum.
∆A/A
-0.02 -0.04 -0.06 750 1500 2250 3000 2 Pump fluence (µJ/cm )
Figure S 10 Linear dependence of the pump fluence for Cu2-xSe NCs
9
∆A/A maximum
1.030 1.025 1.020 1.015 1.010
0 75 150 225 300 375 450 525 600 Time (ps)
Figure S 11 Shift of the maximum of the non-linearity ∆A/A observed for Cu2-xSe NCs.
Two-temperature model The nonlinearities in the dynamics of plasmonic features can be explained by changes in the dielectric function of the investigated metallic structure. These are induced by elevated carrier and lattice temperatures TC and TL, respectively. To investigate the time evolution of the carrier and lattice temperatures the two temperature model (TTM) has been introduced for metal nanoparticles.6 The TTM is represented by two coupled differential equations:
Cc(Tc) is the carrier specific heat capacity and CL the specific heat capacity of the lattice. G is the electron-phonon coupling constant, and PA(t) is the excitation laser energy deposited per unit volume of nanocrystals per unit time. The precise quantitative value of the differential absorption depends on details of the dielectric function and its calculation exceeds the scope of this work. Nevertheless, in a recent work, published during the review process of this paper the TTM has successfully been applied to Cu2-xSe NCs.7 The time dependence of the temperatures TC and TL have been identified. These were then used to fit the differential spectra and determine the carrierphonon coupling factor G for Cu2-xSe (x=0.15) NCs, which is G = 1.2·1016
. Based on
these results we have tried to implement the TTM to our Cu2-xS and Cu2-xTe NCs. The equations given above were solved numerically. 10
∆A normalized
TC+TL
1
Cu2-xS Cu2-xSe Cu2-xTe
0.1
0.01 0
0
5
1
2
3
Time (ps)
10 Time (ps)
15
4
5
20
Figure S 32 Normalized differential absorption spectra of Cu2-xS/Se/Te NCs black, red and green curves, respectively). Inset: the time dependent temperatures calculated according to the TTM for Cu2-xS/Se/Te NCs.
Figure S 3 shows the normalized differential absorption transient of Cu2-xS/Se/Te NCs (black, red and green curves, respectively). The inset displays the time dependent temperatures calculated according to the TTM. PA(t) was kept constant for the simulation of the different copper chalcogenides. For Cu2-xSe NCs the input values for the TTM were chosen according to the data given in a recent publication by Scotognella et al. in Nano Letters. The carrier heat capacity has been estimated to be Cc(Tc) = γTc=rγAuTc with γAu = 63
being the heat
capacity constant of the free carriers in gold and r being the ratio between charge carrier density of the copper chalcogenide (which is estimated from the plasma frequency) and gold carrier density (5.9x1022 cm-3). We applied similar estimations for our new materials Cu2-xS and Cu2-xTe NCs and determined γ to be 1.5 xS/Se/Te
is 2.8·106
, 2.72·106
, and 4·106
, and 5.3
, respectively. CL for Cu2-
. The carrier-phonon coupling factors
for Cu2-xS and Cu2-xTe NCs were obtained by a comparison of the experimental data with the model. The best fit values G for Cu2-xS and Cu2-xTe NCs were 0.8·1016
and 1.4·1016
.
We do not include these simulations to the main text. The constant PA(t) is an assumption which has to be verified in further experiments. Experiments of this kind are kept for future work.
11
References (1) (2)
Mulvaney, P. Langmuir 1996, 12, 788-800. Luther, J. M.; Jain, P. K.; Ewers, T.; Alivisatos, A. P. Nature Materials 2011, 10, 361-
366. (3) Kovalenko, M. V.; Bodnarchuk, M. I.; Talapin, D. V. Journal of the American Chemical Society 2010, 132, 15124-15126. (4) Sang-Hyun Choi; Kwangjin An; Eung-Gyu Kim; Jung Ho Yu; Jeong Hyun Kim; Taeghwan Hyeon Advanced Functional Materials 2009, 19, 1645-1649. (5) Kriegel, I.; Rodríguez-Fernández, J.; Como, E. D.; Lutich, A. A.; Szeifert, J. M.; Feldmann, J. Chemistry of Materials 2011, 23, 1830-1834. (6) Anisimov, S. I.; Kapeliovich, B. L.; Perel'man, T. L. Soviet Physics - JETP 1974, 39, 375377. (7) Scotognella, F.; Della Valle, G.; Srimath Kandada, A. R.; Dorfs, D.; Zavelani-Rossi, M.; Conforti, M.; Miszta, K.; Comin, A.; Korobchevskaya, K.; Lanzani, G.; Manna, L.; Tassone, F. Nano Letters 2011, 11, 4711-4717.
12
Tuning the excitonic and plasmonic properties of copper chalcogenide nanocrystals Ilka Kriegel1&, Chengyang Jiang2&, Jessica Rodríguez-Fernández1, Richard D. Schaller3,4, Dmitri V. Talapin2,3*, Enrico da Como1* and Jochen Feldmann1
1
Photonics and Optoelectronics Group, Department of Physics and CeNS, Ludwig-Maximilians-Universität München, Munich, Germany 2 Department of Chemistry, University of Chicago, Chicago, IL 60637, USA 3 Argonne National Laboratory, Center for Nanoscale Materials, Argonne, IL 60439, USA 4 Department of Chemistry, Northwestern University, Evanston, IL 60208, USA
&
These authors contributed equally to this work. 1
*Corresponding author: [email protected], [email protected] Chemicals. All syntheses were carried out using standard Schlenk line techniques and nitrogen-filled gloveboxes. Copper(II) acetylacetonate (Cu(acac)2, 98%) and trioctylphosphine (TOP, 97%) were purchased from Strem. Ammonium diethyldithiocarbamate, oleic acid (OA, 90%), 1-dodecanethiol (DDT, 98%), copper(I) chloride (99.995%), selenium powder (99.99%), 1-octadecene (ODE, 90%), tellurium shot (99.999%), diisobutylaluminium hydride (1 M in toluene), ammonium sulfide (40-48 wt% in water), didodecyldimethylammonium bromide (DDAB, 98%), hexane (99%, anhydrous), toluene (99.8%, anhydrous), dimethylsulfoxide (DMSO, 99.9%, anhydrous), formamide (FA, 99.5%) and acetonitrile (99.8%, anhydrous) were purchased from Sigma-Aldrich. Oleylamine (OAm, 80-90% C-18 content) was purchased from Acros and arsenic sulfide (99.9%), from Alfa Aesar.
TEM images of Cu2-xS/Se after oxidation TEM images of Cu2-xS/Se NCs confirm that the NCs preserve their original shape and size during oxidation.
Figure S 1 TEM images of Cu2-xS NCs (left image) and Cu2-xSe NCs (right image) after oxidation
2
Size dependence of the LSP band
Figure S 2 Size dependence of the LSP band in the NIR for Cu2-xS NCs of sizes ranging from 4.5 to below 2 nm.
Reduction of Cu2-xTe NCs: addition of the reducing agent [Cu(CH3CN)4][PF6] We have performed additional experiments adding the reducing agent [Cu(CH3CN)4][PF6] (which is milder than DIBAH) to a colloidal solution of Cu2-xTe nanocrystals. Upon addition of [Cu(CH3CN)4][PF6] we have observed the suppression of the plasmon band. Moreover, the band can
Intensity (a.u.)
be retrieved by additional oxidation.
JCPDS #06-0661 30
40
50
2Θ (degree)
3
60
70
Figure S 3 Powder X-ray diffraction patterns measured from four different batches of Cu2-xTe NCs.
Absorbance (a.u.)
as-synthesized Cu2Te NCs Cu2Te NCs reduced with DIBAH
A
B
500 750 1000 1250 1500 1750
Wavelength (nm)
Figure S 4 Reduction of Cu2-xTe NCs with (a) DIBAH and (b) [Cu(CH3CN)4][PF6] followed by subsequent oxidation.
Effect of the solvent refractive index on the LSP resonances of Cu2-xS/Se (x>0) NCs: The sensitivity of the LSP band of non-stoichiometric Cu2-xS (x=0.03) and Cu2-xSe NCs (x=0.2) to the dielectric properties of the surrounding medium was evaluated by first drying 20 µL of a 5 mg/mL NC dispersion in a glovebox and subsequently redispersing the NCs in 2 mL of the respective anhydrous solvent (hexane, toluene or carbon disulfide, with n=1.38, 1.50 and 1.63, respectively). To prevent any further oxidation and hence an oxidation-induced change of the LSP resonance wavelength of the NCs, the NC dispersions were transferred in a glovebox into screwcap- and septum- sealed cuvettes for optical extinction measurements. The position of the LSP band in the copper chalcogenide was actively tuned by changing the dielectric environment of the NCs. In refractive index sensing experiments with Cu2-xSe (x>0) and Cu2-xS (x>0) NCs with well-defined NIR LSPs, both types of NCs were redispersed in organic solvents of different refractive indexes, namely hexane (n=1.38), toluene (n=1.50) and carbon disulfide (CS2, n=1.63). Note that due to the high sensitivity of our NCs to oxygen, the redispersion of the NCs in the different solvents was performed under well-controlled, air-free, conditions in a glovebox and using anhydrous solvents. The results summarized in Figure S demonstrate that both LSPs red-shift with increasing solvent refractive index in agreement with the behavior of LSP resonances in metal nanoparticles1 and with recent work from Alivisatos and co-workers.2 Figure S1c summarizes the LSP peak position of the same Cu2-xS and Cu2-xSe (x>0) NCs plotted versus the refractive index of the solvent clearly showing a red-shift in the maximum of the NIR plasmon band with increasing refractive index of the solvent. 4
a) Extinction (a.u.)
Cu2-xSe
Extinction (a.u.)
b)
CS2 Toluene Hexane Cu2-xS
Toluene 1.496 Hexane 1.375
0.6
c)
CS2 1.62774
0.8 1.0 Energy (eV)
1.2
Energy (eV)
1.0 Cu2-xSe
0.9 0.8 0.7 0.6
Cu2-xS 1.4 1.5 Refractive index
1.6
Figure S 5 (a) and (b) Extinction spectra of Cu2-xSe (x>0) as well as Cu2-xS (x>0) NCs in hexane, toluene, and CS2 with the solvent refractive indices 1.38, 1.50, and 1.63, respectively. (c) Red-shift of NIR LSP with increasing refractive index of the solvent.
Effect of ligands on the LSP in Cu2-xSe (x>0) NCs: Ligand exchange. To cap Cu2-xS (x>0) and Cu2-xSe (x>0) NCs with inorganic-organic hybrid ligands, a two-step procedure was carried out by applying a method developed by Kovalenko et al..3 In a typical experiment, Cu2-xS/Cu2-xSe (x>0) NCs dispersed in 2 mL hexane (~2.5 mg/mL) were mixed with a 1.5 mL DMSO solution of (NH4)3AsS3 (~1.3 mg/mL) under vigorous stirring. Once the NCs were transferred to the DMSO phase, the hexane phase was discarded and the DMSO phase was washed three times with hexane. Next, 3 mL acetonitrile were added to precipitate the NCs. The precipitate was redispersed in FA, yielding a colloidally stable NC dispersion. To transfer the NCs 5
back to toluene, a 2 mL 1 mM solution of DDAB in toluene was mixed with the previous NC dispersion in FA and the two-phase mixture was vigorously stirred for hours before the NCs completely transferred back to the organic phase. This toluene phase was then washed with water and filtered with a 0.2 µm PTFE membrane. The effect of ligands on the NIR plasmon band was investigated. The organic ligand shell was exchanged to inorganic-organic hybrid ligands. Note that here we compare the optical properties of the LSP for NCs in the same solvent but with a different ligand shell. To cap Cu2-xSe (x>0) NCs with inorganic-organic hybrid ligands, a two-step procedure was carried out. First the NCs were transferred to a DMSO solution by exchanging the ligands to charged AsS33-. The back transfer to toluene was carried out by the addition of a 1 mM solution of DDBA to the NCs dispersion. The effect on the LSP is shown in Figure S d. While the NCs do not change shape and size (Figure S a-c), the LSP in the NIR is blue shifting by more than 100 nm and decreasing in intensity. Upon the addition of the hybrid inorganic-organic ligands we introduce charges at the NCs surface. While expecting an increase of the dielectric constant going from organic to hybrid inorganic-organic ligands, entailing a red shift of the LSP band, we observe an intense blue shift instead. This result suggests that the change in the NCs surface charge from neutral (with the organic ligands) to negative (with the inorganic ones), may have a stronger influence on the LSP. By introducing the negative charge of the hybrid organic-inorganic ligand the positive carriers exhibiting the LSP are damped heavily leading to the observed shifts in the spectrum.
6
Figure S 6 TEM images of (A) oleylamine, (B) (NH4)+ - [AsS3]3-, and (C) [DDA]+ - [AsS3]3- capped Cu2-xSe (x>0) NCs. (D) vis-NIR extinction spectra of [DDA]+ - [AsS3]3- capped Cu2-xSe (x>0) NCs (solid curve) and oleylamine-capped Cu2-xSe (x>0) NCs (dashed curve), both dispersed in toluene.
Effect of close-packing on the LSP in Cu2-xS (x>0) NCs: Preparation of Cu2-xS (x=0.03) superlattices. Non-stoichiometric copper sulfide NCS (Cu2-xS, x=0.03) of 12 nm were synthesized as described elsewhere from the thermolysis of a Cu-oleate complex in a mixture of OAm and DDT.4 Sample washing was not performed in the glovebox and carried out by the addition of 10 mL of toluene and 10 mL of ethanol, followed by centrifugation. The resulting deep brown precipitate was afterwards redispersed in 10 mL of toluene and typically consisted of micron-sized superlattices consisting of close-packed (Cu2-xS, x=0.03) NCs. Superlattice disassembly. The careful addition of OAm ligands in excess yielded typically in a colloidal dispersion of disassembled, non-close packed Cu2-xS (x>0) NCs.5 Typically, 40 µL of OAm were added to 2 mL of the superlattices in toluene, followed by bath sonication and mild heating (ca. 40 °C) for several minutes. The effect of NC close packing on the LSP band in Cu2-xS (x>0) NCs of 12 nm size (Figure S b, right panel inset) has been investigated. The synthesis of Cu2-xS NCs, as described elsewhere,4 results in a deep brown precipitate typically consisting of close-packed superlattices consisting of 12 nm sized Cu2-xS (x>0) NCs (Figure S , left TEM image). The disassembly process was monitored via extinction 7
spectroscopy and triggered by the addition of an OAm excess. Figure S shows the evolution of the LSP during NC disassembly. The LSP blue-shifts and decreases in intensity upon superlattice disassembly as recently reported by us.5 LSP coupling is a well studied phenomenon in metallic nanoparticles. It occurs when the particles are close enough to interact with each other. This results in a red-shift of the LSP band with respect to the non-interacting particles. These results further proof on the tunability of the LSP band in the NIR and suggest interesting opportunities for plasmon coupling
Absorbance (a.u.)
phenomena and structures active at telecommunication wavelengths.
0.5 0.4 0.3 0.2 0.1 1200 1600 Wavelength (nm)
2000
Absorbance (a.u.)
Figure S 7 (a) Evolution of the optical properties in the NIR of 12 nm sized Cu2-xS NCs upon controlled disassembly (from the red to the blue curve).
600
800 Wavelength (nm)
1000
Figure S 8 Size-dependent excitonic absorption of Cu2S NCs of < 2 nm to 6 nm (from the black to the green curve). The fluorescence quantum yield Φi has been determined by optical measurements. It has been derived from the equation given below according to which the fluorescence quantum yield of a compound i is determined relative to that of a standard s with known Φ.
8
with fx(λex) = 1 – 10–Ax(λex) and Ax(λex) the absorbance. S describes the reference and i the sample. The absorbance of Cu2S NCs in toluene and the reference IR140 in ethanol was adjusted to be equal at the excitation wavelength (614 nm). Then, fluorescence spectra of both solutions were measured. The fluorescence quantum yield has been corrected for the different solvent refractive indices (n). Extinction IR140 Extinction Cu2S PL IR140 PL Cu2S
PL
Extinction
Excitation
300 350 400 450 500 550 600 650 700 750 800 850 900 950 1000
Wavelength (nm)
Figure S 9 Extinction spectra of Cu2S NCs in toluene (red line) together with the reference spectrum IR140 (black line; not normalized) and PL spectra (dotted lines, respectively), normalized to the maximum.
∆A/A
-0.02 -0.04 -0.06 750 1500 2250 3000 2 Pump fluence (µJ/cm )
Figure S 10 Linear dependence of the pump fluence for Cu2-xSe NCs
9
∆A/A maximum
1.030 1.025 1.020 1.015 1.010
0 75 150 225 300 375 450 525 600 Time (ps)
Figure S 11 Shift of the maximum of the non-linearity ∆A/A observed for Cu2-xSe NCs.
Two-temperature model The nonlinearities in the dynamics of plasmonic features can be explained by changes in the dielectric function of the investigated metallic structure. These are induced by elevated carrier and lattice temperatures TC and TL, respectively. To investigate the time evolution of the carrier and lattice temperatures the two temperature model (TTM) has been introduced for metal nanoparticles.6 The TTM is represented by two coupled differential equations:
Cc(Tc) is the carrier specific heat capacity and CL the specific heat capacity of the lattice. G is the electron-phonon coupling constant, and PA(t) is the excitation laser energy deposited per unit volume of nanocrystals per unit time. The precise quantitative value of the differential absorption depends on details of the dielectric function and its calculation exceeds the scope of this work. Nevertheless, in a recent work, published during the review process of this paper the TTM has successfully been applied to Cu2-xSe NCs.7 The time dependence of the temperatures TC and TL have been identified. These were then used to fit the differential spectra and determine the carrierphonon coupling factor G for Cu2-xSe (x=0.15) NCs, which is G = 1.2·1016
. Based on
these results we have tried to implement the TTM to our Cu2-xS and Cu2-xTe NCs. The equations given above were solved numerically. 10
∆A normalized
TC+TL
1
Cu2-xS Cu2-xSe Cu2-xTe
0.1
0.01 0
0
5
1
2
3
Time (ps)
10 Time (ps)
15
4
5
20
Figure S 32 Normalized differential absorption spectra of Cu2-xS/Se/Te NCs black, red and green curves, respectively). Inset: the time dependent temperatures calculated according to the TTM for Cu2-xS/Se/Te NCs.
Figure S 3 shows the normalized differential absorption transient of Cu2-xS/Se/Te NCs (black, red and green curves, respectively). The inset displays the time dependent temperatures calculated according to the TTM. PA(t) was kept constant for the simulation of the different copper chalcogenides. For Cu2-xSe NCs the input values for the TTM were chosen according to the data given in a recent publication by Scotognella et al. in Nano Letters. The carrier heat capacity has been estimated to be Cc(Tc) = γTc=rγAuTc with γAu = 63
being the heat
capacity constant of the free carriers in gold and r being the ratio between charge carrier density of the copper chalcogenide (which is estimated from the plasma frequency) and gold carrier density (5.9x1022 cm-3). We applied similar estimations for our new materials Cu2-xS and Cu2-xTe NCs and determined γ to be 1.5 xS/Se/Te
is 2.8·106
, 2.72·106
, and 4·106
, and 5.3
, respectively. CL for Cu2-
. The carrier-phonon coupling factors
for Cu2-xS and Cu2-xTe NCs were obtained by a comparison of the experimental data with the model. The best fit values G for Cu2-xS and Cu2-xTe NCs were 0.8·1016
and 1.4·1016
.
We do not include these simulations to the main text. The constant PA(t) is an assumption which has to be verified in further experiments. Experiments of this kind are kept for future work.
11
References (1) (2)
Mulvaney, P. Langmuir 1996, 12, 788-800. Luther, J. M.; Jain, P. K.; Ewers, T.; Alivisatos, A. P. Nature Materials 2011, 10, 361-
366. (3) Kovalenko, M. V.; Bodnarchuk, M. I.; Talapin, D. V. Journal of the American Chemical Society 2010, 132, 15124-15126. (4) Sang-Hyun Choi; Kwangjin An; Eung-Gyu Kim; Jung Ho Yu; Jeong Hyun Kim; Taeghwan Hyeon Advanced Functional Materials 2009, 19, 1645-1649. (5) Kriegel, I.; Rodríguez-Fernández, J.; Como, E. D.; Lutich, A. A.; Szeifert, J. M.; Feldmann, J. Chemistry of Materials 2011, 23, 1830-1834. (6) Anisimov, S. I.; Kapeliovich, B. L.; Perel'man, T. L. Soviet Physics - JETP 1974, 39, 375377. (7) Scotognella, F.; Della Valle, G.; Srimath Kandada, A. R.; Dorfs, D.; Zavelani-Rossi, M.; Conforti, M.; Miszta, K.; Comin, A.; Korobchevskaya, K.; Lanzani, G.; Manna, L.; Tassone, F. Nano Letters 2011, 11, 4711-4717.
12
