(6,5) Single-Wall Carbon Nanotubes
Supporting Information for: 13 nm Exciton Size in (6,5) Single-Wall Carbon Nanotubes Christoph Mann† and Tobias Hertel∗,†,‡ †Institute of Physical and Theoretical Chemistry, Julius-Maximilian University Würzburg, Germany ‡Röntgen Research Center for Complex Material Systems, Julius-Maximilian University Würzburg, Germany E-mail: [email protected] Phone: +49 931 3188969
1
Experimental Methods Sample Preparation Aqueous nanotube suspensions were prepared from commercial CoMoCAT SG65 material (SouthWest NanoTechnologies Inc.) by density-gradient ultracentrifugation (DGU) with sodium deoxycholate (DOC) and sodium dodecyl sulphate (SDS) mixtures as dispersing agents. 1 This process yields suspensions strongly enriched in the semiconducting (6,5) sSWNT species. Subsequently, suspensions were dialyzed with 50 kDa mass cuttoff against aqueous 1.5% wt. sodium cholate (SC) solutions. Five dialysis cycles against 50 ml of fresh SC solutions were carried out with dialysis times increasing from 2 to 12 hours to ensure complete removal of iodixanol density gradient medium as well as of the dispersion additives used in density gradient runs. The resulting suspensions exhibit no iodixanol absorption signature at 245 nm and the first subband exciton absorption at 981 nm indicates complete replaced of the initial surfactant mixture by SC with no noticeable aggregation. For pump-probe investigations we fabricated two types of s-SWNT samples: thin films made of individualized SWNTs and films made of intentionally aggregated SWNTs. The use of thin films for pump-probe studies presents several advantages over the use of suspensions such as better control over excitation densities in the probe volume and the option to remove residual water from films by placement in high-vacuum. s-SWNT suspensions for films with aggregated SWNTs are obtained by addition of MgCl2 solution to the SWNT-SC suspensions. The divalent Mg2+ ions are intended to slightly screen the stabilizing Coulomb repulsion between surfactant-coated SWNTs in the suspensions, but only to a point. The ion concentration of 3 mMol was thus maintained sufficiently low at so as to preserve colloidal stability until hot gelatine solution is introduced in the final film preparation step. Thin s-SWNT-gelatine films were then prepared by mixing 30 µL of a suspension with individualized SWNTs, or of the same suspension containing Mg2+ ions, with 20 µL of a
2
60C hot, 15 wt.% gelatine solution also containing 15 wt.% of SDS. The addition of SDS at this point as well as the prior removal of DOC were found to be essential to help reduce fracturing of the drop-cast gelatine films as well as to prevent clouding during film production. Both, clouding and fracturing are detrimental to pump-probe spectroscopy due to signal contamination by scattered light. This process also ensures that aggregation of SWNTs from the suspensions containing Mg2+ ions 2 occurs only briefly before gelation sets in near room temperature. This prevented the formation of larger floc or an overly heterogeneous distribution of SWNT aggregates in the thin films while at the same time it minimized the risk of inadvertently doping s-SWNTs by exposure to excessive salt ion concentrations. Again, floc would increase the amount of unwanted scattered light. Immediately after deposition the cast gel films contain about 6.0 wt% gelatine, 0.9 wt% SC and 0.8 wt% SDS. Films are eventually allowed to dry for three hours in air before being transferred to an excicator for another 12 hours of drying over dehydrated silica gel.
Optical Setup For pump-probe spectroscopy films are transferred into a high vacuum chamber equipped with a closed cycle cryostat. Experimentation in vacuum helped to prevent water from entering films, which – in combination with atmospheric oxygen – has been suspected of inducing p-doping of s-SWNTs. 3 Degenerate pump-probe spectra were obtained using the output of an optical parametric amplifier (OPA9450, Coherent Inc.) driven at 250 kHz by a regenerative amplifier (RegA9050, Coherent Inc.). The whitelight continuum for probe impulses was generated by focusing 30% of the RegA output into a sapphire crystal. The attenuated and spatially filtered pump beam was focused onto the sample by a 250 mm lens to produce a gaussian excitation profile of (80 ± 2) µm in diameter. The probe beam was focused more tightly into a (40 ± 2) µm diameter spot. 4 3
Coherent artefacts due to probe impulse interference with stray light were minimized by choosing perpendicular polarization of pump and probe impulses unless noted otherwise. In addition, scattered pump light was rejected using a perpendicular oriented polarizer in the transmitted probe beam. Typical pump fluences at the sample were ≈ 100 nJ · cm−2 corresponding to a photon fluence of about ≈ 0.5 · 1012 cm−2 . Data was acquired at 500 Hz readout rate with a spectrograph (Shamrock 303i, Andor Technology PLC) equipped with a CCD camera (Newton DU920P BR-DD, Andor Technology PLC) using 150 line/mm and 600 line/mm gratings. All transient spectra were corrected for the differential transmission signal at negative pump-probe delays, i.e. at -30 ps or -50 ps. The latter account for transient 8.50 cm = 245 pt
spectral changes induced by long-lived states, i.e. with lifetimes in excess of 4µs. 4 All experiments reported here were carried out at room temperature.
Sample Characterization
Figure S1: PLE map of the starting material used for thin film fabrication. A PLE spectrum of the starting (6,5) enriched CoMoCat suspension used for the preparation of (6,5) enriched SWNT films in this study is reproduced in Figure S1. The main emission features seen in the figure can be attributed to the (8,3), (6,5) and (7,5) SWNT species. Fainter and some extremely faint emission features can be attributed to the (6,4), (9,1), (7,3), (9,4), (7,6), (8,4) and (9,2) species. A quantitative determination of abundances 4
8.50 cm = 245 pt
from PLE spectra, however, is challenging because it would require for PL quantum yields (PL-QY) of all species to be known with sufficient certainty. (6,4) (9,1)
(8,3) & (6,5)
(7,5)
3
0
1
2
4
5 6
Figure S2: Absorption spectrum of the starting (6,5) SWNT-enriched suspension along with the results of a multi-peak fit. The band designated as "0" is a phonon sideband of the (6,5) SWNT. Relative abundances of different SWNT species can nevertheless be estimated from absorption spectra using the information on transition energies from the PLE map as input for a multi peak fitting routine. The results of such a multi peak fit in the NIR range of the spectrum for the starting suspension are reproduced in Figure S2. This allows to estimate the abundances of the (6,4), (9,1), (7,5) and of the combined (8,3) and (6,5) absorption features (due to overlap). The estimated abundances then become 3.5% (6,4), 2.5% (9,1), 3% (7,5) and 91% for combined (8,3) and (6,5). If we use the PLE peak intensities of the (8,3) and (6,5) features we can estimate that (8,3) and (6,5) abundances account for 74% and 17% of all SWNTs respectively. In absorption spectra the first subband excitons of (8,3) and (6,5) SWNTs thus evidently show some overlap. The significance of this for the transient absorption signal can be estimated by accounting for the resonant excitation of the (6,5) SWNT at the pump wavelength 5
of 987 nm and correspondingly weak excitation of the (8,3) subband exciton at 960 nm. In combination with the above estimate of relative abundances this leads us to conclude that the transient signal contribution of the first (8,3) subband exciton in our experiments is below about 5% of that of the first (6,5) subband exciton.
References (1) Arnold, M. S.; Green, A. A.; Hulvat, J. F.; Stupp, S. I.; Hersam, M. C. Sorting Carbon Nanotubes by Electronic Structure Using Density Differentiation. Nature Nanotechnology 2006, 1, 60–65. (2) Schöppler, F.; Rühl, N.; Hertel, T. Photoluminescence Microscopy and Spectroscopy of Individualized and Aggregated Single-Wall Carbon Nanotubes. Chemical Physics 2013, 413, 112–115. (3) Aguirre, C. M.; Levesque, P. L.; Paillet, M.; Lapointe, F.; St-Antoine, B. C.; Desjardins, P.; Martel, R. The Role of the Oxygen/Water Redox Couple in Suppressing Electron Conduction in Field-Effect Transistors. Advanced Materials 2009, 21, 3087. (4) Schilling, D.; Mann, C.; Kunkel, P.; Schöppler, F.; Hertel, T. Ultrafast Spectral Exciton Diffusion in Single-Wall Carbon Nanotubes Studied by Time-Resolved Hole Burning. Journal of Physical Chemistry C 2015, 119, 24116–24123.
6
1
Experimental Methods Sample Preparation Aqueous nanotube suspensions were prepared from commercial CoMoCAT SG65 material (SouthWest NanoTechnologies Inc.) by density-gradient ultracentrifugation (DGU) with sodium deoxycholate (DOC) and sodium dodecyl sulphate (SDS) mixtures as dispersing agents. 1 This process yields suspensions strongly enriched in the semiconducting (6,5) sSWNT species. Subsequently, suspensions were dialyzed with 50 kDa mass cuttoff against aqueous 1.5% wt. sodium cholate (SC) solutions. Five dialysis cycles against 50 ml of fresh SC solutions were carried out with dialysis times increasing from 2 to 12 hours to ensure complete removal of iodixanol density gradient medium as well as of the dispersion additives used in density gradient runs. The resulting suspensions exhibit no iodixanol absorption signature at 245 nm and the first subband exciton absorption at 981 nm indicates complete replaced of the initial surfactant mixture by SC with no noticeable aggregation. For pump-probe investigations we fabricated two types of s-SWNT samples: thin films made of individualized SWNTs and films made of intentionally aggregated SWNTs. The use of thin films for pump-probe studies presents several advantages over the use of suspensions such as better control over excitation densities in the probe volume and the option to remove residual water from films by placement in high-vacuum. s-SWNT suspensions for films with aggregated SWNTs are obtained by addition of MgCl2 solution to the SWNT-SC suspensions. The divalent Mg2+ ions are intended to slightly screen the stabilizing Coulomb repulsion between surfactant-coated SWNTs in the suspensions, but only to a point. The ion concentration of 3 mMol was thus maintained sufficiently low at so as to preserve colloidal stability until hot gelatine solution is introduced in the final film preparation step. Thin s-SWNT-gelatine films were then prepared by mixing 30 µL of a suspension with individualized SWNTs, or of the same suspension containing Mg2+ ions, with 20 µL of a
2
60C hot, 15 wt.% gelatine solution also containing 15 wt.% of SDS. The addition of SDS at this point as well as the prior removal of DOC were found to be essential to help reduce fracturing of the drop-cast gelatine films as well as to prevent clouding during film production. Both, clouding and fracturing are detrimental to pump-probe spectroscopy due to signal contamination by scattered light. This process also ensures that aggregation of SWNTs from the suspensions containing Mg2+ ions 2 occurs only briefly before gelation sets in near room temperature. This prevented the formation of larger floc or an overly heterogeneous distribution of SWNT aggregates in the thin films while at the same time it minimized the risk of inadvertently doping s-SWNTs by exposure to excessive salt ion concentrations. Again, floc would increase the amount of unwanted scattered light. Immediately after deposition the cast gel films contain about 6.0 wt% gelatine, 0.9 wt% SC and 0.8 wt% SDS. Films are eventually allowed to dry for three hours in air before being transferred to an excicator for another 12 hours of drying over dehydrated silica gel.
Optical Setup For pump-probe spectroscopy films are transferred into a high vacuum chamber equipped with a closed cycle cryostat. Experimentation in vacuum helped to prevent water from entering films, which – in combination with atmospheric oxygen – has been suspected of inducing p-doping of s-SWNTs. 3 Degenerate pump-probe spectra were obtained using the output of an optical parametric amplifier (OPA9450, Coherent Inc.) driven at 250 kHz by a regenerative amplifier (RegA9050, Coherent Inc.). The whitelight continuum for probe impulses was generated by focusing 30% of the RegA output into a sapphire crystal. The attenuated and spatially filtered pump beam was focused onto the sample by a 250 mm lens to produce a gaussian excitation profile of (80 ± 2) µm in diameter. The probe beam was focused more tightly into a (40 ± 2) µm diameter spot. 4 3
Coherent artefacts due to probe impulse interference with stray light were minimized by choosing perpendicular polarization of pump and probe impulses unless noted otherwise. In addition, scattered pump light was rejected using a perpendicular oriented polarizer in the transmitted probe beam. Typical pump fluences at the sample were ≈ 100 nJ · cm−2 corresponding to a photon fluence of about ≈ 0.5 · 1012 cm−2 . Data was acquired at 500 Hz readout rate with a spectrograph (Shamrock 303i, Andor Technology PLC) equipped with a CCD camera (Newton DU920P BR-DD, Andor Technology PLC) using 150 line/mm and 600 line/mm gratings. All transient spectra were corrected for the differential transmission signal at negative pump-probe delays, i.e. at -30 ps or -50 ps. The latter account for transient 8.50 cm = 245 pt
spectral changes induced by long-lived states, i.e. with lifetimes in excess of 4µs. 4 All experiments reported here were carried out at room temperature.
Sample Characterization
Figure S1: PLE map of the starting material used for thin film fabrication. A PLE spectrum of the starting (6,5) enriched CoMoCat suspension used for the preparation of (6,5) enriched SWNT films in this study is reproduced in Figure S1. The main emission features seen in the figure can be attributed to the (8,3), (6,5) and (7,5) SWNT species. Fainter and some extremely faint emission features can be attributed to the (6,4), (9,1), (7,3), (9,4), (7,6), (8,4) and (9,2) species. A quantitative determination of abundances 4
8.50 cm = 245 pt
from PLE spectra, however, is challenging because it would require for PL quantum yields (PL-QY) of all species to be known with sufficient certainty. (6,4) (9,1)
(8,3) & (6,5)
(7,5)
3
0
1
2
4
5 6
Figure S2: Absorption spectrum of the starting (6,5) SWNT-enriched suspension along with the results of a multi-peak fit. The band designated as "0" is a phonon sideband of the (6,5) SWNT. Relative abundances of different SWNT species can nevertheless be estimated from absorption spectra using the information on transition energies from the PLE map as input for a multi peak fitting routine. The results of such a multi peak fit in the NIR range of the spectrum for the starting suspension are reproduced in Figure S2. This allows to estimate the abundances of the (6,4), (9,1), (7,5) and of the combined (8,3) and (6,5) absorption features (due to overlap). The estimated abundances then become 3.5% (6,4), 2.5% (9,1), 3% (7,5) and 91% for combined (8,3) and (6,5). If we use the PLE peak intensities of the (8,3) and (6,5) features we can estimate that (8,3) and (6,5) abundances account for 74% and 17% of all SWNTs respectively. In absorption spectra the first subband excitons of (8,3) and (6,5) SWNTs thus evidently show some overlap. The significance of this for the transient absorption signal can be estimated by accounting for the resonant excitation of the (6,5) SWNT at the pump wavelength 5
of 987 nm and correspondingly weak excitation of the (8,3) subband exciton at 960 nm. In combination with the above estimate of relative abundances this leads us to conclude that the transient signal contribution of the first (8,3) subband exciton in our experiments is below about 5% of that of the first (6,5) subband exciton.
References (1) Arnold, M. S.; Green, A. A.; Hulvat, J. F.; Stupp, S. I.; Hersam, M. C. Sorting Carbon Nanotubes by Electronic Structure Using Density Differentiation. Nature Nanotechnology 2006, 1, 60–65. (2) Schöppler, F.; Rühl, N.; Hertel, T. Photoluminescence Microscopy and Spectroscopy of Individualized and Aggregated Single-Wall Carbon Nanotubes. Chemical Physics 2013, 413, 112–115. (3) Aguirre, C. M.; Levesque, P. L.; Paillet, M.; Lapointe, F.; St-Antoine, B. C.; Desjardins, P.; Martel, R. The Role of the Oxygen/Water Redox Couple in Suppressing Electron Conduction in Field-Effect Transistors. Advanced Materials 2009, 21, 3087. (4) Schilling, D.; Mann, C.; Kunkel, P.; Schöppler, F.; Hertel, T. Ultrafast Spectral Exciton Diffusion in Single-Wall Carbon Nanotubes Studied by Time-Resolved Hole Burning. Journal of Physical Chemistry C 2015, 119, 24116–24123.
6
