2.3 Characterization
2.3.4 Wide range transmission spectroscopy
Wide range optical transmission measurements on carbon nanotube self-supporting thin films combined with Kramers-Kronig transformation and fitting with the Drude-Lorentz model give detailed information about the optical transitions in the sample.
From these results, quantitative information can be obtained of the contributions of tubes with different diameters. Since addition reactions remove electrons from the nano-tube’s sp2 electronic system, the peak intensity of transitions related to those electrons will decrease. With proper data evaluation this decrease can be quantitatively
deter-mined. From the different degree of decrease of different diameters, we can judge the diameter selectivity of addition reactions.
Thin film preparation
For wide range optical measurements, self-supporting nanotube thin films were pre-pared by vacuum filtration (see Figure 2.4).
Typically 5-10 mg nanotube sample was dispersed in a Triton X-100-water solu-tion (Triton X-100 from VWR, 98+%) to get individually dispersed nanotubes, and was sonicated for at least 2 hours to obtain a homogeneous dispersion. Sonication time is dependent on the sample and can vary in a wide range. The dispersion was left overnight (or several days) to let the bigger bundles precipitate. The top of the dispersion remained homogeneous and contained more or less individually dispersed nanotubes. From the top of the dispersion 10-50 ml was added to the funnel of the fil-ter filled with distilled wafil-ter and filfil-tered on a nitrocellulose, acetone-soluble membrane filter (Millipore, 0.1 µm pore size). Special attention was paid to avoid any movement of the filter. The sample was washed with distilled water, then the membrane filter was carefully dissolved in acetone. The nanotube thin films were placed over a 2 mm diame-ter hole on a graphite disc. Graphite is the only suitable frame madiame-terial, because it has similar thermal dilatation to nanotube thin films. This is very important, because the so-made nanotube thin films were annealed in dynamic vacuum at 200◦C for 12 hours in order to get rid of any volatile and dopant species.
For measuring the thickness of the film by AFM, the same filter was used, but another piece of film was moved on a Si piece and was cut in the middle. The height of this stage was used considering the difference between the value of Si and smooth nanotube surface after flattening the raw AFM image (the fold and crease of the stage was not considered). For precise determination of the thickness, a histogram excluding the stage values was used. Areas used for thickness determination were 10x10µm, and at least 10 areas were used at one sample. The samples were usually 160-290 nm thick.
Figure 2.4: Self-supporting thin film preparation. a) filtering; b) filtrate on soluble membrane filter; c) transparent self-supporting thin film [24].
Measurement parameters
Wide range optical transmission measurements were carried out on self-supporting thin films made of the samples [82]. Transmission data between 25-52500 cm−1 were recorded, by a Bruker IFS 66v/s FT-IR instrument in the far (25-1000 cm−1) and mid-infrared (400-7500 cm−1) region, a Bruker Tensor37 in the near infrared (4000-15000 cm−1), and a Jasco v550 spectrometer in the visible and UV (11100-52500 cm−1).
Measuring in wide spectral range requires using more than one instrument. In dif-ferent spectral ranges difdif-ferent light sources, optical elements, detectors etc. have to be used, and also different substrates are needed to support the sample. Evaluation of the spectra is made difficult by the absorption of the substrate, because there is no substrate transparent in the whole spectral range. To solve the problem, the best solution would be to get rid of the substrates. The special mechanical properties of carbon nanotube thin films permit this [82].
Data evaluation
By using a Matlab-based program written by ´Aron Pekker, Kramers-Kronig trans-formations were performed on the transmittance data to calculate phaseφ:
φ(ω0) = 2πdω0 − Z +∞
0
ln(p
T(ω)/p
T(ω0))
ω2−ω02 dω (2.1)
whered thickness is determined experimentally by AFM measurements. Standard ex-trapolations were used to zero and infinite frequencies: at low wavenumbers frequency-independent, at high wavenumbers power law decrease.
The optical conductivity data were calculated and fitted by Drude-Lorentz oscil-lators. After subtracting the background and all other peaks related to transitions between different Van Hove singularities, we obtained the contribution of the specific peaks [83]. However, these peaks cannot be assigned to single nanotubes, because the transition energies are closer to each other than the width of the peaks.
Out of the available optical functions, optical conductivity ˆσ was chosen since it is additive when several independent processes are involved, like light absorption by different nanotubes [84]. Each process can be modeled by a separate harmonic oscillator:
ˆ From Equations 1.10 and 2.2 the real part of the optical conductivity can be written as: From theσ0(ω) curve, the parameters of the individual oscillators can be determined by standard numerical processes. The fitted Drude-Lorentz oscillators represent the contribution of the transitions between Van Hove singularities of 1D electronic systems (nanotubes), the excitation of the fullπ-electron system (π−π∗transitions in nanotubes and other carbonaceous materials as contaminants). Theπ−π∗transitions (peaks above
∼40000 cm−1) and metallic contaminants, like remaining catalysts are considered as a constant (broad and weak Drude oscillator), other contaminants are considered as a few Lorentzians.
In Figure 2.5 the Drude-Lorentz fit of the optical conductivity spectrum (σ0 vs. ω) is shown, together with the process to obtain the contribution of the specific peaks.
The optical conductivity spectrum has very similar shape as the absorbance spectrum
0 1 0 0 0 0 2 0 0 0 0 3 0 0 0 0 4 0 0 0 0 5 0 0 0 0
Figure 2.5: a) Fitting and background correction of nanotube sample’s optical con-ductivity spectrum; b) extraction of single transitions from the background-corrected optical conductivity spectrum and the fitted Drude-Lorentz oscillators. Own measure-ment of P2 reference sample is used to show the evaluation process.
shown in Figure 1.6. This was the other reason why we have chosen optical conductivity representation out of the available optical functions.
Optical conductivity spectra of the starting materials
Figure 2.6 shows the diameter dependence of S11 and S22 transitions of the start-ing P2, HiPco and CoMoCat nanotubes. It can be clearly seen how the transition wavenumbers change with tube diameter: the smaller the mean diameter, the larger the wavenumber. Thus, the lower wavenumber part of the peak is related to the larger diameter tubes within one sample, the higher wavenumber part to the smaller diameter tubes. When detecting the different changes in the intensities of the lower and higher wavenumber parts within one sample, we can draw conclusions about the diameter selectivity of an addition reaction.
The width of the peaks is due to the diameter distribution of the samples. P2 nanotubes have a more or less narrow, featureless S11 peak indicating that the single components are very close to each other. The S22 peak of P2 shows a more complicated structure, since the energy of these transitions is separated more from each other (see Figure 1.5).
HiPco nanotubes have very wide and much more structured S11 and S22peaks. This indicates that HiPco has a wide diameter distribution but contains quite less types of nanotubes than P2.
CoMoCat nanotubes show also wide peaks, but it is because of their small mean diameter (large difference in transition energies between tubes with close diameters).
Accordingly, these wide peaks are very much structured due to the small number of constituting (n, m) tubes.
2 0 0 0 4 0 0 0 6 0 0 0 8 0 0 0 1 0 0 0 0 1 2 0 0 0 1 4 0 0 0
0
5 0 1 0 0 1 5 0 2 0 0 2 5 0 3 0 0 3 5 0
P 2 ( 1 . 6 0 n m ) H i P c o , x 0 . 2 ( 1 . 0 8 n m ) C o M o C a t ( 0 . 9 0 n m )
Optical conductivity (Ω-1 cm-1 ) W a v e n u m b e r ( c m - 1)
a )
5 0 0 0 1 0 0 0 0 1 5 0 0 0 2 0 0 0 0 2 5 0 0 0
0
5 0 1 0 0 1 5 0 2 0 0
Optical conductivity (Ω-1 cm-1 ) W a v e n u m b e r ( c m - 1)
P 2 ( 1 . 6 0 n m ) H i P c o , x 0 . 2 ( 1 . 0 8 n m ) C o M o C a t ( 0 . 9 0 n m ) b )
Figure 2.6: Diameter dependence of a) S11; and b) S22transitions. The tendency towards lower wavenumbers with increasing mean diameters can be clearly seen.
3 | Results and discussion
In this section I present the results obtained by Raman spectroscopy, thermogravimetry-mass spectrometry (TG-MS), wide range transmission spectroscopy and1H-NMR spectrometry used for characterizing the samples. The samples, reference samples and starting materials were investigated by Raman spectroscopy to determine the D/G mode intensity ratio. This quick measurement gives the first evidence that sidewall functionalization has taken place. TG-MS and 1H-NMR measurements give quantitative information about the H and n-Bu content of the samples, and TG-MS in addition about the diameter dependence by the evolution temperature. Quantitative evaluation of the results obtained by wide range optical spectroscopic measurements yields information about diameter selectivitywithin one sample.
3.1 Hydrogenation reactions on HiPco single-walled carbon nanotubes
On HiPco nanotubes both alkali metal intercalation (by using potassium as inter-calating agent) and modified Birch reduction were performed in three successive steps in order to study the difference in reactivity and diameter selectivity.
The as-made samples were characterized by Raman spectroscopy, TG-MS and wide range optical spectroscopy.
1 2 0 0 1 3 0 0 1 4 0 0 1 5 0 0 1 6 0 0 1 7 0 0
Normalized Raman intensity (a. u.)
R a m a n s h i f t ( c m - 1) D b a n d
G b a n d H i P c o s t a r t i n g m a t e r i a l
1 x 2 x
3 x h y d r o g e n a t e d
Figure 3.1: Raman spectra of HiPco samples hydrogenated by alkali metal (potassium) intercalation. D and G bands are indicated.
1 2 0 0 1 3 0 0 1 4 0 0 1 5 0 0 1 6 0 0 1 7 0 0
Normalized Raman intensity (a. u.)
R a m a n s h i f t ( c m - 1) D b a n d
G b a n d H i P c o s t a r t i n g m a t e r i a l
1 x 2 x
3 x h y d r o g e n a t e d
Figure 3.2: Raman spectra of HiPco samples hydrogenated by modified Birch reduction.
D and G bands are indicated.
Figure 3.3: Typical TG-MS curve of HiPco hydrogenated by intercalating potassium into the bundles. Black solid lines represent the mass loss (TG curve) and its derivative (DTG curve). Notations: (red circle) hydrogen (m/z 2); (blue triangles) methane (m/z 16); (green diamonds) methanol (m/z 31); (magenta squares) m/z 92 toluene.