NANO LETTERS
Influence of Gas Adsorption on Optical Transition Energies of Single-Walled Carbon Nanotubes
2008 Vol. 8, No. 10 3097-3101
Shohei Chiashi,* Satoshi Watanabe, Tateki Hanashima, and Yoshikazu Homma Department of Physics, Tokyo UniVersity of Science, Shinjuku, Tokyo 162-8601, Japan Received April 15, 2008; Revised Manuscript Received July 31, 2008
ABSTRACT The photoluminescence (PL) spectra of suspended single-walled carbon nanotubes (SWNTs) were measured in an ethanol gas atmosphere. When the gas pressure was decreased, the PL peaks were initially blue-shifted to a small extent before a rapid blue-shift took place at a transition pressure that depended on the temperature and diameter of the SWNT being measured. This pressure dependence is due to the adsorption of ethanol molecules on the SWNT surface. The optical transition energies measured below the transition pressure are intrinsic to the SWNT.
Since photoluminescence (PL) measurements of surfactantwrapped single-walled carbon nanotubes (SWNTs) in solution were first performed,1 the possibility of using PL spectra for studying the electronic and optical properties of SWNTs has been examined.2,3 The peaks in a PL map, which is composed of PL emission and excitation spectra, correspond to the optical transition energies between the van Hove singularities (E11, E22, etc.) of semiconducting SWNTs, and are used for determining the chiral indices of SWNTs.4 However, the transition energies measured from wrapped SWNTs in solution depend both on the wrapping materials (surfactants, polymers, and DNA strands have been used) and on environmental conditions such as the temperature and the pH of the solution.5-11 Therefore, it has proved difficult to analyze the PL spectra obtained from wrapped SWNTs and thus to investigate their intrinsic electronic properties. To address this problem, an alternative approach was developed in which the PL spectra of individual SWNTs suspended between microstructures were measured; these can be synthesized by combining lithography techniques and chemical vapor deposition (CVD) methods.12,13 Because the suspended SWNTs are not in contact with any substrate or surrounding medium and because they also emit intense, sharp PL peaks, these structures were deemed more suitable for the investigation of intrinsic optical properties. However, it has been reported that suspended SWNTs are not free from the influence of environmental conditions either.14-16 Finnie et al.15 studied the temperature dependence of the PL peaks of suspended SWNTs; a sudden shift in energy was observed close to room temperature, the origin of which was unclear. * To whom correspondence should be addressed. E-mail: chiashi@ rs.kagu.tus.ac.jp. 10.1021/nl801074j CCC: $40.75 Published on Web 08/30/2008
2008 American Chemical Society
This rapid energy shift depended on the gas atmosphere used during the measurement, and the authors suggested that it was caused by the adsorption of water or oxygen on the SWNT surface. In the present paper, we focus on the changes that occur in the PL and PL excitation spectra of suspended SWNTs when the temperature and gas atmosphere are varied. Ethanol was used as the ambient gas, molecules of which are readily adsorbed on a graphite surface at room temperature,17 and the dependence of the PL spectra on the gas pressure was investigated. We measured the PL spectra of SWNTs suspended between pairs of quartz pillars, as shown in Figure 1A.12 The pillar structures were fabricated on quartz substrates by photolithography techniques. The height and diameter of the pillars were 5 and 2 µm, and the pillar spacing was 5 µm. We used a SWNT synthesis technique based on the alcohol catalytic CVD (ACCVD) method.18 Cobalt catalyst films with a nominal thickness of 0.1 Å were deposited onto the substrates using a vacuum evaporator. Ethanol was used as the carbon source and was buffered with an Ar/H2 gas mixture (3% H2 by volume) using the bubbling method. The substrates were heated under this Ar/H2 atmosphere in a quartz tube (40 mm diameter) with the temperature being controlled using an electronic furnace. After the CVD temperature had been attained, the gas supply was changed from Ar/H2 to buffered ethanol. The total pressure of the CVD gas was approximately 700 Torr. CVD was carried out at 850 °C for approximately 5 min. In the PL measurements, a Ti:Sapphire excitation laser (690-830 nm, Coherent) was used, and the PL signal was detected by an InGaAs multiarray-detector (1024 channels, SpectraPro 2300i, Acton), which was cooled with liquid
Figure 1. (A) Scanning electron microscope image of a SWNT suspended between a pair of quartz pillars. (B) Schematic illustration of the measurement chamber.
nitrogen to -100 °C. The emission spectra were measured from 900 to 1700 nm. The excitation laser was focused using an objective lens (50×) to a spot size of approximately 2 µm in diameter, and the laser power incident on the samples was 30 µW. The PL measurements were performed inside the vacuum chamber shown in Figure 1B, where both the sample temperature and ethanol gas pressure were controlled. The quartz substrate was glued onto a silicon heater using indium metal; a Joule-heating technique with AC voltage was used. Ethanol gas was supplied from a reservoir in which it was dehydrated using a molecular sieve, and the ethanol pressure was controlled using a mass-flow controller combined with a scroll mechanical pump. Because an address number was assigned to each of the pillars, we were able to repeatedly measure PL spectra from the same suspended SWNTs. The study in ref 15 reported that the PL emission peaks of suspended SWNTs rapidly shift in energy at approximately 40 °C in both gas atmospheres and in vacuum. We also observed a rapid blue-shift when our SWNTs were heated in vacuum. After this blue-shift had been induced, ethanol gas was supplied to the vacuum chamber and PL spectra were measured at room temperature. PL maps of a suspended SWNT in (A) ethanol gas (5.0 × 10 Torr) and in (B) vacuum (6.3 × 10-2 Torr) are shown in Figure 2. The emission and excitation energies of the PL peak in Figure 2A closely agree with those obtained previously from suspended SWNTs measured in air,16,19 and enable the chiral indices of our SWNT to be determined (9,8). The PL peak in Figure 2B was blue-shifted by 30 meV with respect to that in Figure 2A for emission energies and by 47 meV for excitation energies. The widths of both the emission and excitation peaks measured in vacuum were the same as those measured in ethanol gas. The PL spectra measured using a 780.0 nm excitation laser at room temperature (28.2 °C) and at various ethanol gas pressures are shown in Figure 2C. Initially, the 3098
Figure 2. PL maps of a suspended (9,8) SWNT in (A) ethanol gas and (B) vacuum. (C) Emission spectra measured at various ethanol gas pressures at 28.2 °C. The excitation laser wavelength was 780.0 nm.
PL emission peak was slightly blue-shifted with decreasing ethanol pressure, before becoming broadened and rapidly blue-shifted at a particular transition pressure (see the spectrum measured at 7.09 × 10-1 Torr). Below this transition pressure, the peak became sharp and did not change further in energy. All of the suspended SWNTs measured showed essentially the same dependence of emission peak wavelength and peak width on ethanol pressure. It was difficult to measure the temperature of the suspended SWNTs directly, and thus the SWNT temperature was assumed to be the same as room temperature. The ethanol gas pressure dependence of a PL emission peak for a suspended (8,6) SWNT measured at three different temperatures is shown in Figure 3A. Although the emission peak wavelength showed a qualitatively similar pressure dependence at all three temperatures, the transition pressure increased with temperature. Below the transition pressure, the emission peak wavelength remained constant and was independent of pressure and temperature. In contrast, above the transition pressure the wavelength increased slightly with pressure. The transition pressure was dependent on the ethanol flow rate when the laser power was high. Laser irradiation with a high power density heated the suspended SWNT, while the flow of ethanol gas cooled the SWNT due to heat transfer. Therefore, the excitation laser power was lowered to avoid undesired heating, and the flow of ethanol gas was stopped during the PL measurements. The ethanol pressure dependence of the PL peaks, including the rapid energy shift and broadening at the transition pressure, can be explained by the adsorption and desorption of ethanol molecules on the SWNT surface. The adsorption rate of gas molecules on the surface (ν+) is given by P
(1 - θ) (1) √2mkT where P is the gas pressure, m is the molecular mass, k is the Boltzmann constant, T is the gas temperature and θ is ν+ )
Nano Lett., Vol. 8, No. 10, 2008
order to obtain the values of A and P0 from the experimental data, eq 3 was rewritten as
( )
log(Pt) - 0.5 log T - log
Figure 3. (A) Dependence of emission wavelength on ethanol gas pressure at various temperatures. The chiral indices of the SWNT were (8,6). (B) Dependence of transition pressure on temperature and tube diameter for SWNTs with four different chiral indices. The dashed line is a fitted line (see main text).
the fraction of surface covered (0 e θ e 1). Because ethanol molecules have a hydroxyl group, we also took into account the interactions between different ethanol molecules (hydrogen bonding) on the SWNT surface. On the basis of the modified Langmuir model,20 which includes interactions not only between an adsorbate molecule and the surface but also among adsorbate molecules, the desorption rate (ν-) is given by
(
ν- ) ν0 exp -
E + zwθ θ kT
)
(2)
where ν0 is a proportional constant, E is the desorption energy at zero coverage, z is the number of neighboring molecules, and w is the lateral interaction energy between adjacent adsorbents. At equilibrium the adsorption rate (ν+) is equal to the desorption rate (ν-), and the pressure becomes P ) P0
θ E + zwθ √T exp 1-θ kT
(
)
(3)
where P0 ) ν0(2πmk)1/2. The critical pressure (Pc), which is the pressure corresponding to the critical coverage θc ) 0.5 in eq 3,20 was taken to be the same as the transition pressure Pt. In the case of SWNTs, the desorption energy at zero coverage (E) is dependent on the tube diameter,21 and was assumed to be E ) Eg - A/dt
(4)
where dt is the tube diameter, A is a constant, and Eg is the desorption energy of ethanol on graphite at zero coverage (29 kJ/mol22). The quantity Eg + zwθc corresponds to the desorption energy of ethanol on graphite (57.6 kJ/mol23). In Nano Lett., Vol. 8, No. 10, 2008
Eg + zwθc θc ) log(P0) + + 1 - θc kT A 1 (5) k Tdt
The transition pressures of four SWNTs with different chiral indices were measured at various temperatures between 23 and 29 °C, and Figure 3B shows the relationship between 1/(Tdt) and the left-hand side of eq 5. The experimental data were fitted well using values of A ) 29 kJ nm/mol and P0 ) 1.6 × 104 Torr/K1/2, as shown by the fitted line in Figure 3B. The dependence of the transition pressure (Pt) on the temperature and tube diameter can thus be explained by the curvature effect of the desorption energy and by the modified Langmuir model. According to the modified Langmuir model, when the value of zw/kT in eq 3 is greater than 4.0, the adsorbents undergo a phase transition. In our case of adsorbed ethanol on a SWNT surface, the value of zw/kT was approximately 20 at room temperature. When the ethanol gas pressure is lower than the transition pressure, the number of adsorbed ethanol molecules on the SWNT surface is extremely small. At the transition pressure, some of the adsorbed ethanol molecules are condensed and others are isolated on the SWNT surface (a coexistence of 2-dimensional liquid and 2-dimensional gas phases). After a monolayer of ethanol in the liquid phase is formed above the transition pressure, the number of adsorbed ethanol molecules increases with pressure and a progressively thicker layered structure is formed. We will now compare our experimental results with those of previous gas desorption studies. In general, gas desorption energies are obtained using thermal desorption spectroscopy (TDS). This technique involves first exposing the sample to the gas and then counting the number of gas molecules desorbed from the sample using a mass spectrometer, while heating under ultrahigh vacuum conditions. By analysis of the thermal desorption spectra, the rate of desorption is given by
( )
Ea dθ ) -νθn exp dt kT
(6)
where ν is the pre-exponential frequency factor, n is the order of desorption and Ea is the activation energy for desorption. In a previous study of ethanol adsorption on graphite, the desorption order was zero, which means that the ethanol molecules exist in both the liquid and gas phases on the graphite surface, and ν was found to be 2 × 1015(2 s-1.24 Equation 6 can then be expressed as dθ d F(θ) 1 ) ν (θ) ) dt dt F(θ ) 1) F(θ ) 1) -
(
)
(7)
where F(θ) is number density of adsorbed ethanol molecules for a coverage θ. The F(θ ) 1) value, which corresponds to the saturated quantity of ethanol in the first adsorbed layer, was found to be 4.92 nm-2 in another study on graphite.25 Although a comparison with our current system is not straightforward and the measured quantities have relatively large uncertainties, the value of ν0θc/F(θ ) 1) corresponds 3099
Table 1. Optical Transition Energies of SWNTs with Different Chiral Indices in Vacuum (n,m)
E11 (eV)
E22 (eV)
(10, 2) (9, 4) (8, 6) (12, 1) (11, 3) (10, 5) (8, 7) (9, 7) (10, 6) (9, 8) (12, 5) (11, 7)
1.246 1.178 1.118 1.094 1.085 1.042 1.032 0.990 0.954 0.927 0.878 0.867
1.746 1.788 1.789 1.626 1.634 1.640 1.751 1.620 1.688 1.590 1.610 1.531
to the pre-exponential frequency factor ν in eq 6, which was calculated to be 1011 s-1 from our experimental results. This is smaller than the value of ν obtained by TDS for the ethanol on graphite system.24 Nevertheless, because the measured values of ν have large uncertainties of 4 orders of magnitude, the two values are comparable. This indicates that the modified Langmuir model is valid in the analysis of our experimental results. Moreover, in a previous TDS study of ethanol gas desorbed from bundled SWNT ensembles, desorption was observed from 240 K to higher than 400 K.24 This is much wider than the temperature range of our measurements, which implies that the bundled SWNT ensembles had an effective surface that was heterogeneous and porous. In contrast, the isolated and suspended SWNTs of our study have an ideal, uniform surface structure, which gives rise to well-defined and rapid adsorption and desorption phenomena, as shown in Figure 3A. PL spectra of suspended SWNTs immersed in various kinds of liquids have previously been measured.26,27 The emission and excitation energies were found to depend on the type of liquid and the shifts in energy of the peaks were explained in terms of the dielectric constants of the liquids. Although the case of immersion in ethanol is not reported in refs 26 and 27, the adsorption of ethanol molecules will also act to increase the dielectric constant of the surroundings of the SWNTs. The effect of the change in dielectric constant on the optical transition energies will increase with the thickness of the ethanol layer. This explains our observation that the emission peak wavelength increases slightly as the ethanol pressure is increased above the transition pressure. The broadening and rapid shifting of the PL peaks in the vicinity of the transition pressure occurs because adsorbed ethanol on the SWNT surface exists in both the liquid and gas phases and thus the value of θ is not uniform. The pressure sensitivity of the PL peak wavelengths in the vicinity of the transition pressure might be a good characteristic for applications in the sensing of gas pressures. Below the transition pressure, very few ethanol molecules are adsorbed on the SWNT surface. Therefore, the PL peak wavelengths do not show any dependence on ethanol pressure in this range. The optical transition energies of suspended SWNTs in ambient air show rapid shifts in energy when the sample temperature changes.15 This indicates that the surfaces of suspended SWNTs exposed to air undergo adsorption by molecules such as water, whose vapor pressure is less than that of ethanol. Therefore, the emission and excitation 3100
Figure 4. Emission and excitation wavelengths of surfactantwrapped SWNTs (crosses), suspended SWNTs in air (open circles) and suspended SWNTs in vacuum (filled circles).
energies measured below the transition pressures are the intrinsic optical transition energies of the SWNTs. Table 1 shows the optical transition energies of a variety of SWNTs in vacuum, and Figure 4 shows the corresponding emission and excitation energies measured in air16,19 and in vacuum (below the transition pressure). For reference, the corresponding values for surfactant-wrapped SWNTs4,28 are also plotted. The suspended SWNTs exposed to air have larger values of E11 and E22 than those of surfactant-wrapped SWNTs. All of the PL peaks of suspended SWNTs measured in vacuum are blue-shifted to a greater extent than those measured in air. The average emission and excitation energy shifts are approximately 25 and 47 meV, respectively. Semiconducting SWNTs can be classified into two types according to their structure [type I with ((2n + m) mod 3 ) 1) and type II with ((2n + m) mod 3 ) 2)]. Although the PL spectra of surfactant-wrapped SWNTs measured at low temperatures were previously found to be dependent on whether the nanotubes were of type I or II,7,8 we observe no such dependence in the data in Figure 4. This indicates that the gas adsorption and desorption processes did not cause structural deformation or strain in our SWNTs.29 In conclusion, we have determined the ethanol gas pressure dependence of the PL spectra of suspended SWNTs. When decreasing the pressure, the emission spectra showed rapid blue-shifting and broadening at a transition pressure. Below this transition pressure, the PL peaks showed no pressure dependence. The transition pressure depends on both the temperature and tube diameter of the SWNT. We attribute the dependence of the PL spectra on ethanol gas pressure to the adsorption and desorption of ethanol molecules on the SWNT surface; the optical transition energies measured below the transition pressure are intrinsic to the SWNT. Acknowledgment. This work was partially supported by a Grant-in-Aid for Scientific Research on Priority Area (No. 19054015) from the Ministry of Education, Culture, Sports, Science, and Technology (MEXT), Japan. References (1) O’Connell, M. J.; Bachilo, S. M.; Huffman, C. B.; Moore, V. C.; Strano, M. S.; Haroz, E. H.; Rialon, K. L.; Boul, P. J.; Noon, W. H.; Kittrell, C.; Ma, J. P.; Hauge, R. H.; Weisman, R. B.; Smalley, R. E. Science 2002, 297, 593–596. (2) Lebedkin, S.; Hennrich, F.; Skipa, T.; Kappes, M. M. J. Phys. Chem. B 2003, 107, 1949–1956. (3) Hartschuh, A.; Pedrosa, H. N.; Novotny, L.; Krauss, T. D. Science 2003, 301, 1354–1356. (4) Bachilo, S. M.; Strano, M. S.; Kittrell, C.; Hauge, R. H.; Smalley, R. E.; Weisman, R. B. Science 2002, 298, 2361–2366. Nano Lett., Vol. 8, No. 10, 2008
(5) Strano, M. S.; Huffman, C. B.; Moore, V. C.; O’Connell, M. J.; Haroz, E. H.; Hubbard, J.; Miller, M.; Rialon, K.; Kittrell, C.; Ramesh, S.; Hauge, R. H.; Smalley, R. E. J. Phys. Chem. B 2003, 107, 6979– 6985. (6) Dukovic, G.; White, B. E.; Zhou, Z. Y.; Wang, F.; Jockusch, S.; Steigerwald, M. L.; Heinz, T. F.; Friesner, R. A.; Turro, N. J.; Brus, L. E. J. Am. Chem. Soc. 2004, 126, 15269–15276. (7) Arnold, K.; Lebedkin, S.; Kiowski, O.; Hennrich, F.; Kappes, M. M. Nano Lett. 2004, 4, 2349–2354. (8) Li, L. J.; Nicholas, R. J.; Deacon, R. S.; Shields, P. A. Phys. ReV. Lett. 2004, 93, 156104. (9) Htoon, H.; O’Connell, M. J.; Cox, P. J.; Doorn, S. K.; Klimov, V. I. Phys. ReV. Lett. 2004, 93, 027401. (10) Hertel, T.; Hagen, A.; Talalaev, V.; Arnold, K.; Hennrich, F.; Kappes, M.; Rosenthal, S.; McBride, J.; Ulbricht, H.; Flahaut, E. Nano Lett. 2005, 5, 511–514. (11) Jeng, E. S.; Moll, A. E.; Roy, A. C.; Gastala, J. B.; Strano, M. S. Nano Lett. 2006, 6, 371–375. (12) Lefebvre, J.; Homma, Y.; Finnie, P. Phys. ReV. Lett. 2003, 90, 217401. (13) Lefebvre, J.; Fraser, J. M.; Finnie, P.; Homma, Y. Phys. ReV. B 2004, 69, 075403. (14) Lefebvre, J.; Finnie, P.; Homma, Y. Phys. ReV. B 2004, 70, 045419. (15) Finnie, P.; Homma, Y.; Lefebvre, J. Phys. ReV. Lett. 2005, 94, 247401. (16) Ohno, Y.; Iwasaki, S.; Murakami, Y.; Kishimoto, S.; Maruyama, S.; Mizutani, T. Phys. ReV. B 2006, 73, 235427.
Nano Lett., Vol. 8, No. 10, 2008
(17) Morishige, K. J. Chem. Phys. 1992, 97, 2084–2089. (18) Maruyama, S.; Kojima, R.; Miyauchi, Y.; Chiashi, S.; Kohno, M. Chem. Phys. Lett. 2002, 360, 229–234. (19) Lefebvre, J.; Fraser, J. M.; Homma, Y.; Finnie, P. Appl. Phys. A 2004, 78, 1107–1110. (20) Adamson, A. W. Physical Chemistry of Surfaces, 4th ed.; John Wiley & Sons: New York, 1982. (21) Pankewitz, T.; Klopper, W. J. Phys. Chem. C 2007, 111, 18917–18926. (22) Walker, P. L. Chemistry and Physics of Carbon; Marcel Dekker: New York, 1970. (23) Shukla, N.; Gui, J.; Gellman, A. J. Langmuir 2001, 17, 2395–2401. (24) Ulbricht, H.; Zacharia, R.; Cindir, N.; Hertel, T. Carbon 2006, 44, 2931–2942. (25) Morishige, K.; Hayashi, K.; Izawa, K.; Ohfuzi, I.; Okuda, Y. Phys. ReV. Lett. 1992, 68, 2196–2199. (26) Miyauchi, Y.; Saito, R.; Sato, K.; Ohno, Y.; Iwasaki, S.; Mizutani, T.; Jiang, J.; Maruyama, S. Chem. Phys. Lett. 2007, 442, 394–399. (27) Ohno, Y.; Iwasaki, S.; Murakami, Y.; Kishimoto, S.; Maruyama, S.; Mizutani, T. Mater. Sci. (cond-mat.mtrl-sci) http://arxiv.org/abs/ 0704.1018v1, 2007. (28) Bachilo, S. M.; Balzano, L.; Herrera, J. E.; Pompeo, F.; Resasco, D. E.; Weisman, R. B. J. Am. Chem. Soc. 2003, 125, 11186–11187. (29) Yang, L.; Han, J. Phys. ReV. Lett. 2000, 85, 154–157.
NL801074J
3101
