Polymer Durability and Radiation Effects - American Chemical Society

---

Chapter 20

Downloaded by UNIV OF CALIFORNIA SAN FRANCISCO on December 7, 2014 | http://pubs.acs.org Publication Date: December 21, 2007 | doi: 10.1021/bk-2007-0978.ch020

Effects of Protons, Electrons, and UV Radiation on Carbon Nanotubes 1

1

3

4

Chong Oh Lee , Ebrahim Najafi , Jae Yong Kim , Song-Hee Han , Takhee Lee , and Kwanwoo Shin * 1

2,

1

Department of Materials Science and Engineering, Gwangju Institute of Science and Technology, Gwangju 500-712, South Korea Department of Chemistry, Sogang University, Seoul, 121-742, South Korea Department of Physics, Hanyang University, Seoul 133-791, South Korea Advanced Photonics Research Institute, Gwangju Institute of Science and Technology, Gwangju 500-712, South Korea

2

3

4

The effects of high energy radiations, including proton, electron and UV, on the physical and chemical properties of carbon nanotubes (CNT) were investigated. The CNTs sheets were irradiated by high energy proton beam of 10 MeV, electron beam of 2 MeV, and UV radiation in ambient environment, and surface functionalization and morphological structures of CNTs before and after high-energy radiations were monitored by X-ray photoelectron spectroscopy, infrared and Raman spectroscopy. When CNTs were irradiated by proton and electron particles, reactions were induced by the transfer of kinetic energy of the bombarding particles while UV-light dissociated adsorbed 0 molecules into atomic oxygens, leading to the surface oxidation to CNTs. The irradiated CNTs exhibited radiation induced changes, distinct from native properties inherited from the classical synthetic routes. 2

232

© 2008 American Chemical Society

In Polymer Durability and Radiation Effects; Celina, M., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2007.

233

Downloaded by UNIV OF CALIFORNIA SAN FRANCISCO on December 7, 2014 | http://pubs.acs.org Publication Date: December 21, 2007 | doi: 10.1021/bk-2007-0978.ch020

Introduction Satellites, spacecraft and equipment used in particle accelerators for high energy nuclear physics involve exposure to high energy radiation and thermal cycling. For example, the radiation environment in space and the atmosphere consists of trapped protons and electrons, cosmic radiation and neutrons. Radiation effects cause extensive damage to a space facility, weakening its performance and lifetime, include latchup, interference, degradation, charging, sputtering, erosion and puncture of materials [1]. As summarized in Figure 1, their energies in space and the atmosphere range from several eV to one GeV equivalent; Solar cosmic rays, with energies of 10 to 10 eV, and galactic cosmic rays, with energies of 10 to 10 eV, long-wave infra red (IR) and ultra violet (UV) radiation in ozonized environments. Without appropriate protection, the exposed surface coatings, especially organic thin films, undergo dissociation, abstraction, and additional reactions that lead to severe mechanical damage, eventually resulting in system failure of the underlying electronics. Further, a single source can induce different reactions on a target, depending on its energy level. Therefore, much effort has been made, on the one hand, to develop advanced synthetic routes by utilizing radiation reactions and, on the other, to test radiation resistant coating surfaces, primarily with the aim of preventing degradation reactions. One of common approaches for strengthening the durability against radiation, is the addition of fillers to the matrix materials. Carbon nanotubes (CNT)s have inspired scientists and engineers to examine a wide range of potential applications [2,3] as exotic filler materials, since the discovery of their extraordinary mechanical, electrical, magnetic and thermal properties [4]. 7

10

10

15

• The elastic modulus is ~ 1 TPa for single-walled nanotubes (SWNT) and ~ 0.3- 1 TPa for multi-walled nanotubes (MWNT). • The strength is 50-500 GPa for SWNT and 10-60 GPa for MWNT. • The resistivity is 5-50 μΩαη. • The thermal conductivity is theoretically as high as 3000 Wm^K' . • The thermal expansion is negligible, while thermal stability is more than 700 °C in air and 2800 °C in vacuum. 1

These properties are mainly originated from the molecular structures of CNTs, which consist of graphene sheets rolled to form hollow cylinders with an extremely high aspect ratio [5]. Moreover, CNTs can be ether metallic or semi­ conducting tubes, depending on their diameter and/or chirality [5]. There are two types of CNTs: SWNTs and MWNTs. Details, such as structure, synthesis methods, application and properties, are well documented [4], [5], [6].

In Polymer Durability and Radiation Effects; Celina, M., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2007.

Downloaded by UNIV OF CALIFORNIA SAN FRANCISCO on December 7, 2014 | http://pubs.acs.org Publication Date: December 21, 2007 | doi: 10.1021/bk-2007-0978.ch020

234 As described above, the unique properties of CNTs provide excellent advantages for many applications and much effort has been focused on developing methods that permit the selective and, more importantly, controlled fiinctionalization of CNTs. Furthermore, since several micrometer-long CNT chains are physically entangled and tightly bundled due to their strong van der Waals interaction, CNTs tend to segregate in most solvents [7]. Consequently, an appropriate chemical modification is necessary to disturb the charge balance around CNTs, or to attach the organic moieties, leading to an enhanced solubility in common organic and aqueous solvents [8-11]. One of the promising methods is to introduce various functional groups on their nonreactive surfaces via oxidative processes [12]. Although many studies have addressed the covalent modification of CNTs in an aqueous environment, little attention has been devoted to dry methods. Dry processes support a mild oxidizing environment and, as a result, are usually less effective than wet methods. Furthermore, dry methods require the accurate control of the temperature, atmosphere, and reaction time during the process [13,14]. In this regard, much research has focused the direct modification of the CNT surface using intense radiation. This has been applied to energetic particles induced chemical and structural modification of CNT and polymer-CNT composites [15, 16, 17]. In particular, proton irradiated graphite shows the induced magnetic ordering [18]. The proton irradiation of CNTs mostly results in morphological damage, such as welding, curves andfractionsof small pieces, and chemical modifications forming C-H bonds [15,19,20]. In this chapter, potential methods for the covalent fiinctionalization of CNTs using exposure to high energy irradiation, including accelerated protons, electrons and UV will be presented. Due to the nature of energetic radiation, some unwanted defects are also highly expected, as shown in Figure 1; a proton beam with Ε ~ keV can cause a sputtering effect while protons with an energy of -100 MeV, can be adsorbed by CNTs. Hence, the collision of energetic particles with a carbon atom will results in displacement of the atom and a number of primary knock-on atoms which leave the tube or displace other atoms in the CNT [17]. In addition, such radiation may impinge on CNT-containing devices or composites, causing undesirable effects, such as electrical interference, charging, erosion and puncture of the target surface. Since the CNTs are one of the most common fillers mixed into polymeric matrices to strengthen the durability against radiation, the question as to whether the reinforcements are fundamental to structural uniqueness of CNTs or whether they are due to specific interactions of the polymer adjacent to the CNT networks has not been answered. In order to address this issue, which is critical to understand the reinforcement mechanism, we must first investigate the radiation induced properties of CNT surface under various high energy radiations, which are comparable to the aerospace environment. Therefore, we investigated chemical and morphological changes of CNTs due to energetic particles, including protons and electrons, and UV radiation on CNTs.

In Polymer Durability and Radiation Effects; Celina, M., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2007.

Downloaded by UNIV OF CALIFORNIA SAN FRANCISCO on December 7, 2014 | http://pubs.acs.org Publication Date: December 21, 2007 | doi: 10.1021/bk-2007-0978.ch020

235

Figure 1. Charged particle energy in a space radiation environment [1] and proton radiation effects.

Experimental Radiation Sources Proton radiation was carried out using the MC-50 Cyclotron at the Korea Institute of Radiological and Medical Sciences (KIRAMS) [21]. The proton energy of the MC-50 can be tuned to 38 MeV by cyclotron adjustment and/or an Al energy degrader placed in front of the target [22]. Figure 2a shows its main components. The degrader can control the exit beam energy and the scatterer in the figure also tunes the beam fluence and flux on the target area in Figure 2a. Accelerated protons with energies of 10, 20 and 35.7 MeV were used in this experiment. Detailed conditions, such as radiation energy, dose, and fluence are listed in Table I. The electron beam were generated from the 2 MeV electron accelerator at the Korea Atomic Energy Research Institute (KAERI). Figure 2b shows a schematic of the electron accelerator. The basic components are as follows: a

In Polymer Durability and Radiation Effects; Celina, M., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2007.

Downloaded by UNIV OF CALIFORNIA SAN FRANCISCO on December 7, 2014 | http://pubs.acs.org Publication Date: December 21, 2007 | doi: 10.1021/bk-2007-0978.ch020

236 300-keV electron gun, a RF bunching cavity, two RF acceleration cavities, a 180-degree bending magnet, and a beam dump [23]. The sample stage was located in the middle between a dump and a kicker magnet. The diameter of the plate of the sample stage was about 20 cm. The nominal kinetic electron energy was 1.5 MeV - 2.0 MeV. The accelerator current was 6 A at peak and 45 mA, on average. Our experiments were performed at an energy of 2 MeV under ambient conditions. The radiation doses were 0, 60, 120, 180 and 240 kGy for each sample. A kilogray (kGy) represents the dose absorbed by the target materials, and is equal to 1 χ 10 joule/kg. All experimental conditions are listed in Table I. Samples for U V / 0 radiation were also prepared by irradiation with UV in the atmosphere for 60 minutes using a conventional U V / 0 cleaner (Model # 42220, Jelight). In order to provide a homogeneous exposure, samples were manually stirred every 10 minutes with a spatula. The UV output of the U V / 0 generator was 28,000 μWatt per cm at a wavelength of254 nm. 3

3

3

3

2

Materials The CNTs, both single-walled nanotubes (SWNT) and multi-walled nanotubes (MWNT) prepared by chemical vapor deposition (CVD) and arcdischarge (AD) methods, respectively, were purchased from Iljin Nanotech, Co., Korea. For proton and electron irradiation experiments, CNTs sheets were prepared as shown in Figure 2 by filtration of the CNT solution mixed in dimethylformamide through a cellulose membrane (pore size: 0.45 μιη). The thickness of the CNT sheets was approximately 0.5 mm, and they were 47 mm in diameter. After drying in a vacuum oven at 80 °C for 24 hours, CNT sheets (Figure 2) were obtained. These sheets were used in the radiation experiments, and were used for analysis such as SEM, Raman spectroscopy and XPS without any further treatment. For a dispersion test, a CNT powder was used instead of the CNT sheets. The morphology of the CNTs was examined by a scanning electron microscopy (SEM), transmission electron microscope (TEM), Raman spectroscopy and X-ray photoelectron spectroscopy (XPS). Micro-Raman spectra for CNTs, in powder and/or sheet were collected by a Renishaw Raman microscope (model-Inviareflex) using the 632.8 nm line of 1.5 mW HeNe laser, focused down to a diffraction limited spot size of about 1 μπι. The incident laser beam was focused onto the specimen surface through a 50 χ objective lens. XPS measurements were performed using a model Multilab ESCA 2000 (VG Co.). The spectrometer is equipped with a non-monochromatic MgKa X-ray source generating photons with energy of a 1253.6 eV. The system pressure was normally maintained below 10" Pa. A l l core level binding energies were referenced to the C l s signal at the binding energy, E , of 284.6 eV. The CNT sheets for XPS measurements were mounted on the probe tip by means of double-sided adhesive tape. 9

b

In Polymer Durability and Radiation Effects; Celina, M., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2007.

In Polymer Durability and Radiation Effects; Celina, M., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2007.

UVO

Proton

Electron

Radiation Source

28,000 (at 254 nm)

UV/O3

0,60,180 min

2

5,10,30,60 min./ 3.83xl0 P/cm /sec

10 MeV 8

60,120,180,240kGy

Radiation Conditions (time/dose /flux)

2 MeV

Radiation Energy

Generator (Jelight)

Electron Accelerator (KAERI) MC-50 Cyclotron (KIRAMS)

Facility

MWNT(CVD)/ Powder

SWNT(AD)/ Sheet

SWNT(AD)/Sheet and powder MWNT /Powder

Sample/Type

Table I. Experimental conditions for proton, electron, and UV radiation

Downloaded by UNIV OF CALIFORNIA SAN FRANCISCO on December 7, 2014 | http://pubs.acs.org Publication Date: December 21, 2007 | doi: 10.1021/bk-2007-0978.ch020

Downloaded by UNIV OF CALIFORNIA SAN FRANCISCO on December 7, 2014 | http://pubs.acs.org Publication Date: December 21, 2007 | doi: 10.1021/bk-2007-0978.ch020

238

In Polymer Durability and Radiation Effects; Celina, M., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2007.

In Polymer Durability and Radiation Effects; Celina, M., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2007.

Figure 2. (a) MC-50 Cyclotron setup at the KIRAMSfor proton radiation [19] (b) the electron beam accelerator at the KAERI for electron radiation [20].

Downloaded by UNIV OF CALIFORNIA SAN FRANCISCO on December 7, 2014 | http://pubs.acs.org Publication Date: December 21, 2007 | doi: 10.1021/bk-2007-0978.ch020

Downloaded by UNIV OF CALIFORNIA SAN FRANCISCO on December 7, 2014 | http://pubs.acs.org Publication Date: December 21, 2007 | doi: 10.1021/bk-2007-0978.ch020

240

Figure 3. SEM image of SWNT sheets for proton and electron radiation.

Surface modifications by electron, proton and UVO Irradiation Chemical changes Since protons and electrons are known to transfer their energy to target materials by direct collision, inducing excitation, ionization and knock-on collision reactions, those types of radiation are often used as a very effective method for modifying the native structures of materials. Reactions on CNTs from proton and electron particles can then be induced by the transfer of kinetic energy of the bombarding particles. In addition, very reactive atomic oxygen is generated when ambient oxygen is dissociated by irradiation at 184.9 nm and ozone at 253.7 nm, thus promoting a degree of oxidative modification of CNTs [26, 27]. Fewer investigations have been devoted to studies of principal radiation effects, which might vary, depending on the radiation sources used and the energies involved. XPS and FT-IR were used to characterize the modified CNT surfaces. The combined results provided quantitative information on the chemical composition and structure of CNTs. For XPS, when an X-ray beam is directed at the SWNT surface, the energy of the X-ray photon is absorbed by a carbon core electron. The core electron escapes from the atom if the photon energy is sufficiently large. Since CNTs are made up of a hexagonal lattice of carbon atoms analogous to the atomic planes of graphite, one can easily obtain the main peak at 285 eV from Cls. However, the raw material usually contains amorphous carbon and various

In Polymer Durability and Radiation Effects; Celina, M., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2007.

Downloaded by UNIV OF CALIFORNIA SAN FRANCISCO on December 7, 2014 | http://pubs.acs.org Publication Date: December 21, 2007 | doi: 10.1021/bk-2007-0978.ch020

241 oxygen containing groups created at the sidewalls, depending on the synthetic method and degree offiinctionalizationon its surface. Consequently, carbonbased contaminants will appear collectively as several peaks around the binding energy of Cls. Therefore, we focused the peaks near 285 eV to determine whether the proton and electron beam, or UV radiation caused any compositional modification. Figure 4a is a characteristic XPS profile showing Cls binding energy spectra of pristine SWNTs. The carbon peak was fitted by several carbon-based surface functional groups, which are indexed as / , 2, 2\ 2" and 3 in the figure. The spectra in Figure 4a clearly indicates that there is a narrow peak (/) at 285.6 eV of Cls atoms where SP (C=C) and SP (C-C) are composition sensitive. For example, the 284.4 eV of C=C and the 285.11 ~ 285.5 eV of C-C are characteristic of highly ordered lattice conformations. When an oxygen molecule is adsorbed to carbon atoms, the binding energy can be shifted higher due to the higher electronegativity of the oxygen [25, 26]. Therefore, the peaks (2, 2' and 2") with higher binding energies at 286.4, 288.6 290.5 eV were assigned; C-O, C=0 and COO, respectively, indicating that the carbon nanotubes are partially oxidized for pristine SWNTs. This indicates that the SWNTs used in our experiments contain various oxygen containing groups which were probably produced during their synthesis or purification during the manufacturing process. The origin of peak (3) near 284 eV, clearly separated from the main peak (i) at 285.6 eV, is not clearly understood, but probably arises from amorphous carbon regions, existing as carbon impurities together with the SWNTs. We cannot rule out that the origin of the peak might be due to the static charging effect, since the sample was mounted on the probe tip by means of insulating double-sided adhesive tape, which might result in a differential static charging at low binding energy [28]. Since this peak was only observed at the pristine sample and the sample exposed to proton radiation (not at the sample exposed to electron radiation) we believe that this peak is due to the nature of a chemical information of SWNTs, not a peak broadeningfroma experimental artifact. We, subsequently compared the results for the same SWNTs irradiated at 2 MeV with electrons and those irradiated at 10 MeV with protons in ambient conditions. Figure 4b shows the XPS spectrum of the electron beam irradiated SWNTs. The radiation dose for this sample was 60 kGy. As shown in thefigure,the small peak (5 in Figure 4a) near 284 eV has essentially disappeared, indicative of the removal of carbon impurities. The peak from oxidized carbon derivatives (2), which may include oxygen containing carbons, remained broad, so a proper deconvolution process could not be performed here. However, the relative concentration of peak 2, obtained from the peak area, decreased upon electron radiation, implying that any significant oxidation did not occur while the carbon impurities were selectively removed by electron radiation. Overall, the spectrum 2

3

In Polymer Durability and Radiation Effects; Celina, M., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2007.

Downloaded by UNIV OF CALIFORNIA SAN FRANCISCO on December 7, 2014 | http://pubs.acs.org Publication Date: December 21, 2007 | doi: 10.1021/bk-2007-0978.ch020

242

shows a main peak from Cls carbon (1), which was similar to that in Figure 4a, without any distinct shoulder peaks nearby. Surprisingly, Figure 4c indicates that a very different reaction due to the collision of protons occurred. Notably, the small peak (5 in Figure 4a) near 284 eV increased significantly while the intensity from the main peak (i) is reduced. The peaks (2, 2\ 2" and 4) with higher binding energies at 286 - 290 eV were remained and can be assigned to a variety of oxidative derivatives of carbons. Again, the origin of this peak (5) is not fully understood and requires further investigation. We, however, postulate that the 10 MeV proton particles generated new sp bonds on the CNT surfaces, corresponding to oxygen-based functional groups such as carbonyl and phenyl groups (which are responsible for the peaks at the higher binding energy region) and preserved the amorphous region of carbon impurities. Consequently, proton collision ruptured C=C bonds around the SWNT, resulting in a very reactive surface which can be oxidized under ambient conditions. We further confirmed that the proton irradiated SWNTs showed an improved solubility in various polar solvents such as DMF, and the solution was very stable even after a month, while raw SWNTs precipitated after only a few minutes. Recently, ultraviolet (UV) treatment has been suggested to modify the CNT surface [24,27,28]. In details, it has been theoretically shown that UV light excites ambient oxygen molecules to a spin-singlet state, resulting in a significant reduction in the activation energy for oxygen molecule chemisorption [24]. Cai et al. have shown that UV-induced atomic oxygens facilitate the surface oxidation of SWNTs [27]. We, recently demonstrated that UV-ozone treatment under ambient conditions can have a dramatic effect on the nature of the surface oxidation of MWNTs, leading to the production of carboxylic acid, phenol, quinine, ester, pyrone, and ketone functional groups [28], as similar to the SWNTs irradiated by protons. Based on the theoretical calculation [24] together with the experimental results, the oxidation mechanism is suggested as follows; the reduction in the activation energy for 0 molecule chemisorption onto the CNT surface increases the rate of nanotubes adsorption. Simultaneously, reactive atomic oxygens, which are dissociated from bonded oxygen (physisorbed and/or chemisorbed oxygens onto CNT surface) molecules by UV radiation, can induce oxidative cleavage of the carbon bonding to form oxidative functional groups more effectively, as shown in Figure 5. Fourier-transform infrared spectroscopy (FT-IR) was carried out to determine the chemical structures and types of functional groups present on MWNTs. A comparison of 60-minute FT-IR spectra of raw and UVO treated CNTs are shown in Figure 6. Several peaks corresponding to bond rotation, vibration, and stretching of various oxygen compounds can be observed. The peaks at 1074 cm" , 1170 cm' , 1840 cm" indicate the presence of ester groups, and the peaks at 1524 cm* , 1636 cm" , 3442 cm" are assigned to the stretching 3

2

1

1

1

1

1

1

In Polymer Durability and Radiation Effects; Celina, M., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2007.

In Polymer Durability and Radiation Effects; Celina, M., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2007.

Figure 4. XPS spectra of irradiated SWNT: (a) before, (b) after 2 Mev, electron beam radiation dose of 60 kGy and (c) after 10 Mev, proton beam radiation for 15 minutes.

Downloaded by UNIV OF CALIFORNIA SAN FRANCISCO on December 7, 2014 | http://pubs.acs.org Publication Date: December 21, 2007 | doi: 10.1021/bk-2007-0978.ch020

244

0 + ho: 184.9 nm 2

0 + hu : 253.7 nm Downloaded by UNIV OF CALIFORNIA SAN FRANCISCO on December 7, 2014 | http://pubs.acs.org Publication Date: December 21, 2007 | doi: 10.1021/bk-2007-0978.ch020

3

[O] + 0 [O

3

Figure 5. An atomic oxygen is generated when molecular oxygen is dissociated by 184.9 nm and ozone by 253.7 nm, and produces oxygen compounds on the CNT surface.

mode of C=C, C=0 of a quinone group, and -OH of a hydroxyl group, respectively [29,30]. Accordingly, the results of Figure 6 illustrate that the UVO treatment of CNTs produced a number of functional groups containing oxygen. Compared to a raw CNT, the dissolution in polar organic solvents (DMF, 1,2dichlorobenzene and pyridine) was improved by as much as 320%, which is comparable to vigorous acid treatment, without any severe alteration of the aspect ratio of CNTs. In summary, depending on the nature of sources and conditions, the radiation effects of the CNTs in a relatively short period of time were very different. 1) The effective surface oxidation was observed for the samples exposed to the proton for 15 min and UV-ozone environment in 60 min, and the oxidative functional groups produced by proton and UV-ozone were essentially similar to those produced by acid-based treatment.[31] 2) The electron radiation could remove the carbon impurities, when the operating conditions are carefully optimized. Nevertheless, these radiation methods described above were performed in a form of a dried CNT powder, thus did not produce any harmful chemical byproducts.

Structural Changes The remarkable structures and properties of carbon nanotubes were reviewed above. Among these, the electrical properties of CNTs have attracted special interest. Depending on the geometrical arrangement of the carbon rings of graphite, CNTs can have either metallic or semiconducting properties. Since electronic properties are strongly related to molecular arrangements, for example, bandgap is inversely related to molecular diameter, the lattice information concerning CNTs should be discussed, as to whether radiation might influence their geometrical structures. Raman spectroscopy normally detects vibration frequencies. Under resonant conditions, the spectra not only provide important

In Polymer Durability and Radiation Effects; Celina, M., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2007.

Downloaded by UNIV OF CALIFORNIA SAN FRANCISCO on December 7, 2014 | http://pubs.acs.org Publication Date: December 21, 2007 | doi: 10.1021/bk-2007-0978.ch020

245

Figure 6. FT-IR spectra obtainedfrom raw MWNT (dotted line) and UVO treated MWNT (line) for 60 min. (Reprintedfrom Colloids and Surface A: Physicochemical and Engineering Aspects, Najafi et al 384-385, 373, Copyright(2006), with permission from Elsevier, [29]) t

information concerning the electronic structure through the strong coupling between electrons and lattice vibrations, but also for this one-dimensional system. In particular, Raman spectra provide information on four important characteristics: radial breathing mode (RBM) associated with the symmetric movement of all carbon atoms in the radial direction, disorder Induced mode (Dband), tangential modes (G-band), second order mode (G'-band). We first focused on the RBM sections in the 160-300 cm" region, where the frequency (