Irradiation studies on carbon nanotube-reinforced boron carbide

Nuclear Instruments and Methods in Physics Research B 272 (2012) 249–252

Contents lists available at ScienceDirect

Nuclear Instruments and Methods in Physics Research B journal homepage: www.elsevier.com/locate/nimb

Irradiation studies on carbon nanotube-reinforced boron carbide Assel Aitkaliyeva a, Michael C. McCarthy b, Hae-Kwon Jeong b, Lin Shao a,c,⇑ a

Department of Materials Science and Engineering, Texas A&M University, College Station, TX 77843, USA Artie McFerrin Department of Chemical Engineering, Texas A&M University, College Station, TX 77843, USA c Department of Nuclear Engineering, Texas A&M University, College Station, TX 77843, USA b

a r t i c l e

i n f o

Article history: Available online 2 February 2011 Keywords: Ion irradiation Carbon nanotubes Ceramics Boron carbide

a b s t r a c t Radiation response of carbon nanotube (CNT) reinforced boron carbide composite has been studied for its application as a structural component in nuclear engineering. The composite was bombarded by 140 keV He ions at room temperature to a fluence ranging from 1  1014 to 1  1017 cm 2. Two-dimensional Raman mapping shows inhomogeneous distribution of CNTs, and was used to select regions of interest for damage characterization. For CNTs, the intensities ratio of D–G bands (ID/IG) increased with fluence up to a certain value, and decreased at the fluence of 5  1016 cm 2. This fluence also corresponds to a trend break in the plot of FWHM (full width at half maximum) of G band vs. ID/IG ratio, which indicates amorphization of CNTs. The study shows that Raman spectroscopy is a powerful tool to quantitatively characterize radiation damage in CNT-reinforced composites. Ó 2011 Elsevier B.V. All rights reserved.

1. Introduction Boron carbide (BC) possesses unique properties such as low specific weight, high elastic modulus, high hardness, and excellent physical and chemical stabilities at high temperatures [1–3]. BC has been used as a neutron absorber in certain types of fission reactors and has been tested for its applications as core components in some fast breeder reactors [4]. But its application is limited by poor ductility. Incorporating carbon nanotubes (CNTs) into BC matrix can enhance materials’ mechanical properties, and increase its resistance to thermal shock and crack growth. However, in order to use CNT-reinforced BC composites as structural components in fission and fusion reactors, material’s radiation tolerance needs to be studied. Neutron interactions with boron atoms lead to creation of energetic particles through the transmutation reaction 10 B(n, a)7Li. Efforts have been made to understand radiation induced structural changes under a high burnup [4,5]. But little is known about radiation responses of CNTs. Similar to other nanomaterials, characterization of damage buildup in CNT-based materials is challenging. Although high resolution transmission electron microscopy, in conjunction with modeling, is able to identify stable defect configurations in CNTs, it does not provide quantitative information [6,7]. Interpretation of electrical resistivity measurements, on the other hand, is difficult because of the mixed semiconducting and metallic properties of CNTs and possible irra⇑ Corresponding author at: Department of Materials Science and Engineering, Texas A&M University, College Station, TX 77843, USA. E-mail addresses: [email protected], [email protected] (L. Shao). 0168-583X/$ - see front matter Ó 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.nimb.2011.01.076

diation induced transitions. In this study, Raman spectroscopy was used to quantitatively characterize radiation damage.

2. Experimental procedure CNT-reinforced BC composite (Nano-Lab Inc., Waltham, MA) was fabricated by hot pressing a mixture of CNTs and BC powders at 2000 °C under 10 ton load, and pressureless sintering at 2000 °C for 2 h. The composite contained about 3–5% of hollow-structured multi-walled carbon nanotubes (MWNTs) with diameter ranging from 15 to 45 nm and average length of 10 lm. MWNTS were produced by standard chemical vapor deposition process, description of which can be found elsewhere [8]. Composite specimens, with a dimension of 4  3  0.5 mm, were irradiated at room temperature with 140 keV He+ ion beam to a fluence ranging from 1  1014 to 1  1017 cm 2. Irradiated samples were characterized by using techniques such as Raman spectroscopy and transmission electron microscopy (TEM). Raman spectra were excited with He–Ne (633 nm) laser line, and measured in backscattering geometry by using Horiba Jobin–Yvon LabRam IR system with spectral resolution of 1.3 cm 1. The laser power was kept constant to prevent sample heating, and a spatial resolution of 20 lm was achieved using a microscope with a 50 objective lens. All peak parameters were acquired using fits to the Lorentzian line shapes in LabRam Analysis Software. Plain view TEM specimens were examined in a JEOL JEM-2010 electron microscope, equipped with Gatan SC1000 ORIUS CCD camera, operated at an accelerated voltage of 200 kV.

250

A. Aitkaliyeva et al. / Nuclear Instruments and Methods in Physics Research B 272 (2012) 249–252

3. Results and discussion Fig. 1 shows a high resolution transmission electron micrograph of unirradiated composite. The boron carbide–carbon nanotube interface, marked by an arrow, suggests a good adhesion between CNTs and ceramic matrix. TEM examination over a large area suggested clustering of CNTs in the form of networking in certain regions, and branching of small CNTs from large nanotubes, which might be caused by the sintering process. Fig. 2 shows Raman mapping of the unirradiated CNT–BC composite. Fig. 2(a) is an optical microscope image of the specimen surface, with the box referring to the mapping domain in which Raman spectra were acquired. The two-dimensional Raman map,

Fig. 1. Transmission electron micrograph of unirradiated CNT-reinforced BC composite. The white arrow marks the BC–CNT interface.

obtained by scanning over the box area, is shown in Fig. 2(b). The green, red and blue colors present the integrated signals over different regions of Raman shifts. Fig. 2(c) shows a threedimensional plot of the mapping region. This 3D map contains two spatial (X and Y coordinates) and one spectral dimensions (intensities of the bands). The pattern observed in an optical image is consistent with the one observed in Raman mapping, which suggests that dispersion of CNTs is not uniform. To determine uniformity of CNTs distribution, Raman spectra were collected from three typical regions marked as 1, 2 and 3, and compared. Raman spectra corresponding to three regions marked in Fig. 2(a) are shown in Fig. 3(a)–(c). Raman spectrum of spot located in the dark region (region 1) of the optical image contains bands positioned at 1346, 1593 and 2664 cm 1. These bands are characteristic to CNTs, which implies that the region corresponds to CNTs enriched area. The band at 1346 cm 1, known as D band, is related to the scattering caused by symmetry-breaking defects, and is typically absent in defect free carbon materials [9]. G mode at 1593 cm 1 is assigned to zone center phonons of E2g symmetry and is known as an intrinsic vibration feature of sp2 carbon sheet [9]. Even though G mode is typically referred to as Raman-allowed C-point mode, latest studies indicate that its origin is also defectinduced and double-resonant [10]. The band at 2664 cm 1 refers to as the D⁄ band, and is caused by double order resonant Raman scattering process. However, its relevance to defects is still debatable [11,12]. Fig. 3(b) shows the spectrum collected in region 2. CNT’s D and G modes are superimposed on a broad spectrum, with a peak at 1090 cm 1. This peak is attributed to BC [13]. The spectrum acquired in region 3 is given in Fig. 3(c). The spectrum has a sharp peak at 1090 cm 1 and does not show any distinctive CNTs bands. This supports our observation that region 3 in Fig. 2(a) corresponds to BC, region 1 to CNTs, and region 2 to the boundary of BC and CNT enriched areas.

Fig. 2. (a) An optical image of unirradiated CNT-reinforced composite; (b) 2-D Raman map and (c) 3-D Raman map of the region marked by the box.

251

A. Aitkaliyeva et al. / Nuclear Instruments and Methods in Physics Research B 272 (2012) 249–252

Region 1

Intensity (arb. units)

1346.1 1593.6

(a)

140 keV He

Intensity (Arb. Units)

(a)

2664

16

2

15

2

14

2

1x10 /cm

1x10 /cm

1x10 /cm

Unirradiated

1000

1500

2000

2500

1000

3000

1500

1342.05

Intensity (arb. units)

1591.55

Region 2

1090.5

1000

1500

2000

2500

(b) 16

2

15

2

1x10 /cm

Unirradiated

1500

2000

Fig. 4. Raman spectra of selected area corresponding to (a) region 1 (CNT enriched) and (b) region 3 (BC matrix), as a function of 140 keV He ion fluence.

17

-2

140 keV He

2500

3000

-1

FWHM of G mode

1x10 cm 16 -2 5x10 cm

2000

3000

-1

Region 3

1500

2500

Raman Shift (cm )

Intensity (arb. units) 1000

2

1x10 /cm

1000

3000

17

1x10 /cm

Raman Shift (cm ) 1090.5

3000

140 keV He

-1

(c)

2500

Raman Shift (cm )

Intensity (Arb. Units)

(b)

2000

-1

-1

Raman Shift (cm )

20 16

10 15

-2

5x10 cm 15

Raman Shift (cm )

14

-2

1x10 cm

-2

1x10 cm

0

Fig. 3. Raman spectra of selected area corresponding to (a) region 1 (CNT enriched), (b) region 2 (boundary), and (c) region 3 (BC matrix), as marked in Fig. 2(a).

1.0

1.2

1.4

ID/IG ratio Fig. 5. A plot of FWHM of G band vs ID/IG ratio of 140 keV He ion-irradiated composite. The arrows indicate increasing fluence.

140 keV He

1.3

ID/IG ratio

Fig. 4(a) and (b) shows the Raman spectra from the regions 1 and 3, respectively, for different ion irradiation fluences. For CNTs, the relative intensity of D mode increases with increasing fluence, and both D and G bands broaden and coalesce, which results in merging of two bands into one large band at high fluence. The ratio of these two modes (ID/IG ratio) is typically used to estimate degree of disorder in graphite and can be used to determine amount of damage induced by irradiation in CNTs [14]. For BC, intensity of the band at 1090 cm 1 decreases with increasing fluence. The significant broadening of this band can be an indication of a partially distorted or amorphous structure. The nature of the band observed at 1470 cm 1 is unclear. It can be attributed to carbon disorder. Fig. 5 plots the full width at half maximum (FWHM) of G band vs. ID/IG ratio for irradiated samples. The arrows indicate increasing ion fluence. From 1  1014 to 1  1016 cm 2, FWHM increases with ID/IG ratio. But above 1  1016 cm 2, the trend changes and a higher FWHM value starts corresponding to lower ID/IG ratio. Similar trend was previously reported in CNTs and in graphite [15,16]. Fig. 6 plots the ID/IG ratio as a function of irradiation fluence. The ID/IG ratio increases with increasing fluence up to 1  1016 cm 2, after which the ratio decreases with fluence.

-2

1x10 cm

1.2

1.1

1.0 0.0

16

3.0x10

16

6.0x10

16

9.0x10

-2

Fluence (cm ) Fig. 6. The intensity ratio of ID/IG as a function of fluence in CNT enriched regions of irradiated composite.

The trend change observed in Figs. 5 and 6 represents an onset of crystalline to amorphous transition. There are several existing

252

A. Aitkaliyeva et al. / Nuclear Instruments and Methods in Physics Research B 272 (2012) 249–252

theories explaining increase of ID/IG ratio upon irradiation: (i) increased density of defects and associated reduction of phonon correlation length; and (ii) ion irradiation induced sp2 ? sp3 transition [10,14,17]. This sp2 ? sp3 transition is promoted by atomic displacements, production of interstitial/vacancy pairs upon momentum transfer from the incident particles, and by the inter-wall coupling from energy deposition [15,18,19]. Initial increase of ID/IG ratio in Fig. 6 can be attributed to the increase of disorder in the structure, while decrease at higher fluence means significantly enhanced vibration of modes over a larger frequency range, which occurs when materials’ structure collapses. Our previous study on carbon buckypapers (CBPs) suggested that amorphization threshold fluence is in the range of 1  1015–1  1016 cm 2, which is consistent with the present study [20]. Such consistency suggests that sintering and dispersion of CNTs does not influence their radiation tolerance. The superior properties of CNTs degrade with crystalline-to-amorphous transition. For application of CNT-reinforced composites, including but not limited to BC, radiation induced degradation of CNTs should be taken into consideration. The present study shows that Raman spectroscopy is a powerful tool for quantitative analysis of the damage buildup in CNTs and CNT-based materials. Raman mapping is valuable essential in identifying regions of interests and determining uniformity and integrity of the CNT-toughened composites. It can be used to study radiation response of CNTs, which is expected to be different from their bulk counterparts. Recently Kumar et al. reported the irradiation response of CNTs under swift ion irradiation [21]. Raman analysis was used to characterize radiation damage. The reported ID/IG ratio increases with increasing ion fluence, which suggests that disorder level was not high enough to cause structural collapse. This is consistent with the ultralow fluences used in their studies. In summary, we have used Raman spectroscopy to characterize radiation damage induced by 140 keV He ions in CNT-reinforced BC composite. CNT enriched regions in the material were detected by Raman mapping. This technique provides quantitative analysis

of disorder and can be used to determine threshold fluence value for amorphization. Acknowledgements The study was supported by National Science Foundation (USA) through grant no. 0846835. The Raman acquisition was supported by the NSF under grant no. BES-0421409. Use of the TAMU Materials Characterization Facility and Microscopy and Imaging Center is acknowledged. References [1] F. Theévenot, J. Eur. Ceram. Soc. 6 (1990) 205. [2] M.W. Chen, J.W. McCauley, J.C. LaSalvia, K.J. Hemker, J. Am. Chem. Soc. 88 (2005) 1935. [3] A. Lipp, Tech. Rundschau 14, 28 (1965) 33. [4] A. Jostsons, C.K.H. Dubose, J. Nucl. Mater. 44 (1972) 91. [5] G.L. Copeland, C.K.H. Dubose, R.G. Donnelly, W.R. Martin, J. Nucl. Mater. 43 (1972) 126. [6] F. Banhart, Rep. Prog. Phys. 62 (1999) 1181. [7] A.V. Krasheninnikov, K. Nordlund, Nucl. Instr. Meth. Phys. Res. B 216 (2004) 355. [8] R. Brukh, S. Mitra, Chem. Phys. Lett. 424 (2006) 126. [9] A.C. Ferrari, J. Robertson, Phys. Rev. B 61 (2000) 20. [10] G. Compagnini, G.A. Baratta, R.S. Cataliotti, A. Morresi, J. Raman Spectrosc. 26 (2005) 917. [11] J. Maultzsch, S. Reich, C. Thomsen, S. Webster, R. Czerw, D.L. Carroll, S.M.C. Vieira, P.R. Birkett, C.A. Rego, Appl. Phys. Lett. 81 (2002) 14. [12] V. Skákalová, J. Maultzsch, Z. Osváth, L.P. Biró, S. Roth, Phys. State. Sol. 1 (2007) 138. [13] U. Kuhlmann, H. Werheit, J. Alloys Compd. 205 (1994) 87. [14] F. Tuinstra, J.L. Koenig, J. Chem. Phys. 53 (1970) 1126. [15] T. Tanabe, Phys. Scr. T64 (1996) 7. [16] K. Niwase, Phys. Rev. B 52 (1995) 15785. [17] K. Nakamura, M. Kitajima, Phys. Rev. B 45 (1992) 78. [18] J. Seldin, C.W. Nezbeda, J. Appl. Phys. 41 (1970) 3389. [19] E. Salonen, A.V. Krasheninnikov, K. Nordlund, Nucl. Instr. Meth. Phys. Res. B 193 (2002) 603. [20] A. Aitkaliyeva, M.C. McCarthy, M. Martin, E.G. Fu, D. Wijesundera, X. Wang, W.-K. Chu, H.-K. Jeong, L. Shao, Nucl. Instr. Meth. Phys. Res. B 267 (2009) 3443. [21] A. Kumar, D.K. Avasthi, J.C. Pivin, P.M. Koinkar, Appl. Phys. Lett. 92 (2008) 221904.