Indian Journal of Pure & Applied Physics Vol. 44, February 2006, pp. 119-124

Transmission electron microscopy of nanomaterials S Neogy, R T Savalia, R Tewari, D Srivastava & G K Dey Materials Science Division, Bhabha Atomic Research Centre, Mumbai 400 085 Received 25 May 2005; accepted 5 December 2005 Transmission electron microscope (TEM) has emerged as a very powerful tool for probing the structure of metals and alloys because of its capability to provide morphological information, crystallographic details, and chemical composition of phases distributed on a very fine scale in a given microstructure. Convergent beam electron diffraction, nano or micro beam diffraction and selected area diffraction techniques have been used to obtain crystallographic information whereas structural information is obtained by high-resolution electron microscopy (HREM). With its multifaceted capabilities such as nanobeam diffraction, composition analysis and imaging abilities at angstrom level, TEM has emerged as an instrument for complete characterization of nano scale microstructure of materials. The use of TEM in the study of various types of nanomaterials is described in this paper. Keywords: Transmission electron microscope, Nanomaterials, High resolution electron microscopy, Microstructural characterization IPC Code: GO1N

1 Introduction Microstructural characterization, on nanometer scale, has become very important for all types of materials in recent times. This microstructural information is required for structure property correlation and for carrying out basic and applied research in materials. Microstructural characterization broadly includes ascertaining the morphology of phases, number of phases, structure of phases, identification of the crystallographic defects and composition of the phases. Transmission electron microscope (TEM) is a tool, which can provide all these aforementioned information from the same region at nanometer level. The resolution of an electron microscope depends not only on the wavelength of the electron beam, but is also limited by aberrations of the image forming lenses present in the electron microscope. In this regard, the spherical aberration of the objective lens, which is the first image forming lens of an electron microscope, is most important. In modern day microscopes, great emphasis is laid on minimizing the spherical aberration coefficient (Cs) of the objective lens so as to improve the resolution of the instrument as much as possible. Point to point resolution of 0.2 nm is very common in most modern TEMs. With the advent of Cs corrected TEMs where the Cs value is drastically reduced by incorporation of beam correcting coils, subangstrom microscopy has become a reality1. The combination of very high-resolution imaging capability along with nanobeam diffraction

and composition analysis ability has made the modern TEM a device for complete microstructural characterization. 2 Examination of Nanostructures in the TEM Nanostructures have been characterized by a variety of techniques. Many of these techniques involve imaging of the nanostructures. However, very few of these techniques as a single tool, can give structural information (atomic arrangement in the nanostructure), composition information and information of the size and the shape of the nanostructure. With its multifaceted capabilities such as nanobeam diffraction and composition analysis and imaging abilities at angstrom level, TEM has emerged as the only tool, which can be used for complete characterization of nanostructured materials. Besides the aforementioned information, it can reveal the nature of crystallographic defects. High resolution imaging (HREM) in the TEM has evolved as a very powerful tool for probing the structure of the phases and the defects. The multilayered structures of contemporaneous interest for their desirable properties from scientific as well technological viewpoints are being examined in considerable detail using this form of microscopy. HREM has been particularly useful in this area because of its ability to resolve the interface structures in these materials. Plethora of scientific literature bears testimony to the fact that nano-dimension can be given not only to

120

INDIAN J PURE & APPL PHYS, VOL 44, FEBRUARY 2006

grains in a crystalline material but also to a variety of other geometries such as tubes, pores and layers. Nano-dimensional tubes or nanotubes have been the subject matter of many investigations because of scientific as well as technological interests. Similarly nanoporous materials and nanolayered materials representing porous materials and multilayered materials having nano-dimensions respectively have also attracted considerable attention because of their scientific and technological importance. TEM has been used to characterize all the aforementioned types of materials to different extents. 2.1 Study of nanotubes

Carbon nanotubes were discovered by electron microscopy in the carbon shoot produced in an electric arc between graphite electrodes, as used in the production of fullerenes2. TEM are frequently used to characterize the diameter of the tubes, their shape and size and also the nature of the nanotubes i.e., whether these are single walled or multiwalled2. Fig. 1 shows a typical micrograph of nanotubes produced by the arc discharge process which gives rise to mostly multiwalled nanotubes. Tubes of various wall thicknesses and overall diameter can be seen in Fig. 1. Such micrographs can be used to ascertain the actual structure of the nanotubes and phenomena like branching of nanotubes. Effect of external parameters like pressure on the shape of nanotubes has been investigated2 by TEM. HREM images of nanotubes have revealed the geometry of individual graphene sheets and their defects in multishell tubes. Nanotubes filled with other elements like Ni and Fe have also

Fig. 1—Bright field TEM micrograph showing multiwalled carbon nanotubes produced by arc discharge process

been examined by TEM. The objective of these studies has been to ascertain the reaction between the tube wall and the element, which has been filled, and also to see the changes occurring in the lattice of element once these have been filled inside the nanotubes. 2.2 Nanoporous material

Porous materials with regularly arranged, uniform mesopores (2 to 50 nm in diameter) are called mesoporous material. Because of their large surface areas, these are useful as adsorbents or catalysts. Activated carbon used as deodorizer and silica gel used as desiccant are well known examples of porous materials. One type of mesoporous materials is the silicon based mesoporous molecular sieves of MCM 41 type having a hexagonal uni-dimensional pore structure whereas in MCM 48 a threedimensional cubic structure is seen3-6. The size and shape of the pores and their arrangement have all been studied in great detail by conventional TEM and HREM using imaging and diffraction. Fig. 2 shows the image of a mesoporous MCM41 material showing parallel arrangement of long pores in the size range of 2.5-3 nm. If this material is examined under microscope at lower magnification, it is not possible to decipher these pores as material appears as a continuum body. High resolution image coupled with nano-diffraction would be able to reveal the true arrangement of the pores.

Fig. 2—HREM micrograph of a mesoporous MCM41 material showing parallel arrangement of long pores in the size range of 2.5-3nm

NEOGY et al.: TRANSMISSION ELECTRON MICROSCOPY OF NANOMATERIALS

Besides the empty structures, the pores could also be filled by different types of species and it is possible to locate these species inside the pores. In one such study, it could be seen that the specimen particles comprised sheet like structures which were more or less rounded3. HREM images taken along the various directions showed that along the [100] direction the pores were found to be arranged in a square grid and in the [111] and the [110] directions the pore arrangements were hexagonal and distorted hexagonal respectively. Substantial distortion and deformation could be seen in the structure of the pores after uranium loading, the extent of change being governed by the amount of uranium put in the pores. Such distortions and deformations as seen in the TEM were in agreement with the small angle X-Ray scattering (SAXS) results on the specimens4-7. Mesoporous structures of the aforementioned type are specially challenging for an electron microscopist. In these materials, the walls of the pores are made of amorphous or noncrystalline material. The pores are, however, periodically arranged. There are diffraction effects, therefore, from the amorphous wall as well as due to the periodic arrangement of the pores. Although the amorphous structure of the walls will give typical diffraction pattern comprising a broad halo, sharp spot patterns can be obtained due to the periodic arrangement of the pore structure. The latter is in such a size range, which will need longer camera lengths for distinct observation of spots, whereas the former will need shorter camera lengths for viewing. Besides diffraction, phase contrast imaging can also be done at two levels. When conventional bright field imaging is done in addition to the transmitted beam, the diffracted beams due to the periodic pore structure which are very close to the transmitted beam also get included in the objective aperture and the image has phase contrast revealing the arrangement of the pores. If a larger objective aperture is used to include the diffraction halo from the amorphous structure, the amorphous structure of the pore walls can be imaged. Besides the aforementioned type of electron microscopy, three-dimensional transmission electron microscopy (3D-TEM) has also been applied for studying this kind of material. Electron tomography consists of collecting many TEM images of an object over a large angular tilt range with a small angular increment (tilt series), after which a 3D reconstruction is carried out from these images. It is shown that this technique can give information with nanometer scale resolution in three dimensions. In a recent study, the

121

3D reconstruction of 5 nm gold particles and 2-3 nm ZrO2 particles, which were deposited inside the mesopores of SBA-15, could be visualized8 . However, the 3D reconstruction has to be interpreted with care because the occurrence of diffraction contrast in the TEM images creates artifacts in the reconstruction8 . 2.3 Nanocrystalline material

Nanocrystalline materials have generated a lot of interest in recent times because of their many attractive properties. For example, nanocrystalline materials due to large grain boundary area in comparison to the volume of the grain exhibit excellent mechanical properties. Besides studies on the structure of the nanocrystals, issues related with the deformation mechanism and nature of grain boundaries in these nanocrystalline materials are also being studied in detail. Out of the several techniques for synthesis of nanocrystals, the production of nanocrystals by transformation of a liquid phase or an amorphous phase to a nanocrystalline phase holds special promise. The production of nanocrystals by crystallization of the amorphous phase is especially attractive as in this case the control of the crystal size is more accurate. With the advent of bulk metallic glasses, it is now possible to produce nanocrystals in bulk and composites of nanocrystals and amorphous phase. A composite material comprising uniform distribution of the nanocrystalline phase in the amorphous matrix offers combination of attractive properties like, high strength, good toughness, etc. In the following sub-sections, the characterization of nanocrystals produced from crystallization of metallic glass by TEM is described. 2.3.1 Crystallization of bulk metallic glass

Figure 3 shows a typical high resolution image of a nanocrystalline material. Typical grain size, shown in the micrograph, is of the order of 30-50 nm. Few points could be noticed from Fig. 3: (i) grain boundaries are devoid of amorphous regions; (ii) the lattice fringes have been found to be distorted and curved in some of the nanograins which was manifested in the form of a slight change in contrast of the lattice fringes near the nanograin boundary and the change in contrast was localized to a very narrow region of approximately 0.5nm along the grain boundary; (iii) lattice could be resolved simultaneously in all the grains in view indicating that grain boundary is parallel to low index planes9;

122

INDIAN J PURE & APPL PHYS, VOL 44, FEBRUARY 2006

Fig. 3—HREM micrograph showing many nano-grain boundaries, stacking fault (indicated by single arrow) and antiphase domain boundary (indicated by double arrow) in crystallized Zr52Ti6Al10Cu18Ni14 bulk metallic glass

(iv) presence of crystallographic defects in nano-grain could be noticed (marked by single and double arrows in Fig. 3). After tilting and imaging these nanograins, the presence of any amorphous region at the grain boundaries was not noticed (Fig. 3). Such an experiment demonstrated that nature of grain boundaries in the nanocrystalline materials is similar to the grain boundaries observed in the crystalline phase. This, in other wards, means that the postulations about the different nature of the nanocrystalline grain boundary are not strictly valid, at least in the present case. Thomas et al.10 reported similar kind of observations in case of nanocrystalline Pd where manifestation of grain boundary structures with random displacements of average magnitude greater than 12 % of the nearest neighbour distance could not be observed. The nanograin boundary was mostly a large angle grain boundary and no coherency could be observed between the planes along any orientation. There was no evidence of voids at the nanograin boundaries, whereas, dislocations could be seen in the specimen, although, their number density was quite small. Subdivision of the nanograins into smaller grains by small angle boundaries could also be noticed (Fig. 4). A regular array of edge dislocations with a periodic distance of about 3 nm and localized strain contrast around the dislocation cores could also be noticed. HREM observation also revealed the presence of twins in many of the nanograins.

Fig. 4—HREM micrograph showing the structure of a nano-grain in crystallized Zr52Ti6Al10Cu18Ni14 bulk metallic glass. A low angle grain boundary within the nanograin could be seen which has been marked by an arrow

Although the structure of the twins observed inside the nanograins was found to be identical to those seen in the case of large grained materials, the propensity of these twins was found to be much larger. Twintwin interaction was also noticed in many cases. These observations are similar to those reported in studies carried out in many other nanocrystalline solids where extensive twinning has been found to occur11. Twins in the nanocrystalline phase is mostly to accommodate the large stress present in them due to their structure and also the growth of the nanograins necessities the formation of twins. Hence, because of coarsening, as the grain size increased the propensity of twins was found to come down since the need for accommodating the stresses generated also came down substantially. 2.3.2 Formation of nano-quasicrystals

In many of the amorphous alloys quasi-crystals could be obtained by crystallization. A large number of alloy systems have now been identified where the quasi-crystalline phase can be obtained in the nanomorphology by crystallization of an amorphous phase. This phenomenon is commonly known as nanoquasicrystallization. Many of the Zr based alloys systems where nano-quasicrystalline phases13 have been found are Zr-Pd (Ref.12), Zr-Ni-Ti and other Ti/Zr-TM (Ref.14) alloys besides multicomponent ZrCu-Ni-Al alloys (ref. 15). Out of these alloys, Zr69.5Cu12Ni11Al7.5 alloy is now very well known for its quasicrystal formation tendency after rapid solidification and crystallization and has also received considerable attention through studies pertaining to

NEOGY et al.: TRANSMISSION ELECTRON MICROSCOPY OF NANOMATERIALS

123

One of the interesting variants of multilayers from electronic materials are quantum wells. The

characterization of such structures has been carried out by a variety of routes including conventional TEM and HREM. In these multilayers, the parameters of interest are morphology, layer thickness, crystallinity, interfaces and defect structures. Conventional TEM is, generally, used for ascertaining the thickness of the multilayers and in judging the uniformity of these layers whereas HREM is used for looking at the structure of the interfaces at atomic level. Figure 6 is a bright field TEM micrograph showing the cross-section of a multi quantum wells (MQW) structure of InP/InGaAs made by MOVPE. The specimens for such examinations have been prepared by the cross-sectional specimen preparation technique17 with four InGaAs quantum wells of nominal widths 15, 7.8, 3.9 and 2.5 nm in order of growth on the substrate, the interface being on edge on orientation. The well thickness was measured by TEM and found to be 15, 7.5, 4.0 and 2.5 nm and the barrier width of 15 nm was found to be in good agreement with the experimental growth parameters18. Figure 7 shows a HREM image of the InGaAs well sandwiched between two InP barrier layers. It could be noticed from the image that firstly the microstructure of either of the phases is clean, as it does not contain any defect and secondly the InP barrier layer has perfect atomic plane matching with no defect in view of image. It was possible to see certain amount of strain in the epilayer induced due to the growth process as revealed by the strain contrast. The atomic arrangement was found to be distorted at many places. In such examinations, it was revealed that the global microstructure may be good but there

Fig. 5—Bright field TEM micrograph showing quasicrystals of nano-dimension obtained after crystallization of Zr69.5Cu12Ni11Al7.5 amorphous alloy

Fig. 6—Bright field TEM micrograph of quantum wells of nominal widths 2.5, 3.9, 7.8, 15nm; barrier width 30nm

hydrogen storage. Fig. 5 shows the presence of a quasicrystalline phase in the nano-morphology obtained by crystallization of an amorphous phase in this alloy. It was difficult to establish the quasicrystalline nature of the particles by selected area electron diffraction since it sampled a large number of nanocrystalline particles. The true quasicrystalline nature could be revealed only by using nano-diffraction technique. One such nano-diffraction pattern is shown as inset in Fig.5 revealing 5-fold symmetry of the pattern. It has been proposed that these alloys have icosahedral short range order in their undercooled melts and in the amorphous phase16. 2.4 Multilayered materials

Multilayers are fabricated in layers built atom-byatom. Multilayers have been grown by a variety of techniques including molecular beam epitaxy (MBE), metal-organic vapour phase epitaxy (MOVPE), etc. These are fully dense and have high concentration of interfaces. The component layers of a multilayer can be as thin as two atomic layers with a strength that nearly approaches theoretical limits. Shortwavelength optics, high-performance capacitors for energy storage, industrial capacitors, integratedcircuit interconnects, tribological coatings, ultra-high strength materials and coatings for gas turbine engines are some of the areas where these materials have application potential. These have very high corrosion and erosion resistance and are ideally suited for aerospace and engine applications. 2.4.1 Electronic multilayer

124

INDIAN J PURE & APPL PHYS, VOL 44, FEBRUARY 2006

material shows the presence of four layer revealing coherent interfaces. Acknowledgement The authors wish to thank Mr B P Sharma, Head, Materials Science Division and Mr A K Grover, Head, Materials Processing Division, Bhabha Atomic Research Centre, Mumbai, for their constant support and encouragement during the course of this study. References

Fig. 7—HREM micrograph of InGaAs quantum well between two InP barrier layers

may be small scale imperfections at many locations which can get revealed by HREM only. 3 Conclusions TEM is a very powerful tool to investigate the microstructure in nanometer scale. Various examples have been shown in this paper to demonstrate the capability of TEM. TEM under HREM mode is, in particular, very useful to study the atomic arrangements in various phases and at the interfaces. HREM images revealed the presence of various defects inside the nano-crystalline grains. The presence of lattice fringes up to the grain boundary shows that nano-grains structure is similar to that of micrograins. High resolution of multilayer InGaAs

1 Hosokawa F, Tomita T, Naruse M, Honda T, Hartel P & Haider M, J Electron Microscopy, 52 (2003) 3. 2 Amelinckx S, Lucas A & Lambin P, Rep Prog Phys, 62 (1999) 1471. 3 Kumar Dharmesh, Bera S, Tripathi A K, Dey G K & Gupta N M, Microporous & Mesoporous Materials, 66 (2003) 157. 4 Kumar Dharmesh, Pillai K T, Sudersanan V, Dey G K & Gupta N M, Chemistry of Materials, 15 (2003) 3859. 5 Kumar Dharmesh, Krishna Vidya, Selvam P, Dey G K & Gupta N M, Advances in Nanoscience & Nanotechnology, (2004) 265. 6 Kumar D, Dey G K & Gupta N M, Phys Chem & Chem Phys, 5 (2003) 5477. 7 Dharmesh Kumar, Bera S, Tripathi A K, Dey G K & Gupta N M, Microporous & Mesoporous Materials (in press). 8 Andries Hendrik Janssen, Three-dimensional transmission electron microscopy of porous catalysts /-[S.l.]: [s.n.], 2003Tekst.-Proefschrift Universiteit Utrecht 9 Williams D B & Carter C B, Transmission electron microscopy III (New York; Plenum Press), 1996, 459. 10 Thomas G J, Siegel R W & Eastman J A, Scripta Metal, 24 (1990) 201. 11 Ch Andra R, Taneja P, John T, Ayuub P, Dey G K & Kulsrestha S K, Nano Structured Materials, 8 (1999) 1171. 12 Murty B S, Ping D H, Hono K & Inoue A, Acta Mater, 48 (2000) 3985. 13 Kim J Y, Majzoub E H, Gibbons P C & Kelton K F, Mat Res Soc Symp Proc, 553 (1999) 483. 14 Kelton K F, Mat Res Soc Symp Proc, 553 (1999) 471. 15 Koster U, Meinhardt J, Roos S & Liebertz H, Appl Phys Lett, 69 (1996) 179. 16 Xing L Q, Hufnagel T C, Eckert J, Loser W & Schultz L, Appl Phys Lett, 77 (2000) 1970. 17 Sridhar Rao D V, Murlidharan K & Dey G K, Proceedings DAE-BRNS Workshop on Thin Film Multilayers, (1999) 30. 18 Sridhar Rao D, Muralidharan K, Dey G K, Halder S K, Bhagayannarayan G, Banerji P, Pal D & Bose D N, Bull of Mater Sci, 22 (1999) 947.