Ultramicroscopy ∎ (∎∎∎∎) ∎∎∎–∎∎∎
1 2 3 4 5 6 7 8 9 10 11 Q2 12 13 14 15 Q1 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66
Contents lists available at ScienceDirect
Ultramicroscopy journal homepage: www.elsevier.com/locate/ultramic
Framed carbon nanostructures: Synthesis and applications in functional SPM tips I.S. Mukhin a,b, I.V. Fadeev b, M.V. Zhukov b, V.G. Dubrovskii a,b,d, A.O. Golubok b,c a
St. Petersburg Academic University, Khlopina 8/3, 194021 St. Petersburg, Russia ITMO University, Kronverksky pr. 49, 197101 St. Petersburg, Russia c Institute for Analytical Instrumentation of the Russian Academy of Sciences, Rizhsky 26, 190103 St. Petersburg, Russia d Ioffe Physical Technical Institute of the Russian Academy of Sciences, Politekhnicheskaya 26, 194021 St. Petersburg, Russia b
art ic l e i nf o
a b s t r a c t
Article history: Received 29 November 2013 Received in revised form 3 October 2014 Accepted 7 October 2014
We present a synthesis method to fabricate framed carbon-based nanostructures having highly anisotropic shapes, in particular, the nanofork and nanoscalpel structures which are obtained systematically under optimized growth conditions. A theoretical model is developed to explain the formation of such nanostructures on Si and W cantilevers exposed to a focused electron beam. We then demonstrate the potentials of these nanostructures as functional tips for scanning probe microscopy. Owing to their anisotropic shapes, such tips can be very useful for nanolithography, nanosurgery of biological objects, and precise manipulation with surface particles. Overall, our method provides a simple and robust way to produce functional scanning probe microscopy tips with variable shapes and enhanced capabilities for different applications compared to standard cantilevers. & 2014 Published by Elsevier B.V.
Keywords: Framed carbon nanostructures Nanoscalpel SPM probes Growth model Nanomanipulation Nanosurgery
1. Introduction Carbon nanomaterials of different types, such as nanodiamonds [1], fullerenes [2,3], graphenes [4,5], nanotubes [6–8], and nanowires [9], have recently gained a rapidly increasing interest for their unique fundamental properties and promising applications. By pursuing advanced nanofabrication methods, rather sophisticated functional nanomaterials can be synthesized which combine carbon nanostructures of different types. In particular, using carbon-based nanotubes or nanowires as building blocks, one can produce functionalized framed carbon nanostructures (FCNSs) with tunable geometry [10]. These FCNSs can be grown from a gaseous carbon environment. Formation process is induced by a focused electron beam, either via the directional catalyzed elongation or by the self-induced mechanism, as in the case of semiconductor nanowires [11,12]. FCNSs present an interesting example of nanographs with tunable orientation and dimensionality. From fundamental viewpoint, such nanographs can become quantum graphs [13] under certain conditions. FCNSs can be used as functional elements of nano-antennas or specialized tips for scanning probe microscopy (SPM). Such tips can be advantageous for visualization of deep channels or vertical walls, nanomodification of solid surfaces, manipulation of micro- and nano-objects, and in nanosurgery. Consequently, this paper is devoted to synthesis, characterization and SPM applications of FCNSs constructed from highly anisotropic carbon nanowires or nanosheets with different arrangements. A high
aspect ratio of “nanoscalpel” geometries considered hereinafter can bring about some new capabilities compared to the standard silicon cantilevers whose aspect ratio is of the order of one. Indeed, a symmetric cantilever has limitations in visualization of deep asymmetric channels, in particular, for the precise determination of the position and length of sharp elongated structures such as submicron surface steps or horizontal nanowires. In lithography, the channel profiles always reflect the geometry of the cantilever tip used and thus it is hard to achieve deep and narrow microchannels as well as produce straight incisions with predefined depth (which is often required in nanosurgery). A small radius of the tip (1–50 nm) does not enable the necessary mechanical stability of the cantilever in contact with a submicron particle. This does not favor a robust micromanipulation with surface particles which need to be relocated over large distances from the origin.
2. Synthesis of carbon structures Growth of FCNSs was performed on the tips of Si or W cantilevers fixed on a scanning electron microscope (SEM) stage, as shown in Fig. 1. The electron beam (1) exposes the cantilever tip (2) as well as the underlying target with a carbon coating (3) which produces a flux of carbon ions (4). The beam is focused either directly on the tip or up to 20 nm away from its edge, after which the beam is scanned on a small surface area (typically 15 15 nm2). FCNSs (5) nucleate on
http://dx.doi.org/10.1016/j.ultramic.2014.10.008 0304-3991/& 2014 Published by Elsevier B.V.
Please cite this article as: I.S. Mukhin, et al., Framed carbon nanostructures: Synthesis and applications in functional SPM tips, Ultramicroscopy (2014), http://dx.doi.org/10.1016/j.ultramic.2014.10.008i
67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97
2
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66
I.S. Mukhin et al. / Ultramicroscopy ∎ (∎∎∎∎) ∎∎∎–∎∎∎
the tip surface exposed to the electron beam, and are monitored by SEM, which allows for a precise in-situ characterization of geometry. In this method, the distance between the SEM tip and target and the type of the carbon film used do not critically influence the growth process. The distance between the tip and the target can be varied within a wide range from 5 to 20 mm, while the carbon coating of the target is accessed via standard thermal sputtering of carbon electrodes or by simply sticking a carbon scotch. It is noteworthy that carbon nanostructures can even be grown without any carbon coating if the residual atmosphere within the SEM vacuum chamber contains the C–H groups which play the role of a carbon precursor. Also, there is no direct correlation between the scan area and the size or geometry of the nano-objects produced. The geometry of the structure is defined by the trajectory of the focused electron beam, where the minimum size ( 50 nm) is set by the area from which the secondary electrons are emitted rather than by the area exposed to the beam. By varying the direction and the moving speed of the electron beam, the accelerating voltage and the current, we are able to tune the growth process of individual carbon nanostructures and to fabricate FCNSs of different types. In particular, varying the moving direction of the beam yields rather complex geometries that will be considered elsewhere. Here, the simplest process is considered where the beam is moved along a fixed direction with a constant speed. In this case, we systematically observe a submicrometer-sized base, followed by several nanowires having 50–100 nm diameters and a parallel alignment, as in Fig. 2(a). We will call this shape “nanofork”. The nanoforks are formed within the 25–30 keV window of the electron beam energy. A lower energy (15–25 keV) usually yields two-dimensional corrugated structures with a corrugation period of 50–100 nm, as in Fig. 2(b) and (c) (nanoscalpel [23]). Below 10 keV, the shape of FCNSs systematically transforms to a rod-like (Fig. 2(d)). X-ray chemical analysis of these FCNSs reveals that they are purely carbon structures. In order to identify the crystal structure of the FCNSs, we performed the transmission electron microscopy (TEM) of the representative samples. The typical TEM images of a carbon nanoscalpel on top of a W cantilever are shown in Fig. 3, along with the corresponding electron diffraction patterns. The diffraction pattern in the insert of Fig. 3(a) reveals the presence of wellresolved bright spots and the two diffuse and low intensity rings. Fig. 3(b) displays the enlarged image of a nanoscalpel part close to its surface with the Fourier transform of the corresponding diffraction pattern shown in the insert. The bright spots on the diffraction pattern represent the (311) crystallographic direction of the W crystal, which is perpendicular to the tip axis. The diffuse
rings in the insert of Fig. 3(a) and the broadened reflexes in the insert of Fig. 3(b) should correspond to a short-range order of C atoms in the nanoscalpel. Therefore, our FCNSs are deemed to have emerged in nearly amorphous form with a short-range crystallinity.
3. Growth mechanism Generally, two kinetic mechanisms for the formation of nanoforks or nanoscalpels could be considered: (i) coalescence of several nanowires into a plane FCNS or (ii) a self-induced decay of a plane FCNS into several parallel nanowires. In Ref. [14], we have proposed a simple model that is capable of describing the formation of flat FCNSs (nanoscalpels) under a focused electron beam. Within the model, 30 keV electrons bombard a carbon target and induce a flux of carbon ions. The emitted ions are confined near the beam axis by the own electric field of the beam. As a result, growth of each FCNS is directed along the beam axis, while the linear displacement of the beam gives rise to a plane nanoscalpel shape. Numerical simulations have confirmed that such plane shapes eventually emerge, while the initial growth step leads to the formation of tapered submicrometer bases similar to that shown in Fig. 2(a). A more detailed modeling should account for different kinetic pathways of growth including the surface diffusion, the radial extension of nanowires and the coalescence mechanisms, as in Refs. [11,15–19]. This will allow for a better control over the entire growth process by technological parameters. Here, we present only the elementary energetic considerations to explain the formation of the fork-like structures. From the SEM image shown in Fig. 2(a), we assume that the tapered base forms at the initial growth step and then transforms into the nanowires. We consider the model geometry shown in Fig. 4(a), with the cantilever tip of the width dn , the tapered FCNS base of the length l0 that extends toward the top with the taper angle α, and n identical rectangular nanowires having the width d ¼ const and height h1 atop the base of widthd0 ¼ dn þ 2l0 tan α. The thickness of the whole structure equals Δ ¼ const. The total volume V1and the surface energy F1 of the nanofork structure in Fig. 4(a) are given by V 1 ¼ dnh1 Δ; F 1 ¼ γ p d0 Δþ 2γ v nh1 ðd þ ΔÞ:
ð1Þ
Here, γ p is the surface energy of the planar interfaces that are parallel to the top surface of the tip and γ v is the surface energy of all the vertical surfaces (see Fig. 4). While a uniform surface energy of all solid-vapor interfaces would be a good approximation for a completely amorphous and isolated FCNS, the asymmetry in surface energies could arise from the fact that the planes are oriented differently with respect to the electron beam (see Fig. 1). In any case, the nanowire fork is energetically preferred to the plane nanoscalpel structure which just continues to extend after the length l0 [Fig. 4(b)] if the surface energy F2 of the upper (dark) part of the structure shown in Fig. 4(b) is larger than F1 for the same volume. Free energy of forming the nanofork is than lower due to a smaller surface energy term for the same volume energy [12]. The volume V2 and the surface energy F2 of the tapered nanoscalpel structure of the height h2 are given by 2
V 2 ¼ ðd0 h2 þ h2 tan αÞΔ; F 2 ¼ γ p ðd0 þ2h2 tan αÞΔ 2
þ 2γ v ðd0 h2 þ h2 tan αÞ þ2γ i h2 Δ= cos α:
Fig. 1. Experimental setup for FCNS growth: (1) the electron beam, (2) the cantilever tip, (3) the underlying target with a carbon coating, (4) the influx of carbon ions, and (5) the growing carbon nanostructure.
ð2Þ
where γ i is the surface energy of the inclined surface as shown in Fig. 4(b). Taking the difference F 1 F 2 under the condition V 1 ¼ V 2 , the nanofork is preferred to the nanoscalpel when the function ! sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi d0 γ p γi 4dh tan α 1þ 1 ð3Þ þ f ðhÞ ¼ h 2 2 γ v γ v sin α d0 is negative, in which h ¼ nh1 is the effective height of the structure.
Please cite this article as: I.S. Mukhin, et al., Framed carbon nanostructures: Synthesis and applications in functional SPM tips, Ultramicroscopy (2014), http://dx.doi.org/10.1016/j.ultramic.2014.10.008i
67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132
I.S. Mukhin et al. / Ultramicroscopy ∎ (∎∎∎∎) ∎∎∎–∎∎∎
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66
3
Fig. 2. Typical SEM images of differently shaped FCNSs on a Si cantilever: (a) nanofork; (b) and (c) nanoscalpel viewed in different directions; and (d) nanorod. The electron beam energy was 30 keV in (a), 25 keV in (b) and (c) and 10 keV in (d).
Fig. 3. (a) TEM image of a nanoscalpel structure with the electron diffraction pattern shown in the insert; (b) Enlarged TEM image of a part of the nanoscalpel, the insert showing the Fourier transform of the diffraction pattern. The diffuse rings in (a) and the broadened reflexes in (b) correspond to nearly amorphous carbon.
when γv o
Fig. 4. Illustrations of nanofork (a) and nanoscalpel (b) structures showing the main parameters of the model. The volume of the dark upper parts of both structures is fixed to a constant value in the model.
pffiffiffi Eq. (3) shows that f ðhÞ ffi h const h 4 0 at large enough h. The nanowire structure is thus energetically suppressed asymptotically due to its developed sidewalls, which is a well-known feature for nanowires on infinite substrates [18]. However, if nanowires are preferred at the initial growth step when they nucleate, the initial shape can pertain at the follow-up stages because the reverse transformation to nanoscalpel requires surface rearrangement of too many atoms and should be kinetically forbidden, as in the case of self-induced GaN nanowires [12]. From Eq. (3), the function f(h) is linear at small h and is negative
d γi ; γ tan α þ d0 p cos α
ð4Þ
In the isotropic case γ v ¼ γ p ¼ γ i , this is reduced to d0 cos α o dð1 þ cos αÞ, the inequality which can be satisfied only if n cos α o 1 þ sin α (since d o d0 =n). The latter condition requires large enough taper angle α and small number of the fork teeth n. As seen from Fig. 2(a) and (b), in some cases the nanoscalpel structure is not entirely flat but rather looks like an initial fork-like shape whose development has been suppressed at an early stage. Thus, our operating conditions should be close to a threshold between the two structures. The typical parameters of our nanoscalpel and nanofork structures (over more than 10 growth experiments) are: d0 ¼500 nm and α ¼ π=6, n ¼ 4, d ¼100 nm for nanoforks. If we take the approximation of a constant surface energy, γ v ¼ γ p ¼ γ i , the condition n cos α o1 þ sin α could not be fulfilled with our experimental parameters so that the scalpel structure always wins. Therefore, within the model, the nanofork is formed due to its lower effective surface energy. This is illustrated in Fig. 5 which shows the function f(h) for the typical experimental dimensions at different values of the control parameter A ¼ γ p =γ v þ2γ i =γ v . Overall, the formation of both the fork-like and the scalpel-like FCNSs is entirely possible and depends on the surface energy values. In the isotropic case with A¼3, the scalpel shape dominates, while for A¼9 the fork formation is preferred at the initial growth step where f(h) is negative. The number of nanowires and their shape will be highly dependent on kinetic factors and require a separate treatment. We also note that nothing prevents nanowires
Please cite this article as: I.S. Mukhin, et al., Framed carbon nanostructures: Synthesis and applications in functional SPM tips, Ultramicroscopy (2014), http://dx.doi.org/10.1016/j.ultramic.2014.10.008i
67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132
I.S. Mukhin et al. / Ultramicroscopy ∎ (∎∎∎∎) ∎∎∎–∎∎∎
4
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66
Fig. 5. Graphs of f(h) obtained from Eq. (3) with d0 ¼ 500 nm, d¼ 100 nm, α ¼ π=6 and different A. For isotropic surface energies (A ¼3), f ðhÞ is positive and the nanoscalpel shape is preferred everywhere. When A is increased to 9, f(h) becomes negative at small h, where the formation of nanofork (with 4 nanowire teeth) is favored.
from growing radially, as in the case of self-induced GaN nanowires on Si [12,19], which will change the surface energetics [12] and finally lead to coalescence. Unlike InAs/GaAs and other Stranski– Krastanow islands [20] or GaN nanowires on AlN interlayer [21], our FCNS should be induced by surface energetics rather than the lattice mismatch since amorphous structures are not expected to have any epitaxial constraint with respect to Si or W.
4. Use of nanoforks in nanolitography Contact and semi-contact SPM modes of force lithography are well-known for standard Si cantilevers [22]. The optimal regime of lithography is realized under the condition P tip 4 P int 4 P material , where P material denotes the plastic limit of a given sample, P int is the pressure in the contact area of the probe with the sample, and P tip is the plastic limit of the probe tip. In practice, the process of SPM lithography is always coupled with SPM visualization using the same probe. In order to improve the resolution of a nanofork probe in the visualization regime, we need to decrease the area of interaction between the sample and the spot. To do so, we incline the long axis of the probe with respect to the substrate surface by a small angle β, as shown in Fig. 6(a). In this configuration, only one of the fork's teeth is used for visualization, while several teeth do lithography. Of course, such an inclination will modify the above criterion of the mechanical stability, but is not expected to change significantly the qualitative picture. Fig. 6(b) shows the SPM image of a laser disc, obtained with the nanofork probe grown on top of a Si cantilever, with the inclination angle β¼151. It is seen that the probe ensures a reasonable resolution for the visualization of the force lithography. Obviously, a long nanofork tip is particularly advantageous compared to the standard one for a more precise visualization of deep lithographical channels. Let us now consider a very important question of the mechanical stability of anisotropic tips under the axial compression while pressing the sample surface [23]. According to Ref. [24], the tip is mechanically stable provided that F c =Scontact o P 3 , where Scontact is the contact area of the tip with the sample surface. The critical force pffiffiffi 2 F c ¼ ð2EΔ3 dÞ=ð3 2h ð1 þ νÞ1=2 Þ ð5Þ determines the mechanical stability under the axial stress, with E as the Young's modulus, ν as the Poisson ratio, d as the tip width
(which equals d0 for nanoscalpel), Δ as the thickness and h as the height. When the long axis of the asymmetric tip is perpendicular to the sample surface, the typical geometrical parameters of our nanoforks yield the value of Fc at approximately 45 μN, which is much larger than the maximum forces used for SPM visualization (usually about 1 μN) and force lithography (less than 35 μN). Eq. (5) shows that the critical mechanical force is proportional to the tip width d and thus the elongated tip shapes are stiffer compared to nanorods. In other words, the long plane of the nanoscalpel plays the role of a stiffening plate with respect to axial compressions and bends which makes the asymmetric tips very suiw for lithography of deep and narrow channels. We mention here the commercially available AFM cantilevers with carbon nanotubes (CNTs) attached at the tips. Of course, single CNT yields a superior resolution [25,26] but is much less robust due to its low mechanical stiffness. In addition, the absence of the pronounced asymmetry of a CNT tip does not allow for a manipulation with nano-objects as will be discussed shortly. Fig. 7 shows the results of lithography of a 10 nm Au film on a polycarbonate surface, obtained with a nanofork tip. The images were obtained with the same probe in the semi-contact mode. The image in Fig. 7(a) reveals a fine-grained structure of the Au film, demonstrating a high spatial resolution of the tip. The lithographical process was performed as follows. The probe modified with the nanofork FCNS was pressed to the Au surface with F 4 1 μN and then moved along the long axis of the fork. After that, the tip was torn off the surface and retained to the initial position in the substrate normal direction, while the in-plane co-ordinate was moved away by 500 nm in each step. In this way, we were able to form the surface dashes with 500 nm spacing [Fig. 7(b)]. The dash width was found to be 100 nm, i.e., equals exactly the width of the tip. The measured dash depths were equal to 10.5 nm showing that the Au film is completely trenched [Fig. 7(c)]. Close examination of different regimes of lithography yields the following optimal parameters: the pressing force¼1.5 μN, the moving velocity¼ 2 μm/s, and the exposition time of 300 μs in each point. Under these conditions, the nanofork tip trenches through the Au film to the substrate with the minimum width of the cut. It should be noted that, on a surface of a soft polycarbonate sample and for a sufficiently low force applied, the “point” regime of lithography also produces a point spot in the Au film. This reveals a small effective contact area, defined by the inclination β. As the applied force increases, the spot starts to elongate along the direction of the long axis of the tip. To investigate the spatial resolution of the nanofork lithography, we performed the modification of a polycarbonate surface according to the template with a varying grating spacing: 150 nm, 100 nm and 80 nm. The results are shown in Fig. 8. When the spacing equals 150 nm [points C and D in Fig. 8(b)] or 100 nm (points E and F), the obtained channels are well separated. The situation changes for the 80 nm spacing (points G and H) where the two channels merge and the total depth is much larger than in points C, D, E and F. This result is expected, because the spatial resolution is limited by the width of the fork itself ( 100 nm), while points G and F relate to the two-pass lithography mode. The same procedure was repeated on the polycarbonate surface covered with a 10 nm gold layer. The lithography parameters were the following: the pressing force was 1.5 μN, the moving velocity was 2 μm/s, and the exposition time of 300 μs in each point. The grating spacing was 250 nm, 150 nm, 100 nm, 50 nm and 25 nm, as shown from left to right in Fig. 9(a). It is seen that the lines are well-separated for the 50 nm spacing and above. Obviously, the linear regime (where the contact area between the tip and the substrate defines the resolution) works only when the displacement of the cantilever occurs is aligned with its long axis.
Please cite this article as: I.S. Mukhin, et al., Framed carbon nanostructures: Synthesis and applications in functional SPM tips, Ultramicroscopy (2014), http://dx.doi.org/10.1016/j.ultramic.2014.10.008i
67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132
I.S. Mukhin et al. / Ultramicroscopy ∎ (∎∎∎∎) ∎∎∎–∎∎∎
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66
5
Fig. 6. (a) Schematic representation of the best tip positioning with respect to the substrate surface. (b) A polycarbonate surface with the pits visualized by the nanofork probe. The inclination angle β was 151 in these particular measurements.
Fig. 7. Surface lithography with a nanofork tip: (a) Atomic force microscopy (AFM) image of a 10 nm Au film on a polycarbonate surface before the lithography. (b) AFM image of a diffraction grating with the 500 nm spacing, 100 nm line width and 10.5 nm depth. (c) Cross-section along the AB segment in (b).
Fig. 8. (a) AFM image of a grating with variable spacing (150, 100 and 80 nm) on a polycarbonate surface, obtained with a nanofork probe at 230 nN force. (b) Cross section along the АВ segment shown in (a).
5. Nanosurgery of erythrocytes Nanofork probes can also be used for nanosurgery of biological objects. In particular, a partial incision of cell membranes enables an investigation of the local structure of the cell without destroying it and, on the other hand, an introduction of different substances (drugs or dyes) through the incision. Fig. 10 shows the results of the surface modification of erythrocytes dried in air for 10 min before surgery. The dissection
was performed by moving the nanofork tip with respect to the erythrocyte surface at a speed of 0.5 μm/s and different pressing forces. Below1 μN, modification is seen only on some parts of the surface. With a 2.5 μN force, the membrane surface is incised all the way, with an incision depth of about 12 nm [Fig. 10(a)]. When the force is raised further, both width and depth of the incisions gradually increases. At 6.5 μN, the membrane starts to flake off the incision line, which is seen at the edges of the modified membrane. The incision depth is larger than 25 nm, while the width
Please cite this article as: I.S. Mukhin, et al., Framed carbon nanostructures: Synthesis and applications in functional SPM tips, Ultramicroscopy (2014), http://dx.doi.org/10.1016/j.ultramic.2014.10.008i
67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132
6
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66
I.S. Mukhin et al. / Ultramicroscopy ∎ (∎∎∎∎) ∎∎∎–∎∎∎
Fig. 9. (a) AFM image of a grating with variable spacing (250, 150, 100, 50 and 25 nm) on a polycarbonate surface covered with a 10 nm gold layer, obtained with a nanofork probe. (b) Cross section along the АВ segment shown in (a).
Fig. 10. AFM visualizations (top) and the corresponding cross sections (bottom) of the modified surface of an erythrocyte, obtained with a nanofork tip with different pressing forces: 2.5 μN (a) and (d), 6.5 μN (b) and (e) and 7.5 μN (c) and (f).
reaches 150 nm [Fig. 10(b)]. At 7.5 μN, Fig. 10(c) reveals the internal structure of the erythrocyte under the membrane. The latter is clearly seen along the border of the dissected region. Higher applied forces typically result in the displacements and conglomerations of the membrane material.
6. Manipulation of nano-objects with nanoforks To perform manipulations with micro- and nano-objects by AFM, the sample is first visualized in the semi-contact mode, after which the probe is placed near the particle and the AFM is converted to the contact mode. Then, the particle is pushed by the tip to the desired spot on the surface. Returning to the semi-contact mode, the particle position is visualized again and the entire cycle is repeated if any additional positioning of the particle is required [27]. Such a method enables a precise relocation of different objects for a distance from several tens of nanometers to several micrometers. It is however
difficult to move a particle at a large distances ( 10 μm) with the standard tips, which is paramount, for example, for manipulations with biological objects and applications in nanosensors. In particular, there is an issue of a particle losing its contact with the probe and falling off the tip when meeting surface inhomogeneities of different types. These issues could be surpassed by using the nanofork tips, owing to a pronounced asymmetry of their shape. Indeed, we have not seen any problem of long-range transfer of particles with our modified probes. In particular, Fig. 11 shows the AFM images of colloidal particles with 650 nm diameter before and after moving some of them with a nanofork tip. The particles were transferred mainly along the long axis of the tip. After the first manipulation, the particles were moved for about 2 μm, as follows from Fig. 11 (b) (the moving distance is AB). The second manipulation enabled moving the particles outside the scanned area [Fig. 11(c)]. It is noteworthy that, in Fig. 11(b) and (c), we can clearly see the AFM contrast revealing the initial position of the particles [labeled A in Fig. 11(b)].
Please cite this article as: I.S. Mukhin, et al., Framed carbon nanostructures: Synthesis and applications in functional SPM tips, Ultramicroscopy (2014), http://dx.doi.org/10.1016/j.ultramic.2014.10.008i
67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132
I.S. Mukhin et al. / Ultramicroscopy ∎ (∎∎∎∎) ∎∎∎–∎∎∎
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 Q3 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
7
Fig. 11. AFM images of an ensemble of colloidal particles on a polycarbonate surface in their initial position (a), after the first (b) and the second (c) manipulation with the nanofork probe. The arrows show the moving direction.
7. Conclusions In summary, we have shown that the nanofork and the nanoscalpel FCNSs can be obtained regularly on the standard Si and W cantilevers by applying a special growth procedure involving the scanning focused electron beam that produces the flat spatial distributions of carbon ions emitted from a carbon target. This gives rise to a directional growth of elongated nanosheets, which can be either uniform (nanoscalpel) or separated into perfectly aligned parallel nanowires of 50–100 nm width (nanofork) depending on technological parameters. The preliminary energetic model is proposed that explains the formation of nanowire teeth under certain conditions. We have demonstrated the use of these nanofork carbon structures as modified SPM tips for nanolithography and visualization of deep channels, nanosurgery of biological objects, and nanomanipulation of colloidal particles. We now plan to perform more growth experiments and modeling to better illuminate the tuning knobs to produce different tip shapes, to control the number and shape of the fork teeth, and the shape transformation from nanoforks to nanoscalpels. More investigations are required into the physical properties of these carbon nanostructures as well as SPM characterization and manipulation with different nano-objects.
Acknowledgment This work was partially supported by the Government of the Russian Federation (Grant # 074-U01), a few grants of the Russian Foundation for Basic Research and the FP7 project FUNPROB. VGD gratefully acknowledges financial support of the Russian Science Foundation under the Grant 14-22-00018. The authors kindly thank Nevedomsky V.N. and Soshnikov I.P. for help with TEM studies. References [1] J.Y. Raty, G. Galli, Ultradispersity of diamond at the nanoscale, Nat. Mater. 2 (2003) 792–795. [2] H.W. Kroto, The stability of the fullerenes Cn, with n¼ 24, 28, 32, 36, 50, 60 and 70, Nature 329 (1987) 529–531. [3] H.W. Kroto, J.R. Heath, S.C. O'Brien, R.F. Curl, R.E. Smalley, C60: Buckminsterfullerene, Nature 318 (1985) 162–163. [4] A.K. Geim, K.S. Novoselov, The rise of grapheme, Nat. Mater. 6 (2007) 183–191.
[5] K.S. Novoselov, A.K. Geim, S.V. Morozov, D. Jiang, M.I. Katsnelson, I.V. Grigorieva, S.V. Dubonos, A.A. Firsov, Two-dimensional gas of massless Dirac fermions in graphene, Nature 438 (2005) 197–200. [6] X. Zhao, Y. Ando, Y. Liu, M. Jinno, T. Suzuki, Carbon nanowire made of a long linear carbon chain inserted inside a multiwalled carbon nanotube, Phys. Rev. Lett. 90 (2003) 187401. [7] S. Iijima, Helical microtubules of graphitic carbon, Nature 354 (1991) 56–58. [8] N.G. Chopra, R.J. Luyken, K. Cherrey, V.H. Crespi, M.L. Cohen, S.G. Louie, A. Zettl, Boron nitride nanotubes, Science 269 (1995) 966–967. [9] L. Qintao, Z. Dezhang, G. Jinlong, Fabrication of carbon nanowire networks by Si ion beam irradiation, Appl. Phys. Lett. 89 (2006) 053107. [10] A.O. Golubok, I.U. Popov, I.S. Mukhin, I.S. Lobanov, Creation and study of 2D and 3D carbon nanographs, Elsevier Phys. E 44 (2012) 976–980. [11] V.G. Dubrovskii, T. Xu, Y. Lambert, J.P. Nys, B. Grandidier, D. Stievenard, W. Chen, P. Pareige, Narrowing the length distribution of Ge nanowires, Phys. Rev. Lett. 108 (2012) 105501. [12] V.G. Dubrovskii, V. Consonni, A. Trampert, L. Geelhaar, H. Riechert, Scaling thermodynamic model for the self-induced nucleation of GaN nanowires, Phys. Rev. B 85 (2012) 165317. [13] M.A.Kokoreva, V.A. Margulis, M.A. Pyataev, Electron transport in a two-terminal Aharonov–Bohm ring with impurities, Elsevier Phys. E 43 (2011) 1610–1617. [14] S.A. Chivilikhin, A.O. Golubok, I.S. Mukhin, Mathematical model of nanowhisker growth under focused electron beam, NTV IFMO 66 (2010) 78–83. [15] M. Tchernycheva, L. Travers, G. Patriarche, F. Glas, J.C. Harmand, G.E. Cirlin, V.G. Dubrovskii, Au-assisted molecular beam epitaxy of InAs nanowires: growth and theoretical analysis, J. Appl. Phys. 102 (2007) 094313. [16] K. Ostrikov, H. Mehdipour, Nanoscale plasma chemistry enables fast, sizeselective nanotube nucleation, J. Am. Chem. Soc. 134 (2012) 4303–4312. [17] K. Ostrikov, H. Mehdipour, Thin single-walled carbon nanotubes with narrow chirality distribution: constructive interplay of plasma and Gibbs–Thomson effects, ACS Nano 5 (2011) 8372–8382. [18] V.G. Dubrovskii, N.V. Sibirev, Growth thermodynamics of nanowires and its application to polytypism of zinc blende III–V nanowires, Phys. Rev. B 77 (2008) 035414. [19] V.G.Dubrovskii, V. Consonni, L. Geelhaar, A. Trampert, H. Riechert, Scaling growth kinetics of self-induced GaN nanowires, Appl. Phys. Lett. 100 (2012) 153101. [20] D. Bimberg, M. Grundmann, N.N. Ledentsov, Quantum Dot Heterostructures, Wiley, New York, 1999. [21] V. Consonni, M. Knelangen, L. Geelhaar, A. Trampert, H. Riechert, Nucleation mechanisms of epitaxial GaN nanowires: Origin of their self-induced formation and initial radius, Phys. Rev. B 81 (2010) 085310. [22] M. Heyde, K. Rademann, B. Cappella, M. Geuss, H. Sturm, T. Spangenberg, H. Niehus, Dynamic plowing nanolithography on polymethylmethacrylate using an atomic force microscope, Rev. Sci. Instrum. 72 (2001) 136–141. [23] J.D. Beard, D.J. Burbridge, A.V. Moskalenko, O. Dudko, P.L. Yarova, S.V. Smirnov, S.N. Gordeev, An atomic force microscope nanoscalpel for nanolithography and biological applications, Nanotechnology 20 (2009) 1–10. [24] L.D. Landau, E.M. Lifshitz, Theory of Elasticity, Pergamon, Oxford, 1970. [25] J. Tang, G. Yang, Q. Zhang, A. Parhat, B. Maynor, J. Liu, L.-C. Qin, O. Zhou, Rapid and reproducible fabrication of carbon nanotube AFM probes by dielectrophoresis, Nano Lett. 5 (2005) 11–14. [26] N.R. Wilson, J.V. Macpherson, Carbon nanotube tips for atomic force microscopy, Nat. Nanotechnol. 4 (2009) 483–491. [27] L.T. Hansenyk, A. Kuhlez, A.H. Sørensen, J. Bohr, P.E. Lindelof, A technique for positioning nanoparticles using an atomic force microscope, Nanotechnology 9 (1998) 337–342.
Please cite this article as: I.S. Mukhin, et al., Framed carbon nanostructures: Synthesis and applications in functional SPM tips, Ultramicroscopy (2014), http://dx.doi.org/10.1016/j.ultramic.2014.10.008i
61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119
1 2 3 4 5 6 7 8 9 10 11 Q2 12 13 14 15 Q1 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66
Contents lists available at ScienceDirect
Ultramicroscopy journal homepage: www.elsevier.com/locate/ultramic
Framed carbon nanostructures: Synthesis and applications in functional SPM tips I.S. Mukhin a,b, I.V. Fadeev b, M.V. Zhukov b, V.G. Dubrovskii a,b,d, A.O. Golubok b,c a
St. Petersburg Academic University, Khlopina 8/3, 194021 St. Petersburg, Russia ITMO University, Kronverksky pr. 49, 197101 St. Petersburg, Russia c Institute for Analytical Instrumentation of the Russian Academy of Sciences, Rizhsky 26, 190103 St. Petersburg, Russia d Ioffe Physical Technical Institute of the Russian Academy of Sciences, Politekhnicheskaya 26, 194021 St. Petersburg, Russia b
art ic l e i nf o
a b s t r a c t
Article history: Received 29 November 2013 Received in revised form 3 October 2014 Accepted 7 October 2014
We present a synthesis method to fabricate framed carbon-based nanostructures having highly anisotropic shapes, in particular, the nanofork and nanoscalpel structures which are obtained systematically under optimized growth conditions. A theoretical model is developed to explain the formation of such nanostructures on Si and W cantilevers exposed to a focused electron beam. We then demonstrate the potentials of these nanostructures as functional tips for scanning probe microscopy. Owing to their anisotropic shapes, such tips can be very useful for nanolithography, nanosurgery of biological objects, and precise manipulation with surface particles. Overall, our method provides a simple and robust way to produce functional scanning probe microscopy tips with variable shapes and enhanced capabilities for different applications compared to standard cantilevers. & 2014 Published by Elsevier B.V.
Keywords: Framed carbon nanostructures Nanoscalpel SPM probes Growth model Nanomanipulation Nanosurgery
1. Introduction Carbon nanomaterials of different types, such as nanodiamonds [1], fullerenes [2,3], graphenes [4,5], nanotubes [6–8], and nanowires [9], have recently gained a rapidly increasing interest for their unique fundamental properties and promising applications. By pursuing advanced nanofabrication methods, rather sophisticated functional nanomaterials can be synthesized which combine carbon nanostructures of different types. In particular, using carbon-based nanotubes or nanowires as building blocks, one can produce functionalized framed carbon nanostructures (FCNSs) with tunable geometry [10]. These FCNSs can be grown from a gaseous carbon environment. Formation process is induced by a focused electron beam, either via the directional catalyzed elongation or by the self-induced mechanism, as in the case of semiconductor nanowires [11,12]. FCNSs present an interesting example of nanographs with tunable orientation and dimensionality. From fundamental viewpoint, such nanographs can become quantum graphs [13] under certain conditions. FCNSs can be used as functional elements of nano-antennas or specialized tips for scanning probe microscopy (SPM). Such tips can be advantageous for visualization of deep channels or vertical walls, nanomodification of solid surfaces, manipulation of micro- and nano-objects, and in nanosurgery. Consequently, this paper is devoted to synthesis, characterization and SPM applications of FCNSs constructed from highly anisotropic carbon nanowires or nanosheets with different arrangements. A high
aspect ratio of “nanoscalpel” geometries considered hereinafter can bring about some new capabilities compared to the standard silicon cantilevers whose aspect ratio is of the order of one. Indeed, a symmetric cantilever has limitations in visualization of deep asymmetric channels, in particular, for the precise determination of the position and length of sharp elongated structures such as submicron surface steps or horizontal nanowires. In lithography, the channel profiles always reflect the geometry of the cantilever tip used and thus it is hard to achieve deep and narrow microchannels as well as produce straight incisions with predefined depth (which is often required in nanosurgery). A small radius of the tip (1–50 nm) does not enable the necessary mechanical stability of the cantilever in contact with a submicron particle. This does not favor a robust micromanipulation with surface particles which need to be relocated over large distances from the origin.
2. Synthesis of carbon structures Growth of FCNSs was performed on the tips of Si or W cantilevers fixed on a scanning electron microscope (SEM) stage, as shown in Fig. 1. The electron beam (1) exposes the cantilever tip (2) as well as the underlying target with a carbon coating (3) which produces a flux of carbon ions (4). The beam is focused either directly on the tip or up to 20 nm away from its edge, after which the beam is scanned on a small surface area (typically 15 15 nm2). FCNSs (5) nucleate on
http://dx.doi.org/10.1016/j.ultramic.2014.10.008 0304-3991/& 2014 Published by Elsevier B.V.
Please cite this article as: I.S. Mukhin, et al., Framed carbon nanostructures: Synthesis and applications in functional SPM tips, Ultramicroscopy (2014), http://dx.doi.org/10.1016/j.ultramic.2014.10.008i
67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97
2
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66
I.S. Mukhin et al. / Ultramicroscopy ∎ (∎∎∎∎) ∎∎∎–∎∎∎
the tip surface exposed to the electron beam, and are monitored by SEM, which allows for a precise in-situ characterization of geometry. In this method, the distance between the SEM tip and target and the type of the carbon film used do not critically influence the growth process. The distance between the tip and the target can be varied within a wide range from 5 to 20 mm, while the carbon coating of the target is accessed via standard thermal sputtering of carbon electrodes or by simply sticking a carbon scotch. It is noteworthy that carbon nanostructures can even be grown without any carbon coating if the residual atmosphere within the SEM vacuum chamber contains the C–H groups which play the role of a carbon precursor. Also, there is no direct correlation between the scan area and the size or geometry of the nano-objects produced. The geometry of the structure is defined by the trajectory of the focused electron beam, where the minimum size ( 50 nm) is set by the area from which the secondary electrons are emitted rather than by the area exposed to the beam. By varying the direction and the moving speed of the electron beam, the accelerating voltage and the current, we are able to tune the growth process of individual carbon nanostructures and to fabricate FCNSs of different types. In particular, varying the moving direction of the beam yields rather complex geometries that will be considered elsewhere. Here, the simplest process is considered where the beam is moved along a fixed direction with a constant speed. In this case, we systematically observe a submicrometer-sized base, followed by several nanowires having 50–100 nm diameters and a parallel alignment, as in Fig. 2(a). We will call this shape “nanofork”. The nanoforks are formed within the 25–30 keV window of the electron beam energy. A lower energy (15–25 keV) usually yields two-dimensional corrugated structures with a corrugation period of 50–100 nm, as in Fig. 2(b) and (c) (nanoscalpel [23]). Below 10 keV, the shape of FCNSs systematically transforms to a rod-like (Fig. 2(d)). X-ray chemical analysis of these FCNSs reveals that they are purely carbon structures. In order to identify the crystal structure of the FCNSs, we performed the transmission electron microscopy (TEM) of the representative samples. The typical TEM images of a carbon nanoscalpel on top of a W cantilever are shown in Fig. 3, along with the corresponding electron diffraction patterns. The diffraction pattern in the insert of Fig. 3(a) reveals the presence of wellresolved bright spots and the two diffuse and low intensity rings. Fig. 3(b) displays the enlarged image of a nanoscalpel part close to its surface with the Fourier transform of the corresponding diffraction pattern shown in the insert. The bright spots on the diffraction pattern represent the (311) crystallographic direction of the W crystal, which is perpendicular to the tip axis. The diffuse
rings in the insert of Fig. 3(a) and the broadened reflexes in the insert of Fig. 3(b) should correspond to a short-range order of C atoms in the nanoscalpel. Therefore, our FCNSs are deemed to have emerged in nearly amorphous form with a short-range crystallinity.
3. Growth mechanism Generally, two kinetic mechanisms for the formation of nanoforks or nanoscalpels could be considered: (i) coalescence of several nanowires into a plane FCNS or (ii) a self-induced decay of a plane FCNS into several parallel nanowires. In Ref. [14], we have proposed a simple model that is capable of describing the formation of flat FCNSs (nanoscalpels) under a focused electron beam. Within the model, 30 keV electrons bombard a carbon target and induce a flux of carbon ions. The emitted ions are confined near the beam axis by the own electric field of the beam. As a result, growth of each FCNS is directed along the beam axis, while the linear displacement of the beam gives rise to a plane nanoscalpel shape. Numerical simulations have confirmed that such plane shapes eventually emerge, while the initial growth step leads to the formation of tapered submicrometer bases similar to that shown in Fig. 2(a). A more detailed modeling should account for different kinetic pathways of growth including the surface diffusion, the radial extension of nanowires and the coalescence mechanisms, as in Refs. [11,15–19]. This will allow for a better control over the entire growth process by technological parameters. Here, we present only the elementary energetic considerations to explain the formation of the fork-like structures. From the SEM image shown in Fig. 2(a), we assume that the tapered base forms at the initial growth step and then transforms into the nanowires. We consider the model geometry shown in Fig. 4(a), with the cantilever tip of the width dn , the tapered FCNS base of the length l0 that extends toward the top with the taper angle α, and n identical rectangular nanowires having the width d ¼ const and height h1 atop the base of widthd0 ¼ dn þ 2l0 tan α. The thickness of the whole structure equals Δ ¼ const. The total volume V1and the surface energy F1 of the nanofork structure in Fig. 4(a) are given by V 1 ¼ dnh1 Δ; F 1 ¼ γ p d0 Δþ 2γ v nh1 ðd þ ΔÞ:
ð1Þ
Here, γ p is the surface energy of the planar interfaces that are parallel to the top surface of the tip and γ v is the surface energy of all the vertical surfaces (see Fig. 4). While a uniform surface energy of all solid-vapor interfaces would be a good approximation for a completely amorphous and isolated FCNS, the asymmetry in surface energies could arise from the fact that the planes are oriented differently with respect to the electron beam (see Fig. 1). In any case, the nanowire fork is energetically preferred to the plane nanoscalpel structure which just continues to extend after the length l0 [Fig. 4(b)] if the surface energy F2 of the upper (dark) part of the structure shown in Fig. 4(b) is larger than F1 for the same volume. Free energy of forming the nanofork is than lower due to a smaller surface energy term for the same volume energy [12]. The volume V2 and the surface energy F2 of the tapered nanoscalpel structure of the height h2 are given by 2
V 2 ¼ ðd0 h2 þ h2 tan αÞΔ; F 2 ¼ γ p ðd0 þ2h2 tan αÞΔ 2
þ 2γ v ðd0 h2 þ h2 tan αÞ þ2γ i h2 Δ= cos α:
Fig. 1. Experimental setup for FCNS growth: (1) the electron beam, (2) the cantilever tip, (3) the underlying target with a carbon coating, (4) the influx of carbon ions, and (5) the growing carbon nanostructure.
ð2Þ
where γ i is the surface energy of the inclined surface as shown in Fig. 4(b). Taking the difference F 1 F 2 under the condition V 1 ¼ V 2 , the nanofork is preferred to the nanoscalpel when the function ! sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi d0 γ p γi 4dh tan α 1þ 1 ð3Þ þ f ðhÞ ¼ h 2 2 γ v γ v sin α d0 is negative, in which h ¼ nh1 is the effective height of the structure.
Please cite this article as: I.S. Mukhin, et al., Framed carbon nanostructures: Synthesis and applications in functional SPM tips, Ultramicroscopy (2014), http://dx.doi.org/10.1016/j.ultramic.2014.10.008i
67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132
I.S. Mukhin et al. / Ultramicroscopy ∎ (∎∎∎∎) ∎∎∎–∎∎∎
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66
3
Fig. 2. Typical SEM images of differently shaped FCNSs on a Si cantilever: (a) nanofork; (b) and (c) nanoscalpel viewed in different directions; and (d) nanorod. The electron beam energy was 30 keV in (a), 25 keV in (b) and (c) and 10 keV in (d).
Fig. 3. (a) TEM image of a nanoscalpel structure with the electron diffraction pattern shown in the insert; (b) Enlarged TEM image of a part of the nanoscalpel, the insert showing the Fourier transform of the diffraction pattern. The diffuse rings in (a) and the broadened reflexes in (b) correspond to nearly amorphous carbon.
when γv o
Fig. 4. Illustrations of nanofork (a) and nanoscalpel (b) structures showing the main parameters of the model. The volume of the dark upper parts of both structures is fixed to a constant value in the model.
pffiffiffi Eq. (3) shows that f ðhÞ ffi h const h 4 0 at large enough h. The nanowire structure is thus energetically suppressed asymptotically due to its developed sidewalls, which is a well-known feature for nanowires on infinite substrates [18]. However, if nanowires are preferred at the initial growth step when they nucleate, the initial shape can pertain at the follow-up stages because the reverse transformation to nanoscalpel requires surface rearrangement of too many atoms and should be kinetically forbidden, as in the case of self-induced GaN nanowires [12]. From Eq. (3), the function f(h) is linear at small h and is negative
d γi ; γ tan α þ d0 p cos α
ð4Þ
In the isotropic case γ v ¼ γ p ¼ γ i , this is reduced to d0 cos α o dð1 þ cos αÞ, the inequality which can be satisfied only if n cos α o 1 þ sin α (since d o d0 =n). The latter condition requires large enough taper angle α and small number of the fork teeth n. As seen from Fig. 2(a) and (b), in some cases the nanoscalpel structure is not entirely flat but rather looks like an initial fork-like shape whose development has been suppressed at an early stage. Thus, our operating conditions should be close to a threshold between the two structures. The typical parameters of our nanoscalpel and nanofork structures (over more than 10 growth experiments) are: d0 ¼500 nm and α ¼ π=6, n ¼ 4, d ¼100 nm for nanoforks. If we take the approximation of a constant surface energy, γ v ¼ γ p ¼ γ i , the condition n cos α o1 þ sin α could not be fulfilled with our experimental parameters so that the scalpel structure always wins. Therefore, within the model, the nanofork is formed due to its lower effective surface energy. This is illustrated in Fig. 5 which shows the function f(h) for the typical experimental dimensions at different values of the control parameter A ¼ γ p =γ v þ2γ i =γ v . Overall, the formation of both the fork-like and the scalpel-like FCNSs is entirely possible and depends on the surface energy values. In the isotropic case with A¼3, the scalpel shape dominates, while for A¼9 the fork formation is preferred at the initial growth step where f(h) is negative. The number of nanowires and their shape will be highly dependent on kinetic factors and require a separate treatment. We also note that nothing prevents nanowires
Please cite this article as: I.S. Mukhin, et al., Framed carbon nanostructures: Synthesis and applications in functional SPM tips, Ultramicroscopy (2014), http://dx.doi.org/10.1016/j.ultramic.2014.10.008i
67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132
I.S. Mukhin et al. / Ultramicroscopy ∎ (∎∎∎∎) ∎∎∎–∎∎∎
4
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66
Fig. 5. Graphs of f(h) obtained from Eq. (3) with d0 ¼ 500 nm, d¼ 100 nm, α ¼ π=6 and different A. For isotropic surface energies (A ¼3), f ðhÞ is positive and the nanoscalpel shape is preferred everywhere. When A is increased to 9, f(h) becomes negative at small h, where the formation of nanofork (with 4 nanowire teeth) is favored.
from growing radially, as in the case of self-induced GaN nanowires on Si [12,19], which will change the surface energetics [12] and finally lead to coalescence. Unlike InAs/GaAs and other Stranski– Krastanow islands [20] or GaN nanowires on AlN interlayer [21], our FCNS should be induced by surface energetics rather than the lattice mismatch since amorphous structures are not expected to have any epitaxial constraint with respect to Si or W.
4. Use of nanoforks in nanolitography Contact and semi-contact SPM modes of force lithography are well-known for standard Si cantilevers [22]. The optimal regime of lithography is realized under the condition P tip 4 P int 4 P material , where P material denotes the plastic limit of a given sample, P int is the pressure in the contact area of the probe with the sample, and P tip is the plastic limit of the probe tip. In practice, the process of SPM lithography is always coupled with SPM visualization using the same probe. In order to improve the resolution of a nanofork probe in the visualization regime, we need to decrease the area of interaction between the sample and the spot. To do so, we incline the long axis of the probe with respect to the substrate surface by a small angle β, as shown in Fig. 6(a). In this configuration, only one of the fork's teeth is used for visualization, while several teeth do lithography. Of course, such an inclination will modify the above criterion of the mechanical stability, but is not expected to change significantly the qualitative picture. Fig. 6(b) shows the SPM image of a laser disc, obtained with the nanofork probe grown on top of a Si cantilever, with the inclination angle β¼151. It is seen that the probe ensures a reasonable resolution for the visualization of the force lithography. Obviously, a long nanofork tip is particularly advantageous compared to the standard one for a more precise visualization of deep lithographical channels. Let us now consider a very important question of the mechanical stability of anisotropic tips under the axial compression while pressing the sample surface [23]. According to Ref. [24], the tip is mechanically stable provided that F c =Scontact o P 3 , where Scontact is the contact area of the tip with the sample surface. The critical force pffiffiffi 2 F c ¼ ð2EΔ3 dÞ=ð3 2h ð1 þ νÞ1=2 Þ ð5Þ determines the mechanical stability under the axial stress, with E as the Young's modulus, ν as the Poisson ratio, d as the tip width
(which equals d0 for nanoscalpel), Δ as the thickness and h as the height. When the long axis of the asymmetric tip is perpendicular to the sample surface, the typical geometrical parameters of our nanoforks yield the value of Fc at approximately 45 μN, which is much larger than the maximum forces used for SPM visualization (usually about 1 μN) and force lithography (less than 35 μN). Eq. (5) shows that the critical mechanical force is proportional to the tip width d and thus the elongated tip shapes are stiffer compared to nanorods. In other words, the long plane of the nanoscalpel plays the role of a stiffening plate with respect to axial compressions and bends which makes the asymmetric tips very suiw for lithography of deep and narrow channels. We mention here the commercially available AFM cantilevers with carbon nanotubes (CNTs) attached at the tips. Of course, single CNT yields a superior resolution [25,26] but is much less robust due to its low mechanical stiffness. In addition, the absence of the pronounced asymmetry of a CNT tip does not allow for a manipulation with nano-objects as will be discussed shortly. Fig. 7 shows the results of lithography of a 10 nm Au film on a polycarbonate surface, obtained with a nanofork tip. The images were obtained with the same probe in the semi-contact mode. The image in Fig. 7(a) reveals a fine-grained structure of the Au film, demonstrating a high spatial resolution of the tip. The lithographical process was performed as follows. The probe modified with the nanofork FCNS was pressed to the Au surface with F 4 1 μN and then moved along the long axis of the fork. After that, the tip was torn off the surface and retained to the initial position in the substrate normal direction, while the in-plane co-ordinate was moved away by 500 nm in each step. In this way, we were able to form the surface dashes with 500 nm spacing [Fig. 7(b)]. The dash width was found to be 100 nm, i.e., equals exactly the width of the tip. The measured dash depths were equal to 10.5 nm showing that the Au film is completely trenched [Fig. 7(c)]. Close examination of different regimes of lithography yields the following optimal parameters: the pressing force¼1.5 μN, the moving velocity¼ 2 μm/s, and the exposition time of 300 μs in each point. Under these conditions, the nanofork tip trenches through the Au film to the substrate with the minimum width of the cut. It should be noted that, on a surface of a soft polycarbonate sample and for a sufficiently low force applied, the “point” regime of lithography also produces a point spot in the Au film. This reveals a small effective contact area, defined by the inclination β. As the applied force increases, the spot starts to elongate along the direction of the long axis of the tip. To investigate the spatial resolution of the nanofork lithography, we performed the modification of a polycarbonate surface according to the template with a varying grating spacing: 150 nm, 100 nm and 80 nm. The results are shown in Fig. 8. When the spacing equals 150 nm [points C and D in Fig. 8(b)] or 100 nm (points E and F), the obtained channels are well separated. The situation changes for the 80 nm spacing (points G and H) where the two channels merge and the total depth is much larger than in points C, D, E and F. This result is expected, because the spatial resolution is limited by the width of the fork itself ( 100 nm), while points G and F relate to the two-pass lithography mode. The same procedure was repeated on the polycarbonate surface covered with a 10 nm gold layer. The lithography parameters were the following: the pressing force was 1.5 μN, the moving velocity was 2 μm/s, and the exposition time of 300 μs in each point. The grating spacing was 250 nm, 150 nm, 100 nm, 50 nm and 25 nm, as shown from left to right in Fig. 9(a). It is seen that the lines are well-separated for the 50 nm spacing and above. Obviously, the linear regime (where the contact area between the tip and the substrate defines the resolution) works only when the displacement of the cantilever occurs is aligned with its long axis.
Please cite this article as: I.S. Mukhin, et al., Framed carbon nanostructures: Synthesis and applications in functional SPM tips, Ultramicroscopy (2014), http://dx.doi.org/10.1016/j.ultramic.2014.10.008i
67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132
I.S. Mukhin et al. / Ultramicroscopy ∎ (∎∎∎∎) ∎∎∎–∎∎∎
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66
5
Fig. 6. (a) Schematic representation of the best tip positioning with respect to the substrate surface. (b) A polycarbonate surface with the pits visualized by the nanofork probe. The inclination angle β was 151 in these particular measurements.
Fig. 7. Surface lithography with a nanofork tip: (a) Atomic force microscopy (AFM) image of a 10 nm Au film on a polycarbonate surface before the lithography. (b) AFM image of a diffraction grating with the 500 nm spacing, 100 nm line width and 10.5 nm depth. (c) Cross-section along the AB segment in (b).
Fig. 8. (a) AFM image of a grating with variable spacing (150, 100 and 80 nm) on a polycarbonate surface, obtained with a nanofork probe at 230 nN force. (b) Cross section along the АВ segment shown in (a).
5. Nanosurgery of erythrocytes Nanofork probes can also be used for nanosurgery of biological objects. In particular, a partial incision of cell membranes enables an investigation of the local structure of the cell without destroying it and, on the other hand, an introduction of different substances (drugs or dyes) through the incision. Fig. 10 shows the results of the surface modification of erythrocytes dried in air for 10 min before surgery. The dissection
was performed by moving the nanofork tip with respect to the erythrocyte surface at a speed of 0.5 μm/s and different pressing forces. Below1 μN, modification is seen only on some parts of the surface. With a 2.5 μN force, the membrane surface is incised all the way, with an incision depth of about 12 nm [Fig. 10(a)]. When the force is raised further, both width and depth of the incisions gradually increases. At 6.5 μN, the membrane starts to flake off the incision line, which is seen at the edges of the modified membrane. The incision depth is larger than 25 nm, while the width
Please cite this article as: I.S. Mukhin, et al., Framed carbon nanostructures: Synthesis and applications in functional SPM tips, Ultramicroscopy (2014), http://dx.doi.org/10.1016/j.ultramic.2014.10.008i
67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132
6
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66
I.S. Mukhin et al. / Ultramicroscopy ∎ (∎∎∎∎) ∎∎∎–∎∎∎
Fig. 9. (a) AFM image of a grating with variable spacing (250, 150, 100, 50 and 25 nm) on a polycarbonate surface covered with a 10 nm gold layer, obtained with a nanofork probe. (b) Cross section along the АВ segment shown in (a).
Fig. 10. AFM visualizations (top) and the corresponding cross sections (bottom) of the modified surface of an erythrocyte, obtained with a nanofork tip with different pressing forces: 2.5 μN (a) and (d), 6.5 μN (b) and (e) and 7.5 μN (c) and (f).
reaches 150 nm [Fig. 10(b)]. At 7.5 μN, Fig. 10(c) reveals the internal structure of the erythrocyte under the membrane. The latter is clearly seen along the border of the dissected region. Higher applied forces typically result in the displacements and conglomerations of the membrane material.
6. Manipulation of nano-objects with nanoforks To perform manipulations with micro- and nano-objects by AFM, the sample is first visualized in the semi-contact mode, after which the probe is placed near the particle and the AFM is converted to the contact mode. Then, the particle is pushed by the tip to the desired spot on the surface. Returning to the semi-contact mode, the particle position is visualized again and the entire cycle is repeated if any additional positioning of the particle is required [27]. Such a method enables a precise relocation of different objects for a distance from several tens of nanometers to several micrometers. It is however
difficult to move a particle at a large distances ( 10 μm) with the standard tips, which is paramount, for example, for manipulations with biological objects and applications in nanosensors. In particular, there is an issue of a particle losing its contact with the probe and falling off the tip when meeting surface inhomogeneities of different types. These issues could be surpassed by using the nanofork tips, owing to a pronounced asymmetry of their shape. Indeed, we have not seen any problem of long-range transfer of particles with our modified probes. In particular, Fig. 11 shows the AFM images of colloidal particles with 650 nm diameter before and after moving some of them with a nanofork tip. The particles were transferred mainly along the long axis of the tip. After the first manipulation, the particles were moved for about 2 μm, as follows from Fig. 11 (b) (the moving distance is AB). The second manipulation enabled moving the particles outside the scanned area [Fig. 11(c)]. It is noteworthy that, in Fig. 11(b) and (c), we can clearly see the AFM contrast revealing the initial position of the particles [labeled A in Fig. 11(b)].
Please cite this article as: I.S. Mukhin, et al., Framed carbon nanostructures: Synthesis and applications in functional SPM tips, Ultramicroscopy (2014), http://dx.doi.org/10.1016/j.ultramic.2014.10.008i
67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132
I.S. Mukhin et al. / Ultramicroscopy ∎ (∎∎∎∎) ∎∎∎–∎∎∎
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 Q3 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
7
Fig. 11. AFM images of an ensemble of colloidal particles on a polycarbonate surface in their initial position (a), after the first (b) and the second (c) manipulation with the nanofork probe. The arrows show the moving direction.
7. Conclusions In summary, we have shown that the nanofork and the nanoscalpel FCNSs can be obtained regularly on the standard Si and W cantilevers by applying a special growth procedure involving the scanning focused electron beam that produces the flat spatial distributions of carbon ions emitted from a carbon target. This gives rise to a directional growth of elongated nanosheets, which can be either uniform (nanoscalpel) or separated into perfectly aligned parallel nanowires of 50–100 nm width (nanofork) depending on technological parameters. The preliminary energetic model is proposed that explains the formation of nanowire teeth under certain conditions. We have demonstrated the use of these nanofork carbon structures as modified SPM tips for nanolithography and visualization of deep channels, nanosurgery of biological objects, and nanomanipulation of colloidal particles. We now plan to perform more growth experiments and modeling to better illuminate the tuning knobs to produce different tip shapes, to control the number and shape of the fork teeth, and the shape transformation from nanoforks to nanoscalpels. More investigations are required into the physical properties of these carbon nanostructures as well as SPM characterization and manipulation with different nano-objects.
Acknowledgment This work was partially supported by the Government of the Russian Federation (Grant # 074-U01), a few grants of the Russian Foundation for Basic Research and the FP7 project FUNPROB. VGD gratefully acknowledges financial support of the Russian Science Foundation under the Grant 14-22-00018. The authors kindly thank Nevedomsky V.N. and Soshnikov I.P. for help with TEM studies. References [1] J.Y. Raty, G. Galli, Ultradispersity of diamond at the nanoscale, Nat. Mater. 2 (2003) 792–795. [2] H.W. Kroto, The stability of the fullerenes Cn, with n¼ 24, 28, 32, 36, 50, 60 and 70, Nature 329 (1987) 529–531. [3] H.W. Kroto, J.R. Heath, S.C. O'Brien, R.F. Curl, R.E. Smalley, C60: Buckminsterfullerene, Nature 318 (1985) 162–163. [4] A.K. Geim, K.S. Novoselov, The rise of grapheme, Nat. Mater. 6 (2007) 183–191.
[5] K.S. Novoselov, A.K. Geim, S.V. Morozov, D. Jiang, M.I. Katsnelson, I.V. Grigorieva, S.V. Dubonos, A.A. Firsov, Two-dimensional gas of massless Dirac fermions in graphene, Nature 438 (2005) 197–200. [6] X. Zhao, Y. Ando, Y. Liu, M. Jinno, T. Suzuki, Carbon nanowire made of a long linear carbon chain inserted inside a multiwalled carbon nanotube, Phys. Rev. Lett. 90 (2003) 187401. [7] S. Iijima, Helical microtubules of graphitic carbon, Nature 354 (1991) 56–58. [8] N.G. Chopra, R.J. Luyken, K. Cherrey, V.H. Crespi, M.L. Cohen, S.G. Louie, A. Zettl, Boron nitride nanotubes, Science 269 (1995) 966–967. [9] L. Qintao, Z. Dezhang, G. Jinlong, Fabrication of carbon nanowire networks by Si ion beam irradiation, Appl. Phys. Lett. 89 (2006) 053107. [10] A.O. Golubok, I.U. Popov, I.S. Mukhin, I.S. Lobanov, Creation and study of 2D and 3D carbon nanographs, Elsevier Phys. E 44 (2012) 976–980. [11] V.G. Dubrovskii, T. Xu, Y. Lambert, J.P. Nys, B. Grandidier, D. Stievenard, W. Chen, P. Pareige, Narrowing the length distribution of Ge nanowires, Phys. Rev. Lett. 108 (2012) 105501. [12] V.G. Dubrovskii, V. Consonni, A. Trampert, L. Geelhaar, H. Riechert, Scaling thermodynamic model for the self-induced nucleation of GaN nanowires, Phys. Rev. B 85 (2012) 165317. [13] M.A.Kokoreva, V.A. Margulis, M.A. Pyataev, Electron transport in a two-terminal Aharonov–Bohm ring with impurities, Elsevier Phys. E 43 (2011) 1610–1617. [14] S.A. Chivilikhin, A.O. Golubok, I.S. Mukhin, Mathematical model of nanowhisker growth under focused electron beam, NTV IFMO 66 (2010) 78–83. [15] M. Tchernycheva, L. Travers, G. Patriarche, F. Glas, J.C. Harmand, G.E. Cirlin, V.G. Dubrovskii, Au-assisted molecular beam epitaxy of InAs nanowires: growth and theoretical analysis, J. Appl. Phys. 102 (2007) 094313. [16] K. Ostrikov, H. Mehdipour, Nanoscale plasma chemistry enables fast, sizeselective nanotube nucleation, J. Am. Chem. Soc. 134 (2012) 4303–4312. [17] K. Ostrikov, H. Mehdipour, Thin single-walled carbon nanotubes with narrow chirality distribution: constructive interplay of plasma and Gibbs–Thomson effects, ACS Nano 5 (2011) 8372–8382. [18] V.G. Dubrovskii, N.V. Sibirev, Growth thermodynamics of nanowires and its application to polytypism of zinc blende III–V nanowires, Phys. Rev. B 77 (2008) 035414. [19] V.G.Dubrovskii, V. Consonni, L. Geelhaar, A. Trampert, H. Riechert, Scaling growth kinetics of self-induced GaN nanowires, Appl. Phys. Lett. 100 (2012) 153101. [20] D. Bimberg, M. Grundmann, N.N. Ledentsov, Quantum Dot Heterostructures, Wiley, New York, 1999. [21] V. Consonni, M. Knelangen, L. Geelhaar, A. Trampert, H. Riechert, Nucleation mechanisms of epitaxial GaN nanowires: Origin of their self-induced formation and initial radius, Phys. Rev. B 81 (2010) 085310. [22] M. Heyde, K. Rademann, B. Cappella, M. Geuss, H. Sturm, T. Spangenberg, H. Niehus, Dynamic plowing nanolithography on polymethylmethacrylate using an atomic force microscope, Rev. Sci. Instrum. 72 (2001) 136–141. [23] J.D. Beard, D.J. Burbridge, A.V. Moskalenko, O. Dudko, P.L. Yarova, S.V. Smirnov, S.N. Gordeev, An atomic force microscope nanoscalpel for nanolithography and biological applications, Nanotechnology 20 (2009) 1–10. [24] L.D. Landau, E.M. Lifshitz, Theory of Elasticity, Pergamon, Oxford, 1970. [25] J. Tang, G. Yang, Q. Zhang, A. Parhat, B. Maynor, J. Liu, L.-C. Qin, O. Zhou, Rapid and reproducible fabrication of carbon nanotube AFM probes by dielectrophoresis, Nano Lett. 5 (2005) 11–14. [26] N.R. Wilson, J.V. Macpherson, Carbon nanotube tips for atomic force microscopy, Nat. Nanotechnol. 4 (2009) 483–491. [27] L.T. Hansenyk, A. Kuhlez, A.H. Sørensen, J. Bohr, P.E. Lindelof, A technique for positioning nanoparticles using an atomic force microscope, Nanotechnology 9 (1998) 337–342.
Please cite this article as: I.S. Mukhin, et al., Framed carbon nanostructures: Synthesis and applications in functional SPM tips, Ultramicroscopy (2014), http://dx.doi.org/10.1016/j.ultramic.2014.10.008i
61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119
