Carbon nanotube plane fastener

The development of large-scale production processes for carbon nanotubes (CNTs) by means of catalytic chemical vapor deposition (CVD) methods has led to the recent availability of stable vertically-aligned CNT brushes.^1–3 Several applications exploiting this brush-like feature of CNTs have been reported to date; one such application is as a base for spinning yarn.^4–6 Since CNTs are highly-aligned, neighboring CNTs can easily make contact at the ends of parallel regions and enable continuous spinning. Another application goes under the popular name of “gecko” tape. Since these surfaces are formed from extended CNT tips, a surface can flexibly fasten to any rough plane by way of van der Waals (vdW) interactions.^7–9 In furtherance to these studies, we investigated whether this brush-like feature could be used in other situations. Indeed, if two brush-like surfaces are pressed together, these engage each other as a plane fastener in a manner not unlike Velcro. Related CNT gecko tape studies have demonstrated significant contact area enhancement between the adhering, although materially distinct, surfaces and thus strong adhesion between pairs of CNT brushes was expected to occur, leading to our plane-fastener concept. We have examined the adhesion properties of pairs of differently-prepared CNT brushes, and investigated the dependence on CNT density and tip shape.

CNT brushes are usually prepared by chemical vapor deposition, with Fe catalyst deposited on a silicon substrate and acetylene gas as carbon source. The CNTs prepared in this study are highly aligned normal to the Si substrate, and on average are 100–160 μm in height and 12 nm in diameter. We prepared two types of CNT brushes, type A with density 0.052 g/cm² and type B with density 0.029 g/cm². Both types were transferred from a Si substrate to a polypropylene (PP) film. The transfer, as illustrated in Fig. 1, was performed using one of two methods. The first involved shaving off a film of aligned CNTs from the substrate using a razor blade and placing it onto the PP thin film heated at ca. 450 K (Fig. 1(b)). The heating softened the PP film enabling the nanotube ends to sink slightly into the film; cooling to room temperature fixed the CNTs to the PP film to form the brush. The preparation produces aligned CNTs, with CNTs fixed in the position they had grown and the growth tips forming the exposed surface. The second method is illustrated in Fig. 1(c). Here the growth-tip surface of CNTs was embedded into the heated PP film. The Si substrate was then peeled off after cooling. In this way we obtained an alternative type of CNT-PP bilayer sheet where the base of the CNTs forms the exposed surface. Figure 2 shows top-view SEM images of the CNT-PP sheets prepared in these two ways; panels (a) and (b) show top-side surfaces prepared from specimens A and B, respectively, whereas panels (c) and (d) show bottom-side surfaces for A and B, respectively. CNTs at the top-side appeared curled, while those at the bottom-side were straight and bundled. Pairs of sheets were then positioned with top-side surfaces facing each other and engaged by pressing with a pressure in excess of 2 N/cm² parallel to the axial direction of the CNTs (normal to the sheet). The entwined sheets were then pulled apart during which time the adhesion strength could be measured using a tensile tester. We investigated the three possible surface configurations, top-top (TT), top-bottom (TB), and bottom-bottom (BB) surfaces, for each pairing of sample type. Repetitive testing in excess of ten cycles was performed for each configuration.

Figure 3(a) shows a typical plot of adhesion strength measured as a function of pulling distance. The photo of inset (i) of the figure shows the two CNT sheets were initially fully engaged. As one of the CNT sheets was pulled, the tensile stress increased reaching a maximum value at a stroke between 50 μm−150 μm. Further pulling resulted in plateau-like steps, as shown by the grey arrows in the plot, between decreasing linear regimes in tensile stress. A suggested explanation might be that the CNTs are partially forming several rope-like structures by connecting with each other, as shown by white arrows in the photo of inset (ii) in Fig. 3(a). The unraveling of these ropes one by one would result in discrete decreases in tensile stress. CNTs finally became detached at ∼500 μm, as shown by the photo of inset (iii) in Fig. 3(a), though most notably the detachment point was in excess of three times the average CNT length. The CNTs prepared for the present experiments generally had wavy features that might have been straightened under tensile stress. Figure 3(b) shows the results of repetitive testing with three different surface configurations (TT, TB, and BB), employing the two CNT specimens of different densities (A and B). Thus, there were six samples in total. Ten repetitions of the tensile tests were conducted for each sample pairing, and the maximum tensile stress values for every cycle were plotted. The TB configuration of type B showed the strongest adhesion of all six samples, with a maximum measured value of approximately 26 N/cm². This is higher than that of conventional fabric hook-and-loop fasteners (∼4N/cm²). Adhesion strength decreased with each repetition for every sample, and was in the range 0.5–3.0 N/cm² over the ten cycles. This is comparable or slightly higher than the range of 0.3−0.4 N/cm² for the Post-it adhesive by 3M^©. SEM images of the CNTs taken after testing are shown in Fig. 4, and generally show very flat surfaces where CNTs have been leveled. It is plausible that adhesion strength progressively weakened as CNTs flattened over the entire adhering area. Over repeated cycles, fewer CNTs remained extended from the surface to engage with the opposing surface.

The CNT fastening system has a different adhesion mechanism to conventional systems, such as the hook and loop type. Adhesion in the CNT system stems from the van der Waals (vdW) interactions or sliding friction between individual CNTs. In a previous in-situ study, the sliding-friction force for two CNTs connected in parallel with 100 nm overlap length was measured at 40 nN. The vdW force between the two parallel CNTs was 1 nN and independent of overlap length.¹⁰ Assuming an ideal CNT fastening system, then all CNT tips would connect to parallel neighbors as two CNT-sheets engaged. In such a case, a very high adhesion strength of up to 1kN/cm² could potentially be exhibited, as the average CNT density estimated from the average length and diameter as well as density of type B specimen (0.029 g/cm²) was ca. 2.8 × 10¹⁰/cm². In the present experiment, values of 0.5−26 N/cm² were observed, suggesting that only ∼2.2 % of CNTs contributed at maximal adhesion (viz. 26 N/cm²). The CNT density over the contact surface area regulated the adhesion properties of the whole system. We can estimate the amount of work by integrating the plot in Fig. 3(a), here showing a maximum adhesion of 2.1 N/cm², was ca. 0.40 J/cm², which is 4×10⁴ times greater than the surface energy of the c-plane of graphite (1.1×10^-4 J/cm² at 20 °C).¹¹ The large energy density is evidence for the large effective surface area contributing to the adhesion, even if only a few percent of CNTs contributed to adhesion. Surface features of the CNT-PP sheet are crucial if the contact area of individual CNTs is to be maximized In the present study, we investigated two surface features, namely density and CNT tip shape. Regarding density, specimen B with lower density showed higher adhesion except in the BB pairing, as shown in Fig. 3(b). In comparing tip shapes of top-side surfaces between specimens A and B, CNT tips in specimen A with the higher density tended to weakly agglomerate to partially form dense regions resembling tufts, typically with size of about 1–2 μm. This resulted in a limited number of CNTs around the tufts able to be engaged with others, and might be the reason for a lower effective adhered surface area.

The difference in CNT tip shapes between the top and bottom-sides were compared. As shown in Figs. 2(a) and 2(b), the former had curled tips whereas the latter had straighter tips. The BB combination exhibited greater adhesion at around 10 N/cm² for the first few cycles, whereas the TT combination exhibited very weak adhesion at less than 3N/cm². Although CNT tips at the bottom-side formed aggregates, these also formed a significant number of small brushes as seen in Figs. 2(c) and 2(d). These were probably formed as CNTs were peeled from the silicon substrate, due to vdW interactions between tips from static electrical or elastic forces present during sample preparation. The CNTs that formed each brush were relatively straight on the bottom-side, and so could have deeper entwining length, resulting in a larger total contact area. The TB configuration showed higher adhesion than the BB configuration for specimen B. If CNTs at the top surfaces were isolated, then an increased contact probability should have been expected.

In summary, the adhesion of CNT brushes was examined. Pairs of differently prepared CNT-PP sheets were intertwined and then under tensile stress separated. The adhesion strength strongly depended on the surface features of the CNT-PP sheets and the CNT entanglement of the two surfaces. CNT plane fastener systems could be applicable in situations where conventional fasteners or adhesive materials might not be able to be used. For example, these can be used under very high temperatures and in contamination-free environments such as in clean rooms or vacuums because they are made of carbon and so do not contain organic or volatile adhesives. CNT brushes can be synthesized on a micron-sized scale at selective areas by photolithographic techniques. These are potentially useful if attaching small components in micron-sized electronic devices as CNTs can effectively transfer heat from circuits to heat sinks because of their high thermal conductivity. CNT fasteners can potentially be used as conductive adhesives when affixed to conductive materials, although the present study employed only non-conductive PP films as base support. More research is required to better understand the adhesive properties of CNT brushes in fastening systems. Structural dependencies on length, diameter, and surface defects need to be examined along with physical parameters such as loading pressure.

The authors thank Dr. T. Nagasaka for discussion and technical support concerning specimens used in this study. We also thank Dr. Y. Maeno and Mr. S. Tabuchi for technical support with adhesion measurements. This work was partially supported by the Osaka Prefecture Collaboration of Regional Entities for the Advancement of Technological Excellence, JST (CREATE-OSAKA).