Objective Carbon nanotubes (CNTs) have become an essential electronic material to replace silicon materials in the post-Moore era due to their unique electrical properties and one-dimensional nanostructures. The micro/nano electronic devices fabricated by CNTs have the advantages of small size, high speed, and low power consumption. The type of effective method to be used to achieve a reliable and effective connection between CNTs and metal electrodes has always been a difficulty and key point in the construction of CNTs electronic devices. To achieve an effective and reliable connection between CNTs and metals, a series of connection technologies have been proposed, such as metal welding, local annealing, ultrasonic welding, electron beam, and ion irradiation. However, some additional elements, such as graphitized carbon, hydrogen, and other solders, are introduced into the interface between CNTs and metal, which affects its connection performances. The process of some methods is complicated, and they required accurate positioning. Thus, an effective and reliable connection technology between CNTs and metal electrodes on a large-scale without damaging the metal electrodes or other structures is highly needed. The laser-processing technology is widely used to manufacture electronic devices with the advantages of high peak power, noncontact processing, and good controllability. Femtosecond pulse laser is considered a cold processing technology since its pulse width is less than the cooling time of the electron. It can avoid damage to the metal electrode structure caused by heat accumulation. Besides, it is an ideal high-energy beam interconnection technology. We use femtosecond pulse laser irradiation technology to realize an effective and repeatable connection between Multi-walled CNTs (MWCNTs) and different metal electrodes (Au and Ni), which provides a certain experimental basis for subsequent large-scale preparation of high-performance CNTs field-effect transistors.
Methods In this experiment, mass concentration of 0.1 mg/mL sodium dodecyl sulfate is applied as a surfactant, and the MWCNTs powder is dispersed uniformly and stably in an aqueous solution using ultrasonic vibration. Spin coating is adopted to deposit the MWCNTs after laying electrodes. In this process, the metal electrode is the carrier of MWCNTs and test probe in the subsequent electrical performance test. Thus, the metal electrode should have a larger area for electrical performance test and smaller channel width for MWCNTs and metal contact. The morphology and size of the electrode are designed. The electrode is a rectangle of 200 μm×200 μm, and the channel's width is 8 μm. The required electrode group is obtained by photolithography, as shown in Fig. 2(a)--Fig. 2(b). Further, MWCNTs are deposited on the metal electrode by the spin coating process with a mass concentration of 0.005 mg/mL MWCNTs dispersion. In this experiment, scanning electron microscope is used to characterize the intrinsic structure, metal electrode structure, and metal morphology of MWCNTs. The electrical properties are tested using a semiconductor device analyzer.
Results and Discussions Femtosecond laser irradiation technology can achieve an effective and repeatable connection between MWCNTs and metal electrodes. The results show that there is no linear relationship between the laser power and irradiation time. Considering MWCNTs and Au as examples, it can be divided into three stages. The first is the nonaction stage. In this stage, even if the laser irradiation time is continually increased, the effect on the Au surface is ignored. The second is the selective modification stage. When the laser power is increased to 220 mW and the irradiation time is set at 60 s, the local plasma enhancement effect between MWCNTs and Au surface modifies the metal surface. When laser irradiates at the Au surface, the free-electron in Au collides with the photon of laser inelastic. Then, the free-electron absorbing photon energy migrates to a high-energy level, which increases the lattice thermal shock. Further, the local area is heated and softened, and micro molten pool formed MWCNTs are “embedded” in the metal electrode, forming a good “embedded” connection (Fig. 4(b)). The Au electrode will be ablated at 65 s in 220 mW. The last stage is material removal. When the laser power increases to 235 mW, the electrode surface will be damaged and ablated in different degrees within 50 ms (Fig. 4(d)). The same state also occurred in the Ni-MWCNTs structure. The results of the electrical test show that the contact resistance between MWCNTs and Au or Ni has been greatly decreased (Fig. 9), indicating that the connection is effective and repeatable.
Conclusions The effects of processing parameters of femtosecond pulse laser, such as irradiation, time, and laser power, on the morphology of different metal electrodes and MWCNTs, are investigated experimentally. When the laser power is 220 mW, and irradiation time is 30 s, an embedded connection formed between Au and MWCNTs. When the laser power is 28 mW and the irradiation time is 30 s, the cladding connection formed between Ni and MWCNTs. Under the same laser power, as the irradiation time continues to increase, the metal electrode surface would be ablated, and the structure of the electrode would be destroyed. The MWCNTs deposited on the electrode surface would be peeled off by the laser shock wave. The contact resistance of Au-MWCNTs-Au structure is reduced from 454--658 kΩ to 78.9--397 kΩ and that of Ni-MWCNTs-Ni structure is reduced from 505--612 kΩ to 21.1--64.6 kΩ. The latter structure is reduced by an order of magnitude, verifying the effectiveness of the connection. It also show that the bonding force and wettability of metal to MWCNTs influence electronic transport before and after interconnection.
