Impact of Manufacturing Defects on Carbon Nanotube Logic Circuits

As CMOS technology enters the nanoelectronics realm (tens of nanometres and below), where quantum mechanical effects start to prevail, conventional CMOS devices are meeting many technological challenges for further scaling. This situation has motivated the emergence of a variety of new nanoelectronic devices [1] [2]. However, as new generations of nanodevices are developed, we become less familiar with their fault mechanisms and the causes behind their failures [3]. On the one hand, the nature of the materials and the physical phenomena used in these technologies are very different from current CMOS. On the other, it is widely acknowledged that the small sizes of resulting devices will cause higher levels of manufacturing defects than those of current CMOS solutions. In addition, in-service (transient and permanent) faults will have to be dealt with [4]. This situation justifies why reliability has become (and will be) a big challenge in the design of current (and future) nanoelectronic devices and architectures. The confident use of these emerging technologies relies on our capacity to better understand their fault mechanisms, and our ability to develop related fault models. Those fault models can be considered as a step forward towards the dependability assessment of emerging architectures for the definition of new and efficient fault mitigation techniques [5]. Among the wide set of new emerging nanoelectronic devices, 1D structures, as Carbon Nanotubes (CNTs) and Silicon Nanowires (SiNWs), are among the most promising for the development of logic circuits [6]. They present the best values for a number of factors that evaluate their potential use, like scalability, gain, operational reliability, performance, room temperature operation, energy efficiency, and CMOS technological and architectural compatibility. In addition, programmable logic array reconfigurable architectures are suggested for 1D structures. As we already analysed SiNWs [7], this study will focus on CNTs. CNTs have mechanical and electrical properties that make them very attractive as nanoelectronic wires and devices. Due to their structure, CNTs have an extraordinary strength and can behave as metallic wires or semiconductors, depending on their chirality and diameter [8] [9]. Metallic wire CNTs can offer superior electrical properties to SiNWs or copper [10]. As semiconductors, CNTs allow the design of field effect transistors (CNTFET). They have a MOSFETs-like structure (which is depicted in Figure 1) but present some potential key advantages: smaller channel and ballistic transport of electrons. Despite the good CNTs properties, the main hurdles concern their fabrication: selectively controlling their electronic properties (e.g. metallic or semiconducting) and placement [8].