TOP-DOWN TECHNIQUES (LITHOGRAPHY) FOR MAKING QUANTUM WIRES
Limin Ji, Lingyun Miao
ECE Dept., University of Rochester
NY, USA 14621
I. Introduction
This paper will mainly focus on the fabrication techniques of the quantum wire. To be specific, the techniques introduced here are all top-down methods. Before the details of the techniques, there are several basic questions that have to be discussed at the beginning. The first question is what is nanotechnology?
Fig 1. A parallel-shaft speed reducer gear which is one of the
largest nanomechanical devices ever modeled in atomic detail. [1]
As it is defined in the encyclopedia, Nanotechnology is “a field of applied science and technology covering a broad range of topics”. The major work is the control of matter on a scale smaller than 100 nanometers, by using the devices on the same length scale. This is a huge area, which includes colloidal science, chemistry, applied physics, material science, mechanical engineering and electrical engineering.
The stimulus for nanotechnology is due to the recovered interest in colloidal science and invention of nano-scale analytical tools such as the atomic force microscope (AFM) and the scanning tunneling microscope (STM). Also, the occurrence of refined processes, such as electron beam lithography and molecular beam epitaxy, realize the deliberate manipulation of nanostructures and thus make the observation of novel phenomena possible. It has already applied in modern use, like the design of computer chip layouts based on surface science.
The history of nanotechnology is pretty young. The concept was first appeared in “There’s Plenty of Room at the Bottom”, a talk given by physicist Richard Feynman at an American Physical Society meeting at Caltech on December 29, 1959 [2]. “ Feynman described a process by which the ability to manipulate individual atoms and molecules might be developed, using one set of precise tools to build and operate another proportionally smaller set, so on down to the needed scale”. In 1974, Professor Norio Taniguchi from Tokyo Science University defined the term “nanotechnology in a paper with title “On the Basic Concept of ‘Nano-Technology”. He defined the term as “mainly consists of the processing of, separation, consolidation, and deformation of materials by one atom or one molecule”. In 1980s’, Dr. K. Eric Drexler completed the definition of nanotechnology through his speeches and books by elevating the technological significance of nano-scale phenomena and devices. Here comes the second question, why does nanotechnology so important?
First of all, nanotechnology is essential to the technology involved in developing the next generation of electronics; besides, it have already done much for science. The famous Moore’s law predicted that computer power will double every 18 months. However, it will probably end between 2010 and 2020 [3]. Since electronics is a $200 billion/year industry, it is crucial “where it will go and how the Moore’s Law curve can be prolonged” [3]. Nanotechnology might be the solution. As it is predicted by the Semiconductor Industry Association, dimension of the extending current technology will be reduced to 50*50 nm till 2010. One example is the Pentium chip. It now contains 40 million transistors and will increase up to 1.5 billion in ten years. Its dimension will reduce to 10*10 cm with a clock speed of 10 GHz and power consumption 175 watts. By then, lithography will be reaching the physical limits. A possible future way is making transistors from individual molecules and then be ‘sprinkled’ onto a pre-prepared surface [3]. Now we face the third question, how far has it gone so far?
Mihail Roco from U.S. National nanotechnology Initiative stated four generations of nanotechnology as shown below:
Fig. 2. Four generations of nanotechnology. [1]
As Roco’s statement, we just entered the third generation of nanotechnology, which are the systems of nanosystems. In the current generation, nanotechnology will have dramatic impact on almost all industries and all areas of society. It is also predicted that nanotechnology would be able to provide better built, longer lasting, cleaner, safer and smarter products for the home, for communications, for medicine, for transportation, for agriculture and for industry in the near future. Quantum Wire, which is one of the basic products, plays a key role in this area. In next section, we will focus back to the topic, the quantum wire and its fabrication technologies. Some background information about the quantum wire will be provided.
II. Background
In Science and Technology Dictionary, Quantum Wire (QWR) is defined as “a strip of conducting material about 10 nanometers or less in width and thickness that displays quantum-mechanical effects such as the Aharanov-Bohm effect and universal conductance fluctuations”. The basic difference between quantum wire and the classical wire is that quantum information cannot be copied; instead, it must be “transported-destroying the information at the source and re-creating it at the destination” [4].
Inside quantum wire, “electrons are confined to a narrow one dimensional channel with motion perpendicular to the channel quantum mechanically frozen out” [5]. Modern semiconductor technologies could be applied here to fabricate the wires, like electron beam lithography and cleaved edge overgrowth. Further more, carbon nanotubes are “self assembled” quantum wires, which has fascinating electronic properties, light weight, small diameter, low chemical reactivity and high tensile strength. The main drawback is the cost. Fig. 3 shows an illustration of carbon nanotube and Fig. 4 is the actual image.
Fig. 3. Illustration of carbon nanotube from www.spacedaily.com/news/nanotech-05zn.html.
Fig. 4. A carbon nanotube between two electrodes from http://www.mb.tn.tudelft.nl.
There are two approaches to produce quantum wire structures, top-down and bottom-up. Roughly speaking, the former could be regarded as modern analogues of ancient methods which develop features down to the sub-100 nm scale. The realization of this type based on the invention of electron-beam writing and advanced lithographic techniques that adopt extreme ultraviolet or hard X-ray radiation [6]. On the contrary, bottom-up methods begin with atoms or molecules to build up nanostructures, in some cases through smart use of self-organization. Fig. 5 shows the difference of those two approaches. In the rest of this section, we will shortly introduce bottom-up methods and then focus on three different types of top-down methods. The specific techniques will be explained in the third part, the main part of this paper.
Fig. 5 Two approaches to control matter at the nanoscale. [6]
For the bottom-up methods, they compile atoms and molecules into the smallest nano-structures with typical dimensions of 2 to 10 nm through carefully controlled chemical reactions, which makes this kind of methods less expensive as the lithographical methods [6]. Sometimes, self-assembling of atoms and molecules is applied to achieve complex structures. However, it is restricted to relatively simple systems. In order to get complex systems, hierarchical self-assembly is applied where “the products of one self-assembly step is a base for the next one” [7]. In addition, biologically inspired self-assembly is playing an important role in nanotechnology. For example, the formation of peptide nanotubes uses biological self-assembling of cylindrical octapeptides as ion channels. Also, the regular and natural nanochannel formation of zeolites could be applied to study dynamical behavior of water adsorbed in bikitaite or for the synthesis of novel materials [7]. In conclusion, the bottom-up methods mainly use naturally formed and rather well-defined structures. As a result, they could not generate designed, interconnected patterns as the produced nanostructures are randomly positioned. Therefore, top-down techniques are required in the manufacturing of longer, better controlled and interconnected structures.
Top-down technologies could be divided into three categories based on the analogy with micromachining: bulk-/film-machining, surface-machining and mold-machining.
In bulk-/film-machining, two ways are applied to make the channel; one is etching trenches in the substrate wafer and the other is doing in the film deposited on the substrate. Usually standard photolithography is applied here and then wet or dry etching is processed on the substrate “in the case of substrate etching and usually chemical etching of the film in the alternative approach” [7]. To close the formed structure, another wafer is bonded on top of the structured substrate or film as shown in Fig. 6.
Fig. 6 Bulk-/film-machining [7]
In surface-machining, first step is putting a bottom layer on the wafer then depositing the sacrificial layer and its patterning. After that, the top layer is deposited on top of the sacrificial layer and patterned (often with irrigation holes, which provide the access to the sacrificial layer. Finally, the nanochannel is formed by removing, i.e., etching the sacrificial layer leaving the bottom and the top layer to form the walls of the nanochannel. The bottom layer, used to form the channel of one material (the same material as the top layer), is not always required. Isotropic sacrificial etching is crucial in nanotechnology because the nanostructures are very sensitive. Dry etching is preferred here since “the nanostructures can be damaged during wet etching by the drag forces during wafer handling and the possibility of sticking of nanostructures during drying” [7]. Furthermore, advantages of this method are presented by the uniformity of the channel height and insensitivity to particles, which would disrupt bonding.
Fig. 7 Surface-machining [7]
In mold-machining, the mold in the inverse shape of the desired structure is formed first by being filled with a structural material. Then the mold can be etched or removed leaving the desired structure behind. It is usually achieved by soft lithography. In soft lithography, photolithography or electron-beam lithography (EBL) is used to produce a pattern in a layer of photoresist on the surface of the silicon wafer to make the mold. Then the mold will be covered by a liquid precursor to poly-dimenthylsiloxane (PDMS) and cured into the rubbery solid. The PDMS stamp is then peeled off the master and further used in different ways to make nanostructures. The process is shown in Fig. 8 as below.
Fig. 8 Mold-machining with soft lithography [7]
There are several other ways to classify those lithography methods but we have no space to discuss them all. However, we did selected 4 typical technologies to discuss in the following sections with very detailed information, beginning with the introduction of lithography history in section three.
III. Lithography Techniques Overview
Generally speaking, lithography can be any technique which can be used to print on a smooth surface. There have been quite a lot of techniques falling into this category since the original invention by Alois Senefelder in 1798. However, unless specially specified, today it usually refers to photolithography, a micro-fabrication technique used to make Ultra Large Scale Integrated (ULSI) circuits and micro-electro-mechanical systems (MEMS).
Photolithography generally uses a pre-fabricated photomask or reticle as a master from which the final pattern is derived. It has been widely regarded as the cornerstone of modern semiconductor industry. However, traditional photolithography meets its resolution limit and other difficulties when the semiconductor pattern feature size approaches the nanometer region. There are two endeavors dedicated to solve this problem. First, considerable research has been done to save this most commercially advanced lithography technology by extending its resolution into sub-100 nm range. Second, other lithography techniques have been proposed and developed over the last decade. As a consequence, a large number of promising microlithographic and nanolithographic technologies exist or are emerging, including electron beam lithography (EBL), nanoimprint lithography (NIL), interferometric lithography (IL), X-ray lithography, extreme ultraviolet lithography (EUVL), and scanning probe lithography (SPL). Some of these techniques have been used successfully in small-scale commercial and important research applications. Some, for example EBL, are even capable of much higher patterning resolution (sometime as small as a few nanometers).
In this documentation, we will first discuss several tricks implemented in traditional photolithography to make this “old” technique suitable for semiconductor quantum wire fabrication. Then EBL, NIL, and SPL are selected to be reviewed because of their representative properties and the length limit of this article.
IV. Traditional Photolithography with Smart Tricks
V-groove heterostructure patterning
As early as in 1987 a heterostructure patterning technique was reported [8], which resulted in near-ideal quantum wire interfaces suitable for stimulated emission from 2-D quantum-confined carriers at room temperature [9]. In this method, V-grooves are first formed on a (100) GaAs substrate. These V-grooves are oriented along the [01¯1] direction using conventional photolithography and wet chemical etching. Multiple-layer heterostructure is grown on the patterned substrate (Fig.9 (a)). Since the AlGaAs cladding layers grow to form a very sharp corner between two {111} crystal planes. The active GaAs quantum well (QW) grows faster along the [100] direction, which results in the formation of a crescent-shaped QW at the bottom of the groove (Fig.9 (b)). The lateral tapering in the thickness of the QW crescent provides lateral variation in the effective band gap due to the increase in the carrier confinement energy with decreasing QW thickness [8]. This results in a 2D potential well which confines the electrons and holes to a quasi-1D quantum wire [9].
Fig.9 Cross section of GaAs/AlGaAs quantum wire heterostructure.
(a) Schematic illustration, (b) dark-field TEM (After Ref. [9])
Quantum wire lasers fabricated using this technique have been demonstrated successfully (Fig.10) [10]. However, this fabrication method has limited application because of the non-planar nature of the processing.
Fig.10 Light vs. current characteristics of a quantum
wire laser with 350 μm cavity length (After Ref. [10])
Double patterning
One of the most common types of double patterning is double-exposure patterning. This technique uses the precise control (at the order of nanometers) of the actual mask position by piezoelectric elements implemented in advanced steppers [11]. It can be employed to efficiently reduce the minimum feature sizes in periodic structures. In this method, the photoresist is exposed a first time with a reduced exposure dose (about 60-80 % of single exposure). Then, the photomask is shifted a certain defined amount (20-30 nm) and the sample is exposed a second time with the same reduced dose. Such a process is only possible if the stepper can have a nanometric control of its reticle movement. Accordingly, the resist feature size is, to first order, determined by the overlap of the masked regions and arbitrarily small resist feature sizes can be produced by this method.