Page 49 -
P. 49

3-4                                                              MEMS: Design and Fabrication


             masks by Niépce in 1822, chemical milling in combination with lithography became a reality, and a new
             level of tolerances were within reach. The major modern applications of lithography-based chemical
             milling are the manufacture of printed circuit boards, started during the Second World War, and by 1961
             the fabrication of Si-integrated circuitry. Photochemical machining is also used for such precision parts
             as color television shadow masks, integrated circuit lead frames, light chopper and encoder discs, and dec-
             orative objects such as costume jewelry [Allen, 1986]. The geometry of a“cut” produced when etching sil-
             icon integrated circuits is similar to the chemical-milling cut of the aerospace industry, but the many
             orders of magnitude difference in size and depth of the cut account for a major difference in achievable
             accuracy. Accordingly, the tolerances for fashioning integrated circuitry are many orders of magnitude
             more precise than in the chemical milling industry.
               In this chapter, we are concerned with lithography and chemical machining used in the IC industry
             and in microfabrication. A major difference between these two fields is in the aspect ratio (height-to-
             width ratio) of the features crafted. The IC industry deals with mostly very small, flat structures with
             aspect ratios of 1to 2. In the microfabrication field, structures typically are somewhat larger, and aspect
             ratios might be as high as 400.
               Isotropic etching has been used in silicon semiconductor processing since its introduction in the early
             1950s. Representative work from that period is the impressive series of papers by Robbins and Schwartz
             [Robbins and Schwartz, 1959, 1960; Schwartz and Robbins, 1961, 1976] on chemical isotropic etching and
             Uhlir’s paper on electrochemical isotropic etching [Uhlir, 1956]. The usual chemical isotropic etchant
             used  for  silicon  was  HF  in  combination with  HNO with  or  without  acetic  acid  or  water  as  diluent
                                                              3
             [Robbins and Schwartz, 1959, 1960; Schwartz and Robbins, 1961, 1976]. The early work on isotropic etch-
             ing in an electrochemical cell (i.e., electropolishing) was carried out mostly in nonaqueous solutions to
             avoid  black  or  red  deposits  that  formed  on  the  silicon  surface  in  aqueous  solutions  [Halloas, 1971].
             Turner showed that if a critical current density is exceeded, silicon can be electropolished in aqueous HF
             solutions without the formation of those deposits [Turner, 1958].
               In the mid-1960s, Bell Telephone Laboratories started work on anisotropic Si etching in mixtures at
             first of KOH, water, and alcohol and later of KOH and water. The need for high-aspect-ratio cuts in sili-
             con arose in the fabrication of dielectrically isolated structures in integrated circuits such as those for
             beam leads. Both chemical and electrochemical anisotropic etching methods were pursued [Stoller and
             Wolff, 1966; Stoller, 1970; Forster and Singleton, 1966; Kenney, 1967; Lepselter, 1966, 1967; Waggener,
             1970; Kragness and Waggener, 1973; Waggener et al., 1967; Waggener et al., 1967; Bean and Runyan,
             1977]. In the mid-1970s, a new surge of activity in anisotropic etching was associated with the work on
             V-groove and U-groove transistors [Rodgers et al., 1977; Rodgers et al., 1976; Ammar and Rodgers, 1980;
             Smith, 1954]. The first use of Si as a micromechanical element can be traced to a discovery and an idea
             from the mid-1950s and early 1960s respectively. The discovery was the large piezoresistance in Si and Ge
             by Smith in 1954 [Smith, 1954]. The idea stemmed from Pfann et al. (1961), who proposed a diffusion
             technique for the fabrication of Si piezoresistive sensors for stress, strain, and pressure. As early as 1962,
             Tufte et al. (1962) at Honeywell followed up on this suggestion. By using a combination of awet isotropic
             etch, dry etching, and oxidation processes, they made the first thin Si piezoresistive diaphragms, of the
             type shown in Figure 3.1, for pressure sensors [Tufte et al., 1962]. In 1972, Sensym became the first to
             make stand-alone Si sensor products. By 1974, National Semiconductor Corporation in California car-
             ried an extensive line of Si pressure transducers in the first complete silicon pressure transducer catalog
             [National Semiconductor, 1974]. Other early commercial suppliers of micromachined pressure sensor
             products were Foxboro/ICT, Endevco, Kulite, and Honeywell’s Microswitch. Innovative nonpressure-sen-
             sor micromachined structures began to be explored by the mid- to late 1970s. Texas Instruments pro-
             duced a thermal print head in 1977 [Texas Instruments, 1977]. IBM produced ink-jet nozzle arrays in
             1977 [Bassous et al., 1977].Western Electric made fiber optic alignment structures in 1974 [Boivin, 1974],
             and Hewlett Packard made thermally isolated diode detectors in 1980 [O’Neill, 1980]. Many Silicon Valley
             and California based microsensor companies (Inset 3.2) played and continue to play a pivotal role in the
             development of the market for Si sensor products. However, from the mid-1990s on, many of the Silicon
             Valley MEMS companies became BIOMEMS oriented, and Si was no longer the preferred substrate.



             © 2006 by Taylor & Francis Group, LLC
   44   45   46   47   48   49   50   51   52   53   54