MRS Bull. Current market outlook — Schafrik, R. Reed, R. The Superalloys Cambridge Univ. Press, Giamei, A. Pollock, T. Power 22 , — Kawagishi, K. In Superalloys eds Huron, E. Elliott, A. A 35 , — Konter, M.
In Superalloys eds Pollock, T. M et al. A 47 , — Gilchrist, A. In Structural Intermetallics eds Hemker, K. National Research Council. Jain, A. APL Mater. Sato, J. Science , 90—91 Makineni, S. Acta Mater. Suzuki, A. Xue, F. Scripta Mater. Dingreville, R. The matrix functions to bind the reinforcement together and protect it. A composite matrix will be one of three types: polymer, metal or ceramic. Norplex-Micarta specializes in a subset of polymer matrix composites called thermosets. Thermosets undergo a high-temperature curing process by which the chemical structure of the polymer is irreversibly cross-linked.
Therefore, thermosets do not melt after curing, as opposed to thermoplastic materials, which do melt. The reinforcement is considered the primary contributor to the strength and stiffness of the composite. Potential reinforcements that can be effectively utilized are nearly infinite. Common reinforcement materials include paper, cotton fabrics, glass, aramids, nylon and carbon fiber. Other materials such as virgin PTFE or rubber can be incorporated into the composite to achieve specific design objectives.
Norplex-Micarta produces pre-preg from several different resin systems such as epoxy, phenolic, melamine, and silicone. These resin systems can be modified through the use of various additives to modify their behavior. These additives can make the material semi to fully conductive, increase the hardness, increase the toughness, or reduce the coefficient of friction, amongst other possible enhancements.
Reinforcements are virtually endless.
From the choice of the fiber, to woven, non-woven, and stitched fabrics, and various hybridizations and combinations of the myriad of inputs, designers have a wide range of potential inputs that can be used to optimize a material for their specific application. Various reinforcements such as carbon, glass, and aramids that can be used alone or in hybrid systems to create new materials that are predictable, reliable, and repeatable.
This can effectively reduce the uncertainty of the measurement error associated with loading. The disadvantage is that relating the measured moment and surface strain directly to the stressstrain curve is difcult. The experimental procedure for microbend experiments is illustrated schematically in Figure 3. As-deposited LIGA Ni beams of different thicknesses prepared by TEM sample preparation methods were bent around a series of tungsten bers of different diameters.
The nal curvature of the deformed beams was measured after the elastic springback that occurred upon the release of the moment applied to the beams. Since the unloading is strictly elastic and the Youngs modulus of LIGA Ni was known, the change in curvature due to elastic springback was used to determine the applied bending moment, M.
The moment curvature relationship of the elastic springback is illustrated in Figure 3. An image acquisition system was utilized to capture the images of specimens during the microbend test to determine the in situ curvature of the beams. From Stolken, J. The images of the deformed beam were analyzed using image correlation and curve tting techniques to determine the nal curvature. Traditionally, hardness is dened as the ability of a material to resist penetration by an indenter.
However, the value of hardness is affected by many factors such as yield strength, strain hardening, surface roughness, and material pile-up and sink-in around the indentation making the interpretation of the results rather complicated. Recently, with the advancement of depth-sensing indentation techniques the indentation depth can be monitored continuously as the load is applied that are commercially available, additional information such as Youngs modulus, can be extracted from the load-displacement curve.
A typical load-displacement curve is shown in Figure 3. E and n are Youngs modulus and Poissons ratio for the specimen, and Ei and ni are the same parameters for the indenter. So if the contact area at the peak load is known, the reduced modulus can be calculated from the measured stiffness.
To determine the area functions for each tip, a series of indents at various contact depths were performed on a standard fused silica specimen and the contact area calculated using Figure 3. A plot of the computed area as a function of contact depth is plotted and a tting procedure is employed to t contact area vs. The value of 3 depends on the geometry of the indenter [10,60]: 3 Z1 for at punch, 3 Z 0.
The hardness of the material is then dened in the usual way  as the mean pressure exerted by the indenter at maximum load. This gives HZ Pmax A 3. Note that the above approach is the classical OliverPharr approach for depth-sensing indentation experiments. However, some modications are needed to account for the effects of tip rounding and Poissons ratio [14,49,53], as well as material pile-up effect on the estimation of contact depth, and thus, contact area [8,15,49,55].
Other modications involve the use of the so-called b and g functions and were also discussed in the literature [16,21,30,43,49]. The force and displacement resolution of the Triboscope system are 1 nN and 0. The noise oor for the force is nN, and the noise oor for the displacement is 0. A sharp North Star indenter a three-sided cubecorner tip and a Berkovich indenter a three-sided pyramidal tip were used in the hardness measurements. Peak load ranges between 50 and 11, mN were applied in an effort to study the effects of impression size from a few microns to tens of nanometers.
Contact mode atomic force microscopy scans were also obtained before and after the indentation tests. Such contact mode AFM images were used for two purposes: rst, to ensure that the indentations were only performed on spots with relatively low roughness. This minimized the possible effects of RMS on hardness measurement . RMS, the root-mean-square height of the surface features, is one of the measurements for surface roughness.
Indents were only introduced into regions with RMS! Also, a load-displacement curve was recorded during each test. Finally, from known area functions of the calibrated tips, hardness and Youngs modulus could be calculated using a modied OliverPharr approach as discussed before . In addition, some special cautions were taken in the experiments to ensure that the results provided a true representation of property measurements.
Also, the tests were done in a quiet room with the temperature maintained at a constant C. An enclosure was also placed around the nanoindentation system to minimize the possible effects of airow currents. Furthermore, relatively short holding periods were used to reduce the possible effects of thermal drift on the measured nanoindentation hardness data. Experimental procedures have already been discussed, and the details of the material processing and characterization of microstructures and microtextures have been published elsewhere [,72].
Thus, in this section, we will focus on a review of recent theoretical developments, as well as some experimental ndings for size-dependent plasticity and the fracture and fatigue behavior of LIGA Ni thin lm materials. Please note that the thickness of these thin lm samples varies from tens of microns to s of microns. This range of thicknesses was chosen to account for the needs of performing mechanical testing on specimens with dimensions that roughly scale with the size of typical LIGA components [20,71].
The term plasticity refers to the permanent shape change of a deformed material, i. Plastic deformation can be accomplished by the diffusion of atoms, movement of dislocations, or homogeneous shear twinning, martensitic transformations. The stress needed to cause irreversible deformation is dened as the ow stress. The theory of plasticity has been studied extensively over the last century and there is a signicant body of literature on deformation mechanisms in bulk solids [31,58,73]. However, as the dimensions of new technologies such as MEMS devices decrease, the understanding is more limited.
For example, it is not clear how materials plastically deform at the micron- and nanoscales. This has stimulated strong interest in the materials science and mechanics communities to develop relevant theories and experiments concerning the sizedependent plasticity of materials [3,39,59]. Moreover, a very careful study of plasticity at small scales is also required for the design of a large number of LIGA Ni MEMS or other metallic-based MEMS applications that involve components such as exural parts that could be subject to plastic deformation in service.
Plastic deformation of metals or crystalline solids is often accomplished by the generation, accumulation, and interaction of dislocations. However, depending on the materials, microstructures, and testing conditions including temperatures , the plasticity of solids may involve a range of mechanisms. These were well summarized in the literature for the different types of solids such as metals, ceramics, and polymers .
The classical theory of plasticity has been developed to fully characterize bulk crystal plasticity [32,58]. It describes the mechanics of plastic deformation in most engineering bulk solids and is based on an experimentally determined constitutive relation between stress and strain. Mathematical expressions have been developed for polycrystalline aggregates under simple loading conditions . These are generally phenomenological in nature. However, the empirical ow rules of the crystal plasticity theory have been veried for their general applicability to a wide range of engineering materials.
Please note that these rules apply largely to isotropic materials. It is also assumed that the stress at one point is only a function of the strain at that point. In the following section, a review of the recently developed strain gradient plasticity SGP theory relevant to size-dependent plasticity at small scales, as well as the reported length scale effects in plasticity will be given. The connections of small-scale plasticity to dislocation theories will also be discussed. Since the pioneering work of Ashby , the existence of the plasticity length scale, i.
The geometrically necessary dislocations are needed to maintain compatibility in the presence of strain gradients. They contribute a dislocation density, rG, in addition to the statistically stored dislocations of density, r s. During plastic deformation, the density of statistically stored dislocations increases due to a wide range of processes that lead to the production of new dislocations.
However, the density of geometrically necessary dislocations is directly proportional to the plastic strain gradient, i. Depending on the relative magnitudes of these two strain 3 increments, plastic strain accumulation may be dominated either by statistically stored dislocations, or by geometrically necessary dislocations. In bulk metals and their alloys, the contributions from statistically stored dislocations to strain accumulations are much greater than those due to geometrically necessary dislocations.
However, at the microscale, the increments in the two dislocations densities may be comparable, i. This length scale, l, has also been determined experimentally. The early experiments for determining length scale parameter were performed by Fleck et al. Fleck and Hutchinson  have also developed phenomenological theories of strain gradient plasticity, which are amenable to nite element implementation.
Subsequent efforts to explain the length scale effects associated with indentation experiments [6,51,61], fracture , and microbend experiments  resulted in the introduction of two length scale parameters: one associated with the rotational gradients, lR, and the other associated with stretch gradients, ls [6,25]. The basic equations of the FleckHutchinson deformation theory as extensions of conventional J2 theory , denitions of length scale parameters, and the mechanismbased strain gradient plasticity MSG approach [22,59,75], discussed here, are based on detailed consideration of the contributions of geometrically necessary and statistically stored dislocations GNDs and SSDs to the overall plasticity.
Given that these topics are well documented in the literature, further discussion will be omitted here due to the scope of the current work. Instead, we will briey summarize the experimental results for LIGA Ni thin lm materials in the following section. As suggested by Nix and Gao , the square of the hardness can be plotted against the reciprocal of the depth of indentation. The microstructural length scale parameter, l, can be computed based on the MSG theory . Using the measured value of H0, and taking the Tabor factor as 3, b as 0.
The extracted microstructural length scale parameter, l, is actually related to the material length scale parameter, l, dened by Fleck and Hutchinson . The results are plotted as a function of surface strain, 3b, for modulus for LIGA Ni EZ four different beam thicknesses 25, 50, , and mm in Figure 3. The regression ts for different thicknesses are also shown. The normalized moments needed to induce similar strains are signicantly larger for thinner beams w25 mm than the thicker beams w mm. In the absence of a size-dependent plastic response, the normalized bending moment curves corresponding to different thicknesses should be coincident.
Accordingly, the differences in material responses are attributed to length scale dependence . The bending moments obtained from the microbend test are utilized to determine the composite length scale parameter, lc, following the strain gradient plasticity analysis developed by Stolken and Evans . From the framework of Fleck and Hutchinson , the strain energy density in the presence of strain gradients can be dened as  3 w Z EP 3 C 2S0 2 3. The nonlinear data tting was performed for the bending moment, M, to all of the bending data, with lc as the only unknown using the built-in functions of the symbolic computation software Mathematica Wolfram, Illinois.
The length scale parameter was thus determined to be lc Z 5. Now, we may extract the rotational gradient length scale parameter as dened before from the obtained composite length scale parameter, lc, using Equation 3.
- Powder Surface Area and Porosity!
- Exploring mechanics of spider silk to design materials with high strength and low density.
- Smart materials for green building;
- National Institute of Standards and Technology (NIST) | Materials Genome Initiative.
- David Marshall?
From the nanoindentation experiments, we just calculated the material length scale parameter to be lw0. This length scale parameter corresponds essentially to the stretch gradient parameter since the stretch gradient plays a dominant role in the nanoindentation tests. The computed parameter is on the same order of what Begley and Hutchinson  obtained from earlier studies on W, Cu, and Ag single crystal, as well as Cu polycrystal.
These gave an estimate of l on the order of 0. From the composite length scale parameter, lc of 5. The measured values of lR and ls can now be incorporated into the extended J2 plasticity theory to obtain the new constitutive law. Incorporation of above measured plasticity length scale parameters into the strain gradient plasticity framework of , therefore, provided a framework for continuum modeling from the micron to the submicron scales.
Within this regime, the extended J2 framework may be incorporated into analytical or numerical models for the determination of: local stressstrain distributions, cracktip elds, and failure criteria for thin lm materials and structures used for MEMS applications. Several factors such as solution conditions including composition, temperature, pH level, organic additives, etc. Thus, the diverse and ample data for mechanical properties measurements on electrodeposited Ni that exist in the literature are due to this complexity.
The microtensile test results obtained for the mm thick LIGA Ni sample are plotted in terms of true stress and true strain in Figure 3. This shows that the plastic deformation starts at very low strain levels w0. However, because of compliance problems with the machine frame, obtaining accurate measurements of the elastic deformation of the sample is very difcult.
In this study, a Youngs modulus of GPa is obtained by a linear t of limited data in an elastic regime from a 50 mm sample test and is used as a material constant in our estimation for a 0. Nevertheless, the ultimate tensile strength measured for the mm sample was MPa. The maximum load applied at the onset of sample failure was w16 N. The ultimate tensile strength of the mm structure was MPa and the maximum load at the onset of sample failure was w Similarly, as shown in Figure 3. The ultimate tensile strength in this case was MPa and the maximum load at the onset of sample failure was w3.
Mazza et al. The difference in numbers is probably because the material they were testing has a different microstructure and is obviously stronger. Nevertheless, the values of tensile properties are in fairly good agreement with the work of Sharpe et al. In all the thin lm samples, there was clear evidence of necking and nal shear fracture at an angle of w to the loading direction Figure 3.
Rough transgranular fracture modes were observed in all cases, along with evidence of secondary cracks Figure 3. In the case of the 50 mm thick structure, a sharp edge is formed, following the onset of necking and shear fracture Figure 3. However, similar sharp edges were not observed in the thicker samples. Fracture in and mm thick foils occurs by a combination of faceted fracture and shear fracture. In the case of the 50 mm foil, the formation of a sharp edge is presumably a result of the combination of necking and cracking along shear planes that are at an angle of w to the loading axis.
The word fatigue is from the Latin expression fatigare which means, to tire . It was believed that a material tires over time when subjected to repeated loading and nally fails. Moreover, as MEMS devices start to push toward higher mechanical power levels and many of them may be subject to very high numbers of fatigue cycles during their service time due to the inherent high operating frequencies, fatigue also become a major concern in the small scales .
Mohr et al. For 10 mm thick and mm long cantilevers, they found that the long-term fatigue behavior is comparable to that known from macroscopic nickel materials. Later, Hemker and co-workers  conducted fatigue tests for dog-bone shape LIGA Ni samples with thicknesses of w mm using a microfatigue tester driven by a voice-coil actuator.
They employed a sinusoidal tensiontension cyclic loading condition with an R ratio of 0. From a scanning electron microscopy SEM fracture surface study, they also proposed that the fatigue crack started at the corner of the microsamples and then propagated across the sample in a transgranular mode. More recently, Boyce et al. The width, thickness, and length of the microbeams were w26, w, and w mm, respectively.
From their fatigue-life study, the endurance limit was determined to be 0. A SEM inspection of the fatigued surface showed evidence of localized extrusions and intrusions associated with persistent slip bands PSBs. Moreover, a thick up to nm oxideNiO layer was found on the surface of the PSBs and thought to be the source of crack initiation.
This thick oxide layer and the associated oxygen-suppression mechanism decreased the fatigue life of LIGA Ni by more than an order of magnitude compared with tests done in low-oxygen environments. The thickening of the oxide is attributed to the disruption of oxide by motions of PSB extrusions and intrusions. Stress-fatigue life data obtained for the 70 and mm thick LIGA Ni specimens are compared with prior results by Cho et al.
Results of Cho et al. The thinner 70 mm thick samples exhibited longer fatigue lives and increased endurance limits w MPa than the thicker samples endurance limits w MPa. The high cycle fatigue results of Cho et al. However, the data from the thicker samples were also close to those reported for bulk Ni in the annealed state .
The results of fatigue tests performed on the thinner w70 mm thick specimens were comparable to those reported for bulk Ni in the hardened condition examined by Mohr et al. However, the increased fatigue strength is associated largely with the increased tensile strength of the thinner samples. This is consistent with prior work by Cho et al. The fatigue fracture modes and the deformed gauge sections are presented in Figure 3. In the case of the mm specimens tested at low stresses, corner fatigue cracks were observed initiating from the surfaces of the specimens.
These propagated inwards at an angle of w stage I fatigue cracks via a at fracture mode Figure 3. A transition to a rougher fracture mode, consisting of fatigue striations was then observed after the cracks reached lengths of w mm Figure 3. Finally, nal failure occurred by a ductile dimpled fracture mode Figure 3. Similar fatigue fracture modes were observed at different stress levels Figure 3. However, the extent of stable fatigue crack growth was less at higher stresses.
At very high stress levels very low fatigue life as shown in Figure 3. In the case of the thinner 70 mm thick specimens, a signicant amount of plastic deformation and necking occurred prior to nal fracture Figure 3. However, the fatigue crack propagation generally occurred on planes that were mainly perpendicular to the loading axis. The fatigue fracture surfaces exhibited clear evidence of transcolumnar fracture Figure 3.
The features observed at lower stresses were more consistent with mechanical fatigue. At intermediate and higher stresses, the fracture surfaces were dominated by evidence of plastic deformation necking down to chisel-shaped wedges , secondary cracking, and faceted trans-granular crack growth. At the highest stresses, extensive necking was observed with limited evidence of rupture by ductile rupture Figure 3.
The fatigue of the thinner samples is therefore dominated by. In contrast, transcolumnar fatigue fracture modes dominate at lower stresses. However, the underlying fatigue deformation and crack growth mechanisms depend on the magnitudes of the applied stress ranges and the sample thickness. In the case of the 70 mm thick structures, more extensive plastic deformation is observed along with predominantly stage II growth by transcolumnar crack growth modes. No evidence of fatigue striations was observed in these structures. However, in the case of the mm thick structures, less extensive but signicant plastic deformation is observed along with stage I and stage II crack growth.
The stage I cracks initiate from the corners and extend inwards until they transform into stage II cracks. Most of the stage II crack growth in the mm thick structures occurs by classical cracktip blunting mechanisms . Crack growth direction a b. After briey reviewing the status of MEMS technology, the focus has been on the development of experimental techniques for the mechanical testing of thin lms and small structures.
The purpose was to demonstrate how these experimental developments could be utilized to further our understanding of the mechanical behavior of materials at small scales, and to help resolve design and reliability issues in MEMS technology. Soboyejo for his encouragement and helpful discussions in the writing of this chapter. Contributions from Dr. Allameh and Dr.
Shrotriya are also gratefully acknowledged. Allameh, S. Arney, S. Arzt, E. Ashby, M. Beams, J. Begley, M. Bishop, D. Bolshakov, A. Boyce, B. Brotzen, F. Brown, S. Buchheit, T. Burbaum, C. Actuators, A, , Cheng, Y. Cho, H. Actuators, A: Phys. Christensen, T. De Guzman, M. Ehrfeld, W. Feynman, R. Fleck, N. Solids, 49, , Florando, J. Gerberich, W. Hay, J. Hertzberg, R. Hill, R. Hommel, M.
Hong, S. Hormes, J. Methods Phys. B, , , Hornbeck, L. Hruby, J.
Optimize Materials Knowledge and Applications with ANSYS Granta
Thesis, Harvard University, Hutchinson, J. Jaeger, R. Kahn, H. Kiesewetter, L. Actuators A, 35, , Klein, M. Koskinen, J. Kraft, O. Madou, M.
- The Political Economy of Liberation: Thomas Sowell and James Cone on the Black Experience.
- High Temperature Materials Design Group - Research Center for Structural Materials (RCSM) | NIMS!
- Chemical complexity promises improved structural alloys for next-gen nuclear energy.
- A Probability Metrics Approach to Financial Risk Measures.
- COMAT 2014 - Recent Trends in Structural Materials.
- Master the Boards USMLE Step 2 CK!
Malzbender, J. Mazza, E. McElhaney, K. Miller, W. Mohr, J. Nadai, A. Poole, W. Read, D. Fatigue, 20, , Rudd, J. Safranek, W. Sharpe, W. Sharpe , W. Shrotriya, P. Spearing, S. Stelmashenko, N. Stolken, J. Tabor, D. Vinci, R. Vlassak, J. Wei, Y. Solids, 45, , Weihs, T. Xiang, Y.
Xie, Z. Miniaturizing technology with benets such as new functionality, cost reduction, and space saving, has helped MEMS applications span over numerous elds including automotive, aerospace, photonics, telecommunications, life sciences, biochemistry, biology, biomedicine, and drug delivery to name a few.
When it comes to sensors and actuators, MEMS is a strong competitor for the conventional manufacturing processes. Whenever a new functionality becomes possible by going small e. This chapter focuses on Si-based MEMS with the main emphases placed on silicon properties, device fabrication, device applications, and the related mechanical and reliability related issues.
This arises mainly from the economic benets due to the well-established semiconductor manufacturing technology that provides the industrial infrastructure needed for MEMS fabrication. This is in addition to the desirable properties of silicon including electrical, optical, and mechanical, linked to various crystal structures.
The well-established micromachining techniques with additive and subtractive processes make the design and mass production of Si-MEMS easy and economical. Si-based MEMS may have other materials that are compatible with silicon. These include silicon oxides, silicon nitrides, silicon carbides, and metals such as Al, W, Cu, and polymers such as polyimide. The type of application and manufacturing process are the main factors that determine the type of silicon used. Single crystal silicon wafers used for MEMS are commonly and mm in diameter with thicknesses in the range of mm for single-sided wafers and mm thinner for the double-sided polished wafers.
Crystal orientation of the wafers are commonly and with dopant type being n or p. Figure 4. This property can be utilized to etch holes with walls making an angle of While silicon is compatible with most service conditions for which MEMS devices are considered, its biocompatibility is being currently investigated. Bare silicon surfaces have shown to be less than favorable for amplication of genetic DNA materials.
- METAL MATRIX COMPOSITES REDUCE WEIGHT FOR SATELLITES.
- The Earths Age and Geochronology.
However, for most biological applications such as implants, Si has been shown to be biocompatible with no toxic products being released. Youngs modulus of silicon varies with orientation, from w GPa in  direction to GPa in  direction. Its Poissons ratio is 0. The strength of crystalline silicon can be understood visualizing the weight of a man w70 kg being suspended from a silicon ber as thin as his hair w mm diameter.
Table 4. For pure single crystal silicon wafers, the mechanical properties are uniform across the whole lot leading to devices. TABLE 4. While single crystal wafers are nearly stress free due to their purity, this is not the case with polysilicon. After deposition, polysilicon must be annealed at high temperatures e. Careful tailoring of deposition temperatures and lm thicknesses in polysilicon multilayer lms, grown alternately at lower and higher temperatures, can lead to nearly stress-free lms. This is particularly important for polysilicon cantilevers where intrinsic stresses can cause curling of the beams if they are not properly annealed.
Mechanical properties of single crystal silicon at elevated temperatures up to C are not signicantly different from those of silicon at room temperature. The onset of softening and plastic deformation takes place at C. Polysilicon, on the other hand, starts to show time-dependent stress annealing at temperatures as low as C leading to instabilities that may interfere with its mechanical functions.
Some of the more important electrical properties of silicon and its alloys are presented in Table 4. Important electrical properties include piezoresistivity, piezoelectricity, and thermoelectricity. Silicon in single crystalline and polycrystalline and even amorphous forms exhibit piezoelectricity characteristics. Piezoresistivity also varies with Youngs modulus, temperature, type, and concentration of dopant. Dened as the change in resistivity due to applied stress, piezoresistivity can be related to the stresses applied parallel and perpendicular to silicon resistor as follows: Dr Z pII sII C pt s r 4.
According to Equation 4. The proportionality constants are functions of temperature and change by K0. For polycrystalline silicon, piezoresistivity is averaged over all grains and all directions yielding an overall effect that is nearly one-third of that of single crystal silicon. A gauge factor, called K, relates the applied stress to the relative change in resistivity and ranges from K30 to C The temperature coefcient of resistivity TCR of polysilicon is 0.
Piezoelectricity effect arises from asymmetry in the primitive unit cell of a crystalline material combined with ionic or semi-ionic bonding of its atoms. Silicon is symmetric in crystal structure diamond crystal structure, cubic with covalent bonding and therefore lacks piezoelectric capability. However, piezoelectric materials may be applied to a silicon substrate as thin lms by either deposition e. Currently, there is no easy way to deposit quartz on Si substrate. Thermoelectric properties of polysilicon make it a suitable material for applications in which the Seebeck effect, Peltier effect, and Thomson effect play the main role.
Seebeck effect used in thermocouples allows for the measurement of temperature differences, while Peltier effect is utilized in thermoelectric coolers and refrigerators. To measure temperature difference between hot and cold sources, two elements with different Seebeck coefcients of a1 and a2 are joined and the joint is placed at the hot source e.
The other ends of the two elements are placed at the cold source e. The temperature difference can be related to the measured electric eld generated between the two elements as follows: DV Z a2 Ka1 Thot KTcold 4. In the Peltier effect, two dissimilar materials are joined together. When an electrical current passes through the couple, a heat ux is generated. This ux cools down one side of the couple relative to the other. While Peltier devices are made of n-type and p-type bismuth telluride elements, it is difcult to implement Peltier effect in thin lm applications.
These numbers can be compared with those of Bi K As seen from Table 4.
The thickening of this layer up to a few tens of nanometers provides the structure with a solid adherent protective layer that effectively prevents further oxidation of silicon. The inert nature of silicon dioxide makes silicon structures very immune to the stress corrosion cracking that affects many other structural materials. Nevertheless, as will be seen later, this will not hold true if the loadbearing structural components made of silicon are exposed to moisture Table 4.
Silicon oxides are widely used in surface micromachining for insulators and sacricial layers to name two applications. Silicon dioxide SiO2 is thermally grown by exposing silicon to oxidizing atmospheres at temperatures above C.
International Graduate Program : Multiscale Materials
Other oxides of silicon are grown by chemical vapor deposition CVD , sputtering, and spin coating. Silicon oxides become soft at temperatures above C and therefore are of limited use as beams and cantilevers at higher temperatures. Silicon nitrides coatings act as barriers to the diffusion of most mobile ions including the sodium and potassium present in biological environments. Youngs modulus of Si3N4 is larger than that of Si making it an excellent choice for coatings on cutting tools. It also makes silicon nitride coating a suitable mask for alkaline etch processes. While Al r Z 2.
Metals such as Ni r Z6. From Allameh, S. Cu rZ 1. Permalloy and Ni are the materials of choice for magnetic transducers. NiCr rZ mU cm is used for laser trimmed resistors. Ti and Ti6Al4V make excellent biocompatible coatings. TiNi is used for shape memory alloy actuation. What distinguishes micromachining from conventional bulk machining is mainly the size, implying that at least one-dimension of the product is in micron range.
Additionally, micromachining usually comprises parallel fabrication of multiple devices, occasionally thousands per wafer with dozens of wafers being processed simultaneously. Micromachining involved in the manufacture of MEMS can be divided into two main groups: bulk and surface micromachining. Structures are rarely made by self-assembly, self-synthesis, or self-organization of species. Usually some kind of patterning will be needed using lithography and pattern transfer followed by micromachining. The basic processes include deposition of layers e.
In addition to deposition methods, additive processes include anodic bonding and fusion bonding. Pattering techniques include optical lithography and double-sided lithography. Finally, etching techniques include wet isotropic, wet anisotropic, plasma, reactive ion etching RIE , and deep reactive ion etching DRIE. The growth techniques include epitaxy, oxidation, sputtering, evaporation, chemical vapor deposition, and spin-on method. An orientation relationship OR between the two crystals at the two sides of the interface determines the crystallographic planes and directions of the overgrowth.