TWI287292B - Method and apparatus transporting charges in semiconductor device and semiconductor memory device - Google Patents

Abstract

A conductor-filter system, a conductor-insulator system, and a charge-injection system are provided. The conductor-filter system provides band-pass filtering function, charge-filtering function, voltage-divider function, and mass-filtering function to charge-carriers flows. The conductor-insulator system provides Image-Force barrier lowering effect to collect charge-carriers. The charge-injection system includes the conductor-filter system and the conductor-insulator system, wherein the filter of the conductor-filter system contacts the conductor of the conductor-insulator system. Method and apparatus on charges filtering and injection are provided for semiconductor device and nonvolatile memory device. Additionally, method and apparatus on charges injection using piezo-ballistic-charges injection mechanism are provided to the charge-injection system and devices operation. Memory cells and array architectures and manufacturing method thereof are provided.

Description

1287292 IX. Description of the invention: [Technical field to which the invention pertains] The present invention relates to a semi-conducting and dithering, and is a method and apparatus for transmitting an electric charge in these devices with respect to an Image Force mechanism. . [Prior Art] Imaging power is a well-known research topic, such as Sze's "Physics of Senriconductor Devices" (Wiley, New Y〇rk, 1981, Ch. This leads to a reduction in the tank, _ splicing and image energy barrier reduction effect (Image_Forcebarriei·loweringeffct), and is the main dominant mechanism for the Schottky Effect used for charge carrier emission. A name published by Lenzlinger and Snow The image force is also discussed in the article "Fulnohan tunneling into the heat of the second J CToweler-Nordheim Timneling into Thermally Grown Si02/5 J. Appl. Phys" 40, pp. 278-283 (1969). In this paper, the image force effect is combined with the Fowler-Nordheim Tunneling (FN) mechanism used by the tunneling of the Thermal Cairiers through the cerium oxide (oxide). Attempts have been made to use image forces together with photocurrent-voltage measurements to trace the distribution of oxides (see Nicollian and Brews for a book on "MOS Physics and Technology" ("MOS Physics and Technology," , Wiley, New York, 1982, Chapter 11, ρ·513)). Imaging forces and these methods have also been applied to study the energy barrier height at the junction of metals and oxides and to study the interface between bismuth (Si) and oxides. In U.S. Patent No. 6,744,111 issued by Wu on June 1, 2004, one has an emitter (Emitter), a base (Base), and a collector (Collector). The three-terminal semiconductor elements are described. Schottky barriers are formed at the junction of the emitter region and the base region and at the junction of the collector region and the base region. This type of component utilizes the Schottky effect (by the image force barrier reduction mechanism) and allows the 0881-A31715TWF(5.0) d 1287292 tunneling current to pass through the Schottky barrier by controlling the voltage in the base region. However, all of the above uses the image force machine Examples and attempts are non-volatile memory-independent applications. Non-volatile semiconductor memory cells with charge storage capability are well known in the art. Charges are typically stored in a floating gate. To define the state of a memory cell. Typically, the state of a 'memory cell can have two levels, or can have more than two levels (used for storage in a multi-level state), such as channel hot electrons (Channel Hot) Electron; CHE), Source_Side Injection (SSI), Fowler-Nordheim Tunneling (FN) and energy introduction into the band (BalKj_t〇_BanciTunneling; BTBT) The mechanism of electron injection can be used to change the state of such memory cells during programming and/or erasing operations. Examples of these mechanisms for memory cell operation can be found in Case Nos. 4,698,787, 5,029,130, 5792,670 and 5,966,329. US patents are discussed separately for the CHEI mechanism, the SSI mechanism, the FN mechanism, and the BTBT mechanism. However, 'all of the above mechanisms and tried injection efficiency (defined as the collected carriers) In addition, these mechanisms require a high voltage to support the memory of L and a voltage of up to 1 〇V is often visible. At such high voltages, it is necessary to strictly control the surrounding net. The quality of the insulator of the brake. Memory operating under these mechanisms is thus susceptible to manufacturing and reliability issues. In view of the above problems, the object of the present invention is to provide - insulation energy barrier - conductor - absolute, body system, scale - insulation age, surface loader can be used to inject efficiency and reduce operation power = ^ invention another purpose It is to provide a carrier (charge_) capable of tight energy distribution and high injection efficiency. The remaining objects of the present invention and the implication of these objects are described in the specification and drawings. J ^ [Summary of the Invention] The purpose of this month is to provide a method for charge filtration and injection in a semiconductor device and a memory. 0881-A31715TWF(5.0) 6 1287292 \ and device. Briefly stated, one embodiment of the invention is a conductor-filter system. The conductor-pass filter system includes a conductor for supplying a normal temperature charge carrier and a filter that is in contact with the conductor. The filter includes a dielectric for providing a filtering function for a polar charge carrier, wherein the filter includes an electrically variably (Potential: Barrier) and is used to control the edge. The flow of a certain polarity charge carrier through the filter in a certain direction. In addition to controlling the charge carriers of a certain polarity, the filter further includes another set of electrically configurable potential energy barriers for controlling the passage in the direction substantially opposite to one of the above directions. The flow of opposite polarity charge carriers. Briefly, another embodiment of the present invention is a conductor-insulator system. The conductor-insulator system includes a conductor having high-energy charge carriers, wherein the high-energy charge carriers have a Schott-Break distribution, and an insulator that interfaces with the conductor. The insulator has an electrically deformable image force potential barrier adjacent the junction, wherein the image potential barrier is electrically configurable to allow the high energy charge carriers to pass over it. In another preferred embodiment, the still-charged charge carriers have an energy spectrum with an energy spectrum ranging from about 3 〇 meV to 300 meV. _ 帛 In other words, another embodiment of the present invention is a charge injection system. The charge injection system includes a conductor-filter system having a conductor to provide a normal temperature charge carrier, and a filter in contact with the conductor, the dielectric comprising a dielectric to provide a charge to a certain polarity. Filter - 侃. The Gu II system includes a group of electrically configurable potential energy barriers for controlling the flow of a certain polarity charge carrier passing through the H-direction in a certain direction, and further including another set of electrically correctable potential energy a barrier to control the flow of opposite polarity charge carriers through the filter in a direction substantially opposite to one of the above directions. The charge injection system further includes a conductor-insulator system. The conductor-insulator system includes a second conductor in contact with the filter and having high energy electrons from the filter, and an insulator connected to the second conductor at an interface, and the insulator There is a _ image force repairing near the junction. The image force potential barrier system 0881-A31715TWF(5.0) 1287292 can be electrically modified to allow these capable charge carriers to pass over it for transmission. Briefly stated, another embodiment of the invention is a memory unit. The memory unit includes a conductor-filter secret having a conductor to provide a constant temperature symmetry, and a filter in contact with the conductor, the dielectric comprising a dielectric to provide a filtering function for a polar charge carrier . The device includes a first group of electrically configurable potentials to control the flow of a polar charge carrier passing through the filter in a certain direction and a second set of electrically correctable potentials. Energy barrier to control • Another opposite polarity through the filter in a direction generally opposite to one of the above directions: the flow of charge carriers. #®单言' Another embodiment of the present invention is a method of forming a memory unit. The method includes the steps of: forming a body having a first conductivity type in a semiconductor substrate, forming a first insulating layer on the substrate, forming an electrical storage layer on the first insulating layer, forming a a second region of the second conductivity type and a second region having a second conductivity type are formed in the body to form a channel region between the main region and the second region, wherein the channel region is substantially Adjacent to and insulated from the charge storage region, a second insulating layer is formed adjacent to the charge insulating layer to form a first conductive region in which at least a portion of the region is adjacent to and electrically insulated from the electrical storage region a filter having a filtering function adjacent to the first conductive region, and a second conductive region adjacent to at least a portion of the first conductive region and insulated by the filament. The simple σ of the invention is another _ real complement system - a method for forming a fresh-keeping array. The method includes the following steps: forming a body having a first conductivity type in the semiconductor substrate, forming a first insulating layer on the substrate, forming bit lines separated from each other and extending in a first direction (4) lmes In the H domain, wherein each of the bit lines (four) is formed in at least a portion of the area of the body: a charge age area machine-insulation layer, and the towel charge storage area is arranged as: an array, wherein The rows of the array extend in the first direction, and the array of the array extends in a second direction perpendicular to the crucible (four) to form a plurality of first regions having a second conductivity type and a plurality of second regions having a second conductivity type a region, forming a plurality of channel regions in the 80881-A31715TWF (5.0) 1287292 sense, within the body, wherein each of the channel regions extends between one of the first regions and the second regions And substantially insulatively adjacent to the adjacent phantoms in the charge regions #, forming a plurality of display H having a filtering function, wherein each filter has a portion of the region and the first conductive regions - Adjacent; And a plurality of second conductive regions, wherein each of the second conductive regions is adjacent to at least a portion of the first conductive regions and is insulated by one of the filters. The present invention is based on other objects and features, and will be apparent from the following detailed description. [Embodiment] The symbol Ν+ used in the present specification represents a heavily doped v-type semiconductor material, and the doping concentration of the cerium-type impurity (for example, arsenic) is typically 102 Å (atoms/cm 3 ). The order of magnitude. The symbol Ρ+ represents a heavily doped ytterbium-type semiconductor material, and the inclusion of the > type impurity (for example, boron) is typically on the order of 10 (atoms/cubic centimeter). Wherever practicable, the description and the following description will be used to refer to the same or similar parts. Fig. 1 shows an energy band diagram of a conductor-insulator system applied under an electric field. The energy band diagram shows a conductor 10 in contact with an insulator 12 and having a Fermi level 16 in its energy band. The figure also shows the conductive strip 18 of the insulator 12 under the effect of image force and the conductive strip 18' under the effect of image force. In addition, the potential energy barrier 24 under the image force effect of the insulator 12 and the potential energy barrier 24' under the imageless force effect have energy barrier heights A 20 and ‘22, respectively. This figure shows that the stress effect changes the shape of the potential barrier from a triangular barrier 24 with a sharp corner at the edge of the barrier to a triangular barrier with a rounded corner 24 (called "image power only" Or "image power barrier"). This effect reduces the energy barrier height of the potential barrier from the barrier height 22 to the energy P early south 20 ’. The difference between the two can be △ 26. An energy barrier peak 28 is also shown at the highest point of the image energy barrier 24, which is a distance from the junction of the conductor 10 and the insulator 12 by 0881-A31715TWF (5.0) 9 1287292 • x- 30. In Fig. 1, the conductor may be a semiconductor, such as an N+ polymorph type of germanium (i.e., polysiclicon), a P+ polycrystalline dream, and a heavily doped polymorphic type of stone lithography (i.e., polycrystalline stone 锗(P〇ly SiGe)), or may be a metal, such as aluminum (Aluminum; A1), platinum (Platinum; Pt), Au, tungsten (Tungsten; W), la (Molybdenum; Mo), ruthenium (Ruthenium; Ru ), Tantalum; Ta), Nickel (Ni), Nitanium Nitride, Titanium Nitride, etc., or alloys of the above materials, such as Ming Shishi (Platinum_Silicide), Tungsten-Silicide, Nickel-Silicide, etc. The insulator can be a dielectric or air. When a dielectric is considered as a material for the insulator, such as an oxide, a nitride, or an oxynitride (SiON) can be used as the dielectric. In addition, dielectrics whose dielectric constant is lower or higher than the dielectric constant (or permittivity) of oxides (referred to as "low-k dielectric" and "high-k dielectric", respectively) may also be considered as The material of the insulator. Such low-k dielectrics may be Fluorinated Silicon Glass (FSG), SiLK, Porous Oxide, such as Nanoporous Cabon-Doped Oxed, and the like. Such high-k dielectrics can be aluminum oxide (Aluminum Oxide; Al2〇3), hafnium oxide (Hafium Oxide; Hf02), titanium oxide (Titanium Oxide; Ti02), oxidized hammer (Zirconium Oxide; Zr02), oxygen • Group (TantalumPen_〇xide; Ta205) and so on. Furthermore, the mixture of these materials or the alloy formed thereof is, for example, an alloy of Hafiiium Oxide-Oxide alloy (Hf〇rSi〇2), an alloy of bismuth-alumina (Hafiiium-Aluminum- Oxide Alloy; HfAlO), "Hafiiium-Oxynitride Alloy; HfSiON... and so on, can be selected as the material of this type of dielectric. In addition, it is not necessary to require the insulator to be a dielectric material having a uniform chemical element, and it is not required to include only a single layer, but the elements therein may be allowed to gradually change, and may be allowed to include more than one layer. Figure 2 shows that electrons 31 are transmitted through a quantum mechanical tunneling mechanism (e.g., Furnohan) to pass the potential barrier of Figure 1. The electrons 31 in the conductor 10 are at normal temperature before being tunneled through the potential barriers 24 and 24, and a_ has no kinetic energy to the Fermi puzzle 16. This type of electron is called "normal temperature 0881-A31715TWF (5.0) 1287292, electronic (Thermal Electrons)", and such charge carriers are called "normal temperature charge carriers" or "normal temperature carriers". When a large electric field (typically greater than 10 MV/cm) is applied to the insulator 12, the ambient temperature electrons 31 can pass through the insulator 12 using a quantum mechanical transmission mechanism. Under such a large electric field, the electrons 31 are shown to tunnel through the insulator 12 and enter the conductive strips 18 and 18', respectively, in the case of image force effects and non-image force effects. When the image force enablement barrier is lowered, it is known that such a tunneling mechanism causes the tunneling rate of electrons 31 to pass through the potential barrier 24 to be higher than the tunneling probability of the transmission through the potential barrier 24'. Fig. 3A shows an energy band diagram of a potential charge carrier (hot electron 32) transmitted over the potential energy barrier of the conductor-insulator system of Fig. 1. A high-energy charge sub-system within a region is defined as a charge carrier with kinetic energy in the Fermi level of the region. For example, in Figure 3A, the hot electrons 32 located within the conductor 10 are shown to have a kinetic energy % relative to the Fermi level of the conductor 1 。. Such an electron transfer mechanism is different from the transfer mechanism of the normal temperature electron 31 in FIG. The kinetic energy % is in the figure and is slightly higher than the energy barrier height of the image force potential barrier 24 by 2〇 and lower than the energy barrier height 22. The figure shows that the hot electrons 32 are moved from the conductor 1 to the insulator 12 in a forward direction 34 (shown by the arrow). When considering a potential barrier that does not have a shirting effect, the kinetic energy % is insufficient to support the thermocouple 32 to cross the potential barrier 24', so that the thermal electrons 32 are blocked by the potential barrier 24 and move along a return path 34. However, under the influence of the image force effect, the reduced energy barrier height 2〇 allows the hot electrons 32 having the same kinetic energy 33 to travel along the forward direction 34 and pass over the image force potential energy barrier to enter its conductive strip 18 . This effect is desirable because when the hot electrons are fabricated in the integrated circuit (10) and related applications of the memory, it can reduce the voltage required to supply the energy of the hot electrons 32. Fig. 3B shows the effect of the image force on the position of the peak of the barrier that changes the height of the barrier and the potential barrier. The energy barrier height and the energy barrier peak position are depicted as a function of the electric field 之外 applied to the insulator 12. Fig. 3B shows that when an electric field ED of about 5 MV/cm is applied to the insulator 12, the barrier height 20 is lowered from 3.1 V to about 3_5 V. Typically, this _ electric field is applied by applying a = across insulator J. For example, an oxide insulator with a thickness of 6 nm is taken as an example. 0881-A31715TWF(5.〇) 11 1287292 • If an electric field of 5 lm is required, it is necessary to cross the 3 3 之 于The oxide. Under the application of this electric field, the force effect wire saves electrons, and the image force effect and the potential energy barrier must be overcome to the distance Xm instead of being overcome to an infinitely long distance. Figure 3B further shows that the distance Xm 3 from the peak of the barrier to the conductor/insulator boundary can be reduced from the infinite length range (ED = 〇 MV / Cm) to less than i nanometer (in the case of ^ 娜 咖 咖 ). In the field of solid state physics, when the transit time of the mobile charge (T_it Time) is shorter than the dielectric polarization time of the medium (for example, the insulator 12 of Fig. 1), _ectric p〇larizati〇n It is known that the polarity of the dielectric does not change closely with the mobile charge. As shown in Figure 3, shortening the length of the barrier peak Xm will reduce the charge transfer time, and this effect is desirable because it provides the dielectric constant of the image force potential barrier 24 (like force dielectric) Constant) = low means to increase the effect of low energy. Other means, such as increasing the rate of charge movement (e.g., increasing charge kinetic energy), may also be considered to reduce the transmission time to achieve the same effect in the present invention. Typically, with such means, the dielectric constant can be reduced from its static value (e.g., the oxide is about 3.9) to a value close to the optical permittivity (e.g., 'oxidation'). The object is 2·2), and thus the image force potential barrier 24 is further reduced by about 14 eV (in terms of oxide). Note that this effect is the result of a short transit time of the carrier (electron) at a distance Xm 30' and this effect is when the carrier transport time is higher than the dielectric polarization time of the insulator 12 and the carrier and other particles Occurs when there is no interaction between them. It is noted that in some cases, the carrier may interact with quantum mechanical particles (e.g., phonons) within distance 30. Such a parent interaction would cause the image dielectric constant of the potential barrier 24 to be slightly greater than its optical permittivity, k and the effect of the reduction of the nucleus would be slightly diminished when the ribs were provided. Fig. 3C shows the relationship between the energy barrier height of the potential barrier calculated from the different dielectric constants K and the electric field applied to the potential barrier using the image force theory. The relationship between the energy barrier height corresponding to the small KH.4 and the electric field Ed is the strongest. For the case of an electric field Ed of about 5 MV/cm, the figure shows that when K=31, the energy barrier height may be reduced to about 2.6 eV, and may be reduced by about 0.2 eV to 2 when Κ=1·4. · 4 eV. The crucible shows 12 〇 881-A317l5TWF (5. 〇) 1287292 - out, by means of an insulator with a lower dielectric constant, and / or by means of the relevant description of Figure 3B to make the image of the power potential barrier The means of reducing the constant to magnify the image force reduces the effect of the energy barrier. Figure 4 is a diagram showing the energy band diagram of the conductor-insulator system provided by the present invention. The embodiment illustrates a group of hot electrons 32 transmitted through the potential-energy barrier 24 of the conductor-insulator system 32 of Figure 1. The conductor-insulator system includes a conductor 10 including a high-energy charge carrier 32 having an energy distribution 36, and an insulator 12 interfacing with the conductor 10 at an interface 14, and the insulator 12 is An image force potential barrier 24 is provided in the vicinity of the junction 14, wherein the image force potential barrier 24 is electrically configurable to allow the high energy charge carrier 32 to be transported over it. The figure shows that the electron 32 has an energy distribution 36 which is a distribution pattern of electrons 32 at different energy levels and this energy distribution % is shown as a Gaussian-Shape having a broad energy spectrum A36. This energy distribution 36 has a peak distribution 36p at the energy level of the kinetic energy 33, which is the same as the kinetic energy level in the description associated with Fig. 3A. Figure 4 further shows that about half (upper half) of electrons 32 have energy above the barrier height of 20, while the other half (lower half) of electrons 32 have energy below the barrier height of 2 〇. In the absence of the effect of reducing the image energy barrier, all of the electrons 32 are shown to be blocked by the potential barrier 24 formed by the conductive strip 18. Under the effect of the reduction of the image and image energy barrier, the electrons 32 in the upper portion of the spectrum are capable of overcoming the image potential potential barrier 24 formed by the conductive strip 8 and in the forward direction 34 (with an arrow) Display) to transfer. • These electrons 32 enter the conductive strip 18 and become electrons 32 having an energy distribution 36. The lower part of the electron 32 is blocked by the image force potential barrier 24 because it does not have sufficient kinetic energy. Therefore, it is important that, as shown, the energy distribution 36 of the electrons 32 reflects only the energy distribution of the upper portion of the electrons 32. In Figure 4, there is another image force effect that is noted and provided herein. It is noted that the lower portion of the electron 32 has a lower kinetic energy than the upper portion. Therefore, when the electron 32 crosses the distance Xm30 and has not reached the peak energy barrier, the transmission time of the lower portion is higher than the transmission time of the upper portion. In some cases, the transmission time is longer than the dielectric relaxation time of the insulator 12 (Dielectrie ° 881-A31715TWF (5.0) 13 1287292 T Relaxati〇n Time), and thus allows the insulator 12 to completely screen and these electrons. Inter-image force interaction. Since the dielectric constant seen by such electrons is large, this results in a weak effect of reducing the image energy barrier. Such effects result in lower energy electrons having to tunnel higher energy barriers 20, thereby blocking the electrons from being able to overcome the potential barrier 24. The image force effect described in Figure 4 provides a filtering effect that allows high energy charge carriers to pass and block low energy charge carriers. The energy level of the carrier (Threshold Energy) can be selected by controlling the energy barrier height 20, and the energy barrier height 2 (^ can be based on the description of the internal energy barrier height and the electric field according to Figure 3B. The relationship between the two is selected by selecting the electric field of the adipose body 12. Taking the 3B picture as an example, when the electric field varies between 〇 and 5, the threshold energy can be adjusted from 3J eV to about 2 Between 5 eV (or equivalently, when the oxide thickness is assumed to be 6 nm, the voltage across the oxide dielectric is applied by applying a voltage of 〇 to 3 V). The electrons of the spectrum are made by mechanisms well known in the art, such as CHEI, SSI, BTBl, etc. Such mechanisms for supplying electron energy typically involve spherical and non-directional collisions with lattice atoms. Therefore, the energy spectrum Δ36 may be in the range of Q.5... to 3^^. Fig. 5 is a view showing another energy band diagram of the conductor-insulator system provided by the present invention, which shows that the high-energy electron 37 is transmitted through the first Figure conductor "insulator system phantom potential month 24. This conductor 'insulator system includes a conductor The high-energy charge carrier 37 of the tarpaulin 38, and an insulator 12, which is adjacent to the conductor knives _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ Tracing 37 can pass through Figure 5, when the still charge carriers (hot electrons 37) are transmitted over the conductor _ _ _ _ image force potential barrier 24, the figure shows the image of the edge system Δ38. This energy band diagram is distributed except that the outer blade is distributed in the narrow-leaf. The balance of the L is the same as that of the fourth diagram. This 0881-A31715TWF(5.0) 1287292 is different in that the hot electrons are no longer The broad energy spectrum δ36 with the energy distribution 36 is changed to the narrower energy spectrum Δ38 with the radiance 38. The peak distribution of the hot electrons 37 and the energy level 33 of the peak distribution of the middle == are interpolated by the fourth graph. The same is true, so that all of the hot electrons 37 are shown to overcome the image force potential barrier 24 formed by the conductive strip 18 and become electrons 37, wherein the electrical energy distribution 38 is similar to the energy distribution 38. Typically The energy spectrum Δ38 is in the range of about 3〇300 meV. This embodiment is unique in that the electron 37 is squeezed into a tight energy distribution. And image force potential energy barrier 24 as a line - "all-pass filter, coffee filter _1_ 1 ^)?", That is, it allows all of the hot electron white b months at lower this to move through it. This unique feature thus provides benefits to the embodiment of the gastric injection efficiency and lower operating voltage. Although the above descriptions with respect to Figures 2 through 5 are described for the case where the high-energy charge carriers are electrons and the energy barrier is a conductive strip, the remaining types of high-energy charge carriers, such as holes, and other species of energy are described. The same description can be easily made with a belt such as a price belt. Fig. 6 is a view showing an embodiment of the energy band diagram of another embodiment of the present invention. In Fig. 6, the conductor-insulator system comprises a conductor 1 包含 comprising a high-energy charge carrier 40 having an energy distribution 48, and an insulator 12 connected to the conductor 1 at an interface 14, The insulator 12 has an image force potential barrier 42 in the vicinity of the interface, wherein the image force potential barrier 42 is electrically configurable to allow high energy charge carriers to pass over it. There are only a few differences between the energy band diagram in Figure 6 and Figure 5, and the rest are the same. One of the differences is that in Fig. 6, the hot electrons 37 are not used as the transport charge carriers, and the high-energy holes 40 (i.e., "thermoelectric holes" 40) are used as the transport charge carriers. In addition, the potential barrier formed by the insulator is now related to the valence band of the insulator. The figure also shows the energy barrier height 41' of a potential barrier 42' in the conductor-insulator system of Figure 1, which is related to the valence band 44 in the absence of an image force effect, and displays an image force potential energy. The barrier height 42 of the barrier 42 is 41. Similar to the energy barrier height 20 in the descriptions of Figures 1, 3B and 3C, the energy barrier height 41 is reduced by the effect of the image force reduction effect when an electric field is applied to the insulator 12. 0881-A31715TWF(5.0) 15 1287292 In Fig. 6, the thermoelectric holes 40 are shown to have an energy distribution 48, and the total system is distributed in a Gaussian shape distribution map having a narrow energy spectrum Δ48. The energy distribution 48 is shown to have a peak distribution 48p and a tail distribution 48t. The peak distribution 48p is shown to have kinetic energy 46 relative to the Fermi level 16 of the conductor 1〇. The kinetic energy 46 is shown to be slightly above the image dynamometer height 41 and below the barrier height 41'. In the absence of an image force barrier reduction effect, the energy of the hole 40 having the energy distribution 48 is shown to be lower than the energy barrier height 41 and thus the potential energy barrier 42 cannot be overcome. However, when there is an image force effect, the figure shows that most of the holes 40 (except for the portion of the tail end distribution of 48t) can overcome the image force potential barrier 42 and transmit in a forward direction 34 to become an energy distribution. 48, the hole 40'. The energy of these holes 40 is higher than the price of the electrical strips 44, so that they can be transported in the same direction within the range of the insulator 12 to reach adjacent materials on the other side of the insulator (not shown). Figure 6 for the hole mapping also shows a high pass filter effect e; g^ect), which is similar to the effect described in the description of Fig. 4 for electronic depiction. As shown, the kinetic energy of the electricity/same 40 in the range of 4 尾 at the end is slightly lower than the energy barrier 41. These holes are blocked from overcoming the image potential energy F 42 and are therefore not included in the energy distribution 48. However, since the hole 4〇 has a close spectrum Δ48, if an additional small voltage (for example, about 1 〇〇 mV) is applied to increase the energy of the hole, the hole 4 in the end distribution is blocked. The situation can be avoided. It can be understood that the present invention adopts the image force barrier reduction effect, and the hot carrier (electron or hole) can transmit through the dying energy barrier with low kinetic energy, and when operating in the memory unit ^ The effect can be reduced when the effect is light. In order to achieve high injection efficiency, it is hoped that in the month b, the knife cloth has a strong class of sharpness and the memory unit can be used together with the image energy barrier reduction effect. It is to be understood that the hairdressings described herein are the actual and social rewards, and cover any examples of any of the changes in the application of the application, although the energy readings of the present invention are 36, 38 and 48. _ is a high-clear shape, but it is known that the general technique of the lie is white, and this 篁 distribution can be extended to any shape and does not need to be symmetrical in energy. 〇 881-A31715TWF (5.0) (g 1287292 - Figure 7 is an embodiment of an energy band diagram of another conductor-filter system of the present invention. In the conductor-filter system shown in Figure 7, a filter is included The device 52 is in contact with a conductor 5 。. The conductor 50 is used to supply electrons 56 for use as room temperature charge carriers. The filter 52 is in contact with the conductor 5 and includes dielectrics 53 and 54 to provide A filtering function for a charge carrier 56 of a certain polarity (negative charge cuter 'electron 56), wherein the filter includes electrically correctable barrier heights 24^ and 2454 to control in a certain direction (forward direction) 34) Flow through a polar charge carrier 56 of the filter 52. ** Figure 7 shows an example of a filter function. The conductor 50 has a Fermi level 1650 and can be * a semiconductor, such as N+ poly Shi Xi, P+ polycrystalline lithox, heavily doped polycrystalline litho structure, poly SiGe, or may be a metal, such as aluminum (4), turn (four), gold (Au), tungsten (W), molybdenum (Mo), yttrium (RU), |g (Ta), nickel (Ni), nickel nitride (TaN), titanium nitride (TiN), etc., or It is an alloy composed of the above materials, such as platinum monomethane, tungsten-telluride, and nickel monolith. The filter 52 is shown to include a tunneling dielectric (hereinafter referred to as td) 53 and a blocking dielectric (hereinafter referred to as BD54). The conductive tape of the tunnel dielectric TD53 has a magical shape of 24 illusions. The conductive strip 18 blocking the dielectric BD 54 is formed from an energy barrier 24 therein, and there is an offset 55 between the conductive strip I854 and the conductive strip I853 of the TD 53. The TD 53 is placed in the vicinity of the conductor 50 and the BD 54 is placed in the vicinity of the TD 53. Typically, the Energy Band Gap of the BD 54 is narrower than the band gap of the TD 53. When the applied voltage is across the filter 52, the band bending of the conductive strips of the filter 52 can be different from each other. The figure shows that the conductive strip 1854 of the BD 54 has a relatively slight band bending than the conductive strip I853 of the TD 53. The body 50 is supplied with a normal temperature electron 56 having an energy distribution 57. The energy distribution 57 of the electron 56 is shown to be lower than the Fermi level 165 〇 and has a peak distribution 57p and a tail distribution 57t on the energy profile. Conductor 50 provides charge carriers with energy less than the Fermi level and is thus used as a "Low-Pass" carrier supply. The figure shows that when the filter 52 is applied with an electric field, the electrons 56 in the peak distribution 57p can pass through the melon 53 by a quantum mechanical transmission mechanism (for example, direct tunneling), and enter the conductive band π% of the BD54 to become An energy distribution 57 having an electron spectrum 56' of a compact energy spectrum Δ 0881-A31715TWF(5.0) 17 1287292 • 57'. In contrast, electrons 56 within 57t of the trailing end distribution are shown to be unable to tunnel through energy barriers 2453 and 2454. The energy barrier 21 of the BD 54 in the filter 52 forms an additional tunneling barrier to block the electrons 56 located within the trailing end distribution 57t, and maintains the energy barrier 2454 at a higher energy than the electrons 56. The energy level ("Ring Energy" 58), the energy barrier 21⁄2 can block these electrons 56. The energy barrier structure of filter 52 thus provides a filtering mechanism that produces a high pass filter effect on tunneling - charge carriers 56. This filtering effect is unique in that it has slightly different filtering effects affecting high energy electrons (e.g., hot electrons 32) in the description associated with Figure 4. Although the Fig. 7 shows that the filter 52 has the TD 53 and the BD 54, this figure is only used as an example, and any additional layer, ^ can be used as long as it has an energy barrier suitable for controlling the charge flow. Such additional layers may be semiconductor or dielectric and may be disposed between TD 53 and BD 54, or may be disposed adjacent to only one of TD 53 or BD 54. The conductor-filter system of Figure 7 is unique in that it is capable of supplying transport charge carriers with a tight energy distribution. This capability is the result of the "low pass" carrier supply function of conductor 50 and the high pass filter function of filter 52. The conductor-filter system of Figure 7 combines these two functions to provide a "Band-Pass" filtering function that allows the transport of charge carriers with a narrow energy spectrum in the energy distribution. This belt pass filter function is another embodiment in which the filter 52 provides a φ filter function. Typically, the energy spectrum is in the range of 30 meV to 300 meV. The filter 52 provides an over-effect of electrons that allow energy to pass above the threshold energy 58. This causes the electrons in the peak distribution 57p to be allowed to pass and the electrons in the tail distribution 57t to be blocked. The energy distribution 57' of the electron 56' is used as an example to illustrate the "bandpass" filtering function provided by the conductor-passing device of Fig. 7, and the energy distribution 57, and the peak distribution of the energy distribution 57 is 57p. Similar to illustrate this effect. In order to achieve the optimal r-bandpass filtering effect, the energy level of the energy distribution 57' can be individually adjusted by adjusting the energy level of the threshold energy 58 to be higher or lower than the energy level shown in Fig. 7. 'Narrow or widen to achieve. The ability to adjust the energy spectrum Δ57 is desirable because it can modulate the "bandwidth" of the band through the filter to create a filtering effect in any practical application. This can be adjusted across the filter 52. Additional voltage or other parameters will be implemented as described in paragraph 0881-A31715TWF(5.〇) 18 (§: 1287292).

In the process paper of the 7th ® shaft and the process, the dielectric constant of the dielectric fiber milk 53 which is based on the town is large. First, this can reduce the BD 54 2 electric field, and the reduction of the electric field in 54 will reduce the penetration rate of the electrons in the tail distribution and increase the electrons. Furthermore, the Nugau-Electrical test = the ship's response time _ 54 has a large dielectric constant, which will allow most of the ranks to cross the TD 53. The extension of the external power supply is called the voltage conversion between the cross-streets of 53. Therefore, it has the (4) contribution to the application of the electric field, the increase of the external power supply, the sensitivity reduction, and the increase of the tail end of the female block. The advantages of the Lin (4) item. In addition to this, during the construction of the filter 52 of Fig. 7, the other parameters are also the energy spectrum Δ57. One of these parameters is the conductive strip offset 55 between the coffee and the fourth. The conductive strip offset 55 can be modified to a unique value conversion of _% by volume, and electrons exceeding 58 in the sample distribution π beyond the threshold month b can be passed through the filter 52. This can be achieved by appropriately selecting ^ and TD = material. In a specific example, when a material with an oxide of td π is selected, the dielectric film of the NOx system (Si〇xN.x) will be an excellent candidate for the 54- because it has a wide range of Confirmed and worthy of film quality and process (4). in

= ization to the expression towel, X" represents the proportion of the oxide of the oxynitride film or an equivalent percentage of the oxide. For example, X = 1 means that the film is a simple oxide; similarly, (4) means that the film is a simple compound. When the portion of the oxide χ changes from 〇 to i, the conductive strip is offset

The amount 55 will change from about 1 eV to 〇eV β. Modifying the proportion X of the Si〇xNi x bismuth compound allows the offset of the conductive strip of the filter 52 to be modified to a desired value, thereby providing adjustment energy. Sy Δ 57 (that is, ▼ passes the "bandwidth out of her") of the device to the method of using the range in the actual application. s such as TD 53 and BD 54 and the Fermi level of the conductor 50 and other parameters can also be used for the method of the level of mmpm* 58 ^ 58 i650, and thus the bandwidth of the bandwidth device. In 〇881-A31715TWF(5.0) 19 d 1287292 - these parameters will be considered during the construction of the conductor-filter of Figure 7. For the purpose of explanation, it is assumed here that polycrystalline germanium, oxide, and nitride are used as the materials of the conductors 5, TD 53, and BD 54, respectively. It is also assumed that the oxide used in TD 53 has a thickness of 30 angstroms (A). Figure 8 shows the energy levels of the threshold energy 58 relative to the Fermi level of 16 s in both cases. The negative energy relative to the Fermi level is a negative value corresponding to the energy level of the threshold energy is lower than the Fermi level, and the difference between the threshold energy and the Fermi level corresponds to the filter passing through the filter. bandwidth". The difference between the two cases is in the Fermi level of the polycrystalline stone (N+ polycrystalline stone versus P+ polycrystalline dream) and the applied voltage Va across the transition 52. The applied voltage Va determines the amount of kinetic energy of the electrons 56 after they have passed through the filter 52. Referring to Fig. 8, for the case where P+ polysilicon and Va are _4, when the thickness (TBD) of the BD 54 is reduced from 30 angstroms to 20 angstroms, the threshold energy 58 is at the Fermi level. The range of variation is approximately 0 eV to 0. 4 eV. In the case of N+ polysilicon and Va being -3V, the figure shows that when the thickness (TBD) of the BD 54 is between 50 angstroms and 20 angstroms, the threshold energy 58 has a Fermi energy level of 165 〇. Large range of variation (approximately 〇·8 eV). It should now be understood that the adjustment of the Fermi energy level of the threshold energy relative to the conductor can be adjusted via the adjustment of the TD thickness and the BD thickness in the filter, and/or can be adjusted via the conductor Fermi level. Adjust it. This method can be used to modify the bandwidth of the transferred charge to the desired level. The kinetic energy of the transfer charge can be used by the square wire (4) and the cat is allowed to apply. The conductor-^ filter system of Figure 7 can be used to provide a filter function for other types of charge carriers. This Wei loader is, for example, electricity; _ such as LightHGles (LH) or heavy holes (Heavy Holes ; HH)). By considering the tunneling of the filter % formed in the energy band of the band, the similar considerations for electrons in the description of the P early and the 7th and 8th pins can be easily applied to these companies. The pie hole has opposite charge polarity with respect to electrons, so by inverting the polarity of the voltage across the filter 52 in Fig. 7, the band pass function of the hole can be implemented. It will also be apparent to those of ordinary skill in the art that, in applying the teachings of the present disclosure, the dielectric of the filter can be altered to thereby modify the energy distribution of the charge to the prison. For example, 0881-A31715TWF(5.0) 20 1287292, although the dielectric constant of the BD 54 is greater than the dielectric constant of the buckle 53, it can be understood that the BD 54 can be used in the teaching of the leak. The material changes the dielectric constant of the wire and the like to effectively allow the passage of charge carriers within the peak distribution during the passage of the transmission. In addition, it is not necessary to find TD 53 fine 54 as a material with uniform chemical elements, but allow the middle elements to be changed ^ and any suitable dielectric material, such as alumina (Al2〇3), yttrium oxide_2) , titanium oxide _, oxidation error division 2), oxidation (lie) ... and so on, can be used to replace oxide, nitride money oxide. Moreover, the responsibility of these materials

What kind of mixture or its Qinglai alloy, lychee oxygen record _ secret material alloy Qing ^, give - oxygen brain combination ) Na), give - the alloy of the alloy _ called .. material, can be used to replace oxide or nitride. Figure 9 illustrates an embodiment of an energy band diagram of a charge injection system for injecting a charge having a tight energy distribution. The energy band structure of this charge injection system is illustrated by the injection of electrons. Referring to Fig. 9, there is shown a conductor-passer system 59 which belongs to the seventh-description_type, conductor-insulator system ω, which is characterized by the type described in Figures 1 and 5, a charge storage. A region (hereinafter referred to as csr) 66, a channel dielectric (hereinafter referred to as CD) 68, and a body 7 〇. Figure 9 shows the complete energy band structure. • For example, in the conductor-filter system 59, in addition to the conductive strips 1853 and 1854 shown in Fig. 7, there are also valence strips 4453 and 4454. The conductor-filter system 59 includes a pass-through (hereinafter referred to as TG) 6b and a charge filter 52. The charge passer 52 includes potential barriers 2453 and 2454, wherein the potential barrier 2454 has a threshold energy 58 to control its filtering effect, as described in the first, ® phase. Charged magnetic field 52 also includes a fine dielectric (TD) 53 and a blocking dielectric (4) 54' as described in relation to Figure 7. The conductor-insulator system 6A includes a ballistic gate (10) and a BG) 62 and a _ _ _ rib 64 as secret conductors and insulators, respectively. The construction of the region of the charge injection system in which the TG 61 extends to the area of the milk 64 is the phase of the conductor 52 in the conductor-filter system 59 and the conductor (BG62) in the conductor-insulator system (9). contact". TG61 and BG62 are metals with a work function, and their work functions are 0881-A31715TWF(5.0) 21 (S) 1287292 - they have Fermi level 1601 and 16 bits. The CSR 66 is shown to utilize a dielectric Rp 64 and a CD 68 to insulate the BG 62 from the body 70, and a semiconductor including an N conductivity type, and the semiconductor has a conductive strip 1866 and a monovalent electrical strip 44. The CSR 66 may include another conductivity type semiconductor (such as a P type) and may include a metal or other material suitable for storing a charge (such as a nano-particles or a well located in a dielectric ( Traps)). The body 70 includes electrical conduction; a strap 1870 and a valence strap 4470, and can be used to modulate an image force barrier 2464 located within the conductor-insulator system 60 by coupling a voltage via 〇68 to 〇81166. The dielectrics RD 64 and CD 68 are shown as a single layer, respectively, but in general may comprise more than one layer to form a mixed layer. Figure 9 is a more detailed description of the formation and implantation of charge with a tight energy distribution. The figure shows that the ambient temperature electrons 56 having a compact energy distribution 57 are supplied by the TG 61 as a supply carrier. These electrons 56 are filtered by the filter 52 during the passage through the filter 52 using the mechanism described in FIG. After being filtered, the ambient temperature electron 56 becomes electron 56, and the energy distribution 57 of the electron 56' is closer to the energy distribution 57 of the normal temperature electron 56 before filtration. These electrons 56' are injected into the conductor-insulator system 60. In one case, a portion of the electrons 56 can be transmitted through the BG 62 with a kinetic energy 33 higher than the Fermi level 1662 of the BG 62 without Scattering (meaning "Ballistic Transport") ), and then becomes an electron 37 at the junction of BG 62 and RD 64. Such electrons 37 (called "Ballistic Electrons") do not experience scattering events with other particles (such as electrons, phonons, etc.) - thus retaining kinetic energy and traveling along the original direction Momentum. In another case, the electric/sub-56' can be partially scattered with other particles when transmitting through the BG 62 (meaning "partial ballistic transmission") and its kinetic energy 33 can still be maintained high enough, and can be maintained to bg 62 travels in the direction of the junction with RD 64, which in turn becomes electron 37. In all cases, such electrons 37 utilize the mechanisms described in relation to Figures 3B and 5 to overcome the barrier height 2〇 of the image force potential barrier 2464, and enter the conductive strip 1864 of the RD 64 and become all the way forward. An electron 37' having an energy distribution 38 is finally collected by the CSR 66 and becomes an electron knife within the conductive strip 1866. This formation and 0881-A31715TWF(5.0) 22 1287292: The process of injecting charge (whether by ballistic transmission or partial ballistic transmission) is called the Ballistic-Charge Injection Mechanism. When electrons are selected as charge carriers, such a mechanism is called ballistic electron injection. The injection efficiency of such electrons (defined as the ratio of the number of collected carriers to the number of supplied carriers) is typically in the range of 1 〇 4 to 10-1. This injection mechanism can be further enhanced by injecting piezoelectric electrons (Piezo-Electrons) (see the piezoelectric ballistic electron injection mechanism in the related description of Section 17B). The ballistic charge injection shown in Figure 9 illustrates ballistic electron injection by applying a voltage between TG 61 and BG 62 to make the kinetic energy 33 of the electron 37 higher than the conductor. The image energy barrier of 60 is achieved by a height of 20. Such a voltage can be reduced by reducing the image force potential barrier 41 as described in the 3A, 3B and 3C diagrams, and the implementation method, for example, can By coupling a positive voltage (say, from about + lv to about +3) to cSR sinking. Alternatively, the work function of the material by selecting CSR66 is greater than the work function of BG 62 (or Fermi level) It is lower than the Fermi level of BG 62) to reduce the image potential energy barrier. The first diagram provides a real complement of the charge-charging diagram of the charge-injection system. An electron with a tight energy distribution. In the conductor_filter system of Figure 1G, the conductor 61 supplies a normal temperature charge carrier 56. The filter 52 is in contact with the conductor 61 and includes dielectrics 53 and 54. To provide a filtering function for a charge carrier 56 (negative charge carrier) of a certain polarity, wherein The buffer 52 has electrically deformable potential energy barriers and controls to control the flow of a certain polarity charge carrier 56 through the filter 52 in a certain direction (toward the other direction 34). In addition to the control of the sub (negative charge carriers, electrons 56), the filter 52 further includes a potential energy barrier 43⁄43 and 4 that can be modified to control the direction substantially opposite to one of the above directions. Through the flow of the opposite polarity charge carriers (positive charge carriers, lh η and book 73) of the filter 52. 〃 / This filtering function allows the charge carriers of a certain polarity to be along a forward direction 34 (ie by Tg6i transmits at BG 62) and blocks charge carriers in a backward direction 74 (i.e., from bg 62 to (4). Therefore, the passer 52 provides a charge that can "purify" the charge flow. 〇881-A31715TWF(5.0) 23 8 1287292 Power Month b. This charge filtering function is another embodiment of the filtering function provided by the filter 52. Except for the energy band diagram of the 10th and the 9th In addition to the differences, the other parties are the same in white. These differences are in the 1G map. The metal is used as the conductor-filter system, and the material of the conductor region in the conductor-insulator system 60. These conductor regions (ie, tg 61 ^ BG62) are now made of semiconductor materials, wherein the semiconductor of TG61 has a conductive strip 1861 and a price. The semiconductor of the electric strip 44β1, and BG a has a conductive strip and a valence band of 44. The TG 61 shown in the figure is a Ρ-type semiconductor, and has electron % as a supply carrier in its valence band a. The electron 56 and its energy distribution 57 undergo a transmission process that is related to the description associated with Fig. 9, so that a portion of the electron 56 can enter the CSR 66 and become an electron 37 having an energy distribution, and finally described in relation to Fig. 9. A similar approach is collected and stored on CSR 66. For the example shown in Figure 10, when a polarity is applied such that the electrons 56 of the TG 61 are injected with a voltage in the 刖 direction 34, the LHs 72 and 73 in the BG 62 are also caused to be in the rearward direction 74. transmission. LH π and booklet 73 transmitted later may cause undesirable problems. For example, when LH 72 and HH 73 are transmitted back into TG 61, the impact may be triggered by the energy of the mussel electric charge ▼ 1601 in the moon ( (this is the case (5) ). In addition, these holes do not contribute to memory operations when ballistic electron injection is used to program memory cells. Therefore, this may waste current and thus consume power. Therefore, it is desirable to block the LH 72 and HH 73 ' from being transmitted back into the Tg 61. The energy band structure of Figure 10 shows that the carriers that are transmitted later (ie, LH 72 and HH 73) must pass through a larger number of energy barriers than the forwardly transmitted carriers (ie, electrons 56) and thus provide a barrier. Afterwards, wear the filter and effect of the carrier. This filtering effect is based on the energy band structure formed by the potential barrier in the filter 52. The first potential barrier 42 that blocks the rearward transmission holes 72 and 73 has barrier heights 4154 and 41, 54 on the entry side and the exit side, respectively. The two barrier heights 4154 and 41, 54 are based on the BD 54 price strip 44^. The second hole energy barrier 4253 has an energy barrier height 4I53 on the eight sides which forms another barrier to the holes 72 and 73. The height of the energy barrier is 4153 0881-A31715TWF (5.0) 24 1287292 — The TD53 price band 4453 at the junction of TD53 and BD54 is used as a reference point. ‘ This filter 52 is based on a highly engineered concept of energy barriers. A particular embodiment of a conductor-filter 59 and a conductor-insulator system 60 for illustrating the high degree of engineering of the barrier includes a P+ polysilicon constituting TG 61 and an oxide layer constituting TD 53. A nitride layer constituting the BD 54 constitutes a polycrystalline germanium of BG, and an oxide layer constituting the RD 64. The use of N+ polysilicon as the material of BG62 comes from several considerations. Among them • The most important consideration is that the s〇lid solubility of N-type impurities (such as arsenic, phosphorus, etc.) is higher than that of P-type impurities (such as boron). It is hoped that the impurity can have a higher solid solubility. Therefore, the doping of the stone is doped with a relatively high concentration, so that the resistance of the doped stone plate is lowered, so that it can be suitably used in an integrated circuit. . In this embodiment, the original S using polycrystalline germanium as the material of TG61 and ^62 is a widely proven benefit, mass productivity, and compatibility with today's 1 (=; test thickness. The oxides of nanometers to 1 nanometers are also derived from the same reason. The thickness of the oxide layer constituting TD 63 may be between about 15 nm and 4 nm, and preferably about 2 nm. Between 3 and 5 nanometers. The thickness range of TD 53 is chosen to be able to transmit the charge carriers (electron, LH or HH weeping is transmitted by direct randomization. BD % thickness is selected as ' When - a moderate voltage in the range of about IV to about 2.5V is applied between the crime 61 and the yang 62. 旎, '旎 阻挡 阻挡 电 电 电 电 电 电 电 电 电 电 电 电 BD 。 。 。 。 。 。 。 。 。 。 。 。 。 Therefore, at higher voltages (above 3V), it is possible to allow certain types of charge carriers (such as 'electron 56) to be transmitted separately, and to block another type of charge carriers (for example, LH% is transmitted later). In the following concept of the obstacle, the thickness of the BD54 is also determined by the dielectric constant of the BD54. Generally speaking, if The device 52 can indeed meet the above requirements, and the thickness of the BD 54 can be 53 (thick) or thick or thin. For example, in this particular embodiment, if the TD 53 is selected to be thick, the card is (ie, 3 angstroms). For oxides, the minimum thickness of the BD 54 can be about 2 nanometers (i.e., 2 angstroms) or thicker. For this particular embodiment, the oxide constituting the butyl (tetra) can be made using conventional product (P) techniques. The high temperature oxide (TempCTatoe oxide; TE0S layer or 疋 using the well-known chemical (orphan _aI 〇 fiber ship) technology made by 0881-A31715TWF (5.0) 25 (ί 1287292 " into a hot gas ( Thermal Oxide. The nitride that makes up BD 54 can be a high-quality nitride without a charge trapping center in the band gap. This high-quality nitride is manufactured by way of example, in ammonia (Ammonia). ; NH3) environment is manufactured by high temperature (for example, 1〇5〇.c) - rapid thermal nitridation (Rapid Nitridati (10); rtn) technology well known in the art. Ballistic charge injection energy barrier high engineering It is now intended to provide details on the high degree of engineering of energy barriers. Figure 11 shows Figure 10 is similar to the energy band diagram of the climbing, but the band bending in the face 62 is relatively slight, in order to reveal more details of the height of the barrier. In addition to showing the area and reference symbols shown in Figure 10, Figure n also shows the energy barrier height 41,53 of the price band offset between the price band 4462 and 4453. The barrier height 41, '"53 is the second transmission to block the LH 72 and HH 73. The potential of the hole potential energy barrier is %. In addition, the figure also shows a first electronic potential energy barrier 24 formed by the TD 53, which has an energy barrier height i and 2 at the entry side and the exit side of the potential energy barrier % of the blocking electron 56 forward. 〇 '53. Also shown is a second electron energy barrier % formed by BD 54 having a barrier height of 2054 and 2, 54 on the entry side and the off side, respectively. The second electrical barrier 2454 also has the effect of blocking the electrons 56 injected into the front φ. As can be seen from the energy band structure provided by the present invention, there are two electronic potential barriers 2453' and 2454 associated with the forwardly transmitted electrons 56. Similarly, there are two hole energy barriers 4254 imaginary 4253 associated with the holes 72 and 73 transmitted in the BG 62. In order to produce a highly efficient ballistic charge, it is desirable that the energy barriers of the first and second electronic energy barriers 2453 and A4 be electrically configurable to assist in the transmission in the forward direction 34. On the contrary, in order to block the hole 72 in the bg 62 from passing back to the TG61, it is desirable to have the energy barrier height of the first hole energy barrier 4254 and the second hole in the entire wire of the piezoelectric ballistic charge. The energy barrier of the energy barrier 4253 is held high enough. Half reference Figure 11 'The energy barrier height of the first electronic energy barrier 2454 is 2〇54 (the main item of Δφ ^ 0881-A31715TWF(5.0) 2 8 1287292 can be expressed by the following formula: one (1) = M> C5, -丨The factory cuts among them

"It is the offset of the conductive band between TG 61 and BD 54 under the condition of flat energy. The voltage drop across the τ during the piezoelectric ballistic electron injection is not the β 疋 test across the TG 61 and BG. The applied voltage of 62 (that is, across the filter ^ is the flat band voltage; ~ is the band gap of 1 ^ 61; ^ and ~ are the dielectric constants of TD53 and BD54 respectively; and the heart and heart are TD respectively The thickness of 53 and BD 54; , like the second hole that blocks the transmission hole, the energy barrier of the second hole can be high. 4153 (anti-sink) can be expressed by the following formula: ΔΦ,

GT ΔΦ,

GT - 丨厂肥丨 where - (2)

"The valence band offset of BG 62 and tD 53帛 under the condition of 疋平 can bring the ribs to the electron injection during the electron injection period and can be expressed as L. The above formulas (1) and (7) can be understood that The height of the barrier is 2〇54(ΔΦ”) and the height of the barrier is 41« (the relationship between the two is different. The relationship between the voltage and the voltage is asymmetric) and depends mainly on the dielectric constant and the dielectric constant. The combined effect of the thickness of the electrical material (ie, the "ε τ effect"). Figure 12 shows an example of a highly engineering concept of energy barriers using the above principles to perform ballistic electron injection. It can be clearly seen that when the buckle 61 and When the voltage is applied to 62, the height of the electron barrier in the TG61 is 2〇54 (secret), which is faster than the barrier height of the LH 72 and ΗΗ73 in the BG 62 = 3 (δφ). Above, when the applied voltage is about 1·5'ν, the energy barrier height 2〇54(ΔΦ−) is transferred to the Fermi level (ie, the energy barrier height is equal to zero), however, 0881-A31715TWF(5.0) 27 8 1287292: At this time, the energy barrier height is 4153 (ΔΦ"), but it is still maintained at a sufficient energy barrier height of about 丄 (4). Reduce the energy band diagram beyond the voltage level. As shown in the figure (7), when the applied voltage falls below this voltage level, the second potential of the blocking electron 56 in the U-shaped sub-potential energy barrier The 2454 is now located below the Fermi level 1601. Therefore, the electrons 56 in the TG 61 with energy above the threshold energy 58 can be transmitted directly through the filter 52 without being blocked by the BD 54. This allows a conductor-filter system The belt of 59 injects electrons moving in the forward direction 34 and having a tightly dense energy distribution 57 through the filter function. In this voltage range, the relationship between the energy barrier height 4153 (Δφ^σΓ) and the applied voltage is much weaker. The second hole potential energy barrier ♦ 4153 (Δφ^-π), which avoids the transmission of holes later, is thus maintained. Therefore, the energy barrier engineering concept described herein can actually provide electrical conductivity for ballistic electron injection. Modify the filter construction method. This filter is unique in that it can filter out unwanted carriers (such as LH 72 and ΗΗ 73 transmitted later) without affecting the transmission of the desired carrier (for example, forward) The transmitted electrons 56). The above statements about formulas (1) and (2) The results shown in Fig. 12 are taken as an example to show the relationship between the energy barriers 2〇54 and 4Ι54 and the voltage. Other energy barrier heights of the filter 52 in Fig. 11 can be easily applied (e.g. The energy barrier heights 24〇, 53 and 20'54 of the energy barrier 24% and 24μ, and the energy barrier heights 41, 43 and 4153 of the second cavity potential energy barrier 42% are the same ρ description. It is understood that the relationship between the energy barrier height of the potential energy barrier for controlling the transmission of the charge carriers and the filter crossover pressure is weaker than the energy barrier height of the potential energy barrier for controlling the forward transmission of the charge carriers. Ρ4 is within the voltage range generally used for ballistic electron injection. It is hoped that the cross-pressure (Gd) of the BD 54 can be smaller than the energy barrier height of 4154. It is desirable to make the heart smaller than the barrier height 4154 because the first hole potential barrier 4254 in the BD54 region can be maintained as a trapezoidal band structure to more effectively block the LH 72 and HH 73 injected later. This energy barrier structure can be more clearly understood by referring to FIG. The figure shows that the barrier height 4154 forms the energy barrier height of one of the first hole potential barriers 4254 (the entry side of the holes 72 and 73), and the barrier height 41'54 forms the first hole potential barrier. The height of the barrier on the other side (the exit side of the holes 72 and 73). The trapezoidal first hole potential energy barrier 4254 is separated from 0881-A31715TWF(5.0) 28 1287292: the main item of the open side energy barrier height 41'54 is equal to ΔΦ', where ΔΦ^_ is still the energy barrier height' 41m. In the specific embodiment of the band structure shown in Fig. 10, when the applied voltage between #TG 61 and BG 62 is -4 V, the energy barrier height 41, 54 is about 〇JeV, and thus the potential of the first hole The trapezoidal structure of the barrier 4254 remains. As appreciated by the above principles, the energy barrier heights 41, 54 can be increased by optimizing the dielectric constant and thickness of the TD 53 and BD 54 _ to reduce & - For this particular embodiment, the voltage range of the TG 61 is selected to be relative to the voltage of the BG 62 - between -3.5V and about -4. 5v to effect ballistic electron injection. As can be seen in the description of Figures 3A, 3B and 3C, such a _ voltage can be further reduced by reducing the image energy barrier height 2〇 of the conductor-insulator system 6〇. This can be achieved by combining - between about IV and about 3V to CSR66. Alternatively, the image force potential barrier can be reduced by selecting the work function of the material of CSR 66 to be lower than the work function of BG 62 (or the Fermi level of BG62). Reducing the image power barrier to reduce the voltage between TG 61 and BG 62 is a desirable effect for the present invention. One of the main advantages is that it can reduce the electric field in the dielectric between TG 61 and BG 62, thus avoiding the problems associated with high electric fields in the dielectric (such as dielectric breakdown, which causes dielectric properties). Permanent damage). " According to another embodiment of the present invention, the filter 52 further provides a voltage splitting function. Voltage • The split function reduces the voltage drop across the dielectric in filter 52. Fig. 12B shows an example of the voltage division function using the above-described energy barrier high engineering concept to perform ballistic electron injection. Referring to Figure 12B, the relationship between the different dielectric cross-pressures and the cross-pressure of the filter 52 is shown. It is apparent that the filter, the crossover of the device 52, that is, the applied voltage between the TG 61 and the BG 62, is divided and shared by several regions in the filter 52. This voltage splitting function provided by the filter phantom allows the applied voltage between TG 61 and BG62 to be split and shared by bd 54 and TD53 without compromising ballistic charge injection. This voltage splitting function reduces the voltage that each dielectric of these dielectrics must suppress, thereby avoiding dielectric breakdown. One of the features of the present invention is the effect provided by the highly engineered concept of energy barriers and the equivalent of O881-A31715TWF(5.0) 29 1287292 • The practice of injecting into the filter. These effects provide voltage splitting and avoid dielectric breakdown problems that may occur during charge injection. In addition, the problem of impact detachment occurring in the TG 60 triggered by the transmission of charge carriers in the future is also effectively avoided by applying a filtering effect to suppress the backward injection of these carriers. • It will thus be appreciated that the filter and band structure described in the present invention can effectively block the transmission of a certain polarity charge carrier during ballistic charge 'Note', while allowing opposite polarity - charge carriers Transfer to. Thus, filter 52 provides a means of over-charging the charge stream "purified". Although not necessary, it is generally desirable that the Fermi level of the BG 62 is located in the center of the energy band gap of the BD 54 in the filter 52 under the condition of the flat band, so as to construct using the band structure and the injection mechanism. In the case of cells, the charge filtering mechanism can be utilized to the fullest extent. The above description of the ballistic charge injection and the highly engineered concept of energy barriers is for electronics. A similar description can be made of the mixed effects achieved by the light and heavy electricity charge. Figure 13 provides an energy band diagram illustrating the ballistic charge injection and effect of the holes in the charge injection system of Figure 1. In the conductor_passer system section shown in Fig. 13, the conductor 61 supplies the normal temperature charge carriers 75 and 76. The filter % is in contact with the conductor & and includes dielectrics 53 and 54 to provide a filtering function for charge carriers 75 and 76 (positive charge carriers) of a certain polarity, wherein the filter 52 has Electrically modified potential energy barrier 42 training and heart, to control; Che H (10) gift amount 34) through the brain 52 of the hybrid f rape 75 and 76 flow. In addition to controlling a polar charge carrier (positive charge carriers 75 and 76), filter 52 also includes an electrically configurable potential energy barrier and 24 tear to control substantially along one of the above directions. The flow in the opposite direction (backward direction 74) through the opposite polarity charge carriers (negative charge carriers, electrons 84) of the filter 52. This kind of jujube scale has a polarity of charge carriers transmitted along the _ front direction 34, and blocks 74 im 〇. "Purification" charge flow charge filtering function. This charge filtering function is provided by the transitioner 52 as a further embodiment of the 0881-A31715TWF (5.0) 1287292: and is similar to the charge filtering function of the description associated with Figure 1G. Referring to Fig. 13, it is shown that LH75 and 76 are located in the valence band of tg6i, and the charge 虬Η 75 and HH76 used as injections are shown to have an energy distribution 77 and are transmitted along the forward direction 34. Although H. 75 and ΗΗ 76 have the same energy distribution in the figure, however, they have no _ effective mass, _ can have no _ energy score _ cloth 0 • • In Fig. 13, LH75 and ΗΗ76 two All use the quantum mechanical transmission mechanism to transmit the energy barrier through the filter 52 to become LH75, and 76. LH 75, in contrast to HH W, has a kinetic energy 46, and this kinetic energy 46 is slightly higher than the energy barrier height 41 of an image energy barrier. When these carriers continue to travel in the forward direction, the transmission behavior in the BG 62 is greatly different due to the difference in effective quality. As far as jjjj% is concerned, the average free path can be very short because its effective mass I is heavier. Therefore, Book 76 is susceptible to scattering events with other particles (such as phonons), resulting in low ballistic transmission efficiency (Ballisticity). In Fig. 13, HH 76, which is experiencing a scattering event and thus loses energy, becomes HH 79. Furthermore, it is also shown that these scattered hh 79 have an energy distribution 81 which is wider than the original energy distribution 77 due to scattering. These holes 79 are transmitted in the figure at an energy level of less than 40, which is an energy barrier of 64 in the price band 4464 of the 64 • price band 4464, and thus are blocked from being able to pass the energy barrier 42 and cannot enter. CSR 66. On the contrary, LH 75 has a lighter effective and quality, so its mean free path is longer than HH 76, and the average free path is long (for example, in 矽^, the average free path of LH is about HH mean free path. three times). In one case, a portion of this kt plant '-hole LH 75' can be transmitted by kinetic energy 46 through BG 62 without scattering (meaning ballistic transmission), and then at the junction of BG 62 and RD 64 LH 78. This type of LH 78 (called "ballistic LH") does not experience scattering events with other particles (such as phonons), thus retaining kinetic energy and impulses along the original direction of travel (M〇mentum), and its energy distribution 80 is the same as the original The energy distribution 77 is similar. In another case, the LH 75 can pass through the BG 62 with partial ballistic transmission, and its kinetic energy 46 can still be maintained high enough to maintain the direction of the boundary between bg 62 and RD 64. 0881-A31715TWF (5.0) 31 8 1287292 'Into, then become LH 78. In all cases, this type of LH 78 utilizes the mechanism described in relation to Figure 6 to overcome the energy barrier height 41 of the image force potential barrier of 42 m, and enters the rainbow > 64 price band 44 and proceeds all the way into The energy distribution 80, LH78, is finally collected by CSR66 and becomes a hole 82 in the valence band 4466. This process of forming and injecting holes (whether using ballistic transmission or partial ballistic transmission) is called the ballistic hole injection mechanism - Injectkm Mechanism. The injection efficiency of such holes (defined as the ratio of the number of carriers collected to the number of carriers supplied) is typically between ΙΟ·6 and ΗΓ3. This injection mechanism can be further improved by injecting a piezoelectric hole (Piezo-Holes) (see the piezoelectric ballistic hole injection mechanism described in Figures 17B and 17C). For a particular embodiment of the materials described in connection with systems 59 and 60 using the first drawing, the voltage range of TG 61 is selected to be between +5V and about +6 〇v relative to the voltage of BG 62 to effect ballistics. Hole injection. Such a voltage can be further reduced by reducing the image energy barrier 41 of the conductor-insulator system 60 as described in relation to Figure 6. For example, this can be accomplished by coupling a voltage between about 3V and about a IV to the CSR 66. Alternatively, the image energy barrier can be reduced by selecting CSR 66 as a material having a smaller work function (or a higher Fermi level) than BG 62. φ By using materials having similar Fermi level to form TG 61 and BG 62, the applied voltage between TG 61 and BG62 can be further reduced. This constitutes another embodiment of the material used for the ballistic hole injection, a filter system 59 and the conductor-insulator system 60. For example, the filter 52 may include a P+ polycrystalline stone constituting TG 61, an oxide layer constituting jd 53, a nitride layer constituting BD 54, and a P+ polycrystalline stone constituting bg 62. The layers, and - constitute the oxide layer of RD64. Such an embodiment allows the voltage of 361 (with respect to the voltage of β (} 62 to be selected in a small voltage range when performing ballistic hole injection (for example, about +4·5 v to about +5.5 V). Figure 13 further shows that while the band structure is biased for transmission in the forward direction 34, the electron spines in the conductive strip_ of the BG 62 can be 0881-A31715TWF along the new rear direction % ( 5.0) 32 1287292 Transmission. This subsequent transmission of electrons 84 can cause problems such as impingement of the TG 61, waste of current and power, etc. These problems are similar to those caused by the transmission of holes later in the description of Figure 10. Therefore, it is desirable to be able to block the electrons 84 by the filter 52 so that it cannot be transported backwards into the band structure of the TG6. Figure 13 shows that the carrier (ie, the electron 84) that is transmitted later must be transported more than the carrier ( That is, LH 75 and HH 76) traverse a large number of energy barriers. The first block to transmit electrons 84 is a potential energy barrier 24, and the entry and exit sides have energy barrier heights of 2〇5413 and 2〇, respectively. 54b. These two energy barriers are 20^1^20^, which are the conductive belts at the junction of BD 54 and BG 62. And the conductive strip I854 at the junction of TD 53 and BD 54 serves as a reference point. The figure shows that the second electron barrier 245% of the entry side has an energy barrier height of 205% and forms another barrier hole 84. The energy barrier south degree 2〇53!3 is the reference point of the TD 53 conductive strip 1853 at the junction of TD 53 and BD 54. The energy barrier height 20'53b (not shown) exists on the exit side of the energy barrier 2453b. And the conductive strip 1853 of the TD 53 at the junction of TG 61 and TD 53 is used as a reference point. In the example illustrated here, the energy barrier is higher than the energy level of the electron 84, and thus is not shown in FIG. The energy barriers 2454] 3 and 2453b form an energy band structure in the conductive strip of the damper 52 to block the electrons 84 that are transported backward. The holes 75 and 76 are in the forward direction 34. Two similar energy barriers. A first potential energy barrier 42 is now formed by the TD 53 and has an energy barrier height 41 and a 41' respectively on the entry side and the exit side. A second potential energy barrier 42 is determined by the BD 54 Formed, and its entry side and exit side respectively have an energy barrier south and a slap (not shown). Both π hunger and 42^ form an energy band structure in the valence band of the filter 52, and have the effect of blocking the forward-passing holes 75 and 76. In Fig. 13, the band structure is injected. The purpose of the hole is to be biased. The two energy barriers and the 41'Mb are located below the energy level of the forward transmission hole, and thus are not shown in Figure 13. Figure 14 shows the energy of the present invention. The obstacle height is called the effect of the bomb hole injection. It shows that the height of the energy barrier that blocks the electrons transmitted backward is 2〇, and the 5th and the transitional device are generally evaluated. ^ stomach 61 and 0881-A31715TWF (5.0) 33 Ϊ 287292 BG 62 The relationship between the voltages is weaker than blocking the relationship between the forward energy barrier height 41 tearing and the filter 52 遂 drop. Therefore, the two energy barrier heights 20'5 and 41Mb can be changed to different degrees by the pressure across the filter 52. As explained in the high-impact engineering, the relationship between the barrier height and the voltage is asymmetric and is dominated by the combined effect of dielectric constant and dielectric thickness (ie, the "εΤ" effect). It can be seen that when the applied voltage between tg 61 and BG 62 is increased, the barrier height 4154b of the holes 75 and 76 in the blocking TG 61 is higher than the barrier height 20 of the electrons 84 in the blocking BG 62. 54b is lowered faster. In fact, when the applied voltage is +3·5ν, the energy barrier height of 41Mb is transferred to the hole energy (ie, the energy barrier height is equal to zero), but at this time the energy barrier height is 20, but it is still maintained at about +2.5 eV. The height of the barrier. Figure 13 shows the band diagram when the applied voltage is increased beyond this voltage level. As shown in Fig. 13, when the applied voltage is increased beyond this voltage level, the second energy barrier 42Mb of the blocking holes 75 and 76 is now under the hole energy. Therefore, the holes 760 and 76 in the TG 61 can be directly transmitted through the filter 52 without being blocked by the BD 54 layer. A much weaker relationship between the energy barrier 20'54b and the applied voltage in this voltage range maintains 245% of the energy barrier of the control electronics 84, thereby preventing electrons from being transmitted later. The description made in Figure 14 above is used as an example to show the relationship between the two barrier heights 2〇, 5 and 41 and the voltage. The same can be easily explained for the other barrier heights of the filter 52 in Fig. 13 (e.g., the energy barrier heights of 20 gamma and 41 hunger, respectively). Therefore, it can be understood that the relationship between the energy barrier height of the potential energy barrier for controlling the transport of the charge carriers and the filter crossover pressure is compared with the energy barrier height and the filter for controlling the potential energy barrier of the forward charge carrier. The relationship between the pressures will be weaker. Although not shown, the filter 52 also provides a voltage splitting function when the voltage polarity between the TG 61 and the BG 62 is set for ballistic hole injection. This voltage splitting function for ballistic hole injection reduces the voltage drop across the dielectric in filter 52 and is primarily dominated by effects similar to those described in Figure 12B for ballistic electron injection. In the case of ballistic hole injection, since the voltage is higher in the description, the voltage division function can reduce the dielectric drop by reducing the voltage of the dielectric in the filter, thereby avoiding dielectric breakdown. 0881-A31715TWF(5.0) 34 8 1287292 Therefore, the concept of scales in this case can actually provide the electrical modification of the ballistic charge. This over-the-top surface can pass the hybrid carrier (for example, the carrier transmitted later) but does not affect the carrier that wants the carrier (for example, the carrier transmitted forward). Day f 'Juss 52 provides another kind of job ^ This kind of over-the-counter, function allows. A chargeer with a certain polarity and a light amount of f (such as LH), and a reverse-polarity The wire with a heavier amount of f (10) passes through _. Therefore, the snooper 52 provides a quality filtering function that filters the flow of charge carriers based on the mass of the carrier. • Figure 15 is a diagram to illustrate the basis for the quality of the filter 52. Refer to Figure 13 for the mastery of this quality magnetic Wei Qijia. In the shirt 13 _ 示之赖-% of the system, the conductor 61 is supplied with a normal temperature charge carrier (Ljj 75 and the volume %. The filter 52 is in contact with the conductor & and includes dielectrics 53 and 54 to provide The filtering function of polar charge carriers 75 and 76 (positive charge carriers), wherein the filter 52 includes an electrically configurable potential energy barrier 42 and 42^ to control >^ a-direction (frontward direction 34) Passing a certain number of polar charge carriers through the filter is a flow with 76. It is known that in quantum mechanics theory, the random rate of charge carriers is a function of its mass, # and heavier carriers (eg, 76) The tunneling rate is lower than that of lighter carriers (such as LH 75). Figure 15 shows the normalized tunneling probability of LH and the book; T^eling Pr〇babqing), And it is shown as a function of the reciprocal of the heart to illustrate the filter, the quality of the device. In this description, the filter 52 is assumed to include a TD « consisting of an oxygen-record having a thickness of 3 nm, and a BD % consisting of a nitride having a thickness of 2 nm. For the voltage range between TG 61 and BG 62 used for ballistic hole injection (+5V to +6V), the tunneling rate of 'HH shown in the figure is about 4 to 8 than that of LH. Magnitude. This difference in the random rate of wear caused by the carrier mass allows the transitioner 52 to perform the mass transition function. Although a hole carrier is used here for illustration, a similar description can be easily derived to other carriers having the same polarity but different heterogeneity (for example, as described in Figures 17B and 17C). Piezoelectric electronics). 〇881-A31715TWF(5.0) 35 (S) 1287292 This mass filtering function is another embodiment of the filtering function provided by the filter 52.

The quality filtering function of filter 52 and its application to LH transmissions bring several benefits to the present invention. For example, this mass filtering function prevents the supply carriers used for ballistic injection in the TG 61 from being wasted. The reason is that most of the hole carriers in 1(}61 are, and HH has a short average mean free path, so it is easy to experience scattering events during transmission through BG. Such HH cannot efficiently make ballistic injection. Contribute, so that it will be wasted when used as a supply carrier. The main supply charge_ is limited to the LH carrier due to the _ amount filtering function provided by the cloister 52. The LH carrier has a longer length. The mean free path, so when transmitted through bg62 via the mechanism described in Figure 13, the more effective axis can be used to record the miscellaneous. The result is that the quality of the function of 52 is characterized by the fact that Wei has a high material. The charge carriers are used as supply carriers to avoid waste of low ballistic carriers on the supply current. The filter 52 of the conductor-filter system 59 provides several unique filtering features. Providing the charge filtering function in the related description of the band passing filter, function, the third (), μ, 13, and 14 figures in the related description of Fig. 7 and the mass filtering in the related description of Fig. 15 In addition to mentioning In addition to the over-the-top function, as described in the related description of Figure 12B, the over-the-counter ^ provides a voltage-dividing type of Wei. It has a white-leaf in the field, and when it is revealed, if it is over The dielectric and/or structure of the device is modified to use it to change the function of the device individually or in a way that is still a fan of the date of the month. For example, the filter can contain two dielectrics f In order to improve its voltage division function, in addition, it is not necessary to require the dielectric chemistry of the filter, and the chemical elements in the middle can be allowed to effectively support these functions while p, ,, 〃, 朗's face It is not limited to the ones described herein and the above-described implementations, and all changes that fall within the scope of the appended claims are covered. Now, please turn to refer to Figure 16. Figure 16 provides the shovel, ', " The band structure is the same. The difference is that the BG62 is not 〇881-A31715TWF(5.〇) 8 36 1287292 = the semiconductor is the material, and the metal having the work function and the work function has a Fermi level - t material, for example, the material used as a conductor in the description of the figure. Figure 6 also shows the charge carriers in the electric strip 4461 of tg6i, that is, the electric ^6, LH75, and the book %, as described in the first, the, ίο diagram, by applying the appropriate size and polarity voltage to the TG 61 and bg weep, the read electron 56 is filtered by the filter 52 and injected on the CSR 66. Similarly, as described in the figure i4, "13" by applying an appropriate amount of electricity and polarity to the D (10) and b (10), ^ (4) The HH 76 is filtered by the filter 52 and injected on the CSR shame. #在上职上的球道充电注人的带带结构的补巾, bg & forming the main ride of ballistic charge transmission, and for the _ The loader can turn at a good speed over it, and the thickness of B(10) generally needs to be several times lower than the average free path of the charge carrier, which is typically about 1 〇 to 邡. When the thickness of the BG 62 layer is required to be so low, BG 62 is inevitably caused to have a souther sheet resistance, and thus leads to 1 (the basic problem of application. For example, in combination with the large r and large c effects, this This can lead to large signal delays (so-called rc delays), which is especially a major problem in unit operations, since this delay may limit the access rate of memory cells in a large memory array. Second, to avoid unselected The memory unit is disturbed, and it is usually required that the ideal set of external touches is applied to those unselected units. However, due to the influence of this delay, the power on the unselected unit may not be the result of the electrical waste value. Unit interference is more likely to occur. In addition, the large R value may be combined with a large current to produce the m effect. When a power county - the wire cap is still, this !r effect will cause the voltage to drop, thus making the - memory unit specified The electrode cannot reach the desired level, and the result is a negative impact on the unit operation. For example, the effect of the IR effect on an unselected unit may be interference. This unselected unit will be deliberately converted from a logical state (such as "〇") to another state (such as "1"). The effect of the IR effect on the selected unit may be the operation of the slowing down unit ( For example, programming, erasing, and reading operations). However, the above problems can be overcome by using the piezoelectric effect described below. Application of piezoelectric effect to ballistic charge injection 0881-A31715TWF(5.0) 37 8 1287292 Piezo-effect is a well-known physical phenomenon in solid state physics. When a mechanical application is applied to a semiconductor material, the piezoelectric effect changes the electrical properties of the semiconductor material (see Pikus and Bir). Symmetry and Strain-Induced Effects in Semiconductors, New York: Wiley, 1974. This mechanical stress may originate from a strain source (also known as a "stressor") inside or outside the semiconductor material. This mechanical stress may occur in a Compressive (ComPressive) version or in a tensioned form ((tensi〇n)' and can cause a Strain, which destroys the symmetry in the crystal lattice, thus deforming the potential in the crystal lattice. Some piezoelectric effects in semiconductors (such as Shi Xi) are well-known applications that include the dust electric resistance effect in the resistor. , Bipolar transistors and Piezo-Junction Effect in the diode, Piezo〇Hall Effect in the sensor, and MOS transistor (" Piezoelectric field effect transistors (Piezo-FETs) in MOSFETS"). The invention further provides the application of piezoelectric effect to ballistic charge carrier injection and transmission. Embodiments of many different memory cells and semiconductor devices will be utilized below to propose a new piezoelectric ballistic charge injection mechanism. Piezoelectric Ballistic Charge Injection Mechanism • It is known that when a strain occurs in a semiconductor, it may separate the energy valley of the conduction band from the degeneracy of the HH and LH valence bands (see Hensel Et al., . "Cyclotron Resonance Experiments in Uniaxially Stressed Silicon: Valence

Band Inverse Mass Parameters and Deformation Potentials, Phys. Rev. 129, pp. 1141-1062", 1963). Figures 17A, 17B, and 17C provide a semiconductor without strain, under Tensile Stress, and Under the Compressive Stress, the Dispersion Relationship between the energy E and the momentum vector k. Figure 17A shows the dispersion relationship of an unstrained semiconductor. One left energy valley 86 and one right energy valley 87 are two conductive bands in the valley, while the conductive band energy valleys 86 and 87 are respectively 0881-A31715TWF (5.O) 38 1287292 - with a minimum value of 86 me and 87 m. The figure shows the smallest The values 86m and 87m are at the same energy level. Since the figure shows that the dispersion curve of the energy valley has different curvatures, the effective mass of electrons in the 86m of the left energy valley is heavier than the effective mass of the electrons within 87m of the right energy valley. It also shows that the LH secondary energy band and the 1111 energy band have two dispersion curves, both of which are filled with holes 90. The LH secondary energy band is 88 and HH times. The energy band 89 is in the figure. Energy degeneracy on the surface 86m • or 87m is separated from the valence band maximum 91 by a band gap 92. A 17B chart shows a dispersion relationship similar to that of the PA chart, but the semiconductor is strained by tensile stress. The lowest value of the energy valley is the upward shift (left energy valley 86) or one downward (right energy valley 87). As a result, the total number of electrons in the two energy valleys will be redistributed. More electrons 85 are concentrated, because the energy level of the conductive band is 87m, the energy level is lower. The electrons are redistributed and mostly concentrated in the energy valley 42 is a promising phenomenon for two reasons. First, due to the conductive energy valley 87 The effective mass of electrons is lighter, so it can produce beneficial effects on electron transport in semiconductors. Second, it is known that energy valley separation can reduce the phenomenon of inter-valley electron scattering. These effects can be specifically explained by using 矽. The strain usually leads to a six-fold degenerate conductive strip that decomposes into a two-fold degenerate and four-fold degenerate energy valley, in which most electrons (nearly one hundred percent of total electrons) are concentrated in the electron transport direction. The two-fold degenerate energy can be found in the valley. It is known that the strain effect is in strain _M0SFET (a piezoelectric field effect transistor, see v〇gelsang et al, "Electron Mobilities and High-Field Drift Velocity in Strained Silicon - on Silicon-Germanium Substrate" ? IEEE Trans, on Electron Devices, pp. — 2641 -2642, 1992) will increase the electron mobility (Mobility) by 50% and the drift rate (Drift Velocity) of 16%. Similar strain effects can be applied to enhance ballistic charge carrier transmission. Therefore, the ballistic electron injection efficiency in the crucible may be improved by the re-aggregation of electrons in the two-fold degenerate energy valley. This can be achieved by applying stress to the crucible to cause strain in the direction of electron transport. It is therefore clear that the piezoelectric effect can result in tightly packed "piezoelectric" electrons (i.e., electrons in a material that is subject to mechanical stress) that have a lighter mass and a lower scattering ratio. In accordance with an embodiment of the present invention, when these effects are combined with ballistic electron transport, it is possible to provide 0881-A31715TWF (5.0) 39 8 1287292 • A piezoelectric ballistic electron injection mechanism. Although the figure does not show that such piezoelectric electrons can be used as the supply electrons of the energy band structure shown in Figs. 9 and 10, the transmission process described therein is experienced. Figure 17B also shows that the stress effect caused by the tensile stress in the semiconductor can also remove the degeneracy of the subvalent energy bands 88 and 89, where the LH secondary energy band is shown as an upward shift, while the jjh secondary energy band is 89. The energy level of the maximum value of 8 lang which is not offset downward is shown to be higher than the maximum value of 91 of the valence band in the i7A figure. The energy level of the maximum value of 89p of the hh sub-band 89 is not lower than the maximum value of 91 of the valence band in the 17A chart. Under the two effects of having this effect offset from the conductive band energy valley 87 with its minimum value of 87 m downward, the band gap 93 may be narrower than the band gap 92 of the unstrained condition of Fig. 17A. Taking 矽 as an example, for a tension-strained shoal layer (for example, forming a 矽 layer on a SiuGex layer), if the 莫 molar ratio X is about 30%, the 矽 has a two-fold defect. The energy level of the degree may be offset by approximately e18eV, while the LH degeneracy may be offset upward by approximately 〇.12ev. This produces an energy band gap 93 of approximately 〇8. Furthermore, the figure shows that an LH and HH band separation phenomenon occurs between the maximum value of the LH sub-band 88 and the maximum value of 89H of the HH sub-band 89. This band separation phenomenon is the effect of removing the degeneracy of LH and HH, and has the effect of reducing the scattering between the LH and HH sub-bands. In addition, the shape change of the subvalent band can reduce the effective quality of the light hole. As a result, the average free path of a strained semiconductor inner ballistic light hole may be longer than the average free path of an unstrained semiconductor inner ballistic light hole. Figure 17B also shows that under the degeneracy of the LH secondary energy band 88 and the HH secondary energy band 89, the hole 90 may re-aggregate from the HH secondary energy band 89 to the LH secondary energy band 88. In fact, when the strain is strained by tensile stress, the total number of holes in the LH secondary energy band may increase by 20% to 90% (see Fischetti et al., Journal of Appl. Physics, vol. 94, ρρ· 1079). -1095, 2003). Furthermore, the scattering ratio of LH is known to be much lower than the scattering ratio of HH (see Hinckley et al" "Hole Transport theory In Pseudomorphic Sii. x Gex Alloys Grown on Si (001) Substrates," Phys Rev B, 41, Ρρ·2912-2926, 1990). The implanter of the present invention 0881-A31715TWF(5.0) 40 1287292 • further considers these effects (such as the LH injection in the related description of Fig. 13) by using the electric hole by TM The hole injection efficiency can be improved by redistributing the secondary energy band to the LH secondary energy band to inject the "piezoelectric" material into the hole. This can be achieved by applying the source region of the hole injection. When LH is heavily inhabited and has high ballisticity, when such a combined effect is applied to ballistic charge injection by such a method, it can provide a method of applying electric current to ballistic charge injection, and It is an embodiment of the piezoelectric ballistic charge and injection mechanism of the present invention. This method improves ballistic hole injection efficiency by injecting a piezoelectric ballistic hole (such as LH). The 帛17C diagram shows the ugly dispersion relationship with the 17B, but the conductor is strained by a compressive stress. Similar to the tensile stress described above, this compressive stress relieves the degeneracy of the subvalent energy bands 88 and 89, but in a manner opposite to that of Figure 17B. The figure shows that the LH secondary energy band is shifted downward and the HH secondary energy band 89 is shifted upward. The deliberate relaxation of HH and LH still reduces the inter-band scattering between the LH and hh charges. Due to the deviation of the subvalent energy band, most of the holes in the figure show that they are concentrated in the HH sub-band. In addition, the figure also shows that the curvature shape of the subvalent band changes when compared to the unstrained example of Fig. 17A. The deformed HH sub-band in Figure 17C reduces the effective mass of the heavy hole and makes it a lighter hole. As a result, φ is within a strained semiconductor, and the average free path of the holes (i.e., piezoelectric holes) is longer than the mean free path of the holes in the unstrained semiconductor. This effect provides an embodiment of another piezoelectric ballistic charge injection mechanism of the present invention. It is known that for a charge in the subvalent energy band where degeneracy is removed, the first order of its effective mass can be linearly offset with stress (see Hensel et al., "Cyclotron Resonance Experiments in Uniaxially Stressed Silicon: Valence Band Inverse

Mass Parameters and Deformation Potentials”, Phys. Rev. 129, pp·1141-1062, 1963, and see Hinckely et al” “Hole Transport Theory in Pseudomorphic Sil-xGex Alloys Grown on Si(001) Substates, M Phys. Rev: B? 41, ρρ·2912-2926, 1990). By applying this linear relationship and the relationship between the effective mass and the mean free path, 0881-A31715TWF(5.0) 41 8 1287292 ^, the average of the charge-viewing electrical charge The method of free and miscellaneous. This & represents another -> 1 turn charge injection _ the real complement, and is paste parallel to the age of electricity to the righteous force and other silk lang. 18th _ slightly slightly mixed average One of the effects of the _ stress system is used as an example to illustrate the effect applied to HH. > See Figure 18 'Vertical axis represents normalization _ smart wrists' mean free path, that is, free dreams are free The ratio of the path to the mean free path of the unstrained cut. It can be clearly seen from this figure that the normalized free path system changes linearly with increasing stress. Furthermore, the mean free path is parallel along the nightmare axis [111]. The direction parallel to the axis along the axis In comparison, the system is much longer. The 19th figure shows the correspondence between the efficiency improvement rate of the repeated electric ballistic hole injection and the Wei stress. This efficiency improvement rate is the ratio of the strain rate to the unstrained rate. , ^ Under moderate stress, for example, about 2GGMegaPaseal ("tender") or lower, the rate of efficiency increase will increase with super-Hneariy, and when the stress is in a higher range (for example, 4〇0MPa or higher), the efficiency improvement rate is almost linearly proportional to the stress. In addition, the rate of efficiency improvement is much more significant along the parallel direction of the Shixi crystal axis [(1)] and the parallel direction along the axis of the Shihejing [[1]. The figure shows that the efficiency is increased by about twenty times and fifty times in parallel directions along the axis of the dream crystal [0〇w[111]. Fig. 20 shows the sensitivity of the unstrained internal efficiency improvement degree to the mean free path (hereinafter referred to as "ιηφ*"). It should be noted that the difference may come from, for example, different concentrations of impurities in the semiconductor. This figure selects the stress in the direction parallel to the direction of the twin axis [001]. Referring to Fig. 20, it can be noted that when the stresses are kept the same, the shorter ratio is 4 nm. The efficiency improvement rate is remarkably increased as compared with the longer πτφ* (for example, 1 〇 nanometer). For example, when a stress of one MPa is applied to a piece having a particle size of 4 nm, the efficiency improvement rate can be as high as (8) times, whereas when the same stress is applied to a dream having 10 nm (five) electricity*, Efficiency may only increase by a factor of 10. The effects presented here are useful for shrinking memory cells under advanced technology, as it is expected that the concentration of germanium impurities in the dream will result in a shorter ιηφ*. This is because the high concentration of impurities in the dream can help the unit to be scaled down to a smaller size (for example, it can avoid the area where the charge is traversed when the memory unit size is 0881-A31715TWF(5.0) 42 1287292 is reduced. Excessive increase). Now Tian can understand that by using piezoelectric ballistic charge, it is possible to change the ballistic mechanism of ballistic carriers ((3), HH or 疋 electrons). It should also be clear to those of ordinary skill in the art that when applying the techniques taught by this method, different types of stresses (such as tensile or compressive stresses) can be selected and the stress axes can be changed, (iv) the distribution and The average freeness increases the injection efficiency in these cases. Although the above discussion is directed to a galvanic hole, it will be apparent to those of ordinary skill in the art that, under similar considerations, the effects and advantages of the piezoelectric hole can be applied to piezoelectric ballistic electron injection. In addition, although the above discussion focuses on semiconductors (such as Shi Xi), the effects and advantages for semiconductors can be used for other types of conductors (because of the branch, hand, %•plate alloy, etc.), in addition Although the above description of the charge injection system is concentrated on the memory-related applications, the effects and advantages of the reliance on the ribs are applicable to other types of semiconductor devices. It is a transistor and an amplifier, etc.). Section 21A shows the relationship between the injection efficiency and the thickness of the active layer (BG 62) for ballistic transport to compare the difference between strain 无 and no strain 矽. As shown, by the piezoelectric ballistic electron injection mechanism, electron energy is injected into the CSR 66 at a much higher efficiency than the normal electrons injected into the unstrained enthalpy. This is due to the fact that ballistic electrons have a lower scattering ratio and a longer average free path, as previously described (see, for example, Figure 17B and its thief). This effect provides one of the features of the present invention by providing a means of solving the large resistance in the prior art. Figure 21B shows the relationship between the sheet resistance of BG62 and the mean free path when the injection efficiency is fixed at one percent. The sheet resistance can be reduced by using a piezoelectric ballistic electron injection mechanism. For example, the sheet resistance is about 250 Ohms/square in an unstrained crucible and drops to about 220 Ohms/square in a strained stone with a similar mean free path. Figure 21B also shows that by using this mechanism, by increasing the mean free path from 10 nm to about 28 nm, the sheet resistance can be lowered even without compromising the injection efficiency. The injection mechanism of the electric axe can be easily applied to the charge injection energy band of the present invention, 0881-A31715TWF(5.0) 43 8 1287292, etc., using the example of the energy band structure shown in FIG. Referring to Fig. 13, since the TG61 is strained, the holes are mostly composed of LH 75. The higher total number of LHs in TG61 is a promising phenomenon 'because it can provide a highly ballistic electric lion to supply current to the charge injection method'. For example, it can be implemented according to a piezoelectric ballistic charge injection mechanism. For example, this is achieved by applying tensile stress to TG 61. Under the influence of stress effect, HH 76, which can coexist with LH 75 in TG 61, accounts for about $% to 20% of the total number of holes compared to lh%. / The main endures 'Although the TG 61 series is strained according to the above-mentioned machine, but bg can be strained according to the conditions of another type of pressure, so that it can pass through the BG 62. The average free path is longer than the sound of the area in which it is located. The embodiment, for example, can be as described in the 17B and 17C, by applying a mechanical stress on the anode 62 to remove the degeneracy of the moon, thereby reducing the LH charge. In the band-to-band scattering event when passing through BG 62, the hole injection efficiency is improved. Memory Unit of the Invention Embodiment 100 Fig. 22 is a cross-sectional view showing an embodiment of a memory cell structure of the present invention. Referring to Fig. 22, the display unit 10 (M system includes a conductor_filter system% belonging to the types described in Figures 7, 9, 1 and 13 and a conductor-insulator system 6〇, which belongs to the A type of charge storage area (hereinafter referred to as CSR) 66 is a type of floating gate (hereinafter referred to as FG) 66100, and a channel dielectric (hereinafter referred to as CD). The conductor-filter system 59 includes a tunneling gate (hereinafter referred to as TG) 61, and a filter %, wherein the TG 61 is a conductor corresponding to the conductor-transitioner system 59. The filter 52 provides a description in FIG. The band filtering function, the charge filtering function in the related descriptions of Figs. 10, 12A, 13, and 14, the voltage dividing function in the related description of Fig. 12B, and the mass filtering function in the related description in Fig. 15. In a preferred embodiment, the filter 52 includes a tunneling dielectric (hereinafter referred to as TD) 53 and a blocking dielectric (hereinafter referred to as 0881-A31715TWF(5.0) 44 (^) 1287292 as described in FIG. BD) 54. Conductor-insulator system 6〇 includes a ballistic gate (hereinafter referred to as BG) 62 and a reserved dielectric (hereinafter referred to as RD) 64, which are respectively used as conductors and insulators of the system. The region of the early element structure from TG 61 to RD 64 is made by a conductor-filter. The filter 52 of the system 59 is constructed in contact with the conductor (BG 62) of the conductor-insulator system 60. The structure thus formed has a TD53 sandwiched between the two regions TG61 and BD54, and has two regions sandwiched between TD53 and BG62. Between the BD 54. The BG 62 is placed in the vicinity of the FG 6610 and is connected to the FG 6610〇.

Preservation of dielectric (RD64) phase insulation. FG661()() is disposed in the vicinity of the main body 70 and insulated from the main body 70 by CD 68. FG 661()() is typically coated and insulated by a dielectric such as RD 64, CD 68, or other adjacent dielectrics. These dielectrics must have suitable thickness and good insulating properties to allow charge. Can keep it without leaking. Typically, the thickness of RD 64 and CD 68 is in the range of from about 5 nanometers to about 20 nanometers. The TD 53 and the BD 54 may include a dielectric having a uniform chemical element or a gradually changing element among them. TD 53 and BD 54 may be included in a group consisting of oxide, nitride, nitrogen oxide, aluminum oxide (Al2〇3), hafnium oxide (HfO2), oxidized hammer (Zr02), and oxidized neon (Ta2〇5). Select it. Furthermore, a composite of any of the above materials or an alloy thereof, such as an alloy of niobium oxide-oxide (Hf〇rSi〇2), an alloy of niobium-oxygen (HfAiO), and a niobium-nitrogen oxide (HfSiON) Alloys, etc., can be used as materials for TD 53 and BD 54. In the preferred embodiment, an oxide dielectric having a thickness of 2 nm to 4 nm and a nitride dielectric having a thickness of 2 nm to 5 nm are selected as materials of TD53 and BD 54, respectively. . The unit 1 of Fig. 22 further provides a source %, a channel %, a drain 97 and a bit = substrate 70. The main body 7G includes a semiconductor having a first-conducting ratio of p-type and doped with a concentration of 1 X 10 atoms/cm 3 to i χ 1 〇 s atoms/cm 3 . Source = 5 and ^ 97 are mixed with the main body 7_ having a channel 96 between the two to persuade the two, and to;; = = conductive type (such as ', cubic A sub-doping concentration for doping These doped regions can be formed using 〇881-A31715TWF(5.0) 45 8 1287292 Thermal Diffusion or by Ion Implantion. In Figure 22, TG61 is shown as overlapping with BG62. Forming an overlap region between the two' wherein at least a portion of the FG 66(10) is below the overlap region. This overlap region has an indispensable position within the cell structure because the supply charge sub-system passes through the overlap region. Filtered for transmission through BG 62, RD 64, and finally into FG 661G(). FG 661()0 is used to collect and store these charge carriers and may be polycrystalline germanium, polycrystalline germanium, or any other species b A semiconductor material that stores electric charge. The conductivity type of j?G 661()() can be N type or P type. The materials of TG61 and BG 62 can be selected as semiconductors, such as n-type polycrystalline germanium, P-type polycrystalline germanium, heavily doped Polycrystalline...etc., or as a metal, for example (A1), platinum (pt), gold (Au), tungsten (W), molybdenum (Mo), ruthenium (RU), group (Ta), nickel (Ni), nitride button (TaN), titanium nitride ( TiN), etc., or an alloy composed of the above materials, such as tungsten telluride, nickel telluride, etc. Although TG61 and BG62 are respectively shown as separate layers in unit 100, the structures of BG62 and Ding 61 For example, the TG 61 may include a nickel-lithium layer formed on a polysilicon layer, so that the TG 61 is a composite layer. The thickness of the TG 61 may range from about SEM to 500 nm. The thickness of the BG 62 can range from about 20 nm to about 2 nm. The energy band structure along the line AA can be of the type shown in Figure 9, the type shown in Figure 1, or the figure shown in Figure 16. The 'programming operation of unit 100 can be achieved using the ballistic electron injection structure mechanism as described in the description of Figures 9 and 10, which is achieved by the piezoelectric ballistic electron injection mechanism in the related description of Figures 17B, 9 and 10. For this particular embodiment, the voltage of TG 61 is selected to be -3.3V to -4.5V with respect to 62 to form a capable injection between the two. The human has the pressure drop of the electrons of the compact sub-energy distribution, and the embodiment can be achieved, for example, by applying the electric Z TG 61 of _3·3ν and the voltage of 0V to the BG 62, so that between the TG 61 and the BG 62 Will yield = a voltage drop of 3.3V. Alternatively, it can be implemented by other voltage combinations, = TG 61 aA + 1.5VM BG 62 〇 3A v 3B ^ 3C, by reducing the image height of the energy barrier The pressure drop between TG 61 and BG 62 can be reduced. ^ 〇 881-A31715TWF (5.0) 46 1287292, by applying a voltage of about 1V to 3·3ν to the source 95, the drain 97 and the body 70, thereby consuming a voltage in the range of about IV to 3V. To CSR 66. For example, assuming that the thickness of the rib 64 is 8 nm, such an image force reduction effect can reduce the voltage drop of TG 61 and bg & 由 from * -3.3 V to about 2.8 V to 3.0 V. After unit 100 accepts programming to a programmed state, CSR66's FG66100 is negatively charged with a *subcarrier. This programming state of cell (10) can be erased by an implementation-erase operation. • The erase operation can be achieved using the ballistic hole injection mechanism in the related description in Figure 13, or by using the piezoelectric ballistic hole injection mechanism in the related descriptions of Figures 17B, 17C, and 13. In this particular embodiment, the electrical susceptibility of s 'TG61 is +5λ^+6ν relative to the power of BG62 to form a voltage drop between the two capable of injecting a light hole having a tight energy distribution. The implementation, for example, can be achieved by applying a voltage of +3V to the voltages of TG 61 and _2V to BG 62, thereby producing TG 61 and BG 62 with +5V. Alternatively, it may be implemented by other voltage combinations, such as applying a voltage of +2·5 ν to a voltage of 61 and _2·5 ν to BG62. As described in Fig. 6, the pressure drop between the redundancy and the 3^62 can be further reduced by lowering the image energy barrier height. The image force barrier is slightly reduced when the Yang 66(8) is negatively charged, and can be reduced more by combining the voltage of about -IV to -3V to the CSR66, and the implementation can be applied by applying A voltage of about one m-3V is achieved by the source 95, the drain 97, and the body 70. For example, assuming that the thickness of the rib is 8 nm, such image force is reduced. The enthalpy drop between TG 61 and BG 62 should be reduced from +5 V to about +4·5 ν to +4.7 V. Finally, in order to read the memory cell, a read voltage of about + lv is applied to the drain of the memory cell, and a voltage of about +2·5 ν (depending on the power supply voltage of the device) is applied. Memory on the BG62. The other areas (ie source 95 and main body 7〇) are in the right position. If the FG661G() is positively charged (ie, the CSR66 is electrically discharged, the channel region 96 is turned on. As a result, a current will flow from the source 95 to the drain 97. This will be the state "1". On the other hand, if FG 66ι〇 ( With negative charge, channel region 96 is either slightly turned on or completely turned off. Therefore, even if both BG62 and drain 97 are pulled up to read voltage, very little or no current can flow through channel 96. For example, 0881-A31715TWF ( 5.0) 47 1287292 - Here, the memory unit is programmed to state "〇" under sensing. 6 隐单单it 100' is to store the charge in a phase opposite to the surrounding electrode. This is illustrated by the condition of the conductive material or semiconductor material (ie, “floating closed”). In the foot-based survival scheme, the charge is evenly distributed throughout the CSR66. It will be understood by those skilled in the art that the present invention is not limited to the above-described embodiments and includes any other type of charge storage scheme. For example, the invention of the invention 70 can be used to include Multiple discrete storage seats (Di Screte

The inside of the storage is like Nano-Particles or Traps in a dielectric layer, as shown in Figures 23 and 24. Embodiment 200 Referring now to Figure 23, it is shown that the memory unit is shifted to a slight variation of the unit 1 of Figure 22. Unit 2 (10) except for one point is the same as the single (10) of Figure 22. The difference is that the unit does not use fg %. . The conductive area is used as the CS secret, and a plurality of nano particles having a nanometer size are used as the c art. After chasing some of the granules 6" turned up to the ellipse _, which is directly in the range of 2 filaments to 1 () filaments, φ is shown in the figure as being in contact with CD 68 and in RD 64. RD 64 is in the figure The middle layer is a single layer, but it can be a stacked layer composed of different dielectrics, such as oxide/nitride/oxide. ===The nanoparticle of the storage base can be the stone granules of each of them. Nanoparticles are ellipsoidal, with 2 filaments to 1G Na·L and well-known (4) technology to make H-storage particles can be used to hide semiconductor materials (such as semiconductor materials) Ge, siGe alloy, etc.), or Hf〇2), or metal (such as Au, Ag, Pt, etc.) y It is common for those skilled in the art to understand that nanoparticle - It does not need to be in the same shape as the surface of the substrate, but can be located on the surface of the substrate or below any of the "degrees", and can have its own micro-discharge. 881-A31715TWF(5.0) 48 8 1287292 The rice granules 662〇0 do not need to be in contact with the ribs 64, nor need to be completely within the ribs, but may be partially located within the RD 64. The portion is located within 768, or may be located entirely within (3) the dip. Embodiment 300 Figure 24 is a cross-sectional view showing an embodiment of another memory unit 3 of the present invention. The unit of Figure 22 is _. The difference is that the unit does not use the conductive area of FG 66100 as the CSR, but the capture dielectric with a plurality of topping centers (well 663〇〇) (4)_ &quot The electric CSR 66 system is a solution of the 663()() as a charge storage seat and can be a nitride layer formed by the low-pressure chemical vapor phase (L_Pre_e_ehen^mpOT_Depc«it_ LPeVDM) well known in the art. The charge storage scheme of both 200 #300 stores the charge in the local charge storage, wherein the partial storage of the unit 200 is a type of nanoparticle 662〇〇, and the partial storage of the unit is the type of the scale 663GG. The unit can operate in a similar manner to the inner unit 100. The advantages of the two unit structures are that the complexity of the process can be reduced, and when such units are placed in a memory array, two adjacent Unit room The interference can be neglected. In addition, if there is a local breakdown event in the insulation layer around one of the storage seats, the charge stored in the other storage holders can still be maintained. The size of the memory unit of the present invention is given The design rules of the Generati〇n are closely related. Therefore, the dimensions of the above defined units and regions are only illustrative examples. However, in general, the size of the memory unit must be available. When the voltage between TG and BG is at a higher absolute value (for example, 3v to 6V), it can be filtered by the filter and can pass through the filter, and the charge can be supplied to the voltage between TG and BG at a lower absolute value. (For example, 2·5ν or less) can be blocked by the filter and cannot pass through the filter. In addition, BG and RD must be sized to allow most of the filtered charge to typically pass through them for collection by the CSR at an injection efficiency of about 10-6 to 10-1. 0881-A31715TWF(5.0) 49 1287292 • “It should be noted that the county is not recorded here and the above implementation _ already, and han-, any all changes that fall within the _ plus patent application. For example, the unit 1. In terms of unit structure and operation, there is no need for a conductor-transitioner system and a conductor-insulator system, but a conductor capable of effectively filtering and transmitting charge carriers to the CSR - a filter-type conductor-insulator The unit structure of the system. ^ • The unit of the array can be aligned with the light path of the tree, and the peripheral circuit of the array can include traditional column address decoding circuits and lines well known in the art. a address decoding circuit, a sense amplifier circuit, an output buffer circuit, and an input buffer circuit. The memory cells of the embodiment of the present invention are typically arranged in a rectangular array having columns and rows within a plurality of § cells in the array. The structure constitutes a structure known in the art, 戋N Na structure. Figure 25 is a schematic diagram showing the structure of one of the two r arrays by taking a memory single illusion as an example. Referring to Figure 25, it is shown There is a word line (w〇rdHnes) 11〇, which contains word lines Μ small M, and M+1, and each-character line is set along a first direction (column direction). There is shown a tunneling line (_Beer 1_) 12〇, which includes a tunneling line 1^,]^, and Ι/Η, and a bit line (bit_Hnes) i3〇, which includes N4, N, N+]l, And Ν+2, all of the tunneling lines 120 and the bit lines 130 are arranged along a second direction (row direction). The BG 62 of each memory unit φ 1〇0 in the same column is by the word line no One is connected to each other. Thereby, the word line Μ+ι is connected to the BG 62 of the lowest column inner mother unit. Each thread 120 is connected to the TG 61 having the memory unit in the same line. Thus, the tunneling line L-1 is connected to the TG 61 of each of the memory cells in the leftmost row in Fig. 25. Similarly, each bit line 13 is connected to the drain 97 of all the memory cells in the same row. Thereby, the bit line N is connected to the drain 97 of each memory cell in the leftmost row in Fig. 25. Since the array presented in this example uses a virtual ground array structure, the leftmost row Memory unit used The bit line N is also used as the source-line N of a neighboring cell in the adjacent row (ie, the center row of Figure 25). Those skilled in the art will appreciate that the source is There is no polarity interchangeable name 'and the source and wave lines or the source line and the bit line are interchangeable names. In addition, the word line 110 is connected to the memory unit 62. 0881-A31715TWF(5.0) 50 8 1287292 Therefore, the BG, and BG lines can also be interchanged with the word line. The NOR array shown in Fig. 25 is a well-known array structure, which is used as an example to illustrate the _type formed by the memory unit of the present invention. . It should be appreciated that although _ only shows the array-small section, the example of Figure 25 illustrates the column size of the bribes formed in the above areas. In addition, the memory unit of the present invention can be applied to the rest of the (10) Han array organization. For example, although each bit and line 130 are arranged to share memory elements with the source lines of adjacent lines, it is possible to arrange the memory lines on each line of the memory array to have their own dedicated source lines. . Furthermore, although the present day and the month are illustrated by a single-memory unit and an N-R array, it will be apparent to those of ordinary skill in the art that a plurality of the present inventions can be arranged in a single column. The rectangular array of rows 'where the plurality of single thins constitute a well-known fine structure in the art, or a mixed structure composed of a combination of -NAND_ and _. The instantaneous erasing operation constructed in accordance with this can be implemented by performing a bit erase on a small group unit (e.g., a single unit for a library-digit character having 8 units). In addition, the erasing operation can also be implemented in a large group (for example, a unit for storing software code, which can include a unit configured as a page, or a plurality of pages constituting an array structure). φ Manufacturing method The present invention further provides a self-freshing technology for forming a memory cell and a memory array (9) Read and transfer technology (TeChmqUeS) and a manufacturing method thereof, and the unit surface (single it 100) and the 25th object shown in FIG. The object is the outline. Miscellaneous here is a single it 1GG to explain, but this way, said bribery examples only, so you can (4) repair andamp; and fine to the rest of the unit. Referring to Fig. 26A, there is shown a plan view of a % of a semiconductor substrate used as a starting material for forming a cell and an array. A cross-sectional view of this material is shown in Fig. 26B' in which the semiconductor substrate 98 is preferably a top-first conductivity type (e.g., p-type). A body 70 is formed within the semiconductor substrate by means of well-known techniques, such as ion implantation, and is assumed to have a first conductivity type. The body 7 can selectively utilize a semiconductor region (e.g., an N-type) having a second conductivity of 0881-A31715TWF(5.0) 51 1287292 to be insulated from the semiconductor substrate 98. The structure shown in Fig. 26B is further processed in accordance with the following steps. A first insulator is formed on the semiconductor substrate 98, and the thickness of the first insulator 68 is preferably between about 5 nm and 50 nm. The insulator 68 can be, for example, an oxide, and the oxide can be deposited using conventional Thermal Oxidation technology, HTO, TEOS, or by in-situ steam generation (ISSG) deposition ( Depostion) process is deposited. Next, a layer of charge storage material 66a, such as polysilicon, is deposited on the structure, for example, by conventional LPCVD techniques, which are doped in a field plus or by subsequent ion implantation. The procedure is achieved in a polycrystalline germanium film. The polycrystalline germanium layer 6 thus formed is used to form the memory cell type (cell 丨(8)) 2 CSR slave shown in FIG. 22, and can utilize the impurity of the second conductivity type to 1 X 1 〇 18 atoms/cm 3 to 5掺杂 1〇2i atomic number/cubic centimeter doping concentration is used for doping. The thickness of the polysilicon layer 66a, for example, may range from 5 nanometers to about 500 nanometers. Preferably, the undulating topography of the polysilicon thus formed is substantially planar. It should be noted that this example selects the polysilicon as the material of the charge storage layer 66a to illustrate the cell 1 〇〇. In general, other suitable materials having charge storage capabilities (e.g., nanoparticle, trapping dielectric) can be used in other units of the invention. Next, a photo-resistance_to_resistant material (hereinafter referred to as light a_t〇_resist) is locally applied to the surface of the structure, and then a conventional masking step is performed. For selectively removing the photoresist, leaving a plurality of photoresist lines extending in the second direction (row direction) on the charge storage layer 66a. The next procedure is to expose the exposed charge storage layer 66a to the insulator 68, which is used as a stop layer. Portions of the charge storage layer that are still under the remaining photoresist layer are not affected by this etching process. This step forms a plurality of polysilicon lines 66b extending in the second direction (i.e., "row direction"), and each of the two polysilicon lines 66b is separated by a first groove. The line width of the polycrystalline dream straight line 66b and the distance between two adjacent polycrystalline dream lions can be equal to: the minimum photolithographic size of the process used. An ion implantation step is then performed to dope the exposed 0881-A31715TWF(5.0) 52 1287292 矽 region for the second conductivity type to form a diffusion region that is self-aligned with the first trench 142. Such a diffusion region forms a bit line 130. The remaining photoresist is then removed using conventional methods. The next procedure is to form a second insulating layer 64a on the exposed charge storage layer 66a. The thickness of the second insulating layer 64a is preferably between about 5 nm and about 50 nm. The insulator may be an oxide deposited by a traversing, enthalpy, TEOS or ISSG deposition technique. The insulator may be in the form of a single layer or a combination of other insulators (for example, a composite layer of oxide and FSG). This second insulator 64a is primarily used to form the RD 64 of the memory cell of the present invention. Next, a conductive layer 62a of polycrystalline germanium is deposited on the structure, for example, 'can be conventional LPCVD technology, the doping method is on-site plus gold) doping or by subsequent -ion cloth The planting procedure was achieved in the polycrystalline dream film. This conductive material is used to form the BG62 of the memory cell and the word line 11 of the memory array. Typically, the conductive layer must be thick enough to fill the first trench 142 and have a thickness in the range of about 2 nanometers to nanometers. Preferably, the undulating topography of the conductive material 62a formed thereby is substantially planar, and a planarization process (i.e., CMP) of a choice of 1 ± can be used to achieve the planar relief topography. Resulting word line

, '〇 structure generally has a thinner area (used as 62 per __ memory unit) in CSR

The upper and the eighth have thicker regions on the bit line diffusion region 13〇 to connect the different units of BG =. It should be noted that polycrystalline dreams are chosen as conductive layers for illustrative purposes (to simplify the process). In general, the remaining conductive materials having low sheet resistance, good trench gap filling capability, and stable properties at high temperatures (e.g., _C) can be utilized. For example, a metallized polycrystalline germanium layer is, for example, a polycrystalline spine having a crane-polycrystalline compound (4) _-P〇lyside, and a well-known CVD technique is used as a conductive material. The crane-polycrystalline suppressor sheet 5 i 1()_square' _ is lower than the severity of the decay _ the resistance of the slab, which is typically about 1 〇〇 to ohm. Materials readily available during the process, such as TiN cattle conductors made of W 4 , can also be considered as conductive material 61a 0881-A31715TWF (5.0) 53 1287292 - material. The next procedure is to form a dielectric M3 on the conductive layer 62a, preferably having a thickness of about 1 nanometer to 4 nanometers. The dielectric (4) can be a nitride deposited using LpcvDv technology well known in the art. - Next, a photoresist is suitably applied to the surface of the structure, and then a photomask step is performed using conventional photolithography techniques to selectively remove the photoresist and leave on the dielectric 143 The lower complex number, the photoresist linear path 14 延伸 extending in the _ direction (column direction). The next procedure is to expose the dielectric 143 to the side and then leave the conductive layer 62a exposed until the insulator is observed to be slicked, and the insulator 6 is used as a -side blocking layer. The charge storage layer (4) and the portion still under the remaining photoresist layer 140 are not affected by this side process. This step forms a plurality of word lines 11 延伸 extending in the first direction (i.e., "column direction"), and each of the two word lines is separated by a second groove 144. The word line 110 read width and the distance between two adjacent characters and lines 11 可以 can be J to be equal to the minimum photolithographic size of the process used. The top view of the resulting structure is shown in Fig. 27, and the cross-sectional views along the lines ^, BB, cc, and dd in this structure are shown in Figs. 27A, 27B, 27C, and 27D, respectively. The next procedure is to engrave the exposed second layer 64a and then etch the exposed charge storage layer 1 until the first insulator 68 is observed, and the first insulator 68 serves as an etch stop layer. The portion of the charge storage layer 66a that is located under the remaining photoresist is not a side program; scale. This step forms a plurality of CSR66s. The light resistance of the fiber is then removed using a transmission method. The top view of the resulting structure is shown in Figure 28, in which the word lines 110 are staggered with the lines of the second grooves 144. The cross-sectional views of the lines from the lines, ribs, cc, and in the structure are shown in Figs. 28A, 28B, 28C, and 28D, respectively. The next procedure is to selectively form an insulating layer such as an oxide (not shown) on the sidewall of the word line no and on the sidewall of the CSR 66 exposed to the second trench 144. The formation of this oxide, which can be formed by the rapid purification of the paste, can be carried out by a thermal oxidation process, and has a thickness of about 2 nm to 8 nm. Next, 0881-A31715TWF(5.0) 8 1287292. A relatively thick dielectric layer (such as an oxide) is formed by a technique well known in the art of conventional LPCVD to fill the second trench. Slot 144. This oxide dielectric is then selectively removed to form an oxide block 146 over the area within trench 144. Preferably, the structure is such that the upper surface of the oxide block 146 is substantially coplanar with the upper surface of the nitride dielectric 143. This can be achieved by using, for example, a chemical-mechanical Polishing (CMP) - to planarize the thick oxide, and then using the nitrogen dielectric 143 as a polishing stop layer and/or etching to prevent the layer. To achieve this by implementing a reactive ion etch (RIE). If it is desired to clean the oxide remaining on the nitride dielectric 143, then a selective oxide step is performed. Thus, this process leaves only the oxide within the second trench 144 to form an oxide block 146 that is self-aligned to the trench 144. The top view of the resulting structure is shown in Fig. 29' where the word line no is staggered with the line of the oxide block 146. The cross-sectional views along the straight lines AA', BB', CC', and DD in this structure are shown in Figs. 29A, 29B, 29C, and 29D, respectively. The next procedure is to perform a step of removing the nitride dielectric 143 (for example, using a Phosphoric add). Next, a filter 52 of a multi-layer structure is formed on the word line no. In a particular embodiment, a third insulator 54a and a fourth insulator 53a φ are considered as a multilayer dielectric of the filter 52. The third insulating layer 54a, such as nitride, is at a temperature of 1050. In the ammonia-containing (NH3) environment of C, it is formed on the word line 11〇 by the rapid thermal nitridation (RaPldJrhennal-Nitridation) program. The third insulator 54a is preferably about 2 nm to 5 nm. The next procedure is to form a fourth insulator 53a, such as an oxide, on the third insulating layer 54a. The fourth insulating layer 53a can be formed by a technique well known using, for example, thermal oxidation, HTO, TEOS, or ISSG. The thickness of the fourth insulating layer 53a is preferably about 2 nm to 4 nm. The third insulating layer 54a and the fourth insulating layer 53a are used as the BD 54 and TD 53 of the memory unit of the present invention, respectively. The top view of the resulting structure is shown in Fig. 30, and the cross-sectional views along the straight lines aa5, BB, CC, and DIT in this structure are shown in Figs. 30A, 30B, 30C, and 30D, respectively. 0881-A31715TWF(5.0) 55 1287292 The field (4) is doped or achieved by subsequent implantation of cesium ions in a private order. This = column wear _ 12 〇 or remember the TG61 ^ fresh, the scale electrical material ^ ^ ^ degree has been flat, rice to 500 nm. Preferably, the undulations of the conductive material 6ia formed thereon are flat, and the _ flat recording (such as CMP) can be undulated by the Wei. It should be noted that the polycrystalline stone system is selected as the material of the conductive material 6ia for the purpose of description... I' as described in the related description of Fig. 27, the other has low sheet resistance, good groove gap filling ability and high temperature. Paste as fresh in the semiconductor manufacturing process, light f-reading materials, such as Wei #, ♦ shop, Deming, Menghua titanium, TiN, TaN, etc., can also be considered as a conductive layer of fairy material, and more Such materials can also be formed on multiple rides to form a composite conductor as a conductive material. Next, a photoresist material (hereinafter referred to as photoresist) is suitably applied to the surface of the structure, and then a photomask process is performed using conventional photolithography techniques to selectively remove the photoresist and conduct electricity. The layer 61a_L leaves a plurality of photoresist linear trajectories extending in the second direction (row direction). The next procedure is to expose the conductive layer 61a to the side until the impurity 53a is observed, and the insulator 53a serves as an etch stop layer. The portion of the conductive layer 61a that is still under the remaining photoresist layer is not affected by this etching process. This step forms a plurality of tunneling lines 120 extending in the second direction (ie, "row direction"), and each of the two lines is separated by a third trench 147 to separate the line width of the alpha tunneling line 12 and two phases. The distance between adjacent tunneling lines 120 can be as small as the minimum photolithographic size of the process used. A top view of the resulting structure is shown in Figure 31, in which the tunneling lines 120 are staggered with the third trenches 147. The cross-sectional views along the lines ΑΑ, ΒΒ, CC, and DD in this structure are shown in Figures 31, 31, 31C, and 31D, respectively. Figure 31 also shows the different areas within the memory cell type (cell 1) of Figure 22. The bit line 13〇 and the word line 13〇2 correspond to the source 95 and the gate 97 of the cell 1〇〇. The figure also shows CD 68, CSR66, RD 64, BG 62, BD 54, TD 53 and TG 61. These areas 0881-A31715TWF(5.0) 56 1287292 - the fields correspond to the corresponding elements in the description in Fig. 22 respectively. The area is exactly the same. The structure of the memory cell and the array is followed by deposition of a mechanically stressed strain material 15 (such as tensile or compressive stress). This strain material 15 is used as a strain source (bump s〇urce) to provide a piezoelectric ballistic charge injection mechanism as described in Figures 17B and 17C, and can be as shown in Figure 31. The strained material 150 is deposited on the structure, or the strained material 150 may be deposited after the insulators 53a and 54b exposed to the third trench 147 are removed by conventional techniques such as the leg. . In the former case, the strained material 150 primarily provides stress to the TG 61. In the latter case, since the strained material is also in contact with the word line 11〇, stress is thus supplied to the TG 61 and BG 62 in each memory unit. The strained material 150 can be a dielectric that provides different stresses and can therefore be used to create a piezoelectric effect in the TG61 and/or BG62 to achieve piezoelectric ballistic charge injection. The stress may be a uniaxial stress whose stress axis is substantially parallel to the surface of the TG 61 and extends in the first direction (column direction). A preferred embodiment of strained material 65 includes a nitride. The stress level (Stress ievei) and physical properties of the nitride can be controlled by thickness and process conditions during its formation. For example, by varying the pressure of a chemical element (such as a silane) during the formation of a nitride, about 50 million Pascals (50 MPa) to about 1 billion Pascals (1 Giga Pascal; 1 GPa) can be achieved. The stress of size. Nitrides with tensile stress or φ compressive stress can be used for chemical vapor deposition (Chemical Vapor)

Deposiotion; CVD) techniques are used to form, for example, thermal-cVD (to form a stress-nitride) or plasma-CVD (to form a compressive stress nitride). In addition, the stress level of the nitride can be modified, and even if necessary, a well-known technique can be used for relaxation, for example, using a dose higher than a threshold concentration (for example, 1 X 1 〇 14 molecules/cm 2 ) The helium ions are implanted with nitride. A top view of the structure produced in the former case is shown in Figure 32 where strained material 150 is deposited over the entire array. Cross-sectional views along the lines aa, BB, CC', and DD' in this configuration are shown in Figs. 32A, 32B, 32C, and 32D, respectively. It will be apparent to those skilled in the art having the benefit of the present disclosure that in the present invention, the strain source that causes the piezoelectric effect on BG 62 and TG 61 does not need to originate from the strained material 15〇, nor does it need to originate from 0881- A31715TWF(5.0) 57 8 1287292 The position shown in the figure can come from any other means and from any other area in the memory unit. Further, the stress does not need to be a uniaxial type, but may be any other type (for example, a biaxial type). For example, when polycrystalline germanium is used as the material of bg 62, the strain source can be derived from BG 62. The reason is that polycrystalline germanium typically provides a tensile stress in the range of about 200 MPa to 500 MPa. Another material that can be used as a strain source is Tungsten-Silicide, which is widely used in the manufacture of semiconductor ICs. The tungsten telluride can provide a stress in the range of about 1·5 GPa to 2 GPa, and can be used alone as a material of BG 62, or can be formed on a polysilicon layer to form BG 62 together with the layer. Other materials, such as Amorphour Silicon, polycrystalline #Poly SlGe, nitrided (TaN), titanium nitride, etc., can also be considered as supporting piezoelectric ballistic charge injection. Material. In addition, the device for introducing strain does not need to be achieved by using a strained material, but can be implanted with heavy atoms (such as 矽, 锗, Shishen, etc.) into the region of the crystal to be strained by other means, such as ion implantation. . Since the implantation of heavy atoms (such as Shi Xi, er, arsenium, etc.) above the threshold dose interferes with the periodicity of the lattice, resulting in a displacement loop (Dislocatioii Loops), strain is generated in this region. . The strain in this region can further provide stress to its vicinity. The stress in the receiving implanted region can be retained by implanted atoms such as nitrogen in the region to avoid being relaxed during the subsequent manufacturing steps of the unit. This ion implantation method has the advantage of simplifying the process 'because it does not require deposition or surname strain materials. In addition, it causes strain in the implanted area, so the strain is limited to the area where the strain effect is most desirable. All of the methods listed above provide the piezoelectric effect of the piezoelectric ballistic charge injection of the present invention. In addition, although only one strain source is shown in the memory unit of the present invention, it will be apparent to those skilled in the art that more than two strain sources can exist simultaneously in the same unit to provide any type of stress. (Zhang stress or compressive stress) are all in different areas of the memory unit within the scope of the appended claims. While the present invention has been described in its preferred embodiments, the present invention is not intended to limit the invention, and the present invention may be modified and modified without departing from the spirit and scope of the invention. The scope of protection is subject to the definition of the scope of the patent application. 0881-A31715TWF(5.0) 58 1287292 [Simple description of the diagram] Figure 1 shows the energy band diagram of a conductor-insulator system. The figure shows that the insulator can carry the part of the conductive strip under the image force effect and the non-image effect; the second figure shows a band diagram showing that the normal temperature electrons follow the conductor-insulation system of Fig. 1 Energy barrier; — Figure 3 shows a band diagram showing the thermal energy tunneling through the potential barrier of the conductor-insulator system of Figure 1; • Figure 3B shows the barrier height of the potential barrier and the peak position of the barrier A functional relationship between the dielectric field and the dielectric field added to the insulator; Figure 3C shows the energy barrier height of the potential barrier as a function of the dielectric field of the dielectric with different dielectric constants; Figure 4 shows an energy A diagram showing the potential energy barrier of a wide-energy spectrum of the electrons transmitted through the conductor-insulator system of the i-th diagram; Figure 5 shows an energy band diagram illustrating the transmission of the hot electrons with a narrow energy spectrum. The potential energy barrier of the energy band of the conductor-insulator system of Fig. i; φ Fig. 6 shows a band diagram illustrating the thermal power hole with a narrow energy spectrum transmitted through the valence band of the conductor-insulator system of Fig. 1. Potential barrier; • Figure 7 shows the invention The energy band diagram of the conductor-filter system; Figure 8 shows the energy of the threshold energy relative to the Fermi level in the case of the applied voltage: the energy level; Figure 9 shows the energy of a charge injection system of the present invention. The embodiment with a diagram illustrates the ballistic electron injection mechanism, and the energy-bearing diagram of another charge injection system of the present invention is shown in FIG. 10, which illustrates the use of the ballistic electron injection mechanism. Filtration effect and image energy barrier reduction effect; Figure 11 shows one of the solar and monthly energy maps, indicating the energy barrier used in the ballistic electron injection mechanism 0881-A31715TWF(5.0) 59 8 1287292 Height Engineering; 12A The figure shows the effect of the high barrier engineering of the present invention on ballistic electron injection, wherein the voltage between the TG and the BG can be used to block the height of the energy barrier for transmitting electrons forward and the barrier height of the backward transmission hole. Different degrees of change; Figure 12B shows the effect of the voltage splitting function of the present invention; Figure 13 shows another embodiment of the energy band diagram of the present invention, illustrating the filtering effect and shadow used by the ballistic light tunnel injection mechanism The force barrier reduction effect; Figure 14 shows the effect of the energy barrier height engineering on the ballistic hole injection, wherein the voltage between the TG and the BG can be used to block the height of the energy barrier from the forward transmission hole and block the back. The height of the energy barrier of the transmitted electrons varies to varying degrees; Figure 15 shows the functional relationship between the normalized tunneling probability of LH and HH and the reciprocal of the TD cross-voltage; Figure 16 is the energy band structure of another charge injection system of the present invention. Example; Fig. 17A is a schematic diagram showing the relationship between the energy E and the impulse k in the unstrained stone; Fig. 17B is a schematic diagram showing the relationship between the energy E and the impulse of the seven times under the tensile stress; Diagram of the relationship between the energy distribution E and the impulse k under compressive stress T5J · circumference, the 18th series is the relationship between the normalized mean free path and stress in the slab strain; ', stress Figure 19 shows the relationship between the stress axis and the efficiency of the disk. The 20th figure shows the average free path scale of the unstrained stone, indicating the compression efficiency and stress. Between Figure; BG film taste resistance and average freedom Figure 21A shows the relationship between injection efficiency and BG thickness. Figure 21B shows the relationship between the piezoelectric electron injection efficiency (9) and the path; 0881-A31715TWF(5.0) 60 1287292 • f 22 — shows a cross-sectional view of an embodiment of the unit structure; f 23 — shows a cross-sectional view of another embodiment of the unit structure of the present invention; and FIG. 24 shows a cross-sectional view of another embodiment of the unit structure of the present invention; - f 25 ® shows the schematic diagram of the hair and fresh elements _ structure; • 帛 26A ® fresh hair _ the first step of the manufacturing memory sheet - the top view of the semiconductor-based r-plate; • the 26B-line shows the 26A is a cross-sectional view of the unit structure along the line CC; FIGS. 27 to 32 are sectional views of the array and the memory cell structure in accordance with the sequence of the subsequent steps of forming an array and the memory unit according to the present invention; FIGS. 27A to 32A are in accordance with The sequence of steps for forming an array and memory cell of the present invention, the display array and the memory cell structure are taken from the cross-sectional view along the lines 27 to 32; and the 27B to 32B are followed by a step of forming a p-column and memory unit according to the present invention. In the order of the steps, the display array and the memory cell structure are along the line 27 to 32, and the cross-sectional view; the 27C to 32C are in the order of the subsequent formation steps of an array and the memory cell according to the present invention, and the display array and the memory cell structure are along Straight line 第 of the 27th to 32th drawings (cross-sectional view; Figs. 27D to 32D are diagrams showing the arrangement of the array and the read cell structure along the subsequent steps of the p-column and the memory cell in accordance with the present invention. Straight line DD, section view. [Major component symbol description] 10~ conductor-insulator system conductor; 14~ conductor and insulator junction; Fermi level in 1650~50; Fermi level in 1662~62; I853~ with image force effect Conductive 12~insulator of 53; W~ Fermi level; Fermi level in 16G1~61; Conductive strip with image force effect; I854~ Conductive strip with image force effect 54; 1801~ The conductive band of the image force effect of 61; ΙΑ2~ with the image force effect of 6, the conductive tape; I%4~ the conductive band with the image force effect of 64; IK6~ the conductive tape with the image force of 66; 0881-A31715TWF(5.0) 1287292 1870~ Conductive tape with image force effect 70; 18'~ conductive tape without image force effect; 20~ energy barrier height with image force effect; 2053b~ energy barrier height; 22~ Image height and energy barrier without image force < 24~ insulator image potential potential barrier; 2453~ energy barrier height; 2453b~ potential barrier; 2454~ energy barrier height; 2454b~ potential barrier; 2464~ imaging force Potential barrier; 24'~electricity of insulator without image force effect Energy barrier;

26~ energy barrier offset; 28~ energy barrier peak; 30~ energy peak to conductor/insulator boundary; 31~ for Furnham tunneling at room temperature; 32~ hot electron; 33~ electron kinetic energy; ~ arrow pointing forward; 34'~ arrow in the backward direction; 36~ broad energy spectrum of electron energy distribution (before energy barrier); 36'~ broad energy spectrum of electron energy distribution (after energy barrier); 37~ single Thermal electrons of energy; 38~ narrow energy spectrum of electron energy distribution (before energy barrier); 38'~ narrow energy spectrum of electron energy distribution (after energy barrier); 40~ single energy thermoelectric hole; 40'~ single energy Thermal hole (after passing the energy barrier); 41~ hole energy barrier height (with image force effect); 4Ι53, 4153b ' 4154 ' 4154b ~ energy barrier height; 41' ~ hole energy barrier height (without image force effect); 41'53, 41'53b, 41'54, 41'54b~ energy barrier height; 42~ image force hole energy barrier; 4253~ second hole potential energy barrier; 4253b~ first potential energy barrier; 4254~ a hole potential energy barrier; 4254b ~ second potential energy barrier; 4264~64 image force barrier; 0881-A31715T WF(5.0) 62 1287292 42'~Electrical hole energy barrier with no image force 丨41⁄2~53 price band; 4401~61 price band; 4464~64 price band, · 447〇~70 price Band; 46~ hole kinetic energy; 48p~48 peak peak decoration; △48~48 narrow energy spectrum; 50~ conductor; 53~ tunneling dielectric; 54~ blocking dielectric; 55~ conductive tape offset 56'~ electrons after tunneling; top sound distribution of 57p~57; energy spectrum of 57~57; energy distribution of △57'~57'; 59~conductor-filter system; 61~through tunnel 64~retention dielectric; 66~ charge storage area; 662〇〇~nano particles; 44~ valence band; 4454~54 price band; 4462~62 price band; 4466~66 price Belt; 44'~ image power strip; 48~ electron energy distribution (before energy barrier); 48t~48 tail distribution; 48'~ electron energy distribution (after energy barrier); 52~ filter; 53a ~ fourth insulator; 54a ~ third insulator; 56 ~ for direct tunneling of normal temperature electrons; 57 ~ tunneling electron energy distribution (before energy barrier); 57t ~ 57 tail distribution; 57'~ tunneling electron energy distribution (after energy barrier); 58~ threshold energy; 60~ conductor-insulator system; 62~ ballistic gate; 64a~second insulator; 6610〇~floating gate; 66300~ card; Channel dielectric / manufacturing process first insulator; 70 ~ body; 71 ~ charge storage area of the electron; 72 ~ light (subtractive); 73 ~ heavy __ miscellaneous; 74 ~ backward direction arrow ' 75~ light hole (injected forward in the gate) 76~ heavy hole (injected in the tunnel); 0881-A31715TWF(5.0) 63 1287292

75'~light hole (injected into the blocking dielectric); 76'~ heavy hole (injected into the dielectric under the barrier); 77~ energy distribution of the hole; 78~ ballistic light Hole (injected in the ballistic gate); 79~scattered heavy hole (through tunnel in the ballistic gate); 80~ ballistic light hole energy distribution; 82~ charge storage area hole; 84 ~ Electron tunneling inside the ballistic gate; 85~ Conductive band can be electrons in the valley; 86~ Conductive band energy can be a mass of electrons; 87~ Conductive band energy can be a light electron; the minimum value of 86m~42; 87m Minimum value of ~44; 88~LH sub-band; 89~HH sub-band; 90~ hole; 91~ valence band degenerate; 92~ no strain condition band gap; 93~ strain strain condition With gap; 94~ compressive strain condition band gap; 95~ source; 96~ no pole; 97~ channel; 98~ substrate; 110~word line; 120~ tunneling; 130~bit line; ~ photoresist; 142~ first trench; 143~ barrier dielectric on nitride dielectric 9 146~ trench oxide filter ; 147 ~ third groove; 150 ~ strain material; ED ~ insulator 12 plus electric field; LH ~ light hole; πιφ * ~ no _ ^ average free path; TBD two and the thickness of the barrier ;; Ttd ~ the thickness of the tunnel dielectric;

Va also, across the filter 52 plus voltage; ~ blocking the dielectric across the voltage; Vid ~ tunneling dielectric across the voltage. 0881-A31715TWF(5.0) 64 (S)

Claims (1)

128 ^^22357 Application for Patent Scope Amendment 10, Patent Application Fan Park: 1. A conductor-material system, including · · 'body, which has electric charge carriers, and t there are - energy points m The device is provided with a material adjacent to the interface with the barrier, and the potential energy barrier is electrically modified to control the transmission of the charge carriers. ...there is no change to 2. The conductor-material system of claim 1 of the patent scope, the charge carrier is a normal temperature charge carrier, and the material system is a filter, wherein the filter system The conductor is in contact with the surface of the crucible and includes a dielectric # to provide a pair of pole loads: a function, wherein the filter system includes an electrical repair: month ^ early to control the passage in a certain direction The flow of the carrier. Polar charge 3· As claimed in the scope of the patent, the filter provides: A material system - a voltage splitting function for reducing the dielectric within the dielectric material. 4. The conductor of claim 2, wherein the filtering function is a belt pass filter function. ~, ',, 5. Conductor-material as described in item 2 of the patent scope, wherein the filtering function is a mass filtering function. ',,,, 6. The conductor of claim 2, wherein the filtering function is a charge filtering function. The conductor-material system of claim 6, wherein the potential energy barrier is a first set of energy barriers, and the filter further comprises a A second set of electrically configurable potential energy barriers for controlling the flow of opposite polarity charge carriers through the filter in substantially opposite directions to the direction. 8. The conductor-material system of claim 7, wherein the relationship between the energy barrier height of the second set of potential energy barriers and the voltage drop across the filter is compared to the first set of potential energy barriers. The relationship between the height of the barrier and the pressure drop across the filter is weak. 9. The conductor-material system of claim 2, wherein the filter comprises: a first dielectric disposed adjacent to the conductor; and a first dielectric; And disposed in the adjacent region of the first dielectric, wherein the energy gap of the first dielectric is narrower than the energy gap of the first dielectric. The conductor-material system of claim 9, wherein the product of the dielectric constant of the second dielectric and the thickness of the first dielectric is substantially greater than the first dielectric The product of the dielectric constant and the thickness of the second dielectric. The conductor-material system of claim 9, wherein the first dielectric system comprises an oxide, and the second dielectric shell comprises a nitride, an oxynitride, an A1203, A material selected from the group consisting of Ti02, Τ &amp; 2〇5' and an alloy composed of the above compounds. 0881 -A31715TWF1 ;(20060816) 66 1287292 Where: 12. The conductor-material system 4, such as the application of the patent scope, is a high-energy charge carrier of the charge carrier, the material is an insulator, and . The EW potential energy barrier system includes a barrier to control the power of the smashing force. ^4 electric carrier over the image of the force potential barrier For example, the conductor-material system described in the scope of Patent Application 帛12, wherein the high-energy electric sciences are doing their best to meet the technical principles*, the material science, and the eight-hearted “have an energy distribution, wherein the knives have an energy of about 3 〇meV to about 300 meV ^ 'As claimed in the patent _ 12 conductor-materials '''8 5 Hai insulation system consists of oxide, FSG, SiLK, CDO, nitride, nitrogen oxides, Al2 〇3, Hf〇2, Ti〇2, Zr〇2, Ta2〇5, X and materials selected from the group consisting of the above materials. Η· A charge injection system, including: a conductor-filter The system includes: a first conductor for supplying a normal temperature charge carrier; and a filter in contact with the first conductor and including a dielectric to provide a filter function for a pair of polar charge carriers, wherein The filter includes an electrically configurable potential energy barrier for controlling the flow of a polar charge carrier passing through the filter in a certain direction; and a conductor-insulator system comprising: a second conductor In contact with the filter and having too much And the same as the electric carrier; and 0881 -A31715TWF1; (20060816) 67 1287292 has a f, a body, which is in contact with the second conductor at an interface, and the image is electric: = 】 = barrier The vicinity of the junction 'where the shadow can be transmitted over it. Change 'to allow the high-energy charge carriers to be filled with the charge injection system described in item 15 of the scope of the application, including a 'two' stand The first set of energy barriers, and the filter device further includes a potential energy barrier that can be modified by the palm of the hand to control the opposite polarity of the filter in the opposite direction to the charge carrier. The charge injection system of item 15, wherein #1:1 carrier has an energy distribution, wherein the energy = has an energy spectrum between about 3 〇 meV and about μ. 1 In the charge injection system described in Item 15 of the patent range, ~ Zhongyu 4-way energy charge sub-system light hole. In the second embodiment of the invention, the charge injection system described in claim 15 of the patent, wherein the special two-charge charge electrons. Applying for the charge injection system described in item 15 of the patent scope, the Bazhong 5 Hai is dedicated to the chargeable sub-system piezoelectric hole. And in the second one, such as the charge injection system described in Item 15 of the application scope of the patent, the electric charge system of the Bazhong 5hai special charge. 22. A memory unit, comprising: a conductor-filter system having: two conductors 2 for supplying a normal temperature charge carrier; and n n-conductor contacts, and comprising a dielectric material for lifting 881-A31715TWFl; 2〇〇6〇816) 68 1287292 For a function of over-the-counter charge of a certain polarity, which causes the over-cry crying system to include: a first set of electrically-correctable potential energy barriers for controlling along a certain a direction of flow through a polarity of charge carriers of the filter, and a second set of electrically configurable potential barriers for controlling the opposite of passing the filter in a direction generally opposite the direction The flow of polar charge carriers. The memory unit of claim 22, wherein the 4-lead system-the first conductor, and the memory unit further comprises a conductor-insulator system, wherein the conductor_insulator system comprises: a two conductor that is in contact with the filter and has a south energy charge carrier from the pass; the '^ insulator is in contact with the second conductor at an interface and has an image force potential barrier to the junction In the vicinity, the image force potential barrier is radio-modified to allow high-energy charge carriers to pass over it. / 24 · If the memory unit described in claim 23 of the patent scope, and whether it is a South China electric charge carrier and a right ^ ^ at the end of the eight y ^ 7 battle with a distribution, wherein the energy distribution has a 3〇meV to about 3〇〇meV energy spectrum. ???, j5. The memory unit of claim 23, wherein the 咼 energy charge sub-system light hole. VIII 26. Such high-energy charge carrier electrons as described in item 23 of the patent scope. As early as the eighth, 27. The memory unit described in claim 23, 0881-A31715TWFl; (2〇〇6〇8l6) 69 1287292 These high-energy charge sub-system piezoelectric holes. It is recognized as the memory described in item 23 of the patent application. The rain can charge the piezoelectric electrons. #凡,中中^ A method of a memory cell, comprising the following steps: forming a body of the ~, 苐-conductivity type. Forming a 笫一绍 绿 昆 绿 卞 卞 卞 ; ; ; ; ; ; ; ; ' ' ' ' ' ' ' ' ' ' ' ' Forming a charge storage region on the first insulating layer; forming a second region having a second conductivity type and a second region in the body; and, a V-electric type, forming a channel region Between the first region and a region, and the channel region is insulated; the storage region is adjacent and the phase: = two = layer adjacent to the charge storage region; y, a cold-electric region, wherein the first conductive region has a y-partial region adjacent to and insulated from the charge storage region; - adjacent to the (fourth) function of the first conductive region of the first conductive region: To the conductive region 'where the second conductive region is adjacent to a portion of the D° s and utilizing the filter phase. 30. The method according to claim 29, wherein the first conductive region and the second conductive region overlap with a plurality of regions; and 0881-A31715TWF1; (20060816) 70 1287292 region region At least a part of the region is disposed on the side of the overlap. In the memory unit formed in claim 29 of the claim, the filter system includes a plurality of dielectric materials and further provides a two-voltage division function for reducing The pressure drop within the dielectric. The formation-memory unit described in item 29 of the patent scope is a filter pass function. In the formation-memory unit described in Item 29 of the patent scope, the filtering function is a quality-based function. The formation/memory unit described in item 29 of the patent scope of Shen 2, and the filtering function is a charge filtering function. The method is specifically for forming a memory unit as described in item 29, and the medium is over 5; the device includes: - ϊ2 1:1' which is disposed adjacent to the conductor; and the domain, wherein the first广八二 is disposed in the vicinity of the first dielectric-intermediate band: ;; r energy energy band gap is compared with the first-dielectric material 36. As in the patent application method, 1 Μ~ The formation of a shell material - memory unit rice particles. The electrical storage region includes a plurality of phase separations. 37. The method of claim 29, wherein the charge storage region comprises: a capture dielectric of the early collapse center. /, there are multiple captures 38. - A method of forming a memory cell array, including the following steps 0881-A31715TWF1; (20060816) 71 1287292 Forming a body of the first isoelectric type in / --- a stack of enthalpy; one thousand conductor substrates; forming a first insulating layer on the semiconductor substrate; , forming separate from each other and in a first direction The extended bit line female distal line is formed in at least a portion of the body, and a plurality of charge storage regions are formed in the 'n' of the 'fourth column' extending along the first direction # and &gt;. Forming a column extending perpendicularly to the second direction of the first direction; forming a plurality of first regions having a second conductivity type in the body. Forming a plurality of second regions having a second conductivity type on the body Forming a plurality of channel regions within the body, each of which is between the plurality of first regions - and between the plurality of ί : domain fields - and each of the channel regions is substantially associated with the hard One of the plurality of (four) storage regions is adjacent and insulated; forming a plurality of first conductive regions, wherein each of the first conductive regions or having at least a partial region adjacent to one of the plurality of charge storage regions Insulating; and a plurality of transitioners having a function of crossing the enthalpy, each of the diapers having at least a portion of the region adjacent to one of the plurality of first conductive regions; forming a plurality of second conductive regions Each of the second conductive regions or regions is adjacent to at least a portion of the plurality of first conductive regions and is insulated by a phase of the plurality of (four). 39. Forming a memory unit 0881-A31715TWF1 as described in claim 38 of the patent scope of claim 4; (20060816) 72 1287292 A plurality of electrical materials are formed in parallel and separated by a wire, and each = word; The two-direction extension traverses the bit-conducting area electrical: the memory unit is connected to the memory unit, and the 40th' is formed as described in item 5% of the patent application: two =, ΐ includes forming a plurality of parallel and Phase separation and by conduction: electricity::: stretch and the second conductive area of the memory, such as the method of claim </ RTI> </ RTI> <RTIgt; Overlapping an overlapping region, wherein each of the electrical domains is disposed adjacent to the overlapping regions. The array unit forms a memory unit as described in item 38 of the specification, wherein the mother line is electrically connected to the second area of some of the memory unit apricots. The method of forming a memory cell array according to claim 38, wherein each m-line system and the second memory regions of the memory cells are electrically connected to the line, and The first regions of some of the records are electrically connected to adjacent rows. 0881 -A31715TWF1 ;(20060816) 73