TWI334196B - Method of manufacturing cmos devices by the implantation of n- and p- cluster ions and negative ions - Google Patents

Description

1334196 IX. INSTRUCTIONS: [References before and after the relevant application] This application claims the priority and advantages of U.S. Provisional Patent Application Nos. 60/392,271 and 60/391,847, both of which are filed on June 26, 2002. . This patent application also requests that the US Patent No. 10/244,617 and the 2nd and 2nd years of the same application in the same application.

The priority of U.S. Patent Application Serial No. 10/251,491, filed on the 2nd of the month. BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an ion implantation system and a semiconductor manufacturing method for implanting an ion beam formed by an N-type dopant group and a negatively charged cluster ion beam. . [Prior Art] The fabrication of a semiconductor device involves the introduction of impurities into a semiconductor substrate to form doped regions. The impurity elements (4) select the semiconductor material for proper bonding to produce an electrical load m to change the conductivity of the semiconductor material. Electricity "(4) produced) or holes (produced by p-type dopants). The concentration of impurity introduced is determined by the resulting region ratio 1 to produce a plurality of material regions to: from the structure and other such electronic structures, These systems are collectively used as two-component semiconductor devices. F" is a traditional method for introducing dopants into semiconductor substrates in cattle. In ion implantation, the inclusion of #& Introducing a departure

O:\119\119484.DOC 1334195 Energy is introduced into the sub-source to ionize the feed-feed material to produce a doped halogen (eg, 75As, 丨丨B, 丨15in, 3 ip . P, or Sb) Ions. An accelerating electric field is provided to extract and accelerate the typically positively charged ions so that ion beams can be generated. As is known in the art, mass analysis is then used to select the species to be implanted and to direct the ion beam to the "semi-(four) substrate. The accelerating electric field will give the ion narration #, and the ding will be carved into the moving cymbal, and the child can allow the ions to penetrate into the inside. Deviation from energy and quality (4) Determining the depth of (4) entering (4),

Because of its higher speed, ions with higher energy and/or lower mass allow deeper penetration! ^. This ion implant line was constructed to carefully control key variables in the implant process, such as ion beam energy, ion beam mass, ion beam current (charge per unit time), and the penetration of ions into the implant. In addition to the total number of dry areas per unit area, in order to maintain the yield of the semiconductor device, the beam angular divergence ((4) the angular variation of the impact base (4)) and the beam space uniformity and range must also be controlled. Currently known (see, for example, Dc Jac〇bs〇n, Konstantin Bourdelle, h_j·Tesman (H_j)

Gossmann), ..Sassowski, MA Albano, V. Bei Lun (v sec), j Μ Poate, love land Aditya Agarwai, Alex Perel, and T〇m fj〇rsky, "Ten Rotten, an alternative to ultra-low energy ion implantation Methodology, IEEE 13th International Conference on Ion Implantation Technology, Alpbach, Austria, 2000; N. Kishimoto et al., "High current anion implanter and its nanocrystals in insulators Manufacturing Applications", IEEE 12th Ion Implantation

O:\119\119484.DOC 1334196 Proceedings of the International Symposium on Technology, Kyoto, Japan, June 22-26, 1998, (1999) 342-345; N. Tsubouchi et al., “Quality separation, low energy, positive Beam characteristics of negative ion deposition equipment, IEEE 12th International Symposium on Ion Implantation Technology, Kyoto, Japan, June 22_26, 1998, (1999) 350-353"; and Junzo Ishikawa et al. Negative ion implantation technology "Nuclear instruments and methods in physics research B 96 (1995) 7-12) implanted negative ion system can provide advantages over implanted positive ions. A very important advantage of negative ion implantation is that it can reduce the surface charge caused by ion implantation in ultra-large integrated circuit devices in the fabrication of complementary MOS semiconductors. In general, positive ion high current (in the order of 1 milliamperes or more) implants a positive potential on the gate oxide layer and other components of the semiconductor device, which is easily higher than the closed oxide layer. Critical damage value. When a positive ion strikes the surface of a semiconductor device, it not only deposits a net positive charge, but also releases secondary electrons, which amplifies the charging effect. Therefore, ion implantation system equipment manufacturers have developed sophisticated charge control devices for so-called electronic overflow guns to enable introduction of low energy electrons into the positively charged ion beam and onto the surface of the device wafer during the implantation process. . This electronic overflow system introduces additional variables into the process and does not completely eliminate yield loss due to surface charging. As the semiconductor device becomes smaller and smaller, the transistor operating voltage and the gate oxide thickness become smaller. Decreasing the damage threshold in the fabrication of the semiconductor device further reduces the yield. Therefore, for many advanced processes, The negative ion implantation system has the potential to improve the yield of the traditional positive ion implant. Unfortunately, this technology has not yet been commercialized, and indeed even in the research and development phase, negative ion implantation has not been used by the author.

O:\119\119484.DOC 1334196 Integral circuit manufacturing. The negative ion source of the prior art relies on so-called negative affinity sputtering targets. An inert gas system such as Xe is fed into a plasma source that produces Xe+ ions. Once it is produced, Xe+ ions will be pulled? The target is sputtered to a negative bias, and the target is coated with helium vapor or other suitable alkaline material. Xe+ ions with energy will splash neutral target atoms, some neutral target atoms will pick up the electricity, and leave the target surface because of the negative electron affinity. Once with a negative charge, the target ions are repelled outside of the target and can be collected and focused by the ion source into a negative ion beam by electrostatic ion optics. Although it is possible to fabricate doped ions of a semiconductor such as boron by this method, the ion current tends to be low, the beam emission tends to be large, and the presence of helium vapor is almost unacceptable for wafer yield. Risk, because alkali metals are considered to be very serious contaminants for the process. Therefore, there is a need for a negative ion source technology that is more commercially viable. A particular interest in semiconductor fabrication is the formation of p_N junctions in semiconductor substrates. This requires the formation of adjacent N-type and P-type doped regions. A general example of junction formation is that an N-type dopant is implanted in a semiconductor region that already contains a uniformly distributed p-type dopant. In this example, an important parameter is the junction depth, which is defined as from the semiconductor surface. The depth, at this depth, N-type and? The type dopants have the same concentration. This junction depth is primarily dependent on the implant dopant mass, energy, and dose. An important aspect of modern semiconductor technology is its continuous evolution to smaller and faster devices. This process is called scaling down. Scale reduction is driven by the continuous development of lithography process improvements, which allow for the inclusion of integrated circuits

O:\119\119484.DOC The size of the semiconductor substrate is getting smaller and smaller. The acceptable shrinkage theory has been developed to guide the wafer manufacturer's material conductor device design while re-positioning the appropriate dimensions in all aspects. That is, at each technology or scale down node. The maximum impact on the scale of the ion implantation process is that the junction is reduced in wood and the shallow junction is gradually increased as the size of the device is reduced. As the integrated circuit technology shrinks, the requirements for increasing shallow junctions turn to the following requirements: as each scale reduction step, the ion implantation energy must be reduced. Recently, the ion energy used for many key implants has been reduced to the point where conventional ion implantation systems that were originally developed to produce higher energy beams are ineffective at providing the necessary implants. These extremely shallow junctions are referred to as "ultra-shallow junctions" or USJ. The limitation of low beam energy conventional ion implantation systems is most evident when the ions are extracted by the ion source and subsequently passed through the beam of the implanter. The ion extraction enthalpy is dominated by the relationship of ChiId-Langmuir, which states that the current density of the extracted beam is proportional to the extraction voltage raised to 3/2 power (that is, the energy of the beam during extraction). . Figure 1 a is the maximum extracted Kun beam current versus extraction voltage diagram for simplicity, assuming that only 75As+ ions are present in the extraction beam. Figure la shows that the extraction current drops rapidly as the energy decreases. In the conventional ion implanter, the 'extraction limit' operation method is found at an energy of less than about 1 Torr to volts. Similar restrictions occur when delivering low energy beams. The low energy ion beam moves at a lower velocity, so that for a given beam current value, the ions will be closer together, that is, the ion density will increase. This can be seen by the J=nev relationship, where J is the ion beam current density expressed in mA/cm2, 'η is the ion density expressed as cnT3, and e is the electron charge (=6.〇2xi〇-〗 9 Coulomb), and ν is the average ion velocity expressed in cm/s. Because of the static electricity between the ions O:\119\119484JDOC 11 plus the current strong ^ (four) flat nuclear anti-*, the two sides replied at low energy and the explosion is more obvious, so the ion beam will be dispersed. This phenomenon is called "beam bursting is positive: the low-energy electrons in the beam line of the implanter, which traps the ion beam during the transfer and contributes to the compensation of the space charge explosion. *,,, and the explosion will still Convergence, etc., are most obvious in the presence of electrostatic focusing mirrors. The falling V and the second are to loosen the bundle from the beam, and the high mobility compensation electrons are stripped off by the special beam t-transfer such as Kun (75). A huge atom is difficult. Because of the M ^ sub-quantity, the ion velocity is lower than that of a light atom. There is also a serious extraction and transfer difficulty for the p-type doped boron. Boron transfer is required by a specific binary process. Extremely low implant energies (eg, less than i仟 electron volts cause difficulties) and most of the ions extracted and delivered by typical BF3 source plasma are not necessarily the result of "ion" and are such as 19{ 7+ and 49 are called + ion fission 'children's fission system is used to increase the charge density of the extracted ion beam and the flat π# large-scale integrated circuit manufacturing front view' when transmitting low energy ^ and ^ effective current' Difficulties will combine to make the formation of USJ non-cracking. The method for the benefit of the Childe-Lanjol equation is that the ion mass is, for example, as shown in Figure 1b, by ionizing the molecule containing the dopant of interest rather than the dopant atom. In this way, although the molecule The kinetic energy is higher during the transfer period, and as it enters the substrate, the molecule breaks into its constituent atoms. The knives are shared among the individual atoms I according to the distribution of their masses, so that the dopants are implanted. The energy is much lower than its original transfer kinetic energy. Consider changing the hetero atom "X" to the free radical "γ" (whether or not Y is the cause of the demonstration). If the ion X Y+ replaces χ+

O:\119\119484.DOC -12· 1334196 is implanted 'then χγ+ must be extracted and transmitted at a higher energy, which increases by a factor equal to {(quality of ΧΥ) / (mass of Χ)}; This ensures that the speed of X remains the same. Since the space-charge effect described by the above-mentioned Childe-Lanmuir equation is superlinear with respect to ion energy, the maximum deliverable ion current is increased. It is known in the art to historically use polyatomic molecules to reflect the problem of low energy implantation. One common example for low-energy boron implantation is the use of BF2+ molecular ions to replace Β+β. This process ionizes the bf3 feed gas into BF2+ ions for implantation. In this way, the mass of the ions is increased to 49 atomic mass units, which allows for an increase in the energy of extraction and transfer, almost five times (i.e., 49/11) coefficients using a single boron atom. However, as the implanted boron energy is reduced to the same coefficient of (49/11). We have noticed that this method does not reduce the current density in the beam because there is only one boron atom per unit charge in the beam. In addition, this method also implants fluorine atoms into the semiconductor substrate along with boron, however, it is known that fluorine may exhibit a negative influence on a semiconductor device. For ion implantation, there are also molecular ion operations using decaborane as a polyatomic molecule, such as the "decaborane, an alternative to ultra-low energy ion implantation" reported by Jacquesson et al., IEEE XIII. The International Symposium on Ion Implantation Technology, "Alpine Bach, Austria, 3G-303 (2000)" and Yamada, "Application of Gas Clusters in the Process of Materials", Materials Science and Engineering A217/ 218, pp. 82 88 (1996). In this case, the implanted particles are ions of decaborane molecules, which contain 10 boron atoms, so the "clustering of this technology does not increase the ion core yield" but at a specific ion current, The decaborane ion Βι〇Ηχ+ has 10 boron atoms per unit, which substantially increases the implantation dose rate.

O:\119\119484.DOC •13· This pair of USJ P-type metal oxide semiconductor (PM〇s) transistors is formed, which in turn is a promising technique for implanting boron, which is very low energy. It effectively reduces the current delivered in the ion beam (1 〇 factor in the case of borax ion), which not only reduces the beam space charge effect, but also the wafer charging effect. Because the wafer is charged by a positive ion beam, especially the gate oxide layer is known to reduce device yield through damage-sensitive gate isolation, which reduces the current through the use of cluster ion beams for USJ device manufacturing. Attractive, the device must gradually provide an ultra-low threshold voltage. It should be noted that these P-type molecular implanted second example towels are clustered by feeding the material <single ionization rather than by the aggregation of the feed material. It should also be / Wang Yi, so far, has not developed a set of techniques for generating N-type doping ions. The success of future complementary MOS (complementary OX) processes is likely to rely heavily on the commercialization of viable N and P multi-implantation technologies. Therefore, it is necessary to solve two obvious problems faced by today's semiconductor manufacturing industry: wafer charging and low energy ion implantation. Traditionally, ion implanters have been divided into three basic types: high current, medium current, and high energy implanters. The 1 (four) high current and medium current implant processes are useful. More specifically, today's high current implanters are primarily used to form low energy high shot areas of transistors such as silk structures, and doping of polycrystalline breaks. These are typically batch-type implanters, that is, they can handle many wafers that are attached to a rotating disk, however, the ion beam remains stationary. High current beam lines tend to be simple and can accept a large number of ion beams; at low energy and high current, the beam at the substrate tends to be large and has a large angle

O:\119\119484.DOC 1334196 • Divergence. The medium current implanter typically combines a series (one wafer at a time) process with a cavity that provides high rotational capability (eg, up to 6 〇.) from the substrate normal: a typical ion beam span The wafer is electromagnetically scanned in an orthogonal direction to ensure uniformity in the agent. In order to meet the commercial implant dose uniformity and reproducibility requirements for only a small percentage change in blood type, the ion beam must have good corners and uniformity (eg, beam on the wafer < 2 degree angular uniformity) ). Because of these requirements, the medium current beam is processed to give superior beam control at the expense of a limited number of degrees. That is to say, the ion transport efficiency through the implanter is limited by the emission of the ion beam. Currently, higher current (about mA amps) ion beams generated at low (<10 仟 electron volts) energy have problems in tandem implanters, which result in wafer yields for specific low energy implants. Into (10) such as 'advanced complementary MOS semiconductor process towel and reduced structure production' is acceptable for broadcast. For batch-type implanters (which handle many wafers attached to the turntable), similar transfer problems exist for low beam energies per ion < 5 仟 electron volts. It is possible to use the beam to transmit optics simply, and the ion beam (4) (space range, spatial uniformity, angular divergence, and angular uniformity) is therefore mostly caused by the ion source itself. (That is the beam characteristics during ion extraction, which are determined by the optical field that the implanter can focus on and control the morphological emission characteristics emitted by the ion source. Using bunching instead of single beam, By transferring two beams of energy and reducing the current carried by the beam, the ion beam emission is effectively enhanced. Therefore, there is a need for a group of polyion and group ion source technologies in semiconductor manufacturing, in addition to being able to provide higher In addition to the efficiency dose rate and higher yield, it can provide - better focus, more accurate and more

OA119\H9484.DOC 1334196 'Ion beam that can be more closely controlled. SUMMARY OF THE INVENTION One object of the present invention is to provide a method of fabricating a semiconductor device capable of forming (ie, accepting) an ultra-shallow impurity doped region of conductivity in a semiconductor substrate, and further capable of high yield The rate of completion completes the move. Another object of the invention is to provide a system and method for ion implantation, by which a negatively charged decaborane (Bl〇Hl4) ion is fabricated into a ruthenium and implanted in a semiconductor A pn junction is formed in the substrate. Another object of the present invention is to provide a method of fabricating a semiconductor device capable of forming an ultra-shallow impurity doped N*p type (i.e., a ruthenium acceptor or a donor) via the use of N and P type clusters of the form AsnHx+. A heterogeneous region in which n-type clusters n = 3 or 4 and 〇 sxsn + 2, while pairs of ρ-type clusters are Βι〇Ηχ%ΐΒι〇Ηχ·. Another object of the invention is to provide a method of implanting arsenic group ions of the form AS3Hx+ and As4Hx+ which is capable of forming an ultra-shallow implanted region of the conductivity type in a semiconductor substrate. • Another object of the invention is to provide a method of producing a phosphorus group polyion in the form of PnHx+, wherein n is equal to 2, 3, or 4 and X is in the range of 〇, which is ionized by feeding ΡΗ3 into the gas, and The phosphorous cluster is subsequently implanted into the semiconductor substrate to complete the doping. Another object of the invention is to provide a method for producing a boron group polyion in the form of ΒηΙίχ+, wherein η is equal to 2, 3, or 4 and X is in the range of 〇 <χ; $6, which is by Β# 6 feeding gas ionization, and subsequently implanting the boron group into the semiconductor substrate to complete the p-type. Another object of the invention is to provide an ion cloth for fabricating a semiconductor device

O:\119\119484.DOC 1334196 implant system designed to form an ultra-shallow impurity holding region of N or P conductivity type in a semiconductor substrate.

According to one aspect of the invention, there is provided a method of implanting a cluster ion comprising the steps of: providing a supply of a dopant atom or molecule to an ionization chamber, the dopant atoms or molecules being combined to form a plurality Clustering of doped atoms and ionizing the dopants into dopant cluster ions, extracting and accelerating the dopant ions by -+ field, mass analyzing the ion beam, and electrically doping The group of polyions are implanted in a semiconductor substrate. The purpose of the invention is to allow a semiconductor device manufacturer to simultaneously extract n pairs of dopant atoms (n=4 in the case of Α%Ηχ+) and not only: implant single-atom' to improve extraction The method of low-energy ion beam difficulty 1 The atoms of the cluster are implanted by the energy of E/n. The group ion ion implantation method can provide low energy, and the single-pump 姑 π 仏 仏 early atom implantation "Equivalent. Therefore, the implanter is operated in the south of the required implantation energy, and the extraction voltage is doubled, which can tolerate the pre-beam current, especially at the time of (iv) formation, the low implantation energy is considered in the ion extraction stage. The improvement obtained by cluster ion implantation can be estimated by estimating the limit of the Childer Langmuer by the following formula: *化成已 (OJmax^ 1.72 (Q/A)1/2V3/2d -2 » Ampere 2 indicates that Q4 ion charge state, Α is expressed in terms of salt unit, ν is the extraction voltage with 仟 砉 Ά Ά d is the gap width to eight metrics - ι 么 仟 表 表 , Divided into 4 equations 〇) with d = mouth 7 common extraction light when "graphics. In fact, the optical system can be made close to this limit by many ion implanters. Extension equation

O:\119\119484.DOC 17 1334196 (1) 'For cluster ion implantation relative to monoatomic implantation, the advantage of the following figure is to define A to quantify the increase in yield or implant dose rate: (2) Δ = n (On/U1)3/2 (mn/m^1/2 △ where Δ is a single atomic implant with n atoms of interest, in terms of relative mass When the energy of the energy UnUn is lower, 'where je V, the dose rate (atoms/second) achieved by implanting the cluster is relatively improved. Among them, unadjusted to give a situation like a single atom (η=ι) In the case of the same implant depth, equation (2) is reduced to: (3) △ = n2. Therefore, the implantation of η doping atom clusters has the potential to provide 2 times higher than the traditional single atom η The implant dose rate. In the As4Hx example ten, for the small sputum, the maximum dose rate is improved by about 16 times. The comparison between low energy As and "implantation is shown in Figure 2a to illustrate this point. The use of implanted people also proposes the transfer of low-energy ion beams. It should be noted that for cluster polymerization, each cluster needs only one charge, not like the traditional situation. Each doping atom carries a charge. The library of dispersion: force: is reduced as the charge density decreases. 'Transmission efficiency (beam transfer) is improved. In addition, clustering is more monomeric. It has a higher quality and is therefore less affected by the internal Coulomb force. Therefore, 'the aggregation of η doping atoms instead of the early one atom implantation can improve the basic transfer problem in low-energy ion implantation. And make the compelling and more productive process feasible. This method is sufficient to implement the formation of cluster ions. It is used in the traditional source of commercial ion implanters, and it produces only very Xiao Shao’s main low-order (eg η=2) clusters, therefore, these implanters cannot

O:\119\119484.DOC •18- 1334196 Effectively achieve low-energy clustering on the top of the column, α complex. It is true that the strong plasma provided by many conventional ion sources will ionize and cluster the knives into several components. The ion source described herein, because of its use of a "soft" ionization process, is capable of generating clustered ions, i.e., ionized by active main electron electron impact. The ion source of the present invention is specifically designed to achieve the purpose of generating and maintaining dopant cluster ions. [Embodiment]

The present invention provides a number of specific embodiments. These specific embodiments are directed to the fabrication of various types of germanium and germanium dopant cluster ions and negatively charged cluster ion beams. Both the erbium and erbium type dopant cluster ions and the negatively charged group ion beam can be generated using the ion source shown in Figures 2a-2f. 2a-2f are schematic views showing the group ion source 1 〇 and its various components. Referring first to Fig. 2a', a cylindrical feed supply bottle 11 such as AsIl3, ρ%, (4) or scented old is provided. The feed material can be stored as a gas at room temperature or can be induced by sublimation of heated solids or vaporization by liquid evaporation.

In. The feed supply 11 is connected to the ionization chamber flow controller 12 via the flow controller 12 and may be a mature computer controlled mass flow controller or simply a connecting tube having a predetermined gas conductivity. In the latter example, the flow rate is changed by the air pressure in the control 11. The controlled dopant-containing gas feed material flow produces a steady gas pressure within the ionization chamber 13, for example, between about 3 χΐ〇 4 Torr and 3 χ 10 _ 3 Torr. The ionization energy 14 provides the temperature of the ionization chamber 13 and all of the components from the source, in the form of a control current having a defined energy or speed, typically controlled to a preferred value. By fine-tuning the source pressure, temperature, current, and electron energy, an environment can be generated in the ionization chamber 3, O:\119\119484.DOC •19· to bind, for example, asH3 dopant atoms or molecules, To form a poly-ion ion containing (iv) a desired hetero atom, for example, a tetramolecular compound AS4Hx+, wherein X is an integer between 〇 and 4. The ionization chamber 13 has an aperture 17 of the material ion (4) in the beam path, 1 is extracted or accelerated by a strong electric field between the ionization chamber 13 and the extraction electrode 15 (4) is generated by a high voltage power source, the high voltage power source The ionization chamber 13 is biased to a voltage v relative to the ground potential, and the extraction electrode is close to

(d) Bit Acceleration % is established in the forward direction to extract positive ions out of the ionization chamber and in the opposite direction when negative ions are required. The accelerated ion system is formed into an ion beam 16 by the extraction electrode 15. The kinetic energy E of the ion beam 16 is expressed by the formula (4): (4) E = /qV/ ,

Where V is the source potential and "the charge per ion. When V is expressed in volts and the q X electron charge unit is not present, then E has an electron volt (eV) unit. According to the invention, a portion of the ion implantation system is formed. The ion source is a sub-impact ionization source. Figure 2b is a schematic cross-sectional view of the ion source according to the present invention, which shows the structure and function of the component constituting the ion source. The cross section is along the direction of the filament beam propagation. The plane is cut and splits the ion source into two halves. The source 10 includes an evaporator 28 and a beam forming region 12 that are joined to each other at the mounting flange 36. The ion source 10 is made via the mounting flange 36. An interface is created between the vacuum chamber of the ion implanter or other process tool. Therefore, the ion source portion of the ion source on the right side of the flange 36 in Fig. 2b is under high vacuum (pressure < 1X10-4 The gas material is introduced into the ionization chamber buckle, wherein the milk knife is ionized by electron impact from one or more electron beams 70a and 7b, and the electron beams are passed through a pair of opposite electric beams. Entering the hole

O:\II9\119484.DOC -20- 1334196 Controls 71a and 71b enter the ionization chamber 44. With this configuration, the ion system is generated in the ion extraction aperture plate 80 and adjacent to an ion extraction aperture 81. The ions are extracted by an extraction electrode (not shown) located before the extraction aperture plate 80 and formed into a high energy ion beam. Various evaporators 28 are suitable for use in the present invention. An exemplary evaporator 28 is shown in Figure 2b. The evaporator 28 is exemplary and can be formed by an evaporator body 3 and a crucible 31 for carrying a solid source feed material 29 such as decaborane BiqHi4 into which an impedance heater can be embedded. The water-cooled line 26 and the mud-type air-cooled line 27 can be configured to be in intimate contact with the evaporator body 3A and to provide a uniform operating temperature above room temperature to the crucible 31. The heat transfer energy between the crucible 31 and the temperature-controlled evaporator body 3〇 is supplied via a pressurized gas, introduced into the crucible-evaporator body interface 34 by the gas feed 41, however, the temperature of the evaporator body 31 is via a Thermoelectrics engage in monitoring. The vaporized decaborane or other gasification material 50 is collected in the stabilizing fluid 51 and passed through the evaporator outlets 39 via a pair of isolation valves 1 and 11 , and through the vapor conduit 32 contained in the source block 35. And entering the ionization chamber 44 via the vapor inlet aperture 33. The isolation valves 丨00 and 丨丨〇, the mounting flange 36, and the source block 35 can also be temperature controlled to a temperature near or above the evaporator temperature to prevent condensation. The ion source oxygen delivery system can include two conduits fed into the ionization chamber 44 by two separate sources. The first source can be a small diameter, low conduction path that is fed to the gaseous material by a source of pressurized gas, such as a gas drum (not shown). The second source of energy is caused by a high conduction path from a low temperature evaporator that vaporizes the solid material. Regardless of the source, the gas delivery system will ionize the chamber 44

The pressure in O:\119\119484.DOC 21 1334196 . is maintained, for example, in a number of Torrents. The evaporator 28 maintains its surface under tight temperature control to maintain contact with the solid material in order to maintain stability of the gas flow into the ionization chamber, and thus the steady gas pressure within the chamber. The isolation valve 11〇 can be closed prior to entering the evaporator 28 to maintain the ion source and ion implanter under vacuum. The isolation valve 1〇〇 can also be closed to keep the vapor 50 contained in the crucible 31. The evaporator 28 can then be safely transferred to the drying hood where the rams 31 can be refilled or cleaned. Before opening 阙i 〇〇, the vent valve 111 welded to the room 100 can be opened to bring the hummer to atmospheric pressure. Once the supply is complete, the chamber 100 can be closed again. By connecting the valve 100 to the valve 110, the evaporator 28 can be assembled to the ion source 10, and the vent valve 111 is subsequently connected to a thick line to The crucible 31 between the valve 100 and the valve 110 and the fixed connector are evacuated. If necessary, the isolation valve 110 can then be opened without compromising the vacuum environment of the ion source and ion implanter. The evaporator assembly 30a is formed by heating and cooling the body 3'' and the removable 坩埚φ 31. It is possible to move the end plate (not shown) on the back of the evaporator 28 to be close to the crucible 31. Once the crucible 28 is removed from the evaporator 28, it can be refilled by removing the cover 341) sealed by the elastomer to the end of the trap and #34 separating the grid 34a of the solid 29 . After refilling, the crucible 31 is inserted into the evaporator body 30 and a vacuum seal is formed at the outlet 39 in front of the evaporator body 3 to heat the enthalpy fluid 51 and the heat present in the 坩埚 evaporator body interface%. The Gas Isolation Exit 39 is used as an outlet for the vaporized gas. In order to achieve the average temperature of 坩埚31, the mechanical fit between 坩埚31 and the evaporator body 3〇 is closed. “The gas can be filled with 坩埚31 and the evaporator body. O:\119\119484.DOC -22· 1334196 Any gap between them to promote heat transfer between the two surfaces. The heat transfer gas 'is entered into the gap via the end plate assembly 28 & and can be at or near atmospheric pressure. The temperature control can be performed by a resistive element such as a proportional integral differential (PID) closed loop control, and the resistive element can be embedded in the evaporator body 3〇. Figure 2f shows a block diagram of a preferred embodiment in which three temperature zones are defined. Area 1 is used for evaporator body 30, zone 2 is used to isolate valves 1 and 11, and #区3 Used for source block h. Each zone can have its own controller, such as the 〇mr〇nE5CK digital controller. In the simplest case, the heating element is used solely to actively control the temperature above room temperature, for example between i8c and 200C. Therefore, the resistance box type heater can be embedded in the evaporator body 30 (heater 1) and the source block 35 (heater 3), and the valve 1 〇〇, ^ can be wound with a 矽 tape heater (heater 2) Where the resistive elements are lines or drop bands. The three thermocouples labeled TC1, TC2, and TC3 in Figure 2f can be lightly embedded in each of the three components 30, 35, 100 (11〇), and read continuously by each of the three dedicated thermostats. The 8 thermostats 1, 2, and 3 are user-programmable with individual temperature set points SP, SP2, and SP3. In one embodiment, the temperature set points are such that SP3 > SP2 > SP1. For example, in the case where the evaporator temperature needs to be 30C, SP2 may be 5〇c and 8]?3 is 7〇 (:. The controller is typically operated when the TC master return does not match the set point, and the controller compares Cooling or heating is initiated as needed. For example, where only heating is used to change the temperature, unless TC1 < SP1, the comparator output is 0. The controller may contain an output power lookup table that is SP1TC1 temperature difference a nonlinear function and feedback the appropriate signal to the heater power of the controller to warm

O:\119\119484.DOC •23- : Smooth adjustment to the tertiary set point value. A typical method of changing the power supply to the heater is by pulse width adjustment of the power supply; this technique can be used to adjust the power supply between 1% and 满 of full scale. The &> controller typically maintains the temperature set point within 0.2C. The evaporating crucible body material can be selected to maintain high temperature uniformity. A small heat leak can be deliberately added to the evaporator body 3 (illustrated to improve the stability of the control system and reduce the settling time by means of a gas line located on the outer surface of the evaporator body 30. The road 27 is buckled around the evaporator body and covered by a plate (not shown). Air can be directed into the manifold system piping and integrated into the evaporator end plate 38 to provide gentle, continuous convective cooling. After the flow control metering valve, the air is fed through the inlet. The air is discharged into the factory discharge port from the composition of the gas. In addition to the air cooling, liquid cooling can also be provided to the evaporator body. For example, The refrigerant is passed through an aperture, for example 1 meter long and 6 mm in diameter, which shuttles through the evaporator body 30. The connection can be accomplished via an assembly that is assembled to the body bore 26. The liquid cooling system provides rapid cooling of the evaporator composition, To provide rapid preventive maintenance when needed. The gas can be fed into the ionization chamber 44 via the gas conduit 33, for example, via a pressurized gas drum. As described above, the solid feed can be in the evaporator 28 Evaporating and feeding vapor into the ionization chamber 44 via the vapor conduit 32. The solid feed 29' located below the perforation separation barrier 34a is also maintained at a uniform temperature by the temperature control of the evaporator body 30 as discussed above. The vapor 5 of the steady fluid 31 is read into the ionization chamber 44 through the outlet 39 and through the shutoff valves 100 and 110 via the vapor conduit 32 located in the source block 35. Thus, with gas O: \119\119484.DOC -24- 1334196 Both body and solid dopant materials can be ionized by this ion source. Figure 2c is a cross-sectional side view showing a multiple electron beam ion source set in accordance with the present invention A basic optical design of the state. In one embodiment of the invention, a pair of spatially separated electron beams 70a and 70b|j are emitted to the spatially separated heating wires n〇a and ll〇b and are deflected by the beam. Or static magnetic fields B 135 & and 135b (as indicated in the direction perpendicular to the paper) are thrown into the ionization chamber 44 by a degree of orbit, first through a pair of substrate apertures 1 〇 6 & 1 and 1 For the separated substrate U) 5a & 1 () 5b, and then through - the electron into the hole imitation 71 a and 71 b. The electrons that pass all the way through the ionization chamber 44 (i.e., through the electrons into the apertures 71a and 71b) are deflected by a beam redirector or static magnetic field 13 and 135b toward a pair of emitters (10)! (4). Electron beaming and 7〇b are transmitted through the substrate aperture 1〇 and 1〇, which enters the ionization chamber 44 by the voltage applied to the substrate 1〇53 and 1〇 (provided by the forward power source 115) ), and the voltage ve to the wire 135aAU5b (provided by the negative power supply (1)) is decelerated. It is important to make the beam formation and transfer region, that is, the electron beam energy outside the ionization chamber, which is significantly higher than that required for ionization. This is due to the space charge effect, which is severe at low energy. The ground reduces the beam current and amplifies the electron beam diameter. Therefore, it is preferred to maintain the electron beam energy in this region between about 15 Å electron volts and 5 Å electron volts. The voltages are all oblique to the ionization chamber 44. For example, if Ve = -0.5 仟V and Va = 1,5 仟, the energy of the electron beam is represented by e(Va-Ve), where e is the electron charge (6.02 x 1〇-19 ware). Therefore, in this example, the electron beam cell or 70b is formed and refracted under 2 electron volts, but enters electrons.

O:\119\119484.DOC -25- 1334196 When the apertures 71a, 71b, they have only an energy of 0.5 volts volts. The table below gives an approximate value of the magnetic field B required to deflect the electron beam by energy E via 90 degrees. Table 1

In the present invention, the magnetic field strength to complete the 90 degree deflection is dependent on the electron energy. The electron energy E is the magnetic field B.

1500 eV 51G 2000 eV 59G 2500 eV 66G The other components shown in Figure 2c include an extracted ion beam 120, a source electrostatic shield 101, and a pair of emitter shields 102a and 102b. These emitter masks 102a and 102b have two items: provide shielding to avoid electromagnetic fields, and provide shielding to avoid stray electrons or ion beams. For example, the emitter masks 102a and 102b shield the electron beams 70a and 70b from the field associated with the potential difference between the substrates 105a and 105b and the source mask 101, or as a stacking field for the stray electron beams from the opposing electrode emitters. The source shield 101 shields the ion beam 120 from the field created by the potential difference between the substrates 105a and 105b and the ionization chamber 44, which is also used to absorb stray electrons and ions that would otherwise strike the ion source component. . For this reason, both the emitter shields 102a and 102b and the source shield 101 are constructed of refractory metal such as molybdenum or graphite. Alternatively, more complete shielding of the ion beam 120 from magnetic fields B 135a and 135b can be achieved by constructing a source of ferromagnetic material such as magnetic stainless steel. Figure 2d is a cross-sectional view showing the mechanical detail and clearly showing how the contents of Figure 2c are incorporated into the ion source of Figure 2b. The electrons are thermally ionically emitted from one or more of the wires O:\119\119484.DOC -26- 1334196 ll〇a and 110b, and accelerated to a pair of corresponding anodes 14a, 14b forming electron beams 70a, 70b. These configurations provide several advantages. First, the wires 11〇3 and 11〇1) can be operated individually or together. Secondly, since the electron beam is generated outside the ionization chamber, the configuration of the transmitter with relatively long life can be extended because the emitter is in a low pressure environment of the vacuum chamber of the implanter, in which the ion beam stays in the environment, And because the emitter can also be effectively protected from ion bombardment. Magnetic flux from a pair of permanent magnets 130a and 13b and a pair of pole assemblies 125a and 125b forming a beam redirector is used to create an even magnetic field across the air gap between the ends of the pole assembly, in which the electron beam is transmitted . The electron beam energies of the magnetic fields 135a and 135b and the electron beams 70a and 70b can be matched such that the electrical pre-beams 70a and 70b are refracted by 90 degrees and passed into the ionization chamber 44 as shown. By refracting electron beams 70a and 70b', for example, via 90 degrees, the line visible between the emitter and the ion-containing chamber containing ions does not exist, thereby preventing the emitter from being bombarded by the lively charged particles. Since Va is positive with respect to the ionization chamber 44, the electron beams 70a, 70b are decelerated as they pass through the gap defined by the substrate apertures 106a and 106b and the electron entry apertures 71a and 71b. Therefore, the combination of the substrate aperture 106a and the electron entrance aperture 71a, and the substrate aperture 106b and the electron entrance aperture 71b, and the gap therebetween, each form an electrostatic lens. In this example, a deceleration lens is formed. The use of a deceleration lens allows the ionization energy of the electron beam to be adjusted without substantially affecting the generation and refraction of the electron beam. The gap may be established by one or more ceramic spacers 132a and 132b that support the substrates 105a and 105b and serve as a support for the source block 35, O:\119\1194S4.DOC -27- 1334196 The holder is below the ionization chamber potential. Ceramic spacers 132 & and 1321) provide both electrical isolation and mechanical support. It should be noted for clarity that the emitter mask 102 and the source shield 101 are not shown in FIG. The electron inlet holes 16a and 6b can restrict the transfer of the electron beams 7a, 70b, and the substrates I05a and 105b can intercept a part of the high-energy electron beams 7a, 7b. Therefore, the substrates 105a, i〇5b must be actively cooled or passively cooled. Active cooling can be achieved by passing a liquid refrigerant such as water through the substrate. Alternatively, passive cooling can be achieved by allowing the substrate to reach its ambient temperature via radiant cooling. This steady state temperature depends on the intercept beam power, the surface area and emission coefficient of the substrate, and the temperature of the surrounding components. When the condensable gas is operated, there is a benefit in allowing the substrates 10a, 5b, 5b to operate at a high temperature such as 200c, which forms a film on the cold surface and forms a film. Figure 2e shows a simplified top view of the electron beam forming region of the source. The wire 110b is at a potential % relative to the ionization chamber 44, such as _〇5 仟 electron volts, while the anode 140b, the magnetic pole composition 125b, the substrate 1〇讣, and the emitter are shielded at the anode potential Va, such as 15 Electronic volts. Therefore, the electron beam energy is 2 volts to volts. The electron beam 7〇b is deflected by the magnetic field 135b in the air gap between the pole assemblies, which allows the electron beam 7〇b to pass through the substrate aperture i〇6b. The substrate apertures 1〇6& and 1〇61>, and the electron entry apertures and their typical values are individually 1 cm in diameter. Figure 3 shows the ion source associated with the critical downstream components, including the proposed clustering and implantation system. Other than the configuration shown in Figure 3, it is possible that the ion source 21 is coupled to the extraction electrode 22 to produce an ion beam 20 containing the cluster ions. Ion beam 2〇 typically contains many ions of different masses.

O:\119\119484.DOC •28· 1334196 ' The subtraction is that the ion-rich species of the ion reserve are produced in the ion source • 21. The ion beam 2〇 then enters the analytical magnet. Inside. The analysis magnet 2 generates a bipolar magnetic field in the ion beam transfer path, which is determined by the current in the neodymium magnet coil; the magnetic field direction is perpendicular to the plane of FIG. The function of the analysis magnet is to spatially separate the ion beam into a small group by the ion beam f, and the radius of the arc depends on the mass-to-charge ratio of the dispersed ions. This brigade is shown in Figure 3 as a bundle of 24' selected ion beams. The magnet 23 is bent along a radius which is given by the equation (5):

(5) R = (2mU)1/2/qB where R is the bending half control, B is the magnetic flux density, m is the ion mass, u is the ion kinetic energy and q is the ion charge state. The selected ion beam is composed of only the energy product ions of (4) Fan Park, which allows the beam to be transmitted through a mass resolution aperture 27 by the radius of the ion beam bending of the magnet. The unselected beam composition will not pass through the mass resolution aperture", but will be intercepted elsewhere. For a beam with a smaller mass-to-charge ratio • m/q compared to the selected beam 25, for example with a mass of 1 or 2 atoms The hydrogen ion composition of the mass unit is less curved by the magnetic field, and the beam is within the radius 璧30 or elsewhere of the magnet cavity. The beam has a larger mass-to-charge ratio than the selected beam 26. The bending radius induced by the magnetic field is large, and the beam system will hit the radius 璧 29 or other places outside the magnet cavity. As has been well established in the art, the combination of the analysis magnet 23 and the mass analytical aperture 27 constitutes a mass. An analysis system capable of selecting a beam 24 selected from a plurality of species bundles 2 extracted from an ion source and subsequently selecting an ion beam 24 for subsequent analysis of the acceleration/deceleration stage 31. This amount 31 can adjust the beam energy to be used for The final energy required for a particular implant process is O:\119\119484.DOC -29- 1334196 • Measured value. The rear stage analysis acceleration/deceleration table 31 can be in the form of, for example, an electrostatic lens or -lINAC (linear accelerator). When being transferred to the wafer • It was charge exchange reaction or (and therefore does not have the correct energy) between the resolving aperture and the wafer mirror, can be "neutral beam filter" or "energy filter" incorporated within this beam path. For example, the rear-end analysis acceleration/deceleration stage 31 can incorporate a factory-shaped or small-angle deflection into the beam path, and the selected ion beam 24 is confined within this shape to subsequently apply a direct current electromagnetic field. However, the composition of the bundle that has become neutral or has multiple charges will inevitably not follow this path. The energy-adjusted beam then enters the beam scanning system 32 in the implant system of Figure 3. The beam scanning system 32 causes the beam to be scanned so that the entire target 28 can be implanted uniformly. Various types of scanning with one-dimensional or two-dimensional, and electrostatic configuration of the magnetic scanning system are possible. The beam then enters into the wafer processing chamber, which is also maintained in a high vacuum environment, with the processing chamber striking the target 28. Various wafer processing chambers and wafer control system configurations are possible, mainly in tandem (one wafer at a time) or batch (many wafers are processed together on a rotating disk). A typical dimension (horizontal or vertical) in a tandem process chamber is mechanically scanned across the beam, which is electromagnetically scanned in orthogonal directions to ensure good spatial uniformity of values. In a batch system, the rotation of the disc provides a mechanical scan in the radial direction, while the vertical or horizontal scan of the turntable can also act simultaneously, keeping the ion beam stationary. For cluster ion implantation that provides the correct dopant configuration, it is contained within the cluster. Each η-doped atom can penetrate the substrate with the same kinetic energy. In the simplest case, the molecular ion is Αη+ form (that is, its only to η

O:\119\l 19484.DOC •30· 1334196 Compensating for the atomic A composition), each n-holes must pass through the semiconductor substrate to the same beam energy fraction 1/η. It has been established, for example, by Shimin in VLSW _, McGraw Hill, pp. 253-254 (1983), when energy splitting occurs when polyatomic molecules hit a solid dry surface. Furthermore, the electrical results of this implant must be the same as the equivalent implant using single atomic ion implantation. This result has been compiled by Jacquesson et al., "Decaborane, an alternative to ultra-low energy ion implantation", ΙΕΕΕ13th International Symposium on Ion Implantation Technology, Alpine Bach, Austria, 3rd 〇 _ 303 (2000) 'detailed examples of decaborane for implantation, and indeed we expect similar results for any dopant clustering. During ion implantation, the dopant atoms penetrate deeper into the semiconductor substrate by tunneling, that is, into the substrate lattice via a symmetric direction containing a low-density lattice atom or a "tunnel". Inside. If the ion orbitals are aligned with the tunnel direction in the semiconductor lattice, ions can be substantially prevented from colliding with the substrate atoms, which can extend the dopant projection range. An effective method to limit or • even prevent tunneling, including the formation of an amorphous layer at the surface of the substrate. One method for producing this layer is to implant ions of the same element of the substrate or ions having the same electrical properties (that is, the same block from the periodic table) into the substrate, so that the implantation process is caused. The crystal damage is sufficient to eliminate the crystal structure of the surface layer of the substrate without subsequently changing the electrical properties of the substrate during the activation step. For example, after implantation of a shallow dopant layer by cluster ion implantation, ruthenium or osmium ions can be implanted into the ruthenium substrate at a dose of 5 χ 1 〇 μ cm·2 at an energy of 20 仟 electron volts. Inside, the amorphous layer is formed in the germanium crystal. 〇:\H9\119484.D〇C • 31 - 1334196 The important application of this method is Li (4) polyion implantation for & complementary MOS process - part of N-type and p-type shallow connection (four) t-complementary oxidized half The guide uses the dominant digital integrated circuit technology, and its name means the formation of both N-channel and P-channel MOS transistors on the same wafer (complementary MOS: both). The success of complementary MOS semiconductors is that circuit designers can create a better circuit by utilizing the complementary characteristics of the opposite transistors, which is particularly advantageous in that the circuit has a lower active power than other techniques. It should be noted that the N-type and p-type proper nouns are based on negative and positive (N-type semiconductors have negative majority carriers, and vice versa), while N-channel and p-channel 4 crystals replicate each other, and each The type of region (polarity) is the opposite. The fabrication of the two types of transistors on the same substrate requires the implantation of N-type impurities and subsequent implantation of p-type impurities, and other types of photoresist shielding devices. It should be noted that the polar regions of the various types of transistors are required for proper operation, but the implantation from the shallow junction is the same as that of the transistor. The Ν type is shallowly implanted into the Ν channel transistor and the ρ type is shallowly implanted into the ρ channel transistor. An example of this method is shown in Figures 乜 and 4b. The figure blunt specifically shows a method for forming an N-channel drain extension 89 through an N-type cluster implant 88, and Figure 4b shows a p-channel drain extension 9〇 formed by a P-type cluster implant 91. It should be noted that both the N and P type transistors require a geometrically similar shallow junction. Therefore, clustering with both N and P types is common to the formation of advanced complementary MOS structures. Its advantages. An application example of this method for forming an NMOS transistor example is shown in the figure & This figure shows a semiconductor substrate 41 that receives certain front-end processing steps for semiconductor device fabrication. The structure is composed of an N-type semiconductor substrate 41 which has been subjected to the steps of 44, 45, via the p-well 43, the trench isolation 42, and the O:\119\119484.DOC -32-1334196 gate stack. The p-well 43 forms a junction with the N-type substrate 々I which provides isolation of the transistor junction in the well. Channel isolation 42 provides lateral dielectric isolation between the N and P wells (e.g., throughout the complementary MOS structure). A gate stack is then formed which includes a gate oxide layer 44 and a polysilicon gate electrode 45 which are patterned to form a gate stack. The photoresist 46 is also coated and patterned so that the area for the transistor is opened, but other areas of the substrate are shielded by the photoresist layer 46. At this point in the manufacturing process, the substrate is ready for a bony extension implant, which is the shallowest doping required for the device process. The typical process required for advanced devices in the 13-micron technology segment is arsenic implant energy between 丨仟 electron volts and 2 仟 electron volts, and an arsenic dose of 5χ1〇ΐ4 cm-2. In this example, the arsenic cluster ion beam 47 is directed to the semiconductor substrate, typically by directing the ion beam in a direction perpendicular to the substrate to avoid being shielded by the gate stack. The energy of the As4Hx+ cluster must be four times the energy required for As + implantation, for example between 4 仟 electron volts and 8 仟 electron volts. These clusters decompose as soon as they collide with the substrate, and the dopant atoms appear stationary in a shallow layer near the surface of the semiconductor substrate, which forms a drain extension 48. It has been noted that the surface layer of the same implant into the gate electrode 49 provides additional doping to the gate electrode. The process described in Figures 5 & therefore is an important application of the present invention. Another application example of this invention is shown in Figure 5b: the formation of deep source/drain regions. This figure shows the semiconductor substrate 41 of Figure 5a after performing a further processing step in the fabrication of a semiconductor device. Additional process steps include the formation of pad oxide 51 and the formation of isolation layer 52 on the sidewalls of the gate stack. Here, the photoresist layer 53 is

O:\119\119484.DOC •33· 1334196 is coated and patterned to expose the implanted transistor, in this example an NMOS transistor. Secondly, the implementation of ion implantation in the formation of source and drain regions. Because of this high dose, it is a suitable application for the proposed clustering method. It is used in the 13 micron technology. The typical implantation parameters are about 6 volts volts per arsenic atom (54) at a dose of 5x1015 cm·2, so it must be 24 volts, 125 U0 U cm, 2 Α 34 Ηχ + implant, 12 volts electrons, 2.5 x 1015 cm. -2AS2Hx+ implant, or 6仟 electron volts, 5xl015 cm-2As+ implant. As shown in Figure 5a, the source and drain regions 55 are formed by this implantation. These regions are interconnected in the circuit (post Defined in the process) and defined by the drain extension 48 to provide a highly conductive connection between the intrinsic transistors connected to the channel region 56 and the gate stacks 44, 45. It should be noted that the gate electrode 45 can be exposed thereto. Implantation (as shown), if so, the source/drain implant provides the main dopant source for the gate electrode. This is shown in Figure 5b as the poly-doped layer 57. It shows the PMOS drain extension 148 and The details of the PMOS source and the gate region 155 are individually shown in Figures 5c and 5d. The results and processes are the same as in Figures 5b and 5c. In Figure 5c, the PMOS drain extension 148 is formed by implanting a boron cluster implant 147. For the 〇 13 micron technology segment 'Typical implant parameters for this implant, at a dose of 5x1014 cm·2, each boron atom will have an implantation energy of 500 eV. Therefore, the b1gHx implant system will be 5 volts electron volts and 5 x 1013 cm _2 At the dose of borax, Figure 5d shows the formation of the PMOS source and drain regions 148, again by implantation of a P-type cluster ion beam 154 such as decaborane. This implant is used for 0.13 micron technology. The typical parameters of the link are about 2 electron volts per boron atom at a boron dose of 5xl015cm·2 (that is, 20 o:\119\119484.DOC-34 at 5xl014min·2 decaborane) · 1334196 volts. In general, ion implantation alone is not sufficient to form an effective semiconductor junction; heat treatment is necessary to electrically activate the implant. After implantation, the semiconductor substrate The crystal structure is severely damaged (the substrate atoms are removed from the lattice position) Implanted dopants are only weakly bonded to the substrate atoms, so the implanted layer has poor electrical properties. Typically at high temperatures (greater than 900. 热处理 heat treatment or annealing to repair the crystal structure of the semiconductor, and blending The heteroatoms are arranged in a substituting manner, that is, at the atomic position of one of the substrates in the crystal structure. This substitution allows the dopant to bond with the substrate atoms and be electrically activated; that is, to change the conductivity of the semiconductor layer. Because of the diffusion of the implanted dopant during the heat treatment, this heat treatment conflicts with the formation of the shallow junction. In fact, boron diffusion during heat treatment is a limiting factor for the USJ to be achieved in the next 1 micrometer specification. Advanced processes that have been developed for this heat treatment to minimize the diffusion of shallow implant dopants are such as "spike pulse annealing." The glitch annealing is a rapid heat treatment in which the residence time at the highest temperature is close to zero: its temperature ramps up and down as fast as possible. In this way, the high temperatures required to activate the implanted dopant can be achieved, and the diffusion of implanted dopants can be minimized. It is our expectation that such advanced heat treatments will be used in conjunction with the present invention in the fabrication of a semiconductor device to maximize its benefits. Fig. 6 shows the generation of a phosphorus group polyion and the formation of a mass-analyzed phosphorus group ion beam. This mass spectrum shows the data obtained during the operation of the ion source of the present invention, which uses phosphine (PH3) as a source to feed the gas. This mass spectrum shows the intensity of the ion current on the vertical scale 61 versus the magnetic field of the analyzer on the horizontal scale 62.

O:\119\119484.DOC •35· 1334196 The field system determines the ion mass-to-charge ratio. The current is measured in a Faraday cup, where the secondary electrical energy is effectively suppressed. Since the two quantities are in the relationship of m/q = aB2 for a specific extraction voltage v, where a is a constant, the horizontal scale is linear with the magnetic field, and the mass-to-charge ratio is nonlinear. This causes higher quality peaks to approach each other on the horizontal scale 62. The observed phosphorus clusters are signals 65, 66 and 67 having two, three and four phosphorus atoms per cluster. This mass spectrometry demonstrates that the ion source of the present invention can support the formation and preservation of clusters during operation. On the left side of the figure, the first group of signals 63 is a hydrogen ion having a mass number 丨 and 2. The hydrogen peak is relatively small, which is much smaller than the phosphorus containing peak. The second group of signals 64 occurs between masses 31 and 35 and corresponds to ions containing a phosphorus atom. During the traditional implantation process, one, several, or all of these peaks may be implanted depending on the selected mass resolution aperture 27 (see Figure 2 for the phantom selection. If the process is sensitive to hydrogen, then some The application may only need to select 31ρ+ peak. In this example, a narrow mass analytical aperture can be applied to exclude the hydride peak ', ie PHX+, where x = 1, 2, 3, or 4. Other processes may need to be implanted All peaks are included in this group to increase the yield. The lower signal group on the right side 65 is composed of phosphorus double body P2, each of which contains two phosphorus atoms. The leftmost effective signal corresponds to the mass number 62. P2+. The adjacent signal to the right is the P2HX+ 'where X is between 1 and 6. We also noticed that the intensity of these signals is reduced compared to the peak value of the individual 64, but observed The intensity depends on the entire set of source input settings and can be optimized to the desired beam conditions, such as 'If you need a bimolecular body, you can maximize the relative height of the peaks. > 2+. The selection system determines how many such bundles need to be implanted during the implantation process. Under side 66 corresponding to the signals classified with three phosphorus-based O: \ 119 \ 119484.DOC -36- 1334196 • atom ⑺ +) of polyionic material. Right-below signal_correspond to ~ a cluster of phosphorus atoms. It should be noted that the strong accumulation of this group is in the -P3Hx+ cluster, while the use of P4+ clustering (the intensity observed by 4x) is more than the implanted Ρ+ or !V, and each implanted phosphorus The energy of an atom is only 1/4 of the nominal ion beam energy. Figure 7 shows the AsH3 mass spectrum using the present invention. The ion beam energy is 19 volts volts, so the effective As implant energy of As4Hx+ is 4 75 仟 electron volts. The beam current of As4Hx+ in Figure 7 is about 25 mA, so the equivalent impurity current is about 1 mA. Figure 7 also shows that the particle current between 0.5 mA and 1 mA is derived from the implantation of an As, AS2, As3, or As4 ion beam by simply adjusting the analysis of the magnet current to select The mass spectrum of the different portions of Figure 7 can also impart an effective implant energy stomach range between about 20 and 5 Torr and volts. Figure 8 shows that As4Hx+ current is a function of cardiac implant energy. The angular extent of the ion beam is in the lateral or scattering direction of llmR, limited by the mass resolution aperture (see, for example, 27 of Figure 3) and the aperture between the Faraday cups to half a degree, or about 6 degrees. The electron volts/former is the lower limit required for arsenic implantation into the USJ device in the semiconductor process. Figure 9 shows the beam current converted to the beam luminance unit of Figure 8, and compared with the "typical" modern medium current value machine. The improvement factor is about 3 〇 (I assume that the current value of the machine is: 4 〇 rad half-angle acceptance, and 200 amps of current at 1 电子 electron volts). Stephens in the ion implantation ‘J. F. Ziegler edition, North Holland, pp. 45 5-499 (I" 2 years), brightness B is defined as:

O:\119\119484.DOC •37· 1334196 • (6) B = 2I/71, 2 (microamperes - mm - 2 millirads - 2), . where 1 is an effective doping beam in microamperes Current, while ε is radiated by a beam of (millilar-millimeters) squared. The radiation is calculated by (7) ε = δα, where δ is the half width of the beam in the scattering plane and α is the half beam angle, which is measured at the imaginary plane, that is, at the position of the analytical aperture. The party's degree is an important characterization of its merits, which allows the beam current to be transferred to a specific incineration specification for quantification, for example by a specific diameter and length. Since the implanted beam has a well-defined acceptance specification, brightness is an important measure of the yield for the light-limited beam. We have noted that this is the main advantage of using cluster ions for monomer ions, as shown in equations (1)-(3). For the eight implants, Equation (3) predicts a 16-fold increase in output, which is Δ = η2. Figure 1 shows the results of the secondary ionization mass spectrometry (SIMS) of a sample of AsHx and As#/ion implanted at 4.75 volts electron volts and 19 angstrom electron volts. The atomic dose is about 16 gong*·2. These data are compared to the full dynamic scattering model TRIM, which is commonly used by the industry in analog ion implantation. This result indicates that we did implant As and As4 at the specified energy. Figure 11 shows the mass spectrum of diborane IH6, which is a gaseous material which is not commonly used in conventional ion implantation but is commercially available. 13 . Figure 11 shows H(H+ . H2+ . H3+) ^ B(B ^ BH+ > BH2+) > β2(Β2+ , β2Η+ , β2Η2+ . b2h3+, b2h4+), b3 (b3, b3h+, b3h2+, b3h3+, b, w+, 22 〇3«4 )> Β4(Β4 'β4η+, Β4Η2+, Β4Η3+, Β4Η4+) groups, and 1 group β Ingusuin u have two kinds of natural hair O:\119\119484.DOC •38· 1334196 The measurement of the octet ι〇 only; ·, and the existence of children and other isotopes expressed in the ratio of 丨丨β to 10β about 4:1, the reaction reflects its own, abundance, the mass spectrum is somewhat complicated to explain. For example, both 'πΒ and (4) are presented in the peak of the heart atomic mass unit. Figure 12 shows the generation of hydrogenated boron clusters and positive group ions in the present invention. This mass spectrum is not derived from the number of enthalpy collected during the operation of the ion source of the present invention, which uses evaporating decaborane. !·^4 is used as a source to feed the material. The boron hydride cluster having a ratio of 1 < 丫 <1 〇 and 〇 < X < 14 forms ByHx+ is separated by 1 atomic mass unit from 1 atomic mass unit to about 124 atomic mass units. It was observed that the maximum signal enthalpy/equivalent to the decaboration group was formed by direct ionization of the decaborane parent molecule. Figure U shows the decaborane anion mass spectrum produced by the ion source of the present invention, which is similar to the mass spectrum of Figure 12. The ionic state formed by the negative decaborane ion is much less, so most (about 9%) ions are included in the original Βι〇Η/peak. Ion implantation using a negative ion in a semiconductor is very beneficial because it substantially eliminates wafer charging observed with positive ion implantation. It is not unusual to produce a sufficient amount of positive and negative ions of a particular material with an ion source; the peak ion currents of Figures 12 and 13 are within the same factor of 2. This is exaggerated in the mass range of the extension shown in Fig. 14. The data collected by the present invention is the same as the ion implantation system of the present invention on the same paper, by collecting the positive ion mass spectrum, inverting the polarity of the ion-check-in power supply, and collecting the negative ions on the same mass range. Completed by mass spectrometry. To collect Figure 14, the Faraday cup current is delivered to a χ-y paper recorder. The important advantages of implanting negative ions instead of positive ions in the case of decaborane are shown below: 1} More useful ions

O:\119\119484.DOC -39· 1334196 _The turbulence is within the symplectic value of interest', which can cause a large singular flux; .2) The original peak mass has a coefficient of almost 2 (5) The negative of the atomic mass unit • the full width of the positive ion mass of the 9 atomic mass unit of the atomic mass), and 3) the elimination of the wafer charge when the negative ion is replaced by the positive ion, which is generally capable of being recipient. Figure 15 shows the SIMS distribution of positive and negative borax ion implants in a ruthenium sample implanted at a dose of 20 volts electron volts of decaborane. The profile is almost identical, and it is expected that if each ion is treated with the same number of boron atoms, it can be implanted into the same projection range. Figure 16 shows the implantation of iSIMS data with negative decaborane, which also shows the hydrogen concentration. The hydrogen dose is 9·9 times the boron dose, which is recommended for the average chemical formula of the minus decaborane BiqH9•. Figure 17 shows how ionization is dependent on electron energy when electrons strike ionization. Ammonia (NH3) is used as an example. The probability is expressed as a cross-section σ in units of 10-i6 cm2. The electron energy (τ) is expressed as electron volts φ. Shown are two sets of labeled BEB (vertical IP) and BEB (adiabatic IP) theoretical curves calculated from the first principle juice' and from Juric et al. (Djurie) (1981) and from Lao (Rao) and Srivas Two sets of experimental data for Tawa (1§]^ and &^) (1992). Figure 17 shows the fact that electron energy in a specific range can generate more ions than other energy ranges. These data are applied to the generation of positive ions, and similar considerations apply to the generation of negative ions: a strong energy is attached as obvious. In general, a positive ion produces a cross-section with a maximum electron impact energy between about 50 electron volts and 500 electron volts, with a peak value of about 100 electron volts. Therefore, the energy of the electron beam entering the ionization chamber 44

O:\119\119484.DOC •40- UJ4196 is an important parameter that affects the operation of the ion source of the present invention. Therefore, the electron beam is transferred to the electron beam to penetrate the ionization chamber. It varies from almost 〇 electron volts to about 5 〇〇〇 electron volts. Figure 2c to Figure. The characteristics of the present invention show how the present invention incorporates electron optics, which is controlled by a broad electron impact ionization energy and can be operated under electron beam formation and refraction zones of an ion source under near-equal conditions. .

Figure 18 is a mass spectrum of n-decaborane ion prepared from the ion source of the present invention. The individual ions that make up the mass spectrum are labeled. In general, the (four) subform is ΒηΗχ +' where 0 and 0 Sxsw. At present, the maximum peak value is the original ι〇Η' ion, and most of the peak intensity is within about 8 original mass units (atoms: units). This original ion is the most likely choice for positive ion implantation. Figure 19 shows the negative and n-decaborane ionomers made with the ion source of the present invention, and the individual ions of the edge-edge mass spectrum are labeled. Negative ion mass spectrometry is much simpler than positive ion mass spectrometry. The special feature is that there is no obvious chlorine or low ionic ions. About 90% of the mass 4 is composed of the original Βι〇Ηχ· ions. Just as the B1〇Hm ion's original ion has a waveguide strength that is mostly within about $ atomic mass units. This original separation is the most likely option for negative implantation. There are several elements of interest in the formation of shallow joints. For the second application, the main replacement is the boron, phosphorus 'arsenic and antimony. Therefore, these elements have the greatest potential applicability to the formation of shallow junctions. Furthermore, 矽 and erroneous implantation are used to form the * 丫 <amorphous region, and therefore, the clustering of 70 s of this temple is useful for the formation of the shallow amorphous region. For compound semiconductors, Lu's are used for shallow junctions or for families. 趣7L includes 矽, 锗, tin, zinc, cadmium and bismuth. Therefore, the enthalpy of these elements is clustered in the manufacture of compound semiconductors. Connect

O:\119\119484.DOC face formation will have a chance. One aspect of this method is the appropriate environment for people and people. This is provided in the ionization chamber used to form the cluster ions, for & i. The various elements have different chemical properties, which are different because of the clothes. Each element and each are selected -θ .. . ^ into the parameters for optimal performance. The parameters that can be obtained include: funeral eclipse, u 忐, wu (3) source pressure controlled by the flow rate of the feed material, electric control system controlled by the temperature control system, w ώ , . Visibility, ionization energy intensity and characteristics, electron beam current and electron energy when ionization 16 is an electron beam. These bases operate and operate to create an optimal environment within the source ionization chamber for forming and ionizing the dopants. Compared with the ion implantation of the single hetero atom in the above, the ion clustering of the doped atomic clusters makes it possible to implant the high-efficiency U-type and P-type in the shallow depth. The invention is illustrated by the following specific examples. The invention is not intended to be limited to the specific embodiments. For example, the possibilities of various modifications, changes, improvements and combinations of the present invention will be apparent to those skilled in the art. It is apparent that many modifications and variations of the present invention are possible by the above teachings. Therefore, it is to be understood that the invention can be practiced within the scope of the appended claims. BRIEF DESCRIPTION OF THE DRAWINGS These and other advantages of the present invention will be readily understood by reference to the following specification and drawings, wherein: Figure 1a is a plot 'which shows the maximum 75As+ according to the law of Childe-Lanmuir Beam current versus extracted energy map.

O:\119\l 19484.DOC • 42- Four-molecule arsenic and monomeric arsenic Figure lb is a graph showing the comparison of the maximum extraction current that can be reached. Figure 2a is a schematic diagram of a clustered ion source in accordance with the present invention. Figure 2b is a schematic diagram of a clustered ion source in accordance with the present invention. Perspective of an exemplary embodiment of the ion source Figure 2c is a side view of the source of the non-ion source of Figure 2b, shown in partial cut-away, with the electron beam and magnetic field shown superimposed thereon. Figure 2d is a partial perspective view of the ion source shown in a partially truncated manner showing the magnetic field and electrical pre-beam source in accordance with the present invention. Figure 2e is a simplified top plan view of an electron beam formed from a source region in accordance with the present invention. Figure 2f is a block diagram of a temperature control system that can be used with the present invention. 3 is a simplified diagram of an exemplary cluster ion implanted in accordance with the present invention. Figure 4a is a NMOS; and a complementary | oxygen semiconductor fabrication sequence diagram during pole extension formation. Figure 4b is a sequential fabrication diagram of a complementary MOS semiconductor during PMOS drain extension formation. Fig. 5a is a view of a semiconductor substrate in the process of the NMOS semiconductor device in the process of extending the implantation step of the drain. Figure 5b is a diagram of a semiconductor substrate during a source/drain implantation step in the fabrication of an NMOS semiconductor device. Figure 5c is a diagram of a semiconductor substrate during a p-type drain extension implantation step in a process of a PMOS semiconductor device. Figure 5d is a diagram of a semiconductor substrate in a source/drain implantation step O:\119\119484.DOC • 43 · 1334196 in a process of a PMOS semiconductor device. Figure 6 is a pH-spectrum mass spectrum produced by the ion source of the present invention. Figure 7 is a mass spectrum of AsH3 produced by the ion source of the present invention. Figure 8 is a graph demonstrating the Α~Ηχ ion current on the wafer in the low energy range. Figure 9 is a graph showing the data converted to beam luminance units in Figure 6. BRIEF DESCRIPTION OF THE DRAWINGS Fig. 1 is a SIMS profile of the erbium concentration immediately after implantation from the AsH/ and As4Hx+ beams using the present invention, and a graph comparing with the TRIM calculation. Figure 11 is a graph showing the mass spectrum of the 制造^6 produced by the ion source of the present invention. Figure 12 is a graph showing the positive ion mass spectrum recorded when the borax 8 feed material is operated in the present invention. Figure 13 is a graph showing the negative ion mass spectrum recorded when the decaborane is fed into the material in the present invention. Fig. 14 is a mass spectrum graph recorded by (iv) sub- and n-boron n successive sampling, which also shows a bimolecular β20ηχ. Figure 15 is a graph showing the SIMS distribution of negative and positive Β 〇Ηχ ions immediately after implantation using the present invention for implantation energy of decaborane at 2 Å electron volts. Figure 16 is a SIMS distribution plot of just implanted 20 Å electron volts decaborane implanted in a crucible showing boron concentration and hydrogen concentration. Figure 17 is a graph of the ionization profile 〇 as a function of ammonia (NH3) electron energy. Figure 18 is a mass spectrum of n-borane ion produced by the ion source of the present invention. Figure 19 is a diagram showing the negative decaborane produced by the ion source of the present invention.

O:\119\119484JDOC • 44 - 1334196 8 10 11 12 13 14 15 16

20 21 22 23 24 25

26 27 27 28 28a 28 29 29 Bungee extended ion source feed supply beam formation region ionization chamber ionization energy stroke electrode group poly ion beam aperture ion beam ion source extraction electrode analysis magnet selected ion beam lower mass beam water-cooled Pipeline higher mass convection air cooling pipe quality analytical aperture evaporator end plate assembly target solid source feed material outer radius 璧

O:\119\119484.DOC -45- 1334196 30 Evaporator body 30 Inner radius 璧30a Evaporator combination 31 坩埚31 Rear analysis Acceleration or deceleration table 32 Vapor conduit 32 Beam scanning system 33 Vapor inlet 33 Wafer processing chamber 34坩埚-Evaporator body interface 34b Cover 35 Source block 36 Mounting flange 39 Evaporator outlet 41 Gas feed 41 N-type substrate 41 η-type substrate 42 Channel isolation 43 Ρ-type well 44 Ionization chamber 44 Gate oxide layer 45 thyristor 46 photoresist 47 Kunqun polyion beam O:\119\119484.DOC -46- 1334196

48 Pole extension 49 Q electrode 50 Vapor 51 塾Oxide 52 Isolation layer 53 Photoresist 54 Kunming poly ion beam 55 Source/nothing 56 Channel 57 Polycrystalline holding 61 Ion current 70a Electron beam 70b Electron beam 71a Electron beam Entering aperture 71b Electron beam entering aperture 80 Ion extraction aperture plate 81 Ion extraction aperture 81 N-well 82 P-well 83 Gate 84 Gate oxide layer 85 Channel isolation 86 Photoresist 87 Photoresist O:\119\119484.DOC -47 - 1334196

88 N-type cluster 89 NMOS immersion extension 90 P Μ 0 S immersion extension 91 P-type cluster 100 isolation valve 101 source shield 102a emitter shield 102b emitter shield 105a substrate 105b substrate 106a substrate aperture 106b substrate aperture 110 isolation valve 110a wire 110b wire 111 vent valve 115 forward power source 116 negative direction power source 120 ion beam 125a magnetic pole composition 125b magnetic pole composition 130a permanent magnet 130b permanent magnet 132a ceramic spacer • 48- O: \119\119484.DOC 1334196

132b ceramic spacer 135a magnetic field 135b magnetic field 135 magnetic field 140a anode 140b anode 141 P-type substrate 141 P-type substrate 142 trench isolation 143 n-type well 144 gate oxide layer 145 gate 146 photoresist 147 boron group polyion beam 148 Extension 149 Gate electrode 151 塾 oxide 152 isolation layer 153 photoresist 154 boron group polyion beam 155 source / drain 156 channel 157 polycrystalline

O:\119\119484.DOC • 49- 1334196 [Description of the Invention] [Embodiment] [Simple description of the diagram] [Description of main component symbols]

O:\119\119484.DOC

Claims (1)

1334196 'Patent Application No. 096109576 Chinese Patent Application Scope Replacement (March 10, 1999, Patent Application Range: 1. A semiconductor substrate by implanting a doping material to a first and second region) A method for fabricating a semiconductor device, the method comprising the steps of: (a) implanting a group of N-type dopants of a first molecular species into a first region of the article to cause the N-type doping of the substrate; and (b) implanting a p-type dopant group ion of a second molecular species into the second region of the base material different from the first region to form a semiconductor The device has N-type cluster ions and p-type group poly ions. 2. A method for fabricating a semiconductor device by implanting a dopant material into a semiconductor substrate, comprising the steps of: generating a first molecular species type a heterogeneous group of polyions; implanting the N-type dopant cluster ions into a first region of the substrate to cause N-type doping of the substrate; generating a negative p-type of the second molecular species Dopant group polyion; and • these "type dopant groups A polyion is implanted into a second region of the substrate to form the semiconductor device. 3. A semiconductor device comprising: a substrate; - a group of N-type molecular dopants in the form of Asjix+ Into the first region of the substrate, wherein χ and η are integers, η is equal to 3 or 4, and A 1 < X < η + 2 ; ^ a Ρ-type dopant of the semiconductor device. A second region of the material to form 119484-990304.doc 1334196. 4. The semiconductor device of claim 3, wherein the p-type dopant is BnHx+ 'where j^x is an integer, and 〇gxS6. The semiconductor device of claim 3, wherein the erbium type dopant is ΒηΗχ· 'where η & χ is an integer ' is equal to 1 〇 and 〇 2 χ $14. 6. The semiconductor device according to claim 3 Wherein the p-type dopant is BnHx+ 'where χ is an integer, η is equal to 1 〇 and 〇gxS14. 7. A semiconductor device comprising: a substrate; a PnHx+ form of N-type dopant clustering An ion implanted in the first region of the substrate, wherein η & And P is an integer, MM4, and hK6; and a P-type dopant implanted in a second region of the substrate to form the semiconductor device. V 8. The semiconductor device according to claim 7 , wherein the p-type dopant is ΒηΗχ+ 'where n and x are integers, 2sns4 and 0 <x; s6. 9. The semiconductor device of claim 7, wherein the p-type dopant is ΒηΗχ- 'where 〇 and \ are integers' η is equal to 1〇 and 〇sxgi4. The semiconductor device of claim 7, wherein the p-type dopant is BnHx+ 'where n&x is an integer and η is equal to 1 〇 and Μΐ£4. 11. A semiconductor device comprising: a substrate having a first and second regions; a first molecular species of N-type dopant cluster ions implanted in a region of the substrate; A P-type dopant cluster ion is implanted in a second region of the substrate to form the semiconductor device. A semiconductor device comprising: a substrate having a first and second regions; And a negative P-type dopant cluster ion of a second molecular species implanted in the second region of the substrate to form the semiconductor device. 119484-990304.doc