Ion implantation with high brightness, low emittance ion source, acceleration-deceleration transport system and improved ion source construction

Use high-brightness and low-emission ion source ion implantation, acceleration and deceleration transportation system and improved ion source structure

<p>a, 35, 548. . . source of ion</p><p>c. . . Impurity gas feed</p><p>e,f. . . Double gasifier</p><p>h. . . cathode</p><p>j. . . Rejection electrode</p><p>k, 1, 41,526. . . Vaporizer feed pipe</p><p>m, n. . . Pressurizer</p><p>D. . . Strong discharge ion current</p><p>b, 36. . . Fixed flange</p><p>d. . . Glow discharge chamber</p><p>g. . . Aluminum zone</p><p>I. . . Uniform magnetic field</p><p>P. . . Discharge power</p><p>B. . . magnetic field</p><p>1. . . Silicon wafer</p><p>2. . . Ion beam</p><p>4. . . Non-parallel ion</p><p>5. . . Depth distribution</p><p>6. . . Launch range</p><p>7. . . Scatter</p><p>8. . . emission</p><p>9. . . source of ion</p><p>10. 630, 640, 660. . . beam</p><p>11. . . Divergence</p><p>12. . . Gate oxide</p><p>13. . . Gate electrode</p><p>14. . . Gate side</p><p>15. . . Drain extension</p><p>16. . . Gate electrode</p><p>17. . . AA</p><p>18. . . Boron concentration</p><p>19. . . Logarithmic ratio</p><p>20n. . . Type doping concentration</p><p>twenty one. . . Side of junction</p><p>26. . . Water cooling channel</p><p>27. . . Gas cooling channel</p><p>28. . . External vaporizer</p><p>28a. . . Chassis device</p><p>29. . . Feed in material</p><p>30. . . Vaporizer body</p><p>30a. . . Vaporizer combination</p><p>31. . . crucible</p><p>32. . . Conduction channel</p><p>34. . . Crucible-vaporizer body interface</p><p>34a. . . Pot sieve</p><p>34. . . b outer cover</p><p>39. . . the inside diameter of</p><p>42. . . Electron gun</p><p>44, 175, 175,, 240. . . Ionization chamber</p><p>45, 45', 174, 174", 544... Aperture</p><p>46. . . Ion extraction aperture narrow hole</p><p>47. . . Electron beam exit aperture</p><p>50. . . Gasification material</p><p>70, 250, 536'. . . Light stop</p><p>74. . . Narrow hole</p><p>80, 260. . . Ion extraction aperture</p><p>90. . . Magnet coil</p><p>100, 110. . . Heating gate valve</p><p>125. . . Electron beam</p><p>130. . . Water-cooled retainer</p><p>143 143'. . . cathode</p><p>144. . . Extractor</p><p>145. . . Beam shaping electrode</p><p>145', 147', 153, 152, 150', 152'. . . lens</p><p>147. . . First cathode</p><p>149. . . Focus electrode</p><p>150 150'. . . Second cathode</p><p>170 170'. . . Long filament</p><p>171. . . Filament core</p><p>172. . . DC power supply</p><p>173. . . With electron beam</p><p>176. . . Ion extraction aperture length</p><p>177. . . Ion extraction aperture plate</p><p>178. . . Beam shaping electrode</p><p>179. . . Grid electrode</p><p>190. . . Vacuum conduction</p><p>200. . . Three-stage vacuum tube</p><p>210. . . Ring electron beam</p><p>220. . . High temperature vaporizer</p><p>225. . . Steam pipe</p><p>235. . . Electron beam</p><p>270. . . With ion beam</p><p>300. . . Lens assembly</p><p>310. . . Holder</p><p>320. . . Cathode combination</p><p>330. . . First cathode</p><p>340. . . Focus electrode</p><p>350. . . Exit lens</p><p>360. . . shell</p><p>360a. . . Bottom edge</p><p>370. . . Holder</p><p>380. . . Feed tube</p><p>390. . . spring</p><p>395. . . Ceramic holder plate</p><p>400. . . Aluminum rod</p><p>410. . . Aluminum spacer</p><p>500. . . Lengthened ionization chamber</p><p>510. . . Lengthened ion extraction aperture</p><p>516'. . . Ionization volume</p><p>520. . . Extraction lens</p><p>528. . . Low temperature vaporizer</p><p>530. . . Beam path</p><p>540, 543. . . Analyzer magnet</p><p>545. . . Turntable</p><p>553, 557. . . electrode</p><p>555. . . Substrate</p><p>587. . . Mirror</p><p>600. . . Dual ion source system</p><p>610. . . Double-slit extraction optical system</p><p>650, 690. . . panel</p><p>670. . . Scan Station</p><p>680. . . direction</p><p>700. . . Vacuum box</p><p>710. . . Function room</p>

Figure 1 is a schematic diagram of a conventional ion source for ion implantation.

Figure 2 is a partial enlarged schematic diagram of the ion source based on the ion implantation in Figure 1.

Figure 3 is a perspective view of the ion implantation source according to the present invention. This figure is a cross-sectional view from the center of the ion implantation source to show the internal structure.

Figure 4 is an enlarged perspective view of the ionization chamber in the ion source based on the ion implantation of Figure 3.

Fig. 4a is a schematic diagram of the electron optical system in the ion implantation source used in the preferred embodiment shown in Fig. 4 according to Fig. 3.

Figure 5 is a schematic diagram of the power supply bias diagram. The power is used to provide the ion source for the ion implantation shown in Figure 3.

Fig. 6a and Fig. 6b are schematic diagrams of the side section and the top section of the ionization chamber in the ion source for ion implantation according to the present invention.

Figures 7a and 7b are perspective views and horizontal cross-sectional views of a device that can improve the electron beam focal length of the ionization chamber in Figure 6.

Figure 7c is a schematic diagram showing the spatial geometric relationship of the device in Figures 7a and 7b.

Fig. 8 shows a perspective view, in which the ion implantation source similar to that in Fig. 3 integrates the device in Fig. 7a and Fig. 7b into the ion implantation source.

Figure 9 is a schematic horizontal cross-sectional view of another example of the ionization chamber according to the present invention.

Figure 10 shows a cross-sectional view of the ionization chamber in Figure 9. The ion implantation source in Figure 3 will be integrated into the ion implantation source.

Figure 11 is a schematic diagram of ion implantation.

Figure 12 is a schematic diagram of the ion beam emitted by the ion source.

Figure 13 is a schematic diagram of ion implantation, which forms a drain to expand the gate adjacent to a target substrate.

Figure 14a is a schematic diagram. The gate side and drain extension layer in the figure are formed by implanting boron ions on a phosphorus-doped silicon substrate. Figure 14b shows the boron along AA in Figure 14a. Logarithmic graph with phosphorus ion concentration.

Figure 15 is a graph showing the relationship between the ion implantation energy of the ion implantation and the prediction of lateral scatter at different ion target incident angles with a function curve.

Figures 16a and 16b are schematic diagrams respectively showing the lateral scatter prediction of ion implantation at normal incidence and a non-parallel incidence angle of 7 degrees.

Fig. 17 is a side view of the device similar to that of Fig. 4, showing another ion implantation source according to the present invention to generate an extended band beam.

Figure 18a is a perspective view of a dual ion source system according to the present invention, which is used in a single ion doping device to implant n-type impurities and p-type impurities.

Figure 18b is a perspective view of the ion source shown in Figure 18a implanting ions on a panel substrate.

Figure 19 is a perspective view of a doping device for a panel display, including the design shown in Figure 18a.

Figure 19a is a diagram showing the crack surface of decaborane.

Figure 20 is a side cross-sectional view of an acceleration and deceleration ion implantation system. The acceleration and deceleration ion implantation system has a fixed beam line, and the wafer is located on a turntable.

Figure 21 is a side view of an ion source suitable for decaborane, etc. The ion source is modified and incorporated into the casing of the traditional acceleration and deceleration ion implantation system shown in Figure 20.

Figure 22 is a side view of another ion source used for decaborane, etc., which is characterized in that the extraction aperture can be extended to generate an initial ion beam with an extended cross-section;

Figure 23 is a vertical cross-sectional view showing an ion source that can use magnetic force to restrict electron beams from passing through an ionization chamber;

Figure 24 is a schematic diagram of ion optics of the extended section of the ionization chamber that draws the ion beam.

Fig. 25 is a perspective view of the lens and the lens holder according to the present invention in combination with an unexplained ratio.

Fig. 26 is an unexplained scale cross-sectional perspective view of the electron gun according to the present invention of Fig. 27.

Figure 27 is an unsealed perspective view of the electron gun according to the present invention.

Field of invention

The present invention relates to an ion source and ion implantation. The present invention particularly relates to an ion implantation, acceleration and deceleration transportation system using a high-brightness and low-emission ion source, and an improved ion source structure.

Background of the invention

The following patent applications are for reference in order to detail the background of the present invention: Provisional patent applications No. 60/267 and 260 were filed in the United States on February 7, 2001 by the inventor Thomas Hess (Thomas N. Horsky), the name of the invention is an ion source for ion implantation; provisional patent applications No. 60/257 and 322 were filed in the United States on December 19, 2000, and the inventor Thomas He The name of the invention is ion implantation; PCT patent application No. US00/33786 was filed on December 13, 2000. The inventor Thomas Hoskey, the name of the invention is ion implantation ion source, system and The method and application for a leaf on November 30, 2000 are also the same for reference. The present invention is a U.S. application, which is a provisional patent application No. 60/170,473 filed on December 13, 1999, and it is currently invalid.

Background: ion implantation

For nearly two decades, ion implantation has been the core technology in semiconductor device manufacturing, and it has also been used in current transistor pn junction manufacturing, especially complementary metal oxide semiconductor (CMOS) devices such as memory and logic chips. By generating impurity elements (for example, ⁷⁵ AS, ¹¹ B, ¹¹⁵ m, ³¹ P,or ¹²¹ Sb) positively charged ions to make transistors such as silicon substrates. The ion implanter can selectively control the structure energy (implantation depth) and ion current (dose) of the imported transistor at the same time. Ion implanters generally use an ion source to generate a ribbon beam that can be as long as about 50mm; the required dose and uniformity of the dose during the process of moving the beam to the substrate are by electromagnetic scanning of the ribbon on the substrate, or by the beam Perform mechanical substrate scanning, or both.

With the advancement of wafer manufacturing technology to 300mm-diameter silicon substrates, how to use traditional ion implanters to produce larger belts to increase wafer yields when manufacturing large substrates has become a hot topic. The lower extension band beam space-charge amplification (space-charge amplification) enables more ion current to migrate in the implanter beam, so the large band beam can bring higher dose rate. Many new planter designs also include a continuous chamber (one wafer at a time), with high tilt capability (e.g., up to 60 degrees to the substrate). The typical practice is to scan the ion beam electromagnetically across the wafer, that is, mechanically scan vertically to ensure uniformity. In order to meet the requirements of implant dose uniformity and specification repetition, the angular position and spatial position of the ion beam must be highly uniform (for example, the beam angular position uniformity of the wafer &lt;ldeg). The generation of beams with these characteristics has many limitations on the beam traveling optics of the implanter. The use of large-scale emission plasma ion sources often results in divergence of the beam diameter and beam angular position, which causes the beam to become a virtual image (Vignetting) due to the aperture. Lost in. At present, it is still very difficult for continuous implanters to generate high-current ion beams at low energy (&lt;2keV). For example, in some low-energy implants (such as the source and drain structures in the front-side CMOS process), wafers The production capacity is extremely low. A similar migration problem also exists in the batch planting machine (the wafer is fixed on the turntable during the process), especially under the condition of low beam energy.

While the optical design of the beam traveling can achieve almost zero errors, the beam characteristics (spatial and angular divergence) are mostly determined by the emission characteristics of the ion source (that is, the characteristics of the beam of ion extraction are determined by the implantation of the ion source after the ion is emitted. Optical can focus and control the degree of beam). The glow discharge plasma source generally used at present has a poor emissivity, which limits the ion implanter to produce ion beams with good focusing ability, precision, and easy control.

Background introduction: ion implantation ion source

The implanter ion source used in the industry standard is an enhanced Bernard ion source (Enhanced Bernard source). As shown in Figure 1, this is a glow discharge ion source with a reflective structure: a hot wire (hot filament) cathode emits thermionic electrons in the ionization chamber (containing impurities fed into the gas), which is restricted by a magnetic field and is located at the anti-cathode Reflection, located at the other end of the ionization chamber. The electrons travel in a spiral orbit between the cathode and the anti-cathode, and produce a high-density plasma (about 10 ¹² Ion/cm ² ). A so-called "plasma column" is formed, which is parallel to the ion extraction aperture narrow hole, and the light beam forms an optical system to extract ions through the extraction aperture narrow hole. By generating high-density plasma and maintaining a discharge current of about 10A, the enhanced Bernard ion source effectively dissociates tightly bound molecular species such as BF ₃ . However, most of this type of ion source emission requires the following plasma-related effects:

1) The plasma voltage (usually about 5V) brings a speed to the ions, and then the angular dispersion of the extracted ions gradually increases.

2) The temperature of ions and electrons in the plasma can reach 10,000K, which brings a warming speed (thermal speed) other than the speed in (1), and at the same time brings an energy dispersion of approximately eV for the ions (according to Maxwell Bozeman distribution), which makes the light beam show dispersion aberration.

3) Coulomb scattering between ions in the plasma brings extra ion energy non-themial spread.

4) Strong extraction flow density is needed to control excessive ions (such as BF ₃ BF in plasma ⁺ , BF ₂⁺ , And P ion), increase the space charge force when drawing, and enhance the emission.

5) A strong magnetic field is needed to operate the glow discharge power source to deflect the beam and enhance the emission of the extracted ion beam, especially under the condition of low beam energy.

6) The high-frequency noise in the plasma entering the die beam is a high-frequency variation of beam flow and light pressure. The sudden loss of low-energy electrons that cause obvious stability or even surrounding the beam (by the positive light pressure) makes it difficult to maintain the beam plasma charge compensation due to the time-varying light pressure, which leads to the amplification of the space charge of the ion beam.

7) The ion extraction aperture cannot be excessively enlarged, for example, 75mm is an example (generally the length is between 20mm and 30mm), because this will increase the required plasma column significantly. If the distance between the cathode and the anti-cathode is large, the Bernard ion source will start to become unstable. The increase in the distance between the cathode and the anti-cathode increases the demand for glow discharge current to maintain plasma stability, which in turn increases power consumption.

Background introduction: ion deceleration

In the design of the traditional ion implanter, the transmission effect of low-energy boron with an energy of less than a few keV is not good. In the manufacture of semiconductor chips with a process of less than 0.18micron, the cost-effectiveness of the weak beam of boron is extremely low. The next generation of implanters has been studied for a long time, and designs that merge the theories of different ion optical systems in recent years have entered the equipment market, trying to solve this low-energy transmission problem. In order to solve the repulsion between the ions that determine the beam migration in the low energy state, a method called "deceleration (dece1)" has been developed so that the implanter can absorb at a higher energy state than the required implantation energy. And move the ion beam, thereby reducing the space charge effect and the influence of glow discharge. And by introducing a deceleration period from the back of the beam line to the wafer terminal station, the ion energy is reduced to the required implantation energy when the ions arrive at the wafer terminal station, for example, 2keV for ion beam extraction and migration, before reaching the wafer When decelerating to 500eV, the beam flow is stronger than when the beam line is limited by the space charge beam in the traditional design without deceleration. However, the use of this deceleration method will bring another serious problem and endanger the practicability of this method. When the ion beam passes through the decelerating lens to the wafer, the spatial distribution of the ion beam is extremely uneven. In addition, when the ion hits the substrate, the range of the incident angle with the wafer surface is too wide, which may cause the "tunneling effect." The spatial and angular dose uniformity of the decelerating beam is more traditional, and the performance of non-decelerating ion implantation is much inferior, which makes it difficult to achieve uniform dose, which affects cost and productivity considerations. Since ion implanters generally only sample a part of the ion beam before or after acting on the wafer plane, the extremely uneven ion beam distribution will also affect the accuracy of the implant dosimetry angle. The dosimetry department is used to control the distribution of implants in a desired range. The accuracy of dosimetry occurs when part of the sampling acceleration and deceleration implantation system beam ion current, the transitional extension and uneven distribution of beam ion current will cause inaccurate implantation, affecting the cost of implantation system capital, wafer quality, and System capacity.

In addition, it is different from other shallow, low-energy implants that have been proposed (but not used in current machines). If the conventional implanter without deceleration uses molecular ion beam (with the required impurity clusters), the deceleration procedure is not required. Decaborane is an example of this type of molecular material.

Wafer manufacturers are currently moving towards the manufacture of complementary metal oxide semiconductor (CMOS) memory and logic chips on 300mm-diameter silicon substrates, hoping that the new process can replace the 200mm substrate process at a more competitive cost. Although this step can only be achieved through the establishment of new factories and the purchase of new semiconductor manufacturing equipment, in addition, these two factors are closely related to the reduction of wafer costs. The multi-billion-dollar equipment purchases are expected to achieve lower-cost manufacturing processes, as well as commercial and front-side (ing-lateral) mass production of semiconductor wafers. Cost reduction means that the wafer production unit value of the wafer fabrication plant equipment must be the same for 300mm and 200mm substrates, which is feasible to a certain extent. However, the use of ion implantation to manufacture ultra-shallow (and ultra-high-density) semiconductor junctions, even the latest acceleration and deceleration implanters, still has the problem of wafer capacity limiting the dose rate, which leads to the use of large-size wafers when semiconductors are processed. The unit yield of crystal cubes did not increase significantly. The problem faced by chip manufacturers is: to set up more planting machines to make up for the reduced unit production rate and increase planting costs (more investment capital, larger floor area maintenance costs of the wafer manufacturing plant, etc.) ), making the idea of reducing the potential cost per unit die by using a large wafer size is actually not feasible.

Background introduction: ion doping

In the past ten years, the research and development of the implantation system of the ion implantation of the super-large substrate has made the flat panel display come out. This type of "ion doping" system transmits a long ribbon ion beam onto a glass or quartz substrate, and uses a static ion beam to perform a typical mechanical scan to complete. For example, for a large substrate with a size of one meter, the ion ribbon beam must also grow with the size to ensure a uniform doping level (usually wider than the substrate). In order to meet this point, a large-volume "cylinder" ion source must be used. The rectangular or cylindrical cylindrical ion source is a space surrounded by a row of permanent magnets, and the plasma is sealed by the magnetic force of the cusp magnetic field. The plasma is generated by one or more RF antennas, which are connected to the RF voltage. The large-diameter ion source emits a beam of light that passes through the extraction lens.

The ion doping system of this scale does not use mass analysis, and the ion species produced by the cylindrical source are migrated and implanted on the substrate. This has caused many problems including the depth of ion implantation and excessive implantation of ion species. In addition, the large ionization volume of the cylindrical source is particularly prone to precipitation. The manufacturer must use a single ion doping system to prevent serious cross-contamination between n-type and p-type impurities. That is to say, the manufacturer must purchase a certain type of impurity (for example, from diboron Design equipment for boron of alkane gas, and another design equipment for n-type impurities (for example, phosphorus from phosphine gas), which leads to double the equipment capital. Not only is the substrate moved in the two systems to make the wafer manufacturing process The time that the medium substrate is exposed to the air increases, which increases the risk of product output.

It can be seen from the above that the conventional bucket ion source technology has the following limitations:

(1) The single grounding area is large (length, width, height).

(2) The cost is too high and the design is too complicated.

(3) The large outer wall surface area and the large ion source volume cause B (from B ₂ H ₆ Feed gas) and P (from PH ₃ The feed gas) dissipates to the outer wall of the ion source, resulting in low ion productivity.

(4) In the ion source related to (3), there will be problems caused by pollution and particle deposition, which will reduce the productivity.

(5) A large number of excessively generated ions are implanted on the substrate, which degrades the implantation process control and equipment performance. For example, the p-type impurity B, which is often used when planting boron, occurs ₂ H ₆ Plasma has a significant H ⁺ With BH _x⁺ And B ₂ H _x⁺ The phenomenon of division.

(6) Since the total ion current delivered to the substrate must be lower than a certain limit to avoid overheating the substrate, the implantation process must be strong H ^＋ Ion implantation (affected by (5) above) will limit the achievable dose rate and productivity.

Summary of the invention:

The object of the present invention is to provide an ion implantation method that generates a high-brightness ion beam along a main axis by gas or vapor of ionized molecules, the molecules including at least one implantable species. The method has the following steps: providing an ionization chamber with a restricted outlet aperture, and providing a pressure in the ionization chamber to make the pressure of the gas or vapor greater than the pressure of the extraction region, and the extraction region is extracted from the ions in the ionization chamber After entering the space; use the main electron to ionize the gas or vapor near the exit aperture of the ionization chamber by the electron impact dissociation method, and generate an ion density of at least 10 at the aperture ¹⁰ cm ^-3 , While controlling the lateral kinetic energy of the ion to be less than 0.7eV; the ionization volume width near the aperture produced by the ion density, which is less than three times the relative exit aperture width; controlling the ionization chamber conditions to avoid the occurrence of arc discharge; by A pumping system, the ions generated in the ionization chamber are pumped from the ionization chamber to the end of the pumping area through the exit aperture; after that, the beam is transferred to the surface of a terminal station by the ion beam optical system; and the delivered ion beam is implanted on On the terminal station.

The implementation of the present invention can also integrate one or more of the following modified features. These changes are characterized by controlling the conditions in the ionization chamber to avoid plasma generation; the brightness of the ion beam during extraction is approximately greater than 1mA-cm- ² -deg- ² x(E/E ₀ ), E system beam energy, E0=10keV; ion current density is at least 1mA/cm ² , Even when the ion mass is 120amu, the x-emissivity (x-emission) of the beam during extraction is less than 70mm-mrad x(E0/E) ^1/2 (E system beam energy, E0=1 0keV); the beam noise of the ion current drawn from the exit aperture is controlled below 1%; any magnetic field strength in the ionization chamber is less than 70gauss; any magnetic field strength in the ionization chamber is less than 20gauss; ionization There is no magnetic field in the chamber; the strength of any magnetic field in the extraction area is less than 20gauss; the gas or vapor consumption is controlled below 10sccm; the main electron system is a light beam generated outside the ionization chamber and guided into the ionization chamber by an electron optical system; Each ionized molecule includes at least two types of implantable atoms, or is composed of at least two types of implantable atoms.

Another object of the present invention is to provide an ion implantation method including the steps of: generating a high-brightness ion beam along a main axis; providing an ionization chamber with an outlet aperture; supplying gas or vapor molecules in the ionization chamber, wherein each ionized molecule is separately Including at least two types of implantable atoms, or consisting of at least two types of implantable atoms; ionizing the molecule and using the ions formed by the molecule to generate a beam of light under controlled conditions, the brightness of the ion beam when the beam is drawn About greater than 1mA-cm ^-2 -deg ^-2 x(E/E ₀ ), where E system beam energy, E0-=10keV, the x-emissivity (x-emission) of the beam when absorbing is less than 70m m-mradx(E0/E) ^1/2 (E-system beam energy, E0=10keV), after that, use ion beam optical system (ionbeam optical system); use ion beam optical system to transfer the beam to the surface of a terminal station; and plant the delivered ion beam at the terminal station superior.

The implementation of the present invention can also be carried out by integrating one or more of the following features. The more dynamic feature is that the molecule is dimer; the molecule includes decaborane; it produces a high-brightness ion beam with low angualr divergence to make the angle between the terminal station and the spindle About less than 1 degree; the step of implanting the delivered ion beam on the terminal station is used to generate a high-brightness, low divergence beam to form a drain extension region of the transistor structure on the terminal station, wherein the transistor structure includes A source, a gate and a drain; the terminal station (gate) further includes a well region (Well) impurities and the gate length of the transistor structure is equal to or less than 0.20um' The drain extension and the gate intersect on the lateral side of the junction, and the lateral steepness of the arsenic implantation of the drain extension is less than or equal to 3nm/decade, where the definition of the lateral steepness of the distribution It is the lateral degree of a decade that is required to complete a decade in the unit concentration of a planted species on the side of the lateral junction. extent)'The side edge of the junction is defined as the area where the implanted ions have the same unit concentration as the impurity in the well area; the drain expands its lateral steepness less than or equal to 2nm/decade; high-brightness, low-divergence angle beam ions are distributed It is planted between the two ends of the gate to clearly define the channel under the gate; clearly defining the channel under the gate includes clearly defining the length of the channel. Decaborane ions with an ion current greater than 0.5mA are implanted in the substrate to dope the pn junction, so as to generate lattice damage in the silicon substrate and reduce channeling (about ions along the lattice The known phenomenon of structure travel and entering distance), to carry out shallow planting.

Another object of the present invention is to provide an ion implantation system for implanting terminal station substrates with low implantation energy. The period is used to accelerate the ion so that the ion transport energy (transport energy) is greater than the proper implantation energy. In the deceleration period, the ion energy is reduced to the proper implantation energy before the terminal station substrate is implanted.

The implementation of the present invention can also integrate one or more of the following modified features. The modified feature is that the ion source includes an electron gun to generate an electron beam with a certain controlled energy by direct electron impact to ionize molecules; the energy of the electron is between 20eV and 500eV; the electron gun located next to the ionization chamber emits an electron beam Pass the ionization chamber to the beamdump; an elongated ionization chamber has a relatively elongated narrow hole extraction aperture, before the beam enters the analyzer, relative to the corresponding extraction aperture, the electron optical system after the aperture is used to reduce the combination Beam beam) distribution length; the electron optical system includes an optical magnifying glass (telescope); the aperture drawn by the ionization chamber is about 6 inches long; the ion implantation system of the present invention can be set up in a batch operation mode, in which a set of wafers is fixed The relative migration of the beam on a carrier affects the scanning progress; the ion implantation system of the present invention can be a continuous ion implanter. The ion source has a vaporizer and a temperature control system of the temperature vaporizer, which are fixed in an ion source ionization chamber together; the ionization chamber electron gun and a light stopper facing the electron beam are thermally insulated from the ionization chamber; combined with the electron beam to stop light The ion implanter is thermally isolated from the rest of the ion source. The ion implantation system includes controlling the temperature of the ionization chamber. The advantage of the ionization chamber is that it is fixed on a heat transfer device with a cooling mass transfer zone. The ion source is a decaborane source, The electron gun is used to provide electron beam energy between about 50 to 1000 eV; the ion source is As ₂⁺ Ion source; the ion source is P ₂⁺ Ion source; the ion source is B ₂⁺ Ion source; the ion source is In ₂⁺ Ion source; the ion source is Sb ₂⁺ Ion source; another object of the present invention is to provide an ion implantation method using any one of the ion implantation systems according to the present invention.

Another object of the present invention is to provide an ion implantation method for implanting terminal station substrates with low implantation energy, which includes generating molecular ions (using molecules with appropriate cluster types for implantation), and accelerating the ion migration The transport energy is greater than the proper implantation energy, and the ion energy is reduced to the proper implantation energy before the terminal station substrate is implanted.

The embodiments of the present invention can also be integrated with one or more of the following modified features. The modified feature is that the ion is decaborane; the ion is P ₂^＋ Ion source; the ion system B ₂^＋ Ion source; the ion is In ₂^＋ Ion source; the ion is Sb ₂^＋ ion. Another object of the present invention is to provide an ion implantation system including: an ion implanter having an ion extraction system; an ion source capable of supplying ions required by the ion extraction system at a commercial ion current level, and the ion source includes an ionization chamber wall surface Surrounded and formed to contain an ionization volume, one of the walls determines the position and the length and width of the extraction aperture, so that the extraction system can extract the ion current from the ionization volume; an electron gun is set in size and position to match the ionization chamber. A main electron beam is emitted along the main axis of the ionization chamber; a stopper is juxtaposed with the electron gun to receive the direction beam, and the stopper maintains a positive voltage relative to the electron gun emitter voltage. The main beam path extends in the direction of the main radial axis Approximately adjacent to the aperture, the direction of the electron beam is the same as the direction corresponding to the width of the extraction aperture, and the size of the electron beam is approximately equal to or greater than the width of the aperture.

The embodiments of the present invention can also be integrated with one or more of the following modified features. Preferably, the system includes two ion sources, one for n-type ion doping and the other for p-type ion doping. The two ion sources are positioned adjacent to each other as if fixed on a double-gate extraction optical system. The ion source generates a substrate with an ion beam implanted on a plane. The modified feature is that the ion implantation system further includes a vaporizer to introduce gas into the ionization volume, a gas channel to enter the gas into the ionization volume, and a control system for controlling the energy of the main electrons. The main electrons are impacted by the electron gun. To ionize individual gas or vapor molecules; the gas includes decaborane; provides a method for forming ions as a ribbon beam; the ribbon beam is shorter than the ion extraction aperture length; the ribbon beam is longer than the ion extraction aperture; the ribbon beam Approximately equal to the length of the ion extraction aperture; the length of the aperture is at least equal to the length or width of the terminal station substrate.

Another object of the present invention is to provide a method for irradiating an extended plane with a predetermined size, which includes generating an ion beam by the ion implantation system of any one of the 46-53 scope of the patent application, and transferring the ion The beam guide extends to a flat surface.

The embodiments of the present invention can also be integrated with one or more of the following modified features. The more dynamic feature is that the extended plane is a flat panel. The method includes completely irradiating the surface of the full panel; the generated ion beam is static, and the ion doping of the panel is completed by mechanical beam scanning; The length is longer than the length of the panel substrate in the vertical scanning direction. The two ion sources are fixed side by side. For example, any one of n-shaped or p-shaped doped ion beams can be directed to the panel through a double-grid extraction optical system arranged on the opposite side of the extraction aperture of the parallel ion source.

The advantages of electron-ion implantation using beam ion source

From a certain level, the present invention reduces the impact of the above-mentioned problem of traditional ion implantation sources. The disclosed invention has the following characteristics, and the ultra-low emission ion source formed by it is particularly suitable for the needs of next-generation ion implanters:

1) Since plasma is not used, there is no problem of plasma voltage.

2) Low ion density (about 10 ¹¹ cm ^-2 Or lower), to reduce the Coulomb scattering between ions, and to reduce the ion energy distribution to a negligible level.

3) The gas molecules are ionized by the direct electron impact ionization method to form "cold" ions, which are similar to the low thermal energy contained in neutral gas molecules, such as less than 0.7eV &lt;leV. This enables the ion source to have high monochromatic dispersion characteristics, and to draw the ion beam with low-angle divergence characteristics.

4) By adjusting the electron impact energy, a high proportion of impurity ions can be generated, reducing the space charge effect.

5) The characteristic of molecular ion that is generally dissociated in glow discharge is retained. For example, using phosphine (PH ₃ ) When feeding gas, most of them can be ionized to PH3 ⁺ Ions (up to 50% extraction flow). In one case of force, decaborane (B ₁₀ H ₄ ) Ionization reached B ₁₀ H _x⁺ The proportion is even higher (>70%). This feature is critical for implanting boron ions under low energy conditions (<1keV), and it also has a significant impact on increasing the dose rate of boron implantation. Plasma ion sources such as enhanced Bernard ion sources, etc. The dissociation caused by the plasma effect and increased outer wall temperature leads to the loss of decaborane molecules, making the enhanced Bernard ion source unable to produce decaborane ions .

6) No magnetic field is required.

7) The low-energy electrons in the beam plasma form a high degree of space charge compensation, so that the high-frequency noise in the original glow discharge source disappears.

8) The ion extraction aperture size is within a measurable range and is distributed between 12mm and 300mm or larger. Therefore, a larger extraction flow can be generated, and it is more in line with the design of a new generation of ion implanter. In fact, the design feature of this ion implanter is not achieved by the previous ion source design.

Advantages of acceleration and deceleration ion implantation

The present invention provides a technology that combines the acceleration and deceleration ion implantation with a variety of molecular implantation ion species, which can greatly enhance the low-energy beam flow and help produce higher-quality low-energy ion implantation and higher productivity. With this combination, the aforementioned long-standing non-uniformity and dosimetry problems in the acceleration and deceleration ion implantation system are solved.

In the molecular ion implantation system of the present invention, the ion beam is formed by a compound containing a variety of required impurity atoms (for example, B, As, P, Sb, or In) to produce more than traditional monomers (such as monoatomic atoms). ) Ion implantation has a higher dose rate and shallower depth implantation. For example, low-energy boron planting can be improved to the original planting monomer B ⁺ The conditions of ion current intensity I and energy E are changed to the new method of implanting decaborane molecular ion B ₁₀ H _x⁺ The ion current intensity is 0.10xI and the energy is 10xE. Extending this example, we can add 5keV, 1mAB10H _x⁺ The ion beam can achieve the same effect of 500eV, 10mAB ⁴ Ion beam. To replace. The results of the implantation depth and the impurity concentration (dose) obtained by the two methods are equivalent, but the decaborane implantation method has more advantages. Due to the migration energy of decaborane ions (mass x velocity ² ) Is ten times that of the same dose of boron ions, and the resulting ion current is one-tenth of the boron ion current. Compared with the single boron implantation, the space charge force caused by beam amplification and beam dissipation is smaller. As mentioned above, the purpose of this method is to increase the effective boron dose rate of the traditional (ie, non-deceleration) ion implanter. On the other hand, another contribution of this method is to develop other more advantages for the application of acceleration and deceleration ion implantation of specific molecular (plex) ions.

According to an embodiment of the present invention, the acceleration/deceleration ion implanter uses decaborane ions to obtain a higher effective boron dose rate than the conventional method, and prevents the single boron ion from decelerating the wafer before implantation. The problem of shrinking beam distribution. According to the present invention, other molecular impurities can also achieve the same effect. The features that contributed to the improvement are detailed in the following paragraphs.

We understand that the space charge effect will limit the acquisition of the beam flow during the initial ion extraction phase and the beam migration phase of the traditional (ie, non-deceleration) ion beam implanter. In the ion extraction stage, the improved characteristics achieved by molecular implantation can be achieved by analyzing the Child-Langmuir limit, which is the maximum space available by the extraction optical system of the ion implanter The charge limits the ion current density, to prove it. Although this limitation is determined to some extent by the optical system design of the implanter, the following analysis can still effectively illustrate the above characteristics:

<maths><img file="TW511113B_D0001.tif" /></maths>

Where J _max Unit mA/cm ² , Q series ion charge state, A series ion mass unit _amu ,U is the unit of drawing voltage inkV, and d is the unit of gap width cm. In practice, this limitation can be achieved by the electrostatic extraction optical system used by the ion implanter. In the extended equation (1), is used to express the improvement degree of space charge limited to molecular implantation relative to monomer implantation, which can be expressed as:

<maths><img file="TW511113B_D0002.tif" /></maths>

A series dose rate (atomic _S / _S ) Relative degree of improvement, the improved condition is molecular implantation, compound mass m _n , Required impurity atom n, acceleration voltage U _n , The conditions before the improvement were single implantation, atomic weight m, and acceleration voltage U. In this comparison, U can be adjusted to the condition of the same depth as that of the single implant to the substrate. Equation (2) can be simplified as: (3) =n ² 。

Therefore, by replacing boron with decaborane in the ion extraction of the traditional (non-deceleration) system, the dose rate can be increased by 100 times.

I found that there is a certain degree of similarity between the deceleration period of the acceleration and deceleration ion implantation system and the operation of the extraction optical system during the extraction period; both use a short distance and strong focus zone. It is also found that equation (1) has an effective performance for comparing molecular ion and monomer ion during the deceleration period. Therefore, I learned that equation (3) can also be used to evaluate the deceleration period. This analysis model can be compared, despite the above-mentioned significant inhomogeneity and dosimetry problems, the traditional acceleration and deceleration planting machine can transmit 2mA of boron monomer to the wafer with the planting energy of 500eV; however, the acceleration and deceleration ion planting machine is used. Decaborane (B ₁₀ H _x^＋ ) Substituting boron monomer, through the technology disclosed in the above-mentioned reference patent related to the present invention, the same dosage rate in the improved method only needs to implant 0.2mA decaborane at 5keV. By reducing the sensitivity of space charge to a certain extent due to the deceleration period, the beam distribution caused by deceleration is not only reduced, but also the uniformity of the deployment of the deceleration system, angular integrity, and dosimetry are also greatly improved.

This new acceleration/molecular ion combination (acceleration/deceleration ion implantation, using molecular beam (plex) ions) can be used to increase the low-energy boron dose rate to an unprecedented level. For example, using 20keV, 3mA decaborane to decelerate decaborane ions to 5keV (4:1 deceleration) and planting at 500eV to achieve a dose rate of up to 30mA is unimaginable in the past! The high boron dose rate makes the mechanical throughput limit of high-dose implants such as PMOS ion source/drain up to 200 wafers per hour (200WPH), even under the 300mm-diameter substrate process (a reference value is 2mA traditional Boron can produce a wafer capacity of about 25WPH under the condition of a dose of 8E14). It will be mentioned later that this type of strong beam stream can also be used in the key process of ultra-shallow junction formation.

This acceleration and deceleration system can also be applied to dimer implantation. Using the high-throughput gasification and ionization technology disclosed in the above-mentioned reference patent related to the present invention, the ion beam containing dimers (previously regarded as inapplicable ion implantation source), and other impurity types can achieve the above effect. For example, AS ₂⁺ ,P ₂⁺ ,B ₂⁺ ,In ₂⁺ , Or Sb ₂⁺ The formation of canbe, as predicted by the deceleration period equation (3), this technology can produce 4 times the deceleration beam, which can improve the maximum dose rate in the same way as the aforementioned method, reduce the non-uniformity, and solve the dosimetry decaborane implantation . The following table 1a is a dimer implant material suitable for the present invention.

<tables><img file="TW511113B_D0003.tif" /></tables>

According to the present system and method system, the dimer compound vaporizes at a temperature lower than its melting point, and the vapor is ionized by the impact action of a wide electron beam to transport the gas.

The disclosed decaborane acceleration and deceleration implantation system opens up a new process for semiconductor manufacturing. Another feature provided by the present invention is that by using the above-mentioned combined deceleration/molecular ion method and system process, many expensive steps in the implantation process are reduced, or the cost is reduced, or the quality of the implantation is improved.

For example, this system can be used for rapid transient diffusion (Transient enhanced diffusion, TED) improvement. In CMOS manufacturing, when producing ultra-shallow pn junctions, special attention should be paid to the manufacturing steps of the PMOS ion source/drain (S/D) structure. Boron is the only P-type impurity with high enough solid solubility and necessary electrical conductivity to form an S/D structure; however, during the intensification cycle that the wafer manufacturing process must go through, boron diffuses quickly on the silicon substrate. This irregular boron diffusion, called rapid transient diffusion (TED), is a limiting factor, especially for pn junction steepness. In the implantation process, it is generally believed that TED can be adjusted by destroying the silicon structure (a way of negative enhancement).

In the formation of the front side, the manufacturer of the ultra-shallow semiconductor chip equipment hopes to achieve the use of low-energy (Sub-keV) boron implantation to form an extremely shallow type-implanted boron distribution, so that it will be ordered by TED Intensified distribution. In order to reduce TED, a low temperature annealing tip (used for rapid thermal annealing or RTA) together with Sub-keV implant can be used to make shallower pn junctions. Since TED is the key to determining the distribution after being activated at low temperature, some people have recently thought that 500eV may be the lowest effective energy boron implantation of boron implantation in order to achieve the minimization of the pn depth junction of the activation. However, I think this assumption lacks corroboration in manufacturing practice. As the depth of boron implantation decreases, the TED effect also decreases linearly. This TED "shallowing" effect comes from the exposed silicon surface acting as a "pool" or a getter, which is used to cause TED defects, so that the implantation can be shallower and the TED effect is reduced.

Since the wafer production capacity is 500eV boron implantation, even the deceleration ion implantation machine, its performance is far below the limit of the machine itself. Since reducing the implantation energy to below 500eV will result in a sharp drop in wafer production capacity, traditional boron implantation in manufacturing will not use conditions below 500eV. Under the premise of building a new fab with a 300mm process and investing a large amount of money in the purchase of equipment, it is even more unable to withstand the impact of reduced production capacity brought about by low temperatures.

However, by using new systems and methods to solve the TED problem with a "shallow" approach, the ultra-shallow planting capacity required by the business can be achieved. In addition, it can also be extended to the new product protection plan related to the density and performance of the treatment and planting, and the volume can be further reduced.

The disclosed acceleration and deceleration system can also be applied to the advanced-amorphization level. In order to help limit the depth of the as-implanted boron profile (as-implanted boron profile), the advance-amorphization (crystalline Lattice destruction) implantation to limit the occurrence of channeling and thus increase the class-plantation depth distribution. Amorphization is accomplished through high-volume germanium or silicon beam implantation. Not only the cost is very high, it also complicates the manufacture of ultra-shallow pn junctions.

Through the application of the aforementioned acceleration and deceleration equipment and cluster molecular impurities, I realize that this problem can be improved by the present invention. In other words, the present invention provides not only the boron dose rate, shallower implantation and performance improvement, but also the damage characteristics of the molecular implantation system can also be used to reduce the expensive Ge or Si pre-amorphous implantation step. need. High-density ion clusters such as decaborane can cause damage to the crystal structure due to inelastic impact when the silicon surface impacts. With a high dose rate (according to the present invention, for example, decaborane between .5 and 3 mA), the resulting damage distribution can reduce or replace the advance-amorphous implantation, reducing the cost of this expensive step in the manufacturing process.

Therefore, the acceleration and deceleration system uses boron implantation to achieve high wafer productivity. For example, decaborane ions can be used in 200mm and 300mm substrate manufacturing processes, and the energy usage can be as low as 100eV. Since the TED effect is further suppressed at very low implantation energy, a pn junction that has never been shallower can be manufactured at very low temperatures.

By electron-impacting ionized molecules to generate molecular ions, heat-sensitive ion source materials, especially solid decaborane and the dimers mentioned above, can be used. By using a wide electron beam guide and a greatly extended extraction aperture parallel to the direction, while using a telephoto optical system, the beam size of the beam line entering the analyzer can be reduced.

The results of the present invention are: 1) High-efficiency boron implantation wafer production capacity, for 200mm and 300mm wafers, the decaborane ion implantation energy is between 100ev and 1keV; 2) By using a high dose rate of ten Borane (e.g. 2x10 per second ¹⁵ Up to 2x10 ¹⁶ + Borane ion), the crystalline structure is damaged without the expensive pre-amorphization implantation step; 3) With ultra-low implantation energy (decaborane energy between 1keV and 5keV, It is equivalent to the boron energy between 100eV and 500eV), ultra-shallow junction, which reduces the broadening of TED's activated boron distribution. 4) By using other cluster molecules, including new-type dimer materials, it can also achieve the same good results with other types of implants.

Therefore, there are fewer steps, greatly reduced costs, shallower pn junctions, higher density, and better performance than conventional ones.

Advantages of electron-beam ion source for ion doping

The present invention can replace the above-mentioned cylindrical ion source in an ion doping system. The advantages of the ion doping system disclosed in the present invention are as follows:

(1) The single grounding area is small-the electron-beam ion source is a small volume ion source that can be extended to the length of the required beam in one direction.

(2) Cost reduction-based on its small size and expandability characteristics, the present invention is simpler and more economical than conventional ones.

(3) High efficiency-due to the smaller volume and surface area, less ions are lost to the outer wall of the ionization volume than conventionally required ions.

(4) Improved process control-it can produce a high proportion of required ions, and can reduce precipitation (the ions required to generate the ion beam at the beginning are reduced accordingly). As the production efficiency of the required impurity ions is improved, the implantation process is better. This is because when the electron-beam ion source generates the main part of the ions with the required ions, the implantation distribution and the accuracy of the dose are significantly improved compared to the conventional ones.

(5) Increase in productivity. The present invention is based on the fact that a larger impurity ion current can be generated than in the prior art, so the productivity can be increased.

(6) Soft ionization-The present invention provides effective molecular ion production such as decaborane, which has significant advantages in terms of productivity and efficiency, for example, in the doping application of diborane ions.

The small footprint of the single-unit grounding area and simplified design make the equipment cost lower, the required building area is smaller, and the productivity is higher. At the same time, the present invention enables the single-ion doping system to be equipped with two ion sources, one for p-type Impurities, the other is for n-type impurities. Using a simple double-slit optical system, the ion doping system can be switched between two ion sources to process multiple substrates at the same time. It not only saves the cost of two sets of complicated design equipment, but also reduces the required floor area by half, and also reduces the production risk of the conventional ion doping system.

Schematic description

Figure 1 is a schematic diagram of a conventional ion source for ion implantation.

Figure 2 is a partial enlarged schematic diagram of the ion source based on the ion implantation in Figure 1.

Figure 3 is a perspective view of the ion implantation source according to the present invention. This figure is a cross-sectional view from the center of the ion implantation source to show the internal structure.

Figure 4 is an enlarged perspective view of the ionization chamber in the ion source based on the ion implantation of Figure 3.

Fig. 4a is a schematic diagram of the electron optical system in the ion implantation source used in the preferred embodiment shown in Fig. 4 according to Fig. 3.

Figure 5 is a schematic diagram of the power supply bias diagram. The power is used to provide the ion source for the ion implantation shown in Figure 3.

Fig. 6a and Fig. 6b are schematic diagrams of the side section and the top section of the ionization chamber in the ion source for ion implantation according to the present invention.

Figures 7a and 7b are perspective views and horizontal cross-sectional views of a device that can improve the electron beam focal length of the ionization chamber in Figure 6.

Figure 7c is a schematic diagram showing the spatial geometric relationship of the device in Figures 7a and 7b.

Fig. 8 shows a perspective view, in which the ion implantation source similar to that in Fig. 3 integrates the device in Fig. 7a and Fig. 7b into the ion implantation source.

Figure 9 is a schematic horizontal cross-sectional view of another example of the ionization chamber according to the present invention.

Figure 10 shows a cross-sectional view of the ionization chamber in Figure 9. The ion implantation source in Figure 3 will be integrated into the ion implantation source.

Figure 11 is a schematic diagram of ion implantation.

Figure 12 is a schematic diagram of the ion beam emitted by the ion source.

Figure 13 is a schematic diagram of ion implantation, which forms a drain to expand the gate adjacent to a target substrate.

Figure 14a is a schematic diagram. The gate side and drain extension layer in the figure are formed by implanting boron ions on a phosphorus-doped silicon substrate. Figure 14b shows the boron along AA in Figure 14a. Logarithmic graph with phosphorus ion concentration.

Figure 15 is a graph showing the relationship between the ion implantation energy of the ion implantation and the prediction of lateral scatter at different ion target incident angles with a function curve.

Figures 16a and 16b are schematic diagrams respectively showing the lateral scatter prediction of ion implantation at normal incidence and a non-parallel incidence angle of 7 degrees.

Fig. 17 is a side view of the device similar to that of Fig. 4, showing another ion implantation source according to the present invention to generate an extended band beam.

Figure 18a is a perspective view of a dual ion source system according to the present invention, which is used in a single ion doping device to implant n-type impurities and p-type impurities.

Figure 18b is a perspective view of the ion source shown in Figure 18a implanting ions on a panel substrate.

Figure 19 is a perspective view of a doping device for a panel display, including the design shown in Figure 18a.

Figure 19a is a diagram showing the crack surface of decaborane.

Figure 20 is a side cross-sectional view of an acceleration and deceleration ion implantation system. The acceleration and deceleration ion implantation system has a fixed beam line, and the wafer is located on a turntable.

Figure 21 is a side view of an ion source suitable for decaborane, etc. The ion source is modified and incorporated into the casing of the traditional acceleration and deceleration ion implantation system shown in Figure 20.

Figure 22 is a side view of another ion source used for decaborane, etc., which is characterized in that the extraction aperture can be extended to generate an initial ion beam with an extended cross-section;

Figure 23 is a vertical cross-sectional view showing an ion source that can use magnetic force to restrict electron beams from passing through an ionization chamber;

Figure 24 is a schematic diagram of ion optics of the extended section of the ionization chamber that draws the ion beam.

Fig. 25 is a perspective view of the lens and the lens holder according to the present invention in combination with an unexplained ratio.

Fig. 26 is an unexplained scale cross-sectional perspective view of the electron gun according to the present invention of Fig. 27.

Figure 27 is an unsealed perspective view of the electron gun according to the present invention.

Symbol description of main components

a, 35, 548. . . source of ion

c. . . Impurity gas feed

e,f. . . Double gasifier

h. . . cathode

j. . . Rejection electrode

k, 1, 41,526. . . Vaporizer feed pipe

m, n. . . Pressurizer

D. . . Strong discharge ion current

b, 36. . . Fixed flange

d. . . Glow discharge chamber

g. . . Aluminum zone

I. . . Uniform magnetic field

P. . . Discharge power

B. . . magnetic field

1. . . Silicon wafer

2. . . Ion beam

4. . . Non-parallel ion

5. . . Depth distribution

6. . . Launch range

7. . . Scatter

8. . . emission

9. . . source of ion

10. 630, 640, 660. . . beam

11. . . Divergence

12. . . Gate oxide

13. . . Gate electrode

14. . . Gate side

15. . . Drain extension

16. . . Gate electrode

17. . . AA

18. . . Boron concentration

19. . . Logarithmic ratio

20n. . . Type doping concentration

twenty one. . . Side of junction

26. . . Water cooling channel

27. . . Gas cooling channel

28. . . External vaporizer

28a. . . Chassis device

29. . . Feed in material

30. . . Vaporizer body

30a. . . Vaporizer combination

31. . . crucible

32. . . Conduction channel

34. . . Crucible-vaporizer body interface

34a. . . Pot sieve

34. . . b outer cover

39. . . the inside diameter of

42. . . Electron gun

44, 175, 175,, 240. . . Ionization chamber

45, 45', 174, 174", 544... Aperture

46. . . Ion extraction aperture narrow hole

47. . . Electron beam exit aperture

50. . . Gasification material

70, 250, 536'. . . Light stop

74. . . Narrow hole

80, 260. . . Ion extraction aperture

90. . . Magnet coil

100, 110. . . Heating gate valve

125. . . Electron beam

130. . . Water-cooled retainer

143 143'. . . cathode

144. . . Extractor

145. . . Beam shaping electrode

145', 147', 153, 152, 150', 152'. . . lens

147. . . First cathode

149. . . Focus electrode

150 150'. . . Second cathode

170 170'. . . Long filament

171. . . Filament core

172. . . DC power supply

173. . . With electron beam

176. . . Ion extraction aperture length

177. . . Ion extraction aperture plate

178. . . Beam shaping electrode

179. . . Grid electrode

190. . . Vacuum conduction

200. . . Three-stage vacuum tube

210. . . Ring electron beam

220. . . High temperature vaporizer

225. . . Steam pipe

235. . . Electron beam

270. . . With ion beam

300. . . Lens assembly

310. . . Holder

320. . . Cathode combination

330. . . First cathode

340. . . Focus electrode

350. . . Exit lens

360. . . shell

360a. . . Bottom edge

370. . . Holder

380. . . Feed tube

390. . . spring

395. . . Ceramic holder plate

400. . . Aluminum rod

410. . . Aluminum spacer

500. . . Lengthened ionization chamber

510. . . Lengthened ion extraction aperture

516'. . . Ionization volume

520. . . Extraction lens

528. . . Low temperature vaporizer

530. . . Beam path

540, 543. . . Analyzer magnet

545. . . Turntable

553, 557. . . electrode

555. . . Substrate

587. . . Mirror

600. . . Dual ion source system

610. . . Double-slit extraction optical system

650, 690. . . panel

670. . . Scan Station

680. . . direction

700. . . Vacuum box

710. . . Function room

Examples:

The following are the terms and their definitions used in this specification.

Lateral kinetic energy (E _T ): The kinetic energy value toward the beam propagation direction, that is, the direction of the draw area. (E _T )=12m <img file="TW511113B_D0004.tif" />_T² , Where V _T It is the speed perpendicular to the beam direction.

Beam noise (N): The fluctuation of the beam flow density relative to the average beam flow ratio, which is generally greater than the frequency value of 100 Hz.

Emission (s): The total emission ε is the product of two types of emission, ε=ε _x ε _y , In a narrow aperture lens placed vertically, ε _x It is emitted in the horizontal direction (along the width of the narrow hole), ε _y It is launched in the vertical direction. Whether used in any lens structure, ε _x With ε _y They are all defined by the horizontal and vertical directions of beam propagation. Another emission factor e is defined as follows: ε _x =2KXα _x , Ε _y =2Kyα _y , Where K=(E ₀ /E) ^1/2 , Where the beam energy of the E system and the E ₀ =10keV; α _x With α _y They are the half-angle (half-angle s) of the beam divergence to the X- and y- directions; and _X And y are the beam sizes in the x- and y-directions, respectively, measured at the same z-position along the propagation direction as the emission variable, which includes at least 70% of the beam flow. The emission value e is expressed in mm-mrad or cm-deg.

Brightness (B): B system beam current I divided by total beam emission: B=Iε _x ε _y

Plasma is defined as a region including the ionization volume that has approximately the same electrical properties, because the ionization volume ion valence number density is approximately equal.

Ion implantation source

Please refer to the drawings. Figures 1 and 2 show a conventional ion source used in ion implantation. The enhanced Bernard ion source is generally suitable for ion implanters with strong ion current, high energy, and medium ion current. The ion source a is fixed to the vacuum system of the ion implanter by a mounting flange b (mounting flange b), and has a vacuum feed pipe for cooling water, thermocouple, impurity gas feed, N ₂ Cooling gas, and power supply. The impurity gas feed c feeds the gas into the glow discharge chamber d, where the gas has been ionized. It is also equipped with a double gasifier e, f, in which solid feed materials such as AS, Sb ₂ O ₃ , And P can be vaporized. The gasification furnace, gas feed and cooling pipelines are located in a mechanical aluminum zone g. The function of water cooling is to control the temperature of the vaporizer and the aluminum zone g to increase sharply. It operates between 100C and 800C, and it can also offset the radiant heat generated by the glow discharge chamber d when the ion source is applied. The glow discharge cell d is fixed on the aluminum zone g in a heat-insulating manner. The ion source is a glow discharge ion source, which maintains continuous glow discharge between the hot filament cathode h and the inner wall glow discharge chamber d. Since the glow discharge generally dissipates more than 300W, and the glow discharge chamber d only emits heat by radiation, the temperature of the glow discharge chamber can be as high as 800C during operation.

The ionization of the gas introduced into the glow discharge chamber d is discharged between the cathode h and the glow discharge chamber d, and is generated by the impact of the electron stream or the glow discharge method. In order to improve the ionization efficiency, as shown in Figure 2, an external magnet coil 90 is required to establish a uniform magnetic field i along the axis between a cathode h and an anti-cathode y to limit the glow discharge electrons. An anti-cathode or repelling electrode j (repeller electrode j) (relative to the cathode h in the glow discharge chamber d) has the same voltage as the cathode h, and is used to reflect the glow discharge electrons restricted by the magnetic field i back and forth to the cathode h and Between the anti-cathode j. The restricted electron orbit is spiral, forming a circular plasma column between the cathode h and the anti-cathode j. Figure 2 shows the possible orbit of an electron between the cathode and the anti-cathode. The orbit is also spiral due to the magnetic field B. The plasma density in the plasma column is extremely high, up to 10 ¹² The progression per cubic centimeter; the neutral and ionized parts in the plasma column are caused to continuously ionize through the action of electrical exchange, resulting in a high current density of extracted ions. The cathode h can be a hot filament or an indirect heating cathode, which emits thermal emission electrons when heated by an external power source. The voltage between the cathode and the counter-cathode is Vc, which is between 60V and 150V, but is lower than the voltage of the ionization chamber d. This method can achieve a strong discharge ion current D of about 10A. Once the glow discharge plasma is turned on, the plasma will form a sheath on the surface of the cathode h (the cathode h is in the glow discharge chamber so it will be in contact with the forming plasma). The sheath has a strong electric field that can effectively absorb the thermionic electron current in the glow discharge; therefore, the high-discharge ion current can be generated by this method.

The heat dissipation of the discharge power P in the glow discharge chamber is P=DVc, which is about several hundred watts. In addition to dissipating heat by the glow discharge, the hot cathode h also radiates heat energy to the outer wall of the glow discharge chamber d. Therefore, for the impurity plasma, the glow discharge chamber d provides an extremely high temperature environment. Moreover, compared with a low temperature environment, by increasing the gas pressure of the glow discharge chamber d and reducing the deposition of impurities on the hot outer wall, the ionization efficiency can be further improved.

If a solid source gasifier furnace e or f is used, the gasification material is fed into the glow discharge chamber d and the pressurizers m and n through the gasifier feed pipes k and 1. The function of the pressurizer is to diffuse the vaporized material into the glow discharge chamber d, and its temperature is approximately the same as that of the glow discharge chamber d. The heat capacity of the vaporizer into the glow discharge chamber prevents the vaporizer from being a source of heat to provide solid feed materials, but can be maintained at equal to or less than 100C. Therefore, only the solid impurity feed material vaporized at a temperature of >100C and decomposed at a temperature of >800C (the standard outer wall temperature of the Bernard ion source) can be vaporized and introduced into the discharge chamber by this method.

Fig. 3 is a preferred embodiment according to the present invention. The related section is also disclosed in the above-mentioned related patent application. Fig. 3 is a cross-sectional view cut along the central axis of the ion source, showing its internal components in detail. The external vaporizer 28 includes a vaporizer body 30 and a crucible 31, in which a solid ion source feed material 29 such as decaborane is located. The resistance heater is embedded in the vaporizer body 30 and is in close contact with the water cooling channel 26 and the convection gas cooling channel 27 and the vaporizer body 30 to provide the crucible 31 with a uniform operating temperature higher than room temperature. The pressurized gas enters the crucible-vaporizer body interface 34 through the gas feed pipe 41, which is the heat conduction between the crucible 31 and the temperature-controlled vaporizer body 30, and at the same time monitors the vaporizer shell temperature through a thermocouple. The vaporized decaborane or other vaporized materials 50 enter the ionization chamber 44 through the vaporizer outlet channel 39, the heating gate valves 100 and 110, and the conduction channel 32. The fixed flange 36 of the ion source and the ion source region 35 are also subject to temperature control to be close to or greater than the temperature of the vaporizer.

The ion source gas transmission system includes two conduits to feed two separate ion sources into the ionization chamber. The first pipe has a smaller diameter and is connected to a pressurized gas ion source, such as a gas column, by a low conduction path. The second conduit is connected from a high-conduction path to a vaporizer that vaporizes solid materials at a low temperature. Regardless of the ion source, the gas concentration in the ionization chamber of the gas transmission system must be maintained at several millitorr. The gasifier needs to maintain strict temperature control of the surface in contact with the solid material to maintain a stable gas concentration in the ionization chamber.

Referring to FIG. 3, the vaporizer assembly 30 a includes a heating and cooling body 30 and a movable crucible 34. The crucible can be opened by moving the bottom plate 28 behind the vaporizer.

Once the crucible is removed from the gasifier, the crucible can be recharged by moving the outer cover 34b attached to the end of the crucible with elastic plastic and opening the grate 34a containing the solid 29.

After the crucible is charged and installed back to the body, the bore 39 at the front end of the body is sealed with air. The inner diameter 39 is the outlet for the gasification gas. The mechanical connection between the crucible and the body needs to be very precise to ensure the uniformity of the crucible temperature. The gap must be filled with gas (cooling gas) to promote heat transfer between the two surfaces. The cooling gas fills the gap through the chassis device 28a.

The temperature control is carried out by the PID closed coil control of the resistance component embedded in the body. The body material has high temperature conductivity to maintain temperature uniformity. A small thermal leak is installed in the main body to control the temperature of the system through the external air passage. The air passage 27 flows through the body of the vaporizer and is covered by multiple outer plates (not shown in the figure). The air is introduced into the system channel and connected with the vaporizer chassis 28 to provide convective cooling. The air passes through the meter valve for flow control and then flows into the inlet, and then the air discharged from the combination finally flows into the casing exhaust pipe.

In addition to air cooling, liquid cooling is also used for the vaporizer body. The refrigerant loops around the body in a 1-meter long 6mm diameter inner diameter tube, and is fixed to the body port 26 by a device. Liquid cooling enables rapid cooling of the vaporizer assembly to provide short-term cooling and can replace solid species.

Referring to Fig. 4, both of them establish good heat conduction between the ionization chamber 44 and the zone 35 by the pressurized gas flowing through the conduit into the interface 36. Gas material, such as the working gas in the gas column such as PH ₃ , Can be introduced into the ionization chamber 44 via the gas feed pipe 33. Generally speaking, the pressure in the ionization chamber 44 is about 1x10 ^-3 Torr, while the outside of the ionization chamber 44 is about 1x10 ^-5 Torr. Referring to Figure 4, the electron beam 125 generated by the electron gun 42 enters the ionization chamber 44 through the aperture 45, parallel to the migration ionization chamber 44 and adjacent to the ion extraction aperture aperture 46, which is located on the ion extraction aperture plate 80, The electron beam 125 passes through the electron beam exit aperture 47 of the ionization chamber 44 and stops at the stopper 70. The light stop 70 is water-cooled by a water-cooled holder 130, which has deionized water with a resistance (>10MΩ-cm). The light stop is electrically insulated by an insulating bracket 56 so that the beam flow intercepted by the light stop 70 can be controlled by the external HV water feed pipe 170 (as shown in FIG. 3). The ion extraction aperture 80 is insulated from the ionization chamber 44 by an electrically insulating, thermally conductive gasket, and has a biased negative voltage relative to the ionization chamber 44. This deflection voltage of the ion extraction aperture 80 establishes a drift electric field to attract ions to the aperture 80, and provides a deflection velocity for other thermionic ions, so that the deflection electric field generates a larger current of extraction ions. The general structure of the ion source is as follows: a 7.5mm diameter circular electron entrance aperture 45, a 10mm diameter electron beam exit aperture 47, a 25mm diameter electron gun assembly 42 with a length of 65mm, and a 67mm high ionization chamber 44. The cut surface 48 of the gun housing 142 allows the part of the electron gun assembly 42 containing the cathode to be exposed to the vacuum environment of the ion source housing to prolong the use time of the cathode 143.

The electron gun optical system includes a cathode 143, a beam shaping electrode 145, a first cathode 147, a focusing electrode 149, a second cathode 150, and an exit lens 152. This lens system draws and restricts the flow of space charge electrons. The downstream 4 component lenses include the first cathode 147, the focusing electrode 149, the second cathode 150, and the exit lens 152 to collimate and accelerate the final energy required for the electron beam at the same time. Can be made of refractory metal or LaB ₆ The formed hot cathode 143 emits thermionic electrons, which can be performed by direct or indirect heating. Electrons are accelerated in the first cathode gap in a Pierce structure. The equipotential of the dielectric cathode 143 and the first cathode 147 can be achieved by the original cone beam-shaping electrode 145 and the first cathode. In this way, the output ion flow can be maximized by the space charge effect, and the extraction amount can reach 5 keV. In addition, the downstream optical system can decelerate the ion current to the final energy, which is between 70eV and 2000eV.

FIG. 4a shows another preferred embodiment of the electron optical system, in which the second cathode 150' and the exit lens 152 are formed by using a puls structure. The inversion of the Puls structure and the Puls extractor 144 are defined by lenses 145' and 147'. The lens 153 connected in series by the second cathode 150' and the lens 152' can be used as a decelerating lens to introduce a low energy (for example, 100eV or lower) to calibrate the electron beam into the ionization chamber 44. Applying the "reverse puls" structure to the deceleration lens 153 helps to limit the space charge electron beam, so that the higher-calibrated low-energy electron flow can be introduced into the ionization chamber 44 accordingly. For example, the pumped electron beam can pass through the lens 147' in the 1keV state, propagate through the lens 150' in the 500eV state, and then decelerate to 100eV through the lens 152', which can be stronger than the extractor 144 in the 100eV state of the beam Electronic ion flow.

The space charge force in the optical system of the electron gun and the part of the decelerating lens 153 can still be improved, and positive ions can be generated deliberately along the electron beam path. The positive space charge ions compensate the negative electron space charge and reduce the net Coulomb repulsion between the electron and the beam, thereby reducing the beam amplification and generating a stronger electron ion current. This is an effective method to reduce the space charge force: because ions have considerable weight, move slowly, and their loss rate is low, if the ratio of ion generation to ion loss is maintained, a certain charge balance can be achieved with. The ions can generate the pressure of the region with the ionizable gas at any point of the electron beam path. For example, consider the regional air pressure P1>P2>P3>P4 shown in Figure 4a (I found that by adjusting individual lens groups and the surrounding real space conduction, the regional pressure Pi can be controlled). The main gas type is decaborane (B ₁₀ H ₁₄ ), a large heavy molecule with an electron-impact ionization profile to generate positive ions. In the ionization chamber 44, the Pi pressure can reach up to 10 ^-3 Torr, so space charge compensation is very effective. This ensures that the electron beam can spread uniformly in the main ionization zone, the ion density near the ion extraction aperture 46 is uniform, and the extraction ion beam can have a uniform ion density. The ambient pressure of the P4 vacuum envelope ion source (usually 10- ⁵ TOrr or less), and because the gas in the ionization chamber 44 propagates through the electron beam into the aperture 45', the extreme value establishes a high pressure gradient. Because the deceleration lens 153 is close to the entrance aperture 45', P2 is relatively high, and the space charge repulsion is greatly reduced. The zone P3 is adjacent to the cathode 143', so P3 and P4 should be maintained at the same level, that is, the arrival rate of decaborane molecules to the cathode surface at a relatively low pressure is less than the desorption rate of the decaborane by-product precipitation on the cathode surface. This is true for low temperature cathodes such as L _a B ₆ The cathode of the field emitter is critical. Generally speaking, refractive metal cathodes can be operated at high temperatures, and fragmentation of precipitated gas will not cause problems.

Figure 5 shows a power supply bias diagram. The power supply includes voltages supplied to the electron gun assembly, ion source and specific instruments. The schema used in Figure 5 is defined as follows:

V _S (Source voltage): 0-40kV pos @ 100 mA. Set the ion beam energy and the bias voltage of the ionization chamber relative to the ground terminal.

V _C (Cathode voltage): 0-2kV neg @ 100 mA. Set the electron beam energy and the bias voltage of the cathode relative to the ionization chamber.

V _F (Filament voltage): 0-5V@50A. Provide heating ion current to directly or indirectly heat the cathode emitter.

V _E (Drained aperture voltage): 0-20Vneg@100mA. The bias of the ion extraction aperture with respect to the ionization chamber.

V _l (Cathode shield voltage): 0-50Vneg@10mA. The bias voltage of the beam-shaping electrode relative to the cathode.

V2 (cathode voltage): 0-5keVpos@50mA. The bias voltage of the first cathode relative to the cathode.

V3 (focus voltage): 0-5keVpos@10mA. The bias voltage of the focusing electrode relative to the cathode.

V4 (outlet lens voltage): 0-2keVpos@50mA. The bias voltage of the exit lens relative to the cathode determines the energy of the electron beam when it leaves the tetrode. The four-stage vacuum tube includes a cathode, a cathode, a focus, and an exit assembly.

V _D (Light stop voltage): 0-2kVpos@100mA. The light stop is biased with respect to the cathode.

M1: The current value of the electron ion leaving the electron gun.

M2: The ion current value emitted by the cathode.

M3: Reach the electron ion current value of the light stopper.

Another embodiment of the present invention is applicable to ion implantation systems with ion extraction slits of 1 to 3 inches long. This embodiment can generate a strong ion current (each impurity beam can generate 5 mA). In this design, a filament that is nearly the same length as the ion extraction slit provides a linear low-energy electron. The position of the filament is parallel to the ion extraction aperture as shown in Figure 6. The electron beam arrangement is the same as in the first embodiment. The characteristics of this structure are: 1) it can generate a strong electron ion current, 2) the electron orbital energy (injection energies) is low while still transmitting a strong electron ion current. From point 2), it can be deduced that since the ionization profile is approximately 100 eV, which is much larger than 5 times that of 2 keV, a strong ion current can be obtained. 3) Finally, since the electrons are drawn in a linear beam, there is no need to use the constraints of magnetic field manufacturing to limit the divergence of the electron beam, which greatly reduces the space charge effect.

Figure 6 shows the simple design of the ion source, where the electrons are injected into the ionization chamber in the same direction as the drawn ion beam. The long filament 170 is heated by the filament core 171 and the DC power source 172 to emit electrons 173 along the length of the filament. The filament can be a ribbon or a thick tungsten wire. The filament 170 should have a lower voltage than the ionization chamber 175. The electrons are accelerated by a square power supply 172 and enter a narrow hole 174 in the center of the ionization chamber 175. Form a first and second vacuum tube design. The top view of this structure is the projection of Figure 6b. The extended electron beams ionize the gas in the ionization chamber 175; the ions are extracted through the ion extraction aperture 176 in the ion extraction aperture plate 177. This design has the advantages of simplicity. The strong electron ion current shown in Fig. 6a can also be generated by uniformly focusing the long filament 170 along the ion extraction aperture 176. The ion beam thus generated is uniform, and the length of the gas path since the electron path through the ionization chamber 175 is equivalent to the length of the ion extraction aperture 176. In addition, since the electron beam is lengthened toward the vertical angle, it is less likely to cause space charge amplification. Compared with the small and circular electron beam, the electron beam produced by this method can produce a stronger total electron ion current transfer at a given energy. To the ionization chamber 175.

In order to improve performance, a grid electrode 179 with a rectangular narrow hole can be embedded between the filament 170 (shown in Figure 7) and the chamber entrance aperture 174 to improve the focusing of the electron beam, thereby forming a three-stage vacuum tube structure. In order to prevent the transition metal contamination of the ionization chamber from entering the aperture 174 due to the evaporation of the filament and eventually causing tungsten or rhenium to enter the ionization chamber 175, the filament can be moved farther away from the ionization chamber to remove easily vaporized constituent materials.

This embodiment is illustrated in Figure 7a. The filament 170 is located far from the ionization chamber 175, and the electron beam is transmitted through a lens with a series of long and square apertures. Figure 7a is a schematic diagram of a three-stage vacuum tube, in which filament 170, beam shaping electrode 178, grid electrode 179, and ionization chamber entrance aperture 174 all have different voltages, but this design is not only limited to three-stage vacuum tubes; a lens is added The number can be adjusted. The horizontal cross-sectional view shows the details of the electron optical system in Fig. 7a, and Fig. 7b shows the detailed electron beam 173 propagation. Although in order to produce the ion optimization required for the implantation process, the lens system V _C By V _S The applicable voltage range is very wide, the general electrode voltage value is: cathode voltage V _C =-100V, beam shaping electrode V ₁ =-102V, grid voltage V ₂ =+100V, source voltage V _S =0 (all volts are expressed relative to the ionization chamber or source voltage), and the ion extraction electrode voltage V _E =-5V. V _E The bias voltage applied to the ion extraction aperture 177 forms a normal drift electric field E, as shown in FIG. 7b. E gives the drift velocity in the ionization chamber to attract positive ions to deviate to the ion extraction aperture plate 177, which can effectively draw an ion beam by the external extraction field. Since ions are generated along the electron beam 173, E imparts kinetic energy to the ions. Figure 8 shows an embodiment of the three-stage vacuum tube embedded filament shown in Figures 7a and 7b applied to an ion source similar to Figure 3. Figure 8 is a detailed description of the ionization chamber 175 and the three-stage vacuum tube 200, where the three-stage vacuum tube is located in the ion source region 35. Other components in Figure 3, such as the low-temperature vaporizer 28, gate valves 100 and 110, fixed flange 36, and gas feed pipe 33 are also included in this embodiment, but not shown in the figure. The electron gun 42, the magnet coil 90, and the light stop 70 are not included in the embodiment in FIG. 8 because the functions of these electron gun components have been replaced by the three-stage vacuum tube 200. The advantages of the embodiment in Figure 8 are as follows: 1) The filament is located far away from the low-pressure location (directly exposed to the surrounding vacuum in the vacuum envelope of the ion source through the vacuum conduction 190), which can increase the filament life; 2) The remote filament can reduce ionization The degree of contamination of the chamber by the filament material; 3) The lens system helps to accelerate and decelerate the movement of the electron beam, so that the ionization chamber produces a stronger electron ion current. 4) The radiant heat generated by the filament 170 can be dissipated by the aluminum region 35 of the water-cooled ion source. 5) The electron beam can be focused on the ion extraction aperture, resulting in high ion extraction efficiency and small lateral momentum. This is because the original electron trajectory moves along the drift electric field E in Figure 7b, so when the electron strikes the ion when it reaches the ion extraction aperture, its basic thermal lateral component kinetic energy will be &lt;&lt;1eV. 6) The embodiment of Fig. 7b and Fig. 8 can generate ion source with low emission and high brightness, improve the ion beam propagation control of implanter, and generate better ion beam space and angle on the wafer substrate than the conventional one Bit uniformity.

Figure 7c shows a four-stage vacuum tube whose structure is similar to that of the three-stage vacuum tube in Figure 7b. The figure also shows its specification information in mm. The four-stage vacuum tube has a focal length adjustment function, so the focusing characteristic of the lens system has nothing to do with the final electron energy. This feature enables the electrons emitted by the refracting filament 170' to be absorbed at a higher energy than when the electrons enter the ionization chamber 175', that is, the deceleration technology is used to transfer a higher-energy electron ion flow that limits the space charge to enter the ionization zone. The representative lens volts are shown in Table A. The setting conditions are that the objective lens distance is 4 times the lens aperture diameter ("D") and the imaging distance is 6D outside the reference plane of the four-stage vacuum tube. At the energy level of 300eV to 100eV, by drawing electrons, the lens is adjusted to inject 100eV electrons into the ionization zone. This adjustment assumes that the electrons are still 100eV when they enter the ionization chamber. Due to the linear nature of the lens, it does not have the characteristic of focusing or confining the beam to the length of the aperture (y-direction). Under strong ion current density, the electron beam expands along the y-axis due to the repulsive force of space charge, which causes the beam to become a virtual image and is lost. Space charge repulsion and beam dispersion can be reduced by generating positive ions through electron-impact ionization along the beam path. The positive space charge ions can compensate for the negative space charge electrons, reduce the net Coulomb repulsion between the electron and the beam, and make the beam magnify, reduce and enhance the electron ion current. In order to prevent the ions drawn by the ionization region 181' from being lost in the four-stage vacuum tube, the lens assembly 180' needs to be kept at a constant relative to V ₀ The voltage of the ionization chamber is sufficiently positive.

Since ions have considerable weight and move slowly, their loss rate is low. If the ratio of ion generation to ion loss is maintained, a certain charge balance can be achieved with and. The ions can generate the pressure of the region with the ionizable gas at any point of the electron beam path. For example, consider the regional air pressure P1>P2>P3>P4 shown in Figure 7c (I found that by adjusting individual lens groups and the surrounding real space conduction, the regional pressure Pi can be controlled). The main gas type is decaborane (B ₁₀ H ₁₄ ), a large heavy molecule with electron-impact ionization profile to generate positive ions. In the ionization chamber 175', the Pi pressure can reach up to 10 ^-3 Torr, so space charge compensation is very effective. This ensures that the electron beam can spread uniformly in the main ionization zone, the ion density near the ion extraction aperture 176' has good uniformity, and the extraction ion beam can have a uniform ion density. The ambient pressure of the P4 vacuum envelope ion source (usually 10 ^-5 Torr or smaller), and because the gas in the ionization chamber 175' travels through the electron beam into the aperture 174", the extreme value establishes a high pressure gradient. Because the decelerating lens 153 is close to 45', P2 is relatively high, and the space charge The repulsive force is thus greatly reduced. Zone P3 is adjacent to the cathode 143', so P3 and P4 should be maintained at the same level, that is, the arrival rate of decaborane molecules to the cathode surface at a relatively low pressure is less than the desorption rate of decaborane by-product precipitation on the cathode surface .This is true for low temperature cathodes such as LaB ₆ The cathode of the field emitter is critical. Generally speaking, refracting metal cathodes can be operated at high temperatures, and fragmentation of precipitated gas will not cause problems.

In another embodiment. Figure 9 shows a top view of a hybrid ion source in which the design and operating features of the first and second embodiments are integrated. It is similar to Figure 7, but this figure shows the intersection of the annular electron beam 210 (located in the plane shown in the figure) and the electron beam 173 emitted by the long filament in the three-stage vacuum tube 200. Therefore, the third embodiment has both a long filament emitter on the main axis and a longitudinal electron beam. Figure 10 is a detailed description of this embodiment, and the ion source is the improved version shown in Figure 3. Not only that, in addition to the low-temperature vaporizer 28 shown in FIG. 3, the ion source in FIG. 10 further includes a high-temperature vaporizer 220 surrounded by the ion source region 35. The vaporizer 210 is located in the ion source region 35 and has no influence on the operation of the vapor conduit 32 or the gas feed pipe 33 in FIG. 3. As shown in FIG. 10, the vapor is introduced into the ionization chamber 175 from the vaporizer 220 through the vapor conduit 225. The purpose of setting the second vaporizer 220 further includes introducing solid impurity compounds such as P, As, Sb, and Sb ₂ O ₃ With InCl ₃ Of steam. Therefore, the steam of solid impurities and the steam of particularly low temperature materials such as B ₁ 0H ₁₄ Trimethyl indium (TMI) can be introduced into the ionization chamber 44 as shown in the embodiment in Figure 10.

The advantages and features of the embodiment in Figures 9 and 10 are: 1) it can generate a strong ion beam; 2) it is matched with the strong electron density beside the extraction aperture, and ions are generated along the path of the drift electric field, so no magnetic field is needed. , And can produce a relatively low emission ion. 3) The solid and gas feeding materials can be used complementarily. The use of two vaporizers allows n-type and P-type impurities to operate simultaneously without using different implanted ion sources; 4) A low-energy electron beam is generated by a filament, together with The high-energy electron beam generated by the electron gun enables ion species of multiple valences to be generated through the same process, among which monovalent species can be generated by low-energy beams, and ion species of multiple valences can be generated by high-energy beam ionization. 5) It has the function of "adjusting" the electron beam energy of the electron source. This function allows the ion source in Figure 9 to flexibly generate different molecules to feed into the crack surface, so that the ion beam composition can be adjusted for specific implantation requirements.

Drain expansion

The technology disclosed in the present invention for the expansion and deployment of the drain electrode of the transistor has its technical characteristics that can greatly improve the performance of the device. The basic concept of the disclosed ion source is to provide a low-emission beam and generate a junction with a smaller lateral steepness, so that the efficiency of the transistor can be improved to meet the needs of the reduction of the transistor.

Having a relatively steep lateral junction of the transistor requires better performance in many aspects. First, the part of the junction area can become a series resistance, and the series resistance formed by the steep junction is smaller, which is beneficial to the driving of the ion current and the conductance of the transistor. The steep junction also makes the conduction secondary threshold (secondary threshold) lower, which is very important when the current drops. This feature has the advantage of reducing the ion current in the power-off state, which can further reduce the static ion current of the entire circuit and prolong the battery life. In addition to the sub-threshold, other advantages of this design are the ability to reduce the open-end voltage without increasing the static ion current. The circuit performance (which can be improved by lowering the open-end voltage) and the backup power (which can be improved by lowering the static current) can be traded off. These are particularly critical when the voltage supply is reduced, especially in the process below 0.25um.

In the ion implantation area, the ion beam is generated and moved to the target, and a part of the energy will make the ions penetrate the target material to a certain depth, as shown in Figure 11. The target can be a single silicon wafer 1 in which a transistor or other structure is formed. The ion beam 2 directed to the silicon wafer 1 can implant ions into the silicon wafer to form the required device. Although the ion paths are different, there is still a distribution that can follow the depth distribution 5 as shown in Figure 1. The depth distribution can be discussed by two parameters: the emission range 6 is the average depth of penetration, and the straggle 7 is a measure of the change in the depth of the atom distribution. The two parameters are highly related to the ion beam conditions in the implantation process, and the distribution of heavier ions or lower energy is shallower. The general ion implantation process analysis can only be discussed on the parallel ion beam 2, which can be calculated through the expected distribution, emission range, and scatter. But in fact, the ion beams are not completely parallel. A considerable part of the ion beams are actually non-parallel, which is referred to as non-parallel ion 4 here. The non-parallel ions 4 hit the silicon substrate at a non-zero angle and the incident angle is 3. Generally speaking, all ion beams have non-parallel parts, and the degree depends on the beam conditions, the status of the implantation equipment, and the adjustment of the implantation machine. The ion beams with different incident angles are discussed below.

The non-parallel ion beam components have a series of several characteristics. First, when the beam 10 is generated and leaves the ion source 9, the angular distribution of the beam 10 can be represented by measuring the emission 8, as shown in Figure 12. This parameter can be used to express the total angular position distribution of the beam when it is drawn by the ion source, expressed in actual angles. Once the ion beam moves along the beamline, the discussion of angular position distribution indicates that the measurement becomes divergence11. Divergence l1 represents the maximum angle of the beam relative to the beam axis. When the beam reaches the target, as described above, each ion generally has its incident angle, because the target is not in its linear position relative to the beam axis. Therefore, the beam divergence 11 produces an incident angle when the beam reaches the target. Note that these parameters are: the beam ion movement angle is different, and a distribution function can be used to describe the divergence angle of the beam density. Using these two points, the ion behavior with the greatest angle will be further discussed below.

The representative ion implantation process discussed here is drain expansion implantation. Referring to Figure 13, the purpose of this step is to form a transistor. The role of drain extension is very important because it determines the function and structure of the transistor. First of all, this indicates that the automatic alignment is completed: the gate electrode 13/gate oxide 12 are stacked to form a gate side 14. After the ion beam 2 is launched, the drain extension is about to be formed. Where the gate stack has been removed, the through silicon is implanted and a implanted layer is formed as a drain extension 15. Where the gate stack is not removed, the penetration gate electrode 16 is implanted. In this process, the drain extension junction is aligned side by side with the gate electrode. This good transistor structure is called an automatic alignment gate. As shown in Figure 3, the following will further discuss the non-parallel ions 4 located under the gate and hitting silicon on the bottom side of the gate.

One of the important developments in semiconductor technology is the shrinking structure. By shrinking, more transistors can be placed in the same silicon area to reduce costs. As far as ion implantation is concerned, what needs to be improved is to reduce the ion energy, because the smaller specification also includes the reduction of ion depth, which can be achieved by reducing the implantation energy. Special attention should be paid to the evolution of new methods, equipment, and materials in semiconductor technology to meet the demand for volume reduction. Naturally, ion implantation must also cooperate. Reducing the size is a big challenge for ion implantation, which means that the junction must be able to achieve a shallower level in order to achieve the purpose of reduction. The boron atom itself is very light and the depth of the cloth implanted in silicon is a key issue in the formation of the P-type junction. In particular, the biggest challenge for P-type drain expansion is that the lowest energy boron beam is already used for its implantation. In addition to the original problem of the ion implanter that transmits a strong ion beam at low energy, the method used to improve the low-energy boron beam has a negative impact on the quality of the transmitted beam and causes the beam to diverge.

In addition to forming the drain extension, there are other steps required for ion implantation. Other steps include temperature treatment or annealing steps that can activate atoms to be implanted. This temperature treatment step needs to be carried out at a high temperature (>900C), at this time it may be affected by the diffusion effect. Diffusion is the part of the implanted ion implantation range that exceeds the distribution of the class-implantation depth, which occurs in the deeper part of the lateral substrate. The challenge is to form a shallow junction (in fact, it should be an ultra-shallow junction or USJ), and the diffusion effect needs to be suppressed. In addition, the diffusion time also needs to be controlled: the so-called TED effect. This effect comes from the silicon cracks in the implanted layer caused by the damage to the silicon crystals caused by the implantation during ion implantation. This results in a very short boron diffusion. However, a short annealing cycle can suppress the diffusion effect to a minimum, so when the TED effect increases, the annealing time should be shortened. We understand that diffusion and TED effects will cause the depth and lateral amplitude of the implanted layer to increase. Therefore, this critical step needs to be carried out with special methods and equipment to produce shallow junctions and increase lateral steepness.

The reduction of boron implantation energy will cause the following problems in boron implantation: weak ion current transmission leads to low productivity. Since a strong ion current is required at low ion energy, the high space charge density of the beam is harmful. When the beam is drawn from the ion source, the space charge density can draw the area and produce an E3/2 effect on the beam flow and energy. This effect greatly reduces the beam flow when the energy is reduced. Not only that, the space charge when the beam moves will be affected by the Coulomb force, pushing the ions to the side of the beam, causing "beam expansion" and beam dispersion as the beam moves along the beam line. This effect is highly energy-related, and it is also a difficult point in delivering low-energy boron (LEB) ion beams to silicon wafer targets.

There are two methods in the implantation industry to solve the LEB beam flow problem, but they will additionally cause the beam to diverge on the wafer. The first method is to minimize the travel distance of the beam line between the ion source and the silicon wafer light, so that a stronger beam can reach the wafer. Short beam lines can be achieved by a wider moving path, so that a stronger beam can be transmitted. In terms of beam divergence, this method will aggravate the problem of beam divergence on the wafer. The second method is to reduce the LEB ion current speed. First, the beam is drawn and moved onto the wafer with high energy, and then the beam speed is reduced to an appropriate energy before implanting on the wafer. This method also has the problem of high wafer divergence, and also brings about the problem of energy pollution of the beam on the wafer.

As shown in Figures 14a and 14b, the side of the junction is one of the most important loops in the depth distribution. The implanted boron layer interacts with the existing doping concentration to produce a P/N junction. With reference to the gate side and drain expansion layer shown in Figure 14a, the lateral cut plane (A-A17) is parallel to but lower than the silicon surface. In Figure 14b, the doping concentration is made along the AA section. In the drain extension region, the boron concentration 18 is extremely high, about 1E20cm-3, and it is strongly shown in the logarithmic ratio 19. Before the LEB implantation of drain extension, the n-type doping concentration 20 already exists. In this example, it is phosphorus, but it can also be other n-type impurity elements. At this time, the junction side 21 is formed where the two doping concentrations are equal. At this time, the concentration is very low and the n-type concentration is about E17-E18cm-3 strong. The boron concentration 18 is also close to this concentration at this time. It should be noted that the boron concentration on the side of the junction is much lower than the drain extension concentration by about 100X. The key point is that even if the beam component is as small as 1%, the side effect of the junction is extremely large, which will be discussed further below. Moreover, the lateral slope of the boron distribution on the side of the junction depends on the lateral steepness of the drain junction. This parameter is usually expressed in units of nm/dccade, or expressed in units of nm of the degree of lateral distribution, which needs to be increased by an order of magnitude when expressing the boron concentration. In the 0.18um process, the general drain lateral steepness is 10nm/dec, and the future demand is &lt;5nm/decade.

When the drain extension is generated, the low divergence beam according to the present invention can improve the lateral steepness of the drain extension and enhance the performance of the transistor. There are two promotion mechanisms: lateral penetration and reduction of lateral scatter.

First, the lateral penetration has a high degree of symmetry. Please refer to Figure 13, ideally a low divergence beam allows ions to enter under the side of the gate in the extended drain region. As shown in FIG. 13, the diverging beam has non-parallel ions. It should be noted that the last position of the non-parallel ions, that is, where the ions penetrate the silicon substrate at the gate side 14. This generally ends up at the mark X, which replaces the drain to extend the main distribution of boron atoms. An incident angle distribution of the astigmatic beam can form a lateral expansion of the distribution on the side of the gate. It can produce a lateral gradient junction with less steepness than that produced by a low divergence beam. As mentioned above, the concentration on the side of the junction is much lower than the maximum concentration of the drain extension, so only a very small ratio (1%) of the light beam will cause a high divergence state on the side of the junction. Moreover, in the angular distribution within the beam, the larger the angle, the lower the density, which will also contribute to the occurrence of lateral gradient junctions. To predict the improvement of the low divergence beam, the sine of the angle of incidence can be multiplied by the side depth of the vertical junction by the estimated lateral displacement. Assuming that the incident angle of the divergent beam is 7 degrees, the lateral displacement is 12% of the junction depth, and it is generally estimated to be 70% of the junction depth of the lateral junction. The resulting lateral junction is 82% of the vertical junction depth, and the junction of the divergent beam is 17% higher than that of the low divergence beam (due to this effect).

The second mechanism related to beam divergence and lateral junction formation is lateral scatter. The reason for the degree of scatter is that the silicon target is a crystalline lattice composed of individual atoms in a fixed pattern at a certain distance, which is inherently non-uniform. The incoming ions may directly hit the silicon atoms or hit the silicon atoms at an angle of incidence, or they may not hit the silicon atoms. These entering boron ion states can be represented by a distribution. The degree of scatter can be regarded as a variation of the vertical depth distribution. In this example, the state of the side can be related to the degree of lateral scatter. The important determinant of the degree of lateral scatter is the angle of incidence, which makes the divergent ions produce greater lateral scatter. For further discussion of this phenomenon, please refer to page 767 of IIT'98 by Nakagawa, Hada, and Thome. Figure 15 refers to this work and describes the lateral scatter as a function of the incident angle Φ and ion energy. Generally 70% of the scatter from lateral to vertical is represented by a dashed line, and an almost horizontal straight line is assumed to be an incident angle of 0 degrees. The data points in the figure are taken under different incident angles and energy conditions. When the ion incident angle is 7 degrees, the lateral scatter is exactly twice that of the traditional model. It should be noted that the larger the angle, the stronger the effect, and the more inclined the distribution is to the side, which is contrary to the ideal of maintaining a steep junction.

With reference to Figure 16a and Figure 16b, in order to determine the relative magnitude of the energy effect, a model is specifically established for this effect under similar conditions. The first step is to reduce the variable and calculate the approximate value: the vertical junction depth is regarded as the emission range (Rp) plus the vertical scatter (Rp) twice, which can be expressed as Xj=Rp+2Rp

In this way, the vertical scatter degree can be expressed by the above formula in terms of the lateral joint position.

In this model, the position of the side of the low-divergence lateral junction is twice the degree of lateral scatter. If we know that the lateral scatter degree is 0.7 times the vertical scatter degree, then the side position of the lateral junction is 1. 4Rp inside the gate side. Now add the divergent beam to the consideration of two conditions, these conditions will also affect the side of the lateral junction. The first condition is to set the lateral scatter to twice that of the normal incident situation, if so, the side position of the junction will be 2.8Rp. The second condition is the occurrence of the symmetry effect. It was previously calculated that its size should be 12% of the depth of the junction, so the side position of the junction is 0.48Rp. Adding the above results, the position of the side of the total junction should be 3.28Rp inside the side of the gate, or 2.3 times the extended lateral direction. Similar ratios can be used to improve the lateral steepness. This advantage is very useful for low-emission ion sources.

In the manufacture of MOS transistors, another key parameter is the channel length. The channel length represents the distance between the ion source and the drain, that is, the lateral distance between the ion source and the drain. The center of the previous article is concentrated on the drain of the transistor. At the same time as the drain is formed, there is an area on the other side of the gate that is also forming an ion source. Note that the lateral expansion of the ion source and drain determines the channel length. Planting parameters related to planting distribution mainly include type and energy, in addition to slope, twist, mask side, and beam divergence. The emission of the beam determines the divergence of the beam, which is therefore related to the formation of the impact transistor. Fewer ion beams penetrate underneath the gate stack, so the channel length will be longer than the low divergence beam. Generally speaking, a longer channel length is not beneficial, but when the channel length is the same as the gate length, the transistor manufactured under the condition of low divergence is quite ideal.

In this case, the problem with the strongly divergent beam is that the divergence patterns are not always consistent. The beam divergence condition will naturally cause this problem because the beam adjustment and setting are different. It comes from that part of the light beam will disappear when the light beam crosses the aperture. In addition, the part of the divergent periphery will also disappear, because the direct relationship between space and angle makes the light beam part disappear when it reaches the substrate. Assume two conditions: one of the beams is set to pass through the aperture, and the second of the apertures shields part of the beam. In the case of concentration, the divergence will also be concentrated as a result, so the beam and the wafer intersect about +/-5 degrees at the proper angle plus or minus the same divergence. In the second case, the side of the beam is shielded by the aperture, making the angular position larger, but it only occurs on one side. Therefore, the angle of reaching the substrate should be an angle and an asymmetric angle plus or minus the same divergence about +2/-5 degrees intersection. Now if the side with less divergence is the side that determines the side channel of the transistor, we can change the junction distribution by reducing the high-angle ions, so that the channel grows due to the weakening of the ability of the ions to penetrate under the side of the gate. This change in channel length is not necessary, and low-emission beams are not affected by this change. It should be noted that the changes discussed are all conditions that can be manipulated by modern general equipment. In today's manufacturing equipment, automated procedures can set various conditions for beam generation, which can be controlled through the aforementioned processes to optimize beam conditions. Although the beam conditions may not be the same every time; the purpose of the system is to produce a beam that meets the requirements, and there are many adjustment methods. These adjustment methods will produce different beam divergence. Unless there are specific requirements, the beam does not have to be a non-central beam or use unusual conditions to create deliberate divergence.

Another advantage of the low-emission ion beam is that it can be designed to make the channel length short and uniform, instead of being caused by incorrect setting of conditions. Since the circuit performance is directly related to the channel length, by designing the channel length to be short and uniform, the ideal of improving performance without changing other conditions can be realized.

Please note that if the channel is too short, it will cause failure, which may cause the drain to source voltage to fail to continue, or make the start voltage inoperable. Therefore, the channel length distribution in the process and circuit design should be considered in terms of acceptability, so that the transistor can operate normally (even the error of only 1ppm is too large). In other words, the smaller the distribution variation of the channel length, and the smaller the design channel length can be on average, it means that the performance can be improved without increasing the cost.

Extension belt-beam

The current extended design of the traditional ion implanter has a feature that produces a stronger beam than in the past. There are three reasons for this: 1) the development trend of large-scale substrates in the industry, such as 300mm-diameter silicon wafers; 2) the development of ultra-large substrates, such as 450mm-diameter silicon wafers, will be used in CMOS and equipment manufacturing. 3) In the design of the continuous planting machine, the beam is extended through the fusion electromagnetic scanning to increase the circular production capacity and improve the uniformity of the substrate dose. In traditional ion implantation, wafer production capacity will shrink as the substrate area increases. The goal of cost reduction achieved by large-size substrates cannot be achieved without stable production capacity coordination. Generating a stronger ion beam is the key technology for ion implantation in silicon equipment manufacturing. The beam flow delivered to the wafer (from the dose rate is proportional to the length of the ribbon beam, the reasons why this feature is suppressed in the conventional ion source are as follows: 1) The conventional ion implantation source can only produce a certain degree of ribbon Beam (approximately 2-3 inches long), 2) If the extended band beam is generated by the beam expansion optical system, the ion current density of the beam will be significantly reduced and the total ion current transmitted to the large substrate will remain unchanged.

Using the related technologies of Figures 3 to 5, Figures 6a to 8 and Figures 9 to 10 of the embodiment, the generated beam is almost equivalent to the beam directly drawn by the ion source. Referring to Figure 17, this can be achieved by reducing the length of the ion source. Figure 17 shows an embodiment similar to the embodiment shown in Figure 4. An electron gun 230 transmits a variable energy electron beam 235 to the ionization chamber 240, which contains gas containing impurities, and is finally driven by a water-cooled light stopper. 250 intercepted. The electron beam propagates parallel to the adjacent ion extraction aperture 260, in which an ion beam is extracted by the extraction optical system. The magnet coil (not shown in the figure) generates an external magnetic field B (not necessary). The longitudinal magnetic field is parallel to the path of the electron beam to restrict the electron beam 235 when the electron beam travels. The path length in Figure 17 is represented by (X+y), where x is the length of the electron gun and y is the length of the ionization chamber (y is also approximately equal to the length of the ion extraction aperture and the ideal length of the extraction belt ion beam 270). It is assumed that the ionization chamber 240 is distributed in a cylindrical shape with a disc shape with a unique extraction aperture.

The ion doping ion source shown in Figure 17 can be used to manufacture panel displays. In this example, the ion extraction aperture 260 can be 850 mm long to produce a 750 mm square panel f with beam implantation. The length y of the ionization chamber in Figure 17 can be longer than 850mm, such as 900mm. The design purpose of the electron gun 230 is to deliver a strong ion current, low energy electron beam 235 to the ionization chamber 240. The general specifications are: cylindrical lens diameter 1inch, electron beam energy 100eV (adjustable from 20eV to 250eV), maximum electron ion current 200mA, electron ion current dynamic range 400 (that is, electron ion current can be adjusted between 500uA and 200mA). The magnet coil generates an external magnetic field B to limit the electron beam of the electron gun and ionization chamber. Since the electron beam path length is quite long after leaving the ion source, a magnetic field B is needed to assist in calibrating the path. By limiting the diameter of the electron beam to propagate in the ionization chamber 240, the low-energy electron beam is affected by the space charge force, and the magnetic flux density is set to be between 50G and 200G, which can maintain the ion generation (ion density) along the ion extraction aperture length of 260 Uniformity. The uniformity of the ions generated along the aperture can be achieved by reducing the feed pressure of the ionization chamber 240 (compared to the conventional ion source pressure of about 4x10 ^-4 Torr and 4x10 ^-³ Torr) to improve, so that the proportion of electrons diverging beam can be reduced to about 1x10 ^-4 Torr is lower. It should be noted that the conventional plasma ion source cannot be operated at a greatly reduced pressure because the plasma cannot be maintained at low pressure. The purpose of this low-pressure operation is to reduce the consumption of the working gas in the ion doping system, and the tool cost of ownership (COO) can be greatly reduced by adjustments of about one order of magnitude.

Figure 18b shows an example of further reduction in equipment cost. The figure shows a dual ion source system 600 with a dual-slit extraction optical system 610. The large-diameter magnet coil 620 provides a uniform magnetic field around the two ion sources. The embodiment of Figure 18b can use a single ion doping device to implant n-type and p-type impurities by arranging n-type (eg, phosphorous) materials in the ion source 1, and p-type materials (eg, boron) in the ion source 1. Source 2. Generally, ions 1 and 2 do not run at the same time. If the situation requires, the two ion sources can also use the same impurity to operate at the same time to generate two beams, increasing the implanted dose rate by a factor of two. Figure 18b shows that the two-band ion beams are simultaneously generated as beam 630 (for example, containing boron) and beam 640 (for example, containing phosphorous).

Fig. 18b shows a situation where the dual ion source of Fig. 18a is doped with a square panel 650 with a beam of light 660. In this example, the panel 650 is fixed to the scanning table 670 and scanned along the direction 680, along the long side of the panel 650. Note that in this example, the ion beam 660 is longer than the short side of the panel.

Figure 19 shows the ion doping system of commonly used doped panels. The panel 690 is fed into the action chamber 710 by the vacuum box 700, and after being rotated 90 degrees, it is scanned in front of the ion beam 660. The dual ion source 600 in Figure 18a can be seen integrated in the system. The details of the magnet coil and the mechanical feed tube are not included due to the simplicity of the picture.

In ion doping, the P-type feed gas can be diborane (B ₂ H ₆ ) And boron fluoride (BF ₃ ). Mass analysis is not used between the ion source and the substrate, and all ions generated by the ion source are implanted on the substrate. This makes BF ₃ The use of fluorine is tricky because fluorine is harmful to oxides and can have a negative effect. In addition, the fluorine content in the ion source plasma is three times that of boron, which will cause too much fluorine to be implanted. Most manufacturers prefer to use B ₂ H ₆ , The effect caused by H implantation is not large (H implantation will cause the substrate to overheat). Generally speaking, there are two shortcomings: 1) The cracking situation becomes larger (many different ions are generated, for example, H ⁺ , BH _X⁺ , B ₂ H _X . The planting range is enlarged because the boron-containing ions with different energy are scattered and planted to penetrate the planting target). 2) Insufficient beam flow leads to low productivity. The surface area of the outer wall and the large ion source volume cause B to be lost to the outer wall of the ion source, resulting in low ion productivity.

Using the ions in Figure 17 to use decaborane as the feed gas material can solve problems 1) and 2), because this method can produce 70% pure B ₁₀ H _X With strong beam ion current (as shown in the decaborane NIST-traceable spectrum in Figure 19a), the surface area of the ion source of the present invention is several orders of magnitude smaller than that of the cylindrical ion source, and the ion source in Figure 17 can verify its high Ion extraction efficiency. Compared with the prior art, the ion source using decaborane in Figure 17 can produce higher productivity, lower equipment cost, and lower particle precipitation (because there is less material accumulated in the ion source).

Acceleration and deceleration technology

For ion implanters that use ion beam deceleration technology before hitting the substrate, generating high-brightness ion beams is an important condition, because the angular divergence and spatial distribution of the decelerated beam will expand after deceleration. To produce a small beam of angular divergence and good spatial uniformity that reaches the target substrate after deceleration, the initial low emission beam is a necessary condition. Since the beam emission (affected by the beam diameter and angular divergence in the two right-angle directions) is inversely proportional to the energy, the upstream beam emission must be smaller than the emission when it reaches the substrate, which is about the same as the reduction ratio. By increasing the serial aperture, the beam emission can be maintained below a certain level, so that the beam flow rate is lower than the acceptable level. Therefore, it is an ideal practice to use a high-brightness ion source, where the brightness can be defined by dividing the beam flow by the emission (that is, the beam flow per unit area per unit angle). The advantage of this method is that the brightness of the serial aperture does not change.

Some implants, such as extended drain implants, require two conditions of low-angle divergence and low-energy ions when reaching the substrate at the same time. If a high-brightness source is not used, the two conditions will be mutually exclusive. When using high-brightness beams, the net implantation dose rate obtained is higher than that obtained when using low-brightness beams, which can be used to increase productivity and reduce equipment costs.

The use of monovalent ion clusters with multiple impurity atoms can produce a higher brightness beam, especially when we replace the beam stream with a dose rate or "effective" beam stream. A monovalent ion cluster composed of n atoms can be accelerated to n times the energy, and its emission is n times smaller than that of an equivalent single-atom beam. Since the dose rate is also n times the ion current, the total brightness of the cluster beam ion current increases to n ² At this time, the brightness is defined as the dose rate divided by the emission. Therefore, the use of a high-brightness ion source capable of generating cluster ion beams can be used for planting under well-controlled decelerating beams, which not only has small angular divergence, good spatial uniformity, and high productivity.

Referring to Figure 20, the first drawing of PCT Patent Application No. US00/33786 in the above reference documents is similar, but the description of this drawing is more detailed. This drawing shows an example of a decelerating planting machine used for traditional boron planting. Figure 1 is a PCT patent application showing a traditional, non-deceleration planting machine. In any case, in the decelerating implanter in Figure 29, the ion source 548 produces ions drawn from a wire aperture (that is, an elongated narrow hole) and then accelerated by the electrode 553 to have mobile energy, which is greater than the final implant energy. The analyzer magnet 543 is injected again to disperse the beam laterally according to the ion charge-to-mass ratio. A mass dissolving aperture (narrow hole) 544 can filter and only allow the required ions to pass through (the ions have the required charge-to-mass ratio) to pass downstream, and measure the ion beam current with a Faraday, or (when no rady is installed) ) To the deceleration electrode 557. The deceleration electrode 557 decelerates the ion beam to the required implantation energy, and then hits the wafer substrate 555. The one shown in Figure 20 is a batch planting machine, which can be mechanically rotated by the turntable 545, and the deceleration technology can also be applied to a continuous planting machine.

Before reaching the decelerating electrode, some ions in the ion beam may interact with the residual gas molecules on the beam line of the implanter or other ions in the beam to cause valence changes. Most of the accelerator and decelerate implanters also have a neutral beam filter design ( Not shown in Figure 20) or other types of energy filters (such as ExB filters, electrostatic reflectors, doglegs, etc.), as disclosed in the prior art, ensure that only ions with a predetermined energy can reach the wafer.

Generally, the batch deceleration planting machine uses a static ion beam, which scans the wafer by mechanical rotation of the turntable 545; however, it can also be implemented in other ways. For example, the present invention can be integrated into a continuous planting machine (processing one wafer at a time) in a single-direction rapid scanning mode (by electrostatic scanning or by a direction magnetic field), while the wafer holder is in the vertical direction Complete a slower mechanical scan. Or it is possible to scan static wafers with dual electromagnets. As far as I know, the commercialization of the continuous deceleration planting machine has not yet been achieved. The advantages of this design are the integration of single-wafer processing and the wafer holder that produces a high oblique angle (up to 60 degrees) (the current cascaded cloth The planting machine has no such design). High oblique angle implantation is important in many manufacturing processes, and it is also popular such as "quad" implantation of well structure, and S/D pole expansion distribution. In addition, the new wafer manufacturing equipment is expected to use single wafer processing to reduce the costly 300mm process wafer batch operations.

Continuous implantation requires higher beam distribution uniformity on implanted wafers than batch implantation. This requirement is more difficult to achieve on 300mm diameter substrates than 200mm diameter substrates. One of the features of the present invention is that the cluster beam implantation and acceleration and deceleration technology are combined by a continuous and strong ion current beamline implantation system. The improved beam distribution can meet the requirements for the uniformity of the beam distribution of the continuous implanter. Increased production capacity.

Figure 21 shows a preferred embodiment of an ion source, which is fixed on the vacuum housing of a conventional ion implanter. This ion source is the same as the aforementioned PCT Patent Application No. US00/33786, and its diagram is shown in Figure 9B. The operating principle of this ion source is different from the traditional ion implantation of commercial ion sources in that the ion generation in this method is not by glow discharge or plasma, but by direct electron-impact ionization, and by transmission in the ionization volume 516, The broad beam of light excites the main electron. This ion source technology is suitable for ionizing molecular compounds, and can provide a strong ion current with decaborane ions and dimer compounds shown in Table 1a. The ion source incorporates a low-temperature vaporizer 528 to generate steam using low-melting solid materials such as indium hydroxide, trimethyl indium, and decaborane. The ion source also incorporates a gas feed 526 to make gaseous compounds such as PH ₃ `AsH <sub>3</sub> `GeH ₄ `B <sub>2</sub> H <sub>6</sub> , And other planting machine gas such as BF <sub>3</sub> , SbF <sub>5</sub> `With PF ₃ To be ionized. The embodiment in Fig. 21 shows that an extended electron gun includes a 90-degree pedestal or a mirror 587 to reduce the single grounding area of the ion source combination, which not only saves space, but also allows the electron optical system to be incorporated into the electron gun, so that the electron energy can be diversified. Chemistry meets the needs of different types of molecules.

Figure 23 shows another embodiment of the ion source and electron gun, which has also been disclosed in the above-mentioned related application. Please refer to the published application in Figure 18b. The electron gun operates on the principle of acceleration and deceleration without using a turning stage. Figure 23 shows that the ion source is fixed to a vacuum enclosure as shown in Figure 21, but the shape of the enclosure has been modified to be incorporated into the electron gun design. The embodiment of Figure 23 also has a set of magnet coils to limit the path of the main electrons in the ionization volume of the ion source. Strong electron ions with controllable variable electron energy flow through the electron gun and enter the ionization chamber; most of the electrons pass through the ionization volume and are intercepted at the stopper 536 (see Figure 22).

In the embodiment of Figure 22, the length of the ionization chamber and the corresponding extended ion extraction aperture have been specially lengthened. The reference application in Figure 18 shows that the extension of the length of the ionization chamber shows that the negative effect is small but positive when operating the ion source. The effect is big. This extension is very different from the Bernard-type glow discharge ion source of traditional ion implantation. If the length of the ionization chamber (the distance between the cathode and the repeller) is designed to be greater than the normal length, the required glow discharge ion current in the Bernard ion source will increase rapidly (the traditional Bernard ion source has a cathode-repeller distance About 2 inches, it can draw up to 5A of glow discharge ion current), and the glow discharge will become unstable. According to the present invention, by greatly extending the ionization volume and the length of the ion extraction aperture, it is possible to extract a stronger ion current than other embodiments. For the problem of beam divergence from the aperture, a special optical system can be solved by shortening the beam distribution length. This method is shown in Figure 24. An elongated ionization chamber 500 having an elongated ion extraction aperture 510, in which the ion beam is extracted and accelerated through the extraction lens 520. I currently set the length of the extraction aperture 510 to 6 inches, which is three times the length of the extraction slit of the conventional glow discharge chamber of the general strong ion implanter. The draw lens 520 has the feature of retractable adjustment; it is a set of two groups of acceleration lenses, in which the second focus of the first lens approximately coincides with the first focus of the second lens to focus the light beam. By setting the lens 520 to achieve a three-to-one order of magnitude, the height of the ion beam can be greatly shortened, and the beam path 530 entering the analyzer magnet 540 of the implanter can be calibrated at the same time. In the past, because the ion extraction aperture of the strong ion implanter was designed to have a longer length to absorb the beam current of a single beam, the space charge force would make the lens 520 unable to operate ideally, and the beam path 530 would not be aligned. In other words, The beam often expands, so the method of using ion optics has not been successful. In fact, the effect is the same as the beam flow obtained through the traditional 2inch aperture and filtered by the analyzer magnet! The reason is that the traditional Bernard type glow discharge ion source used in the traditional strong ion current implanter is greatly affected by the space charge when the beam moves. The ion current generated by the 2inch aperture can reach 50mA (BF ₃ In plasma, about 30% or more of 15mA is B ⁺ ). The ion source of the present invention, on the contrary, can provide an ion current of 1 mA or more from a 2 inch aperture, which accounts for about 7% of decaborane. Therefore, if the aperture is 6 inches, the available ion current will reach 3 mA to 5 mA. The deceleration planter using boron (the deceleration ratio is not an arbitrary ratio, but is determined by the energy pollution and movement restrictions to achieve the ideal beam flow). Because this weak ion flow can absorb higher energy, the ion flow moves as the first Figure 24 is therefore not limited to space charge. Therefore, as shown in Figure 24, the present invention estimates that the beam current can be tripled by lengthening the ion source extraction aperture while maintaining good beam emission characteristics.

In the above example, the width of the ion-carrying beam is shorter than the ion extraction aperture. In another example, the ion-carrying beam is longer than the ion extraction aperture. In other examples, the ion beam is approximately equal to the length of the ion extraction aperture.

Electron gun

The following is an introduction to the electron gun. It should be noted that the disclosed hot cathode electron gun is prone to heat dissipation problems in a vacuum environment. The radiant heat generated by the hot cathode and the heat conduction between the cathode end and the surrounding and the electron impact of individual lens components in particular, the above heat sources Make the heat dissipation of the lens design a key item, and make the lens and the object reach thermal equilibrium through a cooling groove. Create electrical insulation and conduction mechanical paths to allow heat to leave the lens assembly. The most important thing is the cathode assembly because its heat dissipation occupies an important part of the electron gun. In practice, the design of this part of electrical insulation and conduction mechanical path cooling is particularly prone to failure. One of the solutions is to gradually achieve a thermal balance between the temperature of the lens assembly and the environment, but this method also has its problems. Even if the lens assembly is made of refractive material, it can maintain the operating temperature of 1000°C or higher, and the interaction with the working gas makes this method ineffective. Especially when decaborane is used, the environment will crack the decaborane and cause boron to precipitate on the lens assembly to form harmful implanted particles, which may cause a short circuit and greatly shorten the life of the cathode and the ion source. According to the invention, detailed temperature control can be achieved by transferring radiant heat to the cooling body around the lens assembly or, more clearly, around the heat transfer holder, up to 4πsteradians. In addition, the holder clearly regulates the lens components and The relative position between the ion sources.

Referring to FIG. 25, the holder 310 clamps and fixes a lens assembly 300. The holder 310 is made of a rectangular aluminum rod and has a large radiating surface to fix the lens 300 in a clamping manner. In the disclosed embodiment, the lens 300 is placed in the inner diameter by inserting into the stretcher not shown in the figure to unfold the clamp. In addition, the holder 310 and the lens 300 can also be placed in the inner diameter through the extremely cold lens 300, for example, placed in a cold bath in liquid nitrogen, and then the cooled diameter lens 300 is placed in the holder 310, and later when the lens 300 rises It is combined with the inner diameter when it expands to room temperature. These fixed lens components, such as the lens component 300, can produce a heat dissipation effect, effectively dissipate heat from the surface, and allow the surface of the fixture of the radiant heat holder 310 to escape through. To improve the emission ability of the aluminum holder 310, the holder can be plated with a layer of colloidal suspended substance of carbon component such as Aquadag by using cathodic electroplating. Another advantage brought by the combination of the lens holder 310 and the lens assembly 300 is that the holder and the lens assembly can be made of different materials. For example, the lens 300 can be made of chemically inert stainless steel or molybdenum which has good structural properties at high temperatures.

Referring to Figures 26 and 27, the electron gun is composed of 4 non-continuous separate lens components: the cathode assembly 320, the first cathode 330, the focusing electrode 340, and the exit lens 350 in order. The voltage used by each component is different. It is fixed by a separate fixing device 310. There is a gap between the lens assembly 350 and the bottom edge of the housing 360, and the housing surrounds a combination of 4 lens holders.

The temperature of the aluminum housing 360 is much lower than that of the lens components 320-350, and it is formed by radiation coupling with several lens holders 370. The connection is achieved by establishing good heat conduction between the aforementioned water-cooled or temperature-controlled ion source region and the connection bottom edge 360a of the housing 360.

With the thermally conductive elastic sealing edge between the temperature-controlled bottom edge 360a and the ion source area, the advantages of large contact surface area and good thermal coupling enable the housing 360 to maintain a temperature close to the ion source area. Therefore, a temperature difference between the plurality of lens holders 370 and the housing 360 is maintained, so that the radiation energy can be smoothly transmitted from the radiation surface of the lens holder. Moreover, the use of radiant cooling makes the operating temperature of the electron gun stable. This feature has nothing to do with the heat dissipated through the gun assembly. This stability comes from the non-linear effect of radiant cooling, which is less pronounced at low temperatures and more pronounced at high temperatures. This combination has the characteristics of self-regulation and can maintain the electron gun assembly at a consistent operating temperature.

Each lens holder 370 is mostly square, because more surface area can be used to oppose the surface of the adjacent lens assembly holder and the surface of the housing 360. This design has two functions. The first function is the area that is expected to have the greatest heat (that is, the uppermost lens assembly, which includes the cathode assembly 320 and the bottommost lens outlet lens 350), which can directly transfer heat to the nearest housing 360 In the thermal bath, at the same time, the heat energy dissipated by the lens components 330 and 340 shown in FIG. 26 is being transferred between the lens components to make the operating temperature uniform. However, since the four sides of the holder radiately face the surface of the housing, the lens assembly still has a considerable radiation loss area for the cooling housing 360. If the holder can have at least two sides facing adjacent components, the other sides will face the cooling housing 360.

Referring to Figure 25, the general size of the holder 310 is 170 mm in length 1, width w of 26 mm, and height h of 12 mm. The lens assembly 300 generally has an inner diameter of 12.5 mm and an outer diameter of 16 mm.

Refer to Figure 26, the general size of the gun assembly is 6 inches in length L, 3 inches in height H, and 1.5 inches in width W. The feeding tube 380 and each lens assembly are connected with a metal spring 390. The ceramic holder plate 395 establishes electrical insulation. The connection between the fixtures and the components between the lens holder 370 is mechanically fixed to allow electrical conduction and maintain a lens. The order of the components. The aluminum strip 400 coaxially connects the lens components with a plurality of aluminum spacers 410 and is embedded in the lens holder 370. To achieve electrical insulation while controlling the distance and three-dimensional arrangement of the lens components.

The angular deviation of the coaxial position of the lens assembly 320, 330, 340, 350 and any connecting lens must be within tolerance in order to produce effective electron beam focusing and reduce the deviation. The spacing between the lens components is directly affected by the arrangement, especially the polar coordinates determined by the circular symmetry of the lens system (the Z axis is along the mechanical axis, meaning the electron beam propagation direction), and the two-dimensional spatial coordinates generated along this axis. It should be emphasized that the adjustment of the electron beam arrangement direction and focus are within the mechanically achievable range, through the correct assembly of the aluminum rod 400, the individual spacer 410, the lens assembly 320-350, and the correct arrangement of the several holders 370 That's it. The correct arrangement of the lens assembly ensures that the beam can move in the ionization chamber in the correct direction, so that the beam propagates parallel to the axis of the ion extraction aperture. In this way, a small-volume ionization zone can be accurately placed near the ion extraction aperture to achieve high brightness and low emission goals. Moreover, proper temperature control of the lens assembly is an essential element for operating the ion source to prevent the dissociation and precipitation of the gas or vapor from filling the ionization chamber or penetrating the electron gun area. Excessive dissociation and precipitation acting on the lens assembly will shorten the life of the ion source and increase maintenance costs. In order for the ion beam to have stable characteristics after the ion extraction aperture is extracted, the distance between the ionization zone and the ion extraction aperture must be even. The misalignment or out-of-focus of the electron beam will cause the proximity effect and the size of the electron beam to deviate, causing it to move along the ion extraction aperture and adversely affect it. The proper alignment and focusing of the electron beam obtained by the correct alignment and coupling, as well as the stable temperature of the holder, can ensure that the ionization chamber produces a high-brightness ion beam and its lateral ion can be controlled. Not only that, the ions can be mostly in front of the aperture. Resulting, a uniform density is formed along the long axis of the aperture. The brightness of the ion beam drawn from the ion source is proportional to the total ion current and inversely proportional to the ion emission just leaving the ion extraction aperture of the outlet. If the total ion current is a fixed value, the ion beam brightness of the ion source will increase when the ionization volume decreases. It can be understood from the above design that, compared to the electron beams obtained by misfocusing or misalignment in a large ionization zone, the ion source ions have higher ion brightness due to the dense electron beams generated under good control of a small ionization volume.

Among the brightness characteristics in the design of the lens and the lens holder, the ion brightness is derived from the acceleration and deceleration ion implanter and needs to be maximized, because the implantation efficiency of the wafer is proportional to the ion beam brightness. In other words, the planting machine and the production characteristics of the planting machine directly determine the brightness of the ion beam. Therefore, the high-brightness beam design of the acceleration and deceleration type planting machine is the key to the design of improving the acceleration and deceleration efficiency.

Decaborane or the like is used to enhance the brightness of the beam. In various implanters, especially when it is related to the acceleration and deceleration implantation of ions, the temperature control of the various details of the ion source includes the electron gun component part, which is in contact with the decaborane vapor, etc. It is the key to the successful operation of the ion source. It should be noted that this method can produce a strong ion current. In addition, it can also increase the life of the ion source and increase the productivity of the ion source and implantation. When the outer wall temperature is higher than 350C, decaborane will undergo boron dissociation. The decomposed boron particles will precipitate on the cathode and lens assembly. If it deposits too much on the cathode, it will deteriorate the performance of the cathode and reduce ion generation and shorten the life of the cathode. In addition, the boron component will charge the lens assembly, deteriorating the effect of controlling the electron beam, thereby reducing the brightness of the ion beam generated by the ion source. Therefore, only with a good temperature control design can the application of decaborane in acceleration and deceleration applications be successful.