Method for manufacturing a semiconductor device

Abstract

A method (1000) for manufacturing a semiconductor device (800), the method (1000) comprising: Reducing a thickness of a semiconductor substrate (802); and Changing an ion acceleration energy of an ion beam while a relative movement is carried out between the semiconductor substrate (802) and the ion beam impinging on the semiconductor substrate (802), wherein changing the ion acceleration energy involves controlling a temporal change in the acceleration energy dE / dt as a function of a value of an ion beam current density I measured during the relative movement B,meas includes.

Description

BACKGROUNDSome parameters of semiconductor devices may be associated with properties of vertical dopant profiles. For example, vertical power semiconductor devices controlling a load current between a first load electrode at a front side and a second load electrode at a back side of a semiconductor chip include doped regions such as a drift zone, compensation structures, buffer layers and field stop layers with specific vertical dopant profiles, wherein parameters of the vertical dopant profiles of these layers such as uniformity, smoothness and ripple may have a substantial influence on device parameters. Compared to in situdoping during layer deposition, ion implantation enables precise monitoring of both a total dose and a dose rate. Ion implantation typically results in a Gaussian-like distribution of dopants around an end-of-range peak whose distance to a substrate surface is a function of ion acceleration energy of the implanted ions. Production methods with ion implantation are known, for example, from the publications US 2003 / 0 224 612 A1 and US 2015 / 0 371 858 A1, DE 10 2015 115, 173 A1, Laska, T. et al: Ultrathin-Wafer Technology for a new 600V-NPT-IGBT, in: Proc. of the 9th Int. Symp. on Power Semicond. and IC's (ISPSD), pp. 361-364, 1997, issn: 1063-6854, u.s. Pat. No. 6,229,148 B1, Ryssel H.; Rug I.: Ion implantation, 1st Ed. Stuttgart: Teubner, 1978, pp. 124-129, ISBN 978-3-519-03206-9, and also from the post-published document DE 10 2017 119 571 A1. Formation of doping profiles with a low degree of ripple, for example box-like or triangular doping profiles, may be limited by limitations of process technology, for example by a limited thermal balance for backside processes with respect to completed front side structures or by peaks in the range of the penetration depth (end of range peaks) of proton implants.There is a need for an improved ion implantation method and apparatus.SUMMARYThe object is achieved by the teaching of the independent patent claims. Further developments are the subject matter of the dependent claims.Further embodiments are described in the dependent claims. Those skilled in the art will recognize additional features and advantages upon reading the following detailed description and upon viewing the accompanying drawings.BRIEF DESCRIPTION OF THE DRAWINGSThe accompanying drawings are included to provide a further understanding of the present embodiments. The drawings illustrate the present embodiments and together with the description serve to explain principles of the embodiments. Other embodiments and intended advantages will be readily appreciated as they become better understood by reference to the following detailed description. FIG. 1 is a simplified flow diagram illustrating a method of manufacturing a semiconductor device. FIG. 2 is a simplified flow diagram illustrating another method of manufacturing a semiconductor device. FIG. 3 is a graph for illustrating a concentration profile of hydrogen-based donors. FIG. 4 is a graph to illustrate curves r1, r2 of a time change of the ion acceleration and ion implantation energy normalized to the ion beam current density, dE / dt / Ib(t), versus the ion acceleration or ion implantation energy with respect to an energy window ΔE=E 2- E 1. FIGS. 5A and 5B are schematic block diagrams to illustrate a closed loop control process of time variation of ion implantation or ion acceleration energy normalized to the measured ion beam current density, dE / dt / I B( t). FIG. 5C is a schematic graph for illustrating a relationship between a time change of the ion acceleration energy normalized to the ion beam current density, dE / dt / I B( t) versus ion acceleration energy and the doping concentration versus ion acceleration energy. FIG. 6 is a schematic illustration of an ion implantation apparatus based on electrostatic scanning along first and second scanning directions. FIG. 7 is a schematic illustration of an ion implanter based on both electrostatic and mechanical scanning. FIG. 8 is a schematic block diagram for illustrating an ion beam current density control process. FIG. 9 is a schematic cross-sectional view of a semiconductor device.DETAILED DESCRIPTIONIn the following detailed description, reference is made to the accompanying drawings, in which specific embodiments in which the embodiments may be practiced are shown for illustrative purposes. It is to be understood that other embodiments may be utilized and structural or logical changes may be made. For example, features illustrated or described for one embodiment may be used in or in conjunction with other embodiments to arrive at yet another embodiment. The present disclosure is intended to encompass such modifications and variations. The drawings are not to scale and are for illustrative purposes only. Corresponding elements are provided with the same reference numerals in the different drawings, unless stated otherwise.The terms "have," "include," "comprise," "have," and similar terms are open ended terms, and the terms indicate the presence of stated structures, elements, or features, but do not exclude the presence of additional elements or features. The indefinite articles and the definite articles are intended to include both the plural and the singular, unless the context clearly indicates otherwise.FIG. 1 is a schematic flow diagram to illustrate a method 1000 for manufacturing a semiconductor device.It should be understood that while method 1000 is illustrated and described below as a series of steps or events, the illustrated order of such steps or events is not to be interpreted in a limiting sense. For example, some steps may occur in different orders and / or concurrently with other steps or events apart from those illustrated and / or described herein. In addition, not all steps may be required to implement one or more aspects of embodiments of the disclosure herein. Also, one or more of the steps depicted herein may be divided into one or more separate substeps and / or phases.Referring to FIG. 1, the process feature F 10 includes reducing a thickness of a semiconductor substrate. The semiconductor substrate may be a semiconductor wafer or die cut from a semiconductor wafer. The semiconductor substrate may comprise a semiconductor wafer having no, one or more optional epitaxial layers thereon. The semiconductor substrate may also have one or more doped semiconductor regions therein. The semiconductor substrate may be a silicon semiconductor substrate, e.g., a Czochralski (CZ) silicon semiconductor substrate such as a Czochralski (MCZ) magnetic silicon semiconductor substrate. In some embodiments, a material of the semiconductor substrate corresponds to another single crystalline semiconductor material such as silicon carbide (SiC), gallium arsenide (GaAs), gallium nitride (GaN), or another A III B V semiconductor substrate, germanium (Ge), or silicon germanium (SiGe).Referring to FIG. 1, reducing a thickness of the semiconductor substrate may be performed by removing material of the semiconductor substrate from a surface on a second side, e.g. a back side of the semiconductor body opposite a surface on a first side, e.g. a front side. In one or more embodiments, the semiconductor substrate is mechanically supported by a carrier attached to the surface on the first side. Material of the semiconductor substrate may be removed by a chemical process, e.g., by etching such as dry or wet etching, mechanical processes, e.g., abrasive processing such as grinding or polishing, and chemical mechanical processes such as chemical mechanical polishing (CMP). In one or more embodiments, a combination of more than one process for removing material of the semiconductor substrate may be used, e.g., a first process having a greater material removal rate than a second process following the first process. This may allow fine adjustment of a target wafer thickness, for example. The process for removing material of the semiconductor substrate from the second surface may also include a so-called TAIKO process. The TAIKO process is a wafer thinning process in which an outer support ring is not thinned along an edge of the wafer during the thinning process. The outer support ring may support improved thin wafer handling during subsequent processing. For example, wafers thinned using the TAIKO process can typically maintain their strength without being attached to an additional carrier.Referring to FIG. 1, process feature F 20 includes changing an ion acceleration energy (also referred to as ion implantation energy) of an ion beam while performing relative movement between a semiconductor substrate and the ion beam incident on a surface of the semiconductor substrate. Therefore, the ion acceleration energy determined by, for example, an ion acceleration voltage of an ion acceleration unit is changed during a single ion implantation process. A single ion implantation process is an ion implantation process based on a single implantation recipe and is not interrupted by, for example, a tuning period for changing the implantation recipe. In other words, during a single ion implantation process, an ion beam is continuously directed at a target, for example a wafer, i.e., the ion beam is not interrupted in time for tuning purposes. The ion beam may contain ions, for example protons, helium or ions with an atomic number greater than 4, for example ions of nitrogen, aluminum, boron, phosphorus, arsenic, sulfur, selenium, germanium, antimony or oxygen. In one or more embodiments, changing the ion acceleration energy may be subject to a closed loop control process. Process feature F 20 may be performed after process feature F 10. Process feature may also be performed prior to process feature F 10. If process feature F 20 is executed multiple times, process feature F 10 may be executed between process features F 20.FIG. 2 is a schematic flow chart to illustrate another method 1001 of manufacturing a semiconductor device.It should be understood that while method 1001 is illustrated and described below as a series of steps or events, the illustrated order of such steps or events is not to be interpreted in a limiting sense. For example, some steps may occur in different orders and / or concurrently with other steps or events apart from those illustrated and / or described herein. In addition, not all steps may be required to implement one or more aspects of embodiments of the disclosure herein. Also, one or more of the steps depicted herein may be divided into one or more separate substeps and / or phases.Referring to FIG. 2, the process feature F 15 includes forming a doped region in a semiconductor substrate. In one or more embodiments, the doped region is formed on a first side of the semiconductor substrate by introducing dopants through a surface on the first side. The doped region may be formed with a masked or non-masked doping process, e.g. by doping processes such as ion implantation, diffusion from a diffusion source or in situ doping. The doped region may be a body region, a source region, a highly doped body contact region, a channel implantation region of transistors such as IGBTs, an anode or cathode region of a diode, an edge termination structure such as a variation of lateral doping (VLD) region or a junction termination region (JTE), a guard ring or a channel stopper. The process feature F 15 may also include processing the semiconductor substrate on the first side by, for example, processes for forming insulating regions, e.g., thermal oxidation processes, thermal nitriding processes, dielectric layer deposition processes such as chemical vapor deposition (CVD), low-pressure chemical vapor deposition (LPCVD), plasma-enhanced chemical vapor deposition (PEPCVD), atmospheric pressure chemical vapor deposition (APCVD), subatmoshary chemical vapor deposition (SACVD, The chemical vapor deposition may include low density plasma chemical vapor deposition (HDP CVD). Exemplary insulating regions are gate insulation layers, e.g. thermal oxides or nitrides or high-k and low-k dielectrics, field insulation layers, e.g. field oxide layers, insulating intermediate layers, e.g. deposited oxides such as tetraethylorthosilicate glass (TEOS), borophosphosilicate glass (BPSG), borosilicate glass (BSG), phosphosilicate glass (PSG), device insulation layers, e.g. local oxidation of silicon (LOCOS) or trench isolations. The processing of the semiconductor body on the first side may also comprise processes for forming conductive regions, e.g. physical vapor deposition (PVD) of materials such as metals and metal compounds, chemical vapor deposition (CVD) of materials such as doped polycrystalline silicon or tungsten and electrochemical deposition (ECD) of materials such as copper. Exemplary conductive regions are gate electrodes, conductive paths of wiring planes, bond pads, vias, and contacts. Processing the semiconductor body at the first surface may further include processes for removing material from the first surface, e.g., etching processes or abrasive processing such as grinding or polishing and chemical mechanical polishing (CMP). Exemplary structures attributable to material removal are trenches, e.g. gate trenches, or field trenches or trenches of trench isolations, contact holes.Referring to FIG. 2, process feature F 20 is executed as described with reference to FIG. 1.In one or more embodiments, changing the ion acceleration energy comprises controlling a temporal change of the acceleration energy dE / dt depending on a value of a measured ion beam current density I B,meas. The ion beam current density I B,meas can be measured by an ion current detector unit, for example, a Faraday cup. By considering the measured ion beam current density I B,meas the time change of the ion acceleration energy dE / dt normalized to the ion beam current density may be the control variable, thereby making it possible to counteract variations in ion beam current density. This is advantageous in terms of achieving a desired local implantation dose, i.e. an implantation dose at a specific implantation depth, regardless of variations in the ion beam current density.In one or more embodiments, the changing of the ion acceleration energy comprises controlling a temporal change of the acceleration energy dE / dt depending on a value of a measured total implantation dose Dint.When controlling the change over time of the acceleration energy dE / dt, the measured total implantation dose Dint corresponds to the total implantation dose since the beginning of the ion implantation process, e.g. from t=0 s to the time tref, at a specific time tref.In one or more embodiments, the method further comprises converting a target dopant concentration versus depth D L( x p), where x p( E) is a projected range of dopants as a function of ion implantation or ion acceleration energy, into a targeted temporal change in ion acceleration energy normalized to the ion beam current density, dE / dt(E) (I B,norm)set= dE / dt(E) / I B,expected, where I B,expected is an expected ion beam current density. The target doping concentration versus depth D L( x p) may be a desired doping profile to be realized by ion implantation. Ion implantation processes at a fixed ion implantation or ion acceleration energy may need to be performed multiple times to realize the target doping concentration versus depth D L( x p). This may require additional effort and cost in view of tuning cycles between successive ion implantation processes to change implantation formulations. By controlling the change over time of the ion acceleration energy dE / dt depending on a value of the measured ion beam current density I B,meas the target doping concentration versus depth D L( x p) can be realized by a single ion implantation process based on a single implantation recipe. The expected ion beam current density I B,expected may be calculated taking into account parameters set in the implantation formulation, for example a local dose to be introduced at an initial ion acceleration energy.In one or more embodiments, the change over time of the ion acceleration energy dE / dt may be controlled within a window ΔE=E 2- E 1 of the ion acceleration energy, wherein values of an initial ion acceleration energy E 1 and a final ion acceleration energy E 2 may be determined depending on the target doping concentration versus depth D L( x p) For example, the window ΔE of the ion acceleration energy is larger the larger the depth range of dopants to be introduced. Likewise, the greater the target doping concentration at a particular depth, the smaller the change over time in the ion acceleration energy dE / dt.In one or more embodiments, controlling the change over time of the ion acceleration energy dE / dt as a function of the measured ion beam current density I B,meas further comprises comparing, at an ion acceleration energy E, an targeted change over time of the ion acceleration energy normalized to the ion beam current density dE / dt (I B,norm)set= dE / dt target / I B,expected, where I B,expected is an expected beam current density, with a set change over time of an ion acceleration energy normalized to the measured ion beam current density I B,meas, dE / dt set / I B,meas. When the temporal change of the ion acceleration energy dE / dt normalized to the ion beam current density is the control variable, a closed loop control unit has a feedback loop that ensures that the control unit performs a control action to provide an actual process output, i.e., the adjusted temporal change of an ion acceleration energy normalized to the measured ion beam current density, dE / dt set / I B,meas is equal to the so-called reference input or a target value, i.e., dE / dt target / I B,expected, I B,expected. For this reason, the closed loop control unit functions as a feedback control unit. For example, assuming that the ion beam current density falls due to fluctuations during an ion implantation process, the controller will counteract this drop by decreasing the set point of dE / dt, thereby bringing the actual time change of the ion acceleration energy normalized to the measured ion beam current density, i.e., dE / dt set / I B,meas, closer to the targeted time change of the ion acceleration energy normalized to the ion beam current density, dE / dt (I B,norm)set= dE / dt target / I B,expected, I B,expected.The control process therefore enables to counteract variations in ion beam current density, e.g. proton beam current density, while realizing a target doping concentration profile in the semiconductor body based on a single ion implantation recipe. This is advantageous not only in terms of achieving a desired local implantation dose, i.e., an implantation dose at a particular implantation depth, regardless of variations in ion beam current density, but also in terms of simplicity of the ion implantation process by reducing the cost and expense of tuning cycles necessary when the target doping profile is realized by a plurality of separate ion implantation processes at different ion acceleration or ion implantation energies, i.e., wherein each one of the plurality of separate ion implantation processes is based on a separate implantation recipe.In one or more embodiments, the relative movement between the semiconductor substrate and the ion beam incident on the surface of the semiconductor substrate is performed by deflecting the ion beam along a first scan direction and along a second scan direction. The relative movement along the first scanning direction, for example an x-direction, can be carried out by scanning with electrostatic fields. Likewise, the relative movement along the second scanning direction, e.g. a y-direction, can also be carried out by scanning with electrostatic fields.In one or more embodiments, the relative movement between the semiconductor substrate and the ion beam incident on the surface of the semiconductor substrate is performed by deflecting the ion beam along a first scan direction and mechanically moving the semiconductor substrate along a second scan direction. The relative movement along the first scanning direction, for example an x-direction, can be carried out by scanning with electrostatic fields. The relative movement along the second scanning direction, for example a y-direction, may comprise placing the semiconductor substrate on a substrate carrier of a rotating unit and rotating the substrate carrier. A plurality of semiconductor substrates, for example more than 3 or more than 5 or more than 7 or more than 9 or more than 11 and less than 17 or less than 15 or less than 13 wafers, may be placed on the substrate carrier, for example a rapidly rotating disc or wheel.In one or more embodiments, a scan speed of the relative movement between the semiconductor substrate and the ion beam along each of the first scan direction and the second scan direction ranges from 10 m / s to 30 km / s. Therefore, regardless of whether the relative movement between the semiconductor structure and the ion beam is performed by ion beam deflection along both scan directions or by a combination of ion beam deflection and fast mechanical scanning, a fast scan speed is achieved compared to rather slow scan speeds of mechanical scan systems that are in the range of cm / s.In one or more embodiments, the method further comprises, while performing relative movement between the semiconductor substrate and the ion beam incident on the surface of the semiconductor substrate, adjusting input parameters of units for focusing or deflecting the ion beam based on a function where n is an integer greater than two. The function f(E) may also include half-integer exponents, for example, where n and m are integers greater than two. For each ion implantation or ion acceleration energy E(t) set while passing through the window ΔE=E 2- E 1 of the ion acceleration energy, an input parameter p m, where m is an identifier of the input parameter, for example, a voltage or current of a focusing or deflecting unit of an ion implantation device, is set by the function. Thus, the input parameters p m of the focusing or deflecting units are continuously updated during the ion implantation process, which is based on a single implantation recipe.In one or more embodiments, the ion beam corresponds to a proton beam. This may make it possible, for example, to set doping concentration profiles of hydrogen-based donors by proton irradiation and annealing with profiles in which a full width at half maximum FWHM of at least one peak in the concentration profile of the hydrogen-based donors is greater than 5 μm, or even greater than 7 μm, or even greater than 10 μm. Full width at half maximum, FWHM, is a commonly used parameter to describe the width of a "bump" on a curve or function such as the dopant concentration profiles of hydrogen-based donors. It is determined by the distance between points on the curve at which the function reaches half its maximum value, i.e. at which the doping concentration of the hydrogen-based donors reaches half its peak value. The distance refers to a vertical distance along a vertical direction, e.g. along an ion implantation direction or perpendicular to a main surface of the semiconductor substrate. In one or more embodiments, drift zones or field stop zones may be produced having hydrogen-based donor profiles with a full width at half maximum FWHM of at least 1 peak, or multiple peaks or all peaks in the hydrogen-based donor concentration profile that are greater than 5 μm, or greater than 7 μm, or even greater than 10 μm. Profiles produced by proton implantation and healing at a plurality of discrete proton implantation energies typically result in peaks in the end of range or the penetration depth, which have an FWHM of less than 5 μm, or even less than 4 μm, or even less than 3 μm.In one or more embodiments, changing an ion acceleration energy of an ion beam while performing relative movement between a semiconductor substrate and an ion beam incident on the semiconductor substrate is performed with respect to an ion beam incident on a first side of the semiconductor substrate and with respect to an ion beam incident on a second side of the semiconductor substrate, the second side being opposite the first side. This may allow for a precise adjustment of a doping concentration and a doping concentration profile in a drift zone of the semiconductor device, for example.In one or more embodiments, the method 1000 or the method 1001 further comprises reducing a thickness of the semiconductor substrate to a target thickness by removing material of the semiconductor substrate from a surface of the semiconductor substrate on the second side. In addition, the method 1000 or the method 1001 comprises setting a maximum acceleration energy of the ion beam impinging on the second side as a function of the target thickness. This can, for example, prevent an overlap of concentration profiles of dopants introduced from opposite sides of the semiconductor substrate. In one or more embodiments, the ion beam incident on the first side results in a first proton- or hydrogen-based donor profile in the semiconductor substrate and the ion beam incident on the second side results in a second proton- or hydrogen-based donor profile in the semiconductor substrate, wherein a vertical distance between the first and the second proton profile is less than 5 μm, or a vertical overlap between the first and the second proton profile is less than 5 μm. The embodiment is illustrated in the graphical representation of FIG. 3 by depicting a profile of hydrogen-based donors over a vertical direction. The ion beam incident on the first side results in a first hydrogen-based donor profile p 1 and the ion beam incident on the second side results in a second hydrogen-based donor profile p 2. A vertical distance d between the first profile p 1 and the second profile p 2 is less than 5 μm. The profiles p 1 and p 2 can, for example, likewise overlap up to a vertical overlap of less than 5 μm. A full width at half maximum, FWHM, of a peak PK in the concentration profile p 2 of the hydrogen-based donors is less than 5 μm.In one or more embodiments, a drift zone is formed by changing an ion acceleration energy of an ion beam while performing relative movement between a semiconductor substrate and an ion beam incident on the semiconductor substrate.In one or more embodiments, changing an ion acceleration energy of an ion beam while performing a relative movement between the semiconductor substrate and an ion beam incident on the semiconductor substrate is performed after performing all further process steps for introducing dopants into the semiconductor substrate. Thus, the ion beam may be used as a final doping process of the semiconductor substrate, e.g. in view of the thermal budget limitations when doping with hydrogen-based donors.In one or more embodiments, after changing an ion acceleration energy of an ion beam while performing relative movement between the semiconductor substrate and an ion beam incident on the semiconductor substrate, a thermal heating process of the semiconductor substrate is limited to temperatures up to 420° C. and a total duration of up to 4 h. The temperatures can likewise be limited to smaller values, e.g. 400° C., or 380° C. or 350° C. The total duration can likewise be limited to smaller values, e.g. 60 min, or 600 s, or 60 s. In one or more embodiments, the thermal budget defined by a product of heating temperature and duration may be limited to values up to 1600° C., or up to 400° C., or up to 200° C.The above embodiments enable, for example, an improved adjustment of doping profiles based on proton implantation. Since the proton implantation at varying ion acceleration energies of the proton beam can dispense with subsequent diffusion processes of hydrogen, the thermal budget can be lowered, e.g. to minimum annealing temperatures in the range of 350° C. to 380° C. and to minimum annealing durations in the range of 10 min to 60 min or even in the range of seconds, which is typical for rapid thermal annealing (RTP). Thus, thermal annealing parameters may be selected in consideration of requirements of i) long term stability caused by heating during an operating state of the finalised device, and ii) maximum tolerable concentrations of radiation defects stable at low temperatures, e.g. in consideration of comparatively low annealing temperatures and proton implantation from opposite sides of the semiconductor substrate, and in addition target doping concentrations may be produced by comparatively low proton implantation doses in consideration of unnecessary diffusion of hydrogen and vacancies.The methods 1000 and 1001 can additionally enable, for example, better predictable final doping. In view of the reduction of the thermal budget for healing / activating hydrogen-based donors based on proton implantation described with reference to the methods 1000, 1001, semiconductor devices of higher voltage classes, e.g. 1700 V voltage class IGBTs, may be manufactured because the formation of oxygen-based donors substantially occurring in the temperature range between 400° C. and 500° C. may be effectively prevented. If a base doping is formed in the semiconductor body by proton irradiation, e.g. drift zone doping, a first proton irradiation with varying ion acceleration energy may first be carried out through a surface on the first side of the semiconductor substrate before thinning. In addition, a second proton irradiation with varying ion acceleration energy may be performed through a surface on the second side of the semiconductor substrate after thinning the semiconductor substrate to a target thickness. Both proton exposures may be made based on maximum ion acceleration energies less than 2.6MeV, where the value 2.6MeV is a threshold for neutron generation. As a result, radiation protection measures can be dispensed with, for example.Referring to the schematic graph of FIG. 4, schematic curves r 1, r 2 of a temporal change of the ion acceleration energy normalized to the ion beam current density, dE / dt / I B( t) versus ion acceleration energy with respect to the window ΔE=E 2- E 1 of the ion acceleration energy are illustrated, wherein values of an initial ion acceleration energy E 1 and a final ion acceleration energy E 2 may be determined depending on the target doping concentration versus depth D L( x p), to be formed. The change over time of the ion acceleration energy normalized to the measured ion beam current density, dE / dt / I B( t), may correspond to the reference input of the controller, i.e., dE / dt target / I B,expected. The curve r 1 may be associated with a target doping profile with a steep decrease of the doping concentration at a depth corresponding to a projected range at an ion implantation or ion acceleration energy Ex, whereas the curve r 2 may be associated with a target doping profile with a minimum of the doping concentration that is between a maximum of a doping concentration and a surface where dopants enter the semiconductor substrate by means of the ion implantation process.Referring to the schematic diagrams of FIGS. 5A and 5B, a closed loop control process of the time variation of the ion acceleration energy dE / dt / I B( t) normalized to the measured ion beam current density is illustrated. At process phase 101, a target doping concentration profile D L( x) is specified, for example, by a semiconductor process simulation based on TCAD (Technology Computer Aided Design) simulation instruments. The target doping concentration profile D L( x) may then be input to a human machine interface (HMI) or conversion unit 102, which is illustrated in more detail in FIG. 5B. The HMI unit 102 converts the target anti-depth doping concentration D L( x) to a target anti-energy doping concentration D L( E) and the target anti-energy doping concentration D L( E) to a temporal ion acceleration or ion implantation energy dE / dt by considering a functional relationship E(x) between the implantation energy E and a corresponding projected range x. This conversion by the HMI unit 102 may be based on dopant-specific experimental and / or simulation data of profile parameters such as projected range, longitudinal spread (strigle), skewness and curvature, for example. The HMI unit 102 outputs a targeted temporal change of the ion acceleration energy normalized to the ion beam current density, V Etarget= dE / dt target / I B,expected which is input as a reference input to a closed loop control unit 103. An automatic controller unit 1031, which may include an error detector and an amplifier, compares the actual value of the change over time of the ion acceleration energy normalized to the measured ion beam current density, i.e., V Efeedback= dE / dt set / I B,meas, with the reference input, i.e., the target value V Etarget= dE / dt target / I B,expected, determines the deviation, and generates a control signal c that will reduce the deviation between V Efeedback and V Etarget. If V Ffeedback is greater than V Etarget the control signal may cause a decrease in dE / dt to increase a local dose at a particular implantation depth. Similarly, if V Efeedback is less than V Etarget the control signal c may cause an increase of dE / dt to decrease a local dose at a particular implantation depth. The control signal c of the automatic controller unit 1031 is output to an actuator unit 1032, which determines the temporal change of the ion acceleration or ion implantation energy dE / dt according to the control signal c so that V Efeedback will approach V Etarget. The change over time of the ion acceleration energy dE / dt is the output from the actuator unit 1032 to an integration unit 1033 that determines the actual ion acceleration energy E(t). The actual ion acceleration energy E(t) is the output from the integration unit 1033 to units of an ion implantation device part 1034 that are operated based on the ion acceleration energy E(t), for example, an acceleration unit(s), a focusing unit(s), or a deflection unit(s). A feedback control unit 1035 receives as an input a measured value of the ion beam current density I B,meas( t). The ion beam current density I B,meas( t) can be measured by an ion current detector unit, for example, a Faraday cup, which is part of the ion implantation apparatus part 1034. The feedback control unit 1035 further receives the actual time change of the ion acceleration energy dE set / dt and determines V Efeedback. The value V Efeedback is then supplied to the automatic controller unit 1031, thereby closing the feedback loop.FIG. 5C is a graph for illustrating a relationship between a temporal change in ion acceleration energy normalized to the ion beam current density, dE / dt / I B( t) vs. ion acceleration energy and the doping concentration vs. ion acceleration energy D L( E). The illustration is simplified because it is based on, among other things, a linear relationship between the projected range x and the ion acceleration energy E.FIG. 6 shows an ion implantation apparatus 200 including an ion source 205 that generates and emits ions, for example protons, helium, or ions having an atomic number greater than 4, for example ions of nitrogen, aluminum, boron, phosphorus, arsenic, sulfur, selenium, germanium, or oxygen. An accelerating unit 220 may accelerate a selected type of ions and may filter out others. A collimator unit 230 may align the moving directions of the ions in a direction parallel to a beam axis and may direct a collimated ion beam 235 onto a semiconductor substrate 240, which may be temporarily fixed to a substrate carrier 245, e.g., by vacuum. In a plane orthogonal to the beam axis, an ion distribution in the collimated ion beam 235 may be point-symmetric to a beam center point.A cross-sectional area of the ion beam 235 may be on the order of several hundred square micrometers to several square centimeters. A scanning assembly 250 scans the ion beam 235 along a beam track across a main surface of the semiconductor substrate 240 to uniformly spread the ions across the semiconductor substrate 240. The beam trace may include straight portions, zig zag patterns, may form circles, may form a spiral, or any other typical scan pattern.The scanning assembly 250 controls the scan by electrostatic fields, wherein the scanning assembly 250 controls a relative movement between the ion beam 235 and the semiconductor substrate 240 along a first scanning direction x and along a second scanning direction y. The first and second scanning directions x, y may be perpendicular to each other, for example. In the example illustrated in FIG. 6, the scanning assembly 250 includes a first deflection subunit 251 for deflecting the ion beam 235 along the first scanning direction x. The ion beam 235 traverses an area between a pair of first deflection electrodes, which deflect the ion beam 235 along the first scanning direction x. Thereafter, the ion beam passes a pair of second deflection electrodes of a second deflection subunit 252 that deflect the ion beam 235 along a linear second scanning direction y, which may be orthogonal to the drawing plane. The electric fields in the first and second deflection subunits 251, 252 let the ion beam 235 sweep over the entire main surface of the semiconductor substrate 240.The ion implantation apparatus 200 further includes a control unit 260 configured to change an ion acceleration energy E of the ion beam 235 during the relative movement between the semiconductor substrate 240 and the ion beam 235 incident on a surface of the semiconductor substrate 240. The controller 260 may control the change in ion acceleration energy normalized to the ion beam current density, dE / dt / I B as described, for example, with reference to the embodiments illustrated above. The control unit 260 may output the ion acceleration or ion implantation energy E(t) to units of the ion implantation apparatus 200, for example, the acceleration unit 220, the collimator unit 230, and the first and second deflection subunits 251, 252. The ion beam current density I B( t) may be measured by one or more ion current detector units, for example, Faraday cups near the semiconductor substrate 240 on the substrate support 245. The measured ion beam current density I B,meas( t) is output to the control unit 260.FIG. 7 shows an ion implantation apparatus 201 including the ion source 205, the accelerating unit 220, the collimator unit 230, and the control unit 260 as described with reference to the ion implantation apparatus 200 illustrated in FIG. 6. Unlike the ion implantation apparatus 200 of FIG. 6, the ion implantation apparatus 201 is based on both electrostatic scanning and mechanical scanning. Electrostatic scanning along the first scanning direction x is carried out by the first deflection subunit 251 of the scanning assembly 250. Scanning along the second scanning direction y is carried out by moving the semiconductor substrate 240 with respect to the ion beam 235 along the second scanning direction y, which is caused by a rotation of the substrate carrier 245. The substrate carrier 245 may be part of a rotating unit 253. In one or more embodiments, the substrate carrier 245 is disc-like or wheel-like and configured to hold a number of semiconductor substrates using, for example, vacuum. Rotating the substrate carrier 245 along a radial direction 265 causes a relative movement between the semiconductor substrate 240 and the ion beam 235 along the second scan direction y. A rotational speed of the rotating unit 253 may be in the range of hundreds or thousands of rpm, thereby achieving a fast scan with scan speeds of more than 10 m / s or even more than 100 m / s. Ion current detector units 267 may be disposed on the substrate support 245 to measure the ion beam current density I B,meas( t) output to the control unit 260. The ion current detector units 267 may cross an ion beam path along the first direction x.Referring to the schematic diagrams of FIG. 8, an open loop control process of time-varying the ion acceleration energy dE / dt / I B,set normalized to a set ion beam current density is illustrated in combination with a closed loop control process of the ion beam current density. At process stage 201, a target doping concentration profile D L( x) is specified, for example, by a semiconductor process simulation based on TCAD (Technology Computer Aided Design) semiconductor instruments. The target doping concentration profile D L( x) may then be input to a human machine interface (HMI) or conversion unit 202. The HMI unit 202 converts the target anti-depth doping concentration D L( x) to a target anti-energy doping concentration D L( E) and the target anti-energy doping concentration D L( E) to a temporal change in the ion acceleration or implantation energy dE / dt by considering a functional relationship E(x) between the implantation energy E and a corresponding projected range x. This conversion by the HMI unit 202 may be based on dopant-specific experimental and / or simulation data of profile parameters such as, for example, a projected range, a longitudinal spread, a skewness or curvature. The HMI unit 202 outputs a targeted temporal change of the ion acceleration energy normalized to a set ion beam current density, V E= dE / dt / I B,set which is input to a beam guidance system 250. In the beam guidance system 250, the adjusted ion beam current density passes through a closed loop control process. A comparator unit 2031, which may include an error detector and an amplifier, compares the actual value of the ion beam current density I B,meas with the reference input, i.e. I B,set, determines the deviation and generates an error signal err as the input signal to an automatic controller 2032, e.g. a PID controller, which generates a control signal c according to the error signal err, so that I B,meas will approach I B,set. A beam current regulator unit 2033 receives the control signal c as an input signal for adjusting the beam current. The ion beam current density I B,meas can be measured by an ion current detector unit, for example, a Faraday cup, which is part of the ion implantation apparatus, and input to the comparator unit 2031, thereby achieving the closed loop control of the ion beam current density while controlling the temporal change of the ion acceleration or ion implantation energy dE / dt by an open loop process.The ion implantation method and the ion implantation device described above may be used to fabricate semiconductor devices, for example field effect transistors (FETs), insulated gate bipolar transistors (IGBTs), or diodes, in any type of semiconductor substrate material, for example, silicon (Si), silicon carbide (SiC), silicon germanium (SiGe), germanium (Ge), gallium arsenide (GaAs), gallium nitride (GaN).The methods 1000, 1001 can be used to produce field stop zones in silicon substrates by proton implantation. Instead of multiple ion implantations at different implantation energies based on different implantation formulations, a single ion implantation process may be sufficient. Likewise, drift zone doping in SiC may be realized by a single ion implantation process. In Si semiconductor devices for low voltages, drift zone doping and body doping may be realized by a single ion implantation process, in Si or SiC semiconductor devices for medium and high voltages, a single ion implantation process may be performed between successive layer deposits, for example. In Si and SiC semiconductor devices, superjunction structures with a constant or nearly constant doping concentration versus depth may be realized. Deep doping regions at a back side of a semiconductor substrate may also be realized based on a low heat balance, which may be required for back side processing, for example, in view of limitations of the heat balance due to completed front side structures.According to an example, a semiconductor device includes a semiconductor substrate and a field stop zone in the semiconductor substrate. A doping concentration profile of the field stop zone is determined by a concentration profile of hydrogen-based donors and the concentration profile of the hydrogen-based donors has a full width at half maximum, FWHM, of at least one peak, or of a plurality of peaks or all peaks in the concentration profile of the hydrogen-based donors of greater than 5 μm, or greater than 7 μm, or even greater than 10 μm. Full width at half maximum, FWHM, is a parameter commonly used to describe a width of a "bump" on a curve or function such as the dopant concentration profiles of hydrogen-based donors. It is determined by the distance between points on the curve at which the function reaches half its maximum value, i.e. at which the doping concentration of the hydrogen-based donors reaches half its peak value. The distance refers to a vertical distance along a vertical direction, e.g. along an ion implantation direction or perpendicular to a main surface of the semiconductor substrate.For example, the concentration profile of the field stop zone along a vertical direction is steadily decreasing or constant with respect to any portion of an entire vertical extension of the field stop zone.For example, the semiconductor device further comprises a first load terminal contact on a first surface of the semiconductor substrate, a control terminal contact on the first surface of the semiconductor substrate, and a second load terminal contact on a second surface of the semiconductor substrate.FIG. 9 is a schematic cross-sectional view of a semiconductor device 800. The manufacturing of the semiconductor device may include the above-described methods 1000, 1001. The semiconductor device 800 comprises, on a first surface 801 of a semiconductor substrate 802, a first load terminal L 1, for example a source electrode, and a control terminal C, for example a gate electrode. The semiconductor device 800 further comprises, at a second surface 803 of the semiconductor substrate 802, a second load terminal L 2, for example a drain electrode. The semiconductor device 800 also includes a field stop zone 804. A doping concentration profile in the field stop zone 804 is determined from a concentration profile of hydrogen-based donors. Exemplary profiles in the graphical representation of FIG. 9 include a step profile p 1, a Gaussian-like profile p 2, and a continuously decreasing profile p 3.Although specific embodiments are illustrated and described herein, it will be understood by those skilled in the art that a variety of alternative and / or equivalent configurations may be utilized for the specific embodiments shown and described.

Claims

A method (1000) of manufacturing a semiconductor device (800), the method (1000) comprising: reducing a thickness of a semiconductor substrate (802); and changing an ion acceleration energy of an ion beam while performing a relative movement between the semiconductor substrate (802) and the ion beam incident on the semiconductor substrate (802), wherein changing the ion acceleration energy comprises controlling a temporal change of the acceleration energy dE / dt depending on a value of an ion beam current density I B,meas measured during the relative movement.A method (1001) of manufacturing a semiconductor device (800), the method (1001) comprising: forming a doped region in a semiconductor substrate (802); and thereafter changing an ion acceleration energy of an ion beam while performing a relative movement between the semiconductor substrate (802) and the ion beam incident on the semiconductor substrate (802), wherein changing the ion acceleration energy comprises controlling a temporal change of the acceleration energy dE / dt depending on a value of an ion beam current density I B,meas measured during the relative movement, wherein the ion beam is a proton beam, and wherein changing an ion acceleration energy of an ion beam while performing a relative movement between the semiconductor substrate (802, 240) and the ion beam incident on the semiconductor substrate (802, 240), in relation to an ion beam impinging on a first side of the semiconductor substrate (802, 240) and in relation to an ion beam impinging on a second side of the semiconductor substrate (802, 240), wherein the second side is opposite the first side.The method (1000, 1001) of claim 1 or 2, further comprising: converting a target dopant concentration versus depth D L( x p), where x p( E) is a projected range of dopants as a function of ion acceleration energy, into a targeted temporal change in ion acceleration energy normalized to the ion beam current density, dE / dt(E) (I B,norm)set= dE / dt(E) / I B,expected, where I B,expected is an expected ion beam current density.The method (1000, 1001) of claim 1 or 2, wherein controlling the time change of the acceleration energy dE / dt in dependence on a value of the measured ion beam current density I B,meas further comprises comparing, at an ion acceleration energy E, an targeted time change of the ion acceleration energy normalized to the ion beam current density dE / dt (I B,norm)set= dE / dt target / I B,expected, where I B,expected is an expected ion beam current density, with a set time change of the ion acceleration energy normalized to I B,meas, dE / dt set / I B,meas.The method (1000, 1001) according to any of the preceding claims, wherein the relative movement between the semiconductor substrate (802, 240) and the ion beam (235) impinging on the semiconductor substrate (802, 240) is carried out by deflecting the ion beam (235) along a first scanning direction (x) and along a second scanning direction (y).The method (1000, 1001) according to any one of claims 1 to 4, wherein the relative movement between the semiconductor substrate (802, 240) and the ion beam impinging on the semiconductor substrate (802, 240) is performed by deflecting the ion beam (235) along a first scanning direction (x) and by mechanically moving the semiconductor substrate (802, 240) along a second scanning direction (y).The method (1000, 1001) of claim 6, wherein mechanically moving the semiconductor substrate (802, 240) along the second scan direction comprises placing the semiconductor substrate (802, 240) on a substrate carrier (245) of a rotating unit (253) and rotating the substrate carrier (245).The method (1000, 1001) of any one of claims 5 to 6, wherein a scanning speed of the relative movement between the semiconductor substrate (802, 240) and the ion beam (235) along each of the first scanning direction and the second scanning direction ranges from 10 m / s to 30 km / s.The method (1000, 1001) according to any one of claims 6 to 8, further comprising, while performing relative movement between the semiconductor substrate (802, 240) and the ion beam (235) incident on the semiconductor substrate (802, 240), setting input parameters of units for focusing or deflecting the ion beam (235) based on a function f ( E ) = ∑ i = 0 n a i E i, where n is an integer greater than two.The method of claim 2, further comprising: reducing a thickness of the semiconductor substrate (802, 240) to a target thickness by removing material of the semiconductor substrate (802, 240) from a surface of the semiconductor substrate (802, 240) on the second side; and adjusting a maximum acceleration energy of the ion beam incident on the second side depending on the target thickness.The method according to any one of claims 2 or 10, wherein the ion beam impinging on the first side results in a first proton profile in the semiconductor substrate (802, 240) and the ion beam impinging on the second side results in a second proton profile in the semiconductor substrate (802, 240), and wherein a vertical distance (d) between the first and the second proton profile is less than 5 μm, or a vertical overlap between the first and the second proton profile is less than 5 μm.The method according to any one of claims 2 or 10 to 11, wherein a field stop zone or a drift zone is formed by changing an ion acceleration energy of an ion beam while performing relative movement between the semiconductor substrate and the ion beam incident on the semiconductor substrate.The method of any of claims 2 or 10 to 12, wherein changing an ion acceleration energy of an ion beam while performing relative movement between the semiconductor substrate (802, 240) and the ion beam incident on the semiconductor substrate (802, 240) is performed after performing all process steps for introducing dopants into the semiconductor substrate (802, 240).The method according to any one of claims 2 or 10 to 13, wherein, after changing an ion acceleration energy of an ion beam while performing relative movement between the semiconductor substrate (802, 240) and the ion beam incident on the semiconductor substrate (802, 240), a thermal heating process of the semiconductor substrate (802, 240) is limited to temperatures up to 420°C and a total time up to 4h.The method according to any of the preceding claims, wherein changing the ion acceleration energy comprises controlling a temporal change of the acceleration energy dE / dt depending on a value of the total implantation dose D int.

Images