Laser annealing method

Abstract

To provide a laser annealing method capable of improving a carrier density in a deep region of a semiconductor wafer without performing ion implantation of a dopant in the deep region and without fusing a surface layer.SOLUTION: Activation annealing is performed by scanning a surface of a semiconductor wafer, in which a dopant is implanted, with a laser beam for annealing that is a pulse laser beam in a repetition frequency 100 kHz or higher or a continuous wave laser beam. First, when a beam spot passes any fine region in a region to be annealed of the semiconductor wafer, a laser energy cumulative value to be charged into the fine region of the semiconductor wafer is changed, and an activation rate maximization annealing condition to maximize an activation rate of the dopant is determined. On the condition that the laser energy cumulative value to be charged into the fine region is smaller than a laser energy cumulative value on the activation rate maximization annealing condition, the activation annealing of the semiconductor wafer in which the dopant is implanted is performed.SELECTED DRAWING: Figure 7

Description

[Technical Field]

[0001] The present invention relates to a laser annealing method. [Background technology]

[0002] In recent years, there has been a demand for a technology that can add dopants to deeper regions than conventional methods in order to improve the performance of power semiconductor devices. A known method involves shallowly implanting dopants into the surface layer of a semiconductor wafer, then melting the surface layer deeper than the dopant implantation region, and then liquid-phase diffusing the dopant to the melt depth (Patent Document 1). This method can increase the carrier density up to the melt depth to a desired value. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Laid-Open No. 2013-30797 Summary of the Invention [Problem to be solved by the invention]

[0004] For example, when manufacturing semiconductor devices such as insulated gate bipolar transistors (IGBTs), a device structure including a dopant diffusion region is formed on the back surface of the semiconductor wafer (the surface opposite to the melted surface layer) before the surface layer of the semiconductor wafer is melted. If the surface layer is melted to add dopants to deep regions, the back surface of the semiconductor wafer will also become hot, damaging the device structure that has already been formed.

[0005] The dopant can be added to the deep region by ion implanting the dopant into the deep region and then performing activation annealing. To implant ions into the deep region, the ion acceleration energy must be increased, which leads to an increase in the size of the ion implantation equipment.

[0006] An object of the present invention is to provide a laser annealing method that can increase the carrier density in a deep region of a semiconductor wafer without ion-implanting a dopant into the deep region and without melting the surface layer. [Means for solving the problem]

[0007] According to one aspect of the present invention, A laser annealing method for performing activation annealing by scanning a surface of a semiconductor wafer into which a dopant has been implanted with an annealing laser beam, which is a pulsed laser beam or a continuous wave laser beam having a repetition rate of 100 kHz or more, comprising: determining activation rate maximizing annealing conditions that maximize the activation rate of dopants by varying the cumulative value of laser energy applied to an arbitrary micro-region of the semiconductor wafer when the beam spot passes through the said micro-region; There is provided a laser annealing method for performing activation annealing on a semiconductor wafer implanted with dopants under conditions in which the cumulative value of the laser energy input to the minute region is smaller than the cumulative value of the laser energy under the activation rate maximizing annealing conditions. [Effects of the Invention]

[0008] When annealing is performed under conditions where the cumulative value of the laser energy applied to a micro-region is smaller than the cumulative value of the laser energy under the activation rate maximizing annealing conditions, the implanted dopant diffuses to a deep region, thereby increasing the carrier density in the deep region. [Brief explanation of the drawings]

[0009] [Figure 1] FIG. 1 is a schematic perspective view of a laser annealing apparatus used in one embodiment. [Figure 2] FIG. 2 is a schematic diagram showing the movement history of the beam spot 20 on the surface of the semiconductor wafer 30. As shown in FIG. [Figure 3]FIG. 3 is a graph showing the measurement results of the dopant and carrier density distributions before and after activation annealing. [Figure 4] FIG. 4 is a graph showing the measurement results of the dopant density distribution before and after activation annealing, and the carrier density distribution when the moving speed of the beam spot is changed. [Figure 5] 5A to 5C are graphs showing the dopant density before and after annealing, and the carrier density after annealing. [Figure 6] FIG. 6A is a schematic diagram showing the crystalline state of a silicon wafer after dopant is implanted into the silicon wafer, FIG. 6B is a schematic diagram showing the crystalline state of the silicon wafer after annealing under activation rate maximization annealing conditions, and FIG. 6C is a schematic diagram showing the crystalline state of the silicon wafer after annealing under conditions with a lower cumulative laser energy value than the activation rate maximization annealing conditions. [Figure 7] FIG. 7 is a flowchart showing the procedure of the annealing method according to this embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0010] A laser annealing method according to one embodiment will be described with reference to FIGS. FIG. 1 is a schematic perspective view of the laser annealing apparatus used in this embodiment. A laser oscillator 10 outputs a pulsed laser beam with a high repetition rate. For example, a repetition rate of 100 kHz or higher can be called a "high repetition rate." As the laser oscillator 10, for example, a fiber laser oscillator that outputs a pulsed laser beam in the green wavelength range is used.

[0011] A pulsed laser beam output from a laser oscillator 10 passes through a beam expander 11, a folding mirror 12, an attenuator 19, a diffractive optical element 13, a galvanometer scanner 14, an fθ lens 15, and an epi-mirror 16 before being incident on a semiconductor wafer 30. The semiconductor wafer 30 is fixed by suction to a movable table 18. The movable table 18 is moved in one horizontal direction by a linear motion mechanism 17. An xyz Cartesian coordinate system is defined, with the y-axis parallel to the direction of movement of the movable table 18 and the z-axis pointing vertically upward.

[0012] The attenuator 19 attenuates the power of the laser beam. The amount of attenuation of the laser beam by the attenuator 19 is variable. The diffractive optical element 13 shapes the shape and beam profile of the beam spot on the surface of the semiconductor wafer 30.

[0013] The galvanometer scanner 14 scans the pulsed laser beam in one direction. When the pulsed laser beam is scanned, the beam spot moves in the x direction on the surface of the semiconductor wafer 30. By combining the scanning of the pulsed laser beam in the x direction by the galvanometer scanner with the movement of the movable table 18 in the y direction by the linear motion mechanism 17, the pulsed laser beam can be incident on almost the entire surface of the semiconductor wafer 30.

[0014] The control unit 50 controls the laser oscillator 10, the galvanometer scanner 14, and the linear motion mechanism 17. The control unit 50 controls the galvanometer scanner 14, thereby adjusting the movement speed of the beam spot in the x-direction. The average power of the pulsed laser beam output from the laser oscillator 10 can be changed by control from the control unit 50. For example, the average power can be changed by changing at least one of the pulse repetition frequency, peak power, and pulse width. In this embodiment, the peak power of the pulsed laser beam is changed by changing the attenuation amount of the attenuator 19.

[0015] Next, the movement of the beam spot will be described with reference to Fig. 2. Fig. 2 is a schematic diagram showing the movement history of the beam spot 20 on the surface of the semiconductor wafer 30. As an example, the shape of the beam spot 20 is a rectangle that is long in the y direction, with a length L in the y direction of 200 µm and a width W in the x direction of 100 µm. The light intensity distribution is approximately flat in the y direction and a Gaussian distribution in the x direction.

[0016] If the repetition frequency of the laser pulse is denoted as f, the average power of the pulsed laser beam as P, and the pulse energy as E, then E=P / f (1) When the beam spot 20 is scanned in the x direction, the moving speed in the x direction is denoted as v and the width of the beam spot 20 in the x direction is denoted as W, and the overlap ratio OVP is expressed as follows: OVP=(Wv / f) / W (2) It is expressed as:

[0017] The cumulative laser energy Ea applied to a minute region of area dS on the surface of semiconductor wafer 30 is expressed as Ea=(E / (L×W))×(1 / OVP)×dS. That is, under the condition that the length L of the long side and the width W of beam spot 20 are constant, the cumulative laser energy Ea applied to minute region dS on the surface of semiconductor wafer 30 can be adjusted by changing any of the average power P, repetition frequency f, and movement speed v.

[0018] Next, with reference to Figures 3 to 5C, the results of performing activation annealing under various annealing conditions on a semiconductor wafer into which a dopant has been ion-implanted will be described. A silicon wafer was used as the semiconductor wafer, and phosphorus (P) was used as the dopant. The ion-implantation conditions for phosphorus ions are as follows: Acceleration energy: 700 keV Dose: 1×10 13 cm -2

[0019] 3 is a graph showing the measurement results of the dopant density and carrier density distribution before and after activation annealing. The horizontal axis represents depth in units of μm, and the vertical axis represents density in units of cm -3 ]. The dashed line a in the graph of Figure 3 indicates the dopant density before annealing. Secondary ion mass spectrometry (SIMS) was used to measure the dopant density. The solid line b indicates the carrier density after annealing at the average power at which the activation rate is maximized without melting the surface layer of the semiconductor wafer (the average power at this time is defined as 100%). The solid lines c and d indicate the carrier density after annealing at average powers of 90% and 80%, respectively. An average power of 80 W corresponds to 100%. The spreading resistance analysis (SRA) was used to measure the carrier density.

[0020] The annealing conditions other than the average power are as follows: Pulse repetition frequency: 1500kHz Pulse width: 2ns Beam spot length L and width W: 200 μm and 100 μm Beam spot movement speed in x direction: 2.0 m / s Beam profile: Flat in length, Gaussian in width

[0021] When the average power is 80 W, the pulse energy E is approximately 53.3 μJ according to equation (1). When the average power is set to 90% and 80%, the pulse energy is approximately 48 μJ and 42.7 μJ, respectively. The overlap ratio OVP is approximately 98.7% according to equation (2).

[0022] When the average power is 100%, the dopant activation rate is nearly 100%. For example, at the depth where the dopant density is maximum (hereinafter referred to as the "implantation depth"), the dopant density (dashed line a) and the carrier density (solid line b) are nearly equal. Although the carrier density is lower than the dopant density in the region deeper than approximately 1.5 μm, the density at this depth is less than 1 / 100 of the density at the depth where the dopant density is maximum, so this has almost no effect on the activation rate.

[0023] When the average power is set to 90% and 80%, as the average power decreases, the carrier density at the implantation depth decreases, while the carrier density in deeper regions increases. Because the carrier density at the implantation depth is lower than the dopant density, the activation rate is lower than 100%.

[0024] When a pulsed laser beam is scanned, it is incident multiple times on any minute region (any one location) within the region of the semiconductor wafer to be annealed. The total amount of laser energy (hereinafter referred to as the cumulative laser energy value) input into this minute region by multiple incidences of the pulsed laser beam affects the annealing results. Reducing the average power of the pulsed laser beam reduces the cumulative laser energy value input into the minute region. As shown in Figure 3, reducing the cumulative laser energy value increases the carrier density in regions deeper than the implantation depth.

[0025] The accumulated laser energy value also changes when the moving speed of the beam spot is changed under the condition that the average power is constant. Next, with reference to Figures 4 to 5C, we will explain how the carrier density changes when the accumulated laser energy value is changed by changing the moving speed of the beam spot.

[0026] Figure 4 is a graph showing the results of measuring the dopant density distribution before and after activation annealing, and the carrier density distribution when the beam spot moving speed is changed. The horizontal axis represents depth in units of μm, and the vertical axis represents density in units of cm.-3 In the graph of Figure 4, the dashed line a indicates the dopant density before annealing. The solid lines b, c, and d indicate the carrier density after annealing when the beam spot movement speeds were set to 2.0 m / s, 2.5 m / s, and 3.0 m / s, respectively.

[0027] The annealing conditions other than the beam spot moving speed are as follows: Average power: 80W Pulse repetition frequency: 1500kHz Pulse width: 2ns Beam spot length L and width W: 200 μm, 100 μm Under these conditions, the pulse energy E is approximately 53.3 μJ according to equation (1). When the movement speed is set to 2.0 m / s, 2.5 m / s, and 3.0 m / s, the overlap ratio OVP is approximately 98.7%, 98.3%, and 98.0%, respectively, according to equation (2). When the movement speed is increased, the number of laser pulses incident on any minute area decreases, and therefore the cumulative laser energy value decreases.

[0028] A beam spot movement speed of 2.0 m / s corresponds to the condition where the dopant activation rate is maximized. At this time, the activation rate is nearly 100%, as shown by the solid line b. As the movement speed is increased from 2.0 m / s to 2.5 m / s and 3.0 m / s, the carrier density at the implantation depth decreases, while the carrier density in deeper regions increases.

[0029] 5A to 5C are graphs showing the distribution of dopant density before and after annealing, and the distribution of carrier density after annealing. The horizontal axis represents depth in units of μm, and the vertical axis represents density in units of cm -35A to 5C show the dopant density before annealing, the dopant density after annealing, and the carrier density after annealing, respectively. 5A, 5B, and 5C show the densities when annealing was performed with the beam spot moving speeds set to 2.0 m / s, 2.5 m / s, and 3.0 m / s, respectively.

[0030] When annealing was performed under conditions that maximized the activation rate (a travel speed of 2.0 m / s), the dopant density at the implantation depth barely changed, indicating that annealing caused little dopant diffusion. In contrast, when the beam spot travel speed was slowed to 2.5 m / s (Figure 5B) or 3.0 m / s (Figure 5C), the dopant density at the implantation depth decreased, as shown by the solid line Da, indicating that annealing caused the dopant to diffuse in the depth direction. The carrier density also changed in response to the change in dopant density.

[0031] As shown in Figures 3 to 5C, when the cumulative value of laser energy input to an arbitrary microscopic region on the surface of a semiconductor wafer is reduced below the condition that maximizes the activation rate, the carrier density in a region deeper than the implantation depth increases. The reason for this will be explained next with reference to Figures 6A to 6C. The annealing conditions that maximize the activation rate will be referred to as "activation rate maximizing annealing conditions."

[0032] 6A is a schematic diagram showing the crystalline state of a silicon wafer after dopant implantation into the silicon wafer. A crystal lattice is formed by a plurality of silicon atoms 40. When dopant ions are implanted, the implanted dopant atoms 41 do not replace silicon atoms 40 at lattice positions within the silicon crystal, but become interstitial atoms.

[0033] 6B is a schematic diagram showing the crystalline state of a silicon wafer after annealing under the activation rate maximizing annealing conditions. The laser energy input by laser irradiation is converted into thermal energy, and the applied thermal energy causes dopant atoms 41 to replace silicon atoms 40 at their lattice positions. This activates the dopants.

[0034] 6C is a schematic diagram showing the crystalline state of a silicon wafer after annealing under conditions with a lower cumulative laser energy value than the activation rate maximization annealing conditions. Because the thermal energy applied to the crystal is lower than when annealing is performed under the activation rate maximization annealing conditions, some interstitial dopant atoms 41 diffuse in the depth direction before being replaced with silicon atoms 40 at lattice positions. Some of the dopant atoms that diffuse deep into the depths are replaced with silicon atoms 40 at lattice positions.

[0035] 5B and 5C, the fact that the dopant atoms 41 have diffused in the depth direction is inferred from the fact that the dopant density after annealing is higher than that before annealing in a region deeper than the implantation depth. Furthermore, the fact that the dopant atoms 41 diffused into a region deeper than the implantation depth have substituted for silicon atoms 40 at lattice positions is inferred from the fact that the carrier density also increases in accordance with the increase in dopant density in the deep region, as shown in FIGS.

[0036] Next, the procedure of the annealing method according to this embodiment will be described with reference to Fig. 7. Fig. 7 is a flow chart showing the procedure of the annealing method according to this embodiment.

[0037] First, a plurality of semiconductor wafers, for example, silicon wafers, into which dopants have been ion-implanted are prepared (step S1). One annealing condition is set from among a plurality of annealing conditions with different cumulative laser energy values ​​(step S2). Evaluation activation annealing of the semiconductor wafers is performed under the set annealing condition (step S3).

[0038] If there are any annealing conditions under which evaluation activation annealing has not been performed, the annealing conditions are changed (steps S4 and S5), and evaluation activation annealing is performed on the unannealed semiconductor wafers under the changed annealing conditions (step S3).If evaluation activation annealing has been performed under all annealing conditions, the activation rates of multiple annealed semiconductor wafers are measured, and the annealing conditions under which the activation rate is maximized (activation rate maximizing annealing conditions) are determined (step S6).

[0039] Next, the annealing conditions under which the cumulative laser energy value is smaller than the cumulative laser energy value under the activation rate maximization annealing conditions are determined as the annealing conditions for production (step S7). At this time, the annealing conditions under which the desired carrier density distribution is obtained by the activation annealing for evaluation can be determined as the annealing conditions for production. Thereafter, activation annealing of the semiconductor wafer is performed under the determined annealing conditions for production (step S8).

[0040] Next, the excellent effects of this embodiment will be described. In this embodiment, the dopant density can be increased in a region deeper than the implantation depth of the dopant compared to when non-melt activation annealing is performed under annealing conditions that maximize the activation rate. Therefore, the dopant can be efficiently added to a region deeper than the implantation depth of the dopant. In other words, the dopant can be added to a deep region without increasing the acceleration energy of the ion implantation device.

[0041] To achieve a sufficient effect of doping the dopant into the deep region, it is preferable to set the average power of the pulsed laser beam to 90% or less of the average power under the activation rate maximizing annealing conditions, and to achieve a sufficient effect of doping the dopant into the deep region, it is preferable to set the moving speed of the beam spot to 1.25 times or more of the moving speed under the activation rate maximizing annealing conditions.

[0042] Furthermore, since there is no need to heat the surface layer of the semiconductor wafer to a temperature equal to or higher than the melting point of the semiconductor wafer, damage to the element structure formed on the opposite side of the semiconductor wafer from the surface to be annealed is reduced.

[0043] Next, various modifications of the above embodiment will be described. In the above embodiments, a silicon wafer is used as an example of the semiconductor wafer, but a semiconductor wafer other than a silicon wafer may be used. For example, a compound semiconductor wafer may be used. Furthermore, in the above embodiments, phosphorus (P) is used as an example of the dopant, but dopants other than phosphorus, such as arsenic (As) or boron (B), may be used.

[0044] In the above examples, a pulsed laser beam with a high repetition rate was used as the annealing laser beam, but a continuous wave laser beam (CW laser beam) may also be used. When a continuous wave laser beam is used, the cumulative laser energy value can be adjusted by changing the laser beam power, the beam spot movement speed, etc. Furthermore, in the above examples, a laser beam in the green wavelength range was used, but it is also possible to use a laser beam in a wavelength range other than the green wavelength range that is absorbed by the semiconductor wafer to be annealed. Furthermore, in the above examples, a fiber laser oscillator was used as the laser oscillator, but other laser oscillators, such as a laser diode or a solid-state laser such as an Nd:YAG laser, may also be used.

[0045] In the above embodiment, as shown in FIG. 1, the galvanometer scanner 14 is used to move the beam spot in the x direction, and the linear motion mechanism 17 is used to move it in the y direction. However, the galvanometer scanner may be used for both the x and y direction movements.

[0046] The above-described embodiments are merely examples, and the present invention is not limited to the above-described embodiments. For example, it will be obvious to those skilled in the art that various modifications, improvements, combinations, etc. are possible. [Explanation of symbols]

[0047] 10 Laser oscillator 11 Beam Expander 12 Folding mirror 13 Diffractive optical elements 14 Galvanometer Scanner 15 fθ lens 16 Reflection mirror 17 Linear motion mechanism 18 Movable Table 19 Attenuator 20 beam spots 30 Semiconductor wafers 40 silicon atoms 41 dopant atoms 50 control section

Claims

1. A laser annealing method for performing activation annealing by scanning a surface of a semiconductor wafer into which a dopant has been implanted with an annealing laser beam, which is a pulsed laser beam or a continuous wave laser beam having a repetition rate of 100 kHz or more, comprising: determining activation rate maximizing annealing conditions that maximize the activation rate of dopants by varying the cumulative value of laser energy applied to an arbitrary micro-region of the semiconductor wafer when the beam spot passes through the said micro-region; A laser annealing method for performing activation annealing on a semiconductor wafer implanted with dopants under conditions in which the cumulative value of laser energy input to the minute region is smaller than the cumulative value of laser energy under the activation rate maximizing annealing conditions.

2. 2. The laser annealing method of claim 1, wherein the annealing laser beam is a pulsed laser beam, and the cumulative laser energy value when performing the activation annealing is reduced by making the average power of the pulsed laser beam smaller than the average power under the activation rate maximization annealing conditions.

3. 3. The laser annealing method according to claim 2, wherein an average power when performing said activation annealing is set to 90% or less of the average power under said activation rate maximizing annealing conditions.

4. 2. The laser annealing method according to claim 1, wherein the cumulative laser energy value during the activation annealing is reduced by increasing the moving speed of the beam spot of the annealing laser beam on the surface of the semiconductor wafer.

5. 5. The laser annealing method according to claim 4, wherein the moving speed of the beam spot when performing the activation annealing is set to 1.25 times or more the moving speed under the activation rate maximizing annealing conditions.

Images