JP5610493B2 - Semiconductor device manufacturing method and manufacturing apparatus - Google Patents

Description

  The present invention relates to a method and an apparatus for manufacturing a semiconductor device in which a dopant (impurity) is implanted into a semiconductor and activated.

In manufacturing a semiconductor device, a necessary dopant is implanted into a semiconductor, and ion implantation is generally performed. In this ion implantation, an activation process is performed to repair the deviation of the dopant from the crystal lattice position accompanying the ion implantation, and a method using laser light irradiation for the activation process is also known.
In a power device IGBT or the like, a pn junction including an n buffer (field stop) layer and a shallower p layer is formed on the collector side on the back side of the device. For this reason, it is necessary to inject and activate the dopant into regions having different depths.
The example of the manufacturing process of power device IGBT is shown. An n-type silicon wafer is prepared, a predetermined structure is formed on the surface, and grinding is performed from the wafer back side to a predetermined thickness.
A pn junction is formed on the back side of the wafer as described above. Typically, two methods are known.

Conventional Example 1 will be described with reference to FIG.
In this example, phosphorus (P) and boron (B) as dopants are successively implanted at different depths, and then an activation process is performed on both of them (see, for example, Patent Document 1).
Specifically, first, an n-type silicon wafer made by the FZ (floating zone) method is prepared, and the surface structure of the gate and emitter is built. Next, the back surface of the FZ wafer is ground to a predetermined thickness (100 to 150 μm). P is implanted into the wafer to a deep region by ion implantation of 400 keV and a dose of 1E + 13 / cm ² .
FIG. 14 (1) shows the concentration distribution in the depth direction after the implantation of P dopant. With a normal ion implanter, the dopant can be implanted to a depth (range: peak value) of about 1.0 μm. Subsequently, B is implanted into a shallow region by ion implantation of 10 to ²⁰ keV and a dose of 5E + 13 / cm ^{2 to} 1E + 14 / cm ² . FIG. 14 (2) shows the concentration distribution of P and B in the depth direction after implantation of the B dopant. Thereafter, both P and B implantation regions are simultaneously activated by laser light irradiation with a delay time in a double pulse. This laser light irradiation artificially increases the pulse width. The activation process is performed while the surface layer is controlled to the melting point or higher and the sublimation temperature or lower, and as shown in FIG. 14 (3), activation is performed in the process of cooling from the liquid phase because it melts in the shallow B injection region. In the P implantation region deeper than that, activation is performed in the process of cooling from the solid phase (solid phase activation). The collector electrode is formed to complete the IGBT.
The thickness of the n-type semiconductor layer of implanted phosphorus is about 1.0 μm (peak value: 0.5 μm), and the thickness of the p-type semiconductor layer of implanted boron is about 0.2 μm.

In the other conventional example 2, as shown in FIG. 15, after deeply implanting P, the injection region is solid-phase activated, and after injecting B shallowly, the injection region is solid-phase activated (for example, (See Patent Documents 2 and 3).
Specifically, first, an n-type silicon wafer made by the FZ (floating zone) method is prepared, and the surface structure of the gate and emitter is built. Next, the back surface of the FZ wafer is ground to a predetermined thickness (70 to 120 μm). P is implanted into the wafer to a deep region by ion implantation at a dose of 1E + 12 / cm ^{2 to} 1E + 13 / cm ^{2 at} 300 to 600 keV. Next, the surface layer is melted by double-pulse laser annealing, and a deeper region is solid-phase activated. The thickness of the n-type semiconductor layer of the implanted phosphorus is 1 to 1.5 μm. FIG. 15 (1) shows the concentration distribution in the depth direction after the implantation of P dopant. With a normal ion implanter, the dopant can be implanted to a depth (range: peak value) of about 1.0 μm. As shown in FIG. 15B, the P implantation region is solid-phase activated by the laser annealing. Next, B is implanted into a shallow region by ion implantation of 30 to 60 keV and a dose of 1E + 14 / cm ^{2 to} 1E + 16 / cm ² . FIG. 15 (3) shows the concentration distribution of P and B in the depth direction after the implantation of the B dopant. Thereafter, in the same manner as described above, the surface layer is melted by double pulse laser annealing, and only the B injection layer deeper than that is solid-phase activated. The thickness of the p-type semiconductor layer of the implanted boron is 0.1 to 0.5 μm. In this example, as described above, implantation and activation are performed independently for each impurity (see Patent Document 3).

JP 2007-123300 A JP 2007-59431 A JP 2008-85050 A

  By the way, in the power device IGBT, it is necessary to form an n buffer (field stop) layer having a depth of more than 2 μm on the collector side from the viewpoint of further high breakdown voltage characteristics and the like. Deep activation technology is required. Among them, it is no exaggeration to say that the activation technology with a low thermal load for a thin silicon wafer having a thickness of about 100 μm is the key to the bonding process.

  In the conventional manufacturing technique, ion implantation into a deep layer requires an ion implantation machine with MeV class implantation energy, and the occurrence of crystal defects and strain accompanying ion implantation at high energy is more remarkable. Become. However, in the conventional method, since the activation of the deep layer is performed in the solid phase, it is difficult to sufficiently activate. If the activation is performed over a longer time, the heat load on the semiconductor increases. The metal (Al alloy) wiring on the other side is thermally damaged. Further, there is a problem in that the dopant is diluted due to the solid phase diffusion, and the thickness of each layer cannot be formed as designed. Furthermore, since deep activation by a laser with a short irradiation time is performed in a solid phase, the concentration distribution by dopant implantation is almost maintained after activation. When a dopant is implanted with one kind of energy, a concentration distribution having a peak (ideally a Gaussian distribution) is obtained. Therefore, in order to distribute the concentration uniformly in the deep direction as much as possible, the dopant must be injected in multiple stages by changing the energy, and there is a problem that the number of process steps increases.

  The present invention has been made against the background of the above circumstances, and provides a semiconductor device manufacturing method and manufacturing apparatus capable of implanting and activating dopants in deep regions without increasing the thermal load on the semiconductor. Objective.

That is, the first aspect of the method for manufacturing a semiconductor device according to the present invention includes a first step of implanting a dopant shallowly at an implantation depth of 1 μm or less into a semiconductor surface layer, and the surface layer from the dopant implantation region. by also to liquid-phase diffusion to a depth of the said dopant is melted melted above the melting depth deep 2 [mu] m, and a deep implantation and activation of the liquid phase of said dopant to the melt depth A deep process by the liquid phase diffusion in the second process and an activation process thereof, and a continuous laser beam having a wavelength of 650 to 1100 nm is relatively applied to the semiconductor surface. with irradiated while scanning assisted heating, this performed by a composite radiation half-value width is repeatedly duplicated irradiated while relatively scanning the pulsed laser beam having the following pulse waveform 600ns Characterized by the impurity concentration and the box-type distribution profile in.

  In addition, the implantation depth of the dopant in the first step can be set to a depth that exceeds the depth of the maximum concentration of the dopant and becomes 1% of the maximum concentration.

Second method for fabricating a semiconductor device of the present invention, in the first aspect of the present invention, the continuous laser beam is a near infrared laser light generated by a continuous wave laser light source, the pulsed laser beam has a pulse It is a green laser beam generated by an oscillation laser light source.
According to a third aspect of the present invention, there is provided the method for manufacturing a semiconductor device according to the first or second aspect of the invention , wherein the continuous laser beam is a periodic or non-continuous portion in which the power density applied to the semiconductor is minimized. It is characterized by periodicity.
According to a fourth aspect of the present invention, there is provided a method for manufacturing a semiconductor device according to any one of the first to third aspects of the present invention, wherein the irradiation area of the continuous laser beam is set to be larger than the irradiation area of the pulse laser beam on the semiconductor surface. It is characterized by being enlarged.
According to a fifth aspect of the semiconductor device manufacturing method of the present invention, in any one of the first to fourth aspects of the present invention, at least a part or all of the continuous laser light irradiation area on the semiconductor surface is scanned. It is characterized by being positioned beyond the irradiation area of the pulse laser beam on the direction side.
According to a sixth aspect of the present invention, there is provided the method for manufacturing a semiconductor device according to any one of the first to fifth aspects of the present invention, wherein the first step assumes the liquid phase diffusion of the dopant in the second step. The dopant is injected in an amount.
For example, the dopant implantation amount in the first step is adjusted so that a carrier concentration of a certain threshold or more (for example, 1E + 16 / cm 3 or more) can be secured up to a predetermined depth.
According to a seventh aspect of the present invention, there is provided a method for manufacturing a semiconductor device according to any one of the first to sixth aspects of the present invention, wherein the semiconductor is a semiconductor for a power device IGBT, and the surface layer on the back surface side of the semiconductor. A pn junction having a collector structure is formed.
According to an eighth aspect of the present invention, there is provided a semiconductor device manufacturing method according to the seventh aspect , wherein the pn junction is formed in the order of a deep n-type semiconductor layer and a shallow p-type semiconductor layer. The layer is formed by the first step and the second step.
According to a ninth aspect of the present invention, in the eighth aspect of the present invention, the dopant of the n-type semiconductor layer is phosphorus (P), and the second dopant of the p-type semiconductor layer is boron. It is (B).

A semiconductor device manufacturing apparatus of the present invention includes a pulsed laser light source that generates a pulsed laser beam having a pulse waveform with a half-width of 600 ns or less, a continuous wave laser light source that generates a continuous laser beam having a wavelength of 650 to 1100 nm , The pulsed laser light generated by the pulsed laser light source and the continuous laser light generated by the continuous wave laser light source are led to a semiconductor surface in which a dopant is implanted in a shallow region at an implantation depth of 1 μm or less and combined irradiation is performed. An optical system, a scanning device that scans the pulsed laser light and continuous laser light relative to the surface of the crystal material, a power density of continuous laser light applied to the semiconductor surface, and the semiconductor surface irradiated The pulse energy density of the pulse laser beam and the scanning speed of the scanning device are set, and the composite irradiation performs the doppling. A control unit that melts the semiconductor surface layer to a melt depth of 2 μm or more deeper than the injection region of the punt and causes the semiconductor surface layer to liquid-phase diffuse to activate the dopant in the liquid phase to the depth.

  According to the present invention, the dopant is shallowly implanted into the deep region without deeply increasing the thermal load on the semiconductor by melting only the surface layer deeper and the layer into which the dopant has been implanted shallowly. Infusion and activation are performed at the same time. By performing this melting in a short time, the thermal load on the semiconductor can be reduced, and the implantation depth of the dopant can be controlled by the melting depth, so that the dopant can be effectively implanted at a desired depth. Note that the melting / diffusion can be performed in an inert gas (Ar, He, or a mixed gas containing hydrogen (Ar + H, He + H, etc.)) atmosphere or a vacuum atmosphere.

Laser light irradiation is effective in the deep melt diffusion process described above, and combined irradiation with continuous laser light and pulsed laser light is particularly desirable.
In order to realize deep complete melting, irradiation with a pulsed laser beam is important. However, since laser power is limited, laser light with a short pulse width that can take power is desirable to ensure the maximum temperature, and a pulse waveform with a half-value width of 600 ns or less (for example, a green wavelength) is desirable. More desirably, the full width at half maximum is 300 ns or less.
For pulsed laser light irradiation, pulsed laser light output from a pulsed laser light source is used. Examples of the pulsed laser source include those using the second harmonic of an LD-pumped Yb: YAG laser, but the present invention is not limited to a specific pulsed laser source.

However, the light penetration length is limited only by pulsed laser light irradiation, and rapid cooling occurs when the pulse width is shortened, so that crystal defects are likely to occur during cooling after heating. Therefore, from the viewpoint of melting / diffusion to a deep region and control of the cooling rate, assist heating for heating the semiconductor surface layer is indispensable. Assist heating can heat the surface layer of the crystal material for a longer time than pulsed laser light, and can melt to a deep region (eg, 2 μm or more) of the crystal material in combination with pulsed laser light irradiation. The cooling can be moderated.
The assist heating can be performed by laser irradiation with a short wavelength such as ultraviolet light, but is preferably performed using continuous laser light having a longer wavelength. It is desirable that the continuous laser light is irradiated at the same time in accordance with the scanning of the pulse laser light.

The continuous wave laser light source that outputs continuous laser light preferably outputs near infrared laser light such as an LD laser (wavelength 808 nm) having a wavelength longer than the green wavelength.
The near-infrared laser beam can be exemplified as having a wavelength of 650 to 1100 nm. Suitably, the wavelength of 680-825 nm can be shown. In the above wavelength region, light absorption with respect to silicon, which is a general semiconductor material, is good, and a light penetration depth deeper than that of the pulse laser beam can be obtained. As a result, the semiconductor surface layer portion is heated to a deep region, and an assist action is effectively obtained. The assist temperature can be adjusted by controlling the power density of the near infrared laser and the scanning speed.

Note that the present invention only needs to be obtained by combining the action of pulsed laser light irradiation and the assisting action by continuous laser light. The irradiation position in pulsed laser light and continuous laser such as near infrared laser light can be used. The relationship of the light irradiation position is not limited to a specific one. Accordingly, the irradiation areas of the continuous laser beam and the pulse laser beam are irradiated so that they are partially or entirely overlapped on the semiconductor surface, or the respective laser beams are shifted from each other without overlapping each other. Also good. However, if these irradiations are performed completely individually, the effects of the combined irradiation cannot be obtained, so that the effects of the respective irradiations need to influence each other.
The irradiation area can be shown as an area where the energy density of the pulsed laser beam or the power density of the continuous laser beam is 50% (FWHM) with respect to the peak value on the semiconductor surface.

However, in order to effectively obtain the assist action, it is desirable that the irradiation area of continuous laser light such as near infrared laser light is larger than the irradiation area of pulse laser light, and further, the irradiation area of continuous laser light is It is more desirable to cover the irradiation area of the pulse laser beam. In order to obtain an effect as preheating, it is desirable that a part or all of the irradiation area of the continuous laser beam is located at least beyond the irradiation area of the pulse laser beam on the scanning direction side. In order to obtain an effect as heating, it is even more desirable that the irradiation area of the continuous laser beam exceeds the irradiation area of the pulse laser beam on the opposite side in the scanning direction. In addition, it is desirable that the positional relationship between the irradiation areas of both laser beams be symmetric with respect to the scanning direction and the direction orthogonal to the scanning direction. Thus, the same positional relationship can be obtained when the scanning direction is reversed.
That is, it is desirable that the irradiation area of the continuous laser beam has a size exceeding the entire irradiation area of the pulse laser beam. By making the irradiation area of the continuous laser light wider than the irradiation area of the pulse laser light, it is possible to alleviate the heat escape in the lateral direction in the semiconductor layer portion and to heat the semiconductor to a deep region. Further, since the light penetration lengths of the continuous laser light and the pulse laser light are different, the heat escape in the depth direction can be mitigated, and heating to a deep region of the semiconductor is promoted.

  When the irradiation area of the continuous laser light is larger than the irradiation area of the pulse laser light, the beam size (cross section size) of the continuous laser light must be larger than the beam size (cross section size) of the pulse laser light. In this case, the beam size of the continuous laser beam is preferably the beam size of the pulse laser beam + the thermal diffusion length. The maximum beam size of continuous laser light is determined by whether or not the assist temperature determined by the power density and the scanning speed of the substrate is sufficient for deep melting. However, it is desirable that the assist temperature be lower than the melting point of the semiconductor surface material (for example, crystalline silicon) as described above.

Further, it is desirable that the irradiation with the pulse laser beam and the irradiation with the continuous laser beam be performed on the semiconductor surface at the same time. Therefore, both laser beams may be irradiated simultaneously to a predetermined position on the semiconductor surface, or both laser beams may be irradiated to a predetermined position on the semiconductor surface with a time difference. When there is a time difference, the time difference is set so that the assist action of the continuous laser light can be effectively obtained in the irradiation of the pulsed laser light. That is, if the time difference is too large, the continuous laser beam assist function cannot be sufficiently obtained in the irradiation with the pulsed laser beam. Irradiation having the time difference while maintaining the assist action is also included in the irradiation at the same time.
The relationship between the beam size and the irradiation position described above can be adjusted by an optical system. The optical system includes optical materials such as a homogenizer, a lens, and a mirror, and performs shaping and deflection of laser light.

By the combined irradiation of the continuous laser light and the pulse laser light, the semiconductor surface layer can be melted deeply (for example, 2.0 μm or more), and deep implantation and activation of the dopant can be performed collectively by liquid phase diffusion. . This melting depth is determined by setting the power density of the continuous laser beam irradiated on the semiconductor surface, the pulse energy density of the pulse laser beam irradiated on the semiconductor surface, and the scanning speed of the scanning device by the control unit. Can be determined. The power density of the continuous laser beam and the pulse energy density of the pulse laser beam can be adjusted by adjusting the output of the laser light source and / or adjusting the attenuation factor of the attenuator.
The relationship between the above item and the melting depth of the semiconductor is grasped in advance, and the above setting is performed based on this relationship.

  The present invention can be suitably applied to the manufacture of a power device IGBT, and the implantation and activation of a dopant into the surface layer can be used for forming a pn junction having a collector structure on the back surface. However, the application of the present invention is not limited to this, and it can be suitably used for manufacturing a semiconductor device in which a plurality of types of dopants are implanted and activated at different depths. In this case, a third step of implanting the second dopant into the surface layer shallower than the implantation depth in the second step, and melting the surface layer shallowly below the implantation depth of the second dopant to melt the second dopant. And a fourth step of activating the dopant. Even when multiple types of dopants are not used, it can be applied to the manufacture of various semiconductor devices in which dopants are implanted into shallow regions and then liquid phase diffusion is performed in the deep regions to deeply process deep implantation and activation. can do.

As described above, according to the semiconductor device manufacturing method of the present invention, the first step of injecting the dopant shallowly into the surface layer of the semiconductor, and melting the surface layer deeper than the dopant injection region, A second step of performing batch processing of deep implantation of the dopant into the melt depth and activation in the liquid phase by liquid phase diffusion to the melt depth; In the process, the deep implantation by melt diffusion and the activation process are performed by irradiating the semiconductor surface with continuous laser light while relatively scanning and assist heating, and repeatedly repeating while scanning with pulsed laser light. Since it is performed by the compound irradiation to irradiate, the dopant is injected into a shallow region and melted deeper than the implantation depth, so that the dopant is diffused to a deep region with a low heat load, It can be at activation.
Also, without using a MeV class high energy ion implanter, an ion implantation technique with a depth of about 1 μm or less (can be supported by a conventional machine having an implantation energy of 1 MeV or less) and a laser annealing technique that can melt deeper than the implantation depth. In combination, the shallow implanted dopant is deeply diffused by melting and simultaneously activated, whereby a deep n buffer layer of the high voltage IGBT can be formed. As described above, in the present invention, the use of a high energy ion implanter is not necessary, but the use of a high energy ion implanter is not excluded.

Furthermore, as a specific effect when applied to IGBT,
1) It is not necessary to introduce a MeV class high energy ion implanter, and a deep n buffer layer of a high breakdown voltage IGBT can be formed by combining a conventional implanter and a laser annealing machine capable of deep melting. There is no need to invest in an ion implanter, which can reduce manufacturing costs.
2) Near the melting point, a large difference of 1 to 3 digits occurs in the diffusion coefficient between the liquid phase and the solid phase. By utilizing this large difference in diffusion coefficient, the diffusion depth of impurities can be controlled. Therefore, the dopant activation depth can be controlled by optimizing the melting depth.
3) Since the diffusion coefficient in the melt (liquid phase) is larger than that in the solid phase, the shape of the dopant distribution profile in the melt is easily Box-type (the concentration of activated impurities is constant along the depth direction). It can be shaped and has a high activation rate.
4) By increasing or decreasing the initial ion implantation amount, the amount of electrically activated carriers required by the device IGBT (the activation amount of phosphorus in the n buffer layer) can be controlled.
5) By setting the initial impurity implantation to a depth of 1 μm or less and the melting depth to 2 μm or more, Box-type deep activation exceeding a depth of 2 μm can be realized.
6) Conventionally, in order to make the carrier concentration distribution of the n buffer layer uniform deeply (Box type), it has been necessary to implant dopants by changing energy in multiple stages. In the present invention, since a box-type carrier distribution can be obtained by one dopant implantation, the implantation processing time can be shortened.

It is a figure which shows the structure of the laser annealing apparatus used for one Embodiment of this invention. Similarly, it is a figure which shows the time chart of a pulse laser beam and a continuous laser beam. Similarly, it is a figure which shows the time chart of the pulse laser beam in another example, and a continuous laser beam. It is a figure which shows the outline of the IGBT structure to which this invention is applied. Similarly, it is a figure which shows a part of manufacturing process of IGBT to which the method of this invention is applied. Similarly, it is a figure which shows the density | concentration distribution of the dopant in the manufacturing process of IGBT to which the method of this invention is applied. Similarly, it is a figure which shows the irradiation area of the pulse laser beam and continuous laser beam on a semiconductor. Similarly, it is a figure which shows the heating area | region by laser beam irradiation in the cross section of a semiconductor. Similarly, it is a figure which shows the example of a change of the irradiation area of the pulse laser beam and continuous laser beam on a semiconductor. Similarly, in the example, when the P is shallowly implanted (as-implanted), when the hybrid type laser annealing is performed (as-annealed), the concentration distribution in the depth direction of the dopant and the electrically activated state are shown. It is a figure which shows the concentration distribution of the depth direction of the carrier. Similarly, in the examples, it is a diagram showing the concentration distribution of the dopant in the depth direction when B is implanted shallowly and when hybrid laser annealing is performed. Similarly, it is a figure which shows oxygen concentration distribution of the depth direction at the time of performing hybrid type laser annealing using the same sample as FIG. Similarly, it is a figure which shows the density | concentration distribution of the depth direction of a dopant at the time of performing a hybrid type laser annealing by inject | pouring P deeply as a comparative example. It is a figure which shows the density | concentration distribution of the dopant in the process of the conventional dopant injection | pouring and activation process. It is a figure which shows the density | concentration distribution of the dopant in the process of the dopant injection | pouring of other conventional examples, and the activation process.

Hereinafter, an embodiment of the present invention will be described. FIG. 4 shows an example of a cross-sectional structure of an FS type IGBT to be manufactured in the present invention.
A p-type base region 33 in which boron is implanted is formed on the surface side of the semiconductor 30 having a thickness of about 100 μm, and an n + -type emitter region in which phosphorus is implanted into a part of the surface side of the p-type base region 33. 34 is formed. A p + -type collector layer 32 in which boron is implanted is formed in the surface layer on the back side of the semiconductor 30. In a region deeper than the collector layer 32, an n + type buffer layer 31 in which phosphorus is implanted so as to be in contact with the collector layer 32 is formed, and an n− type substrate 35 is located inside thereof. In the figure, 36 is a collector electrode, 37 is an emitter electrode, 38 is a gate oxide film, and 39 is a gate electrode.

FIG. 1 shows an outline of a laser annealing apparatus configuration, which will be described below.
As shown in FIG. 1, the laser neil device includes a processing chamber 2, a scanning device 3 that can move in the XY direction, and a base 4 on the upper portion. . On the base 4, a target object placement base 5 is provided. During the annealing process, the semiconductor 30 to be processed is placed on the target object placement table 5. The scanning device 3 is driven by a motor (not shown) or the like.
A pulse laser light source 10 on which the second harmonic of an LD-pumped Yb: YAG laser is mounted is installed outside the processing chamber 2. The pulsed laser beam 15 that is output by being pulsated by the pulsed laser light source 10 is adjusted in energy density by an attenuator 11 as necessary, and is subjected to beam shaping or optical processing by an optical system 12 including a lens, a reflecting mirror, a homogenizer, and the like. The deflection is made and the semiconductor 30 in the processing chamber 2 is irradiated.
The pulsed laser light 15 output from the pulsed laser light source 10 has a pulse waveform with a half width of 600 ns or less. When the semiconductor laser 30 is irradiated with the pulsed laser light 15, the energy density is adjusted so that only the surface layer of the semiconductor 30 is melted. As described above, the pulse laser beam 15 is shaped into, for example, a spot shape, a circular shape, a square shape, or a long shape by the optical system 12.

  In addition, a continuous wave laser light source 20 including an LD laser light source that generates near-infrared laser light is installed outside the processing chamber 2. The near-infrared laser light 25 output from the continuous wave laser light source 20 is adjusted in power density by an attenuator 21 as necessary, and is shaped and deflected by an optical system 22 including a lens, a reflection mirror, a homogenizer, and the like. And the semiconductor 30 in the processing chamber 2 is irradiated. The near-infrared laser beam 25 is adjusted to a power density that does not reach the melting point of the semiconductor 30 when the semiconductor 30 is irradiated and scanned. As described above, the near-infrared laser beam 25 is shaped into, for example, a spot shape by the optical system 22, and the size thereof is adjusted to be larger than the size of the pulse laser beam 15.

  The laser annealing apparatus has a control unit 6 that controls the entire laser annealing apparatus. The control unit 6 can mainly configure a CPU and a program for operating the CPU. The controller 6 is controllably connected to a power source 7a for driving the pulse laser light source 10 and a power source 7b for driving the continuous wave laser light source 20, and can set the output of each laser light source. Moreover, the control part 6 is connected to the attenuator 11 and the attenuator 21 in a controllable manner, and can set the respective attenuation rates. The control unit 6 is controllably connected to an XY stage drive circuit 8 that outputs a drive signal to the scanning device 3, and can set a moving speed in the scanning device 3. Note that a PC terminal 9 is connected to the control unit 6, and a setting operation, an execution command operation, and the like can be performed on the control unit 6. It is also possible to receive the control information from the control unit 6 and monitor the operating state of the apparatus.

In the laser annealing apparatus, as shown in FIG. 2, a pulse laser beam 15 and a near infrared laser beam 25 are output and are combined and irradiated on the semiconductor 30. In this case, the energy density on the irradiation surface of the pulse laser light, the power density on the irradiation surface of the near-infrared laser light 25, and the scanning speed of the processing object arranging table 5 are set to predetermined values, so that the semiconductor 30 surface layers can be melted to a predetermined depth.
The semiconductor 30 is repeatedly and repeatedly irradiated with the pulsed laser light 15 and combined with the continuous laser light 25, whereby the surface layer portion of the semiconductor 30 is melted over a thickness of 2.0 μm or more. The overlapping ratio (overlap ratio) of the pulse laser beam 15 in the short axis direction and the long axis direction is appropriately selected as necessary (for example, 50 to 90% in the short axis direction, for example, 10% to 50% in the long axis direction). The present invention is not particularly limited. At this time, the semiconductor laser 30 can be scanned with the pulsed laser light 15 and the continuous laser light 25 at a predetermined speed by controlling the moving speed of the object placement table 5 by the scanning device 3. The moving speed of the laser beam in the minor axis direction can be, for example, 30 to 150 mm / second, but is not particularly limited as the present invention.
Note that the near-infrared laser beam, which is a continuous laser beam, always has a constant power density, and may have a discontinuous part as shown in FIG. Thereby, the assist heating in the near-infrared laser beam 25 can be adjusted to adjust the melting depth and the heat load in the surface layer of the semiconductor 30.

In the IGBT, an n-type silicon wafer is prepared as the semiconductor 30 and the surface structure of the gate and the emitter is created, as in the prior art. Next, the back surface of the semiconductor 30 is ground to a predetermined thickness (for example, 100 to 150 μm).
As shown in FIG. 5A, before the collector electrode 36 (see FIG. 4) is formed, P is made shallow in the back surface side of the semiconductor 30 by ion implantation of several tens keV and a dose amount E + 13 / cm ^2. The n + type semiconductor layer 31a is formed by implantation. The depth of the semiconductor layer of implanted phosphorus is zero comma number μm at the 1% position of the peak concentration. The dose amount of phosphorus to be injected is adjusted so as to obtain an electrically activated carrier concentration required after melt diffusion described later.

The semiconductor 30 when the above P is implanted shallowly is shown in FIG. Further, FIG. 6A shows the concentration distribution in the depth direction of the implanted phosphorus at this time. P is distributed at a high concentration in a shallow region.
The semiconductor 30 is subjected to composite irradiation while scanning the pulse laser beam 15 and the near-infrared laser beam 25 at a predetermined speed using the laser annealing apparatus. The surface layer of the semiconductor 30 is melted to a depth of 2.0 μm or more by the composite irradiation, and the previously injected P is liquid phase diffused over the melting depth and injected into a deep region, and with melting and solidification. Activation to deep areas. The semiconductor 30 in this state is shown in FIG. The density distribution at this time is shown in FIG. P is distributed at a stable concentration (Box distribution) over a deep region (depth of 2.0 μm or more). Thereby, for example, an n + type semiconductor layer having a depth of 3.0 μm can be formed.

Thereafter, B is implanted into the shallow region by ion implantation of several tens of keV, 1E + 14 / cm ^{2 to} 1E + 16 / cm ² into the semiconductor 30.
The semiconductor 30 in this state is shown in FIG.
Further, the concentration distribution in the depth direction of boron at this time is shown in FIG. Along with the concentration distribution of P after annealing, B is distributed at a high concentration in a shallow region. The rear surface side of the semiconductor 30 is irradiated with, for example, pulsed laser light to melt the very surface layer. As a result, the region into which B is implanted is activated well and the designed junction depth is maintained (see FIG. 6 (4)). In the B injection region, the surface layer melts very much, but it is substantially solid phase activation.
An IGBT can be obtained by forming the collector electrode 36 on the semiconductor 30.

In the present embodiment, it is possible to melt the semiconductor surface layer deeper by the combined irradiation of the pulsed laser beam and the continuous laser beam.
In this embodiment, as shown in FIG. 7, the irradiation area 25 a when the near-infrared laser light 25 is irradiated onto the semiconductor 30 covers the irradiation area 15 a when the pulse laser light 15 is irradiated onto the semiconductor 30. In addition, the optical systems 12 and 22 are adjusted so as to have a size exceeding the entire irradiation area 15a. In the figure, the arrow in the X-axis direction indicates the scanning direction of the laser beam. In this example, the pulsed laser beam 15 and the near-infrared laser beam 25 are irradiated in the same area on the semiconductor 30 without shifting the irradiation timing.
However, in the present invention, the position of the irradiation area of each laser beam is not limited to the above.

FIG. 8 shows a schematic diagram of thermal diffusion in the depth direction when the semiconductor surface is irradiated with the pulse laser beam and the continuous laser beam.
A temperature assist region is formed in the semiconductor 30 at a deep position of the semiconductor 30 by irradiation with continuous laser light having a light penetration length larger than that of the pulse laser light. For example, with a near-infrared laser beam having a wavelength of 808 nm, a light penetration length of about 10 μm can be obtained in the depth direction. When pulse laser light is irradiated in this state, heat is diffused mainly in the depth direction (Z-axis direction). In this case, the deep temperature assist region reduces the temperature gradient of the shallow heating region of the pulse laser beam, and as a result, the heat escape is reduced and the semiconductor surface is effectively heated to a deep position. In this case, optimization is performed by adjusting the energy density of the pulsed laser beam, the power density of the continuous laser beam, and the scanning speed, thereby effectively heating the semiconductor surface layer.
In addition, when the semiconductor 30 is irradiated only with the pulse laser beam, the temperature gradient in the surface direction and the depth direction is large, and the heat escape is large. For this reason, the heating effect in the depth direction is suppressed, and it becomes difficult to selectively deeply melt the surface layer portion particularly for a thick semiconductor having a large heat capacity.

Next, FIGS. 9A to 9E show examples of changing the irradiation area position of the pulse laser beam and the continuous laser beam.
FIG. 9A shows a rectangular cross section of the pulse at the irradiation position in FIG. 7, and the irradiation area 25a completely covers the irradiation area 15a. In FIG. 9B, the continuous laser light irradiation area 25a has a size exceeding the pulse laser light irradiation area 15a in the long axis direction (Y-axis direction) and the scanning direction (X-axis direction). In the opposite direction, the irradiation areas 15a and 25a overlap and the area ends coincide. In FIG. 9C, the irradiation area 25a does not cover the irradiation area 15a and does not overlap with each other, and the irradiation area 25a is positioned on the scanning direction side of the irradiation area 15a, and adjacent irradiation areas. The edges are in contact with each other. FIG. 9D shows that the irradiation area 25a is separated without covering the irradiation area 15a and without overlapping. However, both are irradiated near each other on the substrate. FIG. 9E shows an irradiation state outside the present invention, and shows a state in which the semiconductor 30 is irradiated with only the pulse laser beam 15 and the semiconductor 30 is processed by the irradiation area 15a. .

  On the semiconductor surface irradiated with the continuous laser light, the temperature gradually increases immediately after irradiation and becomes a steady state. On the other hand, with a pulse laser beam, the temperature rises in a very short time according to the pulse, and the temperature drops in a very short time according to the pulse. When irradiating pulsed laser light, continuous laser light may be irradiated, and the semiconductor laser may be irradiated after the semiconductor surface temperature has reached a steady state. The irradiation timing on the semiconductor surface may be set, for example, by setting a delay time, and after the continuous laser beam irradiation, the pulse laser beam may be irradiated with a delay according to the delay time, or the position of the irradiation area may be set. It is also possible to change the irradiation timing by scanning the composite laser beam with shifting. In the examples of FIGS. 9B, 9C, and 9D, the semiconductor 30 is irradiated with the pulsed laser light 15 after the continuous laser light 25 is irradiated.

Next, examples using the laser annealing apparatus described in the above embodiment will be described.
In the semiconductor 30, phosphorus was ion-implanted into a region having a depth of 1 μm or less at an energy of 40 keV and a dose of 1E + 15 / cm ² . Deep implantation by melt diffusion and activation treatment were performed by a hybrid laser annealing method with a short pulse width.
That is, the energy density of the pulsed laser beam with a wavelength of 515 nm was 14 J / cm ² , the pulse width was 300 ns, the power of the continuous laser beam with a wavelength of 808 nm was 45 W, and the surface of the semiconductor 30 was combinedly irradiated.
After melt diffusion, the concentration distribution profile in the depth direction of phosphorus in the surface layer portion of the semiconductor 30 was measured using SIMS (secondary ion mass spectrometry). Further, the concentration profile in the depth direction of the electrically activated carrier after melt diffusion was measured using SRP (spreading resistance measurement method). A depth profile in the SIMS and SRP is shown in FIG. The graph in the figure shows the change in the concentration distribution of phosphorus in the depth direction and the electrically activated carrier concentration distribution during injection and after melt diffusion.
As is apparent from FIG. 10, the concentration distribution of the electrically activated carrier is a box type (flat concentration distribution) up to the melting depth. The carrier concentration in the flat part is reduced by about two digits compared to the peak concentration of phosphorus at the time of implantation.
Under the same irradiation conditions of a pulse width of 300 ns, a pulse energy density of 14 J / cm ² , and an LD laser power of 45 W, the activation depth reaches 2.75 μm. When the carrier amount / injection amount was expressed as a percentage, a high activation rate of 72% was obtained.
Since a nearly constant activation rate is obtained regardless of the melting depth, a field stop layer suitable for the design can be achieved by adjusting the dose of implanted phosphorus to ensure the necessary carrier concentration for the activation depth. Can be formed.

Also, boron is implanted into the same semiconductor under the same implantation conditions (40 keV, 1E + 15 / cm ² ) as phosphorus, and the same laser annealing method is applied under irradiation conditions of a pulse width of 300 ns, a pulse energy density of 8 J / cm ² , and an LD power of 45 W. Melt and diffuse.
Before and after the melt diffusion, the concentration distribution profile in the depth direction of boron in the surface layer portion of the semiconductor 30 was measured using SIMS (secondary ion mass spectrometry). A depth profile in the SIMS is shown in FIG. The graph in the figure shows the change in boron concentration distribution at the time of injection and after melt diffusion. As is apparent from FIG. 11, even when boron is implanted shallowly, the concentration distribution becomes a box type (flat concentration distribution) up to the melting depth. Therefore, in the present invention, the same effect can be obtained regardless of the type of dopant.
In the laser annealing method of the present invention in this example, it was confirmed that a melting depth of about 2.9 μm was obtained under irradiation conditions of a pulse width of 300 ns, a pulse energy density of 14 J / cm ² , and an LD laser power of 45 W. (See FIG. 12). As shown in FIG. 12, the melting depth was obtained from the intersection position between the slope and the background where the oxygen concentration depth profile after annealing rapidly decreased from the flat portion. In other words, when the diffusion length is in the vicinity of the melting point, a large difference of 1 to 3 digits occurs between the liquid phase and the solid phase, so that the oxygen concentration is rapidly attenuated at the interface between the liquid phase and the solid phase.
A comparison between the activation depth from the SRP depth profile of FIG. 10 and the oxygen SIMS depth profile of FIG. 12 demonstrated that Box-type activation is possible up to the vicinity of the melting depth.

It should be noted that the reason why the deep impurity activation must start from shallow impurity implantation will be described with a specific example (see FIG. 13).
In the example of FIG. 13, phosphorus was implanted deep into the semiconductor surface layer with an energy of 750 keV and a dose of 1E + 15 / cm ² . In the semiconductor, the peak position (range) of the SIMS depth profile is at a depth exceeding 1 μm. This phosphorus-implanted semiconductor was heat-treated using the laser annealing method of the present invention described above under the irradiation conditions of a pulse laser beam having a pulse width of 300 ns, a pulse energy density of 8 J / cm ² , and a continuous laser beam power of 45 W.
As shown in FIG. 13, the SIMS depth profile after melt diffusion does not have a constant impurity concentration in the region shallower than the melt depth of about 1 μm, unlike the box type, regardless of the liquid phase. It is considered that the temperature gradient and impurity concentration gradient at the time of laser irradiation in the region where phosphorus is deeply implanted inhibits the diffusion of impurities in the depth direction. Therefore, in order to deeply diffuse impurities during melting, the process should be started from shallow impurity implantation of 1 μm or less so that the temperature gradient (heat flow) during laser irradiation and the impurity gradient concentration are aligned in the same direction. Is essential.

  As mentioned above, although this invention was demonstrated based on the said embodiment and Example, this invention is not limited to the content of the said description, As long as it does not deviate from this invention, an appropriate change is possible.

Reference Signs List 2 Processing chamber 3 Scanning device 4 Base 5 Processed object placement table 6 Control unit 7a Power supply 7b Power supply 8 XY stage drive circuit 9 PC terminal 10 Pulse laser light source 11 Attenuator 12 Optical system 15 Pulse laser light 15a Irradiation area 20 Continuous oscillation Laser light source 21 Attenuator 22 Optical system 25 Near infrared laser beam 25a Irradiation area 30 Semiconductor 31 n + type buffer layer 32 Collector layer 35 n− type substrate

Claims (10)

A first step of implanting shallow dopants into a surface layer of the semiconductor following implantation depth 1 [mu] m, the surface layer is melted above the melting depth deep 2μm than implanted region of said dopant of the melt the dopant A second step of performing batch treatment of deep implantation of the dopant into the melt depth and activation in the liquid phase by liquid phase diffusion to a depth, and the step in the second step Deep implantation by liquid phase diffusion and its activation treatment are performed by irradiating the semiconductor surface with continuous laser light having a wavelength of 650 to 1100 nm while relatively scanning and assist heating, and a pulse waveform having a half width of 600 ns or less. be characterized in that the impurity concentration of the box-type distribution profile by performing the composite radiation to repeatedly duplicate irradiated while relatively scanning the pulsed laser light having a A method of manufacturing a semiconductor device. The continuous laser beam is a near infrared laser light generated by a continuous wave laser light source, the pulsed laser beam, to claim 1, wherein the pulse is a green laser light generated by the oscillation laser light source The manufacturing method of the semiconductor device of description. 3. The method of manufacturing a semiconductor device according to claim 1, wherein the continuous laser beam makes a discontinuous portion in which a power density irradiated to the semiconductor becomes a minimum is periodic or aperiodic. 4. Wherein on the semiconductor surface, a method of manufacturing a semiconductor device according to any one of claims 1 to 3, characterized in that larger than the irradiation area of the irradiation area of the continuous laser light the pulse laser light. On the semiconductor surface, some or all of the irradiation area of the continuous laser light, more of claims 1-4, characterized in that to position beyond the irradiation area of the pulse laser beam at least the scanning direction A method for manufacturing a semiconductor device according to claim 1. In the said 1st process, the said dopant is inject | poured with the injection quantity which assumed the liquid phase diffusion of the said dopant in the said 2nd process, The manufacturing of the semiconductor device in any one of Claims 1-5 characterized by the above-mentioned. Method. Wherein the semiconductor is a semiconductor for power device IGBT, according to any one of claims 1 to 6, characterized in that the surface layer on the back side of the semiconductor and forms a pn junction of the collector structure A method for manufacturing a semiconductor device. The pn junction is formed in the order of a deep n-type semiconductor layer and a shallow p-type semiconductor layer, and the deep n-type semiconductor layer is formed by the first step and the second step. A method for manufacturing a semiconductor device according to claim 7 . 9. The method of manufacturing a semiconductor device according to claim 8 , wherein the dopant of the n-type semiconductor layer is phosphorus (P), and the second dopant of the p-type semiconductor layer is boron (B). A pulsed laser light source that generates a pulsed laser beam having a pulse waveform with a half-width of 600 ns or less, a continuous wave laser light source that generates a continuous laser beam having a wavelength of 650 to 1100 nm, and the pulsed laser light source generated by the pulsed laser light source An optical system for irradiating a pulsed laser beam and the continuous laser beam generated by a continuous wave laser light source to a semiconductor surface in which a dopant is implanted in a shallow region at an implantation depth of 1 μm or less and irradiating the laser beam, and the pulsed laser beam and A scanning device that scans a continuous laser beam relative to the surface of the crystal material, a power density of the continuous laser beam irradiated on the semiconductor surface, a pulse energy density of the pulsed laser beam irradiated on the semiconductor surface, and the scanning set the scanning speed of the device, the deeper than implanted region of the dopant by the combined radiation 2μm or less Apparatus for manufacturing a semiconductor device is melted to the melting depth, characterized in that it comprises a control unit for activating the dopant in the liquid phase in the deep customers in liquid phase by diffusing the semiconductor surface layer.

Images