JP2010272587A - Method for activating semiconductor impurities - Google Patents

Abstract

Provided is a method for effectively and satisfactorily activating a semiconductor into which impurities are introduced.
A series of continuous pulses with a time difference is applied to a continuous impurity layer in which a first impurity exists at a certain depth from a surface layer of a semiconductor substrate and a second impurity exists in a deeper region. And irradiating the same position with each other in a series of continuous pulses while relatively scanning a predetermined range of the semiconductor substrate, and activating the impurity layer in a non-molten state during the series of continuous pulses. Turn into. Since the heating time is long, heat can be sufficiently diffused to a deep region, and activation of the deep region is facilitated. In addition, since the impurity layer to be activated maintains a solid state, diffusion hardly occurs, and there are semiconductor substrates in which different impurity layers are continuously present, or layers having different implantation conditions even if the same impurity is present. It is effective for activation of continuous semiconductor substrates.
[Selection] Figure 4

Description

  The present invention relates to a method for activating semiconductor impurities by irradiating a pulse laser.

Activation of semiconductor substrates such as insulated gate bipolar transistors (hereinafter referred to as IGBTs) and image sensors using lasers has been attempted for some time. For example, a field stop (hereinafter referred to as FS) type IGBT as shown in FIG. 2 has a gate oxide film 24, a gate electrode 23, and an emitter electrode 22 formed on the surface side of an n − type substrate 15. An n-type buffer layer 11 doped with phosphorus is formed on the back side of the substrate 15, and a p-type collector layer 12 doped with boron and a collector electrode are formed on the buffer layer 11. is there. In the prior art, after the doping of phosphorus and boron in the process of forming the back side of the IGBT, the characteristics of the metal emitter electrode formed on the surface are maintained, and activation at a temperature at which no deterioration occurs is electrically performed. When performed in a furnace, the back side cannot be set to the temperature necessary for activation, and sufficient activation is difficult. In order to make the side hot, activation by local heating using a laser has been attempted.
In particular, in the activation using a pulse laser, since the pulse width (half width of a pulse) of a conventionally used laser is narrow, the heating time is short and it is difficult to perform sufficient activation. For this reason, a method has been proposed in which a plurality of pulses are irradiated continuously to increase the apparent pulse width and activate (see Patent Documents 1 to 5).

  Patent Document 1 proposes a method of irradiating a semiconductor substrate with a laser and performing activation by a double pulse under a condition that a deep region to which an impurity is added is 950 ° C. or higher. At this time, irradiation is performed under the condition that the surface temperature is lower than the sublimation temperature. Continuous pulses are irradiated with a delay time such that a part of the waveforms overlap.

  Patent Document 2 proposes a method of activation by solid phase diffusion. By continuously irradiating pulses, the cooling rate is slowed down, and the generation of defects due to rapid cooling is suppressed.

  Patent Document 3 proposes a method of continuously irradiating a laser with a delay time in order to activate impurity layers having different types or different implantation conditions in the depth direction. Activation is performed without melting the surface layer of the semiconductor substrate. In the example of this document, it is activated to a depth of about 1.5 μm. The upper limit of the pulse width is 1000 ns depending on the number of lasers. Theoretically, 1000 ns or more is possible.

  Patent Document 4 proposes a method of activating a layer that exists continuously in the depth direction with a laser. This laser is called superimposed laser light. The pulse width is set to 100 ns or more and 5 μs or less (200 ns and 500 ns in the embodiment), and can be melted up to a depth of about 0.1 μm, and the work type is selected for activation of a deep region and activation of a shallow region. It is done separately.

  Patent Document 5 proposes a method of activating impurities present in a semiconductor substrate using two lasers. Irradiation is performed at an energy density that does not cause the semiconductor substrate to melt in the previous pulse, and the subsequent pulse is irradiated at an energy density that reaches the same temperature as the previous pulse at a predetermined delay time. The impurity layer is activated without melting the region.

JP 2007-123300 A JP 2006-344909 A JP 2005-223301 A JP 2007-59431 A JP 2008-270243 A

However, the above-described proposed technology still has the following problems.
In the technique proposed in Patent Document 1, since the laser is irradiated so as to be equal to or higher than the melting temperature, the surface layer of the semiconductor substrate is melted. When the irradiated surface is melted, the laser beam is absorbed by the melted layer, and the laser beam does not easily reach a deeper region. For this reason, activation of a region deeper than the melted layer depends on heat diffusion, but since each pulse width is narrow, the heating time is short and the depth that can be activated by heat diffusion is limited.

  In the technique proposed in Patent Document 2, activation is performed by solid phase diffusion, but the surface temperature becomes high because the pulse width is narrow, but in a deep region, heating is performed to a temperature necessary for activation, It is difficult to maintain the temperature. Further, since the surface layer is easily melted due to rapid heating and the energy density to be input cannot be increased, the depth that can be activated is limited.

  In the technique proposed in Patent Document 3, a number of lasers are connected to increase the apparent pulse width. However, the actually widened pulse width is determined by the number of lasers, and the use of many lasers is not a practical method. Therefore, although it can be activated up to about 1.5 μm in the embodiment, it is practically difficult to activate a deeper region. Further, although a substantially rectangular beam is used, in order to shape the oscillation waveform of the solid-state laser, an additional optical system is required in addition to the condensing optical system. Furthermore, by using a substantially rectangular beam, the maximum energy density is reduced, and it may be impossible to obtain the energy density necessary for activation of a deep region. In order to secure a sufficient energy density with a substantially rectangular beam, it is necessary to reduce the beam size, and the productivity is lowered.

  In the technique proposed in Patent Document 4, when an impurity is generally introduced, the laser is easily absorbed in a region where the impurity is introduced, and therefore, it is difficult to penetrate into a region deeper than the region. The activation process by laser irradiation is divided, the previously introduced impurity in the deep region is activated, the impurity is then implanted into the shallow region, and the shallow region is activated again by laser irradiation. However, the manufacturing cost and time for that amount are also required.

  In the technique proposed in Patent Document 5, specifically, since the pulse width is short, rapid heating and rapid cooling occur, and the temperature necessary for activating a deep region exceeding 2 μm without melting is reached, and It is difficult to maintain enough time. In addition, since the pulse width is short, the activation result is easily affected by fluctuations in each pulse and a delay time, and it is difficult to activate a stable deep region.

  In other words, the depth that can be activated by the conventional method is limited, and the pulse width is not sufficiently wide to activate a deep region. It is not a realistic method. In view of the above, an object of the present invention is to provide a practical method in which the impurity layer to be activated maintains a solid state, the impurity hardly diffuses, and the activation can be performed up to a deep region. And

  That is, of the semiconductor impurity activation methods of the present invention, the first invention is a continuous method in which the first impurity is present at a certain depth from the surface layer of the semiconductor substrate, and the second impurity is present in a deeper region. The impurity layer is irradiated with a plurality of pulse lasers at the same position as a series of continuous pulses with a time difference, and the irradiation is performed while relatively scanning a predetermined range of the semiconductor substrate for each series of continuous pulses, The impurity layer is activated in a non-molten state during the series of continuous pulse irradiations.

  In the semiconductor impurity activation method of the second aspect of the present invention, in the first aspect of the present invention, the time width per pulse of the pulse laser (hereinafter referred to as pulse width) is longer than 500 ns and not longer than 2000 ns. Of the series of continuous pulses, the pulse width of the second and subsequent pulse lasers is equal to or longer than the pulse width of the first pulse laser.

  The semiconductor impurity activation method according to the third aspect of the present invention is characterized in that, in the first or second aspect of the present invention, the width of the short axis of the pulse laser is larger than 3 μm and smaller than 50 μm.

The semiconductor impurity activation method of the fourth aspect of the present invention is the semiconductor impurity activation method according to any one of the first to third aspects of the present invention, wherein the energy density of a single pulse on the semiconductor irradiation surface of the pulse laser is 2.0 J / cm ^2. As described above, it is 5.0 J / cm ² or less, and the impurity layer has an energy density that maintains a non-molten state during laser irradiation.

  The semiconductor impurity activation method according to the fifth aspect of the present invention is the semiconductor impurity activation method according to any one of the first to fourth aspects of the present invention, wherein the time difference between the pulses is the same as that of the previous pulse irradiation as the series of continuous pulses. The irradiation timing is set to be longer than the pulse width of the irradiation laser pulse and less than three times.

  According to a sixth aspect of the present invention, there is provided the semiconductor impurity activation method according to any one of the first to fifth aspects of the present invention, wherein the pulse widths of the plurality of pulse lasers irradiated to the same position in the semiconductor substrate having the impurity layer are provided. The total is longer than 1000 ns and shorter than 10 μs.

  In the semiconductor impurity activation method of the seventh aspect of the present invention, in any of the first to sixth aspects of the present invention, the repetition frequency of the pulse laser is irradiated with a plurality of pulse lasers as the series of continuous pulses, Until the next series of continuous pulses are irradiated with a plurality of pulse lasers, the semiconductor is at an interval at a temperature similar to that before irradiation.

  The semiconductor substrate activation method according to an eighth aspect of the present invention is characterized in that in any one of the first to seventh aspects of the present invention, the pulse laser is a Yb: YAG second harmonic.

  That is, according to the present invention, a plurality of pulse lasers that form a series of continuous pulses can be heated to a deep region while maintaining the heating temperature for a long time without melting the impurity layer. It is possible to activate a deep impurity layer that exceeds the depth.

  Hereinafter, the reasons for various conditions defined in the present invention will be described.

The pulse width of the laser is preferably longer than 500 ns and not longer than 2000 ns. FIG. 1 shows changes in temperature when the delay times are 300 ns and 2000 ns with respect to pulse widths of 150 ns and 1000 ns, respectively.
When the pulse width is 150 ns, the peak temperature is reached immediately after the start of irradiation in a short time, so rapid heating is performed. When the pulse width is 1000 ns, the temperature gradient from the start of irradiation until reaching the peak temperature is gentle, and the rapid heating is relaxed. Is done. Furthermore, since heat is transferred to a deep region until the peak temperature is reached and the impurity is warmed, the impurity is more easily activated when the peak temperature is reached, and the region reaching the temperature necessary for activation becomes deeper.

In addition, since the temperature rise is gradual, even if irradiation is performed at an energy density such that the temperature distribution is small and the temperature rises to near the melting point with respect to the laser intensity fluctuation for each pulse, Easy to maintain state. Furthermore, since the cooling rate is also slow, the effect of suppressing the formation of defects that are likely to occur during rapid cooling can be expected. In addition, when the same energy density is irradiated, the peak value of the pulse decreases as the pulse width increases. Therefore, the wider the pulse width, the higher the energy density, and the more the semiconductor substrate. The heat is transmitted to the deeper area, and activation is promoted.
From the above viewpoint, by performing irradiation with an optimal pulse width, even if the impurity layer is not melted at the time of laser irradiation, a deep region can be activated. When the pulse width is 500 ns or less, rapid heating occurs and the heating effect in the depth direction is reduced. When the pulse width exceeds 2000 ns, heat conduction in the depth direction increases, and the semiconductor substrate is formed on the surface side of the semiconductor substrate. The metal wiring will be affected. For these reasons, the pulse width per pulse is preferably longer than 500 ns and not longer than 2000 ns.

  Furthermore, it is desirable that the pulse width of the second pulse is longer than the pulse width of the first pulse in a series of continuous pulses. By irradiating as a series of continuous pulses, the impurity can be raised to a high temperature in the first pulse and maintained at a high temperature in the second and subsequent pulses. At this time, the impurity layer is not melted at the time of laser irradiation. Therefore, it is possible to promote the diffusion of heat in the depth direction by continuously irradiating a pulse having a pulse width wider than that of the first pulse.

The beam size of the pulse laser affects the temperature distribution of the impurity layer during laser irradiation, and in particular the width of the short axis affects the diffusion of heat from the center of the beam in the lateral direction. When the minor axis width is different for the same energy density and the same major axis width, the shorter axis width, the greater the heat diffusion in the lateral direction compared to the heat diffusion in the depth direction. It is difficult to maintain a deep impurity layer at a high temperature because heat is not transmitted to the substrate. If the width of the short axis is such that the gradient of heat from the center of the beam in the lateral direction is gentle, heat diffusion at the center of the beam is likely to be performed in the depth direction. In addition, when the minor axis width of the beam is narrow, the temperature distribution during laser irradiation is easily affected by fluctuations in the beam shape and peak value for each pulse, leading to variations in activation depth and activation concentration, and beam width If the temperature is wider than necessary, irradiation with a high energy density that is high to the melting point or near the melting point easily transmits the temperature to the surface of the semiconductor substrate, warms the metal wiring, and causes deformation and deterioration of characteristics. there is a possibility. Moreover, it becomes difficult to ensure the energy density required for activation by extending the minor axis width too much.
From the above points, by irradiating the width of the short axis of the beam within the optimum range, the influence on the metal wiring formed on the surface of the semiconductor substrate is suppressed, and the activation of impurities existing in the deep region is activated. It can be carried out. In particular, when the minor axis width is longer than 3 μm and shorter than 50 μm, it is possible to suppress excessive heat diffusion and perform stable activation that is not easily affected by variations in beam shape and peak value. be able to.

Furthermore, the energy density at which the impurity layer can maintain non-melting depends on the pulse width and the beam size, particularly the minor axis width of the beam. For example, when the pulse width is 500 ns and the short axis width of the beam is 50 μm, the impurity layer can maintain non-melting, and the energy density that can be increased to the vicinity of the melting point is about 2.0 J / cm ² . If the pulse width is about 500 ns, it is difficult to be affected by fluctuations in the pulse, and the heating and cooling speeds are moderated, so the variation in temperature distribution is small. When the minor axis width is 3 μm, about 2.5 mJ / cm ² is preferable in consideration of diffusion from the center of the beam in the lateral direction. When the pulse width is 2000 ns, the optimum energy density is about 3.5 mJ / cm ² when the minor axis width is 50 μm, and the optimum energy density is about 5.0 J / cm ² when the beam width is 3 μm. ² . When the energy density is further increased, it becomes easy to melt. From the above, the optimal energy density is 2.0 J / cm ² or more 5.0J / cm ^2.

  Next, when irradiating a plurality of pulses as a series of continuous pulses, if the time difference between each pulse is too short, the impurity layer is likely to melt, and if the time difference is too long, the effect as a continuous pulse is not achieved and deep. Activation of the area is difficult. By irradiating with an optimal time difference, the impurity layer can be maintained in a non-molten state and activated up to a deep region. The time difference is preferably longer than the pulse width of the first pulse and less than 3 times.

  When the total pulse width of a series of continuous pulses is 1000 ns or less, activation of a sufficiently deep region is difficult, and when the width is 10 μs or more, it affects metal wiring formed on the surface of the semiconductor substrate when a series of continuous pulses are irradiated. May reach. Therefore, in order to suppress the influence of heat on the surface of the semiconductor substrate and activate a deep region at the time of laser irradiation, the total pulse width of a series of continuous pulses must be wider than 1000 ns and narrower than 10 μs. Is desirable.

  After a series of continuous pulses, if the surface temperature of the semiconductor substrate is higher than before the next series of continuous pulses before the next series of continuous pulses, the repeated series of pulses It is thought that the metal wire is gradually warmed and affects metal wiring. Therefore, the repetition frequency of the laser oscillator is the frequency at which the surface temperature of the semiconductor substrate on which the metal wiring is formed is about the same as before irradiation after the irradiation of a series of continuous pulses. To do.

  A semiconductor substrate to be activated, particularly silicon, has a low absorption coefficient with respect to the Yb: YAG second harmonic and is not easily absorbed by the surface layer, so that the laser beam can easily reach a deep region. For this reason, it can be said that application to activation of impurities existing in a deep region is effective.

According to the present invention, the first impurity is injected from the surface layer to a certain depth, and a plurality of pulse lasers are irradiated to the same position with a time difference to a continuous layer in which the second impurity is further implanted in a deeper region, The impurity layer is activated in a non-molten state when irradiated with a plurality of pulse lasers. Since a single pulse laser having a wide pulse width is used, even if it is in a non-molten state at the time of laser irradiation, since the heating time is long, heat is diffused to a deep region, and activation of the deep region is facilitated. In addition, since the impurity layer to be activated is not melted at the time of laser irradiation, the diffusion of impurities is difficult to occur, and a semiconductor substrate in which different impurities are continuously present, or layers having different implantation conditions are continuous even if the same impurities are present. This is effective for activating deep regions such as semiconductor substrates.
Even in the case of a semiconductor substrate structure in which a p-type impurity layer such as boron is formed in a shallow region from the back surface of the semiconductor substrate, and an n-type impurity layer such as phosphorus is formed in a deeper region, as in the case of IGBT, Since the layers are not melted, activation can be performed to prevent each diffusion at the interface between successive impurity layers.

It is a figure which shows the influence which the pulse width of a pulse laser has on impurity layer temperature. It is a figure which shows the cross-section of FS type IGBT which can apply the method of one Embodiment of this invention. It is a figure which shows the pulse waveform which accumulated the laser irradiated from two lasers in the Example of this invention. Similarly, it is a figure which shows the temperature change of the semiconductor substrate by the difference in a pulse width. Similarly, it is a figure which shows the temperature change of the semiconductor substrate by the time difference of a pulse. Similarly, it is a figure which shows the temperature change of the semiconductor substrate by total pulse time. Similarly, it is a figure which shows the carrier concentration distribution of a boron and phosphorus when irradiated with a Yb: YAG second harmonic laser. Similarly, it is a figure which shows the impurity distribution of the boron before and behind Yb: YAG 2nd harmonic laser irradiation.

In the present invention, an impurity present in a semiconductor substrate is activated by laser irradiation. Below, the example applied to FS type | mold among said IGBT as one Embodiment of this invention is demonstrated.
In a semiconductor substrate used for a power device such as an FS type IGBT (insulated gate bipolar transistor), a p-type impurity layer exists in a shallow region, and an n-type impurity layer exists in a region deeper than the p-type impurity layer. Sufficient activation of the layer is required.

  FIG. 2A is an example of a cross-sectional structure of an FS type IGBT that can be processed in the present invention. A p-type base region 13 in which boron is implanted is formed on the surface side of the semiconductor substrate 10, and an n-type emitter region 14 in which phosphorus is implanted is formed on a part of the surface side of the p-type base region 13. . A p + -type collector layer 12 in which boron is implanted is formed in the surface layer on the back side of the semiconductor substrate 10. An n-type buffer layer 11 in which phosphorus is implanted so as to be in contact with the collector layer 12 is formed in a region deeper than the collector layer 12, and an n-type substrate 15 is located inside thereof. Electrode, 22 is an emitter electrode, 24 is a gate oxide film, and 23 is a gate electrode.

The semiconductor impurity layer is irradiated with the plurality of pulse lasers at the same position with a time difference as a series of continuous pulses. The plurality of pulse lasers are usually emitted from a plurality of laser oscillators with timing adjusted. Therefore, the pulse laser repeatedly irradiated based on the repetition frequency does not correspond to a series of a plurality of pulse lasers referred to here. It is possible to irradiate a series of continuous pulses further based on the repetition frequency following the series of continuous pulses.
As the laser oscillator, one using Yb: YAG second harmonic is preferable. As described above, the pulse laser preferably has a pulse width longer than 500 ns and 2000 ns or less, and the pulse width of the second pulse laser is the same as or longer than the pulse width of the first pulse laser. The pulsed laser, the energy density of one pulse at the irradiated surface by means of an attenuator to be placed in the laser oscillator and an optical path is adjusted to be within the scope of ^{2.0J / cm 2 ~5.0J / cm 2} .

  The pulse laser is appropriately shaped by an optical member or the like, and is preferably applied to the semiconductor substrate in a beam shape having a minor axis width larger than 3 μm and smaller than 50 μm. In irradiation with a series of continuous pulses, each pulse need only satisfy the above-mentioned pulse width, energy density, etc., and need not have the same irradiation conditions.

  In irradiation with a plurality of pulse lasers as a series of continuous pulses, the irradiation timing is set so that the pulse time difference is longer than the pulse width of the laser pulse of the previous irradiation and less than three times. In the irradiation with a plurality of pulse lasers, the number of laser pulses in a series is not particularly limited, but the total pulse width is preferably longer than 1000 ns and shorter than 10 μs.

  In the above-described series of continuous pulse irradiations, it is desirable to set the repetition frequency so that the surface side of the semiconductor substrate has the same temperature as before the irradiation until the next series of continuous pulses is irradiated. This prevents the surface side of the semiconductor substrate from being warmed by repeating a series of continuous pulse irradiations. While the entire surface of the semiconductor substrate is activated while performing a scan using a series of continuous pulses, the overlap ratio between the major axis and the minor axis is not particularly limited as the present invention.

  In this embodiment, the semiconductor substrate 10 has been described as an object to be irradiated with the laser. However, the semiconductor substrate 10 is not limited to the semiconductor of the configuration, and the p-type region and the n-type region exist continuously in the depth direction. Therefore, it can be applied to a semiconductor that needs to be activated to a deeper position. Therefore, either the p-type region or the n-type region may exist on the back side of the semiconductor substrate. In addition, in the impurity layer, the first impurity and the second impurity may be different from each other, and the same type of impurity layer may be continuous, and the impurity concentration and the implantation method are different. It can also be applied to cases.

Example 1
A semiconductor substrate having a thickness of 50 μm and an impurity layer is present on a Yb: YAG second harmonic, and pulse widths 500 ns, 1000 ns, and 2000 ns, and the delay time (pulse time difference) between the two pulse lasers is twice the pulse width. A change in the temperature of the semiconductor when two pulse lasers of 1000 ns, 2000 ns, and 4000 n were continuously irradiated was examined. The delay time at this time is represented by the time from the position where the first pulse is present to the position where the second pulse is present, such as the time from the peak to the peak of the laser intensity of each pulse shown in FIG.
The energy density was set to the maximum energy density at which the impurity layer can maintain a solid state, and the energy density of a single pulse was set to 2.0 J / cm ² , 2.5 J / cm ² , and 3.0 J / cm ² , respectively.

  FIG. 4 shows the back surface of the semiconductor substrate irradiated with laser under the above conditions, the depth of 2.0 μm from the back surface of the semiconductor substrate, and the temperature reached on the surface of the semiconductor substrate. Even so, when the depth is 2.0 μm, the reached temperature is about 800 degrees at a pulse width of 500 ns, reaches 1100 degrees or more at a pulse width of 1000 ns and 2000 ns, and reaches a temperature that is considered necessary for activation. Yes. On the surface side of the semiconductor substrate, it reaches nearly 250 degrees at 2000 ns. When a wider pulse than this is irradiated, there is a concern about the influence on the metal wired on the surface side, so a pulse having a wider pulse width than 2000 ns is not preferable.

(Example 2)
A semiconductor substrate with a thickness of 50 μm and an impurity layer is present on a Yb: YAG second harmonic, with a pulse width of 2000 ns and an energy density of 3.0 J / cm ² . A comparison was made of the reached temperatures when a plurality of pulse lasers of 3 ns, 2000 ns, 4000 ns, and 6000 ns were irradiated.

  FIG. 5 shows the temperature reached on the back surface of the semiconductor substrate irradiated with laser under the above conditions, the depth of 2.0 μm from the back surface, and the surface of the semiconductor substrate. When the delay time is 2000 ns (the pre-irradiation pulse width is 1 time), it can be seen that the temperature of the back surface of the semiconductor substrate exceeds the melting point and is melted. That is, when the energy density is fixed and the time difference in a series of continuous pulses is changed, if the time difference is shortened, the back surface of the semiconductor substrate is likely to reach a high temperature, and the applied energy density must be such that the surface does not melt. . By reducing the energy density, it is possible to prevent melting of the back surface of the semiconductor substrate, but the heat diffusion range is narrowed and the activation depth is shallow. From these facts, it can be said that if the time difference in a series of continuous pulses is too short, it is difficult to activate the impurity layer to a sufficiently deep region without melting. In addition, when the time difference is three times the pulse width of the irradiation pulse, the temperature of the back surface of the semiconductor substrate hardly reaches a high temperature. Therefore, it can be seen that in the vicinity of the depth of 2.0 μm, it is already less than 1000 degrees and it is difficult to activate the deep region. It is possible to give a higher energy density, and the deep region can be the temperature required for activation, but the energy density given is determined by the maximum energy density, and in order to give a higher energy density than that, it is common The beam size is reduced to obtain a desired energy density. However, when the beam size is reduced, the throughput is lowered, which is not preferable. Considering that the time difference is twice the pulse width of the irradiation pulse, there is no influence on the metal wiring on the surface of the semiconductor substrate, and the impurity layer is not melted on the back surface, and is necessary for activation even at about 2.0 μm. It can be seen that the temperature has reached.

(Example 3)
A semiconductor substrate with a thickness of 50 μm and an impurity layer is formed on a Yb: YAG second harmonic, pulse widths of 1000 ns and 2000 ns, and the energy density of a single pulse is 2.0 J / cm ^2, a pulse laser was 3.0 J / cm ^2, were evaluated temperature change of the semiconductor substrate surface when irradiated by repeating each pulse. The result is shown in FIG. As the pulse time of a series of continuous pulses becomes longer, the surface temperature increases. As a result, the temperature approaches the temperature affecting the metal wired on the surface of the semiconductor substrate, so that a deep region can be activated within a range shorter than 10 μs that does not affect the metal wiring.

  In Examples 1 to 3, a case where the thickness is thinner than a conventionally assumed semiconductor substrate is shown. From the above, the present invention can be applied to a thin semiconductor substrate that is difficult to be activated due to a thermal influence on a metal wiring or a resin fixture.

Example 4
A method for manufacturing a semiconductor manufacturing apparatus according to the present invention will be described below. At a tilt angle of 0 degree, p-type impurity boron and n-type impurity phosphorus are implanted into a semiconductor substrate under the conditions of ²⁰ keV, 2E13 / cm ² , and 600 keV, 2E12 / cm ² , respectively, at a frequency of 10 kHz, Yb: YAG. A pulse laser having a pulse width of about 1200 ns, a delay time of about 2400 ns, and an energy density of about 2.5 J / cm ² was irradiated using the second harmonic. FIG. 7 shows the carrier concentration distribution in the depth direction at that time. It can be seen that both the p layer and the n layer are activated, and the n layer is activated to a depth of about 2.5 μm. Moreover, it turns out that it was activated, maintaining the solid phase also near the irradiation surface which reaches high temperature. Further, if this condition is used, it is possible to activate a region of 2.5 μm or less, but it is possible to efficiently activate by changing the pulse width and energy density. Further, from FIG. 8 showing the impurity distribution before and after the irradiation, it can be seen that boron corresponding to the first impurity layer has almost no change in the concentration distribution due to the laser irradiation, does not cause diffusion, and is not melted.
In the above embodiment, activation in the case where different types of impurity layers exist continuously has been described. However, when the present invention is used, a semiconductor substrate in which layers having the same impurity and different implantation conditions exist continuously exist. It is also effective for activation. That is, the present invention can also be applied to the case where the layer existing in the shallow region and the layer existing in the deep region are the same impurity.

10 Semiconductor substrate 11 Buffer layer 12 Collector layer

Claims (8)

  The first impurity is present at a certain depth from the surface layer of the semiconductor substrate, and the plurality of pulse lasers are moved to the same position as a series of continuous pulses with a time difference to a continuous impurity layer in which the second impurity exists in a deeper region. Irradiating and performing the irradiation while relatively scanning a predetermined range of the semiconductor substrate for each series of continuous pulses, and activating the impurity layer in a non-molten state during the series of continuous pulses. A method for activating semiconductor impurities.   The time width per pulse of the pulse laser (hereinafter referred to as pulse width) is longer than 500 ns and not longer than 2000 ns, and the pulse width of the second and subsequent pulse lasers among the series of continuous pulses is first. 2. The semiconductor impurity activation method according to claim 1, wherein the pulse width is equal to or longer than the pulse width of the pulse laser.   3. The semiconductor impurity activation method according to claim 1, wherein the width of the short axis of the pulse laser is wider than 3 [mu] m and smaller than 50 [mu] m. The energy density of a single pulse of the semiconductor irradiated surface of the pulsed laser 2.0 J / cm ² or more and 5.0J / cm ² or less, the energy density of the impurity layer is to maintain the non-molten state at the time of laser irradiation The semiconductor impurity activation method according to claim 1, wherein the semiconductor impurity activation method is provided.   In the irradiation with a plurality of pulse lasers as the series of continuous pulses, the irradiation timing is set so that the pulse time difference is longer than the pulse width of the laser pulse previously irradiated and less than three times. The method for activating a semiconductor impurity according to claim 1.   6. The semiconductor according to claim 1, wherein a total of pulse widths of the plurality of pulse lasers irradiated to the same position in the semiconductor substrate having the impurity layer is longer than 1000 ns and shorter than 10 μs. Impurity activation method.   The repetition frequency of the pulse laser is such that the semiconductor is irradiated with a plurality of pulse lasers as the series of continuous pulses, and the semiconductor is irradiated with a plurality of pulse lasers as the next series of continuous pulses. The semiconductor impurity activation method according to claim 1, wherein the semiconductor impurity activation method is an interval.   8. The semiconductor impurity activation method according to claim 1, wherein the pulse laser is a Yb: YAG second harmonic.

Images