JP2009060125A - Semiconductor device, and manufacturing method thereof - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a semiconductor device wherein a work function of a metal electrode provided on an insulating film made of a material having a higher dielectric ratio than SiO<SB>2</SB>has a desired value. <P>SOLUTION: The semiconductor device includes: a semiconductor substrate 34; a tunnel insulating layer 36 formed on the semiconductor substrate; a floating gate electrode 37 provided on the tunnel insulating layer; an inter-electrode insulating film 38 including a first insulating layer 38a formed on the floating gate electrode and made of a material having a high-dielectric ratio and a second insulating layer 38b formed on the first insulating layer and containing silicon, oxygen, and nitrogen or containing silicon and nitrogen; a control gate electrode 40 formed on the inter-electrode insulating film; an interface layer 44 formed in an interface between the second insulating layer and the control gate electrode and containing a 13-group element; and a source-drain region 35 formed in the semiconductor substrate on both sides of the control gate electrode. As for the number of bonding states of the 13-group elements in the interface layer, the number of metal bonding states is larger than the total number of oxide, nitride, or oxide and nitride bonding states. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a semiconductor device and a manufacturing method thereof.

  Silicon super integrated circuits are one of the fundamental device technologies that will support the advanced information society in the future. High performance due to high functionality, high speed, low power consumption, and the like has been realized by miniaturization of semiconductor elements such as CMOS (Complementary Metal Oxide Semiconductor) logic devices and flash memories. However, in recent years, it has been difficult to achieve high performance by miniaturization of elements due to various physical property limitations.

  In CMOS logic devices, conventional gate electrodes using silicon are exposed to increased gate parasitic resistance as the device operating speed increases, and the effective capacitance of the insulating film is reduced and added due to carrier depletion at the interface with the insulating film. Problems such as variations in threshold voltage due to penetration of impurities into the channel region have been pointed out. In order to solve these problems, metal gate technology has been proposed. In metal gate technology, replacing conventional silicon with refractory metal materials solves problems such as gate parasitic resistance, capacity depletion due to depletion, and removal of impurities all at once, but complexity of device manufacturing technology is of great concern. there were.

  In order to solve this concern, so-called full silicidation, in which a CMOS transistor is formed by conventional silicon gate technology, and then the silicon gate is chemically reacted (silicidation) with the metal to change to silicide to obtain a metal gate. Technology (hereinafter also referred to as FUSI (Fully silicided gate)) has been developed. With this method, the gate processing, post-oxidation, and other steps can be realized with the conventional silicon gate technology, and this is a highly practical technology.

The excellent point of the FUSI technique is that the effective work function of silicide can be modulated by the kind and amount of impurity elements added to the silicon gate before silicidation. During the silicidation, impurities added to the silicon gate are segregated at the interface between the gate insulating film made of SiO ₂ and the silicide due to the so-called snow plowing effect. The original work function of the silicide material is modulated by the action of the segregated impurities. Specifically, in the case of a P-channel MIS transistor, boron and aluminum segregated at the interface increase the work function of silicide, and in the case of an N-channel MIS transistor, phosphorus, arsenic, antimony and the like segregated at the interface work By reducing the function, it becomes easy to set the threshold voltage (Vth) of both transistors to an appropriate value.

On the other hand, for the gate insulating film, introduction of a High-k material having a higher relative dielectric constant is essential for the purpose of suppressing an increase in device power consumption due to an increase in leakage current. Here, in the future product, the combination of the above-mentioned FUSI technology and High-k material is inevitable, and it has been thought that remarkable performance enhancement of the CMOS logic device can be realized. However, in reality, due to the peculiarity of the interface between the silicide and the high-k material, the flat band voltage Vfb of the silicide / high-k material system is completely different from the case of the conventional gate insulating film made of SiO _2. As a result, it has been reported that the control of the threshold voltage Vth of the MOS transistor becomes very difficult (for example, see Non-Patent Document 1). When the material of the insulating film is changed from SiO ₂ to High-k, the apparent work function of silicide is reduced, and the work function modulation action due to impurities is lost.

Unless the problem of difficulty in controlling the threshold voltage Vth is solved, high performance of the CMOS logic device by the FUSI / High-k system cannot be achieved. Here, there is a report that the abnormal shift of the flat band voltage Vfb is improved by arranging the SiO ₂ layer at the interface between the silicide and the insulating film made of High-k. According to this report, HfO ₂ is used as the material of the High-k film, and nickel silicide (NiSi) is used as the silicide. However, as in Non-Patent Document 1, the work function of NiSi is higher than that of SiO _2. In addition to the decrease, the work function modulation effect by impurities such as phosphorus and boron almost disappears, making it difficult to obtain a desired threshold voltage Vth.

On the other hand, a SiO ₂ layer having a thickness of 1.1 nm is disposed at the interface between NiSi and HfO _2, and BF ₂ is ion-implanted into the silicon gate and then silicidated, whereby the gate insulating film is made of SiO _2. It has been reported that the effect of increasing the work function, which is almost the same as the case, can be obtained (for example, see Non-Patent Document 2). However, the reason why such an action is obtained is not described in Non-Patent Document 2.
K. Takahashi et al., “Dual Workfunction Ni-Silicide / HfSiON Gate Stacks by Phase-Controlled Full-Silicidation (PC-FUSI) Technique for 45nm-node LSTP and LOP Devices”, 2004 IEDM, pp91-94 CFHuang and BYTsui, “Analysis of NiSi Fully-silicided Gate on SiO2 and HfO2 for CMOS Application”, 2005 SSDM, pp506-507

Non-Patent Document 2 discloses that a material having a low relative dielectric constant such as SiO ₂ is laminated with a high-k film in order to increase the capacity of the gate insulating film as much as possible in a CMOS logic device for high-speed operation. Such a structure is not practical.

On the other hand, in order to increase the functionality of a stack type memory cell (for example, substrate / tunnel insulating film / floating gate electrode / interelectrode insulating film / control gate electrode) used for flash memory or the like, the control gate electrode is provided with a conventional N ⁺ It is effective to change from polysilicon to metal gate. This is because the depletion layer inside the silicon gate disappears, the voltage drop generated in the depletion layer can be ignored, and the voltage of the control gate electrode can be effectively used by writing / erasing information to the floating gate electrode. . As a specific effect, the multi-value operation is facilitated by widening the memory window width of the threshold voltage Vth.

  As the material for the control gate electrode, as in the case of CMOS logic devices, it is desirable to use a material that is highly compatible with the conventional LSI process. Therefore, the stack type memory cell is formed by the conventional polysilicon gate technology and controlled by the back-end process. A FUSI technique for siliciding the polysilicon of the gate is promising.

  Another requirement for the properties of the control gate electrode is that its work function is as high as possible. This is because there is an effect of reducing the leakage current of the interelectrode insulating film during the write operation, and this makes it possible to enlarge the Vth memory window.

  On the other hand, as an interelectrode insulating film, there is a demand to increase the electrical coupling between the control gate electrode and the floating gate electrode while keeping the leakage current low, and this is achieved by using a High-k material. . In particular, the miniaturization of memory cells in the near future will reach a limit, and the shape of the cells must be changed from a three-dimensional type to a planar type. In the planar type, it is the most effective method to increase the capacitance value of the interelectrode insulating film by the polarization of the insulating film itself, and the use of a high-k material for the interelectrode insulating film is an unavoidable technical flow. .

As described above, as the control gate electrode of the stacked memory cell, it is most effective to use a material having a high work function as much as possible by the FUSI technology that is familiar to the conventional LSI manufacturing process. Is required to use a high-k material. However, as described in the CMOS logic device, it is not easy to set the silicide work function high because of the specificity of the interface between the silicide and the High-k film. As already described, in the case of NiSi, the work function of NiSi is lowered only by changing the material of the insulating film from SiO ₂ to High-k. Even if an attempt is made to increase the work function by segregating boron, aluminum or the like at the interface, the modulation effect due to the impurities is almost lost due to the specificity of the interface between NiSi and the High-k film.

The present invention has been made in consideration of the above circumstances, and a semiconductor device in which the work function of a metal electrode provided on an insulating film made of a material having a dielectric constant higher than that of SiO ₂ has a desired value, and It aims at providing the manufacturing method.

  A semiconductor device according to a first aspect of the present invention includes a semiconductor substrate, a tunnel insulating layer formed on the semiconductor substrate, a floating gate electrode provided on the tunnel insulating layer, and formed on the floating gate electrode An inter-electrode insulating film having a first insulating layer made of a high dielectric constant material, and a second insulating layer formed on the first insulating layer and containing silicon and oxygen and nitrogen or containing silicon and nitrogen; A control gate electrode formed on the interelectrode insulating film; an interface layer including a group 13 element formed at an interface between the second insulating layer and the control gate electrode; and the semiconductor substrate on both sides of the control gate electrode. And the number of bonded states of the group 13 element in the interface layer is greater than the total number of oxidized, nitrided, or oxynitrided bonded states. Characterized by comprising a register.

  The method of manufacturing a semiconductor device according to the second aspect of the present invention includes a step of forming a tunnel insulating layer on a semiconductor substrate, a step of forming a floating gate electrode layer on the tunnel insulating layer, and the floating gate electrode. Forming a first insulating layer made of a high dielectric constant material on the layer; forming a second insulating layer containing at least silicon and nitrogen on the first insulating layer; and silicon on the second insulating layer A step of forming a layer, a step of forming a metal layer on the silicon layer, a step of chemically reacting the silicon layer and the metal layer by heat treatment, and converting the silicon layer into a metal silicide layer, Introducing a group 13 element into the metal silicide layer; diffusing the group 13 element by heat treatment; and at least part of the element is unevenly distributed in a region including an interface between the second insulating layer and the metal silicide layer Characterized in that it comprises a step of, a.

According to the present invention, it is possible to provide a semiconductor device in which the work function of a metal electrode provided on an insulating film made of a material having a dielectric constant higher than that of SiO ₂ has a desired value.

  Embodiments of the present invention will be described below with reference to the drawings. In addition, the same code | symbol shall be attached | subjected to a common structure through embodiment, and the overlapping description is abbreviate | omitted. Each figure is a schematic diagram for promoting explanation and understanding of the invention, and its shape, dimensions, ratio, and the like are different from those of an actual device. However, these are in consideration of the following explanation and known techniques. The design can be changed as appropriate.

  In each embodiment, a CMOS transistor and a stacked memory cell will be described. However, a memory in which the above-described semiconductor elements are integrated, a logic circuit, and a system LSI in which these are mixedly mounted on the same chip are also included in the present invention. Is within the range.

  First, before explaining the embodiments of the present invention, the background to the present invention will be described.

As already described in the prior art, the technical problem in the case where the FUSI electrode is formed directly on the High-k film is described in Non-Patent Document 1 and the like, and between the FUSI electrode and the High-k film. Non-patent document 2 describes a case where an SiO ₂ layer is provided as an interface insulating layer between a FUSI electrode and a High-k film in order to avoid a phenomenon peculiar to the interface.

In future LSIs, it is highly likely that SiON is used instead of SiO ₂ as the material of the gate insulating film. This is because the Si (O) N film has a higher dielectric constant than that of the SiO ₂ film, so that even if the Si (O) N film is stacked with the high-k film, the total gate insulating film thickness is not greatly increased, and the gate capacitance is increased. This is because the leakage current can be reduced and the performance of the CMOS transistor is not impaired.

  Accordingly, the present inventors have arranged the SiON film and the SiN film at the interface between the silicide and the High-k film, thereby preventing the capacitance of the gate insulating film from deteriorating while exhibiting the function of modulating the work function of silicide due to impurities. An experiment was conducted on the assumption that a semiconductor device including a FUSI electrode and a high-k film could be realized.

However, the present inventor has found that even if the conventional impurity addition technique using the snow shoveling effect is combined with the silicide / Si (O) N / High-k layering technique, a desired work function modulation effect due to impurities cannot be obtained. Found. This has been found to be particularly significant in the case of impurities for P-channel MIS transistors such as boron and aluminum. The cause is as follows. When the insulating film is SiO ₂ , boron, aluminum, etc. penetrates into the insulating film in the process of the snow removal effect accompanying silicidation and acts as an electric dipole at that position to increase the work function of silicide. . On the other hand, when nitrogen is contained inside the insulating film, such as SiON, SiN, etc., it has been found that boron or aluminum forms an electrically active defect by forming a compound with nitrogen. This defect caused by nitrogen in the insulating film cancels out the electric dipole, and the desired work function increasing effect cannot be obtained. By the way, for the mechanism of the work function modulation by the impurities in the FUSI / SiO ₂ system, Y. Tsuchiya et al., "Physical Mechanism of Work Function Modulation due to Impurity Pileup at Ni-FUSI / SiO (N) Interface", 2005 IEDM, It is explained in pp637-640.

  The above phenomenon and cause will be described in more detail with reference to FIGS. 1 (a) to 8 (b).

FIG. 1A and FIG. 1B show experimental results when the phenomenon is confirmed. 1A and 1B, the horizontal axis indicates the voltage applied to the gate electrode, and the vertical axis indicates the capacitance of the gate insulating film. In this experiment, NiSi is used as the silicide. When SiO ₂ is used for the gate insulating film, the flat band voltage Vfb is increased in the negative voltage direction when arsenic is added and in the positive direction when boron is added, compared to the case where no impurity is added to the silicon gate. There is a shift. In FIGS. 1A and 1B, the dose amount of both impurities is 1 × 10 ¹⁶ atoms / cm ² . This characteristic is as shown in Non-Patent Documents 2 and 3 (FIG. 1A). On the other hand, when SiON is used as the gate insulating film, a Vfb shift due to arsenic is obtained, but a Vfb shift due to boron cannot be confirmed (FIG. 1B).

This phenomenon can be understood by the following model taking boron as an example (see FIGS. 2A and 2B). When SiO ₂ is used as the gate insulating film, among the boron segregated at the interface with NiSi, those distributed on the NiSi side show a metallic bonding state, and those distributed on the SiO ₂ side show an oxidative bonding state. . Here, the metallic bonding state refers to a case where, when attention is paid to one boron atom, the atom bonded to the boron atom is a metal such as silicon or nickel. On the other hand, an oxidative bond state refers to a case where an atom bonded to a boron atom is oxygen when attention is paid to one boron atom. Metallic boron showing a metallic bonding state has no effect on the work function φ of NiSi, whereas boron oxide showing an oxidized bonding state is present within about 1 nm from the interface at the interface. It produces the functional dipole and increases the work function φ of NiSi. On the other hand, boron oxide existing in the SiO ₂ region separated from the interface by about 1 nm or more has no effect on the work function φ (see FIG. 2A).

On the other hand, when SiON is used for the gate insulating film, there is no change in the state of boron distributed on the NiSi side. On the other hand, since boron distributed in the gate insulating film is combined with nitrogen to obtain a large energy gain, more boron enters the SiON film than in the case of SiO ₂ . Furthermore, since boron bonded to nitrogen acts as a defect having a positive charge and shifts the flat band voltage Vfb to the negative side (see FIG. 2B), this is the work function φ of the electric dipole near the interface. As shown in FIG. 1 (b), no change in the work function φ can be obtained by adding boron.

With regard to the new finding that the work function increasing effect by boron addition cannot be obtained in the case of SiON, we understood the mechanism as described above. Based on this understanding, it was considered that the desired work function modulation can be obtained by appropriately controlling the bonding state of segregated substances such as boron at the silicide and SiON interface. This is the greatest point of one embodiment of the present invention. Specifically, as shown in FIG. 3, by reducing the amount of boron that penetrates deeply into the SiON film and forms positive charges, the action of the electric dipole at the interface is the same as when the SiO ₂ film is used. I thought that it could be demonstrated.

  Here, it has been found that a method of adding boron after silicide formation (described as post-addition) is effective as a technique for creating a boron distribution as described in FIG. FIG. 4 is an experimental result showing that the work function of the NiSi / SiON system was successfully increased by about 0.2 eV by boron post addition. It is also shown that the NiSi / SiON work function can be reduced by about 0.3 eV by post addition of phosphorus (P).

  In FIG. 4, the horizontal axis indicates the film thickness of SiON, and the vertical axis indicates the flat band voltage Vfb. FIG. 5 shows the experimental results obtained by examining the bonding state of boron segregated near the interface at this time by XPS (X-ray photoelectron spectroscopy). This experimental result shows that a NiSi / SiON structure prepared by removing all Si parts by a combination of physical polishing and chemical solution etching is measured from the SiON side by XPS with respect to the NiSi / SiON / Si laminated structure. This is obtained by analyzing the impurity (boron) bonding state at the NiSi / SiON interface. The XPS signal intensity decays exponentially depending on the depth from the sample surface. In the same experiment, the SiON film thickness used in the method of silicidation after adding boron to silicon as in the conventional case, that is, the method of adding boron before silicide formation (pre-addition) is better. The absolute value of the signal intensity of boron that is slightly thicker and buried under the SiON is weaker when pre-added. Therefore, when comparing the pre-addition and post-addition (invention) in the figure, attention is paid only to the relative strength ratios of the individual bonded components. In the pre-addition, a large amount of boron penetrates into the SiON, and the oxidation / nitridation / oxynitride state strength is relatively higher than the metal bond state strength. Oxidation / nitridation / oxynitridation state refers to a state in which, when focusing on one boron atom, the atom bonded to the boron atom is only oxygen, only nitrogen, or both oxygen and nitrogen are bonded simultaneously. Point to. On the other hand, in boron post addition according to an embodiment of the present invention, boron in an oxidized / nitrided / oxynitrided bonded state is relatively smaller than a boron component in a metal bonded state. Considering from the XPS principle that the signal intensity increases exponentially with the component closer to the sample surface, the strength of the oxidized / nitrided / oxynitrided bond state that should be closer to the surface is weaker than the metal bond state strength. This suggests that the structure contains much less oxidized, nitrided and oxynitrided bonds than metal bonds. This inference agrees with the boron distribution / bonding state schematically shown in FIGS.

In the boron post addition method, we used an ion implantation method. In this case, comparing the case of using boron ions and the case of using BF ₂ ions, it was found that only when boron ions were used, the work function increase effect aimed at by one embodiment of the present invention was obtained ( (See FIG. 6). FIG. 7 shows the result of XPS experiment in which the bonding state of the interface boron segregated material in the above experiment was examined. In the BF ₂ ion implantation, the absolute amount of boron reaching the interface is small, and as a result, the work function modulation effect as shown in FIG. 6 is not obtained. On the other hand, when boron ion implantation is performed, since the boron distribution and bonding state as described above can be realized, the work function can be modulated. FIGS. 8A and 8B schematically show the segregation state of boron in each case of B ion implantation and BF ₂ ion implantation.

  Such a bonding state of boron can also be identified by analyzing a device according to an embodiment of the present invention. Specifically, the NiSi / SiON laminated structure with boron segregated at the interface is cross-sectionally observed by a normal TEM (Transmission Electron Microscopy) method, and the electron beam of the TEM is irradiated in the vicinity of the segregation layer at the interface, and the loss spectrum of the electron beam. From this, the bonding state of boron can be known. This is a so-called TEM-EELS (Electron Energy Loss spectroscopy) method. According to the TEM method, the structural interface of NiSi / SiON is clearly shown. Therefore, by irradiating the target area with the electron beam of TEM and acquiring the loss spectrum, boron in a metal-bonded state is about 1 nm from the NiSi / SiON interface in the NiSi side region, and oxidized / nitrided in the about 1 nm region on the SiON side. -It is shown that boron in an oxynitride bond state is a dominant component. As shown in the experimental results of FIGS. 5 and 7, there is an energy difference of about 4 eV between the metal bond state of boron and the oxidized / nitrided / oxynitrided bonded state. Therefore, if there is an energy resolution of the TEM-EELS method, the difference is present. Is clearly distinguished.

  In the above description, boron is taken as an example of the segregation impurity, but the same effect can be obtained in the case of aluminum (Al). Since boron and Al have the same valence electron number of 3, the defect formation method in SiON is very similar. In addition, since the bonding force between Al and nitrogen is extremely strong, a defect formation (inhibition of work function modulation by impurities) similar to that caused by the combination of boron and SiON occurs. The above problem can be solved by appropriately controlling the Al distribution in one embodiment of the present invention. Similarly, the action of one embodiment of the present invention can be expected even with Group 13 elements other than Al. Among group 13 elements, boron and Al are well suited to the LSI manufacturing process, and it is desirable to use these. Furthermore, in the present invention, since the impurity element reaches the NiSi / SiON interface by simple thermal diffusion, which is different from the snow shoveling effect of the conventional method, a high diffusion coefficient in the silicide is required. From this viewpoint, boron having a smaller atomic radius and a larger diffusion coefficient is most preferable. Another reason why boron is most suitable is that boron ion implantation is used in normal LSI manufacturing processes.

  In the above description, SiON is taken as an example, but the same applies to the case of a silicon nitride film. The silicon nitride film contains more nitrogen than SiON, so there are many opportunities for defect formation when aiming at segregation of boron, Al, or other group 13 elements. This problem is solved by appropriately controlling the distribution of these impurities according to the present invention.

  According to the configuration described above, Vth reduction that has been hindered by a phenomenon peculiar to the FUSI / High-k interface can be solved by utilizing the interface SiON layer. Thus, a FUSI gate CMOS combined with a gate insulating film made of a high-k material having an effect of optimizing the threshold voltage Vth of the transistor based on the modulation of the work function due to impurity segregation, which is an advantage of the conventional FUSI technology. A semiconductor device including the device can be provided.

(First embodiment)
A semiconductor device according to a first embodiment of the present invention will be described. The semiconductor device of this embodiment is a CMOS transistor, and a cross section thereof is shown in FIG.

  As shown in FIG. 9, the semiconductor substrate 1 is provided with an N-type well region (N-type semiconductor layer) 2 and a P-type well region (P-type semiconductor layer) 3. The N-type well region 2 and the P-type well region 3 are separated by an element isolation layer 4 having an STI (Shallow Trench Isolation) structure.

  A P-channel MIS transistor 19 is provided in the N-type well region 2. The P-channel MIS transistor 19 includes a P-type diffusion layer 5, a P-type extension layer 6, a gate insulating film 9, a buffer layer 10, an impurity segregation layer 11, and a gate electrode layer 12. The gate insulating film 9 is provided on the N-type well region 2, and the buffer layer 10, the impurity segregation layer 11, and the gate electrode layer 12 are formed on the gate insulating film 9. In the present embodiment, sidewalls 17 made of an insulator are provided on both sides of the laminated structure of the gate insulating film 9, the buffer layer 10, the impurity segregation layer 11, and the gate electrode layer 12. Yes.

  The P-type extension layer 6 is provided in the N-type well region 2 on both sides of the laminated structure of the gate insulating film 9, the buffer layer 10, the impurity segregation layer 11, and the gate electrode layer 12. Are provided in the N-type well region 2 on both sides of the sidewall 17. The P-type diffusion layer 5 is configured so that the junction depth with the N-type well region 2 is deeper than that of the P-type extension layer 6. The P-type diffusion layer 5 and the P-type extension layer 6 become the source / drain regions of the P-channel MIS transistor 19.

  An N channel MIS transistor 20 is provided in the P-type well region 3. The N channel MIS transistor 20 includes an N type diffusion layer 7, an N type extension layer 8, a gate insulating film 13, a buffer layer 14, an impurity segregation layer 15, and a gate electrode layer 16. The gate insulating film 13 is provided on the P-type well region 3, and the buffer layer 14, the impurity segregation layer 15, and the gate electrode layer 16 are formed on the gate insulating film 13. In the present embodiment, sidewalls 17 made of an insulator are provided on both sides of the stacked structure of the gate insulating film 13, the buffer layer 14, the impurity segregation layer 15, and the gate electrode layer 16. .

  The N-type extension layer 8 is provided in the P-type well region 3 on both sides of the laminated structure of the gate insulating film 13, the buffer layer 14, the impurity segregation layer 15, and the gate electrode layer 16, and the N-type diffusion layer 7. Are provided in the P-type well region 3 on both sides of the sidewall 17. The N type diffusion layer 7 is configured so that the junction depth with the P type well region 3 is deeper than that of the N type extension layer 8. The N-type diffusion layer 7 and the N-type extension layer 8 become the source / drain regions of the N-channel MIS transistor 20.

In the first embodiment, the gate insulating films 9 and 13, the buffer layers 10 and 14, and the gate electrode layers 12 and 16 use the same material for both channel MIS transistors. In particular, the buffer layers 10 and 14 are made of either a silicon oxynitride film or a silicon nitride film. When the buffer layers 10 and 14 are silicon oxynitride films, the nitrogen concentration of the films is preferably 3 atomic% or more and 15 atomic% or less. This is because an SiON film with few structural defects can be realized and an electrically extremely stable SiON film can be obtained. In the case of a silicon nitride film (SiN film), a complete silicon nitride film having a composition of Si ₃ N ₄ is desirable. This is because an electrically stable SiN film with few structural defects is realized.

  The impurity segregation layers 11 and 15 are regions having a thickness of about 2 nm centering on the interfaces between the gate electrodes 12 and 16 and the buffer layers 10 and 14, respectively. A large amount of impurities other than the elements constituting the buffer layer are segregated.

  The impurities contained in the impurity segregation layer 11 are composed of any one of group 13 elements (group 3B elements), preferably boron or aluminum. This is because these elements have been used in conventional LSIs and can be easily introduced into production. As a result, the work function of the gate electrode effectively increases.

  The impurities contained in the impurity segregation layer 15 are composed of any one of group 15 elements (group 5B elements), preferably phosphorus, arsenic, or antimony. This is because these elements have been used in conventional LSIs and can be easily introduced into production. As a result, the work function of the gate electrode is effectively reduced.

  In particular, in the present embodiment, the combined state of impurities in the impurity segregation layer 11 is that the total number of metallic combined states that are mainly distributed on the gate electrode 12 side is oxidized, nitrided, It is characterized by more than the sum of the oxynitriding states. The total number of bonded states here is proportional to the area of the peak corresponding to each bonded state in the XPS spectrum of FIG. 5, for example. As a result, the work function of the gate electrode is increased by a group 13 element such as boron or aluminum even on SiON or SiN.

  The gate electrodes 12 and 16 are made of a metal silicide having a work function of 4.55 eV or more and 4.75 eV or less. This property is aimed at the work function in the vicinity of the center of the Si band gap, and the N-channel MIS transistor lowers the work function by segregating different types of impurities around this work function. By reducing the value voltage Vth and increasing the work function in the P-channel MIS transistor, the threshold voltage Vth can be lowered, thereby making it possible to realize a substantially symmetrical threshold voltage Vth. As a metal silicide exhibiting such a work function, a metal silicide containing at least one of Ni and Co can be cited, and Ni is most preferably used from the viewpoint of applicability and performance to LSI. Further, these metal silicides may contain about 5% Pt in terms of metal ratio from the viewpoint of improving heat resistance.

The film thickness of the first and second interface buffer layers 10 and 14 made of SiON or SiN is preferably 0.1 nm or more and 1 nm or less as a film thickness converted to SiO ₂ . This is because when the thickness is smaller than 0.1 nm, it is impossible to solve the two major problems of lowering the work function and disappearing the impurity segregation effect peculiar to the interface between the silicide and the High-k film. Further, if it is thicker than 1 nm, it is impossible to realize the high-speed operation required for the CMOS transistor.

  High-k films are used as the gate insulating films 9 and 13. The most desirable high-k film is nitrogen-added hafnium silicate (HfSiON). This is because HfSiON is extremely compatible with a high-temperature silicon gate process (typically 1000 ° C. or higher) essential for FUSI technology, and is currently a high-k material having the best heat resistance. In addition, the SiON / HfSiON stack or the SiN / HfSiON stack unique to the present embodiment has a thermodynamically extremely stable structure due to almost the same elemental structure, and thus its electrical characteristics. It is also because of the stability.

  As can be seen from the description of the circumstances leading to the present invention, in the present embodiment, the High-k films 9 and 13 are not necessarily required, and the High-k films 9 and 13 may not be included in the gate stack. .

  In the present embodiment, the N-type well region 2 and the P-type well region 3 are made of Si or Ge semiconductor.

(Production method)
Next, a method for manufacturing the semiconductor device of the first embodiment will be described.

  A manufacturing process of the semiconductor device of the first embodiment is shown in FIGS.

  First, an N-type well region 2 and a P-type well region 3 separated by an element isolation layer 4 having an STI structure are formed on a semiconductor substrate 1. Subsequently, a gate insulating film 21, a buffer layer 22, and a silicon layer 23 are deposited on the entire surface of the wafer to obtain the structure shown in FIG. Here, ion implantation for channel impurity control for optimizing transistor characteristics may be performed.

As the gate insulating film 21, a high dielectric oxide is used. Here, as an example, 3 nm thick HfSiON (Hf / (Hf + Si) to 0.5, nitrogen concentration 20 atom%) was deposited by MOCVD (Metal organic chemical vapor deposition). As the deposition method, an ALD (Atomic Layer Deposition) method, an MBE (Molecular Beam Epitaxy) method, a PVD (Physical Vapor Deposition) method, or the like can be used.
The composition of HfSiON is not particularly limited to the above composition as far as the effect of the present embodiment is concerned. However, considering the relative dielectric constant from the viewpoint of maintaining the reliability as an insulating film while typically maintaining a value of 10 or more, Hf / (Hf + Si) is desirably 0.3 to 0.8, The N concentration is desirably 20% or less.

  As the interface buffer layer 22, SiON, SiN, or the like is used. As the deposition method, MOCVD, ALD, MBE, or the like can be used. In this embodiment, the silicon layer 23 is deposited by 70 nm using a normal CVD method.

  Next, with respect to the structure shown in FIG. 10, a gate electrode pattern is formed only on a portion on the P-type well region 3 using normal lithography technology and etching technology. Here, first, a SiN hard mask (not shown) having a thickness of 100 nm is formed on the N-type well region and the silicon layer 23 in the P-type well region 3 by using a normal SiN deposition process, a normal lithography technique and an etching technique. Formed. Using this SiN hard mask as a mask, the silicon layer 23 on the P-type well region 3 is patterned by a commonly used etching gas. At this time, the interface buffer layer 22 and the silicon layer 23 can be simultaneously patterned simultaneously with the silicon layer 23 by using a chlorine or fluorine-based etching gas. Further, the gate insulating film 21 exposed on the P-type well region 3 on both sides of the SiN hard mask is completely removed by wet etching such as HF. Further, As ion implantation for extension formation is performed, and the N-type extension layer 8 is formed by high-temperature spike heat treatment. At this time, since the upper portion of the silicon layer 23 is covered with the SiN hard mask, it is not doped with As ions. Thereafter, the photoresist and the SiN hard mask are removed to obtain the structure shown in FIG.

  Next, a SiN hard mask is formed on the silicon layer 23 in the P-type well region 3 and the N-type well region 2 by using a normal SiN deposition process, a normal lithography technique, and an etching technique, and this SiN hard mask is used. The silicon layer 23 and the interface buffer layer 22 in the N-type well region are patterned. Thereafter, the gate insulating film 21 exposed on the N-type well region 2 on both sides of the SiN hard mask is completely removed by wet etching. Further, B ion implantation for extension formation is performed, and a P-type extension layer 6 is formed by high-temperature spike heat treatment. Further, boron is not doped in the silicon layer 23 whose upper portion is covered with the SiN hard mask. Thereafter, the photoresist and the SiN hard mask are removed to obtain the structure shown in FIG.

  Subsequently, sidewalls 17 are formed, and B ions are implanted into the N-type well region 2 to form the diffusion layer 5, and As ions are implanted into the P-type well region 3 to form the diffusion layer 7. During the ion implantation for forming the diffusion layers 5 and 7, the silicon layer 23 of the N-type well region 2 and the P-type well region 3 is covered with a mask layer. After removing the mask layer, an interlayer insulating film 18 is formed, and polishing for planarizing the interlayer insulating film 18, for example, CMP (Chemical Mechanical Polishing) is performed to obtain the structure shown in FIG.

  In this series of steps, the feature of this embodiment is that no impurity is added to the silicon layer 23 in both channel transistors. In the conventional silicon gate technology, ions are also implanted into the gate electrode during ion implantation for extension or diffusion layer formation. However, in this embodiment, a mask layer is provided on the silicon layer 23 during ion implantation to prevent ion implantation. This is because when the implanted impurities are implanted into the silicon layer 23, the subsequent silicon silicidation reaction of the silicon gate is hindered to prevent the formation of a uniform metal silicide layer between both channel transistors.

  Next, in order to form a metal silicide on the entire surface, a metal film 24 was deposited to obtain the structure shown in FIG. Here, Ni was deposited as the metal film 24 by sputtering to a thickness of 40 nm.

  Subsequently, the silicon layer 23 was entirely changed to the metal silicide 25 by heat treatment. Here, NiSi is formed by heat treatment at 450 ° C. for 1 minute. Unreacted Ni can be selectively removed from the NiSi with a sulfuric acid solution, whereby the structure shown in FIG. 15 was obtained.

Further, a resist mask layer 26 is formed only on the N-type well region 2 having the structure shown in FIG. 15, and a group 15 element is ion-implanted into the entire surface of the structure to obtain the structure shown in FIG. Here, phosphorus as a group 15 element was implanted under the conditions of an acceleration energy of 20 keV and a dose of 5 × 10 ¹⁵ / cm ² .

Further, the resist mask layer 26 on the upper portion of the N-type well region 2 having the structure shown in FIG. 16 was peeled off, and a resist mask layer 27 was formed only on the upper portion of the P-type well region 3. Subsequently, group 13 element ions were implanted into the entire surface of the structure to obtain the structure shown in FIG. Here, boron as a group 13 element was implanted under the conditions of an acceleration energy of 5 keV and a dose of 5 × 10 ¹⁵ / cm ² .

  Next, after removing the resist mask layer 27 from the structure shown in FIG. 17 by a normal process, a heat treatment is performed at 450 ° C. for 1 minute, for example, so that among the group 15 and group 13 elements implanted into the silicide layer 25 At least a part is thermally diffused and segregated at the interface between the silicide layer 25 and the buffer layer 22. As a result, the impurity segregation layers 11 and 15 were formed at the interface between the silicide layer 25 and the buffer layer 22 to complete the structure shown in FIG.

  In the present embodiment, for NiSi having a work function of 4.65 eV, the effective work function is 4.85 eV for the P-channel MIS transistor 19 due to the modulation effect of boron, and 4 for the N-channel MIS transistor 20 due to the modulation effect of phosphorus. An effective work function of .35 eV can be realized. As a result, CMOS transistors having appropriate threshold voltages Vth can be realized.

  According to the present embodiment, the combination of the FUSI technique and the High-k technique can be achieved without a significant increase in the thickness of the insulating film. As described above, according to the first embodiment, it is possible to obtain a CMOS device having a gate electrode that has low resistance and heat resistance and does not have a problem of depletion, impurity diffusion, or penetration.

  Further, by using metal silicide for the gate electrodes 12 and 16, it is possible to prevent the increase in the number of steps when manufacturing a CMOS device as much as possible, and a complicated process is not necessary.

(Modification)
As a modification of the first embodiment, when the gate insulating film is made of only SiON, it is manufactured as follows. The only change is the process of depositing the insulating film. Instead of the process of FIG. 10, a SiON film is deposited as the gate insulating film 21 as shown in FIG. 18, and then a silicon layer 23 to be the gate is deposited.

As a first specific manufacturing method of the SiON film, when a silicon substrate is used as the semiconductor substrate 1, a silicon oxide film is deposited by a normal thermal oxidation method, for example, with a film thickness of 1.0 nm to 2.0 nm. The concentration of 1 at. % To 15 at. About nitrogen is introduced into SiO ₂ to form a SiON film.

Further, as a second specific method for producing the SiON film, the atmosphere in the chamber on which the silicon substrate 1 on which the N-type well region 2 and the P-type well region 3 are formed is mixed with N ₂ and NH ₃ . The surface of the silicon substrate 1 filled with the gas and having the N-type well region 2 and the P-type well region 3 formed thereon is maintained at 700 ° C. Thereby, a silicon nitride film is formed on the N-type well region 2 and the P-type well region 3. Next, the temperature of the silicon substrate 1 is raised to 850 ° C. and held as it is. Subsequently, the atmosphere in the chamber is filled with a mixed gas of N ₂ and O ₂ while maintaining the temperature of the silicon substrate 1 at 850 ° C. As a result, a silicon oxynitride layer containing oxygen is formed between the silicon substrate 1 and the silicon nitride film, and a silicon oxynitride layer containing oxygen is formed on the surface of the silicon nitride film. A gate insulating film comprising a film and a silicon oxynitride layer is formed. That is, in the second specific manufacturing method, a gate insulating film having a three-layer structure including a silicon oxynitride layer, a silicon nitride film, and a silicon oxynitride layer is formed.

  After forming the SiON film in this manner, the structure shown in FIG. 19 is completed through the same steps as those shown in FIGS. According to this structure, even when the FUSI technology and the SiON gate insulating film are combined, threshold voltage modulation by segregating boron and aluminum can be performed as designed.

As described above, according to the first embodiment and the modification thereof, the semiconductor device in which the work function of the metal electrode provided on the insulating film made of the material having a dielectric constant higher than that of SiO ₂ has a desired value. Can be obtained.

(Second Embodiment)
Next, a semiconductor device according to a second embodiment of the invention will be described. The semiconductor device of this embodiment is a CMOS transistor, and differs from the first embodiment in the following points.

  In the second embodiment, the gate electrode 12 of the P-channel MIS transistor shown in FIG. 9 is made of a material having a work function of 4.75 eV or more and 5.10 eV or less, and the gate electrode 16 of the N-channel MIS transistor is a work function of 4.20 eV. It is comprised with the material below 4.55eV.

  With the work functions of these gate electrodes, the threshold voltages of the N-channel MIS transistor and the P-channel MIS transistor can be designed to be lower than those in the first embodiment. It is suitable for a CMOS transistor.

As such a gate electrode, at least one metal of Ni and Co is used as the material of the gate electrode 12 of the P-channel MIS transistor, and when the composition ratio of metal to silicon (metal / Si) is 2 or more, for example, Ni A work function of 4.75 eV or more due to a silicide phase such as ₂ Si, Ni ₃₁ Si ₁₂ , Ni ₃ Si, or Co ₂ Si can be realized. On the other hand, as the material of the gate electrode 16 of the N-channel MIS transistor, the same metal as the gate electrode 12 is used, and when the composition ratio of metal to silicon (metal / Si) is smaller than 1, for example, NiSi _{2 or} CoSi ₂ phase. The resulting work function of 4.55 eV or less can be obtained. Moreover, about 5% of Pt may be contained as a metal ratio with respect to the silicide of these metals. This has the effect of improving the heat resistance of the metal silicide and increases the reliability of the device. The added Pt is bonded to silicon and mainly forms a PtSi phase.

  Next, a method for manufacturing the semiconductor device of the second embodiment will be described.

  The manufacturing process of the semiconductor device of this embodiment is shown in FIGS. The manufacturing method of the semiconductor device of the present embodiment will be described by taking as an example the case where Ni (nickel) is used as the metal material of the gate electrode. Even when Co is used, the obtained performance is almost the same.

As in the case of the first embodiment, as shown in FIG. 20, a CMOS device having a gate laminated structure of a gate insulating film 21 made of silicon layer 23 / buffer layer 22 / High-k is obtained. In the second embodiment, materials having different work functions are used for the metal silicide of the P-channel MIS transistor and the N-channel MIS transistor.
In this embodiment, as shown in FIG. 21, a Ni layer (thin film) 28 is sputter deposited only on the P-type well region 2 shown in FIG. Here, the thickness of the silicon layer 23 is 70 nm. Thereafter, Ni and the silicon layer 23 were reacted by heat treatment at 650 ° C. for 30 seconds to form the NiSi ₂ layer 16, and then unreacted Ni was removed with sulfuric acid. In this way, the structure shown in FIG. 22 is obtained. NiSi ₂ is the most silicon-rich compound among the thermodynamically stable silicide phases of the Ni—Si system, and since it is silicon-rich, the work function has the lowest value of about 4.55 eV or less.

Subsequently, as shown in FIG. 23, a Ni layer (thick film) 29 is deposited 110 nm only on the N-type well region 2. Subsequently, the silicon 23 layer on the N-type well region 2 is converted into Ni-rich Ni silicide, such as Ni ₃ Si, by heat treatment at 450 ° C. for 1 minute. At this time, NiSi ₂ above the P-type well region is also subjected to heat treatment, but NiSi ₂ is the most thermodynamically stable phase in the Ni—Si system, and there is no unreacted Ni layer. There is no phase transition to. The work function of Ni ₃ Si is about 4.85 eV. After removing unreacted Ni with sulfuric acid, the impurity segregation layers 11 and 15 are formed by the same process as described in the first embodiment, whereby the CMOS transistor structure shown in FIG. 9 can be obtained.

In this embodiment, the P-channel MIS transistor 19 uses Ni ₃ Si with a work function of 4.85 eV as the gate electrode 12, and an effective work function of 5.05 eV is realized by the modulation effect of boron. In the MIS transistor, NiSi ₂ having a work function of 4.55 eV is used as the gate electrode 16 and an effective work function of 4.25 eV can be realized by the modulation effect of phosphorus. Thereby, a CMOS transistor having a threshold voltage Vth lower than that of the first embodiment can be realized. Note that Ni ₂ Si having a work function of 4.75 eV may be used as the gate electrode 12 of the P-channel MIS transistor 19, and the effective work function may be 4.95 eV due to the modulation effect of boron.

In this embodiment, the threshold voltage Vth can be set lower than that in the first embodiment. However, since Ni silicide is separately formed for both channel transistors, it is difficult to increase the number of processes. is there. Furthermore, a slight deterioration in device characteristics due to a high temperature process (about 650 ° C.) for producing NiSi ₂ becomes a problem. The process for improving this point is shown below. In contrast to the structure shown in FIG. 20, a silicide reaction control layer 30 made of any one of Ti, Zr, and Hf is deposited only on the P-type well region 3 by sputtering, and a Ni layer 28 is deposited by 30 nm sputtering. Thus, the structure of FIG. 24 is obtained. Here, as an example, Ti having a thickness of 5 nm was deposited by sputtering. This structure is heat treated at 450 ° C. for 1 minute to form Ni silicide. In general, Ni-Si-based silicide reactions occur with Ni being the preferential diffusion species. However, in the structure of this embodiment, the Ti layer 30 exhibits the action of slowing down Ni diffusion, so that the amount of Ni supplied to the silicide reaction front can be suppressed. In addition, the NiSi ₂ phase can be obtained at a temperature lower by about 200 ° C. than a normal silicidation reaction without a Ti layer. Thereafter, the same process as described in the second embodiment is performed. By using the silicide reaction control layer 30, there is a sacrifice of an increase in the number of steps, but the silicidation temperature is lowered, so that it is possible to prevent the performance deterioration of the CMOS transistor. From the viewpoint of the similarity between Ni and Co in the preferential diffusion species during silicide formation, Co can be expected to have almost the same effect as Ni. Further, although Ti is exemplified as the silicide reaction suppression layer, it is known that the similar action can be obtained even with Zr and Hf due to the similarity in chemical properties with Ti.

As described above, according to the second embodiment, it is possible to obtain a semiconductor device in which the work function of a metal electrode provided on an insulating film made of a material having a dielectric constant higher than that of SiO ₂ has a desired value. it can.

(Third embodiment)
Next, a semiconductor device according to a third embodiment of the present invention will be described. The semiconductor device of this embodiment is a CMOS transistor, and differs from the second embodiment in the following points.

  In the third embodiment, the gate electrode 12 of the P-channel MIS transistor shown in FIG. 9 is made of a material having a work function of 4.75 eV or more and 5.10 eV or less, and the gate electrode 16 of the N-channel MIS transistor is a work function of 4.20 eV. Although it is the same as that of 2nd Embodiment until it is comprised with the electrode material of 4.55 eV or less above, electrode materials differ.

  As such a gate electrode, in the present embodiment, silicide containing at least one metal of Pt, Ir, Ru, Pd, and Re is used as a material of the gate electrode 12 of the P-channel MIS transistor. With these materials, a work function of 4.75 eV or more and 5.10 eV or less can be realized. On the other hand, as the material of the gate electrode 16 of the N-channel MIS transistor, silicide containing at least one rare earth metal such as W, Ti, Zr, Hf, Ta, Nb, Er is used. With these materials, a work function of 4.20 eV or more and 4.55 eV or less can be realized.

  Next, a method for manufacturing the semiconductor device of the third embodiment will be described.

  A method of manufacturing the semiconductor device of this embodiment is shown in FIGS.

Similar to the first embodiment, a CMOS device having a gate laminated structure of the gate insulating film 21 composed of the silicon layer 23 / buffer layer 22 / High-k is obtained as shown in FIG. In the third embodiment, similarly to the second embodiment, materials having different work functions are used for the metal silicide of the P-channel MIS transistor and the N-channel MIS transistor.
In this embodiment, as shown in FIG. 26, the Pt layer 31 is deposited only on the N-type well region 2 by the 50 nm sputtering method. Thereafter, the Pt layer 31 and the silicon layer 23 (70 nm) were reacted by heat treatment at 450 ° C. for 1 minute to form the PtSi layer 12. Thereafter, unreacted Pt was removed with aqua regia. Subsequently, an Er layer 32 was deposited to 50 nm only on the P-type well region 3 and a W layer 33 was deposited to 30 nm thereon, to obtain a structure shown in FIG. By subjecting this structure to a heat treatment at 450 ° C. for 1 minute, the Er layer 32 and the silicon layer 23 were reacted to form the Er ₃ Si ₅ layer 16. Thereafter, unreacted Er and W layers were selectively peeled with sulfuric acid. The W layer 33 plays a role of suppressing the Er oxidation reaction during the Er silicide formation heat treatment and stabilizing the formation of the Er ₃ Si ₅ layer 16. Thereby, a CMOS transistor having PtSi as the gate electrode for the P-channel MIS transistor and Er ₃ Si ₅ as the gate electrode for the N-channel MIS transistor can be realized. The subsequent process of introducing the impurity segregation layer was performed in the same manner as described in the first embodiment to obtain the structure of the CMOS transistor shown in FIG.

In this embodiment, the effective work function is 5.15 eV for PtSi having a work function of 4.95 eV for the P-channel MIS transistor due to the modulation effect of boron or the like, and Er ₃ having a work function of 4.35 eV for the N-channel MIS transistor. An effective work function of 4.05 eV can be realized for Si ₅ by a modulation effect such as phosphorus. Thereby, a CMOS transistor having a lower threshold voltage Vth than that of the second embodiment can be realized. Although this embodiment can realize the lowest threshold voltage Vth in the first to third embodiments, since a silicide using a different metal material is used between both channel transistors, a manufacturing process and an apparatus are provided. With the disadvantage of increased complexity.

As described above, according to the third embodiment, it is possible to obtain a semiconductor device in which the work function of a metal electrode provided on an insulating film made of a material having a dielectric constant higher than that of SiO ₂ has a desired value. it can.

  In the first to third embodiments, the present planar type transistor has been described as an example. However, the characteristic of one embodiment of the present invention lies in the stacked structure of the gate electrode / gate insulating film of the CMOS transistor. The range is not limited by transistor shape. Therefore, the present invention can be applied not only to planar transistors but also to three-dimensional transistors such as FIN transistors.

(Fourth embodiment)
Next, a semiconductor device according to a fourth embodiment of the present invention is described. The semiconductor device of this embodiment is a stack type memory, and includes a plurality of memory cells. This memory cell will be described with reference to FIGS. 28 (a) and 28 (b).

  FIGS. 28A and 28B are cross-sectional views of the stacked memory cell according to the fourth embodiment, taken along a word line direction and a plane perpendicular to the word line direction, respectively. As shown in FIGS. 28A and 28B, the memory cell according to this embodiment includes a tunnel insulating film 36, a floating gate electrode 37, an interelectrode insulating film 38, and an impurity segregation layer 44 on a semiconductor substrate 34. The control gate electrode 40 has a gate stack structure (hereinafter also referred to as a memory gate stack) in which the control gate electrodes 40 are stacked in this order. The interelectrode insulating film 38 includes an insulating film 38a made of a High-k material provided on the floating gate electrode 37 and a control gate interface insulating layer 38b provided on the insulating film 38a. A sidewall oxide film 41 is provided on the side of the memory gate stack. The memory gate stack whose sides are covered with the sidewall oxide film 41 is covered with an interlayer insulating film 42. Source / drain diffusion layers 35 are provided on the semiconductor substrate 34 on both sides of the memory gate stack. The individual stacked memory cells are separated from each other by an element isolation insulating film 43.

  In the present embodiment, the control gate electrode 40 is formed of a metal silicide such as Ni. As a result, the stack type memory cell according to the present embodiment can avoid the problem of the control gate voltage drop in the depletion region of the Si gate that occurs in the conventional memory cell, so that the write / erase operation can be performed effectively. As a result, the threshold voltage Vth window becomes larger, and the function of the storage capacity is improved. Further, the interelectrode insulating film 38 is composed of a laminated film of an insulating film 38a made of a high-k material and a control gate interface insulating layer 38b made of a silicon oxynitride film or a silicon nitride film. Sometimes, when a large voltage is applied to the control gate electrode 40, it is possible to limit the leakage current while keeping strong capacitive coupling with the floating gate electrode 37. As a result, the write / erase operation can be performed more effectively than before.

  Here, the feature of the present embodiment is that a film made of SiON or SiN is provided at the interface between the insulating film 38a made of a High-k material and the control gate electrode 40 made of metal silicide. The decrease in the work function of the metal silicide described in the embodiment can be solved. In general, the higher the work function of the control gate electrode 40, the smaller the leakage current with the floating gate electrode 37. This is because the electron barrier between the control gate electrode 40 and the interelectrode insulating film 38 is increased, and electron transport in the interelectrode insulating film 38 is difficult to occur. According to this embodiment, a memory gate stack including the interelectrode insulating film including the insulating film 38a made of the high-k material and the control gate electrode 40 made of metal silicide having a high work function can be realized. 37 and the control gate electrode 40 can be improved at the same time and the leakage current can be suppressed. Note that Co, Ni, or the like can be used as a metal constituting the metal silicide of the control gate electrode 40.

  Furthermore, in this embodiment, the impurity segregation layer 44 made of boron, aluminum or the like is provided at the interface between the control gate electrode 40 and the control gate interface insulating layer 38b, thereby succeeding in further improving the work function of the metal silicide. . This further reduced the leakage current between the control gate electrode 40 and the floating gate electrode 37. As in the first embodiment, the bonding state of impurities contained in the impurity segregation layer 44 is mainly the total number of metallic bonding states that are unevenly distributed on the control gate electrode 40 side, and is mainly unevenly distributed on the control gate interface insulating layer 38b side. The feature of the present embodiment is that it is larger than the sum of the oxidized, nitrided, and oxynitrided states. Thereby, even on the control gate interface insulating layer 38b made of SiON or SiN, the work function of the control gate electrode 40 is increased by a group 13 element such as boron or aluminum.

  Next, a method for manufacturing a semiconductor device according to the fourth embodiment will be described.

  The manufacturing process of the semiconductor device according to the fourth embodiment is shown in FIGS.

First, a thermally oxidized SiO ₂ film having a thickness of 10 nm is formed as a tunnel insulating film 36 on a semiconductor substrate 34 to which a desired impurity is added, and a polycrystalline silicon layer having a thickness of 100 nm is formed as a floating gate electrode 37 by a CVD method or the like. accumulate. Polycrystalline silicon layer 37 is doped N-type with impurities such as phosphorus. Subsequently, a mask material 45 for element isolation processing was deposited on the entire surface of the floating gate electrode 37. Using an element isolation resist pattern (not shown), the mask material 45, the floating gate electrode 37, and the tunnel insulating film 36 are etched by RIE (Reactive Ion Etching) or the like, and the semiconductor substrate 34 is further processed. A 150 nm isolation trench 46 was formed. Thereby, the structure shown in FIGS. 29A and 29B is obtained.

  Subsequently, an element isolation insulating film 43 is deposited on the entire surface of the structure shown in FIGS. 29 (a) and 29 (b). Here, a silicon oxide film was deposited by a CVD method. Thereafter, the silicon oxide film on the surface was removed by a planarization technique such as CMP, and the surface was planarized to obtain the structure shown in FIGS. 30 (a) and 30 (b). At this time, the mask material 45 is exposed.

  Thereafter, the exposed mask material 45 was selectively removed, and the element isolation insulating film 43 was etched to expose the side surface of the floating gate electrode 37. In this embodiment, the silicon oxide film was etched with a dilute HF aqueous solution to form a protrusion of the floating gate electrode 37 having a height of 50 nm. The structure thus obtained is shown in FIGS. 31 (a) and 31 (b).

An insulating film 38a made of a High-k material and an insulating film 38b made of SiON or SiN were deposited as an interelectrode insulating film 38 on the entire surface of the structure shown in FIGS. As the material of the insulating film 38a, hafnium silicate, nitrogen-added hafnium silicate, hafnium aluminate (HfAlO), zirconium silicate, nitrogen-added zirconium silicate, alumina, hafnia, lanthanum aluminate, or the like can be used. In particular, a High-k film made of Hf is suitable as a High-k material of this embodiment because of its high heat resistance and high relative dielectric constant. In the present embodiment, the HfAlO film 38a having a thickness of 15 nm is deposited by the ALD method. At this time, an interface transition layer having a thickness of 1 nm or less may be formed between the floating gate electrode 37 and the high-k film 38a. On the other hand, the insulating layer 38b preferably has a film thickness converted to SiO ₂ of 0.1 nm or more and 2 nm or less, and in this embodiment, silicon having a SiO ₂ converted film thickness of 1 nm and a physical thickness of 1.5 nm. The oxynitride layer 38b was also deposited by the ALD method. Since the silicon concentration of the silicon oxynitride layer 38b is 3 atomic% or more and 15 atomic% or less, a film with few defects can be realized, which is preferable. In this way, the structure shown in FIGS. 32A and 32B was obtained.

  Subsequently, a polycrystalline silicon layer 47 for forming the control gate electrode was deposited on the entire surface of the structure shown in FIGS. 32A and 32B by the CVD method. The film thickness in this embodiment was 100 nm. Thereafter, an RIE mask material 48 is deposited on the entire surface of the polycrystalline silicon layer 47, and a gate resist pattern (not shown) is formed thereon, and then the mask material 48, the polycrystalline silicon 47, the silicon oxynitride layer 38b, High The insulating film 38a made of the -k material, the floating gate electrode 37, and the tunnel insulating film 36 were sequentially etched. In this step, the shapes of the floating gate electrode 37 and the control gate electrode 40 are determined. The cross section at this time is shown in FIGS. 33 (a) and 33 (b).

  Next, after ion implantation for forming the source / drain regions is performed on the structure of FIG. 33, the mask material 48 is removed, and a sidewall oxide film 41 is subsequently formed on the exposed surface of the memory gate stack by 2 nm, for example, by thermal oxidation. did. Subsequently, an activation heat treatment was performed to form the source / drain regions 35, and then the entire surface of the structure was covered with an interlayer insulating film 42 to obtain the structure shown in FIGS. 34 (a) and 34 (b).

  Next, a process for siliciding the control gate electrode is performed. First, a planarization process such as CMP is performed on the structure shown in FIGS. 34A and 34B, and the sidewall oxide film 41 formed on the interlayer insulating film 42 and the polycrystalline silicon layer 47 is removed. Then, the polycrystalline silicon layer 47 is exposed. This is shown in FIGS. 35 (a) and 35 (b).

  Further, a metal layer 49 made of Ni having a film thickness of 70 nm, for example, was deposited on the entire surface of the structure shown in FIGS. 35A and 35B by sputtering (FIGS. 36A and 36B). Here, as the metal layer 49, it is preferable to use either Ni or Co, and make a silicide having a composition in which the composition ratio of the metal to silicon (metal / silicon) is 0.5 to 2. These metal silicide phases have a work function of 4.55 eV to 4.75 eV. Thereby, a silicide electrode having a relatively high work function can be realized. The metal silicide phase can be adjusted by controlling the film thickness ratio between the metal and silicon. Also, by using any of Pt, Ir, Ru, Pd, and Re, or making a silicide having a composition in which the composition ratio of metal to silicon (metal / silicon) is 2 or more using Ni or Co, A work function of 4.75 eV to 5.10 eV can also be realized. In this case, although the leakage current is lower, there is a demerit that it is inferior in thermal stability as compared with a metal silicide having a work function of 4.55 eV to 4.75 eV.

  Thereafter, a heat treatment is performed at 450 ° C. for 1 minute, all of the polycrystalline silicon layer 47 is silicided, and unreacted Ni is selectively removed with sulfuric acid, whereby the structure shown in FIGS. 37 (a) and 37 (b) is obtained. Realized. Since the polycrystalline silicon layer 47 is not doped with impurities, an impurity segregation layer is not formed at the interface between the metal silicide 50 and the insulating layer 38b at this point.

A group 13 element is ion-implanted into the metal silicide 50 having the structure shown in FIGS. 37 (a) and 37 (b). Here, boron was implanted under the conditions of an acceleration energy of 5 keV and a dose of 5 × 10 ¹⁵ / cm ² . Next, at least a part of the boron implanted into the metal silicide 50 serving as the control gate electrode 40 is thermally diffused and segregated at the interface between the metal silicide 50 and the control gate interface insulating layer 38b. In this embodiment, the interface impurity segregation layer 44 is formed at the interface between the control gate electrode 40 and the control gate interface insulating layer 38b by performing heat treatment at 450 ° C. for 1 minute. The impurity segregation layer 44 thus formed has a total number of metallic bonding states of impurities mainly distributed on the control gate electrode 40 side, and the oxidation / nitridation of impurities mainly distributed on the control gate interface insulating layer 38b side. -More than the sum of the oxynitriding states, and works to increase the work function. Thereafter, an interlayer insulating film 42 is formed again, thereby completing the structure shown in FIGS. 28 (a) and 28 (b).

  In this embodiment, the effective work function is 4.85 eV due to the modulation effect of boron with respect to NiSi having a work function of 4.65 eV. As a result, the inter-electrode leakage current during writing / erasing can be remarkably reduced, and the Vth window of the stacked memory cell can be increased.

  Further, in the present embodiment, since a high-k material is used for the interelectrode insulating film, it has the effect of reducing the capacitive coupling strength and leakage current between the electrodes, and further siliciding the control gate electrode 40. Thus, the adverse effect of the voltage drop due to the depletion layer that has become apparent in the case of the conventional polygate can be avoided.

As described above, according to the fourth embodiment, it is possible to obtain a semiconductor device in which the work function of a metal electrode provided on an insulating film made of a material having a dielectric constant higher than that of SiO ₂ has a desired value. it can.

  The composition ratio and work function of the Ni silicide used in the first to fourth embodiments are shown in FIG.

  As mentioned above, although embodiment of this invention was described, this invention is not limited to these, In the category of the summary of the invention as described in a claim, it can change variously. In addition, the present invention can be variously modified without departing from the scope of the invention in the implementation stage. Furthermore, various inventions can be formed by appropriately combining a plurality of constituent elements disclosed in the embodiment.

The figure which shows the experimental result explaining the work function modulation effect | action by the impurity segregation to a silicide / insulating film interface. The schematic diagram which shows the distribution of an impurity in the silicide / insulating film interface, and an atomic bonding state for demonstrating the experimental result of FIG. The schematic diagram which shows the distribution of the impurity in a silicide / insulating film interface, and an atomic bonding state for demonstrating the difference in the work function modulation effect | action by impurity segregation in one Embodiment of this invention. The figure which shows the experimental result at the time of confirming the work function modulation effect by silicide / insulation film interface impurity segregation in this invention. The figure which shows the XPS experimental result which analyzed the difference in the bonding state of the interface segregation impurity by the pre-adding method and the post-adding method. The figure which shows the experimental result at the time of confirming the difference in the work function modulation effect | action by the difference in the introduction method of a boron impurity in one Embodiment of this invention. The figure which shows the XPS experiment result which analyzed the difference in the bonding state of an interface impurity performed in order to confirm the origin of the difference in the work function modulation | alteration effect | action by the difference in the boron impurity introduction | transduction method in one Embodiment of this invention. The schematic diagram which shows the distribution and impurity bonding state of the impurity in a silicide / insulating film interface for demonstrating the difference in the work function modulation effect | action by the difference in the introduction method of a boron impurity in one Embodiment of this invention. Sectional drawing of the gate length direction of CMOSFET by 1st Embodiment of this invention. Sectional drawing which shows the manufacturing process of CMOSFET by 1st Embodiment. Sectional drawing which shows the manufacturing process of CMOSFET by 1st Embodiment. Sectional drawing which shows the manufacturing process of CMOSFET by 1st Embodiment. Sectional drawing which shows the manufacturing process of CMOSFET by 1st Embodiment. Sectional drawing which shows the manufacturing process of CMOSFET by 1st Embodiment. Sectional drawing which shows the manufacturing process of CMOSFET by 1st Embodiment. Sectional drawing which shows the manufacturing process of CMOSFET by 1st Embodiment. Sectional drawing which shows the manufacturing process of CMOSFET by 1st Embodiment. Sectional drawing which shows the manufacturing process of CMOSFET by the modification of 1st Embodiment. Sectional drawing which shows the manufacturing process of CMOSFET by the modification of 1st Embodiment. Sectional drawing which shows the manufacturing process of CMOSFET by 2nd Embodiment. Sectional drawing which shows the manufacturing process of CMOSFET by 2nd Embodiment. Sectional drawing which shows the manufacturing process of CMOSFET by 2nd Embodiment. Sectional drawing which shows the manufacturing process of CMOSFET by 2nd Embodiment. Sectional drawing which shows the manufacturing process of CMOSFET by 2nd Embodiment. Sectional drawing which shows the manufacturing process of CMOSFET by 3rd Embodiment. Sectional drawing which shows the manufacturing process of CMOSFET by 3rd Embodiment. Sectional drawing which shows the manufacturing process of CMOSFET by 3rd Embodiment. Sectional drawing of the stack | stuck type memory cell which concerns on 4th Embodiment. Sectional drawing which shows the manufacturing process of the stack type memory cell which concerns on 4th Embodiment. Sectional drawing which shows the manufacturing process of the stack type memory cell which concerns on 4th Embodiment. Sectional drawing which shows the manufacturing process of the stack type memory cell which concerns on 4th Embodiment. Sectional drawing which shows the manufacturing process of the stack type memory cell which concerns on 4th Embodiment. Sectional drawing which shows the manufacturing process of the stack type memory cell which concerns on 4th Embodiment. Sectional drawing which shows the manufacturing process of the stack type memory cell which concerns on 4th Embodiment. Sectional drawing which shows the manufacturing process of the stack type memory cell which concerns on 4th Embodiment. Sectional drawing which shows the manufacturing process of the stack type memory cell which concerns on 4th Embodiment. Sectional drawing which shows the manufacturing process of the stack type memory cell which concerns on 4th Embodiment. The figure which shows the composition ratio and work function of Ni silicide which are used in each embodiment.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Semiconductor substrate 2 N-type well region 3 P-type well region 4 Element isolation layer 5 P-type diffusion layer 6 P-type extension layer 7 N-type diffusion layer 8 N-type extension layer 9 Gate insulating film 10 Buffer layer 11 Impurity segregation layer 12 Gate Electrode layer 13 Gate insulating film 14 Buffer layer 15 Impurity segregation layer 16 Gate electrode layer 17 Side wall 18 Interlayer insulating layer 19 P channel MIS transistor 20 N channel MIS transistor 21 Gate insulating film 22 Buffer film 23 Silicon layer 24 Metal layer 25 Metal silicide 26 resist mask layer 27 resist mask layer 28 Ni layer (thin film)
29 Ni layer (thick film)
30 Ti layer 31 Pt layer 32 Er layer 33 W layer 34 Semiconductor substrate 35 Source / drain diffusion layer 36 Tunnel insulating film 37 Floating gate electrode 38 Interelectrode insulating film 38a Insulating film 38b made of High-k material Control gate interface insulating layer 40 Control gate electrode 41 Side wall oxide film 42 Interlayer insulating film 43 Element isolation insulating layer 44 Impurity segregation layer 45 Mask material 46 Element isolation trench 47 Polycrystalline silicon layer 48 Mask material 49 Metal layer 50 Metal silicide

Claims (7)

A semiconductor substrate;
A tunnel insulating layer formed on the semiconductor substrate;
A floating gate electrode provided on the tunnel insulating layer;
A first insulating layer formed on the floating gate electrode and made of a high dielectric constant material; and a second insulating layer formed on the first insulating layer and containing silicon and oxygen and nitrogen or containing silicon and nitrogen. An interelectrode insulating film;
A control gate electrode formed on the interelectrode insulating film;
An interface layer including a group 13 element formed at an interface between the second insulating layer and the control gate electrode;
Source / drain regions formed in the semiconductor substrate on both sides of the control gate electrode;
And the number of bonding states of the group 13 element in the interface layer is greater than the total number of oxidation, nitridation, or oxynitride bonding states.   Of the group 13 elements in the interface layer, elements in an oxidized, nitrided, or oxynitride bonded state are mainly distributed on the second insulating layer side, and elements in a metal bonded state are mainly present on the control gate electrode side. The semiconductor device according to claim 1.   3. The semiconductor device according to claim 1, wherein the group 13 element in the interface layer is boron or aluminum.   4. The semiconductor according to claim 1, wherein the control gate electrode is made of a metal silicide containing at least one metal among Ni, Co, Pt, Ir, Ru, Pd, and Re. apparatus.   4. The semiconductor device according to claim 1, wherein the control gate electrode is made of a metal silicide containing at least one of Ni and Co.   The semiconductor device according to claim 1, wherein the first insulating layer contains Hf at least in part. Forming a tunnel insulating layer on the semiconductor substrate;
Forming a floating gate electrode layer on the tunnel insulating layer;
Forming a first insulating layer made of a high dielectric constant material on the floating gate electrode layer;
Forming a second insulating layer containing at least silicon and nitrogen on the first insulating layer;
Forming a silicon layer on the second insulating layer;
Forming a metal layer on the silicon layer;
Chemically reacting the silicon layer and the metal layer by heat treatment to convert the silicon layer into a metal silicide layer;
Introducing a group 13 element into the metal silicide layer;
The step of diffusing the group 13 element by heat treatment, and at least partially disperse it in a region including an interface between the second insulating layer and the metal silicide layer;
A method for manufacturing a semiconductor device, comprising:

Images