JP2007332397A - Electroconductive thin film and manufacturing method therefor - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To form an electroconductive thin film which contains a usable metal for an electrode, and has little unevenness on the surface. <P>SOLUTION: The electroconductive thin film 103 is formed on a substrate 101, for instance, made from monocrystalline silicon, through a lower insulation film 102, for instance, made of a silicon oxide film in an amorphous state. The electroconductive thin film 103 includes ruthenium and nitrogen. The electroconductive thin film 103 may be formed by sputtering a target made from ruthenium with the use of ECR plasma including argon (Ar) gas, xenon (Xe) gas and nitrogen gas. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a method for forming a conductive thin film used for an electrode of a nonvolatile memory element.

  Research and development for the expansion of the multimedia information society and the realization of ubiquitous services are being actively conducted. In particular, a device for recording information (hereinafter referred to as a memory) mounted on a network device or an information terminal is an important key device. The functions required for the memory installed in ubiquitous terminals are high-speed operation, long-term retention period, environmental resistance, low power consumption, and the function that does not erase the stored information even when the power is turned off, that is, non-volatility is essential It is said that.

  Conventionally, many semiconductor devices have been used for memories. As one of them, DRAM (Dynamic Random Access Memory) is widely used. A unit storage element (hereinafter referred to as a memory cell) of a DRAM is composed of one storage capacitor and one MOSFET (Metal-oxide-semiconductor field effect transistor), and the charge stored in the storage capacitor of the selected memory cell. The voltage corresponding to the state is taken out as “on” or “off” of the electrical digital signal from the bit line, thereby reading the stored data (see Non-Patent Document 1).

  However, in the DRAM, when the power is turned off, it is impossible to maintain the state of the storage capacity, and the stored information is erased. In other words, DRAM is a volatile memory element. As is well known, a DRAM requires a refresh operation for rewriting data, and has a drawback that the operation speed is reduced.

  A ROM (Read Only Memory) is well known as a non-volatile memory in which data does not volatilize even when the power is turned off, but the recorded data cannot be erased or changed. As a rewritable nonvolatile memory, a flash memory using an EEPROM (Electrically Erasable Programmable Read Only Memory) has been developed (see Patent Document 1, Non-Patent Documents 1 and 2). Flash memory is used in many fields as a practical non-volatile memory.

  In a memory cell of a typical flash memory, a gate electrode portion of a MOSFET has a stack gate structure including a plurality of layers each having a control gate electrode and a floating gate electrode. In the flash memory, data can be recorded by utilizing the fact that the threshold value of the MOSFET changes depending on the amount of charge accumulated in the floating gate.

  Writing data in the flash memory is performed by hot carriers generated by applying a high voltage to the drain region overcoming the energy barrier of the gate insulating film. In addition, data is written by injecting electric charges (generally electrons) from the semiconductor substrate to the floating gate by applying a high electric field to the gate insulating film to flow an FN (Fowler-Nordheim) tunnel current. Is done. Data is erased by extracting charges from the floating gate by applying a high electric field in the opposite direction to the gate insulating film.

  The flash memory does not require a refresh operation as in a DRAM, but has a problem that the time required for data writing and erasing becomes orders of magnitude longer than that of a DRAM because the FN tunnel phenomenon is used. Further, when data writing / erasing is repeated, the gate insulating film deteriorates, so that the number of rewrites is limited to some extent.

  In contrast to the flash memory described above, as a new non-volatile memory, a ferroelectric memory using ferroelectric polarization (hereinafter referred to as FeRAM (Ferroelectric RAM)) or a ferromagnetic memory using ferromagnetic resistance ( MRAM (Magnetoresist RAM) has been attracting attention and has been actively researched.FeRAM has already been put into practical use, and if it can solve various problems, it will be a portable memory. It is expected that not only logic DRAMs can be replaced.

  Thus, FeRAM expected as a substitute for flash memory is mainly classified into a stack type and an FET type. The stack type is also called a one-transistor one-capacitor FeRAM, and there are a stack type capacitor, a planar type capacitor, and a three-dimensional type capacitor. The stack type includes a one-transistor one-capacitor FeRAM and a two-transistor two-capacitor FeRAM in which two are stacked for stable operation. Practically, stack type FeRAM is the mainstream.

  For example, as shown in FIG. 13, the stack type FeRAM includes a MOS transistor including a gate electrode 1305 provided on a semiconductor substrate 1301 via a source 1302, a drain 1303, and a gate insulating film 1304. A capacitor made of a lower electrode 1311, a dielectric layer 1312 made of a ferroelectric material, and an upper electrode 1313 is connected to 1302. In the example of FeRAM shown in FIG. 13, the capacitor is connected to the source 1302 by the source electrode 1306. A drain electrode 1307 is connected to the drain 1303, and an ammeter is connected to the drain 1303. This FeRAM has a function of extracting “on” or “off” data by detecting the polarization direction of the dielectric layer 1312 made of a ferroelectric material as a current flowing between the source and the drain (channel 1321). Yes.

  What is important in the structure of this FeRAM is the capacitor structure part of the lower electrode-dielectric-upper electrode, and in particular, in terms of memory stability that maintains the insulation of the ferroelectric and retains data for a long time, The interface state of ferroelectrics is very important. For example, if the surface unevenness of the lower electrode is large, it affects the crystallinity of the ferroelectric, and the surface unevenness of the lower electrode causes a leak, lowering the electric withstand voltage of the device and preventing long-term memory retention. It becomes possible. For this reason, in the capacitor structure, the material and the formation method of the lower electrode are very important elemental technologies.

On the other hand, in the field of Si semiconductors and display elements, there is an increasing demand for dielectrics having a dielectric constant higher than that of SiO ₂ (hereinafter referred to as “High-k materials”). In the conventional Si-LSI, SiO ₂ which is an oxide of Si has been used as a dielectric material from the viewpoint of compatibility with the Si substrate and suppression of process contamination. However, by thinning the SiO ₂ film thickness, it tends to have been high performance devices are, SiO ₂ film thickness has become difficult with 1.5nm or less. This is because, in principle, SiO ₂ having a thickness of 1.5 nm or less cannot achieve both of suppressing leakage current and obtaining a desired electric capacity.

Thus the high-k material is used instead of SiO _2, SiO ₂ equivalent thickness (Equivalent oxide thickness: EOT) to be to below 1.5 nm, on which the effective thickness that can suppress the leakage current Thus, studies are underway to obtain a sufficient capacitance for device operation (see Non-Patent Document 3).

However, the forbidden band width (Band gap, E _g ) of a high-k material capable of obtaining a high dielectric constant is in principle smaller than the forbidden band width of SiO _{2, and} a large amount of leakage current tends to flow even with the same film thickness. It is in. Further, in the field of solid capacitors, the leakage current of the high-k material formed on the lower electrode is larger than an ideal value, and the reduction of the leakage current is an issue to be studied from the viewpoint of reliability. In order to reduce the leakage current, it is important to optimize the film formation method of the high-k material, but flattening the surface unevenness of the lower electrode is also a very important issue. Actually, it is also known that when the surface irregularity of the lower electrode is large, the leakage current is larger. For this reason, the material of the lower electrode and the formation method thereof are very important elemental technologies.

  Furthermore, also in the MRAM described above, the formation of a metal thin film serving as an electrode is an important elemental technology. The MRAM is a nonvolatile memory using a ferromagnetic tunnel resistance (TMR) effect at a tunnel junction, and includes a nonvolatile memory, and can perform high-speed reading / writing. MRAM has a high possibility that it can be rewritten indefinitely.

In general MRAM, a lower magnetic layer (also called a free layer), a tunnel insulating layer, and an upper magnetic layer (pinned layer, pinning layer, cover layer) are formed on a substrate insulated by a thermal oxide film or the like. It has a structure. However, a general MRAM structure has a problem that a magnetic field for switching becomes large. For this reason, the switching electric field is reduced by using a multilayer exchange coupling structure (Co ₉₀ Fe ₁₀ / Ru / Co ₉₀ Fe ₁₀ ) in which ruthenium (Ru) is laminated as an antiparallel coupling film on the free layer of the lower magnetic layer. Consideration is ongoing. The Co ₉₀ Fe ₁₀ film forming this multilayer exchange coupling structure has a thickness of 3 nm, and the Ru film has a thickness of 1 nm, which is an extremely thin film (see Non-Patent Document 4).

  In the multilayer exchange coupling structure described above, the metal thin film is a very thin film of 1 nm, and therefore it is important to sufficiently reduce the surface unevenness, and this metal thin film forming method is a very important elemental technology.

  As described above, in the formation (film formation) of a metal thin film, various forming apparatuses and various thin film forming methods have been tried so far. For example, sputtering methods such as vacuum deposition (Evaporation), electron beam deposition (EB deposition), PLD (Pulse laser deposition), DC sputtering, RF sputtering, magnetron sputtering, etc. In addition, ECR sputtering (Electron cyclotron resonance sputtering) can be used.

  Among the film forming methods described above, the EB vapor deposition method is suitable for forming a high melting point material. The EB vapor deposition method is a method of forming a raw material film on a substrate by melting the raw material with an electron beam, and is characterized in that it can be easily formed even if the raw material is a high melting point material. However, there is a problem that it is difficult to form a thin film with small surface irregularities on the substrate depending on the crystallinity of the formed film and the raw material state.

  Further, a PLD method that can form a metal thin film with a good film quality by sputtering a target of a raw material with a powerful laser light source such as an excimer laser has attracted attention. However, in this method, the area of the portion irradiated with the laser in the target surface is very small, and a large distribution occurs in the raw material sputtered and supplied from the laser irradiation portion. For this reason, a large in-plane distribution is generated in the film thickness and film quality of the metal thin film formed on the substrate, and there is a big problem in reproducibility such that even if formed under the same conditions, completely different characteristics are obtained.

  In contrast to the various film forming methods described above, a sputtering method using plasma (also simply referred to as a sputtering method) has attracted attention as a method for forming a ferroelectric film. The sputtering method is a promising film forming apparatus and method because the surface irregularity (surface morphology) of the film to be deposited (film formation) is relatively good without using a high-risk gas or toxic gas as a raw material. It has become one of In sputtering, a reactive sputtering apparatus / method is promising as an excellent apparatus / method for obtaining a film having a stoichiometric composition.

  Among the sputtering methods, the formation of the above-described metal thin film is also studied in the DC sputtering method and the RF sputtering method (conventional sputtering method) that have been used conventionally. In a conventional sputtering method, when a metal thin film is deposited, a target metal target is used, and an inert gas argon is used as a sputtering gas. However, in the conventional sputter method, the input high frequency bias for causing the sputtering phenomenon is generally as large as about 500 W, and there is a phenomenon that the substrate to be deposited and the deposited thin film are sputtered. It was difficult to damage the thin film and obtain sufficiently small film characteristics and surface irregularities.

  On the other hand, as a method capable of forming a good film quality, plasma is generated by electron cyclotron resonance (ECR), and a substrate is irradiated with a plasma flow created by using a divergent magnetic field of the generated plasma. At the same time, there is an ECR sputtering method in which a high frequency or negative DC voltage is applied between the target and the ground, ions in the plasma flow generated by the ECR are drawn into the target, collide, and sputtered to deposit a film on the substrate. is there.

  In the conventional sputtering method, a stable plasma cannot be obtained unless the gas pressure is about 0.1 Pa or more. However, the ECR sputtering method has a characteristic that a stable ECR plasma can be obtained at a pressure of the order of 0.01 Pa. In addition, since the ECR sputtering method performs sputtering by causing particles generated by ECR to collide with a target with a high frequency or a negative DC high voltage, sputtering can be performed at a low pressure.

  Here, the ions in the ECR plasma flow have an energy of 10 eV to several tens of eV due to the divergent magnetic field. In addition, since the plasma is generated and transported at such a low pressure that the gas behaves as a molecular flow, the ion current density of ions reaching the substrate can be increased. Therefore, the ions in the ECR plasma flow give energy to the raw material particles sputtered and flying on the substrate, and promote the binding reaction between the raw material particles and oxygen when forming an oxide film. The film quality of the deposited film is improved.

  As described above, the ECR sputtering method is characterized in that a high-quality film can be formed at a low substrate temperature. See, for example, Patent Document 2, Patent Document 3, and Non-Patent Document 5 for how high-quality thin films can be deposited by ECR sputtering.

  Furthermore, since the deposition rate of the film is relatively stable, the ECR sputtering method is suitable for forming an extremely thin film such as a gate insulating film with good film thickness control. Further, the surface morphology of the film deposited by the ECR sputtering method is flat on the atomic scale order. Therefore, it can be said that the ECR sputtering method is a promising method for forming an electrode film made of metal.

  However, in order to maintain technical progress, it has become difficult to obtain sufficient surface irregularity even when using ECR sputtering. Specifically, in contrast to the situation surrounding the memory as described above, the memory is not realized by the effect of changing the state of the semiconductor (forming a channel) according to the amount of polarization of the ferroelectric material. As shown, a technique has been proposed in which the resistance value of the ferroelectric layer 1403 formed directly on the semiconductor 1401 is changed, and as a result, a memory function is realized. The resistance value of the ferroelectric layer 1403 of the element shown in FIG. 14 is controlled by applying a voltage between the upper electrode 1404 and the lower electrode 1402.

JP-A-8-031960 Japanese Patent No. 2814416 Japanese Patent No. 2779997 Japanese Patent Laid-Open No. 10-152397 JP-A-10-152398 Simon G., Physics of Semiconductor Devices, 1981 (S.M.Sze, "Physics of Semiconductor Devices", John Wiley and Sons, Inc.) Fujio Tsujioka, Applied Physics, Vol. 73, No. 9, p. 1166, 2004. Wilk et al., Journal of Applied Physics, 87, 484, 2000 (Wilk et al., J. Appl. Phys., 87, 484 (2000).) Ryosuke Saito, “Current Status and Future Prospects of MRAM”, FED Review No. 1, No. 25, 2001. Amazawa et al., Journal of Vacuum Science and Technology, Vol. B17, No. 5, pp. 2222, 1999 (J. Vac. Sci. Technol., B17, no. 5, 2222 (1999).

  In the structure shown in FIG. 14, a ferroelectric layer 1403 that causes a resistance switch phenomenon is formed on a metal thin film that is a lower electrode 1402, and a metal thin film is further formed as an upper electrode 1404. In the element structure shown in FIG. 14, by applying a voltage to the upper electrode 1404 and the lower electrode 1402, the resistance value of the ferroelectric layer 1403 is changed from a high resistance state to a low resistance state, and from a low resistance state to a high resistance state. The phenomenon of switching to appears reversibly and stably. This phenomenon is also called giant electro-resonance (CER) effect, and RRAM (Resistance RAM, also called ReRAM), which applies this phenomenon, has attracted attention as a new nonvolatile memory that replaces flash memory. Yes. The resistance switch phenomenon has been observed not only in the above ferroelectric layer but also in the ferromagnetic materials PrCaMnO, TiO, NiO, and the like. Hereinafter, a film that causes a resistance switch phenomenon is referred to as a CER film.

FIG. 15 shows an example of voltage-current characteristics of the resistance switch phenomenon. A multilayer metal layer made of Pt and Ti was formed as a lower electrode on a Si substrate on which SiO ₂ having a thickness of 100 nm was formed, a bismuth titanate layer was formed as a CER film, and a gold electrode was formed as an upper electrode It is a characteristic of things. The Pt, Ti, and bismuth titanate layers were each formed using ECR sputtering. Gold was formed using a vacuum deposition method. The horizontal axis in FIG. 15 is the voltage value applied between the upper electrode and the lower electrode, and the vertical axis is the absolute value of the current value observed when the voltage is applied, and is shown in logarithmic display. However, the voltage values and current values described here are those observed with actual elements as an example. Therefore, this phenomenon is not limited to the following numerical values. Other numerical values may be observed depending on the material and thickness of the film actually used for the element and other conditions.

First, when a positive voltage is applied to the upper electrode, as indicated by (1) in FIG. 15, the current flowing as 10 ⁻⁸ to 10 ⁻⁵ A is very small at 0 to +2.0 V. However, as shown in (2), when + 2.0V is exceeded, a positive current suddenly flows. Actually, a current of 1 × 10 ⁻³ A or more flows but is not observed because this current is not applied to protect the measuring instrument. Subsequently, when a positive voltage is applied to the upper electrode, as shown in (3), the resistance in which a positive current of 1 × 10 ⁻³ A or more flows at about 0.1 V is low. Subsequently, even when a negative voltage of up to −1.5 V is applied to the upper electrode, as shown by (4), the resistance through which a current of 10 ⁻³ to 10 ⁻² A flows is low.

  Thereafter, when a negative voltage of −1.5 V or higher is applied, the value of the flowing current decreases rapidly as shown in (5), and the resistance value increases as shown in (6). Thereafter, even when a negative voltage is applied, the high resistance state (6) is maintained. In this state, when a positive voltage is applied to the upper electrode, as shown in (1), it is in a state of high resistance up to about 0 to 2.0 V, but a voltage of +2.0 V or more is applied and (2) Thus, when the current flows, the low resistance state (3) is obtained. As described above, in the element shown in FIG. 14, a phenomenon of reversibly switching between the “high resistance state” and the “low resistance state” is observed.

However, as shown in FIG. 15, the current value in the low resistance state (hereinafter referred to as On current) is about 10 ⁻² A, which is about the same level as PRAM, which has a lot of On current and is problematic on the circuit. It is. The current value necessary for obtaining a stable element in circuit design is preferably 10 ⁻⁶ A or less achieved in the flash memory, but it is desirable to reduce it to 10 ⁻⁴ A or less, which is about MRAM. If this On current is not reduced, it will be a major barrier to the practical application of RRAM, which is an important issue to be studied.

  Various studies have been made on the reduction of the above On current. A technique that seems to be particularly important is a technique for forming a CER film. However, in addition to this, it has been found that the surface irregularity of the metal thin film serving as the lower electrode serving as the base affects the characteristics of the element that causes the resistance switching phenomenon. It has been found to be very important for film formation.

  The surface roughness of the Pt / Ti lower electrode shown in the example of FIG. 15 has been found to be about 2 nm by observation with an atomic force electron microscope (AFM), a cross-sectional electron microscope (TEM), or the like. It was found that when a CER film is formed on a metal thin film having a slight surface roughness, a large amount of On current flows. As described above, conventionally, it is not easy to reduce the unevenness of the surface of the electrode film constituting the capacitor, and there is a problem that it is difficult to obtain a memory element capable of obtaining stable operation.

  The present invention has been made to solve the above-described problems, and is intended to enable a conductive thin film containing a metal that can be used as an electrode to be formed with less surface unevenness. Objective.

  The conductive thin film according to the present invention is a conductive thin film formed on a substrate, and the conductive thin film is composed of ruthenium and nitrogen, and the A-axis oriented structure is larger than the C-axis oriented structure. It is what. The presence of nitrogen changes the in-plane orientation and reduces the grain size of ruthenium.

  Also, the method for producing a conductive thin film according to the present invention generates a plasma composed of an inert gas and a nitrogen gas supplied at a predetermined composition ratio, and applies a negative bias to a target composed of ruthenium. A step of forming a conductive thin film composed of ruthenium and nitrogen on the substrate by causing the generated particles to collide with the target to cause a sputtering phenomenon and depositing ruthenium constituting the target on the substrate is provided. It is a thing. The plasma may be an electron cyclotron resonance plasma generated by electron cyclotron resonance and given kinetic energy by a divergent magnetic field. Further, the substrate may be heated to a predetermined temperature.

  As described above, according to the present invention, the conductive film is made of ruthenium and nitrogen, and the A-axis oriented structure is larger than the C-axis oriented structure. The conductive thin film can be formed with less surface irregularities.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. FIG. 1 is a configuration diagram showing a configuration of the conductive thin film 103 according to the embodiment of the present invention, and schematically shows a cross section. A conductive thin film 103 shown in FIG. 1 is formed on a substrate 101 made of, for example, single crystal silicon via a lower insulating film 102 made of, for example, an amorphous silicon oxide film. The conductive thin film 103 is composed of ruthenium and nitrogen. Hereinafter, a conductive material composed of ruthenium and nitrogen is referred to as ruthenium nitride for convenience. Further, a state in which nitrogen is introduced into ruthenium is expressed as nitriding. However, these states do not necessarily indicate that nitrogen and ruthenium form a compound.

The lower insulating film 102 is not limited to silicon oxide (silicon dioxide), but is made of a light metal such as LiNbO ₃ such as silicon oxynitride film, alumina, or lithium, beryllium, magnesium, or calcium. There may be. The lower insulating film 102 is made of a fluoride such as LiCaAlF ₆ , LiSrAlF ₆ , LiYF ₄ , LiLuF ₄ , KMgF ₃ , or a transition metal containing scandium, titanium, strontium, yttrium, zirconium, hafnium, tantalum, and a lanthanum series. These oxides and nitrides may be combined. The lower insulating film 102 includes a silicate (a ternary compound of metal, silicon, and oxygen) containing the above elements, an aluminate (a ternary compound of metal, aluminum, and oxygen) containing these elements, Oxides and nitrides containing two or more elements may be used. Furthermore, it may be a multilayered insulating layer made of these insulating films.

  A specific manufacturing method for the conductive thin film 103 described above will be described later, but with an ECR sputtering apparatus as shown in FIG. 2, a target made of ruthenium is converted into argon gas (Ar), xenon (Xe) gas, nitrogen. Sputtering may be performed using ECR plasma made of gas.

  Here, the ECR sputtering apparatus will be described with reference to the schematic cross-sectional view of FIG. The ECR sputtering apparatus shown in FIG. 2 includes a processing chamber 201 and a plasma generation chamber 202 communicating with the processing chamber 201. The processing chamber 201 communicates with an evacuation device (not shown), and the inside of the processing chamber 201 is evacuated together with the plasma generation chamber 202 by the evacuation device.

  The processing chamber 201 is provided with a substrate holder 204 to which the film formation target substrate 101 is fixed. The substrate holder 204 is inclined at a desired angle by a rotation mechanism (not shown) and is rotatable. By tilting and rotating the substrate holder 204, it is possible to improve the in-plane uniformity of the film and the step coverage with the material to be deposited. Further, a ring-shaped target 205 is provided so as to surround the opening region in the opening region into which the plasma from the plasma generation chamber 202 in the processing chamber 201 is introduced.

  The target 205 is placed in a container 205 a made of an insulator, and the inner surface is exposed in the processing chamber 201. Further, a high frequency power source 222 is connected to the target 205 via a matching unit 221 so that, for example, a high frequency of 13.56 MHz can be applied. When the target 205 is a conductive material, direct current may be applied. Note that the target 205 may be not only a circular shape but also a polygonal state as viewed from above.

  The plasma generation chamber 202 communicates with the vacuum waveguide 206, and the vacuum waveguide 206 is connected to the waveguide 208 through a quartz window 207. The waveguide 208 communicates with a microwave generation unit (not shown). In addition, a magnetic coil (magnetic field forming means) 210 is provided around the plasma generation chamber 202 and at the top of the plasma generation chamber 202. These microwave generator, waveguide 208, quartz window 207, and vacuum waveguide 206 constitute a microwave supply means. There is a configuration in which a mode converter is provided in the middle of the waveguide 208.

The operation example of the ECR sputtering apparatus of FIG. 2 will be described. First, after the inside of the processing chamber 201 and the plasma generation chamber 202 is evacuated, Ar gas or Xe gas, which is an inert gas, is introduced from the inert gas introduction unit 211. In addition, a reactive gas (nitrogen gas) is introduced from the reactive gas introduction unit 212 to bring the inside of the plasma generation chamber 202 to a pressure of about 10 ⁻⁵ to 10 ⁻⁴ Pa, for example. In this state, a magnetic field of 0.0875 T (Tesla) is generated in the plasma generation chamber 202 from the magnetic coil 210, and then the plasma generation chamber 202 is passed through the waveguide 208, the quartz window 207, and the vacuum waveguide 206. A 2.45 GHz microwave is introduced into the inside, and an electron cyclotron resonance (ECR) plasma is generated. Note that 1T = 10000 Gauss.

  The ECR plasma forms a plasma flow in the direction of the substrate holder 204 by the divergent magnetic field from the magnetic coil 210. Among the generated ECR plasma, electrons penetrate through the target 205 by the divergent magnetic field formed by the magnetic coil 210 and are extracted toward the substrate 101 and are irradiated onto the surface of the substrate 101. At the same time, positive ions in the ECR plasma are drawn out to the substrate 101 side so as to neutralize the negative charge due to electrons, that is, weaken the electric field, and are irradiated onto the surface of the layer being formed. Thus, while each particle is irradiated, some of the positive ions are combined with electrons to become neutral particles.

  In the thin film forming apparatus of FIG. 2, microwave power supplied from a microwave generation unit (not shown) is once branched in the waveguide 208, and plasma is supplied to the vacuum waveguide 206 above the plasma generation chamber 202. The generation chamber 202 is coupled from the side through a quartz window 207. By doing so, it becomes possible to prevent the scattered particles from adhering to the quartz window 207 from the target 205, and the running time can be greatly improved. In addition, a shutter or the like may be provided between the substrate to be processed and the target 205 to control the arrival of the raw material with respect to the substrate.

  Next, an example of a method for manufacturing the conductive thin film 103 shown in FIG. 1 will be described with reference to FIG. First, as shown in FIG. 3A, a substrate 101 made of p-type single crystal silicon having a main surface of plane orientation (100) and a resistivity of 1 to 2 Ω-cm is prepared. And a mixture of pure water and hydrogen peroxide, as well as a mixture of pure water and dilute hydrogen fluoride, followed by drying.

  Next, the lower insulating film 102 is formed on the cleaned and dried substrate. In forming the lower insulating film 102, the ECR sputtering apparatus shown in FIG. 2 is used, the substrate is fixed to the substrate holder 204 in the processing chamber 201, pure silicon (Si) is used as the target 205, and argon (Ar) is used as the plasma gas. ) And oxygen gas are used to form a metal mode film of Si—O molecules on the substrate 101 so as to cover the surface.

In the ECR sputtering method shown in FIG. 2, first, the inside of the plasma generation chamber 202 is evacuated to a high vacuum state of the order of 10 ⁻⁴ to 10 ⁻⁵ Pa, and then an inert gas introduction unit is introduced into the plasma generation chamber 202. From 211, for example, Ar gas, which is a rare gas, is introduced at a flow rate of about 20 sccm, for example, oxygen gas, which is a reactive gas, is introduced, for example, at about 5 sccm, and the pressure in the plasma generation chamber 202 is set to, for example, 10 ⁻² to 10 ⁻³ Pa Set. Note that sccm is a unit of flow rate and indicates that 1 cm ³ of fluid at 1 atm flows at 0 ° C. per minute. Further, in the plasma generation chamber 202, a magnetic field of electron cyclotron resonance condition is applied by supplying a coil current to the magnetic coil 210 at, for example, 28A. The magnetic flux density in the plasma generation chamber 202 is set to a magnetic field state of about 87.5 mT.

  In addition, a 2.45 GHz microwave (for example, 500 W) is supplied from a microwave generation unit (not shown), and this is supplied to the plasma generation chamber via the waveguide 208, the quartz window 207, and the vacuum waveguide 206. It introduce | transduces in 202, It is set as the state which the plasma of argon was produced | generated in the plasma production chamber 202 by introduction of this microwave.

  The plasma generated as described above is emitted from the plasma generation chamber 202 to the processing chamber 201 side by the divergent magnetic field of the magnetic coil 210. Further, high frequency power (for example, 500 W) is supplied from the high frequency power supply 222 to the target 205 disposed at the outlet of the plasma generation chamber 202. As a result, Ar particles collide with the target 205 to cause a sputtering phenomenon, and Si particles jump out of the target 205.

  After this state, when a shutter (not shown) between the target 205 and the substrate 101 is opened, the Si particles that have jumped out of the target 205 are discharged from the plasma generation chamber 202 and the reactive gas introduction unit 212. It reaches the surface of the substrate 101 together with oxygen gas that is introduced and activated by plasma, and is oxidized by the activated oxygen to form silicon dioxide.

  As described above, the lower insulating film 102 made of silicon dioxide and having a thickness of, for example, about 100 nm can be formed on the substrate 101 (FIG. 3A). After the film is formed to a predetermined thickness, the film formation is stopped by preventing the sputtered raw material from reaching the substrate 101 with the aforementioned shutter closed. Thereafter, the plasma irradiation is stopped by stopping the supply of microwave power or the like, the supply of each gas is stopped, the substrate temperature is lowered to a predetermined value, and the internal pressure of the processing chamber 201 is raised to the atmospheric pressure. Then, the substrate 101 on which the lower insulating film 102 is formed is unloaded from the inside of the processing chamber 201.

  The lower insulating film 102 is insulated so that the voltage is not leaked to the substrate 101 and the desired electrical characteristics are not affected when a voltage is applied to the conductive thin film 103 formed thereon and used as an electrode. It is intended. As long as it has insulating properties, it may be made of an insulating material other than silicon oxide, and the thickness of the lower insulating film 102 is not limited to 100 nm and may be thinner or thicker. The lower insulating film 102 is not heated with respect to the substrate 101 in the above-described film formation by ECR sputtering, but the film may be formed while the substrate 101 is heated. Further, a silicon oxide film formed by oxidizing the surface of the substrate 101 made of silicon by a thermal oxidation method may be used as the lower insulating film 102.

After the lower insulating film 102 is formed as described above, the substrate 101 is carried out from the apparatus to the atmosphere, and then the conductive thin film 103 is formed using pure ruthenium (Ru) as the target 205. The formation of the conductive thin film 103 will be described. The substrate 101 is fixed to the substrate holder 204 of the ECR sputtering apparatus similar to FIG. Subsequently, as shown in FIG. 3B, ruthenium nitride is formed so as to cover the surface of the lower insulating film 102 by ECR sputtering using Xe as the plasma gas and nitrogen (N ₂ ) as the reactive gas. A conductive thin film 103 made of is formed.

The formation of the conductive thin film 103 will be described in detail. In the ECR sputtering method shown in FIG. 2 using the target 205 made of pure ruthenium, first, the inside of the plasma generation chamber 202 is brought into a high vacuum state of 10 ⁻⁴ to 10 ⁻⁵ Pa level. After evacuation, the substrate 101 is heated to, for example, about 400 ° C., and Xe is introduced into the plasma generation chamber 202 from the inert gas introducing unit 211 at 26 sccm, for example, reactive gas such as nitrogen gas is about 8 sccm. Then, the pressure in the plasma generation chamber 202 is set to, for example, 10 ⁻¹ to 10 ⁻² Pa. In addition, a magnetic field of electron cyclotron resonance condition is provided in the plasma generation chamber 202 by supplying a coil current to the magnetic coil 210 at 26 A, for example.

  In addition, a 2.45 GHz microwave (for example, 800 W) is supplied from a microwave generation unit (not shown), and plasma is generated through the waveguide 208, the quartz window 207, and the vacuum waveguide 206. It introduce | transduces in the chamber 202, It is set as the state which the plasma which consists of Xe and reactive gas nitrogen produced | generated in the plasma production | generation chamber 202 by introduction of this microwave.

  The generated plasma is emitted from the plasma generation chamber 202 to the processing chamber 201 side by the divergent magnetic field of the magnetic coil 210. Further, high frequency power (for example, 500 W) is supplied from a high frequency power source 222 to a target 205 made of Ru disposed at the outlet of the plasma generation chamber 202. As a result, Xe particles collide with the target 205 to cause a sputtering phenomenon, and Ru particles jump out of the target 205. The Ru particles jumping out of the target 205 reach the surface of the lower insulating film 102 together with the nitrogen gas activated by the plasma, and the conductive thin film 103 composed of ruthenium nitride is deposited by the activated nitrogen.

  As a result, a state is obtained in which the conductive thin film 103 having a thickness of, for example, about 10 nm is formed on the lower insulating film 102 (FIG. 3B). Note that the thickness of the conductive thin film 103 is not limited to 10 nm, and may be thicker or thinner.

  Incidentally, as described above, when the conductive thin film 103 made of ruthenium nitride is formed by the ECR sputtering method, the substrate 101 is heated to 400 ° C., but this heating is not always necessary. However, when heating is not performed, the adhesion of ruthenium to silicon dioxide is reduced, and thus the conductive thin film 103 may be peeled off. In order to prevent this, it is preferable to form the film by heating the substrate 101. . After the conductive thin film 103 is deposited in a desired thickness as described above, the film formation is stopped by closing the shutter, and the termination process such as stopping the microwave irradiation and stopping the plasma irradiation is performed. Then, the substrate 101 can be carried out.

  Next, the characteristics of the conductive thin film 103 made of ruthenium nitride formed by ECR sputtering will be described in more detail. The inventors have repeatedly observed carefully, so that the surface irregularity, film formation speed, resistivity, and the like of the conductive thin film 103 formed during sputtering using the ECR plasma composed of the sputtering gas and the reactive gas nitrogen gas, It was also found that film properties such as crystallinity can be controlled by the amount (flow rate) of nitrogen gas supplied.

FIG. 4 is a characteristic diagram showing a change in deposition rate with respect to the ratio of introduced Xe gas and nitrogen gas when the conductive thin film 103 is deposited using a ruthenium target in the ECR sputtering method. FIG. 4 shows a case where a substrate 101 made of single crystal silicon is used in which a lower insulating film 102 having a thickness of 100 nm is formed by thermal oxidation, and the flow rate ratio of Xe gas to the total gas flow rate introduced on the horizontal axis. (Xe / (Xe + N ₂ )) is shown, and the film forming speed is shown on the vertical axis. That is, when the horizontal axis is 1, the supplied gas is only Xe gas, and when it is 0, the supplied gas is only nitrogen gas. The supply amount of Xe gas was a constant value of 26 sccm.

  From FIG. 4, it can be seen that when nitrogen is not introduced, the deposition rate, which was 5 nm / min, tends to gradually decrease as the proportion of nitrogen gas increases. When the nitrogen flow ratio is 0.2, the deposition rate is about 3 nm / min.

  Similarly, FIG. 5 shows a change in resistivity of the conductive thin film 103 with respect to the ratio of the introduced Xe gas and nitrogen gas when the conductive thin film 103 is formed using a ruthenium target in the ECR sputtering method. As shown in FIG. 5, when nitrogen is not introduced (Xe flow rate ratio 1), it can be seen that the resistivity of the conductive thin film 103 that was 30 μΩ-cm tends to increase gradually as the proportion of nitrogen gas increases. . The resistivity of the conductive thin film 103 becomes a constant value when the Xe flow ratio is 0.6 or less (nitrogen flow ratio 0.4) or more, and is about 50 μΩ-cm. From the above, it can be seen that the deposition rate and resistivity of the conductive thin film 103 can be accurately controlled by the nitrogen flow rate during sputtering.

  Next, FIG. 6 shows the conductive thin film 103 observed by X-ray diffraction (XRD) with respect to the ratio of the introduced Xe gas and nitrogen gas when the conductive thin film 103 is formed using a ruthenium target in the ECR sputtering method. Shows the change in peak intensity of Ru in the (002), (100), and (101) directions. In FIG. 6, the (002) direction is indicated by a white circle, the (100) direction is indicated by a black circle, and the (101) direction is indicated by a black square. The results (characteristics) in FIG. 6 are obtained when a substrate 101 made of single crystal silicon and having a lower insulating film 102 having a thickness of 100 nm formed by thermal oxidation is used.

In FIG. 6, the flow rate ratio “Xe / (Xe + N ₂ )” of the Xe gas with respect to the total gas flow rate introduced on the horizontal axis is shown, and the vertical axis shows (002), (100), And (101) plane peak intensity in arbitrary units. The lattice spacing on the (002) plane is 0.2142 nm, the lattice spacing on the (100) plane is 0.2343 nm, and the lattice spacing on the (101) plane is 0.2056 nm. When the horizontal axis is 0, only nitrogen gas is used, and when 1 is Xe gas only. The Xe gas was a constant value of 26 sccm.

  From FIG. 6, when nitrogen is not introduced, in other words, the ruthenium film not containing nitrogen is oriented in the (002) plane and is well-known C-axis orientation. On the other hand, in the conductive thin film 103 formed by introducing nitrogen gas at a flow rate ratio of about 0.2 (corresponding to 0.8 in FIG. 6), the peaks of (101) and (100) are higher than the peak of (002). It becomes clear that the crystal structure of the film changes greatly. This phenomenon is not known, and has been made possible for the first time by introducing nitrogen using the ECR sputtering method. As the amount of introduced nitrogen is further increased, the peaks at (101) and (100) become smaller, but the peak at the (002) plane becomes larger. At a nitrogen flow rate ratio of 0.4 (corresponding to 0.6 in FIG. 6), the peak intensity is approximately the same. From these results, it can be seen that the crystallinity of the conductive thin film 103 can be controlled by the nitrogen flow rate during sputtering.

Similarly, Ru (002), (100), and (101) in the conductive thin film observed by X-ray diffraction (XRD) with respect to the ratio of the introduced Xe gas and nitrogen gas when the substrate temperature is heated to 100 ° C. The change in the intensity of the direction peak is shown in FIG. In FIG. 7, the flow rate ratio “Xe / (Xe + N ₂ )” of the Xe gas with respect to the total gas flow rate introduced on the horizontal axis is shown, and (002), (100), and ( The value obtained by standardizing the intensity of the peak in the 101) direction with the peak intensity of (002) is shown in arbitrary units. The Xe gas was set to a constant value of 10 sccm.

  As shown in FIG. 7, when nitrogen is not introduced, in other words, the ruthenium film not containing nitrogen is oriented in the (002) plane and is well-known C-axis orientation. In contrast, in the conductive thin film formed by introducing nitrogen at a flow rate ratio of 0.2 (corresponding to 0.8 in FIG. 7), the peak of (002) is the same as the peak intensity of (100) and (101). It can be seen that the crystal structure of the conductive thin film changes.

  Furthermore, in the conductive thin film formed by introducing nitrogen at a flow rate ratio of 0.4 (corresponding to 0.6 in FIG. 7) or more, it is oriented in the (100) plane from the (002) peak, and the A axis It can be seen that the crystal structure of the conductive thin film changes greatly due to the orientation. Further, when the amount of nitrogen introduced is further increased, the peaks of (101) and (100) become maximum at a nitrogen flow rate ratio of about 0.5 and thereafter gradually decrease. When a nitrogen flow rate ratio of 0.6 (corresponding to 0.4 in FIG. 7) or more is introduced, the flow rate becomes smaller than the (002) peak.

  Although the Xe gas flow rate to be introduced and the sputtering conditions of the substrate temperature are slightly changed, the intensity of the peak showing the (100) A-axis orientation is higher than the peak showing the (002) C-axis orientation by the introduced nitrogen flow rate. A phenomenon of increasing appears, and it can be seen that the crystallinity of the conductive thin film can be controlled by the nitrogen flow rate during sputtering.

FIG. 8 is an overhead view showing an atomic force microscope (AFM) image showing surface unevenness with respect to the flow rate of the introduced nitrogen gas. 8A shows the state of the ruthenium film formed without introducing nitrogen, and FIG. 8B shows the state of the conductive thin film 103 formed by introducing 8 sccm of nitrogen gas. ) Shows the state of the conductive thin film 103 formed by introducing 20 sccm of nitrogen gas. The nitrogen gas flow rate ratio “N ₂ / (Xe + N ₂ )” in FIG. 6 corresponds to 0.82 for introducing nitrogen gas at 8 sccm, and the nitrogen gas flow rate at 20 sccm in FIG. The ratio “N ₂ / (Xe + N ₂ )” corresponds to 0.58. Under any of the conditions, a smooth state with few surface irregularities is obtained, but it can be seen that a flatter film is obtained as the amount of nitrogen gas introduced is increased.

  Furthermore, in order to investigate in detail, the characteristic view which showed the change of the surface unevenness | corrugation with respect to the flow volume of the introduced nitrogen gas is shown in FIG. FIG. 9 shows the flow rate of nitrogen gas introduced on the horizontal axis and the surface irregularities on the vertical axis when the substrate 101 made of single crystal silicon is used to form the lower insulating film 102 having a thickness of 100 nm by thermal oxidation. As a guideline showing the property, the mean square surface roughness (RMS) observed with an atomic force microscope and the average particle diameter are shown. In addition, the average particle height is obtained from the unevenness of the surface unevenness image (topology image) by an atomic force microscope, and binarization processing is performed with a half value of the obtained average particle height as a threshold value. The surface of each particle to be cut from the side is regarded as a circle, the value of this diameter is obtained, and the value of the diameter obtained for each observed particle is averaged to obtain the average particle size.

  From FIG. 9, it can be seen that the RMS of the non-nitrided ruthenium film is about 0.14 nm, and the surface unevenness is quite small. However, when nitrogen is introduced, the RMS value of the formed conductive thin film 103 tends to be further reduced. When the nitrogen introduction amount was 20 sccm, it was 0.13 nm, and it was found that the surface of the conductive thin film 103 was in a very smooth state. As can be seen from FIG. 9, when nitrogen is introduced, the average particle size becomes smaller than a certain amount introduced, so that the in-plane orientation is changed by nitriding and the grain size of ruthenium is reduced. It is done.

  Here, referring to the results shown in FIG. 6 and the results shown in FIGS. 5 and 8, the conductive thin film 103 has a resistivity as long as the peak of (100) is higher than the peak of (002). It can be seen that a higher flatness can be obtained than in the case of ruthenium alone, in a range where is not so high. In other words, the conductive thin film 103 composed of ruthenium and nitrogen has higher flatness than the case of only ruthenium if the A-axis oriented structure is larger than the C-axis oriented structure. Conceivable.

  Furthermore, an observation view of a cross-sectional electron microscope (TEM) is shown in FIG. FIG. 10 shows the above-described ECR sputtering method in which a substrate 101 made of single crystal silicon is formed with a lower insulating film 102 having a thickness of 100 nm formed by thermal oxidation, the Xe gas flow rate is 26 sccm, and the nitrogen gas flow rate is 8 sccm. The case of the conductive thin film 103 formed by the above is shown. FIG. 10 shows a state in which a lower insulating film, a conductive thin film, and a protective film formed thereon for protection are observed. From FIG. 10, it can be seen that the crystals in the conductive thin film 103 have a lateral orientation as well as the C-axis direction. It can also be seen that the surface irregularities are also smooth because the grain size is small.

  Since the surface of the conductive thin film 103 made of ruthenium nitride formed by ECR sputtering film formation was found to be very smooth, a bismuth titanate film was formed as a CER film on the thin film thus formed. Then, an element as shown in FIG. 11 was fabricated and the resistance switch characteristics were confirmed. The element shown in FIG. 11 will be described. For example, a lower insulating film 1102 made of silicon dioxide is provided on a substrate 1101 made of single crystal silicon, and a lower electrode made of ruthenium nitride and having a thickness of 10 nm is formed thereon. Layer 1103 is provided. The lower electrode layer 1103 corresponds to the conductive thin film 103 described above.

The element shown in FIG. 11 includes a CER layer 1104 having a thickness of 30 nm made of a metal oxide containing Bi ₄ Ti ₃ O ₁₂ on the lower electrode layer 1103, and a film made of Au on the CER layer 1104. An upper electrode 1105 having a thickness of 80 nm is provided. As described above, the configurations of the substrate and the insulating film are not limited to this, and other materials can be appropriately selected as long as they do not affect the electrical characteristics.

  The manufacturing of the element shown in FIG. 11 will be briefly described. As described above, the substrate in which the lower insulating film 1102 is formed on the substrate 1101 is taken out into the atmosphere from the ECR sputtering apparatus, and then pure ruthenium. A lower electrode layer 1103 is formed by depositing a conductive thin film on the lower insulating film 1102 by the same ECR sputtering method as described above using a target made of a material. For example, the substrate 1101 is fixed to the substrate holder 204 of the ECR sputtering apparatus shown in FIG. Subsequently, a conductive thin film made of ruthenium nitride is deposited on the lower insulating film 1102 to cover the surface by ECR sputtering using Xe as the plasma gas and nitrogen as the reactive gas. The layer 1103 is formed.

  The substrate 1101 on which the lower electrode layer 1103 made of the conductive thin film of the present invention is formed as described above is carried out into the atmosphere from the inside of the apparatus, and then a sintered body having a ratio of Bi and Ti of 4: 3 as the target 205 ( The substrate 1101 is fixed to the substrate holder 204 of the ECR sputtering apparatus of FIG. 2 using Bi—Ti—O). Subsequently, a CER layer 1104 made of a metal oxide composed of Bi, Ti, and O is formed so as to cover the surface by ECR sputtering using Ar and oxygen as plasma gases.

The formation of the CER layer 1104 made of the metal oxide (ferroelectric material) will be described in detail. First, in the ECR sputtering apparatus shown in FIG. 2 using the target 205 made of Bi—Ti—O, the inside of the plasma generation chamber 202 is first formed. After evacuating to a high vacuum state of 10 ⁻⁴ to 10 ⁻⁵ Pa, the substrate 1101 is heated to 30 ° C. to 700 ° C., and the inert gas introduction unit 211 is provided in the plasma generation chamber 202 by, for example, Ar gas is introduced at a flow rate of 20 sccm, and the pressure in the plasma generation chamber 202 is set to, for example, 10 ⁻² to 10 ⁻³ Pa. In addition, a magnetic field under electron cyclotron resonance conditions is provided in the plasma generation chamber 202 by supplying a coil current to the magnetic coil 210 at 27 A, for example.

  In addition, a 2.45 GHz microwave (for example, 500 W) is supplied from a microwave generation unit (not shown), and this is supplied to the plasma generation chamber via the waveguide 208, the quartz window 207, and the vacuum waveguide 206. It introduce | transduces in 202, It is set as the state by which the plasma was produced | generated in the plasma production chamber 202 by introduction of this microwave.

  The generated plasma is emitted from the plasma generation chamber 202 to the processing chamber 201 side by the divergent magnetic field of the magnetic coil 210. Further, high frequency power (for example, 500 W) is supplied from the high frequency power supply 222 to the target 205 disposed at the outlet of the plasma generation chamber 202. As a result, Ar particles collide with the target 205 to cause a sputtering phenomenon, and Bi particles and Ti particles jump out of the target 205.

  Bi particles and Ti particles jumping out of the target 205 are formed on the surface of the lower electrode layer 1103 together with the plasma released from the plasma generation chamber 202 and the oxygen gas introduced from the reactive gas introduction unit 212 and activated by the plasma. It reaches and is oxidized by the activated oxygen. The oxygen gas may be introduced at about 1 sccm from the reactive gas introduction unit 212, for example. In this example, the target 205 is a sintered body and contains oxygen. However, supply of oxygen can prevent oxygen shortage in the film.

  By forming the film by the ECR sputtering method described above, for example, a state in which the CER layer 1104 having a thickness of about 30 nm is formed can be obtained. Thereafter, the end process is performed in the same manner as described above, and the substrate 1101 can be unloaded.

  Here, in the element shown in FIG. 11, the CER layer 1104 is formed on the lower electrode layer 1103 made of ruthenium nitride. As a result, the smoothness of the surface of the lower electrode layer 1103 formed by the above-described ECR sputtering method is not deteriorated, and the CER layer 1104 can be formed thereon. For example, if the lower electrode is made of a metal material and is easily oxidized, the surface of the lower layer may be partially oxidized during the formation of the above-described metal oxide film, and morphology may deteriorate. On the other hand, according to the element shown in FIG. 11, since the lower electrode layer 1103 made of ruthenium nitride is used as a base, the oxidation of the surface is suppressed, and the surface can be formed with good morphology. A high CER layer 1104 is obtained.

  An upper electrode 1105 made of ruthenium is formed on the surface of the CER layer 1104 formed as described above. For example, the upper electrode 1105 can be formed by processing a ruthenium film by a well-known photolithography technique and etching technique.

  Next, the electrical characteristics of the element (resistive switch element) shown in FIG. 11 manufactured as described above will be described with reference to FIG. This electrical characteristic was investigated by applying a voltage between the lower electrode layer 1103 and the upper electrode 1105. A voltage is applied between the lower electrode layer 1103 and the upper electrode 1105 by a power source, and a current value when the voltage is applied is observed by an ammeter. In FIG. 12, the horizontal axis represents the voltage value applied to the upper electrode 1105, and the vertical axis represents the absolute value of the current value in logarithm.

  In the following, the characteristics shown in FIG. 12 will be described. The voltage values and current values described here are those observed with actual elements as an example. Therefore, this phenomenon is not limited to the following numerical values, and other numerical values may be observed depending on the material and film thickness of the film actually used for the element and other conditions.

First, when a positive voltage is applied to the upper electrode 1105, as shown in (1) of FIG. 12, at 0 to 0.8 V, the current is 10 ⁻⁶ A or less with respect to +0.5 V, and the resistance state is high. . However, as shown in (2), when 0.8V is exceeded, a positive current suddenly flows and a low resistance state is entered. When a voltage of 0 to 0.8 V is applied so that a positive current does not flow suddenly as shown in (2), the high resistance state is maintained as shown in (1).

When a positive voltage is applied again to the upper electrode 1105 after the low resistance state as shown in (2), a positive current of about 1 × 10 ⁻⁵ A flows at about 0.5 V as shown in (3). Subsequently, when a negative voltage is applied to the upper electrode 1105, a current of about 3 × 10 ⁻⁵ A flows at about −0.5 V as shown in (4), indicating that the state is in a low resistance state. However, when a negative voltage exceeding −0.8 V is applied to the upper electrode 1105, current suddenly stops flowing as shown in (5), and a transition is made to the high resistance state.

Thereafter, as shown in (6), even when a negative voltage is applied to the upper electrode 1105, a high resistance state of 10 ⁻⁶ A or less is maintained at −0.5V. Subsequently, when a positive voltage is applied to the upper electrode 1105, it is in a high resistance state up to about + 0.8V as in (1), but transitions to a low resistance state by a positive current exceeding + 0.8V. Hereinafter, a phenomenon in which the high resistance state and the low resistance state are switched reversibly is stably observed.

What should be noted here is the current value of the low resistance state, that is, the current value of the On current. In the prior art shown in FIG. 15, the value of the On current is about 10 ⁻² A. On the other hand, in the element shown in FIG. 11, as shown in FIG. 12, it can be seen that the On current value is reduced to about 10 ⁻⁴ to 10 ⁻⁵ A, 2 to 3 digits. This is presumably because the surface of the conductive thin film (lower electrode layer 1103) made of ruthenium nitride is in a very smooth state.

  Furthermore, the voltage value for switching the resistance is about 1.5 V in the prior art, but according to the element shown in FIG. 11, the voltage is reduced to ± 0.8 V and the power consumption is small. . Low power consumption is very advantageous for devices, for example, mobile communication devices, digital general-purpose devices, digital imaging devices, notebook-type personal computers, and personal digital appliances (PDAs) only In addition, the power consumption of all electronic computers, personal computers, workstations, office computers, large computers, communication units, multifunction devices, and other devices using memory can be reduced.

  By the way, in the element shown in FIG. 11 described above, the bismuth titanate film is used as the CER film, but a CER film made of another material may be used. In the example of the present invention described above, each of the base insulating layer, the metal thin film, and the CER film layer on the silicon substrate is formed by the ECR sputtering method. However, the method for forming each of these layers is not limited to the ECR sputtering method. For example, the base insulating layer on the silicon substrate may be formed by thermal oxidation, chemical vapor deposition (CVD), ALD, conventional sputtering, MOCVD, or the like. The upper electrode may be formed by other methods such as CVD, MBE, IBD, conventional sputtering, and PDL.

  Further, the metal oxide film layer exhibiting CER characteristics may be formed by other methods such as the MOD method, the conventional sputtering method, and the PLD method. However, a flat and good insulating film, metal film, and ferroelectric film can be easily obtained by using the ECR sputtering method.

  Further, in the above embodiment, after each layer is formed, it is once taken out to the atmosphere. However, the processing chambers for realizing each ECR sputtering may be connected by a vacuum transfer chamber by continuous processing. As a result, the substrate to be processed can be transported in a vacuum, and is less susceptible to disturbances such as moisture, leading to improvements in film quality and interface characteristics.

  Further, as shown in Japanese Patent No. 3571679, after each layer is formed, the surface of the formed layer may be irradiated with ECR plasma to improve the characteristics of the film. Further, after each layer is formed, as shown in Japanese Patent Application Laid-Open No. 2004-273730, annealing may be performed in an appropriate gas atmosphere to improve the characteristics.

  Also, the thickness of the conductive thin film is preferably set to an optimal and optimal thickness. For example, in consideration of peeling due to film stress, the conductive thin film desirably has a thickness of 100 nm or less. Further, considering the resistance value as the wiring, it is better to make it thicker than 10 nm. As a result of experiments by the inventors, if the thickness of the ferroelectric is 10 to 100 nm, the operation of the memory is confirmed, and the best state is obtained when the thickness of the ferroelectric layer is 20 nm. .

It is a block diagram which shows the structure of the conductive thin film 103 which concerns on embodiment of this invention. It is a block diagram which shows the structural example of an ECR sputtering apparatus. 5 is a process diagram for explaining an example of a method for manufacturing the conductive thin film 103. FIG. In the ECR sputtering method, when the conductive thin film 103 is formed using a ruthenium target, it is a characteristic diagram showing a change in film formation rate with respect to the ratio of introduced Xe gas and nitrogen gas. FIG. 10 is a characteristic diagram showing a change in resistivity of the conductive thin film 103 with respect to the ratio of introduced Xe gas and nitrogen gas when the conductive thin film 103 is formed using a ruthenium target in the ECR sputtering method. In the ECR sputtering method, when the conductive thin film 103 is formed using a ruthenium target, Ru (002), (100) in the conductive thin film 103 observed by X-ray diffraction with respect to the ratio of the introduced Xe gas and nitrogen gas. , And a characteristic diagram showing a change in peak intensity in the (101) direction. In the ECR sputtering method, when the conductive thin film 103 is formed using a ruthenium target, Ru (002), (100) in the conductive thin film 103 observed by X-ray diffraction with respect to the ratio of the introduced Xe gas and nitrogen gas. , And a characteristic diagram showing a change in peak intensity in the (101) direction. It is an atomic force microscope (AFM) image which shows the surface unevenness | corrugation with respect to the flow volume of the introduced nitrogen gas. It is a characteristic view which shows the change of the surface unevenness | corrugation with respect to the flow volume of the introduced nitrogen gas. It is a cross-sectional electron micrograph of a conductive thin film. It is a block diagram which shows the structural example of the element produced using the conductive thin film. It is explanatory drawing explaining the electrical property of the element shown in FIG. It is a block diagram which shows the structural example of stack type FeRAM. It is a block diagram which shows the structural example of the element which implement | achieves a memory function by the change of the resistance value of a ferroelectric layer. It is explanatory drawing which shows an example of the voltage-current characteristic of the resistance switch phenomenon in the element shown in FIG.

Explanation of symbols

  101 ... Substrate, 102 ... Lower insulating film, 103 ... Conductive thin film.

Claims (4)

A conductive thin film formed on a substrate,
The conductive thin film is composed of ruthenium and nitrogen, and has more A-axis oriented structures than C-axis oriented structures. A plasma composed of an inert gas and a nitrogen gas supplied at a predetermined composition ratio is generated, a negative bias is applied to a target composed of ruthenium, and particles generated from the plasma collide with the target to perform sputtering. A method for producing a conductive thin film, comprising: forming a conductive thin film composed of ruthenium and nitrogen on the substrate by causing a phenomenon and depositing ruthenium constituting the target on the substrate. . In the manufacturing method of the electroconductive thin film of Claim 2,
The method of manufacturing a conductive thin film, wherein the plasma is an electron cyclotron resonance plasma generated by electron cyclotron resonance and given kinetic energy by a divergent magnetic field. In the manufacturing method of the electroconductive thin film of Claim 2 or 3,
The substrate is heated to a predetermined temperature. A method for producing a conductive thin film.