JPH0793290B2 - Fine structure formation method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION <Field of Industrial Application> The present invention is applicable to various electronic devices such as a fine gate or fine wiring in a semiconductor device, a fine width strip line in a microwave device, a switching gate structure or a fine wiring in a Josephson device. The present invention relates to a method for forming various microstructures required to have dimensions on the order of submicron in an apparatus, and various fine structures required to have dimensions on the order of submicron in other optical devices, micromechanical parts, etc. .

<Conventional technology and its problems> The miniaturization of semiconductor devices has been progressing year by year in various electronic devices and optical devices, etc. Optical exposure technology is used.

However, considering that the photolithography technology alone has a limit to miniaturization in principle, various electronic devices will inevitably demand patterns ranging from submicron order to less than that in the future. , It is unlikely that we will be able to follow the momentum of such ultra-miniaturization.

On the other hand, there are electron beam exposure technology, ion beam exposure technology, and X-ray exposure technology as methods that are theoretically considered to be more advantageous for forming fine patterns than the light exposure technology.

Among them, the electron beam exposure technology and the ion beam exposure technology are currently under 0.1 if only fine lines can be drawn.
Although fine lines of μm or less can be exposed, the principle is that the target thin line must be drawn continuously like a single stroke with the beam, so mass productivity is lacking. The device itself is large and expensive.

In particular, since electron beam exposure is 1/100 or less of the wavelength of light in terms of its wavelength, it is easy to think that it is possible to draw in such a minute order as it is, but electrons also have the property of particles. Therefore, the scattering of electrons in the photoresist or the substrate affects, and the size limit of pattern miniaturization as a practical machine is considerably large.

On the other hand, with respect to X-ray exposure, the wavelength itself is considerably short as in the above beam exposure, and therefore it seems that a minute pattern may be drawn only from this, but the X-ray source is a so-called point source. Since it is not, there is penumbra blurring, and fine patterning down to 0.1 μm or less is in the present situation which must wait for future development.

Further, since a very thin membrane of micron order is required for the mask used for this X-ray exposure, in order to prevent the membrane mask from being damaged, it must be placed away from the substrate. However, this also causes the above-described penumbra blurring, which causes an obstacle in fine pattern formation.

In order to eliminate this penumbra blurring, the use of parallel X-rays such as synchrotron orbit synchrotron radiation (SOR) can be considered, but with this, a new problem of diffraction phenomenon will occur in the future, After all, there is a limit to fine patterning.

As described above, since various exposure techniques are inevitably limited in forming fine patterns, fine processing techniques that do not rely only on such exposure techniques have been taken into consideration. Typical examples thereof include a method using side etching and an oblique deposition technique utilizing a shadowing effect.

However, these techniques still have problems in their reproducibility and accuracy.

On the other hand, as another conventional method, only the side surface of the film member formed on the surface of the substance is exposed, and the chemical property of the surface portion of the exposed side surface is changed so as to be different from the internal chemical material. Then, there is a method of obtaining a structure having a fine width (for example, Japanese Patent Laid-Open No. 52-56874). But,
It is difficult to use such a fine width structure as it is alone and in a self-supporting state, and by using it only as a fine width mask, it is possible to form an independent fine structure from surrounding portions having different properties, etc. It was limited. Generally speaking, even if a fine structure is formed, in order to use the fine structure as a member that actually operates in an electronic circuit, such as using itself as a gate of a transistor or a current line, Although it is often requested to integrally connect to a larger structure having an arbitrary shape having the same type of properties as the structure, a rational method to meet this demand has not been disclosed.

In view of the state of the art as described above, the present invention can form a smaller fine structure without leaving the size limit of the fine pattern to only various exposure techniques, and the fine structure thus formed. It is an object of the present invention to provide a rational and novel method for forming a fine structure, which can integrally connect a structure having a larger dimension order to a structure having an arbitrary shape.

<Means for Solving the Problems> In order to achieve the above object, the present inventor first proposes, as a first invention, a fine structure forming method including the following process groups.

A step of forming a layer to be altered and a mask layer in the height direction on the underlayer. A step of removing a part of the mask layer according to a predetermined mask removal pattern to expose the altered layer in the part where the mask is removed. An altered structure forming step of altering the exposed portion of the altered layer to form an altered structure. The mask layer remaining around the altered structure and the altered layer below the patterned layer are patterned into a predetermined shape having a portion connected to the altered structure, or a mask layer patterned into the prescribed shape is formed. Step of patterning the altered layer into a predetermined shape as a mask for further patterning. A fine width alteration process that exerts an alteration action only on the side surface of the altered layer by using the mask layer remaining on the altered layer patterned in a predetermined shape as a mask for selectively altering the side surface of the altered layer. ． Then, a step of removing the unaltered portion of the mask layer and the altered layer, and leaving the side portion altered in the above as a fine width altered structure connected to the altered structure. Altered structure and damage width Using the altered structure as a mask, a structure having a relatively large size having a shape according to the plane shape of the altered structure and a structure having a relatively large size connected to a fine width A step of forming a fine structure having a shape according to the plane shape of the altered structure on the above-mentioned underlayer. On the other hand, the present application also proposes, as a second invention, a method in which the order of the altered structure forming step and the fine width forming step is substantially reversed. Therefore, similarly to the above, the steps of the second invention of the present application will be listed as follows, respectively.

A step of forming a layer to be altered and a mask layer in the height direction on the underlayer. The altered layer is patterned together with the mask layer into a predetermined shape, or the patterned layer is patterned into a predetermined shape and then the modified layer is used as a further patterning mask to define the altered layer. Patterning process. The mask layer remaining on the deteriorated layer patterned into a predetermined shape is used as a mask for selectively changing the side surface of the deteriorated layer, and a fine effect is exerted only on the side surface of the deteriorated layer. Width alteration process. After the fine width alteration step, a part of the mask layer is removed according to a predetermined mask removal pattern so that the alteration layer is exposed including at least a part of the surface of the side surface portion affected by the alteration action in the alteration layer. Process to do. An altered structure forming step of altering the exposed portion of the altered layer to form an altered structure. Then, a step of removing the unaltered portion of the mask layer and the altered layer, and leaving the side portion altered in the above as a fine width altered structure connected to the altered structure. Using the altered structure and the fine width altered structure as a mask,
A structure having a relatively large size according to the plane shape of the altered structure, and a fine structure having a shape according to the plane shape of the fine width altered structure connected to the relatively large size structure. Forming the above-mentioned underlayer. <Operation and Effect> In the above-described constitution of the present invention, “altered” means that “non-altered” portion is formed by adding impurities, changing chemical composition, or changing crystal form. And the subsequent gas phase or liquid phase are different in etching rate in etching (generally smaller than the portion to be altered), which is a function of causing a chemical quality change.

However, in the present invention, as recognized in the above-mentioned gist constitution, when roughly divided, a process group for obtaining a fine width altered structure, and at least finally, an altered structure integrally connected to the fine width altered structure. It consists of a group of steps. Therefore, first of all, regarding the formation of the fine width altered structure, of the area of the altered layer initially patterned to have a predetermined shape, the only remaining one is the side portion of the altered layer that has undergone the alteration action. In addition, since the accuracy of the post-processing after receiving the alteration action does not affect the dimensional accuracy, the altered structure having an extremely fine width can be obtained as the remaining altered portion.

In other words, even when the existing light exposure technique is used for patterning the mask layer into a predetermined shape, the two-dimensional size limit of the fine width altered structure finally formed is the dimension in the light exposure technique. It is not affected by the limit, and it is possible to reduce the dimension limit by one digit or more to the dimension limit of the photolithography technology. And, since this residual altered portion (fine width altered structure) is finally integrally connected to the altered structure having a larger shape and size, at least as in the above-mentioned conventional example, it is simply It is not limited to its use as an independent, fine width mask that is independent and has a different material from the surroundings, and can be used for specific electronic devices and short wavelength optical devices that actually operate fine parts and micromechanical parts. It can be significantly used as a mask when forming a structure connected to another large structure.

Regarding the formation of the fine width altered structure, a little more specifically, for example, when a silicon thin film is selected as the altered layer, the mask layer having a predetermined shape formed thereon is used as a mask for selective oxidation. When only the side surface of the altered silicon layer is used, the etching rate of the side surface portion altered by the oxidation with respect to the etching liquid such as hydrofluoric acid is not altered (not oxidized). Compared to that, it becomes significantly smaller. Therefore, after the fine width alteration step, only the fine width altered side surface portion can be left simply by removing the max layer and immersing the whole substrate in such a hydrofluoric acid solution.

As is clear from such a principle, the mask layer having a predetermined shape formed on the layer to be altered is a planar or two-dimensional structure of the finally-altered altered side surface portion (fine width altered structure). Not only as a patterning mask for defining the overall contour shape, but also as a side-face selective alteration mask for altering only the side face portion of the altered layer as expected. If this mask were not provided, the alteration from the upper surface of the altered layer would proceed, and it would be extremely difficult to make the thickness of the altered side surface uniform in the height direction. Controlling the thickness of a fine structure having an extremely small width is often important. Therefore, particularly when the alteration action depends on oxidation as described above, a material layer containing at least a nitriding component, that is, silicon nitride, oxynitride, or the like is desirable as a mask layer therefor.

However, the alteration action may be based on any principle that is most suitable for the selected material to be altered, and even when a silicon thin film is selected as the alteration layer, the alteration action is not limited to oxidation, but may depend on nitriding. Furthermore, the silicon thin film is a single crystal or a polycrystal, and the concentration of boron added by diffusion or the like is
When it is 10 ¹⁹ atoms / cm ³ or more, the etching rate for an inorganic or organic alkaline aqueous solution of caustic potash, hydrazine, etc. is remarkably reduced as compared with a silicon thin film to which boron is not added. Means can also be used for alteration. In this case, as the mask layer, silicon nitride or the like can be used in addition to silicon oxide, and even if the altered layer is amorphous silicon, it is possible to adopt a means similar to the above, such as an oxidation method. it can.

Next, considering the process group for connecting the fine width altered structure to the altered structure which is good in the conventional dimension order, in the case where the altered structure is produced in advance according to the first invention of the present application, the alteration The mask layer remaining around the structure and the deteriorated layer thereunder are patterned into a predetermined shape having a portion connected to the deteriorated structure, or a mask layer patterned into the predetermined shape is further added. As the patterning mask, the layer to be altered is patterned into a predetermined shape and then subjected to the above-described side surface alteration to the layer to be altered, so that the alteration structure is formed on the basis of the side surface alteration. The fine width altered structure can be physically and integrally connected.

On the other hand, also in the case according to the second invention of the present application, after the fine width alteration step, the alteration layer is exposed so as to include at least a part of the surface of the side surface portion that has been altered in the alteration layer, After removing a part of the mask layer according to a predetermined mask removal pattern, the exposed portion of the altered layer is altered to form an altered structure, so that the altered structure is physically formed on the fine width structure. , An integrally connected structure can be obtained.

As a result, as mentioned earlier, the conventional constraint that a fine width altered structure must be used alone and in a self-supporting state is solved, and a fine altered structure with a larger altered structure in the same altered state is Since the width-altered structure can be integrally formed in a continuous process, as can be seen in Examples described later, for example, a portion of the fine width-altered structure is used as a mask for MOSF.
It is rational to form the gate of ET or its wiring part, on the other hand, the bonding pad and the wide lead line electrically connected to it are constructed by using a modified structure that is good as a conventional size structure as a mask. Various applications will be developed by a simplified manufacturing process.

It should be noted that since the mask layer other than the portion removed according to the mask removal pattern is left as it is during the alteration process for forming an altered structure that is good in the shape of a size larger than that of the fine width altered structure, the alteration action For example, when an oxidative alteration action is caused, the portion exposed through the removed mask portion in the altered layer swells largely due to the entire surface oxidation, so that it can be clearly distinguished from other portions, and if necessary. An altered structure having a sufficient thickness can be obtained.

<Examples> FIG. 1 shows an example of steps for obtaining a fine width altered structure 50 in the fine structure forming method of the present invention.

As shown in FIG. 1 (A), a silicon oxide layer 21 is formed on a silicon substrate 20, and a silicon nitride layer 22 as a protective mask layer is further formed on the silicon oxide layer 21.
Such a three-layer structure 20, 21, 22 constitutes the underlayer 10 referred to in the present invention in this case.

A polycrystalline silicon layer 30 is formed as a modified layer 30 on the silicon nitride layer or the protective mask layer 22, and a silicon nitride layer 40 is further formed thereon as a mask layer for side surface selective modification of the modified layer 30. Has been formed.

For such a starting structure, first, the uppermost silicon nitride layer
40 is patterned into a predetermined shape by a normal photolithography technique or a vapor phase etching technique using a reactive plasma, and the patterned silicon nitride layer 40 is used as a patterning mask for the altered layer 30. The altered layer 30 is patterned into a predetermined shape corresponding to the planar patterning of the silicon nitride layer 40 by liquid phase etching using nitric hydrofluoric acid or the like or reactive ion etching using CF ₄ gas.

As a result, as shown in FIG. 1 (B), the side surface portion exposed to the outside of the altered layer 30 according to the pattern.
35 appears.

Following this step, when the remaining silicon nitride layer 40 is used as a mask for thermal oxidation and exposed to the thermal oxidation atmosphere, only the side surface portion 35 of the exposed alteration layer 30 is selectively oxidized, As shown in FIG. 1C, a silicon oxide portion (altered portion) 50 having a fine width W is formed.

With regard to this step, the silicon nitride layer 22 formed on the surface of the underlayer 10 oxidizes species such as oxygen from the bottom in the above-mentioned oxidation step due to oxidation species such as oxygen diffusing on the underlayer surface. It functions as a protective mask layer to prevent

This is because, if the silicon nitride layer 22 were not present, as shown in FIG. 2, the oxidizing species would diffuse through the oxidizing layer portion such as the silicon oxide layer 21 formed on the surface side of the underlayer. For this reason, the side surface of the altered layer 30 is also oxidized from the bottom surface side as shown by the portion 51, and as a result, the cross section becomes an extremely distorted shape, and the dimensional accuracy is completely eliminated. In the same manner, in this oxidation step, the upper surface region 52 of the substrate 20 may also be accidentally oxidized through the bottom surface side of the silicon oxide layer 21 on the surface of the underlayer. is there.

However, even in the case of oxidation, depending on the oxidation time and degree, such an antioxidation protection mask layer 22 may not be necessary, and as will be described later, in the case of relying on another method for alteration, If desired, a protective mask 22 of a material suitable for each can be used.

On the other hand, however, when the silicon nitride layer and the polycrystalline silicon layer are brought into direct contact with each other as in the case described above, thermal strain is added, which may adversely affect the mechanical strength and electrical characteristics. Therefore, in order to prevent or reduce this, the mask layer may have a multi-layer structure. At this time, although it is desirable to select an oxide film as the layer in contact with the polycrystalline silicon layer, if it is done so, As described above, the dimensional accuracy of the deteriorated layer may be significantly impaired by the diffusion of the oxide film.

In order to avoid this phenomenon, the thickness of the oxide film should be equal to or less than the thickness of the side surface alteration layer. However, more generally speaking, considering the influence of thermal strain, when forming a mask layer having a protective meaning particularly in a multi-layer structure, it is preferable that the entire relative thickness is equal to or less than the side surface altered thickness. Most desirable.

Now, if the structure shown in FIG. 1 (C) is obtained, the remaining silicon nitride layer 40 and the unaltered portions of the altered layer 30 are removed by etching with, for example, a nitric hydrofluoric acid solution. , And possibly the protective masking layer 22 thereunder.
If the remaining deteriorated portion 50 is removed as a mask, as shown in FIG. 1 (D), the fine deteriorated portion 50 or the fine deteriorated portion 50 and the minute portion of the silicon nitride layer 22 thereunder may be removed. A fine structure can be obtained.

The fine structure thus obtained is, as will be understood, not affected by the size limit of the optical exposure technique used for the initial patterning, and can be much finer than that. It can be used as it is as a fine structure, or it can be used as a mask for selective removal of the layer portion (layer 21 and below) below it.

FIG. 3 relates to the fine structure in which the alteration action is left to oxidation as described above, and the oxidation amount to the altered layer 30, that is, (100)
The relationship between the oxidation thickness or the oxidation time of the surface single crystal silicon and the oxidation thickness from the outer end surface of the altered layer side surface 35, that is, the width W in FIGS. 1C and 1D is shown.

As can be seen in this figure, by appropriately selecting the variable mass of the altered layer 30, that is, the oxidation time of the polycrystalline silicon layer 30, the width of the fine structure is linearly related to the square root of the oxidation time. It can be seen that the dimension W can be controlled with good reproducibility.

Further, according to the above method, as shown schematically in FIG. 4, the planar shape 53 of the fine structure to be formed follows the contour 41 of the planar shape of the upper mask layer 40. ,
Even if the predetermined planar shape is irregular, it does not matter at all, and it can be made into a considerably complicated shape.

5 (A) to 5 (C) show that, in addition to the steps for obtaining the fine width altered structure 50, the fine width altered structure 50 has been subjected to the same kind of alteration action, but in the conventional dimension order. An embodiment of the present invention including a group of steps for physically and integrally connecting to the altered structure 60 is shown. The same reference numerals as those used so far indicate corresponding components.

The silicon oxide layer 21 and the silicon nitride layer 22 are formed on the silicon substrate 20 to form the underlayer 10. However, as shown in FIG. 5 (A), this three-layer structure or underlayer has already been formed. 10 is processed into a specific three-dimensional shape including a depression when viewed in the vertical direction, and a polycrystalline silicon layer 30 as the altered layer and a silicon nitride layer 40 as the mask layer referred to in the present invention are laminated thereon. It is formed into a shape.

With respect to such a starting structure, as also shown in FIG. 5 (A), first, a part of the uppermost silicon nitride layer 40 is removed according to a predetermined mask removal pattern, and the removed layer is the altered layer. An opening 42 exposing the polycrystalline silicon layer 30 is formed.

For the sake of convenience, the shape 42 that defines the opening 42 will be referred to as a "second shape", with the predetermined shape relating to the microstructure of the present invention being considered as the first shape.

Next, using the silicon nitride layer 40 thus patterned in the predetermined second shape as a mask, the alteration layer 30 exposed in the patterning opening 42 is planarly oxidized.

Then, the result is as shown in FIG. 5 (B), and a three-dimensional structure 60 in which the polycrystalline silicon layer which is the altered layer 30 is swelled as silicon oxide is obtained.

Although the shape of the three-dimensional structure 60 is also referred to as a second shape for convenience, the structure 60 is constructed based on the opening 42 patterned by the normal photolithography technique as described above. As a result, the dimensional order thereof is generally larger than that of the microstructure finally formed according to the present invention.

After that, as in the case of the previously-described process example, the silicon nitride layer 40 is patterned into a predetermined shape (first shape), and using this as a mask, the polycrystalline silicon layer 30, which is the altered layer, is patterned into the first shape.

However, in this case, as shown by an imaginary line 43 in FIG. 5 (B), in this embodiment, with respect to the second shape, the mask patterns for forming the first shape overlap in a plane. The feature is that

After this, the remaining silicon nitride layer 40 patterned according to the first shape is used as an oxidation mask, and the side surface 35 exposed to the outside of the altered layer 30 thereunder is selectively oxidized.

By doing this, after that, as explained with reference to FIG. 1,
Silicon nitride layer 40, the portion of the altered layer 30 that has not been altered,
Further, by selectively removing the silicon nitride layer 22 as a protection mask except under the altered portion, as shown in FIG. 5 (C), according to the present invention, a fine structure conforming to the desired first shape is obtained. 50a and 50b can be obtained.

However, as described above, in this embodiment, characteristically, for the second shape, because the overlapping patterning of the mask pattern for forming the first shape was attempted, the fine structures 50a and 50b created above were formed. And the structure 60 having a larger dimension order than that is automatically physically connected.

In such a structure, for example, in a MOSFET, the fine silicon oxide structure 50a is used as a mask for forming a fine gate and its wiring structure, and the silicon oxide layer 60 which is a relatively large structure is external to the gate. As a mask for forming a bonding pad for making a connection, the fine structure 50b can be used very effectively by using it as a mask for forming a wiring path to another circuit element, and these structures 50a, 50b and 60 can also function as an inter-wiring insulating layer after the element is completed.

6 (A) and 6 (B) show a more specific embodiment of the present invention. Similarly, the same reference numerals as those used in the above-mentioned embodiment indicate corresponding constituent elements.

FIG. 6 (A) corresponds to the step shown in FIG. 5 (C) described above, and the microstructures 50a and 50b according to the present invention are physically larger than the structure 60 having a size order larger than that. It is formed in a connected state. The difference is that the base layer 10 is
Further, between the silicon substrate 20 and the silicon oxide layer 21 directly below the protective mask 22 made of silicon nitride, which is the uppermost layer, an insulating layer 25 and a polycrystalline silicon layer 24 are sequentially arranged in the height direction from the substrate side.
Is to have.

To put it in advance, this embodiment is a preparation example for obtaining a MOS integrated circuit, in which the relatively thick portion 27 of the insulating layer 25 finally becomes a so-called field insulating film, and the thin portion 26 is the gate insulating film. Becomes

Under the condition shown in FIG. 6A, the polycrystalline silicon layer
When oxidation is performed in an oxidizing atmosphere containing oxygen or water vapor for a time such that 24 is just oxidized, under the microstructures 50a and 50b formed according to the present invention in the polycrystalline silicon layer 24. And the silicon nitride layer or the protective mask 22 left in the lower part of the structure 60 configured according to the second shape pattern serve as an oxidation resistant mask, and are shown in FIG. 6 (B). As described above, only the polycrystalline silicon layer 24 of the portion 24a corresponding to the lower portion retains the property of polycrystalline silicon as it is, and the other portions become silicon oxide.

At this time, the insulating film portion below the original polycrystalline silicon layer 24
If 26 is a suitable thin silicon oxide film, such as 100 Å or less, and if the polycrystalline silicon layer 24 contains suitable donor atoms such as high-concentration phosphorus atoms, such donor atoms will be present between them. Silicon substrate passing through the silicon oxide film 26
The phenomenon of shallow diffusion on the surface of 20 can be caused.

Therefore, if at least the surface of the silicon substrate 10 is made of p-type silicon, it has an n-type source 28 and a drain 29, and its gate is made of polycrystalline silicon, as shown in FIG. 6B. N-channel MOSFE as part 24a
As described above, T can be obtained in a self-aligned manner by adding only one oxidation step.

Obviously, the gate length and channel length of this MOSFET are not affected by the limitations of the photolithography used for patterning the first and second shapes, and can be made sufficiently short. As described with reference to the figure, the designability and reproducibility can be improved.

On the other hand, under the fine structure 50b, the line layer 24b similarly made of polycrystalline silicon having a fine width is formed and remains, so that it can be used as a signal wiring, a high resistance element, etc. The remaining polycrystalline silicon portion (not shown) is
As described above, it can be used as a connection connecting portion with another circuit, a bonding pad, or the like.

However, according to the experiment of the present invention, in particular, the field oxide film 27
If the insulating layer below the polycrystalline line layer 24b is a silicon oxide having a considerably large thickness in the portion of, the oxidizing species from the above and the oxidizing step may wrap around from below and cause the line layer 24b to disappear. I found out.

However, according to the same experiment, in such a case, the insulating layer 25
The above inconvenience can be almost completely eliminated by using at least the surface portion thereof as a chemical constitution including a nitriding component, or by arranging a silicon oxide layer thinner than the polycrystalline silicon layer 24 on the nitride film. Was also verified.

However, even if conventional wet or dry etching is applied from the state of FIG. 6 (A) regardless of the oxidation step as described above, as shown in FIG.
Alternatively, the insulating film 26 can be finely processed, and similarly, the gate structure 24a and the line or resistance structure 24b can be obtained.

Then, by diffusing or implanting an impurity of a predetermined conductivity type in the required area, a self-aligned field effect transistor having a minute gate and a minute channel length can be formed.

In the case of wet etching, the nitride film 22 acts as a mask against a buffer hydrofluoric acid aqueous solution, which is an etching solution for the silicon oxide layer 21 thereunder, and the polycrystalline silicon layer 24, a nitric hydrofluoric acid aqueous solution, etc., and dry etching, particularly In the directional method such as reactive ion etching, the fine structure 50a, 50
b acts as a mask.

FIG. 8 (A) shows a scanning electron micrograph of the main part of the electronic device produced by the embodiment shown in FIGS. 5 and 6, and FIG. 8 (B) shows each part thereof. FIG.

In the case of this preparation example, as a fine structure formed by the present invention, an oxide film having a width of about 0.5 μm and a relatively large second shape structure connected to the oxide film having a width of about 4 to 10 μm. And are formed.

Below this fine oxide film with a width of about 0.5 μm, a width of about 0.2 μm was created by the process shown in FIG. 6 (B).
An ultrafine polycrystalline silicon gate of m is formed.
The channel length of this MOSFET was about 0.1 μm.

Although some embodiments have been described above, the present invention is
As is clear from the principle, the material is not limited to the silicon-based material such as single crystal, polycrystal, and amorphous material, and any material can be selected as the altered layer. The point is that it is only necessary to be able to perform processing so that the etching rate of the altered portion is different from that of the unaltered portion. This, of course, means that there is no limitation on the means itself for causing the alteration action.

Further, in the embodiment, a laminated structure of a silicon nitride layer and a silicon oxide layer is used in the vicinity of the surface of the underlayer, but it is particularly preferable to form a silicon nitride layer having resistance to oxidation in the uppermost layer. In this case, of course, when the silicon oxide layer is not required in the subsequent process, the silicon oxide layer may be omitted and only the silicon nitride layer may be provided.

Also in the above-mentioned embodiments, as described above, the oxidation method is not limited to the thermal oxidation, and other oxidation methods such as plasma oxidation can be adopted, and it is also possible to use a boron addition treatment.

Further, regarding the patterning of the first shape, wet etching was considered in the above-mentioned embodiment, but anisotropic vapor phase etching such as reactive ion etching may be used. In particular, if the mask layer 40 and the altered layer 30 are designed to be continuously patterned together, the dimensional accuracy of the fine structure after alteration treatment such as oxidation is further enhanced, and the substrate 20 is closer to vertical. It is possible to obtain a cross section.

Finally, the present invention can be applied not only to the electronic devices as in the above-described embodiments, but also to various optical devices such as optical waveguides, various micromechanical parts of micron and submicron order, and the like.

[Brief description of drawings]

FIG. 1 is an explanatory view of one step of a method for forming a fine structure related to the present invention, FIG. 2 is an explanatory view of a possible inconvenience, and FIG. 3 is a correlation between the alteration action and the result when the alteration action depends on oxidation. Explanatory drawing, FIG. 4 is an explanatory plan view of a fine width altered structure formed by the technique of FIG. 1, and FIGS. 5 to 7 are explanatory views of other embodiments according to the present invention. FIG. 8 and FIG. 8 are a photomicrograph and an explanatory view of the surface of the main part of an actual electronic device produced according to the present invention. In the figure, 10 is a base layer, 20 is a silicon substrate, 21 is a silicon oxide layer, 22 is a silicon nitride layer which functions as a protective mask under the altered layer, 24 is a polycrystalline silicon layer, 25 is an insulating layer, 30
Is a layer to be altered, 35 is its exposed side surface, 40 is a mask layer, 5
Reference numerals 0, 50a and 50b are altered portions or fine width altered structures obtained by the present invention, and 60 is an altered structure having a second shape.

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Shunsuke Fujita 228-4-201, Kizuki Gion-cho, Nakahara-ku, Kawasaki-shi, Kanagawa Prefecture (72) Inventor, Satoshi Ogaki 2236-10-201, Atsuda, Atsugi-shi, Kanagawa Prefecture (56) References Special Kai 52-56874 (JP, A) JP 52-18181 (JP, A) JP 56-37700 (JP, A)

Claims (2)

[Claims] 1. A step of forming an altered layer and a mass layer in a height direction on an underlayer, the mask layer being partially removed according to a predetermined mask removal pattern, and the mask being removed. A step of exposing the altered layer in the exposed portion; an altered structure forming step of altering the exposed portion of the altered layer to form an altered structure; and the step of remaining around the altered structure The mask layer and the altered layer thereunder are patterned into a predetermined shape having a portion connected to the altered structure, or the altered layer is used as a mask for further patterning. A step of patterning into the predetermined shape; and the altered layer patterned into the predetermined shape,
A fine width alteration step in which the mask layer remaining on the altered layer is used as a mask for selectively altering the side surface of the altered layer to exert an alteration action only on the side surface of the altered layer; A step of removing unaltered portions of the mask layer and the altered layer after the alteration step, and leaving the altered side surface portion as a fine width altered structure connected to the altered structure; and the altered structure And a structure having a relatively large size having a shape according to the plane shape of the modified structure and the structure having a relatively large size, wherein the fine width modified structure is used as a mask. Forming a fine structure having a shape according to the planar shape of the body on the underlayer; 2. A step of forming an altered layer and a mask layer on an underlayer so as to overlap each other in a height direction; patterning the altered layer together with the mask layer into a predetermined shape, or the mask layer. Patterning the modified layer into the predetermined shape and then using the mask layer patterned into the predetermined shape as a further patterning mask to pattern the modified layer into the predetermined shape; with respect to the modified layer patterned into the predetermined shape. ,
A fine width alteration step in which the mask layer remaining on the altered layer is used as a mask for selectively altering the side surface of the altered layer to exert an alteration action only on the side surface of the altered layer; After the alteration step, a portion of the mask layer is exposed according to a predetermined mask removal pattern so that the alteration layer is exposed including at least a part of the surface of the side surface portion that has undergone the alteration action in the alteration layer. A step of removing; an altered structure forming step of altering the exposed portion of the altered layer to form an altered structure; and a step of altering the mask layer and the altered layer after the altered structure forming step. And removing the unaltered portion and leaving the altered side surface portion as a fine width altered structure connected to the altered structure; using the altered structure and the fine width altered structure as a mask, the alteration target The structure having a relatively large size according to the plane shape of the structure and the fine structure having a shape according to the plane shape of the fine width-altered structure connected to the structure having the relatively large size are described above. Forming a base layer; and a fine structure forming method comprising: