JPH11242125A - Silicon substrate and its formation - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a silicon substrate controlled in the porosity, pore diameter, distribution width of the pore diameter in the porous silicon region and the shape of the porous silicon and a method for formation thereof relating to the silicon substrate selectively having the porous silicon within the substrate and a method for formation thereof. SOLUTION: This method for formation of the silicon consists in covering the silicon substrate 1 with a partly opened mask layer 3, immersing the substrate into a chemical conversion liquid and impressing chemical conversion current thereon and anodically chemical converting a part of the silicon substrate 1 from the removed part of this mask layer 3, thereby forming the porous silicon 10 to 11 to a band form within the silicon substrate. The chemical conversion current described above is increased according to the growth degree of the porous silicon in such a manner that the current density at the boundary between the front end of the growth of the porous silicon and the silicon substrate 1 in the process of the anodically chemical conversion is made nearly constant.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a silicon substrate and a method for forming the same, and more particularly, to a method for forming a mask layer on a silicon substrate and selectively anodizing a portion where the mask layer is partially removed. The present invention relates to a silicon substrate having a porous silicon region to be formed and a silicon oxide region obtained by oxidizing the same, and a method for forming the same.

[0002]

2. Description of the Related Art Porous silicon is nanometer (n) in crystalline silicon.
m) A substance in which pores of a size are formed innumerably, for example, formed by anodizing crystalline silicon in a solution containing hydrofluoric acid. Such porous silicon has been studied for its basic preparation method and basic physical properties as a material expected to be applied to various products.

[0003] In particular, in the whole-surface formation method in which the entire surface of a silicon substrate is simultaneously formed, the concentration of hydrofluoric acid (HF) and the formation current of the formation solution are uniformly dispersed and supplied over the entire surface of the substrate, and even if the formation proceeds, the porous silicon and the formation surface are formed. Since the area of the interface with the silicon substrate is always constant, the formation conditions do not change significantly with the formation depth. Therefore, the formation conditions (HF of the formation solution)
(Concentration, formation current, formation current density, etc.), and there are many research reports as an optimal method for studying the relationship between the method of preparing porous silicon and its physical properties basically. Among them, (1) R. Herino, G. Bomchil, K. Barl
a, C. Bertrand, and JL Ginoux, et al., J. Electrochem. Soc. 134, 1994 (1987), describe the relationship between the chemical conditions and the physical properties of the porous silicon produced under full-scale chemical formation. Is described in detail.

Further, a porous silicon is selectively formed by forming a mask layer on a silicon substrate and anodizing the opening around the opening of the mask layer.
(2) P. Steiner and W. Lang, Thin Solid Films 255, 52 (1
995), (3) K. Imai and H. Unno, IEEE Trans. Electron De
vices ED31,297 (1993), (4) VP Bondarenko, AMVari
chenko, A, M. Dorofeev, NM Kazyuchits, VA Labunov,
and VFStel'makh, Tech.Phys.Lett.19,463 (1993),
(5) VPBondarenko, AMDorofeev, and NMKazuchit
s, Microelectronic Engineering 28, 447 (1995).

In these documents, anodization is carried out while maintaining a constant formation current during anodization. However,
When a mask layer is formed on a silicon substrate, a part of the mask layer is selectively removed, and a part of the silicon substrate is selectively formed starting from the portion where the mask layer is removed. However, there is a problem that the area of the interface between the porous silicon and the silicon substrate changes with the progress of chemical formation because the mask layer is concentrated in a narrow portion where the mask layer is removed.

[0006] As prior art for temporally changing the formation current during anodization, (6) PCT / M by BERGER et al.
DE 96/00913 is disclosed. Also, a paper (7) Porous silicon multilayer-optical waveguide by the inventor
s, Thin Solid Film 279,143 (1996) describes a method of creating a porous silicon in which the continuous DC current is changed stepwise at once and the porosity changes discontinuously depending on the current value. .

The prior arts (6) and (7) have high porosity (60
% Or more), the porous silicon itself is used as an optical waveguide, and the porous silicon is oxidized by utilizing the fact that the refractive index of the porous silicon depends on the porosity. However, the main purpose is to create an optical waveguide using silicon oxide in a porous state that has not been melted and densified.

In these documents, although the formation current during anodization is changed with time, the formation current is changed in a stepwise manner. The formation current is kept constant at predetermined intervals to perform anodization. Is what you do.

[0009] Further, as a reference on the subject of performing anodization by pulse current, (8) Xiao-yuan Hou, Hong-lei
Fan, Lei Xu, Fu-long Zhang, Min-quan Li, Ming-ren Yu, A
ppl.Phys.Lett. 68, 2323 (1996), and (9) LVBelyako
v, DNGoryachev, and OMSreseli, Tech.Phys.Lett.22,
97 (1996). Both literatures (8) and (9) compare the effect of a pulse current with that of a continuous constant current in anodizing an entire image of a crystalline silicon substrate. References (8) (9)
In both cases, p-type silicon of about 1 Ωcm was formed under a relatively low current density with a peak current density of about several tens mA / cm ² , and the characteristics of porous silicon as a light emitting diode were examined.

[0010] Next, the main references relating to the oxidation of porous silicon include (10) JJYon, K. Balra, R. Herino, and
J. Appl.phys. 62, 1042 (1987) by G. Bomchil, and (1
1) J. Appl. Ph by K. Balra, R. Herino, and G. Bomchil
ys. 59, 439 (1986).

What is dealt with in these articles is the oxidation of porous silicon formed on the entire surface of a silicon substrate.
It does not relate to selectively formed porous silicon. In the above papers (10) and (11), the importance of porosity control is emphasized because the contraction and expansion of the volume due to oxidation largely depend on the porosity. However, it is stated that in the entire surface formation, even if the porosity is equal to or less than the critical porosity described below, volume expansion due to oxidation affects the increase in the thickness of the silicon oxide film. On the other hand, the volume expansion due to the oxidation of the porous silicon causes a strong internal stress locally as described later. Conversely, if the porosity is too high, the silicon oxide will have a significant volume shrinkage. When applied to an optical waveguide or the like, this volume contraction makes it difficult to form the shape of the waveguide itself. The present invention makes it possible to minimize such volume expansion and contraction.

Further, none of the above-mentioned prior art documents discloses the idea of making the porosity constant over a certain region of the porous silicon selectively formed. That is, when a mask layer having an opening is formed in crystalline silicon and anodization is performed from this opening, the area of the interface between the porous silicon and the crystalline silicon increases as the formation proceeds, so that the formation current is increased. When anodization is performed while keeping the temperature constant, the interface current density at the interface between porous silicon and crystalline silicon becomes relatively smaller as the formation proceeds, and the pore diameter of the porous portion also becomes smaller as the anodization proceeds. There is a problem.

The present invention has been made in view of such a background, and a first object is to selectively form porous silicon on a silicon substrate and to improve the porosity of the porous silicon region. Another object of the present invention is to provide a silicon substrate having a porous silicon region in which the pore diameter, the distribution width of the pore diameter, and the shape of the porous silicon are controlled, and a method for forming the same.
The first method is to directly form porous silicon having a desired porosity, pore size and pore size distribution. The second method is to form porous silicon having a certain value of porosity, pore size and pore size distribution and then partially oxidize the porous silicon to substantially reduce the pore size, This is a method of forming porous silicon having a desired porosity, pore size, and pore size distribution by a combination of steps of increasing the pore size by removing the membrane.
A second object is to establish a silicon substrate in which the surface of silicon oxide formed by oxidizing a porous silicon region is substantially flush with the surface of the first silicon substrate, and a method for forming the same. .

[0014]

According to a first aspect of the present invention, there is provided a silicon substrate in which a porous silicon region is formed in a strip shape on one principal surface side. Was set to 52 to 62%.

Further, in the silicon substrate according to claim 2,
In a silicon substrate in which a silicon oxide region is formed in a band shape on one main surface side, the depth of the silicon oxide region is b, and the difference between the surface of the silicon oxide region and one main surface of the silicon substrate is c + d. When (c + d) / b is ± 6%
To be within the range.

In the method of forming a silicon substrate according to a fourth aspect of the present invention, the silicon substrate is covered with a partially open mask layer, immersed in a chemical conversion solution, a formation current is applied, and the mask layer is removed from the removed portion. In a method for forming a silicon substrate in a band shape in a porous silicon region by anodizing a part of the crystalline silicon substrate, the growth of the porous silicon during the anodizing step may be performed. The formation current is changed in accordance with the degree of growth of the porous silicon so that the current density at the interface between the tip and the silicon substrate is substantially constant.

According to a tenth aspect of the present invention, in the method of forming a silicon substrate, the silicon substrate is covered with a partially open mask layer, immersed in a chemical conversion solution, a formation current is applied, and the mask layer is removed from the removed portion. In the method of forming a silicon substrate in which a porous silicon region is formed in a part of the silicon substrate by anodizing the silicon substrate, an oxide film is formed on a surface of the porous silicon region after the porous silicon region is formed.

[0018]

Embodiments of the present invention will be described below with reference to the accompanying drawings. The basic characteristics of the porous silicon material are determined by the porosity (P), the pore size (R), and the distribution width of the pore size. Here, the porosity (P) is defined as a volume ratio of a void portion in the entire volume of the porous silicon. Further, the diameter of the pore is defined as the pore diameter (R), and the distribution width is defined as the distribution width of the pore diameter.

The porosity, the pore diameter, and the distribution width of the pore diameter of the porous silicon depend on the doping characteristics of the silicon substrate used, the HF concentration of the chemical conversion solution, and the interface current density. When a silicon substrate heavily doped with p-type is used, the porosity, the pore diameter, and the distribution width change as shown in FIG. In FIG. 10, ■ indicates the center value of the pore diameter distribution, and ▲ and ● indicate the half width of the distribution. The upper line of the comment inside the rectangle with rounded corners indicates the HF concentration of the chemical conversion solution, and the lower current value indicates the current density at the interface between porous silicon and crystalline silicon.

FIG. 10 is a graph showing the relationship between the formation conditions and the properties of porous silicon under the entire formation conditions, based on the individual data described in the above-mentioned article (1). It is uniquely configured and indicates the following four important points of porous silicon. First, increasing the interfacial current density increases both porosity and pore size. Second, when the HF concentration of the chemical conversion solution is increased, both the porosity and the pore size decrease. Third, when the HF concentration of the chemical conversion liquid is increased, the distribution width of the pore diameter becomes narrow. In particular, when a high-concentration HF chemical solution is used, the distribution width of the pore diameter becomes very narrow, and the pore diameter becomes very uniform and uniform. Fourth, by selecting the formation conditions of the HF concentration and the interfacial current density of the chemical conversion liquid, it is possible to form porous silicon having any porosity, pore size, and pore size distribution.

The concentration of hydrofluoric acid in the chemical solution used in the present invention is from the usual concentration conventionally used to about 50%, which is the maximum concentration currently supplied industrially, and the higher concentration of hydrofluoric acid. If the acid can be supplied, it is of course contained up to its maximum concentration. Also, the interface current density is tens of amperes /
Includes high current densities of cm ² .

Further, when such a high-concentration HF solution is used, the smoothness of the interface between the porous silicon and the crystalline silicon is improved as compared with the case where the low-concentration HF is used.

When the porous silicon in the silicon substrate of the present invention is oxidized and used for a single mode optical waveguide, the mask opening width w is desirably 20 μm or less. There is no particular condition to limit the minimum value of the mask opening width w.
Any size may be used as long as it has good industrial reproducibility by ordinary fine processing. If the mask opening width w is larger than 20 μm, it becomes difficult to set conditions for single mode propagation. However, the multimode optical waveguide has no upper limit of the mask opening width w.

Similarly, when used as a single mode optical waveguide, the formation depth r is desirably 50 μm or less. Within this range, the function as an optical waveguide can be sufficiently satisfied. This is because if the formation depth r becomes larger than this, the processing area becomes unnecessarily large. However, the multimode optical waveguide does not have this limitation.

The limits of the mask opening width w and the formation depth r are set based on the necessary and sufficient conditions for the optical waveguide when using porous silicon for the optical waveguide. When applied to other fields, depending on the requirements of the application,
These sizes may be freely determined.

FIG. 1 is a view showing a method for forming porous silicon according to claim 4. FIG. 1A is a perspective cross-sectional view, FIG. 1B is a diagram showing a method for controlling the current by changing the current over time, and FIG. 1C is a graph showing the porosity and pore diameter of the formed porous silicon. It is a figure which shows the formation depth dependence. The silicon substrate 1 used in the present invention has a hole concentration of 1 × 10 ¹⁸ cm.
A p-type substrate having a resistivity of about ³ or more, that is, a resistivity of 0.1 Ωcm or less is desirable. However, the condition of the present invention is basically satisfied if the substrate is one in which valence electrons are uniformly supplied to the interface between the porous silicon and the crystalline silicon during the formation. In FIG. 1A, a mask layer 3 made of silicon nitride (SiN _x ) or the like is formed on one main surface 2, which is the first surface of a silicon substrate 1, and the mask layer 3 is elongated in a one-dimensional direction. The opening 4 having a width w is formed by etching or the like. When the substrate thus prepared is used as an anode and anodized in an electrolytic solution containing hydrofluoric acid, the formation depth (radius of the porous region) r is substantially isotropic starting from the end of the mask layer 3. As a result, the porous silicon region 5 grows. Therefore, at a certain point in the growth process, the length L on the cross section of the interface between the porous silicon 5 and the crystalline silicon 1 is approximately L = πr + w.
...
(1) is represented. As is clear from FIG. 1A, the interface current density J at the interface between the porous silicon 5 and the crystalline silicon 1 is J = I / L = I / (πr + w) (2). Here, I is a current per unit length that flows intensively in the mask opening width w.

When the formation proceeds with the current I per unit length being constant, the interface current density J between the porous silicon 5 and the crystalline silicon 1 decreases according to the equation (2). In a region where the formation depth r in the initial formation is smaller than the mask opening width w, the interface current density J is large, but under the condition that the formation depth r is larger than the mask opening width w, ie, L >> w As shown in a comparative example described later, the current density J is significantly smaller than in the initial stage of chemical formation.

On the other hand, in order to make the current density at this interface constant, I = J * L (3) so that the formation current is changed to the interface between the porous silicon 5 and the crystalline silicon 1. It is necessary to increase the area in proportion to the area increase and to control the formation current as a function of time so that I = f (t). A control example of the formation current is schematically shown by a solid line 6 in FIG. A dotted line 7 indicating a constant value in FIG. 1B indicates a controlled interface current density which is controlled in this way.

As shown in equation (3), the current density J at the interface between the porous silicon 5 and the crystalline silicon 1 is kept constant, and is proportional to the increase in the area of the interface between the porous silicon 5 and the crystalline silicon 1. Control to increase the formation current I. When the interface current density is controlled to be constant as described above, the formed porous silicon 5 has both a porosity (P) 8 shown by a solid line and a pore diameter (R) 9 shown by a dotted line in FIG. It has a constant value independent of the depth r.

FIG. 2 is a view showing one embodiment of a method for forming porous silicon according to claim 5. As shown in FIG. 2A, a first porous silicon region 10 having a predetermined pore diameter and a second porous silicon region 11 having a different pore diameter from the first porous silicon region 10. To form At that time, the concentrations of hydrofluoric acid in the electrolytic solution (broken lines) 12 and 13, the current values (solid lines) 14 and 15, and the current densities (dotted lines) 16 and 17
Is shown in FIG. 2 (b). To form the first region 10, an electrolytic solution having a first hydrofluoric acid concentration 12 is used, and a first current value 14 is used. After that, a second porous silicon region 11 is formed with a second current value 15 using an electrolytic solution having a second hydrofluoric acid concentration of 13. In this case, the second hydrofluoric acid concentration 13 is set higher than the first hydrofluoric acid concentration 12, and the second current density 17 is higher than the first current density 16 in accordance with the hydrofluoric acid concentration difference. Then, as shown in FIG. 2C, the porosity (solid line) 20 and 21 of the first and second porous silicon are made substantially equal (P1 = P2), and the first porous silicon region 10 is formed. Can be made larger than the pore diameter 23 of the second porous silicon region 11.

When the formation current is controlled as shown in FIGS. 1 and 2, even if the current density J at the interface between the porous silicon and the crystalline silicon is about several tens mA / cm ² , a mask having a width w is obtained. In the opening 4, a concentrated current of several tens to several hundred times flows as shown in the equation (3). On the other hand, a large amount of gas is generated from the interface between porous silicon and silicon due to the formation current. The path for releasing the large amount of gas is limited to the opening 4 having the width w. Therefore, when the formation current is simply controlled by a continuous DC current, the mask layer 2 is broken by the pressure of a large amount of generated gas, and cracks and peeling occur in the mask. Once the mask layer 2 is damaged or peeled, the intended purpose of selectively forming only a desired portion cannot be achieved. Further, since the supply of the electrolytic solution containing hydrofluoric acid is also restricted to the opening 4 having a small width w, there arises a problem that it is difficult to keep the concentration of the chemical solution constant.

In order to avoid these problems, FIG.
As shown in (a), the effective peak current value I of the unit pulse
p, a pulse current having the same duration tp, and a repetition period T are used. In this case, the effective leading current value Ip of the pulse is calculated by the equation (2), (3), or the current value 14, in FIG.
Control is performed so as to satisfy conditions such as fifteen. Formation is performed by this pulse current, and a high current density at the formation interface can be maintained. On the other hand, the gas generated when the pulse current is applied can be released from the mask opening 4 having the width w even during the pause time of the pulse. In addition, a new electrolytic solution is also supplied to the formation interface with the desorption of gas. In this way, the pulse current value Ip, the duration tp of the unit pulse, and the repetition period T are controlled, and the average current value Ip * tp / T per unit time is set to the mask layer 2.
Is controlled so as not to peel or break. Further, the current value during the pulse pause period (T-tp) does not necessarily need to be controlled to zero, and a negative current may flow.

FIG. 3B shows a pulse current value, an application time width,
The conceptual relationship of the pause time width is schematically shown with respect to the initial, middle and end stages of formation. When the formation current is small, FIG.
The current 14 in (b) is a continuous DC current, and the composite control may be performed such as using a pulse-controlled current as the current 15 when the formation current value becomes large.

The porous silicon prepared as described above is oxidized in, for example, a humid oxygen stream to obtain silicon oxide.

The porosity of the porous silicon is from 52% to 62%.
% Range. The lower limit of 52% is to reduce the volume expansion of non-doped silicon oxide to half or less of the conventional example. The upper limit of 62% is the critical porosity when silicon oxide is doped with about 5 mol% of impurities.

By controlling the porosity as described below, the volume of silicon dioxide can be controlled so that it does not shrink or expand compared to the volume of the initial porous silicon. Therefore, as shown in FIG. 4, the surface 26 of the silicon oxide 25 can be controlled so as to be flush with the surface 2 of the first silicon substrate 1.

As described above, the volume expansion due to the oxidation of silicon fills the volume of the pores of the porous silicon in a complementary manner, whereby the porous silicon region can be changed to the oxide without changing the shape of the initial porous silicon region. . That is, the conditions for this are (a) the volume VSi of 1 mol of silicon is VSi = ZSi / ρSi (5) (b) When this silicon is converted into porous silicon having a constant porosity P, The existing silicon is (1-
P) mole.

(C) When the above (1-P) mole of silicon is oxidized and melted, the volume of the oxide becomes: VSiO ₂ = (1-P) * ZSiO ₂ / ρSiO ₂ (6) (d) If the volumes of (5) and (6) are equalized, that is, if the porosity with zero volume reaction after oxidation is defined as the critical porosity Po, Po = 1− (ρSiO ₂ / ρSi) * (ZSi / ZSiO ₂ ) = 1−A (7) A = (ρSiO ₂ / ρSi) * (ZSi / ZSiO ₂ ) (8) Where ρ is the specific gravity, Z is the molar equivalent, and the subscript Si
And SiO ₂ represent silicon and a silicon oxide film, respectively. By substituting the values of ZSi = 28.0855 g / mol ZSiO ₂ = 60.0843 g / mol ρSi = 2.3291 g / cm ³ ρSiO ₂ = 2.24 g / cm ³ as the physical property values of the above formula (8), The critical porosity Po at which the volume change becomes zero is Po = 55.0%.

Further, when the porosity changes from Po to ΔP, the volume change rate after oxidation ΔVSiO ₂ is ΔVSiO ₂ = − {P * (ZSiO ₂ / ZSi) / (ρSi / ρSiO ₂ ) = −}. P / A = −2.224 ° P (9), and when the porosity changes by 1%, ΔVSiO
₂ = −2.224%.

FIG. 5 shows the relationship between the porosity P of the porous silicon and the volume change rate of SiO ₂ after oxidization of the porosity P. FIG. 5A shows a wide range of porosity from 49% to 61%, and FIG. 5B shows a porosity of 53% to 57%.
Particularly narrow ranges up to% are shown enlarged. FIG. 5 (a)
If ΔP is positive as in (b), volume contraction occurs, and ΔP
If is negative, volume expansion occurs. The coefficient is-△ P / A as described above.

Porous silicon was formed in the same manner as in the embodiment shown in FIGS. However, W = 6 μm and r = 27 μm were adopted as examples of practical parameters. FIG. 6A shows the coordinates of the cross-sectional shape including the silicon oxide 30 after oxidizing and melting the porous silicon. From the intersection of the center of the mask opening width w and the first surface of the silicon substrate 1, the silicon substrate 1
Defines the coordinate -y. Further, coordinates x are defined along the substrate surface 2 from the center of the mask opening width w.

SiO after oxidizing the porous silicon region
FIGS. 6B and 6C show the ΔP dependency of the interface between SiO ₂ and silicon and the surface shape of SiO ₂ . FIG. 6B shows SiO 2
Indicates ₂ and silicon interface 31 overall shape, FIG. 6 (c) shows in detail with particular extracts the shape of the SiO ₂ surface. FIG.
The notes on the curve in (c) show ΔP in%.

As is apparent from these figures, in particular, ΔP
= 0% (P = Po), as shown in the embodiment in FIG. 4, the SiO ₂ surface is the first surface 2 of the silicon substrate.
Is the same plane. FIG. 6 shows an example of the interface shape accompanying volume expansion when ΔP is negative and −1%, −3%, and −5%. Further, an example of the interface shape when ΔP is positive and has + 2%, + 4%, and + 6% and shows volume contraction is also shown. Table 1 shows the y coordinate value of the surface at the center position of the mask opening width w.

[0044]

[Table 1]

As shown in Table 1, if ΔP = ± 6% or less, the displacement is about 3.
Only 6 μm. At ΔP = ± 3%, the maximum displacement is about 1.8 μm. These values are much smaller than those of a comparative example described later.

The above numerical calculation can be performed as follows. First, the shape of the interface (31) between the silicon oxide region and silicon can be substantially expressed by the following two equations.

Y = − (r 2 − (x−w / 2) ² ) ^1/2 [x] ≧ w / 2 (10) y = −r [x] ≦ w / 2 (11) Here, x and y are the coordinates on the position in FIG.
r indicates a radius at the time of forming porous silicon, w indicates a mask opening width, and [x] indicates an absolute value of x.

Next, it is assumed that volume fluctuations such as expansion and contraction due to oxidation are avoided by raising and lowering the surface position of silicon oxide which is a free surface. Thus, the surface position y at an arbitrary x coordinate is expressed by the following equation.

Y = − (r 2 − (x−w / 2) ² ) ^1/2 * △ P / A [x] ≧ w / 2 (12) y = −r * △ P / A [x ] ≦ w / 2 (13) The surface shape shown in FIG. 6C is calculated by Expressions (12) and (13).

The pore size 22 of the first porous silicon region 10 formed as described above is larger than the pore size 23 of the second porous silicon region 11 as shown in FIG. Is controlled. By immersing the silicon substrate 1 on which the porous silicon is formed as described above in, for example, a liquid of an organometallic compound containing a metal element, the organometallic compound is selectively doped into the porous silicon region 10 having a large pore diameter. can do. As the organometallic compound containing a metal element, for example, aluminum (A
l), boron (B), barium (Ba), bismuth (B
i), calcium (Ca), cadmium (Cd), cerium (Ce), cesium (Cs), disprosium (D
y), erbium (Er), europium (Eu), gadolinium (Gd), germanium (Ge), hafnium (Hf), holmium (Ho), indium (I
n), lanthanum (La), lutetium (Lu), magnesium (Mg), niobium (Nb), neodymium (N
d), neighbor (P), promethium (Pm), praseodymium (Pr), rubidium (Rb), antimony (S
b), samarium (Sm), tin (Sn), strontium (Sr), tantalum (Ta), terbium (Tb),
Titanium (Ti), thallium (Tl), thulium (T
m), yttrium (Y), ytterbium (Yb),
Tungsten (W), zinc (Zn), zirconium (Z
There are compounds in the group r). Among them, for example, boron (B) has an effect of reducing the refractive index in silicon oxide. Also, zirconium (Zr) and titanium (Ti)
Increases the refractive index. Further, for example, erbium (E
Rare earth elements such as r) are optically active impurities in silicon oxide and have an effect such as optical amplification. As will be described later, when used in an optical waveguide, the element for reducing the refractive index may be doped in both the porous silicon regions 10 and 11 of FIG. On the other hand, the element for increasing the refractive index is selectively doped only in the region 10 of the porous silicon. The optically active element may be selectively doped only in a first partial region (not shown) of the region 10. When the element having a plurality of functions is selectively doped as described above, the size of the organometallic compound molecule of the element having the effect of reducing the refractive index is minimized, and the compound molecule for increasing the refractive index is increased next. Furthermore, if the compound molecule of the optically active element is maximized, selective doping can be performed according to the pore diameter of each region of the porous silicon. Here, it is possible to change the size of the various organometallic compound molecules used for doping by using a conventional technique as a subject of molecular design. Therefore, selective doping of an element having various functions according to the combination of the pore diameter and the molecular size becomes possible.

The substrate is immersed in a liquid of an organometallic compound containing one or more of these compounds, and then the organic liquid on the surface is removed and oxidized in, for example, a wet oxygen stream.

FIG. 7 is a perspective cross-sectional view after immersion in, for example, an organic material of titanium as an organometallic compound, removing the organic material on the surface, oxidizing at 1150 ° C. in a humid oxygen atmosphere, and further removing the mask layer on the surface. (A). A silicon oxide region 32 doped with titanium and a silicon oxide region 33 having a relatively low titanium concentration are formed in the silicon substrate 1. After oxidation, the regions 32 and 33 do not contract or expand in volume, and their surfaces are flush with the surface 2 of the first silicon substrate 1.

The refractive index of the silicon oxide region 32 is higher than that of the region 33 because titanium is doped. Therefore, an optical waveguide having the low-refractive-index region 33 as a lower clad and the high-refractive-index region 32 as a core can be formed. FIG. 7A shows an optical waveguide having such a configuration. FIG. 7B shows an optical waveguide on which an upper cladding layer 34 is formed as necessary. The optical waveguide structure shown in FIG. 7A is formed under the conditions of w = 0.75 to 6 μm and r = 27 μm, a 1.55 μm laser beam is made incident from one end, and the Near Field Pattern is emitted from the other end. (NFP) was measured.

FIG. 8 shows an example of an NFP of an optical waveguide formed at w = 6 μm. FIG. 8A shows a light intensity distribution in a direction parallel to the substrate surface, FIG. 8B shows a light intensity distribution in a vertical direction, and FIG. 8C shows an equal intensity distribution curve. FIG.
(A) shows very good symmetry in a direction parallel to the substrate surface. On the other hand, it shows asymmetry in the direction perpendicular to the substrate surface,
This is due to a sudden change in the refractive index because the used optical waveguide has the configuration shown in FIG. 7A and does not have a lower clad layer deposited thereon. As is clear from FIG. 8, the transmitted light has its intensity concentrated on a single peak and propagates in a single mode. Mask opening width w is 4-6μ
At m, the NFP had a single peak as shown in FIG. 8, and it was confirmed that light propagated in a single mode. However, the spot size of the transmitted light increased as the mask opening width w decreased.

Mask opening width w of spot size of transmitted light
Is shown in FIG. As shown in FIG. 9, when the mask opening width w was 4 μm to 6 μm, the transmitted light was in the single mode. If it is less than 3 μm, the width of the NFP increases sharply with a decrease in the mask opening width w, and the peak of the emitted light is divided into a plurality, indicating that light propagates in a multimode.

Porous silicon itself is an important device material. This material is also important as an intermediate material for forming silicon oxide. The embodiment shown in FIGS. 4 and 5 has described the implementation method in which the volume fluctuation due to the oxidation of the porous silicon is minimized.

In this embodiment, an embodiment will be described in which porous silicon is doped with an impurity and the volume fluctuation when the silicon is oxidized is minimized. In the optical waveguide shown in FIG. 7, the first silicon oxide region 32 is doped with titanium to increase the refractive index. On the other hand, the second region 33 was used as cladding without being doped. When silicon oxide is used as the optical waveguide in this manner, the refractive index difference between the core and the clad depends on the design conditions of the waveguide, but a waveguide having a high refractive index difference uses a relative refractive index difference of about 2%. . FIG.
(A) shows a known relationship between the concentration and the refractive index when various impurities are added to quartz glass. FIG. 16B shows a similar view when the refractive index of quartz glass is normalized to 1. In order to give a refractive index difference of 2%, ZrO
₂ about 5%, TiO ₂ about 6%, Al ₂ O ₃ and GeO ₂
Requires even higher concentrations.

In the case where oxidation is performed after doping with impurities, the increase in volume due to the oxide of the added impurity element must be taken into consideration in forming porous silicon.

The discussion at this time is similar to the embodiment shown in FIGS. 4 and 5 described above. (A) The volume VS of one mole of silicon
i is VSi = ZSi / ρSi (5) (b) When this silicon is converted into porous silicon having a porosity P, the amount of silicon existing in the volume becomes (1-P) mol, (c) When the above (1-P) mole of silicon is oxidized and melted, the volume of the oxide becomes: VSiO ₂ = (1-P) * ZSiO ₂ / ρSiO ₂ (6) (d) Silicon oxide (1-P) The volume of the oxidized impurity added x mole to the mole is Vimp = x / (1-P) * Zimp / ρimp (14) (e) The sum of the volumes of (6) and (14) is (5) to be equal to the volume of, ZSi / ρSi = (1- P) ZSiO 2 / ρSiO 2 + (x / (1-P) * Zimp / ρimp) ··· (15) i.e., an impurity When the critical porosity at which the volume change becomes zero after oxidation and oxidation is Pimp Pimp = 1− (A / 2 + (A ² / 4−x * B)) ^1/2 (16) Where A = (ZSi / ZSiO ₂ ) * (ρSiO ₂ / ρSi) (8) B = (Zimp / ZSiO ₂ ) * (ρSiO ₂ / ρimp) x is the mole of the added oxide impurity element to silicon oxide Ratio.

Here, when titanium oxide (TiO ₂ ) is used as an example of the added impurity and the density of rutile type crystal is used as the specific gravity ρ, ρimp = ρTiO ₂ = 4.23 g / cm ³ Zimp = ZTiO ₂ = 79.8788 g / Mol.

FIG. 13 shows the critical porosity in the above equation (16) as a function of x. The abscissa x at the 13 shows the molar ratio of titanium oxide to silicon oxide _{(SiO 2) (TiO 2)} , the vertical axis represents the critical porosity when the impurity of the concentration doped.

FIG. 13A shows the range of 0 to 5% as x, and FIG. 13B shows the range of 0 to 2% in detail. As shown in FIG. 13, the critical porosity also increases from 55% to 58% when TiO ₂ is added at about 2 mol%. 5
When mol% is added, it increases to 65%.

An example of selective formation for doping impurities at a high concentration as discussed in the embodiment shown in FIG. 13 will be described. First, as shown in FIG. 14, as in the embodiment shown in FIG. 2, a mask layer 3 was formed on a silicon substrate 1, and an opening 4 was further formed. After that, the first porous silicon 61 was formed under the first chemical condition. However, as shown in FIG. 14B, the current 14 and the current density 1 in FIG.
A current 64 and current density 66 greater than 6 were employed. Second
The same HF concentration, current value and current density as described in the second embodiment were used for the formation of porous silicon. As a result, the porosity was increased by the volume occupied by impurities doped with the porosity 67 of the first porous silicon. In this case, the pore diameter of the first porous silicon also slightly increased.

A series of porous silicon prepared in this way were doped with titanium by 2 to 5% and then oxidized to observe the surface shape. Fig. 11 shows the surface of two types of silicon oxide.
As shown in (a), the substrate could be held substantially in the same plane as the first silicon substrate surface.

FIG. 15 shows the basic concept of this embodiment.
In forming porous silicon for forming a silicon oxide region to be doped with impurities, the porosity is set large in advance, and the critical porosity of each of doped silicon oxide and undoped silicon oxide is used. This is the biggest difference from the discussion of FIG.

[0066]

COMPARATIVE EXAMPLE As in FIG. 1A, a mask layer 3 of a silicon nitride film is formed on a silicon substrate 1 having a specific resistance of 0.01 Ωcm and doped with B at a high concentration. The opening 4 of w was formed. Thereafter, a region of porous silicon was formed while keeping the formation current constant. However, the current value and the concentration of the hydrofluoric acid electrolyte were selected so that the average porosity of the entire region of the porous silicon was 55%. Since the formation current was kept constant, the interface current density at the interface between the porous silicon and the crystalline silicon decreased with the formation time as shown in FIG. 11B (w = 2 μm, w in FIG. 11B).
= 6 μm and w = 15 μm, and the initial interface current density is normalized to 1). Therefore, the porosity 53 of the porous silicon region in this case decreased with the formation depth as shown in FIG. In the initial stage of chemical formation, the interface current density at the interface between porous silicon and crystalline silicon is very high, but the current density decreases as the formation proceeds.
At the end of the formation, the interface current density was significantly lower than the initial value (r = 30 in the example of FIG. 11B).
μm, the interface current density was 2
%, 6%, and 14%). As a result, as is implied in FIG. 10, the porosity is high at the early stage of formation and low at the end of formation. This porous silicon was oxidized and melted in a moist oxygen atmosphere.

FIG. 11A shows a typical example of the cross-sectional shape after oxidizing the porous silicon formed as described above.
As shown in FIG. 11A, volume shrinkage was observed in the central portion 55 of the oxidized region, and the vicinity
In No. 6, the volume expansion of the oxide was remarkably observed. This result indicates that compressive stress acts on the oxide near the interface and conversely tensile stress acts locally on the silicon in contact with the interface. When such a cross section is left at room temperature for about one month and observed with a microscope, FIG.
At a high frequency, chipping due to minute chipping was observed at the interface portion 56 between silicon oxide and silicon, which is projected on the surface of (a). In other words, this indicates that the interface between the porous silicon and the silicon locally contains a large stress.

FIG. 12 shows a typical example of the cross-sectional shape after the SiN _x mask layer is removed from the surface of FIG. 11A by reactive etching (RIE). Table 2 shows the results of measuring the dimensions b, c, and d on a scanning electron micrograph of each sample.

[0069]

[Table 2]

Table 2 shows that the mask opening width w is 6.0 μm.
m, 0.75 μm. The total width a is distributed in a considerably wide range (about 52 to 61 μm). On the other hand, a / b
Is a value slightly higher than 2. Looking further, (a
The value of -w) / b was very close to 2.0, indicating that the formation of porous silicon was isotropic even when experimental errors were taken into account.

On the other hand, the sum of expansion and contraction is (c + d).
Is about 7 μm, and the value of (c + d) / b has an average value of about 27%, indicating that volume contraction and expansion coexist and each of them is extremely large. In comparison,
In the embodiment shown in FIG. 5, the value of the contraction or expansion at ΔP = ± 6% is 3.6 μm, and the formation depth ratio is 13%, which is about の of the value in this comparative example. 4 shows the effectiveness of volume change control after oxidation by porosity control according to the method of the present invention.

Next, another desirable control method of the porosity and pore diameter of the porous silicon region will be described with reference to FIG. FIG. 17 shows the relationship between the ratio of the oxidized porous silicon and the oxidation temperature when the porous silicon is oxidized in a dry oxygen atmosphere for a sufficiently long time (1 hour or more). 300
In the treatment at a temperature of 1000 ° C. (1000 / T = 1.74), almost monomolecular silicon oxide grows on the surface of the microstructure of porous silicon. At this time, the ratio of the oxidized porous silicon is about 15%. In the treatment at 350 ° C. (1000 / T = 1.6), about 30% of silicon is oxidized, and almost two molecular layers of silicon oxide grow. Further, almost all the amount is oxidized by the treatment at 800 ° C. At the intermediate temperature, the oxidation rate is represented by a reciprocal scale of the absolute temperature as shown in the figure. When silicon oxide grows in a monolayer, the substantial pore diameter decreases by about 0.5 nm due to volume expansion due to oxidation. That is, the pore diameter after oxidation can be reduced according to the degree of oxidation.

FIG. 18A schematically shows the fine structure of the porous silicon region. In porous silicon, innumerable pores whose cross-sectional dimensions are substantially the same as those of a silicon column having a nm scale are alternately present. Reference numeral 45 in FIG. 18 indicates this silicon pillar, and the diameter of the pores therebetween is indicated by R1i. When such porous silicon is oxidized at a low temperature in a dry oxygen atmosphere as shown in FIG. 17, the surface of the silicon pillar 45 is oxidized and an oxide film 47 grows as shown in FIG. In the silicon pillar 45, the ratio of the portion consumed as the oxide film 47 and the portion 46 remaining as silicon depends on the oxidation temperature as shown in FIG. When silicon is oxidized, its volume increases about twice as much as before. Therefore,
The sum of the volumes of the remaining silicon pillar 46 and the oxide film 47 increases. As a result, the pore diameter R1o of the partially oxidized porous silicon is substantially reduced. By partially oxidizing the porous silicon in this manner, the pore diameter can be controlled to a value smaller than the initial value.

Further, as shown in FIG. 18B, a portion of silicon oxide may be removed by etching from a partially oxidized porous silicon with a solution containing a small amount of HF. In this case, the remaining portion 46 of the silicon pillar 45 becomes thinner than the initial portion 45. As a result, the substantial pore diameter R2i is equal to the initial pore diameter R2.
1i and the porosity also increases. In this case, the rate of increase in pore diameter and porosity can be controlled by the oxidation rate of partial oxidation.

As described above, by combining the partial oxidation and the removal of the oxide film, it is possible to form porous silicon having a porosity / pore diameter / pore diameter distribution directly formed by chemical formation. In particular, the desired porosity
When forming porous silicon having a pore diameter / pore diameter distribution, a high hydrofluoric acid concentration and a large current density may be required. In such a case, porous silicon is once formed with a high hydrofluoric acid concentration and a small current density, and then a porous silicon having a desired porosity, pore diameter, and pore diameter distribution is obtained by a combination of the above-described partial oxidation and oxide film removal steps. Can be converted to

Although it depends on the treatment method when removing the oxide film with dilute HF, when drying the porous silicon wet with the liquid, reduce the influence of the surface tension of the liquid and take care to prevent the destruction of the porous silicon structure. In this case, the pore size distribution width is particularly small,
If it exceeds 5%, porous silicon having very high porosity and uniform pore diameter can be obtained.

[0077]

As described above, according to the silicon substrate of the first aspect, the porosity of the band-shaped porous silicon region on one main surface side is set to 52 to 62%. When an optical waveguide is formed using high quality silicon, the upper surface of the core and the cladding surrounding the core can be positioned substantially in the same plane as the first surface of the silicon substrate. The silicon substrate with the core surface of the optical waveguide on the same plane as the first surface of the silicon substrate is used as an optical module substrate for mounting optical fibers and other optical elements, and for optical modules made by conventional thick film deposition and etching methods. This is extremely advantageous.

Further, according to the silicon substrate of the second aspect, the depth of the silicon oxide region formed in a belt shape on one main surface side is b, and the depth of the silicon oxide region between the surface portion of the silicon oxide region and the one main surface of the silicon substrate is When the difference is c + d, (c + d) / b is ±
Since it is within the range of 6%, it is extremely advantageous for mounting optical fibers and other optical elements as an optical module substrate, as compared with an optical module formed by a conventional thick film deposition and etching method. It can significantly reduce the local stress included in the interface between silicon oxide and crystalline silicon, and can create a device that is stable in reliability and in which silicon oxide and crystalline silicon are in contact, especially an optical waveguide. Become.

Further, according to the method for forming a silicon substrate according to the fourth aspect, the porous silicon is formed such that the current density at the interface between the growth tip of the porous silicon and the silicon substrate during the anodizing step is substantially constant. Since the formation current is increased in accordance with the growth degree of porous silicon, porous silicon with controlled porosity and pore diameter can be selectively formed in the silicon crystal. Further, it is possible to form multiple layers of porous silicon having the same porosity and different pore diameters, so that one of the multiple layers of porous silicon can be selectively doped with impurities. become. Further, it is possible to form a multilayer porous silicon having a preset porosity and pore diameter. According to this technique, porous silicon having a pore diameter and a porosity set according to an impurity substance to be treated and its doping concentration is formed, and one of these porous silicon layers is selectively doped with an impurity. A method is introduced to introduce critical porosity that minimizes volume fluctuation after oxidation of porous silicon. An optical waveguide is created using this method, and the upper surface of the core and the cladding surrounding the core can be positioned flush with the first surface of the silicon substrate. The silicon substrate with the core surface of the optical waveguide on the same plane as the first surface of the silicon substrate is used as an optical module substrate for mounting optical fibers and other optical elements, and for optical modules made by conventional thick film deposition and etching methods. It is advantageous in comparison. In addition, by using the method of the present invention, local stress included in the interface between silicon oxide and crystalline silicon can be significantly reduced, and a device in which silicon oxide and crystalline silicon are stable in terms of reliability, Above all, it is possible to create an optical waveguide.

According to the method of forming a silicon substrate according to the tenth aspect, after forming a porous silicon region in the silicon substrate by anodizing treatment, an oxide film is formed on the surface of the porous silicon region. Can form a porous silicon region further reduced immediately after chemical formation. Further, by etching an oxide film formed on the surface of the porous silicon region, porous silicon having increased porosity and pore diameter can be formed. By the combination of the partial oxidation and the oxide film removing step, it is possible to provide a second method for forming porous silicon having desired porosity, pore diameter, and pore diameter distribution.

[Brief description of the drawings]

1A and 1B are views showing a method of forming a silicon substrate according to claim 4, wherein FIG. 1A is a perspective sectional view, FIG. 1B is a view showing the formation current and the same current density, and FIG. It is a figure which shows the formation depth dependence of a pore diameter.

FIGS. 2A and 2B are diagrams showing a method of forming a silicon substrate according to claim 5, wherein FIG. 2A is a cross-sectional view, FIG. It is a figure which shows the formation depth dependence of a pore diameter.

FIG. 3 is a diagram showing an application form of a pulse current in the method for forming a silicon substrate according to claim 4 or 5;

FIG. 4 is a view showing a cross-sectional shape after oxidizing a porous silicon region formed by the method of forming a silicon substrate according to claim 4;

FIG. 5 is a diagram showing the relationship between the porosity of porous silicon and the volume fluctuation after porous silicon oxidation.

FIG. 6 is a diagram showing the porosity of porous silicon and the surface shape of silicon oxide.

FIG. 7 is a view showing a basic structure of an optical waveguide formed using the silicon substrate according to claim 2;

FIG. 8 is a diagram illustrating an example of an NFP of light transmitted through a waveguide.

FIG. 9 is a diagram showing an aperture width dependency of a spot size.

FIG. 10 is a view showing a basic concept of a method for forming a silicon substrate according to claims 1 and 2;

FIG. 11 is a diagram showing a comparative example.

FIG. 12 is a view showing a cross-sectional shape after porous silicon oxidation according to a comparative example.

FIG. 13 is a graph showing the dependency of the critical porosity on the concentration of added impurities.

FIG. 14 is a diagram illustrating a form of controlling chemical conversion conditions at the time of adding an impurity.

FIG. 15 is a diagram showing a relationship between porosity and pore diameter assuming addition of impurities.

16A is a diagram showing the relationship between the type and addition amount of impurities and the refractive index, and FIG. 16B is a diagram showing the relationship between the type and addition amount of impurities and the relative refractive index.

FIG. 17 is a diagram showing the relationship between the oxidation rate of porous silicon and the oxidation temperature.

FIG. 18 is a diagram schematically showing partial oxidation of porous silicon.

[Explanation of symbols]

1: silicon substrate, 2: first surface of silicon substrate,
3: Mask layer, 4: Mask opening, 5: Porous silicon region, 6: One form of current control, 7: Si / porous silicon interface current density, 8: Porosity, 9: Pore diameter, 10: No. 1
Porous silicon region, 11: second porous silicon region, 12: non-acid concentration of the first chemical conversion liquid, 13: hydrofluoric acid concentration of the second chemical liquid, 14: first current control, 15: Second current control

Claims (11)

[Claims] 1. A silicon substrate in which a porous silicon region is formed in a strip shape on one main surface side, wherein the porous silicon region has a porosity of 52 to 62%. 2. In a silicon substrate in which a silicon oxide region is formed in a strip shape on one main surface side, a depth of the silicon oxide region is b with respect to one main surface of the silicon substrate.
(C + d) / b is within 6%, where c is the distance to the most concave portion of the surface of the silicon oxide region and d is the distance to the most convex portion of the surface of the silicon oxide region. Silicon substrate. 3. A silicon substrate having a silicon oxide region formed in a strip shape on one principal surface side, wherein the silicon oxide region has a first region having a predetermined refractive index, and the silicon substrate in the first region has a predetermined refractive index. 3. The silicon substrate according to claim 2, further comprising a second silicon oxide region located outside in the inside and having a lower refractive index than the first region. 4. 4. A method in which a silicon substrate is covered with a partially opened mask layer, immersed in a chemical conversion solution, a formation current is applied, and a part of the silicon substrate is anodized from a portion where the mask layer is removed. In the method for forming a silicon substrate in which a porous silicon region is formed in a band shape in the silicon substrate, a current density at an interface between the growth front end of the porous silicon region and the silicon substrate during the anodizing step is substantially constant. Wherein the formation current is changed in accordance with the degree of growth of the porous silicon region. 5. The silicon substrate is immersed in a chemical solution having a predetermined hydrofluoric acid concentration to form at a predetermined current density a predetermined region of the porous silicon region, and then the hydrofluoric acid concentration and the current density are changed. 5. The method according to claim 4, wherein the remaining porous silicon region is formed. 6. The method according to claim 4, wherein the formation current is a pulse current. 7. The method according to claim 4, wherein the formation current is a pulse current, and the application time per unit pulse and / or the repetition time thereof are changed with the growth of the porous silicon region. Item 6. The method for forming a silicon substrate according to Item 5. 8. An opening width of the mask layer is 0.1 to 20 μm.
The method for forming a silicon substrate according to claim 4, wherein m is m. 9. The formation depth in the silicon substrate is 50 μm.
The distance is within m.
3. The method for forming a silicon substrate according to item 1. 10. A silicon substrate, which is covered with a partially open mask layer, immersed in a chemical conversion solution, a formation current is applied, and the silicon substrate is anodized from a portion where the mask layer is removed. In a method of forming a silicon substrate for forming a porous silicon region in a part of the inside,
A method for forming a silicon substrate, comprising: forming an oxide film on a surface of the porous silicon region after forming the porous silicon region. 11. The method according to claim 10, further comprising a step of removing the oxide film.