JPS6083335A - Processing method of wafer - Google Patents

Abstract

PURPOSE:To make the film thickness of a diaphragm constant easily by removing one part of a region up to predetermined depth and simultaneously etching each region in a device in which one part of a semiconductor wafer is etched selectively and changed into a thin-film. CONSTITUTION:A thin-film section 11 is second thickness of t2 is formed in a first region 12 in the first main surface of an silicon wafer 10 with first thickness of t1 through etching. An opening in (a) width is formed to a mask pattern 14 in one part 13 of a third region in the second main surface of the wafer 10 corresponding to a second region except the region 12. One part of the second region except the first region or one part 13 of the third region in the second main surface of the wafer 10 corresponding to the second region is removed up to approximately the same depth as the second thickness t2. The first region 12 and the second or third region are etched simultaneously.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a method for processing a wafer, and more particularly, to a method for selectively etching a portion of a semiconductor wafer to reduce its thickness.

With recent advances in semiconductor technology, semiconductor diaphragms are used in pressure detectors for fluids such as air, water, oil, etc., and pressure/electrical signal converters. So-called.

Diaphragm type silicon pressure sensors are being put into practical use. Figures 1(a) and 1(bl) are diagrams explaining the structure of the sensor;
' is a sectional view at '. In the figure, a diaphragm 2 is provided in the center of a silicon die 1, which is selectively thinned using methods such as ultrasonic machining, electric discharge machining, anisotropic etching, and electrolytic milling. A plurality of gauge resistors 3 are embedded in the diaphragm part. The mosquito resistors are formed by methods such as ion implantation and thermal diffusion, and have a conductivity type opposite to that of the die 1. forming a junction. Furthermore, an insulating film (not shown) made of silicon oxide or the like is deposited on the main surface of the die 1 to increase the reliability of the sensor and stabilize its characteristics. Also,
A metal conductor 4 for detecting the resistance value of 3 is provided on the die 1 and is connected to a pin 7 provided on the package 6 through a bonding wire 5. On the main surface of die 1.

Usually, four gauge resistors are buried, and on the surface,
Alternatively, a Wheatstone bridge circuit is configured outside the package.

W, In the configuration example shown in Figure 1, 2 is 10 to 205
Since the diaphragm 2 has a thickness of m, the diaphragm 2 deforms in response to the pressure difference between the fluids introduced to the upper and lower surfaces of the diaphragm 2. The deformation induces compressive and tensile stress on the surface of the diaphragm 20, and a change in gauge resistance value occurs in proportion to the stress value. This change in resistance value is called a piezoresistance effect. As mentioned above,
It is well known that if a plurality of gauge resistors constitute a Wheatstone bridge, the change in resistance value can be detected as a voltage change. Since this structure has the thin film diaphragm 2, unnecessary stress and distortion occurring in the die 1 and deformation of the diaphragm are easily induced. One of the causes of such stress stress is the difference in thermal expansion coefficient between 1 and 6.

The thermal expansion coefficient of the silicon die 1 is about 3×10 2 /'C, and the thermal expansion coefficient of the package 6 is usually one order of magnitude larger. This difference in coefficients generates compressive and tensile stresses in the diaphragm 2 as the operating ambient temperature of the structure changes, causing the output voltage of the bridge to vary. The pedestal 8 provided between 1 and 6 is intended to reduce such unnecessary stress.

A material having a coefficient of thermal expansion between 1 and 6 or close to either -10,000 is selected. - Examples include single crystal silicon, Pyrex, crystallized glass, etc. The bond between 1.8 and 8.6 is a low melting glass-eutectic alloy.

This is achieved by well-known techniques such as electrostatic coupling. Reference numeral 9 denotes a pressure introduction pipe provided near the center of the package 6 and corresponding to the area of the diaphragm 2.

In FIG. 1, means for supplying fluid pressure from above the diaphragm 2 is omitted. Such supply means are
This is easily achieved by combining 6 with a cap having a second pressure introduction pipe.

All the pressure sensors shown in Figure 1 are made of silicon material, so they can be made with excellent reproducibility and long-term stability. Although conventional semiconductor processes can be used, a new process must be introduced to form the diaphragm 2. As mentioned above, various diaphragm forming methods are known, but all of them are different from conventional semiconductor processes. When the shape of the diaphragm is circular, it is known that the thickness is inversely proportional to the square of the thickness and proportional to one area when the same distributed pressure is applied. Even in the case of the quadrilateral shown in FIG. 1, the stress applied to the gauge resistor can be considered to be inversely proportional to the square of the thickness and proportional to one area. Since the area is determined in the photoresist process, sufficient accuracy can be maintained. However, the thickness of the diaphragm must be controlled during the thinning process. In diaphragm type silicon pressure sensors, silicon 7fl Ever having a thickness of about 200 to 500 μm is usually used as the base material.

If such a wafer is used, most of the thickness is removed in the thinning process.

It is necessary to leave about 10 to 20 μm. moreover,
Since the stress depends on the square of the thickness, film thickness control is required on the order of 1 μm. At the moment.

Various film thickness control methods have been proposed, but none of them are practical and have not been established as a technology for mass production6.If the film thickness varies from device to device, the stress of each device The values are different, ie.

The pressure transduction sensitivity will be distributed over a wide range. Such a distribution of sensitivity causes a serious problem in that the yield rate of the element is reduced and the sensitivity must be corrected individually in the signal processing circuit.

It is an object of the present invention to provide a wafer processing method that allows the thickness of the diaphragm to be easily kept at a constant value.

According to the present invention, the first region K of the first main surface of the wafer having the first thickness is etched by etching or the like.
In processing a wafer to form a thin film portion having a thickness of a first step of removing a portion of the etching layer to a depth approximately equal to the second thickness, and a second step of etching the first region and the second or third region at the same time. A characteristic weno-processing method can be obtained.

Next, a detailed explanation will be given with reference to the drawings.

FIG. 2 is a diagram illustrating a conventional example of controlling the film thickness of a diaphragm, and shows a cross section of a wafer. In the same figure, numeral 10 is a cricon wafer having a first thickness tl. Reference numeral 11 denotes a diaphragm or thin film portion having a second thickness t2 formed from the first region 12 on the first main surface side of 10.

As a means for forming No. 11, there is anisotropic etching using ethylene diamine, pyrocatechol, water, KOH, hydrazine, or the like. If the orientation of the 100 plane is (100), the slope in Figure 2 becomes the (111) plane, and the angle θ in the figure is 54.
．． It becomes 7°. As is well known, when an anisotropic etching solution is used, the etching speed of the (111) plane is extremely slow compared to other planes, so even when performing deep etching of 100 μm or more, side etching does not occur and the pattern is It has the advantage of maintaining accuracy. In the figure, there is also a region 13 on the second main surface of the wafer 10 where a portion is etched. The width a of the area is (2
t2/lan 54.7°).

That is, when 13 regions are anisotropically etched.

As shown in the figure, a triangular groove is formed, and its depth is 1
Etching will automatically stop when the value becomes equal to 2. The above etching is performed simultaneously from one and both surfaces using the mask pattern 14 made of a material such as an oxide film or nitride film provided on both main surfaces of 100 □ as a mask.

As the etching progresses and the groove 13 penetrates the etched area from the first major surface side, the wafer is separated in the etching solution and a die having a size indicated by 15 is created. Separation can be visually observed, so stop etching the moment the die separates.

The thickness of the diaphragm No. 11 is equal to t2.

A preferred method to quickly stop etching is to dilute the etching solution by pouring a large amount of water into it. With this method, the dies separated from the wafer can be subjected to desired post-processing.

Through the assembly process, the structure shown in FIG. 1 is realized. However, this film thickness control technique has serious drawbacks. That is, the thinning process and die separation process for forming the diaphragm are common.

A die collection method similar to “Kingyo Suku!” is essential;
Since die separation is performed in a solution, the dies collide with each other and destroy the diaphragm.

Spin drying, nitrogen blowing, etc. cannot be used during drying of the die, so good drying cannot be achieved, and post-processing must be performed for each die, making it impossible to prevent an increase in man-hours. Furthermore, since a slope is formed on the first main surface side of the wafer opposite to 13, there is also the disadvantage that the die size is large and the number of die accommodated on the wafer is reduced.

The invention of the collar eliminates the drawbacks of the prior art.

FIG. 3 is a diagram for explaining one embodiment of the present invention,
A wafer cross-section similar to FIG. 2 is shown. In this figure, the same numbers as in FIG. 2 indicate the same components.

In the figure, a thin film portion 11 having a second thickness [2] is formed by etching in a first region 12 of the first main surface of a silicon wafer 10 having a first thickness tl. .

In a portion 13 of the third region of the second main surface of the wafer corresponding to the second region excluding the region 12, an opening having a width a is provided in the mask pattern 14. The present invention differs from FIG. 2 in that 13 is limited to a portion within the wafer. An example of the position of 13 on the second main surface of the wafer is shown in FIG.

It is not limited to this. In Fig. 4.

20 is the second main surface of the wafer, 21 is an orientation flat in the <110> direction, for example, indicating the crystal axis direction of the wafer, and 22 is the position where 13 is provided.

An opening 13 is formed in 14, and a mask pattern 14 for forming 11 is also provided on the first main surface of the wafer, after which pressure anisotropic etching is performed. As etching progresses in region 13, triangular grooves are generated.
As the etching progresses further, the etched region from the first main surface opposing the region 13 is penetrated.

Once such penetration is achieved, portion 23 of the wafer separates from the wafer. However, in this embodiment, since the number 13 is limited to one location within the wafer, the gouers whose size is indicated by 25 do not separate from each other and maintain the wafer shape. At the moment when '23 separates, wafer 10 is processed.

Etching will stop if you remove it from the tinde solution and wash it with water. If the width a of 130 is set in the above-described relationship with the desired second thickness t2, it will match the thickness t2 of 11 when etching is stopped. Such wafers undergo post-processing.

Proceeding to the scribing and assembling process, the structure shown in FIG. 1 is realized. In this embodiment, since the film thinning process and the die separation process are separate processes, normal handling is possible in the post-process, and the above-mentioned drawbacks are eliminated.

Since there is no region etched from the first main surface between the grooves and the grooves, the groove size indicated by the size of 25 can be made smaller than in the conventional example, and the number of dies per wafer can also be increased. . In this embodiment, it is necessary to simultaneously form etching mask patterns on the first and second major surfaces of the wafer. A photoresist process is generally used to form such a mask pattern, but
A double-sided exposure machine must be used. A second embodiment of the present invention, which will be described below, shows an improved method that does not require a double-sided exposure machine.

5(a) to 5(e) are diagrams for explaining the second embodiment of the present invention, and show wafer cross-sectional views for each process.In addition, in the same figures, the same numbers as in FIG. 3 are shown. 1 shows the same structure.
A mask 30 made of an oxide film, a nitride film, etc. in which a desired pattern is formed is provided.

A similar mask 31 without a pattern is provided on the second main surface. 32 is a first region in which a diaphragm which becomes a thin film portion is formed by etching, and 33 is a second region that does not include the first region, both of which are not covered with the mask 3o. 34 is a mask covering the wafer other than the second region 33; The same figure (al
The structure in area 3 is immersed in a silicon etching solution.
3 is etched to a depth t2. In this process, the etching solution is not one that performs the above-mentioned anisotropic etching, but may be an isotropic etching solution such as a mixture of nitric acid and hydrofluoric acid. simplifies the process. After the etching is completed, when 34 is removed, the same figure (b) is obtained.
The structure shown in is obtained. Next, regions 32 and 33 are simultaneously etched using the anisotropic etching. As shown in the same figure (bl), in the anisotropic etching, 32 and 33 whose thicknesses differ by t2 are etched at the same time, so after a certain period of time has passed, K as shown in the same figure (C) is etched.
.

All the silicon in region 33 is removed, and at the same time
2, the thickness of silicon is t2. Since the thickness of 31 is usually 0.2 to 1 μm la degrees, it is sufficiently smaller than t2, which is 10 to 20 μm.

Mechanical strength is also weak. That is, part 35 of 31, such as linen, is destroyed by the stirring of the etching solution and separated from the wafer. If 35 is located only at the edge of the wafer, visual observation is possible. No. M. 6 places. -fil is a diagram, but is not limited to this.

In the figure, 40 is the first main surface of the wafer.

41 is a diaphragm formed in the region 32;

In other words, it is a thin film part.The mounting device shown in the same figure pulls up the wafer the moment one end of the wafer breaks off.
If etching is stopped by washing with water, the thickness of the thin film portion becomes t2, which means that film thickness control has been achieved. In this example, two types of eratin masks are used, but one type of mask may be patterned twice as shown in FIGS.
1, a mask 5o is provided on the first main surface, and only the region 33 is opened. If only the region 33 is etched to the depth tl using the etching method described above, the structure shown in FIG. 3(b) is obtained. Next, the region 32 is opened in the photoresist step 50 to obtain the structure shown in the same figure (C1).

By applying anisotropic etching, the structure shown in the same figure (cl) is obtained, and film thickness control is confirmed by separation of μ35. The second embodiment of the present invention shown in FIGS. 5(al-(C)) and FIGS. 7(ta)-(di) is characterized in that the region 33 is etched in advance to a desired depth. Of course, in order to process and remove a part of the wafer to this depth, methods other than etching may be used, such as mechanical methods such as ultrasonic machining, electric discharge machining, and polishing. A non-chemical etching method using particles or the like may also be used.

In addition, A35 is not separated from the wafer,
The etching may be stopped in the integrated state. The mask 31 is transparent to normal light and is shown in FIGS. 8(a) to 8(d).
i is a diagram for explaining the third embodiment of the present invention;
Ueno cross-sectional views are shown for each process. In this figure, the same numbers as in FIG. 5 indicate the same components. Same figure (
In the case of al, the second main surface of the silicon wafer 10 having a thickness tl is covered with a mask 60 serving as an etching mask, and a part thereof has an opening 62. The position of the opening is 3 in Figure 6.
A position similar to 5 is selected. The first main surface of the weno sheet is covered with a full-face mask 61. At 60°61, an oxide film, a nitride film, etc. are selected. next.

The opening 62 is etched to a depth t2i (FIG. 6(b)). Next, 61 is patterned, and the 61 in the area 32 is removed (FIG.
)). Such a structure is immersed in an anisotropic etching solution, and 3
At the time when all the silicon in the region 2.62 is removed from both main surfaces, the film thickness in the region 32 becomes tl. A portion 63 of the region 62 becomes fIIJ, and the relevant time is determined by utilizing the fact that it is destroyed and separated in the liquid, or becomes transparent to light. If the etching is stopped at this time, film thickness control during thin film formation will be realized. In this example, the i wafer, -
The second region of the wafer corresponding to the region excluding the first region 32 of
A portion 62 of the third region of the main surface of the first region is removed by etching, machining, etc. to a desired depth in advance, and then a portion 62 of the third region of the main surface of the first region is removed by etching, machining, etc. to a desired depth. The third area is etched at the same time. There are certain characteristics.

The present invention has been described in detail above. According to the present invention, a thin film portion is formed on a part of the wafer □-, ie.

In the processing method for forming the 7-diameter shim, it is possible to easily control the film thickness of the diaphragm to a predetermined constant value. Compared to other control methods that have been published at the present time, the present invention is capable of providing a distribution once on the woofer, and then growing a silicon single crystal film on an insulating film.
To provide a practical control method suitable for mass production without using 80I wafers. The present invention is not simply applied to a diaphragm type pressure sensor;
It can also be applied to other devices that require a thin film process, such as the process of manufacturing micromechanical devices by processing silicon wafers by anisotropic etching. In addition, by utilizing the extremely slow etching speed of the high concentration layer,
The present invention also includes the use of the present invention in conjunction with a method for automatically stopping etching at a PN junction boundary, an electrochemical etching method for controlling the etching rate using an electric field, and visual monitoring when etching is stopped.

[Brief explanation of the drawing]

Fig. 1(a), (bl is a diagram explaining the structure of a diaphragm type pressure sensor, 1 is a silicon die, 2 is a diaphragm, 3 is a gauge resistor, 4 is a metal conductor, 5 is a bonding wire, 6 is a Package, 7 is pin,
8 is a pedestal, and 9 is a pressure introduction pipe. FIG. 2 is a diagram illustrating a conventional example of controlling the film thickness of a diaphragm, in which 10 is a silicon wafer, 11 is a diaphragm, that is, a thin film portion, 12 and 13 are regions, 14 is a mask pattern, and 15 is a It is the size of the die. FIGS. 3 and 4 are diagrams illustrating one embodiment of the present invention; the same numbers as in FIG. 2 indicate the same components. In the figure, 20 is the second main surface of the wafer, 21 is the orientation flat, 22 is the potential, 23 is the part of the wafer to be separated, and 25 is the size of the die. The same numbers as in the figures indicate the same components. In the figure, 30, 31, 34.50 are masks, 32°33 are regions,
35 is μ, 40 is the first main surface of the wafer, and 41 is a diaphragm, that is, a thin film portion. FIGS. 8(a) to (di) are diagrams explaining the third embodiment of the present invention, and the same numbers as in FIG. 5 indicate the same components. In the same figure, 60-61 are masks; 62 is the area, 63 is the island. (b) 1st 0 20th 4th A 4th C concavity

Claims (1)

[Claims] In processing a wafer in which a thin film portion having a second thickness is formed in a first region of a first main surface of the wafer having a first thickness by means such as etching, the first A first step of removing a part of the second region other than the region or a part of the third region of the second main surface of the wafer corresponding to the region of the cylinder 2 to a depth approximately equal to the thickness of the '2nd part. and a second step of etching the first region and the second or third region at the same time.