Pattern shift measuring method.

0001This invention relates to a method of measuring shift of patterns of a diffused semiconductor wafer after growth of an epitaxial layer thereon, and more particularly to a method of measuring the pattern shift without destroying the semiconductor wafer and in a short period of time.

0002It is essential to control the shift of a buried diffusion pattern after growth of an epitaxial layer in a bipolar transistor in an IC. For this purpose, the growth conditions (such as reaction temperature and reaction speed) should be always controlled to assure constant pattern shift. However, it is very difficult to strictly control the growth conditions since frequent measurement of the pattern shift is inevitable.

0003Conventionally, the angular lapping stain method is used for this purpose. This method comprises: (1) slicing chips having a buried layer in parallel to and perpendicular to the orientation flat by using a dicing saw; (2) angular polishing of o ptional surfaces of the sliced tips ( ϑ = 11° 32'); (3) etching thepolished surfaces (Sirtl, 2 to 3 seconds); and (4) measuring the shift of the patterns of the buried layers by using differential, interference micrographs (x 150).

0004The pattern shift factor is obtained by the following equation: <maths id="math0001" num=""><math display="inline"><mrow><mtext>the pattern shift factor</mtext></mrow></math><img file="EP0484665A2_D0001.tif" /></maths><maths id="math0002" num=""><math display="inline"><mrow><mtext>= the amount of shift ( µm)/the thickness of the epitaxial layer ( µm)</mtext></mrow></math><img file="EP0484665A2_D0002.tif" /></maths>

0005However since it takes approximately two hours to measure the pattern shift and requires destruction of specimens, this conventional method cannot be used so frequently and is rather expensive.

0006To solve the problem of the conventional method, some proposals are made to perform the pattern shift measurement through non-destructive tests as disclosed in such publications as Japanese Patent Laid Open Publications 64-31413 and 64-31414. Under the present circumstances, there is a great demand for a more simplified and reliable method of measuring the pattern shift.

0007It is therefore an object of this invention to provide a method of reliably measuring the pattern shift of a semiconductor wafer in a short period of time without destroying the semiconductor wafer.

0008According to this invention, there is provided a method of measuring a shift between a diffusion pattern and a corresponding pattern on an epitaxial layer of a semiconductor wafer after the epitaxial process. The method comprises measuring a ratio between the line width of a linear pattern vertical to the orientation flat and line width of a linear pattern parallel to the orientation flat to determine the shift between these patterns.

0009The pattern shift can be also determined only by measuring the line width of a linear pattern vertical to the orientation flat.

0010Further, the pattern shift is determined only by measuring the line width of a linear pattern parallel to the orientation flat.

0011According to this invention, the pattern shift can be reliably and inexpensively measured in about five minutes without destroying the semiconductor wafers.

0012Many other objects, advantages and features of the present invention will be better understood from the following description taken in conjunction with the accompanying drawings. <ul id="ul0001" list-style="none"><li>FIG. 1 shows a correlation between a ratio W_V/W_H obtained by the method of this invention and a pattern shift ratio obtained by the conventional method;</li><li>FIG. 2 shows a correlation between W_V of this invention and the pattern shift ratio of the conventional method; and</li><li>FIG. 3 shows a correlation between W_H of this invention and the pattern shift ratio of the conventional method.</li></ul>

0013One embodiment of this invention will now be described with reference to the accompanying drawings.

0014Measurement was performed under the following conditions: Specimens; CZ p-type semiconductor wafer 〈111〉 off-angle 3° 30' in 〈112〉 100 φ , 10-20Ω-cm, OF〈110〉 Buried diffusion layer; Sb, 15Ω/□, depth 8mm Width of the buried diffusion layer; 60 µm Epitaxial layer growth; 10 µm, 1Ω-cm Reaction furnaces used; Cylindrical and vertical furnaces Reaction temperatures; 1150, 1100 and 1050°C Reaction speeds; 0.30 and 0.60 µm /min Reaction pressure; 760 Torr Measurement; A Nomarski differential microscopic interferometer (magnification factor, x200) was used to measure the line width (W_V) of a pattern vertical to the OF(orientation flat), and the line width (W_H) of a pattern parallel to the OF. The measured values and the ratio W_V/W_H are shown in Table 1. <tables id="tabl0001" num="0001"><img file="EP0484665A2_D0003.tif" /></tables>

0015FIGS. 1 to 3 are graphs showing the correlation between the pattern shift ratios determined by the conventional method (angular lapping stain method), and the values W_V /W_H , W_V and W_H determined by the method of this invention.

0016Any of W_V/W_H, W_V and W_H have correlations to the pattern shift ratio of the conventional method (W_V/W_H and W_V have negative correlations and W_H has a positive correlation to the conventional pattern shift ratio). It was confirmed that the pattern shift can be determined by measuring W_V/W_H , and W_V or W_H .

0017In this embodiment, the buried diffusion layer is antimony (Sb). The method of this invention is also applicable to buried diffusion layers of impurities such as boron (B), phosphorous (P) and arsenic (As).

0018According to this invention, it is possible to measure the pattern shift of the epitaxial semiconductor wafer in a short period of time without destroying the wafer. Therefore, measurement can be performed very frequently so that the growth conditions can be always kept optimum.