JP2004193529A - Wafer evaluation method and apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a wafer evaluation method and apparatus which can effectively detect and identify what type of defects or contaminations are present on a wafer and how the defects or foreign mater is distributed on the wafer. <P>SOLUTION: Optical scanning equipment based on a confocal optical system inspects the surface of a wafer to be inspected as a sample, identifies the defects or contaminations on the sample surface detected by the inspection according to its type, outputs and displays the position of the identified defects or contaminations on a wafer screen corresponding to the specimen by a symbol indicative of the type of the defects or contaminations. <P>COPYRIGHT: (C)2004,JPO&NCIPI

Description

【０１１４】
通常は、ａ’≦ａ＜ｂ≦ｂ’であり、８ビット濃淡画像の場合は、ａ’＝０、ｂ’＝２５５となる。
【０１１５】
以下に、欠陥種別の自動分類について図３１を用いて説明する。図３１は欠陥種別の自動分類手順の１例を示すフローチャートで、微小ＬＤＰ、突起、線状突起、キズ、スタッキングフォールトについての分類手順を示す。なお、図３１において、他の欠陥種別に対する判別処理や条件に当てはまらない場合の処理などは省略してある。
【０１１６】
まず、平滑化による雑音除去処理を行う（ステップ３００）。次いで、エッジ抽出による欠陥候補領域の抽出を行う（ステップ３０２）。続いて、欠陥候補領域の個数、大きさ、分布による判別処理の振り分けを行う（ステップ３０４）。この振り分けステップにおいて、欠陥候補領域は、▲１▼小さい候補領域が一つの場合、▲２▼大きい候補領域が一つの場合、▲３▼やや大きめの候補領域が二つの場合、の三つに振り分けられる。
【０１１７】
▲１▼小さい候補領域が一つの場合には、欠陥候補領域における明領域及び暗領域の抽出を行う（ステップ３０６）。次に、明領域及び暗領域の大きさ及び位置関係による欠陥種別の判定を行う（ステップ３０８）。この判定において、明領域が右，暗領域が左にある場合には微小欠陥と判断し、また、明領域が左，暗領域が右にある場合には突起と判断する。
【０１１８】
▲２▼大きい候補領域が一つの場合には、欠陥候補領域における明領域及び暗領域の中点の列を直線近似して近似誤差により欠陥種別を判定する（ステップ３１０）。この判定において、近似誤差が少なく、明領域が左，暗領域が右にある場合には線状突起と判断し、また、近似誤差が少なく、明領域が右，暗領域が左にある場合にはキズと判断する。
【０１１９】
▲３▼やや大きめの候補領域が二つの場合には、２個の欠陥候補領域の形状と位置関係による欠陥種別の判定を行う（ステップ３１２）。この判定において、２個の縦長の長方形が並行に対をなしている場合にスタッキングフォールトと判断する。例えば、このような処理手順がプログラミングされた識別装置により、欠陥や異物の種類を選別し、出力装置によりウエーハ上の欠陥の位置や種類別を表示するようにしてある。
【０１２０】
【実施例】
以下に本発明を実施例をあげてさらに具体的に説明する。これらの実施例は例示的に示されるもので、限定的に解釈されるべきでないことはいうまでもない。
【０１２１】
（実施例１）ＰＷウエーハの評価
直径３００ｍｍのシリコンウエーハについて図１に示したような本発明のウエーハの評価装置を用い、０．０６μｍ以上の欠陥を評価した。この装置は図２に示すような光学系の装置である。
【０１２２】
その結果を図５に示す。これにより、今回評価したシリコンウエーハでは、表示欠陥数は２７９個で、６種類の欠陥や異物が観察された。特にこのウエーハは微小欠陥（結晶欠陥であるＣＯＰ等）がウエーハの中心付近に多く分布していることがわかる。これにより結晶の製造条件を見直すことなど、製造現場へのフィードバックができる。
【０１２３】
（比較例１）
実施例１と同じウエーハをレーザー散乱方式の異物検査装置を用い、０．０６５μｍ以上の欠陥を評価した。
【０１２４】
図７に示すように検出される欠陥の数も少なく。また欠陥の形態（種類）等の区別はできなかった。
【０１２５】
（実施例２）ＥＰＷウエーハの評価
直径３００ｍｍのシリコンウエーハ上にエピタキシャル層を形成したエピタキシャルウエーハについて、実施例１と同様の本発明装置を用い測定した。
【０１２６】
その結果、図６に示すように今回評価したウエーハ表面には、表示欠陥数は２５個で、ＳＦ、マウンド等が観察された。
【０１２７】
（比較例２）
実施例２と同じエピタキシャルウエーハをレーザー散乱方式の異物検査装置を用い、０．１２μｍ以上の欠陥を評価した。
【０１２８】
図８に示すように検出される欠陥の数も少なく。また欠陥の形態（種類）等の区別はできなかった。
【０１２９】
なお、本発明は上記実施の形態に限定されるものではない。上記実施形態は例示であり、本発明の特許請求の範囲に記載された技術思想と実質的に同一な構成を有し、同様な作用効果を奏するものは、いかなるものであっても本発明の技術範囲に包含される。
【０１３０】
【発明の効果】
以上述べたごとく、本発明によれば、ウエーハ上の欠陥や異物の分布状況と各欠陥や異物の成分が一目で判る上、欠陥や異物の分布状況から問題となる工程や装置等が推定できるという効果が達成される。
【図面の簡単な説明】
【図１】本発明のウエーハの評価装置の一つの実施の形態を示す模式的説明図である。
【図２】コンフォーカル光学系によるレーザー顕微鏡の基本構造を示す概略説明図である。
【図３】図２のＡ部分の拡大断面説明図である。
【図４】コンフォーカル光学系によるレーザー顕微鏡を用いたウエーハの検査手順の１例を示すフローチャートである。
【図５】実施例１において観察された欠陥を種類別に示す画像である。
【図６】実施例２において観察された欠陥を種類別に示す画像である。
【図７】比較例１において観察された欠陥を示す画像である。
【図８】比較例２において観察された欠陥を示す画像である。
【図９】コンフォーカル光学系によるレーザー顕微鏡によりウエーハ表面に観察される欠陥の第１の例を示す写真である。
【図１０】コンフォーカル光学系によるレーザー顕微鏡によりウエーハ表面に観察される欠陥の第２の例を示す写真である。
【図１１】コンフォーカル光学系によるレーザー顕微鏡によりウエーハ表面に観察される欠陥の第３の例を示す写真である。
【図１２】コンフォーカル光学系によるレーザー顕微鏡によりウエーハ表面に観察される欠陥の第４の例を示す写真である。
【図１３】コンフォーカル光学系によるレーザー顕微鏡によりウエーハ表面に観察される欠陥の第５の例を示す写真である。
【図１４】コンフォーカル光学系によるレーザー顕微鏡によりウエーハ表面に観察される欠陥の第６の例を示す写真である。
【図１５】コンフォーカル光学系によるレーザー顕微鏡によりウエーハ表面に観察される欠陥の第７の例を示す写真である。
【図１６】コンフォーカル光学系によるレーザー顕微鏡によりウエーハ表面に観察される欠陥の第８の例を示す写真である。
【図１７】コンフォーカル光学系によるレーザー顕微鏡によりウエーハ表面に観察される欠陥の第９の例を示す写真である。
【図１８】コンフォーカル光学系によるレーザー顕微鏡によりウエーハ表面に観察される欠陥の第１０の例を示す写真である。
【図１９】コンフォーカル光学系によるレーザー顕微鏡によりウエーハ表面に観察される欠陥の第１１の例を示す写真である。
【図２０】コンフォーカル光学系によるレーザー顕微鏡によりウエーハ表面に観察される欠陥の第１２の例を示す写真である。
【図２１】コンフォーカル光学系によるレーザー顕微鏡によりウエーハ表面に観察される欠陥の第１３の例を示す写真である。
【図２２】コンフォーカル光学系の原理を説明するための線図である。
【図２３】本発明において用いられる光学式走査装置の原理を説明するための線図である。
【図２４】光スポットが欠陥部分を走査する際の反射ビームの光路を示す図である。
【図２５】本発明において用いられる光学式走査装置の一例の構成を示す線図である。
【図２６】本発明において用いられる欠陥検出装置の一例の構成を示す線図である。
【図２７】欠陥検出回路の一例の構成を示す図である。
【図２８】本発明において用いられる光学式走査装置の変形例を示す線図である。
【図２９】試料上に形成される光スポットアレイと主走査方向との関係を示すである。
【図３０】光検出器の一例を示す平面図である。
【図３１】欠陥種別の自動分類手順の１例を示すフローチャートである。
【符号の説明】
１０：コンフォーカル光学系によるレーザー顕微鏡、１２：顕微鏡本体、１４：レーザー光源、１６，１０２：ビームスプリッタ、１８，１０３：対物レンズ、２０ａ： ピンホール、２０：ピンホール部材、２２：集光レンズ、２４，１０６：光検出器、２６：遮光板、３０：本発明のウエーハの評価装置、４０：識別装置、５０：出力装置、Ｂ：レーザービーム、Ｆ：欠陥、Ｗ：ウエーハ（試料）、Ｗ１：ウエーハ画面、１０１：光源、１０５：空間フィルタ。[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to a wafer evaluation method and apparatus for automatically detecting and identifying defects and foreign substances present on a wafer, for example, a semiconductor wafer such as a silicon wafer.
[0002]
[Related technology]
Generally, a method for manufacturing a semiconductor wafer such as a silicon wafer includes a single crystal manufacturing step of growing a single crystal ingot using a Czochralski (CZ) method or the like, and slicing the single crystal ingot to form at least one principal surface having a mirror surface. And a wafer processing step. Further, the wafer processing step includes a slicing step of slicing a single crystal ingot to obtain a thin disk-shaped wafer, and a chamfering step of chamfering an outer peripheral portion of the wafer obtained in the slicing step to prevent cracking and chipping. A lapping step of flattening the wafer; an etching step of removing processing distortion remaining on the chamfered and wrapped wafer; a polishing (polishing) step of mirror-polishing the wafer surface; and cleaning of the polished wafer. Then, there is a cleaning step for removing abrasives and foreign substances adhered to this. The above steps show the main steps, and other steps such as a heat treatment step are added, the order of the steps is changed, or the steps are performed in multiple stages. The silicon wafer manufactured in this manner is finally subjected to quality inspection, and then put into a container for storing the wafer, and after packaging, the manufactured mirror-polished wafer (hereinafter sometimes referred to as PW) is used. It is sent to a device manufacturing company (device process) to form a device. In addition, in order to add value to the mirror-polished wafer before forming a device, an epi-step for forming an epitaxial layer on the mirror-polished wafer may be further included. A wafer obtained through such a process is called an epitaxial wafer (hereinafter, sometimes referred to as an epi wafer or EPW). In addition, a mirror-polished wafer may be bonded via an oxide film to produce an SOI wafer or the like.
[0003]
In such a manufacturing process, with the miniaturization of devices, the device characteristics to be achieved are becoming increasingly severe, and silicon wafers are required to have further crystal quality completeness and surface cleaning. .
[0004]
Therefore, it is necessary to precisely evaluate the quality of the silicon wafer and to improve the manufacturing process of the silicon wafer and the device manufacturing process. In other words, in the production of semiconductor wafers such as silicon wafers, defects existing on the wafer and foreign substances adhering to the wafer cause a reduction in yield, and the defects and foreign substances adhering to these wafers are inspected to generate defects and foreign substances. In some cases, the source of the defect or foreign matter (generation step) is analyzed by controlling the amount or analyzing the defect or the attached foreign matter present on the wafer.
[0005]
Defects and foreign substances detected in semiconductor wafers such as silicon wafers include defects caused by crystals introduced in the above-described single crystal manufacturing process, defects caused by processing introduced in the wafer processing process, and impurities (particles, heavy metals, etc.). )).
[0006]
Conventionally, an inspection apparatus called a particle counter based on light scattering has been mainly used for inspection of such a silicon wafer surface.
[0007]
This device detects the scattering intensity and detects particles on the wafer.However, if the size is not less than a certain size other than particles, various types of defects are detected, and it is difficult to distinguish between them. Defects and foreign matter detected by this method are counted together, and the amount of defects generated is managed under the name of LPD (Light Point Defect).
[0008]
As a defect that has been a problem in the device process, a COP (Crystal Originated Particle) that appears near the surface layer of the wafer has been known. These are crystal-induced defects that are introduced when the crystal is pulled. As a method of evaluating such a defect, before evaluating each defect, a pretreatment is performed on the silicon wafer itself to improve the sensitivity for a specific defect (to make the defect obvious), and thereafter, visually or by an electron microscope, etc. Is directly observing defects. For example, in the above COP inspection, the COP is a defect having a size of 0.1 μm or less, but becomes apparent by treatment with a solution of ammonia and hydrogen peroxide (also called SC1 solution) and appears as pits on the wafer surface. Detected.
[0009]
In addition, the position of a bright spot on a wafer detected by a particle counter is specified, and the same point as the bright spot is measured by another measuring device to identify a defect.
[0010]
As a method for evaluating the surface of a silicon wafer, a laser microscope using a confocal optical system has recently been used in addition to a particle counter. A confocal optical system focuses laser light on a sample, irradiates it with a small spot, focuses the reflected light again on a pinhole arranged on the entire surface of the receiver, and detects the amount of light passing through the pinhole. It is. In this method, defects and foreign substances on the wafer surface are evaluated with higher sensitivity than a conventional particle counter.
[0011]
[Problems to be solved by the invention]
However, even with the laser microscope using the confocal optical system, the defects detected by the apparatus have been counted together and the total amount and distribution have been observed.
[0012]
Since a laser microscope using a confocal optical system can evaluate with high sensitivity, various types of defects and foreign substances are particularly detected. Although it is effective to control the quality of the wafer by counting the total number of these defects and foreign substances, there is a problem in feeding back the evaluation result to the manufacturing process.
[0013]
At least, it is necessary to distinguish a defect caused by a crystal from a defect caused by processing introduced in a wafer processing step, or a foreign substance such as an impurity (particle or heavy metal).
[0014]
That is, although the number and distribution of defects and foreign substances can be confirmed to some extent in the past, the type of the defect and the positional information (distribution) of each defect have not been clearly grasped.
[0015]
An important point in evaluating a wafer is to clarify what kind of defects and foreign substances are present on the wafer and how each defect and foreign substance is distributed.
[0016]
In particular, it has been found that a characteristic defect which is easily detected when an optical scanning device (for example, a laser microscope) using a confocal optical system is present on the wafer. It is necessary to effectively detect and identify such defects and to improve (feedback) the manufacturing process so that these defects do not exist.
[0017]
When the types of the defects and foreign substances on the wafer and their distribution as described above are found, the state of attachment of the defects and foreign substances to the wafer can be seen at a glance, which makes it easier to clarify the cause of the generation of the defects and foreign substances.
[0018]
The present invention has been made in view of the above-described problems of the related art, and has an effect on what kind of defects and foreign substances are present on a wafer surface and how each defect and foreign substances are distributed on the wafer surface. It is an object of the present invention to provide a method and an apparatus for evaluating a wafer which can be detected and identified in a specific manner.
[0019]
[Means for Solving the Problems]
Using an optical scanning device (for example, a laser microscope) based on a confocal optical system, a close examination of defects observed on the surface of the silicon wafer revealed that a characteristic shape was observed due to the causes of the defects. Was. Therefore, it has been clarified that by performing image processing on a characteristic shape, it is possible to accurately grasp the type of defect and position information (distribution) for each defect.
[0020]
Therefore, the wafer evaluation method of the present invention inspects the surface of a sample, which is a wafer to be inspected, by an optical scanning device using a confocal optical system, and detects defects and foreign substances on the sample surface detected by the inspection by type. The position of the identified defect or foreign matter is output and displayed on a wafer screen corresponding to the sample by a symbol indicating the type of each defect or foreign matter.
[0021]
In particular, it is preferable to identify the type of defect or foreign matter on the surface of the sample (wafer) detected by an optical scanning device using a confocal optical system based on its shape and brightness.
[0022]
In particular, it is effective to use a silicon wafer as the wafer to be inspected. It is preferable to use a laser microscope using a confocal optical system as the optical scanning device using the confocal optical system.
[0023]
In a conventional particle counter (for example, a foreign substance inspection apparatus using a laser scattering method), it is difficult to distinguish a foreign substance or a defect. That is, in order to accurately identify a defect, a method of re-measuring a defective portion measured by a particle counter by another analyzer has been adopted. However, by using an optical scanning device (for example, a laser microscope) using a confocal optical system, it is possible to directly observe the actual state of a defect or a foreign substance in the same apparatus, and thereby to know the distribution of a specific defect or foreign substance. For example, from the distribution of each defect or foreign matter, it is possible to find out the process or apparatus that caused the defect or caused the particles.
[0024]
As an evaluation apparatus for performing such an evaluation method, a first aspect of the wafer evaluation apparatus of the present invention is an optical scanning apparatus using a confocal optical system that inspects the surface of a sample that is a wafer to be inspected, An identification device that analyzes image data of the sample surface obtained by the scanning device and identifies a type of a defect or a foreign substance present on the sample surface, and a position of the identified defect or the foreign substance to each defect or the foreign substance. An output device is provided for displaying on a wafer screen corresponding to the sample by a symbol indicating the type. As an optical scanning device using the confocal optical system, a laser microscope using a confocal optical system can be used.
[0025]
In particular, an optical scanning device using the confocal optical system, a sample stage that supports the sample, a light source that generates a light beam, and an objective lens that focuses the light beam to form a light spot on the sample surface, A photodetector that receives reflected light from the sample surface, and a light beam that is disposed in an optical path between the light source and the objective lens and that is directed from the light source to the sample and a reflected beam that is directed from the sample to the photodetector; And a means for relatively moving the sample and the light spot, scanning the sample surface with the light spot, and obtaining information on the surface area of the sample by reflected light from the sample surface. An optical scanning device for detecting, wherein a light-shielding plate is disposed in an optical path between the beam splitter and the photodetector, and the light-shielding plate is used to detect a light spot scanning direction on the sample surface. Preference has the configuration of shielding the light path of one half in the direction is.
[0026]
It is preferable that the means for relatively moving the light spot and the sample is a means for moving the sample stage with respect to the light spot.
[0027]
It is preferable that the means for relatively moving the light spot and the sample be a beam deflector arranged in an optical path width between the beam splitter and the sample stage.
[0028]
One-dimensional conversion of a light beam generated from the light source into a one-dimensional beam array of m (m is a natural number of 2 or more) sub-beams arranged in a line in an optical path between the light source and the beam splitter. A diffraction grating is arranged, and the sample surface can be scanned by m light spots formed on the sample.
[0029]
In the optical path between the light source and the beam splitter, when m and n are natural numbers of 2 or more, a two-dimensional sub-beam of a radiation beam generated from the light source is arranged in a matrix of m rows and n columns. A two-dimensional diffraction grating for converting into a beam array is arranged, the surface of the sample is scanned by an m × n two-dimensional light spot array, and the light shielding plate is connected to an optical system between the beam splitter and the photodetector. It is preferable to arrange at the position of the pupil.
[0030]
As means for relatively moving the sample and the two-dimensional light spot array, a rotation driving device for rotating the sample stage and a translation driving device for moving the sample stage along an axis orthogonal to the rotation axis are used. Preferably, the sample is spirally scanned with a two-dimensional light spot array.
[0031]
Further, as the optical scanning device using the confocal optical system, another optical system may be used, in which a sample stage supporting the sample, a light source for generating a radiation beam, and m and n are natural numbers of 2 or more. A two-dimensional diffraction grating for converting the radiation beam into a beam array of light beams arranged in a matrix of m rows and n columns; a beam deflecting device for deflecting the beam array in a first scanning direction; An objective lens that focuses the light beams of the array into spots to form an array of m × n light spots on a sample, and is disposed in an optical path between the light source and the objective lens; A beam splitter for separating a light beam directed toward the sample and a reflected beam directed from the sample toward the photodetector, and arranged in a matrix of m ′ rows and n ′ columns when m ′ and n ′ are natural numbers of 2 or more. Receiving A light detector that receives reflected light from a light spot formed on the sample, and a light detector disposed at a pupil position in an optical path between the beam splitter and the detector. A light-shielding plate that shields one half of the optical path in the direction corresponding to the scanning direction of the light spot on the sample surface, and applies the reflected beam from the sample surface to the light-receiving element of a photodetector via the beam deflection device. The light spot array formed on the sample is formed such that the intervals between adjacent light spots are equal when the light spots are projected in a direction orthogonal to the first direction. Can also be adopted.
[0032]
The angle between the axis in the row direction of the light spot array of m rows and n columns formed on the sample and the axis in the direction orthogonal to the first direction is θ, and the spot interval in the row direction of the light spot array is When p1 and the spot interval in the column direction are p2, the following equation (1) is obtained.
[0033]
[Equation 3]
tan θ = (1 / n) × (p1 / p2) (1)
[0034]
Is set to satisfy the condition.
[0035]
The intervals p1 and p2 between the light spots in the row and column directions of the light spot array are set so that p1 = p2, and the following equation (2) is set.
[0036]
(Equation 4)
tan θ = 1 / n (2)
[0037]
The light spot array is formed so as to satisfy the following.
[0038]
In a second aspect of the wafer evaluation apparatus of the present invention, the surface of a sample, which is a wafer to be inspected, is inspected using an optical scanning device using a confocal optical system, and the surface of the sample obtained by the optical scanning device is inspected. A defect detection device that analyzes image data and detects a defect or foreign matter present on the sample surface, and an identification device that identifies the type of defect or foreign matter present on the sample surface obtained by the defect detection device, An output device is provided to output and display the positions of the detected defects and foreign substances on a wafer screen corresponding to the sample by using symbols indicating the types of the respective defects and foreign substances.
[0039]
The defect detection device includes: a sample stage that supports a sample that is the inspection target wafer; a light source that generates a light beam; an objective lens that focuses the light beam to form a light spot on a sample surface; Means for relatively moving the light spot, a photodetector for receiving reflected light from the surface of the sample, and a light detector disposed in an optical path between the light source and the objective lens. A beam splitter for separating a light beam traveling toward the photodetector and a reflected beam traveling from the sample toward the photodetector, and a defect detection circuit for detecting a defect present on the sample surface from an output signal from the photodetector. A defect detection device, which is disposed in an optical path between the beam splitter and the photodetector, and has a light shielding plate that shields an optical path on one half in a direction corresponding to a scanning direction of a light spot on the sample surface. When the defect detection circuit detects that the output signal from the photodetector has a concave peak and a convex peak alternately and continuously occurring in time, a convex or concave defect is present. It is preferable that it is configured to make a determination.
[0040]
One-dimensional diffraction for converting a light beam generated from the light source into a one-dimensional beam array of m (m is a natural number of 2 or more) sub-beams arranged in a line in an optical path between the light source and the beam splitter. A grating is arranged, and the sample surface can be scanned by m light spots formed on the sample.
[0041]
In the optical path between the light source and the beam splitter, when m and n are natural numbers of 2 or more, a two-dimensional beam array of sub beams in which a radiation beam generated from the light source is arranged in a matrix of m rows and n columns A two-dimensional diffraction grating for converting the light spot into a light spot is arranged, and the sample surface can be scanned by a light spot array of m × n light spots.
[0042]
A third aspect of the wafer evaluation apparatus of the present invention includes a sample stage that supports a sample that is a wafer to be inspected, a light source that generates a light beam, and a light spot formed on the sample surface by focusing the light beam. An objective lens, a photodetector that receives reflected light from the surface of the sample, and a light beam that is disposed in an optical path between the light source and the objective lens and travels from the light source toward the sample and a photodetector from the sample. And a beam splitter for separating the reflected beam toward the target, scans the sample surface with a light spot, arranges a light shielding plate in an optical path between the beam splitter and the photodetector, and the light shielding plate An apparatus for evaluating a wafer surface by blocking an optical path on one half in a direction corresponding to a scanning direction of a light spot on the sample surface, wherein the surface information of a specific position on a defective wafer is provided. The output as transformed observed image into two-dimensional image information, characterized by being configured to observe certain defects through the observation image.
[0043]
It is preferable that the contrast of the observation image is corrected in accordance with the type of the defect displayed in the observation image to emphasize the defect.
[0044]
By performing image processing on the image information, the feature of the defect can be extracted and the type of the defect can be automatically classified from the shape and the luminance distribution.
[0045]
It is preferable to display the position and type of the automatically classified defect. A typical example of the wafer to be inspected is a silicon wafer. By designating a specific defect on the sample to be the inspection target wafer, an image of the defect can be output and displayed.
[0046]
With respect to the automatically classified defect, the classified defect is displayed as an image, the type of the defect is confirmed by the operator, and the incorrectly classified defect can be assigned to a proper classification by the operator's judgment. It is even more suitable.
[0047]
BEST MODE FOR CARRYING OUT THE INVENTION
Hereinafter, embodiments of the present invention will be described with reference to the drawings. However, it is needless to say that various modifications can be made without departing from the technical idea of the present invention.
[0048]
FIG. 1 is a schematic explanatory view showing one embodiment of a wafer evaluation apparatus of the present invention. FIG. 2 is a schematic explanatory view showing the basic structure of a laser microscope using a confocal optical system. FIG. 3 is an enlarged sectional explanatory view of a portion A in FIG. FIG. 4 is a flowchart showing an example of a wafer inspection procedure using a laser microscope with a confocal optical system.
[0049]
In FIG. 1, reference numeral 30 denotes an apparatus for evaluating a wafer according to the present invention, which is an optical scanning apparatus using a confocal optical system for inspecting the surface of a sample, in the illustrated example, a wafer W to be inspected, a laser microscope 10 in the illustrated example, and a laser microscope 10 in the illustrated example. And an identification device 40 connected to the laser microscope 10 for identifying defects and foreign substances on the surface of the wafer detected by the laser microscope 10 by type, and a symbol indicating the position of the identified defects and foreign substances by type of each defect and foreign substance. And an output device 50 for displaying an output on a wafer screen W1 corresponding to the inspection target wafer. The operation of the identification device 40 and the output device 50 can be performed by a computer.
[0050]
In FIG. 2, reference numeral 10 denotes an optical scanning device using a confocal optical system, which is a laser microscope in the illustrated example. A laser light source 14 such as an argon laser is provided corresponding to the microscope main body 12.
[0051]
The microscope body 12 includes a laser light source 14, a beam splitter 16, an objective lens 18 for converging a laser beam B on a sample, a wafer W to be inspected in the illustrated example, and a laser beam B reflected from the surface of the wafer W through a pinhole. It comprises a condensing lens 22 that converges on a pinhole 20a of the member 20, and a photodetector 24 that receives the laser beam B passing through the pinhole 20a. In addition, a light shielding plate 26 that shields half of the beam is provided between the beam splitter 16 and the condenser lens 22.
[0052]
The operation principle of such a configuration will be described below. The laser beam B is converged on the surface of the wafer W by the objective lens 18, and irradiates the wafer surface with a spot of, for example, about 0.5 μm.
[0053]
The laser beam B reflected from the surface of the wafer W returns to the optical system, is converged by the condenser lens 22, and enters the photodetector 24 through the pinhole 20a of the pinhole member 20.
[0054]
If the surface of the wafer W has a defect, the wavefront of the reflected light from the defective portion is disturbed, and the spot of the laser beam B is expanded in the photodetector 24, and the light detection signal is reduced.
[0055]
A defect detection circuit (not shown) detects a difference between signals in the photodetector 24, and sets a portion where a signal intensity difference equal to or greater than a set value occurs as a defect portion, and records the size and coordinates of the defect.
[0056]
The inspection is performed while moving at a constant speed, and each beam spot scans the entire wafer W precisely.
[0057]
Subsequently, a procedure for inspecting a sample, for example, a wafer W using the laser microscope 10 by the above-described confocal optical system will be described with reference to a flowchart shown in FIG.
[0058]
The wafer is placed on the wafer cassette and set on the loader (step 200).
[0059]
A recipe such as the order of the wafers to be inspected and the inspection sensitivity is created in the operator console unit (step 201).
[0060]
Alignment is automatically performed, and wafer inspection is continuously performed (step 202).
[0061]
During the wafer inspection, a defect map and a histogram of the inspection result of the inspected area are displayed, and an inspection result file is automatically created (step 203). As a result, defects and foreign substances of 0.06 μm or more are observed. Note that, here, only the “defect map” and the like and the “defect” are expressed, but the defect map naturally includes foreign matter. Hereinafter, a detected object (defect, foreign matter, or the like) obtained using a laser microscope of a confocal optical system may be simply referred to as a “defect” including a foreign matter for convenience.
[0062]
After the completion of the wafer inspection, an image of an arbitrary specified defect portion can be observed from the defect map screen (step 204). When this image is observed, characteristic shapes and image patterns are observed, and each defect can be identified.
[0063]
For example, when the surface of a mirror-polished wafer is observed with a laser microscope using this confocal optical system, defects and foreign matter as shown in FIGS. 9 to 21 are observed. These defects and foreign substances are identified by image processing.
[0064]
In the image processing, the contrast of the obtained grayscale image (input image from the camera) is adjusted so that the defect can be easily identified. The actual procedure corrects the tones so that the difference between the brightest and darkest colors in the resulting image is maximized.
[0065]
Next, features of individual defects obtained by a laser microscope using a confocal optical system will be described. FIGS. 10 to 15 show the forms of defects mainly found in the mirror-polished wafer, FIGS. 18 to 21 show the forms of defects found in the epitaxial wafer, and FIGS. However, the form of the observed defect is shown. This image was observed at a magnification of 7,576. The observation magnification is not particularly limited. However, when the observation is performed at about 8000 times, the shape and the image pattern can be easily recognized and the defect can be easily identified.
[0066]
An optical scanning device using a confocal optical system, for example, as a feature of measurement by the laser microscope 10, when the above-described light shielding plate 26 is provided in the confocal optical system and the imaging system is hidden, the installation position of the light shielding plate 26 is also As shown in FIGS. 2 and 3, if a convex defect F exists on the surface of the wafer W as shown in FIGS. ₁ Light B reflected from ₁ Arrives at the photodetector 24 and is observed brightly. ₂ Light B reflected from ₂ Are observed dark because they cannot pass through the light shielding plate 26. Accordingly, the convex defect F has an image pattern in which the left side is bright and the right side is dark. Conversely, a concave defect has an image pattern in which the left side is dark and the right side is bright. Also, a foreign substance that is transparent to light has an image pattern in which the left side is dark and the right side is bright. The image pattern and the shape of the defect or the foreign matter are mainly used as criteria for identifying the defect or the foreign matter.
[0067]
The defect observed as shown in FIG. 9 is a point-like defect having a concave shape or a minute foreign substance transparent to light. Here, it is called “small defect or small LPD”. These minute defects include conventionally known defects caused by crystals such as COP. This defect is a relatively small (0.5 μm or less) defect. The image data obtained by observing this defect is dark on the left side and bright on the right side. Therefore, a defect can be easily identified by using the image pattern, shape, and defect size (distribution method) as criteria.
[0068]
Further, a point defect having a convex shape as shown in FIG. 10 is also observed. Hereinafter, this defect is referred to as “projection”. These projections are mainly defects caused by processing. The image data obtained from this defect has an image pattern that is bright on the left side and dark on the right side, opposite to the minute defect (small LPD). Defects are easily identified by using the image pattern, shape, and defect size (distribution method) as criteria.
[0069]
Further, defects as shown in FIG. 11 are also observed. Hereinafter, this defect is referred to as “linear projection”. This defect is a defect newly confirmed by the present method, and is characterized by a linear defect having a height of several nm and a length of approximately 0.5 μm or more, unlike crystal defects such as COP. . These linear projections are considered to be mainly defects caused by processing. The image data obtained from this defect is an image pattern in which the same left side as the protrusion is bright and the right side is dark and minute defects are arranged in a straight line. The image pattern, shape (straight line), and its length are determined based on the judgment criteria. Is a defect identified as In particular, such a form of defect has not been observed by other inspection methods, and is one of useful information obtained by the present method. When the linear projection was subjected to an Auger analysis, Ni was detected. Such a defect can be measured even by ordinary particle counter measurement, but it is difficult to separate from the COP, and it is difficult to detect the defect unless it is observed using an AFM or the like. Further, when a wafer having such defects is epitaxially grown, the LPD level of the epi wafer is deteriorated. Further, a decrease in the yield in the device process is expected, and it is necessary to accurately grasp such a form of the defect and to further reduce such a defect.
[0070]
Defects as shown in FIG. 12 are also observed. Hereinafter, this defect is referred to as “pit cluster”. This pit cluster is considered to be a processing-induced defect mainly caused by contamination of heavy metal or the like. When evaluated by a method such as AFM, dents are observed. This defect is not protruding, and the individual minute defects are dark on the left side and bright on the right side, and the minute defects form an aggregate and are not linear.
[0071]
Defects as shown in FIG. 13 are also observed. Hereinafter, this defect is referred to as “linear pit”. These linear pits are considered to be mainly defects caused by processing. This defect is not protruding, and each of the minute defects is dark on the left and bright on the right, and the minute defects are linear.
[0072]
Defects as shown in FIG. 14 are also observed. Hereinafter, this defect is referred to as “scratch”. This flaw is considered to be a defect mainly generated during a processing step or handling. Defects that can be easily identified by using the obtained image data as a criterion based on a continuous image having dark areas on the left and bright areas on the right.
[0073]
Defects as shown in FIG. 15 are also observed. Hereinafter, this defect is referred to as “lap scratch”. This lap scratch is considered to be a defect that is mainly caused by alkali etching of a strain generated in the lapping step and is generated when a polishing allowance is small. When this defect exists, it is identified as dark square image data. It is a clear defect and requires little or no contrast adjustment.
[0074]
Foreign matter as shown in FIG. 16 is also observed. Hereinafter, this foreign substance is referred to as “stain”. It is considered that the stain mainly adheres after a processing step, particularly after washing and drying. This foreign matter is a dark defect whose shape is close to a circle, and is a defect that is identified as a clear defect without adjusting the contrast of an image.
[0075]
Further, in the present method, a foreign matter as shown in FIG. 17 is observed in addition to the defect. Hereinafter, this foreign substance is referred to as “particle”. It is considered that these particles mainly adhere in a processing step. This foreign matter is a dark defect whose shape is not a lap scratch or a stain, and is a defect that is identified as a clear defect without adjusting the contrast of an image.
[0076]
Further, when observing the surface of the epitaxial wafer having the epitaxial layer formed on the surface of the mirror-polished wafer, defects as shown in FIG. 18 are also observed. Hereinafter, this defect is referred to as “SF (Stacking Fault)”. The obtained image data is a defect that can be easily identified by using a square, a rectangle, or a set of two or three small defects included in the rectangle as a criterion.
[0077]
In addition, defects as shown in FIG. 19 are also observed. Hereinafter, this defect is referred to as “dislocation loop”. The obtained image data is a defect which is easily identified by using a bright image pattern on the left side and a dark image pattern on the right side having a certain size as a criterion.
[0078]
In addition, defects as shown in FIG. 20 are also observed. Hereinafter, this defect is referred to as “mound”. This mound is considered to be a defect mainly caused by a foreign substance. The obtained image data is a defect that is clear even if the contrast of the image is not adjusted and is identified as a defect including a part of a square or a rectangle.
[0079]
In addition, a defect whose origin is unknown is also observed as shown in FIG. Hereinafter, this defect is referred to as “EP (epitaxial) defect”.
[0080]
Image processing is performed as identification means for identifying the type of defect or foreign matter based on the above-described determination criteria, so that the defect or foreign matter can be identified, and a map can be represented by a symbol determined for each defect or foreign matter. The symbols used for each defect or foreign matter are not particularly limited, and may be simply alphanumeric characters such as 1, 2, 3, and A, B, and C, in addition to the symbols such as ▼ and □ used herein. May be expressed. Also, the size of the display is not limited, and is set to a size in which the distribution of defects is easily recognizable, and plotted on a wafer screen (defect map) of the output device.
[0081]
Although the algorithm of the analysis by the identification means is not particularly limited, for example, the analysis may be performed in the following procedure.
[0082]
First, the “shape” of the image data of the obtained defect or foreign matter is identified first. This is mainly determined as a point, a circle, a square, or a straight line.
[0083]
Next, the image pattern (right and left light and dark) is confirmed. It is determined whether the obtained defect or foreign matter image is dark on the left and bright on the right, or bright on the left and dark on the right (convex). Thus, the unevenness is estimated. Some of these image patterns cannot be identified.
[0084]
Next, the distribution of the obtained defects is observed. This mainly identifies whether a minute defect exists, whether it exists alone or in an aggregate, and if it exists in an aggregate, whether it exists linearly or circularly.
[0085]
By performing such a determination, the above-described defects and foreign substances can be identified.
[0086]
In addition, there are different defects that are detected in a similar form in PW (mirror polished wafer) and EPW (epitaxial wafer). That is, the lap scratch and the mound have the same form as the defect. These defects may be inspected in different modes depending on whether the inspection target wafer is PW or EPW.
[0087]
If necessary, an application such as displaying only a pattern of a specific defect (for example, a minute defect) as a map is also possible.
[0088]
Hereinafter, an optical scanning device and a defect detection device used in the wafer evaluation device of the present invention will be described in more detail with reference to FIGS. FIG. 22 is a diagram for explaining the principle of the confocal optical system. FIG. 23 is a diagram for explaining the principle of the optical scanning device used in the present invention. 23, the same members as those used in FIG. 22 are denoted by the same reference numerals and described.
[0089]
FIG. 23 (a) shows a state in which a light spot scans an inclined surface that is raised upward, and a convex defect exists on the surface of a sample W which is a wafer to be inspected. Indicates a state in which a gradually lowering inclined surface is scanned. In FIGS. 23 (a) and (b), the light spot scans the sample surface along the direction of the arrow, and the solid line indicates virtual reflected light from a flat sample surface located at the focal plane of the objective lens. The broken line indicates the reflected light from the inclined surface of the defect. In the illustrated optical scanning device, in the present invention, a light shielding plate 107 is disposed in an optical path between the beam splitter 102 and the photodetector 106, and the light in the optical path from the sample W to the photodetector 106 is arranged by the light shielding plate 107. The spot shields the optical path on one side of the front side or the rear side in the scanning direction in which the spot scans the sample surface (in FIG. 23, the front side of the optical path is shielded).
[0090]
When scanning a convex inclined surface where the light spot gradually rises, as shown in FIG. 23A, the reflected beam reflected on the sample surface causes the sample surface to fall on the focal plane of the objective lens according to the law of reflection. It is displaced to the rear side in the scanning direction as compared with the case of scanning the located flat reference plane. On the other hand, when scanning a concave inclined surface where the light spot gradually decreases, as shown in FIG. 23 (b), the reflected beam reflected on the sample surface is compared with the case where the reference surface is scanned by the law of reflection. It is displaced forward in the scanning direction. Therefore, according to the present invention, when the light shielding plate 107 shields the front half of the optical path in the scanning direction, scans a convex inclined surface, and reflects light reflected by the light shielding plate 107 as compared with the case where the reference surface is scanned. The light amount of the beam decreases, and more reflected light enters the photodetector 106. On the other hand, when scanning the concave inclined surface where the light spot gradually decreases, the amount of the reflected beam shielded by the light shielding plate 107 increases compared to when scanning the reference surface, and a smaller amount of reflected light is detected. Incident on the vessel 106.
[0091]
FIG. 24 shows an output signal of the photodetector when the light spot scans the portion where the convex defect exists on the sample surface and an output signal of the photodetector when the light spot scans the portion where the concave defect exists. Curve (a) shows the output signal intensity when scanning a convex defect, curve (b) shows the output signal intensity when scanning a concave defect, and curve (c) shows the case where the light shielding plate 107 does not exist. The output signal intensity when a light spot scans a convex defect or a concave defect is shown. When the light spot scans a portion of the sample surface where a convex defect exists, a convex peak is first generated, and a concave peak is continuously generated temporally. On the other hand, when the light spot scans a portion where a concave defect exists, a concave peak is first generated, and a convex peak is continuously generated temporally. When the light-shielding plate is not used, only a gradually changing concave peak is generated as shown by the curve (c). Therefore, the shape of the defect can be determined from the output signal of the photodetector, and when a convex peak and a concave peak occur alternately, it can be determined that a convex defect exists, and the concave shape can be determined. When a peak and a convex peak occur alternately, it can be determined that a concave defect exists. In addition, when the arrangement position of the light-shielding plate is reversed, that is, when the light-shielding plate is arranged on the rear side in the scanning direction of the optical path, the order of occurrence of the convex peak and the concave peak is only reversed, and the convex defect is When present, a concave peak is generated first, and then a convex peak is generated.
[0092]
FIG. 25 is a diagram showing a configuration of an example of an optical scanning device used in the present invention. The laser beam generated from the laser light source 110 enters the one-dimensional diffraction grating 111. The diffraction grating 111 converts the incident laser beam into n sub beams. It is assumed that these sub-beams are aligned in the paper. The n sub beams emitted from the diffraction grating 111 pass through the first and second relay lenses 112 and 113, and enter the objective lens 115 via the beam splitter 114. The objective lens converges the incident n light beams into a minute spot to form n light spots on the sample to be inspected. The sample stage 117 supporting the sample W is an xy stage having an xy drive mechanism. As the sample stage 117 moves in a direction perpendicular to the plane of the paper, the sample surface is scanned by n light spots.
[0093]
The reflected beam from the light spot formed on the sample W is reflected again by the beam sprinter 114 via the objective lens 115, and is incident on the photodetector 119 via the relay lens 118. The photodetector 119 is a linear image sensor having a plurality of light receiving elements arranged in a line, and reflected beams from each light spot on the sample are incident on the corresponding light receiving element. Each light receiving element of the linear image sensor is separated from each other by a light shielding member, and the light shielding member or its frame functions as a pinhole for restricting light incident on the light receiving element. Therefore, the linear image sensor itself has a spatial filter having a pinhole. In this example, a light shielding plate 120 is arranged in the optical path between the relay lens 118 and the photodetector 119. In addition, for convenience of the drawing, the light shielding plate 120 is shown so as to shield half of one side in a direction orthogonal to the scanning direction of the optical path, but actually, one half of the front side or the rear side in the scanning direction (the direction orthogonal to the paper surface). Shall be shielded from light.
[0094]
FIG. 26 is a diagram showing a configuration of an example of a defect detection device used in the present invention. The laser beam generated from the laser light source 130 is incident on the diffraction grating 131, and is converted into a plurality of light beams aligned in the first direction. It is assumed that the diffracted light is diffracted into the paper for clarity of the drawing. These light beams enter the polarization beam splitter 134 via the first and second relay lenses 132 and 133, pass through the polarization beam splitter 134, and enter the carbano mirror 135. The galvanomirror 135 periodically deflects the incident light beam at a predetermined frequency in a second direction orthogonal to the first direction (scanning direction orthogonal to the paper surface). The plurality of light beams reflected by the galvanomirror enter the objective lens 139 via the third and fourth relay lenses 136 and 137 and the quarter-wave plate 138. The objective lens focuses a plurality of incident light beams into a minute spot and projects the light beam on a sample W to be inspected for defects. Therefore, a plurality of minute light spots are formed on the sample W in a line along the direction corresponding to the first direction. Since these light spots are deflected in a second direction orthogonal to the direction of the light spot row, the sample W is scanned by the plurality of light beams, and thus the sample W is two-dimensionally scanned by the plurality of light beams. .
[0095]
The sample W is arranged on a stage 141 having an xy drive mechanism. The stage 141 moves at a predetermined speed in a direction (first direction) orthogonal to the direction of deflection of the galvanomirror. Therefore, the sample is two-dimensionally scanned by the plurality of light spots. The reflected light from each light spot on the sample is condensed by the objective lens 139, and enters the galvanomirror 135 via the quarter-wave plate 138 and the relay lenses 137 and 136. Then, the light is descanned by the galvanomirror and is incident on the polarization beam splitter 134. Since the incident reflected beam has passed through the quarter-wave plate 138 twice, its polarization plane is rotated by 90 °. As a result, the reflected beam from the sample reflects off the plane of polarization of the polarizing beam splitter and is separated from the illumination beam from the light source. The reflected beam emitted from the polarizing beam splitter enters the linear image sensor 143 via the relay lens 142. The linear image sensor 143 has a plurality of light receiving elements arranged in a line along a direction corresponding to the light spot on the sample.
[0096]
An optical system between the sample and the linear image sensor forms an imaging optical system, and a light shielding plate 144 is arranged in an optical path between the polarizing beam splitter 134 and the photodetector 143. Although the light-shielding plate 144 is illustrated in the drawing so as to shield the optical path in the direction perpendicular to the scanning direction of the galvanometer mirror, the light-shielding plate 144 actually shields the optical path of one half of the optical path in the scanning direction of the galvanometer mirror which is the second direction. To be arranged. The light blocking member or frame that separates each light receiving element of the linear image sensor functions as a spatial filter having a pinhole that regulates a light beam incident on the light receiving element.
[0097]
Since the reflected beam from each light spot on the sample W is descanned by the galvanometer mirror 135, the reflected beam from each light spot on the sample enters each corresponding light receiving element of the linear image sensor 143, and stops. Maintained in state. The output signal from each light receiving element of the photodetector is amplified by the amplifier 145 and supplied to the defect detection circuit 146.
[0098]
FIG. 27 is a circuit diagram showing a configuration of an example of the defect detection circuit. In the present example, a first comparator 150 for comparing an output signal from each light receiving element of the photodetector 143 with a negative limit value and a second comparator 151 for comparing with a positive limit value are provided. The outputs of the first and second comparators are connected to an OR circuit 153. As described above, when a positive peak and a negative peak alternately occur from the light receiving element, the defect is determined as a convex defect, and when a negative peak and a positive peak alternately occur, the defect is determined as a convex defect. I do. If only a negative peak occurs, it is determined that the defect is a steep slope. If only a positive peak is generated, it can be determined that a substance having a high reflectance is locally attached. Therefore, by using this defect detection device, not only the existence of the defect but also the shape and properties of the defect can be determined, and the processing process after the inspection can be accurately determined.
[0099]
It should be noted that a video signal can be generated by reading out the electric charge accumulated in each light receiving element of the photodetector at a predetermined subsequent frequency, so that the optical system shown in FIG. 26 can be used as a microscope.
[0100]
FIG. 28 is a diagram showing a modification of the optical scanning device used in the present invention. The radiation beam emitted from the light source 161 is reflected by the total reflection mirror 162 and enters the diffraction grating 163. This diffraction grating converts an incident radiation beam into a two-dimensional beam array arranged in a matrix of m rows and n columns (m and n are natural numbers of 2 or more). These m × n light beams form a two-dimensional beam array composed of light beams that are equally spaced in the row and column directions. The m × n light beams enter the polarization beam splitter 165 via the Fourier transform lens 164. The beam array passes through the polarization beam splitter 165 and enters the galvanometer mirror 166. The galvanomirror 166 deflects the m × n light beams at a predetermined frequency in a first direction (main scanning direction). The light beam reflected by the galvanomirror enters the objective lens 169 via the relay lens 167 and the quarter-wave plate 168. The objective lens 169 focuses the light beam array of m rows and n columns into a minute spot to form a light spot array of m rows and n columns on the sample W.
[0101]
FIG. 29 is a diagram showing the relationship between the light spot array formed on the sample W and the beam deflection direction (first direction) of the galvanomirror, that is, the main scanning direction. In FIG. 29, a white circle indicates a light spot, and a black circle indicates projection of the light spot on an axis L in a direction orthogonal to the main scanning direction. For clarity of the drawing, a 4-row, 4-column light spot array is shown. The interval between the light spots in each row direction is p1, and the interval between the light spots in the column direction is p2. In order to scan the sample surface without gaps and without overlapping light spots with the m-row and n-column light spot array, the light spots are projected at equal intervals so that the light spots are projected at equal intervals on the axis perpendicular to the main scanning direction. Need to be formed. Hereinafter, this condition will be described. An angle between the axis Lc in the row direction of the light spot array and the axis L in the main scanning direction is defined as θ. The projection length of the interval p1 between the light spots in the row direction with respect to the axis L is p1 × cos θ. Within this projection length, there must be projections of n light spots in the column direction. This condition can be expressed by p1 × cos θ = p2 × sin θ. Therefore, the following equation (1):
[0102]
(Equation 5)
tan θ = (1 / n) × (p1 / p2) (1)
[0103]
When is satisfied, projections of the light spots of the light spot array of m rows and n columns in a direction orthogonal to the main scanning direction are formed at equal intervals. When the intervals between the light spots in the row and column directions are equal (p1 = p2), the light spot array is formed so as to satisfy the following expression (2).
[0104]
(Equation 6)
tan θ = 1 / n (2)
[0105]
Since the light beam array is scanned in the main scanning direction by the galvanometer mirror, the sample surface is scanned by the light spot array arranged in a matrix of m rows and n columns. The stage 171 supporting the sample W is an xy stage, and moves at a predetermined speed in a direction orthogonal to the main scanning direction which is the first direction. With this configuration, the sample surface is two-dimensionally scanned by m × n light spots.
[0106]
The reflected light from the light spot reflected on the sample surface is again incident on the galvanometer mirror 166 via the objective lens 169, the quarter-wave plate 168 and the relay lens 167, is descanned, and is incident on the beam splitter 165. Since this reflected beam has passed through the quarter-wave plate twice, the polarization plane is rotated by 90 °, and is reflected by the polarization plane of the beam splitter. The reflected beam reflected by the beam splitter enters the photodetector 175 via the relay lens 172, the total reflection mirror 172, and the relay lenses 173 and 174.
[0107]
FIG. 30 is a diagram showing a configuration of an example of the photodetector. The photodetector 175 has a two-dimensional matrix of light receiving element arrays 176 (i, j) formed at equal intervals in the row and column directions, and each light receiving element is a positive light from each light spot formed on the sample. Each receives the reflected light. Each light receiving element is constituted by a photodiode, and is separated from each other by a light shielding member 177. The light incident area of each light receiving element is regulated by the light blocking member 176 so that only the specularly reflected light of the light spot formed on the sample surface is incident. Therefore, the optical system of this example constitutes a confocal optical system, and can obtain higher resolution.
[0108]
A light shielding plate 180 is arranged at a pupil position of the imaging optical system between the beam splitter 165 and the photodetector 175. The light-shielding plate shields an optical path on one side in a direction corresponding to the main scanning direction on the sample. By arranging one light shielding plate at the pupil position, one side of the m × n reflected beams can be shielded.
[0109]
The present invention is not limited to the above-described embodiment, and various changes and modifications are possible. For example, in the above-described embodiment, an apparatus for detecting a defect existing on the surface of a sample such as a semiconductor wafer or a photomask blank has been described as an example. However, an optical scanning apparatus used in the present invention is, for example, a laser microscope. The present invention can also be applied to an imaging device that scans a sample surface with a light beam and captures a sample image, and has a higher resolution by arranging a light shielding plate in an optical path of an imaging optical system from the sample to a photodetector. An imaging device can be realized. For example, a video signal can be easily formed by reading out the charge of each light receiving element at a predetermined readout station wave number from the photodetector of the device shown in FIG.
[0110]
In addition, various defect detection circuits can be used as the defect detection circuit. For example, it is possible to compare output signals of adjacent light receiving elements and detect the occurrence of a defect from the comparison result.
[0111]
Further, in the above-described embodiment, an example in which the sample surface is raster-scanned has been described. However, as means for relatively moving the sample and the light beam array, the sample surface is connected to a sample stage supporting the sample, and the stage is rotated. A stage drive having a rotary drive and a translation drive for translating in a direction perpendicular to the axis of rotation can be used. In this case, by rotating and translating the sample stage, the sample is spirally scanned by the m × n-valued light spot array, so that the sample can be scanned at higher speed and the sample surface can be scanned with high resolution. .
[0112]
Next, an example of an identification device for identifying the type of defect or foreign matter will be described. First, the contrast correction of an image will be described. In the captured original image, when the minimum luminance value is a and the maximum luminance value is b, the luminance value z of each pixel is set so that the minimum luminance value after conversion is a ′ and the maximum luminance value is b ′. On the other hand, conversion is performed as described below to obtain a contrast-corrected luminance value z ′ represented by the following equation (3).
[0113]
(Equation 7)

[0114]
Usually, a ′ ≦ a <b ≦ b ′, and in the case of an 8-bit grayscale image, a ′ = 0 and b ′ = 255.
[0115]
Hereinafter, automatic classification of defect types will be described with reference to FIG. FIG. 31 is a flowchart showing an example of a procedure for automatically classifying defect types, and shows a procedure for classifying minute LDPs, projections, linear projections, scratches, and stacking faults. Note that, in FIG. 31, the discrimination processing for other defect types and the processing when the condition is not satisfied are omitted.
[0116]
First, noise removal processing by smoothing is performed (step 300). Next, a defect candidate area is extracted by edge extraction (step 302). Subsequently, the determination process is assigned based on the number, size, and distribution of the defect candidate regions (step 304). In this allocation step, the defect candidate areas are divided into three groups: (1) one small candidate area, (2) one large candidate area, and (3) two slightly larger candidate areas. Can be
[0117]
{Circle around (1)} When there is one small candidate area, a bright area and a dark area in the defect candidate area are extracted (step 306). Next, the defect type is determined based on the size and positional relationship of the bright and dark areas (step 308). In this determination, if the bright area is on the right and the dark area is on the left, it is determined to be a minute defect, and if the bright area is on the left and the dark area is on the right, it is determined to be a protrusion.
[0118]
{Circle around (2)} In the case where there is one large candidate area, the series of the midpoints of the bright area and the dark area in the defect candidate area is linearly approximated, and the defect type is determined based on the approximation error (step 310). In this determination, when the approximation error is small, the bright region is on the left and the dark region is on the right, it is determined to be a linear protrusion. When the approximation error is small, the bright region is on the right and the dark region is on the left, Is judged to be scratched.
[0119]
{Circle around (3)} When there are two slightly larger candidate regions, the defect type is determined based on the shapes and positional relationships of the two defect candidate regions (step 312). In this determination, a stacking fault is determined when two vertically-long rectangles are paired in parallel. For example, the type of defect or foreign matter is selected by an identification device programmed with such a processing procedure, and the position and type of the defect on the wafer are displayed by an output device.
[0120]
【Example】
Hereinafter, the present invention will be described more specifically with reference to examples. It should be understood that these examples are given by way of illustration and should not be construed as limiting.
[0121]
(Example 1) Evaluation of PW wafer
Using a wafer evaluation apparatus of the present invention as shown in FIG. 1 for a silicon wafer having a diameter of 300 mm, defects of 0.06 μm or more were evaluated. This device is an optical system device as shown in FIG.
[0122]
The result is shown in FIG. As a result, in the silicon wafer evaluated this time, the number of display defects was 279, and six types of defects and foreign matters were observed. In particular, it can be seen that this wafer has many small defects (such as COP which is a crystal defect) distributed near the center of the wafer. As a result, feedback to the manufacturing site, such as a review of the crystal manufacturing conditions, can be provided.
[0123]
(Comparative Example 1)
The same wafer as in Example 1 was evaluated for defects of 0.065 μm or more by using a laser scattering foreign substance inspection apparatus.
[0124]
As shown in FIG. 7, the number of detected defects is small. Further, it was not possible to distinguish the form (type) of the defect.
[0125]
(Example 2) Evaluation of EPW wafer
An epitaxial wafer having an epitaxial layer formed on a silicon wafer having a diameter of 300 mm was measured using the same apparatus of the present invention as in Example 1.
[0126]
As a result, as shown in FIG. 6, on the surface of the wafer evaluated this time, the number of display defects was 25, and SF, mound, and the like were observed.
[0127]
(Comparative Example 2)
The defect of 0.12 μm or more was evaluated for the same epitaxial wafer as in Example 2 using a laser scattering type foreign matter inspection apparatus.
[0128]
As shown in FIG. 8, the number of detected defects is small. Further, it was not possible to distinguish the form (type) of the defect.
[0129]
Note that the present invention is not limited to the above embodiment. The above embodiment is an exemplification, and has substantially the same configuration as the technical idea described in the claims of the present invention, and any device having the same operation and effect can be realized by the present invention. Included in the technical scope.
[0130]
【The invention's effect】
As described above, according to the present invention, the distribution state of defects and foreign substances on the wafer and the components of each defect and foreign substance can be determined at a glance, and a problematic process or apparatus can be estimated from the distribution state of defects or foreign substances. The effect is achieved.
[Brief description of the drawings]
FIG. 1 is a schematic explanatory view showing one embodiment of a wafer evaluation device of the present invention.
FIG. 2 is a schematic explanatory view showing a basic structure of a laser microscope using a confocal optical system.
FIG. 3 is an enlarged sectional explanatory view of a portion A in FIG. 2;
FIG. 4 is a flowchart showing an example of a wafer inspection procedure using a laser microscope with a confocal optical system.
FIG. 5 is an image showing defects observed by type in Example 1.
FIG. 6 is an image showing defects observed by type in Example 2.
FIG. 7 is an image showing a defect observed in Comparative Example 1.
FIG. 8 is an image showing a defect observed in Comparative Example 2.
FIG. 9 is a photograph showing a first example of a defect observed on a wafer surface by a laser microscope using a confocal optical system.
FIG. 10 is a photograph showing a second example of a defect observed on a wafer surface by a laser microscope using a confocal optical system.
FIG. 11 is a photograph showing a third example of a defect observed on a wafer surface by a laser microscope using a confocal optical system.
FIG. 12 is a photograph showing a fourth example of a defect observed on a wafer surface by a laser microscope using a confocal optical system.
FIG. 13 is a photograph showing a fifth example of a defect observed on a wafer surface by a laser microscope using a confocal optical system.
FIG. 14 is a photograph showing a sixth example of a defect observed on a wafer surface by a laser microscope using a confocal optical system.
FIG. 15 is a photograph showing a seventh example of a defect observed on a wafer surface by a laser microscope using a confocal optical system.
FIG. 16 is a photograph showing an eighth example of a defect observed on a wafer surface by a laser microscope using a confocal optical system.
FIG. 17 is a photograph showing a ninth example of a defect observed on a wafer surface by a laser microscope using a confocal optical system.
FIG. 18 is a photograph showing a tenth example of a defect observed on a wafer surface by a laser microscope using a confocal optical system.
FIG. 19 is a photograph showing an eleventh example of a defect observed on a wafer surface by a laser microscope using a confocal optical system.
FIG. 20 is a photograph showing a twelfth example of a defect observed on a wafer surface by a laser microscope using a confocal optical system.
FIG. 21 is a photograph showing a thirteenth example of a defect observed on a wafer surface by a laser microscope using a confocal optical system.
FIG. 22 is a diagram illustrating the principle of a confocal optical system.
FIG. 23 is a diagram for explaining the principle of an optical scanning device used in the present invention.
FIG. 24 is a diagram illustrating an optical path of a reflected beam when a light spot scans a defective portion.
FIG. 25 is a diagram showing a configuration of an example of an optical scanning device used in the present invention.
FIG. 26 is a diagram showing a configuration of an example of a defect detection device used in the present invention.
FIG. 27 is a diagram illustrating a configuration of an example of a defect detection circuit.
FIG. 28 is a diagram showing a modification of the optical scanning device used in the present invention.
FIG. 29 is a diagram illustrating a relationship between a light spot array formed on a sample and a main scanning direction.
FIG. 30 is a plan view illustrating an example of a photodetector.
FIG. 31 is a flowchart showing an example of a procedure for automatically classifying defect types.
[Explanation of symbols]
10: Laser microscope with confocal optical system, 12: Microscope main body, 14: Laser light source, 16, 102: Beam splitter, 18, 103: Objective lens, 20a: Pinhole, 20: Pinhole member, 22: Condensing lens , 24, 106: photodetector, 26: light shielding plate, 30: wafer evaluation device of the present invention, 40: identification device, 50: output device, B: laser beam, F: defect, W: wafer (sample), W1: wafer screen, 101: light source, 105: spatial filter.

Claims (26)

The surface of the sample to be inspected is inspected by an optical scanning device using a confocal optical system, defects and foreign substances on the sample surface detected by the inspection are identified by type, and the identified defects and foreign substances are identified. The position of the defect is displayed on a wafer screen corresponding to the sample by a symbol indicating the type of each defect or foreign matter. 2. The wafer evaluation method according to claim 1, wherein a defect or foreign matter on the surface of the sample detected by an optical scanning device using a confocal optical system is classified according to its shape and its type based on its shape and brightness. 3. The method for evaluating a wafer according to claim 1, wherein the wafer to be inspected is a silicon wafer. The method for evaluating a wafer according to any one of claims 1 to 3, wherein the optical scanning device using the confocal optical system is a laser microscope using the confocal optical system. An optical scanning device using a confocal optical system for inspecting the surface of a sample to be inspected, and a defect or foreign matter present on the sample surface by analyzing image data of the sample surface obtained by the optical scanning device And an output device for displaying the position of the identified defect or foreign matter on a wafer screen corresponding to the sample by a symbol indicating the type of each defect or foreign matter. Wafer evaluation device. 6. The wafer evaluation apparatus according to claim 5, wherein the optical scanning device using the confocal optical system is a laser microscope using the confocal optical system. An optical scanning device using the confocal optical system,
A sample stage supporting the sample,
A light source for generating a light beam;
An objective lens that focuses the light beam to form a light spot on the sample surface;
A photodetector that receives reflected light from the sample surface;
A beam splitter disposed in an optical path between the light source and the objective lens, for separating a light beam from the light source toward the sample and a reflected beam from the sample toward the photodetector;
Means for relatively moving the sample and the light spot,
An optical scanning device that scans the sample surface with a light spot and detects information of a surface area of the sample by reflected light from the sample surface,
A light-shielding plate is arranged in an optical path between the beam splitter and the photodetector, and the light-shielding plate shields a half of the optical path in a direction corresponding to a scanning direction of a light spot on the sample surface. 6. The wafer evaluation device according to claim 5, wherein: 8. The wafer evaluation apparatus according to claim 7, wherein the means for relatively moving the light spot and the sample is a means for moving the sample stage with respect to the light spot. 8. The wafer evaluation according to claim 7, wherein the means for relatively moving the light spot and the sample is a beam deflector arranged in an optical path width between the beam splitter and the sample stage. apparatus. One-dimensional conversion of a light beam generated from the light source into a one-dimensional beam array of m (m is a natural number of 2 or more) sub-beams arranged in a line in an optical path between the light source and the beam splitter. 8. The wafer evaluation apparatus according to claim 7, wherein a diffraction grating is arranged, and the surface of the sample is scanned by m light spots formed on the sample. In the optical path between the light source and the beam splitter, when m and n are natural numbers of 2 or more, a two-dimensional sub-beam of a radiation beam generated from the light source is arranged in a matrix of m rows and n columns. A two-dimensional diffraction grating for converting into a beam array is arranged, the surface of the sample is scanned by an m × n two-dimensional light spot array, and the light shielding plate is connected to an optical system between the beam splitter and the photodetector. The wafer evaluation apparatus according to claim 7, wherein the wafer evaluation apparatus is arranged at a position of a pupil. As means for relatively moving the sample and the two-dimensional light spot array, a rotation driving device for rotating the sample stage and a translation driving device for moving the sample stage along an axis orthogonal to the rotation axis are used. The apparatus for evaluating a wafer according to claim 11, wherein the sample is spirally scanned by a two-dimensional light spot array. An optical scanning device using the confocal optical system,
A sample stage supporting the sample,
A light source for generating a radiation beam;
a two-dimensional diffraction grating that converts the radiation beam into a beam array of light beams arranged in a matrix of m rows and n columns, where m and n are natural numbers of 2 or more;
A beam deflecting device for deflecting the beam array in a first scanning direction;
An objective lens for converging the light beams of the beam array into spots to form an array of m × n light spots on a sample;
A beam splitter disposed in an optical path between the light source and the objective lens, for separating a light beam from the light source toward the sample and a reflected beam from the sample toward a photodetector;
When m ′ and n ′ are natural numbers of 2 or more, the light receiving elements are arranged in a matrix of m ′ rows and n ′ columns, and each light receiving element reflects reflected light from a light spot formed on the sample. A photodetector for receiving light,
A light-shielding plate arranged at a pupil position in an optical path between the beam splitter and the detector, and shielding a half-side optical path in a direction corresponding to a scanning direction of a light spot on the sample surface;
The reflected beam from the sample surface is incident on the light receiving element of the photodetector via the beam deflector,
A structure in which the light spot array formed on the sample is formed such that when these light spots are projected in a direction orthogonal to the first direction, intervals between adjacent light spots are equal. The wafer evaluation device according to claim 5, wherein: The angle between the axis in the row direction of the light spot array of m rows and n columns formed on the sample and the axis in the direction orthogonal to the first direction is θ, and the spot interval in the row direction of the light spot array is When p1 and the spot interval in the column direction are p2, the following equation (1) is obtained.

14. The wafer evaluation apparatus according to claim 13, wherein the angle θ is set so as to satisfy the following. The intervals p1 and p2 between the light spots in the row and column directions of the light spot array are set so that p1 = p2, and the following equation (2) is set.

14. The wafer evaluation apparatus according to claim 13, wherein the light spot array is formed so as to satisfy the following. The surface of the sample to be inspected is inspected using an optical scanning device based on a confocal optical system, image data of the sample surface obtained by the optical scanning device is analyzed, and defects existing on the sample surface are analyzed. A defect detection device for detecting foreign matter; an identification device for identifying the type of defect or foreign matter present on the sample surface obtained by the defect detection device; and a position of the detected defect or foreign matter for each defect or foreign matter. A wafer evaluation device having an output device for displaying an output on a wafer screen corresponding to the sample by a symbol indicating the type of the wafer. The defect detection device,
A sample stage that supports a sample that is the inspection target wafer,
A light source for generating a light beam;
An objective lens that focuses the light beam to form a light spot on a sample surface;
Means for relatively moving the sample and the light spot;
A photodetector that receives reflected light from the sample surface;
A beam splitter disposed in an optical path between the light source and the objective lens for separating a light beam from the light source toward the sample and a reflected beam from the sample toward the photodetector;
A defect detection circuit that detects a defect present on the sample surface from an output signal from the photodetector,
A light-shielding plate is disposed in an optical path between the beam splitter and the photodetector, and blocks a half-side optical path in a direction corresponding to a scanning direction of a light spot on the sample surface,
When the defect detection circuit detects that the output signal from the photodetector has a concave peak and a convex peak that occur alternately successively in time, a convex or concave defect is present. 17. The wafer evaluation apparatus according to claim 16, wherein the apparatus is configured to make a determination. One-dimensional diffraction for converting a light beam generated from the light source into a one-dimensional beam array of m (m is a natural number of 2 or more) sub-beams arranged in a line in an optical path between the light source and the beam splitter. 18. The wafer evaluation apparatus according to claim 17, wherein a grating is arranged, and the sample surface is scanned by m light spots formed on the sample. In the optical path between the light source and the beam splitter, when m and n are natural numbers of 2 or more, a two-dimensional beam array of sub beams in which a radiation beam generated from the light source is arranged in a matrix of m rows and n columns 20. The wafer evaluation apparatus according to claim 18, wherein a two-dimensional diffraction grating for converting into a light is arranged, and the sample surface is scanned by a light spot array of m × n light spots. A sample stage that supports a sample that is a wafer to be inspected, a light source that generates a light beam, an objective lens that focuses the light beam to form a light spot on the sample surface, and receives reflected light from the sample surface And a beam splitter disposed in an optical path between the light source and the objective lens to separate a light beam from the light source toward the sample and a reflected beam from the sample toward the photodetector. Scanning the sample surface with a light spot, arranging a light shielding plate in an optical path between the beam splitter and the photodetector, and using the light shielding plate, a direction corresponding to a scanning direction of the light spot on the sample surface. This device evaluates the wafer surface by blocking one half of the optical path, and converts the surface information of a specific position on the defective wafer into two-dimensional image information and outputs it as an observation image. And evaluation device of the wafer, characterized by being configured to allow observation of the specific defect through the observation image. 21. The wafer evaluation apparatus according to claim 20, wherein the defect is emphasized by correcting the contrast of the observed image according to the type of the defect displayed in the observed image. 22. The wafer evaluation apparatus according to claim 20, wherein by performing image processing on the image information, a feature of the defect is extracted, and a type of the defect can be automatically classified based on a shape and a luminance distribution. 23. The wafer evaluation apparatus according to claim 22, wherein the position and type of the automatically classified defect are displayed. 24. The wafer evaluation device according to claim 20, wherein the inspection target wafer is a silicon wafer. The wafer according to any one of claims 20 to 24, wherein an image of the defect can be output and displayed by designating a specific defect with respect to the sample to be the inspection target wafer. Evaluation device. With respect to the automatically classified defects, the classified defects are image-displayed, the type of the defect is confirmed by the operator, and the incorrectly classified defect can be assigned to a valid classification by the operator's judgment. The wafer evaluation apparatus according to any one of claims 22 to 25, wherein:

Images